The abrupt transient centennial to millennial coolings of the North Atlantic region first described in Greenland ice core isotope records by Hans Oeschger and Willy Dansgaard are now widely recognized as recurrent, abrupt weakenings of the Atlantic meridional overturning circulation (AMOC). A hierarchy of climate models has shown that freshening in the North Atlantic can trigger AMOC collapse or condition unforced AMOC variability.  Yet, the existence of a freshwater trigger is largely untested for most events, and it is also uncertain whether a reduction in freshwater fluxes was necessary to permit the recovery of AMOC.  Decadally resolved stalagmite oxygen isotope records from coastal caves in NW Iberia record changes in the δ¹⁸O of the surface eastern North Atlantic which are highly sensitive to the freshwater balance.   Careful refinement of stalagmite carbon isotope proxies to correct for in-cave fractionation effects, has provided an indicator of the abrupt temperature changes caused by AMOC in this region.  These dual indicators resolve the detailed phasing of freshening and past AMOC variations. In stalagmite records spanning late MIS 7 and MIS 6, nearly half of the abrupt coolings have no significant freshening event within a hundred years of the onset of cooling.  In contrast, the millennial coolings at the start of Termination II deglaciation are synchronous with abrupt freshening events.  Thus, triggers of abrupt AMOC weakenings and recoveries may be diverse, and the sensitivity to different triggers may change with the evolution of climatic boundary conditions.  A longer set of stalagmite records provide evidence for more dynamic variations of Northern Hemisphere ice sheets and a succession of positive and negative feedbacks on ice melting rate during deglaciations. 

The recognition of past variation in atmospheric CO₂ in ice core archives, first characterized by Hans Oeschger, catalyzed efforts to estimate past atmospheric CO₂ from indirect proxies in the deeper past, to better estimate the earth system and cryosphere sensitivity to atmospheric CO₂ in warmer than preindustrial periods.  Over the last decades, the challenging path of deriving proxy CO₂ records has led to numerous paradoxes in the relationship between climate variables and estimated CO_2.  From the marine alkenone-based CO₂ proxy, a better representation of the biological processes imprinted on the proxy has improved the relationship between CO₂ and climate trends since the mid Miocene.  Such a representation allows more robust estimates of the scope of changes in atmospheric CO₂ across climate transitions, although estimates of absolute CO₂ still feature high uncertainty.  As orbital resolution CO₂ proxy records emerge for past warmer time periods, we are presented with new questions about the relationship between ice growth on Antarctica and CO₂, and the carbon cycle processes shifting carbon into and out of the atmosphere on orbital timescales.

How to cite: Stoll, H.: Clarifying past AMOC and climate sensitivity with paleoclimate archives, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-4861, https://doi.org/10.5194/egusphere-egu25-4861, 2025.

Stable water isotopes (δ¹⁸O) in precipitation are one of the most abundant paleoclimate proxies and have been used to infer temperature changes at high latitude and hydrological changes in the tropics. In spite of much progress, however, fundamental questions on the paleoclimate interpretation of stable water isotopes still remain open. Combing water isotope observations and an isotope-enabled TRAnsient ClimatE simulation of the last 21,000 years (iTREACE-21), I will discuss some recent progresses towards the understanding of paleoclimatic inferences of  δ¹⁸O.

I will first discuss the δ¹⁸O for the pan-Asian monsoon region. We show that the widespread δ¹⁸O variability that is coherent over the Asian monsoon continental region is accompanied by a coherent hydroclimate footprint, with spatially opposite signs in rainfall. This footprint is generated as a dynamically coherent response of the Asian monsoon system to meltwater forcing and insolation forcing, reinforced by atmospheric teleconnections. As such, a widespread δ¹⁸O depletion in the Asian monsoon region is accompanied by a northward migration of the westerly jet and enhanced southwesterly monsoon wind, as well as increased rainfall from South Asia to northern China, but decreased rainfall in southern China. 

I will then discuss the temperature effect of polar ice core δ¹⁸O, quantitatively, in a new framework called the Unified Slope Equations (USE) that illustrates the general relationship between spatial and temporal δ¹⁸O-temperature slopes. The application of USE to the Antarctica in model simulations and observations shows that the comparable Antarctica-mean spatial slope with deglacial temporal slope in δ¹⁸O-surface temperature is caused accidentally by the compensation responses between the δ¹⁸O-inversion layer temperature relation and the inversion layer temperature itself.  This finding further leads us to propose a paleothermometer that is more accurate and robust than the spatial slope as the present day seasonal slope of -inversion layer temperature, suggesting the possibility of reconstructing past polar temperature changes using present observations.

I will finally discuss the climate interpretation of tropical alpine ice core δ¹⁸O by combining proxy records with climate models, modern satellite measurements and radiative-convective equilibrium theory. I show that the tropical ice core δ¹⁸O is an indicator of the temperature of the middle and upper troposphere, with a glacial cooling of ~7oC . Furthermore, it severs as a Goldilocks indicator of global mean surface temperature change, providing the first estimate of glacial stage cooling that is independent of marine proxies as ~6oC .

How to cite: Liu, Z.: Understanding Paleoclimatic Inference of Stable Water Isotopes using iTRACE Simulation, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-1813, https://doi.org/10.5194/egusphere-egu25-1813, 2025.

Accurately modeling emerging physical climate risks to natural and societal systems—such as global supply chains, the food system, health, and critical infrastructures—is essential for effective preparedness and honest discussions about the consequences of rising greenhouse gas emissions.

A series of anomalous weather events that shattered previous records by wide margins has —yet again—highlighted the need for an improved understanding of the physical processes behind weather and climate extremes, their statistical characteristics, and our ability to project them under future emission scenarios using climate models.

In this Award lecture, I will present an overview of recent studies and preliminary findings that explore the mechanisms and physical drivers of high-impact climate extremes, as well as their statistical characteristics, such as simultaneous or sequential occurrences, which can lead to high societal impacts under current and future climate conditions and will reflect on our capacity to reproduce such events in climate models.

How to cite: Kornhuber, K.: Physical drivers and statistical properties of high impact climate extremes , EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-16597, https://doi.org/10.5194/egusphere-egu25-16597, 2025.

Tomato production is vital to Central India's agricultural output and plays a significant role in the region's economy. However, the escalating impacts of climate change pose a serious threat to the sustainability and productivity of tomato farming in this region. This study assesses the effects of variations in solar radiation and temperature on tomato yields utilizing a calibrated process-based crop simulation model (CSM). Climate forecasts utilizing SSP4.5 and SSP8.5 pathways were applied to model yields in near (2010-2039) and mid-future (2040-2069) scenarios. Significant findings indicate a large reduction in yield potential, particularly under mid-future high-emission scenarios (SSP8.5), accompanied by considerable geographical variability. Regions such as Damoh and Western Nimar demonstrate enhanced resilience owing to advantageous local climatic circumstances, whilst areas like the Kymore Plateau and Bundelkhand Agro-Climatic zone display the most significant decreases. Key developmental phases, including flowering and fruit set, are especially susceptible to elevated temperatures and diminished solar radiation. This research highlights the need for region-specific adaptation techniques to alleviate climate impacts, including modifying planting schedules and adopting heat-tolerant varieties. These insights offer a crucial basis for policymakers and farmers to guarantee the sustainability of tomato production in Central India under changing climate circumstances.

Keywords: Crop Simulation Model (CSM), Tomato Yields, GCMs, Central India, policymakers

How to cite: Singh, P. N. and Srivastava, P. K.: Quantifying Climate Impact on Tomato Production in Central India: A Process-Based Yield Simulation for Near and Mid-Future Scenarios, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-544, https://doi.org/10.5194/egusphere-egu25-544, 2025.

Near-surface wind speed (NSWS) plays a critical role in water evaporation, air quality, and energy production. Despite its importance, NSWS changes in South Asia, a densely populated region, remain underexplored. This study aims to understand and quantify the uncertainties in projections of NSWS over South Asia, particularly in relation to internal variability. Utilizing a 100-member large ensemble simulation from the Max Planck Institute Earth System Model, we identified the Interdecadal Pacific Oscillation (IPO) as the leading mode of internal variability influencing South Asian NSWS in the near future. Our findings reveal that the IPO could significantly impact future NSWS, with its positive phase being linked to strengthened westerly flows and increased NSWS across South Asia. Notably, the study shows that accounting for the IPO's impact could reduce NSWS projection uncertainty by up to 8% in the near future and 15% in the far future. This underscores the key role of internal variability, particularly the IPO, in shaping regional NSWS projections. By reducing uncertainties in these projections, our findings can inform climate adaptation strategies for South Asia, helping optimize wind resource assessments in the context of changing wind patterns.

How to cite: Yuan, H.-S. and Shen, C.: Quanatifying the contributions of internal varibility in South Asian near-surface wind speed, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-1831, https://doi.org/10.5194/egusphere-egu25-1831, 2025.

Accurate rainfall forecasting is crucial for effective urban planning and disaster management in Dhaka, the capital of Bangladesh, a city highly vulnerable to urban flooding and extreme weather events. Traditional forecasting methods often struggle to capture the region's complex rainfall patterns, resulting in inaccurate rainfall forecasts. This study evaluates the performance of two traditional machine learning algorithms, Random Forest Regression and Multi-layer Perceptron (MLP), alongside one deep learning algorithm, the Long Short-Term Memory (LSTM) network. These models are trained and tested to forecast rainfall over 1 to 5-day lead times, emphasizing their ability to handle temporal dependencies in time series data. Atmospheric and hydrologic variables, including temperature, surface pressure, evaporation, solar surface radiation, total column rainwater, large-scale precipitation, and total cloud cover, from the ECMWF (European Centre for Medium-Range Weather Forecasts) Reanalysis v5 (ERA5) dataset, were used as model inputs. Model forecasts were validated against ERA5 rainfall data and compared with the forecasts from the Global Forecast System (GFS) model. Results indicate that the Random Forest model outperforms all others, achieving an RMSE of 6.11 mm and Pearson’s correlation coefficient (R) of 0.74 for a 1-day lead time. The LSTM model achieved an RMSE of 7.46 mm, while the MLP performed less effectively than both RF and LSTM, with an RMSE of 7.61 mm. In comparison, the GFS forecasts displayed an RMSE of 9.16 mm. The RF model outperformed the other models at all lead times; however, its accuracy decreased as the lead time increased. This study highlights the potential of machine learning to improve short to medium range rainfall forecasts, contributing to timely decision-making for urban resilience and resource management.

How to cite: Sarker, M. M., Zobeyer, A. M., Rouf, T., and Rahman, S. M. M.: A Comparative Analysis of Data-Driven Machine Learning Models for Rainfall Forecasting in Bangladesh, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-1959, https://doi.org/10.5194/egusphere-egu25-1959, 2025.

Forecasting extreme rainfall events (EREs) locally is a major difficulty for meteorological organizations in India's diverse topography, including Assam, Uttarakhand, and Himachal Pradesh. Flash floods cause major socioeconomic damage in certain areas. These extremes are increasingly commonplace during the southwest monsoon season in the country and one of the most destructive EREs occurred in June 2022 and 2023 over Assam. This work explores deep learning (DL) models, specifically spatial attention-based U-Net, in conjunction with simulated daily collected rainfall outputs from different parametrization schemes rainfall output from the Weather Research and Forecasting (WRF) model, considering the limitations of deterministic numerical weather models in accurately forecasting these events. The model trained over the districts of Assam for all days (days 1-4) except the districts where the EREs occurred. The suggested model exhibited a greater ability to predict rainfall at the district scale with a mean absolute error of less than 10 mm over four days in June 2022, outperforming both individual and ensemble outputs of WRF. Furthermore, the suggested model had a high prediction accuracy of 91.9% in categorical rainfall prediction, outperforming WRF models by 51.3%. Furthermore, by accurately forecasting EREs at the district level, including Barpeta, Kamrup, Kokrajhar, and Nalbari, the suggested model has shown improved spatial variation when compared to the WRF model. The suggested DL model is tested for real-time ERE events over Assam in June 2023. In the second part, the model has trained for ERE occurred in 2022 and tested for 2023 over Assam at the district level. The district-level performance of the DL and WRF models is compared, and the DL model performs better than the WRF model in capturing EREs, with a noteworthy accuracy of 54.4% compared to only 22.8% for the WRF model. Notably, the DL model accurately represents the amount and severity of rainfall in Assam's western and southern regions. In summary, the study's conclusions directly affect the development of effective strategies for increased preparedness, mitigation, and adaptation measures over complex hilly regions to lessen the loss of life and property, as well as the improvement of early warning systems and related follow-up action.

How to cite: Trivedi, D., Pattnaik, S., and Sharma, O.: Enhancing Heavy Rainfall Predictions Over Vulnerable Regions in Assam Using a Spatial Attention-Based Deep Learning Network, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-2071, https://doi.org/10.5194/egusphere-egu25-2071, 2025.

The Southern Great Plains are subject to fluctuating precipitation extremes that pose significant challenges to agriculture and water management. Despite advancements in forecasting, the mechanisms driving these climatic variations remain incompletely understood. This study investigates the relative contributions of the tropical Pacific and Atlantic Oceans to April-May-June precipitation variability in this region. Using partial ocean assimilation experiments within the Community Earth System Model, we identify a substantial influence of inter-basin interactions, with the Pacific and Atlantic contributing approximately 70% and 30%, respectively, to these variations. Our statistical analysis suggests that these tropical inter-basin contrasts offer a more reliable indicator for late-spring precipitation anomalies than the El Niño-Southern Oscillation. This finding is corroborated by analyses from seven climate forecasting systems in the North American Multi-Model Ensemble, providing a promising outlook for improving real-time forecasting in the Southern Plains. However, the predictive skill of these inter-basin contrasts is currently limited by the lower predictability of the tropical Atlantic, underscoring the need for future research to enhance climate prediction models.

How to cite: Chikamoto, Y., Wang, S., Chang, H.-I., and Castro, C.: Seasonal Predictability of Late-Spring Precipitation in the Southern Great Plains , EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-3766, https://doi.org/10.5194/egusphere-egu25-3766, 2025.

The Eddy-Diffusivity Mass-Flux (EDMF) parameterization (Giordani et al., 2020) offers a new, coherent way to simultaneously parameterize local (diffusivity) and non-local (convective thermal) vertical mixing. This second component parametrizes sub-grid-scale convective plumes propagating through the water column which, through energy conservation, can propagate counter to the stratification gradient. The EDMF scheme is assessed in a 13-year global ¼° coupled NEMO4.2-SI3 simulation, forced by ERA5 atmospheric reanalysis. Its performance in representing observed ocean temperatures is compared to that of a twin simulation using the commonly applied Enhanced Vertical Diffusivity (EVD) parameterization.

The EDMF simulation shows globally reduced temperature biases relative to in-situ observations (0–700 m) compared to the EVD simulation, with similar RMSD (Root Mean Square Deviation) values between the two. By better representing tropical night-time shallow convection, EDMF reduces the cold bias typically observed in EVD simulations within the tropical ocean. We show that the horizontal scales (convective areas), penetration depths and vertical velocities of the simulated plumes agree with measurements of deep convective plumes in the Labrador Sea, and with diurnal convection in the equatorial Pacific Ocean. Additionally, first estimates of convection's contribution to Ocean Heat Content are proposed.

How to cite: Piton, V., Bourdallé-Badie, R., and Giordani, H.: The Eddy-Diffusivity Mass-Flux parameterization: improved representation of convective mixing, global evaluations and implications for Ocean Heat Content, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-4069, https://doi.org/10.5194/egusphere-egu25-4069, 2025.

The TCRE relationship underlies the necessity of net zero emissions for climate stabilization and the utility of carbon budgets as a policy tool. TCRE emerges near universally from Earth system models, and is consistent with observations. However, recent work has systematically dismantled the leading hypothesis explaining the phenomenon, concluding “that this proportionality is not amenable to a simple physical explanation, but rather arises because of the complex interplay of multiple physical and biogeochemical processes.'' (Gillett, 2023). Here we set two intermediate complexity Earth system models (EMICs) to abiotic states, then turn on broad components of Earth's biogeochemical cycles one at a time to see which combination of processes cause TCRE to emerge.

We find that TCRE emerges when ocean alkalinity is set to observed values, without life on land. TCRE likewise emerges independently when the terrestrial biosphere is turned on, with the ocean in an abiotic low alkalinity state. Idealized experiments with the EMICs show that TCRE occurs for configurations of the Earth system where characteristic timescales of carbon absorption and heat absorption are nearly the same. Our results suggest that the emergence of TCRE does in-fact rely on a simple physical mechanism, but why the living components of Earth system are matching the characteristic timescale of carbon absorption to that of heat remains mysterious.

Gillett, N.P.: Warming proportional to cumulative carbon emissions not explained by heat and carbon sharing mixing processes. Nature Communications 14(1), 6466 (2023)

How to cite: MacDougall, A. and MacIsaac, A.: The double emergence of TCRE, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-6364, https://doi.org/10.5194/egusphere-egu25-6364, 2025.

Water scarcity is a growing challenge exacerbated by climate change, particularly in regions like Lombok Island, Indonesia, where water resources are crucial for sustainable development. This research aims to identify suitable locations for Rainwater Harvesting (RWH) by integrating geospatial analysis, the Analytic Hierarchy Process (AHP), and climate projections using CMIP6 data. The study utilizes multiple parameters, including rainfall, land use/land cover (LULC), slope, drainage density, soil texture, and runoff depth, to develop a comprehensive suitability map for RWH.

Historical rainfall data from CHIRPS (1981–2010) and future rainfall projections for mid-century (2031–2060) and end-century (2071–2100) under SSP2-4.5 and SSP5-8.5 scenarios were analyzed to account for climatic variations. Each parameter was processed using geospatial tools, with weights assigned through AHP based on expert input, ensuring a robust multi-criteria decision-making framework. Suitability maps were generated for each temporal scenario, highlighting areas with high to very high potential for RWH, particularly in North and East Lombok.

The results reveal dynamic shifts in RWH site suitability over time, with increasing precipitation under SSP5-8.5 scenarios expanding high-suitability areas. These findings highlight the potential for RWH to manage water resources adaptively in response to projected climate variability. By aligning the outputs with existing water management infrastructure, such as dams, the study provides actionable insights for regional planners and policymakers.

How to cite: Ulfah, A., Renwick, J., and Patria Megantara, R.: Integrating Climate Projections and Geospatial Analysis to Identify Rainwater Harvesting Suitability in Lombok Island, Indonesia, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-7421, https://doi.org/10.5194/egusphere-egu25-7421, 2025.

Precisely predicting weather parameters is crucial for precision horticulture, especially in horticultural lands where timely environmental insights significantly impact crop yield and quality. This study presents a novel hybrid modeling approach employing 1D Transformer networks integrated with traditional machine learning techniques to predict hourly temperature variations. Utilizing the ERA5 reanalysis dataset spanning from 1940 to December 2024, the hybrid model efficiently captures location-specific spatiotemporal dependencies and nonlinear trends in historical weather data.

The predicted weather data generated by the hybrid model is used in FarmD, a web-based user interface developed for farmer-centric applications. FarmD provides real-time visualization of critical weather parameters, including temperature, relative humidity, wind patterns, rainfall, and soil temperature, specifically tailored to horticultural regions. Through its intuitive interface, users can query predicted and historical data by selecting attributes, dates, and times, with an option for location-specific searches to support targeted agricultural decision-making.

This integration of predicted data with an accessible web platform highlights significant advancements in delivering actionable insights to end users. By combining advanced computational methods with user-focused design, FarmD enables horticulturists to adopt data-driven practices, contributing to sustainable and efficient agricultural management.

How to cite: Ramalingam, S.: FarmD: A Web Interface for Visualization of Predicted Weather Parameters Using 1D Transformer Hybrid Models, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-10613, https://doi.org/10.5194/egusphere-egu25-10613, 2025.

Estimates of remaining carbon emissions budgets to limit global warming to 1.5°C or 2°C rely on the near-linear relationship between the change in global mean temperature and total CO₂ emitted since the pre-industrial era. This relationship is known as the Transient Climate Response to cumulative Emissions (TCRE). Previous estimates of TCRE are derived from Earth System Models (ESMs) which are known to lack key processes that affect warming and therefore diagnosed CO₂ emissions. Here we use the UK Earth System Model to quantify, for the first time, the impact on TCRE of including six Earth system processes in isolation (results in parenthesis): fire-vegetation interactions (TCRE increased 14.6%); nitrogen limitation of vegetation (+9.7%); diffuse radiation effects on vegetation (+8.5%); changes in vegetation distribution (-1.5%); climate impacts from wetland methane emissions (+5.1%) and from biogenic volatile organic compounds (-1.4%). From these results we recalculate the TCRE of 11 ESMs of the 6th Coupled Model Intercomparison Project (CMIP6) as though each included all six processes. Averaged over the 11 models, TCRE increased by 23.7%, reducing by 19% the associated remaining carbon budget to both 1.5°C and 2°C.

How to cite: Liddicoat, S., Jones, C., Mercado, L., Robertson, E., Sitch, S., and Wiltshire, A.: Role of Earth system processes in the Transient Climate Response to cumulative Emissions, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-11584, https://doi.org/10.5194/egusphere-egu25-11584, 2025.

Evaporation and potential evapotranspiration are components of hydrological cycle that represent the loss of water from the surface and vegetation to the atmosphere. Potential evapotranspiration is a theoretical index that demonstrates the maximum evaporation and transpiration rates assuming sufficient water availability in soil and canopy. Six identical Regional Climate Models (RCMs) of Euro-CORDEX project were selected in order to obtain a unified ensemble for both characteristics for estimation under RCP2.6, RCP4.5 and RCP8.5 scenarios for the middle (2021-2050) and the end of the 21^st century (2071-2100) for Ukraine. ERA5 observational dataset is used as a baseline climate normal (1991-2020) for tracking the future changes. In this study we applied a quantile mapping approach for bias correction for smoothing systematic errors between observational and simulated datasets. For the baseline period, the sums of evaporation varied mainly between 20-30 mm in winter to 250-290 mm in summer, with the exception of the Carpathians and southern regions near marine coastal areas (more than 300 mm). Climate normals of evapotranspiration were zonally distributed with the exception of mountainous region and varies from 20-50 mm in winter to 290-550 mm in summer. The most tremendous changes of evaporation are expected to occur in winter. In general, during the following 30-year period of 2021-2050, the most significant increase by 8-18% (compared to 1991-2020 baseline) would be expected for RCP4.5 with more pronounced increase during 2071-2100, reaching its highest values up to 40% under RCP8.5 The maximum rates are observed in the Carpathians and the northeast of Ukraine. In contrast, evapotranspiration in winter is expected to increase only by 1-6% during 2021-2050 for all RCPs and 12-22% by the end of the century. The Carpathians will face even a decrease by -4%. Changes in evaporation will be lower for the spring season, with changes by 2-4% in 2021-2050 and 6-12% by the end of the century. The highest spring changes up to 28% also will occur in the Carpathians. The same rates are estimated for evapotranspiration, for which the sharpest changes are 10-16% under RCP 8.5 for 2081-2100 In comparison to winter and spring, summer and autumn seasons will face much slower changes. Moreover, summer season will be characterized by a decrease in evaporation at a rate up to -2..-4% under RCP2.6 and varying within ±1% for other scenarios by the mid-century, showing the typical tendencies for so called “evaporation paradox”. In 2071-2100, the decrease can reach by up to-6% for RCP4.5 and RCP8.5. It must be noted the different tendency for evapotranspiration with an increase by 1-6% in general for all RCPs in 2021-2050, and maximum up to 14% by the end of the century. For autumn the most typical increase in both parameters is within 2-6% for all RCPs, with the highest rates of evaporation in the Carpathians up to 15%.  The obtained results show the importance of considering evaporation in future water management, agriculture and food security in Ukraine, highlighting the seasons and regions with it significant changes.  

How to cite: Rybchynska, V., Pysarenko, L., Pushkar, H., Savenets, M., and Osadchyi, V.: Seasonal changes in evaporation and potential evapotranspiration under different scenarios of climate change on the territory of Ukraine, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-12181, https://doi.org/10.5194/egusphere-egu25-12181, 2025.

Deep learning, a prominent artificial intelligence method, is increasingly applied in research addressing the impacts of global warming in the future. However, it is widely acknowledged that deep learning exhibits limitations in extrapolation, as it typically predicts accurately only within the range of the training data. When future scenarios extend beyond this range, the reliability of predictions can diminish significantly. In Japan, for example, the annual maximum precipitation is reported to be increasing, according to the Japan Meteorological Agency, indicating a potential for future values to exceed historical records. Despite this, limited studies have explored the extent to which deep learning methods can reliably extrapolate beyond the training data range. This study quantitatively evaluates the extrapolation capability of deep learning in hydrology, specifically focusing on rainfall-runoff modeling at the watershed scale. Meteorological data, including precipitation and temperature, are utilized as inputs, while river flow serves as the output. The Long Short-Term Memory (LSTM) model, which is well-suited for time-series data, was employed as the deep learning framework. Data were partitioned into training, validation, and test datasets, with river flow values categorized using threshold percentiles of 90, 95, 97, 98, and 99, rather than conventional time-based splits. This approach allows for a focused investigation into the range of accurate extrapolation beyond the training dataset. Preliminary findings reveal that the LSTM model successfully captured peak river flows up to 250.1% higher than the maximum values of the observed river flow discharge in the training-validation dataset. These results demonstrate the potential for deep learning to extrapolate in hydrological modeling, though further research is necessary to assess the performance of alternative deep learning methods and additional case studies. 

How to cite: Junsei, S.: Evaluating the extrapolation capability of deep learning in rainfall-runoff, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-14457, https://doi.org/10.5194/egusphere-egu25-14457, 2025.

In the context of climate change, a global, widespread shift to increased water limitation is expected over approximately 73% of terrestrial ecosystems, with important implications for food and water security, CO2 uptake, and evaporative cooling. Water-limited regions, exposed to climate-change-related increasing droughts and intense anthropogenic water use, are extremely vulnerable to transitions towards drier eco-hydro-climatological regimes. In the longer term, the ongoing drought conditions may intensify the decline of groundwater levels, threatening groundwater-dependent ecosystems and exacerbating the risk of desertification, thereby amplifying a positive feedback on regional climate change. In some Mediterranean climate-type regions, such as SouthWestern Australia, a dry and warm transition has already been observed. Recent findings are a clear warning that also over the Euro-Mediterranean sector groundwater level may have a negative trend resulting from a decrease in precipitation and/or increasing withdrawal. 

Soil water storage  and groundwater dynamics represent important hydrological processes related to these transitions but they are greatly simplified in state-of-the-art Earth System Models (ESMs). Therefore, it is  essential to improve the representation of hydrological processes and their coupling with the atmosphere and the land surface in ESMs. In this respect, the land surface model included in EC-Earth (ECLAND) still lacks a representation of groundwater and instead implements a free drainage condition at the bottom of the unsaturated soil column. 

In this work, we intend to implement a more realistic groundwater representation in EC-Earth by including a global-scale water table to replace the free drainage bottom boundary condition. As a preliminary measure, the impact of groundwater on the shallow, unsaturated zone is evaluated by constraining the vertical water fluxes with a static water table depth (WTD) derived from a global estimate simulation based on observations. We evaluated the effects of this implementation on water and energy fluxes against a network of stations in land-only simulations from 1979 to the present, with boundary forcing taken from ERA5 reanalysis. First findings suggest that including a WTD has an impact on water exchanges between saturated and unsaturated soil in water-limited regions, particularly in semi-arid and transitional climates, which can not be neglected in Earth system models.

How to cite: Senigalliesi, V., Alessandri, A., Kollet, S., Gelsinari, S., Cherchi, A., and Di Carlo, E.: Enhancing Hydrological Processes in Earth System Models: Implementing Groundwater Dynamics for Improved Climate Representations, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-18989, https://doi.org/10.5194/egusphere-egu25-18989, 2025.

The construction and analysis of daily temperature data series in long enough a time period is important to understand decadal to multi-decadal variability and changing trends in extreme temperature events. This paper reports a new analysis of extreme temperature indices over the last 140 yr in Wuhan, China, with an emphasis on changes in extreme high temperature changes. The daily temperature data from 9 stations from 1881 to 1950 and 1 modern station from 1951 to 2020 were used for the analysis. Based on the data, and the commonly used extreme temperature indices, the variations and long-term trends of extreme high temperature events in Wuhan since 1881 were analyzed. The results show that there was no clear warming trend in maximum temperature, but a quite large inter-annual and inter-decadal fluctuation. In contrast, there was a very significant increase in minimum temperature, with a large upward trend overall. The extreme temperature indices exhibit a periodic variability, and frequent extreme heat events have been experienced over the last 140 yr in Wuhan. Most extreme temperature indices did not exhibit remarkable changes during the first half of the period analyzed. However, the majority of the extreme temperature indices showed significant upward trends over the latter half of the 140 yr period. The possible causes of the observed changes in the extreme high temperature events in the different time periods are also discussed.

How to cite: Zheng, X., Ren, G., He, J., Zhao, Y., Ren, Y., and Yang, G.: Temporal characteristics of extreme high temperatures in Wuhan since 1881, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-456, https://doi.org/10.5194/egusphere-egu25-456, 2025.

Climate change is expected to exacerbate heat stress, particularly in urban areas where the urban heat island (UHI) effect tends to amplify warming compared to surrounding rural regions. Due to the heterogeneity of urban environments, heat stress can vary significantly within cities. Heat vulnerability maps, which combine data on heat sensitivity, heat exposure and adaptive capacity, are valuable tools for identifying areas that should be prioritized for heat stress mitigation measures. One important component of such heat vulnerability maps is data on the spatial distribution of heat. The present study explores the use of satellite data to generate high-resolution temperature maps, addressing two key challenges in the process.

The first challenge arises from the fact that satellites measure land surface temperature (LST) rather than air temperature (AT), whereas the latter is needed as input for most heat stress indicators. While linear models calibrated with weather station data are frequently used to estimate AT from LST, there are cities where the availability of weather stations is insufficient for calibrating models with multiple control variables. Additionally, the LST-AT relationship depends on the prevailing atmospheric conditions. The second challenge of using satellite data is that satellite images are usually not available on an hourly or daily basis due to factors such as satellite scheduling or excessive cloud cover.

To address the first challenge, we adopt a technique introduced by the ECOSTRESS mission, which leverages reanalysis data (GEOS-5) to estimate AT using LST, the normalized difference vegetation index (NDVI), and albedo. We apply this method to spatially downscaled LST data (100m) from the VIIRS instrument aboard the Suomi NPP satellite, AT reanalysis data from ERA5-Land (9km), as well as NDVI and albedo derived from Harmonized Landsat Sentinel (HLS) data (aggregated to 100m). Applying the method to individual satellite images enables day-specific adjustments for varying atmospheric conditions. To overcome the second challenge, we utilize high-resolution AT maps derived from LST images to calculate spatial patterns of air temperature distribution, which are then used to downscale ERA5-Land AT data for those times without satellite images available.

To evaluate the approach described, it is exemplary applied to various cities, whereby the downscaled temperature estimates are validated against (i) temperature estimates based on alternative methods than the ECOSTRESS technique to derive AT from LST, (ii) weather station data, and (iii) existing results from urban climate models.

How to cite: Kortschak, D., Gallaun, H., Kernitzkyi, M., Köberl, J., Miletich, P., and Strohmaier, M.: Spatial downscaling of urban temperatures: evaluation of an approach using satellite and reanalysis data, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-1411, https://doi.org/10.5194/egusphere-egu25-1411, 2025.

Underwater cultural heritage, such as ancient shipwrecks and submerged archaeological sites, faces increasing risks from climate-driven environmental changes. Salinity shifts, temperature anomalies, and biofouling contribute to the degradation of these resources [1]. This study explores deploying two IoT-enabled devices with a crowdsourcing strategy to monitor and address these challenges effectively.

The first device, designed for divers, measures pressure, temperature, and salinity during underwater campaigns and can be placed on the seabed for long-term data collection [2]. The second device, used by local communities like fishers and diving centers, is deployable from boats to 2-3 meters, capturing salinity, temperature, and chlorophyll concentration. Each device incorporates a data logger built on a microcontroller, connected to sensors via robust serial interfaces such as RS485. This configuration ensures reliable communication and minimizes signal degradation in challenging underwater conditions. The microcontroller interfaces with sensors to record measurements, storing data locally until retrieval. Both devices feature a power management system with custom-designed PCBs for efficient energy use.

Data gathered by the devices is stored locally and transferred to a cloud platform via an intuitive mobile app. Communication between the devices and the smartphone uses Bluetooth Low Energy (BLE), while data uploads to the cloud via LTE. This simplifies retrieval and reduces the need for complex equipment or infrastructure.

Community participation plays a central role in this system. Local communities deploy and retrieve boat-based sensors, improving the coverage and frequency of monitoring activities. By pooling data from various contributors, detailed information of environmental conditions near cultural heritage sites is acquired.

The devices undergo rigorous calibration to ensure reliable data collection. Conductivity sensors are standardized with salinity benchmarks, temperature sensors tested with laboratory-grade instruments, pressure sensors calibrated in controlled chambers, and chlorophyll sensors validated using fluorescence references.

Field trials at two underwater sites tested the system under diverse conditions, providing a robust environment to assess device performance and crowdsourcing effectiveness. Feedback from divers, local participants, and heritage professionals refined functionality. Adjustments included stronger enclosures, improved BLE connection stability, and an enhanced mobile app interface.

This study demonstrates the potential of combining smart sensor technology with community engagement to protect underwater heritage. Leveraging IoT devices and collaboration expands monitoring, reduces costs, and fosters local stewardship, offering a scalable, sustainable solution to mitigate environmental impacts on submerged cultural treasures.

References:

[1] P. Michalis, C. Mazzoli, V. Karathanassi, D. I. Kaya, F. Martins; M. Cocco, A. Guy and A. Amditis, "THETIDA: Enhanced Resilience and Sustainable Preservation of Underwater and Coastal Cultural Heritage," IGARSS 2024 - 2024 IEEE International Geoscience and Remote Sensing Symposium, Athens, Greece, 2024, pp. 2208-2211, doi: 10.1109/IGARSS53475.2024.10642229.

[2] L. Pavlopoulos, P. Michalis, M. Vlachos, A. Georgakopoulos, C. Tsiakos and A. Amditis, "Integrated Sensing Solutions for Monitoring Heritage Risks," IGARSS 2024 - 2024 IEEE International Geoscience and Remote Sensing Symposium, Athens, Greece, 2024, pp. 3352-3355, doi: 10.1109/IGARSS53475.2024.10641101.

Acknowledgement:

This research has been funded by European Union’s Horizon Europe research and innovation programme under THETIDA project (Grant Agreement No. 101095253).

How to cite: Vlachos, M., Michalis, P., Mourounas, I., Koukio, P., Gkatzogias, A., Georgakopoulos, A., and Amditis, A.: IoT-Enabled Underwater Devices and Crowdsourcing for Monitoring Climate Risks at Submerged Heritage Sites, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-3194, https://doi.org/10.5194/egusphere-egu25-3194, 2025.

Assessing the risks posed by climate change to cultural heritage (CH) is crucial for developing effective strategies to preserve this non-renewable heritage. This study provides a comprehensive approach to assess climate change-related risks to cultural heritage across five selected sites in Europe: Choirokoitia, Aegina, Epidaurus, Kalapodi, and Ventotene. By applying the Heritage Outdoor Microclimate (HMR_out) and Predicted Risk of Damage (PRD) indices, the study quantifies potential damage to inorganic materials due to long-term changes in temperature and relative humidity (RH). Climate projections are based on high-resolution EURO-CORDEX Regional Climate Model (RCM) simulations under three Representative Concentration Pathways (RCP2.6, RCP4.5, and RCP8.5) for the periods 2021–2050, and 2071–2100. Results reveal a significant increase in temperature and the related indices under all emission scenarios highlighting a warming trend and intensified heat stress across the CH sites. The projected rise in temperature leads to an increase in the HMR_out index across all the CH sites, with the rate of change differing between time periods and scenarios. This rise in the HMR_out index suggests an increase in the predicted risk of damage (PRD) to monuments made of inorganic materials due to heat stress. In contrast, RH and the associated PRD index are expected to decrease. Overall, the projected changes in the HMR_out and PRD indices provide a deeper insight into how climate change may influence preservation of cultural heritage sites constructed from stone and marble.

This work is based on procedures and tasks implemented within the project “Toolbox for assessing and mitigating Climate Change risks and natural hazards threatening cultural heritage - TRIQUETRA”, which is a Project funded by the EU HE research and innovation program under GA No. 101094818.

How to cite: Tringa, E., Georgoulias, A. K., Akritidis, D., Feidas, H., and Zanis, P.: Climate Change and Cultural Heritage: Assessing Future Risks of Damage at Selected European Cultural Heritage Sites, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-3882, https://doi.org/10.5194/egusphere-egu25-3882, 2025.

Crop yield is important for agricultural productivity and country’s economy. Accurate crop yield estimation is critical for policymakers, farmers, and governments because it allows better management techniques, decision making and the implementation of practicable agricultural policies. While crop yield estimation is an essential aspect of modern agriculture, it continues to be one of the most challenging tasks to manage effectively. In this study, we used the Food and Agriculture Organization (FAO) of the United Nations developed AquaCrop model to estimate the crop yields of corn and soybean crops in Illinois, United States (US). Data of various meteorological parameters as precipitation, maximum and minimum temperature, relative humidity, wind speed, solar radiation datasets were collected from NASA Prediction of Worldwide Energy Resources (POWER), for a period of 25-years from 2000 to 2024. Whereas, reference evapotranspiration was calculated by using the modified Hargreaves method. The objective of this study is to assess the accuracy of yield estimation of corn and soybean by using the AquaCrop model. The AquaCrop model was simulated for the growing period of corn and soybean from May to September. Using the AquaCrop model, the maximum and minimum corn yields were found to be 14.49 tons/ha in the year 2022 and 7.60 tons/ha in the year 2005, respectively. Similarly, the maximum yield of soybean was found to be 4.33 tons/ha in the year 2022, while the minimum yield was 2.26 tons/ha in the year 2012. The coefficient of determination (R²) values of 0.72 for maize and 0.76 for soybean, gives a satisfactory level of model accuracy. The model's performance can be improved further by incorporating more ground-truth data and appropriate parameters. This study demonstrates the AquaCrop model's ability to estimate crop production with few input parameters, as well as suggest opportunities for improvement. To improve prediction accuracy and promote informed agricultural planning and food security, future study might use sophisticated methodologies, localized farming practices, crop phenology, and specific soil data. 

Keywords:  AquaCrop, Crop yield, Illinois, Yield Predictions.

How to cite: Gautam, V. and Pathak, S.: Optimizing Corn and Soybean Yield Predictions in Illinois Using the AquaCrop Model , EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-3943, https://doi.org/10.5194/egusphere-egu25-3943, 2025.

Climate change is reshaping the species composition, distribution and extent of forests worldwide. Across vast areas in Central Europe widespread Norway spruce (Picea abies) has exhibited large-scale decline, primarily due to its vulnerability to drought events. Forest management is thus facing important questions related to the replacement of Norway spruce, especially in areas where it was introduced due to its high economic value.

This study investigates the potential of Douglas fir (Pseudotsuga menziesii), a drought- and pest-tolerant non-native species, as a more resilient alternative for use in production forests. At the experimental plot in Jable, central Slovenia where both species coexist, we monitored the xylogenesis of five Douglas firs and five Norway Spruces from March to October 2024 by sampling phloem, cambium and xylem tissue every two weeks using the Trephor tool. Additionally, we collected tree cores from 20 trees of each species to perform dendrochronological analyses. These analyses aim to assess climate-growth correlations and growth-based resilience indicators (resilience, resistance, recovery and recovery period).

 The main objective of this study is to determine whether Douglas fir is to compare 1) interannual growth dynamics, 2) intra-annual growth dynamics of xylem and phloem, 3) climate-growth relationships, and 4) resilience components of both species. We hypothesize that non-native Douglas fir will exhibit greater growth rates and better resilience indicators and could thus be considered as a replacement for Norway Spruce at similar forest sites in central Slovenia and beyond. By addressing critical knowledge gaps regarding the responses of these species to climate variability, this research can provide important insights to support the strategic adaptation of forestry practices and improve the resilience of ecosystems in the face of environmental change.

How to cite: Balzano, A., Partemi, R., Jevšenak, J., Krže, L., and Merela, M.: Douglas Fir (Pseudotsuga menziesii) as an alternative species for the declining Norway spruce (Picea abies) in central Europe: Dendrochronological and xylogenetic insights from Slovenia, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-4269, https://doi.org/10.5194/egusphere-egu25-4269, 2025.

Coastal sea level changes have profound impacts on coastal ecosystems, infrastructure, and communities. Interannual sea level variations along the U.S. East Coast are influenced by a combination of dynamic and thermodynamic processes, including local wind forcing, Gulf Stream variability, regional ocean circulation changes, and thermosteric contributions. These processes are interconnected and strongly modulated by large-scale climate modes such as the North Atlantic Oscillation (NAO), El Niño-Southern Oscillation (ENSO), and Atlantic Multi-decadal Oscillation (AMO). This study leverages machine-learning-based predictive models to quantify and forecast interannual sea level variability by integrating diverse climate indicators. By incorporating indices of large-scale climate modes alongside local and regional oceanographic parameters, the model quantifies the relative contributions of each factor and identifies the dominant processes driving observed variability. The results demonstrate the potential of machine-learning approaches to capture complex nonlinear relationships between climate modes and regional sea level changes. NAO-driven atmospheric forcing and ENSO-related ocean-atmosphere interactions emerge as key predictors, with the models successfully replicating observed variability along different sections of the U.S. East Coast. The findings highlight the importance of integrating large-scale climate dynamics into regional sea level prediction frameworks and suggest new opportunities for improving forecast accuracy at interannual timescales.

How to cite: Ye, Z., Ye, Z., and Zhao, J.: Predicting Interannual Sea Level Variations Along the U.S. East Coast Using Machine Learning and Climate Indicators, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-7561, https://doi.org/10.5194/egusphere-egu25-7561, 2025.

Environmental and socioeconomic factors increase small islands' exposure and vulnerability to climate change. This study reviews issues related to current and future climatic change and its impacts on the small island environments in the Canary Islands. Convection-permitting regionalized projections driven by data from three global climate models included in the Coupled Model Intercomparison Project (CMIP5) have been performed, covering the recent past (1980–2009) and future (2070–2099) periods, under two Representative Concentration Pathways, 4.5 and 8.5. The impact analysis includes water resources, energy, ecosystems and biodiversity, natural hazards, and health issues. We provide a succinct review of sectors that warrant particular attention, due to their weight in the gross domestic product, agriculture and tourism. The concluding section discusses adaptation and response strategies, and the portfolio of research that needs to be addressed. 

How to cite: Carrillo, J., Barrancos, J., Tondreau, P. S., Pérez, J. C., González, A., Expósito, F. J., and Díaz, J. P.: A review of climate change impacts in the Canary Islands, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-8577, https://doi.org/10.5194/egusphere-egu25-8577, 2025.

The aim of this study was to find the connection between the large-scale atmospheric circulation in the winter season and the occurrence of extreme precipitation in the spring months at the regional scale. For the large-scale circulation, climate indices (GBOI and NAOI) associated with the Greenland-Balkan Oscillation and the well-known North Atlantic Oscillation were considered, and for the regional scale, certain representative stations for the middle and lower Danube basins were considered. The tests were carried out for a 120-year interval (1901-2020), by applying the extreme value theory (EVT). The modelling of maximum precipitation was carried out through the generalized extreme value (GEV) distribution. In order to see the impact of the large-scale circulation, the results obtained by incorporating NAOI as covariate into the location parameter of GEV distribution, were compared with the results obtained considering GBOI as covariate. For extreme precipitation in the lower basin area, the influence of GBOI is much more significant than that of NAOI, while for the middle basin area, the differences between the two indices are not so significant.

How to cite: Mares, I., Dobrica, V., Mares, C., and Demetrescu, C.: On the links between large-scale atmospheric circulation and extreme precipitation in the middle and lower Danube basin, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-9608, https://doi.org/10.5194/egusphere-egu25-9608, 2025.

Since Shi et al. proposed that the climate in the drylands of Northwest China experienced a significant transition from a “warming and drying” trend to a “warming and wetting” trend in the 1980s, researchers have conducted numerous studies on the variations in precipitation and humidity in the region and even in arid Central Asia. In particular, the process of the “warming and wetting” trend by using obtained measurement data received much attention. However, there remain uncertainties about whether the “warming and wetting” trend has paused and what its future variations may be. In this study, we examined the spatiotemporal variations in temperature, precipitation, the aridity index (AI), vegetation, and runoff during 1950–2019. The results showed that the climate in the drylands of Northwest China and the northern Tibetan Plateau is persistently warming and wetting since the 1980s, with an acceleration since the 1990s. The precipitation/humidity variations in North China, which are mainly influenced by summer monsoon, are generally opposite to those in the drylands of Northwest China. This reverse change is mainly controlled by an anomalous anticyclone over Mongolia, which leads to an anomalous easterly wind, reduced water vapor output, and increased precipitation in the drylands of Northwest China. While it also causes an anomalous descending motion, increased water vapor divergence, and decreased precipitation in North China. Precipitation is the primary controlling factor of humidity, which ultimately forms the spatiotemporal pattern of the “westerlies-dominated climatic regime” of antiphase precipitation/humidity variations between the drylands of Northwest China and monsoonal region of North China. The primary reasons behind the debate of the “warming and wetting” trend in Northwest China were due to the use of different time series lengths, regional ranges, and humidity indices in previous analyses. Since the EC-Earth3 has a good performance for simulating precipitation and humidity in Northwest and North China. By using its simulated results, we found a wetting trend in the drylands of Northwest China under low emission scenarios, but the climate will gradually transition to a “warming and drying” trend as emissions increase. This study suggests that moderate warming can be beneficial for improving the ecological environment in the drylands of Northwest China, while precipitation and humidity in monsoon-dominated North China will persistently increase under scenarios of increased emissions.

How to cite: Xie, T.: Discussion of the “warming and wetting” trend and its future variation in the drylands of Northwest China under global warming, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-10025, https://doi.org/10.5194/egusphere-egu25-10025, 2025.

The Himalayan region of India is experiencing warmer winters and hotter summers, which are causing reduced yields and putting the production of traditional fruit species in danger. In order to gain an understanding of the thermal growing conditions, it is essential to have chill and heat accumulation monitored. In the current investigation, the Dynamic model is utilized to compute the chill accumulation, while the Growing Degree Days (GDD) method is utilized to compute the heat accumulation. In order to calculate these indices, gridded hourly temperature data from the European Centre for Medium-Range Weather Forecasts (ERA)5 dataset was utilized. The time period covered by this dataset is from 1940 to 2023. The study's findings revealed the best elevation ranges for several of the region's most significant fruits, such as citrus fruits, almond trees, and fresh fruits. Furthermore, places with elevations ranging from 1000 to 2000 are good for growing fresh fruits. This is due to the fact that 70 percent of the Chilling Portion (CP) values are high enough to be greater than 60.

How to cite: Shukla, Y. and Gupta, V.: Assessing Climate Change Effects on Fruit Growing Conditions in the Northwestern Himalayan Region of India, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-14951, https://doi.org/10.5194/egusphere-egu25-14951, 2025.

Fires can alter the properties of soil and other material via heat transfer. The identification of soil heating effects in hearths, for example, has long been a cornerstone in archaeological investigations. However, wildfires can also alter soils, and there is a surprising level of uncertainty into what degree soils are heated and to which depth this occurs in wildfires. This can lead to erroneous assumptions regarding the potential impact of wildfires when attributing heat induced changes in the soil, especially when laboratory heating results are extrapolated to field conditions.

To address this research gap, we compiled and examined new and published field data on maximum temperatures and heating durations for mineral soils during wildfires and prescribed burns in forests, shrublands and grasslands around the globe; and compared these to data obtained from laboratory heating experiments.

Most fires heated only the uppermost centimetres of the mineral soil, rarely exceeding 300 °C below 1 cm depth. Their heat pulses were shorter (<500 s) than those often applied in laboratory studies (1800-3600 s). The highest near-surface temperatures occurred in shrubland wildfires, whereas the longest heating durations in forests with deep organic layers and high fuel loads.

While it is clear that smouldering logs, tree trunks and root systems, or slash pile burns can impart intense heating to substantial depths akin to that under hearths, most landscape-scale fires generate short and shallow heat pulses that are unlikely to lead to detectable lasting changes in the mineral soil. 

How to cite: Doerr, S., Badia-Villas, D., Bryant, R., Matthew, D., Antonio, G.-G., Jorge, M.-S., Jessica, M., Carmen, S.-G., Cristina, S., Cathelijne, S., and Pete, R.: Soil heating under wildfires and prescribed burns and their relevance to archaeological investigations, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-15465, https://doi.org/10.5194/egusphere-egu25-15465, 2025.

