Estimation of monthly-mean global solar radiation using MODIS atmospheric product over China

This paper investigated the potential of MOD08-M3 atmospheric product in estimation of monthly-mean solar radiation. 8 models were developed using cloud fraction (CF), cloud optical thickness (COT), precipitable water vapor (PWV) and aerosol optical thickness (AOT) at 50 stations across China. All the models give reasonable results with average RMSE of 1.247MJm−2 and MAPE of 9.9%. Models have lower RMSE in cool temperature (CT) and warm temperate (WT) zones. In terms of MAPE, models perform better in Qinghai-Tibet plateau climate (QT) zone. Model accuracy can be significantly improved by introducing COT and PWV. The improvements by introducing COT are more pronounced in summer for CT, WT and ST regions. While inclusion of PWV is more effective in summer, autumn, and winter for CT, QT, and ST regions, respectively. However, introducing AOT does not contribute to the improvement in estimation accuracy. The performances of models show seasonal behavior. In terms of MAPE, models perform best in summer for CT and WT regions, and in autumn for ST region. Lowest RMSE are observed in autumn and winter for CT and QT regions, respectively. Models have lower RMSE in both autumn and winter for WT and ST regions.