Small Explosions Interrupt 3-year Quiescence
at Mount St. Helens, Washington

On December 11, 1989, geologists working in the crater at Mount St. Helens discovered two thin layers of ash separated by fresh snow - clear evidence that at least two small explosions had occurred recently. The explosions were neither seen nor heard, but on December 7 scientists suspected that a small ash-producing explosion had occurred when seismometers near the volcano recorded a long explosion-like signal and tilt and displacement meters showed minor deformation of the dome. There were no other large seismic signals to account for the second ash layer, which was most likely associated with one of several smaller signals in early December. The December ash-producing explosions were the first eruptive activity at Mount St. Helens since October 1986.

There have been at least five more ash-producing explosions since December 1989, all without recognized seismic or other geophysical precursors. The ash from these explosions appears to be pulverized pieces of dacite dome without glass shards, which suggests that no new magmatic material was ejected. Several of the explosions were accompanied by snow and rock avalanches, pyroclastic density flows, ballistic showers and debris flows.

These ash-producing explosions are part of a series of at least 28 explosion-like seismic events that began on August 24, 1989. Seismic signals from these events resemble those associated with confirmed ash-producing explosions in April-May 1986. Yet not all of the 1989-1991 events produced ash plumes. Observations during four events clearly showed that neither a steam nor ash plume was generated. There is little information about the other events because they occurred when the mountain was not visible, nor was there physical evidence of ashfall or surface changes when scientists visited the crater days to weeks later. Considerable deformation of the north side of the dome occurred during the series of explosion-like seismic events. Sections of the dome slumped northward and two new vents were formed. However, monitoring the changes associated with individual events was often impossible because several key EDM (electronic distance meter) targets and tiltmeters were destroyed by the series of events.

[Graphic,9K,GIF]
The ash-producing explosions are part of a series of at least 28 explosion-like seismic events that began in August 1989. Long lines indicate the larger events (15 minutes to several hours in duration); short lines indicate small events (<< 15 minutes). Many of the events were not observed and left no physical evidence of their occurrence. The events are coded as follows: visible plume, P; confirmed "no plume", N; recognized tephra deposits, T; flowage deposits (pyroclastic and/or debris), F; ballistic blocks, B.

Only one of the ash deposits found on December 11, 1989, is indicated here. The second ash layer is probably associated with one of several small events (not plotted) in early December. The seismic signal from a small event is easily masked by other signals such as storm noise or rockfall. While seismic records are currently being searched for more small events, it is unlikely that any large signals were missed.

[Graphic,268K,GIF]
Seismic signatures from individual events during the 1989-91 series of explosion-like seismic events are similar to those associated with confirmed ash-producing explosions in April-May 1986. Sections of the seismic record during two of the 1989-91 events and one of the April-May 1986 events show some of the similarities. Tick-marks on the records are one minute apart; adjacent traces are 15 minutes apart. All three signals are from YEL seismometer (in the crater breach north of the dome).

A. Seismic signature of a confirmed ash-producing explosion during April-May 1986. This explosion, which occurred on April 18, 1986, was observed by scientists in the field. The event sent an ash plume to about 1,850 meters above the volcano and peppered the crater floor with ballistic blocks.

B. Signature of a small event from the 1989-91 series of explosion- like seismic events. This event, which occurred on September 14, 1990, was not observed, and no evidence of its occurrence was found when scientists visited the crater 6 days later. The seismic signature is typical of many of the smaller explosion-like signals in the 1989-91 series.

C. Signature of a large event from the 1989-91 series of explosion- like seismic events. This explosion, which occurred on December 7, 1989, was not observed, but instruments in the crater recorded deformation of the dome during the event, and ash was found in the crater several days later. The spacing between 15-minute lines was widened during the event in order to preserve the signature. This signature is similar to those from the smaller events, although higher in over-all amplitude and longer in duration. Plumes were confirmed during several of the larger events, but at least three did not generate a steam or ash plume.

Web note: Photo not on-line
Geologists measure deformation of the dome from a site at the north end of the crater. Many instrument sites close to and on the dome, including a key deformation instrument site near the dome, were destroyed during the series of explosions. Photograph by E. W. Wolfe, February 14, 1991.

Web note: Map not on-line
Schematic map of dome and crater floor showing location of monitoring sites. A, 1989, before the explosion series began. B, 1991, after numerous key monitoring sites were damaged or destroyed. Labeled sites are referenced elsewhere in text.

Chronology

Since May 18, 1980, twenty-one magmatic eruptions have occurred at Mount St. Helens, five predominately explosive and sixteen predominately dome-building. Several of the eruptions involved both explosive and dome-building activity. The last magmatic eruption was a dome-building eruption in October 1986. Between 1980 and 1986, hundreds of small ash-emissions also occurred; scientists called these small eruptions gas-and-ash emissions or ash- producing explosions to distinguish them from the larger explosive magmatic eruptions. These small ash-producing explosions occurred from cracks and small vents on the dome, and sent plumes of steam and ash 100 to 6,500 meters above the volcano. The more vigorous events also produced ballistic showers of hot dome fragments, and some generated small debris flows.

