On a blistering June day in Houma, Louisiana, the local offices of BP—now the Deepwater Horizon Incident Command Center—were swarming with serious men and women in brightly colored vests. Top BP managers and their consultants wore white, the logistics team wore orange, federal and state environmental officials wore blue. Reporters wore purple vests so their handlers could keep track of them. On the walls of the largest "war room," huge video screens flashed spill maps and response-vessel locations. Now and then one screen showed a World Cup soccer match.
Mark Ploen, the silver-haired deputy incident commander, wore a white vest. A 30-year veteran of oil spill wars, Ploen, a consultant, has helped clean up disasters around the world, from Alaska to the Niger Delta. He now found himself surrounded by men he'd worked with on the Exxon Valdez spill in Alaska two decades earlier. "It's like a high school reunion," he quipped.
Fifty miles offshore, a mile underwater on the seafloor, BP's Macondo well was spewing something like an Exxon Valdez every four days. In late April an explosive blowout of the well had turned the Deepwater Horizon, one of the world's most advanced drill rigs, into a pile of charred and twisted metal at the bottom of the sea. The industry had acted as if such a catastrophe would never occur. So had its regulators. Nothing like it had happened in the Gulf of Mexico since 1979, when a Mexican well called Ixtoc I blew out in the shallow waters of the Bay of Campeche. Drilling technology had become so good since then, and the demand for oil so irresistible, that oil companies had sailed right off the continental shelf into ever deeper waters.
To many people in industry and government, spills from tankers like the Exxon Valdez seemed a much larger threat. The Minerals Management Service (MMS), the federal agency that regulated offshore drilling, had claimed that the chances of a blowout were less than one percent, and that even if one did happen, it wouldn't release much oil. Big spills had become a rarity, said Ploen. "Until this one."
In the Houma building, more than a thousand people were trying to organize a cleanup unlike any the world had seen. Tens of thousands more were outside, walking beaches in white Tyvek suits, scanning the waters from planes and helicopters, and fighting the expanding slick with skimmers, repurposed fishing boats, and a deluge of chemical dispersants. Around the spot Ploen called simply "the source," a small armada bobbed in a sea of oil. A deafening roar came from the drill ship Discoverer Enterprise as it flared off methane gas captured from the runaway well. Flames also shot from another rig, the Q4000, which was burning oil and gas collected from a separate line attached to the busted blowout preventer. Nearby, two shrimp boats pulling fire boom were burning oil skimmed from the surface, creating a curving wall of flame and a towering plume of greasy, black smoke. Billions of dollars had already been spent. But millions of barrels of light, sweet crude were still snaking toward the barrier islands, marshes, and beaches of the Gulf of Mexico.
The waters of the Gulf below a thousand feet are a relatively new frontier for oilmen—and one of the toughest places on the planet to drill. The seafloor falls off the gently sloping continental shelf into jumbled basin-and-range-like terrain, with deep canyons, ocean ridges, and active mud volcanoes 500 feet high. More than 2,000 barrels of oil a day seep from scattered natural vents. But the commercial deposits lie deeply buried, often beneath layers of shifting salt that are prone to undersea earthquakes. Temperatures at the seafloor are near freezing, while the oil reservoirs can hit 400 degrees Fahrenheit; they're like hot, shaken soda bottles just waiting for someone to pop the top. Pockets of explosive methane gas and methane hydrates, frozen but unstable, lurk in the sediment, increasing the risk of a blowout.
For decades the exorbitant costs of drilling deep kept commercial rigs close to shore. But shrinking reserves, spiking oil prices, and spectacular offshore discoveries ignited a global rush into deep water. Recent finds in Brazil's Tupi and Guará fields could make that country one of the largest oil producers in the world. Similarly promising deepwater leases off Angola have excited bidding frenzies involving more than 20 companies.
In the Gulf of Mexico, the U.S. Congress encouraged companies to go deep as early as 1995. That year it passed a law forgiving royalties on deepwater oil fields leased between 1996 and 2000. A fleet of new rigs was soon punching holes all over the Gulf at a cost of up to a million dollars a day each. The number of leases sold in waters half a mile deep or more shot up from around 50 in 1994 to 1,100 in 1997.
