14 Sensing

Forty years after he crossed the Greenland ice sheet on skis, Fridtjof Nansen sailed on an ocean liner from Gothenburg, Sweden, to New York City. “The dean of living explorers,” as one newspaper described him, had come seeking support from American officials for a new project that was being hailed as “the greatest exploration of his long career.”1 Nansen was old now and fatigued from a life of travel; his thick moustache and the fringe of hair ringing his bald head were silken and white. This very year, 1929, would prove to be the last of his life. His mood upon arrival in New York was nonetheless upbeat, and his enthusiasm and stature—he was now a Nobel laureate as well as a global celebrity—impressed those who crossed paths with him. His intensity still burned. Wearing a wide-brimmed black hat and speaking in nearly flawless English, Nansen answered reporters’ questions about his new organization, the Aeroarctic association. His goal for the following year, he said, was to launch a scientific expedition from the air to study the geography, oceanography, and meteorology of the Arctic. The work would be conducted from an aerial perch aboard the huge German dirigible known as the Graf Zeppelin.

These kinds of airships—kept aloft by hydrogen-filled tanks and limited to a top speed of about eighty miles per hour—had already been used for polar travel. Several had sailed over the North Pole a few years before.2 But Nansen wasn’t interested in that kind of journey. Instead he made the case to journalists that the weather and ice of the Arctic had profound effects on Europe and the lower latitudes. “We’re thinking of study up there for the population of the rest of the world,” he explained.3 Using an aerial approach for purely scientific purposes, Nansen also seemed to think, would mark a serious technological leap. “Should the Graf Zeppelin flight prove successful,” the American journal Science noted, “it would doubtless be the forerunner of future Arctic flights, on still larger airships, thus exploring thoroughly all the Arctic regions in a far more complete manner than is possible on the ground.”4

As usual, Nansen had seen the future before almost anyone else. And while his death several months later shook colleagues who were helping to plan the Graf Zeppelin journey, the project moved forward anyway. In July 1931, the airship left from Leningrad, flew over the Siberian Arctic, and then reached a point 560 miles from the North Pole before turning east and returning south to Leningrad and Berlin. A team of technicians, using panoramic cameras for surveying and custom-built instruments to measure earth’s magnetism, carried out a hectic five-day scientific program. The journey, as it turned out, did not set off a revolution in Arctic science—at least not right away. In the decades following the Graf Zeppelin’s flight, mapmakers increasingly relied on aerial photography for charting northern islands and the coastline of Greenland, but these instruments were mainly used for military purposes. In the years of World War II, pilots intent on finding targets for bombing runs made detailed photographs of enemy terrain and factories. In the Cold War era that followed, similar techniques were used by American U-2 planes, developed in the mid-1950s, which could cruise at seventy thousand feet, snap detailed photographs of Soviet installations below, and usually evade detection.

There was no precise moment when scientists began borrowing the techniques of military intelligence for their own work. But by the start of the satellite era—beginning with Sputnik in 1957—a number of American and European researchers tried to envision the stratosphere as a vast new playing field. In 1960, NASA started launching observation missions, first with the TIROS satellites, and later with a series of more sophisticated “birds” built under the Nimbus program, which began in 1964. For the most part, the goal was gathering weather data; the early Nimbus satellites—about twelve feet tall and weighing about eight hundred pounds—took photographs of the earth while traveling at an orbit that ranged from two hundred and fifty to six hundred miles up.5 Within a decade, these kinds of satellites were outfitted with exquisitely sensitive cameras and instruments that used microwave and infrared sensors to “read” the surface of the planet beyond the spectrum of visible light. The new equipment could relay information to engineers on the ground about earth’s temperatures, clouds, winds, and sea ice. As a matter of course, those working in the emerging field of “earth observation,” as researchers began to call it in the 1960s, described the purpose of these satellites as “remote sensing.” It was an interesting phrase. To find out about the farthest reaches of the planet, the term suggested, you might not even have to leave your office chair.