1989

On the night of August 24, 1989, the Mount St. Helens seismic network recorded a prolonged seismic signal that began with numerous shallow earthquakes (the largest was magnitude 2.7, depth 1.5 km). The initial signal was followed by several hours of increased background seismicity and hundreds of small earthquakes. Three tiltmeters on the north and west sides of the dome recorded small changes of 10-30 microradians (a microradian is a measure ground-tilt change roughly equal to that caused by raising one end of a kilometer-long beam by the thickness of a dime). EDM surveys of the dome the next day showed no other measurable changes. Although the seismic signal was similar to those associated with confirmed 1986 ash-producing explosions, no evidence of new ash was found . A smaller seismic event the next morning generated neither a steam nor an ash plume.

A 5-hour episode of elevated seismicity near sunset on December 7 was similar to but slightly smaller in average amplitude than the August 24 signal. Instruments in the crater also recorded deformation of the west side of dome during the first hours of the event (16 mm of extension across an existing crack and several hundred microradians of outward tilt at a nearby tiltmeter). Scientists suspected the event was a small explosion, but they were not positive because the August events had not produced ash plumes. Images sent by a slow-scan video camera on Harrys Ridge (5 miles north of the volcano) showed a dark cloud between Harrys and the crater; but it was impossible to tell whether the cloud was from an ash plume or was part of a rapidly approaching weather system. The December 11 discovery of ash layers in the crater snow confirmed scientists' suspicions that ash had been ejected.

Web note: Plots not on-line
Plots showing extension across an existing crack on the west side of the dome and ground tilt at a nearby tiltmeter during the explosion on December 7, 1989, which is indicated by the arrow on the YEL seismic amplitude plot. RSAM units are digital counts representing time-averaged voltage from the output of the seismic data-acquisition system (Endo and Murray, 1991)

1990

Early on the rainy Saturday morning of January 6, a 3-hour explosion-like signal was recorded. The initial signal was larger than in previous events; nevertheless, scientists were unaware of the explosion until several hours later when residents of central Washington began reporting light dustings of ash on their cars. The explosion was apparently directed northward from a new vent or collapse feature about a third of the way up the north flank of the dome. The dome was deformed below the new vent; EDM measurements showed that one station had moved 1.45 meters north and 0.8 meters down. The crater floor north of the dome was coated with sand- to boulder-size ejecta. This event destroyed several deformation targets, two tiltmeters, a magnetometer, and the telemetry for one seismometer. Because this explosion came as a surprise, a seismic alarm system was developed. This system triggers on relatively small seismic events and notifies the duty scientist's 24-hour beeper, permitting scientists to respond rapidly and, when appropriate, to notify the Federal Aviation Administration and emergency-management agencies that are responsible for areas that might be affected.

During the next 10 months, eleven more explosion-like seismic events occurred. Only the September 24 event occurred during a period of good visibility. The slow-scan video camera on Harrys Ridge showed no plume, and two observers on Johnston Ridge, 1 mile west of the camera, also saw no steam or ash plume during the several-hour-long event. A closer inspection of the crater by helicopter revealed no obvious changes, except perhaps a stronger- than-normal sulphur smell.

Seven explosion-like seismic events occurred in November and December; the two largest events (November 5 and December 20) produced visible plumes and considerable public interest. The November 5 explosion, which occurred shortly after 2 a.m., generated a north-directed density flow of hot fragmented rock debris from the dome. The material was apparently launched from a new crater 30 meters in diameter and 5-10 meters deep which formed at the southeast end of the January 6 collapse feature. The rock debris touched down on the lower north and northwest flanks of the dome and swept at least 0.5 kilometers northward on the crater floor. An antenna mast at GDN seismic station (north base of dome) was stripped from Garden rock by the pyroclastic flow, and a 6- meter-high steel antenna tower at YEL seismic station (460 meters north of GDN) was flattened by the density flow or subsequent debris flows. Telemetry from GDN was lost about 1 minute after the seismic event began and telemetry from YEL was lost 16 seconds later, which indicates a flow velocity of 104 km/hr.

The explosion peppered the crater floor north of the dome with ballistic blocks as large as 2 meters in diameter and sent an ash plume 6,500 meters above the volcano. Debris flows caused by the interaction between the pyroclastic density flow, hot ballistic blocks, and old snow, swept across much of the crater floor above YEL and were channelled into existing drainages below YEL. The debris flows traveled from the crater through two main channels (Step and Loowit) and into the North Fork Toutle River. Increases in seismicity at PAK (9 km downstream) and ERT (19 km downstream) 32 and 114 minutes after the explosion respectively, recorded passage of the debris flow.

Unlike the previous explosions, the December 20 event did not vent from the north side of the dome. Instead it appeared to come from a new crack about half way up the east flank of the dome. The explosion sent an ash plume to about 3,000 meters above the volcano; strong winds carried the ash cloud over 80 kilometers SSW. One airline canceled its evening flights in and out of Portland in response to this event.

