Photograph by Norbert Rosing, National Geographic Creative

The physics of ice predicts that sea level will rise twice as much by the end of the century as previously estimated.

There are days when even a born optimist starts to waver in his conviction. The release of a new study projecting that sea level could rise between five and six feet by 2100—when many children born today will still be alive and have been forced to move inland—made Thursday one of those days.

There have been lots of other studies, you might say. True: The last sea-level alarm (in what seems an endless series) came just a month ago. That analysis showed that in the 20th century, sea level rose faster than at any time in the past 2,800 years, and that our fossil-fuel emissions were very likely responsible.

Climate has changed naturally even within human history, that study said, and sea level has changed with it—but not as fast as we’re changing it now.

If you’ve been following climate science for a while, though, that wasn’t terribly surprising.  And if you’re an optimist, that study didn’t knock you off your stride.

In fact there was something almost reassuring about it: Extrapolating out to 2100, it projected a sea level rise of just three to four feet—in line with the most recent and reliably conservative report of the Intergovernmental Panel on Climate Change (IPCC). What’s more, the paper was co-signed by Stefan Rahmstorf, who until then had been a prominent exponent of higher sea-level-rise forecasts. So it was only slightly perverse to say that things were actually looking up.

The worst wasn’t going to happen. With three feet of sea level rise, the United States would “only” be looking at a loss of a land area the size of Massachusetts.

Or at spending many hundreds of billions of dollars to defend the coasts.

Or, just maybe, at cutting carbon emissions in time to keep things from getting so bad.

The Physics of Ice

Comes now the study published in Nature Thursday by Robert Deconto of the University of Massachusetts and David Pollard of Penn State. It’s different from other alarms, and here’s why.

Deconto and Pollard aren’t projecting the future based only on the experience of the past few millennia. They’re projecting it with a computer model of the West Antarctic Ice Sheet and of the Antarctic climate—that is, from the laws of physics.

Just a model, you might say, and translating those laws into an accurate model of an ice sheet is hard. True again: the problem has stumped scientists for decades. They’ve known that ice melts, and that if climate warms enough, the ice sheet will collapse at some point, dumping lot of water into the sea. But they’ve had trouble saying how much warmth is enough and how fast the collapse might proceed. No one has ever watched it happen.

The geologic record offers some test cases. Some 125,000 years ago, for instance, Earth was in an interglacial period, like the one we’re in now, a warm interlude between 100,000-year-long ice ages. The temperature then was about the same as it is today, a degree or two warmer at most. But the best evidence indicates sea level was at least 20 feet higher—which in itself is disconcerting, suggesting as it does that we might be poised on the brink of something big.

Where did 20 feet of water come from? The Greenland ice sheet contains more than enough, but it sits on land and can’t easily fall into the sea.

The West Antarctic Ice Sheet contains enough water to raise sea level 15 feet. And if you could strip away the ice and look at the bedrock, as scientists have done with airborne radar, you’d see how vulnerable it is: Most of the ice sits not on land but on the seabed. It’s a big dome of ice rising out of a seafloor basin, like a soufflé out of a bowl. Beyond the submarine ridges that form the sloping sides of the basin, floating ice shelves extend out to sea. They act like buttresses, propping up the ice dome and keeping it from collapsing and washing away.

Warm Air, Warm Ocean, Bad Juju

Deconto and Pollard’s model shows how the West Antarctic ice could have collapsed 125,000 years ago, generating the high seas of the last interglacial. That lends credence to the model’s forecast for our future.

Three key processes are work, the researchers write. First, as the ocean warms, it melts the floating ice shelf from below, thinning and weakening it.

Second, as the atmosphere warms, it melts the ice shelf from above, generating pools that become crevasses that help break the shelf apart. Scientists saw this happen when the Larsen B ice shelf broke up in 2002, but their ice-sheet models hadn’t fully reflected the importance of the process.

And once the floating ice shelves are gone, and the warm ocean is lapping directly against the face of the grounded ice sheet, and the ice has retreated inside the submarine ridge that forms the edge of the basin, a third process kicks in. Because the seafloor slopes down toward the center of the basin and the ice dome, further retreat exposes an ever larger ice face to the warm water. That accelerates the melting.

Soon tall cliffs of ice are towering above the impinging waves. Meltwater is percolating down into the cliffs and weakening them. But even without the impinging and the percolating, there’s only so tall an ice cliff can get before it becomes unstable.

Richard Alley of Penn State, who collaborated with Deconto and Pollard on an earlier study, has flown along what may be the tallest ice cliff on Earth today, the face of the Jakobshavn glacier on the west coast of Greenland. It is 30 stories tall, he says, and contains 10-story-tall cracks. It is retreating rapidly, by calving giant icebergs, but still there are long periods of waiting between calving events, when the glacier is just slowly thinning and getting ready to launch another floating berg.

The Thwaites glacier in West Antarctica is far more massive than the Jakobshavn glacier. According to another alarming study published last year, it has already become unmoored from the 2,000-foot high submarine ridge that holds it in place. If it begins to retreat down the long slope toward the center of the ice sheet, the cliffs it would produce would be far taller than the Jakobshavn one—and probably not stable.

“Then, rather than break-wait-wait-wait-break, it might switch to break-wait-break or just break-break-break,” Alley says.

That’s what Deconto and Pollard’s model suggests could start happening to the West Antarctic Ice Sheet by the second half of this century, if we don’t curb our carbon emissions: Just break-break-break.  And by 2100, when sea level had risen five or six feet, the breaking would have only just begun.

Time to Yell ‘Fire’?

If we burn all of our fossil fuel reserves, another study last September confirmed, we’ll melt the entire Antarctic and probably all the ice on Earth. (Check out maps showing what that would look like.)

“All of us are fully aware how wrong it is to falsely yell ‘Fire’ in a crowded theater,” Alley writes. “But we are also aware of how wrong it is to sit silently while a fire begins to spread in that theater.

“Right now, I do not believe humanity can continue with unchecked warming while confidently assuming that sea level rise will be limited to roughly three feet in a century. Instead, the recent modeling now favors the view that continuing rapid warming will cause sea level rise to be larger, and perhaps much larger, especially if we look beyond the end of this century.”

On the other hand, according to Deconto and Pollard, if we take vigorous steps to reduce our emissions—of which the steps promised by the recent Paris agreement are only the first—we could still save even the fragile West Antarctic Ice Sheet. Which means we could still save Miami. (Read about how Miami is facing up to the challenge.)

But that may just be the optimism talking.