本文作者：小红猪小分队

Source：http://www.newscientist.com/article/mg21628891.900-the-great-thaw-charting-the-end-of-the-ice-age.html?

Just 20,000 years ago, ice ruled the planet. So why did it relax its grip? Finally, it looks like the answers are in

DURING the summer of 2008, workers excavating Ground Zero in Lower Manhattan dug right down to the bedrock. There, they found something unexpected: a huge pothole more than 10 metres deep, the crevices around it crammed with stones of several different kinds of rock. The consulting geologist immediately recognised these features. The stones had been carried there from many miles away by a glacier that had ground across the bedrock. At some point, a swirling torrent of glacial meltwater had carved out the pothole.

From potholes in New York City to forests beneath the sea, evidence of the time ice dominated the world is all around us. The last great ice age began around 120,000 years ago. One massive ice sheet, more than 3 kilometres thick in places, grew in fits and starts until it covered almost all of Canada and stretched down as far as Manhattan. Another spread across most of Siberia, northern Europe and Britain, stopping just short of what is now London. Elsewhere many smaller ice sheets and glaciers grew, vast areas turned into tundra and deserts expanded as the planet became drier.

With so much ice on land, sea level was 120 metres lower than it is today. Britain and Ireland were part of mainland Europe. Florida was twice the size it is now, with Tampa stranded far from the coast. Australia, Tasmania and New Guinea were all part of a single land mass called Sahul. The planet was barely recognisable.

Then, 20,000 years ago, a great thaw began. Over the following 10,000 years, the average global temperature rose by 3.5 °C and most of the ice melted. Rising seas swallowed up low-lying areas such as the English Channel and North Sea, forcing our ancestors to abandon many settlements. So what drove this dramatic transformation of the planet?

Mysterious changes

We have long known the thaw began with an increase in summer sunlight in the northern hemisphere, melting ice and snow. It is what happened next that has remained mysterious. Soon after the thaw began, for instance, the southern hemisphere began to warm while the northern hemisphere cooled – the opposite of what was expected from the changes in sunshine. Now, after nearly two centuries of wrestling with seemingly contradictory findings, we think we finally understand how the ice age ended.

It all began in the 1830s, when Louis Agassiz noticed that characteristic features created by glaciers, such as scratches in the bedrock and “erratic” rocks dumped far from their place of origin, could be found far from existing glaciers. Similar discoveries were soon being made all over the world, from Canada to Chile. It became clear that there had been a whole series of ice ages.

What had made the ice come and go? In 1864, James Croll proposed that changes in the amount of sunlight reaching different parts of Earth’s surface, due to changes in the planet’s orbit, were responsible. He also suggested that the orbital effects had been amplified by various feedback mechanisms, such as the melting of heat-reflecting snow and ice, and changes in ocean currents.

Croll got many of the details wrong, but he was on the right track. Early in the 20th century, the Serbian astronomer Milutin Milankovitch concluded that summer sunlight in the northern hemisphere must be the crucial factor and spent years painstakingly calculating how this had changed over the past 600,000 years. His ideas weren’t accepted at the time, but in the 1970s studies of ocean-sediment cores revealed that the advances and retreats of the ice ages did indeed coincide with “Milankovitch cycles”.

Yet many enigmas remained. For starters, the changes in sunshine were tiny. Even if they were amplified by more of the sun’s heat being absorbed by the planet as snow and ice melted, it was hard to account for the scale of the global changes. What’s more, when summer sunshine increases in the northern hemisphere, it decreases in the southern hemisphere. This had led Croll to suggest that ice ages alternate between hemispheres: when the north freezes the south thaws and vice versa. But it had long been clear that the whole world had warmed at around the same time.

The answer to these puzzles seemed to emerge in the 1980s, when ice cores drilled in Antarctica revealed an astonishingly close correlation between atmospheric carbon dioxide levels and temperature.

“For the last million years, you see these two going up and down, up and down, together through each ice age, and it’s almost in perfect lockstep,” says Jeremy Shakun of Harvard University. “It’s about as beautiful a correlation as you ever get from nature.”

If CO₂ levels had risen soon after the thaw began in the north, it would explain why the southern hemisphere began to warm too. It would also help to explain the magnitude of the changes. But this promising idea ran into a major problem: by around a decade ago, it had become clear that the Antarctic starting warming a few hundred years before CO₂ levels began to rise. So while soaring CO₂ levels undoubtedly warmed the planet – they are now thought to be responsible for about half of the warming as the ice age ended – they weren’t the initial cause. “Something else was causing Antarctica to warm,” says Daniel Sigman of Princeton University.

Pollen puzzle

This wasn’t the only mystery. In the 1930s, studies of sediments containing the pollen of the alpine flower Dryas octopetala and other plants suggested that almost as soon as Europe began warming, it suddenly got cold again. This cold phase, called the Oldest Dryas or Mystery Interval, lasted from around 17,500 years ago to 14,700 years ago. Ice cores later showed Greenland cooled at the same time.

Yet during this period Antarctica warmed steadily. “On the detailed scale, the south seems to warm before the north,” says Sigman. But what would make the southern hemisphere warm even as the northern hemisphere cooled? It could not be due to orbital changes or rising CO₂ levels – but it could be due to changing ocean currents.

As the vast ice sheets began to melt 19,000 years ago, stupendous quantities of fresh water poured into the North Atlantic (see diagram). Studies of marine sediments off the Irish Sea coast, for example, show that the sea level there rose about 10 metres in just a few hundred years (Science, vol 304, p 1141).

Today in the North Atlantic, salty water arriving from the tropics cools, becomes very dense and sinks to the bottom. These deep, cold waters flow all the way to the southern hemisphere, while on the surface warm water – including the Gulf Stream – flows north. This system of currents is called the Atlantic meridional overturning circulation.

