A review of the geodynamic evolution of flat slab subduction in Mexico, Peru,
and Chile

V.C. Manea, M. Manea, L. Ferrari, T. Orozco, R.W. Valenzuela, A.

Husker, V. Kostoglodov

PII: S0040-1951(16)30591-1
DOI: doi: 10.1016/j.tecto.2016.11.037
Reference: TECTO 127339

To appear in: Tectonophysics

Received date: 7 January 2016

Revised date: 21 November 2016
Accepted date: 29 November 2016

Please cite this article as: Manea, V.C., Manea, M., Ferrari, L., Orozco, T., Valen-
zuela, R.W., Husker, A., Kostoglodov, V., A review of the geodynamic evolution
of flat slab subduction in Mexico, Peru, and Chile, Tectonophysics (2016), doi:
10.1016/j.tecto.2016.11.037

This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.
 ACCEPTED MANUSCRIPT
A REVIEW OF THE GEODYNAMIC EVOLUTION OF FLAT SLAB SUBDUCTION IN

MEXICO, PERU, AND CHILE

V. C. Maneaa,b, M. Maneaa,b, L. Ferrarib ,T. Orozcob, R.W. Valenzuelac, A. Huskerc and

PT
V. Kostoglodovc

RI
a
National Institute for Earth Physics, Măgurele, 12 Călugăreni str., 077125 Ilfov,

SC
Romania
b
Computational Geodynamics Laboratory, Centro de Geociencias, Universidad

NU
Nacional Autónoma de México, Campus Juriquilla, Querétaro, Qro., 76230, Mexico.
MA
c
Departamento de Sismología, Instituto de Geofísica, Universidad Nacional Autónoma

de México, 04510. Mexico City, D. F., Mexico.

D
TE

Abstract
P

Subducting plates around the globe display a large variability in terms of slab
CE

geometry, including regions where smooth and little variation in subduction parameters

is observed. While the vast majority of subduction slabs plunge into the mantle at
AC

different, but positive dip angles, the end-member case of flat-slab subduction seems to

strongly defy this rule and move horizontally several hundreds of kilometers before

diving into the surrounding hotter mantle. By employing a comparative assessment for

the Mexican, Peruvian and Chilean flat-slab subduction zones we find a series of

parameters that apparently facilitate slab flattening. Among them, trench roll-back, as

well as strong variations and discontinuities in the structure of oceanic and overriding

plates seem to be the most important. However, we were not able to find the necessary

and sufficient conditions that provide an explanation for the formation of flat slabs in all

three subduction zones. In order to unravel the origin of flat-slab subduction, it is

1
 ACCEPTED MANUSCRIPT
probably necessary a numerical approach that considers also the influence of

surrounding plates, and their corresponding geometries, on 3D subduction dynamics.

Keywords: Subducting slabs geometry; Flat-slab subduction; Oceanic plate and

overriding plate structure; Numeric modeling of slab flattening.

PT
RI
1. Introduction

SC
In the framework of plate tectonics, subduction zones are among the major

NU
tectonic features where Earth’s lithospheric plates return to the mantle. One tectonic
MA
consequence of the subduction process is the occurrence of inter-plate and intra-plate

earthquakes which define Wadati-Benioff zones, in other words, the present-day shape
D

of sinking slabs. This valuable information combined with tomographic anomalies

TE

provide a powerful tool used to provide a short-term snapshot of the subduction process
P

where the cold oceanic lithosphere sinks into the fluid-like mantle with a variety of dips
CE

and shapes. Traditionally, subduction zones were classified in two main categories: the

Marianas type and the Peru-Chile type (Uyeda and Kanamori, 1979). Whereas the
AC

Marianas subduction zone is characterized by an old oceanic plate that is subducting at

an almost vertical angle, the Peru-Chile type is known for its relatively fast and young

oceanic plate that subducts beneath the South American plate less steeply, including

horizontal slab segments (e.g., Barazangi and Isacks, 1976; Cahill and Isacks, 1992;

Ramos and Folguera, 2009). This particular phenomenon of flat subduction in known to

occur only in three places worldwide: Central Mexico, Peru and Central Chile (Fig. 1). In

other locations where it was previously suggested the slab dip angle is shallow rather

than horizontal (e.g., Cascades, McCrory et al., 2012; Nankai Trough, Matsubara et al.,

2008; Ecuador, Yepes et al., 2016) or with a complex geometry because of the proximity
2
 ACCEPTED MANUSCRIPT
to a triple point (eastern part of the Alaska subduction zone, Ratchkovski and Hansen,

2002, Jadamec and Billen, 2012). On the basis of their particular, but rather similar slab

geometries and subduction settings, in this review we discuss only the Mexican,

Peruvian and Pampean (Chilean) flat-slab subduction zones.

PT
In this paper we review the recent progress in understanding how flat-slabs form,

RI
evolve and interact with the surrounding mantle over geological time scales. Although

SC
clues to the dynamics of the flat-slab subduction processes can be found in several

measurable parameters, as convergence rates or incoming plate ages, their long-term

NU
dynamic evolution is still uncertain due to its transient character. Lifetime of the South
MA
America and Mexico flat-slab segments is less than 20 Myr (Ramos and Folguera, 2009;

Ferrari et al., 2012), and present-day observations are insufficient to understand

D

subduction dynamics without additional insights provided by complementary

TE

observations and techniques, as geochemistry, tectonic records, geochronology and

P

numerical modeling. However, in the last decade, knowledge of the flat-slab subduction
CE

dynamics has considerably increased (e.g. Espurt et al., 2008; Perez Campos e al.,

2008; Gerya et al., 2009; Manea et al., 2012; Skinner and Clayton, 2013; Kusky et al.,
AC

2014; Eakin et al., 2015; Gérault et al., 2015; Hu et al., 2016). While the vast majority of

subducted slabs plunge into the mantle at different, but positive dip angles, the end-

member case of flat-slab subduction seems to strongly defy this rule as the slab moves

horizontally several hundreds of kilometers before diving into the surrounding hotter

mantle. Recent improvement in geological, petrological, and geochemical records,

coupled with computational algorithms and fast growing computing resources, has

contributed to first-order understanding of the evolution of subduction zones in general,

and flat-slab subduction systems in particular. Although several alternative mechanisms

have been proposed to explain this particular type of subduction (e.g. Cross and Pilger,
3
 ACCEPTED MANUSCRIPT
1982; van Hunen et al., 2002; Manea and Gurnis, 2007; Manea et al., 2012), the

analysis of real cases of flat-slab subduction has been somewhat limited.

Here we address advances in the overall understanding of flat-slab structure and

evolution, with a focus on the Mexican, Peruvian, and Chilean flat-slab subduction

PT
zones. By comparing these three particular cases of modern flat-slab subduction we

RI
hope to contribute to the understanding of this rather enigmatic case of subduction. We

SC
first examine the present-day configuration and structure of the Mexican, Peruvian, and

Chilean flat-slabs, and then review the Neogene tectonic history of the Cocos and

NU
Nazca plates. Next, we analyze the geochemical signature and evolution of volcanism
MA
associated with the process of slab flattening. Following this section is a review of

seismic anisotropy and mantle flow in areas of flat subduction, and then we highlight
D

some of the models that best fit the observations. In the end, we will present published
TE

dynamic numeric models that have tried to explain the slab flattening process, and
P

present a comparative summary table of subduction parameters for the three flat-slab
CE

subduction zones.
AC

2. Tectonic setting

Despite the fact that recent geophysical experiments started revealing the hidden

subduction structure at a level of detail never seen before, the tectonic environment

around the subduction of Cocos and Nazca plates beneath North and South America is

still not entirely understood, especially in the flat-slab subduction areas. In this chapter,

we will briefly review the offshore and onshore first-order tectonic elements specific for

the Mexican, Peruvian, and Chilean flat-slabs.

4
 ACCEPTED MANUSCRIPT

2.1. The Mexican flat-slab

PT
In Mexico, the Cocos plate is a remnant of the large Farallon plate who began to

RI
split into a series of smaller plates since ~28 million years ago when the East Pacific

SC
Rise (EPR) began to interact with the North American Plate (Atwater and Stock, 1998).

The approaching of the EPR to the North and South America subduction zones during

NU
the Early Paleogene resulted in an increasingly narrow Farallon plate with a subduction
MA
boundary in excess of 10,000 km length and a lateral variation in slab age and

convergence direction (Wright et al., 2016). Apparently, this produced a highly tensional
D

stress field within the Farallon plate (Wortel and Cloetingh, 1981) that eventually led to
TE

its fission into the Cocos and Nazca plates around 23 million years ago (Lonsdale,
P

2005). Afterward, at ~10 Ma the Rivera plate split from the westernmost part of the
CE

Cocos plate and start acting as an independent microplate since then (DeMets and

Traylen, 2000). The triangular shaped Cocos plate is bordered to the northeast by the
AC

North American plate and the Caribbean plate, to the west by the Pacific plate and to the

south by the Nazca plate (Fig.1 – inset). The present flat-slab area is located along the

central part of the Cocos-North America plate boundary (Fig. 1A,B). Despite the fact that

the convergence rate between Cocos and North America and the plate age increases

only slightly to the southeast along the Middle America Trench (MAT) (∼5 to ∼6

cm/year, ~10 to 18 Ma, respectively) (Schellart et al., 2008; Sdrolias and Muller, 2006),

the dip of the subducting slab varies strongly, from steep to flat (Fig. 1B). In this review

paper, we use velocities for tectonic plates that were calculated using the relative plate

5
 ACCEPTED MANUSCRIPT
motion model in the Indo-Atlantic hotspot reference frame of O’Neill et al. (2005). This

reference frame is widely used in studies where evolution of subducting plates is

investigated by numerical means (Schellart et al., 2008), and it is the one that best

describes plate motions, subduction kinematics and mantle flow (Schellart, 2011).

PT
Using hypocenters of local and teleseismic earthquakes, Pardo and Suárez (1995) first

RI
showed the geometry of the subducting Cocos slab in Mexico and the extension of the

SC
flat-slab segment. However, it was only recently that the accurate geometry of the

Mexican flat-slab was revealed by the receiver function study of Perez-Campos et al.

NU
(2008). The flat-slab segment runs almost perfectly horizontal at ~45 km depth some
MA
300 km inland from the MAT before sinking at a fairly steep angle of ~75° into the

asthenosphere (Husker and Davis, 2009).

D

According to Schellart et al. (2008), the MAT in the flat-slab area retreats (rolls
TE

back) at a rate of 0.6-0.7 cm/yr. The Quaternary erosion rate along the MAT in Mexico is
P

quite small, only 0.1 cm/yr (Mercier de Lépinay et al., 1997; Vannucchi et al., 2013), and
CE

most of the ~200 m thin layer of oceanic sediments enters the subduction zone (Manea

et al., 2003). The Cocos plate contains a series of well-defined oceanic fracture zones
AC

created by the physical extension of transform faults between offset spreading centers

along EPR. These are the Orozco, O’Gorman, and farther south and distant from the

flat-slab area, the Tehuantepec fracture zone (Fig. 1A). Apart from fracture zones, the

Cocos plate morphology is characterized by abundant seamounts that roughen the

oceanic crust surface in different regions. However, offshore the flat-slab area, between

the Orozco and the O’Gorman fracture zones, the oceanic plate surface is rather smooth

(Fig. 1B) compared with the rugged surface of the neighboring regions (Kanjorsky,

2003). The bathymetry along the entire MAT reflects a complex response of the crust to

the subduction process, with the abyssal-hill fault system reactivated due to the plate
6
 ACCEPTED MANUSCRIPT
bending. Ranero et al. (2003) showed that through these faults that cut across the

oceanic crust and uppermost mantle, a large quantity of fluids is transported into the

subduction zone.

The overriding plate is part of the North American plate, which above the flat slab

PT
area moves almost westward at a rate of 1.8-2.0 cm/yr (Schellart et al., 2007). The

RI
continental crust in Mexico is composed by an assemblage of several terranes with

SC
different age and thickness (Sedlock et al., 1993; Ferrari et al., 2012). In particular, the

continental crust underlain by the flat slab area belongs to the Guerrero, the Mixteco, the

NU
Oaxaca and the Xolapa terranes (Fig. 2A). The Guerrero terrane consists of Mesozoic
MA
volcano-sedimentary, low-grade metamorphic assemblages thrusted onto Paleozoic and

Precambrian crystalline rocks of the Mixteco and Oaxaca terranes, which constitute the
D

oldest core of continental Mexico. Bounding the Guerrero, Mixteco and Oaxaca terranes
TE

is the trench parallel belt of low- to medium-grade metamorphic and plutonic rocks of the
P

Xolapa terrane, which is considered to have formed by the exhumation of the terranes to
CE

the north during the Oligocene (Herrmann et al., 1994; Ratschbacher et al., 2009; Ducea

et al., 2004).
AC

2.2. The Peruvian flat-slab

Several seismic studies documented the existence of a quite large (>1500 km in

length) flat-slab segment in central-northern Peru (Barazangi and Isacks 1976;

Hasegawa and Sacks, 1981; Suárez et al. 1983; Grange et al. 1984; Cahill and Isacks

1992). In this region the Nazca plate subducts obliquely beneath the South American

plate at a rate of ~6 cm/yr (Schellart et al., 2008) (Fig. 1C). The Nazca plate contains a

series of well-defined bathymetric highs corresponding to aseismic ridges and fracture

7
 ACCEPTED MANUSCRIPT
zones. Among them, the one that is of particular interest for the Peruvian flat-slab is the

aseismic Nazca ridge (Fig. 1C,D), which has been proposed to contribute to slab

flattening beneath Peru (Pilger, 1981) together with the hypothetical, and now

completely subducted, Inca plateau (Gutscher et al., 1999; Gutscher 2002). Offshore the

PT
region of flat-slab, the age of the Nazca plate increases along the Peru-Chile trench

RI
from ~30 Ma to the north (at ~5°S latitude) to ~45 Ma in the southern (at ~15°S latitude)

SC
part of the flat slab area (Sdrolias and Muller, 2006) (Fig.1C). Strong variations in the

erosion/shortening rates along the trench in the Peru-Chile flat-slab area produced a

NU
counterclockwise rotation of the trench. Whereas in the northern part the trench rolls
MA
back at a rate of ~1.3 cm/yr, in the southern part the trench actually rolls forward at a

high rate of ~-0.7 cm/yr (Schellart et al., 2008). The geometry of the subducted plate in
D

the Peruvian Andes is similar to that reported for the Mexican flat-slab although at
TE

different depth. The slab initially descends at an angle of ~30 down to ~100 km depth,
P

and then it flattens out for several hundred of km beneath Sierras Pampeanas, before
CE

plunging into the mantle (Dorbath et al., 1991; Gutscher et al., 2000).

In the area of flat-slab, the major morphotectonic provinces are, from west to
AC

east, the forearc, the Western Cordillera, the northern part of the Altiplano, the Eastern

Cordillera and the Subandes (Fig. 2B). The basement of the Peruvian Central Andes

consists of sialic Paleozoic basement exposed in the inner Eastern Cordillera and of

Proterozoic to Paleozoic basement exposed in the Arequipa-Antofalla coastal block (Fig.

2B) (Megard, 1987; Chew et al., 2007). This vast region represents a typical Andean-

type orogen whose tectonics have been controlled by long-lived subduction processes,

which records a complex tectonic history, including major changes in the Wadati-Benioff

geometry that accompanied uplift. The development of the present flat-slab subduction

began ~11 Ma (Hampel, 2002) and progressively produced inland migration of

8
 ACCEPTED MANUSCRIPT
magmatism as well as compression that propagated towards the foreland and eventually

uplifted the Peruvian Andes (Sébrier and Soler, 1991; Ramos and Folguera, 2009). In

contrast with the Mexican flat-slab region, the Peruvian flat-slab segment is

characterized by the lack of Plio-Quaternary volcanism. The present intense intra-plate

PT
shallow seismicity and high rate of shortening (0.4 cm/yr) recorded in central Peru has

RI
been considered the effect of flat-slab subduction (Ramos, 1999; Ramos and Folguera,

SC
2009).

Recent seismic experiments constrained the structure of the upper plate and the

NU
flat-slab geometry in great detail. Using ambient noise seismicity and surface wave
MA
analysis, Ma and Clayton (2014) imaged the subducting plate dipping at ~30º up to 150

km from the coast and then becoming horizontal up to 350 km at a depth of ~90 km. The
D

mantle above the flat slab shows relatively high velocity, suggesting that it is part of the
TE

continental lithosphere (Ma and Clayton, 2014).

P
CE

2.3. The Pampean flat-slab

AC

Based mainly on the distribution of seismicity and the absence of arc volcanism,

the pioneer work of Barazangi and Isacks (1976), Isacks et al. (1982) and Jordan et al.

(1983) first identified a flat-slab segment beneath central Chile between ~31º S and

32.5º S. Here, the relatively young Nazca plate (33-38 Ma according to Clouard et al.,

2007) subducts beneath the South American continental plate at a present-day

convergence rate of ~7.1 cm/yr (O’Neill et al., 2005) (Fig.1C,D). Like the MAT, the South

American Trench (SAT) in the flat-slab area rolls back at a rate of ~0.6 cm/yr. Among

the many ridges located on the Nazca plate, the Juan Fernandez Ridge (JFR) is of

particular interest for the Pampean flat-slab. The JFR is a ~900 km long volcanic hotspot
9
 ACCEPTED MANUSCRIPT
chain that originates from a narrow mantle plume (Kopp et al., 2004, Zhao, 2007). The

JFR acts as a barrier in the sediment transportation along the trench (Lowrie and Hey,

1981), and its interaction with the South America plate resulted in several tectonic

effects such as shoreline indentation, crustal uplift and thickening on the overriding

PT
continental plate (Fromm et al., 2004). Recent regional and local seismicity studies show

RI
the subducting Nazca plate descending to ~100-110 km depth, and then flattening out

SC
and remaining at a constant depth for ~300 km before plunging into the mantle to a

more normal subduction angle (Anderson et al., 2007; Linkimer et al., 2011; Marot et al.,

NU
2013). Although not all the aseismic ridges located on the Nazca plate are associated
MA
with flat-slab segments, the obvious spatial correlation between the position of the

Pampean flat-slab and the intersection of the JFR with the SAT has long been used as
D

an argument for suggesting that its extra buoyancy is the main cause for slab flattening
TE

in central Chile (Pilger, 1984; Gutscher et al., 2000; Yañez et al., 2001; Kay and
P

Mpodozis, 2002).
CE

The magmatic, tectonic and sedimentary processes associated with the Pampean

flat-slab have been investigated in many papers (e.g., Jordan et al., 1989, 1993, 1997;
AC

Kay et al., 1988, 1991; Kay and Abruzzi 1996; Kay and Mpodozis, 2002; Ramos et al.,

1991, 2002; Ramos, 1999) and more recently reviewed in Ramos and Folguera (2009).

As in the case of the Mexican flat-slab, the continental crust in central Chile is composed

by an assemblage of several terranes with different age and thickness, orientated

roughly parallel to the trench. The major morphotectonic provinces above the flat-slab

are the forearc, the Main Cordillera, which includes the Principal Cordillera, the Frontal

Cordillera, the Precordillera, and the Sierras Pampeanas (Fig. 2B). The high Andes are

part of the Chilenia terrane, and are separated into two main tectonic units, the Principal

and Frontal Cordillera, both of them consisting of thick- and thin-skinned thrust belts
10
 ACCEPTED MANUSCRIPT
covered by old volcanic rocks (Allmendinger et al., 1990). The Principal Cordillera is

composed by Mesozoic and Tertiary volcanic and sedimentary basement rocks with a

crust thickness up to 55 km. The Frontal Cordillera is characterized by late Paleozoic

marine and early Mesozoic igneous rocks with a crust thickness up to 70 km

PT
(Allmendinger et al., 1990; Gilbert et al., 2006; Heit et al., 2008; Gans et al., 2011).

RI
Farther to the east is the Precordillera, a thin-skinned fold and thrust belt, with a

SC
Grenville age basement covered by Paleozoic shelf carbonates (Cuyania terrane, Astini

et al., 1995; Thomas and Astini, 2003), that accommodated ~70% of the total regional

NU
shortening since 16 Ma (Allmendinger et al., 1990). The most inland tectonic unit is
MA
represented by the Sierras Pampeanas that extend from eastern half of the Cuyania

terrane across the Pampean terrane up to the edge of the Rio de la Plata craton (Fig.
D

2B). The Sierras Pampeanas represent a thick skinned block-faulted belt that extends
TE

more than 800 km inland from the trench, and whose style of deformation is
P

characterized by large basement-cored uplifts separated by sediment-filled basins

CE

(Jordan and Allmendinger, 1986; Gimenez et al., 2000; Alvarado and Ramos, 2011).
AC

3. Geophysical observations

Due to their unique geophysical and geodynamic signature, flat-slab regions offer

a good opportunity to study subduction zones from a different perspective than normal

subduction systems. In the last few decades a wealth of geophysical experiments has

been carried out in areas of flat-slabs, offering the necessary data to study the intriguing

phenomena of flat subduction. In this chapter, we present a review of available

geophysical observations of flat subduction regions. We first review seismic

observations, such as earthquake distribution, Moho depths, mantle tomography and

11
 ACCEPTED MANUSCRIPT
seismic anisotropy, then we analyze results from magnetotelluric experiments, and

finally heat flow observations related with flat subduction in Mexico, Peru and Chile.

3.1. Seismicity and geometry of the flat-slab

PT
RI
To image in detail the flat-slab subduction geometry is essential to understand its

SC
long-term geodynamic and tectonic evolution. Starting with the pioneer work of

Barazangi and Isacks (1976), who were the first to identify flat-slab segments beneath

NU
South America using teleseismic earthquake locations, many seismic studies followed
MA
and nowadays the flat-slab geometries in Mexico, Peru and Chile are well known.

Recent seismological data provide an improved image of the current geometry of the
D

subducting plate below Mexico and South America (Fig. 1). The subducting slabs along
TE

both the Mexican and South American margin show a remarkable variability despite a
P

relatively smooth variation in the incoming plate subduction parameters such as plate
CE

age and convergence rates (O’Neil et al., 2005), for example.

While early studies identified that the Mexican subduction zone presents large dip
AC

variations along strike (Molnar and Sykes, 1969; Stoiber and Carr, 1973), it was Pardo

and Suárez (1995) who, based on relocated hypocenters of local and teleseismic

earthquakes, showed for the first time the entire Cocos and Rivera slab geometry along

MAT, including the flat-slab segment. This study also revealed that the Mexican flat-slab

lacks widespread earthquakes in both the forearc region and within the subducting

Cocos slab (Fig. 3A). For this reason, in the flat-slab area the subducting slab could not

be accurately mapped for distances beyond 250 km from the MAT due to the lack of

intraslab seismicity for depths below 80-100 km. More recently, based on analysis of

data provided by an ultra-dense 2D temporary broadband seismic array located above

12
 ACCEPTED MANUSCRIPT
the flat-slab region (Middle America Subduction Experiment or MASE), Perez-Campos

et al. (2008) used both local and teleseismic waves to show that the subducted Cocos

plate beneath central Mexico becomes almost perfectly horizontal or flat at a distance of

~75 km from the MAT and at ~50 km depth. The slab runs flat for ~175 km then in

PT
plunges steeply at ~75 into the mantle (Fig. 3A).

