Since 2020, the Barry Landslide in Alaska's Prince William Sound has been outfitted with instruments monitoring seismic signals from the area, as researchers hope to catch a destructive, tsunami-generating landslide before it starts.
A team of scientists studying those signals have identified an unusual class of seismic events, characterized by impulsive, high-frequency signals that increase in rate from late summer to mid-winter, before coming to an abrupt halt in the late winter or early spring.
In Seismological Research Letters , Gabrielle Davy of the University of Alaska Fairbanks and colleagues suggest that these signals come from water freezing and thawing within microcracks in the rock located beneath nearby Cascade Glacier. The team is the first to systematically analyze these short-impulsive events near the landslide.
While these events do not correspond to movement of the landslide, they could be helpful in monitoring changes in the overall hydraulic environment behind the landslide, which in turn could be a trigger for slope movement, the researchers note.
The search for seismic precursors for a landslide at Barry Arm is a critical endeavor. The risk of landslides is high, because the landslide slope is steep and rests on a crumbly and heavily faulted bedrock. The slope has also lost an important buttress in Barry Glacier, which has been melting and retreating dramatically over the past century.
"What makes Barry Landslide especially concerning is the size of the landslide," Davy explained. "It's a large, slowly moving mass—on the order of about 500 million cubic meters—that has been creeping for decades."
"If a rapid collapse were to occur, the material would fall directly into the fjord, and that could generate a tsunami with potentially high wave heights," she added. "Barry Arm is visited by kayakers and cruise ships, and nearby communities such as Whittier could be affected, so understanding the hazard is important from both a scientific and a public-safety perspective."
Recognizing the risks, the area of the landslide has been heavily instrumented since 2020. The SRL study by Davy and colleagues is one of the first to sort through the dataset of seismic signals captured by the instruments.
The researchers took on the massive task of manually reviewing a year's worth of continuous seismic waveform data for their study, looking for any signals that might be useful in predicting when and where a landslide might occur.
Manually inspecting the dataset allowed Davy and colleagues to understand the diverse set of signals collected, which can be caused by slope movement, frequent earthquakes, glacier movement and other seismic "noise" from the environment.
"We needed to build a clear baseline understanding of the types of signals that routinely occur in the area, so that any unusual or previously unrecognized signals would stand out. By spending time with the raw data, you train your eye to recognize what 'normal' looks like" before developing classification tools and detection algorithms, Davy explained.
Once the researchers had a way of identifying these unusual short-impulsive events in the seismic records, they used weather and rainfall data and ground-based radar data to measure changes in slope deformation to analyze the pattern and location of the events.
The signals' characteristics, locations and strong temporal patterning suggested a source of small, brittle events that occur seasonally when water freezes and thaws within rock cracks.
"Similar seismic signals have been documented in other settings, although they are not widely reported," Davy said, citing a recent study from Norway that found comparable events near an unstable rock slope that "suggested that their signals may be linked to freeze–thaw processes acting on cracks within the bedrock."
Co-author Ezgi Karasözen said the Alaska Earthquake Center now has a regional landslide detection system in a testing phase at the Barry Landslide, "which will alert us to any slope failures in this area."
"As research on landslide seismology grows, there's increasing recognition that precursor seismic activity—when it does occur—can be an important source of early warning," said Karasözen. "That motivates broader investigations not only at Barry Arm, but also at other sites in southern Alaska where similar hazards exist."
