Bull Volcanol (2011) 73:1455–1486

DOI 10.1007/s00445-011-0511-y

RESEARCH ARTICLE

Geochemical homogeneity of a long-lived, large silicic

system; evidence from the Cerro Galán caldera,
NW Argentina
Chris B. Folkes & Shanaka L. de Silva &
Heather M. Wright & Raymond A. F. Cas

Received: 15 March 2009 / Accepted: 5 June 2011 / Published online: 10 August 2011
# Springer-Verlag 2011

Abstract By applying a number of analytical techniques with amphibole barometry suggest that the grey pumice
across a spectrum of spatial scales (centimeter to microm- originated from potentially hotter and deeper magmas
eter) in juvenile components, we show that the Cerro Galán (800–840°C, 3–5 kbar) than the more voluminous white
volcanic system has repeatedly erupted magmas with nearly pumice (770–810°C, 1.5–2.5 kbar). The grey pumice is
identical geochemistries over >3.5 Myr. The Cerro Galán interpreted to represent the parental magmas to the Galán
system produced nine ignimbrites (∼5.6 to 2 Ma) with a system emplaced into the upper crust from a deeper storage
cumulative volume of >1,200 km3 (DRE; dense rock zone. Most inter-ignimbrite variations can be accounted for
equivalent) of calc-alkaline, high-K rhyodacitic magmas by differences in modal mineralogy and crystal contents
(68–71 wt.% SiO2). The mineralogy is broadly constant that vary from 40 to 55 vol.% on a vesicle-free basis.
throughout the eruptive sequence, comprising plagioclase, Geochemical modeling shows that subtle bulk-rock varia-
quartz, biotite, Fe–Ti oxides, apatite, and titanite. Early tions in Ta, Y, Nb, Dy, and Yb between the Galán
ignimbrite magmas also contained amphibole, while the ignimbrites can be reconciled with differences in amounts
final eruption, the most voluminous Cerro Galán ignimbrite of crystal fractionation from the “grey” parent magma. The
(CGI; 2.08±0.02 Ma) erupted a magma containing rare amount of fractionation is inversely correlated with volume;
amphibole, but significant sanidine. Each ignimbrite con- the CGI (∼630 km 3 ) and Real Grande Ignimbrite
tains two main juvenile clast types; dominant “white” (∼390 km3) return higher F values (proportion of liquid
pumice and ubiquitous but subordinate “grey” pumice. Fe– remaining) than the older Toconquis Group ignimbrites
Ti oxide and amphibole-plagioclase thermometry coupled (<50 km3), implying less crystal fractionation took place
during the upper-crustal evolution of these larger volume
magmas. We attribute this relationship to variations in
Editorial responsibility: K. Cashman magma chamber geometry; the younger, largest volume
This paper constitutes part of a special issue: ignimbrites came from flat sill-like magma chambers,
reducing the relative proportion of sidewall crystallization
Cas RAF, Cashman K (eds) The Cerro Galan Ignimbrite and Caldera:
and fractionation compared to the older, smaller-volume
characteristics and origins of a very large-volume ignimbrite and its
magma system ignimbrite eruptions. The grey pumice clasts also show
evidence of silicic recharge throughout the history of the
Electronic supplementary material The online version of this article
(doi:10.1007/s00445-011-0511-y) contains supplementary material, Cerro Galán system, and recharge days prior to eruption has
which is available to authorized users. previously been suggested based on reversely zoned (OH
C. B. Folkes (*) : H. M. Wright : R. A. F. Cas
and Cl) apatite phenocrysts. A rare population of plagio-
School of Geosciences, Building 28, Monash University, clase phenocrysts with thin An-rich rims in juvenile clasts
Victoria 3800, Australia in many ignimbrites supports the importance of recharge in
e-mail: [email protected] the evolution and potential triggering of eruptions. This
S. L. de Silva
study extends the notion that large volumes of nearly
Department of Geosciences, Oregon State University, identical silicic magmas can be generated repeatedly,
Corvallis, OR 97331, USA producing prolonged geochemical homogeneity from a
1456 Bull Volcanol (2011) 73:1455–1486

long-lived magma source in a subduction zone volcanic regional tectonic regime. Schmitt (2001) suggests that
setting. At Cerro Galán, we propose that there is a zone contractional deformation closes degassing and fracture
between mantle magma input and upper crustal chambers, pathways in the shallow crust, promoting magma accumu-
where magmas are geochemically “buffered”, producing lation and large overpressures, whereas extensional stresses
the underlying geochemical and isotopic signatures. This cause roof rock failure, sudden decompression and erup-
produces the same parental magmas that are delivered tion. Therefore, unraveling the development, longevity, and
repeatedly to the upper crust. A lower-crustal MASH history of upper crustal magmatic systems is of primary
(melting, assimilation, storage, and homogenization) zone importance for understanding not only the conditions
is proposed to act as this buffer zone. Subsequent upper leading to supervolcanic eruptions, but also to the devel-
crustal magmatic processes serve only to slightly modify opment of granite batholiths and the growth and evolution
the geochemistry of the magmas. of the continental crust. Yet, despite decades of study,
surprisingly little is known about the long-term evolution or
Keywords Ignimbrites . Central Andes . Crystal-rich timescales of upper crustal magmatic systems.
rhyodacite . Fractionation . Magma chamber The Cerro Galán Caldera in NW Argentina is an ideal site
for a case study to examine the long-term evolution of a large
silicic magma chamber, as it has produced a series of
Introduction rhyodacite ignimbrites over a long time period (>3.5 Myr),
seemingly with a restricted compositional range (Francis et al.
Large caldera-forming ignimbrite eruptions, now commonly 1989). These deposits offer the rare opportunity to probe the
referred to as super-eruptions (e.g., Rampino and Self 1992), long-term evolution of a batholith-scale system in a series of
have been well documented on the Earth across space and snapshots throughout its lifetime. Herein we report the
time (e.g., Mason et al. 2004). Each eruption records the results of an extensive multiscale petrological and geochem-
state of a batholith-scale magma body at an instant in its ical study that builds upon the initial geochemical work
evolution. Lipman (1984) and Elston (1984) proposed that performed by Francis et al. (1989). They established that
these super-eruptions eviscerate the tops of granite batholiths crystallization of the Galán magmas took place at mid-crustal
and as such, provide a vital window into the plutonic system. levels (at depths in excess of 20 km), ascending immediately
Smith (1979) and Hildreth (1981) first commented in before eruption (c.f., Whitney and Stormer Jr 1985 study of
detail on the nature of these large-volume crystalline felsic the Fish Canyon Tuff magma chamber). Francis et al. (1989)
deposits, termed “monotonous intermediates”. Early studies did not envisage large amounts of melt residing and
found that large silicic ignimbrites were erupted from crystallizing in shallow, upper crustal magma chambers.
magma chambers that were not zoned or only weakly Based on a new understanding of the stratigraphy of the
zoned (Whitney and Stormer Jr 1985; Francis et al. 1989; Galán system (Folkes et al. 2011), we have taken advantage
de Silva 1991). Since then, many studies have confirmed of new analytical techniques to conduct geochemical
and commented on the degree of chemical homogeneity of analyses on scales from whole-rock chemistry to trace
these commonly crystal-rich, calc-alkaline, high-K dacite to element compositions in individual mineral phases, in order
rhyodacite ignimbrites (Hildreth et al. 1991; Chesner 1998; to elucidate the upper crustal development and evolution of
Lindsay et al. 2001; Bachmann et al. 2002; Mason et al. the Cerro Galán magmatic system. Our focus here is the
2004). It is also recognized these magmas have high crystal upper crustal pre-eruptive magmatic evolution, not the
contents (up to 55%) resulting in very high bulk viscosities initial petrogenesis of the Galán magmas (c.f. Kay et al.
(up to 109 Pa s), as crystal fractions approach maximum 2011). In particular, we identify the parental magma that
packing (Spera 2000). Accumulation and eruption of 100s was delivered to the upper crust and show how the erupted
of cubic kilometers of these viscous magmas require a magmas evolved from this. We establish the geochemical,
unique thermal and mechanical environment in the upper petrological, and P-T history of the system through time
crust. A thermally primed and prepared crust where the and show that the Galán system records the repeated
viscoelastic behavior of chamber wall rocks would promote accumulation and evolution of large volumes of remarkably
accumulation over eruption (Jellinek and DePaolo 2003; de similar magma in the upper crust over >3.5 million years.
Silva and Gosnold 2007) appears to be a prerequisite.
Triggering of an eruption may be a natural consequence of Geology of the Cerro Galán Caldera and previous work
the sill-like geometry of the supervolcano “chambers”
(destabilizing the overlying crustal rocks) and the thermal The Cerro Galán Caldera (25°55′S 66°52′W) is located in
and mechanical character of the roof of the chamber (de the southern Central Volcanic Zone of the South American
Silva et al. 2006). The eruption of many large-volume felsic Andes (Fig. 1a), on the eastern edge of the Puna plateau.
ignimbrites is thought to have been influenced by the This plateau is broken up into numerous contractional
Bull Volcanol (2011) 73:1455–1486 1457

a b Biotite 40Ar/39Ar ages

Post-CGI
deposits 2 - 2.27 ± 0.04 Ma
APVC
2.56 ± 0.05 Ma
CGI (2.08 ± 0.02 Ma)
Bolivia
Argentina
Pacific Chile Pre-CGI
Cerro Galán caldera deposits 2.80± 0.04 Ma
Pacific Ocean Cueva 3.77 ± 0.08 Ma
Ocean Negra
Rio de Las Pitas
Vega 4.51 ± 0.11 Ma
Argentina Real 4.68 ± 0.07 Ma
Rio Pirica Grande

Toconquis Group

Merihuaca Formation
Pitas 4.84 ± 0.04 Ma
Upper Merihuaca 5.49 ± 0.11 Ma
Approximate extent
of ignimbrites Middle
N Merihuaca
Lower Merihuaca
5.56 ± 0.10 Ma
5.60 ± 0.20 Ma
0 100 200 km Blanco

Fig. 1 a Regional setting of the Cerro Galán caldera and its with Cerro Galán. New ages are from biotite 40Ar/39Ar age
associated ignimbrites. The Altiplano-Puna Volcanic Complex (APVC) determinations (CGI sanidine age in parentheses; see Folkes et al.
and locations of samples used in this study are also shown. b 2011 for a more detailed description of the stratigraphy)
Representative stratigraphic section of all the ignimbrites associated

“basins and ranges” containing Late Proterozoic and region and was then subject to a series of studies by P.W.
Paleozoic rocks that experienced several periods of defor- Francis and co-workers in the late 1970s and early 1980s.
mation (Allmendinger et al. 1997). A series of NW–SE These early studies recognized a caldera 35×20 km in size,
striking regional strike-slip faults and lineaments permeate with an elevation of ∼4,500 m at the base and ∼6,100 m at
this area of the Puna and the Galán caldera is situated at the SE the peak of the resurgent center. Volcanic activity was
margin of the Archibarca lineament. In combination with the proposed to have commenced at approximately 7 Ma
Altiplano to the North, the Altiplano-Puna plateau is the (Francis et al. 1983), although caldera-forming activity in
highest plateau in the world associated with abundant arc this region may have started much earlier, >12 Ma (Sparks
magmatism (Jordan et al. 1983; Isacks 1988; Allmendinger et et al. 1985). Heit (2005) used geophysical data to show
al. 1997). The ignimbrites and their associated calderas are evidence for a “return flow” model for fluid ascent whereby
located up to 200 km east of the present-day volcanic arc, in fluids causing the observed seismic anomalies beneath the
an area of extremely thick crust, 60–80 km thick (Zandt et al. Puna are generated at deeper levels in the asthenosphere
1994). 200 km to the north of Cerro Galán, the Altiplano- than under the Altiplano and ascend parallel to the oceanic
Puna Volcanic complex (APVC) covers ∼70,000 km2 and slab. A two-branched pathway for fluid ascent is envisaged:
comprises >30,000 km3 of crystal-rich dacitic ignimbrites one to the west associated with the recent magmatic arc,
emplaced during a Late Miocene to Pleistocene ignimbrite and the other associated with back-arc volcanism (including
flare-up (de Silva 1989a; de Silva et al. 2006). Recently Cerro Galán and the APVC) on the east of the plateau
acquired seismic, gravity, and electrical conductivity data in (Schurr et al. 2003; Heit 2005). Crustal delamination has
this region have shown a large area of at least 10–20 vol.% also been recognized beneath the Cerro Galán region, and
partial melt residing in the middle crust (∼17–19 km depth), so like the APVC, the ultimate drive for the magmatism
the Altiplano Puna Magma Body (APMB; Chmielowski et comes from the mantle (Kay et al. 2011).
al. 1999). This is thought to represent the remnants of a large This paper builds upon the petrological and geochemical
sill that acted as a buffer between mantle-dominated lower- work by Francis et al. (1989). An update of the original
crustal magma generation and the pre-eruptive magmas in stratigraphy of Sparks et al. (1985) by Folkes et al. (2011;
the upper crust. The APMB is thought to be responsible for Fig. 1b) is referred to here. Between 5.60 and 4.51 Ma,
the sustained magmatism in the APVC (de Silva et al. 2006; eruptive activity consisted of medium to large-volume
de Silva and Gosnold 2007). Delamination of the lower ignimbrite eruptions (collectively producing the Toconquis
lithosphere may have been the trigger for the elevated mantle Group; starting with the oldest, the Blanco, Lower
power input required to drive the flare-up in this location Merihuaca, Middle Merihuaca, Upper Merihuaca, Pitas,
(e.g., Kay et al. 1999). and Real Grande ignimbrites) along the western margins of
The Cerro Galán caldera was first identified by Friedman the present-day caldera. These Toconquis Group ignim-
and Heiken (1977) with the advent of satellite images of the brites are slightly indurated, pumice-rich, and massive in
1458 Bull Volcanol (2011) 73:1455–1486

texture, with thin layers of pyroclastic surge and ash-fall 67 to 70 wt.%. The predominant mineral assemblage is
deposits at the base of each ignimbrite. The thickness and plagioclase–biotite–quartz–apatite–magnetite–ilmenite
volume of ignimbrites increased over time, culminating in with varying amounts of sanidine and hornblende. The
deposition of the <120 m thick Real Grande Ignimbrite, ignimbrites have 87Sr/86Sr ratios of 0.7108–0.7181 and
representing an estimated magma volume of ∼390 km3 143
Nd/144Nd ratios of 0.51215–0.51225 (Francis et al.
(DRE). 1989). These authors also showed that the Galán ignim-
There was then a hiatus of ∼0.9 Myr until the Cueva brites contain variations in trace elements such as Rb, Sr,
Negra Ignimbrite was erupted on the eastern flanks of the Ba, and Th that would be consistent with eruption from a
present-day caldera. This ignimbrite is massive, pumice- weakly zoned magma chamber. Geochemical modeling
rich, and highly welded throughout most of its thickness. using assimilation and fractional crystallization processes
The first activity associated with the modern Galán caldera showed the Galán magmas were derived from a source
was at 2.80±0.04 Ma with the extrusion of lava domes and enriched in K, Rb, Ba, U, Ta/Sm, Ta/Th, and Sr, with a
small-volume pyroclastic flows. The eruption of the Cerro contaminant similar in radiogenic components to the
Galán Ignimbrite (CGI) at 2.08±0.02 Ma (sanidine age; dacitic ignimbrites.
biotite 40Ar/39Ar age is 2.56±0.05 Ma) produced a caldera,
with deposits reaching at least 40 km from the caldera
structural margins in all directions (Fig. 2 of Folkes et al. Analytical techniques
2011). This ignimbrite is massive, crystal-rich, pumice, and
lithic poor, representing a magma volume estimated at All of the ignimbrites associated with the Galán caldera
∼630 km3 (DRE). Post-caldera activity continued for over system have similar field and petrological characteristics
0.5 Myr after the eruption of the CGI with pyroclastic (Folkes et al. 2011). As a consequence, a detailed
surges and flows dated at 2.27±0.05 and 2.00±0.03 Ma sampling study was undertaken at the thickest and most
(biotite 40Ar/39Ar ages; Fig. 1b). complete sequences of ignimbrite deposits at proximal
The petrology and geochemistry of the CGI (with localities along Rio de Las Pitas and Rio Pirica, ∼12 km
limited analyses of the older Toconquis Formation to the west and 10 km to the south-west of the structural
ignimbrites) were first presented by Francis et al. margin, respectively (Fig. 1a). Samples were selected
(1989) who showed that all samples have similar, high- from these localities for bulk analyses and more detailed
K, dacitic bulk-rock compositions, with SiO2 ranging from geochemistry within the stratigraphic context of the
Fig. 2 Histograms showing 30 18
the bulk-rock compositions of Mean = 69.5 16 Mean = 1.12
all ignimbrite samples analyzed 25 Std Dev = 1.01 Std Dev = 0.18
14
by XRF. A/CNK=Al2O3/(CaO+
Frequency

20 12
Na2O+K2O). All histograms
show a strong Gaussian distri- 10
15
bution with low standard devia- 8
tion values. Maximum 2σ error 10 6
bars are shown 4
5
2
0 0
65 66 67 68 69 70 71 72 0.80 0.90 1.00 1.10 1.20 1.30 1.40 1.50
SiO2 (wt%) MgO (wt%)
25 25
Mean = 1.46 Mean = 0.57
Std Dev = 0.06 Std Dev = 0.08
20 20
Frequency

15 15

10 10

5 5

0 0
1.32 1.36 1.40 1.44 1.48 1.52 1.56 1.60 1.64 0.40 0.50 0.60 0.70 0.80 0.90 1.00
A/CNK TiO2 (wt%)

