IS EUROPE'S central-heating system about to break down, causing climate chaos around the world? Late last year, oceanographers reported a sudden and shocking slowdown in the currents of the North Atlantic, a critical part of the vast system of ocean circulation that influences temperatures and weather around the world. A shutdown could cause famine in south Asia, kill off the Amazon rainforest and plunge western Europe into a mini ice age.

However, if you live in Europe, don't order that snowcat just yet. The conclusions reported last year have been dismissed by many climate scientists, who say their models show the current will keep going for at least another hundred years or so. So what is really going on? Are changes in ocean circulation about to turn our lives upside down, or is this something only our grandchildren will have to cope with?

This vital question is in doubt because the behaviour of ocean currents is still remarkably obscure. On a crude level, the oceans of the world are linked together by a network of currents sometimes called the global conveyor, with warm surface flows connecting to cold deep currents. The conveyor is driven by winds and by a more complicated process called thermohaline circulation - and this is the process that has climatologists worried.

As its name implies, thermohaline circulation depends on heat and salt. An offshoot of the Gulf Stream called the North Atlantic Drift flows all the way to the seas off Greenland and Norway. Evaporation makes the water saltier, so as it is chilled by Arctic winds it becomes denser than the waters underneath it and sinks. It then spills back southward over the undersea ledges between Greenland and Scotland to form a slow, cold, undersea river called the North Atlantic Deep Water. This flows all the way to the Southern Ocean, with some water going as far the Indian Ocean, where it gradually wells up again, perhaps a millennium after it sank.

The weak link is the sinking process. Climate change is injecting ever more fresh water into the Arctic by increasing river flows and accelerating the calving of icebergs from Greenland. This fresh water dilutes the North Atlantic Drift, reducing its density and making it more buoyant. If the fresh water input reaches a critical rate, around 100,000 tonnes per second, sinking could stop entirely. The northern branch of the conveyor would stop, and warm tropical waters would no longer flow past the west coast of Europe.

With that million-gigawatt heat supply switched off, climate models suggest that air temperatures in the region could fall by between 5 and 10 �C, and parts of the US and Canada would suffer too. A switch-off like this is blamed for a cold snap 12,000 years ago called the Younger Dryas, which turned the forests of Scandinavia into tundra.

Could it be happening again? That spectre was raised in December by Harry Bryden of the National Oceanography Centre in Southampton, UK. His team took a ship from Florida to the coast of North Africa, stopping at 120 points en route to lower a bundle of instruments all the way to the sea floor. The researchers compared their results with similar measurements made at irregular intervals since 1957. According to their analysis, the deep, cold return leg of the circulation has weakened by 30 per cent (New Scientist, 3 December 2005, p 6). If that has slowed, they reasoned, then the northward branch of warm water must have slowed too.

In fact, the slowdown seems to have started nearly a decade ago. When the US National Oceanic & Atmospheric Administration made a similar survey of the Atlantic in 1998 it was interested in carbon dioxide levels and did not calculate the flow rate. When Bryden's team did the sums, they found the flow had been relatively steady between 1957 and 1992, dropped off by 1998 and remained low.

Bryden's paper prompted some nervous press coverage. "There were alarming stories saying that the sky is falling," says Carl Wunsch, a physical oceanographer at Massachusetts Institute of Technology. "It's a complicated story reduced to a fairytale". In fact, Bryden's measurements are not proof of imminent cataclysm.

One question mark is whether his team has simply seen short-term fluctuations in the ocean. "The ocean is a very turbulent beast. We tend to assume that at great depth it is quiet, but that's not necessarily so," says Stefan Rahmstorf of the University of Potsdam in Germany. Wunsch likens it to the vagaries of the weather: "It might get colder for a few days in England, but you don't necessarily say we're entering a new ice age." Bryden, however, thinks his team has found more than a stutter. "If we just had the 1998 data we'd be nervous, but 2004 is similar to 1998," he says.

Are the models wrong?

