The ice would melt, expose rock, cool things down, and repeat.
Hundreds of millions of years ago, Earth endured what appeared to be an impossibly long frozen silence — yet life persisted, and the rocks never quite explained how. Harvard researchers have now proposed that the Sturtian glaciation was not a single unbroken catastrophe but a rhythm of freeze and thaw, each cycle driven by the ancient tension between volcanic breath and the slow hunger of weathering stone. In reframing one of Earth science's deepest mysteries, they offer not only an answer about our own planet's survival, but a new lens through which to ask whether life might endure similar ordeals on worlds far beyond our sun.
- For decades, a 56-million-year global ice age defied explanation — standard climate models could not reconcile how life, oxygen, or the ice itself could have lasted so long.
- The contradiction was stark: sedimentary rocks worldwide confirmed the glaciation happened, yet continuous freezing on that scale should have suffocated nearly all life on Earth.
- Harvard's Charlotte Minsky and her team modeled the Franklin Large Igneous Province, a colossal Canadian volcanic system, and discovered a planetary tug-of-war — volcanic CO₂ warmed the planet, melting ice exposed fresh basalt, basalt weathering pulled carbon back down, and the cycle reset.
- Their simulations revealed not one Snowball Earth but a series of million-year freeze-thaw pulses, each brief enough for photosynthetic life to replenish atmospheric oxygen during the warm intervals.
- Published in the Proceedings of the National Academy of Sciences, the findings reframe Earth's most extreme climate period — and open a framework for assessing whether life could survive repeated glaciation cycles on distant exoplanets.
Fifty-six million years of ice. The Sturtian glaciation, which gripped Earth during the Neoproterozoic Era roughly 720 to 635 million years ago, has haunted paleoclimatologists for decades. A planet nearly or entirely encased in snow and ice for that long should have been lethal — oxygen depleted, oceans starved of light, photosynthesis extinguished. Yet the geologic record insisted it happened. The contradiction between what rocks revealed and what climate models could explain stood as one of Earth science's most stubborn mysteries.
Charlotte Minsky and her colleagues at Harvard's School of Engineering and Applied Sciences found a way through it. They modeled the climate effects of the Franklin Large Igneous Province, a massive volcanic system in Canada that erupted more than 700 million years ago. What they identified was a kind of planetary tug-of-war: volcanic activity pumped carbon dioxide into the atmosphere, warming the planet and retreating the ice. But as ice melted, it exposed fresh basalt, which weathered and slowly drew carbon back out of the air. The planet cooled, ice advanced, and the cycle began again.
The result, their simulations showed, was not one continuous deep freeze but a series of shorter glaciation episodes — each lasting roughly a million years — punctuated by warmer intervals. This reframing dissolved the oxygen paradox that had plagued earlier models. In repeated million-year cycles rather than a single 56-million-year lockdown, photosynthetic life could recover between glaciations, keeping the atmosphere breathable. The survival challenge remained extreme, but it became biologically possible.
Published in the Proceedings of the National Academy of Sciences, the findings carry implications beyond Earth's ancient past. The mechanism the Harvard team identified — volcanic carbon release balanced against basalt weathering — could operate on any rocky, water-bearing planet with active volcanism. As astronomers catalog Earth-like worlds, understanding how life might endure repeated global glaciations becomes a question not just of history, but of cosmic habitability.
Fifty-six million years of ice. That's how long the Sturtian glaciation locked Earth in a frozen grip during the Neoproterozoic Era, roughly 720 to 635 million years ago. The planet was nearly or entirely covered in snow and ice—a condition so extreme that scientists have struggled for decades to explain how anything survived it, let alone how the ice age itself could have persisted for so long. Standard climate models simply couldn't account for it. Now researchers at Harvard have proposed a solution that rewrites our understanding of this ancient catastrophe: the Sturtian wasn't one continuous deep freeze, but rather a series of shorter freeze-thaw cycles, each one lasting only a million years or so.
The puzzle has haunted paleoclimatologists for good reason. A global glaciation lasting 56 million years should have been lethal to nearly all life. The ocean would have been starved of light and nutrients. The atmosphere would have been depleted of oxygen as volcanic gases reacted with the air and photosynthesis ground to a halt. Yet the geologic record—sedimentary rocks laid down across the world—shows clear evidence that this glaciation did occur. The contradiction between what the rocks tell us and what climate models predict has been one of Earth science's enduring mysteries.
