Multi-scenario flood modeling in a mountain watershed using data from a NWP model, rain radar and rain gauges

The temporal and spatial distribution of precipitation is the key parameter for flood modelling. The study presents an evaluation of different meteorological data sources to assess their applicability and reliability for flood modeling. Apart from conventional rain gauge data, the information of the Numerical Weather Prediction Model SWISS MODEL (SM) and radar interpreted precipitation taken from the rain radar Fuerholzen, operated by the German Weather Service, has been available. They are used within the framework of an extended and GIS-structured TOPMODEL (Beven and Kirkby, 1979; Beven , 1994; Ludwig and Mauser, 2000), to perform model simulations and forecasts in the Ammer catchment for a hazardous flood event in 1999. The disaggregation and scaling of precipitation data, to meet the requirements of the hydrological model, is of specific interest. A variety of procedures to disaggregate NWP information for a hydrological application is presented, emphasizing the influence of the selected algorithm on the model result. Applying the SM and the rain radar data set, the calculated flood volume is overestimated within a range of 15 to 36%, while the rain gauge data set leads to an underestimated runoff volume of 13%. A sensitivity analysis shows a high variability in the spatial and temporal distribution of predicted and recorded precipitation and its consequent effect on the performance of the hydrological model. However, positive conclusions for future applications of a meteorological and hydrological model synergy can be drawn from the outcome of this study.