Inside the Bold Geoengineering Work to Refreeze the Arctic's Disappearing Ice

Scientific American

20-05-2025

This story was produced in partnership with the Pulitzer Center's Ocean Reporting Network.
A haze of ice crystals in the air created a halo around the low sun as three snowmobiles thundered onto the sea ice on a February morning in far northern Canada. Wisps of snow blew across the white expanse. It was –26 degrees Celsius as we left Cambridge Bay, an Inuit village in a vast archipelago of treeless islands and ice-choked channels. This temperature was relatively warm—six degrees C above average. The winter had been the mildest in 75 years. The sea ice covering the Arctic Ocean was at its smallest extent on record. Scientists predict that within the next 15 years this ice cap will disappear in summer for the first time in millennia, accelerating global warming. The U.K. company Real Ice, whose heavily bundled team was bouncing around on the other two snowmobiles ahead of mine, hopes to prevent that outcome with an effort that has been called extremely ambitious, insane or even dangerous.
At a spot seven kilometers from the village, Real Ice co-founder Cían Sherwin, an Irishman with a red beanie and scraggy goatee, hopped off his snowmobile and started drilling with a long electric auger. A gob of water and frozen shavings sloshed up and out of the hole as he punctured the underside of the ice more than a meter below. Inuit guide David Kavanna widened the opening with a spearlike ice saw, then placed a wood box around it. Sherwin lowered an aluminum pump, which looked like a large coffee urn attached to a curved rubber hose, through the hole. He plugged a cable into a battery pack. After a few seconds water began pouring out of the hose, spilling onto the ice in an ethereal shade of blue. As it congeals, 'the water acts almost like lava,' Sherwin said. 'The ice formation starts almost instantly.'
On supporting science journalism
If you're enjoying this article, consider supporting our award-winning journalism by subscribing. By purchasing a subscription you are helping to ensure the future of impactful stories about the discoveries and ideas shaping our world today.
Thin, broad sheets of ice expand from the ice cap's edges in winter, when it's dark and cold, and melt away in summer, when the sun shines 24 hours a day. The ice acts like a giant mirror, reflecting up to 90 percent of the sun's radiation back toward space. Ocean water, in contrast, absorbs 90 percent of sunlight. The ice cap's core of so-called multiyear ice, which persists year-round, has shrunk by about 40 percent in four decades, kicking off a vicious cycle: as more ice melts, more ocean water is exposed, and that water warms further, melting even more ice. If the ice starts disappearing entirely in summer, global temperatures could rise an extra 0.19 degree C by 2050.
Real Ice is trying to thicken seasonal ice so it lasts longer into the warm months, keeping the planet cool. Sherwin hopes pumping could someday refreeze a million square kilometers of both seasonal and multiyear ice—an area the size of Texas and New Mexico combined and about a fifth of what's now left in summer—to stop the ice cap's death spiral. All it would take, Real Ice says, is half a million ice-making robots.
Polar geoengineering on such an enormous scale could help slow warming until the world finally weans itself off coal, oil and natural gas. Many scientists think it will never work. The researchers at Real Ice argue we no longer have any option but to try; studies suggest that even slashing fossil-fuel use may not save summertime sea ice. 'It's sad that it's ended up that way, but we've got to do something about it,' Sherwin said to me out on the frozen plain. 'Emissions reduction is just not enough anymore.'
Cambridge Bay, which British explorers named for a 19th-century Duke of Cambridge, is a town of 1,800 mostly Inuit inhabitants located across from the Canadian mainland on Victoria Island, one of the world's largest islands. When I landed at the one-room airport on a twin-engine turboprop, I was greeted by a stuffed musk ox and a placard about the 1845 British naval expedition of John Franklin. Cambridge Bay lies along the Northwest Passage, an icy sea route between Europe and Asia sought by explorers for 400 years. Franklin's two ships, Erebus and Terror, were trapped in the polar sea ice that surges down toward Cambridge Bay in winter, buckling into ridges up to 10 meters high. All 129 men onboard died of cold, starvation or disease. These days cruise ships coast through the passage every year, often visiting grave sites of Franklin expedition members.
