Validating Precipitation Estimates Simulated by the Numerical Weather Forecast Models Eta and Mm5 for Use in Flood Forecasting

Accurate rainfall forecasts using numerical weather prediction (NWP) models can provide significant increases in forecast lead-time and thereby give greater opportunity for flood warning and protection measures. One goal of the FLOODRELIEF project is to quantify the uncertainty of forecast precipitation fields from two different NWP models (Eta and MM5) to investigate the impact of precipitation results of NWP models coupled to hydrological models. The different model resolutions in Eta and MM5 models have been tested for the period of July 2002 to September 2002. The grid based hydrological model SEROS was set up for the Odra catchment. It includes a one-dimensional vertical land-surface model and a horizontal routing scheme. Two methods of dynamic coupling of an atmospheric mesoscale model to a hydrological model were tested. In the first case the MM5 model including the Noah land surface scheme was directly coupled to a hydrological routing scheme. In the second case the atmospheric model was coupled to the hydrological model including the land surface model SEWAB and a routing scheme. The best results were achieved with highest model resolution and the coupling to SEROS.