From the air, the western edge of Greenland’s ice sheet looks like an aging elephant’s skin–gray and cracked, as melting at the margins exposes dirt accumulated over tens of thousands of years. A few kilometers inland, however, the gray gives way to blazing white that goes on for hundreds of miles. These vast glaciers contain snow that has fallen and compacted over millennia. The ice sheet–roughly four times the size of California, and more than three kilometers thick in places–is, in essence, a vast frozen reservoir of fresh water.

About 300 kilometers from the coast, high above the Arctic Circle, is the newest polar science station in Greenland, comprising two tents, a hut, two sledge-mounted domes, and a few vehicles. The only way to describe the location is by GPS coördinates: 77º26’54.92885” N, 51º3’19.89396” W. The station sits atop 2,500 meters of ice.

The modest encampment, manned by a nine-person scientific and support crew from a Danish-led international team, might be in the middle of nowhere, but it is also at the center of a growing effort to answer the crucial question of global warming: how quickly, and by how much, will sea levels rise? The two great polar land masses, Greenland and Antarctica, together hold nearly 99 percent of the planet’s landlocked ice, which is capable of raising sea levels when it melts (the melting of sea ice, such as that floating in the Arctic, does not raise the level of the oceans). The giant ice sheets hold enough frozen water to raise sea levels some 80 meters.

If only 10 percent of this ice melted, it would flood the world’s coasts at levels comparable to those seen in post-Katrina New Orleans. While nobody is predicting a catastrophe on that scale anytime soon, scientists are concerned that melting might greatly accelerate as the planet warms. Especially worrisome is a scenario that glaciologists and climate scientists are still piecing together: rather than slowly but steadily melting, the ice sheets could rapidly break apart. Recent observations show that some of the major glaciers on Greenland and the West Antarctic Ice Sheet are sliding ever faster seaward. But the processes involved are not well enough understood to be incorporated into the computer models used to predict how much sea levels will rise in response to climate change.

Gauging the risk that the ice sheets will break apart–and estimating how fast such a breakup would raise sea levels–will require a far better understanding of geology. Not all the bedrock beneath Greenland and Antarctica is mapped. Nobody knows how much liquid water lies under the ice; even a small amount could dramatically speed the breakup of the ice sheets by making the surface below them much more slippery. Across both land masses, scientists are striving to make more precise measurements. Some are busy installing GPS stations on the ice sheets and the bedrock surrounding the coasts to more accurately calculate loss of ice mass. Others are measuring snowfall accumulation and studying how snow compacts into ice. In this way, they are trying to estimate just how much inland ice there is–and by extension, how much has fallen into the ocean.

“We need to go right back to the drawing board on what the ice sheets are about,” says Ted Scambos, lead scientist at the National Snow and Ice Data Center at the University of Colorado at Boulder. “Fifteen years ago, we thought ice sheets wouldn’t respond quickly to global warming because the melting would happen at the surface. This was true, but what we didn’t count on was fracturing. This permits water to get to the base of the ice, all the way through the ice sheet. We were really surprised to see this even where the ice core is well below freezing. The water allows glaciers to flow more rapidly, dumping the ice into the sea.”

A Mystery
The Intergovernmental Panel on Climate Change (IPCC), the international panel of scientists that weighs in every five years or so on the state of the climate, predicted earlier this year that seas would rise between 18 and 59 centimeters this century. While the higher figure is quite worrisome in low-lying coastal areas, the numbers are still small enough–and the prospect far enough in the future–to seem manageable for most of the world. But the IPCC, in explaining its numbers, added that it could provide neither a “best estimate” nor an “upper bound” for how much higher sea levels might rise if the ice sheets disintegrate. (See “Sea-Level Riddle.”)

