Date of Award


Document Type

Open Access Dissertation

Degree Name

Doctor of Philosophy in Geophysics (PhD)

Administrative Home Department

Department of Geological and Mining Engineering and Sciences

Advisor 1

Gregory P. Waite

Committee Member 1

Wayne D. Pennington

Committee Member 2

Simon Carn

Committee Member 3

Diana C. Roman


Full-waveform moment tensor inversion of volcanic seismic signals and travel-time 3D tomography of local earthquakes have been widely used to explore source processes related to magma transport as well as to image the location and size of magma storage systems. However, the inversion solutions and the associated reliability estimates are non-unique and bear intrinsic uncertainties due to simplifying assumptions about the source, inaccuracies in the velocity models, dependence on network configuration, and other a priori constraints imposed by the modeler.

This work addresses the non-uniqueness and uncertainties of the model results by introducing non-linear inversion techniques that allow sampling the model space more effectively. We developed a nonlinear inversion approach for source type that uses a grid search over all possible moment tensor types and orientations to obtain a quantitative measure of the source mechanism reliability. For the tomography inverse problem, the solution space is fully explored using a ‘guided’ Monte-Carlo method in which starting velocity models are randomly selected and, through simulating annealing, only a subset of models that satisfies acceptability criteria is retained. Extensive synthetic tests are employed to test and validate the nonlinear inversion methods. The inversion procedures are then put into practice at Pacaya volcano, Guatemala. First, nonlinear moment tensor inversion is applied to explosion-related, long-period events that were recorded during a temporary installation of four broadband seismic stations in October-November 2013. The derived source reflects a shallow crack-like mechanism that is likely related to bubble-bursting events at the summit. Secondly, nonlinear travel-time local 3D tomography is employed to invert hundreds of local events that were detected during another temporary seismic network installation in January 2015. Re-location of the events using a 3D velocity model reveals the presence of a straight conduit possibly connecting a shallow magma reservoir to the surface.

The inversion approaches proposed in this study allow a comprehensive assessment of the model solution space. This is revealed to be of crucial aid in the determination of the confidence level of model interpretations, especially in cases like Pacaya, where availability of resources and observational data is limited.