Presentation type:
NP – Nonlinear Processes in Geosciences

EGU24-10769 | Orals | MAL27-NP | Lewis Fry Richardson Medal Lecture

On a few characteristics of geophysical turbulent flows  

Annick Pouquet

Our understanding of turbulence has progressed significantly through combining experiments, observations, theoretical developments, direct numerical simulations and modeling. I will discuss briefly three problems (among many) for which our perception has changed: (i) the derivation of a multitude of exact laws stemming from conservation properties, e.g. in fluid, magnetohydrodynamics (MHD) and Hall-MHD turbulence, and their consequences for constraining scaling relations and dynamical evolutions; (ii) the cascade processes of energy in three dimensional strongly rotating stratified turbulent flows (RST) found to be dual, in the sense that the energy can go in a self-similar manner both to the large scales and to the small scales with (different) contant fluxes, a phenomenon also encountered in MHD, including in solar wind observations; and (iii) turbulent fields themselves (velocity, induction, temperature), together with their gradients, can be intermittent with non-Gaussian wings, as in quantum turbulence and shear flows or in MHD and RST.

I will also mention old and new results concerning the propensity for turbulent flows and nonlinear systems to develop sharp, isolated (intermittent) structures at small and large scales in a variety of physical environments. This will be done in the specific context of normalized moments at third-order (skewness S) and fourth-order (kurtosis K). Indeed, intermittency can be evaluated e.g. through the examination of relations between S and K for fields such as the velocity, temperature and magnetic fields as well as for their local rates of dissipation. The field themselves, in general, have small skewness, but in some cases they display high kurtosis, such as for vertical velocities in RST, as observed in the stable nocturnal planetary boundary layer, as well as for the magnetic field in the solar wind, or more recently in the fast dynamo regime in MHD. On the other hand, the local dissipation rates of these fields follow a parabolic K(S) law whose origin may be linked to kinematic constraints, to the applicability of Langevin models to their dynamics, or to self-organized criticality, as suggested by several authors in various physical contexts, from the atmosphere, the ocean and climate, to fusion plasmas, the solar wind and dwarf galaxies [1,2].

Many thanks to all my collaborators, mentors, colleagues, students and post-docs.

[1] Annick Pouquet, Duane Rosenberg, Raffaele Marino and Pablo Mininni: Intermittency Scaling for Mixing and Dissipation in Rotating Stratified Turbulence at the Edge of Instability. Atmosphere 14, 1375 (2023). Special issue in honor of Jack Herring; B. Galperin, A. Pouquet & P. Sullivan Eds..

[2] Yannick Ponty, Hélène Politano and Annick Pouquet: Spatio-temporal intermittency assessed through kurtosis-skewness relations in MHD in the fast dynamo regime. In preparation (2024).

How to cite: Pouquet, A.: On a few characteristics of geophysical turbulent flows , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-10769, https://doi.org/10.5194/egusphere-egu24-10769, 2024.

EGU24-11136 | Orals | MAL27-NP | NP Division Outstanding Early Career Scientist Award Lecture

Exploring space plasma fluctuations at kinetic scales through stochastic process theory 

Simone Benella

During the last decades, space missions provided in situ data of diverse space plasma environments with increasingly higher resolution. This enabled the possibility to investigate peculiar properties of fluctuations in the magnetic field and plasma parameters, transitioning from the magnetohydrodynamic (MHD) to the ion-kinetic regime. The ion-kinetic regime is characterized by a global self-similar scaling of fluctuations, in contrast to the local scale-invariance of the MHD ones. In a series of works, we developed a data-driven approach based on the Langevin equation in order to model statistical features of kinetic fluctuations. In practical terms, the stochastic process thus introduced represents the evolution of the magnetic field fluctuations as a function of the scale. As far as such fluctuations are of the Langevin type, their statistics evolve according to a Fokker-Planck equation. Studying the evolution of fluctuation statistics across the scales, e.g., structure functions, allows us to make predictions about global statistical properties, e.g., scaling exponents.

In this work, we review recent results obtained by using data from the ESA/Cluster mission in near-Earth space. We give evidence that the dynamics of magnetic field increments at kinetic scales can be modeled as a stochastic process of the Langevin type, and that the correct scaling law of the structure functions can be obtained through the stochastic equation in the non-diffusive limit, by linking the drift term of the Langevin equation to the Hölder exponent. Finally, this model allows us to derive the asymptotic limit of individual sets of fluctuations, giving thus predictions on the trend expected at kinetic scales.

How to cite: Benella, S.: Exploring space plasma fluctuations at kinetic scales through stochastic process theory, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-11136, https://doi.org/10.5194/egusphere-egu24-11136, 2024.

NP1 – Mathematics of Planet Earth

EGU24-119 * | ECS | Orals | NP1.1 | Highlight

Advancements in Medicanes Tracking and Forecasting Using Artificial Intelligence 

Javier Martinez-Amaya, Veronica Nieves, and Jordi Muñoz-Marí

Medicanes, tropical-like cyclones in the Mediterranean Sea, pose unexpected challenges to unprepared areas due to their projected increases in intensity. To address these challenges, we proposed: 1) the development of an automatic tracking method in the absence of a comprehensive tracking database for Medicanes; 2) the implementation of a forecasting model for extreme cyclones utilizing artificial intelligence techniques. This is especially beneficial when traditional numerical models struggle to account for nonlinear interactions. We use a K-means algorithm and mean sea level pressure reanalysis data to track storm centers, determining maximum wind speed and position throughout each case’s lifetime. This information categorizes our dataset into storm-like and extreme Medicanes, and facilitates the extraction of spatiotemporal data from infrared satellite images. These features enable us to predict the final classification of Medicanes (whether they are storm-like or extreme) 6 to 36 h before peak wind speed, using an optimized combination of Convolutional Neural Network and Random Forest binary classification methods. By training and testing on Mediterranean data from 1984 to 2020, we successfully diagnosed between 72% and 87% of extreme Medicanes in the studied cases, depending on the lead-time. Our study is the first to employ artificial intelligences for both tracking and forecasting Medicanes, offering a foundational approach to enhance Medicanes preparedness and awareness.

How to cite: Martinez-Amaya, J., Nieves, V., and Muñoz-Marí, J.: Advancements in Medicanes Tracking and Forecasting Using Artificial Intelligence, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-119, https://doi.org/10.5194/egusphere-egu24-119, 2024.

EGU24-1008 | ECS | Posters on site | NP1.1

Statistical properties for a spatially heteregeneous one-dimensional energy balance model perturbed by additive noise 

Gianmarco Del Sarto and Franco Flandoli

A one-dimensional energy balance model (1D-EBM) is a simplified climate model that describes the evolution of Earth's temperature based on the planet's energy budget.

In this study, we examine a 1D-EBM that incorporates a parameter representing the impact of carbon dioxide on the energy balance. Based on empirical studies showing that bistability may occur in Earth's tropics, we consider the planet's ongoing radiation to be latitude-dependent, presenting bistability in low-latitude regions but not in high-latitude ones. This local bistability does not lead to a bifurcation in the entire system, in addition to the classical saddle-node bifurcations between Snowball Earth and the present climate.

We focus on investigating the statistical properties of the system when the model is perturbed with additive noise. Our work is a step towards a clearer understanding of the dynamics in a spatially heterogeneous setting.

How to cite: Del Sarto, G. and Flandoli, F.: Statistical properties for a spatially heteregeneous one-dimensional energy balance model perturbed by additive noise, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-1008, https://doi.org/10.5194/egusphere-egu24-1008, 2024.

The introduction of random perturbations by noise in partial differential equations has proven extremely useful to understand more about long-time behaviour in complex systems like atmosphere and ocean dynamics or global temperature. Considering additional transport by noise in fluid models has been shown to induce convergence to stationary solutions with enhanced dissipation, under specific conditions. On the other hand the presence of simple additive forcing by noise helps to find a stationary distribution (invariant measure) for the system and understand how this distribution changes with respect to changes in model parameters (response theory). I will discuss these approaches with a multi-layer quasi-geostrophic model as example.

How to cite: Carigi, G.: Long-time behaviour of stochastic geophysical fluid dynamics models., EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-1133, https://doi.org/10.5194/egusphere-egu24-1133, 2024.

EGU24-1148 | Orals | NP1.1

Data assimilation and the reconstruction of surface fluxes in quasigeostrophic and transport equations. 

Jochen Bröcker, Giulia Carigi, Tobias Kuna, and Vincent Ryan Martinez

In this contribution, we will analyse simple data assimilation schemes that not only estimate the underlying states of a dynamical system but simultaneously reconstruct unknown components of the dynamics. We focus on quasigeostrophic and transport-diffusion equations (for instance for atmospheric aerosols or tracer gases) and reconstruct forcings or surfacd fluxes, along with the underlying dynamical states. Tracer gases and aerosols play an important role in the dynamics of the atmosphere; aerosols for instance act as condensation nuclei and thus have a major influence on precipitation, while tracer gases such as ozone, methane, or CO2 impact the radiative transfer and are thus linked to important atmospheric phenomena such as the ozone hole and the energy budget of the planet ("greenhouse effect"), respectively. Furthermore, gases as well as aerosols (especially in the lower troposphere) are common pollutants with strong and potentially adverse effects on the environment, human activity, and health. We discuss two algorithms that both apply in the context of the quasigeostrophic as well as the transport-diffusion equations.

How to cite: Bröcker, J., Carigi, G., Kuna, T., and Martinez, V. R.: Data assimilation and the reconstruction of surface fluxes in quasigeostrophic and transport equations., EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-1148, https://doi.org/10.5194/egusphere-egu24-1148, 2024.

EGU24-1469 | Posters on site | NP1.1

On the hydrostatic approximation in rotating stratified flow 

Achim Wirth


Hydrostatic models were and still are the workhorses for realistic simulations of the ocean dynamics, especially for climate applications. The hydrostatic approximation is formally first order in $\gamma=H/L$, where $H$ is the vertical and $L$ the horizontal scale of the phenomenon considered. For stratified rotating flow the dynamics can be separated in balanced flow and wave motion. It is shown that for the linear balanced motion the hydrostatic approximation is exact and for wave motion it is second order, obtaining the leading prefactors. The validity of the hydrostatic approximation therefore also relies on the ratio of the amplitude of wave motion to balanced motion. This ratio adds considerably to the quality of the hydrostatic approximation for larger scale flows in the atmosphere and the ocean.
  
Imposing the divergenceless condition is a linear projection of the dynamical variables in the subspace of divergenceless vector fields, for both the Navier-Stokes and the hydrostatic formalism. Both projections are local in Fourier space.
  Calculating the difference of the two projections, the expression of the error, scaling and prefactors, done by the hydrostatic approximation is obtained. Analyzing the eigen-space of the projector, it is shown that for rotating-buoyant vortical-flow the hydrostatic-approximation is of third order for buoyant forcing, second order for horizontal and first order for vertical dynamical forcing.
  
Using the Heisenberg-Gabor limit it is shown that for large scale ocean dynamics, the difference of the dynamics of the projection-evolution operator between the two formalisms is insignificant. It is shown that the hydrostatic approximation is appropriate for realistic ocean simulations with vertical viscosities larger than $\approx10^{-2}$m$^2$s$^{-1}$. A special emphasis is on unveiling the physical interpretation of the calculations.
  

How to cite: Wirth, A.: On the hydrostatic approximation in rotating stratified flow, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-1469, https://doi.org/10.5194/egusphere-egu24-1469, 2024.

EGU24-2404 | Posters on site | NP1.1

Data Assimilation with Biases & Random Errors 

Juan Restrepo, Jorge Ramirez, Peter vanLeeuwen, and Caio Alves

Assimilating dynamic models and observations, along with their errors using Bayesian estimation method are challenged when the model has both aleatoric and epistemic errors. We devised a diffusion map technique that can filter an observational data stream, stripping it of components that are near statistically stationary, leaving behind what we denote the tendency of the time series. The tendency of the time series can be thought of as an executive summary of the time series. A model constructed on known physical principles may not be able to capture the tendency with fidelity and thus one can identify, from an estimation process on aleatoric fields, the epistemic error. Using machine learning strategies a surrogate model for the epistemic error can be inferred from a comparison of the physics model and the tendency. The surrogate model is thus incorporated into the model dynamics to enhance the fidelity in predictions. The assimilation of the enhanced model and the observations can now be carried out over the aleatoric Bayesian framework. To meet the challenge of the resulting highly nonlinear and non-Gaussian data assimilation we employ a newly developed Stein sampler we call the particle flow filter. In this talk we will describe and demonstrate  this assimilation strategy. 

How to cite: Restrepo, J., Ramirez, J., vanLeeuwen, P., and Alves, C.: Data Assimilation with Biases & Random Errors, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-2404, https://doi.org/10.5194/egusphere-egu24-2404, 2024.

EGU24-2554 | ECS | Orals | NP1.1

Koopman operator theory for enhanced Pacific SST forecasting 

Paula Lorenzo Sánchez, Matt Newman, Antonio Navarra, John Albers, and Aneesh Subramanian

El Niño-Southern Oscillation (ENSO) is a complex climatic phenomenon with significant impacts on global weather patterns and ecosystems. Improving ENSO predictability is therefore an issue of high societal value. However, Global Circulation Models present severe biases when predicting ENSO, and their skill remains comparable to that of vastly simpler empirical models such as Linear Inverse Models (LIMs). LIMs, however, rely on linear dynamics, and they have inherent limitations in capturing the behavior of non-linear phenomena. In this context, Koopman operator theory has emerged as a powerful mathematical framework, offering a novel perspective for analyzing complex non-linear systems, such as ENSO. In this study, we investigate the potential of Koopman operator theory to enhance ENSO forecasting accuracy. Leveraging 2000 years of tropical SST pre-industrial CESM data, we have assessed the skill of the Niño 3.4 index forecasts using the Koopman framework, and compared it to the benchmark set by LIMs. Our analysis includes sensitivity testing of both methods across various parameters, such as retained variability and data length used for operator computations. Our findings reveal nuances in the robustness of Koopman Operator estimates, particularly evident when using shorter training periods, contrasting with more stable LIM counterparts. However, a notable breakthrough emerges as we demonstrate the higher skill of Koopman multimodel ensembles, showcasing consistent improvements over linear models. The comparative analysis highlights the potential of Koopman operator theory in advancing ENSO forecasting beyond linear models. The utilization of Koopman multimodel ensembles emerges as a promising strategy, demonstrating enhanced forecasting capabilities. Yet, challenges in robustness persist, particularly in shorter data spans, signaling avenues for further refinement. Overall, these findings underscore the significance of the Koopman framework and lay the groundwork for future research aimed at refining methodologies for more accurate predictions in complex climatic systems.

How to cite: Lorenzo Sánchez, P., Newman, M., Navarra, A., Albers, J., and Subramanian, A.: Koopman operator theory for enhanced Pacific SST forecasting, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-2554, https://doi.org/10.5194/egusphere-egu24-2554, 2024.

A refined 4D-Var assimilation system within DestinE allows us to assimilate the Meteosat-10/SEVIRI clear-sky radiances over Europe, as well as GOES-16/ABI and GOES-18/ABI, or HIMAWARI/AHI globally at a spatial scale of 75 km instead of the previous 125 km in the ECMWF Integrated Forecasting System (IFS). Higher resolution observations can potentially improve the analysis and therefore the prediction of extreme weather events over Europe, as well as globally. The effects of using higher resolution observations have been investigated with a detailed set of experiments and the impact on wind, temperature, and humidity has been evaluated. A broad range of experiments indicate that exploiting the higher spatial density clear-sky radiances leads to an improvement of humidity sensitive fields in short-range forecasts with the IFS as independently measured for example by instruments on low-earth-orbiting satellites (IASI, CrIS, SSMIS, or ATMS). Due to a reduced displacement and representativeness error, these changes could further lead to improvements in longer range forecasts as these errors propagate upscale nonlinearly. However, so far the impact on the medium range has been mostly neutral.

In addition, pre-processed GOES-16/ABI and GOES-18/ABI observations by NOAA have been assimilated with 10 min sampling rates at 75 km spatial density. Exploring how to best assimilate relatively small spatial and temporal scales for one geostationary satellite, will allow us to approach these smaller scales with other satellites such as HIMAWARI/AHI above the Pacific or MTG-I/FCI above Europe. Data from both satellites will be available for us early in 2024. Preliminary experiments demonstrate the ability of IFS to assimilate observations at the highest available temporal resolution for the GOES-16 and GOES-18 satellites. Higher resolution radiances observed at these shorter time intervals naturally capture smaller scale atmospheric features such as mesoscale convective systems. In our experiments, simultaneously assimilating observations at a higher spatial and temporal resolution leads to an impact that is only marginally better than assimilating higher density observations alone, suggesting a combined investigation of optimal time-assignment, as well as assessment of the observation error are needed to optimise the integration of rapid update measurements in 4D-Var. 

How to cite: Schröttle, J., Lupu, C., and Burrows, C.: Approaching the sub-mesoscale globally at 10 min temporal resolution through assimilating radiances measured by geostationary satellites, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-3222, https://doi.org/10.5194/egusphere-egu24-3222, 2024.

EGU24-4179 | Orals | NP1.1

Initial-Value vs. Model-Induced Forecast Errors: A New Perspective 

Zoltan Toth and Isidora Jankov
Numerical models of the atmosphere are based on the best theory available. Understandably, the theoretical assessment of errors produced by such models is confounding. Without clear theoretical guidance, the experimental separation of the model-induced part of the total forecast error variance is also challenging. In this study, forecast error and ensemble perturbation variances are decomposed. Independent smaller- and larger-scale components separated as a function of lead time are found to be associated with features that completely or only partially lost skill, respectively. For their phenomenological description, the larger-scale variance is further decomposed orthogonally into positional and structural components. An analysis of the various components reveals that chaotically amplifying initial perturbation and error variance predominantly leads to positional differences in forecasts, while structural differences are interpreted as an indicator of model-induced error. The relatively small amplitude of model-induced errors confirms earlier assumptions and limited empirical evidence that numerical models of the atmosphere may be near perfect on the scales they well resolve.

 

How to cite: Toth, Z. and Jankov, I.: Initial-Value vs. Model-Induced Forecast Errors: A New Perspective, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-4179, https://doi.org/10.5194/egusphere-egu24-4179, 2024.

EGU24-5287 | ECS | Posters on site | NP1.1

Exploring the Potential of History Matching for Land Surface Model Calibration 

Nina Raoult, Simon Beylat, James Salter, Frédéric Hourdin, Vladislav Bastrikov, Catherine Ottlé, and Philippe Peylin

With the growing complexity of land surface models used to represent the terrestrial part of wider Earth system models, the need for sophisticated and robust parameter optimisation techniques is paramount. Quantifying parameter uncertainty is essential for both model development and more accurate projections. History matching is an emerging technique in climate science for uncertainty quantification. Using Gaussian process emulators, history matching allows us to rule out parts of parameter space that lead to model outputs being inconsistent with observations. In this presentation, we assess the power of history matching by comparing results to variational data assimilation, commonly used in land surface models for parameter estimation. Although both approaches have different setups and goals, we can extract posterior parameter distributions from both methods and test the model-data fit of ensembles sampled from these distributions. Using a twin experiment, we test whether we can recover known parameter values. Through variational data assimilation, we closely match the observations. However, the known parameter values are not always contained in the posterior parameter distribution, highlighting the equifinality of the parameter space. In contrast, while more conservative, history matching still gives a reasonably good fit and provides more information about the model structure by allowing for non-Gaussian parameter distributions. Furthermore, the true parameters are contained in the posterior distributions. We then consider history matching's ability to ingest different metrics targeting different physical parts of the model, helping to reduce parameter space further and improve model-data fit. We find the best results when history matching is used with multiple metrics; not only is the model-data fit improved, but we also gain a deeper understanding of the model and how the different parameters constrain different parts of the seasonal cycle. We conclude by discussing the potential of history matching in future studies.

How to cite: Raoult, N., Beylat, S., Salter, J., Hourdin, F., Bastrikov, V., Ottlé, C., and Peylin, P.: Exploring the Potential of History Matching for Land Surface Model Calibration, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-5287, https://doi.org/10.5194/egusphere-egu24-5287, 2024.

EGU24-5561 | ECS | Posters on site | NP1.1

Understanding near-surface hydrogeological processes around Lake Velence (Hungary) – using mesh graph neural networks on multidimensional remote sensing data 

Tibor Rapai, Petra Baják, András Lukács, Balázs Székely, and Anita Erőss

Lake Velence is a shallow soda lake in Hungary whose water budget is mainly driven by precipitation and evaporation. The lake has shown a deteriorating tendency recently, including extremely low lake levels and poor water quality, which indicates its vulnerability against changing climatic conditions. At the same time several water usage conflicts appeared in the catchment area. Until recently, the groundwater component in the lake's water budget and the hydrogeological processes in the catchment area have not been taken into consideration. Recent hydrogeological studies, however, show groundwater discharge into the lake.  Thus, further investigating this question is of high importance, hence groundwater could reduce climatic vulnerability.

Our ongoing work aims at developing a model-based evaluation technique, utilizing all map-based geophysical information and time series of different satellite data products, having sufficient spatial resolution and providing information about parameters strongly connected to subsurface processes, showing up on the surface. The basic DEM raster layer is imported from Copernicus GLO-30 dataset, having vertical precision <4 m. The Region Of Interest is a rectangular part of the catchment area: 47.1–47.4N, 18.4–18.8E. The first segmentation of the ROI is done using elevation data combined with lithographic and soil type information, resulting in almost uniform Voronoi-like polygon tessellation, with cells classified by geostructure. Further refinement by land cover type is done using Sentinel-1 SAR data. Other fixed data of point and polygon layers are important terrain features, points of surface inflows, (known) water takeouts and monitoring wells.

The machine learning regression model has time series of measured data at all its layers, daily input from Agárd meteorological station, like precipitation, average temperature, wind speed and relative humidity. Another important input data comes from Sentinel-2 (GREEN-NIR)/(GREEN+NIR)=NDWI spectral index, available in about weekly time steps, varying between 2 days-2 weeks. A crucial feature of all remote sensing data used here is the spatial resolution being better (10 m) or similar to the resolution of the basic DEM model. During training a graph neural network is generated dynamically from the Voronoi tessellation, where cells are nodes and physical processes between neighbouring cells give edge attributes for the graph. We use rectilinear approximations for water runoff/subsurface water exchange between cells, vertical infiltration/discharge under cells and estimated evapotranspiration from them. Learnable parameters governing the intensity of these flows are connected to geostructure and land cover classes. Parameters are optimized with time interval cross validation, with one part of the time series data being left out from optimization in each epoch and used for evaluation against target water level data.

Automatic detection of spatio-temporal patterns, connected to near-surface hydrogeological processes helped visualizing and quantifying estimated physical flows. Comparison with field measurements confirmed theoretic results from MODFLOW basin modelling, proving topography as a driving factor for subsurface flows. Our model is also suitable to handle isotope tracers, and extension to deep learning model promises predictive functionality for water table level.

The research is part of a project which was funded by the National Multidisciplinary Laboratory for Climate Change, (Hungary) RRF-2.3.1-21-2022-00014.

How to cite: Rapai, T., Baják, P., Lukács, A., Székely, B., and Erőss, A.: Understanding near-surface hydrogeological processes around Lake Velence (Hungary) – using mesh graph neural networks on multidimensional remote sensing data, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-5561, https://doi.org/10.5194/egusphere-egu24-5561, 2024.

EGU24-6164 | ECS | Orals | NP1.1 | Highlight

An ensemble based approach for the effect of climate change on the dynamics of extremes 

Dániel Jánosi, Mátyás Herein, and Tamás Tél

In view of the growing importance of climate ensemble simulations, we propose an ensemble approach for following the dynamics of extremes in the presence of climate change. A strict analog of extreme events, a concept based on single time series and local observations, cannot be found. To study nevertheless typical properties over an ensemble, in particular if global variables are of interest, a novel, statistical approach is used, based on a ”zooming in” into the ensemble. To this end, additional sub-ensembles with initially very close members are generated around trajectories of the original ensemble. Plume diagrams initiated on the same day of a year are generated from these sub-ensembles. The trajectories within a plume diagram strongly deviate on the time scale of a few weeks. By defining the extreme deviation as the difference between the maximum and minimum values in a plume diagram, a growth rate for the extreme deviation can be extracted. An average of these taken over the original ensemble (i.e. over all sub-ensembles) characterizes the typical, exponential growth rate of extremes, and the reciprocal of this can be considered the characteristic time of the emergence of extremes. Using a climate model of intermediate complexity, these are found to be on the order of a few days, with some difference between the global mean surface temperature and pressure. Measuring the reciprocal of the growth rate in several years along the last century, results for the temperature turn out to be roughly constant, while a pronounced decaying trend is found in the last decades for the pressure.

How to cite: Jánosi, D., Herein, M., and Tél, T.: An ensemble based approach for the effect of climate change on the dynamics of extremes, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-6164, https://doi.org/10.5194/egusphere-egu24-6164, 2024.

EGU24-6791 | Orals | NP1.1 | Highlight

A Foray of Dynamics into the Realm of Statistics: A Review of Ensemble Forecasting 

Zoltan Toth, Jie Feng, Jing Zhang, and Malaquias Pena

Uncertain quantities are often described through statistical samples. Can samples for numerical weather forecasts be generated dynamically? At a great expense, they can. With statistically constrained perturbations, a cloud of initial states are created and then integrated forward in time. By now, this technique has become ubiquitous in weather and climate research and operations. Ensembles are widely used, with demonstrated value.

 

The atmosphere evolves in a multidimensional phase space. Does a cloud of ensemble solutions encompass the evolution of the real atmosphere? Theoretically, random perturbations in high dimensional spaces have negligible projection on any direction, including the error in the best estimate, therefore consistently degrading it. As the bulk of the perturbation variance lies in the null-space of error, samples in multidimensional space do not contain reality.

 

An evaluation suggests that initial and short-range forecast error and ensemble perturbations are random draws from a high dimensional domain we call the subspace of possible error. Error in any initial condition is a result of stochastic observational and assimilation noise, while perturbations explore other, mostly independent directions from the subspace of possible error that may have resulted from other configurations of stochastic noise. What benefits may arise from the deterministic projection of such noise?

 

How to cite: Toth, Z., Feng, J., Zhang, J., and Pena, M.: A Foray of Dynamics into the Realm of Statistics: A Review of Ensemble Forecasting, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-6791, https://doi.org/10.5194/egusphere-egu24-6791, 2024.

EGU24-7142 | Posters on site | NP1.1 | Highlight

Well-Posedness of the Dynamic Framework in Earth-System Model  

Ruxu Lian, Jieqiong Ma, and Qingcun Zeng

The well-posedness of the dynamic framework in earth-system model (ESM for short) is a common issue in earth sciences and mathematics. In this presentation, the authors will introduce the research history and fundamental roles of the well-posedness of the dynamic framework in the ESM,  emphasizing the three core components of ESM, i.e., the atmospheric general circulation model (AGCM for short), land-surface model (LSM for short) and oceanic general circulation model (OGCM for short) and their couplings. In fact, this system strictly obeys the conservation of energy and is used to make better climate predictions. Then, some research advances made by their own research group are outlined. Finally, future research prospects are discussed.

How to cite: Lian, R., Ma, J., and Zeng, Q.: Well-Posedness of the Dynamic Framework in Earth-System Model , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-7142, https://doi.org/10.5194/egusphere-egu24-7142, 2024.

EGU24-8590 | Orals | NP1.1

A stochastic model for the turbulent ocean heat flux under Arctic sea ice  

Srikanth Toppaladoddi and Andrew Wells

The physics of planetary climate features a variety of complex systems that are challenging to model as they feature turbulent flows. A key example is the heat flux from the upper ocean to the underside of sea ice which provides a key contribution to the evolution of the Arctic sea ice cover. Here, we develop a model of the turbulent ice-ocean heat flux using coupled ordinary stochastic differential equations to model fluctuations in the vertical velocity and temperature in the Arctic mixed layer. All the parameters in the model are determined from observational data. A detailed comparison between the model results and measurements made during the Surface Heat Budget of the Arctic Ocean (SHEBA) project reveals that the model is able to capture the probability density functions (PDFs) of velocity, temperature and heat flux fluctuations. Furthermore, we show that the temperature in the upper layer of the Arctic ocean can be treated as a passive scalar during the whole year of SHEBA measurements. The stochastic model developed here provides a computationally inexpensive way to compute an observationally consistent PDF of this heat flux, and has implications for its parameterisation in regional and global climate models.

How to cite: Toppaladoddi, S. and Wells, A.: A stochastic model for the turbulent ocean heat flux under Arctic sea ice , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-8590, https://doi.org/10.5194/egusphere-egu24-8590, 2024.

EGU24-8958 | ECS | Orals | NP1.1

Capturing the Variability of the Nocturnal Boundary Layer through Localized Perturbation Modeling 

Amandine Kaiser, Nikki Vercauteren, and Sebastian Krumscheid

Numerical weather prediction and climate models encounter challenges in accurately representing flow regimes in the stably stratified atmospheric boundary layer and the transitions between them. This leads to an inadequate depiction of regime occupation statistics and, therefore, to biases in forecasts of near-surface temperature. To explore inherent uncertainties in modeling regime transitions, the response of the near-surface temperature inversion to transient small-scale phenomena is analyzed based on a stochastic modeling approach. A sensitivity analysis is conducted by augmenting a single-column model for the atmospheric boundary layer with deterministic perturbations accounting for small-scale fluctuations in the wind and temperature dynamics and with a stochastic stability function to account for turbulent bursts. The model is a tool to systematically investigate what types of unsteady flow features may trigger abrupt transitions in the mean boundary layer state. Previous research showed that incorporating enhanced mixing, a common practice in numerical weather prediction models, blurs the two regime characteristics of the stably stratified atmospheric boundary layer. Simulating intermittent turbulence through a stochastic stability function is shown to provide a potential workaround for this issue. Including key uncertainty in models could lead to a better statistical representation of the regimes in long-term climate simulation. 

How to cite: Kaiser, A., Vercauteren, N., and Krumscheid, S.: Capturing the Variability of the Nocturnal Boundary Layer through Localized Perturbation Modeling, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-8958, https://doi.org/10.5194/egusphere-egu24-8958, 2024.

EGU24-9569 | ECS | Orals | NP1.1

Impact Predictability: Exploring Extremes in Biosphere Dynamics with Recurrent Neural Networks 

Francesco Martinuzzi, Miguel D. Mahecha, Gustau Camps-Valls, David Montero, Tristan Williams, and Karin Mora

Understanding Earth's terrestrial biosphere dynamics is vital for comprehending our planet's environmental health and sustainability. Recently, the frequency and intensity of extreme climate events have risen, significantly impacting the biosphere. Given the advancements of recurrent neural networks in modeling complex, nonlinear dynamics, we explore the capability of recurrent neural network models to model and predict the impacts of extreme events on biosphere dynamics. In this work, we compare four different recurrent network architectures, each with distinct features: 1) Recurrent Neural Networks (RNNs); 2) Long Short-Term Memory-based networks (LSTMs), known for their efficacy in handling long-term dependencies; 3) Gated Recurrent Unit-based networks (GRUs), which offer a simplified yet effective alternative to LSTMs; and 4) Echo State Networks (ESNs), which are distinguished by fixed internal weights and training based on simple linear regression. Our study found that while recurrent network architectures show similar performance under standard conditions, Echo State Networks (ESNs) show slightly superior performance, particularly under extreme events. However, we also identify limitations in current models under extreme conditions, underscoring the need for specialized approaches to enhance predictive accuracy in these circumstances.

How to cite: Martinuzzi, F., Mahecha, M. D., Camps-Valls, G., Montero, D., Williams, T., and Mora, K.: Impact Predictability: Exploring Extremes in Biosphere Dynamics with Recurrent Neural Networks, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-9569, https://doi.org/10.5194/egusphere-egu24-9569, 2024.

This study aims at proposing a new framework to perform ensemble-based estimations of dynamical trajectories of a geophysical fluid flow system. To perform efficient estimations, the ensemble members are embedded in a set of evolving reproducing kernel Hilbert spaces (RKHS) defining a family of spaces.

The method proposed here is designed to deal with very large scale systems such as oceanic or meteorological flows, where it is out of the question to explore the whole attractor, neither to run very long time simulations. Instead, we propose to learn the system locally, in phase space, from an ensemble of trajectories.

The novelty of the present work relies on the fact that the feature maps between the native space and the RKHS manifold are transported by the dynamical system. This creates, at any time, an isometry between the tangent RKHS at time t and the initial conditions. This has several important consequences. First, the kernel evaluations are constant along trajectories, instead to be attached to a system state. By doing so, a new ensemble member embedded in the RKHS manifold at the initial time can be very simply estimated at a further time. This framework displays striking properties. The Koopman and Perron-Frobenius operators on such RKHS manifold are unitary, uniformly continuous (with bounded generators) and diagonalizable. As such they can be rigorously expended in exponential forms.

This set of analytical properties enables us to provide a practical estimation of the Koopman eigenfunctions. In the proposed strategy, evaluations of these Koopman eigenfunctions at the ensemble members are exact. To perform robust estimations, the finite-time Lyapunov exponents associated with each Koopman eigenfunction (which are easily accessible on the RKHS manifold as well) are determined. On this basis, we are able to filter the kernel by removing contributions of the Koopman modes that exceed the predictability time. We show that it leads to robust estimations of new unknown trajectories. This framework allows us to write an ensemble-based data assimilation problem, where constant-in-time linear combinations coefficients between ensemble members are sought in order to estimate the QG flow based on noisy swath observations.

The methodology is demonstrated on a multilayer quasi-geostrophic model representative of the Gulf Stream area in the North Atlantic at a 10 km resolution and considering 100 training ensemble members. We show the ability of the method to estimate trajectories knowing the initial condition of a new ensemble member. Moreover, ensemble-based data assimilation is performed based on realistic swaths of altimetry observations.

How to cite: Tissot, G., Jaouen, M., and Mémin, E.: Ensemble forecasts in reproducing kernel Hilbert space family: Application on a multilayer quasi-geostrophic numerical simulation, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-10075, https://doi.org/10.5194/egusphere-egu24-10075, 2024.

EGU24-10174 | Posters on site | NP1.1

Alternative approaches to the velocity of climate change: assessment against observations 

Jérôme Kasparian, Laure Moinat, Iaroslav Gaopnenko, and Stéphane Goyette

Climate change causes shifts in distribution ranges of species. The velocity of these shifts is often related to that of climate change, such as the poleward shift of the isotherms. However, species range shifts are not solely determined by climate parameters. Short-term meteorological values, topography, and other barriers also play a role. Moreover, the magnitude and direction of the climate velocity are not defined univocally. They depend on implicit assumptions that underlie the calculations, particularly in determining the direction of the velocity vector. The classical gradient-based definition of climate change [1] displays limitations, in particular local divergence [2], which led us to recently introduce an alternative method maximising the regularity of the velocity field, named Monte-cArlo iTerative Convergence metHod (MATCH) [3].

Since the latter stems from mathematical arguments, its relevance to ecology requires careful assess- ment. We consider North-American birds based on the Audubon Christmas Bird Count as well as marine species recorded in the North-East Atlantic region of the NOAA fisheries survey. For each species, the centroid of the distribution area is determined at two time ranges, and its shifting velocity, in magnitude and direction, is deduced. We also calculate the shift of the isotherms for ground and sea-surface tem- peratures, respectively, at each observation spot, and deduce an average shifting velocity for both the Gradient-based and the MATCH methods.

When comparing the respective shifts of the ranges of species and of climate, we only found a significant positive correlation between latitudinal shift of marine species and their climate counterpart, as calculated with the MATCH approach. Neither the classical gradient method, nor longitudinal shifts, nor bird range shifts displayed significant correlations. Our results therefore suggest that the MATCH approach may provide more ecologically relevant velocity fields. We also confirmed previous observations that marine species better track temperature evolutions than terrestrial ones. Such assessment may help anticipating species range shift and designing conservation strategies.

References

1. S. R. Loarie et al. Nature 462, 1052 (2009)

2. J. Rey, G. Rohat, M. Perroud, S. Goyette, J. Kasparian, Env. Res. Lett. 15, 034027 (2020)

3. I. Gaponenko, G. Rohat, S. Goyette, P. Paruch, J. Kasparian, Sci. Rep., 12, 2997, (2022)

How to cite: Kasparian, J., Moinat, L., Gaopnenko, I., and Goyette, S.: Alternative approaches to the velocity of climate change: assessment against observations, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-10174, https://doi.org/10.5194/egusphere-egu24-10174, 2024.

EGU24-11543 | ECS | Orals | NP1.1

Ensemble Data Assimilation with Model Error Inclusion for Estimating Fault-Slip Occurrence in Large-Scale Laboratory Experiments 

Hamed Ali Diab-Montero, Meng Li, Ylona van Dinther, and Femke C Vossepoel

The forecasting of earthquake occurrence remains a significant challenge in seismology, primarily due to uncertainties in understanding the current state of stress, strength, and the parameters controlling the slip behaviour of faults. Among these parameters, friction parameters are crucial as they control the earthquake recurrence interval and their nucleation size. There is a critical gap in effectively integrating observational data with physics-based models, particularly in the face of parameter bias. This study addresses how ensemble data assimilation methods can be optimized to address these challenges and reduce uncertainties in fault-slip estimates.

Our objective is to enhance the accuracy of earthquake forecasting by incorporating model error into ensemble data assimilation methods, thus improving the estimation of critical state variables such as shear stress and slip velocity. We employed the Ensemble Kalman (EnKF), Adaptive Gaussian Mixture (AGMF), and Particle Flow (PFF) filters, which are integrated with earthquake sequence models. These methods were applied in two stages: 1) Perfect model experiments using 1-D Burridge-Knopoff models to assess the benefits of including model error in estimating non-periodic and chaotic behaviours in low-dimensional systems. 2) Application in a meter-scale direct-shear laboratory setup, assimilating measurements from shear-strain gauges near the fault, to examine the effects of varying normal stress profiles on the fault on the estimates and the impact of including model error.

The perfect model experiments demonstrated improved estimation of shear stress, slip velocity, and the state variable (θ), particularly in estimating non-periodic sequences and chaotic behaviour using stable periodic solutions when confronted with small parameter biases. In the laboratory setup, variations in normal stress profiles significantly influenced the information content. Sensor placement relative to the fault and seismic phase was found to critically impact the observations' informational value, with sensor proximity to the fault being a critical aspect and the information content being higher around the coseismic phase.

This research provides valuable insights into the intricate process of earthquake forecasting, underscoring the role of data-assimilation techniques in enhancing our understanding and forecasting abilities in this field.

How to cite: Diab-Montero, H. A., Li, M., van Dinther, Y., and Vossepoel, F. C.: Ensemble Data Assimilation with Model Error Inclusion for Estimating Fault-Slip Occurrence in Large-Scale Laboratory Experiments, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-11543, https://doi.org/10.5194/egusphere-egu24-11543, 2024.

EGU24-11724 | Orals | NP1.1

Towards a machine learning model for data assimilation and forecasting directly trained from observations 

Christian Lessig, Peter Lean, Tony McNally, Mihai Alexe, Simon Lang, Matthew Chantry, and Peter Dueben

State-of-the-art data assimilation systems, such as the 4DVar system of the European Centre for Medium-Range Weather Forecasts (ECMWF), are highly successful in producing state estimates of the atmosphere constrained by millions of observations. However, existing systems require substantial approximations, e.g. in forward operators, and employ conventional models in their optimization loop. This limits the amount of information that can be extracted from observations, e.g. the assimilation of visible satellite channels is still challenging. The current approach also separates the use of observations into a data assimilation step and a forecasting one, with observations only being used indirectly for forecasting, for example for tuning of parametrizations and for evaluation.

Here, we explore the possibility to train large machine learning models for data assimilation and forecasting directly from observations. In particular, we build a generative transformer neural network that models the joint the probability distribution p(y,x) over output states y for an input x. The input are observations from a temporal window, e.g. 6h or 12h, and the output y can either be an estimate of the state within the window or a short-term forecast, e.g. for another 12h. Different observations are processed by different embedding networks but then fused in the backbone transformer network. To obtain an integrated and consistent representation of the atmospheric state that corresponds to the different input data streams, we train with a variation of the masked token model training objective from natural language processing that impels the network to learn the correlation between the different input streams and channels. To properly represent the statistical nature of the estimation of y given x, our network provides an ensemble prediction as a nonparametric model for the probability distribution over y.

We present results for a network trained with a substantial amount of data, including different satellite observations (such as AMSU-A microwave sounders from NOAA 15-19 and the METOP satellites as well as IASI), radiosondes, and ground station-based measurements. The skill for both data assimilation and forecasting is analyzed and compared to ECMWF’s operational 4DVar system. We also ablate the effect different observations have on the skill of the network output.

How to cite: Lessig, C., Lean, P., McNally, T., Alexe, M., Lang, S., Chantry, M., and Dueben, P.: Towards a machine learning model for data assimilation and forecasting directly trained from observations, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-11724, https://doi.org/10.5194/egusphere-egu24-11724, 2024.

EGU24-12751 | Orals | NP1.1

Data assimilation approaches with iterative ensemble smoothers in coupled nonlinear multiscale models 

Femke Vossepoel, Geir Evensen, and Peter Jan van Leeuwen

Iterative ensemble smoothers, originally developed for parameter estimation in petroleum applications, are effective data assimilation methods in coupled, unstable dynamical systems. In this study, we demonstrate this using a coupled multiscale model based on two Kuramoto-Sivashinsky equations with different spatial and temporal scales, representing two subsystems of an earth-system model. The cross-covariance between the variables of the two subsystems reflects how each subsystem influences the other, leading to unexpected structures that reveal interesting physics of the coupled system. The setup of the data assimilation allows simultaneous updating of both systems, leading to consistent estimates.

A comparative study illustrates the properties of iterative ensemble smoothers and assimilation updates over finite-length assimilation windows. We demonstrate the increased accuracy of the smoothers’ solution compared to that of the standard ensemble Kalman filter and the fast convergence of the iterations related to the efficient handling of nonlinearities by the nonlinear space-time ensemble.

Localisation, whether distance-based or adapted to spatial correlations, can be used in iterative ensemble smoothers to effectively deal with limited ensemble sizes. We discuss the effects of the spatial distribution and temporal frequency of available observations and illustrate how data gaps in one of the two subsystems affect the coupled estimate. Looking forward, we present the possibilities and benefits of a potential implementation of this approach in coupled earth-system models.

How to cite: Vossepoel, F., Evensen, G., and van Leeuwen, P. J.: Data assimilation approaches with iterative ensemble smoothers in coupled nonlinear multiscale models, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-12751, https://doi.org/10.5194/egusphere-egu24-12751, 2024.

The route to chaos in a truncated rotating shallow-water model has been investigated by constructing an autonomous five-mode Galerkin truncated system with complex variables. Two distinct transitions to chaos were observed as the energy injected into the system increased. The first transition is characterized by forming a continuous sequence of bifurcations that follow the usual Feigenbaum path. The second transition, occurring for high values of injected energy, exhibits a sharp transition between quasi-periodic states and chaotic regimes. The first chaotic regime arises since nonlinear interactions are principally dominated by inertial terms, while the second one is related to the increasing importance of free surface elevation in the overall process. By rewriting the system in terms of phase and amplitude, for each variable truncated system, it has been found that phases are locked at the initial value for a certain period of time, followed by a sudden transition due to a simple rotation of $\pm \pi$, even when amplitudes show a chaotic dynamic. The time duration of phase locking decreases as the injected energy increases, and, for high values of injected energy, even phases reach a chaotic regime. This behaviour is observed since, in the nonlinear term of the equations, phases appear through linear combinations of triads of different modes. When the duration of locking periods is different for each mode, the superposition of multiple $\pi$ phases jumps, making the dynamics of the coupled phase triads stochastic, even for small values of the injected energy.

How to cite: Carbone, F. and Dutykh, D.: Route to chaos and resonant triads interaction in a truncated Rotating Nonlinear shallow-water model, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-13180, https://doi.org/10.5194/egusphere-egu24-13180, 2024.

EGU24-13849 | ECS | Posters on site | NP1.1

Spatiotemporally-separated framework for the source reconstruction of atmospheric radionuclide releases 

Yuhan Xu, Sheng Fang, Xinwen Dong, Hao Hu, and Shuhan Zhuang

Determining the source location and release rate is critical in assessing the environmental consequences of atmospheric radionuclide releases, yet this task remains challenging due to the vast multi-dimensional solution space. To address this, we propose a spatiotemporally-separated two-step framework that reduces the dimension of the solution space in each step and enhances the reconstruction accuracy, which is applicable to radionuclide releases. The separating process involves applying a temporal sliding-window average filter to the observations, thereby reducing the influence of temporal variations in the release rate and ensuring that the features of the filtered data are dominated by the source location. Initially, candidate source locations are pre-screened using a correlation-based method. To establish the relationship between the filtered data and candidate source locations, time- and frequency-domain features are extracted from the filtered data and an eXtreme Gradient Boosting algorithm is employed for fitting. The features are further screened out by the Recursive Feature Elimination with Cross-Validation. Utilizing the features of filtered observations, the source location can be determined without the knowledge of the release rate. Subsequently, the release rate is determined using projected alternating minimization with the L1-norm and total variation regularization algorithm.

The proposed method was rigorously tested on two field experiments: the SCK-CEN experiment, featuring local-scale 41Ar releases over two days, and the ETEX-I experiment, involving continental-scale PMCH releases. Validation on the SCK-CEN experiment showed that the lowest source location error fell below 1% and the mean source location error remained under 5%, with temporal variations and peak release rate being accurately reconstructed. Similar accuracy was also observed in the ETEX-I experiment. Compared to traditional correlation-based method and Bayesian method, our method exhibited superior accuracy and a reduced uncertainty range.

Furthermore, comprehensive sensitivity tests were conducted on the SCK-CEN experiment to evaluate the influence of pre-screening range, sliding-window length, feature types, and combinations of observation sites. The results indicated that our method achieved consistent performance across various parameters and conditions, maintaining low error levels even with only a single observation site.

How to cite: Xu, Y., Fang, S., Dong, X., Hu, H., and Zhuang, S.: Spatiotemporally-separated framework for the source reconstruction of atmospheric radionuclide releases, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-13849, https://doi.org/10.5194/egusphere-egu24-13849, 2024.

EGU24-14101 | Orals | NP1.1 | Highlight

Mathematics of Sea Ice and its Ecosystems 

Kenneth Golden

The Arctic and Antarctic sea ice covers form key components of Earth's climate system. Their precipitous declines are impacting the polar marine environment and its ecosystems, with ripple effects felt far beyond the polar regions. As a material sea ice exhibits composite structure on many length scales. A principal challenge is how to use information on small scale structure to find the effective or homogenized properties on larger scales relevant to climate and ecological models. From tiny brine inclusions to rich ice pack dynamics on oceanic scales, and from microbes to polar bears, we'll consider recent advances in modeling sea ice and the ecosystems it hosts. In the spirit of MPE 2013, we’ll focus on the broad range of mathematics and physics being used. Percolation theory and statistical physics, fractal geometry, spectral analysis and random matrix theory, advection diffusion processes, topological data analysis, and uncertainty quantification for dynamical systems will arise naturally in considering various sea ice structures and organisms. This work is helping to advance how sea ice is represented in climate models, and to improve projections of the fate of Earth’s sea ice packs and the ecosystems they support.

How to cite: Golden, K.: Mathematics of Sea Ice and its Ecosystems, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-14101, https://doi.org/10.5194/egusphere-egu24-14101, 2024.

EGU24-14208 | ECS | Orals | NP1.1

Unbiased fully nonlinear data assimilation: the Stochastic Particle Flow Filter 

Hao-Lun Yeh and Peter Jan van Leeuwen

Nonlinearities in numerical models for the geosciences and in observation operators that map model states to observation space have become so strong that they can no longer be ignored. The particle flow filter (PFF) is a fully nonlinear and efficient sequential Monte Carlo filter that removes the weight degeneracy problem in particle filters by iteratively transporting the equal-weighted particles from the prior to the posterior distribution. The deterministic version of the PPF has been successfully applied to high-dimensional systems and is unbiased in the limit of an infinite number of particles. However, with a small number of particles, the ensemble spread can be biased low, especially in the observed part of the state space. This can be partly alleviated by using a so-called matrix-valued kernel in the algorithm, but the fundamental issue remains. To address this challenge, we propose a novel approach, the Stochastic Particle Flow Filter (SPFF), which includes a Gaussian noise in the Stein Variational Gradient Descent dynamics, the amplitude and covariance of which follow directly from theory. With this additional repulsive force between particles, the SPFF guarantees an unbiased posterior pdf, even with a finite number of particles.

We demonstrate the performance of the SPFF using detailed experiments with the 1000-dimensional Loreanz-96 model. Our results demonstrate that SPFF successfully avoids particle collapse of the marginal distributions and accurately captures the evolutions of particles Additionally, and initially unexpectedly, the SPFF exhibits faster convergence than the deterministic PFF and thus improves analysis accuracy compared to the PFF with a matrix-valued kernel at the computational cost. We also show results of its performance on a high-dimensional ocean model demonstrating that we, as a community, are very close to solving the nonlinear data assimilation problem.

How to cite: Yeh, H.-L. and van Leeuwen, P. J.: Unbiased fully nonlinear data assimilation: the Stochastic Particle Flow Filter, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-14208, https://doi.org/10.5194/egusphere-egu24-14208, 2024.

EGU24-15676 | ECS | Posters on site | NP1.1

Observer-based data assimilation for shallow water equations 

Teresa Kunkel, Jan Giesselmann, and Martin Gugat

We consider state estimation for a system described by the one-dimensional shallow water equations. Since in general measurements of the complete state are not available, we assume that we have measurements of only one state variable, e.g. of the water height. In order to estimate the system state from these partial measurements, we construct an observer system that is based on the shallow water equations. Distributed measurements of the water height are inserted into the observer system through source terms of Luenberger type. Our main contribution is to show exponential convergence of the state of the observer system towards the original system state in the long time limit for a 2x2-system of nonlinear hyperbolic balance laws, i.e., we reconstruct the complete system state from measurements of one state variable. The proof is based on estimating the difference between the observer system and the original system via a suitable extension of the relative energy method.
Using energy-consistent coupling conditions and transforming the system to a Hamiltonian formulation, the synchronization result can be extended to star-shaped networks. This might have an application in flood modeling of river systems or control of irrigation systems.

How to cite: Kunkel, T., Giesselmann, J., and Gugat, M.: Observer-based data assimilation for shallow water equations, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-15676, https://doi.org/10.5194/egusphere-egu24-15676, 2024.

EGU24-15777 | Posters on site | NP1.1

Comparison of several approximation schemes on the Cubed Sphere 

Jean-Pierre Croisille, Jean-Baptiste Bellet, and Matthieu Brachet

Approximation, interpolation and quadrature are questions of fundamental 
importance for atmospheric and oceanic problems at planetary scale. 
Computation with spherical harmonics on the sphere is an old mathematical topic; it has a particular interest in geosciences, and is still an active field of research. In this poster, we will show numerical comparisons of several approximation schemes, with a special focus on the Cubed Sphere grid. We test hyperinterpolation, weighted least squares, and interpolation on a series of test functions with various smoothness properties. Our last results include the derivation of explicit formulas for optimal quadrature rules on low resolution Cubed Spheres.

[1] J.-B. Bellet, M. Brachet, and J.-P. Croisille, Interpolation on the Cubed Sphere with Spherical Harmonics, Numerische Mathematik, 153 (2023), pp. 249-278.
[2] J.-B. Bellet and J.-P. Croisille, Least Squares Spherical Harmonics Approximation on the Cubed Sphere, Journal of Computational and Applied Mathematics, 429 (2023),  115213.
[3] C. An and H.-N. Wu, Bypassing the quadrature exactness assumption of hyperinterpolation on the sphere, Journal of Complexity, 80 (2024), 101789.

How to cite: Croisille, J.-P., Bellet, J.-B., and Brachet, M.: Comparison of several approximation schemes on the Cubed Sphere, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-15777, https://doi.org/10.5194/egusphere-egu24-15777, 2024.

EGU24-16847 | ECS | Orals | NP1.1

Simulation-based inference as a paradigm for scientific machine learning in the cryosphere and beyond 

Brian Groenke, Kristoffer Aalstad, Norbert Pirk, Sebastian Westermann, Jakob Zscheischler, Guillermo Gallego, and Julia Boike

Mechanistic or dynamical models based on governing equations are ubiquitous throughout science and engineering. Such models, also referred to as “simulators”, are typically characterized by a forward mapping from some set of inputs or parameters to one or more output quantities of interest. In many cases, the inputs required for the forward model are either unknown or represent approximations of unresolved processes. Bayesian inference provides a natural framework for constraining this model uncertainty using observed data. Such a framework is especially valuable in cryospheric application domains such as permafrost research, where direct observations of many quantities of interest, e.g. subsurface temperature and soil moisture, are only sparsely available. Mechanistic models based on known physics therefore play an indispensable role in filling these gaps. However, virtually all methods for Bayesian inference require repeated evaluation of the forward model which is often computationally challenging, especially for dynamical systems. As a result, computational requirements of statistical inference very quickly become intractable even for systems of only moderate complexity. The burgeoning field of “simulation-based inference” (SBI) aims to leverage modern computational methods from machine learning (ML) and data assimilation (DA) to overcome these challenges and facilitate large scale uncertainty quantification in complex scientific models. In this work, we show how SBI can be seen as a unifying theoretical framework that bridges the gap between existing DA methods (e.g. variants of the ensemble Kalman filter) and full-fledged Bayesian inference with the goal of facilitating hybrid statistical-physical modeling of complex systems. We present a novel set of software tools for making SBI more accessible to researchers along with benchmarks of several methods drawn from both the ML and DA literature. Two of these benchmarks are based on use cases from Arctic land surface modeling: degree day approximation of snowmelt and geothermal inversion of historical climate change from boreholes in Arctic permafrost. We highlight the contributions that SBI can make to solving such inverse problems and discuss further potential applications in the cryosphere and beyond.

How to cite: Groenke, B., Aalstad, K., Pirk, N., Westermann, S., Zscheischler, J., Gallego, G., and Boike, J.: Simulation-based inference as a paradigm for scientific machine learning in the cryosphere and beyond, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-16847, https://doi.org/10.5194/egusphere-egu24-16847, 2024.

EGU24-17015 | Orals | NP1.1

Relative contribution of high-resolution Sentinel-1 data assimilation and modeling choices to improve regional water budget estimates 

Gabriëlle De Lannoy, Louise Busschaert, Sara Modanesi, Devon Dunmire, Isis Brangers, Hans Lievens, Zdenko Heyvaert, Christian Massari, Augusto Getirana, and Michel Bechtold

Land surface models provide self-consistent estimates of the water stored in various components of the land system, i.e. in the soil, vegetation and snow, and of water fluxes. Assimilation of satellite-based data helps to update these estimates to some extent, but it has some limitations when human processes, such as irrigation, are missing in the modeling system. Furthermore, the simulated water distribution heavily depends on the choice of meteorological input and other model choices. In this study, we aim to quantify the relative contribution of (i) Sentinel-1 data assimilation for soil moisture and snow updating, (ii) meteorological input, and (iii) modeling irrigation over the Po river basin in Italy.

The Po river network channels the discharge of snow melt water from the Alps and Apennines, combined with surface and deep subsurface runoff from the hillslopes and valley. During the summer, the river network supplies irrigation water to the large agricultural area in the Po river valley. The Po basin is thus a unique testbed to study various water budget components in an environment with pronounced seasonal water storage dynamics and human water management.

More specifically, we assimilate 1-km Sentinel-1 data into the Noah-MP land surface model coupled to an irrigation module and the hydrological modeling and analysis platform (HyMAP) as runoff routing module. The Noah-MP simulations are forced with meteorological data from either the fifth generation ECMWF atmospheric reanalysis (ERA5) or the Modern-Era Retrospective Analysis for Research and Applications, Version 2 (MERRA-2) for the years 2015-2023. Sentinel-1 snow depth retrievals are assimilated over the mountains in the winter, whereas Sentinel-1 backscatter are mainly assimilated in the valley during the spring, summer and fall. The state updates are applied to snow depth, snow water equivalent, and soil moisture. These updates subsequently trigger updates in estimates of irrigation, leaf area index, discharge and other variables, resulting in a self-consistent re-analysis of the entire water budget of the Po basin. The impact of the Sentinel-1 data assimilation relative to that of the activation of irrigation modeling is quantified using independent in situ and remotely sensed measurements of soil moisture, leaf area index, snow depth, evaporation, irrigation and discharge.

How to cite: De Lannoy, G., Busschaert, L., Modanesi, S., Dunmire, D., Brangers, I., Lievens, H., Heyvaert, Z., Massari, C., Getirana, A., and Bechtold, M.: Relative contribution of high-resolution Sentinel-1 data assimilation and modeling choices to improve regional water budget estimates, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-17015, https://doi.org/10.5194/egusphere-egu24-17015, 2024.

EGU24-17713 | ECS | Orals | NP1.1

The selective boundary crossing in stochastic processes driven by white shot noise 

Giulio Calvani and Paolo Perona

Many physical, chemical, financial, and ecological processes show the presence of a threshold, which may affect their dynamics. For instance, chemical reactions occur when the energy reaches the activation value; for insurance companies and banks, ruin may happen when the balance drops down a minimum amount; in fluvial hydraulics, sediment transport, and morphodynamic processes start when bed shear stresses (i.e., flow discharge) overcome a critical threshold. Other processes may show the presence of additional boundaries. For stochastic processes, whose dynamics depends on the frequency and magnitude of random fluctuations, it is interesting to know the average time the process takes to reach one of the critical boundaries starting from a known value. This quantity is known in the literature as Mean First Passage Times (MFPTs). The quantification of the MFPTs is usually performed by considering one threshold, only. When two or multiple thresholds are present, one may consider the MFPTs of reaching either one of the thresholds, without having passed the other ones. Such a selective condition is referred to in the literature as splitting probability. In this work, we consider stochastic processes governed by a typical Langevin equation with deterministic drift and random instantaneous jumps (white shot noise). We perform a statistical-trajectory-analysis starting from a point between two thresholds and derive exact relationships of the splitting probabilities and the MFPTs of crossing one threshold, only, based on process-dependent dimensionless parameters. Such formulations are then explicitly given for the cases of constant and linear drift functions and both positive and negative jumps. We test the derived formulations against data from MonteCarlo simulations, by varying the process parameters, the starting point, and the values of the thresholds. The comparison shows very good agreement and confirms the correctness of the derived relationships. Additionally, the analysis highlights the role played by the dimensionless parameters. Then, data from flow measurements in a river are considered and we successfully test the formulations against the duration of the raising limb of high-stage events. The derived formulations can be readily applied to calculate both the duration of the raising and the falling limbs of flow events, which are important quantities for engineering applications, as well as for modeling purposes in river eco-morphodynamics.

How to cite: Calvani, G. and Perona, P.: The selective boundary crossing in stochastic processes driven by white shot noise, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-17713, https://doi.org/10.5194/egusphere-egu24-17713, 2024.

EGU24-17846 | ECS | Orals | NP1.1

Temperature at 500hPa, energy budget and the 2003 summer heatwave over Western Europe: a triple-decomposition approach 

Manuel Fossa, Luminita Danaila, and Michael Ghil

Heatwaves represent a major health hazard, as was the case for the 2003 summer heat wave, responsible for more than 70 000 deaths in western Europe. The interplay of mean flow, quasi-periodic and random fluctuations — associated with the westerly jet, Rossby and gravity waves, and eddies — of the large-scale temperature field results in complex heat wave–related amospheric conditions. Understanding how the different physical processes interact is thus crucial for prediction of heat wave events. 
In this study, we use the triple decomposition of turbulent flow (Hussain and Reynolds, JFM, 1972) to compute mean, quasi-periodic and random energy fluctuations of the temperature field, that is the 1-point energy budget of temperatures. This decomposition takes into account all interactions between the zonal jet, Rossby waves, gravity waves, and eddies. Both spectral and dynamical systems analyses are applied to the computed terms. More specifically, the concept of extremal length (Ahlfors, Vol. 371, AMS, 2010) is integrated into the equations to quantify how each term of the energy budget equations contributes to the "trapping" of temperature anomalies over Europe.      
Results show that, amid positive sea surface temperatures and negative soil moisture anomalies, during the first half of August, i.e., the hottest days of the heat wave, quasi-periodic oscillations of polar air increased, resulting in meridional migration of cold air over Canada, and subsequent mixing with warmer air coming from North America. This mixing triggered baroclinic instabilities that led to production of turbulent eddies, which by August 5th suddenly stopped their eastward progression, creating a cyclonically stalled regime over the Mid-Atlantic; this stationary cyclone interacted positively with North African warm air propagating northward over Europe, thus sustaining the dry conditions over France and much of Western Europe. The cyclonic block finally disappeared, stopping the warm air advection from North Africa, with temperatures falling just after that. 
The study reveals that interactions between quasi-periodic and random processes of production, diffusion and dissipation of a scalar field’s energy play an important role in the evolution of a major heatwave. Hence, the 2003 event was not just the result of a superposition of Rossby waves and eddy anomalies. Extremal length analysis thus reveals that the zonal advection of temperature anomalies was blocked by interactions between quasi-periodic and random production and diffusion processes. 
This study highlights the complex turbulent interactions that lead to major heat wave events, and the fact that each such event is thus unique.

How to cite: Fossa, M., Danaila, L., and Ghil, M.: Temperature at 500hPa, energy budget and the 2003 summer heatwave over Western Europe: a triple-decomposition approach, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-17846, https://doi.org/10.5194/egusphere-egu24-17846, 2024.

EGU24-17968 | ECS | Posters on site | NP1.1

Understanding the atmospheric kinetic energy spectrum 

Salah Kouhen, Benjamin Storer, Hussein Aluie, David Marshall, and Hannah Christensen

The Kinetic Energy spectrum of the atmosphere in the mesoscales (10-500 km) is poorly understood. Aircraft measurements in the eighties first revealed that there was a kink in the spectrum, a transition from a slope of -3 to a slope of -5/3, that occurred at scales below around 400 km (Nastrom et al. [1984]). Since that time many possible mechanisms have been posited for the transition but there has been no consensus. We will present a new way of analysing the local scaling laws of geophysical data using coarse-graining, extending the work of Sadek and Aluie [2018]. Our technique allows for the creation of spatial maps of spectral slope, as well as conditioned spectra that can be used to analyse the relationship between different meteorological variables and the atmospheric kinetic energy power spectrum. This enables us to explore causes for the observed shallower slope. We observe shallower spectral slopes in regions of greater convective activity, as well as shallowing in regions of high orographic variability and interesting latitudinal effects. The important implications of our work for the celebrated Nastrom and Gage spectrum (Nastrom et al. [1984]) will be discussed.

 

References: 

GD Nastrom, KS Gage, and WH Jasperson. Kinetic energy spectrum of large-and mesoscale atmospheric processes. Nature, 310(5972):36–38, 1984.

 

Mahmoud Sadek and Hussein Aluie. Extracting the spectrum of a flow by spatial filtering. Physical Review Fluids, 3(12):124610, 2018.

How to cite: Kouhen, S., Storer, B., Aluie, H., Marshall, D., and Christensen, H.: Understanding the atmospheric kinetic energy spectrum, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-17968, https://doi.org/10.5194/egusphere-egu24-17968, 2024.

EGU24-17987 | ECS | Posters on site | NP1.1

Surrogate-based data assimilation for microscale atmospheric pollutant dispersion 

Eliott Lumet, Mélanie Rochoux, Thomas Jaravel, and Simon Lacroix

Microscale pollutant dispersion is a critical aspect of air quality assessment with significant implications for the environment and public health. Designing accurate microscale dispersion models is of paramount importance for predicting air pollution exposure and assessing risks, in particular in emergency situations such as accidents at industrial sites, which are often close to densely-populated urban areas. However, this is a challenging task, as the structure and trajectory of pollutant plumes are strongly influenced by atmospheric flow, which is inherently multi-scale, turbulent, and interacts in complex ways with the built environment. To accurately capture the airflow and dispersion patterns induced by the built environment, large-eddy simulations (LES) are recognized as a high-fidelity numerical approach. However, LES are very costly and remain subject to uncertainties, partly due to the lack of knowledge and variability of the large-scale atmospheric forcing. In emergency situations, it is essential to quantify and reduce these uncertainties in order to better predict where pollutant concentration peaks occur.

In this work, to cope with the computational cost of LES, while providing the best possible information on the processes involved, we design and validate a data assimilation algorithm based on an ensemble Kalman filter (EnKF) that combines in situ concentration measurements with LES information. This LES information is obtained through a surrogate model, based on proper orthogonal decomposition (POD) combined with Gaussian process regression (GPR), which was trained in an offline stage using a large dataset of LES simulations, and which predicts the time-averaged concentration spatial fields for varying large-scale atmospheric conditions. The application of our data assimilation approach to the MUST field-scale experiment provides a proof-of-concept of the system's ability to reduce meteorological parametric uncertainties, correct bias in the model boundary conditions and thereby improve LES pollutant concentration field predictions. The use of the POD-GPR reduced-order model enables generating ensemble predictions that accurately account for the strong nonlinearities of the LES model, in just a few tens of seconds.

In addition, particular attention is paid to the representation of the errors, in particular to the internal variability of the atmospheric boundary layer that induces variability in the LES statistical fields and in the in-situ measurements. We design a bootstrap approach to quantify the significant effect of atmospheric internal variability on microscale dispersion, and we show that GPRs are able to learn this source of noise. Finally, we take internal variability into account in the data assimilation system considering that it is a source of model error and of observation error. This provides a more robust data assimilation framework with a more realistic description of the errors, which will be of interest for dispersion applications in real urban areas.

How to cite: Lumet, E., Rochoux, M., Jaravel, T., and Lacroix, S.: Surrogate-based data assimilation for microscale atmospheric pollutant dispersion, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-17987, https://doi.org/10.5194/egusphere-egu24-17987, 2024.

EGU24-18006 | ECS | Orals | NP1.1 | Highlight

Oscillatory Melancholia state of the Atlantic Meridional Overturning Circulation in an intermediate-complexity climate model 

Reyk Börner, Oliver Mehling, Jost von Hardenberg, and Valerio Lucarini

The Atlantic Meridional Overturning Circulation (AMOC) is considered a tipping element of the earth system featuring bistability: for a given external forcing, a strong and a weak circulation state coexist as competing attracting states of the system. In the presence of random fluctuations, noise-induced transitions between the competing states are possible, posing a risk of abrupt AMOC tipping even without crossing a critical forcing threshold. It is thus crucial to better understand the stability landscape of the earth system with a multistable AMOC, particularly the properties of the boundary separating the basins of attraction of the strong and weak AMOC states. For weak noise, transitions are expected to cross the basin boundary at so-called edge states or "Melancholia states", typically chaotic saddles which are attracting on the boundary but asymptotically unstable. Here we find an edge state between the two stable AMOC states in an earth system model of intermediate complexity, PlaSim-LSG. Our approach is based on an edge-tracking technique that allows to construct a pseudo-trajectory on the chaotic saddle. We characterize the climatic and dynamical properties of this edge state and map out its location in different projections of state space. Near the edge state, the AMOC strength exhibits strong transient oscillations which we link to the ongoing physical processes. We relate our findings to the theory of unstable chaotic sets and discuss implications for the predictability of potential AMOC tipping in the future.

How to cite: Börner, R., Mehling, O., von Hardenberg, J., and Lucarini, V.: Oscillatory Melancholia state of the Atlantic Meridional Overturning Circulation in an intermediate-complexity climate model, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-18006, https://doi.org/10.5194/egusphere-egu24-18006, 2024.

EGU24-18089 | ECS | Orals | NP1.1

Building Response Operators Using Koopman Formalism: the Link between Free and Forced Variability 

John Moroney, Valerio Lucarini, and Niccolo' Zagli

Linking free and forced variability is one of the key challenges in climate science. As the climate system is an out of equilibrium one, the standard application of the fluctuation-dissipation theorem is out of scope. It has been shown in the past that it is possible to construct response operators that can be used to perform climate change projections using a more general formulation of response theory for nonequilibrium systems. Nonetheless, such operators lack the key property of interpretability: one cannot separate the contribution to the total response coming from different modes of natural variability of the system. We show here in a few low-dimensional models how this issue can be taken care of by taking advantage of the Koopman formalism. One can then write the response operator as a sum of terms each associated with a specific mode of variability. The obtained results also shed light on previous findings by Hasselmann and colleagues and on recently proposed data-driven methods aimed at deriving response operators from data.

How to cite: Moroney, J., Lucarini, V., and Zagli, N.: Building Response Operators Using Koopman Formalism: the Link between Free and Forced Variability, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-18089, https://doi.org/10.5194/egusphere-egu24-18089, 2024.

EGU24-18525 | ECS | Posters on site | NP1.1

Characterising Edge States: Measures on chaotic non-attracting invariant sets 

Raphael Roemer and Peter Ashwin

In this work, we explore how to extend the concept of physical measures from attractors to chaotic non-attracting invariant sets. Building on Sweet and Ott’s work from 2000, we make their ideas rigorous by defining a measure on non-attracting sets in terms of Lebesgue Measure and show how to sample it numerically. We discuss its relevance for simple climate models and the sampling techniques’ limitations in the context of more complex and higher dimensional (climate) models. Knowing the measure of a non-attracting set, for example of a saddle or of an edge state (also known as melancholia state), also provides information about its fractal dimension and geometric complexity which can be useful to better understanding tipping phenomena and uncertainty close to a basin boundary.

How to cite: Roemer, R. and Ashwin, P.: Characterising Edge States: Measures on chaotic non-attracting invariant sets, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-18525, https://doi.org/10.5194/egusphere-egu24-18525, 2024.

There is growing concern that various large-scale elements of the Earth system may undergo catastrophic transitions under future climate change. But already at present-day conditions there is a risk of spontaneous transitions to an undesired state induced by stochastic fluctuations, given that any of those elements occupy a multi-stable regime. For many climate sub-systems this potential present-day multi-stability is still uncertain. For instance, it cannot be ruled out that there is a regime of a collapsed Atlantic Meridional Overturning Circulation (AMOC) that is stable under present-day conditions. Assuming such an undesired stable state exists, there also exists an additional unstable state called the edge state. This state anchors the basin boundary separating the desired and undesired regimes, and it lies at the heart of the path taken by the system during a noise-induced transition between the two stable states.

In this work such an edge state lying between the stable regimes of a vigorous and a collapsed AMOC is computed for the first time in a global ocean model using an edge tracking algorithm. The physical characteristics that set this state apart from the usually observed stable regimes are analyzed. This can be useful to detect if a spontaneous collapse of the AMOC induced by stochastic climate variability is underway, or to detect so-called rate-induced tipping. It may be especially helpful if the system is close to the tipping point where the desired state loses stability. Here the desired stable state and the edge state become increasingly similar, but a transition towards the undesired state may nevertheless be detected early-on if specific signatures of the edge state are recognized.

How to cite: Lohmann, J.: Characterization of Edge States as Gateway to a Collapse of the Atlantic Ocean Circulation, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-18809, https://doi.org/10.5194/egusphere-egu24-18809, 2024.

EGU24-19838 | Posters on site | NP1.1

Sensitivity analysis with external stochastic forcings for robust calibration : application to a water and pesticide transfer model 

Claire Lauvernet, Katarina Radisic, and Arthur Vidard

Environmental and hydrological models have become important decision-making tools. The PESHMELBA model [1] is a pesticide transfer model used to simulate and compare possible land-use planning scenarios in order to identify developments that reduce the impact of pesticides in surface waters.

When calibrating environmental models, one of the first steps is a sensitivity analysis [2]. However, this analysis can vary for different realizations of the external conditions (such as rainfall, evapotranspiration, or date of pesticide application) under which the model operates. Indeed, even the external conditions as an uncontrollable stochastic quantity means that the hydrological model itself becomes stochastic. Sobol indices can then be seen as random variables [3], where randomness is given by their dependence on rainfall.

In this study, we calculate the sensitivity indices on two examples: first, on soil hydrodynamic parameters on a vineyard plot in PESHMELBA under several realizations of the same type of rainfall, corresponding to events measured on the field. Second, on a omore complex situation considering the whole watershed and pesticides output variables. The stochasticity of that second case comes from the difference between the rainfall and the date of pesticide application, whicih is a key unknown in pesticide transfer simulations.

We show that the hierarchy of input parameters varies according to the forcings used. In particular, heavier rainfall mainly activates processes in the deep saturated horizon, involving parameters governing saturated soil properties (water content at saturation, for example), which is not the case for lighter rainfall, for which PESHMELBA is essentially influenced by unsaturated soil parameters.

The aim of this work is to take this dependency into account within the sensitivity analysis and to propose a global indice which is valid condidering the forcings uncertainties.

[1] Rouzies, E., Lauvernet, C., Barachet, C., Morel, T., Branger, F., Braud, I., & Carluer, N. (2019). From agricultural catchment to management scenarios: A modular tool to assess effects of landscape features on water and pesticide behavior. Science of The Total Environment, 671, 1144–1160. https://doi.org/10.1016/j.scitotenv.2019.03.060

[2] Mai, J. (2023). Ten strategies towards successful calibration of environmental models. Journal of Hydrology, 620, 129414. https://doi.org/10.1016/j.jhydrol.2023.129414

[3] Hart, J. L., Alexanderian, A., & Gremaud, P. A. (2017). Efficient Computation of Sobol' Indices for Stochastic Models. SIAM Journal on Scientific Computing, 39(4), A1514–A1530. https://doi.org/10.1137/16M106193X

How to cite: Lauvernet, C., Radisic, K., and Vidard, A.: Sensitivity analysis with external stochastic forcings for robust calibration : application to a water and pesticide transfer model, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-19838, https://doi.org/10.5194/egusphere-egu24-19838, 2024.

EGU24-19974 | Orals | NP1.1

Ocean eddy parameterizations: Stochastic and deterministic approaches for kinetic energy injection 

Stephan Juricke, Ekaterina Bagaeva, Sergey Danilov, Christian Franzke, and Marcel Oliver

In this study, we introduce a variety of ocean eddy parameterizations and discuss how they affect the representation of mesoscale turbulence in ocean models. They ultimately all aim at reducing overdissipation at eddy-permitting resolutions by utilizing the inverse energy cascade and energy conversions between potential and kinetic energy.

Mesoscale eddies play a crucial role in the global oceans. They transport tracers and heat, cascade energy across scales and interact with the mean currents and the atmosphere. However, their representation at resolutions close to the Rossby radius of deformation is insufficient. Such eddy-permitting, i.e. barely eddy resolving grids are still commonly applied for decadal climate simulations and will remain state-of-the-art at high latitudes for years to come. These simulations generally suffer from an excessive dissipation of kinetic energy, leading to reduced eddy variability, eddy formation and eddy-mean flow interactions.

Reducing overdissipation via optimized viscous closures is one way forward. Another option is to reinject some of the overdissipated energy back into the resolved flow via so called kinetic energy backscatter parameterizations. We will investigate different methods how to complement our own viscous and backscatter schemes with stochastic components, to account for unresolved chaotic variations of dissipative processes and for scale interactions across the resolution limit. For this purpose, we use data informed approaches such as linear inverse models to generate stochastic patterns based on high resolution reference simulations of idealized channel and double gyre configurations. Our results show that incorporation of such schemes can help to substantially improve the kinetic energy and mean flow characteristics. Furthermore, the varied application of the noise can reveal pathways of energy conversion between potential and kinetic energy, shedding light on the simulated energy cascades at such model resolutions. Aside from the learned construction of the stochastic patterns based on high resolution data, these new schemes come at a small additional computational cost, especially compared to higher resolution simulations. When tuned with caution, they provide a means to incorporate model uncertainty and to reduce systematic biases in ocean models.

How to cite: Juricke, S., Bagaeva, E., Danilov, S., Franzke, C., and Oliver, M.: Ocean eddy parameterizations: Stochastic and deterministic approaches for kinetic energy injection, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-19974, https://doi.org/10.5194/egusphere-egu24-19974, 2024.

EGU24-21689 | ECS | Orals | NP1.1

Intraseasonal atmospheric variability under climate trends 

Bernardo Maraldi, Henk Dijkstra, and Michael Ghil

Low-order climate models have played an important role in understanding the low-frequency variability of the atmospheric circulation and how it can be affected by trends in anthropogenic forcing. A simple quasi-geostrophic model of the midlatitudes’ circulation (Lorenz, Tellus, 1984, 1990) is studied from the perspective of the theory of nonautonomous dynamical systems (NDS: e.g., Ghil et al., Physica D, 2008).

We start with a study of the model’s behavior in the absence of time dependent forcing and determine in this case its steady states. A bifurcation analysis is carried out in order to identify distinct regime behavior types – stationary, periodic and chaotic – in the model’s parameter space. Next, we study the nonautonomous system with a meridional temperature gradient that varies seasonally, according to changes in insolation. The snapshot attractor (Tel et al., JSP, 2020) of the seasonally forced model is compared with the standard forward attractor of the autonomous model for two distinct epochs of the year, at peak summer and peak winter. In both cases, the effects of the change in forcing are reflected in a clear change of shape of the attractor. Predictability is lost in both cases: the summer attractor loses its periodicity when the forcing is seasonal. The winter one favors energy transport through one of the two wave components included in the model. For the same value of the forcing, the structure of the attractor in the autonomous case is quite different from that in the nonautonomous one.

Finally, the meridional forcing is subjected to climate trends, both positive and negative, since the jet intensity changes in opposite directions at low and high altitudes (Lee et al., Nature, 2019). The analysis of the snapshot attractor of the system under climate trends suggests that the model does not follow the geostrophic assumption in certain ranges of the forcing, as the average zonal flow does not always show the expected dependence on the equator-to-pole temperature contrast. On the other hand, the energy transported by the eddies does follow the sign of the climatic trend. Overall, distinct effects are observed. Chaotic behavior can be completely suppressed in favor of a regularly periodic one and vice-versa. At the same time, circulation patterns can change, suddenly disappear, and be restored.

In general, the snapshot attractor proved to be a robust tool in studying the internal variability of the midlatitude circulation, as well as the changes arising in it from anthropogenic forcing trends. The distinct regimes of behavior are being examined more closely by advanced spectral analysis methods (Ghil et al., Rev. Geophys., 2002) to better understand the effects of climate trends on low-frequency variability.

How to cite: Maraldi, B., Dijkstra, H., and Ghil, M.: Intraseasonal atmospheric variability under climate trends, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-21689, https://doi.org/10.5194/egusphere-egu24-21689, 2024.

EGU24-1852 | ECS | Posters on site | ITS4.3/NP1.2

Anticipating critical transitions in multi-dimensional systems driven by time- and state-dependent noise 

Andreas Morr, Keno Riechers, Leonardo Rydin Gorjão, and Niklas Boers

When approaching a one-parameter bifurcation, the feedbacks that stabilise the initial state weaken and eventually vanish; a process referred to as critical slowing down (CSD). This motivates the use of variance and lag-1 auto-correlation as indicators of CSD in order to anticipate bifurcation-induced critical transitions. Both indicators require a prior dimension reduction to a one-dimensional time series. The use of variance is further limited to time- and state-independent driving noise, strongly constraining its generality. Here, we propose a data-driven approach based on deriving a multi-dimensional Langevin equation to detect local stability changes and anticipate bifurcation-induced transitions in systems with generally time- and state-dependent noise. Our approach substantially generalizes the conditions underlying existing early warning indicators, which we showcase in the example of a two-dimensional predator-prey model. This reduces the risk of false and missed alarms significantly and allows for a more holistic understanding of the multi-dimensional system at hand.

How to cite: Morr, A., Riechers, K., Rydin Gorjão, L., and Boers, N.: Anticipating critical transitions in multi-dimensional systems driven by time- and state-dependent noise, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-1852, https://doi.org/10.5194/egusphere-egu24-1852, 2024.

The Atlantic Meridional Overturning Circulation (AMOC) is projected to weaken due to the increase in buoyancy caused by anthropogenic warming and consequent freshening during the 21st century and beyond. Atmosphere-ocean general circulation models simulate this AMOC weakening due to warming of the surface ocean, changes in the hydrological cycle that shift the North Atlantic salt budget, melting sea-ice, and changes in the atmospheric circulation. However, the freshwater contribution from the melting Greenland ice sheet is often either only considered in idealized scenarios or entirely omitted due to computational constraints. This simplification contributes to the large uncertainty surrounding the possibility of the AMOC crossing a tipping point in the forthcoming centuries. Here we employ the fully coupled Earth system model of intermediate complexity Bern3D v3, which dynamically simulates all ice-ocean-atmosphere interactions. We conduct a set of simulations driven by idealized CO2 concentration paths to investigate the impact of the melting Greenland ice sheet on the stability of the AMOC over the next 3000 years. We find that for a slow CO2 increase of 0.5%/yr up to twice pre-industrial levels, the general trends of the AMOC evolution are independent of whether Greenland meltwater is taken into account, with an initial weakening, but long-term recovery. Yet, the additional meltwater results in a further weakening of about 3 Sv after 100 years, but without leading to a full collapse of the circulation. This effect is due to melt rates remaining relatively low for the initial 100 years and only reaching their peak after 500 years. In the long-term, the curtailed AMOC and hence northward heat transport substantially slows down the disintegration of the Greenland ice sheet. Only in scenarios where the melt rates are kept artificially high, the AMOC does not recover. This highlights that the meltwater-induced AMOC weakening stabilizes the Greenland ice sheet, which in turn limits further AMOC weakening. This suggests that the potential for cascading interactions may be limited.

How to cite: Pöppelmeier, F. and Stocker, T. F.: Impact of future Greenland ice sheet melt on the stability of the Atlantic Meridional Overturning Circulation, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-3874, https://doi.org/10.5194/egusphere-egu24-3874, 2024.

EGU24-4021 | ECS | Orals | ITS4.3/NP1.2

Optimal Transition Paths for AMOC Collapse and Recovery in a Stochastic Box Model 

Jelle Soons, Tobias Grafke, and Henk A. Dijkstra

The present-day Atlantic Meridional Overturning Circulation (AMOC) is considered to be a prominent tipping element and its collapse would have grave consequences on the global climate. Hence, it is important to determine probabilities and pathways for noise-induced tipping events. However, as there is no observational evidence for an AMOC transition over the historical period, a noise-induced transition is expected to be a rare event in models and simple Monte Carlo techniques are not suited for such low-probability events. Here, we use Large Deviation Theory to directly compute the most probable transition pathways for the collapse and recovery of the AMOC in a box model of the World Ocean calibrated to the FAMOUS-model, where we added stochastic freshwater forcing. This allows us to determine the physical mechanisms of noise-induced AMOC transitions. We show that the most likely path of an AMOC collapse starts paradoxically with a strengthening of the AMOC followed by an immediate drop within a couple of years due to a short but relatively strong freshwater pulse. The recovery on the other hand is a slow process, where the North Atlantic Ocean needs to be gradually salinified over a course of decades, and its dynamics are quite close to the recovery in a bifurcation tipping event. The proposed method provides several benefits, including an estimate of probability ratios of collapse between various freshwater noise scenarios, showing that the AMOC is most vulnerable to freshwater forcing into the Atlantic thermocline region.

How to cite: Soons, J., Grafke, T., and Dijkstra, H. A.: Optimal Transition Paths for AMOC Collapse and Recovery in a Stochastic Box Model, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-4021, https://doi.org/10.5194/egusphere-egu24-4021, 2024.

EGU24-5870 | ECS | Orals | ITS4.3/NP1.2

Uncertainties too large to predict tipping times of major Earth system components 

Maya Ben Yami, Andreas Morr, Sebastian Bathiany, and Niklas Boers

Observations are increasingly used to detect critical slowing down (CSD) in potentially multistable components of the Earth system in order to warn of forthcoming critical transitions in these components. In addition, it has been suggested to use the statistical changes in these historical observations to extrapolate into the future and predict the tipping time. We argue that this extrapolation is too sensitive to uncertainties to give robust results. In particular, we raise concerns regarding (1) the modelling assumptions underlying the approaches to extrapolate results obtained from analyzing historical data into the future, (2) the representativeness of individual time series representing the variability of the respective Earth system components, and (3) the effect of uncertainties and preprocessing of the employed observational datasets, with focus on non-stationary observational coverage and the way gaps are filled. We explore these uncertainties both qualitatively and quantitatively for the Atlantic Meridional Overturning Circulation (AMOC). We argue that even under the assumption that these natural systems have a tipping point that they are getting closer to, the different uncertainties are too large to be able to estimate the time of tipping based on extrapolation from historical data.

How to cite: Ben Yami, M., Morr, A., Bathiany, S., and Boers, N.: Uncertainties too large to predict tipping times of major Earth system components, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-5870, https://doi.org/10.5194/egusphere-egu24-5870, 2024.

EGU24-5998 | ECS | Orals | ITS4.3/NP1.2

North Atlantic Subpolar Gyre Deep Convection: A Tipping Point Reached Decades Ago? 

Joas Müller, Giuseppe Zappa, and Alessio Bellucci

Recent studies utilizing the CMIP5 and CMIP6 model ensembles reveal that the subpolar North Atlantic (NA) is prone to deep convection collapsing leading to abrupt cooling of sea surface temperatures. Consequently, the latest comprehensive study on tipping points and the first report on global tipping points include the subpolar gyre (SPG) deep convection on the list of core tipping elements of Earth’s climate system.

Here, we investigate the drivers and impacts of a collapse of deep convection in the subpolar NA and the role of internal variability using a coupled climate model large ensemble (namely, the CESM2-LE consisting of 100 ensemble members) under the SSP3-7.0 forcing scenario. We identify that freshening of surface conditions leads to the negative surface density anomaly, eventually resulting in the cessation of deep mixing and the abrupt cooling of sea surface temperatures. The ensemble shows abrupt cooling occurring approximately in 2045 with internal variability leading to a spread of ±11 years. In each ensemble member, the subpolar NA transitions to a new state without deep convection, colder sea surface temperatures, strongly reduced heat loss to the atmosphere, and large circulation changes.

Internal variability does not determine if, but when abrupt cooling occurs, suggesting a forced response to larger-scale changes and a potential tipping point to be reached decades before the prominent abrupt cooling event. We provide evidence for the collapse of deep convection being a component of a positive feedback mechanism resulting in the SPG circulation transitioning to a weaker state. Without deep convection at the center of the circulation, the density gradient-driven part of the gyre circulation vanishes and the circulation strength decreases by approximately 50 %. The tipping point of the subpolar NA is therefore reached decades prior to the abrupt cooling and abrupt cooling is an inevitable consequence of the tipping event.

This points towards a potential misconception concerning drivers of abrupt climate
change in the subpolar NA, connected tipping points, and their thresholds, highlighting
the necessity for clarifying research efforts in the future.

How to cite: Müller, J., Zappa, G., and Bellucci, A.: North Atlantic Subpolar Gyre Deep Convection: A Tipping Point Reached Decades Ago?, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-5998, https://doi.org/10.5194/egusphere-egu24-5998, 2024.

EGU24-7515 | Posters on site | ITS4.3/NP1.2

Structural controllability and management of cascading regime shifts 

Juan Rocha and Anne-Sophie Crépin

Abrupt transitions in ecosystems can be interconnected, raising challenges for science and management in identifying sufficient interventions to prevent them or recover from undesirable shifts. Here we use principles of network controllability to explore how difficult it is to manage coupled regime shifts. We find that coupled regime shifts are easier to manage when they share drivers, but can become harder to manage if new feedbacks are formed when coupled. Simulation experiments showed that both network structure and coupling strength matter in our ability to manage interconnected systems. This theoretical insights calls for an empirical assessment of cascading regime shifts in ecosystems and warns about our limited ability to control cascading effects.

How to cite: Rocha, J. and Crépin, A.-S.: Structural controllability and management of cascading regime shifts, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-7515, https://doi.org/10.5194/egusphere-egu24-7515, 2024.

EGU24-8618 | Orals | ITS4.3/NP1.2

Uncertainty quantification for overshoots of tipping thresholds 

Paul Ritchie and Kerstin Lux-Gottschalk

To tip or not to tip? Many subsystems of the Earth are at risk of undergoing abrupt transitions from their current stable state to a drastically different and often less desired state due to anthropogenic climate change. These so-called tipping events often present severe consequences for ecosystems and human livelihood that are difficult to reverse. One common mechanism for tipping to occur is via forcing and driving a nonlinear system beyond a critical threshold that signifies self-amplifying feedbacks inducing tipping. However, previous work has shown that it is possible to briefly overshoot a critical threshold and avoid tipping. Specifically, the peak distance of an overshoot and the time a system can spend beyond a threshold are governed by an inverse square law relationship. In the real world or complex models, critical thresholds and other system features determining the permitted overshoot are highly uncertain. In this presentation, we look at how such uncertainties affect the probability of tipping from the perspective of uncertainty quantification. We show the importance of constraining uncertainty in the location of the critical threshold and the linear restoring rate to the stable state to reduce the uncertainty in the probability of tipping. Using a simple box model for the Atlantic Meridional Overturning Circulation, we highlight the need to constrain the high uncertainty found in wind-driven fluxes represented by a diffusive time scale within the box model to reduce uncertainty in the tipping probability for overshoot scenarios. 

How to cite: Ritchie, P. and Lux-Gottschalk, K.: Uncertainty quantification for overshoots of tipping thresholds, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-8618, https://doi.org/10.5194/egusphere-egu24-8618, 2024.

EGU24-8927 | Posters on site | ITS4.3/NP1.2

Consistency of resilience indicators in terrestrial vegetation models 

Sebastian Bathiany, Lana Blaschke, Andreas Morr, and Niklas Boers

Terrestrial ecosystems are affected by climate change, deforestation and other human influences. There is concern that the resilience of these ecosystems, i.e. their ability to recover from perturbations, is thereby decreased and that their sensitivity to environmental change is increased. In the extreme case, this sensitivity could diverge at a “tipping point”, and propel systems into alternative states. A prominent example is the potential dieback of the Amazon rainforest and the transition to a savanna-like state.

The notion of resilience is a highly complex and multi-faceted concept. Ecological resilience theory and the mathematical properties of dynamical systems suggest that a number of different resilience quantifiers are related to each other, or even equivalent, which would allow improved “resilience monitoring” from space. For instance, indicators based on the phenomenon of “critical slowing down” (CSD) like variance and autocorrelation, and related indicators have been used to detect changes over time. In contrast to empirical recovery rates, these indicators do not require one to directly observe the recovery from rare extreme disturbances. Also, they do not rely on the observation or attribution of the responsible environmental drivers.

Based on the assumption that fluctuations in remotely sensed proxies of vegetation properties (like biomass or vegetation greenness) behave like iconic one-dimensional stochastic models (most importantly, the Ornstein-Uhlenbeck process), CSD-based indicators should be related to empirical recovery rates after perturbations, to the more general Kramers-Moyal coefficients rooted in statistical mechanics, and to the sensitivity of a dynamical equilibrium state to environmental change. It has been shown that in observations, the theoretically expected relationships between some of these measures roughly hold. At the same time, process-based models, as well as observations, can deviate from such simple stochastic models, e.g. when multiple plant types affect the resilience of an ecosystem but not its sensitivity to environmental change.

In our contribution, we show and discuss examples for such deviations in a global vegetation model LPJ. In addition, we compare resilience indicators across a number of state-of-the-art models from CMIP6 and compare the results to an assessment of observations, in order to separate limitations that are related to the practical measurement process (e.g. uncertainties related to retrieval algorithms) from limitations that are associated with unjustified theoretical assumptions. Our results are meant to guide resilience monitoring toward meaningful indicators and to focus on regions and observable properties that can warn of future loss of ecosystem services.

How to cite: Bathiany, S., Blaschke, L., Morr, A., and Boers, N.: Consistency of resilience indicators in terrestrial vegetation models, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-8927, https://doi.org/10.5194/egusphere-egu24-8927, 2024.

EGU24-11804 | ECS | Orals | ITS4.3/NP1.2

Origin in of the AMOC fresh water transport biases in a state-of-the-art climate model 

Elian Vanderborght, Henk Dijkstra, and René Westen

Recent quasi-equilibrium studies performed in the Community Earth System Model (CESM) have revealed a bi-stable regime of the Atlantic Meridional Overturning Circulation (AMOC) in this model. This suggests that the present-day AMOC might exist in a bi-stable regime, emphasizing the need for accurate predictions regarding the probability of an AMOC collapse over the next decades. However, the CESM exhibits notable biases, with a critical freshwater transport bias at 34°S in the Atlantic emerging as a key determinant of AMOC stability. Specifically, this bias enhances the stability of the AMOC, rendering the CESM unable to accurately predict the likelihood of AMOC tipping.

In this study, we establish a direct connection between the freshwater transport bias in the CESM and a corresponding freshwater content bias in the Indian Ocean. By investigating the detailed freshwater balance, we identify specific regions within the Indian Ocean that exert a significant influence on the Atlantic freshwater transport bias at 34°S. This quantitative analysis enables us to construct an optimal surface-flux correction, which reduces the model biases. This physics-based surface-flux correction allows us to adjust the AMOC to its correct stability regime in the CESM without imposing unrealistic flux adjustments

How to cite: Vanderborght, E., Dijkstra, H., and Westen, R.: Origin in of the AMOC fresh water transport biases in a state-of-the-art climate model, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-11804, https://doi.org/10.5194/egusphere-egu24-11804, 2024.

EGU24-12023 | Posters on site | ITS4.3/NP1.2

Fingerprinting the AMOC and predicting a collapse 

Peter Ditlevsen

 In a recent paper [2] we predicted a collapse of the AMOC as soon as mid-century at odds with assessments based on climate model scenarios. The prediction was based on the sub polar gyre fingerprint as a proxy for the AMOC as proposed by Ceasar et al. [2]. Several other fingerprints have been proposed, all showing early warning signals of a forthcoming tipping point [3]. Here we present a statistical analysis, optimally extracting the common signal in the different fingerprints in order to further solidify the assessments. 

[1] Ditlevsen, P., Ditlevsen, S. Warning of a forthcoming collapse of the Atlantic meridional overturning circulation. Nat Commun 14, 4254 (2023)

[2] Caesar, L., Rahmstorf, S., Robinson, A. et al. Observed fingerprint of a weakening Atlantic Ocean overturning circulation. Nature 556, 191–196 (2018)

[3] Boers, N. Observation-based early-warning signals for a collapse of the Atlantic Meridional Overturning Circulation. Nat. Clim. Chang. 11, 680–688 (2021)

How to cite: Ditlevsen, P.: Fingerprinting the AMOC and predicting a collapse, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-12023, https://doi.org/10.5194/egusphere-egu24-12023, 2024.

EGU24-12187 | ECS | Orals | ITS4.3/NP1.2

Anticipating rate-induced tipping by a deep learning framework 

Yu Huang, Sebastian Bathiany, Peter Ashwin, and Niklas Boers

Rate-induced tipping (R-tipping) occurs when the forcing rate changes too rapidly for the system to track its quasi-equilibrium state, leading to an unexpected collapse. Currently, there is a lack of valid early warning signals (EWS) for R-tipping, particularly in the presence of noise perturbations. To address this deficiency, we employ a deep learning algorithm to extract the high-order structures hidden within time series data before R-tipping occurs. Then the trained neural networks are taken to provide real-time EWS for R-tipping, demonstrating skillful forecasts with a substantially long lead time, surpassing the performance of conventional critical slowing down indicators. Our progress underscores the predictability of R-tipping, offering the potential to improve the ability to deduce the safe operating space for a wider spectrum of complex systems.

How to cite: Huang, Y., Bathiany, S., Ashwin, P., and Boers, N.: Anticipating rate-induced tipping by a deep learning framework, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-12187, https://doi.org/10.5194/egusphere-egu24-12187, 2024.

EGU24-12295 | ECS | Orals | ITS4.3/NP1.2

Analysis of Abrupt Changes in CMIP6 Models Using Edge Detection 

Sjoerd Terpstra, Swinda K.J. Falkena, Robbin Bastiaansen, Sebastian Bathiany, Henk A. Dijkstra, and Anna von der Heydt

Many potential tipping elements have been identified in the climate system over the last decade, although some of them are surrounded by large uncertainties. We perform an updated analysis of abrupt changes in current state-of-the-art climate models to re-evaluate the evidence of these shifts—whether they are tipping points or not. We examine all CMIP6 models (59 in total) under the 1pctCO2 scenario using a Canny edge detection method—adapted for spatiotemporal dimensions—to detect abrupt shifts in climate data. We perform this semi-automatic analysis on 83 two-dimensional variables of the ocean, atmosphere, and land. We aggregate the detected shifts that are connected spatially or temporally. This results in connected regions of abrupt shifts and allows us to map areas that are most at risk of these shifts according to CMIP6 models. We report statistics on number of abrupt changes detected, surface area of abrupt changes, and critical global mean temperature at which these abrupt changes occur. This is done for various climate subsystems and potential tipping elements, such as the Arctic sea ice, Antarctic sea ice and the North Atlantic subpolar gyre. We find evidence for abrupt changes in several systems, but not all models show them equally.

How to cite: Terpstra, S., Falkena, S. K. J., Bastiaansen, R., Bathiany, S., Dijkstra, H. A., and von der Heydt, A.: Analysis of Abrupt Changes in CMIP6 Models Using Edge Detection, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-12295, https://doi.org/10.5194/egusphere-egu24-12295, 2024.

EGU24-16007 | Posters on site | ITS4.3/NP1.2

Investigating Early Warning Signals in Climate Simulations using Complex Networks 

Laure Moinat, Jérôme Kasparian, and Maura Brunetti

Early Warning Signals (EWS) are indicators that can be used to anticipate tipping points i.e. abrupt changes in dynamical systems. Detecting EWS is a crucial part of climate science, especially in the context of climate change. Several methods are used to identify tipping points using time series of climate state variables (e.g. temperature, precipitation, etc), but few consider spatial correlations [1]. Spatial detection could identify the starting location of a transition process from a state to another and be directly applied to satellite observations. We consider different state variables on a numerical grid as a complex network, where grid points displaying correlation are connected and the temporal evolution of this network is studied. 
We seek for network properties that can be used as EWS when approaching the state transition. 

The network is generated and analysed using the pyUnicorn package [2], and compared to classical statistical methods. The networks are constructed using two methods: Pearson correlation coefficient and mutual information, allowing us to compare a linear and a causal approach. Multiple network indicators such as the degree of correlation, the average path length, and the area weighted connectivity are compared. To test the method robustness, we look at the network dependencies in terms of the time window, the interval over which the forcing is changed, and the effect of reducing the extent of the network (limited, for example, over polar or equatorial regions). These indicators show tipping points at the global scale, as simulated in a coupled-aquaplanet configuration with the MIT general circulation model, using as forcing parameter the atmospheric CO2 content or the input of solar energy [3] . The application of such indicators as EWS is discussed.

 

[1] van der Mheen et al. Geophysical Research Letters 40, 11  (2013)

[2] Donges et al. Chaos 25, 113101 (2015)

[3] Brunetti \& Ragon, Physical Review E 107, 054214 (2023)

How to cite: Moinat, L., Kasparian, J., and Brunetti, M.: Investigating Early Warning Signals in Climate Simulations using Complex Networks, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-16007, https://doi.org/10.5194/egusphere-egu24-16007, 2024.

EGU24-17709 | Posters on site | ITS4.3/NP1.2

Applicability of CSD-based resilience analyses to remotely sensed Vegetation Indices in the Tropics 

Lana Blaschke, Andreas Morr, Sebastian Bathiany, Fabian Telschow, Taylor Smith, and Niklas Boers

Tropical forests are vital for climate change mitigation as carbon sinks. Yet, research suggests that climate change, deforestation and other human influences threaten these systems, potentially pushing them across a tipping point where the tropical vegetation might collapse into a low-treecover state. Signs for this trend are reductions of resilience defined as the system's capability to recover from perturbations. When resilience decreases, according to dynamic system theory, a critical slowing down (CSD) induces changes in statistical measures such as the variance and the autocorrelation. This allows to indirectly examine resilience changes in the absence of observations of strong perturbations. Yet, deriving estimates of the statistical measures indicating resilience changes based on CSD impose several assumptions on the system under observation. For tropical vegetation, it is not obvious that these assumptions are fulfilled.

Additionally, the conditions of tropical rainforests pose difficulties on the observation of the vegetation. Among other factors, cloud cover, aerosols, and the dense vegetation hinder the reliable retrieval of Vegetation Indices (Vis), especially from data gathered in the optical spectrum. Thus, such data might not be suitable for resilience analyses based on CSD, even if the theory is applicable in principle.

We investigate the different assumptions of CSD and test them on a diverse set of remotely sensed VIs. Hereby, we establish a framework that allows to decide whether a specific dataset is appropriate for resilience analyses based on CSD.

How to cite: Blaschke, L., Morr, A., Bathiany, S., Telschow, F., Smith, T., and Boers, N.: Applicability of CSD-based resilience analyses to remotely sensed Vegetation Indices in the Tropics, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-17709, https://doi.org/10.5194/egusphere-egu24-17709, 2024.

EGU24-18039 | Orals | ITS4.3/NP1.2

Methodologies for climate tipping points analysis and risk assessments in TIPMIP 

Jonathan F. Donges, Donovan P. Dennis, Sina Loriani, Boris Sakschewski, Nico Wunderling, and Ricarda Winkelmann

The Tipping Point Modelling Intercomparison Project (TIPMIP) is an international initiative that aims to systematically improve our understanding of potential tipping dynamics in different components of the Earth system and to assess the associated uncertainties (www.tipmip.org). By linking and evaluating different models through a systematic framework, TIPMIP aims to fill critical knowledge gaps in Earth system and climate risks by improving their assessment at different levels of anthropogenic forcing and associated long-term commitments and irreversibilities. The Methods and Risk Assessment Working Group of TIPMIP will further develop the methodological foundations of this systematic approach to the study of tipping dynamics in domain-specific and coupled Earth system  numerical models to support future assessment reports and comprehensive risk analyses. In this contribution, we introduce the Methods and Risk Assessment Working Group within TIPMIP, and highlight relevant lines of methodological development to be pursued, including: (i) systematic and automated detection of tipping points and critical transitions in model output and Earth observation data for TIPMIP (e.g, the TOAD framework), (ii) detection of non-linear regime shifts in time series data for TIPMIP beyond amplitude shifts, e.g. transitions between more regular and more erratic variability (e.g. the pyunicorn toolkit), and (iii) probabilistic analysis and emulator approaches of risks for triggering tipping events and cascading tipping dynamics at different levels of anthropogenic forcing (e.g. the pycascades approach).

How to cite: Donges, J. F., Dennis, D. P., Loriani, S., Sakschewski, B., Wunderling, N., and Winkelmann, R.: Methodologies for climate tipping points analysis and risk assessments in TIPMIP, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-18039, https://doi.org/10.5194/egusphere-egu24-18039, 2024.

EGU24-18045 | ECS | Orals | ITS4.3/NP1.2

Tipping cascades, future Earth system trajectories and the prospect of a hothouse: insights from the SURFER model 

Victor Couplet, Marina Martínez Montero, and Michel Crucifix

A tipping cascade is a series of tipping events in the Earth system where transitions in one subsystem can trigger further transitions in other subsystems. A concern for the future is that such a cascade could lock the Earth system in a pathway towards a so-called hothouse state. We investigate this possibility with SURFER, a reduced complexity model with a process-based carbon cycle that can reliably predict CO2 concentrations, global mean temperatures, sea-level rise, and many ocean acidification metrics on timescales from decades to millions of years. We have incorporated in the model a network of interacting tipping elements and their feedback on the climate through albedo changes and additional greenhouse gas emissions. This has allowed for a systematic investigation of the effects of a family of realistic emission scenarios on the future trajectories of the Earth system. Our results show that a permanent shift to a hothouse state within the next few centuries is implausible. On longer time scales, however, tipping cascades can lead to enduring additional warming and particularly sea level rise.

How to cite: Couplet, V., Martínez Montero, M., and Crucifix, M.: Tipping cascades, future Earth system trajectories and the prospect of a hothouse: insights from the SURFER model, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-18045, https://doi.org/10.5194/egusphere-egu24-18045, 2024.

EGU24-18126 | ECS | Posters on site | ITS4.3/NP1.2

Cryosphere tipping elements decisive for tipping risks and cascading effects in the Earth system 

Jonathan Rosser, Ricarda Winkelmann, and Nico Wunderling

The Earth's climate system is a complex system that includes key components such as the Arctic Summer Sea Ice or the El Niño Southern Oscillation as well as climate tipping elements like the continental-scale ice sheets or the Amazon rainforest. Crossing warming thresholds of these elements can lead to a qualitatively different climate state, endangering the stability of human societies. Particularly, the cryosphere elements are vulnerable at current levels of global warming (1.2°C) while also having long response times and large structural uncertainties. Investigating a network of interacting Earth system components using an established conceptual model, we systematically assess which uncertainties of key Earth system component have the largest impacts on tipping risks. We find that the cryosphere tipping elements (the Greenland and the West Antarctica ice sheets) are most decisive for tipping risks and cascading effects within our model. At a global warming level of 1.5°C, neglecting the large cryosphere tipping elements can reduce the mean number of disintegrated Earth system components by as much as 56%. This is concerning as overshooting 1.5°C of global warming is fast becoming inevitable, while current state-of-the-art IPCC-type global circulation models do not (yet) include dynamic ice sheets. Our results suggest that urgent integrated Earth system model development and Earth observation efforts including the large polar ice sheets are necessary and a precautionary measure of meeting stringent climate targets is crucial to limit tipping risks.

How to cite: Rosser, J., Winkelmann, R., and Wunderling, N.: Cryosphere tipping elements decisive for tipping risks and cascading effects in the Earth system, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-18126, https://doi.org/10.5194/egusphere-egu24-18126, 2024.

Extreme cold winters have been projected to decrease in the future, although their impacts on society are still significant. The goal of this study is to assess whether climate change affects the atmospheric mechanisms leading to cold winters.

We first explore the dynamics of 15-day winter cold spells in France, as observed since 1950. We find that the most extreme events tend to have the same atmospheric circulation pattern, consisting of an eastward-shifted NAO- dipole. We calculate an atmospheric index that characterizes this dipole. Then, using a stochastic weather generator with importance sampling, we show that this is a sufficient condition to trigger extremely cold temperatures in France, and that it performs better than a classical North Atlantic Oscillation index. This suggests that a dipole of atmospheric circulation is a necessary and sufficient condition leading to extreme cold spells in France.

We use this atmospheric index to select the CMIP6 models that best reproduce the identified dynamics leading to extreme cold spells of 15 days. Using a stochastic weather generator with importance sampling, we run simulations of worst-case winter cold spells from 2015 to 2100, following different emission trajectories for the selected models. 15-day winter cold spells in France will reach less extreme temperatures at the end of the century, especially in the case of a high-emission scenario (SSP5-8.5). However, the simulated ensembles of extreme cold spells do not show the same warming trend as the mean temperature, and very extreme cold spells are still possible in the near future. The atmospheric circulation prevailing during these events is analyzed and compared with the circulation observed during previous events.

How to cite: Cadiou, C. and Yiou, P.: Assessing changes in the intensity and dynamics of extreme cold spells in France from CMIP6, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-622, https://doi.org/10.5194/egusphere-egu24-622, 2024.

EGU24-1181 | ECS | Posters on site | NP1.3

Evaluation of atmospheric circulation of CMIP6 models for extreme temperature events using Latent Dirichlet Allocation 

Nemo Malhomme, Bérengère Podvin, Davide Faranda, and Lionel Mathelin

Climate models aim at representing as closely as possible the statistical properties of the climate components, including extreme events on which fine-tuning data may be less available. This is a fundamental requirement to correctly project changes in their dynamics due to anthropogenic forcing.
In order to evaluate how closely models match observations, we need algorithms capable of selecting, processing and evaluating relevant dynamical features of the climate components. This has to be reiterated efficiently for large datasets such as those issued from the Coupled Model Intercomparison Project 6 (CMIP6). In this work, we use Latent Dirichlet Allocation (LDA), a statistical soft clustering method initially designed for natural language processing, to extract synoptic patterns from sea-level pressure data and evaluate how close the dynamics of CMIP6 climate models are to ERA5 reanalysis, both in the general case and in the case of extreme temperature events.

LDA allows for learning a basis of decomposition of maps into objects called "motifs". From the ERA5 sea-level pressure data, the method robustly extracts a basis of motifs that are interpretable objects at synoptic scale, i.e. cyclones or anticyclones. Pressure data can be projected onto this basis, yielding motif weights that contain local information about the large-scale atmospheric circulation. LDA decomposition is efficient and sparse: most of the information of a given map is contained in few motifs. It is therefore possible to decompose any map in a limited number of easy-to-interpret synoptic objects. This allows for a variety of new angles for statistical analysis.

The weights statistics can be used to characterize the general and extreme dynamics in reanalysis and model data. By comparing the statistics obtained from reanalysis data with those obtained from a selection of CMIP6 models, we can quantify errors on each localized circulation pattern and identify model-specific and model-agnostic errors. We found that, on average, large-scale circulation is well predicted by the models, but model errors are increased for extreme events such as heatwaves and cold spells. Additionally, Mediterranean motifs were found to be associated with significant model errors for all the considered models in all cases.

How to cite: Malhomme, N., Podvin, B., Faranda, D., and Mathelin, L.: Evaluation of atmospheric circulation of CMIP6 models for extreme temperature events using Latent Dirichlet Allocation, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-1181, https://doi.org/10.5194/egusphere-egu24-1181, 2024.

EGU24-2460 * | Orals | NP1.3 | Highlight

The outstanding European and Mediterranean heatwave activity during summer 2022 

Ricardo Trigo, David Barriopedro, José Garrido-Pérez, Amelie Simon, Sandra Plecha, Ana Teles-Machado, Ana Russo, and Ricardo García-Herrera

The European summer of 2022 has been widely recognized as the warmest since mid-19th century. Our updated analyses of instrumental and reconstructed temperature series since 1500 indicate that the European summer (June-to-August) of 2022 was the warmest on record, exceeding the previous hottest summer of 2021 by a large margin. In fact, the past three summers of 2021–2023 have been among the hottest ones of the last five centuries.

By applying a heatwave (HW) detection algorithm to reanalysis data, we identify three large European HW events that affected ample regions of the continent in mid-June, mid-July and August/early September 2022. These episodes were triggered by high-pressure systems with noticeable differences in their characteristics. Additional analyses confirm that high-latitude blocks were largely responsible for the August 2022 HW, whereas subtropical ridges dominated during the June and July 2022 HWs. These HW events were also accompanied by dry soils and warm Sea Surface Temperatures (SSTs) over the Mediterranean. Indeed, summer 2022 displayed the largest marine heatwave activity of the 1982–2023 period due to an unusually high frequency of long-lasting and intense events, particularly over western Mediterranean.

Taking the June 2022 HW over Iberia as an example, we address the role of dynamical (atmospheric circulation) and thermodynamical (regional soil moisture and western Mediterranean SSTs) drivers in the severity of the event. Flow analogues of the June 2022 heatwave are used to reconstruct the expected temperatures under different combinations of these thermodynamical drivers and assess their separate and combined influences on the intensity of the event. Results show a measurable intensification of the heatwave event (of ~1 °C) by both dry land and warm sea conditions. Although these two drivers are significantly correlated, southwestern European HWs are aggravated if dry soils concur with warm SSTs over western Mediterranean. The magnitude of the Mediterranean SST influence could depend on the soil moisture state, being larger for dry than wet conditions, as well as on the atmospheric circulation.

R.M.T., A.R., S.P. and A.T.M. thank Fundação para a Ciência e a Tecnologia (FCT) I.P./MCTES through national funds (PIDDAC) – UIDB/50019/2020 (https://doi.org/10.54499/UIDP/50019/2020) and LA/P/0068/2020 (https://doi.org/10.54499/LA/P/0068/2020). A.R. and R.M.T. thank also FCT (https://doi.org/10.54499/2022.09185.PTDC, (http://doi.org/10.54499/JPIOCEANS/0001/2019). A.R. was supported by FCT through https://doi.org/10.54499/2022.01167.CEECIND/CP1722/CT0006.

How to cite: Trigo, R., Barriopedro, D., Garrido-Pérez, J., Simon, A., Plecha, S., Teles-Machado, A., Russo, A., and García-Herrera, R.: The outstanding European and Mediterranean heatwave activity during summer 2022, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-2460, https://doi.org/10.5194/egusphere-egu24-2460, 2024.

EGU24-3164 | Orals | NP1.3 | Highlight

Global emergence of regional heatwave hotspots outpaces climate model projections 

Kai Kornhuber, Samuel Bartusek, Richard Seager, and Mingfang Ting

Multiple recent record shattering weather events raise questions about the adequacy of climate models to effectively predict and prepare for unprecedented climate impacts on human life, infrastructure, and ecosystems.

While it appears that  the record breaking global mean temperatures of 2023 are still within the models spread, we show that climate models fail to reproduce the observed emergence of several heatwave hotspots on impact relevant temporal and spatial scales. We identify several heatwave hotspots worldwide in which the warming of the tail of the distribution outpaces the mean warming. In these regions we find that the observed trends by far exceed what is projected by the multi-model mean of state-of-the-art model frameworks. In multiple regions, observed trends are outside of the model spread and persist irrespective of resolution or model design: We investigate models from HighResMip project (49 members) and CAM6 and the ECHAM5 SST forced large ensemble (60 members) and find that biases persist throughout the vast number of climate model experiments with different architectures (fully coupled and forced with observed SSTs), resolutions (25 km – 250km), and large numbers of realizations (Kornhuber et al. submitted).

We discuss potential reasons for the models’ shortcomings. Recently we found that models tend to underestimate the surface anomalies of specific deep atmospheric circulation patterns, which were fundamental to some of the most extreme heat waves in the past (Kornhuber et al. Nat. Comms. 2023). Europe is identified as a heatwave hotspot, driven in part by trends in atmosphere dynamics (Rousi, Kornhuber et al. Nat. Comms. 2022), which models tend to understimate (Vautard et al. Nat. Comms 2023). In addition, non-linear interactions between high pressure, soil moisture deficiencies and temperature have been shown to be pivotal for the record shattering Pacific Northwest heatwave, amplifying its magnitude by 40% (Bartusek, Kornhuber, Ting, Nat. Climate Change 2022). Such dependence structures might not be accurately represented in most models. We conclude that while climate models have been useful and on point regarding their global mean temperature response to increased greenhouse gas concentrations, trends in the far-end tails of the temperature distribution are not well captured, missing emerging high-risk hotspots. This highlights the need to better understand and model the drivers of extreme heat and to rapidly mitigate greenhouse gas emissions to avoid further harm from climate surprises.  

How to cite: Kornhuber, K., Bartusek, S., Seager, R., and Ting, M.: Global emergence of regional heatwave hotspots outpaces climate model projections, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-3164, https://doi.org/10.5194/egusphere-egu24-3164, 2024.

EGU24-3633 | Orals | NP1.3

The Impact of Gulf Stream Moisture Flux Suppression on Atmospheric Blocks 

Jamie Mathews, Arnaud Czaja, Frederic Vitart, and Christopher Roberts

In this study, we explore the impact of oceanic moisture fluxes on atmospheric blocks using the ECMWF Integrated Forecast System (IFS). Artificially suppressing surface latent heat flux over the Gulf Stream region leads to a significant reduction (up to 30%) in atmospheric blocking frequency across the northern hemisphere. Affected blocks show a shorter lifespan (-6%), smaller spatial extent (-10%), and reduced intensity (-0.4%), with an increased detection rate (+14%). These findings are robust across various blocking detection thresholds. Analysis indicates a resolution-dependent response, with resolutions lower than Tco639 (∼18km) showing no significant change in blocking characteristics. Exploring the broader Rossby wave pattern, we observe that diminished moisture flux favors eastward propagation and higher zonal wavenumbers, while oceanic influence promotes stationary and westward-propagating waves with zonal wavenumber 3. This study underscores the critical role of western boundary current’s moisture fluxes in modulating atmospheric blocking and associated Rossby wave dynamics.

How to cite: Mathews, J., Czaja, A., Vitart, F., and Roberts, C.: The Impact of Gulf Stream Moisture Flux Suppression on Atmospheric Blocks, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-3633, https://doi.org/10.5194/egusphere-egu24-3633, 2024.

EGU24-5138 | ECS | Posters on site | NP1.3

Assessment of the predictability of cold-wet-windy Pan Atlantic compound extremes 

Meriem Krouma and Gabriele Messori

Occurrence of cold spells in different North American regions has been related to concurrent wet and windy extremes in Western Europe. This link is driven by an anomalous state of the North Atlantic storm track. Two dynamical pathways have been defined as potential origins of the Pan-Atlantic compound extremes. The first pathway is linked to a Rossby wave train propagating from the Pacific toward the Atlantic, associated with a pronounced Alaskan ridge. The second pathway is characterized by the presence of a high west of Greenland, that favors simultaneously a southward displacement of a trough over eastern USA and an upper-level trough over South western Europe. The aim of this study is to assess the predictability of these two pathways in the ERA5 reanalysis using dynamical systems indicators. These indicators are the local dimension and the persistence of the large-scale atmospheric flow, and can be used as proxies for the predictability of each pathway. We complement this analysis using the ECMWF ensemble reforecasts at different lead times, and computing skill scores for the two pathways.

How to cite: Krouma, M. and Messori, G.: Assessment of the predictability of cold-wet-windy Pan Atlantic compound extremes, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-5138, https://doi.org/10.5194/egusphere-egu24-5138, 2024.

EGU24-5328 | ECS | Orals | NP1.3

Persistence and the practical predictability of surface temperatures 

Emma Holmberg, Steffen Tietsche, and Gabriele Messori

Extreme temperature events can cause severe disruptions to society from negative health consequences to infrastructure damage. Early warning systems are a key element in the mitigation of such impacts, with literature highlighting the potential societal benefit of information from sub-seasonal forecasts. We investigate the relationship between the persistence of an atmospheric state and the practical predictability of surface temperatures focusing on medium to extended range time scales. Persistence is assessed objectively leveraging techniques from dynamical systems theory whilst practical predictability is defined in terms of the forecast error in surface temperature. Atmospheric persistence provides potential value for the practical predictability of temperature in some cases with the results varying depending on season and location. Wintertime temperature forecasts at lead times up to three weeks, and cold spell forecasts up to two weeks in lead time are highlighted as cases where persistence appears to show an association with practical predictability.

How to cite: Holmberg, E., Tietsche, S., and Messori, G.: Persistence and the practical predictability of surface temperatures, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-5328, https://doi.org/10.5194/egusphere-egu24-5328, 2024.

EGU24-5796 | ECS | Orals | NP1.3 | Highlight

Future changes in spatially compounding hot, wet or dry events and their implications for the world's breadbasket regions 

Bianca Biess, Lukas Gudmundsson, Michael Windisch, and Sonia I. Seneviratne

Recent years were characterized by spatially co-occurring hot, wet or dry years around the globe. Spatially compounding extreme events can strongly amplify societal impacts as economic supply chains are increasingly interlinked, highlighting the increasing importance of advancing our knowledge of the effects of human-induced climate change on such events. We assess the occurrence of spatially compounding hot, wet and dry years under different future warming levels of 1.5°C , 2°C, 3°C and higher levels of global warming. We focus our analysis on the top-producing agricultural regions that have historically provided the global food systems with large quantities of wheat, maize, soybean or rice. The occurrence of spatially compounding events and area affected in future climates is determined using Earth System Model simulations from the 6th Phase of the Coupled Model Intercomparison Project (CMIP6). The simulations project a strong increase in the global land area that is concurrently affected by hot, wet and dry extremes under continued global warming. On regional scales, the world’s breadbasket regions are particularly affected by strong increases in the simultaneous occurrence of hot, wet or dry extremes under continued global warming. The spatial extent of agricultural land potentially threatened by climate extremes will increase drastically if global mean temperatures shift from +1.5 °C to +2.0 °C, and will be further amplified with every tenth of degree of warming. This highlights that ambitious climate action needs to be taken in order to limit global warming if we want to keep the global agricultural land in a safe climatic space.

How to cite: Biess, B., Gudmundsson, L., Windisch, M., and Seneviratne, S. I.: Future changes in spatially compounding hot, wet or dry events and their implications for the world's breadbasket regions, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-5796, https://doi.org/10.5194/egusphere-egu24-5796, 2024.

EGU24-6492 | ECS | Posters on site | NP1.3

Statistics of pan-oceanic extreme near-surface winds 

Aleksa Stanković and Rodrigo Caballero

We investigate statistics of extreme 10 m winds over the midlatitude Atlantic, Pacific and Southern Ocean basins, regions associated with the extratropical storm tracks. We compute  statistics of 10 m wind speed using  reanalysis (ERA5) and satellite scatterometer datasets (NOAA NCEI Blended Seawinds), as well as the outputs from CMIP6 climate models. To select the regions with climatologically strong winds, we study the 10 m wind speeds over the oceans only in the regions where the local 98th percentiles exceed 20 ms -1

Annually, the median of 10 m wind speed distribution is the highest in the Southern Ocean, while the extreme winds (starting from 90th percentile) are higher over the oceans in the Northern Hemisphere. The hemispheric differences in the extreme winds are greater and more evident during the respective winter seasons, potentially indicating  differences in the dynamics of extreme winter storms. These findings are consistent over all data products analyzed. Additionally, tails of distributions of winds at 850 hPa in the basins during the winter calculated from reanalysis and observations mirror the patterns observed in 10 m wind distributions, pointing to the influence of large-scale processes in creating stronger extreme winds over the Northern Hemisphere.

How to cite: Stanković, A. and Caballero, R.: Statistics of pan-oceanic extreme near-surface winds, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-6492, https://doi.org/10.5194/egusphere-egu24-6492, 2024.

EGU24-6511 | Posters on site | NP1.3

Development of a New Database of Extreme European Winter Windstorms Derived from Multiple Data Sets 

Clare Flynn, Julia Mömken, Joaquim Pinto, and Gabriele Messori

European winter windstorms can pose significant risks for the safety and lives of people living in their paths as well as to infrastructure and the natural environment. Several storms in the recent past have caused substantial damages, and risks from extreme winter windstorms may increase with climate change. Characterizing the risks and potential losses from such storms and assessing our ability to predict storm economic losses are therefore an utmost priority. To that end, we have developed a new database of extreme European winter windstorm footprints for the extended winter season (ONDJFM) for the period 1995-2015, and have made it publicly available to both the scientific community and industry. In contrast to previously compiled databases, our database includes storm footprints derived from four different data sets and not from a single source: the ERA5 reanalysis, the COSMO-REA6 reanalysis for Europe, simulation output from a regional climate model driven by ERA5 on the EURO-CORDEX domain, and simulation output from the regional climate model COSMO-CLM on an enlarged Germany domain. We included both the footprints themselves, expressed as the relative daily maximum wind gusts associated with a storm event, and the absolute daily maximum wind gusts associated with that footprint. We derived and included the storm footprints associated with the 50 most extreme storms, or Top50 storms, identified within each of the four input data sets. We applied a consistent methodology for identifying storm footprints and assessing their severity across input data sets that does not require downscaling or adjustment with the assistance of an atmospheric or statistical model. This provides for greater comparability among the footprints derived from the different input sources. Lastly, we derived the Top50 storms from each input on its native horizontal resolution, allowing us to characterize the impact that horizontal resolution can have on footprint identification and severity assessment. Our database thus allows for assessment of extreme storms and their impacts from several perspectives, particularly the impacts from use of wind gust data derived on different horizontal resolutions. This complements the existing extreme European winter storm databases, and facilitates scientific research on extreme storms and industry catastrophe modelling assessments.

How to cite: Flynn, C., Mömken, J., Pinto, J., and Messori, G.: Development of a New Database of Extreme European Winter Windstorms Derived from Multiple Data Sets, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-6511, https://doi.org/10.5194/egusphere-egu24-6511, 2024.

EGU24-7539 | ECS | Posters on site | NP1.3

Future Changes in European windstorm loss and seasonal loss clustering in the EURO-CORDEX dataset 

Inovasita Alifdini, Julia Moemken, Alexandre M. Ramos, Aleksa Stankovic, Rodrigo Caballero, and Joaquim G. Pinto

European windstorms are among most important natural hazards for insurance companies. Quantifying with the impact of windstorms becomes more challenging due to seasonal loss clustering, characterized by numerous intense windstorms in a season, leading to exceptionally high seasonal losses. Climate change introduces another level of uncertainty regarding potential losses from European windstorm events.

The EURO-CORDEX dataset is designed to enhance the representation of regional and local weather conditions consists of a set of high-resolution climate simulations at 12.5 km resolution.  In this context, this will allow the assessment of the impact of windstorms for recent and future climates in a finer resolution. To achieve this, we use daily maximum surface wind gusts of 20 global-to-regional climate model chains from EURO-CORDEX (EUR-11 domain). The investigation focuses on the extended winter season (ONDJFM) between the historical period (1976-2005) and future projections under global warming level (GWL) scenarios of +2°C and +3°C, following the Representative Concentration Pathway 8.5 (2006-2100).

The evaluation of windstorm impact is carried out using the Loss Index (LI) method, focusing on the country level. For the historical period, a substantial bias is observed in the 98th percentile of daily maximum wind gusts between EURO-CORDEX and ERA5. This bias is corrected through empirical quantile mapping, resulting in corrected models that show reduced biases in wind gust extremes while maintaining consistency with the climate change signal.

Under the +2°C and +3°C GWLs, the majority of models indicates a reduction in the magnitude and frequency of extreme windstorms over Western Europe and the Iberian Peninsula, leading to decreased European windstorm loss, while an increase over Eastern Europe is expected, contributing to higher loss. In the majority of countries, the occurrence of seasonal loss clustering is expected to decrease under GWL conditions compared to the current climate.

Our study provides valuable insights for insurance companies and policymakers to deal with the uncertainty of the loss of windstorm under future climate conditions.

How to cite: Alifdini, I., Moemken, J., Ramos, A. M., Stankovic, A., Caballero, R., and Pinto, J. G.: Future Changes in European windstorm loss and seasonal loss clustering in the EURO-CORDEX dataset, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-7539, https://doi.org/10.5194/egusphere-egu24-7539, 2024.

EGU24-9093 | ECS | Posters on site | NP1.3

Assessing the climate drivers leading to winter wheat yield shock in Europe 

Federico Stainoh, Bianca Biess, Lukas Gudmundsson, Julia Moemken, Sonia I. Seneviratne, and Joaquim G. Pinto

Ongoing global warming has major implications on the food security and the agricultural sector. Extreme weather such as precipitation extremes, heatwaves, frost and drought can cause severe reductions in crop production. Furthermore, these events can be strongly amplifyed when considered as compound events. In this study, we inquire into the temporally combination of climate extremes leading to substantial drops in winter wheat yield (in the following called “yield shock “) throughout different countries in Europe. Winter wheat is one of the most important crops in Europe in terms of production, and is also acknowledged a major crop globally. We use the Global Data of Agricultural Yield, a satellite-reported yield hybrid dataset of major crops. We categorise the yield as “yield shock” and “no yield shock” in order to reduce the uncertainty and to mainly focus on historical yield plunges. We consider and test different climate indicators (like the number of warm days or cumulative precipitation) that represent weather extreme events at a subseasonal scale. Moreover, we employ a Random Forest to capture any possible nonlinear relation. Our study illustrates the probability of winter wheat yield shock under the occurrence and co-occurrence of subseasonal weather extremes, and the nuances throughout countries with different climatic patterns.

How to cite: Stainoh, F., Biess, B., Gudmundsson, L., Moemken, J., Seneviratne, S. I., and Pinto, J. G.: Assessing the climate drivers leading to winter wheat yield shock in Europe, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-9093, https://doi.org/10.5194/egusphere-egu24-9093, 2024.

EGU24-10153 | ECS | Orals | NP1.3

A Spatio-Temporal Optimization- Based Feature Selection Framework for detecting drivers of heatwaves 

César Peláez-Rodríguez, Jorge Pérez-Aracil, Ronan McAdam, Antonello Squintu, Enrico Scoccimarro, and Sancho Salcedo-Sanz

Machine Learning (ML) encompasses various techniques and algorithms that have proven highly effective in addressing complex climate science tasks. In particular, using ML to detect and forecast extreme events has gained much attention recently. Considering the vast volume of spatial and temporal data available, the employment of data-driven methodologies becomes indispensable for effectively uncovering potential drivers of these events. This study arises with the ambition of proposing a comprehensive and general framework that provides insights about interactions between heatwaves and potential physical drivers across multiple spatio-temporal scales. A novel feature-selection methodology is presented to advance the detection of heatwaves. It is based on a two-step procedure. In the first stage, a non-supervised task is developed for spatially clustering the different variables. Thus, there is a reduced initial pool of driver candidates. In the second stage, a wrapper methodology is applied to determine which time periods are representative for each of the clusters in the occurrence of heatwaves. This algorithm discerns which clusters are selected as drivers and which are discarded, along with the time period, relative to the heatwave occurrence, during which each cluster should be investigated. Thus, the feature selection is developed based on the spatio-temporal distribution of the different variable clusters.

Experiments have been developed for detecting heatwaves in the Lake Como region, a region of key agricultural activity in Northern Italy. A  wide range of ocean and atmospheric variables taken from ERA5 are used (e.g. sea ice concentration, precipitation). The framework allows the identification of the time lag for each variable from short-term to seasonal time scales (up to 180 days). Results have spotted drivers on the subseasonal to seasonal timescale. Important drivers have been identified for short-term periods (less than one week). Local variables are shown to be of much significance for these periods. The method allows for identifying the strong influence of local variables whilst identifying correlations between the heatwave occurrence and different variables and spatial locations. The importance stages of the variable candidates can be established by running the model while removing the most critical variables.

How to cite: Peláez-Rodríguez, C., Pérez-Aracil, J., McAdam, R., Squintu, A., Scoccimarro, E., and Salcedo-Sanz, S.: A Spatio-Temporal Optimization- Based Feature Selection Framework for detecting drivers of heatwaves, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-10153, https://doi.org/10.5194/egusphere-egu24-10153, 2024.

EGU24-10586 | ECS | Orals | NP1.3

Study of extreme heatwave seasons in South Asia using rare event simulations 

Clément Le Priol, Joy M. Monteiro, and Freddy Bouchet

Extreme climate events have major impacts on human societies and ecosystems. The most detrimental events are often extremely rare, with return time of centuries or even millennia. Studying these events in the context of climate change is crucial to help adaptation efforts globally. Yet the study of these extremely rare events is extremely challenging due to the lack of data. Indeed, such events have likely not been observed in the instrumental period. An alternative is to use a global climate model to simulate these extremely rare events. However, this comes at a huge computational cost : gathering good statistics on centennial events would require to run a few thousand years of simulation. 

Rare event algorithms have recently been introduced in the field of climate science to tackle this difficulty [1]. By concentrating the computational effort on the trajectories most susceptible to lead to the extreme event of interest, they allow for the sampling of extremely rare events at a much lower computational cost than standard simulations. 

In this study, we run a rare event algorithm to sample extreme heatwave seasons in a heatwave hotspot of South Asia, using the intermediate complexity model Plasim. We compare the outcome of the algorithm against an extremely long – 8000 years – control run. This comparison allows us to demonstrate that the algorithm not only estimates return times with high precision (as shown in previous work), but also exhibits high precision in the estimation of composite statistics: composite maps conditioned on centennial heatwave seasons estimated from the algorithm, are in very good agreement with the ones from the 8000-year long control simulation. Our results suggest that extreme heatwave seasons in the studied region are associated with a quasi-stationary atmospheric wave-pattern stretching from the North Atlantic towards South Asia. 

