Characterisation of precursory seismic activity towards early warning of landslides via semi-supervised learning
2025

Early Warning of Landslides Using Machine Learning

publication 10 minutes Evidence: moderate

Author Information

Author(s): Murray David, Stankovic Lina, Stankovic Vladimir, Pytharouli Stella, White Adrian, Dashwood Ben, Chambers Jonathan

Primary Institution: University of Strathclyde, Glasgow, UK

Hypothesis

Can machine learning from seismic data predict landslide events?

Conclusion

The study shows that machine learning can effectively predict landslide displacements weeks in advance by analyzing seismic signals.

Supporting Evidence

  • Seismic recordings can identify early stages of slope failure.
  • Machine learning can predict landslide displacements 2 to 4 weeks in advance.
  • The methodology adapts to discover new types of seismic events.

Takeaway

This study found that we can use special computer programs to listen to the ground and tell when a landslide might happen, giving us time to prepare.

Methodology

The study used a semi-supervised Siamese network to analyze seismic recordings from a single station over a period of 10 years.

Limitations

The study's findings are based on data from a single landslide site, which may limit generalizability to other locations.

Digital Object Identifier (DOI)

10.1038/s41598-024-84067-y

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication