Return to the Table of Contents for this chapter.
*Present address: University of Memphis, Center for Earthquake Research and Information, Memphis, Tennessee 38152
The Mount Erebus Volcano Observatory project was established in 1992 to provide digital monitoring and recording of seismic activity associated with the volcanism of Mount Erebus. The volcano is the central feature of Ross Island, with the summit crater located approximately 35 kilometers (km) from McMurdo Station. A permanent convecting crater lake of anorthoclase phonolite magma produces frequent small strombolian eruptions (Kyle 1994). Data from the radiotelemetered seismic network ( figure 1) are digitized and recorded at 100 samples per second using an event-triggered PC-based detection and recording system at McMurdo Station. Data are compressed and automatically transferred daily over the Internet to New Mexico Institute of Mining and Technology and Victoria University for analysis (Skov, Kyle, and Aster 1994). During the study period (November 1994 through June 1996), over 3,000 digitally recorded events were recorded and examined. Continuous analog helicorders are also operated at McMurdo, and a count of events recorded during 1995 was used to examine digital data acquisition efficiency ( figure 2). Analog records indicate that, although many seismic events currently occurring on Erebus are of insufficient size to trigger the digital acquisition system, the larger events which trigger the system do provide a representative sample of the frequency and style of the overall seismic activity. At least four categories of seismic events are recognized: long-period (LP) events, volcano-tectonic (VT) events (Chouet 1996), explosive (E) events (Dibble, O'Brien, and Rowe 1994), and long-period tremor (Chouet 1985). The system also records near-regional, regional, and teleseismic signals originating far outside of the array.
Preliminary locations for the three principal seismic event types ( figure 3) were obtained using a gradient-over-a-half-space velocity model estimated from refraction data by Rowe (1988). The events plotted met the following requirements.
VT events are indicative of the faulting of brittle rock due to perturbations caused by the intrusion or withdrawal of fluids superimposed on the regional stress field. Because they originate in abrupt shear motion along faults, VT events exhibit sharp arrivals for both compressional-wave and shear-wave phases at multiple stations. Only 87 out of the more than 3,000 events examined show clear VT characteristics. We estimate the relative sizes of seismic sources using a simple scale based on the duration of the seismic signal. The mean duration magnitude of located VT events is 1.04 and the majority occur 1 km or more below the summit (figure 3, top). VT events do not occur in temporal swarms, arguing for steady-state stress conditions in the lower volcano.
LP events are related to magmatic source processes and are characterized by emergent first arrivals, relatively long-period sources [typically below approximately 2 hertz (Hz)], and a general deficit of identifiable shear-wave arrivals. LP events make up the largest percentage of the total data set (approximately 65 percent). Over 500 LP events occurred during seismic swarms, accounting for more than one-quarter of the total number of LP events. LP events are commonly difficult to locate accurately due to highly emergent onsets. Additional complications arise in locating events from LP swarms due to the interleaving of arrivals from multiple events, and swarm location estimates are thus not presently included in the location catalog. LP event locations are scattered and are predominantly distributed throughout the upper 1 km of the mountain (figure 3, middle), suggesting that the upper mountain is criss-crossed with numerous small pathways and conduits for movement of magma. The mean duration magnitude of located LP events is 1.97.
E events are directly attributable to volcanic explosions at or beneath the surface of the lava lake. Low-frequency precursors frequently accompany E events and may be caused by oscillations of exsolved gas bubbles prior to the explosion (Dibble et al. 1994, pp. 1-16). E events constitute approximately 20 percent of the total data set. Like LP events, E events are observed to form temporal swarms; over 230 E events occurred during two main periods of explosion swarms, accounting for over one-third of all such activity. E sources locate in the summit lava lake area and in the upper portion of the main conduit (figure 3, bottom). The mean duration magnitude of E events for which a coda could be picked is 0.74. This number is unrepresentative of the population of larger explosion events, however, because the current trigger parameters terminate many of the digital records prior to the time that the coda fades into the background.
Two especially notable swarms of E events occurred during the study period. On 14 April 1996 (julian day 105; figure 2) a swarm of 100 triggered E events occurred near the active crater. A second swarm of 165 triggered E events began on 20 April 1996 (julian day 111; figure 2). Earthquake occurrence rate versus magnitude statistics for these two swarms indicates that a relatively large number of small events occur during E swarms compared to most types of seismic activity, a feature previously noted by Dibble et al. (1984).
An episode of long-period harmonic tremor was recorded on 30 January 1995. The dominant frequency of the tremor changed during the 5 minutes that the tremor was recorded from approximately 5 Hz at the start of the record to approximately 1.5-2 Hz after several minutes. This change could be due to a widening conduit or to changing flow conditions. The dominant interpretation of such tremor episodes (e.g., Chouet 1985) is that they result from the superposed resonant responses of a volcanic conduit to numerous individual pressure pulses during periods of magmatic transport. Thus, it is possible that this episode represents a significant upwelling of magma from depth. The observation that the tremor was clearly recorded even at the McMurdo seismic station, located 35 km from the summit, indicates a deep focus and large-amplitude source.
We thank the Antarctic Support Associates technicians Joe Longo and Joe Pettit for their year-long assistance in the Crary Laboratory at McMurdo Station and in the field during summer fieldwork. Kurt Panter and Ken Sims also assisted with servicing the seismic stations in the field. Helicopter support from VXE-6 was great. This work was supported by National Science Foundation grant OPP 94-19267.
References
Chouet, B.A. 1985. Excitation of a buried magmatic pipe, a seismic source model for volcanic tremor. Journal of Geophysical Research, 90, 1881-1893.
Chouet, B.A. 1996. Long-period volcano seismicity: Its source and use in eruption forecasting. Nature, 380, 309-316.
Dibble, R.R., J. Kienle, P.R. Kyle, and K. Shibuya. 1984. Geophysical studies of Erebus volcano, Antarctica, from 1974 December to 1982 January. New Zealand Journal of Geology and Geophysics, 27, 425-455.
Dibble, R.R., B. O'Brien, and C.A. Rowe. 1994. The velocity structure of Mount Erebus, Antarctica, and its lava lake. In P.R. Kyle (Ed.), Volcanological and environmental studies of Mount Erebus, Antarctica (Antarctic Research Series, Vol. 66). Washington, D.C.: American Geophysical Union.
Kyle, P.R. (Ed.). 1994. Volcanological and environmental studies of Mount Erebus, Antarctica (Antarctic Research Series, Vol. 66). Washington, D.C.: American Geophysical Union.
Rowe, C.A. 1988. Seismic velocity structure and seismicity of Mount Erebus Volcano, Ross Island, Antarctica. (M.S. thesis, University of Alaska, Fairbanks.)
Skov, M.J., P.R. Kyle, and R.C. Aster. 1994. Seismic monitoring of the active Erebus volcano, Ross Island, Antarctica, using digital data acquisition. Seismological Research Letters, 65(1), 57.