Remote sensing held tremendous promise for polar research—these were “powerful new tools” for measuring snow and ice, in the words of one scientific report of the era.6 Sensors could conceivably get carried into an orbit over the poles by satellite, or they could be mounted under airplanes flying over desolate regions of Greenland and Antarctica. They could help survey regions where it was dangerous to launch fieldwork studies or regularly assess areas where it had proven almost impossible to gather a broader understanding of changes in ice, snow, or oceans. In 1968, Henri Bader, the father of ice-core drilling, chaired a symposium in Easton, Maryland, about the future possibilities of remote sensing in the polar regions.7 A year later, at a NASA-sponsored meeting in Williamstown, Massachusetts, a group of academics put forward recommendations for new sensing instruments that would be able to measure the sea levels of oceans and the movements of ice masses on Greenland and Antarctica from the sky.8

It would become clearer, in retrospect, that those imagining the potential for remote sensing in the late 1960s were dreamers. They knew some of their hopes exceeded the ambit of existing technologies. Still, by the late 1970s some of the new equipment—ice-penetrating radar, for instance, carried aloft by aircraft—was proving useful in studying the Greenland ice sheet. Before drilling the Dye-3 and GISP-2 holes, engineers conducted flyover missions with this type of gear, which could essentially see through the island’s white blanket and map the topography of Greenland beneath the ice.9 The idea was to improve the chances of finding an ideal place—level bedrock, which should produce smooth and even layers in the ice sheet—to drill. In this regard, remote sensing became a friendly partner to deep core drilling: To find the best place to dig up the past, one could first look down from high above.

In theory the new sensing technologies could do far more than that, though. Going back to the age of Nansen’s first crossing on skis, the mysteries of Greenland that originally interested explorers were the ice sheet’s origins, extent, and weather. Later came the effort to decode the secrets locked in its deepest layers. By the late twentieth century, researchers had largely realized these goals, but now glaciologists had arrived at the brink of a third mystery: Was the ice sheet growing or shrinking? What sounded simple was in fact the hardest of problems, one that had resisted a sure answer for decades. Due to the size and remote nature of the island, going out into the field to measure its total ice was akin to standing in the midst of a dense forest on an autumn morning and trying to count every fallen leaf. The question had nevertheless tantalized Alfred Wegener and Fritz Loewe, who wondered if Greenland’s ice sheet would eventually face the same fate as the enormous glaciers that had once covered much of Europe and North America, before they’d shattered into the oceans. The question had compelled Carl Benson, who in digging snow pits all over Greenland in the mid-1950s had tried to estimate whether annual snow accumulations on the ice sheet outweighed its annual losses from meltwater and icebergs. The problem had fascinated Paul-Émile Victor, who after his early expeditions in the mid-1950s helped plan another expedition around Greenland to calculate the “mass balance” of the island’s ice.10 Henri Bader, too, made an estimate for the ice in Greenland, in 1961. At the time, he thought the ice sheet might be growing rather than shrinking.11

In truth, there had never been a satisfactory answer. In the late 1960s, Børge Fristrup, a Danish glaciologist who wrote a history of the ice cap, considered the previous attempts to answer whether or not the ice mass of Greenland was in balance and concluded that “it has been assumed that the ice is in a state of equilibrium…in fact little is known about whether this is indeed so.”12 Not much had changed by the mid-1980s, when a NASA report urging the development of remote sensing satellites put it more bluntly. “Despite 25 years of intensive field work in Greenland and Antarctica, and the expenditure of billions of dollars, we are still unable to answer the most fundamental glaciological question: Are the polar ice sheets growing or shrinking?”13 The question was not only scientifically compelling; it was important for practical reasons. At a minimum, a diminishing Greenland ice sheet would mean rising sea levels and imperiled coastal cities.

Scientists working in Greenland around this time had little reason to believe that the ice was in immediate jeopardy; some of the glaciers on the west coast actually seemed to be advancing, meaning the huge ice mass could be expanding. What’s more, the notion of abrupt climate change had not yet become a concern. During the drilling of Dye-3, the Danish scientist Dorthe Dahl-Jensen remembers, “At that time, there was little understanding at all of the melting. We were still talking about: When will the next ice age come?”14 While working on GISP-2 a few years later, Joan Fitzpatrick recalls, “Certainly the people who had proposed the whole drilling project used this as a justification: What if? What if global warming continues? What will be the impacts on the Greenland ice sheet and how will it respond? But it was all very much in the abstract.”15 Without a bigger picture, in other words, and without a sense of urgency, there was a pastiche of observations, speculations, and computer projections. These were only educated guesses as to how ice might behave in a hotter world.

