Bazzi et al. (2026) Observing irrigation using SWOT SAR Ka-band data from daily calibration and validation acquisitions

Identification

• Journal: Science of Remote Sensing
• Year: 2026
• Date: 2026-01-22
• Authors: Henri Bazzi, N. Baghdadi, Cécile Cazals, Sami Najem, Damien Desroches, Frédéric Frappart, Mehrez Zribi, Françoise Ruget
• DOI: 10.1016/j.srs.2026.100378

Research Groups

• UMR TETIS, University of Montpellier, AgroParisTech, INRAE, CIRAD, CNRS, Montpellier, France
• CS Group France, Toulouse, France
• CNES, Toulouse, France
• Interactions Sol Plante Atmosph`ere, UMR1391, Institut National de Recherche Pour l’Agriculture, l’Alimentation et l’Environnement, Bordeaux Science Agro, Villenave d’Ornon, France
• Universit´e de Toulouse, CESBIO, CNES, CNRS, INRAE, IRD, Toulouse, France
• G-EAU Unit, University of Montpellier, AgroParisTech, CIRAD, INRAE, Institut Agro, IRD, Salon de Provence, France

Short Summary

This study investigates the potential of SWOT Ka-band SAR data to detect irrigation events by analyzing its sensitivity to soil moisture variations over an irrigated grassland site. It found that SWOT Ka-band backscatter is sensitive to irrigation, showing an average increase of 4.3 dB on the day of irrigation, and its near-nadir incidence angle allows penetration through dense vegetation.

Objective

• To analyze SWOT's Ka-band sensitivity to soil moisture and determine if it can provide information on soil water status, particularly after irrigation episodes and during periods of well-developed vegetation.

Study Configuration

• Spatial Scale: Fifteen irrigated grassland plots at the "Domaine du Merle" in the Crau plain, southeast France. Geocoded pixel size of 16 m across longitude and 27 m across latitude.
• Temporal Scale: Daily SWOT Ka-band SAR acquisitions from 01 April to 07 July 2023 (Calibration/Validation phase). Irrigation events occurred every 8 to 12 days, lasting from less than an hour to 15 hours.

Methodology and Data

• Models used: No explicit hydrological or land surface models. Processing of SWOTL1BHRSLC data to derive backscatter coefficients (σ0) using specific equations and the PGEL2HRPIXC processing tool for geocoding. Google Earth Engine (GEE) was used for Normalized Difference Vegetation Index (NDVI) calculation.
• Data sources:
  • SWOT Level 1B High-Rate Single-look Complex (SWOTL1BHR_SLC) Ka-band SAR data (KaRIN sensor, near-nadir incidence angle 0° to 4.5°, with study area incidence angles from 3.2° to 3.5°).
  • Sentinel-2 (S2) time series data for Normalized Difference Vegetation Index (NDVI).
  • 30 m SRTM Digital Elevation Model (DEM).
  • Local meteorological station data (rainfall, temperature).
  • In-situ irrigation calendar and mowing dates for 15 grassland plots.

Main Results

• SWOT Ka-band SAR data is sensitive to soil moisture variation induced by irrigation.
• Irrigation events caused an average increase in backscattering of 4.3 dB on the same day of irrigation.
• Extreme increases in Ka-band backscattering (up to >11 dB) were observed in cases of flooded vegetation, attributed to specular reflection and/or double-bounce scattering mechanisms.
• One day after irrigation (following complete infiltration), an average backscattering increase of approximately 2 dB was observed, which returned to pre-irrigation levels two days later due to natural soil drying.
• Despite the Ka-band's short wavelength, SWOT's near-vertical incidence angle (0° to 4.5°) appears to enhance its ability to penetrate dense vegetation cover, allowing detection of soil moisture dynamics.
• The radiometric stability of calibrated SWOTL1BHR_SLC images was better than 0.5 dB (standard deviation of 0.42 dB for forest and 0.48 dB for urban areas).
• The sensitivity of SWOT backscattering to soil moisture under an average NDVI of 0.72 was approximately 0.125 dB per volume percent.
• Harvesting led to an average increase of 1.3 dB in SWOT backscattering, likely due to changes in surface roughness and reduced vegetation attenuation.

Contributions

• This is the first study to exploit the sensitivity of Ka-band SAR data (from SWOT) to soil moisture and irrigation events, particularly under dense vegetation cover.
• It demonstrates that SWOT's near-nadir incidence angle effectively compensates for the Ka-band's short wavelength, enabling soil moisture detection through dense vegetation.
• The findings open new perspectives for leveraging daily SWOT Calibration/Validation acquisitions for large-scale irrigation mapping and agricultural water management.
• The study proposes using SWOT Cal/Val data to create reliable training databases of irrigated/non-irrigated plots for machine learning models, overcoming current limitations in obtaining ground truth data for irrigation classification using other satellite missions (e.g., Sentinel-1 and Sentinel-2).

Funding

• French Space Study Center (CNES, TOSCA, 2025 project)
• National Research Institute for Agriculture, Food, and the Environment (INRAE)

Citation

@article{Bazzi2026Observing,
  author = {Bazzi, Henri and Baghdadi, N. and Cazals, Cécile and Najem, Sami and Desroches, Damien and Frappart, Frédéric and Zribi, Mehrez and Ruget, Françoise},
  title = {Observing irrigation using SWOT SAR Ka-band data from daily calibration and validation acquisitions},
  journal = {Science of Remote Sensing},
  year = {2026},
  doi = {10.1016/j.srs.2026.100378},
  url = {https://doi.org/10.1016/j.srs.2026.100378}
}