Dispersion monitoring services in the Mediterranean Sea: A multi-model statistical approach

The Mediterranean Sea is increasingly impacted by shipborne pollution, necessitating sophisticated oil spill monitoring systems. This study presents a detailed framework for assessing oil spill dynamics in the Mediterranean using various oceanographic models. The first part compares oil dispersion results using sea surface current data from Copernicus Marine Environment Monitoring Service (CMEMS), Naval Hydrographic and Oceanographic Service (SHOM), and French Research Institute for the Exploitation of the Sea (IFREMER). The second part evaluates particle dispersion simulations, improving upon previous results by incorporating Stokes drift and wind drag effects from the Copernicus ERA5 and the MeteOcean model, operational at DICCA Unige. The study also calculates the centroid of each oil slick to track its trajectory over time and assessing its spatial and temporal deformation. To validate the methodology, a practical case study is conducted, comparing the results with Sentinel-1 satellite images from the first days after the accident and with press reports. The results demonstrate that both Stokes drift and wind significantly influence the surface transport of oil spill, improving trajectory predictions when combined with surface currents. In addition, comparison between ocean and climate models revealed substantial variability in dispersion results, highlighting the importance of model selection in operational forecasting or the benefits of the use of a multi-model probabilistic approach.