The Inuit call Cambridge Bay Ikaluktutiak, meaning 'good fishing place.' For millennia their nomadic ancestors came here to fish Arctic char, a silvery-orange cousin of the brook trout. Inuit started living here full-time in the 1940s and 1950s, when the U.S. military hired them to help build a navigation tower and a radar station to detect Soviet bombers coming over the pole. The cold war also led to the idea of controlling the Arctic environment. The U.S.S.R. discussed destroying sea ice with coal dust or explosions and detonated three nuclear devices to try to excavate an Arctic canal. In the U.S., physicist Edward Teller's Project Plowshare nearly got approval to gouge out a harbor in Alaska with atomic bombs.
Geoengineering today is intended to cool Earth to fend off climate change. Some scientists and entrepreneurs are focused on dispersing sulfate particles in the stratosphere to block sunlight, which could lessen heating but also disrupt global weather patterns such as the South Asian monsoon. Mexico recently announced a ban on this solar geoengineering after Silicon Valley start-up Make Sunsets launched two balloons full of sulfur dioxide there. The city of Alameda, Calif., halted an experiment to spray sea-salt particles skyward to make clouds more reflective. Field trials targeting the Arctic, the Antarctic and the 'third pole' of colossal glaciers in the Himalayas have stirred up less controversy, perhaps because unintended consequences would be confined largely to those distant places. In Iceland and India, Silicon Valley nonprofit Bright Ice Initiative has scattered tiny glass beads on glaciers to try to reflect more sunlight and slow the melting. Chinese agencies have blown chemical smoke into clouds with rockets, planes, drones and chimneys to provoke snowfall over glaciers on the Tibetan plateau. Researchers in Scandinavia are developing giant curtains that could be anchored to the seabed to block warm ocean water from melting the undersides of ice shelves in Antarctica. Billions of dollars would be needed to affect the climate.
The idea for thickening ice came from outer space. At a 2012 conference a fractious forum about global warming soured Arizona State University astrophysicist Steve Desch's hopes for quick climate action. Desch, who studies icy bodies such as Pluto's moon Charon, wondered whether we could buy time by making ice in the Arctic. The problem is that sea ice freezes from below. Once the first layer forms, it insulates the seawater from the air, which can be 50 degrees C colder. The thicker the ice gets, the slower it grows. In 2016 Desch published a paper proposing that wind-powered pumps could thicken sea ice by pulling up water from below and spraying it across the top.
Around that time, students at Bangor University in Wales were inspired by a documentary on the Arctic to construct a 're-icing machine,' an ungainly spindle of hoses that twirled like a lawn sprinkler. One of those students was Sherwin. Encouraged by Desch's paper, he and London entrepreneur Simon Woods founded Real Ice in 2022 to see whether sea-ice thickening could scale up. They eventually recruited Desch and several sea-ice scientists as advisers. The company put its first water onto ice in Nome, Alaska, in January 2023, ditching the sprinkler for a commercial pump. They moved to the Canadian High Arctic Research Station in Cambridge Bay the next year to do more. 'It's not exactly the same as a natural process, but it's as close as you can get,' Desch says.
After the team drilled the first hole that February morning and started the pump, we snowmobiled to a destination pinpointed by GPS several hundred meters away. Again the group drilled and inserted a pump, and water began whooshing out. In all, we installed four pumps in four places. As water pooled, the edges crept outward, soaking into the pockmarked snow, which was up to 25 centimeters (10 inches) thick and crusty like hardened white frosting. Within hours the pool would coagulate into electric-blue slush, like a gas station Slurpee.
After a lunch of fruit bars and potato chips around a tiny gas heater in the team's rescue tent, we went to a site the crew had pumped a day earlier. Under a dusting of snow lay flat gray ice. With a drill bit almost as tall as himself, one of the volunteers bored a hole and dropped in a measuring tape with fold-out brass arms at its end. The ice was 152 centimeters thick; almost 30 centimeters of ice had been added, compared with untouched sites they measured.
The ice would thicken further in coming weeks. Because snow is a better insulator than ice—this quality is why igloos work so well—flooding and freezing the snow could allow more cold to penetrate to the ice's underside, creating more ice. After Real Ice thickened 4,100 square meters of ice here in winter 2023–2024, the crew came back in May 2024 to find a significant increase. Across the area they had pumped, ice thickness was 1.9 meters, compared with 1.44 meters in other places. 'Ice growth from below—that's the really efficient part,' Woods told me as he drilled a measuring hole at another refrozen site.
