ONE OF THE MOST TELLING images of our age was taken from Mars in May 2003. It shows the lonely, finite, comfortingly blue ball of Earth and its much smaller moon against a dark sky. Nothing else.
But if the picture had been taken between 600 and 800 million years ago, an influential group of Earth scientists argue, chances are it would have been even more stark – and in black and white. The familiar, friendly blue and green of ocean and continent would be replaced by alien white. Earth would look like a seemingly lifeless snowball hanging in space.
It was an era of global glaciation, these researchers say. Antarctica all over. There was ice at least a kilometre thick at the coastlines, much thicker in the continental interior, and extending right across the oceans.
At temperatures below -40°C, all moisture would have been frozen out of the air and the only clouds would be volcanic in origin. The only sounds would have been the eerie cracking of the creeping glacial ice and the occasional rumble of an earthquake or a volcano.
A little over 10 years after its first publication, in the U.S. journal Science, caused a flurry of debate, the 'snowball Earth' hypothesis still seems bizarre, and contentious. But, as I discovered at an international symposium on the topic, held at the University of Melbourne in September 2008, proponents of the theory believe it is on the way to evolving into an established scientific theory.
"The evidence is [now] a lot stronger," says geologist Paul Hoffman of Harvard University, in Cambridge, Massachusetts, the lead author of the original paper, and still an enthusiastic supporter of the hypothesis.
But doubters are concerned that the data they see does not fit.
"I can't make it work from my point of view," says Kath Grey, another symposium participant and chief palaeontologist at the Geological Survey of Western Australia, in Perth. She has been trying to match up the details of the story she reads in rock cores from Australia with what is being found overseas.
Snowball Earth is a creative idea that, like evolution by natural selection, the Big Bang and continental drift before it, provides an explanation for a set of disparate and unusual observations.
How is it that sediments of glacial origin directly abut rocks typically laid down in the tropics? Why is there evidence pointing to a stagnation of the oceans and the disappearance of almost all life between about 800 and 600 million years ago?
But the real excitement of snowball Earth, and the real source of controversy, is its power to explain one of the great mysteries of the fossil record – the appearance, almost overnight, of large, fully-fledged, multicellular animals, in great diversity.
First, around 575 million years ago, come the weird, mostly soft-bodied marine forms known as Ediacarans. Found in South Australia, Newfoundland, South China and northern Russia, most were attached to the ocean floor, and some reached up to four metres in length.
About 33 million years later, within a short (in evolutionary terms) period of only 20 to 30 million years, the ancestors of almost all modern animals appeared. This sudden burst of diversification is known as the Cambrian explosion.
FEW SCIENTISTS NOW dispute the idea of unusually widespread glaciation before the rise of animals. What's at issue is its extent – whether it covered the planet completely (snowball Earth) or if parts of the ocean remained open (slushball Earth) – and what role it played in evolution. That's the mystery of snowball Earth.
"It's still highly contentious," admits Hoffman. "My own read is that about 10 per cent of the community involved in this research support complete ice cover, about 10 per cent are fiercely opposed to it, and the other 80 per cent of people are sitting on the fence. I don't think there's going to be consensus until the original antagonists cycle out of the system. After all, [the theory of] continental drift took 55 years to be accepted.
"Like any new theory, it's imperfect. So there are a lot of things we think it predicts that are inconsistent with evidence. It may be that in some cases the evidence isn't right, but very often it's because the theory isn't right," he says.
The genesis of the hypothesis has been well documented by Hoffman himself and in a book and articles by science writer, Gabrielle Walker. Interestingly, snowball Earth has roots in Australia. It was the famous Australian Antarctic explorer and geologist, Douglas Mawson, who first considered the idea of global glaciation after coming across sediments in South Australia, clearly glacial in origin and far closer to the equator than he expected.
The sediments deposited by glaciers are very distinctive. As the ice scours underlying rock, it generates mixtures of boulders, stones and silt, particles of all different sizes, flattened and scratched as they are dragged along. Another clear sign of glaciation is the smooth, often very large rocks, known as dropstones, which are carried out to sea inside icebergs and eventually tumble into the seafloor sediments.
Glacial deposits are found all over the world. Many of them seem to have formed about the same time, in the late Neoproterozoic era around 700 million years ago.
Noting this in the '60s, geologist Brian Harland, at the University of Cambridge, suggested that the Earth may have gone through a great ice age. He also recognised that these glacial sediments often occur in rocks that could only have been laid down in the tropics, and inferred that the ice cover was so complete it must have extended into low latitudes. But how could that happen?
