I think about the cosmic snowball theory. A few million years from now the sun will burn out and lose its gravitational pull. The earth will turn into a giant snowball and be hurled through space. When that happens it won’t matter if I get this guy out.
—BASEBALL PITCHER BILL LEE
PUZZLE DOWN UNDER
Doing geology in Australia can be both a blessing and a challenge. On the plus side, most of the continent is a dry desert or scrub habitat, so there are lots of bare, exposed outcrops of rock. There is very little vegetation, unlike wetter parts of the world, where the plants overgrowing nearly everything hampered early British geologists (chapters 4–7) and the other geologists who came after them. Most of my geological career has been spent in deserts and badlands, because these are the only places with suitable exposures of beds to find fossils. On the minus side, Australia is a large, stable block of continental crust that has undergone relatively few great collisions or mountain-building events that force sedimentary basins to sink and rocks to erode, and none since 250 Ma. Most of the sedimentary deposits in much of Australia are very thin and discontinuous, so they don’t lend themselves to the study of long sequences of rocks and fossils through time, unlike many other parts of the world.
Some parts of the geological record are missing or nearly so. Australia has a thick sequence of deposits for much of the Precambrian (such as the stromatolites in chapter 13, the banded iron formations in chapter 14), especially the upper Precambrian rocks that are the subject of this chapter. Most of the Paleozoic units are relatively thin and unfossiliferous compared with many other continents. The Australian Mesozoic record has some bright spots, but no thick stack of richly fossiliferous dinosaur-bearing beds like you find in North and South America and Eurasia and Africa. By the Cenozoic, tectonic activity shut down almost completely, so there are very few mammal-bearing beds and a relatively poor fossil record of the Cenozoic. The exceptions to this rule are unusual deposits like Riversleigh, from the Miocene of Queensland, which consists mostly of fossil mammals and other terrestrial animals that fell into sinkholes in limestone.
Nevertheless, Australian geologists have made the best of what they have. The Precambrian banded iron formations have been intensively studied by many scientists, as have the late Proterozoic soft-bodied fossils of the Flinders Range, or the incredible beautifully preserved Devonian fish of the Gogo Formation.
One of these geologists was the legendary Sir Douglas Mawson (1882–1958), who was also a famous explorer (figure 16.1). He began his career mapping rocks in Melanesia and then in New South Wales, but in 1907 he joined Sir Ernest Shackleton’s expedition to Antarctica, where he stayed for over 2 years, after most of the rest of the explorers had returned. Mawson was the first man to reach the top of Antarctica’s second highest volcano, Mount Erebus, at 3,794 meters (12,448 feet). He also was one of the first to reach the South Magnetic Pole. In a fortunate twist of fate, he turned down a chance to join Robert Falcon Scott’s ill-fated 1910 expedition to the South Pole. They reached their destination only after Norwegian Roald Amundsen had gotten there first, then all of Scott’s crew died trying to return.
Figure 16.1
Photographs of Douglas Mawson: (A) Mawson resting on a sledge at the beginning of the Australasian Antarctic Expedition to Antarctica in 1912. (B) Photo taken in 1913, with a bearded Mawson still suffering from frostbite and malnutrition after he returned as the sole survivor of the expedition. (Courtesy of Wikimedia Commons)
Instead, Mawson led his own Australasian Antarctic Expedition in 1911, which mapped and studied much of East Antarctica. Despite great discoveries, the conditions got bad and most of the explorers perished on their return to home base. After everyone else died, Mawson and his last companion, Xavier Mertz, ate all their sled dogs. They both got hypervitaminosis A from eating too much dog liver, and Mertz died. Alone, Mawson hiked back over hundreds of miles in subfreezing conditions, even falling into a crevasse and dangling over the abyss in the sledge harness. Eventually, inspired by a line of poetry, he managed to pull himself out. As conditions got worse, the soles of his feet were so frostbitten that they detached from the flesh beneath. Then, just hours before he reached home base, his rescue ship sailed away. Even though they were called back by radio, bad weather prevented their return for days. When Mawson was finally rescued, he had been trapped in Antarctica for almost 3 years. The story is told in David Roberts’s book Alone on the Ice: The Greatest Survival Story in the History of Exploration. In his own book, Home of the Blizzard, Mawson recounts the entire grueling experience, including surviving subfreezing winds on Cape Denison that averaged 60 miles per hour, with gusts that approached 200 miles per hour.