There are other reasons to be cautious. Climate models do not predict any substantial slowdown in Atlantic currents until near the end of this century. "It would mean all our models are wrong," says Rahmstorf. Bryden thinks they might well be: "I think if we measure a slowdown, the models will follow."

The trouble for Bryden is that not all observations fit in with his conclusions. If less warm water is flowing north, the seas off western Europe ought to be cooler than normal. They are not. In fact, these waters are slightly warmer than a decade ago. And direct measurements of the cold, deep currents that spill southwards over the ledges joining Scotland, Iceland and Greenland do not show a downward trend. Although these currents did slow between 1995 and 2000, they have picked up again. "We are faced with conflicting evidence," says Rahmstorf.

How to resolve this conflict? It is possible that Bryden's group has got its physics wrong. Like other groups, the team did not measure flow rates directly, but instead calculated them from measurements of temperature and salinity. Rahmstorf and Wunsch both point out that these calculations rely on assumptions that are far from proven.

Or it could be that the currents are changing in ways that no one has anticipated. There is a vast stretch of ocean between Bryden's measurements at 25� north and the overspill at around 65�. In between, the warm surface current becomes meandering and unstable, and difficult to measure, says oceanographer Tore Furevik of the University of Bergen in Norway. "There are certainly large changes going on beneath the surface of the North Atlantic, but we are still missing too many pieces in the puzzle to get the picture clear," he says.

For the moment, this rather unsatisfactory answer is the best we have. Oceanographers and climate scientists agree that thermohaline circulation will slow as the world warms, but most think it will happen later rather than sooner. In its report due next year, the Intergovernmental Panel on Climate Change is likely to predict a slowdown of at most 50 per cent by the end of this century.

This prediction, however, relies on estimates of the freshwater input. While the discharge from Siberian rivers is being monitored, that from Canadian rivers is not. The input from the Greenland ice cap could change too. The glaciers that drain the ice cap are accelerating, and in the past decade the amount of ice they spit into the ocean has doubled. Nobody can predict with confidence what they will do in the coming decades. "At the moment models don't represent the dynamics of Greenland glaciers, which may or may not start moving faster," says Richard Wood of the Hadley Centre for Climate Prediction and Research in Exeter, UK. What all this means is that even if Bryden is wrong and Atlantic circulation is not yet slowing, a shutdown could still happen sooner than most models predict.

Farming hit hard

For Europe, the timing of any slowdown or shutdown is critical. If it does happen soon, the weather will certainly get chilly. Average temperatures would be about 5 �C lower, and winters could be as harsh as those in Newfoundland. In bad years the Thames might freeze over, and even in good years farming will be hit hard.

On the other hand, if currents hold fairly steady until the late 21st century, the cooling effect of a shutdown would help to mitigate warming. There might be drastic changes in other aspects of the climate - not to mention a relatively rapid rise in sea level around the northern Atlantic (see "Rising waters") - but Europeans might escape much of the warming that occurs elsewhere.

Their distant descendants might need those snowcats, though. If greenhouse gases do eventually fall to pre-industrial levels and the world cools down again, there could be a lag of a thousand years before ocean circulation restarts. So Europe still faces the big freeze - just not for a few hundred years.

Coming back to this century, other parts of the world face even more serious consequences than Europe. A slowdown in the thermohaline circulation would reduce the transfer of heat from the southern to the northern hemisphere, shifting the Earth's "thermal equator" to the south. "One of the things that really struck us is that rainfall patterns over the whole world change dramatically," says Wood.

When he tried artificially switching off thermohaline circulation in one climate model, he found that monsoon rains weaken over India, and parts of central and south America lose half their rainfall. "It would have a huge impact on the climate of those regions," says Wood. He estimates that agricultural productivity in parts of India could fall by 30 per cent. And in the Americas? "If you lose the rain then the rainforest tends to die out." Although all of this is based on an imminent shutdown, which climatologists think very unlikely, even a delayed slowdown could seriously disturb rainfall patterns.

