Abstract
The use of meteorological ensembles to produce sets of hydrological predictions has increased the ability to issue flood warnings. However, the spatial scale of a hydrological domain is still much finer than that of a meteorological model, and Numerical Weather Prediction (NWP) models have challenges with misplacement. This study assesses the pre-processing methods with consideration of appropriate ensemble members and a spatial shift of ensemble NWP rainfall fields, in order to improve the accuracy improvement of the ensemble flood forecasting. The analysis shows that appropriate ensemble members of NWP rainfall improves the accuracy of the mean value when ensemble forecasting a flood, whereas the transposition of NWP rainfall fields has a more suitable impact on the accuracy of the best values.