Post-CGI Pre-CGI Pitas Lower Merihuaca

CGI Cueva Negra Upper Merihuaca Blanco
Xl-rich CGI Real Grande Merihuaca Merihuaca
Bull Volcanol (2011) 73:1455–1486 1459

revised stratigraphy (Folkes et al. 2011). Large pumice Mineral LA-ICPMS analyses
clasts (>5 cm diameter, as suggested by Wolff 1985) were
preferentially selected as these cover a wide range of Trace element concentrations in individual minerals were
physical textures and best represent the composition of the analyzed by laser ablation ICPMS in the W.M. Keck
pre-eruptive magma, minimizing the problems of over- Collaboratory for Plasma Spectrometry, Oregon State
sized crystals, magma fragmentation during eruption and University using a NewWave DUV 193 mm ArF Excimer
mechanical fractionation during pyroclastic flow (Wolff laser and VG PQ ExCell Quadrupole ICP-MS. Analyses
1985; de Silva and Francis 1989). Where possible, were performed following the methods of Kent et al.
different geochemical techniques (electron microprobe (2004).
analyses (EPMA) and laser ablation inductively coupled
plasma mass spectrometry (LA-ICPMS) analyses) were
performed on the same pumice clasts to provide a Results
consistent and reliable dataset.
Petrographic reconnaissance and mineral modal analyses Juvenile clast types
were performed at Monash University, Melbourne, Aus-
tralia. Average modal proportions of the main mineral Three main juvenile clasts are present throughout the Galán
phases (quartz, feldspars, biotite, hornblende, and oxides) ignimbrites. By far, the most common type is “white”
as well as some accessory phases (zircon, titanite, and pumice and it accounts for ∼95% by volume of all juvenile
apatite) for each ignimbrite were calculated using a point- clasts and is characterized by its color, high content (up to
counting method with at least 1,000 counts per thin section. 55% on a vesicle-free basis) of large (up to 8 mm diameter)
The analyses were then normalized to vesicle-free propor- crystals and generally large vesicles. Grey pumice clasts are
tions. Relative proportions of individual oxides (ilmenite ubiquitous, but account for only ∼5% by volume of the
and magnetite) were estimated during geochemical analyses juvenile clasts proportion. They are characterized by their
by electron microprobe for 74 thin sections. Bulk-rock x- grey color and finer-grained crystal and vesicle textures.
ray fluorescence (XRF; for major and minor elements) and The physical differences between these juvenile clast types
inductively coupled mass spectrometry (ICPMS; for trace are related to textural differences in vesicle size and the
elements) analyses were performed at Washington State presence (grey pumice) or absence (white pumice) of
University, following the methods of Johnson et al. (1999) microlites, and in the CGI, are coincident with subtle
for XRF analyses and Knaack et al. (1994) for ICPMS variations in bulk Ba, Cu, Pb, Zn concentrations and glass
analyses. Analytical precision was better than ±3% (rela- chemistries (Wright et al. 2011).
tive) for all major elements and ±10% (relative) for trace The third type of juvenile clast is dense, highly
elements. The relative error for REE analysis is considered crystalline (≤90% crystals on a vesicle-free basis), and rare
to be <5%. (<1% of the total juvenile clast population). This crystal-
rich clast type has the same mineral assemblage as the other
Electron microprobe analyses juvenile clasts, with phenocrysts that are intermediate in
size between those of the grey and white pumice clasts
EPMA of minerals and glasses were performed at Oregon (maximum crystal sizes in the crystal-rich juvenile clasts
State University using a CAMECA SX-100 instrument are ≤3 mm in diameter). A more detailed investigation of
equipped with five wavelength dispersive spectrometers this juvenile clast type is presented in the petrology and
and high intensity dispersive crystals for high sensitivity mineral geochemistry section later in this paper.
trace element analysis. Minerals were analyzed using a 15-
keV accelerating voltage, 30-nA sample current and 1-μm General geochemical characteristics of the Galán system
beam diameter. A 12-keV electron beam at 20 nA with a 2–
10-μm diameter defocused beam was used to analyze Whole-rock compositions of 71 pumice samples from all
groundmass glass compositions. Counting times ranged ignimbrites associated with the Cerro Galán caldera were
from 10 to 120 s depending on the element and desired analyzed using XRF and ICP-MS techniques. All major
detection limit. In all cases, 0-time intercept functions were element analyses are normalized to 100% volatile free with
applied to reduce the effects of alkali migration. Analytical representative analyses (along with analytical totals) from
errors were 1–2% (relative) for all major components each ignimbrite presented in Table 1 (the entire dataset is
(>10 wt.%) and 3–10% (relative) for minor components available as Electronic supplementary material). All sam-
(1–10 wt.%). Data reduction was performed online using a ples can be classified as high-K, high-Si dacites to low-Si
stoichiometric PAP correction model (Pouchou and Pichoir rhyolites according to the classification schemes of Le
1985). Maitre (1989). For simplicity, we will henceforth use the
Table 1 Representative bulk-rock analyses of juvenile clasts in each Galán ignimbrite
1460

Ignimbrite Blanco Lower Middle Merihuaca Upper Merihuaca Pitas Real Grande Cueva CGI Post-CGI
Merihuaca Negra
Sample no. CG475g CG 476 CG 83a CG 86a CG 94a CG 279 CG 486 CG 68b CG 484 CG 227b CG 287 CG 122 CG 342 CG298a CG 253 CG 309
Clast type White White Grey White Grey White White Grey White Fiamme White Grey Fiamme XI-rich Dome White
pumice
Location Proximal Proximal Proximal Proximal Proximal Proximal Proximal Proximal Proximal Proximal E Proximal Distsal Medial E Medial E Intra-caldera Proximal/
W W W W W SW W W W SW W medial
Pitas Pitas Pitas Pitas Pitas Pirica Pitas Pitas Pitas E caldera Pirica Antofagasta Patos Patos/Ags E caldera Diamante
Cals Surge

a
SIO2 70.2 69.0 69.2 69.9 69.6 69.3 69.6 69.5 68.4 69.0 69.5 67.0 70.3 63.9 69.7 71.3
TIO2 0.46 0.43 0.52 0.47 0.51 0.58 0.52 0.54 0.57 0.61 0.64 0.56 0.61 1.1 0.58 0.49
Al2O3 15.4 16.0 16.1 15.1 15.5 15.3 15.5 15.6 15.2 15.3 15.2 15.4 15.0 16.5 15.1 14.9
FeO* 2.4 2.3 2.7 2.6 2.6 2.9 2.7 2.8 3.0 3.0 3.2 4.0 2.5 4.9 2.8 2.0
MnO 0.04 0.04 0.04 0.05 0.05 0.05 0.04 0.05 0.05 0.05 0.05 0.05 0.05 0.08 0.05 0.04
MgO 0.91 0.92 1.17 1.01 0.93 1.12 0.96 1.22 1.18 1.22 1.14 1.22 1.20 2.22 1.12 1.10
CaO 2.5 3.3 2.6 2.5 2.7 2.8 2.5 2.5 3.1 2.7 2.6 3.4 2.6 4.1 2.6 2.3
Na2O 3.3 3.8 3.0 3.1 3.4 2.7 3.3 2.8 3.7 3.2 3.2 3.3 3.3 3.4 3.4 3.2
K2O 4.6 4.1 4.5 5.0 4.5 5.0 4.7 4.7 4.6 4.8 4.2 4.7 4.3 3.4 4.5 4.6
P2O5 0.18 0.16 0.17 0.16 0.23 0.22 0.21 0.21 0.20 0.20 0.29 0.29 0.16 0.40 0.20 0.13
TotalNon 96.7 94.7 95.1 96.0 97.0 96.7 96.3 95.8 95.6 96.3 97.9 97.6 97.5 98.4 96.9 96.5
LOI 2.7 4.3 n.d. n.d. n.d. 2.5 2.7 n.d. 3.4 2.3 1.1 n.d. 1.6 1.1 2.1 2.8
Na2O+K2O 7.9 7.9 7.5 8.1 7.9 7.8 8.0 7.5 8.2 7.9 7.4 7.7 7.6 6.9 7.8 7.8
Ca/Na 0.75 0.86 0.89 0.80 0.80 1.03 0.77 0.90 0.86 0.85 0.80 0.77 0.77 1.18 0.78 0.72
b
Sc 4.0 5.4 5.4 4.8 4.7 5.6 4.3 5.7 6.1 6.5 7.4 7.0 6.9 9.2 6.1 6.0
V 47.6 49.7 54.3 55.9 52.7 58.5 56.1 55.2 63.1 65.9 73.5 60.9 56.2 125.8 61.4 43.5
Cr 10.1 9.0 14.7 13.0 8.3 11.4 12.7 14.5 13.9 16.9 15.4 15.0 15.4 17.2 15.3 21.5
Nl 1.7 1.4 5.1 6.4 5.1 0.0 2.3 6.3 3.3 0.5 4.0 8.5 0.1 5.3 0.0 2.9
Cu 9.5 5.9 13.1 19.2 9.3 4.9 10.3 11.3 6.3 11.5 22.4 39.5 4.8 7.6 8.5 7.0
Zn 52.1 50.5 59.5 74.2 66.7 65.2 63.4 74.4 62.6 69.4 57.7 133.3 57.9 104.5 63.6 56.2
Ga 21.0 20.9 21.5 20.5 23.0 22.6 22.2 23.3 23.5 22.3 23.4 22.1 23.3 25.2 21.6 21.8
Cs 23.5 17.0 16.5 16.9 19.2 19.2 15.5 15.9 15.7 15.4 16.8 17.8 25.3 16.3 30.7 134.9
Rb 258.8 212.5 225.0 233.8 232.2 266.4 245.0 263.7 233.6 258.0 258.6 212.0 271.5 252.9 265.8 273.1
Sr 260.4 338.9 292.6 285.9 290.2 286.0 268.0 253.2 309.0 295.7 269.4 264.4 265.9 311.0 285.3 262.5
Ba 407.3 487.0 555.9 587.9 513.0 523.5 518.5 504.9 502.9 592.5 353.9 530.3 334.8 459.1 488.7 499.1
Pb 24.4 23.9 46.4 47.3 25.2 24.9 23.7 23.0 21.5 23.9 23.6 54.4 22.9 11.7 25.5 26.9
Th 23.1 18.8 24.4 21.8 22.5 23.8 25.8 32.4 26.9 34.0 34.4 29.2 31.2 25.4 29.0 30.1
U 11.2 9.2 8.0 8.1 10.2 10.5 8.5 8.1 8.8 10.9 10.7 11.0 21.4 7.5 10.9 19.7
Zr 133.6 116.4 161.7 155.5 160.5 164.4 174.0 176.1 174.2 213.3 195.7 175.9 184.6 232.3 177.7 158.0
Hf 4.2 3.6 4.7 4.8 4.9 4.9 5.2 5.2 5.1 6.2 5.9 5.2 5.6 6.7 5.3 4.8
Bull Volcanol (2011) 73:1455–1486
Table 1 (continued)

Ignimbrite Blanco Lower Middle Merihuaca Upper Merihuaca Pitas Real Grande Cueva CGI Post-CGI
Merihuaca Negra
Sample no. CG475g CG 476 CG 83a CG 86a CG 94a CG 279 CG 486 CG 68b CG 484 CG 227b CG 287 CG 122 CG 342 CG298a CG 253 CG 309
Clast type White White Grey White Grey White White Grey White Fiamme White Grey Fiamme XI-rich Dome White
pumice
Location Proximal Proximal Proximal Proximal Proximal Proximal Proximal Proximal Proximal Proximal E Proximal Distsal Medial E Medial E Intra-caldera Proximal/
W W W W W SW W W W SW W medial
Pitas Pitas Pitas Pitas Pitas Pirica Pitas Pitas Pitas E caldera Pirica Antofagasta Patos Patos/Ags E caldera Diamante
Cals Surge
Bull Volcanol (2011) 73:1455–1486

Nb 14.9 15.3 12.3 15.9 15.3 15.9 16.9 14.4 15.1 16.6 20.9 20.0 20.5 22.5 18.1 17.1
Ta 1.7 1.6 1.5 1.5 2.0 1.7 1.6 1.5 1.6 1.7 2.4 2.5 2.4 1.8 2.0 2.0
La 36.4 32.3 39.6 37.9 50.3 38.8 43.0 38.5 49.6 50.1 58.8 44.0 49.4 69.0 46.7 45.2
Ce 73.5 63.2 80.5 75.0 81.6 77.6 86.9 97.8 98.7 114.6 118.4 88.4 98.1 137.8 92.3 88.7
Pr 8.5 7.3 8.9 8.5 9.6 9.1 10.1 11.2 11.3 13.1 13.4 10.3 11.3 16.1 10.6 10.0
Nd 30.9 26.5 31.4 30.2 35.0 33.1 36.8 39.8 40.7 46.1 47.1 37.5 39.9 58.6 37.8 35.0
Sm 6.1 5.2 5.7 5.7 6.9 6.6 7.1 7.4 7.4 8.2 8.6 7.3 7.6 10.9 7.2 6.4
Eu 1.0 1.1 1.1 1.1 1.2 1.2 1.2 1.2 1.2 1.3 1.3 1.2 1.2 1.6 1.2 1.1
Gd 4.6 4.0 4.1 4.1 5.2 5.0 5.0 5.0 5.2 5.6 6.2 5.6 5.5 8.0 5.3 4.6
Tb 0.61 0.55 0.55 0.54 0.71 0.68 0.66 0.66 0.70 0.75 0.86 0.79 0.79 1.06 0.75 0.64
Dy 2.9 2.8 2.6 2.7 3.5 3.2 3.3 3.1 3.6 2.1 4.3 5.1 4.1 5.3 3.9 3.2
Ho 0.52 0.47 0.45 0.44 0.59 0.57 0.54 0.53 0.61 0.67 0.76 0.75 0.72 0.88 0.69 0.58
Er 1.2 1.2 1.1 1.1 1.4 1.4 1.3 1.3 1.5 1.6 1.9 1.9 1.8 2.1 1.7 1.5
Tm 0.17 0.16 0.15 0.15 0.20 0.18 0.17 0.17 0.20 0.23 0.26 0.27 0.26 0.28 0.24 0.21
Yb 0.9 0.9 0.9 0.9 1.2 1.1 1.0 1.0 1.2 1.2 1.6 2.1 1.6 1.6 1.5 1.3
Lu 0.16 0.14 0.13 0.13 0.17 0.16 0.15 0.15 0.18 0.21 0.24 0.25 0.23 0.23 0.22 0.20
Y 15.5 15.7 12.2 16.8 15.9 16.0 17.0 14.7 16.2 16.9 19.8 20.1 19.1 23.1 18.7 16.5

Total Fe expressed as FeO*

Pitas section at Rio de Las Pitas, Pirica section at Rio Pirica, Diamante surge locality ∼20 km north of the structural margin, n.d. not determined, TotalNon Non-normalized total
a
Major element oxides from XRF analyses renormalized to 100%
b
Trace and rare-earth elements from ICP-MS
1461
1462 Bull Volcanol (2011) 73:1455–1486

term “rhyodacites”, following Francis et al. (1989). Figure 2 major elements and the vast majority of trace elements
shows major element bulk compositions of all analyzed (Figs. 2, 3a–d). For example, 63 of the 71 samples are
ignimbrite samples. All compositional histograms show a restricted to a SiO2 range of 68.5–70.5 wt.%, and a TiO2
strong Gaussian distribution with very low one sigma range of 0.55–0.7 wt.% (Figs. 2, 3a, b, and e). It is
values. Figure 3 presents variation diagrams showing bulk particularly important to note that this chemical homoge-
chemical variations between ignimbrites. The overlap in neity exists both within each ignimbrite and also between
these datasets demonstrates the low degree of variation in different ignimbrites. There appears to be very little major

Fig. 3 Bulk-rock major and a b

trace element variation dia- Rhyolite 6
9
grams of juvenile clasts in the Trachyte
various Galán ignimbrites. Na2O + K2O (wt %) 5 Shoshonite Series
8 Trachyandesite

K2O (wt %)
Analyses in a-c are normalized 4
Basaltic
on a 100% anhydrous basis. 7 Trachyandesite
3 High-K
Maximum 2σ error bars are
6 Dacite
shown. Ignimbrite analyses from 2 Med-K
Francis et al. (1989) are also 5 Basaltic
Andesite
1
plotted. f Eu* is calculated as Andesite Low-K
the geometric mean of normal- 4 0
50 55 60 65 70 75 80 50 55 60 65 70 75 80
ized Sm and Gd
SiO2 (wt %) SiO2 (wt %)

c d 140
2.2
120
MgO (wt %)

1.7 100

V (ppm)
80
1.2
60
0.7
40

0.2 20
14.0 14.5 15.0 15.5 16.0 16.5 17.0 150 200 250 300 350 400

Al2O3 (wt %) Rb (ppm)

e f 800

220
600
Zr (ppm)

Ba (ppm)

180
400

140 200

100 0
0.40 0.45 0.50 0.55 0.60 0.65 0.70 0.40 0.50 0.60 0.70 0.80
TiO2 (wt %) Eu/Eu*

g h
3.4 26

3.2
22
3.0
Y (ppm)
Dy/Yb

CGI
2.8 18
Toconquis Group
Ignimbrites
2.6
14
2.4

2.2 10
1.0 1.5 2.0 2.5 3.0 10 15 20 25
Ta (ppm) Nb (ppm)

