Journal of Petrology | Volume 43 | Number 4 | Pages 663-703 | 2002
© Oxford University Press 2002
Petrogenesis and Implications of Calc-Alkaline Cryptic Hybrid Magmas from Washburn Volcano, Absaroka Volcanic Province, USA
T. C. FEELEY1,*,
M. A. COSCA2 and
C. R. LINDSAY1
1DEPARTMENT OF EARTH SCIENCES, MONTANA STATE UNIVERSITY, BOZEMAN, MT 59717, USA
2INSTITUTE OF MINERALOGY AND GEOCHEMISTRY, UNIVERSITY OF LAUSANNE, BFSH 2, 1015 LAUSANNE, SWITZERLAND
Received
June 16, 2000;
Revised typescript accepted
October 18, 2001
 |
ABSTRACT
|
|---|
The petrogenesis of calc-alkaline magmatism in the Eocene Absaroka
Volcanic Province (AVP) is investigated at Washburn volcano,
a major eruptive center in the low-K western belt of the AVP.
New
40Ar/
39Ar age determinations indicate that magmatism at
the volcano commenced as early as 55 Ma and continued until
at least 52 Ma. Although mineral and whole-rock compositional
data reflect near equilibrium crystallization of modal phenocrysts,
petrogenetic modeling demonstrates that intermediate composition
magmas are hybrids formed by mixing variably fractionated and
contaminated mantle-derived melts and heterogeneous silicic
crustal melts. Nd and Sr isotopic compositions along with trace
element data indicate that silicic melts in the Washburn system
are derived from deep-crustal rocks broadly similar in composition
to granulite-facies xenoliths in the Wyoming Province. Our preferred
explanation for these features is that mantle-derived basaltic
magma intruded repeatedly in the deep continental crust leading
to fractional crystallization, silicic melt production, and
homogenization of magmas, followed by ascent to shallow reservoirs
and crystallization of new plagioclase-rich mineral assemblages
in equilibrium with the intermediate hybrid liquids. The implications
of this process are that (1) some calc-alkaline magmas may only
be recognized as hybrids on purely chemical grounds, particularly
in systems where mixing precedes and is widely separated from
crystallization in space and time, and (2) given the role ascribed
to crustal processes at Washburn volcano, the variation between
rocks that follow calc-alkaline trends in the western AVP and
those that follow shoshonitic trends in the east cannot simply
reflect higher pressures of fractionation to the east in Moho-level
magma chambers in the absence of crustal interaction.
KEY WORDS: petrogenesis; magma mixing; calc-alkaline; Absaroka Volcanic Province; 40Ar/39Ar dates
|
INTRODUCTION
|
|---|
The widespread, but poorly understood Challis–Absaroka
volcanic episode affected large areas of the northwestern USA
in the Eocene following Laramide crustal shortening (i.e. the
‘Challis arc’; Armstrong, 1978
). Volcanic rocks
associated with this event are particularly voluminous in the
Absaroka Volcanic Province (AVP) of northwestern Wyoming and
southwestern Montana, USA, where >20 000 km
3 of calc-alkaline
to shoshonitic magmatic rocks are exposed in the Absaroka, Gallatin,
and Beartooth Ranges (Fig.
1; Absaroka Volcanic Super group
of Smedes & Prostka, 1972
). Because the Absaroka volcanic
rocks have long been presumed to have calc-alkaline compositional
affinities and across-strike enrichments in K
2O similar to magmas
erupted in some continental volcanic arcs (e.g. Dickinson &
Hatherton, 1967
; Chadwick, 1970
), early workers attributed their
origin to shallow subduction of the Farallon plate beneath the
North American plate during the Eocene (e.g. Lipman
et al.,
1972
). However, this interpretation is now controversial for
several reasons. First, spatial and temporal studies have been
unable to decipher any logical time-transgression of magmatic
activity throughout the northwestern USA during the Eocene,
with the result that the ‘Challis arc’ was much
wider than any modern arc and oriented perpendicular to the
present plate margin (Fig.
1). On the basis of this distribution
and geochemical arguments some workers now consider the Challis–Absaroka
volcanic episode as resulting mainly from regional lithospheric
extension and resultant decompression melting (Dudas, 1991
;
Hooper
et al., 1995
; Morris
et al., 2000
). Second, the origins
of magmatic rocks in the AVP are poorly understood because there
exist few detailed studies aimed at deciphering the petrologic
evolution of magmas erupted from individual eruptive centers.
This lack of detailed studies has led to uncertainty on a number
of issues ranging from the role of crustal interaction during
differentiation of the calc-alkaline magmas to the significance
of the across-strike K
2O enrichments in the AVP.
View larger version (68K):
[in this window]
[in a new window]
|
Fig. 1. Map of the Absaroka Volcanic Province showing the stratigraphic units of Smedes & Prostka (1972). Black areas represent locations of principal vent complexes and intrusive centers discussed by Chadwick (1970). Thick dashed line shows the approximate division between western K-poor (calc-alkaline) and eastern K-rich (shoshonitic) belts (after Chadwick, 1970). Single dot-dashed line is the boundary of Yellowstone National Park. Inset shows the locations of early-to-middle Eocene magmatic fields (after Chadwick, 1985; Holder & Holder, 1988; Dudas, 1991; Norman & Mertzman, 1991; Wheeler et al., 1991; Luedke, 1994). Numbers refer to: 1, Sloko Volcanic Province; 2, Francois Lake igneous complex; 3, Colville igneous complex; 4, Clarno volcanics; 5, Challis Volcanic Province; 6, Absaroka Volcanic Province; 7, Montana Alkalic Province; 8, Black Hills. Diagonally ruled field shows inferred extent of Archean cratonic Wyoming Province (Dutch & Nielsen, 1990).
|
|
To address some of these issues and provide a reference point for future studies we present results from our investigation of Washburn volcano, one of the principal eruptive centers for calc-alkaline magmas in the AVP and the type locality for the Washburn Group of the Absaroka Volcanic Supergroup of Smedes & Prostka (1972; Fig. 1). Overall, this work represents the first detailed study of a calc-alkaline (sensu stricto) eruptive center in the AVP. The goals of this paper are (1) to document the ages and ranges of major element, trace element, and isotopic (Sr and Nd) compositions of magmatic products at Washburn volcano, (2) to place constraints from phenocryst compositions on the crystallization history of Washburn rocks, (3) to evaluate the possible roles of different petrologic processes in producing the spectrum of compositions observed, and (4) to use the information in (1)–(3) to develop a generalized model for the evolution of calc-alkaline magmas at Washburn volcano and in the AVP.
