Bayesian Analysis of Oil Spill Statistics

This paper takes a two-step approach to the problem of estimating the likelihood of oil spill frequency and magnitude associated with a given amount of oil transport and production activity. The paper uses Bayesian reasoning plus actual spill data gathered from a variety of sources to generate (a) the density of the number of spills which will occur in the life of a project and (b) the density of the amount of oil which will be spilled in an individual incident. These two densities are then combined to obtain the density of the total volume of oil which will be spilled over a project's life.