Download What drives orogeny in the Andes?

What drives orogeny in the Andes?
S.V. Sobolev GeoForschungsZentrum-Potsdam, Telegrafenberg, 14473 Potsdam, Germany, and Institute of Physics of the
Earth, B. Gruzinskaya 10, Moscow, Russia
A.Y. Babeyko GeoForschungsZentrum-Potsdam, Telegrafenberg, 14473 Potsdam, Germany
ABSTRACT
The Andes, the world’s second highest orogenic belt, were generated by the Cenozoic tectonic shortening of the South American
plate margin overriding the subducting Nazca plate. We use a coupled thermomechanical numerical modeling technique to identify
factors controlling the intensity of the tectonic shortening. From
the modeling, we infer that the most important factor was accelerated westward drift of the South American plate; changes in the
subduction rate were less important. Other important factors are
crustal structure of the overriding plate and shear coupling at the
plates’ interface. The model with a thick (40–45 km at 30 Ma)
South American crust and relatively high friction coefficient (0.05)
at the Nazca–South American interface generates !300 km of tectonic shortening during 30–35 m.y. and replicates the crustal structure and evolution of the high central Andes. The model with an
initially thinner ("40 km) continental crust and lower friction coefficient ("0.015) results in "40 km of South American plate shortening, replicating the situation in the southern Andes. Our modeling also demonstrates the important role of the processes leading
to mechanical weakening of the overriding plate during tectonic
shortening, such as lithospheric delamination, triggered by the
gabbro-eclogite transformation in the thickened continental lower
crust, and mechanical failure of the sediment cover at the shield
margin.
and Molnar, 1987; Somoza, 1998), the beginning of intensive tectonic
shortening in the Andes was associated with the major reorganization
of the plates, followed by an increase of the Nazca–South American
convergence rate ca. 25–30 Ma. Russo and Silver (1996) and Silver et
al. (1998) attributed the Andean orogeny to the Cenozoic increase of
the westward drift rate of the South American plate. Lamb and Davis
(2003) suggested that high shear stress at the interface between Nazca
and the South American plate, caused by the Cenozoic climatecontrolled sediment starvation in the Central Andean trench, played a
leading role in Andean orogeny.
The diversity of the suggested hypotheses reflects the complexity
of the deformation processes responsible for the Andean orogeny, but
it also indicates the lack of quantitative understanding of these pro-
Keywords: Andes, subduction, orogeny, numerical model.
INTRODUCTION
The Andes Mountains extend along the entire western margin of
the South American plate above the subducting Nazca plate. The South
American plate is drifting westward at a rate that has increased from
2 to 3 cm/yr during the past 30 m.y. (Silver et al., 1998). There is a
dramatic difference in structure and evolution between the central Andes (#17$–27$S) and the rest of the Andes. The Altiplano-Puna plateau
of the central Andes is the second highest plateau in the world, after
the Tibetan Plateau, with an average elevation of #4 km and an area
of !500,000 km2 (Fig. 1A). The plateau was formed in the Cenozoic
by as much as 300–350 km of crustal shortening in the western edge
of the South American plate (Isacks, 1988; Allmendinger and Gubbels,
1996; Allmendinger et al., 1997; Lamb et al., 1997; Kley and Monaldi,
1998; Lamb and Davis, 2003; Elger et al., 2005). This shortening generated unusually thick, hot, and felsic continental crust (Allmendinger
et al., 1997; Beck and Zandt, 2002; Yuan et al., 2002). No high plateau
exists in the northern and southern Andes (Fig. 1), where only minor
("50 km) tectonic shortening has been reported (e.g., Allmendinger et
al., 1997; Lamb et al., 1997; Kley and Monaldi, 1998).
Perhaps the key question of the Andean orogeny is why the high
plateau developed only in the central Andes and only in Cenozoic time
(mostly during the past 30 m.y), although the Nazca plate has been
subducted along the entire western margin of the South American plate
during more than 200 m.y. (e.g., Isacks, 1988; Allmendinger et al.,
1997). Several ideas have been proposed to answer this question. Isacks
(1988) suggested that before ca. 25–30 Ma, the central Andes were
underlain by a flat slab that became steeper ca. 25 Ma, causing thermal
weakening and intensive tectonic shortening of the compressed lithosphere of the overriding plate. In another hypothesis (e.g., Pardo-Casas
Figure 1. A: Surface topography of Andes with indicated major
structural features. Trench adjacent to high central Andes has no
sedimentary fill, which may increase friction in subduction channel
(Lamb and Davis, 2003). B: Model setup and boundary conditions.
