Iceberg in Mist; painting by Gerhard Richter
Gerhard Richter: Iceberg in Mist, 1982; from ‘Gerhard Richter: Painting After All,’ an exhibition at the Met Breuer. For more on the exhibition, see Susan Tallman’s essay in this issue.

There is a house-sized boulder in the woods near where I live. I can see it through my window, perched at the top of a steep slope. It seems as though it were set there by a giant’s hand; in a sense, it was. The great rock was deposited by the Laurentide Ice Sheet, which blanketed the northern half of North America roughly 20,000 years ago. By 12,000 years ago, the Laurentide had retreated from my corner of northern New England, leaving Lake Champlain and a scoured, striated landscape in its wake. Around that time, the earth’s temperature spiked rapidly.

We know this thanks to samples, known as ice cores, extracted from Greenland’s two-mile-thick ice sheet, parts of which are almost a million years old. There is a huge variety of information locked in its layers—isotopes of oxygen, trapped bubbles of methane, traces of pollutants—from which entire ancient worlds can be inferred and their climates reconstructed.

In his new book, The Ice at the End of the World, the journalist Jon Gertner chronicles the labors of the scientists who drilled and studied those cores in frigid trenches dug out of the ice sheet itself, occasionally pausing to marvel at layers formed from snow that had fallen when Marcus Aurelius invaded Germany, or that contained traces of volcanic dust from an eruption during Caesar’s reign.

Gertner visited the storage facility in Colorado that now houses the cores. Joan Fitzpatrick, the US Geological Survey scientist who runs the facility, showed him one that dates to 11,700 years ago. “Boom. All of a sudden they get tighter here,” she said, pointing to a sharp transition in the layers of ice that marks a warming spike of about 10 degrees Celsius—the same increase that caused the Laurentide to melt. “Ice age here. Not ice age there. We think this was in the space of a few years. And the whole point is, we all once thought it would take thousands of years.”

There are mounting signs that we are living through a similarly wrenching transition in the global climate—one of our own making. But there may be no clearer signal of the scope and speed of this transformation than the accelerating melting of the cryosphere, the frozen realms of our planet. The cryosphere encompasses glaciers, permafrost, sea ice, snow fields, ice shelves floating on the sea, and ice sheets made from snow piled up over thousands and millions of years. Around 10 percent of Earth’s land is covered by glaciers or ice sheets, which hold 69 percent of the world’s fresh water. Arctic ice helps drive the temperature difference between the poles and lower latitudes that is responsible for the jet stream, the stability of which underpins our global food production system. The permafrost that blankets a quarter of the land in the Northern Hemisphere stores—for now—twice as much carbon as the atmosphere.

For most of human history, the perpetually ice-bound parts of the planet have been an enigma. Concerted efforts to get to know them have been relatively recent. Gertner recounts the obsessive efforts of explorers such as Fridtjof Nansen, Robert Peary, Knud Rasmussen, Peter Freuchen, and Alfred Wegener to cross or map Greenland’s unknown interior in the late nineteenth and early twentieth centuries. These are gripping tales, but Gertner’s main project is describing how Greenland would later become a laboratory for exploring time rather than space. He notes that the French ethnologist Paul-Émile Victor—one of the figures who bridged the adventurers’ era with that of the researchers who came after them, riding enormous diesel tractors called “weasels” instead of dogsleds—organized the first large, post–World War II research expeditions into the island’s interior, motivated by his insight that “Greenland was…a tremendous recording machine of times and climates past.” Over the course of his own research into the history of the world’s largest island, Gertner also comes to regard Greenland’s ice “as an analog for time”—one that “seems capable, too, of telling us how much time we might have left.”

“The Greenland and Antarctic ice sheets are continuing to lose mass at an accelerating rate and glaciers are continuing to lose mass worldwide,” according to the authors of the IPCC Special Report on the Ocean and Cryosphere in a Changing Climate, released by the Intergovernmental Panel on Climate Change in September. The key word is “accelerating.” The rate at which Himalayan glaciers are melting has more than doubled since 2000, while the rate of ice loss from the West Antarctic Ice Sheet (WAIS) has tripled. The rate at which Greenland’s ice is losing mass has increased sevenfold since the 1990s. The volume of meltwater pouring off Greenland’s ice sheet is the highest it has been in the past seven or eight millennia; the rate of surface meltwater runoff over the past twenty years is 33 percent higher than the average for the twentieth century. Greenland alone has boosted global sea level by more than half an inch just since 1972; half of that contribution came in the past eight years.


