Uncovering an Ocean
Much of the change in the Arctic is understood; little of it is reassuring.
NY-ALESUND IS a special place in Arctic science. A huddle of wooden buildings by an icy fjord on Norway’s island of Spitsbergen, high in the Barents Sea, it is the world’s most northerly civilian settlement. Ten countries—Norway, France, Germany, Britain, the Netherlands, Italy, Japan, South Korea, India and China—have research stations in the village, and scientists from many others pass through. All are welcome to the twin wonders of Arctic conditions and Norwegian management.
If you want to study clouds, barnacle geese, seismology, sea ice, zooplankton and any number of pollutants, all indicators of how the Arctic and the world are changing, Ny-Alesund is the place. The only rules are: remove your boots inside, carry a rifle outside (there are often polar bears about) and fly no national flags. The marble lions outside the Chinese station are tolerated but frowned upon.
The melting Arctic is the main subject of discussion in Ny-Alesund, and evident. Your correspondent visited in late winter, yet the fjord was unfrozen. The glacier at its head is rapidly retreating. How unusual is this?
Monitoring sea ice is a fairly recent activity. It began seriously in the 1950s, from aboard nuclear submarines. Satellite monitoring started in 1979. Since then the summer sea ice has shrunk by 12% a decade. That is consistent with the trend predicted by climate-change models over the past three decades, an indication that their mathematical simulations of global warming are roughly right. Using Viking epics, whaling and pollen records, log books, the debris shed by melted ice rafts, diatoms (silicon-armoured algae found in marine sediments), ice cores and tree rings, scientists have constructed a record of the Arctic past which suggests that the summer sea ice is at its lowest level for at least 2,000 years. Six of the hottest years on record—going back to 1880—have occurred since 2004. According to the IPCC, the last time the polar regions were significantly warmer was about 125,000 years ago.
This transformation is in fact happening faster than anyone had predicted. According to an authoritative 2011 assessment for the Arctic Council, “it is now becoming very clear that the cryosphere is changing rapidly and that neither observations nor models are able to tell the full story.” The albedo effect probably explains most of it, as indicated by research published in Nature in 2010. Its American authors found the greatest rise in the Arctic’s temperature during autumn, when the sea ice is at its minimum, over newly exposed sea. More powerful than expected, this positive feedback is active throughout the melting process: when white snow melts to expose darker sea ice; when melt pools form on the ice; and when sea water is exposed. With each change the Arctic surface absorbs more heat.
On thin ice
A simultaneous thinning of the sea ice is also speeding up the shrinkage, because thinner ice is more liable to melt. According to Peter Wadhams of Cambridge University, the average thickness of the pack ice has fallen by roughly half since the 1970s, probably for two main reasons. One is a rise in sea temperatures: in the summer of 2007 coastal parts of the Arctic Ocean measured 7°C—bracingly swimmable. The other was a prolonged eastward shift in the early 1990s in the Arctic’s prevailing winds, known as the Arctic Oscillation. This moved a lot of ice from the Beaufort Gyre, a revolving current in the western Arctic, to the ocean’s other main current, the Transpolar Drift Stream, which runs down the side of Siberia. A lot of thick, multi-year ice was flushed into the Atlantic and has not been replaced.
Attention has recently also focused on a suite of lesser-known greenhouse gases, including ozone and methane, and on soot from diesel exhaust and forest fires. These are known as “short-lived climate forcers”. Though they linger in the atmosphere for a relatively short time, they can have a powerful greenhouse effect. Soot, or black carbon, stays in the atmosphere for an average of six days, whereas carbon dioxide lasts for centuries, even millennia. Yet black carbon has an unusually potent warming effect in the snowy Arctic because the dark soot, after being rained or snowed onto bright snow or ice, continues to absorb heat.
There is speculation that high levels of atmospheric soot from American and European factory flues caused a mini-thaw in the Arctic in the 1930s. The UN’s Environment Programme estimates that reducing black carbon and methane emissions could cut Arctic warming by two-thirds over the next three decades. That would not prevent the disappearance of the summer sea ice, but it might delay it by a decade or two.
