ORIGINAL PAPER

The generation of a diverse suite of Late Pleistocene and Holocenebasalt through dacite lavas from the northern Cascade arcat Mount Baker, Washington

Troy David Baggerman • S. M. DeBari

Received: 25 August 2009 / Accepted: 5 April 2010 / Published online: 1 May 2010

� Springer-Verlag 2010

Abstract Mt. Baker is a dominantly andesitic stratovol-

cano in the northern Cascade arc. In this study, we show that

the andesites are not all derived from similar sources, and

that open-system processes were dominant during their

petrogenesis. To this end, we discuss petrographic observa-

tions, mineral chemistry, and whole rock major and trace

element chemistry for three of Mt. Baker’s late Pleistocene to

Holocene lava flow units. These include the basalt and

basaltic andesite of Sulphur Creek (SC) (51.4–55.8 wt%

SiO2, Mg# 57–58), the Mg-rich andesite of Glacier Creek

(GC) (58.3–58.7 wt% SiO2, Mg# 63–64), and the andesite

and dacite of Boulder Glacier (BG) (60.2–64.2 wt% SiO2,

Mg# 50–57). Phenocryst populations in all units display

varying degrees of reaction and disequilibrium textures

along with complicated zoning patterns indicative of open-

system processes. All lavas are medium-K and calc-alkaline,

but each unit displays distinctive trace element and REE

characteristics that do not correlate with the average SiO2

content of the unit. The mafic lavas of SC have relatively

elevated REE abundances with the lowest (La/Yb)N (*4.5).

The intermediate GC andesites (Mg- and Ni-rich) have the

lowest REE abundances and the highest (La/Yb)N (*6.7)

with strongly depleted HREE. The more felsic BG lavas have

intermediate REE abundances and (La/Yb)N (*6.4). The

high-Mg character of the GC andesites can be explained by

addition of 4% of a xenocrystic olivine component. How-

ever, their depleted REE patterns are similar to other high-

Mg andesites reported from Mt. Baker and require a distinct

mantle source. The two dominantly andesitic units (GC and

BG) are different enough from each other that they could not

have been derived from the same parent basalt. Nor could

either of them have been derived from the SC basalt by

crystal fractionation processes. Crystal fractionation also

cannot explain the compositional diversity within each unit.

Compositional diversity within the SC unit (basalt to basaltic

andesite) can, however, be successfully modeled by mixing

of basalt with compositions similar to the dacites in the BG

unit. Given that the BG dacites erupted at *80–90 ka, and a

similar composition was mixed with the SC lavas at 9.8 ka,

the process that produced this felsic end-member must have

been repeatedly active for at least 70 ka.

Keywords Cascade Arc � Mount. Baker � REE �Geochemistry � Petrology � Andesite � Mixing �Fractionation

Introduction

Calc-alkaline andesite lavas are a characteristic lava type

associated with many arc volcanoes and are the most

common lava type erupted in the northern Cascade arc

(Hildreth 2007). Many types of differentiation processes

have been called upon to explain the development of these

lavas as they migrate through Earth’s crust. These include

crystal fractionation (e.g. Grove and Baker 1984), magma

mixing (e.g. Eichelberger 1975), assimilation of crustal

material (e.g. Mason et al. 1996) and hybrid models with

crystal fractionation coupled with magma recharge and

mixing (Streck et al. 2002; Schaaf et al. 2005). In some

Communicated by T. L. Grove.

Electronic supplementary material The online version of thisarticle (doi:10.1007/s00410-010-0522-2) contains supplementarymaterial, which is available to authorized users.

T. D. Baggerman (&) � S. M. DeBari

Geology Department, MS 9080, Western Washington

University, Bellingham, WA 98225, USA

e-mail: troy22db@gmail.com

123

Contrib Mineral Petrol (2011) 161:75–99

DOI 10.1007/s00410-010-0522-2

cases, complex models are described in which each of the

above processes works in conjunction to form andesite

(Hildreth and Moorbath 1988). Each of the above processes

may leave behind chemical and textural evidence in

andesitic rocks, thereby providing the means to distinguish

between the various models of andesite generation.

Mount Baker is a dominantly andesitic volcano in the

northernmost segment of the Cascade arc. It has been

historically active (Eichelberger et al. 1976; Hildreth et al.

2003) and is situated near several population centers

(Fig. 1), yet studies involving the origin and evolution of

its intermediate lavas are scarce. Previous studies were

inhibited by lack of a stratigraphic framework (e.g.

McKeever 1977; Swan 1980). The breakthrough came with

the detailed eruptive history and geochronology of Mt.

Baker by Hildreth et al. (2003). These authors presented a

chronology of the main eruptive components that includes

locations, descriptions and ages of individual units of the

Mount Baker volcanic field. Their work is summarized in a

geologic map of all volcanic units (Fig. 2). Their work also

provides an extensive major element data set for lavas from

the entire volcanic field along with petrographic observa-

tions of various reaction textures. However, Hildreth et al.

(2003) include little interpretation of the petrologic data

and observed textures. Their chronology and unit descrip-

tions form the foundation for this study.

This study reports distinguishing geochemical charac-

teristics in a set of basalt through dacite lavas erupted from

Mount Baker and describes processes that may be

responsible for the generation of these intermediate lava

types. Presented within this paper are petrographic obser-

vations, whole rock major oxide, trace and REE concen-

trations, along with mineral chemistry, for three Late

Pleistocene and Holocene eruptive units from Mount

Baker. Units include the basalt of Sulphur Creek (51.4–

58.5 wt% SiO2), the andesite of Glacier Creek (58.1–

58.5 wt% SiO2) and the andesite and dacite of Boulder

Glacier (56.6–64.4 wt% SiO2) (units, names and compo-

sitional ranges stated above are from Hildreth et al. 2003).

Geology and tectonic setting

Mount Baker is located in northern Washington State in the

northern Cascade volcanic arc, about 50 km east of Bell-

ingham Bay (Fig. 1). Arc volcanism in the Northern Cas-

cades is associated with the oblique subduction of the warm

and young Juan de Fuca plate beneath the North American

plate. The plates converge at a rate of 3–4 cm/year (Heaton

and Kanamori 1984) with an estimated subduction angle of

20� in Washington and northern Oregon (McCrory et al.

2004). P-wave studies suggest that the crust beneath Mount

Baker is 40–45 km thick (Mooney and Weaver 1989;

Ramachandran et al. 2006) with the subducting slab at a

depth of 100–120 km (Bostock and VanDecar 1995).

The Cascade arc is segmented due to structural changes in

the subduction system and along-arc magmatic trends

(Hildreth 2007; Guffanti and Weaver 1988). Mount Baker is

located in the northernmost segment, the Garibaldi belt

(Fig. 1). The arc as a whole is characterized by a decrease in

mafic magma production to the north, where primitive

magmas become scarce in the Garibaldi belt (Hildreth 2007).

Arc-parallel variations in magmatism have been credited to

variables such as age of the subducting slab (Green and Harry

1999), variations in slab-derived fluid fluxes, degrees of

partial melting of different mantle sources and variations in

mantle H2O (Green and Sinha 2005).

Mount Baker is currently the prominent feature (3,268 m)

of a multivent Quaternary volcanic field (Hildreth et al.

2003), including the 1.15 Ma rhyodacitic Kulshan caldera

(Hildreth 1996). The volcano is blanketed with glaciers

(*40 km2, Post et al. 1971), which are responsible for par-

tial erosion of many of Mount Baker’s volcanic units. The

Fig. 1 Cascade magmatic arc and tectonic setting. Black linesindicate the five segments of the Cascade arc, as described by

Guffanti and Weaver (1988). Mount Baker (MB) is in the northern-

most segment, known as the Garibaldi belt. Locations of major

volcanic centers throughout the Cascade arc are denoted by triangleswith abbreviated labels; MM, Meager Mountain; MC, Mount Cayley;

MG, Mount Garibaldi; MB, Mount Baker; GP, Glacier Peak;

MR, Mount Rainier; MSH, Mount St. Helens; MA, Mount Adams;

SVF, Simcoe Volcanic Field; MH, Mount Hood; MJ, Mount Jefferson;

TS, Three Sisters; NV, Newberry Volcano; CLV, Crater Lake

Volcano; MMc, Mount McLoughlin; MLV, Medicine Lake Volcano;

MS, Mount Shasta; LVC, Lassen Volcanic Center. Rates of subduction

are from McCrory et al. (2004)

76 Contrib Mineral Petrol (2011) 161:75–99

123

volcanic field is built on amphibolite-facies Mesozoic and

Paleozoic rocks of oceanic affinity that were emplaced by

strike slip faulting during the Paleogene (Miller et al. 2003;

Tabor et al. 2003). Volcanic products during the past 1.3 Ma

have been linked to five main stages of volcanic activity that

have been outlined by Hildreth et al. (2003): (1) pre Kulshan

Caldera activity (1.3–1.15 Ma); (2) early post Kulshan

Caldera activity (1.15–0.99 Ma); (3) widely scattered

intracaldera and extracaldera activity, dominantly andesitic

(0.9–0.5 Ma); (4) formation of Black Buttes stratovolcano

and its satellites (0.5–0.2 Ma); and (5) formation of Mount

Baker stratovolcano and its satellites (0.1 Ma to present).

This study focuses on three late Pleistocene to Holocene

units that are used to investigate the generation of inter-

mediate products during the formation of the Mount Baker

stratocone and the associated satellite vents (Fig. 2). The

units are chosen to represent a range of ages and compo-

sitions erupted since the late Pleistocene. The andesite of

Glacier Creek and the andesite of Boulder Glacier

(includes andesite and dacite) are products of the late

Pleistocene stratocone-forming lavas, and the basalt of

Sulphur Creek (includes basalt and basaltic andesite) is the

result of a more recent Holocene flank eruption. We use the

unit names of Hildreth et al. (2003), recognizing that their

compositional range is broader than the unit name.

The basalt of Sulphur Creek is the product of a mono-

genic eruption on the southwest flank of the volcano

7.8 km south of the main edifice (Schreibers Meadow)

(Fig. 2). The eruption location is marked by a heavily

vegetated *100-m tall cinder cone (Schreiber’s Meadow

cinder cone) with two small lakes in the crater marking two

separate vent locations. A lava flow unit emanates from the

cone and extends *12 km down valley to the SE to Baker

Lake. Sub-aqueous lavas are described near the west side

of Baker Lake (Tucker and Scott 2009). Exposures near the

cinder cone are sparse due to a thick layer of Holocene

debris flows (Hildreth et al. 2003), but when located are

blocky spine-shaped outcrops. Other exposures are located

Fig. 2 Geologic map of Mount Baker and related volcanic units from Hildreth et al. (2003). Units discussed in this paper (Sulphur Creek,

Glacier Creek and Boulder Glacier) are highlighted

Contrib Mineral Petrol (2011) 161:75–99 77

123

along road cuts or stream gorges and are generally massive

to blocky and vesicular. The unit consists of a range of

compositions from basalt to andesite (Green 1988; Hildreth

et al. 2003; Moore and DeBari 2009; and this study), with

the most mafic compositions closest to the terminus of the

lava flow. Radiocarbon dates calibrated to calendar years

suggest an age of 9.8 ka (Scott et al. 2001), making it the

youngest lava flow in the Mount Baker volcanic field.

Carbon buried by scoria associated with the Schreibers

Meadow cinder cone has been dated to 8,750 ± 50 and

8,830 ± 50 radiocarbon years bp (Scott et al. 2001).

Green (1988) noted basalt enclaves ranging in diameter

from \1 mm to 15 cm in basaltic andesite of the Sulphur

Creek unit. He used these enclaves to conclude that the

basaltic andesite is a result of mixing and quenching of

basalt within a more differentiated basaltic andesite. The

Sulphur Creek unit has the most Si-poor compositions in

the suites of rocks collected for this study; however, they

are not the most Mg-rich lavas erupted from Mount Baker

(Moore and DeBari 2009).

The andesite of Glacier Creek is interpreted by Hildreth

et al. (2003) to be a single canyon-confined flow (Fig. 2),

but the location of the source vent is unknown. The flow

extends *8 km northwest of the summit of Mount Baker

in the Glacier Creek gorge. Exposures within the gorge are

up to 80 m thick and range from sub-vertical, columnar

jointed to massive or platy outcrops. Exposures at higher

elevations confirmed to be the same unit by whole rock

chemistry can be found as sparse remnants on the western

side of the Coleman glacier. Hildreth et al. (2003) descri-

bed a compositionally similar flow at a locality known as

the Roman Wall. Hildreth et al. (2003) report one K–Ar

date for the Glacier Creek unit of 14 ± 9 ka.

The andesite of Boulder Glacier is a series of stacked

andesite and dacite flows that cap Boulder Ridge and

occupy the upper valley of Boulder Creek near the termi-

nus of the Boulder Glacier (Fig. 2). Exposures consist of

thick stacks (up to 200 m) of multiple columnar jointed to

platy lava flows. Some locations expose basal sections of

flow breccias. Thicknesses of single flows range from 5 m

up to 50 m. Hildreth et al. (2003) estimate the entire

Boulder Glacier unit is composed of at least 25 separate

flows. It is the oldest unit recognized to erupt from the

main edifice of Mount Baker with two K–Ar dates, each

from a different flow. One flow is dated at 80 ± 14 ka, and

the other is dated at 90 ± 52 ka (Hildreth et al. 2003).

Analytical methods

The units were sampled extensively to represent the spatial

distribution as mapped by Hildreth et al. (2003) (sample

locations are shown in ‘‘Supplementary material 6’’).

Samples were crushed by hammer to 2-cm pieces. A

minimum of 50 g of crushed pieces were chipped by

tungsten carbide crushing plates and prepped for X-ray

fluorescence (XRF) and Inductively Coupled Plasma Mass

Spectrometry (ICP-MS) analyses.

XRF

Seventeen samples were analyzed at the Washington State

University (WSU) GeoAnalytical Laboratory for major

element oxides and eighteen trace elements (Sc, V, Ni, Cr,

Ba, Sr, Zr, Y, Rb, Nb, Ga, Cu, Zn, Pb, La, Ce, Th, Nd),

using a ThermoARL Advant’XP? sequential wavelength

dispersive X-ray fluorescence spectrometer. Analytical

methods, accuracy and precision are described by Johnson

et al. (1999). Rock powders were prepared at Western

Washington University (WWU) by grinding fresh rock

chips in a tungsten carbide shatterbox for 4 min. The glass

beads were prepared by fusing a 2:1 ratio of dry flux to

dry rock powder in a pre-heated furnace at 1,000�C for

10 min. Beads were then reground in the tungsten carbide

shatterbox and re-fused as above to ensure complete

homogeneity.