The origin of intermediate composition rocks at the volcano is best explained by mixing between variably fractionated and contaminated mafic magmas and heterogeneous silicic partial melts of Archean granulite-facies rocks in the deep crust. Despite chemical evidence for mixing, there is very little textural or mineral chemistry evidence in the hybrid magmas to support the model. A scenario where hybrid magmas produced in the deep crust ascend to shallow reservoirs and crystallize low-pressure mineral assemblages dominated by plagioclase is preferred to explain these relationships. These results have important implications for the interpretation of igneous textures and their bearing on rock-forming processes and the significance of across-strike geochemical variations in the AVP.
|
TECTONIC AND GEOLOGIC SETTING
|
|---|
Existing geochronologic information indicates that the bulk
of the magmas in the AVP were erupted between 54 and 38 Ma,
following or possibly overlapping the latest phases of Laramide
foreland thrusting, which ended at
55 Ma at this latitude (Love
et al., 1975
; Armstrong, 1978
; Feeley
et al., 1999
; Hiza, 1999
).
Eruption of the Absaroka rocks also appears to temporally coincide
with the onset of regional crustal extension in the northwestern
USA. Evidence supporting this contention comes from abundant
isotopic ages of metamorphic and igneous rocks exposed in the
cores of extensional complexes in Idaho and Washington State
that indicate tectonic elevation and denudation at
45–50
Ma (Burchfield
et al., 1992
), and interbedded tuffs in synextensional
basin-fill deposits that range from 46 Ma to younger than 30
Ma (Janecke & Snee, 1993
). Nevertheless, although temporally
associated with regional crustal extension, there is little
evidence within the AVP for the presence of major early Tertiary
tectonic extensional faults. Several large-displacement extensional
structures involving the volcanic rocks are present, but these
are generally regarded as features related to east-directed
gravitational sliding off the growing volcanic highland (Hauge,
1985
).
At present, the AVP covers 23 000 km2 in northwestern Wyoming and southwestern Montana (Fig. 1). It has been tentatively correlated with other Eocene volcanic rocks in southwestern Montana, the Challis volcanic rocks of Idaho, and the Colville igneous complex of NE Washington, although these fields are not linked by outcrops (Fig. 1; Armstrong, 1978; Dudas, 1991). Undoubtedly, much of the original extent of the volcanic field is eroded or covered by Miocene and younger volcanic rocks of the Snake River Plain–Yellowstone Plateau fields, and continuity with volcanic rocks of similar age may be obscured. However, apparent regional geochemical trends within the AVP, such as increasing K2O contents eastward, suggest it may be separate.
According to correlations advanced by Smedes & Prostka (1972) for Eocene rocks exposed in Yellowstone National Park, three groups make up the Absaroka Volcanic Supergroup. In ascending stratigraphic order these are (1) the Washburn Group, (2) the Sunlight Group, and (3) the Thorofare Creek Group. The areal distribution of the three groups, illustrated in Fig. 1, in part reflects the prevalent thinking at the time of the study of Smedes & Prostka (1972) that volcanic activity migrated SE along two subparallel belts of intrusive–extrusive centers: the ‘K-poor’ (calc-alkaline) western belt and the ‘K-rich’ (shoshonitic) eastern belt (Chadwick, 1970). However, subsequent stratigraphic and geochronologic studies in eastern and southern areas of the volcanic sequence are incompatible with the nomenclature established in Yellowstone National Park (e.g. Brown, 1982; Eaton, 1982; Sundell & Eaton, 1982; Hague, 1985; Decker, 1990; Hiza, 1999). Even within the park itself, Smedes & Prostka (1972) encountered major difficulties in establishing regional stratigraphic subdivisions because many of their formational units consist of monotonous sequences of andesitic volcaniclastic rocks, which are difficult to distinguish at different localities. Furthermore, Hiza (1999) has recently argued that mafic magmas (i.e. <56 wt % SiO2) are characteristically potassic throughout the AVP and that if only these compositions are considered, little regular geographic variation in magma chemistry exists. In spite of these complications we illustrate the inferred location of the K2O dividing line of Chadwick (1970) and the stratigraphic subdivisions of Smedes & Prostka (1972) in Fig. 1 because this facilitates comparison with previous studies. Our work also demonstrates that mafic magmas erupted at Washburn volcano and the Electric Peak–Sepulcher Mountain eruptive center (Lindsay & Feeley, 1999, and this study) are not especially potassium rich. We recognize, however, that the stratigraphy and eruptive history of the AVP are undoubtedly more complex than originally envisioned (e.g. Hiza, 1999).
Seismic refraction studies indicate that the crust beneath the AVP at present is 45–50 km thick (Prodehl & Lipman, 1989). Exposed basement rocks include Archean crystalline rocks of the Wyoming Province (Fig. 1), which are mainly granitoid gneisses that intruded high-grade metasedimentary and metavolcanic rocks at 2·8 Ga, and shallow marine carbonate and clastic sedimentary rocks ranging in age from Cambrian to Cretaceous (Ruppel, 1972; Wooden & Mueller, 1988). Deep- to mid-crustal lithologies are represented by mafic to silicic Archean granulite-facies xenoliths carried to the surface by Eocene alkalic magmas in the Crazy, Bearpaw, and Highwood Mountains (Dudas et al., 1987; Collerson et al., 1989; Joswiak, 1992), and late Cenozoic basaltic magmas of the Snake River Plain (Leeman et al., 1985). Many of these xenoliths are similar to granulite-facies rocks worldwide in that they have relatively low contents of Rb, U, and heavy rare earth elements (HREE), although they are light rare earth element (LREE) enriched (Leeman et al., 1985; Joswiak, 1992).
