Flat Slab Subduction, Trench Suction, and Craton Destruction: Comparison of the North China, Wyoming, and Brazilian Cratons

Abstract

The mechanisms of growth and destruction of continental lithosphere have been long debated. We define and test a unifying plate tectonic driving mechanism that explains the numerous petrological, geophysical, and geological features that characterize the destruction of cratonic lithospheric roots. Data from three Archean cratons demonstrate that loss of their roots is related to rollback of subducted flat slabs, some along the mantle transition zone, beneath the cratons. During flat slab subduction dehydration reactions add water to the overlying mantle wedge. As the subducting slabs roll back, they suck in mantle material to infill the void space created by the slab roll back, and this fertile mantle becomes hydrated. The roll-back causes concomitant lithospheric thinning of the overlying craton so the flux of newly hydrated mantle material inevitably rises causing adiabatic melting, generating new magmas that gradually destroy the roots of the overlying craton through melt-peridotite reactions. Calculated fluxes of new mantle material beneath cratons that have lost their roots range from 2.7 trillion to 70 million cubic kilometers, which is sufficient to generate enough melt to completely replace the affected parts of the destroyed cratons. Cratonic lithosphere may be destroyed in massive quantities through this mechanism, warranting a re-evaluation of continental growth rates with time.

Full text

Flat slab subduction, trench suction, and craton destruction: Comparison
of the North China, Wyoming, and Brazilian cratons
Timothy M. Kusky
a,b,c,
⁎, Brian F. Windley
c,d
, Lu Wang
a,c
, Zhensheng Wang
c,e
, Xiaoyong Li
e,f
,PeiminZhu
e,f
a
State Key Lab for Geological Processes and Mineral Resources, China University of Geosciences, Wuhan, China
b
Three Gorges Research Center for Geohazards, Ministry of Education, China University of Geosciences, Wuhan, China
c
Center for Global Tectonics, China University of Geosciences, Wuhan, China
d
University of Leicester, UK
e
Faculty of Earth Sciences, China University of Geosciences, Wuhan, China
f
Institute of Geophysics and Geomatics, China University of Geosciences, Wuhan, China
abstractarticle info
Article history:
Received 7 March 2014
Received in revised form 19 May 2014
Accepted 24 May 2014
Available online xxxx
Keywords:
Craton destruction
Hydroweakening
Slab rollback
North China craton
Wyoming cra ton
Brazil shield
The mechanisms of growth and destruction of continental lithosphere have been long debated. We define
and test a unifying plate tectonic driving mechanism that explains the numerous petrolog ical, geophysical,
and geological features that characterize the destruction of cratonic lithospheric roots. Data from three
Archean cratons demonstrate that loss of their roots is related to rollback of subducted flat slabs, some
along the mantle transition zone, beneath the cratons. During flat slab subduction dehydration reactions
add water to the overlying mantle wedge. As the subducting slabs roll back, they suck in mantle material
to infill the void space created by the slab roll back, and this fertile mantle becomes hydrated. The roll-back
causes concomitant lithospheric thinning of the overlying craton so the flux of newly hydrated mantle material
inevitably rises causing adiabatic melting, generating new magmas that gradually destroy the roots of the
overlying craton through melt–peridotite reactions. Calculated fluxes of new mantle material beneath cratons
that have lost their roots range from 2.7 trillion to 70 million cubic kilometers, which is sufficient to generate
enough melt to completely replace the affected parts of the destroyed cratons. Cratonic lithosphere may be
destroyed in massive quantities through this mechanism, warranting a re-evaluation of continental growth
rates with time.
© 2014 Elsevier B.V. All rights reserved.
1. Introduction
The rate of continental lithosphere growth and destruction through
time is a long-standing controversial issue in geosciences. Nd and Hf
isotopic data from Archean and Hadean zircons suggest that a large
volume of continental crust, about 50–60% of the current volume, was
extracted by 2.5 Ga (Condie, 2005; Griffin et al., 2013; Rollinson,
2010; Tolstikhin and Kramers, 2008), yet the present volume of
Archean crust is estimated to be less than 30% of the earlier extracted
volume (Condie et al., 2009). Therefore, either large volumes of crust
remain undetected deep in the lithosphere (Griffin et al., 2013)or
continental lithosphere recycling and destruction has been a much
more widespread mechanism than currently appreciated in the
geosciences. Growing data suggests that much of the sub-continental
lithospheric mantle (SCLM) may be Archean in age, even in places
where the overlying crust is much younger (Griffinetal.,2013). There
is also much evidence that appreciable volumes of continental crust
and mantle have been removed by various processes through Earth
history. For example, subduction erosion at trenches has been widely
documented for decades (e.g., Stern, 2011; von Huene and Scholl,
1991). In Japan a whole Paleozoic arc batholith has been removed
(Isozaki et al., 2010), the lower crust of some island arcs (e.g. Talkeetna)
has foundered into the mantle (Jagoutz and Behn, 2013,alsoLallemand,
1995), the lower crust of the Eastern Pontides of Turkey was
delaminated in the late Paleozoic as a result of continent–continent
collision (Dokuz, 2011), the lower lithosphere of the entire Sierra
Nevada mountain range in California was removed during deformation
of the Cordilleran continental margin in the Pliocene (Jones et al., 2004),
and plumes, some related to rifts, have locally eroded the sub-
continental lithospheric mantle (e.g., Tanzania, Yellowstone,
Greenland; Foley, 2008). Other mechanisms of continental lithosphere
destruction are being increasingly understood. Maruyama et al.
(2007b) proposed that large quantities of tonalite–trondhjemite–
granodiorite (TTG) largely of early Precambrian age were subducted
to the mantle transition zone, and from there sank to the core–mantle
boundary, where they make-up a slab graveyard in the form of a new
Tectonophysics xxx (2014) xxx–xxx
⁎Corresponding author at: State Key Lab for Geological Processes and Mineral
Resources, China University of Geosciences, Wuhan, China. Tel.: +86 189 7157 9211.
E-mail addresses: tkusky@gmail.com (T.M. Kusky), brian.windley@btinternet.com
(B.F. Windley), wanglu2005@gmail.com (L. Wang), 979353335@qq.com (Z. Wang),
398021789@qq.com (X. Li), zhupm@126.cn (P. Zhu).
TECTO-126330; No of Pages 14
http://dx.doi.org/10.1016/j.tecto.2014.05.028
0040-1951/© 2014 Elsevier B.V. All rights reserved.
Contents lists available at ScienceDirect
Tectonophysics
journal homepage: www.elsevier.com/locate/tecto
Please cite this article as: Kusky, T.M., et al., Flat slab subduction, trench suction, and craton destruction: Comparison of the North China,
Wyoming, and Brazilian cratons, Tectonophysics (2014), http://dx.doi.org/10.1016/j.tecto.2014.05.028

continent with a high-V anomaly of calc-alkaline material. In this work
we focus on large scale destruction of sub-continental lithospheric
mantle beneath Archean cratons.
