See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/326723311

Magmatism in the Central Andes

Article  in  Elements · August 2018

DOI: 10.2138/gselements.14.4.237

CITATIONS READS

55 2,029

3 authors:

Gerhard Wörner Mirian Mamani

Georg-August-Universität Göttingen Universidad Peruana de Ciencias Aplicadas (UPC)
383 PUBLICATIONS   10,290 CITATIONS    43 PUBLICATIONS   738 CITATIONS   

SEE PROFILE SEE PROFILE

Magdalena Blum-Oeste

21 PUBLICATIONS   144 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Reliability and Validityof the Duruoz Hand Index in an Argentinian Population with Scleroderma View project

Geology of the central andes in northern Chile View project

All content following this page was uploaded by Gerhard Wörner on 27 November 2018.

The user has requested enhancement of the downloaded file.

 August 2018
Volume 14, Number 4

ISSN 1811-5209

Central Andes: Mountains,

Magmas, and Minerals
GERHARD WÖRNER, TAYLOR F. SCHILDGEN, and MARTIN REICH, Guest Editors

Elements of an Extreme Land

Topographic Evolution
Magmatism
Volcanism and Flare-ups
Nitrate Deposits and Hyperaridity
Mineral Resources
 Magmatism
in the Central Andes
Gerhard Wörner1, Mirian Mamani1,2, and Magdalena Blum-Oeste1,3

1811-5209/18/0014-0237$2.50  DOI: 10.2138/gselements.14.4.237

A
ctive continental margins are shaped by subduction-related magma- flat-slab subduction, increased
tism, and the Central Andes of South America are a prime example. The plate coupling and, as a conse-
quence, plate shortening, uplift,
Central Andean orogen has evolved over the past 25 My via magmas erosion, and sedimentation.
ascending from the mantle and interacting with increasingly thickened conti- Second, deposition of plateau-
nental crust. This process is reflected in the volumes and compositional varia- forming ignimbrites (Fig. 1), which
represent large volumes of mixed
tions of the magmas that erupt at the surface. These compositional variations mantle- and crust-derived silicic
can be traced in time and space, and, herein, we provide explanations for magmas containing 70 wt% and
their cause and explore the nature of the Central Andes transcrustal magma 78 wt% SiO2 . Third, the ignim-
brites are locally overlain by flat-
systems that feed the iconic stratovolcanoes today.
lying, phenocryst-poor andesite
Keywords : Central Andean magmatism, isotopes, ignimbrite, magma mixing, shield lavas that may indicate
assimilation, magmatic regimes, transcrustal magma systems hotter and dryer parent magmas.
Fourth, the development of the
VOLCANISM IN THE CENTRAL ANDES iconic andesitic and dacitic strato-
AND ITS GEOLOGICAL CONTEXT volcanoes of the Central Andes, which are characterized by
a composition of 55–68 wt% SiO2 (Fig. 2). True rhyolites
Magmas form in subduction zones by partial melting in
(>69 wt% SiO2 ) are exceedingly rare in stratovolcanoes, yet
the mantle wedge in response to the addition of fluids
it is such rocks that dominate the compositional spectrum
from the down-going oceanic lithosphere. In the Central
of the ignimbrites. Modern (<3 Ma) andesitic edifices can
Andes (Fig. 1), subduction has been active since Jurassic
reach >2,000 m in height and many have summit eleva-
times; however, significant shortening of the crust, crustal
tions well over 6,000 m. Ojos del Salado in northern Chile
thickening, and formation of the Altiplano–Puna Plateau
is the world’s highest active volcano at 6,887 m. These
began only at about 35 Ma (Late Eocene), with acceler-
large clustered volcanoes are the products of intracrustal
ated shortening during the last 10 My. Consequently,
magmatic systems that have typical lifetimes from between
mantle-derived magmas must now traverse the thickest
a few 100 ka to several My (e.g. Hora et al. 2007; Walker
crust (>70 km) of any subduction zone on Earth (Beck et
et al. 2013).
al. 1996). Because of the increasingly arid climate on the
western margin of the Central Andes, volcanic edifices and Fields of smaller, monogenetic volcanoes and related
ignimbrite deposits are extremely well-preserved, and their individual lava flows are rare and concentrated in a few
composition and distribution can be studied back in time. regions, e.g. the Andagua Valley, at Negrillar, as well as
Because the chemical and isotopic composition of magmas in the back-arc region. Out of more than 1,500 analysed
are strongly affected by interaction with crustal material samples, the most mafic magma in the Central Andes
during ascent, and because the thickness of the crust has during Holocene times, and the only true basalt lavas,
changed through time, the Central Andes are an excellent were erupted in the Andagua/Huambo monogenetic field
natural laboratory to study the interaction between crustal (Mg# = 65.3; SiO2 = 51.8 wt%) [Mg# = MgO/(FeOt + MgO)
evolution and magma genesis. × 100, molar] with a few occurrences of shoshonites in the
Peruvian back-arc (Mg# = 69.6; SiO2 = 51.6 wt%) (Mamani
The link between tectonic evolution and magmatism
et al. 2010) (Fig. 2).
is conveniently documented by a typical stratigraphic
sequence of deposits throughout the western slope of the At the other end of the compositional spectrum, the
Central Andes (Wörner et al. 2002). Here, we observe four Central Andes boast one of the largest ignimbrite provinces
general events. First, molasse-type sedimentation that on Earth (de Silva and Kay 2018 this issue). Monotonous
started ~35–25 Ma during a magmatic lull, indicating or crystal-rich dacites to rhyolites of Miocene age contain
individual flows of thousands of cubic kilometres. In this
issue, de Silva and Kay (2018 this issue) discuss how these
1 Abt. Geochemie, Geowissenschaftliches Zentrum
“ignimbrite flare-ups” are related to increased mantle input
Universität Göttingen
Goldschmidtstr. 1 and to a zone of anomalously low seismic velocities in
37077 Göttingen, Germany the middle crust of the southern Central Andes (Ward et
E-mail: gwoerne@gwdg.de al. 2014).
2 Instituto Geológico Minero Metalúrgico (INGEMMET) A simple SiO2 wt% histogram (Fig. 2) is instructive and
Av. Canada 1470
San Borja, Lima 41, Peru summarises the major-element characteristics of magmas
erupted in the Central Andes since Miocene times.
3 2° Investing Initiative,
Compositions more mafic than andesite are rare because
Schönhauser Allee 188,
10119 Berlin, Germany such primitive magmas are too dense and will stagnate,

