Bretwood Higman has spent years scouring Alaska’s mountains for signs of potential landslides. One spot he looked: Tracy Arm Fjord – a popular cruise destination between Juneau and Petersburg.
“I didn't see anything. So I didn't have anything marked,” Higman said. “I’ve been working on mapping all across Alaska.”
So he was caught completely off-guard when a mountainside collapsed into the fjord early one morning last August, generating one of the largest landslide tsunamis in recorded history. The tsunami, which at its highest point was nearly as tall as One World Trade Center in New York City, scoured vegetation off the fjord’s walls as it barreled down the narrow channel.
The landslide came down on Aug. 10 at 5:20 in the morning, so no boats were present. But the area regularly sees more than 20 boats a day in the summer, including large cruise ships that can carry up to 6,000 people.
"This could be really catastrophic if it were to happen at 10 a.m., during the day,” said Ezgi Karasozen, a research seismologist with the Alaska Earthquake Center.
Higman agrees.
“I have tried to encourage my colleagues to look at this and explicitly acknowledge it as a failure,” said Higman, who is a geologist and the executive director of Ground Truth Alaska, an educational nonprofit focused on Alaska’s environment. “We clearly did not identify this beforehand.”
The ordeal kicked-off a months-long research effort to nail down what exactly happened that day. The work culminated in a new paper, published this week in the journal Science, that concluded the landslide caused a megatsunami second only in height to the 1958 tsunami in Lituya Bay.
That the landslide was a complete surprise underscores the difficulty of monitoring for and predicting these types of events – and it shows there’s a long way to go before experts can do so reliably.
It also highlights the widespread nature of the risk, which is only intensifying as climate change fuels heavier precipitation, rapid glacier loss and more landslides.
“That we were all caught off guard, says a lot about these hazards,” said Karasozen, a co-author on the report. “We should be acting more quickly.”
When the mountainside collapsed
Tracy Arm is a popular Southeast Alaska cruise destination because it offers close-up views of the Sawyer and South Sawyer Glaciers from a scenic fjord lined with tall peaks, some towering more than 6,000 feet above sea level.
When the mountainside collapsed, it released so much earth that, if it were re-shaped into a cube, would have been a quarter mile long on each side. When the slide hit the water, it triggered a massive wave that ran 1,578 feet up the opposite side of the fjord.
Within minutes, the wave reached popular glacier viewing points and ripped much of the vegetation off a nearby island.
The tsunami got smaller as it traveled away from the landslide and down the fjord. But even miles away, it caused dangerous waves and currents observed by nearby boats, according to eyewitness accounts provided by the researchers.
One of them was the Blackwood, a 103-foot yacht. Captain Edin Mujanovic reported "significant rolling” while anchored just inside Tracy Arm. When he looked outside, he said “water was rushing out of the bay at a rate that I had never seen before.”
Another boat – a 65-foot, eight passenger tour boat named the David B – reported a “crazy surge” of between five to 10 feet from an inlet off the neighboring Endicott Arm.
“Sorry no pictures,” Captain Jeffrey Smith wrote in a text message around 11 a.m. “Seemed like an emergency situation, not just something to take pictures of.”
A small cruise ship named the National Geographic Venture was also nearby but reported only “minor effects of the wave.”
“I think we lucked out on where we were,” Captain Thomis Morin said in a written testimony included in the research.
Searching for warning signs
Even more important is new information about what happened in the days and hours before the tsunami. Those details are crucial, given they could be used to pinpoint new spots where similar slides might happen down the line.
“I think this event gave us some clues on that end,” Karasozen said.
Typically, when looking for areas that might eventually slide, geologists keep an eye out for something called deformation. That’s when a slope’s bedrock starts cracking and shifting downhill.
When Higman looked at satellite imagery of Tracy Arm before the event, he did see a landslide that likely happened in the 1990’s. But he didn’t see signs of deformation, which is why the area didn’t set off any alarm bells.
In the time since the tsunami, Higman said, a bunch of experts have looked back in time to see if they could identify any of the typical warning signs.
“Even with the knowledge of what's going to happen, we can't really identify it using those approaches,” Higman said. “That's a huge thing of note.”
So what was happening? Perhaps most notable is that the South Sawyer Glacier had retreated rapidly in the weeks leading up to the slide, leaving the base of the slope exposed and without support.
Much of that exposure, the study notes, happened between Aug. 2 and Aug. 5, just days before the mountain collapsed.
“That retreat was exactly at this place where the failure happened,” Higman said. “And so that seems like more than a coincidence, right?”
There were other factors, too. For starters, there was a long spell of moderate rain.
Another possible warning sign: When the researchers looked back at data from the hours before the landslide, they noticed seismic activity that looked like small, distinctive earthquakes happening over and over again. They likely came from the slope itself as it became less and less stable, Higman said.
A stronger grasp on those factors – and what they mean – could help experts catch the next landslide tsunami before it happens. The seismic activity alone, for instance, could have served as a warning a full day before the landslide. That would have been plenty of time for boats to relocate or avoid the area.
A robust monitoring and alert system throughout Alaska would take years to stand up – plus quite a bit of coordination across a long list of agencies and other organizations. But Karasozen said researchers are already thinking about what that might look like.
“There are a couple of things that can be pieced together to issue a color-coded warning,” she said.
The rapid retreat of a glacier, plus a lot of rain, could, for instance, trigger a “yellow” warning. Add in consistent seismic activity, and a more serious alert could go out.
Experts say that understanding those signals better, and monitoring for them, is only becoming more important as glaciers recede – and precipitation intensifies – with climate change. It’s a feat made even more daunting by Alaska’s sprawling coastline.
“We have lots of slopes that we are worried about in Alaska,” Karasozen said.
“The rate of these events is increasing now,” Higman said. “If it's increasing 10 times over what it was not long ago, then it's very important that we look in the right place.”