Stratigraphy
Washburn volcano is a major calc-alkaline eruptive center in the AVP and is the largest Eocene volcanic center exposed in Yellowstone National Park. It is the primary source area for the Lamar River Formation, the eastern member of the Washburn Group of Smedes & Prostka (1972). The Lamar River Formation is particularly well known to visitors of Yellowstone National Park because in it are preserved the famous upright fossil forests (Dorf, 1964). Washburn volcano has been previously mapped by Schultz (1962; 1:30 000) and Prostka et al. (1975; 1:62 500; Fig. 2).
View larger version (81K):
[in this window]
[in a new window]
|
Fig. 2. Simplified geologic map of Washburn volcano and surrounding area (modified from Prostka et al., 1975). •, locations of samples analyzed in this study. Inset shows the location of Washburn volcano in Yellowstone National Park.
|
|
The eroded northern flank of Washburn volcano in the vicinity of Mt. Washburn, and Hedges and Dunraven Peaks in the SW Washburn Range, consists of 1300 m of volcanic vent facies strata, mainly of the Lamar River Formation, that include dikes, lava flows, flow breccias, and debris flow deposits that dip up to 30° away from the primary vent region (Fig. 2; Prostka et al., 1975). The lava flows and dikes are largely pyroxene basaltic andesites and andesites, although numerous amphibole-bearing dacitic lava flows are present near the base of the sequence. With increasing distance from Mt. Washburn and Hedges and Dunraven Peaks the vent-facies rocks grade into alluvial-facies lithologies consisting of epiclastic volcanic conglomerate and breccia, volcanic sandstone and siltstone, and ashfall tuff deposits. The southern flank of Washburn volcano is truncated by the northern segment of the Yellowstone Caldera fault, exposing the interior of the volcano (Fig. 2). Here, fine-grained biotite tonalite of the Sulphur Creek stock intrudes stratigraphically low Lamar River Formation volcanic rocks. This stock is similar in composition and age (see below) to the Eocene volcanic rocks and represents a shallow intrusion related to Washburn volcano.
The samples examined in this study were chosen according to the stratigraphic units of Prostka et al. (1975; Fig 2) from well-exposed, vertical stratigraphic sections dominated by vent-facies rocks. Temporal and spatial variations in bulk chemistry for lava flows within the Washburn volcanic sequence are shown schematically in Fig. 3. In Fig. 3 (and subsequent figures) we designate with different symbols lava flows exposed on Mt. Washburn and those in the SW Washburn Range to the west of the Grand Loop Road because these have different compositional ranges. Lava flows and dikes in the SW Washburn Range consist of a crudely bimodal package of olivine + pyroxene basaltic andesites and amphibole-bearing dacites, whereas dikes, stratigraphically higher lava flows, and the Sulphur Creek stock to the east and NE on Mt. Washburn are predominantly olivine + pyroxene basaltic andesites and pyroxene ± amphibole andesites. Included in this latter sequence are the stratigraphically highest exposed lava flows on Mt. Washburn that Prostka et al. (1975) designated as part of the Langford Formation of the Thorofare Creek Group of Smedes & Prostka (1972; Figs 2 and 3). Although these flows were originally interpreted as younger (middle to upper Eocene) and erupted from vents much more distal than other units at the volcano, our work shows that they are comparable in age and composition with other flows on Mt. Washburn. We therefore consider all exposed units to be derived from the same or very similar magmatic systems.
View larger version (38K):
[in this window]
[in a new window]
|
Fig. 3. Schematic composite stratigraphic column for Washburn volcano combining total thicknesses of strata from the SW Washburn Range and Mt. Washburn areas. Lava flows are indicated by solid patterns; clastic units are indicated by stipple patterns (excluding brecciated autoclastic lava flow tops and bases). Geochemical data panels show the compositional variations of magmas with stratigraphic position. Open symbols are for samples from the SW Washburn Range and filled symbols are for samples from Mt. Washburn. Circles, squares, and triangles are for basaltic andesitic, andesitic, and dacitic composition rocks, respectively. Note: (1) reinterpretation of ‘Langford Formation’ flows on Mt. Washburn as late Washburn volcano units based on data presented in this study; (2) bimodal assemblage of dacitic and basaltic andesitic lavas in lower part of section from SW Washburn Range; (3) dominantly andesitic lavas in upper part of section from Mt. Washburn.
|
|
|
GEOCHRONOLOGY
|
|---|
To ascertain the timing of magmatism at Washburn volcano we
determined
40Ar/
39Ar ages from phenocryst and groundmass samples
of the stratigraphically highest and lowest exposed lava flows
together with a biotite separate from the Sulphur Creek stock
(Table
1). The data are graphically presented in the form of
40Ar/
39Ar age spectra and
39Ar/
40Ar vs
36Ar/
40Ar isochron diagrams
in Fig.
4. All errors on ages and intercepts reported in Fig.
4 are 2
. The errors on individual steps, graphically represented
by the width of rectangular boxes on the age spectrum diagrams,
also represent a 2
level of confidence. The sample localities
are shown in Fig.
2. Details of the analytical procedures are
described in the Appendix and a summary of the results is presented
in Table
1. The full dataset may be downloaded from the
Journal of Petrology Web site at
http://www.petrology.oupjournals.org.
