It's a mystery that has puzzled geologists for years: big glacial stones show up in the middle of the fine-grained clay of sedimentary rocks. It's especially odd when they are associated with carbonate rocks like limestone, which are created in the tropics. "How could they be co-occurring?" asks Hooper professor of geology Paul Hoffman. "The glacier never went south of Cape Cod."
The answer may emerge from a new geological theory, nicknamed the "snowball earth" hypothesis, that Hoffman and his colleagues recently advanced in Science. The theory postulates the occurrence, 750 million years ago, of a class of phenomena that early computer models of climate had described 30 years ago. "It was theoretically possible, but nobody believed it had actually happened," Hoffman says. Now it appears that huge sections of our planet's history may have differed radically from what scientists have believed.
The new theory hinges on a variable called albedo, the reflectiveness of the earth's surface. The amount of heat energy coming in from the sun, minus the amount lost to space, controls the temperature of the planet's surface. If the earth were entirely white and shiny--a state of "high albedo"--that surface would reflect most of the solar energy back into space. Conversely, a dark surface--low albedo--absorbs more heat and reflects less, so the earth gets hotter. Land and water are dark, but ice and snow are reflective.
Now consider the inverse of the greenhouse effect--call it an icebox effect. If carbon dioxide in the atmosphere decreases, the climate gets colder and the polar ice caps grow larger. This raises albedo and means that the earth loses even more heat: the positive feedback cycle continues to cool the planet. If the sea's surface freezes over as far south as Boston (roughly 42 degrees north latitude), the albedo feedback becomes so powerful, says Hoffman, that there's no stopping it: "The earth will freeze over."
This is exactly what the scientists say happened 750 million years ago. From space, Earth would have looked like a giant snowball. "If the oceans freeze over with sea ice, you have no more hydrologic cycle of evaporation and precipitation, so there is no more rain or snow," Hoffman explains. Rivers cease to flow. Photosynthesis nearly stops. This raises a question: If you are in a runaway albedo situation, how do you get out?
One answer: underneath the frozen crust, plate tectonics would continue and the earth's internal heat would affect the icy surface above. As Hoffman depicts the scenario, volcanic eruptions, spewing out rocks, water, and gases such as carbon dioxide and hydrogen sulfide, return carbon--long buried as sediment--to the atmosphere. With no green plants to "fix" carbon dioxide, the gas accumulates in the atmosphere, reaching, after millions of years, a concentration 300 times greater than today's levels. This mega-greenhouse effect reverses the high albedo: ice melts, green water forms-- reducing albedo--and then evaporates, adding water vapor (itself a powerful greenhouse gas) to the atmosphere and reinforcing the warming cycle. "Melting will occur very fast, and the surface gets very warm," Hoffman says. "You might go to an ultra-greenhouse effect in as little as a thousand years."
The scientists propose that at least four cycles of alternating snowball and ultra-greenhouse conditions occurred between 750 million and 550 million years ago. Their theory helps account for some puzzling data from Namibia, where Hoffman and his colleagues have been examining the geologic record since 1993. It is common to find extremely high rates of organic carbon buried in sedimentary rocks that predate glacial periods, since organic matter extracts carbon dioxide from the air, reducing the greenhouse effect and hence cooling the climate. But, says Hoffman, "What's really surprising is seeing that during the glacial period and shortly afterward, you have a carbon isotopic composition that indicates no burial of organic matter at all--the removal of organic carbon is zero, indicating that no organic matter is being produced. And this persists for millions of years. Even at the time of the Cretaceous Tertiary extinction [when dinosaurs disappeared, 65 million years ago], the productivity of the ocean recovered in a few thousand years. Here, we're talking about a productivity collapse that's much more long-lasting--and is associated with glacial deposits in the tropics."
Going back to those volcanic eruptions, one of Hoffman's graduate students, Galen Halverson, calculated that the snowball glaciation in Namibia lasted for at least 10 million years, consistent with estimates by atmospheric scientists that it would take between 4 million and 30 million years to reverse the high albedo, assuming present-day rates of carbon-dioxide emission from volcanoes. The time scale matches well with the geologic record from the carbonate rocks of Namibia. "The snowball hypothesis seems to do the best job of explaining all classes of observations," Hoffman says. "It's a very radical idea, even for geologists--a more extreme and long-lasting event than anyone has postulated in the geologic record before. It's pretty staggering to think that if it did occur, we somehow missed it. The explanation may be that, if you freeze the earth, not much happens geologically. It's a phantom event."