Seasonal Climate Forecast to Brazil Obtained Through Global and Regional Climate Models

Abstract This study evaluates the ability of two global climate models (CPTEC and CFSv2) and a regional climate model (RegCM4) in predicting the seasonal climate in different regions of Brazil. RegCM4 was drove with the outputs of two global climate models: CFSv2 system and CPTEC. Two versions of RegCM were employed in the study: 4.3 and 4.5. RegCM4.3 was ran with six members of CFSv2 system, while RegCM4.5 was ran with a member of the CPTEC model. All forecasts started about two months before the season to be predicted and in all 94 simulations with the regional climate model were carried out. RegCM4 adds value to the predictions of the global climate models, mainly when it is nested to the CPTEC outputs. When RegCM4.5 is drove by CPTEC outputs and Emanuel cumulus convection parameterization is used, there is a good performance of the regional climate model in predicting the precipitation and air temperature in almost all Brazil.