The Wrangell landslide happened in an instant.
Somewhere high on a hillside above Zimovia Highway, the slope started moving. In a matter of seconds, a river of mud fanned out and ran for nearly 4,000 feet, crossing the beach and spilling into the water. On its way, it tore up more than 37 acres of land.
People living there had little warning — just the sound of the slide. It buried two houses, leaving five people dead and one still missing.
According to state geologist Barrett Salisbury with the Alaska Department of Natural Resources, it’s extremely challenging to give warning for disasters like this.
“We’re always going to be surprised by a landslide event,” said Salisbury. “We could give you an estimate of where we think the hotspots for future activity could be — would be — but there’s no guarantee.”
The only guarantee, Salisbury said, is that those hotspots exist all across Southeast Alaska, where many communities have homes nestled on steep mountain slopes. And human-caused climate change could help to trigger even more landslides across the region.
Landslides are incredibly complex. Each one is shaped by the unique geology, hydrology and vegetation on a given slope. That’s the biggest reason why they’re so hard to predict.
But scientists like Gabriel Wolken of the Alaska Division of Geological and Geophysical Surveys say that Southeast mostly lacks the kinds of monitoring that could make people safer.
“We have, you know, this broad understanding of the different ingredients that come into play to aid the development of a landslide,” said Wolken, who manages the state’s climate and cryosphere hazards program. “But we still lack the data.”
How debris flows start
Salisbury was one of a team of geologists who worked to gather data about the Nov. 20 slide in the days after it happened. There are still a lot of unknowns.
“We don’t know where it started. We don’t know why it started exactly,” Salisbury said. “But we do know the moisture had to be there.”
Heavy rain almost always characterizes the “when” of Southeast landslides. Like the Wrangell slide, deadly slides in Sitka in 2015 and in Haines in 2020 happened during strong rainstorms.
On steep slopes, the earth is constantly resisting the force of gravity, thanks to the friction between grains of soil. But grains of soil are lumpy or jagged, and they don’t fit together perfectly.
“Doesn’t matter how tightly you pack them,” Salisbury said. “Those little grains only are in contact on small portions of their total surface area.”
So there’s always space between them. In the 24 hours before the Wrangell landslide, a strong storm dumped three inches of rain. All that water seeped into those spaces between the grains, building pressure in the pores of the soil.
Then, right before the slide, the rain intensified. In six hours, that burst dropped more than an inch of rain. The water kept pushing as the pores filled up completely.
“Eventually, the soil particles will be pushed away from one another,” Salisbury said.
When that happens, solid earth becomes a slurry of mud and water, and it’s primed for a type of landslide called a debris flow.
Different things can cause the earth to give way. An earthquake could shake it loose, or it could fall away from a crack caused by the slope’s topography. Or a falling tree could destabilize the soil as it pulls up its roots. In Wrangell, where wind gusts of at least 60 mph battered the trees during the storm, that’s one possibility.
Geologists do know that the debris flow in Wrangell began on a narrow course, just 95 feet wide. As it picked up speed, the slide fanned out to more than 700 feet at its widest, and the sodden soil got even heavier, gaining even more momentum as it picked up debris in its path.
“Trees, shrubs, homes, cars — anything in the way becomes a part of this debris flow,” said Salisbury.
All that material propelled the slide across Zimovia Highway, where it pooled on the pavement in an enormous pile that took days to clear.
“It is like soup,” Salisbury said. “The big front end loader was trying to stack it higher, and they’d plop it down and it’d go right back where it was.”
Not enough weather stations
Soils in Southeast Alaska can take on a lot of water — they’ve evolved that way by enduring thousands of years of heavy rain. And the amount of rain that came down on Wrangell before the slide was not necessarily unusual.
But it’s incredibly difficult to pinpoint an exact threshold for where there’s enough rain to trigger a landslide — and Southeast’s lack of weather data makes it impossible. For most communities, the most complete and official weather data is collected at airports, on flat ground, close to sea level.