One of the scientists who authored the NASA report asking why billions of dollars had not yielded an answer about the state of the ice sheets was a glaciologist named Bob Thomas. A native of Liverpool, England, Thomas had earned an undergraduate degree in physics but decided he was far more interested in geology. In 1960 he signed up for a scientific tour and traveled by ship from England to the Falkland Islands, and then to Antarctica. The passage there took two months. “We were dumped off for a couple of years at a base on the west side of the Antarctic Peninsula, a place called Galindez Island, around about the Antarctic Circle, just north of the Antarctic Circle,” he would later recall.16 There were thirteen people there. Mostly he took meteorological readings, but from his colleagues and a few visitors he began to learn how to measure glaciers. It happened to be a transitional era in polar studies, a time when premodern customs lingered on. “We still used dogs for transport, so it was good old explorer stuff,” Thomas recalled—although by the time he returned to Antarctica, in 1966, for another two-year stint (this time with the British Antarctic Survey), there were also a few snowmobiles at hand. He was focused by that point on the movement of ice sheets and the behavior of ice shelves, the tongues of large glaciers that extend past the edge of land and float on the ocean, sometimes for many tens of miles. A lot of the time Thomas’s work had involved putting large aluminum or bamboo stakes into the ice and tracking the movements of these stakes over time, which gave insights into how glaciers crept forward and back. Those who knew Thomas considered his energy and stamina almost superhuman. In another era, he might have been the type to sign up with Nansen or Peary to cross Greenland’s ice. As it was, he returned to England after his tours of Antarctica, got a PhD at Cambridge, and entered a life of academic research, mostly in the United States.

In the early 1980s, Thomas landed as a program manager at NASA headquarters in Washington, D.C. His job was to fund promising research ideas that advanced the state of knowledge about the earth’s cryosphere, or its ice-covered areas. One day at NASA headquarters in 1988, a team of NASA technologists, led by a man named William Krabill, gave a presentation to a few administrators. They were trying to see if there might be some interest among upper management—and, possibly, some funding—to expand their research work. Bob Thomas was there that day and listened closely.

Krabill was based at the Wallops Flight Facility, a NASA airfield on Virginia’s sandy eastern shore, which lies several hours east of Washington, D.C. He and a few colleagues had been experimenting over the past year or two with a device called a laser altimeter. Essentially, a laser altimeter is an instrument that’s used to measure the height of something far away; in Krabill’s work, he was measuring Virginia’s Chincoteague Bay from a plane flying above and sending laser pulses down. By calculating the time it took for the light pulses to leave his instrument, reflect from the water surface below, and return to the aircraft, the altimeter could get a fairly exact calculation of the plane’s distance from the water surface. Listening to Krabill talk about this airborne sensing equipment that day in 1988, Bob Thomas became interested. Actually, he became very interested. He suspected that climate might already be having an effect on the polar ice; he was also thinking of his years doing fieldwork on ice sheets that were “so bloody inaccessible,” as he liked to put it, that their titanic size made it impossible to see how they were faring. That day, Thomas told Krabill that if he could make his laser altimeter work to within a resolution of twenty centimeters—meaning that from a plane flying, say, twelve hundred feet above, it could make measurements to help him calculate the elevation of ice below to within eight inches—he would make a major contribution to science. Krabill recalls that he vowed to get the resolution even better, to ten centimeters. “Thomas gave us some seed money,” Krabill says, “and we did some local test flights. Then we took off for Greenland in 1991.”17

