On the Feasibility of Radial Spatial Averaging to Improve Satellite Synthetic Aperture Radar Ocean Wind Retrievals

This work focuses on the feasibility of radial spatial averaging to improve satellite Synthetic Aperture Radar (SAR) ocean Horizontal Wind Speed (HWS) retrievals even in neutral atmosphere situations. The reference for comparison are in-situ HWS measurements from the FINO-1 offshore metmast. The data used from the Sentinel-1 SAR are the Level 2 Ocean (OCN) products for wind, wave, and current applications, retrieving the information from the Ocean Wind Field (OWI) component through Geophysical Model Function (GMF) CMOD-IFR2. The GMF allows for estimation of the equivalent wind speed at 10 meters and, therefore, stable and convective regimes can bias the estimations. The best agreement between SAR and FINO-1 was found -as expected- for neutral atmospheres, which yielded a determination coefficient of R
²
= 0.90 and a root mean squared error (RMSE) of 1.2 m/s. In non-neutral scenarios, statistical indicators showed that fine-tuned spatial averaging is also able to improve SAR HWS retrievals and yielded comparable statistical indicators (R
²
= 0.87, RMSE = 1.1 m/s) for the HWS estimates.