Arctic and Northern Offshore Oil Spill Probabilities

Abstract

Current catastrophic consequences of the Gulf of Mexico blowout have refocused interest on the probabilities of such events in both temperate and northern regions. This paper reviews some of the early studies on oil spill probabilities with emphasis on oil blowouts, and details more recent studies carried out specifically for the Alaskan OCS. Due to the embryonic state of offshore oil development in arctic regions, which has been the case since 1976 to the present, it is not possible to base oil spill probability estimates on empirical data. The early studies relied on a detailed fault tree analysis dealing with the operations as systems without history. More recent studies in northern but not arctic operations use world wide data as a starting point. In the recent and current Alaskan OCS studies, statistically significant non-Arctic empirical data from the US Gulf of Mexico and world-wide sources, together with their variance, were used as a starting point. Next, both the historical non-Arctic frequency distributions and spill causal distributions were modified to reflect specific effects of the Arctic setting, and the resultant fault tree model was evaluated using Monte Carlo simulation to adequately characterize uncertainties treated as probability distribution inputs to the fault tree.