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

net/publication/284669670

Flat subduction dynamics and deformation of the South American plate

Article · January 2008

CITATIONS READS
37 119

7 authors, including:

Nicolas Espurt Francesca Funiciello

Aix-Marseille Université Università Degli Studi Roma Tre
89 PUBLICATIONS   1,503 CITATIONS    117 PUBLICATIONS   3,831 CITATIONS   

SEE PROFILE SEE PROFILE

J. Martinod Benjamin Guillaume

Université Savoie Mont Blanc Université de Rennes 1
159 PUBLICATIONS   6,440 CITATIONS    40 PUBLICATIONS   766 CITATIONS   

SEE PROFILE SEE PROFILE

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

CROCODYL project, INSU/EC2CO/DRIL View project

Research on the Quaternary of Italy View project

All content following this page was uploaded by J. Martinod on 17 June 2019.

The user has requested enhancement of the downloaded file.

 TECTONICS, VOL. 27, TC3011, doi:10.1029/2007TC002175, 2008
Click
Here
for
Full
Article
Flat subduction dynamics and deformation of the South American
plate: Insights from analog modeling
Nicolas Espurt,1,2 Francesca Funiciello,3 Joseph Martinod,1 Benjamin Guillaume,1
Vincent Regard,1 Claudio Faccenna,3 and Stéphane Brusset1
Received 14 June 2007; revised 13 January 2008; accepted 12 March 2008; published 21 June 2008.
[1] We present lithospheric-scale analog models, from analog modeling, Tectonics, 27, TC3011, doi:10.1029/
investigating how the absolute plates’ motion and 2007TC002175.
subduction of buoyant oceanic plateaus can affect both
the kinematics and the geometry of subduction,
possibly resulting in the appearance of flat slab 1. Introduction
segments, and how it changes the overriding plate [2] Flat subduction zones correspond to oceanic slabs
tectonic regime. Experiments suggest that flat subducting horizontally under an overriding continental
subductions only occur if a large amount of a plate. The best known present-day examples of flat subduc-
buoyant slab segment is forced into subduction by tions are located beneath South America, in Peru (3– 15°S),
kinematic boundary conditions, part of the buoyant and central Chile/NW Argentina (27 – 33°S) [e.g., Sacks and
Okada, 1974; Baranzangi and Isacks, 1979; Pilger, 1981;
plateau being incorporated in the steep part of the slab
McGeary et al., 1985; Cahill and Isacks, 1992; Engdalh et
to balance the negative buoyancy of the dense oceanic al., 1998; Gutscher et al., 2000b; Pardo et al., 2002]
slab. Slab flattening is a long-term process (
10 Ma), (Figure 1). In these two zones, the Wadati-Benioff zone
which requires the subduction of hundreds of and available tomographic images show the subducting
kilometers of buoyant plateau. The overriding plate plate dipping approximately 30° from the trench to a depth
shortening rate increases if the oceanic plateau is large of 100– 120 km before proceeding horizontally below the
enough to decrease the slab pull effect. Slab flattening South American continental plate and sinking in the upper
increases the interplate friction force and results in mantle at 700 and 600 km from the trench in Peru and NW
migration of the shortening zone within the interior of Argentina, respectively. These two flat subduction zones are
the overriding plate. The increase of the overriding spatially correlated in the South American continental plate
plate topography close to the trench results from (1) the by (1) an absence of Quaternary volcanism as a conse-
buoyancy of the plate subducting at trench and (2) the quence of the absence of an asthenospheric wedge between
the subducting slab and the continental plate [Nur and Ben-
overriding plate shortening. Experiments are Avraham, 1981; McGeary et al., 1985; Gutscher et al.,
compared to the South American active margin, 2000a; Ramos et al., 2002; Kay and Mpodozis, 2002]; (2) a
where two major horizontal slab segments had seismic energy released within the overriding plate larger
formed since the Pliocene. Along the South than above the steep subduction zones, which may result
American subduction zone, flat slab segments below from the greater contact area between the two plates
Peru and central Chile/NW Argentina appeared at [Gutscher et al., 2000b; Gutscher, 2002]; (3) a concentra-

7 Ma following the beginning of buoyant slab tion of the overriding plate shortening above the flat slab
segments’ subduction. In northern Ecuador and segments [Sébrier et al., 1985; Jordan and Allmendinger,
northern Chile, the process of slab flattening 1986; Pardo et al., 2002; Ramos et al., 2002; Rousse et al.,
resulting from the Carnegie and Iquique ridges’ 2003; Wagner et al., 2005; Siame et al., 2005]; and
subductions, respectively, seems to be active but not (4) vertical topographic motions of the coastal areas
[Macharé and Ortlieb, 1992; Gutscher et al., 1999a; Yañez
completed. The formation of flat slab segments below et al., 2001; Hampel, 2002; Clift et al., 2003; Hampel et al.,
South America from the Pliocene may explain the 2004b; Wipf, 2006], Andean Cordillera [McNulty and
deceleration of the Nazca plate trenchward velocity. Farber, 2002], and back-arc regions on the eastern side of
Citation: Espurt, N., F. Funiciello, J. Martinod, B. Guillaume, the Andes [Dumont, 1996; Ramos et al., 2002; Espurt et al.,
V. Regard, C. Faccenna, and S. Brusset (2008), Flat subduction 2007].
dynamics and deformation of the South American plate: Insights [3] Two main causes have been proposed to explain the
formation of these flat subduction zones:
1
Laboratoire des Mécanismes et Transferts et Géologie, Université de [4] 1. The South American plate moves rapidly toward
Toulouse, CNRS, IRD, OMP, Toulouse, France. the Nazca trench in the hot spot reference frame [Jarrard,
2
Now at Institut Français du Pétrole, Rueil-Malmaison, France.
3
Dipartimento Scienze Geologiche, University Roma TRE, Rome, Italy. 1986; Scholz and Campos, 1995; Silver et al., 1998]. Hence
the active overthrusting of South America above the young
Copyright 2008 by the American Geophysical Union. Nazca oceanic plate exceeds the descent velocity of the slab
0278-7407/08/2007TC002175$12.00

TC3011 1 of 19
TC3011 ESPURT ET AL.: FLAT SLAB AND UPPER PLATE DEFORMATION TC3011

Figure 1

2 of 19
TC3011 ESPURT ET AL.: FLAT SLAB AND UPPER PLATE DEFORMATION TC3011

in the upper mantle and may be responsible for the al., 2004b]. These models show that bathymetric highs and
formation of flat slab segments [Vlaar, 1983; van Hunen oceanic ridges have important consequences on uplift,
et al., 2002b]. tectonic, sedimentary, and erosional processes in the fore-
[5] 2. The subduction of buoyant anomalies (such as arc zone. Similar observations were found using 2-D
bathymetric highs, aseismic ridges, or oceanic plateaus) numerical models [e.g., Chung and Kanamori, 1978;
increases the thickness of the crust and locally reduces the Kodama, 1984; Moretti and Ngokwey, 1985; Collot et al.,
average density of the oceanic plate. Thus Peru and central 1985; Geist et al., 1993]. van Hunen et al. [2002a, 2002b,
Chile/NW Argentina flat subduction zones may result from 2004] reproduce the formation of flat slab segments result-
buoyant oceanic segments (Nazca and Juan Fernandez ing from the subduction of buoyant oceanic plateaus. van
ridges) taken toward subduction by global plate motion Hunen et al. [2002a, 2002b, 2004] show that the trenchward
[Nur and Ben-Avraham, 1981; Pilger, 1981; McGeary et motion of the overriding plate facilitates the formation of
al., 1985; Gutscher et al., 2000b]. Other flat subduction long flat slab segments similar to those observed beneath
zones reported all around the Pacific Ocean (western New South America. From analog experiments modeling the
Guinea, SW Japan, southern Alaska, central Mexico, and subduction of a viscous plate, Martinod et al. [2005] also
Costa Rica) suggest that parameters resulting in the forma- conclude that flat slab segments only occur if the conver-
tion of flat slab segments are not specific to the South gence between subducting and overriding plates is forced by
American margin [Gutscher et al., 2000b]. external boundary conditions.
[6] Nevertheless, causes of the appearance of flat slab [8] The aim of this paper is to present lithospheric-scale
segments are the subject of an active discussion. For analog models using viscous materials and to explore both
instance, it has been recently suggested that the moderate the conditions permitting the appearance of flat subduction
crustal thickness of the Juan Fernandez Ridge is not zones and their consequences on the overriding plate
sufficient to make the buoyant slab responsible for horizon- deformation. Experiments have been specially dedicated
tal subduction [Kopp et al., 2004]. These authors suggest (1) to observe how the trenchward absolute overriding plate
that upper mantle hydration may be an additional mecha- motion and the buoyancy of the subducting plate control the
nism favoring flat subduction. In contrast, the Carnegie formation of a flat slab segment, (2) to study the causal and
Ridge does not result in the formation of any flat subduction temporal relationships existing between the oceanic plateau
zone [Guillier et al., 2001] in spite of the associated broad subduction and the flattening of the slab, and (3) to study
bathymetric anomaly (Figures 1b and 1c). The Iquique the effects of the flat subduction on the overriding plate
Ridge, despite having a more continuous crustal root than tectonic regime. We compare our analog results with geo-
the Juan Fernandez Ridge, is also associated with a steeply logical and geophysical data from the Andean subduction
subducting slab beneath northern Chile [Tassara et al., zone.
2006] (Figures 1b and 1c).
[7] Several kinds of analog and numerical models have
been performed by previous authors to study either the
2. Model Set Up
conditions resulting in the occurrence of horizontal subduc- 2.1. Experimental Procedure
tion zones or the consequence of the subduction of aseismic
[9] We use Newtonian viscous materials within a Plexi-
ridges on the overriding plate. Consequences of bathymetric
glas tank (80  80  20 cm3) to reproduce the subduction
highs and aseismic ridges’ subduction on the fore-arc area
of lithospheric plates in the upper mantle (Figure 2).
have been studied by means of crustal-scale analog models
Lithospheric plates are modeled using high-viscosity sili-
using granular materials [Dominguez et al., 1998; Hampel et