The huge quantities of fresh water pouring into the ocean 19,000 years ago would have diluted the salty water, making it less dense. Result: a slowdown in the overturning circulation. The proof came in 2004 from a study of ocean sediments. The ratio of two heavy elements, which indicates the speed of the deep current, showed that the overturning circulation had almost ground to a halt 17,500 years ago (Nature, vol 428, p 834).

The result was a kind of see-saw effect. With much less heat being carried north by the surface currents, the northern hemisphere cooled. The tropical and subtropical regions of the southern hemisphere, by contrast, began warming as they were losing less heat to the north. This explains many puzzling findings. The slowdown of the Atlantic current can also help explain why CO₂ levels rose during the great thaw (see diagram).

By the 1990s, the search for the source of the CO₂ was focusing on the Southern Ocean. Isotopes in ocean sediments suggested that a huge reservoir of CO₂ had built up in deep waters during the ice age. It is thought that a lack of vertical mixing, along with a cover of sea ice, trapped the gas. During the thaw, however, the ocean was “uncorked” and much of the CO₂escaped into the atmosphere.

Confirmation came earlier this year, thanks to a very detailed isotopic analysis of the CO₂ trapped in ice cores from Antarctica. “The CO₂ must have come from the deep ocean,” says team member Jochen Schmitt of the University of Bern in Switzerland.

Increased vertical mixing in the Southern Ocean is now widely accepted as being behind the release of CO₂. In 2009, for instance, Bob Anderson of the Lamont-Doherty Earth Observatory in New York reported that the Southern Ocean saw big increases in the growth of plankton with silica shells during the Oldest Dryas, when the southern hemisphere began warming (Science, vol 323, p 1443). As the growth of these organisms is limited by how much dissolved silica there is in surface waters, the increases must be due to the upwelling of water rich in silica and other nutrients.

But what caused it? There are two ideas. Sigman points out that Antarctica began warming at almost the same time as the waters just south of the equator. By itself, though, the shutdown of the Atlantic current should only have warmed waters in the tropics, not those as far south as Antarctica. In 2007, his team proposed that when the Atlantic conveyor shut down, it was replaced by a local overturning circulation in the waters around Antarctica. Dense surface water sank and deep water welled up, releasing both heat and CO₂. “That would explain both the Antarctic warming and the CO₂ rise,” says Sigman.

Anderson and his colleagues, however, think that the increased upwelling was driven by changes in winds. Earth has distinct bands of prevailing winds, driven by the temperature differences between the poles and the tropics, coupled with the planet’s rotation. Their positions can change when the temperature differences change.

During the ice age, the band of westerlies in the southern hemisphere – which sailors call the Roaring Forties due to their latitude – would have been further north. The see-saw effect shifted it southwards over the Southern Ocean, warming Antarctica and stirring up the sea around the frozen continent. In particular, the wind-driven circular current would have produced more upwelling in the shallower region between South America and Antarctica.

While the details are still being debated, the big picture now seems clear. “There is still some disagreement about the processes occurring in Antarctica as the last ice age ended,” says Anderson. “But at least the broader features are pretty well accepted.”

Earlier this year, Shakun and colleagues drew together many of these strands of research with an analysis of 80 different records of temperature and atmospheric composition over the past 22,000 years (Nature, vol 484, p 49). Their work pretty much confirms the sequence of events that ended the ice age. It goes like this:

Around 20,000 years ago, the northern ice sheets had spread so far south that just a small increase in sunshine led to extensive melting. As fresh water poured into the North Atlantic, the overturning circulation shut down, cooling the northern hemisphere but warming the southern hemisphere. These changes were mostly due to a redistribution of heat – by 17,500 years ago, the average global temperature had risen just 0.3 °C.

Changing winds or currents, or both, then brought more deep water to the surface in the Southern Ocean, releasing CO₂ that had been trapped for thousands of years. As atmospheric levels climbed above 190 parts per million, the whole planet began to warm. The far north was the slowest to respond, but by around 15,000 years ago, as CO₂ levels approached 240 ppm and the Atlantic overturning circulation sped up again, temperatures started to shoot up. The recovery of the overturning circulation had the opposite effect in the southern hemisphere: warming stalled and the release of CO₂ stopped.

Around 12,900 years ago, the see-saw swung again. Temperatures in northern latitudes suddenly plummeted and remained cold for about 1300 years. This cold snap, called the Younger Dryas, is thought to have been caused by a colossal meltwater lake in North America, which held more water than all the Great Lakes put together, suddenly flooding into the Atlantic and shutting down the overturning circulation once again.

The Southern Ocean, meanwhile, started releasing CO₂ again. Levels in the atmosphere shot up to 260 ppm, causing the whole planet to warm rapidly over the next couple of millennia. By around 10,000 years ago, Earth had been transformed. The ice had retreated, the seas had risen and our ancestors were learning how to farm.

Technically, though, the ice age has not actually ended. The ice has advanced and retreated many times over the past few million years, but some ice has always remained at the poles. Perhaps not for much longer, though. It took just a small increase in sunshine and a gradual, 70-ppm rise in CO₂ to melt the great ice sheets that once covered Eurasia and America. Since the dawn of the industrial age levels have risen by 130 ppm and counting. If we haven’t already pumped enough CO₂ into the atmosphere to melt the ice sheets on Greenland and Antarctica, we might soon.

Fortunately for us, it might take thousands of years for the last great ice sheets to vanish altogether. If it does happen, though, perhaps one day builders in Antarctica will find massive potholes in the bedrock carved by meltwater, and reflect on another dramatic transformation of the planet.

Anil Ananthaswamy is a consultant for New Scientist based in Berkeley, California