Dr. Leif Karlstrom has been granted an NSF Earth Sciences postdoctoral fellowship to carry out a research and education plan at Stanford University. He will investigate magma transport processes in the crust on two different scales through modeling of volcanic tremor and statistical inference of arc-scale eruption patterns. The tremor work aims to identify available mechanisms for the origin of seismicity during volcanic eruptions, and will examine long period seismic signatures of different eruption styles with a coupled fluid flow and elastodynamic numerical model. The larger scale aspect of this work will test models of magmatic plumbing with Quaternary eruptions in the Cascades arc. Network inference techniques and Bayesian statistics will be applied to a newly compiled dataset of recent volcanism throughout the Cascades. The goal of this study is to test for the degree to which magma transport is controlled by dynamic internal organization versus external forcing (glaciation, spatially variable tectonic stresses).
Magma transport is relevant both for understanding Earth history and volcano-human interactions, but it is largely inaccessible to direct observation. As a result, the mechanics of rising and cooling magmas that generate episodic volcanism are poorly constrained. This work probes the deep structure of volcanoes by modeling patterns of radiated seismic waves during eruptions, and using statistical inference techniques to probe the structure of magma plumbing throughout the Cascades volcanic arc. These studies ask the general question: to what extent can theoretical models represent or predict indirect, remote observations? This question has important implications for the degree to which models and data can be synthesized to interpret the geologic record and assess volcanic hazards. Educational activities associated with this project include leading a field trip to Lassen National Park and co-teaching a course on physical volcanology at Stanford.
The transport and storage of magma in the Earth’s crust is responsible generating all igneous rocks exposed at the surface, volcanic landforms and active volcanic eruptions. This project addresses two components of the crustal magma transport system on two quite different spatial and time scales, and using two quite different research tools. Although connected by the common thread of magma transport, these projects address quite aspects magmatism from both a modeling and data analysis perspective. First, on the relatively short timescales associated with active volcanic episodes, I have developed models for the excitation of volcanic conduits by external perturbations with the aim of understanding volcanic tremor (seismic waves generated by volcanic processes). A volcanic conduit in general contains a column of heavily stratified, multiphase fluid, and the propagation of waves through this fluid can be quite complex. This work endeavors to define the characteristic resonant modes of a volcanic conduit under various conditions corresponding to either quiescent or active erupting states. Similar to the case of an organ pipe, the normal modes of a volcanic conduit may be excited by perturbations and the conduit may resonate, generating seismic waves or infrasonic waves that may be observed remotely. Thus, even though they are hidden from view, we may learn about the interior dynamics of volcanic conduits by understanding the characteristic modes of these structures. As Figure 1 indicates, impulsive perturbations such as might occur from rockfall events may excite waves of very complicated character within the conduit, but our models provide a means to understand these perturbations and thus may be useful for hazard forecasting as well as basic physical volcanology research. The second part of this project involves characterizing the distribution of volcanic vents in the Cascades arc (California, Oregon and Washington) that are evidence of eruptions or eruptive centers over the last 4-5 million years. Volcanic vents are the surface signatures of the deeper but hidden crustal magma plumbing system, and the distribution as well as products from volcanic vents provides insight into the spatial and temporal evolution of the crustal magma transport system. I have over the course of this project assembled the most complete dataset to date of distributed Cascades volcanism, including ~2800 individual vents, major element geochemistry of these vents, and auxiliary datasets such as heat flow and crustal seismic velocity at each vent location. In total this new dataset comprises 21 variables at each vent. Such a large dataset is difficult to interpret, so I have developed statistical techniques to analyze the dataset and look for structure that might illuminate controls on Cascades volcanism on the longest spatial and timescales. These techniques involve clustering, to objectively define grouping of vents or vents and associated products, as well statistical methods that look for the most correlated components of the dataset. Although this work is ongoing, a major finding is that surface vent distribution correlates most well with seismic structure of the crust and heat flow, suggesting that the spatial pattern of vents does in fact vary systematically as the structures at depth vary. Figure 2 shows the spatial distribution of vents in the Cascades used in my dataset, along with a representative (average) seismic shear velocity structure from a tomorgraphic model (Porritt et al., 2011).