The North China craton is the world's best example of a craton that
had a thick root in the Precambrian and Paleozoic, and experienced
large-scale root loss in the Mesozoic, with models for the loss ranging
from large-scale delamination or density foundering, to major thermal
erosion mechanisms including melt–peridotite reaction (e.g., Foley,
2008; Gao et al., 2004, 2009). Other cratons, such as the Sahara
metacraton (Liegeois et al., 2013), Wyoming, the Dharwar craton in
India (Griffinetal.,2009), and the Brazilian Shield have lost large
parts of their lithospheric roots, but there is no consensus on how
such large scale recycling of the SCLM was initiated and carried out
(e.g., Zhu et al., 2012). Here we present a new comprehensive model
of large-scale sub-continental lithospheric mantle destruction, using
three examples, and discuss how this mechanism is important for un-
derstanding mechanisms of continental destruction and constraining
models of continental growth through time.
We document and model the relationships between flat slab sub-
duction, trench suction, and craton destruction, using examples from
the North China and Wyoming cratons, each of which locally lost ap-
proximately 100 km of their lithospheric roots in the Cretaceous and
which show spatio-temporal relationships with episodes of flat slab
subduction in the mantle transition zone associated with deep mantle
hydration, coupled with slab rollback and concomitant influx of mantle
fertile material to accommodate the space created by the slab rollback.A
similar process has more recently operated along the western side of
the Brazilian craton where it is thrust beneath the thickened crust of
the Andes in an area of trench rollback. The mutual interaction between
these processes may be more generally applicable than currently per-
ceived. Together with the other processes of subduction erosion and
arc subduction, larger amounts of continental lithosphere may have
been subducted or otherwise returned to the sub-lithospheric mantle
than previously appreciated.
2. Comprehensive model for large-scale sub-continental lithospheric
mantle removal
In this paper we propose a new model of how cratons and their
mantle keels can be destroyed and recycled back into the convecting
mantle. Conventional wisdom is that the prolonged stability of cratonic
lithosphere, specifically the development of a thick, insulating
lithospheric mantle keel, restricts or even prohibits its recycling into
the Earth's mantle (e.g., Durrheim and Mooney, 1994; Jordan, 1988;
Kaban et al., 2003; Trubitsyn et al., 2003). However, if cratonic litho-
sphere can be recycled, it has important implications for crust–mantle
recycling and for understanding crustal growth through time (e.g.,
Artemieva et al., 2002; Bowring and Housh, 1995; Condie et al., 2009;
Foley, 2008; Rino et al., 2004; Taylor and McLennan, 1995; Zhu et al.,
2011, 2012).
Cratons are structurally complex regions that attained prolonged
stability (≥1 Ga) within the continents and so by definition cratons
are Precambrian in age. Indeed, most formed in the Archean
(Goodwin, 1991; Kusky and Polat, 1999; Rudnick, 1995; Windley,
1995). In general, Archean cratons are characterized by cold, thick,
structurally complex lithosphere whose density is offset by its refracto-
ry composition, giving rise to chemical buoyancy (Jordan, 1975, 1981).
The thickness of these cratonic keels (also termed roots, tectosphere,
or sub-continental lithospheric mantle, SCLM) is generally 200–300 km
(Prodehl and Mooney, 2012). The presence of these thick, refractory
and anhydrous peridotite residues beneath Archean crustal regions is
widely held responsible for the inherent stability of Archean cratons
(Durrheim and Mooney, 1994; Griffin et al., 2003a; Kaban et al., 2003;
Pollack, 1986). However, the formation mechanism of these keels has
been controversial (e.g., Griffin et al., 2013)withsomemodelssuggesting
that they represent stacked subducted oceanic slabs (e.g., Helmstaedt and
Gurney, 1995; Kusky, 1993; Stachel et al., 1998) and other models
suggesting that they represent the residue from high degrees of
partial melting of plumes or ambient upper mantle (Herzberg and
Rudnick, 2012).
While traditional models for theevolution of continental lithosphere
suggest that once continents become stable or cratonized, they are inde-
structible and last forever (e.g. Jordan, 1975, 1981, 1988), it is becoming
increasingly clear that in some cases large portions of the sub-
continental lithospheric mantle can be destroyed and returned to the
deep mantle long after their formation When this happens, a craton
loses its cratonic characteristics, and returns to a more orogenic style
of behavior, in a process that has been named the “orogen–craton–
orogen”cycle (Kusky et al., 2007a). However, the processes that lead
to the loss of lithospheric roots have been poorly constrained, and
there are currently a wide variety of ideas about how lithospheric
roots can be lost. Since the SCLM is difficult to sample directly, most
models are based either on geophysical data (Cook et al., 1998), or on
information from xenoliths brought up by kimberlites (e.g., Foley,
2008; Griffin et al., 2003b; Zhu et al., 2012). Although most models
have assumed mechanical detachment or foundering of a lithospheric
root (e.g. O'Reilly et al., 2001), some evidence suggests chemical re-
placement or metasomatic modification of lithospheric roots by upwell-
ing asthenosphere by thermo-chemical processes (e.g. Griffin et al.,
2003b; Xu et al., 2004; Zheng et al., 2005).
We here propose a general plate tectonic and mantle dynamics
model for large-scale craton destruction, based on a comparison of
three cratons that have lost their roots. The model has two parts. The
first is weakening of the SCLM by hydration from long-term subduction
dehydration reactions beneath the cratons. This is followed by
subduction roll-back during episodes of flat slab subduction along the
mantle transition zone, which drives mantle flow into the region
beneath the SCLM by trench suction, because the space created by slab
roll back must be filled by mantle material. This new hotasthenosphere
is re-hydrated by the deep dehydration reactions from the slab in the
transition zone (Windley et al., 2010), then it gradually weakens and
replaces the ancient SCLM by thermal erosion and by melt–peridotite
reactions, replenishing the ancient lithosphere with more fertile
material.
