Uncovering the Secrets of Mammoth Island

One day thousands of years ago, on a tiny island in the middle of the Bering Sea, a woolly mammoth made a fatal misstep. It fell into a pit-like cave with no escape, and there it died.

In 2003, another animal entered the cave — with a ladder. As he explored the space with his colleagues, Russell Graham, a paleontologist from Pennsylvania State University, lifted a rock near the back. There he found a single, pristine tooth from the mammoth, oblong and bumpy and as big as a loaf of bread. “It looked like you had just taken it out of the animal’s mouth,” says Graham, a tall, broad-shouldered man. With a beard and a slightly shuffling gait, he seems a bit of mammoth himself.

He had handled hundreds of mammoth teeth in his career, but this tooth, from the Alaskan island of St. Paul, was special. It would send Graham and a multidisciplinary team of experts on a quest to reconstruct the animal’s environment and solve a mystery — one that has implications for species facing climate-induced extinction today. The mystery began when carbon dating established the tooth was just 6,500 years old. That’s several millennia fresher than any mammoth find on the North American mainland. The date made Graham wonder: After the animals persisted on this wind-battered speck of land for so many millennia after mainland populations perished, what finally killed them off?

Back in the 1960s, Paul Colinvaux, a British-born ecologist, had collected a sediment core from a lakebed on St. Paul. The preserved layers in the core revealed that the site held a sediment record stretching back at least 14,000 years — and the spot, Lake Hill, was only a quarter of a mile away from the cave where Graham had made his discovery. The answers to Graham’s questions might well lie in the dark mud of that lakebed.

A decade after he found the tooth, Graham returned to St. Paul with his team of experts, ready to glean clues from new sediment cores about the intertwined history of the island and its mammoths.

Weather complicates the early spring expedition to the largest of the Pribilofs, a cluster of volcanic islands poking up from the Bering Sea. Upon the team’s arrival to St. Paul, rain quickly turns to snow and they’re forced to stay in the island’s lone town. It’s home to most of St. Paul’s 500 residents, the majority of them Aleuts whose ancestors were brought here by Russian traders more than two centuries ago to work the northern fur seal killing fields.

The team waits for the clouds to lift over the town’s skyline: the Russian Orthodox church’s onion dome, the brightly colored houses and the seafood processing plant. Blue-gray Arctic foxes native to the 40-square-mile island appear without fear outside home base, a research station run by the National Oceanic and Atmospheric Administration.

At last, a bright morning arrives and the researchers head to Lake Hill, a crater lake encircled by a snow-covered rim. Graham’s team drills a hole through a foot of ice. Members thread together meter-long sections of pipe attached to a tube, or corer. With a splash they drive the assembly through the water and push it into the lakebed. Minutes later, the pipe is retrieved, the final meter plugged with sediment layers preserved just as they were deposited.

Within the muck lies a forensic record of what fell, washed or otherwise settled onto the bottom of this lake: fungal spores, fragments of plants, ancient pollen, volcanic ash, the remains of tiny crustaceans — and maybe even DNA from the mammoths themselves, shed while wallowing in the water. The team will use these bits of evidence to build a picture of how the island changed over millennia and, potentially, determine when the mammoths went extinct.

Team members draw the sediment from below, meter by meter, extruding it into plastic tubes. The tubes are then sealed and tucked into a sleeping bag so they don’t freeze, which could distort the layers’ fine resolution.

The work is messy, cold and physical. By day’s end, frozen mud crusts on the researchers’ clothing and clumps in their hair.

Each meter of cored sediment reaches further back in time. As team member Jack Williams of the University of Wisconsin-Madison guides the sixth segment into a tube, he notices the mud changes from a warm brown with a pudding-like texture to a blacker, firmer consistency. The team estimates it corresponds to deposits from roughly 6,000 to 8,000 years ago, spanning the period when Graham’s mammoth died in the cave. That means this segment could include the period of extinction, if mammoth DNA is present in its lower, older layers but absent from the top. “There’s mammoth in there,” Williams predicts.

St. Paul’s landscape of treeless tundra has likely changed little since the end of the last ice age, some 12,000 years ago. It’s easy to imagine where mammoths would have stood on the gentle slope, using their tusks to clear the snow while browsing for vegetation. In the summer, they would have trampled tiny tundra wildflowers as they lumbered to the lake, one of the main freshwater sources on the island.

