Recently, there has been much interest in issuing subseasonal to seasonal (S2S) forecasts, although their skill is often debated. In addition to large systematic errors, ensemble systems are often overconfident, i.e. have incorrect information about the uncertainty of a particular forecast. Stochastic parameterization schemes are used routinely to remedy the problem of overconfidence, but also have the potential to reduce systematic model errors. 

Here, we study the impact of adding a stochastic parameterization scheme in coupled simulations with the climate model CESM.  Physical processes associated with S2S-predictability, like the Madden-Julian  Oscillation (MJO) and Northern Hemispheric blocking are analyzed. In the simulations with a stochastic parameterization scheme, the northward propagation of the MJO is captured better, leading to an improved MJO lifecycle. The impact on other atmospheric fields like precipitation and winds will be discussed. 

How to cite: Berner, J.: Benefits of stochastic parameterizations in subseasonal to seasonal (S2S) forecasts , EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-16321, https://doi.org/10.5194/egusphere-egu21-16321, 2021.

Stochastic subgrid-scale parametrizations aim to incorporate effects of unresolved processes in an effective model by sampling from a distribution usually described in terms of resolved modes. This is an active research area in climate, weather and ocean science where processes evolved in a wide range of spatial and temporal scales. In this study, we evaluate the performance of conditional generative adversarial network (GAN) in parametrizing subgrid-scale effects in a finite-difference discretization of stochastically forced Burgers equation. We define resolved modes as local spatial averages and deviations from these averages are the unresolved degrees of freedom. We train Wesserstein GAN (WGAN) conditioned on the resolved variables to learn the distribution of subgrid flux tendencies for resolved modes and, thus, represent the effect of unresolved scales. Resulting WGAN is then used in an effective model to reproduce the statistical features of resolved modes. We demonstrate that various stationary statistical quantities such as spectrum, moments, autocorrelation, etc. are well approximated by this effective model.

How to cite: Timofeyev, I. and Alcala, J.: Subgrid-scale parametrization of unresolved scales in forced Burgers equation using Generative Adversarial Networks (GAN), EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-6910, https://doi.org/10.5194/egusphere-egu21-6910, 2021.

Intrinsic ocean variability is essential for climate prediction because it is less sensitive to stochastic process, but it is very difficult to be identified due to internal climate variability. Here we use regional interactive ensemble applied on ocean-atmosphere interface (RIE-OA) to suppress atmosphere stochastic variability and to reveal intrinsic variability as well as to understand climate dynamic across multiple timescales. Five atmosphere general circulation models (AGCM) are coupled to an ocean general circulation model (OGCM) over the North Atlantic basin (20^oN to Denmark Strait and Greenland-Scotland ridge). The OGCM interacts with fluxes from a selected AGCM globally except over the North Atlantic basin where the OGCM interacts with the ensemble averaged fluxes from the five AGCMs. The five AGCMs, on the other hand, feel the same ocean states. Hence, the atmosphere stochastic variability impacting the ocean is one-fifth weaker than stand-alone configuration (control case). This leads to reduction of the local climate variability, such as Atlantic Multidecadal Variability, but should not reduce intrinsic variability. Comparing control cases and RIE-OA case, we found the intrinsic ocean variability, a narrow-banded low-frequency (about 8 to 20 years) signal over the North Atlantic Subtropical Gyre, is not influenced by the weakened stochastic variability. More details will be discussed.

How to cite: Shen, M.-L., Keenlyside, N., and Chiu, P.-G.: Understanding intrinsic ocean variability by suppressing regional stochastic variability, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-14346, https://doi.org/10.5194/egusphere-egu21-14346, 2021.

I will present a comprehensive inter-comparison of linear regression, stochastic and deep-learning-based models for reduced-order statistical modelling of the simplified ocean circulation. The reference dataset is provided by the top 150 empirical orthogonal functions (EOFs) and principal components (PCs) of an idealized, eddy-resolving, double-gyre ocean model. Our goal is to have a systematic and comprehensive assessment of the skills, costs and complexities of all the models considered.

The model based on linear regression is considered as a baseline. Additionally, we investigate stochastic models (linear regression plus additive-noise and a multi-level approach), deep-learning models (a feed-forward Artificial Neural Network (ANN), a Long Short Term Memory (LSTM)), and deep-learning augmented linear regression models (also called hybrid models). We also explored stochastically improved deep learning methods by adding spatially correlated white noise in the deep learning models to account for the residuals and left out variance in the discarded PCs. The assessment metrics considered are climatology, variance, RMSE, instantaneous correlation coefficients, frequency map, prediction horizon, and computational costs for training and predictions.  

Until now, we found that the hybrid LSTM models perform the best, followed by the multi-level linear stochastic model and multiplicative white noise model. Additionally, hybrid models found to perform better when augmented by spatially correlated white noise.  This suggests that an amalgam of physics, memory effects, and stochasticity provides the best strategy for low-order representation of oceanic process. However, LSTM was also found to be most expensive to train and forecast amongst all. Skills of simple stochastic models are similar to those of the linear regression model but superior to those of the pure deep learning models, as evidenced by relatively better frequency maps, infinite prediction horizon, and low running cost.

Overall, our analysis promotes multi-level stochastic methods, with memory effects, and stochastic hybrid methods for low-dimensional ocean models as a more practical option when compared to pure deep-learning solutions as they are more accurate, stable, and low-cost. Furthermore, this is an ongoing research project and more updated results will be discussed at the time of presentation.

How to cite: Agarwal, N., Kondrashov, D., Dueben, P., Ryzhov, E., and Berloff, P.: A comparison of data-driven approaches to build low-dimensional ocean models, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-10194, https://doi.org/10.5194/egusphere-egu21-10194, 2021.

Moist equatorial waves, responsible for a large fraction of synoptic and intraseasonal tropical variability, are visible in satellite observations of cloud top temperature and outgoing longwave radiation, with familiar dispersion relations appearing in space time spectra once a background red-like spectrum is removed (Kiladis et al., 2009).

Studies have suggested that the large-scale planetary waves in the equatorial region can be excited by smaller scale gravity waves (Yang & Ingersoll, 2013), baroclinic waves from the extratropics (Wedi & Smolarkiewicz, 2010), or localized synoptic scale heating (Gill 1980, 1982). In this study we examine the possibility that a continuous forcing of anomalies at the mesoscale can, through a turbulent upscale cascade, excites these waves, thus explaining both: the peaks along dispersion relations and the background red spectrum.

Our underlying assumption is that within the tropics (excluding wave forcing from the extratropics), a prerequisite to form coherent heating of ≈1000 km zonal length on an aqua planet is self-aggregation. Over the last two decades, self-aggregation has been studied over a wide range of scenarios up to the atmospheric mesoscale. In this study we examine the aggregation from the mesoscale up to the planetary scale, by applying mesoscale stochastic forcing in idealized spherical shallow water model. In particular, we examine the dependence of the large scale spectra on the field being stochastically forced, and on the existence of moisture.

We find that indeed, continuous stochastic forcing at the mesoscale can excite two-dimensional turbulence and linear tropical wave modes. When vorticity or moisture are forced in the simulations at wavenumber 100, a classical -5/3 slope of the eddy kinetic energy spectrum forms in an upscale energy cascade up to the planetary scale. Furthermore, equatorial waves emerge and Wheeler-Kiladis plots reveal a rich temporal and spatial structure of Rossby, Kelvin, Yanai, and Inertial Gravity waves. On the other hand, stochastic forcing of the divergence, or height fields only leads to a turbulent field when applied at planetary scales. Some explanations for this strong dependence on the type of forcing, and the role of moisture, will be discussed. 

How to cite: Schröttle, J., Suhas, D., Harnik, N., and Sukhatme, J.: Coexistence of equatorial waves & turbulence in moist shallow water simulations , EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-7722, https://doi.org/10.5194/egusphere-egu21-7722, 2021.

Atmosphere and ocean dynamics display many complex features and are characterized by a wide variety of processes and couplings across different timescales. Here we use Multivariate Empirical Mode Decomposition (MEMD; Rehman and Mandic, 2010) to investigate the multivariate and multiscale properties of a low-order model of the ocean-atmosphere coupled dynamics (Vannitsem, 2017). The MEMD allows us to decompose the original data into a series of oscillating patterns with time-dependent amplitude and phase by exploiting the local features of data and without any a priori assumptions on the decomposition basis. Moreover, each oscillating pattern, usually named Multivariate Intrinsic Mode Function (MIMF), can be used as a source of local (in terms of scale) fluctuations and information. This information allows us to derive multiscale measures when looking at the behavior of the generalized fractal dimensions at different scales (Hentschel and Procaccia, 1983) that can be seen as a sort of multivariate and multiscale generalized fractal dimensions (Alberti et al., 2020). With these two approaches, we demonstrate that the coupled ocean-atmosphere dynamics presents a rich variety of common features, although with a different nature of the fractal properties between the ocean and the atmosphere at different timescales. The MEMD results allow us to capture the main dynamics of the phase-space trajectory that can be used for reconstructing the skeleton of the phase-space dynamics, while the evaluation of the fractal dimensions at different timescales characterize the intrinsic complexity of oscillating patterns that can be related to the attractor properties. Our results support the interpretation of the coupled ocean-atmosphere dynamics as well as the investigation of general deterministic-chaotic dissipative dynamical systems in terms of invariant manifolds, bifurcations, as well as (strange) attractors in their phase-space, whose geometric and topological properties are a reflection of the dynamical regimes of the system at different scales. We compare the results obtained for the low-order dynamical model with those derived from the reanalysis data and demonstrate that a similar scale-dependent behavior is found, thus also confirming the suitability of the proposed system to model the ocean-atmosphere dynamics at different timescales and to describe topological and geometrical features of its phase-space.

References

Alberti, T., Consolini, G., Ditlevsen, P. D., Donner, R. V., Quattrociocchi, V. (2020). Multiscale measures of phase-space trajectories. Chaos 30, 123116.

Alberti, T., Donner, R. V., and Vannitsem, S. (2021). Multiscale fractal dimension analysis of a reduced order model of coupled ocean-atmosphere dynamics. Earth Syst. Dynam. Discuss. [preprint], https://doi.org/10.5194/esd-2020-96, in review.

Hentschel, H. G. E., Procaccia, I. (1983). The infinite number of generalized dimensions of fractals and strange attractors. Physica D 8, 435–444.

Rehman, N., Mandic, D. P. (2010). Multivariate empirical mode decomposition. Proceedings of the Royal Society A, 466, 1291–1302.

Vannitsem S., Predictability of large-scale atmospheric motions: Lyapunov exponents and error dynamics, Chaos, 27, 032101, 2017. 

How to cite: Alberti, T., Donner, R., and Vannitsem, S.: On the multiscale fractal features of a low-order coupled ocean-atmosphere model in comparison with reanalysis data, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-135, https://doi.org/10.5194/egusphere-egu21-135, 2021.

It has been well recognized that, for most climatic records, their current states are influenced by both past conditions and current dynamical excitations. However, how to properly use this idea to improve the climate predictive skills, is still an open question. In this study, we evaluated the decadal hindcast experiments of 11 models (participating in phase 5 of the Coupled Model Intercomparison Project, CMIP5) in simulating the effects of past conditions (memory part, M(t)) and the current dynamical excitations (non-memory part, ε(t)). Poor skills in simulating the memory part of surface air temperatures (SAT) are found in all the considered models. Over most regions of China, the CMIP5 models significantly overestimated the long-term memory (LTM) of SAT. While in the southwest, the LTM was significantly underestimated. After removing the biased memory part from the simulations using fractional integral statistical model (FISM), the remaining non-memory part, however, was found reasonably simulated in the multi-model means. On annual scale, there were high correlations between the simulated and the observed ε(t) over most regions of the country, and for most cases they had the same sign. These findings indicated that the current errors of dynamical models may be partly due to the unrealistic simulations of the impacts from the past. To improve predictive skills, a new strategy was thus suggested. As FISM is capable of extracting M(t) quantitatively, by combining FISM with dynamical models (which may produce reasonable estimations of ε(t)), improved climate predictions with the effects of past conditions properly considered may become possible.

How to cite: Xiong, F., Yuan, N., Ma, X., Lu, Z., and Gao, J.: On memory and non-memory parts of surface air temperatures over China: can they be simulated by decadal hindcast experiments in CMIP5?, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-3628, https://doi.org/10.5194/egusphere-egu21-3628, 2021.

One of the most used metrics to gauge the effects of climate change is the equilibrium climate sensitivity, defined as the long-term (equilibrium) temperature increase resulting from instantaneous doubling of atmospheric CO2. Since global climate models cannot be fully equilibrated in practice, extrapolation techniques are used to estimate the equilibrium state from transient warming simulations. Because of the abundance of climate feedbacks – spanning a wide range of temporal scales – it is hard to extract long-term behaviour from short-time series; predominantly used techniques are only capable of detecting the single most dominant eigenmode, thus hampering their ability to give accurate long-term estimates. Here, we present an extension to those methods by incorporating data from multiple observables in a multi-component linear regression model. This way, not only the dominant but also the next-dominant eigenmodes of the climate system are captured, leading to better long-term estimates from short, non-equilibrated time series.

How to cite: Bastiaansen, R., Dijkstra, H., and von der Heydt, A.: Multivariate Estimations of Equilibrium Climate Sensitivity from Short Transient Warming Simulations, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-187, https://doi.org/10.5194/egusphere-egu21-187, 2021.

How to cite: Lembo, V., Lucarini, V., and Ragone, F.: Predicting Climate Change through Response Operators in a Coupled General Circulation Model, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-2221, https://doi.org/10.5194/egusphere-egu21-2221, 2021.