ICP-MS

Thirty-two samples were analyzed at the WSU GeoAna-

lytical Laboratory for 14 rare earth elements (REE) and 13

additional trace elements (Ba, Th, Nb, Y, Hf, Ta, U, Pb,

Rb, Cs, Sr, Sc, Zr) using an Agilent 4500? Inductively

Coupled Plasma Mass Spectrometer (ICP-MS). Samples

were prepared at WWU according to the following proto-

col. Crushed rock material was ground to a fine powder in

an alumina-ceramic shatterbox for 5 min. Glass beads were

prepared by fusing a 1:1 ratio of dry flux to dry rock

powder in a pre-heated furnace at 1,000�C for 10 min.

Beads were then reground to a fine powder in the alumina-

ceramic shatterbox to ensure homogeneity. Analytical

techniques, accuracy and precision for the ICP-MS analy-

ses are available from the WSU GeoAnalytical Laboratory

at (http://www.sees.wsu.edu/Geolab/note/icpms.html).

Electron microprobe

Compositions of olivine, pyroxene, plagioclase and Fe–Ti

oxide minerals were determined by electron microprobe

analyses of nine polished thin sections (three thin sections

for each unit). Thin sections used for microprobe analyses

were chosen to represent all minerals and the range of

textures observed in each unit. Analyzed minerals were

selected in an attempt to collect data from the entire size and

textural range observed in each thin section. The analyses

were performed at the University of Washington, using a

78 Contrib Mineral Petrol (2011) 161:75–99

123

JEOL Superprobe 733 electron microprobe equipped with

four wavelength dispersive spectrometers (WDS) and an

energy dispersive spectrometer (EDS). Calibration was

performed using a standardized set of natural and syn-

thetic minerals. An accelerating voltage of 15 kV was

used for all minerals. A 3 l beam diameter with a beam

current of 10 nA was used for feldspar analysis, while a

beam diameter of \1 l and a beam current of 15 nA was

used for analyses of olivine, pyroxene and oxide minerals.

Peak counting times range from 20 to 40 s and are

dependent on how long it took to achieve a 0.4% statis-

tical error. Analytical error is\3% for major elements and

\8% for minor elements. ZAF corrections were applied

to the data based on the methods of Armstrong (1988),

and Fe–Ti oxides were further corrected using the meth-

ods of Evans et al. (2006).

Petrography and mineral chemistry

Mineral chemistry and textural relationships are complex

and highly variable in all units. It is not uncommon to see

highly corroded phenocrysts with strong chemical disequi-

librium textures situated adjacent to a phenocryst of the

same variety in semi-pristine condition. Petrographic

characteristics of each unit are summarized in Table 1.

Compositions of each phenocryst phase tend to have com-

positional similarity between units and wide compositional

Table 1 Petrographic summary

Unit Phenocryst

(%)

Phenocryst modes Size range

(mm)

% with

reaction

textures

Types of zoning

patterns

Other

basalt of SulphurCreek

20–50 Olivine: 5–15% 0.5–2.5 50 Normal, reverse,

oscillatory

Sometimes mantled by

plag ? opx

Plagioclase: 65–80% 0.5–3.5 70 Normal, reverse,

oscillatory

Clinopyroxene: 5–15% 0.5–3.0 50 Reverse Ophitic texture with

plagioclase

overgrowths

Orthopyroxene: 5–20% 0.5–2.5 50 NA

Fe–Ti oxide: 1–2% \1.0

Trace Cr spinel

andesite of GlacierCreek

50–60 Olivine: 4–8% 1.0–3.0 90 Normal, rare reverse Elevated NiO, often

mantled by

plag ? opx, bimodel

size distribution

Plagioclase: 70–80% 0.1–0.8, 1.2–4.2 65 Normal, reverse,

oscillatory

Bimodal size

distribution,

complicated sieve

textures

Clinopyroxene: 6–10% 0.1–2.0 80 NA

Orthopyroxene:

10–15%

0.1–2.0 80 Normal Ophitic texture with

plagioclase

overgrowths

Fe–Ti oxide: 1–2% \1.0

Trace apatite

andesite of BoulderGlacier

50–60 Olivine: \1% in 6

samples

0.5–1.5 50 Normal Mantled by plag ? opx

Plagioclase: 65–90% 0.2–3.5 75 Normal, oscillatory

Clinopyroxene: 1–10% 0.1–1.5 40 NA Ophitic texture with

plagioclase

overgrowths

Orthopyroxene: 7–30% 0.1–2.0 40 NA Ophitic texture with

plagioclase

overgrowths

Fe–Ti oxide: 2% \1.0

Trace apatite

Percentage values for the Sulphur Creek, Glacier Creek and Boulder Glacier units are based on 1000 point count analysis on thin section samples

Contrib Mineral Petrol (2011) 161:75–99 79

123

ranges within units (Fig. 3; Tables 2, 3, 4). A full list of

mineral compositions can be found in ‘‘Supplementary

materials 1–4’’.

basalt of Sulphur Creek

The Sulphur Creek unit is porphyritic to strongly porphyritic

with 20–50% phenocrysts and microphenocrysts in a

hypocrystalline to cryptocrystalline groundmass. Plagio-

clase constitutes 65–80% of the phenocryst and microphe-

nocryst population and occurs as 0.5–3.5 mm sub- to

euhedral tabular crystals. Equant to elongate orthopyroxene

phenocrysts and microphenocrysts (5–20% of the pheno-

cryst population) are 0.5–3.0 mm in length. Clinopyroxene

phenocrysts (5–15% of the phenocryst population) are also

0.5–3.0 mm in length and occasionally show ophitic textures

with intergrowths of plagioclase grains. The Sulphur Creek

unit is characteristically rich in olivine phenocrysts and

microphenocrysts compared to the other units (5–15% of the

phenocryst population). Olivine crystals are typically

0.5–2.5 mm in length and are sub- to euhedral hopper-

shaped crystals. The groundmass is composed of 70–90%

pilotaxitic plagioclase, 2–10% olivine and 2–10% orthopy-

roxene with 1–5% glass. Oxide minerals make up 1–2% of

the Sulphur Creek unit and occur as groundmass phases or as

inclusions in phenocrysts.

Olivine phenocrysts are typically isolated and often

reveal normal and complicated zoning patterns. For exam-

ple, two grains with multiple point analyses (SC-7TB Oa

and SC-6TB Oa) show oscillatory zoning patterns with

compositional reversals. In these two grains, Fo composi-

tions increase from Fo80 and Fo84 in the cores to Fo84 and

Fo86, respectively, in the middle section of the transect, and

then decrease to Fo77 and Fo78, respectively, at the rim

(‘‘Supplementary material 1’’). Nearly half of the olivine

crystal population displays varying degrees of chemical

disequilibrium textures with corroded and embayed grain

boundaries (Fig. 4). Less commonly, disequilibrium tex-

tures include grains that are mantled with orthopyroxene

and plagioclase microphenocrysts. Nearly all olivine crys-

tals have inclusions of Fe–Ti oxide minerals, while some

have Cr spinel inclusions. Olivine core compositions are

Fo86–64 and rim compositions are Fo77–65. Compositions

decrease by Fo1–14 from core to rim in normally zoned

grains. One olivine groundmass grain in sample SC-6TB is

anomalously Mg-poor, with a composition of Fo60. Sulphur

Creek olivine phenocrysts can be distinguished from those

of Glacier Creek and Boulder Glacier by characteristically

high CaO concentrations (average 0.21 wt% CaO).

Clinopyroxene and orthopyroxene phenocrysts com-

monly display disequilibrium textures such as corroded and

embayed rims (Fig. 4) and contain inclusions of Fe–Ti

oxides. In contrast to olivine and plagioclase, pyroxene

compositions in the Sulphur Creek unit are homogenous.

Orthopyroxene compositions are Wo3–4En65–70Fs26–31 with

Mg#s (Mg# = 100 9 (Mg/(Mg ? FeT)) of 67–73. Clino-

pyroxene compositions are Wo39–41En42–43Fs15–17 with

Mg# 70–75.

Plagioclase phenocrysts ([1 mm) contain inclusions of

Fe–Ti oxides, whereas microphenocrysts (0.5–1.0 mm) are

mostly inclusion-free. About 70% of the plagioclase

55657585% Fo

Boulder GlacierGlacier CreekSulphur Creek

304050607080% An

Boulder Glacier Glacier CreekSulphur Creek

Plagioclase

A

B

Fig. 3 Graphical representation of mineral compositions. a Forsterite

(Fo) content of olivine. b Anorthite (An) content of plagioclase. Solidshapes represent core compositions, while open shapes represent rim

compositions. Brackets connect core and rim analyses from single

phenocrysts

Table 2 Representative olivine compositions

Label SiO2 FeO* MnO NiO MgO CaO Total Fo

Sulphur Creek

SC-3TB OAc 40.1 13.8 0.2 0.2 45.1 0.2 99.6 85.3

SC-3TB OCc 37.1 27.1 0.4 0.0 33.2 0.3 98.0 68.6

SC-6TB OAm 40.4 12.9 0.3 0.2 45.6 0.2 99.6 86.3

SC-6TB OBr 37.5 26.9 0.5 0.1 34.0 0.2 99.2 69.2

SC-7TB OAc 39.2 18.6 0.3 0.1 41.6 0.2 100.0 80.0

SC-7TB OBr 37.9 24.7 0.4 0.1 36.1 0.2 99.4 72.3

Glacier Creek

GC-3TB OAc 39.9 13.5 0.2 0.4 45.5 0.1 99.6 85.7

GC-3TB OAr 38.2 21.2 0.4 0.3 39.0 0.1 99.2 76.7

GC-6TB OBc 39.8 13.7 0.2 0.3 45.4 0.1 99.6 85.5

GC-6TB OCc 40.0 13.5 0.2 0.3 45.5 0.1 99.6 85.8

GC-12TB OAr 37.7 25.0 0.4 0.2 35.9 0.2 99.4 71.9

GC-12TB OBc 37.2 28.3 0.5 0.2 33.0 0.1 99.4 67.5

Boulder Glacier

BG-3TB OAc 38.0 23.4 0.5 0.1 37.7 0.1 99.8 74.2

BG-3TB OAm 38.1 24.9 0.5 0.1 36.4 0.1 100.0 72.3

Individual mineral grains are assigned a letter followed by an addi-

tional letter indicating a location of the analyses (c—core, r—rim).

For example, GC-3TB OBc would be a core analyses in olivine B

from sample GC-3TB. Analytical error is\3% for major elements and

\8% for minor elements

80 Contrib Mineral Petrol (2011) 161:75–99

123

phenocryst population displays reaction textures such as

sieved rims and/or cores, and embayed and corroded rims

(Fig. 4). Microphenocrysts display similar textures but

with less frequency. Ninety percent of plagioclase pheno-

crysts and microphenocrysts display moderate to weak

patchy concentric normal zoning patterns, with *10% of

plagioclase phenocrysts displaying oscillatory and reversed

chemical zoning patterns. Phenocrysts compositions are

An47–70 with An51–70 cores and An47–62 rims. Normally

zoned plagioclase phenocrysts decrease by an average of

An6 from core to rim. Plagioclase microlites in the

groundmass are An45–57 and are comparable to

Table 3 Representative pyroxene and plagioclase compositions

Label SiO2 Al2O3 TiO2 FeO* MnO MgO CaO Na2O Cr2O3 Total Mg# Wo En Fs

Clinopyroxene

SC-3TB CpxBc 51.74 1.93 0.71 10.73 0.34 14.37 19.77 0.36 0.00 99.94 70.48 41.07 41.53 17.40

SC-3TB CpxBr 52.25 2.01 0.81 9.12 0.30 15.13 19.35 0.42 0.02 99.41 74.73 40.72 44.30 14.98

SC-6TB CpxAr 50.94 2.01 0.80 9.65 0.25 14.77 19.48 0.42 0.00 98.31 73.18 40.96 43.21 15.84

GC-3TB CpxAc 53.30 1.14 0.34 10.80 0.40 14.04 19.49 0.35 0.01 99.88 69.86 41.07 41.16 17.76

GC-6TB CpxBc 51.97 1.66 0.59 10.47 0.30 14.30 19.52 0.40 0.02 99.23 70.89 41.02 41.81 17.17

GC-12TB CpxBc 50.96 1.69 0.69 11.45 0.33 13.90 18.95 0.41 0.00 98.40 68.39 40.12 40.95 18.92

BG-3TB CpxBc 53.05 1.23 0.43 10.05 0.34 13.98 20.76 0.36 0.00 100.21 71.26 43.20 40.48 16.32

BG-7TB CpxCc 52.76 1.27 0.45 9.29 0.43 14.29 20.81 0.37 0.00 99.68 73.28 43.40 41.47 15.13

BG-10TB CpxAc 52.54 1.69 0.59 10.56 0.42 13.83 20.26 0.37 0.00 100.26 70.01 42.43 40.30 17.26

Orthopyroxene

SC-3TB OpxAc 52.19 0.94 0.41 19.52 0.62 22.63 1.69 0.05 0.01 98.06 67.39 3.49 65.04 31.47

SC-3TB Opxclot 54.00 0.93 0.36 16.80 0.47 25.25 1.70 0.06 0.00 99.56 72.82 3.40 70.35 26.26

SC-6TB OpxAc 52.72 0.82 0.35 19.98 0.49 22.70 1.69 0.04 0.00 98.79 66.95 3.46 64.63 31.91

GC-6TB OpxBc 53.78 0.88 0.40 19.48 0.46 22.87 1.77 0.05 0.01 99.69 67.67 3.62 65.21 31.16

GC-12TB OpxAc 53.74 0.98 0.32 16.42 0.42 25.17 1.48 0.01 0.00 98.53 73.21 3.01 71.01 25.99

GC-12TB OpxBr 53.85 1.96 0.54 17.88 0.38 23.36 1.44 0.04 0.04 99.50 69.96 3.01 67.85 29.14

BG-3TB OpxAc 51.97 0.70 0.28 22.68 0.64 20.82 1.60 0.02 0.02 98.72 62.07 3.31 60.01 36.68

BG-3TB OpxBc 53.39 0.48 0.12 20.88 0.67 22.47 1.26 0.02 0.01 99.32 65.73 2.58 64.04 33.38

BG-7TB OpxCc 53.50 0.64 0.33 21.43 0.79 21.89 1.38 0.03 0.01 99.99 64.55 2.84 62.72 34.45

Label SiO2 Al2O3 FeO* CaO Na2O K2O SrO Total An

Plagioclase

SC-3TB PAc 54.8 27.4 0.5 10.8 2.5 0.3 0.1 96.5 70.1

SC-3TB PBr 55.9 26.5 0.6 9.3 5.8 0.4 0.1 98.5 46.9

SC-6TB PAc 51.3 30.5 0.4 13.7 3.8 0.1 0.1 100.0 66.9

SC-6TB PAr 53.3 27.8 0.6 10.8 5.2 0.3 0.1 98.0 53.3

SC-7TB PBc 55.7 27.5 0.6 10.1 5.4 0.3 0.2 99.9 50.8

SC-7TB PBr 51.8 28.5 0.7 11.8 4.5 0.2 0.2 97.7 59.1

GC-3TB PAc 51.8 29.6 0.6 12.9 4.2 0.2 0.2 99.3 63.0

GC-3TB PAr 57.0 26.0 0.6 8.9 6.0 0.5 0.2 99.3 45.0

GC-6TB PAc 54.4 27.0 0.6 11.2 5.4 0.3 0.1 99.1 53.4

GC-6TB PBc 48.7 31.6 0.5 15.2 2.8 0.1 0.1 99.0 74.8

GC-12TB PAm 50.6 30.3 0.6 13.6 3.4 0.2 0.2 98.9 68.7

GC-12TB PBc 57.5 26.0 0.6 8.9 6.1 0.5 0.2 99.7 45.0

BG-3TB PBc 51.7 30.0 0.5 13.2 3.8 0.2 0.1 99.4 65.9

BG-3TB PCc 55.9 26.5 0.4 9.7 5.8 0.4 0.1 98.7 47.9

BG-7TB PAc 52.9 29.5 0.5 12.4 4.5 0.2 0.1 100.1 60.5

BG-7TB PAm 60.7 23.8 0.7 6.9 6.9 0.7 0.2 99.8 35.4

BG-10TB PAc 52.6 29.3 0.5 12.4 4.3 0.2 0.1 99.4 61.6

BG-10TB Pgm 64.5 19.6 1.4 4.6 5.4 2.6 0.1 98.2 31.8

Mineral labeling is as described in Table 2. Analytical error is \3% for major elements and \8% for minor elements

Contrib Mineral Petrol (2011) 161:75–99 81

123

compositions of phenocryst rims. Overall, plagioclase

compositions and textures are highly variable and complex.

