高分辨率遥感蒸散数据集的构建受到数据源的限制和云的影响，单一传感器无法达到高时空分辨率覆盖。本文分 析了ETWatch不同尺度遥感蒸散结果的空间特征，通过几种融合方法的比较，分析数据融合前后的数据特征和信息量，将 时空适应性反射率融合模型（STARFM）集成到ETWatch，用于不同尺度遥感蒸散数据的融合，该方法可以很好的结合高 低分辨率数据的空间分布和时间分布信息，在时间上保留了高时间分辨率数据的时间变化趋势，空间上又反映了高空间分 辨率数据的空间细节差异，STARFM融合后的日ET数据与融合前1 km 日ET数据的平均相对误差为1.75%，融合后的月ET 数据与融合前1 km 月ET数据的平均相对误差为0.2%，STARFM适合于不同尺度下遥感ET数据的融合。

Construction of a high-resolution remote-sensing evapotranspiration (ET) dataset is restricted by remote sensing data sources and clouds. Single remote sensor data cannot cover the land with high spatial and temporal resolution. In this paper, we analyzed the spatial characteristics of different scale ET data in ETWatch, compared several common fusion methods, and analyzed the data characteristics and information before and after data fusion. We integrated the spatial and temporal adaptive reflectance fusion model (STARFM) into ETWatch to fuse different scale remote sensing ET data. The results show that the STARFM fusion method effectively can integrate the spatial and temporal distribution information of high & low resolution data, with an average error of 1.75%, compared with input of 1 km daily ET, with a monthly average error of 0.2% compared with input of 1km month ET. The STARFM model is adaptive to fusing different scales of ET data.