Title of article :
Sensitivity testing for improved efficiency of a statistical oil-spill
risk analysis model
Author/Authors :
James M. Price a، نويسنده , , ?، نويسنده , , Walter R. Johnson a، نويسنده , , Zhen-Gang Ji، نويسنده , , Charles F. Marshall a، نويسنده , ,
Gail B. Rainey b، نويسنده ,
Issue Information :
ماهنامه با شماره پیاپی سال 2004
Abstract :
The US Minerals Management Service performs oil-spill risk analyses using, in part, a statistical model of hypothetical oil-spill
trajectories. The “Oil-spill Risk Analysis” (OSRA) model is driven by analyzed sea surface winds and model-generated ocean
surface currents. The OSRA model calculates thousands of oil-spill trajectories over extended areas of US waters and tabulates the
frequencies with which the simulated oil spills contact the geographic boundaries of designated natural resources within a specified
number of days after the simulated spill events are initiated.
The limits of computer mass storage and speed impose constraints on the number of oil-spill trajectory simulations and the time
step of the numerical integration generating the spill trajectories. Also, the model winds and ocean currents have limited spatial
and temporal resolution, producing sensible limits on the spatial density of the simulated oil spills and on the numerical integration
time step. This investigation attempted to determine the smallest practical spatial interval between simulated oil spill release sites
and the shortest integration time step beyond which the frequencies of oil-spill contact essentially do not change. In addition, we
investigated the number of hypothetical oil spills initiated per day from a single location and two methods of numerical integration
to generate the spill trajectories—simple forward time stepping and the fourth order Runge–Kutta method.
Applying the OSRA model to the Gulf of Mexico using 9 years of winds and ocean current data, we found a smallest distance
between simulated spills of 6 nautical miles and a shortest integration time step of 1 h using the fourth order Runge–Kutta integration
method. A 20-min time step would be needed with simple forward time stepping. Initiating hypothetical spills eight times per day
produced essentially the same contact probabilities as initiating spills once per day. These intervals were practical for OSRA model
runs over the US portion of the Gulf of Mexico using commonly available desktop personal computers.
Keywords :
Oil spill model , Statistical oil spill model , Model sensitivity test , Trajectory model , Oil spill risk analysis , Statistical oil spill riskanalysis , Oil spill risk assessment , Statistical oil spill risk assessment
Journal title :
Environmental Modelling and Software
Journal title :
Environmental Modelling and Software