Oxide minerals in the Sulphur Creek unit are repre-

sented by Cr-spinel and titanomagnetite and are \0.5 mm

in diameter. Chromium spinel occurs as a groundmass

phase, while titanomagnetite occurs as groundmass and as

inclusions in all phenocrysts. Titanomagnetite has 69.0–

79.3 wt% FeOT and 18.8–20.1 wt% TiO2 (Table 4).

Andesite of Glacier Creek

The Glacier Creek unit is strongly porphyritic with 47–

61% phenocrysts and microphenocrysts in a holocrystalline

to cryptocrystalline groundmass. Plagioclase dominates the

phenocryst and microphenocryst population at 70–73% of

all phenocrysts and occurs as sub-euhedral, equant to

elongate crystals. The Glacier Creek unit is distinctive in

that plagioclase phenocrysts tend to have a bimodal size

distribution. Approximately 25% of crystals are 1.2–

4.2 mm in diameter, while the remaining population is 0.1–

0.8 mm in diameter. Olivine, the next most abundant

phenocrysts phase (11–12% of the phenocryst population),

occurs as sub- to euhedral, granular or hopper-shaped

crystals that are 0.5–3.0 mm in diameter. Orthopyroxene

(7–10% of the phenocryst population) occurs as 0.1–

2.0 mm sub-euhedral, rounded prismatic to elongate crys-

tals. Clinopyroxene (7–8% of the phenocrysts population)

shares the same textural attributes as orthopyroxene. The

groundmass includes pilotaxitic and sometimes trachytic

plagioclase laths with minor (\1%) orthopyroxene and

glass. Fe–Ti oxides account for 1–2% of the Glacier Creek

unit and occur as phenocrysts, groundmass, or as inclusions

in phenocrysts. Occasional apatite is observed in the

groundmass (\0.5% of the crystal population).

Olivine exists as a sparsely distributed phenocryst phase

in the Glacier Creek unit, often with inclusions of Fe–Ti

oxides and wormy fracture patterns that are filled with

iddingsite. Nearly all olivine phenocrysts display weak to

strong chemical disequilibrium textures that are dominated

by rounded and embayed grain boundaries (Fig. 4) and

mantles of plagioclase and orthopyroxene micro-

phenocrysts. The majority of olivine grains are normally

zoned; however, some display reverse zoning. Olivine core

compositions are Fo86–68 and rims Fo79–68.

Table 4 Temperatures from representative Fe–Ti and pyroxene pair compositions

Oxide pair Ulvospinel Ferrian Ilmenite Results

TiO2 Fe2O3 FeO Al2O3 MnO MgO TiO2 Fe2O3 FeO Al2O3 MnO MgO Temp. (�C) fO2

Glacier Creek Glacier Creek

1 GC-3TB MA 16.38 34.83 42.49 1.88 0.38 2.16 GC-3TB IA 49.33 6.64 37.50 0.09 0.53 3.60 827.25 -14.14

2 GC-6TB MA 8.22 49.36 37.00 0.97 0.40 0.42 GC-6TB IA 45.56 10.46 38.16 0.07 0.84 1.14 785.45 -13.88

3 GC-6TB MB 5.95 52.08 34.93 0.72 0.24 0.31 GC-6TB IB 45.31 7.93 38.84 0.07 0.70 1.25 744.75 -14.82

4 GC-12TB MB 4.68 55.71 33.26 0.68 0.27 0.55 GC-12TB IB 44.93 10.97 36.17 0.08 0.79 1.98 749.56 -14.15

Boulder Glacier Boulder Glacier

5 BG-3TB MA 14.07 39.81 41.49 1.89 0.45 1.59 BG-3TB IA 43.59 15.49 34.74 0.20 0.53 2.23 923.59 -11.19

6 BG-3TB MB 12.24 42.82 39.45 1.61 0.48 1.58 BG-3TB IB 44.08 16.74 34.74 0.22 0.50 2.49 899.85 -11.41

7 BG-7TB MA 13.49 39.89 40.75 1.70 0.45 1.34 BG-7TB IA 46.19 11.98 36.39 0.12 0.74 2.54 881.53 -12.15

8 BG-7TB MB 14.42 39.17 41.62 1.63 0.58 1.54 BG-7TB IB 45.87 11.93 35.85 0.13 0.62 2.74 894.18 -11.96

Pyroxene pair Sample Cpx Opx Results

En Wo En Wo Temp. (�C)

Sulphur Creek

1 SC-3TB 44 41 70 3 1027

2 SC-6TB 42 41 65 3 999

Glacier Creek

1 GC-6TB 43 43 65 4 1013

2 GC-12TB 41 40 68 3 1015

Boulder Glacier

1 BG-3TB 40 43 64 3 994

2 BG-7TB 41 42 62 3 977

Temperatures and fO2 conditions are calculated according to the techniques of Andersen and Lindsley (1985) using the QUILF software package

(Andersen et al. 1993) and the ILMAT spreadsheet of Lepage (2003)

82 Contrib Mineral Petrol (2011) 161:75–99

123

Glacier Creek olivine phenocrysts have a bimodal size

distribution. The larger ([1.0 mm) population represents

50% of all olivine and has more primitive core composi-

tions (three at Fo*86) and more pronounced disequilibrium

textures compared to the small variety. The smaller

(\1.0 mm) olivine has core compositions typically below

Fo78, but are within the range of Fo68–82. The olivine

phenocrysts in the Glacier Creek unit can be distinguished

from olivine phenocrysts in the other units by elevated

concentrations of NiO, with an average of 0.26 wt%,

compared to 0.12 and 0.10 wt% in the Sulphur Creek and

Boulder glacier units, respectively.

Orthopyroxene and clinopyroxene phenocrysts in the

Glacier Peak Andesite have inclusions of Fe–Ti oxides

and display weak to moderate disequilibrium textures such

as rounded and embayed grain boundaries (Fig. 4).

Orthopyroxene is Wo3–4En56–71Fs26–40 with Mg# 64–77.

Clinopyroxene is Wo40–43En41–43Fs15–19 with Mg# 68–74.

Compositions are similar for phenocrysts and groundmass.

Plagioclase phenocrysts are dominated by disequili-

brium textures such as embayed or resorbed rims and sieve

textures. These textures are more pervasive in larger grains.

Sieve patterns are particularly complex with some grains

containing sieved and corroded cores and intact rims, while

others have sieved outer rims with pristine cores (Fig. 4).

Nearly all plagioclase phenocrysts display simple patchy

normal concentric zoning patterns, while some samples

show complex oscillatory or reverse zoning patterns. Pla-

gioclase phenocrysts have core compositions An45–75 and

rim compositions An45–61. Normally zoned plagioclase

Fig. 4 Common reaction textures in plagioclase, olivine and pyrox-

ene phenocrysts in each unit. a Photomicrograph of Sulphur Creek

plagioclase phenocryst with chemical zoning patterns and sieved core

texture with a slightly embayed rim. b Photomicrograph of Glacier

Creek plagioclase phenocryst with a sieved rim. c Photomicrograph of

Boulder Glacier plagioclase phenocryst with a resorbed and sieved

rim with inclusions of Fe–Ti oxide minerals in the core. d Backscatter

electron (BSE) image of chemically zoned and resorbed Sulphur

Creek olivine phenocryst with Fe–Ti oxide inclusions. e BSE image

of chemically zoned and partially resorbed Glacier Creek olivine

phenocrysts. f BSE image of resorbed Boulder Glacier olivine

phenocrysts. g Photomicrograph of Sulphur Creek clinopyroxene

phenocryst with a resorbed rim and inclusions of Fe–Ti oxides. h BSE

image of Glacier Creek orthopyroxene phenocryst with a reaction rim.

i Photomicrograph of resorbed Boulder Glacier clinopyroxene

phenocryst

Contrib Mineral Petrol (2011) 161:75–99 83

123

phenocrysts decrease by an average of An9 from core to

rim. Plagioclase groundmass compositions are slightly

more sodic than phenocryst compositions and are An43–51.

Fe–Ti oxide minerals in the Glacier Creek lavas occur as

isolated grains of titanomagnetite or ilmenite, as coexisting

pairs, or as inclusions in all phenocryst phases. Titano-

magnetite has 73.1–83.4 wt% FeOT and 4.7–16.4 wt%

TiO2. Ilmenite has 43.5–53.3 wt% FeO and 36.9–49.3

TiO2.

Andesite of Boulder Glacier

The Boulder Glacier lavas are strongly porphyritic with

50–60% phenocrysts. Plagioclase is dominant, comprising

65–90% of the phenocryst population, characterized as

either microphenocrysts (0.2–1.0 mm) or phenocrysts (1.0–

3.5 mm) with sub- to euhedral, equant to elongate shapes.

Orthopyroxene is the second most abundant phenocryst

type (7–30%) and occurs as 0.1–2.0 mm equant, prismatic

and elongate, sub- to euhedral crystals. Clinopyroxene (1–

10% of the phenocryst population) occurs as prismatic and,

to a lesser degree, elongate, sub- to euhedral crystals that

are 0.1–1.5 mm in length. Olivine is a minor phase (\1%)

in six of the eleven Boulder Glacier samples. When pres-

ent, it is characterized by 0.5–1.5 mm hopper or granular-

shaped phenocrysts. Apatite microphenocrysts were

observed in some samples in trace amounts (\1%). The

groundmass contains 95% intragranular pilotaxitic plagio-

clase with *5% orthopyroxene. Fe–Ti oxide minerals

occur in the groundmass and as phenocrysts, accounting for

2% of the crystals in the Boulder Glacier lavas.

Olivine was analyzed in only one sample (BG-3TB) in

the Boulder Glacier lavas. The olivine phenocryst has a

core composition of Fo74 and a rim of Fo72. Compared to

smaller grains, larger grains display strong reaction tex-

tures (Fig. 4). Olivine phenocrysts are often mantled by

plagioclase and orthopyroxene, contain inclusions of Fe–Ti

oxide minerals and have a wormy fractured texture with

iddingsite occupying the fractures.

Orthopyroxene is Wo3En60–63Fs33–37 with Mg# 62–66.

Clinopyroxene is Wo42–43En40–41Fs33–37, with Mg# 70–73.

Most pyroxene phenocrysts contain inclusions of Fe–Ti

oxide minerals and plagioclase, while elongate crystals are

sometimes ophitic with plagioclase. Forty percent of the

pyroxene phenocrysts display moderate disequilibrium

textures such as embayed rims and rounded corners

(Fig. 4).

Plagioclase phenocrysts are dominated by chemical

disequilibrium textures such as sieve textures and/or

embayed rims (Fig. 4). Sieve patterns are often observed in

or near the core of crystals and are bounded by intact rims.

In rare cases, plagioclase grains are skeletal and have been

replaced with plagioclase microphenocrysts, opaque

minerals and minor pyroxene. Inclusions of Fe–Ti oxide

minerals occur in most crystals, but are more abundant in

larger phenocrysts. Plagioclase core compositions are

An52–62, while rim compositions are An42–52. Most (but not

all) phenocrysts display normal chemical zoning patterns.

Normally zoned plagioclase phenocrysts decrease by an

average of An9 from core to rim. One groundmass pla-

gioclase grain in sample BG-10TB is An32, which is sig-

nificantly more sodic than any phenocrysts.

Titanomagnetite and ilmenite occur in the Boulder

Glacier lavas as coexisting pairs or as isolated crystals in

the groundmass or as inclusions in plagioclase, pyroxene

and olivine phenocrysts. Titanomagnetite grains have

compositions of 75.2–78.0 wt% FeO* and 12.2–14.1 wt%

TiO2. Ilmenite has compositions of 45.1–49.8 wt% FeO

and 43.0–46.8 wt%.

Crystal clots

Cumulophyric crystal clots up to 15 mm in diameter are a

common feature noted in the units studied in this paper and

are a topic of discussion of several other authors who

studied the petrography of the Mount Baker lavas (Coombs

1939; Stavert 1971; McKeever 1977; Swan 1980; Green

1988 and Hildreth et al. 2003). Cumulophyric crystal clots

seem to be common in Cascade lavas and have been noted

by several other authors at other locations such as Glacier

Peak (Taylor 2001).

Two types of crystal clots with distinct phase compo-

nents and textural characteristics are identified in this

paper. The type of clots commonly found in the Sulphur

Creek lavas are characterized by clusters of plagio-

clase ? orthopyroxene ? clinopyroxene ± olivine that are

\1–15 mm in diameter (Type 1). In contrast, Glacier

Creek lavas have randomly shaped crystal clots with cores

of Fe–Ti oxides and rims of granular orthopyroxene ?

plagioclase microphenocrysts (Type 2). Type 2 crystal

clots occur in half of all Glacier Creek samples observed.