Subducting plate is 45 m.y. old. Initial thickness of continental lithosphere is 100–130 km, with thickest lithosphere in eastern (right)
part of model corresponding to Brazilian shield margin. South American plate is drifting to west (left) with velocity increasing from 2 to
3 cm/yr during past 30 m.y. (Silver et al., 1998). Lower end of Nazca
plate is pulled down, with velocity changing from 5 to 13 cm/yr (Somoza, 1998). C: Shear stress in subduction channel.
! 2005 Geological Society of America. For permission to copy, contact Copyright Permissions, GSA, or [email protected].
Geology; August 2005; v. 33; no. 8; p. 617–620; doi: 10.1130/G21557.1; 4 figures; Data Repository item 2005118.
617
cesses. Each of these hypotheses is based on analyses of many observations, and all have solid observational grounds. However, none have
yet been quantitatively tested by modeling the coupled dynamic and
thermal interaction of the overriding and subducting plates. In this
study we present such models, focusing on dependence of tectonic
deformation of the overriding plate during the past 35 m.y. on (1)
convergence rate, (2) overriding rate, (3) strength of mechanical coupling between subducted and overriding plates, and (4) initial lithospheric structure.
METHOD AND MODEL
Modeling the dynamic interaction between subducting and overriding plates demands realistic rheological models of both plates, including elasticity, plasticity, and temperature- and stress-dependent viscosity. To accomplish this modeling, we use a two-dimensional parallel
thermomechanical finite-element/finite-difference code LAPEX-2D
(see Babeyko et al. [2002] for the description of the previous version
of the code). This code combines the explicit Lagrangian algorithm
FLAC (Cundall and Board, 1988; Poliakov et al., 1993) with the particle technique similar to the particle-in-cell method (e.g., Moresi et
al., 2003). Particles track material properties and full stress tensor, minimizing numerical diffusion related to remeshing. This method allows
the use of realistic temperature- and stress-dependent viscoelastic rheology combined with Mohr-Coulomb plasticity for layered oceanic and
continental lithospheres (Fig. 1B).
The rheological parameters are taken from published experimental
and theoretical rheological studies and are presented in Data Repository
Table DR1.1 Because of the presence of the subduction zone, we have
used rheological parameters for ‘‘wet’’ rocks everywhere in the model
except for the slab and mantle lithosphere of the shield margin. In the
crust, we employ friction- and viscosity-strain softening, which is assumed to be more intensive in the Paleozoic sediments in the Subandean zone (Fig. 1B). The viscous deformation in the mantle is considered to be driven by competing dislocation, diffusion, and Peierls
creep mechanisms. The numerical method routinely includes shear
heating and gabbro-eclogite phase transformation (model details in the
Data Repository; see footnote 1).
A set of models was run for initial crustal structures expected for
the central and southern Andes at 30–35 Ma. Initial crustal structure
for the central Andes (Fig. 2A) contains thick felsic upper crust and
thinner mafic lower crust, with a total crustal thickness of 40–45 km,
assuming that the crust was already significantly shortened by 30–35
Ma (Allmendinger et al., 1997; Lamb et al., 1997). Initial crust for the
southern Andes consists of equally thick upper felsic and lower mafic
layers and has a total thickness of 35–40 km.
The geometry and boundary conditions incorporated in all our
models are schematically shown in Figure 1B and Figure 2A. In all
models, we explore the interaction of the 45 Ma subducting Nazca
plate with the 100–130-km-thick lithosphere of the overriding South
American plate during the past 30–35 m.y. We assume low-angle geometry of the subducting plate, consistent with the present-day structure in the Andes. The model box is 1200 km long and 400 km high
(Fig. 1B), moving to the left (west) together with the overriding plate.
