Abstract：Surface temperature is an important parameter in surface physical processes,and microwave remote sensing is one of the main methods for obtaining surface temperature information under cloud conditions.This study is based on the XGBoost model and uses FY-3D/MWRI brightness temperature data to construct an LST estimation model for the Qinghai Plateau.The performance of this model in typical flood and drought processes is compared and analyzed,and compared with existing business products.The results indicate that the model can well reflect the two low-temperature and two high-temperature regions of the Qinghai Plateau,and it performs stably during floods and droughts.The XGBoost model results have high accuracy.Compared with the CLDAS surface temperature product,the correlation coefficient r is higher by 0.12,the root mean square error RMSE is lower by 3.8 K,and the average absolute error MAE is lower by 3.2 K,indicating that the XGBoost model results have certain advantages in surface temperature inversion on the Qinghai Plateau.
2025, Vol. 42 