RI
Although the Mexican flat-slab lacks extensive seismic activity, the observed

SC
stress distribution shows a preferred orientation in three principal regions parallel to the

trench (Pardo and Suárez, 1995) as following: a shallow (< 25 km depth) thrust faulting

NU
region which overrides a layer characterized by compressional events (40-50 km),
MA
followed by a sparse intra-slab seismicity region with downdip T-axes which ends at

~100 km depth (Pardo and Suárez, 1995). The lateral extent, along the MAT, of the flat-
D

slab in Mexico is still not entirely know, although Pardo and Suárez (1995) suggested
TE

that this area is bounded by the Orozco and the O'Gorman fracture zones (Fig.1A,B).
P

This observation was later reinforced by Manea and Manea (2011a), where they show
CE

that a regional thermal anomaly above the flat-slab in Mexico is bounded by the

prolongation of the Orozco and the O'Gorman fracture zones. It is worth noting that
AC

crustal seismicity in the overriding plate above the flat-slab is almost quite low and no

contractional deformation is observed in the overriding plate unlike other flat subduction

systems, like in Peru and Chile. The existence of a thin (~3–5 km thick) ultra slow

velocity layer (USVL) located on top of the flat-slab segment revealed by Song et al.

(2009), strongly suggests that in Mexico the flat-slab and the overriding plate are

essentially decoupled by a mechanically weak layer of serpentine (Manea et al., 2013)

or talc (Kim et al., 2012a). The origin of the serpentinized layer is closely related with the

presence of fluids. In central Mexico the subducting plate is younger when compared

with the slab in Peru and Chile, and most of the fluids are released before the slab
13
 ACCEPTED MANUSCRIPT
reaches 50-70 km depth (Manea and Manea, 2011b). In addition, the long lasting

Tertiary magmatic activity in the region above the flat slab (Ferrari et al., 2014) may

have created an impermeable lower crust that traps the fluids in the USVL.

Seismicity distribution in Peru has been used in numerous studies aimed at

PT
clarifying the Wadati-Benioff geometry and state of stress within the subducting slab

RI
(Chandra, 1970; Isacks and Molnar, 1971; Pennington, 1981; Jordan et al., 1983; Bevis

SC
and Isacks, 1984; Doser, 1987; Suarez et al., 1990; Norabuena et al., 1994; Tavera and

Buffon, 2001). In central Peru, the map distribution of shallow (<60 km) and intermediate

NU
depth earthquakes (< 350 km) revealed that they are located in the region between the
MA
trench and the coast, as well as inland in an area located between the Cordillera

Occidental and the Subandean zone (Fig. 3B). A nest of deep seismic hypocenters (>
D

350 km) shows a distribution subparallel with the trench and is confined to a band with a
TE

N-S trend. The cross section location of seismicity shows that the depth of earthquakes
P

increases gradually from west to east at an angle of ~30° to a depth of ~150 km where it
CE

remains constant for several hundreds of km (Fig. 3B). In this area, the Nazca plate is

considered to subduct horizontally beneath central Peru. Farther inland, the existence of
AC

a deep nest of seismicity at 500-600 km suggests that the slab descends through the

upper mantle into the transition zone (Gutscher et al., 2000). Similarly to the Mexican

flat-slab, the focal mechanisms associated to the Peruvian flat-slab in each of the three

regions mentioned above show different plane orientations and mechanisms. The

offshore and onshore shallow seismicity revealed the existence of thrusting faults with

planes oriented N-S, parallel to the Cordillera. Tavera and Buffon (2001) suggested that

the offshore horizontal compression is due to the positive convergence between the

oceanic Nazca and continental South America plates, and the onshore compressional

regime may be the consequence of the underthrusting of the Brazilian Shield under the
14
 ACCEPTED MANUSCRIPT
Cordillera Oriental. On the other hand, the intermediate depth as well as deep seismicity

shows horizontal tension axes oriented E-W and parallel with the convergence direction,

suggesting that the subducting Nazca plate is driven by slab pull forces.

According to the distribution of earthquakes in central Chile (Fig. 3C) the

PT
geometry of the Pampean flat-slab shows a remarkable similarity with the flat-slab

RI
region of Mexico and Peru: the slab first dip on average at ~30 down to ~100 km depth,

SC
where it flattens out underneath the overriding lithosphere for several hundreds of

kilometers before sinking into the upper mantle asthenosphere (Fig. 3C). For depths

NU
>200 km the seismicity in this region is nearly absent; only one earthquake cluster can
MA
be identified at ~200 km depth but separated from the shallower intermediate seismicity

by a gap (Gutscher et al., 2000). Although the presence of this aseismic gap is evident,
D

several studies (Isacks and Barazangi, 1973; James and Snoke, 1990, Engdahl et al.,
TE

1995, Pardo et al., 2002) have assumed a continuity of the slab based on tomography
P

data (Gutscher et al., 2000). One feature that distinguish the Chilean flat-slab from
CE

Mexico and Peru, is the presence of a double seismic zone located within the

subducting Nazca slab at intermediate depths (50–200 km) and separated by a 20-25
AC

km thick zone with low seismicity (Marot et al., 2013).

The focal mechanisms of earthquakes in the flat-slab region of central Chile

indicate a compressive regime (thrust and reverse faulting) at depths above 70 km, and

an extensional regime below (normal faulting) (Pardo et al., 2002; Marot et al., 2013).

The seismicity in the overriding plate shows the presence of several clusters located just

below the Sierras Pampeanas at a maximum depth of ~50 km and characterized by

reverse-faulting events.

15
 ACCEPTED MANUSCRIPT
3.2. Moho depths above flat-slabs

Crustal thickness variation above flat-slabs plays an important role not only in arc

volcanism, petrology and deformation of the upper plate (Carr, 1984; Gómez-Tuena et

PT
al., 2007; Wallace and Carmichael, 1999; Ferrari et al., 2012), but also in producing flat-

RI
slab (Manea et al., 2012). Recent receiver function studies provided details of the

SC
structure of the flat-slab subduction systems in Mexico, Peru and Chile, including Moho

depth (Pérez-Campos et al., 2008; Phillips et al., 2012; Phillips and Clayton, 2014; Gans

NU
et al., 2011; Perarnau et a., 2012). In figure 4 we present a summary of crustal thickness
MA
variation above flat-slabs. In Central Mexico, crustal thickness varies from ~35 km in the

Guerrero terrane (initially located ~90 km from the MAT) to ~45 km in the area of flat-
D

slab (located at 130-300 km from the MAT), but reaches 50-55 km in the area of the
TE

Mixteco-Oaxaca terranes beneath the volcanic arc (Fig. 4A) (Ferrari et al., 2012; Manea
P

et al., 2013).
CE

In Peru, a trench-orthogonal transect passing though Cusco, in the southern part

of the flat slab segment, the Moho depth increases from 20 to 40 km in the coastal
AC

Cordillera, reaches ~65 km in the Western Cordillera and reaches a maximum depth of

~75 km beneath the Altiplano (Phillips et al., 2012; Phillips and Clayton, 2014) (Fig. 4B).

In Chile, recent study of Gans et al. (2011) shows that the continental Moho

above the flat slab is rather thick (~70 km), complex and discontinuous in

correspondence with the Andean Cordillera. Also a fairly deep Moho (55-60 km)

characterizes the continental crust beneath the Precordillera and western Sierra

Pampeanas (Regnier et al., 1994; Fromm et al., 2004; Beck et al., 2005; Gilbert et al.,

2006; Alvarado et al., 2007; Heit et al., 2008; Ammirati et al., 2013), but reaches only

16
 ACCEPTED MANUSCRIPT
~35 km in the eastern Sierras Pampeanas (Gilbert et al., 2006; Gans et al., 2011;

Perarnau et a., 2012) (Fig. 4C).

3.3. Slow-slip Phenomena.

PT
RI
Slow slip phenomena comprises a range of phenomena that include Slow-slip

SC
events (SSEs), or slow fault slips that manifest as aseismic deformation, and related

seismic phenomena that include tectonic tremors (TTs), low frequency earthquakes

NU
(LFE’s), and very low frequency earthquakes (LVF’s). They represent a valuable piece
MA
of information that can provide constraints on the pressure and thermal structure of

subduction zones. It is now well established that the location of short term SSE’s and
D

NVT’s correlate well in Japan and Cascadia and the NVT’s are generally accepted to be
TE

a manifestation of the slip produced by the SSE’s (Shelly et al., 2007; Wech et al., 2009;

Obara, 2010; Wech and Bartlow, 2014). Slow slip phenomena are found between 10 –
P
CE

50 km depth in normal subduction zones (e.g. Beroza and Ide, 2011; Obara et al., 2010;

Obara, 2011), nut have not been found in either the Peruvian or Pampean flat slabs
AC

where there are only sparse GPS and seismic networks. However, the current

understanding of the phenomena suggests that it does not occur at greater depths. It is

likely that there is some slow slip phenomena occurring in these regions between 10 –

50 km depth, but its behavior should be like that of a typical subduction zone without

interacting with the deeper flat slab sections.

In Mexico, the flat slab is within the 10 – 50 km depth range for an extended

distance and slow slip phenomena are found to stretch out over a greater area than at

any other place in the world. The SSEs there have now been precisely recorded for

more than a decade, and they include the largest reliably detected events ever
17
 ACCEPTED MANUSCRIPT
(Kostoglodov et al., 2003; Radiguet et al., 2012; Zigone et al., 2012). The SSE’s have

been found in two regions corresponding to flat slab subduction: 1) the central part in the

state of Guerrero where the largest ones occur (Mw ~7.5); and 2) to the east in Oaxaca

where they are still quite large (Mw ~6.6), but have a smaller recurrence interval

PT
(Correa-Mora et al., 2008, 2009). The largest SSE’s recorded in the world occur in

RI
Guerrero, with an average slip of ~10 cm, produce measurable displacements over an

SC
area of ~550x250 km2 every ~4 years and last for 6 months – 1 year (Radiguet et al.,

2012). The majority of the slip from these SSE’s is located between the seismogenic

NU
zone and the flat slab region (~15 km – 40 km depth range) within the Guerrero Seismic
MA
Gap (Radiguet et al., 2012) (Fig. 5). It has been estimated that due to these large SSE’s

the accumulated slip within the Gap is 1/4 that outside of the Gap. This could be the
D

reason that no large earthquake (M ≥ 7) has occurred in more than 100 years within the
TE

Gap. There are also small, short recurrence interval SSE’s located on the flat slab, down

dip of the large SSE’s (Husker et al., 2012; Frank et al., 2015). Their recurrence interval
P
CE

is 50 – 90 days and they last for about a week with a moment magnitude equivalent of

Mw ≤ 6.4 (Frank et al., 2014, 2015).

AC

The TT’s in the flat-slab region in Mexico have a quite complex behavior. All

tremors occur close to the plate interface at ~40 km depth. The area of tremor activity

has been divided into 3 zones (Fig.5C). The transient zone located at the corner of the

slab when it first bends up at 40 km depth (~130 km – 165 km from the trench). Next

there is the buffer zone, which has very little TT and is located ~165 km – 190 km from

the trench. It appears that the small SSE’s occur in the buffer zone, but extend into the

next region (Frank et al., 2015, 2016). The third region (called a Sweet Spot)

concentrates the overwhelming majority of NVT and is located ~190 km – 245 km from

the trench. To the trenchward side of the Sweet Spot and Transient Zone there are high
18
 ACCEPTED MANUSCRIPT
Vp/Vs values observed in the upper part of the slab (Kim et al., 2010). These are most

likely due to high pore fluid pressure that delimits the zones explained above. The high

pore fluid pressure regions may exist due to dewatering of the slab as it changes phases

from the temperature and pressure gradient within the subduction zone (Manea and

PT
Manea, 2011b; Manea et al., 2004).

RI
The flat slab region in Oaxaca (eastern part of the flat slab) does not extend as

SC
far as in Guerrero and the region where slow slip phenomena are found is reduced. Only

one type of SSE has been found there. It seems to occupy a similar location to the large

NU
SSE’s in Guerrero on the dipping portion of the slab, but its recurrence interval (~2 yrs.),
MA
duration (~3 months) and magnitude equivalent (Mw 6.6) are between those of the short

term and long term SSE’s of Guerrero (Correa-Mora et al., 2008, 2009). TT has been
D

located there, but the studies there have relied on sparse networks and so it does not
TE

have information as detailed as in Guerrero. However, it appears to be on the downdip

P

edge of the SSE there as in Guerrero (Brudzinski et al., 2010).

CE

3.4. Seismic tomography

AC

A powerful tool in studying the mantle above subducting slabs, including flat-

slabs, is seismic tomography. Apart for permitting illuminating parts of subduction zones

where no in situ seismicity is available, seismic tomography can be used for determining

the composition, state, and temperature of rocks by linking petrology and temperature to

elastic properties. Thus, identifying regions with anomalous Vp, Vs and Vp/Vs ratios,

offers us a unique opportunity to eventually unravel the hidden structure of subduction

zones. Fortunately, in recent years a series of high-density high-quality seismic

experiments have been carried out in Mexico, Peru and Chile in the regions of flat-slab
19
 ACCEPTED MANUSCRIPT
subduction (MASE experiment in Mexico (Perez-Campos et al., 2008); CAUGHT and

PULSE experiments in Peru (Ward et al., 2013); OVA99, CHARGE, CHARMSE and

CHASE experiments in Chile and Argentina (Alvarado et al., 2009; Anderson et al.,

2004; Wagner et al., 2005).

PT
Previous attempts to identify in great detail the flat-slab geometry and structure

RI
beneath central Mexico were only partially successful because of scarce and sparse

SC
intraslab seismicity. One of the first tomographic images of the flat-slab was produced

only a decade ago by Gorbatov and Fukao (2005). Using global P-wave travel times

NU
they revealed the velocity structure beneath Mexico, including the flat-slab region.
MA
Although the P-wave tomography offered a low-resolution (50x50km) image at that time,

the shape of the Cocos plate in the flat-slab region could be clearly identified. One
D

unexpected feature of this study was the very steep inclination of the Cocos slab behind
TE

the flat-slab segment. This was later confirmed by Husker and Davis (2009) who show
P

that the subducting slab runs flat from the coast to just south of Mexico City, where it
CE

dives sharply into the mantle below the volcanic front of the Trans-Mexican Volcanic Belt

with a dip of ~75°, and can be followed up to ~500 km depth where it is truncated. The
AC

MASE experiment offered a good opportunity to seismically image in great detail the

Cocos plate subduction zone system in central Mexico. Later on, Kim et al. (2012a),

took the advantage of the densely spaced station array of MASE, and using an

advanced 2-D generalized radon transform-based migration method, produced a clear

image of the entire flat-slab subduction system in central Mexico (Fig. 6A). The

subducted oceanic crust can be identified as a thin low-velocity layer (7–8 km thick),

which dips at 15–20° beneath the coast, then flattens out ~300 km from the MAT where

it sharply bends at ~75° degrees underneath the Trans-Mexican Volcanic Belt (TMVB).

Also, Kim et al. (2010) revealed a negative discontinuity (fast-to-slow) located at 65–75
20
 ACCEPTED MANUSCRIPT
km within the lower crust beneath the TMVB, interpreted as a layer characterized by

partial melt (Fig. 6A). The magneto-telluric study of Jödicke et al. (2006) along the same

trace of the MASE transect imaged a low resistivity layer in the same position that they

also interpret as a partially molten lower crust. It is worth mentioning that in Mexico the

PT
Cocos flat slab lies only 3-5 km beneath the North American continental Moho, with no,

RI
or very little, lithospheric mantle in between. Manea et al. (2013) proposed that part of

SC
the low-velocity layer located atop of the flat-slab segment might actually represent a

thin layer of trapped lithospheric mantle that underwent serpentinization.

NU
One distinctive characteristic of the Peruvian, and Chilean, flat-slabs when
MA
compared with the Mexican flat-slab is the presence of upper mantle above the flat-slab

segment. In Peru and Chile, previous regional tomography shows basically that fast
D

features corresponding with the seismicity in the subducted slab, but they did not
TE

provide a more detailed Vp/Vs tomography for the slab and overlying plate (Gutscher,
P

2002; Feng et al., 2004). The Peruvian flat-slab, although is the largest zone of present-
CE

day horizontal subduction, has not been thoroughly investigated in its entire width, and

most studies have focused on the transition from flat-slab to normal dipping slab along
AC

its the southern flank. Other recent studies used receiver function in order to better

constrain the Moho depth and flat-slab geometry (Phillips et al. 2012; Phillips and

Clayton, 2014). Yet, Ma and Clayton (2014) revealed in detail for the first time the

velocity structure in the crust and uppermost mantle above the Peruvian flat-slab using

surface wave analysis. The shear-wave velocity structure above the flat-slab seems to

be fairly simple, with the mantle above the flat-slab segment having a relatively high

velocity structure, and a sill-like low-velocity layer, in the mid-crust interpreted as

partially melted zone (Fig. 6B).

21
 ACCEPTED MANUSCRIPT
In central Chile, recent tomographic studies of the upper mantle above the flat-

slab revealed that seismic velocity structures are rather different from those observed in

normal-dipping subduction zones (i.e. high Vp/Vs ratios) (Wagner et al., 2006), and more

heterogeneous when compared with the Peruvian flat-slab. In Chile, a sliver of

PT
lithospheric mantle above the horizontal segment of the subducting slab is characterized

RI
by low Vp/Vs ratios corresponding to fast S-waves and low P-wave velocities (Wagner et

SC
al., 2005; 2006; 2008). It is generally accepted that low temperature coupled with slab

dehydration induce mantle serpentinization above the subducting slabs, a process that

NU
produce high Vp/Vs ratios in seismic tomography. In central Chile the low Vp/Vs ratio is
MA
unusual for active subduction zones, and has been interpreted as an indication of low

temperature coupled with the absence of mantle hydration (Wagner et al., 2006),
D

probably indicating the presence of ancient depleted continental cratonic lithosphere.

TE

Recently, Porter et al. (2012) combined ambient-noise tomography and earthquake-

P

generated surface wave measurements in order to obtain a more detailed 3D shear

CE

velocity regional model for the Pampean flat-slab region in central Chile. One key finding

of this study revealed that above the flat slab, the mantle wedge is heterogeneous, with
AC

significant horizontal variations in seismic velocities and Vp/Vs ratios (Fig.6C). Whereas

slab velocities increase to the east, velocities in the overlying lithosphere actually

decrease, suggesting that progressively slab dewatering increase the quantity of fluids

in the overlying mantle. Actually, low shear wave velocities are identified in the upper

mantle beneath the Sierras Pampeanas, and are interpreted as hydrated continental

lithosphere (Porter et al., 2012; Marot et al., 2014). At the same time the localized low-

shear wave velocity zones (or a high Vp/Vs ratio) detected within the subducting Nazca

slab can be interpreted as a serpentinized mantle, as is common in other subduction

zones.
22
 ACCEPTED MANUSCRIPT

3.5. Seismic anisotropy

PT
RI
Rocks in the Earth's upper mantle are subject to the long-term deformation

SC
caused by convective flow. The stress produces strain that tends to align the mineral

crystals within the rocks. This alignment in turn produces seismic anisotropy, whereby

NU
waves traveling in different directions propagate at different velocities. Different seismic
MA
waves and methods such as phase-delay tomography, shear wave splitting, surface-

wave scattering, and Ps converted waves have been used to quantify seismic
D

anisotropy (see Park and Levin, 2002 for a review). Shear wave splitting has become a
TE

widely used tool to study the flow, deformation and dynamics of the upper mantle
P

anywhere a seismic station can be deployed (Silver and Chan, 1988, 1991; Silver, 1996;
CE

Savage, 1999; Park and Levin, 2002; Long and Becker, 2010), including subduction

systems (Long and Silver, 2008, 2009; Long, 2013; Long and Wirth, 2013). Anisotropic
AC

structures are formed by deformation and reorientation of minerals, known as lattice-

preferred orientation (LPO), and are usually parallel with the direction of mantle flow.

Field observations and laboratory measurements revealed how the crystallographic

axes of olivine align with shear stress (Jung and Karato, 2001). Since olivine, a naturally

anisotropic mineral, is the dominant element of the upper mantle and has an elongated

crystalline structure, LPO is considered to be the primary cause of anisotropy

observations, although there might be some secondary contribution from other factors

such as the preferred orientation of partial melt pockets (especially in the mantle wedge

below active volcanic arcs) (Zimmerman et al., 1999; Vauchez et al., 2000).
23
 ACCEPTED MANUSCRIPT
The shear wave splitting technique is commonly applied to teleseismic SKS

phases (downgoing S wave through the mantle on the source side, converted to a P

wave through the liquid outer core, and converted to an upgoing S wave through the

mantle and crust on the receiver side) in order to study upper mantle anisotropy. In the

PT
case of subduction zones, the anisotropy is accrued from the point where the P wave is

RI
converted to an S wave upon leaving the outer core, thus including contributions from

SC
the subslab mantle, the slab itself, the mantle wedge, and the overriding plate (Savage,

1999; Long and Silver, 2008, 2009; Long, 2013; Long and Wirth, 2013). Splitting

NU
measurements are often performed using local S waves from intraslab earthquakes
MA
(Yang et al., 1995; Long and Silver, 2008; Long and Becker, 2010; Long, 2013; Long

and Wirth, 2013). These measurements exclude the anisotropic contribution from the
D

subslab mantle. Additionally, source-side anisotropy can be measured using

TE

teleseismic S waves from intraslab events, provided that anisotropy is well characterized
P

under the station, as is often the case from SKS observations (Vinnik and Kind, 1993).
CE

This method is very valuable in studies of the subslab mantle (e. g., Russo and Silver,

1994; Foley and Long, 2011; Eakin and Long, 2013; Lynner and Long, 2013, 2014a,
AC

2014b). The method of shear wave splitting simultaneously determines two parameters.

The polarization direction of fast shear waves, , or fast polarization direction, provides

information about the LPO of olivine (or magmatic lenses). The second parameter is the

delay time, t, between the fast and slow arrivals and is an indicator of the extent and

strength of anisotropy. There are two main olivine fabric types that are related with the

anisotropy patterns observed in subduction zones: A- and B-type fabrics. Additionally,

the C-, D-, and E-type olivine fabrics have also been identified in natural rocks and by a

series of experimental studies (Tommasi et al., 2006; Katayama and Karato, 2006; Jung

24
 ACCEPTED MANUSCRIPT
et al., 2006; Jung, 2009). However, A-, C- and E-type olivine fabrics show the same

shear wave splitting pattern in the case of horizontal shear (Karato et al., 2008),

therefore we will focus our review only on A- and B-type olivine fabrics and their

signature on seismic anisotropy. A-type olivine fabric (the most commonly observed

PT
fabric in anhydrous peridotite) develops an LPO pattern in which the fast axes of

RI
individual olivine crystals tend to align in the direction of shear, and consequently the

SC
fast axis of anisotropy recorded at the seismic station is oriented in the direction of

mantle flow (Karato et al., 2008). This type of olivine fabric can be found in geodynamic

NU
settings where a combination of relatively low stresses, high temperatures, and low
MA
water contents are characteristic (Karato et al., 2008). Therefore, in subduction systems

A-type olivine fabric is often found in the core of the mantle wedge as well as in the
D

subslab mantle (Kneller et al., 2005, 2007; Jung et al., 2006; Long and Silver, 2008).
TE

However it has been demonstrated that LPO geometries are affected by stress,
P

temperature, pressure, and especially by the presence of melt and fluid content. Jung
CE

and Karato (2001) show that in rock samples with a high fluid content and deformed by

simple shear, the olivine fabric is B-type, and so the fast axis of anisotropy observed at
AC

the seismic station is perpendicular to the direction of mantle flow. In subduction

systems B-type olivine fabric is expected to occur in low temperature areas (below 800°

C), rich in volatiles and characterized by high stress, typical of the cold corner of the

mantle wedge (mantle wedge tip) (Kneller et al., 2005, 2007; Jung et al., 2006).

