An explorative object-oriented Bayesian network model for oil spill response in the Arctic Ocean

This paper proposes an Object-Oriented Bayesian Network model to study oil spill mitigative measures using cost-effectiveness as a decision-making factor. The cost-effectiveness is defined as the ratio of dollar spent to the efficiency of oil removal in the spill area. The proposed model considers the complexity, lack of data, and uncertainties in an oil spill and response modelling scenario for Arctic shipping accident occurrence. The crude oil release, slick formation, weathering and transport, ecological impact, and response scenarios are modelled using Bayesian Network. Application of the model is demonstrated using a hypothetical scenario of an oil spill involving a ship in the Arctic region. The results show that a combination of in-situ burning, mechanical and manual recovery, and use of dispersants would cost the highest, while the use of in-situ burning, and dispersants gives the cheapest option. The study provides a deeper understanding of oil spill dynamics and effectiveness of the response techniques. The proposed model could serve as a useful tool for oil spill response decision-making.