The Boulder Glacier lavas contain both populations of

crystal clots. Type 1 is less abundant than in the Sulphur

Creek lavas. Type 2 is observed in nearly every sample, but

is characteristically oval shaped.

Major and trace element compositions

Representative whole rock major, minor and trace elements

analyzed in this study are presented in Table 5. All samples

in this study are medium-K, calc-alkaline basaltic andesites

to dacites based on the criteria of Gill (1981) (Fig. 5). SiO2

concentrations are in the range of 52.5–64.2 wt% (Fig. 6).

The Sulphur Creek unit and the Boulder Glacier unit dis-

play near-linear trends of decreasing TiO2, FeOT, MgO,

84 Contrib Mineral Petrol (2011) 161:75–99

123

Ta

ble

5R

epre

sen

tati

ve

wh

ole

rock

com

po

siti

on

s

Sam

ple

Su

lph

ur

Cre

eku

nit

Gla

cier

Cre

eku

nit

Bo

uld

erG

laci

eru

nit

SC

-3T

BS

C-7

TB

SC

-8T

BS

C-9

TB

QB

SC

-3Q

BS

C-4

GC

-3T

BG

C-5

TB

GC

-6T

BG

C-7

TB

GC

-12

TB

BG

-2T

BB

G-3

TB

BG

-4T

BB

G-7

TB

BG

-10

TB

UT

ME

58

93

69

59

30

98

58

73

74

59

49

69

59

57

57

59

72

44

58

37

14

58

34

14

58

24

97

58

33

13

58

09

88

59

00

68

59

02

16

58

84

67

58

92

06

59

13

04

UT

MN

53

92

02

95

39

20

47

53

94

63

25

39

03

21

53

89

93

25

39

12

90

54

05

07

95

40

55

59

54

05

68

95

40

52

41

54

08

14

35

40

23

97

54

02

43

05

40

30

37

54

02

90

65

40

15

90

Ma

jor

elem

ents

(wt%

)

SiO

25

5.7

85

5.7

05

5.2

15

5.7

75

2.6

05

2.5

25

8.7

05

8.7

45

8.3

45

8.5

15

8.7

26

1.8

06

4.2

26

0.1

86

3.6

36

1.6

6

TiO

21

.27

1.3

41

.38

1.2

61

.65

1.6

70

.87

0.8

50

.85

0.8

40

.83

0.7

40

.65

0.8

50

.65

0.7

5

Al 2

O3

17

.15

17

.11

17

.43

17

.37

19

.04

18

.57

17

.24

17

.56

17

.25

17

.27

17

.55

17

.73

16

.44

17

.73

16

.26

17

.65

FeO

*7

.35

7.5

17

.66

7.1

08

.79

8.8

45

.69

5.9

35

.61

5.7

45

.56

5.0

94

.27

5.0

84

.32

5.1

6

Mn

O0

.14

0.1

50

.15

0.1

40

.16

0.1

70

.11

0.1

10

.10

0.1

10

.10

0.1

00

.09

0.0

90

.09

0.1

1

Mg

O4

.70

4.9

24

.93

4.4

85

.48

5.5

44

.68

4.8

94

.56

4.8

84

.75

2.6

82

.07

3.2

62

.07

2.7

6

CaO

7.2

97

.36

7.5

27

.23

8.4

58

.54

7.0

97

.20

7.0

27

.10

7.1

55

.30

4.4

26

.10

4.3

75

.35

Na 2

O4

.21

4.3

04

.38

4.3

24

.47

4.3

83

.85

3.8

03

.86

3.7

93

.77

4.5

04

.33

3.8

84

.29

4.4

8

K2O

1.3

81

.37

1.2

91

.40

0.8

50

.78

1.5

81

.53

1.5

81

.52

1.5

01

.98

2.4

81

.78

2.4

71

.96

P2O

50

.36

0.3

90

.40

0.3

70

.44

0.4

40

.25

0.2

40

.24

0.2

40

.24

0.2

30

.19

0.2

40

.19

0.2

3

To

tal

99

.64

10

0.1

41

00

.35

99

.44

10

1.9

41

01

.45

10

0.0

41

00

.86

99

.42

10

0.0

01

00

.18

10

0.1

59

9.1

79

9.1

99

8.3

41

00

.11

Mg

#5

7.3

57

.85

7.5

57

.05

6.6

56

.86

3.3

63

.46

3.0

64

.16

4.2

52

.55

0.5

57

.35

0.1

52

.9

XR

Fa

na

lyze

dm

ino

rel

emen

ts(p

pm

)

Ni

39

43

41

46

44

43

70

75

67

77

83

17

12

29

13

23

Cr

94

10

41

00

10

71

07

10

09

19

98

89

91

00

13

11

30

13

45

Sc

25

24

24

25

26

27

19

18

18

19

19

13

11

16

15

20

V1

81

18

41

79

18

12

08

20

81

43

14

11

44

14

61

46

11

48

81

29

11

11

47

Rb

22

22

23

19

88

25

22

24

23

21

34

50

30

43

33

Sr

55

85

39

56

05

43

53

85

50

85

18

53

84

18

41

86

55

46

46

87

85

50

96

02

Zr

18

21

90

18

91

95

19

11

94

14

61

39

14

81

40

13

71

60

20

11

60

18

01

72

Ba

39

93

90

41

03

75

27

32

79

44

54

28

46

34

32

42

15

92

66

15

59

59

75

48

Nb

88

88

77

55

55

58

96

87

Y2

72

82

82

93

13

11

91

81

91

81

82

02

32

22

32

5

Cu

37

36

35

34

41

43

33

22

33

44

22

27

24

30

24

31

Zn

82

11

29

08

38

78

77

17

06

87

46

66

96

67

06

87

3

ICP

-MS

an

aly

zed

tra

ceel

emen

ts(p

pm

)

La

17

.71

7.7

18

.21

7.7

16

.11

6.2

16

.61

6.0

16

.91

5.7

15

.81

7.8

21

.02

1.0

19

.51

9.0

Ce

39

.64

0.2

41

.54

0.7

38

.73

8.8

36

.53

5.0

37

.13

4.6

34

.73

7.4

43

.04

5.2

39

.64

0.4

Pr

5.2

45

.36

5.4

35

.46

5.3

05

.36

4.7

64

.56

4.8

44

.57

4.5

54

.64

5.3

25

.93

5.0

35

.43

Nd

22

.12

2.6

22

.82

3.3

23

.22

3.2

19

.51

8.8

19

.91

8.7

18

.71

8.3

20

.62

3.8

19

.82

2.2

Sm

5.1

05

.35

5.2

65

.52

5.7

15

.74

4.2

23

.99

4.2

64

.04

4.0

04

.00

4.4

45

.07

4.3

24

.96

Eu

1.5

61

.62

1.6

01

.76

1.8

91

.93

1.2

71

.21

1.2

71

.22

1.2

11

.14

1.0

91

.38

1.1

51

.39

Contrib Mineral Petrol (2011) 161:75–99 85

123

Ta

ble

5co

nti

nu

ed

Sam

ple

Su

lph

ur

Cre

eku

nit

Gla

cier

Cre

eku

nit

Bo

uld

erG

laci

eru

nit

SC

-3T

BS

C-7

TB

SC

-8T

BS

C-9

TB

QB

SC

-3Q

BS

C-4

GC

-3T

BG

C-5

TB

GC

-6T

BG

C-7

TB

GC

-12

TB

BG

-2T

BB

G-3

TB

BG

-4T

BB

G-7

TB

BG

-10

TB

UT

ME

58

93

69

59

30

98

58

73

74

59

49

69

59

57

57

59

72

44

58

37

14

58

34

14

58

24

97

58

33

13

58

09

88

59

00

68

59

02

16

58

84

67

58

92

06

59

13

04

UT

MN

53

92

02

95

39

20

47

53

94

63

25

39

03

21

53

89

93

25

39

12

90

54

05

07

95

40

55

59

54

05

68

95

40

52

41

54

08

14

35

40

23

97

54

02

43

05

40

30

37

54

02

90

65

40

15

90

Gd

5.1

15

.24

5.2

15

.37

5.8

55

.91

3.8

73

.74

3.8

73

.72

3.6

13

.77

4.1

24

.57

4.1

44

.82

Tb

0.8

60

.89

0.8

70

.93

0.9

81

.00

0.6

10

.58

0.6

20

.59

0.5

70

.62

0.6

80

.69

0.6

80

.78

Dy

5.2

85

.38

5.3

45

.57

6.0

26

.07

3.6

33

.45

3.5

63

.49

3.4

33

.86

4.1

74

.11

4.1

94

.73

Ho

1.0

81

.10

1.1

11

.15

1.2

21

.23

0.7

20

.70

0.7

20

.71

0.6

90

.78

0.8

70

.84

0.8

60

.96

Er

2.9

43

.06

2.9

33

.14

3.3

43

.41

1.9

31

.91

2.0

01

.97

1.8

52

.15

2.4

22

.27

2.4

32

.65

Tm

0.4

20

.44

0.4

40

.46

0.4

80

.49

0.2

80

.27

0.2

90

.28

0.2

70

.32

0.3

60

.32

0.3

60

.39

Yb

2.6

72

.75

2.7

32

.91

3.0

43

.02

1.7

51

.72

1.7

81

.71

1.6

92

.11

2.3

62

.02

2.2

82

.44

Lu

0.4

30

.44

0.4

40

.46

0.4

70

.48

0.2

90

.27

0.2

90

.27

0.2

70

.33

0.3

80

.32

0.3

70

.39

Th

3.1

63

.03

3.1

82

.70

1.2

61

.25

3.3

93

.05

3.3

73

.15

3.0

03

.88

6.3

64

.02

5.3

64

.00

Hf

4.3

64

.40

4.4

74

.54

4.2

94

.30

3.8

43

.70

3.9

23

.73

3.6

74

.14

5.2

64

.09

4.7

04

.37

Ta

0.5

20

.51

0.5

20

.53

0.4

80

.49

0.3

60

.35

0.3

60

.35

0.3

30

.53

0.6

50

.42

0.5

90

.51

U1

.25

1.2

01

.25

1.0

40

.53

0.5

21

.24

1.1

51

.24

1.1

81

.10

1.6

02

.53

1.4

72

.10

1.6

0

Pb

5.4

15

.10

6.1

74

.74

3.3

22

.80

5.1

99

.93

5.9

95

.76

5.6

36

.06

8.4

04

.84

7.9

26

.62

Cs

0.6

50

.63

0.6

60

.50

0.2

10

.20

0.5

40

.30

0.2

90

.42

0.3

00

.46

1.3

60

.36

1.1

60

.91

An

aly

ses

per

form

edat

the

WS

UG

eoan

aly

tica

lL

abo

rato

ry.

Sam

ple

loca

tio

ns

are

rep

ort

edin

Un

iver

sal

Tra

nsv

erse

Mer

cato

rco

ord

inat

esy

stem

(UT

M)

NA

D8

3.

Acc

ura

cyan

dp

reci

sio

n

esti

mat

esfo

rX

RF

anal

yse

sar

eav

aila

ble

inJo

hn

son

etal

.(1

99

9).

Mg

#=

10

0M

g/(

Mg

?F

e2?

)w

ith

Fe2

?ca

lcu

late

das

0.8

5F

eOT

86 Contrib Mineral Petrol (2011) 161:75–99

123

Na2O, Al2O3 and CaO and increasing K2O with increasing

SiO2. Glacier Creek andesites are tightly clustered on

variation diagrams and are characteristically enriched in

MgO (*5 wt% at 58.5 wt% SiO2) relative to the other

units. Glacier Creek Mg#s (100 9 Mg/(Mg ? Fe2?)) are

comparatively high with values tightly clustered at Mg#

63–64 (Fig. 6). Sulphur Creek samples have a narrow

range of Mg#s at 57–58. The Mg#s of Boulder Glacier

andesites and dacites are relatively low and have the widest

range of values at 50–57.

Trace element patterns on primitive mantle normaliza-

tion diagrams show enrichment in large ion lithophile

elements (LILE) such as Rb, Ba, U, K and Sr compared to

the high field strength elements (HFSE) such as Nb, Ta, Ti

(Fig. 7). Lack of correlation between these elements and

SiO2 (Figs. 7, 8) suggests that these units are not related by

any simple fractionation or mixing process.

Rare earth elements (REE) are plotted in Fig. 9. The most

Si-poor lavas (Sulphur Creek basaltic andesites) have the

highest REE abundances, and the most Si-rich lavas (Boul-

der Glacier andesites and dacites) have intermediate REE

abundances. The intermediate Si lavas (Glacier Creek) have

the lowest REE abundances. Clearly, there is no simple

genetic relationship between these three lava units.

The following sections describe the whole rock chem-

istry for each unit individually and the contrasts between

them.

basalt of Sulphur Creek

Sulphur Creek magmas are distinguished from the more Si-

rich suites of Boulder Glacier and Glacier Creek by lower

concentrations of LILE such as Rb, Ba and K and higher

concentrations of HFSE such as P, Sm, Eu, Tb, Yb and Ti

(Fig. 7). They are also enriched in Sc, V, Cr, Y and Tb for

a given SiO2 concentration relative to the other units

(Fig. 8, Sc, V and Tb are not shown). Trace elements

correlate with SiO2 and generally form sub-linear trends

(Figs. 8, 9, 10). Note that Yb decreases with increasing

SiO2 (Fig. 9).