After recovered from this harrowing experience, Mawson later helped with the search to retrieve the frozen corpses and journals of the ill-fated Scott expedition. He served in the British Army in World War I, and then returned to Australia in 1919 to become a professor of geology at the University of Adelaide until his retirement in 1952. He spent most of his career mapping and studying geology in Australia, especially the Flinders Range in South Australia, now famous for its upper Proterozoic rocks and the earliest megascopic body fossils of the strange creatures known as the Ediacara fauna. Still, Antarctic exploration was in his blood, and in 1929–1931 he led the joint British Australian and New Zealand Antarctic Expedition, which resulted in the formation of the Australian Antarctic Territory. Mawson died at age 76 in 1958, not in some frozen wasteland, but quietly in his bed of old age. He is so famous and so venerated as one of Australia’s greatest explorers and scientists that he appeared on the Australian 100 dollar bill, and his name is on many landmarks in both Australia and Antarctica.
It was lucky that Mawson was so familiar with ice sheets and glacial deposits through his years of experience on the Antarctic continent. In the upper Precambrian rocks of the Flinders Ranges and elsewhere in South Australia, he found thick deposits of what is known as glacial till (figure 16.2). Till is an unsorted, unstratified mass of boulders, gravels, sand, and mud that is randomly dumped by the snout of the glacier as it melts in one place. These sediments are very distinctive, and almost nothing else produces anything like them, so ancient glaciation can be recognized in the rock record. However, many geologists prefer to use the term “diamictite” (meaning “thoroughly mixed” in Greek) or sometimes the word “tilloid” as a noncommittal way of describing a rock with this texture, without directly implying that is it glacial. Following work by Oskar Kulling in 1934 and by Walter Howchin, Mawson was eventually convinced that this was evidence of a global glacial event in the late Proterozoic, because the ancient glacial deposits in Australia were not far from the modern equator. But by the late 1950s and early 1960s, geologists began to dismiss his argument, because plate tectonics had shown that Australia and other continents had moved long distances over time. Conceivably, the Australian Precambrian glacial beds could be from a time when the continent was closer to the South Pole. Ironically, we now know that Australia was right on the equator at that time, even more tropical than Mawson thought, so his evidence was stronger than anyone realized.
Figure 16.2
Photograph of the glacial tills of the late Precambrian Elatina diamictite of Australia. (Courtesy of Paul Hoffman)
THE ICEBED-LIMESTONE SANDWICH
But the idea of a global late Precambrian glaciation wouldn’t die. In 1964 Cambridge geologist W. Brian Harland (1917–2003) published a famous paper that showed that tropical upper Precambrian glacial deposits were not restricted to Australia. Like Mawson, Harland was also familiar with glaciers and ice sheets firsthand, since over his lifetime he spent a lot of time in the Arctic. He established the Cambridge Arctic Shelf Program. He was part of an amazing 43 polar field seasons (of which he led 29) from 1938 to the 1960s, mapping the geology of the archipelago of Svalbard (= Spitzbergen), a group of islands between Norway and Greenland. There he saw not only the deposits of recently melted glaciers, but also upper Precambrian glacial deposits in abundance, not only on Svalbard but also in Greenland and Norway.