The reason for this quickening meltdown is simple. Our greenhouse gas emissions have warmed the globe by about one degree Celsius since the preindustrial era. Polar and high mountain regions are warming much faster than the global average: temperatures in the Arctic, for example, have risen by one degree Celsius just in the past decade. By mid-century, even if global temperature rise is limited to 2 degrees Celsius, the Arctic will have warmed by 4 degrees Celsius.

Models suggest that both the WAIS and the Greenland ice sheet will be locked into a path of irreversible melt at somewhere between 1.5 and 2 degrees Celsius of global warming. Some researchers believe that the collapse of both these ice sheets is already underway, and unstoppable. Most of the world’s mountain glaciers, and much of its polar ice sheets, are going to melt. The only question is how fast. “The fuse is lit,” says Donald Blankenship, an expert on the WAIS. “We’re just running around mapping where all the bombs are.”

Humans have never inhabited a world without ice. In addition to being archives of ancient climate and thermometers with long memories, glaciers and ice sheets are water tanks and weather engines. They are implacable sculptors of the land we tread. They are the mothers of great rivers. They are objects of obsession and of terror. They are inhospitable and indispensable. The Himalayas, the Andes, the Alps, and the frozen North have long shined in our collective compass as fixed, immutable, monolithic. They have also proven surprisingly vulnerable to the ratcheting rise of atmospheric carbon dioxide—now at levels that are unprecedented in the past three million years—and other warming pollutants produced by human activities.

Glaciologists like to joke that their field isn’t rocket science: as it gets hotter, ice melts. But the actual processes are extraordinarily complicated. Vanishing Ice by Vivien Gornitz, a retired research scientist at Columbia University and the NASA Goddard Institute for Space Studies, is a comprehensive survey of the varied ways ice is turning back to liquid, pretty much everywhere. Laden with diagrams and definitions, Vanishing Ice reads like an intro-level college course on the still young science of why glaciers and ice sheets wax and wane over millennia—and how they can disintegrate quickly in a self-reinforcing process.

Gornitz also makes an extended argument for why everyone should care about all this melting, even those who live far from any ice. As such, she helps us prioritize our cryospheric concerns. Care about sea level rise? Focus on Greenland and Antarctica: combined, they contain about 216 feet of it. Among the third of humanity that depends to some degree on water flowing down from mountain ice? Then pay attention to how very sensitive mountain glaciers are to rising temperatures around the world, and how “their relatively small total ice volume belies their importance to global welfare” as irrigation, hydropower, and cultural resources.

In January 2019 an international consortium of researchers published a comprehensive assessment of environmental change in the Himalaya Hindu Kush region: The Hindu Kush Himalaya Assessment. They concluded that a third of all Himalayan ice will vanish by 2100. That is the best-case scenario, the increasingly far-fetched one in which humankind reins in its greenhouse gas emissions quickly enough to stay under 1.5 degrees Celsius of warming. If emissions continue at their current rate, two thirds of Himalayan ice will be gone by the end of this century. The glaciers on and around Mount Everest are likely to disappear no matter what.

“As the dew is dried up by the morning sun, so are the sins of men dried up by the sight of the Himalaya,” according to the Skanda Purana, an ancient Hindu text. “It would be such a catastrophic, shameful thing to do to melt away Mount Everest,” a prominent atmospheric scientist once told me, describing what motivated his work to reduce the air pollution that darkens Himalayan ice with soot, which is, along with greenhouse gases, a major accelerant of glacier loss.1

Whether or not the sight of the naked Himalaya—whose name means “abode of snows”—one day induces shame, or outrage, catastrophes are certain to follow after the disappearance of ice. As glaciers retreat, they reveal long-frozen soil. That, too, will thaw. As it does, mountain slopes will come unglued. This phenomenon is already fissuring rock faces and toppling mountainsides, causing landslides and rock-ice avalanches in the Swiss Alps and beyond.

Some retreating glaciers also feed growing lakes that are dammed by unstable moraines. When these burst, the resulting floods can be immense. From Nepal to Peru, glacial lake outburst floods threaten downstream villages and infrastructure such as hydropower plants. In 1981 this kind of flood destroyed the China–Nepal Friendship Bridge and killed two hundred people. Since then, glacial lakes have grown in size and number; a study in 2010 found at least four thousand across the Himalayan region, with hundreds considered dangerous.


