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 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.

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.

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.

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.

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.

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.

An accurate description of air-sea interaction in atmospheric, ocean and coupled models remains problematic due to unresolved processes in atmospheric models. Systematic differences in winds occur (amongst others) due to undetermined geophysical dependencies. Systematic model errors in ocean winds found on large scale and atmospheric mesoscale propagate to the ocean circulation when used to force ocean models and affect coupled earth system dynamics. 
Geolocated scatterometer-based corrections of wind forcing products already successfully correct for local wind vector biases, but this correction method is highly dependent on sampling. The growing virtual scatterometer constellation is very promising to better capture the detailed forcing errors over the day.  Biases of the order of 0.5m/s in wind speed can introduce a large bias in wind stress, causing significant errors in ocean–atmosphere coupling and climate prediction.
Our focus is on unresolved processes in atmospheric Numerical Weather Prediction (NWP) models, namely systematic errors in boundary layer parameterizations such as lack of ocean currents and/or other biases that persist over time. An improved representation of surface turbulent fluxes relies on better estimates of: the roughness length, the stability function, the sea skin temperature, ocean currents and convective gustiness. 
The goal is to apply model bias reduction schemes with respect to scatterometer winds. Consistent scatterometer corrections will lead to an improved understanding of the coupled atmospheric and oceanic model dynamical processes in the evolving earth system. In addition, corrected model winds reduce errors in ocean forcing and will be helpful in scatterometer data assimilation.

How to cite: Biri, S. and Stoffelen, A.: Coupling Atmospheric Dynamics and Ocean with Winds from Satellites, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-6053, https://doi.org/10.5194/egusphere-egu25-6053, 2025.