Nonlinear atmospheric dynamics produce rare events that are hard to predict and attribute due to many interacting degrees of freedom. A sudden stratospheric warming is a spectacular example in which the winter polar vortex in the stratosphere rapidly breaks down, inducing a shift in midlatitude tropospheric weather patterns that persist for up to 2-3 months.  In principle, lengthy numerical simulations can be used to predict and understand these rare transitions.  For complex models, however, the cost of the direct numerical simulation approach is often prohibitive.  We describe an alternative approach which in principle only requires relatively short duration computer simulations of the system.  Applying this methodology to a classical idealized stratospheric model with stochastic forcing, we compute optimal forecasts of sudden warming events and quantify the limits of predictability.  Statistical analysis relates these optimal forecasts to a small number of easy-to-interpret physical variables.Remarkably, we are able to estimate these quantities using a data set of simulations much shorter than the return time of the warming event.

How to cite: Finkel, J., Webber, R. J., Gerber, E. P., Abbot, D. S., and Weare, J.: Forecasting rare stratospheric transitions using short simulations, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-16323, https://doi.org/10.5194/egusphere-egu21-16323, 2021.

In this work, we consider turbulence closures of LES (Large Eddy Simulation) type for classical decaying 2D turbulence in a priori and a posteriori experiments using explicit filtering approach. According to Germano 1986 decomposition, full subfilter stress for Gaussian filter is decomposed into three Galilean-invariant parts: Leonard, Cross and Reynolds stresses. By analysing spectral transfer of energy and enstrophy, we show that Leonard stress redistributes resolved energy toward large scales and dissipates substantial part of enstrophy, while Cross stress provides an additional enstrophy dissipation at subfilter scales and Reynolds stress predominantly injects energy into middle scales (i.e., Kinetic Energy Backscatter). Substantial part of enstrophy dissipation is located on subfilter scales, and it should be accounted for by choosing base filter wide enough compared to mesh step of LES model. Otherwise, significant fraction of enstrophy dissipation will correspond to subgrid scale stress, which is less universal and harder to approximate. As a result of a priori analysis, we propose LES closure consisting of three parts: SSM (Scale Similarity Model), which is equivalent to Leonard stress, biharmonic Smagorinsky damping as a Cross stress counterpart and ADM (Approximate Deconvolution Model) approximation for Reynolds stress ("backscatter"). The proposed model have two free parameters: Smagorinsky constant and amplitude of the backscatter. These parameters are estimated in a posteriori experiments utilizing dynamic approach and energy-enstrophy balance equation, correspondingly. The proposed model have the following distinctive features: it reproduces energy and enstrophy transfer spectra in accordance to the individual components of the subfilter forces, "reproduces" base filter and reproduces energy growth in accordance to the filtered DNS (Direct Numerical Simulation) solution.

The work was supported by the Russian Foundation for Basic Research (projects 19-35-90023, 18-05-60184) and Moscow Center for Fundamental and Applied Mathematics (agreement with the Ministry of Education and Science of the Russian Federation No. 075-15-2019-1624).

How to cite: Perezhogin, P. and Glazunov, A.: A priori and a posteriori analysis in Large eddy simulation of the two-dimensional decaying turbulence using Explicit filtering approach, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-2382, https://doi.org/10.5194/egusphere-egu21-2382, 2021.

Wetlands, affected by the hydro-climatic condition and human activities, are key elements in providing valuable ecosystem services for ecology, environment, and human. Wetlands can exist in various states (e.g., area, volume, depth, etc.) driven by both natural and human forcing, and are often distributed in a wetlandscape. In these specific landscapes, wetlands (node) and dispersal path (link) of inhabiting species organize ecological networks. Here, we generated the three ecological networks with three dispersal models (threshold distance, exponential kernel, and heavy-tailed dispersal model) and analyzed network characteristics (degree, efficiency and clustering coefficient) associated with the seasonal change of hydro-climatic condition on wetland hydrology. To identify the role of small wetlands, we analyzed two different scenarios in which the sum of wetland areas are similar but their area distributions are distinct. In the first scenario, most of the small wetlands are hydrologically disappeared while the second scenario maintains the small wetlands with a shrunk area of large wetlands. When the area of large wetlands was reduced, a slight decrease in the values of network metrics was observed due to an increase in distances between wetlands. On the other hand, when a number of small wetlands were hydrologically disappeared, all the metric values were significantly decreased compared to the network in which all wetlands were hydrologically maintained. Especially, when the disappeared wetlands were not recovered even after rainfall, possibly due to long-term dehydration of supporting soil, the network characteristics also did not recover even if the total area of wetlands were recovered. However, when the dried small wetlands were hydrologically recovered, the network characteristics also recovered rapidly. Based on our observation, we confirmed that the small wetlands, despite their extremely low areal portion in the entire wetlandscape, play a key role in maintaining the ecological network resilience. Our findings can be used for a decision-making process for wetland conservation and restoration by reflecting the functional importance of small wetlands with physical characteristics requirements such as wetland areas.

How to cite: Kim, B. and Park, J.: The role of small wetlands for resilient ecological networks in a wetlandscape , EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-13712, https://doi.org/10.5194/egusphere-egu21-13712, 2021.

Local properties of chaotic systems can be summarized by dynamical indicators, that describe the recurrences of all states in phase space. Faranda et al. (2017) defined such indicators with the local dimension (d, approximating the local number of degrees of freedom of the system) and the inverse of persistence (θ, approximating the time it takes to leave a local state). It has been conjectured that such indicators give access to the local predictability of systems. The aim of this study is to evaluate how the predictability of climate variables such as temperature and precipitation is related to dynamical properties of the atmospheric flow.

The predictability of a chaotic system can be evaluated through ensembles of simulations, with probability scores (e.g. Continuous Rank Probability Score, CRPS). In this work, we consider ensembles of climate forecasts with a stochastic weather generator (SWG) based on analogs of atmospheric circulation (Yiou and Déandréis, 2019). We are interested in relating predictability scores of European temperatures and precipitation, obtained with this SWG, and the local dynamical properties of the synoptic atmospheric circulation, obtained from the NCEP reanalysis. We show experimentally that the CRPS of local climate variables can be predicted from large-scale (d, \ θ) values of geopotential height fields, for time leads of 5 to 10 days. A practical application is that the predictability of local variables (in Europe) can be anticipated from large-scale dynamical quantities, which can help to dimension the size of ensemble forecasts.

References

Faranda, D., Messori, G., Yiou, P., 2017. Dynamical proxies of North Atlantic predictability and extremes. Sci. Rep. 7, 41278. https://doi.org/10.1038/srep41278

Caby, T. Extreme Value Theory for dynamical systems, with applications in climate and neuroscience. Mathematics [math]. Université de Toulon Sud; Universita dell’Insubria, 2019. English.tel-02473235v1

Yiou, P., Déandréis, C., 2019. Stochastic ensemble climate forecast with an analogue model. Geosci. Model Dev. 12, 723–734. https://doi.org/10.5194/gmd-12-723-2019

Acknowledgments

This project has received funding from the European Union’s Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant agreement No 813844.

How to cite: Krouma, M., Yiou, P., Faranda, D., Thao, S., and Déandréis, C.: Quantification of the relation between dynamical properties of meteorological variables and their predictability, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-9869, https://doi.org/10.5194/egusphere-egu21-9869, 2021.

One of the big challenges today is to appropriately describe heat waves, which are relevant due to their impact on human society. Common characteristics in mid-latitudes involve meanders of the westerly flow and concomitant large anticyclonic anomalies of the geopotential field. These anomalies form the so-called teleconnection patterns, and thus it is natural to ask how robust such structures are in various models and how much data we require to make statistically significant inferences. In addition, it is natural to ask what are the precursor phenomena that would improve forecasting capabilities of the heat waves. In particular, what kind of long term effect does the soil moisture have and how it compares to the respective quantitative contribution to the predictability of the teleconnection patterns.

In order to answer these questions we perform various types of regression on a climate model. We construct the composite maps of the geopotential height at 500 hPa and estimate return times of heatwaves of different severity. Of particular interest to us is a committor function, which is essentially a probability a heat wave occurs given the current state of the system. Committor functions can be efficiently computed using the analogue method, which involves learning a Markov chain that produces synthetic trajectories from the real trajectories. Alternatively they can be estimated using machine learning approach. Finally we compare the composite maps in real dynamics to the ones generated by the Markov chain and observe how well the rare events are sampled, for instance to allow extending the return time plots.

How to cite: Miloshevich, G., Lucente, D., Herbert, C., and Bouchet, F.: Drivers of midlatitude extreme heat waves revealed by analogues and machine learning, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-15642, https://doi.org/10.5194/egusphere-egu21-15642, 2021.

Although the lifecycle of hurricanes is well understood, it is a struggle to represent their dynamics in numerical models, under both present and future climates. We consider the atmospheric circulation as a chaotic dynamical system, and show that the formation of a hurricane corresponds to a reduction of the phase space of the atmospheric dynamics to a low-dimensional state. This behavior is typical of Bose-Einstein condensates. These are states of the matter where all particles have the same dynamical properties. For hurricanes, this corresponds to a "rotational mode" around the eye of the cyclone, with all air parcels effectively behaving as spins oriented in a single direction. This finding paves the way for new parametrisations when simulating hurricanes in numerical climate models.

How to cite: Faranda, D., Messori, G., Yiou, P., Thao, S., Pons, F., and Dubrulle, B.: Hurricane dynamics and rapid intensification via dynamical systems indicators, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-2592, https://doi.org/10.5194/egusphere-egu21-2592, 2021.

The solar wind is characterized by a multiscale dynamics showing features of chaos, turbulence, intermittency, and recurring large-scale patterns, pointing towards the existence of an underlying attractor. However, magnetic field and plasma parameters usually show different scaling regimes with different physical and dynamical properties. Here by using a recent and novel approach developed in the framework of dynamical systems  we investigate the multiscale instantaneous properties of solar wind magnetic field phase space by means of the evaluation of instantaneous dimension and stability. We show the existence of a break in the average attractor dimension occurring at the observed scaling break between the inertial and the dissipative regimes. We further show that sometimes the dynamics is higher dimensional (d>3) suggesting that the phase space is larger than that described by the system variables and invoking for an external forcing mechanism, together with the existence of at least one unstable fixed point that cannot be definitely associated with noise. Instantaneous properties of the attractor therefore provide an efficient way of evaluating dynamical properties and building up improved cascade models.

How to cite: Caby, T., Alberti, T., Faranda, D., Donner, R. V., Consolini, G., Vaienti, S., and Carbone, V.: Instantaneous multiscale properties of solar wind dynamical regimes and their attractors, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-8342, https://doi.org/10.5194/egusphere-egu21-8342, 2021.

We analyze recent trends in extreme precipitation in the Southwestern Alps and link these trends to changes in the atmospheric influences triggering extremes. We consider a high-resolution precipitation dataset of 1x1 km2 for the period 1958-2017. A robust method of trend estimation is considered, based on nonstationary extreme value distribution and a homogeneous neighborhood approach. The results show contrasting extreme precipitation trends depending on the season. Excluding autumn, the significant trends are mostly negative in the Mediterranean area, while the French Alps show more contrasted trends, in particular in winter with significant increasing extremes in the Western and Southern French Alps and decreasing extremes in the Northern French Alps and Swiss Valais. In autumn, most of Southern France shows significant increasing trends, with up to 100% increase in the 20-year return level between 1958 and 2017, while the Northern French Alps show decreasing extremes.
By comparing these trends to changes in the occurrence of the dominant weather patterns triggering the extremes, we show that part of the significant changes in extremes can be explained by changes in the dominant influences, particularly in the Mediterranean influenced region. We also show that part of the trends in extremes are explained by a shift in the seasonality of maxima. 

How to cite: Blanchet, J., Blanc, A., and Creutin, J.-D.: Explaining recent trends in extreme precipitation in the Southwestern Alps by changes in atmospheric influences , EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-14291, https://doi.org/10.5194/egusphere-egu21-14291, 2021.

Extreme analysis, via e.g., GEV, was developed to deal with univariate time series, and is very difficult to extend beyond that dimension. Here we explore a different method, the archetypal analysis, which focuses on multivariate extremes. The method seeks to approximate the convex hull in high-dimensional state space, by identifying corners representing "pure" types, i.e. archetypes. The method, encompasses, in particular, the virtues of EOFs and clustering. The method is presented with a new manifold-based optimization algorithm, and applied to a number of atmospheric fields, including SST and SLP gridded data. The application to SST, in particular, reveals important features related to SST extremes. The strengths and weaknesses of the method and possible future perspectives will be discussed.

How to cite: Hannachi, A. and Trendafilov, N.:          Towards Analysing multivariate weather/climate extremes, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-3288, https://doi.org/10.5194/egusphere-egu21-3288, 2021.

Reliable projections of extremes in near-surface air temperature (SAT) by climate models become more and more important as global warming is leading to significant increases in the hottest days and decreases in coldest nights around the world with considerable impacts on various sectors, such as agriculture, health and tourism.

Climate model evaluation has traditionally been performed by comparing summary statistics that are derived from simulated model output and corresponding observed quantities using, for instance, the root mean squared error (RMSE) or mean bias as also used in the model evaluation chapter of the fifth Assessment Report of the Intergovernmental Panel on Climate Change (IPCC AR5). Both RMSE and mean bias compare averages over time and/or space, ignoring the variability, or the uncertainty, in the underlying values. Particularly when interested in the evaluation of climate extremes, climate models should be evaluated by comparing the probability distribution of model output to the corresponding distribution of observed data.

To address this shortcoming, we use the integrated quadratic distance (IQD) to compare distributions of simulated indices to the corresponding distributions from a data product. The IQD is the proper divergence associated with the proper continuous ranked probability score (CRPS) as it fulfills essential decision-theoretic properties for ranking competing models and testing equality in performance, while also assessing the full distribution.

The IQD is applied to evaluate CMIP5 and CMIP6 simulations of monthly maximum (TXx) and minimum near-surface air temperature (TNn) over the data-dense regions Europe and North America against both observational and reanalysis datasets. There is not a notable difference between the model generations CMIP5 and CMIP6 when the model simulations are compared against the observational dataset HadEX2. However, the CMIP6 models show a better agreement with the reanalysis ERA5 than CMIP5 models, with a few exceptions. Overall, the climate models show higher skill when compared against ERA5 than when compared against HadEX2. While the model rankings vary with region, season and index, the model evaluation is robust against changes in the grid resolution considered in the analysis.

