The added value of regional climate simulations at kilometre-scale resolution to describe daily wind speed: the CORDEX FPS-Convection multi-model ensemble runs over the Alps

In the recent past, the increase in computational resources allowed researchers to run simulations at increasingly horizontal and time resolutions. One such project is the World Climate Research Program’s Coordinated Regional Downscaling Experiments Flagship Pilot Studies (FPS) on convective phenomena. This FPS encompasses a set of simulations driven by the ERA-Interim reanalysis for the period from 2000-2009 (hindcast) and by the Coupled Model Intercomparison Project Phase 5 Global models for the 1996-2005 period (historical). Most models feature a horizontal resolution of 2.2 to 3 km, nested in an intermediate resolution of 12-25 km. An extended Alpine domain is considered for the simulations, due to the complexity of the mountain system together with heavy precipitation events, a large observational network and the high population density of the area. This initiative aims to build first-of-its-kind ensemble climate experiments of convective-permitting models to investigate convective processes over Europe and the Mediterranean.

 

In this study, the Distribution Added Value metric is used to determine the improvement of the representation of all available FPS hindcast and historical simulations for the daily mean wind speed. The analysis is performed on normalized empirical probability distributions and considers station observation data as a reference. The use of a normalized metric allows for spatial comparison among the different altitudes and seasons. This approach permits a direct assessment of the added value between the higher resolution convection-permitting regional climate model simulations against their global driving simulations and respective coarser resolution Regional Model counterparts. Although the complexity of such simulations, those not always reveal an added value. In general, results show that models add value to their reanalysis or forcing global model, but the nature and magnitude of the improvement on the representation of wind speed vary depending on the model, the spatial distribution and the season.