Figure 1. Geodynamic setting of the Nazca-South American convergence system. (a) Map of the eastern Pacific Ocean
and South America. Bathymetric data from the Geosat and European Remote Sensing satellite system 1 (ERS-1) spacecraft
[Smith and Sandwell, 1997] and elevation data from NASA Shuttle Radar Topographic Mission Gtopo 30. Relative plate
motions are indicated by arrows in mm a1 [Gripp and Gordon, 2002]. These velocities are distributed asymmetrically on
the Nazca and South American plates, showing a rapid westward displacement of the South American plate. Earthquake
epicenters are plotted: 0 – 99 km in black; 100– 199 km in dark gray, and 200– 700 km in light gray (taken from U.S.
Geological Survey (USGS) National Earthquake Information Center (USGS National Earthquake Information Center data
are available athttp://earthquake.usgs.gov/regional/neic/neic_bulletins.php). (b) The Nazca plate bears many topographic
anomalies (aseismic ridges or oceanic plateaus marked in gray areas) where the crust is anomalous thick and buoyant. The
estimated subducted lengths of these anomalies beneath the South American plate are shown in light gray areas. Red
triangles mark active andesitic volcanism [Gutscher et al., 2000b]. (c) Profiles of the Nazca slab beneath the South
American plate from earthquake epicenters plotted [Gutscher et al., 2000b]. The subduction of the Carnegie and Iquique
ridges are not marked by any flat slab segment beneath northern Ecuador (marked ‘‘A’’) and northern Chile/Bolivia
(marked ‘‘C’’), respectively. In contrast, there exists a causal relationship between the subduction of the Nazca and Juan
Fernandez ridges and the flat slab segments of Peru (marked ‘‘B’’) and central Chile/NW Argentina (marked ‘‘D’’),
respectively [Gutscher et al., 2000b]. Red arrows show the locus of maximum shortening within the South American plate
above the Peruvian and central Chile/NW Argentina flat slab segments [Gutscher et al., 2000b; Pardo et al., 2002].

3 of 19
TC3011 ESPURT ET AL.: FLAT SLAB AND UPPER PLATE DEFORMATION TC3011

Figure 2. (a) Schematic sketch of the three-experiment set up. The silicone layers model the
lithospheric plates, and the dense glucose syrup models the upper mantle. The bottom of the Plexiglas
tank represents the high-gradient viscosity increase at the upper/lower mantle boundary. The subducting
plate simulates a 50-Ma-old dense oceanic plate followed by a light oceanic plate (aseismic ride or
oceanic plateau). See Table 3 for the values of h and H. (b) Top view of the experimental set up. Silicone
plates present different densities (rd,p,op) and viscosities (hj,p,op) with the overriding continental plate to
the right, the dense oceanic plate in the middle, and the oceanic plateau to the left. The plates are
surrounded by the glucose syrup (see Table 3 for the values of rgs and hgs). An initial regular grid drawn
on the overriding plate allows us to evaluate the deformation within the plates during the experiment. In
the configuration drawn here, the overriding continental plate is pushed at constant speed by a piston
above the subducting plate. See also Table 3 for the values of W, Ld, Lp, and Lop.

4 of 19
TC3011 ESPURT ET AL.: FLAT SLAB AND UPPER PLATE DEFORMATION TC3011

Figure 3. Perspective view of the model surface and workflow for constructing topographic cross
sections. (a) Acquisition of the data with spatial coordinates x, y, and z by laser stereoscopic technique.
(b) Surface generated from digitized points and identification of the overriding plate, oceanic plate, and
the oceanic plateau. (c) Cross section plane perpendicular to the trench allowing us to see the geometrical
and topographical evolution of the experiment.

cone putty. We vary the density of silicone putty to take into of the glucose syrup at a depth of 18 mm, and the syrup is
account the different lithosphere buoyancies. The upper free to move underneath. It pushes the overriding plate
mantle is modeled using a Newtonian low-viscosity glucose toward the trench. A regular grid initially drawn on the
syrup solution. We neglect the role of thermal convection, overriding plate allows us to measure the finite displace-
assuming that the flow within the syrup is only generated by ment field of each single node of the grid along the x and y
plate motion and slab subduction. We assume that the base directions at each time interval (Figure 2b). The cumulative
of the upper mantle (bottom of the Plexiglas tank) acts as an deformation of each element of the grid during the exper-
impermeable barrier where the slab remains temporarily iment is then calculated, integrating the finite displacement
stalled for the adopted experimental timescales (Figure 2a) field over the time. Experiments are monitored with lateral
[Funiciello et al., 2003]. The system is isothermal, which and top view photos, and the topography of some selected
implies that the density contrast between the lithosphere and experiments has been monitored with a topography scanner
the upper mantle is only chemically implemented and kept (Real Scan USB model 300) (Figure 3).
constant during the entire evolution of the experiment (see
the paper by Funiciello et al. [2003] for more details on the 2.2. Similarity Criteria
experimental procedure). [11] Experimental parameters and scaling relationships
[10] Silicone plates that represent the subducting and for the reference experiment (experiment 1) are listed in
overriding plates, respectively, initially float above the Tables 2 and 3, respectively. Materials have been selected to
dense glucose syrup (Figure 2b). The lateral distance respect the standard scaling procedure for length, density,
between plates and the sides of the Plexiglas tank containing viscosity, and stress in a natural gravity field (gModel = gNature)
the experiment is larger than the size of advected cells as described by Weijermars and Schmeling [1986] and Davy
within the upper mantle to minimize lateral boundary and Cobbold [1991].
effects. Thus we assume that lithospheric plates are com- [12] The scale factor for length is 1.5  107, i.e., 1 cm in
pletely surrounded by weak fault zones [Funiciello et al., the experiment corresponds to
66 km in nature. We model
2004]. We take into account neither lateral heterogeneities the subduction of wide and homogenous buoyant plates
within the plates nor the oblique plate convergence. During assuming that (1) the Inca Plateau and the Nazca Ridge
the subduction, the overriding and subducting plates are form a 1500-km-wide buoyant anomaly [Gutscher et al.,
separated by a thin layer of glucose syrup at trench. Hence 1999b] and (2) the Juan Fernandez Ridge was oriented
we assume that the subduction fault plane is weak, with the nearly parallel to the trench in the Lower Miocene [Yañez et
same viscosity as in the upper mantle. Moreover, we always al., 2001] (Figure 1b). In nature, aseismic ridges or oceanic
lubricate the complete surface of the subducting plate with plateaus correspond to topographic anomalies where the
Vaseline, so as to avoid plates sticking. In the initial crust is thicker than the surrounding oceanic plate, increas-
configuration, the leading edge of the silicone plate is forced ing the buoyancy of the oceanic lithosphere [Cloos, 1993;
downward to a depth of 3 cm (corresponding to about Molnar and Gray, 1979]. In our simplified models, we
200 km in nature) inside the glucose syrup to initiate the consider that the thickness of the lithosphere below aseismic
subduction process. In experiments 1 – 3 (Table 3), the ridges or oceanic plateaus is similar. The density of silicone
subduction velocity is not imposed by kinematic boundary plates is scaled to simulate the average density of the
conditions but only results from the negative buoyancy of corresponding natural lithospheric plates (Table 2), and
the subducting slab (i.e., slab pull) [Funiciello et al., 2003]. we reduce the density of the silicone lithospheric plate over
In contrast, in experiments 4 – 8 (Table 3), a constant its whole thickness. Such an approximation should not
convergence rate between subducting and overriding plates modify the general behavior of the subduction zone, as-
is imposed, advancing a rigid piston perpendicular to the suming that the oceanic crust is not delaminated from the
trench (Figure 2a). The piston is plunged into the upper part mantle during the process of subduction [Chemenda et al.,