Fires are among the most significant causes leading to significant alterations, both at the level of the natural and built landscape. These in fact induce significant alterations not only on the vegetation cover, but also on fauna, soil, atmosphere, artifacts and, inevitably, economic losses as well. In the context of the archaeological heritage, fires are a cause of extensive damage especially at the territorial scale, on sites and fragments not yet subject to either excavation or reconnaissance campaigns, but also on known sites that suffer from insufficient protection actions.

Traditional methods of assessing fire severity and property damage incur costs in terms of money and time because of the necessary field survey activities. A combination of geodata science and remote sensing, on the other hand, turns out to be an inexpensive and effective tool for modeling fires, understanding their causes and fire evolution.

In this work we use the potential of geodata science methods applied to spatial and satellite data, to analyse past trends and its correlation with environmental and anthropic factors and to forecast fire risk in the context of climate change, considering the evolution of environmental parameters stated from the Intergovernmental Panel on Climate Change (IPCC, 2022). These findings can be the starting point for the development of forecasting models also with a view to proposing prevention and protection strategies for the archaeological heritage of the Basilicata Region.

Reference

IPCC, 2022: Climate Change 2022: Impacts, Adaptation, and Vulnerability. Contribution of Working Group II to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change [H.-O. Pörtner, D.C. Roberts, M. Tignor, E.S. Poloczanska, K. Mintenbeck, A. Alegría, M. Craig, S. Langsdorf, S. Löschke, V. Möller, A. Okem, B. Rama (eds.)]. Cambridge University Press. Cambridge University Press, Cambridge, UK and New York, NY, USA, 3056 pp., doi:10.1017/9781009325844.

How to cite: Danese, M., Florio, V., Masini, N., and Lasaponara, R.: Impact of fire risk on archaeological heritage in the Age of climate change. Geodata science for prediction and development of strategies for protection., EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-16429, https://doi.org/10.5194/egusphere-egu25-16429, 2025.

When discussing climate change and cultural heritage, the focus often lies exclusively of the vulnerability aspects of the latter. However, cultural heritage can also play an active role in activating strategies and actions to increase its sustainability and mitigate environmental impacts.

Energy rehabilitation and reuse of existing buildings hold the potential to contribute to sustainable heritage conservation while embracing new energy efficiency principles.

According to literature, energy rehabilitation and retrofitting of the building envelope need to be carried out with respect to historic and cultural features and the protection of cultural heritage. This applies as much to listed buildings as to those that, although not formally protected, are part of the historical heritage and define the identity and the skyline of the place (Magrini, Franco, 2016).

In this work, starting from the spatial modeling of the territory and use of satellite data thank to the free-cloud application Google Earth Engine (GEE), it is possible to perform some preliminary analysis. These ones are useful to derive some formal characteristics that directly influence both the energy requirements and the choice of some technological solutions for integrating renewable energy sources (Forster et al.,2025).

References

Forster et al.,2025: Forster, J., S. Bindreiter, B. Uhlhorn, V. Radinger‐peer, and A. Jiricka‐pürrer. 2025. 'A Machine Learning Approach to Adapt Local Land Use Planning to Climate Change', Urban Planning, 10.

Magrini, Franco, 2016: Magrini, A., and G. Franco. 2016. 'The energy performance improvement of historic buildings and their environmental sustainability assessment', Journal of Cultural Heritage, 21: 834-41.

How to cite: Florio, V., Danese, M., and Biscione, M.: Preliminary analysis for energy efficiency assessment. Deriving technical parameters with spatial analysis and GEE, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-16568, https://doi.org/10.5194/egusphere-egu25-16568, 2025.

Monitoring key parameters that drive land-surface processes, such as surface soil moisture (SSM)), is essential for understanding global biogeochemical cycles, including those of water, energy, and carbon. While Earth Observation (EO)-based SSM products have demonstrated significant potential, their practical application is often limited by coarse spatio-temporal resolution. Therefore, downscaling these operational products is a critical scientific challenge for enabling their effective use in regional and local-scale applications.

This study’s aims at presenting an innovative approach for downscaling operational soil moisture products using a variant of the so-called “triangle” method, named the “simplified” triangle. The use of the proposed technique is demonstrated herein using the European Space Agency's (ESA) operational soil moisture product from the Soil Moisture and Ocean Salinity (SMOS) and optical data from ESA’s Sentinel-3 platform. The enhanced spatial SSM estimates are compared against near collocated reference ground data from multiple validated experimental sites across Europe. The results obtained indicate a satisfactory agreement, confirming the proposed approach's promising potential to accurately estimate key land-surface interaction parameters. Conceptually the proposed herein methodological framework is applicable to any operational product, a topic of further investigation.

How to cite: Detsikas, S. E., Petropoulos, G. P., Saviolakis, P., Lekka, C., Karymbalis, E., Katsafados, P., and Georgaki, F.: Downscaling Earth Observation Operational Soil Moisture Products Using multi-sensor Satellite Data: “A Triangle Inversion Approach", EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-19591, https://doi.org/10.5194/egusphere-egu25-19591, 2025.

Advances in geo-information technologies, including Earth Observation (EO), GIS, cloud computing and software tool development, have shown great potential towards addressing key societal challenges faced today associated with the study of land-atmosphere interactions. Accurate information on spatially explicit, distributed estimates of land-atmosphere fluxes and soil surface moisture is essential in a wide range of disciplines, including meteorology, hydrology, agriculture and ecology.

Use of simulation process models has played a key role in extending our abilities to study Earth system processes and enhancing our understanding on how different components of it interplay. A special category of such models includes the so-called Soil Vegetation Atmosphere Transfer (SVAT) models. Those are deterministic simulation models that describe the physical processes controlling energy and mass transport in the soil/vegetation/atmosphere system.

SimSphere is such a software toolkit written in Java for simulating the interactions of soil, vegetation and atmosphere layers of the Earth’s land surface. Its use is at present continually expanding worldwide both as an educational and as a research tool for scientific investigations. It is being used either as a stand-alone application or synergistically with EO data and important advancements particularly in the recent years have been implemented to the model.

Herein, we present state of the art advancements introduced recently to SimSphere SVAT model aiming at making its use more robust when integrated with EO data via the so-called “triangle” method. Use of the recently developed add-on to SimSphere is illustrated herein using a variety of examples that involve both satellite and UAV data. The presented work  is of key significance to the users' community of the model and very timely, given that variants of the so-called “triangle” method being considered for deriving operationally regional estimates of energy fluxes and soil moisture from EO data provided by non-commercial vendors.

KEYWORDS: land surface interactions, geoinformation, earth observation, triangle, SimSphere   Acknowledgements The research presented herein has been conducted in the framework of the project LISTEN-EO (DeveLoping new awareness and Innovative toolS to support efficient waTer rEsources man- agement Exploiting geoinformatiOn technologies), funded by the Hellenic Foundation for Research and Innovation programme (ID 15898). 

How to cite: Gkatzios, G., Petropoulos, G. P., Detsikas, S. E., Lekka, C., Karymbalis, E., and Katsafados, P.: Advancing our understanding of land surface interactions via the development of innovative geoinformation tools , EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-21175, https://doi.org/10.5194/egusphere-egu25-21175, 2025.

Powerful volcanic eruptions inject into the stratosphere sulphur and tephra that may be spread globally and affect the Earth’s climate. Over the last 2500 years, Sigl et al. (2015) made a synthesis of the polar ice core atmospheric sulphur record and climate anomalies from dendrochronological records. Aside from a few historical events, most large eruptions with a bipolar imprint and a significant climate anomaly are from the tropical latitudes, but their sources are unknown.

We analysed the micron-size crytotephra composition accompanying the (stratospheric) sulphate of the 1458 CE and 426 BCE volcanic events recorded in three Antarctic ice cores. The 1458 CE event occurred within a cool climate and was initially attributed to the Kuwae (Vanuatu) eruption. This link is however questioned by Hartman et al. (2019) from their study of a South Pole ice core. The 426 BCE event appears concomitant with a significant global climate cooling, but its source is unknown.

Within the sulphate peak, the particle size distribution, when available, helps documenting the dynamics of the arrival of the stratospheric plume. Cryptotephra are collected by filtration and after carbon-coating, analysed by an EPMA microprobe. We applied the analytical procedure of Narcisi et al. (2019) (who identified the 1257 CE Samalas eruption), adapted to the micron-size of the crytotephra.

For the 1458 CE event, a medium-K dacite to rhyolite composition is consistently observed from Vostok and Dome C ice core samples (218 values). The dacite patch (SiO2~68%) fits well the composition of proximal Kuwae deposits as well as that of an ash layer (~140 values) on Efate Island (Standberg et al, 2023). The rhyolite composition patch (SiO2~72%) is unlikely from a South American source, but appears discretely represented in proximal Kuwae deposits as well as in sediments in the nearby Epi Submarine zone. We suggest that rhyolite is a daughter product from dacite by evolving in the upper layers of the magmatic chamber, and it was spread out first and far away by the eruption.  

 For the 426 BCE event, the cryptotephra composition (220 values) is consistently found within the three ice cores (Vostok, Dome C, Talos Dome) and belongs to high-K rhyodacite. Coincidentally such composition is very close to Kuwae’s (except for higher K) suggesting it was issued from a very similar magmatic chamber. The 10 km wide Ambrym caldera located 50 km north of Kuwae, collapsed ~2000 years ago appears the best candidate. 

References

Hartman et al. (2019). Nature Sci. Rep. 9. https://doi.org/10.1038/s41598-019-50939-x.

Narcisi, B. et al., 2019. Quat. Sci. Rev. 210, 164-174 https://doi.org/10.1016/j.quascirev.2019.03.005.

Sigl, M., et al. Nature 523, 543–549 (2015). https://doi.org/10.1038/nature14565

Strandberg NA et al., (2023). Front. Ecol. Evol. 11: 1087577.doi: 10.3389/fevo.2023.1087577

How to cite: Petit, J.-R., Savarino, J., Delmonte, B., Gautier, E., Ginot, P., and Batanova, V.: Cryptotephra fingerprinting of 1458 CE and 426 BCE volcanic events in East Antarctic ice cores, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-2329, https://doi.org/10.5194/egusphere-egu25-2329, 2025.

The unique mass-occurrence of tiny heteromorph ammonites found in a single layer in the Mt. Cipollara locality of Cerreto d'Esi (Maiolica Formation, Umbro-Marchean Basin, Italy) provides critical insights into the depositional environments of the Cretaceous upper Maiolica Formation. The mechanisms behind the formation and preservation of these ammonite assemblages within black shales remain poorly understood. To solve this knowledge gap, we constrained by means of biostratigraphy and stable Carbon isotopes the whole sections hosting the ammonites-rich layer. The latter was then subjected a high-resolution paleoecological analyses. Samples were collected systematically across multiple stratigraphic levels to ensure comprehensive coverage. The ammonite assemblages were documented, focusing on their morphology, abundance, and associated sedimentary structures. Additionally, sedimentological petrographic examinations were conducted to elucidate depositional processes. Our results reveal a rich assemblage dominated by the family Leptoceratoididae, exhibiting relatively good preservation within a predominantly dysoxic low-energy environment at the bottom. Calcareous nannofossils data suggest the presence of a well-stratified water column, with a low salinity water cap. The multidisciplinary analyses indicates that these black shales served not only as a repository for ammonite remains but also reflected localized paleoecological conditions characterized by reduced turbulence and increased organic deposition. This unique sedimentary context suggests that the deposition of these assemblages could have been influenced by both regional sea-level fluctuations and local hydrographic conditions. In conclusion, the study of the Mt. Cipollara heteromorph ammonites underscores the complexity of Cretaceous paleoenvironments and provides an enhanced understanding of the occurrence of black shales within the Maiolica Formation.

How to cite: Conti, C., Faraoni, P., Mancini, A. M., Luca, M., and Negri, A.: Unusual mass-occurrence of small, uncoiled ammonites in a black shale of the Maiolica Formation in the Umbria-Marche Basin (Central Italy), EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-3477, https://doi.org/10.5194/egusphere-egu25-3477, 2025.

Aeolian mineral dust and diatom influxes at the summit of Roosevelt Island (79.364°S, 161.706°W, 550 m a.s.l.) were investigated over the last 2 kyrs from the RICE ice core (Bertler et al., 2018). Mineral dust at the site is mainly related to large-scale atmospheric circulation patterns within the Eastern Ross and Amundsen Seas, while aeolian diatoms, mainly consisting of Fragilariopsis spp. (F. nana , F. cylindrus, , F. curta), depend on the local oceanic influence of air masses from the marine boundary layer. Thus, the complementarity of these proxies allows appreciating climatic and atmospheric changes experienced at Roosevelt Island over the last 2000 years, in response to some major forcing factors such as ENSO. During the 550-1470 CE period, when higher/less depleted stable water isotope values are observed, the increased importance of blocking ridges in the Amundsen Sea and a weakened Amundsen Sea Low promoted dust-rich air mass advection to RICE. This pattern was accompanied by an increasing trend in snow accumulation and reduced sea ice in the Eastern Ross and Amundsen Seas. At about 1300 CE, the maximum expression of the Ross Sea dipole is reached, with enhanced katabatic outflow in the Western Ross Sea and reactivation of the Ross Sea polynya. At the same time,  the Eastern part of the Ross Sea was still under the influence of blocking ridges promoting maritime air mass advection to RICE and southward shift of the South Westerly Winds. After 1470 CE, unprecedented peaks of aeolian diatom concentration suggest a rapid reorganization of local atmospheric circulation, that probably occurred in relation to the eastward enlargement of the Ross Sea polynya culminating with the opening of the  Roosevelt Island polynya.
For the RICE site, we suggest that several drivers contribute to the long-term dust, sea-ice and polynya variability, but ENSO-driven teleconnections are particularly prominent. On a longer (multidecadal) timescale it seems that El Niño-dominating conditions promoted the establishment of the Ross Sea dipole, while La Niña conditions favored a deeper Amundsen Sea Low and an eastward expansion of the polynya. 

How to cite: Delmonte, B., Lagorio, S., Tetzner, D., Malinverno, E., and Bertler, N.: Aeolian dust and diatoms at Roosevelt Island (Ross Sea, Antarctica) over the last two millennia reveal the local expression of climate changes and the history of the Ross Sea polynya., EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-8574, https://doi.org/10.5194/egusphere-egu25-8574, 2025.

Drought significantly affects the growth and physiological responses of Zagros forests, one of the most important natural habitats in Iran. The Brants oak (Quercus brantii Lindl), a widely distributed and dominant tree species in the Central Zagros Mountains of western Iran, serves as a valuable natural archive for studying historical climate variability and ecological changes. For climate-growth analysis, 30 Q. brantii trees cored from Lordegan area (1820 to 2280 m a.s.l.) in the southwest of Zagros forests of Iran. After preparing the samples, measuring the tree ring widths and cross-dating developed the tree ring chronology (1710-2023) using dplR. The relationships between tree-ring widths (TRW) and monthly mean temperature and precipitation values and the Standardized Precipitation Evapotranspiration Index (SPEI) were analyzed. The strongest climate signal of SPEI was found from previous September until April, representing the pre-growing and early-growing seasons. Among these reconstructions were acknowledged extremely narrow rings in 1870, 1923, 1960, 1964, and 2018, while extremely large rings were found in 1784, 1852, 1957, and 1976. Based on preliminary calculations showing a strong winter SPEI signal, this chronology could be used for climate reconstruction, but further analysis is required. These studies indicate the vulnerability of oak forests in the Zagros Mountains to ongoing climate change and a pressing need for sustainable forest management strategies to preserve these vital ecosystems.

 Keywords: Zagros forests, Iran, Quercus brantii, reconstruction, tree-ring widths

How to cite: Hatami, F., Klesse, S., Treydte, K., Verstege, A., Etemad, V., Pourtahmasi, K., Gessler, A., and Iranmanesh, Y.: Potential for a 300-year drought reconstruction in the Zagros Mountains, Iran based on the tree-ring width of Quercus brantii Lindl. , EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-10067, https://doi.org/10.5194/egusphere-egu25-10067, 2025.

Oases are critical locations for human survival in desert areas. With heavily reliant on runoff from the surrounding mountains, oases in the hyperarid Tarim Basin are especially fragile and sensitive to both climatic–environmental changes and human activity. However, the local evolution process of oases in desert area remains unclear due to strong erosion and contemporaneous heterogeneity, which restricts our understanding of the coupling relationship among climate change, oasis evolution, and human activity. Here, we reconstruct the evolution of oases since the last deglacial period (~15 ka) in the Tarim Basin. The results indicate that oases advanced during 15–11.5 ka, 9–4.5 ka, and 2–1 ka. Through the integration of multiple records of palaeoclimate, palaeoenvironment, archaeology and history, we found that human activity dominated and decoupled oasis evolution from climate change since ~2 ka in the Tarim Basin. Oasis were artificially expanded to sustain the flourishing society during the Han–Tang period, but they declined synchronously afterward. More attention should be paid to the proper management of land and water resources to achieve sustainable development in hyperarid areas.

How to cite: Sun, A.: Human activity has decoupled oasis evolution from climate change since ~2 ka in the Tarim Basin, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-10215, https://doi.org/10.5194/egusphere-egu25-10215, 2025.

While the effects of OAEs are well known for the pelagic successions of the Tethys Ocean, little is known about their impact on the Peri-Tethyan shallow water carbonate systems. Here we present the preliminary results of a study related to the geological mapping of the sheet 390 – Frosinone of the Geological Map of Italy (CARG Project), focussed on the identification and description of the perturbation induced in the Lower Cretaceous shallow water carbonate succession of the Latium-Abruzzi Carbonate Platform by the well-known Early Cretaceous Oceanic Anoxic Events (OAEs).

In the Ernici Mts. (central Apennines, Italy), an Upper Triassic to Upper Cretaceous shallow-water carbonate succession is exposed (Cosentino et al., 2010; Fabbi et al., 2023). This study specifically examines the Lower Cretaceous "calcari ciclotemici a gasteropodi" fm. (CCG - Berriasian p.p. - lower Aptian p.p.), which mainly consists of whitish limestones with intercalations of light grey dolostones. Within this succession, a layer of black dolostone, about ten centimetres thick, has been observed in several outcrops of the dolomitic lithofacies (CCGa) of CCG, at the same stratigraphic position.

Two stratigraphic sections were measured to characterise the microfacies and compositional variations observed between the light-coloured (whitish to light grey) and black layers. SEM images, along with Energy-Dispersive X-ray Spectroscopy (EDS) and Wavelength-Dispersive X-ray Spectroscopy (WDS) analysis indicated the presence of siderite and pyrite aggregates (Meng et al. 2024). These aggregates appear in high concentration starting from the basal part of the blackish dolostone layer and gently decrease towards the upper part of the study interval. TOC and sulphates show similar trends.

Changes in chemical composition between the whitish and blackish dolostones (CCGa) were investigated in situ using the laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) facility at Roma Tre University. The results show a significant increase in elemental concentration of P, Fe, Zn, As, Ba, Pb, and U, as well as in the Fe/Al ratio in the blackish dolostones. These elements are generally considered as redox-sensitive proxies associated with anoxic paleoenvironments (Bodin et al., 2007; Craigie, 2018).

Biostratigraphic calibration performed on the collected samples has established a Hauterivian p.p. age for the investigated CCGa levels. A preliminary attempt for U-Pb dating of the CCGa black dolostone was carried out through LA-ICP-MS investigations at Roma Tre and ETH facilities. In the Tera-Wasserburg diagram, the U-Pb measurements on CCGa black dolostone yielded a lower intercept age of 125.7± 1.8 Ma (MSWD=1.6; N=19). These promising results suggest that the changes in the elemental concentration of the redox-sensitive proxies observed in the CCGa black dolostone were induced by the late Hauterivian Faraoni Oceanic Anoxic Event.

How to cite: Artegiani, F., Cipollari, P., Cosentino, D., Rabiee, A., Guillong, M., Rossetti, F., Cipriani, A., and Fabbi, S.: Early Cretaceous Oceanic Anoxic Events (OAEs) in Peri-Tethyan shallow-water carbonate systems: Evidence from the Latium-Abruzzi Carbonate Platform (Ernici Mts, Central Italy), EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-11624, https://doi.org/10.5194/egusphere-egu25-11624, 2025.

Located in the arid inland of Asia, the eastern part of the Silk Road is marked by certain routes being close to or even crossing large deserts, such as the Taklimakan Desert, one of the largest deserts worldwide. As a result, sand and dust activities have a considerable impact on the transport routes, the desert-oasis ecosystem, and human society along the Silk Road. However, the evolution of dust activity over the past two millennia and its relation to the changes of the Silk Road civilization remains ambiguous. Here, we present a high-resolution (~3 yr) stalagmite record from Xinjiang (northwest China) spanning the past 2,500 years, dated with 19 U/Th ages. Although the stable isotopes and trace elemental ratios of the stalagmite reveal remarkable decadal- to centennial-scale variability of the regional hydroclimate, the Mg/Ca ratio shows a quite different variation pattern compared with other geochemical proxies. Considering various factors that might influence the Mg/Ca ratio of stalagmites, our analysis reached the conclusion that the geographical location close to the desert made the imported dust likely to predominate the increase of Mg/Ca in stalagmites during many characteristic periods. For instance, we found significant increases in the Mg/Ca ratios lasting for more than two centuries during approximately 650-850 CE and 1650-1950 CE (i.e., the Little Ice Age). This generally demonstrates a pattern of reduced dust activities during the Medieval Warm Period and enhanced dust activities throughout the Little Ice Age, which is supported by evidence from the eolian sedimentary section in the southern margin of the Taklimakan Desert that directly reflects dust activity. We further found that the enhanced dust activity during the 650-850 CE might have caused the route shift of the Silk Road from south to north in the Tarim Basin. In addition, the rapid drying of Lop Nur in recent decades could also be attributed to abnormally increased dust activity, as this period was characterized by the most intense dust activity in our records over the last 2,000 years. Our findings further substantiate the argument regarding the association between societal and climatic change along the Silk Road, where the dust production from large deserts poses challenges to sustainable development in the present and the future.

How to cite: Liu, X., Chen, S., Chen, J., Wang, H., Shen, C.-C., Wang, X., and Chen, F.: Reconstruction of dust activity using geochemical proxy from cave stalagmite in the northern Taklimakan Desert, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-12114, https://doi.org/10.5194/egusphere-egu25-12114, 2025.

The Canary Islands, located in the central North Atlantic, provide an exceptional setting for investigating long-term climate dynamics within the Macaronesian region. This study presents sedimentary records from volcanic lacustrine basins across Tenerife, La Gomera and La Palma, analyzed using a multi-proxy approach including magnetic susceptibility, XRF geochemistry, elemental composition (TOC, TN, TS), mineralogy, lipid biomarkers, and updated age models. Preliminary age models suggest that the sequences of La Vega Lagunera (northern Tenerife) extend back up to 400,000 years, and El Malpaís de La Rasca (southern Tenerife), Garajonay (La Gomera), and Playa de Taburiente (La Palma) span the Holocene.
Preliminary results from La Vega Lagunera, a Pleistocene clastic lake, indicate colder conditions during MIS 2 and MIS 4, warmer conditions during the Holocene, MIS 3, and MIS 5, and millennial-scale cycles during MIS 3 and MIS 4. Climate during the Last Glacial Maximum (MIS 2) was notably drier, resembling mid-latitude records. 
Holocene records from paleolacustrine deposits of two closed-drainage basins located in two volcanic craters (La Gomera and Malpaís de La Rasca) and the lacustrine-marsh system of Playa de Taburiente showed coherent patterns of Holocene regional climate variability, with increased fluvial and alluvial activity during the Greenlandian (11.7 to 8.2 ka), a decline during the Northgrippian (8.2 to 4.2 ka), and reduced clastic input during the Meghalayan (last ~4.2 ka). These trends suggest increasing aridity throughout the Holocene. 
These new sedimentary sequences from Tenerife, La Gomera, and La Palma provide further evidence of rapid climate dynamics during glacial and interglacial intervals. Improved age models (OSL, 14C) are still being developed to characterize the cyclic patterns better, while multi-proxy analyses are enhancing our understanding of past climate dynamics. Further research is needed to clarify the roles of regional climate and local factors.
This work is supported by TED2021-129695A-I00 project funded by MICIU/AEI/10.13039/501100011033 and by the European Union NextGenerationEU/PRTR; PALEOMOL (2915/2022) and IVRIPARC (2779/2021), both funded by the Spanish National Parks Organism, and IMPACT (2022CLISA04, Fundación CajaCanarias and Fundación La Caixa).

How to cite: Ocón-Bermúdez, C., Galofre-Penacho, M., Valero-Garcés, B., Armenteros-Armenteros, I., Herrera-Herrera, A., Égüez, N., Martín-Luis, M. C., Casillas Ruiz, R., Vegas, J., Castellano-Rotger, L., Diez-Herrero, A., Casado-Vara, R., and Jambrina-Enríquez, M.: A Multi-Proxy Approach to Reconstructing Long-Term Climate and Environmental Dynamics in the Canary Islands: Inter-Island Comparisons, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-12407, https://doi.org/10.5194/egusphere-egu25-12407, 2025.

Human-induced carbon emissions are altering the modern climate, with severe repercussions on ecosystems. Among others, anthropogenic pressure is causing deoxygenation of the bottom water, with the widespread establishment of hypoxic zones in several Mediterranean areas. The geological archives allow the investigation of past deoxygenation dynamics (sapropel events) and their impact on marine ecosystems. Here, we compare the causes and the evolution of deoxygenation dynamics that occurred during two different time periods (Messinian and Holocene) in different paleoceanographic settings based on their micropaleontological content. The Messinian sapropel events are the result of increased export productivity during a relatively cold and arid context, triggering bottom anoxic conditions. The Holocene sapropel formed in response to weakening/stopping of the thermohaline circulation due to increasing temperature and freshwater input. Our results suggest that the deoxygenation dynamics in the Mediterranean in the near future will not follow the trend characteristic of the Holocene deep-sea sapropel because of the predicted drying trend. Differently, the paleoceanographic setting triggering the Messinian shallow-sea sapropels is comparable with the modern situation in different Mediterranean areas, where human-induced eutrophication is promoting deoxygenation. Based on these results, we suggest that the patchy deoxygenation trend in the Mediterranean Sea caused by climate warming may lead to a drastic change in the ecosystem services which would likely impact human activities.

How to cite: Lozar, F., Mancini, A. M., Morigi, C., Gennari, R., and Negri, A.: Past analogues of deoxygenation events in the Mediterranean Sea: Comparison between shallow and deep settings , EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-15403, https://doi.org/10.5194/egusphere-egu25-15403, 2025.

The Pliensbachian/Toarcian event (P/T-E) and the Toarcian Oceanic Anoxic Event (T-OAE) are two intervals of carbon cycle perturbations linked to massive ¹²C-enriched carbon emissions, causing severe biotic and environmental changes. Here organic carbon isotope, mineralogical composition and sedimentology have been analyzed across the Pliensbachian-Toarcian transition from the Monte Serrone section (Umbria-Marche Basin), which was deposited in a pelagic setting in the western Tethys. A marked negative carbon-isotope excursion occurred across the Pliensbachian-Toarcian boundary and lower Toarcian, respectively, which can be used to identify PTE and T-OAE in the study area. The P/T-E and T-OAE intervals witnessed carbonate production crisis revealed by reduced carbonate contents. We hold that the 0.5 m-thick laminated black shales indicated that the T-OAE was a highly condensed succession because it included the full duration of the T-OAE. Therefore, the T-OAE interval at Monte Serrone coincided not only with diminished carbonate production but also with reduced siliciclastic input, forming quite thin black shale deposition. Abundant marine organisms were present preceding the T-OAE. Nevertheless, none of them survived during the most negative carbon-isotope excursion of the T-OAE, revealing a biotic crisis at this time. Elevated seawater temperature could induce this crisis in the study area. The recovery of benthic foraminifera was delayed at Monte Serrone.

How to cite: Nie, Y., Fu, X., and Rigo, M.: Carbon cycle perturbations during the Pliensbachian-Toarcian transition in the Monte Serrone section (Northern Apennines, Italy), EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-16338, https://doi.org/10.5194/egusphere-egu25-16338, 2025.

The climatic signals recorded by loess sequences vary between different regions, which makes it important to study loess sequences worldwide. The loess deposits in northern Iran are situated in the transitional zone between the European loess and Central Asian loess. However, the depositional dynamics and paleoenvironmental significance of the loess deposits in this region are not well understood, making it difficult to establish detailed correlations with loess deposits elsewhere, partly due to the lack of systematic and high-resolution chronological control. We used K-feldspar pIR₅₀IR₂₉₀ and MET-pIRIR₂₅₀ luminescence dating protocols to date fifty-two K-feldspar samples from the Toshan-19 section in the northern foothills of the Alborz Mountains, northern Iran. These chronological data, along with the climate proxies of magnetic susceptibility and redness, combined with a comparison with published loess records from various regions, indicate the following: (1) K-feldspar luminescence ages obtained using pIRIR and MET-pIRIR protocols are consistent, and their luminescence ages up to ~200 ka are deemed dependable. The loess at Toshan was primarily deposited during 78–24 ka, corresponding to MIS 4–2, and the paleosols developed during 139–78, and 24–1.7 ka, corresponding respectively to MIS 5, and late MIS 2–MIS 1. (2) Drier conditions prevailed during the last glacial and wetter conditions dominated during the last interglacial. Moisture variations during the substages of MIS 5 in this region indicate cold-dry and warm-wet climatic characteristics. The reasons for increased moisture from late MIS 2 onwards in this region still require further investigation. (3) The loess-paleosol records indicate a consistent pattern of climate change over Eurasia on the scale of the last interglacial-glacial cycle. During the substages of MIS 5, warm-wet and cold-dry conditions in northern Iran were in-phase with those on the Chinese Loess Plateau, Europe, and southern Tajikistan; however, they were anti-phased or out-of-phase with those in Xinjiang.  

How to cite: Li, D., zhao, H., and Xie, H.: Loess-paleosol sedimentological characteristics in northern Iran since the last interglacial and their paleoenvironmental significance, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-16630, https://doi.org/10.5194/egusphere-egu25-16630, 2025.

The Lagoa Real uranium province (LRUP) is the main Brazilian target for uranium. Their geochronological studies began in the 80s and provided controversial ages for mineralization. Since then, advances in geochronological methods, increased local petrological data, and knowledge of the uranium cycle have helped geosciences understand crust and mantle behavior over time. As a result, recent geochronological studies developed by CDTN researchers have now begun to reinterpret the evolution of the LRUP.

These studies dated metasomatic  U-ore bodies providing ages between 545 Ma to 520 Ma (in situ U–Pb dating of andradite and titanite, Santos et al., 2023; Journal of South America Earth Science) coeval with the late Pan-African Cycle. Geochronological studies were also carried out on the host rocks (A-Type granites) of the mineralized bodies, providing ages between 1762 Ma to 1741 Ma (U–Pb dating of magmatic Zircon, Amorim et al., 2022; Journal of South America Earth Science), coeval to bimodal magmatism well documented in Brazil and Africa.

Although some of the data obtained suggest that granites might not be the source of uranium, their volcanic expression (metaryolites located to the NW of the LRUP) could be a good candidate. Thus, the uranium mobilization began before the metasomatism, through magmatic processes, coeval with the Post-Archean Uranium Recycling, a global event that incorporates U in the crust from the mantle. Furthermore, preliminary macroscopic and microscopic data from gneisses show evidence of partial melting related to regional metamorphism that may have occurred before metasomatism. This process generated Neoproterozoic uranium deposits in Namibia, at the Southern of the African counterpart of Brazil. Therefore, LRUP could result from overlapping processes in central Brazil accompanied by crustal differentiation episodes leading to a polycyclic evolution.

How to cite: Azevedo, R. A. and Rios, F. J.: The Lagoa Real Uranium Province: polycyclic evolution in the Brazilian geochronological record, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-19922, https://doi.org/10.5194/egusphere-egu25-19922, 2025.

Heatwaves and extreme precipitations are the two prevalent types of weather-related extreme events globally. Compared with univariate extremes, impacts of compound extreme precipitations preconditioned by heatwaves (CHEPs) on the society and economy can be amplified. Previous studies demonstrated that heatwaves can trigger extreme precipitations by enhancing atmospheric instability and moisture-holding capacity. Other studies projected future changes in CHEPs under various greenhouse gas emission scenarios. However, there is a lack of studies assessing the time of emergence (ToE) of CHEP change signals, especially for record-shattering events. Since current water resource management strategies and infrastructures are based on historical data, it is crucial to understand when hydro-meteorological conditions will surpass unprecedented levels to develop effective adaptation and mitigation strategies for climate change.

Here, we present a global analysis of ToE for record-shattering CHEPs as well as their exposed GDP and population (POP). Both the frequency and magnitude of observed CHEPs have substantially increased during the past 65 years at the global scale. Using climate models from Detection and Attribution Model Intercomparison Project, we find that rarer CHEPs are increasingly attributable to anthropogenic greenhouse gas emissions, while aerosol emissions have a mitigating effect on their occurrences. To detect when historical record-shattering events will become normal, we develop a novel framework based on advanced Single Model Initial-condition Large Ensemble simulations. Our results indicate that CHEP hotspots, including East and Southeast Asia, North-central South America, and Central Africa, are likely to experience earlier ToE compared to other regions. In contrast, arid regions, such as North Africa, West Asia, and southwestern Australia, show no signs of ToE until at least 2100. GDP and POP exposure to such events reveal an alarming upward trend throughout the 21st century. By the late 21st century, 41% (29%) of sub-regions defined by the Sixth Assessment Report of the Intergovernmental Panel on Climate Change are projected to experience GDP exposure exceeding 4,000 billion USD (at 2010 purchasing power parity) to record-shattering frequency (magnitude), while 34% (27%) are expected to have POP exposure exceeding 100 million under the SSP2-4.5 scenario. Record-shattering CHEPs pose a distinct threat to the economy between 21.75°N and 53.25°N, with the most significant impact between 35.25°N and 39.75°N. Compared to the GDP exposure, the POP exposure hotspots shift toward lower latitudes, with a broader range extending from 0.75°S to 53.25°N. Additionally, we classify areas based on the Human Development Index and income levels defined by the World Bank. The unequal distribution of GDP and POP exposure reveals the poorest and least developed countries will experience more extended impacts compared to wealthier nations. This study highlights the urgent need for region-specific mitigation and adaptation strategies to combat climate change, especially for the vast high-risk and low-income regions.

How to cite: Liu, J., Chen, J., and Yin, J.: Time of Emergence of Record-shattering Compound Extreme Precipitations Preconditioned by Heatwaves and Their Socio-economic Exposures, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-2350, https://doi.org/10.5194/egusphere-egu25-2350, 2025.

Consecutive heatwave and heavy rainfall (HW-HR) events are occurring with increasing frequency in a warming climate. The time interval affects both environmental conditions and the regional recovery between two consecutive extreme events. However, the dynamics of the transition between consecutive HW-HR events remain poorly understood. In this study, we examine the changes in the time interval of consecutive HW-HR events in China from 1990 to 2019, using meteorological data from over 2,000 stations across mainland China. Our results reveal that the time interval has significantly shortened at 28.2% of the stations. The increased proportion of short-time events (STEs), defined by consecutive events with time intervals of 1 to 2 days, is the primary driver of this trend. From 1990 to 2019, the proportion of STEs increased significantly, at a rate of 2.2% per decade. We also find that climate change-induced anomalies in atmospheric variables during the consecutive HW-HR events may contribute to this rise in the proportion of STEs. Additionally, we assess changes in population exposure to STEs over the past two decades. Exposure has increased at more than three-quarters of the stations, with the increased STEs contributing to over 80% of the rise in exposure. Our findings highlight the need for policymakers to prioritize disaster response during consecutive HW-HR events and implement effective risk management strategies to mitigate population exposure to extreme events.

How to cite: Li, J., Wang, S., Zhu, J., Wang, D., and Zhao, T.: Accelerated Shifts from Heatwaves to Heavy Rainfall in a Changing Climate, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-2385, https://doi.org/10.5194/egusphere-egu25-2385, 2025.

Winter weather events can result in costly damages and severe disruption to affected regions. While compound events research has strongly focused on heat-related events, less focus has been placed on extreme cold hydrometeorological hazards. Cold events impact a range of sectors from energy and agriculture to transport and health. The rail sector is particularly sensitive to cold weather hazards resulting in service delays and cancellations. Snowfall can lead to blocked tracks, points failures and issues with electricity supply. Loss of traction, braking issues and frozen infrastructure can arise from ice formation. The impacts of these cold events can be amplified by the compounding effect of another meteorological variable, such as whether heavy precipitation is present or not, with subsequent impacts dependent on the nature of the cold event. For example, a cold-wet event could incur heavy snowfall, whereas a cold-dry event could result in extreme low temperatures and icy conditions. In this study, we analyse the occurrence of 10,000 rail incidents in Scotland over an extended winter period of October to March for 2006-2023 to investigate the relationship between impacts and compound cold events. We use an impact dataset from Network Rail to categorise high-impact days based on two classifications: (1) days with the highest number of aggregated incidents; (2) days with the highest number of accumulated customer minutes lost. Using daily gridded observations from HadUK-Grid at a 5 km resolution we then apply a localised percentile-based methodology to determine the occurrence of cold-dry and cold-wet events on these high-impact days. Initial results show that the majority of high-impact days consisted of incidents caused by severe snow and icing. Analysis reveal that these incidents occurred under hydrometeorological conditions that can be classified as cold-dry and/or cold-wet events. These findings highlight the importance of considering co-occurring hazards rather than single hazards. The results of this study provide a useful insight into compound cold events for rail sector early warning systems, with valuable information on cold weather event hazard characterisation and their associated impacts across varying timescales.

How to cite: Mattu, K., White, C., Bloomfield, H., and Robbins, J.: Investigating the relationship between compound cold events and impacts on the Scottish rail sector. , EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-4810, https://doi.org/10.5194/egusphere-egu25-4810, 2025.

Probabilistic flood risk assessments (PFRA) and flood early warning systems (FEWS) are essential tools for managing and responding to disastrous flood events, particularly in coastal deltas where flooding is often compound, resulting from the interplay of coastal and riverine water levels and local rainfall. PFRA and FEWS require assessing the compound flood hazard under a broad range of plausible or forecasted hydro-meteorological conditions. While efficient hydrodynamic models for compound flooding have been developed, such as SFINCS (Leijnse et al. 2021; van Ordmondt et al. 2024), trade-offs in the model resolution or number of simulations or stochastic variables are often required for PFRA and FEWS, at the cost of model accuracy.

The physics-guided  hybrid  LSG model (Fraehr et al. 2022, 2023) uses a Sparse Gaussian Process model trained on Empirical Orthogonal Functions (EOF) derived from simulations with a high- and low- resolution hydrodynamic model. For new events to simulate, the approach combines a simulation of the low-resolution hydrodynamic model with the trained surrogate model, to predict high-resolution water depths at low computational costs. While this model has successfully been applied for riverine flooding, it has not yet been used to predict compound flooding from multiple drivers.

This study tests the surrogate SFINCS-LSG model for compound PFRA and FEWS. We investigate the optimal choice of events to train the model and test the model for case studies in Brisbane, Australia and Charleston (NC), USA. We validate the surrogate model against the high-resolution SFINCS model for different historical compound events, hypothetical compound flood scenarios, and compound PFRA.

Based on preliminary results, we find that compared to the course-resolution SFINCS model, the surrogate model provides a significant improvement at very low computation costs, while compared to the high-resolution SFINCS model it achieves a large speedup with only a small drop in accuracy. While the results are promising for individual simulations, the surrogate model struggles to capture the transition zone based on the difference between model simulations. Nonetheless, the surrogate SFINCS-LSG models seem a promising approach to improve compound PFRA and FEWS.

References

Fraehr et al. (2022). Upskilling low-fidelity hydrodynamic models of flood inundation through Spatial analysis and Gaussian Process learning. WRR, https://doi.org/10.1029/2022WR032248

Fraehr et al (2023). Development of a fast and accurate hybrid model for floodplain inundation simulations. WRR, https://doi.org/10.1029/2022WR033836

Leijnse et al. (2021). Modeling compound flooding in coastal systems using a computationally efficient reduced-physics solver: Including fluvial, pluvial, tidal, wind- and wave-driven processes. Coastal Engineering, https://doi.org/10.1016/j.coastaleng.2020.103796

van Ormondt et al. (2024). A subgrid method for the linear inertial equations of a compound flood model, EGUsphere [preprint], https://doi.org/10.5194/egusphere-2024-1839

How to cite: Eilander, D., Fraehr, N., Leijnse, T., and de Goede, R.: Surrogate flood models for compound flood risk assessments and early warning, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-5209, https://doi.org/10.5194/egusphere-egu25-5209, 2025.

Cyclonic systems in the Eastern Mediterranean often produce compound extremes of heavy precipitation and strong winds, significantly impacting socio-economic systems. This study leverages traditional atmospheric analysis and dynamical systems theory to investigate these “wet” and “windy” extremes (Vakrat and Hochman, 2023). Using the co-recurrence ratio (α; De Luca et al., 2020) and persistence (1/θ; Faranda et al., 2017), we quantify atmospheric state dynamics and link them to extreme weather events. Results reveal that compound extremes exhibit higher co-recurrence and persistence than individual extremes, with anomalies in these metrics increasing the likelihood of extreme weather events by up to 18-fold. A case study of the mid-February 2012 Eastern Mediterranean compound event highlights the role of persistent upper-level dynamics in driving these extremes. Our findings emphasize the value of dynamical systems metrics in enhancing the predictability of compound extremes and their application to other regions and extreme weather events (Hochman et al., 2019). 

References

De Luca P, Messori G, Pons FME, Faranda D. Dynamical systems theory sheds new light on compound climate extremes in Europe and Eastern North  America. Quarterly Journal of the Royal Meteorological Society 146: 1636–1650. https://doi.org/10.1002/qj.3757

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

Hochman A, Alpert P, Harpaz T, Saaroni H, Messori G. 2019. A new dynamical systems perspective on atmospheric predictability: eastern Mediterranean weather regimes as a case study. Science Advances 5(6): eaau0936.  https://doi.org/10.1126/sciadv.aau0936 

Vakrat, E. Hochman, A. 2023.Dynamical systems insights on cyclonic compound “wet” and “windy” extremes in the Eastern Mediterranean. Quarterly  Journal of the Royal Meteorological Society 149(757): 3593–3606. https://doi.org/10.1002/qj.4575

How to cite: Hochman, A. and Vakrat, E.: Understanding Cyclonic Compound “Wet” and “Windy” Extremes in the Eastern Mediterranean through Dynamical Systems Theory, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-5532, https://doi.org/10.5194/egusphere-egu25-5532, 2025.

How to cite: Feldmann, M., Domeisen, D. I. V., and Martius, O.: Severe convective outbreaks and heatwaves – a continental-scale compound event, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-6437, https://doi.org/10.5194/egusphere-egu25-6437, 2025.

Near-freezing precipitation (NFP) events, a type of multivariate event compounding temperature and precipitation, are associated with widespread power outages. In a society undergoing an energy transition towards electrification, outages are associated with large impacts. Thus, understanding the impacts and the future evolution of NFP events in a changing climate is increasingly important. In this study, we first establish the relation between outages and NFP events, based on reanalysis data and on a Québec-based utility outage dataset. The highest density of outages in the region is found to occur in association with mixed precipitation near the freezing point. Next, daily NFP totals in various reanalyses are evaluated against Environment Canada weather stations in power-line-dense regions of Québec, to select a gridded reference. The Canadian Surface Reanalysis (CaSR) performs best and is selected. Then, a 28 member CMIP6 ensemble, bias-adjusted using CaSR, is used to evaluate future regional changes in the frequency of near-freezing precipitation events, as a function of time and as a function of local warming. In general, it is found that warmer areas south of Québec see a decline in the frequency of events, while colder northern areas see an increase. The number of days with near-freezing precipitations over 5 mm liquid equivalent varies non monotonically with annual temperature. This number will decrease by up to 40% south of Québec in the future and increase in the north. At the latitude of Montréal, the number of days may first increase and peak before decreasing again at the end of the century, as more wet snow turns to rain. However, rare events show a more uniform pattern of increasing intensity than NFP indicator, with slight decreases mostly near the coasts in the south. For Montréal, end of century NFP increases are more preeminent in ssp245, than in the warmer scenarios, which are likely further past the maximum risk. These findings on the impact of NFP events on the grid, as well as on the future evolution of these events, can directly inform the costly grid-hardening strategies considered for future adaptation.

How to cite: Rousseau-Rizzi, R. and Roy, P.: The impacts and future changes of near-freezing precipitation events in Québec, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-6508, https://doi.org/10.5194/egusphere-egu25-6508, 2025.

High storm surges, extreme sea waves, high river discharges and intense short-term rainfall are flood drivers that make densely populated coastal areas especially vulnerable to flooding. An additional increase of flood hazard and risk in coastal areas is expected due to changes in storminess, mean sea level rise, land subsidence and urbanisation. The simultaneous or consecutive occurrence of two or more flood drivers can lead to an event known as compound flooding.

In Croatia, compound flooding caused by the co-occurrence of high river discharges and storm surges is the only combination of compound flood drivers investigated to date. Consequently, other combinations of compound flood drivers remain unexplored. This study aims to address this gap by conducting further research on compound flooding in Croatia, specifically investigating the co-occurrence of extreme storm surges and rainfall along the Croatian coast. This study will provide insights into the compound flood potential due to extreme storm surges and rainfall at 42 locations along the Croatian coast. The rainfall data was obtained from rain gauge stations and the sea level data was obtained from Coastal Extremes in the Mediterranean Sea reanalysis and has been corrected by tide gauge data using machine learning.

High storm surges and heavy rainfall are flood drivers that often originate from the same weather system. Neglecting their seasonality can lead to a significant underestimation of the dependency and consequently the underestimation of the compound flood potential. With its pronounced seasonality, the Croatian coast is a great example for investigating seasonal correlation and co-occurrence of storm surges and rainfall. By disaggregating the time series into the individual seasons and analysing them separately, we gained a more detailed insight into the co-occurrence patterns of these flood drivers through maps with assigned correlation coefficients and a number of co-occurrences for each location.

As a result of this analysis, we will be able to identify the vulnerable areas with the highest probabilities of co-occurrence of high storm surges and intense rainfall more precisely. The selected locations will be eligible for a more detailed analysis of compound flood risk at the local level in future studies.

How to cite: Gržić, M. M., Petković, I., Ožanić, N., and Krvavica, N.: Joint Occurrence of Extreme Rainfall and Storm Surge along the Croatian Coast: Exploring Seasonal Variations, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-9949, https://doi.org/10.5194/egusphere-egu25-9949, 2025.

Heat stress is projected to intensify with global warming, causing significant socioeconomic impacts and threatening human health. Wet-bulb temperature (WBT), which combines temperature and humidity effects, is a useful indicator for assessing regional and global heat stress variability and trends. However, the variations of European WBT and their underlying mechanisms remain unclear. Using observations and reanalysis datasets, we demonstrate a remarkable warming of summer WBT during the period 1958-2021 over Europe. We find that the increase in European summer WBT is driven by both near-surface warming temperatures and increasing atmospheric moisture content. We identify dominant modes of European summer WBT variability and investigate their linkage with the large-scale atmospheric circulation and sea surface temperature anomalies. The first two leading modes of the European WBT variability exhibit prominent interdecadal to long-term variations, mainly driven by a circumglobal wave train and concurrent sea surface temperature variations. The last two leading modes of European WBT variability mainly show interannual variations, indicating a direct and rapid response to large-scale atmospheric dynamics and nearby sea surface temperature variations. We also present the role of global warming and changes in mid-latitude circulations in the variations of European summer WBT. Our findings can enhance the understanding of plausible drivers of heat stress in Europe and provide valuable insights for future climate adaptation planning.

How to cite: Ma, Q., Chen, Y., and Ionita, M.: Spatiotemporal Variations and Potential Drivers of European Summer Heat Stress, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-9994, https://doi.org/10.5194/egusphere-egu25-9994, 2025.

Denmark has experienced several significant compound flood events in recent years, and in parallel the Danish Meteorological Institute (DMI) has been developing a new flood warning system. This has led to a reassessment of how floods are conceptualized, predicted, and communicated. To support this, we here propose a framework for a flood typology tailored to both single-source and compound flood phenomena, with practical applications for public warnings and communication.

Our methodology builds upon the internationally acknowledged UNDRR/ISC hazard classes, which we filter for flood-related hazards relevant to Northern European coastal, lowland conditions. We then consider the organization of Denmark’s national agencies, local emergency response, and insurance structures. From this, we develop a flood typology for communication. As a part of the study, historical occurrences of compound flood phenomena are meticulously assessed by reviewing textual descriptions from a historical flood register (1990–2020) and conducting detailed case studies of recent events. Additionally, we examine the spatial and temporal overlap of flood-generating processes through a quantitative analysis of historical severe weather warning occurrences (2014–2024) addressing events of rainfall, storms, and sea levels.

