In the spring of 2024, a group of scientists attending a national seismology conference in Anchorage purposely sailed into danger.
Chartering a tour boat, they sailed into Prince William Sound and lingered in a fjord lined by unstable slopes that they recognized could collapse at any moment, triggering localized but massive tsunamis.
“We made jokes about it at the time like, ‘Oh, wouldn’t it be interesting if all the seismologists in the country were killed in the catastrophic landslide?” said Mike West, Alaska’s state seismologist and director of the Alaska Earthquake Center at the University of Alaska Fairbanks.
In September, just five months later, at the aptly named Surprise Inlet, the slope failed, and a succession of five landslides dumped debris into the water, triggering a localized tsunami exactly where the seismologists had paused during their field trip.
And a month ago, at a scenic fjord elsewhere in Alaska, a mountainside collapsed into a fjord called Tracy Arm about 80 miles southeast of Juneau. The Aug. 10 event generated a tsunami that ran at least 1,500 feet up the wall on the other side of the fjord. It was the biggest Alaska landslide since at least 2015, when the collapse of a mountainside in a remote part of Wrangell St. Elias National Park caused one of the biggest slides in North American history and generated a 630-foot-high tsunami at a site called Taan Fjord.
Such landslide-triggered tsunamis can be deadly. In 2017, four people were killed by one in Greenland.
In the recent Alaska tsunamis, the most serious loss was a kayak and other gear that was swept out to sea at Tracy Arm. The affected campers, on an island about 4 miles away from the landslide, reported a 100-foot wave that morning but were otherwise unharmed.
Alaska’s luck may not last, warn scientists with the Massachusetts-based Woodwell Climate Research Center.
The center has gathered several Alaska experts into a program called the Tsunamigenic SLope Instabilities Partnership, or Arctic T-SLIP, to help address the risks of tsunami-generated landslides amid Alaska’s melting glaciers and thawing permafrost,
“Hazards from landslides and landslide-generated tsunamis are increasing. Among all the natural hazards in Alaska, this hazard is the least quantified and integrated into local, state, and federal hazard management practices,” the Arctic T-SLIP website says.
Alaska’s daunting challenges
Other glaciated regions of the world, like Switzerland, Norway, and Iceland, have rigorous landslide monitoring programs. They face their own threats. In Western Norway, for example, an array of instruments initially set up 20 years ago has been monitoring an unstable slope at a site called Åknes that, if it fails catastrophically, could produce a tsunami that would wipe out nearby communities.
But in Alaska, where there are thousands of mountains and tens of thousands of glaciers that are unmonitored and have not even been named, and vast distances and remoteness pose daunting challenges. Scientists, agencies and communities have thus winnowed down their monitoring to the highest priority sites.
That leads to Prince William Sound.
The sound is the site of several communities. It is home to busy commercial and sport fisheries, and traversed by vessels ranging from kayaks and other small recreational watercrafts, to huge cargo ships and oil tankers. It is visited by hundreds of thousands of people who arrive by boat, train or vehicle each year.
On the northeast side, the main focus of landslide-monitoring is Barry Arm, a fjord about 30 miles from the closest coastal town of Whittier, where Barry Glacier has been receding.
The mountain slope there has been moving gradually for decades, since at least the 1930s, historic photographs show. In recent years, with drastic loss of Barry Glacier’s slope-supporting ice, the movement increased significantly, and there are already periodic slides. By 2020, there were episodes of movement reaching 2 meters in a 30-day period, and Barry Glacier’s retreat had exposed 600 million cubic yards of terrain, according to NASA, an amount six times the size of the 2015 Taan Fjord landslide.
Those changes led to a dire warning from scientists that year: A catastrophic collapse of the slope could create a tsunami big enough to send a potentially fatal wave to the town of Whittier, 30 miles to the south and west. The scientists were so alarmed that they penned an open letter about the risks.
“This tsunami could impact areas frequented by tourists, fishing vessels and hunters (potentially hundreds of people at one time),” they wrote. “We believe that it is possible that this landslide-generated tsunami will happen within the next year, and likely within 20 years.”
