What Arctic Fossils Reveal About Our Future Climate

As Earth warms the Arctic faces troubling consequences. To predict that future, scientists are turning to fossils of a warmer past.

By Kendall Powell|Thursday, February 26, 2015
During the mid-Pliocene period the high Arctic was balmier, with boreal-like forests that might have been home to camels, as in this artist's rendering.
Julius Csotonyi/Canadian Museum of Nature

On a warm Thursday in July, Natalia Rybczynski and I paddle our kayaks up the Gatineau River near Ottawa. Fluffy-topped, flat-bottomed clouds speckle the sky, and blooming cornflowers poke through old railroad tracks leading to a dock. We glide past a long-abandoned beaver lodge, its sticks cut at a distinctive angle and spilling helter-skelter down the riverbank into the bourbon-colored water.

Rybczynski (pronounced rib-CHIN-ski) is a big fan of beavers, Canada’s official wildlife mascot; she’s conducted research showing they gnaw sticks with just one front tooth at a time. But our 2013 outing to the river is a consolation trip. Normally she spends her Julys at a beaver pond site that’s millions of years old and 800 miles north of the Arctic Circle, looking for clues about past global warming.

But this summer, Rybczynski, a paleobiologist, is confined to her home base, the Canadian Museum of Nature in Ottawa, as she recovers from a cross-country skiing accident that left her with a concussion. She’d much rather be wrapped in Gore-Tex, wavy brunette hair in a bandana, her dirt-ringed fingernails turning purple on the cold, dry tundra of Ellesmere Island.

Researchers hunt for clues to a 3-million-year-old mystery on Ellesmere Island. Left to right: Travis Mitchell, Marisa Gilbert, Natalia Rybczynski and John Gosse.
Martin Lipman

Today, Ellesmere, which lies next to Greenland on the eastern edge of Canada’s Arctic Archipelago, supports only ankle-high tufts of cotton grass and mossy ground cover; the nearest tree is almost 1,200 miles south. But Rybczynski and her colleagues have unearthed evidence of a balmier Arctic from a slice of time referred to as the mid-Pliocene warm period, roughly 3 million to 3.3 million years ago. The island’s treasure trove of fossils, preserved in permafrost, suggests the area was once an ancient boreal-like forest of larch, cedar and birch grazed by miniature beavers, three-toed horses and black bear ancestors.

The fossils act as a transcript that can be parsed for indicators not only of past climate, but also the climate of the future, making the team’s work as important to climate modelers as it is to paleobiologists.

Monitoring data show that the Arctic has already heated up more dramatically than the rest of the planet. While Earth’s landmass has warmed by about 1 degree Celsius (about 2 degrees Fahrenheit) over the past century, on average, land temperatures in the Arctic have risen almost 2 C (3.6 F). But to date, when climate modelers try to project future warming for the Arctic, the numbers are lower than expected; for reasons not yet fully understood, they don’t reflect the region’s accelerating warming. That means current climate models for the Arctic cannot accurately project the region’s future. If projections are too shaky to tell us what to expect as Earth warms up, the only alternative, Rybczynski points out, is “waiting 100 years to see what happens.”

Helpfully, during the mid-Pliocene warm spell, Earth had the same concentrations of carbon dioxide in the atmosphere and the same average 2 to 3 C (3.6 to 5.4 F) rise in global temperatures we’re headed toward. With its uncanny resemblance to our impending new normal, the mid-Pliocene warm period offers an invaluable proxy for modelers charting future warming, says James White, director of the Institute of Arctic and Alpine Research in Boulder, Colo. “Sometimes the Earth hands you a favor, and this is one.”

While today’s carbon dioxide levels — approaching a global average of about 400 parts per million — are largely due to people burning fossil fuels, the physical factors that directed the climate system 3 million years ago behave the same way today. The Ellesmere fossils are pieces of “ground truth” that can plug the holes in simulations of future climate change. The more we know about what changes to expect — and how fast they’ll come — the better we can prepare.