Harland provided more than just evidence of glaciers. He pointed out that many of the upper Precambrian ice deposits were sandwiched between layers of limestone. This is much more surprising, because today limestones are only formed in warm tropical or subtropical shallow-marine settings like the Bahamas, Florida, Yucatan, the Persian Gulf, and the South Pacific. If this icebed-limestone sandwich had been formed by modern processes, then the glacial deposit surrounded by limestones had to form in the tropics and at sea level. Today there are a few places where tropical glaciers are known, such as the top of Mount Kilimanjaro in Kenya and the Peruvian Andes, but those glaciers are in high mountains. It seems impossible to imagine sea-level tropical glaciers, but Harland’s evidence was inescapable. If the tropics were glaciated, so were the poles, and so was the whole planet in the late Precambrian.
In addition, Harland backed up his conclusions with a new line of evidence: paleomagnetism. He was one of the first to measure the ancient magnetic directions frozen in the rocks since their formation, which can tell us at what latitude a given rock unit was formed. The paleomagnetic directions for all these Svalbard, Greenland, and Norwegian rocks were tropical or subtropical in the Precambrian, so the limestone-till-limestone sandwich was not some fluke. The paleomagnetic data from Australia was not very good then, but later analyses showed that Mawson’s Precambrian icebed-limestone sandwiches were right on the equator. Clearly, something weird was going on if there was indisputable evidence of tropical sea-level ice at that time.
In the 1960s and 1970s, many geologists were still not sure what to make of Mawson’s and Harland’s data and arguments, since it seemed inconceivable that the earth could have frozen all the way to the equator. Despite the evidence, they tended to dismiss this conclusion, because many were not confident that the paleomagnetic data were reliable. In addition, they could visualize less extreme scenarios where individual local regions might shift from limestone to glacial beds without freezing the whole world. The greatest difficulty of all, however, was imagining how the earth could freeze so completely. How could it flip so rapidly from a warm tropical limestone world to a tropical glaciated world, and back from glacial deposits to tropical limestones?
The answer to this came from a surprising direction: climate modeling. In 1969, Russian geophysicist Mikhail Budyko of the Leningrad Geophysical Observatory published a paper that showed how easy it is for a planet to freeze over once the ice sheets start to grow. He pointed to a well-known climate effect known as the albedo feedback loop. “Albedo” is just a fancy word for describing the reflectivity of a surface. As you know if you’ve ever spent time skiing or snowboarding, snow or an ice sheet has high albedo, since it reflects most of the sunlight that hits it. That’s why you need good dark goggles that reduce glare with tinted lenses when you spend time on the ice. By contrast, dark surfaces (like forests or the open ocean) absorb a lot more sunlight and reflect very little.
The albedo feedback system is very sensitive to small changes, which can transform it from frozen to ice-free and back to frozen very quickly. For example, let’s say the earth’s surface is covered by ice, so it has a high albedo and reflects most of the sun’s energy back. But the planet begins to warm slightly, and that ice sheet melts back a bit, exposing dark land and water. This absorbs more sunlight and generates heat, which melts the ice even further. Back and forth these two processes go in a feedback loop that eventually melts the ice in a very short time. Now let’s imagine this dark land and ocean surface has a few really cold winters and the reflective snow and ice layer lasts a bit longer. The increased ice cover reflects more energy back out to space, and the land gets colder, so even more ice sticks around the next few winters, and the ice sheet expands. Before you know it, the entire system has switched back into a complete ice age.
Although scientists knew albedo was a key feature of the polar regions and explained why they were so sensitive to small changes in global temperature, Budyko went further. In what he dubbed an “ice catastrophe,” he showed that if you had even a small ice sheet in the subtropical or tropical latitudes to start, the albedo feedback loop would kick into high gear, and the entire planet could freeze over rapidly. The only dilemma with this model was how to thaw the planet once it is completely frozen and has such a high albedo that most of its energy is reflected back to space. A completely frozen reflective iceball is a dead end, and the warming part of the feedback loop cannot rescue it.