Glass CGI white Cueva Negra Upper Merihuaca Blanco

Post CGI CGI grey Real Grande Middle Merihuaca Francis et al. (1989) igs
Xl-rich CGI Pre-CGI Pitas Lower Merihuaca
Bull Volcanol (2011) 73:1455–1486 1463

element bulk geochemical distinction between the ignim- volcanic system (Eu* is calculated as the geometric mean
brites and therefore little chemical change in these magmas of normalized Sm and Gd; normalizing values from Anders
over the >3.5-Myr time period in which they were erupted. and Ebihara 1982). The CGI possesses the largest Eu
Although TiO2 values overlap, there is a slight variation anomalies at Eu/Eu*=0.5–0.65 compared with the older
between the ignimbrites (Fig. 3e), with the CGI plotting at Toconquis Group ignimbrites varying between 0.55 and
the higher end of the TiO2 field (0.53–0.67 wt.%) compared 0.65 (Fig. 3f).
with the older Toconquis ignimbrites (0.43–0.60 wt.%).
Despite major element homogeneity, bulk-rock trace Petrography and mineral chemistry
element concentrations vary between ignimbrites (Fig. 3g
and h). The young CGI possesses higher concentrations of All of the ignimbrites associated with Cerro Galán possess a
Ta (>2 ppm), Y (>17 ppm), Nb (>17 ppm), and lower Dy/ similar mineral assemblage consisting of, in order of
Yb values (<2.8) than the older Toconquis Group ignim- decreasing abundance, plagioclase>biotite>quartz>apatite>
brites (Ta<2 ppm, Y<17 ppm, Nb<17 ppm and Dy/Yb> magnetite>ilmenite>zircon>titanite>monazite. Amphibole
2.8). Figure 4a presents chondrite-normalized rare-earth and sanidine are also present to varying minor degrees.
element (REE) patterns of the Galán ignimbrites. All Vesicle-free crystallinities vary from 42% to 57% (Table 2).
ignimbrites exhibit enrichment in light REE (LREE) and a The dominant modal mineralogy does not vary significantly
corresponding depletion in heavy REE (HREE) typical of with successive eruptions (Table 2; Fig. 5; Electronic
magmas with an arc affinity. The crystal-rich juvenile clasts supplementary material). For example, average plagioclase,
are the most LREE enriched in the entire sample suite. biotite, and quartz contents vary by less than 5% of the
There is a general progression over time to lower La/Yb vesicle-free total (ranging from 19.0% to 22.4%; 10.6% to
and Dy/Yb ratios (Fig. 4b). All samples possess negative 14.5%, and 8.0% to 10.2% of vesicle-free juvenile clast
Eu anomalies (Eu/Eu* = 0.5–0.66; Fig. 3f), indicating modal proportions, respectively). However, there is an
plagioclase fractionation or partial melting from a inverse relationship between amphibole and sanidine pro-
plagioclase-rich source throughout the history of the portions with the older Toconquis Group ignimbrites all

Fig. 4 a Chondrite-normalized
REE patterns for the Galán
a
1000
ignimbrites. Each line represents CGI
the average value for that ig- Xl-rich Juveniles (CGI)
Cueva Negra
nimbrite. Shaded field represents
Real Grande
the full range of each REE. Pitas
Sample/chondrite

Normalizing values are from Upper Merihuaca

Anders and Ebihara (1982). b 100 Middle Merihuaca
Variations in REE with time. Lower Merihuaca
Maximum 2σ error bars are Blanco

shown

10

1
La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu

b 1.8
2.3
Age (Ma)

2.8
3.3
Post-CGI
3.8 CGI
Pre-CGI
4.3 Cueva Negra
Real Grande
4.8 Pitas
U. Merihuaca
5.3 M. Merihuaca
L. Merihuaca
5.8 Blanco

2.2 2.4 2.6 2.8 3.0 3.2 3.4 15 25 35 45 55

Dy/Yb La/Yb
 1464

Table 2 Modal analyses data for each ignimbrite (proportions in volume percentage)

Ignimbrite CGI CGI CGI CGI Cueva Real Real Real Pitas Pitas Pitas Upper Upper Upper Middle Middle Middle Lower Blanco Blanco Blanco
Negra Grande Grande Grande Merihuaca Merihuaca Merihuaca Merihuaca Merihuaca Merihuaca Merihuaca
Juvenile type White Grey Average XI-rich White Grey Average White Grey Average White Grey Average White Grey Average White White Grey Average
juvenile
clasts
No. of 22 13 35 2 2 7 3 10 5 4 9 1 1 2 2 2 4 4 1 1 2
samplesa

Qtz phenosb 5.6 4.6 5.4 5.6 5.9 4.5 2.8 4.0 5.5 4.8 5.3 6.2 3.4 4.8 8.1 3.0 5.6 8.3 5.1 3.4 4.2
Qtz mlcros 3.1 5.8 4.2 4.5 3.2 3.9 4.4 4.0 1.9 7.2 4.7 3.0 6.6 4.8 3.1 6.2 4.6 0.9 8.3 1.3 4.8
Quartz total 8.7 10.3 9.6 10.0 9.0 8.3 7.2 8.0 7.4 12.0 9.9 9.2 10.0 9.6 11.2 9.2 10.2 9.1 13.4 4.6 9.0
Plag phenos 13.3 6.4 10.9 23.3 12.6 10.5 9.6 10.2 9.9 7.8 8.7 11.9 5.9 8.9 16.8 5.7 11.3 10.7 7.7 16.3 12.0
Plag micros 7.2 11.1 8.7 14.2 6.3 8.2 11.6 9.2 9.3 12.4 10.3 8.1 14.6 11.4 6.4 14.0 10.2 10.1 7.7 13.2 10.4
Plag total 20.6 17.5 19.6 37.5 19.0 18.7 21.1 19.4 19.3 20.2 19.0 20.0 20.6 20.3 23.3 19.7 21.5 20.8 15.4 29.5 22.4
Biotite 13.2 9.3 12.0 32.3 10.6 12.8 12.1 12.6 10.9 12.2 11.7 10.8 13.0 11.9 13.1 12.4 12.7 14.5 10.1 18.2 14.1
Magnetite 1.1 1.3 1.2 2.3 1.0 0.9 0.9 0.9 0.9 1.0 1.0 1.0 1.4 1.2 1.2 0.8 1.0 1.1 1.1 1.1 1.1
Ilmenite 2.1 1.1 1.6 1.1 1.4 1.1 0.9 1.0 1.1 1.2 1.1 0.9 1.0 0.9 1.1 0.7 0.9 1.0 1.1 1.4 1.2
Sanidine 1.5 0.9 1.3 2.8 0.8 0.3 0.5 0.3 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Homblende 0.0 0.0 0.0 0.0 0.5 0.7 0.4 0.6 0.8 0.7 0.7 0.8 2.4 1.6 1.1 2.3 1.7 1.1 1.9 0.6 1.2
Clinopyroxene 0.0 0.0 0.0 0.1 0.3 0.2 0.1 0.2 0.1 0.0 0.0 0.0 0.5 0.3 0.1 0.5 0.3 0.0 0.0 0.0 0.0
Apatite 1.5 0.9 1.3 0.4 1.4 1.6 1.2 1.5 1.2 1.3 1.2 1.8 1.6 1.7 1.5 1.5 1.5 1.8 2.0 1.3 1.7
Zircon 0.4 0.3 0.4 0.7 0.3 0.3 0.2 0.3 0.2 0.2 0.2 0.1 0.3 0.2 0.2 0.2 0.2 0.2 0.3 0.1 0.2
Monazite 0.2 0.1 0.2 0.1 0.1 0.1 0.1 0.1 0.1 0.2 0.1 0.1 0.2 0.1 0.1 0.2 0.1 0.3 0.2 0.3 0.2
Titanite 0.4 0.2 0.3 0.6 0.3 0.2 0.2 0.2 0.1 0.2 0.1 0.2 0.3 0.2 0.1 0.2 0.1 0.2 0.1 0.1 0.1
Total 49.6 42.1 47.6 87.9 44.6 45.2 44.9 45.1 42.0 49.3 45.1 44.9 51.0 48.0 52.8 47.8 50.3 50.1 45.6 57.1 51.4
Percentage 22.5 21.1 21.8 65.5 34.0 25.7 24.7 25.4 23.0 30.8 27.0 26.0 39.0 32.5 27.5 32.0 29.8 28.0 26.0 32.0 29.0
of crystals
Percentage 22.8 29.6 26.2 9.0 42.5 31.3 30.3 31.0 31.7 31.4 32.5 33.0 39.0 36.0 24.5 35.0 29.8 26.0 31.0 24.0 27.5
of matrix
Percentage 54.7 49.3 52.0 25.5 23.5 43.0 45.0 43.6 45.3 37.9 40.5 41.0 22.0 31.5 48.0 33.0 40.5 46.0 43.0 44.0 43.5
of vesicles
Percentage 49.6 42.1 45.9 87.9 44.6 45.2 44.9 45.1 42.0 49.3 45.1 44.1 50.0 47.0 52.8 47.8 50.3 52.3 45.6 57.1 51.4
of crystals
(vescicle free)

a
Data for ignimbrite units represent the average of multiple samples. Each sample represents 1,000 point counts for mineral phases (quartz, plagioclase, sanidine, amphibole, oxides, apatite,
monazite, titanite, and zircon) and proportions of crystals, matrix, and vesicles. Ilmenite and magnetite proportions calculated from a combination of point counting and reconnaissance on an
electron microprobe
b
Phenocrysts are defined as >120 μm; microphenocrysts are between 30 and 120 μm; matrix is <30 μm
Bull Volcanol (2011) 73:1455–1486
Bull Volcanol (2011) 73:1455–1486 1465

Fig. 5 The varying proportions Plagioclase Biotite Sanidine Amphibole

of different minerals (normal-
ized to vesicle-free) within each 2
ignimbrite. Age is shown on the
y-axis. Each data point repre-
sents the average proportion of
3
each mineral in a particular
ignimbrite; one standard error

Age (Ma)
bars are shown. Open diamonds
represent white pumice clasts, 4
black diamonds represent grey
pumice clasts, and the star sym-
bols represent crystal-rich juve-
nile clasts 5

6
10 20 30 40 0 20 0 2 0.0 2.0

Proportion ( vol %)

Magnetite Ilmenite Zircon Titanite

2

3
Age (Ma)

6
0.5 1.5 2.5 0.5 1.0 1.5 0.0 0.5 0.0 0.5

Proportion ( vol %)

containing amphibole (0.6–1.7 modal%) but only trace Plagioclase feldspar

amounts of sanidine (all <0.3 modal%). This is in contrast
to the younger Cerro Galán and Cueva Negra ignimbrites Plagioclase is the dominant mineral phase in the Galán
that have lower average amounts of amphibole (0.5 modal% ignimbrites, accounting for between 19.0% and 22.4% of
in the Cueva Negra Ignimbrite, none in the white CGI average vesicle-free juvenile clast modal proportions
juvenile clasts and very rare crystals in the grey CGI juvenile (Fig. 5). Average plagioclase compositions are shown in
clasts) but higher average proportions of sanidine (0.8– Table 3. Crystal sizes range from <30 μm to ∼8 mm in
1.3 modal%; Fig. 5). diameter. As with some of the other main mineral phases,
Individual crystal phases within juvenile clasts were plagioclase phenocrysts are often broken. All plagioclase
analyzed both by Electron Microprobe and LA-ICPMS crystals selected for geochemical analyses are euhedral,
methods to study their chemical variations in major, fresh, and unaltered. Different phenocryst types were
minor, and trace elements. Mineral phases were chosen identified during petrographic reconnaissance based on
that were common to all ignimbrites; plagioclase, biotite, their zoning textures. Core and rim compositions from
Fe–Ti oxides, and amphibole (except in the CGI). These each phenocryst type in each ignimbrite were then analyzed
phases are described in detail below. Analyses of by electron microprobe to give an indication of the range of
groundmass glass were also undertaken to determine compositions in each sample and between different ignim-
chemical variability. Of the main mineral phases listed brites. The range of minor and trace element concentrations
above, biotite and Fe–Ti oxides show the greatest in plagioclase is relatively constant throughout the eruptive
chemical variability between different phenocrysts within stratigraphy (e.g., Fig. 6a). Trace element chemical varia-
the same ignimbrite. tions between crystals (for example, 20–250 ppm Ba) are
 1466

Table 3 Average microprobe and LA-ICPMS analyses of plagioclase feldspar from each ignimbrite

Ignimbrite Blanco Lower Middle Merihuaca Upper Merihuaca Pitas Real Grande Cueva Negra CGI XI-rich in CGI
Merihuaca

Cores Rims Cores Rims White Rims Grey Rims White Rims Grey Rims White Rims White Rims Grey Rims Cores Rims Whitecores Rims Grey Rims Cores Rims
cores cores cores cores cores cores cores cores
No. of 9 5 31 11 52 20 18 6 8 5 6 4 85 23 121 31 19 10 32 15 150 36 28 12 14 4
analyses

SIO2a 56.2 57.6 60.2 59.2 59.0 58.9 58.8 58.9 56.8 58.3 59.5 57.9 58.7 56.6 58.6 58.1 57.4 60.3 58.3 59.2 57.5 59.3 59.1 59.3 59.6 58.7
Al2O3 27.5 26.5 25.2 25.8 25.6 26.0 25.8 25.8 26.6 25.6 25.8 25.7 25.8 26.7 26.4 26.7 26.5 25.7 26.4 25.8 26.7 25.5 25.7 25.7 25.8 26.3
FeO* 0.25 0.22 0.13 0.15 0.15 0.22 0.17 0.20 0.20 0.22 0.22 0.19 0.18 0.18 0.18 0.22 0.16 0.18 0.16 0.15 0.20 0.17 0.16 0.18 0.17 0.21
CaO 9.4 8.5 6.5 7.3 7.4 7.8 7.0 7.1 8.3 7.5 7.7 8.2 7.5 8.5 7.7 8.0 8.1 6.5 7.6 7.3 8.3 6.9 7.1 7.0 7.4 8.0
Na2O 5.4 5.7 6.9 6.5 6.4 6.2 6.8 6.7 5.8 6.3 6.2 6.5 6.3 5.7 6.2 6.1 6.1 7.0 6.4 6.5 6.0 6.7 6.5 6.5 6.6 6.2
K2O 0.42 0.55 0.77 0.71 0.67 0.62 0.63 0.63 0.56 0.69 0.63 0.56 0.67 0.54 0.69 0.65 0.52 0.70 0.65 0.71 0.56 0.67 0.67 0.70 0.63 0.59
MgO 0.01 0.01 0.01 b.d. 0.02 0.01 n.d. 0.01 0.01 b.d. 0.01 0.02 0.01 0.01 0.01 0.01 b.d. b.d. b.d. b.d. 0.02 0.01 0.01 b.d. b.d. b.d.
P2O5 n.d. n.d. n.d. n.d. 0.02 0.02 b.d. 0.01 0.01 0.03 b.d. b.d. 0.02 0.01 0.02 0.01 0.02 n.d. n.d. n.d. 0.02 0.01 0.01 b.d. n.d. n.d.
Total 99.23 99.23 99.75 99.70 99.34 99.82 99.27 99.45 98.33 98.81 100.10 99.04 99.27 98.40 99.90 99.92 98.90 100.45 99.66 99.70 99.33 99.26 99.36 99.51 100.21 100.01
Ab 49.6 52.9 62.7 59.0 58.5 56.6 61.3 60.6 51.9 57.6 53.5 56.2 57.5 53.1 56.9 55.6 55.7 63.3 57.9 59.0 54.8 61.0 60.1 60.1 59.5 56.5
An 47.8 43.7 32.7 36.8 37.5 39.6 34.9 35.6 44.2 38.2 43.1 39.5 38.7 43.6 39.0 40.5 41.1 32.5 38.2 36.8 41.9 35.0 35.9 35.7 36.8 40.0
Or 2.5 3.4 4.6 4.2 4.0 3.8 3.8 3.8 3.1 4.2 3.4 4.3 3.9 3.3 4.1 3.9 3.1 4.2 3.9 4.3 3.3 4.0 4.0 4.2 3.7 3.5
Lib 18.8 54.0 13.2 7.5 30.5 49.0 10.5 49.7 47.0 10.3 26.0 4.1 6.1
TI 55.9 49.0 97.2 114.9 72.8 60.5 78.2 65.6 40.0 150.5 29.6 77.0 52.1
86
Sr 807.2 959.7 840.2 461.4 847.9 799.8 766.3 631.6 524.7 771.8 543.0 371.2 601.8
88
Sr 830.1 976.7 864.7 479.7 856.4 805.8 782.1 650.1 546.7 718.3 555.0 373.9 610.5
Ba 134.3 149.0 97.2 146.1 107.4 103.5 97.8 116.2 86.5 106.9 73.1 140.6 94.0
La 13.2 8.8 7.8 7.3 7.2 8.0 6.5 10.0 8.3 9.2 7.7 2.7 9.8
Ce 19.0 12.5 13.1 12.0 11.8 11.8 11.7 18.8 13.8 13.5 12.8 4.2 14.1
Nd 4.5 3.2 3.3 2.6 3.1 3.2 2.6 4.5 2.8 3.9 2.6 1.4 3.7
Eu 1.7 1.5 1.1 1.1 1.1 1.3 1.0 1.4 1.3 1.1 1.3 0.4 1.5
Pb 13.7 16.3 14.0 14.8 11.2 13.7 11.2 16.5 17.6 11.5 16.3 9.4 14.6

n.d. not determined, b.d. below detection limit

a
Oxides determined by microprobe analyses (in weight percent)
b
Minor elements determined by LA-ICPMS (in ppm): Ab, An, and Or in mole percent
Bull Volcanol (2011) 73:1455–1486
Bull Volcanol (2011) 73:1455–1486 1467

