MLRG Joined NASA SERVIR Project to Advance Flash Flood Forecasting in West Africa

The Machine Learning Research Group was pleased to join a new NASA-funded project under the SERVIR program: “Machine learning based flash flood forecasting in West Africa with satellite observations” (NASA Grant No. 80NSSC23K0500).

The project is led by Efthymios Nikolopoulos (now, at Rutgers University) and brings together a multi-institutional team with expertise in hydrology, hydrometeorology, remote sensing, operational flood forecasting, stakeholder engagement, and machine learning. The team includes Georgios C. Anagnostopoulos from Florida Institute of Technology, Abdou Ali from AGRHYMET, Jonathan J. Gourley from NOAA’s National Severe Storms Laboratory, Viviana Maggioni from George Mason University, Humberto J. Vergara Arrieta (now, at University of Iowa), and William Amponsah from Kwame Nkrumah University of Science and Technology.

The project aims to develop a flash flood nowcasting and forecasting system for West Africa by combining satellite observations, weather forecasts, machine-learning-based precipitation nowcasting, and distributed hydrologic modeling. The system is designed to support flood prediction across a range of spatial scales, from small urban catchments to larger basins, with Ghana serving as the pilot country for testing and demonstration.

MLRG will contribute to the project’s machine-learning component, focusing on precipitation nowcasting from satellite-based observations. This work aims to improve short-term estimates of precipitation fields that can be used as inputs to hydrologic models, helping extend the lead time and usefulness of flash flood forecasts.

Beyond its technical goals, the project is strongly aligned with SERVIR’s mission of connecting Earth observations with decision support and with the needs of the SERVIR West Africa hub. Through co-development with AGRHYMET and regional partners, the team aims to strengthen institutional capacity for flash flood forecasting and support more timely, actionable information for agencies and communities facing flood risk in West Africa.