Thermal front retreivals from SAR imagery

摘要

Based on linear statistical relationships between components of SST gradients and wind stress variations, a high-resolution methodology is presented to retrieve Gulf Stream thermal front features using only variations in pixel-scale features of the SAR-derived wind stress divergence and curl fields, representing a significant improvement in methodology. It is important to remove small-scale features in divergence and curl wind stress images before they are used to construct the thermal front parameter, TF. We also verified the results with another 42 RADARSAT-2 images acquired at dual-polarization (VV, VH) image mode in the Gulf Stream region. Results indicates that the proposed method works well when retrieved wind speed lies between 5 m/s and 12 m/s, because SST-induced wind gradients can modify the vorticity and divergence fields [O'Neill et al. 2010].