Enhancing Global Flood and Drought Forecasting with SEED-FD: Integrating Remote Sensing for Hydrological Insights

Floods and droughts are among the most destructive hydrological extremes, creating severe socio-economic disruptions worldwide. Many regions, especially those in the Global South, remain highly vulnerable due to inadequate forecasting precision caused by sparse observational networks and limited model capabilities. The European Commission-funded SEED-FD (Strengthening Extreme Events Detection for Floods and Droughts) project under Horizon Europe aims to address these gaps by leveraging advanced Earth observation (EO) and non-EO datasets to strengthen forecasting systems for floods and droughts.

The primary objective of SEED-FD is to enhance the accuracy and global usability of the Copernicus Emergency Management Service (CEMS) Early Warning Systems (EWS). This involves refining key elements of the CEMS hydrological forecasting framework, including the LISFLOOD model’s hydrological processes and calibration strategies, integrating innovative machine learning and data assimilation techniques to improve predictions, and creating new global forecast products. A key focus is on incorporating nontraditional observational data, such as precipitation, soil moisture, and streamflow measurements from EO sources, as well as river discharge data obtained from microstations.

The project adopts a two-step strategy: initial algorithm and method validation in data-rich regions (Danube and Bhima basins) to establish proof of concept, followed by scaling and application in three diverse and vulnerable regions—the Paraná River Basin (Brazil), the Niger River Basin (West Africa), and the Juba-Shebelle Basin (Horn of Africa).

This presentation will cover mid-term findings from SEED-FD, emphasizing progress in hydrological model calibration, improved process representation, data assimilation, and machine learning-based post-processing. These advancements have demonstrated enhanced prediction reliability in the Danube and Bhima basins and offer valuable lessons for scaling solutions to other vulnerable regions.