Shear wave splitting measurements in different subduction zones revealed in

general a certain preference for trench-parallel fast directions for areas close to the tip of

the mantle wedge, and trench-perpendicular fast directions farther away from the trench

(Anderson, 2005; Anderson et al. (2005); Nakajima & Hasegawa, 2004; Long & van der

Hilst 2005; Long and Wirth, 2013). Also, in many cases, fast polarizations for shear
25
 ACCEPTED MANUSCRIPT
waves with paths beneath the slab are parallel to the strike of the subduction zone, or

trench-parallel (Long and Silver, 2008). This observation has often been interpreted as

3-D return flow induced by trench migration (see Long and Silver, 2008 for a review).

The most notable exception to this rule may be the Cascadia subduction zone, where

PT
the fast polarization axes are trench-perpendicular (Currie et al., 2004; Long and Silver,

RI
2008; Eakin et al., 2010). Likewise, in Mexico trench-normal fast polarization directions

SC
have been explained either by past (Obrebski et al., 2006; van Benthem et al., 2008) or

present subduction (van Benthem, 2005; Stubailo and Davis, 2007, 2012a, 2012b, 2014,

NU
2015; Bernal-Díaz et al., 2008; León Soto et al., 2009; Rojo-Garibaldi, 2011; Rojo-
MA
Garibaldi et al., 2011; Ponce-Cortés, 2012; van Benthem et al., 2013; Bernal-López et

al., 2014, 2015; Bernal-López, 2015). In both regions the lithospheric plates being
D

subducted are relatively small, young fragments of the Farallon plate. In the case of
TE

Cascadia these are the Juan de Fuca and Gorda slabs, while the Cocos and Rivera
P

plates are subducting under Mexico. Proposed explanations for subslab, trench-normal
CE

fast polarizations include the absence of a thin decoupling layer under young and hot

slabs (Long and Silver, 2009) and the presence of horizontally aligned sheets of melt
AC

where the slab dips at angles less than 25° (Song and Kawakatsu, 2012). Below we will

review the anisotropy observations for each of the three flat-slab subduction zones, and

propose a series of models that link mantle flow, olivine fabric and observations.

3.5.1. The Mexican flat-slab

In the Mexican subduction zone, several studies now exist of upper mantle shear

wave anisotropy using records of teleseismic SKS phases (van Benthem, 2005; Stubailo

and Davis, 2007, 2012a, 2012b, 2014, 2015; Bernal-Díaz et al., 2008; León Soto et al.,
26
 ACCEPTED MANUSCRIPT
2009; Rojo-Garibaldi, 2011; Rojo-Garibaldi et al., 2011; Ponce-Cortés, 2012; van

Benthem et al., 2013; Bernal-López et al., 2014, 2015; Bernal-López, 2015; Stubailo,

2015), using S waves recorded locally (León Soto et al., 2009; León Soto and

Valenzuela, 2013), and also source-side subslab anisotropy using teleseismic S phases

PT
(Lynner and Long, 2014a). In the area of flat-slab, all fast polarization directions  are

RI
oriented roughly in the direction perpendicular to the MAT with delay times t of about 1s

SC
(Rojo-Garibaldi, 2011; Rojo-Garibaldi et al., 2011; Ponce-Cortés, 2012; van Benthem et

al., 2013; Bernal-López et al., 2014, 2015; Bernal-López, 2015; Stubailo, 2015).

NU
Physical conditions in the subslab mantle are low stress, low water content, and
MA
relatively high temperature, so the LPO of olivine is expected to be A-type (Jung et al.,

2006; Long and Silver, 2008). Therefore, van Benthem et al. (2013) propose that fast
D

polarization directions in the forearc region under the Cocos slab reflect the subslab flow
TE

direction, and are indicative of entrained flow (Fig. 7A). SKS observations using a 100-
P

station linear array found that the fast axes for stations in the forearc are oriented NE-
CE

SW on average, i.e. approximately trench-perpendicular. However, under the backarc,

where the Cocos slab subducts steeply, the fast polarization directions are oriented N-S,
AC

i. e. perpendicular to the strike of the slab, and are interpreted as a result of 2-D corner

flow within the mantle wedge (Rojo-Garibaldi, 2011; Rojo-Garibaldi et al., 2011; Bernal-

López et al., 2014, 2015; Bernal-López, 2015). The differences observed in the

orientation of the fast polarization directions between the forearc and the backarc are

due to the fact that the volcanic arc is not subparallel to the MAT (Bernal-López et al.,

2015). Also, the lack of trench-parallel (or slab strike-parallel) fast polarization

directions for stations located in the arc and backarc region (Rojo-Garibaldi, 2011;

Ponce-Cortés, 2012; van Benthem et al., 2013; Bernal-López et al., 2015) suggests that

27
 ACCEPTED MANUSCRIPT
B-type olivine fabric is not present. It is commonly accepted that one of the regions that

offer proper conditions for B-type olivine fabric is the water-rich, cold nose of the mantle

wedge. However, the absence of B-type olivine fabric would imply that the mantle

wedge associated with flat-slab subduction in Mexico does not offer all the necessary

PT
conditions for this type of fabric to properly develop. Indeed, Manea and Manea (2011b)

RI
show that most of the dehydration processes occur along the flat-slab segment before

SC
the slab plunges in the asthenosphere. In the mantle wedge, the transition from A-type

to B-type olivine fabric is controlled by temperature, and this transition takes place at

NU
~800C (Long, 2013). For the particular geometry of flat-slab subduction in Mexico,
MA
thermal models predict high temperature in the mantle wedge beneath the active

volcanic arc, without a pronounced cold (<800C) mantle wedge (Manea et al., 2005), in
D

agreement with the absence of trench-parallel (or slab strike-parallel) fast axes (Rojo-
TE

Garibaldi, 2011; Ponce-Cortés, 2012; van Benthem et al., 2013; Bernal-López et al.,
P

2015).
CE

Other segments of the MAT in Mexico, outside the region of flat slab subduction,

also show trench-normal fast polarization directions, including the source-side subslab
AC

anisotropy data set of Lynner and Long (2014a). Farther to the northwest, where the

Rivera plate subducts under North America, León Soto et al. (2009) used teleseismic

SKS data and found trench-perpendicular fast axes in the middle of the Rivera slab,

which is an indication that corner flow occurs in the mantle wedge, and also that flow is

entrained underneath the slab (similar to Long & Silver (2008) model). Their use of S

waves from local events, however, measured anisotropy mostly in the continental crust

because of the paucity of deep earthquakes, which kept them from thoroughly sampling

the mantle wedge. In the segment of the MAT southeast of the flat slab, where the slab

28
 ACCEPTED MANUSCRIPT
dips into the mantle at ~25, teleseismic observations found trench-perpendicular

directions (Bernal-Díaz et al., 2008; Ponce-Cortés, 2012; van Benthem et al., 2013; G.

León Soto and R.W. Valenzuela, manuscript in preparation, 2016). Measurements using

PT
S waves from local, intraplate Cocos earthquakes established the existence of two

distinct anisotropy patterns in the mantle wedge, separated by the 100 km isodepth

RI
contour (León Soto and Valenzuela, 2013). It should be noted that in this part of Mexico

SC
no active volcanic arc exists associated to subduction of the Cocos slab, probably due to

a lack of dehydration fluids. For the seismic stations located above the slab with depth

NU
>100 km, the anisotropy measurements show a clear pattern of trench perpendicular
MA
fast directions (average t = 0.36 s), which is consistent with the existence of A-type

olivine and a 2-D corner flow model. On the other hand, for slab depths between 60 and
D

85 km, the data show some trench-parallel fast axes (average t = 0.28 s), but the
TE

pattern is complex given that other measurements with their fast axes oriented in
P

different directions are interspersed throughout the same region, where B-type olivine is
CE

expected. This complexity likely arises from the fact that for shallow earthquakes, the

path through the mantle wedge is shorter, and thus the anisotropy contributions from the
AC

continental crust and also from the slab itself become more significant (León Soto and

Valenzuela, 2013). Teleseismic measurements farther southeast along the MAT, to the

east of the subducted extension of the Tehuantepec Ridge, show small delay times and

variable orientations of the fast polarization directions, indicative of little mantle

anisotropy (Ponce-Cortés, 2012). These observations may suggest that the pattern of

mantle flow driven by trench migration is not coherent (Long and Silver, 2008).

Furthermore, these measurements may mark a transition in the mantle flow regime

given that trench-parallel fast polarization directions are observed farther east in the

29
 ACCEPTED MANUSCRIPT
MAT in Nicaragua and Costa Rica, both below and above the subducted slab (Abt et al.,

2009, 2010). However, the situation in Central America may be more complex. Lynner

and Long (2014a) made source-side splitting measurements and found trench-

perpendicular subslab fast polarization directions under Nicaragua. Lynner and Long

PT
(2014a) suggested that their results differ from those in the work by Abt et al. (2010)

RI
because their data sets sample different volumes of the subslab mantle.

SC
3.5.2. The Andean flat-slabs

NU
MA
The pioneering upper mantle anisotropy study of Russo and Silver (1994) along

the Andean subduction zone revealed that the mantle flow beneath the subducting
D

Nazca plate, with few exceptions, is actually trench-parallel. This observation led to the
TE

conclusion that the subducted lithosphere is mainly decoupled from the underlying
P

asthenosphere, and that trench-parallel mantle flow is a consequence of the retrograde

CE

motion (or slab rollback) of the subducting slab, together with a barrier to entrained

mantle flow at depth, beneath the slab.

AC

The subslab mantle. The above-mentioned exceptions are associated to the

regions of flat-slab subduction in Peru and Chile. Russo and Silver (1994) showed that

in these two particular areas, subslab  directions are trench-perpendicular where the

dip of the slab changes abruptly, as when going from flat to normal subduction. They

further proposed that the change of slab geometry is responsible for local perturbations

of the pressure gradients and associated mantle flow. Recently, Eakin and Long (2013)

conducted a detailed shear wave splitting analysis of the Peruvian flat-slab subduction

system. They used high-quality measurements of SKS, ScS, and local S phases at a

single permanent station located just above the central part of the Peruvian flat-slab.
30
 ACCEPTED MANUSCRIPT
They also made source-side splitting measurements using teleseismic S phases

originating from slab earthquakes in Peru. The anisotropy measurements show that the

mantle beneath the flat-slab is characterized by trench-normal fast splitting directions.

Eakin and Long (2013) and Eakin et al. (2015) interpret these observations as subslab

PT
trench-perpendicular mantle flow characteristic of E- or C-type olivine LPO fabrics (Fig.

RI
7B). Farther south, where the Nazca slab is subducted at a normal dipping angle, the

SC
fast anisotropy directions in the subslab mantle show a preferred trench-sub parallel

orientation (Eakin and Long, 2013; Eakin et al., 2015).

NU
Anderson et al. (2004) applied the shear wave splitting technique using SKS,
MA
SKKS and PKS arrivals in order to obtain fast polarization directions () and splitting

delay times between fast and slow shear waves (t) in the Chilean flat-slab region. They
D

showed that the source of anisotropy in the flat-slab region, as well as the adjacent
TE

areas, is found within and below the slab (i. e., a 75-200 km thick layer located below
P

the subducting oceanic crust). In Chile, the shear wave splitting measurements revealed
CE

a pattern similar to that found for the Mexican and Peruvian flat-slabs. Within the flat-

slab region, the average, subslab  fast directions are oriented E-W, i. e., perpendicular
AC

to the trench, with t of about 1 s, whereas in adjacent areas located west and

southwest of the flat slab the fast axes are oriented N-S, i. e., trench-parallel, and the

delay times, t, are ~0.9 s (Fig. 7C).

The mantle wedge. Several studies (Anderson, 2005; Anderson et al., 2005;

MacDougall et al., 2012) present the results of shear-wave splitting measurements

within the mantle wedge for the Chilean subduction zone in the flat-slab area. They

show trench-perpendicular fast polarization directions above the flat-slab segment and a

31
 ACCEPTED MANUSCRIPT
transition to trench-parallel close to the trench for the neighboring regions where the

Nazca plate dips at a normal angle into the mantle.

Using data from a temporary array, Eakin et al. (2014, 2015) inferred the existence of

a 30-km thick wedge of anisotropic mantle material, located between the Peruvian flat

PT
slab and the continental crust. They found a pronounced change in anisotropy along the

RI
strike of the trench, with trench-parallel fast directions to the north of the subducted

SC
Nazca Ridge and variable orientations to the south. They proposed that southward

migration of the subducted Nazca Ridge produces the observed anisotropy pattern.

NU
MA
3.5.3. Comparative analysis of shear-wave splitting of the Mexican, Peruvian and

Chilean flat-slabs.
D
TE

Shear wave splitting observations in the mantle wedge above subducting slabs
P

worldwide revealed a rather complex distribution, with a pronounced spatial variability

CE

and also significant variations in splitting delay times (Long and Wirth, 2013).

Nonetheless, on a global scale, a common first-order splitting observation is the

AC

existence of trench-parallel fast directions in the mantle wedge and in the subslab

mantle (Long, 2013). In flat-slab regions, however, deviations from this pattern are quite

significant and trench-perpendicular fast directions have been reported. A comparative

analysis of the Mexican, Peruvian and Chilean flat-slabs is of critical importance in order

to interpret the seismic anisotropy observations.

The subslab mantle. In all of these three areas the fast directions in the subslab

mantle are oriented perpendicular to the trench (Fig.7). Additionally, in Peru and Chile

the trench-perpendicular fast axes stand in clear contrast with adjacent regions where

the fast directions are trench-parallel. Russo and Silver (1994) were among the first
32
 ACCEPTED MANUSCRIPT
ones to suggest that slab geometry has a local effect on subslab mantle flow. Later on,

Anderson et al. (2004) proposed that subslab mantle flow in the Chilean subduction

zone is locally controlled by the geometry of the subducting lithosphere. The model of

Anderson et al. (2004) explains the E-W (trench-perpendicular) fast directions observed

PT
beneath Central Chile in the flat-slab area as a consequence of the active flattening

RI
process and the formation of a slab space that allows for the accommodation of the

SC
general eastward mantle flow. Closer to the trench, in the regions to the west and

southwest of the flat slab, however, the subslab mantle flows N-S and is consistent with

NU
the general, trench-parallel trend for South America (Russo and Silver, 1994; Anderson
MA
et al., 2004). Furthermore, Anderson et al. (2004) take the point of view that the slab is

not torn in its transition from flat to normal subduction because flow is oriented parallel to
D

the contours of the subducting slab. In the case of Peru, source-side splitting
TE

measurements reveal trench-perpendicular fast axes beneath the flat-slab (Eakin and
P

Long, 2013; Eakin et al., 2015). The fast directions become trench-oblique to sub-
CE

parallel to the southeast of the flat slab, where the dip angle steepens again (Eakin and

Long, 2013; Eakin et al., 2015), which is broadly consistent with the trench-parallel trend
AC

for South America (Russo and Silver, 1994). Unlike the regions of normal-dipping

subduction adjacent to the flat slabs in South America, no trench-parallel fast axes are

observed in Mexico. The region of flat slab subduction in Mexico runs between 96 and

101°W and shows consistently trench-perpendicular fast polarization directions (van

Benthem, 2005; Rojo-Garibaldi, 2011; Ponce-Cortés, 2012; van Benthem et al., 2013;

Bernal-López, 2015; Bernal-López et al., 2015; Stubailo, 2015). East of 96°W the MAT

makes a bend and changes its trend from WNW-ESE to NW-SE. Likewise, the flat slab

ends and dips at an angle of ~25° (Pardo and Suárez, 1995; Rodríguez-Pérez, 2007;

Melgar and Pérez-Campos, 2011; Kim et al., 2011). The fast polarization directions in
33
 ACCEPTED MANUSCRIPT
this region are rotated ~25° clockwise relative to the stations located over the flat slab to

the west, but they are still generally trench-perpendicular (Bernal-Díaz et al., 2008;

Ponce-Cortés, 2012; van Benthem et al., 2013; G. León Soto and R.W. Valenzuela,

manuscript in preparation, 2016). The rotation of the fast axes may be related to the

PT
change in the trend of the MAT. In stations located still farther east, across the

RI
projected extension of the Tehuantepec Ridge, delay times are short (t ≈ 0.5 s) and

SC
suggest that there is little anisotropy (Ponce-Cortés, 2012). On the other hand, to the

west of the flat slab, i. e. west of 101°W, few splitting measurements are available over

NU
the Cocos slab, but these are generally, trench-oblique (León Soto et al., 2009; Ponce-
MA
Cortés, 2012). Farther west, the Rivera slab dips more steeply than the adjacent Cocos

slab (Pardo and Suárez, 1995), and a tear between the two has been tomographically
D

imaged (Yang et al., 2009). In the central part of the Rivera slab, trench-perpendicular
TE

fast axes are observed and are consistent with both 2-D entrained flow beneath the slab
P

and with corner flow in the mantle wedge (León Soto et al., 2009). The orientation of the
CE

fast axes, however, is consistent with 3-D toroidal flow around the western edge of the

Rivera slab, and with flow through the tear between the Rivera and Cocos slabs (León
AC

Soto et al., 2009).

The mantle wedge. Usually, the best way to quantify mantle anisotropy above

the slab with the shear wave splitting technique is by using S waves from local, intraslab

events. Several studies (Anderson, 2005; Anderson et al., 2005; MacDougall et al.,

2012) show that mantle wedge fast directions in Chile are trench-perpendicular above

the flat slab and trench-parallel in neighboring regions to the south, where the slab dips

more steeply into the mantle. In Peru the southward migration of the subducted Nazca

Ridge controls the anisotropy pattern within a mantle layer sandwiched between the flat

slab and the continental crust. To the north of the subducted Nazca Ridge the fast axes
34
 ACCEPTED MANUSCRIPT
are trench-parallel, whereas to the south the fast axes show variable orientations (Eakin

et al., 2014, 2015). In spite of the occurrence of intraslab earthquakes within the flat

slab segment in Mexico (Pardo and Suárez, 1995; Pacheco and Singh, 2010; Singh et

al., 2014), no shear wave splitting studies have been conducted using local S waves.

PT
Constraints on the anisotropy structure are available, however, from other techniques.

RI
Several studies have reported the existence of a thin (less than 10 km) low velocity zone

SC
(LVZ) between the continental crust and the subducted oceanic mantle. This LVZ has

been interpreted as a low velocity mantle ―wedge‖ (Pérez-Campos et al., 2008) and/or

NU
as low velocity subducted oceanic crust (Pérez-Campos et al., 2008; Song et al., 2009;
MA
Kim et al., 2010, 2012b, 2013; Song and Kim, 2012b). Furthermore, it has been

suggested that this low velocity layer decouples the slab from the overriding plate (Kim
D

et al., 2010). Song and Kim (2012b) found that the upper oceanic crust of the
TE

subducting Cocos slab is a 3 to 5 km thick ultraslow-velocity layer showing anisotropy

P

greater than 5%. Song and Kim (2012a) concluded that the topmost 2 to 6 km of the
CE

Cocos subducted oceanic mantle is an anisotropic high-velocity lid. It has fast

polarization axes oriented in the direction of plate convergence and an anisotropic

AC

strength of about 7%. It was further suggested that anisotropy was frozen in at the time

the plate formed at the Pacific-Cocos ridge spreading center (Song and Kim, 2012a;

Audet, 2013). Work using receiver functions determined the anisotropy in the

continental crust, and both the upper and lower oceanic crust (Castillo et al., 2014;

Castillo-Castellanos, 2015; J. Castillo et al., "Crust and upper mantle seismic anisotropy

variations from the coast to inland in central and southern Mexico", submitted to the

Geophysical Journal International, 2016), and subtracted these values from the SKS

shear wave splitting study by Rojo-Garibaldi (2011). Lastly, Huesca-Pérez et al. (2016)

determined continental crust seismic anisotropy from tectonic tremor and suggested that
35
 ACCEPTED MANUSCRIPT
a highly foliated system of folds-and-thrusts is the most likely cause of the observed

anisotropy.

3.6. Heat flow associated with flat-slabs

PT
RI
Another geophysical observation that improves our understanding of subduction

SC
zone processes in general, and flat-slab subduction in particular, is represented by

regional heat flow. Surface heat flow in the continents comes from two main sources:

NU
radioactive decay mainly from U, Th and K isotopes and from from the mantle.
MA
According to Smithson & Decker (1974), Smithson and Brown (1977), and Allis (1979)

~40% (or ~25 mW/m2) of the average continental heat flow is generated from
D

radioactivity whereas the remaining ~60% comes from the mantle. According to Pollack
TE

(1980), the heat flow from the continents is lower than that from the ocean basins, and,
P

as we discuss in this section, flat-slab areas show ever-lower values of heat flow. At a
CE

global scale, regions with surface heat flow greater than ~70 mW/m 2 and lower than ~40

mW/m2 are considered above and below normal values (Blackwell, 1969; 1971, Roy et
AC

al., 1972). Due to the cooling effect of descending slabs fore arc areas are characterized

by low heat flow (30–40 mW/m2), whereas regional surface heat flow above 100 mW/m2

is common in volcanic arcs, and values of ~70 mW/m2 are reported in back arc areas.

Recent increase in heat-flow data measurements, coupled with significant

improvements in digital geology maps and maps of ocean crustal age permitted the

creation of high-resolution global maps of surface heat flow (e.g. Davies, 2013). A

preliminary analysis of heat flow distribution over flat-slab areas in Mexico, Peru and

Chile show that these regions can be easily identified because of their low heat flow

(Fig. 8A,B). Low heat flow flat slab areas can be observed also in global heat flow maps
36
 ACCEPTED MANUSCRIPT
indirectly derived from seismic models based on surface waves (Shapiro and Ritzwoller,

2004).

In central Mexico, direct heat flow measurements of Ziagos et al. (1985) show

that the flat-slab fore arc region is characterized by below-normal heat flow values of

PT
only 13-22 mW/m2, suggesting low temperatures at the flat-slab interface (~45 km).