The drift of the overriding plate and subduction are generated by pushing the overriding plate at its right boundary and by pulling the slab
from below (Fig. 1B), with the absolute velocities (in hotspot frame)
taken from plate tectonics reconstructions (Silver et al., 1998; Somoza,
1998). All other parts of the model box boundary (Fig. 1B) are open
for free motion of material. We emphasize that the lower end of the
1GSA Data Repository item 2005118, supplementary data, is available at
www.geosociety.org/pubs/ft2005.htm, or on request from editing@geosociety.
org or Documents Secretary, GSA, P.O. Box 9140, Boulder, CO 80301-9140,
USA.
618
Figure 2. Time snapshots of evolution of tectonic shortening for
model of central Andes. Positions of snapshots along horizontal
axis are their true positions in hotspot frame. Color codes correspond to rock types. Approximately 60% of South American western
drift is accommodated by trench rollback and ~40% by tectonic
shortening of South American margin. Note that lower end of the
slab moves to left >200 km during 35 m.y. and hence slab is not
anchored. Note also intensive thickening of felsic upper crust (yellow, orange) and loss of mafic lower crust (green) in South American
plate during past 18 m.y. (model times 17–35 m.y.), while mantle
lithosphere (light green) in South American plate is becoming thinner during tectonic shortening. At ~25 m.y. modeling time, sedimentary cover of shield margin (red) fails and shield begins to underthrust under growing plateau.
slab is not fixed and can be located anywhere between the left and
right boundaries of the model. Therefore, the slab is by no means
anchored, and retains the full dynamic freedom, e.g., in regard to the
retrograde motion. We also note that our boundary conditions do not
prescribe tectonic shortening of the upper plate as it was done in the
previous models (Wdowinski and Bock, 1994; Pope and Willett, 1998).
The interplate interface is modeled as a thin subduction channel,
i.e., an #12-km-thick (three finite elements) layer with the special rheology. For each finite element within the channel, we use either a
frictional (Mohr-Coulomb, brittle) elastoplastic or a temperaturedependant viscous (Peacock, 1996) rheological model, depending on
which of those models requires the lowest shear stress. In this approach, the shallow low-temperature part of the subduction channel has
frictional (brittle) rheology with shear stress increasing with depth (Fig.
1C, black curve). At deeper depth and higher temperature, the viscous
flow mechanism takes over, and shear stress in the channel decreases
with depth (Fig. 1C, red curve). The depth where frictional rheology
changes to viscous rheology depends on the friction coefficient, and it
is deeper if friction is lower (cf. dashed and solid black curves in Fig.
1C). We consider the friction coefficient in the subduction channel to
be a model parameter, and we change it from 0 to 0.15, in agreement
with previous estimates (Bird, 1978; Peacock, 1996; Hassani et al.,
1997).
MODEL FOR THE CENTRAL ANDES
We first found the model that best fits the observations for the
central Andes for the time between 35 Ma to the present day. To do
that, we assumed a thick (40–45 km) crust (Fig. 2A) as a starting
model and used kinematic boundary conditions mimicking subduction
and overriding velocities according to the plate tectonic reconstructions
(Silver et al., 1998; Somoza, 1998). With this setup, we performed
GEOLOGY, August 2005
Figure 3. Evolution of surface topography in central Andes model.
Note formation of high topography and then plateau during last 10
m.y. (model times, 25–35 m.y.).
several numerical experiments, changing the friction coefficient in subduction channel %. All model runs with % ! 0.10 resulted in slab
breakoff and termination of the subduction. At % & 0.05–0.10, subduction survived, but large interplate coupling led to too-strong shortening of the overriding plate.
The model replicates the case of the central Andes most closely
at % # 0.05. In this model, 58% of the westward drift of the South
American plate during the past 35 m.y. is accommodated by the Nazca
slab trench rollback, and the rest (42%) by tectonic shortening of the
South American plate (37%) and subduction erosion (5%) (Fig. 2).
During the shortening, the felsic crust thickness almost doubles, while
the mafic lower crust and mantle lithosphere actually become thinner.
The reason for this is the delamination of the lower crust and mantle
lithosphere driven by gabbro-eclogite transformation in the lower crust,
first discussed in the Andean context by Kay and Kay (1993). Thickening of the crust to more than 45 km switches on mineral reactions
in the mafic lower crust, which increases its density to higher than the
density of mantle peridotite (3300 kg/m3). The bodies of the dense
lower crust and mantle lithosphere tend to sink into the less dense
asthenosphere. While sinking, most of such bodies are moved by the
corner flow toward the trench, join the slab, and are then subducted
into the mantle (Fig. 2B), somewhat similar to the ablative subduction
scenario suggested by Pope and Willett (1998). For more details, see
Data Repository Figure DR1 (see footnote 1).
After #20–25 m.y. model time, the tectonic shortening generates
high topography near the magmatic arc and in the backarc close to the
shield margin (Fig. 3, orange curve). Large topographic gradients initiate intensive flow in the lower and middle crust. They even out crustal
thickness and surface topography and produce a 4-km-high plateau at
30–35 m.y. (Fig. 3), i.e., during the last 5 m.y., in accord with the
timing of the plateau uplift suggested by Gregory-Wodzicki (2000) and
similar to the previous modeling results by Wdowinski and Bock
(1994). At the same time, the tectonic shortening reaches 300–350 km
(curve indicated by filled circles in Fig. 4), in agreement with the
geological estimations of the maximal shortening in the central Andes
(Kley and Monaldi, 1998). The model also predicts failure of the foreland sediments at #25 m.y. model time, followed by underthrusting
of the shield margin and by a switch from a pure-shear to simple-shear
mode of shortening, in agreement with the geological model by Allmendinger and Gubbels (1996).
WHAT DRIVES ANDEAN OROGENY?
Here we examine sensitivity of tectonic shortening to potentially
important factors like overriding rate, strength of mechanical coupling
between subducted and overriding plates, initial lithospheric structure,
and convergence rate. To do that, we modified the central Andes model
described here, ‘‘switching off’’ different factors and examining consequences for the tectonic shortening during the 30 m.y. of evolution.
First we switched off the acceleration of the South American
plate’s westward drift in the central Andes model, leaving all other
GEOLOGY, August 2005
Figure 4. Calculated tectonic shortening versus time for different
models. Numbers near models indicate subduction channel friction
coefficient (first number) and South American western drift velocity
(second group of numbers). Time ranges of some critical processes
in central Andes model are shown below time axis.
parameters unchanged. This modified model (curve denoted by open
boxes in Fig. 4) generates only 60% of the shortening achieved in the
central Andes model during the 30 m.y. Moreover, if the drift velocity
is decreased, to 1 cm/yr, no tectonic shortening happens. These modeling results suggest that the high overriding rate of the South American plate in the Cenozoic may have been a major factor controlling
tectonic shortening in the central Andes, in agreement with previous
studies (Russo and Silver, 1996; Silver et al., 1998). However, we also
conclude that it is unlikely that acceleration of the drift alone could be
responsible for the entire observed tectonic shortening.
Next we switched off the high subduction channel friction in the
central Andes model (setting % to 0.015 instead of 0.05). The resulting
model (Fig. 4, curve indicated by open circles) still generates large
tectonic shortening, i.e., 74% of the shortening achieved in the central
Andes model. However, if we additionally use the mechanically stronger thin-crust (35–40 km) initial model instead of the thick-crust (40–
45 km) model, the tectonic shortening at 30 m.y. model time reduces
to "40 km, which matches the situation in the southern Andes well
(Fig. 4, curve indicated by solid diamonds). These modeling results
suggest that although reducing the subduction channel friction coefficient by four to five times has significant effect on tectonic shortening,
only a combination of this factor with another (e.g., initial crustal structure) could have been responsible for the dramatic difference in the
amount of tectonic shortening between the central and southern Andes.
The shape of the modeled shortening curve for the central Andes
model (solid circles in Fig. 4) shows that the maximal convergence
rate of 15 cm/yr achieved at 10–15 m.y. model time does not have
much effect on the shortening rate. From that, we infer that an increased convergence rate at 25 Ma could not be a reason for the intensive orogeny in the central Andes.
In addition to the factors considered here, other processes can also
influence the rate of tectonic shortening. Active delamination accompanied by the blocking of the corner flow by the delaminated material
(see also Fig. 2B) at 15–20 m.y. model time significantly intensifies
shortening as a result of increased coupling between the plates and
mechanical weakening of the overriding plate. Another increase in the
shortening rate in the central Andes model is associated with the failure
of the foreland sediments, followed by underthrusting of the shield
619
margin and by a switch from a pure-shear to simple-shear mode of
shortening at #25 m.y. model time.