Plagioclase and amphibole phenocrysts and a sample of fine-grained groundmass were separated from a dacitic lava flow at the base of Hedges Peak (MW971), the stratigraphically lowest eruptive unit, to date the initiation of volcanic activity at Washburn volcano. Theoretically, all three samples should yield identical ages because all were at high temperatures immediately before extrusion and cooled rapidly once emplaced. The amphibole sample yielded an 40Ar/39Ar spectrum with evidence of excess argon in the initial heating steps, and most of the gas released defined ages of 60 Ma (Fig. 4a). The integrated total fusion age (roughly equivalent to a K–Ar age) is 59·3 Ma; however, no age plateau was obtained. The corresponding K/Ca ratios determined from the step heating data indicate a relatively homogeneous sample, possibly with some K-rich phyllosilicate contamination consistent with the high K/Ca ratio in the first step. Data from the amphibole sample failed to plot on a statistically meaningful isochron, precluding an independent evaluation of the isotopic ratio of trapped argon and the age of the sample. An internally discordant 40Ar/39Ar age spectrum was obtained from the plagioclase, which has an integrated fusion age of 54·7 Ma (Fig. 4b). Intermediate heating steps defined ages of 50 Ma, but the last 30% of gas released defined ages in excess of 60 Ma. The plagioclase data also failed to plot on a statistically meaningful isochron. The groundmass sample also has an internally discordant 40Ar/39Ar age spectrum, with a integrated fusion age of 54·06 Ma (Fig. 4c). Apart from the initial heating step, the K/Ca ratios are consistent with argon degassing from a relatively homogeneous sample. Isotopic data obtained from intermediate heating steps, representing 50% of the total argon released, plot on an isochron defining a sample age of 55 Ma (Fig. 4d). Because the isochron age is calculated making no assumptions about the isotopic composition of the trapped argon, we consider this age as the most reliable estimate for dating the emplacement of lava flow MW97-01. Moreover, the isochron age is similar to the integrated fusion ages of both the groundmass and the plagioclase. This interpretation suggests that the amphibole age of 59 Ma is unreliable, and it is consistent with no published reports of such old eruptive ages being described from the AVP. We therefore consider the amphibole data as anomalously old, probably because of extraneous argon. A similar result was obtained in the biotite data of sample MW9746 described below.
The 40Ar/39Ar age spectrum obtained for the groundmass of sample MW9743 is internally discordant with an integrated fusion age of 44·4 Ma (Fig. 4e). The highly variable K/Ca ratios are consistent with argon degassing from reservoirs of strongly varying retentivity and composition. When plotted on an isochron diagram, however, the data from heating steps making up 97% of the argon released define a relatively precise (MSWD = 3·9) age of 51·9 Ma for the effusion of this lava (Fig. 4f). Taken together, the isochron ages from the groundmass samples of the stratigraphically lowest and highest samples indicate that the >1 km thick accumulated Washburn volcanic pile was constructed in 3 my.
Biotite from the Sulphur Creek stock sample MW9746 yielded an 40Ar/39Ar age spectrum with evidence of extraneous argon in the first few heating steps with an integrated fusion age of 61·6 Ma and a well-defined age plateau of 53·5 ± 0·4 Ma (Fig. 4g). However, an isochron with these data (MSWD = 3·2) indicates a slightly younger age of 52·6 ± 0·2 Ma (Fig. 4h). The isochron age of 52·6 Ma is our preferred date for the sample, and the 40Ar/36Ar ratio of the trapped argon component (323) is significantly greater than present-day atmosphere (295·5) and is consistent with ‘excess’ argon as previously inferred for the amphibole in sample MW971.
On the basis of the 40Ar/39Ar data presented here we interpret magmatism at Washburn volcano to have commenced as early as 55 Ma and possibly continued until at least 52 Ma. These ages bracket the 53·4 ± 0·3 age reported by Hiza (1999) for a dacitic block and ash flow at the base of Sepulcher Mountain, located 50 km to the NW of Mt. Washburn (Fig. 1). The sequence of rocks exposed on Sepulcher Mountain represents the type section for the Sepulcher Formation, the western member of the Washburn Group of Smedes & Prostka (1972). Because these rocks are compositionally and petrographically identical to rocks exposed at Washburn volcano (Lindsay & Feeley, 1999), we concur with the opinion of Smedes & Prostka (1972) that rocks within the Sepulcher and Lamar River Formations were erupted from similar and nearly contemporaneously active volcanic centers. In addition, our results are also consistent with the suggestion of Hiza (1999) that the oldest rocks in the AVP are calc-alkaline lavas at present exposed in the northwestern part of the field.
|
PETROGRAPHY AND SILICATE MINERAL CHEMISTRY
|
|---|
Washburn igneous rocks investigated in this study are generally
non- to slightly vesicular (<3 vol. % vesicles), porphyritic
to partially glomeroporphyritic lavas and dikes. In thin section
the rocks are hypocrystalline with intersertal to pilotaxitic
groundmass textures. Phenocryst contents range from 44 to 5%
(by volume) with the total decreasing with increasing SiO
2 contents
until
63 wt % SiO
2 (Fig.
5; Table
2). Dacitic rocks have widely
varying phenocryst contents. Changes in mode with bulk composition
are regular throughout the suite. Basaltic–andesitic rocks
contain plag
cpx > ol ± opx; andesitic rocks contain
plag > cpx > opx ± amph ± ol; and dacitic
rocks contain amph + plag ± cpx ± opx ±
bio (Fig.
5). All rocks also contain Fe–Ti oxide microphenocrysts.
Glomerocrysts are common of cpx ± Fe–Ti oxides,
ol + cpx ± Fe–Ti oxides ± plagioclase ±
opx, and plag + amph ± Fe–Ti oxides. Mineral inclusion
patterns and the occurrence of minerals in the glomerocrysts,
together with textural features and compositions of the phenocrysts
described below, suggest the following generalized crystallization
sequences. Basaltic andesitic magmas crystallized olivine followed
by Fe–Ti oxides, clinopyroxene + plagioclase, and orthopyroxene.
Andesitic magmas crystallized plagioclase followed by clinopyroxene
+ orthopyroxene, Fe–Ti oxides and then amphibole in some
cases. Dacitic magmas, in which pyroxenes are rare or absent,
precipitated plagioclase followed by Fe–Ti oxides, amphibole,
and then biotite when present. Groundmass assemblages include
glass (partially to pervasively devitrified), plagioclase, clinopyroxene,
and orthopyroxene. Zircon and apatite are common accessory phases
in the intermediate and silicic lavas. Xenocrysts, identified
by non-equilibrium compositions (see below) or magmatic reaction
textures, are present but rare. Furthermore, we identified no
clear petrographic evidence for mixing or mingling between compositionally
disparate magmas, such as the presence of undercooled blobs
of mafic magma that are frequently found in many andesitic to
dacitic rocks (Bacon, 1986
; Wilcox, 1999
).
