Sentinel-1 soil moisture at 1 km resolution: a validation study
Anna Balenzano,Francesco Mattia,Giuseppe Satalino,Francesco P. Lovergine,Davide Palmisano,Jian Peng,Philip Marzahn,Urs Wegmüller,Oliver Cartus,Katarzyna Dąbrowska-Zielińska,Jan P. Musial,Malcolm W.J. Davidson,Valentijn R.N. Pauwels,Michael H. Cosh,Heather McNairn,Joel T. Johnson,Jeffrey P. Walker,Simon H. Yueh,Dara Entekhabi,Yann H. Kerr,Thomas J. Jackson
DOI: https://doi.org/10.1016/j.rse.2021.112554
IF: 13.5
2021-09-01
Remote Sensing of Environment
Abstract:<p>This study presents an assessment of a pre-operational soil moisture product at 1 <em>km</em> resolution derived from satellite data acquired by the European Radar Observatory Sentinel-1 (S-1), representing the first space component of the Copernicus program. The product consists of an estimate of surface soil volumetric water content Θ [<em>m</em><sup>3</sup>/<em>m</em><sup>3</sup>] and its uncertainty [<em>m</em><sup>3</sup>/<em>m</em><sup>3</sup>], both at 1 <em>km</em>. The retrieval algorithm relies on a time series based Short Term Change Detection (STCD) approach, taking advantage of the frequent revisit of the S-1 constellation that performs C-band Synthetic Aperture Radar (SAR) imaging. The performance of the S-1 Θ product is estimated through a direct comparison between 1068 S-1 Θ images against in situ Θ measurements acquired by 167 ground stations located in Europe, America and Australia, over 4 years between January 2015 and December 2020, depending on the site. The paper develops a method to estimate the spatial representativeness error (SRE) that arises from the mismatch between the S-1 Θ retrieved at 1 <em>km</em> resolution and the in situ point-scale Θ observations. The impact of SRE on standard validation metrics, i.e., root mean square error (RMSE), Pearson correlation (R) and linear regression, is quantified and experimentally assessed using S-1 and ground Θ data collected over a dense hydrologic network (4 − 5 <em>stations</em>/<em>km</em><sup>2</sup>) located in the Apulian Tavoliere (Southern Italy). Results show that for the dense hydrological network the RMSE and correlation are ~0.06 <em>m</em><sup>3</sup>/<em>m</em><sup>3</sup> and 0.71, respectively, whereas for the sparse hydrological networks, i.e., 1 <em>station</em>/<em>km</em><sup>2</sup>, the SRE increases the RMSE by ~0.02 <em>m</em><sup>3</sup>/<em>m</em><sup>3</sup> (70% Confidence Level). Globally, the S-1 Θ product is characterized by an intrinsic (i.e., with SRE removed) RMSE of ~0.07 <em>m</em><sup>3</sup>/<em>m</em><sup>3</sup> over the Θ range [0.03, 0.60] <em>m</em><sup>3</sup>/<em>m</em><sup>3</sup> and R of 0.54. A breakdown of the RMSE per dry, medium and wet Θ ranges is also derived and its implications for setting realistic requirements for SAR-based Θ retrieval are discussed together with recommendations for the density of in situ Θ observations.</p>
environmental sciences,imaging science & photographic technology,remote sensing