It was a Russian geophysicist, Mikhail Budyko, who provided a mechanism. The key was the 'albedo', the extent to which the Earth reflects rather than absorbs the Sun's radiation. The impact of albedo can be dramatic.
Researchers at University of Almería in Spain have recently shown, for instance, that in their southern Spanish province, the local concentration of greenhouses – the world's greatest – has lowered the region's average temperature by 0.3°C in the past 25 years (by absorbing light), while the average temperature in the rest of the country's has risen by 0.5°C.
During glaciation, the Earth's albedo is raised by white, reflective ice. Using a simple model, Budyko was able to show that if the Earth cooled enough for ice sheets to cover half its surface, reaching a latitude of 30° North and South, a positive feedback would set in.
The planet would then be so shiny, reflecting so much solar radiation back into space, that it would begin to cool naturally and freeze over completely – and quickly. Once the feedback took over, it would only be a matter of years, perhaps less, before the Earth iced over, according to Hoffman. But if that happened, how did the planet break free of its icy shackles?
THE ANSWER TO THAT question was provided by Joe Kirschvink of the California Institute of Technology, in Pasadena.
In the late 1980s, through his studies of palaeomagnetism (the positioning of magnetic minerals in rocks) he became convinced of the likelihood of a tropical glaciation about 700 million years ago. Like all magnets, magnetic minerals line up with Earth's magnetic fields. The closer they are to the equator, the more horizontally they lie.
Kirschvink was particularly impressed by glacial deposits in South Australia. Their history had been carefully studied by geologist George Williams, from the University of Adelaide, and there was good evidence to suggest that the magnetic minerals had not been remagnetised since they were first laid down, their tropical signature untouched.
Beside these glacial deposits are rocks made from sediments – sand and mud – affected by tidal action. So the glaciers had clearly been at sea level, not high in the mountains as we find in the tropics today.
Williams explained this anomaly by suggesting that the tilt of the Earth's axis with respect to the plane of its revolution around the Sun (now 23.5°) was then greater than 54°.
At this angle, the poles would receive more solar radiation than the equatorial regions, and what are now the tropics would be much colder than the poles. But little supporting evidence has been found. And the strong seasonality suggested by the idea would mitigate against the formation of glaciers.
Kirschvink thought global glaciation a much simpler explanation of the riddle of Williams' deposits. It was Kirschvink, in fact, who coined the term snowball Earth. And he also presented a way out of the permanent global winter. On snowball Earth, he proposed, although most movement would cease, plate tectonics would still be inexorably churning away beneath the surface, and that would lead to volcanoes belching out gases, especially CO2.
In a complete absence of rain, with the oceans, rocks and any plants covered in ice, there would be little to soak up this outpouring of gas. Over tens of millions of years, CO2 would build up to very high levels in the atmosphere, hundreds of times the present concentration, causing an intense greenhouse effect.
In a reversal of events, all the ice would melt in less than a thousand years as the Earth experienced temperatures of between 40°C and 50°C. Although there is evidence of global glaciations first occurring more than two billion years ago, the climate appears to have been cycling from one extreme to the other through the late Neoproterozoic era.
When Kirschvink published his ideas in the early 1990s, they seemed so bizarre that they were roundly attacked, until Hoffman and his geochemist colleague Daniel Schrag took them up.
Hoffman had been working on carbonate rocks in Namibia, in southern Africa. Carbonates are laid down in the oceans as the result of a reaction between dissolved CO2 and calcium or other metal ions. They are a sign of warmer, tropical waters. But there, among those warm water rocks, about 600 million years ago, were the distinctive glacial sediments.
When Hoffman checked the source of the carbon in the carbonate, he found another odd thing. Natural carbon comes in two varieties or isotopes, a common form known as carbon-12 and the rarer, heavier carbon-13, which has an extra neutron. Plants, which fix the carbon on which life depends, show a distinct preference for carbon-12.
So sea water in which there is life has a slightly higher level of carbon-13 compared with an environment in which there is no life. These concentrations are captured in any carbonate rock formed at the time and can be used to measure living activity.
When Hoffman analysed the isotope concentrations in his Namibian rocks, he found a clear progression from an ocean full of life when the glacial period began, to one nearly devoid of life at the end. Above the glacial period was a thick band of carbonates (now known as cap carbonates), a sign of a warm Earth with plenty of carbon. Similar bands of cap carbonates have since been found around the world.
He called in Schrag to check his carbon isotope results. Schrag not only confirmed what Hoffman had found but recognised that the existence of the cap carbonates fitted nicely with Kirschvink's idea of greenhouse melting. As the water cycle restarted, rain would dissolve and wash vast amounts of acidic CO2 out of the air. This would weather the rock beneath and carry the carbon and minerals into the oceans as the ice melted. In the warming seas, the mixture would react and precipitate to form the cap carbonates.
BUT WHAT TRIGGERED the runaway glaciation that led to snowball Earth? One answer is the same as what caused the more recent ice ages only more so: a wobble in the Earth's orbit took our planet further from the Sun resulting in less intense solar radiation.
But this cooling typically self-corrects, as it slows down the weathering of rocks that would normally absorb CO2. Retention of the gas in the atmosphere leads to global warming, which prevents glaciation from taking hold.
But in the late Neoproterozoic era, Kirschvink suggested, there was one major difference – albedo, again. It turns out that the massive supercontinent Rodinia was breaking up, leaving a great conglomeration of land masses (those now found in the Southern Hemisphere) strung along the equator, which eventually formed the more well-known Gondwana. This increased the albedo of the Earth just where it generally absorbed most solar radiation, and thus lowered the Earth's average temperature further.
The latest supporting evidence for snowball Earth, says Hoffman, comes from work published in Nature in June 2008 by a group led by Huiming Bao of Louisiana State University, in Baton Rouge, on rocks from South China, West Africa and Svalbard (the group of islands containing Spitsbergen, between Norway and the North Pole).
Bao and his colleagues have discovered unusual oxygen isotope signatures in these rocks related to photochemical reactions in the stratosphere. "These could only be preserved at the Earth's surface under conditions of very high CO2 concentrations and almost no photosynthesis and respiration," says Hoffman. "That's exactly what you expect in the snowball Earth state."
Even so, Hoffman thinks the simplest evidence remains the most compelling for the snowball Earth hypothesis. "The idea that there was ice at sea level in the warmest parts of the world, so therefore the colder parts must have been frozen as well, is not going to go away. There is no doubt about tropical glaciation."
Although no one believes the oceans froze solid, whether there was a thick, unbroken, crust of ice is still a bone of contention. This has significant evolutionary implications, because such an ice crust would prevent photosynthesis in the ocean.
Snowball critics argue that this would make life extinct, but supporters say organisms could persist as deep-frozen spores, in pools formed in cracks and under the ice, and in the small areas around hydrothermal vents, such as hot springs, fumaroles and black smokers under the sea.
Life would certainly be put through the wringer, however, having to adapt to tough conditions in small, far-flung patches – a recipe for rapid evolution and speciation. Perhaps enough to prompt the emergence of varied mobile, multicellular creatures; the rise of the animals.
Compared to such a snowball Earth, which he favours, Hoffman says an incompletely frozen slushball Earth would be relatively benign and not as evolutionarily selective. Unfortunately, it's not a question easily settled one way or the other. "We don't have any evidence of what was going on in the open ocean because the ocean floor has all been subducted."
But, he says, the time gap between the last snowball event and the first evidence of multicellular animals in the fossil record has now been reduced to less than two million years. These dates come from microfossils of animal eggs and embryos from the Doushantuo formation in South China, and traces of organic compounds typical of sponges left in rocks in Oman. "There now seems to me to be a much stronger case that these extreme climate events have something to do with the first appearance of multicellular animals."
AMONG EARTH SCIENTISTS, there now appears to be general agreement on three glaciations that potentially could have led to a snowball Earth.
The first happened about 2.2 billion years ago, the other two in the late Neoproterozoic era; the Sturtian (or Cryogenian) period, about 700 million years ago, and the Elatina (or Marinoan) period, about 635 million years ago. But there could have been many more, perhaps even a mix of snowball with slushball glaciations, severe but not complete.
The main concern is the lack of reliable dates. "We need a better handle on the absolute dating of [deposits]," says Kath Grey. "You need volcanics to date, and we just don't have that many in the succession. Finding them is difficult. They are usually little ash beds, very thin, and it's like trying to find a needle in a haystack."
And that's a real problem for many of the doubters. While Hoffman is working from his broad hypothesis down, they are working from the bottom up, trying to fit the evidence to the theory. Most are using forms of stratigraphy, tracing geological history by matching bands or strata of rock in different areas with each other. Presumably for a complete snowball Earth glaciation, you should find a matching band of glacial deposit all over the world.
Grey has spent over 30 years tracking the distributions of fossils in Australia, and matching them across the central and southern sedimentary basins of the country. Thanks to extensive petroleum and mineral exploration, Australia has a reasonably complete record in the form of drillcores from its great sedimentary basins.
