FOR a few minutes David Holland forgets about his work and screams like a kid on a roller coaster. The small helicopter he's riding in is slaloming between towering cliffs of ice - the sheer sides of gigantic icebergs that had calved off Greenland's Jakobshavn glacier. "It was like in a James Bond movie," Holland says afterwards. "It's the most exciting thing I have ever done."
Jakobshavn has doubled its speed in the past 15 years, draining increasing amounts of ice from the Greenland ice sheet into the ocean, and Holland, an oceanographer at New York University, has been trying to find out why. Scientists like him are more than a little astonished at the rate at which our planet's frozen frontiers seem to be responding to global warming. The crucial question, though, is what will happen over the next few decades and centuries.
That's because the fate of the planet's ice, from relatively small ice caps in places like the Canadian Arctic, the Andes and the Himalayas, to the immense ice sheets of Greenland and Antarctica, will largely determine the speed and extent of sea level rise. At stake are the lives and livelihoods of hundreds of millions of people, not to mention millions of square kilometres of cities and coastal land, and trillions of dollars in economic terms.
In its 2007 report, the Intergovernmental Panel on Climate Change (IPCC) forecast a sea level rise of between 19 and 59 centimetres by 2100, but this excluded "future rapid dynamical changes in ice flow". Crudely speaking, these estimates assume ice sheets are a bit like vast ice cubes sitting on a flat surface, which will stay in place as they slowly melt. But what if some ice sheets are more like ice cubes sitting on an upside-down bowl, which could suddenly slide off into the sea as conditions get slippery? "Larger rises cannot be excluded but understanding of these effects is too limited to assess their likelihood," the IPCC report stated.
Even before it was released, the report was outdated. Researchers now know far more. And while we still don't understand the dynamics of ice sheets and glaciers well enough to make precise predictions, we are narrowing down the possibilities. The good news is that some of the scarier scenarios, such as a sudden collapse of the Greenland ice sheet, now appear less likely. The bad news is that there is a growing consensus that the IPCC estimates are wildly optimistic.
The oceans are already rising. Global average sea level rose about 17 centimetres in the 20th century, and the rate of rise is increasing. The biggest uncertainty for those trying to predict future changes is how humanity will behave. Will we start to curb our emissions of greenhouse gases sometime soon, or will we continue to pump ever more into the atmosphere?
Even if all emissions stopped today, sea level would continue to rise. "The current rate of rise would continue for centuries if temperatures are constant, and that would add about 30 centimetres per century to global sea level," says Stefan Rahmstorf of the Potsdam Institute for Climate Impact Research in Germany. "If we burn all fossil fuels, we are likely to end up with many metres of sea level rise in the long run, very likely more than 10 metres in my view."
This might sound dramatic, but we know sea level has swung from 120 metres lower than today during ice ages to more than 70 metres higher during hot periods. There is no doubt at all that if the planet warms, the sea will rise. The key questions are, by how much and how soon?
To pin down the possibilities, researchers have to look at what will happen to all the different contributors to sea level under various emissions scenarios. The single biggest contributor to sea level rise over the past century has been the melting of glaciers and ice caps outside of Greenland and Antarctica, from Alaska to the Himalayas. According to one recent estimate, the continued loss of this ice will add another 10 to 20 centimetres to sea level by 2100. It cannot get much worse than this: even if all this ice melted, sea level would only rise by about 33 centimetres.
Expanding waters
The second biggest contributor has been thermal expansion of the oceans. Its future contribution is relatively simple to predict, as we know exactly how much water expands for a given increase in temperature. A study published earlier this year found that even if all emissions stopped once carbon dioxide levels hit 450 parts per million (ppm) - an unrealistically optimistic scenario - thermal expansion alone would cause sea level to rise by 20 centimetres by 2100, and by another 10 centimetres by 3000. At the other extreme, if emissions peak at 1200 ppm, thermal expansion alone would lead to a 0.5-metre rise by 2100, and another 1.4 metres by 3000 (see "How high, how soon?").
