Operational oil-slick characterization by SAR imagery and synergistic data

A methodology is proposed for the semiautomatic detection, characterization, and classification of slicks detected in C-band Synthetic Aperture Radar (SAR). For the first detection step, automatic algorithms were tested on Environmental Research Satellite (ERS) and Environmental Satellite (EnviSat) images acquired during the Prestige tanker accident. These tests reveal that simple filter or segmentation methods efficiently detect slicks with high contrasts and simple shapes, while a new and more complex multiscale method is able to detect a wider range of slicks. The characteristics of automatically detected slicks are then combined with meteooceanic data in order to eliminate slicks related to wind anomalies and current fronts. The data suggest that slicks in cold upwelling waters are natural, and confirm that slicks are heavy oils when high sea states are present. This detection-classification methodology is validated with aircraft slick-tracking maps. In most cases, joint SAR and environmental data are sufficient to classify the slicks.