RI
Indeed, Currie et al. (2004), and Manea et al. (2004, 2005) show that just a small

SC
amount of shear heating along the flat-slab interface is necessary in order to fit the

modeled heat flow. Such low heat flow records are more common for shield and cratonic

NU
areas (12-14 mW/m2), such as the Grenville Province of the Canadian Shield and the
MA
Appalachians (Lenardic et al., 2000). However, in central Mexico, low heat flow values

above the flat-slab may actually indicate that the contribution from the mantle is lacking.
D

This can happen when there is no mantle between the slab and the crust and the
TE

continental crust is actually shielded from the mantle by the flat-slab immediately below.
P

A clearer picture of the heat-flow distribution and extension above flat-slab in Mexico is
CE

provided by Manea and Manea (2011a). Using spectral techniques applied to

aeromagnetic anomalies they estimated the depths to the Curie isotherm (~600°C),
AC

which in turn are used to estimate the heat flow. The resulting heat flow map of southern

Mexico agrees well with the locations of the main tectonic provinces but also with the

flat-slab area (Fig. 8A). According to the Manea and Manea (2011a), low values of heat

flow above the flat-slab extend over an area of ~230 km along the Pacific coast, and

bounded laterally by the prolongation of the Orozco and O’Gorman oceanic fracture

zones (Fig. 1A). The Trans Mexican Volcanic Belt is characterized by high heat flow

records above 100 mW/m2 (Ziagos et al., 1985), and locally even higher near active

volcanic centers, suggesting the occurrence of temperatures in excess of 1000°C below

the volcanic arc at Moho depths. Farther in the back-arc region, surface heat flow
37
 ACCEPTED MANUSCRIPT
decreases to ~70 mW/m2 still significantly higher than the average heat flow (~40

mW/m2) measured in regions characterized by thickened lithosphere or cratons

(Chapman and Furlong, 1992; Vitorello et al., 1980).

In the Andes, compared with the adjacent regions where the slab is steep, the

PT
flat-slab regions are also distinguishable due to low heat flow (Fig. 8B). For example, in

RI
central Peru low surface heat flow (~ 30 mW/m2) have been attributed to the absence of

SC
asthenospheric mantle wedge above the flat-slab region (Muñoz, 2005). In central Chile,

according to Hamza and Munoz (1996), Hamza et al. (2005) and Cardoso et al. (2010),

NU
heat flow above the flat-slab region is ~40 mW/m2 and increases to only ~60 mW/m2
MA
inland. Although there is relatively little variation of heat flow across the flat-slab area,

the values are well above typical heat flow recorded in continental areas (~25 mW/m2)
D

and likely reflect additional processes such as radioactive heat production within the
TE

overthickened crust, friction along the contact between the subducting Nazca plate and
P

the overriding South American plate, and the presence of lithospheric mantle between
CE

the flat-slab and the crust.

AC

3.7. Magnetotellurics

Electrical resistivity determined from magnetotellurics (MT) surveys are naturally

sensitive to the presence of fluids, whether they are aqueous of partial melts, the more

fluids are present the lower the rock resistivity (Unsworth and Rondenay, 2013). It is

known that subducting slabs, including flat-slabs, play a central role in transporting fluids

into the overlying mantle and crust where they control the chemical properties of the

rocks (R pke et al., 2004; Soyer and Unsworth, 2006). At shallow depths (less < 15 km),

fluids stored in the oceanic crust and sediments are released by compaction of the pore
38
 ACCEPTED MANUSCRIPT
space, while at greater depths, they are discharged gradually (at different P-T

conditions) by metamorphic dehydration reactions of a series of hydrous minerals

(Hacker, 2008). Additionally, slab-derived fluids trigger partial melting of the mantle

wedge that generate arc magmatism. Since electrical resistivity of rocks is quite

PT
sensitive to the presence of fluids, MT surveys can illuminate zones with highly

RI
conductive fluids (water or melt), and effectively broad the image of deep structure of

SC
subduction zones.

There are several MT studies related with flat-slab subduction in particular, the

NU
MT study of Jödicke et al. (2006) for the Mexican flat slab and Booker at al. (2004), Burd
MA
et al. (2013), and Orozco et al. (2013) for the Chilean flat slab. The MT study across the

flat-slab region in Mexico carried out by Jödicke et al. (2006) revealed several high
D

conductivity anomalies: an elongated conductive zone extending >250 km in the lower

TE

crust below the Trans-Mexican Volcanic Belt (TMVB) and two highly conductive isolated
P

anomalies located above the slab interface between the trench and the volcanic front
CE

(Fig. 9A). The elongated highly conductive zone below the TMVB was interpreted as the

presence of partial melts and fluids in the continental crust. By contrast the two isolated
AC

highly conductive spots were interpreted as regions of fluids release due to the

progressive metamorphic dehydration of the basaltic oceanic crust and uppermost

mantle (Jödicke et al., 2006; Manea and Manea, 2011b). On the other hand, the Cocos

slab is imaged as a resistive zone, with decreasin resistivity along the flat slab portion,

probably due to the increase in slab temperature.

For the Chilean flat-slab, the two-dimensional electrical resistivity model of

Booker et al. (2004) reveals a couple of highly resistive areas separated by a vertical

conductor or plume (Fig. 9B). The high resistive regions correspond to a segment of the

Nazca flat-slab, the crystalline basement exposed at surface in the Sierras Pampeanas
39
 ACCEPTED MANUSCRIPT
and the Rio de la Plata craton. According to Booker et al. (2004) the high resistivity

region attributed to the Rio de la Plata craton extends down to 200-300 km, and its

presence at such depth suggests that the further advance of the Nazca flat-slab was

blocked by the collision with the craton. This process would also prevent the

PT
development of a normal hot asthenospheric wedge, consistent with the lack of

RI
volcanism in this region. Also, a 100-km-thick electrically resistive region located

SC
between the craton and the Andes characterize the Sierras Pampeanas lithosphere (Fig.

9). The origin of this highly conductive zone, or plume, rising from deeper mantle to

NU
about 100 km and squeezed between the Nazca slab and the Rio de la Plata craton is
MA
apparently related to partial melts and hydrous fluids released from the Nazca slab

(Booker et al., 2004). However, the recentD MT study of Burd et al. (2013) proposed
D

that a wedge-shaped slab window into the subducting Nazca plate allows the conductive
TE

plume to rise through the mantle up to the base of the lithosphere. Somehow similar with
P

the Mexican flat-slab, low resistivity is also found in the lower continental crust beneath
CE

the highest elevation of the Sierras de Cordoba (Fig. 9B). Interestingly, the location of

crustal electrical conductor corresponds with the edge of the Rio de la Plata craton and
AC

the eastern extent of the Nazca flat-slab, that could be explained by fluids released from

the bending subducting slab (Booker et al., 2004).

In general, flat-slab regions in Mexico and Chile show several similarities, as

highly resistive slabs and highly conductive regions located in the overriding crust,

however, there are also significant differences. The overriding crust below the volcanic

belt shows the presence of a ~250 km long MT low-resistivity anomaly possibly related

with the presence of fluids and partial melts in the lower crust as a result of trenchward

migration of volcanism in the last 10 Myr. Compared with the observed average trench

migration rate of 1.7 cm/yr (Schellart et al., 2011), in Mexico the flat-slab rolled back at a
40
 ACCEPTED MANUSCRIPT
higher rate of ~2.5 cm/yr. This might indicate a higher rate of trench erosion along the

MAT in the past of ~0.8 cm/yr, which is considerably higher than the present-day

assumed erosion rate of ~0.1 cm/yr (Mercier de Lépinay et al., 1997; Vannucchi et al.,

2013). On the other hand, in Chile, there is only one low-resistivity spot 100 km long

PT
located in the crust beneath the Sierras de Cordoba. This likely reflects the absence of

RI
volcanism related with flat-slab subduction in Chile. Although the maximum depth of MT

SC
survey in Mexico is limited to 100 km only, there are other major differences between

the two flat-slab subduction zones such as the presence of a 200-300 km thick resistive

NU
Rio de la Plata craton in Chile and a vertical conductor located above the sinking Nazca
MA
slab.
D

4. Cocos and Nazca Plates tectonic history from Miocene to Present-day

P TE

Another parameter that control slab geometry evolution in time is the

CE

convergence rate. According to Turcotte and Schubert (2014), pressure forces induced

by the mantle flow around subducting slabs contra-balance slab pull forces. The net flow
AC

pressure acting on a descending slab is negative, producing suction forces able to lift

the slab against gravity. Suction force is directly proportional with convergence rates and

mantle wedge viscosity, the higher the convergence rate and mantle wedge viscosity,

the higher the suction force. Present day convergence rates show some modest

variability, from 5.5 cm/yr in central Mexico to 6 cm/yr in Peru and 7 cm/yr in central

Chile, however since mid-Miocene subducting plates in these flat slab regions

experienced significant variations. For example, in central Mexico, Cocos plate

experienced a substantial spike in velocity since mid-Miocene from ~6 cm/yr to ~14

cm/yr in a relatively short time span of ~5 Ma (Fig. 10). Interestingly, the interval from 16
41
 ACCEPTED MANUSCRIPT
Ma to 9 Ma is thought to be the period of the flat-slab onset in central Mexico. On the

other hand, in Peru and Central Chile, we observe only a steady decrease in

convergence rates (Norabuena et al., 1999; Sdrolias and Muller, 2006) since mid-

Miocene from ~14 cm/yr to present day values (Fig. 10). Although in central Mexico the

PT
Cocos plate acceleration during the 17-12 Ma period agrees with the onset of flat-slab,

RI
in Peru and central Chile such a correlation is not observed (Fig. 10B). However, there

SC
is an interesting correlation between the onset of flat-slab in Mexico and Peru (and in

Chile but for a less extent) and the decreasing age of the incoming plate (Fig. 10B).

NU
When comparing the plate age evolution in all three flat-slab areas we can observe that
MA
the onset of flat-slab subduction took place on the descending plate age and velocity

slope, or in the vicinity when the plate age at the trench is at its minimum. The plots
D

shown in Fig. 10B suggest that the incoming plate age should be ~35 Ma or less in order
TE

to induce flat-slab subduction. This might be an important constraint for future studies
P

related with the origin of horizontal subduction.

CE

5. Present-day thermal structure of flat-slabs

AC

It is known that subduction of oceanic lithosphere beneath continents controls the

regional thermal structure (Oxburgh and Turcotte, 1970; Anderson et al., 1978; Honda

and Uyeda, 1983; Molnar and England, 1990; Peacock, 1996, 2003; Davies, 1999),

which in turn has key implications for seismogenic zones, development of slow slip

events and non-volcanic tremors (NVTs), and ultimately arc volcanism. Present-day

thermal structure in subduction zones is commonly investigated numerically using based

on kinematic 2D heat-transfer models that use a fixed slab geometry and time-

independent parameters. These are forced convection models where flow in the mantle
42
 ACCEPTED MANUSCRIPT
wedge is driven only by mechanical coupling with the subducting slab. However, despite

such limitations, these models are well constrained by geological and geophysical

observations, and provide a valuable framework for investigating processes that occur in

subducting slabs, overriding plates and mantle-wedges. Although thermal models for

PT
normal dipping subduction zones have been developed for long time (i.e. Oxburgh and

RI
Turcotte, 1970; Toksoz et al, 1971; Turcotte and Schubert, 1973; Anderson et al., 1978;

SC
Honda and Uyeda, 1983; Cloos, 1985; van den Beukel and Wortel, 1988), models

tailored to present-day flat-slab regions as in Mexico and Chile, have been published

NU
relatively recently (Gutscher, 2002; Currie et al., 2004; Manea et al., 2003, 2004). Here
MA
we will evaluate comparatively the most recent present-day thermal models for these

three flat-slab regions. In figure 11A and B we compare thermal models for central
D

Mexico and central Chile by Manea and Manea (2011B) and Marot et al. (2014). P-T
TE

conditions derived from these thermal models are used in order to investigate stability of
P

hydrous phases and estimate dehydration processes along flat-slab interface, as well as
CE

within the slab (Fig. 11C,D).

In central Mexico most of the oceanic crust and sediments dehydrate in the flat-
AC

slab segment at distances up to ~250 km from the trench (Fig. 11C,D). The oceanic

sediments and crust dehydration pattern show that fluids are released in two separate

clusters, one located at ~150 km from the trench, and the second one located at ~200-

250 km from the trench. The location of these two dehydration clusters correlate

remarkably well with the place where NVTs have been recorded (Fig. 5) suggesting a

direct casual relationship. As discussed in the previous chapters, one of the peculiarities

of southern Mexico is the absence of contractional deformation in the overriding plate as

observed in other flat-slab subduction zones. This has been attributed to a low viscosity

layer (Manea and Gurnis, 2007) located atop of the flat-slab segment, that later was
43
 ACCEPTED MANUSCRIPT
confirmed by seismic studies of Song et al. (2009) and Kim and Clayton (2013). The ~4

km thick ultra-slow velocity layer observed on top the flat-slab interface was interpreted

by Song et al. (2009) and Kim and Clayton (2013) to hydrated minerals as serpentine

and talc, formed probable during slab flattening. Thermal models (Fig.11A) are

PT
consistent with these observations and predict P-T conditions of 400-550°C along the

RI
flat-slab interface at 45 km depth, where serpentine and talc are stable.

SC
In terms of fluids released into the mantle wedge, only a small quantity (<0.5 wt %

H2O) comes from the oceanic crust when it transforms into anhydrous eclogite at depths

NU
of 100-150 km. This would not be enough to produce the abundant arc volcanism
MA
located above the steeply dipping slab behind the flat slab segment. However, if we

assume that the oceanic lithosphere below the oceanic crust retains fluids in form of
D

serpentinization, most of the fluids are released into the overlying mantle wedge. This
TE

hypothesis strongly suggests that the flat-slab oceanic lithosphere beneath central
P

Mexico retains fluids in form of total or partial serpentinized rocks. Actually, the flat-slab
CE

in Mexico seems to be sandwiched between two serpentinzed layers, and this might

represent key information to unlock the cause of slab flattening in Mexico (Manea and
AC

Manea, 2011b).

In figure 11b we present the thermo-mechanical model for central Chile flat slab

region of Marot et al. (2014). The continental mantle both above the normal dipping slab

as well as above the flat-slab segment down to a depth of ~80 km shows temperature

<=600°C, potentially permitting mantle rocks to retain fluids in form of partially

serpentinized rocks. However, low Vp and Vs seismic velocities (typical for fluid bearing

mantle rocks) are recorded above the normal dipping slab, but not above the flat-slab

segment. Here the Vp and Vs are higher (Marot et al., 2014), typical for dry mantle

rocks. Although numeric modeling shows P-T conditions that allow the mantle to be
44
 ACCEPTED MANUSCRIPT
hydrated, the presence of cold and dry rocks above the flat-slab segment suggest that

the subducting slab in this region does not undergo dehydration reactions. It is quite

possible that most of the fluids are released into the overriding mantle before the slab

reaching 100-120 km depth, consistent with progressive increasing of P-T conditions in

PT
the slab and slow Vp and Vs recorded for this region.

RI
Modeling results (Fig. 11D) show P-T conditions within the slab of ~600°C,

SC
temperature that potentially will allow the mantle, but not the oceanic crust, to retain

fluids. The nearly anhydrous character of the oceanic crust corresponding to the flat-slab

NU
segment is also confirmed by the observed fast seismic velocities that agree with dense
MA
and anhydrous eclogites (Marot et al., 2014). It is worth mentioning that this observation

runs contrary to the still common acceptance that flat-slab in central Chile is caused by
D

the positively buoyant JFR. However P-T conditions (500~600°C at ~120 km depth)
TE

within the subducting slab indicate that fluids can be retained within the flat-slab
P

lithospheric segment. Indeed, Marot et al. (2014) show that the subducting Nazca slab
CE

may be actually dehydrating but only at the eastern tip of the flat-slab segment, and the

positive buoyancy created by the partially serpentinized lithospheric mantle in the flat-
AC

slab segment may compensate for negatively buoyant eclogitized oceanic crust.

6. Geochemical signatures of flat-slab subduction

Although magmatism is absent in regions overlying flat slab segments, the age and

location of magmatism, and the composition of magmas emplaced during the process of

slab flattening provide insights into the timing, evolution, and magma sources in such

settings. In the Andean (Chilean) and Mexican flat slab regions, the migration and

broadening of the magmatic arc associated to slab flattening was accompanied by

45
 ACCEPTED MANUSCRIPT
changes in magma composition as slab flattening progressed, but with significant

differences between these areas.

Volcanic rocks emplaced in the central Trans-Mexican Volcanic Belt (99-101°W)

during the progression of slab flattening have subalkaline, medium-K compositions (Fig.

PT
12A) (Gómez-Tuena et al., 2007; Ferrari et al., 2012). As the arc migrated, the

RI
composition of the volcanic products became more differentiated from basaltic andesitic

SC
to andesitic in positions close to the present volcanic front, to andesitic at 65-100 km

from the front, and to andesitic to dacitic at larger distances (Fig. 12A); compositions

NU
also show a poorly developed trend to increasing K 2O contents with time. The
MA
composition of late Miocene (~9-6 Ma) plateau-forming mafic rocks associated to slab

detachment that followed slab flattening is also included for comparison. Volcanic rocks
D

87
also display a trend to lower epsilon Nd and higher Sr/86Sr values with time for rocks
TE

emplaced between ca. 18 and 11 Ma at increasing distance from the modern volcanic
P

front, which reflects an increasing contribution of crustal materials. The youngest of

CE

these samples, with the lowest epsilon Nd values, have also the lowest Ni and Cr

abundances at SiO2 contents similar to those of older rocks, which point to a diminishing
AC

role for mantle melts or a lack of equilibration of crustal melts with peridotitic mantle,

suggesting thus melting in the lower crust. The previous isotopic trend changed to less

isotopically enriched compositions (epsilon Nd between +3.8 and +6.8) in andesite and

dacite emplaced at the largest distance to the actual volcanic front at the end of the slab

flattening process (ca. 11-9 Ma). The latter volcanic rocks have the geochemical

features of adakitic rocks, which reflect a garnet-bearing and plagioclase-free, high-

pressure residual source mineralogy or fractionating mineral assemblage (e.g., low Y

and heavy rare earth elements, Sm/Yb>4, Sr>890 ppm; Fig. 12). Late Miocene adakitic

rocks in the Trans-Mexican volcanic belt have the geochemical and isotopic features of
46
 ACCEPTED MANUSCRIPT
slab melts and have been interpreted as the product of melting of subducted oceanic

crust after prolonged flat subduction (Gómez-Tuena et al., 2003; Mori et al., 2007).

For the Andean flat slab segments, where the flat slab reaches a length of ~ 750 km

form the trench, melting of the slab after prolonged flat subduction has been dismissed

PT
as a mechanism for the generation of adakitic rocks, on the basis of thermal and isotopic

RI
considerations (e.g., Kay and Mpodozis, 2002). Similar thermal considerations apply to

SC
the Mexican flat slab segment, where the relatively old (ca. 30 Ma) oceanic plate that

was subducting in the late Miocene reached only 480 km from the trench. Alternatively,

NU
we suggest slab detachment and associated asthenospheric mantle flow around the
MA
slab edges as a more likely mechanism of slab melting to generate these adakitic rocks.

The Peruvian and Pampean flat slab segments share many common features in their
D

magmatic and tectonic evolution (Ramos and Folguera, 2009). Given the scarcity of
TE

geochemical data for the Peruvian area, the following discussion is mainly based on
P

information available for the Pampean flat slab segment. In this area, magmatism
CE

associated to slab flattening also evolved with time to SiO2- and K2O-richer compositions

(e.g., Kay et al., 1988, 1991; Kay and Mpodozis, 2002;), but with much wider variations
AC

in both parameters (Fig. 12B), toward andesite, dacite and rhyolite, with high-K or even

shoshonitic compositions in some of the youngest volcanic areas (e.g., Kay and

Gordillo, 1994; Kay and Mpodozis, 2002). These rocks also display a trend to increasing
87
crustal contributions with time, indicated by lower epsilon Nd and higher Sr/86Sr values

(e.g., Kay et al., 1991; Goss et al., 2013). In the Peruvian and Pampean flat slab

segments, rocks with adakitic geochemical features have been also reported (e.g., Kay

et al., 1988, 1991; Petford and Atherton, 1996; Kay and Mpodozis, 2002; Coldwell et al.,

2011), but unlike the adakites generated at the end of slab flattening in the Mexican flat

slab area, the isotopic signatures of the adakitic rocks from the Peruvian and Pampean
47
 ACCEPTED MANUSCRIPT
segments indicate a dominant contribution of high-pressure continental crust lithologies.

Two main models have been considered to account for the high-pressure, garnet

signature in these adakite-like rocks: (1) melting at the base of a thickened crust where

garnet is stable (>40–45 km) or mixing of mantle mafic magmas with lower crust

PT
materials (e.g., Hildreth and Moorbath, 1988; Kay et al., 1988, 1991; Haschke et al.,

RI
2002; Kay and Mpodozis, 2002; Goss and Kay, 2009; Coldwell et al., 2011; Mamani et

SC
al., 2010); and (2) contamination of the mantle source with fore-arc crustal rocks as

result of subduction erosion (e.g., Stern, 1991, 2011; Kay et al., 1991; Kay and

NU
Mpodozis, 2002; Litvak et al., 2007; Goss and Kay, 2009; Goss et al., 2013). In the latter
MA
model, melting of garnet-bearing lithologies at high pressure and equilibration with the

ambient peridotitic mantle would explain the high Mg and Cr and Ni abundances in
D

some of the adakitic rocks (Goss and Kay, 2009; Goss et al. 2013). A further argument
TE

given as support for the source contamination model is the transient nature of adakite
P

volcanism coinciding with the migration of the volcanic arc, which is at odds with crustal
CE

thickening processes occurring over long time periods (Kay and Mpodozis, 2002).

Geochemical data suggest that a characteristic feature of volcanism in both the

AC

Mexican and Andean flat slab segments is the trend to increasing contributions of

continental crust materials to the magmas as slab flattening progressed. Nevertheless,

crustal thickening and subduction erosion in the Peruvian and Pampean flat slab

segments played a central role in determining the composition of the magmas generated

during slab flattening, as melting occurred at increasing depths in the crust or mantle,

reaching the stability field of garnet. Such crustal thickening processes apparently did

not take place in the Mexican segment, which suggest decoupling of the converging

plates during the slab flattening phase, as observed in the modern subduction system.

On the other hand, slab detachment at the end of slab flattening in the Mexican segment
48
 ACCEPTED MANUSCRIPT
led to adakitic volcanism associated to slab melting followed shortly after by mafic

volcanism related to asthenospheric upwelling and decompression. Serpentinization of

the oceanic mantle in the bending region behind the trench is a common process in

subduction zones. Deserpentinization of these rocks occurs later once they are brought

PT
at higher depth and temperature. In flat slab areas this process is retarded but not

RI
prevented and if the water is released after the slab plunge again onto the mantle it may

SC
produce partial melting and volcanism. Thermal modelling of the flat slab in Mexico

(Ferrari et al., 2012) indicate that this process may be an important contribution to the

NU
volcanism of the volcanic front. MA
7. Numeric modeling of flat-slabs formation
D
TE

Over the last decade or so, numeric modeling has been employed in order to
P

investigate the first order controlling factors causing flat lying subduction. In this chapter
CE

we will review, in chronological order, numeric and analog modeling studies dedicated to

investigate mechanisms able to induce flat-slab subduction.

