Earth is missing a huge part of its crust. Now we may know why.

The Grand Canyon is a gigantic geological library, with rocky layers that tell much of the story of Earth’s history. Curiously though, a sizeable layer representing anywhere from 250 million years to 1.2 billion years is missing.

Known as the Great Unconformity, this massive temporal gap can be found not just in this famous crevasse, but in places all over the world. In one layer, you have the Cambrian period, which started roughly 540 million years ago and left behind sedimentary rocks packed with the fossils of complex, multicellular life. Directly below, you have fossil-free crystalline basement rock, which formed about a billion or more years ago.

So where did all the rock that belongs in between these time periods go? Using multiple lines of evidence, an international team of geoscientists reckons that the thief was Snowball Earth, a hypothesized time when much, if not all, of the planet was covered in ice.

According to the team, at intervals within those billion or so years, up to a third of Earth’s crust was sawn off by Snowball Earth’s roaming glaciers and their erosive capabilities. The resulting sediment was dumped into the slush-covered oceans, where it was then sucked into the mantle by subducting tectonic plates. (Here’s what will happen when Earth’s tectonic plates grind to a halt.)

Effectively, in many locations, Earth buried the evidence of about a fifth of its geological history, the team argues today in the Proceedings of the National Academy of Sciences. The notion is elegant but provocative, and the authors themselves predict that some geoscientists will express skepticism.

“I think, though, we have extraordinary evidence to support that extraordinary claim,” says study leader C. Brenhin Keller, a postdoctoral fellow at the Berkeley Geochronology Center.

Geochemical ghosts

Although its nuances, triggers, and shutdown mechanisms continue to be debated, the idea that Earth was a giant frigid “snowball” around 700 million years ago is being increasingly accepted by the scientific community. And similar to what we see in Antarctica today, many of Snowball Earth’s glaciers would have been powerful agents of erosion: The overlying pressure of the ice creates wet bases that can move sediment, despite the extremely low temperatures at the surface.

The Great Unconformity has in turn often been suggested as being an erosional feature, but some geologists balked at the idea of such a massive amount of Earth’s crust being so thoroughly wiped away.

Keller, however, found fresh clues hiding in ancient zircons. These resilient minerals lock in the geochemical conditions of their environment when they crystallize, and scientists can pick them apart billions of years later to find out what Earth was once like. (For instance, four-billion-year-old zircons are offering some clues to the origins of life.)

In particular, these zircons contain various radioactive isotopes that act as record-keepers. Uranium isotopes allow researchers to find out the crystals’ formation ages with remarkable precision. Others, like hafnium isotopes, reveal what was happening to the crust and mantle, as certain isotopes prefer one geological setting to another.

Using a cornucopia of zircons, Keller and his team carefully unspooled the geochemical evolution of Earth’s crust across 4.4 billion years of time. They saw that a colossal geochemical shift took place at the theorized beginning of Snowball Earth’s planetwide glaciations, which is only explicable if plenty of Earth’s crust was being recycled into new magma reservoirs.

Oxygen isotopes within these zircons indicated that the crust had gone through low-temperature hydrothermal changes, too. This meant that it was the top of the crust—in contact with water and ice—that was being shaved off and subducted, not the deeper stuff.

Altogether, this evidence suggests that a gargantuan erosional event happened at the surface. Although this erosion didn’t apply evenly across the world, it amounts to an average sediment layer 1.9 to 3.1 miles deep being swept away.

Serendipity in suspicious sediments

The geochemical evidence is powerful, but chance discussions at a recent scientific conference made each future coauthor realize that there’s even more to the story.

For one thing, “around 600 to 700 million years ago, Earth loses its craters,” notes study coauthor Bill Bottke, a planetary scientist and asteroid expert at the Southwest Research Institute in Boulder, Colorado. Some ancient craters still exist on stable continental cores named cratons, but they are few and far between. (Recently, a city-size impact crater was found under the ice in Greenland.)

The easy explanation for this mystery was also a ginormous erosional event, but until now, evidence for one was hard to come by. Unlike many other worlds, “Earth does a really good job at erasing the tracks of its past,” Bottke says. Fortunately, Keller’s geochemistry made it clear that Snowball Earth provides a natural explanation.

Then there’s the huge uptick in sedimentation rates at the start of the Cambrian. All the new sediment required plenty of space to fall into, something that would have only been possible if massive levels of erosion took place beforehand, says coauthor Thomas Gernon, an associate professor of earth science at the University of Southampton.

As the researchers point out, one problem with their data is that there is still a multimillion-year time gap between the predicted end of Snowball Earth and the start of the Cambrian. It’s not clear why the formation of new rock layers took so long to start after all that erosion stopped.

Although it’s likely down to a number of factors, one possibility is that Snowball Earth’s erosion was so significant that there wasn’t much topography left to erode when all was said and done. The planet simply needed to forge more land first, and that takes time.

The theory of everything

Understandably, not everything matches up perfectly, but the study’s narrative is “very plausible,” and its arguments are “pretty clever,” says Ian Fairchild, an emeritus professor of geosciences at the University of Birmingham who wasn’t involved in the research.

Bottke hopes the team is right, but either way, he’s happy that this paper will add to the debate over a prolific geological mystery.

“That conversation is what helps drive the science,” he says.

If validated, the implications of this story could be enormously significant. After all, this study points out that complex life first emerged when Snowball Earth’s monstrous mealtime came to an end. Glaciers would have carved out fjord-like shallow marine areas, which could have been havens for life as the planet warmed back up. This colossal crustal consumption may have also coincided with major geochemical and environmental changes that were potentially beneficial to biological evolution.

In essence, there’s a chance that the diversification of multicellular animals is a direct consequence of ancient glaciers obliterating the planet’s crust.