Water Eased Oil Removal in Gulf

The physical action of water may be partly responsible for the rapid disappearance of plumes of oil and gas formed during the 2010 Deepwater Horizon spill in the Gulf of Mexico.

Deep in the ocean, the plumes disappeared far more quickly than anyone expected. One reason is that Gulf water was already seeded with hydrocarbon-eating microbes that consumed oil and gas from natural seeps. Another reason, a new model suggests, is that the waters around the well mixed and mingled, giving microbial populations not just oxygen and other critical nutrients but also multiple opportunities to take a crack at the oil.

The official estimate from the spill’s Flow Rate Technical Group is that the ruptured well released 2 × 10⁸ gal of oil, or 5.4 × 10⁸ kg, into the Gulf from when the oil rig exploded on April 20, 2010, until the well was capped on July 15, 2010. Along with the oil came approximately 2 × 10⁸ kg of C₁–C₅ alkane gases, according to various estimates. Added to the oil and gas were 7.7 × 10⁵ gal of dispersant, which helped break up the oil into droplets.

It would take years for the hydrocarbons to disappear by natural means, scientists speculated. Yet in some cases, it took only a couple of months.

From the well, 1.5 km below the water surface, the mix of oil, gas, and dispersant rose, drawing in some water as it went. At around 1 km below the surface, a fraction of oil and gas peeled off, driven by density and solubility to form the so-called deep plumes.

Measurements of plume contents were puzzling: Different groups on different boats saw different ratios of methane to propane. And the plumes disappeared faster than anticipated.

A new model explains some of the mysteries. Developed by a group led by David L. Valentine, a professor of earth science at the University of California, Santa Barbara, the model takes into account microbial growth and metabolism, as well as water currents around a depth of 1 km (Proc. Natl. Acad. Sci. USA, DOI: 10.1073/pnas.1108820109).

The work is the first time scientists have combined physical water movement and mixing with models of bacterial growth, and it is an important addition to scientific understanding of microbial response to the spill, says Annalisa Bracco, a professor of oceanography at Georgia Institute of Technology.

According to the model, the oil spill led to staggered population blooms of bacteria that consume different types of hydrocarbons: first the species that chewed on C₃–C₁₅ alkanes, then the species that tackled ethane and methane, and finally the species that used the very long chains.

The model also shows that local currents played an important role. There was no constant stream of freshwater coming toward the well, with plumes stretching away in the opposite direction and building up microbial populations as they went. Instead, the water “sloshed and swirled” around the well, Valentine says, and hydrocarbon-consuming microbial populations that initially thrived on one feast from the well were sustained by repeated exposure to fresh oil and gas.

The phenomenon, which Valentine calls autoinoculation, explains some of the confusing measurements. Propane, for example, is consumed relatively easily and quickly by bacteria, so low amounts measured near the well in June were probably the result of degradation by recirculated bacteria.

Notably, bacterial growth wasn’t limited by oxygen or other essential nutrients, because as the water moved around, the bacteria mixed with water that wasn’t affected by the spill, Valentine says.

The model does have limits. In particular, with grid blocks of 4 km × 4 km × 300 meters, it depicts only an extremely thin horizontal slice of the ocean. Horizontal mixing in the ocean is generally more important than vertical mixing, Bracco says, but the underwater topography of the area likely means that vertical mixing played a significant role.

Richard Camilli, an ocean scientist at Woods Hole Oceanographic Institution, also sees the grid size as a limitation, but for a different reason. He and colleagues documented a plume that was about 2 km wide and about 35 km long, significantly narrower than the model’s grid. Even so, Camilli thinks the model does a good job of matching observational data.

Bracco further notes that the model incorporates a lot of variables. Inputs included 26 hydrocarbon compounds or classes of compounds plus 52 types of hydrocarbon-consuming microbes. And researchers lack good data on all of the microbes to feed into the model, she says.

Despite the limitations, Camilli adds, the model is helpful for understanding the Deepwater Horizon spill because it covers more space and time than was possible for the ships sampling the Gulf during the spill. Having such models available for future spills will help guide how scientists respond. “It shows a very heterogeneous distribution of hydrocarbons in the water and suggests that you really have to think carefully about where you should make your observations,” Camilli says.

Christopher M. Reddy, a marine chemist also at Woods Hole, highlights the work by Valentine and coworkers as a particularly nice example of the scientific process at work. The many papers that have come out over the past 18 months from various groups that worked in the Gulf give different snapshots of the spill. Making sense of that data, he says, “is a serious jigsaw puzzle” that the Valentine group put together.

Terry C. Hazen, an environmental microbiologist who holds a joint appointment on the faculty at the University of Tennessee, Knoxville, and at Oak Ridge National Laboratory, also praises Valentine and colleagues for pulling data from the Gulf into a bigger picture. “It clearly helps us understand things a bit better,” Hazen says, although he and Valentine disagree on how quickly the hydrocarbons degraded.

Even as this particular puzzle seems to come together into a coherent picture, it doesn’t mean that Gulf-spill-related science is coming to an end. Bracco is studying how the physical movement of water in the Gulf dispersed the oil. Camilli is developing robotic systems that will provide more real-time information to aid scientists in the field, now and in the case of future spills. Reddy is examining how oil mixed with coastal sand is degrading.

And Valentine and Hazen still have samples of water and sediment to analyze to better understand the microbial response: to identify the key species, probe how they reacted to the oil, and determine degradation mechanisms. “We have such a tremendous data set,” Hazen says, “I think there are a lot of fine details that we can still learn from the spill.”