AC

One of the first studies dedicated to flat-slab formation was carried out by van Hunen

et al. (2000). They performed 2D Cartesian numerical model experiments for a passive

oceanic lithosphere that is overridden by a continental plate, and investigated in a

systematic manner several major factors that might influence the slab geometry in time

and space. The numeric technique used an incompressible medium with infinite Prandtl

number and assumed the extended Boussinesq approximation (it uses a constant

density, except in buoyancy term). The mantle rheology used was based on a realistic

composite rheological model combining Newtonian and non-Newtonian low,

corresponding to diffusion and dislocation creep. Five distinct parameters were

49
 ACCEPTED MANUSCRIPT
thoroughly investigated, namely: the depth at which the subduction fault locks, the

viscosity of the basaltic crust, the mechanical strength of the mantle, viscous dissipation,

and the overriding plate velocity. The study revealed that the only parameter that plays a

key role in slab flattening is the overriding plate velocity (Fig. 13A). This study was

PT
followed by another work where the effect of buoyant oceanic plateaus in the

RI
development of flat-slab subduction (van Hunen et al., 2002) was analyzed. This was

SC
motivated by the observed correlation between the location of overthickened oceanic

crust regions (Fig. 1) and flat-slab areas in Peru and central Chile. Although these

NU
models are 2D, in nature, they can be applied equally to both buoyant ridges and
MA
plateaus. Since during subduction the basaltic oceanic crust metamorphoses into the

heavier eclogite, one of the main issues related with the overthickened buoyant oceanic
D

crust is the ability to maintain its buoyancy at high pressure and temperature. van Hunen
TE

et al. (2002) numeric models revealed that, even in that case of a kinetic delay of
P

eclogitisation of the oceanic crust, flat slabs as those observed in Peru and central Chile
CE

cannot be obtained without additional mechanisms involved (Fig. 13A). Furthermore,

these models are in 2D, therefore they do not incorporate lateral pulling contribution
AC

from steep slabs that bound the flat slab segment on the sides, which will actually

decrease even further the positive buoyancy of the ridge or plateau. This conclusion led

to another work of van Hunen et al. (2004), where they investigate the effects of

overthrusting of the subducting plate, subduction buoyant oceanic plateau, and slab

suction forces, using similar 2D Cartesian numeric models employed in the two previous

studies. The results show that the minimum condition for slab flattening to occur is the

presence of an overthrusting continental plate, where the length of the flat-slab segment

is primarily controlled by the slab age, overriding plate motion and mantle viscosity (Fig.

13A). Also, based on Fig. 10B we suggest that the incoming plate age at the trench
50
 ACCEPTED MANUSCRIPT
should be 35 Ma or less in order to facilitate slab flattening. The role of plate suction

forces might also be significant, however they were not toughly investigated at that time.

Interestingly, the presence of a buoyant oceanic plateau is much less likely to cause flat

subduction; eventually it plays a significant role but just after the plateau initiates

PT
flattening by other mechanism. Although this chapter is dedicated to numeric modeling

RI
of slab flattening, it is worth mentioning that Martinod et al. (2005) uses 3D analog

SC
models to investigate the dynamical effects of buoyant ridges and plateaus of the slab

dip evolution. These experiments showed that buoyant oceanic features must be quite

NU
large enough to alter the dynamics of subduction, however they were not able to
MA
reproduce flat slabs similar with those observed today. Buoyant ridges that subduct

perpendicular to the trench and similar in size with present-day ridges as Juan
D

Fernandez ridge are not able to lift up the slab and provoke slab flattening. The role of
TE

suction forces on slab dynamics was studied by Manea and Gurnis (2007) where the
P

key parameter investigated was the mantle wedge viscosity and distribution. The
CE

calculations were performed in a 2-D cut through a sphere on a non-deforming grid, by

solving the conservation equations of mass, momentum and energy while making the
AC

Boussinesq approximation (Tan et al., 2006). The mantle rheology used was based on a

Newtonian rheological model, where mantle wedge viscosity depends on temperature

and pressure. This study demonstrated the importance of suction forces in controlling

the slab geometry, the transition of a subduction zone into the flat-lying regime being

achieved when the mantle wedge viscosity is lower than the surrounding asthenosphere

to a depth of ~200 km and within the range of 5 to10×10 19 Pa s. The shallower the low

viscosity wedge and the smaller the viscosity reduction, the higher the negative pressure

in the wedge above the slab and therefore the shallower the slab dip. This study showed

that the slab flattening process is greatly enhanced when the continental plate
51
 ACCEPTED MANUSCRIPT
overthrusts the subducting oceanic plate, in agreement with the previous studies

mentioned above (Fig. 13B). On the other hand, trench roll-forward can only produce

steep slabs (Fig. 13B).

Arcay et al. (2008) used for the first time a thermochemical code of convection

PT
(Christensen, 1992) in order to study the influence of subducting and overriding plate

RI
velocities on slab dip evolution. The calculations are performed in a 2D Cartesian

SC
domain, where rocks are assumed to be incompressible, except for the thermal

buoyancy term in the momentum equation, and the adiabatic heating term in the energy

NU
equation (extended Boussineq approximation). The mantle rheology used was based on
MA
two types of rocks, oceanic crust and mantle, and the rheological model is based on a

non-Newtonian law. Additionally, the models incorporate water transport (using particles
D

or tracers) and rock weakening by diminishing viscosity in the non-Newtonian domain by

TE

decreasing the pre-exponential constant. This study focused on the thermal and
P

mechanical evolution of the arc lithosphere, and its influence on the slab dip evolution.
CE

Modeling results show that when the overriding plate yields under high stress it shortens

and becomes thicker in regions located above the mantle wedge. In agreement with
AC

previous studies, modeling results incorporating upper plate trenchward advance,

revealed extremely high compressive stresses that lead to subducting plate flattening.

The progressive mantle wedge closing actually impedes thermal convection and

asthenospheric corner flow, resulting in dip angle decrease and triggering slab

flattening. This study pointed out for the first time that overriding plate evolution and

morphology (i.e. thickening) in the vicinity of mantle wedge can produce quite efficiently

slab flattening.

Another attempt to reveal the cause of flat-slab subduction was done by Espurt et al.

(2008), where large-scale analog experiments were performed. These experiments

52
 ACCEPTED MANUSCRIPT
showed that flat-slab subduction occur only in a special condition where a large amount

of a buoyant slab segment, or plateau, enters into subduction and actually balances the

negative buoyancy of the dense oceanic slab. However, nowadays flat-slab subduction

correlates spatially with buoyant oceanic ridges small in size, rather than large oceanic

PT
plateaus. Because of this clear spatial correlation between oceanic buoyant features

RI
and flat slabs location (Fig. 1), dynamic effects of aseismic ridge subduction has been

SC
also investigated using advanced and complex numeric codes, that take into account

realistic rheology (i.e. visco-elasto-plastic), phase changes reactions as well as fluid

NU
propagation and partial melting (Gerya et al., 2009). The momentum, continuity, and
MA
heat conservation equations takes into account thermal and chemical buoyancy, and are

solved using the I2ELVIS code (Gerya & Yuen, 2003) based on conservative finite
D

differences and a non diffusive-marker-in-cell technique to track chemical composition.

TE

The numeric model incorporates an oceanic ridge corresponding in size to that of the
P

Nazca ridge (~18 km thick) in the direction orthogonal to the trench. The main
CE

conclusion of this study was consistent with previous findings, namely that subduction of

realistic buoyant oceanic ridges does not result in strong slab flattening and related
AC

decrease of magmatic activity (Fig. 13B). Thermomecanical numeric models were also

performed in order to study margin deformation processes along the Andean subduction

zone, including flat slab regions (Gerbault et al., 2009). They used a 2D Lagrangian

approach (with large-strain Lagrangian formulation) that accounts for elastic-viscous-

brittle behavior and the model setup included initial thermal and rheological structures,

close to the generic present-day Chilean margin conditions. Modeling results showed

that slab flattening is obtained only in the particular case of young and hydrated oceanic

plate forced by the slab-pull to subduct under a resistant continent, and more important,

when rheological properties of the oceanic and continental mantle are switched.
53
 ACCEPTED MANUSCRIPT
However the length of the flat-slab segment obtained was relatively small (~100 km),

whereas intraslab seismicity associated with flat-slab subduction in Peru and central

Chile show larger flat-slab segments up to 300 km or more (Fig. 13B).

The role of the overriding plate thermal state on slab geometry was investigated

PT
numerically by Manea et al. (2012) and Rodriguez-Gonzales et al. (2012). Both studies

RI
used a similar approach where equations of conservation of mass, momentum and

SC
energy were solved for an incompressible 2D fluid domain with high Prandtl number,

which allows us to neglect inertial forces on the momentum equation. Also, these time-

NU
dependent 2D numeric models are similar in terms of boundary conditions, where free
MA
slip boundary conditions were applied to the bottom and side boundaries of the model

domain, while the top of the models used prescribed velocities. In terms of rheology,
D

Manea et al. (2012) used a simple Newtonian rheology coupled with a low viscosity
TE

wedge and channel, whereas Rodriguez-Gonzales et al. (2012) employed a composite

P

rheology based on both diffusion and dislocation creep viscosities. Both studies
CE

provided the same main conclusion, where flat-slab subduction is obtained as a result of

the increased suction force in the mantle wedge when an overthickened and colder
AC

continental lithosphere is located in the vicinity of the subduction system (Fig. 13C).

Flat-slab geometries similar with those observed in central Chile were obtained when

a cratonic lithosphere (200 km or more) is actually combined with trench roll-back

(Manea et al., 2012). This was later confirmed by Taramon et al. (2015) study, where

the influence of cratonic lithosphere on the formation and evolution of flat slabs was

carried out using advanced 3-D time-dependent numeric models (Fig. 13C). In a recent

study of Antonijevic et al. (2015), based on a detailed three-dimensional model of the

structure of shear-wave velocities for the Peruvian flat-slab region, slab flattening is

proposed to be the result of the combined effects of trench retreat, suction, and ridge
54
 ACCEPTED MANUSCRIPT
subduction. This study proposed that the buoyant Nazca ridge plays rather a secondary

role in flat-slab formation, but eventually could support maintaining flat-slab regimes for

longer periods of time.

Although slab flattening seems to occur preferentially in the vicinity of cold and thick

PT
cratons and correlate spatially with buoyant oceanic impactors, as is the case of Peru

RI
and central Chile, in central Mexico the continental lithosphere is very thin or even

SC
missing, and the oceanic Cocos plate lacks large buoyant features. Still flat-slab

subduction has been occurring since middle Miocene. The above-mentioned modeling

NU
studies indicate that whereas buoyant ridges alone are not sufficient for the formation of
MA
flat slabs, the ultimate cause of slab flattening is yet to be found.
D

8. Concluding remark
P TE

As a result of this review we present a comprehensive table (Table 1) comparing a

CE

variety of observations of the three flat-slab subduction regions discussed in this work.

Among the different parameters analyzed we concluded that trench dynamics (i.e. roll-
AC

back) and subducting plate tectonic evolution (i.e. super-fast subduction episodes and/or

young incoming oceanic plates), as well as strong discontinuities in the oceanic and

overriding plates structure are the most important in facilitating slab flattening. However,

we were unable to find a unique set of conditions that provide an explanation for the

formation of flat slabs in all three subduction zones. In this respect we think that the

ultimate cause of slab flattening is yet to be found. Although at a first sight flat-slab

subduction zones do not easily fit into the plate tectonics paradigm, discovering the real

cause of slab flattening offer a great opportunity to better understand subduction zones

and to unravel the complexity of mechanism that drive plate tectonics.

55
 ACCEPTED MANUSCRIPT

References

Abt, D.L., Fischer, K.M., Abers, G.A., Strauch, W., Protti, J.M., González, V., 2009.

Shear wave anisotropy beneath Nicaragua and Costa Rica: Implications for flow in the

PT
mantle wedge. Geochemistry, Geophysics, Geosystems, 10, Q05S15, doi:

RI
10.1029/2009GC002375.

SC
Abt, D.L., Fischer, K.M., Abers, G.A., Protti, M., González, V., Strauch, W., 2010.

Constraints on upper mantle anisotropy surrounding the Cocos slab from SK(K)S

NU
splitting. Journal of Geophysical Research, 115, B06316, doi:10.1029/2009JB006710.
MA
Allis, R.G., 1979. A heat production model for the stable continental crust.

Tectonophysics, 57, 151–165.

D

Allmendinger, R.W., Figueroa, D., Snyder, D., Beer, J., Mpodozis, C., Isacks, B.L.,
TE

1990. Foreland shortening and crustal balancing in the Andes at 30°S latitude.
P

Tectonics, 9, 789– 809.

CE

Alvarado, P., Ramos, V., 2011. Earthquake deformation in the northwestern Sierras

Pampeanas of Argentina based on seismic waveform modeling. Journal of

AC

Geodynamics, 51, 205-218.

Alvarado, P., Pardo, M., Gilbert, H., Miranda, S., Anderson, M., Saez, M., Beck, S.L.,

2009. Flat-slab Subduction and Crustal Models for the Seismically Active Sierras

Pampeanas Region of Argentina. MWR204: Backbone of the Americas: Shallow

Subduction, Plateau Uplift, and Ridge and Terrane Collision. GSA, Boulder, Colorado,

USA. doi: 10.1130/2009.1204(12), pp. 261-278.

Alvarado, P., Beck, S., Zandt, G., 2007. Crustal structure of the south-central Andes

Cordillera and backarc region from regional waveform modeling. Geophysical Journal

International, 121, 170(2), 858–875.

56
 ACCEPTED MANUSCRIPT
Ammirati, J.-B., Alvarado, P., Perarnau, M., Saez, M., Monsalvo, G., 2013. Crustal

structure of the Central Precordillera of San Juan, Argentina (31°S) using teleseismic

receiver functions. Journal of South American Earth Sciences, 46, 100-109.

Anderson, M.L., 2005. Seismic anisotropy, intermediate-depth earthquakes, and

PT
mantle flow in the Chile-Argentina flat-slab subduction zone, Ph. D. thesis, 280 pp.,

RI
University of Arizona, USA.

SC
Anderson, R.N., DeLong, S.E., Schwarz, W.M., 1978. Thermal model for subduction

with dehydration in the downgoing slab. Journal of Geology, 86(6), 731-739.

NU
Anderson, M. L., Zandt, G., Triep, E., Fouch, M., Beck, S., 2004. Anisotropy and
MA
mantle flow in the Chile-Argentina subduction zone from shear wave splitting analysis.

Geophysical Research Letters, 31, L23608, doi:10.1029/2004GL020906.

D

Anderson, M.L., Wagner, L., Gilbert, H., Alvarado, P., Zandt, G., Beck, S., Triep, E.,
TE

2005. Constraints on mantle kinematics influenced by a changing slab geometry:

P

Results from the CHARGE project, South America. Abstract 1000 presented at the
CE

IASPEI General Assembly, Santiago, Chile, October.

Anderson, M., Alvarado, P., Zandt, G., Beck, S., 2007. Geometry and brittle
AC

deformation of the subducting Nazca plate, central Chile and Argentina. Geophysical

Journal International, 171, 419–434.

Antonijevic, S.K., Wagner, L.S., Kumar, A., Beck, S.L., Long, M. D., Zandt, G.,

Tavera, H., Condori, C., 2015. The role of ridges in the formation and longevity of flat

slabs. Nature, 524, 212-215, doi:10.1038/nature14648.

Arcay, D., Lallemand, S., Doin, M.-P., 2008. Back-arc strain in subduction zones:

Statistical observations versus numerical modeling. Geochemistry, Geophysics,

Geosystems, doi: 10.1029/2007GC001875.

57
 ACCEPTED MANUSCRIPT
Astini, R.A., Benedetto, J.L., Vaccari, N.E., 1995. The early Paleozoic evolution of

the Argentine Precordillera as a Laurentian rifted, drifted and collided terrain: a

geodynamical model. Geological Society of America Bulletin, 107, 253– 273.

Atwater, T., Stock, J., 1998. Pacific-North America Plate tectonics of the Neogene

PT
Southwestern United States – An update. International Geological Review, 40, 375-402.

RI
Audet, P., 2013. Seismic anisotropy of subducting oceanic uppermost mantle from

SC
fossil spreading. Geophysical Research Letters 40, 173-177,

doi:10.1029/2012GL054328.

NU
Barazangi, M., Isacks, B.L., 1976. Spatial distribution of earthquakes and subduction
MA
of the Nazca plate beneath South America. Geology, 4, 686–692.

Beck, S.L., Hersh, G., Wagner, L., Alvarado, L., Anderson, M., Zandt, G., Araujo, M.,
D

Triep, E., 2005. The lithosfera structure of the Sierras Pampeanas region of Argentina.
TE

2005 Salt Lake City GSA Annual Meeting, Session no. 253, paper no: 253-5.
P

Bernal-Díaz, A., Valenzuela-Wong, R., Pérez-Campos, X., Iglesias, A., Clayton,

CE

R.W., 2008. Anisotropía de la onda SKS en el manto superior debajo del arreglo VEOX

(abstract). Geos Boletín Informativo de la Unión Geofísica Mexicana, 28 (2), 199-200.

AC

Bernal-López, L.A., 2015. Anisotropía sísmica y flujo del manto producidos por la

placa de Cocos subducida en el sur de México. M. Sc. thesis, 65 pp., Centro de

Sismología y Volcanología de Occidente, Universidad de Guadalajara, Puerto Vallarta,

Jalisco, Mexico.

Bernal-López, L.A., León Soto, G., Valenzuela Wong, R., Núñez Cornú, F.J., 2014.

Anisotropía sísmica y flujo del manto producidos por la placa de Cocos subducida en el

sur de México (abstract). Geos Boletín Informativo de la Unión Geofísica Mexicana 34

(1), 227.

58
 ACCEPTED MANUSCRIPT
Bernal-López, L.A., Garibaldi, B.R., León Soto, G., Valenzuela, R.W., Escudero,

C.R., 2015. Seismic anisotropy and mantle flow driven by the Cocos slab under

southern Mexico. Pure and Applied Geophysics, doi:10.1007/s00024-015-1214-7.

Beroza, G.C., Ide, S., 2011. Slow earthquakes and nonvolcanic tremor. Annu Rev

PT
Earth Planet Sci, 39,271–296.

RI
Bevis, M., Isacks, B.L., 1984. Hypocentral trend surface analysis: probing the

SC
geometry of Benioff Zones. Journal of Geophysical Research, 89, 6153–6170.

Blackwell, D.D., 1969. Heat flow in the northwestern United States: Journal of

NU
Geophysical Research, 74, 992–1077. MA
Blackwell, D.D., 1971. Thermal structure of the continental crust, in Heacock, J.G.,

ed., ―The structure and physical properties of the Earth’s crust‖, Washington, D.C.,
D

American Geophysical Union Monograph, 14, 169–184.

TE

Booker, J., Favetto, A., Pomposiello, C., 2004. Low electrical resistivity associated
P

with plunging of the Nazca flat slab beneath Argentina. Nature, 429, 399–403.
CE

Burd, A.I., Booker, J.R., Mackie, R., Pomposiello, M.C., Favetto, A., 2013. Electrical

conductivity of the Pampean shallow subduction region of Argentina near 33°S:

AC

Evidence for a slab window, Geochemistry, Geophysics, Geosystems, 14, 3192–3209,

doi:10.1002/ggge.20213.

Brudzinski, M.R., Hinojosa ‐ Prieto, H.R., Schlanser, K.M., Cabral ‐ Cano, E.,

Arciniega‐Ceballos, A., Diaz‐Molina, O., DeMets, C., 2010. Nonvolcanic tremor along

the Oaxaca segment of the Middle America subduction zone, Journal of Geophysical

Research (ISSN 0148–0227, eISSN 2156–2202)., v. 115, B00A23,

DOI:10.1029/2008JB006061.

59
 ACCEPTED MANUSCRIPT
Cahill, T., Isacks, B.L., 1992. Seismicity and shape of the subducted Nazca plate.

Journal of Geophysical Research, 97, 17503– 17529.

Cardoso, R.R., Hamza, V.M., Alfaro, C. 2010. Geothermal resource base for South

America: A continental perspective. Proceedings World Geothermal Congress 2010,

PT
Bali, Indonesia, 25-29 April 2010.

RI
Carr, M.J., 1984. Symmetrical and segments variation of physical and geochemical

SC
characteristics of the Central American volcanic front. Journal of Volcanolgy and

Geothermal Research, 20, 231-252.

NU
Castillo-Castellanos, J.A., 2015. Variaciones de la anisotropía sísmica en la corteza
MA
y manto superior en el centro-sur de México. M.Sc. thesis, 147 pp., Instituto de

Geofísica, Universidad Nacional Autónoma de México, Mexico City, Mexico.

D

Castillo, J.A., Pérez-Campos, X., Husker, A.L., Valenzuela-Wong, R., 2014. Crust
TE

and mantle anisotropy variations from the coast to inland in central and southern
P

Mexico, Abstract DI33A-4303 presented at 2014 Fall Meeting, American Geophysical

CE

Union, San Francisco, CA, 15-19 December.

Chandra, U., 1970. The Peru-Bolivia border earthquake of August 15, 1963. Bulletin
AC

of the Seismological Society of America, 60, 639-646.

Chapman, D.S., Furlong, K.P., 1992. Thermal state of the continental lower crust. In:

Fountain, D.M., Arculus, R.J., Kay, R.W. (Eds.), Continental Lower Crust. Elsevier,

Amsterdam, pp. 179– 199.

Chew, D.M., Schaltegger, U., Košler, J., Whitehouse, M.J., Gutjahr, M., Spikings,

R.A., Miškovic., A., 2007. U-Pb geochronologic evidence for the evolution of the

gondwanan margin of the north-central Andes. Geological Society of America Bulletin,

119 (5/6), 697-711.

60
 ACCEPTED MANUSCRIPT
Christensen, U.R., 1992. An Eulerian technique for thermomechanical modeling of

lithospheric extension. Journal of Geophysical Research, 97 (B2), 2015-2036.

Cloos, M., 1985. Thermal evolution of convergent plate margins: Thermal modeling

and reevaluation of isotopic AR‐ages for Blueschists in the Franciscan Complex of

PT
California. Tectonics, 4(5), 421-433

RI
Clouard, V., Campos, J., Lemoine, A., Perez, A., Kausel, E., 2007. Outer rise stress

SC
changes related to the subduction of the Juan Fernandez Ridge, central Chile. Journal

of Geophysical Research, 112, doi: 10.1029/2005JB003999, issn: 0148-0227.

NU
Coldwell, B., Clemens, J., Petford, N., 2011. Deep crustal melting in the Peruvian
MA
Andes: Felsic magma generation during delamination and uplift. Lithos 125, 272-286.

Correa-Mora, F., DeMets, C., Cabral-Cano, E., Marquez-Azua, B., Diaz-Molina, O.,
D

2008. Interplate coupling and transient slip along the subduction interface beneath
TE

Oaxaca, Mexico. Geophysical Journal International, 175, 269-290, doi:10.1111.j.1365-

P

246X.2008.03910.x.
CE

Correa-Mora, F., DeMets, C., Cabral-Cano, E., Diaz-Molina, O., Marquez-Azua, B.,

2009. Transient deformation in southern Mexico in 2006 and 2007: Evidence for distinct
AC

deep-slip patches beneath Guerrero and Oaxaca. Geochemistry, Geophysics,

Geosystems, 10, 1-12.

Cross, T., Pilger, R. 1982. Control of subduction geometry, location of magmatic arcs

and tectonics of arc and back-arc regions. Geological Society of America Bulletin, Vol.

93, p. 545-562.

Currie, C.A., Cassidy, J.F., Hyndman, R.D., and Bostock, M.G., 2004. Shear wave

anisotropy beneath the Cascadia subduction zone and western North American craton.

Geophysical Journal International, 157, 341-353.

61
 ACCEPTED MANUSCRIPT
Davies, J.H., 1999. The role of hydraulic fractures and intermediate-depth

earthquakes in generating subduction-zone magmatism, Nature, 398, 142-145.

