A Bayesian model to predict oil spill consequences of management plans in the Gulf of Mexico

Conference · Tue Dec 31 04:00:00 EST 1996

OSTI ID:478361

Obie, D S; Englehardt, J ^[1]

---

1. Univ. of Miami, Coral Gables, FL (United States)

A Bayesian risk analysis model, comprising of a release assessment module and an exposure assessment module for the oil transportation system in the Gulf of Mexico is described in this paper. The model is used to compute probability distributions for oil spill quantities for 160 grid cells in the Gulf of Mexico, and the volumes of that oil to reach 58 coastline segments over a user-specified planning period. In addition to historical oil spill data, the model can accept subjective information on management alternatives involving changes in the oil transportation system. For example, volumes, tugboat escorts, mechanical equipment and hull design can be altered, and user confidence can be entered concerning how changes will effect spill number and size. The release assessment module uses a predictive Bayesian negative binomial distribution for spill number, and a predictive Bayesian distribution based on the Pareto I distribution for spill size. Conditional transport probabilities developed by the Minerals Management Service and the results of the release assessment module were used in the exposure assessment module. Oil spill data maintained by the US Coast Guard for the years 1991-1995 were analyzed along with two basic oil transportation management scenarios.

• OSTI ID: 478361
• Report Number(s): CONF-961119--
• Country of Publication: United States
• Language: English

Related Subjects

02 PETROLEUM
ENVIRONMENTAL IMPACTS
FORECASTING
MATHEMATICAL MODELS
OIL SPILLS
REMEDIAL ACTION
RISK ASSESSMENT