Advancing Marine Oil Spill Detection in the Arabian Gulf: A Synergistic Approach Using SAR and Optical Satellite Imagery

Detecting oil spills in the Arabian Gulf (AG) is crucial for protecting marine ecosystems and enabling rapid response to pollution events. This study developed an integrated framework for oil spill detection combining Sentinel-1 SAR and Sentinel-2 optical imagery. The novelty lies in deriving the Enhanced Oil Spill Index (E-OSI) from SWIR bands B10 and B11 to overcome spectral overlap between oil and turbid water where traditional visible and near-infrared indices show reduced performance. Validation was conducted through expert-interpreted reference samples encompassing oil spills, ship wakes, and clean water surfaces. The framework integrated SAR Vertical-Vertical (VV) polarization for all-weather monitoring with E-OSI for enhanced optical discrimination, tested across three documented spills (Al-Khafji 2017, Sharjah 2017, Kharag 2023) and extended to the northern Arabian Gulf (NAG) during July 2023 for comprehensive detection. Performance evaluation demonstrated SAR accuracy of 93% (κ = 0.86) and E-OSI accuracy of 88% (κ = 0.82), both exceeding 95% specificity in discriminating look-alike features. Analysis of documented cases revealed variable spill behavior, Al-Khafji covered 130.98 km² spreading 100 km coastward over six days, Sharjah spanned 94.90 km² moving toward sensitive coral habitats, while Kharag decreased from 43.74 to 8.42 km² within three days. Beyond validation, regional detection covered 144.16 km² of unreported spills concentrated near Manifa, Safaniyah production zones (58.72 km²), Bushehr offshore areas (46.35 km²), and major shipping routes (39.09 km²). Spatial correlation analysis linked spill distribution to vessel traffic patterns, while environmental overlay identified critical exposure zones with threatened coral reef ecosystems. The integrated SAR-optical framework demonstrates a validated, scalable, and operationally feasible approach for oil spill detection, area quantification, and environmental risk assessment in data-limited marine regions.