Our findings reveal overlapping definitions within the UNDRR/ISC hazard classes, particularly regarding flood-generating processes and their geographic context. While DMI oversees severe weather warnings, observation networks are divided among four national agencies within the fields of: meteorology, oceanography, inland surface water, and groundwater. The emergency response in Denmark, as managed by 98 municipalities, is generally infrastructure-focused rather than flood-type-specific. For instance, urban water utilities often manage flood operations in cities. The insurance sector distinguishes between pluvial floods (private market) and fluvial or storm surge floods (covered by a national public disaster fund). We propose five general flood types for communication: (1) pluvial, (2) fluvial, (3) coastal, (4) groundwater, and (5) technological hazards (infrastructure failures of pumps, sluice gates, etc.). The historical flood register indicates two predominant compound flood types in Denmark: "coastal + fluvial" (driven by extratropical cyclones in winter) and "pluvial + fluvial" (caused by convective rainfall extremes in summer). It also shows that the key preconditioning variables include soil moisture, snow depth, and Baltic Sea water levels. The analysis of historical weather warnings reveals distinct regional patterns in compound flood risks. The detailed case studies provide storylines of how spatial compounding of flood types can overwhelm both national and local emergency responses. 

By integrating these insights, our study establishes a typology that is locally relevant for the Danish context, enhances the understanding of compound floods, and informs strategies for improved compound forecasting and communication.

How to cite: Pedersen, J. W., Su, J., Ringgaard, I. M., and Larsen, M. A. D.: Developing a flood typology for Denmark with practical applications for public warnings and communication, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-10126, https://doi.org/10.5194/egusphere-egu25-10126, 2025.

Compound events leading to significant coastal flooding have become a major concern in recent years. Storm surges caused by extra-tropical cyclones and coastal precipitation are key drivers of such events. While previous research has developed various methods for storm tracking, these have not sufficiently focused on impact-relevant storm tracking that directly addresses coastal flood risks. Motivated by this research gap, we initiated our analysis by identifying storm surges from 1991 to 2021 and tracked associated low-pressure systems and the associated impact-related compound dynamics. This approach not only enables the understanding of storm impacts but also offers potential for application to downscaled regional climate-scale products.

As a case study, we use Denmark, known for its complex ocean-circulation patterns due the North-Sea/Baltic Sea interface, narrow straits and fjords, and diverse coastline orientation. We clustered 32 sea level stations in Denmark by analyzing the co-occurrence of extreme storm surge events in the period 1990-2023. We then tracked extra-tropical cyclones over Northern Europe using the CERRA mean sea level pressure dataset, by identifying minimas at each time interval and reconstructing tracks by minimizing the distance between candidate points, through the use of Mixed Integer Programming. Finally, we investigate coastal precipitation in different regions of Denmark, as defined by the clustering of sea water level stations, with precipitation estimates from the CERRA-Land dataset.

Our analysis successfully identified storm tracks associated with extreme storm surge events, which were categorized into four distinct clusters. Similarly, Danish water level stations were grouped into three clusters based on the co-occurrence of extreme surge events: (1) the West coast of Jutland, (2) Kattegat and Inner Danish water, and (3) Baltic sea coastlines. By examining the dominant storm track types and station clusters, we revealed significant differences in impacted regions associated with different storm tracks.

We conclude that storm tracks have markedly different impacts on the occurrence of storm surge events across the Danish sub-regions. Precipitation levels associated with these storm surge events, and type of storm track, can uncover the need to consider both storm track characteristics and regional vulnerabilities when assessing compound and multi-variate coastal flood risks as opposed to storm surges in isolation.

How to cite: Agertoft, N., Pedersen, J. W., Su, J., Ringgaard, I. M., and Dahl Larsen, M. A.: Understanding coastal-pluvial compound floods associated with extra-tropical cyclones in Denmark , EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-10158, https://doi.org/10.5194/egusphere-egu25-10158, 2025.

Compound Inland Flooding (CIF) arises from the interactions between multiple hydrometeorological drivers, often magnified by landfalling Atmospheric Rivers (ARs) along the Pacific Northwest coast and interior basins of North America. This study investigates the mechanisms behind two primary CIF types, Rain-on-Snow (ROS) and Saturation Excess Flooding (SEF), using the CanRCM4 large ensemble under global warming levels of +1.5°C, +2°C, and +4°C. By examining the joint occurrence of ARs with ROS and SEF across key sub-regions, including the Cascade Range, Sierra Nevada, and the Great Lakes Basin, we assess the probabilities, seasonal shifts, and hydrological impacts of CIFs in the 21st century. Results show distinct regional patterns, with ROS events projected to decrease in frequency across the Pacific Northwest and Great Lakes Basin but remain significant in high-elevation regions prone to seasonal snowmelt, such as the Canadian Rockies. Conversely, SEF events are projected to increase substantially, particularly in the eastern U.S. and southern Great Lakes, driven by intensified precipitation and persistently saturated soils. The findings indicate that under higher warming levels, the contribution of ROS to extreme runoff can decrease, while SEF-driven flood events become dominant. Signal-to-noise ratio analysis shows that internal climate variability contributes considerable uncertainty to CIF projections in transitional climate zones but is overshadowed by external climate forcing at higher warming levels, particularly in coastal regions. By capturing the compounded effects of precipitation extremes, snowmelt dynamics, and soil moisture conditions, this study underscores the necessity of integrating AR-driven compound events into regional flood risk management strategies. 

How to cite: Najafi, M. R., Fereshtehpour, M., Grgas-Svirac, A., and Cannon, A.: Atmospheric Rivers and Compound Inland Flooding under Climate Change, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-11037, https://doi.org/10.5194/egusphere-egu25-11037, 2025.

This study investigates the impacts of climate change on marine heatwaves and extreme precipitation events associated with atmospheric rivers. First, our findings demonstrate that high-resolution models are more adept at simulating mesoscale eddies in the ocean, thereby facilitating more accurate predictions of future changes in marine heatwaves. Under climate warming, the intensity and annual days of marine heatwaves are projected to increase significantly. Even if organisms within large coastal marine ecosystems fully adapt to long-term mean warming, the escalating intensity of marine heatwaves would nonetheless pose substantial threats to these ecosystems. Furthermore, with global warming, the intensity and annual days of subsurface marine heatwaves are also expected to rise markedly on a global scale. This increase is primarily driven by the long-term rise in subsurface temperatures and changes in their variability. After accounting for the effects of long-term warming, the magnitude of increases in the intensity and annual days of subsurface marine heatwaves is notably greater than those at the surface, further exacerbating the risks posed by global warming to marine ecosystems.

Additionally, the study explores the influence of global warming on atmospheric river events in the Northern Hemisphere. High-resolution Earth system model simulations indicate that, under approximately 4°C of global warming, elevated sea surface temperatures enhance ocean-to-atmosphere moisture flux, thereby intensifying atmospheric river events. This intensification is projected to result in a doubling of the area affected by extreme precipitation events along the western coasts of Europe and North America. By disentangling the thermodynamic and dynamic contributions to intense precipitation associated with atmospheric rivers, the study identifies differences in the direction of vertical wind velocity changes as the primary source of regional disparities in dynamic contributions.

How to cite: Guo, X., Gao, Y., Zhang, S., Cai, W., Leung, R., Zscheischler, J., Thompson, L., Chen, D., Guo, C., Gao, H., and Wu, L.: Impacts of Climate Change on Precipitation and Marine Heatwaves: Insights from High-Resolution Earth System Models, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-11126, https://doi.org/10.5194/egusphere-egu25-11126, 2025.

Extreme weather and climate events are increasingly linked to severe socio-economic impacts, and their combination in space and/or time can further amplify these effects. This has heightened attention on compound events, which are combinations of multiple, potentially non-extreme climate events that collectively result in significant socio-economic consequences. A prominent example of a compound event in agriculture are false spring event. These occur when anomalous warm and wet conditions prevail in late winter, triggering early crop growth, followed by spring frost or severe drought. Such conditions can lead to substantial agricultural losses. To enable early warnings for such events, seasonal predictability is essential, as these phenomena typically unfold over a period of a couple of months. Seasonal predictability typically stems from slowly varying factors, such as sea surface temperatures and teleconnections, which influence the likelihood and timing of such events.

 One of the most globally influential teleconnections is the El Niño–Southern Oscillation (ENSO), with well-documented influence on climate systems worldwide. ENSO's impact on European climate, particularly during late winter, has been extensively studied, raising the question whether ENSO could play a role in triggering false spring events. Investigating these mechanisms offers valuable insights into ENSO's influence on European climate and enhances the potential for improved seasonal predictions of such events. To identify these large-scale patterns and non-linear relationships with other teleconnection patterns and modes of variability, like the North Atlantic Oscillation (NAO), we employ advanced statistical techniques, such as Kernel Regularized Generalized Canonical Correlation Analysis and Bayesian neural networks. By leveraging preimages and Accumulated Local Effect (ALE) plots, we uncover large-scale mechanisms relevant to European climate that exhibit strong interactions with the Niño3.4 region. Finally, we perform a causal analysis to trace the chain of interactions and pathways through which ENSO modulates European false spring events. 

 Our preliminary analysis focused on the first phase of the compound events, late winter. Results revealed significant interactions between the Niño3.4 region and atmospheric circulation patterns in the Euro-Atlantic region. These interactions involve a combination of well-known patterns such as the NAO, the East Atlantic/West Russia pattern, and the Scandinavian pattern. Second-order ALE plots obtained from a Bayesian neural network highlight that the interplay of these components can drive increasingly warm and wet conditions during late winter. These conditions create a favorable environment for the onset of false spring events, advancing our understanding of the mechanisms behind these impactful phenomena.

How to cite: Luther, N., Zorita, E., Luterbacher, J., Vlachopoulos, O., and Xoplaki, E.: Causal Links Between El Niño–Southern Oscillation and European Compound Events: A Focus on False Spring Events, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-13057, https://doi.org/10.5194/egusphere-egu25-13057, 2025.

Global warming may trigger more frequent snow droughts (SD). SD can result from low total precipitation (dry-SD), from high temperature leading to less solid precipitation (warm-SD) or from the combination of both (dry-warm compound SD). Each of those SD type pose different ecological threats. Nevertheless, the regions dominated by SD types, transition patterns, and the future risks under climate change remain unclear. Here, we investigated the dominant SD types and clarify the transition patterns among the three SD types during the historical and the future period. The results suggest a global increase in SD frequency by about 1.5-fold and 2-fold under SSP2-4.5 and SSP5-8.5 respectively. Moreover, the shares of warm SD is increasing and may become dominant by 2050 and probability of dry-warm compound SD may reach 4–10 times that of the historical period. The global transition from dry to warm dominated SD is attributed to greenhouse gases. Those findings provide a scientific reference for addressing climate change risks on SD.

How to cite: Wang, C., Li, Z., Guyennon, N., Chen, Y., and Li, Y.: Patterns of Snow Drought Under Climate Change: From Dry to Warm Dominance, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-13893, https://doi.org/10.5194/egusphere-egu25-13893, 2025.

Compound events (CE), characterized by the combination of climate phenomena that are not necessarily extreme individually, can result in severe impacts when they occur concurrently or sequentially. Understanding past and potential future changes in their occurrence is thus crucial. The present study investigates historical changes in the probability of hot and dry compound events over Europe and North Africa, using ERA5 reanalyses spanning the 1950-2023 period. Two key questions are addressed: (1) Where and when did the probability of these events emerge from natural variability, and what is the spatial extent of this emergence? This is explored through the analysis of “time” and “periods” of emergence, noted ToE and PoE, defined as the year from which and the moments during which changes in compound event probabilities exceed natural variability. The new concept of PoE allows for more in-depth signal analysis. (2) What drives the emergence? More specifically, what are the relative contributions of changes in marginal distributions versus in the dependence structure to the change of compound events probability? The signal is modelled with bivariate copula, allowing for the decomposition of these contributions. A focus on the dependence component is explored to quantify its effect on the signal’s emergence. 

The results reveal clear spatial patterns in terms of emergence and contributions. Five areas are studied in greater depth, selected for their similar signal behaviors. For example, the frequency of hot and dry events sharply increased in Maghreb and in the Iberian peninsula (ToE around 1980) and this rise is mainly due to a change in the marginals. Conversely, in eastern Europe the signal experienced a long PoE lower the natural variability, and this decline of CE probability is mainly driven by a change in the drought index. Although the dependence component is rarely the main contributor to PoE, it remains necessary to detect signal’s emergence. The date of ToE and the duration of PoE can be overestimated as well as underestimated (even more than 20 years) without considering this component. These findings provide new insights into the drivers of CE probability changes and open avenues for advancing attribution studies, ultimately improving assessments of risks associated with past and future climate change. 

How to cite: Schmutz, J., Vrac, M., and François, B.: Spatial structures of emerging hot and dry compound events over Europe from 1950 to 2023, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-13915, https://doi.org/10.5194/egusphere-egu25-13915, 2025.

Drought-flood abrupt alternation (DFAA), characterized by a period of persistent drought followed by sudden heavy precipitation at a certain level, has significant impacts on ecosystems and socioeconomic environment. As previous studies mainly focused on the monthly scale and regional scale, our study proposed a multi-indicator daily-scale method for identifying the DFAA occurrence. Then we applied the method on exploring the DFAA events over China from 1961 to 2018. The results show that: i) The DFAA events mainly occurred in the center and southeast of China. ii) The spatial coverage has a statistically significant (p < 0.05) increasing trend over China, of 0.355 %/decade. iii) The occurrence and spatial coverage of DFAA events increased by decades, and were mainly concentrated in summer (around 85%). Meanwhile, we conducted field experiments in typical area. The measurements revealed that DFAA events increased the soil nitrogen and phosphorus pollution in surface water.

How to cite: Bi, W., Weng, B., Zhang, D., Wang, F., Wang, W., Lin, W., Qi, X., and Lu, M.: Drought-flood abrupt alternation events and their impacts in China, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-14385, https://doi.org/10.5194/egusphere-egu25-14385, 2025.

In recent years, compound events, i.e., events resulting from the interaction between multiple physical drivers, have gained great attention in the scientific community especially as they can lead to greater impacts than events controlled by a main single physical driver. The majority of the studies relied on the use of statistical measures of dependence and multivariate analyses to show potential for compound events across diverse climatic and geographical regions. However, these approaches provide limited insights into the system being investigated and its behavior.

Here, we propose an approach to characterize the 'compoundness' of a system in terms of two components: structural compoundness, which refers to the overall tendency of physical drivers to jointly occur and interact, and transient compoundness, which refers to the specific occurrence or manifestation of interacting physical drivers. We provide example applications of the proposed characterization and discuss their implications for developing climate-resilient adaptation strategies.

How to cite: Ragno, E. and De Michele, C.: Structural and Transient Compoundness in Natural Systems, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-14749, https://doi.org/10.5194/egusphere-egu25-14749, 2025.

Drought episodes combined with hot events usually trigger dramatic impacts on ecosystems and agricultural production. However, most existing studies on climate stress focus primarily on individual events, leading to a neglect of compound information. Based on various combinations of climate conditions, we investigate the impact of 6 modes of events, namely, compound dry and cold events, compound wet and hot events, compound dry and hot events (CDHEs), compound wet and cold events, droughts, and hot events, on maize yield in China. Evidence from both country–level and province–level yield data indicates that CDHEs have emerged as a major threat to maize yield, with higher yield reduction than the other 5 modes of climate events. Negative maize yield anomalies caused by CDHEs have increased over the past decades, partly due to the rising frequency, spatial extent, and severity of compound events. Moreover, the El Niño–Southern Oscillation (ENSO) has recently intensified yield losses associated with CDHEs. Findings from this investigation underscore the urgent need for adaptation strategies to prevent the occurrence of CDHEs, and to mitigate their impacts.

How to cite: Wu, X.: Increasing impact of compound dry and hot events on maize yield in China, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-15475, https://doi.org/10.5194/egusphere-egu25-15475, 2025.

With the expected rapid growth of renewables in the French power system, periods of prolonged low renewable energy generation are expected to have a greater impact on the power system, especially if compounded with high electricity demand. In particular, compound winter low wind and cold events are identified by the French electricity transmission system operator as events that can drive major risks to the future French power system. Using CMIP6 climate simulations, the scope of this study is to characterize the future changes of these climate compound events in the mid- and long-term and assess the associated uncertainties.

To identify compound low wind and cold events, a wind power capacity factor and an electricity demand indices are derived using near-surface wind speed and temperature data from CMIP6 models, including several Single Model Initial-condition Large Ensemble (SMILE), for the 1950-2099 period. Due to large differences between observed and modeled indices, bias adjustment is first applied to raw temperature and near-surface wind speed data. The benefit of multivariate bias adjustment over univariate methods is assessed.

First, we characterize the future changes of compound low wind and cold events frequency in the mid- and the long-term, and which of the marginal characteristics (i.e., cold or low wind events) primarily drive these changes. Then, we assess the associated uncertainties, including uncertainties from internal variability, climate models, emission scenarios, and bias correction methods. Finally, we identify the role of climate drivers, including the global warming level, and exposure drivers, including the installed wind power capacity and the electricity demand parameters. This work demonstrates the relevance of CMIP6 large ensemble of simulations and methodologies currently used in the compound weather and climate events community to assess future risks for the power system.

How to cite: Collet, F., Boé, J., Bador, M., Dubus, L., and Joudier, B.: Future evolution of compound low wind and cold events in winter impacting the French electricity system in CMIP6 climate models, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-15803, https://doi.org/10.5194/egusphere-egu25-15803, 2025.

Compound extreme events are characterized by the combination of two or more events (not necessarily extreme) that can increase their respective individual impact. These phenomena can be of temporal nature (events that occur at the same time or in close succession), spatial nature (what happens in a place affects another) and/or multivariable nature (combination of several variables). This work focuses on the analysis of hot-dry compound extreme events —characterized by the simultaneous occurrence of high daily maximum temperature and low precipitation— and assesses their frequency, duration and severity.

For this purpose, both observational data (for a recent historical period) and climate model simulations provided by the CORDEX initiative, which gathers international efforts devoted to regional climate modeling, are considered. In particular, we use the CORDEX-CORE (CORDEX Coordinated Output for Regional Evaluations) ensemble, which comprises two Regional Climate Models (RCMs) driven by three Global Climate Models (GCMs) under two distinct emission scenarios, covering most continental CORDEX domains at 0.22º spatial resolution (approx. 25km). Systematic biases, typically present in these simulations, have been alleviated with the application of bias adjustment, using a semi-parametric, trend-preserving, quantile mapping method (ISIMIP). 

Our overall results show that hot-dry compound extreme events are enhanced over the next decades, with a general but region-dependent increase in frequency, duration and severity for different levels of global warming (+1.5, +2, +3 and +4 ºC, with respect to pre-industrial conditions), which can have important  impacts across various sectors such as health, economy, tourism and agriculture, among others. 

This work is part of Project COMPOUND (TED2021-131334A-I00) funded by MCIU/AEI/10.13039/501100011033 and by the European Union NextGenerationEU/PRTR. A. C. and R. M. acknowledge support from PID2023-149997OA-I00 funded by MICIU/AEI/10.13039/501100011033 and by ERDF/EU.

Keywords: compound events, regional climate models, climate change, extreme climate.

How to cite: Fuente-Gonzalez, M., Manzanas, R., Diez-Sierra, J., Chantreux, A., and Casanueva, A.: Evaluation and projection of hot-dry compound extreme events in a warmer climate, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-16140, https://doi.org/10.5194/egusphere-egu25-16140, 2025.

Extreme value analysis (EVA) includes a range of methods used to study the frequency and magnitude of rare but catastrophic events, with applications in science and engineering. These methods rely on mathematical theories that assume stable input data over time. However, many long-term datasets, especially those related to natural hazards, show clear changes over time (non-stationarity). With the availability of long-term climate records, there is a need for a reliable approach to analyze non-stationary extreme events that occur together (compound events), which is crucial for hazard assessment. This study introduces a method to analyze non-stationary joint extremes by combining Transformed-Stationary Extreme Value Analysis (tsEVA) with copula theory. This approach accounts for changes in the relationship between variables over time. The method includes sampling strategies to select relevant events, applying tsEVA for non-stationary univariate distributions, and using time-varying copulas to model the evolving relationships between variables. It thus considers all possible sources of non-stationarity that may affect joint extremes. The framework also incorporates statistical tools like the Mann-Kendall test to assess the significance of trends and Monte Carlo resampling for model validation and uncertainty analysis. Using this approach, the joint distribution of extremes in various natural hazards, such as river discharge, wave height, temperature, and drought, was successfully analyzed. The results highlighted the method's effectiveness in addressing diverse sources of non-stationarity and revealed dynamic patterns in variable interrelationships. Furthermore, the methodology developed in this study offers a viable tool for future research focused on generating statistically consistent hazard scenarios to support comprehensive risk assessments.

How to cite: Bahmanpour, M. H., Mentaschi, L., Tilloy, A., Vousdoukas, M., Federico, I., Coppini, G., and Feyen, L.: A generalized method for the analysis of non-stationary joint extremes based on the transformed-stationary extreme value analysis, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-16288, https://doi.org/10.5194/egusphere-egu25-16288, 2025.

Abstract. Droughts significantly affect socioeconomic conditions globally. As a multifaceted phenomenon, droughts are assessed through various indices distinguished in meteorological, agricultural, and hydrological typologies, each designed to capture distinct aspects. So, there is a strong demand for comprehensive drought monitoring tools that integrate multiple aspects to offer a holistic view of drought conditions. Typically, when introducing a new composite drought index, it is evaluated in comparison to existing indices, however this approach cannot allow to evaluate its accuracy in actual conditions. Therefore, shifting the paradigm from model-by-model evaluations to impact-oriented analysis is crucial. This work introduces a drought index based on deep learning where economic losses induced by drought are used as a key metric in assessing the index performance. The introduced index is calculated using cutting-edge deep learning algorithms based on various drought-related variables. Different types of self-supervised learning models, including Convolutional Neural Networks (CNN), Artificial Neural Networks (ANN), and Variational Autoencoders, are employed to enhance the model's accuracy and robustness. We use reanalysis data (ERA5) spanning from 1980 to 2022 for Italy, coupled with the EM-DAT database, to conduct impact analysis. The performance of each model is outlined based on their accuracy in estimating economic losses induced by droughts. 

Keywords: Drought Index, Deep learning, Autoencoder, impact-oriented analysis.

How to cite: Khosh Chehreh, M. and De Michele, C.: Development of a Composite Drought Index using deep learning: A Unified Framework for Multi-Dimensional Drought Characterization, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-16322, https://doi.org/10.5194/egusphere-egu25-16322, 2025.

In early September 2023, Europe experienced a pronounced atmospheric omega-blocking event, which led to spatially compounding precipitation and heat extremes across Europe. Omega-blocking is characterized by a persistent anticyclone at its core, flanked by two low-pressure systems to the southwest and southeast. During the September 2023 event, the center of the omega block was positioned over Central Europe and Southern Scandinavia, which experienced a significant heatwave during the first week of September 2023. Conversely, regions on the southwestern flanks (Spain) and southeastern flanks (Greece, Bulgaria, and subsequently Libya) were affected by extreme precipitation events, leading to severe flooding.

We employ the method of ensemble boosting to explicitly simulate omega-blocking situations with spatially compounding extremes (heatwave and extreme precipitation) with the Community Earth System Model 2 (CESM2). We therefore select analogs to the September 2023 event in a 30-member initial condition large ensemble of the CESM2 and use the model re-initialization approach of ensemble boosting to introduce slight perturbations to initial conditions 10 to 25 days prior to the event. This enables the generation of hundreds of coherent physical event trajectories, supporting the investigation of two key research questions: the first focuses on assessing the capability of the climate model to reproduce the 2023 event in its severity, while the second focuses on identifying the key characteristics of the omega block and its emergence that contribute to the most severe impacts on the ground.

In our talk, we introduce the research concept and address the following research questions: Is the CESM2 model capable of reproducing an omega blocking event with spatially compounding heat and precipitation extremes in the magnitude of the September 2023 event? Could the September 2023 event have been even more devastating by chance? What characteristics of the omega block and its emergence precondition the occurrence of the most extreme spatially compounding impacts in terms of heatwaves and extreme precipitation within the boosted ensemble?

How to cite: Mittermeier, M., Guo, Y., Suarez-Gutierrez, L., Bevacqua, E., and Fischer, E.: Spatially compounding heat and precipitation extremes under omega blocking in Europe, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-17021, https://doi.org/10.5194/egusphere-egu25-17021, 2025.

As global warming intensifies, high-latitude and mid-to-high-elevation watersheds are increasingly experiencing compound low-snow high-temperature events, posing serious challenges to water resources and ecosystem stability. However, the spatiotemporal characteristics of these events and their impacts on vegetation productivity and physiological processes remain insufficiently understood. In this study, drawing on multiple reanalysis datasets and hydrological models, we systematically evaluated the historical and future trajectories of low-snow high-temperature events across the Northern Hemisphere, including their potential lagged effects on ecosystems. By integrating diverse Gross Primary Productivity (GPP) datasets derived from observations, satellite products, and models—and employing an explainable causal machine learning framework—we identified key climatic and plant physiological drivers influencing GPP under these compound conditions. The findings highlight an increasingly frequent and persistent occurrence of low-snow high-temperature events, along with significant effects on vegetation functions, such as water-use efficiency, carbon uptake, and community structural adaptations. Overall, this research not only traces the upward trend of these compound events but also underscores their profound ecological implications, offering valuable insights for advancing global carbon cycle assessments and informing future climate adaptation strategies.

How to cite: Yang, Y.: Rising Compound Low-Snow High-Temperature Events: Drivers, Ecosystem Responses, and Future Outlook, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-17066, https://doi.org/10.5194/egusphere-egu25-17066, 2025.

The complexity of climate risk can lead to cascading impacts across the coupled climate, ecological, agricultural, and socioeconomic systems, which may involve potentially unprecedented outcomes and feedbacks, nonlinear behaviors or tipping points. While advances have been made in understanding such interconnected risks, particularly within specific disciplines, significant gaps remain in our understanding and modelling of such risks, and especially of how they cascade across systems. 

Several of such examples of cascading impacts can be found across the world, just in the last few years. The Australian bushfires of 2019-2020, fueled by extreme heat and prolonged drought, caused massive biodiversity loss, widespread air pollution, and significant economic damages. The 2021 Himalayan glacier collapse led to catastrophic flooding, infrastructure damage, and disruptions to local livelihoods, highlighting the fragility of mountain ecosystems in a warming climate. The global food and energy crisis of 2022, driven by geopolitical conflict and the disruption of supply chains compounded by low crop yields revealed the vulnerability of interconnected supply chains, with far-reaching implications for global stability. The 2024 DANA flooding in Spain, caused by a record-breaking atmospheric instability event and delayed emergency response, resulted in devastating loss of human lives and damage to infrastructure, agriculture, and urban areas, which eventually led to civil unrest in the region. All these examples underscore the need for comprehensive risk assessment, modelling and projection that better captures how shocks may compound and cascade across systems leading to high-impact outcomes larger than the sum of their parts. 

Existing frameworks and methodologies frequently fail to account for nonlinearities and worst-case outcomes or compartmentalize risks, in part to make an extremely complex problem simpler. This limits our ability to capture effects and impacts cascading to and from other sectors and systems, resulting in an incomplete understanding of the systemic nature of risk. Here, we assess to which extent cascading impacts have been included in impact assessments across sectors given our current methodologies and frameworks, to which extent our current methodologies and frameworks are insufficient for the task, and the cases where, even though current technology may allow it, cascading risks may have been overlooked. We reflect on recent examples of cascading impacts and their drivers, and outline critical directions for improving their integration into future risk assessments.

How to cite: Suarez-Gutierrez, L., Bastos, A., and Hegerl, G. C.: Cascading impacts across the coupled climate, ecological, agricultural and socioeconomic systems, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-17367, https://doi.org/10.5194/egusphere-egu25-17367, 2025.

Electricity networks are an important component of critical national infrastructure. Their failure, leading to power outages, can cascade through other infrastructure networks and compromise the function of other critical services. Electricity networks are facing a massive transformation to handle the increased demands placed on them by net zero commitments. Alongside this, future increases in the frequency and intensity of extreme weather will test electricity infrastructure that is already perceived to have insufficient resilience. Transforming networks as part of the net zero transition presents an opportunity to increase their resilience: this requires an understanding of the causes of network failures, the challenges that utility operators face in managing infrastructure risks, and the quantification of weather driven risks.

In this presentation, we present results from two projects. The first project used interviews and round-table discussions with energy industry experts in the UK to understand their needs from climate science as well as to uncover what they consider to be the largest weather and climate risks, the operational difficulties these present to electricity networks and what they believe to be low-regret options for enhancing climate resilience. The second project used statistical analysis to predict damage to electricity infrastructure from key weather hazards (windstorms, heat waves) and assessed how electricity infrastructure risks may change in the future using high-resolution 2.2 km climate simulations.

We discuss the main outcomes, strengths and limitations of both approaches and conclude that 1) expert elicitation provides a detailed and nuanced understanding of the range and severity of societal consequences produced by extreme weather and various compounding factors, and 2) probabilistic impact models that do not include multi-hazard and compounding effects underestimate the potential damages of extreme weather to electricity infrastructure – specifically the effects of wind direction, soil moisture and leaf cover during windstorms.

How to cite: Manning, C., Wilkinson, S., and Fowler, H.: Understanding compound weather and climate risks facing electricity networks, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-17762, https://doi.org/10.5194/egusphere-egu25-17762, 2025.

The Amazon rainforest, a critical global ecosystem, is increasingly threatened by climate change and extreme weather events. Over recent decades, the region has experienced record-high temperatures and unprecedented droughts. Compound drought and heatwave events (CDHWs), characterized by simultaneous dry and hot conditions, along with soil moisture (SM) deficits and high vapor pressure deficits (VPD), exacerbate ecosystem stress and intensify drought severity. This study investigates the climatology of CDHWs and compound low-SM/high-VPD events in the Amazon from 1981 to 2024 using the ERA5 dataset. Most compound events occurred during well-known drought years, including 1983, 1997/1998, 2010, 2015/2016, and 2023/2024. While compound events rarely impacted more than 20% of the region before 2010, subsequent years saw widespread effects, with the 2023/2024 drought ranking as the most extreme on record. During the austral summer of 2023/2024, CDHWs affected 70% of the Amazon's area, compared to 40% in 2015/2016. Similarly, low-SM/high-VPD conditions impacted 30% of the region in 2015/2016 and an unprecedented 60% in 2023/2024. Our results suggest an increase in the frequency, extent, and duration of compound extremes in the Amazon region, particularly over the last two decades, which could have critical implications for ecosystem resilience and climate adaptation strategies. The previous record compound event of 2015/2016 was particularly significant due to its ecological impacts, including tree mortality, biomass growth decline, and reductions in net primary productivity (NPP), gross primary productivity (GPP), and carbon uptake. Therefore, the ongoing record-breaking CDHW and low-SM/VPD conditions in 2023/2024 are expected to have even more severe impacts on the Amazon rainforest.

How to cite: Ferreira, V., Buras, A., Zscheischler, J., Machecha, M., and Ramming, A.: Increasing frequency and intensity of compound droughts in the Amazon region, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-18021, https://doi.org/10.5194/egusphere-egu25-18021, 2025.

How to cite: Xing, T. and De Michele, C.: Spatially compound and local extreme precipitation events: behaviors and trends , EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-18415, https://doi.org/10.5194/egusphere-egu25-18415, 2025.

The increasing frequency and intensity of extreme environmental and climatic stressors, such as heatwaves, droughts, wildfires, and air pollution episodes, highlight the urgency of understanding their interconnected nature. Traditionally studied in isolation, these stressors often interact in complex ways, amplifying their individual and cumulative impacts on ecosystems, economies, and public health. This study explores the global occurrence of compound events involving heatwaves, droughts, wildfires, and poor air quality, identifying their key drivers, spatial distribution, and associated consequences.

ERA5 reanalysis were used to identify drought periods using the Standardized Precipitation-Evapotranspiration Index (SPEI) and detected heatwaves based on temperature anomalies. Fire activity was assessed using Fire Radiative Power (FRP) data from the Moderate Resolution Imaging Spectroradiometer (MODIS) aboard the Terra and Aqua satellites, while air pollution levels, specifically particulate matter (PM2.5), were derived from the Copernicus Atmosphere Monitoring Service (CAMS) global reanalysis (EAC4). The co-occurrence of these phenomena was analyzed to pinpoint regions experiencing compound hot, dry, fire, and pollution events.

Our findings reveal distinct global hotspots where multiple stressors interact. Heatwaves and air pollution events were predominantly observed in regions such as India, the Arabian Peninsula, and eastern China. Meanwhile, the Brazilian Cerrado, northern Australia, and South African savannas frequently experienced simultaneous heatwave and wildfire occurrences. The Mediterranean region, particularly Greece, Portugal, and Italy, exhibited a high prevalence of concurrent heat, drought, wildfire, and air pollution episodes. Notably, in North America and Asia, PM2.5 concentrations reached significantly higher levels during simultaneous extreme events compared to isolated pollution occurrences.

The interplay of compound hot and dry conditions with wildfires, and ultimately with pollution events, presents critical challenges for public health and environmental management. The cascading effects of these interactions underscore the need for integrated approaches that encompass climate adaptation strategies, wildfire risk mitigation, and stringent air quality regulations. Understanding these linkages is essential for formulating policies that enhance climate resilience and safeguard communities against the escalating threats posed by climate-driven extreme events.

This research was funded by the Portuguese Fundação para a Ciência e a Tecnologia, FCT, I.P./MCTES through national funds (PIDDAC): UID/50019/2025 and LA/P/0068/2020 https://doi.org/10.54499/LA/P/0068/2020. This study was conducted within the scope of project https://doi.org/10.54499/2022.09185.PTDC (DHEFEUS) and supported by national funds through FCT. DL and AR acknowledge FCT I.P./MCTES for grants https://doi.org/10.54499/2022.03183.CEECIND/CP1715/CT0004 and https://doi.org/10.54499/2022.01167.CEECIND/CP1722/CT0006, respectively.

How to cite: Russo, A., Bento, V., Lima, D., and Careto, J.: Global Compound Climate Events: Intensified Air Pollution During Simultaneous Extreme Events, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-18746, https://doi.org/10.5194/egusphere-egu25-18746, 2025.

Renewable energy production plays a major role in Norway’s energy sector accounting for approximately 98% of the national electricity production. Unusually little precipitation in southern Norway in year 2021 resulted in reduced filling of the hydropower reservoirs. Accompanied with calm wind conditions over major wind-energy producing areas of Europe this led to exceptionally high electricity prices in Norway during winter 2021/2022.

As most renewable energy sources depend inherently on weather and climate condition, they are sensitive to large-scale weather regimes, natural climate variability, climate change and extreme weather events that can threaten the renewable energy system’s stability and reliability. By transitioning to more renewable sources such as hydro, wind and solar power, societies may expose themselves to an increased risk of potential instabilities and unreliability in the power supply. Within the SusRenew project we expand on the concept of compound events by looking at climate hazards that are specifically relevant for the energy system in Norway, and by looking at the possible joint occurrence of hazard pairs in different regions that are linked in the northern European energy system.

How to cite: Mayer, S., Ridjan Skov, I., Mati, A., Botnen Holm, T., Aall, C., and Deciron, C.: Identifying climate related hazards relevant for the Norwegian Energy System , EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-18782, https://doi.org/10.5194/egusphere-egu25-18782, 2025.

Extreme compound events, defined as the “combination of multiple drivers and/or hazards that contributes to societal or environmental risk”, present a
growing concern for the scientists and the civil society (Zscheischler et al. 2020). Climate models provide physical simulations of the climate until 2100, which permits to better understand the evolution of the extreme events under climate change. This study proposes a novel modeling of bivariate extreme events using bivariate Generalized Pareto Distributions (biGPD), with Extended GPD for the univariate part (EGPD). This novel semi-parametric modeling is applied to an extreme event: the flooding of the Seine and the Loire watersheds in June 2016. This event is a spatially compound event between the accumulated precipitations over the two watersheds. The accumulation of rain over several days is approximated by the Antecedent Precipitation Index (API), and high values of API are considered to lead to flooding. This approach is compared to a more classic copula modeling over simulations in Jacquemin et al. (2025, submitted). As climate simulations often have statistical biases, they must be corrected using bias correction algorithms. This is also the case for their simulations of compound events. This study compares several multivariate bias correction algorithms (CDF-t, dOTC and R2D2) on this event. dOTC seems to perform better than R2D2 for extreme values. The proposed methodology illustrates how compound events can be analyzed, and their evolution in frequency projected. As a perspective, this method can be applied to more diverse compound events, and it could be generalized to events in higher dimensions.

Bibliography:
Zscheischler, J., Martius, O., Westra, S., Bevacqua, E., Raymond, C., Horton, R. M., van den Hurk, B., AghaKouchak, A., Jézéquel, A., Mahecha, M.
D., et al.: A typology of compound weather and climate events, Nature reviews earth & environment, 1, 333–347, 2020.
Jacquemin, G., Vrac, M., Allard, D., and Freulon, X.: Estimating the return period of climate compound events using a non parametric bivariate Generalized Pareto representation. Submitted

How to cite: Jacquemin, G., Vrac, M., Allard, D., and Freulon, X.: Projecting frequencies of extreme rainfall compound events under climate change using bivariate extreme value modeling and multivariate bias corrections, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-19231, https://doi.org/10.5194/egusphere-egu25-19231, 2025.

The urgency for deeper understanding compound hydro-meteorological extreme events is growing, as these events are increasingly recognized for their potential to exacerbate impacts compared to single hazards. Characterized by the concomitance of multiple natural hazards or drivers, compound events are further intensified by climate change, which influences their severity and increases their frequency of occurrence. These hydro-meteorological extremes pose a significant risk to terrestrial ecosystems and have devastating consequences for socio-territorial systems. In Italy, recent extreme events have highlighted this threat, as demonstrated by the unprecedented sequence of heavy precipitation events in 2023-2024 that led to widespread flooding in the region of Emilia Romagna in central-northern Italy. These low-probability events, which resulted in several fatalities and damages amounting to tens of billions of euros, were amplified by antecedent precipitation that saturated soils, significantly enhancing the runoff response and, consequently, flood severity and extension. Indeed, the pre-condition given by soil imbibition preceding heavy rainfall occurrences is crucial in determining the potential severity of the event, but its comprehensive understanding is still limited.

This study investigates the relationship between precipitation and soil moisture conditions in Italy, with the goal to quantitatively characterize their role in the occurrence of historical and plausible compound hydro-meteorological extremes. We employ state-of-the-art reanalysis datasets (ERA5 and ERA5-Land) to analyze a series of representative extreme precipitation events, focusing on their antecedent soil moisture conditions and estimating the typical temporal scales of the associated co-variation. The link between these hydro-meteorological quantities and riverine flood occurrences is assessed considering streamflow discharge data from EFAS hydrological reanalysis dataset. The prevailing large-scale conditions driving these events, in terms of the 500-hPa geopotential height and the integrated water vapor transport column, are explored to identify the key dynamical features responsible for the occurrence of compound flooding. Additionally, a large ensemble of seasonal numerical weather forecasts is employed to sample the phase space of precipitation-soil moisture conditions applying the so-called UNSEEN (Unprecedented Simulated Extremes using ENsembles) approach. Within this framework, the probability of compound precipitation-soil moisture extremes is assessed through a statistical event coincidence analysis to understand the dominant spatio-temporal patterns of their interaction; moreover, physical storylines of rare, yet plausible, extreme flood events will be built through ensemble pooling.

How to cite: Giordani, A., Butera, C., Ruggieri, P., and Di Sabatino, S.: Towards a storyline of compound flood events over Italy: the role of precipitation-soil moisture pre-conditioning, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-19481, https://doi.org/10.5194/egusphere-egu25-19481, 2025.

Compound events occur when multiple drivers or hazards combine, creating societal or environmental risks, with high-impact occurrences such as simultaneous heatwaves and droughts often leading to more severe consequences than isolated incidents. A systematic review of 366 peer-reviewed papers published between 2012 and 2022 reveals an annual average increase of 60% in research focused on compound events, particularly for multivariate (co-occurring) events. Studies primarily focus on Europe, Asia, and North America, while significant gaps remain in Africa, South America, and Oceania. Key modulators, such as the El Niño Southern Oscillation, along with event types like compound floods and high-temperature, low-precipitation events, are highlighted as the most studied within the literature. Recommendations from the review include expanding research in underrepresented regions and studying a broader range of typologies, events and modulators. Furthermore, it also calls for enhanced cross-disciplinary and sectoral collaboration to better understand and manage the growing risks posed by compound events in a changing climate.

How to cite: Brett, L., White, C., Domeisen, D., Ward, P., Zscheischler, J., and van den Hurk, B.: Review article: The growth in compound weather events research inthe decade since SREX (2012-2022), EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-19486, https://doi.org/10.5194/egusphere-egu25-19486, 2025.

As the climate warms, interacting weather extremes such as sequential heat events pose complex risks to societies. Regarding the global food system, laboratory experiments suggest that crop exposure to spring heat may either confer tolerance or enhance vulnerability to subsequent summer heat events. We show, under historic conditions that hot springs benefit crop yield but amplify the impacts of summer heat by 3% to 36% across crops and regions compared to average spring conditions. This increasing sensitivity results in impacts outweighing hot spring benefits when summer temperature anomalies exceed 2-4°C. Analyzing projected temperature increases, we find an eight-fold rise in the frequency of sequential heat extremes under the Shared Socioeconomic Pathway 3-7.0. Accounting for the compounding effect of sequential heat on crop yields increases projected losses by 1 to 71% depending on crop and region. This underlines the emerging nonlinear risk of sequential heat extremes to food security, which can largely be avoided when limiting warming to 1.5°C globally.

How to cite: Hamed, R., Steinmann, C. B., Ma, Q., Balanzategui, D., Broadman, E., Lesk, C., and Kornhuber, K.: Amplified agricultural impacts from increasingly sequential heat extremes, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-19545, https://doi.org/10.5194/egusphere-egu25-19545, 2025.

Reliable climate projections are crucial for informed decision-making in water resource planning and management. However, selecting suitable Global Climate Models (GCMs) remains challenging due to inherent uncertainties and computational constraints. This study introduces a novel hybrid approach for GCM selection, focusing on models that exhibit consistency in projecting future climate changes and skill in representing current climate conditions, including average climate, seasonal patterns, and climatic variations. GCM performance in simulating these critical properties was evaluated for rainfall, maximum temperature, and minimum temperature using the Kling-Gupta Efficiency (KGE) metric, resulting in a structured 3×3 performance matrix for each GCM. The matrix distances, quantifying the disparities between each GCM's performance matrix and the ideal reference matrix, were used to represent overall model performance. GCMs were then ranked based on these differences using the Jenks natural breaks classification method to identify the top-performing models for ensemble construction. The proposed method was tested by selecting GCMs for Nigeria from 19 CMIP6 GCMs. Results indicate that 15 GCMs consistently projected future climate within a 95% confidence interval. Further evaluation reveals that ACCESS.ESM1.5, BCC.CSM2.MR, CMCC.ESM2, and MRI.ESM2.0 are the most suitable for simulating Nigeria's climate. The multi-model ensemble means of the selected GCMs projected a notable increase in rainfall by 10 to 40% over most of the country and maximum and minimum temperatures by 1.0 to 3.5°C and 0.5 to 4.0°C, respectively. The proposed approach offers an effective tool for GCM selection to enhance climate projection reliability.

How to cite: Kumar, S., Choudhary, M. K., and Thomas, T.: Comparative Analysis of GCM Selection Approaches for Climate Change Impact Assessment in India, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-1118, https://doi.org/10.5194/egusphere-egu25-1118, 2025.

Stratospheric aerosols, most of which originate from explosive volcanic sulfur emissions into the stratosphere, are a key natural driver of climate variability. They are thus one of the forcings provided by the Coupled Model Intercomparison Project (CMIP) Climate Forcings Task Team for the CMIP7 Fast Track, a set of climate model experiments designed to deliver the Intergovernmental Panel on Climate Change (IPCC) 7^th assessment cycle. In this work, we document the final version of the stratospheric aerosol forcing datasets delivered to modelling groups for CMIP7 Fast Track. Our datasets cover the 1750-2023 period to meet to the need of modelling groups who might run extended historical simulations starting in 1750 instead of 1850. We produced one volcanic stratospheric sulfur emission dataset catering for the needs of models which have a prognostic interactive stratospheric aerosol scheme, as well as a stratospheric sulfate aerosol optical property dataset required by models that cannot interactively simulate stratospheric sufate aerosols. For the satellite era (from 1979 onwards), sulfur emissions and sufate aerosol optical properties are based on the MSVOLSO2L4 and GloSSAC datasets, respectively. For the pre-satellite era (1750-1978), the emission dataset is based on ice-core datasets complemented by the geological record for small-moderate magnitude eruptions not captured in ice-core records. Although inferring emissions of these eruptions from the geological record is highly uncertain, our approach minimizes an important bias in the pre-satellite era forcing, both in terms of mean and variability. The pre-satellite aerosol optical property dataset is directly derived from emissions using an updated version of EVA_H, a reduced-complexity volcanic aerosol model. This ensures methodological consistency between our emission and optical property datasets, and maximizes consistency with methodologies used in the paleoclimate (PMIP) and volcanic forcing (VolMIP) model intercomparison projects in CMIP6. We will present extensive comparison between our CMIP7 Fast Track dataset and the CMIP6 dataset. Last, we will discuss the main challenges to improve stratospheric aerosol datasets in the future and to move to high frequency (yearly or less) extension and update instead of an ad-hoc production for each CMIP phase.

How to cite: Aubry, T., Toohey, M., Schmidt, A., Kovilakam, M., Sigl, M., Khanal, S., Chim, M. M., Johnson, B., Carn, S., Verkerk, M., Nicholls, Z., and Smith, I.: Historical stratospheric aerosol optical properties and volcanic sulfur emissions for CMIP7 Fast Track, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-3831, https://doi.org/10.5194/egusphere-egu25-3831, 2025.

3D model is a 3D digital representation of objective things, which has been widely applied in fields like urban construction, disaster prevention and mitigation, medical research, biological science, industrial manufacturing, agricultural production, etc. As a special 3D model, 3D geological model possesses the characteristics of 3D model and plays a fundamental role in geological survey, mineral exploitation, underground engineering and smart city construction.With the development of intelligent sensing technology and 3D geological modeling technology, the scale of 3D geological model data increases exponentially. Meanwhile, with the pace of large-scale underground engineering and smart city continuing to increase, 3D geological model with fine large scenes is being eagerly required. The rapid growth of data and the refinement of large application scenes bring new challenges to the real-time dynamic visualization of 3D geological models. These challenges are mainly reflected in the new technical problems related to 3D geological model rendering.This study focuses on 3D geological model rendering and puts forward the corresponding solutions. The validity of the technology has been proved by the simulation test of cluster cloud environment consisting of 5 computers. The technique has been applied in the construction of 3D geological information and visualization system in transparent Xiong’an.Firstly, the data organization mode of two common structures of 3D geological model (3D geological structure model and 3D geological high-precision grid model) is analyzed, and a distributed storage strategy of 3D geological model based on MongoDB is proposed. Aiming at the characteristics of multi-layer data in z-direction of 3D geological structure model, an octree index mechanism is proposed to improve the efficiency of data scheduling according to the z-direction spatial information and layer information. The rendering optimization of a single node 3D geological model is studied. The rendering in the cloud environment still needs the cooperation of each sub-node. Therefore, the overall rendering efficiency in the cloud environment can be improved by adopting efficient rendering optimization strategies for the 3D geological model of each node and selecting an effective node scheduling strategies. Single-node 3D geological model rendering is mainly performed by transferring data from memory to GPU. The communication between memory and GPU is a bottleneck, which will affect the overall rendering efficiency. Through the strategies of visibility elimination, LOD establishment, data merging and instance rendering optimization, this thesis effectively reduces the number of drawing calls and communication times. How to optimize and improve the overall performance of 3D geological model rendering in cloud environment from a global perspective is studied, and a multi-level distributed SCMP framework is proposed, which integrates the advantages of cluster, GPU, distributed storage, etc., to maximize the distributed computing ability of existing machines and improve the rendering efficiency in cloud environment. From the experimental data, the node invocation optimization strategy with “GPU+CPU” can ensure that the frame rate of the four rendering nodes and the end-user scene in the cloud environment is stable at about 35 frames per second, and can achieve satisfactory cluster load balancing effect.

How to cite: Song, Y., Gao, Z., Song, G., and Li, J.: Research on Key Technologies of 3D Geological Model Rendering in Cloud Environment, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-4053, https://doi.org/10.5194/egusphere-egu25-4053, 2025.

As a data-intensive science, earth system science has been focusing on the research of living environment and constituent element’s characteristics, including its forming time, location and evolution. With big data and AI boosting, there are more opportunities and challenges for geological research transformation. And it is more likely to improve geological survey and geological research by means of information method such as AI algorithm, methods, tools, software, and etc.. As for the storage, distribution and application of different format and discipline data, the key is to set up a series of rules and tools to realize the data services’ flexible using in security. It adopts hybrid data management framework to build up an unified index to support the spatial geological data finding. The matching platform is also developed to realize the geological achievements distribution. Moreover, it can significantly benefit faster and more efficient research. In all, the technique has been applied successful on Chinese geological survey information platform with great reuse adaptability on other platforms.