5 of 19
TC3011 ESPURT ET AL.: FLAT SLAB AND UPPER PLATE DEFORMATION TC3011

2000]. In the reference experiment (experiment 1), the 3.2. Subduction of a Plateau Below a Pushed
buoyancy of the silicone plate modeling the dense oceanic Overriding Plate (Experiment 4 (Table 3))
plate is (Drd)Model = 80 kg m3, whereas the buoyancy of [16] In experiment 4, the overriding plate is pushed at a
the plate modeling the plateau-bearing oceanic lithosphere constant velocity of 0.31 mm s1 by a piston, moving
is (Drl)Model = +41 kg m3 (Table 2). If we consider that we perpendicularly toward the trench (Figure 5a). All the other
model a 50-Ma-old dense oceanic lithosphere (maximum parameters are similar to those of the reference experiment
age of the Nazca slab between 10 and 35°S) whose negative (experiment 1). The subducting plate being fixed, the trench
buoyancy is (Drl)Nature = – 35 kg m3 [Cloos, 1993], the velocity is the velocity of subduction (Figure 5b). In this
adopted buoyancy-scaling factor is (Drl)Model/(Drl)Nature = experiment, we observe that a fraction of the velocity
2.3. Thus the buoyancy of the plateau-bearing oceanic imposed by the piston is accommodated by the horizontal
lithosphere corresponding to the light silicone plate in shortening of the overriding plate, and the difference
nature is (Drl)Nature = +18 kg m3, which corresponds to between the piston velocity and trench velocity is the
a 50-Myr-old oceanic lithosphere with a 17-km-thick crust. overriding plate shortening rate (Figures 5b and 5d). How-
Such a plateau would result in a
1900-m-high bathymetric ever, the deformation of the overriding plate being slow
anomaly, similar to that of the Nazca Ridge (see the paper compared to the speed of the piston, the trench motion and
by Martinod et al. [2005] for further details concerning the the velocity of subduction are essentially constrained by the
scaling of buoyant anomalies). advance of the piston (Figure 5b).
[13] The ratio of the oceanic lithosphere viscosity over [17] During the subduction of the dense oceanic plate, the
the upper mantle viscosity is set to an upper bound of dip of the slab is shallower (
49°) (Figures 5a and 5c) than
roughly 104. From the simple scaling equations of Table 2, in the reference experiment because of the forced advancing
we find that 1 min corresponds to 1.4 Ma in nature. of the overriding plate toward the incoming plate. The
plateau reaches the trench after
14 min of run. Initially,
3. Experimental Results plateau subduction initiates without any major change in the
slab geometry (Figures 5a and 5c). When
5.5 cm of
[14] Eight models have been performed using various plateau have been subducted and the tip of the plateau has
combinations of geometrical, kinematical, and rheological reached a depth of
4 cm, the slab geometry reorganizes
parameters, with the aim to explore the subduction of a into two dip domains: The upper part becomes shallower,
negatively buoyant oceanic plate, followed by the subduc- while the lower part of the slab steepens to 57° (Figures 5a,
tion of a buoyant plate modeling a wide oceanic plateau 5c, and 6a). Seven minutes later (
10 Ma in nature), i.e.,
with or without an upper overriding plate (Figure 2 and Table when 11 cm of plateau have subducted (
730 km in
3). The evolution of three selected experiments is described nature), the plateau directly underplates the overriding plate,
in sections 3.1, 3.2, and 3.3 (experiments 1, 4, and 8 (Table 3)). and a flat slab segment occurs. During this phase, the slab
geometry presents three inflection points, showing that it is
3.1. Subduction of a Plateau Below a Free Overriding
easier to fold and unfold the slab to maintain the plateau
Plate (Reference Experiment, Experiment 1 (Table 3))
below the overriding plate than to increase the depth of the
[15] Experiment 1 shows the subduction of a fixed plateau in the inclined segment (Figures 5a, 5c, and 6a). The
oceanic lithosphere followed by a light oceanic plateau tip of the plateau is stabilized in the steep part of the slab at
beneath an overriding plate being free to move above the a depth of roughly 7 cm (
460 km). Until the end of the
subduction zone (Figure 4a). Martinod et al. [2005] have experiment, the length of the horizontal slab segment
already described the behavior of a similar experiment increases to finally reach a length of 4.2 cm, the edge of
without any upper overriding plate (e.g., experiments 2 the flat slab segment being located at 8.9 cm from the trench
and 3 in Table 3). Our experiment shows the typical (corresponding to
620 km in nature).
sequence of events already described by Martinod et al. [18] During the subduction of the dense oceanic plate, the
[2005], i.e., subduction initiation and slab interaction with overriding plate shortening is close to zero (Figures 5d and 7)
the base of the upper/lower mantle discontinuity followed because the imposed convergence velocity is close to the
by a steady state regime of subduction of the dense part of velocity of subduction in the corresponding free subduction
the slab (Figures 4a and 4b). During the steady state regime experiment (reference experiment). The advance of the
and as far as the dense part of the plate is concerned, overriding plate forces the subduction of the plateau, but
subduction is characterized by a velocity of subduction of the overriding plate shortening starts with a delay of about
0.3 mm s1 and a dip of 53°, as in the paper by Martinod et 2 min (Figure 5d). The overriding plate shortening of
4.5 
al. [2005]. As soon as the buoyant plateau starts subducting, 104 s1 shows that the process of subduction is now
the velocity of subduction suddenly decreases, and the slab opposed by the slab buoyancy. It results in a decrease of
steepens (Figures 4a, 4b, and 4c). Finally, the process can be the velocity of subduction (from 0.31 to 0.24 mm s1)
stopped when
4 cm of the buoyant plateau (corresponding (Figures 5b and 5d). Before the formation of the flat slab
to
280 km in nature) have been subducted, enhancing the segment, the overriding plate homogeneously shortens,
progressive verticalization of the slab (Figure 4a). During except close to the trench where the slab and the overriding
the whole experiment, the overriding plate passively follows plate touch closely (Figure 7). In this area (that would
the retreating trench (Figure 4a) and does not deform. correspond to the fore-arc region in nature), the force
exerted by the slab results in a trenchward decrease of the

6 of 19
TC3011 ESPURT ET AL.: FLAT SLAB AND UPPER PLATE DEFORMATION TC3011

Figure 4. Experiment 1. (a) Lateral views of the experiment. The dashed line marks the tip of the
oceanic plateau. The numbering scheme is as follows: (1) t = 01032 (the slab interacts with the bottom of
the box), (2) t = 06016 and 12004 (steady state subduction pursues at constant speed, and the slab dip
maintains a constant value of 53° until the beginning of plateau subduction), (3) t = 06016 and 12004
(steady state subduction pursues at constant speed, and the slab dip maintains a constant value of 53°
until the beginning of plateau subduction), (4) t = 18030 (the oceanic plateau entered in subduction,
decreasing the velocity of subduction and increasing the dip of the slab), and (5) t = 32050 (the buoyancy
of the oceanic plateau stops the subduction process). (b) Amount of subduction and (c) dip of the slab
versus time. Results of experiment 3 realized without any overriding plate are shown for comparison
(dashed lines with white triangles).

7 of 19
TC3011 ESPURT ET AL.: FLAT SLAB AND UPPER PLATE DEFORMATION TC3011

Figure 5. Experiment 4. (a) Lateral views of the experiment. The dashed line marks the tip of the
oceanic plateau. The numbering scheme is as follows: (1) t = 2052 (initiation of subduction), (2) t = 14019
(steady state dense oceanic subduction with a constant dip of 49°), (3) t = 16000 (subduction of the
plateau without any perturbation in the subduction process), (4) t = 17050 and 19026 (the plateau
buoyancy perturbs the subduction process, i.e., the dip of the upper part of the slab decreases, but the dip
of the lower part of the slab increases), and (5) t = 21002 to 27007. (For t = 21002 to 27007, the oceanic
plateau flattens and underplates the overriding plate. Note that the flat slab segment is shorter than the
length of the subducted oceanic plateau.) (b) Amount of subduction, (c) dip of the slab, and (d) overriding
plate strain versus time.

horizontal compressive stress within the overriding plate [19] Laser data of Figure 8 register the evolution of the
[Dalmayrac and Molnar, 1981], explaining why the plate overriding plate topography and location of vertical dis-
shortening is smaller there (Figure 7). When the plateau placements during the experiment. During the subduction of
underplates the overriding plate, the contact between con- the dense plate, the continent does not present any vertical
verging plates progressively spreads out. The locus of displacement. In contrast, the overriding plate topography is
maximum overriding plate shortening is above the termina- strongly affected by the plateau subduction. Uplift of the
tion of the flat slab segment and progressively migrates overriding plate close to the trench occurs as soon as the
within the overriding plate as the flat slab segment broadens plateau subducts, despite the fact that the overriding plate
(Figure 7). The shortening rate of the inner part of the shortening has not begun (Figures 5d, 8a, and 8b). After-
overriding plate (close to the piston) also increases, because wards, the uplift of the overriding plate continues at a
the wider contact between the converging plates results in constant and slower speed until the end of the experiment,
larger friction forces between them. and the uplifted area spreads out within the overriding plate
(back-arc region) (Figures 8a and 8b).

8 of 19
TC3011 ESPURT ET AL.: FLAT SLAB AND UPPER PLATE DEFORMATION TC3011

Figure 6. Kinematic behavior of the slab (dotted lines) in (a) experiment 4 and (b) experiment 8 with
different piston velocities. Flat subduction episode is in red.

3.3. Subduction of a Plateau Below a Faster steady state subduction of the dense oceanic plate (39°
Overriding Plate (Experiment 8 (Table 3)) versus 49° in experiment 4 (Figures 9a and 9c)). Following
[20] In experiment 8, the piston velocity is larger than in plateau subduction, the larger convergence velocity induces
the previous experiment (0.5 mm s1 versus 0.31 mm s1 in the rapid formation (4 min later, i.e.,
7 Ma in nature) of a
experiment 4). The higher trenchward overriding plate flat slab segment (Figures 9a and 9c). As in experiment 4,
velocity results in a decrease of the slab dip during the the flat slab segment required
11 cm of plateau subduction

Figure 7. Overriding plate strain rate fields parallel to the convergence direction in experiment 4 during
the subduction of the dense ocean (dashed line), the beginning of plateau subduction (dotted line), and
flat slab subduction (solid line).

9 of 19
TC3011 ESPURT ET AL.: FLAT SLAB AND UPPER PLATE DEFORMATION TC3011

Figure 8. (a) Cross sections of experiment 4 along the centerline of the overriding plate at different
times showing the propagation of the deformation within the overriding plate in response to the plateau
subduction. (b) Maximum elevation of the overriding plate topography versus time. See section 4.3 for
further explanations.

to occur, and the formation of the flat slab segment is [21] In this experiment, the shortening of the overriding
marked by an increase of the dip of the lower part of the plate starts as soon as the tip of the dense slab touches the
slab, here from 39° to 48° (Figures 9a and 9c). Here, again, bottom of the tank (Figures 9b and 9d). The higher
the tip of the plateau stabilizes within the steep part of the convergence velocity results in a compressional regime in
slab at a depth of
7 cm. the overriding plate during the steady state subduction of the

10 of 19
TC3011 ESPURT ET AL.: FLAT SLAB AND UPPER PLATE DEFORMATION TC3011

Figure 9. Experiment 8. (a) Lateral views of the experiment. The dashed line marks the tip of the
oceanic plateau. The numbering scheme is as follows: (1) t = 2027 (initiation of subduction), (2) t = 11030
(steady state subduction of the dense oceanic subduction with a constant dip of 39°), (3) t = 12035 to
14043 (plateau subduction without any perturbation in the subduction process), and (4) t = 15035 to 19011
(slab flattening and formation of a horizontal slab segment). (b) Amount of subduction, (c) dip of the
slab, and (d) overriding plate strain versus time.

dense plate, its shortening rate being approximately 2.3  (3) the shortening is moderate in the rest of the overriding
104 s1. Following plateau subduction, the shortening rate plate far from the trench.
of the overriding plate increases to
4.5  104 s1
(Figure 9b). As in experiment 4, three domains with
different strain regimes occur in the overriding plate,
4. Interpretation and Discussion: Insights on
following the occurrence of flat subduction: (1) The defor- the Andean Subduction Zone
mation is small close to the trench where the overriding and [22] The three lithospheric-scale analog experiments de-
subducting plates touch closely, (2) the shortening is max- scribed in sections 3.1, 3.2, and 3.3 have been selected to
imum above the termination of the flat slab segment, and

