Can We Refreeze the Arctic? The Real Science Behind Ice Thickening

· hermez's blog

Date · Tags: climate, arctic, geoengineering, science

A UK government agency is funding an experiment in the Canadian Arctic that sounds almost too simple to be real: drill holes in the sea ice, pump seawater onto the surface, and let the brutal polar winter freeze it solid. The goal is to thicken Arctic sea ice and buy time against climate change.

The project is real. The data is encouraging. But the details reveal a story far more complicated than it first appears.

The project #

It is called Re-thickening Arctic Sea Ice, or RASI. It is funded by the Advanced Research and Invention Agency (ARIA), a UK government R&D agency established by act of Parliament in 2022 and launched in 2023 with about £800 million in taxpayer funding. ARIA operates like a civilian DARPA — high risk, high reward, transformative science. It is a publicly funded body overseen by the Department for Science, Innovation and Technology, not a private company or venture.

RASI specifically received a £9.9 million grant over 42 months. The project is led by the University of Cambridge's Centre for Climate Repair under Prof. Shaun Fitzgerald, with field work conducted by two teams: Real Ice (in Cambridge Bay, Nunavut) and Arctic Reflections (in Qikiqtarjuaq, pursuing a different "ice arch" strengthening strategy). The first field season began in January 2026.

How it works #

Arctic sea ice grows naturally from the bottom up. As seawater freezes onto the underside of the ice sheet, it thickens. But ice is also an insulator — the thicker it gets, the more it blocks the cold air from reaching the water below, slowing further growth.

The ice thickening method flips this. Instead of waiting for bottom-up growth, you pump seawater onto the top of the ice. The water freezes rapidly in the -40C air, adding a new layer from above. The technique is not new — it is the same method used to build ice roads in the Arctic. What is new is applying it for sea ice restoration.

There is a secondary benefit. The pumped seawater turns the insulating snow layer on top into slush and then ice, which allows the extreme cold to penetrate through the entire ice sheet, stimulating extra growth on the bottom as well.

The pumps are surprisingly modest. They require less power than a toaster because they only lift water from just beneath the ice to just on top. During the 2024/25 field season, Real Ice ran pumps for 1,080 hours, lifting 50,000 tonnes of seawater onto a test area roughly 450 meters on a side. The result: ice that was about 50 cm thicker than the surrounding natural ice.

The results #

A peer-reviewed paper by Blanchard-Wrigglesworth et al. (2025), published in Earth's Future, presents the first full-season field trial results from Cambridge Bay. The findings were striking:

That brightness matters as much as the thickness. Arctic sea ice reflects about 70% of incoming sunlight back to space. Open ocean reflects only about 7%. This is the albedo effect — the planetary thermostat that sea ice provides. When ice disappears, the Earth absorbs more solar energy, warming faster in a vicious feedback loop.

According to a 2024 study in Geophysical Research Letters, global sea ice has lost 13-15% of its planetary cooling effect since the early 1980s. Arctic sea ice specifically has lost 21-27% of its reflective power since 1980. The albedo forcing from Arctic sea ice loss alone has been about 25% as large globally as direct radiative forcing from increased CO₂ between 1979 and 2011.

Thicker, brighter ice means more sunlight reflected. That is the mechanism by which this project hopes to slow Arctic warming.

Sea ice and sea level rise #

There is a common misconception that melting Arctic sea ice directly causes sea levels to rise. It is a understandable confusion, but the physics say otherwise.

Arctic sea ice is already floating in the ocean, so it displaces its own weight in water. When it melts, it adds only about 2.6% additional volume beyond what was already displaced — a negligible amount. The National Snow and Ice Data Center is explicit: "Sea ice melt is not a significant contributor to sea level rise."

The real drivers of sea level rise are thermal expansion of warming ocean water and the melting of land-based ice — the Greenland ice sheet, Antarctic ice sheet, and mountain glaciers. Greenland alone accounts for about 46% of the Arctic land ice contribution to sea level rise.

That said, sea ice loss and sea level rise are correlated because they share the same cause: warming temperatures. The same heat that melts sea ice also melts land ice and expands ocean water. So while losing Arctic sea ice won't drown coastal cities directly, it is a warning sign that the same warming driving sea level rise is accelerating.

The summer problem #

Here is the catch: the ice did not survive the summer.

When the bay's ice broke apart at the end of the first melt season, the engineered patches went with it. The extra thickness never reached a second year. The melt season erased the winter's gains. This is still early-stage research measuring whether the approach is physically effective, ecologically sound, and scalable — not an operational climate intervention.

ARIA itself is explicit about this limitation. Their FAQ states: "Can re-thickening sea ice solve climate change on its own? No." They describe it as "at best only ever be a sticking plaster to delay loss of the sea ice, and not a solution for climate change. If rapid emissions cuts are not made soon, then no amount of thickening will be able to prevent the Arctic from becoming ice-free."

The scale problem #

The area of summer sea ice that has vanished since 1979 is about 3 million square kilometers — roughly the size of India. An additional 80,000 square kilometers is lost every year on average. The current test area is about 0.2 square kilometers.

Real Ice envisions a future where autonomous underwater drones — hydrogen-powered, costing about $5,000 each — could target key areas at the leading edge of ice loss. Their rough estimate is that halting the annual shrinking of Arctic sea ice would cost $10 billion. For comparison, that is less than two weeks of windfall profits for the top 100 oil companies after recent oil price spikes.

Whether this can scale to climatically meaningful levels is, in the words of University of Washington polar scientist Roger Marchand, "a difficult and open question."

The controversy #

Geoengineering is controversial. A significant number of polar scientists published a critique arguing that sea ice thickening is "unfeasible," "environmentally dangerous," and a dangerous distraction from the core climate need to cut carbon emissions rapidly.

There are real ecological questions. Pumping seawater onto ice brings salt to the surface, along with nutrients and microbes. Natural sea ice is fresh because salt is squeezed out during ice crystal formation. The effect of salting the surface on the Arctic ecosystem — on polar bears, seals, and the food web — is being studied but not yet understood. There are also concerns that flooding the winter snow could prevent polar bears and seals from digging dens, though the Cambridge Bay site is not a denning area.

Prof. Fitzgerald agrees that emission reductions are "critically necessary" but argues it is important to understand additional options given the slow pace of decarbonization. Current policies put the world on track for at least 2.8 degrees of warming above pre-industrial levels.

Bottom line #

The science is real. The method works. The ice gets thicker and brighter. But the ice does not survive the summer yet, the scale is tiny compared to the problem, the ecological impacts are unknown, and even the agency funding the research says it is not a solution — just a possible sticking plaster.

It is one of the most interesting climate experiments on Earth. It is not a rescue plan.

Sources #

last updated: