Lee et al. (2026) A novel approach for soil moisture retrieval from Sentinel-1 SAR via temporal stability-based backscatter analysis

Identification

• Journal: Remote Sensing of Environment
• Year: 2026
• Date: 2026-02-24
• Authors: Seulchan Lee, Seongkeun Cho, Minha Choi
• DOI: 10.1016/j.rse.2026.115319

Research Groups

• Department of Water Resources, Graduate School of Water Resources, Sungkyunkwan University, Suwon 16419, Republic of Korea
• School of Civil, Architectural Engineering & Landscape Architecture, Sungkyunkwan University, Suwon 16419, Republic of Korea

Short Summary

This study developed a novel temporal stability analysis (TSA)-based masking method for Sentinel-1 SAR to improve high-resolution soil moisture retrieval by effectively filtering noisy pixels. The TSA method significantly enhanced correlations and reduced errors compared to existing methods across diverse monitoring networks.

Objective

• To introduce and evaluate a novel masking approach based on temporal stability analysis (TSA) that identifies and retains Sentinel-1 SAR pixels reliably capturing soil moisture dynamics, thereby improving high-resolution soil moisture retrieval and mitigating noise from surface heterogeneities.

Study Configuration

• Spatial Scale: 10 m (SAR backscatter), 1 km (synthetic backscatter), 480 m (cosmic-ray neutron probe evaluation).
• Temporal Scale: Multi-temporal analysis for assessing the stability of soil moisture dynamics.

Methodology and Data

• Models used: Water Cloud Model (WCM), Land Surface Model (LSM).
• Data sources: Sentinel-1 Synthetic Aperture Radar (SAR), in situ soil moisture observations (SMC, TxSON, REMEDHUS networks), LSM-simulated soil moisture and vegetation descriptors, cosmic-ray neutron probe observations.

Main Results

• TSA-based masking improved correlations by up to 0.15 and reduced unbiased Root Mean Square Error (ubRMSE) by up to 0.014 m³/m³ against both LSM-derived and in situ soil moisture, compared to no masking and dynamic masking (DM).
• The method effectively excluded urban and water pixels while retaining pixels where backscatter signals, though influenced by vegetation, were properly compensated by the Water Cloud Model.
• In evaluations with cosmic-ray neutron probe observations, TSA improved the correlation to 0.61 at 480 m resolution.

Contributions

• Introduction of a novel temporal stability analysis (TSA)-based masking approach for Sentinel-1 SAR soil moisture retrieval.
• Demonstrated significant improvements in soil moisture retrieval accuracy (higher correlations, lower ubRMSE) compared to conventional dynamic masking and no masking.
• Enhanced capability to filter out noisy pixels (e.g., urban, water) and effectively manage vegetation-influenced signals, leading to more reliable high-resolution soil moisture products across diverse landscapes.

Funding

• Not specified in the provided text.

Citation

@article{Lee2026novel,
  author = {Lee, Seulchan and Cho, Seongkeun and Choi, Minha},
  title = {A novel approach for soil moisture retrieval from Sentinel-1 SAR via temporal stability-based backscatter analysis},
  journal = {Remote Sensing of Environment},
  year = {2026},
  doi = {10.1016/j.rse.2026.115319},
  url = {https://doi.org/10.1016/j.rse.2026.115319}
}