Climate-caused "end of the world as we know it" stories in recent years have concentrated on a warming planet: cities underwater, permanent droughts, extinctions. It looks more and more as though in the past, however, cold had even more dramatic an impact than the putative warming is predicted to be having now. Glaciers that came as far south as New York and Wisconsin, as some did 18,000 years ago, were not the problem. No, the whole earth — including the oceans — froze over. We were a blinding white Christmas tree ornament in the blackness of space: "snowball earth."
The latest intimation that there might have been such episodes came from geological formations in Africa. In Namibia (and in other places) strange sequences of rocks were laid down on the bottoms of long-vanished oceans between 550 million and 750 million years ago. Layers of jumbled rocks (presumably brought there by glaciers) are overlain by a type of rock called carbonate, which indicates intense weathering. For six summers, Harvard geologist Paul Hoffman labored in the African heat to study these sequences, while his colleague, geochemist Daniel Schrag, interpreted the results of painstaking chemical analyses.
"The carbon isotopes in those rocks tell us that for millions of years the earth was frozen over," says Schrag. "But while the ocean was biologically dead and the land covered, the interior of the planet was working just fine."
To understand this, you have to look at the earth's carbon cycle not from the relatively short timescale of human beings but from the grand view of the geologist. Over millions of years, carbon is brought up from the interior and emitted into the atmosphere as carbon dioxide by volcanoes. When rain falls, some of the carbon dioxide in the air combines with the water to become a weak solution of carbonic acid. This in turn dissolves granite, beginning a process that ends in calcium carbonate (the stuff of seashells and marble) being present in the runoff that eventually reaches the sea. There it precipitates out of the seawater and falls to the ocean bottom, where the tons of pressure turn it into carbonate rock.
"During the great freezing," says Schrag, "carbon dioxide levels in the atmosphere dropped, the earth's temperature fell to around 58 degrees below zero, and the ice covered everything — ocean and land alike." Glaciers moved down from the mountains, dragging along rocks and the rubble that would be left behind when the ice melted.
Once the world iced over, carbon dioxide brought up by volcanoes could no longer be removed from the atmosphere: there were no rivers, no rain or snow and no weathering. For millions of years, carbon dioxide levels climbed, eventually rising to about 300 times what they are today. At this point the greenhouse effect took off with a vengeance and the ice melted back, dropping the glacial rocks onto the ocean floor. Rain began falling again, the resulting carbonic acid dissolved granite again and the whole process cranked up once more. The layer of glacial debris left behind on the ocean floor was covered by another layer of carbonate. These are the layers we see today in places like Namibia. "We have pretty good evidence that the earth went through this cycle at least twice," says Schrag, "and maybe as many as four times."
But if you accept that the snowball event happened (perhaps many times), there are still some questions that need to be answered: How did the earth get into a frozen state in the first place? Could it happen again, and how could any life survive such an event at all?
To discuss these questions, I talked to a man who makes his living producing computer models of planets and their atmospheres. Ray Pierrehumbert, of the University of Chicago, has thought deeply about the evolution of the atmosphere on Mars, as well as the greenhouse effect here at home, and was among the first theoretical scientists to think seriously about the snowball earth scenario.
"The big hole in the [snowball earth] theory right now is understanding how you get the earth to freeze over," he says. An asteroid impact could do it, by throwing up a worldwide layer of dust that would shade and thus cool the planet, but to Schrag this seems an unlikely scenario if snowball events happened repeatedly. Most scientists assume that the freezing had something to do with the levels of carbon dioxide in the atmosphere. Just as adding carbon dioxide can make a planet warm up, removing it can make it cool down. Currently, there is some argument about how low carbon dioxide levels have to be to initiate freezing, but there is a general consensus that low levels are the culprit.
So, should we all start investing in snowmobiles and parkas? Could there be a snowball in our future?