The tsunami didn’t happen in 2020, but the risk remains. Whittier is home to about 275 people, most of whom live in a single high-rise condo building. But the small permanent population belies the town’s significance. It has a deepwater port that buzzes with activity, especially in the summer. Aside from its vessel traffic, it serves Alaska Railroad trains and a variety of vehicles that arrive through a one-of-a-kind single-land tunnel. Other vehicles depart from the Whittier harbor on Alaska Marine Highway ferries connecting the road system to the rest of the sound.
The discoveries at Barry Arm mobilized what is now a state-of-the-art, federally funded monitoring program that has made the site one of the most instrumented mountain slopes in Alaska.
Prior to that, West said, there were just “random university researchers or a good-hearted person at the state survey.”
Partners in the multiagency program include the U.S. Geological Survey, the National Oceanic and Atmospheric Administration and its National Weather Service, the Alaska Division of Geological and Geophysical Surveys and UAF and its Alaska Earthquake Center, among others.
Dennis Staley, a USGS scientist and Alaska landslide project leader, ticked off the equipment there.
The “backbone” is a set of tidal gauges installed by the National Weather Service’s National Tsunami Warning Center, Staley said.
Other equipment includes Alaska Earthquake Center seismic stations, which now serve the dual purposes of detecting earthquakes and detecting landslides. There is a ground-based radar system on the opposite side of the fjord to measure downslope movement. There are weather-tracking devices. There are infrasound devices that capture the sounds of rockfalls and other events. There are remote cameras to capture images over time. And new this year, Staley said, is an array of seismic devices at the top of the landslide area.
Some of the Barry Arm instrumentation can gather information from wider areas, Staley said. Last year’s Surprise Inlet slides, for example, were captured by equipment set up at adjacent Barry Arm.
To maintain the system, scientists travel to the site at least monthly from May to October for tasks like battery replacement and a few times each winter, Staley said.
But Alaska conditions take a toll. The first seismometer installed in 2020 directly on Barry Arm’s landslide area was wiped out by a snow avalanche the following spring, he said.
Adjusting to risks
After the first year of on-site monitoring, scientists downgraded their predicted worst-case scenario of a complete slope collapse to from 30-foot wave at Whittier to a 7-foot wave hitting about half an hour after the landslide.
That would still be dangerous, with a possible runup of over 500 meters on the opposite fjord wall and wave heights of over 200 meters in northern part of Barry Arm and over 5 meters in southern Barry Arm. Harriman Fjord, which is connected to Barry Arm, would be hit by a tsunami within 10 to 15 minutes of a big slide, scientists estimate.
Scientists have been providing periodic updates since the Barry Arm project started.
“Our job as federal scientists, government scientists, is to do our best to better understand what the actual hazard is, using the best available science, and then to make sure that we’re communicating about that hazard to the public,” Staley said.
In Whittier, people have adjusted their activities according to the risks they perceive.
A tour company that has trademarked its popular “26 Glacier Cruise” – even using an advertising jingle to do so — has reconfigured its itinerary to bypass the hazard area.
Phillips Cruises & Tours still brings sightseers to 26 glaciers, but the trip offers a different mix than in past years, said Cody Hanna, the company’s vice president of marine operations. It is not sailing into Barry Arm anymore and, for now, the tour boats are also skipping Surprise Inlet and Surprise Glacier, though the latter has been a big draw for sightseers. Instead, the Phillips boats are taking a little more time on their tours and sailing a little farther out into the sound to a site called College Fjord, where there are multiple other glaciers, including the thinning but advancing Harvard Glacier.
“You can’t predict it, but it’s just grown enough to where it’s just not worth it anymore, so we’re back to Harvard Glacier and College Fjord,” Hanna said of the landslide risk.
Peter Denmark, owner of a kayak tour company, said he has also changed his operations because of landslide risks at Barry Arm. “We are not running any commercial tours up in that area, essentially due to liability concerns,” said Denmark, who also serves on Whittier’s city council and holds the position of vice mayor.
Barry Arm and the other landslide-prone sites around Prince William Sound and elsewhere keep him alert to the dangers, he added. “It is, personally, another hazard just from living in Alaska,” he said.
Brenda Tolman, who has operated a waterfront shop in Whittier for more than 30 years, is less concerned.
