Scientists endured bitter winds to retrieve ancient ice from a blue ice field in the Allan Hills of Antarctica.

Yuzhen Yan

Record-shattering 2.7-million-year-old ice core reveals start of the ice ages

Scientists announced today that a core drilled in Antarctica has yielded 2.7-million-year-old ice, an astonishing find 1.7 million years older than the previous record-holder. Bubbles in the ice contain greenhouse gases from Earth’s atmosphere at a time when the planet’s cycles of glacial advance and retreat were just beginning, potentially offering clues to what triggered the ice ages. That information alone makes the value of the sample “incredible,” says David Shuster, a geochemist at the University of California, Berkeley, who is unaffiliated with the research. “This is the only sample of ancient Earth’s atmosphere that we have access to.”

Described at the Goldschmidt Conference in Paris by Yuzhen Yan, a graduate student at Princeton University, the ice revealed atmospheric carbon dioxide (CO2) levels that did not exceed 300 parts per million, well below today’s levels. Some models of ancient climate predict that such relatively low levels would be needed to tip Earth into a series of ice ages. But some proxies gleaned from the fossils of animals that lived in shallow oceans had indicated higher CO2 levels. If the new result holds up, says Yige Zhang, a paleoclimatologist at Texas A&M University in College Station, the proxies will need to be recalibrated. “We have some work to do.”

The discovery also points the way to finding even older ice, because it comes from a largely ignored “blue ice” area, where peculiar dynamics can preserve old layers. Although blue ice areas offer only a fragmentary view of the past, they may turn into prime hunting grounds for ancient ice, says Ed Brook, a geochemist on the discovery team at Oregon State University in Corvallis. “Ice that’s this old really makes people stand up and notice,” he says. “We’re just scratching the surface.”

Ice cores from Greenland and Antarctica are mainstays of modern climate science. Traditionally, scientists drill in places where ice layers accumulate year after year, undisturbed by glacial flows. The long layer cake records from deep sites in the center of Antarctica reveal how greenhouse gases have surged and ebbed across hundreds of thousands of years. But because heat from the bedrock below can melt the deepest, oldest ice, the approach has not yielded ice any older than 800,000 years, from a core drilled at Antarctica’s Dome C in 2004.

Deep freezes

Rather than drilling deep in Antarctica’s interior, scientists found a 2.7-million-year-old core in shallow “blue ice,” where glacial flows bring old ice up to the surface.

Credits: (Graphic) G. Grullón/Science; (Data) Quantarctica Database, Norwegian Polar Institute; H. Fischer et al. Clim. Past 9, 2489–2505 (2013)

The Princeton-led team went after ancient ice sitting far closer to the surface, in the Allan Hills, a wind-swept region of East Antarctica 200 kilometers from McMurdo Station that is famous for preserving ancient meteorites. In such blue ice areas—just 1% of the continent’s surface—the ice flows across rocky ridges, tipping the record on its side. Deep, old layers are driven up, while wind strips away snow and younger ice, revealing the lustrous blue of compressed ice below. But these contortions also confound the neat ordering of the annual layers—making it impossible to date the ice by counting them.

Michael Bender, a Princeton geochemist on the discovery team, solved the problem by finding a way to date chunks of ice directly from trace amounts of argon and potassium gases they contain. Although not as precise as other dating methods, Bender says, the technique can date ice to within 100,000 years or so.

In 2010, the team drilled their first hole at the Allan Hills, in a place where the ice was shallow and thought to be ascending a hill, with a chance of being stuck against bedrock. They drilled horizontally, toward the hill, in the hopes that the ice would get older as they drilled farther. They ran out of time after 128 meters, before they reached bedrock, but the unfinished core yielded some chunks of ice that were 1 million years old. It was the first sample older than 800,000 years, from a crucial time when glacial periods were switching from occurring every 40,000 years or so to every 100,000 years.

In 2015, the team returned to try again. The environment was harsh, with constant wind shearing their clutch of tents. “Cold is one thing,” says Princeton geochemist John Higgins, “but windy cold is just another beast.” Yet they were able to drill the remaining 20 meters to bedrock, and found the ice that, along with several other new cores, yielded the ancient samples.

It is unlikely that any traditional efforts will match the Allan Hills sample in age, says John Goodge, a geologist at the University of Minnesota in Duluth. Several teams are pursuing sites that may go back as far as 1.5 million years—but even that’s hopeful, says Goodge, a leader of a U.S. team planning to rapidly drill a deep ice site in the Antarctic interior. “In that sense, the Allan Hills ice core may stand on its own for some time,” he says.

Now, the Princeton team wants to go back to the blue ice and drill some more, Brook says—not only to fill in the climate cycles of the last 2.7 million years with a multitude of snapshots, but also to go even deeper in time, before the ice ages, when CO2 levels were higher. There’s evidence that Antarctica has hosted at least some ice for 30 million years. It’s plausible, Brook says, that the next attempt could come back with ice 5 million years old—a time when temperatures are thought to resemble what Earth is heading toward with human-driven warming.

Correction, 18 August, 11:58 a.m.: An earlier version of the story mischaracterized Michael Bender’s role on the discovery team.