It wasn't long before the industry hit pay dirt. New fields with names like Atlantis, Thunder Horse, and Great White came just in time to offset a long-term decline in shallow-water oil production. The Gulf of Mexico now accounts for 30 percent of U.S. production, with half of that coming from deep water (1,000 to 4,999 feet), a third from ultradeep water (5,000 feet or more), and the rest from shallow water. BP's Macondo well, in about 5,000 feet of water and reaching another 13,000 feet beneath the seafloor, wasn't particularly deep. The industry has drilled in 10,000 feet of water and to total depths of 35,050 feet—the latter a world record set just last year by the Deepwater Horizon in another BP field in the Gulf. The U.S. government estimates that the deep Gulf might hold 45 billion barrels of crude. "We're in deep water because that's where the resources are," says Larry Reed, an operations consultant in Houston who has worked with many of the major oil companies. Deepwater wells tend to be highly productive, he adds, like wells in the Middle East.
As technology was taking drillers deeper, however, the methods for preventing blowouts and cleaning up spills did not keep pace. Since the early 2000s, reports from industry and academia warned of the increasing risk of deepwater blowouts, the fallibility of blowout preventers, and the difficulty of stopping a deepwater spill after it started—a special concern given that deepwater wells, because they're under such high pressure, can spout as much as 100,000 barrels a day.
The Minerals Management Service routinely downplayed such concerns. A 2007 agency study found that from 1992 to 2006, only 39 blowouts occurred during the drilling of more than 15,000 oil and gas wells in the Gulf. Few of them released much oil; only one resulted in a death. Most of the blowouts were stopped within a week, typically by pumping the wells full of heavy drilling mud or by shutting them down mechanically and diverting the gas bubble that had produced the dangerous "kick" in the first place.
Though blowouts were relatively rare, the MMS report did find a significant increase in the number associated with cementing, the process of pumping cement around the steel well casing (which surrounds the drill pipe) to fill the space between it and the wall of the borehole. In retrospect, that note of caution was ominous.
Some deepwater wells go in relatively easy. The Macondo well did not. BP hired Transocean, a Switzerland-based company, to drill the well. Transocean's first drill rig was knocked out of commission by Hurricane Ida after just a month. The Deepwater Horizon began its ill-fated effort in February 2010 and ran into problems almost from the start. In early March the drill pipe got stuck in the borehole, as did a tool sent down to find the stuck section; the drillers had to back out and drill around the obstruction. A BP email later released by Congress mentioned that the drillers were having "well-control" problems. Another email, from a consultant, stated, "We have flipped design parameters around to the point that I got nervous." A week before the explosion, a BP drilling engineer wrote, "This has been [a] nightmare well."
By April 20 the Deepwater Horizon was six weeks behind schedule, according to MMS documents, and the delay was costing BP more than half a million dollars a day. BP had chosen to drill the fastest possible way—using a well design known as a "long string" because it places strings of casing pipe between the oil reservoir and the wellhead. A long string generally has two barriers between the oil and the blowout preventer on the seafloor: a cement plug at the bottom of the well, and a metal seal, known as a lockdown sleeve, placed right at the wellhead. The lockdown sleeve had not been installed when the Macondo well blew out.
In addition, congressional investigators and industry experts contend that BP cut corners on its cement job. It failed to circulate heavy drilling mud outside the casing before cementing, a practice that helps the cement cure properly. It didn't put in enough centralizers—devices that ensure that the cement forms a complete seal around the casing. And it failed to run a test to see if the cement had bonded properly. Finally, just before the accident, BP replaced the heavy drilling mud in the well with much lighter seawater, as it prepared to finish and disconnect the rig from the well. BP declined to comment on these matters, citing the ongoing investigation.
