Bias Correction of Remote-Sensed Surface Solar Radiation and Analysis of Meteorological Factor Influences in Plateau Regions: A Case Study of Lhasa

Xizang is characterized by high altitude, low air pressure, strong atmospheric transparency, and complex terrain, while sparse ground stations coexist with continuously available remotely sensed data, and systematic studies on SSR bias correction and meteorological influences under plateau conditions remain limited. This study focuses on a short-term spring case at one urban observation site in Lhasa, using observations collected from 4 to 30 April 2025 to investigate the bias correction of remotely sensed surface solar radiation (SSR) and the influence of meteorological factors. Ground observations and Himawari-8 remotely sensed data were first spatially and temporally matched and preprocessed. Spearman correlation analysis was then used to select key input variables. Support vector regression, random forest, XGBoost, and multiple linear regression models were subsequently developed, followed by a Stacking ensemble model for bias correction. Finally, local sensitivity analysis was conducted to examine the local response of the correction model to selected meteorological variables at a representative baseline point. The results showed that the correlation coefficient between remotely sensed SSR and ground-observed SSR was 0.88 (p<0.001). The Stacking ensemble model achieved the best performance, with a test set R2 of 0.8796, an MAE of 118.54 W/m2, and an RMSE of 152.41 W/m2. Local sensitivity analysis showed that a +10 hPa perturbation in air pressure increased the model output by 173.45 W/m2, while a +10 °C perturbation in air temperature increased the output by 23.76 W/m2. This study provides a reference for improving the accuracy of remotely sensed SSR and for solar resource assessment in plateau regions.