Avaliação da sazonalidade da precipitação no Nordeste brasileiro simulada pelos modelos do IPCC-AR4

Abstract

The global models of the Intergovernmental Panel on Climate Change Fourth Assessment Report (IPCC- AR4) are evaluated for the representation of the precipitation seasonality in the Northern Northeast of Brazil for the period 1901 to 1999. This evaluation is performed using data from the Climatic Research Unit (CRU) and the 20th Century Reanalysis V2 of the National Oceanic and Atmospheric Administration (NOAA). A performance index based on measures of correlation and mean squared error is proposed for evaluation and ranking of the models. Most models provide correlations exceeding 0.85 and mean squared error less than 4% compared to the percentage of monthly rainfall of the climatology averaged over the region. According to the evaluation using the proposed performance index, models BCCR-BCM2.0 (Norwegian) GFDL_CM2_1, UKMO_HADCM3, GISS_MODEL_E_R and CCCMA_CGCM3_1_t63 provide the best representations for the precipitation seasonality in the region.