Soil Moisture SK4

International Soil Moisture Network (ISMN) site within Canada’s Real-Time In-Situ Soil Monitoring for Agriculture (RISMA) network in Saskatchewan, station 4. From 2014–2021, 15-minute in-situ record comprising multi-sensor soil moisture and temperature at six depths (0–5 cm to 150 cm) and collocated meteorology. The site has frequency-domain reflectometry sensors (Stevens HydraProbe II) installed at depths of 5 cm, 20 cm, 50 cm, 100 cm, and 150 cm below the surface. An additional surface measurement of the 0–5 cm layer is also recorded, effectively providing six distinct depth readings (surface 0–5 cm, 5 cm, 20cm, 50 cm, 100 cm, 150 cm). The station also measures air temperature, precipitation, relative humidity, wind speed, and wind direction at a nearby weather station, and it monitors battery voltage for maintenance. Data from this station are logged at 15-minute intervals, providing a high-resolution time series from 2014 through 2021. This data was used to research on "Multi-Scale Soil Moisture Prediction Using Recurrent Neural Networks with Temporal Attention."本数据集为加拿大萨斯喀彻温省农业实时原位土壤监测网络（Real-Time In-Situ Soil Monitoring for Agriculture, RISMA）下属国际土壤湿度网络（International Soil Moisture Network, ISMN）的4号观测站点数据。该站点在2014至2021年间采集了分辨率为15分钟的原位观测记录，包含6个深度区间（0–5 cm至150 cm）的多传感器土壤湿度与温度数据，以及配套的同步气象观测数据。 该站点搭载Stevens HydraProbe II型频域反射法（frequency-domain reflectometry）传感器，分别布设于地表下5 cm、20 cm、50 cm、100 cm及150 cm深度处。此外还额外记录了0–5 cm表层土壤的观测数据，由此共获得6组独立深度的观测值：表层0–5 cm、5 cm、20 cm、50 cm、100 cm及150 cm。 该站点同时同步采集邻近气象站的气温、降水量、相对湿度、风速及风向数据，并监测电池电压以保障运维工作。本数据集的观测数据以15分钟为间隔进行存储，形成了2014年至2021年的高分辨率时间序列。该项数据曾被用于题为“基于带时序注意力（Temporal Attention）机制的循环神经网络（Recurrent Neural Networks）的多尺度土壤湿度预测”的研究。