REE patterns for the Sulphur Creek unit are distin-

guished from the other units by their higher concentrations

0

1

2

3

50 55 60 65Wt. % SiO2

K2O

Glacier Creek

Sulphur Creek

Boulder Glacier

Basaltic Andesite Andesite

Med. K

High K

Fig. 5 K2O vs. SiO2 classification diagram after Gill (1981)

0.0

0.5

1.0

1.5

2.0

50 55 60 65

TiO

2

Boulder Glacier

Sulphur Creek

Boulder Glacier

14

16

18

20

50 55 60 65

lA

2 O3

1

2

3

4

5

6

50 55 60 65

Mg

O

2

4

6

8

10

50 55 60 65

O a

C

2

3

4

5

50 55 60 65

Wt.% SiO2

Na 2

O

40

50

60

70

50 55 60 65

Wt. % SiO2

#g

M

Fig. 6 Whole rock major

element compositions for the

Sulphur Creek (SC), Glacier

Creek (GC) and Boulder Glacier

(BG) units. All major elements

are plotted against SiO2 on the

x-axis. Mg# = Mg/

(Mg ? Fe2?) with Fe2?

calculated as 0.85 FeOT

Contrib Mineral Petrol (2011) 161:75–99 87

123

Trace Element Concentrations

1

10

100

Rb Ba Th U Nb Ta K La Ce Sr P Nd Sm Zr Hf Eu Ti Tb Ho Yb Lu

Sam

ple

/Pri

mit

ive

Man

tle

100

300

500

700

50 55 60 65 70

SiO2 Wt% 50 55 60 65 70

SiO2 Wt%

Ba

pp

m

300

500

700

900

mpp r

S

Glacier Creek Andesite

Sulphur Creek Basaltic Andesite

Boulder Glacier Andesite and Dacite

Fig. 7 Primitive mantle

normalized trace element

diagram. Values are normalized

to primitive mantle

concentrations of Sun and

McDonough (1989). Select

trace elements are plotted

against SiO2 to show how trace

element patterns change with

differentiation. Symbols are the

same as in Fig. 6

10

20

30

40

50 55 60 65 70

Y p

pm

0

20

40

60

80

100

120

50 55 60 65 70

SiO2 Wt%

Cr p

pm

0

20

40

60

80

100

50 55 60 65 70

SiO2 Wt%

Ni p

pm

0

10

20

30

40

50

60

50 55 60 65 70

Rb

pp

m

100

125

150

175

200

225

50 55 60 65 70

Zr

pp

m

4

5

6

7

8

9

50 55 60 65 70

Nb

pp

m

Fig. 8 Select trace elements

plotted against SiO2 on the

x-axis. Symbols are the same

as in Fig. 6

88 Contrib Mineral Petrol (2011) 161:75–99

123

of the middle REEs (Eu–Ho) and heavy REEs (Er, Yb and

Lu) (Fig. 9). The Sulphur Creek unit has the flattest REE

slopes, with average (La/Yb)N = 4.7, whereas Glacier

Creek and Boulder Glacier units have average (La/

Yb)N = 6.7 and 6.3, respectively. Within-flow variations

are characterized by slightly increasing light rare earth

element (LREE) abundances and decreasing heavy rare

earth element (HREE) abundances with increasing SiO2

(Fig. 9). Eu anomalies in the Sulphur Creek unit are either

very weakly negative or non-existent.

A compositional gap within the Sulphur Creek unit

exists between a more mafic subgroup (52.5–53.5 wt%

SiO2, *5.5 wt% MgO) and a more felsic subgroup (55.2–

55.8 wt% SiO2, 4.5–4.9 wt% MgO) (Figs. 5, 6, 7, 8, 9).

These two subgroups are separated spatially, with the more

mafic samples located on the most distal portions of the

lava flow on the west side of Baker Lake (Fig. 2). The

same compositional gap and spatial relationship is

observed in whole rock data and sample locations pre-

sented by Hildreth et al. (2003). Trace element concen-

trations also distinguish the two groups. The mafic group

has lower abundances of LILE such as Rb, Ba, Th, K and

higher abundances of Ti, Nb, Y and HREE compared to the

more felsic group (Figs. 7, 8, 9). Samples with composi-

tions in between the subgroups have not been observed.

The petrologic relationship between these two groups will

be discussed in a later section of this paper.

andesite of Glacier Creek

The andesites of Glacier Creek are intermediate in SiO2,

but are the most enriched in MgO (Mg# 63–65,

MgO = 4.5–5.0 wt%), Ni (67–82 ppm), Cr (89–100 ppm)

and Sr (841–866 ppm) compared to the other units. They

REE Concentrations

10

100

La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Yb Lu

Sam

ple

/Ch

on

dri

te

15

18

20

23

50 55 60 65 70

SiO2 Wt%

mpp aL

1

2

3

4

50 55 60 65 70SiO2 Wt%

mpp bY

Sulphur Creek Basaltic Andesite

Boulder Glacier Andesite and Dacite

Glacier Creek Andesite

Fig. 9 Chondrite normalized

rare earth element diagram.

Values are normalized to

chondrite concentrations of Sun

and McDonough (1989). Select

REEs are plotted against SiO2 to

show how REE patterns change

with differentiation

-16

-15

-14

-13

-12

-11

-10650 750 850 950 1050

T(C)

gol

fO(

2)

Glacier CreekBoulder Glacier

QFMNNO

Fig. 10 Equilibration temperatures and fO2 conditions estimated

from Fe–Ti oxide pairs in the Glacier Creek (GC) and Boulder

Glacier (BG) units. Diagram and location of buffer curves are redrawn

from Eugster and Wones (1962). QFM—Quartz–Faylite–Magnetite

buffer curve, NNO—Nickel–Nickel Oxide buffer curve

Contrib Mineral Petrol (2011) 161:75–99 89

123

also have the steepest REE patterns (La/Yb)N = 6.7, but

the lowest overall REE abundances (Fig. 10). They have

the lowest abundances in most HFSE (Nb, Ta, Zr, Hf)

(Fig. 7) and the highest Sr/Y ratios (40–50). A defining

characteristic of the Glacier Creek samples is their small

range of trace element abundances, excluding Ni, Cu and

Ga. Thus, there is no correlation of trace elements with

SiO2 (Fig. 8).

andesite of Boulder Glacier

The Boulder Glacier andesites and dacites have the highest

abundances of most LILE (Rb, Ba, Th, U, K, La, Ce), the

lowest abundances of some HFSE (P and Ti) and inter-

mediate abundances of HREE compared to the Sulphur

Creek and Glacier Creek units (Fig. 8). This unit is also

distinctly depleted in Ni and Cr (Fig. 8). Some trace ele-

ments correlate well with wt% SiO2 (Figs. 7, 8).

The Boulder Glacier unit displays the widest range of

REE concentrations of the three units, with the highest

LREE abundances (Fig. 9). The HREE form concave up

patterns with abundances intermediate between the other

units. The Boulder Glacier unit is also characterized by

having the most pronounced negative Eu anomaly among

the three units. There is no systematic variation of LREE

with SiO2 (Fig. 9); however, there is a weak positive cor-

relation of intermediate and heavy REE with increasing

SiO2 (Figs. 9). This is in contrast to the negative correla-

tion of HREE with SiO2 in the Sulphur Creek unit.

Discussion

Intensive parameters

The Fe–Ti oxide geothermometer of Andersen and Lindsley

(1985) and Andersen et al. (1993) was used to constrain pre-

eruptive temperatures and assess fO2 conditions for the

Mount Baker magmatic system. Coexisting titanomagnetite

and ferrian ilmenite were found in both the Glacier Creek

lavas and the Boulder Glacier lavas; however, only titano-

magnetite was observed in the lavas of Sulphur Creek. Prior

to calculating temperatures and fO2 conditions, Fe–Ti oxide

pairs (some touching, some not) were tested for equilibrium

following the methods of Bacon and Hirschmann (1988), and

only oxide pairs determined to be in or near equilibrium were

used for the calculations. ILMAT (Lepage 2003) was used to

carry out the equilibrium calculations.

Coexisting oxide pair chemistry along with the calcu-

lated temperatures and fO2 conditions are displayed in

Table 7. Calculated fO2 conditions are near the nickel-

nickel-oxide buffer (Table 4; Fig. 10). Errors using Fe–Ti

oxide pairs for fO2 calculations are reported to be ±0.1 log

units (Andersen et al. 1993). Calculated pre-eruptive tem-

peratures for the Glacier Creek unit are 745–827�C, while

temperatures calculated for the Boulder Glacier unit are

882–924�C (Table 4). The temperature calculations for the

Glacier Creek unit are anomalously low and are not con-

sidered to be representative of pre-eruptive temperatures.

Estimation of temperatures for coexisting pyroxene

pairs was applied using the QUILF software package of

Andersen et al. (1993). A total of six temperatures are

presented in Table 4 with two for each unit discussed in

this paper. Equilibrium condition of the pyroxene pairs was

taken into consideration during the selection of pyroxene

analyses and was based on the difference in Mg# (DMg#)

between each component of the coexisting pairs. Pyroxene

pairs generally have a DMg# \3; however, due to limited

pyroxene data for sample BG-3TB, a DMg# of five was

used, which may represent disequilibrium conditions.

Since pressure estimates in the Mount Baker magmatic

system are unavailable, pyroxene equilibration tempera-

tures reported here are calculated at an estimated pressure

of 3 kbar. The effect of deviating ±1.5 kbar from the

estimated 3 kbar was tested and found to produce \±6�C

difference in the calculated temperatures. The Sulphur

Creek basaltic andesites yield the widest range of pyroxene

temperatures and the hottest temperature, with values of

999�C for SC-6TB and 1,027�C for SC-3TB (Table 4). The

andesites of Glacier Creek yielded a tight range of inter-

mediate temperatures of 1,013�C for sample GC-6TB and

1,015�C for sample GC-12TB (Table 4). The more silicic

lavas of Boulder Glacier yielded lower temperatures of

994�C for sample BG-3TB and 983�C for sample BG-7TB

(Table 4). These are reasonable pre-eruptive equilibration

temperatures for these compositions.

Petrologic relationships between units

As discussed in the following paragraphs, major and trace

element variations between the Sulphur Creek basalt and

the andesites of Glacier Creek and Boulder Glacier cannot

be related by crystal fractionation. This conclusion is based

on the following: (1) the presence of two chemically dis-

tinct andesite types, (2) failed major and trace element

fractionation modeling, and (3) mineral textures and

chemical relationships that are inconsistent with fraction-

ation processes.

Samples with intermediate compositions (57–63 wt%

SiO2) can be separated into two andesite types. Type 1

includes the andesite of Glacier Creek, which is charac-

terized by elevated MgO and Ni, very low variability in

major and trace element concentrations, and low REE

concentrations with elevated La/Yb (Figs. 6, 7, 8, 9). Type

2 includes the andesites and dacites of Boulder Glacier,

which are characterized by comparatively low MgO and

90 Contrib Mineral Petrol (2011) 161:75–99

123

Ni, and sub-linear to scattered trends on element variation

diagrams (Figs. 6, 7, 8, 9). These two separate andesite

types are distinct enough that they could not be produced

by closed-system crystal fractionation from the same mafic

parent.

MELTS, a thermodynamic model of phase equilibria

(Ghiorso et al. 1993), was used to determine whether the

basalts of Sulphur Creek could be parental to one or the

other of the andesitic suites. Fractionation of the most

mafic Sulphur Creek basalt (bsc 442 from Hildreth et al.

2003) consistently produced intermediate major element

compositions that have higher concentrations of Na2O, P2O

and MnO and lower concentrations of MgO and CaO than

the observed Glacier Creek and Boulder Glacier andesite

compositions.

Rayleigh fractionation (Arth 1976) of observed mineral

phases (PLAG ? OPX ? OLV ? OXIDE ± CPX) from

the Sulphur Creek basalt also could not reproduce the trace

element patterns of either of the two andesites. For all

modeled mineral proportions (observed proportions of

pyroxene, olivine, plagioclase and oxide), bulk Ds for REE

are less than one. Hence, REE should increase with dif-

ferentiation of the most mafic Sulphur Creek lava. Instead,

Sulphur Creek lavas show a decrease in HREE with dif-

ferentiation (Fig. 9), and andesites from the other two units

have the same or lower LREE and much lower HREE than

the proposed basalt parent (Fig. 9). Production of Glacier

Creek andesites from Sulphur Creek basalt is even more

problematic, given the higher Ni and Mg# in that andesite

than its proposed Sulphur Creek parent.

In summary, neither the major element or trace element

characteristics permit crystal fractionation as a mechanism

to produce either of the intermediate andesite suites (Gla-

cier Creek or Boulder Glacier) from the basalt of Sulphur

Creek.

The existence of multiple types of andesite at other

Cascade volcanoes has been attributed to an assortment of

open-system processes that include mixing of crustal melts

and various mafic to intermediate magmas (Clynne 1999;

Bullen et al. 1990). Other authors have described lavas that

have been modified by assimilation of crustal material

(Mason et al. 1996) or assimilation of mafic–ultramafic

crystal debris (Streck et al. 2007; Arculus et al. 1983).

Alternatively, the Glacier Creek and Boulder Glacier

units may be the fractionated products of different parental

magmas. If so, a suitable parental magma for the relatively

high-magnesium Glacier Creek andesite must also be

enriched in MgO, Ni and Cr and depleted in REE com-

pared to the parent magma for the andesites of the Boulder

Glacier unit. The description of a compositionally hetero-

geneous mantle that is presumed to be the source of various

types of parental magmas is a common theme in the

Cascade arc (Green and Harry 1999; Bacon et al. 1997;

Leeman et al. 2005; Green and Sinha 2005). Glacier Peak,

Mount Baker’s neighboring volcano to the south, was

found to have at least three chemically distinct parental

magmas (Taylor 2001). Preliminary results by Moore and

DeBari (2009) show that Mount Baker itself has at least

three mantle-derived magma types, at least one of which

has REE characteristics similar to the Glacier Creek

andesites.

As described earlier, disequilibrium mineral textures are

a common characteristic of each unit in this study. These

include phenocrysts with bimodal grain sizes and compo-

sitions, plus reaction textures and complex zoning patterns

(Fig. 4; Table 1). A common explanation of similar tex-

tures has been attributed to magma mixing processes

(Eichelberger 1975). We discuss this possibility in more

detail in the subsequent sections.

Petrologic relationships within units

Basalt of Sulphur Creek (basalt and basaltic andesite)

The Sulphur Creek unit has previously been described as

the result of mixing between basalt and basaltic andesite

(Green 1988). This conclusion was drawn from chemical

and textural relationships manifested in basaltic inclusions

that are sparsely distributed throughout the basaltic

andesite. Interpretations of geochemical data presented in

this study support an additional mixing model for the

Sulphur Creek unit that calls upon a felsic end-member

component similar to the dacite of Boulder Glacier.

Two chemically distinct subgroups within the Sulphur

Creek unit were described earlier (e.g. Figs. 6, 7) and occur

in spatially separate parts of the flow. The more felsic

(andesitic) subgroup has lower mid- to heavy REE abun-

dances than the mafic group (Fig. 11a). However, Raleigh

fractionation of observed modes of plagioclase ? clino-

pyroxene ? orthopyroxene ? olivine in the mafic Sulphur

Creek sub-group predicts increasing abundances in all REE

(Fig. 11b) with fractionation. Thus, the model is inconsis-

tent with the data and suggests that the Sulphur Creek

subgroups are not related by pure crystal fractionation. The

compositional gap between the two subgroups also con-

tradicts a continuous crystal fractionation processes.

Cryptic hornblende fractionation has been considered to

play an important role in arc volcanoes (e.g. Davidson et al.

