Skip to content

A streamlined tephra inversion workflow in the cloud with VICTOR

Qingyuan Yang, Jim Yang, Samuel Krasnoff, Einat Lev

Abstract

The inversion of field observations to infer eruption parameters for past eruptions is a fundamental tool in volcanology. Quantitative inversion requires understanding of statistics, data science, and the limitations of the forward model, a challenge in interdisciplinary research, collaboration, and education. It is thus beneficial to provide the community with access to streamlined workflows for inversion.  Here, we couple the Metropolis-Hastings algorithm with the volcanic ash transport model Tephra2, and present the coupled algorithm as a new method to estimate the Eruption Source Parameters of volcanic eruptions based on mass per unit area or thickness measurements of tephra fall deposits. Outputs of the algorithm are presented as sample posterior distributions for variables of interest. We demonstrate the algorithm and the workflow works for both synthetic data and observations from real eruptions. As a Markov chain Monte Carlo (MCMC) algorithm, Metropolis-Hastings incorporates prior knowledge, quantifies the uncertainty, captures correlations between parameters, and assumes no simplification in sampling from the posterior probability distribution The workflow is implemented on the Volcanology Infrastructure for Computational Tools and Resources (VICTOR), a cloud-based cyberinfrastructure platform supporting the volcanology community. It is streamlined through a Jupyter Notebook, guiding the user through parameter and data input and output analysis. Basic elements in the algorithm and how to implement it are introduced throughout the notebook. By utilizing the features of the shared platform VICTOR, it openly shows the inside of the inversion "black box" in a simplified way, making it accessible for both experienced researchers and students