A Dartmouth-led study by more than 50 climate scientists worldwide provides the first clear projection of how carbon emissions may drive the loss of Antarctica's ice sheet over the next 300 years.

The future of Antarctica's glaciers after 2100 becomes uncertain when looking at existing ice-sheet models individually, the researchers report in the journal Earth's Future. They combined data from 16 ice-sheet models and found that collectively, the projections agree that ice loss from Antarctica will increase, but gradually, through the 21st century, even under current carbon emissions.

But that consistency falls off a cliff after 2100, the researchers found. The models predict that under current emissions, ice in most of Antarctica's western basins begins to retreat rapidly. By 2200, the melting glaciers could increase global sea levels by as much as 5.5 feet. Some of the team's numerical experiments projected a near-total collapse of the Antarctic ice sheet by 2300.

"When you talk to policymakers and stakeholders about sea-level rise, they mostly focus on what will happen up to 2100. There are very few studies beyond that," says Hélène Seroussi, the study's first author and an associate professor in Dartmouth's Thayer School of Engineering.

"Our study provides the longer-term projections that have been lacking," she says. "The results show that beyond 2100, the long-term impact for the regions most susceptible to sea-level rise becomes amplified."

The researchers modeled how Antarctica's ice sheet would fare under both high- and low-emission scenarios through 2300, says Mathieu Morlighem, a Dartmouth professor of earth sciences and a co-author on the study. Dartmouth Engineering alumnus Jake Twarog is also a co-author of the study and contributed as an undergraduate.

"While current carbon emissions have only a modest impact on model projections for this century, the difference between how high- and low-emission scenarios contribute to sea-level rise grows sharply after 2100," Morlighem says. "These results confirm that it is critical to cut carbon emissions now to protect future generations."

The timing of when Antarctica's glaciers would start retreating varied with the ice-flow model the researchers used, Seroussi says. But the speed with which large retreats occurred once a rapid loss of ice began was consistent among the models.

"All the models agree that once these large changes are initiated, nothing can stop them or slow them down. Several basins in West Antarctica could experience a complete collapse before 2200," Seroussi says. "The exact timing of such collapses remains unknown and depends on future greenhouse gas emissions, so we need to respond quickly enough to reduce emissions before the major basins in Antarctica are lost."

The study could lead to further collaborative models that scientists can use to understand and resolve disparities in projections for regions with significant modeling uncertainties, or for the Greenland ice sheet, Seroussi says. Research and computing resources can then be focused on investigating outcomes that those multiple models predict as most likely.

"We're learning from the community of scientists what is going to happen," Seroussi says. "This collaboration means we have a better, more robust assessment of the uncertainty, and we can see where our models agree and where they disagree, so that we know where to focus our future research."