Potassium-argon geochronology of a basalt-andesite-dacite arc system: The Mount Adams volcanic field, Cascade Range of southern Washington MAR^n'^A^^ANPHERE } • Geological Survey, Menlo Park, California 94025 ABSTRACT widely spaced episodes of peak productivity. erupted basaltic scoria and lavas that range Large stratocone systems may remain active widely in composition. In conjunction with High-precision K-Ar dating and detailed for half a million years. Subdivision of com- preparing a detailed geologic map (Hildreth mapping have established an eruptive chro- plex stratovolcanoes into eruptive or con- and Fierstein, 1995), we worked to establish nology for a Cascade stratovolcano and its structional "stages" without detailed map- an eruptive chronology for the whole vol- surrounding array of coeval basaltic cen- ping, geochronology, and compositional canic field, cone and periphery, through an ters. Mount Adams is a 200 km3 andesitic data should be treated skeptically. Discus- ambitious program of high-precision potas- cone that stands at the center of a coeval sion of volumetric eruptive rates can be mis- sium-argon (K-Ar) dating, the results of 1250 km2 Quaternary volcanic field that leading without an adequate time scale. which we report here. The present investi- contains >60 discrete vents. K-Ar ages were Stratovolcanoes need never develop large gation, using the multiple-collector mass measured for 74 samples from 63 of the 136 upper-crustal magma chambers. Basalt spectrometer (Stacey and others, 1981), may volcanic units defined in the course of the erupts peripherally, but its ascent is sup- be the most detailed K-Ar study yet under- mapping. Within analytical error, there are pressed beneath stratovolcanoes by deep- taken of an arc volcanic field where the strat- no discrepancies between K-Ar ages and crustal domains of magma storage and hy- igraphic record is well established. Previous stratigraphic sequence. Major activity be- bridization that form where concentrated attempts to date Quaternary stratovolca- gan in the area ca. 940 ka, and inception of injection of basalt amplifies crustal melting. noes and Quaternary basalt-andesite-dacite the central stratovolcano took place at ca. arc assemblages have met with mixed suc- 520 ka. A plot of cumulative volume erupted INTRODUCTION cess, although efforts comparable to the versus time shows that between 940 and 520 present study are underway in our labora- 3 ka the eruptive rate was <0.04 km /k.y. and Lofty composite cones loom large in per- tory for the Mazama (Bacon and Lanphere, —80% of the products were basaltic. Ande- ceptions of arc magmatism, but remarkably 1990) and Tatara-San Pedro (Dungan and sites are volumetrically dominant and were few of these conspicuous and abundant vol- others, 1993) arc systems. emplaced in three main cone-building epi- canoes have ever been investigated thor- An unnecessary pessimism is widespread sodes centered at 500, 450, and 30 ka—at oughly. Not uncommonly, more is known in the volcanological community concerning eruptive rates of 1.6-5 km3/k.y. At a lower 3 about the isotopic, phenocrystic, or trace- the feasibility of K-Ar age calibration of rate of 0.05-0.1 km /k.y., the magmatic sys- element composition of a stratovolcano Quaternary arc volcanoes, centered perhaps tem remained almost continuously active than about its stratigraphy, longevity, and around the observation that time intervals between the main pulses—although breaks eruptive history. For many active centers, between eruptions are commonly much in activity as long as 30 k.y. are permitted by the historical record is well established, and, shorter than the precision limits typical of the K-Ar data. Andesitic-dacitic activity in for a very few, detailed eruptive histories K-Ar age determinations for mafic and in- the focal region and dominantly basaltic ac- have been reconstructed as far back as the termediate rocks. The present results (Ta- tivity on the periphery have coexisted for limit of routine radiocarbon dating (ca. 40 ble 1) show that such pessimism is unwar- 520 k.y., and their products are interstrati- ka). For hardly any long-lived stratovolca- ranted, at least for a long-lived system. Our fied. The last main episode of cone construc- noes, however, do we have more than a relative success reflects the following fac- tion occurred ca. 40-10 ka, the oldest an- vague impression of what "long-lived" really tors: (1) Performance of the multiple-col- desite identifiably derived from within the means, much less any real measure of epi- lector mass spectrometer, as described in present-day edifice having an age of 33 ± 14 sodicity, fluctuation in eruption rates, or the Appendix. (2) Most products of the ka. Andesites forming the south-summit time-volume-composition relationships. Mount Adams volcanic field are fairly rich in rim and the true summit have ages of 13 ± K20. Although low-K lavas are also present, 8 ka and 15 ± 8 ka, respectively. Mount Adams, in southern Washington (Figs. 1 and 2), is one of the largest Qua- the dominant suite is the most potassic in The time-volume-composition data bear ternary stratovolcanoes (composite cones) the Cascades (Fig. 3); at equivalent Si02 upon several fundamental questions con- in the Cascade Range. The 200 km3 ande- contents, the Mount Adams suite has twice cerning the long-term behavior of arc vol- site-dacite edifice stands at the center of a the K20 content of rocks from Mount St. canoes. Stratovolcanoes commonly grow in coeval 1250 km2 volcanic field that contains Helens (Hildreth and Fierstein, 1985). (3) spurts but can stay active between the >60 discrete vents, many of which have Samples selected for dating were nonvesicu- Geological Society of America Bulletin, v. 106, p. 1413-1429, 7 figs., 1 table, November 1994. 