Bounding the Oil Spill

Last May, Los Alamos ocean scientist Mathew Maltrud was taken aback as he viewed the results of his first supercomputer simulation of the Deepwater Horizon oil spill. The virtual dye he was using to trace the oil's path had been picked up by the swift Loop Current, exited the Gulf of Mexico, and then joined the powerful Gulf Stream flowing up the coast of North America and into the open Atlantic—all within a matter of weeks.

The speed of the currents was something of an eye opener. Having recently completed a global ocean simulation spanning more than a century, Maltrud was accustomed to taking a longer view of events. But at the time, less than a month after the real spill began, the simulation raised the possibility that the oil might wash up on the eastern seaboard or perhaps even reach the coastlines of Europe. The fact that crude oil was still gushing from the seafloor at an unknown rate, with no end in sight, only added to concerns about where the oil might go and when it might get there.

While that first simulation was dramatic, it was not a forecast. It depicted just one possible pathway for the oil. Currents in the Gulf of Mexico are too complex to be accurately predicted more than a couple of weeks into the future. On longer time scales, the dominant feature of the Gulf—the clockwise Loop Current—shifts position over hundreds of miles and sheds eddies at irregular intervals. Such eddies typically travel west and can last anywhere from weeks to months. Along with local winds, these factors alter the overall ocean dynamics in the Gulf in unpredictable ways.

Modeling the Deepwater Horizon oil spill

Modeling the Deepwater Horizon oil spill: The dispersion and dilution of a virtual dye 180 days after the initial spill, as predicted by one of an ensemble of simulations. The color scale shows the dilution factor, the ratio of total amount of dye in the water column to the amount injected at the source.

So to get a handle on the long-term fate of the oil, Maltrud and Los Alamos colleague Phil Jones, with collaborators from the National Center for Atmospheric Research and the University of Kiel in Germany, decided to take a statistical approach. Dozens of simulations were carried out, each beginning with a different but realistic Loop Current scenario, as a way to bracket the ever-changing conditions in the Gulf. When all the simulations were complete, the team averaged the results of the individual simulations to establish where the oil would likely travel.

The scientists used the Parallel Ocean Program, the state-of-the-art ocean circulation model developed at Los Alamos, which divides the world's oceans and seas into computational cells covering a tenth of a degree of longitude and latitude, small enough to rigorously model the Loop Current and associated eddies. The simulations were run on the state of New Mexico's Encanto supercomputer and Oak Ridge National Laboratory's Jaguar supercomputer. On Jaguar, the second fastest computer in the world at the time, each simulation took two or three days to complete.

To simulate the spill, virtual dyes were continuously added at multiple depths above the Macondo well site of the spill for either two or four months, and the simulations were continued for a total of six months or a year. The virtual dyes served as passive tracers, much like food coloring, possessing none of the physical characteristics of real oil. By neglecting coagulation, surface-slick formation, decay, and other processes that would tend to reduce the spread of oil, the team could interpret their predictions as an upper bound for the oil dispersal.

Even with this caveat, the researchers were able to draw a number of timely conclusions that have held up well. No oil was expected to reach the shores of Texas, Mexico, Cuba, or the Bahamas within six months of the initial spill. While some fraction of the oil would almost certainly escape the Gulf, it would be so diluted that large amounts would be unlikely to wash up on the eastern coast of the U.S. or Canada. Oil would be even less likely to reach Europe in detectable amounts. The finding that virtual dye injected below a depth of 800 meters remained within a few hundred miles of the spill is consistent with reports of deep underwater oil plumes lingering in the northern Gulf region.

Despite a brief media flurry when released to the public last June, these conclusions were at least somewhat reassuring in that they suggested that the oil from the Deepwater Horizon spill, which would become the largest marine spill in history, was not likely to cause major deleterious effects beyond the Gulf. This work was published in Environmental Research Letters in August 2010 and serves as just one example of peer-reviewed research that addressed the Gulf oil spill by harnessing capabilities developed for other Department of Energy missions. At the direction of Secretary of Energy Steven Chu, an intense multi-Laboratory effort was formally made to expedite the capping of the Macondo well.

—Craig Carmer


In this issue...

+ More on this article


Contact Us | Careers | Bradbury Science Museum | Emergencies | Inside LANL | Maps | Site Feedback | SSL Portal | Training

Operated by Los Alamos National Security, LLC for the U.S. Department of Energy's NNSA © Copyright 2014 LANS, LLC All rights reserved | Terms of Use | Privacy Policy