The solution was first suggested in a 1981 paper by James Walker, Paul Hays, and James Kasting. They were focused mostly on the way in which weathering of silicate minerals in soils in the landscape can absorb carbon dioxide, but in the last paragraphs of the paper, they talked about Budyko’s models and how an ice cap would shut down the weathering mechanism and lead to Budyko’s “ice catastrophe.” In a brief sentence in the last paragraphs, they suggested another possible mechanism: volcanoes. The earth is unlike any other frozen planet in space (such as Mars and many others that have been found) in that we have an active crust with plate tectonics that powers lots of volcanoes. Volcanic eruptions release lots of gases, especially greenhouse gases like carbon dioxide, water, methane, and sulfur dioxide. If the earth were indeed completely frozen, the volcanic gases would eventually build up and warm the planet through the greenhouse effect, so the ice would finally begin to melt. And once enough dark surface had been exposed, the albedo feedback loop could kick into high gear and quickly melt from a frozen planet to an ice-free subtropical planet with limestones in the tropics.
The idea was in print and discussed by the few people working on the late Precambrian ice deposits, but still not widely known. This all changed with a legendary paper by Joe Kirschvink, now the Nico and Marilyn Van Wingen Professor of Geobiology at Caltech. Joe is one of the most brilliant people I have ever met, and he has more great ideas in a single week than most people have in a lifetime. He is one of the world’s best paleomagnetists, plus he does research on all sorts of problems on the boundary between geology and biology, from magnetic bacteria and butterflies and human biomagnetism, magnetofossils, biomineralization, to innovative ideas about the Cambrian explosion, to climate change and geochemical modeling, to polar wander and reconstructing ancient continental positions. In addition, Joe designs, builds, and maintains his own lab equipment and even writes his own computer programs. Plus, he is an outstanding, provocative, mind-expanding teacher, who challenges his brilliant students at Caltech to push the boundaries. He has won the Feynman Prize for teaching at Caltech and the William Gilbert Award in Paleomagnetism from the American Geophysical Union and has even had an asteroid named after him.
In 1989 Joe put his mind to the problem of how the snowball earth escapes being totally frozen over, and revived to the solution proposed by Walker and Kasting. They were all part of the PPRG (Precambrian Paleobiology Research Group), organized by my good friend Bill Schopf of the University of California–Los Angeles. They held a big PPRG meeting in 1989, where Joe not only revived the volcanic solution to a frozen-over earth, but pointed to evidence from Mawson’s Elatina Formation in Australia that it had actually occurred (see figure 16.2). Most importantly, he coined the phrase “snowball earth,” which was catchy and memorable and shifted the focus from soil weathering to volcanoes and a frozen planet.
Joe has one of the world’s best paleomagnetic labs at his disposal, so he did new analyses of the ancient latitudes of these Proterozoic icebed-limestone sandwiches and showed that many of them (especially Mawson’s Elatina sequence in Australia) were tropical or subtropical. Then he wrote up these ideas in a short paper tucked away in a huge expensive symposium volume from the 1989 PPRG meeting about Precambrian life. After many delays, it was eventually published in 1992 (and it was so expensive that few people owned it or read it), and Joe went on to other problems. Most people would have published such a groundbreaking paper in Nature or Science, but Joe doesn’t need the glory. He has so many great ideas all the time that he doesn’t need to spend all his time promoting each one for long. The idea of a snowball earth was formally named and proposed, with a clear mechanism for how to make it work. Other geologists soon caught on.
THE SNOWBALL BEGINS TO GROW
At Harvard University, geologist Paul Hoffman ran into Kirschvink at the International Geological Congress in Washington, D.C., in 1989 and learned about his snowball earth idea. (I was at that meeting too, but I was focused on other research problems.) When Hoffman began to work upper Precambrian glacial deposits in Namibia in 1993, he realized the importance of the snowball earth hypothesis, and eventually in 1997 he began to really promote it as the Namibian glacial deposits began to dominate his research.
