Deciphering Okmok Volcano's restless years (2002-2005)

• Main Author: Reyes, Celso Guillermo
• Other Authors: McNutt, Stephen, West, Michael, Freymueller, Jeffrey, Larsen, Jessica
• Format: Doctoral or Postdoctoral Thesis
• Language: English
• Published: 2015
• Subjects: Fairbanks; Alaska; Aleutian Islands
• Online Access: http://hdl.handle.net/11122/6120

Dissertation (Ph.D.) University of Alaska Fairbanks, 2015 Okmok Volcano is an active island-arc shield volcano located in the central Aleutian islands of Alaska. It is defined by a 10-km-diameter caldera that formed in two cataclysmic eruptions, the most recent being ~2050 years ago. Subsequent eruptions created several cinder cones within the caldera. The youngest of these, Cone A, was the active vent from 1815 through its 1997 eruption. On July 12 2008 Okmok erupted from new vents located northwest of Cone D. Between 2001 and 2004, geodetic measurements showed caldera inflation. These studies suggested that new magma might be entering the system. In 2002, a newly installed seismic network recorded quasi-periodic ("banded") seismic tremor signals occurring at the rate of two or more episodes per hour. This tremor was a near-continuous signal from the day the seismic network was installed. Although the volcano was not erupting, it was clearly in a state of unrest. This unrest garnered considerable attention because the volcano had erupted just six years prior. The seismic tremor potentially held insight as to whether the unrest was a remnant of the 1997 eruption, or whether it signaled a possible rejuvenation of activity and the potential for eruption. To determine the root cause and implications of this remarkable seismic tremor sequence, I created a catalog of over ~17,000 tremor events recorded between 2003 and mid-2005. Tremor patterns evolved on the scale of days, but remained the dominant seismic signal. In order to facilitate the analysis of several years of data I created a MATLAB toolbox, known as "The Waveform Suite". This toolbox made it feasible for me to work with several years of digital data and forego my introductory analyses that were based on paper "helicorder" records. I first attempted to locate the tremor using the relative amplitudes of the seismograms to determine where the tremor was being created. Candidate tremor locations were constrained to a few locations along a corridor between ...