2007). An alternative REE model (Fig. 12c) attempts to

relate the two Sulphur Creek subgroups through crystal

fractionation that includes hornblende as a cryptic frac-

tionating phase. The model uses mafic Sulphur Creek

sample Qbsc-4 as a starting composition and uses hypo-

thetical fractionating modes adjusted from observed pro-

portions to include hornblende (60% plagioclase ? 20%

hornblende ? 10% orthopyroxene ? 10% clinopyroxene).

Contrib Mineral Petrol (2011) 161:75–99 91

123

Several published sets of Kds for hornblende (Fujimaki

et al. 1984; Sisson 1994) were tested; but intermediate Kds

of Sisson (1994) are used in Fig. 11c. The model shows

that even when hornblende is included as a fractionating

phase, the observed composition of the andesitic Sulphur

Creek subgroup cannot be reproduced.

An alternative hypothesis is magma mixing. As dis-

cussed previously, all units in this study show bimodal

phenocrysts populations; some phenocrysts within a unit

display disequilibrium and reaction textures, while some

phenocrysts lack these textures (see Sulphur Creek

petrography and mineral chemistry). At other arc volca-

noes, textures such as these have been called upon to

support magma mixing (Madeleine et al. 2006; Venezky

and Rutherford 1997). Plagioclase and olivine phenocrysts

in the Sulphur Creek unit have a wide range of composi-

tions that largely overlap with mineral compositions from

the Boulder Glacier and Glacier Creek units (Fig. 3). If

magma mixing processes are responsible for the disequi-

librium textures in these units, the overlapping mineral

compositions may be a result of varying degrees of mixing

between compositions that span the range between two

compositionally distinct end-members.

Mixing calculations for both major and trace elements

show that compositions from the mafic group of the Sul-

phur Creek unit (*5.5 wt% MgO) can be combined with

felsic compositions similar to the Boulder Glacier unit

(*2 wt% MgO) to produce andesitic compositions

comparable to the more evolved Sulphur Creek group

(4.5–5.0 wt% MgO) (Fig. 12). Mixing proportions are

dependent upon samples used in the calculations. Calcu-

lated hybrids that closely match andesitic Sulphur Creek

compositions can be produced by mixing 30–40% Boulder

Glacier dacite with 60–70% Sulphur Creek basalt.

Figure 12 shows the results of a representative mixing

calculation for a hybrid that is produced by combining 70%

Sulphur Creek basalt (Qbsc-4) with 30% Boulder Glacier

dacite (BG-3TB). The resulting hybrid compares favorably

to Sulphur Creek andesite sample SC-7TB (see Fig. 12

inset).

Hybrid compositions produced by the above mixing

calculations are depleted in MgO, Ni and Cr compared to

observed felsic Sulphur Creek samples (Fig. 12 inset).

Elevated MgO, Ni and Cr in these samples may be linked

to the addition of small amounts of olivine (see the fol-

lowing Glacier Creek andesite discussion).

The magma mixing calculations break down when

andesitic rather than dacitic samples from the Boulder

Glacier unit are used as the felsic end-member (see Zr

and Nb in Fig. 8 for a qualitative view of this). Thus, the

magma mixing process appears to rely on a specific

component within the Boulder Glacier unit. The origin of

this felsic component is not constrained in this paper.

However, crustal melting is likely, as discussed further in

the following section on the Boulder Glacier unit. Felsic

end-members in magma mixing scenarios in the Cas-

cades have commonly been attributed to partial melting

of a crustal protolith (Conrey et al. 2001; Streck et al.

2007; Smith and Leeman 1987; Venezky and Rutherford

1997).

The above-described mixing model suggests that the

9.8-ka Sulphur Creek unit is mixed with dacitic composi-

tions similar to the Boulder Glacier unit, which is dated at

80–90 ka. The significant age gap implies that the process

A

10

100

RE

E c

on

cen

trat

ion

/Ch

on

dri

te

QBSC-4

Glacier Creek Avg.

Boulder Glacier Avg.

F=0.8

F=0.6

F=0.4

F=0.2

F=0.1

Increasing Fractionation

Rayleigh FractionationC m/C p=F (D – 1)

Where F =fraction of melt

Fractionating Modes:69% Plag.20% Opx10% Olv1% Mag

Sulphur Creek Compositional Sub-Groups

10

100

La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Yb Lu

La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Yb Lu

La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Yb Lu

sam

ple

/Ch

on

dri

te

Mafic Group

Felsic Group

Mafic Group (basalt to basaltic andesite)Felsic Group (basaltic andesite to andesite)

4

5

6

50 55 60SiO2

M

gO

Mafic Group

Felsic Group

10

100

RE

E c

on

cen

trat

ion

/Ch

on

dri

te

Qbsc-4Sulphur Creek felsic group avg.F=0.8F=0.6F=0.4F=0.2F=0.1

Increasing Fractionation

Rayleigh FractionationC m/C p=F (D – 1)

Where F =fraction of melt

Fractionating Modes:50% Plag20% Hbl15% Opx15% Cpx

B

C

Fig. 11 a Sulphur Creek compositional sub-groups. The Sulphur

Creek unit is separated into a more felsic group and a more mafic

group based on a gap observed in major and trace element

compositions. b Results of a Rayleigh fractionation model with

starting composition of Qbsc-4 and fractionating assemblage of 69%

plagioclase ? 20% orthopyroxene ? 10% olivine ? 1% magnetite. cResults of a Rayleigh fractionation model with starting composition

of Qbsc-4 and a hypothetical fractionating assemblage of 50%

plagioclase ? 20% hornblende ? 15% orthopyroxene ? 14%

clinopyroxene

92 Contrib Mineral Petrol (2011) 161:75–99

123

responsible for generating the Boulder Glacier felsic end-

member is long lived and has been active or recurring at

least until the eruption of the Sulphur Creek lava at 9.8 ka.

Given that the more mafic part of the Sulphur Creek unit

is in the early ejected scoria (Hildreth et al. 2003) and in

the distal part of the lava flow, the early erupted part of the

unit is the least mixed. Continued lava effusion from the

same vent tapped a more mixed source region.

Andesite of Glacier Creek—origin of high-Mg signature

The Glacier Creek andesite has chemical attributes such as

high Mg#s (63–64), high Ni (67–82 ppm) and Cr (89–

100 ppm), elevated [La/Yb]N (*6.7), and depleted REE

compared to average arc andesite. Andesites with these

characteristics are sometimes classified as high-magnesium

andesites or primitive andesites (Grove et al. 2002; Streck

et al. 2007; Kelemen 1994), but the Glacier Creek andesites

are not quite rich enough in MgO, Ni and Cr concentrations

to be classified as such. Crystal fractionation of a high-

magnesium andesite can produce andesites with charac-

teristics such as these (c.f. Grove et al. 2005). Alterna-

tively, the addition of ultramafic or mafic crystalline

material to a typical andesite may also cause elevated

MgO, Ni and Cr concentrations in the whole rock (e.g.

Streck et al. 2007: Arculus et al. 1983).

Thus, we must determine whether the high-Mg nature of

the Glacier Creek andesite is a characteristic of the parental

liquid, or if it is a product of crystal accumulation. A

characteristic feature of the Glacier Creek andesite is that it

contains a population of high-Mg olivine phenocrysts that

have particularly strong disequilibrium textures and are

relatively large ([1 mm) when compared to the remainder

of olivine in Glacier Creek rocks. These olivines may be

xenocrystic in origin. Streck et al. (2007) identified

ultramafic xenocrysts of olivine and orthopyroxene with

strong reaction textures in Mt. Shasta high-Mg andesites.

They proposed that incorporation of these xenocrysts fun-

damentally altered the major element chemistry of a low-

Mg andesite. To distinguish between these possibilities, a

review of the chemistry and petrography of Glacier Creek

olivine is presented below.

As discussed earlier, the Glacier Creek unit has two

olivine phenocryst size populations. Larger, embayed

phenocrysts have more mafic core compositions (up to

Mg# 86) than smaller equilibrium phenocrysts (\Mg# 78)

(Fig. 13). The microprobe data are limited, but are sup-

plemented with additional Scanning Electron Microscope

(SEM) and energy dispersive spectrometer (EDS) analyses.

Figure 13a shows olivine size compared to approximate

core Mg#s from SEM analyses.

We propose that the large mafic olivines (Mg# C 82)

with strong disequilibrium textures are xenocrystic.

Potential sources for the added olivine are outlined below,

followed by a discussion quantifying the amount of added

olivine.

At Mount Baker, the nearest ultramafic source rock is

the Paleozoic Twin Sisters dunite. The dunite crops out

13 km SW of Mount Baker’s summit, comprising two

*2,000 m peaks covering some 90 km2. The body is

dominantly dunitic, but ranges from harzburgite to pure

olivine and is chromite bearing (Tabor et al. 2003). How-

ever, olivine in the dunite is more Mg-rich (Fo90–92) than

the most Mg-rich olivine in the Glacier Creek unit (Fo86)

(Onyeagocha 1978). In addition, the dunite body is bound

by SW dipping high angle thrust faults and is interpreted to

project above Mount Baker (Tabor et al. 2003). The Twin

Sisters dunite itself is thus an unlikely source; however,

components similar to the Twin Sisters dunite may lie

deeper in the crust.

10

100

La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Yb Lu

sam

ple/

Cho

ndrit

e

QBSC-4BG-3TBHybridSC-7TB

70% QBSC-4 + 30% BG-3TB= Hybrid (matched against observed SC-7TB)

0.5

1.0

1.5

SiO2

TiO2

Al2O3

FeO*

MnO MgO

CaO Na2

OK2O

P2O5

Rb Sr Ba Zr Y Cr Ni

Hyb

rid/S

C-7

T

QBSC-4BG-3TBHybridSC-7TB

Fig. 12 Magma mixing model between the mafic Sulphur Creek

compositions and Boulder Glacier dacites. A hybrid liquid composed

of 70% Qbsc-4 and 30% BG-3TB is compared to a measured felsic

Sulphur Creek sample, SC-7TB. REE results are shown on a

chondrite normalized diagram. Inset: major element and select trace

element results are evaluated by comparing calculated hybrid liquid to

measured element concentrations for SC-7TB

Contrib Mineral Petrol (2011) 161:75–99 93

123

Ophiolite bodies underlying volcanic centers (Streck

et al. 2007; Arculus et al. 1983) and mafic plutonic roots

(Dungan and Davidson 2004) have also been proposed as

sources of added mafic/ultramafic crystalline material.

Incorporation of cognate xenoliths is also possible. Mafic

or ultramafic crystalline material derived from any of

these sources could include small percentages of other

minerals phases. However, no elements other than Mg,

Cr and Ni are anomalously high in the Glacier Creek

lavas.

The amount of olivine addition in the Glacier Creek

andesite can be quantified by use of the olivine/liquid

equilibrium relationship of Roeder and Emslie (1970). The

relationship is described as:

Kd ¼ Fe=Mgð ÞOlv= Fe2þ=Mg� �

Liq¼ 0:3

where Fe2? is calculated as 0.85 FeT from fO2 derived

from oxide chemistry. This relationship is used to calculate

Mg#s of hypothetical liquids (Mg#Liq) that are in equili-

brium with known olivine compositions from microprobe

analyses (Fig. 13b; Table 6).

Figure 13b shows that equilibrium liquids calculated

from core compositions of most equilibrium (\1 mm)

olivine grains have lower Mg#s (Mg#Liq = 38–57) than

observed whole rock Mg#s (Mg#WR 63–64). Observed

non-xenocrystic olivines were in Fe/Mg equilibrium with a

more felsic liquid than the observed whole rock composi-

tion. Xenocrystic olivine must be removed to restore whole

rock compositions to an equilibrium liquid composition.

Removal of the xenocrystic olivine is done with olivine

subtraction calculations (Table 7). The amount of xeno-

crystic olivine subtracted is determined by the point where

Fe/Mg equilibrium is achieved between non-xenocrystic

olivine and the new re-calculated whole rock composition.

If the calculations are valid, then the amount of olivine

subtracted from whole rock compositions should be com-

parable to the observed mode of the xenocrystic olivine

population in whole rock samples. The observed mode

from petrographic observation (based on 1,000 point

35 40 45 50 55 60 65 70

Mg#Calculated Mg# of liquid from cores

Calculated Mg# from rims

Measured Mg# of whole rock

40 50 60 70 80 90

SEM Mg#>1mm (core)<1mm (core)

20µ 200µ

- <1mm

- weak to mod. disequilibrium textures

- >1mm

- Strong disequilibrium textures

Felsic Variety Xenocrystic Variety

40 50 60 70 80 90

SEM Mg#>1mm (core)<1mm (core)

20µ20µ 200µ200µ

- <1mm

- weak to mod. disequilibrium textures

- >1mm

- Strong disequilibrium textures

Felsic Variety Xenocrystic VarietyEquilibrium Variety Xenocrystic Variety A

B

Fig. 13 Characteristics of the bimodal olivine population in the

Glacier Creek andesite. a Grey scale images are SEM backscatter

electron images. The SEM Mg# plot shows core compositions of

individual olivine phenocrysts. Compositional data are from semi-

quantitative SEM analyses. b Comparison of Glacier Creek measured

whole Rock Mg#s (open diamonds) and calculated liquid Mg#s from

olivine cores (solid circles) and rims (open circles). Calculations use

compositions of olivine from microprobe analyses and the liquid/

olivine equilibrium constraints described by Roeder and Emslie

(1970)

Table 6 Glacier Creek olivine/liquid equilibrium calculations

Label SiO2 Fe Mg (Fe/Mg)Olv/

(Fe/Mg)WR

Mg#Liq Mg#WR

Whole rock

GC-3TB 58.70 0.07 0.12 0.24–0.45 63

GC-6TB 58.34 0.07 0.11 0.23–0.62 63

GC-12TB 58.72 0.07 0.12 0.34–0.71 64

Olivine core

GC-3TB OAc* 39.94 0.16 1.13 0.24 68

GC-6TB OAc 37.70 0.31 0.88 0.60 46

GC-6TB OBc* 39.77 0.16 1.13 0.25 68

GC-6TB OCc* 40.00 0.16 1.13 0.24 68

GC-12TB OAc 38.70 0.26 0.96 0.47 53

GC-12TB OBc 37.17 0.34 0.82 0.41 42

GC-12TB OCc 38.84 0.24 0.98 0.44 55

GC-12TB Odc 39.22 0.20 1.06 0.34 61

Olivine rim

GC-3TB OAr 38.15 0.25 0.97 0.45 54

GC-6TB OAr 37.81 0.32 0.86 0.62 45

GC-6TB OBr* 37.59 0.30 0.88 0.59 46

GC-6TB OCr* 38.79 0.23 1.00 0.23 57

GC-12TB OAr 37.68 0.30 0.89 0.60 47

GC-12TB OBr 37.25 0.33 0.83 0.71 43

GC-12TB OCr 38.83 0.25 0.98 0.46 54

GC-12TB Odr 38.54 0.25 0.97 0.45 54

Olivine liquid equilibrium calculations follow the methods of Roeder

and Emslie (1970). The relationship is described as Kd = 0.03 =

(Fe/Mg)Olv/(Fe/Mg)Liq. The equation is rearranged to calculate Mg#s

of liquids that are in equilibrium with olivine compositions. The

rearranged equation is as follows: Mg#Liq = 100/[((Fe/Mg)Olv/

0.3) ? 1]. Samples labels with an asterisk are large xenocrystic

olivine. Whole rock Mg#s are calculated as (100 9 (Mg/Mg ?

