Computation of wind direction from SAR images without external a priori information

We present here the follow-up of a previously published work [1], where we described a wavelet based method to characterize the sea surface backscatter structures present in the SAR images. The method relies on the ability of the two-dimensional continuous wavelet technique to detect the spatial structure of the marine atmospheric boundary layer and to isolate wind-related cells and features. The analysis of the cells' geometry, moulded by the radiometric characteristics of the sea surface, permits the identification of the wind direction inside the cells and thus the computation of the wind speed through standard algorithms. About twenty SAR images (ERS-2 and ASAR Wide Swath) over the Mediterranean Sea have been analyzed, and the results compared with NSCAT and QuikSCAT satellites wind fields. These images cover a wide range of meteorological conditions, from low (2 m/s) to moderate winds (12 m/s), presenting many kinds of signature, i.e. wind cells, atmospheric gravity waves, convective structures and radiometric flatness. The main difference of this method, with respect to the majority of those already proposed, is that it does not require a-priori information about the wind direction as well as any periodicity of the backscatter structures. The aliased wind directions are estimated from the texture of the SAR reconstructed map, while the dealiasing is possible due to the asymmetries present in the detected backscatter structures. The resulting SAR derived wind fields have been compared with those provided by satellite scatterometers. Results indicate a good score in detecting the wind direction (≈ 70%). The developed methodology, once tested over an adequate quantity of images to derive statistically reliable results, could be routinely used to enrich SAR images with the wind field, as well as to characterize other backscatter structures displayed by SAR not depending directly on the wind.