How to cite: Thorarinsdottir, T., Sillmann, J., Haugen, M., Gissibl, N., and Sandstad, M.: Evaluating temperature extremes in CMIP6 simulations using statistically proper evaluation methods, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-9722, https://doi.org/10.5194/egusphere-egu21-9722, 2021.

Previous detection and attribution analyses suggest that human-induced increases in greenhouse gases have contributed to observed changes in extreme precipitation. However, all previous detection and attribution studies of observed changes in extreme precipitation i) use station data that have been heavily processed via gridding, transformation, and spatial and temporal averaging or other dimension reduction approaches, as well as using climate models to estimate the responses to external forcing, ii) also use models to estimate the unforced natural variability of extreme precipitation. Both aspects reduce user confidence in detection and attribution results.

We use a novel detection and attribution analysis method that is applied directly to station data in the areas considered without prior processing and use climate models only to obtain estimates of the space-time pattern of extreme precipitation response to external forcing. We use records of the annual maximum one day (Rx1day) or five consecutive days (Rx5day) precipitation accumulations from 5,081 land-based stations spanning the period 1950-2014 with at least 45 years of coverage, including at least 3 years in the period 2010-2014. Expected responses to external forcings are estimated from ALL and NAT forcings simulations from large ensemble simulations performed with CanESM2 and from a multi-model ensemble that participated in phase 6 of the Coupled Model Intercomparison Project (CMIP6). Changes at these stations are evaluated by fitting non-stationary Generalized Extreme Value (GEV) distributions at individual stations to the logarithms of observed Rx1day and Rx5day values. Non-stationarity in the GEV distribution is permitted by using location parameters that are allowed to be linearly dependent on climate model-simulated responses in ln(Rx1day) and ln(Rx5day) to different forcings. We perform the detection and attribution analysis across different spatial scales, including the global, continental, and regional scales.

The influence of anthropogenic forcings on extreme precipitation is detected over the global land area, three continental regions (western Northern Hemisphere, western Eurasia, and eastern Eurasia), and many smaller IPCC regions, including C. North-America, E. Asia, E.C. Asia, E. Europe, E. North-America, N. Europe, and W. Siberia for Rx1day, and C. North-America, E. Europe, E. North-America, N. Europe, Russian-Arctic, and W. Siberia for Rx5day. Consistency between our study and previous studies substantially increases confidence in detection and attribution findings concerning extreme precipitation. The attributed effects of anthropogenic forcing on extreme precipitation include substantially decreased waiting times between extreme annual maximum events in regions where anthropogenic influence has been detected and intensification of extreme precipitation that, at a global scale is consistent with the Clausius-Clapeyron rate of about 7% per 1°C of warming.  

How to cite: Sun, Q., Zwiers, F., Zhang, X., and Yan, J.: Quantifying the human influence on the intensity of extreme 1- and 5-day precipitation amounts at global, continental, and regional scales, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-8471, https://doi.org/10.5194/egusphere-egu21-8471, 2021.

This talk will contrast how U.S. decision makers’ impacts-focused perspective on compound extreme events differs from climate science-based perspectives. Examples from around the U.S. will be provided, with an emphasis on cascading impacts that have spanned multiple regions and sectors. The talk will also propose a path forward for synthesizing ‘top-down’ and ‘bottom-up’ approaches to compound extremes, to facilitate adaptation. Time-permitting, preliminary findings from an analysis of sequential humid heat and extreme precipitation over the U.S. may be shown, as a guiding example. The work described reflects a collaboration of scientists funded by NOAA’s Regional Integrated Sciences and Assessments (RISA) program, charged with co-generating ‘useable science’ by working closely with stakeholders.   

How to cite: Horton, R., Keener, V., Kornhuber, K., Lesk, C., and Walsh, J.: Aligning compound extreme events as defined from climate science and sectoral impact perspectives, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-15761, https://doi.org/10.5194/egusphere-egu21-15761, 2021.

The effects of solar irradiance forcing on weather and climate extremes have received relatively less attention compared to the solar-induced changes in the mean climate. In this respect, here we investigate the possible impact of solar irradiance forcing on the Northern Hemisphere extreme weather and climate variability during summer, from a potential vorticity (PV) perspective. The generation of severe weather events in the extra-tropical regions is often related to intrusions of high PV originating from the polar lower stratosphere. Various two-dimensional PV indices, similar to those characterizing surface temperature and precipitation extremes, are defined to measure the frequency of upper level PV intrusion events. Based on long-term reanalysis data, we show that upper level high PV intrusions over Asia (Europe) are more (less) frequent during high relatively to low solar irradiance summers. Consistent with this PV pattern more (less) frequent surface extreme precipitation events are recorded during high relative to low solar irradiance summers in Asia (Europe). Patterns in the frequency of extreme temperatures are largely opposite to the corresponding extreme precipitation. Furthermore, extreme climate anomaly patterns associated with high solar irradiance forcing are similar to the corresponding patterns associated with strong monsoon circulation over Asia during summer. A preliminary analysis reveals the dominant role of upper level solar related PV anomalies in generation of extreme precipitation in the Asian monsoon region during high solar irradiance summers. A persistent blocking like circulation in the Caspian Sea region during low solar irradiance summers is associated more frequent high PV intrusions and extreme precipitation over Europe. The stability of the solar related extreme precipitation and temperature patterns in the last millennium perspective is also discussed based on proxy data as well as model simulations.

How to cite: Rimbu, N., Ionita, M., and Lohmann, G.: Solar forcing on the Northern Hemisphere weather and climate extremes during summer, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-9411, https://doi.org/10.5194/egusphere-egu21-9411, 2021.

To better assess the future risks associated with Intense Mediterranean Cyclones (IMC) a better understanding of their features, variability, frequency and intensity is required, including a robust detection method. The application of different detection algorithms provides results that are remarkably similar in some aspects but may be very different in others even using the same data. Thus, the selection of a particular method can significantly affect the results. For these reasons it is necessary to use different approaches and datasets to study the sensitivity and robustness of the detection approach. Those approaches often use minima in sea-level pressure (SLP) or extrema in relative vorticity or both to first identify the eye of the cyclone. SLP reflects the atmospheric mass distribution, and is representative of synoptic-scale atmospheric processes. On the other hand, the relative vorticity displays higher variability and is representative of the atmospheric circulation, being able to detect several local extrema (more than one centre), it can reduce uncertainties in the cyclone detection and tracking.

Therefore, within the framework of the EFIMERA project and to detect and track IMC we use a combination of different methods based on previous studies found in the literature. This new list of detected IMC events, together with the observed and well documented ones, are used here to create a new IMC database to be used for the study of their impacts and risk associated.

How to cite: Alvarez-Castro, M. C., Gallego, D., Ribera, P., Peña-Ortiz, C., and Faranda, D.: A new database for Intense Mediterranean Cyclones , EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-15827, https://doi.org/10.5194/egusphere-egu21-15827, 2021.

Previous studies have recognized the societal relevance of climatic extremes on the seasonal timescale and examined physical processes leading to individual high-impact extreme seasons (e.g., extremely wet or warm seasons). However, these findings have not yet been generalizing beyond individual case studies since at any specific location only very few seasonal events of such rarity occurred in the observational record. In this concept paper, a pragmatic approach to pool seasonal extremes across space is developed and applied to investigate hot summers and cold winters in ERA-Interim and the Community Earth System Model Large Ensemble (CESM-LENS). We identify spatial extreme season objects as contiguous regions of extreme seasonal mean temperatures based on statistical modeling. Regional pooling of extreme season objects in CESM-LENS then yields considerable samples of analogues to even the most extreme ERA-Interim events, which allow for climatological analyses of their statistical and physical characteristics.

This approach offers numerous opportunities for analyzing large samples of extreme seasons across regions, and several such analyses are illustrated exemplarily. We perform a climate model evaluation with regard to extreme season size and intensity measures and estimate how often an extreme winter like the cold North American 2013/14 winter is expected anywhere in mid-latitude regions. Moreover, we present a large set of simulated spatial analogues to this event, which allows to study commonalities and differences of their underlying physical processes. Finally, substantial but spatially varying climatological differences in the size of extreme summer and extreme winter objects are identified.

How to cite: Röthlisberger, M., Hermann, M., Frei, C., Lehner, F., Fischer, E. M., Knutti, R., and Wernli, H.: A new framework for identifying and investigating seasonal climate extremes, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-452, https://doi.org/10.5194/egusphere-egu21-452, 2021.

Geophysical fluid dynamics systems generally involve a wide range of spatio-temporal scales. Numerical representation can only simulate some of the scales. The others, at the unresolved scales of motion, must be parameterized for each type of phenomenon (wave, eddy, current), in terms of expected effects on the resolved scales. Most developments then assume that the fluid transport velocity has a time-uncorrelated noisy component with zero mean and stationary statistics. These approximations generally simplify theoretical descriptions, numerical simulations, data comparisons or more recently model error quantifications for data assimilation.

In the present work, we will discuss the applicability of such approximations through two examples: a surface oceanic current dynamics and swell refractions by surface currents.

When the time-decorrelation assumption is valid, we propose simple and tuning-free parametric models to represent the spatial correlations of the white-in-time small-scale velocity to help simulate the geophysical system of interest. These parametric models relies on turbulence space self-similarity and their statistical properties (e.g. spectral slope) can be easily estimated from observations of larger scale fluid velocities.

When the white-in-time approximation is not valid, we extend the previous parametric models to follow self-similarity properties in both time and space.

Numerical simulations will illustrate these theoretical developments along the presentation.

How to cite: resseguier, V., Hascoet, E., Chapron, B., and Fox-Kemper, B.: Validity domains and parametrizations for white-noise and multiscale models in turbulence and wave-turbulence interactions, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-692, https://doi.org/10.5194/egusphere-egu21-692, 2021.

Vertical motions are fundamental to atmospheric dynamics and our understanding of phenomena such as moist convection. A long-standing problem in atmospheric sciences is to understand the mesoscale energy spectra. Several numerical studies show that the vertical velocity spectrum has a homogeneous energy distribution across the mesoscales with a flat spectrum. Compared to the energy spectra of horizontal motion, the mechanisms that govern the spectrum of vertical velocity are less well known. In the troposphere, most of the horizontal mesoscale energy comes from divergent motions. At large scales O(100 km), vertical velocity relates, to a good approximation, to the vertically averaged divergence of horizontal motions by continuity in the incompressible limit. Recent measurements from NARVAL-2 (Next Generation Remote Sensing for Validation Studies) campaign conducted in the tropical Atlantic, unveiled that mesoscale horizontal mass divergence profiles possess a rich vertical structure and high spatio-temporal variability. Although the premise of a radiatively-balanced circulation holds on the long-term average, instantaneous deviations from this equilibrium occur in the form of wave-like oscillations. Numerical studies show that our state-of-the-art models can reproduce the observed variability in mesoscale divergence. We ask the following question in support of the previous arguments: What controls the spectrum of coherent mesoscale vertical motion? We aim to elucidate the mechanisms determining the homogeneous energy distribution across horizontal scales of vertical velocity spectra. This study designs numerical experiments, which include mechanisms-denial simulations employing the Icosahedral Nonhydrostatic (ICON) model. We conducted numerical simulations on a limited-area domain located in the western tropical Atlantic (4°S – 18°N, 64°W – 42°W). This domain has a horizontal resolution of 1.25 km and a lid at 30 km—the analysis period spans 48 hours. The experiments include the following: (i) a control run using DWD NWP physics configuration (ii) a dry atmosphere with all moist processes excluded along with the latent heat surface fluxes (iii) clouds invisible to radiation and, (iv) effects of saturation adjustment on temperature neglected while maintaining surface heat fluxes. Preliminary results show that the divergence profiles horizontally averaged over 200 km present a clear dominance of vertical wavelengths of 3 – 6 km. We found autocorrelation time-scales of around 4 – 6 hours increasing with altitude outside convective areas and consistent among all simulations. All experiments show a systematic decrease of about 50% in the temporal autocorrelation inside convective areas; therefore, moist convective processes modulate divergence's temporal variability. Moreover, we found that local moist processes contribute the most considerable fraction to the energy spectrum at scales < 200 km. The spectral response to moist processes is broad and extends into the free troposphere. The spectral response of surface fluxes instead is confined to the subcloud layer.

How to cite: Morfa-Avalos, Y.: Why is the tropical sky white? Numerical investigations to elucidate the shape of mesoscale vertical velocity spectra., EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-13215, https://doi.org/10.5194/egusphere-egu21-13215, 2021.

The ocean surface wind plays a crucial role in the air-sea exchanges of momentum, heat, and mass, consequently is vital to the controlling of weather and climate. Due to the extremely large range of scales of the motion of the wind field, e.g., flow structures from millimeters to thousands of kilometers, the multiscale dynamics are known to be relevant. In this work, with the help of a Wiener-Khinchine theorem-based Fourier power spectrum estimator, the scaling features of the wind field provided by several satellites, i.e., QuikSCAT, Metop-A, -B, and -C, Haiyang-2B, and China France Oceanography SATellite (CFOSAT), is examined. Power-law scaling behavior is evident in the ranges of 100 to 3000 km with a scaling exponent β varying from 5/3 to 3. The global distributions and seasonal variations of the scaling exponent β have also been considered. The results show that due to the energetic convective activities in the low-latitude zones, the scaling exponents β in these regions are closer to the value of 5/3. As for the mid-latitudes, the values of β are close to 2 and independent of the variation of longitude. Concerning the seasonal variations, for most regions, the scaling exponents measured in winter are larger than those in summer. Furthermore, the seasonal variations of β in low-latitudes are stronger than those in the mid-latitudes. Our preliminary results indicate that all satellites provide a consistent scaling feature of the ocean surface wind field.

How to cite: Gao, Y., Schmitt, F. G., Hu, J., and Huang, Y.: Scaling Feature of the Ocean Surface Wind Field, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-1717, https://doi.org/10.5194/egusphere-egu21-1717, 2021.