All these decisions may have been perfectly legal, and they surely saved BP time and money—yet each increased the risk of a blowout. On the night of April 20, investigators suspect, a large gas bubble somehow infiltrated the casing, perhaps through gaps in the cement, and shot straight up. The blowout preventer should have stopped that powerful kick at the seafloor; its heavy hydraulic rams were supposed to shear the drill pipe like a soda straw, blocking the upward surge and protecting the rig above. But that fail-safe device had itself been beset by leaks and maintenance problems. When a geyser of drilling mud erupted onto the rig, all attempts to activate the blowout preventer failed.
The way BP drilled the Macondo well surprised Magne Ognedal, director general of the Petroleum Safety Authority Norway (PSA). The Norwegians have drilled high-temperature, high-pressure wells on their shallow continental shelf for decades, he said in a telephone interview, and haven't had a catastrophic blowout since 1985. After that incident, the PSA and the industry instituted a number of best practices for drilling exploration wells. These include riserless drilling from stations on the seafloor, which prevents oil and gas from flowing directly to a rig; starting a well with a small pilot hole through the sediment, which makes it easier to handle gas kicks; having a remote-controlled backup system for activating the blowout preventers; and most important, never allowing fewer than two barriers between the reservoir and the seafloor.
"The decisions [BP] made when they had indications that the well was not stable, the decision to have one long pipe, the decision to have only six centralizers instead of 21 to create the best possible cement job—some of these things were very surprising to us here," says Ognedal.
The roots of those decisions lie in BP's corporate history, says Robert Bea, a University of California, Berkeley expert in both technological disasters and offshore engineering. BP hired Bea in 2001 for advice on problems it faced after it took over the U.S. oil companies Amoco and ARCO. One problem, Bea says, was a loss of core competence: After the merger BP forced thousands of older, experienced oil field workers into early retirement. That decision, which made the company more dependent on contractors for engineering expertise, was a key ingredient in BP's "recipe for disaster," Bea says. Only a few of the 126 crew members on the Deepwater Horizon worked directly for BP.
The drilling operation itself was regulated by the MMS (which, in the wake of the accident, was reorganized and renamed the Bureau of Ocean Energy Management, Regulation, and Enforcement). In 2009 the MMS had been excoriated by the U.S. General Accounting Office for its lax oversight of offshore leases. That same year, under the new Obama Administration, the MMS rubber-stamped BP's initial drilling plan for the Macondo well. Using an MMS formula, BP calculated that the worst-case spill from the well would be 162,000 barrels a day—nearly three times the flow rate that actually occurred. In a separate spill-response plan for the whole Gulf, the company claimed that it could recover nearly 500,000 barrels a day using standard technology, so that even a worst-case spill would do minimal harm to the Gulf's fisheries and wildlife—including walruses, sea otters, and sea lions.
There are no walruses, sea otters, or sea lions in the Gulf. BP's plan also listed as an emergency responder a marine biologist who had been dead for years, and it gave the Web address of an entertainment site in Japan as an emergency source of spill-response equipment. The widely reported gaffes had appeared in other oil companies' spill-response plans as well. They had simply been cut and pasted from older plans prepared for the Arctic.
When the spill occurred, BP's response fell well short of its claims. Scientists on a federal task force said in early August that the blown-out well had disgorged as much as 62,000 barrels a day at the outset—an enormous flow rate, but far below BP's worst-case scenario. Mark Ploen estimated in June that on a good day his response teams, using skimmers brought in from around the world, were picking up 15,000 barrels. Simply burning the oil, a practice that had been used with the Exxon Valdez spill, had proved more effective. BP's burn fleet of 23 vessels included local shrimp boats that worked in pairs, corralling surface oil with long fire boom and then igniting it with homemade napalm. In one "monster burn" the team incinerated 16,000 barrels of oil in just over three hours.
"Shrimpers are naturals at doing this," said Neré Mabile, science and technology adviser for the burn team in Houma. "They know how to pull nets. They're seeing that every barrel we burn is a barrel that doesn't get to shore, doesn't affect the environment, doesn't affect people. And where's the safest place to burn this stuff? The middle of the Gulf of Mexico."