2.1. Step 1. Hydroweakening
When oceanic lithosphere subducts, it hydrates the upper mantle
beneath an arc from well-known dehydration reactions (e.g., Karato,
2003; Kawamoto, 2006; Peacock, 2003). However, some hydrous
phases (e.g., Phase A, Phase E, and γ- and β-phase olivine) are stable
to much greater depths and dehydrate even when a slab is in the mantle
transition zone (e.g., Maruyama and Okamoto, 2007; Niu, 2005;
Peacock, 1993; Windley et al., 2010). It is estimated that 40% of the
water subducted in hydrated oceanic crust, mantle, sediments, and
subducted continental material reaches the mantle transition zone be-
tween 410 and 660 km (Maruyama and Okamoto, 2007; Tonegawa
et al., 2008). For instance lawsonite may contain up to 11% water, and
is stable up to 11 GPa (Williams and Hemley, 2001) or about 300 km
(Fig. 2a) and serpentinites can contain up to 13% water and are stable
up to 7 GPa (Ulmer and Trommsdorff, 1995) and after conversion to
denser hydrous phases such as β-phase olivine they can be stable up
to 50 GPa (Frost, 1999; Schmidt, 1995; Williams and Hemley, 2001),
well past the mantle transition zone (Fig. 2b). With increasing temper-
ature (i.e., more time in the transition zone for deep flat slabs) these
phases decompose to less hydrous wadsleyite and ringwoodite with
2.2–3.3 wt.% water, releasing water to the deep mantle, which rises
and hydrates the overlying mantle (Karato, 2003; Richard et al., 2006;
Smyth et al., 2003), as indicated by electrical conductivity and seismic
P-wave velocity data (Ichiki et al., 2006). The water solubility of nomi-
nally anhydrous minerals in the mantle commonly ranges up to tens
of thousands of parts per million H
2
O by weight, constituting a major
2T.M. Kusky et al. / Tectonophysics xxx (2014) xxx–xxx
Please cite this article as: Kusky, T.M., et al., Flat slab subduction, trench suction, and craton destruction: Comparison of the North China,
Wyoming, and Brazilian cratons, Tectonophysics (2014), http://dx.doi.org/10.1016/j.tecto.2014.05.028

reservoir of water that has an important influence on mineral and bulk
mantle properties such as melt relations, rheology and electrical
conductivity (Bromiley et al., 2010). If the transition zone is saturated,
the amount of water locked in these minerals, is potentially four times
larger than all the water in the planet's hydrosphere (Smyth and Frost,
2002). Much of the water released from these phases is concentrated
in the mantle transition zone between 410 and 660 km, which can
lower the melting temperature, leading to the formation of magmas
(Bercovici and Karato, 2003; Maruyama and Okamoto, 2007). These
hydrous phases rise as hydrous magmas that weaken the overlying
SCLM, setting the stage for large-scale root loss (Kusky et al., 2007a;
Niu, 2005; Windley et al., 2010).
In (a) the P–T trajectory is assumed based on a moderately old slab,
and other paths are possible (see Maruyama and Okamoto, 2007).
b) We assume an oceanic slab of 100 km thickness is subducted, and
the surface geotherm follows the curve in (a) with surface heating
from the overlying hot mantle wedge. The oceanic geotherm at 0 Ma
is from Faul and Jackson (2005). We neglect any changes in the slab
thickness from heating, and ignore heating from below for this simple
analysis. The red lines show the change in the geotherm through the
subducting slab with time, assuming a 45° dip angle and a subduction
rate of 3 cm/yr (as we use below). At 7 Ma the slab top is 7 Ma ∗
3cm/y∗sin (45°) km (148.49 km) deep, so the slab bottom is
248.49 km deep. Therefore, the isochronic geotherm at 7 Ma can be
approximated as shown. Similarly, the isochronic geotherms at 14 Ma
and 18 Ma can be drawn roughly. When the slab reaches the transition
zone (we assume at 440 km for the top of the slab)it flattens out, so the
pressure nolonger increases, butthe temperature continues to rise with
time by conductive heat transfer, which releases water from hydrous
phases to the transition zone and overlying mantle by the time-
dependent dehydration reactions. Note that the crustal (MORB + H
2
O)
section becomes anhydrous at about 300 km, but the upper section
of the mantle (peridotite + H
2
0) remains hydrous all the way to
the transition zone.
Fig. 1. (a)3D view of the North Chinacraton, and its position over thePacificflat slab situated in themantle transitionzone. The slab is dehydratingand generating melts in the mantlethat
rise to alter the base of the SCLM (inspired by Zhu et al., 2012). (b) Two perpendicular vertical cross-sections of whole-mantle P-wave tomography beneath eastern Asia showing slow
velocities in red, and fast in blue. The fast velocities outline the flat-lying slabs beneath eastern Asia (modified after Zhao and Ohtani, 2009). (For interpretation of the references to
color in this figure legend, the reader is referred to the web version of this article.)
3T.M. Kusky et al. / Tectonophysics xxx (2014) xxx–xxx
Please cite this article as: Kusky, T.M., et al., Flat slab subduction, trench suction, and craton destruction: Comparison of the North China,
Wyoming, and Brazilian cratons, Tectonophysics (2014), http://dx.doi.org/10.1016/j.tecto.2014.05.028

2.2. Step 2. Slab flattening and rollback causes influx of fertile mantle
beneath SCLM
It is well established that the portion of the NCC that experienced
root destruction is underlain by a horizontal stagnant slab in the mantle
transition zone (Fig. 1;Huang and Zhao, 2006; Zhao et al., 2007), and
several authors have speculated on the causal relationship between
this coincidence (e.g., Zhu et al., 2011, 2012). We here investigate the
physical and geometrical relationships between placing a slab in
the transition zone, slab rollback, lithospheric extension, mantle
replacement, and root loss.
Slab rollback associated with flat slab subduction in the transition
zone is different from slab rollback associated with a normal steeply-
dipping and deeply-penetrating slab at a trench. When a slab becomes
horizontal, that section of the slab no longer exerts slab pull forces on
the overlying slab, and encounters more and more resistance to further
penetration as the flat section grows longer, which will eventually result
in it becoming incapable of further penetration (Fig. 3). When this hap-
pens, the rollback rate must become equal to the subduction velocity,
unless the geometry of the slab changes. There are three end-member
cases to consider (Fig. 3).
To understand this geometrically and kinematically we first must
define some parameters. Fig. 3a(modified and expanded from a
steep-slab subduction system modeled by Becker and Faccenna, 2009)
shows a simple flat-slab subduction system with an overriding plate
and fore-arc sliver. The convergence velocity at the trench (V
C
)is
given by the velocity of the oceanic plate (V
P
, positive towards trench)
plus the velocity of the overriding plate (V
OP
, positive towards trench),
whereas the velocity of the trench (V
T
, positive for rollback) is given by
V
OP
+ the velocity of back arc deformation (V
B
, positive for extension) if
there is no subduction erosion or accretion at the trench (Lallemand,
1995). The velocity of subduction (V
S
) is given by V
P
+V
OP
+V
B
(Fig. 2). Becker and Faccenna (2009) have shown that trench migration
rates (V
T
) globally are typically less than 50% of the convergence rates
(V
C
). This has important implications as discussed below.
Trenches may advance or rollback relative to the overriding plate,
and in this work we only consider the cases of slab rollback. Fig. 3b,c,d
shows possible geometric consequences of what happens when a slab
has a very long flat segment ponded in the transition zone and the
trench is retreating or rolling back. Weak slabs have a more difficult
time than strong slabs to penetrate a viscosity or density contrast at
depth (e.g., Christensen, 1996; Davies, 1995; Kincaid and Olson, 1987)
and thus tend to “pond”or accumulate along the transition zone. For
these ponded slabs, there will come some length where the force
needed to make it continue to penetrate further along the transition
zone exceeds the forces pushing it along the boundary (e.g., van
Hunen et al., 2000), and its forward motion will stop (V
PEN
then = 0),
whence it can be considered anchored. For slabs that lie flat along the
transition zone, the force of slab pull contributing to V
P
will only consist
of gravitational sinking of the part of the slab above the 600 km
discontinuity, which will cause the bend in the lower part of the slab
to have a complimentary “anti-rollback”or flattening (V
FLAT
;Fig. 3)at
the transition from the steep to flat-slab segments (Fig. 3).
