← Back to Sessions

Characterization of the Onset of the 2021 Great Sitkin Dome-Building Eruption Through the Trans-Dimensional Bayesian Inversion of LP Seismicity

Description: 

Long-period (LP) volcanic seismicity and tremor are thought to be triggered by the subsurface transport of magmatic and/or hydrothermal fluids. A mechanistic model of LP events and tremor, called the “leaky gas pocket model”, was proposed by Girona et al. (2019). This model shows that the transport of gas through permeable materials (e.g., lava domes) triggers spontaneous LP/tremor-like pressure oscillations. Building upon the leaky gas pocket model, our project addresses two key questions regarding LP seismicity and its relation to volcanic activity: (1) To what extent can the leaky gas pocket model accurately elucidate the physical properties of a volcanic dome? (2) Can the variation of physical properties during the onset of a dome-building eruption be explained by the inversion of LP seismicity? To explore these questions, we integrate the leaky gas pocket model with a reversible-jump (trans-dimensional) Markov Chain Monte Carlo (RJ-MCMC) framework to invert LP seismicity recorded during the early-stages of the 2021 dome-building eruption of Great Sitkin volcano (Alaska). With this approach, we aim to comprehensively quantify the probability density function of the model parameters for each LP event; these parameters include dome thickness and permeability, gas flow rate and temperature, and seismic attenuation factor, among others. Our preliminary results suggest a gradual accumulation of gas beneath the dome and a possible decrease of dome thickness around the start of the eruption (July 23, 2021). This is consistent with a pre-eruptive gas accumulation and pressurization in the shallow volcanic conduit, which we determine in the range from ~10² Pa to ~10⁴ Pa. The gas flow rate is constrained within the range from 5 ton/day to 200 ton/day, representing the minimum amount of gas supplied to the base of the dome (more gas might be released with no contribution to seismicity). Our analysis represents a pioneering physics-based approach to infer the properties and conditions of volcanic interiors from the inversion of seismic data.

Session: Multidisciplinary Approaches for Volcanic Eruption Forecasting [Poster Session]

Type: Poster

Date: 5/2/2024

Presentation Time: 08:00 AM (local time)

Presenting Author: Kyungmin

Student Presenter: Yes

Invited Presentation: 

Authors