We also show that the algorithm correctly estimates the intensity-duration-frequency statistics of subseasonal heatwaves occuring within centennial heatwave seasons. Thus rare event algorithms could, for instance, be combined with seasonal forecasts to provide information regarding expected number of heatwave days and the distribution of the duration and intensity in an extreme heatwave season, which could be useful for adaptation planning. 

 

References

[1] F. Ragone, J. Wouters, and F. Bouchet, “Computation of extreme heat waves in climate models using a large deviation algorithm,” Proc Natl Acad Sci USA, vol. 115, pp. 24–29, Jan. 2018. 

How to cite: Le Priol, C., Monteiro, J. M., and Bouchet, F.: Study of extreme heatwave seasons in South Asia using rare event simulations, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-10586, https://doi.org/10.5194/egusphere-egu24-10586, 2024.

EGU24-11046 | ECS | Posters on site | NP1.3

Maximal reachable temperatures for Western Europe in current climate 

Robin Noyelle, Yi Zhang, Pascal Yiou, and Davide Faranda

Human bodies, ecosystems and infrastructures display a non-linear sensibility to extreme temperatures occurring during heatwave events. Preparing for such events entails to know how high surface air temperatures can go. Here we examine the maximal reachable temperatures in Western Europe. Taking the July 2019 record-breaking heatwave as a case study and employing a flow analogues methodology, we find that temperatures exceeding 50 C cannot be ruled out in most urban areas, even under current climate conditions. We analyze changes in the upper bound of surface air temperatures between the past (1940–1980) and present (1981–2021) periods. Our results show that the significant increase in daily maximum temperatures in the present period is only partially explained by the increase of the upper bound. Our results suggest that most of the warming of daily maximum surface temperatures result from strengthened diabatic surface fluxes rather than free troposphere warming.

How to cite: Noyelle, R., Zhang, Y., Yiou, P., and Faranda, D.: Maximal reachable temperatures for Western Europe in current climate, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-11046, https://doi.org/10.5194/egusphere-egu24-11046, 2024.

EGU24-11183 | ECS | Posters on site | NP1.3

Impact of horizontal resolution and model time step on European precipitation extremes in the OpenIFS atmosphere model 

Yingxue Liu, Joakim Kjellsson, Abhishek Savita, Joke Lübbecke, and Wonsun Park

Events of extreme precipitation pose a hazard to many parts of Europe but are typically not well represented in climate models. Here, we evaluate daily extreme precipitation over Europe during 1982-2019 in observations (GPCC), reanalysis (ERA-5) and a set of atmosphere-only simulations at low- (100km), medium- (50km) and high- (25km) resolution with OpenIFS (Version43R3). We find that both model simulations and reanalysis underestimate the rates of extreme precipitation compared to observations. The biases are largest for the lowest resolution (100 km) and decrease with increasing horizontal resolution (50 and 25 km) in all seasons. The sensitivity to horizontal resolution is particularly high in mountain regions, likely linked to the sensitivity of vertical velocity to the representation of topography. The sensitivity of precipitation extremes to model resolution increases dramatically with increasing percentiles, which modest biases at the 70th percentile and large biases at 99th percentile.  We also find that extreme precipitation mostly consists of large-scale precipitation (~80%) in winter, while in summer it is mostly large-scale precipitation in Northern Europe (~70%) and convective precipitation in Southern Europe (~70%). Compared to ERA5, the model simulations produce higher large-scale precipitation extremes in winter, but weaker in summer. The discrepancy between OpenIFS simulations and ERA-5 decreases with increasing horizontal resolutions. We also examine the model time step’s effect on extreme precipitation. The results show that the convective contribution to extreme precipitation is more sensitive to the model time step than horizontal resolution. This is likely due to the sensitivity of convective activity to model time step. On the other hand, the large-scale contribution to extreme precipitation is more sensitive to horizontal resolution than model time step, which may be due to sharper fronts and steeper topography at higher resolution. In general, the lowest-resolution and longest time step has overall higher biases than the highest-resolution and shortest time step.

How to cite: Liu, Y., Kjellsson, J., Savita, A., Lübbecke, J., and Park, W.: Impact of horizontal resolution and model time step on European precipitation extremes in the OpenIFS atmosphere model, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-11183, https://doi.org/10.5194/egusphere-egu24-11183, 2024.

EGU24-11807 | ECS | Orals | NP1.3

Drivers of heatwaves in CMIP6 models: an evaluation on historical simulations 

Antonello A. Squintu, Ronan McAdam, César Peláez Rodríguez, Jorge Pérez Aracil, Carmen Alvarez Castro, and Enrico Scoccimarro

Heatwaves heavily affect European public health, society and economy. A full understanding of the drivers behind the occurrence and intensity of heatwaves (HW) is one of the priorities of H2020 CLimate INTelligence (CLINT) project. Particular attention is given to the detection and attribution of HWs in future climate projections. However, it is important to assess the capability of climate models to thoroughly describe relationships between the drivers and the occurrence and intensity of HWs. For this reason, a feature selection framework, based on the Coral Reef Optimization (Salcedo-Sanz et al., 2014) has been developed. This has been applied to ERA5 summer data, using as a target the Lake Como HW occurrence and, as candidate predictors, time series of weather variables calculated on clustered areas on European and global scales. The same algorithm has been applied to historical climate simulations included in CMIP6. The comparison of the results of these two steps has first focused on the similarities in maximum temperature and HW trends in the target region of Lake Como. Then, the selected drivers in each historical climate simulation have been evaluated, using ERA5 results as benchmark. Thanks to this, the models that better resemble the statistical properties and teleconnections described by the reanalysis have been identified. This set of models will be considered for the detection and attribution of heatwaves in future climate projections under different emission scenarios. In this upcoming phase the goal will be to analyse changes in the relationship between the drivers and HW occurrence and intensity, giving an insight about possible future evolutions in heatwaves frequency and magnitude.

How to cite: Squintu, A. A., McAdam, R., Peláez Rodríguez, C., Pérez Aracil, J., Alvarez Castro, C., and Scoccimarro, E.: Drivers of heatwaves in CMIP6 models: an evaluation on historical simulations, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-11807, https://doi.org/10.5194/egusphere-egu24-11807, 2024.

EGU24-11997 | ECS | Orals | NP1.3

Extreme events in a templex 

Gisela Daniela Charó, Davide Faranda, Michael Ghil, and Denisse Sciamarella

Theoretical and numerical studies have shown that transient atmospheric motions leading to weather extremes can be classified through the instantaneous dimension and stability of a state of a dynamical system [Faranda et al., Sci. Rep., 2017]. The asymptotic values of these quantities can be computed theoretically only for specific systems, while their numerical counterpart for climate observables provides information on the rarity, predictability, and persistence of specific states. In this work, we present a first attempt to relate the presence of extreme events with the elements that make up a templex of the system under study, both in the deterministic [Charó et al., Chaos, 2022] and stochastic frameworks [Charó et al., Chaos, 2023]. The templex provides the key characteristics of the topological structure underlying a dynamical system. This work will present results for the classical, deterministic Lorenz [JAS, 1963] attractor and for the Lorenz Random Attractor, dubbed LORA [Ghil & Sciamarella, NPG, 2023].

How to cite: Charó, G. D., Faranda, D., Ghil, M., and Sciamarella, D.: Extreme events in a templex, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-11997, https://doi.org/10.5194/egusphere-egu24-11997, 2024.

Windstorms are extreme midlatitude cyclones and one of the major natural hazards that cause damage and losses in Europe. While the processes involved in their genesis and intensification are generally well understood, there are still considerable uncertainties in the estimation of associated impacts like widespread wind damage and flooding. The compounding characteristics of the events further enhances the complexity of this task. This is even more true for the impact forecasting of windstorms on weather and sub-seasonal time scales. Additionally, there are large uncertainties on how windstorms and their impacts will change in a warmer climate, particularly regarding the role played diabatic processes in a warmer atmosphere. This study presents examples of recent developments regarding windstorms and discusses some new avenues for interdisciplinary research towards bridging the gap between fundamental research and practical applications.

How to cite: Pinto, J. G.: European Windstorms: bridging the gap between fundamental research and practical applications, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-12471, https://doi.org/10.5194/egusphere-egu24-12471, 2024.

The presence of anomalously cold air close to the surface is a prerequisite for the occurrence of cold weather extremes, such as cold spells. However, the process of cold air generation features a substantial variability in space and time, modulated by the rate of energy loss to space by infrared radiation.

Such a variability is investigated using a Lagrangian approach, identifying trajectories that experience rapid non-adiabatic cooling over Eurasia. This approach allows to identify source regions of cold air, and the meteorological conditions that particularly favor its generation and accumulation – which often precedes the most extreme cold spells.

The unraveling of this connection allows to interpret the intra-seasonal and the inter-annual variability in the occurrence of cold extremes, and to gain a mechanistic understanding of how anthropogenic global warming will modify them.

How to cite: Riboldi, J.: Quantifying the generation of cold air during boreal winter and its relevance for cold weather extremes, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-13211, https://doi.org/10.5194/egusphere-egu24-13211, 2024.

EGU24-13496 | ECS | Orals | NP1.3

Persistent winter merged jets over the Atlantic and extreme weather anomalies 

Sohan Suresan, Nili Harnik, and Rodrigo Caballero

Variabilities in the jet streams have a significant influence on our weather and climate, and could potentially increase the likelihood of a range of extreme weather events. The winter of 2009/2010 witnessed an unusual equatorward displacement of the Atlantic jet and its subsequent convergence with the African jet, leading to the emergence of a persistent zonally oriented merged jet. At the same time, intense and prolonged negative phase of the North Atlantic Oscillation (NAO) and unusually cold and extreme weather conditions were reported over the Northern Hemisphere. Such a merging was only observed to occur for a whole winter during the winters of 1968-69 and 1969-70. Preliminary results indicate that such persistent winter merged jets could be more frequent in a future global warming scenario and thus it is important to understand this dynamical regime transition of Atlantic jet and its effects on the weather patterns. In this study, we explore the extreme weather distribution over the northern hemisphere during such winter merged jets and its relation to NAO and ENSO. We show that merged jet winter months have a signature weather pattern distribution that is different from the negative NAO phase. We see a decrease in the surface eddy kinetic energy over the midlatitude during such winter leading to an equatorward shift in storm tracks over the Atlantic region and larger stormtrack density over the western Greenland which could potentially lead to the observed distribution of weather patterns. On comparing the surface temperature anomaly composites between the winters of strong negative NAO, EL Nino, and merged jet months we see that the merged jet winters have a significant persistent temperature distribution signature over the tropics and the Arctics. Similar analysis over the north hemisphere for surface wind, precipitation, and snowfall anomalies also shows a preferred persistent distribution over certain regions during the merged jet-state winters.

How to cite: Suresan, S., Harnik, N., and Caballero, R.: Persistent winter merged jets over the Atlantic and extreme weather anomalies, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-13496, https://doi.org/10.5194/egusphere-egu24-13496, 2024.

EGU24-13645 | ECS | Orals | NP1.3

Underestimating extremes due to random uncertainty 

Nithin Sivadas and David Sibeck

As we attempt to infer a system's response to external driving from measurements, random errors in the measurement of the drivers can lead us to mistakenly infer a non-linear response. In particular, we are likely to underestimate the system's response during extreme and rare driving conditions due to uncertainty in the drivers. We demonstrate this phenomenon for extreme space weather and its impact on Earth's magnetosphere, where due to random errors in the measurements of solar wind drivers, there is a non-linear bias in the magnetosphere's response. We propose that the underlying statistical effect (related to the more well-known regression to the mean effect) is generalizable to the other fields that study different systems' responses to driving, like extreme climate studies. 

How to cite: Sivadas, N. and Sibeck, D.: Underestimating extremes due to random uncertainty, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-13645, https://doi.org/10.5194/egusphere-egu24-13645, 2024.

During the last years, the statistical analysis of compound extremes has gained increasing interest among the scientific community due to the multiple threats presented by such events to society, economy, and the environment. In many situations, the statistics of compound extremes is based on bivariate extreme value theory and measures provided by this framework. Such choice of statistical methodology may however not properly address two relevant aspects: the non-zero duration of such events (which can be rather persistent, e.g. in the case of droughts or heatwaves, which heavily violates the independence assumption of classical extreme value theory) and the fact that not all events of practical revelance can be described as cases falling into the tails of the distribution of some observable of interest.

A general framework addressing the non-extremeness aspect is provided by event coincidence analysis (ECA), which quantifies the empirical frequency of co-occurring events of arbitrary types and allows ist comparison with the values for certain random null models like independent Poisson processes with prescribed event rates. While classical ECA is based on temporal point processes and hence may be criticized for not capturing the statistical characteristics of persistent events very well, I will present a new methodological variant, interval coverage analysis (InCA), as an alternative for specifically addressing co-occurrences of persistent events. In the limit of vanishing event durations, the new interval coverage rates of InCA are identical to the event coincidence rates provided by ECA. By allowing for mutual time shifts between the different types of events under study as well as a temporal tolerance regarding their respective timing, fixed and even distributed time lags can be taken into consideration.

This presentation introduces and compares the basic methodological concepts behind both ECA and InCA (including their extension to studying multivariate and conditional dependency), and demonstrates an example of their respective application in geoscientific contexts. Specifically, the spatial patterns of dependency between the timing of heatwaves and large-scale circulation anomalies of the atmospheric jet stream in the Northern hemisphere is studied. The corresponding analysis reveals specific regions with elevated likelihood of heatwaves along with the emergence of a split (double-banded) jet stream, while the emergence of heatwaves is suppressed at the same time in other regions. The obtained results may thus guide further targeted research regarding the specific mechanisms leading to this regional differentiation in heatwave frequency.

How to cite: Donner, R. V.: Studying statistical dependency among short and persistent events – recent developments and application to mid-latitude circulation anomalies associated with heatwaves, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-14476, https://doi.org/10.5194/egusphere-egu24-14476, 2024.

Over the last decade, researchers have devoted considerable efforts to exploring the feasibility of attaining convection-permitting scales in regional climate model projections. The primary objective has been to comprehend the effects of global warming on climate extremes. Initially conducted as isolated model experiments, these investigations have evolved into coordinated ensemble experiments operating at convection-permitting scales across diverse continents. The implementation of such coordinated ensembles has provided a platform for evaluating model reliability at high resolutions and conducting signal-to-noise analyses on identified climate change signals.

 

Nevertheless, the constraints imposed by limited computational resources have confined these experiments to smaller domains compared to the conventional continental scale employed in dynamical downscaling, as seen in initiatives like the CORDEX community and time slice mode.

 

Despite these inherent limitations, these experiments have successfully showcased the models' capacity to simulate present-day climate conditions. Notably, improvements in various statistical metrics at sub-daily scales have been observed in contrast to parametrized models. Furthermore, the ensemble approach has contributed to reducing uncertainties in assessing both present-day climate and future projections, particularly in terms of frequency, intensity, and extreme precipitation at the hourly time scale.

 

The explicit representation of convection has additionally enabled the study of convective storm system evolution, allowing for an assessment of large-scale feature changes and related physical mechanism driving observed extreme precipitation variations.

 

Preliminary attempts have also been carried on for building convection permitting climate emulator to reduce the computational HPC demand required by the dynamical downscaling models.

These collective findings underscore the necessity of advancing to the next phase of coordination, involving the establishment of multiple coordinated platforms spanning different continents. These platforms will serve as collaborative spaces for discussing model enhancements, aiming to refine existing models by incorporating a more precise representation of Earth system components and defining domains that maximize the number of models capable of generating convection-permitting climate projections.

How to cite: Coppola, E.: Understanding global warming impact on climate extremes by mean of Coordinated Ensemble Experiments of Convection-Permitting Climate Projections, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-16130, https://doi.org/10.5194/egusphere-egu24-16130, 2024.

EGU24-16462 | ECS | Posters on site | NP1.3

Day- and nighttime heatwave clusters over Europe and their physical drivers 

Felicitas Hansen, Frauke Feser, and Eduardo Zorita

Heatwaves rank among the most devastating extreme events over Europe, particularly in terms of mortality and agricultural damage. Heatwaves are often defined as exceedances of thresholds based on daily maximum temperatures, thus considering only daytime effects. However, exceedances of daily minimum temperatures, which often occur at night, are of similar importance, as they can cause additional stress to the human health due to shorter recovery times. For both daytime and nighttime heatwaves, knowledge of the underlying mechanisms is crucial for the successful prediction of the events; however, these mechanisms are not yet fully understood.

Although heatwaves can occur quite locally, the task of heatwave prediction can be simplified by representing heatwaves as recurring large-scale patterns. The aim of the presented work is to identify these dominant coherent spatial patterns over Europe using the SANDRA (Simulated Annealing and Diversified Randomization) clustering method for both daytime and nighttime heatwaves. In a second step, relevant atmospheric, oceanic and land variables are investigated for their physical connection to each of the European heatwave clusters with different time lags.

The results obtained from reanalysis data covering the recent past are compared with those obtained from a long coupled global climate model simulation of the last 2000 years, performed with the MPI-ESM model. The long model simulation is further used to place the recent record-breaking European heatwave of summer 2022 in a historical context.

How to cite: Hansen, F., Feser, F., and Zorita, E.: Day- and nighttime heatwave clusters over Europe and their physical drivers, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-16462, https://doi.org/10.5194/egusphere-egu24-16462, 2024.

EGU24-16848 | Orals | NP1.3

Summer Deep Depressions Increase Over the North Atlantic 

Fabio D'Andrea

Mid-tropospheric deep depressions in summer over the North Atlantic are shown to have strongly increased in the eastern and strongly decreased in the western North Atlantic region.  This evolution is linked to a change in baroclinicity in the west of the North Atlantic ocean and over the North American coast, likely due to the increased surface temperature there. Deep depressions in the Eastern North Atlantic are linked to a temperature pattern typical of extreme heat events in the region. The same analysis is applied to a sample of CMIP6 model outputs, and no such trends are found.  This study suggests a link between the observed increase of summer extreme heat events in the region and the increase of the number of Atlantic depressions. The failure of CMIP6 models to reproduce these events can consequently also reside in an incorrect reproduction of this specific feature of midlatitude atmospheric dynamics.

How to cite: D'Andrea, F.: Summer Deep Depressions Increase Over the North Atlantic, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-16848, https://doi.org/10.5194/egusphere-egu24-16848, 2024.

EGU24-17701 | ECS | Orals | NP1.3

Climate Indices of Extreme Rainfall in the South Asian Monsoon Domain using a Bayesian Hierarchical Model 

Dexter Früh, Felix Strnad, and Bedartha Goswami

Intraseasonal variability of extreme rainfall events (EREs) during the South Asian Summer Monsoon season is modulated by the Boreal Summer Intraseasonal Oscillation (BSISO), a convective system of organised heavy rainfall that moves periodically from the Indian Ocean to the Western Pacific over the subcontinent. The BSISO, in turn, is typically characterised using indices obtained from the leading components of an Empirical Orthogonal Analysis (EOF) of outgoing longwave radiation (OLR) and lower and upper troposhere wind data from the region [1] ). A primary motivation for using OLR and wind data is that the EOF-analysis is not well suited for heavy-tailed data such as precipitation. 

Here, we propose to estimate climate indices directly from EREs in the South Asian Summer Monsoon by applying the Hidden Climate Index (HCI) - framework introduced by Renard et al. (2022) [2]. The method is designed to work with binary, event-like data and utilizes a Bayesian hierarchical model incorporating spatial Gaussian process priors, to capture spatial and temporal interdependencies by sampling from a Bernoulli distribution.

Using the HCI framework, we estimate latent variables that underlie the ERE dynamics in the South Asian Monsoon domain, and show that these are related to large-scale modes of climate variability., We demonstrate that the ERE-based HCIs correlated well to the BSISO and in addition, we find relationships between the observed large-scale spatial ERE patterns  to the El Nino Southern Oscillation and the Silk Road Pattern.

 

[1] Kikuchi, K., Wang, B. & Kajikawa, Y. Bimodal (2012). Representation of the tropical intraseasonal oscillation. Clim. Dyn. 38, 1989–2000.

[2] Renard, B., Thyer, M., McInerney, D., Kavetski, D., Leonard, M., & Westra, S. (2022). A Hidden Climate Indices Modeling Framework for Multivariable Space-Time Data. Water Resources Research, 58

How to cite: Früh, D., Strnad, F., and Goswami, B.: Climate Indices of Extreme Rainfall in the South Asian Monsoon Domain using a Bayesian Hierarchical Model, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-17701, https://doi.org/10.5194/egusphere-egu24-17701, 2024.

EGU24-18308 | ECS | Posters on site | NP1.3

Dynamics of pan-Atlantic winter compound extremes in ERA5 and CMIP6 models 

Richard Leeding, Iana Strigunova, and Gabriele Messori

Recent work has provided robust evidence for the systematic co-occurrence of wintertime cold spells in North America and wet and windy extremes in Europe, which we term compound pan-Atlantic extremes. Both cold spells and wet and windy extremes are individually highly impactful, and their concurrence further amplifies their effects for actors with international exposure who are vulnerable to correlated losses. This study aims to investigate further the atmospheric processes associated with compound pan-Atlantic cold, wet and windy extremes and how these processes are represented in CMIP6 models.

On aggregate, cold spells in different parts of North America statistically co-occur with wind and precipitation extremes in specific European regions. However, North American cold spells can arise from multiple dynamical pathways, altering the location and timing of the associated European extremes for individual cold spells. Here, we use ERA5 reanalysis data (1940-2014) to identify North American wintertime cold spells in three different regions and relate the occurrence of European extremes to Pacific and Atlantic weather regimes. We further analyze the various pathways of North American cold spells by evaluating the relative contribution of planetary (k=1-3) and synoptic (k=4-8) Rossby waves to the resultant weather regimes. The evaluation is performed by partitioning the Rossby wave circulation into different zonal wavenumber ranges using the MODES software, based on the normal-mode function decomposition. This methodology has previously been employed to identify changes in the midlatitude circulation at multiple scales during Eurasian heatwaves, though it is novel in its application to cold spells. Here, we discuss how the wavenumber ranges differ across cold spells and from the climatological state before and during the cold spells. We next compare the CMIP6 historical simulation model data (1940-2014) with the ERA5 results. First, we review the ability of the models to replicate the spatial and temporal pattern of pan-Atlantic extremes for the three cold spell regions. Second, we discuss the performance of the models in capturing the weather regime frequencies and the planetary and synoptic Rossby wave contributions to the North American cold spell pathways.
The results of this study contribute to the evaluation of the model fidelity in reproducing pan-Atlantic compound extremes and the associated circulation, with direct implications for the assessment of climate projections.

How to cite: Leeding, R., Strigunova, I., and Messori, G.: Dynamics of pan-Atlantic winter compound extremes in ERA5 and CMIP6 models, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-18308, https://doi.org/10.5194/egusphere-egu24-18308, 2024.

EGU24-18316 | ECS | Posters on site | NP1.3

Investigating the Influence of Atmospheric Blocking Morphology on Predictability 

Anupama K Xavier, Oisín Hamilton, Davide Faranda, and Stéphane Vannitsem

Low-frequency variability (LFV) encompasses atmospheric and climate processes on time scales from a few weeks to decades.​ This includes atmospheric blockings, heat waves, cold spells, and at longer time scales long-term oscillations like the MJO, the NAO, ENSO….. Better understanding of LFV, could contribute to improved long term forecasts​. 

In the results described in Xavier et al, 2023, who used a reduced order atmosphere-land model (Demaeyer et al, 2020), weather patterns that involve atmospheric blocking to the west of a given topographical feature tend to have reduced predictability and show instability when contrasted with blocking occurrences situated to the east of such topographical elements. This finding aligns with actual meteorological occurrences, such as the persistence of North Pacific blocking patterns (Breeden et al., 2020; Kim and Kim, 2019). The shape and characteristics of the identified blocking events closely resemble North Pacific blocks, where a high-pressure system exists either on the western or eastern side of the underlying topography. In the physical world, these positions correspond to Asian and American continents on either side of the Pacific. Despite quasi-geostrophic models being overly simplified, using such reduced order models in this study allowed us to undertake such mathematical analysis. Thus allowing for a comparison with the real world…

In the current study, we aim to analyze these predictability differences based on the morphology of the blocking situations in the real-world scenario using the CMIP6 dataset. Blocking situations are identified using two different indices, the anomaly-based blocking index proposed by Sausen et al., 1995 and the local reversal of the meridional flow-based index proposed by Davini et al., 2012. Predictability is quantified using local dimension metrics and analogue studies in the identified blocking events. The findings are discussed from the perspective of the current literature on the predictability of blocking.

References

Xavier, A. K., Demaeyer, J., and Vannitsem, S.: Variability and Predictability of a reduced-order land-atmosphere coupled model, EGUsphere [preprint], https://doi.org/10.5194/egusphere-2023-2257, 2023.

Demaeyer, Jonathan & De Cruz, Lesley & Vannitsem, S.: qgs: A flexible Python framework of reduced-order multiscale climate models. Journal of Open Source Software. 5. 2597. 10.21105/joss.02597, 2020.

Breeden, M. L., Hoover, B. T., Newman, M., and Vimont, D. J.: Optimal North Pacific blocking precursors and their deterministic subseasonal evolution during boreal winter, Monthly Weather Review, 148, 739–761, 2020

Kim, S.-H. and Kim, B.-M.: In search of winter blocking in the western North Pacific Ocean, Geophysical Research Letters, 46, 9271–9280, 2019.

Sausen, R., König, W., & Sielmann, F. (1995). Analysis of blocking events from observations and ECHAM model simulations. Tellus A, 47(4), 421–438.

Davini, P., Cagnazzo, C., Gualdi, S., & Navarra, A. (2012). Bidimensional diagnostics, variability, and trends of Northern Hemisphere blocking. Journal of Climate, 25(19), 6496–6509.

How to cite: K Xavier, A., Hamilton, O., Faranda, D., and Vannitsem, S.: Investigating the Influence of Atmospheric Blocking Morphology on Predictability, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-18316, https://doi.org/10.5194/egusphere-egu24-18316, 2024.

The unprecedented changes in Earth's climate are reshaping atmospheric dynamics on a global scale, with profound implications for various sectors, including aviation. One of the critical facets of this transformation lies in the alterations to atmospheric circulation patterns and the concurrent modifications in turbulence characteristics.

We investigate the intricate interplay between climate change, atmospheric circulation patterns, and turbulence modifications over Europe, with a specific focus on their implications for commercial flight operations. Drawing upon ERA5 reanalysis data at 200 hPa pressure level we explore the evolving climatic conditions shaping the European airspace. By jointly looking at the cube-rooted eddy dissipation rate, vorticity, and horizontal divergence we evidence increasing trends of anomalously turbulent conditions over UK and Northern Europe. Otherwise, decreasingly frequency anomalies are associated with light-turbulent conditions over Central and Mediterranean Europe. Overall, increasing trends also correspond to increased severity in terms of turbulence strength levels, with a clear increase in moderate-to-severe turbulence episodes, mainly associated with both converging and diverging atmospheric ridges.

The study highlights the region's vulnerability to significant changes in atmospheric pattern dynamics, emphasizing the potential increase in turbulence-related episodes and severity, impacting aviation safety, fuel efficiency, and passenger comfort. Our analysis aims to provide valuable insights for aviation stakeholders, policymakers, and researchers, contributing to the development of adaptive strategies and operational guidelines.

How to cite: Alberti, T.: Atmospheric circulation changes and turbulence modifications over Europe: implications for commercial flight operations in a changing climate, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-18608, https://doi.org/10.5194/egusphere-egu24-18608, 2024.

EGU24-18695 | Posters on site | NP1.3

Predictability of the low pressure systems leading concurrent events in the Iberian Peninsula. 

Carmen Alvarez-Castro, Davide Faranda, David Gallego, Cristina Peña-Ortiz, Veronica Torralba, and Silvio Gualdi

Filomena was an extratropical cyclone in early January 2021 that was most notable for bringing unusually heavy snowfall to parts of Spain, with Madrid recording its heaviest snowfall in over a century, and with Portugal being hit less severely. Filomena caused severe winds, heavy rainfall and snowfall in different parts of the country but also occurring at the same time. The atmospheric pattern during the event of Filomena was a rare one characterized by a low intrinsic predictability. Despite the number of studies focusing on the detection and characterization of extreme events in Western Europe, our knowledge regarding the predictability of such occurrences remains limited. By studying the intrinsic predictability of an extreme event, we can know the capacity we have to anticipate it, thus being able to take action for the minimization of its impacts by using early warning systems.

In this work, we show first an overview of the intense low pressure systems (including Filomena) that have recently struck specific cities/regions of the Iberian Peninsula with concurrent events. We assess their predictability and, finally, we perform an attribution study of these events to climate change.

How to cite: Alvarez-Castro, C., Faranda, D., Gallego, D., Peña-Ortiz, C., Torralba, V., and Gualdi, S.: Predictability of the low pressure systems leading concurrent events in the Iberian Peninsula., EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-18695, https://doi.org/10.5194/egusphere-egu24-18695, 2024.

EGU24-18866 | ECS | Orals | NP1.3

A Gaussian framework for optimal prediction of extreme heat waves 

Valeria Mascolo, Alessandro Lovo, Corentin Herbert, and Freddy Bouchet

Heat waves are a growing issue in the current climate, causing damage to human societies and other living beings. As climate warms, heat waves are one of the extreme events that will be exacerbated by the rising average temperatures. Understanding the mechanisms that drive heat waves is hence of vital importance to analyze them and to make predictions to mitigate their impacts. However, the rarity of any extreme event makes it particularly hard to study, especially if one wishes to understand the relation between predictors and probability, for instance using machine learning techniques, which are notoriously data-hungry. A possible solution is to use climate model simulations, but they introduce biases and can be prohibitively expensive to run for appropriate dataset lengths.

In this work, we introduce a simple and powerful statistical model, that is able to skillfully predict rare heat waves in a regime of lack of data, as it is the case for the ERA5 reanalysis dataset.

We focus on two-week heat waves over France, and we notice that composite maps of very extreme events are similar to the ones of much less extreme ones. This holds true for an increasing hierarchy of model complexity, including ERA5. We can thus analyze very extreme events by looking at less rare ones, having the advantage of increasing the available statistics. This effect can be explained assuming that the set of predictors and the heat wave amplitude are jointly gaussian. The prediction task can be thus rephrased into estimating the conditional probability of an extreme heat wave happening conditioned on the state of the set of predictors. This quantity, called committor function, is generally hard to estimate given the high dimensionality of the variables involved. Our assumption performs a dimensionality reduction, where an estimate is obtained through an optimal linear projection. The projection map will suggest the most relevant predictors.

We first demonstrate that this Gaussian approximation performs well on a 8000 years run of a model of intermediate complexity, PlaSim, for both analysis and prediction tasks when 500 hPa geopotential height, temperature and soil moisture are used as input fields. When we compare with a machine learning estimation of the committor, which is possible here thanks to the huge amount of available data, our Gaussian model is less skillful, but not by much. We then apply the Gaussian assumption to the 500hPa geopotential height field of the ERA5 dataset, outperforming the neural networks and extending the predictability horizon by several days.

The prediction method we propose is simple and effective. It opens possibilities to achieve skilfull predictions that outperform all known approaches, for instance machine learning, notably when data is scarce. Our method also serves as a better baseline than the climatology to benchmark more complex approaches. Since our statistical model is also interpretable, this framework has potential to go beyond prediction skill only and, thanks to the optimal projection map, foster the study of fast and slow drivers, and the effect climate change has on them.

How to cite: Mascolo, V., Lovo, A., Herbert, C., and Bouchet, F.: A Gaussian framework for optimal prediction of extreme heat waves, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-18866, https://doi.org/10.5194/egusphere-egu24-18866, 2024.

EGU24-20380 | ECS | Posters on site | NP1.3

Temperature and wind statistics with convection-permitting model WRF in the context of heat waves and urban heat island events 

Luminita Danaila, Clement Blervacq, Kazim Sayeed, Manuel Fossa, Nicolas Massei, and Kwok Chun

Recent studies have shown two-way interactions between aloft large-scale structures in the atmosphere
and local features such as surface temperature, wind, and land use. This requires the use of high-resolution land use schemes and convection-permitting models (CPM) for large eddy simulations (LES).  Weather
Research and Forecasting model (WRF) is being increasingly used with resolutions that allow convection
to be fully simulated, and efforts (such as CORDEX FPS URB RC) have been made to introduce more
precise land use schemes to model the impacts of urban zones on temperature and wind statistics. In
this study, we focus on the 2003 summer heat wave and compute temperature and wind statistics from
surface to upper tropospheric pressure levels, ranging from microscales (~50m) to mesoscales (~500km) in
North-western France. The emphasis is put on extreme values of temperature by computing its Probability
Density Function (PDF) over the domain and across different spatial scales. Results pertain to second,
third, and fourth-order moments of temperature and wind reflecting variance, direction of across-scale
interactions, and extreme events' occurrence probability, respectively. Finally, we correlate mean large-scale
temperature gradients with those extreme events. This study provides new insights into the complex and
continuous across scales two-way interactions between local features and large-scale climate.

How to cite: Danaila, L., Blervacq, C., Sayeed, K., Fossa, M., Massei, N., and Chun, K.: Temperature and wind statistics with convection-permitting model WRF in the context of heat waves and urban heat island events, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-20380, https://doi.org/10.5194/egusphere-egu24-20380, 2024.

EGU24-1036 | ECS | Orals | NP1.5

Exploring Noise-induced and CO2-driven AMOC collapses in the PlaSIM-LSG climate model with a Rare Event Algorithm. 

Matteo Cini, Giuseppe Zappa, Francesco Ragone, and Susanna Corti

Earth-system Models of Intermediate Complexity (EMICs) are climate models featuring a simplified representation of climate processes and a much lower computation cost. This makes them particularly suitable for exploring phenomena with a large ensemble simulation approach. Here we use the coupled atmosphere-ocean PlaSIM-LSG EMIC to study the possibility of Atlantic Meridional Overturning Circulation (AMOC) spontaneous collapses and how this is altered in the presence of external anthropogenic forcing. Understanding the stability of the AMOC and its response to anthropogenic forcing is of key importance for advancing climate science. The idea of a “safe-operating space” has been proposed in order to define a threshold on anthropogenic forcing within which the AMOC does not lose stability. This requires understanding the combined action of CO2-driven and noise-induced processes in climate tipping events

 First, we address the occurrence of noise-induced AMOC collapses, i.e. spontaneous abrupt weakening  events induced by chaotic internal climate variability in absence of any external forcing. We address the problem of finding these extreme events via the application of a Rare Event Algorithm, which - via a selective cloning of the most interesting model trajectories -  allows a faster exploration of the model phase space in the direction of an AMOC decrease. The algorithm is applied to a PlaSIM-LSG ensemble simulation run at T21 spectral resolution in the atmosphere, and 3.5 degrees in the ocean, with fixed pre-industrial conditions. A number of collapse events, unseen in the pre-industrial control run, are sampled by the algorithm. Looking at the mechanisms causing the AMOC spontaneous collapse, we find that zonal wind stress over the North Atlantic is the main driver of the initial AMOC slowdown, while the suppression of surface convection in the Labrador sea is the likely cause of the subsequent AMOC collapse. Then, we investigate the influence of increasing CO2 levels on the frequency of these spontaneous AMOC collapses. We show that a higher CO2 not only leads to the well-known weakening of the AMOC mean state, but it also increases the possibility of incurring in abrupt noise-induced transitions. The employment of EMICs, combined with the proposed approach, samples a large number of rare phenomena. This procedure allows us to explore statistical properties that are not accessible with a deterministic approach in state-of-the-art high resolution models.

How to cite: Cini, M., Zappa, G., Ragone, F., and Corti, S.: Exploring Noise-induced and CO2-driven AMOC collapses in the PlaSIM-LSG climate model with a Rare Event Algorithm., EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-1036, https://doi.org/10.5194/egusphere-egu24-1036, 2024.

EGU24-2167 | ECS | Posters on site | NP1.5

Early-Warning Signs for SPDEs under Boundary Noise 

Paolo Bernuzzi, Christian Kuehn, and Henk Dijkstra

The search of early-warning signs able to predict the approach of a parameter to a deterministic bifurcation threshold is relevant in climate as it aims to enable a proper prediction of qualitative changes in the studied models. The observation of such objects in SPDEs (stochastic partial differential equations) permits the consideration of space variables and the ensuing heterogeneity in the behaviour of their solutions.

The presence of Gaussian noise on the boundary of the studied space is used in order to build the signals, whose properties are discussed thoroughly. An example in the form of application of such tools on a climate model is presented and justified. The utility and appropriate use of the results on a more applied perspective are shown.

How to cite: Bernuzzi, P., Kuehn, C., and Dijkstra, H.: Early-Warning Signs for SPDEs under Boundary Noise, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-2167, https://doi.org/10.5194/egusphere-egu24-2167, 2024.

Aeolus 2.0 is an open-source numerical atmosphere model with intermediate complexity designed to capture the dynamics of the atmosphere, especially extreme weather and climate events. The model's dynamical core is built on a novel multi-layer pseudo-spectral moist-convective Thermal Rotating Shallow Water (mcTRSW) model, and it utilizes the Dedalus algorithm, renowned for its efficient handling of spin-weighted spherical harmonics in solving pseudo-spectral problems. Aeolus 2.0 comprehensively characterizes the temporal and spatial evolution of key atmospheric variables, including vertically integrated potential temperature, thickness, water vapor, precipitation, and the influence of bottom topography, radiative transfer, and insolation. It provides a versatile platform with resolutions ranging from smooth to coarse, enabling the exploration of a wide spectrum of dynamic phenomena with varying levels of detail and precision.

The model has been utilized to investigate the adjustment of large-scale localized buoyancy anomalies in mid-latitude and equatorial regions, along with the nonlinear evolution of key variables in both adiabatic and moist-convective environments. Our findings highlight the triggering mechanisms of phenomena such as the Madden-Julian Oscillation (MJO) and the circulation patterns induced by temperature anomalies and buoyancy fields. Furthermore, our simulations of large-scale localized temperature anomalies reveal insights into the impact of perturbation strength, size, and vertical structure on the evolution of eddy heat fluxes, including poleward heat flux, energy, and meridional elongation of the buoyancy field. We observe the initiation of atmospheric instability, leading to precipitation systems, such as rain bands, and asymmetric latent heat release due to moist convection in diabatic environments. This study identifies distinct patterns, including the formation of a comma cloud pattern in the upper troposphere and a comma-shaped buoyancy anomaly in the lower layer, accompanied by the emission of inertia gravity waves. Additionally, the role of buoyancy anomalies in generating heatwaves and precipitation patterns is emphasized, particularly in mid-latitude regions.

In summary, Aeolus 2.0, with its specific capabilities, contributes to our understanding of the complex interactions of moist convection, buoyancy anomalies, and atmospheric dynamics, shedding light on the dynamics of extreme weather events and their implications for climate studies.

References

1. Rostami, M., Zhao, B., & Petri, S. (2022). On the genesis and dynamics of MaddenJulian oscillation-like structure formed by equatorial adjustment of localized heating. Quarterly Journal of the Royal Meteorological Society, 148 (749), 3788-3813. Retrieved from https://rmets.onlinelibrary.wiley.com/doi/abs/10.1002/qj.4388 doi: https://doi.org/10.1002/qj.4388

2. Rostami, M., Severino, L., Petri, S., & Hariri, S. (2023). Dynamics of localized extreme heatwaves in the mid-latitude atmosphere: A conceptual examination. Atmospheric Science Letters, e1188. Retrieved from https://rmets.onlinelibrary.wiley.com/doi/abs/10.1002/asl.1188 doi: https://doi.org/10.1002/asl.1188

How to cite: Rostami, M. and Petri, S.: Exploring Extreme Weather and Climate Events with Aeolus 2.0: A Multi-layer moist-convective Thermal Rotating Shallow Water (mcTRSW) Dynamical Core, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-2867, https://doi.org/10.5194/egusphere-egu24-2867, 2024.

EGU24-4539 | ECS | Posters on site | NP1.5

Extreme value methods in dynamical systems of different complexity 

Ignacio del Amo, George Datseris, and Mark Holland

Extreme value theory provides a universal limit for the extremes of continuous independent and identically distributed random variables and has proven to be robust to generalisation to wider classes of random variables, including stationary processes, some nonstationary processes and even trajectories on deterministic chaotic systems. This universality, together with the fact that these methods require data from only one realization of the system, has been exploited in applications to study many series of climate data.

Fitting a probability distribution to the extreme events of a data series generated by a chaotic dynamical system gives us not only probabilistic predictions of the intensity and return time of the events themselves, but also geometrical information about the local structure of the attractor and the predictability and persistence of the extreme events.

However, these methods are sensitive to the mathematical properties of the dynamical system that generates the data, and are seldomly even mentioned when they are applied to real climate data. One further caveat of these methods is that they are hard to falsify, i.e. we cannot verify easily if an answer is wrong. For these reasons, we explore how these methods respond to different systems with different complexity and different mathematical properties, trying to understand which of the results on the literature could meaningful and which could be numerical artifacts.

How to cite: del Amo, I., Datseris, G., and Holland, M.: Extreme value methods in dynamical systems of different complexity, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-4539, https://doi.org/10.5194/egusphere-egu24-4539, 2024.

Within the climate model hierarchy, simple models usually play the important role of highlighting dynamical processes that can possibly govern climate phenomena. If, in addition, their results are in significant agreement with observations, the processes thus identified are even more likely to regulate the actual phenomena. In this context, the dynamical process of intrinsic variability paced by a deterministic forcing (also called deterministic excitation, DE [1]) is highlighted here by two simple models of different degrees of complexity and set in the different contexts of paleoclimate and physical oceanography. In both cases, despite the simplicity of the models, the results show significant agreement with observations.

The DE mechanism requires the system (i) to possess intrinsic nonlinear relaxation oscillations (ROs) and (ii) to be in the excitable state (i.e., ROs do not emerge spontaneously but can be excited, and therefore paced, by a suitable forcing); moreover, (iii) ROs are excited by a deterministic forcing if a given tipping point is passed.

In the first case [1], the abrupt late Pleistocene glacial terminations are shown by a conceptual model to correspond to the excitation, by the astronomical forcing, of ROs describing glacial-interglacial transitions (e.g., [2]). In the second case [3], ROs describing the Kuroshio Extension low-frequency variability [4] are shown, by a primitive equation ocean model, to be excited remotely by the North Pacific Oscillation. These results show how simple modeling approaches of different complexity advance process understanding and can, therefore, provide theoretical guidelines for interpreting state-of-the-art ESM results.

[1] Pierini S., 2023: The deterministic excitation paradigm and the late Pleistocene glacial terminations. Chaos, 33, 033108.

[2] Gildor H. and E. Tziperman, 2001: A sea ice climate switch mechanism for the 100-kyr glacial cycles. J. Geophys. Res., 106, 9117–9133.

[3] Pierini S., 2014: Kuroshio Extension bimodality and the North Pacific Oscillation: a case of intrinsic variability paced by external forcing. J. Climate, 27, 448-454.

[4] Pierini S., 2006: A Kuroshio Extension System model study: decadal chaotic self-sustained oscillations. J. Phys. Oceanogr., 36, 1605-1625.

How to cite: Pierini, S.: Simple oceanographic and paleoclimate modeling highlights the same dynamical process: intrinsic variability paced by a deterministic forcing, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-4679, https://doi.org/10.5194/egusphere-egu24-4679, 2024.

EGU24-5570 | ECS | Orals | NP1.5

The critical precipitation threshold for the Amazon forest biomass in the LPJmL vegetation model 

Da Nian, Sebastian Bathiany, Boris Sakschewski, Markus Drüke, Lana Blaschke, Maya Ben-Yami, Werner von Bloh, and Niklas Boers

The Amazon rainforest, one of the most important biomes in the world, and recognized as a potential tipping element in the Earth system, has received increasing attention in recent years. Theory and observations suggest that regional climate change from greenhouse gas emissions and deforestation may push the remaining forest toward a catastrophic tipping point.

Despite the urgency to assess the future fate of the Amazon, it remains unclear if state-of-the-art Dynamic Global Vegetation Models (DGVMs) can capture the highly nonlinear dynamics underlying such potentially abrupt dynamics and there is a noticeable scarcity of DGVM evaluations regarding their potential to predict forthcoming tipping points.

In our manuscript, we systematically investigate how the Amazon forest responds in idealized scenarios where precipitation is linearly decreased and subsequently increased between current levels and zero, using the state-of-the-art model LPJmL. We investigate whether large-scale abrupt changes and tipping points occur, and whether early warning signals as expected from theory can be detected. 

Our results indicate a pronounced nonlinearity but reversible behavior between vegetation aboveground biomass (AGB) and mean annual precipitation (MAP) in the LPJmL simulations. In particular, there exists a threshold at a critical rainfall level below which there is a rapid decrease in forest biomass. The value of the threshold is determined by seasonality, evapotranspiration and the adaptive capacity of roots. Significant "early warning signs" can be detected before the transition.

How to cite: Nian, D., Bathiany, S., Sakschewski, B., Drüke, M., Blaschke, L., Ben-Yami, M., von Bloh, W., and Boers, N.: The critical precipitation threshold for the Amazon forest biomass in the LPJmL vegetation model, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-5570, https://doi.org/10.5194/egusphere-egu24-5570, 2024.

EGU24-8104 | ECS | Posters on site | NP1.5

Disentangling the dynamics of the subpolar gyre and its interaction with the AMOC in the CMIP6 ensemble 

Swinda Falkena and Anna von der Heydt

The subpolar gyre (SPG) is one of the climate tipping elements which could have a large impact on the climate in the northern hemisphere. Improving our understanding of its dynamics is key to assessing the likelihood of it passing a tipping point. Some CMIP6 models exhibit abrupt transitions in the sea surface temperature in the SPG region, but the majority does not. The differences in the model response can be related to the stratification bias, with many models having a too strong stratification preventing them from exhibiting bistable gyre dynamics.

To better understand the SPG we study the (lagged) partial correlations between the relevant aspects of its dynamics in the CMIP6 ensemble. In contrast to standard correlations, partial correlations correct for the effect of autocorrelation and the effect of (the past of) other relevant variables. Therefore, it gives a better indication of there being a causal relation. Based on the partial correlation between the sea surface temperature and mixed layer depth we split the ensemble into two groups (strong or negligible relation) and for each select one model to study its dynamics in detail. In addition, we discuss the interaction of the SPG with the Atlantic Meridional Overturning Circulation (AMOC) using the same methods. These results can help in better informing more conceptual climate models of the SPG, AMOC and their interactions, which can be used to study potential tipping dynamics.

How to cite: Falkena, S. and von der Heydt, A.: Disentangling the dynamics of the subpolar gyre and its interaction with the AMOC in the CMIP6 ensemble, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-8104, https://doi.org/10.5194/egusphere-egu24-8104, 2024.

EGU24-8117 | Posters on site | NP1.5

Using a simple model to measure the differences between climate model land surface simulations and FLUXNET observations 

F. Hugo Lambert, Claire Zarakas, Monisha Natchiar S. R., Abigail L. S. Swann, and Charles D. Koven

Complex numerical models of climate consist of simulation of fluid dynamics and thermodynamics on a discrete grid, and parameterizations, which are algorithms that approximate processes smaller than gridscale. Because parameterizations of a given process may be written as different functions of different, potentially non-observable variables, it can be difficult to quantify the process differences between individual climate models and between climate models and the real world.

Here, we attempt to write down a simple linear model that represents the response of the Earth's tropical land surface to atmospheric forcing on monthly timescales in terms of the same observable variables using a technique called continuous structural parameterization. Simulated data are taken from complex General Circulation Models (GCMs) run under the AMIP protocol and a CESM2 perturbed physics ensemble (PPE) of our own devising;  observed measurements are taken from FLUXNET flux tower sites. We find that the simple model captures land surface behaviour well except in mountainous regions.

Establishing a generalised parameter space, we see that most GCMs are in reasonable agreement with FLUXNET at FLUXNET sites, although there is evidence that GCMs consistently slightly overestimate the response of surface turbulent fluxes to downward radiation. Further, it is found that the differences between structurally different AMIP models are considerably greater than the differences between CESM2 PPE members -- even though the PPE parameters are varied across their realistic domain. If the simple model is trained only at GCM spatial gridpoints that contain a FLUXNET site, there is little degradation in simple model performance compared with global training, suggesting that even the few available tropical FLUXNET sites are useful for constraining land surface model response throughout the tropics. This is of course contingent on whether or not point measurements taken by FLUXNET are representative of the wider area around FLUXNET sites.

How to cite: Lambert, F. H., Zarakas, C., Natchiar S. R., M., Swann, A. L. S., and Koven, C. D.: Using a simple model to measure the differences between climate model land surface simulations and FLUXNET observations, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-8117, https://doi.org/10.5194/egusphere-egu24-8117, 2024.

EGU24-8397 | ECS | Orals | NP1.5

AMOC Stability amid Tipping Ice Sheets from Conceptual to Intermediate Complexity Models 

Sacha Sinet, Anna S. von der Heydt, Peter Ashwin, and Henk A. Dijkstra

The Atlantic Meridional Overturning Circulation (AMOC) and polar ice sheets are considered susceptible to critical transitions under climate change. Identified as core tipping elements, their collapse would have global and drastic consequences. Furthermore, the AMOC and polar ice sheets form a complex interacting system, where the collapse of one component can heavily impact the stability of others. In the worst case, this could result in a large-scale domino effect, otherwise known as a cascading tipping event.

In this presentation, our focus is on assessing the stability of the AMOC in the presence of tipping Greenland ice sheet (GIS) and West Antarctica ice sheet (WAIS). While most existing studies agree on the destabilizing impact of a GIS collapse on the AMOC, the consequences of a WAIS collapse remain uncertain. A previous conceptual study suggested that a WAIS tipping event might actually prevent an AMOC collapse against both climate warming and increased GIS meltwater fluxes. Using a better conceptual model of the AMOC, we demonstrate that both the melting rate and natural variability associated with surface meltwater fluxes are decisive factors for this phenomenon to occur. Finally, we present preliminary findings in which the relevance of this stabilizing effect is investigated in the model of intermediate complexity CLIMBER-X.

How to cite: Sinet, S., von der Heydt, A. S., Ashwin, P., and Dijkstra, H. A.: AMOC Stability amid Tipping Ice Sheets from Conceptual to Intermediate Complexity Models, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-8397, https://doi.org/10.5194/egusphere-egu24-8397, 2024.

EGU24-8729 | ECS | Orals | NP1.5

Quantifying risk of a noise-induced AMOC collapse from northern and tropical Atlantic Ocean variability 

Ruth Chapman, Peter Ashwin, Richard Wood, and Jonathan Baker

The Atlantic Meridional Overturning Circulation (AMOC) exerts a major influence on global climate. There is much debate about whether the current strong AMOC may collapse as a result of anthropogenic forcing and/or natural variability. Here, we ask whether internal decadal variability could affect the likelihood of AMOC collapse. We examine natural variability of basin-scale salinities and temperatures in four CMIP6 pre-industrial runs. We fit the CMIP6 variability to several empirical, linear noise models, and to a nonlinear, process-based AMOC model. The variability is weak and its processes inconsistent among the CMIP6 models considered. Based on the CMIP6 variability levels we find that noise-induced AMOC collapse is unlikely in the pre-industrial climate, but plausible if external forcing has shifted the AMOC closer to a threshold, which can be identified for the non-linear model using bifurcation analysis. However the CMIP6 models may systematically underestimate current Atlantic Ocean variability, and we find that substantially stronger variability would increase the likelihood of noise-induced collapse.

How to cite: Chapman, R., Ashwin, P., Wood, R., and Baker, J.: Quantifying risk of a noise-induced AMOC collapse from northern and tropical Atlantic Ocean variability, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-8729, https://doi.org/10.5194/egusphere-egu24-8729, 2024.

EGU24-9253 | ECS | Posters on site | NP1.5

The impact of model resolution on variability in a coupled land atmosphere model 

Oisín Hamilton, Jonathan Demaeyer, Anupama Xavier, and Stéphane Vannitsem

Reduced order quasi-geostrophic land-atmosphere coupled models display qualitatively realistic mid-latitude atmosphere behaviour, meaning that such models can produce typical atmospheric dynamical features such as atmospheric blocking. At the same time, due to a low number of degrees of freedom, they are still simple enough to allow for analysis of the system dynamics. These features mean that these models are well suited to investigating bifurcations in atmospheric dynamics, and use a dynamical systems approach to better understand the corresponding atmospheric behaviour. 

This project introduces a symbolic python workflow for using the flexible  land-atmosphere (qgs, 2020) spectral model with the continuation software AUTO. This work builds on the results of Xavier et al. (2023) to understand how the model variability and predictability is impacted by the model resolution. We also use bifurcation diagrams to better understand how parameters such as atmosphere-land friction impact the atmospheric blocking, and in turn the model atmosphere predictability. This is done for a range of model resolutions to investigate how the number of degrees of freedom impacts both the realism of the model, but also the structures found in the dynamics.

 

Demaeyer, Jonathan & De Cruz, Lesley & Vannitsem, S.: qgs: A flexible Python framework of reduced-order multiscale climate models. Journal of Open Source Software. 5. 2597. 10.21105/joss.02597, 2020. 

 

Xavier, A. K., Demaeyer, J., and Vannitsem, S.: Variability and Predictability of a reduced-order land atmosphere coupled model, EGUsphere [preprint], https://doi.org/10.5194/egusphere-2023-2257, 2023.

How to cite: Hamilton, O., Demaeyer, J., Xavier, A., and Vannitsem, S.: The impact of model resolution on variability in a coupled land atmosphere model, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-9253, https://doi.org/10.5194/egusphere-egu24-9253, 2024.

EGU24-10070 | ECS | Posters on site | NP1.5

Spatial fluctuations of the Arctic sea ice border 

Clara Hummel

Every year, the area of the Arctic sea-ice decreases in the boreal spring and summer and reaches its yearly minimum in the early autumn. Due to global warming, Arctic summer sea ice will most probably disappear. As the sea ice cover decreases, its border is retreating northwards towards the central Arctic. This retreat is not uniform in space and the variability of the border’s movement further North could yield an early warning signal for summer sea ice loss. Here, we track the sea ice border from time series obtained from models of various complexity and observations to study the spatial variability of the border’s movement as Arctic summer sea ice approaches its disappearance.

How to cite: Hummel, C.: Spatial fluctuations of the Arctic sea ice border, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-10070, https://doi.org/10.5194/egusphere-egu24-10070, 2024.

EGU24-11984 | Posters on site | NP1.5

Dependence of simulated variability of surface climate on model complexity – insights from an ensemble of transient simulations of the Last Deglaciation 

Elisa Ziegler, Nils Weitzel, Jean-Philippe Baudouin, Marie-Luise Kapsch, Uwe Mikolajewicz, Lauren Gregoire, Ruza Ivanovic, Paul Valdes, Christian Wirths, and Kira Rehfeld

Climate variability is crucial to our understanding of future climate change and its impacts on societies and the natural world. However, the climate records of the observational era are too short to explore long-term variability. Conversely, an exploration of long transient simulations from state-of-the-art Earth System Models (ESMs) poses high computational demands. It is therefore pertinent to identify the level of complexity sufficient to simulate the variability of surface climate from annual to centennial and longer timescales.

To this end, we use an ensemble of transient simulations of the Last Deglaciation, the last period of significant global warming. The ensemble covers an energy balance model (EBM), models of intermediate complexity (EMICs), general circulation models (GCMs) and ESMs. This constitutes a hierarchy that we categorize based on employed atmosphere and ocean components and their resolution, as well as implemented radiation, land hydrology, vegetation and aerosol schemes.

To investigate the simulated variability of surface temperature and precipitation, we analyze changes in the shapes of their distributions as characterized by their higher order moments – variance, skewness, kurtosis – with warming. These higher order moments relate the tails to the extremes of the distributions. We identify spatial and temporal patterns and how they depend on model complexity. The EMICs can generally match the global and latitudinal changes in temperature variability found in more complex models. However, they lack in precipitation variability. We further find that the EMICs fail to simulate the tails of the precipitation distributions. We observe dependency of variability on the background state, generally increasing with model complexity. However, there is still a large spread between models of similar complexity, some of which can be related to differences in forcings. Furthermore, questions remain on the abilities of models of any complexity to simulate a magnitude of long-term variability similar to that found regionally in proxy reconstructions. Our analysis offers implications as to the complexity needed and sufficient for capturing the full picture of climate change and we offer some first insights into how the findings translate to future projections of climate change.

How to cite: Ziegler, E., Weitzel, N., Baudouin, J.-P., Kapsch, M.-L., Mikolajewicz, U., Gregoire, L., Ivanovic, R., Valdes, P., Wirths, C., and Rehfeld, K.: Dependence of simulated variability of surface climate on model complexity – insights from an ensemble of transient simulations of the Last Deglaciation, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-11984, https://doi.org/10.5194/egusphere-egu24-11984, 2024.

EGU24-12421 | Posters on site | NP1.5

A Novel Process Model of Ocean-Sea-Ice Interaction Using CESM 

Paul Hall, Christopher Horvat, Baylor Fox-Kemper, Samuel Brenner, and Alper Altuntas

We report on the development of a novel process model created to study ocean-sea ice interaction and the dynamics of the upper ocean in the marginal ice zone (MIZ), built using the Community Earth System Model (CESM). Our model uses the MOM6 ocean model and CICE6 sea-ice model as active components within CESM, on a custom ~50km x ~50km grid with a horizontal resolution of ~50m, extending to a depth of 75m (30 vertical layers). The model allows for either reflecting or zonally re-entrant boundary configurations. Atmospheric forcing is imposed through a simplified data atmosphere component that provides constant forcing over the model domain. Results from several simple scenarios are presented and compared to results obtained using the MITgcm.

By working within CESM, we are able to leverage CESM’s existing infrastructure and capabilities, including the use of the Community Mediator for Earth Prediction Systems (CMEPS) for coupling between active components. Furthermore, additional model components that are already available within CESM (e.g., waves, atmosphere) can be incorporated into the process model in a straightforward way. Future work will include incorporation of a modified sea-ice component that allows tracking of individual floes utilizing a discrete element method approach.

How to cite: Hall, P., Horvat, C., Fox-Kemper, B., Brenner, S., and Altuntas, A.: A Novel Process Model of Ocean-Sea-Ice Interaction Using CESM, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-12421, https://doi.org/10.5194/egusphere-egu24-12421, 2024.

EGU24-14067 | ECS | Orals | NP1.5

A Forecast Test for Reducing Dynamical Dimensionality of Model Emulators 

Tongtong Xu, Matthew Newman, Michael Alexander, and Antonietta Capotondi

The climate system can be numerically represented by a set of physically-based dynamical equations whose solution requires substantial computational resources. This makes computationally efficient, low dimensional emulators that simulate trajectories of the underlying dynamical system an attractive alternative for model evaluation and diagnosis. We suggest that since such an emulator must adequately capture anomaly evolution, its construction should employ a grid search technique where maximum forecast skill determines the best reference model. In this study, we demonstrate this approach by testing different bases used to construct a Linear Inverse Model (LIM), a stochastically-forced multivariate linear model that has often been used to represent the evolution of coarse-grained climate anomalies in both models and observations. LIM state vectors are typically represented in a basis of the leading Empirical Orthogonal Functions (EOFs), but while dominant large-scale climate variations often are captured by a subset of these statistical patterns, key precursor dynamics involving relatively small scales are not. An alternative approach is balanced truncation, where the dynamical system is transformed into its Hankel space, whose modes span both precursors and their subsequent responses. Constructing EOF- and Hankel-based LIMs from monthly observed anomalous Pacific sea surface temperatures, both for the 150-yr observational record and a perfect model study using 600 yrs of LIM output, we find that no balanced truncation model of any dimension can outperform an EOF-based LIM whose dimension is chosen to maximize independent skill. However, the dynamics of a high-dimensional EOF-based LIM can be efficiently reproduced by far fewer Hankel modes.

How to cite: Xu, T., Newman, M., Alexander, M., and Capotondi, A.: A Forecast Test for Reducing Dynamical Dimensionality of Model Emulators, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-14067, https://doi.org/10.5194/egusphere-egu24-14067, 2024.

EGU24-14439 | ECS | Posters on site | NP1.5

Nonlocal energy fluxes and fractional operators in updated, stochastic, Budyko-Sellers models 

dustin lebiadowski and shaun Lovejoy

We introduce a stochastic, energy-balance, climate model defined over the macroweather regime (approximately 15 days or longer). Together, the energy balance principle, combined with the model’s natural scaling, demonstrate quite promising results despite the relative simplicity. A special case of the model can also be derived from a very classical basis, and, because of some similarities, we propose this model as a development upon the work of Budyko and Sellers.

When the classical Budyko-Sellers energy balance model is updated by using the (correct) radiative-conductive surface boundary conditions, one obtains the Fractional Energy Balance Equation (FEBE). The FEBE involves fractional space-time operators and its generic solutions are scaling, in agreement with much atmospheric and oceanic data. In time, it implies long range memories that have been successfully used to make both multi-decadal climate projections as well as monthly and seasonal (long range) forecasts. In space, the FEBE is nonlocal so that energy flux imbalances at any location can affect the balance in locations far away. This is possible because the model operates over monthly and longer time scales; over these scales, energy can be both stored and transported in the atmosphere, ocean, and subsurface.

Until now, the FEBE’s full nonlocal space-time interaction operator has been only approximated. Here, by introducing a numerical model, the nonlocal dynamics of the FEBE and corresponding Earth-system FEBE energy flows over the 2D Earth surface are fully detailed.

We propose the FEBE as an alternative to more conventional, deterministic, weather-regime-based climate models. Given the generality of the ideas pursued here - the use of fractional operators; the use of stochasticity and the macroweather regime - there seems a great potential for these to be used much more widely. Hopefully this research, and possibly related works, will encourage a greater diversity of pursuits and be inspiring to others in their own work.

How to cite: lebiadowski, D. and Lovejoy, S.: Nonlocal energy fluxes and fractional operators in updated, stochastic, Budyko-Sellers models, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-14439, https://doi.org/10.5194/egusphere-egu24-14439, 2024.

EGU24-14831 | ECS | Orals | NP1.5

SPEEDY-NEMO: performance and applications of a fully-coupled intermediate-complexity climate model 

Paolo Ruggieri, Muhammad Adnan Abid, Javier Garcia-Serrano, Carlo Grancini, Fred Kucharski, Salvatore Pascale, and Danila Volpi

A fully-coupled general circulation model of intermediate complexity is documented. The study presents an overview of the model climatology and variability, with particular attention for the phenomenology of processes that are relevant for the predictability of the climate system on seasonal-to-decadal time-scales. It is shown that the model can realistically simulate the general circulation of the atmosphere and the ocean, as well as the major modes of climate variability on the examined time-scales: e.g. El Niño-Southern Oscillation, North Atlantic Oscillation, Tropical Atlantic Variability, Pacific Decadal Variability, Atlantic Multi-decadal Variability. We demonstrate the ability of the model in simulating non-stationarity of coupled ocean-atmosphere modes of variability. Potential applications of the model are discussed, with emphasis on the possibility to generate sets of low-cost large-ensemble retrospective forecasts. We argue that the presented model is suitable to be employed in traditional and innovative model experiments that can play a significant role in future developments of seasonal-to-decadal climate prediction.

How to cite: Ruggieri, P., Abid, M. A., Garcia-Serrano, J., Grancini, C., Kucharski, F., Pascale, S., and Volpi, D.: SPEEDY-NEMO: performance and applications of a fully-coupled intermediate-complexity climate model, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-14831, https://doi.org/10.5194/egusphere-egu24-14831, 2024.

EGU24-16958 | Orals | NP1.5

The Challenge of Non-Markovian Energy Balance Models in Climate 

Nicholas Wynn Watkins, Raphael Calel, Sandra Chapman, Aleksei Chechkin, Rainer Klages, and David Stainforth

Hasselmann’s paradigm, introduced in 1976 and recently honoured with the Nobel Prize, can, like many key innovations in the sciences of climate and complexity, be understood on several different levels, both technical and conceptual. It can be seen as a mathematical technique to add stochastic variability into pioneering energy balance models (EBMs) of Budyko and Sellers. On a more conceptual level, it used the mathematics  of Brownian motion to provide an  abstract superstructure linking slow climate variability to fast weather fluctuations, in a context broader than EBMs, leading Hasselmann to posit the need for negative feedback in climate modelling.

Hasselmann's paradigm itself has much still to offer us [e.g. Calel et al, Naure Communications, 2020], but naturally, since the 1970s a number of newer developments have built on his pioneering ideas. One important one has been the development of a rigorous mathematical hierarchy that embeds Hasselmann-type models in the more comprehensive Mori-Zwanzig (MZ) framework  (e.g.  Lucarini and Chekroun, Nature Reviews Physics, 2023). Another has been the interest in long range memory in stochastic EBMs, notably Lovejoy et al’s Fractional Energy Balance Equation [FEBE, discussed in this week’s Short Course SC5.15 ]. These have a memory with slower decay and thus longer range than the exponential form seen in Hasselmann’s EBM. My presentation [based on Watkins et al, in review at Chaos] attempts to build a bridge between MZ-based extensions of  Hasselmann, and the fractional derivative-based FEBE model.  I will argue that the Mori-Kubo overdamped Generalised Langevin Equation, as widely used in statistical mechanics, suggests the form of a relatively simple stochastic EBM with memory for the global temperature anomaly, and will discuss how this relates to FEBE.

How to cite: Watkins, N. W., Calel, R., Chapman, S., Chechkin, A., Klages, R., and Stainforth, D.: The Challenge of Non-Markovian Energy Balance Models in Climate, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-16958, https://doi.org/10.5194/egusphere-egu24-16958, 2024.

EGU24-18360 | Posters on site | NP1.5

Eddy Saturation and Latitudinal Storm Track Shift in a Reduced Two-level Model of the Atmosphere 

Valerio Lucarini, Melanie Kobras, and Maarten Ambaum

We introduce a minimal dynamical system derived from the classical Phillips two-level model with the goal of elucidating the essential mechanisms responsible for the interaction between eddies and mean flow. The choice of a two-level model as starting points allows for appreciating the relative role of barotropic and baroclinic processes. Specifically, we wish to explore the eddy saturation mechanism, whereby, when average conditions are considered, direct forcing of the zonal flow increases the eddy kinetic energy, while the energy associated with the zonal flow does not increase. The eddy-driven jet stream and storm tracks in the mid-latitude atmosphere are known to shift in latitude on various timescales, but the physical processes that cause these shifts are still unclear. Using our low-order model, we aim to understand the link between the structure of the eddies and the shift of the latitudinal maximum of the zonal flow in the mid-latitude atmosphere. Our findings elucidate the basic mechanisms behind baroclinic adjustment and provide insights into the properties of the storm track change between the jet entrance and jet exit regions of the North Atlantic.

How to cite: Lucarini, V., Kobras, M., and Ambaum, M.: Eddy Saturation and Latitudinal Storm Track Shift in a Reduced Two-level Model of the Atmosphere, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-18360, https://doi.org/10.5194/egusphere-egu24-18360, 2024.

EGU24-18412 | ECS | Orals | NP1.5

Rapid Emulation of Spatially Resolved Temperature Response Functions to Effective Radiative Forcing 

Christopher Womack, Noelle Eckley Selin, and Sebastian Eastham

We utilize ideas from signal processing to demonstrate a novel methodology for climate emulation based on the response of the climate system to effective radiative forcing (ERF). While previous work has demonstrated the efficacy of impulse response functions as a tool for climate emulation, these methods are largely non-generalizable to new scenarios and are inaccessible to more general audiences. To remedy this, we propose a generalizable framework for emulation of climate variables such as near-surface air temperature, representing the climate system through the surrogate of spatially resolved impulse response functions. These response functions are derived through the deconvolution of ERF and near-surface air temperature profiles, treating ERF and near-surface air temperature as input and output signals, respectively. Using this framework, new scenarios can be quickly and easily emulated through convolution and other sets of impulse response functions can be derived from any pair of climate variables. We present results from an application to near-surface air temperature based on ERF and temperature data taken from experiments in the sixth phase of the Coupled Model Intercomparison Project (CMIP6). We evaluate the emulator using additional experiments taken from the CMIP6 archive, including the Shared Socioeconomic Pathways (SSPs), demonstrating accurate emulation of global mean and spatially resolved temperature change with respect to the outputs of the CMIP6 ensemble. Global absolute error in emulated temperature averages 0.25 degrees Celsius with a bias ranging from -0.14 to -0.04 degrees Celsius. We additionally show how our emulator can be implemented as a tool for climate education through integration with the En-ROADS platform, providing fast visualizations of spatially resolved temperature change for a number of policy-relevant scenarios. While it is unable to capture state-dependent climate feedbacks, such as the non-linear effects of Arctic sea ice melt in high-warming scenarios, our results show that the emulator is generalizable to any scenario independent of the specific forcings present.

How to cite: Womack, C., Eckley Selin, N., and Eastham, S.: Rapid Emulation of Spatially Resolved Temperature Response Functions to Effective Radiative Forcing, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-18412, https://doi.org/10.5194/egusphere-egu24-18412, 2024.

EGU24-22102 | Orals | NP1.5 | Highlight

From conceptual to complex earth system models: why are models so linear? 

Victor Brovkin, Tobias Stacke, Philipp de Vrese, Thomas Kleinen, and Alexander Winkler

The evolution of the Earth’s climate from the past to the future is explored by a hierarchy of models ranging from conceptual models to full-complexity, high-resolution Earth System Models (ESMs) (Claussen et al., 2002). The strength of conceptual models lies in the clarity of representing the concept of interactions between different climate processes, while ESMs  offer greater realism when it comes to spatial or temporal detail. Intermediate complexity models are somewhere in between, they are able to provide a big picture for long timescales. A common pattern throughout the model hierarchy, except for conceptual models illustrating multiple steady states, is often linearity of model responses to external forcing. This linearity can be visible in transient experiments, but also in equilibrium simulations. The question arises: is this linearity an artefact of our models, or is it reflective of reality?

 

In most cases, the linear response is likely representative of reality. As an example, we will focus on the linearity of land-related processes, such as climate-carbon feedbacks and permafrost-hydrology interactions. Permafrost systems have thresholds at 0°C, leading to nonlinearities at the local scale, but the combined response at large spatial scales tends to be more linear. However, nonlinear and abrupt changes are evident in geological records. For instance, the abrupt onset of the Bölling/Alleröd warming about 14.8 thousand years ago indicates that nonlinear changes on large spatial scales are indeed a real, albeit very rare, phenomenon. We will discuss possible reasons for the predominant linearity of the models and explore whether high-resolution models might show more nonlinear responses than coarse-grid models.

 

Reference:

Claussen, M., Mysak, L., Weaver, A. et al. Earth system models of intermediate complexity: closing the gap in the spectrum of climate system models. Climate Dynamics 18, 579–586 (2002). https://doi.org/10.1007/s00382-001-0200-1

How to cite: Brovkin, V., Stacke, T., de Vrese, P., Kleinen, T., and Winkler, A.: From conceptual to complex earth system models: why are models so linear?, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-22102, https://doi.org/10.5194/egusphere-egu24-22102, 2024.

EGU24-348 | ECS | Orals | CR3.3

Ridge-formation simulations in three dimensions using discrete element methods 

Marek Muchow and Arttu Polojärvi

Sea-ice ridges form as a part of sea-ice deformation, while the ice is moved by winds and ocean currents. While ridging is a localized process, it is assumed to limit the compressive strength of sea ice in large scale. However, formulations of large-scale ice strength, as used in Earth System Models, do not consider individual ridge formation processes in detail. Thus, it is necessary to understand the energy spend in ridge formation and various processes related to generating ice rubble and redistributing it. To investigate ridge formation in detail, we use the Aalto University in-house discrete-element-method (DEM) model. This three-dimensional DEM model features deformable, multi-fracturing, ice floes, which can fail and form ridges when coming into contact, while recording the ridging forces. With this, we discuss why three-dimensional simulations are important to investigate ridge formation process.

How to cite: Muchow, M. and Polojärvi, A.: Ridge-formation simulations in three dimensions using discrete element methods, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-348, https://doi.org/10.5194/egusphere-egu24-348, 2024.

EGU24-1671 | ECS | Posters on site | CR3.3

Monthly Arctic sea ice prediction based on a data-driven deep learning model  

Xiaohe Huan, Jielong Wang, and Zhongfang Liu

There is growing interest in sub-seasonal to seasonal predictions of Arctic sea ice due to its potential effects on midlatitude weather and climate extremes. Current prediction systems are largely dependent on physics-based climate models. While climate models can provide good forecasts for Arctic sea ice at different timescales, they are susceptible to initial states and high computational costs. Here we present a purely data-driven deep learning model, UNet-F/M, to predict monthly sea ice concentration (SIC) one month ahead. We train the model using monthly satellite-observed SIC for the melting and freezing seasons, respectively. Results show that UNet-F/M has a good predictive skill of Arctic SIC at monthly time scales, generally outperforming several recently proposed deep learning models, particularly for September sea-ice minimum. Our study offers a perspective on sub-seasonal prediction of future Arctic sea ice and may have implications for forecasting weather and climate in northern midlatitudes.

How to cite: Huan, X., Wang, J., and Liu, Z.: Monthly Arctic sea ice prediction based on a data-driven deep learning model , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-1671, https://doi.org/10.5194/egusphere-egu24-1671, 2024.

EGU24-2377 | ECS | Posters on site | CR3.3

Multivariate state and parameter estimation using data assimilation in a Maxwell-Elasto-Brittle sea ice model 

Yumeng Chen, Polly Smith, Alberto Carrassi, Ivo Pasmans, Laurent Bertino, Marc Bocquet, Tobias Sebastian Finn, Pierre Rampal, and Véronique Dansereau

In an idealised setup, a dynamics-only sea ice model is used to investigate the fully multivariate state and parameter estimations that uses a novel Maxwell-Elasto-Brittle (MEB) sea ice rheology. In the fully multivariate state estimation, the level of damage, internal stress and cohesion are estimated along with the observed sea ice concentration, thickness and velocity. In the case of parameter estimation, we estimate the air drag coefficient and the damage parameter of the MEB model. The air drag coefficients adjust the strength of the forcing on the sea ice dynamics while the damage parameter controls the mechanical behaviour of the internal property of sea ice. We show that, with the current observation network, it is possible to improve all model state forecast and the parameter accuracy using data assimilation approaches even though problems could arise in such an idealised setup where the external forcing dominates the model forecast error growth.

How to cite: Chen, Y., Smith, P., Carrassi, A., Pasmans, I., Bertino, L., Bocquet, M., Finn, T. S., Rampal, P., and Dansereau, V.: Multivariate state and parameter estimation using data assimilation in a Maxwell-Elasto-Brittle sea ice model, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-2377, https://doi.org/10.5194/egusphere-egu24-2377, 2024.

EGU24-3367 | ECS | Orals | CR3.3

Perscribing Antarctic landfast sea ice in a sea ice-ocean model. 

Noé Pirlet, Thierry Fichefet, Martin Vancoppenolle, Clément Rousset, Pierre Mathiot, Alexander Fraser, Antoine Barthélemy, and Christoph Kittel

The coastal polynyas of the Southern Ocean play a crucial role in the formation of dense water and have an impact on the stability of ice shelves. Therefore, it is important to accurately simulate them in climate models. To achieve this goal, the relationship between grounded icebergs, landfast ice and polynyas appears to be central. Indeed, grounded icebergs and landfast ice are believed to be key drivers of coastal polynyas. However, ESMs do not simulate Antarctic landfast ice. Moreover, at a circumpolar scale, there are no observations of grounded icebergs available. Hence, we must seek model representations that can overcome these issues. To address these gaps, we conducted a study using an antarctic circumpolar configuration of the ocean–sea ice model NEMO4.2-SI–3 at the 1/4° resolution. We ran two simulations for the period 2001–17, with the only difference being the inclusion or exclusion of landfast ice information based on observations. All other factors, including initial conditions, resolution and atmospheric forcings, were kept the same. We then compared the results of these simulations with observations from the advanced microwave scanning radiometer to evaluate the performance of the new simulation. Our analysis allowed us to determine the extent to which prescribing the distribution of landfast ice and setting the sea ice velocity to zero on landfast ice regions influenced various aspects of the sea ice, such as polynyas, landfast ice and sea ice distribution in the model. In the future, we plan to look at the impact on the ocean and to develop a physical parameterization in order to model landfast ice and consequently polynyas on a permanent basis.

How to cite: Pirlet, N., Fichefet, T., Vancoppenolle, M., Rousset, C., Mathiot, P., Fraser, A., Barthélemy, A., and Kittel, C.: Perscribing Antarctic landfast sea ice in a sea ice-ocean model., EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-3367, https://doi.org/10.5194/egusphere-egu24-3367, 2024.

EGU24-4144 | ECS | Orals | CR3.3

A model for ice-mélange based on particle and continuums mechanics 

Saskia Kahl and Carolin Mehlmann

Ice mélange (a mixture of sea ice, bergy bits and icebergs) can have a strong influence on the sea-ice-ocean interaction. So far, ice mélange is not represented in climate models as numerically efficient realizations are missing. This motivates the development of an ice-mélange model based on the viscous-plastic sea-ice rheology, which is currently the most commonly used material law for sea ice in climate models. Starting from the continuum mechanical formulation, we modify the rheology so that icebergs are represented by thick, highly compact pieces of sea ice. These compact pieces of sea ice are held together by a modified tensile strength in the material law. In this framework, the ice mélange is considered as one single fluid, where the icebergs are realised by particles.
Using idealized test cases, we demonstrate that the proposed changes in the material law are crucial to represent icebergs with the viscous-plastic rheology. Similar to the viscous-plastic sea-ice model, the ice-mélange model is highly nonlinear. Solving the model at the resolution needed to represent the typical size of icebergs in ice mélange (< 300m) is therefore challenging. We show that the ice-mélange formulation can be approximated efficiently with a modified Newton's method. Overall, the simple extension of the viscous-plastic sea-ice model is a promising path towards the integration of ice mélange into climate models.

How to cite: Kahl, S. and Mehlmann, C.: A model for ice-mélange based on particle and continuums mechanics, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-4144, https://doi.org/10.5194/egusphere-egu24-4144, 2024.

Rapid decline of Arctic sea ice has created more open water for ocean wave development and highlighted the importance of wave-ice interactions in the Arctic. Some studies have made contributions to our understanding of the potential role of the prognostic floe size distribution (FSD) on sea ice changes. However, these efforts do not capture the full interactions between atmosphere, ocean, wave, and sea-ice. In this study, a modified joint floe size and thickness distribution (FSTD) is implemented in a newly-developed regional atmosphere-ocean-wave-sea ice coupled model and a series of pan-Arctic simulation is conducted with different physical configurations related to FSD changes, including FSD-fixed, FSD-varied, lateral melting rate, wave-fracturing formulation, and wave attenuation rate. Firstly, atmosphere-ocean-wave-sea ice coupled simulations show that the prognostic FSD leads to reduced ice area due to enhanced ice-ocean heat fluxes, but the feedbacks from the atmosphere and the ocean partially offset the reduced ice area induced by the prognostic FSD. Secondly, lateral melting rate formulations do not change the simulated FSD significantly, but they influence the flux exchanges across atmosphere, ocean, and sea-ice and thus sea ice responses. Thirdly, the changes of FSD are sensitive to the simulated wave parameters associated with different wave-fracturing formulations and wave attenuation rates, and the limited oceanic energy imposes a strong constraint on the response of sea ice to FSD changes. Finally, the results also show that wave-related physical processes can have impacts on sea ice changes with the constant FSD, indicating the indirect influences of ocean waves on sea-ice through the atmosphere and the ocean.

How to cite: Yang, C.-Y. and Liu, J.: Understanding influence of ocean waves on Arctic sea ice simulation: A modeling study with an atmosphere-ocean-wave-sea ice coupled model, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-4502, https://doi.org/10.5194/egusphere-egu24-4502, 2024.

EGU24-5177 | ECS | Posters on site | CR3.3

Improving the representation of snow over sea-ice in the SI3 model 

Theo Brivoal, Virginie Guemas, Clement Rousset, and Martin Vancoppenolle

Snow plays a crucial role in the formation and sustainability of sea ice. Due to its thermal properties, snow acts as an insulating layer, shielding the ice from the air above. This insulation reduces the heat transfer between the sea-ice and the atmosphere. Due to its reflective properties, the snow cover also strongly contributes to albedo over ice-covered region, which gives it a significant role in the Earth's climate system.

Current state-of-art climate models use over-simple representations of the snow cover. The snow cover is often represented with a one-layer scheme, assuming a constant density, no wet or dry metamorphism or assuming that no liquid water is stored in the snow. Here, we present the integration of a more advanced snow scheme (ISBA-ES) into the sea-ice model SI3, which serves as the sea-ice component for upcoming versions of the CNRM climate model (CNRM-CM). We compare 1D simulations over the Arctic using this new scheme with observational data and simulations utilizing the previous SI3 snow scheme. Overall, the snow simulated by the ISBA-ES scheme is realistic. We also present a sensitivity analysis of the snow and sea-ice in the SI3 model, exploring various options in the ISBA-ES scheme. Our findings reveal a strong sensitivity of both the snow and the sea-ice to the representation of liquid water in snow and the parameterization employed for calculating snowfall density.

How to cite: Brivoal, T., Guemas, V., Rousset, C., and Vancoppenolle, M.: Improving the representation of snow over sea-ice in the SI3 model, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-5177, https://doi.org/10.5194/egusphere-egu24-5177, 2024.

EGU24-5374 | ECS | Orals | CR3.3

Floe-scale ocean / sea ice energy transfers in the marginal ice zone 

Mukund Gupta, Andrew Thompson, and Patrice Klein

Marginal ice zones are regions where individual sea ice floes interact mechanically and thermodynamically with turbulent ocean currents at the (sub-)mesoscale. Fine scale exchanges of momentum, heat and salinity at the interface between the ocean and the sea ice floes have important effects on upper-ocean energetics, under-ice tracer mixing, and the ice-pack melt rates. The dynamics of these moving floes remain poorly constrained, notably due to the challenge of numerically resolving sub-mesoscale processes and modelling the discrete behavior of sea ice in traditional climate models. 

Here, we use oceanic Large Eddy Simulations (LES), two-way coupled to a Discrete Element Model (DEM) of disk-shaped sea ice floes, to quantify the kinetic energy transfers between ocean and sea ice during summer-like conditions, varying sea ice concentration and floe size distribution. The damping of oceanic currents by floes is found to be important for a sea ice concentration as low as 40%, when the sizes of floes are comparable to the characteristic eddy size. This damping is largely compensated by the generation of kinetic energy due to melt-induced baroclinic instability at the edge of sea ice floes, leading to a net energy sink of approximately 15%, relative to a simulation with no floes. At higher sea ice concentrations, the oceanic kinetic energy production weakens, while energy loss due to ice/ocean damping and floe-floe collisions both increase. These energy fluxes are mediated by the spatial aggregation of sea ice floes that occurs within the high-strain regions surrounding ocean mesoscale eddies. Eddy-driven aggregation can also reduce the melt rate of small floes as they become shielded from warm waters by neighboring larger floes. These results highlight the need for scale-aware, and specifically floe-scale parameterizations of sea ice and its coupling to ocean turbulence, within global climate models.

How to cite: Gupta, M., Thompson, A., and Klein, P.: Floe-scale ocean / sea ice energy transfers in the marginal ice zone, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-5374, https://doi.org/10.5194/egusphere-egu24-5374, 2024.

EGU24-6441 | ECS | Posters on site | CR3.3

Application of mixed least-squares FEM to study sea ice dynamics 

Sonja Hellebrand, Carina Schwarz, and Jörg Schröder

The behavior of sea ice has been studied for many decades. In order to model its viscous-plastic behavior at scales spanning several thousand kilometers, different numerical models have been proposed. Based on the established approach in [1], this contribution presents a simulation model for sea ice dynamics to describe the sea ice circulation and its evolution over one seasonal cycle. In course of that, the sea ice concentration and the sea ice thickness are considered, of which the physical behavior is governed by transient advection equations. Here, the sea ice velocity serves as coupling field.

Recently developed approaches base on a finite element implementation choosing a (mixed) Galerkin variational approach, see e.g. [2] and [3]. But therein, challenges may occur regarding the stability of the numerically complex scheme, especially when dealing with the first-order advection equations. Thus, we propose the application of the mixed least-squares finite element method, which has the advantage to be also applicable to first-order systems, i.e., it provides stable and robust formulations even for non-self-adjoint operators, such as the tracer equations (for sea ice thickness and sea ice concentration).

For solving the instationary sea ice equation the presented least-squares finite element formulation takes into account the balance of momentum and a constitutive law for the viscous-plastic flow. The considered primary fields are the stresses σ, the velocity v, the concentration Aice and the thickness Hice. In relation, four residuals are defined for the derivation of a first-order least-squares formulation based on the balance of momentum, the constitutive relation for the stresses, and two tracer-equations. Different approaches can be made with respect to the approximation functions of the primary fields, i.e., choosing e.g. conforming (H(div) interpolation functions) or non-conforming (Lagrangian interpolation functions) stress approximations, while Lagrangian interpolation functions are chosen for the remaining fields. In order to compare such approaches, the box test case is utilized, cf. [3], which is well described in literature.

References:

[1] W.D. Hibler III. A dynamic thermodynamic sea ice model. Journal of Physical Oceanography, 9(4):815-846, 1979.

[2] S. Danilov, Q. Wang, R. Timmermann, M. Iakovlev, D. Sidorenko, M. Kimmritz, T. Jung. Finite-Element Sea Ice Model (FESIM), Version 2. Geoscientific Model Development, 8:1747-1761, 2015.

[3] C. Mehlmann and T. Richter. A modified global Newton solver for viscous-plastic sea ice models. Ocean Modelling, 116:96-107, 2017.

How to cite: Hellebrand, S., Schwarz, C., and Schröder, J.: Application of mixed least-squares FEM to study sea ice dynamics, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-6441, https://doi.org/10.5194/egusphere-egu24-6441, 2024.

EGU24-6569 | ECS | Posters on site | CR3.3

Using discrete element methods to understand in-plane fragmentation of sea ice floes 

Adam Bateson, Daniel Feltham, David Schröder, Scott Durski, Jennifer Hutchings, Rajlaxmi Basu, and Byongjun Hwang

Sea ice floe size can impact several processes that determine the evolution of the Arctic sea ice, including lateral melt volume, momentum exchange, and rheology. Floe size distribution (FSD) models are applied within continuum sea ice models to capture the evolution of the FSD through parameterisations of the processes that modify floe size such as lateral melting and wave break-up of floes. FSD models do not yet adequately resolve in-plane fragmentation processes of floes such as the breakup of floes under wind forcing, through interactions between neighbouring floes, or through thermal weakening. It is challenging to characterise and therefore parameterise these in-plane floe breakup processes due to limited availability of in-situ observations. Discrete element models (DEMs) offer an alternative way to understand the different mechanisms of floe fragmentation. By resolving relevant properties such as shear and normal stress and sea ice strength at the sub-floe scale, it is possible to use DEMs as a virtual laboratory and directly simulate the break-up of floes into smaller fragments.

In this study, we describe how in-situ observations of sea ice can be combined with output from sea ice DEMs to develop parameterisations of in-plane breakup of floes that can then be applied in continuum models. We then discuss the necessary model developments in order to apply a sea ice DEM to floe fragmentation at smaller scales. We will also present results from a series of DEM simulations used to model the fracture of sea ice under different forcing conditions and with varying sea ice states to identify the important sea ice parameters and processes in determining the size of the floes that form from in-plane breakup events.

How to cite: Bateson, A., Feltham, D., Schröder, D., Durski, S., Hutchings, J., Basu, R., and Hwang, B.: Using discrete element methods to understand in-plane fragmentation of sea ice floes, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-6569, https://doi.org/10.5194/egusphere-egu24-6569, 2024.

Arctic sea ice has experienced a differential decline in speed due to the same anthropogenic greenhouse gas forcing, as evidenced by rapid decline after the end of the last century. Our convergent observations, last-millennium reanalysis, and model analyses have revealed that large tropical volcanic eruptions can lead to a decadal increase in Arctic sea ice, and the 1982 and 1991 large volcanic eruptions slowed down the decline of Arctic sea ice during the last century. The models, selected based on the observed sensitivity of Arctic sea ice to volcanic eruptions, suggest that the earliest ice-free summer year in the Arctic will be around 2040 in high-emission sceneria of SSP585. These findings emphasized the crucial need to incorporate volcanic influences when projecting future Arctic changes amid global warming.

How to cite: Wang, X.: Historical volcanic eruptions slowed down rapid decline in Arctic sea ice linked to global warming, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-9554, https://doi.org/10.5194/egusphere-egu24-9554, 2024.

EGU24-9802 | ECS | Posters virtual | CR3.3

The sea ice component of MUSE, the unstructured-mesh global ocean model of CMCC 

Francesco Cocetta, Lorenzo Zampieri, and Doroteaciro Iovino

The rapidly evolving sea ice cover requires novel modeling approaches and mathematical techniques to accurately simulate the sea ice dynamics, thermodynamics, and its interactions with the atmosphere and ocean at varying spatiotemporal resolutions. In this context, the CMCC is developing the Multiscale Unstructured model for Simulating the Earth’s water environment (MUSE), a novel global ocean-sea ice model on unstructured meshes.

MUSE employs a finite-element numerical discretization on unstructured meshes, aiming at offering flexibility in simulating the global ocean for various applications, ranging from physical process understanding to operational sea ice predictions. The ongoing implementation of the sea ice component utilizes the traditional continuous sea ice formulation and the 2+1 split assumption, meaning that the sea ice dynamics and advection are solved for horizontal motions while the thermodynamics and radiative processes are parameterized at the subgrid scale.   

MUSE employs a modified elastic-viscous-plastic (mEVP) solver for the sea ice dynamics and a Flux Corrected Transport (FCT) advection scheme, alongside the state-of-the-art column physics package "Icepack" maintained by the CICE consortium.

Here, we describe the global implementation of the sea ice component in MUSE and its coupling with the ocean. We present the resulting representation of vertical thermodynamic processes and horizontal dynamics of sea ice.

How to cite: Cocetta, F., Zampieri, L., and Iovino, D.: The sea ice component of MUSE, the unstructured-mesh global ocean model of CMCC, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-9802, https://doi.org/10.5194/egusphere-egu24-9802, 2024.

EGU24-10098 | Posters on site | CR3.3

Best of SIDFEx: Highlights and lessons learned from six years of sea-ice drift forecasting 

Simon F. Reifenberg, Valentin Ludwig, and Helge F. Goessling and the SIDFEx Team

We showcase the Sea Ice Drift Forecast Experiment (SIDFEx) database. SIDFEx is a collection of close to 225,000 lagrangian drift forecasts for the trajectories of assets (mostly buoys) on the Arctic and Antarctic sea ice, at lead times from daily to seasonal with mostly daily resolution. The forecasts are based on systems with varying degrees of complexity, ranging from free-drift forecasts to forecasts by fully coupled dynamical general circulation models. Combining several independent forecasts allows us to construct a best-guess consensus forecast, with a seamless transition from systems with lead times of up to 10 days to systems with seasonal lead times. The forecasts are generated by 13 research groups using 23 distinct forecasting systems and sent regularly to the Alfred-Wegener-Institute, where they are archived and evaluated. Many groups send forecasts operationally in near-real time.

In our presentation, we will introduce the motivation behind and setup of SIDFEx, as well as an overview on the general forecast skill. We will focus on selected highlights, comprising the operational support of research cruises, short-term predictions of sea-ice deformation and regular contributions to the Sea Ice Outlook competition.

How to cite: Reifenberg, S. F., Ludwig, V., and Goessling, H. F. and the SIDFEx Team: Best of SIDFEx: Highlights and lessons learned from six years of sea-ice drift forecasting, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-10098, https://doi.org/10.5194/egusphere-egu24-10098, 2024.

EGU24-11288 | Orals | CR3.3

Towards improving numerical sea ice predictions with data assimilation and machine learning 

William Gregory, Mitchell Bushuk, Yongfei Zhang, Alistair Adcroft, and Laure Zanna

In this presentation we highlight recent developments in the implementation of Machine Learning (ML) algorithms into the large-scale sea ice model, SIS2. Specifically, we show how a Convolutional Neural Network (CNN) can be used to systematically reduce global sea ice biases during a 5-year ice-ocean simulation. The CNN has been trained to learn a functional mapping from model state variables to sea ice concentration Data Assimilation (DA) increments. Therefore, during model integration, the CNN ingests information about the numerical model's atmosphere, ocean, and sea ice conditions, and predicts the appropriate correction to the sub-grid category sea ice concentration terms (without seeing any actual sea ice observations). We also show how this combined DA+ML approach leads to a natural framework for augmenting training data for neural networks; one which can lead to significant improvements in online performance, without the need for direct online learning. The bias reductions over the 5-year simulation period for this CNN correction scheme are even competitive with the bias reductions achieved from DA. These findings therefore suggest that our approach could be used to reduce systematic sea ice biases in fully coupled climate model predictions on seasonal-to-climate timescales.

How to cite: Gregory, W., Bushuk, M., Zhang, Y., Adcroft, A., and Zanna, L.: Towards improving numerical sea ice predictions with data assimilation and machine learning, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-11288, https://doi.org/10.5194/egusphere-egu24-11288, 2024.

EGU24-11413 | Posters virtual | CR3.3

Sea ice strength in SI3 

Emma Fiedler, Ed Blockley, Clement Rousset, and Martin Vancoppenolle

The NEMO sea ice model, SI3, includes the simple formulation of Hibler (1979; H79) to parameterise the compressive strength of sea ice. This assumes that thick and compact sea ice has more strength than thin and low concentration sea ice. However, the H79 strength scheme does not consider physical assumptions around energy conservation. The strength scheme of Rothrock (1975; R75) is based on the amount of potential energy gained and frictional energy dissipated during ridging, and has been introduced to SI3. Additionally, the option for a negative exponential redistribution of ridged ice among thickness categories, to better approximate observations and improve stability compared to the existing uniform redistribution when using R75, has been included. The R75 strength formulation is stable and works well in SI3 at version 4.2 with an EVP rheology, under a Met Office forced NEMO/SI3 model configuration. Sea ice strength is generally reduced for the R75 scheme compared to H79. The most notable effect on the model output is a greater number of, and sharper, features in the resulting modelled ice field when using the R75 scheme compared to the H79 scheme, which are particularly apparent in the ice thickness field. An increase in the model effective resolution is therefore demonstrated.

How to cite: Fiedler, E., Blockley, E., Rousset, C., and Vancoppenolle, M.: Sea ice strength in SI3, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-11413, https://doi.org/10.5194/egusphere-egu24-11413, 2024.

EGU24-11908 | ECS | Orals | CR3.3

A data-driven sea-ice model with generative deep learning 

Tobias Sebastian Finn, Charlotte Durand, Flavia Porro, Alban Farchi, Marc Bocquet, Yumeng Chen, and Alberto Carrassi

The current generation of sea-ice models with Brittle rheologies can represent the observed temporal and spatial scaling of the sea-ice dynamics at resolutions of around 10 km. However, running those models is expensive, which can prohibit their use in coupled Earth system models. The promising results of neural networks for the fast prediction of the sea-ice extent or sea-ice thickness offer an opportunity to remedy this shortcoming. Here, we present the development of a data-driven sea-ice model based on generative deep learning that predicts together the sea-ice velocities, concentration, thickness, and damage. Trained with more than twenty years of simulation data from neXtSIM, the model can extrapolate to previously unseen conditions, thereby exceeding the performance of baseline models.

Relying on deterministic data-driven models can lead to overly smoothed predictions, caused by a loss of small-scale information. This is why the ability to perform stochastic predictions can be instrumental to the success of data-driven sea-ice models. To generate stochastic predictions with neural networks, we employ denoising diffusion models. We show that they can predict the uncertainty that remains unexplained by deterministic models. Furthermore, diffusion models can recover the information at all scales. This resolves the issues with the smoothing effects and results in sharp predictions even for longer horizons. Therefore, we see a huge potential of generative deep learning for sea-ice modelling, which can pave the way towards the use of data-driven models within coupled Earth system models.

How to cite: Finn, T. S., Durand, C., Porro, F., Farchi, A., Bocquet, M., Chen, Y., and Carrassi, A.: A data-driven sea-ice model with generative deep learning, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-11908, https://doi.org/10.5194/egusphere-egu24-11908, 2024.

EGU24-12451 | ECS | Posters on site | CR3.3

Development of ship navigation risk indicator in sea ice-infested water 

Xinfang Zhang

There's increasing transpolar shipping in both the Arctic and Antarctic as a result of the reduction of sea ice and the desire from social economics.  Sea ice is a hazard for shipping in ice-infested water, Ship navigability in ice-covered sea depends on sea ice concentration, ice thickness, fraction of pressure ridges, and multi-year ice as well as ice speed and compression, it also depends on the vessel ice class. IMO introduced Risk Index Outcome(RIO) to provide guidelines for safe navigation, calculation of RIO requires accurate sea ice information including sea ice concentration and thickness. We developed a method similar to RIO to calculate navigation risk indicators using forecasting models including ECMWF S2S data, Copernicus data, and DMI data. Other than conventional sea ice parameters sea ice concentration and sea ice thickness, ice salinity, and ice age are also taken into account in risk indicator calculation. We select the time March 2019 -Oct 2020 and adopt the initial condition of the model forecast for sea ice to demonstrate the capabilities of seasonal forecasting of this navigation risk indicator in different models. In future, the calculation method will be implemented within the ClimateDT environment.

How to cite: Zhang, X.: Development of ship navigation risk indicator in sea ice-infested water, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-12451, https://doi.org/10.5194/egusphere-egu24-12451, 2024.

EGU24-12804 | Posters on site | CR3.3

 A laboratory model of fragmentation of a 2D membrane by waves. Analogies and differences with sea ice. 

Michael Berhanu, Louis Saddier, Mathéo Aksil, Palotai Ambre, and Michel Tsamados

The marginal ice zone is the transition region between the dense floating ice pack and the open ocean. In this zone, the interaction of surface waves with sea ice is highly complex. The sea ice is broken up into fragments, the floes, which can split into smaller parts and drift under the action of waves and underwater current. Although the downscaling is challenging, laboratory model experiments can contribute to a better understanding of this process coupling fluid and solid mechanics on a large range of time and space scales. We propose to study the fragmentation of a floating membrane, made up of 10 µm graphite particles arranged in a monolayer, by gravity surface waves with a wavelength of around 15 cm [1]. For a sufficiently strong wave amplitude, the raft progressively breaks up, developing cracks and producing fragments whose sizes decrease over a time scale that is long relative to the wave period. We then study the distribution of the fragments produced during the fragmentation process. The visual appearance of the size-distributed fragments surrounded by open water bears a striking resemblance to the floes produced by the fracturing of sea ice by waves. The fragmentation concepts and morphological tools developed for sea ice floes can be applied to our macroscopic analog. Although the mechanics of the two systems differ in their physical properties and in the fracture process, our experiment provides a model laboratory system for studying the fragmentation of floating 2D materials

 

[1] Saddier, L., Palotai, A., Aksil, M., Tsamados, M., & Berhanu, M. (2023). Breaking of a floating particle raft by water waves. In arXiv preprint arXiv:2310.16188.

How to cite: Berhanu, M., Saddier, L., Aksil, M., Ambre, P., and Tsamados, M.:  A laboratory model of fragmentation of a 2D membrane by waves. Analogies and differences with sea ice., EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-12804, https://doi.org/10.5194/egusphere-egu24-12804, 2024.

EGU24-12912 | Orals | CR3.3

Improvements in September Arctic sea ice predictions via assimilation of summer CryoSat-2 sea ice thickness observations 

Yong-Fei Zhang, Mitch Bushuk, Michael Winton, Bill Hurlin, William Gregory, Jack Landy, and Liwei Jia

Because of a spring predictability barrier, the seasonal forecast skill of Arctic summer sea ice is limited by the availability of melt-season sea ice thickness (SIT) observations. The first year-round SIT observations, retrieved from CryoSat-2 from 2011 to 2020, are assimilated into the GFDL ocean–sea ice model. The model's SIT anomaly field is brought into significantly better agreement with the observations, particularly in the Central Arctic. Although the short observational period makes forecast assessment challenging, we find that the addition of May–August SIT assimilation improves September local sea ice concentration (SIC) and extent forecasts similarly to SIC-only assimilation. Although most regional forecasts are improved by SIT assimilation, the Chukchi Sea forecasts are degraded. This degradation is likely due to the introduction of negative correlations between September SIC and earlier SIT introduced by SIT assimilation, contrary to the increased correlations found in other regions.

How to cite: Zhang, Y.-F., Bushuk, M., Winton, M., Hurlin, B., Gregory, W., Landy, J., and Jia, L.: Improvements in September Arctic sea ice predictions via assimilation of summer CryoSat-2 sea ice thickness observations, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-12912, https://doi.org/10.5194/egusphere-egu24-12912, 2024.

EGU24-13528 | ECS | Orals | CR3.3

A MAGICC Arctic Sea Ice Emulator 

Sian Chilcott, Malte Meinshausen, and Dirk Notz

CMIP6 models present our best understanding of the Earth system, yet they currently fail to simulate a plausible evolution of sea ice area to changes in the global-mean temperature. We aim to assess whether correcting the temperature and Arctic Amplification biases between CMIP6 models and observations can simulate a sensitivity of sea ice loss to global warming that is within the plausible range. To do this, we develop an emulator that is calibrated to physically-based CMIP6 models and then constrained to observations. Such a tool efficiently translates the global-mean temperature of a specific year into a physically-based and observationally constrained probabilistic ensemble of SIA in each month. This setup allows our emulator to capture the core physical processes of CMIP6 projections, while capturing the observed sensitivity of sea ice loss to global warming through the observational constraint of Arctic Amplification. While there are many application possibilities of our emulator, we use our model here to probabilistically diagnose the timing of an ice-free Arctic Ocean. We find that under a high (SSP5-8.5), medium (SSP2-4.5) and low (SSP1-2.6) emission scenario, an ice-free September Ocean is ‘likely’ at 1.73 of global warming above the pre-industrial level, however we note that the probability in the lower emission scenario reduces to ‘unlikely’ in the late 21st century as the global temperature partially recovers. Our projections suggest that the probability of an ice-free summer ocean rises rapidly from ‘unlikely’ at 1.5 of global warming to ‘likely’ at 2 of global warming, stressing the importance of preventing global temperatures rising above 1.5, as the probability of losing sea ice coverage in September rises sharply thereafter. For March, we also find that the observational constraints increase the probability of an ice-free ocean under SSP5-8.5, becoming ‘likely’ in early 2200, while the probability remains very low under SSP2-4.5 and SSP1-2.6 as less than 5% of models reach ice-free conditions. Our projections suggest an ice-free summer ocean could occur at 0.5 cooler levels than the CMIP6 multi-model ensemble mean implies. Likewise, our approach suggests the probability of an ice-free Arctic Ocean year-round is increased when constraining the Arctic Amplification to observations.

How to cite: Chilcott, S., Meinshausen, M., and Notz, D.: A MAGICC Arctic Sea Ice Emulator, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-13528, https://doi.org/10.5194/egusphere-egu24-13528, 2024.

EGU24-15156 | Posters on site | CR3.3

Calibration of a  hybrid sea-ice model based on particle and continuums mechanics 

Carolin Mehlmann and Thomas Richter

Presently, climate models employ a continuum approach to describe sea ice. This approach assumes that statistical averages can be derived from a large number of ice floes. However, employing continuum rheological models at or below the scale of individual floes is only valid if the failure mode of a single floe aligns with that of an aggregate of floes. Initially, continuum models were designed for a grid resolution of 100 km. With recent advancements in computing power, sea-ice models are frequently operated at higher mesh resolutions, potentially leading to grid cells that no longer contains a representative sample of sea-ice floes.

We are addressing these shortcomings of current continuum sea-ice models by developing a hybrid model. The idea of the hybrid approach is to nest a particle model into a continuum sea-ice model in order to predict sea ice on fine spatial scales in a region of interest. An important component of particle models is a drag law to describe the influence of ocean and atmospheric currents on the floes. Measurements obtained onboard the Polarstern expedition PS 138 have shown that the correlation cannot be described fully locally, in regions with strongly heterogenous ice cover. Instead, larger surrounding flows have a substantial effect on the motion of small ones. Detailed numerical simulations of idealised test cases do confirm these findings.   

How to cite: Mehlmann, C. and Richter, T.: Calibration of a  hybrid sea-ice model based on particle and continuums mechanics, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-15156, https://doi.org/10.5194/egusphere-egu24-15156, 2024.

EGU24-15487 | Posters on site | CR3.3

Extended seasonal forecast of Antarctic Sea Ice using ANTSIC-UNet 

Ziying Yang, Jiping Liu, and Rune Grand Graversen

Antarctic sea-ice variability affects the ocean and atmosphere both locally through thermodynamic processes and beyond the Antarctic regions remotely through dynamic processes, which may all change due to global warming. In this study, we develop the ANTSIC-UNet, a deep-learning model trained on physically enriched climate variables, to predict the extended seasonal Antarctic sea ice concentration of up to 6 months in advance. We assess the predictive skill of ANTSIC-UNet as regards linear trend prediction and anomaly persistence prediction in the Pan- and regional Antarctic areas using comparative analyses with two baseline models. Our results exhibit superior performance of ANTSIC-UNet for the extended seasonal Antarctic forecast. The predictive skill of ANTSIC-UNet is notably season-dependent, showing distinct variations across regions. Optimal prediction accuracy is found in winter, while diminished skill found during the summer can be largely attributed to the ice-edge error. High predictive skills are found in the Weddell Sea throughout the year, which suggests that regional Antarctic sea-ice predictions beyond 6 months are possible. We further quantify variable importance through a post-hoc interpretation method which indicates that ANTSIC-UNet has learned the relationships between SIC and other climate variables and the method therefore provides information on the physics of the model. At short lead times, on timescales of up to two months, ANTSIC-UNet predictions exhibit heightened sensitivity to sea surface temperature, radiation conditions and vertical atmospheric circulation conditions in addition to the sea-ice itself. At longer lead times, predictions are dependent on stratospheric circulation patterns at 7-8 months lead in addition to sea-ice. Furthermore, we discuss the potential of implementing physical constraints to enhance sea-ice-edge predictability.

How to cite: Yang, Z., Liu, J., and Grand Graversen, R.: Extended seasonal forecast of Antarctic Sea Ice using ANTSIC-UNet, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-15487, https://doi.org/10.5194/egusphere-egu24-15487, 2024.

EGU24-18506 | ECS | Orals | CR3.3

Direct Numerical Simulation of shear turbulence interacting with a melting-freezing ice layer 

Diego Perissutti, Francesco Zonta, Alessio Roccon, Cristian Marchioli, and Alfredo Soldati

When a turbulent flow of water interacts with an ice boundary at near-freezing temperature, the fluid can undergo freezing or melting, depending on the local temperature. The turbulence structures that develop in proximity to the ice layer can affect the convective heat transport patterns, leading to the formation of complex phase-boundary morphologies. The ice layer evolves as part of the solution and modifies the near-boundary fluid structures, resulting in heat transfer perturbations. We investigate these ice-water interactions at small scales by performing Direct Numerical Simulations of an open channel flow at shear Reynolds number in the range between 10^2 and 10^3. The upper section of the channel is occupied by ice, while free shear conditions are applied at the bottom. Temperature is imposed on both walls. The ice melting/freezing is simulated using a phase field method [1] combined with a volume penalization immersed boundary method. A pseudo-spectral scheme [2] is used to solve the equations for momentum and energy transport and for phase evolution. We investigated how the behavior of the system changes with the flow conditions (i.e. Reynolds number), with a specific focus on characterizing the features of the ice morphology. In particular, we observed a remarkable influence of turbulence intensity on the ice morphology: at low shear Reynolds, the typical streamwise-oriented canyons already reported in similar studies [3] are present. However, at higher shear Reynolds, spanwise instabilities are triggered, making the final ice morphology more complex.

FIgure1: Render view from below of the open channel flow at a low Reynolds number. On the top section of the channel, the corrugated ice layer is shown. On the ice boundary, the normalized heat flux passing through it is displayed (high heat flux is shown in red, low heat flux in blue). The local temperature field is reported on the side domain boundaries. The typical streamwise-oriented canyons at the ice interface are visible and the heat flux correlates well with those patterns (the heat flux is higher inside the canyons).

[1]R. Yang et al., Morphology evolution of a melting solid layer above its melt heated from below, Journal of Fluid Mechanics, 956, A23, 2023.

[2]F Zonta et al., Nusselt number and friction factor in thermally stratified turbulent channel flow under non-Oberbeck–Boussinesq conditions, International journal of heat and fluid flow, 44:489–494, 2013.

[3]L. A. Couston et al., Topography generation by melting and freezing in a turbulent shear flow, Journal of Fluid Mechanics, 911, A44, 2021.

How to cite: Perissutti, D., Zonta, F., Roccon, A., Marchioli, C., and Soldati, A.: Direct Numerical Simulation of shear turbulence interacting with a melting-freezing ice layer, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-18506, https://doi.org/10.5194/egusphere-egu24-18506, 2024.

EGU24-19333 | Orals | CR3.3

Recent progress in nesting a DEM- based regional sea ice model within a continuum model 

Wenjun Lu, Andrei Tsarau, Yuan Zhang, Raed Lubbad, and Sveinung Løset

Understanding sea-ice dynamics at the floe scale is crucial to improve regional ice forecast and comprehend the polar climate systems. Continuum models are commonly used to simulate large-scale sea-ice dynamics. However, they have both theoretical and computational limitations in accurately representing sea-ice behaviour at small scales. Discrete Element Models (DEMs), on the other hand, are well-suited for modelling the behaviour of individual ice floes but face limitations due to computational constraints. To address the limitations of both approaches while combining their strengths, we explored the feasibility of nesting a DEM within a continuum model. This paper reports recent progresses in addressing two challenges associated with this method: 1) how to couple a discrete element method (DEM) – based model (a Lagrangian model explicitly tracking each element in space) into a continuum model (a Eulerian model with fixed spatial mesh transferring state variables within); 2) how to explicitly model fracture of sea ice at large scales. Based on our assessment, integrating DEM and continuum model simulations showed potential for offering accurate, high-resolution predictions of sea ice, particularly in coastal areas and near islands. Simulating fracture of sea ice still poses great computational challenges. However, we see a potential in a data-driven approach to accelerate the computational efficiency in resolving floe-scale ice fractures.  

How to cite: Lu, W., Tsarau, A., Zhang, Y., Lubbad, R., and Løset, S.: Recent progress in nesting a DEM- based regional sea ice model within a continuum model, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-19333, https://doi.org/10.5194/egusphere-egu24-19333, 2024.

EGU24-22361 | Posters on site | CR3.3

neXtSIM-DG – A next-generation discontinuous Galerkin sea ice model 

Einar Ólason, Timothy Spain, and Thomas Richter and the The neXtSIM team

We present neXtSIM-DG, the novel sea ice model created as part of the Scale Aware Sea Ice Project (SASIP). NeXtSIM-DG is a continuum sea ice model that combines several new model paradigms at once: besides established rheologies, we use the newly developed Brittle Bingham–Maxwell rheology. The discretization is based on higher-order continuous and discontinuous finite elements. We take advantage of the object orientation of the C++ implementation of the model to create a flexible, maintainable, and easily modifiable code base ready for adaptation and adaptation by the user. Finally, the C++ implementation uses modern data structures that allow for efficient shared-memory parallelization and are ready for GPU acceleration. These aspects reflect better the different scales of sea ice dynamics in space and time. In this poster, we review the basic modelling features and present some details of numerical realization. In particular, we study the effect of high-order discretization and the role of different rheologies. 

How to cite: Ólason, E., Spain, T., and Richter, T. and the The neXtSIM team: neXtSIM-DG – A next-generation discontinuous Galerkin sea ice model, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-22361, https://doi.org/10.5194/egusphere-egu24-22361, 2024.

NP2 – Dynamical Systems Approaches to Problems in the Geosciences

EGU24-175 | ECS | Orals | CL2.4

Why do oceanic nonlinearities play a weak role in Extreme El Niño events? 

Fangyu Liu, Jérôme Vialard, Alexey V. Fedorov, Christian Éthé, Renaud Person, and Matthieu Lengaigne

Extreme El Niño events exhibit outsized impacts worldwide and considerably enhance the El Niño Southern Oscillation (ENSO) warm/cold phase asymmetries. While many mechanisms were proposed, no consensus has been reached and the relative role of atmospheric and oceanic processes remains to be illustrated. Here we quantitatively assess the contribution of oceanic nonlinearities through a state-of-the-art oceanic general circulation model, which realistically simulates extreme El Niño related characteristics and the oceanic nonlinear processes responsible for ENSO skewness. An effective way is developed to isolate sea surface temperature (SST) nonlinear response based on paired experiments forced with opposite wind stress anomalies. We demonstrate that the overall oceanic nonlinearities play a marginal role on extreme El Niño amplitude, which largely arises from the compensation between positive contributors from tropical instability waves (TIWs) and nonlinear dynamic heating (NDH) and negative contributors from subsurface processes and air-sea fluxes. The physical processes keep robust when using the other mixing scheme or mixed layer option for the heat budget. Our findings quantitively reveal the subtle contribution of oceanic nonlinearities, yielding strong evidence for the paramount role of atmospheric nonlinearities in shaping extreme El Niño events.

How to cite: Liu, F., Vialard, J., V. Fedorov, A., Éthé, C., Person, R., and Lengaigne, M.: Why do oceanic nonlinearities play a weak role in Extreme El Niño events?, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-175, https://doi.org/10.5194/egusphere-egu24-175, 2024.

EGU24-626 | ECS | Posters on site | CL2.4

Dynamical evolution of ENSO in a warming background: A review of recent trends & future projections 

Sreevathsa Golla, Joël Hirschi, Jennifer Mecking, Adam Blaker, Stephen Kelly, and Robert Marsh

The wide-spread implications of El Niño–Southern Oscillation (ENSO) on global and regional climate necessitates a better understanding of how the underlying interannual dynamics have changed over recent years. Year-to-year changes in ENSO impact terrestrial and marine habitats, water availability, food security and social stability (Santoso et al., 2017). With abundant evidence of a warming climate, it is imperative to understand how a large-scale climatic oscillation such as ENSO is evolving and influencing changes in large-scale atmospheric circulation patterns (Alizadeh et al., 2022; Cai et al., 2021). Furthermore, quantifying the influence of the ocean on changes in this climatic pattern is an interesting and important question to answer. Evaluating the ability of models to appropriately represent the underlying physics and dynamical changes impacting the spatiotemporal extent and the intensity of ENSO is crucial to understanding ocean-climate teleconnections and changes in climatic extremes. In this study, we review and evaluate the representation of ENSO in several high-resolution CMIP6 and HighResMIP models and forced ocean-only simulations focusing on the ability of current state-of-the-art models to represent central equatorial pacific warming and cooling. This evaluation involves looking at the development and propagation of warm temperature anomalies on surface and sub-surface levels in the equatorial Pacific and understanding the differences in simulating surface heat budget and exchange with the overlying atmosphere and the deeper ocean. Surface and sub-surface (up to 200m depth) temperature anomalies in the Niño 3.4 region were calculated from modelled data and were then compared with anomalies from observational and reanalysis temperature datasets (like EN4, ORAS5). We find good agreement in the timing and vertical structure of surface/sub-surface temperature anomalies in the forced model simulations, particularly during strong ENSO years. Moreover, the genesis of sub-surface anomalies and their further propagation to the surface was well simulated in the forced simulations. The vertical coherence of temperature anomalies was relatively more pronounced in forced ocean-only simulations than in coupled high-resolution model runs. Furthermore, we comment on the shortcomings and suggest potential improvements that can be made in the models that could improve the model’s ability to capture ENSO strength and variability.

How to cite: Golla, S., Hirschi, J., Mecking, J., Blaker, A., Kelly, S., and Marsh, R.: Dynamical evolution of ENSO in a warming background: A review of recent trends & future projections, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-626, https://doi.org/10.5194/egusphere-egu24-626, 2024.

EGU24-696 | ECS | Orals | CL2.4

Tropical SST Impacts on the Subtropical Atmospheric Circulation and Regional Precipitation 

Weiteng Qiu, Mat Collins, Adam Scaife, and Agus Santoso

The tropical Pacific Ocean hosts the Earth’s most prominent year-to-year climate fluctuation, the El Niño-Southern Oscillation (ENSO), which exerts strong impacts on remote regions of the globe through atmospheric teleconnection. In this study, we use reanalysis data and Coupled Model Intercomparison Project Phase 6 (CMIP6) historical simulations to investigate the relationship between tropical and subtropical atmospheric circulation, and the tropical SST patterns and regional precipitation.   

We find dynamical relationships between subtropical high intensity, the Hadley and Ferrel Circulation intensity, and the Eady Growth Rate from the reanalysis. A poleward shift of the maximum in Eady Growth Rate is associated with a strengthening of the descending branches of the Ferrel and Hadley Cells, with subtropical troposphere adiabatic warming and an increased intensity and poleward movement of the subtropical highs. Shifts in the poleward Eady Growth Rate are dominated by changes in vertical wind shear which, in turn, are in thermal wind balance with variations and trends in temperature. The mechanism for the intensification of the subtropical highs involves feedbacks from high-frequency transient eddies. Strong North Pacific and South Pacific Subtropical highs are associated with La-Niña conditions. We also show that the mechanisms for interannual variations are similar to those for trends in the highs.

We further analysed the performance of the coupled models in reproducing the trends (1979-2014) of the tropical zonal wind and regional precipitation. The CMIP6 historical simulations do not capture the intensification of trade winds within the Niño 4 region, and they also fail to reproduce the statistically significant precipitation trends over the Southern North America and the Amazon. However, a linear adjustment, based on ENSO teleconnections, can be applied to the coupled models to make the precipitation trends much closer to observations. The relationship between SST patterns and precipitation trends are confirmed by looking at atmosphere-only simulations. This study provides further evidence of the importance of reconciling observed and modelled SST patterns in the tropical Pacific.

How to cite: Qiu, W., Collins, M., Scaife, A., and Santoso, A.: Tropical SST Impacts on the Subtropical Atmospheric Circulation and Regional Precipitation, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-696, https://doi.org/10.5194/egusphere-egu24-696, 2024.

EGU24-1547 | ECS | Orals | CL2.4

Variable-resolution global atmospheric models are sensitive to driving SST in ENSO/IOD-Australian rainfall teleconnections 

Ying Lung Liu, Lisa Alexander, Jason Evans, and Marcus Thatcher

We have investigated the sensitivity of a global climate model to driving sea surface temperatures (SST) in simulating Australian rainfall characteristics, including the El Niño-Southern Oscillation (ENSO)- and Indian Ocean Dipole (IOD)-related rainfall variability. We employed the Conformal Cubic Atmospheric Model (CCAM), a global atmospheric model characterized by variable resolution, CCAM was forced by two SST datasets with different spatiotemporal resolutions: the 0.25° daily Optimum Interpolation Sea Surface Temperature (CCAM_OISST) version 2.1 and the 2° monthly Extended Reconstruction SSTs Version 5 (CCAM_ERSST5). A benchmarking framework was employed to appraise model performance, revealing strong agreement between the simulations and the Australian Gridded Climate Data (AGCD) in climatological rainfall spatial patterns, seasonality, and annual trends. It is noted that both simulations tend to overestimate rainfall amount, with CCAM_OISST exhibiting a larger bias.

Moreover, CCAM's performance in capturing ENSO and IOD correlations with rainfall was assessed during Austral spring (SON) using a novel hit rate metric. The findings underscore that only CCAM_OISST effectively reproduces observed SON ENSO- and IOD-rainfall correlations, achieving hit rates of 86.6% and 87.5%, respectively, in contrast to 52.7% and 41.8% for CCAM_ERSST5. Noteworthy disparities in sea surface temperatures were observed along the Australian coastline between OISST and ERSST5 (the so-called “Coastal Effect”). These disparities may be attributed to spatial interpolation errors arising from the differences in resolution between the model and driving SST. An additional experiment within CCAM, masking OISST with ERSST within a 5° proximity to the Australian continent, underscores the pronounced impact of the “Coastal Effect” on IOD-Australian rainfall simulations. Conversely, its influence on ENSO-Australian rainfall was constrained. Therefore, realistic local SSTs are important if model simulations are to reproduce realistic IOD-rainfall responses over Australia. Additionally, even though an SST product with a longer time span is preferred in simulating IOD-related variability, circumspection is warranted in the analysis of the impact of IOD on Australian rainfall when utilizing climate model output with a substantial discrepancy in spatial resolutions between the model and the driving SST. After showing CCAM’s ability to simulate ENSO- and IOD-rainfall, our future research will involve pacemaker experiments to isolate remaining climate modes and investigate their independent impact on Australian rainfall.

How to cite: Liu, Y. L., Alexander, L., Evans, J., and Thatcher, M.: Variable-resolution global atmospheric models are sensitive to driving SST in ENSO/IOD-Australian rainfall teleconnections, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-1547, https://doi.org/10.5194/egusphere-egu24-1547, 2024.

EGU24-1749 | ECS | Orals | CL2.4

Seasonality of Feedback Mechanisms Involved in Pacific Coastal Niño Events 

Daniel Rudloff, Sebastian Wahl, and Joke Lübbecke

The 2017 Pacific Coastal Niño Event was the strongest of its type. It caused torrential rainfall and devastating flooding in Peru and Ecuador and thus rapidly caught the attention of the scientific community. Multiple studies have been conducted focusing on the causes and consequences of this event. While the strong connection between SST anomalies and local rainfall, especially during boreal spring, is well established, the causes of the extreme warming are still a subject of discussion. In this study, we focus on the seasonality of the effectiveness of mechanisms and feedbacks involved in coastal Niño Events, utilising reanalysis products and historical model simulations from the Flexible Ocean and Climate Infrastructure (FOCI).

The 2017 event stands out due to its strength and timing as it occurred earlier in the year than most other events. We find that the atmospheric conditions during this time of year are very different due to the presence of atmospheric convection which modulates the SST-cloud feedback. Further, the event coincided with the season of strongest wind-driven upwelling. This combination enables a different forcing of a short but strong event. Additional model sensitivity experiments are performed for a better understanding of underlying mechanisms. We show how the same local wind stress forcing acts differently in different seasons, with its strongest impact during the months of strongest entrainment. Events forced by local heat fluxes and wind stress forcing only do not show any subsurface warming, which is found to be the main reason for their rapid decay. Even though the atmospheric response to a coastal warming varies seasonally, without any subsurface forcing, e.g., the events cannot be sustained through atmospheric feedbacks.

How to cite: Rudloff, D., Wahl, S., and Lübbecke, J.: Seasonality of Feedback Mechanisms Involved in Pacific Coastal Niño Events, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-1749, https://doi.org/10.5194/egusphere-egu24-1749, 2024.

EGU24-2133 | Orals | CL2.4

The El Niño Southern Oscillation (ENSO) recharge oscillator conceptual model : past achievements, future prospects. 

Jérôme Vialard and the CLIVAR ENSO conceptual model Working Group

The Recharge Oscillator (RO) is a simple mathematical model of the El Niño Southern Oscillation (ENSO). It is based on two ordinary differential equations that describe the evolution of eastern Pacific sea surface temperature and western Pacific oceanic heat content. These equations are based on physical principles that operate in nature: (i) the air-sea interaction loop known as the Bjerknes feedback, (ii) a delayed negative feedback arising from the slow oceanic response to near-equatorial winds, (iii) state-dependent stochastic forcing from intraseasonal wind variations known as Westerly Wind Events, and (iv) nonlinearities such as those related to deep atmospheric convection and oceanic advection. These elements can be combined in different levels of RO complexity. The RO reproduces the ENSO key properties in observations and climate models: its amplitude, dominant timescale, seasonality, warm/cold phases asymmetries, and the seasonal predictability decrease known as the “spring barrier”. We then discuss the RO in view of timely research questions. First, the RO can be extended to account for pattern ENSO diversity (with events that either peak in the central or eastern Pacific). Second, the core RO hypothesis that ENSO is governed by tropical Pacific dynamics is discussed under the perspective of research suggesting an influence from other basins. Finally, we discuss the RO relevance for studying ENSO response to climate change, and underline that accounting for diversity and better linking the RO parameters to the long term mean state are important research avenues. We end by proposing a list of ten important RO-based research problems.

How to cite: Vialard, J. and the CLIVAR ENSO conceptual model Working Group: The El Niño Southern Oscillation (ENSO) recharge oscillator conceptual model : past achievements, future prospects., EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-2133, https://doi.org/10.5194/egusphere-egu24-2133, 2024.

EGU24-2166 | Orals | CL2.4

Mechanisms of Tropical Pacific Decadal Variability 

Antonietta Capotondi and the CLIVAR Tropical Pacific Decadal Variability Working Group

Naturally-occurring variability in the Tropical Pacific at timescales in the 7-70 years range, defined here as Tropical Pacific Decadal Variability (TPDV), modulates ENSO characteristics and its global impacts, and is linked to the rate of change of the globally-averaged surface temperature. Thus, understanding TPDV is integral to robustly separate the forced climate response from internally-generated climate variability and thereby produce reliable projections of the tropical Pacific and global climate. Several oceanic mechanisms have been proposed to explain TPDV, including off-equatorial Rossby wave activity, propagation of spiciness anomalies from the subtropical to the tropical regions, and changes in the strength of the shallow upper-ocean overturning circulations, known as “Subtropical Cells”. However, uncertainties remain on the relative importance of these oceanic mechanisms. Another critical source of uncertainty concerns the nature and origin of the atmospheric forcing of those oceanic processes. Anomalous wind forcing could arise as a response to tropical Pacific sea surface temperature (SST) anomalies, be induced by Pacific extra-tropical influences or result from tropical basin interactions. This presentation critically reviews the nature and relative importance of the oceanic and atmospheric processes driving TPDV. Although uncertain, the tropical oceanic adjustment through Rossby wave activity is likely a dominant source of variability at decadal timescales. A deeper understanding of the origin of TPDV-related winds is a key priority for future research.

How to cite: Capotondi, A. and the CLIVAR Tropical Pacific Decadal Variability Working Group: Mechanisms of Tropical Pacific Decadal Variability, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-2166, https://doi.org/10.5194/egusphere-egu24-2166, 2024.

EGU24-2466 | ECS | Posters on site | CL2.4

Asymmetric Influences of ENSO Phases on the Predictability of North Pacific Sea Surface Temperature 

Zhaolu Hou, Jianping Li, and Yina Diao

The North Pacific sea surface temperature (SST) exerts profound climatic influence. El Niño-Southern Oscillation (ENSO) significantly impacts North Pacific SST, yet the influence from ENSO’s distinct phases on SST predictability remains unclear. Overcoming model limitations, this study assesses SST predictability under diverse ENSO phases using reanalysis. Quantifying predictability limits (PL), results unveil asymmetry: El Niño PL at 5.5 months, La Niña at 8.4 months, and Neutral at 5.9 months. This asymmetry mirrors contemporary multimodal prediction skills. Error growth dynamics reveal La Niña's robust signal strength with slow error growth rate, contrasting El Niño's weaker signal and faster error growth. Neutral exhibits intermediate signal strength and elevated error growth. Physically, predictability signal strength aligns with SST variability, whereas error growth rate correlates with atmospheric-ocean heating anomalies. La Niña, inducing positive heating anomalies, minimizes atmospheric noise impact, resulting in lower error growth. The results are beneficial for improving North Pacific SST predictions.

How to cite: Hou, Z., Li, J., and Diao, Y.: Asymmetric Influences of ENSO Phases on the Predictability of North Pacific Sea Surface Temperature, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-2466, https://doi.org/10.5194/egusphere-egu24-2466, 2024.

EGU24-2993 | Posters on site | CL2.4

El Niño Southern Oscillation and Tropical Basin Interaction in Idealized Worlds 

Dietmar Dommenget and David Hutchinson

In this study we discuss a set of fully coupled general circulation model simulations with idealised geometries of the tropical ocean basins and land with a focus on important characteristics of El Niño Southern Oscillation (ENSO) type of variability and tropical basin interaction. In a series of 15 simulations we first vary the zonal width of a single tropical ocean basin from 50o to 360o, while the rest of the tropical zone is set as land. Further we discuss different simplified configurations of two or three tropical ocean basins. The results show remarkable changes in ENSO characteristics as function of basin width and due to the interaction with other basins that challenge our current understanding of ENSO dynamics. A single basin ENSO has an optimal basin width of about 150o at which ENSO preferred period is the longest, the wind stress feedback is the strongest and variability is stronger than in all other basin widths, expect for the 350o basin. Tropical basin interactions substantially affect ENSO strength, periodicity, feedbacks, non-linearity, spatial scale and pattern. In experiments with two or three identical ocean basins we find highly synchronized ENSO modes that are identical between basins and far more energetic and oscillatory then the single basin modes. The results suggest that tropical basin interaction is an essential part of ENSO. The framework of these experiments can help to better understand the atmospheric dynamics of ENSO and should help to formulate an ENSO theory that incorporates tropical basin interactions as a core element.

How to cite: Dommenget, D. and Hutchinson, D.: El Niño Southern Oscillation and Tropical Basin Interaction in Idealized Worlds, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-2993, https://doi.org/10.5194/egusphere-egu24-2993, 2024.

This study investigates the delayed influence of the Indian Ocean dipole (IOD),  isolated and combined with ENSO, on the early winter North Atlantic-European (NAE) circulation.  Results reveal that a positive IOD induces a strong response in the NAE region during December, leading to a positive North Atlantic Oscillation (NAO)-like pattern. This circulation response also induces a north-south precipitation dipole and a positive temperature anomaly over Europe. The underlying physical mechanism involves a rainfall dipole response to the IOD in the Indian Ocean, persisting into early winter, which triggers a perturbation in the zonal wind within the subtropical South Asian jet (SAJET) region. This initiates a wave-train that propagates northeastward into the North Atlantic. Additionally, a positive IOD enhances transient eddy activity in the European region. Transient eddy forcing provides strong positive feedback to the NAO-like anomaly. While the ECMWF-SEAS5 seasonal hindcast system reproduces the sign of the response, its magnitude is considerably weaker. The possible reasons for this weak response are investigated. The model can reproduce the delayed rainfall dipole response to the IOD, however, the structure of the response shows some differences with the re-analysis. The zonal wind perturbation in ECMWF-SEAS5 in the SAJET region is only about half of the re-analysis magnitude. Moreover, the wave propagation into the stratosphere, as estimated by the 100h𝑃𝑎 eddy heat fluxes, plays a minor role in the re-analysis and the model.

How to cite: Kucharski, F., Raganato, A., and Abid, M. A.: The combined  impact of Indian Ocean dipole and ENSO on the North Atlantic-European circulation during early boreal winter in re-analysis and in the ECMWF-SEAS5 hindcast , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-3110, https://doi.org/10.5194/egusphere-egu24-3110, 2024.

EGU24-3728 | Orals | CL2.4 | Highlight

Super El Niño: A product of three-ocean interactions  

Chunzai Wang, Jiazhen Wang, and Hanjie Fan

El Niño, the largest climate phenomenon on Earth, profoundly influences global climate, weather, ecosystems, and human societies. Super (or extreme) El Niño, in particular, has a significant impact on climate and extreme weather events, but its formation mechanism remains unknown. This presentation utilizes observations, climate model outputs, and coupled model experiments to demonstrate that interactions among the tropical Pacific, Indian, and Atlantic Oceans contribute to the development of super El Niño. The early onset of El Niño imparts sufficient strength in the summer and fall to trigger the Atlantic Niña and Indian Ocean dipole. Subsequently, the Atlantic Niña and Indian Ocean dipole alternately generate additional westerly wind anomalies over the equatorial western-central Pacific, reinforcing El Niño through the Bjerknes feedback and leading to the emergence of super El Niño. This novel mechanism is termed the Indo-Atlantic booster. The findings emphasize super El Niño as a product of three interactions, suggesting that incorporating both the Indian and Atlantic Oceans and their teleconnections with the Pacific will significantly enhance predictions of super El Niño and climate.