If the resistive forces in the transition zone cause V
PEN
to become
zero and the slab is anchored, then the slab will be forced to retreat
and V
FLAT
(anti-rollback) should equal V
T
(rollback), and the trench
Fig. 2. Phase diagrams for MORB + H
2
O (a) and peridotite + H
2
O systems (b) simulating a subducted slab reaching pressures equivalent to the mantle transition zone, showing how
hydrous phases (e.g. lawsonite) are stable in the MORB section until approximately 300 km, and stable in the peridotite section into the transition zone (modified with re-calculations
after Maruyama and Okamoto, 2007). In (a) the P–T trajectory is assumedbased on a moderatelyold slab, and other pathsare possible (see Maruyama and Okamoto,2007). b) We assume
that an oceanicslab of 100 km thicknessis subducted, and the surfacegeotherm follows the curve in (a) with surfaceheating from the overlying hot mantlewedge. The oceanic geotherm
at 0 Ma is from Faul and Jackson(2005). We neglect any changes in the slabthickness fromheating, and ignoreheating from belowfor this simpleanalysis. The redlines show the change in
the geotherm through the subducting slab with time, assuming a 45 degree dip angle and a subduction rate of 3 cm/yr (as we use below). At 7 Ma the slab top is 7Ma∗3cm/yr∗sin
(45°) km (148.49 km) deep, so the slab bottom is 248.49 km deep. Therefore, the isochronic geotherm at 7 Ma can be approximated as shown. Similarly, the isochronic geotherms at
14 Ma and 18 Ma can be drawn roughly. When the slab reaches the transition zone (we assume at 440 km for the top of the slab) it flattens out, so the pressure no longer increases,
but the temperature continues to risewith time by conductive heat transfer, which releases waterfrom hydrous phasesto the transitionzone and overlyingmantle by the time-dependent
dehydration reactions. Note that the crustal (MORB + H
2
O) sectionbecomes anhydrous atabout 300 km, but the upper section of themantle (peridotite+ H
2
0) remains hydrous all the
way to the transition zone.(For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
4T.M. Kusky et al. / Tectonophysics xxx (2014) xxx–xxx
Please cite this article as: Kusky, T.M., et al., Flat slab subduction, trench suction, and craton destruction: Comparison of the North China,
Wyoming, and Brazilian cratons, Tectonophysics (2014), http://dx.doi.org/10.1016/j.tecto.2014.05.028

will rapidly retreat away from the overriding plate. In this case, if V
OP
is
zero, then back-arc extension (V
B
) will also equal the rollback velocity
(Fig. 3b). However, Becker and Faccenna (2009) have shown that V
T
never exceeds V
P
, and is typically less than 50% of V
P
. This implies that
if V
PEN
has gone to zero, that V
FLAT
should be greater than Vt, and the
slab will become steeper (δwill decrease) during flat slab anchoring
(Fig. 2c). This can only continue until the slab is vertical, since there
are no known slabs that are overturned above the transition zone. If
V
T
were greater than V
FLAT
, then the slab would become flatter above
thetransitionzone(Fig. 3d).
It is likely that flat-lying slabs do not actually penetrate and move
along the transition zone, but simply hit the density and rheological
contrast there, and cannot penetrate if they are weak. This is likely
because there is no slab pull force driving the slab forward, and the
ridge push and basal traction forces on the slab at this point are much
smaller (approximately 30% of slab pull) according to Lithgow-
Bertelloni and Richards (1998). This would induce slab rollback
(increase V
T
) and a simple flattening of the slab with concomitant
extension of the overriding plate (V
B
). A simple geometric analysis
shows that if a 45°-dipping slab flattens out along the transition zone
at 600 km, and stops without moving horizontally (i.e., V
PEN
is zero)
then if V
T
is zero, the slab can flatten out leaving a 600 km long flat
slab alongthe transition zonebefore the overlying slab becomes vertical
(note: we use 600 km for these simple calculations since the base of the
transition zone is about 660 km, the slab is 50–100 km thick, and the
base of the slab can flatten out anywhere between 410 and 660 km).
Since V
T
rarely exceeds 50% of V
P
, then another 300 km of the
subducting slab can be flattened out without penetration along the
transition zone before the slab is vertical. This length could be increased,
if limited true slab penetration along the transition zone is allowed,
perhaps at a rate of up to 30% of V
P
, assuming that only the forces
not driven by slab pull (gravity) are pushing the slab. This could allow
another 180 km of slab penetration for a slab hitting the 600 km discon-
tinuity, resulting in a flat slab section 1100 km long. Interestingly, the
flat part of the Pacific slab underlying the area of cratonic root loss in
eastern Asia is approximately 1500 km long (Huang and Zhao, 2006).
The crux of why this is related to craton destruction is that geomet-
rically, if the slab has moved away from the original site of the trench by
600–900 km, it has carried the mantle above the slab with it (since voids
cannot be created in the mantle), causing new mantle material to flow
sideways below the overriding plate to replace the moving mantle;
we call this flow the mantle flux (flux
m
in Figs. 3, 4). The mantle flux
can be huge —if we take a typical example of a 1000 km-wide
(measured parallel to the trench) subduction zone, that has rolled
back by 900 km with a flat slab section in the transition zone at
600 km, then 540 million cubic kilometers of mantle need to flow into
the space between the overriding plate and transition zone to fill the
space created by the retreating slab (Fig. 4). This new mantle is then
hydrated by the dehydration reactions from the slowly heating and
dehydrating wadslyeite and ringwoodite in the slab, inducing partial
melting (Fig. 4). The extension of the overriding plate (V
B
)inducedby
the slab rollback has caused lithospheric thinning, so the new mantle
material must also rise in addition to flowing sideways, causing
additional adiabatic melting (e.g., de Smet et al., 1999; Vlaar, 1983).
Together with this mantle hydration, the movement of new material
above the flat slab induced by the slab suction of the rolling back slab,
and adiabatic decompression forms enough melts that can cause a
melt–peridotite reaction to thermochemically erode the base of a
craton, resulting in large-scale craton destruction (e.g., Foley, 2008;
Gao et al., 2004, 2009).
3. North China Craton
The North China craton is the world's best example of a craton that
had a thick root in the Archean, and lostthe eastern half of the root dur-
ing Mesozoic tectonomagmatic events. There are numerous reviews on
the current geometry of the root beneath the craton (S.L. Li et al., 2011;
Tian and Zhao, 2013; Wang et al., 2013; Xu et al., 2011; Y.Y. Li et al.,
2011), and the evidence for its prior existence and loss (e.g., Griffin
et al., 1998; Kusky et al., 2007a, 2007b; Menzies et al., 1993; Wilde
et al., 2003; Wu et al., 2003; Yang, 2003; Zhu et al., 2011, 2012); we
only briefly summarize this evidence here.
