Seasonal forecasting of water resources in the Blue Nile Basin.

At the recent European Geosciences Union (EGU) General Assembly 2025 in Vienna, SPS Blue Nile team (from KIT, University of Potsdam, partner institutions from Ethiopia and Sudan, and Berlin-based company Adelphi) presented their research.

PhD student Rebecca Wiegels gave a talk on Seasonal AFNOCast, a deep learning method using an Adapted Fourier Neural Operator (AFNO) -based model to bias-correct and downscale seasonal forecasts. The AI-based system showed strong performance, offering high-resolution, skillful predictions crucial for sectors like agriculture and water management through applications as shown by SPS Blue Nile colleague (“Seasonal AFNOCast: A Deep Learning Approach for Enhanced Regional Seasonal Predictions”, https://doi.org/10.5194/egusphere-egu25-4729).

Prof. Dr. Axel Bronstert and Dr. Morteza Zargar presented a poster on seasonal hydrological and sediment modelling and forecasting, highlighting the use of the WASA-SED model in combination with ECMWF-SEAS5 climate data to improve water resource planning in the Upper Blue Nile Basin—home to the Grand Ethiopian Renaissance Dam. Their work emphasizes the importance of evaluating rainfall datasets for reliable hydrological modeling and supports long-term, locally grounded water management strategies (“Seasonal Streamflow and Sediment Forecast in the Upper Blue Nile Basin, Ethiopia”, https://doi.org/10.5194/egusphere-egu25-6789). It further includes the sediment seasonal forecasting of water and sediment discharges in the Blue Nile system.

PhD student Yasir Hageltom introduced a sub-seasonal crop yield forecasting framework combining WRF model with process-based crop modelling (WRF-NoahMP-Crop). The research aims to improve early warning capabilities for farming decisions in a region increasingly affected by climate variability and rising food demand (https://doi.org/10.5194/egusphere-egu25-11319).

The novel techniques captured the attention of researchers which resulted in great discussions.