Kevin Macumber wanted to be a forester. Today he manages about 4,000 acres of longleaf pine in Mississippi—not for the timber, but for what lies far beneath the woods. It’s black gold: oil, deep underground. And the key to getting it out is the same molecule that lets all those trees grow: carbon dioxide.

“Another day in paradise,” says Macumber as we meet at a Chevron gas station in southeastern Mississippi, about the closest thing to a landmark around here. We’re headed to the old trailer home that’s become the operational headquarters for Tellus Operating Group, a wildcat oil company with some old oil fields in this neck of the woods. I follow his black Chevy pickup down country byways until we finally turn off on a dirt road that winds through forests to the company trailer, where the coffee is fresh and Macumber can banter with a few of his workers on this warm sunny day.

The secret about old oil reservoirs below the surface is that they still have oil, sometimes a lot, but it no longer comes out easily. Companies can pump large volumes of CO2, piped in from natural deposits belowground, down into the wells, forcing out the oil that would otherwise stay put. Macumber used to work at Occidental, one of the large oil companies that helped pioneer this “enhanced oil recovery” technique in Texas. But now he’s helped found Tellus—named after one of the Roman Earth goddesses—to do the same thing in Mississippi.

Right now, Tellus gets its CO2 from a deposit called the Jackson Dome in western Mississippi, and other oil companies are using a similar approach at hundreds of old wells around the country. But Tellus is one of two U.S. oil companies that hopes to try something entirely new, any day now. Instead of piping in natural CO2, it will use the greenhouse gas captured at a coal-fired power plant just completed nearly 100 miles north of here and send it down into the reservoir, pushing oil out and leaving the greenhouse gas deep below, safely locked away from the atmosphere, so it does not add to global warming.

At least three coal-fired power plants are under construction in the U.S. that are designed to have their CO2 emissions captured and sent to an oil field for enhanced oil recovery, including the Kemper County Energy Facility up the road from here. More arrangements like this are being made worldwide. The scheme is vital: The only way nations can meet the targets in the Paris Agreement to combat climate change is to eliminate the burning of fossil fuels or to capture emissions and find a place to store them besides the atmosphere. Sequestering the gas belowground costs money, and the only way to pay for it on a scale large enough to slow global warming is for oil companies worldwide to buy the CO2 for enhanced oil recovery. The coal plants, in return, would make money selling their CO2.

But there’s a new flaw in this game plan: cheap oil. Oil companies that today pay for CO2 to be delivered from natural deposits are in danger of losing money, because the current price of oil is so low. For now, company’s like Tellus and others have to keep CO2 flowing into their old reservoirs because if they shut down, they lose the underground pressure they spent so much money to create. But if oil prices stay too low for too long, they will no longer be able to afford to keep purchasing CO2. Then the world will lose the one prospective way it currently has for paying to keep CO2 out of the atmosphere.

Macumber says Tellus loses money on oil produced with CO2 when a barrel of oil sells below $50. Oil is currently selling for around $40 a barrel, thanks to a world awash in petroleum, perhaps because producers are scrambling to pump as much as possible, before economies move away from oil in an attempt to limit climate change. The industry has survived price swings before, but the low price now could be an early sign of a long-term decline.

One big piping system

To show me how enhanced oil recovery works in Tellus’s Raleigh oil field, Macumber takes me down the dirt road to the pipeline that makes it all possible. It carries CO2 from Jackson Dome, but the same line will one day carry CO2 captured at the Kemper facility. The only signs of the pipe are a gap in the trees that stretches for kilometers, a tiny yellow marker that says “Warning: Carbon Dioxide Pipeline,” and a little shack that houses the pumps that keep the CO2 moving underground and pressurized.

The process is not as easy as just getting CO2 and dumping it down a well, however. A kind of mini factory needs to be built—pumps, compressors, generators for electricity, among other kit. The entire old field and wells need to be refitted to cope with CO2 going down and, more importantly, to cope with CO2-laced water (otherwise known as carbonic acid) coming back up, which eats away at machinery and metals that are not properly protected. Then there are the big electric bills that come from running all those pumps and compressors—the single largest expense, which doesn’t include the $1 million or more a mile it costs to lay an electric line through these woods to power the equipment.

Before the CO2 goes down into a reservoir it has to be pressurized up to as much as 4,500 pounds per square inch. Millions of cubic feet of the gas must be pumped underground each day. All of that costs billions of dollars in initial outlay, plus daily operating expenses. The oil company behind it must be patient, because it can take years to put enough CO2 down the hole to build up sufficient pressure to push up an extra barrel of oil out the other end. Since 1999 the largest enhanced oil recovery (EOR) company operating in Mississippi, Denbury, estimates it has spent more than $5 billion to build and operate oil facilities and pipelines using CO2 in the Magnolia State alone. A 15 percent federal tax credit helps offset some of that cost, as do various state tax exemptions, but it’s still a significant chunk of change.