How to cite: cui, N. and gao, Z.: Research and application of key technologies of geological data platform, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-5383, https://doi.org/10.5194/egusphere-egu25-5383, 2025.

Climate forcings are the input drivers to coupled climate models (AGOCMs) and earth system models (ESMs). They are routinely used as part of the coupled model intercomparison project (CMIP). Here we present the historical greenhouse gas forcing used in the seventh phase of CMIP (CMIP7) and compare it to its predecessors from CMIP6 and CMIP5. We show that revised methods and input data have had little effect on historical estimates of greenhouse gas forcing, but that greenhouse gas forcing has continued to increase since 2015 (the end of the CMIP6 historical experiment), even if forcing from some specific gases has decreased. Beyond the greenhouse gas forcings, there are a number of other forcings involved in CMIP. Following on from our involvement in the CMIP Forcings Task Team, we present an outline for moving towards sustained, roughly annual, releases of these forcings and discuss the challenges for realising this possibility.

How to cite: Nicholls, Z., Pflüger, M., and Meinshausen, M.: CMIP7 historical greenhouse gas forcing and steps towards sustained releases, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-5765, https://doi.org/10.5194/egusphere-egu25-5765, 2025.

With the major challenges posed by climate change and significant shifts in Earth systems, the need for high-precision and diverse climate predictions has grown. These predictions aim to explore a variety of scenarios, such as the Shared Socioeconomic Pathways (SSPs). Advances in computational power have enabled the development of sophisticated coupled physical-biogeochemical-ecological models of marine systems. However, these models remain computationally intensive and energy-demanding, raising questions about the appropriate level of complexity relative to the availability of independent data for accurate calibration, and calls for simplification to reduce execution time. Here, we aim to simplify the Eco3M-MED model, which is a complex biogeochemical model representing the low trophic levels (up to mesozooplankton) in the ocean through 37 state variables, and which is intended to be run at the scale of the Mediterranean basin.

Common simplification methods include conservation analysis, lumping, time exploration, and sensitivity analysis. Since most of these simplification methods reduce or even penalize the ability to interpret model results, or require complex implementation, we have chosen a simple, classic method, based on the local sensitivity analysis (One-Factor-At-A-Time, OFAT) method that does not impair this ability. This work's originality lies in the approach adopted to obtain different declinations of the reduced model. This approach indeed benefits from an original strategy for parametrizing the Eco3M-MED model, initiated several years ago and recently implemented in practice. This strategy consists of the construction of a set of consistent parameters, resulting in the establishment of relations between the so-called core parameters and dependent parameters. Core parameters are perturbed based on the level of knowledge of each parameter. The main objective of this study is to apply this novel approach to identify the biogeochemical processes that can be removed with minimal impact on model performance, thereby enabling model simplification and reducing computational costs. We also apply the principle that a single simplified model is not necessarily the best solution, and aim instead to derive a family of simplified models associated with different usage objectives, ensuring that the simplified model reproduces certain quantities well in particular.  The criteria used to derive a simplified model from the sensitivity analysis are also subject to analysis to identify their influence on the degree of simplification. Finally, the computational efficiency and accuracy of simplified models were compared with the full model to determine optimal simplification for specific applications. Future research will focus on performing global sensitivity analysis on high-impact core parameters to assess uncertainties in both the full and simplified models.

How to cite: Zhang, Y., Baklouti, M., and Brasseur, P.: Sensitivity-driven simplification of complex ecosystem models: Integrating mechanistic insights for cost reduction and predictive accuracy, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-6300, https://doi.org/10.5194/egusphere-egu25-6300, 2025.

Earth system reconstructions are key to understanding past climate dynamics, environmental variability, and biogeochemical cycles, revealing how Earth responds to natural and human influences. These reconstructions require integrating geological, geophysical, and geochemical data with advanced computational models.

Recent advancements in Large Language Models (LLMs) enhance the processing of complex datasets, improving the accuracy and predictive power of Earth system reconstructions. To leverage this, we develope GeoGPT, a domain-specific LLM for geosciences, trained on open-source data. This open-source, non-profit project encourages broad collaboration among experts in broad branches of the geoscience and AI. GeoGPT helps advance Earth system reconstructions, offering deeper insights into Earth's past and future, and guiding responses to environmental challenges.

By harnessing the combined strengths of Geoscientists, AI experts, and the broader research community, GeoGPT aspires to unlock new avenues of exploration, accelerate breakthrough discoveries.

How to cite: Chen, H.: GeoGPT and Its Potential Applications for Earth system Reconstructions, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-7628, https://doi.org/10.5194/egusphere-egu25-7628, 2025.

Active margin topography has profoundly shaped Earth’s climate, biodiversity, and natural resource distribution over geological time. However, reconstructing paleotopography in these regions remains challenging due to the sparse and uneven distribution of proxies like stable isotope paleoaltimetry, palynology, paleobotany, and thermochronology. These traditional methods often leave large spatial and temporal gaps, with uncertainties in paleoelevation estimates reaching up to 2,000–3,000 m. To address these challenges, we introduce an innovative workflow utilizing artificial intelligence to reconstruct paleotopography at active margins since the Devonian. Using Explainable Boosting Machines (EBMs), we identify key factors such as plate kinematics, mantle dynamics, and climate that govern active margin topography. Insights from the EBM analysis guided the development of a Random Forest (RF)-based regressor which was then used to predict paleotopography through time. Our RF model achieved a mean error of 554 m when validated against present-day ETOPO elevation data. Our model highlights time-evolving subduction flux, trench migration rates, and upper mantle temperature as the primary controls on active margin topography. To validate our approach, we compare our reconstructions with existing paleotopographic models and geological proxies in two regions: the Cenozoic Andes and Mesozoic-Cenozoic Eastern China. For the Andes, our model closely matches the existing reconstructions, highlighting a ~4,000 km rapid rise of the Altiplano since the late Oligocene, driven by an increase in subduction flux (from 0.03 km³/yr to 0.10 km³/yr) and a transition in trench migration from retreating (2 cm/yr) to stationary, likely due to slab anchoring. In Eastern China, our model predicts sustained high topography (>2,500 m) during much of the Cretaceous, attributed to high subduction flux (>0.12 km³/yr) from the Pacific Plate and an advancing trench. A subsequent shift to trench retreat (-2 cm/yr) in the Late Cretaceous–Early Cenozoic led to back-arc extension and a decline in elevation to ~1000 m. Our study offers a transformative approach to bridging gaps in paleotopographic constraints, improving our understanding of the interplay between surface and interior processes. By providing a robust framework for reconstructing past landscapes, our model has significant implications for studying ecosystems, biodiversity evolution, and the metallogenesis of convergent margins.

How to cite: Singh, S. P., Seton, M., Zahirovic, S., and Wright, N. M.: Reconstructing the Topographic Evolution of Active Margins Since the Devonian Using Artificial Intelligence, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-9233, https://doi.org/10.5194/egusphere-egu25-9233, 2025.

Earth system modelling is an important instrument to investigate climate change in an integrated way, taking into account the interactions between the different compartments of the Earth system. It is also an important tool to perform climate projections for different climate scenarios in order to take appropriate mitigation and adaptation measures. Such climate simulations are coordinated internationally as part of the World Climate Research Programme’s (WCRP) Coupled Model Intercomparison Project Phase 7 (CMIP7).

Climate forcings are key for defining the main drivers of climate change in climate simulations. A very important aspect of the CMIP7 intercomparison is that all participating models were run under similar experimental conditions. In particular, in using the same climate forcings in the different models.

The Alfred Wegener Institute (AWI) will participate in the CMIP7 project with two state-of-the-art Earth system models AWI-CM3 and AWI-ESM3. This is being done as part of the German contribution to the Coupled Model Intercomparison Project (CAP7). The AWI contribution to CAP7 includes the adaptation of the AWI-CM3 model to be able to use different forcing data (such as greenhouse gases, solar forcing, O₃, and aerosol forcing) to fulfil the requirements of CMIP7. One task is therefore the implementation of the climate forcing dataset for anthropogenic aerosols MACv2-SP (currently available for CMIP6plus [Fiedler and Sudarchikova, 2024]) provided for CMIP7. For this purpose, an aerosol interface will be implemented in the AWI-CM3 climate model to read and process the aerosol forcing data provided by the MACv2-SP dataset.

In this presentation we will discuss the implementation of the MACv2-SP data into the AWI-CM3 climate model and present first results of the responses to these forcings.

How to cite: Wieters, N., Streffing, J., Goessling, H., and Jung, T.: Preparing AWI-CM3 for CMIP7: Implementing anthropogenic aerosol forcing (MACv2-SP), EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-10752, https://doi.org/10.5194/egusphere-egu25-10752, 2025.

General circulation models (GCM) can reasonably reproduce the global mean temperature trend during the historical period. In this study, we examine the performance of CMIP6 models in reproducing the changes in land-ocean temperature contrast between 1979-2014 during the comprehensive satellite observation era. The observed land-ocean warming contrast, defined as the land warming trend divided by the ocean warming trend, is completely outside the model spread of the historical scenario, indicating that the models severely underestimate land temperature increase relative to global mean temperature warming. Even when sea surface temperatures are prescribed to observations in AMIP experiments, the land warming trend remains outside the model spread, particularly between 15S and 15N. This was shown to be because GCM overestimates the increase in specific humidity on tropical land and underestimates the drying trend on tropical land. Since future projections over land have a significant impact on human activity, improving the representation of tropical land surface processes in GCMs is essential.

How to cite: Toda, M., Kang, S., and Shaw, T.: Climate Models Underestimate Satellite Era Land-ocean Warming Contrast in the Tropics, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-11291, https://doi.org/10.5194/egusphere-egu25-11291, 2025.

The CMIP ensembles represent a partial exploration of our uncertainty in the future physical climate under various scenarios for future greenhouse gas emissions. They are thus valuable tools for exploring the potential consequences of climate change for society. One aspect of this is the impact on global and national economies. In the economics literature this is often addressed through a “damage function” which relates economic damages to national, regional or global changes in temperature.

Here we will present an assessment of the economic damages implied by the CMIP6 ensemble for various nations/regions, different Shared Socio-Economic pathways, and, crucially, a variety of different damage functions. A number of important factors will be highlighted including:

• The uncertainty in economic damages which arises from the chaotic nature of the climate system, characterised by those CMIP6 models with relatively large initial condition ensembles.
• The relative consequences for economic assessments of uncertainty in the damage function, the choice of CMIP6 model (model uncertainty), chaotic uncertainty (initial condition uncertainty), and scenario uncertainty.
• How these factors vary by country and region.

CMIP6 only represents a limited exploration of uncertainty in the physical climate response and there is also considerable uncertainty in the damage functions beyond that currently explored in the literature. These represent deep uncertainty. We will present plans for future work to embed more thorough explorations of epistemic uncertainty into future analyses, including the consequences of crossing tipping points. These considerations are valuable when considering the design and implementation of the CMIP7 project. What would be the most useful design characteristics if the target were economic assessments? We will address this question in terms of both the size of initial condition sub-ensembles, the diversity of models included, and the value of a mixture of higher and lower resolution model implementations.

How to cite: Rosser, J. and Stainforth, D.:  Economic consequences of CMIP diversity and targets for CMIP7, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-12178, https://doi.org/10.5194/egusphere-egu25-12178, 2025.

A common set of simulation is important for intercomparison between different models. The ‘entry-card’ to participate in CMIP is to perform the baseline DECK simulations (1850-control, 1pctCO2, abrupt-4xCO2, historical-amip). However, performing the control and historical simulations from an 1850 baseline is prohibitively expensive for high-resolution, fully-coupled, general-circulation-models (GCMs). Therefore, HighResMIP chose to use a shorter experimental protocol based on 1950 conditions alongside a shorter spin-up length and simplified aerosols. Because of this difference in protocol, it is not clear exactly how the HighResMIP simulations relate to the other CMIP simulations. In this study we analyse the control and historical simulations of the HadGEM3-GC3.1 model, which performed control and historical simulations based on both 1950 and 1850 baselines. Our results show that the absolute temperature is sensitive to the different experimental protocol, but the anomalies are much more comparable. This opens an interesting discussion on whether climate change should be discussed in terms of absolute values or anomalies. The difference in the absolute value (and mean state) is largely due to the different aerosol scheme used in CMIP and HighResMIP for this particular model. The second phase of HighResMIP no longer require models to use Easy Aerosol, so modelling centres should use the same aerosol scheme if they would like their HighResMIP simulations to be comparable to CMIP simulations.

How to cite: Lai, M. and Roberts, M.: 1950-control vs 1850-control: How do HighResMIP simulations relate to CMIP simulations?, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-12926, https://doi.org/10.5194/egusphere-egu25-12926, 2025.

The projected changes in the hydrological cycle under global warming remain highly uncertain across current climate models. Here, we demonstrate that the observational past warming trend can be utilized to effectively constrain future projections in mean and extreme precipitation on both global and regional scales. The physical basis for such constraints relies on the relatively constant climate sensitivity in individual models and the reasonable consistency of regional hydrological sensitivity among the models, which is dominated and regulated by the increases in atmospheric moisture. For the high-emission scenario, on the global average, the projected changes in mean precipitation are lowered from 6.9% to 5.2% and those in extreme precipitation from 24.5% to 18.1%, with the inter-model variances reduced by 31.0% and 22.7%, respectively. Moreover, the constraint can be applied to regions in middle-to-high latitudes, particularly over land. These constraints result in spatially resolved corrections that deviate substantially and inhomogeneously from the global mean corrections. This study provides regionally constrained hydrological responses over the globe, with direct implications for climate adaptation in specific areas.

How to cite: Dai, P., Nie, J., Yu, Y., and Wu, R.: Constraints on regional projections of mean and extreme precipitation under warming, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-14119, https://doi.org/10.5194/egusphere-egu25-14119, 2025.

The global sea surface warming pattern emerged since the early 1980s is characterized by a boomerang-shape warming in the western Pacific, the basin-wide warming in the Indian Ocean north of 30°S, and a triple-stripe warming in the North Atlantic. This pattern can be obtained with or without El Niño/La Niña signals, indicating the independence of El Niño/La Niña, and is the leading EOF with the El Niño/La Niña signals removed. A negative phase of this pattern started emerging in the early 1980s, switched to positive phase in the 1990s, and has been becoming more prominent for the past few years.

CMIP models have been found to have difficulty simulating observed global sea surface temperature (SST) trend, especially the cooling trend in the tropical eastern Pacific. However, the cooling trend in the tropical eastern Pacific in the past four decades is statistically insignificant in our trend analysis adopting a more stringent signal detection method (namely, the False Discovery Rate, FDR). By applying the same trend detection and EOF approach to the simulated SST in the historical simulations of forty CMIP6 models by removing El Niño/La Niña signals, we detected in the ensemble mean SST a trend pattern closely resembling the observed, which also changes from negative to positive phases in the late 1990s and continues becoming more positive into 2014. Whereas each model has slightly different performance in simulating this trend pattern, the ensemble time series of corresponding trend pattern in each model correctly reflects the emergence and enhancement of the warming pattern in the past four decades. However, this model ability seems to be masked by the large fluctuations of El Niño/La Niña, an intrinsic climate mode contributing large internal variability to the global domain, and its temporal fluctuations cannot be synchronized in the coupled models in the historical experiments, which are strongly driven by continuously increasing radiative effect of greenhouse gases concentration. On the other hand, The models seem to be capable of simulating the emergence of the global ocean warming pattern in response to the prescribed increasing greenhouse gas concentration.

How to cite: Hsu, H.-H. and Chen, Y.-L.: CMIP6 Models Properly Simulate the Emergence of Global Ocean Warming Pattern, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-14140, https://doi.org/10.5194/egusphere-egu25-14140, 2025.

GeoGPT is a non-profit domain-specific Large Language Model for geosciences, trained based on open-source data. It provides an effective solution to the challenges of managing large data volumes, complex formats, and low efficiency in the utilization of books and papers in the field of paleontology. Its powerful data extraction capabilities will significantly enhance the efficiency of extracting, analyzing, and building databases for data of various formats, sizes, and origins. This enables scientists to construct online fossil datasets and empowers paleontologists to develop innovative tools such as paleontological classification assistants. Not only does this accelerate scientific research progress, but it also makes the acquisition and application of paleontological data, such as invertebrate fossils, more convenient, ultimately driving comprehensive progress in the field of paleontology.

How to cite: Xiang, Z.: GeoGPT: Transforming Paleontology with AI-Powered Data Extraction and Analysis, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-14239, https://doi.org/10.5194/egusphere-egu25-14239, 2025.

One of the major challenges faced by the geotectonic community is how to determine the paleolongitude of continents and tectonic plates as we try to reconstruct Earth’s tectonic history back in time, because classic paleomagnetic record is only sensitive to paleolatitude.  Torsvik et al. (2014) previously used mantle structure as a reference frame for palaeolongitude constraints back in Earth history, assuming that the two equatorial and antipodal large low shear velocity provinces (LLSVPs) observed in present-day Earth’s lower mantle are fixed and stable ancient structures unrelated to plate tectonic history and subduction geometry. However, such an assumption is inconsistent with true polar wander (TPW) record (Li et al., 2004, 2023), the cyclic occurrence of global mantle plume activity coupled with the supercontinent cycle (Li et al., 2008; and Zhong, 2009), and geodynamic modelling results (Zhong et al., 2007; Zhang et al., 2010; Flament et al., 2017).

In a recent paper of Li et al. (2023), we utilized palaeomagnetically interpreted TPW record, particularly inertia interchange true polar wander (IITPW) events, and global mantle plume record, to develop a dynamic global mantle reference frame that not only provides a first-order mantle dynamic evolution for the past 2 billion years, but also for the first time provides a way to trace the longitudinal change of continents and tectonic plates back in time. In particular, through the recognition of newly-defined type-1 and type-2 IITPW events coupled with plume record checking, we are now able to hypothesis that: (1) in periods with type-1 IITPW, the concerned supercontinent had developed its own degree-2 mantle structure (e.g., the antipodal LLSVPs divided by concurrent circum-supercontinent subduction girdle); (2) in periods with type-2 IITPW, a young supercontinent or multiple plates during the assembly of that supercontinent were moving over a legacy degree-2 mantle structure of the immediate ancestor supercontinent prior to the maturity of its own mantle structure. In our model, Nuna (lifespan 1600–1300 Ma) assembled at about the same longitude as the latest supercontinent Pangaea (lifespan 320–170 Ma), with an equatorial degree-2 mantle structure starting to exist as early as ca. 1700 Ma. Rodinia (lifespan 900–720 Ma) formed through introversion assembly over the legacy Nuna subduction girdle either ca. 90^◦ to the west or to the east before the subduction girdle surrounding it generated its own degree-2 mantle structure by ca. 780 Ma (but not before 800 Ma). Pangea assembled over the subduction girdle of legacy Rodinian degree-2 mantle structure, with its own degree-2 mantle structure (the one we still observe today) formed no much earlier than 270 Ma.

References

Flament, N., Williams, S., Müller, R.D. et al., 2017. Nat. Commun. 8, 14164.

Li, Z.-X., Liu, Y. and Ernst, R., 2023. Earth-Sci. Rev. 238, 104336.

Torsvik, T.H., van der Voo, R., Doubrovine, P.V. et al., 2014. Proc. Natl. Acad. Sci. 111 (24), 8735–8740.

Zhang, N., Zhong, S., Leng, W., Li, Z.-X., 2010. J. Geophys. Res. Solid Earth 115(B6), B06401.

How to cite: Li, Z.-X.: Absolute longitudinal constraints for palaeogeographic reconstruction based on a dynamic mantle reference frame, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-15897, https://doi.org/10.5194/egusphere-egu25-15897, 2025.

Back in 2017, solar forcing recommendations for the 6th round of the Coupled Model Intercomparison Project (CMIP) were provided which covered, for the first time, all relevant solar irradiance and energetic particle contributions. Since then, this dataset has been extensively used in climate model experiments and has been tested in various intercomparison studies. Further, new datasets have been come available. An International Space Sciene Institute (ISSI) Working Group has been established to review these recent achievements in order to define the strategy for building a revised solar forcing dataset for the 7th round of CMIP. After receiving community feedback on this strategy, a historical solar forcing dataset for CMIP7 has been recently constructed. Major changes with respect to CMIP6 include the adoption of the new Total and Spectral Solar Irradiance Sensor (TSIS-1) solar reference spectrum for solar spectral irradiance and an improved description of top-of-the-atmosphere energetic electron fluxes, as well as their reconstruction back to 1850 by means of geomagnetic proxy data. Solar irradiance varaibility in the reference forcing dataset is based on historical reconstructions generated with the new empirical NASA NOAA LASP (NNL) Solar Spectral Irradiance Version 1 model, NNLSSI1. In adition, an alternative solar irradiance dataset, based on SATIRE, is provided for sensitivity experiments. In this talk we will discuss the applied modifications with respect to CMIP6 and their implication for climate modeling. Ongoing activities on solar forcing uncertainty quantification and the construction of future solar forcing scenarios will also be summarized.

How to cite: Funke, B., Dudok de Wit, T., Haberreiter, M., Marsh, D., Ermolli, I., Kinnison, D., Nesse, H., Seppälä, A., Sinnhuber, M., Usoskin, I., Asikainen, T., Bender, S., Chatzistergos, T., Coddington, O., Koldoboskiy, S., Lean, J., van de Kamp, M., and Verronen, P.: Solar forcing for CMIP7, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-15908, https://doi.org/10.5194/egusphere-egu25-15908, 2025.

Paleomagnetism provides the main quantitative tool for reconstructing Earth’s paleogeography. Apparent polar wander paths (APWPs), derived from paleomagnetic data, trace the motion of tectonic plates relative to the Earth’s rotation axis through geological time, providing a paleogeographic framework for studying the evolution of Earth’s interior, surface, and atmosphere. Traditionally, APWPs are calculated from study-mean paleomagnetic poles that are assigned equal weight, regardless of the number of paleomagnetic sites used to compute it and the uncertainties in the position or age of the pole. Here, we introduce the next generation of APWPs that are calculated from site-level paleomagnetic data instead of from study-mean poles. This alternative approach assigns larger weight to larger data sets and allows the incorporation of spatial and temporal uncertainties. We demonstrate the advantages of this new method with recently published APWPs based on compiled (Gallo et al., 2023) and simulated site-level data (Vaes et al., 2023). We show how the latter, a global APWP for the last 320 Ma, provides more reliable estimates of the apparent polar wander rate of all major tectonic plates, and discuss its implications for the rate and magnitude of true polar wander since 320 Ma. In addition, we introduce APWP-online.org: an online, open-source environment that provides user-friendly tools to compute site-level APWPs and to use them to quantify relative paleomagnetic displacements. We showcase how these tools are currently used to compute site-level APWPs, e.g., for the North China block and Tibetan terranes. Finally, we provide future directions for the construction of APWPs and highlight opportunities for improving their quality and resolution.

How to cite: Vaes, B. and van Hinsbergen, D.: Reconstructing paleogeography using site-level apparent polar wander paths, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-16299, https://doi.org/10.5194/egusphere-egu25-16299, 2025.

The provision of solar forcing datasets for CMIP7 includes a dataset with scenarios from the present to 2300. This dataset contains daily values of the same variables as in the historical solar forcing for CMIP7, namely: solar spectral irradiance, medium energy electrons, solar energetic protons and galactic cosmic rays. In contrast to CMIP6, which had only two scenarios, for CMIP7 we will provide a large ensemble of scenarios to avoid selection bias.

Let us stress that we are providing scenarios, not forecasts: the reconstructions vary randomly in time, but their statistical and spectral properties are fully consistent with historical variations, providing realistic surrogates for solar forcing.

In this presentation we explain how historical observations are used to build these surrogate reconstructions. This process involves several steps, starting with the 14C reconstructions of past solar activity. These will be described in detail, together with the first version of the dataset. 

How to cite: Dudok de Wit, T., Funke, B., Haberreiter, M., Marsh, D., Ermolli, I., Kinnison, D., Nesse, H., Seppälä, A., Sinnhuber, M., Usoskin, I., Asikainen, T., Bender, S., Chatzistergos, T., Coddington, O., Koldoboskiy, S., Lean, J., van de Kamp, M., and Verronen, P.: Solar forcing for CMIP7: making of future scenarios, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-17597, https://doi.org/10.5194/egusphere-egu25-17597, 2025.

A Digital Twin (DT) is a data-driven model of a physical entity with two-information flows that enables the direct interaction between both. DTs of the natural environment are typically constructed by fusing multi-modal measurements of some physical phenomena using Artificial Intelligence (AI) methods. The physical entity interacts with the DT through natural changes whilst the DT interacts with the physical entity through automated changes in sensing systems and through decision-making processes. Large-scale DTs of the Earth system are currently in development through initiatives such as Destination Earth (DestinE) whilst small-scale DTs for local monitoring are in development for numerous applications such as hazard warning, agriculture and eco-hydrology. Currently these systems are being developed independently yet combining them offers opportunities for calibrating large-scale DTs and improving the resolution of large-scale DTs by replicating the dynamics of smaller systems using AI methods. In this contribution, we develop a new concept through which to link small- and large-scale DTs in order to automate an agile sensing system that can respond to natural environmental variability and directly measure changes of interest. Large-scale DTs are built primarily through Earth Observation (EO) data and describe regional to global scale changes in the Earth system whilst small-scale DTs simulate local variability using in situ sensors such as Terrestrial Laser Scanners (TLS). Linking the two means the large-scale DT can inform small-scale DTs by adapting their measurements (e.g. spatial and temporal resolution, focus area of interest, specific physical measurements) in response to regional changes in, for example, weather patterns. We focus on the following components: 1) using the small-scale DT to downscale the large-scale DT and ‘zoom’ into areas of interest; 2) using both the small- and large-scale DT to automatically detect changes in the environment and acquire new measurements without human intervention; and 3) using the small-scale DTs to calibrate large-scale DTs. With the increasing development of digital twin technology in the environmental sciences, our new concept will enable better integration of DTs and improve monitoring performance, which can improve decision-making. 

How to cite: Durrant, A., Harcourt, W. D., Höfle, B., Weiser, H., and Tabernig, R.: Linking small- and large-scale Digital Twins: A concept , EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-18187, https://doi.org/10.5194/egusphere-egu25-18187, 2025.

Paleoclimate information has played a key role in demonstrating how the Earth System responds to a variety of external forcings and how the earth’s climate is tightly related to atmospheric greenhouse gas concentrations. Although no strict analogue of possible future climate states exists, testing our understanding of the earth system, as embedded in earth system models, for conditions widely different from the historical period, is made possible by the existence of paleoclimate and paleoenvironmental reconstructions. Since its start in 1995, PMIP, the Paleoclimate Modelling Intercomparison Project (https://pmip.lsce.ipsl.fr/), has fostered and coordinated model-model and model-data comparisons for key periods: the mid-Holocene, ~6000 years ago, the Last Glacial Maximum (LGM), 21,000 years ago, the last two millennia, the last interglacial, the mid-Pliocene warm period (MPWP) were the key periods for PMIP4-CMIP6, with specific targets for each period. For instance, the enhanced monsoons and response of the northern high latitudes for the mid Holocene, the fate of Arctic sea ice and climate of the last interglacial, large spatial gradients and equilibrium climate sensitivity for the LGM and MPWP. In addition, each of these periods stood as reference for further PMIP experiments aimed to better understand the response of the climate system to external forcings.

For the next CMIP phase, PMIP continues to contribute studies on the responses to external forcings. This poster will present the targets for the FastTrack last interglacial experiment (abrupt-127k) as well as future opportunities related to other periods (e. g. Kageyama et al., 2024). We look forward to discuss with the CMIP and PMIP communities to plan further cross-cutting work and analyses.

Acknowledgements and cited reference.

We are acknowledging the help of the PMIP community in building PMIP over the years.

Kageyama M, et al., (2024) Lessons from paleoclimates for recent and future climate change: opportunities and insights. Front. Clim. 6:1511997. doi: 10.3389/fclim.2024.1511997

How to cite: Kageyama, M., Brierley, C., and Peterschmitt, J.-Y. and the the PMIP community: Earth system responses to external forcings : opportunities from paleoclimate studies and the Paleoclimate Modelling Intercomparison Project (PMIP) for CMIP, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-18925, https://doi.org/10.5194/egusphere-egu25-18925, 2025.

We present the CMIP7 solar forcing dataset and its validation both for the effects of solar irradiance and particle forcing. In particular we present the results from first simulation runs that use the new CMIP7 as well as the previous CMIP6 dataset for the period 2002-2012, covering two solar maxima and a deep solar minimum. Specifically, we present simulation runs carried out with the chemistry-climate models WACCM, SOCOL, EMAC, ICON and KASIMA to determine the response to the solar SSI and particle forcing. The performance of the CMIP7 recommendations with respect to atmospheric radiative heating and composition will be evaluated both compared to the CMIP6 recommendations, and to satellite observations of atmospheric trace gases. The different responses and their implications will be discussed.

How to cite: Haberreiter, M. and the CMIP7 Solar Forcing Validation Team: CMIP7 solar forcing validation, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-20295, https://doi.org/10.5194/egusphere-egu25-20295, 2025.

The past topographic evolution of the Tibetan-Himalayan orogen holds the key to understanding interactions between Earth, Climate and Life processes since deep times. This has been hindered so far by the lack of accurate paleogeographic reconstructions of the orogen through time based on well-dated reliable proxies of past elevations. In the sedimentary archives of the basins formed in the orogen during the collision, recovered fossil content including pollen, fish and mammals yielded first order estimates on elevations based on environmental conditions of nearest living relatives while leaf physiognomies provided more direct constraints. Stable isotope composition from ancient meteoric waters preserved in pedogenic carbonates and biomarkers have been recovered and interpreted in terms of paleoelevations assuming past meteoric lapse rates. Outside of the basins in the high massifs, synkinematic hydrous silicates preserving ancient rainfalls have been used for paleoaltimetry purpose, notably in the Himalayas. Despite these significant efforts, the new paleoelevation datasets have led more to controversy than consensus. Fierce debates currently involve several international groups. Widely different topographic growth scenarios have been proposed with end-members ranging from a high Plateau prior to the onset of the India-Asia collision (“Proto-Tibetan Plateau”), to a much more recent - mostly Miocene - uplift and the preservation of broad low elevation valleys late until the Neogene.

As part of the starting TIBETOP project (funded by the french ANR) we propose here a state-of-the-art review of paleoelevation proxies across the Tibetan-Himalayan orogen, ranging from surface records in sedimentary basins to deeper crustal rocks now exhumed in the relief and mountain belts bordering these basins. We present a compilation and reappraisal of the existing regional paleoelevation data including revised provenance, stratigraphy dating, and stable isotope data in basin records as well as structural context, exhumation and fluid-rock deformation interactions at different interfaces of the continental crust. The TIBETOP project thus aims to produce a set of interactive paleogeographic reconstructions through time with associated datasets constraining the Himalayan-Tibetan orogen since the India-Asia collision. These will be improved through the project to include new data and updated paleogeographic reconstructions made available to modelers of climatic, biotic and surface processes to enable testing the above cited fundamental hypotheses on the role of mountain and plateau building on Earth System processes over geologic time.

How to cite: Dupont-Nivet, G., Fidalgo, J.-C., Moreau, K., Rivera, L., Feng, Z., Fang, X., Lavé, J., and Licht, A.: Tibetan Plateau paleogeographic reconstructions during the India-Asia collision: from paleoelevation proxies to geodynamic models, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-21621, https://doi.org/10.5194/egusphere-egu25-21621, 2025.

The integration of big data, cloud models, and extensive knowledge to drive new knowledge discovery through data is a new paradigm for research in the field of Earth sciences. Although the advancement of big data technologies and infrastructures has simplified data acquisition, deep-time geoscience still faces challenges such as fragmented data, difficulties in visualization, and insufficient computing power. To assist the broad community of geoscientists, we propose the "Deep Platform," a one-stop online research platform that utilizes cloud computing and advanced technologies. The platform provides open access to deep-time geoscientific data, knowledge, models, and computing power. It is designed to promote collaborative innovation and discovery among global geoscientists. The "Deep Platform" represents a significant advancement in geoscientific exploration, fostering global collaboration and advancing a data-driven research paradigm within the framework of open science.

How to cite: Hu, L.: DEEP Platform: Empowering Global Geoscientists  in Data-Driven Research Era, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-21952, https://doi.org/10.5194/egusphere-egu25-21952, 2025.

Climate action (SDG 13) and reducing inequalities (SDG 10) are central goals of sustainable development. However, the distribution of climate risks and carbon emissions across regions is uneven, and this disparity poses significant challenges in global climate change governance. To address this issue, this study defines the concept of climate risk and introduces the "Mismatch Responsibility Index" to quantify the imbalance between carbon emissions (carbon footprint) and climate risk burdens. The study further examines the socio-economic and technological factors that drive this imbalance. The key findings include: (1) Climate risks and carbon footprints exhibit significant spatial and temporal variability, with the gap between cities expanding over time; (2) In China, more than half of the prefecture-level cities experience a significant mismatch in climate responsibility, with underdeveloped regions facing disproportionately high climate risks; (3) The main factors contributing to this mismatch are energy consumption patterns, population size, and the level of technological innovation. Further policy analysis indicates that local government policies, the promotion of regional green energy transitions, and technological innovation are essential to narrowing the gap in responsibility distribution. (4) Using simulations of different policy scenarios, the study proposes several recommendations, including strengthening local government climate policies, supporting green energy transitions, promoting technological innovation, and reallocating international climate finance. These measures are expected to reduce regional disparities in climate responsibility and contribute to more equitable climate governance.

How to cite: Li, Y., Liu, X., Hasi, E., Ji, R., Zhang, S., and Hao, Y.: The Climate Risk and Regional Carbon Emission Responsibility in China from the Perspective of "Mismatch Responsibility": Temporal-Spatial Variability and Driving Factors Analysis, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-383, https://doi.org/10.5194/egusphere-egu25-383, 2025.

Floods, as one of the most devastating natural disasters, have far-reaching impacts on property, safety, and mental health. This study employs Structural Equation Modeling (SEM) to explore the pathways linking flood experiences to psychological distress, focusing on mediating factors such as property loss, recovery efforts, and socioeconomic conditions. Drawing on data from the 2021 floods in Germany, the analysis provides insights into how direct and indirect factors interact to shape mental health outcomes.

Flood experience is conceptualized as direct exposure to flood hazards, including water depth, flow velocity, and contamination. These factors collectively capture the intensity and severity of the flood event. Key findings reveal that flood experience significantly predicts property loss (Estimate = 0.254, p < 0.001) and direct impacts such as self-injury, family injury, and uncertainty about the safety or whereabouts of family members or close friends during flood, which, in turn, exacerbate psychological effects. These direct impacts, alongside property loss, drive psychological impacts, measured through post-traumatic stress disorder (PTSD) screening and ongoing mental health effects, including persistent thoughts about the event and whether it continues to affect individuals' daily lives (Estimate = 2.227, p = 0.006). Socioeconomic factors, such as income and property ownership, influence recovery efforts, which mitigate psychological distress (Estimate = 0.294, p < 0.001). While recovery efforts mitigate distress (Estimate = 0.294, p < 0.001), property loss remains a substantial stressor. The total indirect effect of flood experience on psychological burden (Estimate = 0.304, p = 0.002) underscores the cumulative impact of material loss, immediate threats, and recovery challenges.

The model achieves strong fit indices (χ²/df = 2.15, RMSEA = 0.048, CFI = 0.925), validating its conceptual framework. These findings emphasize the critical role of flood experience in shaping mental health outcomes and the need for holistic disaster response strategies that address immediate impacts and foster long-term psychological recovery. By emphasizing both direct and cascading effects, this study informs policies aimed at enhancing resilience and mental health support in flood-prone areas.

How to cite: Pham, T. T. T. and Sairam, N.: Understanding the Psychological Impacts of Flooding: A Structural Equation Modeling Approach from the 2021 German Floods, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-830, https://doi.org/10.5194/egusphere-egu25-830, 2025.

Evidence shows an ongoing increase in the frequency and severity of global heatwaves, raising concerns about the future impacts of climate change and the associated socio-economic costs. Here, we develop a disaster footprint analytical framework by integrating climate models, epidemiological and hybrid input-output, and computable general equilibrium global trade models to estimate the mid-century socioeconomic impacts of heat stress. We consider health costs related to heat exposure, the value of heat-induced labor productivity loss, and indirect losses due to economic disruptions cascading through supply chains. We find that the global heatwave days would increase by 104% in 2060 compared to 2022 under SSP585, and the global average annual number of heat-induced deaths would increase to around 1.12 million (0.85 ~ 1.39 million). For economic impacts, we show that the global annual incremental loss increases exponentially from 0.03±0.01 (SSP245) ~ 0.05±0.03 (SSP585) percentage points during 2030 – 2040 to 0.05±0.01 ~ 0.15±0.04 percentage points during 2050 – 2060. By 2060, the expected global economic losses reach a total of 0.6% ~ 4.6% with losses attributed to health loss (37%~45%), labor productivity loss (18%~37%), and indirect loss (12%~43%) under different SSPs. Small and medium-sized developing countries in Southeast Asia and Africa suffer the most from heat risks as well as regional supply chain disruptions.

How to cite: Zhu, S., Sun, Y., Wang, D., and Guan, D.: Health and Economic Costs of Future Extreme Heat Risk, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-1449, https://doi.org/10.5194/egusphere-egu25-1449, 2025.

Projections of macroeconomic damage from future climate change tend to suggest mild to moderate impacts. This leads to welfare-optimal climate policies in Integrated Assessment Models (IAMs) that recommend very slow emissions reductions over the coming decades, in sharp contrast with the ambitions of the Paris Agreement. These econometric models assume that weather impacting a single country is all that affects the economy of that country. We examine whether the addition of global weather conditions in the empirical modelling of economic growth affects the projections of the impact of climate change on global GDP. In effect, we explore whether the interconnectedness of the global economy makes individual countries vulnerable to weather changes that impact other countries. Using three influential econometric models we add global weather to the regressions. We find that this leads to significant worsening of the projections of macroeconomic damage for given future emissions scenarios. Damage to world GDP in 2100 under SSP5-8.5, averaged across both econometric models and climate models increases from ~11%  under models without global weather to ~40% if global weather is included. Further, we demonstrate that when the damage function used in IAMs is estimated from empirical models augmented with global weather conditions, they reduce the welfare-optimal amount of climate change from ~2.7C to ~1.7C which is consistent with the Paris Agreement targets. Our results highlight the need for econometric modelling and climate science’s understanding of extreme events to be integrated much more consistently to ensure the costs of climate change are not underestimated. 

How to cite: Neal, T., Newell, B., and Pitman, A.: Reconsidering the macroeconomic damages of severe warming, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-3401, https://doi.org/10.5194/egusphere-egu25-3401, 2025.

Hail is a main contributor to weather-related damages to buildings, cars, and agriculture in Switzerland, demanding actionable information on hail risks and forecasts across sectors. The research project scClim addresses this demand by establishing a seamless model chain from observing, modelling and forecasting hail events to the quantification of hail impacts, including simulations to compare hail occurrence in current and future climate.

Within the project, we study several types of impact-based hail forecasts and warnings for Switzerland, addressing the interests of different stakeholder groups. We employ ensemble weather forecasts by the Swiss Meteorological Office combined with (a) impact-informed vulnerability thresholds to produce local hail warnings, and (b) information about exposed assets and their calibrated vulnerability to produce aggregated hail impact forecasts. While the impact-based forecasts would have to be thoroughly validated before operational use, the forecast products highlight how varying demands of different stakeholder groups shape the forecast product and the provided information.

How to cite: Gebhart, V., Schmid, T., and Bresch, D. N.: Impact-Based Hail Forecasts for Switzerland in the scClim Project, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-4445, https://doi.org/10.5194/egusphere-egu25-4445, 2025.

The aging agricultural labor force presents significant challenges to farm productivity and sustainability, particularly when compounded by climate change. This issue is especially critical in countries like Thailand, where the agricultural workforce is aging rapidly. Notably, 2.84 million older adults are engaged in the agricultural sector, representing 59.2% of all older adults participating in the economy. A Swiss Re Institute (2021) study ranked Thailand as the fifth most vulnerable country to GDP impacts from climate change. Furthermore, projections indicate that climate change could inflict cumulative damages on Thailand's agricultural sector, totaling 0.61–2.85 trillion baht from 2011 to 2045, averaging 17.9–83.8 billion baht annually. Without adequate preparation, Thailand risks significant economic downturns driven by agricultural productivity and production losses.

Effective local governance is one of the most crucial determinants of coping with the crisis above. This study used a mixed-methods approach, combining quantitative and qualitative research. The quantitative research included a survey of 2,500 older farmers in 2024 from three provinces in Thailand: Chiang Rai in the northern region, Buriram in the Northeastern region, and Uthai Thani in the central part, where extreme drought exists. The qualitative approach involved focus groups with older farmers and in-depth interviews with policymakers and older farmers.

The findings revealed that flexible and adaptable local governance is among the most critical factors contributing to the resilience of older farmers. Drought management for older farmers in Thailand requires coordination among multiple agencies with distinct roles, emphasizing the need for integration to ensure effective communication. Agencies must collaborate to share information and coordinate efforts to disseminate accurate, comprehensive, and timely information. Examples include broadcasting weather forecasts from the Meteorological Department, coordinating cloud-seeding operations, and providing water and resource management guidance through the Irrigation Department. More importantly, its communication strategies must specifically target older farmers. Various communication channels should be utilized, particularly platforms that older farmers can readily access, such as community radio, village loudspeakers, and local media. Additionally, digital platforms and social media can be leveraged to inform younger family members, who can relay the information to older adults. To ensure accessibility, communication materials should be simple, straightforward, and audience-specific. This includes tailored communication, such as using local dialects or translating complex information into user-friendly formats, as older farmers may have difficulty understanding formal language or technical terms. This effectively supports their resilience in drought.  After implementing drought management measures, the government must assess the effectiveness of its communication strategies. This evaluation should determine whether the information was delivered efficiently to older farmers and identify gaps or barriers in the communication process. Based on these insights, necessary adjustments should be made to enhance future communication efforts, ensuring they are tailored to this demographic's needs, preferences, and limitations.

Keywords: Adaptive Governance, Older Farmers, Risk Reduction, Resiliency, Drought

How to cite: Swangsilp, S.: Local Governance for Enhancing Resilience to Climatic Challenges Among Older Farmers: A Case Study from an Extreme Drought-Prone Area in Thailand, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-5150, https://doi.org/10.5194/egusphere-egu25-5150, 2025.

Climate change is increasing the frequency and intensity of extreme events and hazards, posing serious risks to societies and ecosystems worldwide. These phenomena do not only threaten economic systems but also broader dimensions of human well-being, including inequality, health, and education. Despite a growing recognition of these risks, the global mechanisms linking climate extremes to human development remain poorly understood. Furthermore, besides GDP, explicit estimation of future climate change damages and extremes on socio-economic projections remain limited. 

In this study, we focus on the impacts of climate extremes on human development, analyzing their effects on three main components of the Human Development Index (HDI): life expectancy, expected years of schooling, and gross national income per capita. Using a dataset covering 1,773 sub-national regions over three decades from 1990 to 2020, we employ high-resolution climate data to examine immediate and lagged socio-economic responses to extreme events and hazards, particularly rainfall extremes, heatwaves, and droughts. By exploiting fixed effects panel modeling, our approach accounts for the simultaneous inclusion of multiple extremes in damage functions and evaluates the integration of an adaptation proxy to capture regional differences in vulnerability.

Finally, we apply the derived impact functions to Shared Socioeconomic Pathways (SSP) scenarios, providing projections of climate-driven damages on HDI across different development and climatic narratives, capturing the key climatic and social uncertainties. 

How to cite: Mastropietro, M., Spinoni, J., and Tavoni, M.: Past and Projected Climate Extremes Impacts on Human Development , EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-5881, https://doi.org/10.5194/egusphere-egu25-5881, 2025.

The "Early Warning for All" initiative, launched at COP27, aims to ensure global coverage by early warning systems (EWS) for hazardous weather, water, or climate events by 2027. This study evaluates the effectiveness of EWS in Northeast China focusing on six meteorological disasters: heavy rain, cold waves, high winds, high temperatures, hail, and frost over the past five years.

We analyze the evolution in the timeliness and content of early warnings, correlating these with the integration of new technologies. Our findings reveal significant variations in EWS performance across different disaster types and geographical areas. For instance, while some systems provide warnings with substantial lead times for events like heavy rain and cold waves, others, particularly for hail and frost, show less temporal advance or accuracy.

This research highlights the disparity between scientific advancements in EWS and their practical application, underscoring the need for improved communication and decision-making processes within the warning system framework. We discuss the reasons for these imbalances, such as technological adoption rates, regional infrastructure variances, and policy implementation challenges.

Our study suggests that while technological capabilities have advanced, the translation into operational EWS effectiveness remains uneven. We propose several strategies to bridge the gap between scientific potential and operational reality, aiming to enhance the third pillar of EWS—effective communication and decision-making. These insights are crucial for refining EWS to better protect communities from natural hazards, contributing to the global aim of universal early warning coverage by 2027.

Keywords: Early Warning Systems, Meteorological Disasters, Northeast China, Timeliness, Accuracy, Communication Strategies, Decision-Making, Climate Resilience

How to cite: li, C.: Enhancing Timeliness of EWS with new technology: A case study in Northeast China , EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-7739, https://doi.org/10.5194/egusphere-egu25-7739, 2025.

Hailstorms have shown rising severity and frequency in recent years, posing a growing threat to crops and presenting significant challenges for the agricultural and insurance sectors in the face of climate change. As part of an interdisciplinary project (scCLIM, Seamless coupling of kilometer-resolution weather predictions and climate simulations with hail impact assessments for multiple sectors), this study focuses on assessing the impact of future hail occurrence on wheat across Europe.

We utilize results from high-resolution climate simulations with a grid spacing of 2.2 km, which were conducted using the COSMO regional climate model for both current and future climate. The future climate simulation, targeting a 3°C global warming scenario, was performed using the pseudo-global warming approach. Hail activity was simulated using the hail growth model HAILCAST, which was embedded within COSMO. A model of wheat phenology was used to estimate the wheat harvest dates based on COSMO outputs, enabling an assessment of the present and future exposure of wheat to hail.  By integrating high-resolution climate simulations with a crop phenology model, this approach bridges the gap between agricultural production and climate risks associated with extreme events. 

In this contribution, we examine the temporal and spatial alignment between hail events and crop development, with a particular focus on assessing the sensitivity of future risk of hail damage to wheat with respect to the interplay between changes in hail occurrence and earlier harvest dates. The results reveal regional variations in hail impacts on wheat across Europe, offering valuable insights into crop management, climate change adaptation strategies, and risk assessment within the insurance sector.

How to cite: Cui, R., Portmann, R., Thurnherr, I., and Calanca, P.: Assessing the impact of hail on wheat production in Europe under climate change, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-8356, https://doi.org/10.5194/egusphere-egu25-8356, 2025.

Wildfires are an emerging peril in traditional natural hazard risk assessment. Increasingly extreme fire behavior, unprecedented mega-fires and rising economic damages are commonly attributed to a combination of climatic shifts, expansion in areas where human development meets natural landscapes (wildland-urban interface), and an accumulation of fuel. 

Remote sensing products provide the most comprehensive data source for the global assessment of wildfires and their impacts. However, scientists and practitioners in Disaster Risk Reduction are faced with several fire products from different satellite missions, whose differences, advantages and limitations can be difficult to assess and understand, especially for users outside the remote sensing domain. At best, this issue complicates the process of identifying the most appropriate dataset, making it a challenging and time-consuming endeavor; at worst, it can result in inaccurate results. 

We address these issues by offering a concise overview of remote sensing fire products and clarifying terms that are interpreted differently across scientific communities, with a focus on their application in risk assessment. Our analysis centers on products representing burned area and active fire locations. While burned area products leverage several satellite overpasses and reflect the area affected by large fires best, active fire location products provide the fire radiative power, a measure of the fire intensity, which is an important metric linked to impacts. 

We present a historic wildfire hazard set, which combines burned area data and fire radiative power recorded by the Moderate Resolution Imaging Spectroradiometer (MODIS) satellite product for the years 2002–2023. We join this hazard set with exposure datasets (representing physical assets and population) and damage records to calibrate socio-economic vulnerabilities to wildfires. This forms the basis for estimating wildfire impacts and risks, necessary for prioritising adaptation options and the pricing of insurance.

How to cite: Steinmann, C. B., Koh, J., Kropf, C. M., Bresch, D. N., and Hantson, S.: Data requirements for assessing global socio-economic wildfire impacts and risks, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-8464, https://doi.org/10.5194/egusphere-egu25-8464, 2025.

How to cite: Keune, J., Di Giuseppe, F., Wetterhall, F., and Barnard, C.:  A return period-based early warning index for extreme precipitation , EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-8483, https://doi.org/10.5194/egusphere-egu25-8483, 2025.

