Regionalización estadística del sistema de predicción estacional del ECMWF (SEAS5) en el marco de los servicios climáticos

Abstract

Quality seasonal forecasts properly downscaled to the resolution needed by users is a key issue for planning of water resources or agriculture sectors. Together with a better understanding of the mechanisms driving climate variability at this timescale, a downscaling post-process has been developed by AEMET (Petisco, 2008a; Petisco 2008b; Amblar, 2017) for improving the usage of ECMWF seasonal forecasting system (SEAS5). The algorithm makes successive use of an analogue technique -based on a Euclidean distance- and regression to downscale maximum and minimum temperature and precipitation. One important ingredient of this algorithm is the usage of a new observational gridded dataset with 5 km of horizontal resolution covering the peninsular Spain and the Balearic Islands (Peral et al., 2017). Results show small error metric values, allowing to conclude that at least over the Iberian Peninsula –where the high resolution observational grid is available- the here described downscaling algorithm performs satisfactory.