Once underground, the tiny greenhouse gas molecule mixes with the bigger molecules that make up the toxic stew known as oil, both helping them flow better and restoring the subterranean pressure that had been reduced by the original tapping of the petroleum. CO2 also acts kind of like dishwashing liquid; it scours out oil lurking between grains of sand. The CO2 goes down, pressure goes up and less viscous oil flows back to the surface, ready to be sold. The oil industry calls the whole operation a recycling facility: burn oil, produce CO2, capture that CO2 and use it to force out more oil to burn. If the CO2 can be captured from a coal-fired power plant, even if it ends up producing a few more barrels of oil, the overall approach can help to reduce emissions, slowing global warming.

Cheap gas a must

The trick to EOR is getting the CO2 on the cheap. Nationwide, the oil industry injects roughly 60 million metric tons of CO2 into old oil wells each year. That’s the equivalent of the pollution from 20 coal-fired power plants. Roughly 70 million barrels of oil per year are produced this way. Since 1972 CO2 has been injected into U.S. oil reservoirs continuously, resulting in an extra two billion barrels of oil and a billion metric tons of CO2 stored underground. Today about one quarter of that CO2 comes from industries that happen to be located close to old oil fields and produce lots of CO2 as a by-product, such as fertilizer manufacturing plants or cement kilns. The other three quarters is naturally occurring CO2, which simply transfers the gas from one underground reservoir to another.

Right now, CO2 can cost less than $10 per metric ton. In Mississippi roughly one ton of CO2 yields almost two barrels of oil, so $5 of CO2 per barrel is significant if oil is selling for less than $40 a barrel—and that’s before adding in all the capital costs for the equipment. The cost of dealing with CO2 can be as much as half the cost of recovering a barrel of oil this way, according to the National Energy Technology Laboratory.

Yet the practice is widespread, in part because oil prices have been much higher in recent years and because it is hard to find new multimillion barrel reservoirs these days, especially in the picked over U.S. Denbury, based in Plano, Texas, controls more than 1,000 miles of CO2 pipelines and has published reserves of 17 trillion cubic feet of the greenhouse gas, used to pump more than 70,000 barrels of oil a day. It is using CO2 at roughly 170 wells at the Tinsley oil field here in Mississippi. Everywhere the trees try to grow in around the wells, sequestering CO2 the biological way. “Grappling with the jungle is part of the business out here,” says Greg Schnacke, executive director of governmental relations for Denbury, on a tour of Tinsley.

All told, there are maybe 5,000 miles of CO2 pipeline in the U.S. Despite what sounds like big numbers, “it’s not a big industry,” Schnacke notes. “It’s roughly 3 percent of U.S. [oil] production.”

Designing coal for carbon capture

If the Jackson Dome did not exist underground in Mississippi, the Kemper County Power Facility to burn coal and capture the CO2 would not exist either. In fact, geology dictates everything about this particular power plant and oil scheme: There are billions of metric tons of the dirtiest form of coal—lignite—literally underneath where Kemper was built by Mississippi Power and its corporate parent Southern Co. The lignite was laid down by an old iteration of the Gulf Coast millennia ago. Similarly, the ancient coast left lots of oil deposits, salt domes and the like.

The Kemper facility is the world’s first full-scale coal-fired power plant designed for carbon capture. The costly plant will clean the dirtiest form of coal by first turning it into a gas, then stripping off the various pollutants—acid rain–causing sulfur, smog-forming nitrogen and globe-warming carbon—before any burning. To pay for this expensive proposition, each of the pollutants gets turned into a product: sulfur into sulfuric acid for the pulp and paper industry, nitrogen into ammonia for agriculture and carbon into pure CO2 for oil companies, traveling down a new specially built pipeline 60 miles to interconnect with the existing network.

For the CO2 captured at Kemper to work for enhanced oil recovery, oil prices need to go up. Whether they do depends as much on geopolitics as geology—the vagaries of market speculation on future oil prices and how much economic pain private oil companies can take compared with their national oil company counterparts, like Saudi Arabia’s Aramco. As it stands, long-term futures contracts suggest that oil is headed to more than $40 per barrel in 2018. The world may be awash in oil at present, making it cheap, but such gluts have not lasted forever in the past—and the oil industry is gambling that it will not last forever this time either. Cheap oil usually boosts demand, which then consumes available supply, driving prices back up over time—or so it has been over the course of the 20th and early decades of the 21st centuries.

Driving back from Tinsley, Schnacke and I pass a coal train. “That’s our competition,” he observes, meaning the coal is destined for a different power plant.

“That’s your future CO2 source,” I counter. For a system of enhanced oil recovery fed by coal plants designed for carbon capture to pay off, Denbury, Tellus and every other oil company must survive current low oil prices.

The last time oil prices stayed low for a very long time was the 1980s and 1990s. But this is not the same oil industry. Now the only big finds are offshore or, perhaps, in the Arctic. These are places that require not just substantial planning but also a huge amount of money invested before the first barrel of oil appears, not unlike EOR with CO2, only with a much less certain payoff. In fact, Denbury might use the low oil prices to acquire more reservoirs on the cheap. “A lot of oil fields are capable of accepting CO2,” Schnacke notes. “It requires capital investment and time.”

Does carbon capture pay?

