Soil moisture is an integral part of the climate system. There are numerous monitoring networks operated by state and federal agencies that provide real‐time soil moisture measurements in the United States. These measurements have value for a variety of applications, such as operational drought monitoring. This paper describes the methods that are used to collect and quality control soil moisture measurements from multiple networks, to convert volumetric water content to soil moisture percentiles, and to interpolate these data to a national 4‐km grid in the contiguous United States. The accuracy of three interpolation methods (inverse distance weighting [IDW], ordinary kriging [OK], and regression kriging [RK]) is evaluated, and we conclude that RK provides the most accurate approach for producing national soil moisture percentile maps. At 5 cm, the 12‐yr mean absolute error (MAE) for RK of the soil moisture percentiles is 0.17, followed by OK (0.19) and IDW (0.27). At 20 cm, the MAEs for RK, OK, and IDW are 0.18, 0.19 and 0.26, respectively. Two case studies, the 2011 drought and the 2015 summer floods in Oklahoma and West Texas, were used to demonstrate the utility of these soil moisture percentile maps.

中文翻译：

---

制定和评估国家土壤湿度百分位图

土壤水分是气候系统不可或缺的一部分。在美国，有许多州和联邦机构运营的监控网络可提供实时的土壤湿度测量。这些测量对于多种应用具有价值，例如可操作的干旱监测。本文介绍了用于收集和质量控制来自多个网络的土壤湿度测量值，将体积水含量转换为土壤湿度百分位数，并将这些数据插值到美国连续4公里的国家网格中的方法。评估了三种插值方法（反距离权重[IDW]，普通克里格法[OK]和回归克里格法[RK]）的准确性，我们得出结论，RK为提供全国土壤水分百分比图提供了最准确的方法。在5厘米处 土壤湿度百分位数RK的12年平均绝对误差（MAE）为0.17，其次是OK（0.19）和IDW（0.27）。在20厘米处，RK，OK和IDW的MAE分别为0.18、0.19和0.26。使用两个案例研究（2011年干旱和2015年俄克拉荷马州和西德克萨斯州的夏季洪水）来证明这些土壤湿度百分位图的实用性。