View larger version (20K):
[in this window]
[in a new window]
|
Fig. 5. Variation of modal abundances of phenocryst phases in Washburn lavas and dikes (Sulphur Creek Stock excluded) vs wt % SiO2. Symbols as in Fig. 3. Large symbols in pyroxene panel represent clinopyroxene; small symbols represent orthopyroxene. All rocks also contain microphenocrysts of Fe–Ti oxides and a few dacitic rocks contain small amounts of biotite phenocrysts.
|
|
View this table:
[in this window]
[in a new window]
|
Table 2: Major element, trace element, Sr and Nd isotopic ratios, and modal data for Washburn volcano igneous rocks
|
|
The Sulphur Creek stock is tonalitic to quartz dioritic with plag > qtz > bio > cpx = opx > Fe–Ti oxide > amph. Texturally, the stock is fine grained, phaneritic, and subophitic in that late-crystallizing anhedral quartz partially encloses elongate plagioclase and biotite grains.
Olivine
Olivine generally occurs as equant, euhedral to subhedral phenocrysts in basaltic andesitic lavas. In the majority of these samples the phenocrysts range in size from about 4·0 to 0·5 mm, although populations in individual samples typically have a much narrower range. In a few samples the size variation is continuous from a maximum of 1·0 mm to a minimum of 0·1 mm. Additionally, a few andesitic and dacitic rocks contain small amounts of olivine (Fig. 5). Although we did not analyze any of these grains, they are probably xenocrysts or relicts because grain boundaries are ragged, variably resorbed, and have corona of clinopyroxene or orthopyroxene plus plagioclase. In nearly all rocks olivine is partially altered to bowlingite, particularly along crystal faces and fractures.
The total analyzed variation in the compositions of olivine in the basaltic andesitic rocks ranges from Fo85 to Fo72; within single phenocrysts, rims are normally zoned 
7 mol % Fo content relative to core compositions (Table 3). Fe2+/Mg ratios of olivine phenocryst cores are plotted versus those of three whole rocks in Fig. 6a. In this figure the whole-rock Fe2O3/FeO weight ratio is assumed to be 0·24 based on the Kilinc et al. (1983) expression for Fe speciation, an f(O2) of nickel–nickel oxide (NNO; Huebner & Sato, 1970), and a temperature of 1000°C (see below). In general, minimum olivine core Fe2+/Mg ratios increase with increasing rock Fe2+/Mg ratios and appear to indicate equilibrium between the olivine and the bulk rock based on a KD (=[Fe/Mg]ol/[Fe/Mg]rock) = 0·3 ± 0·03 (Roeder & Emslie, 1970; Wagner et al., 1995).
Pyroxene
Clinopyroxene occurs in Washburn volcanic rocks as isolated, euhedral to subhedral phenocrysts up to 6 mm across, as compact, often rounded glomerocrystic aggregates up to 1 cm in diameter, as a microphenocryst phase in the groundmass, and as subhedral to euhedral rims surrounding orthopyroxene cores in basaltic andesitic rocks. In dacitic rocks clinopyroxene phenocrysts may be surrounded by corona of amphibole. Orthopyroxene occurs as: (1) individual subhedral to euhedral lath-shaped phenocrysts 0·5–1·0 mm across; (2) as crystals in monomineralogic clots and the glomerocrysts described above; (3) as smaller crystals together with plagioclase and Fe–Ti oxides in reaction coronas surrounding amphibole phenocrysts of some dacitic rocks; (4) as microphenocrysts in the groundmass.
Pyroxene compositions are typical of calc-alkaline rock series. Clinopyroxenes (Wo34–46En43–52Fs5–16) are mainly augite, although a few phenocrysts have endodiopsidic cores, particularly in more magnesian-rich rocks, and orthopyroxenes are bronzite–hypersthene (Wo2–5En67–81Fs16–30; Table 4). All Washburn clinopyroxenes have low Al (Al/6 oxygens < 0·18) and Ti (Ti/6 oxygens < 0·02), suggesting crystallization under low-pressure, shallow-crustal conditions (Fig. 7a). Pyroxene zoning patterns are generally normal. Relative to cores, most rims are enriched in ferrosilite content by an amount <9%, and they have lower contents of Cr, Al, and less commonly Ti. In addition, a few analyzed pyroxene grains are reversely zoned. The majority of these are clinopyroxene phenocrysts with rimward enrichments in Mg (7% En) and less typically Ca (4% Wo) and Ti. Reverse zoning is less common in orthopyroxene and typically at the limit of analytical resolution (1% En). The origin of the Mg-rich clinopyroxene rims may be related to recharge and mixing with less evolved magmas or, as suggested by the lack of reverse zonation in the Wo component for many of these grains, late precipitation of Fe–Ti oxides in the groundmass and lowering of FeO in the liquid.
Compositions of pyroxene phenocryst cores are compared with whole-rock compositions in Fig. 6b and c. With the exception of an obviously xenocrystic grain in a low-Mg dacite (MW974; the only cpx grain observed in this particular thin section), the relationships illustrated are consistent with an equilibrium KD (=[Fe/Mg]pyx/[Fe/Mg]rock) between 0·2 and 0·3 for clinopyroxene, which is similar to values (0·20–0·25) found experimentally in 1 atm, silica-saturated and undersaturated arc lavas (Grove & Baker, 1984; Kennedy et al., 1990). Equilibrium KD values for orthopyroxene in Mg-rich basaltic andesitic rocks appear higher than for clinopyroxene, but probably reflect early precipitation of clinopyroxene followed by orthopyroxene in mafic magmas. Magmatic temperatures calculated for rims of touching clinopyroxene–orthopyroxene pairs in three lavas using the methods of Wood & Banno (1973) and Wells (1977) are in the range of 1012–1017°C for a basaltic andesite (MW9730), 981–999°C for an andesite (MW9721), and 967–968°C for a high-silica andesite (MW9743; Table 4).
