THE IMPORTANCE OF SCATTEROMETER WINDS IN THE EXTRACTION OF WIND FIELD FROM SAR IMAGES

The wind field extraction from Synthetic Aperture Radar (SAR) images is still problematic due to the difficulty of inferring the wind direction without external information. The wind direction may be inferred by means of numerical models, or retrieved from the backscatter signature in the SAR image, using spectral analysis, spatial local gradient, or wavelet related methods [1,2,3,4]. Unfortunately, the variety of geophysical phenomena imaged by SAR often interferes with the selection of the features related to the wind, and even more in coastal areas and regional seas, where meteorology may be strongly influenced by local circulations and orography [5]. Thus, the validation of the wind fields obtained from SAR images against a truth data-set is the premise to acknowledge SAR winds methodologies. In our opinion, the only realistic data set to carry out this fundamental validation is delivered by scatterometer observations. Despite their resolution is coarser than that obtainable from SAR, they have been proved to be reliable, sometimes more than the atmospheric hindcasts by Limited Area Models [6]. An automatic near real time system to perform SAR wind evaluation from ENVISAT Advanced SAR (ASAR) Wide Swath (WS) SAR images has been set up at the premises of the Institute of Atmospheric Sciences and Climate (ISAC) in Padua, Italy. Here the ASAR images, downloaded from the European Space Agency (ESA), are analyzed and the sea surface wind vector computed by the 2-D Continuous Wavelet Transform (2DCWT) methodology, while scatterometer data are also collected to verify the SAR wind against remote sensing observations.