Davies, J.H., 2013. Global map of solid Earth surface heat flow. Geochemistry,

Geophysics, Geosystems, doi: 10.1002/ggge.20271.

PT
DeMets, C., Traylen, S., 2000. Motion of the Rivera plate since 10Ma relative to the

RI
Pacific and North American and the mantle. Tectonophysics, 318, 119–159.

SC
doi:10.1016/S0040-1951(99)00309-1.

Dorbath, L., Dorbath, C., Jimenez, E., Rivera, L., 1991. Seismicity and tectonic

NU
deformation in the Eastern Cordillera and the sub-Andean zone of central Peru. Journal
MA
of South American Earth Sciences, 4 (1/2), 13-24.

Doser, D.I., 1987. The Ancash, Peru, earthquake of 1946 November 10: evidence for
D

low-angle normal faulting in the high Andes of northern Peru. Geophysical Journal
TE

International, 91(1), 57-71.

P

Ducea, M., Valencia, V.A., Shoemaker, S., Reiners, P.W., DeCelles, P.G., Campa,
CE

M.F., 2004. Rates of sediment recycling beneath the Acapulco trench: Constraints from

(U‐Th)/He thermochronology. Journal of Geophysical Research: Solid Earth (1978–

AC

2012), 109(B9).

Eakin, C.M., Long, M.D., 2013. Complex anisotropy beneath the Peruvian flat slab

from frequency-dependent, multiple-phase shear wave splitting analysis. Journal of

Geophysical Research, 118, 4794-4813, doi:10.1002/jgrb50349.

Eakin, C.M., Obrebski, M., Allen, R.M., Boyarko, D.C., Brudzinski, M.R., Porritt. R.,

2010. Seismic anisotropy beneath Cascadia and the Mendocino triple junction:

Interaction of the subducting slab with mantle flow. Earth and Planetary Science Letters

297, 627-632.

62
 ACCEPTED MANUSCRIPT
Eakin, C.M., Long, M.D., Beck, S.L., Wagner, L.S., Tavera, H., Condori, C., 2014.

Response of the mantle to flat slab evolution: Insights from local S splitting beneath

Peru. Geophysical Research Letters 41, 3438-3446, doi: 10.1002/2014GL059943.

Eakin, C.M., Long, M.D., Wagner, L.S., Beck, S. L., Tavera, H., 2015. Upper mantle

PT
anisotropy beneath Peru from SKS splitting: Constraints on flat slab dynamics and

RI
interaction with the Nazca Ridge. Earth and Planetary Science Letters, 412, 152-162,

SC
doi:10.1016/j.epsl.2014.12.015.

Engdahl, E.R., van der Hilst, R.D., Berrocal, J., 1995. Imaging of subducted

NU
lithosphere beneath South America. Geophysical Research Letters, 22 (16), 2317-2320,
MA
doi: 10.1029/95GL02013.

Espurt, N., Funiciello, F., Martinod, J., Buillaume, B., Regard, V., Faccenna, C.,
D

Brusset, S., 2008. Flat subduction dynamics and deformation of the South American
TE

plate: Insights from analog modeling. Tectonics, 27, TC3011,

P

doi:10.1029/2007TC002175.
CE

Feng, M., Assumpção, M., Van der Lee, S., 2004. Group-velocity tomography and

lithospheric structure of the South American continent. Physics of the Earth and
AC

Planetary Interiors, 147, 315-331.

Ferrari, L., Orozco-Esquivel, T., Manea, V.C., Manea, M., 2012. Review Article: The

dynamic history of the Trans-Mexican Volcanic Belt and the Mexico subduction zone,

Tectonophysics, 522-523, 122-149.

Ferrari, L., ergomi, M., Martini, M., Tunesi, A., Orozco- Esquivel, M., L pez

Mart nez, M., 2014. Late Cretaceous - Oligocene magmatic record in southern Mexico:

the case for a temporal slab window along the evolving Caribbean-North America-

Farallon triple boundary. Tectonics, v. 33, p. 1738–1765, DOI: 10.1002/2014TC003525.

63
 ACCEPTED MANUSCRIPT
Foley, B.J., Long, M.D., 2011. Upper and mid-mantle anisotropy beneath the Tonga

slab, Geophysical Research Letters 38, L02303, doi:10.1029/2010GL046021.

Frank, W.B., Shapiro, N.M., Husker, A.L., Kostoglodov, V., Romanenko, A.,

Campillo, M., 2014. Using systematically characterized low-frequency earthquakes as a

PT
fault probe in Guerrero, Mexico. Journal of Geophysical Research,119 (10), 1686–7700.

RI
Frank, W.B., Shapiro, N.M., Husker, A.L., Kostoglodov, V., Bhat, H.S., Campillo, M.,

SC
2015. Along-fault pore-pressure evolution during a slow-slip event in Guerrero, Mexico,

Earth and Planetary Science Letters. 413, 135-143.

NU
Frank, W.B., Shapiro, N.M., Husker, A.L., Kostoglodov, V., Gusev, A.A., Campillo,
MA
M., 2016. The evolving interaction of low-frequency earthquakes during transient slip.

Science Advances, 2(4), e1501616. http://doi.org/10.1126/sciadv.1501616.

D

Fromm, R., Zandt, G., Beck, S.L., 2004. Crustal thickness beneath the Andes and
TE

Sierras Pampeanas at 30°S inferred from Pn apparent phase velocities. Geophysical

P

Research Letters 31, L06625, doi:10.1029/2003GL019231.

CE

Gans, C.R., Beck, S.L., Zandt, G., Gilbert, H., Alvarado, P., Anderson, M., Linkimer,

L., 2011. Continental and oceanic crustal structure of the Pampean flat slab region,
AC

western Argentina, using receiver function analysis: new high-resolution results.

Geophysical Journal International, 186(1), 45-58.

Gerbault, M., Cembrano, J., Mpodozis, C., Farias, M., Pardo, M., 2009. Continental

margin deformation along the Andean subduction zone: thermo-mechanical models.

Physics of the Earth and Planetary Interiors, 177(3-4), 180-205.

Gérault, M., Husson, L., Miller, M.S., Humphreys, E.D., 2015. Flat-slab subduction,

topography, and mantle dynamics in south-western Mexico, Tectonics, 34, doi:

10.1002/2015TC003908.

64
 ACCEPTED MANUSCRIPT
Gerya, T.V., Fossati, D., Cantieni, C., Seward, D., 2009. Dynamic effects of aseismic

ridge subduction: numerical modelling. European Journal of Mineralogy, 21, 649-661.

Gerya, T.V., Yuen, D.A., 2003. Characteristics-based marker-in-cell method with

conservative finite-differences schemes for modeling geological flows with strongly

PT
variable transport properties. Phys. Earth Planet. Interiors, 140, 293-318.

RI
Gilbert, H.J., Beck, S., Zandt, G., 2006. Lithospheric and upper mantle structure of

SC
central Chile and Argentina. Geophysical Journal International, 165 (1), 383–398.

Giménez, M., Martinez, M.P. Introcaso, A., 2000. A Crustal Model based mainly on

NU
Gravity data in the Area between the Bermejo Basin and the Sierras de Valle Fértil-
MA
Argentina. Journal of South American Earth Sciences, 13 (3), 275-286.

Gómez-Tuena, A., LaGatta, A., Langmuir, C., Goldstein, S., Ortega-Gutiérrez, F.,
D

Carrasco-Núñez, G., 2003. Temporal control of subduction magmatism in the Eastern

TE

Trans-Mexican Volcanic Belt: mantle sources, slab contributions and crustal

P

contamination. Geochemistry, Geophysics, Geosystems, 4(8),

CE

doi:10.1029/2003GC000524.

Gómez-Tuena, A., Langmuir, C.H., Goldstein, S.L., Straub, S.M., Ortega-Gutiérrez,

AC

F., 2007. Geochemical Evidence for Slab Melting in the Transmexican Volcanic Belt.

Journal of Petrology, 48, 537-562, doi:10.1093/petrology/egl071

Gorbatov, A., Fukao, Y., 2005. Tomographic search for missing link between the

ancient Farallon subduction and the present Cocos subduction. Geophysical Journal

International, 160 (3), 849-854.

Goss, A.R., Kay, S.M., 2009. Extreme high field strength element (HFSE) depletion

and near-chondritic Nb/Ta ratios in Central Andean adakite-like lavas (~27°S, ~68°W).

Earth and Planetary Science Letters, 270, 97-109.

65
 ACCEPTED MANUSCRIPT
Goss, A.R., Kay, S.M., Mpodozis, C., 2013. Andean adakite-like high-Mg andesites

on the northern margin of the Chilean–Pampean flat-slab (27–28.5°S) associated with

frontal arc migration and fore-arc subduction erosion. Journal of Petrology, 54, 2193-

2234.

PT
Grange, F., Hatzfeld, D., Cunningham, P., Molnar, P., Roecker, S.W., Suarez, G.,

RI
Rodrigues, A., Ocola, L., 1984. Tectonic implications of the microearthquake seismicity

SC
and fault plane solutions in southern Peru. Journal of Geophysical Research, 89 (B7),

6139-6152.

NU
Gurnis, M., Turner, M., Zahirovic, S., DiCaprio, L., Spasojevic, S., Müller, R.D.,
MA
Boyden, J., Seton, M., Manea, V.C., Bower, D.J., 2012. Plate tectonic reconstructions

with continuously closing plates, Computers and Geosciences, 38, 35-42,

D

doi:10.1016/j.cageo.2011.04.014
TE

Gutscher, M.-A., Olivet, J.-L., Aslanian, D., Eissen, J.-P., Maury, R., 1999. The lost
P

Inca Plateau: cause of flat subduction beneath Peru? Earth and Planetary Science
CE

Letters, 171, 335– 341.

Gutscher, M.A., 2002. Andean subduction styles and their effect on thermal structure
AC

and interplate coupling. Journal of South America Earth Sciences, 15 (1), 3–10.

Gutscher, M.-A., Spakman, W., Bijwaard, H., Engdahl, E.R., 2000. Geodynamics of

flat subduction: seismicity and tomographic constraints from the Andean margin.

Tectonics, 19, 814– 833.

Hacker, B.R., 2008. H2O subduction beyond arcs. Geochemistry, Geophysics,

Geosystems, 9(3), Q03001, doi: 10.1029/2007GC001707.

Hampel, A., 2002. The migration of the Nazca Ridge along the Peruvian active

margin: a re-evaluation. Earth and Planetary Science Letters, 203(2), 665-679.

66
 ACCEPTED MANUSCRIPT
Hamza, V.M., Muñoz, M., 1996. Heat flow map of South America. Geothermics, 25,

599– 646.

Hamza, V.M., Silva Dias, F.J.S., Gomes, A.J.L., Delgadilho Terceros, Z.G., 2005.

Numerical and functional representations of regional heat flow in South America.

PT
Physics of the Earth and Planetary Interiors, 152, 223–256.

RI
Haschke, M.R., Siebel, W., Günther, A., Scheuber, E., 2002. Repeated crustal

SC
thickening and recycling during the Andean orogeny in north Chile (21°–26°S). Journal

of Geophysical Research, 107, 1-18.

NU
Hasegawa, A., Sacks, I.S., 1981. Subduction of the Nazca plate beneath Peru as
MA
determined from seismic observations. Journal of Geophysical Research, 86, 4971–

4980.
D

Heit, B., Yuan, X., Bianchi, M., Sodoudi, F., Kind, R., 2008. Crustal thickness
TE

estimation beneath the southern central Andes at 30°S and 36°S form S wave receiver
P

function analysis. Geophysical Journal International, 174(1), 249-254.

CE

Herrmann, U.R., Nelson, B.K., Ratschbacher, L., 1994. The origin of a terrane: U/Pb

zircon geochronology and tectonic evolution of the Xolapa complex (southern Mexico).
AC

Tectonics, 13(2), 455-474.

Hildreth, W., Moorbath, S., 1988. Crustal contributions to arc magmatism in the

Andes of Central Chile. Contributions to Mineralogy and Petrology 98, 455-489.

Honda S., and Uyeda S., 1983. Thermal process in subduction zones—a review and

preliminary approach on the origin of arc volcanism. In: Shimozuru D., Yokoyama I.,

editors. Arc Volcanism—Physics and Tectonics. Tokyo: Terrapublishing, 117-140.

Hu, J., Liu, L., Hermosillo, A., Zhou, Q., 2016. Simulation of late Cenozoic South

American flat-slab subduction using geodynamic models with data assimilation. Earth

and Planetary Science Letters, 438, 1-13.

67
 ACCEPTED MANUSCRIPT
Huesca-Pérez, E., Valenzuela, R.W., Ortega R., 2016. Crustal anisotropy from

tectonic tremor in Guerrero, Mexico. Geochemistry Geophysics Geosystems 17, 2323-

2335, doi:10.1002/2016GC006358.

Husker, A., Davis, P.M., 2009. Tomography and thermal state of the Cocos plate

PT
subduction beneath Mexico City. Journal of Geophysical Research, 114. B04306,

RI
http://dx.doi.org/ 10.1029/2008JB006039.

SC
Husker, A.L., Kostoglodov, V., Cruz-Atienza, V.M., Legrand, D., Shapiro, N.M.,

Payero, J.S., Campillo, M., Huesca- P rez, E., 2012. Temporal variations of non-

NU
volcanic tremor (NVT) locations in the Mexican subduction zone: Finding the NVT sweet
MA
spot. Geochemistry, Geophysics, Geosystems, 13, Q03011,

doi:10.1029/2011GC003916.
D

Isacks, B.L., Barazangi, M., 1973. High frequency shear waves guided by a
TE

continuous lithosphere descending beneath western South America. Geophysical

P

Journal International, 33(2), 129-139.

CE

Isacks, B.L., Molnar, P., 1971. Distribution of stresses in the descending lithosphere

from a global survey of focal mechanism solutions of mantle earthquakes. Reviews of

AC

Geophysics and Space Physics, 9(1), 103-174.

Isacks, B., Jordan, T., Allmendinger, R.W., Ramos, V.A., 1982. La segmentación

tectónica de los Andes Centrales y su relación con la placa de Nazca subductada, Vth

Congreso Latinoamericano de Geología, Universidad de Buenos Aires, Buenos Aires.

Jadamec, M.A., Billen, M.I. 2012. The role of rheology and slab shape on rapid

mantle flow: Three‐dimensional numerical models of the Alaska slab edge. Journal of

Geophysical Research: Solid Earth, 117(B2).

68
 ACCEPTED MANUSCRIPT
James, D.E. Snoke, J.A., 1990. Seismic evidence for continuity of the deep slab

beneath central and Eastern Peru. Journal of Geophysical Research, 95,

doi:10.1029/89JB03314. issn: 0148-0227.

Jödicke, H., Jording, A., Ferrari, L., Arzate, J., Mezger, K., Rupke, L., 2006. Fluid

PT
release from the subducted Cocos plate and partial melting of the crust deduced from

RI
magnetotelluric studies in southern Mexico. Implications for the generation of volcanism

SC
and subduction dynamics. Journal of Geophysical Research, 111, B08102, doi:

10.1029/2005JB003739.

NU
Jordan, T.E., Allmendinger, R.W. 1986. The Sierras Pampeanas of Argentina: A
MA
modern analogue of Rocky Mountain foreland deformation. American Journal of

Science, 286(10), 737-764.

D

Jordan, T.E., Allmendinger, R.W., Brewer, J.A., Ramos, V.A., Ando, C.J., 1983.
TE

Andean tectonics related to geometry of subducted Nazca plate: Geological Society of

P

America Bulletin, 94, 341–361, doi: 10.1130/00167606.

CE

Jordan, T., Zeitler, P., Ramos, V.A., Gleadow, A.J.W., 1989. Thermochronometric

data on the development of the basement peneplain in the Sierras Pampeanas,

AC

Argentina. Journal of South American Earth Sciences, 2, 207–222.

Jordan, T.E., Drake, R., Nasser, Ch., 1993. Estratigraf a del Cenozoico medio en la

Precordillera a la latitud del R o J chal, San Juan, Argentina, in Actas, Congreso

Geol gico Argentino, 12th, and Congreso de Exploraci n de Hidro- carburos, 2nd,

Mendoza: uenos Aires, Asociaci n Geol gica Argentina, 2, 132-141.

Jordan, T.E., Reynolds III, J.H., Erickson, J.P., 1997. Variability in age of initial

shortening and uplift in Central Andes, 16-33°30’S. in Tectonic uplift nad climate

change, Part II, Ruddiman, W.F. (ed), Springer US, 41-61, doi: 10.1007/978-1-4615-

5935-1_3.
69
 ACCEPTED MANUSCRIPT
Jung, H., Karato, S., 2001. Water-induced fabric transitions in olivine. Science,

293(5534), 1460-1463, doi: 10.1126/science.1062235.

Jung, H., Katayama, I., Jiang, Z., Hiraga, T., Karato, S., 2006. Effect of water and

stress on the lattice preferred orientation (LPO) of olivine. Tectonophysics, 421(1), 1-22.

PT
doi: 10.1016/j.tecto.2006.02.011

RI
Jung, H., 2009. Deformation fabrics of olivine in Val Malenco peridotite found in Italy

SC
and implications for the seismic anisotropy in the upper mantle. Lithos, 109(3), doi:

10.1016/j.lithos.2008.06.007.

NU
Kanjorski, N.M., 2003. Cocos Plate structure along the Middle America Subduction
MA
Zone off Oaxaca and Guerrero, Mexico: influence of subducting plate morphology on

tectonics and seismicity. PhD Thesis, University of California, San Diego.

D

Karato, S., Jung, H., Katayama, I., Skemer, P., 2008. Geodynamic Significance of
TE

Seismic Anisotropy of the Upper Mantle: New Insights from Laboratory Studies. Annual
P

Review of Earth and Planetary Sciences, 36(1), 59-95, doi:

CE

10.1146/annurev.earth.36.031207.124120

Kay, S. M., Maksaev, V., Moscoso, R., Mpodozis, C., Nasi, C., Gordillo, C.E., 1988.
AC

Tertiary Andean magmatism in Chile and Argentina between 28°S and 33°S: Correlation

of magmatic chemistry with a changing Benioff zone. Journal of South American Earth

Sciences, 1, 21-38.

Kay, S. M., Mpodozis, C., Ramos, V. A., Munizaga, F., 1991. Magma source

variations for mid to late Tertiary volcanic rocks erupted over a shallowing subduction

zone and through a thickening crust in the Main Andean Cordillera (28–33°S). In:

Harmon, R.S., Rapela, C. (Eds.), Andean Magmatism and its Tectonic Setting.

Geological Society of America, Special Paper 265, 113–137.

70
 ACCEPTED MANUSCRIPT
Kay, S.M., Abbruzzi, J.M., 1996. Magmatic evidence for Neogene lithospheric

evolution of the central Andean ―flat-slab‖ between 30°S and 32°S. Tectonophysics, 259,

15– 28.

Kay, S.M., Mpodozis, C., 2002. Magmatism as a probe to the Neogene shallowing of

PT
the Nazca plate beneath the modern Chilean flat-slab. Journal of South American Earth

RI
Sciences, 15, 39–57.

SC
Katayama, I., Karato, S., 2006. Effect of temperature on the B- to C-type olivine

fabric transition and implication for flow pattern in subduction zones. Physics of the

NU
Earth and Planetary Interiors, 157, 33-45.
MA
Kim, Y., Clayton, R.W., Jackson, J.M., 2010. Geometry and seismic properties of the

subducting Cocos plate in central Mexico. Journal of Geophysical Research, 115,

D

B06310, doi:10.1029/2009JB006942.
TE

Kim, Y., Clayton, R.W., Keppie, F., 2011. Evidence of a collision between the
P

Yucatán block and Mexico in the Miocene. Geophysical Journal International, 187, 989-
CE

1000, doi: 10.1111/j.1365-246X.2011.05191.x.

Kim, Y., Clayton, R.W., Jackson, J.M., 2012a. Distribution of hydrous minerals in the
AC

subduction system beneath Mexico. Earth and Planetary Science Letters, 341-344, 58-

67.

Kim, Y., Miller, M.S., Pearce, F., Clayton, R.W., 2012b. Seismic imaging of the

Cocos plate subduction zone system in central Mexico. Geochemistry Geophysics

Geosystems 13, Q07001, doi:10.1029/2012GC004033.

Kim, Y., Clayton, R.W., Asimow, P.D., Jackson, J.M., 2013. Generation of talc in the

mantle wedge and its role in subduction dynamics in central Mexico. Earth and

Planetary Science Letters 384, 81-87, doi:10.1016/j.epsl.2013.10.006.

71
 ACCEPTED MANUSCRIPT
Kneller, E.A, van Keken, P.E., Karato, S.-i., Park, J., 2005. B-type olivine fabric in

the mantle wedge: Insights from high-resolution non-Newtonian subduction zone

models, Earth and Planetary Science Letters 237, 781-797,

doi:10.1016/j.epsl.2005.06.049.

PT
Kneller, E.A., van Keken, P.E., Katayama, I., Karato, S., 2007. Stress, strain, and B-

RI
type olivine fabric in the fore-arc mantle: Sensitivity tests using high-resolution steady-

SC
state subduction zone models, Journal of Geophysical Research, 112, B04406,

doi:10.1029/2006JB004544.

NU
Kopp, H., Flueh, E.R., Papenberg, C., Klaeschen, D., 2004. Seismic investigations of
MA
the O'Higgins Seamount Group and Juan Fernandez Ridge: Aseismic ridge

emplacement and lithosphere hydration, Tectonics, 23.

D

Kostoglodov, V., Singh, S.K., Santiago, J.A., Franco, S.I., Larson, K.M., Lowry, A.R.,
TE

Bilham, R., 2003. A large silent earthquake in the Guerrero seismic gap, Mexico,
P

Geophysical Research Letters, 30(15), 1807, doi:10.1029/2003GL017219.

CE

Kostoglodov, V., Husker, A., Shapiro, N.M., Payero, J.S., Campillo, M., Cotte, N.,

Clayton, R., 2010. The 2006 slow slip event and nonvolcanic tremor in the Mexican
AC

subduction zone, Geophys. Res. Lett., 37, L24301. doi:10.1029/2010GL045424.

Kusky, T. M., Windley, B. F., Wang, L., Wang, Z., Li, X., Zhu, P., 2014. Flat slab

subduction, trench suction, and craton destruction: Comparison of the North China,

Wyoming, and Brazilian cratons. Tectonophysics, 630, 208-221.

Le n Soto, G., Ni, J.F., Grand, S.P., Sandvol, E., Valenzuela, R.W., Guzmán

Speziale, M., Gómez González, J.M., Dominguez Reyes, T., 2009. Mantle flow in the

Rivera-Cocos subduction zone. Geophysical Journal International, 179(2), 1004-1012,

doi:10.1111/j.1365-246X.2009.04352.x.

72
 ACCEPTED MANUSCRIPT
Le n Soto, G., Valenzuela, R.W., 2013. Corner flow in the Isthmus of Tehuantepec,

Mexico inferred from anisotropy measurements using local intraslab earthquakes,

Geophysical Journal International, 195, 1230-1238, doi: 10.1093/gji/ggt291.

Lenardic, A., Guillou-Frottier, L., Mareschal, J.-C., Jaupart, C., Moresi, L.-N., Kaula,

PT
W.M., 2000. What the mantle sees: the effect of continents on mantle heat flow. In:

RI
Richards, M.A., Gordon, R.G., van der Hilst, R.D. (Eds.), The History and Dynamics of

SC
Global Plate Motions, Geophysical Monograph. vol. 121. American Geophysical Union,

Washington, DC, 95–112.

