Author(s): H. Madsen; M. A. Sunyer
Linked Author(s): Henrik Madsen
Keywords: Climate change; Statistical downscaling; Weather generator; Extreme events; Uncertainty assessment
Abstract: One of the greatest challenges in climate change impact assessments is to downscale the information from climate model projections to provide reliable estimates of the expected changes in climate variables at the local or regional scale of application. In this study, an ensemble of regional climate model projections from the European ENSEMBLES data base is downscaled to estimate the changes in precipitation. Different statistical downscaling procedures based on a common change factor methodology are analysed and compared. Changes in statistical characteristics (e.g. mean, variance, skewness, probability of dry days, etc.) of precipitation from the present to the future climate are extracted from the climate model results and used to project statistical characteristics of the precipitation representing the local scale. Results from the different downscaling methods are compared with special emphasis on the estimation of extreme precipitation. The paper discusses the limitations and advantages of different statistical downscaling methods and the uncertainties in downscaling climate change projections.