“We’re trying to understand how what we’re doing to the Earth’s atmosphere and oceans will play out in the future,” says Bette Otto-Bliesner, who runs a full-complexity climate model — and its 1.5 million lines of code — through a supercomputer named Yellowstone at the National Center for Atmospheric Research in Boulder. “How much we trust our model depends on how well we can reproduce the climate of yesterday. In this case, yesterday is the Pliocene.”

Rybczynski and her team have their work cut out for them. Capturing the full picture of the mid-Pliocene’s doppelganger climate requires reconstructing the whole Arctic paleoenvironment from sparse bits of remains. That means cataloging everything from the creatures and vegetation that co-existed to the rainfall and forest fire patterns. It means tracking landscape shifts and the swirl of ocean currents, and conjuring the ghost of Pliocene sea ice.

Piecing together a climate from a smattering of fossils is the ultimate forensic work. But this case, Rybczynski knows, is critically important to solve — and time is not on her side.

Natalia Rybczynski awaits a radio call at the Beaver Pond site.
Martin Lipman

While the worst consequences of climate change are still to come, evidence of amplified warming already abounds in the Arctic. At a hamlet on the southern end of Ellesmere called Grise Fiord, whose Inuit name means “the place that never thaws out,” the Inuit have watched the sea ice that supports their traditional seal, polar bear and whale hunting decrease every year. The Intergovernmental Panel on Climate Change has warned it’s likely that the Arctic will be sea ice-free in summertime before 2050.

“The changes to the Arctic represent the single most important natural history event of our lifetime,” says Meg Beckel, CEO of the Canadian Museum of Nature, which helps support the team’s $80,000 polar expeditions. “It’s dramatic. What will it look like 100 years from now?”

Reanimating the Past

Rybczynski’s love affair with Arctic fossils began when she was a high school student in the late 1980s, learning about paleobiology from her mentor, C. Richard “Dick” Harington. But it wasn’t until she graduated from college in 1994 that she joined Harington — the museum’s curator of Quaternary zoology at the time — on a research expedition to Ellesmere Island. Two years earlier, he made a startling discovery at the island’s Beaver Pond site. Geologist John Fyles, who discovered the site, had shown Harington beaver-cut sticks from Ellesmere — a clear indication that the spot was once home to something other than polar bears — and Harington wanted to see the site for himself.

On that first trip, on an outcrop a few hundred yards from the mouth of Strathcona Fiord, he immediately picked up three bones lying on the surface.

One was from a small beaver and another from a bear. The abundant beaver skeletons were Dipoides, a species two-thirds the size of modern beavers that had been found at sites in China and Idaho dated to the Pliocene. The site gave up more beaver-chewed sticks and saplings. That, plus the diversity of other wildlife found — frogs, fish, deerlets, hares — led Harington to conclude that around 3 million to 4 million years ago, the site was a beaver-dammed watering hole.

At the time, it was a rather radical idea. “People see the Arctic as an ancient landscape that’s been that way forever,” says Rybczynski. But Harington was suggesting “right up until the last ice age, you had a forest and this great diversity of mammals, and a frog and fish and everything!” The fossil evidence was hard to ignore. It seemed to hail from the rough stretch of time that included the mid-Pliocene warming event. But they weren’t entirely sure.

By the early 2000s, Rybczynski was at Duke University, studying beaver evolution for her dissertation. But her mind kept returning to the high Arctic. Were she and Harington reading those bones properly? Were they really from the mid-Pliocene? A forest, she knew, would require more than a couple of degrees of warmth to thrive. Just how high did the thermostat rise?

During much of the past 5 million years, Earth’s climate was in a cooling trend (estimated here from studies of ancient marine fossils). But during the mid-Pliocene, the planet warmed by 2 to 3 degrees Celsius — the same rise we’re headed toward.