Hoffman is a tall, gaunt, bearded, athletic field geologist who prefers to spend most of his time trekking the Canadian Arctic or the Namibian or Australian desert, looking for outcrops. He’s a serious cross-country runner, so he loves hiking. He was familiar with a variety of examples of icebed-limestone sandwiches in the Proterozoic of Canada, having spent much of his career mapping the protocontinents that made up the Archean precursors of Canada. These were later assembled into the Proterozoic core that was the nucleus around which North America grew. Hoffman recruited a number of brilliant colleagues with skills in geochemistry (such as Dan Schrag), and soon they set the geological community on fire with talk about the snowball earth.
Hoffman, Schrag, and other geologists analyzed some beautifully exposed icebed-limestone sandwiches, such as those in the deserts of modern Namibia in southwestern Africa (figure 16.3). There, the limestones on top of the glacial till are particularly thick and well developed, and they showed some peculiar geochemical and mineralogical characteristics. Hoffman and Schrag suggested that these “cap carbonates” that sit on top of the glacial tills are products of inorganic precipitation of limestones once the ocean geochemistry, saturated with dissolved carbonate, was released from the grip of the ice. They are clearly not the normal kind of limestones formed today, which are precipitated by organic activity. Modern limestones are made largely by the shells of corals, molluscs, and other marine creatures, as well as calcareous algae.
Figure 16.3
Geologists Dan Schrag (left) and Paul Hoffman (right) standing on the thick glacial Ghaub Formation full of glacial boulders in Otavi, Namibia, pointing to the boundary between the glacial bed and the overlying “cap carbonate.” (Courtesy of Paul Hoffman)
Another suggestive piece of evidence is the brief return of banded iron formations (BIFs) during the peak of the late Proterozoic snowball conditions. Kirschvink pointed out that this would make sense if the earth were frozen over, because it would shut down the oceans and make them anoxic and saturated with dissolved carbonate, so they would become highly acidic oceans (as is happening to our oceans now thanks to greenhouse gases). Without runoff from the sediment flowing down rivers (now completely frozen), the sulfate input to the oceans would be shut off, and this would result in abundant dissolved iron in these acidic, low-oxygen, low-sulfur oceans. Under these conditions, iron could accumulate on the bottom as it did between 3.7 and 1.7 Ga (see figure 14.5).
So the main snowball earth model runs like this: something causes the planet to begin to cool down dramatically until large ice sheets begin to form. At that time, without abundant and complex life (as we have now) to regulate the carbon cycle and keep pumping carbon dioxide into the atmosphere (chapter 6), the planet would begin to experience a runaway albedo feedback loop and eventually freeze over from the poles to the equator. Once it was a frozen snowball, it would be stuck in that state for millions of years, just like Mars is completely frozen now after having had liquid surface water with oceans and rivers in the past. Oceanic circulation would shut down, BIFs would accumulate on the anoxic seafloor, and lots of carbon would be frozen into little cages of ice known as methane hydrates in the seafloor sediments. If nothing else happened, the earth would have stayed frozen, and we would not be here.
Unlike Mars or any other planet, however, Earth has plate tectonics and volcanoes that over a long time erupted enough greenhouse gases to finally begin to warm the planet. Once that occurred, another runaway albedo feedback loop kicked in, and the ice melted rapidly, until it was almost all gone. The carbon caged in ice in methane hydrates on the seafloor released huge amounts of methane, further accelerating the global warming. The ocean geochemistry was now so rich in dissolved carbonate that huge deposits of calcite precipitated directly out of seawater to form cap carbonates. Finally, the planet was stable again with warm tropics and cooler poles.
Further research revealed that there were at least two or three separate events in the late Proterozoic, and one in the early Proterozoic (about 2 Ga), known as the Huronian (figure 16.4). It was based on the well-known Gowganda tillite on the shores of Lake Huron and showed that snowball earth conditions are not unique but can happen multiple times if the conditions are right.
