Controlled oil-on-water field exercises provide rare opportunities to validate marine pollution monitoring tools under realistic spill conditions. Here we use multi-temporal and multi-channel airborne and spaceborne synthetic aperture radar (SAR) observations from the NORSE experiment and the NOFO oil-on-Water campaigns in the North Sea to study (i) slick evolution and (ii) the separability of hazardous petroleum/mineral oils, including different emulsion states, from biogenic or vegetable look-alike films. The temporal analysis shows that spreading can be rapid at early stages, reinforcing the need for frequent satellite screening during response operations. At the same time, look-alike films can mimic low-backscatter oil signatures, so operational systems must retain enough polarimetric information to reduce false alarms. We therefore develop an unsupervised compact hybrid-polarimetric approach and introduce the Model-based Effective Reflection Coefficient (MERC), a physically motivated descriptor derived from hybrid-polarimetric scattering components. Across both airborne and satellite datasets, MERC improves oil detection and oil-type characterization compared with commonly used intensity and compact-polarimetric descriptors, supporting scalable, wide-swath SAR surveillance. These findings are relevant for routine monitoring of shipping- and offshore-related spills and for Arctic-adjacent waters, where remote location, limited accessibility, and extended periods of darkness increase reliance on all-weather satellite observations.
🌊
