Recent results in model-based wind retrieval

From multiple measurements of the normalized radar backscatter (/spl sigmaspl deg/) made by a spaceborne scatterometer, the near-surface wind over the ocean can be inferred using a geophysical model function relating /spl sigmaspl deg/ and the vector wind. Recently, a model-based retrieval technique has been developed. The technique avoids many of the problems associated with traditional point-wise wind retrieval and ambiguity removal. The model-based retrieval can also produce vorticity and divergence fields as auxiliary products. In this paper the method is applied to ERS-1 scatterometer data. Comparisons in performance between the traditional two-step pointwise wind retrieval/ambiguity removal method and the model-based retrieval method are presented using simulated and actual ERS-1 measurements. The results suggest that model-based retrieval can produce more accurate estimates of the wind field than pointwise wind retrieval-particularily in low wind speed regions where the C-band model function results in low wind accuracy for traditional wind estimation. The authors consider the spectra of the wind and wind vorticity over scales of from 50 to 1000 km.<>