Many people think of oil deposits as vast underground lakes, and oil wells as big straws that suck up the liquid gold. Eventually, the lake is drained, so the oil company packs up its gear and moves on. If this were really how things worked, the dwindling size of recent oil and natural gas discoveries would be alarming -- feeding gloomy predictions that crude oil production is approaching its peak.
In fact, when major oil outfits abandon wells, they usually leave a lot behind. The reservoirs under most U.S. wells drilled in the past century may still hold twice the amount of oil that has been sucked to the surface. The reason is Geology 101: Oil is locked in the pores of rock layers deep below ground. Sink a hole into such a layer, and the pressure of the earth above squeezes out the oil. But as it oozes out, pressure on the remaining oil diminishes. After three or four decades the flow of oil drops so low that big oil producers, such as Exxon Mobil Corp. (XOM ) or France's Total (OT ), no longer want to bother collecting it. So they sell the well to a smaller company, such as Anadarko Petroleum Corp. (APC ).
Now, many engineers are reassessing the riches that may lie hidden under such wells -- including the 400,000 U.S. wells that produce, on average, just 2.2 barrels a day. These still account for almost 15% of domestic U.S. oil production, or 7% of total U.S. consumption. Using new, enhanced recovery techniques, the output of some low-flow wells can be increased dramatically. It's even possible to revive old wells that aren't producing a drop.
In addition, new supercomputer systems are powerful enough to simulate not just individual wells but also entire oil fields. These improved models are revealing subterranean details that geologists and oil-field engineers have never seen before. They often spot isolated pockets of oil and gas that can be tapped by extending a nearby well. And with better models of the underground oil-bearing structures, simulations can help determine which of the new -- and expensive -- recovery methods would work best for a given deposit.
The oil barons of the early 20th century rarely pulled up more than 10% of a reservoir's bounty. Things got better after World War II, when the oil industry developed secondary "lifting" techniques. These restore underground pressure by pumping water back down into the earth or returning the natural gas that normally bubbles up with the oil. Secondary techniques enabled oil companies to withdraw 30% of the oil in deposits below.
More recently, engineers have unleashed a third, or tertiary, wave of recovery methods, using gas, chemicals, and even colonies of specially engineered microbes, to rejuvenate old wells. These approaches can double extraction potential, to 60% or 70%. Under ideal conditions, some companies claim they'll hit 80%. In other places tertiary methods may not work at all. But if 60% proves to be a rough average, that would "virtually double the known reserves of oil," says Olivier Le Peuch, president of Schlumberger Information Solutions, a unit of oil-services giant Schlumberger Ltd. (SLB ).
For now, such technologies aren't in widespread use. Tricks such as boosting underground pressure with down-the-hole injections of carbon dioxide don't come cheap -- some can easily double extraction costs. So in the 1990s, when oil prices were low, major producers had little incentive to adopt such methods. But the recent runup in oil prices has spurred smaller, independent producers to embrace high tech. Overall, though, the industry is still chugging along at "an average recovery of about 30%," says James "Jeb" Blackwell, manager of exploration at Chevron Energy Technology Co.
That may soon change. Tapping the "new" oil is a relative snap, notes William Bartling, head of market strategy for the oil and gas industry at Silicon Graphics Inc. (SGI ), which builds supercomputers for oil-field simulations. "You know where the oil is, the wells are already drilled, and the equipment is amortized," says Bartling. "So getting the extra oil is very margin-rich."
That's why Anadarko is budgeting $684 million for long-term operations at the Salt Creek oil field, which it bought for $265 million in 2003. Salt Creek, north of Casper, Wyo., is dotted with 4,000 wells that have pumped out some 700 million barrels of oil. But Anadarko thinks at least 150 million more barrels can be extracted. By injecting CO2, Anadarko expects to boost output to 28,000 barrels a day, from 2003's level of 5,000 barrels a day.
Obtaining CO2 isn't a major obstacle. About 125 miles away, in Shute Creek, Wyo., Exxon operates a natural gas processing plant that vents tons of CO2 annually. Some of it is now arriving at Salt Creek through a new $45 million pipeline. By 2008 all of Exxon's CO2, and then some, will be pumped down into the earth, permanently disposing of 2 million tons a year of greenhouse gas while unlocking acres of black gold for Anadarko.
Similar schemes are already working in the Southwest. "We have 1,000 miles of CO2 pipeline feeding enhanced-recovery oil wells," says David J. Borns, manager of geotechnology research at Sandia National Laboratories. Some 25 million tons a year of CO2 have helped boost the region's oil production by about 500,000 barrels a day.
The latest idea is called MEOR, for microbial enhanced oil recovery. Various labs around the world are engineering special bugs that generate CO2 biologically, along with detergent-like chemicals that help flush oil out of rocks. The microbes can be cultivated underground or in well-side vats. Because they grow explosively, the Energy Dept., which is funding several research projects, says MEOR technology may be the most cost-effective of all tertiary processes.
MEOR is already used in Venezuela, China, Indonesia, and the U.S. to treat deposits of heavy oil -- a molasses-thick form of oil. Researchers at Oak Ridge National Laboratory hope to develop new armies of bioengineered bugs that can infiltrate underground rocks and turn the gunky stuff into the sweet-flowing crude that erupts like the gushers in Hollywood movies.
Another technique involves injecting steam into wells to reduce the viscosity of heavy oil. This might also help unleash the energy trapped in Canada's famous tar sands. To get at these deposits without employing strip-mining methods, engineers need to predict how the thinned-down oil will flow underground. The software that simulates flow in regular oil reservoirs doesn't work well when sand granules are mixed in with the oil. But an upcoming software tool, dubbed Intersect, will handle oil-and-sand mixes. Development is being funded by Schlumberger, Total, and Chevron, which owns a large portfolio of heavy-oil deposits.
When a commercial version of Intersect is unveiled later this year, it could spur faster development of Canada's immense oil-sand deposits. These hold recoverable reserves estimated at around 179 billion barrels -- with the potential of topping even Saudi Arabia's proven reserves of 261 billion.
Since 1990, after the oil giants began selling off scads of low-flow wells, small and midsize companies have hit jackpots. Marathon Oil Corp. (MRO ) and Arco (now part of BP PLC (BP )) used some of the world's fastest supercomputers to reanalyze old seismic survey data they bought from big producers. "Marathon and Arco were able to see a little more detail," says SGI's Bartling, "and they spotted oil in places where others hadn't."
Small outfits such as Devon Energy (VN ), Newfield Exploration (NFX ), and Spinnaker Exploration (SKE ) also pounced on advanced computer models and the new recovery options. "They worked hard to make $10 oil economic," Bartling says. "Now they're getting $60. It's the American dream -- with lots of new Ferraris in company parking lots." Playing this new high-tech oil game, he bets, will keep on paying off well beyond 2010.
By Otis Port in New York