“We don’t have extensive [weather] stations around the Southeast,” Salisbury said. “Especially not on a small island like Wrangell, where there are lots of homes across many different types of terrain.”
The topography of Southeast Alaska creates microclimates, where weather behaves very differently across small distances. A microclimate at mile 11.2 of Zimovia Highway may explain, in part, the cause of the slide.
“It could have been a cloud literally was denser 10 miles away from town, and it rained more,” Salisbury said. “Or because of the shape of the mountains and the channels, the high winds were focused to trigger that landslide. But we don’t know.”
There are no weather records to show what happened on that ridge, but the state Department of Transportation did set up a rain gauge at the slide zone during the search. One night, that gauge measured twice as much rain as the one at the airport.
The “where” of landslides
Determining the “where” of a landslide can be even harder. But according to Wolken, history can be a good rule of thumb.
“There is this repetition on the landscape,” Wolken said. “Where one landslide has occurred in the past, there is then the possibility that future landslides tend to occur in the same places.”
Residents across Wrangell Island can recall stories of smaller slides that happened in their own backyards, or their neighbors’ backyards, leaving behind piles of dirt or jumbles of boulders. More overgrown slide scars can be seen from planes or boats, including one big scar that Salisbury’s team spotted to the south of the recent slide. That old slide stopped short of houses.
Salisbury said it’s hard to pinpoint why the hill above mile 11.2 gave way. Along the ridge on either side, the slope is similar in its geology, topography and plant cover. But the Nov. 20 slide might have created new weaknesses in the hillside.
As the slide came down, it slammed into a series of bedrock ridges, including the bench of an old logging road. Those collisions caused tendrils of the liquified earth to split off on either side of the main slide — two to the north and one to the south.
And some of those smaller slides tapped into stream beds, which funneled them even further down the hillside, where they too stopped short of houses. Now those paths are prone to slide again.
Studying slides in a changing climate
Despite knowing the ingredients of a landslide, Wolken said predicting them with any precision is far off for Southeast Alaska.
“Warning systems can be developed in places where instrumentation and data are available in really high concentrations,” Wolken said. “But in Alaska, that’s a really challenging thing to achieve.”
Developing more complete inventories and maps of past slides can help. In Wrangell, the Department of Natural Resources did two aerial surveys — one this summer, before the slide, and another in late November, after the slide.
Those datasets will be used to build models of the slopes across Wrangell, which will help reveal channelized areas, extremely steep slopes, or newly unstable slopes that may have been loosened by the same Nov. 20 storm.
And at a state level, the department’s landslides hazard program is working with the U.S. Geologic Survey to compile a statewide landslide database, which will include as many historic slides as possible.
In the near term, the Alaska Department of Transportation will be installing two new weather stations at the site of the slide — one at the bottom of the slope and another on the ridge. Those will collect data about soil moisture, temperature, wind and rainfall at the slide area.
Wolken said that data is especially important as climate change, caused by the burning of fossil fuels, intensifies the risk of slides.
“We’re in a changing environment right now, with changing climate conditions including increases in heavy rain, snowfall and rapid temperature changes,” Wolken said.
Efforts to develop tools that can help people prepare for climate change in a slide-prone Southeast are just getting underway. After a 2015 slide killed three, people in Sitka were scared of the steep slopes all around them.
Researchers at the Sitka Sound Science Center, with collaborators at other scientific agencies, developed a simple, risk-based warning system in an attempt to give the community some peace of mind. It took four years to install more climate stations and develop a model for forecasting landslide conditions — a much more attainable goal than predicting landslides. Many Southeast communities, including Wrangell, may look to that system as a model.
“And when it comes to predicting a landslide, it’s always going to be really very challenging, and especially with absolute certainty,” Wolken said. “But if we have data, we at least have a way to assess and monitor.