Figure 16.4
Early Proterozoic (Huronian) glacial deposits from the Gowganda till, near Lake Huron in Canada. These represent an early snowball episode, about 2.1 Ga. (Courtesy of Wikimedia Commons)
Geologists, like all scientists, are naturally skeptical of new ideas, especially those that seem beyond the norm. Over the past 25 years, the snowball earth model has piled up an increasing volume of data, so most of the geological community has no choice but to accept the obvious conclusion that something like a snowball earth must have happened at least three or four times.
Still, there are dissenters. A number of geologists accept that there were equatorial sea-level glaciers in the late Proterozoic, but not that the entire tropical region had frozen over so the earth was a frozen snowball. They prefer a slightly less extreme idea, nicknamed the “slushball.” In this model, there was some glaciation on the equator (the data demand it), but much of the tropical region was cold but ice-free. They point to geological evidence of sediments that could only be formed in water, not ice. However, even Kirschvink’s original model allowed for some ice-free regions in the tropics, so this is not new. Also, many geologists see evidence that the snowball earth episodes had rapid fluctuations of glacial-interglacial cycles, as the most recent Ice Ages did (chapter 25), so this allows for both glacial sediments and sediments formed in running water and unfrozen oceans. Most importantly, the dating of the separate snowball events in the latest Proterozoic shows that they were globally synchronous and occurred from pole to equator at the same time. This favors a more extreme snowball earth rather than a slushball, because in slushball models the ice lines retreat when the carbon dioxide levels go up—but that is not what we see in the late Proterozoic snowball model.
One of the implications of the snowball earth model is that the late Proterozoic deep freeze apparently caused a big change in life. Before the snowball, we find abundant fossils of the spores of eukaryotic algae (known as acritarchs) in the marine sedimentary record. Then they apparently went through a mass extinction, because after the snowball earth ends, most of the diversity of acritarchs is gone. Instead, in the latest Proterozoic we see evidence of life returning to the earth with much more complex forms, including many that were multicellular. By the end of the Proterozoic, the first large multicellular animals appear all around the earth. Known as the Ediacara biota, they were first discovered by Reg Sprigg in Mawson’s Flinders Ranges of Australia, then described by paleontologist Martin Glaessner. Once these creatures flourished, the diversification of multicellular animal life (such as trilobites) launched into high gear. It has the misleading name “Cambrian explosion” but it’s more like a “Cambrian long fuse,” since it took 70 million years from the earliest Ediacaran creatures to the earliest trilobite.
Most fitting of all, one of the strange jellyfish-like impressions from the Ediacara fauna is named Mawsonites spriggi—after the two men whose mapping of the Flinders Ranges brought the Ediacara fauna to light.
FOR FURTHER READING
Hazen, Robert M. The Story of the Earth: The First 4.5 Billion Years from Stardust to Living Planet. New York: Penguin, 2013.
Macdougall, Doug. Frozen Earth: The Once and Future Story of Ice Ages. Berkeley: University of California Press, 2013.
Mawson, Douglas. Home of the Blizzard: A Heroic Tales of Antarctic Exploration and Survival. New York: Skyhorse, 2013.
Roberts, David. Alone on the Ice: The Greatest Survival Story in the History of Explorations. New York: Norton, 2014.
Schopf, J. William. Cradle of Life: The Discovery of Earth’s Earliest Fossils. Princeton, N.J.: Princeton University Press, 1999.
Schopf, J. W., and Cornelis Klein, eds. The Proterozoic Biosphere: A Multidisciplinary Study. Cambridge: Cambridge University Press, 1992.
Shaw, George H. Earth’s Early Atmosphere and Oceans, and the Origin of Life. Berlin: Springer, 2015.
Walker, Gabrielle. Snowball Earth: The Story of a Maverick Scientist and His Theory of Global Catastrophe That Spawned Life as We Know It. New York: Broadway, 2004.
Ward, Peter, and Joe Kirschvink. A New History of Life: The Radical New Discoveries About the Origin and Evolution of Life on Earth. New York: Bloomsbury, 2015.