E lements , V ol . 14, pp. 237–244 237 A ugus t 2018

 Distribution of stratovolcanoes (Miocene to Holocene)
Figure 1
and monogenetic volcanic centers (Pliocene to
Holocene) in the Central Andes. Active volcanoes mentioned in the
text are marked in red; the volcanoes of Parinacota (P), Taapaca (T)
and Aucanquilcha (A) are marked in large blue letters. Large-
volume ignimbrites are color-coded according to age (age column
cool, and crystallize during ascent to the surface. This
highlights the effective crustal density filter in processing
mantle-derived magmas through magmatic differentiation
and crustal assimilation in this thick-crust continental
arc. Subsequent assimilation and compositional differen-
tiation leads to magmas of more evolved compositions.
Appropriately, andesites with a range from 55–68 wt%
SiO2 and those formed by differentiation, assimilation,
and mixing in (trans-)crustal magma systems represent the
most abundant magma types. Magmas having 68–72 wt%
SiO2 rarely erupt in the Central Andes, but where they do
they typically form crystal-rich domes (or “tortas”) indica-
tive of high magma viscosities. These domes may represent
the crystal mushes from which more silicic and voluminous
ignimbrites are derived by melt extraction. However, the
maximum of the SiO2 distribution between 64–67 wt%
SiO2 for (older) intrusive rocks falls close to this minimum
in composition for erupted lavas. Further differentiation
and mixing with crustal melts produce large volumes
of silicic magmas that can feed large-volume ignimbrite
eruptions (Fig. 2).
There are three main questions with respect to Andean
magmatism. First, how do magmas form beneath the
Central Andes? Second, how do magmatic trace-element
and isotopic compositions reflect changing conditions of
magma evolution during the past 35 My of Andean orogeny
and increasingly thickened continental crust? Third, what
is the role of the Andean crust in explaining the variation
in isotopic ratios in lavas that are spatially distinct? We
E lements
bottom right); caldera structures are outlined in yellow. The
location of the mid-crustal Altiplano–Puna Magmatic Body (red
dashes) is based on geophysical data (Zandt et al. 2003).
Ignimbrites from the Southern Peruvian Volcanic Complex are
mostly between 20 Ma and 5 Ma. Figure modified from Freymuth et al .
(2015).
address these questions separately for andesite magmas and
for magmas that form large-volume ignimbrites because the
flux of andesites is continuous and evenly distributed in
space whereas ignimbrites erupt from major calderas and
caldera clusters during discrete flare-up episodes (de Silva
and Kay 2018 this issue).
COMPOSITIONAL CHANGES IN TIME
AND SPACE
Andesites through Time
Investigating how the chemical signatures in andesites
change through time is addressed by compiling trace-
element and isotopic data for volcanic rocks erupted over
the last 180 My of active continental margin evolution
(Figs .  3 and 4). Such an analysis shows that the trace-
element signatures of andesites change systematically
through time and suggests that this change is related to
the shortening and thickening of the continental crust.
In igneous geochemistry, certain trace-element ratios
are indicative of a prominent role for particular igneous
minerals in magma genesis. For example, the Sm/Yb and
Sr/Y ratios for garnet can be characteristic because garnet
is known to fractionate middle from heavy rare-earth
elements (HREEs). Thus, these ratios increase during high-
pressure magma evolution. Amphibole fractionation will
decrease the Dy/Yb ratio in andesites, whereas titanite
suppresses this ratio in silicic magmas. At low pressures,
plagioclase and clinopyroxene will be the dominant
238 A ugus t 2018

Distribution of SiO2 in three broad categories of
Figure 2
Central Andean magmas: the intrusive igneous rocks
and the two types of extrusive rocks, the lavas and the ignimbrites.
Note the distinct bimodal SiO2 wt% values between lavas (blue)
and ignimbrites (yellow), which are controlled by density, viscosity,
and the maximum SiO2 of the eutectic compositions. Older intru-
sive rocks (red speckled) have a wide compositional range but their
maximum is between the modes of the lavas and the ignimbrites.
A B
C D
Trace-element signatures (Sr/Y, Dy/Yb, and Sm/Yb),
Figure 3
including versus age, of erupted magmas in the
Central Andes through time, from the Jurassic (200 Ma) to today.
Color code of magma type and its age are given at top of figure.
Abbreviations are as follows: cpx = clinopyroxene; plag = plagio-
clase feldspar. (A) Ratio of Sr/Y plotted with respect to wt% SiO2.
(B) Ratio of Sm/Yb plotted with respect to age (Ma). (C) Ratio of
Dy/Yb plotted with respect to wt% SiO2. (D) Ratio of Dy/Yb
E lements
fractionating phases and Sr/Y ratios will remain low. These
general principles can be used to link trace-element signa-
tures in magmas to the pressure (i.e. depth within the
crust) where magma evolution takes place. In thin-crust
settings, where plagioclase, pyroxene, and amphibole
dominate the equilibrium assemblage in andesites, the
ratios of Sr/Y, Sm/Yb, and Dy/Yb will be low. Increasing
pressure of fractional crystallization and/or assimilation
processes tends to increase Sr/Y, Sm/Yb and Dy/Yb (Kay
et al. 1988, 1999; Coira et al. 1993; Mamani et al. 2010).
Figure 3 shows these three trace-element ratios and their
evolution through time. Figures 3A and 3C clearly indicate
that – with few exceptions – the maximum Sr/Y and Dy/Yb
ratio are found only in intermediate andesites (55–68 wt%
SiO2 ) that are younger than 5 Ma (Pliocene, blue symbols
in Fig. 3). Basalts (<52 wt% SiO2 ) are rare and lack the high
Sr/Y, Dy/Yb, and Sm/Yb ratios indicative of residual garnet.
Thus, a strong garnet signature is found in intermediate
magmas only after the youngest phase of crustal thickening
(<10 Ma) (Fig. 3B). Findings such as these have resulted
in a recent proliferation of publications linking averaged
trace-element ratios to crustal thickness in a quantitative
way (e.g. Farner and Lee 2017; Hu et al. 2017). However, as
Figure 3 demonstrates clearly, low Sm/Yb, Dy/Yb and Sr/Y
ratios occur throughout the Central Andes at any time,
plotted with respect to Sm/Yb. Arrows indicate compositional varia-
tions caused by the distinct preference for certain trace elements in
the different residual mineral phases during fractional crystallization
and/or crustal melting and assimilation. Colors: Quaternary/
Pliocene = blue; Miocene/Oligocene = yellow and orange; Eocene/
Paleocene = red; Cretaceous = green; Jurassic = black. Modified from
Mamani et al. (2010).
239 A ugus t 2018