Fig. 3. (a) Kinematic framework of flat slab subduction associated with slabrollback. VC = convergence rateat trench, V
OP
= velocity of overriding plate (+towardstrench), V
B
=back-
arc deformation rate (+for extension), V
T
= velocityof trench (+for rollback), V
P
= velocityof oceanic plate (+to trench), V
S
= sinking velocity of slab, δ= slabdip angle, R = bending
radius at trench,V
PEN
= velocityof slab penetration,V
FLAT
= velocityof slab flattening (+for rollback,as with V
T
), flux
m
= amount of new mantlematerial neededto fill space created by
slab rollback. Some vectors after Becker and Faccenna (2009).(b–c–d) show how the geometry of the slab changes with variations between V
T
and V
FLAT
.
5T.M. Kusky et al. / Tectonophysics xxx (2014) xxx–xxx
Please cite this article as: Kusky, T.M., et al., Flat slab subduction, trench suction, and craton destruction: Comparison of the North China,
Wyoming, and Brazilian cratons, Tectonophysics (2014), http://dx.doi.org/10.1016/j.tecto.2014.05.028

The North China craton is divisible into the Eastern and Western
blocks (Fig. 1), separated by the Central Orogenic belt that represents
an Archean–Paleoproterozoic collisional orogen (Kusky, 2011; Zhao
and Zhai, 2013). The Central Orogenic belt also roughly corresponds to
a topographic and gravity gradient that separates the Eastern block
with thin lithosphere from the Western block with thick lithosphere,
although the NS-trending gravity lineament extends far beyond the
borders of the NCC lying parallel to the Pacific margin and outboard
subduction zone (e.g., Niu, 2005). The Western block (also known as
the Ordos platform) of the NCC is a stable craton with a thick mantle
root (up to 215 km deep), few earthquakes, low heat flow, and has
experienced little deformation since the Precambrian (Yuan, 1996;
Zhai and Liu, 2003; Zhu et al., 2012). The Eastern block is different —it
has high heat flow, numerous sometimes large earthquakes, active vol-
canoes, and has undergone significant deformation especially since the
Mesozoic (J.L. Liu et al., 2011; Zhang et al., 2011). Geophysical data (Zhu
et al., 2012) show that the eastern half of the NCC now has a thin litho-
sphere (as thin as 60–65 km) and no preserved lithospheric root (Yuan,
1996; Zhu et al., 2012). However,data from xenoliths in kimberlites and
lavas erupted in the Paleozoic and Mesozoic show that the Eastern block
once had a thick root developed in the Archean, but lost it sometime in
the Mesozoic (Fan and Menzies, 1992; Gao et al., 1998, 2002, 2004,
2009; Griffin et al., 1998; Menzies et al., 1993; Wu et al., 2003, 2008;
Zheng and Wu, 2009; Zhu et al., 2011, 2012). The best constraint on
the age of the root loss is 132–117 Ma (Zhang et al., 2014)withapeak
at about 125 Ma (e.g., Zhu et al., 2011). The loss of the eastern part of
the lithospheric root of the NCC took place during two stages.
3.1. Stage 1. Hydroweakening associated with long-term sub-craton
subduction
The North China craton has experienced a long and complex
evolution, including an arc/continent and maybe continent/continent
collisions in the Archean, leading to the formation of a thick Archean
mantle root (see reviews by Kusky et al., 2007a,b; Kusky, 2011; J.G. Liu
et al., 2011; J.L. Liu et al.,2011; Zhai and Santosh, 2011). Paradoxically,
even though the mantle root formed in the Archean, many workers
suggest that the craton did not amalgamate until about 1.8 Ga in the
Paleoproterozoic (e.g., Zhao and Zhai, 2013; Zhao et al., 2001).
Alternatively, Kusky and Li (2003),Kusky et al. (2007a),andKusky
and Santosh (2009) suggested that the 2.3 to 1.9–1.8 Ga tectonic events
in the NCC were focused along the northern margin of the craton during
its life as an old passive margin converted to an Andean margin after an
arc/continent collision, then a major continent–continent collision
during amalgamation with the Columbia (Nuna) supercontinent at
1.9–1.8 Ga. This model also explains how the Archean mantle root
along the north margin of the craton was replaced by fertile
Paleoproterozoic mantle at 1.85 Ga (J.G. Liu et al., 2011). In any case
the NCC experienced a long period of subduction beneath the craton
in the Proterozoic, potentially adding water and hydroweakening
the mantle above the subducting slabs, pre–setting the stage for
the generation of melts to interact with and destroy the overlying SCLM.
After 1.8 Ga, the NCC underwent a period of relative calm until
~700 Ma with subduction under the craton until ~250 Ma (Maruyama
et al., 2007a, 2007b) from Dabie Shan in the south from 500 to
Fig. 4. New comprehensi ve model for craton d estruction throug h flat slab dehydration, slab rollback, mantle influx, melt-generation, and melt SCLM peridotite reaction. See text
for discussion.
6T.M. Kusky et al. / Tectonophysics xxx (2014) xxx–xxx
Please cite this article as: Kusky, T.M., et al., Flat slab subduction, trench suction, and craton destruction: Comparison of the North China,
Wyoming, and Brazilian cratons, Tectonophysics (2014), http://dx.doi.org/10.1016/j.tecto.2014.05.028

250 Ma, from subduction under the craton from the Solonker suture in
the north, and from 200 Ma to present with subduction from the Pacific
and paleo-Pacific margin (see Fig. 5 and reviews by Kusky et al., 2007a,
b; Windley et al., 2010). It is estimated that more than 18,000 km ofoce-
anic lithosphere were subducted beneath the NCC, perhaps more than
any other craton on Earth (Kusky et al., 2007a,; Windley et al., 2010).
This prolonged subduction is suggested to have led to pronounced
hydroweakening, as described in the model above.
3.2. Stage 2. Flat slab subduction, roll back, and influx of new mantle material
Plate reconstructions of the paleo-Pacific realm or Panthalassic
Ocean (e.g., Engebretson et al., 1985; Lithgow-Bertelloni and Richards,
1998; Norton, 2007; Seton et al., 2012; Smith, 2003; Utsonomiya et al.,
2007; van der Meer et al., 2012; Yang, 2013) show that the continental
margin of East Asia has experienced several episodes of subduction of
oceanic lithosphere since the early Paleozoic (Maruyama et al., 2007a,
2007b), as well as several ridge subduction events (e.g., Isozaki et al.,
2010). Van der Meer et al. (2012) presented a new reconstruction for
Triassic–Jurassic times and suggested that the Panthalassic Ocean was
divided into distinct paleo-oceanic plate systems, the Pontus in the
west and the Thalassa in the east, separated by the Telkhinia system of
subduction zones. By the Late Triassic–Early Jurassic the Thalassa
Ocean (Fig. 6) was divided into the Izanagi, Farallon, and Phoenix plates
whose motion carried several exotic terranes across the ocean to collide
with Asia and North America (van der Meer et al., 2012; Yang, 2013).