Planktonic copepods are tiny crustaceans, with a typical size of the order of mm, living in suspension in marine or freshwaters during their entire life cycle. They have swimming and jumping abilities and are known to be well adapted to their turbulent environment. Turbulence is known to increase their contact rate and feeding flux. However too intense turbulence is believed to have a negative effect so that a qualitative bell-shape is classically invoked to represent the contact rate of copepods versus turbulence intensity. In this framework, the objective of this work is to quantify the influence of ambient turbulence on copepod’s behavior, using trajectory analysis.

In this work, the motions of copepods were filmed using an infrared high-speed camera (1000 fps) in a turbulent environment, in the dark to avoid phototropism. The custom-made experimental set-up has been built-up in order to obtain in a central zone an isotropic and homogeneous turbulence representative of the natural environment. The flow was characterized with different tracer sizes at different turbulence intensities.

Copepods are filmed and the trajectories are extracted using signal processing routines. The instantaneous velocity, tangential and centripetal accelerations, and the local curvature are extracted for each trajectory. Their pdfs are computed, as well as different statistical moments: these indicators are studied at varying the turbulence intensity level (Reynolds number). Particles of different sizes (100 and 600 microns of mean diameters) and dead copepods are compared to living copepods statistics. This strategy allows to precisely characterize the copepods behavioral activity in relation with ambient turbulence. Ecological interpretations are drawn from the experimental results.

How to cite: Le Quiniou, C., Schmitt, F., Huang, Y., Calzavarini, E., and Souissi, S.: Copepods in turbulence: laboratory velocity and acceleration studies using high speed cameras, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-4860, https://doi.org/10.5194/egusphere-egu21-4860, 2021.

The surface circulation in the Gulf of Tonkin (GoT) was analyzed using 2.5 year-long dataset from the High-Frequency radar (from April 2014 to October 2016). High temporal resolution of the measurements and large coverage from HFR dataset enable us to characterize the variability of surface circulation in the GoT in a wide range of scales: from tidal to annual scale. A number of techniques of data including rotary spectral analysis (RSA), principal component analysis (PCA), harmonic tidal analysis, coherent analysis, etc. were used to identify the dominant modes of variability. The tidal motions, accounting for approximately 62% of the total variability, revealed the dominance of diurnal components (K1 and O1) with 4 times larger magnitude than that of semi-diurnal constituent (M2). At seasonal scale, the monsoon wind plays an important role in driving the surface circulation in the GoT. This was supported by a tight correlation (0.7) between the wind stress and current velocities and by a large contribution (more than 50%) of the Ekman-driven component to the total variability of currents in the offshore area. Along the shore, large seasonal variability of circulation was highlighted. During the year, the seaward extension of the coastal current is primarily controlled by the cross-shore wind stress while the flow intensity is modulated by the Red River discharge.

How to cite: Tran, M. C., Sentchev, A., and Nguyen, K. C.: Multi-scale variability of surface currents in the Gulf of Tonkin derived from HF radar observations, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-16284, https://doi.org/10.5194/egusphere-egu21-16284, 2021.

Lake Geneva is one of the largest bodies of water in western Europe and the largest one in the Alps region. Besides its obvious touristic importance it supplies drinking water for a large portion of Switzerland and work for hundreds of commercial fishermen. It has been under constant monitoring since the 1970's, for the impact of human activities on its water quality and biodiversity. The lake is known to be a warm monomictic lake, thermally stratified through most of the year with the exception of winter, when small thermal vertical gradients permit mixing from top to bottom. In lake Geneva, thermal stratification is one of the main environmental drivers of phytoplankton communities which are widely used as bioindicators for freshwater ecosystems. Studies on thermal stratification are thus essential to better predict phytoplankton seasonality and the development of harmful species blooms. In this work we examine more than 20 years of surveillance data from the INRAE (National Research Institute for Agriculture, Food and Environment) regarding temperature vertical profiles and meteorological data. We review both the climatology and the temperature stratification history of the lake and refine the temperature depth profiles obtaining the yearly progressions of the mixed layer depths. We then discuss the fitting of the depth profiles through the use of power-law and exponential functions, finding that in 66% of the cases the power-law better describes the experimental data, and we report the probability density function of the related statistics throughout the seasons.  Finally, we discuss the implications of our results for the modelling of the lake turbulent regime and phytoplankton seasonality.

How to cite: Beltram Tergolina, V., Jiang, Y., Schmitt, F., Berti, S., Calzavarini, E., and Anneville, O.: Power-law thermal stratification in lake Geneva and its seasonal evolution, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-8821, https://doi.org/10.5194/egusphere-egu21-8821, 2021.

Greenland Block Index (GBI) and North Atlantic Oscillation (NAO) are climate indices widely used for climatological studies especially over the Greenland Ice Sheet (GrIS). Particularly in summer, they are highly and negatively correlated; both have a strong relationship to near surface processes around the GrIS; their magnitude creates non-linear feedbacks and influences the low troposphere, shaping spatial accumulation and ablation patterns.

NAO is a measure of the surface pressure difference over the North Atlantic, providing insight of intensity and location of the jet stream. GBI denotes the general circulation over Greenland at the 500-hPa level and depending on its signal promotes heat and moist advection towards inland.

Based on the 1959-2019 period, the extreme summer melt of 2019 recorded the highest mean summer GBI while the extreme summer melt of 2012 recorded the lowest mean summer NAO. Their impact, however, goes beyond the melting season since the inter-seasonal phase change of these two indices may enhance/ postpone early melt/late refreezing and vice-versa.

Supported by 62 years of high-resolution regional climate model output (RACMO2.3p2), this work uses a statistical approach to analyze inter-seasonal variability of climate oscillations and their impact on the surface energy budget components over the GrIS. Also, teleconnection changes in a changing climate are hypothesized.

How to cite: Silva, T., Abermann, J., Shahi, S., Schöner, W., and Nöel, B.: Is the impact of climate oscillations changing over the Greenland Ice Sheet?, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-14893, https://doi.org/10.5194/egusphere-egu21-14893, 2021.

Pseudo-proxy experiments test the skill and sensitivity of two extended climate field reconstruction (CFR) methodologies in reconstructing northern North Atlantic Summer sea surface temperatures (SSTs). The Summer target data originate from one millennium-long simulation of the CESM LME (Otto-Bliesner et al. 2016) .  The experiments test the reconstruction skill systematically for input data mimicking SST marine proxies and instrumental observations, including characteristics such as sparse distribution in space, varying signal-to noise ratio and age uncertainties.

The Bayesian hierarchical model BARCAST assumes implicitly that the target variable is described as an AR(1) process in time (Tingley & Huybers 2010), while the proxy surrogate reconstruction (PSR) method makes no such assumption (Graham et al. 2007). The PSR selects climate analogues from the simulated instrumental period in our study.

Results show that both methodologies generate skillful reconstructions for perfectly dated input data, and the PSR is superior when realistic noise levels are chosen for the input data. When the input is perturbed with age-uncertainties, the methodologies are unable to generate acceptable skillful reconstructions. Facilitating in form of data clustering is tested for both methodologies in the attempt of improving reconstruction skill. This proves successful for the PSR methodology, with the best skill obtained using n=3 clusters over the reconstruction region.

Additionally, and addressed as a topic for discussion, we detect weak temporal persistence in the input data and the BARCAST reconstructions. The lack of SST persistence is found to be partly due to the input data sampling frequency: Summer means (June, July, August) averaged for every year. Analyses show that the simulated SST data exhibit weaker memory from one Summer to the next, compared to year-to-year variability based on annual means. Similar results are also found for instrumental observations. This finding stands in contrast to results of previous studies on terrestrial reconstruction, where climate reconstructions and individual proxy records exhibit strong persistence properties, also targeting the Summer season (Werner et al. 2018, Nilsen et al. 2018)^,.

References:

Graham, N.E. et al. (2007), Clim. Change, 83, 241-285, doi: 10.1007/s10584-007-9239-2

Otto-Bliesner, B.L. et al (2016), Bull. Amer. Meteor. Soc., 97, 735-754, doi: 10.1175/BAMS-D-14-00233.1

 Nilsen, T. et al. (2018), Clim. Past, 14, 947-967, doi: 10.5194/cp-14-947-2018

Tingley, M. P. and P. Huybers (2010a), J. Clim., 23, 2759–2781, doi: 10.1175/2009JCLI3015.1

Werner, J. P. et al. (2018), Clim. Past, 14, 527-557, doi: 10.5194/cp-14-527-2018

How to cite: Nilsen, T. and Talento, S.: Climate field reconstruction of North Atlantic sea surface temperatures, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-9865, https://doi.org/10.5194/egusphere-egu21-9865, 2021.

There is emergent evidence that abrupt shifts of the Atlantic Meridional Overturning Circulation (AMOC) have occurred during interglacial periods, with recent observations and model simulations showing that we may have over-estimated its stability during warm climates. In this study, we present a multi-proxy reconstruction of deep-water characteristics from the Rockall Trough in the Eastern North Atlantic to assess the variability of Nordic seas and Labrador Sea deep-water formation during past interglacial periods MIS 1, 5, 11, and 19. To test the warm climate stability hypothesis and to constrain the variability of deep-water formation for past warm climates, we performed geochemical analysis on planktic (Nd isotopes) and benthic foraminifera (δ¹⁸O and δ¹³C) along with sedimentological analysis. This approach allows us to reconstruct paleocurrent flow strength, as well as the origin and contribution of different water masses to one of the deep-water components of the AMOC in the Rockall Trough. We found that deep-water properties varied considerably during each of our chosen periods. For example during the Holocene εNd variability is smaller (1.8 per mil) when compared to variability during MIS 19 (3.3 per mil), an interglacial that experienced very similar orbital boundary conditions. Our results confirm that deep-water variability in the eastern North Atlantic basin was more variable in previous interglacial periods when compared to our current Holocene and provide new insight into the relative contribution of Nordic Seas Deep Water and Labrador Sea Water in the Rockall trough.

How to cite: Murphy O' Connor, M., Colin, C., and Morley, A.: Assessing the stability of the AMOC during past warm climates, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-15298, https://doi.org/10.5194/egusphere-egu21-15298, 2021.

One of the main contributors to palaeoclimate variability on millennial timescales is understood to be Dansgaard-Oeschger (D-O) cycles. Our awareness of these phenomena arises primarily from quasi-periodic, abrupt transitions of large magnitude detected in δ¹⁸O records from Greenland ice cores (e.g. Dansgaard et al, 1982; Johnsen et al, 1992), although there is evidence of similar variability in other archives and regions. D-O cycles have plenty to capture the imagination:

• the strength and rapidity of climate changes over Greenland,

• their regularity throughout MIS3 (~60 to 30 thousand years before present) and occurrence during the last deglaciation contrasting with their relative absence during the Last Glacial Maximum and Holocene,

• their opposed characteristics in Greenland and Antarctica,

• and that different models require different boundary conditions to reproduce this phenomena, if they can reproduce it at all.

This talk characterises D-Olike cycles in two different models: Planet Simulator (PlaSim, an Earth System Model with simplified atmospheric physics, thermodynamic sea ice, and simplified ocean dynamics), and COSMOS (a CMIP3-era ESM). We identify four phases to D-O cycles and commonalities and differences in their representations in these models. Finally, we examine which phases of this type of variability continue to contribute to climate variability today and what that looks like.

How to cite: Andres, H. and Tarasov, L.: Characterising Dansgaard-Oeschger cycles: from MIS3 to today, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-13778, https://doi.org/10.5194/egusphere-egu21-13778, 2021.

To describe earth’s former and predict the expected future climate in a general way we need to understand at least two basic characteristics of the distribution of earth’s temperature, its mean state and its temporal and spatial variance of temperature. There is some confidence in the projection of the mean state but the characteristics and changes of climate variability, especially on multi-decadal and longer time-scales are less known.

To characterize climate variability on these time scales, the instrumental record is too short. Climate proxies such as oxygen isotopes from foraminifera retrieved from marine sediments provide long records but do not exclusively carry information about the climate signal of interest. The decomposition of proxy time series into climate and non-climate components is challenging and depends on the adequate representation of the major involved biological and physical processes influencing the record. But even with a reasonable representation of the combined processes as fluctuations in proxy seasonality, bioturbation and errors in the age model, a proxy record still appears as the combination of these effects.

As a proxy record is only a single representation of this sum of effects we work on replicate measurements as a tool to characterize and separate the variability components. We therefore analysed oxygen isotopes and Mg/Ca in replicated measurements from the same sample, in replicated samples from the same sediment layer and in nearby sediment cores spanning the Holocene.  
If we compare two records the relation of them will determine the commonness of the underlaying processes. As records for example come from the same core or from cores of nearby located sites, they share the same climate signal. In the case they are from the same core they also share the errors in the age model and the time uncertainty introduced by bioturbation. Combining different types of replicates allows us the analyse the effect of different combinations of shared and independent errors.

The first two cores that we work on come from about 10 km apart located sites in the Indonesian Sea. GeoB 10054-4 was drilled in a water depths of 1076 meters, at longitude of 112°40.10’E and latitude 8°40.90’S and its average sedimentation rate was estimated as 20 cm/kyears. GeoB 100537 was drilled in a water depths of 1372 meters, at longitude 112°52.30’E and at latitude 8°40.56’S and its average sedimentation rate is estimated as 45cm/kyears.

In the presentations we will show first results of the analysis of intra core and inter core variability.

How to cite: Dyck, H., Laepple, T., Dolman, A., Groeneveld, J., and Mohtadi, M.: Separating Climate Variability and Non-Climate Noise in Proxy Records - A case study on replicate marine sediment records, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-13495, https://doi.org/10.5194/egusphere-egu21-13495, 2021.

Synoptic timescale forecasts over Equatorial Africa are important for averting weather-and climate-related disasters and the resulting agricultural losses. Observational studies have shown that rainfall anomalies often propagate eastward across Equatorial Africa, and that there is a linkage between synoptic-scale eastward-propagating precipitation and Convectively Coupled Kelvin Waves (CCKWs) over this region. We explore the mechanisms in which CCKWs modulate the propagation of precipitation from West to East over Equatorial Africa. We examine the first Africa-wide climate simulation from a convection permitting model (CP4A) along with its global driving-model simulation (G25) and evaluate both against observations (TRMM) and ERA-Interim (ERA-I), with a focus on precipitation and Kelvin wave activity.

