Simulation of Diffuse Solar Radiation with Tree-Based Evolutionary Hybrid Models and Satellite Data.
Shuting Zhao,Youzhen Xiang,Lifeng Wu,Xiaoqiang Liu,Jianhua Dong,Fucang Zhang,Zhijun Li,Yaokui Cui
DOI: https://doi.org/10.3390/rs15071885
IF: 5
2023-01-01
Remote Sensing
Abstract:Diffuse solar radiation (Rd) provides basic data for designing and optimizing solar energy systems. Owing to the notable unavailability in many regions of the world, Rd is traditionally estimated by models through other easily available meteorological factors. However, in the absence of ground weather station data, such models often need to be supplemented according to satellite remote sensing data. The performance of Himawari-7 satellite inversion of Rd was evaluated in the study, and hybrid models were established (XGBoost_DE, XGBoost_FPA, XGBoost_GOA, and XGBoost_GWO), so as to improve the satellite data and achieve a better utilization effect. The meteorological data of 14 Rd stations in mainland China from 2011 to 2015 were used. Four input combinations (L1–L4) and eight input combinations (S1–S8) of meteorological factors corresponding to satellite remote sensing data were used for model simulation, while two optimal combinations (S7 and S8) were selected for cross-station application. The results revealed that the accuracy of Himawari-7 satellite Rd data was low, with RMSE, R2, MAE, and MBE values of 2.498 MJ·m−2·d−1, 0.617, 1.799 MJ·m−2·d−1, and 0.323 MJ·m−2·d−1, respectively. The performance of these coupled models based on satellite data was significantly improved. The RMSE and MAE values increased by 15.5% and 9.4%, respectively, while the R2 value decreased by 10.9 %. Compared with others based on satellite data, the XGBoost_GOA model exhibited optimal performance. The mean values of RMSE, R2, and MAE were 1.63 MJ·m−2·d−1, 0.76 and 1.21 MJ·m−2·d−1, respectively. The XGBoost_GWO model exhibited optimal performance in the cross-station application, and the average RMSE value was reduced by 2.3–10.5% compared with the other models. The meteorological factors input by the models exhibited different levels of significance in different scenarios. Rd_s was the main meteorological parameter that affected the model based on satellite data, while RH exhibited a significant improvement in the XGBoost_FPA and XGBoost_GWO models based on ground weather stations data. Accordingly, the present authors believe that the XGBoost_GOA model has excellent ability for simulating Rd, while the XGBoost_GWO model allows for cross-station simulation of Rd from satellite data.