In the thick of a Greenland summer of field work in 2015, Benjamin Hmiel and his team drilled into the massive ice sheet’s frozen innards, periodically hauling up a motorcycle-engine-sized chunk of crystalline ice. The ice held part of the answer to a question that had vexed scientists for years: How much of the methane in the atmosphere, one of the most potent sources of global warming, comes from the oil and gas industry?

Previously, geologic sources like volcanic seeps and gassy mud pots were thought to spit out about 10 percent of the methane that ended up in the atmosphere each year. But new research, published this week in Nature, suggests that natural geologic sources make up a much smaller fraction of the methane in today’s atmosphere. Instead, the researchers say, that methane is most likely attributable to industry. Added up, the results indicate we’ve underestimated the methane impacts of fossil fuel extraction by up to 40 percent.

That’s both bad news for climate change and good, says Hmiel, the lead author of the study and a researcher at the University of Rochester. Bad, because it means that oil and gas production has had a messier, bigger impact on the greenhouse gas budget than scientists knew. But Hmiel finds the result encouraging for almost the same reason: The more of the methane emissions that can be pinpointed to human activity like oil and gas extraction, the more control it means policymakers, businesses, and regulators have to fix the problem.

“If we think of the total methane in the atmosphere as slices of a pie—one slice is from ruminants, this other is from wetlands. The slice is we used to think was from geologic methane was too big,” says Hmiel. “So what we’re saying is that the fossil fuel pie slice is larger than what we thought, and we can have a bigger influence on the size of the slice, because it’s something we can control.”

Methane, the “bridge” fuel—but a bridge to where?

A potent greenhouse gas, methane’s carbon core and hydrogen arms are arranged in a configuration that makes it exceptional at absorbing heat. On a 20-year timescale, a methane molecule is roughly 90 times more effective at trapping heat in the atmosphere than a molecule of carbon dioxide, the greenhouse gas that wields the most control over Earth’s future warming in the long-term.

Methane’s atmospheric concentrations have increased by at least 150 percent since the Industrial Revolution. Because of its potency, the more of it there is in the air the harder it will be to keep the planet’s temperatures from soaring past global climate goals.

Methane is also the protagonist in a planet-wide, decades-long scientific mystery: Where, exactly, does all the extra methane heating up the atmosphere today all come from? Is it cow burps or rice paddies? Leaks from oil and gas production? Burbling gassy mud volcanoes or seeps along the Earths shifting seams?

Over the past few decades, as calls to reduce carbon dioxide emissions have grown louder and natural gas collection technologies like fracking have gotten cheaper, many coal-fired power plants across the United States and abroad have retired. In the U.S. over 500 coal-fired power plants have closed since 2010. In many cases they are replaced by natural gas (which is made up primarily of methane gas) plants, which now produce nearly 40 percent of the U.S.’s energy needs.

Methane burns more efficiently than coal, making it a better option, carbon-cost-wise and air-pollution-wise, than coal. It also sticks around in the atmosphere for much less time than CO2—an average of nine years, compared to CO2’s hundreds.

Because of its characteristics, natural gas has been often been touted as a “bridge fuel” to help smooth the transition to a carbon-neutral energy future. Natural gas plants fill energy needs today while renewable or carbon-negative technologies develop.

“The question is: Is this a bridge fuel, or is it going to be around for a very long time?” says Sheila Olmstead, an environmental economist at the University of Texas at Austin. “The market is telling us it’s probably going to be around for a long time.”

However, the climate cost of natural gas has relied on a basic assumption: There are less total carbon emissions from natural gas than from other sources. But in recent years a flotilla of scientific studies have brought that assumption into question, primarily by looking at how much gas is lost during the production process.

If there are very few leaks or losses along the way—less than a few percent of the total amount of gas recovered—the math breaks even or comes out ahead. But if that “leakage rate” climbs over more than about 1 percent of the total gas recovered, the budget gets fuzzy, says Robert Howarth, a climate scientist at Cornell.

One recent study found that the widely used “leakage rate” of gas in the U.S. natural gas production process could be over 2 percent. Others, looking at specific “super emitters” in major drilling regions of the US, have found even more leakage.

