Answers are grounded in this paper and cite their sources.
Soil moisture regulates land–atmosphere coupling, agricultural productivity, hydrological processes, and ecosystem functions; however, accurately estimating it over heterogeneous semi-arid regions remains challenging. Here, we present a multi-sensor approach that integrates radar backscatter from Sentinel-1 and Sentinel-2, optical reflectance from Landsat-8, and ancillary products from the Soil Moisture Active Passive (SMAP) mission to estimate surface soil moisture. The analysis focuses on the Thamirabarani river basin for the period 2015–2025, and SMAP products were used as the independent reference data for evaluation. A multiple linear regression formulation yields robust performance, with a Nash-Sutcliffe efficiency of 0.78, a root-mean-square error of 0.040 m 3 /m 3 , and a mean absolute error of 0.032 m 3 /m 3 . Overall, accuracy decreases during the pre-monsoon period and improves in the post-monsoon season because increased precipitation raises soil moisture and alters vegetation structure. High soil moisture (0.41–0.48 m 3 /m 3 ) was primarily found in the basin’s northeastern and western regions, where dense forest cover enhances water retention. Sensitivity diagnostics indicate that radar-vegetation interaction and preceding precipitation account for most of the variability in the retrieval. Comparative evaluation against alternative machine learning algorithms yields similar predictive accuracy while preserving physical interpretability. Uncertainty mainly arises from SMAP products and from spatial-scale discrepancies between coarse satellite grids and sub-grid heterogeneity. The present results demonstrate that an interpretable, multi-sensor integration approach can resolve regional-scale soil moisture dynamics in semi-arid environments and support land and water resource management in areas with sparse in situ observations.
No takes yet. Share an insight, caveat, or question.
Harani et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: