USGS/CVO Logo, click to link to National USGS Website
USGS/Cascades Volcano Observatory, Vancouver, Washington

Mount St. Helens, Washington

Excerpt from:
Swanson, Cameron, Evarts, Pringle, and Vance, 1989,
IGC Field Trip T106: Cenozoic Volcanism in the Cascade Range and Columbia Plateau, Southern Washington and Northernmost Oregon: American Geophysical Union Field Trip Guidebook T106, p.21-24.

Mount St. Helens

Mount St. Helens is young. Its oldest known deposits were erupted about 50-40 thousand years ago, and the cone that partly collapsed in 1980 is only 2200 years old. Since its birth it has produced more than 60 tephra layers (Mullineaux, 1986), several tens of volcanically-induced debris flows (at least six of which entered the Columbia River 100 kilometers downstream (Scott, 1988)), and the equivalent of 60 cubic kilometers of dacitic lava (Smith, 1987). It has been the most active volcano in the Cascades during the Holocene, and for that reason its eruption in 1980 came as no surprise. The volcano has been studied intensively, and its eruptive history is know with greater clarity than that of any other Cascade volcano.

Eruptive History

Crandell (1987) divides the history of the volcano into "four extended stages of intermittent activity, each lasting two thousand years or longer. The volcano is now in such a stage that began about 4,000 radiocarbon years ago" (Table 2). The stages are separated by dormant intervals of thousands of years. Each stage contains eruptive periods with durations of decades to centuries; those periods for the current stage are named, but those for past stages are not. Many carbon-14 and dendrochronologic ages (Crandell, et.al., 1981; Mullineaux, 1986; Yamaguchi, 1983, 1985) provide a well-constrained context within which to interpret the volcanic history.

Table, Simplified Eruptive History of Mount St. Helens, click to enlarge [Table,27K,InlineGIF]
Table 2:
Simplified Eruptive History of Mount St. Helens. Simplified from: Crandell, 1987

Mullineaux (1986) recognized at least one tephra set, comprising several layers of tephra, in each eruptive stage and period except the Sugar Bowl and Goat Rocks periods, each of which has only one layer. The tephra sets and some single tephra layers form distinctive markers, recognizable by the assemblage and relative proportions of ferromagnesian phenocrysts (hypersthene, hornblende, biotite, cummingtonite, olivine, and augite). The tephra sets and some layers can be correctly identified in the field with a mortar, pestal, and binocular microscope using methods developed by D.R. Mullineaux.

The tephras are the most useful means of placing deposits of Mount St. Helens in proper stratigraphic context. They are also used to date events far from the volcano. For example, set S, about 13 thousand years ago, is interbedded with deposits of the Missoula floods on the Columbia Plateau and provides one of the best ages for the floods (Mullineaux et al., 1978; Waitt, 1985). Tephra layer Cs, about 37 thousand years ago, spread southward into the Lake Lahontan basin of Nevada, where it is probably equivalent to the Marble Bluff Bed (Davis, 1978; Mullineaux, 1986). Layer Yn, about 4 thousand years ago, occurs near Entwhistle, Alberta, 900 kilometers north-northeast of Mount St. Helens (Westgate, et.al., 1970; Mullineaux, 1986). Layer Ye occurs in northeast Oregon, 300 kilometers from the volcano (Borchardt, et.al., 1973). Layers Wn (1480 A.D.) and We (1482 A.D.) (Yamaguchi, 1983, 1985) are known 400 kilometers northeast and east of the volcano, respectively (Smith, et.al., 1977). Layer T (1800 A.D.) occurs in northwestern Montana, 375 kilometers away (Okazaki, et.al., 1972).

Pyroclastic flows and lahars formed in each eruptive stage and period (Mullineaux and Crandell, 1981; Crandell, 1987). Many pyroclastic flows carry lithic fragments probably derived from coeval domes. Lava flows are known with certainty from only the Cougar stage and Castle Creek and Kalama periods, although the "floating island" lava flow (Lawrence, 1941; Hoblitt et al., 1980) probably erupted in the 19th century (Crandell, 1987).

All chemical analyses reported by Smith and Leeman (1987; see also Smith 1984) and Crandell (1987) from pre-Castle Creek tephras are silicic andesite or dacite, mostly with water-free SiO2 contents of 6-68% but rarely as low as 57%.

The lack of mafic tephra suggests that no basaltic andesite or basalt was erupted before Castle Creek time. Proximal exposures are needed to test this suggestion, however, because such flows can be erupted with little tephra. Large dacite or silicic andesite domes of Pine Creek age, but no mafic flows, occur in the lower half of the present crater wall, so it is unlikely that mafic flows were produced during Pine Creek time. Three dacite or silicic andesite domes of possible Smith Creek age or older crop out in the northern part of the crater; their age assignment is based on the presence of cummingtonite, which is absent in younger deposits. However, domes and tephras of a given period need not have the same phenocryst assemblage. Aside from these questionably older domes, no pre-Pine Creek deposits occur in the crater. Near-vent mafic flows, perhaps in a small cone (Finn and Williams, 1987), could underlie the Pine Creek domes. No direct evidence exists for such flows, but gabbro inclusions resembling those in the present (Heliker, 1983, 1984) occur in Cougar and Pine Creek deposits (J.S.Pallister, written commun,, 1988) and suggest that mafic magma was in the reservoir system during much of Mount St. Helens history, possibly even before 50 ka (Williams, et.al., 1987). Nonetheless, the volcano has surely been dominated by silicic domes throughout most of its history.

The major compositional change occurred at the onset of Castle Creek time about 2200 years B.P. The period began, was dominated by, and ended with basalt, basaltic andesite, and lesser andesite, although dacite tephras, pyroclastic flows, and possibly a dome also formed (Mullineaux and Crandell, 1981; Smith and Leeman, 1987; Crandell, 1987). Numerous mafic lava flows, mostly borderline trachybasalt and trachybasaltic andesite (Le Bas, et.al., 1986) although more often called olivine basalt and two-pyroxene basaltic andesite, were erupted on all sides of the volcano, especially the southwest and north. The Cave Basalt issued from a vent probably near the southwest base of the cone about 1700 carbon-14 years B.P.; it contains 3.4-kilometer-long Ape Cave, the longest known uncollapsed segment of a lava-tube, as part of its 8.3-kilometer-long tube system (Greeley and Hyde, 1972).

Sugar Bowl dome on the north flank of the cone formed about 1200 carbon-14 years B.P. Two lateral blasts, the largest of which threw lithic lapilli 10 kilometers northeast of the vent, probably accompanied dome growth (Crandell and Hoblitt, 1986). East dome, at the east foot of the volcano, chemically resembles Sugar Bowl; it is undated, but bracketing ages allow it to be of Sugar Bowl age. The rhyodacitic compositions of both domes are most silicic (69-70%) yet found at Mount St. Helens (Smith and Leeman, 1987).

The Kalama eruptive period began in winter or early spring of 1479-1480, as deduced from dendrochronologic dating of the dacitic Wn tephra (Yamaguchi, 1983, 1985), the most voluminous tephra from Mount St. Helens since the Y tephras about 4000 years B.P. Another widespread tephra, We, fell in winter or early spring of 1481-1482. Episodic activity thereafter produced voluminous silicic andesite lava flows. The symmetric pre-1980 cone, known as North America's Fuji, was built by these and the older Castle Creek flows. A dacite dome formed at the summit during Kalama time, and lahars and pyroclastic flows were frequently produced.

The Goat Rocks eruptive period began in 1800 A.D. with the dacitic tephra layer T and ended in 1857 (Crandell, 1987). The "floating island" silicic andesite flow was erupted before 1838, and an explosion sent lithic ash 100 km downwind in 1842. The Goat Rocks dome was extruded on the northwest flank of the volcano 600-700 meters below the summit within several years after the 1842 explosion, possibly during or before 1847, when Paul Kane, a Canadian artist, painted a famous canvas of its growth. A large fan of debris spread downslope from the Goat Rocks dome; prismatic jointing is common in blocks of the fan, and paleomagnetic measurements indicate that some of the blocks were deposited above the Curie point and others below (R.P.Hoblitt in Crandell (1987)). Contemporary accounts indicate activity several times during the 1840's and 1850's but are non-specific and even contradictory. The last significant activity before 1980 was "dense smoke and fire" in 1857, although minor, unconfirmed eruptions were reported in 1898, 1903, and 1921 (Majors, 1980).

Volcanic Activity, 1980-1988

This activity has been thoroughly documented and is familiar to most volcanologists. See especially papers in Lipman and Mullineaux (1981), a series of nine papers in Science (v.221,no.4618, 1983), and Swanson, et.al. (1987) for general summaries. Only a synopsis is given here; other specifics are mentioned in the road log.

Seismicity began several days before March 20, 1980, when an earthquake (M=4.2) centered under the volcano commanded wide attention. The first of a series of small phreatic explosions occurred on March 27, accompanying the opening of a crater within a horseshoe-shaped graben concave northward at the summit of the cone. Strong seismicity continued, at times with bursts of deep volcanic tremor (Endo, et.al., 1981; Quamar, et.al., 1983); deep tremor was felt state-wide in early April but died away without returning. By mid-April a bulge was obvious on the north flank of the volcano; geodetic measurements began shortly thereafter and documented horizontal growth of the bulge at a steady maximum rate of >1.5 meters/day (Lipman, et.al., 1981a). The bulge was surface evidence of a cryptodome intruding the volcano. Seismicity continued into May, with fewer but larger earthquakes, and phreatic activity was intermittent. No magmatic gas was detected, although new fumaroles appeared in the crater and at the head of the bulge.

At 0832 on May 18, a complex earthquake (M=5.1) shook the volcano, probably causing (but possibly caused by) a huge, 2.7-cubic-kilometer-landslide that in three different blocks successively removed the bulge and upper 400 meters of the volcano (Voight, et.al., 1981, 1983), leaving a 600-meter-deep crater 2 kilometers wide rim-to-rim. The landslide quickly developed into a debris avalanche that sped at 110-240 kilometers/hour for 24 kilometers down the North Fork Toutle River; arms of the avalanche entered spirit Lake, 8 kilometers from the summit, and overtopped 300-380-meter high Johnston Ridge north of the Toutle. The avalanche buried the Toutle valley to a depth of nearly 50 meters. Its hummocky deposit is distinctive; similar morphology at other volcanoes has been reinterpreted in light of its observed origin (Siebert, et.al., 1987).