As Ny-Alesund’s spring research season began, a team of French soot researchers were among the first arrivals, after the Northern fulmars gliding over the fjord. They headed straight up Mount Zeppelin, which rises 473 metres behind the village and has a small laboratory on its summit, filled with whirring dials and a hum of motors, sucking in air from the winds racing overhead. With luck, the Frenchmen will reveal the role of soot in the Arctic melt. Yet the main reason is already known. A dial in the lab showing atmospheric carbon dioxide reads 398 parts per million, a 40% increase since the start of the industrial revolution.
A sea of worries
Around 125,000 years ago—when the IPCC thinks the Arctic was last much warmer than today—polar meltwater raised the sea level by 4-6 metres. If that happened again it would displace a billion people and inundate most of the world’s biggest cities, including New York, London and Mumbai. The chances of that are uncomfortably high—though it might take a couple of centuries—because of another Arctic surprise.
The IPCC’s prediction in 2007 of a rise in sea levels of up to 59cm by the end of this century was attended by large uncertainties over the world’s big ice sheets. Greenland’s is up to 3km (1.9 miles) deep and contains enough water to raise the sea level by 7.5 metres; the Antarctic ice sheets are much bigger and could potentially cause a 57-metre rise. In recent years they have also been monitored by satellite, with the first data obtained in 1992. Until about 2000 the ice sheets seemed to be more or less stable, with increased snowfall on their tops compensating for increased melting at the margins.
But in Greenland something has changed. The ice sheet’s recent rate of loss—around 200 gigatonnes a year—represents a fourfold increase on a decade ago. Half this melting is thought to be due to the warming atmosphere. The other half is due to warmer seawater, caused by global warming or a shift in Atlantic currents, or both. As a result, the sea is eating away the edge of the ice sheet at a faster rate. Between 2002 and 2007 the Jakobshavn Isbrae, a big glacier in western Greenland, retreated by 3km a year, shedding a total of over 36 billion tonnes of ice.
If the climate stabilises soon, the ice cap might resettle at a slightly lower mass than it has now, raising sea levels by only a few centimetres. But if the warming continues, the ice cap will continue to melt. Sooner or later, it is thought, a tipping-point would be reached when the decline would become terminal. “How long will the ice cap last?” asks Dorthe Dahl-Jensen of the University of Copenhagen, who has studied the subject in depth. “We don’t know because we don’t fully understand what determines the velocity of its ice streams.” But such a collapse has happened before. Around 15,000 years ago the Barents Sea ice sheet, which stretched from northern England to Siberia, disintegrated in perhaps less than 1,000 years, probably because of warming seas.
Melting glaciers point to another fear: disruption to the ocean’s overturning circulation. As warm surface waters flow north they cool and get saltier, which makes them denser. This denser water sinks and returns to the south at depth. The Gulf Stream and its northern extension, the North Atlantic Drift, are part of this circulation system, which is the main reason why Norway’s coast is 20°C warmer in winter than Canada’s at the same latitude.
Previous shutdowns of this system, deep-sea sediments suggest, were caused by massive infusions of freshwater into the North Atlantic; one of them, around 8,200 years ago, was caused by the sudden emptying of a vast North American lake. That would have made the Atlantic’s upper layers less dense, so less liable to sink. And now the Arctic’s upper layers are getting less dense, for several reasons: melting Arctic glaciers, rising surface-water temperatures, increased precipitation and an absence of salt concentrations resulting from sea-ice formation.
A paper published in Nature in 2005, by Harry Bryden of the National Oceanography Centre in Southampton and colleagues, presented the first data from an array measuring the overturning circulation. They seemed to suggest that the system might have slowed by 30% compared with 1957. As more data have become available, however, it is becoming clear that the difference can be explained largely by seasonal fluctuations. No climate model has predicted a complete shutdown during this century, though many suggest that a slowdown is possible.
