Six Degrees: Our Future on a Hotter Planet. Mark Lynas

unabated) most of the ice sheet will disappear over the next 1,000 years, still giving humanity plenty of time to prepare for the full 7 metres of inundation-even though lower-lying areas would go under much sooner.

      In addition, some of the melting will be offset by increased snowfall, leading to thicker ice in the centre. This is another result of rising temperatures, as a warmer atmosphere can hold more water vapour. Much of Antarctica and the interior of Greenland is classed as ‘polar desert’ because it is simply too cold for snow to fall in significant amounts. Already evidence suggests that areas of the Greenland ice sheet above the 1,500-metre contour line are accumulating snow and new ice (at 6 cm a year according to one study). It has even been suggested that a thicker Greenland ice sheet could offset rising sea levels.

      But real-world evidence runs counter to these optimistic scenarios, suggesting that James Hansen may be right after all. The models on which the predictions of Greenland ice melt are based operate by estimating the difference between water loss from future melting and ice accumulation from future snowfall. There is much more to ice sheet dynamics than just melting and snowfall, however. Vast quantities of ice are constantly flowing away from Greenland's centre in gigantic glaciers, which surge through fjords and discharge icebergs into the sea. These glaciers can have a rapid effect on the stability of the ice sheet, yet they aren't properly accounted for in the models. ‘Current models treat the ice sheet like it's just an ice cube sitting up there melting, and we're finding out it's not that simple,’ says Ian Howat, an expert on Greenland's glaciers.

      In particular, as melting proceeds on the surface, whole rivers plunge down through icy sink-holes called moulins onto the bedrock beneath the ice sheet. This meltwater then acts as a lubricant underneath the ice, speeding up the glaciers as they proceed towards the sea. As one glaciologist told Nature: Along the coasts, all the glaciers are thinning like mad, and they're also flowing faster than they ought to. Changes initiated in coastal regions will propagate inland very quickly.' When Byron Parizek and Richard Alley, two glaciologists at Penn State University in the US, had a first stab at including meltwater lubrication in an ice sheet model, they found that this did indeed lead to a thinner ice cap over Greenland, and a greater contribution to sea level rise.

      Greenland's glaciers are also changing much quicker than anyone expected. The largest outflow glacier on the whole landmass, Jacobshavn Isbrae in the south-west, is so huge that it alone has a measurable impact on global sea levels, accounting for 4 per cent of the twentieth-century rise. Not only has the gigantic river of ice thinned by a phenomenal 15 metres annually since 1997 (that's about four office block storeys every year), but its flow rate more than doubled between 1997 and 2003, suggesting that an increased quantity of Greenland's ice is being sucked out into the sea. As if to emphasise the abnormal shift, Jacobshavn Isbrae's floating ice shelf has now suffered almost complete disintegration, spawning an armada of icebergs along the coast.

      On the eastern side of the ice cap, a second glacier has also suffered dramatic changes. A US-based research team led by Ian Howat studied satellite photos of the Helheim Glacier's behaviour between 2000 and 2005, and was stunned to discover that not only had the ice flow speeded up, but the glacier had thinned by over 40 metres and retreated several kilometres up its fjord. About half of the thinning is due to increased surface melt: in recent years ever-wider areas of Greenland have been rising above freezing, and thousands of blue meltwater pools now pepper the ice surface in the summer. But the remainder is due to the speeding-up of glacial flow rates, to the distinctly unglacial pace of 11 kilometres per year.

      This faster flow draws more ice down the valley, thinning the glacier just as a rubber band gets thinner when you stretch it. In the process, now being repeated in glaciers right around the ice cap, billions of tonnes more ice are dumped in the North Atlantic, raising sea levels still further. According to Howat, the thinning has reached a ‘critical point’ which has begun ‘drastically changing the glacier's dynamics’. His conclusion is devastating: ‘If other glaciers in Greenland are responding like Helheim, it could easily cut in half the time it will take to destroy the Greenland ice sheet.’