Pulses of ignimbrite magmatism expressed as
Figure 4
magmatic flux (km3/per N–S km along arc distance
per My) and average shortening rates in the Central Andes (from
Oncken et al. 2006, thick blue line). Note that ignimbrite pulses
migrate in space from north to south and follow phases of intense
shortening (yellow stippled zones) by about 5–2 My. Modified from
Freymuth et al. (2015).
even in the youngest (<1 Ma) andesites erupted through
maximum crustal thickness. Therefore, the application of
any ­calibration between crustal thickness and Sm/Yb or Sr/Y
cannot be used reliably to extract paleocrustal thicknesses
from individual compositional data. The Central Andes are
the best example of low Sm/Yb, La/Yb (not shown) and
Dy/Yb – the hallmarks of high-pressure magmatic evolu-
tion (Hildreth and Moorbath 1988) – occuring at any time,
even after extensive crustal thickening. Even in thick-crust
settings, shallow, low-pressure magmatic evolution can be
a common process. Michelfelder et al. (2013) and Muir et
al. (2014) have shown that the mid-crustal Altiplano–Puna
Magmatic Body (Ward et al. 2014) plays a prominent role
in the assimilation of magmas in this region at relatively
low-pressure, where garnet is not a stable residual phase.
Increasing crustal thickness and increasing degrees of
assimilation are also reflected in increasing modification
of isotopic compositions with time. This observation has
been used by Scott et al. (2018) to constrain the temporal
and spatial variation of crustal thickening and assimilation.
Their results, supporting earlier observations, suggest that
significant crustal thickening occurred as early as 20 Ma
(Early Miocene) and then migrated southwards, in similar
fashion to that suggested for the ignimbrites (see below).
Changing Volumes and Composition
of Ignimbrites
Plateau ignimbrites in southern Peru and northernmost
Chile (14°S to 19°S) have ages between 24 Ma and 14 Ma
(Thouret et al. 2016) and tend to be younger towards the
south. The massive outpouring of ignimbrite-forming
magmas of the Altiplano–Puna Volcanic Complex still
further south between 22°S and 24°S (Guzmán et al.
2014; de Silva and Kay 2018 this issue) occurred mostly
between 10 Ma (Late Miocene) and 5 Ma (lower Pliocene).
E lements
Volume, timing, and composition of ignim-
brite magmatism are directly related to two
main controlling parameters. First, ignim-
brites follow the north-to-south passage of
the Juan Fernández Ridge when the Nazca
Plate changed from flat to steep subduction
angles. Steepening of the slab caused an
influx of hot asthenospheric mantle, decom-
pression melting in the mantle wedge, and
an increased magma flux into the base of
the crust (de Silva and Kay 2018 this issue).
A second major control is exerted by the
change in thickness and thermal regime
of the overlying Andean crust. Significant
compositional differences were found by
Brandmeier and Wörner (2016) between
the older (>14 Ma) large-volume plateau-
forming ignimbrites in northernmost
Chile and southern Peru and the younger
ignimbrites (<10 Ma) of the Altiplano–Puna
Volcanic Complex in the south (de Silva
et al. 1989). Older ignimbrites are less
depleted in HREEs and less radiogenic in
Sr isotopes (<0.708; Freymuth et al. 2015),
indicating lesser crustal contributions
during magma evolution in the thinner
and thermally immature crust. In contrast,
the Altiplano–Puna Volcanic Complex and
northern Puna ignimbrites in the southern-central Andes
are younger and have relatively high 87Sr/ 86 Sr (>0.708).
These geochemical signatures reflect interaction of mafic
magmas with thermally matured, thick crust, providing
ideal conditions for partial melting and assimilation of
crustal rocks into mantle-derived magmas. The mid-crustal
geophysical anomaly of the Altiplano–Puna Magmatic
Body between 21°S and 24°S (Fig. 1) is the present-day
expression of such partial crustal melting. The Altiplano–
Puna Magmatic Body is the source region for ignimbrite
magmas in the Altiplano–Puna Volcanic Complex and for
strongly contaminated andesites, such as those at Uturuncu
Volcano on the Altiplano (Muir et al. 2014).
Overall, the large volume outpourings of mixed crustal and
mantle-derived magmas migrate from north to south, occur
in distinct pulses of two to three million years, and change
in composition with a higher crustal contribution from
deeper sources in younger ignimbrites towards the south
(Fig. 4). Thus, temporal and spatial variations in volume
and composition not only correlate to the N–S passage of
the Juan Fernández Ridge across the entire Central Andes,
but the composition and the amount of crustal melting
involved is also controlled in time and space by crustal
shortening and thickening, with a resulting increase in
average crustal temperatures over the past 26 My.
Spatial Variations in Isotope Compositions:
Distinct Crustal Domains
Modern magmas in the Central Andes represent mixtures
of mantle-derived, fractionated magmas and assimilated
continental crust. The isotopic compositions of Pb, Sr,
Nd (and other radiogenic isotopes) cannot be changed
by melting or fractional crystallisation and, thus, directly
reflect the composition and proportion of magmatic
components. Compared to the large variations in isotope
composition of old continental crust, the isotope composi-
tion of the mantle source below arcs can be safely assumed
to be relatively invariable. Any observed isotopic varia-
tion should, therefore, reflect the addition of crustal
materials either to the magma source via subduction
erosion (Stern 1991) or through assimilation of conti-
nental crust. Mantle source contamination by subduction
240 A ugus t 2018
erosion prior to melting will only have a small effect on
magmatic O-isotope compositions, whereas the effect on
Sr isotopes will be strong. By contrast, crustal assimila-
tion of magmas will affect both O- and Sr-isotope ratios.
Therefore, the strongest argument against contamination
of the mantle source as the main process to explain elevated
Sr-isotope values is the elevated O-isotope ratios of magmas
throughout the Central Andes. Combined O-, Sr-, Nd- and
Pb-isotope measurements indicate that modern andesite
magmas (<1 Ma) in the Central Andes have assimilated up
to 20% of old continental crust during their long ascent
from mantle source to surface. This degree of assimilation
has since been confirmed by many studies and seems to
be a robust value for typical Andean andesites from the
active volcanic front. Increased crustal contributions are
observed in andesites from stratovolcanoes that overly the
Altiplano–Puna Magmatic Body (e.g. Uturuncu Volcano)
(Michelfelder et al. 2013; Muir et al. 2014).
Lead isotopes are the most sensitive tracer for crustal assim-
ilation because of the strong leverage that crustal rocks
have on assimilated magmas. Accordingly, magmas will
tend to acquire the Pb-isotope composition of the crust
through which they ascend even for small amounts of
assimilation. This observation led Mamani et al. (2010) to
map out crustal domains of different Pb-isotope composi-
tions in erupted magmas and rocks from the metamorphic
basement (Fig. 5). The Arequipa Domain stands out because
of its unusually nonradiogenic 206Pb/204Pb ratios, gener-
ally below 18.5, indicating long-term depletion in U/Pb
after widespread Grenvillian (Mesoproterozoic) granulite
metamorphism. Not only do metamorphic rocks from the
