JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 99, NO. B3, PAGES 4341-4354, MARCH 10, 1994 Seismicity at Mount St. Helens, 1987-1992: Evidence for repressurization of an active magmaticsystem Seth C. Moran Geophysics Program, University of Washington, Seattle Abstract. Several changes have occurred in seismicity associated with Mount St. Helens since it last erupted in October 1986. Perhaps themost significant is therecurrence of earthquakes deeper than 3 km, previously observed only in themonths following thecataclysmic eruption of May 18, 1980, and briefly in March1982. Events located below 6.5 km define a circular aseismic zone,similarin location to one observed afterMay 18, 1980. Focal mechanisms are computed for twodepth ranges, 4-6.5km (Intermediate) and 6.5-10 km (Deep), using both polarity and amplitude ratio data. Mechanisms for Intermediate events aremostly strike-slip, and many P and T axes point in directions --80 ø offset fromtheregional stress field, indicating that earthquakes in this zone are caused by magmatic aswell astectonic forces.Mechanisms for Deep events arealso mostly strike-slip, and P and T axes forma pattern suggestive of a pressure increase within theaseismic zone. The observed Deep patterns arecompared with theoretical stress fields generated by a numerical model of a pressurized hole in an elastic plate. Results from modeling support the pressure increase hypothesis for 1987-1992 seismicity, in contrast to a pressure drop for post-May 18, 1980, earthquakes. Thisapparent repressuriza- tionis proposed to be a result of the sealing of the shallow conduit system. Introduction Mount St. Helens (MSH), a Quaternary dacitic-andesite vol- cano located in southwestern Washington, last erupted in October1986. Sincethis last eruption therehave beena number of changes in seismic activity associated with the volcano.The significance of these changes in terms of the evolution of the MSH magmatic system and possible future eruptiveactivity is, however, unclear. In this paper I attemptto address this ques- tion, usingobservations and results from a stress field studyto contrast the 1980-1986 and 1987-1992 time periods.In particu- lar, I continue a studyby Barker and Malone [1991 ], who used a cylindrical magma chamber model to attempt to reproduce focal mechanism patterns observed immediately after the cata- clysmiceruption of May 18, 1980. I performa similar analysis for the 1987-1992 data set, remodel their 1980 data set, and use these results in conjunction with other studies to propose a modelfor changes in the magma conduit system since1986. Seismicity,1980-1992 Thirteen years of seismic history have accumulated at MSH sincethe Univerisityof Washington (UW) digital recording sys- tem began operation in early 1980. Duringthis time roughly 8000 digitally recorded events have been located beneaththe volcano. In this paper5347 hypocenters are presented, includ- ing only those eventswith picked phasearrivals at sevenor morestations, azimuthal gaps of less than 135 ø , relative vertical errorsof lessthan 1 km, and estimated relative epicentral errors of lessthan0.5 km. Figure 1 summarizes this record. Earth- quakes are located by using a one-dimensional layeredvelocity Copyright 1994by the American Geophysical Union. Paper number 93JB02993. 0148-0227/94/93J B-02993 $05.00 model (Figure 2). The MSH station configuration has changed duringthis time. However,of the 17 current stations within 50 km of MSH, 7 were in operationby May 1980, 12 by Sep- tember 1981, and 15 by 1984. Thus locationdifferences associ- ated with changing station configuration are thought to be small, and comparisons on the scale madein this paperare legitimate. Figure 1 shows several changes in seismicity from 1980-1986 to 1987-1992. Perhaps the most obvious is a gradualincrease in seismic activity beneath the volcano,from an averageof-2 locatable events per month in the first half of 1987 to a peakof -90 per month in the first half of 1990. Seismicactivity has sincedeclined,graduallydropping to an average of-20 events per month in the latterhalf of 1992. Increases in seismicity fre- quently precede eruptions at volcanic centers and are thought to be related to stress changes due to movement of magma beneath the surface. The recentMSH episode differs from "typical"pre- cursory patterns observed at MSH and othervolcanic centers in two important respects. First, many of the recent earthquakes have been deep. Earthquakes in precursory swarms commonly are almost exclusively shallow; at MSH only the May 18, 1980, and March 1982 precursory swarms contained earthquakes with hypocenters deeperthan 3 km [Weaver et al., 1983; Jonientz- Trislerand Zollweg, 1987]. Second, the increase took placeover 3 years, not the days-to-weeks timescale typical of precursory swarmso Thus the significance of the recentseismic increase is unclear. Another changewas the recurrence of deep (> 3 km) earth- quakes beneath the volcano. After the summer of 1980, most seismic activity was restricted to shallow(< 3 km) depths, with the exception of a 2-monthswarmof small, deep earthquakes pre6eding the explosive March 1982 eruption[Weaver et al., 1983]. Several months after the last eruption in Octobert986, however, earthquakes deeper than 3 km were againobserved. Earthquakes first occurred in two clusters, one centered around 2 km depth and the other between 4 and 6 km (Figure 2). 4341