A new algorithm could help detect landslides in minutes

Landslides can be truly devastating, killing people and animals that can’t get out of the way in time and washing away property. Landslides generally occur during earthquakes, volcanic eruptions, or massive rainfall that make a sloped section of land like a cliff unstable. Now, a team from the University of Alaska Fairbanks (UAF) have developed a new method that could be used to remotely detect large landslides within minutes and tell if the slide is a tsunami hazard. Their method is described in a study published February 9 in the journal The Seismic Record.

[Related: California wildfires may give way to massive mudslides.]

Monitoring Alaska’s glacial fjords for danger

The study cites a 2015 landslide that sent 100 million cubic yards of rock into Alaska’s Taan Fjord. It generated a tsunami that stripped vegetation as high as 620 feet above the waterline.

In response, the team helped develop a prototype system capable of real-time detection that has been in place since August 2023 around the Barry Arm section of Prince William Sound. The system uses data from seismic stations already in Alaska’s monitoring network.

State and federal disaster agencies worry that a landslide and tsunami could occur at Barry Arm. The Barry Glacier has significantly retreated and left behind an unsupported fjord wall, or the slope of rock or ice rising up from the water to the top of the landmass. Over the past several years, the wall has slumped about 650 feet. Additional analysis of seismic station data on the region revealed that three landslides occurred in 2020 and 2021.

“The warming climate is causing glaciers to retreat, leaving behind valleys whose mountainsides and hillsides have lost their support,” UAF research seismologist Ezgi Karasözen said in a statement. “This is important, especially in regions like southern coastal Alaska, because huge masses of land can and do spill into water and cause tsunamis.”

According to the study, this instability has disaster agencies concerned that a catastrophic failure of the glacier wall could create a tsunami of waves several feet high that reach nearby communities in only 20 minutes.

Looking for long-period waves

Scientists monitor seismic activity that shows up in jagged waves on seismographs. When a landslide begins, it generally registers on seismic sensors as short-period waves. As the slide accelerates, identifiable long-period waves show up. Landslides eventually produce more of these long-period waves than other sources of energy like earthquakes. Most earthquake ruptures only last a few seconds, while landslides can go for a minute or more.

Seismograms recorded by several stations reflect the Barry Arm 3 landslide. (a) Long-period seismograms appear in blue, and short-period seismograms appear in gray. (b) Long-period seismograms only. CREDIT: Ezgi Karasözen and Michael West.

The detection method in the study involves quickly identifying a landslide’s long-period waves among seismic data that is crowded with short-period waves. Since glaciers can create hundreds of daily seismic events that produce waves, coastal fjords like Barry Arm create a challenge for landslide detection. Far away seismic stations also do not allow for real-time assessment, since it takes time for the seismic waves to reach those stations.

“With an earthquake, there are instruments that measure ocean wave heights, and tsunami warning centers are on alert after an earthquake,” Karasözen said. “But landslides aren’t systematically monitored in Alaska or elsewhere in the world. If a landslide-triggered tsunami were to happen, we wouldn’t know. That’s a major concern.”

A landslide algorithm and five-minute warnings

To create the new monitoring method, the team developed an algorithm that continually scans seismic data to detect a landslide’s long-period wave signature. If the system finds a match, it will estimate the slide’s location and volume. In areas that are well monitored, the location of the landslide can be estimated to within a few miles.

[Related: New AI-based tsunami warning software could help save lives.]

The team analyzed data of the three recent Barry Glacier landslides and six additional landslides to build the algorithm. The end goal is to build a larger system to alert tsunami and seismology agency personnel, but more work must be done to create this system.

While additional researchers have shown that landslide seismograms can be used to estimate location and volume, those efforts usually were typically unique to a specific region, required constant updates, and were not designed to be used in real-time.

“The potential for real-time monitoring of large landslides is one important component of the interagency effort underway to address Alaska’s landslide challenge,” Michael West, study co-author and director of the Alaska Earthquake Center at UAF’s Geophysical Institute, said in a statement.

According to the team, this new method of determining a landslide’s location, volume, and potential is quick enough to support the National Oceanic and Atmospheric Administration’s major goal of issuing a tsunami warning within five minutes of a landslide.