The landslide removed confining pressure on the cryptodome and its surrounding hydrothermal system. Juvenile gas was rapidly released from the cryptodome and superheated groundwater flashed to steam, causing a blast that exploded laterally from the collapsing north flank. The blast, called a "stone wind" by local journalists, knocked down most trees (the equivalent of about 150,000 houses) in a 600-square-kilometer area. Its maximum velocity may have been supersonic (Kieffer, 1981b). The degree to which juvenile gas or flashed groundwater drove the blast is debated (Eichelberger and Hayes, 1982; Kieffer, 1981a, b; see also Brugman, 1988), as is the question of whether the blast was, in volcanologic terms, a low-aspect-ratio ignimbrite (Walker and McBroome, 1983) or a surge (Hoblitt and Miller, 1984; Waitt, 1984b).

Soon after the blast, a lahar rushed down the South Fork Toutle River and several streams draining the south and east flanks of the volcano. Whether water for these lahars came from snowmelt or from groundwater ejected by the eruption is hotly argued. The largest lahar, down the North Fork Toutle, did not start until early afternoon; it was fed as the debris avalanche dewatered (Janda, et.al., 1981). In addition to causing havoc along the rivers themselves, the lahars fed so much debris into the columbia River that 31 ships were stranded in upstream ports until the 4-meter-deep channel was dredged to its pre-eruption depth of 12 meters -- the first in a series of similar dredgings to maintain Portland (Oregon) as a seaport.

Juvenile dacite pumice and ash mixed with lithic debris began erupting soon after the blast (Criswell, 1987), perhaps from the shallow root of the cryptodome. The flux increased about noon, apparently with arrival of pumice from a 7-10-kilometer-deep reservoir (Rutherford, et.al., 1985; Carey and Sigurdsson, 1985; Scandone and Malone, 1985). Experimental work suggests that just before eruption this reservoir was a a pressure of 220 +/- 30 MPO, Pwater was 0.5-0.7 Ptotal, and the temperature was 930 +/- 10 degrees C (Rutherford, et.al., 19885). Pyroclastic flows fed by the eruption column covered the debris avalanche in the upper North Fork Toutle basin, forming the pumice plain. Hydorexplosions created phreatic pits on the pumice plain, possibly as the pyroclastic flows covered the dewatering debris avalanche (Moyer and Swanson, 1987).

Plinian to subplinian explosions took place on May 25, June 12, July 22, August 7, and October 16-18, 1980. Products of the explosions decreased in SiO2 content with time from 65 to 63 % or less, possibly because they tapped a magma body zoned chemically, mineralogically, and in gas content.

Small domes grew in June and August but were destroyed by the next eruption. The current dome began growing after the last major explosion on October 18. As of December 1988, 17 episodes of dome growth have taken place, each lasting several days to 1 year (1983-1984); the latest was in October 1986. The dome stands 267 meters above its vent and 350 meters above its north base, is 860 meters by 1060 meters across, and contains 74 million cubic meters. The dome is highly phyric (30-35% plagioclase, 5% hypersthene, 1-2% hornblende, 1-2% Fe-Ti oxides, and <0.5% clinopyroxene), about 50% crystalline, and has about 63% SiO2. Geodetic and seismic precursors enabled prediction of growth events. The geometry and volumetric rate of growth of the dome followed consistent patterns with time. Several hundred small explosions occurred from the dome between 1980 and 1986, and a few large rockfalls spawned minor lithic pyroclastic flows and surges, none of which had sufficient volume to leave the crater. Several small lahars formed when snow melted during explosions and rockfalls.

Petrologic Interpretation

Smith and Leeman (1987) found that the St. Helens dacite has similar or even lower contents of many incompatible elements than the basalt and andesite from the volcano but is relatively enriched in Ba, Rb, K, Cs, and Sr. The unusual depleted nature of the dacite, and low bulk distribution coefficients for numerous trace elemets, preclude an origin by fractionation of the basalt or andesite. Smith and Leeman (1987) favor an interpretation involving melting of metabasaltic crustal rocks enriched in Ba, Rb, Cs, and Sr owing to interbedded sedimentary rocks or metasomatic enrichment of the source region. They consider mantle-derived basaltic magma to be the heat source for crustal melting.


Return to:
[Report Menu] ...
[Mount St. Helens Menu] ...
[Mount St. Helens Eruptive History Menu] ...



ButtonBar

URL for CVO HomePage is: <http://vulcan.wr.usgs.gov/home.html>
URL for this page is: <http://vulcan.wr.usgs.gov/Volcanoes/PacificNW/AGU-T106/msh.html>
If you have questions or comments please contact: <GS-CVO-WEB@usgs.gov>
09/21/98, Lyn Topinka