crustal basement share the same isotopic characteristics as
Spatial distribution of Pb-isotopes in Central Andean
Figure 5
magmatic rocks. The Arequipa Domain (red shading)
stands out as a region with unradiogenic Pb isotope compositions
(i.e. those falling below 18.5 206Pb/204Pb).
E lements
the andesites, but so too do ore deposits, locally derived
sediments, and soils. Even llama teeth carry the Pb isotopic
signature of the local crust.
TRANSCRUSTAL MAGMA SYSTEMS
IN THE CENTRAL ANDES
Magmas from the Earth’s mantle tend to stall, accumulate,
cool, and differentiate within the lower crust until their
density is reduced, which in turn later facilitates further
ascent higher into the crust. If the magma supply from
the mantle is sufficiently high, a system of interconnected
conduits and storage reservoirs can develop at intermediate
and shallow levels in the crust where magma accumula-
tion, crystallisation, and assimilation can take place. Such
a transcrustal magmatic system produces mixed magmas
formed at various stages and depths and containing an
older crystal cargo derived from previous intrusions.
Applying the concept of a transcrustal magma system
to the andesitic volcanoes of the Central Andes raises
several questions. What are the compositions of the
different magmas involved and where do these magmas
form, stagnate, differentiate, mix or mingle before they
erupt as andesites? Answers to these questions can be
found in the petrography and geochemistry of lavas from
individual stratovolcanoes. The stratovolcanoes of Taapaca,
Parinacota, and Aucanquilcha in northern Chile provide
excellent examples.
Taapaca Volcano (18°06’S/69°30’W) (Fig. 1) is a dacitic
dome complex (35 km3) located on the western edge of the
Altiplano with its present summit at 5,850 m. During the
past 1.5 My, Taapaca has generated only crystal-rich dacites
of uniform composition. Abundant mafic enclaves and the
frequent presence of sanidine megacrysts indicate mixing
and mingling between mafic andesites and sanidine-
bearing rhyodacites.
Parinacota Volcano (18°09’S/69°08’W) (Fig. 1) is a youthful
volcano, large in size (50 km3), and with a much more
complex evolution than Taapaca. Its conical edifice
(6,350 m) was built in ~9 ky by the eruption of uniform
andesitic lavas over the glaciated remnant of an older
cone. This older cone, built between 160 ka and 40 ka,
was destroyed after a catastrophic sector collapse (Wörner
et al. 1988; Hora et al. 2007). The old-cone lavas constantly
changed between basaltic andesite to rhyolite, whereas the
younger cone comprises only uniform mafic andesites.
Significant changes in lava composition and eruption rates
are, thus, linked temporally to edifice collapse.
Aucanquilcha Volcano (6,175 m; 21°13’12’S, 68°28’6’W) is
a massive stratovolcano that forms part of a slightly W–E
elongated chain of stratovolcanoes (12 km × 10 km) known
as the Aucanquilcha Cluster. Aucanquilcha itself is similar
in size and composition to Taapaca. The Aucanquilcha
Cluster is a group of long-lived andesitic to dacitic strato-
volcanoes that span more than 10 My of volcanic activity.
Considering the cluster as an evolving magmatic system,
Walker et al. (2013) proposed it to be the surface expression
of a slowly evolving plutonic body at depth. Early products
(10–5 Ma) show low eruption rates and are dominated by
andesites. Increased eruption rates after 4 Ma produced a
range of lava compositions from mafic andesite to dacite,
whereas the central Aucanquilcha stratovolcano (>1 Ma)
has only erupted porphyritic dacites with abundant mafic
inclusions (e.g. Walker et al. 2014), similar to Taapaca
Volcano.
241 A ugus t 2018
 Magma Mixing and the Composition
of End-members
Petrographic evidence suggests that the intermediate to
silicic magmas of the Central Andes form by mixing of
(at least) two contrasting mafic a nd s ilicic e nd-member
magmas. This evidence includes abundant inclusions of
more mafic magma that show variable degrees of disintegra-
tion into the more SiO2 -rich dacitic host. On a microscopic
scale, phenocrysts are often reversely zoned, indicating
resorption and regrowth from melts of more mafic compo-
sition following mixing. These observations indicate that
most, if not all, andesites and dacites are hybrid magmas
of which mafic inclusions represent mafic recharge into
a resident, more silicic, magma reservoir. The analysis of
such mafic i nclusions, a nd o f r are m afic la vas, pr ovides
evidence that the mafic end-member in mixing was compo-
sitionally highly variable. Based on these observations,
A of the crust that inhibits eruption of the basalts, plus the
B magmas are evolved melts of andesite to dacitic composi-
Three end-member compositions of Central Andean
Figure 6
magmas as defined in a Sr versus silica graph.
(A) First, basaltic enclaves in the hybrid dacites of Taapaca Volcano
(Chile) (light-blue diamonds) show highly variable concentration in
Sr at similar SiO2 wt% values. Tie lines between individual mafic
enclaves (dark blue diamonds) with their host dacite (the dark blue
square “deep crustal rhyodacite”) from Taapaca define a common
silicic end-member of rhyodacite composition. Second, a calk-
alkaline basaltic andesite end-member (purple triangle) is similar to
melt inclusions in olivine (yellow circles) from Llaima Volcano
(Chile) in the Southern Volcanic Zone (Bouvet de Maisonneuve et
al. 2013). Third, there is the mafic end-member magmas of a
Sr-(and incompatible element) enriched “shoshonitic” basalt (green
circle). These three end-members encompass (almost) the entire
compositional variation of magmas erupted in the Central Andes
(background grey symbols). (B) Chondrite-normalized rare-earth
element patterns highlight the distinct trace-element signatures of
the three end-members: enriched shoshonitic basalt, calc-alkaline
basaltic andesite, and deep crustal rhyodacite. Figures after
B lum -O este and Wörner (2016).
E lements
Blum-Oeste and Wörner (2016) proposed that three types
of end-member magmas dominate andesite formation in
the Central Andes (Fig. 6). The first is a mafic magma of
calc-alkaline basaltic andesite that is typical throughout the
Andes and represents “normal” arc basalts formed through
flux melting of the mantle wedge. The second mafic magma
type is a “shoshonitic” basalt that is enriched in incompat-
ible trace-elements (Sr, light REEs) and could be derived
either from melting old enriched lithospheric mantle or
foundered amphibole-bearing cumulate rocks from earlier
magmatic periods. The third end-member is a rhyodacite
magma that is depleted in middle-to-heavy REEs, indicating
the role of garnet, amphibole, and titanite in shaping its
trace-element signatures. Such rhyodacite magmas likely
form deep in the crust through melting of crustal rocks
and hybridization with mantle-derived mafic magmas
(Hildreth and Moorbath 1988). These three end-members
rarely erupt at the surface because of the density filter effect
highly viscous nature of the rhyodacite (Fig. 2). Mixing/
mingling processes between such end-members are, thus,
fundamental in making andesites in the Central Andes.
Fractional crystallisation may be involved but appears to
play a subordinate role.
Magmatic Regimes
There are three main elements of the transcrustal magma
systems in the Central Andes that feed the stratovolcanoes.
First is the basaltic input from the mantle wedge, caused