Breakthroughs don’t necessarily happen because of one big technological jump. More often it’s because a cluster of new technologies and ideas suddenly coalesce around a difficult problem, along with a headstrong person. That seemed to be the case with remote sensing. Without question, a reliable aircraft—a Navy P-3 surveillance aircraft—that could fly long distances was essential to making Thomas’s plan work. Also, Krabill’s laser altimeter, a sophisticated tool bolted to the floor of the plane, could now help researchers make fairly precise ice elevation measurements over tens of thousands of square miles. But the device was only useful for measuring the changes in the ice sheet because a host of new navigation tools had also become available. “GPS came along,” Thomas later explained, “and for the first time we knew where an airplane was.”18 In other words, GPS could tell Krabill exactly where the plane was in the sky, which was crucial to measuring the ice. What’s more, it ensured that a pilot followed a precise route over the ice sheet—not just during a single year of research, but over several years. This was vital as well. To understand how the ice sheet was changing, you couldn’t measure its elevation once. You had to make surveys of exactly the same routes over Greenland a few years apart and then compare them. If the height of the ice dropped in a place that was measured repeatedly, it would suggest the ice cap was thinning. And that in turn could mean Greenland’s ice was shrinking rather than growing.

The mission became known as PARCA, which stood for Program for Arctic Regional Climate Assessment. And there was no certainty that Thomas and Krabill’s approach would succeed. “Bob basically bet his career on this,” John Sonntag, a NASA engineer who was part of the team flying routes over Greenland, recalls.19 Hundreds of hours were logged over Greenland in 1993 and 1994, and again in 1998 and 1999—long and sometimes monotonous flights over the most desolate stretches of icescape and rock. Krabill and his laser altimeter surveyed the surface below and collected data. The flight tracks of the PARCA study, superimposed over the ice sheet by the flight planners, began to make Greenland look like a map of crisscrossing interstate highways.

In 1999, the PARCA group, led by Krabill, published an article in Science that noted “Rapid Thinning of Parts of the Southern Greenland Ice Sheet.” A year later, in the same journal, they reported that their measurements of the ice sheet over a six-year period showed it was losing about 51 cubic kilometers of ice per year.20 This was akin to an ice cube, 2.3 miles long on each side, falling into the ocean every year. “That was the first time a mass balance assessment of a large ice sheet had ever been done,” Sonntag would later explain. For those who cared about the earth’s cryosphere, it was a significant scientific breakthrough. “What that taught us was the ice sheet was not in balance, even then, in the 1990s,” Sonntag says. “But it also showed us that the ice sheet was growing a little bit in the center but it was thinning a lot around the edges. And the thickening in the center did not offset the thinning on the edges. In other words it was, in the net, delivering sea level rise.”21

As Thomas and Krabill were working on measuring the mass balance of Greenland by plane, a number of scientists were contemplating a bigger step. By the 1990s, research projects in the Arctic were classified by American science agencies into three different categories. The first involved traveling to a site to do fieldwork. This might entail measuring snow accumulations by digging snow pits, for instance, or using stakes to track the movement of individual glaciers. Regardless of what tools technologists could invent, that kind of work remained crucial, since it could lead to deep scientific insights about the ice sheet. That was precisely what deep drilling had accomplished and—even further back—what the researchers at Eismitte and Station Centrale had achieved. The biggest shortcoming was that field experiments couldn’t cover large areas. Also, they usually could not be sustained over long periods of time.

But there was now a second area of research—using an airplane to measure the ice on Greenland. This approach could not collect the same depth of detail as a team in the field. On the other hand, there were no supplies to drag over the ice, no snowmobiles, no tents, no stoves, no drills, blizzards, or crevasses. Most crucially, an aircraft could range effortlessly over a tremendous area and take one flight after another, season after season. While the expenses could be significant, so were the potential breakthroughs.

Satellites promised to take Arctic research to a third level. Circling the earth from three hundred or five hundred miles up, a remote sensor would likely provide scientists with even less detail than aircraft surveys. But satellites would cover a far greater area and have an observational frequency—orbiting the earth fifteen times a day for years—that meant a constant stream of data. If a glacier at the end of the earth melted, or if some big piece of ice broke off the wildest and most desolate nub of north Greenland, someone—something—would notice.

