Avaliação do Desempenho das Simulações Históricas dos ´Modelos do CMIP6 na Representação do Regime de Precipitação no Sul do Brasil

This study assesses the performance of historical simulations from CMIP6 models in representing precipitation patterns in southern Brazil between 1982 and 2014, comparing these simulations with ERA5 reanalysis data. The results demonstrate a generally good agreement in the amplitudes of seasonal precipitation between CMIP6 and ERA5, except for Winter, where climate models underestimated rainfall percentiles. During the Summer, the ACCESS-CM2 and BCC-ESM1 models exhibited the strongest correlations. In the Winter and Autumn seasons, the NorESM2-MM model demonstrated the best performance in terms of correlation and standard deviation. However, for the Spring season, despite some models exhibiting strong correlation, there were disparities in the standard deviation.