by hydrous flux melting such as in any subduction zone
around the world. However, mafic magmas entering the
crust appear to be variable in composition with distinct
trace-element signatures. Second is advective heat from
basaltic intrusions, which causes partial melting of older
crustal lithologies and hybridization of mantle and crustal
melts in the lower to middle crust. The resulting hybrid
tion that ascend to shallow levels where they stagnate,
accumulate, and differentiate to rhyodacite magmas. Third
are the mafic magmas that bypass the lower crustal melting
zone with less assimilation and differentiation and that
rise directly to shallow levels to recharge the evolved,
shallow reservoirs. This heat and mass input keeps the
shallow reservoirs thermally “alive”, and the mafic magmas
mix with the older silicic resident magmas to form the
abundant andesites and dacites (Fig. 2). The Altiplano–Puna
Magmatic Body likely represents such accumulated mixed
magmas. Pressure of emplacement, storage, differentiation,
and mixing below stratovolcanoes can be derived from the
composition of amphibole and plagioclase crystals that
formed as phenocrysts in these magmas. Pressure estimates
for Aucanquilcha, Parinacota, and Taapaca volcanoes (e.g.
Walker et al. 2013), and Central Andes stratovolcanoes
more generally, indicate that the feeding reservoirs are
located at depths of 5–20 km (Fig. 7).
Figure 7 summarizes our model of transcrustal magma
feeding systems below different types of stratovolcanoes
in the Central Andes using our four example volcanoes.
Recharge of hotter, less evolved, magmas from depth into the
shallow reservoirs critically controls the rate and composi-
tion of magmas erupted to the surface and defines different
magmatic regimes. At low recharge rates (e.g. Aucanquilcha
and Taapaca volcanoes), the main magma types erupted
are uniform hybrid dacites and volcano growth rates are
slow. At depth, an increasingly larger plutonic body can
accumulate over millions of years (accumulation regime)
(Fig. 7). At increased mafic recharge, during the “activation
regime”, the magmatic products become compositionally
and thermally less uniform, resulting in increased mixing
and mingling. The growth of more mafic reservoirs that are
242 A ugus t 2018
spatially and temporally close to more slowly cooled and
more evolved magma batches feed eruptions with basaltic
to rhyolitic compositions at variably higher eruption rates.
The old cone of Parinacota Volcano is a good example.
At still higher recharge, this complex of diverse magma
reservoirs is bypassed in a “break-through” regime when
mafic-to-intermediate andesites are produced during
times of frequent mafic recharge by mixing/mingling and
fractional crystallisation. The high frequency of recharge
does not allow for the cooling and advanced differentia-
tion required to form more evolved compositions. These
lavas show very little evidence of mixing with a resident
evolved magma, but when such evidence exists it is in the
form of rare antecrysts entrained from more silicic crystal
mushes [antecrysts are crystals that did not crystallize
directly from the host magma in which they are entrained,
though they are, nevertheless, genetically related to that
magma]. Resulting magmas are uniform, “steady state”,
mafic-to-intermediate andesites exemplified by the young
cone of Parinacota. Deeper and higher rates of recharge are
also indicated by the sudden change to a high eruption
rate at the surface. The change in magmatic regime from
activation to breakthrough at Parinacota correlates in time
to a sector collapse, indicating a cause-and-effect relation
that needs to be explored further.
A completely different magmatic regime is represented by
rare fields of monogenetic scoria cones (Fig. 1) that occur
along the volcanic front of the Central Andes. They are
Schematic reconstruction of four distinct magmatic
Figure 7
regimes feeding Central Andean stratovolcanoes
characterised by increasing recharge rates and decreasing volcano
lifetimes. The magmatic regimes are exemplified by the volcanoes
E lements
unrelated to the stratovolcanoes and their magmas show
no evidence for mixing with resident silicic magmas in
the form of crystal cargo. Monogenetic centers represent
magmas that ascend through the crust without intercepting
or developing large magma reservoirs and associated bodies
of crystal mush.
At the other end of the spectrum, when recharge rates
are sufficiently high, large masses of eruptible, highly
evolved, and SiO2 -rich rhyolitic magmas can be mobilized
from previously accumulated magmatic bodies in relatively
short periods of time. These magmas then erupt in great
volumes (>1,000 km3) from caldera centers, such as in the
Altiplano–Puna Volcanic Complex (de Silva and Kay 2018
this issue).
The concept of different “magmatic regimes” in a
transcrustal magma system may help to better under-
stand the variability of different types of volcanoes in the
Central Andes. We observe long-lived (several My) clusters
of evolved, mostly uniform, amphibole-bearing dacitic
lavas as one end-member that contrasts with abundant,
archetypal stratovolcanoes constructed from monotonous
pyroxene andesites over a short period of time (<100 ka) as
the other. Magmatic regimes may change back-and-forth
from one extreme to the other depending on the rate of
mafic recharge from below and the size and temperature
of resident, evolved magmas at shallow levels.
of Aucanquilcha, Taapaca, and the “old” and “young” cones of
Parinacota (see Fig. 1 for locations). Abbreviation: cpx =
clinopyroxene.
243 A ugus t 2018
SUMMARY AND OUTLOOK
The Central Andes represent an exemplar of changing
magma compositions through time and space related to
the evolution of an active continental margin and thick-
ening orogenic crust. The Central Andes have sparked
many studies relating major and trace-element para-
meters to crustal thickness that, despite their
quantitative components, are essentially empirical. Such
approaches are informative but should be treated with
caution.
Volume estimates for ignimbrites in the Central Andes
(e.g. Brandmeier and Wörner 2016) remain imprecise.
Significantly more geochronologic and isotopic data are
needed at both the regional and the individual stratovol-
cano scale to produce reliable rates for magma production,
crustal recycling, and net magma input from the mantle
through space and time. Such data are critical to mass
balance calculations of magmatic processes and the thermal
evolution of the thickened crust in the Central Andes, and
for ancient orogens more generally. Ultimately, we need to
better link quantitative information on magma production
rates in the mantle, the stress pattern and thermal condi-
tions of the continental crust that controls the location
and depth of magma reservoirs, and the composition and
volume of erupted magmas through time and space.
ACKNOWLEDGMENTS
We thank Jon Blundy and Bernie Wood for the idea, encour-
agement, and critical feedback for this issue for Elements.
Many colleagues and students contributed significantly to
our Andean endeavors; far too many to name them all.
I (GW) thank Russ Harmon for getting me into studying
volcanoes in the Central Andes in the first place, the late
Jon Davidson for making sense of it all, and my Peruvian
wife Lilia for giving me a place to come home to.

