Seismic-based characterization for Volcano-Induced landslides: a feasibility study based on the 2018 Anak-Krakatau volcano flank collapse

The 22 December 2018 Anak-Krakatau volcanic flank collapse triggered a devastating tsunami along the Sunda Strait coast, highlighting the need for tsunami early warning systems to detect both volcanic collapse and landslide tsunami triggering from real-time seismic data. Seismically, landslides are deficient in high-frequency radiation compared to tectonic earthquakes. In this study, carried out in the framework of the TSUNAMI_RISK project, we implemented spectral analysis and back-projection to locate low-frequency events (40-70 s) along the Indonesian coast, specifically volcano-induced landslides. Using the Indonesian seismic network, we successfully detected and located the Anak-Krakatau volcanic flank collapse. The back-projection method allowed us to estimate the location about 100 seconds after the event initiation, with detection achieved with stations up to 1° from the source. To assess potential biases in our approach, we evaluated two weeks of continuous long-period seismic data (11-25 November 2019) from high-risk volcano tsunami areas in eastern Indonesia, such as Batu Tara, Iliwerung, Gamalama, and Sangeang Api. We conclude by discussing the expected false alarm rates for various warning thresholds based on the current network configuration in Indonesia and explore the routes to implementing this approach for operational tsunami early warning.
The 28th IUGG General Assembly (IUGG2023) (Berlin 2023)