The past few years have seen increasingly frequent and intense floods, culminating in 2024 with a year characterized by widespread and devastating inundations worldwide. In Europe despite major advancements in flood forecasting and flood protection measures undertaken in the past, in2024 heavy rainfall events resulted in severe flood impacts and massive socio-economic losses, claiming over three hundred lives. Extreme precipitation, breaking observed records, is expected to have an increased likelihood under global warming all the more we should question and scrutinize our knowledge about the frequency and severity of floods. Reliable estimates of these are challenging even for the current climate conditions, and disaster gaps can lead us to an underestimation of the risks. The use of the UNSEEN method (Thompson et al. 2017) has been proved to be very valuable in estimating both, unprecedented but plausible extreme floods and droughts.

Here we present a simple method to expand the UNSEEN method to develop storylines under various global warning levels. We selected precipitation scenarios with different spatial patterns for estimated return periods between 100 and 1000 years from pooled re-forecasts from ECMWF (ENSext and SEAS5), providing 8400 years of plausible weather sequences. The selected climate scenarios are perturbed by increasing the precipitation intensity according to the Clausius-Clapeyron relation for five different global warming levels, and used to run coupled hydrologic-hydraulic simulations. The results show that record-breaking, high-impact river floods are possible under the current atmospheric conditions, and climate change substantially aggravates flood impacts, as the relative increase in peak discharge can be significantly larger than the increase in precipitation, leading to a disproportionally high flood impact increase. The development of storylines of extreme flood events with a high spatial and temporal resolution are a valuable tool to explore, describe, and communicate extreme events and their dynamics. Such instruments are key for developing an informed vision and comprehensive protective measures in terms of flood risk management and emergency response.

References

Thompson, V., Dunstone, N. J., Scaife, A.A., Smith, D. M., Slingo, J. M., Brown, S. and Belcher, S.E.: High risk of unprecedented UK rainfall in the current climate. Nat Commun 8, 107, https://doi.org/10.1038/s41467-017-00275-3, 2017.

How to cite: Kauzlaric, M., Munz, L., Mosimann, M., Martius, O., and Zischg, A. P.: A simple approach for developing storylines of flood impacts under various global warming levels, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-8731, https://doi.org/10.5194/egusphere-egu25-8731, 2025.

Society has become more dependent on reliable electricity infrastructure to function normally. The ascending trend of blackouts in recent years suggests that today’s power system is becoming increasingly vulnerable to severe weather and puts an accent on an emerging issue that deals with power system resilience. Resilience, in this context, refers to the system's capacity to limit the extent, severity, and duration of service disruptions following extreme events.

To better understand and improve power system resilience, this study presents a comprehensive analysis of outage statistics in Sweden from 2007 to 2021, utilizing data from Energiföretagen Sverige. The findings reveal that approximately 26% of all outages are attributable to weather-related events, affecting nearly one-third of customers and contributing significantly to customer outage durations. These disruptions directly undermine the reliability and resilience of the power grid.

This research examines the correlation between specific weather phenomena—such as storms, heavy snowfall, and high winds—and the frequency and severity of power outages. The analysis identifies a strong connection between severe weather patterns and prolonged outages, particularly in rural and forested regions where overhead power lines are more vulnerable. By analyzing spatial and temporal patterns, this study identifies vulnerable areas within Sweden's power infrastructure and emphasizes the need for targeted resilience strategies. Proposed measures include enhanced vegetation management, infrastructure reinforcement, and the adoption of advanced grid technologies to mitigate the impacts of extreme weather events. These insights contribute to developing a more robust and reliable electricity system, better equipped to withstand future climate challenges.

How to cite: Duvnjak Zarkovic, S. and Messori, G.: Impact of Climate Extremes on Power Systems, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-8932, https://doi.org/10.5194/egusphere-egu25-8932, 2025.

Assessing climate risk due to climate change for the present and future periods has been the focus of both academic and applied research in recent years, reflecting its critical importance. In this study, we evaluated climate risks for the Marmara Region in northwestern Türkiye by integrating high-resolution climate projections with socio-economic data, aiming to inform and support regional climate policies.

To achieve this, we generated climate projections at a 0.025° x 0.025° resolution using the convection-permitting COSMO-CLM model, driven by EC-Earth3-Veg from CMIP6. These projections cover both the reference period (1995–2014) and a future period (2050–2059) under the SSP3-7.0 scenario) for a broader western part of Türkiye. The Marmara Region was selected as a focal area due to its vital economic significance, its diverse and densely populated urban centers, and its extensive agricultural areas. This approach allows for a comprehensive assessment of climate impacts on a region with critical socio-economic importance, providing actionable guidance to inform policy development and adaptation strategies.

We conducted a comprehensive climate risk assessment by integrating hazard data with components of sensitivity, vulnerability, and adaptive capacity components, which were derived from reliable socio-economic datasets provided by institutions such as the Turkish Statistical Institute and the Turkish State Meteorological Service. For the weighting phase, we employed multiple methodologies, including the Analytic Hierarchy Process (AHP), Principal Component Analysis (PCA), and variance-based distribution methods, to investigate their respective contributions to the final risk evaluation.

Preliminary findings reveal city-level climate risks for both the present and future periods, offering critical insights for key vulnerabilities and areas of concern. These results provide essential guidance for regional policymakers, enabling the identification of specific risk hotspots and developing targeted strategies that address the region-specific challenges. These results serve as a foundation for developing targeted strategies to mitigate climate risks, strengthening resilience, and enhance adaptation capacity in the Marmara Region.

How to cite: Moral, A. C., Yürük Sonuç, C., and Ünal, Y.: Climate Risk Analysis for Marmara Region, Türkiye, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-8933, https://doi.org/10.5194/egusphere-egu25-8933, 2025.

The increase in the number of heat days caused by climate change leads to intensified thermal heat stress for the human population, especially for vulnerable groups such as the elderly, children, and people with chronic illnesses. As climate change progresses, the demand for healthcare services will rise sharply in the coming years, considering that the number of heat days (Tmax > 30 °C) and tropical nights (Tmin > 20 °C) has already doubled or tripled in recent decades in Austria.

A transdisciplinary team of health, climate, and complexity scientists is needed to comprehensively investigate the effects and risks of climate change, with the focus on heat, on the health system. In a first step, the correlations between meteorological conditions (temperature, humidity, etc.) and health outcomes are analysed. To assess the effect of heatwaves on hospital admissions and deaths, data of daily maximum temperature, deaths, and hospital admissions per care region in Austria for the months June-September of the period 2007-2019 are used. In the detailed analyses, various definitions of heat waves, latency periods, and other factors are examined.

The investigated correlations between prevailing climate conditions and their effects on health are used to investigate future climate scenarios with respect to their conditions. Thus, projections can be made about imminent risks for people and consequently healthcare organisations. For this purpose, the different impacts of heat stress on staff, clients and management assessed with the participating healthcare organizations of the research project. Climate impact chains are developed and applied to ensure a systemic understanding of the risk, exposure and vulnerabilities. Derived adaptation measures  are subsequently identified at an institutional level. In addition, areas are identified in which the institutions have no influence and need support, for example through urban planning (e.g. greening and unsealing of outdoor areas not owned by the institutions, shaded path to an existing cooling center).

The results of the correlation analysis show significantly higher risk ratios for deaths in hospitals and for hospital admissions during heatwaves. This applies both to the population as a whole and to elderly people (>= 75 years). However, the increased burden is not only noticeable for clients, but also for healthcare staff, as analysed with the healthcare organisations within climate impact chains. The results indicate that there are some fields of action in which the institutions can take measures, such as regular training on the topic of heat, adapted uniforms, or adapting work processes and medication during heatwaves.

However, there are also areas in which healthcare organisations are dependent on the support and implementation of measures at city/regional level. For example, Nature-based solutions (Nbs) such as large-scale greening and unsealing are measures to reduce heat stress in the long term, thus reducing the strain on people – positively impacting health conditions. Furthermore, outdoor retreats are created in this way, reducing the burden of poor living standards.

How to cite: Bügelmayer-Blaschek, M., Ledebur, K., Hochebner, A., Schneider, M., and Klimek, P.: Healthcare organizations under heat stress: Risk assessment and solutions in Austria , EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-9152, https://doi.org/10.5194/egusphere-egu25-9152, 2025.

An early warning system Stromynazeleznici.cz (trees on railway tracks) has been developed to assist the national rail infrastructure administrator (Správa železnic, SZ) in managing the hazard of tree falls. A forecast of the tree-fall hazard on a 3-hour basis for the following three days is provided. The model incorporates data from weather forecasts (Aladin model) and a tree-fall susceptibility layer which delimits the locations where falling trees are capable of crossing railway tracks.

The tree-fall susceptibility layer is prepared from the raster of a normalized digital surface model. One-meter cells contain information about the absolute height of the surface above the relief model. All non-vegetated areas (all types of buildings, tall objects, bridges, masts, etc.) and areas with low vegetation that do not pose a hazard are filtered out. Impact zone buffers are defined for the remaining vegetation areas according to the actual height of the vegetation. The final output is a proportion of the length of railway lines per unit section which are threatened by falling vegetation.

Stromynazeleznici.cz contains tree fall evidence for recording, presenting, and exporting incidents. The forecast is based on a regression model programmed in R (server solution Project R). A multivariate logistic regression was chosen as the most suitable approach to construct the model according to cross-validation results and practical requirements. The following characteristics were selected as explanatory variables in the logistic regression: maximum daily wind gust, soil saturation index, snow index, the occurrence of thunderstorms, the season, the range of altitudes in the vicinity of the rail track, the median height of trees along the railway tracks, and the length of the rail track section with trees along the rail track.

The hazard level of tree falls is calculated for the "hectolines" (i.e., 100-meter segments) of the railway track. These are then aggregated into three levels of administrative units defined by SZ. The hazard level is calculated for three-hour intervals, covering a 45-hour forecast period – resulting in 15 time slots for each hectoline (the rail network in Czechia consists of 94,759 hectolines). The forecast is updated four times a day as new meteorological data become available.

The data is stored in a database and presented in the form of graphs, tables, and an interactive map. Hazard information can be found on the map: the tree-fall hazard level is represented by a five-level colour scale for individual administrative units. When zooming in, the risk is shown in relation to the hectolines. A timeline is located at the bottom of the screen, allowing users to switch between different time slots or aggregated time windows. Clicking on an administrative unit or hectoline will display the forecast and details for the selected element. The map also offers additional thematic layers — fallen trees, a layer showing vegetation susceptibility to falling onto the railway track, a tree health layer (derived from the Sentinel-2 data), and a forest tree species layer.

How to cite: Bíl, M., Nezval, V., Andrášik, R., Kubeček, J., Cícha, V., and Lepka, Z.: Early Warning System for Tree-Fall Hazards on Railways: An Example of a System Developed for the Czech Railway Infrastructure Administrator, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-9793, https://doi.org/10.5194/egusphere-egu25-9793, 2025.

One of the pressing challenges of our time is bridging the gap between climate science and decision-making to effectively manage risks from climate change, with weather and climate extremes being central to climate-related risk. Traditional climate science predominantly employs probabilistic approaches, generating large model ensembles to explore likely ranges of future conditions. While valuable, this approach often neglects low-likelihood, high-impact events that pose profound risks to society [1].

Strengthening the connection between climate science and decision-making is increasingly critical, particularly as the frequency and severity of extreme weather events rise. Integrated risk assessment and management require a holistic approach encompassing robust knowledge of potential impacts, hazard identification, risk monitoring, early warning and effective communication. While uncertainties in climate projections and predictions are unavoidable, they should not result in decision paralysis. Instead, the focus should be on interdisciplinary collaboration and enhancing links between climate science and decision-making through a better and more decision-relevant understanding of climate impacts [2].

This talk will address recent approaches, highlighting the importance of bridging disciplines and incorporating user-needs to address the complex challenges posed by climate risks. For instance, event-based storylines considering high-impact events, integrating system vulnerability and exposure to better assess risk will be discussed. When co-developed by climate scientists and stakeholders, storylines informed by physical climate and impact modeling provide actionable insights tailored to specific contexts.

References

[1] Sillmann J, Shepherd TG, van den Hurk B, Hazeleger W, Martius O, Zscheischler J, 2021: Event-based storylines to address climate risk, Earth’s Future, 9, doi: 10.1029/2020EF001783.

[2] Sillmann J, Raupach TH, Findell KL, Donat M, Alves LM, Alexander L, Borchert L, Borges de Amorim P, Buontempo C, Fischer EM, Franzke CL, Guan B, Haasnoot M, Hawkins E, Jacob D, Mahon R, Maraun D, Morrison MA, Poschlod B, Ruane AC, Shampa, Stephenson T, van der Wel N, Wang Z, Zhang X and Županić J, 2024: Climate extremes and risks: links between climate science and decision-making. Front. Clim. 6:1499765. doi: 10.3389/fclim.2024.1499765.

How to cite: Sillmann, J.: Climate Extremes and Risk: Connecting Climate Science and Decision-Making via Interdisciplinary Approaches Focusing on Climate Impacts, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-9881, https://doi.org/10.5194/egusphere-egu25-9881, 2025.

As the frequency and intensity of extreme weather events continue to increase due to climate change, risk mitigation has become a critical aspect of climate change adaptation in cities. The impacts of extreme weather events in cities are extremely diverse. Consequently,  integrative, systems-based approaches have been praised given their capacity to structure holistic risk assessments, account for both qualitative and quantitative data, and for accounting for the interactions between system components. Given the diversity and complexity of urban systems, interdisciplinary knowledge integration is critical in order to account for varied perspectives related to the impacts of extreme weather events on urban systems. Despite advances made to integrate different strands of knowledge through systems-based approaches, few methods exist to contextualize, analyse and evaluate its diversity. Assessing knowledge diversity exposes varying ways in which stakeholders identify and problematize the impacts of extreme weather events uncovering knowledge gaps and dominant knowledge framings that might hinder risk governance processes.  This study presents a novel methodology that integrates mental models and the social-ecological-technological systems (SETS) framework to assess and compare individual stakeholder perceptions of urban systems under the lens of an extreme weather event. By classifying system components and interactions into social, ecological, and technological domains, mental models enable the visualization of knowledge diversity, as well as the identification of potential gaps and silos in stakeholder understanding. The methodology is applied to New York City as a case study, engaging 20 stakeholders from diverse disciplines and sectors involved in mitigating the impacts of extreme precipitation. Findings reveal significant variability in how stakeholders emphasize SET domains and interactions. This methodology offers a transferable framework for assessing knowledge diversity in urban climate adaptation, emphasizing the importance of reflecting on stakeholder perspectives to identify gaps and synergies. By supporting more holistic and inclusive co-production processes, this approach provides a theoretical and empirical foundation for advanced modelling efforts that are capable of addressing the multifaceted challenges posed by climate change in urban environments.

How to cite: Herreros Cantis, P., Khromova, S., Olazabal, M., McPhearson, T., Langemeyer, J., and Neumann, M.: Knowledge Diversity for Climate Change Adaptation: A Social-Ecological-Technological Systems (SETS) Approach to Mental Models, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-11120, https://doi.org/10.5194/egusphere-egu25-11120, 2025.

Crop models currently have a limited capacity to accurately simulate the impacts of extreme climate events (ECEs), and there is considerable uncertainty across different models. Consequently, the assessment of food security risks from future climate extremes based on existing frameworks is less reliable. To address this issue at global scale, we are developing an advanced hybrid model that integrates process-based crop models with information on the occurrence of extreme climate events and a deep learning framework. Specifically, our model uses outputs from multiple crop models provided by the third round of the Inter-Sectoral Impact Model Intercomparison Project (ISIMIP 3a) as the initial input. The second input will consist of the daily occurrence of four types of extreme events: two related to temperature (heatwaves and coldwaves) and two related to precipitation (droughts and pluvials). We employ a Long Short-Term Memory (LSTM) network with an attention mechanism designed to dynamically capture the varying impacts of ECEs at different crop growth stages. The results are expected to offer a more precise simulation and deeper understanding of how ECEs affect food security. This study highlights the potential of AI-hybrid modeling to enhance the accuracy of crop impact assessments under climate change.

How to cite: Shan, B., Qian, H., Guan, X., and De Michele, C.: Enhancing crop yield simulations under extreme climate events using a hybrid model, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-12074, https://doi.org/10.5194/egusphere-egu25-12074, 2025.

In order to advance new theoretical and practical integration of Earth and Social Sciences to address the climate crisis and its impacts on society, the World Climate Research Programme has created the My Climate Risk (MCR) Lighthouse activity. The goal of MCR is to develop and mainstream a bottom-up approach to climate risk. To progress in this path, MCR has assembled regional centres (‘Hubs’) from institutions/researchers with knowledge in the field of climate risk and that allow this approach to be taken to local and regional scales. These Hubs comprise a variety of forms and modes of operation depending on the local interests and needs. In March 2022, the MCR CONICET Argentina Regional Hub was created (https://sites.google.com/view/mcrhubconicet). Through MCR CONICET Argentina Regional Hub its members learn, participate and motivate a scientific-technical and social perspective to promote adaptation and face climate extremes in Argentina employing the co-production of knowledge, storylines and multiple lines of evidence.

This work aims to share initiatives and projects from the Global South that are rooted in inter- and transdisciplinary dialogue and the inclusion of actors and institutions of the region, to address climate risk research. The case studies presented here address Argentina's need to improve hydrometeorological services availability, accessibility and interpretation. The first case study presents the coproduction cycle that led to a subseasonal novel local prediction product in northeastern Argentina, co-produced between climatologists, anthropologists and family farming actors within the framework of the CLIMAX project. The successful experience of this development highlights the importance of involving local communities in the development of climate information products that can be socially appropriated. A case of use of climate storylines as a tool for improving decision making is presented. Physical Climate Storylines was put in dialogue with Socio-anthropological Narrative Analysis around a drought event in Southeastern South America. Finally, the strategy of multiple lines of evidence is used, showing results of the “A River All Waters” project, which integrated transversal lines of work to address the impact of climate change on the Chubut River in Argentine Patagonia. This project shows a reduction in precipitation and an increase in temperature since 1960, which caused a decrease in river flows. These three case studies showed the need to explore novel methodologies that favour a bottom-up approach to regional and local climate risk.

How to cite: Cappelletti, L. M., Cánneva, J., Díaz, L., Fossa Riglos, M. F., Gulizia, C., Hernández, V., Incicco, C., Hurtado de Mendoza, M. S., Mindlin, J., Pansa, D., Pessacg, N., Prudente, C., Rivera, J. A., Robledo, F., Rosales, D. A., Ruscica, R. C., Sörensson, A. A., and Testani, N.: A bottom-up approach to climate risk from the Global South: the case of the CONICET Argentina My Climate Risk Hub, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-12539, https://doi.org/10.5194/egusphere-egu25-12539, 2025.

The "Hamburg Pluvial Flood Risk Map" aims to improve our understanding of the drivers, dynamics and interactions of climate-induced (disaster) risks in Hamburg. Following the risk framework of the IPCC, we calculate a risk index based on hazard, exposure and (social) vulnerability. In this sense, we combine data from the previously published Social Vulnerability Index (von Szombathely et al., 2023) with novel meter-scale hydraulic simulations of urban flooding provided by the heavy rain hazard map of the city of Hamburg (BKG/FHH 2023). We have enhanced the modeling of social vulnerability by applying the TOPSIS method and the Shannon Entropy procedure. and propose a high-resolution exposure modeling designed for urban flooding, with different exposure layers threatening health and restricting mobility and accessibility. We show that fundamentally new spatial patterns emerge for pluvial flood risk in Hamburg, which differ from familiar socio-economic urban structures and at the same time differ clearly from a pure representation of the hazard. Presented through high-resolution spatial maps, this analysis aids in identifying adaptation needs and prioritizing policy measures for climate change adaptation.

References:

BKG/FHH 2023. Eine Starkregen-Gefahrenkarte für Deutschland. https://www.business-geomatics.com/2023/02/02/eine-starkregen-gefahrenkarte-fuer-deutschland/

von Szombathely M., Hanf F. S., Janka B., Meier L., Ossenbrügge J., Pohl T. 2023. An Index-Based Approach to Assess Social Vulnerability for Hamburg, Germany: International journal of disaster risk science. 14, 5, p. 782-794 13 p. DOI: 10.1007/s13753-023-00517-7

How to cite: von Szombathely, M., Behrens, J., Hanf, F. S., Lennartz, M., Poschlod, B., Vogelbacher, A., and Sillmann, J.: Hamburg Pluvial Flood Risk Map, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-13534, https://doi.org/10.5194/egusphere-egu25-13534, 2025.

Vegetables are full of nutrients that are difficult to obtain from meat or grains, such as vitamins, minerals, and dietary fiber, but they are vulnerable to abiotic stress, making it difficult to obtain consistent yields. Climate extreme events have caused a decline in vegetable production, often leading to elevated vegetable prices. Here, we investigate how climatic factors influence vegetable price changes in China, focusing on colder months when extreme weather impacts are more pronounced. We found three major patterns in vegetable consumer price index (VCPI) data, including data from 31 provinces in China from 2003 to 2023. The first empirical orthogonal function (EOF) mode shows that vegetable prices in all provinces vary together, and this is linked with temperature variations in China. The second EOF mode has a north-south dipole spatial pattern, and it is linked to low-temperature events in southern China, which are closely linked to Arctic warming during colder months and central Pacific La Niña occurrences, especially in December. In addition to temperature, precipitation also affects vegetable prices, with cold rain and snow contributing to VCPI increases resulting from the third EOF mode. Also, the third mode, showing an east-west dipole pattern, is associated with eastern Pacific El Niño occurrences during January and February. Major VCPI patterns and relevant climate factors will facilitate the prediction of vegetable prices on a seasonal time scale and can be used as scientific evidence to prepare for a surge in vegetable prices by combining with seasonal climate forecasts. As China accounts for half of the world’s vegetable production and fluctuations in its vegetable prices can profoundly affect global food security, our findings would be useful to support stable vegetable production, ensure food security, and minimize economic losses globally.

How to cite: Qiu, Y. and Kim, J.: Impact of climatic variability on vegetable price in China during colder months, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-14763, https://doi.org/10.5194/egusphere-egu25-14763, 2025.

This work investigates how the damages in cross-sectoral climate change risk hotspots can be assessed drawing on methodologies developed and applied to eight case studies conducted within the EU CROSSEU project. Hotspot analysis represents several key challenges in the assessment of the impacts of climate hazards and the focus here is on extreme climate events rather than on the impacts of gradual climate change. Hotspots are defined as areas where climate events are likely to generate high potential damages. The hotspot identification methodology provides a framework for identification of context specific vulnerabilities due to a combination of factors, including the magnitude of extreme climate events (physical aspects), the presence of critical infrastructure and vulnerable populations (socio-economic aspects), and the sector specific vulnerabilities as well as interconnectedness of different sectors (cross-sectoral aspects). The identification of hotspots is based on a combination of quantitative and qualitative data, including climate projections, socio-economic data, and stakeholder consultations.

The hotspot methodological framework is applied to a range of case study sectors and geographical settings. The case studies cover heat waves in Czech Republic and United Kingdom; drought in regions of Germany, Czech Republic, Poland, and Romania; floods in Denmark, Germany, and Italy; and snow avalanches in the Alps and Carpathian Mountains. While three other case studies addressed climate change impacts and spillover effects in the Lower Danube region and across Europe- particularly on renewable energy infrastructure and agriculture.

In terms of physical vulnerabilities, the case studies demonstrate that Prague and Southern Moravia in the Czech Republic, and London in the UK, are hotspots for heat-related mortality and morbidity, and specific social and structural vulnerabilities in these areas are related to high population densities, aging populations, and the urban heat island effect.  Several regions in Germany, Czech Republic, Poland, and Romania are identified as hotspots for drought. The economic vulnerability of these regions is primarily due to the reliance of agriculture on rainfed water sources. Coastal cities in Southern Denmark and Northern Germany are vulnerable to storm surges, impacting thousands of residents by disrupting daily life, socioeconomic activities, restricting movement and even necessitating temporary relocation. The mountainous areas of the Trentino Alto Adige region in Italy are hotspots for debris flows and flash floods, and are vulnerable due to their low-lying coastal areas, high population densities, and critical infrastructure. The Italian Alps and the Făgăraș Mountains in the Romanian Carpathians are hotspots for snow avalanches with potential high economic losses for tourism. The Lower Danube region is a hotspot for both droughts and floods, posing significant risks to a unique biodiversity ecosystem, as well as to agriculture, energy infrastructure, and human settlements.

This hotspot analysis in the CROSSEU project provides key comparative risk assessment measures, contributing to the establishment of effective adaptation strategies in the EU and also at regional levels.

This research received funds from the project “Cross-sectoral Framework for Socio-Economic Resilience to Climate Change and Extreme Events in Europe (CROSSEU)” funded by the European Union Horizon Europe Programme, under Grant agreement n° 101081377.

How to cite: Some, S., Halsnæs, K., Cheval, S., Micu, D., Skougaard Kaspersen, P., Adamescu, M., Arhire, G., Borga, M., Calzadilla, A., Charousset, S., Dessens, O., Falcescu, V., Franceschinis, C., Giucă, R., Igescu, D., Jenkins, K., Vasilakos, N., Nielsen, K., Nertan, A., and Oueslati, B. and the et al.: Cross-Sectoral Climate Change Risk Hotspots in Europe: Insights from CROSSEU Case Studies, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-15952, https://doi.org/10.5194/egusphere-egu25-15952, 2025.

In emergency management, the gap between scientific risk knowledge and operational decision-making remains a persistent challenge for early warning systems. While vast amounts of data—ranging from risk assessments to historical event records—are available, they are often underutilized due to the complexity and fragmentation of information sources. This research proposes an innovative approach to bridge this gap by integrating Retrieval-Augmented Generation (RAG) with domain-specific knowledge graphs to enhance situational awareness and decision support in emergency operations centers.

The proposed solution focuses on developing a graph-based RAG pipeline that interacts with an external repository of risk data on Italy, specifically tailored for emergency response personnel, including civil protection agencies and the Italian Red Cross. The repository incorporates emergency plans, historical events, risk assessments, civil protection guidelines and legislation, and real-time updates from external sources such as news and media. By structuring the data through a knowledge graph aligned with established risk frameworks (e.g., RISK INFORM), the system enables precise, explainable, and contextual information retrieval.

Key features of the tool include an explainability module for transparency, a PDF parser for document integration, and a web interface that allows users to interact with the system through natural language queries. For example, an analyst responding to severe floods in Northern Italy could query the system for demographic data, flood risk hotspots, and critical infrastructure at risk, receiving actionable insights grounded in both historical and live data.

The project demonstrates how AI-driven approaches, when combined with structured domain knowledge, can make early warning systems more effective by improving accessibility, scalability, and interoperability across sectors. The use of knowledge graphs ensures data explainability and traceability, addressing key challenges in emergency management, such as trust in AI outputs and timely decision-making. The platform, currently under development, aims to serve as a proof-of-concept for future applications in multi-hazard early warning systems.

This research contributes to the evolving field of AI-enhanced early warning systems, offering a novel, trans-disciplinary methodology that combines data science, emergency management, and humanitarian operations to improve anticipatory action and disaster preparedness.

How to cite: Bove, J.-B., Rudari, R., Trasforini, E., D'Andrea, M., Massucchielli, L., and Gioia, A.: Bridging Risk Knowledge and Operational Outcomes through Retrieval-Augmented Generation and Knowledge Graphs for Early Warning Systems, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-16239, https://doi.org/10.5194/egusphere-egu25-16239, 2025.

Low- and middle-income countries are often vulnerable to extreme weather events and simultaneously have limited access to insurance markets, leaving damages largely uncovered. To address the issue, the World Bank advocates for a risk-layering approach. Part of this approach is the issuance of catastrophe bonds (CAT bonds), especially for high-risk layers and when capital requirements are substantial. However, access to the CAT bond market remains limited. Market entry of potential issuers, but also investors, is hindered by the extensive, and most often, expensive technical knowledge needed. Additionally, high investor premium demands pose a great challenge for low- and middle-income countries, a situation likely to worsen with climate change intensification.
We propose three interdependent solutions to enhance countries and investors access to the CAT bond market. First, we develop an open-source and -access CAT bond tool and implement it within the CLIMADA environment, an open-source global risk assessment platform. The tool allows for the design and evaluation of either potential or already existing CAT bonds and was tested in a tropical cyclone insurance case study in Samoa. Second, we propose a multi-country CAT bond design, which pools risk across nations. We apply this design to a case study of tropical cyclone risk in Small Island States (SIDS). We find that such pooling allows to decrease both capital requirements and premiums by up to 27% and 17%, respectively, while still offering competitive returns to investors. Third, we introduce a financial scheme addressing premium support, capital supply, and greenhouse gas reduction incentives. We apply the scheme in a case study to the SIDS utilizing the previously developed CAT bond tool and the presented pooling approach. These solutions together aim to expand access to risk transfer for vulnerable countries, offering a more
sustainable and affordable pathway to disaster resilience.

How to cite: Bergmüller, K. O., Håkansson, V. W., Juhel, S., Bresch, D. N., and Kropf, C. M.: Closing the Insurance Gap – Enhancing Access to the CAT Bond Market, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-17829, https://doi.org/10.5194/egusphere-egu25-17829, 2025.

Myanmar is highly vulnerable to climatic changes in extreme weather such as increased precipitation and extreme temperatures. During the dry season of the last two years, Myanmar has already suffered such events. High-resolution climate model simulations are urgently needed to understand the complexity of future impacts of extreme weather and climate change in Myanmar. While global climate model simulations cover the region with a horizontal resolution of around 100 km, most regional climate models available over this region have a resolution of up to 25 km. This is still not enough to accurately assess the vulnerability to climate change for such a diverse country with complex topography, and thus, new high-resolution simulations are needed to understand the effect of climate change on regional to local scales.

We are conducting dynamically downscaled climate simulations across a large part of Myanmar using the Weather Research and Forecasting (WRF) model. Downscaled climate data are generated for five simulation periods: one for the present (1981-2010) and two for each of two future periods (2031-2060 and 2071-2100), under both the intermediate-emission shared socioeconomic pathway (SSP2-4.5) and the very high-emission pathway (SSP5-8.5). The simulations are performed at a high spatial (5 km) and temporal resolution. Through these simulations, we can achieve more realistic precipitation patterns and detailed information on local precipitation and temperature extremes, considering also the daily cycle. 

We will present our preliminary findings from the downscaled modelling of weather extremes and information about heat stress and drought indices. This will provide insight into potential impacts on food security and fragility to climate change in general, both of great implications for local society and economy.

How to cite: Messmer, M., González-Rojí, S. J., Mo Aung, N. C., Hunt, G., and Leonard, S.: Assessment of Climate Fragility in Myanmar based on high-resolution regional climate model simulations, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-18200, https://doi.org/10.5194/egusphere-egu25-18200, 2025.

The private sector and industry are increasingly accessing physical climate data to a) identify and disclose climate risk as required by regulations and b) seek to estimate and limit both present and future economic impacts of physical risk on their business.

Projections of physical climate risk and associated changes in estimates in losses are typically provided to private sector stakeholders in isolation from the wider social and systemic picture. Whilst individual private sector stakeholders can take some measures to minimise impacts from climate hazards, a collective approach in conjunction with local communities and the public sector may be more effective, in terms of both risk reduction, and upfront cost.

We provide physical climate storylines and narratives as a complement to the typical physical risk data provided to private stakeholders. We work with many private stakeholders and are seeking interdisciplinary discussion and collaboration with a view to exploring and quantifying the cost- benefit of individual stakeholder action versus collective action.

How to cite: Sagoo, N. and Leach, N.: Exploring how physical climate storylines and narratives impact private stakeholder behaviour, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-20300, https://doi.org/10.5194/egusphere-egu25-20300, 2025.

The China-Pakistan Economic Corridor (CPEC), a cornerstone of China’s Belt and Road Initiative (BRI), has significantly enhanced economic connectivity, infrastructure, and mobility between China and Pakistan, with investments exceeding $60 billion. However, this rapid transformation raises critical questions about the intersection of climate, trade mobility, and public health. Increased connectivity under CPEC may amplify the risk of Japanese Encephalitis (JE), a zoonotic, mosquito-borne disease endemic to several Asian countries, including China. JE transmission is influenced by complex ecological and climatic factors, including temperature, precipitation, and land-use changes, which impact mosquito vectors (Culex tritaeniorhynchus) and their habitats.

This study evaluates the risk of JE outbreaks in Pakistan through a One Health framework, highlighting the interplay of climate, mobility, and health. Specifically, it focuses on cross-sectoral collaboration across public health, veterinary, and environmental agencies to mitigate emerging threats. Objectives include assessing JE transmission risks along CPEC and proposing climate-sensitive, One Health interventions for prevention and control.

The risk assessment integrates data from human health, veterinary, and environmental sectors using interdisciplinary methodologies:

• Climate and environmental mapping of Culex breeding sites along CPEC using satellite imagery and meteorological data, identifying that over 50% of CPEC-associated regions, particularly in Sindh and Punjab, have optimal conditions for mosquito breeding due to rice paddies, irrigation systems, and seasonal climatic variability.
• Analysis of trade and mobility data, showing a 240% increase in human and animal movement along CPEC, intensifying vector and amplifying host exposure.
• Stakeholder interviews, revealing critical gaps in JE surveillance, real-time communication, and coordinated climate-informed response strategies.

Findings highlight that approximately 60% of identified high-risk areas are vulnerable to JE outbreaks, driven by favorable climatic and environmental conditions for vector proliferation. This underscores the urgent need for integrated strategies that account for climate variability and its impacts on vector dynamics.

Key recommendations include:

• Developing GIS-based vector surveillance systems to monitor climate-driven changes in mosquito breeding habitats.
• Establishing real-time data-sharing platforms for JE surveillance between China and Pakistan, incorporating climate and environmental data.
• Promoting JE vaccination programs for vulnerable populations in high-risk, climate-sensitive areas.
• Enhancing diagnostic and response capacity across public health and veterinary laboratories through climate-informed training initiatives.
• Raising community awareness through public health campaigns on vector control, emphasizing climate adaptation strategies.
• Formulating a joint JE outbreak preparedness and response framework, integrating climate projections, vector control, and rapid response teams.

This study demonstrates the critical need for interdisciplinary approaches to address health risks at the nexus of climate, trade mobility, and emerging infectious diseases. By integrating climate science into One Health strategies, the research underscores how transdisciplinary collaboration can build resilience against the multifaceted challenges of climate change and global connectivity.

How to cite: Ahmad, S. and Idrees, F.: From Trade Routes to Transmission Routes: Climate, Mobility, and Risk Assessment of Japanese Encephalitis Outbreak in Pakistan under the China-Pakistan Economic Corridor: A One Health Approach, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-529, https://doi.org/10.5194/egusphere-egu25-529, 2025.

Growing evidence demonstrated that exposure to ambient fine particulate matter (PM_2.5) increases mental health risk via neuroinflammation and oxidative stress. However, less is known about the relative contribution of PM_2.5 originating from different emission sources on mental health, such as depression and anxiety, particularly in low- and middle-income countries like India. Therefore, we examined the associations of short- and long-term exposure to total and source-specific PM_2.5 with depression and anxiety in Indian adults.

A cross-sectional analysis has been conducted in 12 Indian states using data from the National Mental Health Survey (NMHS), 2015-16, a nationally representative and population-based study in India. This study includes a total of 34,357 participants, 18 years and older. The 1-month and 12-month mean exposure to PM_2.5 and its source originating from 8 emission sources were assessed using a 1 km x 1 km high-resolution satellite-derived database and the WRF-CMAQ model, respectively, at participants' residential addresses before the NMHS interview date. The Mini International Neuropsychiatric Interview (MINI) version 6.0.0 was used to evaluate depression and anxiety disorders in adults. Adjusted odds ratios (ORs) were estimated for depression and anxiety per IQR increase in PM_2.5 using a logistic mixed-effects regression model after adjusting for the individual and household level covariates.

In this study, the weighted prevalence of the current depressive and anxiety disorders among adults was 2.69% (95% CI-2.66-2.72) and 2.96% (95% CI-2.93-2.99), respectively. The estimated mean PM_2.5 exposure for 1-month and 12-months was 55.8±19.6 and 44.3±13.5 µg/m³ respectively. Each IQR increase in PM_2.5 exposure was significantly and strongly associated with depressive disorder (OR = 1.13; 95% CI: 1.05–1.21) for a 1-month exposure window and anxiety disorder (OR = 1.16; 95% CI: 1.07–1.26) for a 12-month exposure after adjusting for potential confounders. PM_2.5 originating from different emission sectors was associated with mental health outcomes, with the strongest associations for power, transport, international transboundary, and domestic sources for at least one health endpoint, whereas agricultural sources showed protective associations with both outcomes. Subgroup analyses showed stronger associations among individuals with lower household incomes and lower education.

Our study suggests that interventions to reduce PM_2.5 from key emitting sources may reduce the burden of mental health in India, although cohort studies are recommended to determine the causal relationship.

How to cite: Kundu, P., Dey, S., Krishnan, A., Ghosh, S., N Rao, G., Benegal, V., Varghese, M., and Gururaj, G.: Association between exposure to fine particulate matter from different emission sources and mental health outcomes in India, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-687, https://doi.org/10.5194/egusphere-egu25-687, 2025.

The spread of malaria is a major health burden, affects many people in Africa, depends on climate but also socio-economic conditions. Thus, it is important to gauge the impact of anthropogenic global warming on malaria and attribute anthropogenic causes. Here we compute the Time Of Emergence (TOE) of vector density and of the Entomological Inoculation Rate (EIR) in the SSP3-7.0 scenario using 50 bias-corrected members of Community Earth System Model version 2 (CESM2) Large Ensemble simulations. This reveals that vector density, which depends on climate conditions, and EIR, which depends on both climate and population density, will rise significantly and permanently above the pre-industrial background variability due to anthropogenic causes in Africa. Both the vector density and EIR have areas, mainly in central Africa, where anthropogenic causes have already significantly changed, and many more areas will experience anthropogenic caused changes in the 2030 and 2040s and towards the end of this century. Our simulations also show clear evidence that extremes of vector density and EIR increase in the future by almost 100%, suggesting that major malaria epidemic outbreaks will become much more likely. We also perform simulations with constant population and with no climate change which partly reveal underlying malaria dynamics. Our results highlight the need to prepare for an expansion and intensification of the malaria burden if no health interventions are being taken.

How to cite: Franzke, C. and Parihar, R. S.: Time of Emergence and Future Projections ofExtremes of Malaria Infections in Africa, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-2188, https://doi.org/10.5194/egusphere-egu25-2188, 2025.

While increasing heat is a direct impact of climate change on health, epidemiological studies are quite limited in India. Here we examined the impact of heat on neonatal (children less than 28 days) and very early (children died on the first day) neonatal mortality using the fifth round of the National Family Health Survey (NHFS-5) dataset for 2019-2021. For heat exposure, we used a global daily temperature dataset at 1-km by 1-km saptial scale. First, we evaulated the global temperature dataset with station-based measurements for India and found reasonable accuracy for further application. In the NHFS-5, health and demographic information was collected from 30456 clusters spanning across urban and rural India covering every district. The estimated very early neonatal mortality and neonatal mortality values were 17.1 and 23.4 per 1000 live births, respectively. We then assinged exposure to daily maximum and minimum temperature at household level, and using a generalized logistic regression model, estimated the effect of heat after adjusting for the covariates. For every 1 degree increase in maximum and minimum temperature, very early neonatal mortality increase by 2.7% (95% CI: 1.6-3.8) and 2.0% (1.0-2.9), respectively. We found larger effect of heat on neonates born in the 'poorest' households (3.3% and 3.0% highesr risk for every 1 degree increase in maximum and minum temperature) with the effect declining (but still significant) with an increase in wealth index. We also found larger effect on male child than on female child, and on neonates in rural region than in urban region, and the effect fizzles out with a few days lag. As temperature is expected to rise further due to climate change, adequate adaptation startegy is required to protect the most vulnerable group; without which India cannot meet the sustainable development goal of reducing very early neonatal mortality and neonatal below 7 and 10 per 1000 live births, respectively, by 2030.          

How to cite: Dey, S., Gaur, G., and Philip, S.: Impact of heat on neonatal mortality in India, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-2681, https://doi.org/10.5194/egusphere-egu25-2681, 2025.

Plastic burning can significantly contribute to the overall particulate matter (PM) burden in developing countries, where inadequate waste management and low public awareness often result in open refuse burning. However, their chemical composition and health-related properties are largely unelucidated. In this study, we generated PM through controlled combustion of five widely used plastic materials. Our findings reveal that metals and polycyclic aromatic hydrocarbons (PAHs) detected in the plastic samples may drive oxidative stress through ROS formation. We observed significant quantities of EPFRs and ROS in the aqueous extracts of the PM. Additionally, plastic burning PM showed excessively high levels of reactive chlorine species (RCS). The oxidative potential, a key metric for PM toxicity, was assessed using acellular assays- OP-DTT and OP-OH. A kinetic box model was employed to simulate OP-OH, focusing on the rate of hydroxyl radical (•OH) formation. The model integrated reactions involving PAHs, metals, EPFRs, ROS, and RCS, using rate constants from established literature. It reasonably predicted •OH formation rates for the five types of plastics tested. Our results suggest that radical production is driven by complex chemical mechanisms, including redox cycling of active components, ROS cycling, Fenton chemistry, and organic oxidation reactions. Given the widespread use of plastics and growing environmental concerns around plastic pollution, this study highlights the urgent need for stricter regulations and improved waste management practices, especially in developing countries. Further details will be presented.

How to cite: Salim, R., Kapur, S., Schervish, M., Edwards, K., Gerritz, L., Raghunathan, R., A. Nizkorodov, S., S. Gunthe, S., and Shiraiwa, M.: The health effects of plastic burning particulate matter- the role of metals, Polycyclic aromatic hydrocarbons, environmentally persistent free radicals, reactive oxygen and chlorine species in inducing oxidative stress in human body, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-4125, https://doi.org/10.5194/egusphere-egu25-4125, 2025.

How to cite: Jiang, S., Li, C., and Wu, X.: Impact of global warming on the anemia among women of reproductive age in the global south, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-4778, https://doi.org/10.5194/egusphere-egu25-4778, 2025.

Recent research has reported an increase in Ground Level Ozone (GLO) concentrations in South Asia, with ongoing climate change being one of the contributing factors to this rise. In the lower-middle income economies such as India, studies related to cognitive impacts of GLO are insignificant and scarce. This time-series study aims to quantify the risk of cognitive disorders associated with ozone exposure for Delhi. The high-resolution gridded (5km*5km) daily maximum 8-hour mean ozone concentration data (MDA8) retrieved from WRF-Chem simulation were linked to geocoded-anonymized daily hospital admissions data of several cognitive disorders (e.g., depression, anxiety, Parkinson's disease, etc.). The WRF-Chem model was simulated for the Delhi domain over a four-year (2016-2019), for the same period hospital admissions data were collected. The generalized additive model (GAM) with Poisson distribution was utilized for examine an association of O₃ exposure with cognitive disorders. The delayed effect of exposure was assessed employing 20-days lag. The results of relative risk (RR) against lag days showed inverted-U shape curve with highest RR of 1.0092 (95% CI: 1.0051-1.0134) on 10^th lag day. The age-gender-stratified analysis revealed that females (RR: 1.0085, lag-day: 17) exhibited slightly higher risk compared to males (RR: 1.0071, lag-day: 9), while the younger demographic (age≤60 years) were at marginally elevated risks than elderly (age>60 years). In India, the mitigation measures and policies are predominantly aimed at reducing particulate matter pollution. The findings of this study are pertinent to present and future contexts, whereby evidences of intensifying effects of climate change on ozone are more pronounced than particulate matter. The research offers significant insights into the relationship between ‘public health’ and ‘air pollution’, contributing to the existing literature on highly polluted urban environment like Delhi.

Keywords: Ground level ozone (GLO); Cognitive impacts; Generalized Additive Model (GAM); WRF-Chem

How to cite: Parmar, P. D., Chandra, M., Sharma, S., and Kota, S. H.: Cognitive impacts of Ground Level Ozone (GLO) exposure in Delhi: Estimating a risk in the highly polluted urban environment, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-4878, https://doi.org/10.5194/egusphere-egu25-4878, 2025.

Background: Invasive alien plant species offer enormous ecological and public health risks worldwide, with Kenya experiencing some of the most severe consequences. Non-native flora outcompete indigenous species, reducing local biodiversity, agricultural production, and grazing areas, affecting food security and rural livelihoods. Parthenium hysterophorus (Asteraceae), a highly invasive weed, poses considerable concern due to its capacity to alter ecological systems. Climate change exacerbates these difficulties by altering rainfall patterns and temperatures, enabling invasive species to spread and thrive. As a result, these modifications frequently increase mosquito-breeding areas, which exacerbates the transmission of malaria, dengue, and other arbovirus diseases. Female mosquitoes, the primary vectors of these pathogens, require either blood meals or plant-derived sugars, despite the widespread acknowledgment that arboviral illnesses are highly recognized as serious public health concerns, little is known about how invasive plant species affect mosquito populations or arboviral transmission. This study examines the influence of P. hysterophorus on mosquito vector abundance, diversity, and arbovirus dynamics in the Kenyan Rift Valley area.

Methods: Mosquitoes were collected from six villages with varying levels of P. hysterophorus infestation—three heavily invaded and three free from P. hysterophorus. Using a combination of trapping techniques, approximately 50,000 mosquitoes representing 48 species were captured and identified. This comprehensive survey evaluated mosquito abundance and diversity, providing critical insights into the ecological impacts of invasive alien species on arboviral vector populations.

Conclusions: The findings will elucidate the complex interplay between invasive alien plants, land-use changes, and mosquito vector dynamics, shedding light on the mechanisms driving arbovirus transmission. This study will inform precise vector control strategies and deepen our understanding of the ecological impacts of invasive species on public health, including their role in the spread of diseases. This study will not only guide more targeted vector control strategies but also enhance our understanding of the broader ecological and public health impacts of invasive species in Kenya, particularly in disease spread.

How to cite: Osman, T., Chiuya, T., Fevre, E., and Borgermeister, C.: Weed, Mosquito, Virus: The Ecological Triad Shaping Disease Transmission in Kenya, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-5254, https://doi.org/10.5194/egusphere-egu25-5254, 2025.

Climate change is a key determinant of public health, influencing disease patterns, and public and environmental well-being. Mosquito population dynamics are largely determined by climatic factors and water availability. Therefore, understanding the linkage between local water resources and mosquito dynamics is crucial for better predicting current and future health risks, and informing effective disease control and health risk reduction. Here, we investigate how temporal and spatial distribution of water availability affects mosquito populations in a natural wetland area under current and future climate r scenarios. The study was conducted in the natural park of the  Aiguamolls de l’Empordà,  connected to La Muga and El-Fluvia river basins in Catalonia, Northeast Spain. Empirical data on river runoff and local water levels were collected from several discharge stations, while abundance estimates of mosquito populations were obtained from mosquito traps spread across the study area. The hydrological assessment is carried out with the Soil Water Assessment Tool (SWAT). This model uses observed rainfall and air temperature from the gridded earth observation dataset over Europe (E-OBS) to simulate streamflow and hydrological responses of the study area. Based on the in situ data and hydrological simulation outputs, we derive a relationship between water availability and mosquito population abundance that can be used to predict future disease risk in the study area. Our results will be integrated within the “Infectious Disease decision-support tools and Alert systems to build climate Resilience to emerging health Threats (IDAlert)” project funded by the European Union. This decision-support tool plays a critical role in targeted interventions in water management and the health sector, directly contributing to reducing health risks due to mosquito-borne diseases.

Keywords: Health risk, SWAT, Spatial and temporal distribution of water, Mosquito populations, IDAlert

How to cite: Sirisena, J., Stiles, P., Rodriguez, J., Berenguer, S. B., Bartumeus, F., Costa, M. M., and Bouwer, L. M.: Developing a prediction model for the relationship between climate, water resources, and mosquito dynamics: application to a case study in  a human-impacted Mediterranean wetland, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-5618, https://doi.org/10.5194/egusphere-egu25-5618, 2025.

