English
The risks of river flooding are expected to increase due to climate change. The consequences of river flooding can be reduced through early-warning systems in such a way as to enable timely evacuation and implementation of mitigation measures. Hydrodynamic models provide accurate overviews of the areas at risk, but their long computational times mean they cannot be used for short-term forecasting. Therefore, this project develops a fast machine-learning approach producing real-time flood probability maps. Using the latest weather forecasts as input, this means that crisis teams can make informed decisions on evacuation strategies and mitigation measures during river flooding events.