REFERENCES
Beck SL and 5 coauthors (1996) Crustal-
thickness variations in the central
Andes. Geology 24: 407-410
Blum-Oeste M, Wörner G (2016) Central
Andean magmatism can be constrained
by three ubiquitous end-members. Terra
Nova 28: 434-440
Bouvet de Maisonneuve C, Dungan
MA, Bachmann O, Burgisser A (2013)
Petrological insights into shifts in
eruptive styles at Volcán Llaima (Chile).
Journal of Petrology 54: 393-420
Brandmeier M, Wörner G (2016)
Compositional variations of ignimbrite
magmas in the Central Andes over the
past 26 Ma — a multivariate statistical
perspective. Lithos 262: 713-728
Coira B, Kay SM, Viramonte J (1993)
Upper Cenozoic magmatic evolu-
tion of the Argentine Puna—a model
for changing subduction geometry.
International Geology Review 35:
677-720
de Silva SL (1989) Altiplano-Puna
volcanic complex of the Central Andes.
Geology 17: 1102-1106
de Silva SL, Kay SM (2018) Turning up
the heat: high-flux magmatism in the
Central Andes. Elements 14: 245-250
Farner MJ, Lee C-TA (2017) Effects of
crustal thickness on magmatic differen-
tiation in subduction zone volcanism:
a global study. Earth and Planetary
Science Letters 470: 96-107
Freymuth H, Brandmeier M, Wörner G
(2015) The origin and crust/mantle
mass balance of Central Andean ignim-
brite magmatism constrained by oxygen
and strontium isotopes and erupted
volumes. Contributions to Mineralogy
and Petrology 169, doi: 10.1007/
s00410-015-1152-5
Guzmán S and 7 coauthors (2014) Spatial-
temporal distribution of explosive
volcanism in the 25-28°S segment of
the Andean Central Volcanic Zone.
Tectonophysics 636: 170-189
Hildreth W, Moorbath S (1988) Crustal
contributions to arc magmatism in the
Andes of central Chile. Contributions to
Mineralogy and Petrology 98: 455-489
Hora JM, Singer BS, Wörner G (2007) Scott EM and 6 coauthors (2018) Andean
Volcano evolution and eruptive flux on surface uplift constrained by radio-
the thick crust of the Andean Central genic isotopes of arc lavas. Nature
Volcanic Zone: 40Ar/39Ar constraints Communications 9, doi: 10.1038/
from Volcán Parinacota, Chile. s41467-018-03173-4
Geological Society of America Bulletin
Stern CR (1991) Role of subduction
119: 343-362
erosion in the generation of Andean
Hu F, Ducea MN, Liu S, Chapman JB magmas. Geology 19: 78-81
(2017) Quantifying crustal thickness
Thouret J-C, Jicha BR, Paquette
in continental collision belts: global
J-L, Cubukcu EH (2016) A 25 myr
perspective and a geologic applica-
chronostratigraphy of ignimbrites
tion. Scientific Reports 7, doi: 101038/
in south Peru: implications for the
s41598-017-07849-7
volcanic history of the Central Andes.
Kay SM and 5 coauthors (1988) Tertiary Journal of the Geological Society, doi:
andean magmatism in Chile and 10.1144/jgs2015-162
Argentina between 28°S and 33°S:
Walker BA Jr and 5 coauthors (2013)
correlation of magmatic chemistry with
Crystal reaming during the assembly,
a changing Benioff zone. Journal of
maturation, and waning of an eleven-
South American Earth Sciences 1: 21-38
million-year crustal magma cycle:
Kay SM, Mpodozis C, Coira B (1999) thermobarometry of the Aucanquilcha
Neogene magmatism, tectonism, and Volcanic Cluster. Contributions to
mineral deposits of the Central Andes Mineralogy and Petrology 165: 663-682
(22°S to 33°S Latitude). In: Skinner BJ
Ward KM, Zandt G, Beck SL, Christensen
(ed) Geology and Ore Deposits of the
DH, McFarlin H (2014) Seismic imaging
Central Andes. Society of Economic
of the magmatic underpinnings
Geology Special Publication 7, pp 27-59
beneath the Altiplano-Puna volcanic
Mamani M, Wörner G, Sempere T (2010) complex from the joint inversion of
Geochemical variations in igneous surface wave dispersion and receiver
rocks of the Central Andean orocline functions. Earth and Planetary Science
(13°S to 18°S): tracing crustal thick- Letters 404: 43-53
ening and magma generation through
Wörner G and 8 coauthors (1988) The
time and space. Geological Society of
Nevados de Payachata volcanic region
America Bulletin 122: 162-182
18°S/69°W, N. Chile) I. Geological,
Michelfelder GS, Feeley TC, Wilder AD, geochemical, and isotopic observations.
Klemetti EW (2013) Modification of Bulletin of Volcanology 30: 287-303
the continental crust by subduction
Wörner G, Ulig D, Kohler I, Seyfried H
zone magmatism and vice-versa: across-
(2002) Evolution of the West Andean
strike geochemical variations of silicic
Escarpent at 18°S (N. Chile) during the
lavas from individual eruptive centers
last 25 Ma: uplift, erosion and collapse
in the Andean Central Volcanic Zone.
through time. Tectonophysics 345:
Geosciences 3: 633-667
183-198
Muir DD, Blundy JD, Hutchinson MC,
Zandt G, Leidig M, Chmielowski J,
Rust AC (2014) Petrological imaging
Baumont D, Yuan X (2003) Seismic
of an active pluton beneath Cerro
detection and characterization of the
Uturuncu, Bolivia. Contributions
Altiplano-Puna magma body, Central
to Mineralogy and Petrology 167:
Andes. Pure and Applied Geophysics
980-1004
160: 789-807
Oncken O and 5 coauthors (2006)
Deformation of the Central Andean
Upper Plate System – facts, fiction,
and constraints for plateau models. In:
Oncken O and 7 coeditors (eds) The
Andes. Frontiers in Earth Sciences.
Springer, Berlin, pp 3-27