Bob Thomas’s boss at NASA was a glaciologist named Jay Zwally. From the mid-1980s on, Zwally had lobbied within the agency for building a satellite that could measure the polar regions through laser sensing. After making the case repeatedly for the importance of a constant stream of data, he recalls that “the response would be, ‘Ice sheets are a long-term climate problem. You only need to measure them every ten years or so.’ ”22

But in the mid-1990s, concerns about the possible impacts of global warming intensified. As the data from Bob Thomas’s Greenland flights filtered in, news about abrupt climate changes in the past, gleaned from the deep cores drilled in Greenland, was spreading around academic circles. Zwally began making progress with his idea. In 1999, NASA committed about $280 million to a satellite ice project. Led by Zwally, the program was christened ICESat, which stood for Ice, Cloud, and land Elevation Satellite. The main goal was to use a laser altimeter to measure elevation changes on the ice sheets of Greenland and Antarctica. The satellite would keep constant watch on the ice caps for five years—and it would never blink. At the time, Zwally noted that the technical challenge of the mission was to launch a one-ton satellite 365 miles above the earth, which performed a laser measurement “40 times a second [while] moving 16,000 miles per hour.”23 But NASA now viewed that kind of technology as perfectly doable. Following several years of planning and assembly—along with a few delays—the mission was sent into orbit with a Delta-II Rocket launched from Vandenberg Air Force Base in California in January 2003.

It made sense that Thomas and Zwally were pushing ahead with remote sensing. They were glaciologists. Both had spent time in the field (Zwally had been involved in the Greenland drilling during the 1980s), and they knew from cold, hard experience the difficulties of working on ice sheets and the advantages of surveying them from high above. They understood a warming world would put the ice in a precarious position. What was more difficult to understand, though, was the project that a young NASA scientist on the West Coast was mapping out at around the same time. Mike Watkins hadn’t ever been to Greenland or Antarctica. He had just earned a PhD in engineering from the University of Texas and come to work at the Jet Propulsion Lab in Pasadena, California. Watkins wasn’t thinking much about ice, or about lasers, in the early 1990s when he first started working through an idea for a pair of satellites that would circle the earth together, one following the other, going around and around. Mainly, he was thinking about gravity.

Ever since the late 1960s, scientists at NASA had aspired to create a remote-sensing satellite that could take a good measurement of the planet’s gravity field.24 But other projects had taken priority—sensors for the atmosphere, ocean, and soil—and decade after decade, a gravity satellite project kept getting pushed aside. Still, the idea of measuring earth’s gravity field with extreme precision remained a point of discussion, in part because such a measurement could have practical as well as scientific value. The planet’s gravity is not consistent everywhere. The pull on an object moving overhead can be greater wherever the mass is denser—for instance, above mountain ranges like the Rockies or Alps, or over vast ice sheets like Greenland’s. These gravitational variations can have subtle but important effects. They can influence the paths of satellites and ballistic missiles, for instance, which is something the U.S. military cares deeply about. They can also affect the oceans, since sea levels can be distorted in some locations by the gravitational effects of what lies far beneath the water’s surface. Down below, there are deep trenches, submerged mountain ranges, and the remnants of lost continents that slid under the seafloor hundreds of millions of years ago. If you could gather an improved measurement of the planet’s mean gravity field, the data could prove useful in fields ranging from aeronautics to oceanography.

Watkins’s proposal for the gravity project was called GRACE, for Gravity Recovery and Climate Experiment; it was greenlighted by NASA administrators in 1997, with Watkins appointed as the project scientist, and his former mentor, Byron Tapley, as the research leader. To those who weren’t familiar with the project, it was not immediately obvious what a gravity satellite could discover about climate, let alone about ice. There were no cameras or radars directed toward earth on these machines. But Watkins was optimistic about what GRACE might do. He understood that earth’s gravity fields vary not only from place to place but from season to season—and almost always because of the movement of water from one region to another. He believed that “time variability,” as he thought of it, might yield some intriguing results. “I started to do simulations,” he recalls, “and I’d ask, ‘How much can we measure changes in the Greenland and Antarctic ice sheet? How well can we measure aquifer changes in groundwater?’ And we started to realize that this was the thing that was going to break the mission wide open. And so we wrote the proposal, we said, we’ll get the mean gravity field of earth, that’s a slam dunk. But here’s this other cool thing we can do.”25

