Zan et al. (2026) Validation of high-resolution surface soil moisture time series retrieved by means of SAR interferometry

Identification

• Journal: Remote Sensing of Environment
• Year: 2026
• Date: 2026-01-29
• Authors: Francesco De Zan, Paolo Filippucci, L. Brocca
• DOI: 10.1016/j.rse.2026.115266

Research Groups

• delta phi remote sensing GmbH, Gilching, Germany
• National Research Council, Research Institute for Geo-Hydrological Protection, Perugia, Italy

Short Summary

This paper introduces and validates a novel, computationally efficient algorithm for high-resolution surface soil moisture retrieval using Synthetic Aperture Radar (SAR) interferometry and closure phases from Sentinel-1 data. The resulting products demonstrate strong correlations with modeled and alternative remote-sensing soil moisture products, particularly in areas characterized by high interferometric coherence.

Objective

• To develop and validate a novel, computationally efficient algorithm for high-resolution surface soil moisture retrieval based on Synthetic Aperture Radar (SAR) interferometry and closure phases.

Study Configuration

• Spatial Scale: Local processing windows of approximately 240 meters by 240 meters (79x18 pixels); final product aggregated to a 1 kilometer grid spacing; spatially filtered product has a resolution of 5 kilometers (full width at half power). Test sites ranged from 23,000 square kilometers to 42,000 square kilometers.
• Temporal Scale: Multi-year time series, with soil moisture measurements generated for each SAR acquisition (typically 3 to 6 days for Sentinel-1). Study periods for test sites ranged from 1 year to 3 years (e.g., Eastern Spain: January 2019 to December 2021).

Methodology and Data

• Models used:
  • Novel algorithm based on SAR interferometry and closure phases, building on models from De Zan et al. (2014) and De Zan and Gomba (2018).
  • Exponential filter (Brocca et al., 2010) for temporal smoothing.
  • 2-D Gaussian smoothing kernel for spatial filtering.
  • RT1 model (Quast et al., 2019, 2023) for RT1 soil moisture product.
  • Land Parameter Retrieval Model (Van der Schalie et al., 2017) for SMAP 1 km product.
  • Global Land Evaporation Amsterdam Model (GLEAM, Miralles et al., 2011a,b) for GLEAM product.
• Data sources:
  • Primary: Sentinel-1 Synthetic Aperture Radar (SAR) imagery (C-band, VV polarization).
  • Validation/Comparison:
    • In situ soil moisture measurements from a network of 20 stations in central Italy (5 centimeter depth).
    • ERA5 Land reanalysis soil moisture (10 kilometer spatial resolution, 7 centimeter depth).
    • RT1 satellite soil moisture product (Sentinel-1 derived, 1 kilometer spatial resolution).
    • SMAP 1 kilometer soil moisture product (passive microwave, 1 kilometer spatial resolution).
    • GLEAM soil moisture product (satellite observations, surface soil moisture).
    • S1-Copernicus (S1-Cop) soil moisture dataset (Sentinel-1 derived).
  • Software: ISCE2 (NASA/JPL) for SAR image co-registration.

Main Results

• The novel InSAR algorithm efficiently processes long SAR time series with minimal computational cost, generating high-resolution soil moisture measurements for each acquisition.
• InSAR soil moisture products show strong Spearman correlations with modeled soil moisture (ERA5 Land, GLEAM) and alternative remote-sensing products (SMAP 1 km), particularly in areas characterized by high interferometric coherence.
• Performance is limited in regions with low interferometric coherence (e.g., due to dense vegetation or snow cover), complex topography, or urbanized areas.
• In central Italy, the InSAR soil moisture product performs satisfactorily against in situ observations, and against ERA5 Land, Spearman correlations range from 0.53 to 0.87.
• Spatially and temporally filtered InSAR soil moisture products show improved performance, with an effective resolution of 5 kilometers.
• In Spain, Turkey, Greece, and Libya, the InSAR soil moisture product performs better than RT1 and S1-Copernicus soil moisture products when compared to SMAP 1 km as a reference, with average correlations exceeding 0.75 in Spain and Turkey.
• In northern Italy, the RT1 soil moisture product performs better than InSAR and S1-Copernicus soil moisture.
• A low correlation (typically less than 0.46) was observed between InSAR soil moisture and backscatter-derived Sentinel-1 products (RT1, S1-Copernicus), suggesting a strong complementarity between the two retrieval approaches.
• A 3-year dataset for Spain demonstrated that the filtered InSAR soil moisture product accurately reproduces both seasonal and short-term soil moisture temporal variability, with Spearman correlations greater than 0.8 in most cases against ERA5 Land.
• The median Spearman correlation of InSAR soil moisture against ERA5 Land in Spain was 0.81, which is notably high compared to other Sentinel-1 derived products reported in scientific literature.

Contributions

• Presents a novel, computationally efficient algorithm for high-resolution soil moisture retrieval using SAR interferometry and closure phases, overcoming the quadratic computational complexity limitations of previous interferometric methods.
• Demonstrates the feasibility of deriving multi-year, high-resolution (1 kilometer, with a filtered version at 5 kilometers) surface soil moisture products from Sentinel-1 SAR data using a phase-based interferometric technique.
• Provides extensive validation of the new InSAR soil moisture product across seven diverse Mediterranean test sites, utilizing in situ measurements, land surface models, and other satellite-derived soil moisture products.
• Highlights the strong complementarity between phase-based (InSAR) and backscatter-based SAR soil moisture retrieval methods, suggesting potential for future integration to enhance product robustness and accuracy.
• Identifies key directions for future research, including adapting the algorithm for upcoming SAR missions (e.g., L-band NISAR, ROSE-L, or geosynchronous C-band platforms) and integrating backscatter information with phase data.

Funding

• ESA 4DMED-DEMETRAS project (contract number 4000136272/21/I-EF CCN. No 3)
• ESA Hydroterra+ Earth Explorer 12 Phase 0 Science and Requirements Consolidation Study (contract number 4000146182/24/NL/IB/ar)

Citation

@article{Zan2026Validation,
  author = {Zan, Francesco De and Filippucci, Paolo and Brocca, L.},
  title = {Validation of high-resolution surface soil moisture time series retrieved by means of SAR interferometry},
  journal = {Remote Sensing of Environment},
  year = {2026},
  doi = {10.1016/j.rse.2026.115266},
  url = {https://doi.org/10.1016/j.rse.2026.115266}
}