Then there are the great ice sheets of Greenland and Antarctica, which hold enough water to raise sea level by about 70 metres. Until recently, their contribution to sea level rise was negligible, and the IPCC predicted that Greenland would contribute 12 centimetres at most to sea level rise by 2100, while Antarctica would actually gain ice overall due to increased snowfall. "A lot of new results have been published since then to show that this very conservative conclusion does not hold," says Eric Rignot of the University of California, Irvine.
To study the ice sheets, Rignot and colleagues have combined satellite-based radar surveys, aircraft altimetry and gravity measurements using NASA's GRACE satellite. They found that ice loss is increasing fast. Greenland is now losing about 300 gigatonnes of ice per year, enough to raise sea level by 0.83 millimetres. Antarctica is losing about 200 gigatonnes per year, almost all of it from West Antarctica and the Antarctic Peninsula, raising levels by 0.55 millimetres. "The mass loss is increasing faster than in Greenland," Rignot says. "It'll overtake Greenland in years to come."
If this trend continues, Rignot thinks sea level rise will exceed 1 metre by 2100. So understanding why Greenland and Antarctica are already losing ice faster than predicted is crucial to improving our predictions.
The main reason for the increase is the speeding up of glaciers that drain the ice sheets into the sea. One cause is the knock-on effect of warmer air melting the surface of the ice: when the surface ice melts, the water pours down through crevasses and moulins to the base of glaciers, lubricating their descent into the sea. Fears about the impact of this phenomenon have receded somewhat, though: Antarctica is thought to be too cold for it to be a big factor, and even in Greenland it is only a summertime effect. "It's significant, but I don't think it's the primary mechanism that would be responsible for dramatic increases in sea level," says glaciologist Robert Bindschadler at the NASA Goddard Space Flight Center in Greenbelt, Maryland.
There is another way for surface melt to affect sea level, though. Meltwater fills any crevasses, widening and deepening the cracks until they reach all the way down to the base of the ice. This can have a dramatic effect on floating ice shelves. "Essentially, you are chopping up an ice shelf into a bunch of tall thin icebergs, like dominoes standing on their ends," says Bindschadler. "And they are not very stable standing that way." They fall over, and push their neighbours out to sea.
The most famous break-up in recent times - that of the Larsen B ice shelf on the Antarctic Peninsula in 2002 - likely happened this way. While the break-up of floating ice shelves does not raise sea level directly, the disintegration of Larsen B had consequences that models at the time failed to predict. With little to resist their advance, glaciers behind Larsen B immediately began to move up to eight times faster. Five smaller ice shelves in the rapidly warming Antarctic Peninsula have also broken up and many others are disintegrating.
What lies beneath
Surface melt poses little threat in West Antarctica, as it is so much colder. Here the danger comes from below. Take the ice shelf holding back the massive Pine Island glacier, which is thinning in a strange pattern. Radar scans have revealed giant "ripples" up to 100 metres deep on its underside.
Bindschadler thinks that the currents created by winter winds raise relatively warm water from a few hundred metres down in the Amundsen Sea off West Antarctica. This melts the underside of the ice shelf and gets trapped in the space it carves out, thus continuing to melt the ice from below over a few seasons. As the ice shelf thins, the Pine Island glacier behind it is speeding up, from 3 kilometres per year three years ago to over 4 kilometres per year according to the latest unpublished measurements by Ian Joughin of the University of Washington in Seattle.
What does this have to do with global warming? Climate change, aided and abetted by the loss of ozone, has strengthened the winds that circle Antarctica. This is speeding up the Antarctic circumpolar current and pushing surface waters away from the coast, causing deeper, warmer water to well up.
Along with the Thwaites glacier and some smaller ones, Pine Island glacier drains a third of the West Antarctic ice sheet. This ice sheet is particularly vulnerable to ocean heat because much of it rests on the seabed, a kilometre or more below sea level. This submarine ice will not raise sea level if it melts, but if it goes a lot of higher-level ice will end up in the ocean. The vulnerable parts contain enough ice to raise sea level 3.3 metres - less than the 5 metres that was once estimated but more than enough to have catastrophic effects.