EpiOutlook is an epidemiological indicator platform currently under development as part of the IDAlert project, a research consortium taking a OneHealth approach to understanding the impacts of climate change on the emergence and spread of infectious diseases in Europe and Bangladesh. The aim of the platform is to provide short-term early warning indicators of epidemiological risks, including those related to extreme weather and climate-sensitive infectious diseases (CSIDs). Currently, one climatic extreme indicator is operational for use in EpiOutlook which relates to drought, and makes use of the Standardised Precipitation-Evapotranspiration Index. Here, we propose two new climate extreme indicators currently under development, one related to heat and a second related to flood risk. We make use of fine-scale climate data (0.25x0.25) to categorise grid cells by the two proposed indicators (heat and flooding), which can then be extrapolated to the scale of interest. The impacts of extreme heat on health are well documented (e.g., extreme low and high temperatures and humidity leading to more adverse health outcomes), particularly for vulnerable groups such as pregnant women and children. Less well established are specific temperature ranges which puts people at risk to the highest number of climate-related health risks including CSIDs. We propose making use of our current CSID indicators (malaria, tick-borne diseases, leishmaniasis, Vibrio spp., West Nile Virus and Aedes-borne diseases), all of which consider the impacts of temperature and humidity. We aim to categorise temperature and humidity ranges which create ideal conditions for the highest number of CSIDs, weighed against the non-communicable disease impacts such as heat stress/stroke and adverse pregnancy outcomes, to provide a comprehensive spatial and temporal outlook for the effects of heat on health. Flooding is a major climatic risk in Europe, leading to destruction of property and livelihoods and infectious disease risk. We aim to develop a simple categorisation of flood risk via fluvial and pluvial flooding over Europe, incorporating several elements of the traditional water balance model, but producing an output which will be more interpretable by a wider range of end users. Risk will be assessed based on precipitation, elevation, land cover, potential evapotranspiration/soil moisture, groundwater recharge rate, proximity to a river, and river runoff/flow. Coastal flooding will not be considered at this stage, due to different flooding mechanisms, and instead proximity to a coastline will be seen as preventative as a source of drainage. Flood risk will be validated using flood data, and if the categorisation is proved accurate in representing flood risk, the occurrence of a flood in the preceding years will be considered in the categorisation. We aim to use the results from the flood indicator to provide valuable input to a leptospirosis indicator, a water-borne disease which is closely related to flooding. We believe that these indicators will provide easy to interpret quantification of climate extremes which relate to health in Europe, useful for public health decision-makers to make necessary adjustments to the current and near future risks posed by climate change.

How to cite: Charnley, G. E. C., Ball, E., Llabrés-Brustenga, A., San José Plana, A., Colgate, A., and Lowe, R.: Developing climate extreme indicators for EpiOutlook, a climate-informed subseasonal-to-seasonal forecast platform for epidemiological risks, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-6073, https://doi.org/10.5194/egusphere-egu25-6073, 2025.

Vector-borne diseases are responsible for over 700,000 deaths annually and are expected to spread to new regions and become more frequent due to climate change. This is primarily because vectors (such as insects and ticks) are ectothermic ("cold-blooded") and highly influenced by environmental conditions.

We are broadly interested in better understanding how climate change, in combination with land use changes, will affect the spread and frequency of vector-borne diseases. This will be achieved by using machine learning models, such as random forest and deep learning algorithms, to predict disease spread and frequency. By adopting a One Health approach, where we consider human health as interconnected with animal and environmental health, we will integrate multiple data sources (e.g., climate, land use, socio-economic factors, human health, and animal health) to improve our predictions.

As an example, in the EU project Planet4Health, we will employ various machine learning models to predict outbreaks of vector-borne diseases (leishmania and mosquito-borne diseases) in the Iberian Peninsula. The project aims to identify the model that gives the most accurate and meaningful predictions, and later incorporate it as part of an early warning systems for predicting such outbreaks.  

How to cite: Fossen, E.: Using machine learning to predict vector-borne diseases in a changing climate, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-6441, https://doi.org/10.5194/egusphere-egu25-6441, 2025.

In the framework of the Next Generation EU program, the Vitality project was founded to 
develop different research activities related to the sustainability and environmental protection. 
One of them, called One health: Telemedicine and Environment coordinated by the University 
‘G. d’Annunzio’, Italy, is focused on the impacts of climate changes and pollutant changes, on 
the evolution of some human health diseases. Here we report the infrastructure development to 
study how temperature and other meteorological parameters, air pollutant, such us ozone, 
nitrogen oxides, PM10, impact the evolution of diabetes. One of the main activities is putting 
together the last five years of meteorological and composition data and those of hospital 
admissions and clinical analyses of more that 13,000 patients. Another activity is to study the 
real life of diabetes patients monitoring continuously their physiological parameters, 
atmospheric parameters of the region where they live and indoor air quality of their houses. 
First results of the project, strengths and weakness will be discussed.

How to cite: Di Carlo, P., Aruffo, E., Mascitelli, A., and Chiacchiaretta, P.: Impact of the atmospheric composition and climate changes on the evolution of diabetes in Central Italy: the Vitality project, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-7076, https://doi.org/10.5194/egusphere-egu25-7076, 2025.

As part of the AdaptNet project, which aims to adapt and network general practitioner and specialist medical care to the health impacts of climate change, interactive maps are being produced for Germany that estimate current and future health risks. For heat, flooding, air quality, allergens, vectors and forest fires, it will be possible to obtain corresponding hazard levels at the level of districts and independent cities (corresponding to the NUTS3 regions in Germany). Estimating the health risks associated with climate change helps to avoid over- and under-adaptation of ambulant care to the consequences of climate change.

The methodology developed is based on the assessment of the most important factors for each hazard. The high spatial resolution requires a correspondingly high-resolution data base to be able to represent regional characteristics in the risk assessment. For the assessment of the current situation, data from recent years was used to include the already advanced climate change of the early 21st century. The future estimates refer to data around the year 2050.

The methodology was evaluated using two test regions (urban and rural). Very complex and data-intensive risk assessments were carried out for the two test regions and compared with a simpler approach, which was then applied to the whole of Germany.

When developing risk assessments relevant to emergency and disaster risk management in the health sector, WHO recommends that three factors be considered: hazard, exposure and vulnerability. We ensured that hazard and exposure were covered by factors in the risk assessment itself. Vulnerable groups were deliberately not included in the risk assessment, because they are individually targeted in an adaptation toolbox developed in the AdaptNet project.

How to cite: Kaspar-Ott, I., Álvarez, F., and Hertig, E.: Assessment of health risks due to climate change in Germany, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-8023, https://doi.org/10.5194/egusphere-egu25-8023, 2025.

Unlike natural dust (NDust), which primarily affects sparsely populated areas, mitigating health disparities from anthropogenic dust (ADust) fine particulate matter (PM_2.5) is crucial. ADust PM_2.5 has significant effects on public health and socio-economic conditions. With internal economic inequality widening within countries globally, urbanization, and aging populations exacerbating social vulnerability, assessing the health burden of ADust PM_2.5 pollution is crucial for achieving Sustainable Development Goal 3.9. This study integrates annual population and economic data with dust (include ADust and NDust) PM_2.5 concentrations to evaluate mortality due to this exposure and its relationship with income inequality. Our findings reveal a significant association between income inequality and mortality due to dust PM_2.5 exposure, considering variables such as the Gini index, GDP per capita, and exposed population structure. Greater income inequality and significant demographic change amplify the public health impacts of dust PM_2.5 pollution. Addressing wealth distribution inequalities is essential in pollution risk research and policy-making. Optimizing wealth distribution and enhancing control of ADust can effectively reduce health risks, fostering sustainable social development.

How to cite: Lian, L., Chen, S., and Huang, J.: Wealth inequality amplified the anthropogenic dust mortality, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-8028, https://doi.org/10.5194/egusphere-egu25-8028, 2025.

We used wavelet transform cross-correlation analysis to inform the model of the number of sand flies as a function of meteorological and environmental variables. To that end we used historical sand fly monitoring datasets from several past and ongoing collaborations in Europe, Turkey and Israel (projects EDENext, VectorNet, CLIMOS and PLANET4HEALTH), and correlated those with the corresponding temperature, precipitation, and soil moisture data.

We were looking into how the number of these disease vectors depends on all these variables and were interested to define the time lags between the changes of the meteorological and environmental drivers and change (particularly rise) in numbers of sand flies. We were additionally interested in how the change in climatic suitability for sand fly development will influence their spread in Europe. Finally, we researched if the modelled behavior can be universal across the sand fly species, or should be developed separately by species, and climatic regions.

Our results should assist development of the early warning systems for the spread of sand fly borne diseases that can be used by public health authorities for efficient and effective preparedness.

Funding: The CLIMOS consortium is co-funded by the European Commission grant 101057690 and UKRI grants 10038150 and 10039289. The six Horizon Europe projects, BlueAdapt, CATALYSE, CLIMOS, HIGH Horizons, IDAlert, and TRIGGER, form the Climate Change and Health Cluster. The PLANET4HEALTH consortium is co-funded by the European Commission grant 101136652. The five Horizon Europe projects, GO GREEN NEXT, MOSAIC, PLANET4HEALTH, SPRINGS, and TULIP, form the Planetary Health Cluster.

How to cite: Blesic, S. M., Tosic, M., Matic, V., Waitz, Y., Kirstein, O., Antoniou, M., and Maia, C.: Modeling climate drivers of the current and future spread of sand flies in Europe and neighboring countries with the use of wavelet transform analysis, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-8258, https://doi.org/10.5194/egusphere-egu25-8258, 2025.

The perception of green and blue spaces has been widely recognized for its positive impact on health and can be assessed through surveys that capture individuals’ experiences of their surrounding environment. While such surveys provide data that can be seen as ground truth, their implementation is often constrained by privacy concerns, time limitations, and inefficiencies. To address these challenges, quantitative datasets—such as the Normalized Difference Vegetation Index (NDVI) and land use data—can serve as inputs for spatial measurement methods, including buffer models, street view analyses, and viewshed analyses, to estimate green and blue space exposure. However, existing spatial measurement methods often fail to align with how people perceive green and blue spaces in their environment.

This study aims to address the question: How can green and blue space perception be modeled using spatial exposure measurement methods? To explore this, three spatial measurement approaches are applied: Euclidean buffer models, Streetview analyses, and viewshed analyses. These results are converted into Spearman correlation coefficients. Additionally, survey data collected in the Netherlands, where participants assessed green and blue spaces within their residential surroundings, are also analyzed using Spearman correlations. The correlations derived from spatial measurement methods are compared with those from the survey data to evaluate how well these methods capture perceived green and blue space exposure.

The findings aim to identify which spatial measurement methods best model individuals’ perceptions and offer insights into improving urban planning and policy. By enhancing the alignment between spatial models and human perception, this research contributes to more effective evaluations of green and blue space distribution in the Netherlands and highlights areas that may benefit from additional green and blue infrastructure.

How to cite: Vamos, C., Huss, A., Scheider, S., and Vermeulen, R.: Modeling the perception of green and blue space using spatial exposure measurement methods, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-9016, https://doi.org/10.5194/egusphere-egu25-9016, 2025.

Sultriness is formed by the interaction of several weather factors. It is the state of the atmosphere when the water vapor pressure exceeds 18.7 hPa. This condition has adverse physiological effects on plants, animals and especially on the human body. For this reason, in this research, emphasis was placed on the time evolution of sultriness at the meteorological station Hurbanovo in the Slovak Republic. The paper will examine the 40-year period (1981 – 2020). The study is a continuation of the work of Štefan Kveták, who examined the previous 30-year period (1951 – 1980). We hypothesized that the number of sultry days is also increasing due to climate change. The basis of the whole assumption was hourly data from meteorological stations in the database of the Slovak Hydrometeorological Institute. As the scientific goals of the project, we preferred the categorization of sultriness according to various criteria, the evaluation of their frequency and time trends of occurrence, and we compared their development with the previous period.

How to cite: Szabóová, K.: Number of Sultry Days in the Territory of Slovakia, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-9655, https://doi.org/10.5194/egusphere-egu25-9655, 2025.

Exposure to atmospheric compounds increases the risk of type 2 diabetes. In our study, we will show a derived exposure-response curve from the relative risk to develop type 2 diabetes because of exposure to different pollutants, i.e. particulate matter (PM2.5 and PM10), nitrogen dioxide (NO2), and ozone (O3). The curve is used to estimate a worldwide map of relative risk and the percentage of the attributable burden for each pollutant, using high resolution dataset of atmospheric pollutants from satellite observations. Finally, we will show the validation of the model comparing the modeled percentage of the numbers of patients that are affected by type 2 diabates also because of pollutants exposure with a regional analysis of the attributable patients affected by type 2 diabetes.

How to cite: Aruffo, E., Mascitelli, A., Chiacchiaretta, P., Carrieri, F., Baldassarre, M. P. A., Formoso, G., Consoli, A., and Di Carlo, P.: A worldwide study to estimate the relative risk to develop type 2 Diabetes Mellitus because of atmospheric pollutants exposure, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-12304, https://doi.org/10.5194/egusphere-egu25-12304, 2025.

Jordan is one of the most water-scarce regions in the world, facing climate change impacts on water, energy, and food—the core components of the WEF Nexus. Health, as an additional dimension of the nexus, is being investigated through the NIH-funded Global Center on Climate Change, Water, Energy, Food, and Health Systems (GC3WEFH). A key component of the Center is its Data Hub, which focuses on providing analytical access to datasets that reflect the WEFH nexus components and assembling an open-source software ecosystem to support integrative research while adhering to FAIR (Findable, Accessible, Interoperable, Reusable) principles.

This presentation demonstrates how large language models (LLMs) are transforming our ability to explore the complex interdependencies within the WEFH Nexus. By extracting insights from interdisciplinary sources—such as scientific articles, policy documents, and environmental and health datasets—LLMs provide powerful tools for integrated data analysis and decision-making across these critical domains. Built on the SuAVE (Survey Analysis via Visual Exploration, suave.sdsc.edu) platform, the Data Hub catalog enables intuitive browsing, querying, and faceted searches of Jordan-specific datasets. To enhance accessibility, LLM-based applications are integrated into the hub, allowing natural language queries to generate tables, maps, and visualizations, revealing interrelationships among nexus indicators such as the effects of climate change on water quality and health outcomes. Additional tools evaluate the AI-readiness of datasets and implement strategies to improve their usability for machine learning applications. These innovations enable deeper insights into the WEFH Nexus, supporting simulations of system sustainability and assessing the health impacts of water, food, and energy-focused strategies in environmentally stressed regions.

To further understand the global research landscape of the nexus, we constructed and analyzed a global co-authorship network of research articles referencing all four nexus components in their titles or abstracts. Using OpenAlex, an open-access bibliographic database, and the network analysis extension of the SuAVE platform, we visualized and examined the evolution of research collaborations, emerging topics, and knowledge gaps. Our analysis revealed that over 60% of WEFH-related publications have been produced in the last four years, reflecting a rapidly expanding but still fragmented field. The co-authorship network exhibits higher clustering and fragmentation compared to more established research areas, such as the Water-Energy-Food Nexus, which is characteristic of emerging disciplines. Key topics identified within the WEFH Nexus emphasize sanitation, water quality, and water treatment (water); wellness, safety, and public health systems (health); crop yields, food security, and nutrition (food); and renewable energy and emissions reduction (energy).

While the United States leads global contributions, accounting for nearly 30% of publications in the field, significant opportunities remain to foster stronger global collaborations and reduce fragmentation in the network. The GC3WEFH is leading this effort through a multi-institutional, international collaboration focused on modeling the climate impacts on vulnerable communities in water-scarce areas of Jordan.

This work is supported by the US National Institutes of Health, Fogarty International Center, under award # 1P20TW012709-01.

How to cite: Zaslavsky, I., Al-Delaimy, W., Mohtar, R., and Kirkpatrick, C.: Leveraging AI, Large Language Models, and Co-Authorship Network Visualization to Globally Understand the Water-Energy-Food-Health Nexus, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-12362, https://doi.org/10.5194/egusphere-egu25-12362, 2025.

Impact of Heat Exposure during Pregnancy in Ethiopian Cities

Desalew Meseret Moges^1*, Per-Ola Olsson², Ebba Malmqvist³, Masresha Tessema¹, Eleni Papadopoulou⁴, Kristoffer Mattisson³

¹ Nutrition, Environmental Health and Non-communicable Disease Research Directorate, Ethiopian Public Health Institute, Addis Ababa, Ethiopia.

² Department of Physical Geography and Ecosystem Science, Lund University, Sweden.

³ Division of Occupational and Environmental Medicine, Lund University, Sweden.

⁴ Global Health Cluster, Norwegian Institute of Public Health, Oslo, Norway.

Abstract
Climate change poses a significant public health threat, particularly for vulnerable groups such as pregnant women and children. Heat stress, when the body struggles to regulate its internal temperature due to high temperatures, presents increased health risks during pregnancy. Exposure to heat stress during pregnancy can result in adverse health outcomes for both the mother and fetus, including preterm birth, low birth weight, stillbirth, and pregnancy complications. However, research on the effects of heat exposure in epidemiological studies remains limited and inconsistent in low-resource countries like Ethiopia. This is mainly due to a lack of comprehensive data and resources. These regions often face limited infrastructure, scarce ground monitors, unreliable data collection systems, and insufficient technological support.

To address these gaps, this heat exposure study, which is part of the EU-funded ENABLE (Enabling Environments for Non-communicable Disease (NCD) risk reduction in Ethiopia) project, with the overarching aim to investigate the impact of urban heat exposure on maternal health outcomes in four Ethiopian cities: Addis Ababa, Jimma, Adama, and Harar. The present study's primary objective is to utilize remote sensing data to evaluate heat exposure.

Land Surface Temperature (LST), which measures the Earth's surface temperature, and the Discomfort Index (DI), which combines air temperature and humidity, will be used to assess heat stress. Data will be collected from satellite sensors (Landsat, MODIS; Moderate Resolution Imaging Spectroradiometer), climate data (ERA5; the fifth generation of the European Centre for Medium-Range Weather Forecasts atmospheric reanalysis of the global climate), and ground measurement from PurpleAir monitors. Heat stress will be assessed using hot days or heat waves when LST and DI exceed the 95th, 97th, or 99th percentiles for two consecutive days. This will involve creating high spatial resolution maps of heat exposure hotspots in Ethiopian cities.

The results from the present study will later be used in the ENABLE project to assess individual exposure to heat stress and effects on pregnancy outcomes. The planned epidemiological studies will include pregnant women recruited within the ENABLE project, with a target enrollment of 5000 participants, following their pregnancies from initiation till birth. Pregnancy outcomes collected from hospitals and public health records will be linked to heat metrics using GPS data from maternal residential addresses. This research provides critical insights into the intersection of climate change and urban heat stress in Ethiopia. The results can potentially inform Ethiopia’s climate-resilient urban planning and maternal health policies.

How to cite: Moges, D. M., Olsson, P.-O., Malmqvist, E., Tessema, M., Papadopoulou, E., and Mattisson, K.: Impact of Heat Exposure during Pregnancy in Ethiopian Cities, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-12913, https://doi.org/10.5194/egusphere-egu25-12913, 2025.

Poor ambient air quality poses a significant global health concern. However, accurate measurement remains challenging, particularly in countries like India, where ground monitors are scarce despite high expected exposure and health burdens. This lack of precise measurements impedes understanding of changes in pollution exposure over time and across populations, limiting effective public health responses. India faces severe air pollution issues, with fine particulate matter (PM_2.5) levels consistently exceeding the World Health Organization (WHO) guidelines, leading to various health problems, including respiratory and cardiovascular diseases, injuries, and deaths. Existing health impact research on PM_2.5 in India is limited, particularly for pediatric populations in diverse and socioeconomically varied regions.

In this study, we developed an open-source daily PM_2.5 dataset at a 10 km resolution for India from 2005 to 2023 using a two-stage machine learning model. This model integrates data from satellite sensors, meteorological variables, and land-use information, validated against held-out monitor data to generate accurate daily PM_2.5 estimates. We then linked this dataset with over one million pediatric ambulance dispatch records across 11 states in India from 2013 to 2015 to investigate the short-term effects of PM_2.5 exposure on pediatric emergency health outcomes. We employed a fixed-effects Poisson regression model combined with an instrumental variable (IV) approach to address potential endogeneity issues, such as reverse causality and omitted variable bias. The primary instrument used is thermal inversion, a meteorological phenomenon associated with elevated PM_2.5 levels. Our outcome measure is the number of ambulance dispatches per 100,000 people per day, categorized by cause (illness or injury) to reduce misclassification bias. Our fixed-effects model controls for time-invariant differences and temporal confounders, isolating effects of PM_2.5. Using thermal inversion as an instrument further confirms the robustness of the causal link between short-term exposure to PM_2.5 and increased ambulance dispatches.

Our analysis reveals significant associations between short-term PM_2.5 exposure and increased pediatric ambulance dispatches. For all-cause and illness-related calls, we observed more than a 2% increase in ambulance dispatches per 10 μg/m³ increase in PM_2.5 exposure, with cumulative lagged effects up to 7 days. Furthermore, for injury-related dispatches, there was more than a 5% increase associated with a 10 μg/m³ increase in PM_2.5 exposure, with cumulative effects observed within just 0 to 1 day of exposure. These findings emphasize the severe public health implications of PM_2.5 exposure on vulnerable populations, particularly children, underscoring the necessity for stringent air quality regulations and public health interventions across India.

How to cite: Kawano, A., Heft-Neal, S., Janagama, S., Strehlow, M., and Bendavid, E.: Short-term health impacts of PM2.5 exposure on pediatric ambulance dispatches in India using air quality data developed by machine learning, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-13356, https://doi.org/10.5194/egusphere-egu25-13356, 2025.

Background: Climate change has been associated with changes in pollen allergenicity, plant phenology, and overall pollen production levels highlighting its potential implications for public health. These changes can further lead to shifts in the duration, timing, and intensity of pollen seasons, affecting both allergenic and non-allergenic plant species. 

Objective: We analysed changes in grass and other pollen concentrations, pollen seasons and daily maximum temperatures over 32 years (1990 to 2023) in Melbourne.

Methods: Daily pollen counts were collected at Parkville, Melbourne every year for three months, October to December. Pollen was categorized as grass and other with other being trees and weeds. Seasonal trend decomposition was used to analyse long term trends in daily maximum temperatures and daily pollen concentrations. Linear regression was used to analyse changes in start, end and duration of core pollen season.

Results and discussion: According to preliminary results, the daily maximum temperature increased (Est slope = 0.0001/day, p <0.01) in Melbourne over the study years while the daily pollen concentrations depicted decreasing trend (p < 0.01). Core pollen season in Melbourne had an earlier start date (Est slope = -0.34 day/year, p < 0.01) and a longer duration (p < 0.01) over the decades 1990 to 2023. The results suggest climate change might be affecting the pollen seasons but the effect on pollen concentrations may have been masked by other environmental and climatic factors. These insights could have significant implications for vulnerable population, healthcare, research and urban planning.

How to cite: Dhankhar, A., Darssan, D., Dey, S., R Lampugnani, E., and J Osborne, N.: Analysing changes in temperature and pollen concentrations in Melbourne over 30 years, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-14502, https://doi.org/10.5194/egusphere-egu25-14502, 2025.

Mosquito-borne arboviruses pose a grave threat to millions of people worldwide each year, with climate change rapidly expanding hotspots of deadly Aedes-related diseases. Aware of potential compound effects regarding other important diseases, it has become imperative for health authorities to maintain a detailed surveillance of key variables that can trigger Aedes-borne epidemic episodes. Disease transmission is generally conditioned by multiple socio-economic factors, and among them, the environmental suitability for vectors and viruses to proliferate is a necessary –although not sufficient– condition that needs to be closely monitored. As such, a comprehensive service that allows stakeholders to detect and predict environmental suitability on affected hotspots is crucial for communities to better prepare in the case of present and future outbreaks.

To this end, AeDES2 is a next generation climate-and-health operational service that reproduces and improves computation of Aedes-borne Diseases Environmental Suitability over its previous version (Muñoz et al., 2020), expanding its temporal and spatial scope while simultaneously enhancing observational and forecasting quality of Aedes-related disease transmissibility. Users can consult the historical evolution of the environmental suitability values on any grid point of interest, as well as the expected future evolution up to three seasons in advance. Aside from environmental suitability values, health authorities can additionally utilize AeDES2 to analyse the estimated percentage of population at risk –crucial for governing bodies to implement control measures in order to reduce the spread of the disease.

AeDES2 incorporates four different environmental suitability models, translating temperature and precipitation values into environmental suitability outputs while considering epidemiological factors for transmission probability. Its monitoring system generates an up-to-date 12-member ensemble reference, providing a continuously updated historical sequence of environmental suitability values. On the forecasting side, AeDES2 builds on its predecessor’s pattern-based multi-model calibration, by assimilating state-of-the-art calibration methods such as causality-based calibration or multi-calibration techniques, aiming to reliably reproduce key non-linear patterns that are used as predictors in the cross-validated forecast system.

How to cite: Corvillo Guerra, J., Torralba, V., González Romero, C., Pérez-Zanón, N., Llabrés-Brustenga, A., Riviére-Cinnamond, A., and Garikoitz Muñoz, Á.: From climate variables to health information - Predicting and monitoring mosquito-borne disease outbreaks with AeDES2, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-15204, https://doi.org/10.5194/egusphere-egu25-15204, 2025.

How to cite: Bonfiglio, D., Bianco, S., Maragliano, M., Corcione, V., Rodi, G. C., Marangoni, S., Roberto, P., and Mosca, A.: Predicting Vector-Borne Disease Risk using Earth Observation and Machine Learning: A Case Study in northern Italy, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-16611, https://doi.org/10.5194/egusphere-egu25-16611, 2025.

In large parts of Africa, including the Lake Victoria Basin, malaria continues to be a major public health challenge, causing significant morbidity and mortality despite advancements in treatment and prevention. Climate change has the potential to exacerbate the problem because the disease vectors depend on non-permanent open water bodies for mosquito breeding and certain temperature thresholds for larval development. Climate-induced changes in hydrology, such as shifts in the timing and intensity of rainy seasons, combined with rising temperatures, may extend malaria risk to higher altitudes and new areas, necessitating preventive measures to counteract new transmission patterns.

In this study, we assess the impact of climate change on malaria transmission in the Lake Victoria Basin in Kenya, focusing on changes in mosquito breeding sites and temperature. To improve the representation of the breeding sites, the malaria transmission model VECTRI was coupled with the eco-hydrological model SWIM, which was enhanced with a pond module to capture non-permanent water bodies. We evaluated the model performance using Sentinel-1 satellite-derived water occurrence data and malaria incidence data obtained from health records in Kenya's Lake Victoria basin. To obtain high-resolution insights into the future of malaria transmission, projections were made using an ensemble of nine CMIP6 GCMs, downscaled to 1 km using the CHELSA downscaling method, covering several SSP scenarios.

The results of the coupled approach were promising in simulating water occurrence patterns and malaria incidence, demonstrating its potential as a valuable tool for predicting the effects of climate change on malaria transmission by capturing the interplay between climate, hydrology, and malaria dynamics. This research can guide the development of targeted public health interventions and adaptation strategies to mitigate the effects of climate change in malaria-endemic and at-risk regions.

How to cite: Engelhardt, H., Dieng, M. D. B., Schwarz, M., Volk, M., and Hattermann, F. F.: Assessing the Impact of Climate Change on Malaria Transmission in Kenya's Lake Victoria Basin, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-17438, https://doi.org/10.5194/egusphere-egu25-17438, 2025.

Tropical deforestation causes local climate warming and is a potential risk to human health. Previous studies have shown tropical deforestation causes increased heat stress and reduces safe outdoor working hours, but the excess mortality due to warming from deforestation has not been quantified. Here we use remote sensing Earth observations to make the first pan-tropical assessment of the population-weighted warming due to tropical deforestation and the associated heat-related mortality burden. We focus our analysis on tropical deforestation that has occurred during 2001 to 2020. We use spatially explicit satellite datasets of annual forest cover change and land surface temperature to identify areas of surface warming that are co-located with forest loss and use data on population distribution to map population-weighted exposure to this warming. We use data on non-accidental mortality combined with relationships between heat exposure and excess mortality from the literature, to estimate the heat-attributable excess mortality due to nearby tropical deforestation. We examine how population exposure to deforestation-induced warming varies by region and by the degree of tropical forest loss. Overall, our analysis shows tropical deforestation during 2001 to 2020 exposed over 350 million people to local climate warming with population-weighted daytime land surface warming of 0.27°C. We estimate this warming results in around 28,000 additional deaths per year, accounting for 39% of the total heat-related mortality burden caused by global climate change and deforestation combined. The impacted populations (those living near deforested areas) are predominantly from lower-income groups, often traditional and indigenous communities, with limited access to adaptive measures to protect against the impacts of climate warming. Our analysis provides important evidence of the negative human health impacts of tropical deforestation at local, regional and national scales.

How to cite: Reddington, C., Smith, C., Butt, E., Baker, J., Oliveira, B., Yamba, E., and Spracklen, D.: Impacts of tropical deforestation on local climate and human health, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-18273, https://doi.org/10.5194/egusphere-egu25-18273, 2025.

Building systems resilient to the societal and health impacts of future climate extremes requires actionable, context-based scenarios. Historically, public health and epidemiology have relied on retrospective analyses, which can be inadequate for preparing for unprecedented events.

To overcome this, we propose a methodology to develop context-based scenarios of health impacts (e.g. cardiovascular mortality) from future climate extremes also considering adaptation mechanisms (e.g. early warning system, health care improvements). This builds upon a methodology introduced by Shepherd et al. in 2018, which has been further developed for use on societal impacts including population health. The approach uses qualitative integration of various components to develop context-based scenarios. Here are some examples of these components:

• Historical and Future Climate Data: Using historical climate data and numerical projections to create geographically situated scenarios of extreme weather events.
• Analysis of Past Extremes: Considering health impacts from past extreme events of different magnitudes within the same geographic area.
• Cross-contextual Analysis: Considering health impacts from past extreme events in different settings and conceptually applying those scenarios, while considering contextual differences.
• Awareness: Considering the level of awareness within the population regarding climate extremes and their potential health impacts captured using semi-structure interviews, which can influence community preparedness and adaptation.

This approach is designed to leverage the insights from natural and critical social sciences while making room for methodological and epistemological differences. The integration of quantitative and qualitative data will occur through an iterative process, where both types of data complement each other in developing context-based scenarios. Quantitative data will provide the statistical foundation (e.g., projected cardiovascular mortality), while qualitative data will add depth by capturing social dynamics and adaptation strategies. The two will be synthesized in the final scenarios to ensure a comprehensive understanding of the impacts of climate extremes on different population groups.

How to cite: Raffetti, E., Messori, G., and Rusca, M.: Storyline approaches to characterize population health impacts of future climate extremes, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-19538, https://doi.org/10.5194/egusphere-egu25-19538, 2025.

How to cite: Di Sabatino, S. and the TRIGGER Consortium: Building a pathway to improve climate and health research: the case of the TRIGGER project , EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-19612, https://doi.org/10.5194/egusphere-egu25-19612, 2025.

Extremes in the Earth System are significant drivers of adverse population health outcomes. To fully understand potential future impacts on our health, Earth System Models (ESMs) and their output are increasingly integrated with and connected to Planetary Health applications.

To emphasize the role of ESM in understanding interactions between natural systems and their implications for human health, we conduct a systematic literature review focusing on the linkage between Earth System Modeling and Planetary Health applications.  By analyzing the use of ESM data in health applications, we identify variables across different Earth System spheres, evaluate their reliability, and highlight gaps in translating ESM outputs into health applications. Variables such as temperature, precipitation, and air quality are explored for their direct and indirect effects on health outcomes, including increased risks of infectious diseases, heat stress, and malnutrition.

The reviewed studies employ diverse Earth System Models (ESMs) and dynamic downscaling techniques to project future health scenarios, mainly relying on simple linkage rather than fully coupling Planetary Health applications. Key findings reveal substantial increases in mortality and morbidity rates linked to cardiovascular and respiratory diseases, exacerbated by prolonged exposure to extreme heat and degraded air quality. For instance, regional analyses indicate significant health risks in densely populated urban areas and low-income regions, emphasizing the need for tailored mitigation strategies. Notably, applications such as simulating the impacts of heatwaves on mortality in Europe and assessing adaptation measures like green space-based cooling systems exemplify the need for integration of ESM and Planetary Health.

Our synthesis highlights the critical interplay of socioeconomic, demographic, and Earth System factors in shaping health vulnerabilities, underscoring the importance of intersectionality in climate health research. Advancing the integration of ESM and Planetary Health is crucial for promoting climate resilience and equity in health outcomes.

How to cite: Thiele-Eich, I., Rahmen, M., and Falkenberg, T.: Linking Earth System Modeling and Planetary Health: A Systematic Literature Analysis of Interactions and Impacts on Human Health, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-19710, https://doi.org/10.5194/egusphere-egu25-19710, 2025.

For understanding localized hydrological and climatological processes, downscaling gridded precipitation data to finer spatial resolutions is a crucial prerequisite. For a densely populated country like India, accurate downscaled data is crucial for building resilience to climate change impacts, supporting adaptation efforts, and enhancing disaster management. In recent years, deep learning (DL) has emerged as a powerful tool for advancing Earth system modelling and climate data downscaling. This study presents a comprehensive intercomparison of deep learning architectures, for downscaling precipitation across India. A few efficient DL architectures from recent studies are chosen for intercomparison such as simple dense, simple convolutional neural network, Fast Super Resolution Convolutional Neural Network (FSRCNN), Super Resolution Deep Residual Network (SRDRN), U-Net, and Nest-U-Net. The experiments are designed in synthetic style by using coarsened ECMWF Reanalysis version 5 (ERA5; 1^ox1^o) daily variables as the inputs and high-resolution Indian Monsoon Data Assimilation and Analysis reanalysis (IMDAA; 0.12^ox0.12^o) daily precipitation as training labels and benchmarks for the evaluation. Training and validation are conducted for the period 1980-2014, afterwards the trained models are evaluated on data from 2015-2020. To reduce the biases induced by the highly positive-skewed precipitation data and to enhance the model performance on extreme events, a weighted mean absolute error is implemented for training. The performance of the DL models is also compared with the Bias Correction and Spatial Disaggregation (BCSD), a renowned statistical downscaling method. The results indicate that all deep learning DL models outperformed the BCSD method. Among the DL models, U-Net and Nest-U-Net demonstrated superior performance in capturing fine-scale precipitation patterns and extreme precipitation events, owing to their encoder-decoder architecture, which effectively learns spatial features at different scales. In contrast, the FSRCNN and SRDRN produced results with slightly lower precision than the U-Net models, but at a significantly reduced inference time, making them more efficient for faster data generation. The findings underscore the potential of deep learning for improving regional precipitation downscaling across India, offering a promising alternative to traditional statistical methods like BCSD in handling complex, non-linear relationships inherent in climate data.

How to cite: Murukesh, M. and Kumar, P.: Comparative analysis of deep learning architectures trained for downscaling gridded precipitation across India, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-676, https://doi.org/10.5194/egusphere-egu25-676, 2025.

Renewable resources are strongly dependent on local and large-scale weather situations. Skillful subseasonal to seasonal (S2S) forecasts -beyond two weeks and up to two months- can offer significant socioeconomic advantages to the energy sector. In particular, accurate wind speed forecasts result in optimized generation of wind-based electric power. This study aims to enhance wind speed predictions using a diffusion model with classifier-free guidance to downscale S2S forecasts of surface wind speed. We propose DiffScale, a diffusion model that super-resolves spatial information for continuous downscaling factors and lead times. Leveraging weather priors as guidance for the generative process of diffusion models, we adopt the perspective of conditional probabilities on sampling super-resolved S2S forecasts. We aim to directly estimate the density, associated with the target S2S forecasts at different spatial resolutions and lead times without auto-regression or sequence prediction, resulting in an efficient and flexible model. Synthetic experiments were designed to super-resolve wind speed S2S forecasts from the European Center for Medium-Range Weather Forecast (ECMWF) from a coarse resolution to a finer resolution of data from ERA5, which serves as a high-resolution target, derived from reanalysis data. We achieve a significant increase in the quality of predictions, utilizing the proposed diffusion model for continuous downscaling and bias correction of the ECMWF forecasts.

How to cite: Springenberg, M., Otero Felipe, N., and Ma, J.: DiffScale: Towards Continuous Downscaling and Bias Correction in Subseasonal Wind Speed Forecasts, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-1650, https://doi.org/10.5194/egusphere-egu25-1650, 2025.

Understanding the complex relationship between trace gases as well as undestanding various source and sink pathways in the atmsophere need good qualtity continuous and reliable datasets. However, obtaining comprehensive long-term profiles for key trace gases is a significant challenge. We have initiated a new research strand to consrtuct  long term data using machine learning. Output from a  Chemical Transport Model (CTM) and observational data from satellite instruments (such as HALOE and ACE-FTS) is merged using machine learning. This integration results in the creation of daily, gap-free datasets for six crucial gases: ozone (O3), methane (CH4), hydrogen fluoride (HF), water vapour (H2O), hydrogen chloride (HCl), and nitrous oxide (N2O) from 1991 to 2021.

Chlorofluorocarbons (CFCs) are a critical source of chlorine that controls stratospheric ozone losses. Currently, ACE-FTS is the only instrument that provides sparse but daily measurements of these gases. Monitoring changes in these ozone-depleting substances, which are now banned, helps assess the effectiveness of the Montreal Protocol. We have initiated the construction of gap-free stratospheric profile data for CFC-11 as a subsequent step.

We use a regression model to estimate the relationship between various tracers in a CTM and the differences between the CTM output field and the observations, assuming all errors are due to the CTM setup. Once the regression model is trained for observational collocations, it is used to estimate biases for all the CTM grid points. To enhance accuracy, we employed various regression models and found that XGBoost regression outperforms other methods. ACE-FTS v5.2 data (2004-present) is used to train (70%) and test (30%) the XGBoost performance.

Our results demonstrate excellent agreement between the constructed profiles and satellite measurement-based datasets. Biases in TCOM data sets, when compared to evaluation profiles, are consistently below 10% for mid-high latitudes and 50% for the low latitudes, across the stratosphere. The constructed daily zonal mean profile datasets, spanning altitudes from 15 to 60 km (or pressure levels from 300 to 0.1 hPa), are publicly accessible through Zenodo repositories.

     CH4:       https://doi.org/10.5281/zenodo.7293740   
     N2O:          https://doi.org/10.5281/zenodo.7386001
     HCl :         https://doi.org/10.5281/zenodo.7608194
     HF:        https://doi.org/10.5281/zenodo.7607564
     O3:         https://doi.org/10.5281/zenodo.7833154 
     H2O:          https://doi.org/10.5281/zenodo.7912904
     CFC-11:    https://doi.org/10.5281/zenodo.11526073  
     CFC-12:      https://doi.org/10.5281/zenodo.12548528
     COF2:        https://doi.org/10.5281/zenodo.12551268

In an upcoming iteration, we are enhancing the algorithm as well as add more species in the current setup. We believe these data sets would provide valuable insights into the dynamics of stratospheric trace gases, furthering our understanding of their behaviour and impact on the climate.

References:

Dhomse, S. S., et al.,: ML-TOMCAT: machine-learning-based satellite-corrected global stratospheric ozone profile data set from a chemical transport model, Earth Syst. Sci. Data, 13, 5711–5729, https://doi.org/10.5194/essd-13-5711-2021, 2021.

Dhomse, S. S. and Chipperfield, M. P.: Using machine learning to construct TOMCAT model and occultation measurement-based stratospheri
c methane (TCOM-CH4) and nitrous oxide (TCOM-N2O) profile data sets, Earth Syst. Sci. Data, 15, 5105–5120, https://doi.org/10.5194/essd-15-5105-2023, 2023.

How to cite: Dhomse, S. and Chipperfield, M.: Machine Learning to Construct Daily, Gap-Free, Long-Term Stratospheric Trace Gases Data Sets, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-2394, https://doi.org/10.5194/egusphere-egu25-2394, 2025.

Analogue has been a widely-used concept in atmospheric science, particularly useful in weather forecasting and climate-related studies. The underlying idea is straightforward. An analogue is identified by determining its level of similarity to a reference weather or climate condition, traditionally, via computing a Euclidean distance. Recently, a deep-learning based framework, called ClimaDist, was proposed for climate analogue identification, found to outperform traditional Euclidean distance metrics. Despite the promising performance, similarly to many deep-learning models, it is challenging to estimate the uncertainty of the analogue searching process undertaken by ClimaDist. This hinders its applicability to real-world operations, especially for those requiring decision making.

To address this challenge, this study extends the capabilities of ClimaDist through incorporating a uncertainty quantification method, together with explainable AI (XAI) techniques. Specifically, the Evidential Deep Learning (EDL) approach is applied to the analogue searching process undertaken by the ClimaDist. This enables effective quantification of the uncertainty associated with data and model, respectively, while exploring their relationship with overall model performance. Two distinct scenarios are applied to these two models using data that were seen and unseen during the training processing. 

An experiment has been designed to verify the proposed approach using ERA5 data over a square domain centred at the Nettebach (Germany) covering the geographic range of 55°N to 47°N and 3°E to 11°E. Two ClimaDist models, one with the best validation performance and the other one best training performance, respectively, are used for comparison. These models are assessed based on the similarity of the found analogues and via under two distinct scenarios –with input data seen and unseen during the training process, respectively. Preliminary results suggest that the integration of uncertainty quantification enhances the interpretability and reliability of analogue identification, enabling improved downstream applications. Specifically, high model uncertainty can be highlighted by the proposed approach while fully unseen data is used as input. This not only provides valuable insight in knowing the capacity of the underlying model but also allows the optimization of resource usage.

How to cite: Chen, C., Chen, P.-C., Tseng, C.-Y., and Wang, L.-P.: Uncertainty quantification through the climate analogue identification process by ClimaDist, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-3610, https://doi.org/10.5194/egusphere-egu25-3610, 2025.

Databases of high-resolution interpolated climate data are essential for analyzing the impacts of past climate events and for developing climate change adaptation strategies for managed and natural ecosystems.  To enable such efforts, we contribute an accessible, comprehensive database of interpolated climate data for Africa that includes monthly, annual, decadal, and 30-year normal climate data for the last 120 years (1901 to present) as well as multi-model CMIP6 climate change projections for the 21st century. The database includes variables relevant for ecological research and infrastructure planning, and comprises more than 25,000 climate grids that can be queried with a provided ClimateAF software package. In addition, 30 arcsecond (~1km) resolution gridded data, generated by the software, are available for download (https://tinyurl.com/ClimateAF). The climate grids were developed with a three-step approach, using thin-plate spline interpolations of weather station data as a first approximation, subsequent fine-tuning with deep neural networks to capture medium-scale local weather patterns, and lastly dynamic lapse-rate based downscaling to a user-selected resolution, or to scale-free point estimates with the ClimateAF software package. The study contributes a novel deep learning approach to model orographic precipitation, rain shadows, lake and coastal effects, including the influences of wind direction and strength. The climate estimates were optimized and cross-validated with a checkerboard approach to ensure that training data was spatially distanced from validation data. We conclude with a discussion of applications and limitations of this database.

How to cite: Namiiro, S., Hamann, A., Wang, T., Castellanos-Acuña, D., and Mahoney, C.: Climate data interpolation with deep neural networks: a comprehensive dataset of historical and future climate for Africa, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-4595, https://doi.org/10.5194/egusphere-egu25-4595, 2025.

The use of machine learning (ML) for climate science has attracted considerable attention within the last few years. A number of recent studies have used ML to extract information from global climate data (e.g. regional downscaling), predict future states of the climate system and evaluate models against observations. In particular, Brunner and Sippel (2023) showed that low-resolution global climate models and observations can reliably be distinguished based on the global distribution of daily temperature, even after removing the mean model bias. ML is thus able to isolate fundamental differences between models and observations even in the presence of substantial internal variability. This raises the questions of whether ML can also distinguish between model and observational data on a regional scale, whether ML is as successful for km-scale models as for coarse-resolution models, and whether more complex bias correction methods reduce the success of ML.

To answer these questions, we use daily temperature fields over Austria, a topographically very complex domain. As training data, we use 200 different, randomly drawn days from each of the 13 ÖKS15 bias-corrected EURO-CORDEX models with an output resolution of 1km, resulting in 2600 samples labeled “model” which are matched by the same number of random days labeled “observation” from the SPARTACUS observation dataset. We use the binary classification approach to distinguish between the two classes of models versus observations. A logistic regression classifier is trained to determine the probability that a daily temperature field belongs to one of the two classes. In order to evaluate the ML algorithm subsequently, all days from the out-of-sample 10-year period 2005-2014 are used as test data.

The ML algorithm succeeds in correctly identifying the overwhelming majority of the test data for the setup used, resulting in an accuracy of 99%. The  results remain consistent even when a different sample of 2x2600 random training days is used. In contrast to more complex classifiers, such as a convolutional neural network (CNN), the learned coefficients from the logistic regression allow insights into the spatial patterns that are crucial for distinguishing between models and observations. While the performance of climate models is typically evaluated on climatological timescales, our results highlight that such classifiers can be used to identify patterns of structural model biases. Our method hence offers a computationally efficient approach for model evaluation, especially when handling km-scale climate model data on a regional domain.

References:
Brunner L. and Sippel S. (2023): Identifying climate models based on their daily output using machine learning, Environmental Data Science, https://doi.org/10.1017/eds.2023.23

How to cite: Meindl, M., Voigt, A., and Brunner, L.: Using machine learning to distinguish km-scale climate models and observations on a regional scale, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-6462, https://doi.org/10.5194/egusphere-egu25-6462, 2025.

North Atlantic sea surface temperatures (NASST), particularly in the subpolar region, exhibit some of the highest predictability across global oceanic systems. However, the relative contributions of atmospheric versus oceanic influences on the long term NASST variability remains ambiguous. In this study, we utilize neural networks (NNs) to assess the significance of various atmospheric and oceanic predictors in forecasting the state of NASST within the CANARI Large Ensemble, which employs the Met Office CMIP6 physical climate model (HadGEM3-GC3.1) at a high-resolution atmospheric scale (N216, approximately 60 km at midlatitudes) and a 1/4° resolution for oceanic data. The ensemble comprises forty members, driven by CMIP6 historical data and SSP3-7.0 scenarios for the period from 1950 to 2099. First, we evaluate the ability of the NNs to anticipate the phases of long term (multidecadal variability) using observational datasets, thereby investigating the consistency of physical processes influencing NASST variability between modeled predictions and real-world observations. Second, the research delves into how the interplay between oceanic and atmospheric predictors, alongside external forcings and internal variability (atmospheric noise), impacts the machine learning-based predictions and we use explainable AI techniques to identify the sources of predictability and to pinpoint physical mechanisms and regions crucial for accurate NN forecasts.

How to cite: Colfescu, I.: Explainable neural nets for disentangling sources of predictability in the North Atlantic Sea Surface Temperature (NASST), EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-6680, https://doi.org/10.5194/egusphere-egu25-6680, 2025.

Complex multiscale dynamics of the atmosphere in extratropical latitudes includes various persistent atmospheric regimes with the residence time up to several weeks. Identification, simulation and prediction of such dynamics remains one of the challenging problems. In this work we use a stochastic recurrent neural network (RNN) with specific architecture to address this problem, appealing to RNN’s ability to handle memory effects well. The proposed RNN connects two types of variables: (i) a low-dimensional representation of the physical variables via Principal Component Analysis (PCA), and (ii) Kernel PCA variables which serve to better represent the target atmospheric regimes [1]. The stochastic component of the RNN has a simple form which allows us to analytically write Bayesian log-posterior and log-likelihood functions to train and cross-validate the model given the particular dataset.
Using the observed and climate-model-generated winter geopotential height data in the Northern Hemisphere, we show that the proposed stochastic model is able to reproduce/predict various dynamical properties and distributions of the target regimes in the kernel space, as well as to reconstruct kernel variables from a low-dimensional representation of the original spatio-temporal field.

References
1. Mukhin et al. (2022). Revealing recurrent regimes of mid-latitude atmospheric variability using novel machine learning method. Chaos: An Interdisciplinary Journal of Nonlinear Science, 32(11). 

How to cite: Gavrilov, A., Mukhin, D., Safonov, S., and Samoilov, R.: Stochastic recurrent neural network for modeling atmospheric regimes, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-7404, https://doi.org/10.5194/egusphere-egu25-7404, 2025.

In hydrology, the use of machine learning (ML) has gained traction due to its ability to provide alternative or complementary approaches to traditional process-based modelling. These models identify numerical patterns in time series data without needing to solve conservation equations. This flexibility enables hydrological calculations in areas where data sources are incomplete or non-existent.
Studies benchmarking ML models (SVM, RNN, CNN) against process-based models have shown that ML models deliver promising results with lower computational cost and less information about the physical processes they are modelling. Consequently, they can effectively utilize spatially discretized physical data on a large scale.
This study designs a Long Short-Term Memory (LSTM) neural network to learn sequential relationships between atmospheric, climatic and geographic features and daily streamflow data from 39 headwater gauging stations in the northern Ebro river basin. LSTM models include an internal state that can store information and learn long-term dependencies, enabling them to model sequential data effectively. However, the numerical patterns identified by LSTM models do not inherently respect universal physical laws, such as the conservation of mass.
To address the limitation, the Mass-Conserving LSTM (MC-LSTM) model has been employed and compared with the standard LSTM model. The MC-LSTM model introduces a modified cell structure that adheres to conservation laws by extending the learning bias to model the redistribution of mass.
This analysis highlights not only the high accuracy of LSTM models in predictive hydrologic modelling but also the critical importance of integrating physics-based features to enable ML models to effectively capture the hydrological dynamics of the basin.
Acknowledgments: This work is funded by the European Research Council (ERC) through the Horizon Europe 2021 Starting Grant program under REA grant agreement number 101039181-SED@HEAD.

