Bayesian identification of oil spill source parameters from image contours

Area of Science:

• Environmental Science
• Remote Sensing
• Oceanography

Background:

• Oil spills present significant risks to marine and coastal ecosystems.
• Accurate identification of oil pollution sources is crucial for effective environmental management and incident investigation.
• Satellite imagery offers a valuable tool for monitoring and analyzing oil slick distribution at sea.

Purpose of the Study:

• To develop and validate a Bayesian approach for estimating oil spill source parameters from satellite-detected oil slick contours.
• To determine the location, time, duration, and quantity of oil released during a spill.
• To assess the method's performance using synthetic spill scenarios.

Main Methods:

• A Bayesian statistical framework was employed to model the inverse problem of spill source estimation.
• Oil slick contours derived from remotely sensed data were used as input.
• Two synthetic experiments were conducted: one with a fully observed contour and another with partially observed contours at different times.

Main Results:

• The proposed Bayesian method successfully estimated the five key spill parameters: 2D source coordinates, spill time, spill duration, and oil quantity.
• The method provided reliable estimates even with partial and time-delayed observations of oil slick contours.
• Quantified uncertainties in parameter estimates reflect the confidence in the results.

Conclusions:

• The Bayesian approach is effective for identifying oil spill sources using satellite imagery.
• This method enhances the capability to investigate unreported spills and manage marine pollution.
• The technique provides robust source parameter estimation with associated uncertainty quantification.