The trick was in how the satellites would measure the planet’s changes. GRACE’s two modules would orbit earth about every ninety minutes, chasing each other around the sky in measured pursuit, about 135 miles apart, one following the other like cat and mouse. They were designed to be sleek: Clad in solar panels, they had a resemblance to oversized gold bricks, each about the size of a small sports car. “We wanted a very compact design,” Watkins recalls, “so that nothing moves, nothing changes, the center of gravity doesn’t move. We never turn anything on or off on the spacecraft. They’re just sailing along, sailing just as quietly as they can, so they’re mostly only affected by gravity.” But the gravity from below could vary. That was the key. Where there’s more water mass in one place on earth, there is more gravity; where there’s less water mass, there is less gravity. That meant that when the first GRACE spacecraft approached the space above the Greenland ice sheet, with its mass of about three quadrillion tons, it would respond to the subtle gravitational pull of the ice. It would be pulled slightly forward and away from its trailing partner, and the distance between them—those 135 miles—might increase by less than a human hair. But because the two modules were in constant contact with each other through a microwave communication link, the minuscule change in distance could be precisely measured. This could be measured over and over again, month after month, year after year. If the ice on Greenland kept pouring into the ocean and its ice sheet grew smaller, the gravitational pull would therefore change, too. Scientists on the ground could conceivably convert that distance measurement into a calculation of ice loss.

As Watkins conceived of it, GRACE would constantly weigh the ice sheet, which is why one of his colleagues began calling it “the scale in the sky.”26 And in this respect, it would be different than almost any other satellite. GRACE’s movement—or more exactly, its change in movement from one month to the next, and from one year to the next—was the measurement itself. Or at least that was the plan. Watkins was sure the satellite would work in terms of its engineering and its capability to collect scientific data. “But we weren’t quite certain what we were going to see,” he says. “Is Greenland changing? Are there big changes in aquifers underground? What if those signals are really small?” He worried that if the ice sheets were close to equilibrium, meaning they were gaining as much mass from snow every winter as they were losing in summer, GRACE would not prove particularly interesting to glaciologists or climate scientists. It might turn out to be a very boring set of observations, Watkins thought.27

GRACE was authorized during a late-1990s era at NASA when some science missions were approved on the condition that they satisfy an agency directive to be “faster, better, cheaper.” The joke around NASA at the time was that the first two parts of that motto were fine; it was the last part that made life difficult. What ultimately carried GRACE to completion and a launch date was a partnership that the American scientists struck with several German science agencies to share costs. The German team agreed to pay for a rocket to launch the satellite and conduct GRACE’s mission operations while it was in orbit. NASA’s Jet Propulsion Lab meanwhile did the bulk of the satellite design and paid for the instrumentation. The two spacecraft were assembled by Airbus, just outside of Munich. Ultimately, the cost amounted to $97 million for NASA and about $30 million for Germany.

Using a Russian rocket, GRACE was sent into space without a hitch from the Plesetsk Cosmodrome, a launchpad about five hundred miles north of Moscow, on March 17, 2002. As planned, the two modules of the satellite moved into orbit and, using onboard tanks of nitrogen for acceleration and positioning, eventually separated by the requisite 135 miles. GRACE was meant to sail in a circumpolar orbit, meaning that rather than tracing the equator it would cruise over the South and North Poles. The two spacecraft soon began to transmit their measurements several times per day, usually to a ground station located in Svalbard, which lies in the Arctic Ocean north of Norway. From there, the data would be routed to the German science team, near Munich, and to Americans at the Jet Propulsion Lab in California. In those early days, as the raw data began coming back—data that no one had ever really seen before—scientists didn’t immediately gape in wonder. Mostly, they scratched their heads and tried to figure out how to make sense of it.