If Barry Arm hasn’t caused a localized tsunami in the past, it probably will not do so now, she figures. “I don’t see it as any big threat,” she said.
She does not worry much about the slides in the nearby mountains because they have been happening for decades, she said. “You’ll hear the crash when they come down,” she said.
Tsunami awareness
In seismically active Alaska, people associate tsunamis with earthquakes. The 1964 Great Alaska Earthquake, which was the second most powerful in the historical record killed 139 people, including 11 in Whittier. Most deaths were from the tsunamis that slammed against coastlines.
The 1964 earthquake has a profound legacy in Alaska. That includes preparations for future earthquake-generated tsunamis; several communities are designated by the NWS as “TsunamiReady,” with detailed safety systems to prepare for earthquake-generated tsunamis.
There is also a history in Alaska of dangerous landslide-induced tsunamis. In Lituya Bay, an ellipse-shaped fjord cut into the Gulf of Alaska coast, a landslide shaken loose by a magnitude 7.8 earthquake in 1958 triggered a megatsunami. It sent water 1,720 feet up a mountain slope and killed five people. Thirty-six years later, in 1994, a freak underwater slide in the Southeast Alaska town of Skagway caused a localized tsunami that killed a construction worker at a dock.
The slides and tsunami risk being studied in Prince William Sound are different.
They are caused not by earthquakes but long-term weakening of slopes that are no longer supported by glacier ice. A contributing factor sometimes is the thaw of high-altitude permafrost that binds slabs of rock together. Both changes are linked to long-term climate warming.
As glaciers at the bases of mountains rapidly retreat, they are de-buttressing, said Jillian Nicolazzo, the Alaska Division of Geological and Geophysical Survey’s landslide program coordinator.
“So that the ice that had been holding all that rock in place is not there,” she said. “It’s too steep, so it has to find an equilibrium.”
The thawing high-altitude permafrost that contributes to slope weakness known as rock permafrost. Though lingering snow patches offer hints, it is hard to determine exactly how much of it exists on mountains or where it is located, Nicolazzo said. What is known is that it is sensitive to warming and implicated in landslides elsewhere.
“When that ice does start to melt, the space that it occupied becomes a void, and rock and soil fill that void,” she said.
Rain can be another trigger. Surprise Inlet’s landslides last year got their final push from heavy rainfall that loosened the already unstable slope, research shows.
For residents, there is a pressing question: Can people be warned to evacuate from a landslide tsunami just as they are with earthquake-generated tsunamis, which can take hours to hit shores?
“Potentially, yes. The difficulty is you have to know that the landslide occurred and where it occurred, to start with,” Nicolazzo said. “We might be talking minutes, but it takes the scientists that long to triangulate where it came from,” Nicolazzo said.
Seismic signals are used to pinpoint locations. Scientists at the Alaska Earthquake Center can spot the differences between landslide seismic signals and earthquake signals. Earthquake signals are more varied in their frequencies and more sudden; landslide signals tend to be in low frequencies and are more gradual.
The center’s seismic network proved vital in the Tracy Arm event, Nicolazzo said.
“They saw it pretty quickly and they were able to triangulate,” she said. “But the timing does make it difficult. The landslide happens, and then water starts moving.”
Good news, bad news
The downgrading of the severity of a Barry Arm collapse is a piece of good news that has come from the new monitoring. The bad news – as demonstrated by the Surprise Inlet landslides — is that there are numerous other Prince William Sound sites at risk.
A 2024 USGS study used satellite data to identify 43 sites in the sound where landslides occurred between 2016 and 2022. Fourteen were previously known. Eleven of the identified sites have potential for generating tsunamis.
Rising to the top of that list is Maynard Mountain, said Gabriel Wolken, a geologist who heads the Division of Geophysical and Geological Surveys’ Climate and Cryosphere Hazards Program.
Maynard Mountain, through which the single-land Anton Anderson Memorial Tunnel runs, looms directly over Whittier and a new cruise dock that started operations this year. Shrinking Learnard Glacier clings near the top, while rocks frequently tumble off its lower flanks.
