Journal
GEOPHYSICAL JOURNAL INTERNATIONAL
Volume 219, Issue 2, Pages 1138-1147Publisher
OXFORD UNIV PRESS
DOI: 10.1093/gji/ggz354
Keywords
North America; Numerical solutions
Categories
Funding
- French National Research Agency (ANR) [ANR-15-CE04-0009, ANR-17-CE23-0015-01]
- TelluS-ALEAS program of the French National Institute of Sciences of the Universe (CNRS-INSU)
Ask authors/readers for more resources
Quantifying landslide activity in remote regions is difficult because of the numerous complications that prevent direct landslide observations. However, building exhaustive landslide catalogues is critical to document and assess the impacts of climate change on landslide activity such as increasing precipitation, glacial retreat and permafrost thawing, which are thought to be strong drivers of the destabilization of large parts of the high-latitude/altitude regions of the Earth. In this study, we take advantage of the capability offered by seismological observations to continuously and remotely record landslide occurrences at regional scales. We developed a new automated machine learning processing chain, based on the Random Forest classifier, able to automatically detect and identify landslide seismic signals in continuous seismic records. We processed two decades of continuous seismological observations acquired by the Alaskan seismic networks. This allowed detection of 5087 potential landslides over a period of 22 yr (1995-2017). We observe an increase in the number of landslides for the period and discuss the possible causes.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available