Soil moisture (SM) retrieved from satellite observations has become available at a global scale with relatively high spatial–temporal resolution, and the satellite-derived SM can be useful data sources where in-situ measurements are scarce or not available. In this study, the SM data from two different satellite sensors, the Advanced Scatterometer (ASCAT) and Advanced Microwave Scanning Radiometer 2 (AMSR2), are evaluated through the comparison with in-situ observation collected from twelve sites over a three-year period (2013–2015) in South Korea. The results reveal that the ASCAT descending overpass (09:30, the local equatorial crossing time) shows a better correlation with the in-situ observation than the ascending overpass (21:30, the local equatorial crossing time), while no significant difference in performance is found for AMSR2. Moreover, ASCAT SM retrieval shows a generally better agreement with in-situ observation. Considering the spatial mismatch and different measurement depths, a cumulative distribution function (CDF) matching method, as well as an exponential filter method, are employed to improve the applicability of satellite-derived SM. Specifically, the observation operators based on CDF matching are derived to find the optimal temporal period and tested by cross-validation. It is found that the CDF matching method split into two groups (i.e., growing and non-growing season) outperforms the other temporal groups. Additionally, considering different observation depths between the in-situ (>10 cm) and the satellite products (the top soil layer), the root-zone SM (RZSM) is derived from satellite surface SM by using the exponential filter method. For this study, a characteristic time length (T) at each observation depth is optimized by maximizing the r value between the SWI and the in-situ observation. Although the optimal T value generally increases with observation depth, it is clearly seen that T values are highly location-dependent. Given an encouraging improvement of the satellite SM estimation when scaling and filtering method applied, the results obtained in this study show that the satellite SM products have the useful potential for operational applications.

从卫星观测中获得的土壤水分（SM）已在全球范围内以相对较高的时空分辨率获得，并且在缺乏或不提供原位测量的情况下，卫星衍生的SM可以成为有用的数据来源。在这项研究中，通过与三年期间从十二个站点收集的现场观测结果进行比较，评估了来自两个不同卫星传感器，高级散射仪（ASCAT）和高级微波扫描辐射仪2（AMSR2）的SM数据（ 2013–2015）在韩国。结果表明，ASCAT下降立交桥（09:30，当地赤道穿越时间）显示出与实地观测的相关性比上升立交桥（21:30，当地赤道穿越时间）更好，而在找到AMSR2的性能。此外，ASCAT SM检索显示出与现场观察通常更好的一致性。考虑到空间失配和不同的测量深度，采用累积分布函数（CDF）匹配方法以及指数滤波方法来提高源自卫星的SM的适用性。具体而言，派生基于CDF匹配的观察算子以找到最佳时间周期，并通过交叉验证进行测试。发现CDF匹配方法分为两组（即生长季节和非生长季节）优于其他时间组。此外，考虑到原位（> 10厘米）和卫星产品（表土层）之间的不同观察深度，通过使用指数滤波方法从卫星表面SM得出根区SM（RZSM）。对于这项研究，SWI和原位观测值之间的r值。尽管最佳T值通常随观察深度而增加，但可以清楚地看到T值高度依赖于位置。考虑到采用缩放和滤波方法时卫星SM估计的令人鼓舞的改进，本研究获得的结果表明，卫星SM产品具有可操作应用的潜力。