likely due to the relatively large analyzed spot-size clase, zircon, ilmenite and magnetite, and no apparent core to
(∼40 μm) for these trace elements using LA-ICPMS rim compositional zonation. Representative major element
methods. A large spot-size is likely to have intersected oxides (recalculated as cations per formula unit) and trace
different growth stages and therefore slightly different element concentrations in biotite are shown in Table 4 and
chemistries of the plagioclase phenocrysts based on relative presented graphically in Fig. 6b, for each ignimbrite. The
position with respect to the true phenocryst core. In general, Mg-number of biotite remains fairly constant at 0.58–0.66 in
we find successive ignimbrite eruptions produce a remark- all ignimbrites (Table 4). Nb concentrations are slightly
ably similar range of plagioclase compositions (Fig. 6). elevated in the CGI (Fig. 6b), corresponding with the higher
Three main plagioclase populations have been identified concentration of this element in the whole-rock analyses
based on texture, composition, and zoning profiles (Fig. 7). outlined above.
Type 1 plagioclase crystals are the most common, account-
ing for approximately 70% of the total plagioclase Quartz
population. These crystals are non-zoned, with a restricted
composition throughout the width of the crystal at An32-36 Quartz is present in all ignimbrites, accounting for between
(Fig. 7a). Type 2 plagioclases are the next most common 8.0% and 10.2% of total juvenile clast modal proportions.
and these normally zoned phenocrysts account for approx- There is a large variation in crystal sizes from <30 μm to
imately 25% of the total phenocryst population. There is a ∼8 mm in diameter. Quartz crystals are resorbed, sub-
much wider range in compositions of type 2 phenocrysts, rounded to rounded, often exhibiting rounded embayments
which typically contain calcic cores of An46 - 52 with rare in their outer margins and frequently contain melt inclu-
cases in the CGI reaching as high as An70 (Fig. 7b; outliers sions. Trace element concentrations were not determined
shown in Fig. 6a). Most crystals exhibit a gradual decrease for quartz crystals, due to the detection limits for the
in An-content from the cores to a 100–200 μm An-poor rim operating conditions outlined above.
(∼An35; a similar composition to the type 1 plagioclase
crystals). Type 1 and 2 plagioclase crystals are found in all Amphibole
ignimbrites. Type 3 phenocrysts are rarer than types 1 and 2
plagioclase, accounting for only ∼5% of the total plagio- Amphibole is a minor ferromagnesian phase in all the Galán
clase population. They are recognized in all ignimbrites ignimbrites, accounting for <1.7% of average vesicle-free
with the exception of the CGI. These crystals remain non- juvenile clast modal proportions (Fig. 5). Where present (i.e.,
zoned (∼An35) for the majority of their crystal width, with a in the Toconquis Group and Cueva Negra ignimbrites),
thin (20–50 μm), calcic-rich (An50-55) rim (Fig. 7c). amphibole crystals are generally fresh and unaltered, but
relatively small; the largest crystals are less than 500 μm in
Biotite diameter but more commonly less than 100 μm. However,
amphibole crystals are occasionally broken-down and altered
Biotite has been shown to be one of the most useful minerals to plagioclase, spinel, and clinopyroxene throughout the
for demonstrating geochemical variations and “fingerprint- entire crystal width. There is no evidence for chemical
ing” of seemingly “monotonous” Andean ignimbrites (de zonation in fresh, unaltered phenocrysts. Amphibole compo-
Silva and Francis 1989). A detailed analysis of biotite sitions include dominant hornblende and edenite, according
phenocrysts in all the Galán ignimbrites was undertaken in to the International Mineralogical Association classification
order to ascertain if different ignimbrites could be geochem- in Deer et al. (1992). Table 5 shows representative amphibole
ically fingerprinted. Biotite is the most common ferromag- analyses with a selection of major and minor element data
nesian phase in the Galán ignimbrites, accounting for 10.6– shown in Fig. 6c. Two populations of amphibole crystals are
14.5% of average vesicle-free juvenile clast modal propor- delineated based on Al2O3 and SiO2 contents; one with
tions (Fig. 5). Rare cases of bent and fractured crystals are <2 Al p.f.u (per formula unit), >6.9 Si p.f.u and the other
present in the Galán samples. Biotite crystals with a similar with >2 Al p.f.u., <6.9 Si p.f.u. (Fig. 6c). There is no
morphology are found in the Fish Canyon Tuff and have correlation with different juvenile clasts types, with both
been shown to originate from Precambrian wall rocks around populations found in grey and white pumice samples. For
the associated magma chamber (on the basis of in situ Sr available analyses, there is no chemical trend or variation
isotopes; Charlier et al. 2007). Determining this for the over time within each population (Fig. 6c).
Galán ignimbrites is beyond the scope of this study and so
these crystals are excluded from analyses. Crystals are up to Fe–Ti oxides
4 mm in diameter in the CGI and Cueva Negra Ignimbrite
and up to 2 mm in diameter in the Toconquis Group Fe–Ti oxides account for between 1.9% and 2.8% of the
ignimbrites. They often contain rare inclusions of plagio- average vesicle-free juvenile clast modal proportions
1468 Bull Volcanol (2011) 73:1455–1486

a Plagioclase
White pumice Grey Pumice
2.0

3.0

4.0
Age (Ma)

4.5

5.0

5.5

6.0
0.25 0.35 0.45 0.55 0.65 0.75 0.25 0.35 0.45 0.55 0.65 0.75 0 100 200 300
An An Ba (ppm)

b Biotite
2.0

3.0

4.0
Age (Ma)

4.5

5.0

5.5

6.0
0.4 0.44 0.48 0.52 0.56 2.3 2.5 2.7 2.9 200 300 400 0 20 40 60 80 100
Ti (pfu) Al (pfu) V (ppm) Nb (ppm)

c 3.0
Amphibole

3.5

4.0
Age (Ma)

4.5

5.0

5.5

6.0

1 2 3 0.2 0.4 0.6 0.8 6.2 6.7 7.2 7.7 0.0 0.2 0.4
Al (pfu) Na (pfu) Si (pfu) Ti (pfu)

d Magnetite e Ilmenite
2.0

3.0
Age (Ma)

4.0

4.5

5.0

5.5

6.0
0 0.01 0.02 0.03 0.04 2.9 3.3 3.7 4.1 0.01 0.02 0.03 0.04 0.7 0.8 0.9
Mn (pfu) Fe (pfu) Mn (pfu) Ti (pfu)
 Bull Volcanol (2011) 73:1455–1486 1469

ƒ Fig. 6 Electron microprobe and LA-ICPMS analyses of various chemical distinction between oxides in the successive
mineral phases in the Galán ignimbrites. Oxide analyses are
recalculated as cations per formula unit (pfu) a Plagioclase phenoc-
ignimbrites. Representative analyses of magnetite and
rysts (formula units calculated on the basis of eight oxygens). b ilmenite crystals from each ignimbrite (Fig. 6d and e) are
Biotite phenocrysts (formula units calculated on the basis of 22 similar, illustrating the homogeneity of this >3.5-Myr
oxygens). c Amphibole phenocrysts (formula units calculated on the eruptive sequence.
basis of 24 oxygens). d Magnetite (formula units calculated on the
basis of four oxygens). e Ilmenite (formula units calculated on the
basis of three oxygens). Note: Only plagioclase An (anorthite)- Glass analyses
contents include rim analyses (cross symbols). All other analyses are
of core and/or core and rim compositions due to LA-ICPMS beam Matrix glasses were analyzed from all ignimbrites and
size. Open boxes and symbols denote white pumice clasts, grey boxes
and black symbols denote grey pumice clasts and striped boxes on
representative analyses are presented in Table 7 and in
amphibole plots denote a mixture of white and grey analyses. White Fig. 3a and b. All glass analyses are rhyolitic (70.5–
boxes and star symbols denote the crystal-rich juvenile clasts in the 80.8 wt.% SiO2) and more evolved than their associated
CGI. For all plots, diamond (and star) symbols represent mean values, bulk-rock compositions (68.5–71 wt% SiO2). All major
the boxes represent the 25% and 75% quartiles and the whiskers
represent the maximum range of data
element glass compositions are homogenous except for grey
pumice in the CGI. This observation must be used with
caution, however, due to the paucity of glass analyses for the
(Fig. 5). This population consists exclusively of magnetite older grey pumice clasts. In general, grey pumice glass
(Usp10–19) and ilmenite (Hem21–25). The average proportion compositions are more variable and less silica-rich (range of
of magnetite varies between ignimbrites at 0.9–1.2 modal 71–78.9 wt.%, average of 73 wt.% SiO2) compared to a
%. As described previously, there is a higher average range of 76.4–80.8 wt.% for all other analyses.
proportion of ilmenite in the CGI pumice (>1.6 modal%)
compared with the older Toconquis Group ignimbrites Crystal-rich juvenile clasts
(<1.2%; Fig. 5). Representative electron microprobe anal-
yses of ilmenite and magnetite are shown in Table 6. Like Crystal-rich (75–90% crystals, vesicle free) juvenile clasts
the other major minerals in the assemblage, there is little form a minor proportion (<1%) of the overall juvenile clast

Fig. 7 The three main types of

plagioclase crystals in each of
the Galán ignimbrites. Repre- a
50
sentative photomicrographs Type 1
(crossed nicols) are shown along
with core-to-rim anorthite (An) 40
An

content profiles of these plagio-

clase types determined by elec- 30
tron microprobe. The transects
are shown as black lines termi- 20
nating at the rim of the crystals. 0 100 200 300 400
Scale bars for photomicrographs
are all 200 microns in length b
50

40
An

30 Type 2
20
0 100 200 300 400 500

c
50 Type 3
40
An

30

20
0 50 100 150 200
Distance from Rim (µm)
 1470

Table 4 Average microprobe and LA-ICPMS analyses of biotite from each ignimbrite

Ignimbrite CGI CGI CGI Cueva Negra Real Grande Real Grande Pitas U. Merihuaca U. Merihuaca M. Merihuaca M. Merihuaca L. Merihuaca Blanco
White Grey XI-rich White Grey White White Grey White Grey Average Average
Total analyses 53 96 39 32 49 20 18 20 25 42 18 38 37

Na2Oa 0.33 0.27 0.33 0.56 0.43 0.34 0.38 0.44 0.56 0.49 0.42 0.32 0.43
MgO 13.3 13.5 12.6 16.7 14.2 13.7 14.1 14.4 15.8 14.7 13.9 13.5 14.1
SiO2 36.6 36.9 36.7 38.0 36.8 37.0 36.3 36.9 37.3 36.5 37.0 37.5 37.4
K2O 9.0 8.9 9.2 9.0 8.9 8.7 9.1 9.0 9.0 8.9 8.9 8.1 7.9
CaO 0.02 0.04 0.01 0.01 0.03 0.11 0.02 0.02 0.00 0.12 0.03 0.24 0.25
TiO2 4.2 4.3 4.2 4.6 4.5 4.1 4.3 4.3 4.4 4.2 4.3 4.4 4.4
MnO 0.24 0.24 0.26 0.13 0.18 0.17 0.22 0.20 0.15 0.15 0.19 0.19 0.13
FeO*b 17.7 16.9 19.1 13.0 16.1 17.2 17.6 16.5 14.4 15.3 17.2 16.8 16.1
Al2O3 14.1 14.4 14.4 14.6 14.3 14.4 14.3 14.4 14.7 14.3 14.6 14.1 14.2
F2O 0.92 n.d. n.d. n.d. 1.15 0.69 1.19 0.86 n.d. n.d. 0.77 n.d. n.d.
Cl2O 0.19 n.d. n.d. 0.19 0.26 0.18 0.20 0.20 n.d. n.d. 0.18 n.d. n.d.
Total 96.63 95.44 96.65 96.69 96.85 96.51 97.75 97.21 96.31 94.72 97.49 95.17 95.00
Mg#c 0.58 0.59 0.54 0.69 0.61 0.62 0.60 0.62 0.66 0.63 0.62 0.59 0.61
Scd 27.0 30.1 43.0 33.4 19.0 19.2 20.6 21.3 33.9 22.3 23.2 32.7 25.7
V 349.0 300.4 382.5 367.9 285.0 366.5 367.5 362.4 350.4 383.4 332.1 327.1 377.4
Cu 15.1 26.2 7.0 17.4 12.1 7.2 5.1 6.6 16.4 18.3 8.2 12.2 4.8
Zn 507.9 414.6 456.4 270.9 458.0 499.7 452.3 465.3 427.2 388.2 487.9 381.3 338.8
Rb 971.6 721.4 827.1 804.6 1,865.6 826.3 749.3 834.2 943.2 747.4 745.5 1,118.7 1,343.9
86
Sr 10.8 22.6 8.1 22.7 36.5 16.5 7.4 8.1 20.1 30.6 12.1 34.8 33.0
88
Sr 9.2 21.3 7.7 21.6 33.7 13.6 4.7 6.2 19.2 29.5 10.3 34.3 32.8
Y 0.54 2.53 0.72 0.33 0.69 0.27 0.11 0.24 0.36 0.63 0.48 1.92 1.01
Zr 2.0 10.4 3.3 7.0 2.7 1.4 2.1 2.4 3.6 3.3 3.0 4.2 5.7
Nb 65.0 59.2 68.5 66.2 43.1 46.8 52.8 53.9 60.0 43.6 49.6 51.5 35.5
Ba 417.9 742.8 778.4 706.2 612.1 466.1 843.3 581.8 603.0 842.8 782.8 934.8 1,092.7
Pb 6.5 22.2 4.2 4.1 13.9 5.2 3.8 5.5 11.4 13.8 9.5 39.7 44.4

n.d. not determined

a
Oxides determined by microprobe analyses (in weight percent)
b
FeO* denotes total iron
c
Mg#=Mg/(Mg+Fe)Molar
d
Minor elements determined by LA-ICPMS
Bull Volcanol (2011) 73:1455–1486
Bull Volcanol (2011) 73:1455–1486 1471

Table 5 Representative microprobe analyses of amphibole from each ignimbrite

Ignimbrite Cueva Cueva Real Pitas Pitas Middle Middle Lower Lower
Negra Negra Grande Merihuaca Merihuaca Merihuaca Merihuaca
Sample CG 520 CG 227a CG 113a CG 499 CG 499 CG 157aa CG 87ab CG 289a CG 289a
Crystal Hbl4 Hbl 1 Hbl2 Hbl 1 Hbl 1 Hbl2 Hbl1 Hbl 2 Hbl 1a
number

SiO2a 44.5 49.4 42.9 44.8 47.2 43.2 43.1 43.8 47.5
TiO2 1.1 0.6 2.5 2.2 1.5 2.6 2.4 2.6 1.1
Al2O3 11.0 7.1 14.0 10.5 8.6 11.5 11.7 12.0 7.5
FeO* 13.6 11.9 12.5 12.0 12.9 12.0 12.9 12.2 14.0
MgO 14.2 16.8 13.3 15.3 15.5 14.2 13.5 14.3 14.3
MnO 0.37 0.34 0.10 0.13 0.27 0.15 0.17 0.10 0.41
CaO 10.9 11.4 10.8 11.4 11.2 12.1 11.0 11.1 11.4
Na2O 1.8 1.3 2.4 1.9 1.6 1.9 1.9 2.0 1.3
K2O 0.96 0.70 1.10 0.91 0.77 0.89 0.98 1.00 0.76
Cl 0.08 0.11 0.14 0.08 0.10 0.07 0.07 0.04 0.07
Total 98.55 99.64 99.90 99.18 99.72 98.69 97.79 99.10 98.32
%Ab 0.60 0.60 0.65 0.60 0.60 0.65 0.65 0.65 0.65
T (°C)b 789.2 751.4 805.4 819.4 794.7 803.7 803.6 811.6 763.0
Alc 1.9 1.2 2.4 1.8 1.5 2.0 2.0 2.1 1.3
P (kbar)d 3.7 1.7 5.2 2.5 1.9 3.7 3.9 3.7 1.9

n.d. not determined

a
Oxides determined by microprobe analyses (in weight percent)
b
Temperatures calculated by hornblende-plagioclase thermometry (Holland and Blundy 1994) with average rim compositions of co-existing
plagioclase phenocrysts (%Ab)
P
c
Total Al cations ð AlÞ is calculated based on 23 oxygen and 13 cations
d
Pressures calculated by Al-in-hornblende barometry using the method of Anderson and Smith (1995)

Table 6 Representative microprobe analyses of magnetite and ilmenite crystals in grey and white pumice from the Galán ignimbrites

Ignimbrite Pitas-white U. Merihuca-grey M. Merihuaca-white M. Merihuaca-grey L. Merihuaca-white Blanco-grey

Sample CG98aa CG98aa CG 483 CG 483 CG87a CG87a CG 481 CG 481 CG 476 CG 481 CG 475g CG 475g
Mineral Mag2 Ilm7 Mag1 Ilm1 Mag5 Ilm1a Mag5 Ilm2 Mag1 Ilm1 Mag8 Ilm1c

SiO2a 0.04 0.02 0.04 -0.01 0.06 0.02 0.07 0.04 0.03 0.05 0.07 0.02
TiO2 5.27 36.03 4.41 37.57 5.91 41.72 5.80 39.19 4.76 39.50 6.22 40.52
Al2O3 1.65 0.07 1.51 0.04 1.72 0.16 1.21 0.06 1.67 0.09 1.90 0.07
FeO* 83.25 52.48 81.47 53.55 80.67 48.60 80.47 50.39 81.64 48.46 80.93 49.59
MnO 0.71 1.44 0.46 0.97 0.62 0.64 0.47 1.29 0.59 1.44 0.35 0.75
MgO 0.80 2.33 0.92 0.94 1.04 1.67 0.41 1.46 0.94 1.33 0.89 1.79
CaO b.d 0.00 0.02 0.01 0.01 0.02 0.02 0.02 0.03 0.04 b.d 0.31
NiO 0.04 b.d n.d. n.d. 0.02 0.01 n.d. n.d. n.d. n.d. n.d n.d.
ZnO 0.15 0.09 n.d. n.d. 0.26 0.02 n.d. n.d. n.d. n.d. n.d n.d.
V2O3 0.42 0.27 n.d. n.d. 0.46 0.25 n.d. n.d. n.d. n.d. n.d n.d.
Total 92.33 92.75 88.83 93.09 90.77 93.09 88.44 92.44 89.67 90.92 90.37 93.07
T (°C)b 780 721 748 774 747 780
fO2(ΔNNO) 1.4 1.5 1.0 1.1 1.5 1.06

FeO* denotes total Fe, n.d not determined

a
Oxides in weight percent
b
Temperatures and oxygen fugacities calculated by Fe–Ti oxide thermometry using the model of Andersen and Lindsley (1988)
1472 Bull Volcanol (2011) 73:1455–1486

Table 7 Average microprobe analyses of matrix glass from the Galán ignimbrites

Ignimbrite CGI Real Grande Pitas U. Merihuaca M. Merihuaca L. Merihuaca

Pumice type White Grey White Grey White Grey White Grey White
No. of analyses 57 23 17 15 13 2 7 4 19

