Data for the article: "Root zone soil moisture estimation with Random Forests"

The dataset contains information on soil moisture, meteorological conditions, land cover, and soil hydrological groups for the soil moisture stations installed within the Raam soil moisture network. The network has a total of 15 stations within the Raam catchment, located in the southeastern portion of the Netherlands. The datasets covers the periods from April 2016 to December 2018. It was used for predicting root zone soil moisture using a Random Forest model and a 1-dimensional process-based model.<br> For each station, daily in situ measurements of surface soil moisture (SSM) at 5 cm and zone weighted depth-averaged root zone soil moisture (RZSM) are given. The meteorological conditions are obtained from daily datasets available from KNMI. First, the measurements from KNMI meteorological stations are interpolated in order to get spatially distributed values covering the study sites. The values from the interpolated maps were extracted for each point in the Raam network. <br><br>The vegetation characteristics are represented by leaf area index (LAI) obtained from MODIS. The crops at each station for each year are obtained from fieldwork data. The BOdemFysische Eenheden Kaart (BOFEK2012, Wosten et al., 2013), which is a map of soil hydro-physical properties for the Netherlands, was the basis for the information on the soil groups at the study sites. <br> <br>

本数据集涵盖荷兰拉姆（Raam）土壤湿度监测网络内布设的各土壤湿度站点的土壤湿度、气象条件、土地覆被及土壤水文分组相关信息。该网络在荷兰东南部的拉姆流域内共计布设15个监测站点，数据集时间跨度为2016年4月至2018年12月，曾被用于依托随机森林（Random Forest）模型与一维过程驱动模型开展根区土壤湿度预测任务。<br>针对每个监测站点，数据集提供了5cm深度表层土壤湿度（SSM）以及分区加权深度平均根区土壤湿度（RZSM）的逐日原位实测数据。气象数据源自荷兰皇家气象研究所（KNMI）的公开逐日数据集：首先对KNMI气象站点的实测数据进行空间插值，以获取覆盖研究区域的空间分布气象值，随后从插值生成的栅格地图中提取拉姆监测网络各站点对应位置的气象数据。<br><br>植被特征由从中分辨率成像光谱仪（MODIS）获取的叶面积指数（LAI）表征；各站点每年的作物种植信息均来自实地调查数据。荷兰土壤水文物理属性图《BOdemFysische Eenheden Kaart》（BOFEK2012，Wosten等，2013）为研究区域的土壤水文分组信息提供了核心数据基础。