CLARIS Project: towards climate downscaling in South America

Abstract

We explore the bias in monthly and seasonal mean precipitation simulated by ensembles of different regional climate models over South America within the context of the EU-FP6 CLARIS project (A Europe-South America Network for Climate Change Assessment and Impact Studies). We briefly described two series of coordinated simulations: (i) Case studies of anomalous months for south-eastern South America performed with an ensemble of six models, and (ii) A multiyear simulation of the period 1991-2000 performed by four models. The models have been forced with the European Centre for Medium Range Weather Forecasting Reanalysis (ERA-40) and are compared to observational data compiled by the Climatic Research Unit (CRU). The ensemble-mean bias can be large when simulating particular extreme periods in La Plata Basin. Our multi-model analysis suggests that even though the ten-year ensemble mean is able to capture the major regional characteristics of seasonal mean precipitation for South America, models individually display considerable precipitation biases especially in tropical areas. The relatively good performance of the multimodel annual average over La Plata Basin results from the cancelation of offsetting errors in the individual models.