Wind Speed Retrieval Based on Sea Surface Roughness Measurements from Spaceborne Microwave Radiometers

Abstract Wind speed is the main factor responsible for the increase in ocean thermal emission because sea surface emissivity strongly depends on surface roughness. An alternative approach to estimate the surface wind speed (SWS) as a function of surface roughness is developed in this study. For the sea surface emissivity, the state-of-the-art forward Fast Microwave Emissivity Model, version 3 (FASTEM-3), which is applicable for a wide range of microwave frequencies at incidence angles of less than 60°, is used. Special Sensor Microwave Imager and Advanced Microwave Scanning Radiometer (AMSR-E) observations are simulated using FASTEM-3 and the Global Data Assimilation and Prediction System operated by the Korea Meteorological Administration. The performance of the SWS retrieval algorithm is assessed by comparing its SWS output to that of the Global Data Assimilation System operated by the National Centers for Environmental Prediction. The surface roughness is computed using the Hong approximation and characteristics of the polarization ratio. When compared with the Tropical Atmosphere–Ocean data, the bias and root-mean-square error (RMSE) of the SWS outputs from the proposed wind speed retrieval algorithm were found to be 0.32 m s −1 (bias) and 0.37 m s −1 (RMSE) for the AMSR-E 18.7-GHz channel, 0.38 m s −1 (bias) and 0.42 m s −1 (RMSE) for the AMSR-E 23.8-GHz channel, and 0.45 m s −1 (bias) and 0.49 m s −1 (RMSE) for the AMSR-E 36.5-GHz channel. Consequently, this research provides an alternative method to retrieve the SWS with minimal a priori information on the sea surface.