Plagioclase
Plagioclase is the most abundant mineral in Washburn rocks, with the exception of a few dacitic and basaltic andesitic samples in which amphibole and clinopyroxene, respectively, are slightly more abundant. In all bulk-rock compositions plagioclase occurs as large (0·3–4·0 mm), euhedral to subhedral phenocrysts in isolation or in glomerocrysts, and as equant microphenocrysts and lath-shaped microlites in the groundmass. Phenocryst habits are generally lath-shaped, although grains become increasingly equant in more silica-rich rocks. There appears to be no compositional distinction between plagioclase grains in isolation or in glomerocrysts. Compositions of plagioclase phenocryst cores for the Washburn suite span from An83 to An40 (Table 5), although greater than two-thirds of phenocryst cores are between An65 and An50 (Fig. 8).
View larger version (27K):
[in this window]
[in a new window]
|
Fig. 8. Ternary components in plagioclase phenocryst cores and rims and groundmass (gmass) plagioclase as a function of host rock composition.
|
|
Although oscillatory zoning of small magnitude is common, overall core to rim zoning patterns of plagioclase phenocrysts are relatively uncomplicated; over half of all crystals analyzed are normally zoned <15 mol % An and <20% display reverse zoning of any magnitude. The maximum extent of reverse zoning observed is <12 mol % An and with very few exceptions it is <5 mol %. Groundmass crystals in all rock types have compositional ranges that are similar to the ranges of rim compositions and thus are either identical to, or slightly more Na rich than core compositions (Fig. 8).
The relationships between plagioclase phenocryst cores and whole-rock Ca/Na ratios are illustrated in Fig. 6d. With the exception of a few Ca-rich phenocrysts in the dacitic lavas, the relationships illustrated are consistent with an equilibrium KD (=[Ca/Na]plag/[Ca/Na]rock) between one and three for the andesitic and dacitic lavas. These values suggest early crystallization of most plagioclase cores under low to moderate magmatic water contents, based on the data of Sisson & Grove (1993), who showed that KD varies with melt water content, from 1·0 at anhydrous conditions to 5·5 for melts with 6% water at 2 kbar. The more Ca-rich cores may represent grains crystallized under higher water contents, unequilibrated crystals from higher temperature or pressure stages of evolution, xenocrysts derived from mixing with more mafic magmas, or low-density refractory solids retained from zones of deep-crustal melting and silicic magma production (see Feeley & Dungan, 1996). Because dacitic and andesitic rocks have relatively high proportions of these Ca-rich cores, whereas they are less abundant in more mafic rocks (e.g. Fig. 6d), we favor the latter hypothesis (see below). The relationships in Fig. 6d for the basaltic andesitic magmas suggest equilibrium KD values around one for most plagioclase cores. These values suggest crystallization from magmas with lower water contents relative to more evolved magmas or precipitation of plagioclase concurrently or following saturation of high-Ca clinopyroxene. Also shown in Fig. 6d are equilibrium relationships between plagioclase phenocryst rim and liquid (i.e. groundmass) Ca/Na ratios determined for five rocks by electron microprobe analyses. The plagioclase phenocryst rims in these rocks straddle or plot slightly above the 3·0 KD reference line. These relationships suggest equilibrium growth of plagioclase rims under higher water contents than for cores, perhaps because of residual water build-up during crystallization of predominantly anhydrous mineral assemblages.
Amphibole
Amphibole is present in more silica-rich Washburn rocks as phenocrysts ranging in size up to 4 mm. In many of these rocks amphibole is rimmed by opacite or less commonly by a fine-grained mass of plagioclase + Fe–Ti oxides ± pyroxenes, presumably as a result of breakdown during decompression (Foden & Green, 1992). In addition, amphibole phenocrysts in many samples are partially altered along cleavage surfaces and cracks to a fine-grained mass, perhaps bowlingite, and a few samples have crystals with cavernous interiors, presumably as a result of incomplete crystal growth. In these latter crystals alteration can be substantial. All Washburn amphiboles have (Ca + Na)M4 > 1·34 and (Na)M4 < 0·67 and thus are calcic amphiboles according to the classification scheme of Leake (1978) and Hawthorne (1981). Amphibole phenocrysts in andesitic rocks range from pargasite to pargasitic hornblende, whereas those in dacitic rocks are pargasitic hornblende to edenite (Leake, 1978; Hawthorne, 1981; Table 6).
Figure 7b plots AlIV vs (Na + K) cations in the A site of amphibole phenocrysts in Washburn rocks. Experimental studies demonstrate that AlIV exhibits correlated increases with (Na + K)A with rising crystallization temperature. The higher AlIV and (Na + K)A in amphiboles from the andesitic rocks relative to most grains in the dacitic rocks are consistent with their less evolved nature and inferred higher liquidus temperatures. A few amphibole crystals from the dacitic rocks have cores with high AlIV and (Na + K)A that plot at the high-temperature end of the experimental data arrays. These cores probably represent unequilibrated regions from higher-temperature (or pressure) stages of evolution because they are compositionally similar to phenocrysts in the andesitic rocks and are surrounded by normally zoned rims similar to most other amphibole crystals in the dacitic lavas.
|
CHEMICAL COMPOSITIONS
|
|---|
We analyzed 46 rocks from Washburn volcano for major and trace
element abundances by X-ray fluorescence (XRF) spectrometry.
Sample locations are shown in Fig.
2. On the basis of the XRF
analyses, a set of 17 rocks was selected for additional determination
of rare earth element (REE) and other trace element contents
by instrumental neutron activation analysis (INAA) and a further
subset of these (
n = 9) for determination of Sr and Nd isotopic
ratios by thermal ionization mass spectrometry (TIMS). All data
are reported in Table
2 and details of the analytical procedures
used are described in the Appendix.
