Author: Gwyneth Retta Hughes

Publisher:

Published: 2010

Total Pages:

ISBN-13:

Approximately 10% of the global population, some 550 million people, live within 100 km of an active volcano, making it imperative that the causes of magma accumulation in the crust and the factors affecting the scale of subsequent volcanic eruptions are investigated and modeled. This PhD dissertation examines magmatic processes at two scales: silicic calderas represent the large but infrequent end of the continuum, whereas mafic dikes are small in scale but common. Many previous authors have presented hypotheses for how large silicic systems develop. In order to test and refine these hypotheses, I undertook a global compilation that empirically examines how the characteristics of 140 young silicic calderas reflect their crustal-tectonic setting. Results indicate that the size and geochemistry of silicic calderas are affected by the nature of the underlying crust, the tectonic setting, and the local stress regime. For example, large, rhyolitic calderas tend to occur in continental settings under extension. There are, however, few true "rules, " and exceptions may prove useful in analyzing how silicic magma chambers form. Based on this compilation, I present a probabilistic method for determining the tectonic-crustal setting of a given caldera from its diameter, and eruption geochemistry. Focusing specifically on arc settings, this study demonstrates that (1) the abundance of silicic calderas in a given arc is proportional to the trench-normal convergence rate, except in arcs with back-arc spreading; and (2) silicic calderas in continental arcs tend to occur farther behind the volcanic front than do more typical arc volcanoes, possibly because of the abundance of pre-existing structures in continental margins. At the opposite end of the volcanic spectrum, this dissertation examines the intrusion of two mafic dikes. The first lies beneath Mammoth Mountain, California, and was associated with a 1989 seismic swarm. Based on an inversion of leveling data constrained by relocated earthquakes, I propose that a dike 2 km long, 8 km high, with 1 m of opening was intruded at 9 km depth beneath the south side of Mammoth Mountain. The second dike investigation focused on the eruption of Miyakejima, Japan, in 2000, associated with more than 10,000 earthquakes, several small eruptions and progressive caldera collapse. Displacements recorded by GPS stations and pre- and syn- event seismicity were used to determine a geological explanation of the event. In the proposed model, a shallow dike propagated ~30 km away from Miyakejima for one week, stopped propagating laterally after intersecting a pre-existing fault zone, and then continued to open and grow vertically for nearly two months.