Simulation of floods caused by intense rain events with the water balance model LARSIM

The distributed water balance model LARSIM has been successfully used for real-time forecasting of fluvial floods for two decades. In order to improve the simulation of infiltration excess, Horton Overland Flow (HOF) and (local) pluvial floods due to intense rainfall, LARSIM was amended by the infiltration module of the research model RoGeR. The newly introduced infiltration module accounts for matrix infiltration as well as infiltration via macropores and shrinkage cracks by applying the physically based Green-Ampt approximation. First, we derived an approach to parameterize the infiltration module with a high spatial resolution on the basis of a digital soil map and land use data for the federal state of Baden-Wurttemberg. Secondly, this approach was transferred to regions in North-Rhine Westphalia and Luxembourg using different soil data. The infiltration module and the parametrization approaches were evaluated using seven intense rain events in five different mesoscale catchments in Baden-Wurttemberg, North-Rhine Westphalia and Luxembourg. Introducing the infiltration module improved simulation results for all seven mainly HOF dominated events. When using sufficiently accurate rain input, the measured discharges of the events (volume and peak) could generally be reproduced well with the model. These findings suggest that the infiltration module as well as the parametrization approach are broadly valid, transferable and robust. The successful model improvement lays an important foundation for a possible future real-time forecasting of pluvial floods. However, our investigations also showed that the uncertainty of radar based rain may typically cause under- or overestimations of the resulting discharge by a factor of up to two. Moreover, LARSIM should be further improved with respect to simulating the simultaneous occurrence of HOF and saturation driven runoff. Consequently, besides further refining LARSIM, the quantitative estimation and forecasting of precipitation with weather radar also needs to be further improved in order to enhance the simulation and prediction of floods caused by intense rain events.