Monthly mean solar radiative fluxes and cloud forcing over South America in the period of 1986-88: GCM results and satellite-derived data

Abstract

The incident solar radiative fluxes, simulated by an atmospheric general circulation model over South America for the period 1986-88, are compared with the satellite-derived surface fluxes provided by the Surface Radiation Budget (SRB) datasets. The comparison shows that the model systematically overestimates both all-sky and clear-sky SRB fluxes while representing well their latitudinal variations. In order to analyze the reasons for the bias, the shortwave radiation code employed in the model is tested with more comprehensive techniques in a stand-alone mode. The results of testing demonstrate that the code underestimates solar radiation absorption in the clear-sky atmosphere due to trace gases and aerosols. The underestimation of the absorption due to aerosols contributes noticeably to the surface flux bias. The impact of clouds on the surface fluxes is estimated by calculating cloud radiative forcing, defined as the difference between the net surface fluxes in all-sky and clear-sky conditions. The comparison of model-simulated and satellite-derived values of cloud radiative forcing over South America demonstrates that the model simulates fairly well its latitudinal variations and annual cycles as compared with SRB data. However, the model overestimates the SRB surface cloud radiative forcing over the tropical region of South America and underestimates it over the extratropical region in both January and July. The comparisons of the incident surface fluxes simulated by the model at the grid points with those measured at three Amazonian observational sites show good agreement at one site and large discrepancies at the other two sites.