Water availability will change in huge river basins and irrigation networks. In the Himalaya, glaciers regulate the flow of the Indus, Brahmaputra, Mekong, Ganges, Yangtze, and other rivers that nearly two billion people depend on. These glaciers’ seasonal runoff is an essential complement to monsoon rains. About 40 percent of the dry-season water in the Indus River basin, shared by India and Pakistan, comes from glaciers. In the near term, warming temperatures will cause glacial runoff to spike, and water flowing through the Indus’s tributary rivers will increase. And then, at some point later this century, the flow of meltwater will decline, irrevocably, and slow to a trickle.2

For the 670 million people around the world who live in high mountain regions, from the Andes to the Caucasus, this reduction of glacial meltwater will profoundly disrupt their agriculture, livelihoods, and ways of life. For the 680 million people who live in the world’s low-lying coastal regions, the meltdown threatens to wash away the land under their feet.

The vast majority of the earth’s ice lies at the poles. While mountain glaciers hold less than one percent of the world’s ice, equivalent to a potential 1.4 feet of sea level rise, the Antarctic continent is home to 87 percent of the world’s ice by volume, or almost two hundred feet of potential sea level rise. Greenland has 10 percent of the world’s ice—and over twenty feet of sea level rise.

When it comes to these great ice sheets, what should worry us isn’t so much the prospect of ice slowly melting into the indefinite future, but the mechanics and probability of abrupt collapse. As the geophysicist David Archer warned in his slim, superb book The Long Thaw (2009), “there are reasons to worry that real ice sheets can melt in ways that would evade the current state-of-the-art model forecasts…. Ice knows a few tricks for melting quickly that glaciologists are not predicting in advance, but only discovering as they happen.” A decade later, the authors of the IPCC special report have vindicated him. They admit that previous IPCC projections “have tended to err on the side of caution” when it comes to sea level rise: “Significant sea level rise contributions from Antarctic ice sheet mass loss…, which earlier reports did not expect to manifest this century, are already being observed.”

Both Gertner and Gornitz highlight the research of Eric Rignot, a scientist at UC Irvine and NASA, who has closely studied glaciers in both Greenland and West Antarctica. He doesn’t think current models capture the “threshold behavior” exhibited by large, fast-changing glaciers that terminate in the ocean, such as Greenland’s Jakobshavn and the West Antarctic’s Thwaites Glacier. Recent expeditions have shown that warmer ocean water is lapping at the ice shelves that buttress these enormous seaward-flowing glaciers, chewing at them from below, and lubricating them at the point where ice meets bedrock.

Rignot believes that he and his fellow scientists are witnessing the early stages of their collapse. The rate of ice discharge from the Florida-sized Thwaites, which keeps an enormous amount of upstream ice in the WAIS locked in place, was four times higher between 2003 and 2010 than it had been in 1970. A new study by Rignot and his colleagues at UC Irvine suggests that the floating shelf attached to Thwaites—which holds back the rest of the glacier—could disappear within a few years or decades. After that, the entire glacier (containing enough water to raise seas by two feet) would melt completely within sixty years (worst-case scenario) to a couple centuries (best-case). The glacier will be doomed if its “grounding line”—the boundary between the glacier on land and the attached ice shelf floating on water—retreats past a certain subglacial ridge, beyond which lies a deep, downward-sloping basin that could funnel warmer seawater even farther under the glacier. If Thwaites and its neighbors like Pine Island Glacier go, most of the remaining WAIS would vanish in a self-sustaining process. Seas would rise by ten feet globally.

Rignot’s research also challenges the conventional wisdom that Greenland’s melt will be gradual—as reflected in the IPCC’s latest projection that it will take millennia for large portions of the ice sheet to go. “It’s not like we’re looking at the big taps opening up in the next century,” Rignot says. “They’re going to be opening up in the next decade or so.” The models used by the IPCC suggest we’re in for about three feet of global sea level rise by 2100, with 1.5 degrees Celsius of warming. But Rignot deems those projections too conservative. Thanks to the accelerating melt in Greenland and Antarctica, he thinks it’s more likely that seas will rise five feet by 2100. The total potential sea level rise could be much higher. “There’s not so much we’re going to be able to do to change that,” Rignot says.