Na2Oa 1.8 2.6 2.3 1.4 2.6 1.9 2.3 1.7 2.5
SiO2 78.4 71.6 77.2 78.9 77.4 77.3 77.7 78.9 77.9
Al2O3 13.4 16.1 12.5 13.0 12.8 13.4 13.1 12.4 12.3
MgO 0.05 0.64 0.11 0.06 0.04 0.14 0.06 0.04 0.08
TiO2 0.10 0.49 0.15 0.15 0.12 0.11 0.15 0.11 0.10
MnO 0.04 0.09 0.03 0.02 0.01 0.05 0.03 0.03 0.03
P2O5 b.d. 0.19 b.d. 0.00 b.d. b.d. b.d. b.d. b.d.
FeO 0.43 1.5 0.75 0.33 0.24 0.64 0.34 0.31 0.68
K2O 4.8 5.0 5.8 5.3 5.9 5.4 5.4 5.7 5.3
CaO 0.84 1.9 0.85 0.74 0.85 0.82 0.94 0.77 0.92
Total Alkalis 6.6 7.6 8.2 6.7 8.5 7.3 7.7 7.4 7.8
Totalb 96.8 96.4 95.5 96.9 96.8 98.0 97.1 96.7 97.5

Analyses performed at the University of Oregon (c.f. Wright et al. 2011 for methods)
b.d. below detection limit
a
Oxides in weight percent
b
Average analytical totals

population found in the Galán ignimbrites. These pumice Temperature and fO2
clasts are more mafic than their grey and white equivalents
with SiO2 ∼64 wt.%, MgO ∼2.2 wt.% and FeO ∼4.9 wt.% To estimate the temperature and fO2 of magmas, composi-
(Table 1 and presented in Fig. 3a and b). These clasts have tions of co-existing Fe–Ti oxides were compared. Oxides
a hypidiomorphic granular texture. The mineral assem- that exhibited signs of exsolution and alteration were
blage of these juvenile clasts is identical to that of the excluded from chemical analyses. Care was taken to select
grey and white pumices except that the crystal-rich clasts euhedral mineral pairs either in close proximity or direct
have a higher average abundance of biotite (32.3 vol.%, contact. Equilibrium compositions were checked using the
compared to 12.0 vol.% in CGI grey and white pumice) criteria of Bacon and Metz (1984) and Bacon and
and plagioclase (37.5 vol.% compared to 19.6 vol.%). Hirschmann (1988). These compositions are presented in
Plagioclase and biotite mineral compositions for these Table 6 and were then used to calculate temperatures and
crystal-rich juvenile clasts are compositionally identical oxygen fugacities according to the solution model of
to those in white pumice clasts (Tables 4 and 6, Ghiorso and Evans (2008). Temperatures for white and grey
respectively). pumice clasts are distinguished within each ignimbrite.
White pumice clasts in the Merihuaca Formation ignimbrites
have mean values between 739°C and 751°C, the Pitas
Magma storage conditions Ignimbrite has a mean of 771±34°C (2σ deviation from the
mean), whereas the Real Grande Ignimbrite has a mean of
To better constrain minor variations between ignimbrite- 761±10°C (Fig. 8a). For those ignimbrites where grey
forming magma chambers at Cerro Galán, we use co- pumice was analyzed alongside white pumice (CGI, Real
existing oxide pairs and hornblende-plagioclase pairs to Grande, Middle Merihuaca), the Fe–Ti oxide temperatures in
determine magmatic temperatures, Al-in-hornblende to grey pumice are on average ∼10–25°C hotter than their
determine crystallization pressures, and the relative white equivalents, although the ranges of calculated temper-
abundance or absence of sanidine (and preliminary atures overlap. Fe–Ti oxides in the CGI pumice (both grey
melt inclusion data) to speculate on magmatic water and white) and in white Cueva Negra samples produce
contents. Only mineral rim compositions were used anomalously low mean temperatures between 650°C and
in model calculations, as these are thought to best 710°C (Fig. 8a). We hypothesize that these low temperature
reflect the magmatic conditions immediately prior to estimates are due to oxide compositions that lie outside the
eruption. calibration range of the Ghiorso and Evans solution model.
Bull Volcanol (2011) 73:1455–1486 1473

Fig. 8 Results of various geo-

thermometers. a Two-oxide ther-
a Fe-Ti oxide thermometry b Hbl-Plag thermometry
Ghiorso & Evans (2008) Andersen & Lindsley (1988) Holland & Blundy (1994)
mometers. The data is split 2.0 2.0
between the solution models of CGI
Andersen and Lindsley (1988)
and Ghiorso and Evans (2008). b
Hornblende-plagioclase thermom- 3.0 3.0

Age (Ma)
eter based on the solution model
of Holland and Blundy (1994). In
Cueva Negra
all plots, open boxes and symbols 4.0 4.0
represent white pumice clasts and
grey boxes and black symbols
represent grey pumice clasts. For Real Grande
Pitas
all plots, diamond symbols repre- 5.0 5.0
sent mean values, the boxes rep- Upper Merihuaca
Middle Merihuaca
resent the 25% and 75% quartiles Lower Merihuaca
Blanco
and the whiskers represent the
6.0
maximum range of data 700 800 700 800 700 800

Temperature (ºC)

Log fO2 values for the CGI and white Cueva Negra samples amphibole crystals; compositions do not greatly vary
are in excess of 1.5 units above the NNO buffer, resulting in between ignimbrites at Ab60–65. The full range of temper-
temperature uncertainties of at least 50°C and as much as atures calculated by this method for all samples is 709–
100°C using the Ghiorso and Evans (2008) model. These 844°C (Table 6 and Fig. 8b), where maximum uncertainty
same oxide pairs in the CGI and Cueva Negra Ignimbrite lie of the technique has been estimated to be ±40°C over the
within the calibration range of the Andersen and Lindsley calibration range of 400–1,000°C and 1–15 kbar (Holland
model (1988; Fig. 8a), yielding mean temperatures between and Blundy 1994). Mean temperatures (with 2σ deviation
790±8 and 817±6°C (respectively). We therefore view the from the mean) are as follows; Lower Merihuaca—781±39
unrealistically low temperatures for the CGI and white (white), Middle Merihuaca—803±10 (white) and 815±24
Cueva Negra samples generated by the Ghiorso and Evans (grey), Pitas—802 ± 34 (grey), Real Grande—799 ± 56
(2008) model with caution. Oxides in the Toconquis Group (grey) and Cueva Negra—784±42°C (grey). Amphibole
ignimbrites are consistently ∼30°C hotter using the Andersen in the Middle Merihuaca Ignimbrite was analyzed in both
and Lindsley (1988) model compared with the Ghiorso and grey and white pumice clasts. The grey pumice returned a
Evans (2008) model, therefore the Andersen and Lindsley mean temperature 10–15°C hotter than the white pumice
temperatures for the CGI and Cueva Negra Ignimbrite are (Fig. 8b), although the two populations are within error of
considered maxima. Log fO2 values for all ignimbrites using each other. Therefore, based on two-oxide and hornblende-
the Ghiorso and Evans (2008) solution model show a plagioclase thermometry, the preferred crystallization tem-
relatively restricted range, from 1.0 to 1.7 units above the peratures for the white pumice clasts are 790–820°C for the
NNO buffer, similar to oxygen fugacities from the La Pacana Toconquis Group ignimbrites, 790–800°C for the Cueva
ignimbrites (Lindsay et al. 2001). Negra Ignimbrite and a maximum of 790°C for the CGI.
The hornblende-plagioclase geothermometer (Holland The uncertainties in using these geothermometers for the
and Blundy 1994) was applied to the Lower and Middle Galán ignimbrites are 15–40°C (Putirka 2008). The
Merihuaca, Pitas, Real Grande, and Cueva Negra ignim- difference in temperatures between the grey and white
brites containing amphibole phenocrysts of adequate size pumice clasts are similar or less than these error estimates
with little alteration. Only two white pumice samples (CG using the above solution models, but are consistently hotter
87a—Middle Merihuaca and CG 289—Lower Merihuaca) than white pumice temperatures by 10–20°C. We believe
contain amphibole that conformed to these conditions. This this may represent an appreciable difference and propose
geothermometer is pressure dependent, therefore temper- the grey pumice clasts within each ignimbrite could
atures were calculated iteratively, using “starting” pressures represent crystallization from slightly hotter magmas.
calculated from Al-in-hornblende geobarometry (based on
the calibration of (Schmidt 1992; maximum uncertainties Pressure
are considered to be ±0.6 kbar). We use the second
calibration of Holland and Blundy (1994); edenite+albite Matrix glass compositions can be used to calculate the
= richterite+anorthite, as suggested by Andersen (1996). pressure conditions at which the glass equilibrated. Tuttle
Plagioclase compositions were taken from plagioclase and Bowen (1958) first showed that the typical evolution of
crystals in direct contact or within close proximity to the matrix glass to SiO2-rich (>77% anhydrous) compositions
1474 Bull Volcanol (2011) 73:1455–1486

requires crystallization under very low pressures (<1 kbar). Depth (km)
4 8 12 16 20
Unfortunately, the glass compositions of the Galán ignim- 3.0
brites do not allow the direct use of the haplogranite quartz-
albite-orthoclase (Qz–Ab–Or) ternary system developed by
3.5
Johannes and Holtz (1996) and Blundy and Cashman
(2001), due to elevated normative corundum compositions Cueva Negra
4.0

Age (Ma)
affecting the position of the eutectic point (all compositions
contain >3 wt.% normative corundum whereas the calibra-
tion range requires <1 wt.%). We can qualitatively surmise, 4.5
Real Grande
based on the findings of Tuttle and Bowen (1958), that the
Pitas
high-Si glass present in all Galán ignimbrites is an 5.0
indication that the final stages of crystallization of these
magmas occurred at pressures <1 kbar. 5.5 Middle Merihuaca
Lower Merihuaca
The only mineral barometer applicable to the assemb-
lages in these ignimbrites is Al-in-hornblende (e.g., 1.0 2.0 3.0 4.0 5.0

Johnson and Rutherford 1989a; Anderson and Smith Pressure (kbar)

1995). Amphibole phenocrysts are present in sufficient Fig. 9 Pre-eruptive pressure estimates of the magmas responsible for
sizes to use this barometer from the Lower and Middle Galán Ignimbrite eruptions. Pressures are estimated from the Al-in-
Merihuaca, Pitas, Real Grande, and Cueva Negra ignim- hornblende barometer (Anderson and Smith 1995). In the box and
brites. Although amphibole is present in the Blanco and whisker plots, open boxes represent white pumice clasts, grey boxes
represent grey pumice clasts, and striped boxes represent a mixture of
Upper Merihuaca ignimbrites, the few samples collected white and grey analyses. For all plots, diamond symbols represent
only contain crystals altered by post-depositional processes. mean values, the boxes represent the 25% and 75% quartiles and the
The method of Anderson and Smith (1995) was employed whiskers represent the maximum range of data
by using 53 amphibole rim compositions (either in close
proximity or directly touching plagioclase crystals) from Viscosity and density
eight samples. Iterative calculations were carried out using
the temperatures obtained from hornblende-plagioclase Estimates of magma viscosity and density can be performed
thermometry (Holland and Blundy 1994; see above). using various model calculations based on the temperature,
Several ignimbrites (Lower Merihuaca, Pitas, and Cueva water, and bulk-rock compositions. Melt densities were
Negra) contain distinct clusters of amphibole crystallization calculated according to the method of Spera (2000), taking
pressures in both white and grey pumice clasts (Fig. 9). The into account pressure, temperature, and water contents. All
population with lower pressures lies at a roughly similar ignimbrites have melt densities of 2,300–2,320 kg/m3 (at
depth range (within errors), from an average of 1.96± 4 wt.% H2O, 3 kbar, and 800°C). Crystal-free anhydrous
0.34 kbar for the Lower Merihuaca Ignimbrite to 1.77± magma densities are 2,450 kg/m3 for all ignimbrites. Melt
0.64 kbar for the Cueva Negra Ignimbrite (errors are 2σ viscosities were calculated according to Giordano et al.
deviation from the mean). This corresponds to a crustal (2008), returning values of 105.9–107.2 Pa s for all
depth of 7.4 and 6.7 km, respectively (assuming a crustal ignimbrites (at 3–4 wt.% H2O, 800°C). Taking into account
density of 2,700 kg/m3 and lithostatic pressure=magma a crystallinity of 35–55 vol.%, magma (bulk) viscosities
chamber pressure). The higher pressure amphibole popula- range from 106.9 to 109.9 Pa s (Spera 2000).
tion is present in the Lower Merihuaca (average of 3.82±
0.76 kbar), Middle Merihuaca (3.75±0.76 kbar) and Pitas Summary and interpretation of the mineral chemistry
(2.98±0.62 kbar) ignimbrites. This corresponds to a crustal
depth of 11.3–14.4 km. A deeper crystallization level is The available mineral chemistry data show that each
inferred for the Real Grande Ignimbrite (average of 4.68± ignimbrite contains plagioclase, biotite, amphibole (except
1.04 kbar; 17.7±3.9 km depth). The shallower depth the CGI), and Fe–Ti oxides with nearly identical composi-
population agrees with estimates from other large caldera tions. Distinguishing between the ignimbrites based on
systems such as Long Valley, Yellowstone, and La Pacana geochemical fingerprinting (e.g., de Silva and Francis
for which pressures of 2–3 kbar have been calculated 1989) is not possible because of the large degrees of
(Hildreth 1981; Christiansen 2001; Lindsay et al. 2001), overlap in nearly all elements, except for elevated bulk-rock
placing the tops of their magmatic systems in the upper concentrations of Nb, Ta, Y, and lower Dy/Yb ratios in the
8 km of the earth's crust. Amphibole crystals returning CGI. This geochemical homogeneity also transcends
higher pressures are inferred to have crystallized from magma volumes, including the smallest volume Blanco
magmas at deeper levels in the crust (up to ∼18 km depth). and Lower Merihuaca ignimbrites (<10 km3 DRE) and the
Bull Volcanol (2011) 73:1455–1486 1475

largest volume CGI (∼630 km3 DRE; Folkes et al. 2011). (Toconquis Group) to sanidine-phyric (CGI) magmas. This
This suggests each crystallizing assemblage originated from change may result from a decrease in water content causing
magmas nearly identical in composition and pre-eruptive the sanidine liquidus to shift to higher pressures and
histories. That is, the same upper crustal magma accumu- temperatures (e.g., Naney 1983) and/or a decrease in
lation and crystallization processes occurred to produce temperature and pressure of the magma chambers. For
these magmas over the entire >3.5-Myr history of ignim- example, in the phase diagram for the dacitic Fish Canyon
brite generation. Tuff, sanidine defines a negative Clapeyron curve such that
Several lines of evidence (textural, compositional, and it will begin to crystallize as pressure and temperature
mineral thermobarometry), show there to be two distinct decrease (Johnson and Rutherford 1989b). Preliminary data
magma batches represented in each of the Galán ignim- from melt inclusions trapped within quartz crystals show no
brites. Shallow magmas resided in near-surface chambers systematic difference in water contents across the temporal
(770–800°C, 6–10 km depth), prior to their eruption. A history of the Galán ignimbrites (all ignimbrites generally
second deeper (14–18 km depth) and potentially hotter contain a range of 2–5 wt.% H2O; Wright et al. 2011; H.
(800–840°C) magma is implied to exist by a second Wright, unpublished data). We therefore interpret the
population of phenocrysts. These pumice clasts have change in mineralogy to indicate evolution to a shallower
elevated Cu, Pb, and Zn concentrations (Wright et al. magmatic system, before commencement of activity asso-
2011) and less-evolved glass compositions. We propose ciated with the modern caldera-forming episode (∼2.80 Ma;
these clasts record an intermediate storage depth for the Folkes et al. 2011).
grey magmas “parental” to the Galán ignimbrites that
ascended repeatedly to the shallow chambers prior to the Geochemical homogeneity
eruption of each ignimbrite.
The geochemical evidence presented in this study shows
that the Cerro Galán system generated similar rhyodacite
Discussion magmas repeatedly over a >3.5-Myr history. We have built
upon the existing work of Harmon (1981) and Francis et al.
A two-tiered plumbing and storage system is the generally (1989) to show that the geochemical homogeneity of these
accepted architecture for large silicic magma systems and is ignimbrites extends to minor, trace, and rare-earth elements
consistent with geophysical and petrologic constraints. Many in individual minerals such as biotite, plagioclase, Fe–Ti
workers have shown that shallow upper crustal magma oxides and hornblende, as well as glass analyses in all
chambers must have been present prior to large eruptions ignimbrites. Different eruption volumes and repose periods
and caldera-forming episodes (Smith 1979; Hildreth 1981; between eruptions do not appear to affect the overall
Christiansen 1983; Grunder and Boden 1987; Lipman et al. geochemical homogeneity. Volumes calculated for all the
1997; Lindsay et al. 2001; Bachmann et al. 2002). At Galán, Galán ignimbrites erupted since 6 Ma (Fig. 10) show the
evidence for large shallow chambers comes from amphibole first ignimbrites erupted from the Toconquis Group were
barometry (with crystallization occurring at ∼1.5–2 kbar, ∼6– small volume (total combined volume of ∼70 km3 DRE)
7.5 km depth) and is also implied by the high crystallinity and erupted over <200 kyr (see Folkes et al. 2011). The
(up to 55%) of the pumice clasts requiring protracted final three ignimbrites (Real Grande, Cueva Negra, and
residence at low pressures. These shallow chambers are CGI) have a much larger combined volume (>1,000 km3
represented by the abundant white pumice clasts. DRE), each separated by ∼1 Myr of quiescence. Therefore,
In the APVC, these shallow chambers are thought to be the observed chemical homogeneity was repeated in
fed from a deeper chamber where a remnant signal has been successive ignimbrite eruptions, regardless of the volume
imaged by geophysics at depths of 17–19 km (Chmielowski or length of time separating the eruptions.
et al. 1999). Wright et al. (2011) propose that the deeper, The chemical homogeneity of the Galán magmatic
grey pumice magma in the CGI provides evidence of a system through time raises an interesting question; how
silicic recharge body. Rising magma, potentially “recharge” do these magmas become homogenized and what is the
grey magma from intermediate depths, could have also cause of this homogenization? Both the deeper, potentially
provided the heat and volatiles necessary to trigger eruption hotter grey “parental” magma and the shallower, cooler
from these shallow chambers at Cerro Galán. white pumice magma are nearly identical in petrology and
Evidence from mineral chemistries and proportions show geochemistry. Therefore the chemical homogeneity of the
that the younger ignimbrites erupted from the Galán system Galán magmas (as a whole) must have been imparted on
(the Cueva Negra and CGI) crystallized at shallower depths them before they were delivered to upper crustal magma
than the older Toconquis Group ignimbrites. Evidence for chambers. We suggest that there was a chemical “buffer
this comes from the change from amphibole-phyric zone” between the mantle and the upper crust, which
1476 Bull Volcanol (2011) 73:1455–1486