Bindschadler has calculated that a change in ocean currents could potentially deliver up to 1019 joules of heat per year to the continental shelf off West Antarctica - and only about 109 joules per year would be required to melt the ice shelves that hold back the Pine Island and Thwaites glaciers. "The ocean has an enormous amount of heat compared to the atmosphere," he says.
Even in Greenland, where the ice sheet rests on land above sea level, ocean heat still matters. When not dodging giant icebergs, Holland has been trying to find out why Greenland's Jakobshavn glacier started moving faster in 1997, speeding up from around 6 kilometres per year to more than 9 kilometres per year by 2000 and 13 kilometres per year by 2003. The glacier continues to drain ice from the Greenland ice sheet at a higher rate than before.
The increase had been attributed to lubrication by meltwater, but Holland's team recently stumbled across data from local fishing boats, which deploy thermometers in bottom-trawling nets. One fact stood out: the temperature of the subsurface waters around West Greenland jumped in 1997, prior to the massive calving of Jakobshavn.
As the team reported last year, though, the real trigger lay in what happened in 1996. That year, the winds across the North Atlantic weakened, slowing down the warm Gulf Stream. The weakened current meandered aimlessly and hit west Greenland. "A modest change in wind gives you a big bang in terms of ice sheet dynamic response," says Holland.
Findings like these suggest that predicting sea level rise is even trickier than previously thought. If relatively small changes in winds and currents could have a big impact on ice sheets, we need extremely good models of regional climate as well as of ice sheets. At the moment we have neither - and while regional climate models are improving, ice sheet models are still too crude to make accurate predictions.
"They are coarse models that don't include mechanisms that allow glaciers to speed up," says Rignot. "And what we are seeing today is that this is not only a big missing piece, this could be the dominant piece. We can't really afford to wait 10 to 20 years to have good ice sheet models to tell people, 'Well, sea level is actually going to rise 2 metres and not 50 centimetres', because the consequences are very significant, and things will be pretty much locked in at that point."
So climate scientists are looking for other ways to predict sea level rise. Rahmstorf, for instance, is treating the Earth as one big black box. His starting point is the simple idea that the rate of sea level rise is proportional to the increase in temperature: the warmer Earth gets, the faster ice melts and the oceans expand. This held true for the last 120 years at least. "There is a very close and statistically highly significant correlation between the rate of sea level rise and the temperature increase above the pre-industrial background level," says Rahmstorf.
Extrapolating this to the future, based on IPCC emissions scenarios, suggests sea level will rise by between 0.5 and 1.4 metres - and the higher estimate is more likely because emissions have been rising faster than the IPCC's worst-case scenario. Rahmstorf's study got a mixed reception when it first appeared, but he can feel the winds of change. "I sense that now a majority of sea level experts would agree with me that the IPCC projections are much too low," he says.
Could even Rahmstorf's estimate be too low? It assumes the relation between temperature and sea level is linear, but some experts, most prominently James Hansen of NASA's Goddard Institute for Space Studies in New York, argue that because there are multiple positive feedbacks, such as the lubrication of glaciers by meltwater, higher temperatures will lead to accelerating ice loss. "Why do I think a sea level rise of metres would be a near certainty if greenhouse gas emissions keep increasing?" Hansen wrote in New Scientist (28 July 2007, p 30). "Because while the growth of great ice sheets takes millennia, the disintegration of ice sheets is a wet process that can proceed rapidly."
Hansen has made no specific prediction, however. So just how bad could it get? Tad Pfeffer of the University of Colorado in Boulder decided to work backwards from some of the worst-case scenarios: 2 metres by 2100 from Greenland, and 1.5 metres from West Antarctica, via the Pine Island and Thwaites glaciers. Just how fast would the glaciers have to be moving for the sea level to rise by these amounts? Pfeffer found that glaciers in Greenland would need to move at 70 kilometres per year, and Pine Island and Thwaites glaciers at 50 kilometres per year, from now until 2100. Since most glaciers are moving at just a few kilometres per year, to Pfeffer and many others, these numbers seem highly unrealistic.