How to cite: Wang, C., Wang, J., and Fan, H.: Super El Niño: A product of three-ocean interactions , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-3728, https://doi.org/10.5194/egusphere-egu24-3728, 2024.

The El Niño-Southern Oscillation (ENSO) is one of the most significant integrated interannual oscillations with coupled atmosphere-ocean processes in the tropical Pacific. Most coupled climate models are weak in depicting ENSO asymmetry over equatorial Pacific subsurface. And it is still unclear how the stand-alone ocean model contributes to this bias. In this study, we found that most ocean models from the Ocean Model Intercomparison Project (OMIP), driven by JRA55, underestimate the asymmetry of ENSO in the equatorial western Pacific subsurface. We investigated the primary factors contributing to this bias using composite analysis and diagnostics, and found that the weaker responses in upwelling and stronger responses in downwelling to westerly and easterly wind stress anomalies in the models are mainly responsible for the bias. Furthermore, the underestimation of zonal current variability over western Pacific subsurface, influenced by the gradient of mean state of sea surface height along the equatorial Pacific, leads to an opposite relationship between asymmetry and the zonal component of nonlinear dynamic heating in the western Pacific subsurface comparing to that in the eastern Pacific subsurface. Our study emphasizes the importance of accurately modeling ocean currents to capture the characteristics of ENSO nonlinearity and highlights the significance of nonlinear dynamic responses to external forcing.

How to cite: Li, J. and Yu, Y.: Underestimated ENSO Asymmetry and Zonal Currents over the Equatorial Western Pacific in OMIP2 experiments , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-4811, https://doi.org/10.5194/egusphere-egu24-4811, 2024.

EGU24-4820 | ECS | Posters on site | CL2.4

Synchronous Decadal Climate Variability in the Tropical Central Pacific and Tropical South Atlantic 

Chao Liu, Soon-Il An, Soong-Ki Kim, Malte Stuecker, Wenjun Zhang, Fei-Fei Jin, Jae-Heung Park, Leishan Jiang, Aoyun Xue, Xin Geng, Hyo-Jin Park, Young-Min Yang, and Jong-Seong Kug

The El Niño-Southern Oscillation (ENSO), the strongest interannual climate signal, has a large influence on remote sea surface temperature (SST) anomalies in all three basins. However, a missing map piece in the widespread ENSO teleconnection is the Equatorial Atlantic, where the ENSO footprint on local SST is less clear. Here, using reanalysis data and partially coupled pacemaker experiments, we show that the tropical Pacific SST anomalies, manifested as a Central Pacific (CP) ENSO-like structure, synchronize the tropical South Atlantic (40°W-10°E, 15°S-0°) SST anomalies over the last seven decades, but on a quasi-decadal (8-16 year) timescale. Such a decadal connection is most evident during the boreal spring-summer season, when the CP ENSO-like decadal SST anomalies induce a cooling of the South Atlantic SSTs through atmospheric teleconnections involving both Southern Hemisphere extratropical Rossby waves and equatorial Kelvin waves. The resulting subtropical South Atlantic low-level anticyclonic circulation and easterlies at its northern flank cause local ocean-atmosphere feedback and strengthen the Pacific-to-Atlantic teleconnections. In contrast, the concurrent tropospheric temperature teleconnection is less destructive to the above Atlantic SST response due to the weaker and more west decadal Pacific SST anomalies compared to the interannual ENSO counterpart. Pacific-driven coupled simulations reproduce key observational features fairly well, while parallel Atlantic-driven simulations show little forcing into the Pacific. Our results show that the tropical Central Pacific is an important source of decadal predictability for the tropical South Atlantic SST and the surrounding climate.

How to cite: Liu, C., An, S.-I., Kim, S.-K., Stuecker, M., Zhang, W., Jin, F.-F., Park, J.-H., Jiang, L., Xue, A., Geng, X., Park, H.-J., Yang, Y.-M., and Kug, J.-S.: Synchronous Decadal Climate Variability in the Tropical Central Pacific and Tropical South Atlantic, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-4820, https://doi.org/10.5194/egusphere-egu24-4820, 2024.

EGU24-5122 | Orals | CL2.4

The mechanism of multi-year La Niña events and their impact on spring precipitation over southern China 

Licheng Feng, Guangliang Li, and Ronghua Zhang

By diagnosing and analyzing the frequent occurrence of multi-year La Niña events in recent years, this study reveals the process and mechanism of the Southeast Pacific subsurface cold water triggering multi-year La Niña events. Revealing for the first time the propagation channels and physical processes of multi-year La Niña events triggered by subsurface cold water. In late spring and early summer, the anomalous eastward wind strengthens in the central equatorial Pacific, while abnormal wind stress divergence occurs in the eastern Pacific, which strengthens and spreads westward over time. The weak negative sea surface temperature anomaly in the eastern equatorial Pacific is accompanied by upwelling, providing a source of cold water for the surface. As the season progresses, the weakened equatorial undercurrent and the enhanced southern equatorial current cause cold water to spread westward and accumulate in the central Pacific, thereby extending upwards to expose the sea surface. The exposed cold water causes a cooling of the sea surface and triggers local sea atmosphere interactions, leading to abnormal development of sea atmosphere and ultimately forming a multi-year La Niña events. Composite analyses were performed in this study to reveal the differences in spring precipitation over southern China during multiyear La Niña events from 1901-2015. It was found that there is significantly below normal precipitation in the first boreal spring, but above normal in the second year. The differences in spring precipitation over southern China are correlative to the changes in anomalous atmospheric circulations over the northwest Pacific, which can in turn be attributed to different anomalous sea surface temperatures (SSTs) over the tropical Pacific. During multiyear La Niña events, anomalous SSTs were stronger in the first spring than those in the second spring. As a result, the intensity of abnormal cyclones (WNPC) in the western North Pacific Ocean (WNP) in the first year is stronger, which is more likely to reduce moisture transport, leading to prolonged precipitation deficits over southern China. In contrast, the tropical SST signal is too weak to induce appreciable changes in the WNPC and precipitation over South China in the second year. The difference in SST signals in two consecutive springs leads to different spatial patterns of precipitation in southern China by causing different WNPC.

How to cite: Feng, L., Li, G., and Zhang, R.: The mechanism of multi-year La Niña events and their impact on spring precipitation over southern China, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-5122, https://doi.org/10.5194/egusphere-egu24-5122, 2024.

Understanding external drivers of the El-Nino Southern Oscillation (ENSO) is essential for predicting its future evolution. Orbital precession has been identified as a driver of ENSO variability through both proxy records and climate model simulations, yet the exact mechanics remain unclear. This orbital cycle moderates the seasonal timing of insolation relative to Earth's revolution around the Sun, thereby adjusting the magnitude of the seasonal cycle experienced by each hemisphere. Here, we analyze output from a suite of simulations in NCAR CESM 2.1.1 designed to analyze ENSO under different precessional extremes that significantly modify the meridional temperature gradients and the cold tongue seasonal cycle in the Pacific ocean. Variations in orbital precession have a strong impact on the magnitude, periodicity, and spatial expression of tropical Pacific variability. We find a critical role for both the North and South Pacific Meridional Modes (NPMM and SPMM) in explaining changes in ENSO and decadal variability by propagating subtropical anomalies to the equatorial Pacific along with a shift in the meridional structure of equatorial winds. As an example, when the perihelion of orbit occurs during boreal winter creating a dampened (strengthened) seasonal cycle in the Northern (Southern) Hemisphere, the SPMM becomes significantly more active while the NPMM weakens. This precessional state experiences a shift toward amplified decadal variability and a greater prevalence of Eastern El Nino events in comparison with the other orbital configurations tested. Understanding the precessional control of tropical variability via subtropical pathways may help explain developments that have occurred in the past, as well as future changes which may be observed due to shifts in meridional temperature gradients.

How to cite: Persch, C. and Sanchez, S.: A Critical Role for Meridional Modes in Determining the Equatorial Pacific Response to Orbital Precession, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-6660, https://doi.org/10.5194/egusphere-egu24-6660, 2024.

The winter sea surface temperature (SST) anomalies in the Kuroshio and adjacent regions (KAR), which greatly influence the East Asian–North Pacific–North American climate, are closely related to El Niño–Southern Oscillation (ENSO). This SST relationship between the KAR and the equatorial eastern-central Pacific is widely assumed to be symmetric between El Niño and La Niña. Compared to previous studies indicating the significant and strong KAR warming during El Niño winters, this study indicates weakly negative KAR SST anomalies in the composite analysis for all La Niña events. Positive winter KAR SST anomalies unexpectedly appear in approximately half of La Niña events, which counteract negative SST anomalies in the rest of La Niña events. Further analysis suggests that the impact of La Niña on KAR SST anomalies is modulated by the East Asian winter monsoon (EAWM) during early winter. The weaker-than-normal EAWM offsets the anomalous northeasterly winds in the KAR induced by La Niña and then reinforces the KAR warming through warm oceanic advection. As for strong EAWM, it enhances the northeasterly winds to the west of an anomalous Philippine Sea cyclone associated with La Niña, leading to KAR cooling with more latent heat flux loss from the ocean and anomalous cold oceanic advection. Additionally, when the EAWM is independent of ENSO and is associated with the western Pacific pattern, it also can exhibit a pronounced influence on the KAR SST anomalies via the major processes of surface latent flux and horizontal heat advection in the ocean, accompanied by a change in Kuroshio transport.

How to cite: Chen, S., Chen, J., Wang, X., and Xiao, Z.: Varying Relationship between La Nin a and SST Anomalies in the Kuroshio and Adjacent Regions during Boreal Winter: Role of the East Asian Winter Monsoon, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-7307, https://doi.org/10.5194/egusphere-egu24-7307, 2024.

EGU24-7849 | ECS | Orals | CL2.4

On the decadal changes of Atlantic-Pacific interactions and the effects of external forcing 

Soufiane Karmouche, Evgenia Galytska, Gerald A. Meehl, Jakob Runge, Katja Weigel, and Veronika Eyring

We show the results of a study investigating the predominant role of external forcing in steering Atlantic and Pacific ocean variability during the latter half of the 20th (and early 21st) century. By employing the PCMCI+ causal discovery method, we analyze reanalysis data, pacemaker simulations, and a CMIP6 pre-industrial control run. The results reveal a gradual (multi)decadal change in the interactions between major modes of Atlantic and Pacific interannual climate variability from 1950 to 2014. A sliding window analysis identifies a diminishing El Niño-Southern Oscillation (ENSO) effect on the adjacent Atlantic basin through the tropical route, coinciding with the North Atlantic trending toward and maintaining an anomalously warm state after the mid-1980s. In reanalysis, this is accompanied by the prevalence of an extra-tropical pathway connecting ENSO to the tropical Atlantic. Meanwhile, causal networks from reanalysis and pacemaker simulations indicate that increased external forcing might have contributed to strengthening ENSO’s opposite sign response to tropical Atlantic variability during the 1990s and early 21st century, where warming tropical Atlantic sea surface temperatures induced La Niña-like easterly winds in the equatorial Pacific. The analysis of the pre-industrial control run underscores that modes of natural climate variability in the Atlantic and Pacific influence each other also without anthropogenic forcing. Modulation of these interactions by the long-term states of both basins is observed. This work demonstrates the potential of causal discovery for a deeper understanding of mechanisms driving changes in regional and global climate variability.

 

Karmouche, S., Galytska, E., Meehl, G.A., Runge, J.,Weigel, K.,& Eyring,V. (2023b, in review). Changing effects of external forcing on Atlantic-Pacific interactions. EGUsphere, 2023, 1–36. https://doi.org/10.5194/egusphere-2023-1861

How to cite: Karmouche, S., Galytska, E., Meehl, G. A., Runge, J., Weigel, K., and Eyring, V.: On the decadal changes of Atlantic-Pacific interactions and the effects of external forcing, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-7849, https://doi.org/10.5194/egusphere-egu24-7849, 2024.

Processes leading to the onset and development of an El Niño event in the tropical Pacific remain elusive. Observed data and Ocean General Circulation Model (OGCM) simulations are used to reveal a well-defined pattern of sea surface temperature (SST) perturbations along the mean North Equatorial Countercurrent (NECC) pathways in association with the onset and evolution of some El Niño events. The OGCM-based sensitivity experiments are conducted to illustrate how a warm SST anomaly (SSTA) on the equator can result from a thermal forcing that is prescribed north of 10°N, similar to observed SST anomalies in December 1988. Within approximately one year, the imposed SST anomaly north of 10°N tends to be transported to the dateline region on the equator by the mean ocean circulation in the western Pacific (the low-latitude western boundary current (LLWBC) and the NECC). In due course, an upper-layer ocean warming is generated off the equator at 6-10°N and then on the equator, which acts to induce a westerly wind anomaly response; a simple statistical atmospheric wind stress model is then used to depict an expected westerly wind response. These resultant SST and surface wind perturbations can couple together over the western tropical Pacific, forming air-sea interactions and setting up a stage for El Niño onset. As such, this pathway mechanism can reasonably well explain the appearance of a warm SST anomaly on the equator in the dateline region and the corresponding development of westerly wind anomalies over the western Pacific in association with El Niño onset.

 

How to cite: Gao, C. and Zhang, R.: A Mechanism from Pathway Perspective for the Generation of a Warm SST Anomaly in the Western Equatorial Pacific, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-8442, https://doi.org/10.5194/egusphere-egu24-8442, 2024.

EGU24-8581 | ECS | Posters on site | CL2.4 | Highlight

Increased predictability of extreme El Niño from decadal interbasin interaction 

Xuan Ma, Rizhou Liang, Xiaosong Chen, Fei Xie, Jinqing Zuo, Cheng Sun, and Ruiqiang Ding

Predicting extreme El Niño–Southern Oscillation (ENSO) events remains a formidable task. Utilizing eigen microstates (EMs) of complex systems, we elucidate the interplay of two key sea surface temperature (SST) anomaly modes, the newly identified North Atlantic–west Pacific Mode (NAPAM) and discovered Victoria Mode (VM). Our findings demonstrate that a cold NAPAM phase coupled with a positive VM phase markedly elevates the probability of extreme El Niño events; NAPAM's decadal variability serves as a key modulator of extreme El Niño events' frequency. Our empirical model, capitalizing on these modes, achieves robust forecasts with a 6–8 month lead time and boasts a 0.73 correlation with the observed ENSO index in hindcasts. Notably, the model precisely forecasts the intensity of four landmark extreme El Niño episodes: 1982/1983, 1987/1988, 1997/1998, and 2015/2016. Our findings offer promising avenues for refining ENSO predictive frameworks and deepen our understanding of the key climatic drivers.

How to cite: Ma, X., Liang, R., Chen, X., Xie, F., Zuo, J., Sun, C., and Ding, R.: Increased predictability of extreme El Niño from decadal interbasin interaction, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-8581, https://doi.org/10.5194/egusphere-egu24-8581, 2024.

EGU24-9096 | ECS | Orals | CL2.4 | Highlight

Effects of Niño1+2 and Niño3.4 ENSO Events over Euro-Mediterranean Climate Variability  

Ece Yavuzsoy-Keven, Yasemin Ezber, and Omer Lutfi Sen

El Niño Southern Oscillation (ENSO) is a climate phenomenon that affects the atmospheric circulation of the Northern Hemisphere and causes short-term variability in temperature and precipitation patterns. ENSO impacts over the Euro-Mediterranean (EM) region are commonly defined by using Niño3.4 and Niño3 indices. However, some recent studies indicate that the ENSO event represented by both Niño1+2 and Niño3.4 indices (shared ENSO) is more effective over EM region climate.

In this study, we examine the response of the EM climate to ENSO events detected by Niño1+2 and Niño3.4 regions. NCEP/NCAR Reanalysis surface air temperature, precipitation, 500 hPa geopotential height, 850 hPa wind, and 300 hPa zonal wind datasets and SST-based ENSO indices from ERSSTv4 were used in the analysis for boreal winters between 1950 and 2019. For composite analysis, we separated ENSO events as El Niño and La Niña according to those observed in Niño1+2, Niño3.4, and both regions. We also tried to understand if there is any relation between ENSO and teleconnection patterns such as NAO, East Atlantic (EA), Trough Displacement Index for the Mediterranean Trough (TDI_MedT), and East Atlantic/Western Russia (EAWR) by using the cross-correlation analysis. Additionally, investigate the winter (December, January, February, DJF) ENSO’s possible lagged impacts on the teleconnection patterns in the subsequent seasons, spring (March-April-May, MAM), summer (June-July-August, JJA), and autumn (September-October-November, SON).

The major finding of this study is that the shared ENSO event is more effective over the EM climate than the ENSO events detected only by Niño1+2 or Niño3.4 indices. Further, it is also important for the predictability of the EM climate. In the shared El Niño event, the Middle East and much of North Africa tend to become colder than climatology while Europe becomes warmer. The anticyclonic wind anomaly over western Europe causes drier air in southern Europe and wetter air in northern Europe. The shared El Niño event also modulates the westerly flows at the upper troposphere. The westerly flow accelerates over high latitudes while decelerates over European mid-latitudes, causing northern Europe to be wetter and the Mediterranean Basin to be drier. The cross-correlation analysis including all SST-based ENSO indices and teleconnection indices that the EA index has a significant correlation with the Niño1+2 index across all seasons.

How to cite: Yavuzsoy-Keven, E., Ezber, Y., and Sen, O. L.: Effects of Niño1+2 and Niño3.4 ENSO Events over Euro-Mediterranean Climate Variability , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-9096, https://doi.org/10.5194/egusphere-egu24-9096, 2024.

EGU24-9334 | ECS | Orals | CL2.4

Characterizing Nonlinearities in ENSO Dynamics Using Hybrid Machine Learning Models 

Jakob Schlör, Jannik Thuemmel, Antonietta Capotondi, Matthew Newman, and Bedartha Goswami

Event-to-event differences of the El Niño Southern Oscillation (ENSO) result in different patterns of extreme climate conditions globally, which requires ENSO forecasts that accurately predict both the likelihood and the type of an event. One question regarding predictable ENSO dynamics is the extent to which they may be captured by multivariate linear dynamics and, relatedly, whether predictable nonlinearities must be accounted for or may be treated stochastically.

In this study, we combine Recurrent Neural Networks with the Linear Inverse Model (LIM) to assess the role of predictable nonlinearities and non-Markovianity in the evolution of tropical Pacific sea surface temperature anomalies. We observe that modeling nonlinearities significantly enhances the forecast accuracy, particularly in the western tropical Pacific within a 9 to 18-month lag time. Our results indicate that the asymmetry of warm and cold events is the main source of the nonlinearity. Moreover, we demonstrate that the predictability of the Hybrid-model can be reliably inferred from the theoretical skill of the LIM whereas a similar assessment is not possible in pure deep learning models.

How to cite: Schlör, J., Thuemmel, J., Capotondi, A., Newman, M., and Goswami, B.: Characterizing Nonlinearities in ENSO Dynamics Using Hybrid Machine Learning Models, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-9334, https://doi.org/10.5194/egusphere-egu24-9334, 2024.

The interannual variability of boreal summer sea surface temperature (SST) in the tropical Atlantic displays two dominant modes, the Atlantic zonal mode highlighting SST variations in the equatorial–southern tropical Atlantic (ESTA) region and the northern tropical Atlantic (NTA) mode focusing on SST fluctuations in the NTA region except in the Gulf of Guinea. Observational evidence indicates that both the boreal summer ESTA and NTA warming are accompanied by a pair of anomalous low-level anticyclones over the western tropical Pacific, and the NTA-related anticyclone is more obvious than the ESTA-related one. Both atmosphere-only and partially coupled experiments conducted with the Community Earth System Model version 1.2 support the observed NTA–Pacific teleconnection. In contrast, the ESTA-induced atmospheric circulation response is negligible over the tropical Pacific in the atmosphere-only experiments, and although the response becomes stronger in the partially coupled experiments, obvious differences still exist between the simulations and observation. The ESTA-induced atmospheric circulation response features an anomalous low-level cyclone over the western tropical Pacific in the partially coupled experiments, opposite to its observed counterpart. It is found that the ESTA warming coincides with significantly La Ni ñ a–like SST anomalies in the central–eastern equatorial Pacific,the influence of which on the tropical atmospheric circulation is opposite to that of the ESTA warming, and therefore contributes to difference between the ESTA-related simulations and observation. Moreover, the cold climatological mean SST in the ESTA region is unfavorable to enhancing the ESTA–Pacific teleconnection during boreal summer

How to cite: Ren, H.: The Impact of Tropical Atlantic SST Variability on the Tropical Atmosphere duringBoreal Summer, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-9772, https://doi.org/10.5194/egusphere-egu24-9772, 2024.

EGU24-10200 | ECS | Orals | CL2.4 | Highlight

Roles of Tropical-Pacific Interannual–Interdecadal Variability in Forming the Super Long La Niña Events 

Run Wang, Hong-Li Ren, and Minghong Liu

The super long La Niña phenomenon, which has an extremely long duration, like the recent 2020–2023 La Niña event, is less concerned than the super El Niño. In this study, we identify five super long La Niña events after 1950 and investigate roles of the 2–3-year quasi-biennial (QB) and 3–7-year low-frequency (LF) ocean–atmosphere coupled processes of El Niño–Southern Oscillation (ENSO), and the interdecadal background in forming the basin-scale prolonged negative sea surface temperature anomalies (SSTAs) during these events. We group the five events into the thermocline-driven type (the 1983–1986 and 1998–2002 events) and the wind-driven type (the 1954–1957, 1973–1976, and 2020–2023 events). The former inherited a sufficiently discharged state of equatorial upper-ocean heat content from the preceding super El Niño and dominated by the thermocline feedback, leading to a LF oceanic dynamical adjustment to support the maintenance of negative ENSO SSTAs. The latter were promoted by the relatively more important zonal advective feedback and Ekman pumping feedback and deeply affected by a strongly negative equatorial zonal wind stress background state that sourced from the strong negative phase of the Interdecadal Pacific Oscillation. Besides, the QB ENSO variability with casual contributions during these events is less important. Results show that both the LF ENSO variability and the interdecadal Pacific background could assist to the genesis of such elongated La Niñas.

How to cite: Wang, R., Ren, H.-L., and Liu, M.: Roles of Tropical-Pacific Interannual–Interdecadal Variability in Forming the Super Long La Niña Events, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-10200, https://doi.org/10.5194/egusphere-egu24-10200, 2024.

EGU24-11374 | ECS | Orals | CL2.4 | Highlight

The El Niño response to tropical volcanic eruptions and geoengineering  

Clarissa Kroll and Robert Jnglin Wills

Following tropical volcanic eruptions and in response to geoengineering efforts in climate models, the occurrence of El Niño is notably enhanced. However, the precise mechanisms leading to the preference of the El Niño state remain a subject of ongoing debate. In this study, we explore the El Niño response within the context of stratospheric aerosol injection experiments using the Community Earth System Model version 1, with the Whole Atmosphere Community Climate Model atmospheric component (CESM1 WACCM). Our investigation is centered around the Stratospheric Aerosol Geoengineering Large Ensemble Dataset encompassing three distinct scenarios: a simulation of the RCP8.5 scenario as baseline climate change scenario, a geoengineering scenario, in which surface temperature increases are completely compensated and a scenario focusing solely on the stratospheric heating derived from the geoengineering approach. Our analysis reveals that the El Niño response is primarily linked to the heating in the tropical tropopause layer and lower stratosphere, and notably, it occurs independently of tropospheric cooling effects. We explain the increased occurrence of El Niño after volcanic eruptions and simulated geoengineering interventions by a slow down of the tropical atmospheric circulation, which is caused by increases in gross moist stability due to aerosol heating in tropical tropopause layer.

How to cite: Kroll, C. and Jnglin Wills, R.: The El Niño response to tropical volcanic eruptions and geoengineering , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-11374, https://doi.org/10.5194/egusphere-egu24-11374, 2024.

EGU24-11643 | ECS | Orals | CL2.4

Dynamical systems analysis of the "El Niño Southern Oscillation" phenomenon  

Julia Mindlin, Gabriel B Mindlin, and Pedro di Nezio

Since the 1980s, when the World Meteorological Organization launched the TOGA (Tropical Ocean-Global Atmosphere Program) program, great advances have been made in understanding ENSO by studying a hierarchy of models (Dijkstra, 2005). At the most complex end of this hierarchy are the Global Climate Models (GCMs), with which simulations of the entire climate system are performed, while at the most elementary end are the simple dynamical models that involve the minimum number of modes necessary to generate the phenomenon and therefore represent the dominant physical processes. Conceptually, two different ways of understanding the irregular oscillations of ENSO are still valid: it could be either a self-sustained oscillator of a chaotic nature or a stable mode excited by atmospheric noise. 

In this work, we use methods from complex systems to revisit the ideas regarding two plausible dynamics of ENSO. We ask if the dynamics can be better represented as a self-sustained oscillator of a chaotic nature or a stable mode excited by noise. For this, we analyzed the sea surface temperatures (SSTs), one of the output variables of the simulations generated with GCMs, the most complex simulations available from the extended system. This temperature field averaged in a particular region of the eastern equatorial Pacific (Niño 3.4) gives rise to a temporal signal widely used for ENSO monitoring and as a proxy for the study of the oscillation. In order to analyze the dynamics of the system, we reconstruct the phase space from an embedding of the temporal signal. We find that three modes are enough to recover the ENSO dynamics of the extended system, in principle of infinite dimension. Our conceptual model is based on the existence of a self-sustaining oscillation with a critical slowing down in phase space; that is, the system traverses a region of phase oscillation with a critical slowing down in phase space; that is, the system traverses a region of phase space more slowly, and includes a periodic forcing that gives rise to chaotic behavior for certain values of the parameters. We validate the model with a topological and statistical analysis of the periodic orbits in the system and, in addition, we show that the complexity of the signal is better represented as a self-sustained oscillator of a chaotic nature than as a stable mode excited by noise (Wang, 2018).

Dijkstra, HA, Nonlinear Physical Oceanography, volume 28. Springer, 2nd revised edition, 2005.

Wang C., A review of ENSO theories, National Science Review, Volume 5, Issue 6, November 2018, Pages 813–825

How to cite: Mindlin, J., Mindlin, G. B., and di Nezio, P.: Dynamical systems analysis of the "El Niño Southern Oscillation" phenomenon , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-11643, https://doi.org/10.5194/egusphere-egu24-11643, 2024.

EGU24-12873 | ECS | Orals | CL2.4

The Dynamics and Propagation of Westerly Wind Bursts 

Inko Bovenzi, Minmin Fu, and Eli Tziperman

Westerly wind bursts (WWBs), a westerly anomaly in equatorial winds in the Pacific, occur before every major El Niño event, yet major aspects of their mechanism are still not fully understood. Proposed mechanisms include cyclones approaching the equator, eastern-propagating convective heating, and wind-induced surface heat exchange, which amplifies WWBs near their peaks (Fu and Tziperman, 2019). To better understand WWB dynamics, we study their composite momentum budget using reanalysis and examine the role of convective heating and other factors. We find that many WWBs are not directly explained by nearby tropical cyclones or convective precipitation. We study their momentum budget before, during, and after the peak of the event, finding different balances at each stage. A comparison of the deduced balance to that in atmospheric general circulation climate models should add confidence in their ability to simulate this important factor in El Niño's development.

How to cite: Bovenzi, I., Fu, M., and Tziperman, E.: The Dynamics and Propagation of Westerly Wind Bursts, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-12873, https://doi.org/10.5194/egusphere-egu24-12873, 2024.

EGU24-12936 | Orals | CL2.4

A Regime View of ENSO Flavors Through Clustering in CMIP6 Models 

Pradeebane Vaittinada Ayar, David Battisti, Camille Li, Martin King, Mathieu Vrac, and Jerry Tjiputra

El Niño-Southern Oscillation (ENSO) flavors in the tropical Pacific are studied from a regime perspective. Five recurring spatial patterns or regimes characterizing the diversity of ENSO are established using a clustering approach applied to the HadISST sea surface temperature (SST) anomalies. Compared to previous studies, our approach gives a monthly characterization of the diversity of the warm and cold phases of ENSO established from observations but commonly applied to models and observations. Two warm (eastern and central El Niño), two cold (basin wide and central La Niña) and a neutral reference regimes are found. Simulated SST anomalies by the models from the latest Coupled Model Intercomparison Project Phase 6 are then matched to these reference regimes. This allows for a consistent assessment of the skill of the models in reproducing the reference regimes over the historical period and the change in these regimes under the high-warming Shared Socio-economic Pathway (SSP5.8.5) scenario. Results over the historical period show that models simulate well the reference regimes with some discrepancies. Models simulate more intense and spatially extended ENSO patterns and have issues in capturing the correct regime seasonality, persistence, and transition between regimes. Some models also have difficulty simulating the frequency of regimes, the eastern El Niño regime in particular. In the future, both El Niño and central La Niña regimes are expected to be more frequent accompanied with a less frequent neutral regime. The central Pacific El Niño and La Niña regimes are projected to increase in amplitude and variability. 
Reference:
Vaittinada Ayar, P.Battisti, D. S.Li, C.King, M.Vrac, M., & Tjiputra, J. (2023). A regime view of ENSO flavors through clustering in CMIP6 modelsEarth's Future11, e2022EF003460. https://doi.org/10.1029/2022EF003460

How to cite: Vaittinada Ayar, P., Battisti, D., Li, C., King, M., Vrac, M., and Tjiputra, J.: A Regime View of ENSO Flavors Through Clustering in CMIP6 Models, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-12936, https://doi.org/10.5194/egusphere-egu24-12936, 2024.

In recent decades, a growing body of research has highlighted the intricate interplay between the El Niño-Southern Oscillation (ENSO) and various climatic patterns across multiple ocean basins. Several studies have highlighted the significance of the South Atlantic Subtropical Dipole (SASD) and its association with ENSO.

This investigation examines the interaction between SASD and ENSO, focusing on the critical role of the South Pacific High in these dynamics. Our study proposes that the onset of the South American Monsoon (SAM) plays a crucial role in this connection, challenging the traditional perception of land's passive role in tropical interbasin interactions.

We identified two eastern Pacific and two central Pacific ENSO precursors from SAM onset period using ERA5 reanalysis data along with 1200-year CESM2 PI run. Applying partial linear regressions revealed the following patterns: initially, warm Southwestern Tropical Atlantic (SWTA) and basin-wide low pressure in the equatorial and subequatorial Atlantic, evolving into cold Southeastern Tropical Pacific (boreal spring); then, negative South Pacific Oscillation (SPO) during the following boreal summer, culminating in La Niña conditions between 12 and 15 months later (SON and DJF of the following year).

We hypothesize that anomalous upper-level divergent monsoonal circulation acts as a bridge connecting the two ocean basins. Ekman dynamics arguably transfers and amplify atmospheric signals from the SAM and SPO to the equatorial Pacific Ocean.

Random Forest and Support Vector Machines for regression analysis yielded results consistent with those from the linear model; superior skill was noted in La Niña prediction compared to under-predicted El Niño events.

Moving forward, we intend to construct causal networks to disentangle the complex interplays described herein while ensuring independence from other known teleconnections; alternatively, we plan to design appropriate numerical experiments using coupled GCMs.

This study's preliminary results present exciting opportunities to enhance early ENSO prediction by considering the impact of the South American Monsoon on aligning the variability of the tropical South Atlantic and South Pacific oceans.

How to cite: Bellacanzone, F. and Bordoni, S.: Enhancing early ENSO prediction: how the South American Monsoon onset connects the South Atlantic Subtropical Dipole and the South Pacific Oscillation, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-13140, https://doi.org/10.5194/egusphere-egu24-13140, 2024.

EGU24-13513 | ECS | Posters on site | CL2.4 | Highlight

Impact of summer-persistent ENSO events on the global climate and the occurrence of extreme weather events 

Anna Schultze, Zhengyao Lu, Qiong Zhang, Minjie Zheng, and Thomas Pugh

El Niño Southern Oscillation (ENSO), the most prominent climate variability in the tropical Pacific Ocean, significantly influences global climate and weather patterns, impacting ecosystems and societies worldwide. Our study focuses on the underexplored aspect of summer-persistent ENSO events, their global climatic impacts, and their role in triggering extreme weather occurrence.

ENSO events follow a distinct cycle, with El Niños more tightly bound to this cycle, while some La Niñas tend to fall below the ENSO threshold during the summer and then re-intensify in the following winter, resulting in multi-year La Niña events. However, there have been cases of slower ENSO decay, where sea surface temperature anomalies (SSTA) exceeding the ENSO threshold values into the northern-hemisphere summer, have been observed. The 2018/2019 El Niño, persisting until July, is a recent example, linked to significant events like the severe Australian bushfires in 2020 and the longest heatwave in history in the North Pacific in 2019. The El Niño was followed by a triple-dip La Niña, linked to extreme weather events in Africa, Australia and the United States. This highlights the importance of understanding the summer-persistent ENSO events.

Our study is structured based on three aims: identifying past summer-persistent ENSO events, assessing their impacts on global temperature and precipitation patterns, and examining their linkage to extreme weather events. Utilizing the Oceanic Niño Index calculated from the extended reconstructed sea surface temperature (ERSSTv5), we categorised ENSO events into conventional, summer-persistent, and multi-year summer-persistent types. The latter two were defined by events in which the Oceanic Niño Index exceeded the ENSO threshold until June for one or two consecutives summer seasons, respectively. We identified 12 summer-persistent ENSO events since 1940, separated into four summer-persistent El Niños, five summer-persistent La Niñas, and three multi-year summer-persistent La Niñas. Analyzing ERA5 reanalysis composites of 2-m temperature and precipitation, we compared the climatic impacts of these ENSO variants across winter and summer. This study advances our understanding of the climatic consequences of summer-persistent ENSO events, providing insights crucial for developing mitigation strategies for their impacts on global climate and extreme weather occurrences.

How to cite: Schultze, A., Lu, Z., Zhang, Q., Zheng, M., and Pugh, T.: Impact of summer-persistent ENSO events on the global climate and the occurrence of extreme weather events, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-13513, https://doi.org/10.5194/egusphere-egu24-13513, 2024.

The El Niño Southern Oscillation (ENSO) dominates tropical climate variability. While it is defined by alterations in sea surface temperatures in the eastern and central tropical Pacific, ENSO influences temperature and precipitation patterns across the globe through a network of atmospheric and oceanic teleconnections. Whether ENSO is controlled or responds to external climate factors has long remained elusive, in large part due to the lack of paleoclimate evidence of tropical variability during different climate states. Here we utilize the geochemical signatures of planktic foraminifera to reconstruct eastern and central tropical variability during the last glacial maximum (LGM), some 20-25,000 years ago. Climate conditions during the LGM were very different, featuring atmospheric CO2 concentrations, global temperatures, and sea level all substantially lower than today. However, precessional forcing, thought to be a potential control on ENSO expression, was similar to modern orbital configuration. Our reconstruction spans the central and eastern tropical Pacific during this key time frame and assesses how the patterns of variability - or ENSO ‘flavors’ - may have changed. We compare our spatial reconstructions of variability to changes in the equatorial Pacific thermocline and test hypotheses of thermocline control of ENSO. We explore the evolution of the eastern and central Pacific thermocline, and how their relationship may be an additional factor in influencing ENSO expression. Our results provide key insights into the evolution and history of tropical variability under differing background climate states, providing context for modern ENSO behavior and prediction.

How to cite: Rustic, G., Rosenheim, E., Slotter, J., and Hill, K.: Reconstructing Tropical Pacific Variability During the Last Glacial Maximum Using Individual Foraminifera: An Investigation of ENSO Flavors , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-13790, https://doi.org/10.5194/egusphere-egu24-13790, 2024.

EGU24-13992 | ECS | Orals | CL2.4

Oceans outside the tropical Pacific influence ENSO when ENSO predictability is poor 

Jemma Jeffree, Nicola Maher, Dillon Amaya, and Dietmar Dommenget

Various studies demonstrate that the El Niño Southern Oscillation is influenced by each of the Atlantic Ocean, Indian Ocean, extra-tropical Pacific Ocean and Southern Ocean. However, there is no cohesive picture of the relative importance of different ocean basins. Furthermore, even when considering only one basin, there is disagreement over the strength of it's influence on ENSO. Differences between previous studies likely arise from differences in their design. Untangling interbasin influences is non-trivial, due to  the need to distinguish between correlation and causation. Investigating these interbasin interactions is additionally complicated by model bias, and computational expense limiting the breadth of model studies.

We investigate the interbasin influences on ENSO from a new angle. We use analogue forecasting instead of initialised ensemble forecasting: we select analogues similar to some target state from a long model run (e.g. pre-industrial control or single model initial-condition large ensemble), rather than initialising from that target state. The analogue forecasts, made by following the selected analogues through time in the model run, have been previously evaluated to show similar skill to an initialised forecast. These forecasts are much faster than traditional initialised forecasts, allowing us to explore multiple models, lead times and initialisation months. We explore whether these analogue forecasts are improved by considering information from regions outside the tropical Pacific, and then infer how these regions contribute to ENSO evolution.

When ENSO forecasts are skilful, before the Spring Predictability Barrier, outside influences have little impact on ENSO forecast skill. When ENSO forecasts cross the Spring Predictability Barrier and are poor, then considering information from outside the Tropical Pacific Ocean improves forecasts. We conclude that when ENSO is in a growth phase it dominates the climate system, but in a decay phase ENSO is influenced by regions outside the tropical Pacific. This behaviour is consistent across at least two global coupled climate models, despite large variability in the way these models represent ENSO's seasonal evolution. We intend to expand this investigation to more models, and to compare the impacts of verifying forecasts against observational or model data.

How to cite: Jeffree, J., Maher, N., Amaya, D., and Dommenget, D.: Oceans outside the tropical Pacific influence ENSO when ENSO predictability is poor, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-13992, https://doi.org/10.5194/egusphere-egu24-13992, 2024.

EGU24-15294 | ECS | Orals | CL2.4

Towards a better understanding of ENSO diversity: a paleoclimate perspective 

Isma Abdelkader Di Carlo, Pascale Braconnot, Matthieu Carré, Mary Elliot, and Olivier Marti

El Niño-Southern Oscillation (ENSO) events are hard to put in one category because they differ in intensity, spatial pattern, and temporal evolution. Studies have characterized events into two main categories: central Pacific (CP) and eastern Pacific (EP) events. The indicators used to compute EP and CP events are varied, from area-averaged regions to Empirical Orthogonal Function (EOF) analysis. In the recent climatic period, they all show similar results. However, future projections show differing results when using two different methods of computing EP and CP events. Since the observational period is too short, we use paleoclimate reconstructions, which provide unique and quantitative measures of past climate changes over long time scales. We will first synthesize previous studies and discuss how they have used paleoclimate modeling and/or data to provide clues into how ENSO diversity may have been shaped in past climates. Our results indicate that many apparent inconsistencies in future projection studies are due to misleading use of ENSO diversity indicators and that investigating ENSO diversity with a climate change perspective requires assessing both changes in the climate mean state (annual mean and seasonality) and changes in variability. 

How to cite: Abdelkader Di Carlo, I., Braconnot, P., Carré, M., Elliot, M., and Marti, O.: Towards a better understanding of ENSO diversity: a paleoclimate perspective, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-15294, https://doi.org/10.5194/egusphere-egu24-15294, 2024.

EGU24-17071 | ECS | Posters on site | CL2.4

Present and future of Extreme El Niño teleconnections over North America in CMIP6 models 

Margot Beniche, Jérôme Vialard, and Matthieu Lengaigne

Previous studies did suggest a diversity of the ENSO teleconnection pattern, with an eastward shifted pattern for El Niño relative to La Niña or for “eastern Pacific” (EP) relative to “central Pacific” (CP) El Niño events. Recently, Beniche et al. (in revision) demonstrated that extreme El Niño events (i.e. the strongest EP events, such as those in 1982/83, 1997/98, and 2015/16) were the only events leading to a clear eastward shift of the winter ENSO teleconnection pattern over North America. This specific teleconnection is also associated with reproducible wet (warm) anomalies over the western USA coast (northern USA and Canada). They did however demonstrate it based on the limited observational dataset, and a single AMIP CNRM-CM6.1 ensemble.

The current study aims at evaluating the robustness of these results using the broader AMIP6 and CMIP6 datasets. The specificity of the Extreme El Niño North American winter teleconnection pattern, and its inter-event and inter-member reproducibility, are robust across 23 historical AMIP ensembles (1979-2014). These events are associated with 73% chances of warm conditions over the Northern USA and Canada and 68% chances of wet conditions over the Western US coast across the AMIP ensemble. The stronger reproducibility of the extreme El Niño teleconnections can be explained by a more favourable Signal to Noise (SNR) ratio (mainly due to stronger signal).

We further evaluate the realism of these teleconnections patterns in presence of the systematic biases that are present in CMIP6. We only select CMIP6 models that reproduce Extreme El Niño events based on the precipitation-index of Cai et al. (2014). In agreement with previous studies using CMIP5 (e.g. Bayr et al., 2019), we find that models with stronger cold climatological SST bias are unable to simulate extreme Niño3 rainfall anomaly events. CMIP6 models that reproduce extreme El Niño tropical rainfall reasonably also reproduce the specific extreme El Niño 500 hPa geopotential height and surface temperature winter teleconnection pattern over North America. They however do not reproduce well the specific wet anomalies over the west American coast associated with those events, casting doubt on the CMIP6 ability to project precipitation changes over this region. We end by discussing the relevance of these results for understanding projected changes in ENSO teleconnections over North America in the context of different Shared Socioeconomic Pathways (SSPs) scenarii.

How to cite: Beniche, M., Vialard, J., and Lengaigne, M.: Present and future of Extreme El Niño teleconnections over North America in CMIP6 models, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-17071, https://doi.org/10.5194/egusphere-egu24-17071, 2024.

EGU24-17210 | ECS | Posters on site | CL2.4 | Highlight

Crying wolf with the 2023 El Niño: a predicted event that failed to materialize? 

Sandro Carniel, Gian Luca Eusebi Borzelli, Aniello Russo, and Cosimo Enrico Carniel

The El Niño–Southern Oscillation (ENSO) is a phenomenon that involves the redistribution of heat in the tropical Pacific Ocean, resulting in irregular oscillations in the sea surface temperature (SST) between warm (El Niño) and cold (La Niña) phases, and impacting the global planetary climate. In July 2023 the World Meteorological Organization, formally responsible to declare the onset of El Niño, officially announced its onset to the media, urging governments to prepare for potential high impacts on health, ecosystems and economies. However, the analysis of long-term meteorological and oceanographic data updated to the end of 2023 shows that while the eastern Pacific was warmer than normal in the second half of the year, the overall configuration of the tropical Pacific climate system did not indicate a strong El Niño event. Our findings show that the 2023-24 El Niño event, initially predicted to be at least moderate and possibly strong, turned out to be weak and, de facto, the year closed confirming it as a weaker than expected event. Based on historical records, we hypothesize that the state of the Pacific climate system at the end of 2023, following the unusual 2023-24 El Niño, may lead to the development of a strong or very strong El Niño by mid-2024.

How to cite: Carniel, S., Eusebi Borzelli, G. L., Russo, A., and Carniel, C. E.: Crying wolf with the 2023 El Niño: a predicted event that failed to materialize?, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-17210, https://doi.org/10.5194/egusphere-egu24-17210, 2024.

EGU24-20761 | ECS | Orals | CL2.4

Visualizing the transition from LaNiño to ElNiño from NASA's model outputs 

Atousa Saberi and Gregory Shirah

The ENSO affects global weather. We used NASA GEOS Subseasonal to Seasonal (S2S) Coupled ocean-atmosphere model, NASA MERRA‐2 reanalysis, along with NOAA Niño3.4 SST anomaly index time series to visualize the transition from  LaNiño 2021 to ElNiño 2023. The visualization is a comprehensive model explainer showing changes in the top 300 meters of the Pacifc Ocean (such as thermocline flattening, movements of the temperature anomalies) coupled with the Walker Circulation and the continous coupled interaction between the ocean and the atmosphere. It's the first effort in visualizing the Walker Circulation and the moving convective branch across the Pacific without schematic plots but rather with climate model outputs.  We will also cover the effect of the two phases of ENSO on the global weather pattern. This visualization will be narrated and released to the public in the future.

How to cite: Saberi, A. and Shirah, G.: Visualizing the transition from LaNiño to ElNiño from NASA's model outputs, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-20761, https://doi.org/10.5194/egusphere-egu24-20761, 2024.

EGU24-21415 | Posters on site | CL2.4

How closely related are the Interdecadal Pacific Oscillation and El Niño-Southern Oscillation? 

Tim Cowan, Hanna Heidemann, Scott B. Power, and Benjamin J. Henley

Sea surface temperature (SST) patterns in the Pacific Ocean cause climate variability in many parts of the world. This is due to the El Niño-Southern Oscillation (ENSO) on interannual timescales and the Interdecadal Pacific Oscillation (IPO) acting on decadal to interdecadal timescales, modifying ENSO teleconnections. However, how both ENSO, ENSO diversity and the IPO interact with each other still requires further clarification. In this study, we use observations of Pacific Ocean SSTs from 1920 to 2022 to explore the statistical relationships between decadal ENSO variability and the IPO. More specifically, we show how ENSO event characteristics of both central and eastern Pacific El Niño, as well as all La Niña events varies between their occurrence in warm (positive), compared to cool (negative) phases of the IPO. We further show that up to 60% of the variability in the IPO Tripole Index can be reconstructed by using simple ENSO metrics such as the relative frequency of El Niño and La Niña events. While statistically a clear relationship between ENSO and the IPO exists, some of the IPO’s key features, especially North Pacific SSTs, cannot be explained by decadal ENSO variability.  

How to cite: Cowan, T., Heidemann, H., Power, S. B., and Henley, B. J.: How closely related are the Interdecadal Pacific Oscillation and El Niño-Southern Oscillation?, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-21415, https://doi.org/10.5194/egusphere-egu24-21415, 2024.

Monsoon rainfall and year-to-year variability play an important role in Africa’s energy, agriculture, and other societal sectors. Within the African continent, east African countries are affected much by higher degrees of variability in seasonal monsoon precipitation. Two large-scale climate drivers, the Indian Ocean Dipole (IOD) and El Niño Southern Oscillation (ENSO) are studied in this regard. A strong connection starting from a season ahead is identified for early austral summer (Oct-Nov-Dec, OND) monsoonal rain in eastern Africa.  This has been examined using various data sources, detrending data beforehand, analysing either recent or earlier time periods - covering two decades each, and using the analyses of regression. Results of compositing also suggested a strong significant anomaly in OND rain covering that region of east Africa (named here as region A:18˚S-12˚N, 25˚E-52˚E).  When IOD and ENSO are both negative in July-August-September(JAS) there is a significant deficit in OND rainfall, while an excess rain when both are positive. The Walker circulation plays a key role via altering descending and ascending branches in two circumstances. Based on this analysis, it is possible to deliver an estimation of cumulative rain in terms of median value, range and distribution, one season in advance, at a point location or average over a region. Results are further verified for recent two years of 2022 and 2023, where drivers were of same sign, either both negative (2022) or positive (2023). Classifications based on two drivers, starting from JAS, are not only modulating cumulative rain but also influencing onset dates; excess (deficit) rain and early (late) onset are associated with positive (negative) phases of both drivers. Interestingly, regions of east Africa, south of that box region show a complete reverse pattern in OND and that pattern continues till Dec-Jan-Feb. In terms of mechanisms, apart from Walker circulation, ocean also plays a key part.      

            Some results of compositing are confirmed for longer records (1940-2021) too and further classification of drivers, based on a threshold value (+0.4) is tested. In the recent year 2023, as both drivers were strongly positive in JAS, more analyses in such cases are presented.  We note, if either of the drivers is weak positive and lies in the range of 0 to +.04, the signal in region A weakens substantially on the eastern side of the box. The strongest weakening happens when both the drivers are of low magnitude in JAS (i.e.,  between 0 to +0.4). Rainfall (OND) variability of region A, at intra-decadal, decadal and multi-decadal scales are studied by applying the method of centered moving averages of 5-year, 11-year and 21-year respectively. A decreasing trend is noted in all situations and major peak/trough years are identified. For multi-decadal analyses, a shift at around 1958 is identified when the trend of OND rain is reversed and switched from increasing to decreasing. Our results have implications for future planning in optimizing energy and agricultural outputs and the livelihood of millions of east Africans will be impacted.   

How to cite: Roy, I. and Troccoli, A.: Important drivers of October to December rainfall season in eastern Africa and relevant mechanisms, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-21764, https://doi.org/10.5194/egusphere-egu24-21764, 2024.

NP3 – Scales, Scaling and Nonlinear Variability

EGU24-2344 | ECS | Orals | NP3.3

A North Atlantic Ocean-originated mode of the AMOC multicentennial variability 

Kunpeng Yang, Haijun Yang, and Mengyu Liu

A multicentennial oscillation (MCO) of the Atlantic meridional overturning circulation (AMOC) is exhibited in a CESM1 control simulation. It primarily arises from internal oceanic processes in the North Atlantic, potentially representing a North Atlantic Ocean-originated mode of AMOC multicentennial variability (MCV) in reality. Specifically, this AMOC MCO is mainly driven by salinity variation in the subpolar upper North Atlantic, which dominates local density variation. Salinity anomaly in the subpolar upper ocean is enhanced by the well-known positive salinity advection feedback that is realized through anomalous advection in the subtropical-subpolar upper ocean. Meanwhile, mean advection moves salinity anomaly in the subtropical intermediate ocean northward, weakening the subpolar upper salinity anomaly and leading to its phase change. This mechanism aligns with a theoretical model we proposed earlier. In this theoretical model, artificially deactivating either the anomalous or mean advection in the AMOC upper branch prevents it from exhibiting AMOC MCO, underscoring the indispensability of both the anomalous and mean advections in this North Atlantic Ocean-originated AMOC MCO. In our coupled model simulation, the South Atlantic and Southern Ocean do not exhibit variabilities synchronous with the AMOC MCO; the Arctic Ocean’s contribution to the subpolar upper salinity anomaly is much weaker than the North Atlantic. Hence, this North Atlantic Ocean-originated AMOC MCO is distinct from the previously proposed Southern Ocean-originated and Arctic Ocean-originated AMOC MCOs. 

How to cite: Yang, K., Yang, H., and Liu, M.: A North Atlantic Ocean-originated mode of the AMOC multicentennial variability, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-2344, https://doi.org/10.5194/egusphere-egu24-2344, 2024.

EGU24-3084 | ECS | Posters on site | NP3.3

Spatio-temporal climate fingerprint in palaeoclimate data vs models 

Vanessa Skiba, Andrew Dolman, Raphaël Hébert, Mara McPartland, and Thomas Laepple

Knowledge on natural climate variability is pivotal for making future climate projections. Previous studies demonstrated that centennial to millennial temperature variability is lacking in climate model simulations and that this bias is spatially heterogeneous. Various mechanisms have been proposed that might be important to modulate this low-frequency variability such as the ocean circulation, the meridional temperature gradient or external forcing and climate sensitivity to that forcing, but the evidence to identify the main driver(s) is still debated. Here, we provide preliminary insights on the respective importance of those mechanisms in driving long-term climate variability by investigating spatial patterns of low-frequency climate variability.

Low-frequency variability beyond multi-decadal timescales cannot be studied using only instrumental data due to data limitations and the confounding impact of anthropogenic forcing. Consequently, noisy and biased palaeoclimate proxy observations have to be utilised in order to investigate spatio-temporal patterns of climate change. Using a multi-archive and -proxy approach, we characterise the first-order spatial pattern of low-frequency climate variability of interglacial periods. By combining information on the spatio-temporal fingerprint derived from various archives and proxies with different characteristics, we aim to identify the common climate variability signal and assess the ability of climate models to explain the proxy-based spatial pattern of low-frequency variability.

How to cite: Skiba, V., Dolman, A., Hébert, R., McPartland, M., and Laepple, T.: Spatio-temporal climate fingerprint in palaeoclimate data vs models, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-3084, https://doi.org/10.5194/egusphere-egu24-3084, 2024.

EGU24-3157 | ECS | Posters on site | NP3.3

Power law error growth in a more realistic atmospheric Lorenz system with three spatiotemporal scales 

Hynek Bednar and Holger Kantz

Inspired by the Lorenz (2005) system, we mimic an atmospheric variable in one dimension, which can be decomposed into three spatiotemporal scales. This is motivated by and consistent with scale phenomena in the atmosphere. When studying the initial error growth in this system, it turns out that small scale phenomena, which contribute little to the forecast product, significantly affect the ability to predict this product. In other words, a more precise knowledge of the initial condition does not translate into a longer closeness of the forecast to the truth. Lorenz gave a sketch of such error growth. After a fast growth of the small scale errors with saturation at these very same small scales, the large scale errors continue to grow at a slower rate until even these saturate. We will present that scale dependent error growth can be translated into power law error growth. We will explain how parameter values of the power law are related to the error growth properties of the individual scales. We apply the results to the initial error growth of numerical weather prediction systems and show that the validity of the power law would imply a finite prediction horizon.

How to cite: Bednar, H. and Kantz, H.: Power law error growth in a more realistic atmospheric Lorenz system with three spatiotemporal scales, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-3157, https://doi.org/10.5194/egusphere-egu24-3157, 2024.

EGU24-4074 | ECS | Posters on site | NP3.3

A model-data comparison of North Atlantic climate variability and its responses to natural forcing over the last millennium  

Qin Tao, Jesper Sjolte, and Raimund Muscheler

North Atlantic climate variability is to a large extent governed by the recurring modes of atmospheric circulation, also exhibiting impacts of volcanic and solar activities. These factors emphasize the importance of evaluating the leading variability modes and their responses to natural forcing in climate models for assessing the North Atlantic-European climate predictions. The recent availability of spatial field reconstructions of atmospheric circulation over the last millennium offers a unique opportunity for the paleo-evaluation of CMIP-PMIP models for these purposes across annual to centennial timescales. Particularly, with the possibility of comparing the spatial structure of variability.


In this study, we perform a model-data comparison of the North Atlantic climate focusing on the leading variability modes (North Atlantic Oscillation, NAO; East Atlantic Pattern, EA; Scandinavian Pattern, SCA) and the imprints of major natural forcing over the last millennium. We first develop an updated version of climate field reconstructions covering the past 700 years by assimilating proxy records into isotope-enabled simulations. This new version shows improved skills in reproducing the leading variability modes to serve as a reference for the comparisons with the past1000 runs. We then evaluate the multidecadal spatial variability in winter modes from the last millennium to the end of the 21st century. The models generally have a good representation of the average spatial structures of the NAO, EA and SCA patterns, but with persistent biases in their spatial variability. Particularly, the underestimated spatial shift in the NAO centres of action is directly related to the biases in regional temperature and precipitation changes. Furthermore, we examine the volcanic and solar imprints over the last millennium. Although not all the models can reproduce the significant NAO responses to volcanic eruptions as shown in the reconstructions, they do capture some NAO-like signals mixed with the EA and SCA patterns. Overall, our model-data comparison presents some potential uncertainties in climate projections over the North Atlantic sector, which remain challenging for the reliability of future projections. Also, this model-data comparison framework presents a pathway for future studies aiming to select the better-performing models for regional climate studies.

How to cite: Tao, Q., Sjolte, J., and Muscheler, R.: A model-data comparison of North Atlantic climate variability and its responses to natural forcing over the last millennium , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-4074, https://doi.org/10.5194/egusphere-egu24-4074, 2024.

EGU24-4158 | ECS | Orals | NP3.3

Changes in deep-water formation amplify the Earth's Equilibrium Climate Sensitivity on multi-centennial time scales 

Fernanda DI Alzira Oliveira Matos, Christian Stepanek, Gerrit Lohmann, Qiong Zhang, Katherine Elizabeth Power, Jan Streffing, and Tido Semmler

Quantifying the Earth's climate system response to changes in atmospheric carbon dioxide (CO2) concentrations is crucial for understanding the impact of greenhouse gases on the Earth's past, present, and future climate. The sensitivity of the Earth's climate to increasing CO2 levels will largely determine the environmental conditions faced by human societies, fauna, and flora in the years to come. Projected future climate conditions depend on the sensitivity of the numerical models employed. Therefore, a comprehensive understanding of model sensitivity to radiative forcing across various temporal and spatial scales is essential. Towards this goal, we employ the newly developed AWI-CM3 model, which will be used for future climate projections in CMIP7, to examine Equilibrium Climate Sensitivity (ECS) across different time scales. 

Our quasi-equilibrium simulations span 2,000 model years, subjected to atmospheric CO2 concentrations of 280, 400, 560, and 1120 ppmv. The highest concentration simulation is inspired by the CMIP6 abrupt4xCO2 protocol, designed to assess climate response to an abrupt change in radiative forcing. Notably, our simulations run much longer than the CMIP6 suggested 150-year duration. The lower concentration simulation represents the pre-industrial period (PI), while the remaining were designed to investigate the climate with CO2 concentrations similar to the current climate and with a doubling of PI levels, respectively.

The ECS derived from AWI-CM3 stands at 3.95ºC, ranking it as medium-range sensitivity compared to the CMIP6 ensemble. A key finding is that ECS increases by up to 1.5ºC when simulations are extended beyond the CMIP6 minimum runtime requirement. This change in ECS correlates to alternations in deep water formation in both the North Atlantic and Southern Oceans. Throughout the simulations, we note adjustment processes in the overall climate and multi-centennial variability in the strength of the Atlantic Meridional Overturning Circulation (AMOC) due to changes in North Atlantic Deep Water (NADW) and Antarctic Bottom Water (AABW) formation. 

The simulations also reveal a progressive weakening and shallowing of the AMOC and a strengthening of the AABW as CO2 concentrations increase. Beyond 200 years, under adjusted radiative forcing, the AMOC recovers, but the resultant circulation pattern features persistently shallower NADW and a weaker, more northward-extending AABW in the Atlantic and Pacific Oceans. Our results highlight the intricate relationship between deep water formation and Earth's equilibrium climate sensitivity. Furthermore, our findings suggest a need to reevaluate the current framework for deriving ECS in the standard CMIP6 methodology. Prolonged simulations not only enhance our understanding of the underlying mechanisms driving climate sensitivity to changing radiative forcing but also provide valuable insights into the time required for the Earth's climate to adjust to these changes.

How to cite: Oliveira Matos, F. D. A., Stepanek, C., Lohmann, G., Zhang, Q., Power, K. E., Streffing, J., and Semmler, T.: Changes in deep-water formation amplify the Earth's Equilibrium Climate Sensitivity on multi-centennial time scales, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-4158, https://doi.org/10.5194/egusphere-egu24-4158, 2024.

Multi-centennial climate variability, evident in paleoclimate proxy records and observed in both forced transient and unforced control simulations with numerous fully coupled climate models, presents a significant yet elusive phenomenon in climate dynamics. This study, utilizing a coupled climate model EC-Earth3-LR, identifies and analyzes a prominent multi-centennial climate variability with a distinct 200-year cycle in a pre-industrial (PI, with atmospheric CO2 concentration of 280 ppmv) control simulation. This oscillation originates predominately from the North Atlantic and displays a strong association with the Atlantic Meridional Overturning Circulation (AMOC). 

We pinpoint the crucial interplay between salinity advection feedback and vertical mixing in the subpolar North Atlantic as key roles in providing the continuous internal energy source to maintain this multi-centennial oscillation. The perturbation flow of mean subtropical-subpolar salinity gradients serves as positive feedback that sustain the AMOC anomaly, while the mean advection of salinity anomalies and the vertical mixing acts as negative feedback, constraining the amplitude of AMOC anomaly.
 
In warmer climate conditions, with atmospheric CO2 concentrations elevated to 400 ppmv and 560 ppmv, we observe an expected stabilization of the water column in the North Atlantic deep-water formation regions, potentially leading to a reduction in the AMOC. These conditions are simulated to assess the evolution of unforced internal multi-centennial variability under higher CO2 levels. Results show that while multi-centennial climate variability persists in these warmer climate states, oscillation amplitudes are diminished. Despite the reduced intensity, the most pronounced effects remains in the North Atlantic and the Arctic, hypothesized to be driven by AMOC fluctuations. In contrast to the PI simulation, where the Arctic and subtropical fluxes exhibit aligned power spectra peaks, the warmer climate scenarios reveal longer timescales and reduced amplitudes in multi-centennial climate variability, suggesting a climate state dependence in the subtropical mechanism. Notably, while the subtropical salinity feedback is coupled with the Arctic mechanism in the PI state, it evolves into a weaker, slower, and self-sustaining mechanism in warmer climates.

How to cite: Cao, N., Zhang, Q., and Power, K.: Multi-centennial variability of the Atlantic Meridional Overturning Circulation: underlying mechanisms and its response to elevated CO2 levels , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-4605, https://doi.org/10.5194/egusphere-egu24-4605, 2024.

Momentum-scalar coupling turbulence, such as buoyancy-driven turbulence and electrohydrodynamic (EHD) turbulence, involves the transportation of multicomponent scalars under the strong interplay of multiphysics. For instance, in the atmosphere, the temperature gradient can induce buoyancy, driving the flow to form thermal convection. At the same time, electric body force can be generated on droplets, dust, and moisture gradients through spatial electric fields, resulting in air flow into EHD turbulence. Additionally, charged species move and create electric current, leading to Lorentz force due to the magnetic field of the earth, which may induce magnetohydrodynamic (MHD) turbulence. These physical mechanisms generate the diverse phenomena on our beautiful planet. This study theoretically explores how multiphysical mechanisms interplay, governing the cascades of turbulent kinetic energy and multicomponent scalars. Some new scaling properties, which differ from those predicted in buoyancy-driven turbulence, EHD, and MHD, emerge when two mechanisms and scalar components exist simultaneously. The quad-cascade processes of such turbulent systems are again validated. Unfortunately, when three or more mechanisms are taken into account at the same time, the problem becomes unattainable to close. This research endeavors to shed light on the diverse observation in momentum-scalar coupling turbulence across various scenarios.

How to cite: Zhao, W.: Emergence of scaling properties in momentum-scalar coupling turbulence: Exploring interplay of multiphysics mechanisms, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-4818, https://doi.org/10.5194/egusphere-egu24-4818, 2024.

EGU24-5384 | Orals | NP3.3

Continuous-time state-space time series models for delta-O-18 and delta-C-13 

Eric Hillebrand, Mikkel Bennedsen, Kathrine Larsen, and Siem Jan Koopman

Time series analysis of delta-O-18 and delta-C-13 measurements from benthic foraminifera for purposes of paleoclimatology is challenging. The time series reach back tens of millions of years, they are relatively sparse in the early record and relatively dense in the later, the time stamps of the observations are not evenly spaced, and there are instances of multiple different observations at the same time stamp (Westerhold et al., 2020, Science 369 p. 1383). The time series appear non-stationary over most of the historical record with clearly visible temporary trends of varying directions. In this paper, we propose a continuous-time state-space framework to analyze the time series prepared in Westerhold et al. (2020). State space models are uniquely suited for this purpose, since they can accommodate all the challenging features mentioned above. We specify univariate models and joint bivariate models for the two time series of delta-O-18 and delta-C-13. The models are estimated using maximum likelihood by way of the Kalman filter recursions. The suite of models we consider has an interpretation as an application of the Butterworth filter (Gomez 2001 [JBES 19 p. 365], Harvey & Trimbur 2003 [REStat 85 p. 244]). We propose model specifications that take the origin of the data from different studies into account and that allow for a partition of the total period into sub-periods following Westerhold et al. 2020, which we have been able to confirm with a statistical method (Larsen et al. 2024: Estimating Breakpoints between Climate States in Paleoclimate Data, abstract submitted to EGU General Assembly Session CL3.2.3). The models can be used, for example, to generate evenly time-stamped data by way of Kalman filtering. They can also be used, in future work, to analyze the relation to proxies for CO2 concentrations.

How to cite: Hillebrand, E., Bennedsen, M., Larsen, K., and Koopman, S. J.: Continuous-time state-space time series models for delta-O-18 and delta-C-13, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-5384, https://doi.org/10.5194/egusphere-egu24-5384, 2024.

EGU24-6672 | ECS | Posters on site | NP3.3

Unified scaling framework for Holocene, Quaternary and Phanerozoic geochronology variability 

Rhisiart Davies, Shaun Lovejoy, Raphael Hebert, Fabrice Lambert, and Andrej Spiridonov

With few exceptions, paleodata are irregularly sampled; this poses numerous challenges for the statistical characterization of paleoindicators, this includes the indicators needed to understand the climate and macroevolution.  The key variable is the measurement density - the number of measurements per unit time (r(t)).  Our study used 27 paleoindicators collectively spanning time scales from years to hundreds of millions of years.

Using Haar fluctuation analysis and for all the series, we show that r(t) has two scaling regimes.  At high frequencies, there is a low intermittency (quasi-Gaussian) scaling regime (intermittency parameter C1 ≈ 0).  Over this regime, the fluctuation exponent H is negative implying that the chronologies become more uniform at longer time scales, r(t) is commonly close to a Gaussian white noise (H = -1/2).  In contrast, at low frequencies, r(t) is highly intermittent (large C1), but it also has positive H so that fluctuations tend to grow with scale but in a highly intermittent fashion.  In this this regime, “gaps” at all scales are important. 

The two regimes have simple physical interpretations: the high frequency behaviour can be explained by fairly smooth (but scaling) sedimentation rates, whereas the low frequencies can be explained by scaling erosion processes that introduce gaps over a wide range of scales (in conformity with the Sadler effect). To confirm this interpretation, we introduce a simple multiplicative sedimentation -  erosion model that is close to the data.  Finally, we empirically show that the gaps typically have extreme power law probability tails so that the series are not only scaling in time, but also in probability space.

A key issue for paleontologists is the effect of variable r(t) on the paleoindicator estimates themselves (e.g. on paleotemperatures T(t)).  Using Haar fluctuations we determined the fluctuation - fluctuation correlation R(Δt) = < Δ r(Δt) ΔTt) >.  When R(Δt) is small, the measurements and indicators are statistically independent so that the biases due to r(t) variability on paleoindicator statistics are easy to correct.  However, at large Δt, the correlations are frequently large, and this poses additional difficulties in data interpretation.  Strong correlations were observed in the Quaternary, but not the Holocene or Phanerozoic.

Our study spans more than 8 orders of magnitude in time scale and it shows that it is wrong to theorize paleoseries as being fundamentally regularly sampled but interspersed with occasional data “holes” that can be dealt with using conventional techniques such as interpolation.  While Haar fluctuation analysis is insensitive to the chronology variability - and if needed can easily be statistically corrected for any biases that it introduces -  this is not true of existing spectral estimators that are extremely sensitive to scaling data gaps. 

How to cite: Davies, R., Lovejoy, S., Hebert, R., Lambert, F., and Spiridonov, A.: Unified scaling framework for Holocene, Quaternary and Phanerozoic geochronology variability, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-6672, https://doi.org/10.5194/egusphere-egu24-6672, 2024.

EGU24-6829 | ECS | Orals | NP3.3

Tropical-polar teleconnections: Impacts of North Atlantic meltwater forcing on the Indian Ocean 

Benjamin H. Tiger, David McGee, and Caroline Ummenhofer

Across all future IPCC Shared Socioeconomic Pathways, the strength of the Atlantic Meridional Overturning Circulation (AMOC) is projected to decline. However, there is much less certainty about the impacts of AMOC decline further afield. Evidence from paleoclimate archives and simulations suggests eastern African monsoons weakened under periods of high meltwater forcing in the North Atlantic, particularly during the most recent deglaciation. To explore the dynamics of this high- to low-latitude teleconnection, we use a compilation of ~30 sea surface temperature (SST) records from the tropical Indian Ocean spanning the last 30 ka. The zonal Indian Ocean SST gradient calculated from this compilation shows a remarkable similarity with North Atlantic 231Pa/230Th records of AMOC strength, particularly during intervals of variable meltwater forcing such as the Younger Dryas, Bølling-Allerød, and Heinrich stadials. A weaker AMOC is associated with cooler western Indian Ocean and a warmer eastern Indian Ocean, suggesting a tight linkage between AMOC strength and zonal Indian Ocean variability. To better understand this teleconnection, we analyzed a meltwater single-forcing scenario from a transient simulation of the Last Glacial Maximum to present (TraCE, 22ka-0ka). Under simulated meltwater forcing events, the tropical zonal Indian Ocean SST gradient intensifies (i.e., relative cooling in the west and warming in the east), in agreement with SST paleorecords. This response stems from an intensification of the subtropical high over Southern Europe which drives northerly surface wind anomalies across Arabia and the Horn of Africa, with cooler Northern Hemisphere anomalies extending as far south as Madagascar. This cools the surface western Indian Ocean, particularly in the Arabian Sea, enhancing the Bjerknes feedback and strengthening the Walker circulation across the basin. This effect is strongest in austral summer (DJF) when the Somali Jet reverses and northerly winds advect cool northern air into the deep tropics. Anomalous northerly winds and western Indian Ocean cooling were also found to be common feature of eight hosing experiments under preindustrial boundary conditions from the North Atlantic Hosing Model Intercomparison Project (NAHosMIP). Overall, we hypothesize an atmospheric mechanism connecting the high-latitude North Atlantic and tropical Indian Ocean under meltwater forcing, with the western Indian Ocean playing an outsized role in steepening the zonal SST gradient across the basin which weakens monsoon systems in eastern Africa.

How to cite: Tiger, B. H., McGee, D., and Ummenhofer, C.: Tropical-polar teleconnections: Impacts of North Atlantic meltwater forcing on the Indian Ocean, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-6829, https://doi.org/10.5194/egusphere-egu24-6829, 2024.

Hydroclimatic systems are complex nonlinear dynamically-evolving systems, often made up of a large number of interconnected components that change both in space and in time. Therefore, any effort towards reliable modeling and forecasting of hydroclimatic systems requires proper selection of scientific concepts and methods. Many different scientific concepts and methods have been proposed in the literature and applied to numerous hydroclimatic systems, processes, and problems around the world. Among such, concepts and methods based on chaos theory, complex networks, and fractal theory have been found to offer unique and useful avenues for studying hydroclimatic systems and, thus, have been finding widespread applications in recent times. The purpose of the present study is to discuss the advances in the applications of these concepts to hydroclimatic systems and to look toward the future. This is done through: (1) presenting some key aspects of chaos theory, complex networks, and fractal theory and their relevance to hydroclimatic systems; (2) reviewing various applications of these concepts to hydroclimatic systems, processes, and problems; (3) addressing important data-related issues in the applications of these concepts to hydroclimatic systems; and (4) offering specific directions to advance these concepts and applications further, especially in the context of future grand challenges associated with hydroclimatic systems.

How to cite: Sivakumar, B.: Complexity, Connectivity, and Scale in Hydroclimatic Systems, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-7105, https://doi.org/10.5194/egusphere-egu24-7105, 2024.

EGU24-7152 | ECS | Posters on site | NP3.3

Thermohaline Circulation Determines the Multi-centennial Variability of Earth's Climate System 

Fengli An, Mingjun Tong, and Haijun Yang

Long-term proxy data have shown that there is significant multi-centennial variability in Earth's climate system. However, the causes and mechanisms of this variability are still a major scientific problem for climate scientists and archaeologists. From the middle of the Holocene until the Industrial Revolution, there was little change in the Earth's external forcing, so it is important to study the long-term natural oscillations of our climate system during this period. In our research, we designed a series of experiments using CESM1.0 to explore the sources of multi-centennial variability of climate system. In some experiments, the thermohaline circulation was turned off to see if its presence would affect the oscillation of climate system. We finally conclude that thermohaline circulation is likely to determine the multi-centennial variability of Earth's climate system.

How to cite: An, F., Tong, M., and Yang, H.: Thermohaline Circulation Determines the Multi-centennial Variability of Earth's Climate System, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-7152, https://doi.org/10.5194/egusphere-egu24-7152, 2024.

EGU24-7409 | ECS | Posters virtual | NP3.3

Fractal approach in the Analysis of climate change due to the ozone layer hole 

Meenakshi Murugan

The ozone layer acts as the planet's natural sunscreen, protecting people, plants, and animals from harmful UV-B rays. In Antarctica, British scientists discovered the hole in the ozone layer in 1985. The effects of climate change have been experienced by all living hoods through various kinds of natural calamities due to this hole. Many researchers dedicated their time to solving this problem and saving the planet. This article explores Antarctica's post-1985 climate changes.  The authors have to Investigate the time series data for the global temperature, precipitation, and Antarctica ice sheet mass balance through analysis utilizing the fractal analysis tool.  Additionally, the nonlinear dynamical data's chaotic feature is verified.

How to cite: Murugan, M.: Fractal approach in the Analysis of climate change due to the ozone layer hole, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-7409, https://doi.org/10.5194/egusphere-egu24-7409, 2024.

EGU24-8571 | ECS | Posters on site | NP3.3

Unravelling succolarity to quantify multiscale petrophysical properties beyond porosity 

Bo Li, Ciprian Panaitescu, Paul Glover, Kejian Wu, Piroska Lorinczi, and Bingsong Yu

Characterising the complexity of spatial patterns and their underlying physics using a nonlinear approach is growing in many fields. Many features in geomaterials, such as pore and fracture systems, exhibit scaling behaviour, allowing their properties to be characterised using fractal theory.

The widely used fractal dimension is a ratio that compares how the level of detail in a structure varies with its size, measuring its space-filling ability. Lacunarity, derived from the Latin word "lacuna," meaning "gap," quantifies the “voidness” of a texture. Nevertheless, neither fractal dimension nor lacunarity can characterise the percolating properties of a fractal. Mandelbrot coined the concept of succolarity. Given that "percolare" in Latin translates to "to flow through," the term "succolare" (sub-colare) aptly conveys the concept of "to nearly flow through" in neo-Latin. A succolating fractal is characterised by almost containing the connecting paths that permit percolation, i.e., one below the percolation threshold. However, it remains a less known notion than the other two fractal counterparts. In the last ten years, succolarity has evolved from an idea to a computable parameter. It has characterised many patterns in different scales and fields, such as medical objects, material surfaces, and networks from nano-micropores to rivers.

In this contribution, we aim (i) to understand the physical meaning of succolarity and how it relates to pore networks and other petrophysical properties across different scales, and (ii) to provide new approaches to succolarity calculation. We implemented the succolarity algorithm using the gliding box-counting method. We then re-examined the published datasets for validation and comparison. The succolarity for 2/3D images of rock samples and synthetic models with various porous structures was also calculated for deeper understanding. Finally, we correlated the succolarity results with porosity, permeability, and other petrophysical parameters.

Our findings reveal that (i) succolarity contains information about a structure's anisotropy, phase fraction (e.g., porosity in the case of pore space), and percolation information. (ii) It is susceptible to connectedness. As we cut out smaller pores of a structure, succolarity decreases linearly until a pore size (porosity) threshold is reached; it drops significantly and follows a power law. (iii) Succolarity (Su) and permeability are fitted to an exponential relation: k = aebSu. The computation of succolarity excludes isolated pores for a given flooding direction, allowing it to reflect the flow properties better than porosity alone. (iv) Moreover, it is worth noting that using pressure or velocity field in the succolarity calculation algorithm would endow it with a clearer physical meaning than being a proxy for porosity.

How to cite: Li, B., Panaitescu, C., Glover, P., Wu, K., Lorinczi, P., and Yu, B.: Unravelling succolarity to quantify multiscale petrophysical properties beyond porosity, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-8571, https://doi.org/10.5194/egusphere-egu24-8571, 2024.

EGU24-8588 | ECS | Orals | NP3.3

 Power-spectra of turbulent buoyant jets from laboratory measurements 

Konstantinos Gkoutis, Ilias Papakonstantis, Panagiotis Papanicolaou, and Panayiotis Dimitriadis

In total, 11 experiments of turbulent buoyant jets were carried out in an experimental apparatus, which includes a tank with dimensions 1.00 m x 0.80 m x 0.70 m. Specifically, in a stationary homogeneous ambient fluid, six (6) experiments were performed, with a temperature at the outlet significantly higher than the ambient water, and five (5) experiments with the same temperature but in an ambient saltwater environment of initial density difference between 18.4 and 19.2 kg/m3. The nozzle diameter was equal to 1.5 cm in all experiments, the densimetric Froude number was ranging between 1.72 and 3.73, and the Reynolds number ranging between 1222 and 3136. The experiments included flow visualization and concentration measurements based on the Laser Induced Fluorescence (LIF) technique using Rhodamine 6G as fluorescent tracer. A planar laser sheet was created and the experiments were recorded using a suitable video camera. The energy-spectra of the concentration were estimated using Fast Fourier Transformation and were compared to theoretical arguments, such as the K41 model.

How to cite: Gkoutis, K., Papakonstantis, I., Papanicolaou, P., and Dimitriadis, P.:  Power-spectra of turbulent buoyant jets from laboratory measurements, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-8588, https://doi.org/10.5194/egusphere-egu24-8588, 2024.

EGU24-9251 | ECS | Orals | NP3.3

Complex network-based analysis of the spatial evolution of the flash droughts over India 

Akshay Pachore and Renji Remesan

Flash droughts, characterized by a rapid decline in soil moisture, are short-term drying events that can cause significant damage to crops when they occur during the growing season. Understanding the spatial and temporal evolution properties of flash droughts is crucial for the effective mitigation and management of this extreme event. The present study employed the complex network theory to assess the spatio-temporal properties of the flash drought which was quantified using the soil moisture percentile drop (SMPD) based definition during the period from 1981 to 2020 over the entire Indian region. An event synchronization (ES)-based complex network is constructed and the spatial propagation of the flash droughts is analyzed using the unidirectional and directed complex networks-based metrics i.e., strength, direction, and distance. Initial results gave insights into how the flash drought hotspots are connected in space and their temporal evolution pattern. From the result of the distance metrics, it was observed that flash drought propagates for longer distances in the central-eastern and southern peninsular regions as compared to the rest of the regions over India. Inference gained from the present analysis can be useful for building an early warning system for flash drought in terms of onset and spatial propagation along with insights on the spatially connected flash drought vulnerable regions.           

How to cite: Pachore, A. and Remesan, R.: Complex network-based analysis of the spatial evolution of the flash droughts over India, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-9251, https://doi.org/10.5194/egusphere-egu24-9251, 2024.

EGU24-10640 | Orals | NP3.3

A synchronized solar dynamo model and its consistency with climate data 

Frank Stefani, Gerrit Horstmann, Martins Klevs, George Mamatsashvili, and Tom Weier

We examine a remarkable consistency of the power spectrum of paleoclimatic varved sediment data from Lake Lisan [1] with that of a novel solar dynamo model that is doubly synchronized by tidal effects and the revolution of the Sun around the barycenter of the solar system [2,3]. We support and specify the latter model by quantifying the tidal excitation of magneto-Rossby waves [4] at the solar tachocline and by estimating the effects of spin-orbit coupling. Typical time series resulting from this dynamo model are then utilized in a double regression of solar and anthropogenic influences on the global temperature of the past 170 years in order to quantify various climate sensitivities [5].
 

[1] S. Prasad et al., Geology 32 (2004), 581
[2] F. Stefani et al., Solar Phys. 296 (2021), 88
[3] F. Stefani et al., arXiv:2309.00666
[4] G. Horstmann et al., Astrophys. J. 944 (2023), 48 
[5] F. Stefani, Climate 9 (2021), 163

How to cite: Stefani, F., Horstmann, G., Klevs, M., Mamatsashvili, G., and Weier, T.: A synchronized solar dynamo model and its consistency with climate data, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-10640, https://doi.org/10.5194/egusphere-egu24-10640, 2024.

EGU24-10986 | Orals | NP3.3

Tropical mountain ice core  : A Goldilocks indicator for global temperature change 

Zhengyu Liu, Yuntao Bao, Lonnie Thompson, Ellen Mosley-Thompson, Clay Tabor, Guang Zhang, Mi Yan, Marcus Lofverstrom, Isabel Montanez, and Jessica Oster

 Tropical mountain ice core : A Goldilocks indicator for global temperature change   Zhengyu Liu1,2,3, , Yuntao Bao1, Lonnie G. Thompson3,4, Ellen Mosley-Thompson1,3,  Tabor Clay5, Guang J. Zhang6, Mi Yan2, Marcus Lofverstrom7, Isabel Montanez8,  Jessica Oster9  1.      Department of Geography, The Ohio State University, Columbus, OH 2.      School of Geography Science, Nanjing Normal University, Nanjing, China. 3.      Byrd Polar and Climate Research Center, The Ohio State University, Columbus, OH 4.      School of Earth Sciences, The Ohio State University, Columbus, OH 5.      Department of Earth Sciences, University of Connecticut, Storrs, CT 6.      Scripps Institute of Oceanography, University of California San Diego, San Diego, CA

  • Department of Geosciences, University of Arizona, Tucson, AZ
  • Department of Earth and Planetary Sciences, University of California–Davis, Davis, CA
  • Department of Earth and Environmental Sciences, Vanderbilt University, Nashville, TN

 

Very high tropical alpine ice coresprovide a distinct paleoclimate record for climate changes in the middle and upper troposphere. However, the climatic interpretation of a key proxy, the stable water oxygen isotopic ratio in ice cores (), remains an outstanding problem. Here, combining proxy records with climate models, modern satellite measurements and radiative-convective equilibrium theory, we show that the tropical  is an indicator of the temperature of the middle and upper troposphere, with a glacial cooling of -7.35+-1.1oC (66% CI). Moreover, it severs as a “Goldilocks-type” indicator of global mean surface temperature change, providing the first estimate of glacial stage cooling that is independent of marine proxies as -5.9+-1.2oC. Combined with all estimations available gives the maximum likelihood estimate of glacial cooling as -5.85+10.51oC .

 

 

How to cite: Liu, Z., Bao, Y., Thompson, L., Mosley-Thompson, E., Tabor, C., Zhang, G., Yan, M., Lofverstrom, M., Montanez, I., and Oster, J.: Tropical mountain ice core  : A Goldilocks indicator for global temperature change, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-10986, https://doi.org/10.5194/egusphere-egu24-10986, 2024.

EGU24-11058 | ECS | Orals | NP3.3

Analysing the multifractality and cross correlations of river stage records using detrended fluctuation principles 

Sumayah Santhoshkhan, Athira Madhu, Muraleekrishnan Bahuleyan, and Susan Mariam Rajesh

This study presents the application of Multifractal detrended fluctuation analysis (MFDFA) for analyzing the multifractal properties of river stage time series of Indian rivers. Initially, the MFDFA method was applied to detect long-range correlations and multifractal behaviour of river stage time series of 81 locations from 11 basins of Peninsular India. The study found that all the Hurst exponent (H) values are found to be more than 0.5 indicating the long-range power law correlations in the stage data of Indian rivers. The different datasets indicated strong multifractal degree and a strong association between H and Holder exponent supported by a strong correlation over 0.98. Most of the multifractal spectra (93 %) indicated a positive asymmetry showing the frequent low fluctuations and localized high fluctuations. Basin wise analysis showed the strongest long-term persistence (LTP) and highest degree of multifractality for the datasets of Cauvery basin. In order to get an insight on the multifractality, MFDFA was applied to the daily data of corresponding period as that of stage observations. The analysis indicated that unlike the case of stage data 68 % of the data showed LTP while rest of the data displayed STP in the analysis. The multifractality of stage series is more than that of stream flow series at all river basins in India. Multifractal cross correlation analysis performed between daily river stage data and discharge data of same period indicated a strong correlation (>0.8) in majority of cases for different scales, despite the absence of a definite pattern in the correlation behavior for the data of different stations. This analysis is proven to be a very essential and useful prerequisite for developing stage-discharge relationships in a multifractal perspective, which may eventually help in proper flood management of Indian basins' changing climate scenario.

Keywords: Persistence, multifractality, Stage, Streamflow, Correlation, Scale

How to cite: Santhoshkhan, S., Madhu, A., Bahuleyan, M., and Rajesh, S. M.: Analysing the multifractality and cross correlations of river stage records using detrended fluctuation principles, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-11058, https://doi.org/10.5194/egusphere-egu24-11058, 2024.

EGU24-11382 | Orals | NP3.3

Multifractal correlation of rainfall and wind fields and consequences on wind power production 

Auguste Gires, Jerry Jose, Angel Garcia-Gago, Ioulia Tchiguirinskaia, and Daniel Schertzer

Rainfall and wind exhibit extreme variability over wide range of space-time scales. Such features are naturally transferred to wind turbine torque and ultimately to wind energy production. Improving our understanding of wind power production requires better accounting for the impact of these small scale fluctuations. This is much needed in order to achieve UN’s (United Nations) Sustainable Development Goal 7 (affordable and clean energy for all) and in a context of increasing global transition towards renewable and carbon neutral energy.

The project RW-Turb (https://hmco.enpc.fr/portfolio-archive/rw-turb/; supported by the French National Research Agency, ANR-19-CE05-0022) was developed to address this challenge and to understand better the correlation across scales between rainfall and wind fields and its impact on wind power production. A high resolution measurement campaign was set up between 12/2020 and 07/2023 with two 3D sonic anemometers (manufactured by Thies), two mini meteorological stations (manufactured by Thies), and two disdrometers (Parsivel2, manufactured by OTT) installed on a meteorological mast at 75 and 45 m respectively in the wind farm of Pays d’Othe (110 km south-east of Paris, France; operated by Boralex). The framework of Universal Multifractals (UM) is used to carry out this analysis. It is a physically based and mathematically robust framework that enables to characterize and simulate the extreme variability of geophysical fields across scales. It is furthermore parsimonious since it relies on the use of only three parameters.

In a first step multifractal analysis of the available fields (wind velocity, power available at the wind farm, power produced by wind turbines, air density, and rainfall) is implemented. Event based analysis enabled to observe differences in UM parameters depending on whether it is raining or not. In general, a slightly stronger variability is found when it rains. In a second step, a joint multifractal analysis is implemented to further quantify correlation across scales between the studied fields. An increase in correlation exponent of the various fields with increase in rain rate is found.

Numerical simulations are then used as a complement to data analysis. More precisely, 3D space plus time vector fields which realistically reproduce observed spatial and temporal variability of wind fields are generated with multifractal tools. Then, they are used as input into three modeling chains of increasing complexity to simulate wind turbine torque. The simplest model uses average wind field over swept area, while a more realistic one computes the torque as an integral over the blades of the turbine enabling to account for the space-time variability of wind. Finally, OpenFAST, which is widely used by researchers and practitioners is implemented. UM analysis on the simulated torque time series were performed to quantify the impact of small scale fluctuations on wind power production, as well as the ability of the various models to account for it.

How to cite: Gires, A., Jose, J., Garcia-Gago, A., Tchiguirinskaia, I., and Schertzer, D.: Multifractal correlation of rainfall and wind fields and consequences on wind power production, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-11382, https://doi.org/10.5194/egusphere-egu24-11382, 2024.

EGU24-11718 | Orals | NP3.3 | Highlight

An ocean-atmosphere paradox, Phase 2 

Brian Durham and Christian Pfrang

Last year we posed the question:  Given Earth’s ocean-atmosphere gas equilibrium, why do measured atmospheric carbon dioxide (CO2) curves rise more steeply against solvent temperature than predicted by Henry’s Law? (https://presentations.copernicus.org/EGU23/EGU23-6069_presentation.pdf). We here develop our experimental procedure to simulate more closely the ocean-atmosphere gas exchange in the lab, seeking to better understand the relationship between atmospheric COand average sea surface temperatures, with implcations for past and future climate variability in the Earth System. 

To this end, we previously reported provisional trends when water and natural seawater samples are equilibrated with an atmospheric ratio of CO2 in air. We also outlined a narrower interest in the published offset in annual CO2 cycles between marine and terrestrial stations which record atmospheric CO2 levels (Ye Yaun et al 2019).

Provisional results were compared with published values from seawater that had been `killed’ and acidified (Li and Tsui 1971 and Weiss (1974). Working at atmospheric partial pressures of CO2 however, a definitive value for the respective Henry constant was complicated by the difficulty of predicting an equilibrium asymptote in either water or seawater determinations.

We therefore listed a number of modifications to be adopted in future campaigns to address this issue. One proposed modification was to investigate alternative catalysts. Sodium dodecyl sulphate (SDS) is therefore replaced with a natural enzyme complex, generically carbonic anhydrase (CA), described as efficient in the reversible hydration of CO2 to bicarbonate.  CA is seen as an enzyme family with several independent evolutions across the phylogenetic tree, abundant in plants, diatoms, eubacteria and archaea (Supuran, 2016). The metallo-proteins are described as including a reaction space that combines one half hydrophilic and the opposing half hydrophobic, `allowing these enzymes to act as some of the most effective catalysts known in nature’.

In Phase 2 we therefore compare water samples with and without dosing with an infusion of terrestrial soil biota, while for seawater, being a living medium, we use a freshly-unfrozen sample of UK Atlantic coast water for each determination.

How to cite: Durham, B. and Pfrang, C.: An ocean-atmosphere paradox, Phase 2, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-11718, https://doi.org/10.5194/egusphere-egu24-11718, 2024.

Solar energy is an intricate phenomenon, especially within tropical insular locations, where this energy source demonstrates significant fluctuations across various short-term timeframes and spatial dimensions. Research on the stochastic characteristics of solar energy is gaining momentum in the scientific literature, revealing signs of scaling properties despite its inherent complexity. This paper sequentially delves into the examination of temporal fluctuations scaling and multifractal properties of irradiance for tropical insular sites (Guadeloupe, Réunion, Hawaï). By analogy with Taylor law performed on several complex process, an analysis of temporal fluctuations irradiance scaling properties is proposed. The results showed that the process of intradaily variability obeys Taylor’s power law for every short time scales and several insolation conditions. This approach elucidates the relationship between the variance of fluctuations and the mean with exponent between 1 and 2. This could confirm the relevance of Tweedie Convergence Theorem in a manner related to the central limit theorem; a mathematical basis for Taylor’s power law, 1/f noise and multifractality according to Kendal and Jørgensen [1].
Through various multifractal analysis techniques, including MFDFA, wavelet leader, structure functions, and arbitrary order Hilbert spectral analysis, on global solar radiation sequences, the intermittent and multifractal properties inherent in global solar radiation data have been brought to light across  scales ranging from one second to few hours and all intensities.
The understanding the dynamics of irradiance fluctuations is essential in various fields, including atmospheric science, remote sensing, and renewable energy. The results of these properties can help improve the modeling and prediction, which is crucial to optimally integrate PV onto electrical grids.

 

Reference

W. S. Kendal and B. Jorgensen, Tweedie convergence: A mathematical basis for Taylor's power law, 1/f noise, and multifractality. Phys. Rev. E 84, 066120, 2011. DOI. https://doi.org/10.1103/PhysRevE.84.066120.

 

How to cite: Calif, R. and Andre, M.: Temporal fluctuations scaling, multifractality and Tweedie distributions of solar energy in insular context, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-13680, https://doi.org/10.5194/egusphere-egu24-13680, 2024.

EGU24-13694 | ECS | Posters on site | NP3.3

Multifractal analysis of Cn2  scintillometer data 

Sitian Zhu, Auguste Gires, Cedo Maksimovic, Ioulia Tchiguirinskaia, and Daniel Schertzer

The cooling efficacy of green roofs in mitigating the urban heat island (UHI) effect within dense cities is largely attributed to evapotranspiration (ET) processes. Hence, accurate understanding and quantification of ET are pivotal for optimizing this cooling effect. ET estimation can be achieved either directly (weighing lysimeters) or indirectly (e.g., Penman-Monteith equation). Micro-meteorological approaches have been developed in recent years. Among which scintillometer can evaluate ET by its measurement parameter   which corresponds to the fluctuations of air refractive index  ) in combination with surface energy balance and Monin-Obukhov similarity theory. Hence, improvement in  data as well as understanding of its variability across wide range of space-time scale would result in better ET estimation and ultimately optimization. Yet it is often overlooked, and little research has focused on it and notably its variability.

This study explores the ET estimation on a wavy and vegetated green roof covering an area of 1 ha, known as the Blue Green Wave, which is located in Ecole des Ponts Paristech campus. Data from a large aperture scintillometer with 10-minute timestep during December 2019 and January 2020 is adopted.   data variability across scales was analysed with the help of structure function and Universal Multifractal model (UM). The UM framework, widely employed for characterizing and simulating geophysical fields extremely variable across wide range of space-time scales, relies on two parameters with physical interpretation: the mean intermittency codimension  and multifractality index  (, indicates monofractal; , indicates log-normal model.) An additional one, which is needed for non-conservative fields such as ET is the non-conservativeness parameter H.

Both structure function and UM approaches reveal good scaling behaviour on scales ranging from 10 min to 2h, confirming the relevance of the framework and demonstrating the potential for upscaling and downscaling. UM analysis conducted through Trace Moment and Double Trace Moment methods, provided similar values for UM parameters around   H is approximately 0.44 in our case, which deviates from traditional scaling laws due to the intricate composition of the fluxes and requires further investigations. Indeed  is influenced by temperature, humidity, air pressure and wind speed. To interpret properly structure function analysis from UM analysis, it is necessary to introduce a parameter denoted a. It corresponds to the power to which the assumed conservative underlying field should be raised before fractional integration to account for non-conservativeness to retrieve the studied field.  Here, we observed that a is around 0.76 to ensure the highest consistency of the outcome from both the structure function and UM analyses. A better understanding of the underlying complexity and variability of Cn2 is achieved by our analysis. This, in turn, improves our understanding of the underlying physical processes generating variability and temporal-spatial dynamics in ET, which paves the way for future applications.

 

How to cite: Zhu, S., Gires, A., Maksimovic, C., Tchiguirinskaia, I., and Schertzer, D.: Multifractal analysis of Cn2  scintillometer data, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-13694, https://doi.org/10.5194/egusphere-egu24-13694, 2024.

EGU24-14135 | ECS | Posters on site | NP3.3

Spatial coherence as a key metric for interpreting marine records of Holocene temperature variability 

Rebecca Cleveland Stout, Cristian Proistosescu, and Gerard Roe

Constraining forced and unforced climate variability impacts interpretations of past climate variations and predictions of future warming. However, comparing general circulation models (GCMs) and Holocene hydroclimate proxies reveals significant mismatches between simulated and reconstructed low-frequency variability on multi-decadal to multi-centennial timescales. Using a combination of GCMs and energy balance models, we have previously identified robust differences in the spatial pattern and magnitude of forced and unforced temperature variability on these long timescales. Our work suggests that not only is it important to understand variance, but also the spatial correlation between temperature at different sites. In principle, the spatial correlation at low frequencies is strongly related to the nature of variability. Now, we apply this dynamical understanding to the proxy record—specifically, across 49 globally-distributed Holocene sediment core sites with Mg/Ca and Uk37-based temperature reconstructions. We identify spatiotemporal statistics of forced and unforced variability using GCMs, and then use proxy-system models to assess how variability and spatial correlation are filtered by Mg/Ca and Uk37. Understanding these spatial correlations provides extra targets for interpreting these cores. Ultimately, we seek to characterize the forced and unforced components of slow modes of climate adjustment across the Holocene. 

How to cite: Cleveland Stout, R., Proistosescu, C., and Roe, G.: Spatial coherence as a key metric for interpreting marine records of Holocene temperature variability, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-14135, https://doi.org/10.5194/egusphere-egu24-14135, 2024.

EGU24-16097 | ECS | Posters on site | NP3.3

Climatic and environmental impacts of an Oruanui-like supereruption in the Southern Hemisphere extratropics 

Christina Brodowsky, Simon Barker, Michael Sigl, and Kirstin Krüger

Explosive volcanic eruptions have disrupted the climate system dramatically in the past. Recent volcanological fieldwork suggests that at least four VEI 8 events took place in the past 100’000 years, depositing large amounts of volcanic volatiles onto polar ice sheets, each one with potentially significant impacts on human life on Earth. Previous studies on this research topic and time period tend to focus either on tropical eruptions or only consider changes in radiative forcing due to orbital parameters, solar variability, or changes in atmospheric CO2. Here, we seek to evaluate the climatic and environmental impacts of the ~25.5 ka Oruanui eruption (Taupō caldera, 38°S, 175°E, New Zealand). We thereby refine our understanding of the volcanic forcing based on volcanological and ice core data to provide a basis for long-term climate simulations. We use existing emission details for an idealized Oruanui-like eruption scenario. We run an ensemble of CESM2/WACCM simulations with 1850 pre-industrial conditions and instantaneously emit 260 Tg SO2, and the corresponding halogen load derived from petrological estimates into the stratosphere. We then analyze the climatic effects in the decades following the eruption compared with available paleo proxies. Our overarching goal is to provide comprehensive insights into the climatic and environmental repercussions of an Oruanui-like eruption, with a specific emphasis on the differences to tropical events of comparable magnitude. By comparing these two distinct types of eruptions, we aim to contribute to a refined understanding of volcanic impacts on Earth's climate and human life.

How to cite: Brodowsky, C., Barker, S., Sigl, M., and Krüger, K.: Climatic and environmental impacts of an Oruanui-like supereruption in the Southern Hemisphere extratropics, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-16097, https://doi.org/10.5194/egusphere-egu24-16097, 2024.

EGU24-16400 | Posters on site | NP3.3

Uncertainties in modelling diagenetic self-organisation in limestone-marl sequences 

Hanno Spreeuw, Johan Hidding, Niklas Hohmann, and Emilia Jarochowska

Rhythmic variations in the properties of sediments are commonly used as archives of paleoclimate changes driven by variation in insolation caused by the changes in the Earth's orbit and the tilt of its axis. But rhythmicity can also arise from diagenetic self-organization. Distinguishing between these two drivers requires simulating self-organization. We started with an attempt to reproduce the main results from a paper by Ivan L'Heureux (2018)¹ - who proposed a mathematical model of a nonlinear dynamical system, in which self-organized oscillations arise from homogenous initial sediment and result in sediment layers with different compositions. The model consists of five stiff differential equations, for the composition of calcite and aragonite, two mineral polymorphs of CaCO3, of which aragonite is metastable, for the concentrations of calcium and carbonate ions in the pore water and for the porosity, as functions of depth and time. The self-organized patterns are in this model the result of two processes happening at different temporal scales: rapid dissolution of aragonite and slow sediment compression in response to increased porosity as aragonite is removed from the solid phase. Reproducing the steady-state distributions along depth required a major effort, mostly with regard to understanding what triggers numerical instabilities, but was finally successful.  
Currently, we have not yet succeeded in reproducing oscillations, that L'Heureux predicted
, without requiring an external force, for high initial and boundary sediment porosity.  It is essential that we are able to determine for which inital and boundary conditions oscillations should occur, beyond the uncertainties introduced by numerical algorithms for solving partial differential equations, e.g. for many sets of parameters the integrations over time can easily "derail". We have formulated two questions that we want to share with the audience in order to seek help. These are our questions: 

1) Do the five differential equations describe the underlying physics adequately?  
2) Our current software implementation of the five differential equations does not yield any oscillations, is that a flaw on our side, or does this agree with mathematical insights?

  • "Diagenetic Self-Organization and Stochastic Resonance in a Model of Limestone-Marl Sequences" by Ivan L'Heureux (2018). https://doi.org/10.1155/2018/4968315

How to cite: Spreeuw, H., Hidding, J., Hohmann, N., and Jarochowska, E.: Uncertainties in modelling diagenetic self-organisation in limestone-marl sequences, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-16400, https://doi.org/10.5194/egusphere-egu24-16400, 2024.

EGU24-16778 | ECS | Orals | NP3.3

Probability distributions as indicators of dissipative dynamics in river chemistry 

Caterina Gozzi, Axel Kleidon, and Antonella Buccianti

The chemistry of rivers plays a crucial role in comprehending the evolution of weathering processes, especially in the context of climate change and human activities. As weathering proceeds within river catchments, chemical concentrations tend to move towards saturation, or thermodynamic equilibrium. However, thermodynamic equilibrium is extremely difficult to achieve in an open system where matter and energy are continuously exchanged.

The speed of weathering processes and the associated probability distributions of concentrations values differ among geochemical species. We demonstrate that these differences are characterized by the rate of entropy production associated with the mixing of groundwater enriched with weathering products with the less saturated river water.

Based on river chemistry and discharge data observations in the Arno River basin in central Italy, we distinguish two groups of chemical variables, reflecting different levels of dissipative behavior. We show that Calcium (Ca2+) and Bicarbonate (HCO3-) concentrations are close to saturation along most of the downstream length of the Arno River, with decreasing dissipation rates and a (log)normal distribution, while Sodium (Na+) and Chlorine (Cl) concentrations increase substantially downstream, showing increased dissipation rates and being power-law distributed. This supports our hypothesis that power law distributions appear to be indicative of dissipative systems far from thermodynamic equilibrium, while (log)normal distributions indicate weakly dissipative systems close to equilibrium. This suggests that the frequency distributions of environmental variables are intricately connected to their thermodynamic state, and the degree of disequilibrium constrains the range over which power-law scaling can be observed. These results should contribute to a more comprehensive understanding of the characteristics and underlying mechanisms that lead to these types of distributions, allowing to better classify variability in systems based on how dissipative they are.

How to cite: Gozzi, C., Kleidon, A., and Buccianti, A.: Probability distributions as indicators of dissipative dynamics in river chemistry, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-16778, https://doi.org/10.5194/egusphere-egu24-16778, 2024.

EGU24-17721 | ECS | Posters on site | NP3.3

Multifractal analysis of aerosol particle concentration during rain and dry conditions in nm and µm range 

Jerry Jose, Yelva Roustan, Auguste Gires, Ioulia Tchiguirinskaia, and Daniel Schertzer

Below cloud scavenging by rain is known to be a very efficient sinking mechanism for aerosols in atmosphere. Since this scavenging depends on interaction between aerosol particles as well as the scavening raindrops, and notably their respective size ranges, it is interesting to examine both fields together across various size ranges and across temporal scales. Towards this, a 4 month long data was used from Cherbourg-Octeville, France from 01/11/2010 to 12/03/2011 from the experimental station managed by Institut de Radioprotection et de Sûreté Nucléaire (IRSN). Here, simultaneous and continuous measurement of size resolved particle concentration (14.6 to 478.3 nm and 0.523 to 19.81 µm) range has been done using Scanning Mobility Particle Sizer (SMPS) and Aerodynamic Particle Sizer (APS), and rain measurement using a disdrometer.

Variation of total aerosol concentration in nm and µm range, as well as individual number concentration in small size bins were analyzed according to rain and dry events, using the framework of Universal Multifractals (UM). UM is widely used, as a physically based scale invariant framework, for characterizing and simulating extreme variability and intermittency in geophysical fields. From initial analysis, the total concentration showed scaling properties (1 min to 1 hr), in both rain and dry events, regardless the scavenging efficiency of event. This was further explored in individual concentration ranges and they showed similar scaling properties in different rain types. However, while considering the different stages of rain, say start and end, the values of UM parameters showed some variation. To understand the behavior more clearly, few sizes were selected from nm and µm range, and efforts were made to extract the field which is devoid of scavenging by rain. Understanding the correct transformation required to extract accurate UM values and comparing the scavenging and non scavenging fields will improve understanding of particle concentration variation, and eventually understanding of scavenging coefficient.

How to cite: Jose, J., Roustan, Y., Gires, A., Tchiguirinskaia, I., and Schertzer, D.: Multifractal analysis of aerosol particle concentration during rain and dry conditions in nm and µm range, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-17721, https://doi.org/10.5194/egusphere-egu24-17721, 2024.

EGU24-17984 | ECS | Orals | NP3.3

Observation of different multifractal phase transitions over three typhoon events 

Ching-Chun Chou, Auguste Gires, Li-Pen Wang, Ioulia Tchiguirinskaia, and Daniel Schertzer

Rainfall is extremely variable in both space and time, which makes its analysis complex. A widely used framework to properly handle these features is Universal Multifractals (UM), which is a physically based and mathematically robust framework. It relies on three parameters, meaning it is parsimonious. Two types of multifractal phase transitions can affect the analysis of a series: (i) the divergence of moments, which is related to the singular limit of the underlying cascade process at small scales and notably explains the power law fall-off observed on numerous geophysical fields, (ii) sampling limitations, which is related to the fact that great moments cannot be observed on finite series.
 
This study employs UM to analyse the time series of rainfall intensities observed by the Parsival2 disdrometer at the 10-second resolution from three distinct typhoons over the period of July to October 2022, revealing differences and limitations in their statistical characteristics. It enables us to illustrate the  two previously mentioned concepts of divergence of moments and sampling limitations and their impact on the analysis of rainfall data.

The analysis of typhoon Hinnamnor exhibited limitations due to the sampling dimension, indicating that the current data length was insufficient to capture the multifractal nature of the rainfall events for large moments, reducing the robustness of the analysis for moments greater than 5.43. It reflects sampling limitations, leading to a constrained understanding of extreme events.

For typhoon Nalgae, our analysis highlighted the occurrence of divergence of moments, i.e. a limitation associated with a critical moment. As higher-order moments were calculated, we observed statistical values tending towards infinity, suggesting that extreme rainfall events significantly influenced this typhoon and pointing out the inadequacy of traditional statistical methods in such scenarios. Such multifractal phase transition is seldom observed on individual series, highlighting the interest of studying this typhoon series. 

Finally, the analysis of the typhoon Nesat presented a behavior affected by both multifractal phase transitions resulting in a more complex interpretation. 

Our findings provide new insights into the multifractal analysis of typhoon rainfall intensities, emphasising the importance of considering multifractal theory and its associated phase transitions when dealing with natural phenomenon data. These discoveries lay a crucial methodological foundation for more accurate prediction and response to extreme weather events. In case of rainfall, it then has some hydrological consequences notably in terms of stormwater management or optimization of dams for hydraulic production.

How to cite: Chou, C.-C., Gires, A., Wang, L.-P., Tchiguirinskaia, I., and Schertzer, D.: Observation of different multifractal phase transitions over three typhoon events, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-17984, https://doi.org/10.5194/egusphere-egu24-17984, 2024.

EGU24-19055 | ECS | Posters on site | NP3.3

Role of size and height of ice sheet on millennial-scale climate variability 

Yuta Kuniyoshi, Ayako Abe-Ouchi, Sam Sherriff-Tadano, and Wing-Le Chan

Unlike the interglacial stable climate, glacial climate was dominated by millennial-scale variability, which is strongly associated with changes in the Atlantic meridional overturning circulation (AMOC). The development of the North American ice sheet has been shown to have a significant impact on the strength of the AMOC through surface cooling and enhanced surface winds. However, the impact of mid-glacial ice sheet involved in millennial-scale variability of the AMOC are still elusive. Here, using a coupled atmosphere-ocean model MIROC4m, we perform several climate simulations under mid-glacial ice sheet configurations. We use Marine Isotope Stage (MIS)-5a and MIS-3 ice sheet configurations as boundary conditions, which are derived from the simulation of an ice sheet model, IcIES-MIROC. These volumes are the 40 m sea level equivalent for MIS5a (approximately 33% of the LGM) and the 96 m sea level equivalent for MIS3 (approximately 80% of the LGM). To account for uncertainty in the altitude of the ice sheet, we also conduct experiments under topographic conditions in which only the altitude was changed, but not the extent, for each ice sheet configuration. As a result, self-sustained oscillations of millennial-scale periodicity in the climate and AMOC are simulated for both ice sheet cases. The result suggests that the millennial-scale climate variability could occur as long as the North American ice sheet exists, even if the ice sheet is small. The expansion of the North American ice sheet from MIS5a to MIS3 have an influence of shortening the weak AMOC period (stadial) and lengthening of the strong AMOC period (interstadial), because the stronger surface winds over North Atlantic enhance retreat of sea-ice during the stadial and increase salt transport via wind-driven ocean circulation during the interstadial. Meanwhile, one of the other simulations under the ice sheet condition with MIS3-equivalent extent but altitudes as low as 50% results in a persistent stadial state, which is due to the large cooling effect. Our results show that the relative strength of surface wind and surface cooling effects depends on the ice sheet configuration, which could modify the length of stadial and interstadial.

How to cite: Kuniyoshi, Y., Abe-Ouchi, A., Sherriff-Tadano, S., and Chan, W.-L.: Role of size and height of ice sheet on millennial-scale climate variability, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-19055, https://doi.org/10.5194/egusphere-egu24-19055, 2024.

EGU24-19127 | ECS | Orals | NP3.3 | Highlight

Climate legacies in macroevolutionary dynamics 

Gregor Mathes

Biodiversity is critically endangered by anthropogenic climate change. One of the core goals of ecological research and conservation science is therefore to enhance the mechanistic understanding of the processes that cause species to go extinct, particularly in light of anthropogenic climate change. However, the presence of non-linearities, multiple equilibria, thresholds, and internal feedbacks within ecological and climatic systems often impedes a mechanistic comprehension. One fundamental issue for extinction studies using contemporary data is that this data is always dependent on past conditions. Within ecology, the dependence of contemporary biodiversity dynamics on past climate is generally termed “climate legacy”. Climate legacies can arise from a multitude of ecological processes, such as time lags, niche conservatism, physiological thresholds, or cascading effects. Further, climate legacies can be assumed to be present in all ecological systems as a consequence of the dynamic nature of ecological patterns and processes. If not accounted for, climate legacies can hinder or even prevent the detection of true ecological responses to climate change. However, few studies on the relationship between extinction dynamics and climate include these climate legacies. Even less studies reach beyond merely discussing potential impacts of climate legacies and include them in their empirical framework. Those studies where climate legacies were included and quantified found a large impact of these legacy effects on extinction dynamics. Here I introduce a methodical framework for the quantification of effects arising from climate legacies in biotic systems of any temporal scale. I first introduce the concept of climate interactions, which describe and quantify the potential dependence of extinction risk on the long-term climatic context. Climate interactions might create a characteristic pattern in extinction dynamics and can arise from climate legacies acting over days to millions of years. They therefore provide a unifying framework for studying the consequences of climate legacies in ecosystems. The expected characteristic pattern consists of higher extinction risk, or related measures, when climatic changes add to previous trends in the same direction (such as a short-term warming adding to a long-term warming trend). It is hypothesized that these synergistic climate interactions first lead to environmental conditions increasingly different from initial adaptations of taxa, which then result in a higher extinction risk for these taxa. An antagonistic climate interaction, where a short-term climate change reverses a previous long-term trend (such as short-term cooling adding to a long-term warming trend), might result in a generally lower extinction risk through climatic conditions being more similar to initial adaptations of taxa. The emergence of expected patterns are then tested in  a variety of ecosystems, both marine and terrestrial, taking advantage of the fossil record with its rich information of past responses of organisms to climatic changes. 

How to cite: Mathes, G.: Climate legacies in macroevolutionary dynamics, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-19127, https://doi.org/10.5194/egusphere-egu24-19127, 2024.

EGU24-926 | ECS | Posters on site | CR2.2

Simulating the impact of an AMOC weakening on the Antarctic Ice Sheet using a coupled climate and ice sheet model 

Anna Höse, Moritz Kreuzer, Willem Huiskamp, Torsten Albrecht, Stefan Petri, Ricarda Winkelmann, and Georg Feulner

Many model studies show that a shutdown of the Atlantic meridional overturning circulation (AMOC) causes reduced northward heat transport into the North Atlantic and a warming Southern Ocean in addition to shifts in large-scale atmospheric circulations. How these changing climate conditions could influence the present-day state of the Antarctic Ice Sheet is little studied even though observational data of AMOC strength show a slowdown trend over the last decades. The ocean current as well as the Antarctic Ice Sheet might reach climate tipping points triggering irreversible processes with consequences already on human time-scales. It's unclear whether increasing Southern Ocean temperatures due to a AMOC shutdown could accelerate basal melting rates, the critical parameter which in turn may induce tipping of the West Antarctic Ice Sheet.

Here, a freshwater hosing that forces the shutdown of the AMOC is applied to the North Atlantic in a global climate model with an interactive ice sheet model for Antarctica. This model framework consists of the Parallel Ice Sheet Model (PISM) that is coupled to the CM2Mc global Earth system model via the ice shelf cavity model PICO (Potsdam Ice-shelf Cavity mOdel). PISM is interactively coupled to the ocean module in order to investigate feedbacks at the ice-ocean boundary, while the atmospheric forcing is prescribed. Preliminary results show that an AMOC shutdown results in warming sea surface temperatures in the southern hemisphere along with a small shift in the mid-latitude westerlies due to reduced northward heat transport, which is in line with previous studies. Antarctic marginal temperatures decrease, however, resulting in a reduction of Antarctic mass through increased calving and decreased basal melting.

How to cite: Höse, A., Kreuzer, M., Huiskamp, W., Albrecht, T., Petri, S., Winkelmann, R., and Feulner, G.: Simulating the impact of an AMOC weakening on the Antarctic Ice Sheet using a coupled climate and ice sheet model, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-926, https://doi.org/10.5194/egusphere-egu24-926, 2024.

EGU24-966 | ECS | Orals | CR2.2

Greenland Ice Sheet evolution during the Last Interglacial with an improved surface mass balance modeling approach  

Thi Khanh Dieu Hoang, Aurélien Quiquet, Christophe Dumas, Andreas Born, and Didier M. Roche

The Last Interglacial period (LIG) (130 - 116 kaBP), characterized by higher global mean temperature and sea levels compared to the present-day due to the Earth’s orbit configuration, has been well-studied as a recent example of a climate period warmer than today. There is particular interest in studying the ice sheet-climate interactions in view of our current climate change. However, the extent of the ice sheet and its contribution to the rise of sea levels during the LIG remain debatable as different approaches suggest a wide range of estimations. In order to cover such a long period, some processes are simplified in the modeling approach by using prescribed forcings, simple surface mass balance (SMB) schemes, or equilibrium simulations, which all affect the numerical estimation of ice sheet evolution. 

In our work, to perform transient simulations, we use an Earth system model of intermediate complexity (iLOVECLIM), which has been widely used to study various long-timescale periods. Additionally, we use a physically-based energy and mass balance model with 15 vertical snow layers BESSI (BErgen Snow Simulator) to account for the effect of insolation changes as well as snow-albedo feedback. The climate forcings of the snow model are obtained by running iLOVECLIM transiently from 135 to 115 kaBP, downscaled over the Northern Polar region. Using the SMB computed by BESSI, we then simulate the ice sheet evolution during the LIG with GRISLI - the ice sheet model in the iLOVECLIM framework. 

To assess the benefits of using a physically-based SMB model in the ice sheets simulation, the outputs of GRISLI-BESSI are compared to the current SMB scheme of iLOVECLIM, a simple parametrization called ITM (Insolation Temperature Melt). The Greenland ice sheet volume simulated by the two SMB models reaches the minimum value at 127.5 kaBP, around 500 years after the peak of global mean temperature. The magnitude of ice sheet retreat and its contribution to the sea level in ITM simulations are significantly higher than in BESSI due to an overestimation of the zones of ablation. 

The findings suggest that, compared to a parameterization, we have more confidence in the ice sheet estimation with a physically-based SMB model. Further works with fully interactive ice sheet modeling that takes into account the melt-elevation feedback can improve the simulation of the ice sheet-climate interactions of long-time scales. 

How to cite: Hoang, T. K. D., Quiquet, A., Dumas, C., Born, A., and Roche, D. M.: Greenland Ice Sheet evolution during the Last Interglacial with an improved surface mass balance modeling approach , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-966, https://doi.org/10.5194/egusphere-egu24-966, 2024.

EGU24-1991 | ECS | Orals | CR2.2 | Highlight

When will the Antarctic ice shelves not be viable anymore? 

Clara Burgard, Nicolas C. Jourdain, Christoph Kittel, Cyrille Mosbeux, Justine Caillet, and Pierre Mathiot

The Antarctic contribution to sea-level rise in the coming centuries remains very uncertain, due to the possible triggering of instabilities such as the Marine Ice Sheet Instability (MISI) and Marine Ice Cliff Instability (MICI). These instabilities are mainly linked to the evolution of the floating ice shelves, which usually buttress the ice flow from the ice-sheet to the ocean. However, these are currently thinning. Better understanding the evolution of ice shelves in the next decades to centuries is therefore important and crucial to better anticipate the evolution of sea-level rise.

In this study, we investigate the viability of ice shelves for a number of climate models and scenarios. This is estimated from the emulation of the surface and basal mass balance of MAR and NEMO respectively, and from high-end dynamical ice flows obtained through Elmer/Ice. We then use a Bayesian calibration to give weight to members closer to observations. We find that large uncertainties remain, mainly because of the uncertainty in basal melt, and that viability limits vary largely depending on the ice-shelf location.

How to cite: Burgard, C., Jourdain, N. C., Kittel, C., Mosbeux, C., Caillet, J., and Mathiot, P.: When will the Antarctic ice shelves not be viable anymore?, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-1991, https://doi.org/10.5194/egusphere-egu24-1991, 2024.

EGU24-3666 | Orals | CR2.2

Deciphering Antarctic Ice Sheet Mass Loss: A Modeling Approach to Distinguish Climate Change from Natural Variability 

Johanna Beckmann, Hélène Seroussi, Lawrence Bird, Justine Caillet, Nicolas Jourdain, Felcity McCormack, and Andrew Mackintosh

The Antarctic Ice Sheet (AIS) is currently undergoing accelerated mass loss, significantly contributing to rising sea levels (SLR). Despite numerous observations, uncertainties persist in understanding the drivers and dynamic responses of AIS mass loss. Climate variability strongly influences AIS dynamics, but limited observational data hinders precise attribution to climate change or natural variability. This study addresses this gap by employing advanced modeling techniques to assess the extent to which observed and future AIS mass loss can be attributed to climate change versus variability. Utilizing a unique "initialization method" with the ISSM model, we approximate the AIS state circa 1850, a period minimally affected by anthropogenic forces. From this baseline, we project AIS development using UKESM1 forcing, comparing scenarios with and without anthropogenic influence. This investigation aims to enhance our understanding of the impact of climate change on the AIS and its implications for future SLR.

How to cite: Beckmann, J., Seroussi, H., Bird, L., Caillet, J., Jourdain, N., McCormack, F., and Mackintosh, A.: Deciphering Antarctic Ice Sheet Mass Loss: A Modeling Approach to Distinguish Climate Change from Natural Variability, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-3666, https://doi.org/10.5194/egusphere-egu24-3666, 2024.

EGU24-4093 | ECS | Posters on site | CR2.2

Interactions between ocean circulation and the Northern Hemisphere ice sheets at 40 ky B.P. in an Earth System Model (iLOVECLIM-GRISLI) 

Louise Abot, Claire Waelbroeck, Aurélien Quiquet, Casimir Delavergne, and Nathaelle Bouttes

During the last glacial period, the climate went through rapid fluctuations together with changes in ocean circulation and ice sheets volume accompanied by iceberg discharges. These rapid climate variations, namely Dansgaard-Oeschger events, are still not fully explained. This study’s aim is to contribute to their better understanding, focusing on interactions between ice sheets and ocean circulation. To this end, we use the iLOVECLIM-GRISLI coupled climate-ice sheet model and run two different perturbation experiments related to the ice sheet and ocean components. Starting from a quasi equilibrium corresponding to 40 ky B.P. greenhouse gas concentration, incoming solar radiation and ice sheet volume, the first experiment consists in imposing either constant or amplified sub-shelf melt rates in comparison with the control simulation. In the second experiment, we focus on the interface between the ice sheets and the bedrock. The basal friction coefficient values are imposed following the same procedure. These two experiments are similar to freshwater hosing experiments but here the water comes directly from the interactively computed ice sheets change. For each experiment, the perturbation is imposed for 500 years before returning to the unperturbed conditions for one thousand years and its impacts on the climate system are investigated. Our results highlight feedbacks that may help to explain the abrupt nature of the climate transitions observed during the last glacial period. 

How to cite: Abot, L., Waelbroeck, C., Quiquet, A., Delavergne, C., and Bouttes, N.: Interactions between ocean circulation and the Northern Hemisphere ice sheets at 40 ky B.P. in an Earth System Model (iLOVECLIM-GRISLI), EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-4093, https://doi.org/10.5194/egusphere-egu24-4093, 2024.

EGU24-4802 | Orals | CR2.2

A synchronously coupled global model iOM4: a new modeling tool for simulations of the ocean-cryosphere interactions  

Olga Sergienko, Matthew Harrison, Alexander Huth, and Nicole Schlegel

How to cite: Sergienko, O., Harrison, M., Huth, A., and Schlegel, N.: A synchronously coupled global model iOM4: a new modeling tool for simulations of the ocean-cryosphere interactions , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-4802, https://doi.org/10.5194/egusphere-egu24-4802, 2024.

EGU24-5104 | ECS | Posters on site | CR2.2

Simulating Antarctic Ice Sheet evolution through the mid-Pleistocene transition 

Christian Wirths, Antoine Hermant, Christian Stepanek, Johannes Sutter, and Thomas Stocker

Unravelling the main drivers of the mid-Pleistocene transition (MPT; around 1.2–0.8 million years ago) remains a significant challenge in paleoclimate research. Noteworthy changes that occurred in the climate system during that time include a pronounced shift from 41-kyr to 100-kyr periodicity of glacial cycles and the emergence of much larger ice sheets. While a number of studies have focused on the interplay between the climate system and northern hemispheric ice sheets during the MPT, the role of Antarctica in driving and responding to climate change at that time remains largely unknown. This is particularly relevant as, consequently, the response of Antarctica’s vast ice sheets to a major transition in Quaternary climate, and their potential role in shaping the transition, remain uncertain. 

Here, we use the Parallel Ice Sheet Model (PISM) to simulate the transient evolution of the Antarctic Ice Sheet through the MPT. Computation of the evolution of ice sheets in PISM is enabled by means of a climate index approach that is based on snapshots of climatic conditions at key periods. The climate index approach interpolates between individual climate snapshots based on various paleo-proxy records. Further, we test Antarctica's response to different pre-MPT GCM snapshots of different CO2, orbital, and land-sea mask configurations. Climate snapshots are derived from the Community Earth System Models (COSMOS), a general circulation model that simulates atmosphere, ocean, sea ice and land vegetation in dependence of reconstructions of paleogeography, orbital configuration, and greenhouse gas concentrations.  

Our study aims to better understand the evolution of the Antarctic Ice Sheets during the MPT and to constrain potential dynamical transitions in the climate-cryosphere system. Furthermore, we seek to clarify the influence of different pre-MPT ice sheet configurations on simulated characteristics of this transition.  

The findings from this study will contribute to an improved understanding of cryospheric changes that occurred during the Quaternary. Furthermore, we aim to provide insights into potential future Antarctic trajectories under anthropogenic climate change. 

How to cite: Wirths, C., Hermant, A., Stepanek, C., Sutter, J., and Stocker, T.: Simulating Antarctic Ice Sheet evolution through the mid-Pleistocene transition, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-5104, https://doi.org/10.5194/egusphere-egu24-5104, 2024.

EGU24-5525 | Orals | CR2.2

Modeling the Antarctic Surface Mass Balance with a coarse temporal resolution 

Enrico Maiero, Florence Colleoni, Cécile Agosta, Carlo Barbante, and Barbara Stenni

Sublimation is the most important ablation term in the Antarctic Surface Mass Balance (SMB) (Agosta et al., 2019), while it is currently negligible for both Greenland and mountain glaciers (prevailing surface melt). Since simple parameterized SMB models are usually developed for Greenland and Alpine glaciers, they mostly misrepresent sublimation. To face this problem, we developed EBAL, a new parameterized Energy SMB model for Antarctica based on SEMIC (Krapp et al., 2017), which is an Energy SMB model developed for Greenland whose main innovations are a sinusoidal parameterization for the diurnal cycle to assess melt and refreezing and an albedo dependence on snow depth. EBAL was calibrated with both MAR (Kittel et al., 2022) and RACMO (Wessem et al., 2018) outputs for the period 1979-2000 and for the period 2075-2099 under the SSP5-8.5. EBAL can reproduce the statistical properties of MAR and RACMO sublimation time series and spatial distribution even if it uses a coarse time step (1 day). However, our final aim is to use EBAL for paleoclimate simulations, for which the temporal resolution of the inputs is even coarser, as often only monthly data is available. Thus, we have tested the idea of superimposing the present day-to-day variability on the MAR monthly atmospheric forcing of SSP5-8.5. Simulated SMB with EBAL forced with MAR original daily SSP5-8.5 inputs leads to a 210 Gt/yr sublimation, and to a 1425 Gt/yr melt. When forcing EBAL with monthly means only (linearly interpolated), we obtain a 113 Gt/yr sublimation and a 620 Gt/yr melt. When adding present-day variability to linearly interpolated monthly inputs, EBAL computes a 175 Gt/yr sublimation and a 1386 Gt/yr melt. Those latter numbers are very similar to those obtained when forcing with daily inputs. We propose to use this method to test EBAL for paleoclimate applications.

References

  • Agosta, C. et al., (2019). “Estimation of the Antarctic surface mass balance using the regional climate model MAR (1979–2015) and identification of dominant processes”. The Cryosphere. 13,  pp. 281-296. 10.5194/tc-13-281-2019. 
  • Kittel, C. et al., (2022). “Clouds drive differences in future surface melt over the Antarctic ice shelves”. The Cryosphere. 16, pp. 2655-2669. 10.5194/tc-16-2655-2022.
  • Krapp, M et al., (July 2017). “SEMIC: an efficient surface energy and mass balance model applied to the Greenland ice sheet”. The Cryosphere 11.4, pp. 1519–1535. 10.5194/tc-11-1519-2017
  • Wessem, J. M. et al., (Apr. 2018). “Modelling the climate and surface mass balance of polar ice sheets using RACMO2 – Part 2: Antarctica (1979–2016)”. The Cryosphere 12, pp. 1479–1498. 10.5194/tc-12-1479-2018

How to cite: Maiero, E., Colleoni, F., Agosta, C., Barbante, C., and Stenni, B.: Modeling the Antarctic Surface Mass Balance with a coarse temporal resolution, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-5525, https://doi.org/10.5194/egusphere-egu24-5525, 2024.

EGU24-5584 | Orals | CR2.2

Future Greenland melt in coupled ice sheet-climate CESM simulations: feedbacks, thresholds, reversibility 

Miren Vizcaino, Thirza Feenstra, Michele Petrini, Raymond Sellevold, Georgiou Sotiria, Katherine Thayer-Calder, William Lipscomb, and Julia Rudlang

Estimates of future Greenland ice sheet (GrIS) melt are mostly based on regional climate modelling for a fixed GrIS topography or on ice sheet modelling with forcing from climate models. This prevents the modelling of climate and GrIS feedbacks and other types of interaction. Here we examine a set of multi-century simulations with the Community Earth System Model featuring an interactive GrIS to explore future relationship between global climate change and ice sheet change. To this end, we compare a set of coupled CESM-CISM 1% CO2 increase simulations until stabilization at two, two and a half, three and four times pre-industrial CO2 levels to examine the sensitivity of the GrIS to emission mitigation. Here we find a large role of ocean circulation weakening and associated regional climate changes on GrIS melt for moderate emission scenarios and large melt differences between the three times and four times CO2 stabilization scenarios. In addition, we examine the role of feedbacks on ice sheet evolution by comparing a 1% to 4xCO2 coupled simulation with a simulation where the GrIS topography and meltwater fluxes to the ocean are prescribed as pre-industrial. Finally, we explore the effects on GrIS melt rates of a fast 5% CO2 reduction from four times to pre-industrial levels, with a focus on restoration of high latitude climate, GrIS albedo, surface energy fluxes and refreezing capacity.  

How to cite: Vizcaino, M., Feenstra, T., Petrini, M., Sellevold, R., Sotiria, G., Thayer-Calder, K., Lipscomb, W., and Rudlang, J.: Future Greenland melt in coupled ice sheet-climate CESM simulations: feedbacks, thresholds, reversibility, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-5584, https://doi.org/10.5194/egusphere-egu24-5584, 2024.

EGU24-5698 | ECS | Posters on site | CR2.2

Geoengineering's role in reducing future Antarctic mass loss is unclear 

Mira Adhikari, Daniel Martin, Tamsin Edwards, Antony Payne, James O'Neill, and Peter Irvine

Using the BISICLES ice sheet model, we compare the Antarctic ice sheet’s response over the 22nd century in a scenario where idealised large scale, instantaneous geoengineering is implemented in 2100 or 2050 (geoengineering), with scenarios where the climate forcing is held constant in the same year (stabilisation). Results are highly climate model dependent, with larger differences between models than between geoengineering and stabilisation scenarios, but show that geoengineering cannot prevent significant losses from Antarctica over the next two centuries. If implemented in 2050, sea level contributions under geoengineering are lower than under stabilisation scenarios. If implemented in 2100, under high emissions, geoengineering produces higher sea level than stabilisation scenarios, as increased surface mass balance in the warmer stabilisation scenarios offsets some of the dynamic losses. Despite this, dynamic losses appear to accelerate and may eventually negate this initial offset, indicating that beyond 2200, geoengineering could eventually be more effective.

How to cite: Adhikari, M., Martin, D., Edwards, T., Payne, A., O'Neill, J., and Irvine, P.: Geoengineering's role in reducing future Antarctic mass loss is unclear, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-5698, https://doi.org/10.5194/egusphere-egu24-5698, 2024.

EGU24-6140 | ECS | Orals | CR2.2

Long term ice-sheet albedo feedback constrained by most recent deglaciation 

Alice Booth, Philip Goodwin, and Bb Cael

Slow climate feedbacks that operate on timescales of more than a century are currently underrepresented in model assessments of climate sensitivity, and this continues to hinder efforts to accurately predict future climate change beyond the end of the 21st Century. As such, the magnitude of multi-centennial and millennial climate feedbacks are still poorly constrained. We utilise recent reconstructions of Earth’s Energy Imbalance (EEI) to estimate both the total climate feedback parameter and the ice-sheet albedo feedback since the Last Glacial Maximum. This new proxy-based record of EEI facilitates the first opportunity to simultaneously calculate both the magnitude and timescale of Earth’s climate feedback over the most recent deglaciation using a purely proxy data-driven approach, and without the need for simulated reconstructions. We find the ice-sheet albedo feedback to have been an amplifying feedback reaching an equilibrium magnitude of 0.55 Wm-2K-1, with a 66% confidence interval of 0.45 Wm-2K-1 to 0.63 Wm-2K-1. The timescale for the ice-sheet albedo feedback to reach equilibrium is estimated as 3.61Kyrs, with a 66% confidence interval of 1.88Kyrs to 5.48Kyrs. These results provide new evidence for the timescale and magnitude of the amplifying ice-sheet albedo feedback that will continue to drive anthropogenic warming for millennia to come, further increasing the urgency for an effective climate change mitigation strategy to avoid serious long-term consequences for our planet and its ecosystems.

How to cite: Booth, A., Goodwin, P., and Cael, B.: Long term ice-sheet albedo feedback constrained by most recent deglaciation, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-6140, https://doi.org/10.5194/egusphere-egu24-6140, 2024.

EGU24-7415 | ECS | Orals | CR2.2 | Highlight

Stability regimes and safe overshoots in West and East Antarctica 

Ann Kristin Klose and Ricarda Winkelmann

Earth's climate will likely exceed a warming of 1.5°C in the coming decades. Maintaining such warming levels for a longer period of time may pose a considerable risk of crossing critical thresholds in Antarctica and, thereby, triggering self-sustained, potentially irreversible ice loss, even if the forcing is reduced in a temperature overshoot. Due to the complex interplay of several amplifying and dampening feedbacks at play in Antarctica, the duration and amplitude of such warming overshoots as well as their eventual 'landing' climate will determine the long-term evolution of the ice sheet.