Fe2?)). Whole rock Fe2? is 0.85 FeT

94 Contrib Mineral Petrol (2011) 161:75–99

123

counts) is 2.5–3.7% xenocrystic olivine by volume in a

given Glacier Creek sample.

Subtraction of 4% xenocrystic olivine from Glacier

Creek whole rock compositions achieves the desired equi-

librium Mg#s (Fig. 14). This amount is close to observed

xenocrystic olivine modes, and the subtraction model is

considered to be successful. The model is sensitive to many

parameters, and the small discrepancy (\1% by volume)

could be a result of several possibilities: (1) errors in the

point counting of xenocrystic olivine in the Glacier Creek

samples; (2) added olivine may have been absorbed into the

host liquid post addition, making the observed amount

lower than the total amount added; (3) subtracted olivine

compositions are not representative of the added olivine

component; (4) inaccurate Fe concentrations and ultimately

Mg#s due to assumed oxygen fugacity conditions that are

misrepresentative of the magmatic system; (5) the added

mafic/ultramafic material is not pure olivine and may con-

tain other minerals.

If we accept the olivine addition model, then composi-

tions of Glacier Creek andesites after olivine subtraction

should be representative of liquid compositions prior to

olivine addition. The subtraction will reduce Glacier Creek

andesite MgO contents from *4.5 to *3 wt% and Mg#s

from *65 to *55. Calculated concentrations of FeOT are

not very different than before. Other major element com-

positions (TiO2, K2O, Na2O, P2O5 and MnO, Table 7) are

also relatively unaffected by olivine subtraction. CaO

compositions, however, are slightly elevated.

Given that olivine is poor in incompatible trace ele-

ments, the REE patterns of the re-calculated Glacier Creek

lavas will be no different than before. They are still much

more REE depleted than the other two units (Fig. 9). No

suitable parent for such a REE-depleted magma is descri-

bed in this study. However, the basaltic andesite of Tarn

Plateau, a high-magnesium andesite (Moore and DeBari

2009), is the only primitive magma at Mt. Baker that has

the REE signature and major element characteristics that

could be parental to the Glacier Creek andesite. The rela-

tionship between these two units will be discussed sepa-

rately (Moore and DeBari, in preparation).

Boulder glacier andesite and dacite

Models describing the generation of intermediate to felsic

volcanic rocks in the Cascade arc can be placed into two

different categories, crustal melting and crystal fraction-

ation. At Mount St. Helens, Smith and Leeman (1987) have

interpreted dacites as the result of partial melting of a lower

crustal basaltic amphibolite source. Similar interpretations

have been made at Mount Jefferson where two types of

crustal melting to form intermediate to felsic rocks have

been identified; melting of a basaltic amphibolite to form

Table 7 Glacier Creek olivine subtraction calculations

Label SiO2 TiO2 Al2O3 FeO* MgO CaO Na2O K2O Total Mg#

Original whole rock composition

GC-3TB 58.70 0.87 17.24 5.69 4.68 7.09 3.85 1.58 100.06 63

GC-6TB 58.34 0.85 17.25 5.61 4.56 7.02 3.86 1.58 99.41 63

GC-12TB 58.72 0.83 17.55 5.66 4.75 7.15 3.77 1.50 100.27 64

Subtracted olivine composition

38.00 15.00 40.00 7.00

New whole rock compositions after subtraction of 4.0% xenocrystic olivine

GC-3TB 59.56 0.91 17.96 5.30 3.21 7.09 4.01 1.65 99.69 56

GC-6TB 59.19 0.89 17.97 5.22 3.08 7.02 4.02 1.65 99.03 55

GC-12TB 59.58 0.86 18.28 5.27 3.28 7.16 3.93 1.56 99.93 57

30

40

50

60

70

58 59 59 60 60

Wt.% SiO2

Mg

#

Whole Rock Mg#

Whole-Rock Mg# minus Olivine

Whole Rock minus 4.0% Olivine

Range of Mg#Liq calculated from evolved

olivine cores

Fig. 14 Glacier Creek olivine subtraction model. Dashed-in area

indicates calculated range of liquid Mg#s in equilibrium with evolved

olivine cores. Calculations use the liquid/olivine equilibrium con-

straints described by Roeder and Emslie (1970). See Table 6 for

Mg#Liq values. Solid diamonds represent measured Glacier Creek

whole rock compositions. ‘‘X’’ symbols represent modified Glacier

Creek whole rock compositions after subtraction of 4.0 wt% xeno-

crystic olivine. These modified compositions are in the range of

liquids calculated to be in equilibrium with cores from evolved

olivine phenocrysts

Contrib Mineral Petrol (2011) 161:75–99 95

123

rhyodacite and melting of a MORB-like granulite to form

Sr-rich andesite (Conrey et al. 2001). At Mount Rainer,

crustal melting processes are postulated for the formation

of a felsic end-member, but the melting source is not

identified (Venezky and Rutherford 1997). A less common

model describes Mount St. Helens dacites not as partial

melts of lower crust, but as partial melts of the subducting

slab (Defant and Drummond 1993). In contrast to crustal

melting models, Hildreth (2007) proposes that the majority

of dacites in the Cascade arc are slightly evolved products

of an andesitic liquid. Bacon and Druitt (1988) also favor a

crystal fractionation model for generation of rhyodacites at

Crater Lake Caldera.

The Boulder Glacier unit is clearly not a fractionation

product of the basalts from the Sulphur Creek unit

(Fig. 11). But could it be a fractionation product of the

more Si-poor Glacier Creek andesite? The Glacier Creek

andesite has a distinctive REE pattern that has lower

abundances than the Boulder Glacier unit (Fig. 9). An

alternative Rayleigh fractionation model using the Glacier

Creek andesite sample as a starting composition is shown

in Fig. 15. In this model, fractionation of observed phases

(70% plagioclase ? 18% orthopyroxene ? 10% oliv-

ine ? 2% magnetite ? 0.1% apatite) predicts composi-

tions that are comparable in the light to mid-REE (La-Sm)

abundances to observed Boulder Glacier samples. How-

ever, this is after high amounts of fractionation (*70–

80%). In contrast, modeled concentrations of the HREE

and Eu are lower than observed Boulder Glacier compo-

sitions after the same amount of fractionation. Such high

amounts of fractionation and inconsistent REE concentra-

tions do not support this simple REE crystal fractionation

model. Additionally, fractionation modeling does not pre-

dict the steep drop in Sr content (Fig. 9), nor the steep

increase in Rb content (Fig. 8) observed in the Boulder

Glacier unit.

Fractional crystallization of the Glacier Creek unit was

also tested using whole rock major element compositions

with MELTS (Ghiorso et al. 1993). Fractionation of sample

GC-6TB at the NNO buffer and pressures ranging from 1 to

3 kbar consistently produced liquid compositions with

lower concentrations of MgO and Al2O3 and higher con-

centrations of TiO2 and K2O than the observed Boulder

Glacier compositions.

On the basis of major element and REE modeling, it is

unlikely that the Boulder Glacier andesites and dacites are

fractionated products of either the Sulphur Creek basaltic

andesite or the Glacier Creek andesite. Fractional crystal-

lization from other parent compositions cannot be ruled

out, but are not considered likely given the lack of corre-

lation between REE and SiO2 contents within this unit

(Fig. 9) and the wide compositional range of some trace

elements.

The felsic end-member of the Boulder Glacier unit could

well be a melt of the lower crust, which is *45 km in this

region (Ramachandran et al. 2006). The ‘‘deep crustal hot

zone’’ of Annen et al. (2006) may apply here, where silicic

and intermediate melts are generated by emplacement of

hot, mantle-derived basalt sills and dikes into the lower

crust.

What would this lower crust look like? At Mount Baker,

adjacent and potentially underlying basement rocks consist

of multiple accreted terrains of the NW Cascades Thrust

System (Tabor et al. 2003). These terranes consist of a

variety of juvenile oceanic units with both metaigneous and

metasedimentary protoliths. In addition, Miller et al. (2003)

describe the North Cascades crystalline core (presumed to

underlie the NW Cascades Thrust System) as generally

composed of amphibolite-facies metamorphic rocks that

have been successively intruded by Cretaceous plutons.

Melting of crust hosting a variety of lithologies such as

these can produce medium-K crustal melts with variable

REE characteristics depending on the presence or absence

of garnet and amphibole (e.g. Wolff and Wyllie 1994;

Rapp and Watson 1995; Sen and Dunn 1994; Conrey et al.

2001). The Boulder Glacier lavas do have low MREE and

HREE abundances, suggestive of an influence from garnet

and/or amphibole. Thus, the geochemical characteristics of

the Boulder Glacier lavas are consistent with melting of

rocks found within the crustal column at Mt. Baker.

The range of compositions from dacite to andesite in the

Boulder Glacier unit may be due to several factors. The

most important consideration is that the Boulder Glacier

unit consists of multiple flows with unclear age relation-

ships (Hildreth et al. 2003). The lack of REE correlation

with SiO2 may be due to sources or processes that varied

over the time that the unit was erupted. These variations

10

100

La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Yb Lu

Ree

Co

nce

ntr

atio

n/C

ho

nd

rite

GC-5TBBG-10TBF=0.8F=0.6F=0.4F=0.2F=0.1

Rayleigh FractionationC m/C p=F (D – 1)

Where F =melt fraction

Fractionating Modes:70% Plag18% Opx10% Olv2% Mag0.1% Apt

Fig. 15 Rayleigh fractionation model with Glacier Creek andesite

(GC-5TB) as parent. Model shows results for various values of F(melt fraction). Model results can be compared to Boulder Glacier

sample BG-10TB. The fractionating assemblage is 70% plagio-

clase ? 18% orthopyroxene ? 10% olivine ? 2% magnetite ? 0.1%

apatite. The best match is after 80–90% fractionation (F = 0.10–0.20)

96 Contrib Mineral Petrol (2011) 161:75–99

123

over time could include the following: 1) melt input from

variable crustal sources through time; 2) variable amounts

of fractionation; and 3) variable amounts of mixing with

multiple mafic end-members.

As discussed previously, the process or source that

produced the most felsic Boulder Glacier compositions

must have been active repeatedly for at least 70–80 ka of

Mt. Baker history. This includes the eruption of the

Boulder Glacier lavas at 80–90 ka to the mixing of a

similar component with the Sulphur Creek lavas at 9.8 ka.

Summary and conclusions

The late Pleistocene lavas described in this study, the

andesite of Glacier Creek, the andesite of Boulder Glacier

and the basalt of Sulphur Creek, contain chemical and

textural characteristics that are distinctive to each unit.

These distinctive features are the result of a diverse range

of open-system differentiation processes.

Neither of the two andesite units can be related to the

Sulphur Creek basaltic lava by fractionation processes.

Their geochemistry is distinct enough that they must be

generated from compositionally heterogeneous mantle-

derived parents and/or crustal melting. Nor can even the

range of compositions within the Sulphur Creek unit be

explained by crystal fractionation. Pervasive disequili-

brium textures coupled with chemical data indicate that the

bimodal compositions in the Sulphur Creek unit are the

result of varying degrees of mixing with a dacitic end-

member that is similar to the dacitic compositions of the

Boulder Glacier unit. All Sulphur Creek samples display

textural evidence, indicating that they have been mixed to

some degree; however, the samples in the more felsic

group require mixing proportions of *70% mafic Sulphur

Creek magma and *30% Boulder Glacier dacite.

The Glacier Creek andesite is intermediate in SiO2

between the Sulphur Creek and the Boulder Glacier unit. It

is enriched in MgO, Cr and Ni and is very depleted in REE

with steep patterns ([La/Yb]N = 6.7). This unit is a distinct

magma type, perhaps related to the more primitive Tarn

Plateau high-Mg andesite of Moore and DeBari (2009). We

show that the addition of *4% xenocrystic olivine from an

unconstrained mafic/ultramafic source modified the whole

rock composition from its original liquid composition.

Olivine addition is supported by elevated MgO, Ni and Cr

abundances, the presence of two distinct olivine phenocryst

phases, and strong disequilibrium textures. The source for

the xenocrystic olivine is unconstrained, but the most likely

possibilities include olivine-rich cumulates or dunitic

country rock.

The andesites and dacites of Boulder Glacier cannot be

related to the Sulphur Creek or Glacier Creek units through

crystal fractionation. The unit consists of at least 25

stacked flows whose stratigraphy and age relationships are

not clearly worked out. Geochemical trends within the unit

do not support a simple closed system process for the

generation of all the lava flows. The most felsic composi-

tions in the unit are more likely derived by partial melting

of a crustal protolith. A specific protolith is not constrained,

but amphibolite with residual amphibole and garnet and

possibly apatite could produce the depleted REE concen-

trations observed in Boulder Glacier samples. Processes

responsible for the generation of the Boulder Glacier unit

are long lived and active for up to 70–80 ka. If consistent

with the majority of crustal melting models in the

Fig. 16 Schematic model of

intermediate magma generation

at Mount Baker. Underplating

of mantle-derived basalt (pink)

induces partial melting of an

amphibolite crustal protolith to

generate felsic crustal melts

(green). Mixing of these end-

members at various levels in the

crust produce intermediate

compositions (blue). Interaction

and addition of mafic crystal

debris (green pattern) may

further modify intermediate

magma compositions

Contrib Mineral Petrol (2011) 161:75–99 97

123

literature, it is likely that melting occurs in lower crustal

levels where mantle-derived basalts provide a heat flux that

induces melting. In this model, the influx of basalt and

partial melting may occur consistently or periodically, over

long time periods of at least 70–80 ka.