Major elements
Variations in major element compositions of Washburn rocks are illustrated with respect to SiO2 in Fig. 9. According to the classification schemes of Peacock (1931), Peccerillo & Taylor (1976), and Le Maitre (1989), Washburn igneous rocks form a medium- to high-K, calc-alkaline suite (alkali–lime index 60; not shown) ranging in composition from basaltic andesitic through dacitic. In these respects the compositions of Washburn rocks share broad overall similarities with other Eocene calc-alkaline igneous rocks in the Absaroka Volcanic Province and elsewhere in the ‘Challis arc’ (e.g. Norman & Mertzman, 1991; Hooper et al., 1995; Dostal et al., 1998; Lindsay & Feeley, 1999; Morris et al., 2000). As is typical of calc-alkaline suites, MgO, Fe2O3, CaO, TiO2, and MnO decrease with increasing SiO2, whereas Na2O and K2O increase. Trends for Al2O3 and P2O5 are diffuse and show no obvious correlation with SiO2.
With respect to identification of magmatic sources, differentiation processes, and compositional evolution of the Washburn magmatic system, the important major element features are: (1) samples from the SW Washburn Range consist mainly of a bimodal package of basaltic andesitic and dacitic rocks, whereas stratigraphically higher units on Mt. Washburn range continuously from basaltic andesitic through dacitic rocks, but are dominated by andesitic compositions; (2) trends with respect to SiO2 for many elements are linear, except for MgO, Al2O3, and P2O5, which have more diffuse distributions of data; (3) at a given SiO2 content, rocks from Mt. Washburn tend to have lower MgO and higher Al2O3, respectively, than rocks from the SW Washburn Range.
Meen (1985) emphasized plots of K2O vs K2O/MgO and MgO for discriminating rocks at Independence volcano (Fig. 2), believed to be related exclusively by high-pressure fractional crystallization of mantle-derived parents in the absence of crustal contamination from those that interacted with crustal materials. In theory, fractional crystallization of magnesian minerals plus plagioclase leads to production of evolved liquids that fall within narrow coronal bands characterized by large increases in K2O compared with smaller increases in K2O/MgO and decreases in MgO (Fig. 10a). In contrast, contamination of magmas with low-K2O/MgO crustal melts produces trajectories with low slopes on such plots. Figure 10a illustrates a plot of K2O vs K2O/MgO for Washburn volcano rocks. Many samples from the SW Washburn Range define a narrow coronal band nearly identical to those formed by suites of rocks at Independence volcano inferred by Meen (1985) to be uncontaminated. The SW Washburn rocks that define this band cannot represent a single uncontaminated liquid line of descent, however. As discussed below, variations in trace element contents and isotopic ratios are inconsistent with this proposition. Furthermore, nearly all Washburn rocks, including the most mafic samples, form a shallow array in Fig. 10b that cuts Independence ‘uncontaminated’ trends at high angles. We interpret the Washburn array as a mixing line between high-MgO basaltic magmas and low-MgO crustal melts. This implies that the compositions of all Washburn rocks were affected by crustal interaction and none can be assumed to be direct uncontaminated differentiates of primary mantle-derived magmas.
Trace elements
Variations in trace element abundances of Washburn volcanic rocks display an interesting set of relationships that are similar to, but somewhat more complex than for major elements. The important features are as follows.
Many trace elements (e.g. Ba, Rb, Y, Zr, Sc) display overall linear, but diffuse variations with respect to SiO2, although highly compatible trace elements such as Cr and Ni scatter widely so that some intermediate composition rocks have relatively high concentrations of these elements (Fig. 11). These features do not reflect xenocrystic olivine as there is no correlation between modal percent olivine and Ni and Cr contents in the andesitic and dacitic rocks (Table 2). They are also not due to contamination during sample preparation because the samples were crushed in tungsten-carbide grinding bowls, which does not cause analytically significant contamination for Ni and Cr (or MgO; Johnson et al., 1999). It is noteworthy that, analogous to variations in MgO, rocks from Mt. Washburn generally have lower Ni and Cr contents than most equivalent bulk composition rocks from the SW Washburn Range. Moreover, the samples that form a narrow coronal band in Fig. 10a do not define a single liquid line of descent resulting from fractionation of magnesian minerals plus plagioclase in diagrams illustrating variation of Ni and Cr with SiO2. Accordingly, diagrams such as Fig. 10a may not uniquely identify magmas strictly related by simple closed-system fractional crystallization (see Meen, 1985; Meen & Eggler, 1987).
View larger version (31K):
[in this window]
[in a new window]
|
Fig. 11. Selected trace element contents in Washburn igneous rocks vs SiO2. Symbols as in Fig. 9. Symbols with center point in Ni and Cr diagrams are those defining coronal band in Fig. 10a. Dacitic rock types are distinguished on the basis of location and Sr, Zr, Y, and Rb contents. Dashed arrows through dacitic composition rocks in Ni and Cr vs SiO2 diagrams are linear regression lines.
|
|
Both Sr and Zr increase slightly with increasing SiO2 contents, indicating that fractionation of plagioclase and zircon (or crustal melting with these phases residual) were not major processes in the evolution of the Washburn suite as a whole. In contrast, Y decreases with increasing SiO2, indicating that fractionation of large amounts of clinopyroxene or amphibole, or contamination by crustal melts with low Y contents was important. On the basis of location and Sr, Zr, Y, and Rb contents, three groups of dacitic rocks can be distinguished at Washburn volcano; two from the SW Washburn Range, designated types 1 and 2, and one from Mt. Washburn, designated type 3. On a comparative basis, type 1 dacites (including one high-silica andesitic lava) have higher Rb, Zr, and Y, and lower Sr contents than type 2 dacites (Fig 10). Type 3 dacites have similarities to and differences from types 1 and 2 in that they have Y contents similar to those of type 2 dacites, Sr contents similar to or greater than those of type 1 dacites, and Zr and Rb contents intermediate between those of both types.