“Over the past few years of research I’d say the whole argument for methane for a bridge fuel is really gone,” says Howarth. “But if we go back and say we really do need natural gas for a while, that calculation depends on methane’s break-even point. And we’re not sure we’re close to that.”

It’s critical to phase out CO2 emissions, stresses Jessika Trancik, an energy expert at MIT, because that’s the stuff that will keep the planet locked in for long-term warming. But for the climate goals the world is scrambling to hit right now—keeping air temperatures from soaring the 3.6 degrees Fahrenheit (2 degrees Celsius) temperature goals from the 2015 Paris Agreement—it’s also critical to keep any extra methane from leaking into the atmosphere.

“It’s impossible to hit those climate targets with methane in the mix,” says Lena Höglund Isaksson, a greenhouse gas expert at Austria’s International Institute for Applied Systems Analysis.

(See methane flowing from a leak at a natural gas storage field near Los Angeles).

The ice has answers

It’s remarkably difficult to figure out how much of the methane in the atmosphere comes from human sources, like oil and gas drilling or burning, how much comes from other human-influenced source like agriculture, and how much comes from natural sources like volcanic seeps.

Where it comes from determines what humans can do about it. If it’s oil and gas, we can fix the systems to produce less. If it’s volcanoes, we might be less able to manage the emissions.

“It’s like a detective story,” says Höglund Isaksson.

In the past, scientists made estimates of how much so-called natural methane comes from geologic sources by trekking to a particular seep or muddy volcano and very carefully measuring its emissions. Then the scientists would scale up those observations to make an estimate for the entire planet. Using that strategy, most estimates put the annual contribution of natural geology-sourced methane at about 50 teragrams per year, around 10 percent of the total annual amount of methane emitted. Recent estimates put the total annual methane contribution from acquiring and burning fossil fuels at just under 200 teragrams.

Hmiel’s team suspected that the geologic sources might actually be even smaller—and they had a place to test that suspicion: the wide, flat Greenland ice sheet. The ice there, buried over 100 meters below the surface, dated from before the Industrial Revolution got underway in the 1800s, and so it had pre-industrial methane trapped inside tiny air bubbles in its frozen lattice.

They dug up over 2,000 pounds of ice. Then they sucked the methane-containing air out of the bubbles trapped in the ice.

Methane from natural geologic sources has a slightly different chemical makeup than methane from other sources, like wetlands. The methane sucked out of the 250-year-old ice contained traces of only a tiny amount of geologic methane. And because the samples were from before the start of the Industrial Revolution and the concurrent increase in methane from coal and oil, there were no traces of methane from fossil fuels.

In contrast, samples from after the Industrial Revolution started showed a telltale fingerprint of fossil fuels.

But the key finding was about how little methane from geologic sources there was in the ice: the equivalent of no more than about 5 teragrams of methane released to the atmosphere per year, in those pre-fossil-fuel-dependent days. It’s unlikely that the geology has changed in that short a time, so that estimate is, Hmiel says, a good assumption for what geology is contributing today, as well.

Crucially, that contribution is 10 times smaller than other estimates—including those used by the U.S. Environmental Protection Agency and the Intergovernmental Panel on Climate Change—used to make scientific assessments and policy decisions.

Overall, scientists have long known exactly how much methane there is in the atmosphere. That number hasn’t changed: There’s still about 570 teragrams of methane collecting in the atmosphere each year. But if there’s a lot less from the natural geologic sources, some other source must make up the difference. The team could also demonstrate that the most likely source is oil and gas operations.

If oil and gas operations have had a much bigger footprint on methane emissions than previously known, Hmiel thought, that also means they can clean up those emissions—both by reducing the amount of gas used and by cleaning up the leaks, flares, and other wasted gas from the process.

“Power utilities that are currently choosing whether to focus on wind and solar or gas—if they choose gas, it’s crucial to understand that that plant is going to be around for decades,” says Olmstead.

“They have real staying power well beyond what the nameplate expiration date is. Knowing that, does that change the decisions we make today? That we’ll have effects on methane emissions 10, 20, 30, 40 years into the future?”