In the present study the role of soil moisture (SM) in a statistical downscaling framework for precipitation in the EuroMediterranean domain is assessed. Different settings of the statistical downscaling models, differing in terms of the predictor variables used, are compared to quantify the influence of SM on the downscaling results. Results indicate an improvement of the skill of the statistical models when using SM information. This improvement is only moderate when averaged over the whole Euro-Mediterranean domain, but for specific regions the gain in performance is substantial. Regional projections of precipitation under the RCP4.5 and RCP8.5 scenario are considerably modified when SM is used as additional predictor in comparison with results based on atmospheric predictors alone.

No abstract