But snow is also a better mirror than ice, which could complicate the picture. Sea ice covered by snow reflects 90 percent of solar radiation, whereas bare sea ice reflects 50 to 70 percent. Real Ice would need snow to accumulate in spring to replenish the snow it flooded in winter, or the process could increase melting.
That's just one way flooding snow could backfire. As seawater freezes, the salt in it is ejected from the ice crystals, leading ever saltier pockets of brine to form on the surface. Salt lowers the freezing point of ice, whether on winter roads or the sea. If pumping seawater leaves more salt on the surface in summertime, it could end up accelerating the disappearance of the ice.
So far this doesn't seem to be happening. On another morning Woods put a hollow red barrel on the drill and bored into the ice at a refrozen site to extract an ice core about as long and thick as his arm. He held it up to the pale sun, which illuminated hairline channels where the salty fluid had eaten its way through the ice. 'This natural process helps the brine to migrate back into the ocean,' he said.
It's still not clear how ice thickening will affect sea life, starting with the microscopic algae that grow on the underside of the ice. They're eaten by zooplankton, which are eaten by fish, which are eaten by mammals. On a different morning I snowmobiled with University of Alaska Fairbanks marine biologist Brendan Kelly to a low ridge formed by two enormous plates of ice pushing together. A polar science adviser in President Barack Obama's administration, Kelly has studied seals and polar bears for more than four decades. In that time he's also watched fossil-fuel emissions march steadily upward. So despite his discomfort with geoengineering, he agreed to advise Real Ice.
In the hazy light, the monochrome landscape seemed devoid of life. But as we crunched on foot along the snow caking the ridge, Kelly stopped to point out an Arctic fox footprint. Farther on he found a urine stain, then desiccated green scat, then a small pit. 'Seal, probably,' Kelly said. In spring, ringed seals claw holes through snowdrifts. They hide their fuzzy white pups in these lairs while they dive for fish and crustaceans. Foxes and polar bears dig around to try to find the pups. Kelly scooped at the firm snow, tiny icicles swaying from his white mustache, but couldn't find a lair.
Polar bears also depend on snow. They excavate dens in larger drifts to warm their cubs, which are born the size of a guinea pig. Most Arctic snow tends to fall in late autumn. It's unknown whether enough new snow would build up after wintertime ice thickening for bears and seals to make dens in spring. Of course, polar bears and seals are already expected to decline as their sea-ice habitat melts away. Is Real Ice doing more harm than good by pumping seawater into this environment, melting the snow? 'We don't know that,' Kelly said. 'But we need to know it.'
Across two months last winter, Real Ice pumped water through almost 200 holes. Drills and snowmobiles broke, team members got frostnip, and Arctic foxes chewed through long, thin thermistor cables used to measure temperature in the snow and ice. The researchers thickened 250,000 square meters of sea ice. The ice cap is losing 300,000 times that area every year.
The key to scaling up is to 'bring the engineering underwater,' Sherwin told me later. The Sant'Anna School of Advanced Studies in Pisa, Italy, is developing an underwater drone two meters long that will bore through ice from below with a heated pipe and start pumping water up through it. In renderings, it looks like a folding pocketknife with a pipe instead of a blade. Real Ice hopes to test a prototype this year, says co-CEO Andrea Ceccolini, an Italian computer scientist and investor who joined the company in 2022.
The plan is to thicken 100 square kilometers of sea ice in winter 2027–2028 to demonstrate the technique to governments and investors. The approach verges on the fantastic. A swarm of 50 drones would melt holes in minutes and pump water as their infrared cameras monitored the progress. Technicians on a floating or onshore hub would swap out the drones' batteries, plugging the old ones into chargers powered by wind turbines or by green hydrogen or ammonia brought in by ship. Tapping into electricity from Canada's Nunavut region would contribute to climate change because most of it is generated from diesel fuel.
The ultimate goal of thickening one million square kilometers of sea ice would take an estimated 500,000 drones, which would consume two terawatt-hours of electricity and require 20,000 people to service them, according to rough math Ceccolini has done. The cost would be $10 billion annually. The drones would vastly exceed the 3,800 Argo robot sensors circulating in oceans worldwide, and drone experts say a revolution in battery technology would be needed.