Government agencies lack the resources to do a Barry Arm-like instrumentation and monitoring program, Wolken said. But what agency scientists can do is watch the mountain, whether from land or remotely through satellite imagery captured from space.
“We did that at Barry Arm. We’ve learned a lot from that exercise,” Wolken said.
To better understand Maynard’s contours and landslide potential, the division was planning a survey of the mountain this fall using lidar, a technology that measures distance using light beams. This year’s planned lidar survey follows one performed at Maynard in 2022.
The probability of a tsunami-generating landslide at Maynard is considered to be very low, but the results if one were to happen would be “cataclysmic,” said Bretwood Higman, a Seldovia-based geologist who is part of the Woodwell Climate Research Center team.
“If a landslide failed catastrophically, you’ve got two minutes. There’s nothing you can do in two minutes, basically,” Higman said.
Lessons learned at Barry Arm are also being applied on the other side of the Whittier tunnel, where retreating Portage Glacier is exposing an unstable slope that is already the site of frequent rockslides.
As with Maynard Mountain, lidar surveys are planned at Portage, Wolken said. There are also plans for ice-penetrating radar surveys to better understand the glacier’s thickness and the characteristics of the bedrock below, he said.
Higman, who spends a lot of time scouting the Portage Glacier area, said two slide areas by the melting glacier are combining to reshape the slope. “The very bottom of the slope is just falling apart and crumbling onto the glacier,” he said.
There is no permanent community at Portage Glacier, but the area is a popular destination for tourists and recreational users. For tourists, Portage Glacier’s retreat is part of the attraction. At the Forest Service’s Begich, Boggs Visitor Center, big windows that once provided close-up views of glacier ice now overlook open water. People wanting to see the namesake glacier must now hike up a mountain pass or boat across the lake or, in winter when the lake is frozen, ski or skate across it.
New device could provide communities with advanced warning
Higman, who has experimented with a series of measuring devices at different Southcentral Alaska mountain sites, has come up with what he believes could be a valuable tool for community-based landslide safety.
It is a simple, glass-encased device using the type of radar signals that are employed on fishing boats. The devices send radio signals, “like a walkie-talkie,” through the internet to nearby databases. The goal is to detect changes in the slope. Since small movements precede big slides, the devices could give advance warning opportunities.
The devices promise to be more practical than other equipment Higman and his partners have positioned on various slopes, he said. They do not need to be embedded into holes dug into bedrock, as is necessary with some movement-measuring equipment. They are relatively cheap and easy to retrieve and replace, if needed. They could be managed by members of the public in affected communities. And they might avoid the problems with equipment he has installed at Columbia Glacier, a Prince William Sound landmark, where mountain goats have gnawed on or possibly eaten expensive communications and GPS devices.
Higman said he has nearly a dozen of the devices, and he has deployed them in a few places, mostly at Portage and at a Matanuska River Valley site called Glacier View, where a slide threatens a small subdivision.
It’s only a start, he said.
“I’m aiming for 100 sites. That’s what we need,” he said.
Beyond the sound
On the long list of landslide-prone glaciated areas is the coastal town of Seward, a popular tourist and recreation destination just outside of Prince William Sound on the Kenai Peninsula.
The town of 2,800 has recent experience with landslide-generated tsunamis.
Last year, a section of mountain face fell into a marine site called Pedersen Lagoon, triggering a local tsunami that reached a height of about 56 feet. The Aug. 7, 2024, event did not injure anyone, but it damaged a private lodge’s wooden boardwalks and affected a Kenai Fjords National Park camping area. Previously, a 2022 landslide in Seward caused a small tsunami that was detected by equipment aboard the UAF-operated research vessel Sikuliaq.
Wolken, who loves to visit the picturesque coastal town with his family, laughed when asked if the precarious conditions of the slopes there dampens his enjoyment.
Family trips now to Seward are punctuated with safety lectures delivered to his two young daughters, and his watchfulness is constant, he said.
“I’m always looking at the slopes. I know the destruction that’s possible from gravity,” he said. “I don’t zone out. I’m always scanning.”
This story has been supported by the Solutions Journalism Network, a nonprofit organization dedicated to rigorous and compelling reporting about responses to social problems, http://solutionsjournalism.org.