Bridging the gap between cyclone wind and wave by C-band SAR measurements

Active microwave remote sensing of hurricane-strength wind is a challenging task due to the saturation of col-polarization backscattering signal under such condition. Here we take advantage of the fact that wind sea wave growth does not saturate at high wind and the intrinsic relationship among wind-wave triplets (sea surface wind speed, significant wave height, and peak wave period) within a tropical storm to derive the wind speed. Three Sentinel-1 (S-1) and nine RADARSAT-2 (R-2) C-band synthetic aperture radar (SAR) images acquired between 20 and 40 m/s winds are collected in this study. The S-1 and R-2 SAR-derived winds are compared with those measured by coincident National Oceanic and Atmospheric Administration Stepped-Frequency Microwave Radiometer (SFMR) and simulated by Symmetric Hurricane Estimates for Wind (SHEW) model. Validations against SFMR winds for S-1 show Root-Mean-Square Error (RMSE) of 1.7 m/s with a 0.2 m/s bias at the left side of cyclone centers and RMSE of 2.9 m/s RMSE with a 0.56 bias at the backside of cyclone centers. R-2 SAR-derived winds against SHEW model results show a RMSE of 2.4 m/s with a 0.3 m/s bias and 2.6 m/s with a 0.35 m/s bias at the right side and the left side of cyclone centers, while the RMSE is 3.9 m/s with a 0.1 m/s bias at the backside of cyclone centers. The wave-information-based wind retrieval method works well at the left side and right side, but less accurately at the backside of a tropical storm when wind wave and swell are mixed.