11 of 19
TC3011 ESPURT ET AL.: FLAT SLAB AND UPPER PLATE DEFORMATION TC3011

anomaly stops the process of subduction, producing a slab

verticalization (Figure 4a). Although the average buoyancy
of the slab is still slightly negative at the end of experiment 1,
its magnitude is too small to maintain the subduction active
[Martinod et al., 2005]. In natural cases, plateau subduction
decreases subduction velocity, and simultaneous slab steep-
ening may be followed by slab break-off (see the papers by
Regard et al. [2005] and Conrad et al. [2004] and Mediter-
ranean examples described by Wortel and Spakman [2000],
for instance). However, the simplified Newtonian viscous
rheology adopted to simulate the experimental lithosphere
does not favor the occurrence of this phenomenon in our
models.
[24] So far, we have only considered the specific case in
which plates’ motion only results from forces acting in the
subduction zone (slab pull force). However, in nature, the
absolute motion of lithospheric plates can be also influenced
by global far-field boundary conditions that are able to
partly control the velocity of subduction. For example, the
rapid westward motion of the South American lithosphere
above the Nazca plate [Marrett and Strecker, 2000] seems
to be controlled by the opening of the Atlantic Ocean and
the motion of Africa [Silver et al., 1998].
[25] Our experimental results suggest that the trench
migration imposed by the trenchward motion of the over-
riding plate does not modify the vertical component of slab
velocity. In the entire experimental set (see Table 3) and
during the steady state subduction of the dense oceanic
plate, we observe that Vz = Vsub tan(a) is proportional to
Figure 10. Dip of the slab (a) versus Dr/Vsub in Dr, where Vsub is the velocity of subduction, Vz the vertical
experiments 1 – 8 (see Table 3), where Dr is the negative component of the slab velocity, a the dip of the slab, and
buoyancy of the subducting plate, and Vsub is the velocity of Dr the slab buoyancy (see Figure 10). Hence experimental
subduction. The linear relationship between Dr/Vsub and results suggest that the appearance of flat slab segments
tan(a) shows that the vertical component of the slab cannot be produced even if we increase the overriding plate
velocity does not depend on the overriding plate motion in trenchward velocity as long as the slab is negatively buoy-
this set of experiments. ant, since the vertical component of the slab velocity is
maintained independently of trench motion. Heuret et al.
[2007] also observed that plate kinematics can strongly
show the evolution of the slab geometry as a function of the
affect the evolution of the subduction process. In particular,
overriding plate motion and subducting plate buoyancy. In
they noted that the motion of the overriding plate toward the
sections 4.1, 4.2, and 4.3, we show how their evolution
trench enhances trench rollback and slab shallowing, despite
gives insights to understand processes resulting in flat
the fact that the subducted lithosphere is dense and heavy
subductions beneath the Andes and the large-scale associ-
[Scholz and Campos, 1995; Heuret et al., 2007]. However,
ated tectonic processes within the overriding South Amer-
the dense slab can never reach a dip inferior to 18°, even if it
ican plate.
is strongly forced to subduct [Heuret et al., 2007].
4.1. Parameters Controlling the Flat Subduction [26] In fact, our experiments suggest that the presence of
a buoyant subducting plateau and an overriding plate
[23] Experimental results confirm that the subduction of a pushed toward the trench by lateral boundary conditions
buoyant lithosphere (such as an oceanic plateau) is not the are necessary for flat slab segments to occur (experiments 4
only element that results in the formation of a flat slab and 8) (Figure 6). For instance, the buoyant effect of the
segment [Martinod et al., 2005]. Experiment 1 shows that Inca Plateau and Nazca Ridge subduction beneath the
part of subducted buoyant plateaus can easily sink in the trenchward advancing of South America plate would be
upper mantle and be incorporated in the steep lower part of responsible for the 1500-km-wide Peruvian flat slab seg-
the slab (Figure 4a). The buoyancy of the plateau is ment [Gutscher et al., 1999b]. In southern South America,
compensated by the negative buoyancy of the dense oceanic the Juan Fernandez Ridge coincides with the central Chile/
slab. If the overriding plate is free to move, the dip of the NW Argentina flat slab segment [Pilger, 1981]. In these two
slab increases following plateau subduction as the velocity zones, the present-day flat slab segments are clearly illus-
of subduction decreases [Bellahsen et al., 2003; Martinod et trated by seismicity data (Figure 1c) [Gutscher et al., 2000b;
al., 2005]. Flat subduction cannot occur, and the buoyant Pardo et al., 2002]. Earthquakes at
120-km depth show

12 of 19
TC3011 ESPURT ET AL.: FLAT SLAB AND UPPER PLATE DEFORMATION TC3011

Table 1. Ridge/Plateau Subduction Characteristics Along the Andean Subduction Zone

Crustal Beginning of Length of
Ridge/ Width Height Thickness Symmetric on Subduction Subduction Flat Slab Upper Plate
Plateau (km) (km) (km) Origin Pacific Plate (Ma) (km) Segment Uplift

Carnegiea 280 2 13 – 19 Galapagos - 1
70(?) no yes

hot spot (only fore arc)
b
Inca
300 - - Pacific Marquesas 13
900(?) yes no
spreading ridge plateau
hot spot
Nazcac 200 1.5 18 ± 3 Pacific Tuamotu 11.2
900 yes yes
spreading ridge plateau
hot spot
Iquiqued 130 1.5 15 Pacific Austral 2
150(?) no no
spreading ridge plateau
hot spot
J. Fernandeze <100 4 15 hot spot located - 12
850 yes yes
near the Alejandro
Selkirk island
a
[Lonsdale and Klitgord, 1978; De Vries, 1988; Deniaud et al., 1999; Gutscher et al., 1999a; Guillier et al., 2001; Spikings et al., 2001; Sallarès et al.,
2005; Pedoja et al., 2006].
b
[Gutscher et al., 1999b].
c
[Soler and Bonhomme, 1990; Dumont, 1996; Hsu, 1992; Macharé and Ortlieb, 1992; Gutscher et al., 1999b; Woods and Okal, 1994; Le Roux et al.,
2000; Latrubesse and Rancy, 2000; McNulty and Farber, 2002; Hampel, 2002; Rousse et al., 2003; Hampel et al., 2004a, 2004b; Rosenbaum et al., 2005;
Tassara et al., 2006; Espurt et al., 2007].
d
[Gutscher et al., 1999b; Rosenbaum et al., 2005; Tassara et al., 2006].
e
[Allmendinger et al., 1983; Jordan and Alonso, 1987; Reynolds et al., 1990; von Huene et al., 1996; Pardo et al., 2002; Ramos et al., 2002; Kopp et al.,
2004; Tassara et al., 2006; Farias et al., 2008].

the flat slab segments advancing beneath the overriding have been larger. Thus we do not think it is necessary to
plates at more than 600 km east of the trench. In contrast, involve upper mantle hydration to explain the central Chile/
the eastern intermediate slab (between 200 and 600 km) is NW Argentina flat subduction zone, although this phenom-
steeper (
60°) than elsewhere beneath South America enon may undeniably result in lighter oceanic lithosphere
(
30°). The slab flattening is accompanied by an increase [Kopp et al., 2004].
of the dip of the intermediate part of the slab, as observed in
experiments 4 and 8 (Figure 6). 4.2. Timing of the Flat Subduction
[27] Kopp et al. [2004] recently found that the Juan
Fernandez Ridge may not have sufficient buoyancy to [28] The appearance of a flat slab segment is a process
generate a flat slab segment. The Juan Fernandez Ridge is that initiates after several Ma following the beginning of
a discontinuous topographic anomaly
4 times smaller than plateau subduction (experiment 4). This delay decreases
the Nazca Ridge (see Figure 1a and Table 1). Kopp et al. when the overriding plate velocity increases (experiment 8).
[2004] suggest that upper mantle hydration along ridge- In fact, experiments show that slab flattening requires time
parallel fault zones decreases the density of the oceanic to occur because at the beginning of plateau subduction its
plate and consequently favors the buoyancy of the slab. In buoyancy is compensated by the negative buoyancy of the
fact, the reconstruction of the Juan Fernandez Ridge geom- lower part of the slab (Figures 11 and 12). We observe that
etry suggests a 20°N ridge segment (i.e., almost trench flat subduction occurs when the average buoyancy of the
parallel), which subducted below northern Chile between 25 inclined part of the slab turns positive. The delay between
and 12 Ma [Yañez et al., 2001]. The effect of trench-parallel the beginning of plateau subduction and the formation of a
buoyant anomalies on the dynamics of subduction is sig- flat slab segment has also been observed by van Hunen et
nificantly larger than the effect of trench-perpendicular al. [2002a, 2002b] (Figure 11). The delay is larger in our
ridges [see Martinod et al., 2005]. Thus this ridge segment experimental set, probably because the numerical setup of
may also explain the formation of the central Chile/NW van Hunen et al. [2002a, 2002b] does not take into account
Argentina flat slab segment in spite of the modest size of the the negative buoyancy of the slab below 400-km depth.
Juan Fernandez Ridge. In the absence of any symmetric of [29] The Nazca Ridge subduction beneath the Peruvian
the Juan Fernandez Ridge on the Pacific plate, it is impos- coast started
11.2 Ma ago [Hampel, 2002], and the
sible to evaluate the volume of the ridge segment that volcanism has been inactive above the ridge segment since
subducted since the Lower Miocene below Chile. Despite
4 Ma, in response to the flattening of the slab (Figure 12)
the fact that the volume of the Juan Fernandez Ridge [Soler and Bonhomme, 1990; Rosenbaum et al., 2005;
segment now visible in the Pacific seafloor is modest (see Espurt et al., 2007]. Yañez et al. [2001] showed that near
Figure 1a), the volume of magma that resulted from the 30°S, the Juan Fernandez Ridge subduction started approx-
Juan Fernandez hot spot activity (Table 1) in the past may imately 12 Ma ago. Its subduction is followed by the end of

13 of 19
TC3011 ESPURT ET AL.: FLAT SLAB AND UPPER PLATE DEFORMATION TC3011

Figure 11. Length of the flat slab segment versus length of subducted plateau in experiments 4 and 8.
Numerical results without overriding plate motion [van Hunen et al., 2002a] and with overriding plate
motion [van Hunen et al., 2002b] are shown for comparison.