In June the Discoverer Enterprise and the Q4000 began collecting oil directly at the busted blowout preventer, and by mid-July they had ramped up to 25,000 barrels a day—still far less, even when the efforts of the skimmers and the burn team were added, than the nearly 500,000 barrels a day BP had claimed it could remove. At that point the company finally succeeded in placing a tight cap on the well, halting the gusher after 12 weeks.
In 1990, after the Exxon Valdez spill, Congress's Office of Technology Assessment analyzed spill-response technologies and found them lacking. "Even the best national response system will have inherent practical limitations that will hinder spill-response efforts for catastrophic events—sometimes to a major extent," wrote OTA's director, John H. Gibbons. "For that reason it is important to pay at least equal attention to preventive measures as to response systems The proverbial ounce of prevention is worth many, many pounds of cure."
Just weeks before the Macondo blowout, the Obama Administration had announced with some fanfare an expansion of offshore drilling. By summer the administration was struggling in court to preserve a moratorium on deepwater drilling until such time as it could be deemed safe. "In some cases I'm not confident that the industry is tapping these resources safely," says Bea. "We can expect more of these in the future."
By early August BP seemed on the verge of plugging the Macondo well permanently with drilling mud and cement. The federal task force's estimate of the amount of oil released stood at 4.9 million barrels. Government scientists estimated that BP had removed a quarter of the oil. Another quarter had evaporated or dissolved into scattered molecules. But a third quarter had been dispersed in the water as small droplets, which might still be toxic to some organisms. And the last quarter—around five times the amount released by the Exxon Valdez—remained as slicks or sheens on the water or tar balls on the beaches. The Deepwater Horizon spill had become the largest accidental spill into the ocean in history, larger even than the Ixtoc I blowout in Mexico's Bay of Campeche in 1979. It is surpassed only by the intentional 1991 gulf war spill in Kuwait.
The Ixtoc spill devastated local fisheries and economies. Wes Tunnell remembers it well. The tall, 65-year-old coral reef expert at Texas A&M University–Corpus Christi earned his doctorate studying the reefs around Veracruz in the early 1970s, and he kept studying them for a decade after the spill coated them with oil. Tunnell wrote an early report on the impact there and on Padre Island in Texas. In early June, after the new disaster had once again raised the question of how long the impact of a spill can last, he returned to Enmedio Reef to see if any Ixtoc I oil remained. It took him three minutes of snorkeling to find some. "Well, that was easy," he said.
Tunnell stood in the clear, waist-deep water of the protected reef lagoon holding what appeared to be a three-inch-thick slab of sandy gray clay. When he broke it in two, it was jet black on the inside, with the texture and smell of an asphalt brownie. Here on the lagoon side, where the reef looked gray and dead, the Ixtoc tar mat was still partially buried in the sediments. But on the ocean side of the reef, where winds and waves and currents were stronger, no oil remained. The lesson for Louisiana and the other Gulf states is clear, Tunnell thinks. Where there is wave energy and oxygen, sunlight and the Gulf's abundant oil-eating bacteria break it down fairly quickly. When oil falls to the bottom and gets entrained in low-oxygen sediments like those in a lagoon—or in a marsh—it can hang around for decades, degrading the environment.
Fishermen in the nearby village of Antón Lizardo hadn't forgotten the spill either. "The Ixtoc spill about destroyed all the reefs," said Gustavo Mateos Moutiel, a powerful man, now in his 60s, who wore the trademark straw hat of the Veracruzano fishermen. "Octopus gone. Urchins gone. Oysters gone. Conch gone. Fish almost all gone. Our families were hungry. The petroleum on the beach was halfway up our knees." Though some species, such as Bay of Campeche shrimps, recovered within a few years, Moutiel, along with several other fishermen who had gathered on the beach, said it took 15 to 20 years for their catches to return to normal. By then two-thirds of the fishermen in the village had found other jobs.