Using the Parallel Ice Sheet Model, we systematically test for the reversibility of committed large-scale ice-sheet changes triggered by warming projected over the next centuries, and thereby explore (1) the stability regimes of the Antarctic Ice Sheet and (2) the potential for safe overshoots of critical thresholds in Antarctica.

We demonstrate crucial features of the Antarctic Ice Sheet's stability landscape for its long-term trajectory in response to future human actions: Given ice-sheet inertia, an early reversal of climate may allow for avoiding self-sustained ice loss that would otherwise be irreversible (for the same reduction in warming) due to multistability of the ice sheet at the basin- and continental scale. While we show that such safe overshoots of critical thresholds in Antarctica may be possible, it is also clear that limiting global warming is the only viable option to evade the risk of widespread ice loss in the long term.

How to cite: Klose, A. K. and Winkelmann, R.: Stability regimes and safe overshoots in West and East Antarctica, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-7415, https://doi.org/10.5194/egusphere-egu24-7415, 2024.

EGU24-8333 | ECS | Orals | CR2.2

Coupled ensemble simulations of the Northern Hemisphere ice sheets at last two glacial maxima  

Violet Patterson, Lauren Gregoire, Ruza Ivanovic, Niall Gandy, Stephen Cornford, and Sam Sherriff-Tadano

Coupled climate-ice sheet models can capture important interactions between the ice sheets and the climate that can help us better understand an ice sheet's response to changes in forcings. In this respect, they are a useful tool for simulating future ice sheet and sea level changes as a result of climate change. However, such models have large uncertainties related to the choice of climate and ice sheet parameters used. The same processes that operate today, also occurred in glacial times, and previous work has shown that simulating the North American ice sheet at the Last Glacial Maximum (LGM; ~21 ka BP) provides a strong benchmark for testing coupled climate-ice sheet models and recalibrating uncertain parameters that control surface mass balance and ice flow (Gandy et al., 2023).

Here, we build on this work by performing the first coupled FAMOUS-BISICLES simulations of the last two glacial maxima, including all Northern Hemisphere ice sheets interactively. The ice sheet component of this model is capable of efficiently simulating marine ice sheets, such as the Eurasian ice sheet, despite the high computational cost of higher order physics. We simulate and compare both the LGM and the Penultimate Glacial Maximum (PGM; ~140 ka BP), since both periods displayed major differences in the distribution of ice between Eurasia and North America. Uncertainty is explored by running ensembles of 120 simulations, randomly varying the uncertain parameters controlling ice sheet dynamics and climate through Latin Hypercube Sampling. We also work on improving the representation of ice streams in the model through performing internal ice temperature spin ups and sensitivity tests varying till water drainage properties. The ensemble members are evaluated against empirical data on ice sheet extent and ice stream location to find combinations of parameters that produce reasonable simulations of the North American and Eurasian ice sheets for both periods. We determine the impact of the uncertainty in these parameters on the result and whether both ice sheets show similar sensitivities to the model parameters. These simulations will provide a starting point for analysing some of the interactions between the climate and the ice sheets during glacial periods and how they may have led to different ice sheet evolutions.

How to cite: Patterson, V., Gregoire, L., Ivanovic, R., Gandy, N., Cornford, S., and Sherriff-Tadano, S.: Coupled ensemble simulations of the Northern Hemisphere ice sheets at last two glacial maxima , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-8333, https://doi.org/10.5194/egusphere-egu24-8333, 2024.

The dynamics of the ice sheets on glacial-interglacial time scales are highly controlled by interactions with the solid Earth, i.e., glacial isostatic adjustment (GIA). Particularly at marine ice sheets, competing feedback mechanisms govern the migration of the ice sheet’s grounding line and hence the ice sheet stability.

In this study, we run coupled ice sheet–solid Earth simulations over the last two glacial cycles. For the ice sheet dynamics we apply the Parallel Ice Sheet Model PISM and for the load response of the solid Earth we use the three-dimensional viscoelastic Earth in view of sea-level and vertical displacement changes we apply the Viscoelastic Lithosphere and Mantle Model VILMA.

With our coupling setup we evaluate the relevance of feedback mechanisms for the glaciation anddeglaciation phases in Antarctica considering different 3D Earth structures resulting in a range of load-response time scales. For rather long time scales, in a glacial climate associated with far-field sea level low stand, we find grounding line advance up to the edge of the continental shelf mainly in West Antarctica, dominated by a self-amplifying GIA feedback, which we call the ‘forebulge feedback’. For the much shorter time scale of deglaciation, dominated by the Marine Ice Sheet Instability, our simulations suggest that the stabilizing GIA feedback can significantly slow-down grounding line retreat in the Ross sector, which is dominated by a very weak Earth structure (i.e. low mantle viscosity and thin lithosphere).

The described coupled framework, PISM-VILMA, allows for defining restart states to which to run multiple sensitivity simulations. It can be easily implemented in Earth System Models (ESMs) and provides the tools to gain a better understanding of ice sheet stability on glacial time scales as wellas in a warmer future climate.

How to cite: Albrecht, T., Bagge, M., and Klemann, V.: Feedback mechanisms controlling Antarctic glacial cycle dynamics simulated with a coupled ice sheet–solid Earth model, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-9032, https://doi.org/10.5194/egusphere-egu24-9032, 2024.

EGU24-10162 | ECS | Orals | CR2.2

A new climate and surface mass balance product for the Antarctic and Greenland ice sheet using RACMO2.4.1 

Christiaan van Dalum, Willem Jan van de Berg, Srinidhi Nagarada Gadde, and Michiel van den Broeke

Recent progress in parameterizations of surface and atmospheric processes have led to the development of a major update of the polar version of the Regional Atmospheric Climate Model (RACMO2.4.1). Here, we present a new high-resolution climate and surface mass balance product by applying RACMO2.4.1 to the Antarctic and Greenland ice sheet for the historical period (starting in 1960 and 1945, respectively). In addition, RACMO output is now available for the first time on a pan-Arctic domain, starting in 1980. We assess these products by comparing model output of the surface mass balance and its components and the near-surface climate with in-situ and remote sensing observations, and study differences with the previously operational RACMO iteration, RACMO2.3p2. 

Among other changes, RACMO2.4.1 includes new and updated parameterizations related to surface and atmospheric processes. Most major updates are part of the physics package of cycle 47r1 of the Integrated Forecast System (IFS) of the European Center for Medium-Range Weather Forecasts (ECMWF), which is embedded in RACMO2.4.1. This includes updates to the cloud, radiation, convection, turbulence, aerosol and lake scheme. Other major changes are directly related to the cryosphere, such as the introduction of a new spectral albedo and radiative transfer scheme for glaciated snow, fixes to the snow drift scheme, a new multilayer snow scheme for seasonal snow and an updated ice mask. These updates lead to changes in the near-surface climate. For example, the horizontal transport of snow that is present in the atmosphere leads to a redistribution of snowfall. Furthermore, the spatial resolution for the Antarctic domain is increased to 11 km, which is also used for the pan-Arctic domain, while 5.5 km is used for Greenland. Here, we also discuss the impact that aforementioned changes have on the climate of the polar regions and the surface mass balance and its components of the ice sheets.

How to cite: van Dalum, C., van de Berg, W. J., Nagarada Gadde, S., and van den Broeke, M.: A new climate and surface mass balance product for the Antarctic and Greenland ice sheet using RACMO2.4.1, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-10162, https://doi.org/10.5194/egusphere-egu24-10162, 2024.

EGU24-10256 | ECS | Orals | CR2.2

Reconstructing the Greenland ice Sheet during the last two deglaciations 

Majbritt Kristin Eckert, Mikkel Lauritzen, Nicholas Rathmann, Anne Solgaard, and Christine Hvidberg

The Parallel Ice Sheet Model (PISM) is used to build up a glacial Greenland ice sheet, simulate the evolution of the Greenland ice sheet through glacial terminations I and II and investigate the evolution during previous warmer climates, the Eemian and the Holocene thermal maximum. During the Holocene, surface elevation changes derived from ice cores suggest a large thinning in the North, suggesting that the Greenland ice sheet was connected to the North American ice sheet in Canada during the last glacial. By including Canada in the modelling domain this thinning in the early Holocene as the connecting ice bridge broke up will be investigated. 

How to cite: Eckert, M. K., Lauritzen, M., Rathmann, N., Solgaard, A., and Hvidberg, C.: Reconstructing the Greenland ice Sheet during the last two deglaciations, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-10256, https://doi.org/10.5194/egusphere-egu24-10256, 2024.

EGU24-12773 | ECS | Orals | CR2.2

Improved treatment of snow over ice sheets in the NASA GISS climate model: towards ice sheet–climate coupling 

Damien Ringeisen, Patrick Alexander, Lettie Roach, Ken Mankoff, and Igor Aleinov

Representing the interactions between ice sheets and climate is essential for more accurate prediction of climate change and sea level rise. Ice sheets interact with the overlying atmosphere via the accumulation of snow and its compaction into firn, then ice, as well as the melting of surface snow and ice and the creation of runoff water. Getting an adequate representation of heat transfer, compaction, and melting processes is essential for an accurate representation of snow on land ice in global climate models. We are implementing an improved snow model on top of land ice as part of an effort to couple the NASA GISS climate model with the PISM ice sheet model. The new snow model includes additional layers and processes that are not currently incorporated (e.g., liquid water retention, percolation and refreezing, and snow densification), and mass and energy transfer methods that are consistent with both static ice sheets (with implicit iceberg fluxes) and interactive ice sheets (with explicit dynamics). We are tuning the densification scheme of this snow model with temperature and density data from common FirnCover and SumUp observations at locations in the accumulation zone of Greenland, and we compare the resulting density profiles to other SumUp density profiles in Greenland and Antarctica. We will assess the impact of this new snow model in climate model simulations with a static ice sheet compared with the previous (simpler) 2-layer snow model. Finally, we aim to use the non-coupled simulations as a baseline to assess the impact of dynamic coupling with an interactive ice sheet model.

How to cite: Ringeisen, D., Alexander, P., Roach, L., Mankoff, K., and Aleinov, I.: Improved treatment of snow over ice sheets in the NASA GISS climate model: towards ice sheet–climate coupling, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-12773, https://doi.org/10.5194/egusphere-egu24-12773, 2024.

EGU24-13618 | ECS | Orals | CR2.2

Reconstructing the coupled Greenland Ice Sheet–climate evolution during the Last Interglacial warm period 

Matt Osman, Jessica Tierney, and Marcus Lofverstrom

During the Last Interglacial (LIG), approximately 130-118 thousand years ago (ka), the Arctic experienced relative warmth and global sea levels considerably higher than modern.  While this interval is thus considered key for understanding long-term ice–climate feedbacks under warm-state climate conditions, large uncertainties remain surrounding i. the magnitude and spatial expression of LIG global temperature change, ii. the relative contributions of the Antarctic vs. Greenlandic Ice Sheets (GrIS) to LIG sea level rise, and iii. the sensitivity of the GrIS to centennial- to millennial-scale ocean-atmospheric forcing.  Here, we present, to our knowledge, a first attempt at reconstructing the coupled GrIS–climate evolution during the LIG using an internally consistent offline “paleoclimate data assimilation” approach.  Our methodology combines a newly compiled database of nearly 400 chronologically consistent marine geochemical and ice sheet-derived climate-proxy records (spanning 250 sites globally) with recently developed, state-of-the-art transient simulations of the LIG using the coupled Community Earth System Model v2 featuring an interactive Community Ice Sheet Model v2 (CESM2-CISM2).  Our preliminary assimilations suggest LIG peak global mean surface warming of +0.1-0.5˚C (±2 range) above the pre-industrial state, arising from enhanced and widespread (>2-5˚C) high Arctic warming.  Leveraging our CESM2-coupled CISM2 results, we further identify a max GrIS contribution of 2.0 (±0.6) meters of sea level rise equivalent at around 125 ka, nearly ~two millennia after peak LIG climate forcing.  These initial results provide a new proxy-model integration framework for reconciling past GrIS contributions to global sea level rise and benchmark the potential long-term sensitivity of the GrIS to ongoing Arctic warming.

How to cite: Osman, M., Tierney, J., and Lofverstrom, M.: Reconstructing the coupled Greenland Ice Sheet–climate evolution during the Last Interglacial warm period, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-13618, https://doi.org/10.5194/egusphere-egu24-13618, 2024.

Mass loss from ice sheets under the ongoing anthropogenic warming episode is a major source for sea-level rise. Due to the slow responses of ice sheets to changes in atmospheric and oceanic boundary conditions, ice sheets are projected to undergo further retreat as the climate reaches a new equilibrium, producing a long-term commitment to future sea-level rise that is fulfilled on multi-millennial scale. Future projections of ice sheets beyond 2100 have routinely employed end-of-the-century atmosphere-ocean conditions from climate model output under specified emission scenarios. This approach, however, does not account for long-term responses of the climate system to external forcings. Here we analyze the long-term atmospheric and oceanic responses to a variety of emission scenarios in several climate models and show that polar climates may see substantial changes after the atmospheric CO2 level stabilizes. With a 3-D ice sheet model, we demonstrate that the long-term climate responses are crucial for evaluating ice sheets' commitment to future sea-level rise.

How to cite: Li, D.: Effects of long-term climate responses on ice sheets' commitment to future sea-level rise, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-15160, https://doi.org/10.5194/egusphere-egu24-15160, 2024.

EGU24-15323 | ECS | Posters on site | CR2.2

Investigating the evolution and stability of the Greenland ice sheet using remapped surface mass balance forcing 

Charlotte Rahlves, Heiko Goelzer, and Michele Petrini

Surface mass balance (SMB) forcing for projections of the future evolution of the Greenland ice sheet with stand-alone modeling approaches is commonly produced on a fixed ice sheet geometry. As changes of ice sheet geometry become significant over longer time scales, conducting projections for the long-term evolution and stability of the Greenland ice sheet usually requires a coupled climate-ice sheet modeling setup. In this study we use an SMB remapping procedure to capture the first order feedbacks of a coupled climate-ice sheet system with a stand-alone modeling approach. Following a remapping procedure originally developed to apply SMB forcing to a range of initial ice sheet geometries (Goelzer et al., 2020), we produce SMB forcing that adapts to the changing ice sheet geometry as it evolves over time. SMB forcing from a regional climate model is translated from a function of absolute location to a function of surface elevation depending on 25 regional drainage basins, thereby reducing biases that would arise by applying the SMB derived from a fixed ice sheet geometry. We use forcing for different emission scenarios from the CMIP6 archive to compare results from the remapping approach with results from commonly used methods of parameterizing the SMB-height feedback, as well as with results from a semi-coupled climate-ice sheet simulation.

How to cite: Rahlves, C., Goelzer, H., and Petrini, M.: Investigating the evolution and stability of the Greenland ice sheet using remapped surface mass balance forcing, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-15323, https://doi.org/10.5194/egusphere-egu24-15323, 2024.

EGU24-15401 | Posters on site | CR2.2

Development and implementation of a refined climate index forcing for paleo ice-sheet modeling applications  

Antoine Hermant, Christian Wirths, and Johannes Sutter

The contribution of the Antarctic Ice Sheet (AIS) to sea-level rise under future scenarios remains uncertain. Simulations of the AIS covering past-climate periods provide valuable insights into its response to a range of climatological background states and transitions, as well as its past contributions to sea-level change. However, data to constrain the modelled ice-flow and the paleo-climate forcing is often lacking, leading to considerable uncertainties with respect to paleo ice sheet evolution. Here, we implement and test a framework to provide paleo-climate scenarios for continental scale ice sheet models. Our approach involves the use of an improved climate index based on ice-core records to translate paleo forcing snapshots from Earth System Models and regional models into transient paleo-climate scenarios, specifically to simulate the dynamics of the AIS throughout the last glaciation and deglaciation. Additionally, we refine paleo-accumulation scenarios by introducing a regionally-specific and temperature-dependant scaling of accumulation. Our study aims to enhance our understanding of AIS dynamics on glacial-interglacial time-scales and provide improved paleo-informed initializations for AIS projections. 

How to cite: Hermant, A., Wirths, C., and Sutter, J.: Development and implementation of a refined climate index forcing for paleo ice-sheet modeling applications , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-15401, https://doi.org/10.5194/egusphere-egu24-15401, 2024.

EGU24-15987 | ECS | Posters on site | CR2.2

Assessing Antarctic Ice Sheet dynamics under temporary overshoot and long-term temperature stabilization scenarios   

Emma Spezia, Fabrice Kenneth Michel Lacroix, Vjeran Visnjevic, Christian Wirths, Antoine Hermant, Thomas Frölicher, and Johannes Sutter

Current projections of Antarctic Ice Sheet dynamics during the next centuries are subject to large uncertainties both reflecting the ice sheet model setup as well as the climate pathways taken into consideration. Assessing both we present model projections of the Antarctic Ice Sheet’s response to a range of temporary temperature overshoot and stabilization scenarios until the year 2500 accounting for various ice sheet sensitivities. We employ the ice sheet model PISM at continental scale forced by Earth system model data tailored to specific global temperature scenarios via an adaptive greenhouse gas emissions approach. These scenarios reflect both emission pathways which result in a transient temperature overshoot during the 21st and 22nd century as well as stabilization of global temperatures without overshoot. We contrast these simulations with the well- known CMIP6 scenarios to illustrate the diverse potential pathways of Antarctic Ice Sheet dynamics under uncertain future climate trajectories. 

How to cite: Spezia, E., Lacroix, F. K. M., Visnjevic, V., Wirths, C., Hermant, A., Frölicher, T., and Sutter, J.: Assessing Antarctic Ice Sheet dynamics under temporary overshoot and long-term temperature stabilization scenarios  , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-15987, https://doi.org/10.5194/egusphere-egu24-15987, 2024.

EGU24-16455 | ECS | Posters on site | CR2.2

Ice-dammed lake-glacier interactions: Modelling the impact on Fennoscandian Ice Sheet retreat 

Ankit Pramanik, Sarah Greenwood, Carl Carl Regnéll, and Richard Gyllencreutz

Ice-dammed lakes expedite glacier retreat, leading to the expansion of lakes and an elevated risk of Glacial Lake Outburst Floods (GLOFs), and delay the freshwater inflow to the ocean. The escalating number of ice-dammed lakes in Greenland, High Mountain Asia, and Patagonia, driven by the swift retreat of glaciers amid rapid warming, poses a significant threat of natural disasters. In the geological record, evidence indicates the rapid retreat of the Fennoscandian ice sheet, marked by the formation, expansion, and drainage of large (10s-1000s km2 surface area and up to 100s m deep) ice-dammed proglacial lakes along the entire length of the late-deglacial ice margin. The deglaciation and ice-lake interactions of the Fennoscandian Ice Sheet (FIS) provide a valuable analogue for projecting the future retreat of the Greenland ice sheet, where a manifold increase in the number and volume of ice-dammed lakes is anticipated.

Despite extensive research on marine-terminating glaciers, the dynamics of lacustrine-terminating glaciers remain poorly understood. While there are some notable differences in thermo-mechanical processes between marine and lacustrine glaciers, a significant contrast lies in the fact that the calving of lake-terminating glaciers is governed by the stress balance induced by rapidly fluctuating lake levels and thermodynamics inherent of lakes. Our study delves into accessing the impact of critical factors, such as lake size and bathymetry, on the retreat of the Fennoscandian Ice Sheet, using the Ice-sheet and Sea-level System Model (ISSM). Furthermore, we aim to evaluate the influence of calving, subaqueous melt, and rapidly fluctuating lake levels on the FIS retreat. The model's accuracy will be ensured through calibration and validation against geologically reconstructed ice sheet boundaries and lake levels.

How to cite: Pramanik, A., Greenwood, S., Carl Regnéll, C., and Gyllencreutz, R.: Ice-dammed lake-glacier interactions: Modelling the impact on Fennoscandian Ice Sheet retreat, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-16455, https://doi.org/10.5194/egusphere-egu24-16455, 2024.

EGU24-16702 | ECS | Posters on site | CR2.2

Isochronally constrained ice flow evolution of Dronning Maud Land, Antarctica during the Last Glacial Period 

Vjeran Visnjevic, Julien Bodart, Antoine Hermant, Christian Wirths, Emma Spezia, and Johannes Sutter

To improve the robustness of future simulations of ice flow across the Antarctic continent as well as the projections of sea-level rise accompanying it, it is necessary to improve our understanding of the past evolution of ice dynamics. This is specially the case considering the lack of constraints on climate and basal conditions on the regional scale. To address this, we use high resolution regional ice flow modeling combined with radar obtained repositories of internal reflection horizons and ice core data, to constrain the ice flow evolution of both grounded and floating ice across the Dronning Maud Land during the Last Glacial Period. Combining the modeling results obtained using the ice sheet model PISM with radar and ice core data will enable us to improve our knowledge of conditions at the ice base, but also provide an opportunity to test and compare a range of potential climate reconstructions. The presented workflow will further be expanded to other basins in Antarctica as well as to the interglacial-glacial transition, and the results will be used to improve future simulations of ice flow across Antarctica.

How to cite: Visnjevic, V., Bodart, J., Hermant, A., Wirths, C., Spezia, E., and Sutter, J.: Isochronally constrained ice flow evolution of Dronning Maud Land, Antarctica during the Last Glacial Period, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-16702, https://doi.org/10.5194/egusphere-egu24-16702, 2024.

EGU24-17391 | ECS | Orals | CR2.2

Critical thresholds of the Greenland Ice Sheet from the LGM to the future 

Lucía Gutiérrez-González, Jorge Alvarez-Solas, Marisa Montoya, Ilaria Tabone, and Alexander Robinson

In recent decades the Greenland Ice Sheet (GrIS) has undergone accelerating ice-mass loss. The GrIS is thought to be a tipping element of the Earth system due to the existence of positive feedbacks with the climate. Previous work has shown threshold behavior in the system, and its stability has been studied in a range of temperatures of the present to a global warming of +4K. However, there is still no consensus on the values of its critical thresholds for the future. Furthermore,  its stability at  lower temperatures hasn’t been studied yet. Here we use the ice-sheet model Yelmo coupled with the regional climate model REMBO and a parametrization of the ice-ocean interactions to obtain the bifurcation diagram of the GrIS from temperatures representative of the LGM (-12K) to a warmer scenario (+4K). The preindustrial simulated equilibrium volume is larger than the observations, a feature common to many other ice-sheet models. This could indicate model biases, but also that the GrIS could currently not be fully in equilibrium with the preindustrial forcing, with implications for future projections. To investigate this issue, we simulated the transient evolution of the GrIS since the LGM to the present day in the context of the bifurcation diagram, with equilibrium states acting as attractors. 

How to cite: Gutiérrez-González, L., Alvarez-Solas, J., Montoya, M., Tabone, I., and Robinson, A.: Critical thresholds of the Greenland Ice Sheet from the LGM to the future, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-17391, https://doi.org/10.5194/egusphere-egu24-17391, 2024.

EGU24-18501 | Posters on site | CR2.2

Protocol for a Last Interglacial Antarctic ice-sheet model inter-comparison 

Lauren Gregoire, Maxence Menthon, Edward Gasson, and Louise Sime

During the last interglacial, geological records show evidence that the sea level peaked between 6 and 9 m above pre-industrial sea level, with a major contribution from the Antarctic ice sheet. However, ice-sheet models give a very large range of values due to a lack of understanding of the mechanisms leading to the Antarctic ice sheet retreat during the Last Interglacial

Here, we propose a protocol to apply systematically to multiple ice-sheet models to better understand the climate and ice-sheet model uncertainties as well as mechanisms leading to a smaller Antarctic ice sheet. We present the climate forcing choices and methodology, ice-sheet model requirements and the group of simulations suggested. The protocol includes transient penultimate deglaciation and last interglacial equilibrium simulations to make it accessible to all types of ice-sheet models. The protocol includes also sensitivity experiments such as hosing.

Inputs from the community are welcome to improve the protocol under development and make it relevant to all ice-sheet modelling groups interested in participating!

How to cite: Gregoire, L., Menthon, M., Gasson, E., and Sime, L.: Protocol for a Last Interglacial Antarctic ice-sheet model inter-comparison, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-18501, https://doi.org/10.5194/egusphere-egu24-18501, 2024.

EGU24-19165 | Posters on site | CR2.2

Oceanic gateways in Antarctica - Impact of relative sea-level change on sub-shelf melt 

Moritz Kreuzer, Torsten Albrecht, Lena Nicola, Ronja Reese, and Ricarda Winkelmann

Relative sea level (local water depth) on the Antarctic continental shelf is changing by the complex interplay of processes associated with Glacial Isostatic Adjustment (GIA). This involves near-field visco-elastic bedrock displacement and gravitational effects in response to changes in Antarctic ice load, but also far-field interhemispheric effects on the sea-level pattern. On glacial time scales, these changes can be in the order of several hundred meters, potentially affecting the access of ocean water masses at different depths to Antarctic grounding lines and ice sheet margins. Due to strong vertical gradients in ocean temperature and salinity at the continental shelf margin, basal melt rates of ice shelves could change significantly just by variations in relative sea level alone.
Based on a coupled ice sheet – GIA model setup and the analysis of bathymetric features such as troughs and sills that regulate the access of open ocean water masses onto the continental shelf (oceanic gateways), we conduct sensitivity experiments to derive maximum estimates of Antarctic basal melt
rate changes, solely driven by relative sea-level variations.
Under Last Glacial Maximum sea-level conditions, this effect would lead to a substantial decrease of present-day sub-shelf melt rates in East Antarctica, while the strong subsidence of bedrock in West Antarctica can lead up to a doubling of basal melt rates. For a hypothetical globally ice-free sea-level
scenario, which would lead to a global mean (barystatic) sea-level rise of around +70 m, sub-shelf melt rates for a present-day ice sheet geometry can more than double in East Antarctica, but can also decrease substantially, where bedrock uplift dominates. Also for projected sea-level changes at the
year 2300 we find maximum possible changes of ±20 % in sub-shelf melt rates, as a consequence of relative sea-level changes only.

How to cite: Kreuzer, M., Albrecht, T., Nicola, L., Reese, R., and Winkelmann, R.: Oceanic gateways in Antarctica - Impact of relative sea-level change on sub-shelf melt, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-19165, https://doi.org/10.5194/egusphere-egu24-19165, 2024.

EGU24-20197 | ECS | Posters on site | CR2.2

Constraining projections of future freshwater fluxes from Antarctica  

Violaine Coulon, Javier Blasco, Qing Qin, Jan De Rydt, and Frank Pattyn

As global temperatures rise, Antarctica's grounded ice sheet and floating ice shelves are experiencing accelerated mass loss, releasing meltwater into the Southern Ocean. This increasing freshwater discharge poses significant implications for global climate change. Despite these consequences, interactive ice sheets and ice shelves have generally not been included in coupled climate model simulations, such as those in CMIP6. Consequently, CMIP6 projections lack a detailed representation of spatiotemporal trends in ice-sheet freshwater fluxes and their impact on the global climate system, introducing major uncertainties in future climate and sea-level projections. To address this, we provide future Antarctic freshwater forcing data and uncertainty estimates for climate models. These are derived from an ensemble of historically calibrated standalone ice sheet model projections, produced with the Kori-ULB ice flow model, under different climate scenarios up to 2300. Here, we analyse spatiotemporal trends in calving rates, ice shelf basal melt and surface mass balance for all Antarctic ice shelves. 

How to cite: Coulon, V., Blasco, J., Qin, Q., De Rydt, J., and Pattyn, F.: Constraining projections of future freshwater fluxes from Antarctica , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-20197, https://doi.org/10.5194/egusphere-egu24-20197, 2024.

EGU24-20332 | ECS | Orals | CR2.2

The effect of Pacific climatology on the North American Ice Sheet at the Last Glacial Maximum 

William J. Dow, Sam Sherriff-Tadano, Lauren J. Gregoire, and Ruza Ivanovic

Surface ocean conditions and atmospheric dynamics can affect the surface mass balance (SMB) of remote ice sheets via their influence on heat and moisture transport. Here, we use the FAMOUS-ice coupled climate-ice sheet model, coupled to a slab ocean, to simulate the Last Glacial Maximum (LGM). The model was run hundreds of times to produce a large ensemble that captures a range of uncertain model inputs (parameter values). We investigate the range of simulated atmospheric circulation patterns in the 16 ‘best’ ensemble members based on constraints, such as global temperature, their relationship to sea surface conditions in the North Pacific and the interactions with the North American ice sheet. We present evidence of upper tropospheric planetary waves that facilitate communication between the tropical Pacific and extratropical Laurentide ice sheet region, yet there are clear differences in upper tropopsheric dynamics when compared to recent historical period. There is limited evidence for this tropical-extra-tropical relationship being directly responsible for regional differences in Laurentide SMB evolution.

How to cite: Dow, W. J., Sherriff-Tadano, S., Gregoire, L. J., and Ivanovic, R.: The effect of Pacific climatology on the North American Ice Sheet at the Last Glacial Maximum, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-20332, https://doi.org/10.5194/egusphere-egu24-20332, 2024.

EGU24-21079 | Orals | CR2.2

Understanding conditions leading to WAIS collapse, from the Last Interglacial to the modern 

Mira Berdahl, Gunter Leguy, Eric Steig, William Lipscomb, Bette Otto-Bliesner, Nathan Urban, Ian Miller, and Harriet Morgan

It is virtually certain that the West Antarctic Ice Sheet (WAIS) collapsed during past warm periods in Earth’s history, prompting concerns about the potential recurrence under anthropogenic climate change. Despite observed ice shelf thinning in the region, the combination of climate forcing and ice sheet sensitivity driving these changes remains unclear. Here, we investigate the joint effects of climate forcing and ice sheet sensitivity to evaluate conditions leading to WAIS collapse. We run ensembles of the Community Ice Sheet Model (CISM), spun up to a pre-industrial state, and apply climate anomalies from the Last Interglacial (LIG, 129 to 116 yr ago), and the future (SSP2-4.5).  Forcing is derived from Community Earth System Model (CESM2) global simulations. We find that only modest ocean warming is required to cause significant WAIS mass loss, though such loss takes multiple centuries to millennia to manifest.

How to cite: Berdahl, M., Leguy, G., Steig, E., Lipscomb, W., Otto-Bliesner, B., Urban, N., Miller, I., and Morgan, H.: Understanding conditions leading to WAIS collapse, from the Last Interglacial to the modern, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-21079, https://doi.org/10.5194/egusphere-egu24-21079, 2024.

EGU24-651 | ECS | Posters on site | HS7.1

Multi-scale comparison of rainfall measurement in Paris area between two optical disdrometers of different working principles 

Marcio Matheus Santos de Souza, Auguste Gires, and Jerry Jose

A disdrometer is an instrument designed to assess both the size and velocity of descending hydrometeors. The applications of rainfall measurements retrieved with the help of disdrometers are diverse, spanning areas such as traffic control, scientific research, airport observation systems, and hydrology. Modern disdrometers leverage microwave or laser technologies that have increased the accuracy of the measurements with each iteration. Still, the quality of measurements fluctuates depending on factors such as raindrop size, wind velocity, and rain rate. A comprehension of these variations is needed to better understand the level of reliability of each device depending on the specific rain conditions.

In this study, we compare the performance of two optical disdrometers : 3D Stereo disdrometer (manufactured by Thies Clima) and Parsivel2 (manufactured by OTT). Both devices provide size resolved measurement of rainfall along with velocity of falling drops. Parsivel is set to record data every 30 seconds over a sampling area of 54 cm² and arranges the information in 32 x 32 classes of drop size and velocity. Unlike the Parsivel, 3D Stereo does not discretize measurements, and directly provides the diameter and velocity of each falling drop in a sampling area of 100 cm² with a measuring resolution of 0.08 mm and 0.2 m/s respectively, and a temporal resolution of 1 millisecond. This finer resolution data enables us to study rainfall variability at very small scales which are not usually available.

Here, we used continuously and simultaneously measured data since 21/08/2023, from TARANIS observatory of ENPC (https://hmco.enpc.fr/portfolio-archive/taranis-observatory/). The initial comparison of the data was done using a time series of rain-rate for rainfall events in between a dry period of at least 15 minutes and total depth >0.7 mm. This revealed an unexpected disparity in the water volume collected between the devices. Parsivel collected more than 3D Stereo on every instance, and the disparity got bigger as the rain rate increased. With the purpose of studying the source of this disparity, the sampling area of the 3D Stereo was divided into 8 sections and compared with each other. This showed that the estimate of rainfall parameters such mean diameter, mean velocity of the drops (which were expected to be uniform over long periods regardless of the section where drops are measured) were not the same for the sections studied, and exhibited clear trends. To understand this discrepancy in a scale invariant way, and to evaluate the performance of devices across scales and not only at a single scale, the widely used framework for studying variability of geophysical fields – Universal Multifractals (UM) was employed for assessing the scaling behavior of fields. Rainfall from both devices showed previously reported average scaling behavior from 30 s to 30 min. The difference between rain events and also the behavior at finer scales, which can be accessed from 3D stereo disdrometer were also studied using the UM framework and will be discussed.

Authors acknowledge the Ra2DW project (supported by the French National Research Agency - ANR-23-CE01-0019), for partial financial support.

Keywords: rainfall; disdrometer; multifractals;

How to cite: Santos de Souza, M. M., Gires, A., and Jose, J.: Multi-scale comparison of rainfall measurement in Paris area between two optical disdrometers of different working principles, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-651, https://doi.org/10.5194/egusphere-egu24-651, 2024.

EGU24-2655 | Posters on site | HS7.1

Designing the TUDS rainfall observatory in northern Ghana 

Nick van de Giesen, Frank Annor, Sylvester Ayambila, Richard Dogbey, Vincent Hoogelander, Gordana Kranjac-Berisavljevic, Kingsley Kwabena, Rob Mackenzie, Marc Schleiss, and Remko Uijlenhoet

Convective rainfall in West Africa is poorly monitored and understood. There are large gaps between remote sensing rainfall products and what is observed on the ground. There are several reasons for these gaps. First, satellites and rain gauges measure at very different scales so one would expect that remote sensing products contain more events at lower intensities than small gauges. Second, a lot happens between the clouds observed by satellites and the ground. Rainfall may evaporate and move with the wind, causing further disconnects between space and ground observations. There are also indications that clouds in West Africa contain many small drops due to the presence of many aerosols, thereby possibly “misleading” satellite products. Finally, it is likely that there are further factors that are not yet accounted for.

In order to tackle this disconnect between ground and space observations, we plan to build the TUD - UDS, or TUDS, rainfall observatory near Tamale and Nyankpala in northern Ghana. The following are initial ideas that we would like to discuss at the EGU. It will be a multi-scale observatory, starting at a grid of nine gauges on a 500m grid (1km x 1km total). This small grid should capture the inherent spatial variability of convective rainfall events with convective cells of 2km or less. The largest grid would also contain nine gauges and have an extent of 10km x 10km, or larger. This outer grid would capture the movement of convective cells, including those contained within so-called line squalls. An intermediate grid may complete this picture. The structure will look, more or less, like the one in the picture below.

Different instruments will be at our disposal, from simple totalling rain gauges to disdrometers. There will be five Thies disdrometers, one Ott Parsivel, and several TAHMO stations and/or tipping bucket rain gauges. Also experimental intervalometers will be placed in the grid to better understand rainfall structure over time and space. Several instruments will be co-located to examine strengths and weaknesses of the different methods.

We explicitly invite comments and contributions.  

 

TEMBO Africa: The work leading to these results has received funding from the European Horizon Europe Programme (2021-2027) under grant agreement n° 101086209. The opinions expressed in the document are of the authors only and no way reflect the European Commission’s opinions. The European Union is not liable for any use that may be made of the information.

How to cite: van de Giesen, N., Annor, F., Ayambila, S., Dogbey, R., Hoogelander, V., Kranjac-Berisavljevic, G., Kwabena, K., Mackenzie, R., Schleiss, M., and Uijlenhoet, R.: Designing the TUDS rainfall observatory in northern Ghana, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-2655, https://doi.org/10.5194/egusphere-egu24-2655, 2024.

Precipitation droplets are influenced by environmental fields and transform in time and space, following cloud microphysical processes. Accordingly, a raindrop size distribution (DSD) changes shape in a various form. However, DSDs cannot be calculated directly in radar or bulk models and are expressed using an approximate function. Exponential and gamma distribution are well-known as approximation functions, but there are DSDs of shapes that cannot be represented by these functions. One of them is a bimodal DSD with two peaks. Previous modeling studies have indicated that the bimodal DSD is formed when the collision-breakup process reaches equilibrium. On the other hand, recent observation-based studies have discussed the influence of convective activity within the precipitation system on forming the bimodal DSD. However, observations have not been able to quantitatively study the microphysical changes of individual particles and have yet to reveal the formation mechanisms within the precipitation system. In this study, we investigated quantitatively the process of the formation of the bimodal DSD by two-dimensional simulation of multicellular convection with the bin method. The simulation results showed that the bimodal DSD was formed during the updraft and downdraft in the mature stage of the multicell. Additionally, the bimodal DSD was formed at lower altitudes where there was inflow into the precipitation system. Particles that constituted the maximum of the bimodal DSD were found to have been advected by the inflow. Particles that constituted the local maximum dropped against the updraft. In contrast to these, particles that constituted the local minimum were less affected by the inflow and had difficulty dropping against the updraft. These results suggested that the bimodal DSD was formed by horizontal and vertical size sorting because of inflow and updrafts in the mature multicellular convection. In the future, it is necessary to simulate the reproduction of observed cases and compare them with observations.

How to cite: Okazaki, M., Yamaguchi, K., Yanase, T., and Nakakita, E.: Spatiotemporal structure of raindrop size distribution due to flow field in a convective precipitation system simulated by bin cloud microphysics model., EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-6387, https://doi.org/10.5194/egusphere-egu24-6387, 2024.

EGU24-6767 | Posters on site | HS7.1

Microphysical properties of the stratiform precipitation in Kyiv city based on OTT Parsivel2 and pluviograph data  

Svitlana Krakovska, Liudmyla Palamarchuk, and Anastasiia Chyhareva

Precipitation detailed characteristics, namely spectrum of particles by their sizes, phase and precipitation intensity with high-resolution timestep, still need to be investigated due to the complexity of their direct instrumental measurements but necessity for improving forecast for different applications including hydrological and emergency service. Our study is focused on the stratiform precipitation associated with cloud system (Ns-As) of warm front during prolonged and intense precipitation event on the 25th October 2023 in Kyiv, Ukraine. This warm front cloud system was connected with an occluded low over Poland which developed on the East periphery of a huge depression (970 hPa) over the Northern Atlantic.

We analyzed the OTT Parsivel² - Laser Weather Sensor measurement data with 10sec time steps. Parsivel² was installed nearby regular meteorological station, which is a part of the WMO network, and its measurements were used for verification. Precipitation intensity and raindrop distributions had wavy character, where we can distinguish a few waves of precipitation enhancement. The average intensity of the minimum wave was 0.02mm/min that corresponds to 30 raindrops with size varying from 0.5 to 1.5mm and maximum falling speed 4m/s for the largest raindrops. The average intensity of maximum precipitation enhancement wave was 0.15mm/min with around 100 raindrops per 10sec with sizes mainly from 0.5 to 2.5mm (with some raindrop sizes up to 3.5mm) and average falling speed 5-6m/s. Total amount of 26-hour precipitation event was 24.2mm according to OTT Parsivel² measurements and 26mm according to SYNOP data from Kyiv WMO station (ID 33345). We should note that in modern climate condition in Kyiv such prolonged frontal precipitation even in autumn is rather rare event in respect to previous decades.  

Gained results were compared with previous studies based on 20-year measurement by pluviograph at the same Kyiv WMO station. For stratiform precipitation, average maximum precipitation intensity within precipitation enhancement waves was around 0.11mm/min. Duration of main precipitation enhancement waves was around 21 minutes. Characteristics of precipitation enhancements waves are key for assessment of surface runoff value. The significant fraction of water on the ground that forms surface runoff goes mainly from such precipitation enhancement waves, when around 60 up to 90% of the maximum surface runoff can be formed.

In conclusion, OTT Parsivel² Laser Weather Sensor was used in Ukraine for the first time and demonstrated good performance versus the city station accumulation measurements and historical pluviograph data at the station. At the moment this instrument is under way to the Ukrainian Antarctic station Akademik Vernadsky where further exploitation will allow to test and obtain measurement data for different phase of precipitation, mostly mixed and solid and compare with data from Micro Rain Radar Pro. Obtained and future results will extend our understanding of precipitation formation, their microphysics and dynamics, interconnections between precipitation intensity and size/fall speed of raindrops and solid particles. Future studies could help to evaluate the transformation of cloud and precipitation formation processes under the climate change for better parameterization in numerical models, to study the microphysical structure and composition of precipitation.

How to cite: Krakovska, S., Palamarchuk, L., and Chyhareva, A.: Microphysical properties of the stratiform precipitation in Kyiv city based on OTT Parsivel2 and pluviograph data , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-6767, https://doi.org/10.5194/egusphere-egu24-6767, 2024.

EGU24-7802 | Posters on site | HS7.1

Cloud based tool to enhance urban resilience with the Fresnel Platform using the Multi-Hydro Model 

Guillaume Drouen, Daniel Schertzer, Auguste Gires, and Ioulia Tchiguirinskaia

The aim of the Fresnel platform of École des Ponts ParisTech is to foster research and innovation in multiscale urban resilience. Studying the hydrological response of such complex urban areas accounting also for small scale spatio-temporal precipitation variability requires adapted tools. For these reasons, RadX provides a user-friendly graphical interface to run simulations using a fully distributed and physically based model: Multi-Hydro. RadX is designed as a Software as a Service (SaaS) platform, allowing users to work with data across a wide range of space-time scales and the appropriate tools for analyzing and simulating this data.

The hydrological model, developed at École des Ponts ParisTech, integrates four open-source software applications previously used and validated independently by the scientific community as well as practitionners. Its modular structure includes a surface flow module, sewer flow module, a ground flow module and a precipitation module. It is able to simulate the quantity of runoff and rainwater infiltrated into unsaturated soil layers from any space-time varying rainfall event at any location of the studied peri-urban watersheds, as well as depth and flow in all the pipes and nodes of the sewer network.

Users can launch hydrological simulations using the Multi-Hydro model directly from their web browser, while they are run on dedicated servers. They can adjust two key input parameters: the land use of the studied catchment and the rainfall data. Dedicated tools have been developed to enable users to modify the land use of the catchment with the same ease as using a raster graphic editor. Users can either choose real rainfall events captured by the X-band weather radar located at École des Ponts ParisTech or utilize user-defined synthetic rainfall as input. Data from other radar can also easily be integrated. 

For the simulation output, the interface provides users with different tools to study in detail the impact of the chosen input parameters. For instance, by simply selecting two sewer junctions on an interactive map, users can generate a sewer path between these two points and display an interactive representation of the water level heights in sewer conduits and junctions along the user-defined sewer network path.

Additional components can be integrated into RadX to meet specific requirements using visual tools and forecasting systems, including those from third parties. Developments are still in progress, with a constant loop of requests and feedback from the scientific and professional world.

How to cite: Drouen, G., Schertzer, D., Gires, A., and Tchiguirinskaia, I.: Cloud based tool to enhance urban resilience with the Fresnel Platform using the Multi-Hydro Model, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-7802, https://doi.org/10.5194/egusphere-egu24-7802, 2024.

EGU24-8714 | ECS | Orals | HS7.1

Improvements in rain gauge design and measurements to minimise under-catch errors 

Mark Dutton and Domenico Balsamo

Precipitation measurements provide historic and near real-time data for Met Services and ground truth references for modelling and forecasting.  Current methods suffer from well-known under-catch problems1.  These are caused by wind effect2 on the gauge, out-splash, evaporation, and internal tipping bucket (‘counting’) errors.  Thereby causing water-balance errors for Hydrology scientists.  Good gauge design and correct siting can minimise these errors but not eliminate them.

Over 10 years of research, into the best aerodynamic shape for a precipitation gauge, was carried out to minimize out-splash and maximize catch3.  Comparison field work1 and Computational Fluid Dynamic4 (CFD) research was undertaken between standard straight-sided, ‘chimney’ shaped, aerodynamic shaped and pit-installed (out of the wind) gauges.  This research demonstrated that it may be possible to quantify under-catch using gauge rim-based wind data, drop-size and drop-type information.  Field comparison between the “new instrument” and pit gauge will be needed.  Once quantified at source, it can then be used to accurately correct live data.

This new instrument uses ultrasonic wind sensors and Doppler-Shift measuring techniques to obtain wind versus rainfall catch data.  Also using optical and/or impact sensing techniques we can measure the individual drop size and count the drops involved in a rain event.  By adding weighing technology to the tipping bucket design and improving calibration methods, we can improve resolution and detect evaporation losses.  Also power efficient and controlled heating to allow the inclusion of solid precipitation measurements.  Then finally use machine learning (ML) techniques to correct the errors.

Therefore, the aim of this project is to design a simple to use intelligent instrument to minimise and possibly eliminate under-catch measurement errors balancing out the water budget.  Allow installation of the instruments at ground and raised levels without increase in errors caused predominately by the wind.  Create near real-time and historic field precipitation data, both corrected and non-corrected to be use by Met Services and Hydrology modelling scientists.

References

1. Sevruk, B. Methods of correction for systematic error in point precipitation measurement for operational use, World Meteorological Organization - Operational Hydrology, Report No. 21, 1982.

2. Pollock, M. D., et al. Quantifying and mitigating wind induced undercatch in rainfall measurements, Water Resources Research, 54, 2018.

3. Strangeways, Ian. Improving precipitation measurement. International Journal of Climatology. 24. 1443 - 1460. 10.1002/joc.1075, 2004.

4. Colli, M., et al.  A Computational Fluid-Dynamics Assessment of the Improved Performance of Aerodynamic Rain Gauges. Water Resources Research. 54. 10.1002/2017WR020549, 2018.

How to cite: Dutton, M. and Balsamo, D.: Improvements in rain gauge design and measurements to minimise under-catch errors, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-8714, https://doi.org/10.5194/egusphere-egu24-8714, 2024.

EGU24-8789 | ECS | Orals | HS7.1

Merging personal weather stations with real-time radar rainfall estimates at the catchment scale 

Nathalie Rombeek, Markus Hrachowitz, Davide Wüthrich, and Remko Uijlenhoet

Real-time flood forecasting and warning during extreme rainfall events remains challenging since accurate and real-time available data are critical. Nowcasting based on radar rainfall can be utilized for this, as it has a high spatial and temporal resolution (i.e. typically 1 km and 5 min). However, the quantitative precipitation estimates (QPE) from the radar, upon which radar rainfall nowcasting is based, often contains substantial uncertainty and bias. While the QPE are usually corrected with official rain-gauge networks, these networks are sparse, and not always available in (near) real-time.

Instead, personal weather stations (PWS) can be used, as they have a much higher density and are available in real time. While PWS are prone to several sources of error, quality control algorithms can be used to improve their accuracy. Previous research already showed that merging quality controlled PWS with radar rainfall estimates reduces the underestimation for 1-hour accumulated rainfall at the pan-European scale. However, this has not yet been investigated at the catchment scale. This research aims to investigate the potential of merging PWS data with radar rainfall estimates for different catchments in the Netherlands, by considering multiple rainfall events starting from 2018. The goal is to quantify the performance in relation to rainfall type, quality control algorithms and catchment properties, validated against the climatological gauge-adjusted radar dataset from the KNMI. The insights obtained from this research have the potential to be utilized for real-time radar rainfall nowcasting and consequently flood forecasting.

How to cite: Rombeek, N., Hrachowitz, M., Wüthrich, D., and Uijlenhoet, R.: Merging personal weather stations with real-time radar rainfall estimates at the catchment scale, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-8789, https://doi.org/10.5194/egusphere-egu24-8789, 2024.

EGU24-10819 | Orals | HS7.1

Spatial and temporal structure of normal and extreme rainfall 

András Bárdossy

The space time behaviour of precipitation is very complex. The knowledge of the dependence structures in space and time is very important for the assessment of flood risks. In this contribution the dependence structures of normal and extreme events are compared. Both rain gauges with high temporal resolution and radar images are investigated. Spatial and temporal copulas are used for this investigation. Due to the large number of zero observations, especially for short temporal aggregations an indicator approach is used to detect structural differences. The results show, that the temporal dependence structure of rainfall gradually changes with increasing intensity. Similar behaviour can be detected for the spatial structure with the addition of advection related differences in both ranges and angles of anisotropy. The findings indicate that metagaussian approaches which only consider spatial and temporal correlations are not appropriate for the description and the simulation of rainfall extremes. Finally a new structural simulation method using non-Gaussian dependence is presented.

How to cite: Bárdossy, A.: Spatial and temporal structure of normal and extreme rainfall, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-10819, https://doi.org/10.5194/egusphere-egu24-10819, 2024.

EGU24-12007 | Orals | HS7.1

Revisiting nonterminal hydrometeors: Refining instrument uncertainty 

Michael Larsen, Andrei Vakhtin, and Anthony Gomez

The fall velocities of rain and drizzle drops are often assumed to be a deterministic function of their size. These diameter-fall speed relationships are intrinsically assumed in the retrievals provided by some commercial rain measurement instruments (e.g. the Joss-Waldvogel Disdrometer (Distromet), Micro Rain Radar (METEK), and 1-Dimensional Video Disdrometer (Joanneum Research)).

Some disdrometers are capable of independently measuring droplet size and fall-speed and provide evidence that not all drops adhere to the assumed size/fall-speed relationship. The ubiquity and magnitude of these deviations are still an area of some debate; clear identification of drizzle and rain drops falling at speeds different than their expected terminal fall velocities is muddied by conservative estimates of disdrometer resolution and performance. For a long time the bulk of observed non-terminal drop fall speeds were assumed to be instrumental artifacts and, even now, most investigators conclude drops falling at non-terminal speeds do not have a large impact on rain measurement science.

To date, uncertainties in disdrometer-derived drop sizes and fall speeds have usually been derived from the manufacturer estimates. Here, we improve on these estimates by using a field calibration source (the new ``Large Drop Generator'' from Mesa Photonics) that permits user-selectable generation of droplets with known sizes and fall speeds. From these data, empirical estimates of disdrometer sizing and fall velocity bias and uncertainty can be determined. This, then, allows for a more reliable estimate of the fraction of non-terminal drops in natural rain and a more reliable assessment of the impact of non-terminal drizzle and rain drops in data derived from instruments that assume a specific drop size/fall-speed relationship.

How to cite: Larsen, M., Vakhtin, A., and Gomez, A.: Revisiting nonterminal hydrometeors: Refining instrument uncertainty, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-12007, https://doi.org/10.5194/egusphere-egu24-12007, 2024.

EGU24-12054 | Posters on site | HS7.1

Testing a new Radar QPE methodology for winter events with a low melting layer 

Raquel Evaristo, Ju-yu Chen, Alexander Ryzhkov, and Silke Trömel

The RY precipitation product of the German Weather Service (DWD) is severely affected by the presence of the low melting layer and frequently shows circular features of enhanced precipitation around the radar sites during the winter time. 
The radars tend to be installed at relatively high terrain and to scan at elevations at a minimum of 0.5° in order to avoid beam blockage and ground clutter. In doing so two problems arise:

1) the difference between the ground and the radar beams becomes a problem especially at large distances from the radar, and consequently precipitation processes in the lowest layers are not observed.
2) the radar beam often reaches the melting layer and may even cross it where it is sampling the snow above.As a result problems arise when deriving surface QPE from the radar: regions of enhanced QPE in ring shapes around the radar sites, and underestimation of the precipitation beyond the melting layer.

A new methodology (PVPR - Polarimetric Vertical Profile of Reflectivity) developed by Ryzhkov et al. 2022 is tested here for which the radar reflectivity (ZH) is reconstructed to correct for the effect of the melting layer and snow beyond. In this methodology the melting layer is detected independently for each azimuth based on the values of ZH and ρHV (cross-correlation coefficient between horizontal and vertically polarized radar waves). In particular the range bin at which the melting layer was reached is recorded (mlb_r). The strength of the melting layer (ML_S) is defined based on how much the value of ρHV  dropped within the melting layer. The values of ML_r and ML_S at a specific elevation are considered sufficient to characterize the melting layer, and are then compared with lookuptables which were generated by simulations of the melting layer effect on the radar beam. A correction factor is then applied based on the lookuptables to the ZH profile within and beyond the melting layer. Visually the result shows a smoother field of reflectivity without the obvious bright band and decreased values associated with snow at farther ranges.

In this study the PVPR methodology was used to correct ZH which in turn was used to calculate rain rates and rain accumulations in a few winter events in Germany.  The results show a strong improvement in the quality of the QPE when compared to rain gauges. The quality of the resulting QPE depends on the event and on the location of the radar. More specifically, the quality decreases when the melting layer is very low, at heights comparable to the radar height, and when the difference between the beam and the surface increases. These problems will be analyzed and potential solutions will be tested in order to improve the quality of the rainfall product.

Ryzhkov, Alexander, Pengfei Zhang, Petar Bukovčić, Jian Zhang, and Stephen Cocks. 2022. "Polarimetric Radar Quantitative Precipitation Estimation" Remote Sensing 14, no. 7: 1695. https://doi.org/10.3390/rs14071695 

How to cite: Evaristo, R., Chen, J., Ryzhkov, A., and Trömel, S.: Testing a new Radar QPE methodology for winter events with a low melting layer, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-12054, https://doi.org/10.5194/egusphere-egu24-12054, 2024.

EGU24-12374 | ECS | Posters on site | HS7.1

Implications of the rainfall spatial variability for the real-time modeling of runoff triggering stony debris flows 

Mauro Boreggio, Matteo Barbini, Martino Bernard, Matteo Berti, Massimiliano Schiavo, Alessandro Simoni, Sandivel Vesco Lopez, and Carlo Gregoretti

In a mountainous environment, high-intensity and short-duration precipitation can generate sudden and abundant runoff at the base of rocky cliffs. This runoff, upon impacting the debris deposits present there, can trigger debris-flow phenomena. In the province of Belluno, in the Boite River valley, a network of rain gauges has been set up to monitor precipitation in the Rovina di Cancia site, where 12 debris-flow events have occurred in the last 10 years. The rain gauges are strategically placed both upstream and downstream of the debris-flow initiation area. In most cases, the precipitation showed significant spatial variability in both planimetric and altimetric aspects. This variability is crucial when simulating the runoff that triggers stony debris flows. The simulation of the peak runoff that triggered the 12 occurred events using a single rain gauge presented a high scatter compared to the simulation performed with the spatially recorded rainfall, except when the chosen rain gauge was close to the rocky cliffs. Furthermore, modelling using radar estimates as rainfall input also displayed significant variability based on the rain gauge used to correct the radar data. Essentially, accurate real-time simulation of runoff triggering debris flows requires the presence of rain gauges upstream of the initiation area, particularly in close proximity to the rocky cliffs.

How to cite: Boreggio, M., Barbini, M., Bernard, M., Berti, M., Schiavo, M., Simoni, A., Vesco Lopez, S., and Gregoretti, C.: Implications of the rainfall spatial variability for the real-time modeling of runoff triggering stony debris flows, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-12374, https://doi.org/10.5194/egusphere-egu24-12374, 2024.

EGU24-13111 | ECS | Orals | HS7.1

Identification of wet and dry periods in commercial microwave link observations via information theory framework 

Anna Špačková, Martin Fencl, and Vojtěch Bareš

Commercial microwave links (CML) have already demonstrated their promising potential in rainfall observation and sensing. The CMLs enable indirect monitoring of path-averaged rainfall intensity as the transmitted signal is attenuated along the link path mainly by raindrops. However, the signal is also attenuated during dry weather periods and is affected by both atmospheric and hardware conditions. Faulty separation of wet and dry periods can easily lead to incorrect rainfall estimates and remains challenging to estimate due to irregular fluctuations of the attenuated signal.

This study aims to use information theory approach to estimate wet and dry periods in the CML signal attenuation observation, which is achieved by evaluating individual predictors and combinations of predictors. The method enables any data to be used as predictors without the need for parameters to describe relations between different variables, as the discrete probability distributions are applied. The model that provides the strongest information content to the wet and dry classification is binarized using an optimized threshold and validated. Thiesen et al. (2019) recently applied this approach to identify rainfall-runoff events in discharge timeseries.

Data of non-winter periods between 2014 and 2016 are used with a temporal resolution of 1 minute. For one CML in the Prague network, wet and dry periods were defined manually as reference (target). Predictors included raw CML data (signal attenuation), as well as derived timeseries such as signal attenuation shifted in time, relative magnitude of attenuation, gradient of the signal attenuation and signal deviation. In addition, external predictors such as temperature deviation, rain gauge precipitation observations or synoptic types are used as additional predictors.

By selecting different predictors, it is possible to compare effectiveness in estimating the reference wet and dry periods. Variation in the strength of the relations between the target and the predictors allows ranking the suitability of available predictors and their combinations for the task. Subsequently, having the best performing predictor, it is combined with others and their collective performance was iteratively evaluated to find the most accurate combination of three predictors described in a multidimensional discrete distribution model. The resulting predictor combination was then converted into binary form and validated. A method comparison is performed with separation of constant and moving average baseline attenuation for wet periods identification as well as wet/dry classification using a threshold for rolling standard deviation of the signal.

Having sufficient data amount for data-driven models enables utilizing the relationships within the dataset without being limited by parametric or operational assumptions, which are often embedded part of wet/dry in classification methods.

References
Thiesen, S., Darscheid, P., and Ehret, U.: Identifying rainfall-runoff events in discharge time series: a data-driven method based on information theory, Hydrol. Earth Syst. Sci., 23, 1015–1034, https://doi.org/10.5194/hess-23-1015-2019, 2019.

This work was supported by the Grant Agency of the Czech Technical University in Prague, grant no. SGS23/048/OHK1/1T/11.

How to cite: Špačková, A., Fencl, M., and Bareš, V.: Identification of wet and dry periods in commercial microwave link observations via information theory framework, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-13111, https://doi.org/10.5194/egusphere-egu24-13111, 2024.

EGU24-14062 | Posters on site | HS7.1

Upward transport in a canopy assisted by raindrop impacts on plant leaves 

Tristan Gilet and Loïc Tadrist

The interception of raindrops by plant leaves induces a redistribution of water, nutrients, and micro-organisms, from the surface of these leaves to their surroundings. It consequently shapes the plant ecosystem. For example, in wheat fields (as in most major crops), splashing raindrops are the main mechanism of spore dispersal for fungal diseases at the epidemic stage, with severe consequences on crop yield. Surprisingly, the observed dispersal is not only downward (wash off / dripping) or outward (splash), but also upward, which may considerably speed up the fungus propagation. Other nutrients and microorganisms might also benefit from such upward transport external to the plant.

In this work, we unravel an efficient and universal mechanism of upward transport: after a raindrop splashed on a plant leaf, the residual water on the leaf can be shot upward as the leaf springs back. We illustrate this phenomenon with several plant leaves. Then we present results obtained from systematic experiments with artificial leaves, thanks to which both the mechanics of rain-induced leaf motion and the fluid dynamics of leaf-induced droplet ejections are elucidated. We identify the range of mechanical properties of the leaf that makes upward shooting fully effective. Finally, we show that the efficiency of this upward transport increases more than proportionally with rain intensity. Its occurrence and role in shaping ecosystems will be largely amplified in the case of an increased frequency of extreme rain events.

How to cite: Gilet, T. and Tadrist, L.: Upward transport in a canopy assisted by raindrop impacts on plant leaves, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-14062, https://doi.org/10.5194/egusphere-egu24-14062, 2024.

EGU24-16024 | ECS | Orals | HS7.1

Quantifying precipitation intermittency for Bergen, Norway, from measurements and models across a wide range of time scales 

Ingrid O. Bækkelund, Mari B. Steinslid, and Harald Sodemann

Intermittency of rainfall is an important property, for example in the context of urban flooding. There is currently a lack of information about the ability of numerical weather prediction models to represent precipitation intermittency for different weather situations, in particular at high resolution in space and time. Here we present a new way to quantify rainfall intermittency based on a near-continuous, high-resolution precipitation dataset from Bergen, Norway, one of the rainiest cities in Europe. 

We quantify precipitation intermittency from a precipitation dataset acquired at the Geophysical Institute, Bergen, spanning the period 2019-2022 at a 1 min time resolution. Precipitation rates were obtained from a Total Precipitation Sensor TPS-3100 (Yankee Environmental Systems Inc., USA) and a Parsivel2 disdrometer (OTT Hydromet GmbH, Germany). In addition, we use precipitation output at 1 min resolution from the regional high-resolution weather forecasts model HARMONIE-AROME for selected events. Precipitation intermittency is then identified for a range of minimum inter-event times (MIT) from 1 min to 24 h, and precipitation event durations from 1 min to 33 days. Next, the precipitation events for different intermittencies are related to average meteorological characteristics during the events with respect to air temperature, pressure, wind speed, rain rate and amount, and corresponding weather regimes.  

We compile the intermittency information into a 2-dimensional heat map that can be considered as a characteristic fingerprint for precipitation in Bergen. Particular frequency maxima and minima appear to be related to different precipitation processes and weather regimes. A scale gap between 30 min and 2 h event duration for MIT larger than 12 h indicates that separate factors control precipitation processes at these time scales. Weather regimes show a clear influence on the precipitation characteristics, with a markedly higher probability for long-duration rain events in the zonal flow regime for longer event durations at high MITs compared to the Scandinavian trough regime. A comparison between precipitation intermittency simulated by HARMONIE-AROME shows reasonable agreement with observed event characteristics for events lasting more than 1h, while events with durations of 30 min and less are poorly represented. 

How to cite: Bækkelund, I. O., Steinslid, M. B., and Sodemann, H.: Quantifying precipitation intermittency for Bergen, Norway, from measurements and models across a wide range of time scales, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-16024, https://doi.org/10.5194/egusphere-egu24-16024, 2024.

EGU24-17471 | Orals | HS7.1

Wind-induced bias of catching-type precipitation gauges and their overall collection efficiency 

Luca G. Lanza, Arianna Cauteruccio, and Enrico Chinchella

In windy conditions, the measurement of liquid and solid atmospheric precipitation is still a challenge even using the most advanced automatic instrumentation (Cauteruccio et al., 2021). The measurement accuracy is affected by various environmental sources of bias, including siting issues and exposure. These add to the instrumental bias, which can be minimized in case of accurate instrument calibration. Wind is however recognised as the most impactful source of environmental bias, outperforming by 3 to 50 times the total impact of all other environmental factors.

Computational Fluid Dynamics simulation with embedded liquid (raindrops) and solid (snowflakes) particle tracking is here used to quantify the wind-induced bias of catching-type precipitation gauges. Starting from the numerically calculated catch ratios, six common commercial gauges having different outer geometry are compared in terms of their expected performance under various precipitation intensity and wind speed conditions. Preliminary wind tunnel experiments allowed full validation of the simulated aerodynamic behaviour and its effect on water drop trajectories.

The overall collection efficiency is shown to depend on the precipitation intensity and its functional dependence is quantitatively derived as a measure of the instrument performance under a wind climatology characterised by a uniform probability density function. A less pronounced diversion of hydrometeor trajectories is shown – at any given size – by instruments with aerodynamic design than in case of more traditional geometry.

Chimney-shaped instruments rank low in case of liquid precipitation measurements, while a high performance is shown by inverted conical and Nipher shielded instruments and the investigated quasi-cylindrical gauges have intermediate behaviour, which depends on their specific aerodynamic features. All instruments rank low at light to moderate precipitation intensity for the measurement of solid precipitation, except the Nipher shielded gauge.

This work provides the basic information needed to apply adjustments to the measured data and supports manufacturers in upgrading instruments with an existing design by introducing on-board adjustments of the measured precipitation. These would only require contemporary measurement of the wind velocity (often included in typical meteorological stations). The full work and the numerically derived adjustments for the six investigated commercial gauges are published in Cauteruccio et al. (2024).

References

Cauteruccio, A., Colli, M., Stagnaro, M., Lanza, L.G. & Vuerich, E. (2021). In situ precipitation measurements. In T. Foken (Ed.), Handbook of Atmospheric Measurements (359-400). Switzerland, Springer Nature. ISBN 978-3-030-52170-7, https://doi.org/10.1007/978-3-030-52171-4_12.

Cauteruccio, A., Chinchella, E. and L.G. Lanza (2024). The overall collection efficiency of catching-type precipitation gauges in windy conditions. Water Resour. Res., in press. https://doi.org/10.1029/2023WR035098.

How to cite: Lanza, L. G., Cauteruccio, A., and Chinchella, E.: Wind-induced bias of catching-type precipitation gauges and their overall collection efficiency, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-17471, https://doi.org/10.5194/egusphere-egu24-17471, 2024.

EGU24-18921 | Orals | HS7.1

Unveiling the Geodetic Distribution of Temporal Characteristics in Rainstorm Events across Republic of Korea 

Hoyoung Cha, Jongjin Baik, Hyeon-Joon Kim, Jinwook Lee, Jongyun Byun, and Changhyun Jun

Abstract

This study analyzed geodetic distribution about temporal characteristics in rainstorm (> 1 hour) observed at approximately 600 rainfall stations across Republic of Korea. Utilizing minute-scale precipitation data observed by rainfall stations from 2000 to 2022, independent rainstorm events separated from rainfall data per unit time (i.e., 10, 20, 30, and 60 minutes) and Inter-Event Time Definition (IETD) (i.e., 2, 3, 4, and 6 hours). The significant variations in rainfall characteristics are defined as the number of independent rainstorm events, rainfall duration (hour), amount (mm), and intensity (mm/hour) for quantifying the temporal characteristics across rainfall stations. We quantified temporal characteristics among rainfall characteristics observed by rainfall stations based on latitude and longitude. The number of independent rainstorm events varies significantly depending on unit time and IETD, and the occurrence of events was frequently observed in areas characterized by island features. The rainfall amount for independent rainstorm events obscured significant characteristics, excluding Halla Mountain on Jeju Island. The geodetic distribution for the duration and intensity per rainstorm event varied depending on the characteristics of the region (i.e., island, mountain, etc.). Based on these results, it was confirmed that certain temporal characteristics vary according to regional features. In future research, we intend to utilize this information to cluster rainfall stations based on temporal characteristics.

Keywords: Independent Rainstorm Events, Temporal Characteristics, Geodetic Distribution, Regional Features, Republic of Korea

Acknowledgment

This research was supported by Korea Environment Industry & Technology Institute (KEITI) funded by Korea Ministry of Environment (RS-2022-KE002032 and 2022003640001) and was also supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (No. NRF-2022R1A4A3032838 and No. RS-2023-00250239).

How to cite: Cha, H., Baik, J., Kim, H.-J., Lee, J., Byun, J., and Jun, C.: Unveiling the Geodetic Distribution of Temporal Characteristics in Rainstorm Events across Republic of Korea, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-18921, https://doi.org/10.5194/egusphere-egu24-18921, 2024.

EGU24-20231 | Posters on site | HS7.1

A new method for disaggregating path-averaged rain rates from commercial microwave links 

Martin Fencl and Marc Schleiss

Commercial microwave links (CMLs) serve as point-to-point radio connections in cellular backhaul and offer a promising way to measure rainfall opportunistically. Raindrops along the CML path attenuate electromagnetic waves, allowing the conversion of this attenuation into path-averaged rain rates. Wide coverage of CML networks, high density in urban areas, and cost-effective operation present clear advantages over traditional rain gauges and radar networks. However, the integrated nature of CML data poses a challenge. When transforming this data into spatially representative rainfall estimates, such as 2D maps, path-integrated rain rates need to be converted into point data and interpolated to a regular two-dimensional Cartesian grid. The most direct method involves reducing each CML observation to a single-point measurement at the path's center, followed by interpolation using techniques like kriging or inverse distance weighted (IDW) interpolation. Yet, past studies indicate that for longer CMLs (several kilometers) and intense localized rain showers, this approach can introduce significant biases and unrealistic rainfall distributions due to the substantial spatial and temporal variability of rainfall.

In this contribution, we introduce a new disaggregation method employing random cascades. The method redistributes rainfall amounts along CML paths across progressively smaller scales using a discrete, conservative multiplicative random cascade. Inspired by the EVA (Equal-volume area) cascade developed by Schleiss (2020) for disaggregating spatially intermittent rainfall fields, our approach involves splitting each CML segment into two new segments with different path-lengths but identical path-integrated rainfall. We call this new method CLEAR (CML segments with equal amounts of rain). CLEAR is tested for CML network of 77 CMLs located in Prague, CZ. First, the disaggregation is evaluated using simulated CML observations and, second, CML rain rates derived from real attenuation data.

Our findings demonstrate that CLEAR surpasses reconstruction algorithms that reduce CML observations into a single point. It accurately replicates the highly diverse rainfall distributions observed along CMLs, including their intermittency. Moreover, the stochastic nature of the cascade enables the quantification of uncertainty associated with the spatial redistribution of rainfall rates along CMLs.

References

Schleiss, Marc. “A New Discrete Multiplicative Random Cascade Model for Downscaling Intermittent Rainfall Fields.” Hydrology and Earth System Sciences 24, no. 7 (July 23, 2020): 3699–3723. https://doi.org/10.5194/hess-24-3699-2020.

How to cite: Fencl, M. and Schleiss, M.: A new method for disaggregating path-averaged rain rates from commercial microwave links, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-20231, https://doi.org/10.5194/egusphere-egu24-20231, 2024.

EGU24-20898 | Posters on site | HS7.1

Comparative analysis of rainfall characteristics for two distinct research plots 

Jürgen Komma, Borbala Szeles, Katarina Zabret, Mojca Šraj, and Juraj Parajka

In natural environments, rainfall causes soil erosion, which has a significant impact on the agricultural production and the ecological conditions of the streams. Due to different types of vegetation, their unique characteristics and seasonality, there are still a lot of open scientific questions about how rainfall interception process influences the rainfall erosivity and soil erosion. With the aim of improving knowledge about rainfall interception by different vegetation and its impact on the rainfall erosivity, an interdisciplinary and international research team (Faculty of Civil and Geodetic Engineering at the University of Ljubljana, Slovenian Forestry Institute and Technical University of Vienna) work together in the research project entitled “Evaluation of the impact of rainfall interception on soil erosion”. In the scope of the project, drop size distribution measurements above and below selected plants will be conducted in combination with classical measurements of rainfall partitioning. The measurements are ongoing in the small urban park in Ljubljana, Slovenia and in the experimental catchment with mainly agricultural land use in Lower Austria (The Hydrological Open Air Laboratory HOAL in Petzenkirchen). To evaluate the differences in rainfall characteristics for the two research plots, a comparative analysis on rainfall event properties such as rainfall amount, duration and intensity, size and velocity distribution of raindrops is performed. The aim of the presentation is to introduce the project and presents the first comparison of the rainfall characteristics at research plots in Austria and Slovenia.

Acknowledgments: This contribution is part of the ongoing research project entitled “Evaluation of the impact of rainfall interception on soil erosion” supported by the Slovenian Research and Innovation Agency (project J2-4489) and the Austrian Science Fund (FWF) I 6254-N.

How to cite: Komma, J., Szeles, B., Zabret, K., Šraj, M., and Parajka, J.: Comparative analysis of rainfall characteristics for two distinct research plots, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-20898, https://doi.org/10.5194/egusphere-egu24-20898, 2024.

NP4 – Time Series and Big Data Methods

EGU24-24 | Orals | NP4.1

The fractional Sinusoidal wavefront Model (fSwp) for time series displaying persistent stationary cycles 

Gael Kermarrec, Federico Maddanu, Anna Klos, and Tommaso Proietti

In the analysis of sub-annual climatological or geodetic time series such as tide gauges, precipitable water vapor, or GNSS vertical displacements time series but also temperatures or gases concentrations, seasonal cycles are often found to have a time-varying amplitude and phase.

These time series are usually modelled with a deterministic approach that includes trend, annual, and semi-annual periodic components having constant amplitude and phase-lag. This approach can potentially lead to inadequate interpretations, such as an overestimation of Global Navigation Satellite System (GNSS) station velocity, up to masking important geophysical phenomena that are related to the amplitude variability and are important for deriving trustworthy interpretation for climate change assessment.

We address that challenge by proposing a novel linear additive model called the fractional Sinusoidal Waveform process (fSWp), accounting for possible nonstationary cyclical long memory, a stochastic trend that can evolve over time and an additional serially correlated noise capturing the short-term variability. The model has a state space representation and makes use of the Kalman filter (KF). Suitable enhancements of the basic methodology enable handling data gaps, outliers, and offsets. We demonstrate our method using various climatological and geodetic time series to illustrate its potential to capture the time-varying stochastic seasonal signals.

How to cite: Kermarrec, G., Maddanu, F., Klos, A., and Proietti, T.: The fractional Sinusoidal wavefront Model (fSwp) for time series displaying persistent stationary cycles, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-24, https://doi.org/10.5194/egusphere-egu24-24, 2024.

On some maps of the first military survey of the Habsburg Empire, the upper direction of the sections does not face the cartographic north, but makes an angle of about 15° with it. This may be due to the fact that the sections were subsequently rotated to the magnetic north of the time. Basically, neither their projection nor their projection origin is known yet.

In my research, I am dealing with maps of Inner Austria, the Principality of Transylvania and Galicia (nowadays Poland and Ukraine), and I am trying to determine their projection origin. For this purpose, it is assumed, based on the archival documentation of the survey, that these are Cassini projection maps. My hypothesis is that they are Graz, Cluj Napoca or Alba Julia and Lviv. I also consider the position of Vienna in each case, since it was the main centre of the survey.

The angle of rotation was taken in part from the gufm1 historical magnetic model back to 1590 for the assumed starting points and year of mapping. In addition, as a theoretical case, I calculated the rotation angle of the map sections using coordinate geometry. I then calculated the longitude of the projection starting point for each case using univariate minimization. Since the method is invariant to latitude, it can only be determined from archival data.

Based on these, the starting point for Inner Austria from the rotation of the map was Vienna, which is not excluded by the archival sources, and since the baseline through Graz also started from there, it is partly logical. The map rotation for Galicia and Transylvania also confirmed the starting point of the hypothesis.  Since both Alba Julia and Cluj Napoca lie at about the same longitude, the method cannot make a difference there; and the archival data did not provide enough evidence. In comparison, the magnetic declination rotations yielded differences of about 1°, which may be due to an error in the magnetic model.

On this basis, I have given the assumed projections of the three maps with projection starting points, and developed a method for determining the projection starting points of the other rotated grid maps. The results suggest that there is a very high probability that the section network was rotated in the magnetic north direction, and thus provide a way to refine the magnetic declination data at that time.

With this method I managed to give new indirekt magnetic declinations data from Central-East Europe, which can help to improve the historical magnetic field models. The main reason for this is that we don’t have any measurement from that region.

Furthermore the difference beetwen the angle of the section north and the declination data from gufm1 always 0.8-1°. Maybe there are systematical data error at that region.

Supported by the ÚNKP-23-6 New National Excellence Program of the Ministry for Culture and Innovation from the source of the National Research, Development and Innovation Fund.

How to cite: Koszta, B. and Timár, G.: A possible cartographical data source for historical magnetic field improvement: The direction of the section north of the Habsburg first military survey, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-582, https://doi.org/10.5194/egusphere-egu24-582, 2024.

EGU24-1988 | ECS | Posters on site | NP4.1

Predictive ability assessment of Bayesian Causal Reasoning (BCR) on runoff temporal series 

Santiago Zazo, José Luis Molina, Carmen Patino-Alonso, and Fernando Espejo

The alteration of traditional hydrological patterns due to global warming is leading to a modification of the hydrological cycle. This situation draws a complex scenario for the sustainable management of water resources. However, this issue offers a challenge for the development of innovative approaches that allow an in-depth capturing the logical temporal-dependence structure of these modifications to advance sustainable management of water resources, mainly through the reliable predictive models. In this context, Bayesian Causality (BC), addressed through Causal Reasoning (CR) and supported by a Bayesian Networks (BNs), called Bayesian Causal Reasoning (BCR) is a novel hydrological research area that can help identify those temporal interactions efficiently.

This contribution aims to assesses the BCR ability to discover the logical and non-trivial temporal-dependence structure of the hydrological series, as well as its predictability. For this, a BN that conceptually synthesizes the time series is defined, and where the conditional probability is propagated over the time throughout the BN through an innovative Dependence Mitigation Graph. This is done by coupling among an autoregressive parametric approach and causal model. The analytical ability of the BCR highlighted the logical temporal structure, latent in the time series, which defines the general behavior of the runoff. This logical structure allowed to quantify, through a dependence matrix which summarizes the strength of the temporal dependencies, the two temporal fractions that compose the runoff: one due to time (Temporally Conditioned Runoff) and one not (Temporally Non-conditioned Runoff). Based on this temporal conditionality, a predictive model is implemented for each temporal fraction, and its reliability is assessed from a double probabilistic and metrological perspective.

This methodological framework is applied to two Spanish unregulated sub-basins; Voltoya river belongs to Duero River Basin, and Mijares river, in the Jucar River Basin. Both cases with a clearly opposite temporal behavior, Voltoya independent and Mijares dependent, and with increasingly more problems associated with droughts.

The findings of this study may have important implications over the knowledge of temporal behavior of water resources of river basin and their adaptation. In addition, TCR and TNCR predictive models would allow advances in the optimal dimensioning of storage infrastructures (reservoirs), with relevant substantial economic/environmental savings. Also, a more sustainable management of river basins through more reliable control reservoirs’ operation is expected to be achieved. Finally, these results open new possibilities for developing predictive hydrological models within a BCR framework.

How to cite: Zazo, S., Molina, J. L., Patino-Alonso, C., and Espejo, F.: Predictive ability assessment of Bayesian Causal Reasoning (BCR) on runoff temporal series, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-1988, https://doi.org/10.5194/egusphere-egu24-1988, 2024.

EGU24-3857 | ECS | Posters on site | NP4.1 | Highlight

Spatial-Temporal Analysis of Forest Mortality 

Sara Alibakhshi

Climate-induced forest mortality poses an increasing threat worldwide, which calls for developing robust approaches to generate early warning signals of upcoming forest state change. This research explores the potential of satellite imagery, utilizing advanced spatio-temporal indicators and methodologies, to assess the state of forests preceding mortality events. Traditional approaches, such as techniques based on temporal analyses, are impacted by limitations related to window size selection and detrending methods, potentially leading to false alarms. To tackle these challenges, our study introduces two new approaches, namely the Spatial-Temporal Moran (STM) and Spatial-Temporal Geary (STG) approaches, both focusing on local spatial autocorrelation measures. These approaches can effectively address the shortcomings inherent in traditional methods. The research findings were assessed across three study sites within California national parks, and Kendall's tau was employed to quantify the significance of false and positive alarms. To facilitate the measurement of ecosystem state change, trend estimation, and identification of early warning signals, this study also provides "stew" R package. The implications of this research extend to various groups, such as ecologists, conservation practitioners, and policymakers, providing them with the means to address emerging environmental challenges in global forest ecosystems.

How to cite: Alibakhshi, S.: Spatial-Temporal Analysis of Forest Mortality, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-3857, https://doi.org/10.5194/egusphere-egu24-3857, 2024.

Iram Parvez1, Massimiliano Cannata2, Giorgio Boni1, Rossella Bovolenta1 ,Eva Riccomagno3 , Bianca Federici1

1 Department of Civil, Chemical and Environmental Engineering (DICCA), Università degli Studi di Genova, Via Montallegro 1, 16145 Genoa, Italy (iram.parvez@edu.unige.it,bianca.federici@unige.it, giorgio.boni@unige.it, rossella.bovolenta@unige.it).

2 Institute of Earth Sciences (IST), Department for Environment Constructions and Design (DACD), University of Applied Sciences and Arts of Southern Switzerland (SUPSI), CH-6952 Canobbio, Switzerland(massimiliano.cannata@supsi.ch).

3 Department of Mathematics, Università degli Studi di Genova, Via Dodecaneso 35, 16146 Genova, Italy(riccomag@dima.unige.it).

The deployment of hydrometeorological sensors significantly contributes to generating real-time big data. The quality and reliability of large datasets pose considerable challenges, as flawed analyses and decision-making processes can result. This research aims to address the issue of anomaly detection in real-time data by exploring machine learning models. Time-series data is collected from IstSOS - Sensor Observation Service, an open-source software that stores, collects and disseminates sensor data. The methodology consists of Gated Recurrent Units based on recurrent neural networks, along with corresponding prediction intervals, applied both to individual sensors and collectively across all temperature sensors within the Ticino region of Switzerland. Additionally, non-parametric methods like Bootstrap and Mean absolute deviation are employed instead of standard prediction intervals to tackle the non-normality of the data. The results indicate that Gated Recurrent Units based on recurrent neural networks, coupled with non-parametric forecast intervals, perform well in identifying erroneous data points. The application of the model on multivariate time series-sensor data establishes a pattern or baseline of normal behavior for the area (Ticino). When a new sensor is installed in the same region, the recognized pattern is used as a reference to identify outliers in the data gathered from the new sensor.

How to cite: Parvez, I.: Exploring Machine Learning Models to Detect Outliers in HydroMet Sensors, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-4280, https://doi.org/10.5194/egusphere-egu24-4280, 2024.

EGU24-5268 | ECS | Orals | NP4.1

Unveiling Geological Patterns: Bayesian Exploration of Zircon-Derived Time Series Data 

Hang Qian, Meng Tian, and Nan Zhang

For its immunity to post-formation geological modifications, zircon is widely utilized as chronological time capsule and provides critical time series data potential to unravel key events in Earth’s geological history, such as supercontinent cycles. Fourier analysis, which assumes stationary periodicity, has been applied to zircon-derived time series data to find the cyclicity of supercontinents, and wavelet analysis, which assumes non-stationary periodicity, corroborates the results of Fourier Analysis in addition to detecting finer-scale signals. Nonetheless, both methods still prognostically assume periodicity in the zircon-derived time-domain data. To stay away from the periodicity assumption and extract more objective information from zircon data, we opt for a Bayesian approach and treat zircon preservation as a composite stochastic process where the number of preserved zircon grains per magmatic event obeys logarithmic series distribution and the number of magmatic events during a geological time interval obeys Poisson distribution. An analytical solution was found to allow us to efficiently invert for the number and distribution(s) of changepoints hidden in the globally compiled zircon data, as well as for the zircon preservation potential (encoded as a model parameter) between two neighboring changepoints. If the distributions of changepoints temporally overlap with those of known supercontinents, then our results serve as an independent, mathematically robust test of the cyclicity of supercontinents. Moreover, our statistical approach inherently provides a sensitivity parameter the tuning of which allows to probe changepoints at various temporal resolution. The constructed Bayesian framework is thus of significant potential to detect other types of trend swings in Earth’s history, such as shift of geodynamic regimes, moving beyond cyclicity detection which limits the application of conventional Fourier/Wavelet analysis.

How to cite: Qian, H., Tian, M., and Zhang, N.: Unveiling Geological Patterns: Bayesian Exploration of Zircon-Derived Time Series Data, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-5268, https://doi.org/10.5194/egusphere-egu24-5268, 2024.

Semi-enclosed freshwater and brackish ecosystems, characterised by restricted water outflow and prolonged residence times, often accumulate nutrients, influencing their productivity and ecological dynamics. These ecosystems exhibit significant variations in bio-physical-chemical attributes, ecological importance, and susceptibility to human impacts. Untangling the complexities of their interactions remains challenging, necessitating a deeper understanding of effective management strategies adapted to their vulnerabilities. This research focuses on the bio-physical aspects, investigating the differential effects of spring and summer light on phytoplankton communities in semi-enclosed freshwater and brackish aquatic ecosystems.

Through extensive field sampling and comprehensive environmental parameter analysis, we explore how phytoplankton respond to varying light conditions in these distinct environments. Sampling campaigns were conducted at Müggelsee, a freshwater lake on Berlin's eastern edge, and Barther Bodden, a coastal lagoon northeast of Rostock on the German Baltic Sea coast, during the springs and summers of 2022 and 2023, respectively. Our analysis integrates environmental factors such as surface light intensity, diffuse attenuation coefficients, nutrient availability, water column dynamics, meteorological data, Chlorophyll-a concentration, and phytoplankton communities. Sampling encompassed multiple depths at continuous intervals lasting three days.

Preliminary findings underscore significant differences in seasonal light availability, with summer exhibiting extended periods of substantial light penetration. These variations seem to impact phytoplankton abundance and diversity uniquely in each ecosystem. While ongoing analyses are underway, early indications suggest distinct phytoplankton responses in terms of species composition and community structure, influenced by the changing light levels. In 2022 the clear water phase during spring indicated that bloom events have occurred under ice cover much earlier than spring, while in the summer there were weak and short-lived blooms of cyanobacteria. The relationship between nutrient availability and phytoplankton dynamics, however, remains uncertain according to our data.

This ongoing study contributes to understanding the role of light as a primary driver shaping phytoplankton community structures and dynamics in these environments.  Our research findings offer insights for refining predictive models, aiding in ecosystem-specific eutrophication management strategies, and supporting monitoring efforts of Harmful Algal Blooms.

How to cite: Kaharuddin, A. and Kaligatla, R.: Comparative Study of Spring and Summer Light Effects on Phytoplankton Communities in Semi-Enclosed Fresh- and Brackish Aquatic Ecosystems., EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-5733, https://doi.org/10.5194/egusphere-egu24-5733, 2024.

EGU24-6065 | ECS | Orals | NP4.1

Magnetospheric time history:  How much do we need for forecasting? 

Kendra R. Gilmore, Sarah N. Bentley, and Andy W. Smith

Forecasting the aurora and its location accurately is important to mitigate any potential harm to vital infrastructure like communications and electricity grid networks. Current auroral prediction models rely on our understanding of the interaction between the magnetosphere and the solar wind or geomagnetic indices. Both approaches do well in predicting but have limitations concerning forecasting (geomagnetic indices-based model) or because of the underlying assumptions driving the model (due to a simplification of the complex interaction). By applying machine learning algorithms to this problem, gaps in our understanding can be identified, investigated, and closed. Finding the important time scales for driving empirical models provides the necessary basis for our long-term goal of predicting the aurora using machine learning.

Periodicities of the Earth’s magnetic field have been extensively studied on a global scale or in regional case studies. Using a suite of different time series analysis techniques including frequency analysis and investigation of long-scale changes of the median/ mean, we examine the dominant periodicities of ground magnetic field measurements at selected locations. A selected number of stations from the SuperMAG network (Gjerloev, 2012), which is a global network of magnetometer stations across the world, are the focus of this investigation.

The periodicities retrieved from the different magnetic field components are compared to each other as well as to other locations. In the context of auroral predictions, an analysis of the dominating periodicities in the auroral boundary data derived from the IMAGE satellite (Chisham et al., 2022) provides a counterpart to the magnetic field periodicities.

Ultimately, we can constrain the length of time history sensible for forecasting.

How to cite: Gilmore, K. R., Bentley, S. N., and Smith, A. W.: Magnetospheric time history:  How much do we need for forecasting?, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-6065, https://doi.org/10.5194/egusphere-egu24-6065, 2024.

EGU24-6151 | Posters on site | NP4.1

Using information-theory metrics to detect regime changes in dynamical systems 

Javier Amezcua and Nachiketa Chakraborty

Dynamical systems can display a range of dynamical regimes (e.g. attraction to, fixed points, limit cycles, intermittency, chaotic behaviour) depending on the values of parameters in the system. In this work we demonstrate how non-parametric entropy estimation codes (in particular NPEET) based on the Kraskov method can be applied to find regime transitions in a 3D chaotic model (the Lorenz 1963 system) when varying the values of the parameters. These infromation-theory-based methods are simpler and cheaper to apply than more traditional metrics from dynamical systems (e.g. computation of Lyapunov exponents). The non-parametric nature of the method allows for handling long time series without a prohibitive computational burden. 

How to cite: Amezcua, J. and Chakraborty, N.: Using information-theory metrics to detect regime changes in dynamical systems, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-6151, https://doi.org/10.5194/egusphere-egu24-6151, 2024.

EGU24-9367 | ECS | Orals | NP4.1

Fractal complexity evaluation of meteorological droughts over three Indian subdivisions using visibility Graphs 

Susan Mariam Rajesh, Muraleekrishnan Bahuleyan, Arathy Nair GR, and Adarsh Sankaran

Evaluation of scaling properties and fractal formalisms is one of the potential approaches for modelling complex series. Understanding the complexity and fractal characterization of drought index time series is essential for better preparedness against drought disasters. This study presents a novel visibility graph-based evaluation of fractal characterization of droughts of three meteorological subdivisions of India. In this method, the horizontal visibility graph (HVG) and Upside-down visibility graph (UDVG) are used for evaluating the network properties for different standardized precipitation index (SPI) series of 3, 6 and 12 month time scales representing short, medium and long term droughts. The relative magnitude of fractal estimates is controlled by the drought characteristics of wet-dry transitions. The estimates of degree distribution clearly deciphered the self-similar properties of droughts of all the subdivisions. For an insightful depiction of drought dynamics, the fractal exponents and spectrum are evaluated by the concurrent application of Sand Box Method (SBM) and Chhabra and Jenson Method (CJM). The analysis was performed for overall series along with the pre- and post-1976-77 Global climate shift scenarios. The complexity is more evident in short term drought series and UDVG formulations implied higher fractal exponents for different moment orders irrespective of drought type and locations considered in this study. Useful insights on the relationship between complex network and fractality are evolved from the study, which may help in improved drought forecasting. The visibility graph based fractality estimation evaluation is efficient in capturing drought and it has vast potential in the drought predictions in a changing environment.

Keywords:  Drought, Fractal, SPI, Visibility Graph

How to cite: Rajesh, S. M., Bahuleyan, M., Nair GR, A., and Sankaran, A.: Fractal complexity evaluation of meteorological droughts over three Indian subdivisions using visibility Graphs, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-9367, https://doi.org/10.5194/egusphere-egu24-9367, 2024.

EGU24-9537 | Posters on site | NP4.1

Wavelet-Induced Mode Extraction procedure: Application to climatic data 

Elise Faulx, Xavier Fettweis, Georges Mabille, and Samuel Nicolay

The Wavelet-Induced Mode Extraction procedure (WIME) [2] was developed drawing inspiration from Empirical Mode Decomposition. The concept involves decomposing the signal into modes, each presenting a characteristic frequency, using continuous wavelet transform. This method has yielded intriguing results in climatology [3,4]. However, the initial algorithm did not account for the potential existence of slight frequency fluctuations within a mode, which could impact the reconstruction of the original signal [4]. The new version (https://atoms.scilab.org/toolboxes/toolbox_WIME/0.1.0) now allows for the evolution of a mode in the space-frequency half-plane, thus considering the frequency evolution of a mode [2]. A natural application of this tool is in the analysis of Milankovitch cycles, where subtle changes have been observed throughout history. The method also refines the study of solar activity, highlighting the role of the "Solar Flip-Flop." Additionally, the examination of temperature time series confirms the existence of cycles around 2.5 years. It is now possible to attempt to correlate solar activity with this observed temperature cycle, as seen in speleothem records [1].

[1] Allan, M., Deliège, A., Verheyden, S., Nicolay S. and Fagel, N. Evidence for solar influence in a Holocene speleothem record, Quaternary Science Reviews, 2018.
[2] Deliège, A. and Nicolay, S., Extracting oscillating components from nonstationary time series: A wavelet-induced method, Physical Review. E, 2017.
[3] Nicolay, S., Mabille, G., Fettweis, X. and Erpicum, M., A statistical validation for the cycles found in air temperature data using a Morlet wavelet-based method, Nonlinear Processes in Geophysics, 2010.
[4] Nicolay, S., Mabille, G., Fettweis, X. and Erpicum, M., 30 and 43 months period cycles found in air temperature time series using the Morlet wavelet, Climate Dynamics, 2009.

How to cite: Faulx, E., Fettweis, X., Mabille, G., and Nicolay, S.: Wavelet-Induced Mode Extraction procedure: Application to climatic data, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-9537, https://doi.org/10.5194/egusphere-egu24-9537, 2024.

EGU24-10258 | Orals | NP4.1

New concepts on quantifying event data 

Norbert Marwan and Tobias Braun

A wide range of geoprocesses manifest as observable events in a variety of contexts, including shifts in palaeoclimate regimes, evolutionary milestones, tectonic activities, and more. Many prominent research questions, such as synchronisation analysis or power spectrum estimation of discrete data, pose considerable challenges to linear tools. We present recent advances using a specific similarity measure for discrete data and the method of recurrence plots for different applications in the field of highly discrete event data. We illustrate their potential for palaeoclimate studies, particularly in detecting synchronisation between signals of discrete extreme events and continuous signals, estimating power spectra of spiky signals, and analysing data with irregular sampling.

How to cite: Marwan, N. and Braun, T.: New concepts on quantifying event data, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-10258, https://doi.org/10.5194/egusphere-egu24-10258, 2024.

EGU24-10415 | ECS | Orals | NP4.1

Application of Transfer Learning techniques in one day ahead PV production prediction 

Marek Lóderer, Michal Sandanus, Peter Pavlík, and Viera Rozinajová

Nowadays photovoltaic panels are becoming more affordable, efficient, and popular due to their low carbon footprint. PV panels can be installed in many places providing green energy to the local grid reducing energy cost and transmission losses. Since the PV production is highly dependent on the weather conditions, it is extremely important to estimate expected output in advance in order to maintain energy balance in the grid and provide enough time to schedule load distribution. The PV production output can be calculated by various statistical and machine learning prediction methods. In general, the more data available, the more precise predictions can be produced. This poses a problem for recently installed PV panels for which not enough data has been collected or the collected data are incomplete. 

A possible solution to the problem can be the application of an approach called Transfer Learning which has the inherent ability to effectively deal with missing or insufficient amounts of data. Basically, Transfer Learning is a machine learning approach which offers the capability of transferring knowledge acquired from the source domain (in our case a PV panel with a large amount of historical data) to different target domains (PV panels with very little collected historical data) to resolve related problems (provide reliable PV production predictions). 

In our study, we investigate the application, benefits and drawbacks of Transfer Learning for one day ahead PV production prediction. The model used in the study is based on complex neural network architecture, feature engineering and data selection. Moreover, we focus on the exploration of multiple approaches of adjusting weights in the target model retraining process which affect the minimum amount of training data required, final prediction accuracy and model’s overall robustness. Our models use historical meteorological forecasts from Deutscher Wetterdienst (DWD) and photovoltaic measurements from the project PVOutput which collects data from installed solar systems across the globe. Evaluation is performed on more than 100 installed PV panels in Central Europe.

How to cite: Lóderer, M., Sandanus, M., Pavlík, P., and Rozinajová, V.: Application of Transfer Learning techniques in one day ahead PV production prediction, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-10415, https://doi.org/10.5194/egusphere-egu24-10415, 2024.

EGU24-11897 | Posters on site | NP4.1

Results of joint processing of magnetic observatory data of international Intermagnet network in a unified coordinate system 

Beibit Zhumabayev, Ivan Vassilyev, Zhasulan Mendakulov, Inna Fedulina, and Vitaliy Kapytin

In each magnetic observatory, the magnetic field is registered in local Cartesian coordinate systems associated with the geographic coordinates of the locations of these observatories. To observe extraterrestrial magnetic field sources, such as the interplanetary magnetic field or magnetic clouds, a method of joint processing of data from magnetic observatories of the international Intermagnet network was implemented. In this method, the constant component is removed from the observation results of individual observatories, their measurement data is converted into the ecliptic coordinate system, and the results obtained from all observatories are averaged after the coordinate transformation.

The first data on joint processing of measurement results from the international network of Intermagnet magnetic observatories in the period before the onset of magnetic storms of various types, during these storms and after their end are presented. There is a significant improvement in the signal-to-noise ratio after combining the measurement results from all observatories, which makes it possible to isolate weaker external magnetic fields. A change in the shape of magnetic field variations is shown, which can provide new knowledge about the mechanism of development of magnetic storms.

How to cite: Zhumabayev, B., Vassilyev, I., Mendakulov, Z., Fedulina, I., and Kapytin, V.: Results of joint processing of magnetic observatory data of international Intermagnet network in a unified coordinate system, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-11897, https://doi.org/10.5194/egusphere-egu24-11897, 2024.

We introduce the CLEAN algorithm to identify narrowband Ultra Low Frequency (ULF) Pc5 plasma waves in Earth’s magnetosphere. The CLEAN method was first used for constructing 2D images in astronomical radio interferometry but has since been applied to a huge range of areas including adaptation for time series analysis. The algorithm performs a nonlinear deconvolution in the frequency domain (equivalent to a least-squares in the time domain) allowing for identification of multiple individual wave spectral peaks within the same power spectral density. The CLEAN method also produces real amplitudes instead of model fits to the peaks and retains phase information. We applied the method to GOES magnetometer data spanning 30 years to study the distribution of narrowband Pc5 ULF waves at geosynchronous orbit. We found close to 30,0000 wave events in each of the vector magnetic field components in field-aligned coordinates. We discuss wave occurrence and amplitudes distributed in local time and frequency. The distribution of the waves under different solar wind conditions are also presented. With some precautions, which are applicable to other event identification methods, the CLEAN technique can be utilized to detect wave events and its harmonics in the magnetosphere and beyond. We also discuss limitations of the method mainly the detection of unrealistic peaks due to aliasing and Gibbs phenomena.

How to cite: Inceoglu, F. and Loto'aniu, P.: Using the CLEAN Algorithm to Determine the Distribution of Ultra Low Frequency Waves at Geostationary Orbit, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-12928, https://doi.org/10.5194/egusphere-egu24-12928, 2024.

EGU24-12938 | Posters on site | NP4.1

Applying Multifractal Theory and Statistical Techniques for High Energy Volcanic Explosion Detection and Seismic Activity Monitoring in Volcanic Time Series 

Marisol Monterrubio-Velasco, Xavier Lana, Raúl Arámbula-Mendoza, and Ramón Zúñiga

Understanding volcanic activity through time series data analysis is crucial for uncovering the fundamental physical mechanisms governing this natural phenomenon.

In this study, we show the application of multifractal and fractal methodologies, along with statistical analysis, to investigate time series associated with volcanic activity. We aim to make use of these approaches to identify significant variations within the physical processes related to changes in volcanic activity. These methodologies offer the potential to identify pertinent changes preceding a high-energy explosion or a significant volcanic eruption.

In particular, we apply it to analyze two study cases. First, the evolution of the multifractal structure of volcanic emissions of low, moderate, and high energy explosions applied to Volcán de Colima (México years 2013-2015). The results contribute to obtaining quite evident signs of the immediacy of possible dangerous emissions of high energy, close to 8.0x10^8 J. Additionally, the evolution of the adapted Gutenberg-Richter seismic law to volcanic energy emissions contributes to confirm the results obtained using multifractal analysis. Secondly, we also studied the time series of the Gutenberg-Richter b-parameter of seismic activities associated with volcanic emissions in Iceland, Hawaii, and the Canary Islands, through the concept of Disparity (degree of irregularity), the fractal Hurst exponent, H, and several multifractal parameters. The results obtained should facilitate a better knowledge of the relationships between the activity of volcanic emissions and the corresponding related seismic activities.  

How to cite: Monterrubio-Velasco, M., Lana, X., Arámbula-Mendoza, R., and Zúñiga, R.: Applying Multifractal Theory and Statistical Techniques for High Energy Volcanic Explosion Detection and Seismic Activity Monitoring in Volcanic Time Series, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-12938, https://doi.org/10.5194/egusphere-egu24-12938, 2024.

EGU24-13593 | ECS | Posters on site | NP4.1

Characterizing Uncertainty in Spatially Interpolated Time Series of Near-Surface Air Temperature 

Conor Doherty and Weile Wang

Spatially interpolated meteorological data products are widely used in the geosciences as well as disciplines like epidemiology, economics, and others. Recent work has examined methods for quantifying uncertainty in gridded estimates of near-surface air temperature that produce distributions rather than simply point estimates at each location. However, meteorological variables are correlated not only in space but in time, and sampling without accounting for temporal autocorrelation produces unrealistic time series and potentially underestimates cumulative errors. This work first examines how uncertainty in air temperature estimates varies in time, both seasonally and at shorter timescales. It then uses data-driven, spectral, and statistical methods to better characterize uncertainty in time series of estimated air temperature values. Methods for sampling that reproduce spatial and temporal autocorrelation are presented and evaluated. The results of this work are particularly relevant to domains like agricultural and ecology. Physical processes including evapotranspiration and primary production are sensitive to variables like near-surface air temperature, and errors in these important meteorological inputs accumulate in model outputs over time.

How to cite: Doherty, C. and Wang, W.: Characterizing Uncertainty in Spatially Interpolated Time Series of Near-Surface Air Temperature, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-13593, https://doi.org/10.5194/egusphere-egu24-13593, 2024.

EGU24-13879 | ECS | Posters on site | NP4.1

Understanding the role of vegetation responses to drought in regulating autumn senescence 

Eunhye Choi and Josh Gray

Vegetation phenology is the recurring of plant growth, including the cessation and resumption of growth, and plays a significant role in shaping terrestrial water, nutrient, and carbon cycles. Changes in temperature and precipitation have already induced phenological changes around the globe, and these trends are likely to continue or even accelerate. While warming has advanced spring arrival in many places, the effects on autumn phenology are less clear-cut, with evidence for earlier, delayed, or even unchanged end of the growing season (EOS). Meteorological droughts are intensifying in duration and frequency because of climate change. Droughts intricately impact changes in vegetation, contingent upon whether the ecosystem is limited by water or energy. These droughts have the potential to influence EOS changes. Despite this, the influence of drought on EOS remains largely unexplored. This study examined moisture’s role in controlling EOS by understanding the relationship between precipitation anomalies, vegetation’s sensitivity to precipitation (SPPT), and EOS. We also assess regional variations in responses to the impact of SPPT on EOS.

The study utilized multiple vegetation and water satellite products to examine the patterns of SPPT in drought and its impact on EOS across aridity gradients and vegetation types. By collectively evaluating diverse SPPTs from various satellite datasets, this work offers a comprehensive understanding and critical basis for assessing the impact of drought on EOS. We focused on the Northern Hemisphere from 2000 to 2020, employing robust statistical methods. This work found that, in many places, there was a stronger relationship between EOS and drought in areas with higher SPPT. Additionally, a non-linear negative relationship was identified between EOS and SPPT in drier regions, contracting with a non-linear positive relationship observed in wetter regions. These findings were consistent across a range of satellite-derived vegetation products. Our findings provide valuable insights into the effects of SPPT on EOS during drought, enhancing our understanding of vegetation responses to drought and its consequences on EOS and aiding in identifying drought-vulnerable areas.

How to cite: Choi, E. and Gray, J.: Understanding the role of vegetation responses to drought in regulating autumn senescence, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-13879, https://doi.org/10.5194/egusphere-egu24-13879, 2024.

EGU24-16981 | ECS | Orals | NP4.1

A machine-learning-based approach for predicting the geomagnetic secular variation 

Sho Sato and Hiroaki Toh

We present a machine-learning-based approach for predicting the geomagnetic main field changes, known as secular variation (SV), in a 5-year range for use for the 14th generation of International Geomagnetic Reference Field (IGRF-14). The training and test datasets of the machine learning (ML) models are geomagnetic field snapshots derived from magnetic observatory hourly means, and CHAMP and Swarm-A satellite data (MCM Model; Ropp et al., 2020). The geomagnetic field data are not used as-is in the original time series but were differenced twice before training. Because SV is strongly influenced by the geodynamo process occurring in the Earth's outer core, challenges still persist despite efforts to model and forecast the realistic nonlinear behaviors (such as the geomagnetic jerks) of the geodynamo through data assimilation. We compare three physics-uninformed ML models, namely, the Autoregressive (AR) model, Vector Autoregressive (VAR) model, and Recurrent Neural Network (RNN) model, to represent the short-term temporal evolution of the geomagnetic main field on the Earth’s surface. The quality of 5-year predictions is tested by the hindcast results for the learning window from 2004.50 to 2014.25. These tests show that the forecast performance of our ML model is comparable with that of candidate models of IGRF-13 in terms of data misfits after the release epoch (Year 2014.75). It is found that all three ML models give 5-year prediction errors of less than 100nT, among which the RNN model shows a slightly better accuracy. They also suggest that Overfitting to the training data used is an undesirable machine learning behavior that occurs when the RNN model gives accurate reproduction of training data but not for forecasting targets.

How to cite: Sato, S. and Toh, H.: A machine-learning-based approach for predicting the geomagnetic secular variation, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-16981, https://doi.org/10.5194/egusphere-egu24-16981, 2024.

EGU24-17344 | Posters on site | NP4.1

Introducing a new statistical theory to quantify the Gaussianity of the continuous seismic signal 

Éric Beucler, Mickaël Bonnin, and Arthur Cuvier

The quality of the seismic signal recorded at permanent and temporary stations is sometimes degraded, either abruptly or over time. The most likely cause is a high level of humidity, leading to corrosion of the connectors but environmental changes can also alter recording conditions in various frequency ranges and not necessarily for all three components in the same way. Assuming that the continuous seismic signal can be described by a normal distribution, we present a new approach to quantify the seismogram quality and to point out any time sample that deviates from this Gaussian assumption. To this end the notion of background Gaussian signal (BGS) to statistically describe a set of samples that follows a normal distribution. The discrete function obtained by sorting the samples in ascending order of amplitudes is compared to a modified probit function to retrieve the elements composing the BGS, and its statistical properties, mostly the Gaussian standard deviation, which can then differ from the classical standard deviation. Hence the ratio of both standard deviations directly quantifies the dominant gaussianity of the continuous signal and any variation reflects a statistical modification of the signal quality. We present examples showing daily variations in this ratio for stations known to have been affected by humidity, resulting in signal degradation. The theory developed can be used to detect subtle variations in the Gaussianity of the signal, but also to point out any samples that don't match the Gaussianity assumption, which can then be used for other seismological purposes, such as coda determination.

How to cite: Beucler, É., Bonnin, M., and Cuvier, A.: Introducing a new statistical theory to quantify the Gaussianity of the continuous seismic signal, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-17344, https://doi.org/10.5194/egusphere-egu24-17344, 2024.

EGU24-17566 | ECS | Posters on site | NP4.1

Unveiling Climate-Induced Ocean Wave Activities Using Seismic Array Data in the North Sea Region 

Yichen Zhong, Chen Gu, Michael Fehler, German Prieto, Peng Wu, Zhi Yuan, Zhuoyu Chen, and Borui Kang

Climate events may induce abnormal ocean wave activities, that can be detected by seismic array on nearby coastlines. We collected long-term continuous array seismic data in the Groningen area and the coastal areas of the North Sea, conducted a comprehensive analysis to extract valuable climate information hidden within the ambient noise. Through long-term spectral analysis, we identified the frequency band ranging from approximately 0.2Hz, which appears to be associated with swell waves within the region, exhibiting a strong correlation with the significant wave height (SWH). Additionally, the wind waves with a frequency of approximately 0.4 Hz and gravity waves with periods exceeding 100 seconds were detected from the seismic ambient noise. We performed a correlation analysis between the ambient noise and various climatic indexes across different frequency bands. The results revealed a significant correlation between the North Atlantic Oscillation (NAO) Index and the ambient noise around 0.17Hz.

Subsequently, we extracted the annual variation curves of SWH frequency from ambient noise at each station around the North Sea and assembled them into a sparse spatial grid time series (SGTS). An empirical orthogonal function (EOF) analysis was conducted, and the Principal Component (PC) time series derived from the EOF analysis were subjected to a correlation analysis with the WAVEWATCH III (WW3) model simulation data, thereby confirming the wave patterns. Moreover, we conducted the spatial distribution study of SGTS. The spatial features revealed that the southern regions of the North Sea exhibit higher wind-wave energy components influenced by the Icelandic Low pressure system and topography, which explains the correlation between ambient noise in the region and the NAO index. Furthermore, spatial features disclosed a correlation between the first EOF mode of the North Sea ocean waves and the third mode of sea surface temperature anomalies. This research shows the potential of utilizing existing off-shore seismic monitoring systems to study global climate variation and physical oceanography.

How to cite: Zhong, Y., Gu, C., Fehler, M., Prieto, G., Wu, P., Yuan, Z., Chen, Z., and Kang, B.: Unveiling Climate-Induced Ocean Wave Activities Using Seismic Array Data in the North Sea Region, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-17566, https://doi.org/10.5194/egusphere-egu24-17566, 2024.

EGU24-18061 | ECS | Orals | NP4.1

A new methodology for time-series reconstruction of global scale historical Earth observation data 

Davide Consoli, Leandro Parente, and Martijn Witjes

Several machine learning algorithms and analytical techniques do not allow gaps or non-values in input data. Unfortunately, earth observation (EO) datasets, such as satellite images, are gravely affected by cloud contamination and sensor artifacts that create gaps in the time series of collected images. This limits the usage of several powerful techniques for modeling and analysis. To overcome these limitations, several works in literature propose different imputation methods to reconstruct the gappy time series of images, providing complete time-space datasets and enabling their usage as input for many techniques.

However, among the time-series reconstruction methods available in literature, only a few of them are publicly available (open source code), applicable without any external source of data, and suitable for application to petabyte (PB) sized dataset like the full Landsat archive. The few methods that match all these characteristics are usually quite trivial (e.g. linear interpolation) and, as a consequence, they often show poor performance in reconstructing the images. 

For this reason, we propose a new methodology for time series reconstruction designed to match all these requirements. Like some other methods in literature, the new method, named seasonally weighted average generalization (SWAG), works purely on the time dimension, reconstructing the images working on each time series of each pixel separately. In particular, the method uses a weighted average of the samples available in the original time series to reconstruct the missing values. Enforcing the annual seasonality of each band as a prior, for the reconstruction of each missing sample in the time series a higher weight is given to images that are collected exactly on integer multiples of a year. To avoid propagation of land cover changes in future or past images, higher weights are given to more recent images. Finally, to have a method that respects causality, only images from the past of each sample in the time series are used.

To have computational performance suitable for PB sized datasets the method has been implemented in C++ using a sequence of fast convolution methods and Hadamard products and divisions. The method has been applied to a bimonthly aggregated version of the global GLAD Landsat ARD-2 collection from 1997 to 2022, producing a 400 terabyte output dataset. The produced dataset will be used to generate maps for several biophysical parameters, such as Fraction of Absorbed Photosynthetically Active Radiation (FAPAR), normalized difference water index (NDWI) and bare soil fraction (BSF). The code is available as open source, and the result is fully reproducible.

References:

Potapov, Hansen, Kommareddy, Kommareddy, Turubanova, Pickens, ... & Ying  (2020). Landsat analysis ready data for global land cover and land cover change mapping. Remote Sensing, 12(3), 426.

Julien, & Sobrino (2019). Optimizing and comparing gap-filling techniques using simulated NDVI time series from remotely sensed global data. International Journal of Applied Earth Observation and Geoinformation, 76, 93-111.

Radeloff, Roy, Wulder, Anderson, Cook, Crawford, ... & Zhu (2024). Need and vision for global medium-resolution Landsat and Sentinel-2 data products. Remote Sensing of Environment, 300, 113918.

How to cite: Consoli, D., Parente, L., and Witjes, M.: A new methodology for time-series reconstruction of global scale historical Earth observation data, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-18061, https://doi.org/10.5194/egusphere-egu24-18061, 2024.

EGU24-18197 | ECS | Orals | NP4.1 | Highlight

The regularity of climate-related extreme events under global warming 

Karim Zantout, Katja Frieler, and Jacob Schewe and the ISIMIP team

Climate variability gives rise to many different kinds of extreme impact events, including heat waves, crop failures, or wildfires. The frequency and magnitude of such events are changing under global warming. However, it is less known to what extent such events occur with some regularity, and whether this regularity is also changing as a result of climate change. Here, we present a novel method to systematically study the time-autocorrelation of these extreme impact events, that is, whether they occur with a certain regularity. In studies of climate change impacts, different types of events are often studied in isolation, but in reality they interact. We use ensembles of global biophysical impact simulations from the Inter-Sectoral Impact Model Intercomparison Project (ISIMIP) driven with climate models to assess current conditions and projections. The time series analysis is based on a discrete Fourier transformation that accounts for the stochastic fluctuations from the climate model. Our results show that some climate impacts, such as crop failure, indeed exhibit a dominant frequency of recurrence; and also, that these regularity patterns change over time due to anthropogenic climate forcing.

How to cite: Zantout, K., Frieler, K., and Schewe, J. and the ISIMIP team: The regularity of climate-related extreme events under global warming, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-18197, https://doi.org/10.5194/egusphere-egu24-18197, 2024.

EGU24-18210 | ECS | Posters on site | NP4.1

Long-term vegetation development in context of morphodynamic processes since mid-19th century 

Katharina Ramskogler, Moritz Altmann, Sebastian Mikolka-Flöry, and Erich Tasser

The availability of comprehensive aerial photography is limited to the mid-20th century, posing a challenge for quantitatively analyzing long-term surface changes in proglacial areas. This creates a gap of approximately 100 years, spanning the end of the Little Ice Age (LIA). Employing digital monoplotting and historical terrestrial images, our study reveals quantitative surface changes in a LIA lateral moraine section dating back to the second half of the 19th century, encompassing a total study period of 130 years (1890 to 2020). With the long-term analysis at the steep lateral moraines of Gepatschferner (Kauner Valley, Tyrol, Austria) we aimed to identify changes in vegetation development in context with morphodynamic processes and the changing climate.

In 1953, there was an expansion in the area covered by vegetation, notably encompassing scree communities, alpine grassland, and dwarf shrubs. However, the destabilization of the system after 1980, triggered by rising temperatures and the resulting thawing of permafrost, led to a decline in vegetation cover by 2020. Notably, our observations indicated that, in addition to morphodynamic processes, the overarching trends in temperature and precipitation exerted a substantial influence on vegetation development. Furthermore, areas with robust vegetation cover, once stabilised, were reactivated and subjected to erosion, possibly attributed to rising temperatures post-1980.

This study demonstrates the capability of historical terrestrial images to enhance the reconstruction of vegetation development in context with morphodynamics in high alpine environments within the context of climate change. However, it is important to note that long-term mapping of vegetation development through digital monoplotting has limitations, contingent on the accessibility and quality of historical terrestrial images, as well as the challenges posed by shadows in high alpine regions. Despite these limitations, this long-term approach offers fundamental data on vegetation development for future modelling efforts.

How to cite: Ramskogler, K., Altmann, M., Mikolka-Flöry, S., and Tasser, E.: Long-term vegetation development in context of morphodynamic processes since mid-19th century, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-18210, https://doi.org/10.5194/egusphere-egu24-18210, 2024.

EGU24-19601 | ECS | Posters on site | NP4.1

Discrimination of  geomagnetic quasi-periodic signals by using SSA Transform 

Palangio Paolo Giovanni and Santarelli Lucia

Discrimination of  geomagnetic quasi-periodic signals by using SSA Transform

  • Palangio1, L. Santarelli 1

1Istituto Nazionale di Geofisica e Vulcanologia L’Aquila

3Istituto Nazionale di Geofisica e Vulcanologia Roma

 

Correspondence to:  lucia.santarelli@ingv.it

 

Abstract

In this paper we present an application of  the SSA Transform to the detection and reconstruction of  very weak geomagnetic signals hidden in noise. In the SSA Transform  multiple subspaces are used for representing and reconstructing   signals and noise.  This analysis allows us to reconstruct, in the time domain, the different harmonic components contained in the original signal by using  ortogonal functions. The objective is to identificate the dominant  subspaces that can be attributed to the  signals and the subspaces that can be attributed to the noise,  assuming that all these  subspaces are orthogonal to each other, which implies that the  signals and noise  are independent of one another. The subspace of the signals is mapped simultaneously on several spaces with a lower dimension, favoring the dimensions that best discriminate the patterns. Each subspace of the signal space is used to encode different subsets of functions having common characteristics, such as  the same periodicities. The subspaces  identification was performed by using singular value decomposition (SVD) techniques,  known as  SVD-based identification methods  classified in a subspace-oriented scheme.The  quasi-periodic variations of geomagnetic field  has been investigated in the range of scale which span from 22 years to 8.9 days such as the  Sun’s polarity reversal cycle (22 years), sun-spot cycle (11 years), equinoctial effect (6 months), synodic rotation of the Sun (27 days) and its harmonics. The strength of these signals vary from fractions of a nT to tens of nT. Phase and frequency variability of these cycles has been evaluated from the range of variations in the geomagnetic field recorded at middle latitude place (covering roughly 4.5 sunspot cycles). Magnetic data recorded at L'Aquila Geomagnetic observatory (geographic coordinates: 42° 23’ N, 13° 19’E, geomagnetic coordinates: 36.3° N,87°.2 E, L-shell=1.6) are used from 1960 to 2009.

 

 

How to cite: Paolo Giovanni, P. and Lucia, S.: Discrimination of  geomagnetic quasi-periodic signals by using SSA Transform, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-19601, https://doi.org/10.5194/egusphere-egu24-19601, 2024.

EGU24-22262 | ECS | Posters on site | NP4.1

Temporal Interpolation of Sentinel-2 Multispectral Time Series in Context of Land Cover Classification with Machine Learning Algorithms 

Mate Simon, Mátyás Richter-Cserey, Vivien Pacskó, and Dániel Kristóf

Over the past decades, especially since 2014, large quantities of Earth Observation (EO) data became available in high spatial and temporal resolution, thanks to ever-developing constellations (e.g.: Sentinel, Landsat) and open data policy. However, in the case of optical images, affected by cloud coverage and the spatially changing overlap of relative satellite orbits, creating temporally generalized and dense time series by using only measured data is challenging, especially when studying larger areas.

Several papers investigate the question of spatio-temporal gap filling and show different interpolation methods to calculate missing values corresponding to the measurements. In the past years more products and technologies have been constructed and published in this field, for example Copernicus HR-VPP Seasonal Trajectories (ST) product.  These generalized data structures are essential to the comparative analysis of different time periods or areas and improve the reliability of data analyzing methods such as Fourier transform or correlation. Temporally harmonized input data is also necessary in order to improve the results of Machine Learning classification algorithms such as Random Forest or Convolutional Neural Networks (CNN). These are among the most efficient methods to separate land cover categories like arable lands, forests, grasslands and built-up areas, or crop types within the arable category.

This study analyzes the efficiency of different interpolation methods on Sentinel-2 multispectral time series in the context of land cover classification with Machine Learning. We compare several types of interpolation e.g. linear, cubic and cubic-spline and also examine and optimize more advanced methods like Inverse Distance Weighted (IDW) and Radial Basis Function (RBF). We quantify the accuracy of each method by calculating mean square error between measured and interpolated data points. The role of interpolation of the input dataset in Deep Learning (CNN) is investigated by comparing Overall, Kappa and categorical accuracies of land cover maps created from only measured and interpolated time series. First results show that interpolation has a relevant positive effect on accuracy statistics. This method is also essential in taking a step towards constructing robust pretrained Deep Learning models, transferable between different time intervals and agro-ecological regions.

The research has been implemented with the support provided by the Ministry of Culture and Innovation of Hungary from the National Research, Development and Innovation Fund, financed under the KDP-2021 funding scheme.

 

Keywords: time series analysis, Machine Learning, interpolation, Sentinel

How to cite: Simon, M., Richter-Cserey, M., Pacskó, V., and Kristóf, D.: Temporal Interpolation of Sentinel-2 Multispectral Time Series in Context of Land Cover Classification with Machine Learning Algorithms, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-22262, https://doi.org/10.5194/egusphere-egu24-22262, 2024.

EGU24-971 | ECS | Posters on site | ITS1.11/NP4.2

Terrestrial Water Storage Reconstruction: A Causal Inference Approach 

Vivek Kumar Yadav and Bramha Dutt Vishwakarma

The water availability in a region is driven by the water cycle, which is changing quickly in response to climate change and direct human interventions. The water cycle is defined and controlled by the variation in water fluxes such as Precipitation (P), Evapotranspiration (Et), Runoff (R), and Storage change (ΔS). Out of these water fluxes, ΔS is a key variable for ecosystem habitability and surviving droughts. It is an important parameter in drafting water management policy, but due to lack of long and reliable data the impact of climate change on ΔS is yet to be understood. The only Global observations of Terrestrial water storage (TWS) are available from GRACE satellite mission since 2002 at monthly scale.

Although GRACE data has transformed hydrological science significantly, its short time series limits usage of GRACE for climate change analysis of hydrological fluxes (closing the multidecadal water budget and sea level budget, understanding the spatiotemporal evolution of water availability, and so on). To tackle this, several studies have attempted reconstructing ΔS prior to GRACE period. These studies employ either hydrological modelling of ΔS, statistical regression,  or machine learning techniques. While machine learning methods have been assessed superior, they suffer from issues such as a lack of explainability, failure to identify causal drivers of TWS change, and use of short time series for feature extraction and training leading to poor or no representation of decadal natural variability.

Furthermore, in all the studies till now, representation of local human activities, such as ground water extraction or reservoir operation,  was either absent or assumed to be a linear trend. Here we revisit a reconstruction method by Humphrey et al., 2017 and show that these approximations have a considerable impact on the quality of reconstruction. Then we propose a multivariate regression model that relates selected hydrometeorological variables with TWS. These variables are identified from causal analysis of JULES model outputs. We show that temperature has a very weak relation with TWS compared to precipitation. The causal inference based model is able to capture realistic variability in reconstructed TWS. Our TWS reconstruction for the Ganges basin outperforms the contemporary attempts and is able to identify the drivers for interannual changes in TWS . The results bring historical perspective to the current state of water resources in the basin and provide context for design of future water resources policy.

How to cite: Yadav, V. K. and Vishwakarma, B. D.: Terrestrial Water Storage Reconstruction: A Causal Inference Approach, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-971, https://doi.org/10.5194/egusphere-egu24-971, 2024.

EGU24-1838 | Posters on site | ITS1.11/NP4.2

A comparison of two causal methods in the context of climate analyses 

David Docquier, Giorgia Di Capua, Reik V. Donner, Carlos A. L. Pires, Amélie Simon, and Stéphane Vannitsem

Correlation does not necessarily imply causation, and this is why causal methods have been developed to try to disentangle true causal links from spurious relationships. In our study, we use two causal methods, namely the Liang-Kleeman information flow (LKIF) and the Peter and Clark momentary conditional independence (PCMCI) algorithm, and apply them to four different artificial models of increasing complexity and one real-case study based on climate indices in the North Atlantic and North Pacific. We show that both methods are superior to the classical correlation analysis, especially in removing spurious links. LKIF and PCMCI display some strengths and weaknesses for the three simplest models, with LKIF performing better with a smaller number of variables, and PCMCI being best with a larger number of variables. Detecting causal links from the fourth model is more challenging as the system is nonlinear and chaotic. For the real-case study with climate indices, both methods present some similarities and differences at monthly time scale. One of the key differences is that LKIF identifies the Arctic Oscillation (AO) as the largest driver, while El Niño-Southern Oscillation (ENSO) is the main influencing variable for PCMCI. More research is needed to confirm these links, in particular including nonlinear causal methods.

How to cite: Docquier, D., Di Capua, G., Donner, R. V., Pires, C. A. L., Simon, A., and Vannitsem, S.: A comparison of two causal methods in the context of climate analyses, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-1838, https://doi.org/10.5194/egusphere-egu24-1838, 2024.

Causality analysis is an important and old problem lying at the heart of scientific research. Causality analysis based on data, however, is a relatively recent development. Traditionally causal inference has been classified as a field in statistics. Motivated by the predictability problem in physical science, it is found that causality in terms of information flow/transfer is actually a real notion in physics that can be derived ab initio, rather than axiomatically proposed as an ansatz, and, moreover, can be quantified. A comprehensive study with generic systems (both deterministic and stochastic) has just been fulfilled, with explicit formulas attained in closed form (Liang, 2016). These formulas are invariant upon nonlinear coordinate transformation, indicating that the so-obtained information flow should be an intrinsic physical property. The principle of nil causality that reads, an event is not causal to another if the evolution of the latter is independent of the former, which all formalisms seek to verify in their respective applications, turns out to be a proven theorem here. In the linear limit, its maximum likelihood estimator is concise in form, involving only the commonly used statistics, i.e., sample covariances. An immediate corollary is that causation implies correlation, but the converse does not hold, expressing the long standing philosophical debate ever since Berkeley (1710) in a transparent mathematical expression.

The above rigorous formalism has been validated with benchmark systems like baker transformation, Hénon map, stochastic gradient system, and with causal networks in extreme situations such as those buried in heavy noises and those with nodes almost synchronized (e.g., Liang, 2021), to name a few. They have also been applied to real world problems in the diverse disciplines such as climate science, dynamic meteorology, turbulence, neuroscience, financial economics, quantum mechanics, etc., with interesting new findings. For example, Stips et al. (216) found that, while CO2 emission does drive the recent global warming, on a paleoclimate scale, it is global warming that drives the CO2 emission; PNA, a teleconnection pattern related to the inclement weather in North America, may trace a part of its origin to a rather limited local marginal sea far away in Asia. Besides, with the above causality analysis, pollution sourcing (particularly PM2.5) may be conducted in a rather effective way via causal graph reconstruction. If time permits, I will also present an ongoing application to the development of causal AI algorithms to overcome the interpretability crisis, and a recent remarkable exercise with such an algorithm in the forecasting of El Niño Modoki, a climate mode linked to hazards in far-flung regions of the globe.

 

References:

Liang, 2014: Unraveling the cause-effect relation between time series. Phy. Rev. E,  90, 052150.

Liang, 2016: Information flow and causality as rigorous notions ab initio. Phys. Rev. E, 94, 052201.

Liang, 2021: Normalized multivariate time series causality analysis and causal graph reconstruction. Entropy, 23, 679.

Liang et al., 2021: El Niño Modoki can be mostly predicted more than 10 years ahead of time. Nature Sci. Rep. 11:17860

 

How to cite: Liang, X. S.: Causality as a real physical notion ab initio, and its applications in Earth system sciences, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-2618, https://doi.org/10.5194/egusphere-egu24-2618, 2024.

EGU24-2948 | Orals | ITS1.11/NP4.2 | Highlight

Causal methods for climate extremes 

Sebastian Engelke

The talk discusses a critical topic in climate science: understanding how interventions on our climate system influence the likelihood of extreme events. The focus is on methodologies that enable causal attribution of such events to specific drivers, rather than merely predicting their occurrence. We discuss common practices and highlight the use of recent statistical methods that are applicable when only observational data is available, as opposed to model-based data. The talk defines the concept of a causal effect of a treatment (such as changes in flood infrastructure or increased CO2 emissions) on extreme outcomes (like a one in 100 year flood). We also cover the estimation of these effects amidst confounding factors and the assessment of associated uncertainties. Finally, we discuss the inherent challenges of applying causal inference to extreme climate events. 

How to cite: Engelke, S.: Causal methods for climate extremes, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-2948, https://doi.org/10.5194/egusphere-egu24-2948, 2024.

EGU24-4191 | ECS | Posters on site | ITS1.11/NP4.2

Causal effects of teleconnection patterns on soil moisture through different climate paths over the Greater Horn of Africa 

Wen Zhuo, Shibo Fang, Xinran Gao, Ricardo B. Lourenco, Yanru Yu, Jiahao Han, and Alemu Gonsamo

Soil moisture is undoubtably a vital variable of the climate system. Understanding the interactions among atmosphere, climate, and soil is necessary for water resource management, drought monitoring, and disaster prevention. However, evaluation of those interactions so far primarily focused on typical correlation analysis which often fail to imply causal relationship due to autocorrelation and high dimensionality within time series variables. Here, we used a data driven causal inference method called PCMCI+ to discover causal relationships among teleconnection patterns (El Niño Southern Oscillation (ENSO) and Indian Ocean Dipole (IOD)), climate variables (precipitation and temperature) and soil moisture during 1980-2022 over Great Horn of Africa (GHOA), where is a susceptible region to suffer from severe drought. Further, we quantitative calculated the causal effects of teleconnection patterns on SM through different climate paths. Results suggest that IOD generally presents higher causal effects on climate variables (temperature and precipitation) or on soil moisture through both precipitation and temperature paths than ENSO over most parts of GHOA. Moreover, precipitation performs shorter lag effect and greater causal effect on soil moisture in GHOA. Our study provides the first attempt to quantitatively analyze the causal effects of teleconnection patterns on SM through both precipitation path and temperature path, and it highlights the causal relationships within atmosphere-climate-soil interactions, which could help for better understanding of climate change impact on drought over GHOA.

How to cite: Zhuo, W., Fang, S., Gao, X., Lourenco, R. B., Yu, Y., Han, J., and Gonsamo, A.: Causal effects of teleconnection patterns on soil moisture through different climate paths over the Greater Horn of Africa, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-4191, https://doi.org/10.5194/egusphere-egu24-4191, 2024.

EGU24-4315 | Orals | ITS1.11/NP4.2

Evaluation of Shannon Entropy-based Information transfer in nonlinear systems  

Carlos Pires, Stéphane Vannitsem, and David Docquier

We present a general theory for computing and estimating Shannon entropy-based information transfer in nonlinear stochastic systems driven by deterministic forcings and additive and/or multiplicative noises, by extending the Liang-Kleeman framework of causality inference to nonlinear cases. The method presents effective and computable formulas of the rates of information transfer between sets of causal and consequential system variables, relying on the evaluation of conditional expectations of the deterministic and stochastic forcings (Causal Sensitivity Method: CSM). The CSM can work with a) ensemble model runs, b) system time series in ergodic conditions and c) time series without a priori knowledge of model equations. The CSM also allows to express the information transfer parcels, which are attributable either to one-to-one interactions or to synergies across groups of variables and assess where the information is more relevant in the state space. The CSM is tested in two proof-of-concept low-order models: 1) a nonlinear model derived from a potential function and 2) the classical chaotic Lorenz model, both forced by additive and/or multiplicative noises. The CSM is also tested with a nonlinear regression model of the ice-cover time evolution, forced by radiation. The CSM estimation is much more robust and efficient than methods using the stochastic model’s full probability density function and its derivatives, whose estimation is rather unreliable in case of short data availability. The analysis also demonstrates that the CSM estimation is computationally cheap in the different experiments, providing evidence of the possibilities and generalizations offered by the method, thus opening new perspectives on real-world applications. This work was funded by the Portuguese Fundação para a Ciência e a Tecnologia (FCT) I.P./MCTES through national funds (PIDDAC) – UIDB/50019/2020(https://doi.org/10.54499/UIDB/50019/2020),UIDP/50019/2020(https://doi.org/10.54499/UIDP/50019/2020) and LA/P/0068/2020 (https://doi.org/10.54499/LA/P/0068/2020) and the project  JPIOCEANS/0001/2019 (ROADMAP).

 

How to cite: Pires, C., Vannitsem, S., and Docquier, D.: Evaluation of Shannon Entropy-based Information transfer in nonlinear systems , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-4315, https://doi.org/10.5194/egusphere-egu24-4315, 2024.

EGU24-4693 | ECS | Posters on site | ITS1.11/NP4.2

Granger causality in tail 

Juraj Bodik

Granger causality plays a pivotal role in uncovering directional relationships among time-varying variables and enhancing decision-making in complex systems. While this notion gains heightened importance during extreme events in highly volatile periods,
state-of-the-art methods primarily focus on causality within the body of the distribution. We introduce a new rigorous mathematical framework for “Granger causality in tail,” designed to evaluate whether an extreme event in one time series causes a corresponding extreme event in another. Moreover, we describe how we can quantify the magnitude of the causal impact of an extreme event on other variables. 

We establish equivalences between our Granger causality in tail and other causal concepts, including “classical Granger causality,” “Sims causality,” and “structural causality.” By proving the key properties of Granger causality in tail, we assert its usefulness in high-dimensional complex systems with potential hidden confounders. Here, to model the tails of the variables, we utilize the “extreme value theory” framework. We also propose an inference method for detecting the presence of Granger causality in tail and provide insights into the asymptotic properties of our estimator within the framework of a stochastic recurrence equation (SRE) model.