How to cite: González Planet, I. and Juez, C.: Streamflow forecasting in the Ebro river basin using Machine Learning (ML) and a physical mass constraint, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-8105, https://doi.org/10.5194/egusphere-egu25-8105, 2025.

Accurate climate projections are critical for understanding climate change and to design adaptation and mitigation strategies. Weighting schemes that aggregate a range of climate model projections are widely used to provide more reliable estimates of future climate conditions. Recently, causal discovery has been successfully introduced in the weighting schemes to constrain uncertainties in climate model projections based on the performance and interdependence of climate models. However, the previous methodologies typically (and strongly) only utilize a single metric, the F1 score of performance and similarity between each climate model and observational data,  to compare the different models' causal structures. Here, we introduce alternative and more sophisticated causal weighting schemes inspired by the theory of kernel methods and Gaussian processes to compare causal graphs directly in suitable reproducing kernel Hilbert spaces. In addition, we propose alternative causal weighting schemes that rely on interventions, graph-based distances, and counterfactual evaluations. We will evaluate the causal weighting strategies in various synthetic and CMIP6 model datasets. 

How to cite: Camps-Valls, G., Debeire, K., Varando, G., Runge, J., and Eyring, V.: Causal Weighting for Climate Projections, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-8363, https://doi.org/10.5194/egusphere-egu25-8363, 2025.

Atmospheric ozone is a crucial absorber of solar radiation and an important greenhouse gas. However, explicitly representing ozone in climate models is computationally expensive. A recent study introduced a simple linear machine learning-based ozone parameterization scheme (mloz) for daily ozone prediction based on temperature. Here we develop and implement the mloz in the UK Earth System Model (UKESM) for long-term idealized climate simulations. It produces stable ozone predictions over 50 years with a computational cost of less than 0.5% of the total runtime. The scheme accurately predicts ozone distribution, with climatology field errors of less than 10% in the stratosphere. It also realistically represents ozone variabilities, including seasonal and Quasi-Biennial Oscillation-related variabilities, despite a slight underestimation of amplitudes over the stratospheric polar regions. Additionally, we further demonstrated its generalizability by successfully transferring the mloz trained on UKESM to the ICOsahedral Nonhydrostatic model (ICON). Over 30 years of climate sensitivity tests indicate that it can effectively represent the response of ozone to the sudden quadrupling of CO₂, significantly outperforming the simplified linearized ozone photochemistry scheme (Linoz) in the troposphere. This implies that the mloz can be transferred to other climate models without a full chemistry module to enable an efficient explicit ozone simulation.

How to cite: Ma, Y., Abraham, L., Versick, S., Ruhnke, R., Braesicke, P., and Nowack, P.: A Highly Efficient Machine Learning-based Ozone Parameterization for Climate Models, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-9358, https://doi.org/10.5194/egusphere-egu25-9358, 2025.

Assessing precipitation impacts due to anthropogenic climate change relies on accurate and high-resolution numerical Earth system model (ESM) simulations. However, such simulations are computationally too expensive, and their discretized formulation can introduce systematic errors. These can, for example, lead to an underestimation of spatial intermittency and extreme events.
Generative machine learning has been shown to skillfully downscale and correct precipitation fields from numerical simulations [1].
However, these approaches require separate training for each Earth system model, making corrections of large ESM ensembles computationally costly.
Here, we follow a diffusion-based approach [2] by training an unconditional generative consistency model [3] on high-resolution ERA5 precipitation data. Once trained, a single generative model can be used to efficiently downscale arbitrary ESM simulations in an uncertainty-aware and scale-adaptive manner. Using three different climate models, GFDL-ESM4 [4], POEM [5], and SpeedyWeather [6], we evaluate the performance and generalizability of our approach.

[1] Harris, L., McRae, A.T., Chantry, M., Dueben, P.D. and Palmer, T.N., 2022. A generative deep learning approach to stochastic downscaling of precipitation forecasts. Journal of Advances in Modeling Earth Systems, 14(10), e2022MS003120.
[2] Hess, P., Aich, M., Pan, B., and Boers, N., 2024. Fast, Scale-Adaptive, and Uncertainty-Aware Downscaling of Earth System Model Fields with Generative Machine Learning. arXiv preprint arXiv:2403.02774.
[3] Song, Y., Dhariwal, P., Chen, M., and Sutskever, I. 2023.  Consistency Models. In International Conference on Machine Learning (pp. 32211-32252).
[4] Dunne, J.P., Horowitz, L.W., Adcroft, A.J., Ginoux, P., Held, I.M., John, J.G., Krasting, J.P., Malyshev, S., Naik, V., Paulot, F. and Shevliakova, E., 2020. The GFDL Earth System Model version 4.1 (GFDL‐ESM 4.1): Overall coupled model description and simulation characteristics. Journal of Advances in Modeling Earth Systems, 12(11), e2019MS002015.
[5] Drüke, M., von Bloh, W., Petri, S., Sakschewski, B., Schaphoff, S., Forkel, M., Huiskamp, W., Feulner, G. and Thonicke, K., 2021. CM2Mc-LPJmL v1.0: biophysical coupling of a process-based dynamic vegetation model with managed land to a general circulation model. Geoscientific Model Development 14, 4117–4141.
[6] Klöwer, M., Gelbrecht, M., Hotta, D., Willmert, J., Silvestri, S., Wagner, G.L., White, A., Hatfield, S., Kimpson, T., Constantinou, N.C. and Hill, C., 2024. SpeedyWeather.jl: Reinventing atmospheric general circulation models towards interactivity and extensibility. Journal of Open Source Software, 9(98), 6323.

How to cite: Hess, P., Aich, M., Pan, B., and Boers, N.: Downscaling precipitation simulations from Earth system models with generative machine learning, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-9742, https://doi.org/10.5194/egusphere-egu25-9742, 2025.

Subseasonal-to-seasonal (S2S) forecasts are crucial for decision-making and early warning systems in extreme weather. However, the chaotic nature of atmospheric dynamics limits the predictive skill of climate models on S2S timescales. Teleconnections can provide windows of improved predictability, but leveraging these external drivers to enhance S2S forecast skill remains challenging. This study introduces a spatio-temporal neural network (STNN) designed to predict weekly North Atlantic European (NAE) weather regimes at lead times of one to six weeks during boreal winter. The STNN integrates a stacked vision transformer (ViT) encoder and a long short-term memory (LSTM) decoder to capture short- and medium-range variability. By incorporating spatio-temporal data on the stratospheric polar vortex, tropical outgoing longwave radiation, and 1D NAE regime time series, the network can access patterns linked to teleconnections of key drivers of European winter weather. Its modular design enables the application of mechanistic interpretability, providing novel neuron-level insights into the prediction behavior. The improved predictive skill beyond lead week three and enhanced accuracy for specific regimes suggest novel learned patterns of external drivers. Using Activation Maximization (AM), we analyze these learned representations, and by incorporating gradient-based explanations of correct predictions, we infer additional insights into prevalent teleconnections. 

How to cite: Bommer, P. L., Kretschmer, M., Spurler, F., Bykov, K., Boehnke, P., and Hoehne, M. M.-C.: Combining spatio-temporal neural networks with mechanistic interpretability to investigate teleconnections in S2S forecasts, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-9753, https://doi.org/10.5194/egusphere-egu25-9753, 2025.

Local climate information is crucial for impact assessment and decision-making, yet coarse global climate simulations cannot capture small-scale phenomena. Current statistical downscaling methods infer these phenomena as temporally decoupled spatial patches. However, to preserve physical properties, estimating spatio-temporally coherent high-resolution weather dynamics for multiple variables across long time horizons is crucial. We present a novel generative approach that uses a score-based diffusion model trained on high-resolution reanalysis data to capture the statistical properties of local weather dynamics. After training, we condition on coarse climate model data to generate weather patterns consistent with the aggregate information. As this inference task is inherently uncertain, we leverage the probabilistic nature of diffusion models and sample multiple trajectories. We evaluate our approach with high-resolution reanalysis information before applying it to the climate model downscaling task. We then demonstrate that the model generates spatially and temporally coherent weather dynamics that align with global climate output.

How to cite: Schmidt, J., Schmidt, L., Strnad, F., Ludwig, N., and Hennig, P.: Spatiotemporally Coherent Probabilistic Generation of Weather from Climate, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-11083, https://doi.org/10.5194/egusphere-egu25-11083, 2025.

Recent years have seen rapid advancements in large-scale data-driven models for weather forecasting. Several of these models can now compete with, and in some respects outperform, physics-based numerical models for medium-range forecasting. They offer significant computational savings and potential forecasting accuracy improvements approximately equivalent to a decade of progress in traditional methods. This progress has prompted announcements from weather institutes across the world about plans to integrate AI-driven models into their operational workflows in the near future.

As data-driven models become integral to operational forecasting, critical questions about fairness and equity remain. Studies reveal substantial variations in forecast quality across regions, particularly for extreme weather. Unlike physical models, the disparities in data-driven models often stem from passive design decisions, such as inductive biases and weighting schemes, which may be reassessed and changed, if needed. Moreover, ensuring equitable access to these models, along with the means to effectively utilise and improve them, is essential so that both high- and low-income countries can share in their benefits.

This work explores fairness in data-driven weather forecasting, with a focus on outcome-based perspectives. We begin by defining fairness from both process and outcome viewpoints. We then analyse the performance of current data-driven models across different regions and socio-economic groups globally. Our findings reveal significant disparities that may exacerbate pre-existing socio-economic and climate-related vulnerabilities. To address these challenges, we advocate for a deliberate focus on fairness and equity in data-driven model development, emphasising the importance of active design choices to promote equitable outcomes.

How to cite: Olivetti, L. and Messori, G.: Whose weather is it? A fairness perspective on data-driven weather forecasting, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-11668, https://doi.org/10.5194/egusphere-egu25-11668, 2025.

Machine learning (ML) parameterisations for climate models are emerging as a promising approach for capturing subgrid-scale processes, which are not explicitly resolved in climate models due to limitations on resolution. These ML parameterisations are typically trained on datasets generated by high resolution climate models or existing parameterisations (“offline”), but evaluated based on their performance when coupled into an existing climate model (“online”). Quantifying uncertainties associated with ML parameterisations is crucial for gaining insights into the reliability of hybrid ML-climate models.

I will discuss uncertainties associated with an ML parameterisation for atmospheric GWs, focusing on the parametric uncertainties which originate during the training process. I will show how these can propagate when coupled online, becoming a significant source of uncertainty in climate model circulation that we must consider carefully when building ML parameterisations.  

How to cite: Mansfield, L. and Sheshadri, A.: Uncertainty Quantification of Machine Learning Parameterisations, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-12129, https://doi.org/10.5194/egusphere-egu25-12129, 2025.

High-resolution climate projections are vital for understanding the local impacts of climate change in fields like agriculture, hydrology, energy production, and disaster risk management. However, Earth System Model (ESM) output often lacks the spatial detail needed to capture regional to local-scale variability, while showing large biases when compared to observational data. Statistical downscaling (SD) is commonly used to address such issues by refining the coarse spatial resolution of ESM output. While the current ISIMIP3 (Inter-Sectoral Impact Model Intercomparison Project, third round) SD algorithm is robust and computationally efficient, it struggles with increasing differences between source and target resolutions. To address these limitations, we applied deep-learning-based SD methods to create a globally consistent, high-resolution dataset for near-surface climate variables.

Using the perfect prognosis approach, we combined ERA5 as large-scale atmospheric predictors with ERA5-Land as high-resolution predictands (target resolution of ~10 km) to create accurate transfer functions (TFs) that align with ISIMIP's requirements, such as trend preservation and inter-variable consistency. These TFs are subsequently applied to ESM output to generate downscaled climate forcings. The resulting framework is both scalable and computationally efficient, making it suitable for multi-model applications. The results were compared with similar methodologies and its improvements were demonstrated in a cross-validation framework, particularly in capturing local-scale features.

Our approach offers a robust tool for generating high-resolution climate data, providing valuable insights to researchers and decision-makers working on climate impact assessments and adaptation planning. This work contributes to the next iteration of ISIMIP and to OptimESM, targeting the CMIP7-based modeling framework. The derived high-resolution projections are designed to complement CMIP7 datasets, enabling the creation of downscaled ensembles that conform with ISIMIP's objectives and support a wide range of impact modeling applications.

How to cite: Quesada-Chacón, D., Sauer, I., Mengel, M., and Frieler, K.: A deep learning approach to statistical downscaling and its potential to increase the resolution of the impact model simulations within ISIMIP, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-12175, https://doi.org/10.5194/egusphere-egu25-12175, 2025.

Estimation of the Reference Evapotranspiration (ET₀) is critical in water resources management under climate change, especially for agricultural water management in arid and semi-arid regions. Thus, estimating baseline ET₀ poses significant challenges, particularly in inadequate climatological monitoring regions. In this study, a hybrid modeling approach based on the incorporation of empirical models, Particle Swarm Optimization (PSO), and XGBoost algorithm (Empirical-PSO-XGBoost) was developed and evaluated to forecast ET₀ under limited climate variables. The results showed the Empirical-PSO-XGBoost outperformed the purely calibrated empirical and Temperature-PSO-XGBoost models for estimating monthly (daily) ET₀ with NSE reaching 0.99 (0.86) and 0.98 (0.67) for the calibration and validation phases, respectively. Besides, up to 63 CMIP6 projections were coupled with Empirical-PSO-XGBoost for forecasting the long-term ET₀ under SSP245 and SSP585 climate change scenarios. Thus, the simulation showed a significant increase in ET₀ and seasonal patterns compared to the baseline ET₀ where the change in range of [+5, +10] % is associated with probability values of 0.65 and 0.78 for SSP245 and SSP585, respectively. Overall, the developed framework is useful for implementing adaptation strategies to mitigate climate change effects on water resource allocation and agricultural management. It provides the ET₀ associated with Exceedance probability for each month which is useful for assessing the water availability-related-risk in scheduling irrigation and sowing date of crops.

How to cite: Elbilali, A., Hadri, A., Taleb, A., EL Khalki, E. M., Tanarhte, M., and Kharrou, M. H.: A Novel Modeling Framework based on Empirical models, PSO, XGBoost, and multiple GCMs for the projection of Long-Term Reference Evapotranspiration, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-12646, https://doi.org/10.5194/egusphere-egu25-12646, 2025.

Climate variability and change significantly influence crop production, presenting challenges that extend from understanding the basic crop growth principles to evaluating the effects of extreme weather events on crop development. Addressing this requires effective agro-management strategies guided by tailored climate services. However, a critical gap exists between scientific insights and their practical application. This study introduces and evaluates an AI-driven methodology designed to simulate crop growth and predict grain maize yields across Europe. Specifically, nested Recurrent Neural Networks (RNNs) are tested as a computationally efficient surrogate model for the process-based ECroPS model developed by the European Commission’s Joint Research Centre. Traditional mechanistic crop models, like ECroPS, require numerous meteorological inputs and significant computational resources, limiting scalability for applications such as large-scale climate simulations or ensemble modeling that explore variables like climate projections and CO₂ effects. In contrast, the surrogate AI model relies on just three weather inputs—daily minimum and maximum temperatures and daily precipitation—trained using ECMWF-ERA5 reanalysis data. This streamlined approach demonstrates the potential to bridge the gap between resource-intensive crop modeling and scalable, data-driven solutions for climate impact assessments.

How to cite: Vlachopoulos, O., Luther, N., Ceglar, A., Toreti, A., and Xoplaki, E.: Surrogate impact modelling for crop yield assessment with nested RNNs, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-13061, https://doi.org/10.5194/egusphere-egu25-13061, 2025.

How to cite: Laanti, T., Ezhova, E., Lintunen, A., Noe, S., Kulmala, M., Miles, V., and Heljanko, K.: XAI finds signs of clouds in the Net Ecosystem Exchange of boreal forest , EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-13234, https://doi.org/10.5194/egusphere-egu25-13234, 2025.

Making projections of possible future climates with models is essential to improve our understanding of the causes and implications of anthropogenic climate change. While Earth system models are currently the most complete description of the Earth system, these models are computationally expensive. Simpler models (emulators) are therefore useful to explore the large space of possible future climate scenarios and to generate large ensembles. One class of emulators are simple climate models (SCMs), which model the Earth system with simplified physics. A second class of emulators are statistical models, which learn relationships directly from correlations in climate model data. In this preliminary work, we seek to combine the benefits of the physical grounding of SCMs with those of purely statistical emulators, using tools from causal representation learning. The resulting causal climate emulator may allow exploration of the effect of various interventions on the Earth system, including the effect of changing forcings.

The goal of causal representation learning (CRL) is to simultaneously learn low-dimensional latent representations from high-dimensional data, and a causal graph between these latent representations. In the context of climate model data, we aim to infer latent variables representing regions with shared climate variability from fine-grid climate model data, and causal teleconnections between these regions, representing climate dynamics. We build on recent previous work by Boussard et al., which illustrated how a CRL method, Causal Discovery with Single-parent Decoding (CDSD), may be used for this task. CDSD is a continuous optimization method to learn a distribution over latent variables such that every grid-point observation is driven by a single latent variable, and a causal graph between these latents is also learned. 

We illustrate that on surface fields of monthly pre-industrial climate model data, CDSD learns physically-reasonable latent variables but learning a robust causal graph between the latent variables remains a challenge. We evaluate our models on synthetic data that approximate the spatiotemporal structures that we observe in climate model data. By autoregressively rolling out the model we can then generate an ensemble of future climate trajectories with the learned generative model. We develop a Bayesian filter to maintain a constant spatial spectrum throughout our autoregressive rollout, and show that it leads to stable climate prediction. Finally, we explore approaches for including the effect of forcings such as greenhouse gasses in the model.

How to cite: Boussard, J., Hickman, S., Trajkovic, I., Kaltenborn, J., Gurwicz, Y., Nowack, P., and Rolnick, D.: Causal climate emulation, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-13307, https://doi.org/10.5194/egusphere-egu25-13307, 2025.

Earth system models (ESMs) are widely used to study climate changes resulting from both anthropogenic and natural perturbations. Over the past years, significant advances have been made through the development of new numerical schemes, refined physical parameterizations, and the use of increasingly powerful computers. Despite these advances, tuning ESMs to accurately reproduce historical data remains largely a manual process, and persistent errors and biases continue to challenge their accuracy. Reducing uncertainties in long-term climate projections and accurately estimating the spread of climate simulations continue to be critical challenges.

Recent advances in machine learning have motivated the development of learning-based methods for the calibration of ESMs. One emerging area of research is the design of hybrid modeling approaches, which combine a physical core with a machine learning model. Training these hybrid models end-to-end (or online) has the potential to unify various challenges in ESMs development, ranging from building subgrid scale parameterizations, to bias correction and parameter tuning.

Training hybrid models online requires working with an optimization problem that depends on the numerical integration of the system. Solving this optimization problem using gradient-based approaches requires the system to be differentiable, or to have access to the adjoint of the numerical model, which is not the case for most of the large-scale physical models. Beyond the need for differentiability, developing hybrid models requires interfacing a physical core that is implemented in low-abstraction languages that are running on CPUs, with AI-based models that are developed using high-abstraction, rapidly evolving languages that run on GPUs. While this interface is not a problem at inference time, doing this interface at calibration time, which is necessary when doing online learning, is not trivial as it would require an iterative communication between components that are implemented on different architectures.

In this work, we aim to investigate online learning and hybrid models to develop new computing paradigms, tools, and calibration methods for designing numerical models that are closely aligned with observations. We study the potential of online learning for deriving efficient and scalable solutions to the above-mentioned problems for applications that include both short-term forecasting and long-term simulations, which require stability considerations of the resulting hybrid systems. We explore learning configurations that include both fully differentiable and black-box physical cores. The latter configuration aims at evaluating the extent to which differentiable programming frameworks can upscale modeling capabilities in terms of accuracy, computational efficiency, and adaptability to represent diverse physical processes.

How to cite: Ouala, S., Meunier, E., Fablet, R., and Le Sommer, J.: Leveraging Differentiable Programming and Online Learning for the design of Hybrid Numerical Models, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-13717, https://doi.org/10.5194/egusphere-egu25-13717, 2025.

Clouds influence Earth’s climate by reflecting sunlight and trapping heat, but their role in climate change remains uncertain, causing major unpredictability in models. Global 3D cloud data can improve predictions.

Observations from NASA’s CloudSat mission have advanced our understanding of cloud structures but are limited by long revisit times and narrow coverage. Imaging instruments offer broader, faster coverage but lack vertical information.

In [1] a deep learning approach addressed this challenge by combining MSG/SEVIRI satellite imagery with CloudSat profiles to extrapolate vertical cloud structures beyond observed tracks. Using geospatially-aware Masked Autoencoders, models were pre-trained on a year of MSG data (2010) and fine-tuned with CloudSat tracks as ground truth. This self-supervised training improved reconstruction, outperforming previous methods and simpler architectures [2].

In this work, we explore to what degree including information of the temporal dynamics of clouds can further improve the quality of the 3D cloud reconstruction. Instead of using a single image  as input we use a temporal sequence of MSG/SEVIRI images, spanning a period of several hours before and after the target cloud vertical profile. We use a combination of the geospatial encodings used in [1] and the temporal encoding used in [3] to embed these spatiotemporal MSG/SEVIRI cubes in rich, general purpose latent space. We then use a finetuning model as in [1] to map the embeddings into 3D radar reflectivity maps. 

We perform a sensitivity analysis to explore how the quality of the reconstruction varies as a function of the amount of temporal information included. We also explore the relative strengths of different pre-training strategies with respect to the quality of the 3D reflectivity reconstruction and cloud type segmentations. With this, we provide insights on self-supervised learning for atmospheric applications.

References

• Stella Girtsou et al. “3D Cloud reconstruction through geospatially-aware Masked Autoencoders” 2024. arXiv: 2501.02035 [cs.CV]. URL: https://arxiv.org/abs/2501.02035.
• Sarah Brüning et al. “Artificial intelligence (AI)-derived 3D cloud tomography from geostationary 2D satellite data”. en. In: Atmos. Meas. Tech. 17.3 (Feb. 2024), pp. 961–978.
• Yezhen Cong et al. SatMAE: Pre-training Transformers for Temporal and Multi-Spectral Satellite Imagery. 2023. arXiv: 2207.08051 [cs.CV]. URL: https://arxiv.org/abs/2207.08051.

How to cite: Diaz, E., Bintsi, K.-M., Castiglioni, G., Eisinger, M., Freischem, L., Girtsou, S., Johnson, E., Jones, W., Jungbluth, A., and Massant, J.: Reconstructing 3D vertical cloud profiles using cloud dynamics, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-13734, https://doi.org/10.5194/egusphere-egu25-13734, 2025.

Reducing climate model biases is crucial for decreasing uncertainties in future climate projections. Despite recent efforts, improvements between the latest generations of Earth System Models (ESMs) have been modest, primarily due to the continued reliance on subgrid-scale parametrizations. These parametrizations are necessary because the model resolutions in CMIP6 are too coarse to explicitly simulate too small-scale processes such as ocean mesoscale eddies and deep atmospheric convection, which significantly influence regional and global climate patterns. Recent advances in computational power have enabled higher-resolution models, allowing for some of these processes to be simulated explicitly, reducing the need for parametrization. Here, we combine a convolutional neural network (CNN) classifier and explainable AI (XAI) to investigate the role of increased resolution in simulating winter surface temperature fields. The CNN is used to classify ESMs with varying resolutions based on snapshots of their surface temperature fields, while the XAI approach explains which regions and features the CNN relies on to make these distinctions, providing deeper insights into ESM performance. Results indicate that models with similar ocean grids are more frequently confused by the CNN than those from similar modeling centers, emphasizing the crucial role of ocean resolution, particularly the presence of mesoscale eddies, in shaping climate simulations. Although the analysis is restricted to surface air temperature, the XAI approach offers a more nuanced understanding of model differences compared to traditional bias analyses. This methodology can be extended to other climate variables and ESM features, offering a powerful tool for enhancing model intercomparison and evaluating ESM performance.

How to cite: Michel, S., Strommen, K., and Christensen, H.: Unravelling the role of increased model resolution on surface temperature fields using explainable AI, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-15256, https://doi.org/10.5194/egusphere-egu25-15256, 2025.

Climate change is profoundly affecting ecosystems and societies. Impacts on hydrological regimes, water resources, and urban heatwaves are particularly important, emphasizing the need for a detailed understanding of these changes at local scales to inform effective adaptation strategies. Achieving this requires reliable, high-resolution projections of future climate conditions. However, current climate models operate at coarse spatial resolutions, limiting their ability to capture small-scale processes and extreme weather events. To bridge this gap, robust downscaling techniques are essential for refining the outputs of global and regional climate models.

We propose a multivariate super-resolution (SR) approach to downscale temperature and precipitation data in Switzerland to improve the representation of localized patterns, particularly in Alpine regions, while simultaneously capturing the interdependencies between temperature and precipitation, which are crucial for hydrological applications. We leverage advanced machine learning techniques, including Generative Adversarial Networks (GANs) and Diffusion models, to overcome the limitations of classical methods in capturing inter-variable dependencies. These models provide an ensemble framework, providing multiple possible realizations, to account for downscaling uncertainties, resulting in more robust and reliable outputs for impact modeling and decision-making. We test different loss functions, like a regional CRPS, to allow for variability in the generated meteorological fields.

We compare the performance of GANs and Diffusion models along with the differences between univariate and multivariate settings. Our approach includes applying a multivariate bias correction prior to downscaling. The downscaled results are compared to a setting based on univariate bias correction. Additionally, we present the pipeline, which integrates bias correction and downscaling and is intended to be open source.

How to cite: Horton, P., Samarin, M., Otero, N., Allen, S., and Volpi, M.: Multivariate climate downscaling using deep learning models, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-15363, https://doi.org/10.5194/egusphere-egu25-15363, 2025.

Understanding the potential impacts of climate change on global vegetation dynamics is crucial for effective environmental management and biodiversity conservation. This study employs a machine learning-based framework to analyze historical NDVI data and project future vegetation growth under different climate scenarios. Utilizing the GIMMS NDVI dataset (1981–2000) for model training and CMIP6 climate projections (2021–2100) for scenario analysis, the study evaluates changes in vegetation growth across four Shared Socioeconomic Pathways (SSPs). Results indicate a significant near-term increase in global mean NDVI (2021–2040) under all scenarios, followed by divergent trends. While SSP126 and SSP245 sustain modest increases, SSP370 and SSP585 show sharp declines in NDVI over the long term, driven by adverse temperature effects. Regional analyses reveal contrasting patterns: NDVI values in Africa, South America, and Oceania decline under most scenarios, while North America, Europe, and Asia exhibit potential increases, except under high-emission scenarios like SSP585. These findings underscore the importance of targeted interventions to mitigate climate impacts and highlight the role of machine learning in predicting vegetation responses to environmental changes. The study provides actionable insights for policymakers, emphasizing the need for sustainable land management practices and greenhouse gas reduction strategies to preserve global ecosystems.

How to cite: Nguyen, A. K. and Chen, W.: Machine Learning Projections of Climate Change Impacts on Global Vegetation Dynamics, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-15622, https://doi.org/10.5194/egusphere-egu25-15622, 2025.

Earth’s seasonality profoundly influences nearly every aspect of life on our planet. It plays a key role in driving vegetation cycles and shaping wildlife behavior. Seasonality also impacts human life significantly, affecting health, mood, social dynamics, and cultural patterns. Despite its importance, seasonality is still traditionally defined by astronomical seasons—equal-length divisions applied uniformly across the Earth. Although this division is simple and intuitive, it overlooks crucial seasonal patterns influenced by atmospheric weather. In this study, we propose a data-driven approach to redefining seasons using objective clustering. We develop an algorithm that segments various meteorological factors into meaningful seasonal clusters. Building on this algorithm, we objectively define seasons for each region globally and analyze the effects of Climate Change on these clusters. We find that seasonality is driven by different meteorological factors in different regions on Earth. Additionally, we observe that Climate Change has significantly altered the duration and onset of Earth’s seasons.

How to cite: Shmuel, A., Magaritz-Ronen, L., Raveh-Rubin, S., and Milo, R.: Revisiting Earth’s Seasonality using Machine Learning Models, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-16234, https://doi.org/10.5194/egusphere-egu25-16234, 2025.

Machine learning (ML), and deep learning (DL) in particular, hold the potential to solve long-standing challenges in understanding and modeling the Earth system. Earth system model (ESM) development is reluctant to implement DL algorithms because they are considered intransparent, meaning it is unclear how these models extrapolate to unseen conditions, e.g., under a changing climate. Still, machine learning is often used to extrapolate into the future,  which can lead to misleading results.
We demonstrate these limitations and the dangers of performing naive extrapolation by using a set of deep neural networks to emulate simulated data of gross primary production (GPP). We use a process-based model (PBM) that simulates photosynthetic CO2 uptake as a product of radiation (PAR), stress from daily meteorology (f_Tmin , f_VPD , f_SM ), vegetation state (fPAR), and CO2 (ε_max(CO2 )). It is given by

GPP = ε_max (CO2 ) · PAR · fPAR · f_Tmin · f_VPD · f_SM + ε.                                                           (1)

The PBM contains many of the typical challenges when using ML for Earth’s system science. It accounts for stochastic noise (ε), is capable of exhibiting multi-year memory, and the predictors are highly correlated on multiple time scales. Further, this model exhibits interesting extrapolation behavior as some of the factors  (f_Tmin , f_VPD , f_SM , f_PAR) saturate in extreme meteorological conditions while others (PAR, ε_max ) do not. We feed the PBM with predictors obtained from historical and future climate simulations of a comprehensive Earth system model. The training dataset contains the predictors and predictions of the PBM for various locations in a similar climate zone but different continents and for the historical time frame (1850-present) together with a spurious predictor, namely, surface wind speed. To obtain a set of independent models, each of the co-authors separately implements a custom architecture, without knowing which predictor is which. This results in four different models, namely a linear model, a multi-layer perceptron, a long-short term memory (LSTM), and an attention-based model.
We find that all models show strong prediction performance in cross-validation (Normalized Nash–Sutcliffe Efficiency (NNSE) > 0.9), decent performance when extrapolating to sites on different continents (NNSE > 0.7), but three out of four models show virtually no skill when predicting to a changed climate (NNSE < 0.6). Additionally, most models emit gradients in the same order of magnitude as the PBM when ignoring values where  some factors saturate. This indicates that the networks did not learn the saturation behavior from the data. Further, the model that extrapolates best is the LSTM, a model that has a built-in maximum output and, hence, has to saturate.
In conclusion, strong spatial generalization and cross-validation performance do not guarantee decent extrapolation for neural networks even in relatively simple, stable systems. These findings highlight the importance of selecting architectures in line with the expected extrapolation behavior when predicting Earth’s system processes under climate change conditions.

How to cite: Reimers, C., ElGhawi, R., Kraft, B., and Winkler, A. J.: Limitations of Machine Learning Models in Extrapolating to a Changing Climate, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-16360, https://doi.org/10.5194/egusphere-egu25-16360, 2025.

The performance of a deep convolutional neural network in predicting near-surface air temperature (T2m) and total precipitation (P) over Europe is assessed, comparing its results with the Copernicus European Regional Reanalysis (CERRA) and the regional dynamical model HCLIM. The ML-model accurately captures broad seasonal temperature and precipitation patterns, with minor biases in summer and more pronounced warm biases in winter. While the model effectively reproduces the probability density functions (PDFs) of daily temperature and precipitation, it underestimates extreme cold events and the high precipitation extremes in some regions. Climate indices, including cold extremes (TM2PCTL), warm extremes (TM98PCTL), consecutive dry days (CDD), and consecutive wet days (CWD), highlight that the ML model aligns closely with CERRA. However, it slightly underestimates CDD and overestimates CWD, particularly in mountainous and Mediterranean regions. Analyses of spatio-temporal variability demonstrate high correlations with CERRA for temperature, exceeding 0.99 for spatial patterns and 0.95 for temporal correlations, while correlations for precipitation are lower, with underestimated temporal variability. The ML model generally outperforms HCLIM, particularly in aligning with observed data, although challenges remain in capturing extremes and reducing biases in certain regions. These results further highlight the potential of the ML model for regional climate downscaling and impact studies, while emphasizing the need for further refinement to enhance its representation of extreme events and improve spatial accuracy.

How to cite: Fuentes-Franco, R., Ivanov, M., Koenigk, T., Krus, K., Aldama Campino, A., and Wang, F.: Assessment of a Pan-European high-resolution downscaling through Deep Learning, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-17278, https://doi.org/10.5194/egusphere-egu25-17278, 2025.

Despite the remarkable strides made by AI-driven models in modern precipitation forecasting, these black-box models cannot inherently deepen the comprehension of underlying mechanisms. To address this limitation, we propose an AI-driven knowledge discovery framework known as genetic algorithm-geographic weighted regression. Through this framework, we have constructed an iterative optimization of knowledge generation and utilization. On the one hand, new explicit equations are discovered to describe the intricate relationship between precipitation patterns and terrain characteristics. Experiments have shown that the discovered equations demonstrate remarkable accuracy when applied to precipitation data, outperforming conventional empirical models. Notably, our research reveals that the parameters within these equations are dynamic, adapting to evolving climate patterns. On the other hand, these previously undisclosed equations have contributed new knowledge about terrain-precipitation relationships, which can be embedded into the AI model for better interpretability and climate projection accuracy. Specifically, the unveiled equations can enable fine-scale downscaling for precipitation predictions using low-resolution future climate data. This capability offers invaluable insights into the anticipated changes in precipitation patterns across diverse terrains under future climate scenarios, which enhances our ability to address the challenges posed by contemporary climate science.

How to cite: Xu, H., Chen, Y., Zeng, Z., Li, N., Li, J., and Zhang, D.: Exploring Terrain-Precipitation Relationships with Interpretable AI for Advancing Future Climate Projections, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-17628, https://doi.org/10.5194/egusphere-egu25-17628, 2025.

In recent years, machine learning has emerged as a promising alternative to numerical weather prediction models, offering the potential for cost-effective and accurate forecasts. However, a significant limitation of current machine learning methods for weather forecasting is the lack of principled and efficient uncertainty quantification—a key element given the complexity of the Earth's climate system and the challenges in modeling its processes and feedback mechanisms. Inadequate uncertainty quantification and reporting undermines trust in and the practical use of current weather forecasting methods (Eyring et al., 2024).

Uncertainty quantification methods for weather forecasting typically use prediction intervals and can be categorized into Bayesian and frequentist approaches. Bayesian methods, while theoretically appealing, often involve restrictive assumptions and do not scale well to the complexity of spatio-temporal data. Frequentist approaches, such as ensemble-based methods, are widely used in weather forecasting and include techniques like perturbing initial states with noise (Bi et al., 2023; Scher et al., 2021), varying neural network parameters (Graubner et al., 2022), or training generative models (Price et al., 2023). However, most frequentist methods provide only asymptotically valid prediction intervals, which may not suffice in all weather forecasting applications.

Conformal prediction (CP) is a promising uncertainty quantification framework that delivers valid and efficient prediction intervals for any learning algorithm, without requiring assumptions about the underlying data distribution (Vovk et al., 2005). Despite its growing popularity in the machine learning and statistics communities, traditional CP methods are not tailored to spatio-temporal data in weather forecasting. This is due to challenges arising from spatial and temporal dependencies—such as spatial autocorrelation and temporal dynamics—that violate the exchangeability assumption underlying standard CP methods. Several recent studies attempted to address these challenges by introducing new CP algorithms specifically designed for various types of non-exchangeability (Oliveira et al., 2024). However, these adaptations face several limitations, including high computational complexity, asymptotic guarantees, and/or the need for recalibration of prediction intervals.

In this presentation, we will evaluate CP methods in the context of weather forecasting and discuss several limitations. In addition, we will highlight recent advances and discuss potential future directions that could address challenges underlying the use of CP in weather forecasting.

References:

Eyring, V., et al. Pushing the Frontiers in Climate Modelling and Analysis with Machine Learning. Nature Climate Change, 2024.

Leutbecher, M., et al. Ensemble Forecasting. JCP, 2008.

Bi, K., et al. Accurate Medium-range Global Weather Forecasting with 3D Neural Networks. Nature, 2023.

Scher, S., et al. Ensemble Methods for Neural Network-based Weather Forecasts. JAMES, 2021.

Graubner, A., et al. Calibration of Large Neural Weather Models. NeurIPS, 2022.

Price, I., et al. Probabilistic Weather Forecasting with Machine Learning. Nature, 2025.

Vovk, V., et al. Algorithmic learning in a random world. New York: Springer, 2005.

Oliveira, R.I., et al. Split Conformal Prediction and Non-exchangeable Data. JMLR, 2024.

How to cite: Mortier, T., Decancq, C., Sale, Y., Javanmardi, A., Waegeman, W., Hüllermeier, E., and Miralles, D. G.: Valid Prediction Intervals for Weather Forecasting with Conformal Prediction, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-17783, https://doi.org/10.5194/egusphere-egu25-17783, 2025.

Earth System Models (ESMs) are fundamental to understanding and projecting climate change. While they have demonstrated continuous improvements over the last decades, systematic errors and large uncertainties in their projections remain. A large contribution to these uncertainties stems from the representation of unresolved processes such as clouds and convection that occur at scales smaller than the model grid spacing. This impacts the models’ ability to accurately project global and regional climate change, climate variability, and extremes. High-resolution models with horizontal grid spacing of a few kilometers or less alleviate many biases of coarse-resolution models, but at high computational costs. Yet short simulations from high-resolution models can be used to inform machine learning (ML)-based parameterizations that are then incorporated into hybrid (physics+ML) ESMs. This new generation of hybrid models promises to reduce systematic errors and enhance projection capabilities compared to current state-of-the-art ESMs [1, 2]. In an effort to design a comprehensive hybrid ESM, the ICOsahedral Non-hydrostatic (ICON) model is equipped with a variety of physics-aware ML parameterizations, including moist convection, cloud cover and radiation. This talk will present an overview of the modelling activities undertaken within this framework, with a special focus on the developed ML-based cloud cover parameterization. This parameterization takes the form of an interpretable non-linear equation discovered through a combination of ML techniques including symbolic regression and sequential feature selection [3]. We demonstrate that, with this new parameterization, ICON runs stably over several decades and reduces global biases in cloud cover and radiation metrics. In addition, the new equation is controlled by only 10 free parameters that we automatically calibrate to achieve more accurate climate projections. This approach of discovering a low-dimensional data-driven equation for a parameterization with subsequent tuning of the hybrid model can be used in any host ESM provided suitable training data.

References:

[1] Eyring, V., Collins, W.D., Gentine, P. et al., Pushing the frontiers in climate modeling and analysis with machine learning, Nat. Climate Change, doi:10.1038/s41558-024-02095-y, 2024.

[2] Eyring, V., Gentine, P., Camps-Valls, G., Lawrence, D.M., and Reichstein, M., AI-empowered Next-generation Multiscale Climate Modeling for Mitigation and Adaptation, Nat. Geosci., doi:10.1038/s41561-024-01527-w, 2024.

[3] Grundner, A., Beucler, T., Gentine, P. and Eyring, V., Data-driven equation discovery of a cloud cover parameterization, J. Adv. Model. Earth Sys., doi:10.1029/2023MS003763, 2024.

How to cite: Savre, J., Schwabe, M., Grundner, A., Hefner, K., Heuer, H., Klamt, J., Pastori, L., Schlund, M., Gentine, P., and Eyring, V.: Towards a prototype hybrid ICON-ML model with physics-aware machine learning parameterizations, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-17999, https://doi.org/10.5194/egusphere-egu25-17999, 2025.

Aerosols affect the Earth’s energy budget by both scattering and absorbing solar radiation. Measuring parameters that separately quantify the two components, such as the aerosol absorption optical depth (AAOD), is key to better understanding the aerosol direct climate effect. As most satellite instruments can only retrieve the total aerosol extinction signal, the most reliable source of global AAOD observations is the ground-based AERONET sensor network. AERONET comprises hundreds of stations worldwide; however, their spatial distribution is uneven and coverage remains sparse in many relevant regions. To effectively reduce our uncertainty related to absorbing aerosols and efficiently expand the network, new stations should be placed in locations that maximise measurement informativeness. In this study, we address the problem of optimal sensor placement using convolutional neural processes (ConvNPs). ConvNPs are meta-learning models that use convolutional neural networks to learn maps from heterogeneous input datasets to a context-dependent Gaussian predictive model. We train ConvNPs using reanalysis data to learn to model daily global AAOD from sparse point observations given at station locations and additional gridded auxiliary data. The model’s probabilistic predictions are then harnessed in an active learning framework to sequentially propose new observation locations that optimally reduce model uncertainty and improve the network's informativeness. Our subsequent analysis considers further practical factors that might trade off with informativeness in the selection of new station locations, such as cloudiness and remoteness. The resulting proposed placements identify locations that would optimally enhance ground-based AAOD observation and can inform and focus future network expansion efforts.

How to cite: Pelucchi, P., Coca-Castro, A., Andersson, T. R., Vicent Servera, J., and Camps-Valls, G.: Optimal Sensor Placement for Aerosol Absorption Optical Depth with Convolutional Neural Processes, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-18735, https://doi.org/10.5194/egusphere-egu25-18735, 2025.

How to cite: Atkinson, J., O'Gorman, P., Berner, J., and Weinzierl, M.: Coupling a new convection parameterisation trained using high-resolution simulations to the Community Atmospheric Model, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-19177, https://doi.org/10.5194/egusphere-egu25-19177, 2025.

Neural operators are transforming computationally intensive scientific disciplines such as weather forecasting and climate modeling, accelerating simulations by several orders of magnitude. However, they often fail to respect fundamental physical principles, such as conservation laws, during long autoregressive rollouts. We introduce an efficient correction layer that enforces global conservation constraints in neural operators. For initial conditions approximately satisfying the constraints, we prove that conservation can be guaranteed while only moderately increasing the total runtime. In a number of fluid dynamics experiments, our method produces physically realistic simulations while maintaining the computational advantages of neural operators. Our results enable the development of reliable and efficient climate model emulators by ensuring that crucial physical balance equations, such as mass and energy, are preserved during extended simulations.

How to cite: White, A., Duruisseaux, V., Bonev, B., Azizzadenesheli, K., Anandkumar, A., and Boers, N.: Enforcing Conservation Laws in Neural Operators for Earth System Modeling, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-19180, https://doi.org/10.5194/egusphere-egu25-19180, 2025.

Generative Deep Learning architectures, such as Diffusion models, offer an alternative to traditional physical modeling and regression models due to their ability to produce stochastic ensembles with a single run. Even though these models are capable of downscaling coarse data, they are often trained in contained regions, which can lead to severe spatial overfitting as the model learns location-specific patterns rather than generalizable physical relationships. In practice, the usability of the models is constrained to the area where they were originally trained, and their predictive capabilities degrade significantly when applied to regions outside the training domain, even if these regions share similar characteristics.
This study presents a one-step and two-step diffusion model capable of downscaling 2-meter temperature from ERA5 to higher-resolution grids in large areas, such as the Contiguous United States or Europe, without spatially overfitting. We use CONUS404, a reanalysis dataset created using simulations of the Weather Research and Forecasting (WRF) model over the Contiguous United States, as our target data and ERA5 and constants involved in the creation of CONUS404, such as altitude and land use, as our input. The model has been trained over the whole area using 10 years of 3-hourly data, and two years have been used for testing. To study the spatial generalization capabilities of the model, we reserve an area of the study region solely for testing and compute evaluation metrics separately for this area to ensure meaningful results. We compare the results of training in large and small areas and the number of years. In addition, we discuss the usefulness of ensemble prediction and the effect that the number of ensemble members has on the performance of the downscaling. Future steps include applying this methodology for downscaling EURO-CORDEX to EMO1 and multivariate downscaling.

How to cite: Benitez Benavides, M., Rodríguez, M., Margalef, T., Panadero, J., and Dutta, O.: Stochastic diffusion model for large-scale temperature downscaling, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-19397, https://doi.org/10.5194/egusphere-egu25-19397, 2025.

We consider a data-driven framework for climate prediction tasks in which the dynamics are learnt in a low-dimensional latent space. We rely on dimensionality reduction techniques --- linear principal component analysis and nonlinear autoencoders and their variants --- to then learn dynamical evolution  in the corresponding latent space using disparate methodologies --- linear inverse modeling, dictionary-based sparse regression, reservoir computing, neural differential equations, attention-based transformers, etc. In this setting, we seek to better understand the interplay between the spatial and temporal representations of variability and how they affect prediction skill.

Balu Nadiga, Los Alamos National Laboratory and Kaushik Srinivasan, University of California Los Angeles

How to cite: Nadiga, B. and Srinivasan, K.: Climate Prediction Based on Latent Space Dynamics, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-21270, https://doi.org/10.5194/egusphere-egu25-21270, 2025.

Seismic data forms the backbone of what we understand in the subsurface, and seismic data interpretation is still usually done by hand. Automatic seismic interpretation with deep learning is very promising, but there the problem is a lack of labelled training data. In this study, we use forward stratigraphic modelling and show how forward modelling can be advantageously used in deep learning.

Specifically, we focus on shelf-edge trajectories as the geological representations of lateral and vertical shifts in sediments’ position through time. They provide continuous tracks of changes in relative sea-level as well as sediment stacking patterns and depositional geometries. Mapping these trajectories and measuring their changing angles help in quantifying the sequence stratigraphic analysis and predicting ancient depositional environments.

Here, we evaluate the ability of deep learning models, trained on synthetic seismic data, to identify clinoforms and their rollover points for shelf-edge trajectories mapping. The synthetic training dataset generated using geological processed-based forward modelling represents different depositional slope scenarios. Controlling the different parameters that govern shelf-edges and shelf-edge trajectories (such as bathymetry, sediment supply, eustatic sea-level changes and subsidence) gave us a better chance to mimic realistic and diverse depositional setting, which helps in generalizing the deep learning model. In addition, the ground truth (labels) for the created synthetic seismic data is automatically generated by the forward model, without the need of manual labelling seismic data.

Higher accuracy score on both validation and testing datasets demonstrates the power and effectiveness of using synthetic as training dataset. This study shows that synthetic data can play a major role in bridging the gap between traditional seismic interpretation and automating the process using machine learning. It also shows that forward modelling is a powerful technique to combine with data modelling, such as machine learning.

How to cite: AlGharbi, W., Bell, R., and John, C.: Forward Stratigraphic Modelling to Generate Synthetic Seismic Training Dataset for Deep Learning: A Case Study to Predict Shelf-Edge Trajectories, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-288, https://doi.org/10.5194/egusphere-egu25-288, 2025.

This study aims to develop a CNN-LSTM hybrid network model integrated with a coupled self-attention mechanism, based on deep learning techniques, to simulate flood processes in the Inner Harbor area of Macau. With global climate change and accelerated urbanization, Macau, a low-lying coastal city, frequently experiences urban flooding due to typhoons and heavy rainfall. While traditional hydrological and hydrodynamic models can accurately predict flooding processes, their computational intensity and lack of real-time responsiveness make them unsuitable for emergency disaster warnings. To address these limitations, this paper proposes a convolutional long short-term memory (ConvLSTM) model enhanced with a coupled self-attention mechanism. The model leverages an encoder-decoder structure to predict the evolution of flood processes under 4–10 hours of heavy rainfall in the Inner Harbor area of Macau.

The model integrates CNN components for extracting spatial features, LSTM components for capturing temporal features, and a coupled self-attention mechanism to dynamically reweight spatial-temporal representations, improving the model's sensitivity to key flood patterns. The encoder encodes input sequences into fixed-length vectors, while the decoder translates these vectors into target sequences. The self-attention mechanism ensures the model focuses on critical spatial and temporal regions, further enhancing prediction accuracy and robustness.

The training and testing datasets were constructed from simulation data generated by hydrological-hydrodynamic models and static geographical information data, following preprocessing and normalization. Evaluation metrics, including mean squared error (MSE), Nash-Sutcliffe efficiency coefficient (NSE), and relative error, were used to assess model performance. Results demonstrate that the proposed hybrid model, augmented by the coupled self-attention mechanism, effectively simulates maximum water depth distribution and flood evolution processes, achieving high consistency with hydrodynamic simulation data while providing improved predictive performance.

How to cite: Wangqi, L.: Flood Process Simulation in Macau's Inner Harbor Area Based on CNN-LSTM, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-2197, https://doi.org/10.5194/egusphere-egu25-2197, 2025.

AI enhanced environmental modelling workflows: Towards Automated Scientific Exploration in Hydrology

Authors: Darri Eythorsson, Kasra Keshavarz, Cyril Thébault, Mohamed Ismaiel Ahmed, Raymond Spiteri, Alain Pietroniro and Martyn Clark

Modern hydrological modeling has evolved into a complex scientific endeavour requiring sophisticated workflows that span multiple scales, processes, and computational paradigms. While existing workflow solutions address specific technical challenges, the field lacks comprehensive frameworks that can support end-to-end modeling while maintaining reproducibility and scalability. This works introduces two complementary frameworks that aim to address these fundamental challenges: CONFLUENCE (Community Optimization Nexus For Large-domain Understanding of Environmental Networks and Computational Exploration) and INDRA (the Intelligent Network for Dynamic River Analysis).