Web note: Photo not on-line
View looking south at crater and dome on November 5, 1990. The arrow indicates the location of the new vent. Photograph by R. B. Waitt.

Web note: Photo not on-line
Close-up view of the November 5, 1990, vent. Note the geologist standing on the rim of a small crater which formed at the east end of the vent. Photograph by E. W. Wolfe.

Web note: Photo not on-line
A 6-meter-high steel tower at Yellow rock (YEL, 0.5 km north of the dome) was ripped from its base, knocked over, twisted and abraded by the November 5, 1990, explosion and debris flow. Photograph by E.W. Wolfe.

[Graphic,12K,GIF]
Seismic-amplitude plot for the morning of November 5, 1990 showing the timing of the explosion and subsequent debris flow at three sites. Peak A is the explosion and peak B is the debris flow. REM is on the dome, PAK is 9 km downstream, and ERT is 19 km downstream. The increased separation of B from A reflects the time it took the debris flow to move from the dome to the lower flanks of the volcano and the North Fork Toutle River.

1991

As of October 1991, there have been four explosion-like seismic events. The two largest (Feb 5 and Feb 14) damaged several seismic stations and produced ash plumes, pyroclastic density flows, and small debris flows. The February 5 event is perhaps the more interesting because both new vents on the dome apparently were active. The explosion began at 07:47 a.m. About 2 1/2 minutes into the event, the seismic amplitude increased sharply and signals from REM (upper east side of dome) and GDN were lost. An observer south of the mountain reported that a white plume shot up 1,800 meters above the volcano and turned dark grey 2-3 minutes later. Convection and winds aloft carried the plume 3,500 meters higher and southeastward; ashfall was reported just south of Mount Adams, 55 km away. Field investigations several hours after the event revealed a spectacular pattern of ballistic impact craters extending more than 1 km north-northeast from the vent on the north flank of the dome and a small pattern of impacts around the east vent (the December 20, 1990 vent). The explosion also sent a pyroclastic density flow down the north flank of the dome and caused a rock-and-snow avalanche from the dome and secondary debris flows.

Web note: map not on-line
Simplified map of the deposits formed by the February 5, 1991, explosion. The north vent generated a ballistic shower that hurled rocks over 1 km from the dome. The east vent produced only a very small ballistic shower. A snow avalanche, pyroclastic density flow and debris flow were also generated during this event.

Web note: Photo not on-line
Ballistic showers accompanied several of the explosions. On February 5, 1991, some of the ballistic blocks were 1-2 meters in diameter and created large impact craters. Photograph by R.B. Waitt.

Discussion

These explosions and explosion-like seismic events have caused considerable speculation among scientists regarding their cause, and the similarities and differences among them. The ash produced by the explosions consists of fragmented dome rock rather than juvenile magma, which suggests the explosions are related to the shallow hydrothermal system within or beneath the dome. Is groundwater percolating down to hotter rocks and flashing to steam? Or is rain or melting snow trickling down to areas of hotter rocks through an extensive crack system within the dome (see article by Larry Mastin)? Rain and snow induced cooling could be accelerating the formation of these cracks. The north side of the dome has slumped down and northward during this series of events. Are the explosions causing the slumping, or is the cracking and slumping triggering the explosions by depressurizing trapped pockets of steam or super-heated water? And why aren't steam or ash plumes generated during every event?

The shallow seismicity (1.5-2.0 km) associated with these events and the lack of seismic precursors suggests that magma is not rising into the dome or through the conduit system below. But in late 1987, small (< M=2.5) earthquakes began occurring in the deeper (4-8 km) part of the system (Malone, 1990). The frequency of these earthquakes gradually increased through mid-1990 (Moran and Malone, 1990) and then began decreasing, which suggests that changes may have occurred in the deeper part of the magma system. Is there a correlation between the deeper earthquakes and the shallow-seated explosions on the dome? Could the gas driving these explosions originate at the deeper level? Scientists may never know the answers to all these questions. But, if the explosions continue, careful monitoring of groundwater, rainfall, seismicity, and dome deformation may provide some answers.

[Graphic,17K,GIF]
Time vs. depth plot of seismic activity recorded at Mount St. Helens from 1980 through mid-October 1991. The earthquakes range in magnitude between 0.0 and 5.0. Note the virtual absence of any events below ~ 3 km between 1982 and 1987. Plot provided by the University of Washington Geophysics Program.

References Cited

Endo, E.T., Murray, T., 1991, Real-time Seismic Amplitude 
Measurement (RSAM): a volcano monitoring and prediction tool: 
Bulletin of Volcanology, v. 53, p. 533-545

Malone, S.D., 1990, Mount St. Helens, the 1980 Re-awakening and 
Continuing Seismic Activity: Geoscience Canada, v. 17, p. 146-150

Moran, S.C., Malone, S.D., 1990, Recent Micro-Seismic Activity 
at Mt. St. Helens and its Implications for the Evolution of the 
Deeper Magmatic System: EOS, v. 71, p. 1693-1694