A coupling ocean-atmosphere climatic modelling study for rainfall and soil moisture simulations on the São Francisco River basin

Abstract

The aim of this study was to obtain a better understanding of the Ocean-Atmosphere Global Circulation Coupling Model (CGCM) performance for forecasting the interannual rainfall variability on the São Francisco River Basin, during austral summer (DJF) 1997-2007. In addition, the rainfall predictions and calculated potential vapor transpiration were the input variables for the Hydrological Balance Model (HBM) experiments to obtain soil moisture estimations. Simulations using CGCM were compared with forecastings based on the Atmosphere Global Circulation Model (AGCM), which has been used previously for this purpose. Even though there were systematic errors of rainfall over estimations for the Basin, the CGCM had better performance than the AGCM at the spatial representation and showed positive correlation coefficients with observation values. These facts corroborate that ocean-atmosphere coupling is an important mechanism to be taken into account for rainfall forecasting at the Brazilian southeast zone. On the other hand, the HBM-AGCM and the HBM-CGCM were quite similar in terms of correlation coefficients (0,6) for soil moisture estimation. This suggests that the corrected estimated precipitation and potential evapotranspiration (ETP) resulting from climate modeling and dynamics of the AGCM and CGCM, as input data for water balance models in the seasonal scale, can be used to provide support to the best practices for the management of surface water in the basin of the São Francisco River.