Researchers estimate that mammoths disappeared from North America’s mainland 10,000 to 14,000 years ago. But they persisted on St. Paul for millennia after that. The animals also survived — for even longer — on Wrangel Island, deep in the Russian Arctic. Researchers there have found teeth that are only 4,000 years old; the Wrangel mammoths were alive as Egypt erected its great pyramids.

Like Wrangel, St. Paul hasn’t always been an island. At the peak of the last ice age, around 21,000 years ago, it was a volcanically active spot on the southern edge of the Bering Land Bridge. Mammoths, saber-toothed cats, short-faced bears and other large animals, or megafauna, roamed the passage between North America and Eurasia. Then, around 11,000 years ago, the climate began to warm and sea level rose, swallowing the land bridge and turning the area into an island over the next 2,000 years. The mammoths were trapped, but their very isolation may have protected them — at least for a while.

Researchers debate what caused the mammoths to die out on the mainland at the end of the last ice age. Some say it was the changing climate, while others argue that humans were the culprits, hunting mammoths to extinction.

Even within Graham’s team, there’s a friendly difference of opinion. “To me, one of the telling patterns is that the timing of extinction around the world roughly corresponds to the timing of human migration dispersion around the world,” says Williams, who sees humans as the major culprits. On the mainland of North America, the arrival of humans, climatic changes and the mammoth’s extinction happened at roughly the same time. But mammoths weathered many other climatic shifts over the hundreds of thousands of years they roamed Earth, Williams notes.

Graham, on the other hand, thinks climate was the driving force. Humans may have sped things up, he says, but mammoths and other Ice Age megafauna were on their way out anyway. “I think the extinctions would have occurred whether humans came or not,” he says. He believes each climate shift that the mammoths survived actually weakened the species, pushing them closer to extinction.

The cores of sediment from Lake Hill won’t settle the general megafauna extinction debate: St. Paul’s isolation likely sets it apart from whatever influence humans had on the mainland mammoth population. There’s no evidence that humans ever made it to this dot of an island before Russian fur traders arrived in the late 1700s.

On St. Paul, at least, the more likely suspect is climate change, though it’s also possible that the final straw for mammoths here was habitat loss: With increasing sea levels, the island may have become too small to sustain its megafauna population.

Determining what happened to St. Paul’s mammoths isn’t purely a historical exercise. Many species worldwide currently face the same pressures: changing climate, human encroachment, rising seas. “This is hugely relevant to today,” Williams says. “Can climate change alone cause species to go extinct? Or is the story that climate change is a stressor that makes populations more susceptible to extinction when combined with other stressors?” The most recent major die-off event, at the end of the Pleistocene some 11,000 years ago, “has direct carryover to how we think about the current wave of extinctions and how to minimize them,” Williams says.

After four days of coring — interrupted at one point by a blizzard — the team has 400 pounds of mud in 45 meters of tubing to ship to a Minnesota lab, where the next chapter in the mystery unfolds.

It’s May in Minneapolis when the team is reunited with their mud at the University of Minnesota’s National Lacustrine Core Facility — as the name suggests, a repository for lake-based sediment cores. Staff at LacCore have already sliced the cores in half and taken high-resolution photos. Graham and colleagues from various disciplines will spend three days cutting the cores into tiny pieces using stainless steel spatulas. They’ll divide the thousands of samples among little plastic boxes to take back to their respective labs, which spread across North America, as the mammoths once did.

Beth Shapiro at the University of California, Santa Cruz, gets first dibs on the samples since her team will take on the most sensitive test, at greatest risk of contamination during handling: the search for ancient mammoth DNA in the layers of sediment.

With the help of a graduate student, Williams will look for a kind of mammoth proxy: sporormiella, a fungus that lives in the dung of large herbivores, and grains of ancient pollen that can reveal what kind of vegetation once grew around the lake.

At the University of Alaska Fairbanks, Matthew Wooller and colleagues will analyze the mud for the remains of diatoms, water fleas and other tiny life to learn about the water temperature and clarity over time.

