Evaluation of precipitation in the ENSEMBLES regional climate models over the complex orography of Slovenia

This study aims to assess the capability of regional climate models (RCMs) to simulate precipitation climatology over the southeastern Alpine flanks, where complex orography prevails. Precipitation simulations from 14 different 40-year downscaling experiments (1961–2000) driven by the ERA-40 reanalysis from the ENSEMBLES project are analysed in terms of realistic reproduction of spatial and temporal seasonal precipitation patterns over Slovenia. The best RCM performance was found for winter when precipitation is mainly caused by large-scale processes. A significantly lower model performance was observed for summer when precipitation is mainly caused by convective processes. Indeed, the largest spread of simulated precipitation between ensemble members was observed in summer, which is likely due to differences in the simulation of convective processes by the models. In addition, the simulations were analysed using partial least-square regression to quantify dependence of the simulated precipitation on physiographic factors. The applied regression model showed that orography and longitude are the most important variables associated with the spatial variability of precipitation in Slovenia. The empirical orthogonal functions were used to analyse the spatial patterns of inter-annual precipitation variability. The analysis showed that all models reproduced the most dominant mode of variability, but significant differences occurred between the models, especially in summer. In general, a thorough analysis has shown that the quality of simulations is highly variable among different regions as well as for different seasons.