To figure out those answers, Rybczynski turned to Ash Ballantyne, another grad student at Duke, who was using oxygen isotopes from tree cores to determine historical temperatures. Meeting him at his lab bench one day, she handed him a small disc of larch tree trunk with about 20 growth rings. “Cool specimen,” Ballantyne recalls thinking, but not remarkable. “Then she told me it was maybe 3.5 million years old and asked, could I possibly extract some climate information out of this?”

Judging by the wood’s apparent freshness, Ballantyne thought so. It still had bark on it. The natural deep-freeze of Ellesmere’s permafrost had mummified the wood. That meant Ballantyne could isolate cellulose, a structural molecule that keeps a record of temperatures in the plant’s lifetime. The ratio of two oxygen isotopes in the larch’s cellulose would give Ballantyne a temperature range.

He calculated that the Beaver Pond larch thrived at a yearly average of minus 5.5 C (22 F), about 14 degrees warmer than today’s average. “It was warmer than previous estimates [for the mid-Pliocene Arctic], but it seemed to make more sense,” says Ballantyne, now a climate scientist at University of Montana in Missoula. For a productive forest to grow, Ballantyne explains, temperatures have to remain above freezing for half the year. “We intuitively knew this,” he says, and the larch data confirmed it.

Rybczynski and Ballantyne realized the larch might have lived through a period of increased Arctic warmth similar to the one we’re heading into, and it could also hold clues about what else was happening on the globe — if the larch really was from the mid-Pliocene. Rybczynski realized the team needed stronger evidence that the sites were indeed Pliocene, including any more fragments the permafrost might give up.

The research team’s 2008 camp at the Fyles Leaf Bed site near Strathcona Fiord on Ellesmere Island, in the Canadian Arctic.
Martin Lipman

A Shocker in the Shards

In 2003, Rybczynski returned to Ottawa, essentially taking over for Harington at the museum when he semi-retired. With support from the museum, her team revisited Ellesmere in 2006, working in an area just 6 miles south of the Beaver Pond site, called the Fyles Leaf Bed.

The three-person team crouched on the steep ridge, scouring layers of sandy gravel and leaf litter. On the last day, following her own motto — “Pick up everything!” — Rybczynski plucked a fragment from the dirt that looked like either bone or mummified wood. She wrapped it in toilet paper and hiked it back to camp. Sitting at the kitchen tent table, she confirmed the shard was, indeed, bone.

Encouraged by that find, the team returned two summers later and turned up a few more slivers — enough to start piecing them together in the kitchen tent like the family jigsaw puzzle. Rybczynski could tell she had the entire thickness of the bone, from the outer surface to the hollow. That anatomical measurement, cortical bone thickness, scales with overall body size.

Whatever this thing was, it was huge.

When Rybczynski returned to the lab and sawed a tiny slice off the end of one shard, she caught the faintest unmistakable whiff of singed flesh. “Ooh, that’s collagen,” she recalls thinking. “That’s really interesting.”

Like DNA, collagen holds a sequence that might help reveal an animal’s identity. Rybczynski hoped she could place another creature among the larch needles and cones on the forest floor. “This had the possibility of being different than anything we knew [from the Beaver Pond site], even bigger than the bear,” she says. “Of course, I was kind of curious.”

The following year, in 2009, Rybczynski met just the right person to unlock the answer. Michael Buckley had perfected a new collagen fingerprinting technique to identify fossils. Rybczynski asked if it would work on her mystery beast. Collagen probably could survive in the high Arctic for a few million years, he said, but no one had successfully tested anything that old.

Rybczynski sent samples to Buckley’s University of Manchester lab in the U.K. When he compared the collagen readout with 32 species he’d already sequenced, Buckley saw it aligned most closely with a group not typically associated with the Arctic.

He emailed Rybczynski: “It looks like a camelid. Is that good?”

“Good? It is incredible!” came her reply.

