流域蒸散发定量估算一直是水科学领域的研究前沿,水文模型和遥感反演是当前估算区域蒸散发的常用手段。研究通过数据同化,集成水文模型和遥感模型的优势,耦合遥感蒸散发到水文模型中以实现多源数据下的蒸散发数据同化。选择北京市沙河流域为研究区,分布式时变增益水文模型作为模型算子,基于集合卡尔曼滤波同化算法,利用双层遥感模型模拟的蒸散发同化水文模型,并基于地面通量站观测的日蒸散发进行验证。结果表明,同化结果与观测数据相比平均绝对百分比误差较同化前减少,精度进一步提升,且当遥感观测输入频繁时精度改善明显。研究证明基于水文模型的蒸散发数据同化系统,是一种可实现输出精度更高和时序连续的区域蒸散发的新型模式。该成果将进一步丰富创新蒸散发估算的学科内容,为准确理解区域水循环规律提供科学依据。

The quantitative estimation of watershed Evapotranspiration (ET) has been an international frontier in water sciences for a long time. Hydrological models and remote sensing ET models are usually used to estimate regional ET at different spacetime scales, but these two methods are obviously insufficient to obtain precise and continuous regional ET. The hydrological models have the capability to simulate time-continuous daily or monthly ET processes, but the accuracy is not high compared with remote sensing ET models. The applicability of remote sensing ET models based on surface energy balance is restricted by the lack of high frequency and high resolution thermal data. A compromise between these two methodologies is represented by improving the optimization of hydrological models on the basis of a new ET series, which are produced by Data Assimilation (DA) scheme combining sparse remote estimates into the hydrological model. This study aimed to integrate the advantages of the two models to simulate the daily ET processes in Shahe River basin, Beijing. For this progect, the distributed hydrological model was fist constructed and the daily hydrological processes of 19992007 simulated. Then, the Ensemble Kalman Filter (EnKF) was used to assimilate the ET series calculated by remote sensing retrieval into the hydrological model to adjust the simulation. The results show that the ET estimation accuracy is improved after the data assimilation, and the MAPE between the DSMbased ETs and LASbased ETs in the study area is reduced. The integrated method is proved better, and improves the hydrology modeling accuracy. Therefore, the project successfully develops a new land surface ET mode with the advantages of hydrological model and remote sensing ET model, and the study founds the new method could simulate regional ET with high accuracy and continuous time series. The new land surface ET model not only follows the surface energy balance, but also meets the regional water balance, and has more perfect water thermal coupling mechanism. The study will further enrich the content of ET estimation disciplines, and provide a scientific basis for better understanding of the laws of regional water cycle.