A new report focuses on using remote sensing technologies to identify landslide hazards in a large area of coastal Alaska. Of the identified landslides, preliminary analysis suggests that 11 of the 43 could result in potential tsunamis if they were to catastrophically fail.

Small changes over time add up

Sources/Usage: Public Domain. View Media Details

As glaciers in Alaska continue to melt and retreat due to warming temperatures, they may leave behind unstable slopes. No longer frozen, some of these slopes could start to move, possibly resulting in landslides. When slope stability is lost near water, there’s a potential to generate tsunamis as well.  When landslides near water move at greater than ~4 m/s or 10 mph (a speed faster than most humans can run) they have the potential to generate tsunamis that can be highly destructive not only in the immediate vicinity, but also over broader regions. These landslide-generated tsunamis can impact life, marine traffic, and the built and natural environment.

One area experiencing rapid glacier retreat since the mid-1800's is the Prince William Sound region in southcentral Alaska. Prince William Sound is surrounded by the glacier-covered Chugach Mountains, which rise from sea level to nearly 4,000 m over less than 20 km. This region has 25,000 km² of waterways and over 150 tidewater and valley glaciers.  The steep, snow-covered landscape is prone to landslides with the potential to generate tsunamis (referred to as tsunamigenic or “tsunami inducing”). One example of this cascading hazard is found in Barry Arm, where the retreat of the Barry Glacier has exposed the toe of a slow-moving landslide. Should the landslide undergo rapid catastrophic failure, it could generate a wave that would have devastating consequences and produce life-threatening conditions.

Since 2021, the USGS, National Tsunami Warning Center, Alaska Division of Geological & Geophysical Surveys, Alaska Earthquake Center, and other partners have dedicated time and resources to characterize the hazard posed by the Barry Arm landslide and develop a system to warn for a tsunami generated by its failure. In addition, the interagency team of scientists has focused efforts towards identifying other landslides in the broader Prince William Sound region, where the land-area total exceeds that of the state of Maryland.

This information allows for a better understanding of how slope stability conditions are changing over space and time due to glacial processes, climate change, and other factors that may contribute to increased landslide risks.  "The data in the report are intended to provide a preliminary means of assessing the potential hazard of the landslides using PSInSAR and to assist in the prioritization of future research, surveillance, and hazard assessment." said Lauren Schaefer, the lead USGS author and researcher. "Combined with aerial imagery and previous landslide inventory records, the addition of this technology continues to enhance our understanding and knowledge of landslides in the study area."

Within the area of interest, we identified and confirmed 43 landslides of variable sizes and velocities that moved between 2016 and 2022. Landslides identified in this report occur throughout Prince William Sound, but more are found in the western portion of the study area. 

Of these landslides, 14 were previously documented through other means (e.g., aerial imagery) although they were not necessarily identified as recently moving. Previously identified landslides are referred to by the names assigned by previous authors, and new landslides were named based on nearby geographic features. Of the identified landslides, our preliminary analysis suggests that 11 have tsunamigenic potential if they were to rapidly slide into adjacent waters. Of these, three are situated directly above water and the other eight are situated directly above glaciers or land with some distance to the nearest waterbody.

Although our estimate of tsunamigenic plausibility is preliminary and can be refined with additional observations and analyses, it can be used to prioritize ongoing and future hazard assessment, surveillance, and research efforts.

Getting a satellite's view of small changes

For years, researchers have gathered information through using aerial imagery from plane flights, using historical records and other research projects.  The PSInSAR effort adds another way to "see" different changes and ground deformation.  PSInSAR is most effective at detecting slow-moving changes that may not be readily visible.

Using satellite remote sensing data and a method known as persistent scatterer interferometric synthetic aperture radar (PSInSAR), data were collected. Persistent Scatterer Interferometry (PSI) is a powerful remote sensing tool that allows us to measure and monitor changes in the Earth’s surface over time. Using satellites, radar waves are reflected off the Earth’s surface, measure changes in the land surface altitudes, and then “bounce” back to the satellites. Imagery is produced by measuring the time it takes for the waves to travel back to the satellite. The taller the object on the surface, the faster the radar waves bounce back; a tall mountain produces a shorter bounce and a deep canyon, a longer one. By stitching the information together, we get a depiction of the Earth’s surface as measured by the radar. The larger the number of passes a satellite makes, the more images (called interfereograms) are collected for analysis.

We used PSInSAR to identify ground deformation in the Prince William Sound region using Sentinel-1 satellite images acquired during snow-free months (typically June-October) from 2016 through 2022.  Results were then compared with previously gathered aerial imagery and landslide inventory records, leading to a more comprehensive understanding of changes over time.

Sources/Usage: Public Domain. View Media Details

Using this technique allows us to study wide swaths of land while providing a high degree of resolution and detail.  Because satellites make multiple passes around the Earth, we can take multiple readings of the Earth’s surface and detect where there have been changes over time. This measurable shift, called displacement, tells us what has changed and where. The more images collected with the same sensors/equipment over shorter timeframes, the more accurately the rate and degree of change will be measured.

Having a lot of data points yields a more accurate depiction of how the landscape has changed over that time period. This is exceptionally valuable because the changes can occur at different speeds within a timeframe, helping us to identify if thing are speeding up or slowing down.

Another factor impacting our ability to measure changes in the landscape using PSInSAR is the landscape itself. It may be hard to collect data in vegetated, forested, snow-covered, and steep terrain, all of which are found in SE Alaska.  These landscape qualities can interfere with the radar “bounce”, leading to less accurate measurements. Despite some of these limitations, using satellites and other airborne craft provide benefits in that they can gather data in places where people can’t easily get to, such as the wild, rugged slopes of southeast Alaska. In turn, we get more high-quality information, albeit with some constraints, while reducing personnel time, costs and dangers.

One of the ways to increase accuracy is to collect data, make a lot of interferograms and look at information gathered over a longer time, six years in this case. We also confirmed landslide presence and estimated landslide area using a combination of ascending and descending radar tracks, terrain and landscape mapping using optical imagery and digital terrain data, and previous landslide inventories. (While slow-moving landslides do not necessarily represent a hazard, they have been seen to rapidly accelerate prior to catastrophic failure in other locations.)

“Although the estimate of tsunamigenic plausibility is preliminary and can be refined with additional observations and analyses, it can be used to prioritize ongoing and future hazard assessment, surveillance, and research efforts, “said Jonathan Godt, USGS Landslide Hazards Program Coordinator.  “The more information we gather about changes in the landscape around us, the better decisions we can make with regards to future risks.”