Temporal changes in stress preceding the 2004–2008 eruption of Mount St. Helens, Washington Heather L. Lehto a, ⁎, Diana C. Roman a , Seth C. Moran b a University of South Florida, Department of Geology, 4202 E. Fowler Ave., SCA 528, Tampa, FL 33620-5201, United States b Cascades Volcano Observatory, United States Geological Survey, 1300 SE Cardinal Court, Building 10, Suite 100, Vancouver, WA 98683, United States abstract article info Article history: Received 28 May 2010 Accepted 12 August 2010 Available online 20 August 2010 Keywords: Mount St. Helens VT earthquakes fault plane solutions stress eruption forecasting stress tensor inversion Cascades The 2004–2008 eruption of Mount St. Helens (MSH), Washington, was preceded by a swarm of shallow volcano-tectonic earthquakes (VTs) that began on September 23, 2004. We calculated locations and fault- plane solutions (FPS) for shallow VTs recorded during a background period (January 1999 to July 2004) and during the early vent-clearing phase (September 23 to 29, 2004) of the 2004–2008 eruption. FPS show normal and strike-slip faulting during the background period and on September 23; strike-slip and reverse faulting on September 24; and a mixture of strike-slip, reverse, and normal faulting on September 25–29. The orientation of σ 1 beneath MSH, as estimated from stress tensor inversions, was found to be sub-horizontal for all periods and oriented NE–SW during the background period, NW–SE on September 24, and NE–SW on September 25–29. We suggest that the ephemeral ~90° change in σ 1 orientation was due to intrusion and inflation of a NE–SW-oriented dike in the shallow crust prior to the eruption onset. © 2010 Elsevier B.V. All rights reserved. 1. Background and motivation Significant seismic activity has been observed at Mount St. Helens (MSH), Washington, USA, during both non-eruptive and eruptive periods since multi-station seismic monitoring began in 1980 (Moran et al., 2008). Seismicity at MSH has been monitored since 1980 by a dense (14–20 stations) network of single-component, short-period seismometers (Fig. 1). Shallow (b 3 km depth; all depths are referenced to a datum of 2.2 km ASL, the altitude of the highest MSH seismic station (Thelen et al., 2008)) seismicity at MSH prior to the 2004–2008 unrest has been characterized by swarms recorded during dome-building and phreatic events from 1980–1986, and a continuously active cluster of seismicity located at ~3 km depth between mid-1987 to September 2004 (Fig. 4 of Moran et al., 2008). Swarms of deeper events have occurred in 1980, 1987–1992, 1996, 1998, and 2002 (Moran et al., 2008). Unrest preceding the 2004–2008 eruption at MSH began on September 23, 2004, with a swarm of shallow volcano-tectonic earth- quakes (VTs, which have dominant frequencies N 5 Hz and clear P- and S- waves; Lahr et al., 1994; Moran et al., 2008) accompanied by deeper (~5– 12 km depth) deflation of the volcano (Dzurisin et al., 2008; Lisowski et al., 2008). Precursory VTs recorded on September 23 are located at a depth of 2 km using the 1-D velocity model of Thelen et al. (2008), indicating a sudden ~1 km upward shift in the depths of hypocenters at the onset of unrest. The rate of VTs increased through September 23, reached a maximum on September 24, and then declined early on September 25 (Moran et al., 2008). The first long period (LP) events (which have dominant frequencies b 5 Hz and emergent waveforms; Lahr et al., 1994) were observed on September 25, and a gradual transition from mostly VTs to a mix of hybrid events (which have high-frequency onsets and low- frequency codas; Lahr et al., 1994) and LP events occurred between September 25 and October 5 (Moran et al., 2008). This decrease in VT activity and concurrent increase in the number of LP and hybrid events likely marks the complete formation of the fault structure accommodating magmatic intrusion and extrusion. The first phreatic explosion occurred on October 1, followed by the extrusion of dacite lava beginning on October 11 and continuing through January 2008. Analyses of VT fault-plane solutions (FPS) have been used to investigate systematic changes in the orientation of the local stress field at MSH (e.g., Barker and Malone, 1991; Moran, 1994; Musumeci et al., 2000) as well as at other restless volcanoes (e.g., Roman and Cashman, 2006). Such studies frequently reveal ~ 90° changes in the orientation of the principal stress axes that appear to reflect magma intrusion (e.g., Roman and Cashman, 2006, and references therein). The ~90° rotation of the stress field is believed to result from the inflation of a magma-filled dike as it ascends through the brittle crust (Fig. 2). In this model, a dike inflates in the direction of regional or background minimum compression (σ 3 ), inducing a local stress field in which local maximum compression is orientated perpendicular to regional or background maximum compression (σ 1 ). Stresses from dike inflation produce VTs with FPS whose orientations reflect a local stress field that is rotated by ~90° compared to the regional or background stress field. Previously published analyses of FPS for mid- crustal (5–12 km deep) VTs recorded at MSH between 1980 and 1998 Journal of Volcanology and Geothermal Research 198 (2010) 129–142 ⁎ Corresponding author. E-mail address: hlehto@mail.usf.edu (H.L. Lehto). 0377-0273/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.jvolgeores.2010.08.015 Contents lists available at ScienceDirect Journal of Volcanology and Geothermal Research journal homepage: www.elsevier.com/locate/jvolgeores