Ocean current estimation using ScanSAR data

It is possible to estimate the radial velocity of a target by comparing the Doppler centroid, as estimated from the return from the target, with a "reference" Doppler centroid. Of particular interest is the estimation of one component of an ocean current, where the target is an area of ocean a few square kilometres in extent. Studies using single beam SAR products have shown this technique to be effective in estimating the location of the Gulf Stream and the Loop Current in the Gulf of Mexico, but the size of these features means that a swath width of greater than 100 km would be desirable. This immediately suggests working with ScanSAR data, where a wider swath is formed by combining several beams. However, adapting the technique to such data presents a number of difficulties. The spectral information needed for Doppler estimation is not available from standard products, thus custom processing of raw data is necessary. For each beam we have burst mode rather than continuous data, which complicates the spectral shape of the target return but this problem can be avoided by the appropriate choice of azimuth patch size. Combining estimates from several beams involves problems with independent errors and beam edge corruption. The "reference" Doppler centroid must be estimated over a large range and azimuth extent. This estimation is accomplished by fitting a quasi-physical model, involving orbital information, satellite attitude and information about the satellite antenna. The fitting problem is examined, with specific reference to RADARSAT and ENVISAT. Examples of ocean current estimation using RADARSAT ScanSAR data are shown.