How to cite: Bodik, J.: Granger causality in tail, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-4693, https://doi.org/10.5194/egusphere-egu24-4693, 2024.

EGU24-6535 | Orals | ITS1.11/NP4.2

Unfolding the Manifold Flavours of Causality 

Rui A. P. Perdigão

The present communication provides a contribution to an overarching cross-methodological causality investigation, encompassing a methodological synergy among physical, analytical, information-theoretic and systems intelligence approaches to causal discovery and quantification in complex system dynamics. These efforts methodologically lead to the emergence of a broader causal framework, valid not only in classical recurrence-based dynamical systems, but also on the generalized information physics of non-ergodic coevolutionary spatiotemporal complexity.

This study begins with a comprehensive cross-examination of causality metrics derived from these diverse domains, by synthesizing causality insights from information theory, which enables the quantification of information flow among variables; differential geometry, which captures the curvature and structure of causal relationships; dynamical systems, which analyze the temporal evolution of systems and associated kinematic geometric properties; and fundamental physical metrics, which elucidate causal connections in the physical world from fundamental thermodynamic principles. Through this analysis, we aim to deepen our understanding of causality in complex systems, with physical process understanding and geophysical applications in mind.

Having laid out some of the key methodological flavours of causality, the present communication introduces new metrics further contributing to a broader non-Shannonian information theoretic causality pool of methods, along with additional advances on quantum thermodymamical, nonlinear statistical mechanical, differential geometric and topologic approaches on causality. Positioning ourselves in a broader nonlinear non-Gaussian non-ergodic setting to tackle far-from-equilibrium structural-functional coevolution and synergistic emergence in complex system dynamics, our derivations further contribute to a new generation of information theoretic, dynamical systems and non-equilibrium thermodynamic causality approaches, along with their synergistic articulation towards a unified framework. This brings out further cross-methodological comparability, portability and complementary insights on dealing with the intricate causality of complex multiscale far-from-equilibrium Earth system dynamic phenomena.

By unveiling manifold flavours of causality and their interconnections, this study brings out their commonalities, synergies, and further potential synergistic applications across disciplines. This interdisciplinary approach not only enhances our theoretical understanding of causality but also provides practical implications for applications in fields such as data science, network theory, and complex systems analysis, with direct relevance across the Earth system sciences and beyond.

How to cite: Perdigão, R. A. P.: Unfolding the Manifold Flavours of Causality, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-6535, https://doi.org/10.5194/egusphere-egu24-6535, 2024.

EGU24-6584 | Posters on site | ITS1.11/NP4.2

Causal discovery among wind-related variables in a wind farm under extreme wind speed scenarios: Comparison of results using Granger causality and interactive k-means clustering 

Katerina Schindlerova (Hlavackova-Schindler), Kejsi Hoxhallari, Luis Caumel Morales, Irene Schicker, and Claudia Plant

Using the era5 meteorological reanalysis data from 2000 to 2020 [1], we investigate temporal effects of ten wind related processes in time intervals of extreme wind speed values, extracted and corrected towards wind turbine locations for a wind farm in Andau, Austria.  We approach the problem by two ways, by the Granger causal inference, namely by the heterogeneous Graphical Granger model (HMML) [2] and by clustering, namely by the interactive k-means clustering (IKM) [3].

We investigate six scenarios based on the hydrological half-year, a moderate wind speed and time intervals of low or high extreme wind speed in the farm. In case of HMML, we discover causal variables and their values for each scenario.  Regarding the method IKM, it is used for three clusters (clusters for a moderate wind speed and for a low and high extreme wind speed) to find coefficient representations of each interacting variable with respect to the wind speed in each of the six scenarios.   We compare the results of both methods in terms of the values of causal variables and of the values of the coefficients of representation and evaluate the interpretability of the discovered causal connections with the expert meteorological knowledge.

 [1]  https://cds.climate.copernicus.eu/cdsapp#!/dataset/reanalysis-era5-pressure levels?tab=overview   

[2] Hlaváčková-Schindler, K., Plant, C. (2020) Heterogeneous graphical Granger causality by minimum message length, Entropy, 22(1400). pp. 1-21 ISSN 1099-4300 MDPI (2020).

[3] Plant, C., Zherdin, A., Sorg, C., Meyer-Baese, A., Wohlschläger, A. M. Mining interaction patterns among brain regions by clustering. IEEE Transactions on Knowledge and Data Engineering, 26(9):2237–2249, 2014.

How to cite: Schindlerova (Hlavackova-Schindler), K., Hoxhallari, K., Caumel Morales, L., Schicker, I., and Plant, C.: Causal discovery among wind-related variables in a wind farm under extreme wind speed scenarios: Comparison of results using Granger causality and interactive k-means clustering, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-6584, https://doi.org/10.5194/egusphere-egu24-6584, 2024.

EGU24-7546 | Posters on site | ITS1.11/NP4.2

Quantifying the influence of cloud controlling factors with causal inference 

Lisa Bock, Adrian McDonalds, Axel Lauer, and Jakob Runge

As a key component of the hydrological cycle and the Earth’s radiation budget, clouds play an important role in both weather and climate. Our incomplete understanding of clouds and their role in cloud-climate feedbacks leads to large uncertainties in climate simulations. Using causal discovery as an unsupervised machine learning method we aim to systematically analyse and quantify causal interdependencies and dynamical links between cloud properties and their controlling factors. This approach goes beyond correlation-based measures by systematically excluding common drivers and indirect links. By estimating the causal effect of each of the cloud controlling factors for different cloud regimes we expect to be able to better understand the dominant processes which determine the micro- and macro-physical properties of clouds.

Specifically, causal inference is used to investigate the links between cloud properties such as cloud cover, cloud water path, cloud top height and cloud radiative effects and so-called cloud controlling factors, i.e., quantities that impact cloud formation and temporal evolution of the cloud (e.g., sea surface temperature, water vapour path and lower tropospheric stability). For this, causal networks are calculated from time series of these variables from satellite and reanalysis datasets averaged over different geographical regions and cloud regimes in order to quantify the strength of the individual links in the resulting causal graph by applying causal effect estimation.

How to cite: Bock, L., McDonalds, A., Lauer, A., and Runge, J.: Quantifying the influence of cloud controlling factors with causal inference, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-7546, https://doi.org/10.5194/egusphere-egu24-7546, 2024.

Many approaches to infer causal relations from time series in Earth sciences have been proposed and applied in order to identify diverse interactions, such as the influence of large-scale circulation modes on local temperature and precipitation, variability of Euroasian winters due to changing Arctic Sea ice cover, or interactions of solar activity and interplanetary medium conditions with the Earth’s magnetosphere-ionosphere systems. The methods usually depend on “dimensions” in which the understanding of underlying phenomena is located: The phenomena or processes can be linear or nonlinear; deterministic, or random. The third abstract “dimension” is the actual dimensionality of the problem, given either by the dimension of the state space of the underlying mechanism or the number of involved variables. We will conduct a short flight inside these “dimensions,” shedding light on some of the shades, comparing some of the causality inference methods using model and real data from the Earth sciences.

This study was supported by the Czech Academy of Sciences, Praemium Academiae awarded to M. Paluš and the Czech–Chinese Academies of Sciences Mobility Plus Project NSFC-23-08.

How to cite: Palus, M.: Many shades in three dimensions and parallel universes of causality analysis, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-8450, https://doi.org/10.5194/egusphere-egu24-8450, 2024.

EGU24-10448 | Posters on site | ITS1.11/NP4.2

Exploring Global and Local Water Scarcity Dynamics through Causal Discovery and Structural Causal Models 

Myrthe Leijnse, Marc F.P. Bierkens, and Niko Wanders

Water scarcity is driven by diverse natural and anthropogenic factors and represents a critical global challenge. Structural Causal Models are powerful tools to reveal the intricate interactions among social, ecological and hydrological components within human-water systems affected by water scarcity. This study integrates causal thinking into statistical and data-driven hydrological modelling, offering a different perspective on understanding system dynamics affecting water resources in water-scarce regions, the so-called water scarcity hotspots.

In this work we apply causal discovery methods to independent timeseries of sectoral water demand, social-economic variables, meteorological drivers and groundwater depletion to obtain a causal network representing human-water system interactions at global water scarcity hotpots. To derive this network we use global datasets and advanced causal network learning algorithms, specifically (Joint-)PCMCI (Runge et al., 2023). Recognizing the importance of large data sample sizes for a robust global causal network, we further extend our approach to construct a causal network specific to one of the water scarcity hotspots (California), using more detailed local data. Therefore, our framework provides a comprehensive understanding of water scarcity dynamics including both global and local scales. Through a comparative analysis of network outcomes derived from global datasets with those specific to California, we evaluate the effectiveness of our causal inference modelling framework.

After conducting and evaluating the causal networks at global and local scale, we applied methods from structural causal modelling and statistical machine learning to estimate causal effects of anthropogenic or natural system changes on water availability at water scarcity hotspots. This framework allows us to answer important (counterfactual) questions, such as understanding how the rate of unsustainable groundwater abstraction is affected by shifts in water management practices e.g., a reduction in irrigated cropland area.

As such, this work contributes to understanding how using causal inference methods are valuable to modelling of water scarcity, ultimately providing input to informed decision-making in water resource management and finding strategies to mitigate water scarcity impacts.

Runge, J., Gerhardus, A., Varando, G., Eyring, V., & Camps-Valls, G. (2023). Causal inference for time series. Nature Reviews Earth & Environment4(7), 487-505.

How to cite: Leijnse, M., Bierkens, M. F. P., and Wanders, N.: Exploring Global and Local Water Scarcity Dynamics through Causal Discovery and Structural Causal Models, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-10448, https://doi.org/10.5194/egusphere-egu24-10448, 2024.

EGU24-11714 | ECS | Posters on site | ITS1.11/NP4.2

Subseasonal prediction of heatwaves in the Iberian Peninsula using causality-based transformer networks. 

Cas Decancq, Daniel Hagan, Victoria Deman, Akash Koppa, and Diego Miralles

Subseasonal prediction of heatwaves, although highly valuable for risk reduction, is challenging because heatwave onsets and propagation are complex processes with both fast and slow drivers from local to global scale. Traditionally, subseasonal forecasting relies heavily on dynamical model ensembles, which are complex and of high computational cost. As an alternative, machine learning provides potentially performant solutions that may match or even outperform these physical-based models. Transformers, in particular, are the current state-of-the-art deep learning infrastructures, and using multi-head-attention allows them to keep track of long-term complex dependencies in timeseries data. However, to better forecast heatwaves subseasonally, it is essential to move beyond purely predictor-to-target associative measures when identifying the sources of predictability. Such endeavours require causal frameworks that provide directionality and explainable power for the predictor-to-target relationships.

This study seeks to implement the PCMCI+ (Runge, 2020) framework to identify causal drivers of heatwaves on the Iberian Peninsula on a subseasonal scale. Causally-selected predictors are employed to forecast the occurence of heatwaves up to six weeks in advance using transformer networks, both for different seasons and sub-regions in the Iberian Peninsula. Preliminary results reveal heatwaves can be predicted with reasonable accuracy with a forecast window of six weeks, particularly in water limited regions, using causality-based machine learning.


Reference:

Runge, J. (2020). Discovering contemporaneous and lagged causal relations in autocorrelated nonlinear time series datasets. In Conference on Uncertainty in Artificial Intelligence, pages 1388–1397. PMLR.

How to cite: Decancq, C., Hagan, D., Deman, V., Koppa, A., and Miralles, D.: Subseasonal prediction of heatwaves in the Iberian Peninsula using causality-based transformer networks., EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-11714, https://doi.org/10.5194/egusphere-egu24-11714, 2024.

EGU24-12204 | Orals | ITS1.11/NP4.2

Some alternative metods for causal discovery 

András Telcs

Causal inference is indeed a challenging endeavor, particularly when applied to observational studies of interacting systems. Perl's theory, along with the PC algorithm on directed acyclic graphs, and its extensions PCMCI and FCI, are powerful tools. However, their application to time series is time-consuming, and they still struggle to distinguish Markov-equivalent scenarios.

In our talk, we will present some methods based on principles that are partly or fully different from those underlying the aforementioned tools. Due to time constraints, we will focus on the main principles that allow the discovery of causal relations between a pair of systems, including hidden common causes (referred to as common drivers or confounders in different schools of thought). We won't delve into the numerous technical challenges due to the time limit.

How to cite: Telcs, A.: Some alternative metods for causal discovery, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-12204, https://doi.org/10.5194/egusphere-egu24-12204, 2024.

In 2022, La Niña and negative Indian Ocean Dipole (IOD) coincided, causing abnormally warm sea surface conditions in the eastern Indian Ocean (near Indonesia). This provided additional moisture to feed monsoon depressions, resulting in heavy rainfall in Pakistan. El Niño-Southern Oscillation (ENSO) and Indian Ocean Dipole (IOD) are two modes of sea surface temperature variability that can significantly impact precipitation in Pakistan's Upper Indus Basin. The current study used in situ observations and reanalysis ERA 5 precipitation data to determine the causal influence of ENSO and IOD on precipitation variability using an information-theoretic generalization of Granger causality. The predicted causal effect and causal delay obtained using conditional mutual information, a.k.a. transfer entropy, were further validated using conditional means ("composites") - precipitation means computed for different ENSO states; El Niño (positive), La Niña (negative), and neutral. Uncovering the causal and delayed effects of ENSO and IOD, as well as associated mechanisms, on subsequent precipitation in the UIB could provide a stronger foundation for improving seasonal climate predictions with a longer lead time, as well as understanding how regional and large-scale drivers affect regional precipitation.

This study was supported by the Czech Academy of Sciences, Praemium Academiae awarded to M. Paluš and the Czech–Chinese Academies of Sciences Mobility Plus Project NSFC-23-08.

How to cite: Latif, Y. and Palus, M.: Causal information flow and information transfer delay from ENSO and IOD to precipitation variability in the Upper Indus Basin, Pakistan, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-12884, https://doi.org/10.5194/egusphere-egu24-12884, 2024.

EGU24-13220 | Posters on site | ITS1.11/NP4.2

Scaling properties of irreversibility indices in turbulence 

François G. Schmitt

In 3D turbulence there is a flux of energy from large to small scales in the inertial range, associated with irreversibility, i.e. a breaking of the time reversal symmetry (Pumir, 2016). Such turbulent flows are characterized by scaling properties and we consider here how irreversibility depends on the scale. Two indicators of irreversibility for time series are tested involving triple correlations in a non-symmetric way. The first one proposed by Pomeau (1982, 2004) is: Po(r)=<X(t)X(t+r)X(t+3r)>-<X(t)X(t+2r)X(t+3r)>, where r is an increment and X(t) is the turbulent velocity which is stationary with zero mean. The second indicator has been proposed in the finance literature (Ramsey and Rothman, 1996), and was called symmetric bicovariance function: γ(r) = <X2(t)X(t+r)>-<X(t)X2(t+r)>. For time reversible processes, both indicators are zero, whereas their departure from 0 is an indicator of irreversibility.

We study these indicators applied to fully developed turbulence time series, from flume tank, wind tunnel and atmospheric turbulence databases. It is found that irreversibility occurs in the inertial range and has scaling properties with slopes close to one. A maximum value is found around the injection scale. This confirms that the irreversibility is associated with the turbulent cascade in the inertial range and shows that the irreversibility is maximal at the injection scale, the largest scale of the turbulent cascade.

This is published in Schmitt, F.G., Scaling analysis of time-reversal asymmetries in fully developed turbulence, Fractal and Fractional, 7(8), 630, 2023.  https://doi.org/10.3390/fractalfract7080630

Cited references: Pumir et al., Phys. Rev. Lett.. 116, 124502 (2016); Pomeau, J. de Physique 43, 859 (1982); Pomeau, Lect. Notes Phys. 644, 425 (2004); Ramsey and Rothman, J. Money Credit Bank. 28, 1 (1996).

How to cite: Schmitt, F. G.: Scaling properties of irreversibility indices in turbulence, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-13220, https://doi.org/10.5194/egusphere-egu24-13220, 2024.

EGU24-14429 | ECS | Posters on site | ITS1.11/NP4.2

Compression-complexity based estimation of Causality: Applications in Earth and Climate Sciences 

Aditi Kathpalia, Pouya Manshour, and Milan Paluš

Many approaches to time series causality exist and have been inspired from fields such as statistics, information theory, physics and topology. We have proposed a method called compression-complexity causality (CCC) [1] inspired from the field of data compression in computer science. It is based on the idea that the compressibility of the ‘effect’ time series changes when the ‘cause’ time series is considered in the evolution of the future dynamics of the effect. Compressibility is estimated using compression-complexity estimator for time series called ‘effort-to-compress’, which employs a lossless data compression algorithm for complexity estimation. CCC makes minimal assumptions on given time series data and has been seen to work well for short length data, irregularly sampled data as well as data with low temporal resolution. We have also introduced a multidimensional version of CCC, called Permutation CCC (PCCC) [2], which uses Takens’ embedding for appropriate high dimensional representation of time series. This representation is subsequently encoded using ordinal patterns before computation of CCC. PCCC formulation retains the original robustness of CCC. This is demonstrated with its application on simulated multidimensional systems. We apply this formulation to infer causality between CO2 emissions – temperature recordings on three different time scales, El Niño–Southern Oscillation phenomena – South Asian Summer Monsoon on two different time scales, as well as North Atlantic Oscillations – European temperature recordings on two different time scales. These paleoclimate and climate datasets suffer from the issues of missing samples, low temporal resolution and short length data and so a reliable inference of these climatic interactions requires a robust causality estimator.  
Finally, we explore another variation of CCC which can help to infer causality in the multivariate cases. This variation helps to infer the existence of causal influences to a particular variable (from each other variable considered) while conditioning out the additional variables present. The presence of causal influences to each variable is decided by choosing the model of least compression-complexity which can help to explain the evolution of the future of that particular variable. In case more than one model has least complexity, the smallest model is chosen. We apply this formulation to understand interactions in space-weather system, particularly the solar wind-magnetosphere-ionosphere system interactions, which manifest as geomagnetic storms and substorms. We compare the performance of CCC formulations with existing methods in case of simulations as well as real data applications. 

This study is supported by the Czech Academy of Sciences, Praemium Academiae awarded to M. Paluš.

References:
[1] Kathpalia, A., & Nagaraj, N. (2019). Data-based intervention approach for Complexity-Causality measure. PeerJ Computer Science, 5, e196.
[2] Kathpalia, A., Manshour, P., & Paluš, M. (2022). Compression complexity with ordinal patterns for robust causal inference in irregularly sampled time series. Scientific Reports, 12(1), 14170.

How to cite: Kathpalia, A., Manshour, P., and Paluš, M.: Compression-complexity based estimation of Causality: Applications in Earth and Climate Sciences, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-14429, https://doi.org/10.5194/egusphere-egu24-14429, 2024.

EGU24-15830 | ECS | Posters on site | ITS1.11/NP4.2

Validating ENSO Feedbacks in Climate Models Using a Causal Discovery Method 

Emma Schultz, Dim Coumou, and Michael Massmann

The El Niño-Southern Oscillation (ENSO) stands out as the dominant driver of climate fluctuations on interannual timescales. As ENSO causes extreme weather events in the Pacific region and beyond, it has wide ranging socio-economic impacts. Over the past decades, a strengthening in the temperature gradient is observed between the Western and Eastern Pacific. However, climate model simulations do not depict this strengthening trend. Here we explore if the Bjerknes feedback is well represented in climate models, and if not whether this could explain the discrepancy between the observed and modeled trends. The Bjerkness feedback represents the dominant feedback processes between atmosphere and ocean that drive ENSO variability. A causal discovery method, the PCMCI algorithm, is used to construct causal networks of key variables in the Bjerknes feedback: near surface temperatures, sea level pressure and trade winds across the Pacific Ocean. Causal networks are constructed for time periods 1950-1982 and 1982-2014, based on both reanalysis data and climate model simulations. The observed changes between causal networks based on the early and later period are examined. The analysis reveals a strengthening causal influence of trade winds on sea level pressure and temperatures in networks based on reanalysis data. This significant strengthening trend is absent in networks based on climate model simulations. As an increased influence of the trade winds would have a cooling effect on Central and Eastern Pacific, this might explain why there is no observed warming in the Central and Eastern Pacific over the past decades, and thus a strengthened temperature gradient. The lack of this strengthening causal influence of trade winds in climate models might thus explain why the models do show a warming over the Eastern Pacific, weakening the temperature gradient.

How to cite: Schultz, E., Coumou, D., and Massmann, M.: Validating ENSO Feedbacks in Climate Models Using a Causal Discovery Method, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-15830, https://doi.org/10.5194/egusphere-egu24-15830, 2024.

EGU24-15950 | ECS | Posters on site | ITS1.11/NP4.2

Leveraging Information Flow for Data-Driven Subseasonal Forecasting of Sahelian Hot Extremes 

Victoria M. H. Deman, Daniel F. T. Hagan, Damián Insua-Costa, Akash Koppa, and Diego G. Miralles

The semi-arid Sahel region has witnessed an increase in extreme weather conditions such as repeated drought cycles, desertification, heatwaves and floods in recent decades. These events pose existential threats to the already vulnerable population and natural ecosystem. Addressing the underexplored potential of subseasonal forecasting in the Sahel, data-driven models offer an alternative to traditional dynamical approaches. These models – distinguished by enhanced computational efficiency, reduced sensitivity to initial conditions, the ability to learn intricate relationships from data, and the ability to capture nonlinear dynamics – represent an asset in building resilience in the region. 

This study investigates the potential of employing a rigorous causality framework based on the Liang–Kleeman information flow for predictor selection. Previous research has underscored the pitfalls of using correlations for predictor selection when forecasting using machine learning models, as spurious correlations may lead to the selection of predictors without any physical connection. In response, our research investigates the potential of this information flow causality to select predictors within a vast array of predefined variables, including coupled ocean–atmospheric oscillation indices, sea-surface temperatures, vegetation indices and soil moisture. Subsequently, our focus is directed towards predicting summer maximum temperature extremes with lead times of 2, 4, 8 and 16 weeks using the selected predictors and a variety of deep learning techniques. Despite the challenge of predicting short-lived heatwaves in a region characterised by the high baseline temperatures, our results indicate that the information flow causality effectively reduces dimensionality, and enables a selection of features with causal relationships that facilitates subsequent forecasting. In the following, the causal knowledge from the predictor selection step will be quantitatively transferred into the machine learning models themselves, thereby providing an interpretable framework for the prediction of the hot extremes in the region. 

How to cite: Deman, V. M. H., Hagan, D. F. T., Insua-Costa, D., Koppa, A., and Miralles, D. G.: Leveraging Information Flow for Data-Driven Subseasonal Forecasting of Sahelian Hot Extremes, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-15950, https://doi.org/10.5194/egusphere-egu24-15950, 2024.

EGU24-17312 | ECS | Orals | ITS1.11/NP4.2

Causal evaluation of humanitarian aid on food security 

Jordi Cerdà-Bautista, José María Tárraga, Vasileios Sitokonstantinou, and Gustau Camps-Valls

In a world where climate change is rapidly accelerating, droughts are becoming more frequent and severe, posing a serious challenge to food security in the most vulnerable regions. The Horn of Africa has witnessed a rise in acute malnutrition, affecting 6.5 million people in 2022 [1]. Prolonged dry spells significantly contribute to this crisis [2], yet it is crucial to recognize that droughts are not the sole driver. Various factors, including hydrological conditions, food production capabilities, market access, insufficient humanitarian aid, conflicts, and displacement, play a significant role [3,4]. Understanding the underlying causes of food insecurity is pivotal for improving the effectiveness of humanitarian actions, yet in this context, the study proves to be complex, involving multiple variables, scales, and non-linear relationships. Predictive Machine Learning (ML) techniques are not suited to understanding the causes and estimating the causal effect by default [5,6], instead, this study focuses on causal inference to quantify the impacts of climate and socioeconomic factors on food insecurity. Our key contributions involve discerning causal relationships within the intricate food security system, integrating a comprehensive database including socio-economic, weather and remote sensing data, and estimating the causal effect of humanitarian interventions on the food security index, the outcome of interest. The causal discovery task is performed via time series methods accounting for nonlinear and nonstationary relations, like the PCMCI algorithm and nonlinear Granger causality [7,8], identifying the drivers in the data that are causally linked to the outcome. Besides, the causal effect estimation task is performed via a Conditional Average Treatment Effect (CATE), gaining insights into the spatiotemporal heterogeneity of the impact of humanitarian interventions on the outcome [9]. Such endeavors are crucial for facilitating more efficient future interventions and policies, thereby improving transparency and accountability in humanitarian aid.

References

[1] WFP, “Impacts of the Cost of Inaction on WFP Food Assistance in Eastern Africa (2021 & 2022),” https://docs.wfp.org/api/documents/WFP-0000148305/download/, 2023.

[2] Coughlan de Perez E., et al, “From rain to famine: assessing the utility of rainfall observations and seasonal forecasts to anticipate food insecurity in East Africa,” Food Secur., vol. 11, no. 1, pp. 57–68, 2019.

[3] Maxwell D. et al, “Viewpoint: Determining famine: Multi-dimensional analysis for the twenty-first century,” Food Policy, vol. 92, 2020.

[4] Guy A. J. et al, “Climate, conflict and forced migration” Global Environmental Change, vol. 54, no. 4, 2019.

[5] Pearl J., “Causality: Models, reasoning, and inference,” Cambridge University Press, vol. 19, 2000.

[6] Peters J., Janzing D., and Schlkopf B., Elements of Causal Inference: Foundations and Learning Algorithms, The MIT Press, 2017.

[7] Runge, J.. "Discovering contemporaneous and lagged causal relations in autocorrelated nonlinear time series datasets." Conference on Uncertainty in Artificial Intelligence. PMLR, 2020.

[8] Camps-Valls, G. et al, “Discovering causal relations and equations from data”, Physics Reports 1044 :1--68, 2023

[9] Giannarakis, G. et al, (2022). Personalizing sustainable agriculture with causal machine learning. arXiv preprint arXiv:2211.03179.

How to cite: Cerdà-Bautista, J., Tárraga, J. M., Sitokonstantinou, V., and Camps-Valls, G.: Causal evaluation of humanitarian aid on food security, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-17312, https://doi.org/10.5194/egusphere-egu24-17312, 2024.

EGU24-19242 | ECS | Orals | ITS1.11/NP4.2

Multi-model comparison of causal relationships between atmospheric and marine biogeochemical variables 

Germain Bénard, Marion Gehlen, and Mathieu Vrac

Time series of in situ observations and remote sensing data suggest variability in epipelagic ecosystems at seasonal to multiannual time scales. These go along with changes in physical-biogeochemical conditions. While a consensus exists on the proximate causes of observed ecosystem variability (e.g. mixed layer variability, availability of nutrients, grazing pressure), the role of large-scale drivers (e.g. natural climate modes) still needs to be better understood. Moreover, differences in the implementation of marine ecosystem processes exist among Earth System Models, and it is important to understand the uncertainty around the representation of specific interactions via inter-model comparison.

We use output from 5 multi-centennial Earth system model simulations under pre-industrial climate to identify modes of low-frequency biogeochemical properties and the importance of individual drivers. The study focuses on the North Atlantic subpolar gyre (NASPG), a region of high primary productivity and considerable observed natural variability in physical and biogeochemical conditions. We explore causality between modes of climate variability, ocean physics and biogeochemistry by applying a Knowledge-Data-Discovery method, PCMCI. This method enables causal links with a potential time lag to be established between different domains. It proposes a novel way for the comparison of differences between model dynamics.

First, six geographic subregions are identified, based on their physical-biogeochemical characteristics (e.g. deep convection zones, intensity of spring bloom), followed by by the selection of physical and biogeochemical variables. These variables are the maximum winter mixed layer depth due to the role in supplying nutrients to the surface fueling the spring bloom, the North Atlantic Oscillation (NAO), a dominant natural mode climate variability, for its contribution to sea surface temperature (SST) and nutrient variability in the subpolar gyre, and the Gyre Strength, an index reflecting the response of the NASPG to wind forcing. We focus on one micronutrient (Iron) and one macronutrient (Nitrate). They were chosen because both can limit the primary production in this region. 

Next, PCMCI is applied to search for the temporal relationships (potentially lagged) between different regions and variables. These relationships are computed from partial correlations which, for gaussian distributed data, is equivalent to a causal link. The application of this method allows networks of causality to be identified, highlighting drivers of nutrient variability under varying natural climate forcing. The approach enables the quantification of intermodel differences either by analyzing one link after another or by looking directly at the entire causal graphs with a newly proposed method to quantify the dissimilarity between two models.

This method verified expected interactions such as the role of mixed layer depth for nutrient supply and quantified the strength of this interaction across the models. It also highlighted model-specific dynamics such as the role of temperature (via sea-ice formation) for iron in two biogeochemical models out of 5. 



 

How to cite: Bénard, G., Gehlen, M., and Vrac, M.: Multi-model comparison of causal relationships between atmospheric and marine biogeochemical variables, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-19242, https://doi.org/10.5194/egusphere-egu24-19242, 2024.

EGU24-20089 | ECS | Posters on site | ITS1.11/NP4.2

Causal Discovery of Stochastic Dynamical Systems: A Markov Chain Approach 

Marcell Stippinger, Attila Bencze, Ádám Zlatniczki, Zoltán Somogyvári, and András Telcs

Exploring causal relationships among stochastic dynamic systems based solely on observed time series of their states poses a challenging problem. In this context, we present a novel method for causal discovery within stochastic dynamic systems, specifically designed to overcome the limitations of existing methods, particularly in detecting hidden and common drivers. Our proposed approach is based on a straightforward observation: a process generated by a stochastic dynamical system follows a Markov chain if and only if all external influences are independent and identically distributed (i.i.d.). Consequently, the primary tool in our proposed causal discovery scheme involves testing whether the process generates a Markov chain, as opposed to relying on the "classical" causal Markov property or d-separation.

Our method is nonparametric, requiring no intervention, and is built on a reasonably small number of assumptions. We tested our model both on simulated Markov chains of finite state space and structural vector autoregressive processes. To demonstrate the efficacy of our model, we apply it to weather data consisting of solar irradiation and daily average air temperature. Through our method, we successfully identify the ground truth, revealing that irradiation drives temperature. Furthermore, we adeptly pinpoint the true lag while eliminating spurious lags in the autocorrelation function.

How to cite: Stippinger, M., Bencze, A., Zlatniczki, Á., Somogyvári, Z., and Telcs, A.: Causal Discovery of Stochastic Dynamical Systems: A Markov Chain Approach, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-20089, https://doi.org/10.5194/egusphere-egu24-20089, 2024.

EGU24-20353 | Orals | ITS1.11/NP4.2

A Reply to “On Spurious Causality, CO2, and Global Temperature” 

Adolf Stips, San Liang, and Diego Macias-Moy

Stips et al (2016) demonstrated the existing causal relationship between Green House Gases (GHG) concentrations and Global Mean Surface Temperature (GMTA) based on the Information Flow (IF) methodology. Critics on the application of the Information Flow concept as developed by Liang (2008, 2016) has focused on the underlying assumption of uncorrelated residuals (noise) between the time series. However, this assumption can only make sense for a system with two components, as for a multi-dimensional system unobserved noise may well exist. Fundamentally, there can be no such thing like correlated noise at all. It can seemingly only appear because of some hidden process(es). For investigating this in detail a multivariate information flow analysis has been developed. We will show that in our tests using processes with correlated noises, the preset causalities can be well reproduced. Further, it will be demonstrated that reducing autocorrelation within the time series by pre-whitening, confirms the achieved causality directions. Finally, we question the validity of the proposed alternative measure using forecast error variance decomposition based on vector autoregression by Goulet and Goebel (2021), because in their method causal directions can be simply reversed by reordering.  A physically faithful causal measure should be generally independent of ordering.

 

Coulombe, P. G. and Goebel, M. 2021. On Spurious Causality, CO2, and Global Temperature.  Econometrics9(3), 33.

Liang, X. S. 2008. Information Flow within Stochastic Dynamical System. Phys. Rev. E 78: 031113.

Liang, X. S. 2016. Information Flow and Causality as rigorous Notions ab initio. Physical Review E 94: 05220.

Stips, A., D. Macias, C. Coughlan, E. Garcia-Gorriz, and X. S. Liang. 2016. On the Causal Structure between CO2 and Global Temperature. Scientific Reports 6: 21691.

How to cite: Stips, A., Liang, S., and Macias-Moy, D.: A Reply to “On Spurious Causality, CO2, and Global Temperature”, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-20353, https://doi.org/10.5194/egusphere-egu24-20353, 2024.

EGU24-20548 | Orals | ITS1.11/NP4.2 | Highlight

Some Thoughts on Causal Inference, the Scientific Method, and Data Assimilation 

Michael Ghil, Alberto Carrassi, and Olivier de Viron

Causal inference is at the heart of the scientific method as usually practiced. Still, Karl Popper (The Logic of Scientific Discovery, 1935/1959)  tells us that a theory in the empirical sciences can never be proven: it can only be falsified, meaning that it can, and should, be scrutinized with decisive experiments. Even so, nobody that I know writes or publishes papers to disprove one’s own theory, only an opposing theory. And the debate rages on.

At the heart of this session lies the question of whether, and how, one can prove, rather than just disprove, a causal link between phenomena in the empirical sciences. The session deals specifically with statistical, as opposed to dynamical methods. These methods have the advantage that they are essentially indifferent to any laws of, or other accumulated heuristic ideas on, the field to which they are being applied: whether the time series one considers are from the environmental sciences, biology or medicine does not matter, only their length and accuracy does.

Judea Pearl (e.g., Stat. Surveys, 2009) made an important observation on how to transcend the saying that “Correlation is not causation” by pointing out that standard methods of statistical analysis rely on the stationarity hypothesis of the phenomena being examined. Crucial questions, however, like the causal role of anthropogenic forcing in climate change, deal precisely with the causes of nonstationarity. In particular, Pearl suggested counterfactual analysis as an essential approach in establishing criteria for the necessary and sufficient character of a given cause for a given phenomenon. Thus, the common approach of detection and attribution in the climate sciences only covers the sufficiency aspect of anthropogenic forcing, and more can be done (Hannart et al., BAMS, 2016; Clim. Change, 2016).

The present talk will cover four specific aspects of these broad issues: (i) the distinction between information transfer, including both linear correlations and nonlinear extensions thereof, and true causation; (ii) the divergent results of some widely, and not so widely, used methods of studying information transfer (Krakovska et al., PRE, 2018; Kossakowski et al., Psychol. Methods, 2021; Delforge et al., HESS, 2022); (iii) shared variability of climatic time series (De Viron, GRL, 2013; ); and (iv) the uses of data assimilation in applying counterfactual theory to nonstationary phenomena (Carrassi, QJRMS, 2017; Metref et al., QJRMS, 2019).

Conclusions will include the obvious one that statistical studies of causal inference have to be complemented by dynamical ones.

How to cite: Ghil, M., Carrassi, A., and de Viron, O.: Some Thoughts on Causal Inference, the Scientific Method, and Data Assimilation, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-20548, https://doi.org/10.5194/egusphere-egu24-20548, 2024.

EGU24-21883 | Orals | ITS1.11/NP4.2 | Highlight

Large Language Models for Causal Discovery in the Earth Sciences 

Gustau Camps-Valls, Kai-Hendrik Cohrs, Emiliano Diaz, Vasileios Sitokonstantinou, and Gherardo Varando

Causality is essential for understanding complex systems like the Earth and climate, where a plethora of intertwined variables and processes happen in the wild. Constructing causal graphs often relies on either data-driven or expert-driven approaches, both fraught with challenges. The former methods, like the celebrated Peter-Clark (PC) algorithm, face issues with data requirements and assumptions of causal sufficiency, while the latter demand substantial time and expertise.

This work explores the capabilities of Large Language Models (LLMs) as an alternative to domain experts for causal graph generation. We frame conditional independence queries as prompts to LLMs and employ the PC algorithm with the answers. The performances of the LLM-based conditional independence oracle on systems with known causal graphs show a high degree of variability. We improve the performance through a proposed statistical-inspired voting schema that allows control over false-positives and false-negatives rates. We apply our chatPC algorithm to understand the causal relations between complex sets of variables (social, economic, conflicts, environmental, and climatic factors) in two pressing problems: population displacement and food insecurity in Africa. We find plausible graphs as corroborated by experts in the humanitarian sector, finding traces of causal reasoning in the model's answers. We posit that LLM-based causality is a new, promising, alternative avenue for automated causality, especially indicated for rapid response and data-scarce regimes.

How to cite: Camps-Valls, G., Cohrs, K.-H., Diaz, E., Sitokonstantinou, V., and Varando, G.: Large Language Models for Causal Discovery in the Earth Sciences, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-21883, https://doi.org/10.5194/egusphere-egu24-21883, 2024.

EGU24-22158 | ECS | Posters on site | ITS1.11/NP4.2

Spatiotemporal Causal Effect Estimation 

Rebecca Herman and Jakob Runge

Causal discovery and effect estimation for time series provide scientists with a way to extract causal information from observational studies when possible. But the high dimensionality of raw climate data causes computational problems for most analysis methods, and causal inference is no exception. To address this problem, climate scientists usually pre-process climate data using dimension reduction techniques (including seasonal and regional averaging and principle component analysis) that may result in the loss of valuable information before the true analysis even begins. For example, climate scientists often represent El Niño Southern Oscillation variability (ENSO) using the uni-variate Nino3.4 index, which cannot distinguish between central Pacific and eastern Pacific El Niño events, which are believed to impact global climate varaibility in different ways. This study introduces a method for avoiding premature data dimension reduction in causal effect estimation, implemented in tigramite. The method allows the researcher to define multi-variate climate indices, reducing the dimensionality of the causal inference problem via the causal assumptions instead of losing information from the data itself. To investigate the performance of this approach on climate data, we examine the effect of ENSO on the North Atlantic Oscillation (NAO) in simulated data from the Coupled Model Intercomparison Project, phase 6. We choose this as our case study because different types of El Nino are believed to have very different effects on NAO, to the extent that the impact may be completely undetectable in observations when no distinction between the types of ENSO is made. By comparing estimated effects using uni- and multi-variate climate indices, we demonstrate that this method retains valuable information that would be lost in uni-variate analysis, and make recommendations for best practices when using multi-variate climate indices in causal effect estimation.

How to cite: Herman, R. and Runge, J.: Spatiotemporal Causal Effect Estimation, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-22158, https://doi.org/10.5194/egusphere-egu24-22158, 2024.

EGU24-262 | Orals | HS3.5

Differentiable modeling for global water resources under global change 

Chaopeng Shen, Yalan Song, Farshid Rahmani, Tadd Bindas, Doaa Aboelyazeed, Kamlesh Sawadekar, Martyn Clark, and Wouter Knoben

Process-based modeling offers interpretability and physical consistency in many domains of geosciences but struggles to leverage large datasets efficiently. Machine-learning methods, especially deep networks, have strong predictive skills yet are unable to answer specific scientific questions. A recently proposed genre of physics-informed machine learning, called “differentiable” modeling (DM, https://t.co/qyuAzYPA6Y), trains neural networks (NNs) with process-based equations (priors) together in one stage (so-called “end-to-end”) to benefit from the best of both NNs and process-based paradigms. The NNs do not need target variables for training but can be indirectly supervised by observations matching the outputs of the combined model, and differentiability critically supports learning from big data. We propose that differentiable models are especially suitable as global hydrologic models because they can harvest information from big earth observations to produce state-of-the-art predictions (https://mhpi.github.io/benchmarks/), enable physical interpretation naturally, extrapolate well (due to physical constraints) in space and time, enforce known physical laws and sensitivities, and leverage progress in modern AI computing architecture and infrastructure. Differentiable models can also synergize with existing global hydrologic models (GHMs) and learn from the lessons of the community. Differentiable GHMs to answer pressing societal questions on water resources availability, climate change impact assessment, water management, and disaster risk mitigation, among others. We demonstrate the power of differentiable modeling using computational examples in rainfall-runoff modeling, river routing, forcing fusion, as well applications in water-related domains such as ecosystem modeling and water quality modeling. We discuss how to address potential challenges such as implementing gradient tracking for implicit numerical schemes and addressing process tradeoffs. Furthermore, we show how differentiable modeling can enable us to ask fundamental questions in hydrologic sciences and get robust answers from big global data.

How to cite: Shen, C., Song, Y., Rahmani, F., Bindas, T., Aboelyazeed, D., Sawadekar, K., Clark, M., and Knoben, W.: Differentiable modeling for global water resources under global change, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-262, https://doi.org/10.5194/egusphere-egu24-262, 2024.

Streamflow can be affected by numerous factors, such as solar radiation, underlying surface conditions, and atmospheric circulation which results in nonlinearity, uncertainty, and randomness in streamflow time series. Diverse conventional and Deep Learning (DL) models have been applied to recognize the complex patterns and discover nonlinear relationships in the hydrological time series and incorporating multi-variables in deep learning can match or improve streamflow forecasts and hopes to improve extreme value predictions. Multivariate approaches surpass univariate ones by including additional time series as explanatory variables. Deep neural networks (DNNs) excel in multi-horizon time series forecasting, outperforming classical models. However, determining the relative contribution of each variable in streamflow remains challenging due to the black-box nature of DL models.

 

We propose utilizing the advanced Temporal Fusion Transformers (TFT) deep-learning technique to model streamflow values across various temporal scales, incorporating multiple variables. TFT's attention-based architecture enables high-performance multi-horizon forecasting with interpretable insights into temporal dynamics. Additionally, the model identifies the significance of each input variable, recognizes persistent temporal patterns, and highlights extreme events. Despite its application in a few studies across different domains, the full potential of this model remains largely unexplored. The study focused on Sundargarh, an upper catchment of the Mahanadi basin in India, aiming to capture pristine flow conditions. QGIS was employed to delineate the catchment, and daily streamflow data from 1982 to 2020 were obtained from the Central Water Commission. Input variables included precipitation, potential evaporation, temperature, and soil water volume at different depths. Precipitation and temperature datasets were obtained from India Meteorological Department (IMD) datasets, while other variables were sourced from the ECMWF fifth-generation reanalysis (ERA-5). Hyperparameter tuning was conducted using the Optuna optimization framework, known for its efficiency and easy parallelization. The model trained using quantile loss function with different combinations of quantiles, demonstrated superior performance with upper quantiles. Evaluations using R2 and NSE indicated good performance in monthly streamflow predictions for testing sets, particularly in confidently predicting low and medium flows. While peak flows were well predicted at certain timesteps, there were instances of underperformance. Unlike other ML algorithms, TFT can learn seasonality and lag analysis patterns directly from raw training data, including the identification of crucial variables. The model underwent training for different time periods, checking for performance improvement with increased length of data. To gain a better understanding of how distinct sub-processes affect streamflow patterns at various time scales, the model was applied at pentad and daily scales. Evaluation at extreme values prompted an investigation into improving predictions through quantile loss function adjustments. Given the computational expense of daily streamflow forecasting using TFT with multiple variables, parallel computing is employed. Results demonstrated considerable accuracy, but validating TFT's interpretive abilities require testing alternative ML models.

 

How to cite: Mohan, M. and Kumar D, N.: Multivariate multi-horizon streamflow forecasting for extremes and their interpretation using an explainable deep learning architecture, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-451, https://doi.org/10.5194/egusphere-egu24-451, 2024.

EGU24-2211 | ECS | Posters on site | HS3.5

Staged Learning in Physics-Informed Neural Networks to Model Contaminant Transport under Parametric Uncertainty 

Milad Panahi, Giovanni Porta, Monica Riva, and Alberto Guadagnini

Addressing the complexities of groundwater modeling, especially under the veil of uncertain physical parameters and limited observational data, poses significant challenges. This study introduces an approach using Physics-Informed Neural Network (PINN) framework to unravel these uncertainties. Termed PINN under uncertainty, PINN-UU, adeptly integrates uncertain parameters within spatio-temporal domains, focusing on hydrological systems. This approach, exclusively built on underlying physical equations, leverages a staged training methodology, effectively navigating high-dimensional solution spaces. We demonstrate our approach through application of reactive transport modeling in porous media, a problem setting relevant to contaminant transport in soil and groundwater. PINN-UU shows promising capabilities in enhancing model reliability and efficiency, and in conducting sensitivity analysis. Our approach is designed to be accessible and engaging, offering insightful contributions to environmental engineering, and hydrological modeling. It represents a step toward deciphering complex geohydrological systems, with broad implications for resource management and environmental science.

How to cite: Panahi, M., Porta, G., Riva, M., and Guadagnini, A.: Staged Learning in Physics-Informed Neural Networks to Model Contaminant Transport under Parametric Uncertainty, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-2211, https://doi.org/10.5194/egusphere-egu24-2211, 2024.

EGU24-2850 | ECS | Orals | HS3.5

Development of a Distributed Physics-informed Deep Learning Hydrological Model for Data-scarce Regions 

Liangjin Zhong, Huimin Lei, and JIngjing Yang

Climate change has exacerbated water stress and water-related disasters, necessitating more precise runoff simulations. However, in the majority of global regions, a deficiency of runoff data constitutes a significant constraint on modeling endeavors. Traditional distributed hydrological models and regionalization approaches have shown suboptimal performance. While current data-driven models trained on large datasets excel in spatial extrapolation, the direct applicability of these models in certain regions with unique hydrological processes may be challenging due to the limited representativeness within the training dataset. Furthermore, transfer learning deep learning models pre-trained on large datasets still necessitate local data for retraining, thereby constraining their applicability. To address these challenges, we present a physics-informed deep learning model based on a distributed framework. It involves spatial discretization and the establishment of differentiable hydrological models for discrete sub-basins, coupled with a differentiable Muskingum method for channel routing. By introducing upstream-downstream relationships, model errors in sub-basins propagate through the river network to the watershed outlet, enabling the optimization using limited downstream runoff data, thereby achieving spatial simulation of ungauged internal sub-basins. The model, when trained solely on the downstream-most station, outperforms the distributed hydrological model in runoff simulation at both the training station and upstream stations, as well as evapotranspiration spatial patterns. Compared to transfer learning, our model requires less training data, yet achieves higher precision in simulating runoff on spatially hold-out stations and provides more accurate estimates of spatial evapotranspiration. Consequently, this model offers a novel approach to hydrological simulation in data-scarce regions with unique processes.

How to cite: Zhong, L., Lei, H., and Yang, J.: Development of a Distributed Physics-informed Deep Learning Hydrological Model for Data-scarce Regions, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-2850, https://doi.org/10.5194/egusphere-egu24-2850, 2024.

EGU24-3028 | Orals | HS3.5 | Highlight

Spatial sensitivity of river flooding to changes in climate and land cover through explainable AI 

Louise Slater, Gemma Coxon, Manuela Brunner, Hilary McMillan, Le Yu, Yanchen Zheng, Abdou Khouakhi, Simon Moulds, and Wouter Berghuijs

Explaining the spatially variable impacts of flood-generating mechanisms is a longstanding challenge in hydrology, with increasing and decreasing temporal flood trends often found in close regional proximity. Here, we develop a machine learning-informed approach to unravel the drivers of seasonal flood magnitude and explain the spatial variability of their effects in a temperate climate. We employ 11 observed meteorological and land cover time series variables alongside 8 static catchment attributes to model flood magnitude in 1268 catchments across Great Britain over four decades. We then perform a sensitivity analysis to understand how +10% precipitation, +1°C air temperature, or +10 percentage points of urbanisation or afforestation affect flood magnitude in catchments with varying characteristics. Our simulations show that increasing precipitation and urbanisation both tend to amplify flood magnitude significantly more in catchments with high baseflow contribution and low runoff ratio, which tend to have lower values of specific discharge on average. In contrast, rising air temperature (in the absence of changing precipitation) decreases flood magnitudes, with the largest effects in dry catchments with low baseflow index. Afforestation also tends to decrease floods more in catchments with low groundwater contribution, and in dry catchments in the summer. These reported associations are significant at p<0.001. Our approach may be used to further disentangle the joint effects of multiple flood drivers in individual catchments.

How to cite: Slater, L., Coxon, G., Brunner, M., McMillan, H., Yu, L., Zheng, Y., Khouakhi, A., Moulds, S., and Berghuijs, W.: Spatial sensitivity of river flooding to changes in climate and land cover through explainable AI, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-3028, https://doi.org/10.5194/egusphere-egu24-3028, 2024.

EGU24-4105 | ECS | Orals | HS3.5

Global flood projection and socioeconomic implications under a physics-constrained deep learning framework 

Shengyu Kang, Jiabo Yin, Louise Slater, Pan Liu, and Dedi Liu

As the planet warms, the frequency and severity of weather-related hazards such as floods are intensifying, posing substantial threats to communities around the globe. Rising flood peaks and volumes can claim lives, damage infrastructure, and compromise access to essential services. However, the physical mechanisms behind global flood evolution are still uncertain, and their implications for socioeconomic systems remain unclear. In this study, we leverage a supervised machine learning technique to identify the dominant factors influencing daily streamflow. We then propose a physics-constrained cascade model chain which assimilates water and heat transport processes to project bivariate risk (i.e. flood peak and volume together), along with its socioeconomic consequences. To achieve this, we drive a hybrid deep learning-hydrological model with bias-corrected outputs from twenty global climate models (GCMs) under four shared socioeconomic pathways (SSPs). Our results project considerable increases in flood risk under the medium to high-end emission scenario (SSP3-7.0) over most catchments of the globe. The median future joint return period decreases from 50 years to around 27.6 years, with 186 trillion dollars and 4 billion people exposed. Downwelling shortwave radiation is identified as the dominant factor driving changes in daily streamflow, accelerating both terrestrial evapotranspiration and snowmelt. As future scenarios project enhanced radiation levels along with an increase in precipitation extremes, a heightened risk of widespread flooding is foreseen. This study aims to provide valuable insights for policymakers developing strategies to mitigate the risks associated with river flooding under climate change.

How to cite: Kang, S., Yin, J., Slater, L., Liu, P., and Liu, D.: Global flood projection and socioeconomic implications under a physics-constrained deep learning framework, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-4105, https://doi.org/10.5194/egusphere-egu24-4105, 2024.

EGU24-4238 | ECS | Posters on site | HS3.5

Letting neural networks talk: exploring two probabilistic neural network models for input variable selection 

John Quilty and Mohammad Sina Jahangir

Input variable selection (IVS) is an integral part of building data-driven models for hydrological applications. Carefully chosen input variables enable data-driven models to discern relevant patterns and relationships within data, improving their predictive accuracy. Moreover, the optimal choice of input variables can enhance the computational efficiency of data-driven models, reduce overfitting, and contribute to a more interpretable and parsimonious model. Meanwhile, including irrelevant and/or redundant input variables can introduce noise to the model and hinder its generalization ability.

Three probabilistic IVS methods, namely Edgeworth approximation-based conditional mutual information (EA), double-layer extreme learning machine (DLELM), and gradient mapping (GM), were used for IVS and then coupled with a long short-term memory (LSTM)-based probabilistic deep learning model for daily streamflow prediction. While the EA method is an effective IVS method, DLELM and GM are examples of probabilistic neural network-based IVS methods that have not yet been explored for hydrological prediction. DLELM selects input variables through sparse Bayesian learning, pruning both input and output layer weights of a committee of neural networks. GM is based on saliency mapping, an explainable AI technique commonly used in computer vision that can be coupled with probabilistic neural networks. Both DLELM and GM involve randomization during parameter initialization and/or training thereby introducing stochasticity into the IVS procedure, which has been shown to improve the predictive performance of data-driven models.

The IVS methods were coupled with a LSTM-based probabilistic deep learning model and applied to a streamflow prediction case study using 420 basins spread across the continental United States. The dataset includes 37 candidate input variables derived from the daily-averaged ERA-5 reanalysis data.

Comparing the most frequently selected input variables by EA, DLELM, and GM across the 420 basins revealed that all three models select a similar number of input variables. For example, the top 15 input variables selected by all methods included nine variables that were similar.

The input variables selected by EA, DLELM, and GM were then used in the LSTM-based probabilistic deep learning models for streamflow prediction across the 420 basins. The probabilistic deep learning models were developed and optimized using the top 10 variables selected by each IVS method. The results were compared to a benchmark scenario that used all 37 ERA-5 variables in the prediction model. Overall, the findings show that the GM method results in higher prediction accuracy (Kling-Gupta efficiency; KGE) compared to the other two IVS methods. A median KGE of 0.63 was obtained for GM, whereas for the EA, DLELM, and all input variables’ scenario, KGE scores of 0.61, 0.60, and 0.62 were obtained, respectively.

DLELM and GM are two AI-based techniques that introduce elements of interpretability and stochasticity to the IVS process. The results of the current study are expected to contribute to the evolving landscape of data-driven hydrological modeling by introducing hitherto unexplored neural network-based IVS to pursue more parsimonious, efficient, and interpretable probabilistic deep learning models.

How to cite: Quilty, J. and Jahangir, M. S.: Letting neural networks talk: exploring two probabilistic neural network models for input variable selection, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-4238, https://doi.org/10.5194/egusphere-egu24-4238, 2024.

EGU24-4325 | ECS | Posters on site | HS3.5

Towards learning human influences in a highly regulated basin using a hybrid DL-process based framework 

Liangkun Deng, Xiang Zhang, and Louise Slater

Hybrid models have shown impressive performance for streamflow simulation, offering better accuracy than process-based hydrological models (PBMs) and superior interpretability than deep learning models (DLMs). A recent paradigm for streamflow modeling, integrating DLMs and PBMs within a differentiable framework, presents considerable potential to match the performance of DLMs while simultaneously generating untrained variables that describe the entire water cycle. However, the potential of this framework has mostly been verified in small and unregulated headwater basins and has not been explored in large and highly regulated basins. Human activities, such as reservoir operations and water transfer projects, have greatly changed natural hydrological regimes. Given the limited access to operational water management records, PBMs generally fail to achieve satisfactory performance and DLMs are challenging to train directly. This study proposes a coupled hybrid framework to address these problems. This framework is based on a distributed PBM, the Xin'anjiang (XAJ) model, and adopts embedded deep learning neural networks to learn the physical parameters and replace the modules of the XAJ model reflecting human influences through a differentiable structure. Streamflow observations alone are used as training targets, eliminating the need for operational records to supervise the training process. The Hanjiang River basin (HRB), one of the largest subbasins of the Yangtze River basin, disturbed by large reservoirs and national water transfer projects, is selected to test the effectiveness of the framework. The results show that the hybrid framework can learn the best parameter sets of the XAJ model depicting natural and human influences to improve streamflow simulation. It performs better than a standalone XAJ model and achieves similar performance to a standalone LSTM model. This framework sheds new light on assimilating human influences to improve simulation performance in disturbed river basins with limited operational records.

How to cite: Deng, L., Zhang, X., and Slater, L.: Towards learning human influences in a highly regulated basin using a hybrid DL-process based framework, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-4325, https://doi.org/10.5194/egusphere-egu24-4325, 2024.

EGU24-4768 | ECS | Orals | HS3.5

HydroPML: Towards Unified Scientific Paradigms for Machine Learning and Process-based Hydrology 

Qingsong Xu, Yilei Shi, Jonathan Bamber, Ye Tuo, Ralf Ludwig, and Xiao Xiang Zhu

Accurate hydrological understanding and water cycle prediction are crucial for addressing scientific and societal challenges associated with the management of water resources, particularly under the dynamic influence of anthropogenic climate change. Existing work predominantly concentrates on the development of machine learning (ML) in this field, yet there is a clear distinction between hydrology and ML as separate paradigms. Here, we introduce physics-aware ML as a transformative approach to overcome the perceived barrier and revolutionize both fields. Specifically, we present a comprehensive review of the physics-aware ML methods, building a structured community (PaML) of existing methodologies that integrate prior physical knowledge or physics-based modeling into ML. We systematically analyze these PaML methodologies with respect to four aspects: physical data-guided ML, physics-informed ML, physics-embedded ML, and physics-aware hybrid learning. PaML facilitates ML-aided hypotheses, accelerating insights from big data and fostering scientific discoveries. We initiate a systematic exploration of hydrology in PaML, including rainfall-runoff and hydrodynamic processes, and highlight the most promising and challenging directions for different objectives and PaML methods. Finally, a new PaML-based hydrology platform, termed HydroPML, is released as a foundation for applications based on hydrological processes [1]. HydroPML presents a range of hydrology applications, including but not limited to rainfall-runoff-inundation modeling, real-time flood forecasting (FloodCast), rainfall-induced landslide forecasting (LandslideCast), and cutting-edge PaML methods, to enhance the explainability and causality of ML and lay the groundwork for the digital water cycle's realization. The HydroPML platform is publicly available at https://hydropml.github.io/.

[1] Xu, Qingsong, et al. "Physics-aware Machine Learning Revolutionizes Scientific Paradigm for Machine Learning and Process-based Hydrology." arXiv preprint arXiv:2310.05227 (2023).

How to cite: Xu, Q., Shi, Y., Bamber, J., Tuo, Y., Ludwig, R., and Zhu, X. X.: HydroPML: Towards Unified Scientific Paradigms for Machine Learning and Process-based Hydrology, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-4768, https://doi.org/10.5194/egusphere-egu24-4768, 2024.

EGU24-6378 | ECS | Posters on site | HS3.5

Seasonal forecasts of hydrological droughts over the Alps: advancing hybrid modelling applications 

Iacopo F. Ferrario, Mariapina Castelli, Alasawedah M. Hussein, Usman M. Liaqat, Albrecht Weerts, and Alexander Jacob

The Alpine region is often called the Water Tower of Europe, alluding to its water richness and its function of supplying water through several important European rivers flowing well beyond its geographical boundaries. Climate change projections show that the region will likely experience rising temperatures and changes in precipitation type, frequency, and intensity, with consequences on the spatiotemporal pattern of water availability. Seasonal forecasts could supply timely information for planning water allocation a few months in advance, reducing potential conflicts under conditions of scarce water resources. The overall goal of this study is to improve the seasonal forecasts of hydrological droughts over the entire Alpine region at a spatial resolution (~1 km) that matches the information need by local water agencies, e.g., resolving headwaters and small valleys. In this study we present the progress on the following key objectives:

  • Improving the estimation of distributed model (Wflow_sbm) parameters by finding the optimal transfer function from geophysical attributes to model parameters and upscaling the information to model resolution.
  • Combining physical-hydrological knowledge with data-driven (ML/DL) techniques for improving accuracy and computational performance, without compromising on interpretation
  • Integrating EO-based hydrological fluxes, like streamflow, surface soil moisture, actual evapotranspiration, and snow waters equivalent, with the aim of regularizing the calibration/training, tackling the problem of model parameters equifinality.

Our work is part of the InterTwin project that aims at developing a multi-domain Digital Twin blueprint architecture and implementation platform. We build on the technological solutions developed in InterTwin (e.g. openEO, CWL and STAC) and fully embrace its inspiring principles of open science, reproducibility, and interoperability of data and methods.

How to cite: Ferrario, I. F., Castelli, M., Hussein, A. M., Liaqat, U. M., Weerts, A., and Jacob, A.: Seasonal forecasts of hydrological droughts over the Alps: advancing hybrid modelling applications, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-6378, https://doi.org/10.5194/egusphere-egu24-6378, 2024.

EGU24-6656 | ECS | Orals | HS3.5

Exploring Catchment Regionalization through the Eyes of HydroLSTM 

Luis De La Fuente, Hoshin Gupta, and Laura Condon

Regionalization is an issue that hydrologists have been working on for decades. It is used, for example, when we transfer parameters from one calibrated model to another, or when we identify similarities between gauged to ungauged catchments. However, there is still no unified method that can successfully transfer parameters and identify similarities between different regions while accounting for differences in meteorological forcing, catchment attributes, and hydrological responses.

Machine learning (ML) has shown promising results in the generalization of its results at temporal and spatial scales for streamflow prediction. This suggests that ML models have learned useful regionalization relationships that we could extract. This study explores how the HydroLSTM representation, a modification of traditional Long Short-Term Memory, can learn meaningful relationships between meteorological forcing and catchment attributes. One promising feature of the HydroLSTM representation is that the learned patterns can generate different hydrological responses across the US. These findings indicate that we can learn more about regionalization by studying ML models.

How to cite: De La Fuente, L., Gupta, H., and Condon, L.: Exploring Catchment Regionalization through the Eyes of HydroLSTM, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-6656, https://doi.org/10.5194/egusphere-egu24-6656, 2024.

EGU24-6965 | ECS | Posters on site | HS3.5

A Machine Learning Based Snow Cover Parameterization  for Common Land Model (CoLM)  

Han Zhang, Lu Li, and Yongjiu Dai

Accurate representation of snow cover fraction (SCF) is vital for terrestrial simulation, as it significantly affects surface albedo and land surface radiation. In land models, SCF is parameterized using snow water equivalent and snow depth. This study introduces a novel machine learning-based parameterization, which incorporates the light-GBM regression algorithm and additional input features: surface air temperature, humidity, leaf area index, and the standard deviation of topography. The regression model is trained with input features from the Common Land Model (CoLM) simulations and the labels from the Moderate Resolution Imaging Spectroradiometer (MODIS) observations on a daily scale. Offline verification indicates significant improvements for the new scheme over multiple traditional parameterizations.

Moreover, this machine learning-based parameterization has been online coupled with the CoLM using the Message Passing Interface (MPI). In online simulations, it substantially outperforms the widely used Niu and Yang (2007) scheme, improving the root mean square errors and temporal correlations of SCF on 80% of global grids. Additionally, associated land surface temperature and hydrological processes also benefit from the enhanced estimation of SCF. The new solution also shows good portability as it also demonstrates similar enhancements when it is directly used in a global 1° simulation, even though it was trained at a 0.1° resolution.

How to cite: Zhang, H., Li, L., and Dai, Y.: A Machine Learning Based Snow Cover Parameterization  for Common Land Model (CoLM) , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-6965, https://doi.org/10.5194/egusphere-egu24-6965, 2024.

Land-atmosphere coupling (LAC) involves a variety of interactions between the land surface and the atmospheric boundary layer that are critical to are critical to understanding hydrological partitioning and cycling. As climate change continues to affect these interactions, identifying the specific drivers of LAC variability has become increasingly important. However, due to the complexity of the coupling mechanism, a quantitative understanding of the potential drivers is still lacking. Recently, deep learning has been considered as an effective approach to capture nonlinear relationships within the data, which provides a useful window into complex climatic processes. In this study, we will explore the LAC variability under climate change and its potential drivers by using Convolutional Long Short-term Memory (ConvLSTM) together with explainable AI techniques for attribution analysis. Specifically, the variability of the LAC, defined here as a two-legged index, is used as the modeling target, and variables representing meteorological forcing, land use, irrigation, soil properties, gross primary production, ecosystem respiration, and net ecosystem exchange are the inputs. Our analysis covers global land with a spatial resolution of 0.1° × 0.1° every one day during the period 1979–2019. Overall, the study demonstrates how interpretable machine learning would help us understand the complex dynamics of LAC under changing climatic conditions. We expect the results to facilitate the understanding of terrestrial hydroclimate interactions and hopefully provide multiple lines of evidence to support future water management.

How to cite: Huang, F., Shangguan, W., and Jiang, S.: Identifying potential drivers of land-atmosphere coupling variation under climate change by explainable artificial intelligence, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-7202, https://doi.org/10.5194/egusphere-egu24-7202, 2024.

EGU24-7950 | ECS | Posters on site | HS3.5

Improving streamflow prediction across China by hydrological modelling together with machine learning 

wang jiao and zhang yongqiang

Predicting streamflow is key for water resource planning, flood and drought risk assessment, and pollution mitigation at regional, national, and global scales. There is a long-standing history of developing physically or conceptually catchment rainfall-runoff models that have been continuously refined over time to include more physical processes and enhance their spatial resolution. On the other hand, machine learning methods, particularly neural networks, have demonstrated exceptional accuracy and extrapolation capabilities in time-series prediction. Both approaches exhibit their strengths and limitations. This leads to a research question: how to effectively balance model complexity and physical interpretability while maintaining a certain level of predictive accuracy. This study aims to effectively combine a conceptual hydrological model, HBV, with machine learning (Transformer, Long Short-Term Memory (LSTM)) using a differentiable modeling framework strategy, tailored to predicting streamflow under diverse climatic and geographical conditions across China. Utilizing the Transformer to optimize and replace certain parameterization processes in the HBV model, a deep integration of neural networks and the HBV model is achieved. This integration not only captures the non-linear relationships that traditional hydrological models struggle to express, but also maintains the physical interpretability of the model. Preliminary application results show that the proposed framework outperforms traditional HBV model and pure LSTM model in streamflow prediction across 68 catchments in China. Based on the test results from different catchments, we have adjusted and optimized the model structure or parameters to better adapt to the unique hydrological processes of each catchment. The application of self-attention mechanisms and a differentiable programming framework significantly enhances the model's ability to capture spatiotemporal dynamics. It is likely that the proposed framework can be widely used for streamflow prediction somewhere else.

How to cite: jiao, W. and yongqiang, Z.: Improving streamflow prediction across China by hydrological modelling together with machine learning, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-7950, https://doi.org/10.5194/egusphere-egu24-7950, 2024.

EGU24-9319 | ECS | Posters on site | HS3.5

Developing hybrid distributed models for hydrological simulation and climate change assessment in large alpine basins 

Bu Li, Ting Sun, Fuqiang Tian, and Guangheng Ni

Large alpine basins on the Tibetan Plateau (TP) provide abundant water resources crucial for hydropower generation, irrigation, and daily life. In recent decades, the TP has been significantly affected by climate change, making it crucial to understand the runoff response to climate change are essential for water resources management. While limited knowledge of specific alpine hydrological processes has constrained the accuracy of hydrological models and heightened uncertainties in climate change assessments. Recently, hybrid hydrological models have come to the forefront, synergizing the exceptional learning capacity of deep learning with a rigorous adherence to hydrological knowledge of process-based models. These models exhibit considerable promise in achieving precision in hydrological simulations and conducting climate change assessments. However, a notable limitation of existing hybrid models lies in their failure to incorporate spatial information and describe alpine hydrological processes, which restricts their applicability in hydrological modeling and climate change assessment in large alpine basins. To address this issue, we develop a set of hybrid distributed hydrological models by employing a distributed process-based model as the backbone, and utilizing embedded neural networks (ENNs) to parameterize and replace different internal modules. The proposed models are tested on three large alpine basins on the Tibetan Plateau. Results are compared to those obtained from hybrid lumped models, state-of-the-art distributed hydrological model, and DL models. A climate perturbation method is further used to evaluate the alpine basins' runoff response to climate change.Results indicate that proposed hybrid hydrological models can perform well in predicting runoff in large alpine basins. The optimal hybrid model with Nash-Sutcliffe efficiency coefficients (NSEs) higher than 0.87 shows comparable performance to state-of-the-art DL models. The hybrid distributed model also exhibits remarkable capability in simulating hydrological processes at ungauged sites within the basin, markedly surpassing traditional distributed models. Besides, runoff exhibits an amplification effect in response to precipitation changes, with a 10% precipitation change resulting in a 15–20% runoff change in large alpine basins. An increase in temperature enhances evaporation capacity and changes the redistribution of rainfall and snowfall and the timing of snowmelt, leading to a decrease in the total runoff and a reduction in the intra-annual variability of runoff. Overall, this study provides a high-performance tool enriched with explicit hydrological knowledge for hydrological prediction and improves our understanding about runoff’s response to climate change in large alpine basins on the TP. 

How to cite: Li, B., Sun, T., Tian, F., and Ni, G.: Developing hybrid distributed models for hydrological simulation and climate change assessment in large alpine basins, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-9319, https://doi.org/10.5194/egusphere-egu24-9319, 2024.

In facing the challenges of limited observational streamflow data and climate change, accurate streamflow prediction and flood management in large-scale catchments become essential. This study introducing a time-lag informed deep learning framework to enhance streamflow simulation and flood forecasting. Using the Dulong-Irrawaddy River Basin (DIRB), a less-explored transboundary basin shared by Myanmar, China, and India, as a case study, we have identified peak flow lag days and relative flow scale. Integrating these with historical flow data, we developed an optimal model. The framework, informed by data from the upstream Hkamti sub-basin, significantly outperformed standard LSTM, achieving a Kling-Gupta Efficiency (KGE) of 0.891 and a Nash-Sutcliffe efficiency coefficient (NSE) of 0.904. Notably, the H_PFL model provides a valuable 15-day lead time for flood forecasting, enhancing emergency response preparations. The transfer learning model, incorporating meteorological inputs and catchment features, achieved an average NSE of 0.872 for streamflow prediction, surpassing the process-based model MIKE SHE's 0.655. We further analyzed the sensitivities of the deep learning model and process-based model to changes in meteorological inputs using different methods. Deep learning models exhibit complex sensitivities to these inputs, more accurately capturing non-linear relationships among multiple variables than the process-based model. Integrated Gradients (IG) analysis further demonstrates deep learning model's ability to discern spatial heterogeneity in upstream and downstream sub-basins and its adeptness in characterizing different flow regimes. This study underscores the potential of deep learning in enhancing the understanding of hydrological processes in large-scale catchments and highlights its value for water resource management in transboundary basins under data scarcity.

How to cite: Ma, K. and He, D.: Streamflow Prediction and Flood Forecasting with Time-Lag Informed Deep Learning framework in Large Transboundary Catchments, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-9980, https://doi.org/10.5194/egusphere-egu24-9980, 2024.

EGU24-11159 | ECS | Orals | HS3.5

Uncovering the impact of hydrological connectivity on nitrate transport at the catchment scale using explainable AI 

Felipe Saavedra, Noemi Vergopolan, Andreas Musolff, Ralf Merz, Carolin Winter, and Larisa Tarasova

Nitrate contamination of water bodies is a major concern worldwide, as it poses a risk of eutrophication and biodiversity loss. Nitrate travels from agricultural land to streams through different hydrological pathways, which are abstrusely activated under different hydrological conditions. Certainly, hydrological conditions can alter the connection between different parts of the catchment and streams, in many cases independent of the discharge levels, leading to modifications in transport dynamics, retention, and nitrate removal rates in the catchment. While enhanced nitrate transport can be linked to high levels of hydrological connectivity, little is known about the effects of the spatial patterns of hydrological connectivity on the transport of nutrients at the catchment scale.

In this study, we combined daily stream nitrate concentration and discharge data at the outlet of 15 predominantly agricultural catchments in the United States (191–16,000 km2 area, 3500 km2 median area, and 77% median agriculture coverage) with soil moisture data from  SMAP-Hydroblocks (Vergopolan et al., 2021). SMAP-Hydroblocks is a hyperresolution soil moisture dataset at the top 5 cm of soil column at 30-m spatial resolution and 2-3 days revisit time (2015-2019), and it is derived through a combination of satellite data, land-surface and radiative transfer modeling, machine learning, and in-situ observations.

We configured a deep learning model for each catchment, driven by 2D soil moisture fields and 1D discharge time series, to evaluate the impact of streamflow magnitude and spatial patterns of soil moisture on streamflow nitrate concentration. The model setup comprises two parallel branches. The first branch incorporates a Long Short-term Memory (LSTM) model, the current state-of-the-art for time-series data modeling, utilizing daily discharge as input data. The second branch contains a Convolutional LSTM network (ConvLSTM) that incorporates daily soil moisture series, the fraction of agriculture of each pixel, and the height above the nearest drainage as a measurement of structural hydrological connectivity. Finally, a fully connected neural network combines the outputs of the two branches to predict the time series of nitrate concentration at the catchment outlet.

Preliminary results indicate that the model performs satisfactorily in one-third of the catchments, with Nash-Sutcliffe Efficiency (NSE) values above 0.3 for the test period, which covers the final 25% of the time series, and this is achieved without tuning the hyperparameters. The model failed to simulate nitrate concentrations (resulting in negative NSE values) typically in larger catchments. Using these simulations and explainable AI, we will quantify the importance of different inputs, in particular, we tested the relative importance of soil moisture for simulating nitrate concentrations. While the literature shows most of the predictive power for nitrate comes from streamflow rates, we show how soil moisture fields add value to the prediction and understanding of hydrologic connectivity. Finally, we will fine-tune the model for each catchment and include more predictors to enhance the reliability of model simulations.

How to cite: Saavedra, F., Vergopolan, N., Musolff, A., Merz, R., Winter, C., and Tarasova, L.: Uncovering the impact of hydrological connectivity on nitrate transport at the catchment scale using explainable AI, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-11159, https://doi.org/10.5194/egusphere-egu24-11159, 2024.

EGU24-11778 | ECS | Orals | HS3.5

How much data is needed for hydrological modeling?  

Bjarte Beil-Myhre, Bernt Viggo Matheussen, and Rajeev Shrestha

Hydrological modeling has undergone a transformative decade, primarily catalyzed by the groundbreaking data-driven approach introduced by F. Kratzert et al. (2018) utilizing LSTM networks (Hochreiter & Schmidhuber, 1997). These networks leverage extensive datasets and intricate model structures, outshining traditional hydrological models, albeit with the caveat of being computationally intensive during training. This prompts a critical inquiry into the requisite volume and complexity of data for constructing a dependable and resilient hydrological model.


In this study, we employ a hybrid model that amalgamates the strengths of classical hydrological models with the data-driven approach. These modified models are derived from the LSTM models developed by F. Kratzert and team, in conjunction with classical hydrological models such as the Statkraft Hydrology Forecasting Toolbox (SHyFT) from Statkraft and the Distributed Regression Hydrological Model (DRM) by Matheussen at Å Energi. The models were applied to sixty-five catchments in southern Norway, each characterized by diverse features and data records. Our analysis assesses the performance of these models under various scenarios of data availability, considering factors such as:


- Varying numbers of catchments selected based on size or location.
- The duration of the data records utilized for model calibration.
- Specific catchment characteristics and outputs from classical models employed as inputs 
(e.g., area, latitude, longitude, or additional variables).


Preliminary findings indicate that model inputs can be significantly stripped down without compromising model performance. With a limited set of catchment characteristics, the performance approaches that of the model with all characteristics, mitigating added uncertainty and model complexity. Additionally, increasing the length of data records enhances model performance, albeit with diminishing returns. Furthermore, our study reveals that augmenting catchments in the model does not necessarily yield a commensurate improvement in overall model performance. These insights contribute to refining our understanding of the interplay between data, model complexity, and performance in hydrological modeling.


The novelty in this research is that the hybrid models can be applied in a relatively small area, with few catchments and a limited number of climate stations and catchment characteristics compared to the CAMELS setup, used by Kratzert and still achieve improved results. 

How to cite: Beil-Myhre, B., Matheussen, B. V., and Shrestha, R.: How much data is needed for hydrological modeling? , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-11778, https://doi.org/10.5194/egusphere-egu24-11778, 2024.

EGU24-12068 | ECS | Orals | HS3.5

Hybrid Neural Hydrology: Integrating Physical and Machine Learning Models for Enhanced Predictions in Ungauged Basins 

Rajeev Shrestha, Bjarte Beil-Myhre, and Bernt Viggo Matheussen

Accurate prediction of streamflow in ungauged basins is a fundamental challenge in hydrology. The lack of hydrological observations and the inherent complexities in ungauged regions hinder accurate predictions, posing significant hurdles for water resource management and forecasting. Over time, efforts have been made to tackle this predicament, primarily utilizing physical hydrological models. However, these models need to be revised due to their reliance on site-specific data and their struggle to capture complex nonlinear relationships. Recent work by Kratzert et al. (2018) suggests that nonlinear regression models such as LSTM neural networks (Hochreiter & Schmidhuber, 1997) may outperform traditional physically based models. The authors demonstrate the application of LSTM models to ungauged prediction problems, noting that information about physical processes might not have been fully utilized in the modeling setup.

In response to these challenges, this research explores a novel approach by introducing a Hybrid Neural Hydrology (HNH) approach by fusing the strengths of physical hydrological models like Statkraft Hydrology Forecasting Toolbox (SHyFT), developed at Statkraft and the Distributed Regression Hydrological Model (DRM), developed by Matheussen at Å Energi with machine learning model, specifically Neural Hydrology, developed by F. Kratzert and team. By combining the information and structural insights of physically based models with the flexibility and adaptability of machine learning models, HNH seeks to leverage the complementary attributes of these methodologies. The combination is achieved by fusing the uncalibrated physical model with an LSTM based model. This hybridization seeks to enhance the model's adaptability and learning capabilities, leveraging available information from various sources to improve predictions in ungauged areas. Furthermore, this research investigates the impact of clustering catchments based on area to improve model performance.

The data used in this research includes dynamic variables such as precipitation, air temperature, wind speed, relative humidity, and observed streamflow obtained from sources such as the internal database at Å Energi, The Norwegian Water Resources and Energy Directorate (NVE), The Norwegian Meteorological Institute (MET), ECMWF (ERA5) and static attributes such as catchment size, mean elevation, forest fraction, lake fraction and reservoir fraction obtained from CORINE Land Cover and Høydedata (www.hoydedata.no).

This study presents HNH as a novel approach that synergistically integrates the structural insights of physical models with the adaptability of machine learning. Preliminary findings indicate promising outcomes from testing in 65 catchments in southern Norway. This suggests that information about physical processes and clustering catchments based on their similarities significantly improves the prediction quality in ungauged regions. This discovery underscores the potential of using hybrid models and clustering techniques to enhance the performance of predictive models in ungauged basins.

How to cite: Shrestha, R., Beil-Myhre, B., and Matheussen, B. V.: Hybrid Neural Hydrology: Integrating Physical and Machine Learning Models for Enhanced Predictions in Ungauged Basins, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-12068, https://doi.org/10.5194/egusphere-egu24-12068, 2024.

EGU24-12574 | ECS | Orals | HS3.5 | Highlight

Analyzing the performance and interpretability of hybrid hydrological models 

Eduardo Acuna, Ralf Loritz, Manuel Alvarez, Frederik Kratzert, Daniel Klotz, Martin Gauch, Nicole Bauerle, and Uwe Ehret

Hydrological hybrid models have been proposed as an option to combine the enhanced performance of deep learning methods with the interpretability of process-based models. Among the various hybrid methods available, the dynamic parameterization of conceptual models using LSTM networks has shown high potential. 

In this contribution, we extend our previous related work (Acuna Espinoza et al., 2023) by asking the questions: How well can hybrid models predict untrained variables, and how well do they generalize? We address the first question by comparing the internal states of the model against external data, specifically against soil moisture data obtained from ERA5-Land for 60 basins in Great Britain. We show that the process-based layer can reproduce the soil moisture dynamics with a correlation of 0.83, which indicates a good ability of this type of model to predict untrained variables. Moreover, we compare this method against existing alternatives used to extract non-target variables from purely data-driven methods (Lees et al., 2022), and discuss the differences in philosophy, performance, and implementation. Then, we address the second question by evaluating the capacity of such models to predict extreme events. Following the procedure proposed by Frame et al (2022), we train the hybrid models in low-flow regimes and test them in high-flow situations to evaluate the generalization capacity of such models and compare them against results from purely data-driven methods. Both experiments are done using large-sample data from the CAMELS-US and CAMELS-GB dataset.

With these new experiments, we contribute to answering the question of whether hybrid models give an actual advantage over purely data-driven techniques or not.

References

Acuna Espinoza, E., Loritz, R., Alvarez Chaves, M., Bäuerle, N., & Ehret, U.: To bucket or not to bucket? Analyzing the performance and interpretability of hybrid hydrological models with dynamic parameterization. EGUsphere, 1–22. https://doi.org/10.5194/egusphere-2023-1980, 2023.

Frame, J. M. and Kratzert, F. and Klotz, D. and Gauch, M. and Shalev, G. and Gilon, O. and Qualls, L. M. and Gupta, H. V. and Nearing, G. S., :Deep learning rainfall--runoff predictions of extreme events, Hydrology and Earth System Sciences, 26 ,3377-3392, https://doi.org/10.5194/hess-26-3377-2022, 2022

Lees, T., Reece, S., Kratzert, F., Klotz, D., Gauch, M., De Bruijn, J., Kumar Sahu, R., Greve, P., Slater, L., and Dadson, S. J.: Hydrological concept formation inside long short-term memory (LSTM) networks, Hydrology and Earth System Sciences, 26, 3079–3101, https://doi.org/10.5194/hess-26-3079-2022,  2022.

How to cite: Acuna, E., Loritz, R., Alvarez, M., Kratzert, F., Klotz, D., Gauch, M., Bauerle, N., and Ehret, U.: Analyzing the performance and interpretability of hybrid hydrological models, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-12574, https://doi.org/10.5194/egusphere-egu24-12574, 2024.

EGU24-12981 | ECS | Orals | HS3.5

Using Temporal Fusion Transformer (TFT) to enhance sub-seasonal drought predictions in the European Alps 

Annie Yuan-Yuan Chang, Konrad Bogner, Maria-Helena Ramos, Shaun Harrigan, Daniela I.V. Domeisen, and Massimiliano Zappa

In recent years, the European Alpine space has witnessed unprecedented low-flow conditions and drought events, affecting various economic sectors reliant on sufficient water availability, including hydropower production, navigation and transportation, agriculture, and tourism. As a result, there is an increasing need for decision-makers to have early warnings tailored to local low-flow conditions.

The EU Copernicus Emergency Management Service (CEMS) European Flood Awareness System (EFAS) has been instrumental in providing flood risk assessments across Europe with up to 15 days of lead time since 2012. Expanding its capabilities, the EFAS also generates long-range hydrological outlooks from sub-seasonal to seasonal horizons. Despite its original flood-centric design, previous investigations have revealed EFAS’s potential for simulating low-flow events. Building upon this finding, this study aims to leverage EFAS's anticipation capability to enhance the predictability of drought events in Alpine catchments, while providing support to trans-national operational services.

In this study, we integrate the 46-day extended-range EFAS forecasts into a hybrid setup for 106 catchments in the European Alps. Many studies have demonstrated Long Short-Term Memory (LSTM)’s capacity to produce skillful hydrological forecasts at various time scales. Here we employ the deep learning algorithm Temporal Fusion Transformer (TFT), an algorithm that combines aspects of LSTM networks with the Transformer architecture. The Transformer's attention mechanisms can focus on relevant time steps across longer sequences enabling TFT to capture both local temporal patterns as well as global dependencies. The role of the TFT is to improve the accuracy of low-flow predictions and to understand their spatio-temporal evolution. In addition to EFAS data, we incorporate features such as European weather regime data, streamflow climatology, and hydropower proxies. We also consider catchment characteristic information including glacier coverage and lake proximity. By incorporating its various attention mechanisms, makes TFT a more explainable algorithm than LSTMs, which helps us understand the driving factor for the forecast skill. Our evaluation uses EFAS re-forecast data as the benchmark and measures the reliability of ensemble forecasts using metrics like the Continuous Ranked Probability Skill Score (CRPSS).

Preliminary results show that a hybrid approach using the TFT algorithm can reduce the flashiness of EFAS during drought periods in some catchments, thereby improving drought predictability. Our findings will contribute to evaluating the potential of these forecasts for providing valuable information for skillful early warnings and assist in informing regional and local water resource management efforts in their decision-making.

How to cite: Chang, A. Y.-Y., Bogner, K., Ramos, M.-H., Harrigan, S., Domeisen, D. I. V., and Zappa, M.: Using Temporal Fusion Transformer (TFT) to enhance sub-seasonal drought predictions in the European Alps, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-12981, https://doi.org/10.5194/egusphere-egu24-12981, 2024.

EGU24-13417 | ECS | Orals | HS3.5

Evaluating physics-based representations of hydrological systems through hybrid models and information theory 

Manuel Álvarez Chaves, Eduardo Acuña Espinoza, Uwe Ehret, and Anneli Guthke

Hydrological models play a crucial role in understanding and predicting streamflow. Recently, hybrid models, combining both physical principles and data-driven approaches, have emerged as promising tools to extract insights into system functioning and increases in model predictive skill which are beyond traditional models.

However, the study by Acuña Espinoza et al. (2023) has raised the question whether the flexible data-driven component in a hybrid model might "overwrite" the interpretability of its physics-based counterpart. On the example of conceptual hydrological models with dynamic parameters tuned by LSTM networks, they showed that even in a case where the physics-based representation of the hydrological system is chosen to be nonsensical on purpose, the addition of the flexible data-driven component can lead to a well-performing hybrid model. This compensatory behavior highlights the need for a thorough evaluation of physics-based representations in hybrid hydrological models, i.e., hybrid models should be inspected carefully to understand why and how they predict (so well).

In this work, we provide a method to support this inspection: we objectively assess and quantify the contribution of the data-driven component to the overall hybrid model performance. Using information theory and the UNITE toolbox (https://github.com/manuel-alvarez-chaves/unite_toolbox), we measure the entropy of the (hidden) state-space in which the data-driven component of the hybrid model moves. High entropy in this setting means that the LSTM is doing a lot of "compensatory work", and hence alludes to an inadequate representation of the hydrological system in the physics-based component of the hybrid model. By comparing this measure among a set of alternative hybrid models with different physics-based representations, an order in the degree of realism of the considered representations can be established. This is very helpful for model evaluation and improvement as well as system understanding.

To illustrate our findings, we present examples from a synthetic case study where a true model does exist. Subsequently, we validate our approach in the context of regional predictions using CAMELS-GB data. This analysis highlights the importance of using diverse representations within hybrid models to ensure the pursuit of "the right answers for the right reasons". Ultimately, our work seeks to contribute to the advancement of hybrid modeling strategies that yield reliable and physically reasonable insights into hydrological systems.

References

  • Acuña Espinoza, E., Loritz, R., Álvarez Chaves, M., Bäuerle, N., & Ehret, U. (2023). To bucket or not to bucket? analyzing the performance and interpretability of hybrid hydrological models with dynamic parameterization. EGUsphere, 1–22. https://doi.org/10.5194/egusphere-2023-1980

How to cite: Álvarez Chaves, M., Acuña Espinoza, E., Ehret, U., and Guthke, A.: Evaluating physics-based representations of hydrological systems through hybrid models and information theory, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-13417, https://doi.org/10.5194/egusphere-egu24-13417, 2024.

EGU24-14280 | Orals | HS3.5

Quantifying Evapotranspiration and Gross Primary Productivity Across Europe Using Radiative Transfer Process-Guided Machine Learning 

Sheng Wang, Rui Zhou, Egor Prikaziuk, Kaiyu Guan, René Gislum, Christiaan van der Tol, Rasmus Fensholt, Klaus Butterbach-Bahl, Andreas Ibrom, and Jørgen Eivind Olesen

Accurately quantifying water and carbon fluxes between terrestrial ecosystems and the atmosphere is highly valuable for understanding ecosystem biogeochemical processes for climate change mitigation and ecosystem management. Remote sensing can provide high spatial and temporal resolution reflectance data of terrestrial ecosystems to support quantifying evapotranspiration (ET) and gross primary productivity (GPP).  Conventional remote sensing-based ET and GPP algorithms are either based on empirical data-driven approaches or process-based models. Empirical data-driven approaches often have high accuracy for cases within the source data domain, but lack the links to a mechanistic understanding of ecosystem processes. Meanwhile, process-based models have high generalizability with incorporating physically based soil-vegetation radiative transfer processes, but usually have lower accuracy. To integrate the strengths of data-driven and process-based approaches, this study developed a radiative transfer process-guided machine learning approach (PGML) to quantify ET and GPP across Europe. Specifically, we used the Soil Canopy Observation, Photochemistry, and Energy fluxes (SCOPE, van der Tol et al. 2009) radiative transfer model to generate synthetic datasets and developed a pre-trained neural network model to quantify ET and GPP. Furthermore, we utilized field measurements from 63 eddy covariance tower sites from 2016 to 2020 across Europe to fine-tune the neural networks with incorporating physical laws into the cost function. Results show that PGML can significantly improve the SCOPE simulations of net radiation (R2 from 0.91 to 0.96), sensible heat fluxes (R2 from 0.43 to 0.77), ET (R2 from 0.61 to 0.78), and GPP (R2 from 0.72 to 0.78) compared to eddy covariance observations. This study highlights the potential of PGML to integrate machine learning and radiative transfer models to improve the accuracy of land surface flux estimates for terrestrial ecosystems.

How to cite: Wang, S., Zhou, R., Prikaziuk, E., Guan, K., Gislum, R., van der Tol, C., Fensholt, R., Butterbach-Bahl, K., Ibrom, A., and Olesen, J. E.: Quantifying Evapotranspiration and Gross Primary Productivity Across Europe Using Radiative Transfer Process-Guided Machine Learning, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-14280, https://doi.org/10.5194/egusphere-egu24-14280, 2024.

Deep learning models for streamflow prediction have been widely used but are often considered as "black boxes" due to their lack of interpretability. To address this issue, the field has recently focused on Explainable Artificial Intelligence (XAI) methods to improve the transparency of these models. In this study, we aimed to investigate the influence of precipitation uncertainty on data-driven modeling and elucidate the hydrological significance of deep learning streamflow modeling in both temporal and spatial dimensions by Explainable Artificial Intelligence techniques. To achieve this, an LSTM model for time series prediction and a CNN-LSTM model for fusion spatial-temporal information are proposed. These models are driven by five sets of reanalyzed datasets. The contribution of precipitation before peak flow to runoff simulation is quantified, in order to identify the most important processes in runoff generation for each river basin. In addition, visualization techniques are employed to analyze the relationship between the weights of the convolutional layers in our models and the distribution of precipitation features. By doing so, we aimed to gain insights into the underlying mechanisms of the models' predictions.

The results of our study revealed several key findings. In the high-altitude areas of the Yangtze River's upper reaches, we found that snowmelt runoff, historical precipitation, and recent precipitation were the combined causes for floods. In the middle reach of the Yangtze River, floods were induced by the combined effect of historical and recent precipitation, except for the Ganjiang River, where historical precipitation events played a major role in controlling flood events. Through the visualization of convolutional layers, we discovered that areas with high convolutional layer weights had a greater impact on the model's predictions. We also observed a high similarity between the weight distribution of the convolutional layers and the spatial distribution of multi-year average precipitation in the upper reach river basins. In the middle reach, the weight distribution of the model's convolutional layers showed a strong correlation with the monthly maximum precipitation in the basin. Overall, this study provides valuable insights into the potential of deep learning models for streamflow prediction and enhances our understanding of the impacts of precipitation in the Yangtze River Basin.

How to cite: Tian, Y., Tan, W., and Yuan, X.: Revealing the key factors and uncertainties in data-driven hydrological prediction using Explainable Artificial Intelligence techniques, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-14666, https://doi.org/10.5194/egusphere-egu24-14666, 2024.

EGU24-16235 | ECS | Orals | HS3.5

Flow estimation from observed water levels using differentiable modeling for low-lying rivers affected by vegetation and backwater 

Phillip Aarestrup, Jonas Wied Pedersen, Michael Brian Butts, Peter Bauer-Gottwein, and Roland Löwe

Simulations of river flows and water levels are crucial for flood predictions and water resources management. Water levels are easy to observe using sensors, while the mapping between water levels and flows in rivers is usually derived from rating curves. However, rating curves frequently do not include geometry, backwater effects, and/or seasonal variations, which can limit their applicability – especially in stream systems that are affected by seasonal vegetation and backwater effects. To address this, we propose a differentiable model that merges a neural network with a physically based, steady-state implementation of the Saint-Venant equations. 

In the setup, the neural network is trained to predict seasonal variations caused by vegetation growth in Manning’s roughness based on inputs of meteorological forcing and time, while the physical model is responsible for converting flow estimates into water levels along the river channel. The framework efficiently estimates model parameters by tracking gradients through both the physical model and the neural network via backpropagation. This allows us to calibrate parameters for both the runoff and the Manning’s roughness from measured water levels, thus overcoming rating curve limitations while accounting for backwater, river geometry, and seasonal variations in roughness. 

We tested the model on a 20 km stretch of the Vejle River, Denmark, which is both heavily vegetated and affected by backwater from the coast. The model was trained across five water level sensors using two years of data (2020-2022). When evaluated against 10 years of observed flow measurements (2007-2017), the model demonstrated a Mean Absolute Relative Error (MARE) of 10% compared to manually gauged discharge observations. This is comparable to the estimated uncertainty of 10% in the discharge measurements.  

The framework enables a calibration of dynamic Manning roughness within a few hours, and therefore offers a scalable solution for estimating river flows from water levels when cross-section information is available. Potential applications span across many disciplines in water resource management. 

How to cite: Aarestrup, P., Pedersen, J. W., Butts, M. B., Bauer-Gottwein, P., and Löwe, R.: Flow estimation from observed water levels using differentiable modeling for low-lying rivers affected by vegetation and backwater, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-16235, https://doi.org/10.5194/egusphere-egu24-16235, 2024.

EGU24-17842 | ECS | Orals | HS3.5 | Highlight

Deep learning based differentiable/hybrid modelling of the global hydrological cycle 

Zavud Baghirov, Basil Kraft, Martin Jung, Marco Körner, and Markus Reichstein

The integration of machine learning (ML) and process based modeling (PB) in so-called hybrid models, also known as differentiable modelling, has recently gained popularity in the geoscientific community (Reichstein et al. 2019; Shen et al. 2023). The approach aims to address limitations in both ML (data adaptive but difficult to interpret and physically inconsistent) and PB (physically consistent and interpretable but biased). It holds significant potential for studying uncertain processes in the global water cycle (Kraft et al. 2022).

In this work, we developed a differentiable/hybrid model of the global hydrological cycle by fusing deep learning with a custom PB model. The model inputs include air temperature, precipitation, net radiation as dynamic forcings, and static features like soil texture as input to a long short-term memory (LSTM) model. The LSTM represents the uncertain and less understood spatio-temporal parameters which are directly used in a conceptual hydrological model. Simultaneously, we use fully connected neural networks (FCNN) to represent the uncertain spatial parameters. In the hydrological model we represent key water fluxes (e.g. transpiration, evapotranspiration (ET), runoff) and storages (snow, soil moisture and groundwater). The model is constrained against the observation-based data, like terrestrial water storage (TWS) anomalies (GRACE), fAPAR (MODIS) and snow water equivalent (GLOBSNOW).

Building upon previous work (Kraft et al. 2022), we improved the representations of key hydrological processes. We now explicitly estimate vegetation state that is directly used to partition ET into transpiration, soil and interception evaporation. We also estimate rooting-zone water storage capacity—a key hydrological parameter that is still highly uncertain. To asses the robustness of the estimated parameters, we quantify equifinality by training multiple models with random weight initialisation in a 10-fold cross validation setup.

The model learns reasonable spatial and spatio-temporal patterns of critical, yet uncertain, hydrological parameters as latent variables. For example, we assess and show that the estimations of global spatial patterns on rooting-zone water storage capacity and transpiration versus ET are plausible. Equifinality quantification indicates that the dynamic patterns of the modelled water storages are robust, while there is a large uncertainty in the mean of soil moisture and TWS.

References

Kraft, Basil, Martin Jung, Marco Körner, Sujan Koirala, and Markus Reichstein. 2022. “Towards Hybrid Modeling of the Global Hydrological Cycle.” Hydrology and Earth System Sciences 26 (6): 1579–1614.

Reichstein, Markus, Gustau Camps-Valls, Bjorn Stevens, Martin Jung, Joachim Denzler, Nuno Carvalhais, et al. 2019. “Deep Learning and Process Understanding for Data-Driven Earth System Science.” Nature 566 (7743): 195–204.

Shen, Chaopeng, Alison P Appling, Pierre Gentine, Toshiyuki Bandai, Hoshin Gupta, Alexandre Tartakovsky, Marco Baity-Jesi, et al. 2023. “Differentiable Modelling to Unify Machine Learning and Physical Models for Geosciences.” Nature Reviews Earth & Environment 4 (8): 552–67.

How to cite: Baghirov, Z., Kraft, B., Jung, M., Körner, M., and Reichstein, M.: Deep learning based differentiable/hybrid modelling of the global hydrological cycle, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-17842, https://doi.org/10.5194/egusphere-egu24-17842, 2024.

EGU24-20112 | ECS | Posters on site | HS3.5

Data-driven global projection of future flooding in 18.5 million river reaches 

Boen Zhang, Louise Slater, Simon Moulds, Michel Wortmann, Yinxue Liu, Jiabo Yin, and Xihui Gu

Reliable flood projection is crucial for designing suitable flood protection structures and for enhancing resilience in vulnerable regions. However, projections of future flooding suffer from cascading uncertainties arising from the climate model outputs, emission scenarios, hydrological models, and the shortage of observations in data-sparse regions. To overcome these limitations, we design a new hybrid model, blending machine learning and climate model simulations, for global-scale projection of river flooding. This is achieved by training a random forest model directly on climate simulations from 20 CMIP6 models over the historical period (1985−2014), with extreme discharges observed at approximately 15,000 hydrologic stations as the target variable. The random forest model also includes static geographic predictors including land cover, climate, geomorphology, soil, human impacts, and hydrologic signatures. We make the explicit assumption that the random forest model can ‘learn’ systematic biases in the relationship between the climate simulations and flood regimes in different regions of the globe. We then apply the well-calibrated random forest model to a new vector-based, global river network in approximately 18.51 million reaches with drainage areas greater than 100 km2. Global changes in flood hazard are projected for the 21st century (2015−2100) under SSP2-4.5 and SSP5-8.5. We show that the data-driven method reproduces historical annual maximum discharges better than the physically-based hydrological models driven by bias-corrected climate simulations in the ISIMIP3b experiment. We then use the machine learning model with explainable AI to diagnose spatial biases in the climate simulations and future flood projections in different regions of the globe.

How to cite: Zhang, B., Slater, L., Moulds, S., Wortmann, M., Liu, Y., Yin, J., and Gu, X.: Data-driven global projection of future flooding in 18.5 million river reaches, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-20112, https://doi.org/10.5194/egusphere-egu24-20112, 2024.

Machine learning has long been restricted by the mystery of its black box, especially in the fields like geosciences that emphasizes clear expressions of mechanisms. To deal with that issue, we provided a fundamental framework combining two branches, clusters and regressions in machine learning, specifically, spectral clustering in unsupervised clustering methods and artificial neural networks in regression models, to resemble calculations in process-based models. With a case study of evapotranspiration, it was demonstrated that our framework was not only able to discern two processes, aerodynamics and energy, similar to the process-based model, i.e., Penman-Monteith formula, but also provided a third space for potential underrepresented process from canopy or ecosystems. Meanwhile, with only a few hundred of training data in most sites, the simulation of evapotranspiration achieved a higher accuracy (R2 of 0.92 and 0.82; RMSE of 12.41W/m2 and 8.11 W/m2 in training set and test set respectively) than commonly used machine learning approaches, like artificial neural networks in a scale of 100,000 training set (R2 of 0.85 and 0.81; RMSE of 42.33W/m2 and 46.73 W/m2). In summary, our method provides a new direction of hybridizing machine learning approaches and mechanisms for future work, which is able to tell mechanisms from a little amount of data, and thus could be utilized in validating the known and even exploring the unknown knowledge by providing reference before experiments and mathematical derivations.

How to cite: Hu, Y. and Jiang, Y.: Interpretably reconstruct physical processes with combined machine learning approaches, a case study of evapotranspiration, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-20579, https://doi.org/10.5194/egusphere-egu24-20579, 2024.

EGU24-20602 | ECS | Orals | HS3.5

Enhanced Continental Runoff Prediction through Differentiable Muskingum-Cunge Routing (δMC-CONUS-hydroDL2) 

Tadd Bindas, Yalan Song, Jeremy Rapp, Kathryn Lawson, and Chaopeng Shen

Recent advancements in flow routing models have enabled learning from big data using differentiable modeling techniques. However, their application remains constrained to smaller basins due to limitations in computational memory and hydrofabric scaling. We propose a novel methodology to scale differentiable river routing from watershed (HUC10) to continental scales using the δMC-CONUS-hydroDL2 model. Mimicking the Muskingum-Cunge routing model, this approach aims to enhance flood wave timing prediction and Manning’s n parameter learning across extensive areas. We employ the δHBV-HydroDL model, trained on the 3000 GAGES-II dataset, for streamflow predictions across CONUS HUC10 basins. These predictions are then integrated with MERIT basin data and processed through our differentiable routing model, which learns reach-scale parameters like Manning’s n and spatial channel coefficient q via an embedded neural network. This approach enhances national-scale flood simulations by leveraging a learned Manning’s n parameterization, directly contributing to the refinement of CONUS-scale flood modeling. Furthermore, this method shows promise for global application, contingent upon the availability of streamflow predictions and MERIT basin data. Our methodology represents a significant step forward in the spatial scaling of differentiable river routing models, paving the way for more accurate and extensive flood simulation capabilities.

How to cite: Bindas, T., Song, Y., Rapp, J., Lawson, K., and Shen, C.: Enhanced Continental Runoff Prediction through Differentiable Muskingum-Cunge Routing (δMC-CONUS-hydroDL2), EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-20602, https://doi.org/10.5194/egusphere-egu24-20602, 2024.

EGU24-21725 | ECS | Posters on site | HS3.5

Machine Learning Insights into Aquifer Recharge: Site suitability analysis in season water availability scenarios 

Valdrich Fernandes, Perry de Louw, Coen Ritsema, and Ruud Bartholomeus

Groundwater models are valuable tools for optimising decisions that influence groundwater flow. Spatially distributed models represent groundwater levels across the entire area from where essential information can be extracted, directly aiding in the decision-making process. In our previous study, we explored different machine learning (ML) models as faster alternatives to predict the increase in stationary groundwater head due to artificial recharge in unconfined aquifers while considering a wider spatial extent (832 columns x 1472 rows, totalling 765 km2) than previous ML groundwater models. The trained ML model accurately estimates the increase in groundwater head within 0.24 seconds, achieving a Nash-Sutcliffe efficiency of 0.95. This allows quick analysis of site suitability at potential recharge rates. This study extends the approach to incorporate seasonal variation in water availability, illustrating the concept of storing excess water during winter to meet heightened demands during summer, when water availability is minimal. Additionally, we quantify the impacts of the local properties, geohydrological and surface water network properties, on the storage capacity by training ML models on estimating the summer decay rate of stored water in hypothetical aquifer recharge sites.  

Among 720 hypothetical recharge sites, we vary its location, recharge rate and size to capture various combinations of local properties in the catchment. Artificial recharge is modeled using a MODFLOW-based groundwater model, representing the geo-hydrological properties and the surface water network in the Baakse Beek catchment in the Netherlands. The recharge is simulated from October 2011 till February 2012 with the remainder of the year simulated without any artificial recharge. Based on the modeled heads, the decay rate of stored water is calculated for the period until October. This calculated decay rate, in combination with the local properties are used to train and evaluate the ML model. The relative contributions of properties to the decay rate are quantified using the latest developments in explainable AI techniques. Techniques such as permutation importance and Ceteris paribus profiles not only help categorize the suitability of potential recharge sites but also quantify the relative contribution of each property. By leveraging these insights, water managers can make informed decisions regarding site improvement measures. 

How to cite: Fernandes, V., de Louw, P., Ritsema, C., and Bartholomeus, R.: Machine Learning Insights into Aquifer Recharge: Site suitability analysis in season water availability scenarios, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-21725, https://doi.org/10.5194/egusphere-egu24-21725, 2024.

EGU24-566 | ECS | Orals | HS3.4

A Transformer-Based Data-Driven Model for Real-Time Spatio-Temporal Flood Prediction 

Matteo Pianforini, Susanna Dazzi, Andrea Pilzer, and Renato Vacondio

Among the non-structural strategies for mitigating the huge economic losses and casualties caused by floods, the implementation of early-warning systems based on real-time forecasting of flood maps is one of the most effective. The high computational cost associated with two-dimensional (2D) hydrodynamic models, however, prevents their practical application in this context. To overcome this drawback, “data-driven” models are gaining considerable popularity due to their high computational efficiency for predictions. In this work, we introduce a novel surrogate model based on the Transformer architecture, named FloodSformer (FS), that efficiently predicts the temporal evolution of inundation maps, with the aim of providing real-time flood forecasts. The FS model combines an encoder-decoder (2D Convolutional Neural Network) with a Transformer block that handles temporal information. This architecture extracts the spatiotemporal information from a sequence of consecutive water depth maps and predicts the water depth map at one subsequent instant. An autoregressive procedure, based on the trained surrogate model, is employed to forecast tens of future maps.

As a case study, we investigated the hypothetical inundation due to the collapse of the flood-control dam on the Parma River (Italy). Due to the absence of real inundation maps, the training/testing dataset for the FS model was generated from numerical simulations performed through a 2D shallow‐water code (PARFLOOD). Results show that the FS model is able to recursively forecast the next 90 water depth maps (corresponding to 3 hours for this case study, in which maps are sampled at 2-minute intervals) in less than 1 minute. This is achieved while maintaining an accuracy deemed entirely acceptable for real-time applications: the average Root Mean Square Error (RMSE) is about 10 cm, and the differences between ground-truth and predicted maps are generally lower than 25 cm in the floodable area for the first 60 predicted frames. In conclusion, the short computational time and the good accuracy ensured by the autoregressive procedure make the FS model suitable for early-warning systems.

How to cite: Pianforini, M., Dazzi, S., Pilzer, A., and Vacondio, R.: A Transformer-Based Data-Driven Model for Real-Time Spatio-Temporal Flood Prediction, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-566, https://doi.org/10.5194/egusphere-egu24-566, 2024.

EGU24-571 | ECS | Posters on site | HS3.4

Hydrological Significance of Input Sequence Lengths in LSTM-Based Streamflow Prediction 

Farzad Hosseini Hossein Abadi, Cristina Prieto Sierra, Grey Nearing, Cesar Alvarez Diaz, and Martin Gauch

Abstract

Hydrological modeling of flashy catchments, susceptible to floods, represents a significant practical challenge.  Recent application of deep learning, specifically Long Short-Term Memory networks (LSTMs), have demonstrated notable capability in delivering accurate hydrological predictions at daily and hourly time intervals (Gauch et al., 2021; Kratzert et al., 2018).

In this study, we leverage a multi-timescale LSTM (MTS-LSTM (Gauch et al., 2021)) model to predict hydrographs in flashy catchments at hourly time scales. Our primary focus is to investigate the influence of model hyperparameters on the performance of regional streamflow models. We present methodological advancements using a practical application to predict streamflow in 40 catchments within the Basque Country (North of Spain).

Our findings show that 1) hourly and daily streamflow predictions exhibit high accuracy, with Nash-Sutcliffe Efficiency (NSE) reaching values as high as 0.941 and 0.966 for daily and hourly data, respectively; and 2) hyperparameters associated with the length of the input sequence exert a substantial influence on the performance of a regional model. Consistently optimal regional values, following a systematic hyperparameter tuning, were identified as 3 years for daily data and 12 weeks for hourly data. Principal component analysis (PCA) shows that the first principal component explains 12.36% of the variance among the 12 hyperparameters. Within this set of hyperparameters, the input sequence lengths for hourly data exhibit the highest load in PC1, with a value of -0.523; the load of the input sequence length for daily data is also very high (-0.36). This suggests that these hyperparameters strongly contribute to the model performance.

Furthermore, when utilizing a catchment-scale magnifier to determine optimal hyperparameter settings for each catchment, distinctive sequence lengths emerge for individual basins. This underscores the necessity of customizing input sequence lengths based on the “uniqueness of the place” (Beven, 2020), suggesting that each catchment may demand specific hydrologically meaningful daily and hourly input sequence lengths tailored to its unique characteristics. In essence, the true input sequence length of a catchment may encapsulate hydrological information pertaining to water transit over short and long-term periods within the basin. Notably, the regional daily sequence length aligns with the highest local daily sequence values across all catchments.

In summary, our investigation stresses the critical role of the input sequence length as a hyperparameter in LSTM networks. More broadly, this work is a step towards a better understanding and achieving accurate hourly predictions using deep learning models.

 

Keywords

Hydrological modeling; Streamflow Prediction; LSTM networks; Hyperparameters configurations; Input sequence lengths

 

References:

Beven, K. (2020). Deep learning, hydrological processes and the uniqueness of place. Hydrological Processes, 34(16), 3608–3613. doi:10.1002/hyp.13805

Gauch, M., Kratzert, F., Klotz, D., Nearing, G., Lin, J., and Hochreiter, S. (2021). Rainfall–runoff prediction at multiple timescales with a single Long Short-Term Memory network, Hydrol. Earth Syst. Sci., 25, 2045–2062, DOI:10.5194/hess-25-2045-2021.

Kratzert, F., Klotz, D., Brenner, C., Schulz, K., & Herrnegger, M. (2018). Rainfall--runoff modelling using Long Short-Term Memory (LSTM) networks. Hydrology and Earth System Sciences, 22(11), 6005–6022. DOI:10.5194/hess-22-6005-2018.

 

How to cite: Hosseini Hossein Abadi, F., Prieto Sierra, C., Nearing, G., Alvarez Diaz, C., and Gauch, M.: Hydrological Significance of Input Sequence Lengths in LSTM-Based Streamflow Prediction, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-571, https://doi.org/10.5194/egusphere-egu24-571, 2024.

EGU24-811 | ECS | Orals | HS3.4

A deep learning approach for spatio-temporal prediction of stable water isotopes in soil moisture 

Hyekyeng Jung, Chris Soulsby, and Dörthe Tetzlaff

Water flows and related mixing dynamics in the unsaturated zone are difficult to measure directly, so stable water isotope tracers have been used successfully to quantify flux and storage dynamics and to constrain process-based hydrological models as proxy data. In this study, a data-driven model based on deep learning was adapted to interpolate and extrapolate spatio-temporal isotope signals of δ18O and δ2H in soil water in three dimensions. Further, this was also used to help quantify evapotranspiration and groundwater recharge processes in the unsaturated zone. To consider both spatial and temporal dependencies of water isotope signals in the model design, the output space was decomposed into temporal basis functions and spatial coefficients using singular value decomposition. Then, temporal functions and spatial coefficients were predicted separately by specialized deep learning models in interdependencies among target data, such as the LSTM model and convolutional neural network. Finally, the predictions by the models were integrated and analyzed post-hoc using XAI tools.

Such an integrated framework has the potential to improve understanding of model behavior based on features (e.g., climate, hydrological component) connected to either temporal or spatial information. Furthermore, the model can serve as a surrogate model for process-based hydrological models, improving the use of process-based models as learning tools.

How to cite: Jung, H., Soulsby, C., and Tetzlaff, D.: A deep learning approach for spatio-temporal prediction of stable water isotopes in soil moisture, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-811, https://doi.org/10.5194/egusphere-egu24-811, 2024.

EGU24-2872 | ECS | Posters on site | HS3.4

Runoff coefficient modelling using Long Short-Term Memory (LSTM) in the Rur catchment, Germany 

Arash Rahi, Mehdi Rahmati, Jacopo Dari, and Renato Morbidelli

This research examines the effectiveness of Long Short-Term Memory (LSTM) models in predicting runoff coefficient (Rc) within the Rur basin at the Stah outlet (Germany) during the period from 1961 to 2021; monthly data of temperature (T), precipitation (P), soil water storage (SWS), and total evaporation (ETA) are used as an input. Because of the complexity in predicting undecomposed Rc time series due to noise, a novel approach incorporating discrete wavelet transform (DWT) to decompose the original Rc at five levels is proposed.

The investigation identifies overfitting challenges at level-1, gradually mitigated in subsequent decomposition levels, particularly in level-2, while other levels remain tuned. Reconstructing Rc using modelled decomposition coefficients yielded Nash-Sutcliffe efficacy (NSE) values of 0.88, 0.79, and 0.74 for the training, validation, and test sets, respectively. Comparative analysis highlights that modelling undecomposed Rc with LSTM yields to a minor accuracy, emphasizing the pivotal role of decomposition techniques in tandem with LSTM for enhanced model performances.

This study provides novel insights to address challenges related to noise effects and temporal dependencies in Rc modelling; through a comprehensive analysis of the interplay between atmospheric conditions and observed data, the research contributes in advancing predictive modelling in hydrology.

How to cite: Rahi, A., Rahmati, M., Dari, J., and Morbidelli, R.: Runoff coefficient modelling using Long Short-Term Memory (LSTM) in the Rur catchment, Germany, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-2872, https://doi.org/10.5194/egusphere-egu24-2872, 2024.

EGU24-2939 | ECS | Orals | HS3.4

Probabilistic streamflow forecasting using generative deep learning models 

Mohammad Sina Jahangir and John Quilty

The significance of probabilistic hydrological forecasting has grown in recent years, offering crucial insights for risk-based decision-making and effective flood management. This study explores generative deep learning models, specifically the conditional variational autoencoder (CVAE), for probabilistic streamflow forecasting. This innovative approach is applied for forecasting streamflow one to seven days (s) ahead in 75 Canadian basins included in the open-source Canadian model parameter experiment (CANOPEX) database. CVAE is compared against two benchmark quantile-based deep learning models: the quantile-based encoder-decoder (ED) and the quantile-based CVAE (QCVAE).

Over 9000 deep learning models are developed based on different input variables, basin characteristics, and model structures and evaluated regarding point forecast accuracy and forecast reliability. Results highlight CVAE‘s superior reliability, showing a median reliability of 92.49% compared to 87.35% for ED and 84.59% for QCVAE (considering a desired 90% confidence level). However, quantile-based forecast models exhibit marginally better point forecasts, as evidenced by Kling-Gupta efficiency (KGE), with a median KGE of 0.90 for ED and QCVAE (compared to 0.88 for CVAE). Notably, the CVAE model provides reliable probabilistic forecasts in basins with low point forecast accuracy.

The developed generative deep learning models can be used as a benchmark for probabilistic streamflow forecasting due to the use of the open-source CANOPEX dataset. Overall, the results of this study contribute to the expanding field of generative deep learning models in hydrological forecasting, offering a general framework that applies to forecasting other hydrological variables as well (precipitation and soil moisture).

How to cite: Jahangir, M. S. and Quilty, J.: Probabilistic streamflow forecasting using generative deep learning models, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-2939, https://doi.org/10.5194/egusphere-egu24-2939, 2024.

EGU24-6432 | ECS | Orals | HS3.4

A Step Towards Global Hydrologic Modelling: Accurate Streamflow Predictions In Pseudo-Ungauged Basins of Japan 

Hemant Servia, Frauke Albrecht, Samuel Saxe, Nicolas Bierti, Masatoshi Kawasaki, and Shun Kurihara

In addressing the challenge of streamflow prediction in ungauged basins, this study leveraged deep learning (DL) models, especially long short-term memory (LSTM) networks, to predict streamflow for pseudo ungauged basins in Japan. The motivation stems from the recognized limitations of traditional hydrological models in transferring their performance beyond the calibrated basins. Recent research suggests that DL models, especially those trained on multiple catchments, demonstrate improved predictive capabilities utilizing the concept of streamflow regionalization. However, the majority of these studies were confined to geographic regions within the United States.

For this study, a total number of 211 catchments were delineated and investigated, distributed across all four primary islands of Japan (Kyushu - 32, Shikoku - 13, Honshu - 127, and Hokkaido - 39) encompassing a comprehensive sample of hydrological systems within the region. The catchments were obtained corresponding to the streamflow observation points and their combined area represented more than 43% of Japan's total land area, after accounting for overlaps. After cleaning and refining the streamflow dataset, the remaining catchments (198) were divided into training (~70%), validation (~20%), and holdout test (~10%) sets. A combination of dynamic (time-varying) and static (constant) variables were obtained on a daily basis corresponding to the daily streamflow data and provided to the models as input features. However, the final model accorded higher significance to dynamic features in comparison to the static ones. Although the models were trained on daily time steps, the results were aggregated to monthly timescale. The main evaluation metrics included the Nash-Sutcliffe Efficiency (NSE) and Pearson’s correlation coefficient (r). The final model achieved a median NSE of 0.96, 0.83, & 0.78, and a median correlation of 0.98, 0.92, & 0.91 corresponding to the training, validation, and test catchments, respectively. For the validation catchments, 90% exhibited NSE values greater than 0.50, and 97% demonstrated a correlation surpassing 0.70. Correspondingly, these proportions were observed at 77% and 91% for the test catchments.

The results presented in this study demonstrate the feasibility and efficacy of developing a data-driven model for streamflow prediction in ungauged basins utilizing streamflow regionalization. The final model exhibits commendable performance, as evidenced by high NSE and correlation coefficients across the majority of the catchments. Importantly, the model's ability to generalize to unseen data is highlighted by its remarkable performance on the holdout test set, with only a few instances of lower NSE values (< 0.50) and correlation coefficients (< 0.70).

How to cite: Servia, H., Albrecht, F., Saxe, S., Bierti, N., Kawasaki, M., and Kurihara, S.: A Step Towards Global Hydrologic Modelling: Accurate Streamflow Predictions In Pseudo-Ungauged Basins of Japan, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-6432, https://doi.org/10.5194/egusphere-egu24-6432, 2024.

EGU24-6846 | ECS | Orals | HS3.4

Towards Fully Distributed Rainfall-Runoff Modelling with Graph Neural Networks 

Peter Nelemans, Roberto Bentivoglio, Joost Buitink, Ali Meshgi, Markus Hrachowitz, Ruben Dahm, and Riccardo Taormina

Fully distributed hydrological models take into account the spatial variability of a catchment, allowing for a more accurate representation of its heterogeneity, and assessing its hydrological response at multiple locations. However, physics-based fully distributed models can be time-consuming when it comes to model runtime and calibration, especially for large-scale catchments. On the other hand, deep learning models have shown great potential in the field of hydrological modelling, outperforming lumped rainfall-runoff conceptual models, and improving prediction in ungauged basins via catchment transferability. Despite these advances, the field still lacks a multivariable, fully distributed hydrological deep learning model capable of generalizing to unseen catchments. To address the aforementioned challenges associated with physics-based distributed models and deep learning models, we explore the possibility of developing a fully distributed deep learning model by using Graph Neural Networks (GNN), an extension of deep learning methods to non-Euclidean topologies including graphs and meshes.

We develop a surrogate model of wflow_sbm, a fully distributed, physics-based hydrological model, by exploiting the similarities between its underlying functioning and GNNs. The GNN uses the same input as wflow_sbm: distributed static parameters based on physical characteristics of the catchment and gridded dynamic forcings. The GNN is trained to produce the same output as wflow_sbm, predicting multiple gridded variables related to rainfall-runoff, such as streamflow, actual evapotranspiration, subsurface flow, saturated and unsaturated groundwater storage, snow storage, and runoff. We show that our GNN model achieves high performance in unseen catchments, indicating that GNNs are a viable option for fully distributed multivariable hydrological models capable of generalizing to unseen regions. Furthermore, the GNN model achieves a significant computational speedup compared to wflow_sbm. We will continue this research, using the GNN-based surrogate models as pre-trained backbones to be fine-tuned with measured data, ensuring accurate model adaptation, and enhancing their practical applicability in diverse hydrological scenarios.

How to cite: Nelemans, P., Bentivoglio, R., Buitink, J., Meshgi, A., Hrachowitz, M., Dahm, R., and Taormina, R.: Towards Fully Distributed Rainfall-Runoff Modelling with Graph Neural Networks, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-6846, https://doi.org/10.5194/egusphere-egu24-6846, 2024.

This research created a deep neural network (DNN)-based hydrologic model for an urban watershed in South Korea using multiple LSTM (long short-term memory) units and a fully connected layer. The model utilized 10-minute intervals of radar-gauge composite precipitation and temperature data across 239 grid cells, each 1 km in resolution, to simulate watershed flow discharge every 10 minutes. It showed high accuracy during both the calibration (2013–2016) and validation (2017–2019) periods, with Nash–Sutcliffe efficiency coefficient values of 0.99 and 0.67, respectively. Key findings include: 1) the DNN model's runoff–precipitation ratio map closely matched the imperviousness ratio map from land cover data, demonstrating the model's ability to learn precipitation partitioning without prior hydrological information; 2) it effectively mimicked soil moisture-dependent runoff processes, crucial for continuous hydrologic models; and 3) the LSTM units displayed varying temporal responses to precipitation, with units near the watershed outlet responding faster, indicating the model's capability to differentiate between hydrological components like direct runoff and groundwater-driven baseflow.

How to cite: Kim, D.: Exploring How Machines Model Water Flow: Predicting Small-Scale Watershed Behavior in a Distributed Setting, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-7186, https://doi.org/10.5194/egusphere-egu24-7186, 2024.

EGU24-8102 | ECS | Orals | HS3.4

Improving the Generalizability of Urban Pluvial Flood Emulators by Contextualizing High-Resolution Patches 

Tabea Cache, Milton S. Gomez, Jovan Blagojević, Tom Beucler, João P. Leitão, and Nadav Peleg

Predicting future flood hazards in a changing climate requires adopting a stochastic framework due to the multiple sources of uncertainties (e.g., from climate change scenarios, climate models, or natural variability). This requires performing multiple flood inundation simulations which are computationally costly. Data-driven models can help overcome this issue as they can emulate urban flood maps considerably faster than traditional flood simulation models. However, their lack of generalizability to both terrain and rainfall events still limits their application. Additionally, these models face the challenge of not having sufficient training data. This led state-of-the-art models to adopt a patch-based framework, where the study area is first divided into local patches (i.e., broken into smaller terrain images) that are subsequently merged to reconstruct the whole study area prediction. The main drawback of this method is that the model is blind to the surroundings of the local patch. To overcome this bottleneck, we developed a new deep learning model that includes patches' contextual information while keeping high-resolution information of the local patch. We trained and tested the model in the city of Zurich, at spatial resolution of 1 m. The evaluation focused on 1-hour rainfall events at 5 min temporal resolution and encompassing extreme precipitation return periods from 2- to 100-year. The results show that the proposed CNN-attention model outperforms the state-of-the-art patch-based urban flood emulator. First, our model can faithfully represent flood depths for a wide range of extreme rainfall events (peak rainfall intensities ranging from 42.5 mm h-1 to 161.4 mm h-1). Second, the model's terrain generalizability was assessed in distinct urban settings, namely Luzern and Singapore. Our model accurately identifies water accumulation locations, which constitutes an improvement compared to current models. Using transfer learning, the model was successfully retrained in the new cities, requiring only a single rainfall event to adapt the model to new terrains while preserving adaptability across diverse rainfall conditions. Our results suggest that by integrating contextual terrain information with local terrain patches, our proposed model effectively generates high-resolution urban pluvial flood maps, demonstrating applicability across varied terrains and rainfall events.

How to cite: Cache, T., Gomez, M. S., Blagojević, J., Beucler, T., Leitão, J. P., and Peleg, N.: Improving the Generalizability of Urban Pluvial Flood Emulators by Contextualizing High-Resolution Patches, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-8102, https://doi.org/10.5194/egusphere-egu24-8102, 2024.

EGU24-9190 | ECS | Orals | HS3.4

Learning Catchment Features with Autoencoders 

Alberto Bassi, Antonietta Mira, Marvin Höge, Fabrizio Fenicia, and Carlo Albert

By employing Machine Learning techniques on the US-CAMELS dataset, we discern a minimal number of streamflow features. Together with meteorological forcing, these features enable an approximate reconstruction of the entire streamflow time-series. This task is achieved through the application of an explicit noise conditional autoencoder, wherein the meteorological forcing is inputted to the decoder to encourage the encoder to learn streamflow features exclusively related to landscape properties. The optimal number of encoded features is determined with an intrinsic dimension estimator. The accuracy of reconstruction is then compared with models that take a subset of static catchment attributes (both climate and landscape attributes) in addition to meteorological forcing variables. Our findings suggest that attributes gathered by experts encompass nearly all pertinent information regarding the input/output relationship. This information can be succinctly summarized with merely three independent streamflow features. These features exhibit a strong correlation with the baseflow index and aridity indicators, aligning with the observation that predicting streamflow in dry catchments or with a high baseflow index is more challenging. Furthermore, correlation analysis underscores the significance of soil-related and vegetation attributes. These learned features can also be associated with parameters in conceptual hydrological models such as the GR model family.

How to cite: Bassi, A., Mira, A., Höge, M., Fenicia, F., and Albert, C.: Learning Catchment Features with Autoencoders, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-9190, https://doi.org/10.5194/egusphere-egu24-9190, 2024.

EGU24-9446 | ECS | Orals | HS3.4

Skilful prediction of mid-term sea surface temperature using 3D self-attention-based neural network 

Longhao Wang, Yongqiang Zhang, and Xuanze Zhang

Sea surface temperature (SST) is a critical parameter in the global ocean-atmospheric system, exerting a substantial impact on climate change and extreme weather events like droughts and floods. The precise forecasting of future SSTs is thus vital for identifying such weather anomalies. Here we present a novel three-dimensional (3D) neural network model based on self-attention mechanisms and Swin-Transformer for mid-term SST predictions. This model, integrating both climatic and temporal features, employs self-attention to proficiently capture the temporal dynamics and global patterns in SST. This approach significantly enhances the model's capability to detect and analyze spatiotemporal changes, offering a more nuanced understanding of SST variations. Trained on 59 years of global monthly ERA5-Land reanalysis data, our model demonstrates strong deterministic forecast capabilities in the test period. It employs a convolution strategy and global attention mechanism, resulting in faster and more accurate training compared to traditional methods, such as Convolutional Neural Network with Long short-term memory (CNN-LSTM). The effectiveness of this SST prediction model highlights its potential for extensive multidimensional modelling applications in geosciences.

How to cite: Wang, L., Zhang, Y., and Zhang, X.: Skilful prediction of mid-term sea surface temperature using 3D self-attention-based neural network, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-9446, https://doi.org/10.5194/egusphere-egu24-9446, 2024.

Traditional hydrological models have long served as the standard for predicting streamflow across temporal and spatial domains. However, a persistent challenge in modelling lies in mitigating bias inherent in streamflow estimation due to both random and systemic errors in the employed model. Removal of this bias is pivotal for effective water resources management and resilience against extreme events, especially amidst evolving climate conditions. An innovative solution to address this challenge involves the integration of hydrological models with deep learning methods, known as hybridisation. Long Short-Term Memory networks (LSTM), have emerged as a promising and efficient approach to enhancing streamflow estimation. This study focuses on coupling LSTM with a physically distributed model, Wflow_sbm, to serve as a post-processor aimed at reducing modelling errors. The coupled Wflow_sbm-LSTM model was applied to the Boyne catchment in Ireland, utilising a dataset spanning two decades, divided into training, validation, and testing sets to ensure robust model evaluation. Predictive performance was rigorously assessed using metrics like Modified Kling-Gupta Efficiency (MKGE) and Nash-Sutcliffe Efficiency (NSE), with observed streamflow discharges as the target variable. Results demonstrated that the coupled model outperformed the best-calibrated Wflow_sbm model in the study catchment based on the performance measures. The enhanced prediction of extreme events by the coupled Wflow_sbm-LSTM model strengthens the case for its integration into an operational river flow forecasting framework. Significantly, Wflow is endorsed by the National Flood Forecast Warning Service (NFFWS) in Ireland as a recommended model for streamflow simulations, specifically designed for fluvial flood forecasting. Consequently, our proposed Wflow_sbm-LSTM coupled model presents a compelling opportunity for integration into the NFFWS. With demonstrated potential to achieve precise streamflow estimations, this integration holds promise for significantly enhancing the accuracy and effectiveness of flood predictions in Ireland.

How to cite: Mohammed, S. and Nasr, A.: Advancing Streamflow Modelling: Bias Removal in Physically-Based Models with the Long Short-Term Memory Networks (LSTM) Algorithm, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-9573, https://doi.org/10.5194/egusphere-egu24-9573, 2024.

EGU24-10506 | Posters on site | HS3.4

Enhancing Hydrological Predictions: Feature-Driven Streamflow Forecasting with Sparse Autoencoder-based Long Short-Term Memory Networks 

Neha Vinod, Arathy Nair Geetha Raveendran, Adarsh Sankaran, and Anandu Kochukattil Ajith

In response to the critical demand for accurate streamflow predictions in hydrology, this study introduces a Sparse Autoencoder-based Long Short-Term Memory (SA-LSTM) framework applied to daily streamflow data from three-gauge stations within the Greater Pamba River Basin of Kerala, India, which was the worst affected region by the devastating floods of 2018. The SA-LSTM model addresses the challenge of feature selection from an extensive set of corresponding 1 to 7 days lagged climatic variables, such as precipitation, maximum and minimum temperatures, by incorporating a sparsity constraint. This constraint strategically guides the autoencoder to focus on the most influential features for the prediction analysis. The prediction process involves training the SA-LSTM model on historical streamflow data and climatic variables, allowing the model to learn intricate patterns and relationships. Furthermore, this study includes a comparative analysis featuring the Random Forest (RF)-LSTM model, where the RF model is employed for feature extraction, and a separate LSTM model is used for streamflow prediction. While the RF-LSTM combination demonstrates competitive performance, it is noteworthy that the SA-LSTM model consistently outperforms in terms of predictive accuracy. Rigorous evaluation metrics, including Correlation Coefficient (R2), Root Mean Square Error (RMSE), Mean Square Error (MSE), and Mean Absolute Error (MAE), highlight the SA-LSTM's forecasting accuracy across the three stations. Notably, the R2 values surpass 0.85, RMSE values remain under 12 cubic meters per second (m³/s), MSE values are below 70 (m³/s), and MAE values approach 8 m³/s. The detailed comparison between the above models underscores the superior capabilities of the SA-LSTM framework in capturing complex temporal patterns, emphasizing its potential for advancing hydrological modeling and flood risk management in flood-prone regions.

 

Key words : Streamflow, LSTM, Sparse Autoencoder, Flood, Greater Pamba

How to cite: Vinod, N., Geetha Raveendran, A. N., Sankaran, A., and Kochukattil Ajith, A.: Enhancing Hydrological Predictions: Feature-Driven Streamflow Forecasting with Sparse Autoencoder-based Long Short-Term Memory Networks, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-10506, https://doi.org/10.5194/egusphere-egu24-10506, 2024.

EGU24-11506 | ECS | Posters on site | HS3.4

Forecasting reservoir inflows with Long Short-Term Memory models 

Laura Soncin, Claudia Bertini, Schalk Jan van Andel, Elena Ridolfi, Francesco Napolitano, Fabio Russo, and Celia Ramos Sánchez

The increased variability of water resources and the escalating water consumption contribute to the risk of stress and water scarcity in reservoirs that are typically designed based on historical conditions. Therefore, it is relevant to provide accurate forecasts of reservoir inflow to optimize sustainable water management as conditions change, especially during extreme events, such as flooding and drought. However, accurate forecasting the inflow is not straightforward, due the uncertainty of the hydrological inputs and the strong non-linearity of the system. Numerous recent studies have employed approaches based on Machine Learning (ML) techniques, such as Artificial Neural Networks (ANN), Long Short-Term Memory (LSTM), and Random Forest (RF), with successful examples of providing skilful site-specific predictions. In particular, LSTM have emerged among the pool of ML models for their performance in simulating rainfall-runoff processes, thanks to their ability to learn long-term dependencies from time series. 
Here we propose an LSTM-based approach for inflow prediction in the Barrios de Luna reservoir, located in the Spanish part of the Douro River Basin. The reservoir has a dual role, as its water is used for irrigation during dry summer periods, and its storage volume is used to mitigate floods. Therefore, in order to operate the reservoir in the short-term, Barrios de Luna reservoir operators need accurate forecast to support water management decisions in the daily and weekly time horizons. In our work, we explore the potential of a LSTM model to predict inflow in the reservoir at varying lead times, ranging from 1 day up to 4 weeks. Initially, we use as inputs past inflow, precipitation and temperature observations, and then we include meteorological forecasts of precipitation and temperature from ECMWF Extended Range. For the latter experiments, different configurations of the LSTM are tested, i.e. training the model with observations and forecasts together and training the model with observations only and fine tune it with forecasts.
Our preliminary results show that precipitation, temperature and inflow observations are all crucial inputs to the LSTM for predicting inflow, and meteorological forecast inputs seem to improve performance for the longer lead-times of one week up to a month.
Predictions developed will contribute to the Douro case study of the CLImate INTelligence (CLINT) H2020 project.

