AmericanChemicalSociety.com

Results from a study of shear-wave splitting in regional and local earthquakes recorded by a broadband seismic network on the island of Montserrat have confirmed an hypothesized link between changes in fluid pressure within an active magmatic conduit system and changes in the orientation of local stresses within the rock surrounding the conduit. Montserrat is home to the Soufriere Hills Volcano, which has been erupting since 1995 in a sequence of ~1-4 year-long phases separated by periods of quiescence, making it an ideal place to study temporal changes in stress between periods of eruption and quiescence. A network of broadband seismometers has operated on Montserrat since October 1996, making it possible to analyze shear-wave splitting (from which the orientation of maximum horizontal compression. SH_max, can be determined) prior to and during phases of eruption from 1996-1998, 1999-2003, and 2005-present. We have completed analysis of waveforms of regional earthquakes recorded on a) four broadband seismometers on Montserrat (MBBY, MBGB, MBGH, MBRY), which operated nearly continuously between October 1996 and September 2007, and b) two three-component seismometers (SKI and BPA), operating during the same period and located on the neighboring islands of St. Kitts and Antigua. By including analyses from off-island stations in our study, we are able to assess whether any observed stress field reorientations are related to local (volcanic) or regional (tectonic) phenomena.

Results of the shear-wave splitting analysis provide strong evidence for a volcanically-induced change in the local stress field orientation beneath Montserrat during the study period. For regional stations SKI and BPA, measurements of SH_max are oriented predominantly NE-SW, or arc-normal, and there is no indication of a temporal change in SH_max beneath St. Kitts or Antigua during the study period. In contrast, combined measurements of SH_max from local (Montserrat) seismic stations show a far more complex pattern which varies contemporaneously with previously observed changes in VT FPS P-axis orientations through early 2005. A closer examination of SH_max on individual stations during April 1998-April 2005 (when the highest number of measurements were obtained) indicates that local stress field changes during May-November 1999 (the six months leading up the onset of the second phase of the eruption) are station-dependent. Stations MBGB and MBGH, both located northwest of the Soufriere Hills vent, record a ~90° change in SH_max starting in May 1999, station MBRY (east of the vent) records a ~70° change in SH_max, and station MBBY (southwest of the vent) does not record a rotation. We fit a numerical model of SH_max orientations and magnitudes produced by an inflating dike by maximizing agreement between modeled SH_max trajectories to the distribution of SH_max observed on Montserrat during May-November 1999. The best-fit model is a 1-km-long, meter-wide dike oriented N30°E, roughly parallel to the regional orientation of maximum compressive stress. The strong agreement between modeled SH_max vectors and SH_max observations during May-November 1999 indicates that the local stress field reorientation apparent in both shear-wave splitting analysis results and VT earthquake FPS data may be confidently attributed to inflation of a NE-trending dike during the six months preceding the onset of eruption in November 1999.

This study has had a major impact on my career in that the results have clearly justified a greater attempt to understand stress changes beneath active volcanoes, as well as the state of stress beneath quiescent volcanoes (a critical piece of information for using volcanic stress field analysis as a forecasting tool.) As a result of the work funded by this starter grant, I have begun a major NSF-funded study of changes in stress beneath Aleutian volcanoes, with the aim of understanding 'failed' volcanic eruptions, or stalled mid-crustal intrusions of magma. This study involves two graduate students who have been given responsibility for major components of the study. A second phase of the PRF-funded study that is now in progress is a study of background stress field orientation at Crater Lake, Oregon. The aim of this study is to critically examine the assumption that regional stress field orientation can be taken as a proxy for local background stress field orientation at a volcano, and hopefully to demonstrate the need for focused baseline stress studies at potentially active volcanoes. Finally, my colleagues and I are now in the process of preparing and submitting manuscripts on our results from Soufriere Hills for publication.