Indeed, the IPCC’s sea-level estimates are based on math that takes into account only a few well-understood processes. One is the expansion of seawater as it warms. Another is the melting of mountain glaciers in temperate zones–places like the Alps, Andes, and Himalayas. Third is the melting of the ice sheets’ surfaces and the glaciers’ seaward migration under the pull of gravity, though this may be partly balanced by increased snowfall that occurs because warmer air holds more moisture. The problem is that other processes may actually prove far more consequential. Warmer and higher oceans undermine the glaciers that flank Greenland and Antarctica, yanking them away from their seabed moorings. With those bulwarks weakened, inland glaciers slide much faster toward the sea. In a complementary process, water gushes down through fractures and holes in the Greenland ice sheet, making it easier for its glaciers to slide. As the glaciers reach lower (and therefore warmer) elevations, the melting and sliding will further accelerate.

And that’s exactly what appears to be happening. Recent observations have shown that the movement of major glaciers in both Greenland and the West Antarctic Ice Sheet is in fact accelerating. To pick just two examples, by 2005 a vast glacier in Greenland, the Jakobshavn, was slipping seaward twice as fast as it had in 1996, and another, the Kangerdlugssuaq, was slipping three times as fast as it had in 2000. “The current dynamical changes that we are seeing on the ice sheet are not captured in any climate model,” says Prasad Gogineni, director of the Center for Remote Sensing of Ice Sheets at the University of Kansas, which is a participant in the ­Danish-­led effort. “That seems to indicate a huge uncertainty.” Today’s climate models, he says, simply can’t be relied on to predict what will happen to the great ice sheets.

Greenland’s melting ice sheet is now contributing more than half a millimeter per year to sea-level rise, according to a study coauthored by Eric Rignot, a senior research scientist at NASA’s Jet Propulsion Laboratory who has collaborated with Gogineni and others at the University of Kansas. That seemingly small figure is noteworthy because it’s more than twice the upper limit of Greenland’s contribution as estimated by the IPCC in its report earlier this year. “The [existing models] don’t assume any change in velocity in the glaciers, except on very long time scales,” Rignot says. “What we are seeing today is that those glaciers do speed up in a significant fashion in response to climate warming.”

The melting of the ice sheets is really just getting started, says Rignot. The current surge in the velocity of the glaciers has accompanied the 0.74 ºC of warming the planet has seen over the past 100 years. The IPCC is predicting that global temperatures will rise far more in the next 100 years–1.8 ºC to 4 ºC, depending on future emissions of greenhouse gases. The resulting loss of ice will dwarf any increases in snowfall, Rignot says. But Rignot disagrees with the conclusions drawn by the IPCC: he believes oceans will rise more than a full meter before the end of the century, nearly twice the upper bound of the IPCC’s predictions. “We have to acknowledge that we don’t have reliable ways to predict what ice sheets will do, but that they will certainly react much more strongly to climate warming in the future,” he says. “There is no reason to alarm people that the end of the world is coming. But there is no reason to reassure them, either, that there is nothing to worry about with the ice sheets.”

Drilling Down
With advanced radar technology, researchers at the north Greenland site are producing the first detailed pictures of large areas of the ice-sheet base, with particular attention to pockets of water. Previous technology could detect large lakes underneath the ice. But the new technology, honed at the University of Kansas and deployed by researchers from both Kansas and the University of Copenhagen, can detect even a few millimeters of water, which are just as dangerous. What’s more, whereas previous radar equipment took measurements only of the ice directly beneath it, the new technology also provides information about ice layers and the ice sheet’s base in a three-kilometer-wide swath of ice cap.

During the winter of 2005-2006–summer in Antarctica–scientists from Kansas lugged the new system down to the West Antarctic Ice Sheet and collected data on a 30-by-10-­kilometer grid. Early results show much more detail about which parts of the ice sheet’s base are sitting on water and which are still frozen to the ground. Whether the water came from geothermal heat, friction from ice in motion, or accumulation of surface meltwater is not yet clear. But the new data should help improve ice-sheet models, says Claude Laird, a research scientist at the University of Kansas and a member of both the Antarctic and the north Greenland expeditions. This summer, Laird and the other scientists used the technology to map a 370-kilometer swath of Greenland. When the results start coming in, they should give a clear picture of that swath, and of how much water lies within the ice or beneath the sheet.

