Methods of empirical statistical downscaling and error correction of regional climate models: calibration and testing
Three statistical downscaling techniques (Linear Scaling, LOCI and Quantile Mapping) were applied to correct temperature (minimum and maximum) and precipitation simulations from four Regional Climate Models (RCMs). The simulated series were corrected with respect to the climatology provided by a collection of daily series over Italy. The RCMs simulations cover the period 1961-2000, with boundary conditions provided by the ECMWF reanalysis ERA-40. The performance of the MOS (Model Output Statistics) methods was assessed using three different validation procedures. The results enable to evaluate the performance of the applied MOS methods, in terms of their ability to correct RCM simulations. After the validation phase, these downscaling techniques can be applied to future simulations to improve or correct, at the local scale, climate projections produced by regional models for different emission scenarios.
Publication available only on-line