We conclude that the major factor controlling Andean orogeny is
likely accelerating westward drift of the South American plate. However, we also infer that neither this drift nor any other factor was alone
responsible for the orogeny in the central Andes. Our modeling suggests that the extreme orogeny in the Andes took place at the time and
in the place when and where at least three major conditions were met:
(1) a high overriding rate of the South American plate (achieved during
the past 30 m.y. and especially during the past 10 m.y.); (2) a thick
crust (40–45 km) in the backarc that resulted from tectonic shortening
before 30 Ma, including enhanced shortening during the flat-slab episode; and (3) a friction coefficient of #0.05 in the subduction channel
(both latter conditions were likely achieved only in the central Andes).
The models also demonstrate the important role of several accompanying processes in causing the internal mechanical weakening of the
South American plate during the tectonic shortening. The main accompanying processes are lithosphere delamination, triggered by the
gabbro-eclogite transformation in the thickened continental lower crust,
and mechanical failure of the sediments covering the shield margin.
Note, however, that here we do not consider changes in slab geometry
in time and space, which could potentially also affect deformation of
the overriding plate. Consideration of these effects will require threedimensional modeling, which is beyond the scope of this paper.
Based on our modeling results for the central Andes, we can infer
some general remarks on the deformation of a continental margin overriding a subducting plate. The fast overriding (!1–2 cm/yr) leads to
the low-angle subduction (van Hunen et al., 2004) and associated
strong coupling between the plates. The overriding plate undergoes
compression with the maximal stresses achieved in regions of highest
interplate friction and of increased buoyancy of the slab due to subducted ridges or ocean plateaus. Those are the places where the continental crust first reaches the critical thickness of #45 km when the
gabbro-eclogite–driven lithospheric delamination may be activated, intensifying the tectonic shortening. Later, the less compressed parts of
the overriding plate also become involved in this process. Hence, a
large tectonic shortening of the overriding plate margin is probably an
unavoidable consequence of a fast and long overriding. The rate of
shortening depends, however, on a series of parameters. A suitable
combination of the most important parameters, i.e., high overriding
velocity, high interplate friction, and prethickened crust, generates
spectacular orogens like the Cenozoic central Andes, while in other
cases (like in the southern Andes) the orogeny remains subtle.
ACKNOWLEDGMENTS
This work is a part of the collaborative research program SFB-267 Deformation Processes in the Andes, supported by the Deutsche Forschungs Gemeinschaft and GeoForschungsZentrum-Potsdam. We thank Onno Oncken, Tim
Vietor, and members of the SFB-267 team for fruitful discussions. The paper
benefited from the comments of Shimon Wdowinski, Chris Beaumont, and an
anonymous reviewer.
Note in press: We dedicate this paper to the memory of Peter Giese who initiated the Berlin-Potsdam multidisciplinary project in the central Andes.
REFERENCES CITED
Allmendinger, R.W., and Gubbels, T., 1996, Pure and simple shear plateau uplift,
Altiplano-Puna, Argentina and Bolivia: Tectonophysics, v. 259, p. 1–13.
Allmendinger, R.W., Jordan, T.E., Kay, S.M., and Isacks, B.L., 1997, The evolution of the Altiplano-Puna plateau of the central Andes: Annual Review
of Earth and Planetary Sciences, v. 25, p. 139–174.
Babeyko, A.Y., Sobolev, S.V., Trumbull, R.B., Oncken, O., and Lavier, L.L.,
2002, Numerical models of crustal-scale convection and partial melting
620
beneath the Altiplano-Puna plateau: Earth and Planetary Science Letters,
v. 199, p. 373–388.
Beck, S.L., and Zandt, G., 2002, The nature of orogenic crust in the central
Andes: Journal of Geophysical Research, v. 107, p. 2230.
Bird, P., 1978, Stress and temperature in subduction zones: Tonga and Mariana:
Royal Astronomical Society Geophysical Journal, v. 55, p. 411–434.
Cundall, P.A., and Board, M., 1988, A microcomputer program for modelling
large-strain plasticity problems, in Swoboda, G., ed., 6th International
Conference in Numerical Methods in Geomechanics: Brookfield, Vermont,
A.A. Balkema, p. 2101–2108.
