An automatic procedure for improving sea surface wind field estimation from X-band COSMO-SkyMed SAR data

Sea surface wind field variations are associated with the interaction between ocean surface and atmospheric phenomena. In this context the observation, monitoring and quantification of wind field over sea surface is of great relevance to better understand, characterize and predict behavior of atmospheric variables involved in these processes. Microwave remote sensing has a great impact for the study of the ocean-atmospheric interaction processes and the estimation of sea surface wind field. Synthetic Aperture Radar (SAR) is a sensor that provides high resolution day- and night-time remotely sensed measurements of the ocean surface that range between few meter and more than hundred meters. The large spatial-temporal coverage provided by SAR makes this sensor a very valuable tool for measuring geophysical parameters such as ocean surface wind field. The development of several SAR constellations greatly enhanced both space and time SAR observation of the sea surface providing that reliable methodologies for wind field extraction from SAR images are available. In particular, the use of X-band COSMO-SkyMed SAR data is highly innovative and is of great relevance in the modeling of oceanographic processes. The system consists of a constellation of four LEO mid-sized satellites, each equipped with a multimode high-resolution SAR operating at X-band, which ensures both a wide area coverage and a small revisit time. In order to supply for a wide variety of applications the COSMO-SkyMed SAR payload can acquire a scene in different modes. For the purpose of such a study only the ScanSAR HugeRegion mode will be accounted for, which ensures a wide image area coverage. Unfortunately, such an acquisition mode makes the wind field estimation a non-trivial task due to the presence of SAR processing artifacts, which are not related to geophysical phenomena. Moreover, wind field estimation and the modeling of related oceanographic processes can be corrupted by the presence of atmospheric fronts that strongly affect X-band SAR data. The main purpose of this paper is to outline an automatic processing chain which improves SAR image data quality making such data suitable for several applications such as wind field estimation, in which processing artifacts and atmospheric fronts provide misinterpretation.