Just as glaciologists import terms from medicine and biology—glaciers are “thinning” or “wasting”; when they lose big chunks they are “calving”—climate science more broadly seems to borrow metaphors from the way that ice and water flow. These reports are rife with mention of “cascading effects” and “tipping points,” invisible thresholds past which natural systems topple, often irreversibly, into new states.

Many such tipping points may have already been crossed. Greenland’s accelerating melt is sending fresh water into the North Atlantic. Because it’s less dense than saltwater, that infusion could be contributing to the observed diminishment of the Atlantic Meridional Overturning Circulation, the system of ocean currents that transport heat northward from the tropics, sequesters huge amounts of carbon, and keeps Britain temperate. If this system of currents, already at its weakest point in the past millennium, slows significantly or shuts down, it could lead to more heat being kept in the southern oceans, which would further destabilize the great ice sheets of Antarctica. A weaker system of currents in the Atlantic also leads to less rainfall over the Amazon, and hence more fires—potentially shifting the world’s largest rainforest from carbon storehouse to carbon emitter.

Feedback mechanisms—effects of warming that drive more warming—threaten to send parts of the cryosphere into a death spiral. For example, as glaciers recede, they expose dark rocks near their edges, which soak up the sun’s heat, which spurs more melting to reveal more and more dark land. The shrinking of sea ice creates the same result: more dark ocean water soaks up more heat, melting more ice. Reduced sea ice will warm the Arctic further, possibly fueling the “dieback” of boreal forests—as rates of regeneration are outpaced by tree loss from spreading fires and pests—and accelerating permafrost thaw.

Photograph of birds on an iceberg in Greenland

Ulrike Crespo

Photograph by Ulrike Crespo from her book Grönland (Greenland), just published by Kehrer

Even if these thresholds aren’t crossed anytime soon, the forecast for the frozen Arctic tundra is about as grim as the prognosis for mountain glaciers. About a quarter of permafrost will thaw by 2100 even if warming is limited to well below 2 degrees; if emissions continue at their current rate, close to 70 percent of permafrost will melt. The most visible consequences are heaving house foundations and “drunken forests”—trees tilting at wild angles as the ice wedges under their shallow roots melt and soil subsides—which can already be found in Alaska and Siberia, along with the rapid spread of “thermokarst” lakes forming in some spots where permafrost melts.

The hidden consequences are much more alarming. As permafrost thaws, carbon in frozen soils will be digested by microbes and released as carbon dioxide or methane.3 If a significant fraction of that carbon is exhaled as methane—a super-potent warming agent—then temperatures could spike sharply over a short time period. There is little evidence that this has begun. But it’s the kind of “carbon climate feedback” that keeps climate modelers awake at night, partly because most climate models don’t tend to include this mechanism. (As one climate modeler once told me, “If we can’t quantify something very well, we tend to ignore it.”) Even if this enormous methane “pulse” doesn’t happen, permafrost is scheduled to continue adding carbon to the atmosphere for centuries to come. One permafrost expert, Kevin Schaefer, estimates permafrost carbon feedback alone will increase temperatures by 0.2 degrees Celsius by 2100, and—inevitably—more beyond that point.4

In The End of Ice, Dahr Jamail recounts a series of pilgrimages to the doomed glaciers of the Alaskan Range, the Cascades of Washington state, and Glacier National Park in Montana. He visits the Pribilof Islands in Alaska’s Bering Sea, and listens to locals describe how rising ocean temperatures and dwindling sea ice are killing off the seabirds, seals, and other wildlife that make subsistence there possible. (Ice, it’s worth noting, is not the only thing that ends in The End of Ice; Jamail also travels to warmer climes to document and foretell the end of coral, the end of forests in the American West, the end of the Amazon, the end of Miami Beach.)

Jamail’s spare prose at times veers into cliché (“But now, these frozen rivers of time are themselves running out of time”) and maudlin personal reflections. But he ably renders moments of grief and outrage, through moving testimony from indigenous inhabitants of the far north and brutally candid assessments from the dozens of scientists he interviews. Toward the end of The End of Ice, Jamail travels to Utqiagvik, a 1,500-year-old Inupiat village (for a while known as Barrow) in Alaska’s North Slope. The northernmost settlement in the US, it’s a stop-off point for flights carrying workers heading to the oil fields of Prudhoe Bay. Elders there have watched in recent years as storm surge has chewed away at the berms protecting the village from the encroaching Chukchi Sea.