How much global warming could be countered through sea-ice preservation depends on numerous variables affecting sunlight and melt dynamics. Preserving a million square kilometers of sea ice for one additional summer month would cool Earth as much as removing 930 million metric tons of carbon dioxide from the atmosphere over 20 years, Real Ice estimates. For these results, $10 billion is actually cheap, Ceccolini says, and the cooling would be immediate. Capturing that much CO 2 from the air with existing machines would currently cost at least $465 billion. For perspective, humanity emits 910 million tons of CO 2 every eight days, with no end in sight. Thickening sea ice is a Band-Aid 'while you cure the patient—the planet—properly,' Ceccolini says.
Every day in Cambridge Bay, after three or four hours of flooding, the team used an ice ax to hack each pump out of its hole. Fat white frost flowers formed on top of the solidifying surface. The speed of the freeze was striking—but so was the immensity of the frozen plain stretching to the horizon. It was hard to envision hundreds of thousands of drones popping through the ice day after day, all winter long, for decades.
The only highway in Cambridge Bay is the sea ice. In winter and spring, Inuit residents snowmobile hundreds of kilometers over it to go ice fishing and hunt musk ox and seals, local hunter Brandon Langan told me in his living room, a black musk ox hide hanging behind him. He works part-time for Real Ice flying airborne drones to monitor the thickened ice's reflectivity. When the ice recedes in summer, the Inuit fish the Arctic char that run into the bay from lakes and streams. When ice returns in fall, they hunt the caribou that cross it to the mainland. Two out of every three meals are fish or game. 'Sea ice, to us, it's life,' Langan said. 'It helps us get our food. It gives us our clothing.'
Now hunters who used to start moving on the ice in October have to delay until December. A few have even fallen through. In spring, the ice cracks and melts sooner. Ice loss has diminished the local caribou herd by 90 percent; the animals go hungry waiting to cross, and when bottled up for too long on the shore, they're easier prey for wolves. Hundreds of caribou have drowned after breaking through the ice. One hunter who had previously been a guide for Real Ice told me at the cultural center in the high school library that he hopes ice thickening could rejuvenate game populations.
Inuit insights are vital. Throughout history scientists and explorers often ignored Indigenous knowledge of the Arctic. The last time anyone saw the Franklin expedition was when Inuit hunters encountered starving sailors dragging a lifeboat across the ice in Washington Bay, clad in wool rather than furs. Franklin was dead, and the remaining explorers traded the Inuit beads for seal meat—they apparently didn't know how to hunt seal themselves. Later, native people found mutilated bodies farther south, indicating the explorers had resorted to cannibalism. Charles Dickens dismissed these reports, which were later confirmed, as 'the wild tales of a herd of savages' and suggested the Inuit had slain Franklin's men. This colonial mindset would persist as the Arctic came under government control. Canada and Alaska took Indigenous children, including some from Cambridge Bay, away to be reeducated in abusive residential schools, where thousands died.
A week after I met with Langan I talked with Inuit Circumpolar Council international chairwoman Sara Olsvig, who has spoken out against testing the seabed-curtain idea in her native Greenland. She says governments need to start regulating geoengineering, and researchers need to seek the free, prior and informed consent of local communities. When somebody claims, ''We need your piece of land in the name of a greater good,' that's exactly what happened when we were colonized,' Olsvig says.
Real Ice obtained permits from Nunavut's Inuit government, as well as the Cambridge Bay hunters-and-trappers organization. Ceccolini says the operation will shut down if ice thickening proves ineffective or damaging. Although the company may patent its drone technology, its articles of association prohibit it from distributing profits. But it would consider founding a new company with Indigenous part-ownership if it decided to scale up, Ceccolini says.
Local elders are hesitant. They gather at the cultural center on weekdays to sew fur boots and mittens and speak the local language. During a break for black tea and candied Arctic char, I asked three of them about the sea-ice thickening. They would be concerned about the drones if Real Ice conducted its 100-square-kilometer demonstration, a key step to scaling up, in the strait near Cambridge Bay rather than farther north. 'If they start doing that under the water, we're going to get no more fish, no more seal, no more nothing,' said Annie Atighioyak, who was born in an igloo on the sea ice in 1940.
As global fossil-fuel emissions keep rising, attitudes about geoengineering are starting to change. Two weeks before I arrived in Cambridge Bay I went to an annual Arctic science conference at the Oceanographic Institute of Monaco. Frederik Paulsen, a Swedish pharmaceutical billionaire in a tailored suit and rimless glasses, took the podium. Though not a scientist, Paulsen was the first person to reach all eight poles—the geographic, geomagnetic, magnetic and least accessible poles of each hemisphere—and was onboard one of the two submersibles that planted a Russian flag on the North Pole seabed in 2007.