How to cite: Soncin, L., Bertini, C., van Andel, S. J., Ridolfi, E., Napolitano, F., Russo, F., and Ramos Sánchez, C.: Forecasting reservoir inflows with Long Short-Term Memory models, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-11506, https://doi.org/10.5194/egusphere-egu24-11506, 2024.

EGU24-11768 | ECS | Posters on site | HS3.4

High-Efficiency Rainfall Data Compression Using Binarized Convolutional Autoencoder 

Manuel Traub, Fedor Scholz, Thomas Scholten, Christiane Zarfl, and Martin V. Butz

In the era of big data, managing and storing large-scale meteorological datasets is a critical challenge. We focus on high-resolution rainfall data, which is crucial to atmospheric sciences, climate research, and real-time weather forecasting. This study introduces a deep learning-based approach to compress the German Radar-Online-Aneichung (RADOLAN) rainfall dataset. We achieve a compression ratio of 200:1 while maintaining a minimal mean squared reconstruction error (MSE). Our method combines a convolutional autoencoder with a novel binarization mechanism, to compress data from a resolution of 900x900 pixels at 32-bit depth to 180x180 pixels at 4-bit depth. Leveraging the ConvNeXt architecture (Zhuang Liu, et al., 'A ConvNet for the 2020s'), our method learns a convolutional autoencoder for enhanced meteorological data compression. ConvNeXt introduces key architectural modifications, such as revised layer normalization and expanded receptive fields, taking inspiration from Vision Transformer to form a modern ConvNet. Our novel binarization mechanism, pivotal for achieving the high compression ratio, operates by dynamically quantizing the latent space representations using a novel magnitude specific noise injection technique. This quantization not only reduces the data size but also preserves crucial meteorological information as our low reconstruction MSE demonstrates. Beyond rainfall data, our approach shows promise for other types of high-resolution meteorological datasets, such as temperature, humidity, etc. Adapting our method to these modalities could further streamline the data management processes in meteorological deep learning scenarios and thus facilitate efficient storage and processing of diverse meteorological datasets.

How to cite: Traub, M., Scholz, F., Scholten, T., Zarfl, C., and Butz, M. V.: High-Efficiency Rainfall Data Compression Using Binarized Convolutional Autoencoder, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-11768, https://doi.org/10.5194/egusphere-egu24-11768, 2024.

Machine learning has extensively been applied to for flow forecasting in gauged basins. Increasingly, models generating forecasts in some basin(s) of interest are trained using data from beyond the study region. With increasingly large hydrological datasets, a new challenge emerges: given some region of interest, how do you select which basins to include among the training dataset?

There is currently little guidance on selecting data from outside the basin(s) under study. An intuitive approach might be to select data from neighbouring basins, or basins with similar hydrological characteristics. However, a growing body of research suggests that including hydrologically dissimilar basins can in fact produce greater improvements to model generalisation. In this study, we use clustering as a simple yet effective method for identifying temporal and spatial hydrological diversity within a large hydrological dataset. The clustering results are used to generate information-rich subsets of data, that are used for model training. We compare the effects that basin subsets, that represent various hydrological characteristics, have on model generalisation.
Our study shows that data within individual basins, and between hydrologically similar basins, contain high degrees of redundancy. In such cases, training data can be heavily undersampled with no adverse effects – or even moderate improvements to model performance. We also show that spatial hydrological diversity can hugely benefit model training, providing improved generalisation and a regularisation effect.

How to cite: Snieder, E. and Khan, U.: Towards improved spatio-temporal selection of training data for LSTM-based flow forecasting models in Canadian basins, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-12293, https://doi.org/10.5194/egusphere-egu24-12293, 2024.

EGU24-13353 | Posters on site | HS3.4

Flood Prediction Using Deep Neural Networks Across a Large and Complex River System  

Mostafa Saberian and Vidya Samadi

Accurately predicting streamflow poses a considerable challenge particularly for intense storm events occurring across complex river systems. To tackle this issue, we developed multiple deep neural network models including a Long Short-term Memory (LSTM) and Neural Hierarchical Interpolation for Time Series Forecasting (N-HiTS) to predict short duration (1-hour) flood hydrographs. LSTM excels in preserving prolonged dependencies in structured time series data, while N-HiTS introduces an innovative deep neural architecture characterized by backward and forward residual links and a deep stack of fully connected layers. In addition, N-HiTS employs a combination of multi-rate sampling and multi-scale synthesis of predictions, resulting in a hierarchical forecasting structure that reduces computational requirements and enhances accuracy. Our goal was to evaluate the robustness and effectiveness of these advanced algorithms by comparing them with the National Water Model (NWM) forecast, across a large and complex river system i.e., the Wateree River Basin in South Carolina, USA. The models were trained and tested using precipitation, temperature, humidity, and solar radiation data during the periods 01/01/2009 to 09/30/2022 and 10/1/2022 to 01/01/2024, respectively.  Analysis suggests that N-HiTS showcased state-of-the-art performance and enhanced hourly flood forecasting accuracy by approximately 10% compared to LSTM and NWM with a negligible difference in computational costs. N-HiTS was able to more accurately forecast time to peak and peak rate values of hourly flood hydrographs compared to the LSTM and NWM. Our extensive experiments revealed the importance of multi-rate input sampling and hierarchical interpolation approaches designed within the N-HiTS model that drastically improved the flood forecasting and interpretability of the predictions.

How to cite: Saberian, M. and Samadi, V.: Flood Prediction Using Deep Neural Networks Across a Large and Complex River System , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-13353, https://doi.org/10.5194/egusphere-egu24-13353, 2024.

We propose a hybrid deep learning model that combines long short-term memory networks (LSTMs) to capture both spatial and temporal dependencies in the river system. The LSTM component processes spatial information derived from topographical data and river network characteristics, allowing the model to understand the physical layout of the river basin. Simultaneously, the LSTM component exploits temporal patterns in historical dam release and rainfall data, enabling the model to discern the dynamics of flood propagation. In comparison of previous study, previous results accepted only hydrological models such as HECRAS, FLDWAV, FLUMEN. But, this study accept combination of HECRAS and Deep Learning algorithm, LSTM. The goal of this study is to predict the river highest level and travel time by dam release 3 to 6 hours in advance throughout the Seomjin river basin. In order to achieve, this study conducted hydrological modeling (HECRAS) and developed a deep learning algorithm (LSTM). Afterward, the developed model combining HECRAS and LSTM was verified at six flood alert stations. Finally, the models will provide the river highest level and travel time information up to 6 hours in advance at six flood alert stations. To train and validate the model, we compile a comprehensive dataset of historical dam release events and corresponding flood travel times from a range of river basins. The dataset includes various hydrological and meteorological features to ensure the model's robustness in handling diverse scenarios. The deep learning model is then trained using a subset of the data and validated against unseen events to assess its generalization capabilities. Preliminary results indicate that the hybrid HECRAS-LSTM model outperforms traditional hydrological models in predicting flood travel times. The model exhibits improved accuracy, particularly in cases of complex river geometries and extreme weather events. Additionally, the model demonstrates its potential for real-time forecasting, as it can efficiently process and assimilate incoming data. In conclusion, our study showcases the effectiveness of using a hybrid HECRAS-LSTM model for forecasting flood travel time by dam release. By leveraging the power of deep learning, we pave the way for more precise and reliable flood predictions, contributing to the overall resilience and safety of communities located downstream of dam-controlled river systems.

How to cite: Kang, J., Lee, G., Park, S., Jung, C., and Yu, J.: The Development of Forecasting System Flood Travel Time by Dam Release for Supplying Flood Information Using Deep Learning at Flood Alert Stations in the Seomjin River Basin, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-13848, https://doi.org/10.5194/egusphere-egu24-13848, 2024.

EGU24-14765 | ECS | Posters virtual | HS3.4

Optimizing Groundwater Forecasting: Comparative Analysis of MLP Models Using Global and Regional Precipitation Data 

Akanksha Soni, Surajit Deb Barma, and Amai Mahesha

This study investigates the efficacy of Multi-Layer Perceptron (MLP) models in groundwater level modeling, specifically emphasizing the pivotal role of input data quality, particularly precipitation data. Unlike prior research that primarily focused on regional datasets like those from the India Meteorological Department (IMD), our research explores the integration of global precipitation data, specifically leveraging the Integrated Multi-satellitE Retrievals for Global Precipitation Measurement (IMERG) dataset for MLP-based modeling. The assessment was conducted using two wells in Dakshina Kannada, evaluating four MLP models (GA-MLP, EFO-MLP, PSO-MLP, AAEO-MLP) with IMERG and IMD precipitation data. Performance metrics were employed, including mean absolute error, root mean square error, normalized Nash-Sutcliffe efficiency, and Pearson's correlation index. The study also includes convergence analysis and stability assessments, revealing the significant impact of the precipitation dataset on model performance. Noteworthy findings include the superior performance of the AAEO-MLP model in training with IMD data and the GA-MLP model's outperformance in testing at the Bajpe well with both datasets. The stability of the GA-MLP model, indicated by the lowest standard deviation values in convergence analysis, underscores its reliability. Moreover, transitioning to the IMERG dataset improved model performance and reduced variability, providing valuable insights into the strengths and limitations of MLP models in groundwater-level modeling. These results advance the precision and dependability of groundwater level forecasts, thereby supporting more effective strategies for international groundwater resource management.

How to cite: Soni, A., Barma, S. D., and Mahesha, A.: Optimizing Groundwater Forecasting: Comparative Analysis of MLP Models Using Global and Regional Precipitation Data, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-14765, https://doi.org/10.5194/egusphere-egu24-14765, 2024.

EGU24-15248 | ECS | Orals | HS3.4

Estimation of Small Stream Water Surface Elevation Using UAV Photogrammetry and Deep Learning 

Radosław Szostak, Mirosław Zimnoch, Przemysław Wachniew, Marcin Pietroń, and Paweł Ćwiąkała

Unmanned aerial vehicle (UAV) photogrammetry allows the generation of orthophoto and digital surface model (DSM) rasters of a terrain. However, DSMs of water bodies mapped using this technique often reveal distortions in the water surface, thereby impeding the accurate sampling of water surface elevation (WSE) from DSMs. This study investigates the capability of deep neural networks to accommodate the aforementioned perturbations and effectively estimate WSE from photogrammetric rasters. Convolutional neural networks (CNNs) were employed for this purpose. Three regression approaches utilizing CNNs were explored: i) direct regression employing an encoder, ii) prediction of the weight mask using an encoder-decoder architecture, subsequently used to sample values from the photogrammetric DSM, and iii) a solution based on the fusion of the two approaches. The dataset employed in this study comprises data collected from five case studies of small lowland streams in Poland and Denmark, consisting of 322 DSM and orthophoto raster samples. Each sample corresponds to a 10 by 10 meter area of the stream channel and adjacent land. A grid search was employed to identify the optimal combination of encoder, mask generation architecture, and batch size among multiple candidates. Solutions were evaluated using two cross-validation methods: stratified k-fold cross-validation, where validation subsets maintained the same proportion of samples from all case studies, and leave-one-case-out cross-validation, where the validation dataset originates entirely from a single case study, and the training set consists of samples from other case studies. The proposed solution was compared with existing methods for measuring water levels in small streams using a drone. The results indicate that the solution outperforms previous photogrammetry-based methods and is second only to the radar-based method, which is considered the most accurate method available.

This research was funded by National Science Centre, Poland, project WATERLINE (2020/02/Y/ST10/00065), under the CHISTERA IV programme of the EU Horizon 2020 (Grant no 857925).

How to cite: Szostak, R., Zimnoch, M., Wachniew, P., Pietroń, M., and Ćwiąkała, P.: Estimation of Small Stream Water Surface Elevation Using UAV Photogrammetry and Deep Learning, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-15248, https://doi.org/10.5194/egusphere-egu24-15248, 2024.

EGU24-16234 | ECS | Posters on site | HS3.4

A bottom-up approach to identify important hydrological processes by evaluating a national scale EA-LSTM model for Denmark 

Grith Martinsen, Niels Agertoft, and Phillip Aarestrup

The utilization of data-driven models in hydrology has witnessed a significant increase in recent years. The open-source philosophy underpinning much of the code developed and research being conducted has facilitated widespread access to the the hydrological community to sophisticated machine learning models and technology (Reichstein 2019). These data driven approaches to hydrological modelling has witnessed growing interest after multiple studies has shown how machine-learning models were able to outperform nationwide traditional physics-based hydrological models (Kratzerts et al. 2019). The latter often demands substantial man-hours for development, calibration and fine-tuning to accurately represent relevant hydrological processes.

In this national-scale explorative study we undertake an in-depth examination of Danish catchment hydrology. Our objective is to understand what processes and dynamics are well captured by a purely data driven model without physical constraints, namely the Entity-Aware Long Short-Term Model (EA-LSTM). The model code was developed by Kratzerts et al. (2019) and the analysis build on top of a newly published national CAMELS data set covering 301 catchments in Denmark (Koch and Schneider, 2022), with an average resolution of 130 km2.

Denmark, spanning an area of around 43 000 km2, demonstrates a relatively high data coverage. Presently more than 400 stations record water level measurements in the Danish stream network, while a network of 243 stations have collected meteorological data since 2011. These datasets maintained by the Danish Environmental Protection Agency and the Danish Meteorological Institute, respectively, and are publicly available.

Despite Denmark’s data abundance, Koch and Schneider (2022) demonstrated that the data-driven EA-LSTM model, trained with the CAMELS dataset for Denmark (from now on referred to as the DK-LSTM) were not able to outperform the traditional physics-based hydrological model, against which it was benchmarked. Consequently, performance of the DK-LSTM model could be increased by pre-training it with simulations from a national physics-based model indicating that dominating hydrological processes are not described by the readily available input data in the CAMELS dataset.

This study conducts a comprehensive analysis of Danish catchment hydrology aiming to explore three aspects: 1) the common characteristics of the catchments where the DK-LSTM performs well or encounters challenges, 2) the identification of hydrological characteristics, that exhibit improvement when informing the data-driven model with physics-based model simulations, and 3) an exploration of whether the aforementioned findings can guide us in determining necessary physical constraints and/or input variables that explains the hydrological processes for the data-driven model approach at a national scale, using the example of DK-LSTM.

 

Koch, J., and Schneider, R. Long short-term memory networks enhance rainfall-runoff modelling at the national scale of Denmark. GEUS Bulletin49. https://doi.org/10.34194/geusb.v49.8292, 2022.

Kratzert, F., Klotz, D., Shalev, G., Klambauer, G., Hochreiter, S., and Nearing, G.: Towards learning universal, regional, and local hydrological behaviors via machine learning applied to large-sample datasets, Hydrol. Earth Syst. Sci., 23, 5089-5110, https://doi.org/10.5194/hess-23-5089-2019, 2019.

Reichstein, M., Camps-Valls, G., Stevens, B. et al. Deep learning and process understanding for data-driven Earth system science. Nature 566, 195–204. https://doi.org/10.1038/s41586-019-0912-1, 2019.

How to cite: Martinsen, G., Agertoft, N., and Aarestrup, P.: A bottom-up approach to identify important hydrological processes by evaluating a national scale EA-LSTM model for Denmark, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-16234, https://doi.org/10.5194/egusphere-egu24-16234, 2024.

EGU24-16474 | ECS | Orals | HS3.4

Short- and mid-term discharge forecasts combining machine learning and data assimilation for operational purpose 

Bob E Saint Fleur, Eric Gaume, Michaël Savary, Nicolas Akil, and Dominique Theriez

In recent years, machine learning models, particularly Long Short-Term Memory (LSTM), have proven to be effective alternatives for rainfall-runoff modeling, surpassing traditional hydrological modeling approaches 1. These models have predominantly been implemented and evaluated for rainfall-runoff simulations. However, operational hydrology often requires short- and mid-term forecasts. To be effective, such forecasts must consider past observed values of the predicted variables, requiring a data assimilation procedure 2,3,4. This presentation will evaluate several approaches based on the combination of open-source machine learning tools and data assimilation strategies for short- and mid-term discharge forecasting of flood and/or drought events. The evaluation is based on the rich and well-documented CAMELS dataset 5,6,7. The tested approaches include: (1) coupling pre-trained LSTMs on the CAMELS database with a Multilayer Perceptron (MLP) for prediction error corrections, (2) direct discharge MLP forecasting models specific for each lead time, including past observed discharges as input variables, and (3) option 2, including the LSTM-predicted discharges as input variables. In the absence of historical archives of weather forecasts (rainfall, temperatures, etc.), the different forecasting approaches will be tested in two configurations: (1) weather forecasts assumed to be perfect (using observed meteorological variables over the forecast horizon in place of predicted variables or ensembles) and (2) use of ensembles reflecting climatological variability over the forecast horizons for meteorological variables ensembles made up of time series randomly selected from the past. The forecast horizons considered range from 1 to 10 days, and the results are analyzed in light of the time of concentration of the watersheds.

 

References

1. Kratzert F, Klotz D, Brenner C, Schulz K, Herrnegger M. Rainfall–runoff modelling using Long Short-Term Memory (LSTM) networks. Hydrol Earth Syst Sci. 2018;22(11):6005-6022. doi:10.5194/hess-22-6005-2018

2. Bourgin F, Ramos MH, Thirel G, Andréassian V. Investigating the interactions between data assimilation and post-processing in hydrological ensemble forecasting. J Hydrol (Amst). 2014;519:2775-2784. doi:10.1016/j.jhydrol.2014.07.054

3. Boucher M ‐A., Quilty J, Adamowski J. Data Assimilation for Streamflow Forecasting Using Extreme Learning Machines and Multilayer Perceptrons. Water Resour Res. 2020;56(6). doi:10.1029/2019WR026226

4. Piazzi G, Thirel G, Perrin C, Delaigue O. Sequential Data Assimilation for Streamflow Forecasting: Assessing the Sensitivity to Uncertainties and Updated Variables of a Conceptual Hydrological Model at Basin Scale. Water Resour Res. 2021;57(4). doi:10.1029/2020WR028390

5. Newman AJ, Clark MP, Sampson K, et al. Development of a large-sample watershed-scale hydrometeorological data set for the contiguous USA: data set characteristics and assessment of regional variability in hydrologic model performance. Hydrol Earth Syst Sci. 2015;19(1):209-223. doi:10.5194/hess-19-209-2015

6. Kratzert, F. (2019). Pretrained models + simulations for our HESSD submission "Towards learning universal, regional, and local hydrological behaviors via machine learning applied to large-sample datasets", HydroShare, https://doi.org/10.4211/hs.83ea5312635e44dc824eeb99eda12f06

7. Kratzert, F. (2019). CAMELS Extended Maurer Forcing Data, HydroShare, https://doi.org/10.4211/hs.17c896843cf940339c3c3496d0c1c077

How to cite: Saint Fleur, B. E., Gaume, E., Savary, M., Akil, N., and Theriez, D.: Short- and mid-term discharge forecasts combining machine learning and data assimilation for operational purpose, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-16474, https://doi.org/10.5194/egusphere-egu24-16474, 2024.

EGU24-17502 | ECS | Posters on site | HS3.4

Towards improved Water Quality Modelling using Neural ODE models 

Marvin Höge, Florian Wenk, Andreas Scheidegger, Carlo Albert, and Andreas Frömelt

Neural Ordinary Differential Equations (ODEs) fuse neural networks with a mechanistic equation framework. This hybrid structure offers both traceability of model states and processes, as it is typical for physics-based models, and the ability of machine learning to encode new functional relations. Neural ODE models have demonstrated high potential in hydrologic predictions and scientific investigation of the related process in the hydrologic cycle, i.e. tasks of water quantity estimation (Höge et al., 2022).

This explicit representation of state variables is key to water quality modelling. There, we typically have several interrelated state variables like nitrate, nitrite, phosphorous, organic matter,…  Traditionally, these states are modelled based on mechanistic kinetic rate expressions that are often only rough approximations of the underlying dynamics. At the same time, this domain of water research suffers from data scarcity and therefore solely data-driven methods struggle to provide accurate predictions reliably. We show how to improve predictions of state dynamics and to foster knowledge gain about the processes in such interrelated systems with multiple states using Neural ODEs. 

Höge, M., Scheidegger, A., Baity-Jesi, M., Albert, C., & Fenicia, F.: Improving hydrologic models for predictions and process understanding using Neural ODEs. Hydrol. Earth Syst. Sci., 26, 5085-5102, https://hess.copernicus.org/articles/26/5085/2022/

How to cite: Höge, M., Wenk, F., Scheidegger, A., Albert, C., and Frömelt, A.: Towards improved Water Quality Modelling using Neural ODE models, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-17502, https://doi.org/10.5194/egusphere-egu24-17502, 2024.

EGU24-17543 | Orals | HS3.4

Deep-learning-based prediction of damages related to surface water floods for impact-based warning 

Pascal Horton, Markus Mosimann, Severin Kaderli, Olivia Martius, Andreas Paul Zischg, and Daniel Steinfeld

Surface water floods are responsible for a substantial amount of damage to buildings, yet they have received less attention than fluvial floods. Nowadays, both research and insurance companies are increasingly focusing on these phenomena to enhance knowledge and prevention efforts. This study builds upon pluvial-related damage data provided by the Swiss Mobiliar Insurance Company and the Building Insurance of Canton Zurich (GVZ) with the goal of developing a data-driven model for predicting potential damages in future precipitation events.

This work is a continuation of a previous method applied to Swiss data, relying on thresholds based on the quantiles of precipitation intensity and event volume, which, however, resulted in an excessive number of false alarms. First, a logistic regression has been assessed using different characteristics of the precipitation event. Subsequently, a random forest was established, incorporating terrain attributes to better characterize local conditions. Finally, a deep learning model was developed to account for the spatio-temporal properties of the precipitation fields on a domain larger than the targeted 1 km cell. The deep learning model comprises a convolutional neural network (CNN) for 4D precipitation data and subsequent dense layers, incorporating static attributes. The model has been applied to predict the probability of damage occurrence, as well as the damage degree quantified by the number of claims relative to the number of insured buildings.

How to cite: Horton, P., Mosimann, M., Kaderli, S., Martius, O., Zischg, A. P., and Steinfeld, D.: Deep-learning-based prediction of damages related to surface water floods for impact-based warning, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-17543, https://doi.org/10.5194/egusphere-egu24-17543, 2024.

EGU24-18073 | ECS | Orals | HS3.4

Operational stream water temperature forecasting with a temporal fusion transformer model 

Ryan S. Padrón, Massimiliano Zappa, and Konrad Bogner

Stream water temperatures influence aquatic biodiversity, agriculture, tourism, electricity production, and water quality. Therefore, stakeholders would benefit from an operational forecasting service that would support timely action. Deep Learning methods are well-suited for this task as they can provide probabilistic forecasts at individual stations of a monitoring network. Here we train and evaluate several state-of-the-art models using 10 years of data from 55 stations across Switzerland. Static features (e.g. station coordinates, catchment mean elevation, area, and glacierized fraction), time indices, meteorological and/or hydrological observations from the past 64 days, and their ensemble forecasts for the following 32 days are included as predictors in the models to estimate daily maximum water temperature for the next 32 days. We find that the Temporal Fusion Transformer (TFT) model performs best for all lead times with a cumulative rank probability score (CRPS) of 0.73 ºC averaged over all stations, lead times and 90 forecasts distributed over 1 full year. The TFT is followed by the Recurrent Neural Network (CRPS = 0.77 ºC), Neural Hierarchical Interpolation for Time Series (CRPS = 0.80 ºC), and Multi-layer Perceptron (CRPS = 0.85 ºC). All models outperform the benchmark ARX model. When factoring out the uncertainty stemming from the meteorological ensemble forecasts by using observations instead, the TFT improves to a CRPS of 0.43 ºC, and it remains the best of all models. In addition, the TFT model identifies air temperature and time of the year as the most relevant predictors. Furthermore, its attention feature suggests a dominant response to more recent information in the summer, and to information from the previous month during spring and autumn. Currently, daily maximum water temperature probabilistic forecasts are produced twice per week and made available at https://drought.ch/de/allgemeine-lage/wassertemperatur/fliessgewaesser-1-1.html. 

How to cite: Padrón, R. S., Zappa, M., and Bogner, K.: Operational stream water temperature forecasting with a temporal fusion transformer model, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-18073, https://doi.org/10.5194/egusphere-egu24-18073, 2024.

EGU24-18154 | ECS | Orals | HS3.4

Can Blended Model Improve Streamflow Simulation In Diverse Catchments ? 

Daneti Arun Sourya and Maheswaran Rathinasamy

Streamflow simulation or rainfall-runoff modelling has been a topic of research for the past few decades which has resulted in a plethora of modelling approaches ranging from physics models to empirical or data driven approaches. There are many physics-based (PB) models available to estimate streamflow, but still there exists uncertainty in model outputs due to incomplete representations of physical processes. Further, with advancements in machine learning (ML) concepts, there have been several attempts but with no/little physical consistency. As a result, models based on ML algorithms may be unreliable if applied to provide future hydroclimate projections where climates and land use patterns are outside the range of training data. 

Here we test blended models built by combining PB model state variables (specifically soil moisture) with ML algorithms on their ability to simulate streamflow in 671 catchments representing diverse conditions across the conterminous United States.

For this purpose, we develop a suite of blended hydrological models by pairing different PB models (Catchment Wetness Index, Catchment Moisture Deficit, GR4J, Australian Water Balance, Single-bucket Soil Moisture Accounting, and Sacramento Soil Moisture Accounting models) with different ML methods such as Long Short Term Memory network (LSTM), eXtreme Gradient Boosting (XGB).

The results indicate that the blended models provide significant improvement in catchments where PB models are underperforming. Furthermore, the accuracy of streamflow estimation is improved in catchments where the ML models failed to estimate streamflow accurately.

How to cite: Sourya, D. A. and Rathinasamy, M.: Can Blended Model Improve Streamflow Simulation In Diverse Catchments ?, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-18154, https://doi.org/10.5194/egusphere-egu24-18154, 2024.

EGU24-18762 | Orals | HS3.4

Benchmarking hydrological models for national scale climate impact assessment 

Elizabeth Lewis, Ben Smith, Stephen Birkinshaw, Helen He, and David Pritchard

National scale hydrological models are required for many types of water sector applications, for example water resources planning. Existing UK national-scale model frameworks are based on conceptual numerical schemes, with an emerging trend towards incorporating deep learning models. Existing literature has shown that groundwater/surface water interactions are key for accurately representing future flows, and these processes are most accurately represented with physically-based hydrological models.

In response to this, our study undertakes a comparative analysis of three national model frameworks (Neural Hydrology, HBV, SHETRAN) to investigate the necessity for physically-based hydrological modelling. The models were run with the full ensemble of bias-corrected UKCP18 12km RCM data which enabled a direct comparison of future flow projections. We show that whilst many national frameworks perform well for the historical period, physically-based models can give substantially different projections of future flows, particularly low flows. Moreover, our study illustrates that the physically-based model exhibits a consistent trajectory in Budyko space between the baseline and future simulations, a characteristic not shared by conceptual and deep learning models. To provide context for these results, we incorporate insights from other national model frameworks, including the eFlag project.

How to cite: Lewis, E., Smith, B., Birkinshaw, S., He, H., and Pritchard, D.: Benchmarking hydrological models for national scale climate impact assessment, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-18762, https://doi.org/10.5194/egusphere-egu24-18762, 2024.

EGU24-20636 | ECS | Orals | HS3.4

Can Attention Models Surpass LSTM in Hydrology? 

Jiangtao Liu, Chaopeng Shen, and Tadd Bindas

Accurate modeling of various hydrological variables is important for water resource management, flood forecasting, and pest control. Deep learning models, especially Long Short-Term Memory (LSTM) models based on Recurrent Neural Network (RNN) structures, have shown significant success in simulating streamflow, soil moisture, and model parameter assessment. With the development of large language models (LLMs) based on attention mechanisms, such as ChatGPT and Bard, we have observed significant advancements in fields like natural language processing (NLP), computer vision (CV), and time series prediction. Despite achieving advancements across various domains, the application of attention-based models in hydrology remains relatively limited, with LSTM models maintaining a dominant position in this field. This study evaluates the performance of 18 state-of-the-art attention-based models and their variants in hydrology. We focus on their performance in streamflow, soil moisture, snowmelt, and dissolved oxygen (DO) datasets, comparing them to LSTM models in both long-term and short-term regression and forecasting. We also examine these models' performance in spatial cross-validation. Our findings indicate that while LSTM models maintain strong competitiveness in various hydrological datasets, Attention models offer potential advantages in specific metrics and time lengths, providing valuable insights into applying attention-based models in hydrology. Finally, we discuss the potential applications of foundation models and how these methods can contribute to the sustainable use of water resources and the challenges of climate change.

How to cite: Liu, J., Shen, C., and Bindas, T.: Can Attention Models Surpass LSTM in Hydrology?, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-20636, https://doi.org/10.5194/egusphere-egu24-20636, 2024.

EGU24-20907 | Orals | HS3.4

Revolutionizing Flood Forecasting with a Generalized Deep Learning Model 

Julian Hofmann and Adrian Holt

The domain of spatial flood prediction is dominated by hydrodynamic models, which, while robust and adaptable, are often constrained by computational requirements and slow processing times. To address these limitations, the integration of Deep Learning (DL) models has emerged as a promising solution, offering the potential for rapid prediction capabilities, while maintaining a high output quality. However, a critical challenge with DL models lies in their requirement for retraining for each new domain area, based on the outputs of hydrodynamic simulations generated for that specific region. This need for domain-specific retraining hampers the scalability and quick deployment of DL models in diverse settings. Our research focuses on bridging this gap by developing a fully generalized DL model for flood prediction.

FloodWaive's approach pivots on creating a DL model that can predict flood events rapidly and accurately across various regions without requiring retraining for each new domain area. The model is trained on a rich dataset derived from numerous hydrodynamic simulations, encompassing a wide spectrum of topographical conditions. This training is designed to enable the model to generalize its predictive capabilities across different domains and weather patterns, thus overcoming the traditional limitation of DL models in this field.

Initial findings from the development phase are promising, showcasing the model's capability to process complex data and provide quick, accurate flood predictions. The success of this fully generalized DL modeling approach could revolutionize applications of flood predictions such as flood forecasting and risk analysis. Regarding the later, real-time evaluation of flood protection measures could become a reality. This would empower urban planners, emergency response teams, and environmental agencies with the ability to make informed decisions quickly, potentially saving lives and reducing economic losses.

While this project is still in its developmental stages, the preliminary results point towards a significant leap in flood forecasting technology. The ultimate goal is to offer a universally deployable, real-time flood prediction tool, significantly enhancing our ability to mitigate the impact of floods worldwide.

  

How to cite: Hofmann, J. and Holt, A.: Revolutionizing Flood Forecasting with a Generalized Deep Learning Model, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-20907, https://doi.org/10.5194/egusphere-egu24-20907, 2024.

One of the latent difficulties in the fields of climatology, meteorology, and hydrology is the scarce rainfall information available due to the limited or nonexistent instrumentation of river basins, especially in developing countries where the establishment and maintenance of equipment entail high costs relative to the available budget. Hence, the importance of generating alternatives that seek to improve spatial precipitation estimation has been increasing, given the advances in the implementation of computational algorithms that involve Machine Learning techniques. In this study, a multitask convolutional neural network was implemented, composed of an encoder-decoder architecture (U-Net), which simultaneously estimates the probability of rain through a classification model and the precipitation rate through a regression model at a spatial resolution of 2 km2 and a temporal resolution of 10 minutes. The input modalities included data from rain gauge stations, weather radar, and satellite information (GOES 16). For model training,  validation, and testing, a dataset was consolidated with 3 months of information (February to April 2021) with a distribution of 70/15/15 percent, covering the effective coverage range of the Munchique weather radar located in the Andean region of Colombia. The obtained results show a Probability of Detection (POD) of 0.59 and a False Alarm Rate (FAR) of 0.39. Regarding precipitation rate estimation, it is assessed with a Root Mean  Square Error (RMSE) of 1.13 mm/10min. This research highlights the significant capability of deep learning algorithms in reconstructing and reproducing the spatial pattern of rainfall in tropical regions with limited instrumentation. However, there is a need to continue strengthening climatological monitoring networks to achieve significant spatial representativeness, thereby reducing potential biases in model estimations. 

How to cite: Barrios, M., Rubiano, H., and Guevara-Ochoa, C.: Implementation of deep learning algorithms in the sub-hourly rainfall fields estimation from remote sensors and rainfall gauge information in the tropical Andes, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-21431, https://doi.org/10.5194/egusphere-egu24-21431, 2024.

EGU24-196 | ECS | Posters on site | GM3.2

Remote sensing and geomorphometry application in riverscapes evolution in the south-eastern Arabian Peninsula (Sultanate of Oman) 

Andrea Pezzotta, Alessia Marinoni, Mohammed Al Kindi, Michele Zucali, and Andrea Zerboni

Riverscapes in arid and semi-arid environments serve as crucial archives, enabling us to understand the landscape evolution and the active and fossil geomorphological processes that shape the Earth's surface. Such environmental contexts are generally wide, and these settings are routinely investigated with remote sensing tools. We selected two distinct study areas from the south-eastern margin of the Arabian Peninsula (Sultanate of Oman) to detect climate and tectonic imprints over landform development: 1) Jebel Akhdar (JAK), and its surrounding areas, located in the Al-Hajar Mountains (to the North), is a wide anticline formed by the Late Cretaceous obduction of the Semail Ophiolite and the associated time-equivalent tectonics, followed by the Cenozoic tectonic events; and 2) Jebel Qara (JQA), situated in the Dhofar Mountains (to the South), is placed along the Gulf of Aden transform margin, featuring transtensional faults giving rise to stepped escarpments and grabens. The extant landscapes of both regions are characterized by a network of narrow and deep canyons that incised limestone massifs, while the surrounding plain areas show the development of important alluvial fan systems.

The application of remote sensing is essential for investigating the development of fluvial systems at a regional scale, combined with field survey to validate specific sites of interest, thereby understanding the geomorphological evolution at various scales. Specifically, remote sensing techniques include the processing of satellite imagery and the comparison with the available historical imagery and maps to detect changes in geomorphic processes. Remote sensing and field survey allow the recognizing of different geomorphological features; the dominant ones are represented by elements and landforms related to structural setting, fluvial activity, and karst processes. The associations of the abovementioned landforms make it possible to assess the structural influence on drainage and karstic network development. Data collected from remote sensing implements the geomorphometric quantification of geomorphological processes, mostly considering changes in topography and river network analyses. The most meaningful morphometric indices applied (such as drainage divide stability, normalized steepness index, knickpoint detection, and swath profiles…) suggest their values strongly vary along faults in JAK, highlighted even with the alignment of knickpoints; while, in JQA, values show little changes in correspondence of faults and knickpoints are controlled both by karst and structural settings. In this way, the combination of remote sensing and morphometrical analyses permits to quantify the central role of litho-structural influence on the development of riverscapes in the south-eastern Arabian Peninsula. This approach facilitates the identification of the primary geomorphological processes that have shaped the landscape in arid and semi-arid contexts of the Sultanate of Oman, making it a versatile method that can be applied to understand the riverscapes evolution processes in analogous regions.

How to cite: Pezzotta, A., Marinoni, A., Al Kindi, M., Zucali, M., and Zerboni, A.: Remote sensing and geomorphometry application in riverscapes evolution in the south-eastern Arabian Peninsula (Sultanate of Oman), EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-196, https://doi.org/10.5194/egusphere-egu24-196, 2024.

EGU24-1613 | ECS | Posters on site | GM3.2

Simulating 4D scenes of rockfall and landslide activity for improved 3D point cloud-based change detection using machine learning 

Ronald Tabernig, Vivien Zahs, Hannah Weiser, and Bernhard Höfle

Terrestrial Laser Scanning (TLS) systems have been refined to automatically and continuously scan defined areas with high temporal resolution (sub-hourly), leading to the development of Permanent Laser Scanning (PLS). This temporal resolution requires the development of new methods for efficient extraction of change information. The creation of labeled 4D point clouds (3D+time), classified by surface change type, remains time-consuming. This hinders the evaluation of change detection methods and the training of machine learning (ML) and deep learning (DL) models.

This study explores how synthetic 4D point clouds can be effectively utilized for detecting and classifying spatiotemporal changes. We combine simplified process path simulations, simulated PLS, and change detection methods (e.g. M3C2) [1]. This combination is used to automatically evaluate calculated distances compared to a pre-defined reference. It also generates labeled 4D training datasets for ML/DL approaches.

We adapted the Gravitational Process Path model (GPP) [2] to create gravity-influenced process paths for our PLS simulations. Utilizing these paths, we simulate two different scenarios, 1) including a forest situated on top of a large landslide and 2) an outcrop with rockfall activity. For the forest scenario, a constant velocity is applied to each tree to simulate slope movement. The velocity of the objects in the rockfall scene is determined by the GPP model. Dynamic 3D scenes are generated from these scenarios and used as input for Virtual Laser Scanning (VLS). Realistic simulation of LiDAR surveys (of these virtual scenes) is achieved by using the open-source simulator HELIOS++ [3]. This workflow allows for the determination of the accurate position of each object at any given time. It provides reference data that is usually unavailable in real data acquisitions. In the rockfall scenario, M3C2 distances are calculated, and areas of similar change are clustered. For the forest located on the landslide, 2D and 3D displacement vectors are derived from the displacement of the tree trunks. These changes are then compared to the actual change occurring between epochs. Furthermore, the time steps between each epoch can be chosen arbitrarily, enabling the exploration of various scenarios and processes using labeled point clouds at any temporal resolution.

Preliminary results suggest that this workflow can assist in determining the scan resolution required to detect changes of a specific size and magnitude. We establish a simulation-based error margin for each method used by comparing the results to the reference data. This enables direct evaluation of method performance during implementation.

We demonstrate the potential of combining process simulation and laser scanning simulation for resource efficient planning of TLS and PLS campaigns, geographically sound generation of dynamic point clouds, the evaluation of change detection and quantification methods, and generating labeled point clouds as training data for 4D ML/DL methods. 

References:
[1] py4dgeo: https://github.com/3dgeo-heidelberg/py4dgeo
[2] Wichmann, V. (2017): https://doi.org/10.5194/gmd-10-3309-2017.
[3] Winiwarter, L. et al. (2022): https://doi.org/10.1016/j.rse.2021.112772. 

How to cite: Tabernig, R., Zahs, V., Weiser, H., and Höfle, B.: Simulating 4D scenes of rockfall and landslide activity for improved 3D point cloud-based change detection using machine learning, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-1613, https://doi.org/10.5194/egusphere-egu24-1613, 2024.

EGU24-4207 | Orals | GM3.2

Incorporating ontological characteristics for global landform classification based on 30 meters DEM 

Xin Yang, Chenghu Zhou, Sijin Li, Junfei Ma, Yang Chen, Xingyu Zhou, Fayuan Li, Liyang Xiong, Guoan Tang, and Michael Meadows

Landform classification and mapping provide fundamental data for Earth science research, natural resource management, environmental monitoring, urban planning, and various other domains. Despite the availability of DEMs with 1-arc second resolution, global-scale studies on landform classification and mapping are inconsistent in terms of general classification systems and methods.

Landforms represent not only assemblages of morphological characteristics but also encompass the human understanding of the Earth, which is constrained by the nature and scale of quantitative analysis. Here, we propose a novel framework for global landform mapping to significantly improve the quantitative evaluation of geomorphological features.

The proposed framework incorporates geomorphological ontology that takes account of their conceptualization to construct classified objects. We propose the accumulated slope (AS) and mountain uplift index (MUI) to emphasize the integrity and continuity of geomorphological units, providing more precise results compared to traditional methods. Aggregating local terrain features into global metrics, AS effectively overcomes the potential negative influence of increased resolution on landform integrity. MUI aligns better with human perception of mountainous morphology and surpasses the limitations of window-based computing.

In presenting the new framework, we have developed and made available a public dataset, Global Basic Landform Unit (GBLU), which incorporates a comprehensive set of objects that constitute the range of landforms on Earth. In emphasizing the integration of classification with quantitative analysis, GBLU highlights the connection between natural objects and human understanding in geomorphology and the Earth sciences. The GBLU outperforms previous datasets (the basic landform classification and global mountain assessment) in expressing landform details. GBLU can be downloaded at https://geomorph.deep-time.org. It serves as a valuable resource in facilitating a deeper understanding of landform spatial distribution and evolution, and supporting research in a diverse range of fields.

How to cite: Yang, X., Zhou, C., Li, S., Ma, J., Chen, Y., Zhou, X., Li, F., Xiong, L., Tang, G., and Meadows, M.: Incorporating ontological characteristics for global landform classification based on 30 meters DEM, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-4207, https://doi.org/10.5194/egusphere-egu24-4207, 2024.

The influence of temperature as a key factor in slope stability, particularly in temperate regions, remains insufficiently explored. This study investigates the thermo-hydro-mechanical (THM) response of expansive soils, focusing on the thermally-induced activity in clay landslides.

Establishing a representative thermal variable for broad-scale assessments poses challenges due to material heterogeneities and the intricate nature of THM processes. Our research employs landslide spatial modelling in Italy, concentrating on clay-rich areas with shallow landslides on gentle slopes. Utilizing geo-lithological and geological maps and the Italian National Inventory (IFFI), we apply a Generalized Additive Model (GAM) based on slope units to capture nonlinearities in the temperature-shear strength relationship. A decade-long dataset of Land Surface Temperature (LST) from MODIS, accessible in Google Earth Engine, serves as a key input.

The study produces spatial probability maps for clay deposits across Italy, revealing a positive correlation between landslide occurrence and LST on warmer, gentle slopes, especially in Southern Italy. This aligns with the observation that higher temperatures reduce soil and water viscosity, amplifying shear creep rates in clay-rich materials. By elucidating the temperature-slope stability relationship, this study contributes to understanding landslide dynamics in temperate climates, facilitating the development of effective risk recognition strategies.

How to cite: Loche, M. and Scaringi, G.: Exploring Temperature-Shear Strength Dynamics: A Spatial Modelling Approach for Clay Landslide Susceptibility in Italy, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-5190, https://doi.org/10.5194/egusphere-egu24-5190, 2024.

EGU24-6250 | ECS | Orals | GM3.2

Three-Dimensional Stress Analysis of Mountain Ranges: A Novel Approach Using Marching Volume Polytopes Algorithm and Finite Cell Method  

Viktor Haunsperger, Jörg Robl, Andreas Schröder, and Stefan Hergarten

The negative feedback between relief formation due to valley incision, increasing topographic stress towards a critical stress state dependent on rock strength, and consequently relief-destroying (and stress-reducing) landslides determines the geometry of alpine landscapes. Hence, the computation of topographic stresses for entire mountain massifs is crucial to identify potential landslide hotspots at steep landforms close to rock failure, determining the maximum strength of rocks and rock sequences at the mountain scale, and explaining contrasting geometries of alpine landscapes in dependence on the prevailing rock types. Traditional 2D stress and displacement calculations on valley cross-sections tend to oversimplify the complicated stress pattern, particularly where valleys converge or around ridges and peaks. 3D stress calculations based on standard finite element methods are computationally expensive and not feasible for entire mountain massifs at a reasonable expense.

Our study addresses this limitation by employing a novel three-dimensional approach, utilizing the Marching Volume Polytopes Algorithm for mesh generation and the Finite Cell Method as an alternative to the widely used finite element method. Incorporating an octree-like structure and advancing-front meshing techniques, the Marching Volume Polytopes Algorithm accurately represents given surface data through a tetrahedral mesh. In the Finite Cell Method representing a fictitious domain approach, the difficulty of generating adequate grids for physical domains with complicated geometry is transformed into the problem of specifying an adequate integration scheme for the finite cells and thus saving degrees of freedom. The computational efficiency of our approach is particularly advantageous when dealing with equidistant grids such as digital elevation models for mesh generation.

In a first study, we use our model to compute the 3D topographic stress distribution for the three Austrian UNESCO Global Geoparks known for over-steepened valley flanks and high landslide activity. Initial results show high shear stress maxima occurring predominantly at over-deepened glacial valleys bordered by rock faces, with stress maxima at valley flanks but also at or slightly below the valley floors. Unexpected stress patterns occur in areas with a complicated landscape geometry, where valleys converge, or intersecting ridge lines form pyramid peaks. Lithological contrasts of the investigated mountain massifs are reflected in very different stress patterns, with shear stress maxima showing the highest values in carbonate-dominated units.

In addition to local topographic metrics, the spatial distribution of observed landslides and the rock types that occur, modelled topographic stresses provide a new data set for assessing landslide potential. Beyond that, modeling topographic stresses of entire mountain massifs offers new insights into the evolution of alpine landscapes in the competition between relief-forming and relief-destroying processes.

How to cite: Haunsperger, V., Robl, J., Schröder, A., and Hergarten, S.: Three-Dimensional Stress Analysis of Mountain Ranges: A Novel Approach Using Marching Volume Polytopes Algorithm and Finite Cell Method , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-6250, https://doi.org/10.5194/egusphere-egu24-6250, 2024.

Mapping benthic reefs at high resolution and accuracy is vital for the management and conservation of coral habitats. Optical remote sensing data has emerged as a valuable tool for large-scale reef mapping in the past decades, with numerous data sets and methods being utilised and developed. In this study, we present a comprehensive comparison of optical remote sensing based bathymetry and benthic mapping methods. We use different optical data including WorldView-2 stereo and Sentinel-2 imagery to map the water depths of coral reef areas in the Xisha region of the South China Sea. Bathymetry data derived from photogrammetric and linear regression methods are compared to the reprocessed Ice, Cloud and land Elevation Satellite-2 (ICESat-2) data. We find that the linear regression method (root-mean-square-error, RMSE=0.60 m) outperforms photogrammetry (RMSE=1.02 m), and the higher resolution WorldView-2 data yields less systematic biases than Sentinel-2 data. Considering that water depths reflect changes in temperature and light, which are critical factors influencing coral reef distribution, we propose to use satellite-derived bathymetry as a feature for coral reef classification. We demonstrate that combining topography and spectral information can improve the overall mapping accuracy, particularly for compositions characterised by sharp boundaries.

How to cite: Liu, Y., Zhou, Y., and Yang, X.: Bathymetry derivation and slope-assisted benthic mapping using optical satellite imagery in combination with ICESat-2, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-7114, https://doi.org/10.5194/egusphere-egu24-7114, 2024.

EGU24-7859 | Posters on site | GM3.2

The Performance of the Man-Kendall Test in the Analysis of Coastal Changes along Cliff Sections on the Baltic Sea 

Michael Fuchs, Lars Tiepolt, Karsten Schütze, and Jewgenij Torizin

Airborne Light Detecting and Ranging (LiDAR) surveys became essential in tracking the evolving coastal landscapes of Mecklenburg-Vorpommern on the Baltic Sea for more than one decade, producing a data series of Digital Terrain Models (DTMs) crucial for estimating coastal erosion along the exposed cliffs. Although change detection based on differences between these DTMs is supposed to represent erosion and deposition accurately, a detailed analysis indicates that the initial and final DTMs in the data series sometimes fail to capture the full extent of changes due to various factors. So, natural phenomena, such as the movement of cliff materials (rolling, sliding, creeping), human activities aimed at coastal protection, and errors in DTM processing may disturb clear trends, introducing uncertainties and, in particular, making the data series appear alternating.

To address these issues, we proposed to apply the robust Mann-Kendall test, a non-parametric statistical method used to identify trends in a data series without assuming any particular data distribution. It focuses on determining the direction and consistency of trends (ascending or descending), rather than the change’s magnitude. By implementing this approach, we can pinpoint areas that exhibit clear trends, thereby significantly improving the accuracy of coastal retreat estimations. In regions where trends are not readily apparent, it becomes crucial to investigate potential contributing factors thoroughly by exploring natural environmental dynamics, assessing the impact of human activities, and scrutinizing any errors in data processing. Such a comprehensive analysis ensures a more holistic understanding of the factors influencing these zones.

We employed the proposed approach across four distinguished shore areas characterized by the distinct geological composition of the cliffs, delving into the trends of coastal retreat over the past ten years. As expected for areas with clear trends, the estimation of the dimensions of the recent coastal retreat was in good agreement with historically recorded data. Additionally, in areas exhibiting no discernible trends, we were able to identify the underlying reasons, shedding light on the intricacies of coastal dynamics.

How to cite: Fuchs, M., Tiepolt, L., Schütze, K., and Torizin, J.: The Performance of the Man-Kendall Test in the Analysis of Coastal Changes along Cliff Sections on the Baltic Sea, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-7859, https://doi.org/10.5194/egusphere-egu24-7859, 2024.

EGU24-8222 | ECS | Posters on site | GM3.2

Automated and flexible measuring of grain size and shape in images of sediment with deep learning 

David Mair, Guillaume Witz, Ariel Henrique Do Prado, Philippos Garefalakis, and Fritz Schlunegger

The size and shape of sediment particles record crucial information on erosion, transport, and deposition mechanisms during sedimentary processes. Therefore, data on grain morphometry is a critical component in understanding sediment production and transport dynamics in various environments, such as fluvial or hillslope settings. However, traditional field methods are labor-intensive, and results may suffer from a limited number of observations. At the same time, remote measurements in images or point clouds still need improvements to counter low accuracy or the need for time-consuming manual corrections (e.g., Steer et al., 2022). These persisting challenges impede the capability of routinely obtaining size and shape information.

Here, we present a new and automated approach (Mair et al., 2023) for obtaining morphometric information on coarse sediment particles from segmented images. To do so, we tap into the capability for transfer learning of deep neural networks. In particular, we use state-of-the-art deep learning, developed to find cells in biomedical images, to segment individual grains in pictures of various sediments and image types. Our method validation includes assessing segmentation performance against ground truth from annotated images and evaluating the measurement quality by comparing results to independent measurements in the field and in images. This approach facilitates precise and rapid grain segmentation and outperforms existing methods. In addition, we observe that higher segmentation quality directly leads to improved precision and accuracy for grain size and shape data. Furthermore, any model of the used architecture can easily be re-trained for new image conditions, which we successfully did for several different settings. This highlights the potential for easy adapting to different environments and scales with comparatively small datasets.

References

Mair, D., Witz, G., Do Prado, A. H., Garefalakis, P., and Schlunegger, F.: Automated detecting, segmenting and measuring of grains in images of fluvial sediments: The potential for large and precise data from specialist deep learning models and transfer learning, Earth Surf. Process. Landforms, 1–18, https://doi.org/10.1002/esp.5755, 2023.

Steer, P., Guerit, L., Lague, D., Crave, A., and Gourdon, A.: Size, shape and orientation matter: fast and semi-automatic measurement of grain geometries from 3D point clouds, Earth Surf. Dyn., 10, 1211–1232, https://doi.org/10.5194/esurf-10-1211-2022, 2022.

How to cite: Mair, D., Witz, G., Do Prado, A. H., Garefalakis, P., and Schlunegger, F.: Automated and flexible measuring of grain size and shape in images of sediment with deep learning, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-8222, https://doi.org/10.5194/egusphere-egu24-8222, 2024.

EGU24-10314 | ECS | Posters on site | GM3.2

A deep learning-based super-resolution DEM model for pluvial flood simulation 

Yue Zhu, Paolo Burlando, Pauy Yok Tan, Christian Geiß, and Simone Fatichi

High-resolution Digital Elevation Model (DEM) data provides essential information for pluvial flood simulation. Although the increased accessibility and quality of publicly available DEM datasets can facilitate geospatial analysis at various scales, existing DEM datasets with global coverage mostly lack sufficient spatial resolution for pluvial flood simulations, which require detailed topographic information to be included in the simulation. Simulating flood scenarios with low-resolution DEMs (>30m) can result in substantial deviations from real cases. This issue becomes even more severe for flood-prone areas in data-scarce developing countries.

Image super-resolution is a technique for reconstructing low-resolution information into high-resolution data. Various deep-learning models have been employed for this task, primarily focusing on generating high-resolution natural-colour images. However, the effects of these deep learning models on enhancing the resolution of DEM data have not been extensively investigated. One of the state-of-the-art super-resolution models, the Residual Channel Attention Network (RCAN), has gained popularity due to its accuracy and efficiency. Leveraging publicly available low-resolution global DEM data and high-resolution regional DEM data, this study assesses the performance of RCAN models in a DEM super-resolution task. The experimental results suggest that, compared to conventional interpolation methods, the tested RCAN model exhibits superior performance in constructing high-resolution DEM data. The generated super-resolution DEM data were then tested in pluvial flood simulations and achieved substantially higher realism in modelling floodwater distribution. The proposed method for constructing super-resolution DEMs opens up the possibility of simulating flooding at hyper-resolution globally.

How to cite: Zhu, Y., Burlando, P., Tan, P. Y., Geiß, C., and Fatichi, S.: A deep learning-based super-resolution DEM model for pluvial flood simulation, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-10314, https://doi.org/10.5194/egusphere-egu24-10314, 2024.

EGU24-11527 | ECS | Orals | GM3.2

Identification of river channel bankfull geometry from topographic indicators extracted from high-resolution digital elevation datasets  

Valeria Ruscitto, Michele Delchiaro, Wolfgang Schwanghart, Eleonora Brignone, Daniela Piacentini, and Francesco Troiani

River channel bankfull geometry and discharge are important features providing valuable insights into fluvial monitoring and flood recurrency. The bankfull stage represents the riverbank position that approximates the level at which water overflows onto the floodplain. Bankfull discharge is considered the channel-forming discharge, with a recurrence interval of approximately 1.5 years. Bankfull floods are significant, as they are highly effective in changing channel shape and characteristics. Their recurrence intervals can be used for stream assessment and have implications for infrastructure design and flood mapping. Additionally, gaining insights into the factors influencing floodplain inundation across various time periods is crucial, as the frequency of flood events is predicted to rise with the increase in global temperatures.

In this contribution, we present a novel approach to identify the bankfull geometry through a set of dedicated MATLAB functions. A Digital Elevation Model (DEM) with ground resolution of 1 m/pixel is used as input elevation dataset, obtained with airborne LiDAR (Light Detection and Ranging) survey. The selected river channels are divided in regularly spaced sampling sections, where the bankfull geometry is extracted. Then, the hydraulic depth function that plots the elevation above the river thalweg vs. the ratio between the area and the width is computed for every section. Then, the elevation above river associated to the lowest and the most prominent peaks of the function, corresponding respectively to the bankfull stage or bankfull/floodplain inflection point and to the floodplain, are automatically extracted for each section. Manning’s equation is then applied to the hydraulic geometry corresponding to the lowest peaks elevation to compute the bankfull discharge at every river channel section. The validation process includes the comparison between the results obtained through the automatic bankfull geometry and discharge estimation and discharge data available from river hydrological gauges. Results demonstrate that the developed approach is effective to delineate the bankfull geometry from high-resolution DEMs and complements traditional qualitative field observations. Thus, our approach represents a cost-effective alternative for mapping detailed spatial variations over large spatial extents that are difficult to cover with traditional fieldwork.

How to cite: Ruscitto, V., Delchiaro, M., Schwanghart, W., Brignone, E., Piacentini, D., and Troiani, F.: Identification of river channel bankfull geometry from topographic indicators extracted from high-resolution digital elevation datasets , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-11527, https://doi.org/10.5194/egusphere-egu24-11527, 2024.

EGU24-12288 | Posters on site | GM3.2

TopoToolbox 3 – avenues for the future development of a software for terrain analysis 

Wolfgang Schwanghart, William Kearney, Anna-Lena Lamprecht, and Dirk Scherler

The Earth’s surface results from the interplay of tectonic and erosive forces, and the action of organisms and humans. To gain a deeper understanding of these interactions, accurate monitoring and analysis of topography is essential. Digital elevation models (DEMs) are powerful tools for achieving this goal and are available at ever increasing spatial resolution. TopoToolbox is a research software that provides a “laboratory” for the analysis of DEMs, enabling customized, automated analysis, prototyping and creative method development. Its high computational efficiency, ease-of-use and extensive documentation have attracted a worldwide user base across multiple research disciplines.

Over the last ten years, TopoToolbox, now in version 2, has undergone numerous changes and additions. The development of version 3 of TopoToolbox seeks to build on those past successes and take the software to the next level. Specifically, our goals are (1) to improve usability and accessibility, (2) to enhance quality assurance in the software’s development process, and (3) to increase community involvement in the ongoing development of TopoToolbox. We strive to achieve these goals in a recently funded 2-year project, in which community involvement is a key aspect. In this presentation, we aim to interact with other researchers interested in terrain analysis to discuss avenues for future developments and activities that improve TopoToolbox's usability, expand its usage, and increase its impact in a new version 3.

How to cite: Schwanghart, W., Kearney, W., Lamprecht, A.-L., and Scherler, D.: TopoToolbox 3 – avenues for the future development of a software for terrain analysis, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-12288, https://doi.org/10.5194/egusphere-egu24-12288, 2024.

After successfully applying segmentation and machine learning for landform identification and delineation for concave, convex, and generic landforms (landslides, floodplains), the used approach is generalized as a framework. The approach can be implemented in any GIS software that allows scripting and is based on four steps: (i) object-based segmentation based on a specific geomorphometric variable, (ii) contextual merging if the landform is composed of multiple shapes, (iii) selection of the training data segments, (iv) statistical classification by machine learning. The framework refers to creating a set of rules for various scenarios of landform types to allow the implementation of the approach for various landforms and areas around the globe. One of the main requirements regarding the DEM is that its feature resolution be high enough to allow at least a segment to cover the target landform spatially. This requires either LiDAR or RADAR DEMs, with medium or high resolution. We tested COPDEM in areas where there is no vegetation cover and the results show that landslides, floodplains, gullies, sinkholes, and closed depressions can be depicted by the approach.

How to cite: Niculita, M.: A generic framework for the identification and delineation of landforms from high-for DEMs using segmentation, contextual merging, and machine learning, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-13034, https://doi.org/10.5194/egusphere-egu24-13034, 2024.

EGU24-13135 | Posters on site | GM3.2

Spectral analysis as proxy for lineament spatial distribution: validation and case study 

Anna Maria Dichiarante, Tim Redfield, Espen Torgersen, Anne Kathrine Svendby, and Volker Oye

Spectral analysis (SA) is a technique commonly used in signal and image processing that makes use of the Fast Fourier Transform to compute the 2D power spectrum, which is a representation of the magnitude of each frequency component of the signal or image. SA can be similarly performed on a topographic map, and the orientation, frequency and magnitude (or power) of general topographic trends can be automatically retrieved and displayed in the 2D power spectrum. Recent studies have shown that spectral analysis can be successfully used to characterize repetitive and spatially homogeneous features or landforms, such as ridge and valley or glacial lineations. However, although these repetitive features dominate the 2D power spectrum, all the topographic information of the map is still present. Therefore, SA can be used on heterogenous and complex topographic map as a proxy for lineament analysis.

Lineament analysis is broadly used in a wide number of applications which include tectonic studies, exploration for groundwater, hazard evaluation for tunnel excavation, rockfalls or waste repository etc. Here, we propose a new methodology for lineament analysis based on spectral analysis and we demonstrate that this is a fast and effective way to derive lineament spatial distribution from images that can be visualized as rose diagrams. To validate our methodology, we stochastically generated 1000 synthetic lineament networks and numerically compared the rose diagrams derived from the power spectra to known lineament distribution. The comparison held a similarity of 94%.

The methodology was also applied to the Oslo region and compared to automatically extracted lineaments from OttoDetect software (developed by the Geological Survey of Norway). Results on three pre-selected areas characterized by different topographic patterns showed similarity of 97%, 95%, and 90%, respectively.

One of the pitfalls of spectral analysis is the lack of positioning on the original map of the signatures in the power spectrum. To locate the main signature on the map, we used the orientation of the main signatures from the power spectrum and used cross-correlation and clustering methods on topographic profiles.

How to cite: Dichiarante, A. M., Redfield, T., Torgersen, E., Svendby, A. K., and Oye, V.: Spectral analysis as proxy for lineament spatial distribution: validation and case study, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-13135, https://doi.org/10.5194/egusphere-egu24-13135, 2024.

EGU24-13867 | ECS | Orals | GM3.2

The effect of correcting the projection error in Digital Terrain Models on Earth surface processes 

Anne Voigtländer, Aljoscha Rheinwalt, and Stefanie Tofelde

Hiking up a steep mountain, in comparison to walking on a flat beach, is unarguably different. But the horizontal distance made, estimated using a Digital Terrain Model (DTM), might be the same. The projection of 3D landscapes onto 2D grids in DTMs leads to a slope-dependent, inhomogeneous sampling of the surfaces, and a first-order error in topographic metrics. Using the slope dependency of this error, we can quantify and revert it. Foremost, correcting the projection error allows for more accurate estimates of area and volume, e.g., to quantify natural hazards; and enables the use of the full slope distribution to define the physical space of surface processes at any scale.

We quantify the projection error using synthetic landscapes for which analytical solutions of slope angles and surface area are known. In applying the correction to DTM data of a real landscapes, we can address geomorphological processes in physically more meaningful ways. The corrected extracted topographic proxies, here exemplary, the erosional response to uplift in the Mendocino Triple Junction (MTJ) area, California, USA, provide two aspects for interpretation of geomorphic processes. First, as all slope angles are now represented equally, the variations in slope distribution by region of uplift rate is more pronounced. Second, the erosional response causes not only a steepening but narrow slope distribution in the regions of high uplift. The transient response is visible in a broadening of the distribution towards the lower slope angles, as deposition becomes more prevalent. In this example, we also find that the surface area ratio, enables determining the effectiveness of Earth surface processes, by increasing or decreasing the differential between the standard-planform and the surface area. Earth surface processes, that involve transport and volume along the surfaces, if not referenced in time, the ratio between the planform and surface area can provide a spatial reference and could be explored further. Correcting topographic metrics also allows addressing additional questions, like, which slope angles characterize which process domains, which processes create steepening, which lowering of slopes, where, and to what extent? Or, which parts of landscapes, maybe not the steepest, correlate to the highest potential to erode?

 

How to cite: Voigtländer, A., Rheinwalt, A., and Tofelde, S.: The effect of correcting the projection error in Digital Terrain Models on Earth surface processes, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-13867, https://doi.org/10.5194/egusphere-egu24-13867, 2024.

The integration of point-cloud data in geo- and environmental sciences has become increasingly pivotal, with applications ranging from UAVs, spaceborne and airborne lidars to ground-based lidars and stereo-photogrammetric techniques. This session seeks contributions that delve into challenges related to classification, segmentation, and noise removal in the context of point-cloud data, crucial for facilitating change detection studies. Our study focuses on the Navigational Branch of the ERDC Coastal Hydraulics Laboratory tasked with developing a Digital Twin model for a Dam, exemplifying the complexities involved in creating CAD models of terrain and structures.

To address the intricacies of point-cloud data processing, we employed both open-source and proprietary software solutions—Cloud Compare and Autodesk ReCAP— for noise reduction, ensuring the prepared data is seamlessly integrated into CAD modeling software, specifically Inventor. Surface modeling involved the strategic application of planes on cloud points to generate a foundation for sketching and subsequent solid surface extrusion.

Classification of data points was initiated through the implementation of regions in the noise removal software, facilitating the depiction of various areas on the model. Further, color and material assignment in the CAD software enhanced the identification of distinct part areas. Microstation TopoDOT played a pivotal role in creating a detailed terrain model, complete with physical landmarks and water bodies specific to the Dalles dam site.

The resulting models were exported in the desired file format, ensuring compatibility with sponsor requirements. This case study not only showcases the practical challenges encountered in working with point-cloud data but also highlights effective strategies for noise reduction, classification, and model exportation. The presented methodologies contribute to the broader spectrum of geo- and environmental sciences, emphasizing the significance of accurate point-cloud processing for comprehensive modeling endeavors.

How to cite: Krapac, M.: Advancements in Point-Cloud Processing for Geo-Environmental Modeling: A Case Study of The Dalles Dam Digital Twin Creation, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-14080, https://doi.org/10.5194/egusphere-egu24-14080, 2024.

EGU24-14955 | Posters on site | GM3.2

Multi-Technique Analysis and Landscape Evolution: Implications for Landslide-Fluvial Cascading Hazards Assessment 

Marta Guinau, Celeste Fernández-Jiménez, Anna Barra, Marc Viaplana-Muzas, Ariadna Flores, Maria Ortuño, Marta González, Jordi Pinyol, and Clàudia Abancó

The interaction between slope instability processes and river dynamics often triggers a cascade effect. Sediment influx from slopes can obstruct rivers, leading to upstream flooding and potential catastrophic flash floods downstream upon dam breakage. In addition, the incision of the drainage network steepens the valley hillslopes, further exacerbating slope instability processes, modifying the geomorphology and the sedimentary fluxes and increasing the occurrence of landslide-derived hazards.

In this regard, a comprehensive and updated landslide inventory, especially focusing on the interconnection between landslides and drainage networks, is crucial for effective hazard assessment considering these cascading effects induced by slope and fluvial processes.

This study presents advancements in landslide mapping by integrating data from Multi-Temporal Synthetic Aperture Radar (MT-InSAR) and landscape evolution analysis through geomorphological indices such as Chi, Normalized Channel Steepness Index (Ksn) and Stream Length-Gradient Index (SL). Identification of anomalies along rivers using Ksn and SL (knickpoints or knickzones) aided in pinpointing abnormal slopes due to sediment influx from landslides. Additionally, active areas were delineated using the ADAfinder tool, extracting data from MT-InSAR provided by the European Ground Motion Service (EGMS). This multi-technique analysis highlighted the slopes of interest. Landslides identified with these techniques were delimited and characterized in terms of type assignment, using 2x2 m DTM hillshades derived from airborne LiDAR data and field observations.

The upper catchments of the Garona and Noguera Pallaresa rivers (central Pyrenees-NE Spain) were selected as study cases. The study highlights the disequilibrium in the watershed divide between Noguera Pallaresa and Garona basins, suggesting a transition toward equilibrium favouring a main divide migration towards the Noguera Pallaresa due to hillslope processes. The assessment of the equilibrium profile geometry of the Noguera Pallaresa river at a regional scale suggests at least two main knickpoints. The river sections downstream of the knickpoints are associated with landslides triggered by post-glacial dynamics and incision wave effects. Combining SL and Ksn curves with Active Deformation Areas (ADA) underscores areas with potentially reactivating deep-seated landslides, signifying potential high damages in case of low-probability but catastrophic reactivations.

In conclusion, the integration of diverse methodologies shed light on the spatial relationship between transient features in the landscape (knickpoints) and landslide occurrence, emphasizing the need for a comprehensive approach to mitigate landslide and fluvial risks in the Noguera Pallaresa and Garona river basins.

How to cite: Guinau, M., Fernández-Jiménez, C., Barra, A., Viaplana-Muzas, M., Flores, A., Ortuño, M., González, M., Pinyol, J., and Abancó, C.: Multi-Technique Analysis and Landscape Evolution: Implications for Landslide-Fluvial Cascading Hazards Assessment, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-14955, https://doi.org/10.5194/egusphere-egu24-14955, 2024.

EGU24-15001 | Orals | GM3.2 | Highlight

Applying photogrammetry to time-lapse imagery for geomorphological change detection 

Anette Eltner, Xabier Blanch, Oliver Grothum, Lea Epple, Eliisa Lotsari, Katharina Anders, and Melanie Elias

Cameras that capture images in time-lapse mode of the earth surface enable great opportunities for change detection and thus potential process identification and understanding. The camera systems can range from simple and robust game cameras to complex and synchronised full frame cameras. The main workflow of calculating digital elevation models from overlapping images is similar for the different types of systems; automatically matching the images, performing bundle adjustment considering either calibrated or non-calibrated cameras, geo-referencing the data by automatic ground control point (GCP) measurement, densifying the point cloud and eventually calculating point cloud differences. However, adapted pre-and post-processing steps are needed due to the varying observation conditions considering the camera qualities and the objects of interest. The time-series of point cloud-based change information can be further processed, for example, with time-series clustering approaches to disentangle overlapping processes.

We will introduce three different case studies in the field of fluvial geomorphology, soil erosion research and rockfall assessment. Thereby, different camera systems are utilized. Four low-cost time-lapse cameras are applied in arctic environments to study changes of a river bank at a distance of about 60 m. The high robustness of the cameras encompasses the trade-off of low quality images. In addition, challenging lighting conditions and enduring snow cover complicate the photogrammetric processing. The images are captured with a frequency of two hours, and six permanent GCPs are used to geo-reference the measurements.

Digital SLR cameras are used in moderate climate to measure soil surface changes either due to rainfall simulations or due to natural rainfall events. During the rainfall simulation we use images that are captured by up to ten cameras with a frequency of 10 to 20 seconds and at an object distance between 3 to 4 m. And at the field plot we installed three special camera rigs that encompass five cameras each that are event-controlled by a micro-controller and single board computer solution, which trigger the cameras each time a rain collector bucket is tipping in addition to daily captured images. Challenges for change detection arise from vegetation present at the plots and from runoff water covering the soil surface. Eventually, the derived models of change are used to validate physical based soil erosion models.

The last case study utilizes five full-frame system cameras in the Mediterranean to detect single rockfall events. Images are captured three times a day by an ad-hoc system at a distance of about 100 m. The data is transferred via a locally installed network module. Many areas within the field of view remain stable throughout the measurement period allowing for a time-SIFT approach that matches the images from different points in time. Machine learning algorithms are applied to automatically identify rockfalls in the final 4D dataset. Thereby, we showcase the great potential of time-lapse photogrammetry for different applications of geomorphological change detection.

How to cite: Eltner, A., Blanch, X., Grothum, O., Epple, L., Lotsari, E., Anders, K., and Elias, M.: Applying photogrammetry to time-lapse imagery for geomorphological change detection, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-15001, https://doi.org/10.5194/egusphere-egu24-15001, 2024.

EGU24-15418 | ECS | Posters on site | GM3.2 | Highlight

A definition of land surface geomorphodiversity across different scales 

Martina Burnelli, Laura Melelli, Francesco Bucci, Michele Santangelo, Federica Fiorucci, and Massimiliano Alvioli

Geodiversity is “the variety of abiotic features and processes of the land surface and subsurface” [1,2]. Consensus is growing that geodiversity is the geosphere counterpart of what biodiversity represents within the biosphere, atmosphere, and hydrosphere [2]. Thus, it is potentially relevant to ecosystem functions and services [2]. Since the introduction of geodiversity, several scholars studied it from the theoretical and practical points of view, with different approaches, assumptions and purposes. Methods to define diversity of the geosphere are quantitative, qualitative, or a combination of the twos, with the occasional addition of heuristics [3].

Here, we describe a quantitative derivation of a subset of geodiversity, namely, geomorphodiversity. The effort stems from the need of an objective method, apt to providing easy to understand results, readily available for subsequent applications. To that end, requirements are in order about the data included in the analysis: they should be widely available, to allow reproduction of the analysis in most geographical locations, and they should contain enough information to approximate real-world geodiversity.

Geomorphodiversity is one implementation fulfilling the requirements, obtained in the literature by different groups, for different locations [4,5], using simple geomorphometry. Data for the method implemented in Italy [6] are a digital elevation model (EUDEM, 25 m resolution), and a lithological map at 1:100,000 scale [7]. DEM provides derived quantities such as slope, drainage network, landforms [8] and slope units [9], all of which contribute in different ways to produce partial diversity maps. We eventually combine partials into an overall geomorphodiversity raster index, GmI, distinguishing five classes of land surface diversity.

The inherent parameter dependence in the existing implementations of GmI, partially resolved in [6], is one issue to overcome. Free parameters are embedded in the size of neighborhoods (moving windows, or focal statistics) used to calculate the variety, the arbitrary output resolution, and procedures to polish the final raster diversity map from artifacts. We suggest a multiple assessment of the variety of partial abiotic parameters with a full range of different neighborhood sizes, and a-posteriori statistical selection of local values of diversity. This results in a parameter-free approach to GmI, also allowing a custom resolution of the output, with the lower bound of DEM resolution.

We consider a parameter-free geomorphodiversity as a measure of the potential of morphological evolution of the landscape, useful to investigate natural and human-induced diversity in urban areas [10], in combination with accurate, local mapping of geomorphological landforms [11].

 

References

[1] Gray, (2004) Geodiversity: valuing and conserving abiotic nature. ISBN 978–0–470-74215-0

[2] Schrodt et al., PNAS (2019) https://doi.org/10.1073/pnas.1911799116

[3] Zwoliński et al., Geoheritage (2018) https://doi.org/10.1016/B978-0-12-809531-7.00002-2

[4] Benito-Calvo et al, Earth Surf Proc Land (2009) https://doi.org/10.1002/esp.1840

[5] Melelli et al., Sci Tot Env (2017) https://doi.org/10.1016/j.scitotenv.2017.01.101

[6] Burnelli et al., Earth Surf Proc Land (2023) https://doi.org/10.1002/esp.5679

[7] Bucci et al., Earth System Science Data (2022) https://doi.org/10.5194/essd-14-4129-2022

[8] Jasiewicz et al., Geomorphology (2013) https://doi.org/10.1016/j.geomorph.2012.11.005

[9] Alvioli et al., Geomorphology (2020) https://doi.org/10.1016/j.geomorph.2020.107124

[10] Alvioli, Landscape and Urban Planning (2020) https://doi.org/10.1016/j.landurbplan.2020.103906

[11] Del Monte et al., Journal of Maps (2016) https://doi.org/10.1080/17445647.2016.1187977

How to cite: Burnelli, M., Melelli, L., Bucci, F., Santangelo, M., Fiorucci, F., and Alvioli, M.: A definition of land surface geomorphodiversity across different scales, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-15418, https://doi.org/10.5194/egusphere-egu24-15418, 2024.

EGU24-15904 | Orals | GM3.2

Mapping gold mines under the French Guiana rainforest: return of experience with different mobile lidar systems 

Thomas Dewez, Sébastien Linares, Silvain Yart, Florian Masson, Marie Collignon, Lucas Rivera, Caroline Bedeau, and Matthieu Chevillard

Gold is abundant in the greenstone belts of the Guiana shield, in South America, leading to alluvial mining in river sediments and in in-situ rocks. In French Guiana, legal mining takes place under strict environmental regulations and controls, but illegal operations also occur uncontrolled in the vast expanses of the rainforest. Here we describe a successful range of mobile lidar systems, acquisition schemes and processes to map the ground and underground mining operations in a rainforest context. We seek to detect illegal operations, supply and transportation pathways and base camps, using crewed planes and helicopters, uncrewed fixed-wing and multi-copter vehicles (UAV) and handheld lidar systems.

To sense ground elevation below the canopy, airborne lidar systems face three challenges: tree heights (some trees exceed 70 m in height), incised topography (requires performant terrain following capabilities), dark and wet ground surface largely absorbs lidar pulses requiring powerful sources. Tested uncrewed airborne vehicles (UAV) did not yet meet all of the flying autonomy, terrain-following capability, lidar range and on-board decision systems. At present, crewed systems adapt better to conditions and achieve mission objectives.

Over forested areas, observed canopy penetration rates is of the order of 1 ground point for 250 lidar pulses (0.4%). To generate a 1-m/pixel Digital Terrain Model (DTM) with a minimum of occluded pixels, acquisition density should exceed 250 pts/m² at canopy level everywhere. In Dorlin (central French Guiana), a helicopter flew 85-m-above ground-level, 70 % side-lap and 90° cross-lines, using a Riegl VUX-1LR lidar. Targeting 400 pts/m² at canopy-top for 95 % of the 220 ha territory, it reached a canopy-top density of 1400 +/- 750 pts/m² and 43 pts/m² ground density overall. On fully forested areas, ground density dropped to 22.4+/-22.6 pts/m² with 5% of the surface never receiving points at 1 m² level. This enabled interpolation of a 25cm/pixel DTM, which revealed narrow paths, quad tracks, and shaft platforms and head frames under the forest. 2-m kernel high-pass filtering enhanced features better than a standard hill shading. Base camp hut structures, invisible in DTM, are retrievable from native point clouds in a 4 to 5 m-above-ground elevation range. Huts covered in black tarpaulins stand out as rectangular hollow patches due to lidar photon absorption. But even without tarpaulin, hut wooden frames stand out particularly well when point cloud subsets are lit up with the PCV filter of Cloud Compare. Ore-bearing quartz stockpiles however are too small and occluded for a reliable detection and volume computation.

Instead, SLAM-based handheld lidar systems (GeoSLAM Zeb-Revo and Zeb-Horizon) complement the detailed mapping of quartz stockpiles volume, shaft conduit geometry and gallery entrances. Then real-time, SLAM-based quadcopter UAV lidar (Flyability Elios 3) safely penetrates shafts from the surface to explore the undergound gallery network. These new millimetre-scale density point clouds critically reveal spacing, orientation and dimensions of ore-bearing veins, which improves the metallogical understanding of the site and uniquely documents the way artisanal illegal miners operate.

Lidar acquisitions and processing are now being streamlined for systematic use in law enforcement operations and environmental protection actions.

How to cite: Dewez, T., Linares, S., Yart, S., Masson, F., Collignon, M., Rivera, L., Bedeau, C., and Chevillard, M.: Mapping gold mines under the French Guiana rainforest: return of experience with different mobile lidar systems, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-15904, https://doi.org/10.5194/egusphere-egu24-15904, 2024.

EGU24-16317 | Orals | GM3.2

Debris flow catchments and landscape evolution in the northern Colombian Andes 

Edier Vicente Aristizábal Giraldo and Oliver Korup

Fans are cone-shaped depositional landforms composed of a mixture of sediments, mainly derived from debris flow processes at the catchment scale. In mountainous terrains located in humid climates, debris flows are fundamental agents of landscape evolution and a highly destructive natural hazard. In the northern Colombian Andes, fans have been traditionally occupied by human settlements, which has also produced a long history of disasters in many settlements located on fans. For example, a debris flow on November 13, 1985, devastated the city of Armero, killing approximately 22,000 people and causing economic losses totaling over $US 339 million. In 2017, the city of Mocoa was affected by a debris flow where 333 people died, 130 houses were destroyed, and 1461 were partially affected.

Debris-flow risk is likely to increase as a consequence of the increasing magnitude and frequency of extreme weather and rapid population growth over the past few decades. Hence, identifying fan spatial distribution and debris flow occurrences is important for land use planning. In this study, we implemented geomorphometric analyses in the northern Colombian Andes to understand debris flow occurrence in terms of landscape evolution. Using digital elevation models, fan inventory, morphometric parameters, and geomorphic indices associated with the drainage network at the catchment scale, the close interconnection between debris-flow hazards and landscape evolution is explained.

The results show a clear spatial pattern of fans location and debris-flow-prone basins with knickpoint upstream migration and transient-state catchments, those characterized by high values of Ksn, hypsometric index and constraint values of 𝛘. Those findings suggest that landscape evolution indexes could improve debris flow susceptibility assessment at regional scale.

How to cite: Aristizábal Giraldo, E. V. and Korup, O.: Debris flow catchments and landscape evolution in the northern Colombian Andes, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-16317, https://doi.org/10.5194/egusphere-egu24-16317, 2024.

EGU24-16412 | ECS | Orals | GM3.2

Deep-Image-Matching: an open-source toolbox for multi-view image matching of complex geomorphological scenarios 

Francesco Ioli, Luca Morelli, Livio Pinto, and Fabio Remondino

Geomorphometry and geomorphological mapping are essential tools for understanding landscape changes. The recent availability of 3D imaging sensors and processing techniques, including Artificial Intelligence, is offering interesting solutions for gemorphometric analyses and processes understanding. Photogrammetry stands as a pivotal image-based tool in geomorphology, enabling accurate 3D reconstruction of complex natural environments and effective tackling of multi-temporal monitoring challenges. A key step in photogrammetry is the identification of corresponding points between different images, traditionally achieved through the extraction and matching of local features such as SIFT and ORB. However, these methods face difficulties when using images of complex environments scenarios. Deep Learning (DL) methods have recently emerged as powerful tools to address challenges such as strong radiometric variations and viewpoint changes (Morelli et al., 2022; Ioli et al., 2023). However, their practical application in photogrammetry is hindered by the lack of libraries integrating DL matching into standard SfM pipelines.

The presentation will introduce the recently developed Deep-Image-Matching, an open-source toolbox designed for multi-view image matching using DL approaches, specifically tailored for 3D reconstruction in complex scenarios (https://github.com/3DOM-FBK/deep-image-matching). This tool can be used to achieve a 3D reconstruction with wide camera baselines and strongly varying viewpoints (e.g., with ground-based monitoring cameras), with datasets involving varying illumination or weather conditions typical of multi-temporal monitoring, with historical images, or in low-texture situations (e.g., snow or bare ice).

Deep-Image-Matching provides the flexibility to choose from a variety of local feature extractors and matchers. Supported methods include traditional local feature extractors, such as ORB or SIFT, as well as learning-based methods, such as SuperPoint, ALIKE, ALIKED, DISK, KeyNet + OriNet + HardNet, and DeDoDe. Matcher choices range from traditional nearest neighbor algorithms to state-of-the-art options like SuperGlue and LightGlue. Available semi-dense matching solutions include the detector-free matchers LoFTR and RoMa.

To handle high-resolution images, the tool offers a tiling process. In case of strong image rotations, such as aerial stripes, images are automatically rotated before matching. Image pairs for matching can be selected by exhaustive brute-force matching, sequential matching, low-resolution guided pairs selection, or global descriptor-based image retrieval. Geometric verification is used to discard outliers among matched features. The extracted image correspondences are stored in a COLMAP database for further processing (i.e. bundle adjustment and dense reconstruction) or can be exported in other formats useful for other open-source and commercial software.

The presentation will highlight how image-based geomorphometry and geomorphological mapping could benefit of the realized tool and how complex environmental scenarios (landslides, glaciers, etc.) could be analysed and monitored with the support of deep learning.

References:

Ioli, F., Bruno, E., Calzolari, D., Galbiati, M., Mannocchi, A., Manzoni, P., Martini, M., Bianchi, A., Cina, A., De Michele, C. & Pinto, L. (2023). A Replicable Open-Source Multi-Camera System for Low-Cost 4D Glacier Monitoring. Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci., 48, 137-144

Morelli, L., Bellavia, F., Menna, F., & Remondino, F. (2022). Photogrammetry Now and Then - From Hand-Crafted to Deep Learning Tie Points. Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci., XLVIII-2/W1-2022, 163–170

How to cite: Ioli, F., Morelli, L., Pinto, L., and Remondino, F.: Deep-Image-Matching: an open-source toolbox for multi-view image matching of complex geomorphological scenarios, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-16412, https://doi.org/10.5194/egusphere-egu24-16412, 2024.

EGU24-18657 | ECS | Posters on site | GM3.2

Tailoring slope units delineation according to different natural phenomena for institutional land use planning at the regional scale 

Rossana Napolitano, Michele Delchiaro, Leonardo Maria Giannini, Claudia Masciulli, Giandomenico Mastrantoni, Marta Zocchi, Massimiliano Alvioli, Paolo Mazzanti, and Carlo Esposito

The Latium region (Central Italy) is currently updating the institutional hydro-geological plan, one of the main planning tools to prevent geo-hydrological hazard at regional scale. The plan focuses on landslides, erosion and hydraulic hazard assessment using both conventional and innovative approaches. This analysis required different scales of study, according to the different processes acting on slopes, and their broader physiographic context. In this multiscale approach, slope units represent the most suitable territorial units of analysis and mapping, considering their morpho-hydrological representativeness and scalability.

Slope units are a particular type of terrain units, characterized by internal homogeneity and external heterogeneity, delineated from a digital elevation model considering the natural setting of the territory. A widely used tool for slope unit delineation is the software ‘’r.slopeunits’’ [1,2]. The parameters controlling the delineation are both morphological and hydrological, derived from a digital elevation model. The software implements an iterative and adaptive process, depending on the aforementioned parameters, resulting in slope unit sets optimized for the local morphology. The accurate selection of input parameters requires careful consideration, but it also allows extra flexibility in defining the proper scale of the output slope unit map.

Here, we aim at obtaining a new way to select the values of the software’s input parameters, considering their relations with the different processes, to single out the proper scale of analysis. Specifically, we provide additional terrain analysis methods to find “good” parameter ranges, implemented in simple computer scripts that make use of r.slopeunits. The workflow is organized as follow. First, the geomorphological domains (i.e. hillslope, unchanneled, and fluvial domain) are discriminated by the implementation of the slope – area function, with the area weighed by the runoff values available from the GIS-based model BIGBANG [3]. Next, the flow paths related to the hillslope and unchanneled domains and related basins are hierarchized using Strahler ordering. Then, delineation of basins and half-basins for every path order is computed. Finally, implementation of zonal statistics functions on the half-basins of every path order and calculation of the parameters ranges that for slope unit delineation is performed.

Implementation of a multi – scale derivation of slope units with a range of input parameters, customized according to the type of natural phenomena (landslide, flooding, erosion etc.), allows an adaptive multi – scale approach, specific for each process, for a comprehensive multi-hazard evaluation. One of the future applications of the research is the application of this approach for the definition of ‘’buffer zones’’ covered by natural or semi-natural vegetation, capable of counteracting slope instabilities. In the context of the hazard and risk mitigation management, these outcomes could represent an efficient aid for regulating urban development in a proper and secure manner.

 

References

[1] Alvioli et al. (2016). Geosci Mod Dev, https://doi.org/10.5194/gmd-9-3975-2016

[2] Alvioli et al. (2020). Geomorphology, https://doi.org/10.1016/j.geomorph.2020.107124

[3] BIGBANG model, https://www.isprambiente.gov.it/pre_meteo/idro/BIGBANG_ISPRA.html

 

How to cite: Napolitano, R., Delchiaro, M., Giannini, L. M., Masciulli, C., Mastrantoni, G., Zocchi, M., Alvioli, M., Mazzanti, P., and Esposito, C.: Tailoring slope units delineation according to different natural phenomena for institutional land use planning at the regional scale, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-18657, https://doi.org/10.5194/egusphere-egu24-18657, 2024.

EGU24-18943 | ECS | Posters on site | GM3.2

Reconstructing ancient coastal landscapes and sea-level stands in Southern Italy (Cilento coast): a geostatistical approach 

Alessia Sorrentino, Gaia Mattei, and Pietro Patrizio Ciro Aucelli

This research aims to obtain coastal paleo-environmental reconstructions through the analysis of direct and indirect paleo sea-level markers (SLMs, i.e., SLIPs, TLPs, MLPs) by GIS-aided geostatistics.  

In this work, we used classical SLMs combined with a caves inventory in the Cilento area in the Campania Region (Southern Italy). In this area, mainly characterized by carbonatic rocks, numerous emerged and submerged caves are present along active and fossil cliffs as evidenced in the papers of Antonioli et al., 1994 and Esposito et al., 2002.  

As reported in Ferranti, 1998 and Florea et al., 2007, coastal caves can be considered positively correlated to the glacial-hydro-eustatic sea-level oscillations, especially on the carbonatic substratum.  

Therefore, caves cannot be classified as sea-level markers (SLMs) strictu sensu, anyway, they can be considered as a mark of ancient sea-level position, especially when the occurrence of floor elevation is well-distributed all along the coast (in the case of areas characterised by homogeneous tectonic behaviour). In detail, in this work, the floor elevation of the cave entrances was correlated with tidal notches, wave-cut platforms, Lithophaga burrows, and marine deposits deriving both from previous knowledge and new direct and indirect surveys carried out through classic geomorphological investigations and using robotic technologies and remote sensing.  

All collected data were used to produce a specific geodatabase “PALEOScape (PALEO SeasCAPE)” (Sorrentino et al., 2023) structured based on international standards for sea-level studies. Caves information was obtained from an existing caves’ Inventory (Federazione Speleologica Campana; Russo et al., 2005) integrated by field surveys. Thanks to the well-documented tectonic stability of the area, it was possible to ascribe at the same age SLMs having the equal altimetric position.

These records were analysed by a geostatistical approach by correlating the cave entrances to known sea-level stands increasing the information available on paleo sea-level stands along the examined coast.

By integrating this approach with a new method for semi-automatic landform recognition and classification, it was possible to reconstruct ancient coastal landscapes related to known sea level stands, but also to some new altimetric positions not previously reported in the area.

REFERENCES

Antonioli, F., Cinque, A., Ferranti, L., & Romano, P. 1994. Emerged and Submerged Quaternary Marine Terraces of Palinuro Cape (Southern Italy). Memorie Descrittive Carta Geologica d’Italia, 52, 237–260.

Federazione Speleologica Campana https://www.fscampania.it/catasto-2/catasto/  

Ferranti, L. 1998. Underwater cave systems in carbonate rocks as semi-proxy indicators of paleo-sea levels. Il Quaternario-Italian Journal of Quaternary Sciences, 11(1), 41-52.

Florea, L. J., Vacher, H. L., Donahue, B., Naar, D. 2007. Quaternary cave levels in peninsular Florida. Quaternary Science Reviews, 26(9-10), 1344-1361.

Russo, N., Del Prete, S., Giulivo, I., Santo, A. 2005. Grotte e speleologia della Campania : atlante delle cavità naturali. Elio Sellino Editore.

Sorrentino, A., Maratea, A., Mattei, G., Pappone, G., Tursi, M. F., Aucelli, P. P. 2023. A GIS-based geostatistical approach for palaeo-environmental reconstructions of coastal areas: the case of the Cilento promontory (southern Italy). In 2023 IEEE International Workshop on Metrology for the Sea; Learning to Measure Sea Health Parameters (MetroSea) (pp. 488-493). IEEE.

 

How to cite: Sorrentino, A., Mattei, G., and Aucelli, P. P. C.: Reconstructing ancient coastal landscapes and sea-level stands in Southern Italy (Cilento coast): a geostatistical approach, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-18943, https://doi.org/10.5194/egusphere-egu24-18943, 2024.

EGU24-20335 | ECS | Posters on site | GM3.2

 Bivariate mountain definition: a case study for the turkish mountain system  

Neslihan Dal and Tolga Görüm

Türkiye, 61% of which consists of mountains, has an extremely rugged topography. Anatolia, which is located in the collision zone of plates with different characteristics, exhibits a morphological character with different stages of mountain formation due to the Paleotectonic and Neotectonic movements it has been exposed to during geological times. In Anatolia, where the main physiographic character is mountains, the proportion and boundaries of mountains and mountainous areas have not been quantitatively defined and there has not been a geomorphometric approach to this until now. In this study, the mountain definition obtained from the pixel-based and multi-scale basic data matrix was subjected to various analyzes with the modeling created in geographic information systems. In addition, how the mountain definition and classification change at varying scales and thresholds is revealed.

The characterization has two main purposes: To determine the framework of the methodology in the definition of macro landforms and to determine the most optimum model that quantitatively defines mountain and mountainous area. According to the model, mountains cover 61% of Türkiye. In this context, in addition to developing a model to geomorphometrically define mountain and mountainous area characterization, the thesis approaches mountains, which are a macro morphological landform, from an ontological perspective and approaches the questions we asked at the beginning in terms of geographical epistomology. In this respect, the thesis is a contribution to traditional geomorphology.  A bivariate map of 16 classes to visualize the relationships between morphological variables and a combination of mean elevation and topographic relief classifies mountains. The classification shows a transition from low rugged and low mountains, to moderate rugged and moderate height mountains, to high rugged and high mountains, to very high rugged and very high mountains. Within the framework of the classification, according to four different ruggedness ratios in Türkiye, low rugged mountains occupy 37%, moderate rugged mountains 33%, high rugged mountains 20% and very high rugged mountains 9%.

How to cite: Dal, N. and Görüm, T.:  Bivariate mountain definition: a case study for the turkish mountain system , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-20335, https://doi.org/10.5194/egusphere-egu24-20335, 2024.

EGU24-20923 | Orals | GM3.2

Do Mature, Fluvial Landscapes Obey Hamilton's Principle? 

Scott D. Peckham

Students of physics typically take a theory course on classical mechanics in which they learn about Hamilton's Principle and how it can be used to derive many well-known physical laws that describe the motion of objects from particles, to light rays, to celestial bodies, including Newton's laws and Snell's Law from geometrical optics.  This powerful principle has also been shown to apply to fields (i.e., continuous systems) such as the electromagnetic and gravitational fields, and it is a foundational concept in quantum physics.  Hamilton's principle states that the dynamics of a physical system will optimize a functional (in our case, an integral over a spatial domain) of the system's Lagrangian, which is typically the difference between its kinetic and potential energies.  Many previous authors have postulated that fluvial landscapes may evolve in such a way that local and/or global kinetic energy dissipation or stream power is minimized, and this is the basis of the optimal channel network (OCN) simulation models that have been widely studied.  However, Hamilton's Principle suggests that these formulations are lacking an important piece, namely the global introduction of potential energy into the fluvial system by rainfall.  The author will show that by introducing this missing piece, Hamilton's Principle and the Euler-Lagrange theorem lead to a partial differential equation (PDE) for idealized, steady-state landforms.  This same PDE can also be derived from conservation of mass and an empirical slope-discharge formula.  These connections therefore point to a new theoretical framework for understanding the interplay between function and form in mature, fluvial landforms;  that is, an explanation for why these landforms take the forms we observe.  The author will also present ideas and algorithms for analyzing digital elevation models (DEMs), in an effort to test for agreement with Hamilton's principle.

How to cite: Peckham, S. D.: Do Mature, Fluvial Landscapes Obey Hamilton's Principle?, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-20923, https://doi.org/10.5194/egusphere-egu24-20923, 2024.

EGU24-22282 | ECS | Orals | GM3.2

A data-driven approach to understanding esker morphogenesis 

Meaghan Dinney and Tracy Brennand

Eskers are ubiquitous features on previously glaciated landscapes, recording the configuration and dynamics of the channelized meltwater system. Studies of esker composition and form have resulted in a variety of genetic interpretations surrounding the ice, water, and sediment characteristics under which they may develop. However, issues of apparent equifinality currently limit the usefulness of eskers for reconstructing broad-scale glacial hydrology. Although some authors have attempted to asses esker morphogenesis, previous studies are limited by their small sample size and/or use of qualitative morphometric indices.

This project aims to explore whether eskers have a distinct morphogenetic signature using data science techniques. Published research has been mined for empirical studies of esker composition and structure. These data were compiled into a database summarizing the genetic interpretations commonly invoked for eskers (e.g., depositional environment, meltwater flow regime) as well as the supporting evidence for such inferences (e.g., sedimentary logs). Semi-automated methods will be tested to map eskers from high resolution (1-2 metres) LiDAR digital terrain models and to extract their morphometry. A range of planform- and profile-scale morphometric indices will be employed and new indices that can more precisely quantify esker morphometry will be developed.

The resulting highly-dimensional dataset can be analyzed using machine learning techniques in order to assess the relationships between sedimentologic, morphometric, and genetic variables. Preliminary results from database development and analysis will be presented and methodological concerns will be discussed.

How to cite: Dinney, M. and Brennand, T.: A data-driven approach to understanding esker morphogenesis, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-22282, https://doi.org/10.5194/egusphere-egu24-22282, 2024.

NP5 – Predictability

This study introduces an innovative approach aimed at enhancing the accuracy of regional weather forecasts from the AROME model, covering northern Algeria. By leveraging AROME analysis, a refined representation based on real observations and widely used for monitoring and validating our model, our primary objective was to precisely correct surface parameters, including temperature at 2 meters, humidity, wind force, and sea-level atmospheric pressure (MSLP). This correction was performed based on their forecast ensemble, all while preserving spatial resolution.

This methodology has yielded promising results, demonstrating a significant improvement in the accuracy of regional weather forecasts. The presentation will delve into the detailed integration process of the Convolutional Neural Network (CNN) and AROME analysis, highlighting the successes achieved in correcting essential surface parameters. These advancements strengthen the reliability of regional meteorological models, with positive implications for resource planning and management in the northern region of Algeria.

How to cite: Bousri, I.: Optimisation of Regional Weather Forecasts for Northern Algeria Using a Convolutional Neural Network and AROME Model Analysis., EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-1665, https://doi.org/10.5194/egusphere-egu24-1665, 2024.

EGU24-2005 | Posters on site | NP5.2

Enhancing member-by-member post-processing with neural networks 

Sebastian Lerch, Jakob Freytag, Thomas Muschinski, and Sam Allen

Using post-processing methods to correct systematic errors of ensemble forecasts has become standard practice in research and operations. During recent years, a new focal point of research interest has been the use of modern machine learning methods to allow for more flexible post-processing methods that incorporate additional input predictors. In particular, neural network (NN) models have been shown superior predictive performance in various case studies [1-3].

In contrast, the member-by-member (MBM) post-processing approach [4] adjusts each ensemble member individually using a relatively simple statistical model. This has the advantage that the post-processed ensemble forecasts are not only calibrated, but physically consistent over time, space and different weather variables. Therefore, multivariate dependencies are preserved even if MBM is applied separately for each component. The drawback is that MBM has no straightforward way of incorporating additional input variables (beyond ensemble predictions of the target variable) and therefore typically fails to perform as well as NN-based post-processing approaches [3].

To address this shortcoming, we propose a novel NN-enhanced MBM post-processing approach (“MBM-NN”), which combines the basic idea of MBM with a neural network for incorporating additional predictors to leverage advantages of both approaches. In case studies on probabilistic wind gust forecasting over Germany and on the EUPPBench dataset [5], we demonstrate that the MBM-NN model achieves significant improvements over the standard MBM approach, and reaches comparable performance to state-of-the-art NN-based post-processing models, while retaining multivariate dependencies.

References

[1] Rasp, S. and Lerch, S. (2018). Neural networks for post-processing ensemble weather forecasts. Monthly Weather Review, 146, 3885-3900

[2] Vannitsem, S., Bremnes, J.B., Demaeyer, J., Evans, G.R., Flowerdew, J., Hemri, S., Lerch, S., Roberts, N., et al. (2021). Statistical Postprocessing for Weather Forecasts - Review, Challenges and Avenues in a Big Data World. Bulletin of the American Meteorological Society, 102, E681-E699

[3] Schulz, B. and Lerch, S. (2022). Machine learning methods for postprocessing ensemble forecasts of wind gusts: A systematic comparison. Monthly Weather Review, 150, 235-257

[4] Van Schaeybroeck, B. and Vannitsem, S. (2015). Ensemble post‐processing using member-by-member approaches: theoretical aspects. Quarterly Journal of the Royal Meteorological Society 141, 807-818

[5] Demaeyer, J., Bhend, J., Lerch, S., Primo, C., Van Schaeybroeck, B., Atencia, A. Ben Bouallègue, Z., Chen, J., Dabernig, M., Evans, G., Faganeli Pucer, J., Hooper, B., Horat, N., et al. (2023). The EUPPBench postprocessing benchmark dataset v1.0. Earth System Science Data, 15, 2635-2653

How to cite: Lerch, S., Freytag, J., Muschinski, T., and Allen, S.: Enhancing member-by-member post-processing with neural networks, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-2005, https://doi.org/10.5194/egusphere-egu24-2005, 2024.

EGU24-2145 | ECS | Orals | NP5.2 | Highlight

Quantification of the practical predictability of thunderstorm occurrence using machine learning 

Kianusch Vahid Yousefnia, Tobias Bölle, Isabella Zöbisch, and Thomas Gerz

While the statistical post-processing of numerical weather prediction (NWP) data constitutes a powerful ingredient of many forecasting suites of severe weather, post-processing for thunderstorm occurrence becomes ever more difficult as the lead time of the NWP forecast increases. In terms of identifying thunderstorm occurrence as a function of lead time, this increased difficulty is reflected in a decay of skill for which even sophisticated machine learning (ML) models cannot fully compensate. In this work, we propose how the time scale of skill decay of supervised ML models can be studied as a function of the spatiotemporal label resolution used for training. If the label is constructed from lightning observations, label resolution is modified by varying the time and radius thresholds by which strokes of lightning are associated with NWP data. We exemplify our method using SALAMA, a feedforward neural network model which we have developed for identifying the probability of thunderstorm occurrence in NWP data. The model has been trained on convection-resolving ensemble forecasts over Central Europe and lightning observations. We show for SALAMA that the time scale for skillful thunderstorm predictions increases linearly with label resolution, which underlines the practical ability of our method to quantify the predictability of thunderstorm occurrence.

How to cite: Vahid Yousefnia, K., Bölle, T., Zöbisch, I., and Gerz, T.: Quantification of the practical predictability of thunderstorm occurrence using machine learning, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-2145, https://doi.org/10.5194/egusphere-egu24-2145, 2024.

EGU24-2361 | ECS | Posters on site | NP5.2

Precipitation forecast post-processing: blending deterministic NWPs with machine learning 

Luca Monaco, Roberto Cremonini, and Francesco Laio

Direct model output forecasts by Numerical Weather Prediction models (NWPs) present some limitations caused by errors mostly due to sensitivity to initial conditions, sensitivity to boundary conditions and deficiencies in parametrization schemes (i.e. orography).
These sources of error are unavoidable, and atmosphere chaotic dynamics makes prediction errors to spread rapidly in time in the course of the forecast, inducing both systematic and random errors.
Nonetheless, in the last 50 years NWPs had a significant decrease in the impact of these source of errors, even in the long-term forecast, thanks for instance to an ever-increasing computational capability, but still their relevance is not neglectable.
Moreover, different NWPs present specific different pros and cons which are findable empirically. For instance, in the case of precipitation forecast in the north-west Italy, low spatial resolution models (e.g. ECMWF-IFS) tend to be more reliable in terms of space and time in predicting the average precipitation, while high resolution models (e.g. COSMO-2I) tend to forecasts the maximum precipitation better. Research purposes apart, actual limitations must be seen in an operational context, where weather forecasts’ skillfulness and associated uncertainty are information of the utmost importance to the forecaster and in general to the user of a certain forecasts system.
In order to tackle the limitations of NWPs and the need of an uncertainty-quantified meteorological forecast, we propose a machine learning based multimodel post-processing technique for precipitation forecast. We focus on precipitation since it is the most important variable in the issue of spatially localized weather alert notice by the Italian Civil Protection’ system and at the same time it is one of the most challenging variables to forecast.

We use different Convolutional Neural Networks (CNNs) to obtain both deterministic and probabilistic forecast grids over 24h up to 48h focusing in the North-West Italy, using different high and low resolution deterministic NWPs as input and using high resolution rain-gauge corrected radar observations as ground truth for the training. We use constrainted linear regressions as a mean of deterministic benchmark, and ECMWF-EPS as a mean of probabilistic benchmark. The test phase show decent improvements in terms of RMSE for every season.

How to cite: Monaco, L., Cremonini, R., and Laio, F.: Precipitation forecast post-processing: blending deterministic NWPs with machine learning, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-2361, https://doi.org/10.5194/egusphere-egu24-2361, 2024.

EGU24-2794 | ECS | Orals | NP5.2

Improving Distributional Regression Forests for Post-processing Extreme Wind Gust Forecasts  

Bastien François, Harun Kivril, Maurice Schmeits, and Kirien Whan

Extreme events, such as wind gusts or extreme precipitation, can generate huge impacts on our society. Accurate predictions of such events are thus vital for taking preventive measures. In spite of continued scientific progress in weather forecasting, ensemble forecasts exhibit biases and under-dispersion and have to be calibrated using observations before being used, e.g., in hydrological or renewable energy applications. Several post-processing techniques have therefore been developed and applied over the last decades in order to improve forecast quality. Many existing post-processing methods are parametric, i.e. they assume that the predictive distribution belongs to a class of known probability distributions. Parameters of the assumed distribution are then modeled as functions of predictors obtained from numerical weather prediction models, for example using nonhomogeneous regression or more advanced tree-based methods. One of the main limitations of such methods is that a suitable family of probability distributions has to be selected to describe the distribution of the target variable. This implies that intermediate and high values are modeled with the same parametric distribution, which can lead to suboptimal results for extremes. We propose to adapt an existing distributional tree-based technique (Distributional Regression Forests) used for ensemble post-processing to overcome this limitation by allowing the method to choose different statistical distributions to model intermediate and extreme values. The proposed method is applied to forecasts of 6-hourly maximum wind gusts from 2018 to 2022 over the Netherlands using the ECMWF-IFS ensemble data. Results are compared against several state-of-the-art parametric and non-parametric post-processing methods. In comparison with these alternatives, the proposed algorithm reasonably corrects intermediate values and presents the largest skill improvements for wind gust extremes depending on lead times, stations and thresholds. However, it remains difficult to beat the raw forecasts of extremes. Therefore, it encourages further research on adding more flexibility to parametric methods for the post-processing of extreme weather forecasts.

How to cite: François, B., Kivril, H., Schmeits, M., and Whan, K.: Improving Distributional Regression Forests for Post-processing Extreme Wind Gust Forecasts , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-2794, https://doi.org/10.5194/egusphere-egu24-2794, 2024.

EGU24-2861 | ECS | Orals | NP5.2

D-Vine GAM Copula based Quantile Regression with Application to Ensemble Postprocessing 

David Jobst, Annette Möller, and Jürgen Groß

Temporal, spatial or spatio-temporal probabilistic models are frequently used for weather forecasting. The D-vine (drawable vine) copula based quantile regression (DVQR) is a powerful tool for this application field, as it incorporates important predictor variables from a large set by a data-driven sequential forward selection procedure and is able to model complex nonlinear relationships among them. However, the current DVQR does not always explicitly and economically allow to account for additional covariate effects, e.g.  temporal or spatio-temporal information. Consequently, we propose an extension of the current DVQR, where we parametrize the bivariate copulas in the D-vine copula through Kendall's Tau which can be linked to additional covariates. The parametrization of the correlation parameter allows generalized additive models (GAMs) and spline smoothing to detect potentially hidden covariate effects. The new method is called GAM-DVQR, and its performance is illustrated in a case study for the postprocessing of 2m surface temperature forecasts. We investigate a constant as well as a time-dependent Kendall's Tau. The GAM-DVQR models are compared to the benchmark method gradient-boosted Ensemble Model Output Statistics (EMOS-GB). The results indicate that the GAM-DVQR models are able to identify time-dependent correlations as well as relevant predictor variables and significantly outperform the state-of-the-art method EMOS-GB. Furthermore, the introduced parameterization allows using a static training period for GAM-DVQR, yielding a more sustainable model estimation in comparison to DVQR using a sliding training window. 

How to cite: Jobst, D., Möller, A., and Groß, J.: D-Vine GAM Copula based Quantile Regression with Application to Ensemble Postprocessing, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-2861, https://doi.org/10.5194/egusphere-egu24-2861, 2024.

EGU24-4357 | Orals | NP5.2

Autoregressive extensions of EMOS with application to surface temperature ensemble postprocessing 

Annette Möller, David Jobst, and Jürgen Groß

Two extensions of the autoregressive EMOS (AR-EMOS) which are based on the idea of smooth EMOS (SEMOS) model are proposed: The de-seasonalized EMOS (DAR-SEMOS) approach models time series behavior in the mean and variance of the predictive distribution separately, the standardized AR-SEMOS (SAR-SEMOS) method attempts to incorporate both effects jointly by fitting a time series model to the standardized forecast errors. The proposed modifications both allow to incorporate seasonal and trend effects as well as autoregressive behavior into the mean and variance parameter of the predictive distribution. Due to this explicit modelling of seasonal and trend behavior a rolling training period is not required anymore, and a longer (static) training period can be utilized for model fitting. The extended models can postprocess ensemble forecasts with arbitrary forecast horizons. In a case study for 2m surface temperature the extensions DAR- and SAR-SEMOS yield substantial improvements over AR-EMOS and SEMOS, for all considered forecast horizons and at the majority of observations stations. Overall, the SAR-SEMOS model yields the most noticeable improvements. At the same time its seamless approach of jointly modelling the time series behavior in the mean and variance parameter makes it appealing for practical and possibly operational use.

How to cite: Möller, A., Jobst, D., and Groß, J.: Autoregressive extensions of EMOS with application to surface temperature ensemble postprocessing, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-4357, https://doi.org/10.5194/egusphere-egu24-4357, 2024.

By the end of 2022, the renewable energy share of the global electricity capacity reached 40.3% and the new installations were dominated by solar energy, showing a global increase of 21.7%. Due to the high volatility of photovoltaic energy sources, their successful integration into the electrical grid requires accurate short-term power forecasts. These forecasts are obtained from the predictions of solar irradiance, where the most advanced method is the probabilistic approach based on ensemble forecasts.  However, ensemble forecasts are often underdispersive and subject to systematic bias. Hence, they require some form of statistical post-processing, where parametric models provide full predictive distributions of the weather variables at hand.

We propose a general two-step machine learning-based approach to calibrating ensemble weather forecasts, where, in the first step, improved point forecasts are generated, which then together with various ensemble statistics serve as input features of the neural network estimating the parameters of the predictive distribution [1]. In a case study based on global horizontal irradiance forecasts of the operational ensemble prediction system of the Hungarian Meteorological Service, the predictive performance of this novel method is compared with the forecast skill of the raw ensemble and the state-of-the-art ensemble model output statistics approaches [2]. We show that at least up to 48h, statistical post-processing substantially improves the predictive performance of the raw ensemble for all forecast horizons considered; the maximal gain e.g. in terms of the mean continuous ranked probability score is above 20%. Furthermore, the proposed two-step machine learning-based approach outperforms in skill its competitors.

References

1.  Baran, Á, Baran, S., A two-step machine-learning approach to statistical post-processing of weather forecasts for power generation.   Q. J. R. Meteorol. Soc. (2023), doi:10.1002/qj.4635.

2.  Schulz, B., El Ayari, M., Lerch, S., Baran, S., Post-processing numerical weather prediction ensembles for probabilistic solar irradiance forecasting. Sol. Energy 220 (2021), 1016-1031.

*Research is supported by the Hungarian National Research, Development and Innovation Office under grant no. K142849

How to cite: Baran, S. and Baran, Á.: Machine learning-based parametric post-processing of solar irradiance ensemble forecasts, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-4404, https://doi.org/10.5194/egusphere-egu24-4404, 2024.

EGU24-4610 | ECS | Posters on site | NP5.2

Statistical post-processing and generation of spatially correlated precipitation forecasts with convolutional neural networks 

Syed Ahmad Siffat, Quan Jun Wang, Kirien Whan, and Erik Weyer

The raw forecasts from a numerical weather prediction (NWP) model cannot be directly used because of systematic biases. Statistical calibration is performed to produce reliable and accurate ensemble forecasts. However, this is usually done on a grid-cell by grid-cell basis, followed by the use of empirical copula to embed a realistic spatial structure in the calibrated ensemble members. One drawback of these approaches is that it is difficult to select the empirical copula. In this paper, we propose Convolutional Neural Network (CNN) based models for post-processing precipitation forecast fields and generating ensemble forecasts. Unlike the traditional approaches which are applied to individual grid-cells, the model is applied to the whole precipitation field. Monte-Carlo (MC) dropouts are used to estimate uncertainty and generate ensemble forecasts. These ensemble forecasts preserve the inherent spatial structure, thereby eliminating the need for ensemble reordering. The model is applied to NWP forecasts of Brisbane drainage basin in eastern Australia. It is evaluated on all precipitation events, including no, low and high precipitation amounts. The results show that, for all levels of precipitation, the ensemble forecasts are skillful at both the grid-cell and basin scale, and the uncertainty is estimated reliably.

How to cite: Siffat, S. A., Wang, Q. J., Whan, K., and Weyer, E.: Statistical post-processing and generation of spatially correlated precipitation forecasts with convolutional neural networks, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-4610, https://doi.org/10.5194/egusphere-egu24-4610, 2024.

EGU24-5541 | ECS | Posters on site | NP5.2

Machine learning-based discrete post-processing of visibility ensemble forecasts 

Maria Nagy-Lakatos and Sandor Baran

In aviation meteorology, as well as in water and road transportation, the accurate and
reliable prediction of visibility is of utmost importance. Despite various meteorological
services offering ensemble forecasts for visibility, the predictive accuracy and reliability for
this parameter are notably lower compared to variables like temperature or wind speed.
Therefore, it is strongly recommended to implement some form of calibration, typically
involving the estimation of the predictive distribution through parametric or non-parametric
methods, including machine learning techniques. The World Meteorological Organization
suggests that visibility observations should be reported in discrete values, turning the
predictive distribution into a discrete probability law. Consequently, the calibration process
can be simplified to a classification problem. This study investigates the predictive
performance of locally, semi-locally, and regionally trained proportional odds logistic
regression (POLR) and multilayer perceptron (MLP) neural network classifiers using
visibility ensemble forecasts from the European Centre for medium-range weather
forecasts. The findings reveal that while climatological forecasts surpass the raw
ensemble, post-processing leads to a substantial improvement in forecast skill. Overall,
POLR models exhibit superiority over their MLP counterparts.

Reference
Baran, S., Lakatos, M., Statistical post-processing of visibility ensemble forecasts.
Meteorol. Appl. 30 (2023), paper e2157, doi:10.1002/met.2157.

*Research is supported by the ÚNKP-23-3 New National Excellence Program of the
Hungarian Ministry for Culture and Innovation from the source of the National Research,
Development and Innovation Fund and the Hungarian National Research, Development
and Innovation Office under Grant No. K142849.

How to cite: Nagy-Lakatos, M. and Baran, S.: Machine learning-based discrete post-processing of visibility ensemble forecasts, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-5541, https://doi.org/10.5194/egusphere-egu24-5541, 2024.

EGU24-6281 | ECS | Posters on site | NP5.2

Tail calibration of probabilistic forecasts 

Sam Allen, Jonathan Koh, and Johanna Ziegel

Probabilistic forecasts comprehensively describe the uncertainty in the unknown future outcome, making them essential for decision making and risk management. While several methods have been introduced to evaluate probabilistic forecasts, existing evaluation techniques are ill-suited to the evaluation of forecasts for extreme events, which are often of particular interest due to the impact they have on forecast users. In this work, we reinforce previous results related to the deficiencies of proper scoring rules when evaluating forecasts for extreme outcomes, demonstrating that classes of scoring rules cannot distinguish between forecasts with the incorrect tail behaviour. Alternative methods to evaluate forecasts for extreme events are therefore required. To this end, we introduce several notions of tail calibration for probabilistic forecasts, which allow forecasters to assess the reliability of their predictions for extreme outcomes. We study the relationships between these different notions, and provide several examples. We then demonstrate how these tools can be applied in practice by implementing them in a case study on European precipitation forecasts.

How to cite: Allen, S., Koh, J., and Ziegel, J.: Tail calibration of probabilistic forecasts, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-6281, https://doi.org/10.5194/egusphere-egu24-6281, 2024.

EGU24-6959 | ECS | Posters on site | NP5.2

Can neural networks outperform quantile mapping for post-processing seasonal weather forecast variables over the Alpine region? 

Sameer Balaji Uttarwar, Sebastian Lerch, Diego Avesani, and Bruno Majone

The possibility to use seasonal weather forecasts is of paramount importance in hydrological and socio-economical applications. However, current seasonal weather forecasts from global numerical weather prediction (NWP) models inherit systematic biases resulting from inaccurate representation and parameterization of local to global scale environmental processes. Therefore, the hydrological community frequently uses the quantile mapping (QM) statistical postprocessing for bias correction and downscaling of the meteorological inputs (i.e., daily precipitation and temperature) to hydrological models. The QM often assumes a linear and static relationship between quantiles of observed and simulated data over time. These limitations can be relaxed by employing a Neural Network (NN) based postprocessing method. In this context, the objective of this study is to compare the accuracy of QM and NN statistical postprocessing of ensemble seasonal weather forecasts over the Trentino-South Tyrol region (north-eastern Italian Alps), characterised by complex topography. 

The study uses the latest fifth-generation seasonal weather forecast system (SEAS5) total precipitation and 2m-temperature dataset produced by European Centre for Medium-Range Weather Forecast (ECMWF), available at a horizontal grid resolution of 0.125° x 0.125° with 25 ensemble members in a re-forecast period from 1981 to 2016. The respective reference dataset is a high-resolution gridded observation (250 m x 250 m) over the region of interest. The QM method derives a functional relationship between the variable of interest and the corresponding predictor, whereas the NN based methods can be used with a set of predictors to learn the linear and non-linear relationships in a data-driven way.

The analysis is divided into training (1981 – 2010, 30 years) and testing (2011 – 2016, 6 years) period to compare the cumulative ranked probability scores (CRPS) of both the statistical postprocessing methods. The statistical postprocessing is implemented univariately on the daily dataset (2m temperature and precipitation) over a month for each lead time. The raw forecasts and postprocessed forecasts are compared with particular focus on the effects of the forecast lead time and location, as well as diurnal and seasonal cycles in forecast performance. The postprocessed forecasts revealed a significant improvements compared to the raw forecasts.

How to cite: Uttarwar, S. B., Lerch, S., Avesani, D., and Majone, B.: Can neural networks outperform quantile mapping for post-processing seasonal weather forecast variables over the Alpine region?, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-6959, https://doi.org/10.5194/egusphere-egu24-6959, 2024.

EGU24-7702 | ECS | Posters on site | NP5.2

On statistical calibration of dual-resolution precipitation forecasts 

Marianna Lakatos-Szabó, Estíbaliz Gascón, and Sándor Baran

Recently, all leading meteorological centers release ensemble forecasts that vary in terms of ensemble size and spatial resolution, even when covering the same area. These factors significantly impact the forecast accuracy and computational resources required. In the last few years, the plans of upgrading the configuration of the Integrated Forecast System of the European Centre for Medium-Range Weather Forecasts (ECMWF) from a single forecast with 9 km resolution and a 51-member ensemble with 18 km resolution induced an extensive study of the forecast skill of both raw and post-processed dual-resolution predictions comprising ensemble members of different horizontal resolutions.

We investigate the predictive performance of the censored shifted gamma (CSG) [1] ensemble model output statistic (EMOS) approach for statistical post-processing with the help of dual-resolution 24h precipitation accumulation ensemble forecasts over Europe with various forecast horizons. The high-resolution operational 50-member ECMWF ensemble is supplemented by a 200-member low-resolution (29-km grid) experimental forecast. The various dual-resolution combinations, which are equivalent in computational cost to the operational ensemble, show improved forecast skill after EMOS post-processing compared with raw ensemble combinations [3]. Additionally, the differences between these combinations are significantly reduced as a result of this post-processing technique. Moreover, the semi-locally trained CSG EMOS is fully able to catch up with the state-of-the-art quantile mapping [2] and provides an efficient alternative without requiring additional historical data essential in determining the quantile maps.

References:

[1] Baran, S. and Nemoda, D. (2016). Censored and shifted gamma distribution based EMOS model for probabilistic quantitative precipitation forecasting. Environmetrics 27, 280–292.
[2] Gascón, E., Lavers, D., Hamill, T. M. , Richardson, D. S., Ben Bouallègue, Z., Leutbecher, M. and Pappenberger, F. (2019). Statistical postprocessing of dual-resolution ensemble precipitation forecasts across Europe. Quart. J. Roy. Meteor. Soc. 145, 3218–3235.
[3] Szabó, M., Gascón, E. and Baran, S. (2023) Parametric post-processing of dual-resolution precipitation forecasts. Weather Forecast., 38(8), 1313–1322.

How to cite: Lakatos-Szabó, M., Gascón, E., and Baran, S.: On statistical calibration of dual-resolution precipitation forecasts, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-7702, https://doi.org/10.5194/egusphere-egu24-7702, 2024.

The new programs of EUMETNET have now been launched for 5 years (2024-2028). Within this context, new activities on statistical postprocessing have been proposed, supported by 16 National Meteorological Services. One key activity that will continue is the benchmark of different statistical techniques for correcting weather forecasts. Another one is the development of ready-to-use ensemble calibration techniques that will be available to the community, in particular using machine learning techniques. Finally, several workshops will be organized on the topic during the duration of the project. In this poster, we will discuss the past achievement on statistical postprocessing using the benchmark developed in the context of the previous phase of the project, the current activities, and the future plans in comparing the statistical methods.

How to cite: Vannitsem, S. and Demaeyer, J.: The past, current and future activities on statistical postprocessing in the context of the European Meteorological Network (EUMETNET) , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-7878, https://doi.org/10.5194/egusphere-egu24-7878, 2024.

EGU24-8137 | Orals | NP5.2

Temperature Forecasting with Markov Expert Aggregations 

Olivier Wintenberger, Leo Pfitzner, and Olivier Mestre

A multitude of Numerical Weather Prediction (NWP) models, along with their associated Model Output Statistics (MOS), are readily available. Expert Aggregation (EA) algorithms combine them in an online and adaptive manner. While EA competes optimally against the best-fixed combination of experts (Wintenberger 2017), it falls short in handling rapid changes. We introduce the class of Markov-EA algorithms, extending the seminal work of Mourtada and Maillard (2017) on Exponentiated Weights to other EA algorithms such as BOA and ML-Poly. Understanding how and when to adjust the weights is crucial for obtaining optimal second-order regret bounds. Assuming a (non-homogeneous) Markovian dynamic, we enhance the EA predictions of short and poorly predicted events, such as the cold event in the Chamonix valley, using weight sharing and strategies involving sleeping experts. This work is done in collaboration with Leo Pfitzner and Olivier Mestre (Météo France).

How to cite: Wintenberger, O., Pfitzner, L., and Mestre, O.: Temperature Forecasting with Markov Expert Aggregations, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-8137, https://doi.org/10.5194/egusphere-egu24-8137, 2024.

EGU24-8807 | ECS | Orals | NP5.2

The fractions skill score for ensemble forecast verification 

Ludwig Wolfgruber, Tobias Necker, Lukas Kugler, Martin Weissmann, Manfred Dorninger, and Stefano Serafin

This work explores how the Fractions Skill Score (FSS), originally developed for deterministic forecasts of binary events, can be used for probabilistic forecast verification. By comparing a selection of four ensemble-based methods to compute the FSS, we highlight their distinct behaviour with ensemble size, neighbourhood size, and frequency of occurrence of the forecast event. Our study emphasizes that only a specific variant of the FSS, which we refer to as "probabilistic FSS", demonstrates reasonable behaviour with ensemble size. We reveal that the probabilistic FSS depends on ensemble size in a similar way as the Brier Skill Score, despite performing a neighbourhood-based instead of a grid-point-based forecast evaluation. We derive a formula that describes the expected behaviour of the probabilistic FSS with changes in ensemble size. Finally, utilizing a unique dataset of high-resolution 1000-member ensemble precipitation forecasts for Germany, we explore the impact of ensemble and neighbourhood size on the predictive skill by studying various subsamples of the full ensemble.

How to cite: Wolfgruber, L., Necker, T., Kugler, L., Weissmann, M., Dorninger, M., and Serafin, S.: The fractions skill score for ensemble forecast verification, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-8807, https://doi.org/10.5194/egusphere-egu24-8807, 2024.

EGU24-9326 | ECS | Orals | NP5.2

Leveraging deterministic weather forecasts for in-situ probabilistic predictions via deep learning 

David Landry, Anastase Charantonis, and Claire Monteleoni

We propose a neural network approach to produce probabilistic weather forecasts from a deterministic numerical weather prediction. The developed method is applicable to any gridded forecast including the recent machine learning weather prediction model outputs. To postprocess multiple lead times using a single model, we introduce a lead time embedding that encodes the shift in biases as the forecast progresses. We apply our approach to operational outputs from the Global Deterministic Prediction System up to ten-day lead times. The model is trained to predict METAR in-situ surface temperature observations in Canada and the United States. The resulting forecasts have a mean CRPS below 2.5 K at 10 days lead time while maintaining a spread-error ratio of approximately 0.9, suggesting appropriate calibration. For extreme temperatures, the model’s biases are comparable to that of the underlying deterministic forecast. Our approach increases the utility of a deterministic forecast by adding information about the uncertainty, without incurring the cost of simulating multiple trajectories. It requires no information regarding forecast spread and can be used to generate probabilistic predictions from any deterministic forecast.

How to cite: Landry, D., Charantonis, A., and Monteleoni, C.: Leveraging deterministic weather forecasts for in-situ probabilistic predictions via deep learning, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-9326, https://doi.org/10.5194/egusphere-egu24-9326, 2024.

EGU24-10787 | ECS | Posters on site | NP5.2

Statistical Post-Processing for Wind Gust and Precipitation Extremes: Insights from a Pre-Operational System 

Harun Kıvrıl, Bastien François, Maurice Schmeits, Kirien Whan, Eva van der Kooij, and Antonello Squintu

Effective forecasting of weather events, especially extremes, is critical for minimizing potential damage and ensuring public safety. Yet, ensemble forecasts that allow to represent uncertainty in weather predictions like ECMWF-ENS suffer from biases and dispersion issues. These shortcomings decrease the forecasting skill and introduce the need for statistical post-processing methods like Ensemble Model Output Statistics (EMOS) and Quantile Regression Forest (QRF) to enhance the forecast quality. While these methods increase overall forecasting performance in general, their capability to handle extreme events varies. To strengthen the forecasting of these critical events, a closer examination and potential refinement of the statistical post-processing methods are necessary, along with the development of methods tailored to extreme weather events. This study analyzes the performance of pre-operational post-processing models of ECMWF-ENS wind gust and precipitation forecasts in the Netherlands using EMOS and/or QRF techniques. The skill of the models (using Continuous Probability Ranked Skill Score (CRPSS) and Brier Skill Score (BSS)) for both wind gusts and precipitation is demonstrated. Besides, by focusing on windstorm Poly (July 5th, 2023) and a heavy rain case (June 22nd, 2023), the capabilities of the methods in forecasting specific extreme events are investigated. While the QRF forecasts for the heavy rain case show better skill than for the raw forecasts, for the Poly storm case the post-processed models performed worse than the raw forecasts initially. Thus, this work concentrates on the underlying reasons for the limited performance on that event and proposes an error modeling approach to enhance the post-processing performance of the storm event without compromising the overall forecasting performance. 

How to cite: Kıvrıl, H., François, B., Schmeits, M., Whan, K., van der Kooij, E., and Squintu, A.: Statistical Post-Processing for Wind Gust and Precipitation Extremes: Insights from a Pre-Operational System, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-10787, https://doi.org/10.5194/egusphere-egu24-10787, 2024.

EGU24-10797 | ECS | Posters on site | NP5.2 | Highlight

Improving probabilistic wind speed forecasts with deep learning 

Katharina Klein, Daniel Tolomei, Sjoerd Dirksen, Kirien Whan, and Maurice Schmeits

This project aims at developing post-processing models for deriving probabilistic weather forecasts from NWP forecast data using deep learning techniques.

The first part of the project involves improving probabilistic wind speed forecasts for consecutive lead times using an autoregressive model. Post-processing multiple lead times simultaneously is challenging because of the inherent temporal dependencies. Classical approaches often involve processing lead times individually and subsequently employing empirical copula methods to handle such dependencies. Building on previous work, we instead consider the ARMOS model which incorporates temporal dependencies through the autoregressive property of forecast errors and can be used to obtain an explicit multivariate probability distribution for the weather variable in question. As such, it is a generalization of the widely used Ensemble Model Output Statistics (EMOS) used for estimating marginal distributions.

For the purpose of this project, the model is applied to deterministic forecasts from the Harmonie-Arome model of KNMI, yielding a multivariate parametric forecast distribution for hourly wind speeds up to 48 hours ahead. We model the marginal conditional distributions as truncated normal distributions, and the model parameters are estimated both linearly and as the output of a neural network with convolutional and optional LSTM layers which can detect spatial patterns and temporal dependencies, respectively. We compare the resulting models to a variant of EMOS adapted to deterministic forecasts that is paired with a copula method.

The ARMOS model has so far shown good performance in modeling temporal dependencies explicitly without the need to use a copula method. Moreover, the network models outperform the classical approach of estimating the distribution parameters linearly. We will provide an update on these results as well as an outlook on planned future work.

How to cite: Klein, K., Tolomei, D., Dirksen, S., Whan, K., and Schmeits, M.: Improving probabilistic wind speed forecasts with deep learning, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-10797, https://doi.org/10.5194/egusphere-egu24-10797, 2024.

EGU24-11096 | Orals | NP5.2

The Environmental Modeling Center Verification System (EVS): Real-time verification of Unified Forecast System (UFS) models 

Jason Levit, Geoffrey Manikin, Alicia Bentley, Logan Dawson, and Tara Jensen

Currently in the final stages of development, the Environmental Modeling Center Verification System (EVS) is a real-time software system that, once implemented on NCEP supercomputers, will provide verification statistics and graphics for NCEP operational forecast systems. Using the Model Evaluation Toolkit (METplus) suite of verification software, the EVS will use both real-time forecast output and environmental observations to create information on the performance of all NCEP environmental models and products, which are either derived or directly created from the Unified Forecast System (UFS) suite of models. The EVS will generate hundreds of metrics for the suite, which when viewed on EMC webpages will provide a comprehensive and up-to-date overview of the performance of NCEP forecast systems for the general public, model developers, researchers, and decision makers. As EMC works with the international modeling community to develop modeling systems that are based on the Unified Forecast System (UFS), EMC is now evolving towards using the EVS for a single, unified verification software system for real-time verification analysis, evaluation of new and upgraded systems proposed for NWS operations, and as a capability to determine systematic errors and biases that illustrate areas for potential model improvements. This project, led by EMC’s Verification, Post-Processing, and Product Generation Branch (VPPPGB) aims to unify EMC’s verification strategy under one maintainable and supportable software system, and to use performance metrics that have been vetted and peer-reviewed by the UFS community via the Developmental Testbed Center’s 2021 UFS Metrics Workshop. 

How to cite: Levit, J., Manikin, G., Bentley, A., Dawson, L., and Jensen, T.: The Environmental Modeling Center Verification System (EVS): Real-time verification of Unified Forecast System (UFS) models, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-11096, https://doi.org/10.5194/egusphere-egu24-11096, 2024.

EGU24-11394 | ECS | Orals | NP5.2

Attention-based postprocessing of ensemble weather forecasts for renewable energy applications by leveraging inter-ensemble relationships of multiple predictors 

Aaron Van Poecke, Ruoke Meng, Jonathan Demaeyer, Joris Van den Bergh, Geert Smet, Piet Termonia, Peter Hellinckx, and Hossein Tabari

Indirect models for renewable energy forecasting rely heavily on accurate weather predictions. Operational weather forecasting today is mainly based on numerical weather prediction models, often employing ensembles to estimate the day-to-day forecast uncertainty. To correct for errors due to simplifications in these models, inaccurate initial conditions, and representativeness problems, statistical postprocessing becomes necessary for these ensemble forecasts. Current postprocessing techniques often disregard possible inter-ensemble relationships by correcting each member separately, or employ a distributional approach that requires extra multivariate methods to restore spatio-temporal and inter-variable correlations. In this work, we tackle these shortcomings with an innovative, attention-based member-by-member approach which postprocesses each member individually while simultaneously integrating information from other ensemble members. Variables required for renewable energy forecasting are postprocessed at the station level by regressing ensemble forecasts of multiple predictors, including the forecasted variable itself, against observational data. The training data utilized is sourced from the EUPPBench dataset, which contains ensemble predictions from the integrated forecasting system of the ECMWF and corresponding observations. Transformer modules built around Self-Attention are employed to capture dependencies between different predictors, such as temperature and total cloud cover, next to significant relationships between the ensemble members themselves. Additionally, our model postprocesses the forecasts for all lead times simultaneously, taking into account the correlation between the postprocessed variable and  forecasts generated at earlier and later lead times. This results in postprocessing techniques that can be employed in downstream applications for conversion to renewable energy forecasts.

How to cite: Van Poecke, A., Meng, R., Demaeyer, J., Van den Bergh, J., Smet, G., Termonia, P., Hellinckx, P., and Tabari, H.: Attention-based postprocessing of ensemble weather forecasts for renewable energy applications by leveraging inter-ensemble relationships of multiple predictors, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-11394, https://doi.org/10.5194/egusphere-egu24-11394, 2024.