Fig. 10 A cumulative erupted Different fractionating assemblages

volume-time graph (c.f. Folkes
0.9
et al. 2011), along with F values
(proportion of melt remaining)
of each ignimbrite as calculated Yb
from geochemical modeling Dy 0.8
Ta
(Table 8 and see text). The lower Y
set of F values corresponds to a Nb
constant fractionating assem-
0.7
blage in all ignimbrites, whilst

Cumulative erupted volume (DRE, km3)

the upper set of F values corre- Same fractionating assemblages
0.9
sponds to a fractionating assem-

F (Fraction of liquid remaining;

from Rayleigh Fractionation)
blage with a 25% increase in
1200
zircon, ilmenite, and titanite
fractionation in the Toconquis 0.8
Group ignimbrites (see text) Yb
Ta
Dy
800 Nb 0.7
Y Real
Grande
CGI
Cueva
Negra
0.6

400
Pitas
0.5
Merihuaca
Fm
Blanco
0
6 5 4 3 2
Biotite 40Ar/39Ar age (Ma)

maintained a similar body of melt that was tapped rhyodacites after they accumulated. We argue that although
repeatedly to produce the Galán magmas. the Galán magma bodies would have assimilated minor
We propose that this “buffer zone” could be a deep amounts of upper crustal material and experienced recharge
(>35 km) melt zone in the lower crust, similar to the MASH by more primitive magma from below, they essentially
zone (melting, assimilation, storage, and homogenization; evolved by shallow closed-system fractionation (as sug-
Fig. 11) hypothesis of Hildreth and Moorbath (1988). This gested for the Fish Canyon Tuff by Charlier et al. 2007).
MASH zone would impart the underlying isotopic, major, The geochemical homogeneity we observe at Cerro Galán
and minor element signatures (as suggested by Hildreth and is similar to the general lack of chemical variations observed
Moorbath 1988 and Annen et al. 2006 for various Andean in other examples of monotonous intermediates (e.g., Hildreth
volcanic systems) onto the Galán magmas, repeatedly 1981). However, although these large-volume ignimbrites
producing homogeneous melts (from a mixture of mantle- commonly exhibit a general lack of chemical variations,
and crustal-derived melts). These are then delivered to the there are often subtle geochemical differences in minor and
upper crust (Fig. 11) where shallow level crystallization trace elements of bulk rock and/or mineral and glass
produces the minor textural and chemical variations seen in compositions. Because of the lack of variation in all other
the Galán ignimbrite suite prior to eruption. The depth of geochemical data, any subtle, but distinct differences warrant
this MASH zone, determined by thermomechanical factors a detailed analyses and explanation, and explaining these
in the crust (de Silva and Gosnold 2007) and/or delamina- differences becomes integral to understanding the evolution
tion of the lower crust (e.g., Kay et al. 1999) could explain of large volumes of magma in the upper crust. At Cerro
the chemical variations in the Galán ignimbrites also Galán, we find small variations between different ignimbrites
observed by Kay et al. (2011). in bulk-rock trace element compositions (e.g., Fig. 3g and h).
We therefore assume each ignimbrite eruption from the We now address the origin of these subtle differences in the
Cerro Galán system tapped very similar rhyodacite magmas magma compositions seen in the juvenile clasts.
in the upper crust. While the magmas were most likely fed
from a much deeper system driven by thermal energy from Closed system evolution of the Cerro Galán magmas
mantle-derived basalts and assimilation of deep crust as
suggested by Francis et al. (1989) and also Kay et al. Certain bulk-rock trace elements in juvenile clasts show
(2011), our focus is the upper crustal evolution of these small, but appreciable differences among the ignimbrites.
Bull Volcanol (2011) 73:1455–1486 1477

Depth Satellite
Caldera collapse
Pre-eruption tant role in crystallization in the CGI magma (e.g., KdBa
(km) scoria cone magma chamber
for sanidine is ∼11.5 compared to ∼0.3 in plagioclase;
0 Magma Rollinson 1993). This conclusion is strengthened by higher
composition
5 proportions of sanidine in the CGI than the older Toconquis
White pumice
10 magma
Group ignimbrites (>0.9% as opposed to <0.3% of the total
Rhyodacite
phenocryst population; Fig. 5), perhaps coinciding with the
Grey pumice
20 magma onset of crystallization of this mineral phase.
Andesite
The variations in Nb and Ta can also be explained by the
30 km differences in mineral proportions between the ignimbrites.
MASH Basaltic- Nb and Ta concentrations in magmas are extremely
zone andesite sensitive to the crystallization of ilmenite and titanite
50 km (sphene), DNb & Ta >>1 (Rollinson 1993) for these phases.
Moho
Basalt Figure 5 shows that the CGI contains the highest proportion
of these two minerals, which could account for the higher
Basaltic flux from mantle Nb and Ta values seen in the CGI bulk-rock analyses.
(subduction and delamination driven)

Crystal-rich juvenile clasts—evidence for crystal

Fig. 11 A conceptual model of the crust beneath Cerro Galan fractionation
(modified from de Silva et al. 2006). A MASH (melting, assimilation,
storage, and homogenization; Hildreth and Moorbath 1988) zone
exists in the lower crust where basaltic magmas assimilate crust and Rare juvenile crystal-rich clasts comprise <1% of the total
homogenize with time. As the middle crust warms and becomes more juvenile clast population. Given the inferred high crystal
ductile, magmas rise, fractionate, evolve, and pond in larger melt contents of the rhyodacite magmas and the important role
zones (producing the “grey pumice” magmas). More buoyant melts
periodically rise to shallow crustal magma chambers (producing the of crystal fractionation discussed above, the presence of
“white pumice” magmas). These crystallize, cool, and form crystal these clasts may be a valuable link to other parts of the
“mushes”, and are eventually erupted as the crystal-rich ignimbrites. magmatic system that may be less readily erupted. For
Grey pumice magma recharge to the shallow white pumice magma instance, recent work emphasizes crystal-rich monotonous
chambers is also proposed to trigger eruptions (Wright et al. 2011)
ignimbrites sample long-lived “crystal mushes” that may
not be able to erupt due to their high viscosities (Smith
The youngest ignimbrite (the CGI) has high Nb, Ta, Y, and 1979; Hildreth 1981; Bachmann and Bergantz 2004;
Yb values and relatively low middle to heavy REE ratios Christiansen 2005). Melt extraction is proposed to occur
(e.g., Dy/Yb; Figs. 3g, h and 4b). These trace element from sill-like, non-convecting mushes (Bachmann and
concentrations are highly sensitive to variations in the Bergantz 2004; Hildreth 2004). The eruptible part of these
modal proportions of accessory phases. The CGI possesses systems are thought to have a crystal fraction of 45–60 vol.
the highest proportions of zircon crystals (>0.4% compared %, values overlapping with crystallinities of white and grey
to <0.3% for the older Toconquis ignimbrites; see Fig. 5). juvenile clasts in the Galán ignimbrites. In this model,
This would account for the elevated Yb and Y (DYb =527 in crystal-rich juvenile clasts might represent original, wholly
zircon; Rollinson 1993) and lower middle to heavy REE preserved parts of semi-solid crystal mushes that have been
ratios (e.g., Dy/Yb; Fig. 4b) in the CGI. This is consistent incorporated into the erupting material. These clasts have a
with Zr concentrations; the older Toconquis ignimbrites higher proportion of plagioclase (37.5 vesicle-free vol.%)
contain the lowest Zr bulk-rock values compared with the and biotite (32.3 vol.%) compared with the more ubiquitous
CGI (Fig. 3e). The lower proportion of hornblende in the white and grey juvenile clasts in the CGI (19.6 and
CGI samples would also account for the lower Dy/Yb ratios 12.0 vol.%, respectively; Table 2). This may reflect higher
seen in the whole-rock composition, as hornblende prefer- degrees of crystal accumulation, and would also explain the
entially concentrates the middle rare-earth elements (Gd, elevated LREE concentrations in the crystal-rich clasts (as
Tb, Dy) in its crystal structure as most recently shown by the LREE are more compatible in biotite and plagioclase
Davidson et al. (2007). than the middle and heavy REE; Rollinson 1993).
Furthermore, an explanation for the positive correlation The textural and bulk trace element compositions of
of Eu/Eu* with Ba concentrations is the onset of sanidine similar clasts and inclusions found in APVC ignimbrites
into the crystallization assemblage. An increase in sanidine have been interpreted as margin accumulations of crystals
crystallization would lead to larger Eu anomalies (lower Eu/ (de Silva 1989b). The modal increase in biotite and
Eu*) with decreasing Ba concentrations. The CGI also has plagioclase proportions of these clasts is consistent with
the largest Eu-anomalies at corresponding lower Ba con- conditions likely to be obtained at the margins of the
centrations (Fig. 3f), indicating sanidine played an impor- magma chamber (e.g., Marsh 2000). The crystal-rich
1478 Bull Volcanol (2011) 73:1455–1486

juvenile clasts in the Galán ignimbrites contain crystal sizes 32% crystallization of the fractionating assemblage. These
that are intermediate between the white and grey pumice values overlap with those of previous attempts to model
clasts, consistent with their crystallization from the grey aspects of the Galán system (Francis et al. 1989; Mantovani
magma. Although chamber margin cumulates could have and Hawkesworth 1990). F values for the CGI white pumice
formed in the Galán magma chambers, we interpret the clasts range from 0.87 to 0.90, corresponding to 10–13%
crystal-rich juvenile clasts as parts of the semi-solid crystal fractionation. Although this range is slightly higher than that
mush that formed with progressive crystal fractionation of expected from the major element fractionation (up to 8%;
the grey parent magma. To place relative quantitative Fig. 12), we believe that fractionation can explain the major
constraints on the degree of fractionation with respect to and trace element variations between the CGI samples and
the observed chemical variations, we further present can therefore be applied to the older Toconquis Group
simplified models of closed system fractionation, using ignimbrites. Figure 10 shows the relationship between the
both a common set of fractionating phases for all amounts of fractionation needed to produce the overall trace
ignimbrites and varying the fractionating assemblage. element concentrations in each ignimbrite with erupted
volumes over time. It is interesting to note that F values
Geochemical modeling correlate with ignimbrite volumes. For example, F values for
each trace element are highest in the largest volume CGI
We pursue a more quantitative approach to explain the bulk- (∼630 km3) and Real Grande ignimbrite (∼390 km3), in
rock chemical variations in Ta, Y, Nb, Dy, and Yb between contrast to lower F values for the smaller-volume Blanco and
each ignimbrite magma using geochemical modeling. A Merihuaca ignimbrites (<40 km3). This suggests lower
starting (parent composition) was selected based on the above degrees of crystal fractionation took place in the larger
hypothesis that grey pumice clasts most closely represent the volume ignimbrites. However, this modeling assumed the
composition of the magmas originally emplaced into the fractionating assemblage contained the same modal propor-
upper crust. We have selected sample CG 122 (from the CGI; tions of minerals in all the Galán magmas. We have shown
Table 1) with the most elevated Cu, Pb, and Zn concen- that the bulk-rock variations in Nb, Ta, Y, and Yb can be
trations (see Wright et al. 2011). Variations from the starting explained by different zircon, ilmenite, and titanite propor-
composition were modeled to test whether the chemical tions in the final crystallization assemblages of the ignim-
variability in the above trace elements is consistent with the brites. Assuming a common “parental” grey magma
modal mineralogy of the suite. We have used the fractional composition for all ignimbrites, the higher modal proportions
crystallization model of Ersoy and Helvaci (2010) to see if of these minerals in the CGI and Cueva Negra ignimbrite
the CGI white pumice major element (bulk rock) composi- suggest that lower amounts of fractionation of these phases
tions can be produced from this starting composition. The took place in these magmas in their early evolution in the
fractionating assemblage was taken from the average modal upper crust. Conversely, greater amounts of crystal fraction-
proportions of minerals in the CGI crystal-rich juvenile ation of these mineral phases in the older Toconquis Group
clasts. Figure 12 shows the results of this modeling using a ignimbrites could explain their lower bulk-rock Nb, Ta, Y,
selection of major element oxides. In all cases, the white and Yb concentrations.
CGI compositions can be produced by fractionating the We did not take any samples of crystal-rich juvenile clasts
selected grey CGI parent composition by up to 8%. from the Toconquis Group ignimbrites. We therefore simulate
The starting composition (CG 122) was then used to the effect of increased fractionation of these mineral phases in
model the selected trace element abundances (those that the modeling. Modal proportions of magnetite, zircon,
showed variations between ignimbrites: Ta, Y, Nb, Dy, and ilmenite, apatite, and titanite were increased by 25% in the
Yb), in each of the ignimbrites by Rayleigh fractionation. fractionating assemblage for the Toconquis Group ignim-
Bulk partition coefficients (D) were calculated by multi- brites. Altering the fractionating assemblage by this amount
plying the modal proportions of minerals in the crystal-rich does not change the overall trend of increasing F values with
juvenile clasts (the fractionating assemblage) by the increasing ignimbrite volume (Fig. 10, Table 8). We therefore
partition coefficient (KD) of each trace element for that conclude that bulk-rock compositions of the larger volume
specific mineral. KD values were selected from the literature ignimbrites can be explained by lower amounts of fraction-
(see Table 8 for references). The amount of melt remaining ation from the parental magma emplaced in the upper crust
(F) was used to indicate the amount of fractionation compared with the smaller-volume ignimbrites.
required (1−F) to produce the observed trace element
concentrations in each ignimbrite (daughter compositions). Open-system processes—recharge
The results of the trace element modeling are presented in
Table 8 and Fig. 10. F values range from 0.68 to 0.90, In many volcanic deposits, there is evidence for the input of
indicating fractionation of the parental composition by 10– hotter, less-evolved magma acting as the driving force or
Bull Volcanol (2011) 73:1455–1486 1479

1.60 5.0

4.5

1.40
4.0

Na2O (wt%)
MgO (wt%)

CGI Grey CGI Grey 6% 8%

4%
(CG 122) 2% 4% 6% 8%
(CG 122) 2% 3.5
1.20
3.0

1.00 2.5

2.0

17.0 6.0
Al2O3 (wt%)

CGI Grey

K2O (wt%)
16.0 CGI Grey (CG 122) 2%
5.0
4% 6% 8%
(CG 122)
2% 4% 6% 8%

15.0 4.0

14.0 3.0

13.0 2.0
65.0 66.0 67.0 68.0 69.0 70.0 71.0 72.0 66.0 67.0 68.0 69.0 70.0 71.0 72.0

SiO2 (wt%) SiO2 (wt%)

Fig. 12 Major element fractionation diagrams showing the amount of crystallization (in percent) required to produce the CGI white pumice clasts
from the selected CGI grey “parent” (CG 122). Maximum 2σ error bars are shown for the bulk-rock analyses

trigger for the eruption of near-surface magma chambers. rysts by Boyce and Hervig (2008) could have been
This can consist of a wide variety of indicators, such as generated by the input of the proto-grey pumice magma
mafic enclaves or glomerocrysts within the erupted magma, in the CGI, which also contributed to triggering the
phenocryst rims showing a markedly different composition eruption (see Wright et al. 2011).
to their interiors, or compositionally zoned deposits The degassing and recharge event documented by Boyce
showing a progression to more primitive compositions as and Hervig (2008) in the CGI apatites has not yet been
eruption ensues and empties the magma chamber (e.g., investigated in the older Toconquis Group ignimbrites.
Bacon and Druitt 1988; Wolff et al. 1990). Recent work has However, we suggest that a similar recharge event in the
shown the importance of silicic magma recharge in older Toconquis Group may be represented as reversely
providing the heat and volatiles necessary to trigger zoned plagioclase crystals (type 3), with thin (<50 μm)
eruption in continental arc settings (e.g., de Silva et al. anorthite-rich (An50–55) rims (Fig. 6c). Reversely zoned
2008; Smith et al. 2009), but evidence can be cryptic, plagioclase phenocrysts were first suggested to indicate
involving subtle textural and chemical variations. recharge events (e.g., MacDonald and Katsura 1965), and
Boyce and Hervig (2008) have shown that OH and Cl although only a limited crystal population was investigated
growth zonation in apatite phenocrysts from the CGI record in this study, we suggest that this mineral might be
a multistage magmatic history just prior to eruption. Based recording a similar recharge process to that documented
on apatite growth rates, they attribute these zoning profiles by Boyce and Hervig (2008). Using experimental studies of
to H2O degassing and recharge <400 days prior to eruption. plagioclase rim growth (e.g., Larsen 2005), the 10–50-μm
These conclusions are consistent with our data suggesting rims would have taken 10–60 days to form, in accordance
that the grey pumice in the Galán ignimbrites indicates with the recharge timescales suggested by zoning in apatite
recharge by deeper and potentially hotter magma (and in phenocrysts. However, it must be noted that plagioclase
the case of the CGI may have provided an additional composition is not only a function of the melt composition
volatile input; Wright et al. 2011). We suggest that the from which it is crystallizing, but is also sensitive to a wide
degassing and recharge event recorded in apatite phenoc- variety of other properties including magmatic temperature,
Table 8 Closed-system fractional crystallization modeling parameters and results
1480

Fractionating assemblage volume percentage (from crystal-rich juvenile clasts in CGI)

