Mt. Spurr's 1992 Eruptions


by the Alaska Volcano Observatory

Eos, Transactions, American Geophysical Union, 74:19, pp 217 and 221-222.

NOTE: This version is lacking the final edits by the Eos editors (mostly punctuation).

On 27 June, 1992, the Crater Peak vent on the south side of Mt. Spurr awoke from 39 years of dormancy and burst into sub-plinian eruption after 10 months of elevated seismicity. Two more eruptions followed in August and September. The volcano lies 125 km west of Anchorage, Alaska's largest city and an important international hub for air travel. The Alaska Volcano Observatory (AVO) was able to warn communities and the aviation industry well in advance of these eruptions.


Background
Chronology Chemistry and Petrology
Gas Monitoring
Discussion
Acknowledgments
Principal Investigators
References

Background


The Spurr massif is a large andesitic stratovolcano which collapsed in the late Pleistocene to produce an ashflow-covered debris avalanche and a 5 by 6 km caldera. Post-collapse effusions formed a central silicic-andesite summit dome, and Crater Peak, a basaltic-andesite cone in the caldera breach (Figure 1 ; Riehle, 1985; Nye and Turner, 1990). Crater Peak produced some 40 major mid- to late-Holocene tephra eruptions and the only other historic eruption in 1953 (Riehle, 1985; Juhle and Coulter, 1955). The volcano has been monitored seismically for 10 years by a system which now incorporates technical advances developed in response to the eruptions at Redoubt Volcano in 1989-90 and Mt. Pinatubo in 1991 (March and Power, 1990; Murray et al., 1991). Eight seismometers were in place, including one on the crater rim, just before the June eruption. Normally, about 10 earthquakes were located per month, but swarms of scores of events shook the north caldera rim in 1982 and 1989.

AVO monitors potentially hazardous Alaskan volcanoes, particularly those in Cook Inlet. When an eruption is believed to be imminent or is confirmed, AVO conducts an emergency call-down to key federal, state, and municipal agencies. AVO also communicates hazard information through periodic public updates, as well as through scientific publications.

Chronology

All times used in this report are ADT (UT-8 hours) unless otherwise noted.

Pre-eruption:
The first sign of reawakening was a swarm of small volcano-tectonic (VT) earthquakes beneath Crater Peak in August 1991. Following two months of relative calm, seismicity beneath the volcano increased to about ten times the pre-August level (Figure 2). Pulses of seismicity occurred in February, May, and June 1992. Most events were concentrated beneath Crater Peak and Spurr's summit, but some occurred under the north caldera rim. On 3 June AVO issued an advisory to government agencies and the airline industry to plan for an eruption. The daily number of located events reached 28 on 5 June, after which VT earthquakes declined to about 3-5 events/day and bursts of shallow tremor began. An overflight on 8 June revealed upwelling in the crater lake, which had turned from green to gray. One to 24 shallow tremor bursts per day, each 1-10 minutes long, were recorded at stations within 10 km of Crater Peak between 6 and 26 June. On 11 June, a field party found the lake temperature to be 50 C, pH 2.5, and SO4/Cl 2.7; temperature and pH were similar to the 1970s, but SO4/Cl had increased one hundred fold. Small new geysers erupted in the talus pile at the base of the crater wall.

On behalf of AVO, the Alaska State Seismologist briefed the Alaska Division of Emergency Services on 17 June. Tremor duration increased abruptly at 1534 ADT (all times in this article are Alaska Daylight Time, UT-8 hours, unless otherwise noted) on 24 June with the onset of a tremor episode which lasted 154 minutes. Twelve hours later a similar episode lasted 142 minutes. Eight additional tremor bursts occurred within the following 18 hours.

