INFERRING MULTISCALE STRUCTURE IN ATMOSPHERIC-TURBULENCE USING SATELLITE-BASED SYNTHETIC-APERTURE RADAR IMAGERY

This paper describes the multiscale structure in synthetic aperture ra dar (SAR) backscatter seen in an image of a cold-air outbreak as being due to the influence on the surface of the ocean of three scales of i nteracting atmospheric turbulence. A low-resolution (pixels have side lengths of 100 m) SAR image shows the first class of backscatter patte rn to be broken lines of enhanced radar backscatter (sigma), of order 1 dB above the background, with an average cross-wind scale of 5 km an d downwind lengths of many tens of kilometers. These are associated wi th quasi-two-dimensional roll vortices. The second class of radar back scatter pattern is two-dimensional backscatter regions whose linear ag gregrations constitute the lines of backscatter. They are, on average, 1 km wide in the cross-wind direction and are 2.5 km long in the down wind direction. A higher-resolution image (pixels have side lengths of 12.5 m) of the same area, shows a third class of structure: smaller-s cale regions of enhanced backscatter up to several hundred meters in l ength downwind and greater than 190 m across wind that are associated with 7-dB variations in sigma in the unaveraged image. The modulated, fine-scale backscatter structure compares favorably in both scale and effect to observations of the modulation of ''microfronts'' in the atm ospheric surface layer by kilometer-scale ''inactive eddies,'' the lat ter being of uncertain origin. Therefore the hypothesis offered here i s that the fine-scale SAR structure is caused by the influence of atmo spheric microfronts on the ocean surface, while the two-dimensional, k ilometer-scale backscatter features are the signature of inactive eddi es linearly aggregated by or possibly intrinsic to the roll vortices.