"I'm urging caution about snowball Earth, for the simple reason that the Australian data that we've got so far – and the data's limited – doesn't necessarily fit with data that's coming in from overseas. And if it turns out that these glaciations are not synchronous, then we've got quite a big problem."
Within Australia everything seems to correlate quite nicely, she says, but there are issues about the actual age of the Elatina glaciation.
If you place this band at 635 million years ago, which correlates with what has been found in China, she says, you end up with four significant glaciations in Australia, whereas if you place it later, together with a well-dated glaciation from northeast Tasmania and King Island in Bass Strait at 580 million years ago, you only have three. There's no good evidence one way or the other, and either way you end up with more than the two generally accepted glaciations.
Grey sees other significant differences. She has been particularly interested in 'acritarchs', microscopic organic structures in the fossil record, which she thinks are probably planktonic green algae. They can be tracked for more than a billion years, but during the Neoproterozoic era, a new, larger, spiny form appears.
In China, this happens straight after their last major glaciation, a match with sites in Norway, Svalbard and the Himalayas. But in Australia there are different species, and they come in rapidly, long after glaciation and close to the debris from a huge asteroid impact in South Australia known as the Acraman event. This matches sites in East Siberia and Eastern Europe.
Another issue is the publication of a precise radiometric dating of the older Sturtian glaciation in South Australia by Mark Fanning of the Australian National University, in Canberra, and Paul Link of Idaho State University in Pocatello, USA. They put it at 659 million years ago, at odds with the estimated 700 million years that has been accepted until now.
"It certainly creates a problem," says Hoffman. "It doesn't leave much time between the two glaciations. I think it will eventually sort itself out, but it's certainly a bit of a complication at the moment."
WHATEVER THE OUTCOME of deliberations over snowball Earth, the theory has already served as a great stimulus and focus of the debate over the rise of animals, fostering many new ideas.
Patricia Vickers-Rich, a palaeontologist at Monash University in Melbourne, has switched from research on the origin of birds in past 65 million years to studying where animals came from, in part because of the excitement of working in a developing field.
"I'm just really pleased to be working in this area. It's fun to work in immature science, because you really feel that you might make a discovery that could make a fundamental difference in the way people look at things."
She believes the sudden appearance of mobile, complicated, multicellular animals at the end of the Neoproterozoic-era glaciations can be related to three events, or maybe a mix of them. One is an increase in oxygen in the atmosphere. The second is an injection of nutrients such as calcium and phosphates into the ocean, possibly due to the erosion of a huge Transgondwanan mountain range.
But the idea in which she is most interested at present is the freshening of the oceans as the ice melted. During significant glaciations the great bulk of freshwater is frozen as ice.
The majority of liquid water, underneath the ice cap in the oceans, would be highly saline, too salty for most marine life and certainly for multicellular animals. So Rich believes they would have had to retreat to less saline refugia, evolving and waiting for the great melt to make the oceans more habitable for them to rush out and colonise.
Another idea to emerge in the past couple of years – independent of snowball Earth but clearly a significant factor at the time of the rise of animals – is the formation and erosion of a huge Transgondwanan mountain range, between about 635 and 515 million years ago. A spin-off of the coming together of Gondwana, it would have been more than 8,000 km long and 1,500 km wide.
Evidence for its existence springs from the work of geologist Rick Squire, also at Monash University. He has studied the age of the mineral zircon within quartz grains in sands and sandstones all over the world. Wherever he went across what would have been Gondwana, he found a signal of rocks laid down within a narrow band between 650 and 500 million years ago.
And there is a pattern consistent with massive erosion during that time which moved about 100 million km3 of material into the oceans – enough to cover the entire contiguous part of Australia to a depth of about 10 km.
Squire believes this must have greatly increased levels of nutrients in the ocean, and that it also buried huge amounts of organic matter, allowing a build-up of oxygen in the atmosphere, necessary for complex animal life.
Wherever you stand on the snowball Earth idea, the glaciations of the late Neoproterozoic era provide a huge lesson in the power of the greenhouse effect, and the risks of modern climate change. It was real, it was fast, and it appears to have had a dramatic effect on evolution.
But snowball Earth itself is unlikely to happen again, says Hoffman, at least not in the near future.
"It gets more difficult because the Sun is getting closer and brighter all the time. Joe Kirschvink attributed the cooling to the fact that there were a lot of continents at low latitude, increasing albedo and weathering. So that's a very different geography from what we have today. We can't rule it out in the distant future, but that would be tens or hundreds of million of years away."
Tim Thwaites is a science journalist in Melbourne and the president of the Australian Science Communicators.