NU
Linkimer, L., Beck, S.L., Zandt, G., Alvarado, P.M., Anderson, M.L., Gilbert, H.J.,
MA
Olsen, K. 2011. Appendix C: Geometry of the Wadati-Benioff Zone and Deformation of

the Subducting Nazca Plate in the Pampean Flat Slab of West- Central Argentina in
D

Lithospheric Structure of Pampean Flat Slab (Latitude 30- 33°S) and Northern Costa
TE

Rica (Latitude 9-11°N) Subduction Zones. Linkimer L. PhD Dissertation. University of

P

Arizona, Tucson, 241 pp.

CE

Litvak, V., Poma, E., Kay, S.M., 2007. Paleogene and Neogene magmatism in the

Valle del Cura region: New perspective on the evolution of the Pampean flat slab, San
AC

Juan province, Argentina. Journal of South American Earth Sciences 24, 117-137.

Long, M.D., Silver, P.G., 2008. The subduction zone flow field from seismic

anisotropy: A global view. Science 319, 315-318.

Long, M.D., Silver, P.G., 2009. Mantle flow in subduction systems: The subslab flow

field and implications for mantle dynamics. Journal of Geophysical Research 114,

B10312, doi:10.1029/2008JB006200.

Long, M.D., Becker, T.W., 2010. Mantle dynamics and seismic anisotropy. Earth and

Planetary Science Letters, Frontiers, 297, 341-354.

73
 ACCEPTED MANUSCRIPT
Long, M.D., 2013. Constraints on subduction geodynamics from seismic anisotropy.

Reviews of Geophysics, 51, 76-112.

Long, M.D., Wirth, E.A., 2013. Mantle flow in subduction systems: The wedge flow

field and implications for wedge processes. Journal of Geophysical Research, 118, 583-

PT
606, doi:10.1002/jgrb.50063.

RI
Long, M.D., van der Hilst, R.D., 2005. Upper mantle anisotropy beneath Japan from

SC
shear wave splitting. Physics of the Earth and Planetary Interiors, 151, 206-222.

Lonsdale, P., 2005. Creation of the Cocos and Nazca plates by fission of the

NU
Farallon plate. Tectonophysics, 404, 237- 264.
MA
Lowrie, A., Hey, R., 1981. Geological and geophysical variations along the western

margin of Chile near latitude 33°S and their relation to Nazca plate subduction. Memoir
D

Geological Society of America, 54, 741-754.

TE

Lynner, C., Long, M.D., 2013. Sub-slab seismic anisotropy and mantle flow beneath
P

the Caribbean and Scotia subduction zones: Effects of slab morphology and kinematics,
CE

Earth and Planetary Science Letters 361, 367-378, doi:10.1016/j.epsl.2012.11.007.

Lynner, C., Long, M.D., 2014a. Sub-slab anisotropy beneath the Sumatra and
AC

circum-Pacific subduction zones from source-side shear wave splitting observations.

Geochemistry, Geophysics, Geosystems 15, 2262–2281, doi:10.1002/2014GC005239.

Lynner, C., Long, M.D., 2014b. Testing models of sub-slab anisotropy using a global

compilation of source-side shear wave splitting data, Journal of Geophysical Research

Solid Earth 119, 7226–7244, doi:10.1002/2014JB010983.

Ma, Y., Clayton, R.W., 2014. The crust and uppermost mantle structure of Southern

Peru from ambient noise and earthquake surface wave analysis. Earth and Planetary

Science Letters, 395, 61-70.

74
 ACCEPTED MANUSCRIPT
MacDougall, J.G., Fischer, K.M., Anderson, M.L., 2012. Seismic anisotropy above

and below the subducting Nazca lithosphere in southern South America. Journal of

Geophysical Research 117, B12306. http://dx.doi.org/10.1029/2012JB009538.

Mamani, M., Wörner, G., Sempere, T., 2010. Geochemical variations in igneous

PT
rocks of the Central Andean orocline (13°S to 18°S): Tracing crustal thickening and

RI
magma generation through time and space. Geological Society of America Bulletin, 122,

SC
162-182.

Manea, M., Manea, V.C., Kostoglodov, V., 2003. Sediment fill of the Middle America

NU
Trench inferred from the gravity anomalies, Geofisica Internacional, 42(4), 603-612.
MA
Manea, V. C., Manea, M., Kostoglodov, V., Currie, C.A., Sewell, G., 2004. Thermal

structure, coupling and metamorphism in the Mexican subduction zone beneath

D

Guerrero, Geophysical Journal International, 158(2), 775–784, doi:10.1111/j.1365-

TE

246X.2004.02325.x
P

Manea, V.C., Manea, M., Kostoglodov, V., Sewell, G., 2005. Thermo-mechanical
CE

model of the mantle wedge in Central Mexican subduction zone and a blob tracing

approach for the magma transport, Physics of the Earth and Planetary Interiors, 149,
AC

165-186, doi:10.1016/JPEPI2004.08.024

Manea, V.C., Gurnis, M., 2007. Subduction zone evolution and low viscosity wedges

and channels, Earth and Planetary Science Letters, 264(1-2), 22 – 45.

Manea, M., Manea, V.C., 2011a. Curie point depth estimates and correlation with

subduction in Mexico, Pure and Applied Geophysics, doi: 10.1007/s00024-010-0238-2

Manea, V.C., Manea, M., 2011b. Flat-slab thermal structure and evolution beneath

Central Mexico, Pure and Applied Geophysics, doi: 10.1007/s00024-010-0207-9.

75
 ACCEPTED MANUSCRIPT
Manea, V.C., Perez-Gussinye, M., Manea, M., 2012. Chilean flat slab subduction

controlled by overriding plate thickness and trench rollback, Geology, 40 (1), 35-38; doi:

10.1130/G32543.1.

Manea, V.C., Manea, M., Ferrari, L., 2013. Review Article: A Geodynamical

PT
Perspective on the Subduction of Cocos and Rivera plates beneath Mexico and Central

RI
America. Tectonophysics, doi: 10.1016/j.tecto.2012.12.039.

SC
Marot, M., Monfret, T., Pardo, M., RanalIi, G., Nolet, G., 2013. A double seismic zone

in the subducting Juan Fernandez Ridge of the Nazca Plate (32°S), central Chile.

NU
Journal of Geophysical Research-Solid
MA Earth, 118(7), 3462–3475,

doi:10.1002/jgrb.50240.

Marot, M., Monfret, T., Gerbault, M., Nolet, G., Ranalli, G., Pardo, M., 2014. Flat
D

versus normal subduction zones: a comparison based on 3-D regional traveltime

TE

tomography and petrological modeling of central Chile and western Argentina (29° -
P

35°S). Geophysical Journal International, 199(3), 1633-1654.

CE

Martinod, J., Funiciello, F., Faccenna, C., Labanieh, S., Regard, V., 2005.

Dynamical effects of subducting ridges: insights from 3-D laboratory models,

AC

Geophysical Journal International, 163, 1137-1150.

Matsubara, M., Obara, K., Kasahara, K., 2008. Three-dimensional P-and S-wave

velocity structures beneath the Japan Islands obtained by high-density seismic stations

by seismic tomography. Tectonophysics, 454(1), 86-103.

McCrory, P.A., Blair, J.L., Waldhauser, F., Oppenheimer, D.H., 2012. Juan de Fuca

slab geometry and its relation to Wadati ‐ Benioff zone seismicity. Journal of

Geophysical Research: Solid Earth, 117(B9).

76
 ACCEPTED MANUSCRIPT
Mégard, F., 1987. Cordilleran Andes and Marginal Andes: a review of Andean

geology, north of the Arica elbow (18°S). in: J.W.H. Monger & J. Francheteau Eds.,

Circum-Pacific orogenic belts and evolution of the Pacific Ocean Basin. American

Geophysical Union, Geodynamics Series. 18, 71-95.

PT
Melgar, D., Pérez-Campos X., 2011. Imaging the Moho and subducted oceanic

RI
crust at the Isthmus of Tehuantepec, Mexico, from receiver functions. Pure and Applied

SC
Geophysics, 168, 1449-1460, doi:10.1007/s00024-010-0199-5.

Mercier de Lepinay, B., Michaud, F., Calmus, T., Bourgois, J., Poupeau, G., Saint-

NU
Marc, P., The Nautimate Team, 1997. Large Neogene sub- sidence event along the
MA
Middle America Trench off Mexico (18°-19°N): Evidence from submersible observations,

Geology, 25, 387-390.

D

Molnar, P., Sykes, L.R., 1969. Tectonics of the Caribbean and Middle America
TE

regions from focal mechanisms and seismicity. Geological Society of America Bulletin,
P

80, 1639-1384.
CE

Molnar, P., England, P., 1990. "Late Cenozoic uplift of mountain ranges and global

climate change: chicken or egg?" Nature, 346.6279, 29-34.

AC

Mori, L., Gómez-Tuena, A., Cai, Y., Goldstein, S.L., 2007. Effects of prolonged flat

subduction on the Miocene magmatic record of the central Trans-Mexican Volcanic Belt.

Chemical Geology, 244 (3-4), 452-473.

Muñoz, M., 2005. No flat Wadati-Benioff zone in the central and southern central

Andes. Tectonophysics, 395, 41-65.

Nakajima, J., Hasegawa, A., 2004. Shear-wave polarization anisotropy and

subduction-induced flow in the mantle wedge of northeastern Japan. Earth and

Planetary Science Letters, 225(3-4), 365-377.

77
 ACCEPTED MANUSCRIPT
Norabuena, E.O., Snoke, J.A., James, D.E., 1994. Structure of the subducting Nazca

plate beneath Peru. Journal of Geophysical Research, 99, 9215–9226.

Norabuena, E.O., Dixon, T.H., Stein, S., Harrison, C.G.A., 1999. Decelerating Nazca-

South America and Nazca-Pacific Plate motions, Geophysical Research Letters, 26,

PT
3405-3408.

RI
Obara, K., Tanaka, S., Maeda, T., Matsuzawa, T., 2010. Depth dependent activity of

SC
non-volcanic tremor in southwest Japan, Geophysical Research Letters, 37, L13306,

doi:10.1029/2010GL043679.

NU
Obara K., 2011. Characteristics and interactions between non-volcanic tremor and
MA
related slow earthquakes in the Nankai subduction zone, southwest Japan. Journal of

Geodynamics, 52: 229–248. doi:10.1016/j.jog.2011.04.002.

D

Obrebski, M., Castro, R.R., Valenzuela, R.W., Van Benthem, S., Rebollar, C.J.,
TE

2006. Shear–wave splitting observations at the regions of northern Baja California and
P

southern Basin and Range in Mexico. Geophysical Research Letters, 33, L05302,
CE

doi:10.1029/2005GL024720.

O’Neill, C., Müller, D., Steinberger, B., 2005. On the uncertainties in hot spot
AC

reconstructions and the significance of moving hot spot reference frames,

Geochemistry, Geophysics, Geosystems, 6, Q04003, doi:10.1029/2004GC000784.

Orozco, L.A., Favetto, A., Pomposiello, C., Rossello, E., Booker, J., 2013. Crustal

deformation of the Andean foreland at 31°30’S (Argentina) constrained by

magnetotelluric survey. Tectonophysics, 582, 126-139.

Oxburgh, E.R., Turcotte, D.L., 1970. Thermal structure of island arcs. Bulletin of the

Geological Society of America, 81(6), 1665-1688.

78
 ACCEPTED MANUSCRIPT
Pacheco, J.F., Singh, S.K., 2010. Seismicity and state of stress in Guerrero

segment of the Mexican subduction zone. Journal of Geophysical Research, 115,

B01303, doi:10.1029/2009JB006453.

Pardo, M. Suárez, G., 1995. Shape of the subducted Rivera and Cocos plates in

PT
southern Mexico: Seismic and tectonic implications. Journal of Geophysical Research,

RI
100(B7), 12,357-12,373.

SC
Pardo, M., Comte, D., Monfret, T., 2002. Seismotectonic and stress distribution in the

central Chile. Journal of South American Earth Sciences, 15(1), 11-22.

NU
Park, J., Levin, V., 2002. Seismic anisotropy: Tracing plate dynamics in the mantle.
MA
Science, 296, 485-489.

Peacock, S.M., 1996. Thermal and petrologic structure of subduction zones, in

D

Subduction: Top to Bottom, edited by Bebout, G. E., et al., 119-133, AGU Geophysical
TE

Monograph 96, Washington, D.C.

P

Peacock, S.M., 2003. Thermal structure and metamorphic evolution of subducting

CE

slabs, Geophysical Monograph-American Geophysical Union, 138, 7-22.

Pennington, W.D., 1981. Subduction of the eastern Panama basin and

AC

seismotectonics of northwestern South America. Journal of Geophysical Research, 86,

10,753-10,770.

Perarnau, M., Gilbert, H., Alvarado, P., Martino, R., Anderson, M., 2012. Crustal

structure of the Eastern Sierras Pampeanas of Argentina using high frequency local

receiver functions. Tectonophysics, 580, 208-217.

Pérez-Campos, X., Kim, Y., Husker, A., Davis, P.M., Clayton, R.W., Iglesias, A.,

Pacheco, J.F., Singh, S.K., Manea, V.C., Gurnis, M., 2008. Horizontal subduction and

truncation of the Cocos Plate beneath central Mexico. Geophysical Research Letters 35,

L18303, doi:10.1029/2008GL035127.
79
 ACCEPTED MANUSCRIPT
Petford, N., Atherton, M.P., 1996. Na-rich partial melts from newly underplated

basaltic crust: the Cordillera Blanca batholith, Peru. Journal of Petrology 37, 1491-1521.

Phillips, K., Clayton, R.W., Davis, P., Tavera, H., Guy, R., Skinner, S., Stubailo, I.,

Audin, L., Aguilar, V., 2012. Structure of the subduction system in southern Peru from

PT
seismic array data, Journal of Geophysical Research, 117, B11306,

RI
doi:10.1029/2012JB009540.

SC
Phillips, K., Clayton, R.W., 2014. Structure of the subduction transition region from

seismic array data in southern Peru. Geophysical Journal International, 196, 1889–

NU
1905.
 MA
Pilger Jr, R.H., 1981. Plate reconstructions, aseismic ridges, and low-angle

subduction beneath the Andes. Bulletin of Geological Society of America, 92(7), 448-
D

456.
TE

Pilger Jr., R.H., 1984. Cenozoic plate kinematics, subduction and magmatism: South
P

American Andes. J. Geol. Soc. (Lond.) 141, 793– 802.

CE

Pollack, H.N., 1980. The heat flow from the Earth: a review. In Mechanisms of

Continental Drift and Plate Tectonics (pp. 183-192). Academic Press London.
AC

Ponce-Cortés, J.G., 2012. Medición de la anisotropía de las ondas SKS en el manto

superior, debajo de las estaciones permanentes del Servicio Sismológico Nacional

instaladas a partir del año 2005. B. Sc. thesis, 79 pp., Facultad de Ingeniería,

Universidad Nacional Autónoma de México, Mexico City, Mexico.

Porter, R., Gilbert, H., Zandt, G., Beck, S., Warren, L., Calkins, J., Alvarado, P., and

Anderson, M., 2012. Shear wave velocities in the Pampean flat-slab region from

Rayleigh wave tomography: implications for slab and upper mantle hydration. Journal of

Geophysical Research 117, B11301. http://dx.doi.org/10.1029/2012JB009350.

80
 ACCEPTED MANUSCRIPT
Radiguet, M., Cotton, F., Vergnolle, M, Campillo, M., Walpersdorf, A., Cotte, N.,

Kostoglodov, V., 2012. Slow slip events and strain accumulation in the Guerrero gap,

Mexico. Journal of Geophysical Research, 117(B4), B04305, doi:

10.1029/2011JB008801.

PT
Ramos, V.A., Munizaga, F., Kay, S.M., 1991. El magmatismo cenozoico a los 33ºS

RI
de latitud: Geocronología y relaciones tectónicas. VI Congreso Geológico Chileno:

SC
Actas 1, pp. 892–896.

Ramos, V.A., 1999. Plate tectonic setting of the Andean Cordillera. Episodes, 22,

NU
183-190. MA
Ramos, V.A., Cristallini, E.O., Pérez, D.J., 2002. The Pampean flat slab of the

Central Andes. Journal of South America Earth Sciences. 15, 59– 78

D

Ramos, V.A., Folguera, A., 2009, Andean flat-slab subduction through time. In:
TE

Murphy, J.B. Keppie, J.D., Hynes, A.J. (Eds.), Ancient Orogens and Modern Analogues.
P

Geological Society, London, Special Publications 327, 31-54. doi:10.1144/SP327.3

CE

Ranero C.R., Phipps Morgan J., McIntosh K., Reichert C., 2003, Bending-related

faulting and mantle serpentinization at the Middle America trench: Nature, v. 425, p.
AC

367–373, doi:10.1038/nature01961.

Ratchkovski, N.A., Hansen, R.A., 2002. New evidence for segmentation of the

Alaska subduction zone. Bulletin of the Seismological Society of America, 92(5), 1754-

1765.

Ratschbacher, L., Franz, L., Min, M., achmenn, R., Martens, U., Stanek, K.,

Stubner, K., Nelson, .K., Herrmann, U., Weber, ., L pez-Mart nez, M., Jonckheere,

R., Sperner, B., Tichomirowa, M., McWilliams, M.O., Gordon, M., Meschede, M., Bock,

P., 2009. The North American–Caribbean plate boundary in M xico–Guatemala–

Honduras. In: James, K.H., Lorente, M.A., Pindell, J.L. (eds.). The origin and Evolution
81
 ACCEPTED MANUSCRIPT
of the Caribbean Plate. Geological Society of London, 328 (Special Publications), 219-

293.

Regnier, M., Chiu, J.M., Smalley Jr., R., Isacks, B.L., Araujo, M., 1994. Crustal

thickness variation in the Andean foreland, Argentina, from converted waves. Bulletin of

PT
the Seismological Society of America 84 (4), 1097–1111.

RI
Rodriguez-Gonzales, J., Negredo, A.M., Billen, B.I., 2012. The role of the overriding

SC
plate thermal state on slab dip variability and on the occurrence of flat subduction,

Geochemistry Geophysics Geosystems, 13, Q01002, doi:10.1029/2011GC003859.

NU
Rodríguez-Pérez Q., 2007. Estructura tridimensional de velocidades para el sureste
MA
de México, mediante el análisis de trazado de rayos sísmicos de sismos regionales. M.

Sc. thesis, 83 pp., Instituto de Geofísica, Universidad Nacional Autónoma de México,

D

Mexico City, Mexico.

TE

Rojo-Garibaldi, B., 2011. Anisotropía de las ondas SKS en el manto superior debajo
P

de un arreglo sísmico entre Guerrero y Veracruz. B. Sc. thesis, 84 pp., Facultad de

CE

Ciencias, Universidad Nacional Autónoma de México, Mexico City, Mexico.

Rojo-Garibaldi, B., Ponce Cortés, G., Valenzuela Wong, R., Stubailo, I., Davis, P.M.,
AC

Husker, A., Pérez Campos, X., Iglesias, A., Clayton, R.W., 2011. Anisotropía de la onda

SKS en el manto superior debajo del arreglo MASE y las nuevas estaciones del Servicio

Sismológico Nacional (abstract). Geos Boletín Informativo de la Unión Geofísica

Mexicana 31 (1), 116.

Roy, R.F., Decker, E.R., Blackwell, D.D., 1972. Continental heat flow, in Robertson,

E.C., ed., The nature of the solid Earth: New York, McGraw-Hill, 506–543.

Russo, R., Silver, P.G., 1994. Trench-Parallel Flow Beneath the Nazca Plate from

Seismic Anisotropy. Science, 263, 1105-1111, doi:10.1126/science.263.5150.1105.

82
 ACCEPTED MANUSCRIPT
Rüpke, L.H., Morgan, J.P., Hort, M., Connolly, J.A.D., 2004. Serpentine and the

subduction zone water cycle. Earth and Planetary Science Letters, 223, 17–34.

Savage, M.K., 1999. Seismic anisotropy and mantle deformation: What have we

learned from shear wave splitting? Review of Geophysics, 37, 65–106.

PT
Schellart, W,P., Freeman, J., Stegman, D.R., Moresi, L., May, D., 2007. Evolution

RI
and diversity of subduction zones controlled by slab width. Nature. 446:308-311.

SC
Schellart, W.P., Stegman, D.R., Freeman, J., 2008. Global trench migration

velocities and slab migration induced upper mantle volume fluxes: constraints to an

NU
Earth reference frame based on minimizing viscous dissipation. Earth-Science Reviews,
MA
88, 118–144, doi: 10.1016/j.earscirev.2008.01.005.

Schellart, W.P., Stegman, D.R., Farrington, R.J., Moresi L., 2011. Influence of
D

lateral slab edge distance on plate velocity, trench velocity, and subduction partitioning.
TE

Journal of Geophysical Research-Solid Earth. 116

P

Sdrolias, M., Muller, R.D., 2006. Controls on back-arc basin formation.

CE

Geochemistry, Geophysics, Geosystems 7(4), Q04016, doi: 10.1029/2005GC001090

Sebrier, M., Soler, P., 1991. Tectonics and magmatism in the Peruvian Andes from
AC

upper Oligocene to present. Geological Society of America, Special Paper 265 "Andean

magmatism and its tectonic setting", R. Harmon & C. Rapela, eds., 259-278.

Sedlock, R., Ortega-Gutiérrez, F., Speed, R., 1993. Tectonostratigraphic terranes

and the tectonic evolution of Mexico. Geological Society of America Special Paper 278,

153.

Shapiro, N.M., Ritzwoller, M.H., 2004. Thermodynamic constraints on seismic

inversions. Geophysical Journal International 157, doi:10.1111/j.1365-

246X.2004.02254.x.

83
 ACCEPTED MANUSCRIPT
Shelly, D.R., Beroza, G.C., Ide, S., 2007. Non-volcanic tremor and low-frequency

earthquake swarms. Nature, 446 (7133), 305–307, doi:10.1038/nature05666.

Skinner, S.M., Clayton, R.W., 2013. The lack of correlation between flat slabs and

bathymetric impactors in South America. Earth and Planetary Science Letters, 371, 1-5.

PT
Silver, P.G., Chan, W.W., 1988. Implications for continental structure and evolution

RI
from seismic anisotropy. Nature 335, 34-39.

SC
Silver, P.G., Chan, W.W., 1991. Shear wave splitting and subcontinental mantle

deformation. Journal of Geophysical Research 96, 16,429-16,454.

NU
Silver, P.G., 1996. Seismic anisotropy beneath the continents: Probing the depths of
MA
Geology. Annual Reviews of Earth and Planetary Science 24, 385-432.

Singh, S.K., Pérez-Campos, X., Espíndola, V.H., Cruz-Atienza, V.M., Iglesias, A.,
D

2014. Intraslab earthquake of 16 June 2013 (Mw 5.9), One of the closest such events to
TE

Mexico City. Seismological Research Letters 85, 268-277, doi:10.1785/0220130179.

P

Smithson, S.B., Decker, E.R., 1974. A continental crustal model and its geothermal
CE

implications. Earth and Planetary Science Letters, 22, 215-225.