Jamail visits Vladimir Romanovsky, a permafrost specialist doing research in the area. Over his thirty-five years of temperature monitoring, Romanovsky has seen a staggering rise of 3 degrees Celsius in permafrost twenty meters below the surface. At this rate of warming, the permafrost will rise above 0 degrees Celsius by mid-century at the latest. “Nobody was expecting this, and most people would be surprised to see this happen so soon,” he tells Jamail. Since that conversation, Romanovsky has published a study documenting thawing at permafrost sites in the Canadian Arctic to depths that weren’t expected, according to most models, until 2090.

In the most affecting passage of The End of Ice, Jamail visits Wesley Aiken, the ninety-two-year-old town elder. Aiken describes the changes he has seen in Utqiagvik. The main one is that it used to be “far, far colder.” The walruses don’t show up as often because the sea ice is less reliable and doesn’t last as long. Snows start not in August, as they used to, but in October:

“All the ice is melting,” he says and just looks at me to let the weight of his statement sink in. He speaks slowly. He is from a world where there was never a need to rush anything. “The ice used to hang around here all summer when I was young. The ocean is now eroding the coast. The waves are getting bigger and rolling into the coast. I think we’ll have no more Point Barrow before much longer.” …He tells me the permafrost is thawing, that while it used to be only a foot below the surface, it is now four and even six feet down. “I can see more green grass out there in the tundra,” he says, pointing out the window.

The passage brought to mind an eerily similar conversation I had several years ago with a man named Ishay Paldan, the eldest resident of the one-thousand-year-old Himalayan village of Kumik, in north India. As we sat in his kitchen, he pointed out his window and described how a now-distant glacier—the sole source of water for Kumik—once covered the rock and grass on the slopes above. “When I was a child, there were no problems with water,” he told me. “The snow line then almost came down to the top of the village. Now look.”

“It’s all changing,” Aiken concludes, echoing the message of so many scientists who study the world’s ice. “We know this is happening…and I don’t think it’s going to stop.”

Climate models and emissions projections tend to treat the year 2100 as some kind of deadline or denouement. But time, of course, doesn’t stop at 2100. Most of the carbon dioxide that we emit today will still be in the atmosphere, trapping heat, in the century to come. The oceans will absorb some of that heat, and will nibble at the ice sheets that abut them. The warming will continue long after. The seas will keep rising. Permafrost will go on thawing and releasing its long-stored carbon. The Himalaya will lose most of its ice. The ice sheets will continue to lose mass. It’s not alarmism to acknowledge all this. It is simply what will happen.

The authors of the IPCC special report dryly note that all of these changes “occur on spatial and temporal scales that may not align with existing governance structures and practices.” Some will play out over millennia; some may come much sooner. Regardless, we aren’t remotely prepared—logistically or psychically—for any of them.

“Sometimes, the ice sheet has also struck me as the photographic negative of an ocean,” writes Gertner, reflecting on his many visits to Greenland. “Rather than darkness streaked with white foam, it is lightness streaked with silt and dust. Even over the course of a few years, I could see it thin and recede.” He writes of watching the appearance of “new land”—a technical term, apparently, describing the barren till that is so loose it blows into the eyes, mouths, and instruments of researchers—revealed as glaciers recede. And as that new land emerges, thousands of miles away some old, beloved land succumbs to newly formed seawater.

The irreversibility of these changes might provoke despair, a sense of futility. It would be wrong, and dangerous, to let our response stop there. All of these studies arrive at the same, now familiar conclusion: reducing emissions of greenhouse gases is the only way to slow down and perhaps shrink the magnitude of this loss. For the next few years, maybe a decade or two, we can still influence the rate of melting through swift action.5

In industrialized societies, we don’t tend to think of ourselves as ancestors—in the deep, many-generations sense—but that’s what the moment demands. The project before us is paradoxical: thinking on long time scales while acting with furious urgency.

In the closing pages of his book, Gertner notes that Greenland’s ice cores contain traces of lead dating to ancient Roman smelters and to the dawn of the Industrial Revolution in the early eighteenth century. And then, in ice cores from the 1980s, one can read the signs of another abrupt transition: the moment when the US decided to phase out lead in gasoline. The ice got clean. The mirror of ice offers this other kind of reflection, too. It reminds us that we are capable of our own “threshold behavior.”