Left, fossilized camelid bone shards. Right, the shards pieced together formed part of a single camel tibia bone 30 percent larger than a modern camel tibia.
Left: Martin Lipman; Right: Alex Tirabasso/Canadian Museum of Nature, adapted by permission from Macmillan Publishers LTD [Nature Communications], (DOI: 10.1038/ncomms2516) 2013

Rybczynski knew this family, which includes camels, llamas and alpacas, had evolved in North America before spreading to South America, Asia and eventually Africa. But this animal lived 745 miles farther north than any of its previously found relatives. They might have discovered the first forest-roaming camel.

To be certain, in 2010 she headed back to the Fyles Leaf Bed once more in search of more bone. This time, the expedition team recovered more shards — bringing the total to 30 — which yielded some anatomical clues. In analyzing the shards, Rybczynski realized they made up a single tibia from a giant camel about 30 percent larger than its modern cousins.

The 9-foot-tall camel might have sideways-chewed birch leaves, its splayed feet lumbering over snow in half a year of darkness and half a year of midnight sun. It’s an odd image for most of us, but to Rybczynski, it’s more evidence that in the past the Arctic was warmer and tree-filled and had more biodiversity.

“If you were casually walking through” that past Arctic, Rybczynski says, “it would have the feel of a boreal forest with a few differences — including a camel.”

A Pinpoint in Time

Rybczynski had used every scrap of inference and cutting-edge technology to pull the camel’s identity out of those sad-looking shards. Yet she still couldn’t be sure it was from the mid-Pliocene. Although the Ellesmere fossils matched other specimens from the Pliocene Epoch, it’s a span of about 3 million years. Without a more precise date, it would be impossible to construct a meaningful climate story from the fossils.

A region’s climate rests on many variables: land and sea temperatures, the shape of the landmass, how ocean currents mix globally, even the trajectory of Earth’s orbit. Being off by even a million years — a blink in geologic time — could dramatically change those factors. The team needed a narrower window.

To get that time stamp, Rybczynski sought the help of geologist John Gosse at Dalhousie University in Nova Scotia. He had refined a dating technique that could reliably clock sand as old as 8 million years, and she hoped he could extract some exactness from the Ellesmere sands. But Gosse had his doubts about the site’s age.

“Skeptical? I definitely was,” he recalls. In his mind, the site snuggled too closely to glacial deposits from the last ice age that lay atop the sloping Martian-esque hills. That would make it merely 35,000 years old. And, he says, the site was sandy, “like your favorite beach” with none of the compaction typical of sites millions of years old.

Geologist John Gosse collects sand samples from the lower levels of the Fyles Leaf Bed site on Ellesmere Island.
Martin Lipman

Gosse carefully tunneled into the hillside below the camel bone find and extricated a 2-kilogram (4.5-pound) chunk of sand. When he got it to his laboratory, Gosse isolated the quartz from the sample and then chemically squeezed out traces of aluminum-26 and beryllium-10 isotopes. The aluminum decays twice as fast as the beryllium; the longer sands are buried, the lower theirratio will be.

The samples’ ratios told him the camel roamed 3.8 million years ago and the beavers set up their dam 3.4 million years ago, give or take half a million years — age ranges accurate enough to place them in the middle of the mid-Pliocene warm period.

“The age was right on what Dick Harington had predicted,” says Gosse. “I was extremely relieved to know we had an independent way of verifying it. I’m glad I didn’t bet them.”

At the same time, Ballantyne and Rybczynski gathered collaborators who knew multiple ways to nail down past temperatures, calculating proxies from plant and soil bacteria fossils. Three distinct data sets pointed to the same number: an average yearly temperature 34 F warmer than today’s Arctic.

Such an average would have extended the above-freezing growing season from May to September. The Pliocene animals, plants and microbes all told the same story: They lived in a boreal-like forest ecosystem not quite like any that exists today. (The nearest modern example would be the Eastern Siberian taiga, dominated by larch and home to elk, weasels and wolverines.)

While average global temperatures in the mid-Pliocene rose only 3.6 to 5.4 F, the Arctic was a totally different world. “So the question is, what was amplifying temperatures in the Arctic?” Ballantyne asks.