27 June Eruption:
On the morning of 26 June a field party reported that the crater lake had mostly disappeared and that lithic blocks, presumably forcefully ejected, had impacted the residual mudflat. At 1204 continuous tremor began. Meanwhile, at 1630, AVO issued formal warning of concern and went on 24-hour staffed duty. The final stage of the prelude began at 0300 on 27 June with a swarm of VT at 0-2 km depth, which built to rates of 1 event every 2 minutes, and a few long-period (LP) earthquakes. This swarm probably reflected the forceful injection of magma into the upper conduit. Abrupt doubling of tremor amplitude at 0705 heralded onset of eruption, which soon destroyed the closest seismic station, 400 m from the vent. Tremor amplitude gradually increased throughout the eruption, peaking between 0935 and 1025, and was recorded on stations more than 100 km away. Midmorning pilot reports estimated the plume to be as high as 9 km, and the National Weather Service (NWS) estimated a plume height of 14.5 km based on C-band radar measurements (pers. comm.). Small volumes of hot pyroclastic debris mixed with snow late in the eruption, forming flows which swept down the south side of the cone to the Chakachatna River, 6 km from the crater. The eruption lasted 4 hours. Seismicity decreased and by 8 July was below pre-August 1991 levels. In intensity and brevity, the volcano had repeated its 1953 performance (Juhle and Coulter, 1955).

The 27 June airfall-tephra formed a narrow black stripe that passed northward into sparsely populated areas, broadening in the lee of the Foraker-McKinley massif in the Alaska Range (Figure 3). Ash thickness was about 2mm in Denali National Park, 260 km downwind, and 1mm in Manley Hot Springs, 420 km downwind. Tephra volume was about 50x106 m3 [20x106 m3 dense-rock equivalent (DRE)], and was mostly juvenile andesite. The plume passed north to the Beaufort Sea, turned southeast into Canada and the contiguous United States, and drifted eastward.

18 August Eruption:
Seismicity remained low through July and the first half of August. Seismic monitoring of the volcano was somewhat compromised by the destruction of the crater rim station. Despite repeated attempts to reinstall the crater-rim station, the closest seismometer was now 4.8 km away. Only one shallow and two deep events were recorded between 12 August and 17 August. Perhaps the 27 June eruption "opened" the conduit, and allowed magma to rise undetected. At 1538 on 18 August a 16-min episode of weak tremor including several LP events began. At 1548 a pilot reported an ash-rich plume. The main eruption began at 1642 when strong tremor was recorded on all Spurr stations. By 1658 a subplinian column thrust through low clouds to reach 11 km altitude. Large bombs were thrown 750 m above the vent Ultimately, the radar-determined plume top reached about 14 km -- pilot reports were higher. Small pyroclastic flows descended the east and southeast flanks of Crater Peak. Some flows were dry and hot, and left coarse, clast-supported deposits with lobate, steep-fronted margins. Other flows mixed with snow and ice high on the cone and were more mobile and cooler. A late shower of mostly lithic blocks as large as 1 m were hurled as far as 3.8 km southeast of Crater Peak. The southeastward distribution of these deposits was controlled by the position of the vent against the northwest crater wall. More than 170 lightning strikes were detected by the AVO lightning detection system during the second half of the eruption. Eruption ended after 3 hours and 28 minutes at 2011, but intermediate and deep crustal seismicity increased afterward to levels comparable to those of mid-June.

The volume of August tephra is about 110x106 m3 (40x106 m3 DRE). Upper-level winds took the tephra plume east-southeast directly over Anchorage (Figure 3, Figure 4), where sand-sized ash fell as thick as 3 mm. Beyond Anchorage, the axis of the plume crossed the Chugach Mountains and followed the coast toward Yakutat Bay. At Yakutat, 550 km downwind, ashfall was significant; at Juneau, 1000 km downwind, the plume was opaque enough to disrupt air traffic. Ashfall forced the closing of Anchorage International Airport for 20 hours. Air-quality alerts were issued during the ashfall and on the following day, as vehicular traffic resuspended the ash.