CONFLUENCE implements a modular architecture that enforces workflow reproducibility through a unified configuration system while maintaining the flexibility needed to support diverse modeling applications. The framework provides comprehensive solutions for four critical workflow components: (1) flexible geospatial domain definition and discretization, (2) model-agnostic data acquisition and preprocessing, (3) extensible model setup and parameterization capabilities, and (4) comprehensive evaluation and optimization tools. This systematic approach enables efficient, reproducible, and transparent hydrological modeling across scales.

INDRA augments this foundation by implementing a network of specialized AI expert agents that support various components of the hydrological modeling workflow. Through structured dialogue between domain experts (including AI specialists in hydrology, hydrogeology, meteorology, data science, and geospatial analysis), INDRA provides context-aware guidance while maintaining complete provenance of modeling decisions and their justification. This AI-assisted approach helps address three critical challenges: (1) the growing complexity of modelling decisions, (2) the need for reproducible workflows and detailed documentation, and (3) the technical barriers limiting broader adoption of advanced modeling practices.

The integration of these frameworks aims to explore how automation and AI assistance can enhance rather than disrupt traditional modeling practices. By maintaining clear documentation of decisions and their justifications, these systems help build trust in model results while creating opportunities for recursive learning from previous modeling experiments. Our case studies, spanning scales from individual catchments to continental domains, showcase the frameworks' capabilities while highlighting their potential to transform how researchers’ interface with complex environmental modeling workflows.

This work aims to advance both operational and research oriented hydrological modeling practices, offering a foundation for reproducible, scalable, and interoperable modeling while maintaining scientific rigor and flexibility. The framework’s open-source nature and modular design create opportunities for community-driven development and extension, potentially accelerating scientific discovery in hydrological sciences.

How to cite: Eythorsson, D., Keshavarz, K., Thébault, C., Ahmed, M., Spiteri, R., Pietroniro, A., and Clark, M.: AI enhanced environmental modelling workflows: Towards Automated Scientific Exploration in Hydrology, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-4565, https://doi.org/10.5194/egusphere-egu25-4565, 2025.

This study explores integrating reinforcement learning (RL) with idealised climate models to address key parameterisation challenges in climate science. Current climate models rely on complex mathematical parameterisations to represent sub-grid scale processes, which can introduce substantial uncertainties. RL offers capabilities to enhance these parameterisation schemes, including direct interaction, handling sparse or delayed feedback, continuous online learning, and long-term optimisation. We evaluate the performance of eight RL algorithms on two idealised environments: one for temperature bias correction, another for radiative-convective equilibrium (RCE) imitating real-world computational constraints. Results show different RL approaches excel in different climate scenarios with exploration algorithms performing better in bias correction, while exploitation algorithms proving more effective for RCE. These findings support the potential of RL-based parameterisation schemes to be integrated into global climate models, improving accuracy and efficiency in capturing complex climate dynamics. Overall, this work represents an important first step towards leveraging RL to enhance climate model accuracy, critical for improving climate understanding and predictions. Code accessible at https://github.com/p3jitnath/climate-rl.

How to cite: Nath, P., Moss, H., Shuckburgh, E., and Webb, M.: RAIN: Reinforcement Algorithms for Improving Numerical Weather and Climate Models, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-5159, https://doi.org/10.5194/egusphere-egu25-5159, 2025.

Machine Learning is gaining increasing attention from the scientific community in hydrological and hydraulic research. However, this field faces a consistent challenge in applying data-driven approaches due to the evident poor generalization capabilities, which are partly a result of inherent data scarcity.

We propose incorporating expert knowledge into data-driven models for river hydraulics and flood mapping by integrating physically-based information without relying on the underlying mathematical formulation (e.g., the calculation of the residuals of differential equations). This approach appears to be particularly valuable for flood simulations, where hydraulically relevant distributed parameters such as roughness, lithology, topography etc. pose significant uncertainties. The method is versatile and applicable to physical systems and scenarios in which the underlying mathematical formulation is not fully known, but expert knowledge enables the introduction of meaningful, physically-inspired constraints.

Specifically, the physical information is integrated into the model by including an additional term, weighted by the hyperparameter in the guise of a regularization term in the loss function :

Here, represents the data-driven error metric, while the physical loss term is an error metric that depends not only on the true and predicted outputs ( ), but also potentially on the inputs . Indeed, this term employs physical principles, laws, and quantities, which are not explicitly formulated in the original dataset. In this sense, we can note its similarity to data augmentation, a widely used technique in machine learning that extracts additional insights by offering alternative interpretations of the same dataset.

We clarify that this approach does not aim to replace numerical solvers or serve as an alternative numerical model, as Physics-Informed Neural Networks do: indeed, their similarity is limited to the formulation of the modified loss function.

We assessed the methodology and empirically quantified the effectiveness of the method in a simplified, well-controlled problem, evaluating the gain in generalisation capability of Neural Networks (NNs) in the reconstruction of the steady state, one-dimensional, water surface profile in a rectangular channel. We found improved predictive capabilities, even when extrapolating beyond the boundaries of the training dataset and in data-scarce scenarios. This kind of assessment is of great relevance to the application of NNs to flood mapping, where cases featuring values of the observed quantities falling out of the range of the recorded series need to be predicted.

New experiments have been also conducted on two-dimensional domains. The data-driven model was trained on a single catchment and tested on its ability to determine flooded areas in unseen catchments. Preliminary results show that an encoder-decoder model with convolutional layers exhibits improved generalization when a physical training strategy is employed. Future applications could include flood mapping for ungauged basins, leveraging similarities with other basins.

How to cite: Guglielmo, G. and Prestininzi, P.: Physically-Enhanced Training of Neural Networks for Hydraulic Modelling of Rivers and Flood Events, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-5991, https://doi.org/10.5194/egusphere-egu25-5991, 2025.

Background: Agriculture is increasingly leveraging machine learning (ML) to enhance yield predictions and optimize agronomic practices. Maize, a staple crop in Ghana, offers a valuable case study for evaluating the effectiveness of diverse ML models in yield prediction and resource management.

Objective: This study aims to evaluate the predictive performance of four ML models namely Random Forest (RF), Support Vector Machine (SVM), K-Nearest Neighbours (KNN), and Extreme Gradient Boosting (XGBoost) for maize yield and agronomic efficiency prediction. It also compares variable importance across these models to determine key explanatory variables.

Methods: The study utilized 4,496 georeferenced maize trial datasets from various agroecological zones in Ghana. Thirty-five explanatory variables included soil properties, climate, topography, crop management practices, and fertilizer application datasets. Model performance was evaluated using leave-one-out, leave-site-out, and leave-agroecological-zone-out cross-validation techniques. Metrics including Mean Error (ME), Root Mean Squared Error (RMSE), and Model Efficiency Coefficient (MEC) were used to compare model accuracy, while a permutation-based approach was employed to assess variable importance.

Results: XGBoost emerged as the most accurate model, achieving the lowest RMSE for yield (639.5 kg ha⁻¹) and agronomic efficiency (11.6 kg kg⁻¹), particularly for nitrogen (AE-N). RF demonstrated competitive performance, while KNN and SVM yielded inconsistent results under rigorous cross-validation conditions. Key explanatory variables identified across models included nitrogen fertilizer, rainfall, and crop genotype, underscoring their critical role in yield and agronomic efficiency outcomes.

Conclusion: XGBoost was the most robust and accurate model for maize yield and agronomic efficiency predictions, offering a reliable tool for data-driven agricultural planning in diverse agroecological settings. The findings underscore the transformative role of advanced ML techniques in modern agriculture, particularly in optimizing staple crop production in sub-Saharan Africa.

How to cite: Asamoah, E., Heuvelink, G., Chairi, I., Bindraban, P., and Logah, V.: Modelling Maize Yield and Agronomic Efficiency Using Machine Learning Models: A Comparative Analysis, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-9987, https://doi.org/10.5194/egusphere-egu25-9987, 2025.

Expert interpretation of borehole data is a critical component of geological modelling, offering essential insights into the spatial distribution of geological units within the subsurface. For large-scale regional mapping efforts, expert interpretation of all available data is impractical due to the sheer volume of boreholes. Therefore, many 3D geological subsurface models rely only on a small portion of all available data. Machine learning (ML) models can be used to automate borehole data interpretation, increasing data density. However, these automated interpretations must adhere to strict spatial and stratigraphical relationships to be consistent with the established geological knowledge of the area. Using a dataset of 1,400 boreholes with expert interpretations from the Roer Valley Graben (Southeast Netherlands), we explore how ML models can be integrated into geological modelling workflows, highlighting the challenge of ensuring compatibility with geological principles and known spatial relationships. We evaluate the model performance using traditional metrics such as accuracy, Cohen's kappa and F1 Score and newly proposed geology-inspired metrics to quantify the ability of Random Forest and Neural Network models to interpret borehole data into lithostratigraphic units while preserving key geological relationships. Our results demonstrate that while many models achieve accuracy values of 75% to 80%, Neural Networks perform significantly better in capturing the expected sequential relationships between geological units, achieving up to 96% of geological transitions between geological units that are plausible, compared to 65% for the best-performing Random Forest model selected based on traditional metrics. This study underscores the need for domain-specific metrics in evaluating model performance and the potential for ML to increase the volume of data incorporated in subsurface models.

How to cite: Garzón, S., Dabekaussen, W., De Boever, E., Busschers, F., Mehrkanoon, S., and Karssenberg, D.: Assessing the Geological Plausibility of Machine Learning Borehole Interpretations: A Case Study in the Roer Valley Graben, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-10915, https://doi.org/10.5194/egusphere-egu25-10915, 2025.

Abstract:
Alluvial terraces in watersheds are geomorphic features formed by river incision and sediment deposit, representing former floodplain levels. They serve as valuable records of fluvial dynamics, climatic changes, and tectonic activity. Mapping methods for terraces that rely on field-acquired data, often involving physical or chemical analyses, are not feasible for large-scale applications. When aiming to map at the national scale, the development of a methodology that eliminates the need for such detailed information enhances scalability and broadens applicability.

We proposed a semi-automatic predictive mapping method for watershed terraces using 25m Digital Earth Model (DEM) provided by the IGN (French geographical service) and derived variables such as curvature, slope, and the difference from a base level (Raingeard et al., 2019). This method demonstrated meaningful results in the Pyrenean Piedmont for the Baïse and Ousse rivers, with the predicted map showing strong alignment with the geological reality.

In this study, we propose an automated approach for identifying alluvial terraces using relative height. Relative height is defined as the difference between the elevation derived from a DEM and the base level. Our methodology is based on the hypothesis that terraces are represented as flat areas in the relative height, where pixels exhibit similar statistical distributions. To capture these patterns, we employ a Gaussian Mixture Model, a probabilistic framework that approximates data as a combination of multiple Gaussian distributions. In this context, each Gaussian distribution corresponds to a specific alluvial terrace.

We conducted experiments on the study areas used by Raingeard et al. (2019), and the results are consistent with both the semi-automatic method and the geological reality. These outcomes provide promising prospects for the predictive mapping of superficial deposits

Reference:

Raingeard A., Tourlière B., Lacquement. F, Baptiste. J, Tissoux. H. Semi-automatic quaternary alluvial deposits mapping - Methodology for the predictive mapping of flat terrains within a watershed, by semi-automatic analysis of the Digital Elevation Model. INQUA 2019, Jul 2019, Dublin, Ireland. 2019.

How to cite: Rakotonirina, H., Lohier, T., Raingeard, A., Lacquement, F., Baptiste, J., and Tissoux, H.: Mapping Alluvial Terraces in Watersheds Using Gaussian Mixture Model on Relative Height., EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-11031, https://doi.org/10.5194/egusphere-egu25-11031, 2025.

Stochastic parameterisations have seen widespread use in atmospheric models. These schemes represent uncertainty by adding terms that include a random noise component directly to the equations that describe the time evolution of the model. Stochastic parameterisation development thus involves the following questions: what are the sources of uncertainty, how do we represent them, and how precisely should we formulate stochastic terms to quantify them? Common methods to quantify the uncertainty inherent in parameterisation include applying multiplicative perturbations to physics tendencies (such that the larger the tendency due to subgrid processes, the more uncertainty we should have in the tendency) or applying perturbations to physical parameters (our physics schemes often rely on parameters whose values we do not know precisely, and have complicated nonlinear responses to perturbing this set of parameters, so perturbing each one within its own specified range during the model run allows this uncertainty to feed back into the model state).

In this work we present a different approach to stochastic parameterisation. We perturb the thermodynamic profiles that constitute the inputs to parameterisation schemes. The perturbations are scaled by the degree of subgrid inhomogeneity. A representation of the subgrid inhomogeneity is estimated by a machine learning model which has been trained on coarse-grained high resolution (dx=~1.5km) model output from the Met Office Unified Model. The scheme is implemented in LFRic, the UK Met Office’s next generation modelling system, and we present results of experiments ran in single column model mode.

How to cite: Reid, H. and Morcrette, C.: A new stochastic physics scheme incorporating machine-learnt subgrid variability, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-11747, https://doi.org/10.5194/egusphere-egu25-11747, 2025.

Accurately forecasting water table rise (WTR) is essential for effective water resource management, infrastructure development, flood risk mitigation, and environmental conservation. This research employed multiple machine learning (ML) models, namely Ridge Linear Regression (RLR), Radial Basis Function Support Vector Machine (RBF-SVM), Linear SVM (LSVM), Random Forest (RF), and a hybrid deep learning Transformer (TR) with Bi-Long Short-Term Memory (BiLSTM), to forecast WTR one and two weeks ahead in the Muscat Governorate, Oman. A total of 19,465 high-resolution datasets, measured at half-hour intervals between December 2017 and January 2019, were utilized. The data were divided into training and testing sets, with 90% (17,976 datasets) used for training and the remaining 10% (1,489 datasets) reserved for testing. A two-way time series analysis was employed to analyze dynamic interactions between two time-dependent behaviors over time. Additionally, the rolling forecasting method was used alongside the models to capture patterns and provide updated predictions based on the most recent data trends. The results demonstrated that RLR outperformed both the individual ML models and the hybrid deep learning TR-BiLSTM models, as indicated by the NSE and RSR statistical metrics applied to the testing data. Furthermore, the one-week step-ahead forecasting achieved greater accuracy in predicting WTR compared to the two-week step-ahead forecast. However, the average computational time of the hybrid deep learning TR-BiLSTM models was notably higher compared to the standalone models. Linear models such as RLR and LSVM demonstrated accurate forecasting results due to their ability to prevent overfitting in correlated features and effectively capture the simplicity of the relationship between the data. The approach presented in this research can be effectively useful to various arid regions worldwide that are influenced by WTR.

How to cite: Elzain, H. E., Abdalla, O., Al-Maktoumi, A., Kacimov, A., and Chen, M.: Water table rise forecasting using machine and deep learning models in arid regions, Oman, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-12464, https://doi.org/10.5194/egusphere-egu25-12464, 2025.

Differentiable programming enables automatic differentiation (AD) tools to compute gradients through code without manually defining derivatives. AD tools can differentiate through entire software stacks, composing many functions and algorithms via the chain rule. With models that incorporate differentiable programming long-standing challenges like systematic calibration, comprehensive sensitivity analyses, and uncertainty quantification can be tackled, and machine learning (ML) methods can be integrated directly into the process-based core of earth-system models (ESMs) to incorporate additional information from observations. Through the advent of ML, several AD tools are gaining traction. A new generation of powerful tools like JAX, Zygote and Enzyme enable differentiable programming for models of varying complexity including highly complex coupled ESMs. Here we present an overview about the perspectives of differentiable programming for ESMs, using experience from two of our applications in atmospheric modelling. First off, we set up PseudoSpectralNet, a differentiable quasi-geostrophic model in Julia with Zygote. This is a hybrid model combining neural networks with a dynamical core, showcasing how the stability and accuracy of ML models is improved by integrating a process-based dynamical core into our model. Additionally, ongoing work uses Enzyme to achieve a differentiable version of the significantly more complex SpeedyWeather.jl atmospheric model. We will discuss advantages of both approaches and give an outlook into future possibilities with differentiable models. 

How to cite: Gelbrecht, M., Klöwer, M., and Boers, N.: Differentiable Programming for Atmospheric Models: Experiences and Perspectives , EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-12601, https://doi.org/10.5194/egusphere-egu25-12601, 2025.

Accurate representation of turbulent processes remains a critical challenge in atmospheric modelling. Large Eddy Simulations (LES) serve as valuable tools for understanding atmospheric turbulence by explicitly resolving energy-containing eddies while parameterizing smaller-scale motions through subgrid-scale (SGS) models. In their most complex forms, these SGS parameterizations can significantly influence LES performance and computational efficiency, making their improvement useful for advancing atmospheric modelling capabilities. Neural Network based emulation of such parametrizations have proven effective in reducing the computational cost, while maintaining accuracy and stability.

Building upon recent advances in physics-informed neural networks (NN) for atmospheric modelling and emulation of physics-based processes, we present a physics-guided NN architecture for emulation of the SGS turbulence parameterizations that introduces several key innovations. Our approach uniquely combines scale-specific normalization with multi-scale feature extraction through parallel convolutional paths, distinguishing it from existing physics-guided machine learning frameworks. The deep learning-based (DL) model also incorporates physically-motivated constraints across different spatial scales while simultaneously ensuring conservation of momentum and energy.

Unlike earlier studies that focus on single aspects of physical conservation, our architecture implements a comprehensive physics-informed framework that combines Richardson number gradient handling for stability constraints, with explicit treatment of diffusion and viscosity coefficients, and scale-specific normalization for different atmospheric variables. The model was trained on limited high-resolution Radiative-Convective Equilibrium (RCE) simulations from the Met Office-Natural Environment Research Council (NERC) Cloud model (MONC), employing physics-based loss functions that enforce both conservation laws and stability constraints.

The training dataset consisted of 3-D diagnostics data from the RCE simulations, with a 64x64 km² domain and 1 km grid spacing in the horizontal. While the original simulations had 99 vertical levels with varying vertical resolution, the DL model was trained on random slices (vertical levels) chosen from the original data volume. The inputs consisted of the resolved state variables like velocity components (u, v, w) from the previous time step, the perturbations to potential temperature, mixing ratios and Richardson number, whereas the targets for the DL model were the SGS tendencies of the model prognostic fields resulting from the Smagorisnky parameterization and the coefficients of viscosity and diffusion.

The DL model's cross-regime applicability was evaluated through multiple independent test cases: (200-second sampling frequency) and different atmospheric conditions from the Atmospheric Radiation Measurement (ARM) program. The simulations from ARM atmospheric settings were mainly targeted at simulating shallow convection, with different grid/domain configurations. Results from the off-line tests demonstrate promising performance in predicting SGS and transport coefficients (viscosity and diffusion) across these varied conditions, particularly in maintaining physical consistency during regime transitions.

Our preliminary findings indicate that this enhanced multi-scale, physics-informed architecture can effectively learn SGS parameterizations from limited training data while maintaining physical fidelity across different atmospheric conditions and spatio-temporal resolutions. This approach demonstrates the potential for the development of high-fidelity, generalizable parameterizations for weather and climate models, suggesting a route forward for reducing the greater computational costs associated with more complex SGS parameterization schemes.

How to cite: Panda, S. K., Jones, T., Shahzad, M., Ellis, A.-L., and Lawrence, B.: Physics-Guided Deep Learning-Based Emulation of Subgrid-Scale Turbulence Parameterization for Atmospheric Large Eddy Simulations, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-13920, https://doi.org/10.5194/egusphere-egu25-13920, 2025.

Given our current computational resources, a state-of-the-art ocean model can achieve a grid resolution in the order of 10 km for realistic global simulations, meaning that small-scale convection and wind-driven mixing near the surface of the ocean with length scales of roughly 1 m cannot be explicitly resolved. However, these microturbulent processes play a fundamental role in setting the structure of the ocean stratification, govern air-sea fluxes exchange as well as tracer transport with the interior of the ocean. Therefore, we use parameterizations, models which approximate small-scale processes using large-scale variables, to represent their effects in climate simulations.

In this work, we propose NORi: a novel, physically-principled, and data-driven approach to parameterizing ocean vertical mixing using neural ordinary differential equations (NODEs). NORi uses neural ODEs (NO) to augment a simple eddy-diffusivity closure based on the local gradient Richardson number (Ri). The Ri-based diffusivity closure captures local convective- and shear-driven mixing, while the neural ODEs augment the base model with nonlocal entrainment fluxes due to convection using neural networks. NORi is designed for realistic seawater's nonlinear equation of state using TEOS-10 and explicitly represents temperature and salinity fluxes. When compared against high fidelity large-eddy simulations (LES) of different convective strengths, background stratifications, and shear conditions, NORi demonstrates excellent prediction and generalization capabilities. By design, NORi automatically satisfies tracer invariance and conservation. It also exhibits high numerical stability and accuracy owing to its online training paradigm, where neural networks are calibrated against time-integrated field variables of interest rather than on instantaneous, time-independent turbulent fluxes. When compared against other parameterizations, NORi produces deeper mixed layers which are in better agreement with the LES solution. NORi is implemented straightforwardly into Oceananigans.jl, the fastest ocean model to date without intermediate wrappers. This can be achieved owing to the cutting-edge paradigm of the Julia programming language as well as the simple, modern and flexible interface of Oceananigans.jl. Using large-scale simulations, we demonstrate that NORi is numerically stable for at least 100 years despite being trained with only a 2-day integration, is computationally efficient, and produces realistic fields which are comparable to existing parameterization.

How to cite: Lee, X. K., Ramadhan, A., Souza, A., Wagner, G., Silvestri, S., Marshall, J., and Ferrari, R.: NORi: A Novel, Physically-Principled Approach to Parameterization of Upper Ocean Turbulence using Neural Ordinary Differential Equations, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-14742, https://doi.org/10.5194/egusphere-egu25-14742, 2025.

The added value of machine learning for weather and climate applications is measurable through performance metrics, but explaining it remains challenging, particularly for large deep learning models. Inspired by climate model hierarchies, we propose that a full hierarchy of Pareto-optimal models, defined within an appropriately determined error-complexity plane, can guide model development and help understand the models' added value. We demonstrate the use of Pareto fronts in atmospheric physics through three sample applications, with hierarchies ranging from semi-empirical models with minimal parameters (simplest) to deep learning algorithms (most complex). First, in cloud cover parameterization, we find that neural networks identify nonlinear relationships between cloud cover and its thermodynamic environment, and assimilate previously neglected features such as vertical gradients in relative humidity that improve the representation of low cloud cover. This added value is condensed into a ten-parameter equation that rivals deep learning models. Second, we establish a machine learning model hierarchy for emulating shortwave radiative transfer, distilling the importance of bidirectional vertical connectivity for accurately representing absorption and scattering, especially for multiple cloud layers. Third, we emphasize the importance of convective organization information when modeling the relationship between tropical precipitation and its surrounding environment. We discuss the added value of temporal memory when high-resolution spatial information is unavailable, with implications for precipitation parameterization. Therefore, by comparing data-driven models directly with existing schemes using Pareto optimality, we promote process understanding by hierarchically unveiling system complexity, with the hope of improving the trustworthiness of machine learning models in atmospheric applications.

Preprint: https://arxiv.org/abs/2408.02161

How to cite: Beucler, T., Grundner, A., Shamekh, S., Ukkonen, P., Chantry, M., and Lagerquist, R.: Distilling Machine Learning's Added Value: Pareto Fronts in Atmospheric Applications, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-15200, https://doi.org/10.5194/egusphere-egu25-15200, 2025.

The training of machine learning models for weather and climate on multiple datasets, including local high-resolution reanalyses and level-1 observations, is one of the frontiers of the field. It promises to allow for models that are no longer constrained by the capabilities of equation-based models, as is currently still largely the case when training on global reanalyses. For example, level-1 observations contain the feedback from arbitrary scale processes and hence do not suffer from the closure problem of equation-based models. Training on observations might hence lead to machine learning-based Earth system models with reduced systematic biases, in particular for long-term climate projections. Local reanalyses are only available for a small set of regions, mainly over Europe and North America. Appropriate training might allow one to generalize the detailed process information in these to other regions or even globally. 

In this talk, we present results on the effective training with a combination of global and local reanalysis as well as level-1 observations. We consider different pre-training protocols to learn the correlations between datasets, which is critical to obtain a benefit through their combination. We use a forecasting task as baseline and study the effectiveness of different variants of masked-token modeling and more sophisticated approaches that exploit the latent space of the machine learning models. We also study different fine-tuning strategies to extract a best state estimate from multiple datasets and to generalize regional datasets globally. For this, we build on the extensive results on fine-tuning of large language models that have been developed in the last years. Our results aim to determine general principles which combination of datasets is beneficial. We also perform a detailed analysis of the physical consistency and physical process representation in the model output. Through this, we believe our work provides an important stepping stone for the next generation of machine learning-based models for weather and climate.

How to cite: Lessig, C.: Towards next generation machine learning-based Earth system models that exploit a wide range of datasets, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-16817, https://doi.org/10.5194/egusphere-egu25-16817, 2025.

Mesoscale atmospheric processes that are neither resolved nor parameterized in global climate models, such as slantwise convection, can have a significant impact on climate variability and change. An example of such mesoscale influence on the climate is shown by Wills et al. 2024,  who demonstrate that circulation responses to surface anomalies are increased through heat and momentum fluxes by mesoscale processes. To enable longer simulations in comprehensive global climate models that include information about subgrid mesoscale processes, a machine learning (ML) parameterization could be applied at relatively low computational cost. So far, such ML parameterizations have been primarily applied to idealized geographies (e.g., aquaplanets), and they have not targeted midlatitude mesoscale processes in particular.

In this work, we focus on midlatitude mesoscale processes over the Gulf Stream region, as simulated by variable resolution CESM2 simulations, which have 14-km resolution over the North Atlantic. Learning subgrid fluxes from this model allows a targeted parametrization of mesoscale processes leading to vertical fluxes, namely slantwise convection and frontogenesis. We use an artificial neural network to predict vertical profiles of subgrid fluxes of momentum, heat and moisture. The features (inputs) for the ML models in this work include coarse-grained atmospheric state variables at each grid point, such as the vertical profiles of horizontal winds, temperature and their horizontal shear as well as surface pressure. The vertical profile of the specific humidity and the value of convective available potential energy are included to assess the importance of moist dynamics in the determination of subgrid convectional fluxes. Our results show that moisture variables have a rather small impact, suggesting that the subgrid fluxes can be explained by dry dynamics. A greater importance is found in the horizontal differences of neighbouring momentum and temperature columns. This suggests that neighbouring column information may be essential in the prediction of subgrid-scale fluxes, e.g., through the action of shear instabilities or conditional symmetric instability. Combined with information about the vertical localization relationship of the inputs and outputs, the goal is to feed this information into an equation discovery approach, which could lead to deeper physical understanding of mesoscale momentum and energy fluxes in midlatitudes.

How to cite: Ismaili, E., Beucler, T., and Jnglin Wills, R.: Prediction and Understanding of Subgrid-Scale Vertical Fluxes by Missing Midlatitude Mesoscale Processes Using a Machine Learning Approach, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-17078, https://doi.org/10.5194/egusphere-egu25-17078, 2025.

Accurate oil spill predictions are crucial for mitigating environmental and socioeconomic impacts. Numerical models, like MEDSLIK-II (De Dominicis et al., 2013), simulate oil advection, dispersion, and transformation, but their performance depends heavily on the configuration of physical parameters, often requiring labor-intensive manual tuning based on expert judgment.

To address this limitation, we integrate MEDSLIK-II with a Bayesian Optimization (BO) framework to systematically identify the optimal parameter configuration, ensuring simulations closely match observed spatiotemporal oil spill distributions. Our optimization focuses on horizontal diffusivity and drift factor parameters, using the Fraction Skill Score as the objective metric to maximize, thus reducing the overlap between simulations and observations.

The approach is validated on the 2021 Baniyas (Syria) oil spill, demonstrating improved accuracy, reduced biases and lower computational costs compared to the standalone numerical model.

By integrating BO with the MEDSLIK-II numerical model, our method enhances oil spill prediction capabilities and provides a transferable, physically consistent optimization framework applicable to a wide range of geophysical challenges.

This work is conducted within the framework of the iMagine European project, which leverages Artificial Intelligence, including AI-assisted image generation, to advance a series of use cases in marine and oceanographic science.

References

De Dominicis, M., Pinardi, N., Zodiatis, G., & Lardner, R. (2013). MEDSLIK-II, a Lagrangian marine surface oil spill model for short-term forecasting – Part 1: Theory. Geoscientific Model Development, 6, 1851–1869. https://doi.org/10.5194/gmd-6-1851-2013

How to cite: De Carlo, M. M., Accarino, G., Atake, I., Elia, D., Epicoco, I., and Coppini, G.: Improving Oil Spill Numerical Simulations through Bayesian Optimization, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-17333, https://doi.org/10.5194/egusphere-egu25-17333, 2025.

Vertical energy transport from the land surface into the atmosphere in the form of sensible and latent heat flux must be well represented in numerical weather prediction models to allow accurate estimates of near-surface atmospheric variables. Traditionally, these heat fluxes are parameterized relying on Monin-Obukhov Similarity Theory (MOST), which is based on differences in wind speed, air temperature, and humidity between adjacent measurement or model levels. Recently, Wulfmeyer et al. (2024) estimated heat flux with machine learning at much higher accuracy compared to MOST. Their ML model proposed the incorporation of additional predictor variables when estimating latent heat flux (such as solar radiation), which stands in contrast to the classical MOST approach. However, the analysis in Wulfmeyer et al. (2024) is based on a rather short data period in August 2017 at three nearby locations in Oklahoma, USA, which limits the generalizability of the results. Here, we replicate and expand the findings from Wulfmeyer et al. (2024) on a dataset from the boundary layer field site (GM) Falkenberg of the German Meteorological Service over a period of twelve years, covering various seasons and synoptic weather situations. Our findings support the role of incoming shortwave radiation not only for latent but also for sensible heat flux estimates, particularly for other parts of the year. The results thus underline the potential to develop more advanced flux parameterizations beyond MOST. In future research, we intend to investigate the role of other predictor variables, such as vapor pressure deficit or soil moisture, to assess the generalizability of the relations, to judge their performance under extreme conditions, and to derive simple but universally applicable parameterizations.

How to cite: Butz, M. V., Karlbauer, M., Beyrich, F., and Wulfmeyer, V.: Replicating Sensible and Latent Heat Flux Diagnosis with Multilayer Perceptrons on Multi-Year Falkenberg Tower Data , EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-17542, https://doi.org/10.5194/egusphere-egu25-17542, 2025.

Vegetation plays a crucial role in the Earth's climate system via a number of key processes, affecting the exchange of carbon, water and energy between surface and atmosphere. The complex relationship between climate change and vegetation highlights the importance of accurate and reliable vegetation models that fully capture these interactions. 

Traditional vegetation models are primarily designed to operate on CPU architectures, which restricts their ability to exploit advancements in modern parallel computing architectures such as GPUs. Furthermore, limited process knowledge and the absence of direct observations and/or quantitative theories for certain processes hinder accurate representation of these processes, which introduces uncertainties in the model results, leading to discrepancies when compared to observations. The rigid structure of these traditional models also makes integration of new processes challenging and hinders the application of advanced optimization techniques for automatic parameter tuning and objective calibration using abundant observational data due to their non-differentiable nature.

This work follows a new paradigm in vegetation modeling that integrates the robustness of traditional models with the adaptive power of machine learning techniques. The goal is to combine reliable physical components with machine learning components. As opposed to classical vegetation models, the resulting hybrid model is differentiable and the parameters of both the physical and the neural network components can be optimized jointly and efficiently using observational data.

In the proposed hybrid vegetation model, machine learning can be used to improve the computational efficiency of the model by emulating computationally expensive routines. We have implemented the key processes related to photosynthesis in LPJ in Julia. This minimal model setup is used to explore the potential of machine learning to replace the computationally expensive root-finding algorithm used in computing the optimal ratio of intercellular to ambient CO2 concentration, and hence stomatal conductance.

Machine learning can also be used for better process representation. The recently developed neural or universal differential equations offer a particularly promising methodological framework for learning the dynamics of carbon allocation to different vegetation pools using observations. The dynamics of carbon allocation to different plant components can be effectively modeled using a neural ODE approach, which utilizes observations of observable variables (e.g., Above Ground Biomass (AGB), Leaf Area Index (LAI)) to learn the dynamics of unobservable variables such as vegetation carbon pools.

How to cite: Badri, M., Hess, P., Lin, Y., Bathiany, S., Gelbrecht, M., and Boers, N.: Towards a Hybrid Vegetation Model, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-17958, https://doi.org/10.5194/egusphere-egu25-17958, 2025.

Modelling microplastic transport through porous media, such as soils and aquifers, is an emerging research topic, where existing hydrogeological models for (reactive) solute and colloid transport have shown limited effectiveness thus far. This perspective article draws upon recent literature to provide a brief overview of key microplastic transport processes, with emphases on less well-understood processes, to propose potential research directions for efficiently modeling microplastic transport through the porous environment. Microplastics are particulate matter with distinct physicochemical properties. Biogeochemical processes and physical interactions with the surrounding environment cause microplastic properties such as material density, geometry, chemical composition, and DLVO interaction parameters to change dynamically, through complex webs of interactions and feedbacks that dynamically affect transport behavior. Furthermore, microplastic material densities, which cluster around that of water, distinguish microplastics from other colloids, with impactful consequences that are often underappreciated. For example, (near-)neutral material densities cause microplastic transport behavior to be highly sensitive to spatio-temporally varying environmental conditions. The dynamic nature of microplastic properties implies that at environmentally relevant large spatio-temporal scales, the complex transport behavior may be effectively intractable to direct physical modeling. Therefore, efficient modeling may require integrating reduced-complexity physics-constrained models, with stochastic or statistical analyses, supported by extensive environmental data. This is a sub-project (focusing on microplastics in the environment) of the Digital Waters Flagship funded by the Research Council of Finland, where we aim to create a digital ecosystem for machine learning aided hydrological modelling of various hydrosphere processes across all environmental compartments, focusing particularly on the critical zone.

How to cite: Tang, D. and Yang, X.: Modeling microplastic transport through porous media: challenges arising from dynamic transport behavior, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-18139, https://doi.org/10.5194/egusphere-egu25-18139, 2025.

Permeability should be dispersed conveniently to control the aquifer's type and quality. Permeability in a variety of porous media can be determined using different methods depending on the environment and the scope of the porosity media. These days, permeability of core samples and well logging data with greater aquifer heterogeneity, artificial intelligence algorithms are well-known for estimating permeability. Machine learning and artificial intelligence have gained popularity and credibility across all scientific fields. To address the dearth of resources in geosciences generally and hydrology specifically.

As soft computing techniques, Artificial Neural Networks (ANNs) have demonstrated the capacity to estimate acceptable outputs with tolerable outcomes. The ANN model uses basic processing units, which are networks of interconnected neurons. The simplest approach is the Feed-Forward Artificial Neural Network (FF-ANN). The Middle Jurassic Hugin Formation may have been deposited as a mouth bar setting during the period of general transgression, as evidenced by fluctuating permeability values brought on by changes in the sediment supply, which varying porosity values brought on by variations in the amount of clay and size of grains.

Keywords: Artificial Neural Network, Feed-Forward Artificial Neural Network, Volve oilfield, Hugin Formation, Permeability estimation.

How to cite: Ghattas, K. and Buday, T.: Artificial Neural Network Approaches for Permeability Estimation, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-18810, https://doi.org/10.5194/egusphere-egu25-18810, 2025.

Machine learning (ML) has the potential to reduce systematic uncertainties in Earth System Models by replacing or complementing existing physics-based parameterizations of sub-grid processes. However, after decades of research, ensuring generalization and stability of ML-based parameterizations remains a major challenge.  We aim to minimize both epistemic and aleatoric sources of uncertainty via physically inspired, vertically recurrent neural networks (RNN) which offer key benefits such as parametric sparsity and efficient modeling of non-locality in a column. To address aleatoric uncertainty, we furthermore incorporate stochasticity and convective memory into the ML architecture. We present preliminary results using the ClimSim framework, where the physically inspired ML framework replaces a superparameterization in a low-resolution climate model.

How to cite: Ukkonen, P., Mansfield, L., and Christensen, H.: Emulation of sub-grid physics using stochastic, vertically recurrent neural networks, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-19319, https://doi.org/10.5194/egusphere-egu25-19319, 2025.

How to cite: Barre, F., Bouman, E., Hertwich, E., and Moran, D.: High resolution economic modelling for climate risk assessments: An application to coastal storm surges in Norway, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-269, https://doi.org/10.5194/egusphere-egu25-269, 2025.

Understanding the spatiotemporal variations and driving forces of groundwater-dependent ecosystems (GDEs) resilience can provide scientific evidence for GDEs protection under natural and anthropogenic perturbations. However, the differences in the spatiotemporal variations of GDEs and non-GDEs resilience and their responses to climatic and anthropogenic disturbances are still unclear. Here, we applied lag-1 month temporal autocorrelation (AR(1)) based on kernel Normalized Difference Vegetation Index (kNDVI) to explore the spatiotemporal pattern of GDEs and non-GDEs resilience, and used propensity score matching (PSM) to identify the difference. XGBoost and Shapley model are applied to spatially quantify the marginal contributions from each single drivers. We found that over a third of both GDEs and non-GDEs experienced a shift from an increasing to a declining resilience trend from 1982 to 2022, with the resilience decline in GDEs being 7.5% slower than in non-GDEs. GDEs resilience are mostly responsive to precipitation and VPD, while non-GDEs resilience are mostly responsive to temperature and PET variations. Plant biodiversity significantly boosts GDEs resilience, which has a different impacting threshold compared with non-GDEs resilience. The impact of stocking density on resilience is much higher in GDEs than in non-GDEs. These findings highlight the urgent need for policy interventions to protect and manage groundwater and plant biodiversity in GDEs to maintain its resilience.

How to cite: Wu, T. and Liu, Y.: The resilience of global groundwater dependent ecosystems (GDEs) declined less than non-GDEs in the last forty years under differentiated spatial driving forces, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-1845, https://doi.org/10.5194/egusphere-egu25-1845, 2025.

Nature degradation directly impacts company portfolio performance by disrupting ecosystem services critical to operations (such as water availability, pollination, etc.), while on the other hand company activities (such as deforestation, pollution, and resource extraction) significantly contribute to nature's degradation. This reciprocal relationship has intensified the need for robust methodologies to assess the underling nature-related financial risks as a pathway to allow companies to engage in mitigation and adaptation activities. While various approaches are currently in use, both business-focused (e.g nature value at risk) and nature-centered (e.g earth system index), significant gaps remain in harmonised methodologies that are comprehensive, user-friendly, and replicable, especially within biodiversity and ecosystems services at company level. An examination of existing approaches and contextual applications to company or industry archetype reveals both advantages and limitations in representing the double-materiality of risk associated with businesses. Here, we explore a potential framework for companies at the sector archetype that can assist in assessing current and potential nature-related financial risks. This is done by integrating NRFR metrics multidimensionally from the perspective of nature-related dependencies, exposures, and pressures across high climate impact industry archetypes, including agriculture, energy, and the built environment. However, the inherent challenges in representing complex and adaptive systems like nature through a metric approach should be held with caution. Regardless, this approach offers optimal direction on which companies can adopt for their individual NRFR assessments.

Keywords: Risk Assessment, Nature degradation, nature-related financial risks, business

How to cite: Nyasulu, M. K., Sheikh, H., Christiaen, C., Lockwood, P., Jean-Pierre, J.-P., Wassenius, E., and Quek, C.: Nature-Related Financial Risks Assessments framework for Company and Industry Archetype: Metrics Review, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-4495, https://doi.org/10.5194/egusphere-egu25-4495, 2025.

How to cite: Bansod, S., Verschuur, J., and Comes, T.: High-frequency survey data reveal complex impact pathways of climate-related disasters on household welfare in Malawi , EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-5930, https://doi.org/10.5194/egusphere-egu25-5930, 2025.

There is mounting evidence that we are living in a time of turbulence, with many disruptive events becoming increasingly common and intense. In these times, understanding the dynamics of past shocks can help us better prepare and potentially prevent severe impacts in the future. The Shock Tracker is a living database of cases, encompassing everything from wildfires, to floods, disease outbreaks and conflict. The case studies are formulated as storylines, describing the event and its multiple drivers and impacts, through a standardized reporting protocol. The database currently has over 100 documented cases and, thanks to its living nature, it is growing every day. The cases are submitted by people from diverse backgrounds who become part of our growing Shock Tracker Network. All cases then undergo rigorous review before being added to the final archive. There is a particular focus in the protocol on how the shock was shaped by the interactions between people and nature. Through the case studies, the Shock Tracker highlights how anthropogenic climate change contributed to shock events, where mild but multiple drivers led to extreme impacts, and what the role of human action and agency were in both driving and mitigating these events and their impacts. The Shock Tracker is therefore a collection of cases that show that climate-induced events are already happening and are not only a future problem, that they are not only caused by extreme conditions, and that they are not only natural but often triggered by social decisions. We hope that the Shock Tracker can be a source of both direct learning from past events to better prepare us for the future and a useful resource for academic research into the patterns of drivers and impacts of shocks.

How to cite: Wassénius, E. and Rubin, G.: Shock Tracker: A living database of shocks, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-6470, https://doi.org/10.5194/egusphere-egu25-6470, 2025.

Between 1998 and 2017, tropical cyclones (TCs) caused 233,000 deaths, affected approximately 726 million people globally, and led to an average of 9.3 million human displacements annually between 2017 and 2020 (Kam et al., 2024). Within countries impacted by TCs, economically disadvantaged populations are disproportionately affected (Jing et al., 2024). Adaptation to TCs is impregnated with uncertainty within a global context where coastal adaptation efforts are unbalance distributed (Magnan et al., 2023).

While adaptive capacity varies widely at subnational levels (Magnan et al., 2023), adaptive information is provided at the national level or, at best, at the second administration level (e.g., states). There is a lack of local adaptation information specifically related to TCs. As part of the TRANSCLIMA project (https://transclima.ihcantabria.com/), we developed global, local risk indicators at the fourth administration level, based on changes in TC characteristics, exposed population, and TC-related adaptive capacity.

This study identifies TC regions where changes in intensity and frequency are observed. Based on these changes, regions where minor or major TCs shift and assesses whether TCs may become an unprecedented hazard, leading to emergent risks. These hazard indicators resulted from analysing TC characteristics under two climatological periods: a baseline climate (1980-2017) and a future high-emission climate scenario, Shared Socioeconomic Pathway SSP8.5 (2015-2050). We used synthetic tracks datasets of four Global Climate Models (CMCC, CNRM, EC-Earth, and HadGEM3)  (Bloemendaal, et al., 2022). Population data were obtained from the fourth version of the gridded population of the world with a 1 km resolution of the Socioeconomic Data and Application Center for the base year 2000 and for the years 2040 and 2050 under the SSP5 scenario (Center For International Earth Science Information Network-CIESIN-Columbia University, 2017), calculated for each coastal locality (Odériz et al., 2024). We assessed the adaptive capacity of each TC region using an index that combines local adaptive capacity, such as indicator local experience based on IBTrACS data (Knapp, 2018; K. R. Knapp et al., 2010). Additionally, we proposed a national-level insurance coverage indicator and a national-level adaptation readiness indicator.

Using this global, local-resolution risk assessment, we provided a detailed overview of the adaptation status of countries, considering subnational levels, that can be used to identify hotspots for financial adaptation plans.

How to cite: Odériz, I. and Losada, I.: Local-resolution risk assessment for tropical cyclones: toward global adaptation , EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-6917, https://doi.org/10.5194/egusphere-egu25-6917, 2025.

With a recent green climbdown in global finance including the world’s largest money manager BlackRock leaving the high-profile Net Zero Asset Managers group, the debate on the financial risks from climate change is reignited. While there is an agreement on the catastrophic potential of unmitigated climate change itself, many important players in the global financial system have reevaluated to what degree climate risk equals investment risk. A key point in this debate is who owns the financial risks from climate change and who is subsequently responsible for managing them. This study looks into the question of climate change risk ownership by examining two megatrends unfolding in the global financial system over the last three decades: the financialization of climate change and the parallel evolution of the asset management industry. The analysis shows how the interconnection between these two trends has resulted in a financial system where climate risk disclosure demanded by newly introduced regulations such as TCFD and intended to enhance risk management, actually enables private entities to shield profits from climate-related losses, while leaving systemic risks unaddressed. Drawing on the literature from the financialization of nature, risk ownership, and climate risk assessment, the study highlights how technological advancements in climate risk models and government incentives for low-carbon investments create adverse selection and moral hazards for both physical and transition risks from climate change. Introducing the concept of ‘climate risk ownership’ through case studies on renewable energy investments and disaster insurance, the study highlights the gaps in the management of the financial risks of climate change between public and private entities.

How to cite: Rözer, V.: Climate risk ownership in the age of asset management, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-7026, https://doi.org/10.5194/egusphere-egu25-7026, 2025.

There have been rapid developments in mandatory Environment, Social, and Governance (ESG) information disclosure in recent years. Under the requirements that have been developed, companies, including financial institutions, are required to analyse not only the climate risk they face but also the risk from biodiversity loss and ecosystem degradation, and their interactions with climate risk. For EU nations and other European nations including Norway, companies are subject to the Corporate Sustainability Reporting Directive (CSRD) requirements that came into effect last year and require such disclosures. Many financial institutions lack the expertise to sufficiently manage these requirements and need support not only to satisfy the requirements, but also to proactively manage their assets against the risks they face.

Under such regulatory developments, we have been working with financial institutions in Norway to support their ESG information disclosure activities, focusing on physical climate risk, nature risk, and their potential connections. In this talk, we highlight our approach, focusing on the opportunities we see for both financial institutions and the scientific community. These will be presented through a case study we conducted, focusing on data flow and availability, methodologies developed on bridging climate and nature research, and on the limits we faced as academic researchers. The collaboration has led to the development of methods that could position the financial institutions as leaders in the sector regarding risk management and building resilience towards climate and nature risk. Given the disclosure requirements, transparent methods and coherent data generation on physical and transition risks will be an opportunity for enhancing awareness of climate and nature risk, and for getting a comprehensive picture of the economic impacts of climate change (including the impacts of extreme events) and the benefits of avoided impacts via mitigation and adaptation actions.

How to cite: Kunimitsu, T., Daloz, A. S., Kusch, E., and Sillmann, J.: Interactions of physical climate risk and nature risk for ESG information disclosure in the financial sector, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-10005, https://doi.org/10.5194/egusphere-egu25-10005, 2025.

Floods cause large disruptions to society by causing both direct and indirect damages. These impacts will be further exacerbated by climate change and socioeconomic development. In addition to direct impacts, businesses may face indirect losses resulting from disruptions to their operations, adding extra complexity to business risk assessments. Additionally, business closures can have far-stretching economic repercussions. Flood insurance is an instrument to reduce the impact of floods for businesses by spreading the risk over space and time. While the (future-) increase in flood damages puts pressure on businesses, insurance systems tailored to businesses remain underexplored.

This research applies and extends the ‘Dynamic Integrated Flood Insurance’ (DIFI) model to analyse flood insurance for businesses in the Netherlands, taking into account both insurance against direct damages and insurance against business interruption damages. We analyse the responses of various insurance systems to changes in flood risk. These systems include voluntary insurance, solidarity-based insurance, and public-private partnership insurance. In addition, we assess the effect of adaptation on the viability of flood insurance by allowing businesses to take building-level measures to reduce their flood risk.

To facilitate the insurance analysis, flood damages are estimated using an object-based approach that takes high resolution (25m x 25m) inundation maps as input. To simulate the insurance uptake, company-level financial data obtained from the Dutch Chamber of Commerce is used in a subjective expected utility framework. This module is calibrated on actual insurance uptake numbers and takes risk misperception into account. DIFI simulations until 2080 show how premiums, insurance uptake, and policyholder adaptation efforts develop over time for various insurance market structures. These projections provide valuable insights into the viability and effectiveness of different insurance market structures in the face of climate change and shifting socioeconomic conditions.

The novelty of this research lies not only in incorporating businesses into the insurance analysis, but also in introducing a focus on business interruption damages, offering a more comprehensive perspective on flood impacts for businesses. Initial results reveal that, in certain sectors, flood-related business interruption damages are nearly as high as, or even exceed, direct damages. These findings offer new insights into the impact of flooding on businesses and the challenges of insuring such damages.

Consequently, the findings are relevant for policymakers and insurers by identifying which insurance market structures are more resilient to the increasing flood risk, providing guidance on designing financially sustainable insurance framework. Moreover, the study highlights the need for targeted insurance incentives to encourage business-level adaptation, and it informs decisions regarding potential government involvement in the insurance system to ensure equitable access to flood insurance.

How to cite: Ingels, M., Botzen, W., Aerts, J., Brusselaers, J., and Tesselaar, M.: Business-level flood insurance coverage and adaptation under climate change in the Netherlands, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-11087, https://doi.org/10.5194/egusphere-egu25-11087, 2025.