Sediment cores pulled from 500 meters below Prudhoe Dome in northwest Greenland contain pollen and plant matter. That sounds mundane until you understand what it means: sunlight hit that ground roughly 7,000 years ago. The ice was gone. The GreenDrill team published this finding on April 18, 2026, and a predictable wave of coverage followed, some of it framing the discovery as evidence that Greenland has survived warm periods before and bounced back fine.
That reading is wrong, and the numbers in the study itself explain why.
What the Sediment Actually Says
Prudhoe Dome melted during the early Holocene when summer temperatures in the Arctic ran 3 to 5°C above today's levels, driven by stronger solar insolation and warmer North Atlantic circulation. Lead author Caleb Walcott-George at the University of Kentucky put it plainly: some projections show Prudhoe Dome reaching those same temperature thresholds by 2100 under moderate emissions scenarios. The past melt is not a reassurance. It is a calibration point.
The critical difference between then and now is the forcing mechanism. Orbital cycles and ocean heat drove the Holocene warming over thousands of years, giving ice systems time to respond gradually. Today's forcing is CO2 at 420 ppm, a concentration unprecedented in the ice record, accumulating over decades. The physics of ice loss does not care which forcing caused the warming, but the speed matters enormously for coastal planning. A slow Holocene retreat gave coastlines millennia to adjust. A 2100 collapse gives Miami about 74 years.
Fair point to the skeptics: the entire Greenland ice sheet did not melt 7,000 years ago. The central dome persisted. That distinction is real and the study makes it clearly. But Prudhoe Dome is a peripheral dome, and its vulnerability under mild Holocene warmth is precisely what makes the finding alarming rather than comforting. If a high-elevation region in the northwest went ice-free at 3-5°C of local warming, the question is not whether peripheral domes are vulnerable. We now know they are. The question is how many of them we are willing to sacrifice.
The Numbers That Should Drive Policy
Greenland has been losing 250 to 270 gigatons of ice per year since 2002. That is already contributing 0.7 to 0.8 mm annually to sea-level rise, 4 times the rate recorded in the 1990s. Prudhoe Dome's collapse alone adds 20 to 50 cm of sea-level rise by 2100 in moderate scenarios. If major domes follow, the long-run number reaches 7 meters. The East Coast of the United States and the river deltas of Bangladesh and the Netherlands do not have a geological timescale to adapt to that.
I hold a tension here I will name directly: the study's authors are right that sub-ice drilling gives us better local sea-level predictions, and that investment in monitoring is worth making. But monitoring is not mitigation. Knowing exactly when Prudhoe Dome will go does not stop it from going.
The 2023 melt season, comparable in scale to the record years of 2012 and 2019, showed that both dry and cloudy conditions can drive extreme melt through different mechanisms, margins in dry years, central ice in cloudy ones. The ice sheet has multiple failure modes. Betting on stability because the Holocene precedent exists is like reading a crash report and concluding the car is safe because it survived previous accidents.
The GreenDrill findings are a direct argument for staying below 1.5°C of global warming. Not because the science demands optimism, but because the sediment record has now told us exactly what 3°C costs.