When combined, the story of each unit in this study

provides a petrogenetic model for the generation of inter-

mediate lavas at Mount Baker (Fig. 16). Magma mixing is

a principal process that influences the chemical signatures

and textural characteristics of intermediate compositions

studied here and may also be a significant process for

intermediate lavas generated in the last 100 ka at Mount

Baker. In the mixing model, there must be at least three

compositionally distinct end-members: (1) a mafic end-

member that is similar to, but more primitive than, the

Sulphur Creek basalt, (2) a felsic end-member that is

similar to the dacitic compositions of the Boulder Glacier

unit, (3) a third end-member that has inherently depleted

REE concentrations, but high La/Yb, in order to produce

Glacier Creek type REE patterns. Mafic end-members are

presumably mantle-derived, while felsic end-members are

likely generated by partial melting of a crustal protolith.

Addition of xenocrystic mafic crystal debris may have a

significant effect on intermediate magma compositions

post-mixing events. Intermediate compositions may devi-

ate slightly from the main mixing trend due to small

amounts of differentiation, but the prevailing geochemical

signature is controlled by mixing and other open-system

processes.

Acknowledgments Funding for this project was provided by the

Cascade Volcano Observatory through the Jack Kleinman Internship

for Volcano Research, Sigma Xi’s Grants in Aid of Research pro-

gram, the Western Washington University Geology Department and

the Funds for the Enhancement of Graduate Research provided by the

Western Washington University Research and Sponsored Programs

department. Analytical work for this study was made possible with

the assistance from George Mustoe, Scott Kuehner, Diane Johnson

and Charles Knaack. This manuscript was significantly improved by

reviews from Pete Stelling, Scott Linneman, Michael Dungan and

Charles Bacon.

References

Andersen DJ, Lindsley DH (1985) New (and final!) models for the Ti-

magnetite/ilmenite geothermometer and oxygen barometer. EOS

Trans Am Geophys Union 66(18):416

Andersen DJ, Lindsley DH, Davidson PM (1993) QUILF; a Pascal

program to assess equilibrium among Fe-Mg-Mn-Ti oxides,

pyroxenes, olivine, and quartz. Comput Geosci 19(9):1333–

1350

Annen C, Blundy JD, Sparks RSJ (2006) The genesis of intermediate

and silicic magmas in deep crustal hot zones. J Petrol 47(3):505–

539

Arculus RJ, Johnson RW, Chappell BW, McKee CO, Sakai H (1983)

Ophiolite-contaminated andesites, trachybasalts, and cognate

inclusions of Mount Lamington, Papua New Guinea; anhydrite-

amphibole-bearing lavas and the 1951 cumulodome. J Volcanol

Geotherm Res 18(1–4):215–247

Armstrong JT (1988) Quantitative analysis of silicate and oxide

materials: comparison of Monte Carlo, ZAF, and /(qz) proce-

dues. In: Newbury DE (ed) Microbeam analysis-1988, San

Francisco Press, San Francisco, pp 239–246

Arth JG (1976) Behavior of trace elements during magmatic

processes; a summary of theoretical models and their applica-

tions. J Res US Geol Surv 4(1):41–47

Bacon CR, Druitt TH (1988) Compositional evolution of the zoned

calcalkaline magma chamber of Mount Mazama, Crater Lake,

Oregon. Contrib Mineral Petrol 98(2):224–256

Bacon CR, Hirschmann MM (1988) Mg/Mn partitioning as a test for

equilibrium between coexisting Fe-Ti oxides. Am Mineral 73(1–

2):57–61

Bacon CR, Bruggman PE, Christiansen RL, Clynne MA, Donnelly-

Nolan JM, Hildreth W (1997) Primitive magmas at five Cascade

volcanic fields; melts from hot, heterogeneous sub-arc mantle.

Can Mineral 35(2):397–423

Bostock MG, VanDecar JC (1995) Upper mantle structure of the

northern Cascadia subduction zone. Can J Earth Sci 32(1):1–12

Bullen TD, Clynne MA, Muffler LJP (1990) Trace element and

isotopic constraints on magmatic evolution at Lassen volcanic

center. J Geophys Res 95(B12):19671–19691

Clynne MA (1999) A complex magma mixing origin for rocks

erupted in 1915, Lassen Peak, California. J Petrol 40(1):105–132

Conrey RM, Hooper PR, Larson PB, Chesley J, Ruiz J (2001) Trace

element and isotopic evidence for two types of crustal melting

beneath a High Cascade Volcanic Center, Mt. Jefferson, Oregon.

Contrib Mineral Petrol 141(6):710–732

Coombs HA (1939) Mount Baker, a Cascade volcano. Geol Soc Am

Bull 50(10):1493–1509

Davidson J, Turner S, Handley H, Macpherson C, Dosseto A (2007)

Amphibole ‘‘sponge’’ in arc crust? Geology 35(9):787–790

Defant MJ, Drummond MS (1993) Mount St. Helens; potential

example of the partial melting of the subducted lithosphere in a

volcanic arc. Geology 21(6):547–550

Dungan MA, Davidson J (2004) Partial assimilative recycling of the

mafic plutonic roots of arc volcanoes; an example from the

Chilean Andes. Geology 32(9):773–776

Eichelberger JC (1975) Origin of andesite and dacite; evidence of

mixing at Glass Mountain in California and at other Circum-

Pacific volcanoes. Geol Soc Am Bull 86(10):1381–1391

Eichelberger JC, Heiken F, Widdicombe R, Wright D, Keady CJ,

Cobb DD (1976) New fumarolic activity on Mount Baker;

observations during April through July, 1975. J Volcanol

Geotherm Res 1(1):35–53

Eugster HP, Wones DR (1962) Stability relations of the ferruginous

biotite, annite. J Petrol 3(1):82–125

Evans BW, Scaillet B, Kuehner SM (2006) Experimental determina-

tion of coexisting iron-titanium oxides in the systems FeTiAlO,

FeTiAlMgO, FeTiAlMnO, and FeTiAlMgMnO at 800 and 900

degrees C, 1–4 kbar, and relatively high oxygen fugacity.

Contrib Mineral Petrol 152(2):149–167

Fujimaki H, Tatsumoto M, Aoki K-I (1984) Partition coefficients of

Hf, Zr, and REE between phenocrysts and groundmasses.

J Geophys Res 89(Suppl B):662–672

Ghiorso MS, Sack RO, Anonymous (1993) MELTS; software for the

thermodynamic analysis of phase equilibria in magmatic

systems. Abstr Programs Geol Soc Am 25(6):96

Gill JB (1981) Orogenic andesites and plate tectonics, minerals and

rocks. Springer, Berlin xiv, 385 p

Green NL (1988) Basalt-basaltic andesite mixing at Mount Baker

Volcano, Washington; estimation of mixing conditions.

J Volcanol Geotherm Res 34(3–4):251–265

98 Contrib Mineral Petrol (2011) 161:75–99

123

Green NL, Harry DL (1999) On the relationship between subducted

slab age and arc basalt petrogenesis, Cascadia subduction

system, North America. Earth Planet Sci Lett 171(3):367–381

Green NL, Sinha AK (2005) Consequences of varied slab age and

thermal structure on enrichment processes in the sub-arc mantle

of the northern Cascadia subduction system. J Volcanol

Geotherm Res 140(1–3):107–132

Grove TL, Baker MB (1984) Phase equilibrium controls on the

tholeiitic versus calc-alkaline differentiation trends. J Geophys

Res 89(B5):3253–3274

Grove TL, Parman SW, Bowring SA, Price RC, Baker MB (2002)

The role of an H2O-rich fluid component in the generation of

primitive basaltic andesites and andesites from the Mt. Shasta

region, N California. Contrib Mineral Petrol 142(4):375–396

Grove TL, Baker MB, Price RC, Parman SW, Elkins-Tanton LT,

Chatterjee N, Muentener O (2005) Magnesian andesite and

dacite lavas from Mt. Shasta, Northern California; products of

fractional crystallization of H2O-rich mantle melts. Contrib

Mineral Petrol 148(5):542–565

Guffanti M, Weaver CS (1988) Distribution of late Cenozoic volcanic

vents in the Cascade Range; volcanic arc segmentation and regional

tectonic considerations. J Geophys Res 93(B6):6513–6529

Heaton TH, Kanamori H (1984) Seismic potential associated with

subduction in the northwestern United States. Bull Seismol Soc

Am 74(3):933–941

Hildreth W (1996) Kulshan caldera: a Quaternary subglacial caldera in

the North Cascades, Washington. Geol Soc Amer Bull

108(7):786–793

Hildreth W (2007) Quaternary magmatism in the Cascades; geologic

perspectives. USGS Professional Paper, no. P 1744, p 125

Hildreth W, Moorbath S (1988) Crustal contributions to arc magmatism

in the Andes of central Chile. Contrib Mineral Petrol 98(4):455–489

Hildreth W, Fierstein J, Lanphere MA (2003) Eruptive history and

geochronology of the Mount Baker volcanic field, Washington.

Geol Soc Am Bull 115(6):729–764

Johnson DM, Hooper PR, Conrey RM (1999) XRF analysis of rocks

and minerals for major and trace elements on a single low

dilution Li-tetraborate fused bead. Adv X-Ray Anal 41:843–867

Kelemen PB (1994) Origin of high Mg# andesite and the continental

crust. Mineral Mag 58A(A-K):464–465

Leeman WP, Lewis JF, Evarts RC, Conrey RM, Streck MJ (2005)

Petrologic constraints on the thermal structure of the Cascades

Arc. J Volcanol Geotherm Res 140(1–3):67–105

Lepage L (2003) ILMAT: an Excel worksheet for ilmenite-magnetite

geothermometry and geobarometry. Comput Geosci 29:673–678

Madeleine CSH, Blundy JD, Sparks RSJ (2006) Magma evolution

and open-system processes at Shiveluch Volcano; insights from

phenocryst zoning. J Petrol 47(12):2303–2334

Mason PRD, Downes H, Thirlwall MF, Seghedi I, Szakacs A, Lowry

D, Mattey D (1996) Crustal assimilation as a major petrogenetic

process in the East Carpathian Neogene and Quaternary

continental margin arc, Romania. J Petrol 37(4):927–959

McCrory PA, Blair JL, Oppenheimer DH, Walter SR (2004) Depth to

the Juan de Fuca slab beneath the Cascadia subduction margin—

a 3-D model for sorting earthquakes, Series 91, U.S. Geological

Survey Data, CD-Rom 1.2

McKeever D (1977) Volcanology and geochemistry of the south flank

of Mount Baker, Cascade Range, Washington. Western Wash-

ington University, M.S. Thesis, 126 p

Miller RB, Matzel JP, Paterson SR, Stowell H (2003) Cretaceous to

Paleogene Cascades Arc; structure, metamorphism, and time-

scales of magmatism, burial, and exhumation of a crustal section.

GSA Field Guide 4:107–135

Mooney WD, Weaver CS (1989) Regional crustal structure and

tectonics of the Pacific coastal states; California, Oregon, and

Washington. Memoir Geol Soc Am 172:129–161

Moore NE, DeBari SM (2009) Origin and geochemical evolution of

mafic magmas from the Cascade arc, Mount Baker, Washington:

probes into mantle and crustal processes. Abstr Programs Geol

Soc Am 41(7):191

Onyeagocha AC (1978) Twin Sisters Dunite; petrology and mineral

chemistry. Geol Soc Am Bull 89(10):1459–1474

Post A, Richardson D, Tangborn WV, Rosselot FL (1971) Inventory

of glaciers in the North Cascades, Washington. USGS Profes-

sional Paper 705a, p 26

Ramachandran K, Hyndman RD, Brocher TM (2006) Regional P

wave velocity structure of the northern Cascadia subduction

zone. J Geophys Res 111:15

Rapp RP, Watson EB (1995) Dehydration melting of metabasalt at 8–

32 kbar: implications for continental growth and crust-mantle

recycling. J Petrol 36:891–931

Roeder PL, Emslie RF (1970) Olivine-liquid equilibrium. Contrib

Mineral Petrol 29(4):275–289

Schaaf P, Stimac J, Siebe C, Macias JL (2005) Geochemical evidence

for mantle origin and crustal processes in volcanic rocks from

Popocatepetl and surrounding monogenetic volcanoes, central

Mexico. J Petrol 46(6):1243–1282

Scott KM, Macias JL, Naranjo JA, Rodriguez S, McGeehin JP (2001)

Catastrophic debris flows transformed from landslides in volca-

nic terrains; mobility, hazard assessment, and mitigation strat-

egies. USGS Professional Paper 1630, p 59

Sen C, Dunn T (1994) Dehydration melting of a basaltic composition

amphibolite at 1.5 and 2.0 GPa: implications for the origin of

adakites. Contrib Minerol Petrol 117(4):394

Sisson TW (1994) Hornblende-melt trace-element partitioning mea-

sured by ion microprobe. Chem Geol 117(1–4):331–344

Smith DR, Leeman WP (1987) Petrogenesis of Mount St. Helens

dacitic magmas. J Geophys Res 92(B10):10313–10334

Stavert LW (1971) A geochemical reconnaissance investigation of

Mount Baker andesites. Western Washington University, M.S.

Thesis, 60 p

Streck MJ, Dungan MA, Malavassi E, Reagan MK, Bussy F (2002)

The role of basalt replenishment in the generation of basaltic

andesites of the ongoing activity at Arenal Volcano, Costa Rica;

evidence from clinopyroxene and spinel. Bull Volcanol

64(5):316–327

Streck MJ, Leeman WP, Chesley J (2007) High-magnesium andesite

from Mount Shasta; a product of magma mixing and contam-

ination, not a primitive mantle melt. Geology 35(4):351–354

Sun SS, McDonough WF (1989) Chemical and isotopic systematics

of oceanic basalts; implications for mantle composition and

processes. Geol Soc Spec Publ 42:313–345

Swan VL (1980) The petrogenesis of the Mount Baker volcanics,

Washington. Washington State University, Ph.D

Tabor RW, Haugerud RA, Hildreth W, Brown EH (2003) Geologic

map of the Mount Baker 30- by 60-minute quadrangle,

Washington. Publications of the U. S. Geological Survey, scale

1:100,000

Taylor DD (2001) Petrology and geochemistry of mafic lavas near

Glacier Peak, North Cascades, Washington. Western Washing-

ton University, M.S. Thesis, 86 p

Tucker DS, Scott KM (2009) Structures and facies associated with the

flow of subaerial basaltic lava into a deep freshwater lake: the

Sulphur Creek lava flow, North Cascades, Washington.

J Volcanol Geotherm Res 185:311–322

Venezky DY, Rutherford MJ (1997) Pre-eruption conditions and

timing of dacite-andesite magma mixing in the 2.2 ka eruption at

Mount Rainier. J Geophys Res 102(B9):20069–20086

Wolff MB, Wyllie PJ (1994) Dehydration melting of amphibolite at

10kbar: the effects of temperature and time. Contrib Minerol

Petrol 115:369–383

Contrib Mineral Petrol (2011) 161:75–99 99

123