Duane Froese at the University of Alberta will look for layers of volcanic ash in the sediment that can be linked to known eruptions in the region. This, along with carbon dating of tiny plant pieces found in the mud, will allow the team to put firm dates at points along the mud core, calibrating it in time. Ash layers may also implicate or rule out a volcanic eruption as the cause of the mammoths’ extinction.

With the mud distributed, the researchers disperse. They will spend more than two years analyzing and discussing the muck as they try to pinpoint when the mammoths went extinct, and to understand why.

Shapiro’s team first compares unclassified DNA sequences found in the sediment samples with the genome of the mammoth’s closest surviving relative, the African elephant. It’s a match. Even better: In 2015, they’re able to compare the sediment DNA with newly-sequenced mammoth DNA from a separate team of researchers. It was another match in all core samples from 5,650 to 10,850 years old, the oldest sediment they’d collected on St. Paul.

To rule out a random match, the team compares the sediment DNA with that of a two-toed sloth, an animal with zero chance of being found in the Arctic in the last 12,000 years. As a further check on their work, they also compare the unclassified sediment DNA to that of the animals most likely to contaminate a sample, including humans. The DNA sequences don’t align, which makes Shapiro and her team confident that what’s in the sediment really belonged to a mammoth.

Meanwhile, Williams’ team finds the sporormiella they are looking for, as well as two other fungal spores associated with dung that can serve as additional proxies for an actual mammoth.

“Of the three spore types, two disappear from the core’s sediment layers within 2 centimeters of each other, which would be the equivalent of a few decades,” Williams says of the mammoth dung-loving fungus. The ancient mammoth DNA disappears at the same point. “Having the proxies match that well is what you always hope for but rarely attain,” he says.

When Froese’s lab identifies a core sample layer of volcanic ash from a known eruption in the area 3,595 years ago, it gives added confidence to the teams’ timeline. The multidisciplinary effort establishes that the mammoth’s extinction occurred 5,600 years ago, plus or minus a century.

“We have one of the best constrained times of extinction for mammoths anywhere — and maybe for any prehistoric extinction,” Williams says.

With the end-date for the St. Paul mammoths now confidently established, the researchers can use it to solve the biggest part of the mystery: What finally killed the animals?

Graham and several of the other team members begin by ruling out possible causes, such as a change in vegetation to something less mammoth-friendly. Pollen in the core samples reveal that herbaceous plants on the island gave way to shrubs, but only after the mammoths’ extinction. It’s possible that a climate shift changed the composition of the plant communities — or the shrubs may have become dominant simply because the mammoths were no longer stomping around, keeping their growth in check.

The data does show a modest rise in temperature that predates extinction. With the melting ice and rising seas of the End-Pleistocene, the island began to shrink rapidly until about 9,000 years ago. At that point, the rate of land loss slowed until St. Paul reached its current size, smaller than the city of San Francisco, around 6,000 years ago. The shrinking island would have downsized the mammoth population, too, as resources grew scarce.

But the final straw, the team concludes, is something of a surprise.

Beginning about 7,850 years ago, the variety and number of diatoms and water fleas found in the sediment change in a way that indicates the water was becoming shallower and murkier. The lake began to dry up — the modest bump in temperature was apparently enough to increase evaporation. The mammoths were losing their main watering hole. “Elephants need fresh water every day,” Graham says. “They drink 70 to 200 liters per individual. There’s no reason to think mammoths were different.”

Thanks to a combination of diminished island size, rising temperature and perhaps the animals themselves causing erosion and fouling their own water, the mammoths died of thirst.

The team solved the mystery of the St. Paul mammoths’ demise. But why they persevered may, paradoxically, be harder to answer. It’s reasonable to propose that their isolation protected them from human activity — but that idea is difficult to prove, and Graham suggests mainland vegetation may have changed in ways St. Paul’s didn’t. Either way, their isolation eventually came with a downside: When environmental changes made the island less habitable, they had nowhere to go. It’s a cautionary tale about the vulnerability of island populations today.

“Usually we’d say this was an insignificant climate change, but on St. Paul, it turned out not to be,” Graham says. While water was the crucial factor there, it could be other limitations — habitat or availability of food — for other species in other places. Graham adds: “This research suggests that many island populations, not just in the Bering Strait but throughout the world, could be jeopardized by coming climate change.”

[This article originally appeared in print as "Mammoth Island."]