the arc volcanism at
5 Ma [Ramos et al., 2002]. Thus the slab segment is always much shorter than the amount of
delay between the subduction of the buoyant anomalies and subducted plateau (Figures 5a and 9a). Numerical experi-
the appearance of the two present-day flat slab segments ments of Hassani et al. [1997] also show that the slab
below South America is
7 Ma, which is equivalent to the flattening only starts when a sufficient length of buoyant
delay observed in the experiments (Figure 12). It may also slab is engaged in the subduction process despite the fact
explain why broader ridges such as the Carnegie or Iquique that their models do not incorporate preexisting dense
ridges do not result in the formation of any modern flat subducting slab.
subduction zone. For Lonsdale and Klitgord [1978], the [31] These experimental results are consistent with the
Carnegie Ridge subduction only began 1 Ma ago. The subducted length of the Nazca Ridge below Peru. In fact,
analysis of seismicity distribution suggests that the slab the paleomagnetic reconstruction of the Nazca Ridge sub-
plunges below northern Ecuador with a dip close to 30° duction shows that the subducted length of the Nazca Ridge
down to a depth of at least 200 km [Guillier et al., 2001] below South America is
900 km [Hampel, 2002]. The flat
(Figure 1c). In fact, the broadening of the volcanic zone and slab segment being
700 km long,
200 km of ridge are
the magmatic evolution in northern Ecuador [Gutscher et located in the steep part of the slab, as observed in experi-
al., 1999a; Bourdon et al., 2003] suggest that slab flattening ments 4 and 8 (Figures 4a and 9a).
is active but not achieved, because the subduction of the
Carnegie Ridge would be too recent. The Iquique Ridge 4.3. Deformation of the Overriding Plate
may also be associated with a steep subducting slab below [32] Our models offer the possibility to study relation-
northern Chile because its subduction only began
2 Ma ships between the subducting plate buoyancy and the
ago [Rosenbaum et al., 2005] (Figure 1c). related overriding plate deformation (Figure 12). When
[30] The comparison between experiments 4 and 8 shows the trenchward overriding plate velocity is equal to the
that the amount of subducted plateau when flat subduction spontaneous trench retreat velocity, the overriding plate
initiates is similar (
11 cm) despite different convergence does not deform (experiments 1 and 4). The overriding
velocities (Figure 11). These experiments are scaled to plate shortens when kinematic boundary conditions impose
represent the subduction of a 50-Ma-old oceanic plate larger convergence velocities and when buoyant plateaus
overlain by a
2-km-high plateau (Table 2), and the flat are forced to subduct (experiments 4 and 8). This agrees
subduction occurs following the equivalent of
700 km of with the numerical results by Buiter et al. [2001] and
plateau subduction (Figure 11). Part of the oceanic plateau Hampel and Pfiffner [2006].
is incorporated in the lower steep slab segment, and the flat

14 of 19
TC3011 ESPURT ET AL.: FLAT SLAB AND UPPER PLATE DEFORMATION TC3011

Figure 12. Schematic sketches showing evolution of the flat slab subduction process illustrated by
analog models, applied to the Andean subduction zone (modified after Gutscher et al. [2000a]). (a) Steep
subduction (>30° dip) during the subduction of the dense Nazca plate. (b) Subduction of buoyant and
overthickened oceanic crust (e.g., ridge or oceanic plateau) with uplift of the fore-arc area. The buoyant
slab is forced into subduction by the trenchward motion of the overriding plate. The dip of the slab
decreases, and the volcanism migrates landward. (c) The upper slab starts to underplate the overriding
plate, while the lower slab is steeper. Experiments suggest that it is necessary to involve a large amount of
buoyant plate to obtain a flat slab segment, since part of the buoyant plateau is incorporated in the steep
part of the slab to balance the negative buoyancy of the dense oceanic lithosphere. The strong plate
contact is translated by the increase of the overriding plate shortening. The flat slab process cancels the
vertical component of the slab favoring the uplift of the overriding plate. (d) Propagation of the flat slab
segment beneath the overriding plate. The horizontal and buoyancy forces transmitted to the back-arc
region cause back arc thrusting and uplift far from the trench.

[33] The topographic data presented here are purely horizontal forces transmitted to the back-arc region may
qualitative. The effects of erosion and sedimentation have activate inherited weak zones and cause back arc thrusting
not been considered. Experiment 4 shows that the subduc- in natural cases.
tion of a plateau does not increase immediately the over- [34] These experimental results show that the flat sub-
riding plate shortening rate because plateau subduction does duction process induces a widespread shortening and uplift
not result immediately in a decrease of the slab pull force in the overriding plate, which is consistent with geological
(Figure 5d). In contrast, close to the trench, the overriding observations along the Andean subduction zone. The Ecua-
plate topography uplifts as soon as the plateau enters in the dorian fore arc exhibits Pliocene to Pleistocene marine
subduction zone (Figure 8a). This is interpreted as an terraces (so-called tablazos), which are exposed at 200–
isostatic adjustment accommodating the subduction of the 300 m above the Carnegie Ridge [De Vries, 1988; Deniaud
plateau [Moretti and Ngokwey, 1985], while the slowly et al., 1999; Pedoja et al., 2006]. Gutscher et al. [1999a]
growing long-term uplift observed in the topographic high also show that the Carnegie Ridge is linked to a regional
developing close to the trench results from overriding plate uplift of the fore arc more than 110 km landward. The
thickening (Figure 8b). The shortening rate in the overriding subduction of the Carnegie Ridge is uplifting the coastal
plate also increases following the appearance of the flat slab area, even if the flat subduction process is not achieved
segment (Figure 4d). Cross sections demonstrate that over- (Figure 12b). In Peru, the southward migration of the Nazca
riding plate thickens and uplifts far from the trench, in the Ridge between 11 and 17°S has strongly influenced the
back-arc region, because the longer contact increases fric- geomorphology and tectonic activity of the Peruvian fore
tion forces between plates (Figures 7 and 8). Consequently, arc system [Sébrier et al., 1985; Hsu, 1992; Macharé and

15 of 19
TC3011 ESPURT ET AL.: FLAT SLAB AND UPPER PLATE DEFORMATION TC3011

Table 2. Scaling of Parameters for the Reference Experiment (Experiment 1) (Table 3)

Parameter Meaning Unit Model Nature

g gravitational acceleration m s2 9.81 9.81

Thicknesses
h silicone/lithospheric plates m 0.013 80,000
H glucose syrup/upper mantle 0.10 660,000
length scale factor
LModel/Nature = 1.5  107
Buoyancies
Drd = rgs - rd 50-Ma-old oceanic lithosphere kg m3 80 35
Drp = rgs - rp 50-Ma-old ocean, 1900-m-high +41 +18
plateau
Drop = rgs - rop overriding plate +102 +45
buoyancy scale factor (normal continent:
DrModel/DrNature = 2.3 +96)
Viscosities
hd dense subducting silicone/ Pa s 3.5  105
1024
50-Ma-old oceanic lithosphere
hp oceanic plateau 1.8  105
1024
hop overriding lithosphere 1.6  105
1023
hgs upper mantle 22
1020
hlm lower mantle 1
1022
viscosity scale factor
hModel/hNature
2  1019
t: (tModel/tNature) = (DrgL)Nature/ characteristic time s 42 (42 s) 3.16  1013
(DrgL)Model  (hModel/hNature) = (1 Ma)
1.3  1012
U: (UModel/UNature) = (tNature  characteristic velocity m s1 3.3  104 2.9  109
LModel)/(tModel  LNature) = (0.33 mm s1) (9.3 cm a1)
1.1  105

Ortlieb, 1992; Le Roux et al., 2000; Hampel, 2002; Clift et [Macharé and Ortlieb, 1992]. In the Andes, the Nazca
al., 2003; Hampel et al., 2004b; Wipf, 2006]. Pliocene to Ridge flat segment may control the shape of the modern
Pleistocene marine terraces developed above the Nazca topography of the cordillera. For example, the recent
Ridge indicate an uplift of
900 m of the coastal margin extensional collapse in the Cordillera Blanca [McNulty

Table 3. Experimental Parametersa

Experiment

1 2 3 4 5 6 7 8

Width of the Plates (W) 0.305 0.2 0.3 0.305 0.3 0.3 0.3 0.3
Thickness of Silicone Plates (h) 0.013 0.012 0.013 0.013 0.013 0.012 0.013 0.013
Thickness of Glucose Syrup (H) 0.1 0.104 0.1 0.1 0.1 0.1 0.1 0.1
Piston Velocity 0 0 0 0.31 0.31 0.31 0.16 0.5

Dense Subducting Plate

Length (Ld) 0.32 0.4 0.37 0.3 0.3 0.3 0.3 0.3
Density (rd) 1496 1497 1496 1496 1496 1434 1434 1496
Viscosity (hd) 3.5  105 2.7  105 3.5  105 3.5  105 3.5  105 3.5  105 3.5  105 3.5  105

Light Subducting Plate ‘‘Plateau’’

Length (Lp) 0.17 0.2 0.22 0.25 0.25 0.25 0.25 0.25
Density (rp) 1375 1416 1375 1375 1375 1375 1375 1375
Viscosity (hp) 1.8  105 2.6  105 1.8  105 1.8  105 1.8  105 1.8  105 1.8  105 1.8  105

Overriding Plate
Length (Lop) 0.225 - - 0.18 0.19 0.2 0.2 0.15
Density (rop) 1314 - - 1314 1314 1314 1314 1314
Viscosity (hop) 1.6  105 - - 1.6  105 1.6  105 1.6  105 1.6  105 1.6  105

Glucose Syrup
Density (rgs) 1416 1422 1416 1416 1416 1416 1416 1416
Viscosity (hgs) 22 12 22 22 22 22 22 22
a
See also Figure 2 for the description of W, h, H, Ld, Lp, and Lop. Lengths are given in m, viscosities are given in Pa s, densities are given in kg m3, and
velocities are given in mm s1. The evolutions of experiments 1, 4, and 8 are described in details in section 3. The results of experiments 2, 3, 5, 6, and 7 are
used to build Figure 10.