Lagged composites show that both simulations capture the eastward propagating precipitation signal seen in observational studies, though G25 has a weaker signal. Composite analysis on high-amplitude Kelvin waves further shows that both simulations capture the connection between the eastward propagating precipitation anomalies and Kelvin waves. In comparison to TRMM, however, the precipitation signal is weaker in G25, while CP4A is more realistic. As the Kelvin wave activity is also well represented in both simulations when compared to ERA-I, the weak precipitation signal in G25 may be partly associated with the weak coupling between the precipitation and Kelvin waves. We show that CCKWs modulate the eastward propagation of convection and precipitation across Equatorial Africa through at least two related physical processes. Firstly, an enhancement of the low-level westerlies leads to increased low-level convergence; secondly, westerly moisture flux anomalies amplify lower-to-mid-tropospheric specific humidity. Results show that both CP4A and G25 generally simulate the key horizontal features of CCKWs, with anomalous low-level westerlies in phase with positive precipitation anomalies. However, both models show a weakness in capturing the vertical profile of anomalous specific humidity, and the zonal-vertical circulation is too weak in G25 and incoherent in CP4A compared to ERA-I.

In both ERA-I and the simulations, Kelvin wave-induced convergence and the westward tilt with height of anomalous zonal winds and specific humidity tends to weaken to the east of the East African highlands. It appears that these highlands impede the coherent eastward propagation of the wave-precipitation coupled structure.

How to cite: Ayesiga, G., Holloway, C., Williams, C., Yang, G.-Y., Stratton, R., and Roberts, M.: Linking Equatorial African precipitation to Kelvin wave processes in the CP4-Africa convection-permitting regional climate simulation, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-8419, https://doi.org/10.5194/egusphere-egu21-8419, 2021.

The North Atlantic Oscillation (NAO) is a high-frequency oscillation that has known influences on the climatology of weather patterns across the eastern United States. This study explores the relationship between the daily North Atlantic Oscillation index with observed high-wind events from 391 first-order weather stations across the eastern U.S. from 1973-2015. These events were determined following typical National Weather Service high-wind criteria: sustained winds of at least 18 m•s-1 for at least 1 hour or a wind gust of at least 26 m•s-1 for any duration. Since research literature shows high-wind events are often connected to parent mid-latitude cyclone tracks, and since the NAO has been shown to influence these storm tracks, it is hypothesized that changes in NAO phases are connected to spatial shifts and frequencies in high-wind observations. Initial results show a preferred southwesterly direction during each NAO phase. Variance in high-wind directions appears to increase (decrease) during negative (positive) NAO phases. Further, the greatest spatial difference in the mean center of high-wind observations was between positive and negative NAO phases. Overall, these preliminary findings indicate changes in high-wind observations may be linked to NAO phases.

How to cite: Stinnett, S., Durkee, J., Gilliland, J., Murley, V., Black, A., and Goodrich, G.: The Relationship between the North Atlantic Oscillation and High-Wind Observations Across the Eastern United States , EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-13792, https://doi.org/10.5194/egusphere-egu21-13792, 2021.

Earth's climate can be understood as a dynamical system that changes due to external forcing and internal couplings. It can be characterized from the evolution of essential climate variables, such as surface air temperature. Yet, the mechanisms, amplitudes, and spatiotemporal patterns of global and local temperature fluctuations around its mean, called temperature variability, are insufficiently understood. Discrepancies exist between temperature variability from model and paleoclimate data at the temporal scale of years to centuries and at the local scale, both of which are important socio-economic scales for long-term planning.
Here, we clarify whether global and local temperature signals from the last millennia show a stationary variance on these timescales and thus behave in a stable manner or not. Therefore, we contrast power spectral densities and their scaling behaviors using simulated, observed, and reconstructed temperatures on periods between 10 and 200 years. Despite careful consideration of possible spectral biases, we find that local temperatures from paleoclimate data tend to show unstable behavior, while simulated temperatures almost exclusively show stable behavior. Conversely, the global mean temperature tends to be stable. We explain this by introducing the gain as a powerful tool to analyze the forced temperature response, based on a novel estimate of the joint power spectrum of radiative forcing.
Our analysis identifies main deficiencies in the properties of temperature variability and offers new insights into the linkage between raditative forcing and temperature response, relevant to the understanding of Earth’s dynamics and the assessment of climate risks.

How to cite: Ellerhoff, B. and Rehfeld, K.: Timescale-dependent stability of surface air temperature and the forced temperature response, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-1939, https://doi.org/10.5194/egusphere-egu21-1939, 2021.

The highly successful Budyko-Sellers energy balance models are based on the classical continuum mechanics heat equation in two spatial dimensions. When extended to the third dimension using the correct conductive-radiative surface boundary conditions, we show that surface temperature anomalies obey the (nonclassical) Half-order energy balance equation (HEBE, with exponent H = ½) implying heat is stored in the subsurface with long memory. 

Empirically, we find that both internal variability and the forced response to external variability are compatible with H ≈ 0.4.  Although already close to the HEBE and classical continuum mechanics, we argue that an even more realistic “effective media” macroweather model is a generalization: the fractional heat equation (FHE) for long-time (e.g. monthly scale anomalies).  This model retains standard diffusive and advective heat transport but generalize the (temporal) storage term.  A consequence of the FHE is that the surface temperature obeys the Fractional EBE (FEBE), generalizing the HEBE to 0< H ≤1.  We show how the resulting FEBE can be been used for monthly and seasonal forecasts as well as for multidecadal climate projections.  We argue that it can also be used for understanding and modelling past climates.

How to cite: Lovejoy, S.: Budyko-Sellers 2.0: the classical and fractional heat equations, and the fractional energy balance equation, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-7877, https://doi.org/10.5194/egusphere-egu21-7877, 2021.

Natural climate variability is partially attributed to the solar radiative forcing. The scope of this work is to increase the scientific understanding of the relative role of solar variations on the terrestrial climate. The applied methodology examines initially the variation of multiple climatic parameters (temperature, zonal wind, relative and specific humidity, sensible and latent surface heat flux, cloud cover, precipitation) in response to the 11-year solar cycle. An additional goal is to estimate the response of the climate system’s parameters to the solar forcing in multiple forecasting horizons and to evaluate the behavior of the climate system in shorter time scales. The adopted methodology includes the development of linear regression models which calculate the dependency of the climatic parameters to solar variations for each grid point of the global dataset on a monthly time scale. The solar indicator used in this study is the 10.7-cm solar radio flux (F10.7) provided by NOAA, while the climate data are extracted from the NCEP/NCAR Reanalysis 1 project with a spatial resolution of 2.5^o X 2.5^o for 67 years. Regarding the climate system’s response forecasting, an Artificial Neural Network has been trained for modeling and forecasting the solar indicator time series for a few time steps in advance and the effect on climatic parameters is estimated using the established regression equations. The results exhibit that the variation of the climatic parameters can be partially attributed to the 11-year solar cycle. Statistically significant areas with relatively high solar cycle signal were found in multiple pressure levels and geographical regions. Furthermore, the results indicate that the identification of a clear solar signal in the climatic data is a difficult task due to the climate system’s complexity; advanced non-linear methods could be applied in order to obtain a more accurate understanding of this research field.

How to cite: Tzanis, C. G., Benetatos, C., and Philippopoulos, K.: The impact of solar variability on climatic parameters, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-3000, https://doi.org/10.5194/egusphere-egu21-3000, 2021.

What is the spatial scale of climate fluctuations, and how does this scale depend on the timescale under consideration? To answer this question, the spatio-temporal correlation structure of global surface temperature fields is characterized, for the period 1850-present, by estimating frequency spectra of the effective spatial degrees of freedom (ESDOF). These ESDOF spectra serve as a simple summarizing metric of the frequency-dependent spatial auto-correlation function. ESDOF spectra are estimated from: (a) the HadCRUT global gridded temperature anomaly dataset, based exclusively on instrumental measurements, and including detailed error variance estimates; (b) the NOAA 20th Century Reanalysis; and (c) a large ensemble of CMIP historical climate model simulations. When comparing (i) error corrected ESDOF spectra from the instrumental data to (ii) those obtained from the reanalysis and the model simulations, with HadCRUT data gaps imposed, results are found to be highly consistent among the three data sources. When the analysis is applied to the entire globe, the ESDOF spectra exhibit an almost uniform power-law frequency scaling with about 100 ESDOFs at monthly timescales and only about 2 ESDOFs at multidecadal timescales. Second-order differences in this scaling behaviour are found when the analysis is restricted to various spatial subdomains of the globe, namely, the tropics, extra-tropics, land areas, and ocean areas. A few implications of the diagnosed ESDOF reduction towards the longer timescales are briefly discussed.

How to cite: Kunz, T. and Laepple, T.: Power-law frequency scaling of surface temperature spatial degrees of freedom – estimated from instrumental data, reanalysis and climate model simulations, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-13589, https://doi.org/10.5194/egusphere-egu21-13589, 2021.

We present a user-friendly open-source Matlab package for stochastic data analysis. This package enables to perform a standard analysis of given timeseries like scaling analysis of structure functions and energy spectral density, estimation of correlation functions or investigation of the PDF’s of increments including Castaing fits. Also, this package can be used to extract the stochastic equations describing scale-dependent processes, such as the cascade process in turbulent flows, through Fokker-Planck equations and concepts of non-equilibrium stochastic thermodynamics. This stochastic treatment of scale-dependent processes has the potential for a new way to link to fluctuation theorems of non-equilibrium stochastic thermodynamics and extreme events (small scale intermittency, structures of rogue waves). 

The development of this user-friendly package greatly enhances the practicability and availability of this method, which allows a comprehensive statistical description in terms of the complexity of time series. It can also be used by researchers outside of the field of turbulence for the analysis of data with turbulent like complexity, including ocean gravity waves, stock prices and inertial particles in direct numerical simulations. Support is available: github.com/andre-fuchs-uni-oldenburg/OPEN FPE IFT, where questions can be posted and generally receive quick responses from the authors.

This package was developed by the research group Turbulence, Wind energy and Stochastics (TWiSt) at the Carl von Ossietzky University of Oldenburg. We acknowledge funding by Volkswagen Foundation. 

How to cite: Fuchs, A., Kharche, S., Wächter, M., and Peinke, J.: An open source Matlab package for stochastic data analysis in the context of Fokker-Plank Equation and Integral Fluctuation Theorem, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-9608, https://doi.org/10.5194/egusphere-egu21-9608, 2021.

Despite many airborne measurements and research campaigns our understanding of turbulence in free atmosphere is still far from sufficient. Part of the problem is the limited amount of measurement data, another part is measurement errors and last, but not least element is inadequate or not satisfactory data analysis. This presentation addresses some aspects of this last issue. The simplest way to characterize turbulence is to define/measure characteristic velocity U and length L (or time T) scales of turbulent  eddies. Two quantities necessary to estimate them are the turbulence kinetic energy K and the turbulence kinetic energy dissipation rate. A universal scaling relation between dissipation rate, turbulence kinetic energy and the turbulence length scale follows from the classical picture of the equilibrium Richardson-Kolmogorov cascade. There, the energy is transported between scales in a downward cascade until it is dissipated into heat by the smallest eddies. This universal scaling is a basis of many turbulence models, also in the context of atmospheric applications. However, a number of recent papers suggest that a universal, although different from the classical, scaling could also be observed in unsteady turbulent flows, which ate typical in the free atmosphere. In this study we investigate the nonequilibrium scaling relation between the integral length scale of turbulence and dissipation rate using velocity signals from various airborne measurements of atmospheric cloud turbulence, including that in and around convective clouds. A research aircraft measures 1D intersection of turbulent velocity field as a time series collected along the flight tajectory. Hence,  information on  temporal behavior (decay or development) of turbulence is not directly available. In this study we show how this important information can be recovered based on the observed scaling relations.

How to cite: Waclawczyk, M., Wójtowicz, J., and Malinowski, S.: Nonequilibrium scaling vs. the temporal evolution of turbulence, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-7572, https://doi.org/10.5194/egusphere-egu21-7572, 2021.

Extinction coefficient (σ_e) is a measure of light attenuation in the atmosphere, due to absorption and scattering properties of constituent gases and aerosols. In meteorological context, σ_e is used to understand transparency of the atmosphere, by estimating visibility or meteorological observable range (MOR). An accurate representation of visibility is required for safe functioning of various domains such as transport sectors, free optic communication, etc., and for understanding regional variations in air quality and climate. As the measurement of visibility is subjective and dependent on the instrument and range of measurement, here we attempt to characterize the same using extinction coefficient. σ_e was investigated under the framework of universal multifractals (UM), which is widely used to analyze and characterize geophysical fields that exhibit extreme variability over measurement scales.

For this study, σ_e was extracted from forward scattering visibility data by disdrometer (Campbell Scientific PWS100) located in the Paris area (France), operated by Hydrology, Meteorology, and Complexity laboratory of École des Ponts ParisTech (HM&Co, ENPC). As governing nonlinear equations of the atmosphere such as Navier-Stokes possess scale invariance, it was assumed here that the behavior of light attenuating particles should inherit similar scaling properties and hence be treated as multifractal fields. σ_e extracted from MOR measured at Paris-Charles de Gaulle airport was also subjected to multifractal analysis during the same time period for comparison. With direct analysis and simulations, it was found that σ_e exhibits multifratcal properties but are influenced by upper limit of visibility range in the instrument used for measurement. From the study, we suggest usage of extinction coefficient (σ_e) for characterizing atmospheric visibility as the former is a more physically relevant quantity which is objectively measured by instruments and directly related to particles in the atmosphere; while emphasizing the need to consider biases from instrumental range.

How to cite: Jose, J., Gires, A., Tchiguirinskaia, I., and Schertzer, D.: Multifractal analysis of extinction coefficient and its consequences in characterizing atmospheric visibility, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-11242, https://doi.org/10.5194/egusphere-egu21-11242, 2021.