Qtz Plag Biotite Kspar Magnetite Ilmenite Hbl Cpx Apatite Zircon Titn

12.4 42.7 34.1 2.6 1.3 2.0 0.0 0.1 3.2 0.6 0.7
Theoretical fractionating assemblage for Toconquis Group ignimbrites
Qtz Plag Biotite Kspar Magnetite Ilmenite Hbl Cpx Apatite Zircon Titn
10.4 42.7 34.1 2.6 1.6 2.5 0.0 0.1 4.0 0.8 0.9
KD valuesa
Qtz Plag Biotite Kspar Magnetite Ilmenite Hbl Cpx Apatite Zircon Titn
Ta 0.01 0.04 1.57 0.01 0.00 106.00 0.00 0.75 0.00 47.50 16.50
Y 0.00 0.13 1.23 0.00 2.00 0.00 6.00 3.10 40.00 0.00 0.00
Nb 0.00 0.06 6.37 0.00 2.50 0.00 4.00 0.80 0.10 0.00 6.30
Dy 0.02 0.11 1.72 0.06 0.00 4.90 13.00 7.30 50.70 101.50 19.00
Yb 0.02 0.09 1.47 0.03 0.00 4.10 8.38 6.37 23.90 527 10.00
Bulk D’s Parent (ppm)b Daughters (ppm-average from bulk-rock analyses of white pumice clasts)
CG 122 Blanco Lower Merihuaca Middle Merihuaca Upper Merihuaca Pitas Real Grande Cueva Negra CGI
3.07 3.10 1.70 1.60 1.60 1.80 1.70 2.00 1.70 2.30
1.79 21.00 15.50 15.70 15.80 16.50 16.70 17.50 17.00 19.40
2.28 23.00 14.90 15.30 15.20 15.80 16.70 18.70 16.60 20.10
3.11 5.40 2.90 2.80 2.70 3.10 3.40 3.60 2.90 4.20
4.63 2.50 0.90 0.90 0.90 1.00 1.10 1.20 1.00 1.60
F (proportion of melt remaining) with same fractionating assemblage in all ignimbrites
Blanco Lower Merihuaca Middle Merihuaca Upper Merihuaca Pitas Real Grande Cueva Negra CGI
Ta 0.75 0.73 0.73 0.77 0.75 0.81 0.75 0.87
Y 0.68 0.69 0.70 0.74 0.75 0.79 0.76 0.90
Nb 0.71 0.73 0.72 0.75 0.78 0.85 0.77 0.90
Dy 0.74 0.73 0.72 0.77 0.80 0.82 0.74 0.89
Yb 0.75 0.75 0.75 0.78 0.80 0.82 0.78 0.88
F (proportion of melt remaining) with theoretical fractionating assemblage
Blanco Lower Merihuaca Middle Merihuaca Upper Merihuaca Pitas Real Grande Cueva Negra CGI
Ta 0.80 0.78 0.78 0.82 0.80 0.85 0.80 0.90
Y 0.76 0.77 0.77 0.80 0.81 0.85 0.83 0.93
Nb 0.72 0.73 0.73 0.75 0.78 0.85 0.78 0.90
Dy 0.80 0.79 0.78 0.82 0.84 0.86 0.80 0.91
Yb 0.80 0.80 0.80 0.82 0.84 0.85 0.82 0.91

a
KD values from Rollinson (1993)
b
The parental composition for the models is CG 122
Bull Volcanol (2011) 73:1455–1486
Bull Volcanol (2011) 73:1455–1486 1481

volatile contents, and crystallization depth (Housh and Luhr rysts. Although we argue that each ignimbrite magma in the
1991; Singer et al. 1995; Tepley III et al. 2000; Blundy et Galán system originated from a non-zoned magma cham-
al. 2006). Nonetheless, the indication of recharge between ber, variations in crystal fractionation are proposed to have
these two approaches is intriguing. produced the minor geochemical variations between the
At present, reversely zoned plagioclase phenocrysts have Galán ignimbrites.
not been categorically documented in the CGI. Two
possible explanations are proposed. The CGI came from The influence of magma chamber geometry on crystal
the largest volume of magma, and so the limited population fractionation
of plagioclase crystals analyzed in this study may simply
not have sampled any of the rare type 3 (reversely zoned) Many studies have examined the role of chamber morphol-
crystals. However, we have observed small microphenoc- ogy on crystal fractionation and settling (de Silva and Wolff
rysts and microlites (<60 μm) in the CGI white pumice that 1995; Miller and Smith 1999; Maughan et al. 2002;
contain bright (high-An) rims in BSE (back-scatter elec- Christiansen 2005). We agree with de Silva and Wolff
tron) images that could be recording such a recharge event. (1995) that in large-volume systems, crystal fractionation
Alternatively, the recharge event recorded in the apatite will be dominated by sidewall crystallization and that
phenocrysts by Boyce and Hervig (2008) may not have fractional crystallization is the determining factor in magma
been as clearly recorded in the plagioclase crystals analyzed differentiation. The degree of chemical homogeneity in
in this study. This could be because apatite is more sensitive monotonous intermediates may be related to magma
than plagioclase to changes in volatile components, chamber geometry or aspect ratio. We use aspect ratio here
particularly with respect to Cl, OH, and F (Brenan 1993; as a ratio of the height to width of the chamber. Low aspect
Boyce and Hervig 2008). ratios are thought to inhibit extensive sidewall crystalliza-
tion, with most crystallization occurring within the well-
Comparison with other large silicic systems mixed magma reservoir (de Silva 1991; de Silva and Wolff
1995), perhaps dominated by “crystal-mushes” (Bachmann
There are few detailed geochemical studies of deposits from and Bergantz 2004; Hildreth 2004). It has been theorized
the repeated eruption of crystal-rich monotonous intermedi- that larger volume ignimbrites originate from magma
ates over a protracted time period. Perhaps the best chambers with low aspect ratios that produce sill-like
documented example of a large crystal-rich silicic system morphologies (e.g., de Silva 1991; de Silva and Wolff
is the Fish Canyon Tuff (∼5,000 km3) erupted from the La 1995; Miller and Smith 1999); equilibration pressures of
Garita caldera in SW Colorado, USA ∼27.8 Ma (Hildreth H2O- and CO2-rich melt inclusions support this theory
1981; Whitney and Stormer Jr 1985; Riciputi et al. 1995; (Schmitt 2001). Furthermore, modeling by Blundy and
Lipman et al. 1997; Bachmann et al. 2002). These studies Cashman (2008) shows that sill-like geometries will limit
conclude that there is chemical homogeneity in magma vertical variations in crystallinity because of a limited
composition at a chamber-wide scale. However, discrete but variation in equilibrium H2O. Magma produced from
important chemical and textural differences exist at the reservoirs of this shape would thereby become composi-
centimeter to millimeter scale within individual crystals. tionally homogeneous. This is in contrast to smaller-
These variations include resorbed quartz, oscillatory and volume, upright bottle-shaped magma chambers that expe-
reverse zoning in amphibole crystals and complex zoning rience more sidewall crystallization and therefore higher
profiles in plagioclase phenocrysts (Bachmann et al. 2002; degrees of crystal fractionation and differentiation (de Silva
2005). 1991; Maughan et al. 2002). Whereas this process
The large-volume, crystal-rich Lund Tuff in the Great undoubtedly also occurs in large sill-like chambers, it does
Basin of Nevada and Utah, USA (Maughan et al. 2002) so to a lesser proportion due to their relatively small vertical
lacks systematic compositional zoning and contains a extent (de Silva and Wolff 1995). This variation in magma
virtually identical mineral assemblage throughout the chamber geometry could help explain the small-scale
deposit. There are small-scale variations in the Lund Tuff, chemical variations described above for other monotonous
including minor and trace element bulk-rock compositions large-volume ignimbrite deposits.
and modal proportions of crystallizing phases, similar to the The variations in fractionation as highlighted by our
Galán ignimbrites. These authors postulate these subtle geochemical modeling may also be explained by variations
chemical variations result from origins in multiple magma in magma chamber geometry. The largest volume eruptions
chambers, with the same bulk composition but slightly in the Galán system (the CGI and Real Grande ignimbrites
differing degrees of crystallization. There is some evidence with volumes >300 km3) are likely to have originated from
for similar processes in the Galán ignimbrites, including flat sill-like magma chambers. In contrast, the smaller-
multiple zoning patterns in plagioclase and apatite phenoc- volume ignimbrite eruptions (Blanco, the Merihuaca For-
1482 Bull Volcanol (2011) 73:1455–1486

mation, and Cueva Negra ignimbrites with volumes been generated in the case of the other large-volume,
<50 km3) may have come from magma chambers with crystal-rich rhyodacitic magma systems (e.g., Huber et al.
higher aspect ratios (e.g., de Silva and Wolff 1995). This 2009).
would have produced a greater relative proportion of
sidewall crystallization, leading to higher amounts of
crystal fractionation (lower F values), as suggested by the Conclusions
geochemical modeling (Fig. 10). However, the smaller-
volume ignimbrites are still largely compositionally homo- The results of this geochemical study of the ignimbrites of
geneous. Therefore high degrees of crystal fractionation the Cerro Galán caldera system yield the following
producing highly differentiated magmas with effective conclusions:
separation of crystals and melt could not have occurred,
which resulted in non-zoned, well-mixed magma chambers, 1. The bulk-rock geochemistries of juvenile clasts be-
consistent with the crystal-mush model (Bachmann and tween and within each ignimbrite are very composi-
Bergantz 2004; Hildreth 2004). tionally similar, confirming these rhyodacites as highly
In contrast, magmas associated with the La Pacana homogeneous, in agreement with previous geochemical
caldera (northern Chile) show evidence of a single chamber work on the Galán magmatic system (e.g., Francis et al.
being segregated into two distinct crystallinity regions 1989). This study has shown that chemical homogene-
(Schmitt et al. 1999; Lindsay et al. 2001). La Pacana is a ity extends to the mineral scale. A variety of minerals
caldera complex in the APVC of the Central Andes common to all ignimbrites exhibit very little chemical
(Fig. 1a) that erupted a series of large-volume felsic variation over the >3.5 Myr time period encompassed
ignimbrites with similar compositions and volumes to the by these eruptions. We find no evidence to support the
Galán ignimbrites from 5.6 to 2.5 Ma (the Pujsa Ignimbrite assertion of Francis et al. (1989) that the CGI was
—5.6±0.2 Ma, >500 km3; the Toconao Ignimbrite—4.5 to derived from a zoned magma chamber. We suggest that
4.0 Ma, >100 km3; the Atana Ignimbrite—3.96±0.02 Ma, the parental magmas feeding these ignimbrites were
2,200 km3; and the Pampa Chamaca/Talabre ignimbrites— generated under near-identical conditions, undergoing
2.52±0.06, 100 km3; de Silva and Gosnold 2007 and repeated processes before their eruption at the Earth's
references therein). The large Atana Ignimbrite shares many surface. A lower-crustal melt or MASH region is
similarities with the Galán ignimbrites and is geochemically proposed to exist beneath Cerro Galán acting as a
homogeneous throughout its entire deposited thickness chemical “buffer” zone that generates the underlying
(average outflow thickness of 30–40 m; Lindsay et al. chemical signatures before minor chemical variations
2001). However, the magma chamber responsible for the are applied in upper crustal chambers where final
production of the Atana Ignimbrite had an associated, accumulation of the homogenized magmas take place.
more-evolved cap that produced the geochemically distinct, 2. There are two main juvenile clasts (white and grey
smaller-volume cogenetic crystal-poor Toconao Ignimbrite pumice) present within the Galán ignimbrites. We
with a volatile-rich, rhyolitic composition (Lindsay et al. extend the conclusions of Wright et al. (2011) that grey
2001). Products from a similar, evolved crystal-poor cap are pumice clasts in the CGI originated from magmas
not observed in the Galán system or the other examples of inferred to be parental to the more voluminous white
crystal-rich ignimbrites described above. This difference pumice magmas, to all of the Galán ignimbrites. The
could be a function of repose periods between eruptions. If grey pumice magma is speculated to originate from a
an ignimbrite eruption is triggered before there is sufficient slightly deeper and potentially hotter source than the
time to fractionate a rhyolitic cap, e.g., due to a “pulse” in white pumice magma, and is inferred to be an
higher magmatic-power input, only crystal-rich, monoto- intermediate-depth (up to 18 km) magma body that
nous dacites will erupt (Lipman 2007). Lindsay et al. was repeatedly ‘tapped’ to produce the more volumi-
(2001) also proposed the La Pacana system had an unusual nous white magma that crystallized in shallow-depth
magma chamber geometry, with a prominent cupola above (6–10 km) chambers. Rare examples of crystal-rich
the large sill-like magma chamber into which the residual juvenile clasts in all ignimbrites provide evidence for
melt of the crystal-rich dacite segregated and was pre- the incorporation of material from the growing crystal
served, eventually generating the Toconao ignimbrite. mush portions of the Galán magma chambers.
Perhaps this feature was absent from Cerro Galán and the 3. Mineral barometry shows that the Galán ignimbrites
other examples cited above, precluding the occurrence of a contain a population of crystals that equilibrated and
Toconao-like magma body. Alternatively, a late reheating crystallized from magmas in the upper 10 km of the
and defrosting event could trigger whole-chamber convec- crust (<3 kbar). We disagree with the hypothesis of
tion, mixing back any crystal-poor volatile-rich cap that had Francis et al. (1989) that upper crustal processes did not
Bull Volcanol (2011) 73:1455–1486 1483

play a dominant role in the Cerro Galán system. We these large silicic, crystal-rich systems still remains conten-
argue that each magma was delivered repeatedly to tious. We propose that the underlying chemical signatures
shallow crustal levels (represented by the white pumice (and therefore the chemical homogeneity) are imparted on
clasts) experiencing crystallization prior to eruption. these magmas before they are delivered to the upper crust.
This agrees with the protracted residence times of Shallow level crystal fractionation processes serve only to
crystal mushes in the upper crust (Bachmann and modify the magma chemistries to a minor degree. A long-
Bergantz 2004) and other depth estimates of large- lived lower-crustal melt body (similar to a MASH zone
volume silicic systems that show evidence for shallow hypothesized by Hildreth and Moorbath 1988) is proposed
magma chambers beneath large calderas. to act as a chemical buffer zone between more mafic
4. Subtle variations in bulk-rock trace element proportions magma addition from below and subsequent shallow crustal
(Ta, Y, Nb, Dy, and Yb) between ignimbrites are magma chambers where only minor geochemical and
reconciled with small differences in the amount of petrological variations take place.
crystal fractionation calculated by geochemical model- We suggest that any subtle geochemical variations
ing. Each ignimbrite magma can be generated from a observed in other large-volume ignimbrites (‘monotonous
grey ‘parent’ magma originally emplaced in the upper intermediates’; Hildreth 1981), with a general pattern of
crust. Using these trace elements, the most voluminous restricted chemical variability, should be investigated and
CGI requires higher F values (amount of melt remain- explained in detail as this will be integral to any meaningful
ing) than smaller-volume Toconquis Group ignimbrites, discussion on the magmatic evolution of these large-volume
suggesting more crystal fractionation took place in the systems. In particular, other examples of crystal-rich,
smaller-volume ignimbrites during their upper-crustal monotonous intermediates with evidence for spatially
evolution. Although each ignimbrite would have restricted, repeated eruptions should be investigated more
erupted from highly viscous and homogenized crystal thoroughly with respect to their geochemical evolution. For
mushes, the smaller-volume ignimbrites would have example, recent work by Lipman (2007) has shown a
come from magma chambers with lower aspect ratios, number of large, silicic crystal-rich intermediates were
experiencing slightly greater amounts of sidewall erupted between ∼30 and 27 Ma in the San Juan volcanic
crystallization and therefore crystal fractionation. field of Colorado. It will be interesting to ascertain whether
5. The grey pumice magma provides evidence for open- the remarkable degree of homogeneity seen in the Galán
system processes throughout the history of the Galán ignimbrites is a feature also seen in the San Juan volcanic
system and these clasts have been proposed to act as system and other large silicic systems around the world
recharge magma, as suggested for the CGI by Wright et with a protracted history of ignimbrite generation.
al. (2011). The presence of reversely zoned (OH and
Cl) apatite phenocrysts in the CGI (Boyce and Hervig Acknowledgments This research was funded by an Australian
2008), and plagioclase phenocrysts with thin An-rich Research Council Discovery Program Grant DP0663560 to the
research team led by R. Cas. Partial support for this work came from
rims in many juvenile clasts from the Galán ignimbrites
NSF grant EAR 0710545 to S. de Silva that is gratefully acknowl-
suggest significant degassing and/or recharge events edged. We thank Monash University, Oregon State University and
<400 days before eruption. The grey pumice magma Salta University, Argentina for access to the various facilities required
could be reflecting recharge events of slightly deeper, to undertake this research. Journal reviews from Olivier Bachmann
and Eric Christiansen and suggestions from the editors for this special
more volatile-rich magma, perhaps acting as an
issue helped to improve this manuscript.
eruption trigger.

