Flood forecast combining hydrological modelling and Machine
Boris Gratadoux presents an automatic flood forecast tool that combines two kinds of information used as input of a Machine Learning algorithm.
Thales has developed an automatic flood forecast tool combining two kinds of information used as input of a Machine Learning algorithm:
- - Forecasted flows at the outlet of the watershed of interest, obtained with an ensemble data assimilation using a particle filter;
- - Current soil state information from space-based observations;
The prototype allows forecasting time series of flood occurrence associated with a confidence index for a 6-day flood forecast with a 6-hour time-step. It has been tested and validated on a French watershed with good results.