Elger, K., Oncken, O., and Glodny, J., 2005, Plateau-style accumulation of
deformation—The Southern Altiplano: Tectonics (in press).
Gregory-Wodzicki, K.M., 2000, Uplift history of the Central and Northern Andes: A review: Geological Society of America Bulletin, v. 112,
p. 1091–1105.
Hassani, R., Jongmans, D., and Chery, J., 1997, Study of plate deformation and
stress in subduction processes using two-dimensional numerical models:
Journal of Geophysical Research, v. 102, p. 17,951–17,965.
Isacks, B.L., 1988, Uplift of the Central Andean plateau and bending of the
Bolivian orocline: Journal of Geophysical Research, v. 93, p. 3211–3231.
Kay, R., and Kay, S., 1993, Delamination and delamination magmatism: Tectonophysics, v. 219, p. 177–189.
Kley, J., and Monaldi, C., 1998, Tectonic shortening and crustal thickness in
the central Andes: How good is the estimate?: Geology, v. 26, p. 723–726.
Lamb, S., and Davis, P., 2003, Cenozoic climate change as a possible cause for
the rise of the Andes: Nature, v. 425, p. 792–797.
Lamb, S., Hoke, L., Kennan, L., and Dewey, J., 1997, Cenozoic evolution of
the central Andes in Bolivia and northern Chile, in Burg, J.-P., and Ford,
M., eds., Orogeny through time: Geological Society [London] Special
Publication 121, p. 237–264.
Moresi, L.-N., Dufour, F., and Muhlhaus, H.B., 2003, A LaGrangian integration
point finite element method for large deformation modeling of viscoelastic
geomaterials: Journal of Computational Physics, v. 184, p. 476–497.
Pardo-Casas, F., and Molnar, P., 1987, Relative motion of the Nazca (Farallon)
and South American plates since Late Cretaceous time: Tectonics, v. 6,
p. 233–248.
Peacock, S., 1996, Thermal and petrologic structure of subduction zones, in
Bebout, G., et al., eds., Subduction, top to bottom: American Geophysical
Union Geophysical Monograph 96, p. 119–134.
Poliakov, A.N., Cundall, P.A., Podladchikov, Y.Y., and Lyakhovsky, V.A., 1993,
An explicit inertial method for the simulation of the viscoelastic flow: An
evaluation of elastic effects on diapiric flow in two- and three-layer models, in Stone, D.B., and Runcorn, S.K., eds., Flow and creep in the solar
system: Observations, modelling and theory: Dordrecht, Kluwer Academic Publishers, p. 175–195.
Pope, D.C., and Willett, S.D., 1998, Thermo-mechanical model for crustal thickening in the central Andes driven by ablative subduction: Geology, v. 26,
p. 511–514.
Russo, R., and Silver, P.G., 1996, Cordillera formation, mantle dynamics, and
the Wilson cycle: Geology, v. 24, p. 511–514.
Silver, P.G., Russo, R.M., and Lithgow-Bertelloni, C., 1998, Coupling of South
American and African plate motion and plate deformation: Science,
v. 279, p. 60–63.
Somoza, R., 1998, Updated Nazca (Farallon)—South America relative motions
during the last 40 m.y.: Implications for mountain building in the central
Andean region: Journal of South American Earth Sciences, v. 11,
p. 211–215.
van Hunen, J., van den Berg, A.P., and Vlaar, N.J., 2004, Various mechanisms
to induce shallow flat subduction: A numerical parameter study: Physics
of the Earth and Planetary Interiors, v. 146, p. 179–194.
Wdowinski, S., and Bock, Y., 1994, The evolution of deformation and topography of high elevated plateaus: 2. Application to the central Andes: Journal of Geophysical Research, v. 99, p. 7121–7130.
Yuan, X., Sobolev, S.V., and Kind, R., 2002, New data on Moho topography
in the Central Andes and their geodynamic implications: Earth and Planetary Science Letters, v. 199, p. 389–402.
Manuscript received 31 January 2005
Revised manuscript received 13 April 2005
Manuscript accepted 18 April 2005
Printed in USA
GEOLOGY, August 2005