Worst case
So what is possible? For scenarios based on conservative assumptions, such as a doubling of glacier speeds, Pfeffer found sea level will rise by around 80 centimetres by 2100, including thermal expansion. "For the high end, we took all of the values we could change and we pushed them forward to the largest numbers we imagined would be reasonable," says Pfeffer. The answer: 2 metres.
These estimates fit well with recent studies of comparable periods in the past, which have found that sea level rise averaged up to 1.6 metres per century at times. There is a huge caveat in Pfeffer's number crunching, though. "An important assumption we made is that the rest of West Antarctica stays put. And this is the part of West Antarctica that is held behind the Ross ice shelf and the Ronne ice shelf," says Pfeffer. "Those two ice shelves are very big, and very thick, and very cold. We don't see a way to get rid of those in the next century."
Holland is not so sure. He has been studying computer models of ocean currents around Antarctica, and he doesn't like what he sees. The subsurface current of warm water near the frozen continent, known as the circumpolar deep water, branches near the coast, and one branch hits Pine Island - which is probably why the ice there is thinning and speeding up. "Another branch of it comes ever so close to the Ross ice shelf," says Holland. "In some computer simulations of the future, the warm branch actually goes and hits Ross."
While it is impossible to predict exactly what will cause this, the lessons from Jakobshavn show that a small change in the wind patterns over Antarctica might be enough to shift the warm current towards and eventually underneath the Ross ice shelf. Then even this gigantic mass of ice - about the size of France - becomes vulnerable, regardless of how cold the air above it is. Pfeffer agrees that the Ross and Ronne ice shelves are the wild cards. "If we pull the plug on those two, then we create a very different world."
Is there really a danger of a collapse, which would cause a sudden jump in sea levels? Paul Blanchon's team at the National Autonomous University of Mexico in Cancun has been studying 121,000-year-old coral reefs (pictured above) in the Yucatan Peninsula, formed during the last interglacial period when sea level peaked at around 6 metres higher than today. His findings suggest that at one point the sea rose 3 metres within 50 to 100 years.
We just don't know if this could happen again in the 21st century. What is clear, though, is that even the lowest, most conservative estimates are now higher than the IPCC's highest estimate. "Most of my community is comfortable expecting at least a metre by the end of this century," says Bindschadler.
And it will not stop at a metre. "When we talk of sea level rising by 1 or 2 metres by 2100, remember that it is still going to be rising after 2100," Rignot warns.
All of which suggests we might want to start preparing. "People who are trying to downplay the significance say, 'Oh, the Earth has gone through changes much greater than this, you know, in the geological past'," says Pfeffer. "That's true, but it's completely irrelevant. We weren't there then."
What it all means
If a 1 metre rise in sea level doesn't sound like much, consider this: about 60 million people live within 1 metre of mean sea level, a number expected to grow to about 130 million by 2100.
Much of this population lives in the nine major river deltas in south and southeast Asia. Parts of countries such as Bangladesh, along with some island nations like the Maldives
, will simply be submerged.
According to a 2005 report, a 1-metre rise in sea level will affect 13 million people in five European countries and destroy property worth $600 billion, with the Netherlands the worst affected. In the UK, existing defences are insufficient to protect parts of the east and south coast, including the cities of Hull and Portsmouth.
Besides inundation, higher seas raise the risk of severe storm surges and dangerous flooding. The entire Atlantic seaboard of North America, including New York, Boston and Washington DC, and the Gulf coast will become more vulnerable to hurricanes. Today's 100-year storm floods might occur as often as every four years - in which case it will make more sense to abandon devastated regions and towns than to keep rebuilding them.
Anil Ananthaswamy is a contributing editor for New Scientist