Smithson, S.B., Brown, S.K., 1977. A model for the lower continental crust Earth and
AC

Planetary Science Letters, 35, 134-144.

Song, T.-R.A., Kawakatsu, H., 2012. Subduction of oceanic asthenosphere:

Evidence from sub-slab seismic anisotropy. Geophysical Research Letters 39, L17301,

doi:10.1029/2012GL052639.

Song, T.-R.A., Kim, Y., 2012a. Anisotropic uppermost mantle in young subducted

slab underplating central Mexico. Nature Geoscience 5, 55-59, doi:10.1038/ngeo1342.

Song, T.-R.A., Kim Y., 2012b. Localized seismic anisotropy associated with long-

term slow-slip events beneath southern Mexico. Geophysical Research Letters 39,

L09308, doi:10.1029/2012GL051324.
84
 ACCEPTED MANUSCRIPT
Song, T.A., Helmberger, D.V., Brudzinski, M.R., Clayton, R.W., Davis, P., Pérez-

Campos, X., Singh, S.K., 2009. Subducting slab ultra-slow velocity layer coincident with

silent earthquakes in Southern Mexico. Science 324, 502–506, doi:

10.1126/science.1167595.

PT
Soyer, W., Unsworth, M., 2006. Deep electrical structure of the northern Cascadia

RI
(British Columbian, Canada) subduction zone: implications for the distribution of fluids.

SC
Geology, 34(1), 53-56.

Stern, C.R., 1991. Role of subduction in the generation of Andean magmas. Geology

NU
19, 78-81. MA
Stern, C.R., 2011. Subduction erosion: rates, mechanisms, and its role in arc

magmatism and the evolution of the continental crust and mantle. Gondwana Research,
D

v. 20, p. 284-308
TE

Stoiber, R.E., Carr, M.J., 1973. Quaternary volcanic and tectonic segmentation of
P

Central America. Bulletin of Volcanology., 37-3, 304-325.

CE

Stubailo, I., 2015. Seismic anisotropy below Mexico and its implications for mantle

dynamics, Ph.D. thesis, 119 pp., University of California, Los Angeles, CA, USA.
AC

Stubailo, I., Davis, P., 2007. Shear wave splitting measurements and interpretation

beneath Acapulco-Tampico transect in Mexico, Eos Transactions of the American

Geophysical Union 88 (52), Fall Meeting Supplement Abstract T51B-0539.

Stubailo, I., Davis, P.M., 2012a. Anisotropy of the Mexico subduction zone based on

shear-wave splitting analysis (abstract), Seismological Research Letters 83 (2), 379.

Stubailo, I., Davis, P.M., 2012b. Anisotropy of the Mexico subduction zone based on

shear-wave splitting and higher modes analysis, Abstract T11A-2538 presented at 2012

Fall Meeting, American Geophysical Union, San Francisco, CA, 3-7 December.

85
 ACCEPTED MANUSCRIPT
Stubailo, I., Davis, P., 2014. Seismic anisotropy of the Mexican subduction zone

based on the surface waves, shear wave splitting, and higher modes, Abstract DI33A-

4302 presented at 2014 Fall Meeting, American Geophysical Union, San Francisco, CA,

15-19 December.

PT
Stubailo, I., Davis, P.M., 2015. The surface wave, shear wave splitting, and higher

RI
mode seismic anisotropy comparison of the Mexican subduction zone (abstract),

SC
Seismological Research Letters 86 (2B), 677.

Suárez, G., Molnar, P., Burchfiel, B.C., 1983. Seismicity, fault plane solutions, depth

NU
of faulting, and active tectonics of the Andes of Peru, Ecuador and southern Colombia.
MA
Journal of Geophysical Research, 88, 10403– 10428.

Suárez, G., Gagnepain, J., Cisternas, A., Hatzfeld, D., Molnar, P., Ocola, L.,
D

Roecker, S.W., Viodé, J.P., 1990. Tectonic deformation of the Andes and the
TE

configuration of the subducted slab in central Peru: results from a microseismic

experiment. Geophysical Journal International, 103, 1 – 12.

P
CE

Tan, E., Choi, E., Thoutireddy, P., Gurnis, M., Aivazis, M., 2006. GeoFramework:

coupling multiple models of mantle convection within a computational framework.

AC

Geochemistry Geophysics Geosystems 7. doi:10.1029/2005GC001155 Q06001, 14 pp.

Taramon, J.M., Rodriguez-Gonzalez, J., Negredo, A.M., Billen, M.I., 2015. Influence

of cratonic lithosphere on the formation and evolution of flat slabs: insights from 3-D time

dependend modeling. Geochemistry, Geophysics, Geosystems, 16,

doi:10.1002/2015GC005940

Tavera, H., Buforn, E., 2001. Source mechanism of earthquakes in Peru. Journal of

Seismology 5, 519– 539.

86
 ACCEPTED MANUSCRIPT
Thomas, W.A., Astini, R.A., 2003. Ordovician accretion of the Argentine Precordillera

terrane to Gondwana: a review. Journal of South American Earth Sciences 16, 67–79,

doi: 10.1016/S0895-9811(03)00019-1.

Toksöz, M., Nafi, J., Minear, W., Julian, B.R., 1971. Temperature field and

PT
geophysical effects of a downgoing slab. Journal of Geophysical Research, 76, 1113-

RI
1138.

SC
Tommasi, A., Vauchez, A., Godard, M., Belley, F., 2006. Deformation and melt

transport in a highly depleted peridotite massif from the Canadian Cordillera:

NU
Implications to seismic anisotropy above subduction zones. Earth and Planetary
MA
Science Letters, 252, 245-259.

Turcotte, D.L., Schubert, G., 1973. Frictional heating of the descending lithosphere.
D

Journal of Geophysical Research, 78(26), 5876-5886.

TE

Turcotte, D.L., Schubert, G., 2014. Geodynamics, Cambridge University Press.

P

Unsworth, M., Rondenay, S., 2013. Actively observing fluid movement in the mid to
CE

deep crust and lithospheric mantle utilizing geophysical methods, solicited chapter. In:

Harlov D, Austrheim H (eds) Metasomatism and metamorphism: the role of fluids in

AC

crustal and upper mantle processes. Lecture Notes in Earth System Sciences, Springer,

535–598, ISSN: 2193-8571.

Uyeda, S., Kanamori, H., 1979. Back-arc opening and the mode of subduction.

Journal of Geophysical Research, vol. 84, no. B3. 1049-1061.

van Benthem, S.A.C., 2005. Anisotropy and flow in the uppermantle under Mexico.

M. Sc. thesis, 41 pp., Utrecht University, Utrecht, The Netherlands.

van Benthem, S.A.C., Valenzuela, R.W., Obrebski, M., Castro, R.R., 2008.

Measurements of upper mantle shear wave anisotropy from stations around the

southern Gulf of California. Geofísica Internacional, 47 (2008), pp. 127–143

87
 ACCEPTED MANUSCRIPT
van Benthem, S.A.C., Valenzuela, R.W., Ponce, G.J, 2013. Measurements of upper

mantle shear wave anisotropy from a permanent network in southern Mexico. Geofísica

Internacional 52, 385-402.

van den Beukel, J., Wortel, R., 1988. Thermo-Mechanical modelling of arc-trench

PT
regions. Tectonophysics, 154, 177-193.

RI
van Hunen, J., van den Berg, A.P., Vlaar, N.J., 2000. A thermomechanical model of

SC
horizontal subduction below an overriding plate. Earth and Planetary Science Letters,

182, 157– 169.

NU
van Hunen, J., van den Berg, A.P., Vlaar, N.J., 2002. On the role of subducting
MA
oceanic plateaus in the development of shallow flat subduction. Tectonophysics 352,

317– 333.
D

van Hunen, J., van den Berg, A.P., Vlaar, N.J., 2004. Various mechanisms to induce
TE

present ‐ day shallow flat subduction and implications for the younger Earth: A
P

numerical parameter study, Earth and Planetary Science Letters, 146, 179–194.
CE

Vannucchi, P., Sak., P.B., Morgan, J.P., Ohkushi, K., Ujiie, K., the IODP Expedition

Shipboard Scientists, 2013. Rapid pulses of uplift, subsidence, and subduction erosion
AC

offshore Central America: implications for building the rock record of convergent

margins. Geology, 41 (9), 995-998.

Vauchez, A., Tommasi, A., Barruol, G., Maumus, J., 2000. Upper mantle deformation

and seismic anisotropy in continental rifts. Physics and Chemistry of the Earth, Part A:

Solid Earth and Geodesy, 25(2), 111-117.

Vinnik, L.P., Kind, R., 1993. Ellipticity of teleseismic S-particle motion. Geophysical

Journal International 113, 165-174.

88
 ACCEPTED MANUSCRIPT
Vitorello, I., Hamza, V.M., Pollack, H.N., 1980. Terrestrial heat flow in the Brazilian

highlands. Journal of Geophysical Research, 85, 3778-3788.

Wagner, L.S., Beck, S., Zandt, G., 2005. Uppermantle structure in the south central

Chilean subduction zone (30° to 36°S). Journal of Geophysical Research 110, doi:

PT
10.1029/2004JB003238.

RI
Wagner, L.S., Beck, S.L., Zandt, G., Ducea, M., 2006. Depleted lithosphere, cold,

SC
trapped asthenosphere, and frozen melt puddles above the flat slab in central Chile and

Argentina, Earth and Planetary Science Letters, 245, 289-301.

NU
Wagner, L.S., Anderson, M., Jackson, J.M., Beck, S.L., Zandt, G., 2008. Seismic
MA
evidence for orthopyroxene enrichment in the continental lithosphere, Geology, 36, 935-

938.
D

Wallace, P., Carmichael, I., 1999. Quaternary volcanism near the Valley of Mexico:
TE

implications for subduction zone magmatism and the effects of crustal thickness
P

variations on primitive magma compositions. Contributions to Mineralogy and Petrology,

CE

135, 291–314, doi: 10.1007/s004100050513.

Ward, K.M., Porter, R.C., Zandt, G., Beck, S.L., Wagner, L.S., Minaya, E., Tavera,
AC

H., 2013. Ambient noise tomography across the Central Andes. Geophysical Journal

International, doi: 10.1093/gji/ggt166.

Wech, A.G., Creager, K.C., Melbourne, T.I., 2009. Seismic and geo- detic constraints

on Cascadia slow slip, Journal of Geophysical Research, 114, B10316,

doi:10.1029/2008JB006090.

Wech, A.G., Bartlow, N.M., 2014. Slip rate and tremor genesis in Cascadia.

Geophysical Research Letters, 41(2), pp. 392-398, doi: 10.1002/2013GL058607.

Wortel, R., Cloetingh, S., 1981. On the origin of the Cocos-Nazca spreading center.

Geology, 9, 425-430.
89
 ACCEPTED MANUSCRIPT
Wright, N M., Seton, M., Williams, S.E., Müller, R.D., 2016. The Late Cretaceous to

recent tectonic history of the Pacific Ocean basin. Earth Science Reviews, 154, 138–

173. http://doi.org/10.1016/j.earscirev.2015.11.015

Yang, X., Fischer, K.M., Abers, G.A., 1995. Seismic anisotropy beneath the

PT
Shumagin Islands segment of the Aleutian-Alaska subduction zone. Journal of

RI
Geophysical Research, 100, 18,165-18,177.

SC
Yang, T., Grand, S.P., Wilson, D., Guzmán-Speziale, M., Gómez-González, J.M.,

Domínguez-Reyes, T., Ni, J., 2009. Seismic structure beneath the Rivera subduction

NU
zone from finite-frequency seismic tomography. Journal of Geophysical Research, 114,
MA
B01302, doi:10.1029/2008JB005830.

Yáñez, G., Ranero, C., von Huene, R., Díaz, J., 2001. Magnetic anomaly
D

interpretation across the southern Central Andes (32°-33.5°S): the role of the Juan
TE

Fernández ridge in the late Tertiary evolution of the margin. Journal of Geophysical
P

Research, 106, 6325-6345.

CE

Yepes, H., Audin, L., Alvarado, A., Beauval, C., Aguilar, J., Font, Y., Cotton, F.,

2016. A new view for the geodynamics of Ecuador: Implication in seismogenic source
AC

definition and seismic hazard assessment. Tectonics.

Zhao, D., 2007. Seismic images under 60 hotspots: Search for mantle plumes.

Gondwana Research, 12, 335-355.

Ziagos, J., Blackwell, D., Mooser, F., 1985. Heat flow in southern Mexico and the

thermal effects of subduction. Journal of Geophysical Research 90, 5410–5420.

Zigone, D., Rivet, D., Radiguet, M., Campillo, M., Voisin, C., Cotte, N., Walpersdorf,

A., Shapiro, N., Cougoulat, G., Roux, P., Kostoglodov, V., Husker, A., Payero, J.S.,

2012. Triggering of tremors and slow slip event in Guerrero, Mexico, by the 2010 Mw 8.8

90
 ACCEPTED MANUSCRIPT
Maule, Chile, earthquake. Journal of Geophysical Research, 117 (B09304), doi:

10.1029/2012JB009160

Zimmerman, M. E., Zhang, S. Q., Kohlstedt, D. L., Karato, S., 1999. Melt distribution

in mantle rocks deformed in shear, Geophysical Research Letters, 26, 1505 – 1508,

PT
doi:10.1029/1999GL900259.

RI
SC
NU
MA
D
P TE
CE
AC

91
 ACCEPTED MANUSCRIPT
Figure Captions.

PT
RI
SC
NU
MA
D
P TE

Figure 1. The Mexican, Peruvian and Chilean flat-slab bathymetric and

CE

topographic maps (A, C). Three-dimensional visualization of the Mexican and Andean
AC

subduction zones (B, D). Large curved gray arrows are used to facilitate viewing of

subducting slabs geometry beneath Mexico and South America. Surface relief is shown

as a semi-transparent layer. Labeled black contours indicate depths to the slab surface

from the Earth’s surface. Red arrows show the direction of the Cocos and Nazca plates

movement relative to North (NAM) and South America (SAM) plates. EPR - East Pacific

Rise, MAT - Middle America Trench, SAT - South America Trench, NR – Nazca Ridge,

IQR – Iquique Ridge, JFR – Juan Fernandez Ridge. Central inset – global view of

present day tectonic plates: JF – Juan de Fuca, Co – Cocos, R – Rivera, NA – North

America, P – Pacific, SA – South America, CA – Caribbean, N – Nazca.

92
 ACCEPTED MANUSCRIPT

PT
RI
SC
Figure 2. A. Main crustal units corresponding to the Mexico flat-slab region:

NU
TMVB - Trans Mexican Volcanic Belt, GT – Guerrero Terrane, MxT – Mixteco Terrane,

MyT – May Terrane, XT – Xolapa Terrane, OxT – Oaxaca Terrane. B. Major

MA
morphotectonic provinces in South America corresponding to flat-slab areas in Peru and

Chile: FA - Forearc, WC - Western Cordillera, EC - Eastern Cordillera, SA – Subandes,

D
TE

AP- Altiplano, AA- Arequipa-Antofalla coastal block, PN – Puna, MC - Main Cordillera,

PC - Principal Cordillera, SP - Sierras Pampeanas, Pc. – Precordillera, FC - Frontal

P

Cordillera.
CE
AC

93
 ACCEPTED MANUSCRIPT

PT
RI
SC
NU
MA
D
P TE
CE
AC

Figure 3. Vertical cross-sections through the Mexican (earthquakes from SSN

(Servicio Sismologico Nacional)), Peruvian and Chilean flat-slabs based on Kim et al.

(2012b), Phillips and Clayton (2014) and Marot et al. (2014). Light blue surface depicts

the oceanic crust; red surface represents the oceanic lithosphere; purple dots show the

earthquakes location.

94
 ACCEPTED MANUSCRIPT

PT
RI
SC
NU
MA
D
P TE
CE
AC

Figure 4. Moho geometry for Mexico, Peru and Chile flat-slab regions based on

seismic experiments of Perez-Campos et al. (2008), Kim et al. (2012b), Phillips and

Clayton (2014), and Marot et al. (2014).

95
 ACCEPTED MANUSCRIPT

PT
RI
SC
NU
MA
D
P TE
CE
AC

Figure 5. A. GPS stations and total displacements (red vectors) produced by the

2006 SSE in central Mexico (Kostoglodov et al., 2010); white dashed square represents

the Guerrero Seismic Gap, and yellow dashed square represents the Sweet Spot. B.

Displacements records at CAYA permanent GPS station in Mexico for a time span of

96
 ACCEPTED MANUSCRIPT
more that 16 years. Note the regular recurrence of SSEs in the flat-slab area in central

Mexico. Gray band depicts the 2006 SSE shown in A. C. Temporal distribution (mid

2005-2007) of the NVT energy along a profile over the flat-slab in central Mexico and the

good correlation with high-NVT energy release during the 2006 SSE. D. Seismogenic,

PT
transit-SSEs and NVTs locations along the flat-slab interface in central Mexico.

RI
SC
NU
MA
D
P TE
CE
AC

97
 ACCEPTED MANUSCRIPT

PT
RI
SC
NU
MA
D
P TE
CE
AC

Figure 6. A. Composite dVs/Vs image for a 500 km transect across central

Mexico from Kim et al. (2010). Note the presence of a low-velocity region beneath the

active volcanic arc. B. S-wave velocity structure for the Peruvian flat-slab from Ma and

Clayton (2014). Note the clear presence of a mid-crustal low-velocity zone above the

flat-slab segment. C. Model perturbations for S-wave through the Chilean flat-slab of

98
 ACCEPTED MANUSCRIPT
Marot et al. (2014). Similarly as for Peru, there are several low velocity anomalies

located above the flat-slab segment.

PT
RI
SC
NU
MA
D
P TE
CE
AC

99
 ACCEPTED MANUSCRIPT

PT
RI
SC
NU
MA
D
TE
P
CE
AC

100
 ACCEPTED MANUSCRIPT
Figure 7. Schematic views of the flat-slab geometries in central Mexico, Peru and

central Chile subduction zones, and the relationship with observed seismic anisotropy

(purple arrows). Blue arrows represent the oceanic plates velocities, and yellow arrows

show the overriding plates velocities. Note the consistent trench perpendicular sub-slab

PT
anisotropy observed for all three flat-slabs. A. Trench-perpendicular shear-wave fast

RI
axes are observed in Mexico beneath the slab, and perpendicular to the strike of the

SC
slab for the mantle wedge above the slab. Notice that the trench and the volcanic arc

are not subparallel. B. In Peru the southward migration of the Nazca ridge (variable blue

NU
tone ribbons) through time seems to induce a complicated pattern of mantle anisotropy
MA
above the flat-slab segment. To the south, the sub-slab mantle is trench-subparallel. C.

In central Chile the mantle anisotropy distribution is comparable with central Mexico. RP
D

– Rio de la Plata craton.

P TE
CE
AC

101
 ACCEPTED MANUSCRIPT

PT
RI
SC
NU
MA
D
P TE
CE
AC

Figure 8. A. Heat flow map of central Mexico based on spectral analysis of

aeromagnetic anomalies (Manea and Manea, 2011a). B. Heat flow map of South

America based on spherical harmonic expansion to degree 36 of the global heat flow

(Hamza et al, 2005; Cardoso et al., 2010). Note the reduced heat flow values for all

three regions located above the flat-slabs.

102
 ACCEPTED MANUSCRIPT

PT
RI
SC
NU
MA
D
TE

Figure 9. A. Magnetotelluric profile above the Mexican flat-slab from Jödicke et al.
P

(2006). Note the presence of a large conductive region beneath the TMVB and above
CE

Moho. B. Magnetotelluric profile across the eastern edge of the Sierras Pampeanas

above the Chilean flat-slab from Booker et al. (2004). Note the presence of a lower crust
AC

conductor beneath the Sierras de Cordoba (SC), and also of a large vertical conductor

at mantle depths above the slab. Semitransparent thin gray band represents the slab

surface. White dashed square marks the high resistivity region located between the

Andes and the Sierras Pampeanas. Other notations are like in figure 2.

103
 ACCEPTED MANUSCRIPT

PT
RI
SC
NU
Figure 10. A. Global paleoreconstructions models based on G-Plates (Gurnis et
MA
al., 2012) and the location and evolution of flat-slab areas in Mexico, Peru and Chile

(gray ovals). Plate notations are as in figure 1-inset. MxFS – Mexican flat-slab, PFS –
D

Peruvian flat-slab, ChFS – Chilean flat-slab. B. Cocos and Nazca incoming plate age
TE

and velocity development at flat-slab latitudes. Gray band represents the onset of flat-

slab.
P
CE
AC

104
 ACCEPTED MANUSCRIPT

PT
RI
SC
NU
Figure 11. Present-day thermal structure of flat-slabs in central Mexico (A)

(Manea and Manea, 2011b) and Chile (B) (Marot et al., 2014). C,D – Dehydration
MA
models along the slab interface.
D
P TE
CE
AC

105
 ACCEPTED MANUSCRIPT

PT
RI
SC
NU
MA
D
P TE
CE
AC

Figure 12. A. Migration of the Miocene magmatic arc during slab flattening in the

Mexican subduction zone. Locus of volcanism migrated ca. 220 km between early and

late Miocene to distances of about 480 km from the trench. B. Migration of the magmatic

arc in the Pampean flat segment from Oligocene to present. Locus of magmatism

106
 ACCEPTED MANUSCRIPT
reached a distance of ca. 750 km from the trench in the late Miocene to Pliocene. Based

on Kay and Mpodozis (2002).

PT
RI
SC
NU
MA
D
P TE
CE
AC

107
 ACCEPTED MANUSCRIPT

PT
RI
SC
NU
MA
D

Figure 13. A. Comparative view of 2D numeric dynamic models of subduction that

TE

produce flat slabs (adapted from van Hunen et al. (2000, 2002, 2004)). Horizontal

eastward mantle flow, buoyant oceanic plateaus and overthrusting facilitate slab
P
CE

flattening. Note how overthrusting (or trench roll-back) is one of the most efficient ways

to promote flat-slabs.
AC

108
 ACCEPTED MANUSCRIPT

PT
RI
Figure 13. B. The strong effect of trench dynamics (rollback or rollforward) on slab

SC
geometry is also demonstrated by Manea and Gurnis (2007). Advanced 3D thermo-

mechanical numeric models petrologically show that buoyant ridges (as Nazca ridge)

NU
alone are not able to induce slab flattening (Gerya, 2009). Gerbault et al. (2009)
MA
demonstrate that slab flattening is produced when rheological properties of the

subducting oceanic plate and the overriding continental plate are interchanged.
D
P TE
CE
AC

109
 ACCEPTED MANUSCRIPT

PT
RI
SC
NU
MA
D
P TE
CE

Figure 13. C. The effect of overriding plate structure (i.e. thickness) on slab geometry
AC

evolution is demonstrated by Manea et al. (2012), Rodriguez-Gonzales et al. (2012) (2D

models) and confirmed recently by 3D models of Taramon et al. (2015).

110
 ACCEPTED MANUSCRIPT

PT
RI
SC
NU
MA
D
P TE
CE
AC

Table 1. Comparative table of the observed physical parameters specific for flat-

slab subduction regions in Mexico, Peru and Chile. Among them, trench rollback,

subducting plate structure and tectonic evolution, as well as strong discontinuities in the

overriding plate structure are the most important in facilitating slab flattening.
111