KI-gestützte Optimierung der Vorhersage und Steuerung von Hochwasserereignissen
Start Date: 01.10.2025 | End Date: 30.09.2028
The RIVERCAST project aims to develop various methods for improving flood protection planning and operations, and to implement these in practice within the southern Erft catchment area. The project’s focus is closely aligned with the 10-point action plan of the MULNV (now MUNV) and addresses known shortcomings in flood management and flood protection planning. A central element is the use of modern AI models to address the issues outlined. These models utilise data from existing and additional measuring stations to be installed, and integrate further available information (radar-based precipitation, precipitation forecasts, conductivity, soil moisture, etc.).
In this way, the aim is, on the one hand, to improve inflow forecasts for existing flood retention basins. To our knowledge, the use of AI-based forecasts to supplement precipitation-runoff-based discharge forecasts is a novel approach and may, in the long term, also contribute to the improvement of a nationwide flood forecasting system.