The Pontus Ocean gradually closed by subduction in the Telkhinia sub-
duction zones as the Izanagi, Farallon, and Phoenix plates grew. By the
Early Cretaceous (150–140 Ma) the Izanagi plate was subducting be-
neath the entire East Asian margin, as the Pacific plate began to grow
from the triple junction between the three oceanic plates. The Izanagi
plate continued to subduct beneath East Asia until about 90 Ma, when
the Izanagi–Pacific triple junction migrated northwards along the mar-
gin that was subductingthis ridge,and this led to the Pacific plate being
subducted beneath East Asia since 90 Ma. Thus, the slabs visible in the
mantle beneath Asia (Fig. 1b; Huang and Zhao, 2006; Zhao et al.,
2007) should contain a record of 60 Ma of subduction of the Izanagi
plate from 150 to 90 Ma, and 90 Ma of subduction of the Pacific plate
to the present day. If we assume rates of trench-perpendicular subduc-
tion of 3 cm/yr, (a reasonable lower-end estimate, see Seton et al.,
2012) there should be beneath Asia a 1800 km-long slab of the Izanagi
plate, and a 2700 km-long slab of the Pacificplate.
Fig. 5. Geometry and location of the North China craton above subducting slabs from the Early Paleozoic to Tertiary. From Windley et al., 2010, with permissions.
7T.M. Kusky et al. / Tectonophysics xxx (2014) xxx–xxx
Please cite this article as: Kusky, T.M., et al., Flat slab subduction, trench suction, and craton destruction: Comparison of the North China,
Wyoming, and Brazilian cratons, Tectonophysics (2014), http://dx.doi.org/10.1016/j.tecto.2014.05.028

Tomographic data (Fig. 1b; Huang and Zhao, 2006;Zhao et al., 2007)
show a high-velocity anomaly interpreted to be the Pacific (and per-
haps older) slab dipping beneath Asia, and flattening out into and
along the mantle transition zone, and extending almost to the NS grav-
ity lineament and westernextent of root loss of the NCC (Fig. 1). The flat
part of the slab is abnormally thick (it could be thicker representing two
stacked slabs, or it could bean artifact of the data) and is about 1500 km
long, whereas the dipping part is about 1200 km long (measured paral-
lel to dip). Thus, the 2700 km of preserved slabs beneath Asia accord
well with the idea that they represent both the Izanagi and Pacific
slabs, especially if the flat section has a doubled thickness or is stacked
(Fig. 1b). This concept of the mantle homeof this long-lived subduction
slab system was well portrayed by Maruyama et al. (Fig. 5 in. 2007a) as
the Gondwana slab graveyard under the present Pacific Ocean.
If the Izanagi plate started subducting at 150 Ma at a rate of 3 cm/yr
along the same trajectory (dip, or δin Fig. 2) with a 1200 km-long
dipping segment, then it would take 40 Ma for the slab to reach the tran-
sition zoneat 110 Ma. If the slab dip was verticalit would reach the tran-
sition zone with a length of 600 km at 20 Ma after subduction initiation
at 130 Ma, shortly before the peak time of root loss for the NCC at
125 Ma. These times would be less if the slabs flattened out at 400 km
instead of 600 km, or if the subduction rates were faster. We assume
that the Izanagi plate was a weak slab and unable to penetrate the tran-
sition zone(as shown by the fact that it is currently flat lying within the
transition zone), and that it flattened out at this time. This would hap-
pen if the penetration velocity (V
PEN
) of the Izanagi plate was zero,
and further subduction (V
S
) was accommodated by V
FLAT
or flattening
of the slab at depth, with concomitant rollback of the slab on the surface.
This dynamic situation set up the lateral flow of mantle material to
beneath the craton (Fig. 4), to fill in the space above the slab created
by the slab rollback as described above, bringing in new fertile
mantle, that was hydrated by the water released from dehydrating
minerals in the flat slab, and as this new hydrated fertile mantle
rose it partially melted, reacted with the lithospheric root, and
thermochemically eroded the root (e.g., Foley, 2008; Gao et al.,
2004, 2009; Xu et al., 2004; Zheng et al., 2005). We calculate the
mantle flux (flux
m
) associated with this case by approximating a
1500 km-long flattened slab segment at 600 km, for a 3000 km slab
width (measured parallel to the trench), and for a mantle flux of
2.7 trillion cubic km of new mantle that moved in beneath East
Asia due to trench rollback and slab flattening. Schellartetal.
(2008) made global estimates of the mantle flux (flux
m
in our
terminology) of 456–539 km
3
per year, which if extended for the
time period equivalent for the approximately 30 Ma of slab rollback
beneath Asia (and assuming 500 km
3
/yr), would yield 15 trillion
cubic km of slab-rollback induced mantle flux globally, meaning
that the rollback of the Pacific slab beneath Asia accounted for about
20% of the total global rollback-related mantle flux in that period.
This model has another fundamental difference from other percep-
tions of how flat slab subduction operates. Many researchers assume
that the subduction zone and slab have stayed essentially in place,
that the slab subducted to the tra nsition zone, then became flat and con-
tinued to penetrate horizontally by moving along the transition zone. In
our model, we use the opposite end-member, and assume that the slab
subducted to the transition zone, then flattened out because it was too
weak to penetrate the rheological/phase boundary, then it progressively
rolled back along the transition zone and the surface (Fig. 3b–d). Thus,
the trench at 150–120 Ma ago was located much closer to the Asian
mainland, and moved away as the slab flattened and rolled back, open-
ing a “gap”1500 km wide and 400–600 km deep that had to be filled by
new mantle material moving in laterally from the west (flux
m
in Fig. 4).
This process has the potential to generate enormous amounts of melt,
explaining not only the loss of the root of the NCC, but also much of
the magmatism in eastern China (e.g., Niu, 2005).
Fig. 6. Plate reconstructions of the paleo-Pacific (Panthalasia) Ocean and adjacent continents from the Late Triassic to Late Cretaceous.
Modified from Yang, 2013.
