河流流量是水文循环过程的关键变量，在洪水预警、水资源调度和生态环境管理中具有重要意义。本论文选取黄河源区的唐乃亥水文站为试点，提出一种融合卫星遥感和机器学习方法的河流流量监测技术。首先利用哨兵2号遥感影像提取河流的水面宽度，联合全球陆面数据同化系统模拟的5种气象水文变量（蒸散发、土壤湿度、温度、陆地水储量和径流量）作为驱动因子，分别构建了基于四种统计方法（线性函数、幂函数、指数函数和多项式函数）和四种机器学习方法（XGBoost、Random Forest、LightGBM和CatBoost）的流量监测模型，评估不同模型监测结果间的差异并利用SHAP方法量化不同驱动因子的重要性。结果表明，四种统计方法中，多项式函数模型在测试期的效果优于其它三种模型；相比于传统统计方法，机器学习方法在模拟精度和稳定性方面有明显的改善，决定系数提高了46.15%，均方根误差（RMSE）和平均绝对误差（MAE）分别减少了54.61%和55.65%。Random Forest模型在测试期的模拟效果优于其它三种模型，其决定系数、RMSE和MAE分别为0.96, 172.81 m3/s和147.33 m3/s。SHAP方法表明水面宽度在流量监测模型中具有最显著的贡献（189.02），其次为土壤湿度（145.11）和温度（97.41）。本研究证实了联合卫星遥感和机器学习方法在复杂地形与资料匮乏区域开展高精度流量估算的可行性和优越性。

River discharge is a pivotal variable within the hydrological cycle, holding significant importance for flood warning, water resource allocation, and eco-environmental management. Traditional ground-based methods are limited by sparse station distribution and high costs of data acquisition, particularly in areas with complex terrains or remote regions, making it difficult to meet the demands of precise water resource management. Satellite remote sensing technology offers extensive coverage and high spatiotemporal resolution, providing new data sources and methodologies for river discharge monitoring. Machine learning (ML) approaches can accurately simulate complex relationships between river discharge and multiple driving factors, offering novel avenues for processing intricate hydrological data and optimizing models. By integrating machine learning algorithms with remote sensing and in-situ river discharge, it can provide the innovative measure for the efficient and reliable of river discharge monitoring. This study selected the Tangnaihai Hydrometry Station as the study area, and proposed a river discharge monitoring method by integrating satellite remote sensing and ML methods. Firslty, the Sentinel-2 imagery was utilized to extract river water surface width based on the Google Earth Engine cloud platform. The GLDAS v2.2 model-simulated five variables were served as predictor variables, namely evapotranspiration, soil moisture, temperature, terrestrial water storage and runoff. Discharge monitoring models were subsequently developed based on four statistical methods (linear function, power function, exponential function, and polynomial function) and four ML algorithms (XGBoost, Random Forest, LightGBM, and CatBoost). The discrepancies among different models were assessed, and the Shapley Additive Explanation (SHAP) method was employed to quantify the importance of different input variables. The results demonstrate that the polynomial function model demonstrates superior performance over other three statistical models during the testing period, with an R2of 0.67, and its error metrics (MAE: 319.01 m3/s, RMSE: 393.14 m3/s) were lower than those of the other three statistical models. Compare to traditional statistical approaches, ML models exhibit significant improvements overall in both simulation accuracy and stability, and the coefficient of determination (R2) increased by 46.15%, while the root mean square error (RMSE) and mean absolute error (MAE) decreased by 54.61% and 55.65, respectively. Notably, the Random Forest model achieved the optimal performance in the testing phase, with the R2 of 0.96, RMSE of 172.81 m3/s, and MAE of 147.33 m3/s, reflecting robust generalization capability and stability. SHAP analysis revealed that water surface width contributed most significantly to the discharge monitoring model (189.02), followed by soil moisture (145.11) and temperature (97.41). The runoff variable exhibited the minimal degree of influence on the river discharge monitoring model with the value of 14.14%. This study confirms the feasibility and superiority of integrating satellite remote sensing and ML approaches for high-accuracy discharge estimation in regions characterized by complex topography and data scarcity. Future work could be optimized by integration of higher-resolution satellite imagery and mechanistic models with physical processes.