Long-Term Surface Soil Moisture Dataset for the Huai River Basin (1951–2016)

This dataset provides a long-term (1951–2016) retrospective reconstruction of surface soil moisture (SSM_RF) at 0–5 cm depth for the Huai River Basin, China. It addresses the temporal limitations of the SMAP_L4 SSM product, which, despite its high resolution and accuracy, only spans from 2015 onwards. The reconstruction was developed using a machine learning framework based on the Random Forest algorithm. Inputs include ERA5-Land meteorological variables, SMAP ancillary data, and static geographic features such as soil texture and topographic indices.The SSM_RF dataset was validated against SMAP_L4 SSM, in-situ observations, and land surface model (LSM) outputs. Results show strong agreement in the northern and central basin (correlation > 0.6, ubRMSE < 0.04 m³/m³). This GeoTIFF-format dataset ( 9 km resolution) is suitable for drought monitoring, flood risk assessment, and climate-resilient water management. By extending SMAP-like data back to 1951, it provides a valuable resource for long-term hydrological studies under changing climate conditions.