Skip to content

Eruption probabilities from seismic data assimilation: Insights from the 2023 paroxysms of Shishaldin Volcano

Társilo Girona 1, Kyungmin Kim1, Matthew Haney2, David Fee1, John Power2, Taryn Lopez1

  • Affiliations: 1Alaska Volcano Observatory, Geophysical Institute, University of Alaska Fairbanks, Fairbanks, AK, USA. 2US Geological Survey (USGS), Alaska Volcano Observatory, Anchorage, AK, USA. 

  • Presentation type: Poster

  • Presentation time: Monday 16:30 - 18:30, Room Poster Hall

  • Poster Board Number: 272

  • Programme No: 2.4.43

  • Theme 2 > Session 4

Abstract

Connecting geophysical observables to subsurface processes is crucial for interpreting volcanic unrest and forecasting eruptions. Tremor, a persistent ground vibration often recorded at active volcanoes, is a key pre-eruptive indicator. Variations in tremor, such as changes in dominant frequency, overtones, or seismic amplitude, are sometimes observed before eruptions. However, similar changes can also occur during quiescence or declining activity, raising questions about how tremor reflects the physical and geometrical conditions of shallow volcanic conduits and whether it can reliably indicate pressure changes and eruption probabilities. To address this, we introduce a novel data assimilation framework that integrates seismic data, a physics-based tremor model, an inversion method using Reversible Jump Markov Chain Monte Carlo (RJ-MCMC) to constrain model parameters, and a genetic algorithm for real-time optimization of eruption probabilities. Our model, extending from Girona et al. (2019), posits that tremor arises from gas entering shallow magma systems, accumulating temporarily in the conduit, and ultimately escaping via permeable flow. We tested this framework on the 13 paroxysms of Shishaldin Volcano, Alaska, between July and November 2023, all preceded by tremor variations. Our results suggest these events were driven by a combination of magma ascent, increased gas flux (consistent with satellite observations), and partial conduit sealing, all leading to pressure rises of several MPa and increased eruption probabilities within hours. This approach demonstrates the potential for near-real-time, physics-based eruption forecasting using tremor data. Girona, T., Caudron, C., & Huber, C. (2019), https://doi.org/10.1029/2019JB017482.