Rainfall and wind are both known to exhibit extreme variability over a wide range of spatio-temporal scales which makes such fields complex to characterize, simulate and even measure. In this paper, we present a database that will enable to characterize the Interactions between rainfall and wind turbulence. Preliminary analysis using the framework of Universal Multifractals, commonly used to analyse and simulate these fields will also be presented.

The data collected during a high resolution measurement campaign on a meteorological mast will be used. More precisely the wind, temperature, pressure, humidity and rainfall fields are collected using 3D sonic anemometers (manufactured by Thies), mini meteorological stations (manufactured by Thies), and disdrometers (Parsivel2, manufactured by OTT). The latter gives access to the size and velocity of drops falling through its sampling area. The temporal resolution is of 100 Hz for the 3D sonic anemometers, 1 Hz for the meteorological stations and 30 s for the disdrometers. The devices are installed at two heights (approx. 45 m and 80 m), which enables to assess effects of altitude.

Initial results will be presented, notably with regards to the scaling features of the various fields, their characteristic parameters, and their correlation across scales.

Authors acknowledge the RW-Turb project (supported by the French National Research Agency - ANR-19-CE05-0022), for financial support. This project aims to quantify the impact of atmospheric turbulence and rainfall on wind power production.

How to cite: García Gago, Á., Schertzer, D., and Gires, A.: Interactions between rainfall and wind turbulence in a Universal Multifractal framework, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-10799, https://doi.org/10.5194/egusphere-egu21-10799, 2021.

Abstract

Hydrological applications such as flood design usually deal with and are driven by region-specific reference rainfall regulations, generally expressed as Intensity-Duration-Frequency (IDF) values. The meteorological module of hydro-meteorological models used in such applications should therefore be capable of simulating these reference rainfall scenarios. The multifractal cascade framework, since it incorporates physically realistic properties of rainfall processes such as non-homogeneity (intermittency), scale invariance, and extremal statistics, seems to be an appropriate choice for this purpose. Here we suggest a rather simple discrete-in-scale multifractal cascade based approach. Hourly rainfall time-series datasets (with lengths ranging from around 28 to 35 years) over six cities (Paris, Marseille, Strasbourg, Nantes, Lyon, and Lille) in France that are characterized by different climates and a six-minute rainfall time series dataset (with a length of around 15  years) over Paris were analyzed via spectral analysis and Trace Moment analysis to understand the scaling range over which the universal multifractal theory can be considered valid. Then the Double Trace Moment analysis was performed to estimate the universal multifractal parameters α,C₁ that are required by the multifractal cascade model for simulating rainfall. A renormalization technique that estimates suitable renormalization constants based on the IDF values of reference rainfall is used to simulate the reference rainfall scenarios. Although only purely temporal simulations are considered here, this approach could possibly be generalized to higher spatial dimensions as well.

Keywords

Multifractals, Non-linear geophysical systems, Cascade dynamics, Scaling, Hydrology, Stochastic rainfall simulations.

How to cite: Ramanathan, A., Versini, P.-A., Schertzer, D., Tchiguirinskaia, I., Perrin, R., and Sindt, L.: Simulating reference rainfall scenarios for hydrological applications using a multifractal approach, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-6605, https://doi.org/10.5194/egusphere-egu21-6605, 2021.

The urban population growth requires an improvement in the resilient behavior of these areas to extreme weather events, especially heavy rainfall. In this context, well-developed urban planning should address the problems of infrastructure, sanitation, and installation of communities, primarily related to insufficiently gauged locations. The main objectives of this study were to analyze the impacts of in-situ rain gauges’ distribution associated with the elaboration of a spatial diagnosis of the occurrence of floods in the municipality of Itaperuna, Rio de Janeiro – Brazil. The methodology consisted of the spatial analysis of rain gauges’ distribution with the help of the fractal dimension concept and investigation of flood susceptibility maps prepared by the municipality based on transitory factors (which consider precipitation in the modeling) and on permanent factors (natural flood susceptibility). Both maps were validated by the cross-tabulation method, crossing each predictive map with the recorded data of flood spots measured during a major rainfall event. The results pointed that the fractal analysis of the rain gauges’ distribution presented a scaling break behavior with a low fractal dimension at the small-scale range, mostly concerned in (semi-)urban catchments, highlighting the incapacity of the local instrumentation to capture the spatial rainfall variability. Thereafter, the cross-tabulation validation method indicated that the flood susceptibility map based on transitory factors presented an unsatisfactory probability of detection of floods when compared to the map based on permanent factors. These results allowed us to take into account the hydrological uncertainties concerning the insufficient gauge network and the impacts of the sparse distribution on the choice and elaboration of flood susceptibility maps that use rainfall data as input. Finally, we performed a spatial analysis to estimate the population and habitations that can be affected by floods using the flood susceptibility map based on permanent factors.

How to cite: Campos, P. C. D. O., Paz, I., Marques, M. E. S., Tchiguirinskaia, I., and Schertzer, D.: Fractal analysis of rain gauges’ inhomogeneous distribution and the associated hydrological impacts: the case study of Muriaé River basin, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-13856, https://doi.org/10.5194/egusphere-egu21-13856, 2021.

How to cite: Santos-Duarte-Ramos, V.: Precipitation nowcasting based on multifractal advection and deformation , EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-16526, https://doi.org/10.5194/egusphere-egu21-16526, 2021.

The general goal of the Fresnel platform of Ecole des Ponts ParisTech is to develop research and innovation on multiscale urban resilience. It is therefore conceived as a SaaS (Sofware as a Service) plaform providing data over a wide range of space-time scales and  appropriate softwares to analyse and simulate them over this range. 

The most recent development is the radar component RadX V3.0 that is now operational at https://radx.enpc.fr. It provides an easy access to various products based on precipitation measurement performed at the radial scale of 128 m by the ENPC polarimetric X-band radar. Using reliable and open source libraries it features a real-time radar display available to the general public and professionals who can freely access the precipitation data over a large part of Île-de-France region from their web browser (desktop and mobile). Another major component is the "analysis" section where scientists and managers  can define and select rainfall events in a interactive calendar and then analyse rainfall data throught different tools such as an interactive map with time control and dynamically genetared hyetograms.
For more refined spatial analysis registered users can also introduce their own shapefiles containing catchments and subcatchments, as well as to extract data and maps. They can also pinpoint a radar pixel to display hyetogram from a local area, this versatily on spatial and temporal selections allows for very precise analysis.

The application allows for different radar product analysis like DPSRI (Dual Polarization Surface Rainfall Intensity) and SRI (Surface Rainfall Intensity).These complementary products enhance the case studies analysis and give weather scientists more tools directly available from their web browser. 

Further software developments include high resolution hydrological modeling and a multifractal toolbox to estimate the scale invariant features of the precipitation and other fields (e.g. landuse). These developments are performed in close contacts and feedbacks from the scientific and professional world and they greatly benefit from the support of the Chair “Hydrology for Resilient Cities” (https://hmco.enpc.fr/portfolio-archive/chair-hydrology-for-resilient-cities/) endowed by the world leader industrial in water management and from previous EU framework programmes.

How to cite: Drouen, G., Schertzer, D., Monier, L., Willinger, B., and Tisserand, B.: The Fresnel Platform for Greater Paris: online tool to Dynamically Manage Multiscale Urban Resilience, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-12388, https://doi.org/10.5194/egusphere-egu21-12388, 2021.

In the process of producing grid data using observation data, the density of the stations were found to have the greatest influence on spatial (Hwang and Ham, 2013). Currently, the resolution of Korea’s ground detection network is about 12 to 15km additional stations need to be set up to improve spatial accuracy. However, indiscriminate installation of observatories is an objective challenge because of the enormous cost and the various factors to consider. It is important to select major observation points on an objective basis based on the existing KMA (Korea Meteorological Administration)'s AWS(Automatic Weather System), ASOS(Automated Synoptic Observing System)  data to increase the representative and reliability of the observation data. However, the establishment of an observatory so far has been chosen for subjective observation purposes, which may make it difficult to derive scientific data. In this study there is identified the long-term variability of urban meteorological data using the Hurst exponent (H) obtained through Rescaled range analysis (R/S analysis). And additional observation points are proposed for each meteorological element through network analysis.

R/S analysis is an analysis that measures the variability of time series by standardizing observations over time to make them in a dimensionless ratio and analyze the changes according to the length of the data used. H between 0 and 1 provides a criterion for distinguishing the measure of correlation that a time series has. H = 0.5 means that the present event does not affect subsequently, however the other values are correlated, not independent, and continuum of influence (Hwang and Cha 2004). The meteorological factors data were obtained from SK planet, AWS, ASOS installed in Seoul. As a result, long-term relativity between temperature and humidity are shown to be at a minimum of 0.750 and a maximum of 0.941.

Key words :  R/S analysis, Hurst exponent, long-term relativity

How to cite: kang, E.-B., Kang, D.-D., and Lee, D.-I.: Analysis of Long-term Variability through Temperature and Humidity Data in Urban Meteorological Observation Network, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-11434, https://doi.org/10.5194/egusphere-egu21-11434, 2021.

The need to adapt and increase the resilience of urban areas regarding the issues induced by urbanization and climate change effects (e.g. floods, Urban Heat Island, pandemics, etc.), has led to propose several strategies as the Natural-Based Solutions (NBS), which are focus on restoring natural processes such as infiltration and evapotranspiration (ET) in urban areas. In consequence, the increasing interest on NBS installation (highly supported by the H2020 program of the European Commission) by urban planners, decision-makers, researchers and the residing population has conducted to question the most efficient ways of NBS deployment. In this context, the urban dynamics (e.g. population density, land use patterns, transport network, etc.) and the distribution of green areas at different spatial scales play a key role that characterise the urban development in the territory.

Based on the study of an urban agglomeration named Est-Ensemble, located at the east of Paris (France), this research aims to: i) determinate the fractal dimension of the built-up and green areas by using 2 different box-counting methods; ii) set the potential areas to install NBS, through the development of an iterative downscaling scheme over the built-up structure with the software Fractalopolis, and following a polycentric approach inspired on the urban form of Ile de France Region, and iii) assess the population access to the nearest green spaces and deficit of green spaces.

Further, from local scale measurements of ET made close to Est-Ensemble agglomeration, the authors carried out a multifractal analysis of the ET data to better evaluate the observed scaling behaviour. This will be coupled with spatial approach developed above to evaluate the impact of temperature reduction of different land use scenarios. This research is partly supported by the French ANR EVNATURB project.

How to cite: Castellanos Diaz, L. A., Bonin, O., Versini, P. A., and Tchiguirinskaia, I.: Analysis of spatial dimensions and explicit multifractal modelling for the deployment of green areas in an urban agglomeration., EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-10732, https://doi.org/10.5194/egusphere-egu21-10732, 2021.

The fully distributed and scalable model Multi-Hydro (MH) enables a high resolution hydro-dynamic modelling of surface flow, infiltration, sewer flow and their interactions, including the retroaction of sewer overflows on surface. Its modular structure simplifies the introduction of new numerical engines, e.g., to simulate air quality or microclimate, to test implementation methodologies and/or develop user-friendly tools for urban management and design, going well beyond the flood control purposes.

Several extensions of MH were recently developed and greatly widened its functionalities. To give an example, they could be used for modelling and visualisation of climatic stress in the built environment and the resulting outdoor comfort, with an identification of cool corridors and climate safe paths.

By considering the high-resolution distributed rainfall, but also the layout of impervious and green areas on a range of representative streetscapes of a (semi-)urbanised catchment, this presentation addresses the questions on efficiency of additional ecosystem performances related to water availability (as cooling effect). Several scenarios were considered regarding possible adaption strategies, with a particular emphasis on  a multiscale analysis of :

• the confrontation between models and experimental data;
• the model induced structural choices and resulting limitations, in particular on the relevant space-time scales, as well their capacity to represent the extreme heterogeneity of the fields;
• the modelling of environmental variability across scales rather than at a given scale.

How to cite: Ribeiro-de-Moura, R.: Multiscale modelling of a (semi-)urbanised catchment with the help of the Multi-Hydro model, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-16534, https://doi.org/10.5194/egusphere-egu21-16534, 2021.

How to cite: Otranto-da-Silva, G.: Fractal tools to analyse the spatial variability of Blue Green Solutions in an urban area, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-16533, https://doi.org/10.5194/egusphere-egu21-16533, 2021.

In the last decades, Nature-Based Solutions (NBS) have become widely considered as a sustainable development strategy for the development of urban environments. Many previous studies only focused on the hydrological performances of NBS, whose economic impacts were not considered. Some studies considered both hydrological performances and economic costs to design cost-effective NBS scenarios, but only at a single catchment scale. Thus, a comprehensive investigation of NBS in terms of both hydrological performances and Life cycle costs (LCC) within the Universal Multifractal (UM) framework is significant for improving the multi-scale resilience of cities. In this study, the hydrological response of a 5.2 km² semi-urban watershed is investigated under various NBS scenarios and highly spatially variable rainfall events. First, the heterogeneous spatial NBS distribution in each scenario is quantified using their fractal dimension. Then, the hydrological responses are assessed with the help of the fully-distributed and physically-based model (Multi-Hydro) with a spatial resolution of 10 m. To evaluate the cost-effectiveness of NBS scenarios across scales, the statistical scale-independent “maximum probable singularity” γ_s, as defined in the UM framework, is combined with the economic indicator (LCC) to obtain the scale-independent cost-effectiveness (scale-independent CE) indicator for designing cost-effective NBS scenarios. The effective maximum singularity γ_max of each simulation is combined with LCC at different scales to obtain a scale-dependent cost-effectiveness (scale-dependent CE) indicator to be compared with the scale-independent CE. Results show that CEs obtained by both methods are strongly correlated, especially over the small-scale range. Therefore, the scale-independent CE based on UM framework is considered as an appropriate indicator to design NBS implementation at different scales.

Overall, this study presents a new approach for designing cost-effective NBS scenarios. This approach is based on the UM framework and enables to quantify the NBS scenario cost-effectiveness across a range of scales with the help of a scale-independent CE indicator. This approach can be efficiently applied to urban planning across various scales.