At the nearby Oil Field Café, everybody I ask chuckles when I observe that the fight against climate change seems to rely on burying CO2 to bring up more oil—which then gets burned to create more CO2. Yet, putting CO2 capture on coal-fired power plants and other big industrial polluters seems less a question of whether and more a question of when. The real question is: Who will pay for it?

Dumping CO2 in the sky is free, but capturing it—and even more so, storing it underground—costs money. This gap between market realities and action to combat climate change is where the government comes in, in theory. Indeed, the Clean Power Plan proposed by the Obama administration to clean up CO2 emissions from power plants relies on capture and storage to allow coal-fired power plants to continue to produce electricity, but with less climate-changing pollution. In the long run even power plants that burn natural gas will need to capture CO2.

Companies that do this get a federal tax credit of $10 for every ton up to 75 million tons, but that does not defray the massive initial expense. Kemper has cost more than $6 billion to build. “EOR is the best choice,” says Rich Esposito, a geologist turned chemist at electric utility Southern Co.

Kemper will capture roughly three million metric tons of CO2 each year. Tellus and Denbury need so much CO2 that they’ve contracted with Kemper to purchase all of it.

For the climate to benefit, however, enough of that CO2 has to remain sequestered underground after it’s done scouring out more oil, a complex calculation that also depends on how many barrels of oil—and of what quality—are ultimately produced. “All of this is for naught if we don’t get the CO2 certified as permanently sequestered,” Esposito says. Industry practice and academic research suggest that one third of the CO2 pumped underground stays there—trapped in the same microscopic holes in the rock that once held the oil—and two thirds comes back up with the petroleum. That two thirds is then topped up with fresh CO2 from the pipeline and sent back underground to scour out more oil. Few such EOR operations have come to the end of their useful lives, which means few have closed in a well that used CO2. So a full accounting of how much CO2 really gets stored is postponed for some future reckoning.

Still, CO2 for oil recovery can hardly be worse than simply dumping the greenhouse gas directly into the atmosphere, where it has already accumulated in sufficient quantity to stave off the next ice age for millennia. “We frankly believe that probably 99-plus percent of what we purchase and put into EOR remains behind,” says Dan Cole, Denbury’s vice president of commercial development. That remains to be proved.

Ordinary folks fear the CO2 will simply leak upward, worrying about CO2 mingling with water to form carbonic acid that leaches heavy metals and other contaminants out of the deep. Or about it escaping directly to the surface and settling in a smothering cloud on a home or town, as happened in 1986 when Lake Nyos in Cameroon burped out a pure, invisible cloud of natural CO2 that killed more than 1,700 people. The storage benefit is obviated if there’s even a small leak that provides a path back to the atmosphere. Old wells that connect underground to newer ones could prove a problem, some of which were plugged with nothing better than a tree stump or have simply been forgotten and lost.

And in the end there is only so much CO2 the oil industry can use. Even all the oil reservoirs in the world could not handle the more than 13 billion metric tons of CO2 that come from burning coal each year, even if pipelines and the rest could be built. “We can only take a certain amount of CO2 that could potentially be captured in the future,” Cole says. “EOR can’t be the end-all answer for CO2 capture.”

The U.S. has the most oil recovered with CO2 but it is really China that needs the technology. China is the world’s largest source of CO2 pollution yet it is less capable of affording the technology to clean up its coal-burning habit. Getting more oil out of the ground, in the Jilin or Shengli petroleum reservoirs for example, could help defray the cost.

One can imagine an underground network of CO2 pipelines for EOR—and ultimately underground storage of the greenhouse gas—that grows to the size of the underground and aboveground network of oil and gas pipelines that currently exists, one that covers most continents and even extends offshore to where CO2 can most safely be buried under the seafloor. But it is also easy to imagine how many trillions of dollars it would take to build such a vast, sprawling industrial infrastructure to clean up the vast, sprawling industrial infrastructure that already exists to burn fossil fuels. And, besides EOR, there is no current way to make money from that infrastructure.

Perhaps low oil prices are a practice run for life under climate change, when fossil fuels have to be forsaken, driving down their prices and driving companies to produce cleaner forms of energy. A carbon tax could make the dirtiest fossil fuels unprofitable. Low oil prices may render impossible petroleum found in difficult environments like the Arctic or far offshore in the oceans or found in difficult forms like tar sands. Even flooding old oil reservoirs with CO2 could prove too expensive to sustain in a world where oil costs $30 a barrel rather than the $120 a barrel of a few years back.

Economics is one rubric that works to keep fossil fuels in the ground. Another is the government and regulations like the Clean Power Plan. The Paris Agreement may also help. But the U.S. is in the vanguard of the effort to have coal and burn it, too. As it stands, oil fields like Tinsley or Macumber’s Raleigh will make or break prospects for cleaning the dirtiest power plants.

Fracking and CO2 are the best available routes to more oil here in the U.S., Macumber and others argue. “This is our golden child in the future,” the oil industry veteran says of his Mississippi assets. CO2 “is a fairly inexpensive commodity for what it does,” he adds. “It’s hard to beat.”