Even in the turbulent, highly oxygenated waters of France's Breton coast, it took at least seven years after the 1978 Amoco Cadiz spill for local marine species and Brittany's famed oyster farms to fully recover, according to French biologist Philippe Bodin. An expert on marine copepods, Bodin studied the long-term effects of the spill from the grounded tanker. He believes the impact will be far worse in the generally calmer, lower-oxygen waters of the Gulf, particularly because of the heavy use of the dispersant Corexit 9500. BP has said the chemical is no more toxic than dishwashing liquid, but dispersants were used extensively on the Amoco Cadiz spill, and Bodin found them to be more toxic to marine life than the oil itself. "The massive use of Corexit 9500 in the Gulf is catastrophic for the phytoplankton, zooplankton, and larvae," he says. "Moreover, currents will drive the dispersant and the oil plumes everywhere in the Gulf."
In May, scientists in the Gulf began tracking plumes of methane and oil droplets drifting up to 30 miles from the broken well, at depths of 3,000 to 4,000 feet. One of those scientists was University of Georgia biogeochemist Mandy Joye, who has spent years studying hydrocarbon vents and brine seeps in the deep Gulf. She found a plume the size of Manhattan, and its methane levels were the highest she had ever measured in the Gulf. As bacteria feast on spilled oil and methane, they deplete the water of oxygen; at one point Joye found oxygen levels dangerously low for life in a water layer 600 feet thick, at depths where fish usually live. Since waters in the deep Gulf mix very slowly, she said, such depleted zones could persist for decades.
BP was using old DC-3s set up like giant crop dusters to spray Corexit 9500 onto surface slicks. But for the world's first major deepwater spill, the company also got permission from the U.S. Environmental Protection Agency and the Coast Guard to pump hundreds of thousands of gallons of dispersant directly into the oil and gas spewing from the well, a mile beneath the surface. That helped create the deepwater plumes.
"The whole goal is to keep oil off the beaches, because that's what drives the economy," Joye said one day in June as she ran samples through her gas chromatograph aboard the R.V. F. G. Walton Smith. The little research ship was bobbing in an oily sheen a few miles from the busted well. "But now you've got all this material in the water column that no one is seeing and that you can't get rid of. If oil gets to the surface, about 40 percent evaporates. You can skim it, you can burn it, you can do something with it. But these tiny particles in the water column will persist for God knows how long."
Oceanographer Ian MacDonald at Florida State University worries not only about the plumes but also about the sheer volume of spilled oil. He believes it could have a major impact on the overall productivity of the Gulf—not just on pelicans and shrimps in the Louisiana marshes, but on creatures throughout the region, everything from zooplankton to sperm whales. He's particularly concerned about bluefin tuna, which spawn only in the Gulf and in the Mediterranean; the tuna population was already crashing due to overfishing. "There is a tremendous amount of highly toxic material in the water column, both at the surface and below, moving around in one of the most productive ocean basins in the world," MacDonald said.
During their June cruise Joye's team sampled water within a mile of the Discoverer Enterprise, close enough to hear the apocalyptic roar of its huge methane flare. Researchers and crew members stood on the back deck of the Walton Smith and quietly took pictures. The caustic vapors of oil, diesel, and asphalt burned their lungs. As far as the eye could see, the cobalt blue waters of the deep Gulf were stained brownish red. When Joye went back inside she was in a pensive mood.
"The Deepwater Horizon incident is a direct consequence of our global addiction to oil," she said. "Incidents like this are inevitable as we drill in deeper and deeper waters. We're playing a very dangerous game here. If this isn't a call to green power, I don't know what is."
Americans burn nearly 20 million barrels of oil a day. In early August the U.S. Senate adjourned for the summer without taking up an energy bill.
Correction: A revision has been made in the identification of the dispersant used after the 1978 Amoco Cadiz spill off the coast of Brittany. We incorrectly identified it as Corexit 9500, a dispersant used extensively in the Gulf of Mexico after the Deepwater Horizon disaster. The manufacturer of Corexit, the Nalco Company, points out that while other brands of dispersants were used in 1978, Corexit 9500, which was developed later, was not. Three thousand metric tons of dispersant were used by the French Navy to combat the Amoco Cadiz spill. Scientific studies show that it took as much as two decades for some species to recover to pre-spill levels.