16 of 19
TC3011 ESPURT ET AL.: FLAT SLAB AND UPPER PLATE DEFORMATION TC3011

and Farber, 2002] and Neogene counterclockwise rotations segments also increase friction between plates, and the plot
[Rousse et al., 2003] would be related to the Nazca Ridge of the Nazca-Farallon/South America relative motion indi-
(Figure 12c). Similarly, the subduction of the Juan Fernan- cates two periods of slow convergence rates: one in the
dez Ridge below central Chile is marked by a Pliocene fore Oligocene between 35 and 26 Ma and the other one in the
arc uplift [Farias et al., 2008]. last 10 Ma [Pardo-Casas and Molnar, 1987]. These two
[35] In northern South America, part of the Amazonian periods are marked by the formation of large flat slab
basin located in front of the Peruvian flat slab segment segments, below southern Peru and northern Chile in the
displays deformation patterns related to the advance of the Oligocene [James and Sacks, 1999] and below Peru and
flat slab segment (Figures 1c and 12c) [Dumont, 1996; central Chile/NW Argentina from the Pliocene, suggesting a
Latrubesse and Rancy, 2000; Espurt et al., 2007]. This zone significant effect of the formation of flat segments on the
exhibits very wide sub-Andes, several uplifted basement subduction velocity.
blocks (Shira Mountains, Cushabatay High, and Moa Divi-
sor range), and regional uplift (Fitzcarrald Arch uplift)
above the eastern edge of the Nazca Ridge flat slab segment
5. Conclusions
[Espurt et al., 2007] (Figure 12d). Similarly, the eastward [38] Lithospheric-scale analog experiments suggest that
migration of the Juan Fernandez flat slab segment spatially the appearance of flat slab segments can result only from
coincides with the propagation of the deformation front of large buoyant ridges/plateaus forced to subduct by the
the precordillera thrust belt and the Pliocene uplift of the trenchward advance of the overriding plate, as previously
Sierras Pampeanas [Allmendinger et al., 1983; Ramos et al., suggested by Gutscher et al. [2000b] and van Hunen et al.
2002; Giambiagi and Ramos, 2002; Siame et al., 2005] [2002a, 2002b]. As a matter of fact, when negatively
(Figure 12d). In addition, the subduction of the northern buoyant slab segments subduct below a trenchward advanc-
trench-parallel segment of the Juan Fernandez Ridge may ing overriding plate, our experiments suggest that the
explain the excess of 100 km of shortening exhibited by overriding plate motion does not modify the vertical com-
balanced cross sections during the Neogene to the north of ponent of the slab velocity. Then, although increasing the
the present-day flat slab segment [Allmendinger et al., 1983; overriding plate’s trenchward motion results in a decrease of
Jordan and Alonso, 1987; Reynolds et al., 1990]. Similarly, the dip of the slab, it is unable to create flat slabs, as
Rousse et al. [2003] also suggest that the subduction of the observed below Peru and central Chile/NW Argentina.
northern segment of the Juan Fernandez Ridge during the [39] In our experiments modeling the subduction of a
Lower Miocene may explain the particularly intense short-
2-km-high plateau on a 50-Ma-old oceanic lithosphere,
ening in the Bolivian Orocline at that time. flat subduction only occurs after
700 km of buoyant
[36] We show that vertical displacements above the plateau subduction because part of the plateau is incorpo-
Andean flat slab segments are essentially deciphered in rated in the steep part of the slab to balance the negative
the fore-arc and back-arc regions. In fact, models suggest buoyancy of the dense underlying oceanic slab. Hence the
that vertical topographic motions related to the flat slabs delay between the plateau subduction and the flattening of
would be observed in the entire of the Andean Cordillera, if the slab is controlled by the trenchward velocity of the
we could eliminate the mountain-building processes. overriding plate.
[37] Finally, experiments suggest that the flat subduction, [40] In South America, the Peruvian flat slab segments
increasing the friction force between plates, may result in a occurred
7 Ma, following the beginning of the Nazca
decrease of the subduction velocity. This phenomenon is not Ridge-Inca Plateau subduction. The same delay is observed
clearly observed here because the boundary conditions we between the subduction of the Juan Fernandez Ridge and
apply are purely kinematic and because in these experi- the appearance of flat subduction below central Chile/NW
ments, the length of the superficial part of both plates is Argentina. In northern Ecuador and northern Chile, the
small when the flat slab segment occurs. Gutscher et al. process of the slab flattening could be active but not
[2000b] note that the seismic energy released within South completed, supporting the idea that the slab flattening needs
America is 3– 5 times larger above flat slab segments than time to occur.
above normal inclined segments. Seismicity clusters ob- [41] The overriding plate shortening rate increases fol-
served in the overriding plate of Peru and Argentina, east of lowing the subduction of an oceanic plateau as a result of
the flat slab segments, are consistent with both the greater the smaller slab pull force. In addition, it increases follow-
contact and friction between plates (Figures 1c and 12d) ing the appearance of the flat slab because of a larger
[Gutscher et al., 2000b]. Thus the formation of two major interplate friction force. The locus of maximum shortening
flat slab segments during the Pliocene below the South migrates within the interior of the overriding plate when the
American plate may explain the decrease of the eastward slab flattens.
velocity of the Nazca plate during the last
5 Ma [e.g., [42] The overriding plate topography close to the trench
Pardo-Casas and Molnar, 1987; Somoza, 1998; Norabuena is immediately affected by the subduction of a buoyant
et al., 1999]. This decreasing velocity has been attributed by plateau. We observe that two processes control the elevation
Iaffaldano et al. [2006] to the uplift of the Altiplano Plateau, of the overriding plate close to the active margin: (1) isostatic
the load of which would consume a significant amount of adjustments resulting in rapid topographic changes, depend-
the driving force responsible for the trenchward motion of ing on the buoyancy of the subducting plate at trench, and
the Nazca plate. In fact, the experiments show that flat slab (2) overriding plate thickening. Finally, the appearance of

17 of 19
TC3011 ESPURT ET AL.: FLAT SLAB AND UPPER PLATE DEFORMATION TC3011

two major flat slab segments below South America
5 Ma the National Research Council of Italy, and by other national funding
agencies participating in the Third Memorandum of Understanding, as well
ago may thus explain the decrease of the convergence as from the EC Sixth Framework Programme. Analog models have been
velocity between the Nazca plate and the South American performed in the Laboratory of Experimental Tectonics (LET) of University
plate since that time. ‘‘Roma TRE’’. Midland Valley is acknowledged for providing ‘‘3DMove’’
for model building and 3D visualization. We also thank M. Espurt, G. Saldi,
and R. Martinez for their useful comments and suggestions. Constructive
[43] Acknowledgments. This research project has been financed by reviews by O. Oncken and an anonymous reviewer greatly improved this
the French INSU-CNRS (Institut National des Sciences de l’Univers-Centre manuscript.
National de la Recherche Scientifique) program DyETI (Dynamique et
Evolution de la Terre Interne). F. Funiciello has been supported by Euro-
horcs/ESF-European Young Investigators Awards Scheme, by funds from