16-17 September Eruption:
Seismicity remained elevated after the August outburst, but signs of an impending eruption remained so weak, even on the newly reinstalled station 400 m from the vent, that AVO had a crew in the crater on the afternoon of 16 September. At about 1930 on 16 September, seismic activity - both discrete events and weak tremor - began to increase at the crater rim station. Tremor amplitude increased again at 2225. An eruption, accompanied by brief incandescence, began at 2236 but lasted only 11 min. Weak tremor followed for the next hour. At 0004 on 17 September the main phase of the eruption began impulsively, accompanied by intermittent bright incandescence which could be seen both from Anchorage and with AVO's telemetered slow-scan camera system, 130 km to the south-southeast. This eruption lasted 3 h 36 min. Pyroclastic flows swept down the south, east, and east-northeast flanks of Crater Peak, entraining snow to become lahars. Other flows moved down the south flank. Although these flows look like pyroclastic flows, they were cool and water-saturated by the afternoon of 17 September. A narrow ballistic field extends at least 10 km east from the vent. This time a strong swarm of about 50 VT shocks occurred between depths of 5-10 km during the last part of, and for a few hours following the eruption, perhaps reflecting readjustment of the system after magma withdrawal.

The ash cloud rose to a radar-determined height of nearly 14 km (NWS, pers. comm.) and curved eastward across Alaska, only slightly dusting Anchorage but depositing enough ash to prompt air quality alerts and air traffic disruptions in Palmer, Wasilla, and surrounding communities (Figure 3). There was substantial ashfall in Glenallen, 350 km east, and detectable ashfall at Burwash Landing, 700 km east. Tephra volume is about 50x106 m3 (20x106 m3 DRE).

Post-eruption:
Rates of seismicity remained high through 1992, with two notable seismic crises. Nearly 24 hours of continuous tremor on 1-2 October followed by 72 hours of intermittent, quasi-periodic, tremor drove AVO to its highest concern-code, although no eruption occurred. AVO again went to its highest concern-code at 2207 AST on 9 November, a half hour into a 3.5 hr swarm of more than 170 earthquakes, but once again no eruption followed. These earthquakes had mixed frequency contents, occurred at three locations a few hundred meters apart and 1.2 km below Crater Peak, and had magnitudes up to 1.7. They probably were caused by the shallow intrusion of magma. The duration of this swarm was similar to those of the three eruptions. Seismicity is gradually returning to background level at this time.

Chemistry and Petrology

Most proximal bombs are cauliflower-like porphyritic hornblende-bearing andesite with brown, microlite-rich andesitic groundmass glass. Subordinant light grey-green andesite scoria has clear rhyolitic glass. Rare light-colored pumice composed of nearly microlite-free rhyolitic glass with a few percent of plagioclase phenocrysts and rare quartz grains also occurs. Much of the distal tephra appears to be the same as the proximal bomb material. Despite large variations in groundmass glass composition, the andesite is uniform in major and trace element composition throughout the eruptions. The 1992 andesite differs from 1953 andesite in having similar or lower concentrations of highly incompatible elements at higher SiO2 (Figure 5). 1992 andesite is therefore not a simple fractionate of magma that fed the previous eruption, but represents a new batch of magma. Prehistoric Crater Peak lava flows record the rise of small, chemically unrelated, batches of magma which fed only a few eruptions.

August and September pyroclastic flows contain several percent of gneissic xenoliths that are partly melted and highly inflated. Smaller proportions of similar material were also found in proximal June 27 deposits. These xenoliths consist of clear, vesicular, rhyolitic glass with crystals of plagioclase, orthopyroxene, cordierite, sillimanite, garnet, and spinel. Subordinate xenolith lithologies include white plagioclase-quartz-glass rock, which is siliceous (76 wt.% SiO2), yet, compared to the andesite, strongly depleted in all incompatible elements except U, and fewer xenoliths of garnet-plagioclase-wollastonite skarn. All xenoliths are metamorphic, which is remarkable because most of the country-rock exposed near Spurr is granitic. Cordierite-bearing samples fall off the trend of Spurr lavas towards high concentrations of both compatible transition metals and K, Cs, Rb, Ba, REE, Nb, Ta Y, U, Th, and Pb. Nye and Turner (1990) suggested that silicic "sweat" from the upper crust forms a significant component of all Spurr magmas. These xenoliths in bulk cannot be this component, because of their high concentrations of compatible transition metals, but their interstitial melt (as yet unanalyzed) may be. The xenoliths provide direct evidence of extensive partial melting of country rock beneath Spurr.

