The role of regional climate model setup in simulating two extreme precipitation events in the European Alpine region

In this study we have investigated the role of domain settings and model's physics in simulating two extreme precipitation events. Four regional climate models, all driven with a re-analysis dataset were used to create an ensemble of 61 high-resolution simulations by varying physical parameterization schemes, domain sizes, nudging and nesting techniques. The two discussed events are three-day time slices taken from approximately 15-months long climate simulations. The results show that dynamical downscaling significantly improves the spatial characteristics such as correlation, variability as well as location and intensity of maximum precipitation. Spatial variability, which is underestimated by most of the simulations can be improved by choosing suitable vertical resolution, convective and microphysics scheme. The results further suggest that for studies focusing on extreme precipitation events relatively small domains or nudging could be advantageous. However, a final conclusion on this issue would be premature, since only two extreme precipitation events are considered.