All Washburn volcanic rocks have LREE-enriched, chondrite-normalized REE patterns (Fig. 12). For the suite as a whole andesitic and dacitic rocks tend to have lower concentrations of HREE, higher concentrations of LREE, and flatter middle (MREE) to HREE patterns (e.g. Tb–Lu) than basaltic andesitic rocks. As a result of the first two features chondrite-normalized REE plots display fanning (cross-over) patterns with fulcrums centered between Nd and Sm. All Washburn rocks have either no Eu anomaly or small positive anomalies that increase from 1·00–1·03 for basaltic andesitic rocks to 1·01–1·15 for andesitic and dacitic rocks. These relationships are consistent with the increasing Sr concentrations with SiO2, in that both indicate that fractionation of plagioclase was not a major process during evolution of the Washburn suite. Furthermore, the decreasing HREE contents with increasing SiO2 wt % virtually require that differentiation at Washburn volcano involved mixing with, or assimilation of, crustal melts with low HREE contents. They cannot reflect fractionation of substantial amounts of amphibole, the only major phenocryst phase in Washburn rocks with mineral–melt Kd values >1 for the MREE and HREE. Fractionation of substantial amphibole results in depletion of MREE relative to HREE and the production of concave-upward REE patterns.
View larger version (28K):
[in this window]
[in a new window]
|
Fig. 12. Rare earth element abundances of Washburn rocks normalized to nonvolatile C1 chondrite from Anders & Ebihara (1982). Gd* is extrapolated from the LREE trend. Symbols as in Fig. 9.
|
|
Mid-ocean ridge basalt (MORB)-normalized trace element plots for Washburn basaltic andesitic rocks are characterized to some extent by features considered diagnostic of subduction-related magmas (Fig. 13; Pearce, 1982). For example, relative to MORB, basaltic andesites have strong enrichments in large ion lithophile elements (LILE; Sr, K, Rb, Ba, and Th) and depletions in the high field strength elements (HFSE) Ta, Nb, and Ti. The HFSE Zr and Hf, however, are not depleted relative to MORB. High-K2O basaltic rocks from the shoshonitic Sunlight volcano, eastern AVP (Fig. 1), are further enriched in LILE, but not HSFE, relative to Washburn basaltic andesites (Fig. 13b). Least evolved (MgO 5·93 wt %; Meen & Eggler, 1987) high-alumina tholeiitic basaltic andesite from Independence volcano, eastern AVP, shows variable enrichments and depletions in LILE relative to Washburn basaltic andesites, although with the exception of Ti, HFSE are depleted. Plots for Washburn andesitic and dacitic rocks are enriched in both LILE and Zr and Hf relative to basaltic andesitic rocks, although they are depleted in Ti, Y, and Yb.
View larger version (46K):
[in this window]
[in a new window]
|
Fig. 13. MORB-normalized trace element diagrams. (a) Trace element abundance patterns for rocks from Mt. Washburn. (b) Trace element abundance patterns for rocks from the SW Washburn Range contrasted with mafic basaltic andesite from Independence volcano (2080; Meen & Eggler, 1987) and high-K basalt from Sunlight volcano (T. C. Feeley, unpublished data, 2001). (c) The range in trace element abundances for model Washburn primary magma (method of calculation in text) contrasted with subalkaline mafic rocks from the Colville Igneous Complex (Morris et al., 2000), the Challis volcanic field (Norman & Mertzman, 1991), and the Crazy Mountains (Dudas, 1991). (d) Model Washburn primary magma contrasted with high-K basalt from Sunlight volcano (T. C. Feeley, unpublished data, 2001), basaltic andesite from Independence volcano (Meen & Eggler, 1987), and alkaline basalt from the Crazy Mountains (Dudas, 1991). In (a) and (b), symbols as in Fig. 9. It should be noted in (c) and (d) that not all elements are available for all samples. In all diagrams normalizing values and element order after Pearce (1982). Also shown for reference in all diagrams are typical patterns for tholeiitic ocean-island basalt (OIB) and calc-alkaline volcanic arc basalt (Pearce, 1982).
|
|
Sr and Nd isotopes
Initial 87Sr/86Sr and 143Nd/144Nd isotopic ratios corrected for 53 my of in situ growth of radiogenic Sr and Nd for nine Washburn rocks are presented in Table 2 and illustrated in Fig. 14. Relative to ‘bulk Earth’ all rocks have high Sr and low Nd isotopic ratios and thus plot within the ‘enriched’ quadrant in Fig. 14a. The data define a negatively correlated array extending from a field defined by ultramafic xenoliths brought up in Eocene alkalic magmas in the Crazy Mountains, 150 km north of Washburn volcano, to fields defined by Wyoming Province Archean granulite-facies rocks brought up as xenoliths in Cenozoic magmas (Fig. 14a; Leeman et al., 1985; Dudas et al., 1987) and Archean supracrustal amphibolites and granitoids that dominate the exposed basement near Washburn volcano (Meen, 1987a; Wooden & Mueller, 1988). Washburn mafic rocks have slightly higher 87Sr/86Sr ratios but similar 143Nd/144Nd ratios compared with mafic rocks from Independence and Sunlight volcanoes in the eastern belt of the AVP (Meen & Eggler, 1987; T. C. Feeley, unpublished data, 2001). Although the Washburn rocks have MORB-normalized trace element patterns characterized by depletions in Ta and Nb (Fig. 13), they plot well below the field defined by modern subduction-related basalts because of their low Nd isotopic ratios. This feature is characteristic of many early- to late-Cenozoic magmatic rocks in the Wyoming Province and is considered to reflect derivation from or interaction with ancient subcontinental lithospheric mantle (e.g. Fraser et al., 1985; Dudas et al., 1987; Meen & Eggler, 1987; Scambos, 1991; MacDonald et al., 1992; O’Brien et al., 1995).
The Nd and Sr isotopic data indicate that crustal material was important in the genesis of Washburn magmas, and differences in crustal composition may provide an explanation for the contrasts b
TOP
PETROGRAPHY AND SILICATE MINERAL...
CHEMICAL COMPOSITIONS