EGU24-11694 | Posters on site | NP5.2

Optimal blending of ensemble forecasts for probabilistic precipitation nowcasting 

Martin Weissmann, Simon Köhldorfer, Tobias Necker, and Alexander Kann

Precipitation nowcasting is essential for mitigating the negative effects of severe weather, necessitating accurate and timely forecasts. Our study introduces a novel Local Ensemble Transform Kalman Filter (LETKF)-based blending approach that effectively combines ensemble forecasts to provide cheap probabilistic nowcasts. Our method optimally weights ensemble probabilities by considering precipitation/radar observations as the ground truth. It shifts computed weights in time and refines neighborhood probabilities (NPs). The upscaling step for computing NPs introduces two free parameters, neighborhood size and precipitation rate, which can be selected based on the forecasters needs. Demonstrated within GeoSphere Austria's regional forecasting system using AROME ensemble forecasts over Austria, our technique was especially beneficial for short lead times of up to 2 hours. Longer lead times require to incorporate signal propagation, which is possible by shifting weights in space and time. Our study underscores the effectiveness of data assimilation techniques in enhancing ensemble blending. The proposed approach affordably improves the robustness and accuracy of short-term probabilistic forecasts and holds the potential for extending it to multi-model ensemble blending.

How to cite: Weissmann, M., Köhldorfer, S., Necker, T., and Kann, A.: Optimal blending of ensemble forecasts for probabilistic precipitation nowcasting, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-11694, https://doi.org/10.5194/egusphere-egu24-11694, 2024.

EGU24-16157 | ECS | Posters on site | NP5.2

Enhancing Renewable Energy Forecasting: A Comprehensive Evaluation of Weather Forecast Models and Post-Processing Methods for Belgium 

Ruoke Meng, Aaron Van Poecke, Geert Smet, Jonathan Demaeyer, Hossein Tabari, Peter Hellinckx, Joris Van den Bergh, and Piet Termonia

As renewable energy sources continue to account for an increasing proportion of Belgium's energy production, decision making in renewable energy production increasingly relies on accurate numerical weather prediction forecasts. For general applications, forecast validation often focuses on direct comparisons to observations for the whole domains of interest, while in this study we assess model performance specifically related to renewable energy productions. We perform extended verification of relevant variables (wind speed, temperature, solar radiation, etc.) from multiple high-resolution deterministic and ensemble weather forecast models operated in Belgium for the period of May 2021 - June 2023. The forecasts are verified with observational datasets collected from on- and offshore weather stations, masts, lidars, and wind farm observations to comprehensively understand the capabilities of the models, making use of various deterministic and probabilistic skill scores. The results show that during lead times up to two days, although verification metrics differ among models, there are systematic errors in their forecasts for different observation sites. Such errors can often be eliminated by post-processing techniques. Therefore, we extend our verification dataset, with post-processed forecasts corrected by several methods including member-by-member and AI-based approaches. The results of this work will lead to an enhanced understanding of current forecasting skills of the operational models, help to evaluate the effectiveness of goal-oriented post-processing methods, and provide a reference for Belgian sustainable energy stakeholders.

How to cite: Meng, R., Van Poecke, A., Smet, G., Demaeyer, J., Tabari, H., Hellinckx, P., Van den Bergh, J., and Termonia, P.: Enhancing Renewable Energy Forecasting: A Comprehensive Evaluation of Weather Forecast Models and Post-Processing Methods for Belgium, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-16157, https://doi.org/10.5194/egusphere-egu24-16157, 2024.

EGU24-18642 | Posters on site | NP5.2

Copula-based statistical post-processing for multi-site temperature forecasts 

Elisa Perrone, Maurits Flos, and Irene Schicker

Modern weather forecasts are typically in the form of an ensemble of forecasts obtained from multiple runs of numerical weather prediction models. Ensemble forecasts are usually biased and affected by dispersion errors, and they should be statistically corrected to gain accuracy. This is often done following a two-step approach: first, we correct the univariate forecasts, and then, we reconstruct the dependence structure non-parametrically via empirical copulas. The parametric correction of the dependence structure is limited to Gaussian copula-based methods. In this work, we propose a novel approach based on a more general parametric class of copulas called Archimedean copulas. We test the new method in both a simulated scenario and a case-study setting for multi-site temperature forecasts from the ALADIN-LAEF ensemble system in Austria. Our findings show that the state-of-the-art non-parametric techniques perform well in the simulation study. However, Archimedean copulas outperform the existing techniques, especially Gaussian copula approaches, and output well-calibrated forecasts in the real-case study. Our analysis demonstrates the usefulness of including advanced parametric copula methods in the post-processing context and the need of a more realistic simulated framework to test new methodology.

How to cite: Perrone, E., Flos, M., and Schicker, I.: Copula-based statistical post-processing for multi-site temperature forecasts, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-18642, https://doi.org/10.5194/egusphere-egu24-18642, 2024.

EGU24-19254 | Posters on site | NP5.2

Post-processing for an on-demand extremes digital twin – a multi-model approach for wind and solar energy production 

Irene Schicker, Petrina Papazek, Pascal Gfäller, Iris Odak Plenkovic, Ivan Vujec, Alexander Kann, and Kristian Horvath

 

The amount of wind and solar energy fed into the European power grid increases rapidly and with the transition to a fossil fuel-free energy production, relying heavily on renewable energy sources, more accurate predictions for both high-resolution temporal and spatial scales are needed to ensure, most of all, grid stability. This is even more the case with extreme events, both extremes in weather across the nowcasting to weeks ahead time scale and combined and non-necessarily extreme weather, events such as Dunkelflaute or longer lasting solar/wind droughts. Accurate, frequently updated and especially on-demand available predictions of the expected power production are needed. Post-processing methods enable targeted forecasts of meteorological parameters both at site-location and regional level, which can server for a conversion to power production, particularly a direct conversion of NWP predictions and observations to power production.

For an on-demand extreme digital twin forecasting system, fast and transferable post-processing methods, able to account for the upper/lower bounds of the respective distributions are needed. Furthermore, they need to be able to either generate on-the-fly (semi-)synthetic power production data or a reduced set of both observation and NWP input data. The latter is essential when moving towards hyper-resolution NWP simulations with only a limited set of training data available.

Within the on-demand extreme digital twin initiative, several post-processing methods, statistical and (deep) machine learning, were implemented and applied to selected use cases for on/offshore wind and solar production extreme events. Here, we demonstrate (i) the capability of the Kalman filter, the analogs method, IrradPhyDNet, a sequence-2-sequence LSTM, Random Forest, and other machine learning/statistical methods in extreme event prediction, (ii) evaluate the methods skills by using heterogeneous (multiple NWP models, observations, climatologies, etc.) and varying length input data and real/(semi-)synthetic power data as target, and (iii) present a workflow for an on-demand prediction for wind and solar energy production including user-interaction.

How to cite: Schicker, I., Papazek, P., Gfäller, P., Odak Plenkovic, I., Vujec, I., Kann, A., and Horvath, K.: Post-processing for an on-demand extremes digital twin – a multi-model approach for wind and solar energy production, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-19254, https://doi.org/10.5194/egusphere-egu24-19254, 2024.

EGU24-1340 | ECS | PICO | NP5.3

The global interdependence patterns of extreme-rainfall events 

Zhen Su, Henning Meyerhenke, and Jürgen Kurths

As a powerful data-driven technology, the complex network paradigm has contributed significantly to the studies of spatio-temporal patterns of climate phenomena at different scales, such as El Niño–Southern Oscillation, Indian Ocean Dipole, and monsoon. In this work, we study the global extreme-rainfall patterns, which can potentially be used to improve the predictability of extreme events. The idea is to identify regions of similar extreme-rainfall patterns. For this, we propose a network-based clustering workflow which includes unsupervised learning. More precisely, this workflow combines consensus clustering and mutual correspondences. By applying this workflow to two satellite-derived precipitation datasets, we identify two main global interdependence structures of extreme rainfall, during boreal summer. These two structures are consistent and robust. From a climatological point view, they explicitly manifest the primary intraseasonal variability in the context of the global monsoon, in particular, the “monsoon jump” over both East Asia and West Africa, and the mid-summer drought over Central America and southern Mexico. We highlight the advantage of network-based clustering in (i) decoding the spatio-temporal patterns of climate variability and in (ii) the intercomparison of these patterns, especially regarding their spatial distributions over different datasets.

How to cite: Su, Z., Meyerhenke, H., and Kurths, J.: The global interdependence patterns of extreme-rainfall events, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-1340, https://doi.org/10.5194/egusphere-egu24-1340, 2024.

EGU24-6251 | ECS | PICO | NP5.3 | Highlight

An evolving network approach to assess compounding heat and dry extremes in Europe. 

Domenico Giaquinto, Giorgia Di Capua, Warner Marzocchi, and Jürgen Kurths

The probability of incidence of compound events is increasing due to human-induced climate change: in particular, there is high confidence that concurrent heatwaves and droughts will become more frequent with increased global warming1. Hereby, understanding the aggregated impact of multiple and synchronized compound hot and dry events at different spatial regions is a pressing issue, especially when it comes to predicting these extremes. In order to assess the evolution of these climate hazards, it is crucial to identify the synchronization structures of compound hot and dry events. To achieve this goal,  we highlight the hotspot regions where extremes are increasing and analyse the atmospheric precursors driving these anomalous conditions. Complex networks represent a promising tool in this perspective. In this work, we present an evolving network approach to assess the time evolution of synchronized compound hot and dry extremes due to global warming in continental Europe. Under this framework, we identify those regions where the frequency of these events has increased in the past 80 years and we describe their atmospheric drivers. Using ERA5 reanalysis data2 and focusing on the extended summer seasons (from April to September) of the period 1941-2020, we construct an evolving network constituted by 51 consecutive layers. Each layer models the synchronization structure in space of compound hot and dry events for a certain time window. Once the evolving network is established, the 51 layers are analysed to highlight the main changes in the graph structure. In particular, by looking at different centrality and clustering metrics and their evolution, we identify hotspot regions, and consequently we describe the atmospheric conditions which drive the compound events at these key locations. Climate complex networks prove to be a powerful tool to reveal hidden features of climate processes; this approach indeed brings out key aspects concerning the spatial dynamics of hot and dry events, laying the foundations to build a forecasting method for these extremes.

References

1) S.I. Seneviratne, X. Zhang, M. Adnan, W. Badi, C. Dereczynski, A. Di Luca, S. Ghosh, I. Iskandar, J. Kossin, S. Lewis, et al. Weather and climate extreme events in a changing climate; climate change 2021: The physical science basis. contribution of working group i to the sixth assessment report of the intergovernmental panel on climate change, 2021.

2) H. Hersbach, B. Bell, P. Berrisford, G. Biavati, A. Horányi, J. Muñoz Sabater, J. Nicolas, C. Peubey, R. Radu, I. Rozum, D. Schepers, A. Simmons, C. Soci, D. Dee, and J-N. Thépaut. Era5 hourly data on single levels from 1940 to present. Copernicus Climate Change Service (C3S) Climate Data Store (CDS), 2023.

How to cite: Giaquinto, D., Di Capua, G., Marzocchi, W., and Kurths, J.: An evolving network approach to assess compounding heat and dry extremes in Europe., EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-6251, https://doi.org/10.5194/egusphere-egu24-6251, 2024.

EGU24-7007 | PICO | NP5.3

Fractional Integral Statistical Model: A new way for climate prediction and projection from the perspective of scaling 

Naiming Yuan, Christian Franzke, Da Nian, Zuntao Fu, Kairan Ying, Feilin Xiong, and Wenjie Dong

It is well recognized that climate memory is one the origins for climate predictability, but how to include the concept of climate memory into the climate prediction, is still an open question. Here in this work, we suggest the Fractional Integral Statistical Model (FISM), a generalized stochastic climate model, as a new way for this purpose. With FISM, one can extract the “forcing-induced direct component ε(t)” and the “memory-induced indirect component M(t)” from a given variable x(t). By predicting ε(t), one can further obtain the predicted x(t) using FISM. Different from traditional prediction approaches which normally focus on x(t), here this new strategy based on FISM clarifies the climate memory impacts. From this new perspective, we have quantified the climate memory induced predictability, and developed a temperature response model that can project the future warming trend. Compared to CMIP6 simulations, our approach projects lower global warming levels over the next few decades. A further examination indicates that many CMIP6 models overestimated the climate memory, which might contribute to the overestimated future warming trend.

How to cite: Yuan, N., Franzke, C., Nian, D., Fu, Z., Ying, K., Xiong, F., and Dong, W.: Fractional Integral Statistical Model: A new way for climate prediction and projection from the perspective of scaling, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-7007, https://doi.org/10.5194/egusphere-egu24-7007, 2024.

EGU24-10026 | PICO | NP5.3 | Highlight

Modeling the number of hospital admissions for malaria in South Africa by using climate variables as disease drivers 

Suzana Blesic, Milica Tosic, Neda Aleksandrov, Thandi Kapwata, and Caradee Wright

Recently we proposed a regression model for the number of hospital admissions for malaria in the Limpopo province of South Africa. We developed our model using the available weekly epidemiological reports from five districts in this province, in the period 2000-2020. We analyzed number of hospitalizations for malaria time series in relation to time series of temperature, rainfall and evaporation from bare soil ground or satellite data from the same geographical area and developed an algorithm that links combined changes in these three variables with the changes in number of malaria hospitalizations. We used wavelet spectral analysis to determine time lags in their cross-correlations.  

We used this model to provide projections for the Limpopo malaria cases for the next five years (2025-2029). Since there are no future projections available for evapotranspiration, we used three different methods to estimate future values of this variable in our model: 1) a combination of temperature and rainfall data, 2) use of total soil moisture content records and their projections, and 3) use of Hargreaves empirical formula. We will present and compare our results for all three cases.

Our calculations can be used for public health preparedness.  

How to cite: Blesic, S., Tosic, M., Aleksandrov, N., Kapwata, T., and Wright, C.: Modeling the number of hospital admissions for malaria in South Africa by using climate variables as disease drivers, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-10026, https://doi.org/10.5194/egusphere-egu24-10026, 2024.

EGU24-12314 | ECS | PICO | NP5.3

Spatially coherent structure of forecast errors – A complex network approach 

Shraddha Gupta, Abhirup Banerjee, Norbert Marwan, David Richardson, Linus Magnusson, Jürgen Kurths, and Florian Pappenberger

The quality of weather forecasts has improved considerably in recent decades as models can better represent the complexity of the Earth’s climate system, benefitting from assimilation of comprehensive Earth observation data and increased computational resources. Analysis of errors is an integral part of numerical weather prediction to produce better quality forecasts. The Earth’s climate, being a highly complex interacting system, often gives rise to significant statistical relationships between the states of the climate at distant geographical locations. Likewise, correlated errors in forecasting the state of the system can arise from predictable relationships between forecast errors at various regions resulting from an underlying systematic or random process. Estimation of error correlations is very important for producing quality forecasts and is a key issue for data assimilation. However, the size of the corresponding correlation matrix is larger than what is possible to represent on geographical maps in order to diagnose its full spatial variation.

In this work, we propose an approach based on complex network theory to quantitatively study the spatiotemporal coherent structures of medium-range forecast errors of different climate variables. We demonstrate that the spatial variation of the network measures computed from the error correlation matrix can provide insights into the origin of forecast errors in a climate variable by identifying spatially coherent patterns of regions having common sources of error. Notably, the network topology of forecast errors of a climate variable is significantly different from those of random networks corresponding to a deterministic phenomenon which the model fails to simulate adequately. This is especially important to reveal the spatial heterogeneity of the errors – for example, the forecast errors of outgoing long-wave radiation in tropical regions can be correlated across very long distances, indicating an underlying climate mechanism as the source of the error. Additionally, we highlight that these structures of forecast errors may not always be directly derivable from the spatiotemporal co-variability pattern of the corresponding climate variable, contrary to the expectations that the patterns should resemble each other. We further employ other common statistical tools such as, empirical orthogonal functions, to support these findings. Our results underline the potential of complex networks as a very promising diagnostic tool to gain better understanding of the spatial variation, origin, and propagation of forecast errors.

 

How to cite: Gupta, S., Banerjee, A., Marwan, N., Richardson, D., Magnusson, L., Kurths, J., and Pappenberger, F.: Spatially coherent structure of forecast errors – A complex network approach, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-12314, https://doi.org/10.5194/egusphere-egu24-12314, 2024.

EGU24-13178 | PICO | NP5.3

Reliable El Niño forecasting before the spring predictability barrier 

Josef Ludescher, Armin Bunde, and Hans Joachim Schellnhuber

The El Niño Southern Oscillation (ENSO) is the most consequential driver of interannual global climate variability and can lead to extreme weather events like drought or flooding in various parts of the world. Current operational forecasts are hampered by the so-called spring predictability barrier (SPB), which makes forecasts before or during the boreal spring particularly challenging. 

In recent years, we developed several methods based on complex system science that can provide reliable El Niño forecasts well before the SPB, thus about doubling the pre-warning time. The first of these methods is based on a dynamical climate network (CN) consisting of nodes that are reanalysis grid points in the Pacific, and links between them, whose strength is characterized by the cross-correlations of the atmospheric surface temperatures at the grid points. In the calendar year before an El Niño event, the links between the eastern equatorial Pacific and the rest of the tropical Pacific tend to strengthen such that the average link strength exceeds a certain threshold. This property serves as a precursor to forecast the onset of El Niño events. In particular, the CN-based method has already provided 12 real-time forecasts, 11 of which turned out to be correct (p = 5.1*10-3). Here, we discuss an improvement of the CN method as well as the combination with other El Niño forecasting methods. 

Approaches based on information entropy and the zonal temperature gradient in the western Pacific provide additional forecasts with about 1 year lead time for the magnitude and the type of an upcoming El Niño event, respectively. Combining the three methods provides not only more information about an upcoming El Niño, particularly about the risk exposure of a given geographical location, but concurring forecasts can support each other and lead to higher overall confidence in the forecast. This was the case, for instance, at the end of 2022, when the combined method correctly forecasted a moderate-to-strong El Niño of eastern Pacific type for 2023.  

How to cite: Ludescher, J., Bunde, A., and Schellnhuber, H. J.: Reliable El Niño forecasting before the spring predictability barrier, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-13178, https://doi.org/10.5194/egusphere-egu24-13178, 2024.

We apply potential forecasting [1,2] to the WISE database that contains water accounts of European river basins [3]. We identify basins under stress and discuss various scenarios of water use. The complexity of water inflows and abstractions introduces sources of uncertainty that require analysis of geophysical, climatic, agricultural and social factors of water use, and this data represents an important case study for development of multivariate data science techniques. We report our findings and projections of hydrological dynamics in European regions.

[1] Livina et al, Physica A 2013

[2] Billuroglu & Livina, Journal of Failure Analysis and Prevention 2022

[3] WISE database, European Environmental Agency, https://www.eea.europa.eu/en/datahub

 

How to cite: Livina, V. N.: Potential forecasting of water accounts of European river basins, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-14571, https://doi.org/10.5194/egusphere-egu24-14571, 2024.

EGU24-15381 | ECS | PICO | NP5.3 | Highlight

Understanding and predicting the spread of Phlebotomine sand flies in Europe 

Danyang Wang, Anouschka Hof, Kevin Matson, and Frank van Langevelde

Climate change influences the transmission of vector-borne diseases by affecting the distribution and survival of disease vectors. Numerous diseases are transmitted by phlebotomine sand flies (SFs), including Leishmaniasis. Several major sand fly-borne diseases (SFBDs) are responsible for high global disease burdens and high socio-economic costs. In Europe, 22 known SF vector species are largely confined to the Mediterranean Basin, yet global warming is predicted to drive the spread of SFs to large areas of Europe in the 21th century, an effect likely to be exacerbated by anthropogenic variables. However, the constraints to the geographic distributions of SFs are not well understood. This study aims to increase the understanding of the drivers of the spatial distributions of SFs. To achieve this, we use species distribution modelling (SDM) to assess the role of climate, land-use and socio-economic drivers in shaping the geographic distributions of all endemic SF vectors in Europe. With this knowledge, we predict future hotspots of SFs in Europe. Our predictions are spatially explicit, scenario-based, and informative for surveillance efforts.

How to cite: Wang, D., Hof, A., Matson, K., and van Langevelde, F.: Understanding and predicting the spread of Phlebotomine sand flies in Europe, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-15381, https://doi.org/10.5194/egusphere-egu24-15381, 2024.

EGU24-17215 | ECS | PICO | NP5.3

Forecasting of Precipitation-Induced Landslides Using Atmospheric Rivers: Opportunities and Challenges 

Sara M. Vallejo-Bernal, Lisa Luna, Norbert Marwan, and Jürgen Kurths

Landslides are particularly costly disasters, causing about 4,500 fatalities and US$20 billion in damages worldwide each year. In Western North America, where intense and frequent precipitation events interact with complex topography and steep slopes, precipitation-induced landslides (PILs) are a serious geological hazard. Recently, it has been revealed that the majority of PILs in the region are triggered by precipitation from atmospheric rivers (ARs), transient channels of intense water vapor flux in the troposphere. However, the synoptic conditions differentiating landslide-triggering and non-triggering ARs remain unknown. In this study, we explore opportunities for improved landslide forecasting in Western North America using catalogs of land-falling ARs and PILs, along with ERA5 climatological data, from 1996 to 2018. First, we employ event synchronization, a non-linear measure specially tailored for event series analysis, to identify landslide-triggering ARs. Based on the AR-strength scale, which ranks ARs in levels from 1 to 5, we further characterize landslide-triggering ARs in terms of intensity and persistence. Subsequently, we spatially resolve the conditional probability of PIL occurrence given the detection of AR-attributed precipitation in the antecedent week, revealing the contribution of each AR level. Lastly, using hourly estimates of integrated water vapour transport, geopotential height, and precipitation at 0.25° spatial resolution, we differentiate the spatio-temporal evolution of synoptic conditions preceding landslide-triggering and non-landslide triggering ARs. Our results constitute a first, fundamental, and necessary step toward AR-based landslide forecasts, contributing crucial insights to improve forecasting accuracy at the short and early medium-range (1–7 days).

How to cite: Vallejo-Bernal, S. M., Luna, L., Marwan, N., and Kurths, J.: Forecasting of Precipitation-Induced Landslides Using Atmospheric Rivers: Opportunities and Challenges, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-17215, https://doi.org/10.5194/egusphere-egu24-17215, 2024.

Urban environments are especially sensitive to global warming due to their characteristic man-made surfaces and decreased vegetation cover. Elevated temperature in cities can facilitate the pole-wards expansion of arthropod disease vectors, including Phlebotomine sand flies (SFs). No study to date has yet been done to understand the effects of elevated urban temperatures on the distribution range shifts of SFs on continental scale. This study fills that gap and tests the role of urban heat island (UHI) in driving distribution range shifts of Phlebotomus perniciosus in Europe under two climatic scenarios. We find that P. perniciosus can occur more northly in summer due to UHI under both scenarios. Our study suggests that arthropod disease vectors can occur in cities where they are not expected due to UHI.

How to cite: Tak, V., Wang, D., and Matson, K.: The urban heat island effect aggravates the impact of climate change on the spatial distribution shifts of Phlebotomus perniciosus in Europe, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-17798, https://doi.org/10.5194/egusphere-egu24-17798, 2024.

The recurrence of similar states is a fundamental property of the processes that shape and influence our living and non-living world. There are numerous examples of geological and climatic processes on both short and long time and spatial scales, such as the regular activity of geysers within minutes, the more irregular but still recurrent occurrence of earthquakes (on time scales between weeks and years), the El Niño climate phenomenon occurring every three to five years, the glacial cycles (thousands of years), or the Milanković cycles, which periodically force climate changes up to hundreds of thousands of years. The recurrence of states in such dynamic processes generates typical recurrence patterns that can be used to detect regime changes, to classify the dynamics, or even to predict future changes. I will report on recent achievements in recurrence analysis in recent years, including methodological developments tailored for challenging data in the geosciences, such as irregularly sampled data or extreme event data. The overview includes further important and innovative developments, such as conceptual recurrence plots, ideas for parameter selection, multiscale recurrences, correction schemes, and new perspectives by combining recurrence analysis with machine learning.

How to cite: Marwan, N.: Advances in Recurrence Analysis for Predictive Modeling and Dynamic Classification, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-21587, https://doi.org/10.5194/egusphere-egu24-21587, 2024.

EGU24-21913 | PICO | NP5.3

Predicting strong local wind with high-resolution nonhydrostatic numerical weather prediction model 

Vladimir Djurdjevic, Milica Tosic, and Irida Lazic

The nonhydrostatic multiscale model on the B grid (NMMB) was employed to forecast an episode of intense local Kosava wind in northeast Serbia. Kosava, a vigorously turbulent local wind, originates from the east or southeast near the Danube's "Iron Gate," moves westward over Belgrade, and then extends northward into the regions of Romania and Hungary. Typically attributed to a jet-effect wind within the narrow gorge of the "Iron Gate," it can reach maximum speeds exceeding 30 m/s. The NMMB model, with a horizontal resolution of 1.2 km, was utilized for the 2019 Kosava episode forecast. The high resolution, that surpasses the typical standards in numerical prediction models used by national meteorological services and other centers, can be crucial for accurately predicting strong wind gusts and capturing the specific dynamics and characteristics of the wind associated with the narrow gorge. The NMMB model results are compared with measured wind data and the results from models with lower resolutions.

How to cite: Djurdjevic, V., Tosic, M., and Lazic, I.: Predicting strong local wind with high-resolution nonhydrostatic numerical weather prediction model, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-21913, https://doi.org/10.5194/egusphere-egu24-21913, 2024.

EGU24-539 | ECS | Orals | HS3.9

Revisiting the common approaches for hydrological model calibration with high-dimensional parameters and objectives  

Songjun Wu, Doerthe Tetzlaff, Keith Beven, and Chris Soulsby

Successful calibration of distributed hydrological models is often hindered by complex model structures, incommensurability between observed and modeled variables, and the complex nature of many hydrological processes. Many approaches have been proposed and compared for calibration, but the comparisons were generally based on parsimonious models with limited objectives. The conclusions could change when more parameters are to be calibrated with multiple objectives and increasing data availability. In this study four different approaches (random sampling, DREAM, NSGA-II, GLUE Limits of acceptability) were tested for a complex application - to calibrate 58 parameters of a hydrological model against 24 objectives (soil moisture and isotopes at 3 depths under vegetation covers). By comparing the simulation performance of parameter sets selected from different approaches, we concluded that random sampling is still usable in high-dimensional parameter space, providing comparable performance to other approaches despite of the poor parameter identifiability. DREAM provided better simulation performance and parameter convergence with informal likelihood functions; however, the difficulty in describing model residual distribution could possibly result in inappropriate formal likelihood functions and thus the poor simulations. Multi-criteria calibration, taking NSGA-II as an example, gave ideal model performance/parameter identifiability and explicitly unravelled the trade-offs between objectives after aggregating them (into 2 or 4); but calibrating against all 24 objectives was hindered by the “curse of dimensionality”, as the increasing dimension exponentially expanded the Pareto front and increased the difficulty to differentiate parameter sets. Finally, Limits of acceptability also provided comparable simulations; moreover, it can be regarded as a learning tool because detailed information about model failures is available for each objective at each timestep. However, the limitation is the insufficient exploration of high-dimensional parameter space due to the use of Latin-Hypercube sampling.

Overall, all approaches showed benefits and limitations, and a general approach to be easily used for such complex calibration cases without trial-and-error is still lacking. By comparing those common approaches, we realised the difficulty to define a proper objective function for many-objective optimisation, either for aggregated scalar function (due to the difficulty of assigning weights or assuming a form for the residual distribution) or the vector function (due to the expansion of the Pareto front). In this context, the Limits of Acceptability approach provided a more flexible way to define the “objective function” for each timestep, though it introduces extra demands in understanding data uncertainties and deciding on what should be considered acceptable. Moreover, in such many-objective optimisation, it is possible that not a single parameter set can capture all the objectives satisfactorily (not in 8 million run in this study).  The non-existence of any global optimal in the sample suggests that the concept of equifinality should be embraced in using an ensemble of comparable parameters to represent such complex systems.

How to cite: Wu, S., Tetzlaff, D., Beven, K., and Soulsby, C.: Revisiting the common approaches for hydrological model calibration with high-dimensional parameters and objectives , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-539, https://doi.org/10.5194/egusphere-egu24-539, 2024.

EGU24-1745 | Posters on site | HS3.9

Predictive uncertainty analysis using null-space Monte Carlo  

Husam Baalousha, Marwan Fahs, and Anis Younes

The inverse problem in hydrogeology poses a significant challenge for modelers due to its ill-posed nature and the non-uniqueness of solutions. This challenge is compounded by the substantial computational efforts required for calibrating highly parameterized aquifers, particularly those with significant heterogeneity, such as karst limestone aquifers. While stochastic methods like Monte Carlo simulations are commonly used to assess uncertainty, their extensive computational requirements often limit their practicality.

The Null Space Monte Carlo (NSMC) method provides a parameter-constrained approach to address these challenges in inverse problems, allowing for the quantification of uncertainty in calibrated parameters. This method was applied to the northern aquifer of Qatar, which is characterized by high heterogeneity. The calibration of the model utilized the pilot point approach, and the calibrated results were spatially interpolated across the aquifer area using kriging.

NSMC was then employed to generate 100 sets of parameter-constrained random variables representing hydraulic conductivities. The null space vectors of these random solutions were incorporated into the parameter space derived from the calibrated model. Statistical analysis of the resulting calibrated hydraulic conductivities revealed a wide range, varying from 0.1 to 350 m/d, illustrating the significant variability inherent in the karstic nature of the aquifer.

Areas with high hydraulic conductivity were identified in the middle and eastern parts of the aquifer. These regions of elevated hydraulic conductivity also exhibited high standard deviations, further emphasizing the heterogeneity and complex nature of the aquifer's hydraulic properties.

How to cite: Baalousha, H., Fahs, M., and Younes, A.: Predictive uncertainty analysis using null-space Monte Carlo , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-1745, https://doi.org/10.5194/egusphere-egu24-1745, 2024.

Remote sensing observations hold useful prior information about the terrestrial water cycle. However, combining remote sensing products for each hydrological variable does not close the water balance due to the associated uncertainties. Therefore, there is a need to quantify bias and random errors in the data. This study presents an extended version of the data-driven probabilistic data fusion for closing the water balance at a basin scale. In this version, we implement a monthly 250-m grid-based Bayesian hierarchical model leveraging multiple open-source data of precipitation, evaporation, and storage in an ensemble approach that fully exploits and maximizes the prior information content of the data. The model relates each variable in the water balance to its “true” value using bias and random error parameters with physical nonnegativity constraints. The water balance variables and error parameters are treated as unknown random variables with specified prior distributions. Given an independent set of ground-truth data on water imports and river discharge along with all monthly gridded water balance data, the model is solved using a combination of Markov Chain Monte Carlo sampling and iterative smoothing to compute posterior distributions of all unknowns. The approach is applied to the Hindon Basin, a tributary of the Ganges River, that suffers from groundwater overexploitation and depends on surface water imports. Results provide spatially distributed (i) hydrologically consistent water balance estimates and (ii) statistically consistent error estimates of the water balance data. 

How to cite: Mourad, R., Schoups, G., and Bastiaanssen, W.: A grid-based data-driven ensemble probabilistic data fusion: a water balance closure approach applied to the irrigated Hindon River Basin, India , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-2267, https://doi.org/10.5194/egusphere-egu24-2267, 2024.

EGU24-2300 | ECS | Posters on site | HS3.9

Representing systematic and random errors of eddy covariance measurements in suitable likelihood models for robust model selection  

Tobias Karl David Weber, Alexander Schade, Robert Rauch, Sebastian Gayler, Joachim Ingwersen, Wolfgang Nowak, Efstathios Diamantopoulos, and Thilo Streck

The importance of evapotranspiration (ET) fluxes for the terrestrial water cycle is demonstrated by an overwhelming body of literature. Unfortunately, errors in their measurement contribute significantly to (model) uncertainties in quantifying and understanding ecohydrological systems. Measurements of surface-atmosphere fluxes of water at the ecosystem scale, the eddy covariance method can be considered a powerful technique and considered an important tool to validate ET models. Spatially averaged fluxes of several hundred square meters may be obtained. While the eddy-covariance technique has become a routine method to estimate the turbulent energy fluxes at the soil-atmosphere boundary, it remains not error free. Some of the inherent errors are quantifiable and may be partitioned into systematic and stochastic errors. For model-data comparison, the nature of the measurement error needs to be known to derive knowledge about model adequacy. To this end, we compare several assumptions found in the literature to describe the statistical properties of the error with newly derived descriptions, in this study. We are able to show, how sensitive the assumptions about the error are on the model selection process. We demonstrate this by comparing daily agro-ecosystem ET fluxes simulated with the detailed agro-hydrological model Expert-N to data gathered using the eddy-covariance technique.

How to cite: Weber, T. K. D., Schade, A., Rauch, R., Gayler, S., Ingwersen, J., Nowak, W., Diamantopoulos, E., and Streck, T.: Representing systematic and random errors of eddy covariance measurements in suitable likelihood models for robust model selection , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-2300, https://doi.org/10.5194/egusphere-egu24-2300, 2024.

EGU24-4140 | ECS | Orals | HS3.9

Integrating Deterministic and Probabilistic Approaches for Improved Hydrological Predictions: Insights from Multi-model Assessment in the Great Lakes Watersheds 

Jonathan Romero-Cuellar, Rezgar Arabzadeh, James Craig, Bryan Tolson, and Juliane Mai

The utilization of probabilistic streamflow predictions holds considerable value in the domains of predictive uncertainty estimation, hydrologic risk management, and decision support in water resources. Typically, the quantification of predictive uncertainty is formulated and evaluated using a solitary hydrological model, posing challenges in extrapolating findings to diverse model configurations. To address this limitation, this study examines variations in the performance ranking of various streamflow models through the application of a residual error model post-processing approach across multiple basins and models. The assessment encompasses 141 basins within the Great Lakes watershed, spanning the USA and Canada, and involves the evaluation of 13 diverse streamflow models using deterministic and probabilistic performance metrics. This investigation scrutinizes the interdependence between the quality of probabilistic streamflow estimation and the underlying model quality. The results underscore that the selection of a streamflow model significantly influences the robustness of probabilistic predictions. Notably, transitioning from deterministic to probabilistic predictions, facilitated by a post-processing approach, maintains the performance ranking consistency for the best and worst deterministic models. However, models of intermediate rank in deterministic evaluation exhibit inconsistent rankings when evaluated in probabilistic mode. Furthermore, the study reveals that post-processing residual errors of long short-term memory (LSTM) network models consistently outperform other models in both deterministic and probabilistic metrics. This research emphasizes the importance of integrating deterministic streamflow model predictions with residual error models to enhance the quality and utility of hydrological predictions. It elucidates the extent to which the efficacy of probabilistic predictions is contingent upon the sound performance of the underlying model and its potential to compensate for deficiencies in model performance. Ultimately, these findings underscore the significance of combining deterministic and probabilistic approaches for improving hydrological predictions, quantifying uncertainty, and supporting decision-making in operational water management.

How to cite: Romero-Cuellar, J., Arabzadeh, R., Craig, J., Tolson, B., and Mai, J.: Integrating Deterministic and Probabilistic Approaches for Improved Hydrological Predictions: Insights from Multi-model Assessment in the Great Lakes Watersheds, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-4140, https://doi.org/10.5194/egusphere-egu24-4140, 2024.

EGU24-5219 | ECS | Posters on site | HS3.9

Quantifying Uncertainty in Surrogate-based Bayesian Inference 

Anneli Guthke, Philipp Reiser, and Paul-Christian Bürkner

Proper sensitivity and uncertainty analysis for complex Earth and environmental systems models may become computationally prohibitive. Surrogate models can be an alternative to enable such analyses: they are cheap-to-run statistical approximations to the simulation results of the original expensive model. Several approaches to surrogate modelling exist, all with their own challenges and uncertainties. It is crucial to correctly propagate the uncertainties related to surrogate modelling to predictions, inference and derived quantities in order to draw the right conclusions from using the surrogate model.

While the uncertainty in surrogate model parameters due to limited training data (expensive simulation runs) is often accounted for, what is typically ignored is the approximation error due to the surrogate’s structure (bias in reproducing the original model predictions). Reasons are that such a full uncertainty analysis is computationally costly even for surrogates (or limited to oversimplified analytic cases), and that a comprehensive framework for uncertainty propagation with surrogate models was missing.

With this contribution, we propose a fully Bayesian approach to surrogate modelling, uncertainty propagation, parameter inference, and uncertainty validation. We illustrate the utility of our approach with two synthetic case studies of parameter inference and validate our inferred posterior distributions by simulation-based calibration. For Bayesian inference, the correct propagation of surrogate uncertainty is especially relevant, because failing to account for it may lead to biased and/or overconfident parameter estimates and will spoil further interpretation in the physics’ context or application of the expensive simulation model.

Consistent and comprehensive uncertainty propagation in surrogate models enables more reliable approximation of expensive simulations and will therefore be useful in various fields of applications, such as surface or subsurface hydrology, fluid dynamics, or soil hydraulics.

How to cite: Guthke, A., Reiser, P., and Bürkner, P.-C.: Quantifying Uncertainty in Surrogate-based Bayesian Inference, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-5219, https://doi.org/10.5194/egusphere-egu24-5219, 2024.

EGU24-6157 | ECS | Orals | HS3.9

Analyzing Groundwater Hazards with Sequential Monte Carlo  

Lea Friedli and Niklas Linde

Analyzing groundwater hazards frequently involves utilizing Bayesian inversions and estimating probabilities associated with rare events. A concrete example concerns the potential contamination of an aquifer, a process influenced by the unknown hydraulic properties of the subsurface. In this context, the emphasis shifts from the posterior distribution of model parameters to the distribution of a particular quantity of interest dependent on these parameters. To tackle the methodological hurdles at hand, we propose a Sequential Monte Carlo approach in two stages. The initial phase involves generating particles to approximate the posterior distribution, while the subsequent phase utilizes subset sampling techniques to evaluate the probability of the specific rare event of interest. Exploring a two-dimensional flow and transport example, we demonstrate the efficiency and accuracy of the developed PostRisk-SMC method in estimating rare event probabilities associated with groundwater hazards.

How to cite: Friedli, L. and Linde, N.: Analyzing Groundwater Hazards with Sequential Monte Carlo , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-6157, https://doi.org/10.5194/egusphere-egu24-6157, 2024.

EGU24-7610 | Posters on site | HS3.9

Parameter estimation of heterogeneous field in basin scale based on signal analysis and river stage tomography 

Bo-Tsen Wang, Chia-Hao Chang, and Jui-Pin Tsai

Understanding the spatial distribution of the aquifer parameters is crucial to evaluating the groundwater resources on a basin scale. River stage tomography (RST) is one of the potential methods to estimate the aquifer parameter fields. Utilizing the head variations caused by the river stage to conduct RST is essential to delineate the regional aquifer's spatial features successfully. However, the two external stimuli of the aquifer system, rainfall and river stage, are usually highly correlated, resulting in mixed features in the head observations, which may cause unreasonable estimates of parameter fields. Thus, separating the head variations sourced from rainfall and river stage is essential to developing the reference heads for RST. To solve this issue, we propose a systematic approach to extracting and reconstructing the head variations of river features from the original head observations during the flood periods and conducting RST. We utilized a real case study to examine the developed method. This study used the groundwater level data, rainfall data, and river stage data in the Zhuoshui River alluvial fan in 2006. The hydraulic diffusivity (D) values of five observation wells were used as the reference for parameter estimation. The results show that the RMSE of the D value is 0.027 (m2/s). The other three observation wells were selected for validation purposes, and the derived RMSE is 0.85(m2/s). The low RMSE reveals that the estimated D field can capture the characteristics of the regional aquifer. The results also indicate that the estimated D values derived from the developed method are consistent with the sampled D values from the pumping tests in the calibration and validation processes in the real case study. The results demonstrate that the proposed method can successfully extract and reconstruct the head variations of river features from the original head observations and can delineate the features of the regional parameter field. The proposed method can benefit RST studies and provide an alternative mixed-feature signal decomposition and reconstruction method.

How to cite: Wang, B.-T., Chang, C.-H., and Tsai, J.-P.: Parameter estimation of heterogeneous field in basin scale based on signal analysis and river stage tomography, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-7610, https://doi.org/10.5194/egusphere-egu24-7610, 2024.

EGU24-7820 | Orals | HS3.9

Data-driven surrogate-based Bayesian model calibration for predicting vadose zone temperatures in drinking water supply pipes 

Ilja Kröker, Elisabeth Nißler, Sergey Oladyshkin, Wolfgang Nowak, and Claus Haslauer

Soil temperature and soil moisture in the unsaturated zone depend on each other and are influenced by non-stationary hydro-meteorological forcing factors that are subject to climate change. 

The transport of both heat and moisture are crucial for predicting temperatures in the shallow subsurface and, as consequence, around and in drinking water supply pipes. Elevated temperatures in water supply pipes (even up to 25°C and above) pose a risk to human health due to increased likelihood of microbial contamination. 

To model variably saturated flow and heat transport, a partial differential equation (PDE)-based integrated hydrogeological model has been developed and implemented in the DuMuX simulation framework.  This model integrates the hydrometeorological forcing functions via a novel interface condition at the atmosphere-subsurface boundary. Relevant soil properties and their dependency on temperatures have been measured as time series at a pilot site at the University of Stuttgart in detail since 2020. 

Despite these efforts on measurements and model enhancement, some uncertainties remain. These include capillary-saturation relationships in materials where they are difficult to measure, especially in the gravel-type materials that are commonly used above drinking water pipes. 

To enhance our understanding of the underlying physical processes, we employ Bayesian inference, which is a well-established approach to estimate uncertain or unknown model parameters. Computationally cheap surrogate models allow to overcome the limitations of Bayesian methods for computationally intensive models, when such surrogate models are used in lieu of the physical (PDE)-based model. Here, we use the arbitrary polynomial chaos expansion equipped with Bayesian regularization (BaPC).  The BaPC allows to exploit latest (Bayesian) active learning strategies to reduce the number of model-runs that are necessary for constructing the surrogate model.  

In the present work, we demonstrate the calibration of a PDE-based integrated hydrogeological model using Bayesian inference on a BaPC-based surrogate.  The accuracy of the calibrated and predicted temperatures in the shallow subsurface is then assessed against real-world measurement data. 

How to cite: Kröker, I., Nißler, E., Oladyshkin, S., Nowak, W., and Haslauer, C.: Data-driven surrogate-based Bayesian model calibration for predicting vadose zone temperatures in drinking water supply pipes, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-7820, https://doi.org/10.5194/egusphere-egu24-7820, 2024.

EGU24-8007 | ECS | Orals | HS3.9

Investigating the divide and measure nonconformity  

Daniel Klotz, Martin Gauch, Frederik Kratzert, Grey Nearing, and Jakob Zscheischler

This contribution presents a diagnostic approach to investigate unexpected side effects that can occur during the evaluation of rainfall--runoff models.

The diagnostic technique that we use is based on the idea that one can use gradient descent to modify the runoff observations/simulations to obtain warranted observations/simulations. Specifically, we show how to use this concept to manipulate any hydrograph (e.g., a copy of the observations) so that it approximates specific NSE values for individual parts of the data. In short, we follow the following recipe to generate the synthetic simulations: (1) copy the observations, (2) add noise, (3) clip the modified discharge to zero, and (4) optimise the obtained simulation values by using gradient descent until a desired NSE value is reached.

To show how this diagnostic technique can be used we demonstrate a behaviour of Nash--Sutcliffe Efficiency (NSE) that appears when evaluating a model over subsets of the data: If models perform poorly for certain situations, this lack of performance is not necessarily reflected in the NSE (of the overall data). This behaviour follows from the definition of NSE and is therefore 100% explainable. However, from our experience it can be unexpected for many modellers. Our results also show that subdividing the data and evaluating over the resulting partitions yields different information regarding model deficiencies than an overall evaluation. We call this phenomenon the Divide And Measure Nonconformity or DAMN.



How to cite: Klotz, D., Gauch, M., Kratzert, F., Nearing, G., and Zscheischler, J.: Investigating the divide and measure nonconformity , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-8007, https://doi.org/10.5194/egusphere-egu24-8007, 2024.

Groundwater heads are commonly used to monitor storage of aquifers and as decision variables for groundwater management. Alluvial gravel aquifers are often characterized by high transmissivities and a corresponding strong seasonal and inter-annual variability of storage. The sustainable management of such aquifers is challenging, particularly for already tightly allocated aquifers and in increasingly extreme and potentially drier climates, and might require the restriction of groundwater abstraction for periods of time. Stakeholders require lead-in time to prepare for potential restrictions of their consented takes.

Groundwater models have been used in the past to support groundwater decision making and to provide the corresponding predictions of groundwater levels for operational forecasting and management. In this study, we benchmark and compare different model classes to perform this task: (i) a spatially explicit 3D groundwater flow model (MODFLOW), (ii) a conceptual, bucket-type Eigenmodel, (iii) a transfer-function model (TFN), and (iv) three machine learning (ML) techniques, namely, Multi-Layer Perceptron models (MLP), Long Short-Term Memory models (LSTM), and Random Forrest (RF) models. The model classes differ widely in their complexity, input requirements, calibration effort, and run-times. The different model classes are tested on four groundwater head time series taken from the Wairau Aquifer in New Zealand (Wöhling et al., 2020). Posterior parameter ensembles of MODFLOW (Wöhling et al., 2018) and the EIGENMODEL (Wöhling & Burbery, 2020) were combined with TFN and ML variants with different input features to form a (prior) multi-model ensemble. Models classes are ranked with posterior model weights derived from Bayesian model selection (BMS) and averaging (BMA) techniques.

Our results demonstrate that no “model that fits all” exists in our model set. The more physics-based MODFLOW model is not necessarily providing the most accurate predictions, but can provide physical meaning and interpretation for the entire model region and outputs at locations where no data is available. ML techniques have generally much lower input requirements and short run-times. They show to be competitive candidates for groundwater head predictions where observations are available, even for system states that lie outside the calibration data range.

Because the performance of model types is site-specific, we advocate the use of multi-model ensemble forecasting wherever feasible. The benefit is illustrated by our case study, with BMA uncertainty bounds providing a better coverage of the data and the BMA mean performing well for all tested sites. Redundant ensemble members (with BMA weights of zero) are easily filtered out to obtain efficient ensembles for operational forecasting.

 

References

Wöhling T, Burbery L (2020). Eigenmodels to forecast groundwater levels in unconfined river-fed aquifers during flow recession. Science of the Total Environment, 747, 141220, doi: 10.1016/j.scitotenv.2020.141220.

Wöhling, T., Gosses, M., Wilson, S., Wadsworth, V., Davidson, P. (2018). Quantifying river-groundwater interactions of New Zealand's gravel-bed rivers: The Wairau Plain. Goundwater doi:10.1111/gwat.12625

Wöhling T, Wilson SR, Wadsworth V, Davidson P. (2020). Detecting the cause of change using uncertain data: Natural and anthropogenic factors contributing to declining groundwater levels and flows of the Wairau Plain Aquifer, New Zealand. Journal of Hydrology: Regional Studies, 31, 100715, doi: 10.1016/j.ejrh.2020.100715.

 

How to cite: Wöhling, T. and Crespo Delgadillo, O.: Predicting groundwater heads in alluvial aquifers: Benchmarking different model classes and machine-learning techniques with BMA/S, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-8818, https://doi.org/10.5194/egusphere-egu24-8818, 2024.

EGU24-8872 | Orals | HS3.9

Characterization and modeling of large-scale aquifer systems under uncertainty: methodology and application to the Po River aquifer system 

Monica Riva, Andrea Manzoni, Rafael Leonardo Sandoval, Giovanni Michele Porta, and Alberto Guadagnini

Large-scale groundwater flow models are key to enhance our understanding of the potential impacts of climate and anthropogenic factors on water systems. Through these, we can identify significant patterns and processes that most affect water security. In this context, we have developed a comprehensive and robust theoretical framework and operational workflow that can effectively manage complex heterogeneous large-scale groundwater systems. We rely on machine learning techniques to map the spatial distribution of geomaterials within three-dimensional subsurface systems. The groundwater modeling approach encompasses (a) estimation of groundwater recharge and abstractions, as well as (b) appraisal of interactions among subsurface and surface water bodies. We ground our analysis on a unique dataset that encompasses lithostratigraphic data as well as piezometric and water extraction data across the largest aquifer system in Italy (the Po River basin). The quality of our results is assessed against pointwise information and hydrogeological cross-sections which are available within the reconstructed domain. These can be considered as soft information based on expert assessment. As uncertainty quantification is critical for subsurface characterization and assessment of future states of the groundwater system, the proposed methodology is designed to provide a quantitative evaluation of prediction uncertainty at any location of the reconstructed domain. Furthermore, we quantify the relative importance of uncertain model parameters on target model outputs through the implementation of a rigorous Global Sensitivity Analysis. By evaluating the spatial distribution of global sensitivity metrics associated with model parameters, we gain valuable insights into areas where the acquisition of future information could enhance the quality of groundwater flow model parameterization and improve hydraulic head estimates. The comprehensive dataset provided in this study, combined with the reconstruction of the subsurface system properties and piezometric head distribution and with the quantification of the associated uncertainty, can be readily employed in the context of groundwater availability and quality studies associated with the region of interest. The approach and operational workflow are flexible and readily transferable to assist identification of the main dynamics and patterns of large-scale aquifer systems of the kind here analyzed.

How to cite: Riva, M., Manzoni, A., Sandoval, R. L., Porta, G. M., and Guadagnini, A.: Characterization and modeling of large-scale aquifer systems under uncertainty: methodology and application to the Po River aquifer system, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-8872, https://doi.org/10.5194/egusphere-egu24-8872, 2024.

EGU24-10517 | Orals | HS3.9

Lock-ins and path dependency in evaluation metrics used for hydrological models 

Lieke Melsen, Arnald Puy, and Andrea Saltelli

Science, being conducted by humans, is inherently a social activity. This is evident in the development and acceptance of scientific methods. Science is not only socially shaped, but also driven (and in turn influenced) by technological development: technology can open up new research avenues. At the same time, it has been shown that technology can cause lock-ins and path dependency. A scientific activity driven both by social behavior and technological development is modelling. As such, studying modelling as a socio-technical activity can provide insights both in enculturation processes and in lock-ins and path dependencies. Even more, enculturation can lead to lock-ins. We will demonstrate this for the Nash-Sutcliffe Efficiency (NSE), a popular evaluation metric in hydrological research. Through a bibliometric analysis we show that the NSE is part of hydrological research culture and does not appear in adjacent research fields. Through a historical analysis we demonstrate the path dependency that has developed with the popularity of the NSE. Finally, through exploring the faith of alternative measures, we show the lock-in effect of the use of the NSE. As such, we confirm that the evaluation of models needs to take into account cultural embeddedness. This is relevant because peers' acceptance is a powerful legitimization argument to trust the model and/or model results, including for policy relevant applications. Culturally determined bias needs to be assessed for its potential consequences in the discipline. 

How to cite: Melsen, L., Puy, A., and Saltelli, A.: Lock-ins and path dependency in evaluation metrics used for hydrological models, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-10517, https://doi.org/10.5194/egusphere-egu24-10517, 2024.

EGU24-10770 | Orals | HS3.9 | Highlight

Uncertainty and sensitivity analysis: new purposes, new users, new challenges 

Francesca Pianosi, Hannah Bloomfield, Gemma Coxon, Robert Reinecke, Saskia Salwey, Georgios Sarailidis, Thorsten Wagener, and Doris Wendt

Uncertainty and sensitivity analysis are becoming an integral part of mathematical modelling of earth and environmental systems. Uncertainty analysis aims at quantifying uncertainty in model outputs, which helps to avoid spurious precision and increase the trustworthiness of model-informed decisions. Sensitivity analysis aims at identifying the key sources of output uncertainty, which helps to set priorities for uncertainty reduction and model improvement.

In this presentation, we draw on a range of recent studies and projects to discuss the status of uncertainty and sensitivity analysis, focusing in particular on ‘global’ approaches, whereby uncertainties and sensitivities are quantified across the entire space of plausible variability of model inputs.

We highlight some of the challenges and untapped potential of these methodologies, including: (1) innovative ways to use global sensitivity analysis to test the ‘internal consistency’ of models and therefore support their diagnostic evaluation; (2) challenges and opportunities to promote the uptake of these methodologies to increasingly complex models, chains of models, and models used in industry; (3) the limits of uncertainty and sensitivity analysis when dealing with epistemic, poorly bounded or unquantifiable sources of uncertainties.

How to cite: Pianosi, F., Bloomfield, H., Coxon, G., Reinecke, R., Salwey, S., Sarailidis, G., Wagener, T., and Wendt, D.: Uncertainty and sensitivity analysis: new purposes, new users, new challenges, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-10770, https://doi.org/10.5194/egusphere-egu24-10770, 2024.

EGU24-11414 | ECS | Posters on site | HS3.9

Single vs. multi-objective optimization approaches to calibrate an event-based conceptual hydrological model using model output uncertainty framework. 

Muhammad Nabeel Usman, Jorge Leandro, Karl Broich, and Markus Disse

Flash floods have become one of the major natural hazards in central Europe, and climate change projections indicate that the frequency and severity of flash floods will increase in many areas across the world and in central Europe. The complexity involved in the flash flood generation makes it difficult to calibrate a hydrological model for the prediction of such peak hydrological events. This study investigates the best approach to calibrate an event-based conceptual HBV model, comparing different trials of single-objective, single-event multi-objective (SEMO), and multi-event-multi-objective (MEMO) model calibrations. Initially, three trials of single-objective calibration are performed w.r.t. RMSE, NSE, and BIAS separately, then three different trials of multi-objective optimization, i.e., SEMO-3D (single-event three objectives), MEMO-3D (mean of three objectives from two events), and MEMO-6D (two events six objectives) are formulated. Model performance was validated for several peak events via 90 % (confidence interval) CI-based output uncertainty quantification. The uncertainties associated with the model predictions are estimated stochastically using the ‘relative errors (REs)’ between the simulated (Qsim) and measured (Qobs) discharges as a likelihood measure. Single-objective model calibration demonstrated that significant trade-offs exist between different objective functions, and no unique parameter set can optimize all objectives simultaneously. Compared to the solutions of single-objective calibration, all the multi-objective calibration formulations produced relatively accurate and robust results during both model calibration and validation phases. The uncertainty intervals associated with all the trials of single-objective calibration and the SEMO-3D calibration failed to capture observed peaks of the validation events. The uncertainty bands associated with the ensembles of Pareto solutions from the MEMO-3D and MEMO-6D (six-dimensional) calibrations displayed better performance in reproducing and capturing more significant peak validation events. However, to bracket peaks of large flash flood events within the prediction uncertainty intervals, the MEMO-6D optimization outperformed all the single-objective, SEMO-3D, and MEMO-3D multi-objective calibration methods. This study suggests that the MEMO_6D is the best approach for predicting large flood events with lower model output uncertainties when the calibration is performed with a better combination of peak events.

How to cite: Usman, M. N., Leandro, J., Broich, K., and Disse, M.: Single vs. multi-objective optimization approaches to calibrate an event-based conceptual hydrological model using model output uncertainty framework., EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-11414, https://doi.org/10.5194/egusphere-egu24-11414, 2024.

EGU24-12676 | ECS | Posters on site | HS3.9

Physics-Informed Ensemble Surrogate Modeling of Advective-Dispersive Transport Coupled with Film Intraparticle Pore Diffusion Model for Column Leaching Test 

Amirhossein Ershadi, Michael Finkel, Binlong Liu, Olaf Cirpka, and Peter Grathwohl

Column leaching tests are a common approach for evaluating the leaching behavior of contaminated soil and waste materials, which are often reused for various construction purposes. The observed breakthrough curves of the contaminants are affected by the intricate dynamics of solute transport, inter-phase mass transfer, and dispersion. Disentangling these interactions requires numerical models. However, inverse modeling and parameter sensitivity analysis are often time-consuming, especially when sorption/desorption kinetics are explicitly described by intra-particle diffusion, requiring the discretization along the column axis and inside the grains. To replace such computationally expensive models, we developed a machine-learning based surrogate model employing two disparate ensemble methods (stacking and weighted distance average) within the defined parameter range based on the German standard for column leaching tests. To optimize the surrogate model, adaptive sampling methods based on three distinct infill criteria are employed. These criteria include maximizing expected improvement, the Mahalanobis distance (exploitation), and maximizing standard deviation (exploration).
The stacking surrogate model makes use of extremely randomized trees and random forest as base- and meta-model. The model shows a very good performance in emulating the behavior of the original numerical model (Relative Root Mean Squared Error = 0.09). 
Our proposed surrogate model has been applied to estimate the complete posterior parameter distribution using Markov Chain Monte Carlo simulation. The impact of individual input parameters on the predictions generated by the surrogate model was analyzed using SHapley Additive exPlanations methods.

How to cite: Ershadi, A., Finkel, M., Liu, B., Cirpka, O., and Grathwohl, P.: Physics-Informed Ensemble Surrogate Modeling of Advective-Dispersive Transport Coupled with Film Intraparticle Pore Diffusion Model for Column Leaching Test, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-12676, https://doi.org/10.5194/egusphere-egu24-12676, 2024.

EGU24-13393 | ECS | Posters on site | HS3.9

Datasets and tools for local and global meteorological ensemble estimation 

Guoqiang Tang, Andrew Wood, Andrew Newman, Martyn Clark, and Simon Papalexiou

Ensemble gridded meteorological datasets are critical for driving hydrology and land models, enabling uncertainty analysis, and supporting a variety of hydroclimate research and applications. The Gridded Meteorological Ensemble Tool (GMET) has been a significant contributor in this domain, offering an accessible platform for generating ensemble precipitation and temperature datasets. The GMET methodology has continually evolved since its initial development in 2006, primarily in the form of a FORTRAN code base, and has since been utilized to generate historical and real-time ensemble meteorological (model forcing) datasets in the U.S. and part of Canada. A recent adaptation of GMET was used to produce multi-decadal forcing datasets for North America and the globe (EMDNA and EM-Earth, respectively). Those datasets have been used to support diverse hydrometeorological applications such as streamflow forecasting and hydroclimate studies across various scales. GMET has now evolved into a Python package called the Geospatial Probabilistic Estimation Package (GPEP), which offers methodological and technical enhancements relative to GMET. These include greater variable selection flexibility, intrinsic parallelization, and especially a broader suite of estimation methods, including the use of techniques from the scikit-learn machine learning library. GPEP enables a wider variety of strategies for local and global estimation of geophysical variables beyond traditional hydrological forcings.  This presentation summarizes GPEP and introduces major open-access ensemble datasets that have been generated with GMET and GPEP, including a new effort to create high-resolution (2 km) surface meteorological analyses for the US. These resources are useful in advancing hydrometeorological uncertainty analysis and geospatial estimation.

How to cite: Tang, G., Wood, A., Newman, A., Clark, M., and Papalexiou, S.: Datasets and tools for local and global meteorological ensemble estimation, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-13393, https://doi.org/10.5194/egusphere-egu24-13393, 2024.

We consider the optimal inference of spatially heterogeneous hydraulic conductivity and head fields based on three kinds of point measurements that may be available at monitoring wells: of head, permeability, and groundwater speed. We have developed a general, zonation-free technique for Monte Carlo (MC) study of field recovery problems, based on Karhunen-Loève (K-L) expansions of the unknown fields, whose coefficients are recovered by an analytical adjoint-state technique. This allows unbiased sampling from the space of all possible fields with a given correlation structure and efficient, automated gradient-descent calibration. The K-L basis functions have a straightforward notion of period, revealing the relationship between feature scale and reconstruction fidelity, and they have an a priori known spectrum, allowing for a non-subjective regularization term to be defined. We have performed automated MC calibration on over 1100 conductivity-head field pairs, employing a variety of point measurement geometries and quantified the mean-squared field reconstruction accuracy, both globally and as a function of feature scale.

We present heuristics for feature scale identification, examine global reconstruction error, and explore the value added by both groundwater speed measurements and by two different types of regularization. We show that significant feature identification becomes possible as feature scale exceeds four times measurement spacing and identification reliability subsequently improves in a power law fashion with increasing feature scale.

How to cite: Hansen, S. K., O'Malley, D., and Hambleton, J.: Feature scale and identifiability: quantifying the information that point hydraulic measurements provide about heterogeneous head and conductivity fields, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-14219, https://doi.org/10.5194/egusphere-egu24-14219, 2024.

EGU24-14805 | Orals | HS3.9

Sensitivity analysis of input variables of a SWAT hydrological model using the machine learning technique of random forest 

Ali Abousaeidi, Seyed Mohammad Mahdi Moezzi, Farkhondeh Khorashadi Zadeh, Seyed Razi Sheikholeslami, Albert Nkwasa, and Ann van Griensven

Sensitivity analysis of complex models, with a large number of input variables and parameters, is time-consuming and inefficient, using traditional approaches. Considering the capability of computing importance indices, the machine learning technique of the Random Forest (RF) is introduced as an alternative to conventional methods of sensitivity analysis. One of the advantages of using the RF model is the reduction of computational costs for sensitivity analysis.

The objective of this research is to analyze the importance of the input variables of a semi-distributed and physically-based hydrological model, namely SWAT (soil and water assessment tool) using the RF model. To this end, an RF-based model is first trained using SWAT input variables (such as, precipitation and temperature) and SWAT output variables (like streamflow and sediment load). Then, using the importance index of the RF model, the ranking of input variables, in terms of their impact on the accuracy of the model results, is determined. Additionally, the results of the sensitivity analysis are examined graphically. To validate the ranking results of the RF-based approach, the parameter ranking results of the Sobol G function, using the RF-based approach and the sensitivity analysis method of Sobol’ are compared. The ranking of the model input variables plays a significant role in the development of models and prioritizing efforts to reduce model errors.

Key words: Sensitivity analysis, model input variables, Machine learning technique, Random forest, SWAT model.

How to cite: Abousaeidi, A., Moezzi, S. M. M., Khorashadi Zadeh, F., Sheikholeslami, S. R., Nkwasa, A., and van Griensven, A.: Sensitivity analysis of input variables of a SWAT hydrological model using the machine learning technique of random forest, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-14805, https://doi.org/10.5194/egusphere-egu24-14805, 2024.

EGU24-16086 | ECS | Posters on site | HS3.9

Disentangling the role of different sources of uncertainty and model structural error on predictions of water and carbon fluxes with CLM5 for European observation sites 

Fernand Baguket Eloundou, Lukas Strebel, Bibi S. Naz, Christian Poppe Terán, Harry Vereecken, and Harrie-Jan Hendricks Franssen

The Community Land Model version 5 (CLM5) integrates processes encompassing the water, energy, carbon, and nitrogen cycles, and ecosystem dynamics, including managed ecosystems like agriculture. Nevertheless, the intricacy of CLM5 introduces predictive uncertainties attributed to factors such as input data, process parameterizations, and parameter values. This study conducts a comparative analysis between CLM5 ensemble simulations and eddy covariance and in-situ measurements, focusing on the effects of uncertain model parameters and atmospheric forcings on the water, carbon, and energy cycles.
Ensemble simulations for 14 European experimental sites were performed with the CLM5-BGC model, integrating the biogeochemistry component. In four perturbation experiments, we explore uncertainties arising from atmospheric forcing data, soil parameters, vegetation parameters, and the combined effects of these factors. The contribution of different uncertainty sources to total simulation uncertainty was analyzed by comparing the 99% confidence
intervals from ensemble simulations with measured terrestrial states and fluxes, using a three-way analysis of variance.
The study identifies that soil parameters primarily influence the uncertainty in estimating surface soil moisture, while uncertain vegetation parameters control the uncertainty in estimating evapotranspiration and carbon fluxes. A combination of uncertainty in atmospheric forcings and vegetation parameters mostly explains the uncertainty in sensible heat flux estimation. On average, the 99% confidence intervals envelope >40% of the observed fluxes, but this varies greatly between sites, exceeding 95% in some cases. For some sites, we could identify model structural errors related to model spin-up assumptions or erroneous plant phenology. The study guides identifying factors causing underestimation or overestimation in the variability of fluxes, such as crop parameterization or spin-up, and potential structural errors in point-scale simulations in CLM5.

How to cite: Eloundou, F. B., Strebel, L., Naz, B. S., Terán, C. P., Vereecken, H., and Hendricks Franssen, H.-J.: Disentangling the role of different sources of uncertainty and model structural error on predictions of water and carbon fluxes with CLM5 for European observation sites, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-16086, https://doi.org/10.5194/egusphere-egu24-16086, 2024.

EGU24-16361 | ECS | Orals | HS3.9

Estimating prior distributions of TCE transformation rate constants from literature data 

Anna Störiko, Albert J. Valocchi, Charles Werth, and Charles E. Schaefer

Stochastic modeling of contaminant reactions requires the definition of prior distributions for the respective rate constants. We use data from several experiments reported in the literature to better understand the distribution of pseudo-first-order rate constants of abiotic TCE reduction in different sediments. These distributions can be used to choose informed priors for these parameters in reactive-transport models.

Groundwater contamination with trichloroethylene (TCE) persists at many hazardous waste sites due to back diffusion from low-permeability zones such as clay lenses. In recent years, the abiotic reduction of TCE by reduced iron minerals has gained attention as a natural attenuation process, but there is uncertainty as to whether the process is fast enough to be effective. Pseudo-first-order rate constants have been determined in laboratory experiments and are reported in the literature for various sediments and rocks, as well as for individual reactive minerals. However, rate constants can vary between sites and aquifer materials. Reported values range over several orders of magnitude.

To assess the uncertainty and variability of pseudo-first-order rate constants, we compiled data reported in several studies. We built a statistical model based on a hierarchical Bayesian approach to predict probability distributions of rate constants at new sites based on this data set. We then investigated whether additional information about the sediment composition at a site could reduce the uncertainty. We tested two sets of predictors: reactive mineral content or the extractable Fe(II) content. Knowing the reactive mineral content reduced the uncertainty only slightly. In contrast, knowing the Fe(II) content greatly reduced the uncertainty because the relationship between Fe(II) content and rate constants is approximately log-log-linear. Using a simple example of diffusion-controlled transport in a contaminated aquitard, we show how the uncertainty in the predicted rate constants affects the predicted remediation times.

How to cite: Störiko, A., Valocchi, A. J., Werth, C., and Schaefer, C. E.: Estimating prior distributions of TCE transformation rate constants from literature data, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-16361, https://doi.org/10.5194/egusphere-egu24-16361, 2024.

Deeper insights on internal model behaviors are essential as hydrological models are becoming more and more complex. Our study provides a framework which combines the time-varying global sensitivity analyses with data mining techniques to unravel the process-level behavior of high-complexity models and tease out the main information. The extracted information is further used to assist parameter identification. The physically-based Distributed Hydrology-Soil-Vegetation Model (DHSVM) set up in a mountainous watershed is used as a case study. Specifically, a two-step GSA including time-aggregated and time-variant approaches are conducted to address the problem of high parameter dimensionality and characterize the time-varying parameter importance. As we found difficulties in interpreting the long-term complicated dynamics, a clustering operation is performed to partition the entire period into several clusters and extract the corresponding temporal parameter importance patterns. Finally, the clustered time clusters are utilized in parameterization, where each parameter is identified in their dominant times. Results are summarized as follows: (1) importance of selected soil and vegetation parameters varies greatly throughout the period; (2) typical patterns of parameter importance corresponding to flood, very short dry-to-wet, fast recession and continuous dry periods are successfully distinguished. We argue that somewhere between “total period” and “continuous discrete time” can be more useful for understanding and interpretation; (3) parameters dominant for short times are much more identifiable when they are identified in dominance time cluster(s); (4) the enhanced parameter identifiability overall improves the model performance according to the metrics of NSE, LNSE, and RMSE, suggesting that the use of GSA information has the potential to provide a better search for optimal parameter sets.

How to cite: Wang, L., Xu, Y., Gu, H., and Liang, X.: Investigating dynamic parameter importance of a high-complexity hydrological model and implications for parameterization, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-18569, https://doi.org/10.5194/egusphere-egu24-18569, 2024.

EGU24-18804 | ECS | Orals | HS3.9

Accelerating Hydrological Model Inversion: A Multilevel Approach to GLUE 

Max Rudolph, Thomas Wöhling, Thorsten Wagener, and Andreas Hartmann

Inverse problems play a pivotal role in hydrological modelling, particularly for parameter estimation and system understanding, which are essential for managing water resources. The application of statistical inversion methodologies such as Generalized Likelihood Uncertainty Estimation (GLUE) is often obstructed, however, by high model computational cost given that Monte Carlo sampling strategies often return a very small fraction of behavioural model runs. There is a need, however, to balance this aspect with the demand for broadly sampling the parameter space. Especially relevant for spatially distributed or (partial) differential equation based models, this aspect calls for computationally efficient methods of statistical inference that approximate the “true” posterior parameter distribution well. Our study introduces multilevel GLUE (MLGLUE), which effectively mitigates these computational challenges by exploiting a hierarchy of models with different computational grid resolutions (i.e., spatial or temporal discretisation), inspired by multilevel Monte Carlo strategies. Starting with low-resolution models, MLGLUE only passes parameter samples to higher-resolution models for evaluation if associated with a high likelihood, which poses a large potential for substantial computational savings. We demonstrate the applicability of the approach using a groundwater flow model with a hierarchy of different spatial resolutions. With MLGLUE, the computation time of parameter inference could be reduced by more than 60% compared to GLUE, while the resulting posterior distributions are virtually identical. Correspondingly, the uncertainty estimates of MLGLUE and GLUE are also very similar. Considering the simplicity of the implementation as well as its efficiency, MLGLUE promises to be an alternative for statistical inversion of computationally costly hydrological models.

How to cite: Rudolph, M., Wöhling, T., Wagener, T., and Hartmann, A.: Accelerating Hydrological Model Inversion: A Multilevel Approach to GLUE, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-18804, https://doi.org/10.5194/egusphere-egu24-18804, 2024.

EGU24-19966 | Orals | HS3.9

Operational Sensitivity Analysis for Flooding in Urban Systems under Uncertainty 

Aronne Dell'Oca, Monica Riva, Alberto Guadagnini, and Leonardo Sandoval

The runoff process in environmental systems is influenced by various variables that are typically are affected by uncertainty. These include, for example, climate and hydrogeological quantities (hereafter denoted as environmental variables). Additionally, the runoff process is influenced by quantities that are amenable to intervention/design (hereafter denoted as operational variables) and can therefore be set to desired values on the basis of specific management choices. A key question in this context is: How do we discriminate the impact of operational variables, whose values can be decided in the system design or management phase, on system outputs considering also the action of uncertainty associated with environmental variables? We tackle this issue upon introducing a novel approach which we term Operational Sensitivity Analysis (OSA) and set within a Global Sensitivity Analysis (GSA) framework. OSA enables us to assess the sensitivity of a given model output specifically to operational factors, while recognizing uncertainty in the environmental variables. This approach is developed as a complement to a traditional GSA, which does not differentiate at the methodological level the nature of the type of variability associated with operational or environmental variables.

We showcase our OSA approach through an exemplary scenario associated with a urban catchment where flooding results from sewer system failure. In this context, we distinguish between operational variables, such as sewer system pipe properties and urban area infiltration capacity, and environmental variables such as, urban catchment drainage properties and rain event characteristics. Our results suggest that the diameter of a set of pipes in the sewer network is the most influential operational variable. As such, it provides a rigorous basis upon which one could plan appropriate actions to effectively manage the system response.

How to cite: Dell'Oca, A., Riva, M., Guadagnini, A., and Sandoval, L.: Operational Sensitivity Analysis for Flooding in Urban Systems under Uncertainty, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-19966, https://doi.org/10.5194/egusphere-egu24-19966, 2024.

EGU24-20013 | ECS | Orals | HS3.9

Field-scale soil moisture predictions using in situ sensor measurements in an inverse modelling framework: SWIM² 

Marit Hendrickx, Jan Diels, Jan Vanderborght, and Pieter Janssens

With the rise of affordable, autonomous sensors and IoT (Internet-of-Things) technology, it is possible to monitor soil moisture in a field online and in real time. This offers opportunities for real-time model calibration for irrigation scheduling. A framework is presented where realtime sensor data are coupled with a soil water balance model to predict soil moisture content and irrigation requirements at field scale. SWIM², Sensor Wielded Inverse Modelling of a Soil Water Irrigation Model, is a framework based on the DREAM inverse modelling approach to estimate 12 model parameters (soil and crop growth parameters) and their uncertainty distribution. These parameter distributions result in soil moisture predictions with a prediction uncertainty estimate, which enables a farmer to anticipate droughts and estimate irrigation requirements.

The SWIM² framework was validated based on three growing seasons (2021-2023) in about 30 fields of vegetable growers in Flanders. Kullback–Leibler divergence (KLD) was used as a metric to quantify information gain of the model parameters starting from non-informative priors. Performance was validated in two steps, i.e. the calibration period and prediction period, which is in correspondence with the real-world implementation of the framework. The RMSE, correlation (R, NSE) and Kling-Gupta efficiency (KGE) of soil moisture were analyzed in function of time, i.e. the amount of available sensor data for calibration.

Soil moisture can be predicted accurately after 10 to 20 days of sensor data is available for calibration. The RMSE during the calibration period is generally around 0.02 m³/m³, while the RMSE during the prediction period decreases from 0.04 to 0.02 m³/m³ when more calibration data is available. Information gain (KLD) of some parameters (e.g. field capacity and curve number) largely depends on the presence of dynamic events (e.g. precipitation events) during the calibration period. After 40 days of sensor data, the KGE and Pearson correlation of the calibration period become stable with median values of 0.8 and 0.9, respectively. For the validation period, the KGE and Pearson correlation are increasing in time, with median values from 0.3 to 0.7 (KGE) and from 0.7 to 0.95 (R). These good results show that, with this framework, we can simulate and predict soil moisture accurately. These predictions can in turn be used to estimate irrigation requirements.

Precipitation radar data was primarily considered as an input without uncertainty. As an extension, precipitation forcing error can be treated in DREAM by applying rainfall multipliers as additional parameters that are estimated in the inverse modelling framework. The multiplicative error of the radar data was quantified by comparison of radar data to rain gauge measurements. The prior uncertainty of the logarithmic multipliers was described by a Laplace distribution and was implemented in DREAM. The extended framework with rainfall multipliers shows better convergence and acceptance rate compared to the main framework. The calibration period shows better performance with higher correlations and lower RMSE values, but a decrease in performance was found for the validation period. These results suggest that the implementation of rainfall multipliers leads to overfitting, resulting in lower predictive power.

How to cite: Hendrickx, M., Diels, J., Vanderborght, J., and Janssens, P.: Field-scale soil moisture predictions using in situ sensor measurements in an inverse modelling framework: SWIM², EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-20013, https://doi.org/10.5194/egusphere-egu24-20013, 2024.

In recent years, Machine Learning (ML) models have led to a substantial improvement in hydrological predictions. It appears these models can distill information from catchment properties that is relevant for the relationship between meteorological drivers and streamflow, which has so far eluded hydrologists.
In the first part of this talk, I shall demonstrate some of our attempts towards understanding these improvements. Utilising Autoencoders and intrinsic dimension estimators, we have shown that the wealth of available catchment properties can effectively be summarised into merely three features, insofar as they are relevant for streamflow prediction. Hybrid models, which combine the flexibility of ML models with mechanistic mass-balance models, are equally adept at predicting as pure ML models but come with only a few interpretable interior states. Combining these findings will, hopefully, bring us closer to understanding what these ML models seem to have 'grasped'.
In the second part of the talk, I will address the issue of uncertainty quantification. I contend that error modelling should not be attempted on the residuals. Rather, we should model the errors where they originate, i.e., on the inputs, model states, and/or parameters. Such stochastic models are more adept at expressing the intricate distributions exhibited by real data. However, they come at the cost of a very large number of unobserved latent variables and thus pose a high-dimensional inference problem. This is particularly pertinent when our models include ML components. Fortunately, advances in inference algorithms and parallel computing infrastructure continue to extend the limits on the number of variables that can be inferred within a reasonable timeframe. I will present a straightforward example of a stochastic hydrological model with input uncertainty, where Hamiltonian Monte Carlo enables a comprehensive Bayesian inference of model parameters and the actual rain time-series simultaneously.

How to cite: Albert, C.: Advances and prospects in hydrological (error) modelling, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-20170, https://doi.org/10.5194/egusphere-egu24-20170, 2024.

EGU24-3548 | Posters on site | AS1.2

Improving the Completion of Weather Radar Missing Data with Deep Learning 

Aofan Gong, Haonan Chen, and Guangheng Ni

Weather radars commonly suffer from the data-missing problem that limits their data quality and applications. Traditional methods for the completion of weather radar missing data, which are based on radar physics and statistics, have shown defects in various aspects. Several deep learning (DL) models have been designed and applied to weather radar completion tasks but have been limited by low accuracy. This study proposes a dilated and self-attentional UNet (DSA-UNet) model to improve the completion of weather radar missing data. The model is trained and evaluated on a radar dataset built with random sector masking from the Yizhuang radar observations during the warm seasons from 2017 to 2019, which is further analyzed with two cases from the dataset. The performance of the DSA-UNet model is compared to two traditional statistical methods and a DL model. The evaluation methods consist of three quantitative metrics and three diagrams. The results show that the DL models can produce less biased and more accurate radar reflectivity values for data-missing areas than traditional statistical methods. Compared to the other DL model, the DSA-UNet model can not only produce a completion closer to the observation, especially for extreme values, but also improve the detection and reconstruction of local-scale radar echo patterns. Our study provides an effective solution for improving the completion of weather radar missing data, which is indispensable in radar quantitative applications.

How to cite: Gong, A., Chen, H., and Ni, G.: Improving the Completion of Weather Radar Missing Data with Deep Learning, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-3548, https://doi.org/10.5194/egusphere-egu24-3548, 2024.

EGU24-5373 | ECS | Orals | AS1.2 | Highlight

Convective environments in AI-models - What have AI-models learned about atmospheric profiles? 

Monika Feldmann, Louis Poulain-Auzeau, Milton Gomez, Tom Beucler, and Olivia Martius
The recently released suite of AI-based medium-range forecast models can produce multi-day forecasts within seconds, with a skill on par with the IFS model of ECMWF. Traditional model evaluation predominantly targets global scores on single levels. Specific prediction tasks, such as severe convective environments, require much more precision on a local scale and with the correct vertical gradients in between levels. With a focus on the North American and European convective season of 2020, we assess the performance of Panguweather, Graphcast and Fourcastnet for convective available potential energy (CAPE) and storm relative helicity (SRH) at lead times of up to 7 days.
Looking at the example of a US tornado outbreak on April 12 and 13, 2020, all models predict elevated CAPE and SRH values multiple days in advance. The spatial structures in the AI-models are smoothed in comparison to IFS and the reanalysis ERA5. The models show differing biases in the prediction of CAPE values, with Graphcast capturing the value distribution the most accurately and Fourcastnet showing a consistent underestimation.
By advancing the assessment of large AI-models towards process-based evaluations we lay the foundation for hazard-driven applications of AI-weather-forecasts.

How to cite: Feldmann, M., Poulain-Auzeau, L., Gomez, M., Beucler, T., and Martius, O.: Convective environments in AI-models - What have AI-models learned about atmospheric profiles?, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-5373, https://doi.org/10.5194/egusphere-egu24-5373, 2024.

EGU24-5571 | ECS | Orals | AS1.2

SHADECast: Enhancing solar energy integration through probabilistic regional forecasts 

Alberto Carpentieri, Doris Folini, Jussi Leinonen, and Angela Meyer

Surface solar irradiance (SSI) is a pivotal component in addressing climate change. As an abundant and non-fossil energy source, it is harnessed through photovoltaic (PV) energy production. As the contribution of PV to total energy production grows, the stability of the power grid faces challenges due to the volatile nature of solar energy, predominantly influenced by stochastic cloud dynamics. To address this challenge, there is a need for accurate, uncertainty-aware, near real-time, and regional-scale SSI forecasts with forecast horizons ranging from minutes to a few hours.

Existing state-of-the-art SSI nowcasting methods only partially meet these requirements. In our study, we introduce SHADECast [1], a deep generative diffusion model designed for probabilistic nowcasting of cloudiness fields. SHADECast is uniquely structured, incorporating deterministic aspects of cloud evolution to guide the probabilistic ensemble forecast, relying only on previous satellite images. Our model showcases significant advancements in forecast quality, reliability, and accuracy across various weather scenarios.

Through comprehensive evaluations, SHADECast demonstrates superior performance, surpassing the state of the art by 15% in the continuous ranked probability score (CRPS) over diverse regions up to 512 km × 512 km, extending the state-of-the-art forecast horizon by 30 minutes. The conditioning of ensemble generation on deterministic forecasts further enhances reliability and performance by more than 7% on CRPS.

SHADECast forecasts equip grid operators and energy traders with essential insights for informed decision-making, thereby guaranteeing grid stability and facilitating the smooth integration of regionally distributed PV energy sources. Our research contributes to the advancement of sustainable energy practices and underscores the significance of accurate probabilistic nowcasting for effective solar power grid management.

 

References

[1] Carpentieri A. et al., 2023, Extending intraday solar forecast horizons with deep generative models. Preprint at ArXiv. https://arxiv.org/abs/2312.11966 

How to cite: Carpentieri, A., Folini, D., Leinonen, J., and Meyer, A.: SHADECast: Enhancing solar energy integration through probabilistic regional forecasts, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-5571, https://doi.org/10.5194/egusphere-egu24-5571, 2024.

EGU24-5849 | ECS | Posters on site | AS1.2

Towards seamless rainfall and flood forecasting in the Netherlands: improvements to and validation of blending in pysteps 

Ruben Imhoff, Michiel Van Ginderachter, Klaas-Jan van Heeringen, Mees Radema, Simon De Kock, Ricardo Reinoso-Rondinel, and Lesley De Cruz

Flood early warning in fast responding catchments challenges our forecasting systems. It requires frequently updated, accurate and high-resolution rainfall forecasts to provide timely warning of rainfall amounts that will reach a catchment in the coming hours. The Netherlands is a typical example, with polder systems below sea level, a high level of urbanization and catchments with short response times. The need for better short-term rainfall forecasts is clearly present, but this is generally not feasible with numerical weather prediction (NWP) models alone. Hence, an alternative rainfall forecasting method is desirable for the first few hours into the future.

Rainfall nowcasting can provide this alternative but quickly loses skill after the first few hours. A promising way forward is a seamless forecasting system, which tries to optimally combine rainfall products from nowcasting and NWP. In this study, we applied the STEPS blending method to combine rainfall forecasts from ensemble radar nowcasts with those from the Harmonie-AROME configuration of the ACCORD NWP model in the Netherlands. This blending method is part of the open-source nowcasting initiative pysteps. To make blending possible in an operational setup, including the needs of involved water authorities, we made several adjustments to the blending implementation in pysteps, for instance:

  • We reduced the computational time by using a faster preprocessing and advection scheme.
  • We improved the noise initialization (needed for generating ensemble members) to allow for stable forecasts, also when one or both product(s) contain(s) no rain.
  • We enabled a dynamic disaggregation of the 1-hour resolution NWP forecasts to match the temporal resolution of the radar nowcast.

We operationalized the updated blending framework in the flood forecasting platforms of the involved water authorities. Given a forecast duration of 12 hours for the blended forecast and a 10-minute time step, average computation times are 3.4 minutes for a deterministic run and 12.3 minutes for an ensemble forecast with 10 members on a 4-core machine. Preprocessing takes approximately 10 minutes and only needs to occur when a new NWP forecast is issued. We tested the implementation for an entire, rainy summer month (July 15 to August 15, 2023) and analyzed the results over the entire domain. The results demonstrate that the blending method effectively combines radar nowcasts with NWP forecasts. Depending on the statistical score considered (such as RMSE and critical success index), the blending method performs either better or on par with the best-performing individual product (radar nowcast or NWP). A consistent finding is that the blending closely tracks the nowcast quality during the initial 1 to 2 hours of the forecast (in this study, the nowcast had lower errors than NWP during the first 2 – 2.5 hours), after which it gradually transitions into the NWP forecast. At longer lead times, the seamless product retains local precipitation structures and extremes better than the NWP product. It does this by leveraging information from the radar nowcast and the stochastic perturbations. Based on these results, a seamless forecasting approach can be regarded as an improvement for the involved water authorities.

How to cite: Imhoff, R., Van Ginderachter, M., van Heeringen, K.-J., Radema, M., De Kock, S., Reinoso-Rondinel, R., and De Cruz, L.: Towards seamless rainfall and flood forecasting in the Netherlands: improvements to and validation of blending in pysteps, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-5849, https://doi.org/10.5194/egusphere-egu24-5849, 2024.

EGU24-5909 | ECS | Posters on site | AS1.2

Impact of Spatial Density of Automatic Weather Station Data on Assimilation Effectiveness in WRF-3DVar Model 

Zeyu Qiao, Bu Li, Aofan Gong, and Guangheng Ni

Implementing the 3-Dimensional Variational (3DVar) data assimilation technique using high-density automatic weather station (AWS) observations substantially improves the precipitation simulation and forecast capabilities in the Weather Research and Forecasting (WRF) model. Given the impact of spatial distribution and quantity of observation data on assimilation effectiveness, there is a growing need to assimilate the most efficient amount of observation data to improve the precipitation forecast accuracy, especially in the context of the proliferation of data from diverse sources. This study investigates the impacts of spatial density of assimilated data on enhancing model predictions, focusing on a squall line event in Beijing on 2 August 2017 which has approximately 2400 AWSs in the simulation domain. Seven experiment groups assimilating varying proportions of AWS data (3.125, 6.25, 12.5, 25, 50, 75, and 100 percent of total AWSs) were conducted, comprising 10 experiments per group. The results were then compared with the experiment without data assimilation (CTRL) and the observations. Results show that while the WRF model roughly captured the evolution of this event, it overestimated the precipitation amount with significant deviations in precipitation locations. A general positive correlation was observed between the spatial density of assimilated data and the enhancement in model performance. However, there is a notable threshold beyond which additional data ceases to enhance forecast accuracy. The model performs best when the ratio of the number of assimilated AWSs to the model simulated area reaches 1/40 km-2. Moreover, significant variations in improvement effects across experiments within the same group indicate the substantial impact of spatial distribution of assimilated AWSs on forecast outcomes. This study provides a reference for devising more efficient and cost-effective data assimilation strategies in numerical weather prediction.

How to cite: Qiao, Z., Li, B., Gong, A., and Ni, G.: Impact of Spatial Density of Automatic Weather Station Data on Assimilation Effectiveness in WRF-3DVar Model, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-5909, https://doi.org/10.5194/egusphere-egu24-5909, 2024.

EGU24-6155 | ECS | Orals | AS1.2

Enhanced Foundation Model through Efficient Finetuning for Extended-Range Weather Prediction 

Shan Zhao, Zhitong Xiong, and Xiao Xiang Zhu

Weather forecasting is a vital topic in meteorological analysis, agriculture planning, disaster management, etc. The accuracy of forecasts varies with the prediction horizon, spanning from nowcasting to long-range forecasts. The extended range forecast, which predicts weather conditions beyond two weeks to months ahead, is particularly challenging. This difficulty arises from the inherent variability in weather systems, where minor disturbances in the initial condition can lead to significantly divergent future trajectories.

Numerical Weather Prediction (NWP) has been the predominant approach in this field. Recently, deep learning (DL) techniques have emerged as a promising alternative, achieving performance comparable to NWP [1, 2]. However, their lack of embedded physical knowledge often limits their acceptance within the research community. To enhance the trustworthiness of DL-based weather forecasts, we explore a transformer-based framework which considers complex geospatial-temporal (4D) processes and interactions. Specifically, we select the Pangu model [3] with a 24-hour lead time as the initial framework. To extend the prediction horizon to two weeks ahead, we employ a low-rank adaptation for model finetuning, which saves computation resources by reducing the number of parameters to only 1.1% of the original model. Besides, we incorporate multiple oceanic and atmospheric indices to capture a broad spectrum of global teleconnections, aiding in the selection of important features.

Our contributions are threefold: first, we provide an operational framework for foundation models, improving their applicability in versatile tasks by enabling training rather than limiting them to inference stages. Second, we demonstrate how to leverage these models with limited resources effectively and contribute to the development of green AI. Last, our method improves performance in extended-range weather forecasting, offering enhanced prediction skills, physical consistency, and finer spatial granularity. Our methodology achieved reduced RMSE on T2M, Z500, and T850 for 0.13, 139.2, and 0.52, respectively, compared to IFS. In the future, we plan to explore other settings, such as predicting precipitation and extreme temperatures.

REFERENCES
[1] Nguyen, Tung, et al. "ClimaX: A foundation model for weather and climate." arXiv preprint arXiv:2301.10343 (2023).
[2] Lam, Remi, et al. "Learning skillful medium-range global weather forecasting." Science (2023): eadi2336.
[3] Bi, Kaifeng, et al. "Accurate medium-range global weather forecasting with 3D neural networks." Nature 619.7970 (2023): 533-538.

How to cite: Zhao, S., Xiong, Z., and Zhu, X. X.: Enhanced Foundation Model through Efficient Finetuning for Extended-Range Weather Prediction, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-6155, https://doi.org/10.5194/egusphere-egu24-6155, 2024.

EGU24-6545 | Posters on site | AS1.2

Improving precipitation nowcasting using deep generative models: a case-study and experiences in R2O  

Kirien Whan, Charlotte Cambier van Nooten, Maurice Schmeits, Jasper Wijnands, Koert Schreurs, and Yuliya Shapovalova

Precipitation nowcasting is essential for weather-dependent decision-making. The combination of radar data and deep learning methods has opened new avenues for research. Deep learning approaches have demonstrated equal or better performance than optical flow methods for low-intensity precipitation, but nowcasting high-intensity events remains a challenge. We use radar data from the Royal Netherlands Meteorological Institute (KNMI) and explore various extensions of deep learning architectures (i.e. loss function, additional inputs) to improve nowcasting of heavy precipitation intensities. Our model outperforms other state-of-the-art models and benchmarks and is skilful at nowcasting precipitation for high rainfall intensities, up to 60-min lead time. 

Transferring research to operations is difficult for many meteorological institutes, particularly for new applications that use AI/ML methods. We discuss some of these challenges that KNMI is facing in this domain. 

How to cite: Whan, K., Cambier van Nooten, C., Schmeits, M., Wijnands, J., Schreurs, K., and Shapovalova, Y.: Improving precipitation nowcasting using deep generative models: a case-study and experiences in R2O , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-6545, https://doi.org/10.5194/egusphere-egu24-6545, 2024.

EGU24-6856 | Posters on site | AS1.2

Very Short-Range Precipitation Forecast in Korea Meteorological Administration 

ho yong lee, Jongseong Kim, Joohyung Son, and Seong-Jin Kim

Korea Meteorological Administration (KMA) has been providing the public with an hourly precipitation forecast updated every 10 minutes for the next 6 hours since 2015. This forecasts, named as the Very Short-Range Forecast (VSRF), differs from other longer forecasts ? such as short-range and medium-range forecasts issued by forecasters. The VSRF is automatically generated by a system based on two different models: MAPLE (McGill Algorithm for Precipitation nowcasting by Lagrangian Extrapolation) and KLAPS (Korea Local Analysis and Prediction System). 

MAPLE, based on Variational Echo Tracking (VET) from radar observations, has an intrinsic disadvantage: its performance decreases rapidly. On the other hand, numerical weather prediction systems like KLAPS are not initially as effective as MAPLE due to model balancing factors such as spin-up, but they maintain initial skill for a slightly longer period. Therefore, to provide the best predictions to the public, it is necessary to merge the two models properly. KMA conducted tests to determine the optimal way to utilize both models and established weights for each model based on their performance and precipitation tendencies. According to a 4-year evaluation, MAPLE outperforms for up to 2 hours, while KLAPS performs better after 4 hours. Consequently, the two models were merged with a hyperbolic tangent weight applied between 2 and 4 hours, and we named it as the best guidance. 

The best guidance was verified against precipitation observed by 720 raingauges over South Korea during the summer seasons from 2020 to 2023. It demonstrated better skill compared to both MAPLE and KLAPS. The average threat scores, with a rain intensity threshold of 0.5 mm/h throughout the forecast period, were 0.40 for the best guidance, 0.38 for MAPLE, and 0.35 for KLAPS.

The best guidance depends on both MAPLE and KLAPS. Therefore, KMA is actively working to improve the performance of each model. Additionally, a very short-range model based on AI is currently under development and running in semi-operations.

How to cite: lee, H. Y., Kim, J., Son, J., and Kim, S.-J.: Very Short-Range Precipitation Forecast in Korea Meteorological Administration, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-6856, https://doi.org/10.5194/egusphere-egu24-6856, 2024.

EGU24-6873 | Posters on site | AS1.2

Does more frequent Very Short-Range Forecast provide more useful information? 

Joohyung Son, Jongseong Kim, and Seongjin Kim

The Very Short-Range Forecast (VSRF) for precipitation from the Korea Meteorological Administration (KMA) is released every 10 minutes, providing forecasts for the next 6 hours at 10-minute intervals. However, when the forecast is provided to the public, it is updated at 10-minute interval, but only provides up to 6 hours at every hour. Consequently, from the public's perspective, forecasts for specific times may change every 10 minutes. While this allows users to access the latest updates, it also poses a challenge in terms of reduced reliability due to constantly changing predictions.

This study aims to assess the prediction performance and variability between forecasts released at 10-minute intervals and those at 1-hour intervals. We evaluated with the Very Short-Range Forecast numerical model KLAPS in VSRF and seek to determine which approach offers more valuable information from the public's standpoint. The assessment focuses on two distinct types of precipitation. The first involves convective showers, which sporadically appear over short durations, driven by atmospheric instability during the Korean Peninsula's summer. The second relates to systematic precipitation associated with a frontal boundary accompanying a medium-scale low-pressure system. For convective showers, the 1-hour interval exhibits better performance and continuity, particularly as the forecast time extends. In the case of systematic precipitation, the 1-hour interval remains superior, though the skill is not as prominent as with convective showers. This highlights that an abundance of information doesn't always equate to high-quality information.

How to cite: Son, J., Kim, J., and Kim, S.: Does more frequent Very Short-Range Forecast provide more useful information?, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-6873, https://doi.org/10.5194/egusphere-egu24-6873, 2024.

EGU24-7086 | Posters on site | AS1.2

Development of stadium-specific numerical forecast guidance for weather forecast for the 2024 Gangwon Winter Youth Olympic Games 

Yeon-Hee Kim, Eunju Cho, Sungbin Jang, Junsu Kim, Hyejeong Bok, and Seungbum Kim

The 2024 Gangwon Winter Youth Olympic Games (GANGWON 2024) will be held in the province of Gangwon in the Republic of Korea from January 19 to February 1, 2024, which already hosted the Olympic Winter Games PyeongChang 2018. In order to successfully host these first Winter YOG to be held in Asia, which will be held for the first time in Asia, it is necessary to provide customized weather information for decision-making in game operation and support in establishing game strategies for athletes and their teams. Accordingly, the Korea Meteorological Administration develops point-specific numerical forecast guidance for major stadiums and provides it to the field to support successful hosting of YOG and improvement of performance. Numerical forecast guidance is the final data delivered to consumers or forecasters as post-processed numerical model data that has been corrected by applying altitude correction and statistical methods to produce highly accurate forecasts. For a total of 13 forecast elements (temperature, minimum/maximum temperature, humidity, wind direction/speed, precipitation, new snow cover, sky conditions, precipitation probability, precipitation type), we developed user-customized numerical forecast guidance specialized for competition points  (Gangneung Olympic Park, Pyeongchang Alpensia Venue, Biathlon Center, Olympic Sliding Center departure/arrival, Wellyhilli departure/arrival, High1 departure/arrival). Through the process of Perfect Prognostic Method (PPM), Model Output Statistics (MOS), optimization, and optimal merging, the systematic errors inherent in the numerical model are removed, and the optimal data (BEST) with improved forecasting performance is provided as customized numerical forecast guidance specific to stadium locations.  In the prediction performance evaluation for the period of December 2023, the accuracy (improvement rate) compared to the average of available models was temperature 1.49℃ (18%), humidity 12% (25%), wind speed 1.87m/s (33%), and visibility 12.8km (17%).

How to cite: Kim, Y.-H., Cho, E., Jang, S., Kim, J., Bok, H., and Kim, S.: Development of stadium-specific numerical forecast guidance for weather forecast for the 2024 Gangwon Winter Youth Olympic Games, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-7086, https://doi.org/10.5194/egusphere-egu24-7086, 2024.

EGU24-7091 | Posters on site | AS1.2

The Development of precipitation model modifed with ECMWF IFS and XGBoost and its performance verification 

Eunju Cho, Yeon-Hee Kim, Seungbum Kim, and Young Cheol Kwon

This study was conducted to develop a modified precipitation model for its amount and existence by combining machine learning method, Extreme Gradient Boosting(XGBoost), with ECMWF IFS(Integrated forecasting system) and, finally, estimate the related performance.

According to the analysis of regional precipitation characteristic, prior to its development, the ratio of precipitation existence was various on a basis of a forecast’s district and its season. These different patterns on each district makes it necessary to develop the regional and seasonal model respectively.

And, the first attempt at the machine learning showed the importance of each feature as input-variables, as a result of which cloud physics-related features, for example large-area precipitation, total precipitation, visibility and what not, proved so significant. However, the insufficient amount of these feature’s data seemed to result in overfitting. And therefore, the feasible features, except for cloud physics-related things, of IFS data were used. In addition, auxiliary features and their gradient for every lead-time were calculated and added: relative vorticity, divergence, equivalent potential temperature, main 6 patterns for Korean summer and so on. The number of features amounted to around 144 with which for the 9-year training set, 2013~2021, based learning to be conducted regionally, followed by using validation-set of 2022.

As a result of validation for precipitation existence and its amount up to 135 hours ahead on the 10 regions at 00UTC in summer of 2022, Critical Success Index(CSI) was more improved by 10.3% than before. Accuracy(ACC) for each lead-time rose by 6% and its fluctuation also decreased. And the correction by this machine learning alleviated the overfitting trend of precipitation forecast amount produced by the original model, and improved correlation and linearity between observation and forecast. In particular, while the machine learning prevailed over the original model up to 100 hours ahead, from then on, both of them showed similar performance or that of the former was downward slightly. If the above-mentioned cloud physics features are used to further sharpen machine learning technique, its performance should be enhanced more and more.

How to cite: Cho, E., Kim, Y.-H., Kim, S., and Kwon, Y. C.: The Development of precipitation model modifed with ECMWF IFS and XGBoost and its performance verification, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-7091, https://doi.org/10.5194/egusphere-egu24-7091, 2024.

EGU24-7291 | ECS | Posters on site | AS1.2

Improvements in fog predictions via a modified reconstruction of moisture distribution using the Weather Research and Forecasting(WRF) model 

Eunji Kim, Soon-Young Park, Jung-Woo You, and Soon-Hwan Lee

Since fog is an important weather phenomenon affecting the traffic safety, accurate fog forecasting should be attained to minimize meteorological disasters. Most fog forecasts determine only the presence or absence of fog based on less visibility than 1 km, which is known as the visibility diagnostic method. During this process, fog could be predicted by the visibility calculated in the numerical weather prediction (NWP) model using the cloud liquid water content (LWC) near the surface. In this study, we investigated to increase the accuracy of fog forecast by optimizing the reconstruction of moisture distribution method, which can simulate the intensity of fog as well as the presence or absence of fog. The performances of the fog simulations were examined by modifying the relative humidity threshold at a height of 2 m and the stability parameters which affect turbulence and also one of the important criteria for fog occurrence. When we applied the optimize parameters to fog prediction in the winter seasons, the probability of detection (POD) has been increased significantly from 0.21 to 0.54. These improvements were attributed to the corrected relative humidity threshold and the stability parameters. Although the false alarm rate (FAR) remained almost unchanged, the critical success index (CSI) has been improved slightly lesser than those of the POD. When we analyzed the life cycle of fog, it takes time for the NWP model to simulate water droplets in the fog-developing stage. Therefore, the accuracy of the fog simulation is intimately related to the reconstruction of moisture distribution. The NWP model, however, showed a better performance in the process of fog dissipation than the reconstruction of moisture distribution method that was sensitive to temperature and turbulence. In conclusion, the reconstruction of moisture distribution led to a considerable improvement of the fog prediction in the generation and development stage since we used the optimized humidity threshold. It is also expected that accurate fog prediction could be achieved in the future by considering the aerosol effects, which is another importance factor for the fog generation.

How to cite: Kim, E., Park, S.-Y., You, J.-W., and Lee, S.-H.: Improvements in fog predictions via a modified reconstruction of moisture distribution using the Weather Research and Forecasting(WRF) model, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-7291, https://doi.org/10.5194/egusphere-egu24-7291, 2024.

EGU24-7536 | Orals | AS1.2

Nowcasting with Transformer-based Models using Multi-Source Data  

Çağlar Küçük, Apostolos Giannakos, Stefan Schneider, and Alexander Jann

Rapid advancements in data-driven weather prediction have shown notable success, particularly in nowcasting, where forecast lead times span just a few hours. Transformer-based models, in particular, have proven effective in learning spatiotemporal connections of varying scales by leveraging the attention mechanism with efficient space-time patching of data. This offers potential improvements over traditional nowcasting techniques, enabling early detection of convective activity and reducing computational costs. 

In this presentation, we demonstrate the effectiveness of a modified Earthformer model, a space-time Transformer framework, in addressing two specific nowcasting challenges. First, we introduce a nowcasting model that predicts ground-based 2D radar mosaics up to 2-hour lead time with 5-minute temporal resolution, using geostationary satellite data from the preceding two hours. Trained on a benchmark dataset sampled across the United States, our model exhibits robust performance against various impactful weather events with distinctive features. Through permutation tests, we interpret the model to understand the effects of input channels and input data length. We found that the infrared channel centered at 10.3 µm contains skillful information for all weather conditions, while, interestingly, satellite-based lightning data is the most skilled at predicting severe weather events in short lead times. Both findings align with existing literature, enhancing confidence in our model and guiding better usage of satellite data for nowcasting. Moreover, we found the model is sensitive to input data length in predicting severe weather events, suggesting early detection of convective activity by the model in rapidly growing fields. 

Second, we present the initial attempts to develop a multi-source precipitation nowcasting model for Austria, tailored to predict impactful events with convective activities. This model integrates satellite- and ground-based observations with analysis and numerical weather prediction data to predict precipitation up to 2-hour lead time with 5-minute temporal resolution.  

We conclude by discussing the broad spectrum of applications for such models, ranging from enhancing operational nowcasting systems to providing synthetic data to data-scarce regions, and the challenges therein.

How to cite: Küçük, Ç., Giannakos, A., Schneider, S., and Jann, A.: Nowcasting with Transformer-based Models using Multi-Source Data , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-7536, https://doi.org/10.5194/egusphere-egu24-7536, 2024.

EGU24-7753 | ECS | Orals | AS1.2

On the usefulness of considering the run-to-run variability for an ensemble prediction system 

Hugo Marchal, François Bouttier, and Olivier Nuissier

The run-to-run variability of numerical weather prediction systems is at the heart of forecasters' concerns, especially in the decision-making process when high-stakes events are considered. Indeed, forecasts that are brutally changing from one run to another may be difficult to handle and can lose credibility. This is all the more true nowadays, as many meteorological centres have adopted the strategy of increasing runs frequency, some reaching hourly frequencies. However, this aspect has received little attention in the literature, and the link with predictability has barely been explored.

In this study, run-to-run variability is investigated through 24h-accumulated precipitations forecasted by AROME-EPS, Météo-France's high resolution ensemble, which is refreshed 4 times a day. Focusing on the probability of some (warning) thresholds being exceeded, results suggest that how forecasts evolve over successive runs can be used to improve their skill, especially reliability. Various possible aspects of run sequence have been studied, from trends to rapid increases or decreases in event probability at short lags, also called "sneaks" or "phantoms", as well as the persistence of a non-zero probability through successive runs. The added value provided by blending successive runs, known as lagging, is also discussed.

How to cite: Marchal, H., Bouttier, F., and Nuissier, O.: On the usefulness of considering the run-to-run variability for an ensemble prediction system, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-7753, https://doi.org/10.5194/egusphere-egu24-7753, 2024.

EGU24-8449 | ECS | Orals | AS1.2

Radiation fog nowcasting with XGBoost using station and satellite data 

Michaela Schütz, Jörg Bendix, and Boris Thies

The research project “FOrecasting radiation foG by combining station and satellite data using Machine Learning (FOG-ML)” represents a comprehensive effort to advance radiation fog prediction using machine learning (ML) techniques, with focus on the XGBoost algorithm. The nowcasting period is up to four hours into the future.

The initial phase of the project involved developing a robust classification-based model that could accurately forecast the occurrence of radiation fog, a challenging meteorological phenomenon. Radiation fog is particularly difficult to predict because it depends on a complex interplay of factors such as ground cooling, humidity, and minimal cloud cover. It often forms rapidly and in local areas. This required careful analysis of the chronological order of the data and consideration of autocorrelation to increase the effectiveness of model training.

Building upon this foundation, the next two phases concentrated on improving the model’s forecasting performance for visibility classes (step 2) and for absolute visibility values (step 3). The main focus was then on a nowcasting period of up to two hours. This nowcasting period is critical in fog prediction as it directly impacts transportation planning and safety. The use of ground-level observations in step 2 and integration of satellite data in step 3 provided a rich dataset that allowed for more nuanced model training and validation.

In the latest phase of research, satellite data has been incorporated to further refine the prediction model, especially regarding the fog formation and dissipation. Satellite imagery provides additional variables of atmospheric data that are not readily available from ground-based observations. This integration aims to address one of the inherent limitations in fog forecasting methods, particularly in areas where ground-based observations are sparse.

Throughout the different stages, the project emphasized the need for thorough data processing and validation. This included the implementation of cross-validation techniques to assess the generalizability of the models and the use of various metrics to gauge their predictive power. This has also included the incorporation of trend information, which has proven to be crucial for forecasting with XGBoost. Our research has also shown that not only the overall performance, but also the performance of the transitions (fog formation and resolution) should be analyzed to get a complete picture of the model performance. This finding was consistent throughout the entire study, regardless of classification-based forecast or regression-based forecast.

We have been able to significantly improve the performance of our nowcasting model with each step. We will be presenting the key findings and latest results from this research at EGU24.

All results from step 1 can be found in “Current Training and Validation Weaknesses in Classification-Based Radiation Fog Nowcast Using Machine Learning Algorithms” from Vorndran et al. 2022. All results from step 2 can be found in “Improving classification-based nowcasting of radiation fog with machine learning based on filtered and preprocessed temporal data” from Schütz et al. 2023.

How to cite: Schütz, M., Bendix, J., and Thies, B.: Radiation fog nowcasting with XGBoost using station and satellite data, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-8449, https://doi.org/10.5194/egusphere-egu24-8449, 2024.

EGU24-9528 | ECS | Orals | AS1.2

Ensemble forecast post-processing based on neural networks and normalizing flows 

Peter Mlakar, Janko Merše, and Jana Faganeli Pucer

Ensemble weather forecast post-processing can generate more reliable probabilistic weather forecasts compared to the raw ensemble. Often, the post-processing method models the future weather probability distribution in terms of a pre-specified distribution family, which can limit their expressive power. To combat these issues, we propose a novel, neural network-based approach, which produces forecasts for multiple lead times jointly, using a single model to post-process forecasts at each station of interest. We use normalizing flows as parametric models to relax the distributional assumption, offering additional modeling flexibility.We evaluate our method for the task of temperature post-processing on the EUPPBench benchmark dataset. We show that our approach exhibits state-of-the-art performance on the benchmark, improving upon other well-performing entries. Additionally, we analyze the performance of different parametric distribution models in conjunction with our parameter regression neural network, to better understand the contribution of normalizing flows in the post-processing context. Finally, we provide a possible explanation as to why our method performs well, exploring per-lead time input importance.

How to cite: Mlakar, P., Merše, J., and Faganeli Pucer, J.: Ensemble forecast post-processing based on neural networks and normalizing flows, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-9528, https://doi.org/10.5194/egusphere-egu24-9528, 2024.

EGU24-9659 | Posters on site | AS1.2

Application Research of Multi-source New Detection Data in Snow Depth Prediction for Beijing Winter Olympics 

Jia Du, Bo Yu, Yi Dai, Sang Li, Luyang Xu, Jiaolan Fu, Lin Li, and Hao Jing

According to the demand of the Winter Olympic Organizing Committee for snow depth prediction, the application of multi-source new data in snow depth was studied based on densely artificial snow-depth measurement, microscopic snowflake shape observation and PARSIVEL data. The specific conclusions are as follows: (1) Most of the Snow-Liquid-Ratio(SLR) in Beijing competition zone was between 0.69 and 1.43 (unit: cm/mm, the same below), while that in Yanqing zone was between 0.53 and 1.17. But 7.5% of the SLRs in Yanqing zone exceeded 3.5, which all occurred in the same period of the key service time of 2022 Beijing Winter Olympics, making it more difficult to predict new snow depth. (2) The higher the SLR, the lower the daily minimum surface temperature and lowest air temperature.  Plate or column ice crystals, rimed snowflakes, and dendritic snowflakes were observed, whose corresponding SLRs increased. The average falling speed of particles falling below 2m/s can be used as an indicator of phase transfer. (3) The vertical distributions of temperature and humidity with SLR <1 or >2 were summarized. It was found that when the cloud area coincided with the dendritic growth zone with height close to Yanqing zone, the SLR would be more than 2, higher than that of Beijing zone. (4) A weather concept model generating large SLR was extracted. Snow in Beijing is often accompanied by easterly winds in boundary layer, which is easy to form a wet and ascending layer in the lower troposphere due to the blocking of western mountain. In the late winter season, helped by the temperature’s profile, it tends to produce unrimed dendritic snowflakes, leading to a great SLR.

How to cite: Du, J., Yu, B., Dai, Y., Li, S., Xu, L., Fu, J., Li, L., and Jing, H.: Application Research of Multi-source New Detection Data in Snow Depth Prediction for Beijing Winter Olympics, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-9659, https://doi.org/10.5194/egusphere-egu24-9659, 2024.

EGU24-9935 | ECS | Posters on site | AS1.2

Machine and Deep Learning algorithms to improve weather forecasts over a complex orography Mediterranean region 

Luca Furnari, Umair Yousuf, Alessio De Rango, Donato D'Ambrosio, Giuseppe Mendicino, and Alfonso Senatore

The rapid development of artificial intelligence algorithms has generated considerable interest in the scientific community. The number of scientific articles relating to applying these algorithms for weather forecasting has increased dramatically in the last few years. In addition, the recent operational launch of products such as GraphCast has put this area of research even more in the spotlight. This work uses different Machine Learning and Deep Learning algorithms, namely ANN (Artificial Neural Network), RF (Random Forest) and GNN (Graph Neural Network), with the aim to improve the short-term (1-day lead time) forecasts provided by a physically-based forecasting system. Specifically, the CeSMMA laboratory, since January 2020, has been producing daily forecasts accessible via the https://cesmma.unical.it/cwfv2/ webpage related to a large portion of southern Italy. The NWP (Numerical Weather Prediction) system is based on the WRF (Weather Research and Forecasting) model, with boundary and initial conditions provided by the GFS (Global Forecasting System) model. The AI algorithms post-process the NWP output, applying correction factors achieved by a two-year training considering the observations of the dense regional monitoring network composed of ca. 150 rain gauges.

The results show that the AI is able to improve daily rainfall forecasts compared to ground-based observations. Specifically, the ANN reduces the average MSE (Mean Square Error) by approximately 29% and the RF by 21% with respect to the WRF forecast for the whole study area (about 15’000 km2). Moreover, the GNN applied to a smaller area (considering only 22 rain gauges) further reduces the MSE by 35% during the heaviest rainfall months.

In addition to improving the performance of the forecast, the AI-based post-processing provides reasonable precipitation spatial patterns, reproducing the main physical phenomena such as the orographic enhancement since it is not a surrogate model and benefits from the original physically-based forecasts.

 

Acknowledgements. This work was funded by the Next Generation EU - Italian NRRP, Mission 4, Component 2, Investment 1.5, call for the creation and strengthening of ‘Innovation Ecosystems’, building ‘Territorial R&D Leaders’ (Directorial Decree n. 2021/3277) - project Tech4You - Technologies for climate change adaptation and quality of life improvement, n. ECS0000009. This work reflects only the authors’ views and opinions, neither the Ministry for University and Research nor the European Commission can be considered responsible for them.

How to cite: Furnari, L., Yousuf, U., De Rango, A., D'Ambrosio, D., Mendicino, G., and Senatore, A.: Machine and Deep Learning algorithms to improve weather forecasts over a complex orography Mediterranean region, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-9935, https://doi.org/10.5194/egusphere-egu24-9935, 2024.

On May 17, 2019, a rare severe convective weather occurred in Beijing, accompanied by local heavy rainstorm, hail, thunderstorm and gale. This severe convective weather occurred significantly earlier than normal years, bringing great challenge to the forecast. Using multiple observation data and radar four-dimensional variational assimilation products to analyze the triggering and development evolution of this severe convection. Under the conditions of no obvious weather scale system and local high potential unstable energy, the eastward advancement of the sea breeze front was the main factor triggering strong convection. As the northwest wind in the air increasing, the environmental conditions became stronger vertical wind shear, which was beneficial for the storm to maintain for a longer period of time. The supercell was the main cause of the convective weather. During the development of storms, they split into two parts and moved counterclockwise. The southern echo gradually weakened as it moved northward, while the northern echo moved southward, strengthening and developing into a super cell accompanied by a mesocyclone. The significant fluctuations in the height of the 0 ℃ layer within a small range resulted in different melting rates of hail during its descent, leading to the formation of spiky hail.

How to cite: Yu, B.: Analysis of a rare severe convective weather event in spring in Beijing of China, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-10809, https://doi.org/10.5194/egusphere-egu24-10809, 2024.

EGU24-12420 | ECS | Orals | AS1.2

Can convection permitting forecasts solve the tropical African precipitation forecasting problem? 

Felix Rein, Andreas H. Fink, Tilmann Gneiting, Philippe Peyrille, James Warner, and Peter Knippertz

Forecasting precipitation over Africa, the largest landmass in the tropics, has been a long standing problem. The unique conditions of the West African monsoon result in large and long lasting mesoscale convective systems. Global numerical weather prediction (NWP) models have gridsizes in the 10s of kilometers, particular when run in ensemble mode, leaving convection to be parameterized. This often results in precipitation being forecast on too large scales, in the wrong places, and with too weak intensity, ultimately leading to little to no skill in tropical Africa.


It has been argued that convection permitting (CP) NWP forecasts would cure some of the problems described above but those have only recently become feasible in an operational setting, although ensembles are still deemed to be too expensive. Here, we systematically compare regional deterministic CP and global ensemble forecasts in the region over multiple rainy seasons for the first time. We analyze CP forecasts from AROME and Met Office Tropical African Model, and seven global ensemble forecasts from the TIGGE archive, both individually and as a multi-model ensemble. In order to create an uncertainty estimate, we create neighborhood ensembles from CP forecasts at surrounding grid points, which allows for a fair comparison to the ensembles and a probabilistic climatology. Considering both precipitation occurrence and amount, we use the Brier score (BS) and the continuous ranked probability score (CRPS), along with their decompositions in discrimination, miscalibration and uncertainty, for evaluation.


Using neighborhood methods, deterministic forecasts are turned into probabilistic forecasts, allowing a fair comparison with ensembles. All numerical forecasts benefit from Neighborhoods, improving their BS and CRPS in terms of both miscalibration and discrimination. We find all individual forecasts to have skill over most of tropical Africa, with some ensemble models lacking skill in some regions and the multi model showing the most overall skill. The CP forecasts TAM and AROME outperform non-CP forecasts mainly in the region of the little dry Season and the Soud. However, large areas of low skill in terms of CRPS remain and even with high resolution, numerical models still struggle to predict precipitation in tropical Africa. 

How to cite: Rein, F., Fink, A. H., Gneiting, T., Peyrille, P., Warner, J., and Knippertz, P.: Can convection permitting forecasts solve the tropical African precipitation forecasting problem?, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-12420, https://doi.org/10.5194/egusphere-egu24-12420, 2024.

EGU24-12855 | Posters on site | AS1.2

Project IMA: Lessons Learned from Building the Belgian Operational Seamless Ensemble Prediction System 

Lesley De Cruz, Michiel Van Ginderachter, Maarten Reyniers, Alex Deckmyn, Idir Dehmous, Simon De Kock, Wout Dewettinck, Ruben Imhoff, Esteban Montandon, and Ricardo Reinoso-Rondinel

 

In recent years, several national meteorological services (NMSs) have invested considerable resources in the development of a seamless prediction system: rapidly updating forecasts that integrate the latest observations, covering timescales from minutes to days or longer ahead (e.g. DWD's SINFONY; FMI's ULJAS, MetOffice's IMPROVER and Geosphere Austria's SAPHIR) [1]. This move was motivated mainly by rising expectations from end users such as hydrological services, local authorities, the renewable energy sector and the general public. The development of seamless prediction systems was made possible thanks to the increasing availability of high-resolution observations, continuing advances in numerical weather prediction (NWP) models, nowcasting algorithms, and improved strategies to combine multiple information sources optimally. Moreover, the rise of AI/ML techniques in forecasting and nowcasting can further reduce the computational cost to generate frequently updating seamless operational forecast products.

 

We present the journey of building the Belgian seamless prediction system at the Royal Meteorological Institute of Belgium, with the working title "Project IMA". IMA uses both the deterministic INCA-BE and the probabilistic pysteps-BE systems to combine nowcasts with the ALARO and AROME configurations of the ACCORD NWP model. In the lessons learned along the way, we focus on what is often omitted, moving from research to operations, and integrating what we learn from operations back into research. We discuss the benefits of integrating new developments within the free and open-source software (FOSS) pysteps [2]. Our experience shows that using and contributing to FOSS not only leads to more transparency and reproducible, open science; it also enhances international collaboration and can benefit other users, including developing countries, bringing us a step closer to the ambitious goal of Early Warnings for All by 2027 [3].

 

References

 

[1] Bojinski, Stephan, et al. "Towards nowcasting in Europe in 2030." Meteorological Applications 30.4 (2023): e2124.

[2] Imhoff, Ruben O., et al. "Scale‐dependent blending of ensemble rainfall nowcasts and numerical weather prediction in the open‐source pysteps library." Quarterly Journal of the Royal Meteorological Society 149.753 (2023): 1335-1364.

[3] WMO, "Early warnings for all: Executive action plan 2023-2027", 8 Nov 2022,  https://www.preventionweb.net/quick/75125.

How to cite: De Cruz, L., Van Ginderachter, M., Reyniers, M., Deckmyn, A., Dehmous, I., De Kock, S., Dewettinck, W., Imhoff, R., Montandon, E., and Reinoso-Rondinel, R.: Project IMA: Lessons Learned from Building the Belgian Operational Seamless Ensemble Prediction System, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-12855, https://doi.org/10.5194/egusphere-egu24-12855, 2024.

Moderate to heavy rain produced by slantwise ascent of moist air above the cold high is prevalent in cold season in East China. The slantwise ascent is usually characterized by a southwest moist flow aroused by the so-called southern branch trough of 500hPa level to the south of the Qinghai-Tibet Plateau, while the cold high is usually formed by cold air damming, which is familiar to weather forecasters due to topographic feature of East China. The routine short-range forecast skill for this kind of precipitation of weather forecasters is usually limited by model performance. Through large sample model verification, our study indicates that, for the rainfall produced by southwesterly moist flow ascending above the cold high, the ECMWF model always underestimates the rainfall amount on the northeastern part of the rainfall belt, which could be taken as a systematic bias of the state-of-the-art global model. Our case studies indicate that the underestimation of rainfall amount is related to the weaker slant ascent of moist southwest flow forecast by ECMWF model than observation or reanalysis. The southwest flow above the northeastern flow induced by the cold high forms strong wind shear and warm-moist advection, which favors the occurrence of conditional symmetric instability producing strong slantwise ascent not well reflected by global model.

How to cite: Hu, N. and Fu, J.: Investigating Model Forecast Bias for Rainfall Produced by Slantwise Ascent above Cold High, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-13809, https://doi.org/10.5194/egusphere-egu24-13809, 2024.

EGU24-13853 | Orals | AS1.2 | Highlight

A Research Agenda for the Evaluation of AI-Based Weather Forecasting Models 

Imme Ebert-Uphoff, Jebb Q. Stewart, and Jacob T. Radford and the CIRA-NOAA team

Over the past few years purely AI-driven global weather forecasting models have emerged that show increasingly impressive skill, raising the question whether AI models might soon compete with NWP models for selected forecasting tasks. At this point these AI-based models are still in the proof-of-concept stage and not ready to be used for operational forecasting, but entirely new AI-models emerge every 2-3 months, with rapidly increasing abilities. Furthermore, many of these models are orders of magnitude faster than NWP models and can run on modest computational resources enabling repeatable on-demand forecasts competitive with NWP. The low computational cost enables the creation of very large ensembles, which better represent the tails of the forecast distribution, which, if an ensemble is well calibrated, allows for better forecasting of rare and extreme events.

However, these AI-based weather forecasting models have not yet been rigorously tested by the meteorological community, and their utility to operational forecasters is unknown. In this presentation we propose several studies to address the above issues, grouped into two central foci:

(1) Nature of AI models: AI-based models have very different characteristics from NWP models. Thus, in addition to applying evaluation procedures developed for NWP models, we need to develop procedures that test for AI-specific weaknesses. For example, NWP models and their physics backbone guarantee certain properties - such as dynamic coupling between fields - that AI-based models are not required to uphold. Developing suitable tests is based on a fundamental understanding of the AI-based models.

(2) Forecaster Perspective: Evaluation of weather forecasting models should be performed with respect to particular applications of weather forecasts, and it is critical to have research meteorologists and operational forecasters involved in the evaluation process. Our initial evaluation of AI-based models in CIRA weather briefings revealed that these models have characteristics that make interpretation of their forecasts fundamentally different from the physics-based NWP model predictions meteorologists are familiar with. For example, the increasing “blurriness” of AI-based predictions with longer lead times is not a reflection of weaker atmospheric circulations, but rather a reflection of uncertainty. Evaluations aimed at specific meteorological phenomena and atmospheric processes will allow the community to make informed decisions in the future regarding in what environments and for which applications AI-based weather forecasting models may be safe and beneficial to use.

In summary, AI-based weather forecasts have different characteristics from familiar dynamically-based forecasts, and it is thus important to have a robust research plan to evaluate many different characteristics of the models in order to provide guidelines to operational forecasters and feedback to model developers. In this abstract we propose a number of characteristics to evaluate, present results we already obtained, and suggest a research plan for future work.

How to cite: Ebert-Uphoff, I., Stewart, J. Q., and Radford, J. T. and the CIRA-NOAA team: A Research Agenda for the Evaluation of AI-Based Weather Forecasting Models, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-13853, https://doi.org/10.5194/egusphere-egu24-13853, 2024.

Large-eddy simulations of an idealized tropical cyclone (TC) were conducted as benchmarks to provide statistical information about subgrid convective clouds at a convection-permitting resolution over a TC convection system in different stages. The focus was on the vertical and spatial distributions of the subgrid cloud and associated mass flux that need to be parameterized in convection-permitting models. Results showed that the characteristics of the subgrid clouds varied significantly in various parts of the TC convection system. Statistical analysis revealed that the subgrid clouds were mainly located in the lower troposphere and exhibited shallow vertical extents of less than 4 km. The subgrid clouds were also classified into various cloud regimes according to the maximum mass flux height. Local subgrid clouds differed in mass-flux profile shape and magnitude at various regimes in the TC convection system.

How to cite: Zhang, X. and Bao, J.-W.: Statistics of the Subgrid Cloud of an Idealized Tropical Cyclone at Convection-Permitting Resolution, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-14232, https://doi.org/10.5194/egusphere-egu24-14232, 2024.

EGU24-14541 | Posters on site | AS1.2

Status and Plan of Standard Verification System for the NWP model in Korea Meteorological Administration 

Sora Park, Hyeja Park, Haejin Lee, Saem Song, Jong-Chul Ha, and Young Cheol Kwon

The Korea Meteorological Administration (KMA) has established and operated a standard verification system of the operational NWP models to evaluate the predictive performance of NWP model and compare them with other NWP models operated by domestic and foreign organization. This secures the objectivity of the verification results by applying the verification standards (WMO-No.485) presented by World Meteorological Organization (WMO), and being able to compare the performance with the numerical forecasting models of other institutions under the same conditions. The NWP models to be verified is a global, a regional, very short-range, and an ensemble prediction system and verification against analyses and observations are performed twice a day (00 UTC, 12 UTC). In addition to standard verification, precipitation, typhoon and various verification indexes (CBS index, KMA index, jumpiness index) are verified and used to evaluate the utilization of NWP models. The Korea Integrated Model (KIM), which is developed for Korea’s own NWP model, has been in operation since April 2020. Since the start of operation, the RMSE of 500hPa geopotential height (in Northern Hemisphere) has decreased every year, showing that forecast performance is improving. In addition, it can be seen that the 72-hour prediction accuracy for 12-hour accumulated precipitation (1.0 mm or more) in the Korean Peninsula area (75 ASOS stations) is also improving. As such, this study intends to discuss the predictive performance of the numerical forecast model based on the standard verification system and plans to improve the verification system in the future. 

How to cite: Park, S., Park, H., Lee, H., Song, S., Ha, J.-C., and Kwon, Y. C.: Status and Plan of Standard Verification System for the NWP model in Korea Meteorological Administration, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-14541, https://doi.org/10.5194/egusphere-egu24-14541, 2024.

EGU24-15431 | ECS | Orals | AS1.2 | Highlight

Nowcasting of extreme precipitation events: performance assessment of Generative Deep Learning methods 

Gabriele Franch, Elena Tomasi, Rishabh Umesh Wanjari, and Marco Cristoforetti

Radar-based precipitation nowcasting is one of the most prominent applications of deep learning (DL) in weather forecasting. The accurate forecast of extreme precipitation events remains a significant challenge for deep learning models, primarily due to their complex dynamics and the scarcity of data on such events. In this work we present the application of the latest state-of-the-art generative architectures for radar-based nowcasting, focusing on extreme event forecasting performance. We analyze a declination for the nowcasting task of all the three main current architectural approaches for generative modeling, namely: Generative Adversarial Networks (DGMRs), Latent Diffusion (LDCast), and our novel proposed Transformer architecture (GPTCast). These models are trained on a comprehensive 1-km scale, 5-minute timestep radar precipitation dataset that integrates multiple radar data sources from the US, Germany, the UK, and France. To ensure a robust evaluation and to test the generalization ability of the models, we concentrate on a collection of out-of-domain extreme precipitation events over the Italian peninsula extracted from the last 5 years. This focus allows us to assess the improvements these techniques offer compared to extrapolation methods, evaluating continuous (MSE, MAE) and categorical scores (CSI, POD, FAR), ensemble reliability, uncertainty quantification, and warning lead time. Finally, we analyze the computational requirements of these new techniques and highlight the caveats that must be considered when operational usage of these methods is envisaged. 

How to cite: Franch, G., Tomasi, E., Wanjari, R. U., and Cristoforetti, M.: Nowcasting of extreme precipitation events: performance assessment of Generative Deep Learning methods, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-15431, https://doi.org/10.5194/egusphere-egu24-15431, 2024.

EGU24-16617 | Posters on site | AS1.2

Forecasting extreme events with the crossing-point forecast  

Zied Ben Bouallegue

The crossing-point forecast (CPF) is a new type of ensemble-based forecast developed at the European Centre for Medium-Range Weather Forecasts. The crossing point refers to the intersection between the cumulative probability distribution of a forecast and the cumulative probability distribution of a model climatology. Originally, the CPF has emerged as a consistent forecast with the diagonal score, a weighted version of the continuous ranked probability score targeting high-impact events. Ranging between 0 and 1, the CPF can serve as an index for high-impact weather and thus directly be compared with the well-established extreme forecast index. The CPF is also interpretable in terms of a return period and conveys a sense of a “probabilistic worst-case scenario”.  Using a recent example of an extreme event affecting Europe, we illustrate and discuss the performance and specificities of this new type of forecast for extreme weather forecasting.

Ben Bouallegue, Z (2023).  Seamless prediction of high-impact weather events: a comparison of actionable forecasts. arXiv:2312.01673

How to cite: Ben Bouallegue, Z.: Forecasting extreme events with the crossing-point forecast , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-16617, https://doi.org/10.5194/egusphere-egu24-16617, 2024.

EGU24-17158 | Orals | AS1.2 | Highlight

AIFS – ECMWF’s Data-Driven Probabilistic Forecasting  

Zied Ben Bouallegue, Mihai Alexe, Matthew Chantry, Mariana Clare, Jesper Dramsch, Simon Lang, Christian Lessig, Linus Magnusson, Ana Prieto Nemesio, Florian Pinault, Baudouin Raoult, and Steffen Tietsche

In just two years, the idea of an operational data-driven system for medium-range weather forecasting has been transformed from dream to very real possibility. This has occurred through a series of publications from innovators, which have rapidly improved deterministic forecast skill. Our own evaluation confirms that these forecasts have comparable deterministic skill to NWP models across a range of variables. However, on medium-range timescales probabilistic forecasting, typically achieved through ensembles, is key for providing actionable insights to users. ECMWF is building on top of these recent works to develop a probabilistic forecasting system, AIFS. We will showcase results from our progress towards this system and outline our roadmap to operationalisation.

How to cite: Ben Bouallegue, Z., Alexe, M., Chantry, M., Clare, M., Dramsch, J., Lang, S., Lessig, C., Magnusson, L., Prieto Nemesio, A., Pinault, F., Raoult, B., and Tietsche, S.: AIFS – ECMWF’s Data-Driven Probabilistic Forecasting , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-17158, https://doi.org/10.5194/egusphere-egu24-17158, 2024.

The increasing integration of renewable energy resources to the national grids necessitates
accurate prediction of power generation from those sources in terms of secure operation of
electricity grid system and energy trading. Electricity generation of renewable energy power
plants such as wind and solar are inherently affected by weather conditions. The wind condition
particularly is affected by surface characteristics such as orography and vegetation, therefore it is
the one of the near surface atmospheric variables having the strongest local variability. The high-
resolution Numerical Weather Prediction (NWP) models are utilized to take the local conditions
into account. WRF model is the one of the most common NWP models having been widely
investigated by various researchers. On the other hand, The Model for Prediction Across Scales
(MPAS) is a relatively new NWP model utilizing non-uniform mesh structures, developed by the
National Center for Environmental Predictions (NCEP). However, there are limited studies in the
literature which compare the prediction performance of WRF and MPAS model in terms of
surface wind speed. This study evaluates the prediction accuracy of near surface wind of two
downscaled NWP models namely, WRF-ARW and MPAS. Both models are configured with
almost identical physics suites and initialized with 3 hourly 00-UTC initialization of Global
Forecast System (GFS) data. The model outputs are obtained at 10 minutes interval for 48 hours
horizon. Hourly averaged model results are compared with observations from 104 on-site
meteorological stations located in Turkiye having different complexity in terms of correlation
coefficient and RMSE.

How to cite: Yalcin, R. D., Yilmaz, M. T., and Yucel, İ.: Evaluation of the Impact of Uniform and Non-Uniform Resolution Implementations in Numerical Weather Prediction Models over the Accuracy of Short-Term Wind Prediction, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-17339, https://doi.org/10.5194/egusphere-egu24-17339, 2024.

