结合洋河水库实测水质数据和 Sentinel-2（哨兵 2 号）卫星遥感影像数据，基于线性回归方法，建立遥感波段 反射率与水体总磷实测质量浓度之间的反演模型，反演洋河水库 2016—2019 年总磷质量浓度，并分析库区总磷 质量浓度时空变化特征。研究结果表明：线性回归遥感反演模型决定系数达到 0.818，相对误差平均值在 20% 以 内，计算结果较稳定，模型适用性较好。在时间分布上，洋河水库水质整体基本较好，2016—2018 年洋河水库水 质较好，达到Ⅲ类，部分月份水质较差，超过Ⅲ类，主要集中在秋季；2019 年洋河水库水质较差，整体达到Ⅳ类，部 分月份出现Ⅴ类，超标现象集中在夏秋季节。在空间分布上，非汛期洋河水库总磷质量浓度自东、西入口及取水 口向水库中心区逐渐升高，汛期靠近西入口的水域总磷质量浓度明显高于其他区域。利用哨兵 2 号卫星遥感影 像对总磷质量浓度进行反演，可为水库的水质监管提供技术支撑。

The measured data of Yanghe Reservoir and Sentinel-2 satellite image data were used to construct an inversion model of total phosphorus mass concentration in the water body, and the temporal and spatial variation characteristics of total phosphorus concentration in Yanghe Reservoir are analyzed, which provided a new way for water quality monitoring of the reservoir, and provided a new method for the reservoir area. Scientific basis was provide for water environment management.Based on the linear regression method, an inversion model between the reflectivity of the remote sensing band and the measured mass concentration of total phosphorus in water was established, and the total phosphorus mass concentration in Yanghe Reservoir from 2016 to 2019 was inverted, and the spatiotemporal characteristics were analyzed.In terms of time distribution, the average concentration of total phosphorus mass concentration in the reservoir in spring and winter is lower than that in summer and autumn, and the water quality in summer and autumn is worse than other seasons. From 2016 to 2018, the water quality of Yanghe Reservoir was relatively good, reaching Class Ⅲ, while the water quality in some months was poor, exceeding Class Ⅲ, mainly in autumn. In 2019, the water quality of Yanghe Reservoir was poor, reaching Class Ⅳ as a whole, and Class Ⅴ in some months, and the excesses were concentrated in summer and autumn. In terms of spatial distribution, the mass concentration of total phosphorus in Yanghe Reservoir in non-flood seasons gradually increased from the east and west entrances and water intakes to the central area of the reservoir. During the flood season, the mass concentration of total phosphorus in the waters near the west entrance was significantly higher than that in other areas. The optimal Sentinel-2 satellite band combination (B3/B5) and the field measured data of the same period were used to establish a total phosphorus mass concentration inversion model. The coefficient of determination (R2) of the established model reaches 0.818, which indicates good applicability. The error analysis between the inversion value and the measured value shows that the average relative error is 19.5%, and the root mean square error is 0.008 67 mg/L, indicating a good model accuracy. However, due to the limitation of geographical conditions, the selected sampling points are few. In the follow-up research, it is necessary to further accumulate the measured data and optimize the regression algorithm. The reason for this phenomenon is that endogenous pollution may be the most important factor in the excess of total phosphorus concentration in Yanghe Reservoir. At the same time, according to the estimation results of pollution sources in the Yanghe Reservoir basin, the total phosphorus non-point source pollution load of the main tributaries on the west side and the Xiyang River accounts for nearly 2/3 of the whole basin. And it is mainly discharged from rural life and livestock and poultry breeding. It also reflects the reason for the higher total phosphorus mass concentration on the west side of the reservoir during the flood season.