Wood's model also predicts that a shutdown would warm the southern hemisphere by 0.2 �C on average - not much, but against a background of rising global temperatures any extra warming will hardly be welcome. In 2004, Brazil was hit by the first ever hurricane recorded in the South Atlantic, perhaps a consequence of rising sea-surface temperatures. Could a slowdown in circulation have contributed? "Theoretically, this is possible," says Michael Mann of Pennsylvania State University in Philadelphia, "but I think it would be a leap to tie any observed change to thermohaline circulation. It could just be fluke."

While the precise effects on the climate remain uncertain, there is little doubt that a shutdown will wreak serious damage beneath the waves, since upwelling waters supply vital nutrients to the phytoplankton that are the basis of ocean food chains. A study last year predicted that the productivity of the world's oceans would fall by a fifth if the Atlantic thermohaline circulation shuts down.

There's more. At the moment, the oceans are soaking up a lot of the excess carbon dioxide pouring into the atmosphere. Without the thermohaline circulation, however, surface waters will soon become saturated and greenhouse gases will build up faster still. Any reduction in carbon fixing as forests and ocean ecosystems fail would only compound the effect.

A lot therefore depends on what is really happening in the Atlantic. To find out, the UK has launched a project called RAPID, an unprecedented effort to monitor North Atlantic currents. In 2004, on the same voyage that found the controversial signs of a slowdown, Bryden's team planted a series of 22 moorings along a line from Africa to America (see Diagram). Cables fixed to the seafloor tether instrument packages that are constantly measuring ocean properties such temperature and salinity.

With continuous measurements now coming in from the Atlantic, it should be possible to distinguish between short-term fluctuations and a longer-term trend. "Soon we'll find what seasonal variability there is and know whether what we said was a 30 per cent slowdown was above the noise level. I'm hoping we'll eventually be able to get 10 years' worth of measurements."

Then again, monitoring a single cross section of one ocean might not be enough. "People are obsessed with the North Atlantic and it's only 10 per cent of the ocean," Wunsch says. "There is a danger we're neglecting the rest." For example, there are areas of sinking ocean water around Antarctica that also help to drive the global ocean circulation. "Recent papers have started to suggest there are changes happening in the Southern Ocean," says oceanographer Steve Rintoul, based in Hobart, Australia.

Some studies show that the water in some Antarctic seas is getting less salty, for instance, but the picture in the Southern Ocean is even less clear than it is in the Atlantic. "We have to observe this system globally and indefinitely," says Wunsch. "But how the devil do you get governments to do that?"

From issue 2547 of New Scientist magazine, 15 April 2006, page 42

Rising waters

In the Hollywood disaster movie The Day After Tomorrow, the shutdown of the Atlantic circulation somehow results in a giant wave crashing over New York. That is nonsense, of course, but a shutdown of the Atlantic thermohaline current would raise sea levels in the North Atlantic.

At present, sinking waters in the Arctic produce a kind of plughole effect, lowering the sea level of this region and slightly raising it elsewhere. If sinking stops, sea levels will fall in the Southern Ocean and rise by up to a metre on some North Atlantic coasts.

A group led by Anders Levermann of the Potsdam Institute for Climate Impact Research in Germany has run a simulation that shows northern Norway getting a metre rise by the time the circulation stops altogether, while the UK, north-eastern Canada and the US see rises of up to 80 centimetres (see Map). With global sea levels also rising as the oceans warm up and glaciers melt, these regions face total rises of well over a metre by the end of the century.

One NASA study shows that if the sea level was just a metre higher on the New York coast, storm surges that now occur just once in a century would happen almost every year, flooding some low-lying suburbs. In the Thames estuary, planners believe they can handle total rises of up to two metres, but if there is also a large increase in the frequency and ferocity of storms, as some models predict, then an extra half-metre rise from a thermohaline shutdown could mean abandoning some areas to the sea.