However, it would take centuries of exposure to waters warmer than currently exist for such a situation to occur, they report today in Nature Climate Change.
If all the ice sheets of Antarctica melted, they would raise sea level by 60 meters (200 feet). Even with anthropogenic climate change, there's little chance that will happen. But a rise even a fraction of that amount would be devastating.
Scientists are beginning to understand the dynamics of the marine ice of West Antarctica, but less so that of East Antarctica, where there's enough frozen water to raise sea level 53 meters (174 feet). The new study focused on one part of East Antarctica, the bowl-like Wilkes Basin, which sits below sea level and holds enough ice to raise sea levels 3 to 4 meters (10 to 13 feet).
An ice sheet isn't static. Ice is added from the land and ice exits into the sea by melting and calving of icebergs. If these two amounts are equal, the ice sheet can be said to be in "balance," with its mass staying about the same from year to year.
Matthias Mengel and Anders Levermann of the Potsdam Institute for Climate Impact Research in Germany simulated what would happen if the ocean warmed enough to "imbalance" the Wilkes Basin ice shelf, tinkering with both the ice input and output. In particular, they tested what would happen if they increased the ice losses to the sea without any more inflow from the land.
When they did this, the researchers discovered that there is a zone of ice near the seaward end of the ice shelf that is wedged against ridges of rock. These wedges of ice acted like a plug for the entire shelf.
If that narrow ice plug were to melt, the ice shelf would become unstable and disintegrate. The ice currently held within the basin would flow into the ocean and raise global sea level by several meters.
"This is unstoppable when the plug is removed," said Levermann. "The speed [of removal] we don't know, but it's definitely a threshold."
A Distant Threat
In the simulation, the ice shelf retreated a bit from the sea at the beginning, but the situation was not irreversible. When the ice plug threshold was crossed, however, the melting continued, and the ice shelf disintegrated.
The ice plug effect "was a little twist that was surprising in our simulation," Levermann said. Such a situation is plausible, however; a similar scenario seems to be playing out at the much smaller West Antarctic Pine Island Glacier, which lost its plug when it broke free of its seafloor ridge in the 1970s and is already contributing to sea level rise.
Ice plugs in East Antarctica might also explain why massive amounts of ice were able to melt and raise the oceans to levels much higher than those seen today during the very warm late Pliocene, 4.8 to 3.5 million years ago.
There is no danger of the Wilkes Basin emptying itself of ice anytime soon, though. Mengel and Levermann's simulations involved scenarios of 400 to 800 years and waters 1 to 2.5°C warmer than today.
"They are also talking about temperatures much higher than they are now," said Ian Joughin of the University of Washington's Polar Science Center in Seattle. While it's true that in some places, such as at the small Pine Island Glacier, warmer water is moving in and melting smaller ice shelves, larger ice shelves are harder to melt.
"Big ice shelves have quite a lot of cold water under them," said Joughin. "In West Antarctica, a lot of times the winds have changed and pulled warm water in." In East Antarctica there are also warmer waters offshore. But those waters would require a few thousand years to eat away enough East Antarctic ice to endanger the ice plugs and the entire ice shelves, he said.
"This is fairly good news," Joughin said, because it means the huge reserve of ice in the Wilkes Basin won't be melting anytime soon. However, he noted, "it might be a huge problem a thousand years from now."