Some very strange events have left their signs in the rocks of Precambrian time, the nine-tenths of Earth's history before fossils became common. Various observations point to times when the whole planet appears to have gripped by colossal ice ages. Big-thinker Joseph Kirschvink first assembled the evidence in the late 1980s, and in a 1992 paper he dubbed the situation "the snowball earth."
Evidence for the Snowball Earth
What did Kirschvink see?
- Many deposits of Neoproterozoic age (between 1000 and about 550 million years old) show the distinctive signs of ice agesyet they involved carbonate rocks, which are made only in the tropics.
- Magnetic evidence from these ice-age carbonates showed that indeed they were very near the equator. And there is nothing to suggest that the Earth was tilted on its axis any differently from today.
- And the unusual rocks known as banded iron formation appeared at this time, after an absence of more than a billion years. They have never reappeared.
These facts led Kirschvink to a wild surmise glaciers had not just spread over the poles, as they do today, but had reached all the way to the equator, turning the Earth into a "global snowball." That would set up feedback cycles reinforcing the ice age for quite some time:
- First, white ice, on land and upon the ocean, would reflect the sun's light into space and leave the area cold.
- Second, the glaciated continents would emerge as the ice took water from the ocean, and the newly exposed continental shelves would reflect sunlight rather than absorbing it as dark seawater does.
- Third, the huge quantities of rock ground into dust by the glaciers would take carbon dioxide from the atmosphere, reducing the greenhouse effect and reinforcing the global refrigeration.
These tied in with another event: the supercontinent Rodinia had just broken apart into many smaller continents. Small continents are wetter than large ones, hence more likely to support glaciers. The area of continental shelves must have increased, too, thus all three factors were reinforced.
The banded iron formations suggested to Kirschvink that the sea, blanketed in ice, had gone stagnant and run out of oxygen. This would allow dissolved iron to build up instead of circulating through living things as it does now. As soon as the ocean currents and continental weathering resumed, the banded iron formations would be quickly laid down.
The key to breaking the glaciers' grip was volcanoes, which continually emit carbon dioxide derived from old subducted sediments (more on volcanism). In Kirschvink's vision, the ice would shield the air from the weathering rocks and allow CO2 to build up, restoring the greenhouse. At some tipping point the ice would melt, a geochemical cascade would deposit the banded iron formations, and snowball Earth would return to normal Earth.
The Arguments Begin
The snowball earth idea lay dormant until the late 1990s. Later researchers noted that thick layers of carbonate rocks capped the Neoproterozoic glacial deposits. These "cap carbonates" made sense as a product of the high-CO2 atmosphere that routed the glaciers, combining with calcium from the newly exposed land and sea. And recent work has established three Neoproterozoic mega-ice ages: the Sturtian, Marinoan and Gaskiers glaciations at about 710, 635 and 580 million years ago respectively.
The questions arise as to why these happened, when and where they happened, what triggered them, and a hundred other details. A wide range of experts found reasons to argue against or quibble with the snowball earth, which is a natural and normal part of science.
Biologists saw Kirschvink's scenario as looking too extreme. He had suggested in 1992 that metazoansprimitive higher animalsarose through evolution after the global glaciers had melted and opened new habitats. But metazoan fossils were found in much older rocks, so obviously the snowball earth had not killed them. A less extreme "slushball earth" hypothesis has arisen that protects the biosphere by positing thinner ice and milder conditions. Snowball partisans argue their model cannot be stretched that far.
To an extent, this appears to be a case of different specialists taking their familiar concerns more seriously than a generalist would. The more distant observer can easily picture an icelocked planet that has enough warm refuges to preserve life while still giving the glaciers the upper hand. But the ferment of research and discussion will surely yield a truer and more sophisticated picture of the late Neoproterozoic. And whether it was a snowball, slushball or something without a catchy name, the type of event that seized our planet at that time is impressive to contemplate.
PS: Joseph Kirschvink introduced the snowball earth in a very short paper in a very large book, so speculative that the editors didn't even have someone review it. But publishing it was a great service. An earlier example is Harry Hess's groundbreaking paper on seafloor spreading, written in 1959 and circulated privately before it found an uneasy home in another large book published in 1962. Hess called it "an essay in geopoetry," and ever since the word has had a special significance. I do not hesitate to call Kirschvink a geopoet as well. For instance, read about his polar wander proposal.