On a clear night, if the alignment of the earth was just right and you knew exactly where to look—and then, if you walked into a field on the dark outskirts of some town, somewhere on earth—you could gaze up at a certain moment and watch GRACE fly by, two tiny fleeting dots. “It was whoosh, whoosh, like streaks in the sky,” one engineer working on the project would later say. The satellites looked like streaks because they were moving at about 17,200 miles per hour.28

For the first few years after the mission started, scientists pored over the data transmissions and tried to estimate changes in the polar ice caps. It was not a simple matter. Isabella Velicogna, a physicist with the Jet Propulsion Lab working at the University of Colorado Boulder, built complex mathematical models to try to analyze all the factors—mountains under the ice, for instance, and the weight of the atmosphere—that might affect the gravitational forces that GRACE was sensing in its orbit. “If you want to look at the total mass of the ice sheet,” she reasoned, “you have to separate it from any other signal.”29 Velicogna focused on trying to “peel away” all the noise except for the tugging effect of ice on the spacecraft. Other teams of scientists were doing similar calculations, and while some came up with different results, all of the estimates pointed in the same direction: The ice caps were in serious decline. When Velicogna published her first results for Greenland in 2006, she calculated that the ice sheet was losing well over one hundred billion tons of ice per year, and that Antarctica seemed to be losing substantial amounts as well.30 The article floored a number of prominent polar researchers. “I remember reading their first paper, and I literally couldn’t believe it,” says Berrien Moore, a scientist who managed NASA missions on and off for several decades.31 A few years later, Velicogna updated her findings and noted that according to GRACE readings between 2002 and 2009, “the mass loss of the ice sheets [was] not a constant but accelerating with time.” Greenland had gone from losing about 137 billion tons of ice per year to losing about 286 billion tons a year.32 It wasn’t easy to get your head around that kind of number, but it was about equivalent to two Mount Everests falling into the sea every year. “It’s not a ‘Run for the hills’ kind of story,” Waleed Abdalati, a high-ranking scientist at NASA, said at the time, “but it’s ‘Wow, this is serious.’ ”33

To Abdalati, it meant that Greenland’s ice, and to some extent Antarctica’s, was pushing up sea levels—not yet by leaps and bounds, but by small and steady increments. And this seemed revelatory. For years, global sea levels had been measured two ways: by tidal gauge markers in harbors around the world, and by satellites that NASA had put up in the sky. Both types of measurements had shown that the ocean had been rising over the past few decades. Both types were in close agreement. Still, the evidence of this slow and steady rise—a few millimeters per year—couldn’t explain why the tide was creeping up.34 Oceanographers understood that in a warming world, sea levels could rise for two distinct reasons: The first was that the oceans would expand in volume as they gained heat; the second was that ice melting from Greenland and Antarctica, along with ice melting from mountain glaciers around the world, would trickle into the sea. What GRACE could do was calculate the mass of ice pouring off Greenland and Antarctica and mountain glaciers. In effect, it gave scientists a new way to solve an old problem. For the first time, they could factor in how much of the rise in sea levels was resulting from additional mass, from the melting glaciers, and how much from additional volume caused by warmer temperatures.

It was obviously bad news. And in the years following, it didn’t get any better. GRACE kept circling and sending down its results, and Isabella Velicogna’s charts on Greenland’s annual ice decrease began to resemble a steep staircase going down, with losses comprising more than a trillion tons. As observations started coming in from ICESat, the satellite measuring changes in ice sheet elevation, the picture of Greenland looked similar to what GRACE was showing: Ice was going down, down, down. But it wasn’t as though an army of researchers was yet being dispatched to examine the situation; it was still a small-scale effort at understanding a large-scale problem. Writing in the British scientific journal Nature in 2008, and quoting Ohio State University glaciologist Ian Howat, Alexandra Witze noted:

Very few researchers have Greenland as the main focus of their scientific work. Decades from now, this could turn out to be one of the most short-sighted allocations of resources that began the twenty-first century. Climate change elsewhere in the Arctic has been swifter than anticipated. The remarkable shrinkage of the sea ice is “the largest change in Earth’s surface that humans have probably ever observed,” Howat points out. Trying to get any and every handle on how that affects the poised mass of [Greenland’s] ice next door must surely be a priority, he says.35

The average summer temperatures in Greenland were now sometimes 7 to 9 degrees Fahrenheit warmer than normal. And all kinds of distressing evidence about the Arctic was filtering in from the ground and from the air.