1413 Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/106/11/1413/3381918/i0016-7606-106-11-1413.pdf by guest on 01 October 2021 HILDRETH AND LANPHERE structional stages, usually,: on the basis of structural, morphologic, or compositional evidence. Without adequate geochronology, the reality or integrity of Such stages, how- 47°N ever useful as a first measure, has to remain suspect: shifting vents and deposition over rugged terrain can make 103 and 105 yr un- conformities indistinguishable. (4) Strato- volcanoes are built up largely from the prod- ucts of central-vent eruptions, but many also have flank vents, and some (like Mount Adams) are surrounded by extensive arrays of peripheral vents. Without good age con- While Swan trol one can seldom do better than speculate about possible relationships among the magmas erupted at central, flank, and pe- ripheral locations. Assignment of undated flank-vent units to main-cone stages can be very tenuous. Neighboring cinder cones are sometimes called "parasitic," but rarely are the age-and-origin assumptions implicit in that word well founded. (5) Speculation about "cycles" of eruptive behavior, pro- gressive evolution in size or composition of a magmatic system, or stages in develop- ment of a volcano cannot be tested or in- vestigated adequately without detailed age ' Port/and control. (6) Delineations within complexvol- A MOUNT HOOD canic fields of petrologic suites, of lines of magmatic descent, or of unrelated but co- 122°W 121 °W existing magma types gain plausibility when scrutinized for age relationships as well as by Figure 1. Regional location map shows distribution of Mount Adams, Indian Heaven compositional criteria. (IHVF), and Simcoe Mountains (SMVF) volcanic fields and emphasizes four river systems that drain Mount Adams. Symbols: # = selected towns; A = large stratovolcanoes; -k = Even the best geochronology won't re- selected lesser volcanic centers, principally those that help define Quaternary volcanic solve all these problems, but extensive ap- front. In southern Washington, the Quaternary volcanic zone has an east-west width of 150 plication of precise dating methods—tied km. to detailed stratigraphic relationships—re- mains the most promising (and still under- lar, nondiktytaxitic, and (to the extent pos- vals in a volcano's past, and to estimate char- utilized) approach to understanding how sible) holocrystalline—taken generally from acteristic rates of magmatic differentiation. volcanic systems work. glass-poor massive interior zones of lava But good age control also enables us to ad- flows. Selection was based typically on thin- dress recurrent uncertainties about funda- VOLCANIC GEOLOGY section examination of several samples of mental processes and properties of mag- any particular unit, scrutinized to exclude matic systems. For example: (1) In thinking Geologic Setting porosity, glass, secondary minerals, and about subvolcanic plumbing systems, age in- posteruptive alteration. In some cases, re- formation is as important as composition for Mount Adams lies 50 km north of the Co- collection in successive field seasons was interpreting spatially separated eruptive lumbia River and 50 km east of Mount St. necessary to obtain appropriate material. units as comagmatic products of a common reservoir, the duration and configuration of Importance of Geochronology in Volcanology which are seldom definable. (2) If evolved • magma (say, dacite) has erupted repeatedly Comprehensive geochronology enor- at a particular volcano, age and composi- Figure 2. Location map for Mount Adams mously enhances the value and interpret- tional data are equally essential for assessing volcanic field. Abbreviations: PP = Pikers ability of stratigraphic, geochemical, and whether such behavior represents small Peak (south summit); SC = Sunrise Camp; petrologic data sets for any complex volcan- batches recurrently injected from the deep TS = The Spearhead; Ck = Creek; FK = ic field. Good age control obviously im- crust or intermittent tapping of a shallow Fork. • = summit of Mount Adams. A = proves our ability to recognize past episodes fractionating chamber of considerable size approximate center of deeply eroded Hell- of elevated (or diminished) eruptive fre- and longevity. (3) Eruptive products of long- roaring stratovolcano (ca.