GCM Driven COSMO-CLM Post-Processed Precipitation Over Italy: Control and Future Scenarios

Abstract

In Turco et al. (2013a, [19]), three MOS methods (Linear-Scaling, Quantile-Mapping and MOS Analog) have been proposed and they have shown skill to downscale daily precipitation outputs of the ERA40-driven COSMO-CLM model over three Italian domain (Orvieto, Po river basin, Sardinia). This study extend this previous analysis [19] applying these MOS techniques to the entire model chain (CMCC-CM driven COSMO-CLM model) in control and future periods. These methods clearly outperform the uncalibrated RCM outputs, regardless of the region, the season and precipitation index. Generally, better results have been obtained with the quantile mapping method. Finally, the application of these MOS techniques to future RCM scenarios is also discussed. Generally the projected scenarios are quite consistent among the dynamical model and the MOS methods, giving some confidence in the robustness of the MOS in changing climate conditions. The stronger climate change signal that appears, regardless of the scenario (RCP4.5 or RCP8.5), domain, downscaling method or precipitation indices, is the decreasing trend in summer. This report presents the preliminary results of combined downscaling methods for precipitation, adopted by CMCC within the framework of the GEMINA project (product P93b).