Ice-Albedo Effect
Ice reflects solar energy; open water absorbs it. Less ice means warmer oceans, hindering ice formation.
Alison Mackey/Discover

His best guess is that sea ice, or rather the lack of it, played a big role, thanks to something called the albedo effect. Shining white sea ice reflects most of the sun’s energy back into space. Where it’s absent, the dark ocean waters absorb 90 percent of the energy coming in, heating it up. That warmth ultimately limits the reforming of sea ice and also has a warming influence on nearby land.

Ballantyne’s next challenge is to sleuth out the mechanisms that accelerated Pliocene sea ice melting. It won’t be easy, Ballantyne says. Past sea ice tends to leave even fewer traces than giant camels. But an intriguing new study suggests he’s on the right track. In the summer of 2013, Ballantyne and his former adviser, White at the Institute of Arctic and Alpine Research, ran a “what if” experiment.

“Well, hell, let’s just take the ice away and see what happens,” White recalls the team thinking. They ran a Pliocene climate model with all ice removed from the Arctic year-round. Unlike previous Pliocene models, this “no ice” version returned temperatures 18 to 27 F warmer than today’s average annual temperatures for the Canadian Arctic and Greenland, coming closer to what the historical data pulled from the ground said.

The researchers must resolve the mismatch between the Pliocene models and reality. “It’s a key question for the future,” says White, “because we are headed for a Pliocene climate.”

Arctic Sea Ice Minimum Cover, 2014 vs. 1981-2010
This satellite image from September 2014 (Canada appears in the foreground) shows Arctic sea ice at its annual minimum of about 2 million square miles — about 480,000 square miles less than the 1981-2010 average (outlined in red).
NASA/Goddard Scientific Visualization Studio

Turning Back to the Future

While Rybczynski recuperated in Ottawa in 2013, Gosse and the expedition team continued the hunt for puzzle pieces, this time traveling to westernmost Banks Island to search for Pliocene dwellers and date sites there.

The team brought back enough bones to fill two cupped hands and several buckets of sands for dating. They think one bone might be from another camel, but it’s not yet clear whether it dates to the Pliocene or the more recent ice age. These hard-won bits may not seem impressive to an outside observer. But, as Gosse explains, each one provides a data point for a condition happening locally that can be tied to the global record.

Once his team places all the sites across the high Arctic in time, then all the rich data — the field notes, the fossils, the lists of insects and vegetation — can help climate modelers fix their big problem. They’ll have a detailed picture of the mid-Pliocene paleoclimate. And then, they can solve for X to find the climate conditions that created that warmth, much like the no-ice experiment.

“We’re constantly developing the model to make it better and better,” says Otto-Bliesner. She and Ballantyne next want to include how ash from forest fires affected Pliocene climate. Rybczynski’s sites could hold the answer.

She’s been itching to return to the Fyles Leaf Bed, where the charcoal-black layers of leaf litter represent annual cycles spanning thousands of years. Hidden among them is the tale of how vegetation and wildfire seasons changed with temperature.

Time seems to be speeding up for Rybczynski and the others trying to solve this climate puzzle. On May 9, 2013, the Mauna Loa observatory in Hawaii recorded carbon dioxide levels climbing above 400 ppm for the first time. We are steadily marching toward a global yearly average above that threshold — and a Pliocene climate. The future is starting to look a lot like the past; only this time, 7 billion people will feel the heat.

And that includes Rybczynski. Thawing permafrost in the Arctic both dismays and delights her. Slumping hillsides leave slashes of broken-open earth, revealing new excavation sites. But the thaw could also spell the end for Ellesmere’s exquisitely preserved fossils.

“That is the double-edged sword. We might lose the Beaver Pond site, but we find these new sites,” says Rybczynski. “It definitely gives me a sense of urgency — getting this work done and getting it right.”

[This article originally appeared in print as "Cold Case."]

Comment on this article