Charlotte Minsky and her colleagues at Harvard's School of Engineering and Applied Sciences approached the problem by modeling the climate effects of the Franklin Large Igneous Province, a massive volcanic region in Canada that erupted more than 700 million years ago and remained active for roughly two million years. This wasn't a typical volcano. It was one of the largest magmatic episodes in Earth's history, pumping enormous quantities of carbon dioxide into the atmosphere. But here's where the team's insight becomes crucial: they recognized a kind of planetary tug-of-war. When volcanic activity released carbon, the atmosphere warmed and ice retreated. As ice melted, it exposed fresh basalt rock. That newly exposed basalt weathered away, slowly removing carbon from the air. As carbon declined, the planet cooled and ice advanced again. The cycle repeated.
Their simulations revealed something unexpected. Rather than a single, unbroken Snowball Earth, the Sturtian consisted of multiple episodes—short bursts of global glaciation followed by brief warm interludes, then back to ice. Each cycle lasted roughly a million years, not 56 million. The geologic record, when viewed through this lens, suddenly made sense. The sedimentary rocks showed evidence of glacial advance and retreat, not because the models were wrong, but because the researchers had been imagining the wrong scenario.
This reframing solves the oxygen problem that had plagued earlier explanations. In a continuous 56-million-year Snowball, the atmosphere would have lost its oxygen as volcanic gases consumed it and photosynthesis ceased in the frozen ocean. Life would have suffocated. But in a world of repeated shorter cycles, the atmosphere never had time to become fully depleted. Between each glaciation, during the warmer periods, photosynthetic organisms could replenish oxygen. The biosphere only had to endure million-year-scale ice ages, not a half-century-long freeze. That's a fundamentally different survival challenge—still extreme, but not impossible.
Minsky and her team published their findings in the Proceedings of the National Academy of Sciences, titled "Repeated snowball–hothouse cycles within the Neoproterozoic Sturtian glaciation." The work addresses not just how life persisted through Earth's most violent climate swings, but also how such cycles might occur on distant exoplanets. As astronomers discover more Earth-like worlds, understanding the conditions under which life can survive repeated global glaciations becomes increasingly relevant. The mechanism the Harvard team identified—the interplay between volcanic carbon release and basalt weathering—could operate on any rocky planet with active volcanism and liquid water. It's a reminder that Earth's ancient climate catastrophes aren't just historical curiosities. They're a key to understanding which worlds beyond our solar system might harbor life.
Notable Quotes
The biosphere would have only had to persist through shorter—million year timescale—Snowballs, allowing atmospheric oxygen to persist without being fully depleted.— Charlotte Minsky and colleagues, Harvard SEAS
This could help explain how aerobic life persisted through such an extreme interval.— Charlotte Minsky, lead author
The Hearth Conversation Another angle on the story
Why did scientists think the Sturtian was one long ice age instead of multiple cycles?
The geologic evidence was clear—rocks showed glaciation happened. But the climate models couldn't explain how a single 56-million-year freeze was possible. It seemed to violate the laws of planetary physics. So people accepted the evidence but couldn't explain the mechanism.
And the Franklin volcanic province was the missing piece?
Not exactly missing—it was known. But nobody had modeled what happens when you have that much volcanic carbon being released while simultaneously exposing fresh basalt that weathers and removes carbon. It's a feedback loop that naturally creates cycles.
So the ice would melt, expose rock, the rock would cool things down again, and the cycle repeats?
Exactly. And each cycle is short enough that life doesn't get completely suffocated. The atmosphere never fully loses its oxygen because there are warm periods in between where photosynthesis can recover.
That seems almost elegant—a planetary thermostat.
It is. And it's not unique to Earth. If this mechanism works here, it could work on exoplanets too. That's why the Harvard team thinks their model has implications beyond ancient history.
Does this change how we think about whether life could survive on other worlds?
It expands the possibilities. If a planet can cycle through repeated Snowballs rather than getting locked in one, it gives life more windows to persist. That makes habitability more plausible in scenarios we might have thought were hopeless.