[Paper Review] Managing large oil Spills in the Mediterranean

J. A. Jiménez Madrid, Antonio Garcı́a-Olivares+2 more|arXiv (Cornell University)|Oct 1, 2015

Oil Spill Detection and Mitigation15 references3 citations

TL;DR

This paper proposes probabilistic beaching maps for large oil spills across the Mediterranean Sea using Lagrangian particle simulations driven by a hindcast ocean circulation model. By simulating oil spill trajectories from 185 source points—mainly tanker routes and offshore platforms—it estimates the likelihood and timing of oil reaching specific coastal segments, offering long-term risk assessment to complement short-term spill forecasts and improve emergency decision-making.

ABSTRACT

For the first time a statistical analysis of oil spill beaching is applied to the whole Mediterranean Sea. A series of probability maps of beaching in case of an oil spill incident are proposed as a complementary tool to vulnerability analysis and risk assessment in the whole basin. As a first approach a set of spill source points are selected along the main paths of tankers and a few points of special interest related with hot spot areas or oil platforms. Probability of beaching on coastal segments are obtained for 3 types of oil characterised by medium to highly persistence in water. The approach is based on Lagrangian simulations using particles as a proxy of oil spills evolving according the environmental conditions provided by a hincast model of the Mediterranean circulation.

Motivation & Objective

• To develop long-term probabilistic assessments of oil spill beaching in the Mediterranean Sea, complementing short-term forecasting systems.
• To identify the most vulnerable coastal segments to large oil spills based on historical ocean circulation patterns.
• To support decision-making in oil spill emergencies by providing statistical likelihoods of oil reaching specific coastlines.
• To evaluate the long-term transport pathways and arrival times of oil slicks to sensitive coastal areas.
• To create a decision-support tool for marine safety and environmental protection authorities in the Mediterranean region.

Proposed method

• Lagrangian particle simulations are used to model oil spill trajectories, treating particles as proxies for oil slicks.
• Simulations are driven by a high-resolution hindcast ocean circulation model (NEMO-MED12) of the Mediterranean Sea.
• 185 source points are selected along major tanker routes and near offshore oil platforms and terminals.
• Three oil types with medium to high persistence in water are simulated to assess beaching probabilities.
• Beaching probability is calculated as the fraction of particles reaching each coastal segment over time.
• Results are aggregated into probability maps showing the percentage of spilled oil expected to reach each coastal segment, along with arrival timing.

Experimental results

Research questions

• RQ1 What is the long-term probability that an oil spill from a given source will reach specific coastal segments in the Mediterranean?
• RQ2 How do arrival times of oil slicks vary across different coastal areas following a large spill?
• RQ3 Which coastal regions are most vulnerable to oil spill beaching based on historical ocean circulation patterns?
• RQ4 How do different oil types (with varying persistence) affect the likelihood and timing of beaching?
• RQ5 To what extent can statistical beaching probability maps improve decision-making in oil spill emergencies?

Key findings

• The methodology successfully generates probabilistic beaching maps for 740 simulations across 185 source points, covering the entire Mediterranean basin.
• For a 30,000-ton spill, coastal segments can expect up to 30 tons per kilometer of oil deposition if the beaching probability is 5%.
• Arrival times of oil slicks to different coastlines range from as early as 3 days to as late as 16 days after a spill event.
• The maps identify key transport barriers and persistent pathways in the Mediterranean circulation that influence long-term spill dispersion.
• The results are intended for integration into the MEDESS-4MS Decision Support System and will be publicly available online.
• The approach is robust to model limitations and can be updated with improved hindcasts, such as the new Copernicus reanalysis, to enhance accuracy.