Alaska’s landscape is changing. One striking example: During the past six decades, the number of thermokarst lakes in Fairbanks, the state’s third-largest city, has doubled. These are lakes that form when permafrost thaws, causing the ground to subside or in some cases collapse and then fill with water. Increased warming means more thawing and more lakes. And because the Arctic is warming three times faster than the rest of the planet, these formations are becoming an increasingly common feature across the state, especially in the interior, where the permafrost layer is often just below the 32-degree threshold required to keep the soil frozen. Even modest increases in temperature can lead to landscape-scale changes or abrupt thaw events such as slumps and landslides.

In mid-June, reporting a story for the current issue of Sierra on the impacts of a warming Arctic on Alaska’s built environment, I visited one of these thermokarst lakes, a small, unassuming body of water outside Fairbanks. It was close to a frequently traveled two-lane road, a power line, and residential property, all of which could be affected in the coming decade as the permafrost continues to thaw and the lake expands. Homes might need to be relocated and the power line and road rebuilt to withstand the influx of water and shifting ground. This is easy enough to imagine.

But the thawing permafrost in Alaska is connected to other big questions scientists are scrambling to understand: How much methane and carbon dioxide will be emitted from ancient carbon stored in the ground as it heats up, and what kind of impact will this have on future warming?

The process of permafrost thaw leading to increased emissions is relatively straightforward. As the ground warms, organic matter, mostly dead plants and animals compressed and frozen for thousands of years, is made available to microorganisms, which convert the material into CO2, methane, or nitrous oxide. Scientists have a pretty good idea how much carbon is stored in northern soils: more than twice what humans have already emitted since the beginning of the Industrial Revolution.

But researchers know far less about how much of that carbon will be released over time and precisely what form it will take: methane or carbon dioxide. On a 20-year timescale, methane is about 80 times more potent a greenhouse gas than CO2, so it has greater effects in the short term but does not remain in the atmosphere as long. Adding to the complexity is the fact that permafrost varies considerably across the Arctic region. The Fairbanks area, for example, is on “discontinuous permafrost,” meaning it is not uniformly frozen. (Permafrost is defined as ground that remains at or below freezing for two consecutive years.) But on Alaska’s North Slope, the ground is solidly frozen, and the ice-rich permafrost can be hundreds of feet deep. The rate of thaw and subsequent emissions can also be shaped by vegetation type and the presence of water.

“We have widespread evidence that the permafrost is changing and degrading very rapidly,” said Charles Miller, Deputy Science Lead for NASA’s Arctic Boreal Vulnerability Experiment, who has done extensive monitoring in the Arctic. “This would suggest from our models that we should see large increases in the amount of methane. But we haven’t really seen those yet. And that’s a big puzzle.”

Parts of that puzzle are beginning to take shape. Katey Anthony, an aquatic ecosystem ecologist at the University of Alaska Fairbanks, has spent more than 20 years studying thermokarst lakes and the role they play in emitting methane. She has mapped emissions from hundreds of lakes across Alaska, including the one I visited. The results are worrisome. Thermokarst lakes, she’s found, are hot spots for emissions, particularly methane.

“These lakes have the highest emissions of any kind of land surface type in the Arctic,” Anthony told me. While thermokarst lakes make up only a small percentage of the Arctic landmass, they could still be a significant source of added methane, on top of what humans are already putting into the atmosphere via oil and gas emissions and agriculture. According to Anthony, if methane emissions from thermokarst lakes were included in models—currently they aren’t—the climate feedback from permafrost thaw would double over the next 80 years. Even though it remains a relatively small figure compared with emissions from burning fossil fuels, the greenhouse gases released from thawing permafrost will make it that much more challenging to mitigate the worst impacts of climate change. As Anthony noted in a 2018 paper on the subject, the release of methane from thaw events is “likely to amplify climate warming beyond most current Earth system model projections.”

There are several reasons why thermokarst lakes are such large emitters. One is that they are mostly atop what’s called yedoma, carbon-rich permafrost that formed during the Pleistocene more than 10,000 years ago. Unlike soil that has previously been exposed to warming, there’s a huge amount of organic material for microbes to consume. In addition, water itself is good at storing heat, which disrupts the annual freeze-thaw cycle that has defined the region for thousands of years. Lake surfaces still freeze, in most cases, but the water below the surface acts like a layer of insulation even during the depths of winter. This means the permafrost underneath lake beds is thawing at a much faster rate than it normally would—meters instead of millimeters a year, according to Miller. This is what scientists refer to as abrupt thaw, a process that has not been fully incorporated into climate modeling and which could pose the greatest risk to the climate as permafrost disappears. At the same time, the water also creates an anoxic environment, which leads to the production of methane as well as carbon dioxide.

“It is not surprising that we find the occurrence of hot spots to be closely linked with distance to standing water,” Miller told me.

Miller’s research is also focused on how much methane and CO2 will likely be released into the atmosphere from thawing permafrost in the coming decades. He is part of a new collaboration between NASA and the European Space Agency that is working to develop a more comprehensive set of data for the region, which covers approximately 5.5 million square miles and includes parts of the United States, Canada, Greenland, Iceland, Scandinavia, and Russia.

Scientists know far more about Alaska than the rest of the Arctic, Miller said, leaving important questions unanswered. Siberia for example, which makes up 70 percent of the Arctic land area, has not received as much attention. And that makes it difficult for researchers to generalize about the scale of emissions—current or future—from thawing permafrost. The 2019 IPCC Special Report on the Ocean and Cryosphere in a Changing Climate noted that there is “medium evidence with low agreement whether northern permafrost regions are currently releasing additional net methane and CO2 due to thaw.”

But in the case of Alaska, according to studies Miller will publish in the near future, the balance has already shifted: The state has become a net emitter of carbon rather than a place that stores it.

Ted Schuur, professor of ecosystem ecology at Northern Arizona University and cofounder of the Permafrost Carbon Network, has arrived at a similar conclusion. “I’m willing to say with the data we have on hand, there’s a strong suggestion where we’re already underway with this feedback cycle,” Schuur told me. “We have net carbon emissions.” Schuur says that the idea of a “methane bomb” being released as a result of thawing permafrost isn’t all that helpful: The methane and carbon aren’t going to be released all at once, he said, but more likely gradually over time. But the net effect, according to Schuur, is just as troubling.

What does this all mean for efforts to address global climate change?

Most obvious, it only underscores what the science has long affirmed: Human emissions will have to be substantially reduced to avoid catastrophic warming. According to Miller, “We need to be at not only zero emissions but really negative emissions territory to help stabilize the environment.” The uncertainty also leaves open the possibility that methane emissions from thawing permafrost will trigger a feedback cycle that could dramatically accelerate warming. Even if only 10 percent of the carbon stored in permafrost is released into the atmosphere over the next 80 years, Miller pointed out, it would be equal to about one-quarter of what humans have already emitted since the Industrial Revolution.

“I’m not trying to be alarmist but just presenting the information as we understand it better,” Miller told me. “I would say that there is a great urgency in terms of people understanding what’s going on with the climate system, especially carbon feedback and the Arctic. And understanding that there are now some natural events that will take place over which we will have very little control but that we will have to take into account in our planning for the future.”