Probabilistic estimates of oil spill occurrences are used in the development of environmental assessments for hypothetical developments in the US Chukchi and Beaufort Seas. Due to the limited offshore oil development in this region, it was not possible to base these oil spill probability estimates on empirical data from that region. Rather, statistically significant non-Arctic empirical data from the US Gulf of Mexico (GOM) 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. A series of studies, associated with successive lease sale scenarios between 2000 and 2006, was carried out directed at developing and applying the fault tree methodology. In addition, a study directed solely at updating the GOM data was carried out. The series of studies consisted of five Beaufort Sea and/or Chukchi Sea application studies and the GOM data update studies. This report summarizes the methodology and gives results of its application to the estimation of oil spill probabilities and their characteristics for the Chukchi and Beaufort Seas region based on the most recent studies and statistics through 2006 Ongoing studies will incorporate OCS data through 2010 and will capture the loss of well control frequencies including the Macondo well.

This content is only available via PDF.
You can access this article if you purchase or spend a download.