How to cite: Qiu, Y., Schertzer, D., Tchiguirinskaia, I., Monier, L., Willinger, B., and Tisserand, B.: A scale-independent cost-effective design of Nature-Based Solutions within a multifractal framework, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-7465, https://doi.org/10.5194/egusphere-egu21-7465, 2021.

The mapping of faults and fractures is a problem of high relevance in Earth Sciences. However, their identification in digital elevation models is a time-consuming task given the resulting networks' fractal nature. The effort is especially challenging in submarine environments, given their inaccessibility and difficulty in collecting direct observations. Here, we propose a semi-automated method for detecting faults in high-resolution gridded bathymetry data (~1 m horizontal and ~0.2 m vertical) of the Pescadero Basin in the southern Gulf of California, which were collected by MBARI's D. Allan B autonomous underwater vehicle. This problem is well suited to be explored by machine learning and deep-learning methods. The method learns from a model trained to recognize fault-line scarps based on key morphological attributes in the neighboring Alarcón Rise. We use the product of the mass diffusion coefficient with time, scarp height, and root-mean-square error as training attributes. The method consists of projecting the attributes from a three-dimensional space to a one-dimensional space in which normal probability density functions are generated to classify faults. The LDA implementation results in various cross-sectional profiles along the Pescadero Basin show that the proposed method can detect fault-line scarps of different sizes and degradation stages. Moreover, the method is robust to moderate amounts of noise (i.e., random topography and data collection artifacts) and correctly handles different fault dip angles. Experiments show that both isolated and linkage fault configurations are detected and tracked reliably.

How to cite: Vega Ramirez, L. A., Splez Madero, R. M., Contreras Perez, J., Caress, D., Clague, D. A., and Paduan, J. B.: A new Method for Fault-Scarp Detection Using Linear Discriminant Analysis (LDA) in High-Resolution Bathymetry Data From the Alarcón Rise and Pescadero Basin, Gulf of California., EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-417, https://doi.org/10.5194/egusphere-egu21-417, 2021.

Random forests is a supervised machine learning algorithm which has witnessed recently an exponential increase in its implementation in water resources. However, the existing implementations have been restricted in applications of Breiman’s (2001) original algorithm to regression and classification models, while numerous developments could be also useful for solving diverse practical problems. Here we popularize random forests for the practicing hydrologist and present alternative random forests based algorithms and related concepts and techniques, which are underappreciated in hydrology. We review random forests applications in water resources and provide guidelines for the full exploitation of the potential of the algorithm and its variants. Relevant implementations of random forests related software in the R programming language are also presented.

How to cite: Tyralis, H., Papacharalampous, G., and Langousis, A.: Random forests in water resources, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-2105, https://doi.org/10.5194/egusphere-egu21-2105, 2021.

Nowadays a wide range of methods and tools to study and forecast time series is available. An important problem in forecasting concerns embedding of time series, i.e. construction of a high dimensional space where forecasting problem is considered as a regression task. There are several basic linear and nonlinear approaches of constructing such space by defining an optimal delay vector using different theoretical concepts. Another way is to consider this space as an input feature space – IFS, and to apply machine learning feature selection (FS) algorithms to optimize IFS according to the problem under study (analysis, modelling or forecasting). Such approach is an empirical one: it is based on data and depends on the FS algorithms applied. In machine learning features are generally classified as relevant, redundant and irrelevant. It gives a reach possibility to perform advanced multivariate time series exploration and development of interpretable predictive models.

Therefore, in the present research different FS algorithms are used to analyze fundamental properties of time series from empirical point of view. Linear and nonlinear simulated time series are studied in detail to understand the advantages and drawbacks of the proposed approach. Real data case studies deal with air pollution and wind speed times series. Preliminary results are quite promising and more research is in progress.

How to cite: Kanevski, M.: Empirical analysis of time series using feature selection algorithms, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-6697, https://doi.org/10.5194/egusphere-egu21-6697, 2021.

Multi-dimensional arrays (also known as raster data, gridded data, or datacubes) are key, if not essential, in many science and engineering domains. In the case of Earth sciences, a significant amount of the data that is produced falls into the category of array data. That being said, the amount of data that is produced daily from this field is huge. This makes it hard for researchers to analyze and retrieve any valuable insight from it. 1-D sensor data, 2-D satellite imagery, 3-D x/y/t image time series and x/y/z subsurface voxel data, 4-D x/y/z/t atmospheric and ocean data often produce dozens of Terabytes of data every day, and the rate is only expected to increase in the future. In response, Array Databases systems were specifically designed and constructed to provide modeling, storage, and processing support for multi-dimensional arrays. They offer a declarative query language for flexible data retrieval and some, e.g., rasdaman, provide federation processing and standard-based query capabilities compliant with OGC standards such as WCS, WCPS, and WMS. However, despite these advances, the gap between efficient information retrieval and the actual application of this data remains very broad, especially in the domain of artificial intelligence AI and machine learning ML.

In this contribution, we present the state-of-art in performing ML through Array Databases. First, a motivating example is introduced from the Deep Rain Project which aims at enhancing rainfall prediction accuracy in mountainous areas by implementing ML code on top of an Array Database. Deep Rain also explores novel methods for training prediction models by implementing server-side ML processing inside the database. A brief introduction of the Array Database rasdaman that is used in this project is also provided featuring its standard-based query capabilities and scalable federation processing features that are required for rainfall data processing. Next, the workflow approach for ML and Array Databases that is employed in the Deep Rain project is described in detail listing the benefits of using an Array Database with declarative query language capabilities in the machine learning pipeline. A concrete use case will be used to illustrate step by step how these tools integrate. Next, an alternative approach will be presented where ML is done inside the Array Database using user-defined functions UDFs. Finally,  a detailed comparison between the UDF and workflow approach is presented explaining their challenges and benefits.

How to cite: Campos Escobar, O. J. and Baumann, P.: Towards AI in Array Databases, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-7409, https://doi.org/10.5194/egusphere-egu21-7409, 2021.

The quest to understand cause and effect relationships is at the basis of the scientific enterprise. In cases where the classical approach of controlled experimentation is not feasible, methods from the modern framework of causal discovery provide an alternative way to learn about cause and effect from observational, i.e., non-experimental data. Recent years have seen an increasing interest in these methods from various scientific fields, for example in the climate and Earth system sciences (where large scale experimentation is often infeasible) as well as machine learning and artificial intelligence (where models based on an understanding of cause and effect promise to be more robust under changing conditions.)

In this contribution we present the novel LPCMCI algorithm for learning the cause and effect relationships in multivariate time series. The algorithm is specifically adapted to several challenges that are prevalent in time series considered in the climate and Earth system sciences, for example strong autocorrelations, combinations of time lagged and contemporaneous causal relationships, as well as nonlinearities. It moreover allows for the existence of latent confounders, i.e., it allows for unobserved common causes. While this complication is faced in most realistic scenarios, especially when investigating a system as complex as Earth's climate system, it is nevertheless assumed away in many existing algorithms. We demonstrate applications of LPCMCI to examples from a climate context and compare its performance to competing methods.

Related reference:
Gerhardus, Andreas and Runge, Jakob (2020). High-recall causal discovery for autocorrelated time series with latent confounders. In Advances in Neural Information Processing Systems 33 pre-proceedings (NeurIPS 2020). 

How to cite: Gerhardus, A. and Runge, J.: LPCMCI: Causal Discovery in Time Series with Latent Confounders, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-8259, https://doi.org/10.5194/egusphere-egu21-8259, 2021.

The Earth’s climate is a highly complex and dynamical system. To better understand and robustly predict it, knowledge about its underlying dynamics and causal dependency structure is required. Since controlled experiments are infeasible in the climate system, observational data-driven approaches are needed. Observational causal inference is a very active research topic and a plethora of methods have been proposed. Each of these approaches comes with inherent strengths, weaknesses, and assumptions about the data generating process as well as further constraints.
In this work, we focus on the fundamental case of bivariate causal discovery, i.e., given two data samples X and Y the task is to detect whether X causes Y or Y causes X. We present a large-scale benchmark that represents combinations of various characteristics of data-generating processes and sample sizes. By comparing most of the current state-of-the-art methods, we aim to shed light onto the real-world performance of evaluated methods. Since we employ synthetic data, we are able to precisely control the data characteristics and can unveil the behavior of methods when their underlying assumptions are met or violated. Further, we give a comparison on a set of real-world data with known causal relations to complete our evaluation.

How to cite: Käding, C. and Runge, J.: A Benchmark for Bivariate Causal Discovery Methods, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-8584, https://doi.org/10.5194/egusphere-egu21-8584, 2021.

Surface winds — both wind speed and vector wind components — are fields of fundamental climatic importance. The character of surface winds greatly influences (and is influenced by) surface exchanges of momentum, energy, and matter. These wind fields are of interest in their own right, particularly concerning the characterization of wind power density and wind extremes. Surface winds are influenced by small-scale features such as local topography and thermal contrasts. That is why accurate high-resolution prediction of near‐surface wind fields is a topic of central interest in various fields of science and industry. Statistical downscaling is the way for inferring information on physical quantities at a local scale from available low‐resolution data. It is one of the ways to avoid costly high‐resolution simulations. Statistical downscaling connects variability of various scales using statistical prediction models. This approach is fundamentally data-driven and can only be applied in locations where observations have been taken for a sufficiently long time to establish the statistical relationship. Our study considered statistical downscaling of surface winds (both wind speed and vector wind components) in the North Atlantic. Deep learning methods are among the most outstanding examples of state‐of‐the‐art machine learning techniques that allow approximating sophisticated nonlinear functions. In our study, we applied various approaches involving artificial neural networks for statistical downscaling of near‐surface wind vector fields. We used ERA-Interim reanalysis as low-resolution data and RAS-NAAD dynamical downscaling product (14km grid resolution) as a high-resolution target. We compared statistical downscaling results to those obtained with bilinear/bicubic interpolation with respect to downscaling quality. We investigated how network complexity affects downscaling performance. We will demonstrate the preliminary results of the comparison and propose the outlook for further development of our methods.

This work was undertaken with financial support by the Russian Science Foundation grant № 17-77-20112-P.

How to cite: Rezvov, V., Krinitskiy, M., Gavrikov, A., and Gulev, S.: Comparison of AI-based approaches for statistical downscaling of surface wind fields in the North Atlantic, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-8844, https://doi.org/10.5194/egusphere-egu21-8844, 2021.

Recently, Unmanned Aerial Vehicles (UAVs) equipped with high-resolution imaging sensors have become a viable alternative for ecologists to conduct wildlife censuses, compared to foot surveys. They cause less disturbance by sensing remotely, they provide coverage of otherwise inaccessible areas, and their images can be reviewed and double-checked in controlled screening sessions. However, the amount of data they generate often makes this photo-interpretation stage prohibitively time-consuming.

In this work, we automate the detection process with deep learning [4]. We focus on counting coastal seabirds on sand islands off the West African coast, where species like the African Royal Tern are at the top of the food chain [5]. Monitoring their abundance provides invaluable insights into biodiversity in this area [7]. In a first step, we obtained orthomosaics from nadir-looking UAVs over six sand islands with 1cm resolution. We then fully labelled one of them with points for four seabird species, which required three weeks for five annotators to do and resulted in over 21,000 individuals. Next, we further labelled the other five orthomosaics, but in an incomplete manner; we aimed for a low number of only 200 points per species. These points, together with a few background polygons, served as training data for our ResNet-based [2] detection model. This low number of points required multiple strategies to obtain stable predictions, including curriculum learning [1] and post-processing by a Markov random field [6]. In the end, our model was able to accurately predict the 21,000 birds of the test image with 90% precision at 90% recall (Fig. 1) [3]. Furthermore, this model required a mere 4.5 hours from creating training data to the final prediction, which is a fraction of the three weeks needed for the manual labelling process. Inference time is only a few minutes, which makes the model scale favourably to many more islands. In sum, the combination of UAVs and machine learning-based detectors simultaneously provides census possibilities with unprecedentedly high accuracy and comparably minuscule execution time.

Fig. 1: Our model is able to predict over 21,000 birds in high-resolution UAV images in a fraction of time compared to weeks of manual labelling.

References

1. Bengio, Yoshua, et al. "Curriculum learning." Proceedings of the 26th annual international conference on machine learning. 2009.

2. He, Kaiming, et al. "Deep residual learning for image recognition." Proceedings of the IEEE conference on computer vision and pattern recognition. 2016.

3. Kellenberger, Benjamin, et al. “21,000 Birds in 4.5 Hours: Efficient Large-scale Seabird Detection with Machine Learning.” Remote Sensing in Ecology and Conservation. Under review.

4. LeCun, Yann, Yoshua Bengio, and Geoffrey Hinton. "Deep learning." nature 521.7553 (2015): 436-444.

5. Parsons, Matt, et al. "Seabirds as indicators of the marine environment." ICES Journal of Marine Science 65.8 (2008): 1520-1526.

6. Tuia, Devis, Michele Volpi, and Gabriele Moser. "Decision fusion with multiple spatial supports by conditional random fields." IEEE Transactions on Geoscience and Remote Sensing 56.6 (2018): 3277-3289.

7. Veen, Jan, Hanneke Dallmeijer, and Thor Veen. "Selecting piscivorous bird species for monitoring environmental change in the Banc d'Arguin, Mauritania." Ardea 106.1 (2018): 5-18.

How to cite: Kellenberger, B., Veen, T., Folmer, E., and Tuia, D.: Deep Learning Enhances the Detection of Breeding Birds in UAV Images, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-12157, https://doi.org/10.5194/egusphere-egu21-12157, 2021.

How to cite: Sudmanns, M., Augustin, H., van der Meer, L., Baraldi, A., and Tiede, D.: The Sen2Cube.at national semantic Earth observation data cube for Austria , EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-12486, https://doi.org/10.5194/egusphere-egu21-12486, 2021.

Total cloud score is a characteristic of weather conditions. At the moment, there are algorithms that automatically calculate cloudiness based on a photograph of the sky These algorithms do not know how to find the solar disk, so their work is not absolutely accurate.

To create an algorithm that solves this data, the data used, obtained as a result of sea research voyages, is used, which is marked up for training the neural network.

As a result of the work, an algorithm was obtained based on neural networks, based on a photograph of the sky, in order to determine the size and position of the solar disk, other algorithms can be used to work with images of the visible hemisphere of the sky.

How to cite: Borisov, M. and Krinitskiy, M.: Artificial neural networks in the problem of determining the position and size of the solar disk in wide-angle photographs of the sky, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-13051, https://doi.org/10.5194/egusphere-egu21-13051, 2021.

Within the framework of meteorology and oceanology, the importance of the cloud mass and the type of clouds cannot be underestimated. When describing and studying weather, precipitation and the movement of air masses over the ocean, the amount and type of clouds determines the flows of precipitation, their intensity, helps to predict the weather and the content of various impurities in the air, which makes the study of the properties of cloud cover one of the key aspects of meteorological and oceanological research.

The types of clouds are determined by the specialist, visually comparing the picture of the sky over the ocean with the guideline documents, the use of which reduces the possibility of the human factor affecting the determination of these parameters.

For an accurate study, study of the dynamics and dependence of climatic models on the conditions of cloud types, long-term measurements of the same type and the continuity of their methods are required. However, all these data are very unevenly distributed over the Earth's surface, and the number of ship observations is greatly reduced.

Thus, taking into account the importance of reliable determination of data related to cloudiness and the problems of their accuracy, the relevance and need to automate the determination of cloud types are obvious.

As a result of the work, an algorithm was obtained that allows classifying cloud types based on photographs taken during long-term sea expeditions.

How to cite: Veremev, N.: The use of artificial neural networks in the problem of classifying cloud types in wide-angle images of the visible hemisphere of the sky., EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-14470, https://doi.org/10.5194/egusphere-egu21-14470, 2021.

The use of artificial intelligence and specifically deep learning (DL) approaches in the domain of remote sensing is increasing. Such methods provide excellent results and show great potential for future applications. Earth observation sensors are able to deliver data with higher spatial, spectral and temporal resolutions. In this project, we use Sentinel-2 multispectral data and couple this input with a crowd annotated very high resolution (VHR) map which is generated in the video-game Cerberus, developed by the company BlackShore. In Cerberus, players are able to map features, like buildings, forest and specific types of crop fields, that are subsequently used as input for the Machine Learning (ML) pipeline. The ML pipeline is applied to classify crop fields in a larger region.

The main objective of this research is to study the accuracy of a model in detecting and describing the type of crop and whether the addition of a temporal dimension increases the accuracy. We will be experimenting with different methods that take their root in DL. The study region shown to Cerberus-players is Oromia in Ethiopia, south of the capital Addis Ababa. Using Sentinel-2 data, we aim to extend the generated maps to cover Ethiopia.

First, we will implement two DL methods; Random Forest (RF), and a 3D Convolutional Neural Network (CNN) that do not make use of the temporal dimension in order to have a baseline of the expected accuracy from a single multi-spectral image. Next, we will investigate four models that make use of time series: 1) a hybrid convolutional neural network-random forest (CNN-RF); 2) a 3D CNN that takes as input the output of a stack of 3D CNNs; 3) a model based on Recurrent Neural Networks (RNNs) performing pixel-based classification; and 4) an innovative method that combines the strength of RNNs, CNNs and Generative Adversarial Networks. 

We are now implementing the methods and shall report on results at EGU April 2021. For future research, it could be a very interesting case to study the possibility of generalizing the combined approach of crowd annotated training data with extended classification over larger regions and generalizing to other areas.

How to cite: Guillot, A., You, S., van't Woud, H., Perenboom, M., Kruijver, A., and Foing, B.: Crowd, Crops and Machines: How crowdsourced annotations can help towards crops classification, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-14553, https://doi.org/10.5194/egusphere-egu21-14553, 2021.

Remote sensing (RS) imagery, generated by e.g. cameras on satellites, airplanes and drones, has been used for a variety of applications such as environmental monitoring, detection of craters, monitoring temporal changes on planetary surfaces.

In recent years, researchers started applying Computer Vision [TP1] methods on RS data. This led to a steady development of remote sensing classification, providing good results on classification and segmentation tasks on RS data.  However, there are still problems with current approaches. Firstly, the main focus is on high-resolution RS imagery. Apart from the fact that these data are not accessible to everyone, the models fail to generalize on lower resolution data. Secondly, the models fail to generalize on more fine-grained classes. For example, models tend to generalize very well on detecting buildings in general, however they fail to distinguish if a building belongs to a fine-grained subclass like residential or commercial buildings. Fine-grained classes often appear very similar to each other, therefore, models have problems to distinguish between them. This problem occurs both in high-resolution and low-resolution RS imagery, however the drop in accuracy is much more significant when using lower resolution data.

For these reasons, we propose a Multi-Task Convolutional Neural Network (CNN) with three objective functions for segmentation of RS imagery. This model should be able to generalize on different resolutions and receive better accuracy than state-of the-art approaches, especially on fine-grained classes.

The model consists of two main components. The first component is a CNN that transforms the input image to a segmentation map. This module is optimized with a pixel-wise Cross-Entropy loss function between the segmentation map of the model and the ground truth annotations. If the input image is of lower resolution, this segmentation map will miss out on the complete structure of input images. The second component is another CNN to build a high-resolution image from the low-resolution input image in order to reconstruct fine-grained structure information. This module essentially guides the model to learn more fine-grained feature representations. The transformed image from this module will have much more details like sharper edges and better color. The second CNN module is optimized with a Mean-Squared-Error loss function between the original high-resolution image and the transformed image. Finally, the two images created by the model are then evaluated by a third objective function that aims to learn the distance of similarity between the segmented input image and the super-high resolution segmentation. The final objective function consists of a sum of the three objectives mentioned above. After the model is finished with training, the second module should be detached, meaning high-resolution imagery is only needed during the training phase.

At the moment we are implementing the model. Afterwards, we will benchmark the model against current state of the art approaches. The status will be presented at EGU 2021.­

How to cite: Zamarialai, S., Perenboom, T., Kruijver, A., Shi, Z., and Foing, B.: Monitoring Temporal Developments from Remote Sensing Data using AI Fine-Grained Segmentation, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-15917, https://doi.org/10.5194/egusphere-egu21-15917, 2021.

In order to make predictions on how the climate would respond to changes in global and regional emissions, we typically run simulations on Global Climate Models (GCMs) with perturbed emissions or concentration fields. These simulations are highly expensive and often require the availability of high-performance computers. Machine Learning (ML) can provide an alternative approach to estimating climate response to various emissions quickly and cheaply. 

We will present a Gaussian process emulator capable of predicting the global map of temperature response to different types of emissions (both greenhouse gases and aerosol pollutants), trained on a carefully designed set of simulations from a GCM. This particular work involves making short-term predictions on 5 year timescales but can be linked to an emulator from previous work that predicts on decadal timescales. We can also examine uncertainties associated with predictions to find out where where the method could benefit from increased training data. This is a particularly useful asset when constructing emulators for complex models, such as GCMs, where obtaining training runs is costly. 

How to cite: Mansfield, L., Nowack, P., and Voulgarakis, A.: Predicting climate model response to changing emissions, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-15944, https://doi.org/10.5194/egusphere-egu21-15944, 2021.

When working with ungridded remote sensing data, such as swath surface reflectance like Moderate Resolution Imaging Spectroradiometer (MODIS) MOD09 or Visible Infrared Imaging Radiometer Suite (VIIRS) VNP09, extracting targeted information of interest from a collection of granules can be a challenging exercise. Given a region of interest (ROI), it is tedious both to determine the subset of granules that intersect the ROI, as well as identifying, within the granules, the individual instantaneous field of views (IFOVs) contained by the ROI.

The SpatioTemporal Adaptive-Resolution Encoding (STARE) is an indexing scheme that recursively divides the Earth's surface into quadtree hierarchies, allowing triangular elements ("trixels") of varying sizes (resolutions) to be identified with unique index values. STARE is also a software library that operates on STARE indices. It can efficiently determine the spatial relationship between two trixels, by evaluating their index values, if the trixels share a common path in the STARE tree structure. By representing geographical regions as the sets of trixels with adaptive resolutions that tesselating them, STARE provides an elegant method to determine geospatial coincidence of arbitrarily shaped geographic regions, with accuracy up to ~7-8 cm in length. 

In this presentation, we introduce STARELite, a SQLite STARE extension and its use for cataloguing volumes of remote sensing granules that researchers often possess in their local storage. In this application, STARELite is used to determine subsets of granules intersecting arbitrary ROIs. Further, STARELite can be used for the inverse search problem: Determining all spatially coincident granules of an individual granule. STARELite leverages other components of the STARE ecosystem; namely STARE sidecars, which hold the trixel index values of each iFOV and a set of trixels representing the cover of each granule; STAREMaster, which is used to generate STARE sidecar files; and STARPandas, a Python Pandas extension used to bootstrap STARELite databases.

Given the limitations of SQLite, STARELite is to be understood as a proof of concept for the integration of STARE into relational databases in general. 

How to cite: Griessbaum, N., Rilee, M., Frew, J., and Kuo, K.-S.: Integrating STARE with relational databases, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-16410, https://doi.org/10.5194/egusphere-egu21-16410, 2021.

El Niño Southern Oscillation (ENSO) index has been shown as a non-Gaussian and nonlinear stochastic process. Here we assess the statistical significance of non-Gaussianity and non-linearity through the analysis of third-order statistics of El Niño 3.4 index in the period 1870–2018, namely the bicovariance (lagged third-order moments) and bispectrum (its 2D Fourier transform). The analysis of bicovariance reveals a tendency for extreme (weak) ENSO signal in the Boreal Spring to be followed by la Niñas (El Niños) in the forthcoming Boreal Winter, thus contributing for a nonlinear attenuation of the ENSO Spring Predictability Barrier. The bispectrum provides a spectral decomposition of skewness in a similar way of the spectral decomposition of variance.  Positive and negative real bispectrum values identify triadic phase synchronizations (at frequencies f1, f2 and f1+f2, mostly in the period range 2–6 years) contributing respectively to extreme El Niños and La Niñas. The known positive ENSO skewness and the main features of the ENSO bicovariance and bispectrum are shown to be well reproduced by fitting a bilinear stochastic model where the influence of non-observed variables is simulated by a delayed multiplicative noise, being able to generate non-Gaussianity and non-linearity. The model shows improved forecasts, with respect to benchmark linear models, up to four trimesters ahead, especially of the amplitude of extreme El Niños. The authors would like to acknowledge MISU (Meteorological Institute at Stockholm University) and the financial support FCT through project   UIDB/50019/2020 – IDL and project JPIOCEANS/0001/2019 (ROADMAP: ’The Role of ocean dynamics and Ocean–Atmosphere interactions in Driving cliMAte variations and future Projections of impact–relevant extreme events’).

How to cite: Pires, C. and Hannach, A.: Bicovariance and Bispectrum of ENSO index and its impact in nonlinear predictability , EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-8439, https://doi.org/10.5194/egusphere-egu21-8439, 2021.

The Madden-Julian Oscillation (MJO) is the dominant mode in the tropical atmosphere on sub-seasonal time scales, with a strong influence on the tropical weather and impacts on higher latitudes. Although it is an in depth-studied phenomenon, its intensification and attenuation mechanisms are not fully understood. The purpose of this communication is to analyse the statistics of MJO events using the Wheeler and Hendon index.
In our framework an MJO event takes place when the amplitude of the index is above a threshold for a certain number of days, depending on the averaging of the signal. With this, we define the maximum amplitude of an event, its duration and its size which is the sum of the amplitudes along the duration of an event.
We then analyse how the statistical properties change under variations in the definition of events. We further explore whether the tails of the event distributions are heavy tailed. As MJO interacts with other phenomena and has impacts on higher latitudes we compare the statistics of the MJO with other atmospheric indices. These statistical analyses may contribute to the knowledge of the intensification and attenuation processes that constitute the basic dynamics of the MJO.

How to cite: Minjares, M., Barreiro, M., and Corral, Á.: Statistical Analysis of Madden-Julian Events Using Time Series Indices., EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-13469, https://doi.org/10.5194/egusphere-egu21-13469, 2021.

The nature of the climate system is very complex: a network of mutual interactions between ocean and atmosphere lead to a multitude of overlapping geophysical processes. As a consequence, the same process has often a signature on different climate variables but with spatial and temporal shifts. Orthogonal decompositions, such as Canonical Correlation Analysis (CCA), of geophysical data fields allow to filter out common dominant patterns between two different variables by maximizing cross-correlation. In general, however, CCA suffers from (i) the orthogonality constraint, which tends to produce unphysical patterns, and (ii) the use of direct correlations, which leads to signals that are merely shifted in time being considered as distinct patterns.

In this work, we propose an extension of CCA, complex rotated CCA (crCCA), to address both limitations. First, we generate complex signals by using the Hilbert transforms. To reduce the spatial leakage inherent in Hilbert transforms, we extend the time series using the Theta model, thus creating an anti-leakage buffer space. We then perform the orthogonal decomposition in complex space, allowing us to detect out-of-phase signals. Subsequent Varimax rotation removes the orthogonal constraints to allow more geophysically meaningful modes.

We applied crCCA to a pair of variables expected to be coupled: Pacific sea surface temperature and continental precipitation. We show that crCCA successfully captures the temporally and spatially complex modes of (i) seasonal cycle, (ii) canonical ENSO, and (iii) ENSO Modoki, in a compact manner that allows an easy geophysical interpretation. The proposed method has the potential to be useful especially, but not limited to, studies on the prediction of continental precipitation by other climate variables. An implementation of the method is readily available as a Python package.

How to cite: Rieger, N., Corral, A., Turiel, A., and Olmedo, E.: Complex rotated CCA: a method to correlate lagged geophysical variables, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-9695, https://doi.org/10.5194/egusphere-egu21-9695, 2021.