Meanwhile, the scientists are seeking clues from the past. At the north Greenland base station, amid the huts and vehicles, an aluminum pole is staked to the ice cap. Next summer, the scientists will return to the spot and start drilling out an ice core, boring about 2,500 meters to bedrock. They are particularly interested in one key geologic period, called the Eemian interglacial. During this stretch of time, from around 130,000 to 115,000 years ago, the planet warmed. Greenland hit temperatures 7 to 8 ºC warmer than today’s, and sea levels surged at least three to five meters higher than they are now. If this happened today, much of south Florida, Bangladesh, and many other low-lying coastal areas and islands would be submerged.

The warming during the Eemian period was caused by eccentricities in Earth’s orbit that periodically allow more solar energy to hit the planet. An understanding of how the climate and ice sheets responded during the Eemian warm-up should sharpen our picture of how they’ll respond today. Already, a record of Eemian sea levels is available from existing geological sites. But scientists would like to know more about short-term climatic variations–coolings and warmings–within the Eemian period. That should help them better understand current climate changes and more accurately predict how sea levels will rise.

Glaciers accrete over time, and different layers contain records of Earth’s past climate. A ­sample of ice spanning the whole Eemian period–never before found in the northern hemisphere–would provide a wealth of information. Identifying oxygen isotopes within water molecules can reveal what temperatures prevailed when snow fell. Trapped air bubbles inside ice contain samples of the old atmosphere. The thickness of ice layers can reveal how much snow fell. And bits of dust and organic matter will allow accurate dating. After conducting radar analysis, the researchers at north Greenland think the spot marked by the aluminum pole–dubbed NEEM, for “North Greenland Eemian”–will contain a 120-meter-thick chunk of ice representing the entire period (see “Quest for Ancient Ice”). “We can get even better data on these fast climate oscillations,” says Laird. “And we can get some forecasting about what climate change will mean.”

Stuck
In north Greenland, the science is done for the season. It had taken the team of scientists nine days to reach the site–a 370-kilometer slog, dragging thousands of kilograms of equipment, fuel, and food and using two snowcats, three snowmobiles, and a Toyota Land Cruiser outfitted with tracks. Their trip was plagued by delays: at one point, a blizzard had them hunkering down for days; at another, two of the Land Cruiser’s tracks broke. After more than four weeks in the field, the scientists waited to be evacuated. Though it was only August, dangerous weather loomed, and they were anxious to get home and analyze data they’d gathered during their weeks of work.

But first they had to get off the ice sheet. On a sunny afternoon–with temperatures reaching -4 ºC–a ski-equipped cargo plane made a soft touchdown. After disembarking, the pilot, in his olive-green jumpsuit and wraparound sunglasses, kicked worriedly at a new layer of snow.

Three hours later, the nine scientists and crew members had boarded the cargo plane. The plane labored up and down the snowy runway. But just as the pilot had feared, the snow was too soft for the plane to reach takeoff speed. Departure would need to wait 12 hours, until 4:00 a.m., the coldest time of day. At the appointed hour, they tried again. This time, ice was frozen to the bottom of the plane’s front ski. The pilot’s efforts to shake it loose, using hydraulics to move the ski up and down against the snow, broke a pin that held the hydraulics in place. Another plane needed to deliver the replacement parts.

Finally, around 11:00 p.m. that night, a second cargo plane landed. It was nearly 20 ºC below zero. A crew from the second plane sprinted across the ice, fixed the ski, and took off again with a rocket-assisted flourish. And finally, the original plane followed suit, skating quickly across the icy surface. Amid the roar of side-mounted rocket motors, the researchers and the crew made it aloft at 2:30 a.m., a day and a half after their first try. Fortunately, the ice sheet was frozen solid. For now.

David Talbot is Technology Review’s chief correspondent.