In 2023, while flying over Greenland in an ultralight aircraft, Paulsen was startled to notice that the once brilliant ice sheet was turning darker as less fresh snow fell. He said he decided only 'more drastic solutions' could avoid catastrophic climate impacts, given our failure to rein in fossil fuels. It's not enough to just study climate, he scolded the scientists. 'Now is the time to act.'
The University of the Arctic, a network of educational institutions that Paulsen chairs, has rated the feasibility of 61 polar interventions, from spraying glaciers with ski-resort snow guns to cables that stop icebergs from drifting south. At the conference, John Moore, a University of Lapland glaciologist, presented the seabed-curtain idea. Also there was Fonger Ypma of Arctic Reflections, a Dutch sea-ice-thickening company that has done field trials in Newfoundland and Svalbard. Last year he went to Cambridge Bay to learn from Real Ice, but he hopes to deploy large movable pumping platforms rather than drones.
The surge of interest has created a schism in polar science. In October 2024 a preprint paper by 42 top glaciologists condemned ice thickening and other polar geoengineering techniques as dangerous and unfeasible. Seabed curtains could disrupt nutrient flows to CO 2 -consuming phytoplankton. Spreading tiny glass spheres across mountain glaciers could decrease the reflectivity of ice covered in fresh snow. But the overarching concern is that geoengineering fixes are 'making decarbonization a lot less attractive,' according to Heidi Sevestre, one of the paper's authors, who visited an Arctic Reflections trial last year. What they are offering 'doesn't attack the cause of the issue, the fossil fuels,' she says.
At the Monaco conference, Kim Holmén, a Norwegian Polar Institute climate scientist who has spent more than three decades on Svalbard, the fastest-warming place on Earth, argued that trying to undo the harms of our technology with even more technology was folly. 'I make mistakes every day, you make mistakes every day, and if we create a system where it must work every day, it will fail,' he said. Critics say it would be more effective to put geoengineering funds toward cutting emissions.
The amount being spent on Arctic geoengineering is small but growing. Arctic Reflections has raised $1.1 million and Moore $2 million. Real Ice's directors have committed $5 million to its ice-thickening project, only a fraction of the cost for a 100-square-kilometer demonstration. It's hard to imagine state agencies allocating $10 billion a year for sea-ice thickening, especially with China, Russia and reportedly the U.S. looking to develop Arctic shipping routes; for them, less ice is better. Brazil's Amazon Fund for rainforest conservation, which Real Ice has held up as a possible funding model, has collected only $780 million from governments.
Private investors might mean fewer political obstacles. In March, Paulsen, who's offering a €100,000 prize for projects 'reversing Arctic change,' convened a dinner in Geneva to introduce geoengineering researchers—including a Real Ice adviser—to two dozen potential donors. He wants to hold similar 'adopt-a-billionaire' gatherings in the U.S. He also claims he's discussed geoengineering with officials from the Trump administration, which withdrew the U.S. from the Paris climate agreement. The administration didn't respond to a request for comment.
Real Ice and Arctic Reflections would like to eventually sell 'cooling credits,' a strategy used by Make Sunsets. Individuals and companies that want to compensate for their CO 2 emissions pay that start-up to launch balloons full of sunlight-blocking sulfur dioxide. Carbon credits have paid for the planting of trees to remove tens of millions of tons of CO 2 from the atmosphere. But the credit arrangements also have been criticized as a 'license to pollute' to avoid fossil-fuel reductions. Some of the biggest buyers are tech firms such as Microsoft, whose emissions are swelling as the company builds AI data centers.
Kelly, the former White House adviser, is indeed worried that geoengineering could be 'hijacked' by oil or tech companies as an excuse to continue business as usual. But he's more concerned about the gigantic geoengineering experiment he says we're already conducting by emitting tens of billions of tons of greenhouse gases every year. Ice companies just have to be willing to shut down their technology if it starts harming nature or undermining climate goals, he told me as we drank coffee at the research station in Cambridge Bay. They have to be willing to turn back, unlike Franklin and other overconfident explorers who came here to discover the Northwest Passage. 'We all have to keep asking ourselves and checking ourselves and one another: Have we slid into hubris?' Kelly said. 'The alternative is to think [we humans] know what's best, that we can get through the passage.'