      Other scientists concur. Speaking at the Fall Meeting of the American Geophysical Union in December 2005, the University of Maine's Dr Gordon Hamilton also noted ‘very dramatic changes’ on eastern Greenland's Kangerdlugssuaq Glacier. In just one year, between April 2004 and April 2005, this enormous glacier both doubled in speed and simultaneously retreated by 4 kilometres. If other large glaciers begin to go the same way as Helheim and Kangerdlugssuaq-which have now doubled the rate at which they together dump ice in the ocean from 50 to 100 cubic kilometres a year-it could ‘pull the plug’ on Greenland, Dr Hamilton warned.

      The newest evidence suggests that all is not yet lost however. In March 2007 Ian Howat and colleagues reported in Science magazine that results from their latest survey work might be more reassuring. Although both the Helheim and Kangerdlugssuaq glaciers had indeed doubled their rate of mass loss in 2004, as previously reported, two years later-by 2006-they had returned to something more like normality. However, the glaciologists Martin Truffer and Mark Fahnestock, writing in Science magazine in March 2007, are at pains to point out that this most recent change ‘does not mean that [the glaciers] have stabilised’. Instead, ‘the question remains whether changes in the past five years have left the system as a whole more vulnerable’. Clearly these giant rivers of ice are complex beasts that scientists are still struggling to understand.

      But whatever the behaviour of individual outlet glaciers, widerscale satellite studies of Greenland's entire ice cap do suggest that major changes are afoot. Scientists working on the GRACE satellite programme (Gravity Recovery and Climate Experiment) reported in November 2006 that big losses are now under way. Whilst the ice sheet was probably in balance for much of the 1990s, between 2003 and 2005 it shed about 100 billion tonnes a year of ice, enough to raise global sea levels by 0.3 mm a year.

      Jim Hansen, for one, is continuing his battle to get the world to wake up to the threat of melting ice caps, and attempts by the Bush administration and Hansen's bosses in NASA to silence him have been given characteristically short shrift. (When NASA public relations staff ordered him to stop giving lectures or talking to journalists without clearing material with them first, Hansen went straight to the press with the story, sparking embarrassing ‘NASA censorship’ headlines across the US media.) His publications-even in weighty scientific journals-are now increasingly peppered with strong words such as ‘dangerous’ and ‘cataclysm’, ignoring the usual convention that scientists must muzzle themselves with emotionless jargon. One Hansen paper published (with five leading co-authors) in May 2007 warns bluntly in the abstract that ‘recent greenhouse gas emissions place the Earth perilously close to dramatic climate change that could run out of our control, with great dangers for humans and other creatures’-a clear statement of fact, but one which must have raised disapproving eyebrows throughout the halls of academe.

      However, Hansen's contention that the world's ice sheets could collapse much more rapidly than the IPCC suggests does have a solid base in physics. In order to explain the rapid real-world fragmentation of ice sheets at the end of the last ice age, Hansen outlines a process called the ‘albedo-flip’-something which, if repeated today, could destroy the remaining ice sheets much faster than conventional projections suggest. This albedo-flip is worryingly simple: as snow and ice melt they become wet, making the surface darker and therefore more able to absorb sunlight. This raises temperatures further, sparking wider melting, in a classic positive feedback. The albedo-flip is why, Hansen suggests, ice sheet disintegration can be ‘explosively rapid’ rather than a more stately process taking millennia to play out. And given that large areas of Greenland and west Antarctica are already bathed in summer meltwater, Hansen suggests that this ‘trigger mechanism’ of darker, wet snow is already engaged today.

      So how fast might sea levels rise? The IPCC's 2007 report suggests only 18 to 59 cm-reassuring figures to those who live close to the coast. However, it also introduces a caveat, admitting that uncertainties about ice sheet response times could make this figure higher. But it doesn't say how much higher, and no one else in the scientific community has ventured to offer an estimate-except, once again, James Hansen whose warning about global warming as long ago as 1988 suggests remarkable foresight. In a paper entitled ‘Scientific reticence and sea level rise’ and published in