EGU24-18548 | ECS | Posters on site | AS1.2

Enhancing Regional NWP Model with GNSS Zenith Total Delay Assimilation: A WRF and WRFDA 3D-Var Approach in the Greater Region of Luxembourg 

Haseeb Ur Rehman, Felix Norman Teferle, Addissu Hunegnaw, Guy Schumann, Florian Zus, and Rohith Muraleedharan Thundathil

Compared to alluvial floods, flash or pluvial floods are difficult to predict because they result from intense and brief periods of extreme precipitation. Luxembourg has a history of being impacted by floods, with notable occurrences in January 2011, May 2016, December 2017, January 2018, February 2019, and February 2020. However, July 2021 stands out as the most severe flood year on record in the region. In this study we are aiming to develop, a high-resolution numerical weather prediction (NWP) model for effective local heavy rainfall prediction in a nowcasting scenario and provide real time for flood simulation. The modeling relies on the Weather Research and Forecasting (WRF) model, which incorporates local Global Navigation Satellite System (GNSS) data assimilation and local precipitation observations to simulate small-scale, high-intensity convective precipitation.

As part of this, we will also test run the LISFlood flood model in an operational inundation forecast mode, meaning that the flood model will be run with the WRF precipitation forecasts as inputs.

The WRF model was configured for the Greater Region, utilizing a horizontal grid resolution of 12 km and incorporating high-resolution static datasets. Meteorological data i.e. July 13 -14 2021, from the Global Forecast System (GFS) were employed in the model setup as initial boundary condition. Zenith Total Delay (ZTD) data collected from various GNSS stations (112) across Germany and Luxembourg were assimilated into the model. Additionally, observational datasets including Surface Synoptic Observations (SYNOP), Upper Air Data, Radiosonde measurements (TEMP), and Tropospheric Airborne Meteorological Data Reporting (TAMDAR) were assimilated. Following this integration, an sensitivity analysis of various meteorological parameters such as precipitation, surface temperature (T2), and relative humidity was performed.

 

Keywords: NWP, WRF, Flash flood, LISFlood, Weather forecast, High-Resolution, GNSS, ZTD

How to cite: Rehman, H. U., Teferle, F. N., Hunegnaw, A., Schumann, G., Zus, F., and Muraleedharan Thundathil, R.: Enhancing Regional NWP Model with GNSS Zenith Total Delay Assimilation: A WRF and WRFDA 3D-Var Approach in the Greater Region of Luxembourg, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-18548, https://doi.org/10.5194/egusphere-egu24-18548, 2024.

EGU24-18938 | Posters on site | AS1.2

Forecasting tropical high-impact rainfall events using a hybrid statistical dynamical technique based on equatorial waves 

Samantha Ferrett, Gabriel Wolf, John Methven, Tom Frame, Christopher Holloway, Oscar Martinez-Alvarado, and Steve Woolnough

Recent work within the WCSSP FORSEA project and its successor FORWARDS has demonstrated that a hybrid statistical-dynamical forecasting technique combining model ensemble forecasts of equatorial waves with climatological rainfall statistics conditioned on wave phase and amplitude can provide additional skill in predicting high impact weather. The underlying rationale for the technique is twofold. Firstly that high impact rainfall events in the tropics are commonly associated with presence of equatorial waves; and secondly that while global models can adequately predict the evolution of dynamical structure of equatorial waves on time-scales of several days they do not predict the relationship between waves and rainfall well. In tests using the Met Office Global and Regional Forecasting System (MOGREPS) the hybrid forecast is found to outperform model rainfall forecasts from both the global and regional convection permitting versions of MOGREPS, however a weighted blend of the MOGREPS forecasts and the hybrid forecast was found to have the highest skill and further improvements in the method may be obtained by taking into consideration the effects of wave-superposition and interaction. To ascertain whether forecasts can be further improved by better predictions of wave amplitude and phase we compare to hypothetical best-case hybrid forecast computed using wave amplitudes and phases taken from reanalysis. This best-case scenario indicates that errors in forecasting all wave types diminish the hybrid forecast's skill, with the most significant reduction observed for Kelvin waves, suggesting that a significant improvement in the prediction of the propagation of equatorial waves would have a significant impact on rainfall prediction in the tropics. 

How to cite: Ferrett, S., Wolf, G., Methven, J., Frame, T., Holloway, C., Martinez-Alvarado, O., and Woolnough, S.: Forecasting tropical high-impact rainfall events using a hybrid statistical dynamical technique based on equatorial waves, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-18938, https://doi.org/10.5194/egusphere-egu24-18938, 2024.

EGU24-19257 | ECS | Orals | AS1.2

Dynamic Locally Binned Density Loss 

Jan Prosi, Sebastian Otte, and Martin V. Butz

In the field of precipitation nowcasting recent deep learning models now outperform traditional approaches such as optical flow [1,2]. Despite their principled effectiveness, these models and their respective training setups suffer from particular shortcomings.  For instance, they often rely on pixel-wise losses, which lead to blurred predictions by which the model expresses its uncertainty [2]. Additionally, these losses can negatively impact training dynamics by overly penalizing small spatial or temporal discrepancies between predictions and actual observations, i.e., the double penalty problem [3]. Generative methods such as discriminative losses or diffusion models do not suffer from the blurring effect as much [1, 4]. However, training these methods is complicated because training success is highly sensitive to the network architecture as well as to the learning setup and its parameterization [5].

Previous research has shown that spatial verification methods such as the fractions skill score offer an easy-to-implement alternative to solve the problem of pixel-wise losses [6, 7]. However, the fact that each pixel within the neighborhood of a spatial kernel is weighted equally poses a limiting factor to their performance and potential. Inspired by theories of cognitive modeling and in relation to the fractions skill score loss, we introduce a dynamic locally binned density (DLBD) loss: Forecasting target is not the actual precipitation in a grid cell but a target distribution, which encodes the density of binned precipitation values in a locally weighted area of grid cells. The loss is then determined via the cross-entropy of the predicted and the target distribution. We show that our novel prediction loss avoids the double penalty problem.  It thus diminishes the negative impact of small spatial offsets. Moreover, it enables the learning model to gradually shift focus towards progressively more accurate predictions.

We achieve best performance by simultaneously training on multiple concurrent forecasting targets that cover different local extents. We schedule the weighting of the loss terms such that the focus shifts from larger to smaller neighborhoods over the course of training. This way, the DL model first learns density dynamics and basic precipitation shifts. Later, it focuses on minimizing small spatial deviations, tuning into the local dynamics towards the end of training.  Our DLBD loss is easy-to-implement and shows great performance improvements.  We thus believe that DLBD losses can also be used by other forecasting architectures where the current forecasting loss precludes smooth loss landscapes.

 


1: Leinonen et al. 2023: Latent diffusion models for generative precipitation nowcasting with accurate uncertainty quantification
2: Espeholt et al. 2022: Deep learning for twelve hour precipitation forecasts
3: Grilleland et al. 2009: Intercomparison of spatial forecast verification methods.
4: Ravuri et al. 2021: Skilful precipitation nowcasting using deep generative models of radar
5: Mescheder et al. 2018: Which training methods for GANs do actually converge?
6: Roberts et al. 2008: Scale-selective verification of rainfall accumulations from high resolution forecasts of convective events.
7: Lagerquist et al. 2022: Can we integrate spatial verification methods into neural-network loss functions for atmospheric science?

How to cite: Prosi, J., Otte, S., and Butz, M. V.: Dynamic Locally Binned Density Loss, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-19257, https://doi.org/10.5194/egusphere-egu24-19257, 2024.

EGU24-19321 | ECS | Orals | AS1.2

Viability of satellite derived irradiance data for ML-based nowcasts 

Pascal Gfäller, Irene Schicker, and Petrina Papazek

Photovoltaic (PV) power production is increasingly becoming a central pillar in the shift to renewable power sources. The use of solar irradiance has great potential, as it is practically limitless and globally provides magnitudes more energy to the Earth than currently or foreseeable required. Solar irradiance as a power source does, however come with certain downsides. Besides the effects of seasonality and day-night-cycles on its usable potential, it´s broad use suffers mostly from uncertainty through its volatility. The actual extent of solar irradiance at the surface of the Earth is strongly influenced by a variety of atmospheric phenomena, most prominently clouds and atmospheric turbidity. The forecasting of near-future solar irradiance can thereby be beneficial in the estimation of PV power production in itself and with the goal of maintaining a stable equilibrium in electrical grids.

To achieve nowcasts on a larger grid scope, forecasting of solar irradiance from satellite data can substitute forecasting of power output for individual sites. Satellite data, in contrast to ground-based data sources or NWP model estimates, is less reliant on the proper workings of a wide range of externalities. General-purpose spatiotemporal neural networks can be adapted to this task and provide predictions within a very short timeframe, with no requirement of HPC-infrastructure. A sparse model relying on a single satellite-based data source has less points of failure that could affect its forecasting performance and can be very efficient, but this sparsity could also reduce the achievable predictive accuracy. Benefits of smaller and simpler forecasting pipelines therefore may need to be balanced with requirements in terms of accuracy.

To gather more meaningful and reliable results, a variety of spatiotemporal neural networks is implemented and tested to provide a more meaningful foundation. The models were selected and evaluated with respect to their different architectural patterns and designs, to get a notion of architectures beneficial to this task and achieve a more generalizable argument concerning the use satellite data as the sole basis of solar irradiance nowcasting.

In an attempt of improving the viability of satellite-based nowcasting a commonly occurring flaw in near-real-time satellite data sources, missing or skipped frames, solutions to mitigate issues in operational nowcasting are considered. In place of ad-hoc preprocessing such as interpolation of missing data frames, an attempt to condition the models to missing frames is made.

How to cite: Gfäller, P., Schicker, I., and Papazek, P.: Viability of satellite derived irradiance data for ML-based nowcasts, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-19321, https://doi.org/10.5194/egusphere-egu24-19321, 2024.

EGU24-19377 | ECS | Posters on site | AS1.2

Advancing Spatiotemporal Rainfall Nowcasting through Deep Learning Techniques 

Ahmed Abdelhalim, Miguel Rico-Ramirez, Weiru Liu, and Dawei Han

For weather forecasters and hydrologists, predicting rainfall in the short term – minutes to a few hours – is crucial for a range of applications. While traditional nowcasting methods excel in operational settings, they face limitations in predicting convective storm formation and high-intensity events. Enter deep learning, a powerful tool transforming numerous fields. Convolutional neural networks, in particular, have shown promise in improving nowcasting accuracy. These networks can learn complex patterns and relationships within data, like the intricate tapestry of rainfall variations observed in historical radar sequences. However, capturing long-term dependencies in this data remains a challenge, resulting in fuzzy nowcasts and underestimating high-intensity events. This study proposes a novel deep learning model that goes beyond simple extrapolation, effectively capturing both the spatial correlations and temporal dependencies within rainfall data. Our hybrid convolutional neural network architecture tackles this challenge through three key components: Decoder & Encoder: These modules focus on unraveling the intricate spatial patterns of rainfall and a temporal Module to learn the subtle long-term evolutions and interactions between rain cells over time. By capturing these temporal dependencies, the model can produce more accurate forecasts. To evaluate the model performance, it is compared against both deep learning and optical flow baselines. This presentation will introduce the model and provide a summary of its performance in spatiotemporal rainfall nowcasting.

Keywords: deep learning; spatiotemporal encoding, rainfall nowcasting; radar; optical flow

How to cite: Abdelhalim, A., Rico-Ramirez, M., Liu, W., and Han, D.: Advancing Spatiotemporal Rainfall Nowcasting through Deep Learning Techniques, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-19377, https://doi.org/10.5194/egusphere-egu24-19377, 2024.

EGU24-19699 | ECS | Posters on site | AS1.2

Evaluation of seamless forecasts for severe weather warnings  

Verena Bessenbacher, Jonas Bhend, Lea Beusch, Daniele Nerini, Colombe Siegenthaler, Christoph Spirig, and Lionel Moret

At MeteoSwiss, NWP and ML-based models are run operationally on a daily basis to provide weather forecasts and weather warnings for the general public. These forecasts come from various models that differ in lead times, initialization frequency, spatial resolution, and extents. We aim at combining those sources into a probabilistic, gridded weather forecast that is seamless in space and time. Creating a seamless forecast needs careful post-processing so as not to introduce cut-offs or unphysical behavior at the seams between the model runs. This includes using multiple forecast sources and forecast initializations (called lagged ensembles) and combining these using comprehensive blending methods. 

The first minimal viable product of a seamless forecast is currently being produced at MeteoSwiss, and will soon be available to the forecasters in real time. 

We evaluate the merit of these forecasts in terms of warning thresholds for rain and wind gusts. To do so, we compare reforecasts and observations from ground stations as well as rain radar observations from a set of past severe weather events over Switzerland. We benchmark the seamless forecast with individual forecast sources and post-processed products to evaluate the added value of seamlessly combining different forecast sources into one blended product. We furthermore plan to compare different methods for blending between sources soon.

How to cite: Bessenbacher, V., Bhend, J., Beusch, L., Nerini, D., Siegenthaler, C., Spirig, C., and Moret, L.: Evaluation of seamless forecasts for severe weather warnings , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-19699, https://doi.org/10.5194/egusphere-egu24-19699, 2024.

Integrating the hybrid and multiscale analyses and the parallel computation is necessary for current data assimilation schemes. A local data assimilation method, Local DA, is designed to fulfill these needs. This algorithm follows the grid-independent framework of the local ensemble transform Kalman filter (LETKF) and is more flexible in hybrid analysis than the LETKF. Local DA employs an explicitly computed background error correlation matrix of model variables mapped to observed grid points/columns. This matrix allows Local DA to calculate static covariance with a preset correlation function. It also allows using the conjugate gradient (CG) method to solve the cost function and allows performing localization in model space, observation space, or both spaces (double-space localization). The Local DA performance is evaluated with a simulated multiscale observation network that includes sounding, wind profiler, precipitable water vapor, and radar observations. In the presence of a small-size time-lagged ensemble, Local DA can produce a small analysis error by combining multiscale hybrid covariance and double-space localization. The multiscale covariance is computed using error samples decomposed into several scales and independently assigning the localization radius for each scale. Multiscale covariance is conducive to error reduction, especially at a small scale. The results further indicate that applying the CG method for each local analysis does not result in a discontinuity issue. The wall clock time of Local DA implemented in parallel is halved as the number of cores doubles, indicating a reasonable parallel computational efficiency of Local DA.

How to cite: Wang, S. and Qiao, X.: A Local Data Assimilation Method (Local DA v1.0) and its Application in a Simulated Typhoon Case, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-21770, https://doi.org/10.5194/egusphere-egu24-21770, 2024.

EGU24-21772 | Orals | AS1.2

Calibration of Convective-scale Hourly Precipitation Based on the Frequency-Matching Method 

Xiaoshi Qiao, Shizhang Wang, and Mingjian Zeng

Calibration of convective-scale hourly precipitation based on the frequency-matching method was carried on using CMPASS observation and CMA-MESO 3km forecast data. The character of hourly precipitation bias was studied.The effect of frequency-matching method (FMM) on the bias correction of CMA-MESO 3km hourly precipitation forecasts was analyzed. In the bias characteristic analysis, the differences in precipitation intensity in different regions of the country and the differences in precipitation in different months were considered. The whole country was divided into 7 sub-regions for monthly analysis. In the bias correction based on the frequency-matching method, the daily variations of precipitation bias and the impact of increasing and decreasing precipitation values on the corrected precipitation scores were analyzed. The results show that CMA-MESO 3km forecasts have a wet bias in light rainfall in the cold season, while a dry bias dominates in moderate to heavy rainfall. In the warm season, except for the Tibet region, the hourly precipitation forecast bias of CMA-MESO 3km shows significant daily variations, with more precipitation in the afternoon and less at night and in the morning, especially for heavy rainfall. Therefore, whether to consider the daily variations of precipitation bias in the use of FMM correction mainly reflects in the summer, especially at night and in the morning. Considering the daily variations of precipitation bias is beneficial to improving the forecast skills (TS scores) for nighttime and morning in the summer. Further analysis shows that the positive contribution of FMM correction to forecast scores mainly comes from the increase in frequency adjustment, especially for heavy rainfall. However, for light rainfall with wet bias, FMM often results in negative contribution. Therefore, FMM has a significant improvement effect on heavy rainfall in winter and nighttime rainfall in summer. The reason for this result is that the hit rate of CMA-MESO hourly precipitation forecast is low, and the false alarm rate is generally high, especially for heavy rainfall. In this case, the increased precipitation significantly increases the hit rate, while the false alarm rate increases to a lesser extent, thereby improving the precipitation scores.

How to cite: Qiao, X., Wang, S., and Zeng, M.: Calibration of Convective-scale Hourly Precipitation Based on the Frequency-Matching Method, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-21772, https://doi.org/10.5194/egusphere-egu24-21772, 2024.

The latest assessment report (AR6) of the Intergovernmental Panel on Climate Change includes a new element to climate research, i.e. the Interactive Atlas (IA), which is very useful for users from different sectors. As the new CMIP6 global climate model simulations use the brand-new SSP-scenarios paired with the RCP-scenarios, the latest climate change projections should be evaluated in order to update the regional and national adaptation strategies. Keeping this in mind we focused on Europe, with a special emphasis on Hungary in our study.

Our aim was to analyse the potential future changes of different temperature indices for Europe, in order to recognize spatial patterns and trends that may shape our climate in the second half of the 21st century. For this purpose, multi-model mean simulation data provided by the IPCC AR6 WG1 IA were downloaded on a monthly base. We chose two climate indices beside the mean temperature values, which represent temperature extremes, namely, the number of days with maximum temperature above 35 °C and the number of frost days (i.e. when daily minimum temperature is below 0 °C). We focused on the end of the 21st century (2081–2100) with also briefly considering the medium-term changes of the 2041–2060 period (both compared to the last two decades of the historical simulation period, i.e. 1995–2014 as the reference period). For both future periods we used all scenarios provided in the IA, namely, SSP1-2.6, SSP2-4.5, SSP3-7.0 and SSP5-8.5.

Several zonal and meridional segments over the continent were defined, where we analysed the projected changes of the indices. The zonal segments provide an insight on two different effects that may induce spatial differences between future regional changes. (i) Continentality can be recognized as an increasing effect from the western parts of the segment towards the east. (ii) Topography also appears as the influence of mountains, plains, and basins emerge. The meridional segments provide information about the north-to-south differences as well, as the effects of sea cover. The changes in the indices are plotted on diagrams representing the different months, where the differences in the scenarios are also shown. These diagrams are compared to their respective landscape profiles, furthermore, statistical parameters were calculated. In addition, a monotony index was defined as the cumulative direction of differences between the neighbouring grid cells and analysed within the study.

Our results show that in the changes of mean temperature, both the zonal location and sea cover will play a key role in forming spatial differences within Europe. However, for the extreme temperature indices, topography and continentality are likely to become more dominant than sea cover, while the zonal location remains an important factor. 

Acknowledgements: This work was supported by the Hungarian National Research, Development and Innovation Fund [grant numbers PD138023, K-129162], and the National Multidisciplinary Laboratory for Climate Change [grant number RRF-2.3.1-21-2022-00014]. 

How to cite: Divinszki, F., Kis, A., and Pongrácz, R.: Analysing the projected monthly changes of temperature-related climate indices over Europe using zonal and meridional segments based on CMIP6 data, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-389, https://doi.org/10.5194/egusphere-egu24-389, 2024.

EGU24-868 | ECS | Posters on site | CL4.3

Relationship of the predictability of North Pacific Mode and ENSO with predictability of PDO 

Jivesh Dixit and Krishna M. AchutaRao

PDO and ENSO are most prominent variability modes in the Pacific Ocean at decadal and interannual timescales respectively. Mutual independence between ENSO and PDO is questionable (Chen & Wallace, 2016). Linear combination of the first two orthogonal modes of SST variability in our Study Region (SR; 70oN - 20oS, 110oE - 90oW) i.e. mode 1 (interannual mode, we call it, IAM; ENSO like variability) and mode 2 (North Pacific Mode (NPM; Deser & Blackmon (1995)); a decadal mode) produces a PDO like variability (Chen & Wallace, 2016). It suggests that PDO is not independently hosted in the Pacific Ocean and can be represented by two linearly independent variability modes.

To produce credible and skillful climate information at multi-year to decadal timescales, Decadal Climate Prediction Project (DCPP), led by the Working Group on Subseasonal to Interdecadal Prediction (WGSIP), focuses on both the scientific and practical elements of forecasting climate by employing predictability research and retrospective analyses within the Coupled Model Intercomparison Project Phase 6 (CMIP6). Component A under DCPP experiments concentrates on hindcast experiments to examine the prediction skill of participating models with respect to actual observations.

As linear combination of  IAM and NPM in SR produces PDO pattern and timescales efficiently, we compared the  ability of DCPP-A hindcasts to predict  IAM, NPM, and  PDO. In this analysis we use output from 9 models (a total of 128 ensemble members), initialised every year from 1960 to 2010. To produce the prediction skill estimates.

At lead year 1 from initialisation, the prediction of NPM,  IAM and PDO is quite skillful as the models are initialised with observations. In subsequent years, skill of either IAM or NPM or both drop significantly and that leads to drop in skill of predicted PDO index. Both the deterministic estimates and probabilistic estimates of prediction skill for DCPP hindcast experiments suggest that the ability of hindcast experiments to predict NPM governs the prediction skill to predict PDO index.

Keywords: PDO, ENSO, NPM, CMIP6, DCPP, hindcast

References

Chen, X., & Wallace, J. M. (2016). Orthogonal PDO and ENSO indices. Journal of Climate, 29(10), 3883–3892. https://doi.org/10.1175/jcli-d-15-0684.1

Deser, C., & Blackmon, M. L. (1995). On the Relationship between Tropical and North Pacific Sea Surface Temperature Variations. Journal of Climate, 8(6), 1677–1680. https://doi.org/10.1175/1520-0442(1995)008<1677:OTRBTA>2.0.CO;2

How to cite: Dixit, J. and AchutaRao, K. M.: Relationship of the predictability of North Pacific Mode and ENSO with predictability of PDO, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-868, https://doi.org/10.5194/egusphere-egu24-868, 2024.

EGU24-1757 | Posters on site | CL4.3

Is the NAO signal-to-noise paradox exacerbated by severe winter windstorms? 

Lisa Degenhardt, Gregor C. Leckebusch, Adam A. Scaife, Doug Smith, and Steve Hardiman

The signal-to-noise paradox is known to be a limitation in multiple seasonal and decadal forecast models where the model ensemble mean predicts observations better than individual ensemble members. This ‘paradox’ occurs for different parameters, like the NAO, temperature, wind speed or storm counts in multiple seasonal and decadal forecasts. However, investigations have not yet found the origin of the paradox. First hypotheses are that weak ocean – atmosphere coupling or a misrepresentation of eddy feedback in these models is responsible.

Our previous study found a stronger signal-to-noise error in windstorm frequency than for the NAO despite highly significant forecast skill. In combination with the underestimation of eddy feedback in multiple models, this led to the question: Might the signal-to-noise paradox over the North-Atlantic be driven by severe winter windstorms?

To assess this hypothesis, the signal-to-noise paradox is investigated in multiple seasonal forecast suites from the UK Met Office, ECMWF, DWD and CMCC. The NAO is used to investigate the changes in the paradox depending on the storminess of the season. The results show a significant increase of the NAO-signal-to-noise error in stormy seasons in GloSea5. Other individual models like the seasonal model of the DWD or CMCC do not show such a strong difference. A multi-model approach, on the other hand, shows the same tendency as GloSea5. Nevertheless, these model differences mean that more hindcasts are needed to conclusively demonstrate that the signal-to-noise error arises from Atlantic windstorms.

How to cite: Degenhardt, L., Leckebusch, G. C., Scaife, A. A., Smith, D., and Hardiman, S.: Is the NAO signal-to-noise paradox exacerbated by severe winter windstorms?, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-1757, https://doi.org/10.5194/egusphere-egu24-1757, 2024.

EGU24-1940 | ECS | Orals | CL4.3

Study of the Decadal Predictability of Mediterranean Sea Surface Temperature Based on Observations 

Xiaoqin Yan, Youmin Tang, and Dejian Yang

Sea surface temperature (SST) changes in the Mediterranean Sea have profound impacts on both the Mediterranean regions and remote areas. Previous studies show that the Mediterranean SST has significant decadal variability that is comparable with the Atlantic multidecadal variability (AMV). However, few studies have discussed the characteristics and sources of the decadal predictability of Mediterranean SST based on observations. Here for the first time we use observational datasets to reveal that the decadal predictability of Mediterranean SST is contributed by both external forcings and internal variability for both annual and seasonal means, except that the decadal predictability of the winter mean SST in the eastern Mediterranean is mostly contributed by only internal variability. Besides, the persistence of the Mediterranean SST is quite significant even in contrast with that in the subpolar North Atlantic, which is widely regarded to have the most predictable surface temperature on the decadal time scale. After the impacts of external forcings are removed, the average prediction time of internally generated Mediterranean SST variations is more than 10 years and closely associated with the multidecadal variability of the Mediterranean SST that is closely related to the accumulated North Atlantic Oscillation forcing.

How to cite: Yan, X., Tang, Y., and Yang, D.: Study of the Decadal Predictability of Mediterranean Sea Surface Temperature Based on Observations, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-1940, https://doi.org/10.5194/egusphere-egu24-1940, 2024.

EGU24-3190 | ECS | Orals | CL4.3

Seasonal forecasting of the European North-West shelf seas: limits of winter and summer sea surface temperature predictability 

Jamie Atkins, Jonathan Tinker, Jennifer Graham, Adam Scaife, and Paul Halloran

The European North-West shelf seas (NWS) support economic interests and provide environmental services to several adjacent populous countries. Skilful seasonal forecasts of the NWS would be useful to support decision making. Here, we quantify the skill of an operational large-ensemble ocean-atmosphere coupled dynamical forecasting system (GloSea), as well as a benchmark persistence forecasting system, for predictions of NWS sea surface temperature (SST) at 2-4 months lead time in winter and summer. We also identify sources of- and limits to NWS SST predictability with a view to what additional skill may be available in the future. We find that GloSea NWS SST skill is generally high in winter and low in summer. Persistence of anomalies in the initial conditions contributes substantially to predictability. GloSea outperforms simple persistence forecasts, by adding atmospheric variability information, but only to a modest extent. Where persistence is low – for example in seasonally stratified regions – both GloSea and persistence forecasts show lower skill. GloSea skill can be degradeded by model deficiencies in the relatively coarse global ocean component, which lacks a tidal regime and likely fails to properly fine-scale NWS physics. However, using “near perfect atmosphere” tests, we show potential for improving predictability of currently low performing regions if atmospheric circulation forecasts can be improved, underlining the importance of development of atmosphere-ocean coupled models for NWS seasonal forecasting applications.

How to cite: Atkins, J., Tinker, J., Graham, J., Scaife, A., and Halloran, P.: Seasonal forecasting of the European North-West shelf seas: limits of winter and summer sea surface temperature predictability, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-3190, https://doi.org/10.5194/egusphere-egu24-3190, 2024.

EGU24-4538 | ECS | Orals | CL4.3

Statistical downscaling of extremes in seasonal predictions - a case study on spring frosts for the viticultural sector 

Sebastiano Roncoroni, Panos Athanasiadis, and Silvio Gualdi

Spring frost events occurring after budburst of grapevines can damage new shoots, disrupt plant growth and cause large economic losses to the viticultural sector. Frost protection practices encompass a variety of vineyard management actions across timescales, from seasonal to decadal and beyond. The cost-effectiveness of such measures depends on the availability of accurate predictions of the relevant climate hazards at the appropriate timescales.

In this work, we present a statistical downscaling method which predicts variations in the frequency of occurrence of spring frost events in the important winemaking region of Catalunya at the seasonal timescale. The downscaling method exploits the seasonal predictability associated with the predictable components of the atmospheric variability over the Euro-Atlantic region, and produces local predictions of frost occurrence at a spatial scale relevant to vineyard management.

The downscaling method is designed to address the specific needs highlighted by a representative stakeholder in the local viticultural sector, and is expected to deliver an actionable prototype climate service. The statistical procedure is developed in perfect prognosis mode: the method is trained with large-scale reanalysis data against a high-resolution gridded observational reference, and validated against multi-model seasonal hindcast predictions.

Our work spotlights the potential benefits of transferring climate predictability across spatial scales for the design and provision of usable climate information, particularly regarding extremes.

How to cite: Roncoroni, S., Athanasiadis, P., and Gualdi, S.: Statistical downscaling of extremes in seasonal predictions - a case study on spring frosts for the viticultural sector, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-4538, https://doi.org/10.5194/egusphere-egu24-4538, 2024.

EGU24-4873 | ECS | Orals | CL4.3

Why does the Signal-to-Noise Paradox Exist in Seasonal Climate Predictability? 

Yashas Shivamurthy, Subodh Kumar Saha, Samir Pokhrel, Mahen Konwar, and Hemant Kumar Chaudhari

Skillful prediction of seasonal monsoons has been a challenging problem since the 1800s. However, significant progress has been made in Indian summer monsoon rainfall prediction in recent times, with skill scores reaching 0.6 and beyond, surpassing the estimated predictability limits. This phenomenon leads to what is known as the “Signal-to-noise Paradox.” To investigate this paradox, we utilized 52 ensemble member hindcast runs spanning 30 years.

Through the application of ANOVA and Mutual Information methods, we estimate the predictability limit globally. Notably, for the boreal summer rainfall season, the Indian subcontinent exhibited the paradox, among several other regions, while the Equatorial Pacific region, despite demonstrating high prediction skill, does not have the Signal-to-Noise paradox. We employed a novel approach to understand how sub-seasonal variability and their projection in association with predictors are linked to the paradoxical behavior of seasonal prediction skill.

We propose a new method to estimate predictability limits that is free from paradoxical phenomena and shows much higher seasonal predictability. This novel method provides valuable insights into the complex dynamics of monsoon prediction, thereby creating opportunities for expanded research and potential improvements in seasonal forecasting skill in the coming years.

How to cite: Shivamurthy, Y., Saha, S. K., Pokhrel, S., Konwar, M., and Chaudhari, H. K.: Why does the Signal-to-Noise Paradox Exist in Seasonal Climate Predictability?, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-4873, https://doi.org/10.5194/egusphere-egu24-4873, 2024.

EGU24-7134 | ECS | Orals | CL4.3

Towards the Predictability of Compound Dry and Hot Extremes through Complexity Science 

Ankit Agarwal and Ravikumar Guntu

Compound Dry and Hot Extremes (CDHE) have an adverse impact on socioeconomic factors during the Indian summer monsoon, and a future exacerbation is anticipated. The occurrence of CDHE is influenced by teleconnections, which play a crucial role in determining its likelihood on a seasonal scale. Despite the importance, there is a lack of studies unravelling the teleconnections of CDHE in India. Previous investigations specifically focused on teleconnections between precipitation, temperature, and climate indices. Hence, there is a need to unravel the teleconnections of CDHE. This study presents a framework combining event coincidence analysis (ECA) with complexity science. ECA evaluates the synchronization between CDHE and climate indices. Subsequently, complexity science is utilized to construct a driver-CDHE network to identify the critical drivers of CDHE. A logistic regression model is employed to evaluate the proposed drivers' effectiveness. The occurrence of CDHE exhibits distinct patterns from July to September when considering intra-seasonal variability. Our findings contribute to the identification of drivers associated with CDHE. The primary driver for Eastern, Western India and Central India is the indices in the Pacific Ocean and Atlantic Ocean, respectively, followed by the indices in the Indian Ocean. These identified drivers outperform the traditional Niño 3.4-based predictions. Overall, our results demonstrate the effectiveness of integrating ECA and complexity science to enhance the prediction of CDHE occurrences.

How to cite: Agarwal, A. and Guntu, R.: Towards the Predictability of Compound Dry and Hot Extremes through Complexity Science, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-7134, https://doi.org/10.5194/egusphere-egu24-7134, 2024.

EGU24-8028 | ECS | Orals | CL4.3

Constraining near to mid-term climate projections by combining observations with decadal predictions 

Rémy Bonnet, Julien Boé, and Emilia Sanchez

The implementation of adaptation policies requires seamless and relevant information on the evolution of the climate over the next decades. Decadal climate predictions are subject to drift because of intrinsic model errors and their skill may be limited after a few years or even months depending on the region. Non-initialized ensembles of climate projections have large uncertainties over the next decades, encompassing the full range of uncertainty attributed to internal climate variability. Providing the best climate information over the next decades is therefore challenging. Recent studies have started to address this challenge by constraining uninitialized projections of sea surface temperature using decadal predictions or using a storyline approach to constrain uninitialized projections of the Atlantic Meridional Overturning Circulation using observations. Here, using a hierarchical clustering method, we select a sub-ensemble of non-initialized climate simulations based on their similarity to observations. Then, we try to further refine this sub-ensemble of trajectories by selecting a subset based on its consistency with decadal predictions. This study presents a comparison of these different methods for constraining surface temperatures in the North-Atlantic / Europe region over the next decades, focusing on CMIP6 non-initialized simulations.

How to cite: Bonnet, R., Boé, J., and Sanchez, E.: Constraining near to mid-term climate projections by combining observations with decadal predictions, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-8028, https://doi.org/10.5194/egusphere-egu24-8028, 2024.

EGU24-9049 | Posters on site | CL4.3

Constraining internal variability in CMIP6 simulations to provide skillful near-term climate predictions 

Rashed Mahmood, Markus G. Donat, Pablo Ortega, and Francisco Doblas-Reyes

Adaptation to climate change requires accurate and reliable climate information on decadal and multi-decadal timescales. Such near-term climate information is obtained from future projection simulations, which are strongly affected by uncertainties related to, among other things, internal climate variability. Here we present an approach to constrain variability in future projection simulations of the coupled model intercomparison project phase 6 (CMIP6). The constraining approach involves phasing in the simulated with the observed climate state by evaluating the area-weighted spatial pattern correlations of sea surface temperature (SST) anomalies in individual members and observations. The constrained ensemble, based on the top ranked members in terms of pattern correlations with observed SST anomalies, shows significant added value over the unconstrained ensemble in predicting surface temperature 10 and also 20 years  after the synchronization with observations, thus extending the forecast range of the standard initialised predictions. We also find that while the prediction skill of the constrained ensemble for the first ten years is similar to the initialized decadal predictions, the added value against the unconstrained ensemble extends over more regions than the decadal predictions. In addition, the constraining approach can also be used to attribute predictability of regional and global climate variations to regional SST variability.

How to cite: Mahmood, R., G. Donat, M., Ortega, P., and Doblas-Reyes, F.: Constraining internal variability in CMIP6 simulations to provide skillful near-term climate predictions, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-9049, https://doi.org/10.5194/egusphere-egu24-9049, 2024.

There is an ongoing discussion about the contributions from forced and natural sources to the Atlantic Multi-decadal Variability (AMV).  As the AMV influences the general climate in large regions, this question has important consequences for climate predictions on decadal timescales and for a robust estimation of the influence of climate forcings.

Here, we investigate the Atlantic Multi-decadal Variability (AMV) in observations and in a large CMIP6 historical climate model ensemble. We compare three different definitions of the AMV aimed at extracting the variability intrinsic to the Atlantic region. These definitions are based on removing from the Atlantic temperature the non-linear trend, the part congruent to the global average, or the part congruent to the multi-model ensemble mean of the global average. The considered AMV definitions agree on the well-known low-frequency oscillatory variability in observations, but show larger differences for the models. In general, large differences between ensemble members are found.

We estimate the forced response in the AMV as the mean of the large multi-model ensemble.  The forced response resembles the observed low-frequency oscillatory variability for the detrended AMV definition, but this definition is also the most inefficient in removing the forced global mean signal. The forced response is very weak for the other definitions and only few of their individual ensemble members show oscillatory variability and, if they do, not with the observed phase.

The observed spatial temperature pattern related to the AMV is well captured for all three AMV definitions, but with some differences in the spatial extent. The observed instantaneous connection between NAO and AMV is well represented in the models for all AMV definitions. Only non-significant evidence of NAO leading the AMV on decadal timescales is found.

How to cite: Christiansen, B., Yang, S., and Drews, A.: The Atlantic Multi-decadal Variability in observations and in a large historical multi-model ensemble: Forced and internal variability, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-9100, https://doi.org/10.5194/egusphere-egu24-9100, 2024.

EGU24-9274 | ECS | Orals | CL4.3 | Highlight

The Role of the North Atlantic for Heat Wave Characteristics in Europe 

Sabine Bischof, Robin Pilch Kedzierski, Martje Hänsch, Sebastian Wahl, and Katja Matthes

The recent severe European summer heat waves of 2015 and 2018 co-occurred with cold subpolar North Atlantic (NA) sea surface temperatures (SSTs). However, a significant connection between this oceanic state and European heat waves was not yet established.

We investigate the effect of cold subpolar NA SSTs on European summer heat waves using two 100-year long AMIP-like model experiments: one that employs the observed global 2018 SST pattern as a boundary forcing and a counter experiment for which we removed the negative NA SST anomaly from the 2018 SST field, while preserving daily and small-scale SST variabilities. Comparing these experiments, we find that cold subpolar NA SSTs significantly increase heat wave duration and magnitude downstream over the European continent. Surface temperature and circulation anomalies are connected by the upper-tropospheric summer wave pattern of meridional winds over the North Atlantic European sector, which is enhanced with cold NA SSTs. Our results highlight the relevance of the subpolar NA region for European summer conditions, a region that is marked by large biases in current coupled climate model simulations.

How to cite: Bischof, S., Pilch Kedzierski, R., Hänsch, M., Wahl, S., and Matthes, K.: The Role of the North Atlantic for Heat Wave Characteristics in Europe, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-9274, https://doi.org/10.5194/egusphere-egu24-9274, 2024.

EGU24-9690 | ECS | Orals | CL4.3

Hybrid statistical-dynamical seasonal prediction of summer extreme temperatures over Europe 

Luca Famooss Paolini, Paolo Ruggieri, Salvatore Pascale, Erika Brattich, and Silvana Di Sabatino

Several studies show that the occurrence of summer extreme temperatures over Europe is increased since the middle of the twentieth century and is expected to further increase in the future due to global warming (Seneviratne et al., 2021). Thus, predicting heat extremes several months ahead is crucial given their impacts on socio-economic and environmental systems.

In this context, state-of-the-art dynamical seasonal prediction systems (SPSs) show low skills in predicting European heat extremes on seasonal timescale, especially in central and northern Europe (Prodhomme et al., 2022). However, recent studies have shown that our skills in predicting extratropical climate can be largely improved by subsampling the dynamical SPS ensemble with statistical post-processing techniques (Dobrynin et al., 2022).

This study assesses if the seasonal prediction skill of summer extreme temperatures in Europe in the state-of-the-art dynamical SPSs can be improved through subsampling. Specifically, we use a multi-model ensemble (MME) of SPSs contributing to the Copernicus Climate Change Service (C3S), analysing di hindcast period 1993—2016. The MME is subsampled by retaining a subset of members that predict the phase of the North Atlantic Oscillation (NAO) and the Eastern Atlantic (EA), typically linked to summer extreme temperatures in Europe. The subsampling relies on spring predictors of the weather regimes and thus allows us to retain only those ensemble members with a reasonable representation of summer heat extreme teleconnections.

Results show that by retaining only those ensemble members that accurately represent the NAO phase, it not only enhances the seasonal prediction skills for the summer European climate but also leads to improved predictions of summer extreme temperatures, especially in central and northern Europe. Differently, selecting only those ensemble members that accurately represent the EA phase does not improve either the predictions of summer European climate or the predictions of summer extreme temperatures. This can be explained by the fact that the C3S SPSs exhibits deficiencies in accurately representing the summer low-frequency atmospheric variability.

Bibliography

Dobrynin, M., and Coauthors, 2018: Improved Teleconnection-Based Dynamical Seasonal Predictions of Boreal Winter. Geophysical Research Letters, 45 (8), 3605—3614, https://doi.org/10.1002/2018GL07720

Prodhomme, C., S. Materia, C. Ardilouze, R. H. White, L. Batté, V. Guemas, G. Fragkoulidis, and J. Garcìa-Serrano, 2022: Seasonal prediction of European summer heatwaves. Climate Dynamics, 58 (7), 2149—2166, https://doi.org/10.1007/s00382-021-05828-3

Seneviratne, S., and Coauthors, 2021: Weather and Climate Extreme Events in a Changing Climate, chap. 11, 1513—1766. Cambridge University Press, https://doi.org/10.1017/9781009157896.013

How to cite: Famooss Paolini, L., Ruggieri, P., Pascale, S., Brattich, E., and Di Sabatino, S.: Hybrid statistical-dynamical seasonal prediction of summer extreme temperatures over Europe, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-9690, https://doi.org/10.5194/egusphere-egu24-9690, 2024.

EGU24-9905 | ECS | Orals | CL4.3

Optimization-based driver detection and prediction of seasonal heat extremes 

Ronan McAdam, César Peláez Rodríguez, Felicitas Hansen, Jorge Pérez Aracil, Antonello Squintu, Leone Cavicchia, Eduardo Zorita, Sancho Saldez-Sanz, and Enrico Scoccimarro

As a consequence of limited reliability of dynamical forecast systems, particularly over Europe, efforts in recent years have turned to exploiting the power of Machine Learning methods to extract information on drivers of extreme temperature from observations and reanalysis. Meanwhile, the diverse impacts of extreme heat have driven development of new indicators which take into account nightime temperatures and humidity. In the H2020 CLimate INTelligence (CLINT) project, a feature selection framework is being developed to find the combination of drivers which provides optimal seasonal forecast skill of European summer heatwave indicators. Here, we present the methodology, its application to a range of heatwave indicators and forecast skill compared to existing dynamical systems. First, a range of (reduced-dimensionality) drivers are defined, including k-means clusters of variables known to impact European summer (e.g. precipitation, sea ice content), and more complex indices like the NAO and weather regimes. Then, these drivers are used to train machine learning based prediction models, of varying complexity, to predict seasonal indicators of heatwave occurrence and intensity. A crucial and novel step in our framework is the use of the Coral Reef Optimisation algorithm, used to select the variables and their corresponding lag times and time periods which provide optimal forecast skill. To maximise training data, both ERA5 reanalysis and a 2000-year paleo-simulation are used; the representation of heatwaves and atmospheric conditions are validated with respect to ERA5. We present comparisons of forecast skill to the dynamical Copernicus Climate Change Service seasonal forecasts systems. The differences in timing, predictability and drivers of daytime and nighttime heatwaves across Europe are highlighted. Lastly, we discuss how the framework can easily be adapted to other extremes and timescales.



How to cite: McAdam, R., Peláez Rodríguez, C., Hansen, F., Pérez Aracil, J., Squintu, A., Cavicchia, L., Zorita, E., Saldez-Sanz, S., and Scoccimarro, E.: Optimization-based driver detection and prediction of seasonal heat extremes, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-9905, https://doi.org/10.5194/egusphere-egu24-9905, 2024.

EGU24-10539 | ECS | Orals | CL4.3

Exploring multiyear-to-decadal North Atlantic sea level predictability using machine learning and analog methods 

Qinxue Gu, Liwei Jia, Liping Zhang, Thomas Delworth, Xiaosong Yang, Fanrong Zeng, and Shouwei Li

Long-term sea level rise and multiyear-to-decadal sea level variations pose substantial risks of flooding and erosion in coastal communities. The North Atlantic Ocean and the U.S. East Coast are hotspots for sea level changes under current and future climates. Here, we employ a machine learning technique, a self-organizing map (SOM)-based framework, to systematically characterize the North Atlantic sea level variability, assess sea level predictability, and generate sea level predictions on multiyear-to-decadal timescales. Specifically, we classify 5000-year North Atlantic sea level anomalies from the Seamless System for Prediction and EArth System Research (SPEAR) model control simulations into generalized patterns using SOM. Preferred transitions among these patterns are further identified, revealing long-term predictability on multiyear-to-decadal timescales related to shifts in Atlantic meridional overturning circulation (AMOC) phases. By combining the SOM framework with “analog” techniques based on the simulations and observational/reanalysis data, we demonstrate prediction skill of large-scale sea level patterns comparable to that from initialized hindcasts. Moreover, additional source of short-term predictability is identified after the exclusion of low-frequency AMOC signals, which arises from the wind-driven North Atlantic tripole mode triggered by the North Atlantic Oscillation. This study highlights the potential of machine learning methods to assess sources of predictability and to enable efficient, long-term climate prediction.

How to cite: Gu, Q., Jia, L., Zhang, L., Delworth, T., Yang, X., Zeng, F., and Li, S.: Exploring multiyear-to-decadal North Atlantic sea level predictability using machine learning and analog methods, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-10539, https://doi.org/10.5194/egusphere-egu24-10539, 2024.

The inter-annual to multi-decadal variability of recurrent, synoptic-scale atmospheric circulation patterns in the Northern Hemisphere extratropics, as represented by the Jenkinson-Collison classification scheme, is explored in reanalysis data spanning the entire 20th century, and in global climate model (GCM) data from the historical, AMIP and DCPP experiments conducted within the framework of CMIP6. The aim of these efforts is to assess the effect of coupled vs. uncoupled and initialised vs. non-initialized GCM simulations in reproducing the observed low-frequency variability of the aforementioned circulation patterns.

Results reveal that the observed annual counts of typical recurrent weather patterns, such as cyclonic or anticyclonic conditions and also situations of pronounced advection, exhibit significant oscillations on multiple time-scales ranging between several years and several decades. The period of these oscillations, however, is subject to large regional variations. This is in line with earlier studies suggesting that the extratropical atmospheric circulation’s low frequency variability is essentially unforced, except in the Pacific-North American sector where the forced variability is enhanced due to ENSO teleconnections. Neither the periods obtained from historical nor those obtained from AMIP experiments align with observations. Likewise, not even the periods obtained from different runs of the same GCM and experiment correspond to each other. Thus, in an non-initialized model setup, ocean-atmosphere coupling or the lack thereof essentially leads to the same results. Whether initialization and/or augmenting the ensemble size can improve these findings, will also be discussed.

Acknowledgement: This work is part of project Impetus4Change, which has received funding from the European Union’s Horizon Europe research and innovation programme under grant agreement No 101081555.

How to cite: Brands, S., Cimadevilla, E., and Fernández, J.: Low-frequency variability of synoptic-scale atmospheric circulation patterns in the Northern Hemisphere extratropics and associated hindcast skill of decadal forecasting systems, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-10551, https://doi.org/10.5194/egusphere-egu24-10551, 2024.

EGU24-10574 | Orals | CL4.3 | Highlight

Will 2024 be the first year above 1.5 C? 

Nick Dunstone, Doug Smith, Adam Scaife, Leon Hermanson, Andrew Colman, and Chris Folland

Global mean surface temperature is the key metric by which our warming climate is monitored and for which international climate policy is set. At the end of each year the Met Office makes a global mean temperature forecast for the coming year. Following on from the new record 2023, we predict a high probability of another record year in 2024 and a 35% chance of exceeding 1.5 C above pre-industrial. Whilst a one-year temporary exceedance of 1.5 C would not constitute a breech of the Paris Agreement target, our forecast highlights how close we are now to breeching this target. We show that our 2024 forecast can be largely explained by the combination of the continuing warming trend of +0.2 C/decade and the lagged warming affect of a strong tropical Pacific El Nino event. We further highlight 2023 was significantly warmer than forecast and that much of this warming signal came from the southern hemisphere and requires further understanding.

How to cite: Dunstone, N., Smith, D., Scaife, A., Hermanson, L., Colman, A., and Folland, C.: Will 2024 be the first year above 1.5 C?, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-10574, https://doi.org/10.5194/egusphere-egu24-10574, 2024.

EGU24-11485 | ECS | Orals | CL4.3

Summer drought predictability in the Mediterranean region in seasonal forecasts 

Giada Cerato, Katinka Bellomo, and Jost von Hardenberg

The Mediterranean region has been identified as an important climate change hotspot, over the 21st century both air temperature and its extremes are projected to rise at a rate surpassing that of the global average and a significant decrease of average summer precipitation is projected, particularly for the western Mediterranean. On average, Mediterranean droughts have become more frequent and intense in recent years and are expected to become more widespread in many regions. These prolonged dry spells pose a substantial threat to agriculture and impact several socio-economic sectors. In this context, long-range weather forecasting has emerged as a promising tool for seasonal drought risk assessment. However, the interpretation of the forecasting products is not always straightforward due to their inherent probabilistic nature. Therefore, a rigorous evaluation process is needed to determine the extent to which these forecasts provide a fruitful advantage over much simpler forecasting systems, such as those based on climatology. 

In this study, we use the latest version of ECMWF’s seasonal prediction system (SEAS5) to understand its skill in predicting summer droughts. The Standardized Precipitation Evapotranspiration Index (SPEI) aggregated over different lead times is employed to mark below-normal dryness conditions in August. We use a comprehensive set of evaluation metrics to gain insight into the accuracy, systematic biases, association, discrimination and sharpness of the forecast system. Our findings reveal that up to 3 months lead time, seasonal forecasts show stronger association and discrimination skills than the climatological forecast, especially in the Southern Mediterranean, although the prediction quality in terms of accuracy and sharpness is limited. On the other hand, extending the forecast range up to 6 months lead time dramatically reduces its predictability skill, with the system mostly underperforming elementary climatological predictions. 

This approach is then extended to examine the full ensemble of seasonal forecasting systems provided by the Copernicus Climate Change Service (C3S) to test their skill in predicting droughts. Our findings can help an informed use of seasonal forecasts of droughts and the development of related climate services.

How to cite: Cerato, G., Bellomo, K., and von Hardenberg, J.: Summer drought predictability in the Mediterranean region in seasonal forecasts, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-11485, https://doi.org/10.5194/egusphere-egu24-11485, 2024.

EGU24-11930 | ECS | Posters on site | CL4.3

A global empirical system for probabilistic seasonal climate prediction based on generative AI and CMIP6 models  

Lluís Palma, Alejandro Peraza, Amanda Duarte, David Civantos, Stefano Materia, Arijit Nandi, Jesús Peña-Izquierdo, Mihnea Tufis, Gonzalo Vilella, Laia Romero, Albert Soret, and Markus Donat

Reliable probabilistic information at the seasonal time scale is essential across various societal sectors, such as agriculture, energy, or water management. Current applications of seasonal predictions rely on General Circulation Models (GCMs) that represent dynamical processes in the atmosphere, land surface, and ocean while capturing their linear and nonlinear interactions. However, GCMs come with an inherent high computational cost. In an operational setup, they are typically run once a month and at a lower temporal and spatial resolution than the ones needed for regional applications. Moreover, GCMs suffer from significant drifts and biases and can miss relevant teleconnections, resulting in low skill for particular regions or seasons. 

In this context, the use of generative AI methods that can model complex nonlinear relationships can be a viable alternative for producing probabilistic predictions with low computational demand. Such models have already demonstrated their effectiveness in different domains, i.e. computer vision, natural language processing, and weather prediction. However, although requiring less computational power, these techniques still rely on big datasets in order to be efficiently trained. Under this scenario, and with sufficiently high-quality global observational datasets spanning at most 70 years, the research trend has evolved into training these models using climate model output. 

In this work, we build upon the work presented by Pan et al., 2022, which introduced a conditional Variational Autoencoder (cVAE) to predict global temperature and precipitation fields for the October to March season starting from July initial conditions. We adopt several pre-processing changes to account for different biases and trends across the CMIP6 models. Additionally, we explore different architecture modifications to improve the model's performance and stability. We study the benefits of our model in predicting three-month anomalies on top of the climate change trend. Finally, we compare our results with a state-of-the-art GCM (SEAS5) and a simple empirical system based on the linear regression of classical seasonal indices based on Eden et al., 2015.

 

Pan, Baoxiang, Gemma J. Anderson, André Goncalves, Donald D. Lucas, Céline J.W. Bonfils, and Jiwoo Lee. 'Improving Seasonal Forecast Using Probabilistic Deep Learning'. Journal of Advances in Modeling Earth Systems 14, no. 3 (1 March 2022). https://doi.org/10.1029/2021MS002766.


Eden, J. M., G. J. van Oldenborgh, E. Hawkins, and E. B. Suckling. 'A Global Empirical System for Probabilistic Seasonal Climate Prediction'. Geoscientific Model Development 8, no. 12 (11 December 2015): 3947–73. https://doi.org/10.5194/gmd-8-3947-2015.

How to cite: Palma, L., Peraza, A., Duarte, A., Civantos, D., Materia, S., Nandi, A., Peña-Izquierdo, J., Tufis, M., Vilella, G., Romero, L., Soret, A., and Donat, M.: A global empirical system for probabilistic seasonal climate prediction based on generative AI and CMIP6 models , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-11930, https://doi.org/10.5194/egusphere-egu24-11930, 2024.

EGU24-12969 | ECS | Orals | CL4.3

How unusual is the recent decade-long pause in Arctic summer sea ice retreat? 

Patricia DeRepentigny, François Massonnet, Roberto Bilbao, and Stefano Materia

The Earth has warmed significantly over the past 40 years, and the fastest rate of warming has occurred in and around the Arctic. The warming of northern high latitudes at a rate of almost four times the global average (Rantanen et al., 2022), known as Arctic amplification, is associated with sea ice loss, glacier retreat, permafrost degradation, and expansion of the melting season. Since the mid-2000s, summer sea ice has exhibited a rapid decline, reaching record minima in September sea ice area in 2007 and 2012. However, after the early 2010s, the downward trend of minimum sea ice area appears to decelerate (Swart et al., 2015; Baxter et al., 2019). This apparent slowdown and the preceding acceleration in the rate of sea ice loss are puzzling in light of the steadily increasing rate of greenhouse gas emissions of about 4.5 ppm yr−1 over the past decade (Friedlingstein et al., 2023) that provides a constant climate forcing. Recent studies suggest that low-frequency internal climate variability may have been as important as anthropogenic influences on observed Arctic sea ice decline over the past four decades (Dörr et al., 2023; Karami et al., 2023). Here, we investigate how unusual this decade-long pause in Arctic summer sea ice decline is within the context of internal climate variability. To do so, we first assess how rare this is deceleration of Arctic sea ice loss is by comparing it to trends in CMIP6 historical simulations. We also use simulations from the Decadal Climate Prediction Project (DCPP) contribution to CMIP6 to determine if initializing decadal prediction systems from estimates of the observed climate state substantially improves their performance in predicting the slowdown in Arctic sea ice loss over the past decade. As the DCPP does not specify the data or the methods to be used to initialize forecasts or how to generate ensembles of initial conditions, we also assess how different formulations affect the skill of the forecasts by analyzing differences between models. This work provides an opportunity to attribute this pause in Arctic sea ice retreat to interannual internal variability or radiative external forcings, something that observation analysis alone cannot achieve.

How to cite: DeRepentigny, P., Massonnet, F., Bilbao, R., and Materia, S.: How unusual is the recent decade-long pause in Arctic summer sea ice retreat?, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-12969, https://doi.org/10.5194/egusphere-egu24-12969, 2024.

EGU24-14341 | Posters on site | CL4.3

Compound Heat and Dry Events Influenced by the Pacific–Japan Pattern over Taiwan in Summer 

Szu-Ying Lin, Wan-Ling Tseng, Yi-Chi Wang, and MinHui Lo

Compound dry and hot events, characterized by elevated temperatures and reduced precipitation, pose interconnected challenges to human social economics, necessitating comprehensive strategies for mitigation and adaptation. This study focuses on the Pacific-Japan (PJ) pattern, a significant climate variability influencing summer climates in East Asia. While previous research has explored its impact on Japan and Korea, our investigation delves into its effects on Taiwan, a mountainous subtropical island with a population of approximately 24 million. Utilizing long-term temperature and rainfall data, along with reanalysis dynamic downscaling datasets, we examine the interannual impacts of the PJ pattern on summer temperature and compound heat and dry events. Our findings reveal a significant temperature increase during the positive phase of the PJ pattern, characterized by anticyclonic anomalous circulation over Taiwan. Additionally, both the Standardized Precipitation Index and soil water exhibit a decline during this phase, reflecting meteorological and hydrological drought conditions. A robust negative correlation (-0.7) between drought indices and temperature emphasizes the compound effect of heat and dry events during the PJ positive phase. This study enhances the understanding of the PJ pattern as a climate driver, describing its role in hot and dry summers over Taiwan. The insights gained, when integrated into seasonal prediction and early warning systems, can aid vulnerable sectors in preparing for potential heat and dry stress hazards.

How to cite: Lin, S.-Y., Tseng, W.-L., Wang, Y.-C., and Lo, M.: Compound Heat and Dry Events Influenced by the Pacific–Japan Pattern over Taiwan in Summer, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-14341, https://doi.org/10.5194/egusphere-egu24-14341, 2024.

EGU24-14379 | Posters on site | CL4.3

Linkage between Temperature and Heatwaves in Summer Taiwan to the Pacific Meridional Mode 

Chieh-Ting Tsai, Wan-Ling Tseng, and Yi-Chi Wang

Over the past century, Taiwan has gradually recognized the hazards posed by extreme heat events (EHT), prompting the development of mid-term adaptation strategies to address challenges in the coming decades. However, our understanding of decadal-scale temperature variations remains insufficient, requiring further research into influencing factors. Our study reveals the crucial role of the Pacific Meridional Mode (PMM) in modulating decadal-scale variations in summer temperatures in Taiwan. During the positive phase of PMM, warm sea surface temperature anomalies trigger an eastward-moving wave train extending into East Asia. This leads to the development of high-pressure circulations near Southeast Asia and Taiwan, enhancing the temperature increase. This mechanism has been reproduced in experiments using the Taiwan Earth System Model. Moreover, our study utilizes the calendar day 90th percentile of maximum temperature (CTX) as the threshold for extreme high-temperature events (EHT), while also employing the heatwaves magnitude scale (HWMS) as the criterion for defining heatwaves. During the positive phase of PMM, the frequency and duration of EHT increase, with variations observed across different regions. The overall intensity of heatwave events also strengthens, primarily due to extended durations. Notably, in a single city, this results in exposure of up to 800,000 person-days to EHT, presenting a tenfold increase compared to the annual effect observed in the long-term warming trend. These findings on the decadal-scale relationship between summer temperatures in Taiwan and PMM contribute to a deeper understanding of EHT and heatwaves events impacts, providing more nuanced insights for future regional strategies in mitigating heatwave disasters.

How to cite: Tsai, C.-T., Tseng, W.-L., and Wang, Y.-C.: Linkage between Temperature and Heatwaves in Summer Taiwan to the Pacific Meridional Mode, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-14379, https://doi.org/10.5194/egusphere-egu24-14379, 2024.

EGU24-14688 | ECS | Orals | CL4.3

Exploring ML-based decadal predictions of the German Bight storm surge climate 

Daniel Krieger, Sebastian Brune, Johanna Baehr, and Ralf Weisse

Storm surges and elevated water levels regularly challenge coastal protection and inland water management along the low-lying coastline of the German Bight. Skillful seasonal-to-decadal (S2D) predictions of the local storm surge climate would be beneficial to stakeholders and decision makers in the region. While storm activity has recently been shown to be skillfully predictable on a decadal timescale with a global earth system model, surge modelling usually requires very fine spatial and temporal resolutions that are not yet present in current earth system models. We therefore propose an alternative approach to generating S2D predictions of the storm surge climate by training a neural network on observed water levels and large-scale atmospheric patterns, and apply the neural network to the available model output of a S2D prediction system. We show that the neural-network-based translation from large-scale atmospheric fields to local water levels at the coast works sufficiently well, and that several windows of predictability for the German Bight surge climate emerge on the S2D scale.

How to cite: Krieger, D., Brune, S., Baehr, J., and Weisse, R.: Exploring ML-based decadal predictions of the German Bight storm surge climate, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-14688, https://doi.org/10.5194/egusphere-egu24-14688, 2024.

Atlantic meridional overturning circulation (AMOC) is one of the mechanisms for climate predictability and one of the properties that decadal climate predictions are attempting to predict. The starting point for AMOC decadal predictions is sensitive to the underlying data assimilation and/or initialization procedure. This means that different choices during the data assimilation procedure (e.g., assimilation method, assimilation window, data sources, resolution, nudging terms and strength, full field vs anomaly initialization/assimilation, etc) can result in a different mean and even variability of reconstructed ocean circulation. How coherent the AMOC initial states should be among the CMIP-like decadal prediction experiments? How good in general should the initial AMOC be for decadal predictions? And do initialization issues of the ocean circulation influence the prediction skill of other variables that are of interest for application studies? These are the questions that we were attempting to address in our study, where we analyzed twelve decadal prediction systems from the World Meteorological Organization Lead Centre for Annual-to-Decadal Climate Prediction project. We identify that the AMOC initialization influences the quality of predictions of the subpolar gyre (SPG). When predictions show a large initial error in their AMOC, they usually have low skill for predicting the internal variability of the SPG five years after the initialization.

How to cite: Polkova, I. and the Co-Authors: Initialization shock in the ocean circulation reduces skill in decadal predictions of the North Atlantic subpolar gyre, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-15358, https://doi.org/10.5194/egusphere-egu24-15358, 2024.

EGU24-15476 | Posters on site | CL4.3

Statistics of sudden stratospheric warmings using a large model ensemble 

Sarah Ineson, Nick Dunstone, Adam Scaife, Martin Andrews, Julia Lockwood, and Bo Pang

Using a large ensemble of initialised retrospective forecasts (hindcasts) from a seasonal prediction system, we explore various statistics relating to sudden stratospheric warmings (SSWs). Observations show that SSWs occur at a similar frequency during both El Niño and La Niña northern hemisphere winters. This is contrary to expectation, as the stronger stratospheric polar vortex associated with La Niña years might be expected to result in fewer of these extreme breakdowns. We show that this similar frequency may have occurred by chance due to the limited sample of years in the observational record. We also show that in these hindcasts, winters with two SSWs, a rare event in the observational record, on average have an increased surface impact. Multiple SSW events occur at a lower rate than expected if events were independent but somewhat surprisingly, our analysis also indicates a risk, albeit small, of winters with three or more SSWs, as yet an unseen event.

How to cite: Ineson, S., Dunstone, N., Scaife, A., Andrews, M., Lockwood, J., and Pang, B.: Statistics of sudden stratospheric warmings using a large model ensemble, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-15476, https://doi.org/10.5194/egusphere-egu24-15476, 2024.

EGU24-15709 | ECS | Orals | CL4.3

Predicting Atlantic and Benguela Niño events with deep learning  

Marie-Lou Bachelery, Julien Brajard, Massimiliano Patacchiola, and Noel Keenlyside

Extreme Atlantic and Benguela Niño events continue to significantly impact the tropical Atlantic region, with far-reaching consequences for African climate and ecosystems. Despite attempts to forecast these events using traditional seasonal forecasting systems, success remains low, reinforcing the growing idea that these events are unpredictable. To overcome the limitations of dynamical prediction systems, we introduce a deep learning-based statistical prediction model for Atlantic and Benguela Niño events. Our convolutional neural network (CNN) model, trained on 90 years of reanalysis data incorporating surface and 100m-averaged temperature variables, demonstrates the capability to forecast the Atlantic and Benguela Niño indices with lead times of up to 3-4 months. Notably, the CNN model excels in forecasting peak-season events with remarkable accuracy extending up to 5 months ahead. Gradient sensitivity analysis reveals the ability of the CNN model to exploit known physical precursors, particularly the connection to equatorial dynamics and the South Atlantic Anticyclone, for accurate predictions of Benguela Niño events. This study challenges the perception of the Tropical Atlantic as inherently unpredictable, underscoring the potential of deep learning to enhance our understanding and forecasting of critical climate events. 

How to cite: Bachelery, M.-L., Brajard, J., Patacchiola, M., and Keenlyside, N.: Predicting Atlantic and Benguela Niño events with deep learning , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-15709, https://doi.org/10.5194/egusphere-egu24-15709, 2024.

EGU24-15974 | ECS | Posters virtual | CL4.3

Recalibrating DWD’s operational climate predictions: towards a user-oriented seamless climate service 

Alexander Pasternack, Birgit Mannig, Andreas Paxian, Amelie Hoff, Klaus Pankatz, Philip Lorenz, and Barbara Früh

The German Meteorological Service's (Deutscher Wetterdienst DWD) climate predictions website  (www.dwd.de/climatepredictions) offers a centralized platform for accessing post-processed climate predictions, including subseasonal forecasts from ECMWF's IFS and seasonal and decadal predictions from the German climate prediction system. The website design was developed in collaboration with various sectors to ensure uniformity across all time frames, and users can view maps, tables, and time series of ensemble mean and probabilistic predictions in combination with their skill. The available data covers weekly, 3-month, 1-year, and 5-year temperature means, precipitation sums and soil moisture for the world, Europe, Germany, and particular German regions. To achieve high spatial resolution, the DWD used the statistical downscaling method EPISODES. Moreover, within the BMBF project KIMoDIs (AI-based monitoring, data management and information system for coupled forecasting and early warning of low groundwater levels and salinisation) the DWD provides climate prediction data of further hydrological variables (e.g. relative humidity) with corresponding prediction skill on a regional scale.

However, all predictions on these time scales can suffer from inherent systematic errors, which can impact their usefulness. To address these issues, the recalibration method DeFoReSt was applied to decadal predictions, using a combination of 3rd order polynomials in lead and start time, along with a boosting model selection approach. This approach addresses lead-time dependent systematic errors, such as drift, as well as inaccuracies in representing long-term changes and variability.

This study highlights the improved accuracy of the recalibration approach on decadal predictions due to an increased polynomial order compared to the original approach, and its different impact on global and regional scales. It also explores the feasibility of transferring this approach to predictions with shorter time horizons of the provided variables.

How to cite: Pasternack, A., Mannig, B., Paxian, A., Hoff, A., Pankatz, K., Lorenz, P., and Früh, B.: Recalibrating DWD’s operational climate predictions: towards a user-oriented seamless climate service, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-15974, https://doi.org/10.5194/egusphere-egu24-15974, 2024.

EGU24-16366 | ECS | Orals | CL4.3

Decadal predictions outperform projections in forecasting winter precipitation over the Mediterranean region 

Dario Nicolì, Silvio Gualdi, and Panos Athanasiadis

The Mediterranean region is highly sensitive to climate change, having experienced an intense warming and drying trend in recent decades, primarily due to the increased concentrations of anthropogenic greenhouse gases. In the context of decision-making processes, there is a growing interest in understanding the near-term climate evolution of this region.

In this study, we explore the climatic fluctuations of the Mediterranean region in the near-term range (up to 10 years ahead) using two different products: projections and decadal predictions. The former are century-scale climate change simulations initialized from arbitrary model states to which were applied anthropogenic and natural forcings. A major limitation of climate projections is their limited information regarding the current state of the Earth’s climate system. Decadal climate predictions, obtained by constraining the initial conditions of an ensemble of model simulations through a best estimate of the observed climate state, provide a better understanding of the next-decade climate and thus represent an invaluable tool in assisting climate adaptation.

Using retrospective forecasts from eight decadal prediction systems contributing to the CMIP6 Decadal Climate Prediction Project (CMIP6 DCPP) and the corresponding ensemble of non-initialized projections, we compare the capabilities of the state-of-the-art climate models in predicting future climate changes of the Mediterranean region for some key quantities so as to assess the added value of initialization. 

Beyond the contribution of external forcings, the role of internal variability is also investigated since part of the detected predictability arises from internal climate variability patterns affecting the Mediterranean. The observed North Atlantic Oscillation, the dominant climate variability pattern in the Euro-Atlantic domain, as well as its  impact on wintertime precipitation over Europe are well reproduced by decadal predictions, especially over the Mediterranean, outperforming projections. We also apply a sub-sampling method to enhance the respective signal-to-noise ratio and consequently improve precipitation skill over the Mediterranean.

How to cite: Nicolì, D., Gualdi, S., and Athanasiadis, P.: Decadal predictions outperform projections in forecasting winter precipitation over the Mediterranean region, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-16366, https://doi.org/10.5194/egusphere-egu24-16366, 2024.

EGU24-16985 | Posters on site | CL4.3

Investigating signals in summer seasonal forecasts over the North Atlantic/European region 

Julia Lockwood, Nick Dunstone, Kristina Fröhlich, Ramón Fuentes Franco, Anna Maidens, Adam Scaife, Doug Smith, and Hazel Thornton

The current generation of seasonal forecast models struggle to skilfully predict dynamical circulation over the North Atlantic and European region in boreal summer.  Using two different state-of-the-art seasonal prediction systems, we show that tropical rainfall anomalies drive a circulation signal in the North Atlantic/Europe via the propagation of Rossby waves.  The wave, however, is shifted eastwards compared to observations, so the signal does not contribute positively to model skill.  Reasons for the eastward shift of the Rossby wave are investigated, as well as other drivers of the signal in this region.  Despite the errors in the waves, the fact that seasonal forecast models do predict dynamical signals over the North Atlantic/Europe signifies seasonal predictability over this region beyond the climate change trend, and understaning the cause of the errors could lead to skilful predictions.

How to cite: Lockwood, J., Dunstone, N., Fröhlich, K., Fuentes Franco, R., Maidens, A., Scaife, A., Smith, D., and Thornton, H.: Investigating signals in summer seasonal forecasts over the North Atlantic/European region, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-16985, https://doi.org/10.5194/egusphere-egu24-16985, 2024.

EGU24-17418 | Posters on site | CL4.3

Strengthening seasonal forecasting in the Middle East & North Africa (MENA) through the WISER Programme. 

Stefan Lines, Nicholas Savage, Rebecca Parfitt, Andrew Colman, Alex Chamberlain-Clay, Luke Norris, Heidi Howard, and Helen Ticehurst

In this presentation, we introduce the WISER MENA projects SeaFOAM (Seasonal Forecasting Across MENA) and SeaSCAPE (Seasonal Co-Production and Application in MENA). These projects explore both the improvement to the regional-level seasonal forecast in the MENA region, as well as how to tailor the information in ways useful to a range of climate information stakeholders. SeaFOAM works alongside Maroc Meteo, Morocco's National Meteorological and Hydrological Service (NMHS) and the Long Range Forecasting node of the Northern Africa WMO Regional Climate Centre (RCC), to develop a framework for objective seasonal forecasting. This approach will blend techniques such as bias correction via local linear regression and canonical correlation analysis (CCA), with skill-assessed sub-selected models, to improve forecasting accuracy. Multiple drivers of rainfall variability, including the North Atlantic Oscillation (NAO) and Mediterranean Oscillation (MO), are investigated for their calibration potential. SeaSCAPE works with the WMO and various partners across MENA to understand the use of seasonal information in multiple sectors, exploring existing gaps and needs. Through stakeholder engagement workshops, training and bespoke support for the Arab Climate Outlook Forum (ArabCOF), SeaSCAPE operates collaboratively to tailor regional and national-level climate information to improve accessibility and usability of climate information on seasonal timescales.

How to cite: Lines, S., Savage, N., Parfitt, R., Colman, A., Chamberlain-Clay, A., Norris, L., Howard, H., and Ticehurst, H.: Strengthening seasonal forecasting in the Middle East & North Africa (MENA) through the WISER Programme., EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-17418, https://doi.org/10.5194/egusphere-egu24-17418, 2024.

EGU24-17585 | Orals | CL4.3

Skill of wind resource forecasts on the decadal time scale 

Kai Lochbihler, Ana Lopez, and Gil Lizcano

Accurate forecasts of the natural resources of renewable energy production have become not only a valuable but a crucial tool for managing the associated risks of specific events, such as wind droughts. Wind energy, alongside with solar power, now provide a substantial part to the renewable energy share of the global energy production and growth in this sector will most likely further increase. The naturally given fluctuations of wind resources, however, pose a challenge for maintaining a stable energy supply, which, at the end of the chain, can have an impact on the energy market prices.
Operational short-term forecasting products for the wind energy sector (multiple days) are already commonly available and seasonal to sub seasonal forecasting solutions (multiple months) can provide valuable skill and are gaining in popularity. On the other side of the spectrum, typically on a time scale of multiple decades, we find risk assessment based on climate change projections. In between the long and short term time scales, however, there is a gap that still needs to be filled to achieve seamless prediction of risks that are relevant for the energy sector: decadal predictions.

Here, we present the results of an evaluation study of a multi-model decadal prediction ensemble (DCPP) for a selection of wind development regions in Europe. The evaluation is based on multiple decades long hindcasts and carried out with a focus on the skill of predicting specific event types of wind resource availability in a probabilistic context, alongside with basic deterministic skill measures. We further investigate specific event constellations and their large-scale drivers that, in combination, can provide windows of opportunity with enhanced predictive skill. We conclude with a discussion on how this hybrid approach can be used to potentially increase not only forecast skill but also the trust of the end user.

How to cite: Lochbihler, K., Lopez, A., and Lizcano, G.: Skill of wind resource forecasts on the decadal time scale, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-17585, https://doi.org/10.5194/egusphere-egu24-17585, 2024.

EGU24-19229 | ECS | Orals | CL4.3

Comparing the seasonal predictability of Tropical Pacific variability in EC-Earth3 at two different horizontal resolutions 

Aude Carreric, Pablo Ortega, Vladimir Lapin, and Francisco Doblas-Reyes

Seasonal prediction is a field of research attracting growing interest beyond the scientific community due to its strong potential to guide decision-making in many sectors (e.g. agriculture and food security, health, energy production, water management, disaster risk reduction) in the face of the pressing dangers of climate change.

Among the various techniques being considered to improve the predictive skill of seasonal prediction systems, increasing the horizontal resolution of GCMs is a promising avenue. There are several indications that higher resolution versions of the current generation of climate models might improve key air-sea teleconnections, decreasing common biases of global models and improving the skill to predict certain regions at seasonal scales, e.g. in tropical sea surface temperature.

In this study, we analyze the differences in the predictive skill of two different seasonal prediction systems, based on the same climate model EC-Earth3 and initialized in the same way but using two different horizontal resolutions. The standard (SR) and high resolution (HR) configurations are based on an atmospheric component, IFS, of ~100 km and ~40 km of resolution respectively and on an ocean component, NEMO3.6, of ~100 km and ~25 km respectively. We focus in particular on the Tropical Pacific region where statistically significant improvements are found in HR with respect to SR for predicting ENSO and its associated climate teleconnections. We explore some processes that can explain these differences, such as the simulation of the tropical ocean mean state and atmospheric teleconnections between the Atlantic and Pacific tropical oceans. 

A weaker mean-state bias in the HR configuration, with less westward extension of ENSO-related SST anomalies, leads to better skill in ENSO regions, which can also be linked to better localization of the atmospheric teleconnection with the equatorial Atlantic Ocean. It remains to be assessed if similar improvements are consistently identified for HR versions in other forecast systems, which would prompt their routine use in seasonal climate prediction.

How to cite: Carreric, A., Ortega, P., Lapin, V., and Doblas-Reyes, F.: Comparing the seasonal predictability of Tropical Pacific variability in EC-Earth3 at two different horizontal resolutions, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-19229, https://doi.org/10.5194/egusphere-egu24-19229, 2024.

EGU24-19251 | Orals | CL4.3 | Highlight

The opportunities and challenges of near-term climate prediction 

Hazel Thornton

Accurate forecasts of the climate of the coming season and years are highly desired by many sectors of society. The skill of near-term climate prediction in winter in the North Atlantic and European region has improved over the last decade associated with larger ensembles, improving models and boosting of the prediction signal using intelligent post processing. International collaboration has improved the availability of forecasts and promoted the uptake of forecasts by different sectors. However, significant challenges remain, including summer prediction, understanding the risk of extremes within a season, multi-seasonal extremes and how best to post process the forecasts to aid decision making. This talk will summarise recent near-term climate prediction research activities at the UK Met Office and will detail our experience of providing such forecasts to the energy and water sectors.  

How to cite: Thornton, H.: The opportunities and challenges of near-term climate prediction, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-19251, https://doi.org/10.5194/egusphere-egu24-19251, 2024.

This study focuses on applying machine learning techniques to bias-correct the seasonal temperature forecasts provided by the Copernicus Climate Change Service (C3S) models. Specifically, we employ bias correction on forecasts from five major models: UK Meteorological Office (UKMO), Euro-Mediterranean Center on Climate Change (CMCC), Deutscher Wetterdienst (DWD), Environment and Climate Change Canada (ECCC), and Meteo-France. Our primary objective is to assess the performance of our bias correction model in comparison to the original forecast datasets. We utilise temperature-based indices recommended by the Expert Team on Climate Change Detection and Indices (ETCCDI) to evaluate the effectiveness of the bias-corrected seasonal forecasts. These indices served as valuable metrics to gauge the predictive capability of the models, especially in forecasting natural cascading hazards such as wildfires, droughts, and floods. The study involved an in-depth analysis of the bias-corrected forecasts, and the derived indices were crucial in understanding the models' ability to predict temperature-related extreme events. The results of this research contribute valuable information for decision-making and planning across various sectors, including disaster risk management and environmental protection. Through a comprehensive evaluation of machine learning-based bias correction techniques, we enhance the accuracy and applicability of seasonal temperature forecasts, thereby improving preparedness and resilience to climate-related challenges. 

How to cite: Mbuvha, R. and Nikraftar, Z.: Machine Learning Approaches to Improve Accuracy in Extreme Seasonal Temperature Forecasts: A Multi-Model Assessment , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-19297, https://doi.org/10.5194/egusphere-egu24-19297, 2024.

EGU24-19359 | ECS | Posters on site | CL4.3

Seasonal forecast of the late boreal winter temperature based on solar forcing and QBO 

Mikhail Vokhmianin, Antti Salminen, Kalevi Mursula, and Timo Asikainen

The ground temperature variability in the Northern Hemisphere winter is greatly influenced by the state of the polar vortex. When the vortex collapses during sudden stratospheric warmings (SSWs), rapid changes in stratospheric circulations propagate downward to the troposphere in the subsequent weeks. The ground effect following SSWs is typically manifested as the negative phase of the North Atlantic Oscillation. Our findings reveal a higher frequency of cold temperature anomalies in the Northern part of Eurasia during winters with SSWs, and conversely, warm anomalies in winters with a strong and stable vortex. This behavior is particularly evident when temperature anomalies are categorized into three equal subgroups, or terciles. Recently, we developed a statistical model that successfully predicts SSW occurrences with an 86% accuracy rate. The model utilizes the stratospheric Quasi-Biennial Oscillation (QBO) phase and two parameters associated with solar activity: the geomagnetic aa-index as a proxy for energetic particle precipitations and solar irradiance. In this study, we explore the model's potential to provide a seasonal forecast for ground temperatures. We assess the probabilities of regional temperature anomalies falling into the lowest or highest terciles based on the predicted weak or strong vortex state. Additionally, we demonstrate that the QBO phase further enhances the forecast quality. As the model provides SSW predictions as early as preceding August, our results carry significant societal relevance as well, e.g., for the energy sector, which is highly dependent on prevailing weather conditions.

How to cite: Vokhmianin, M., Salminen, A., Mursula, K., and Asikainen, T.: Seasonal forecast of the late boreal winter temperature based on solar forcing and QBO, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-19359, https://doi.org/10.5194/egusphere-egu24-19359, 2024.

EGU24-995 | ECS | Posters on site | CL4.10 | Highlight

Assessing the predictability of Euro-Mediterranean droughts through seasonal forecasts 

Thomas Dal Monte, Andrea Alessandri, Annalisa Cherchi, and Marco Gaetani

Droughts are characterized by prolonged and severe deficits in precipitation that can extend in time, over a season, a year or more. They are confined to specific climatic zones but can manifest in both high and low rainfall regions. Contributing factors include temperatures, strong winds, low relative humidity, and the characteristics of rainfall. Drought events are characterized through indices that can be categorized based on the specific impacts they are associated with, such as meteorological, agricultural, or hydrological effects. Using such indices for drought characterization serves multiple purposes, including detection, assessment, and representation of drought conditions within a particular region. Seasonal precipitatio is essential for social and economic development and activities, hence. Reliable seasonal forecasts, especially regarding extreme precipitation events, become crucial for sectors like agriculture and insurance. Europe, and in particular the Mediterranean region, is expected to be considerably affected under climate change. The northern regions are anticipated to exhibit higher variability, increasing the risk of floods, while the southern areas may face decreased rainfall, prolonged dry spells, and intensified evaporation, potentially leading to more frequent drought occurrences.

This research aims to evaluate the prediction skill for extreme drought events at the seasonal time-scale using the SPI and SPEI indices over the EURO-Mediterranean area. The use of SPEI also takes into account the effect of temperature on the water balance, given by the calculation of potential evapotranspiration within it, which can be crucial in a context of global warming. We consider the seasonal forecasts provided by the Copernicus multi-system and we use the Brier Skill Score metric for the assessment of the performance. The objective is to understand potential predictability factors of these indices within the study area. The results show a positive performance for most of the areas examined, between 60 and 80 percent of the entire area for both indices. This led us to investigate possible optimization strategies to increase the skill in the area.

Using the multi-model approach we optimize the prediction skill obtaining considerable performance in forecasting drought conditions. Different multi-model strategies are compared, including the selection or aggregation of available forecasts to achieve the best overall performance in the area. We show that multi-model optimization can indeed provide valuable probabilistic predictions of seasonal drought events in many areas of the Euro-Mediterranean that could be useful for the decision-making process of the affected end users.

How to cite: Dal Monte, T., Alessandri, A., Cherchi, A., and Gaetani, M.: Assessing the predictability of Euro-Mediterranean droughts through seasonal forecasts, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-995, https://doi.org/10.5194/egusphere-egu24-995, 2024.

EGU24-1120 | ECS | Orals | CL4.10

Effects of the realistic vegetation cover representation on the large-scale circulation and predictions at decadal time scale. 

Emanuele Di Carlo, Andrea Alessandri, Fransje van Oorschot, Annalisa Cherchi, Susanna Corti, Giampaolo Balsamo, Souhail Boussetta, and Timothy Stockdale

Vegetation is a highly dynamic component of the Earth System. Vegetation plays a significant role in influencing the general circulation of the atmosphere through various processes. It controls land surface roughness, albedo, evapotranspiration and sensible heat exchanges among other effects. Understanding the interactions between vegetation and the atmosphere is crucial for predicting climate and weather patterns. This study explores how better representation of vegetation dynamics affects climate predictions at decadal timescale and how surface characteristics linked to vegetation affect the general circulation at local, regional and global scales. We used the latest satellite datasets of vegetation characteristics and developed a new and improved parameterization for effective vegetation cover. We implemented the new parameterization in the land surface scheme Hydrology Tiled ECMWF Scheme for Surface Exchanges over Land (HTESSEL), which is embedded in the EC-Earth model. 

The enhancement of the model's vegetation variability significantly improves the prediction skill of the model for several parameters, encompassing both surface and upper-level elements such as 2-metre temperature, zonal wind at 850 hPa and mean sea level pressure. The improvement is particularly evident over Euro-Asian Boreal forests. In particular, a large-scale effect on circulation emerges from the region with the most 2-metre temperature improvement, over Eastern Europe. 

The incorporation of an effective vegetation cover also introduces heightened realism in surface roughness and albedo variability. This, in turn, leads to a more accurate representation of the land-atmosphere interactions. The regression analysis of surface roughness and albedo with 2-metre temperature, mean sea level pressure and wind (both at surface and 850 hPa) reveals a robust relationship across the entire northern hemisphere. This relation between the surface and the atmosphere is notably absent in the standard configuration model, where the vegetation is prescribed by a dynamical vegetation module.

These findings underscore the substantial impact of vegetation cover on the general circulation, particularly in the northern hemisphere, and emphasise its crucial role in improving prediction skills. Furthermore, they highlight the challenges faced by modern earth system models in accurately representing several processes connecting the land surface and the atmosphere.

How to cite: Di Carlo, E., Alessandri, A., van Oorschot, F., Cherchi, A., Corti, S., Balsamo, G., Boussetta, S., and Stockdale, T.: Effects of the realistic vegetation cover representation on the large-scale circulation and predictions at decadal time scale., EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-1120, https://doi.org/10.5194/egusphere-egu24-1120, 2024.

EGU24-1407 | ECS | Posters on site | CL4.10 | Highlight

Time Lag and Cumulative Effects of Extreme Climate on Coastal Vegetation in China 

Dong Tong and Dahai Liu

Rapid global changes are altering regional hydrothermal conditions, especially in ecologically vulnerable regions such as coastal areas of China. The response of vegetation growth to extreme climates and the time lag-accumulation relationship still require further exploration. We characterize the vegetation growth status by solar-induced chlorophyll fluorescence (SIF), analyzed the vegetation dynamic in coastal areas of China from 2000 to 2019, also explored the spatiotemporal pattern of vegetation, and assessed the response of vegetation to extreme climate in term of time lag-accumulation by combines gradual analysis and abrupt analysis. The results showed that (1) Coastal areas of China were sensitive to global climate change, with extreme high temperatures and extreme precipitation increasing from 2000 to 2019, and the warming in high latitudes was greater than in low latitudes, while the increase in precipitation was concentrated in the southern regions, which are already water-rich. (2) The vegetation in coastal areas of China improved significantly, with gradual analysis showed that the vegetation improvement area accounts for 94.12% of the study area, and the abrupt analysis showed that the majority (69.78%) of the vegetation change types were "monotonic increase", with 11.77% showing "increase with negative break" and 9.48% "increases to decreases." (3) Significant lag-accumulation relationships were observed between vegetation and extreme climate in coastal areas of China, and the time-accumulation effects was stronger than time-lag effects. The accumulation time of extreme temperatures was typically less than one month, and the accumulation time of extreme precipitation was 2-3 months. These findings contribute to filling gaps in understanding the time lag-accumulation effects of extreme climates on vegetation in sensitive coastal regions. It provides a foundational basis for predicting the growth trend of coastal vegetation, environmental changes and ecosystem evolution, which is essential for a comprehensive assessment of coastal ecological security.

How to cite: Tong, D. and Liu, D.: Time Lag and Cumulative Effects of Extreme Climate on Coastal Vegetation in China, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-1407, https://doi.org/10.5194/egusphere-egu24-1407, 2024.

EGU24-3134 | Orals | CL4.10

Decadal predictability of seasonal temperature distriubutions 

André Düsterhus and Sebastian Brune

Climate predictions focus regularly on the predictability of single values, like means or extremes. While these information offer important insight into the quality of a prediction system, some stakeholders might be interested in the predictability of the full underlying distribution. These allow beside evaluating the amplitude of an extreme also to estimate their frequency. Especially on decadal time scales, where we verify multiple lead years at a time, the prediction quality of full distributions may offer in some applications important additional value.

In this study we investigate the predictability of the seasonal daily 2m-temperature on time scales of up to ten lead years within the MPI-ESM decadal prediction system. We compare yearly initialised hindcast simulations from 1960 onwards against estimates for climatology and uninitialised historical simulations. To verify the predictions we demonstrate a novel approach based on the non-parametric comparison of distributions with the integrated quadratic distance (IQD).

We show that the initialised prediction system has advantages in particular in the North Atlantic area and allow so to make reliable predictions for the whole temperature distribution for two to ten years ahead. It also demonstrates that the capability of initialised climate predictions to predict the temperature distribution depends on the season. Finally, we will also discuss potential opportunities and pitfalls of such approaches.

How to cite: Düsterhus, A. and Brune, S.: Decadal predictability of seasonal temperature distriubutions, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-3134, https://doi.org/10.5194/egusphere-egu24-3134, 2024.

EGU24-3274 | ECS | Orals | CL4.10

 A Multi-year Climate Prediction System Based on CESM2 

Yong-Yub Kim, June-Yi Lee, Axel Timmermann, Yoshimitsu Chikamoto, Sun-Seon Lee, Eun Young Kwon, Wonsun Park, Nahid A. Hasan, Ingo Bethke, Filippa Fransner, Alexia Karwat, and Abhinav R.Subrahmanian

Here we present a new seasonal-to-multiyear earth system prediction system which is based on the Community Earth System Model version 2 (CESM2) in 1° horizontal resolution. A 20- member ensemble of temperature and salinity anomaly assimilation runs serves as the initial condition for 5-year forecasts. Initialized on January 1st of every year, the CESM2 predictions exhibit only weak climate drift and coupling shocks, allowing us to identify sources of multiyear predictability. To differentiate the effects of external forcing and natural climate variability on longer-term predictability, we analyze anomalies calculated relative to the 50-member ensemble mean of the CESM2 large ensemble. In this presentation we will quantify the extent to which marine biogeochemical variables are constrained by physical conditions. This analysis provides crucial insights into error growth of phytoplankton and the resulting limitations for multiyear predictability.

How to cite: Kim, Y.-Y., Lee, J.-Y., Timmermann, A., Chikamoto, Y., Lee, S.-S., Kwon, E. Y., Park, W., A. Hasan, N., Bethke, I., Fransner, F., Karwat, A., and R.Subrahmanian, A.:  A Multi-year Climate Prediction System Based on CESM2, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-3274, https://doi.org/10.5194/egusphere-egu24-3274, 2024.

EGU24-4083 | Posters virtual | CL4.10 | Highlight

Enhancing Subseasonal Climate Predictions through Dynamical Downscaling: A Case Study in the Southern Plains of the United States 

Yoshimitsu Chikamoto, Hsin-I Chang, Simon Wang, Christopher Castro, Matthew LaPlante, Bayu Risanto, Xingying Huang, and Patrick Bunn

Predicting extreme precipitation events at subseasonal timescales is a critical challenge in Earth system science. This study advances climate predictability by employing dynamical downscaling, specifically focusing on convection-permitting modeling in the Southern Plains of the United States. Two contrasting extreme precipitation periods in Texas, the extremely dry May of 2011 and the abnormally wet May of 2015, were selected for analysis. To enhance subseasonal climate forecasting, we integrated the Weather Research and Forecasting (WRF) model with the decadal climate prediction system based on the Community Earth System Model (CESM). Evaluating the impact of dynamical downscaling on the prediction of extreme precipitation events, our study demonstrates how high-resolution downscaling enhances model skill in capturing these events. The findings hold the potential to significantly contribute to improving climate predictions and assessing regional climate-related risks, aligning with the session's goals.

How to cite: Chikamoto, Y., Chang, H.-I., Wang, S., Castro, C., LaPlante, M., Risanto, B., Huang, X., and Bunn, P.: Enhancing Subseasonal Climate Predictions through Dynamical Downscaling: A Case Study in the Southern Plains of the United States, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-4083, https://doi.org/10.5194/egusphere-egu24-4083, 2024.

Accurate seasonal streamflow forecasts (SSF) are crucial for disaster prevention, water management, agriculture, and hydropower generation. A global approach becomes imperative in regions lacking forecast systems. The Météo-France seasonal prediction system (MF System 8 - SYS8), contributing to Copernicus Climate Change Services (C3S), employs a fully coupled Atmosphere-Ocean General Circulation Model (AOGCM) with an advanced river routing component (CTRIP) interacting with the ISBA land-surface scheme. This study evaluates the skill of the SYS8 global SSF through hindcast river discharges. This work is part of the European project CERISE, which aims to enhance the C3S seasonal forecast portfolio by improving land initialisation methodologies.

SYS8 derives land initial conditions from a historical initialisation run where land (such as soil moisture and river discharges) is weakly constrained, contrasting with the atmosphere and ocean counterparts, which are nudged to the ERA5 and GLORYS re-analysis. This study improves the initialisation run by relaxing soil moisture to fields reconstructed from an offline land simulation.  Daily streamflow ensemble hindcasts of 25 members are generated in a  0.5° grid, with a lead time of up to 4 months initialised on the first day of May and August between 1993-2017. May and August initialisations allow forecasting of summer (JJA) and fall (SON) seasons. Actual forecast skill is assessed against streamflow observations in 1608 monitored basins worldwide (with areas > 3000 km2) using deterministic and probabilistic metrics. The classical Ensemble Streamflow Prediction approach (ESP) serves as a benchmark to evaluate the control SYS8 SSF skill and the additional skill of soil moisture nudging.

Globally, hindcast skill improves with enhanced land-surface initial conditions, especially during summer. Lower latitudes (<50°N) exhibit increased skill, while higher and cooler latitudes may lead to overestimated streamflow magnitude and oscillation amplitude due to soil moisture constraints. Local skill degradation will be discussed. Still, positive results support ongoing efforts to enhance land initialisation through a global land data assimilation system.

How to cite: Narváez, G. and Ardilouze, C.: Global Streamflow Seasonal Forecasts: Impact of soil moisture initialization in a novel two-way AOGCM-River Routing coupling approach, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-5484, https://doi.org/10.5194/egusphere-egu24-5484, 2024.

EGU24-6494 | Posters virtual | CL4.10 | Highlight

Seasonal predictions of summer humid heat extremes in the southeastern United States driven by sea surface temperatures 

Liwei Jia, Thomas Delworth, and Xiaosong Yang

Humid heat extreme (HHE) is a type of compound extreme weather event that poses severe risks to human health. Skillful forecasts of humid heat extremes months in advance are essential for developing strategies to help communities build more resilience to the risks associated with extreme events. This study demonstrates that the frequency of summertime HHE in the southeastern United States (SEUS) can be skillfully predicted 0-1 months in advance in the SPEAR (Seamless system for Prediction and EArth system Research) seasonal forecast system. The sea surface temperature (SST) at the tropical North Atlantic (TNA) basin is found as the primary driver of the prediction skill. The responses of large-scale atmospheric circulation and winds to anomalous warm SSTs in TNA favor the heat and moisture flux transported from the gulf of Mexico to the SEUS. This research demonstrates the role of slowly-varying sea surface conditions in modifying large-scale environments that contribute to the predictions of HHE in SEUS. The results are potentially applicable for developing early warning systems of HHE. 

How to cite: Jia, L., Delworth, T., and Yang, X.: Seasonal predictions of summer humid heat extremes in the southeastern United States driven by sea surface temperatures, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-6494, https://doi.org/10.5194/egusphere-egu24-6494, 2024.

“Synergistic Observing Network for Ocean Prediction (SynObs)” is a project of the United Nations Decade of Ocean Science for Sustainable Development. SynObs aims to find the way to extract maximum benefits from the combination among various ocean observation platforms, including satellite and in situ observations. A major ongoing effort led by SynObs is the international multi-system OSEs/OSSEs. In this activity, various operational centers and research institutes participating will conduct Observing System Experiments (OSEs) and Observing System Simulation Experiments (OSSEs) using a variety of ocean or coupled ocean-atmosphere prediction systems with the common setting to evaluate ocean observation impacts which are robust for most ocean prediction systems. More than 10 ocean prediction systems with various model resolutions and diverse data assimilation methods are used in this activity, and impacts of various observation data, including satellite sea surface temperature and height, Argo floats, and tropical mooring buoys, will be evaluated.

The activity is divided into two parts. The first part is the ocean prediction OSEs. In this part, we run several ocean reanalysis runs assimilating different observation datasets at least for 2020 (preferably extended to 2022), and conduct 10-day ocean predictions from the reanalysis fields of every 5 days. Three-dimensional oceanic temperature, salinity, and velocity fields with the 1/10-degree resolution, and several two-dimensional diagnostics with the 1/4-degree resolution will be analyzed. The second part is the subseasonal-to-seasonal (S2S) OSEs. Here, we run several ocean reanalysis runs for 2003-2022, and conduct 1-month (4-month) coupled predictions from the reanalysis fields of every month (twice a year). We will evaluate the impacts of ocean observation data on the long-term reanalysis and S2S predictions using the coupled prediction systems. We also plan to conduct OSSEs using multiple ocean prediction systems in order to assess newly emerging or future observing systems, such as SWOT, ocean gliders, etc. 

We are currently conducting the S2S OSEs using a Japanese operational global ocean data assimilation and coupled prediction system for S2S forecasts. We are now conducting OSEs assimilating no in situ observations and withholding temperature and salinity profiles observed by Argo floats. In the presentation, we will introduce the results and the perspective of the collaborative activities.

How to cite: Fujii, Y., Ishikawa, I., and Hirahara, S.: Early results of OSEs conducted for the SynObs international multi-system OSE effort using an Japanese operational system, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-6970, https://doi.org/10.5194/egusphere-egu24-6970, 2024.

EGU24-7918 | ECS | Orals | CL4.10

Generation of sea ice initial conditions for the next Météo-France seasonal forecasting system 

Fousiya Thottuvilampil Shahulhameed, Jonathan Beuvier, and Damien Specq

Research and development activities around the current Météo-France operational seasonal forecasting system (System 8) are underway to upgrade it to the next version (System 9), along with efforts to improve the initialization of its components. Among these components, sea ice is particularly challenging to initialize. At present, a coupled-nudged initialisation strategy, based on a high-resolution configuration of the CNRM-CM6 climate model, is employed to initialise the System 8, except for the sea-ice. In order to get initial states of sea ice that are consistent with the forecasting model, our procedure consists in making a preliminary continuous run where the ocean and sea ice models are integrated in stand-alone mode, with forcing at the surface from an atmosphere reanalysis.

However, in the current operational System 8 – based on the NEMO 3.6 ocean model and the GELATO sea ice model – the initial states of sea ice generated with this procedure are not fully realistic. Results show that the sea ice thickness over the Arctic region in the System 8 initial states is underestimated compared to the reference data. Numerous sensitivity experiments were carried out with the current NEMOv3.6-GELATO system, leading to some minor improvements. Thus, an upgraded version of the ocean model (NEMO version 4.2) coupled to a new sea-ice component (SI3) has been tested (in stand-alone mode, not coupled to the atmosphere) to see if the use of more recent versions of ocean and sea-ice models leads to some improvements in the Arctic sea ice representation. The results are encouraging as the representation of sea ice variables in the Arctic is improved compared to the old version.

This incites our team to foresee that System 9 will indeed incorporate the NEMO4.2 and SI3 models, and that the same initialization procedure as before (using these new models) will provide sea-ice initial states closer to those observed.

 

 

How to cite: Thottuvilampil Shahulhameed, F., Beuvier, J., and Specq, D.: Generation of sea ice initial conditions for the next Météo-France seasonal forecasting system, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-7918, https://doi.org/10.5194/egusphere-egu24-7918, 2024.

EGU24-11927 | Posters virtual | CL4.10 | Highlight

Seasonal prediction of solar energy resources in the United States 

Xiaosong Yang, Thomas Delworth, Liwei Jia, Nathaniel Johnson, Feiyu Lu, and Colleen McHugh

Solar energy plays a crucial role in the transition towards a sustainable and resilient energy future. One challenge that remains is the considerable year-to-year variation in solar energy resources. As a result, precise seasonal solar energy predictions become pivotal for effective energy system planning and operation.  This study employs GFDL’s GFDL’s Seamless System for Prediction and Earth System (SPEAR) to evaluate seasonal solar irradiance prediction across the United States.  Notably, SPEAR demonstrates high skill in predicting solar irradiance particularly in the western United States. Furthermore, we conduct an advanced predictability analysis to pinpoint the underlying physical drivers contributing to this skillful solar energy prediction.  The outcomes of this research offer substantial potential benefits to stakeholders within the energy sector by providing predictable information regarding year-to-year fluctuations in solar energy resources.

How to cite: Yang, X., Delworth, T., Jia, L., Johnson, N., Lu, F., and McHugh, C.: Seasonal prediction of solar energy resources in the United States, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-11927, https://doi.org/10.5194/egusphere-egu24-11927, 2024.

EGU24-11948 | Posters on site | CL4.10

What is the Target for Multi-Model and Perturbed-Physics Ensembles? 

David Stainforth

Much effort goes into studying the causes of systematic errors in Earth System Models (ESMs). Reducing them is often seen as a high priority. Indeed, the development of Digital Twin approaches in climate research is founded on the idea that a sufficiently good model would be able to provide reliable and robust, conditional predictions of climate change (predictions conditioned on scenarios of future greenhouse gas emissions). Here, “reliable” encapsulates the idea that the predictions are suitable for use by society in anticipating and planning for future climate change, and “robust” encapsulates the idea that they are unlikely to change as the models are improved and developed.

Such an approach, however, begs the question, when is a model sufficiently realistic to be able to provide reliable, detailed predictions? A physical processes view of current ESMs suggests that they are not close to this level of realism while a nonlinear dynamical systems perspective raises questions over whether it will ever be possible to achieve such reliability for the types of regionally-specific, extrapolatory, climate change predictions that we may think society seeks.

Given this context, multi-model and perturbed-physics ensembles are often seen as a means to quantify uncertainty in conditional, climate change predictions (commonly referred to as “projections” in the scientific community). In the IPCC atlas (https://interactive-atlas.ipcc.ch/) the most easily accessible output is the multi-model median with the 10th, 25th, 75th and 90th percentiles of the multi-model distribution also prominent. This presentation in terms of probabilities implies that the probabilities themselves have meaning to the users of the data - most users are likely to take them as probabilities of different outcomes in reality. Unfortunately multi-model ensembles cannot be interpreted that way because we have no metric for the shape of model space nor any idea of how to explore it, so the ensemble members cannot be taken as independent samples of possible models. Perturbed-parameter ensembles work in a more defined space of possible model-versions but the shape of that space is also undefined and as a result the ensemble-based probabilities are again arbitrary.

When seeking the best possible information for society, multi-model and perturbed physics ensembles would benefit from targeting diversity: the greatest possible range of responses given a particular model structure. Model emulators could be used to systematise this process. Such an approach would provide more reliable information. It changes the question, however, from “when is a model sufficiently realistic” to “how unrealistic does a model have to be to be uninformative about extrapolatory future climatic behaviour?”

In this presentation I will discuss and elaborate on these issues.

 

References:

Stainforth, D., “What we do with what we’ve got”, Chapter 21 in “Predicting Our Climate Future: What we know, what we don’t know and what we can’t know”, Oxford University Press, 2023.

Stainforth, D.A. et al., Confidence, uncertainty and decision-support relevance in climate predictions, Phil.Trans.Roy.Soc., 2007.

Stainforth, D.A. et al., Issues in the interpretation of climate model ensembles to inform decisions, Phil.Trans.Roy.Soc., 2007.

How to cite: Stainforth, D.: What is the Target for Multi-Model and Perturbed-Physics Ensembles?, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-11948, https://doi.org/10.5194/egusphere-egu24-11948, 2024.

EGU24-12988 | ECS | Posters on site | CL4.10

A CNN-based Downscaling Method of C3S Seasonal Forecast: Temperature and Precipitation 

Qing Lin, Yanet Díaz Esteban, Fatemeh Heidari, Edgar Fabián Espitia Sarmiento, and Elena Xoplaki

Copernicus Climate Change Service provides seasonal forecasts for meteorological outlooks several months in advance and can provide indications of future climate risks on a global scale. Using downscaling techniques, global variables can be transferred to the high-resolution regional scale, allowing the information to be elaborated for extreme events detection and further implementing and coupling with hydrological models for regional hazard prediction, thus serving agriculture and energy, improving planning for tourism and other sectors.

In this study, we applied a new CNN-based architecture for temperature and precipitation downscaling. Both variables are downscaled from 1 degree to 1 arcminute to fulfill the requirements as an input to the hydrological models. The architecture implements an auto-encoder/decoder structure to extract the data relations. The system is trained with seasonal forecast inputs and observation data to establish the relation between both scales. The model is then evaluated with the validation period from the observation data to achieve the best performance, changing network structures and tuning different network hyper-parameters. The results show a good fit for the observation data on the monthly scale, providing enough details in the downscaling product. Finally, the best-performing networks for downscaling temperature and precipitation are selected and could be extended for further utilization.

How to cite: Lin, Q., Díaz Esteban, Y., Heidari, F., Espitia Sarmiento, E. F., and Xoplaki, E.: A CNN-based Downscaling Method of C3S Seasonal Forecast: Temperature and Precipitation, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-12988, https://doi.org/10.5194/egusphere-egu24-12988, 2024.

EGU24-13811 | ECS | Posters on site | CL4.10

Estimating Seasonal to Multi-year Predictability of Statistics of Climate Extremes using the CESM2-based Climate Prediction System 

Alexia Karwat, June-Yi Lee, Christian Franzke, and Yong-Yub Kim

Climate extremes, such as heat waves, heavy precipitation, intense storms, droughts, and wildfires, have become more frequent and severe in recent years as a consequence of human-induced climate change. Estimating the predictability and improving prediction of the frequency, duration, and intensity of these extremes on seasonal to multi-year timescales are crucial for proactive planning and adaptation. However, climate prediction at regional scales remains challenging due to the complexity of the climate system and limitations in model accuracy. Here we use a large ensemble of simulations, assimilations, and reforecasts using Community Earth System Model version 2 (CESM2) to assess the predictability of statistics of climate extremes with lead times of up to 5 years. We show that the frequency and duration of heat waves during local summer in specific regions are predictable up to several months to years. Sources of long-term predictability include not only external forcings but also modes of climate variability across time scales such as El Niño and Southern Oscillation, Pacific Decadal Variability, and Atlantic Multidecadal Variability. This study implies opportunities to deepen our scientific understanding of sources for long-term prediction of statistics of climate extremes and the potential for the associated disaster management.

How to cite: Karwat, A., Lee, J.-Y., Franzke, C., and Kim, Y.-Y.: Estimating Seasonal to Multi-year Predictability of Statistics of Climate Extremes using the CESM2-based Climate Prediction System, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-13811, https://doi.org/10.5194/egusphere-egu24-13811, 2024.

EGU24-15488 | ECS | Orals | CL4.10

Phytoplankton predictability in the Tropical Atlantic - triggered by nutrient pulses from the South 

Filippa Fransner, Marie-Lou Bachèlery, Shunya Koseki, David Rivas, Noel Keenlyside, Nicolas Barrier, Matthieu Lengaigne, and Olivier Maury

The variability and predictability of the Tropical Atlantic primary productivity remains little explored on interannual-to-decadal time scales. Here, we  present the results of two studies, in which find a decadal scale variability in phytoplankton abundance that can be predicted three years ahead. The predictions are made with NorCPM, which is a fully coupled climate prediction model with ocean biogeochemistry that assimilates temperature and salinity to reconstruct past variability. From these reconstructions, predictions are initialized that are run freely ten years ahead. We find that the predictability is a result of nutrient pulses that are advected with the southern branch of the South Equatorial Current from the most southern part of the Atlantic, and that then get caught in the Equatorial undercurrent before they reach the surface in the Tropical Atlantic Ocean. A more detailed analysis is being done in order to pinpoint the underlying mechanisms in a forced ocean model, where we find a link to the Pan-Atlantic decadal oscillation.

How to cite: Fransner, F., Bachèlery, M.-L., Koseki, S., Rivas, D., Keenlyside, N., Barrier, N., Lengaigne, M., and Maury, O.: Phytoplankton predictability in the Tropical Atlantic - triggered by nutrient pulses from the South, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-15488, https://doi.org/10.5194/egusphere-egu24-15488, 2024.

EGU24-15829 | Posters on site | CL4.10

The role of realistic vegetation variability in climate predictability and prediction 

Andrea Alessandri, Emanuele Di Carlo, Franco Catalano, Bart van den Hurk, Magdalena Alonso Balmaseda, Gianpaolo Balsamo, Souhail Boussetta, and Tim Stockdale

Vegetation is a relevant and highly dynamic component of the Earth system and its variability – at seasonal, interannual, decadal and longer timescales – modulates the coupling with the atmosphere by affecting surface variables such as roughness, albedo and evapotranspiration. In this study, we investigate the effects of improved representation of vegetation dynamics on climate predictability and prediction at the seasonal timescale. To this aim, the observational constraints from the latest generation satellite dataset of vegetation Leaf Area Index (LAI) have been integrated in the modeling, including a parameterization of the effective vegetation cover as a function of LAI. The improved vegetation representation is implemented in HTESSEL, which is the land surface model included in the seasonal forecasting (ECMWF SEAS5) systems used in this work.

Our results show that the realistic representation of vegetation variability has significant effects on both potential predictability and actual prediction skill at the seasonal time scale. It is shown a significant improvement of the skill in predicting boreal winter (December-January-February; DJF) 2m Temperature (T2M) at 1-month lead time especially over Euro-Asian boreal forests; the improvement is at least in part due to the more realistic representation of the interannual albedo variability that is related to the changes in vegetation shading over snow. Remarkably, from the region with the most considerable T2M improvement originates a large-scale ameliorating effect on circulation encompassing Northern Hemisphere middle-to-high latitudes from Siberia to the North Atlantic. The results indicate that the coupling with the improved vegetation might operate by amplifying locally the signal originating from the North Atlantic sector, therefore improving both potential predictability and actual skill over the region. Concurrently, the improved predictability and skill over the Euro-Asian forests appears to feedback to the large-scale circulation enhancing the representation of the circulation pattern and associated interannual anomalies.

How to cite: Alessandri, A., Di Carlo, E., Catalano, F., van den Hurk, B., Balmaseda, M. A., Balsamo, G., Boussetta, S., and Stockdale, T.: The role of realistic vegetation variability in climate predictability and prediction, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-15829, https://doi.org/10.5194/egusphere-egu24-15829, 2024.

EGU24-16402 | Orals | CL4.10

On the stationarity of the global spatial dependency of heat risk on drought. 

Matteo Zampieri, Karumuri Ashok, Andrea Toreti, Davide Bavera, and Ibrahim Hoteit

Compound climate anomalies pose escalating risks in the context of climate change, with anomalous heat and drought presenting significant stressors to both ecosystems and society. The simultaneous occurrence of these events can be influenced by land surface processes such as the soil moisture – air temperature coupling. However, the long-term variability of this coupling remains unexplored. Here, using a combination of observations and multi-model ensemble forecasts dating back to the 1980s, we examine the global land exposure to higher than normal probabilities of concurrent hot temperature anomalies and drought on a monthly scale. Our findings confirm that drought substantially shapes the spatial distribution of heat-related risks on a global scale, offering a crucial predictive factor for these combined events. Traditionally, defining heat anomalies for non-adaptive systems involves fixed reference temperature thresholds. Using this method, our analysis reveals that the portion of global land experiencing drought-conditioned hot temperature anomalies has tripled in less than three decades. Surprisingly, the global level of spatial coupling appears to be declining. However, this outcome heavily depends on the specific definition of heat risk employed. By employing a time-dependent temperature threshold that considers changes in the climate's mean state due to both global warming and natural variability, a different picture emerges. Using the latter method, the level of spatial coupling demonstrates persistence and stability. Importantly, this method is better suited to assessing risks for adaptive systems and is more consistent with our current understanding of the underlying processes. Our study strongly advocates for tailoring hazard definitions to the specific processes and systems under investigation. Additionally, it underscores the pivotal role of operational sub-seasonal and seasonal forecasts in early warning systems, crucial for societal adaptation in the face of global warming.

How to cite: Zampieri, M., Ashok, K., Toreti, A., Bavera, D., and Hoteit, I.: On the stationarity of the global spatial dependency of heat risk on drought., EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-16402, https://doi.org/10.5194/egusphere-egu24-16402, 2024.

EGU24-16456 | Orals | CL4.10

Advancements and Challenges in Assessing and Predicting the Global Carbon Cycle Variations Using Earth System Models 

Hongmei Li, Tatiana Ilyina, István Dunkl, Aaron Spring, Sebastian Brune, Wolfgang A. Müller, Raffaele Bernardello, Laurent Bopp, Pierre Friedlingstein, William J. Merryfield, Juliette Mignot, Michael O'Sullivan, Reinel Sospedra-Alfonso, Etienne Tourigny, and Michio Watanabe

The imperative to comprehend and forecast global carbon cycle variations in response to climate variability and change over recent decades and in the near future underscores its critical role in informing the global stocktaking process. Our study investigates CO2 fluxes and atmospheric CO2 growth through ensemble decadal prediction simulations using Earth System Models (ESMs) driven by CO2 emissions with an interactive carbon cycle. These prediction systems provide valuable insights into the global carbon cycle and, therefore, the variations in atmospheric CO2. Assimilative ESMs with interactive carbon cycles effectively reconstruct and predict atmospheric CO2 and carbon sink evolution. The emission-driven prediction systems maintain comparable skills to conventional concentration-driven methods, predicting 2-year accuracy for air-land CO2 fluxes and atmospheric CO2 growth, with air-sea CO2 fluxes exhibiting higher skill for up to 5 years. Our multi-model predictions for the next year, along with assimilation reconstructions, for the first time contribute to the Global Carbon Budget 2023 assessment. We plan regular updates and the involvement of more ESMs in future assessments. Ongoing efforts include implementing seasonal-scale predictions for skill improvement. Furthermore, we assess uncertainty contributions to CO2 flux and growth predictions, revealing the comparable impacts of internal climate variability and diverse model responses, particularly at a lead time of 1-2 years. Notably, the effect of CO2 emission forcing rivals internal variability at a 1-year lead time. Large uncertainties in CO2 responses to initial states of ENSO are observed, stemming from both model responses and internal variability. The challenge lies in addressing the scarcity and uncertainty of data for initialization and obtaining precise external forcings to enhance the reliability of predictions. The further advancements involve not only addressing comprehensive bias correction but also implementing statistical methods to enhance dynamical predictions.

How to cite: Li, H., Ilyina, T., Dunkl, I., Spring, A., Brune, S., Müller, W. A., Bernardello, R., Bopp, L., Friedlingstein, P., Merryfield, W. J., Mignot, J., O'Sullivan, M., Sospedra-Alfonso, R., Tourigny, E., and Watanabe, M.: Advancements and Challenges in Assessing and Predicting the Global Carbon Cycle Variations Using Earth System Models, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-16456, https://doi.org/10.5194/egusphere-egu24-16456, 2024.

EGU24-16842 | Posters on site | CL4.10 | Highlight

Exploring Sources of Multi-year Predictability of Terrestrial Ecosystem 

June-Yi Lee, Yong-Yub Kim, and Jeongeun Yun

The demand for decision-relevant and evidence-based near-term climate information is increasing. This includes understanding and explaining the variability and changes in ecosystems to support disaster management and adaptation choices. As climate prediction from seasonal to decadal (S2D) expands to encompass Earth system dimensions, including terrestrial and marine ecosystems, it is crucial to deepen our scientific understanding of the long-term predictability sources for ecosystem variability and change. Here we explore to what extent terrestrial ecosystem variables are driven by large-scale - potentially predictable -climate modes of variability and external forcings or whether regional random environmental factors are dominant. To address these issues, we utilize a multi-year prediction system based on Community Earth System Model version 2 (CESM2).  The system consists of 50-member uninitialized historical simulations, 20-member ocean assimilations, and 20-member hindcast initiated from every January 1st integrating for 5 years from 1961 to 2021. The key variables assessed are surface temperature, precipitation, soil moisture, wildfire occurrence, and Gross Primary Productivity. Our results suggest that land surface processes and ecosystem variables over many parts of the globe can be potentially predictable 1 to 3 years ahead originating from anthropogenic forced signals and modes of climate variability, particularly El Nino and Southern Oscillation and Atlantic Multi-decadal variability. These global modes of climate variability shift regional temperature and precipitation patterns, leading to changes in soil moisture, wildfire occurrence, and terrestrial productivity.  

How to cite: Lee, J.-Y., Kim, Y.-Y., and Yun, J.: Exploring Sources of Multi-year Predictability of Terrestrial Ecosystem, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-16842, https://doi.org/10.5194/egusphere-egu24-16842, 2024.

EGU24-18766 | Orals | CL4.10

Deciphering Prediction Windows of Opportunity: A Cross Time-Scale Causality Framework   

Stefano Materia, Constantin Ardilouze, and Ángel G. Muñoz

While subseasonal forecasts often exhibit limited skill across mid-latitudes, occasional improvements are observed in specific locations during certain periods, known as "windows of opportunity." Understanding the causal factors behind these windows is complex due to the diverse and interdependent nature of predictors, their spatial and temporal variability, and the challenges in establishing causality relationships. 

Traditional lagged-correlations methods provide only a partial view, lacking insights into causality. Based on previous work on the role of land surface processes, multi-model subseasonal model skill assessment and the use of causality metrics in predictions across timescales (e.g. Ardilouze et al., 2020, 2021; Materia et al 2020, 2022; Muñoz et al., 2023), here we propose an approach based on the Liang-Kleeman information flow, allowing the assessment of statistically significant causal links across various lead times.

Applied to reforecast and reanalysis data, our framework successfully identifies significant predictability drivers -involving sea-surface temperatures, atmospheric circulation and remote and local land-surface processes-, revealing their interference (interplay), evolving patterns and prevalence from seasonal to subseasonal scales. 

Furthermore, the comparison between reanalysis and reforecast results aids in assessing the capability of models to capture these causality features, suggesting additional ways to conduct model diagnostics. We illustrate here the theoretical background by showcasing the causal factors influencing a window of opportunity identified from a multimodel subseasonal reforecast.

 

References

Ardilouze, C., Materia, S., Batté, L., Benassi, M., & Prodhomme, C. (2020). Precipitation response to extreme soil moisture conditions over the Mediterranean. Climate Dynamics, 1, 1–16. https://doi.org/10.1007/S00382-020-05519-5/TABLES/2

Ardilouze, C., Specq, D., Batté, L., & Cassou, C. (2021). Flow dependence of wintertime subseasonal prediction skill over Europe. Weather and Climate Dynamics, 2(4), 1033-1049. https://doi.org/10.5194/wcd-2-1033-2021 

Materia, S., Muñoz, Á. G., Álvarez-Castro, M. C., Mason, S. J., Vitart, F., & Gualdi, S. (2020). Multi-model subseasonal forecasts of spring cold spells: potential value for the hazelnut agribusiness. Weather and Forecasting. https://doi.org/10.1175/waf-d-19-0086.1 

Materia, S., Ardilouze, C., Prodhomme, C., & et al. (2022). Summer temperature response to extreme soil water conditions in the Mediterranean transitional climate regime. Climate Dynamics, 58, 1943–1963. https://doi.org/10.1007/s00382-021-05815-8

Muñoz, Á. G., Doblas-Reyes, F., DiSera, L., Donat, M., González-Reviriego, N., Soret, A., Terrado, M., & Torralba, V. (2023). Hunting for “Windows of Opportunity” in Forecasts Across Timescales? Cross it. EGUGA, EGU-15594. https://doi.org/10.5194/EGUSPHERE-EGU23-15594 

How to cite: Materia, S., Ardilouze, C., and Muñoz, Á. G.: Deciphering Prediction Windows of Opportunity: A Cross Time-Scale Causality Framework  , EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-18766, https://doi.org/10.5194/egusphere-egu24-18766, 2024.

NP6 – Turbulence, Transport and Diffusion

EGU24-1427 | ECS | Orals | NP6.2

Experimental study of gravity current propagation over rough tilted surfaces 

Mostafa Shehata, Marie Rastello, Florence Naaim, and Herve Bellot

Numerous previous studies have been done to understand the physics of gravity currents. Some considered the propagation over smooth or rough horizontal surfaces (Tokyay et al. (2014), Zhou et al. (2017)), and others studied inclined surfaces without any roughness. What about the more complex situation of inclined rough surfaces? We have studied, experimentally, how a finite volume of heavy fluid (salted water with a clay suspension) rushing down a slope (20° & 30°) is affected by the multiple obstacles it counters on its way. Our setup is a classic volume release configuration in a 2D flume immersed in a 20 m3 water tank at INRAE Grenoble. The study tests (352 experiments per tilt) different initial conditions of the released flow (volume and density) and various surface conditions from smooth to rough:

  • For the initial conditions: the experiments show that the front velocity is monotonically related to its initial volume and density. Although the initial mass of the flow is the product of its density times volume, the mass effect on the flow front velocity cannot come in replacement of both volume and density effects, as it was found that the front velocity is non-monotonically related to its initial mass.
  • For the surface conditions: besides testing the smooth case, we have covered a wide range of roughness configurations using obstacles with different shapes, heights, and spacings. Walls or barriers blocking the whole width of the flume have been used (see Fig.1-a). Testing various heights and spacings shows that higher barriers decrease the flow front velocity, while non-monotonic relations were found when the spacing between successive barriers in the flow direction is changed. Flow propagation over and through an array of obstacles has also been studied with various obstacles arrangement (in-line and staggered) and different obstacles' cross-sections (rectangular and circular). For circular obstacles, the (x𝑓-t) curve is no longer smooth but takes the shape of stairsteps, and they are found to be more efficient in decelerating the flow (see Fig.1-b).

Studying both 20° and 30°-flume tilts enables us to look through the slope effect. The analysis shows that, in general, increasing the slope results in higher front velocity values. Nevertheless, the degree of influence is dependent on diverse factors (volume, density, bed surface conditions). In addition, we have studied the effect of the initial flow parameters on the flow height just after the lock release (at an accurate predetermined distance from the lock chosen based on 252 experiments). This height depends only on the initial volume and density effect is negligible. Determination of this height is essential for our non-dimensional analysis: to study the temporal evolution of the non-dimensional front position (𝑥𝑓−𝑥o)/𝑥o versus the non-dimensional time (t𝑓/to). Indeed, it will enable us to avoid using the initial flow depth at the lock that is highly dependent on the inclination angle, or an estimated virtual height after the lock that would be less representative (see Fig.1-c).

Fig. 1: 

 

 

How to cite: Shehata, M., Rastello, M., Naaim, F., and Bellot, H.: Experimental study of gravity current propagation over rough tilted surfaces, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-1427, https://doi.org/10.5194/egusphere-egu24-1427, 2024.

EGU24-1871 | Posters on site | NP6.2

Internal solitary wave energy transformations under ridged ice cover 

Kateryna Terletska, Vladimir Maderich, and Elena Tobisch

Propagation of internal solitary waves (ISW) under the edge of the ice cover may lead to their
destabilization through overturning and breaking events. Factors such as ice cover depth, ridging
intensity, and internal wave amplitudes play crucial roles in the evolution and disintegration of ISW
beneath the ice cover. In the study, a numerical investigation of the transformation of ISW
propagating from open water in the stratified sea under ridged ice cover is carried out. A
nonhydrostatic numerical model, that is based on the Reynolds averaged Navier-Stokes equations in
the Boussinesq approximation for a continuously stratified fluid, was used in the investigation. The
study focused on an idealized scenario with a vertical distribution of potential density anomalies
designed to replicate the summer profile of potential density observed over the Yermak Plateau in
the Arctic Ocean. In the numerical experiments, number of ice keels were placed beneath a
uniform-thickness ice layer. The ice keel shape was approximated by the Versoria function. It is
carried out calculations with a different ridging intensity, that is, the ratio of the maximum height of
the keel to the distance between the keels. In present calculations, it varies from 1/1000 for
moderately ridged ice to 1/20 for heavily ridged ice, which is broadly consistent with the ocean
values. The transformation of ISW of depression is additionally governed by the blocking
parameter β for a single keel, which is the ratio of the height of the minimum thickness of the upper
layer under the ice keel to the incident wave amplitude. An important characteristic of the ISW-
ridged ice interaction is the loss of kinetic and available potential energy during the ISW
transformation. Energy transformation due to mixing leads to the transition of energy to background
potential energy and energy dissipation. To characterize the dependence of energy loss on keel
height and distance between keels, we introduced the parameter, which is the ratio of the sum of
submerged ice thickness and maximal keel penetration to the distance between keels. An energy
loss was estimated based on a budget of depth-integrated pseudoenergy before and after the wave
transformation. The results revealed that the energy loss increases with a decrease in distance
between keels or an increase in keel height. The level of energy loss is highest for β values near
zero. For values β greater than 0.8, interaction is moderate or weak, and distance between the keels
no longer affects energy loss.

How to cite: Terletska, K., Maderich, V., and Tobisch, E.: Internal solitary wave energy transformations under ridged ice cover, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-1871, https://doi.org/10.5194/egusphere-egu24-1871, 2024.

EGU24-2296 | Posters on site | NP6.2

Obliquely interacting breathers in a three-layer fluid 

Keisuke Nakayama and Kevin Lamb

In a three-layer system with equal upper and lower thin layers with the same density jump across each interface, fully nonlinear governing equations have revealed that breathers exist under the Boussinesq approximation. Also, it has been demonstrated that breathers may occur in the Baltic Sea. Additionally, in previous studies, it has been shown that the larger the upper- and lower-layer thicknesses, the more the breathers behave like solitons and that phase shifts occur after two breathers interact, with a forward/backward shift of the faster/slower breather, while the properties of the breathers are preserved. Still, the oblique interaction of breathers has yet to be explored. Thus, we aimed to investigate oblique breather interactions in a three-layer system by using fully nonlinear numerical simulations to clarify the characteristics of breathers. The ratio of the thin layer thicknesses to the total depth was 0.25 in this study. Breathers have two significant parameters, p and q, corresponding to the wavelength of a breather and the envelope amplitude. So, we had several configurations to clarify the influence of incident angles and amplitudes on the breather interactions by changing the parameters p and q. Stably progressing breathers, where p and q are 0.025 and 0.006, were examined by changing the incident angles from 10 to 40 degrees to estimate a critical angle. Also, the oblique breather interactions with a larger envelope amplitude were simulated to analyse the amplitude dependence of the critical angle. A Mach stem was found to occur in oblique breather interactions. Also, the critical angle was revealed to decrease as the envelope amplitude decreases. The behaviour of obliquely-interacting breathers provides further evidence that breathers in a three-layer fluid have soliton-like characteristics.

How to cite: Nakayama, K. and Lamb, K.: Obliquely interacting breathers in a three-layer fluid, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-2296, https://doi.org/10.5194/egusphere-egu24-2296, 2024.

EGU24-3505 | ECS | Orals | NP6.2

Direction Symmetry of Wave Field Modulation by Tidal Current and its Consequences for Extreme Nonlinear Waves 

Saulo Mendes, Ina Teutsch, and Jérôme Kasparian

Theoretical studies on the modulation of unidimensional regular waves over a flat bottom due to a current typically assign an asymmetry between the effects of opposing/following streams on the evolution of major sea variables, such as significant wave height. The significant wave height is expected to monotonically increase with opposing streams and to decrease with following streams. To some extent, observations on data sets containing a few thousand of waves or over a continuous series of about a day confirm this prediction. Here we show that in very broad-banded seas with high directional spread, the asymptotic behavior of sea variables over large data sets is highly non-trivial and does not follow the theoretical predictions, especially at high values of the ratio between tidal stream and group velocity. Furthermore, we analyze the anomalous statistics originating from both forward and opposing non-stationary currents. Despite the sea states being dominantly broad-banded and featuring a large directional spread, we found that anomalous statistics are of the same order of magnitude of those observed in unidirectional laboratory experiments and symmetrical in regard to the orientation of the tidal current.

How to cite: Mendes, S., Teutsch, I., and Kasparian, J.: Direction Symmetry of Wave Field Modulation by Tidal Current and its Consequences for Extreme Nonlinear Waves, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-3505, https://doi.org/10.5194/egusphere-egu24-3505, 2024.

We investigate the properties of, and can carry out a stability analysis of a baroclinic current in, a stratified thermal rotating shallow-water model for the subinertial dynamics of the upper ocean, a key player in the global climate system where most of the ocean variability is concentrated affecting the lateral transport of floating material such as plastic garbage, oil, and Sargassum seaweed.  Unlike the standard thermal model, the model considered here includes linear buoyancy variation in the vertical, still maintaining its two-dimensional structure and that of the adabatic (constant density) model.  Like the standard thermal model, the stratified thermal model produces submesoscale circulations resembling those observed in satellite imagery, yet taking longer to manifest.  Our study is motivated by this numerical observation. The model possesses a Lie--Poisson Hamiltonian structure.  A particular aspect of the model is that it supports motion integrals which neither form the kernel of the corresponding bracket nor are related to any explicit symmetries via Noether's theorem.  Among other things, we investigate the role of these conservation laws in constraining the growth of finite-amplitude perturbations to a zonal flow with quadratic vertical shear. Joint work with Maria J. Olascoaga.

How to cite: Beron-Vera, F.: Properties and baroclinic instability of stratified thermal ocean flow, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-3654, https://doi.org/10.5194/egusphere-egu24-3654, 2024.

Internal waves have a wide range of scales but are typically unresolved in climate or global models. With an unprecedented capability of observing and simulating these processes, they are becoming increasingly important to quantify the upscale effect of these processes. As the largest marginal sea in the western Pacific, the South China Sea has the most energetic and frequent internal waves around the world. These waves are also affected by multiscale processes, climate changes, and anthropogenic impacts. There have been considerable advances in exploring the generation and propagation of internal waves in recent years. However, the understanding of the formation and fate of internal solitary-like waves on the continental shelf is still very limited. It is widely accepted that these internal waves generally originate from the Luzon Strait. They usually have regular occurrence and are phase-locked to tidal forcing in the Luzon Strait. However, we present field measurements showing an irregular occurrence of nonlinear internal waves on the northern shelf of the South China Sea. This irregular occurrence is in striking contrast to the prominent predictability of internal waves originating from the Luzon Strait. We reveal that the intermittent nature of the occurrence is due to the local generation of nonlinear internal waves on the continental shelf, in addition to the fission of shoaling internal waves. The results reported here are expected to apply to other shelf regions of the world's oceans.

How to cite: Bai, X.: Intermittent Generation of Nonlinear Internal Waves on Continental Shelf, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-4294, https://doi.org/10.5194/egusphere-egu24-4294, 2024.

EGU24-5897 | ECS | Orals | NP6.2

On the turbulent structures generated by intruding downslope rotating gravity currents 

Tassigny Axel, Negretti Maria Eletta, and Wirth Achim

Gravity currents play a crucial role in the formation of deep waters in the ocean, contributing to the vorticity and energy transfers towards the ocean interior. We present results from an experimental study on downslope intruding and rotating gravity currents into an initially two-layer stably stratified ambient at high buoyancy Reynolds numbers. A new turbulent process of downslope transport, intermittent and localized, is identified, taking the form of cascades. The lifetime of cascades presents a power law relationship, and the related transport does not exhibit any characteristic length scale, suggesting self-organized criticality. Cascades reveal to be the main contributor to the vorticity and turbulence in the ocean interior, with a dependence on the Coriolis parameter and the density anomaly to the surrounding ambient. Vorticity is produced both by the spreading of the cascade into the interior, and by the meandering and the break up of the deep boundary current (formed from downward Ekman transport). When the intrusion spreads at the pycnocline only, anticyclonic eddies are formed in the intrusion and top layers, whereas for intrusions spreading through the full bottom layer, vortices of both signs are generated due to bottom friction. The turbulence in the receiving ambient reveals to be horizontally isotropic, non-stationary and non-homogeneous. In the intrusion area close to the slope, the turbulence is forced by energy injection at the penetration length scale through the cascades. The central area far from the boundaries is characterized, instead, by freely evolving two-dimensional turbulence, forced at large scales.

 

How to cite: Axel, T., Maria Eletta, N., and Achim, W.: On the turbulent structures generated by intruding downslope rotating gravity currents, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-5897, https://doi.org/10.5194/egusphere-egu24-5897, 2024.

EGU24-5954 | Posters on site | NP6.2 | Highlight

A physical model of the Gibraltar Strait: the HERCULES experiment 

Maria Eletta Negretti, Axel Tassigny, and Louis Gostiaux

Gravity currents are one of the key sub-mesoscale processes that drive energy transfer, impact the thermohaline structure and the vertical exchange of water masses in the ocean. At present, their representation remains difficult in numerical models. The targeted study area is the Strait of Gibraltar between the Mediterranean and the Atlantic Ocean. We present preliminary results of experiments obtained using the first realistic implementation of the Strait of Gibraltar with the adjacent Gulf of Cadiz and Alboran Sea, including all main forcings: the density difference, the barotropic tide, the Earth’s rotation and the realistic topography, scaled using available in-situ data. Detailed measurements of the velocity and density fields reveal that the large-scale circulation and the further faith of the Mediterranean waters flowing into the Atlantic Ocean are strongly influenced by the turbulent processes at small scale that take place in the main control areas, i.e. the Camarinal and Espartell sills. Two-dimensional velocity and densit