The Cerro Galán caldera provides an excellent opportu- References

nity to examine the repeated eruption of large-volume
ignimbrites over a protracted time period. The results from
Allmendinger RW, Jordan TE, Kay SM, Isacks BL (1997) The
this study have important implications for large caldera- evolution of the Altiplano-Puna plateau of the Central Andes.
forming ignimbrite eruptions in general. We have shown Annu Rev Earth Planet Sci 25:139–174
that the Cerro Galán system has repeatedly erupted Anders E, Ebihara M (1982) Solar system abundances of the elements.
Geochim Cosmochim Acta 46:2363–2380
geochemically near-identical magmas over >3.5 Myr. This
Andersen JL (1996) Status of thermobarometry in granitic batholiths.
is all the more fascinating because these magmas are Trans R Soc Edin 87:125–138
thought to have originated at depths in excess of 35 km, Andersen JL, Lindsley DH (1988) Internally consistent solution
interacting with crustal material as they ascended (see Kay models for the Fe-Mg-Mn-Ti oxides, Fe-Ti oxides. Am Mineral
73:714–726
et al. 2011). This geochemical homogeneity is present
Anderson JL, Smith DR (1995) The effects of temperature and
regardless of volumes or repose period between each fO2 on the Al-in-hornblende barometer. Am Mineral 80:549–
eruption. The cause and location of the homogeneity of 559
1484 Bull Volcanol (2011) 73:1455–1486

Annen C, Blundy JD, Sparks RSJ (2006) The genesis of intermediate de Silva SL (1991) Styles of zoning in central Andean ignimbrites;
and silicic magmas in deep crustal Hot zones. J Petrol 47(3):505– insights into magma chamber processes. In: Harmon SR, Rapela
539 CW (eds) Andean Magmatism and its Tectonic Setting. Geolog-
Bachmann O, Bergantz GW (2004) On the origin of crystal-poor ical Society of America Special Paper, pp 217–232
rhyolites: extracted from batholithic crystal mushes. J Petrol de Silva SL, Francis PW (1989) Correlation of large ignimbrites; two
45:1565–1582 case studies from the Central Andes of northern Chile. J Volcanol
Bachmann O, Dungan MA, Lipman PW (2002) The Fish Canyon Geotherm Res 37:133–149
magma body, San Juan volcanic field, Colorado: rejuvenation de Silva SL, Gosnold WD (2007) Episodic construction of batholiths:
and eruption of an upper-crustal batholith. J Petrol 43:1469– insights from the spatiotemporal development of an ignimbrite
1503 flare-up. J Volcanol Geotherm Res 167:320–335
Bachmann O, Dungan MA, Bussy F (2005) Insights into shallow de Silva SL, Wolff JA (1995) Zoned magma chambers: the influence
magmatic processes in large silicic magma bodies: the trace of magma chamber geometry on sidewall convective fraction-
element record in the Fish Canyon magma body, Colorado. ation. J Volcanol Geotherm Res 65:111–118
Contrib Mineral Petrol 149:338–349 de Silva SL, Zandt G, Trumbull R, Viramonte JG, Salas G, Jimenez N
Bacon CR, Druitt TH (1988) Compositional evolution of the zoned, (2006) Large ignimbrite eruptions and volcano-tectonic depres-
calcalkaline magma chamber of Mount Mazama, Crater Lake, sions in the Central Andes: a thermomechanical perspective. In:
Oregon. Contrib Mineral Petrol 98:224–256 Troise C, de Natale G, Kilburn CRJ (eds) Mechanisms of activity
Bacon CR, Hirschmann MM (1988) Mg/Mn partitioning as a test for and unrest at large caldera. Geological Society, London, Special
equilibrium between coexisting Fe-Ti oxides. Am Mineral 73:57– Publications, pp 47–63
61 de Silva SL, Salas G, Schubring S (2008) Triggering explosive
Bacon CR, Metz J (1984) Magmatic inclusions in rhyolites, eruptions—The case for silicic magma recharge at Huaynaputina,
contaminated basalts, and compositional zonation beneath the southern Peru. Geology 36:387–390
Cosa volcanic field, California. Contrib Mineral Petrol 85:346– Deer WA, Howie RA, Zussman J (1992) An introduction to the rock-
365 forming minerals. 2nd edition. Prentice Hall, p 696
Blundy JD, Cashman K (2001) Ascent-driven crystallisation of dacite Elston W (1984) Subduction of young oceanic lithosphere and
magmas at Mount St. Helens, 1980–1986. Contrib Mineral Petrol extensional orogeny in southwestern North America during
140:631–650 mid-Tertiary time. Tectonics 3:229–250
Blundy J, Cashman K (2008) Petrologic reconstruction of magmatic Ersoy Y, Helvaci C (2010) FC-AFC-FCA and mixing modeler: A
system variables and processes. Rev Mineral Geochem 69:179– Microsoft Excel spreadsheet program for modeling geochemical
239 differentiation of magma by crystal fractionation, crustal assim-
Blundy JD, Cashman K, Humphreys M (2006) Magma heating by ilation and mixing. Comput Geosci 36:383–390
decompression-driven crystallization beneath andesite volcanoes. Folkes CB, Wright HM, Cas RAF, de Silva SL, Lesti C, Viramonte JG
Nature 443:76–80 (2011) A re-appraisal of the stratigraphy and volcanology of the
Tuttle OF, Bowen NL (1958) Origin of granite in the light of Cerro Galán volcanic system, NW Argentina. In:Cas RAF,
experimental studies in the system NaAlSi3O8-KAlSi3O8-SiO2- Cashman K (eds) The Cerro Galan Ignimbrite and Caldera:
H2O. Geol Soc Am Mem 74:153 characteristics and origins of a very large volume ignimbrite and
Boyce JW, Hervig RL (2008) Magmatic degassing histories from its magma system. Bull Volcanol doi:10.1007/s00445-011-0459-y
apatite volatile stratigraphy. Geology 36(1):63–66 Francis PW, O'Callaghan LJ, Kretschmar GA, Thorpe RS, Sparks RSJ,
Brenan J (1993) Kinetics of fluorine, chlorine, and hydroxyl exchange Page RN, de Barrio RE, Gillou G, Gonzalez OE (1983) The
in fluorapatite. Chem Geol 110:195–210 Cerro Galan ignimbrite. Nature 301:51–53
Charlier BLA, Bachmann O, Davidson JP, Dungan MA, Morgan DJ Francis PW, Sparks RSJ, Hawkesworth CJ, Thorpe RS, Pyle DM, Tait
(2007) the upper crustal evolution of a large silicic magma body: SR, Mantovani MS, McDermott F (1989) Petrology and
evidence from crystal-scale Rb-Sr isotopic heterogeneities in the geochemistry of volcanic rocks of the Cerro Galan caldera,
Fish Canyon Magmatic System, Colorado. J Petrol 48(10):1875– northwest Argentina. Geol Mag 126(5):515–547
1894 Friedman J, Heiken G (1977) Report in: Skylab explores the Earth. In:
Chesner CA (1998) Petrogenesis of the Toba Tuffs, Sumatra, N.A.S.A. Publication 380: pp. 137–170
Indonesia. J Petrol 39(3):397–438 Ghiorso MS, Evans BW (2008) Thermodynamics of rhombohedral
Chmielowski J, Zandt G, Haberland C (1999) The central Andean oxide solid solutions and a revision of the Fe-Ti two-oxide
Altiplano-Puna magmatic body. Geophys Res Lett 26:783–786 geothermometer and oxygen barometer. Am J Sci 308:957–
Christiansen RL (1983) Yellowstone magmatic evolution; Its bearing 1039
on understanding large volume explosive volcanism. In: Boyd Giordano D, Russell JK, Dingwell DB (2008) Viscosity of magmatic
FR (ed) Explosive volcanism. National Academy of Sciences, liquids: a model. Earth Planet Sci Lett 271:123–134
U.S, pp 84–95 Grunder AL, Boden DR (1987) Comment on '…Magmatic
Christiansen RL (2001) The Quaternary and Pliocene Yellowstone Conditions of the Fish Canyon Tuff, Central San Juan Volcanic
Plateau volcanic field of Wyoming, Idaho, and Montana. US Field, Colorado' by Whitney and Stormer, 1985. J Petrol
Geol Surv Prof Pap 729-G, 145 p., 3 plates, scale 1:125,000 28:737–746
Christiansen EH (2005) Contrasting processes in silicic magma Harmon RS (1981) Petrogenesis of Andean andesites from combined
chambers: evidence from very large volume ignimbrites. Geol Sr-O isotope relationships. Nature 290:396–399
Mag 142(6):669–681 Heit B (2005) Teleseismic tomographic images of the Central Andes at
Davidson J, Turner S, Handley H, Macpherson C, Dosseto A (2007) 21° S and 25.5° S: an inside look at the Altiplano and Puna
Amphibole "sponge" in arc crust? Geology 35(9):787–790 Plateaus. In: Freie Universitat. Berlin, p 137
de Silva SL (1989a) Altiplano-Puna volcanic complex of the Central Hildreth W (1981) Gradients in silicic magma chambers: implications
Andes. Geology 17:1102–1106 for lithospheric magmatism. J Geophys Res 86:10153–10192
de Silva SL (1989b) The origin and significance of crystal rich Hildreth W (2004) Volcanological perspectives on Long Valley,
inclusions in pumices from two Chilean ignimbrites. Geol Mag Mammoth Mountain, and Mono Craters: several contiguous but
126(2):159–175 discrete systems. J Volcanol Geotherm Res 136:169–198
Bull Volcanol (2011) 73:1455–1486 1485

Hildreth W, Moorbath S (1988) Crustal contributions to arc magma- Lipman PW (1984) The roots of ash flow calderas in western north
tism in the Andes of Central Chile. Contrib Mineral Petrol America: windows into the tops of granitic batholiths. J Geophys
98:455–489 Res 89(B10):8801–8841
Hildreth W, Halliday AN, Christiansen RL (1991) Isotopic and Lipman PW (2007) Incremental assembly and prolonged consolida-
chemical evidence concerning the genesis and contamination of tion of Cordilleran magma chambers: evidence from the Southern
basaltic and rhyolitic magma beneath the Yellowstone Plateau Rocky Mountain volcanic field. Geosph 3:42–70
volcanic field. J Petrol 31(1):63–138 Lipman PW, Dungan MA, Bachmann O (1997) Comagmatic grano-
Holland T, Blundy J (1994) Non-ideal interactions in calcic amphib- phyric granite in the Fish Canyon Tuff, Colorado; implications
oles and their bearing on amphibole-plagioclase thermometry. for magma-chamber processes during a large ash-flow eruption.
Contrib Mineral Petrol 116:433–447 Geology 25:915–918
Housh TB, Luhr JF (1991) Plagioclase-melt equilibria in hydrous MacDonald GA, Katsura T (1965) Eruption of Lassen Peak, Cascade
systems. Am Mineral 76:477–492 Range, California in 1915: example of mixed magmas. Geol Soc
Huber C, Bachmann O, Manga M (2009) Homogenization processes Am Bull 76:475–482
in silicic magma chambers by stirring and mushification (latent Mantovani MSM, Hawkesworth CJ (1990) An inversion approach to
heat buffering). Earth Planet Sci Lett 283:38–47 assimilation and fractional crystallisation processes. Contrib
Isacks BL (1988) Uplift of the central Andean plateau and bending of Mineral Petrol 105:289–302
the Bolivian orocline. J Geophys Res 7614:3325–3346 Marsh BD (2000) Magma chambers. In: Sigurdsson H, Houghton BF,
Jellinek AM, DePaolo DJ (2003) A model for the origin of large McNutt SR, Rymer H, Stix J (eds) Encyclopedia of volcanoes.
silicic magma chambers: precursors of caldera-forming eruptions. Academic, San Diego, California, pp 191–206
Bull Volcanol 65:363–381 Mason BG, Pyle DM, Oppenheimer C (2004) The size and frequency
Johannes W, Holtz F (1996) Petrogenesis and experimental petrology of the largest explosive eruptions on Earth. Bull Volcanol 66:735–
of granitic rocks. Springer, Berlin, p 335 748
Johnson MC, Rutherford MJ (1989a) Experimental calibration of Maughan LL, Christiansen EH, Best MG, Gromme CS, Deino AL,
the aluminium-in-hornblende geobarometer with application Tingey DG (2002) The Oligocene Lund Tuff, Great Basin, USA:
to Long Valley Caldera (California) volcanic rocks. Geology a very large volume monotonous intermediate. J Volcanol
17:837–841 Geotherm Res 113:129–157
Johnson MC, Rutherford MJ (1989b) Experimentally determined Miller DS, Smith RB (1999) P and S velocity of the Yellowstone
conditions in the Fish Canyon Tuff, Colorado, Magma Chamber. volcanic field from local earthquake and controlled-source
J Petrol 30(3):711–737 tomography. J Geophys Res 104:15105–15121
Johnson DM, Hooper PR, Conrey RM (1999) XRF analysis of rocks Naney MT (1983) Phase equilibria of rock-forming ferromagnesian
and minerals for major and trace elements on a single low silicates in granitic systems. Am J Sci 283:993–1033
dilution Li-tetraborate fused bead. Advances in X-ray Analysis Pouchou JL, Pichoir F (1985) ‘PAP’ Procedure for improved
41:843–867 quantitative microanalysis. Microbeam Analysis Proceedings.
Jordan TE, Isacks BL, Allmendinger RW, Brewer JA, Ramos VA, In: Armstrong JT (ed) Microbeam Analysis. San Francisco Press,
Ando CJ (1983) Andean tectonics related to geometry of the pp 104–106
subducted Nazca plate. Geol Soc Am Bull 94:341–361 Putirka KD (2008) Thermometers and barometers for volcanic
Kay SM, Mpodozis C, Coira B (1999) Neogene magmatism, systems. Rev Mineral Geochem 69:61–120
tectonism, and mineral deposits of the central Andes (22° to Rampino M, Self S (1992) Volcanic winter and accelerated glaciation
33°S latitude). In: Skinner BJ, Holland R (eds) Geology and Ore following the Toba super-eruption. Nature 359:50–52
Deposits of the Central Andes. Society of Economic Geologists Riciputi LR, Johnson CM, Sawyer DA, Lipman PW (1995) Crustal
Special Publications, pp 27–59 and magmatic evolution in a large multicyclic caldera complex:
Kay SM, Coira B, Wörner G, Kay RW, Singer BS (2011) isotopic evidence from the central San Juan volcanic field. J
Geochemical, isotopic and single crystal 40Ar/39Ar age Volcanol Geotherm Res 67:1–28
constraints on the evolution of the Cerro Galán ignimbrites. Rollinson H (1993) Using geochemical data. Longman Group,
In:Cas RAF, Cashman K (eds) The Cerro Galan Ignimbrite Harlow, pp 1–352
and Caldera: characteristics and origins of a very large Schmidt MW (1992) Amphibole composition in tonalite as a function
volume ignimbrite and its magma system. Bull Volcanol of pressure: an experimental calibration of the Al-in-hornblende
doi:10.1007/s00445-010-0410-7 barometer. Contrib Mineral Petrol 110:304–310
Kent AJR, Stolper EM, Francis D, Woodhead J, Frei R, Eiler J (2004) Schmitt AK (2001) Gas-saturated crystallization and degassing in
Mantle heterogeneity during the formation of the North Atlantic large-volume, crystal-rich dacitic magmas from the
Igneous Province: Constraints from trace element and Sr-Nd- Altiplano-Puna, northern Chile. J Geophys Res 108(B12):
Os-O isotope systematics of Baffin Island picrites. Geochem, 30561–30578
Geophys, Geosystems 5(11):1–26 Schmitt AK, Lindsay JM, Emmermann R (1999) Pre-eruptive magma
Knaack C, Cornelius SB, Hooper PR (1994) Trace element analyses of storage conditions and evidence for externally controlled eruption
rocks and minerals by ICP-MS. Technical Notes for WSU of large-volume central Andean ignimbrites. EOS Trans Am
GeoAnalytical Lab Geophys Union 80:F982–F983
Larsen JF (2005) Experimental study of plagioclase rim growth Schurr B, Asch G, Rietbrock A, Trumbull R, Haberland C (2003)
around anorthite seed crystals in rhyodacitic melt. Am Mineral Complex patterns of fluid and melt transport in the central
90:417–427 Andean subduction zone revealed by attenuation tomography.
Le Maitre RW (1989) A classification of igneous rocks and glossary of Earth Planet Sci Lett 215:105–119
terms. Blackwell, Oxford Singer BS, Dungan MA, Layne GD (1995) Textures and Sr, Ba, Mg,
Lindsay JM, Schmitt AK, Trumbull RB, de Silva SL, Siebel W, Fe, K, and Ti compositional profiles in volcanic plagioclases:
Emmermann R (2001) Magmatic evolution of the La Pacana clues to the dynamics of calc-alkaline magma chambers. Am
Caldera System, Central Andes, Chile: compositional variation of Mineral 80:776–798
two cogenetic, large-volume felsic ignimbrites. J Petrol 42 Smith RL (1979) Ash-flow magmatism. Geol Soc Am Spec Paper
(3):459–486 180:5–27
1486 Bull Volcanol (2011) 73:1455–1486

Smith VC, Blundy JD, Arce JL (2009) A temporal record of Wolff JA (1985) The effect of explosive eruption processes on
magma accumulation and evolution beneath Nevado de geochemical patterns within pyroclastic deposits. J Volcanol
Toluc, Mexico, preserved in plagioclase phenocrysts. J Petrol Geotherm Res 26:189–201
50:405–426 Wolff JA, Worner G, Blake S (1990) Gradients in physical
Sparks RSJ, Francis PW, Hamer RD, Pankhurst RJ, O'Callaghan LO, parameters in zoned felsic magma bodies: implications for
Thorpe RS, Page RN (1985) Ignimbrites of the Cerro Galan evolution and eruptive withdrawal. J Volcanol Geotherm Res
caldera, NW Argentina. J Volcanol Geotherm Res 24:205–248 43:37–55
Spera FJ (2000) Physical properties of magma. In: Sigurdsson H, Wright HMN, Folkes CB, Cas RAF, Cashman KV (2011) Heteroge-
Houghton BF, McNutt SR, Rymer H, Stix J (eds) Encyclopedia neous pumice populations in the 2.08 Ma Cerro Galán ignim-
of volcanoes. Academic, San Diego, California, pp 171–190 brite: implications for magma recharge and ascent preceding a
Tepley FJ III, Davidson JP, Tilling RI, Arth JG (2000) Magma mixing, large volume silicic eruption. In:Cas RAF, Cashman K (eds) The
recharge and eruption histories recorded in plagioclase phenoc- Cerro Galan Ignimbrite and Caldera: characteristics and origins
rysts from El Chichon volcano, Mexico. J Petrol 41:1397–1411 of a very large volume ignimbrite and its magma system. Bull
Whitney JA, Stormer JC Jr (1985) Mineralogy, petrology, and Volcanol doi:10.1007/s00445-011-0525-5
magmatic conditions from the Fish Canyon Tuff, central San Zandt G, Velasco AA, Beck SL (1994) Composition and thickness of the
Juan Volcanic Field, Colorado. J Petrol 26:726–762 southern Altiplano crust, Bolivia. Geology 22:1003–1006