遥感土壤水分数据与水文模型的同化研究，是提供流域尺度土壤水分数据的有效手段，也是提升流域水文模型模拟性能的基础。山区土壤水分时空变异性十分显著，在山区流域开展相关研究尤为重要。但由于山区实测资料匮乏，遥感数据存在一定不足，在山区大流域尺度上的同化研究更具挑战性。本研究将以典型高寒山区--祁连山区的两个重点流域(黑河上游和疏勒河上游)为研究区，对SMOS遥感土壤水分数据进行不确定性分析及离散化处理后，确定其与水文模型SWAT的最佳同化方案，采用EnKF (Ensemble Kalman filter)方法进行同化；开发SWAT-SM模型，直接利用同化后的表层土壤水分数据集进行模拟，评估土壤水分数据同化对分布式水文模型在高寒山区模拟性能的影响。项目旨在高寒山区大流域尺度上开展遥感土壤水分数据与水文模型的同化及影响研究，可为准确模拟高寒山区流域水文过程提供重要手段，并推动内陆河流域的水资源管理。

Assimilation of remote sensing soil moisture data in distributed hydrological model, is an effective way to provide accurate soil moisture data at large scale, and also the basis to improve the performance of distributed hydrological model. As soil moisture varies significantly at both temporal and spatial scale in mountain area, the assimilation is very necessary but challenging in these areas. However, due to lack of observations and defects of remote sensing data, little has been done at large watershed scale in mountain area. This study aims to assimilate remote sensing soil moisture data in distributed hydrological model and evaluate the effect of assimilation on model performance in high elevation and cold mountain area. The research area consists of two inland river watersheds, the upper reach of the Heihe River watershed and Shule River watershed in the Qilian mountain area. Based on the uncertainty analysis, SMOS data is disaggregated to improve its accuracy and spatial resolution. The best assimilation scheme is determined by comparing four different tempo-spatial scale matching schemes through EnKF methods. SWAT-SM model will be developed to simulate hydrological processes directly using the assimilated soil moisture dataset of surface layer. The assimilation effect on model performance in high elevation and cold mountain area is evaluated by comparing the simulation results of SWAT, SWAT-SM and the observations. This study will provide an effective way for simulating hydrological processes in high elevation and cold mountain area, and promote scientific management of water resources in inland watersheds.