Suche
Mein Konto
AI against floods: New flood forecast model for Germany!
Researchers at KIT are using AI to develop a flood prediction model for small river basins in Germany, funded with 1.8 million euros.
AI against floods: New flood forecast model for Germany!
On March 9, 2025, researchers from the Karlsruhe Institute of Technology (KIT) via an innovative flood prediction model based on artificial intelligence (AI). The project, called KI-HopE-De, aims to significantly improve the prediction of flood events in small river basins in Germany. These areas, which cover an area of approximately 5 to 500 square kilometers, are particularly vulnerable to sudden flood events due to their structure.
The research is funded by the Federal Ministry of Education and Research with 1.8 million euros. In current practice, the federal states mainly publish regional warning levels, which affects the interpretation and ability to react in the event of a flood. This will change in the future with the new national, probabilistic flood forecast model.
Interdisciplinary collaboration for better predictions
As part of the project, a comprehensive hydro-meteorological data set will be created that will be accessible to the public. This data set consists of measurement and prediction data not only from KIT, but also from German Weather Service (DWD) and various state environmental offices. Dr. Ralf Loritz from KIT emphasizes that modern AI methods are able to recognize complex relationships in weather and water level data, which makes the forecasts significantly more precise.
The project is interdisciplinary and brings together experts from the fields of hydrology, meteorology and machine learning. Partners like that State Office for Nature, Environment and Consumer Protection North Rhine-Westphalia (LANUV NRW) as well as that Rhineland-Palatinate State Office for the Environment (LfU RP) work closely with the researchers to ensure that the models developed meet the requirements of future users, mainly state authorities.
Innovative approaches to improving warning systems
Current warning systems for floods in Germany are often inaccurate and are limited to regional warnings. This method is not effective enough considering flood events that often occur unexpectedly. The KI-HopE-De program aims to enable forecasts up to 48 hours in advance in order to sustainably improve public safety and flood protection.
The close collaboration between science and practice is central to the success of the project. Professor Peter Knippertz from KIT emphasizes the creation of an innovative platform for all flood forecast centers in Germany. The research activities at KIT are part of a larger effort to not only increase technological possibilities, but also to raise awareness of flood protection measures among the population.
Overall, the KI-HopE-De project represents a significant step in the use of modern technology to predict future flood events more precisely and thus make a valuable contribution to security in Germany.