Not every Arctic feedback will result in further warming. Less ice on rivers, lakes and the sea will lead to more evaporation, probably increasing cloud cover at mid-levels, which should have a cooling effect. Increased evaporation from the exposed sea should also mean more snowfall to compensate for the ice cap’s losses (though together with rapid erosion at the margins this could increase the gradient of the ice cap, making it less stable.) But this is small consolation for another big fear: rapid thawing of the Arctic permafrost.
Best kept cold
Roughly a quarter of the northern hemisphere, including most of the Arctic land, is covered by this layer of frozen rock, soil and organic carbon. Formed over millennia, it varies in depth from a few centimetres to up to 1,500 metres in Siberia. Much of the Arctic’s shallow continental shelf is also covered by permafrost. According to an estimate made in 2009, terrestrial permafrost holds about 1.7 trillion tonnes of carbon, roughly twice as much as the atmosphere. By another estimate subsea permafrost stores an additional 0.5 trillion tonnes. And underlying it there may be another 0.8 trillion tonnes in the form of methane hydrates, an icy white material discovered in the 1960s.
Though tricky to get at, methane hydrates could be a massive energy source. Globally they are estimated to contain more energy than all known deposits of fossil fuels. Yet if even a small portion of the methane contained in them were to be abruptly emitted, the warming effect could be catastrophic. Methane is a short-lived greenhouse gas—it stays in the atmosphere for 6-10 years before being oxidised—but it is 25 times more efficient than carbon dioxide at trapping heat. And no one is sure how stable the hydrates are.
Once unfrozen, the permafrost’s organic matter is either swiftly broken down by microbes that emit carbon dioxide or, in waterlogged conditions, which are common, it is eaten by a group of bacteria called methanogens that release methane. Either way the permafrost becomes a source of greenhouse gas. Yet the bacteria also release nitrates, which stimulate plant growth, so the thawing ground may sometimes become a carbon sink. All three things are now happening.
A 2011 study for America’s National Oceanic and Atmospheric Administration found evidence of thawing permafrost in most parts of the Arctic. Record high temperatures were logged at a three-decade-old permafrost station in Alaska, though there was no recorded change to the frozen surface in western Canada. Last year an American-Russian research team observed plumes of methane, some more than a kilometre across, bubbling up from the Siberian Sea, suggesting that its permafrost layer was no longer sealing the methane hydrates below. Such methane release might be due to a range of causes, including the resettling of subsea sediments, thermal contractions and seismic activity as well as thawing permafrost, and in this instance it was not clear what was responsible.
Watch our animation of the receding Arctic ice-shelf and the shipping routes it could unlock
One of the best available guides to this risk is a survey of 41 permafrost scientists published in Nature last year. They predicted that at the current rate of global warming between 48% and 63% of terrestrial permafrost would be thawed to a depth of 3 metres by 2100. In the process, they expected between 7% and 11% of its stored organic matter to be released into the atmosphere. Only a little over 2% of that would be in the form of methane, but this would be responsible for 30-50% of the resultant warming. It would be impossible to prevent these emissions: they would probably continue for centuries.
In 2009 and 2010 Europe and America experienced two of the coldest winters on record. That was not because global warming is not happening, but because the climate system is complicated. Arctic warming changes the temperature gradient between the tropics and the poles, which affects weather patterns across the northern hemisphere. Several recent papers, including one published by American and Chinese researchers in the Proceedings of the National Academy of Sciences in February, have suggested this could be behind the chilly winters. The theory is that they were caused by warmish, wet air rising from the exposed Arctic Ocean in autumn, creating areas of high pressure which destabilised the Arctic boundary layer. This sent snow-laden winds blowing into Europe and Siberia. It is too early to be sure about this. But compared with the disasters that a melting Arctic could bring, grumbling about the weather might seem like a comforting diversion.
Given the scale of these risks, it is extraordinary how little research has been done on permafrost. “There are a lot of white spots in our knowledge,” admits Leonid Yurganov, a permafrost expert at the University of Maryland—Baltimore County. But a lot has recently been learned, which suggests that an explosive methane release is very unlikely. Ice cores going back 800,000 years show no trace of such an event. Nonetheless, the release of permafrost or subsea carbon could be gradual and still cause a lot of warming, and that does seem likely.