E lements 244 A ugus t 2018

 PARTING SHOTS

WHAT’S YOUR NEXT DREAM?

Gerhard Wörner1

DOI: 10.2113/gselements.14.4.286

While working on extinct volcanoes in Germany, I always wanted to

see Hawaii. This dream came true in 1984 when I was on my way to
study volcanoes in Antarctica when we passed through Hawaii … and
I got the thrill of seeing Pu‘u ‘Ō ‘ō erupt. Later, over a beer, I was telling
my friend, Russ Harmon, how my fieldwork adventures had fulfilled
a long-held dream. “So, what’s your next dream?”, he asked. “I want
to work on the volcanoes of the Atacama in South America”, was my Figure 2 Two-peso coins ‘minted’ by the nitrate mine company, Compania de
Salitre de Antofagasta. These coins are made from bakelite (the first
answer, not realizing that he and his colleagues were planning a field plastic made from synthetic components). These coins could only be spent in the
campaign to work on Parinacota Volcano in northern Chile the next company’s store. Photo : G. Wörner.
year. I was invited to join.
This was 1986 and Chile’s dictator, Augusto Pinochet, was still in power.
the surviving workers were forced to return to the mines. The military
The shaggy border town of Arica, our departing spot to the high Andes,
colonel who was responsible for this massacre was even awarded for
had just one lonely place to hang out before departing to the field:
restoring order. Those mines became ghost towns after the synthetic
Café 21 on calle 21. The café was the center of town, an uneasy place
fixation of nitrogen from air (the Haber–Bosch process) was invented.
where students would protest and police would arrest them, and where
Later, Pinochet’s torturers used the abandoned mines as prisons for
dollar-dealers with big pockets would be the only ones to change your
those who opposed his regime.
currency. In the early mornings, a snarling voice offered daily newspa-
pers telling stories of terrorist attacks against the regime and question- This border region of the Atacama is today crossed by many trucks that
able weapons finds in the desert. These stories gave Pinochet all the deliver and export goods to and from Bolivia. Can you imagine old,
reason he needed to violently crack down on protesters. unsafe trucks overloaded with timber from the Bolivian rainforest of
the Amazon basin, crossing the Andes to make their way down this
Going into the field meant climbing a winding desert road that had
treacherous road to the Pacific harbor at Arica? In 1986, this road was
the biggest potholes imaginable. We drove from sea level to 4,500 m in
Bolivia’s only connection to the world’s markets. And, unfortunately,
one day, into the winter of the Cordillera at −18 °C at night. We had to
drug dealers and smugglers abound.
pass four police stations on the way, and each time the police recorded
the numbers of all our documents into thick books.
Unique Geology and Ecology
Disputed Land But what magnificent volcanic geology! The road from Arica to the
high Andes climbs along the flanks of deep canyons and arrives at
Arica, and the Atacama, is a disputed border land. Land mines can be
vast plains covered with extensive ignimbrites. Tectonic forces have
found at every road crossing between Chile, Peru, and Bolivia (Fig. 1).
thrust the Cordillera up to 5,500 m. Incision by rivers, fed from glacial
Chile took this territory during the “War of the Pacific” (1879–1884) in a
waters from peaks on the horizon, cut canyons almost 2,000 m deep.
sweeping advance to Lima (Peru). With the help of the British, Chile had
The combined tectonic and erosional processes have produced over-
fought and won this war, which was all about natural resources: nitrates,
steepened slopes and frequent earthquakes have caused slope failures.
gold, copper, and other riches to fill the pockets of the wealthy and
Gigantic landslides have displaced rock sections that are more than
treasuries of the European empires. The
600 m thick up to 40 km downslope.
Bolivians lost their access to the ocean,
but they still threaten to take it back at Arriving on the Altiplano, the views are breathtaking: volcanoes right
any time. They even have a navy for that and left, snow-capped glaciated peaks rising from the Altiplano to well
purpose, sailing on Lake Titicaca. above 6,000 m. Parinacota is the highest volcano in Lauca National
Park (Fig. 3). The little streams from these glaciated volcanoes feed
The nitrates of the Atacama were a
the “bofedales”, wetlands in the desert, that are rich in green pasture
precious commodity, the source for the
for the alpacas and llamas that provide wool and meat to the Aymara
explosives that killed millions during
Indians (Fig. 4). The elegant vicuñas and endemic (and monogamous)
World War I. Workers were lured into the
Andean geese live here. Coots build their nests in shallow lake waters,
desert mines with promises of paying
and giant condors circle overhead.
off their debts, which only increased
because workers had to pay for their Little jumping rabbit-like animals, with a long tail, that nurse their
shovels and other equipment in the babies in a belly pocket, are the viscachas, marsupials similar to
Figure 1 Minefield alert at a company store. Regular punishments chinchillas. Viscachas enjoy the andesitic block lavas that provide ideal
small road near the
Chilean town of Ollagüe, close further increased worker debts. The hiding spaces for rest and breeding. Their nesting material and feces
to the border with Bolivia. Photo : mine company controlled the railway are an archive for pollen, which tell stories of past climates. Wetter
G. Wörner. and even the money (F ig. 2). There climate conditions during past ice ages (and even deeper in geological
was no escape from the treacherous time) have left their traces on this land. Lake sediments and shorelines
conditions. Late in December 1907, ~18,000 mine workers, many with are evidence of vast lakes and changing climates on the Altiplano. The
families, marched the >100 km from the mines in the desert to the extensive salt flats (salars) are their remnants. The salars and their saline
coastal town of Iquique to protest against the difficult working condi- pools are home to the several species of flamingos, some endemic, that
tions and low pay. Workers in the harbor, where the nitrate was shipped, breed here.
joined them. The dispute ended in tragedy. Thousands were killed, and
Lake Chungará, near the border of Bolivia, at 4,518 m, is the highest
lake for its size (21 km2) on our planet. Lake Chungará developed after a
small river was blocked by a flank collapse between growing Parinacota
1 GZG, Abteilung Geochemie
Universität Göttingen, Germany Volcano and Ajoya Volcano, its much older glaciated neighbor. This
E-mail: gwoerne@gwdg.de blockage severed the connection to the surrounding outflowing rivers

