The permanence of frozen ground in the Arctic is no longer guaranteed as Earth’s temperatures continue to climb. But how much the degradation of so-called permafrost will worsen climate change is still unclear, according to the Intergovernmental Panel on Climate Change’s (IPCC’s) Sixth Assessment Report, released this week. The uncertainty leaves researchers with a frustrating hole in their climate projections.

Permafrost covers a quarter of the Northern Hemisphere’s land and stores around 1.5 trillion metric tons of organic carbon, twice as much as Earth’s atmosphere currently holds. Most of this carbon is the remains of ancient life encased in the frozen soil for up to hundreds of thousands of years.

In recent decades, permafrost has thawed because of global warming from heat trapped primarily by carbon dioxide released to the atmosphere from burning fossil fuels. Arctic warming is rising at twice the global average rate since 2000, according to the National Oceanic and Atmospheric Administration. As that increase accelerates the thaw of permafrost, the organic carbon contained within it breaks down and releases carbon dioxide, exacerbating climate change.

But climate scientists are unsure how much carbon will be released from permafrost and when, which is reflected in the wide range of estimates provided by this week’s IPCC report. This uncertainty hampers climate change projections, making it harder to know whether the world’s nations are on track to meet targets designed to limit global warming established in the 2015 Paris Agreement. Creating policies to meet those targets hinges on a precise understanding of how much carbon dioxide enters the atmosphere each year.

Most existing climate models do not currently account for carbon released from permafrost in their simulations. The IPCC’s latest report instead makes a best estimate of the range of carbon that permafrost could potentially expel. It then accounts for that range when estimating the world’s remaining carbon budget—the amount of CO2 that the can still be emitted—for meeting Paris Agreement targets, says Charlie Koven, a carbon cycle scientist at Lawrence Berkeley National Laboratory and one of the lead authors of the report. While not ideal, this approach is “a reflection of the urgency of the climate crisis,” he says. “We don’t have time for a perfect solution. We need to act on the knowledge that we do have.”

Meanwhile researchers are trying to better comprehend permafrost’s contribution to global warming. The key is Earth’s carbon cycle, wherein carbon is exchanged among the land, water, and air. In the warming Arctic, two opposing influences are changing the cycle. As the soil thaws, bacteria, fungi and other microbes that live in the soil consume exposed organic matter and belch carbon into the atmosphere. And microbial communities better thrive in a warmer Arctic, increasing both their numbers and their appetite. This potentially creates a so-called feedback loop: climate change causes more carbon dioxide to be released, which worsens the problem globally and thus sparks the release of more carbon. Plants typically grow better in warmer temperatures and carbon-rich air, however. That growth pulls more carbon from the air and deposits it into the soil when the plants die. So as Earth’s permafrost warms, “you’re asking, ‘Who’s winning?’ We don’t immediately know,” says Ted Schuur, a professor at Northern Arizona University, who studies Arctic ecosystems.

For the past 20 years, Schuur has operated a research site near Denali National Park in Alaska to try to identify the victor by monitoring the exchange of carbon dioxide between the soil and air. This spring he reported something concerning in the journal JGR Biogeosciences: the soil microbes are winning, meaning that the small area surrounding Schuur’s study site is a net source of CO2.

At this point, “it’s hard to imagine a scenario in which the growth of plants is able to overtake the loss of carbon from permafrost soils,” Koven says. But to better pinpoint the amount of permafrost carbon that could be released throughout the 21st century, “there’s a lot more work to be done.”

While other CO2-monitoring sites have also recently shown the microbes are winning—for example, in findings published this month in Environmental Research Letters—there are too few sites to confidently represent all of Earth’s combined permafrost regions, which span about 22 million square kilometers. Very few of these research sites are in Siberia, the world’s largest region of permafrost, making it “a black box,” says Jennifer Watts, an Arctic systems researcher at the Woodwell Climate Research Center in Falmouth, Mass.

The Arctic’s harsh, remote landscape makes it challenging for researchers to venture out and set up more sites. One possible remedy is to observe the thawing permafrost from afar with satellites. Scientists have already used satellites to determine that vegetation cover has been increasing across most of the Arctic, although remotely monitoring carbon release requires more technological finesse.

Sparse sampling is not the only reason for uncertainty about thawing permafrost’s contribution to climate change. Arctic scientists also continue to discover alarming new twists in the carbon cycle story. In 2019 Watts and other scientists reported in Nature Climate Change that microbes at many study sites remain somewhat active even during winter, thus releasing some amount of carbon dioxide year-round. Separately, in Nature Geoscience the following year, Schuur, Koven and other researchers tied thaw lakes—expanding pools of meltwater from thawing ice-rich permafrost—to the release of bubbles of methane, a carbon compound with a more potent greenhouse effect than carbon dioxide. “Without those bubbles, you underestimate the carbon feedback,” says Katey Walter Anthony, a professor of ecosystem ecology at the University of Alaska Fairbanks and a co-author of the thaw lakes study.

Even today methane release from the warming Arctic remains “a huge question mark,” Watts says. “We know there’s methane. We don’t know how much, and we absolutely don’t know what it’s going to look like in the future.”

Although the globally averaged temperature has already increased by more than one degree Celsius from preindustrial levels, “that’s relatively small, compared to the possible changes that could await us,” Koven says. Climate scientists’ go-to strategy for better understanding the permafrost implications of Earth’s warmer future has been to examine its warmer past. Over the past couple of million years, temperatures have occasionally climbed slightly higher than today. “Each of those warm periods is a great natural experiment,” says Alberto Reyes, an associate professor at the University of Alberta, who studies how ancient permafrost responded to these past warm periods.

This spring Reyes and others analyzed Canadian Arctic and sub-Arctic cave deposits, such as flowstones and stalagmites, and determined that Arctic permafrost soils thawed quite a bit during some ancient warm periods. But based on records from ice cores from the Antarctic and Greenland ice sheets that contain trapped air bubbles from the distant past, no spikes in atmospheric carbon dioxide or methane seemed to follow in response.

“Where are those greenhouse gases?” Reyes says. The answer is not necessarily that they weren’t emitted from thawing permafrost. The oceans in the ancient past had more time to suck up carbon dioxide released from permafrost than they will this time around. Today’s levels are much higher than those of the past two million years, and they are rising at a faster rate. “We’ve pushed the system so far already,” he says. “Humans have created a nonanalogue condition.”