Using the Horizontal-to-Vertical Spectral Ratio Method to Estimate Thickness of the Barry Arm Landslide, Prince William Sound, Alaska

Open-File Report 2024-1071
Landslide Hazards Program
By: , and 

Links

Abstract

Conducting detailed investigations of large landslides is difficult, especially in the subsurface, largely due to environmental factors such as steep slopes, difficult access, and numerous objective hazards. These factors have made it challenging to accurately estimate the depth to the failure surface of the Barry Arm landslide, a large (roughly 108 cubic meters), deep-seated bedrock landslide in Prince William Sound, Alaska, recognized in 2019. The landslide has exhibited accelerated movement in recent years and poses a potential tsunamigenic hazard if rapid failure occurs. Failure surface depth, equivalent to landslide thickness, is a necessary metric for landslide-volume calculations and associated tsunami wave models. In this report, we used seismic noise recorded by a seismometer located on the Barry Arm landslide in Alaska to calculate the horizontal-to-vertical spectral ratio (HVSR) to investigate the site fundamental frequency (f0) and depth of the failure surface. To ensure that observed peak frequencies in the spectral ratio were related to the underlying stratigraphy (and not caused by other noise sources like nearby glaciers, topographic resonance, weather, or human activities), we also calculated HVSRs using earthquake signals, HVSRs at other seismic stations within a 2.5-kilometer radius, and a standard spectral ratio between the landslide station and other sites. We observed multiple peaks in the landslide HVSR curves at 1.5 hertz (Hz), 4–5 Hz, and 7–11 Hz. The frequencies of these peaks were consistent at the landslide site through time and across methods and were dissimilar to those identified at other seismic stations in the area, making it unlikely the peaks were caused by local noise.

Directional HVSRs calculated at 15-degree intervals showed amplification of the higher frequency peaks in the direction parallel to slip, indicating two-dimensional site effects. We used the distinct frequency peaks in the seismic record to develop a 4-layer conceptual model of the landslide wherein the top of the deepest layer represents the primary failure surface, or the boundary between damaged (mobile) and undamaged material. We inverted Rayleigh wave ellipticity curves within this 4-layer configuration with constraints on S-wave velocity and layer thickness based on analogous material properties identified in the literature. This was necessary absent any site-specific subsurface S-wave velocity data. The best-fitting models indicate a mean slope-normal depth to the failure surface of 188 (±9) meters (m), with additional stratigraphic boundaries at 4 and 20 m below ground surface, potentially representing layered motion. These results agree with and improve upon ranges estimated by previous studies and can support future modeling and assessment efforts at Barry Arm.

Suggested Citation

Collins, A.L., Allstadt, K.E., and Staley, D.M., 2024, Using the horizontal-to-vertical spectral ratio method to estimate thickness of the Barry Arm landslide, Prince William Sound, Alaska: U.S. Geological Survey Open-File Report 2024–1071, 25 p., https://doi.org/10.3133/ofr20241071.

ISSN: 2331-1258 (online)

Study Area

Table of Contents

  • Acknowledgments
  • Abstract
  • Introduction
  • Site Setting
  • Methods
  • Results and Discussion
  • Conclusion
  • References Cited
Publication type Report
Publication Subtype USGS Numbered Series
Title Using the horizontal-to-vertical spectral ratio method to estimate thickness of the Barry Arm landslide, Prince William Sound, Alaska
Series title Open-File Report
Series number 2024-1071
DOI 10.3133/ofr20241071
Year Published 2024
Language English
Publisher U.S. Geological Survey
Publisher location Reston VA
Contributing office(s) Geologic Hazards Science Center - Landslides / Earthquake Geology
Description vii, 25 p.
Country United States
State Alaska
Other Geospatial Barry Arm landslide
Online Only (Y/N) Y
Google Analytic Metrics Metrics page
Additional publication details