This research proposes a physical algorithm for retrieving sea surface wind speed using a combination of passive microwave satellite observations and simulations, namely Advanced Microwave Scanning Radiometer (AMSR-E) observations for vertical polarization, and a simulation based on the fast Radiative Transfer for TOVS with Global Data Assimilation System data for horizontal polarizations. A regression relationship between sea surface roughness and sea surface wind speed was derived using the Hong approximation and polarization ratio near Brewster's angle. The sea surface wind speed algorithm was assessed using the AMSR-E data. Comparing sea surface wind speed from the proposed algorithm using AMSR-E 6.925, 10.65, and 18.7 GHz data with Tropical Atmospheric Ocean (TAO) data from 1 January 2010 to 31 December 2010 shows that the estimated bias and root mean square error (RMSE) were -0.126 and 1.191ms(-1), -0.094 and 1.152ms(-1), and -0.085 and 1.338ms(-1), respectively. The current official AMSR-E sea surface wind speed shows the bias and RMSE were -0.614 and 1.330ms(-1), respectively, when compared with TAO buoy observations, which was similar to those of the proposed algorithm using AMSR-E 18.7 GHz. This study provides a useful sea surface wind speed retrieval algorithm with high accuracy on the sea surface, which can be applied to a variety of spaceborne passive microwave radiometers.