There wasn’t a single dramatic change—a disruptive moment—between, say, 2005 and 2010. But there were trend lines now, showing the Arctic was under a steady siege from warmer weather. And then one day in July 2012, Tom Wagner, a polar scientist and administrator at NASA’s Washington, D.C., headquarters, came home from work and greeted his wife, Renee Crain, who was also a polar scientist.36 In addition to overseeing a number of the agency’s earth science programs, Wagner spent a good amount of time looking at data coming in from NASA’s satellites.

Crain said to Wagner: “Hey, what are your satellites seeing in Greenland?”

Wagner asked why.

“There’s melt going on all over the place,” she said.

Together they logged on to a computer and watched YouTube videos about rivers of meltwater coming down off the Greenland ice sheet. Floods were washing away bridges and heavy equipment in towns like Kangerlussuaq, the Inuit village formerly known as Sondrestrom, where the country’s main airport was located.

Wagner immediately got on the phone with a colleague at the Jet Propulsion Lab in California who monitored the Arctic with an orbiting satellite that uses a sensor known as a scatterometer. In glaciology, a scatterometer is useful because it can measure a thin film of meltwater atop an ice sheet. Wagner recalls, “I talked to our guy who looks at the surface of Greenland, and he said, ‘Oh, when I looked at the recent scatterometry image over Greenland, I saw all red. I just assumed the satellite was broken.’ ” But it wasn’t broken. The scatterometer showed the extent of water atop Greenland’s ice sheet because water and snow have different ways of “scattering” radio waves back to the spacecraft. Red meant water.

“Anyway,” Wagner says, “that’s when we realized the entire surface of the Greenland ice sheet had melted.”37

Glaciologists who entered their field in the 1970s or 1980s sometimes tell a similar story: They had begun their work under the assumption that changes to the ice sheets, if they happened at all, would be slow and modest. These were natural features of the earth that changed in long spans—tens of thousands of years—and not in the time it takes for a life to begin and end. At the start of their careers, some of these scientists had been warned by wizened university advisors that they were pursuing an area of study that could prove painfully dull and possibly a dead end. As GRACE and ICESat began sending back observations, however, it was already clear this was not at all true. Bob Thomas, writing in the journal Science with a younger NASA glaciologist named Eric Rignot, noted: “Perhaps the most important finding of the past 20 years has been the rapidity with which substantial changes can occur on polar ice sheets. As measurements become more precise and more widespread, it is becoming increasingly apparent that change on relatively short time scales is commonplace.” So now they knew. Glaciers could suddenly accelerate and thin. And all the while ice sheets could appreciably deteriorate. The idea—the old idea—that polar ice caps existed in some kind of “steady-state” needed to be discarded.38

That made the work more interesting and more urgent. But it left ice scientists in a somewhat bewildering place, too. They were struggling to understand a situation they had never considered before—as if they had sailed into a peripheral region, like those marked on the edges of the maps of old cartographers, who in drawing features of an unexplored place (the north of Greenland, for instance) called it terra incognita, an unknown land. In the summer of 2012, one thing glaciologists could be sure about was that colleagues who studied climates of the past had concluded that immense ice sheets once covered Europe and North America, that those ice sheets had shattered and melted, and that at various points in earth’s history the climate was warmer and sea levels were much higher.39 Yet whether Greenland and Antarctica, in their diminishment, would follow some kind of historical pattern or timeworn design was a question without a definite answer. Ice cores couldn’t explain that. There was no recorded history of how an ice sheet disintegrates. “We’ve never seen it,” Eric Rignot would later remark. “No human has ever seen that.”40

It therefore stood to reason that if remote sensing had solved the question of whether Greenland’s ice was growing or shrinking, it had also led into another problem, one that seemed even more difficult. It was the mystery of how nature might move much of that ice into the sea. Perhaps it would melt from the heat, or shatter as it softened, or be eroded by intrusions of warming salt water. Or perhaps all three forces in combination would undo it. Meanwhile, there was the worry: Would the polar glaciers reach a condition of degradation that would make it difficult to stop the changes that had been set in motion? The ice sheets in Greenland and in Antarctica were so big, and so complicated, that it was difficult to say where a point of no return might be.

It was around this moment that some glaciologists began to talk more freely about the notion of something they were calling “ice sheet collapse.” It was also at this point that some began to ask: How fast could it happen?