E lements 286 A ugus t 2018

 PARTING SHOTS

indigenous Aymara Indians know their varied uses. Llareta extracts

have antibacterial effects, its resin can be used as glue or as a fragrance,
and the dried plant is the only fuel to be found at such elevations.
The Lauca National Park with the old town of Parinacota at its center,
and the Vicuñas National Reserve have been set up to preserve the
natural wonders of the region. Plans to lower the level of the lake to
feed water to an electrical power plant were abandoned and mining
activities have been restricted. Consequently, many animals thrive here.
In contrast, across the border in Bolivia, the viscachas and vicuñas are
hunted down and eaten.

Parque Nacional de Lauca (Lauca National Park) in northernmost

Figure 3
Chile with the twin volcanoes Parinacota (front center) and
Pomerape (rear). Photo : G. Wörner.
Figure 5
An example of the
llareta (Azorella
compacta) plant
growing more than
a meter thick.
Photo : G. Wörner.

Global climate change and globalization have taken their toll. Glaciers
have retreated dramatically over the past 27 years that I have conducted
fieldwork in the region. The level of Lake Chungará has dropped,
exposing tens of meters of muddy ground along its shallow shores
(Fig. 6). Climate change combined with excessive borax mining in the
National Reserve has lowered the water level of Salar de Surire and now
threatens the endemic Andean flamingos. Road traffic has exploded
with globalization. The new tarmac road from Arica to La Paz has litter
everywhere, plastic bags and bottles being tossed out by truck drivers.

Figure 6
The water levels of
Alpacas and llamas are the economic base of Aymara Indians on the Lake Chungará
Figure 4
Altiplano. The animals feed on the swampy grasslands known locally (Chile) have
as “bofedales”, which themselves receive water from higher snow melt. dropped markedly
Photo : G. Wörner. over the past few
decades. The
Lake Chungará therefore still has its original livestock of small endemic formerly pristine
lake shore is now
killfish. Biologists once told me the story of these fish: the genetic littered with
code of the endemic fish indicates their origin in the salt waters of the plastic, and the
Atlantic Ocean where their closest relatives now live. They adapted to coots use the
the freshwater rivers that drained the Amazon and lived happily until plastic bags to
make their nests.
tectonic forces changed and displaced the drainage systems. With the Photo : G. Wörner.
formation of the endorheic basin of the Altiplano at about 10 Ma, some
rivers (and some fish, apparently) were caught up between the eastern
and western cordilleras. Increasing aridity, due to blockage of rain from This arid Atacama ecosystem has not had the same level of support
the east and the development of the cold Humboldt Current along the and the attention as that of our planet’s rain forests or Antarctic jewels.
west coast, created large salars that are fed by small freshwater rivers. But, it is just as precious, unique, and vulnerable. It should not need
The killifish had to adapt and retreat to the upper reaches of these rivers a geologist’s view and almost 30 years of observation to recognize the
where they encountered fresh glacial meltwaters, such as the small river true value and beauty of these arid lands. Working there has become
south of Parinacota Volcano. When Parinacota’s flank failed, these fish my life’s true dream and it is saddening to realize that such beauty
were isolated from the rest of their native waters on the Altiplano. Since will soon be gone.
then, their genetic code has deviated from populations beyond the lake.
What’s your next dream?
This is a striking example of geological and biological coevolution.
This place is full of wonders. The llareta (yareta) cushion plant, for
example, can grow for hundreds of years at altitudes of between 4,400 m
and 4,800 m, forming bulges more than a meter high (Fig. 5). The

E lements 287 A ugus t 2018

View publication stats