It's the most tangible intangible disaster of the past decade.
Following the undersea blowout of the Deepwater Horizon two years ago, a majority of the oil and gas escaping from the rig's out-of-control well never surfaced. Instead, it flowed in a diffuse layer to the southwest, thousands of feet below sea level. Largely invisible, this snaking "plume" nevertheless entered the imaginations of millions of people -- at least until its demise to the Gulf's vast size and a host of hungry microbes.
A compelling image -- but it never happened. At least, not the way scientists imagined.
Rather than flowing in a tidy path to the southwest, pulled along by a steady current, the Deepwater Horizon plume was a mess of swirl and slosh. Virgin water exposed to the spill, rather than whisking away permanently, would return after weeks, carrying with it microbes already primed to chew hydrocarbons, according to a study published yesterday in the Proceedings of the National Academy of Sciences.
The study presents a unified theory of the plume, its model results matching many of the often contradictory observations made by scientists during the first months of the spill. Understanding the bathtub circulation of the Gulf of Mexico suddenly sorted these findings into a comprehensible whole, said Dave Valentine, the paper's lead author and a microbial geochemist at the University of California, Santa Barbara.
"There's no perfect way of explaining something as amorphous and ever-changing as this was," he said. It is almost irreducibly complex, he added. "It's almost like an enclosed bay. It's not a simple current where things move from point A to point B."
The Valentine study comes at a crucial time for Gulf research. A government study published last month confirmed that nearly half of the oil -- and almost all of the gas -- released from the BP well likely remained trapped in deep waters. In all, some 33,000 barrels of oil a day remained in the deep, the study found, an estimate in line with a chemical study of the oil's fate also released yesterday.
Folding these mature estimates of the released oil, along with evidence of microbial degradation, into a plausible theoretical framework is essential to the government's ongoing investigation of the spill's environmental damage, according to NOAA Administrator Jane Lubchenco, who found time from her high-profile job to edit Valentine's study.
"These results may help us better understand the variability in the rapid rates of hydrocarbon consumption by bacteria in the plume, as observed by several groups of researchers," she said in a release to Greenwire, "while our scientists continue to examine the impacts of the Deepwater Horizon spill on the Gulf ecosystem."
This is not Valentine's first foray into the plume. Previously, his work uncovered the large amount of gas that remained trapped underwater during the spill (Greenwire, Sept. 17, 2010). Valentine also found that, much to his amazement, the recalcitrant methane had vanished, degraded by bacteria, during a follow-up cruise in the early fall (Greenwire, Jan. 7, 2011).
Scientists who studied the plume found Valentine's model convincing. While it did not match every observation perfectly, and its resolution was somewhat coarse, those are simply improvements that can be made on what seems like a foundational step.
"Their approach is holistic and does an excellent job of explaining large-scale patterns observed in the Gulf of Mexico following the spill," said John Kessler, a chemical oceanographer at Texas A&M University and one of the plume's chief researchers.
"This is probably a slam-dunk understanding of how the plume worked," added Chris Reddy, a chemist at the Woods Hole Oceanographic Institution. "The plume activity is a lot more complicated than we really thought."
Reddy was part of the Woods Hole team that, early on, helped shape perception of the plume with a report published in Science the month after BP's well was capped (Greenwire, Aug. 20, 2010). They described what seemed like a diffuse cloud of hydrocarbons -- on average, the plume had a concentration of 1 part hydrocarbon to every million parts water -- lurking underwater, stretching over an area the size of Manhattan.
Their view then, and really until Valentine's study, was that the oil came out of the well "and took a right-hand turn," Reddy said. It was a simplistic idea, in retrospect, he said.
The model helps explain several confounding findings, added Rich Camilli, the lead author of the Woods Hole study. Their cruise arrived at the Deepwater Horizon site just before warnings of an incipient hurricane. And while they saw signs of hydrocarbons to the well's northeast, the lead was not strong enough.
"We had limited time on site, limited resources and a hurricane coming at us," Camilli said with some regret. "We had to focus our energies. And we focused on the southwest because it seemed to be a bigger signal."
The most important finding from Valentine's model was its discovery of microbial priming, several scientists said. For deep waters not previously exposed to the spill, the carbon-hungry bugs followed a predictable pattern, one species after another blooming to consume its favored hydrocarbon, said Terry Hazen, the microbiologist who gained fame by identifying an oil-eating bug feasting on the plume (Greenwire, Aug. 24, 2010).
In unexposed water, the easiest-to-digest hydrocarbons would go first, Hazen said. To put it in human terms: "The candy went away first," he said. "Then we got into the meat and potatoes. And then we got into the gristle."
This pattern changed once water previously exposed to the spill, after sloshing in deep spirals that could stretch for 50 miles, returned to the wellhead. Their bibs already on, the host of microbes began eating the candy (propane), meat (alkanes) and gristle (aromatic hydrocarbons) all at the same time. It was a smorgasbord.
Valentine suspects this priming dynamic happens all the time in waters home to oil and gas seeps. But no one has been able to find it, he said, "largely because we've never the controlled release [necessary] -- or in this case, an uncontrolled release."
The layering of old and new water also explains observation differences recorded by the Woods Hole group, Valentine and Samantha Joye, a biochemical oceanographer at the University of Georgia. Both the Woods Hole group and Joye had found similar ratios of propane to methane in their samples, while Valentine had contradictory data.
"And Dave is not a hack," Reddy said. "We were going, 'How can we have this discrepancy? Our data was solid.' We used a lot of brain power trying to figure out why Dave's data was different."
Reddy even began giving presentations about the differences between Valentine and Joye's data to study confusion about the plume. Then, finally, Reddy and Valentine were sitting together on another ocean research cruise, and Reddy remarked, "Dave, how do we figure out this propane shit?" In 30 seconds, Valentine sketched out his new model, where consumption rates would vary with old and new water.
"He really unified theories about the plume," Reddy said.
Navy models play key role
That unification could not have happened without some elaborate modeling, however, including heavy mathematical lifting by one of Valentine's co-authors, Igor Mezic, an engineer at Santa Barbara.
Mezic adapted the Gulf models used by the Navy to keep their gliders from running into the seafloor, adding mixing diagnostics he had previously applied to describing the oil's movement on the surface. Combined with the huge amount of data recorded during the spill -- 10 times the normal amount -- a model that is typically used for short-term predictions becomes far more rigorous.
"It's an approach that really showed where the action was," Valentine said.
The group then seeded the physical model with both the hydrocarbons erupting from the well and 52 theoretical bacteria types, each tuned to a different feedstock. Tracking the movement of these bugs, which included exemplars of the microbes previously discovered in the plume by Hazen and Valentine, revealed how important the microbial "memory" of the plume became after the spill's first few weeks.
"It seems like in the early stages, the first week and first month, there were more dramatic swings and blooms of variability, then things stabilize a bit," Valentine said, thanks to the layered presence of multiple primed bacteria.
While microbial degradation was an important part of the plume's demise, that does not mean all of the hydrocarbons were consumed, Valentine added. Oil contains a host of complex chemicals like polycyclic aromatic hydrocarbons, which many bacteria find difficult to break down. It is quite possible those plume components vanished due to the dilution over the Gulf of Mexico's vast expanse, rather than any bacterial work.
"We don't really know what happened to a lot of that stuff," Valentine said.