References
Allmendinger, R. W., V. A. Ramos, T. E. Jordan, Deniaud, Y., P. Baby, C. Basile, M. Ordoñez, G. Montenegro, phys. Res. Lett., 28, 3749 – 3752, doi:10.1029/
M. Palma, and B. L. Isacks (1983), Paleogeography and G. Mascle (1999), Ouverture et évolution 2001GL013257.
and Andean structural geometry, northwest Argentina, tectono-sédimentaire du golfe de Guayaquil: Bassin Gutscher, M.-A. (2002), Andean subduction styles and
Tectonics, 2, 1 – 16, doi:10.1029/TC002i001p00001. d’avant-arc néogène et quaternaire du sud des their effect on thermal structure and interplate coupling,
Baranzangi, M., and B. L. Isacks (1979), Subduction of Andes équatoriennes, C. R. Acad. Sci. Ser. 2, 328, J. South Am. Earth Sci., 15, 3 – 10, doi:10.1016/
the Nazca plate beneath Peru: Evidence from spatial 181 – 187. S0895-9811(02)00002-0.
distribution of earthquakes, Geophys. J. R. Astron. DeVries, T. J. (1988), The geology of late Cenozoic Gutscher, M.-A., J. Malavieille, S. Lallemand, and J. Y.
Soc., 57, 537 – 555. marine terraces (tablazos) in northwestern Peru, Collot (1999a), Tectonic segmentation of the North
Bellahsen, N., C. Faccenna, F. Funiciello, J. M. Daniel, J. South Am. Earth Sci., 1, 121 – 136, Andean margin: Impact of the Carnegie Ridge col-
and L. Jolivet (2003), Why did Arabia separate doi:10.1016/0895-9811(88)90030-2. lision, Earth Planet. Sci. Lett., 168, 255 – 270,
from Africa? Insights from 3-D laboratory experi- Dominguez, S., S. E. Lallemand, J. Malavieille, and doi:10.1016/S0012-821X(99)00060-6.
ments, Earth Planet. Sci. Lett., 216, 365 – 381, R. von Huene (1998), Upper plate deformation Gutscher, M.-A., J. L. Olivet, D. Aslanian, J. P. Eissen,
doi:10.1016/S0012-821X(03)00516-8. associated with seamount subduction, Tectonophysics, and R. Maury (1999b), The ‘‘lost Inca Plateau’’:
Bourdon, E., J. P. Eissen, M.-A. Gutscher, M. Monzier, 293, 207 – 224, doi:10.1016/S0040-1951(98)00086-9. Cause of flat subduction beneath Peru?, Earth
M. L. Hall, and J. Cotten (2003), Magmatic re- Dumont, J. F. (1996), Neotectonics of the Subandes- Planet. Sci. Lett., 171, 335 – 341, doi:10.1016/
sponse to early aseismic ridge subduction: The Brazilian craton boundary using geomorphological S0012-821X(99)00153-3.
Ecuadorian margin case (South America), Earth data: The Marañon and Beni basins, Tectono- Gutscher, M.-A., R. Maury, J. P. Eissen, and E. Bourdon
Planet. Sci. Lett., 205, 123 – 138, doi:10.1016/ physics, 259, 137 – 151, doi:10.1016/0040- (2000a), Can slab melting be caused by flat subduc-
S0012-821X(02)01024-5. 1951(95)00200-6. tion?, Geology, 28, 535 – 538, doi:10.1130/0091-
Buiter, S. J. H., R. Govers, and M. J. R. Wortel (2001), Engdalh, E. R., R. D. van der Hilst, and R. Buland 7613(2000)28<535:CSMBCB>2.0.CO;2.
A modelling study of vertical surface displacements (1998), Global teleseismic earthquake relocation Gutscher, M.-A., W. Spakman, H. Bijwaard, and E. R.
at convergent plate margins, Geophys. J. Int., 147, with improved travel times and procedures for Engdalh (2000b), Geodynamics of flat subduction:
415 – 427, doi:10.1046/j.1365-246X.2001.00545.x. depth determination, Bull. Seismol. Soc. Am., 88, Seismicity and tomographic constraints from the
Cahill, T., and B. L. Isacks (1992), Seismicity and 722 – 743. Andean margin, Te ctoni cs, 19, 814 – 833,
shape of the subducted Nazca plate, J. Geophys. Espurt, N., P. Baby, S. Brusset, M. Roddaz, W. Hermoza, doi:10.1029/1999TC001152.
Res., 97, 17,503 – 17,529, doi:10.1029/92JB00493. V. Regard, P.-O. Antoine, R. Salas-Gismondi, and Hampel, A. (2002), The migration history of the Nazca
Chemenda, A. I., J. P. Burg, and M. Mattauer (2000), R. Bolaños (2007), How does the Nazca Ridge sub- Ridge along the Peruvian active margin: A re-eva-
Evolutionary model of the Himalaya-Tibet system: duction influence the modern Amazonian foreland luation, Earth Planet. Sci. Lett., 203, 665 – 679,
Geopoem based on new modelling, geological and basin?, Geology, 35, 515 – 518, doi:10.1130/ doi:10.1016/S0012-821X(02)00859-2.
geophysical data, Earth Planet. Sci. Lett., 174, G23237A.1. Hampel, A., and A. Pfiffner (2006), Relative importance
397 – 409, doi:10.1016/S0012-821X(99)00277-0. Farias, M., R. Charrier, S. Carretier, J. Martinod, of trenchward upper plate motion and friction along
Chung, W. Y., and H. A. Kanamori (1978), Mechanical A. Fock, D. Campbell, J. Caceres, and D. Comte the plate interface for the topographic evolution of
model for plate deformation associated with aseis- (2008), Late Miocene high and rapid surface uplift subduction-related mountain belts, in Analogue and
mic ridge subduction in New Hebrides Arc, Tecto- and its erosional response in the Andes of central Numerical Modelling of Crustal-Scale Processes,
nophysics, 50, 20 – 40. Chile (33° – 35°S), Tectonics, 27, TC1005, edited by S. J. H. Buiter and G. Schreurs, Spec.
Clift, P. D., I. Pecher, N. Kukowski, and A. Hampel doi:10.1029/2006TC002046. Publ. Geol. Soc. London, 253, 105-115.
(2003), Tectonic erosion of the Peruvian forearc, Funiciello, F., C. Faccenna, D. Giardini, and K. Regenauer- Hampel, A., N. Kukowski, J. Bialas, C. Huebscher, and
Lima Basin, by subduction and Nazca Ridge Lieb (2003), Dynamics of retreating slabs: 2. Insights R. Heinbockel (2004a), Ridge subduction at an ero-
collision, Tectonics, 22(3), 1023, doi:10.1029/ from three-dimensional laboratory experiments, sive margin: The collision zone of the Nazca Ridge
2002TC001386. J. Geophys. Res., 108(B4), 2207, doi:10.1029/ in southern Peru, J. Geophys. Res., 109, B02101,
Cloos, M. (1993), Lithospheric buoyancy and colli- 2001JB000896. doi:10.1029/2003JB002593.
sional orogenesis-subduction of oceanic plateaus, Funiciello, F., C. Faccenna, and D. Giardini (2004), Hampel, A., J. Adam, and N. Kukowski (2004b), Re-
continental margins, island arcs, spreading ridges, Role of lateral mantle flow in the evolution of sub- sponse of the tectonically erosive south Peruvian
and seamount, Geol. Soc. Am. Bull., 105, 715 – 737, duction systems: Insights from laboratory experi- forearc to subduction of the Nazca Ridge: Analysis
doi:10.1130/0016-7606(1993)105<0715:LBACOS> ments, Geophys. J. Int., 157, 1393 – 1406, of three-dimensional analogue experiments, Tec-
2.3.CO;2. doi:10.1111/j.1365-246X.2004.02313.x. tonics, 23, TC5003, doi:10.1029/2003TC001585.
Collot, J. Y., J. Daniel, and R. V. Burne (1985), Recent Geist, E. L., M. A. Fisher, and D. W. Scoll (1993), Hassani, R., D. Jongmans, and J. Chery (1997), Study
tectonics associated with the subduction/collision of Large-scale deformation associated with ridge subduc- of plate deformation and stress in subduction pro-
the Entrecasteaux zone in the central New Hebrides, tion, Geophys. J. Int., 115, 344 – 366, doi:10.1111/ cesses using two-dimensional numerical models,
Tectonophysics, 112, 325 – 356, doi:10.1016/0040- j.1365-246X.1993.tb01191.x. J. Geophys. Res., 102, 17,951 – 17,965, doi:10.1029/
1951(85)90185-4. Giambiagi, L. B., and V. A. Ramos (2002), Structural 97JB01354.
Conrad, C. P., S. Bielk, and C. Lithgow-Bertelloni evolution of the Andes in a transitional zone be- Heuret, A., F. Funiciello, C. Faccenna, and S. Lallemand
(2004), Great earthquakes and slab pull: Interaction tween flat and normal subduction (33°30 0 – (2007), Plate kinematics, slab shape and back-arc
between seismic coupling and plate-slab coupling, 33°450S), Argentina and Chile, J. South Am. Earth stress: A comparison between laboratory models and
Earth Planet. Sci. Lett., 218, 109 – 122, doi:10.1016/ Sci., 5, 101 – 116, doi:10.1016/S0895-9811(02) current subduction zones, Earth Planet. Sci. Lett., 256,
S0012-821X(03)00643-5. 00008-1. 473 – 483, doi:10.1016/j.epsl.2007.02.004.
Dalmayrac, B., and P. Molnar (1981), Parallel thrust Gripp, A. E., and R. G. Gordon (2002), Young tracks of Hsu, J. T. (1992), Quaternary uplift of the Peruvian
and normal faulting in Peru and constraints on the hotspots and current plate velocities, Geophys. coast related to the subduction of the Nazca Ridge:
state of stress, Earth Planet. Sci. Lett., 55, 473 – J. Int., 150, 321 – 361, doi:10.1046/j.1365-246X.2002. 13.5 to 15.6 degrees south latitude, Quaternary Int.,
481, doi:10.1016/0012-821X(81)90174-6. 01627.x. 15 – 16, 87 – 97, doi:10.1016/1040-6182(92)90038-4.
Davy, P., and P. R. Cobbold (1991), Experiments on Guillier, B., J. L. Chatelain, E. Jaillard, H. Yepes, Iaffaldano, G., H.-P. Bunge, and T. H. Dixon (2006),
shortening of a 4-layer continental lithosphere, G. Poupinet, and J. F. Fels (2001), Seismological Feedback between mountain belt growth and plate
Tectonophysics, 188, 1 – 25, doi:10.1016/0040- evidence on the geometry of the orogenic system in convergence, Geology, 34, 893 – 896, doi:10.1130/
1951(91)90311-F. central-northern Ecuador (South America), Geo- G22661.1.

