A synergic study on estimating surface downward shortwave radiation from satellite data

Surface downward shortwave radiation (DSR) is a fundamental variable in determining the Earth's balance and essential many applications. Considerable efforts have been devoted to algorithm development, product generation, validation. However, few studies focused on comparing retrieval approaches, examining their strengths weaknesses, identifying most suitable scenarios for each approach. In this study, we implemented evaluated five representative DSR algorithms, including forward parameterization approach, two physical inversion methods (look-up table (LUT) optimization), statistical (direct estimation neural networks). We then proposed an algorithm-integration framework that combined results of these further improve accuracy consistency. To validate retrievals, used in-situ data collected at 25 stations Baseline Radiation Network (BSRN) over one year. Validation revealed consistently performed best, with overall root mean square error (RMSE) 91.7 W/m 2 or relative RMSE 16.9%, although it generated fewest valid retrievals. For identical set, LUT approach comparable those parameterization. The network-based reduced by 11.0 2.0%, compared best individual algorithm. Our analysis demonstrates integration promising way obtain are superior estimates from any • Five approaches were compared. Parameterization but An was combine algorithms. Algorithm produced