Gas Monitoring

AVO routinely monitors SO2 emission from Cook Inlet volcanoes by airborne UV correlation spectroscopy (COSPEC). In addition, airborne IR spectroscopy of CO2 was started on 25 September, 1992. There were slight increases in SO2 during premonitory seismic activity in August 1991 and May-June 1992, but COSPEC-determined SO2 concentrations were slight on 8 June. Surprisingly, SO2 concentrations were down to the detection limit two days after the June eruption, despite the fact that Total Ozone Mapping Spectrometer (TOMS) data indicate that the eruption cloud contained 185 kilotonnes of SO2 (Global Volcanism Network, 1992a). SO2 values were low after the August eruption, and before and after the September eruption, although again TOMS data indicate that the plumes contained 300 and 190 kilotonnes of SO2, respectively (Global Volcanism Network 1992b). The lack of SO2 between eruptions, despite the S-rich nature of the magma, may reflect absorption of SO2 by an active hydrothermal system during non-eruptive degassing. Alternatively, magma may have remained deep, below the depth of gas-saturation, until just before each eruption.

Discussion

Spurr has departed from its 1953 behavior by erupting three times instead of once. Yet all four eruptions share a commonality in volume and duration, as if controlled by some unchanging parameters involving initial overpressure, conduit volume, and magma-production dynamics in the deep source. What is not shared is the associated pattern of seismicity -- a vexing problem for would-be forecasters. We have observed most possible permutations: seismicity preceding but not following eruption, seismicity following but not preceding eruption, and seismicity both preceding and following eruption. We also observed intense seismicity without eruption in October and November. This fascinating story, and the character of magmatic behavior it implies, continues to unfold.

Acknowledgments

This work was supported by the Alaska Volcano Observatory. AVO is a cooperative program of the US Geological Survey (USGS, 4200 University Drive, Anchorage, AK 99508 and 345 Middlefield Road, Menlo Park, CA 94025 ), University of Alaska Fairbanks Geophysical Institute (UAFGI, University of Alaska Fairbanks, Fairbanks, AK 99775), and Alaska Division of Geological and Geophysical Surveys (ADGGS, 794 University Avenue #200, Fairbanks, AK 99709) under the Volcano Hazards Program of the USGS. The State of Alaska provided additional funding during the Spurr crisis through its Division of Emergency Services.

Principal Investigators

Principal contributors to the work summarized here are, from the USGS, T Casadevall, B Chouet, J Dorava, M Doukas, I Ellersieck, C Gardner, R Hoblitt, A Jolly, T Keith, J Lahr, T Mattox, B May, G McGimsey, D Meyer, T Miller, C Neal, R Page, J Paskievitch, J Power, C Stephens, D Trabant, and R Waitt. From the UAF/GI, J Beget, J Davies, K Dean, J Eichelberger, M Harbin, J Kienle, S McNutt, S Swanson, and G Tytgat. From ADGGS, R Motyka, C Nye, and G March. Requests for reprints may be sent to Chris Nye at UAFGI.

References

Global Volcanism Network, Bulletin of the Global Volcanism Network, Smithsonian Institution, 17, 6-8 , 1992a.

Global Volcanism Network, Bulletin of the Global Volcanism Network, Smithsonian Institution, 17,-3, 1992b.

Juhle, W. and H. Coulter, The Mt. Spurr eruption, July 9, 1953; Transactions, Amer. Geophys. Union, 36, 199-202, 1955.

March, G.D. and J.A. Power, A networked computer configuration for seismic monitoring of volcanic eruptions, U.S. Geol. Survey Open-file Report 90-442, 19 pp, 1990.

Murray, T.L., J.A. Power, G.D. March, A.B. Lockhart, J.N. Marso, and A. Miklius, Application of a real-time data acquisition and analysis system in response to activity at Mount Pinatubo, Phillippines, 1991 (abstract), Eos, 72, 47, 1991.

Nye, C.J., and D.L. Turner, Petrology, geochemistry, and age of the Spurr volcanic complex, eastern Aleutian arc, Bull. Volcanol., 52,05-226, 1990.

Riehle, J.R., A reconnaissance of the major Holocene tephra deposits in the upper Cook Inlet region, Alaska, J. Volcanol. Geotherm. Res.,6, 37-74, 1985.