The accuracy improvement of clear-sky surface shortwave radiation derived from CERES SSF dataset with a simulation analysis

Abstract Towards the Xiaotang region along the northern margin of the China's largest desert, a quantitative assessment of the precision of clear-sky satellite observations (the Single Scanner Footprint TOA/Surface Fluxes and Clouds downward surface shortwave radiation product of Clouds and the Earth's Radiant Energy System (CERES), DSSRCER) is conducted, the localized inversion mode of “absolutely clear-sky” downward surface shortwave radiation (DSSR) is established, and the “absolutely clear-sky” DSSR in Xiaotang during 2005–2018 is simulated by the Santa Barbara Discrete Atmospheric Radiative Transfer (SBDART) model. In general, under the “absolutely clear-sky” condition of Xiaotang region, there is a significant error in DSSRCER, and the simulated results of SBDART (DSSRSBD) with same input parameters as DSSRCER is better and more comparable. Single scattering albedo (SSA), asymmetry parameter (ASY) and aerosol optical depth (AOD) play crucial roles in deciding the accuracy of DSSR, and after parameter adjustment, the DSSRSBD is better than the initial, which is improved remarkably with all indexes of the fitting results greatly improved. The temporal variation of the DSSR during 2005–2018 indicates that the highest annual average value is found in 2008 (770.00 W·m−2), while the lowest appears in 2010 (600.97 W·m−2). Besides, the highest seasonal mean DSSR appears in summer, which between 860.6 and 935.07 W·m−2, while reaches the lowest in winter (403.79–587.53 W·m−2). Moreover, the monthly average DSSR changes as a curve with a single peak and is close to normal distribution, the highest appears in June (934.61 W·m−2), while the minimum with the value of 390.34 W·m−2 is found in December. All of the solar elevation angle, the characteristics of climate and aerosol particles in different seasons may contribute to the temporal variation.