18 of 19
 TC3011 ESPURT ET AL.: FLAT SLAB AND UPPER PLATE DEFORMATION TC3011

James, D. E., and I. S. Sacks (1999), Cenozoic forma- Andean Convergence, edited by L. D. Kulm et al., Somoza, R. (1998), Updated Nazca (Farallon)-South
tion of the central Andes: A geophysical perspec- Mem. Geol. Soc. Am., 154, 729-740. America relative motions during the last 40 My:
tive, in Geology and Ore Deposits of the Central Pardo, M., D. Comte, and T. Monfret (2002), Seismo- Implications for mountain building in the central
Andes, vol. 7, edited by B. J. Skinner, pp. 1 – 26, tectonic and stress distribution in the central Chile Andean region, J. South Am. Earth Sci., 11, 211 –
Soc. Econ. Geol., Littleton, Colo. subduction zone, J. South Am, Earth Sci., 15, 11 – 215, doi:10.1016/S0895-9811(98)00012-1.
Jarrard, R. D. (1986), Causes of compression and ex- 22, doi:10.1016/S0895-9811(02)00003-2. Spikings, R. A., W. Winkler, D. Seward, and R. Handler
tension behind trenches, Tectonophysics, 132, 89 – Pardo-Casas, F., and P. Molnar (1987), Relative motion (2001), Along-strike variations in the thermal and
102, doi:10.1016/0040-1951(86)90027-2. of the Nazca (Farallón) and South America plate tectonic response of the continental Ecuadorian An-
Jordan, T. E., and R. W. Allmendinger (1986), The since late Cretaceous times, Tectonics, 6, 233 – des to the collision with heterogeneous oceanic crust,
Sierras Pampeanas of Argentina: A modern analo- 248, doi:10.1029/TC006i003p00233. Earth Planet.Sci. Lett., 186, 57 – 73, doi:10.1016/
gue of Rocky Mountain foreland deformation, Am. Pedoja, K., J. F. Dumont, M. Lamothe, L. Ortlieb, J.-Y. S0012-821X(01)00225-4.
J. Sci., 286, 737 – 764. Collot, B. Ghaleb, M. Auclair, V. Alvarez, and Tassara, A., H.-J. Götze, S. Schmidt, and R. Hackney
Jordan, T. E., and R. Alonso (1987), Cenozoic strati- B. Labrousse (2006), Plio-Quaternary uplift of (2006), Three-dimensional density model of the Nazca
graphy and basin tectonic of the Andes Mountains. the Manta Peninsula and La Plata Island and the plate and the Andean continental margin, J. Geophys.
20° – 28°, south latitude, Am. Assoc. Pet. Geol. subduction of the Carnegie Ridge, central coast of Res., 111, B09404, doi:10.1029/2005JB003976.
Bull., 71, 49 – 56. Ecuador, J. South Am. Earth Sci., 22, 1 – 21, van Hunen, J., A. P. van den Berg, and N. J. Vlaar
Kay, S. M., and C. Mpodozis (2002), Magmatism as a doi:10.1016/j.jsames.2006.08.003. (2002a), On the role of subducting oceanic plateaus
probe to the Neogene shallowing of the Nazca plate Pilger, R. H. (1981), Plate reconstructions, aseismic in the development of shallow flat subduction, Tec-
beneath the modern Chilean flat-slab, J. South ridges, and low-angle subduction beneath the Andes, tonophysics, 352, 317 – 333, doi:10.1016/S0040-
Am.Earth Sci., 15, 39 – 57, doi:10.1016/S0895- Geol. Soc. Am. Bull., 92, 448 – 456, doi:10.1130/ 1951(02)00263-9.
9811(02)00005-6. 0016-7606(1981)92<448:PRARAL>2.0.CO;2. van Hunen, J., A. P. van den Berg, and N. J. Vlaar
Kodama, K. (1984), A simple model calculation of the Ramos, V. A., E. O. Cristallini, and D. J. Pérez (2002), (2002b), The impact of the South American plate
shear stress and surface profile caused by the per- The Pampean flat-slab of the central Andes, motion and the Nazca Ridge subduction on the flat
turbation on subducting plate, J. Seismol. Soc. Jpn., J. South Am. Earth Sci., 15, 59 – 78, doi:10.1016/ subduction below South Peru, Geophys. Res. Lett.,
37, 647 – 654. S0895-9811(02)00006-8. 29(14), 1690, doi:10.1029/2001GL014004.
Kopp, H., E. R. Flueh, C. Papenberg, and D. Klaeschen Regard, V., C. Faccenna, J. Martinod, and O. Bellier van Hunen, J., A. P. van den Berg, and N. J. Vlaar
(2004), Seismic investigations of the O’Higgins (2005), Slab pull and indentation tectonics: Insights (2004), Various mechanisms to induce present-day
Seamount Group and Juan Fernandez Ridge: Aseis- from 3D laboratory experiments, Phys. Earth shallow flat subduction and implications for the
mic ridge emplacement and lithosphere hydration, Planet. Inter., 149, 99 – 113, doi:10.1016/j.pepi. younger Earth: A numerical parameter study, Phys.
Te c t o n i c s , 2 3 , T C 2 0 0 9 , d o i : 1 0 . 1 0 2 9 / 2004.08.011. Earth Planet. Inter., 146, 179 – 194, doi:10.1016/
2003TC001590. Reynolds, J. H., T. E. Jordan, N. M. Johnson, J. F. j.pepi.2003.07.027.
Latrubesse, E. M., and A. Rancy (2000), Neotectonic Damanti, and K. D. Tabbutt (1990), Neogene Vlaar, N. J. (1983), Thermal anomalies and magmatism
influence on tropical rivers of southwestern Ama- deformation of the flat-subduction segment of the due to lithospheric doubling and shifting, Earth
zon during the late Quaternary: The Moa and Ipix- Argentine-Chilean Andes: Magnetostratigraphic Planet. Sci. Lett., 65, 322 – 330, doi:10.1016/
una river basins, Brazil, Quaternary Int., 72, 67 – constraints from Las Juntas, La Rioja province, 0012-821X(83)90170-X.
72, doi:10.1016/S1040-6182(00)00022-7. Argentina, Geology, 102, 1607 – 1622. von Huene, R., I. Pecher, and M.-A. Gutscher (1996),
Le Roux, J. P., C. Tavares Correa, and F. Alayza (2000), Rosenbaum, G., D. Giles, M. Saxon, P. G. Betts, R. F. Development of the accretionary prism along Peru
Sedimentology of the Rimac-Chillon alluvial fan at Weinberg, and C. Duboz (2005), Subduction of the and material flux after subduction of Nazca Ridge,
Lima, Peru, as related to Plio-Pleistocene sea level Nazca Ridge and the Inca Plateau: Insights into the Tectonics, 15, 19 – 33, doi:10.1029/95TC02618.
changes, glacial cycles and tectonics, J. South Am. formation of ore deposits in Peru, Earth Planet. Sci. Wagner, L. S., S. Beck, and G. Zandt (2005), Upper
Earth Sci., 13, 499 – 510, doi:10.1016/S0895- Lett., 239, 18 – 32, doi:10.1016/j.epsl.2005.08.003. mantle structure in the south central Chilian sub-
9811(00)00044-4. Rousse, S., S. Gilder, D. Farber, B. McNulty, P. Patriat, duction zone (30° to 36°S), J. Geophys. Res., 110,
Lonsdale, P., and K. D. Klitgord (1978), Structure and V. Torres, and T. Sempere (2003), Paleomagnetic B01308, doi:10.1029/2004JB003238.
tectonic history of the eastern Panama Basin, Geol. tracking of mountain building in the Peruvian Weijermars, R., and H. Schmeling (1986), Scaling of
Soc. Am. Bull., 89, 981 – 999, doi:10.1130/0016- Andes since 10 Ma, Tectonics, 22(5), 1048, Newtonian and non-Newtonian fluid dynamics
7606(1978)89<981:SATHOT>2.0.CO;2. doi:10.1029/2003TC001508. without inertia for quantitative modelling of rock
Macharé, J., and L. Ortlieb (1992), Plio-Quaternary Sacks, I. S., and H. Okada (1974), A comparison of the flow due to gravity (including the concept of rheo-
vertical motions and the subduction of the Nazca anelasticity structure beneath western South Amer- logical similarity), Phys. Earth Planet. Inter., 43,
Ridge, central coast of Peru, Tectonophysics, 205, ica and Japan, Phys. Earth Planet. Inter., 9, 211 – 316 – 330.
97 – 108, doi:10.1016/0040-1951(92)90420-B. 219, doi:10.1016/0031-9201(74)90139-3. Wipf, M. A. (2006), Evolution of the Western Cordil-
Marrett, R., and M. R. Strecker (2000), Response of Sallarès, V., P. Charvis, E. R. Flueh, and J. Bialas lera and coastal margin of Peru: Evidence from
intracontinental deformation in the central Andes (2005), Seismic structure of the Carnegie ridge low-temperature thermochronology and geomor-
to late Cenozoic reorganization of South American and the nature of the Galápagos hotspot, Geophys. phology, Ph.D. thesis, Swiss Fed. Inst. of Technol.,
plate motions, Tectonics, 19, 452 – 467, doi:10.1029/ J. Int., 161, 763 – 788, doi:10.1111/j.1365-246X. Zürich, Zürich, Switzerland.
1999TC001102. 2005.02592.x. Woods, M. T., and E. A. Okal (1994), The structure of the
Martinod, J., F. Funiciello, C. Faccenna, S. Labanieh, Scholz, C. H., and J. Campos (1995), On the mechan- Nazca Ridge and Sala y Gomez seamount chain from
and V. Regard (2005), Dynamical effects of sub- ism of seismic decoupling and back-arc spreading dispersion of Rayleigh waves, Geophys. J. Int., 117,
ducting ridges: Insights from 3-D laboratory mod- at subduction zones, J. Geophys. Res., 100, 205 – 222, doi: 10.1111/ j.1365-246X.1994.
els, Geophys. J. Int., 163, 1137 – 1150. 22,103 – 22,115, doi:10.1029/95JB01869. tb03313.x.
McGeary, S., A. Nur, and Z. Ben-Avraham (1985), Sébrier, M., J. L. Mercier, F. Mégard, B. Laubacher, Wortel, M. J. R., and W. Spakman (2000), Subduction
Special gaps in arc volcanism: The effect of collision and E. Carey-Gailhardis (1985), Quaternary normal and slab detachment in the Mediterranean-Car-
or subduction of oceanic plateaus, Tectonophysics, and reverse faulting and the state of stress in the pathian region, Science, 290, 1910 – 1917,
119, 195 – 221, doi:10.1016/0040-1951(85)90039-3. central Andes of south Peru, Tectonics, 4, 739 – doi:10.1126/science.290.5498.1910..
McNulty, B., and D. Farber (2002), Active detachment 780, doi:10.1029/TC004i007p00739. Yañez, G. A., C. R. Ranero, R. von Huene, and J. Dı́az
faulting above the Peruvian flat slab, Geology, 30, Siame, L. L., O. Bellier, M. Sébrier, and M. Araujo (2001), Magnetic anomaly interpretation across the
567 – 570, doi:10.1130/0091-7613(2002)030<0567: (2005), Deformation partitioning in flat subduction southern central Andes (32° – 34°S): The role of the
ADFATP>2.0.CO;2. setting: Case of the Andean foreland of western Juan Fernández Ridge in the late Tertiary evolution
Molnar, P., and D. Gray (1979), Subduction of conti- Argentina (28°S – 33°S), Tectonics, 24, TC5003, of the margin, J. Geophys. Res., 106, 6325 – 6345,
nental lithosphere: Some constraints and uncertain- doi:10.1029/2005TC001787. doi:10.1029/2000JB900337.
ties, Geology, 7, 58 – 62, doi:10.1130/0091-7613 Silver, P. G., R. M. Russo, and C. Lithgow-Bertelloni
(1979)7<58:SOCLSC>2.0.CO;2. (1998), Coupling of South American and African
Moretti, I., and K. Ngokwey (1985), Aseismic ridge plate motion and plate deformation, Sciences N. Y., S.
Brusset, B. Guillaume, J. Martinod, and V. Regard,
subduction and vertical motion of overriding plate, 279, 60 – 63. Laboratoire des Mécanismes et Transferts et Géologie,
paper presented at the Symposium Paris, Géodyna- Smith, H. F. W., and D. T. Sandwell (1997), Global Université de Toulouse, CNRS, IRD, OMP, 14 Avenue
mique des Caraı̈bes, Paris, 5 – 8 Feb . seafloor topography from satellite altimetry and E. Belin, F-31400 Toulouse, France.
Norabuena, E. D., T. H. Dixon, S. Stein, and C. G. A. ship depth soundings, Science, 277, 1956 – 1961, N. Espurt, Institut Français du Pétrole, 1 et 4 Avenue
Harrison (1999), Decelerating Nazca-South America doi:10.1126/science.277.5334.1956. de Bois-Préau, F-92852 Rueil-Malmaison Cedex,
and Nazca-Pacific plate motions, Geophys. Res. Lett., Soler, P., M. G. Bonhomme (1990), Relation of mag- France. ([email protected])
26, 3405 – 3408, doi:10.1029/1999GL005394. matic activity to plate dynamics in central Peru C. Faccenna and F. Funiciello, Dipartimento
Nur, A., and Z. Ben-Avraham (1981), Volcanic gaps from Cretaceous to present, in Plutonism From Ant- Scienze Geologiche, University Roma TRE, L. S.
and the consumption of aseismic ridges in South arctica to Alaska, edited by S. M. Kay and C. W. Leonardo Murialdo 1, I-00146 Rome, Italy.
America, in Nazca Plate: Crustal Formation and Rapela, Geol. Soc. Am. Spec. Pap., 241, 173-192.

19 of 19

View publication stats