Fifty-six million years is a long time to be frozen. It is longer than the entire Cenozoic, longer than the reign of the dinosaurs, longer, frankly, than most scientists thought physically possible for a global glaciation. And yet the geological record is unambiguous: roughly 717 million years ago, ice swallowed the Earth, and it did not fully let go for nearly 60 million years. The event is called the Sturtian glaciation, and for decades it has been quietly embarrassing climate science.
The problem is not that Snowball Earth happened. The evidence for that is solid enough. The problem is that every serious climate model ever built says it should have lasted maybe a few million years, not 56. Something in the standard picture is badly wrong, and nobody has convincingly said what.
Charlotte Minsky, a graduate student at Harvard’s John A. Paulson School of Engineering and Applied Sciences, has been sitting with this discrepancy. Working with Robin Wordsworth, the Gordon McKay Professor of Environmental Science and Engineering, and co-authors David T. Johnston and Andrew H. Knoll, she built a coupled model of the ancient climate and the global carbon cycle and ran it forward through the Cryogenian period, that strange interval between roughly 720 and 635 million years ago when the planet was doing things it has never done since. What came back from the simulations was not what the standard story predicts. Earth, the model suggests, never locked into a single unbroken freeze at all.
A Planetary Thermostat Gone Wrong
Here is what Minsky’s model proposes instead. Just before the Sturtian began, a colossal volcanic system in what is now northern Canada, the Franklin Large Igneous Province, erupted across vast swaths of land and left behind enormous fields of fresh basalt. Basalt, when exposed to air and rain, weathers chemically and as it does so pulls carbon dioxide out of the atmosphere. The Sturtian glaciation, in this picture, began not as a single catastrophic freeze but as the first in a long series of pulses: CO2 gets drawn down, temperatures plummet, ice covers the planet, weathering grinds to a halt. Then volcanoes and other geological processes slowly rebuild atmospheric CO2. Temperatures rise, ice retreats, fresh basalt is again exposed at the surface. Weathering kicks in again, CO2 falls again, and the whole thing freezes over once more.
A limit cycle. A planetary thermostat stuck not at one setting but oscillating between two extremes, hothouse and snowball, for tens of millions of years.
The elegance of this hypothesis is that it resolves not one paradox but several at once. The duration problem disappears: instead of asking how a single Snowball state persisted for 56 million years (a question standard models cannot answer), you simply ask how long the oscillation continued, and the answer follows naturally from the carbon cycle dynamics. The observed patterns in the sedimentary record from that period, which have always been a bit awkward to explain under the canonical model, fit too. And there is a third puzzle the new work addresses, perhaps the strangest one of all.
How Did Anything Survive?
Life was not absent from the planet during the Sturtian. Simple organisms, including early aerobic life that needed oxygen to function, were around before the glaciation began and were still around after it ended. Maintaining oxygen in the atmosphere through 56 million years of global freeze, with the photosynthetic plankton that produce most of it presumably buried under ice, has been a persistent headache for researchers. The limit cycle model offers at least a partial answer. Each time the planet swung back to a hothouse state, the ice retreated, photosynthesisers got sunlight again, and oxygen was replenished before the next freeze hit. “This could help explain how aerobic life persisted through such an extreme interval,” Minsky said.
The work, published in the Proceedings of the National Academy of Sciences, leans on computational modeling rather than new field data, and the researchers are careful about that. Box models of the carbon cycle are useful precisely because they are simple enough to run across geological time, but simplicity is also a limitation: they smooth over regional variation, local chemistry, and the full complexity of ancient ocean circulation. The sedimentary record from the Sturtian is fragmented and hard to read at the resolution needed to confirm or deny rapid cycling between states. Direct evidence of the oscillations Minsky’s model predicts, perhaps in the form of isotopic signatures from carbonate rocks laid down during Sturtian warm intervals, remains to be found.
There are competing hypotheses, too. The “Slushball” model, which proposes that equatorial oceans stayed open even during peak glaciation, has its own advocates, and the debate about exactly how frozen the planet got is not settled. What Minsky and Wordsworth are not claiming is that the limit cycle picture is definitive. What they are claiming is that it is consistent with what we know, that it resolves discrepancies the other models do not, and that it deserves serious attention.
What the study quietly changes is the framing. If the model holds up, the Sturtian glaciation was less an entombment than a kind of geological breathing: the planet repeatedly seizing up and then gasping back to warmth, driven by the chemistry of a single vast volcanic field that happened to erupt at exactly the wrong moment. It is a stranger story than the standard one, and in a way a more hopeful one. The Earth did not simply lock up for 56 million years with life clinging on somehow. It cycled. It recovered. It froze again.
That those early aerobic organisms made it through at all, surfing repeated swings between catastrophe and reprieve, now looks less like a mystery and more like something almost explicable. Almost.
Frequently Asked Questions
If Snowball Earth only lasted a few million years at a time, why do we call it one event?
The geological record doesn’t have fine enough resolution to distinguish individual freeze-thaw cycles within the Sturtian, so the whole 56-million-year period has traditionally been lumped together as a single glaciation. The new research suggests that label may be misleading: what we’re looking at could be a rapid series of global freezes and thaws, each lasting perhaps hundreds of thousands of years, stacked one on top of the other. Whether future work in the rock record can actually separate them out remains to be seen.
Why would a volcanic eruption cause the Earth to freeze rather than warm up?
The Franklin Large Igneous Province wasn’t pumping CO2 into the atmosphere so much as leaving behind a massive landscape of fresh basalt, a rock that hungrily absorbs CO2 from the air as it weathers. The eruption itself may have warmed things briefly, but the long aftermath, centuries and millennia of chemical weathering across an enormous surface area, was what pulled CO2 levels down far enough to trigger glaciation. It’s a counterintuitive chain of cause and effect that depends entirely on what happens after the lava cools.
Could something like this happen again today?
The conditions that set off the Sturtian were fairly specific: a vast volcanic province erupting at low latitudes, where warm rain accelerates basalt weathering, combined with atmospheric CO2 levels and continental configurations quite different from today’s. That said, the underlying carbon cycle dynamics the model describes are real and still operating. What the research underlines is that the climate system contains feedback loops capable of extreme, self-sustaining behaviour under the right circumstances, which is not an entirely comfortable thought.
How do we know early aerobic life existed before and after the Sturtian?
The fossil and geochemical record from the Neoproterozoic is incomplete but not silent. Biomarkers and isotopic signatures in rocks predating and postdating the Sturtian point to the presence of oxygen-dependent organisms on both sides of the glaciation. The puzzle has never been whether life survived, but how it managed to, given the conditions. The limit cycle model’s prediction of repeated warm intervals gives those organisms somewhere to be during the freeze.
ScienceBlog.com has no paywalls, no sponsored content, and no agenda beyond getting the science right. Every story here is written to inform, not to impress an advertiser or push a point of view.
Good science journalism takes time — reading the papers, checking the claims, finding researchers who can put findings in context. We do that work because we think it matters.
If you find this site useful, consider supporting it with a donation. Even a few dollars a month helps keep the coverage independent and free for everyone.