During the cold Arctic winter, drill holes through existing sea ice floes and use pumps (powered by gasoline in the small tests, with plans for renewables like wind or hydrogen in larger concepts) to bring up seawater from below and spread it across the surface.

The frigid air freezes the water quickly, adding a layer of ice on top. This also compacts overlying snow, which can reduce insulation and promote additional ice growth from below as heat conducts upward more efficiently.

Proponents, including startups like Real Ice and Arctic Reflections, and researchers at the University of Cambridge’s Centre for Climate Repair, frame it as “restoration” rather than full geoengineering—enhancing a natural process to make ice thicker and potentially more likely to survive summer melt.

The goal is to increase albedo (reflectivity) and slow the ice-albedo feedback loop: Less ice means more dark ocean absorbs sunlight, warming the region further and accelerating melt.

Two small tests delivered proof-of-concept results:

Svalbard lagoon (Norway): Researchers used a gasoline-powered pump delivering ~3,500 liters of seawater per minute. They flooded a limited area and saw rapid surface freezing plus extra bottom-up growth as the added layer compacted snow and improved heat conduction.

Cambridge Bay, Nunavut (Canada): Real Ice’s team (with partners) used battery-powered pumps to flood up to 250,000 m² (about 62 acres) of sea ice. They added roughly 30 cm (12 inches) of new ice thickness compared with nearby control sites—equivalent to decades of natural thinning caused by warming. Drone and satellite data even showed the engineered patches were initially ~40% brighter.

By June, the thickened ice had melted at the same rate as untouched areas. Ocean heat from below, surface slush, and ponding erased the gains. The engineered patches broke up when the surrounding bay ice fractured—exactly like the controls.

Independent glaciologists called even these tiny-scale efforts “enormous.” Scaling to cover meaningful portions of the Arctic (millions of square kilometers) would require fleets of floating, renewable-powered pumps operating through the polar night—logistically and energetically daunting. One researcher described it as “using a teaspoon to hold the ocean back.”

Arctic sea ice has been shrinking dramatically for decades (summer minimum extent down ~13% per decade). These experiments show the technique works locally in winter, but they also highlight the gap between small tests and climate impact. Broader reviews of polar geoengineering (including this approach, reflective glass beads, and underwater curtains) conclude none are currently viable at the speed or scale needed—and some carry environmental risks.

The results were published last week in the Journal of Geophysical Research: Oceans and a preprint in Earth’s Future.

_____________________________________________________________________________________

Here are the two specific studies referenced in the Science article (published around April 17, 2026):

1. Journal of Geophysical Research: Oceans (published April 2026)

Title: Field Observations of Sea Ice Thickening by Artificial Flooding

Authors: Lead author includes Hammer et al. (Svalbard team, involving Arctic Reflections and collaborators like The University Centre in Svalbard and Delft University)

DOI / Link: https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2025JC022738 (open access)

Key details from the paper:

• Conducted in Vallunden Lagoon (Van Mijenfjord), Svalbard, from March to June 2024.
• They artificially flooded a small area (~1,500 m²) with seawater pumped from below during April 11–15, 2024.
• Results: Added ~26 cm of ice thickness on top of the original ~90 cm first-year ice. The flooding heated and increased salinity in the upper ice layers. Surface albedo changed due to slush, snow drift, and melt events.
• Positive short-term effects: Thicker ice and a 6-day delay in “rotten ice” formation.
• Limitation: The extra thickness did not delay the final disappearance of the ice in summer compared to the non-flooded reference site. Ocean heat and surface processes erased the gains.

This is the smaller-scale Svalbard lagoon test mentioned in the coverage.

2. Preprint in Earth’s Future (now accepted for publication)

Title: Artificial flooding leads to thicker and brighter Arctic sea ice (or similar; full title describes the Cambridge Bay fieldwork campaign)

Link to preprint: https://essopenarchive.org/doi/full/10.22541/essoar.176556285.57463888/v1 (December 2025 version; the peer-reviewed version is accepted in Earth’s Future)

Key details:

Conducted in Cambridge Bay, Nunavut (Canada), over the 2024/2025 sea ice season (Real Ice collaboration with University of Cambridge Centre for Climate Repair and others).

Larger test: Flooded up to ~250,000 m² using battery-powered pumps in multiple sessions.

Results: Flooded areas were up to 30 cm thicker than control (unflooded) sites by early May, with slightly thinner snow cover. Sites flooded twice thickened more than those flooded once. Drone/satellite data showed the thickened ice was initially ~40% brighter (higher albedo).

During melt: Brighter appearance and initially slower melt rates, but the ice ultimately broke up when the surrounding bay ice fractured in summer—similar fate to controls.

Both studies demonstrate that artificial flooding can thicken sea ice and temporarily increase reflectivity in winter/early spring, but the added thickness did not persist through summer melt under the tested conditions.

Access notes

The JGR: Oceans paper is open access (full PDF available directly via the Wiley link above).

The preprint is freely downloadable from ESS Open Archive.

If you have institutional access or are willing to pay for the final Earth’s Future version once fully published, that will be the polished peer-reviewed form.

These are the exact publications behind the recent Science news piece. The findings align with what was summarized: promising local winter thickening (~20–30 cm, equivalent to reversing decades of thinning), but no lasting summer survival advantage in these small trials, highlighting the massive scaling challenges for any meaningful Arctic-wide impact.