8T.M. Kusky et al. / Tectonophysics xxx (2014) xxx–xxx
Please cite this article as: Kusky, T.M., et al., Flat slab subduction, trench suction, and craton destruction: Comparison of the North China,
Wyoming, and Brazilian cratons, Tectonophysics (2014), http://dx.doi.org/10.1016/j.tecto.2014.05.028

An additional aspect of slab rollback is that as the subduction
zone retreats from the continental margin, the overriding plate will
experience extension if the velocity of the overriding plate (V
OP
)is
not greater than the velocity of rollback (V
T
). Upper plate extension
is described by V
B
, and it is intriguing to note that during this interval
of slab flattening and trench rollback, the upper crust of the NCC
underwent massive extension and the formation of metamorphic core
complexes (e.g., Darby et al., 2004; J.G. Liu et al., 2011; J.L. Liu et al.,
Fig. 7. (a) Map of the Wyoming craton and surrounding orogens. Part of the root of the western part of the craton has been lost. Map redrawn after Foster et al., 2006. (b) Geodynamic
model showing the position of the Farallon slab beneath North America at 70 Ma and present. From Steinberger (2008), with data from Liu et al. (2008) (with permissions). Mantle
temperatures are shown along the vertical section (drawn along the 41 N latitude), clearly showing the position of the Farallon plate. The surface shows dynamic topography, with
blues representing depressions and green to yellow higher topography.(c) Map of South America showing locations of present and past flat slab subduction segments, and Precambrian
basement. Modified from Beck and Zandt, 2002,andKusky, 2010. (d) Cross section across the Altiplano showingthe roots of the Brazilian shield being eroded from mantle influx during
slab rollback and steepening (modified from Beck and Zandt, 2002). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
9T.M. Kusky et al. / Tectonophysics xxx (2014) xxx–xxx
Please cite this article as: Kusky, T.M., et al., Flat slab subduction, trench suction, and craton destruction: Comparison of the North China,
Wyoming, and Brazilian cratons, Tectonophysics (2014), http://dx.doi.org/10.1016/j.tecto.2014.05.028

2011; Lin et al., 2011). These extend across much of eastern Asia, and
correspond in time to the loss of the lithospheric root. Extension also
has the possibility of creating fault-related pathways for magmas to mi-
grate upwards through the crust, aiding the thinning processes, and
explaining much of the surface magmatism during this interval.
4. Wyoming Craton
The Wyoming craton in northwestern USA (Fig. 7a) consists
predominantly of Late Archean rocks, with an older division in the NW
(The Montana Metasedimentary Terrane) comprised of 3.2–3.5 Ga
gneisses intercalated with metasedimentary rocks that underwent a
major tectonic event at 2.8 Ga (Chamberlain et al., 2003; Foster et al.,
2006; Wooden and Mueller, 1988). The southern part of the craton is
known as the Southern Accreted Terranes with ages of 2.65–2.63 Ga,
reflecting the accretion of juvenile arc and other terranes to the craton
(Frost et al., 2006; Mueller and Frost, 2006). It is bounded by the 1.92–
1.77 Ga Trans-Hudson orogen on the east, the 1.86–1.77 Ga Great Falls
tectonic zone on the north, the ~2.4–1.6 Ga Selway and Farmington
zones (and the younger Belt Basin and Idaho Batholith) in the west, and
the 2.0–1.7 Ga Mojave and Yavapai orogens on the south (Fig. 7a).
Miocene to Recent volcanic rocks of the Snake River Plain and intrusives
related to the Yellowstone plume affected the western part of the craton.
The Wyoming craton was amalgamated into the cratonic core of
Laurentia by the collisions of arcs and continental fragments along the
Trans-Hudson orogen at 1.85–1.78 Ga (Hoffman, 1988; Whitmeyer
and Karlstrom, 2007). In contrast to the North China craton, the
Wyoming craton seems to have been dominated by subduction away
from the craton during the Proterozoic, with accretion of arcs and
other blocks along its margins (Whitmeyer and Karlstrom, 2007).
Thus, mantle hydration and early root loss of the Wyoming craton did
not start until much later, during the Laramide orogeny in the
Cretaceous when the craton was underlain by flat slabs that hydrated
the sub-continental lithospheric root (e.g., Humphreys et al., 2003).
In general, the Wyoming craton is characterized by thick and strong
lithosphere with a distinct velocity structure (Dueker et al., 2001), but
there is some indication that the Wyoming craton has been gradually
decratonized since the mid-Cretaceous at about 100 Ma. Results from
the DEEP PROBE (Snelson et al., 1998) and CD-ROM (Yuan and
Dueker, 2005, 2010) seismic profiles, and xenolith data, reveal that
the crust of the Wyoming craton is 50–60 km thick, with an unusually
thick lower mafic crust that was underplated at circa 1.8 Ga (Keller,
2008). The Wyoming craton was stable from 1.8 Ga until the Laramide
orogeny in the Cretaceous, generally attributed to flat-slab subduction
of the Kula and Farallon plates (e.g., English and Johnston, 2004). During
the Laramide, the Wyomingcraton saw the formation of widespreadba-
sins separated by basement uplifts (Fig. 7a) bounded by steep thrust
faults, and deep tomographic images show that the typical thick North
American cratonal root is not fully preserved beneath the craton
(Keller, 2008; Pavlis, 2011; Snelson et al., 1998). Currently, magmatism
from the Yellowstone plume and associated Snake River Plain lavas are
further eroding the root of the Wyoming craton.
The Wyoming craton is therefore very similar to the NCC, in that it
formed in the Archean, had additional terranes accreted to its margins,
saw a major event with new mantle underplating at 1.8 Ga (compare
with J.G. Liu et al., 2011) during incorporation into a larger continent,
was relatively stable (excepting some rifting along the margins) until
the Cretaceous when flat slab subduction (shallow in this case) started,
and since then the craton has been progressively decratonized and acts
like an orogen again, repeating the orogen–craton–orogen cycle (Kusky
et al., 2007a). The latest phase of decratonization is related to impinge-
ment of the Yellowstone plume on the base of the craton (Stachnik et al.,
2008).
The relict, flat, subducted Kula–Farallon slab is still visible
tomographically beneath western North America and under the
Wyoming craton (Fig. 7b). It is interesting that the slab is presently in
the mantle transition zone between 410 and 600 km beneath the
Wyoming craton, but farther east it steepens and drops into the lower
mantle beneath the Superior craton, which preserves its thick root.
Calculating the mantle flux (flux
m
) of the Wyoming craton is not as
straightforward as for the NCC. The Farallon slab was flat at shallow
levels beneath the Wyoming craton at 70 Ma, and has since sunk and
flattened out into the transition zone, before plunging deeper into
the mantle farther to the east (Fig. 7b). Thus we estimate that a
1600 km-long slab (Fig. 7b) with a width of 1500 km (parallel to
the paleo-trench) has flattened and sunk 400 km to the transition
zone, yielding a mantle flux of 960 million cubic km of mantle material
that had to flow sideways into the region above the sinking slab,
explaining the partial root loss of the Wyoming craton.
5. Destruction of the Brazilian Craton beneath the Andes
In a somewhat similar process to the slab rollback and lateral influx
of fertile mantle beneath the NCC and Wyoming Cratons described
above, Beck and Zandt (2002) and Ramos and Folguera (2009)
described delamination of the subcrustal lithospheric mantle of the
Brazilian craton beneath the Altiplano, and similar processes elsewhere
in the crustally-thickened Andes (Fig. 7c). The Andes presently are
segmented into regions of flat slab subduction, and regions of steep
subduction (Barazangi and Isacks, 1976), and some of these such as
the Peru and Chile flat slabs have been shown to be actively retreating
(Manea et al., 2012), even though South America is moving towards the
trench (van Hunen et al., 2004). The trench outboard of the Chilean flat
slab has been actively retreating for the past 25 Ma (Schellart et al.,
2007). Over the history of the Andes, different subduction segments
have changed from steep to flat, and flat to steep. Ramos and Folguera
(2009) found that when subducting slabs change from being flat to
being steep, if the overriding plate has a thick crust, they respond with
delamination of the root of overlying crustal material, with basaltic
underplating, crustal extension (Allmendinger et al., 1997), higher heat
flow and thermal uplift (Allmendinger et al., 1997), and felsic; (rhyolitic)
volcanism (Kay and Kay, 1993). If slab steepening affects an overlying thin
crust, then flood basalts erupt in areas where the lithosphere has already
been thinned (Kay et al., 1987). Although the slabs beneath the Andes are
flat at relatively shallow levels (i.e., not in the mantle transition zone), the
processes may be similar to those in the NCC. As the slab steepens, fertile
mantle material must flow in laterally from the Atlantic or deep beneath
South America, and then flows into a region hydrated by slab dehydra-
tion, partially melts, and erodes the lithospheric root.
Beck and Zandt (2002) showed that the SCLM of the Brazilian craton
is thrust beneath the Eastern Cordillera of the Andes on the edge of the
Aliplano (Fig. 7c). The root of the western part of the craton has been
thinned from a normal thickness of about 200 km to about 100 km in
this zone where the subcrustal lithosphere is delaminating piece by
piece (Fig. 7c). The lower crust is flowing in beneath the strong upper
crust towards the trench, aiding the delamination. The reason that the
destruction of the root of the Brazilian craton is much more limited in
scope than the NCC is because the flat slab beneath it is shallow, and
as it rolls back into a steep attitude, much less fertile mantle material
or ductile lower crust flows in to replace the space created by the slab
steepening/rollback.
6. Discussion
6.1. Relationships between flat slab subduction, trench suction and craton
destruction and the Cretaceous superplume, mantle avalanches, and hot
mantle temperatures
AbovewedescribedthepartialdestructionoftheNorthChina,
Wyoming and Brazilian cratons during slab rollback events in the
Cretaceous and Cenozoic. The early to middle Cretaceous was a period
of upwelling of the mid-Pacific superplume, and mantle avalanches
10 T.M. Kusky et al. / Tectonophysics xxx (2014) xxx–xxx
Please cite this article as: Kusky, T.M., et al., Flat slab subduction, trench suction, and craton destruction: Comparison of the North China,
Wyoming, and Brazilian cratons, Tectonophysics (2014), http://dx.doi.org/10.1016/j.tecto.2014.05.028

related to the closure of Tethys (e.g., Larson, 1991; Maruyama et al.,
2007a; Utsonomiya et al., 2007). This superplume probably followed
the major mantle avalanche events that started at about 180 Ma,
whose thermal effects heated the mantle to about 50 K above its normal
temperature, peaking at 125 Ma, and that lasted for about 30 Ma
(Machetel and Humler, 2003), which was also the peak time of destruc-
tion of the NCC. Also, during this time interval and the Cretaceous Nor-
mal Superchron (125–84 Ma), plate motions greatly accelerated (Seton
et al, 2012), as shown by the time (132–117 Ma) of superfast-spreading
of the Pacific plate (Zhang et al., 2014), suggesting that all these phe-
nomena may be somehow inter-related. We speculate that the higher
mantle temperatures and faster plate motions aided craton destruction
simply by making the slab rollback and mantle influx more efficient.
Using Fig. 3, if plate velocities V
S
increase, and slab penetration (V
PEN
)
is still held at zero, then the slab rollback (V
T
)andslabflattening
(V
FLAT
) will increase and the slab will rollback and flatten at the rate
of the increased plate velocity. To balance the volume, then the new
fertile mantle will have to flow into the newly created space above
the rolling-back slab at a faster rate (flux
m
), made easier by the higher
temperature-buffered mantle viscosities, enhancing the rise of melts
to interact with and thermochemically erode the roots of cratons.
6.2. Implications for models of crustal and lithospheric growth
Crustal growth models and curves have long been important to
produce, but contentious in their interpretations, largely because
they so often fail to take account of all the multitudes of variables
that inevitably relate to the extremely complex make-up and
development of the continental crust. The concept of subduction
erosion and crustal recycling at subduction zones has only been
topical in the last decade, when it became apparent that more material
was removed from the margins of continents or accreted arcs than
previously supposed (e.g. Clift et al., 2009; Scholl and von Huene,
2010). Thus, it is essentially difficult to estimate the true continental
or crustal growth rate, because of the poorly known volume of the
recycled materials (Rino et al., 2004). The concept of subduction zone
recycling does not take account of the delamination and removal of
the lithospheric rootsof thickened cratonsand orogens. If large volumes
of thickened roots of cratonic lithospheric can be recycled into the
mantle, as we propose here, this has important implications for crust–
mantle recycling and for understanding crustal growth through time
(e.g., Artemieva et al., 2002; Bowring and Housh, 1995; Condie et al.,
2009; Foley, 2008; Rino et al., 2004; Taylor and McLennan, 1995; Zhu
et al., 2011, 2012). O'Reilly et al. (2009) pointed out that high-resolution
global seismic tomography (Vs) models reveal high-velocity domains
in Africa that are interpreted as Archean depleted buoyant continental
roots that remain attached to overlying thinned continental crust.
Such high-velocity buoyant roots and overlying thinned crust locally
extend, often as fragments, well out under the deep Atlantic Ocean,
and interpreted by O'Reilly et al. (2009) as remnant fragments of litho-
sphere that were separated from main continental regions by episodes
of rifting. If this is the case, then thickened continental roots were more
extensive than normally perceived, and also disruption and removal of
the roots to the mantle has likewise been more pervasive. Clearly crust-
al growth curves are in need of recalculation and re-assessment.
6.3. Growth of the “second”and “third”continents through subduction
erosion, arc subduction, and craton destruction
Most traditional models for the evolution of continental lithosphere
are based on the premise that once continents form, they are forever
permanent features that remain on the surface, but can be broken up,
re-arranged in the supercontinent cycle, and remelted forming more
evolved magmas. Continental crust in particular is largely regarded as
unsubductable because of its low density and buoyancy relative to the un-
derlying mantle. Over the past two decades this view has been gradually
changing. It is now recognized that approximately the same volume of
material that is added to continental crust at subduction zones is brought
back into the mantle by gradual subduction erosion (e.g., von Huene and
Scholl, 1991), and in some cases entire crustal sections of immature arcs
are subducted and returned to the mantle (e.g., Kawai et al., 2013). The
reason the crustal material can accumulate there is that phase transitions
below 270 km forming jadeite-bearing assemblages make the continental
crust more dense than surrounding mantle between 270 and 660 km
(Kawai et al., 2009)Kawai et al. (2009, 2013) suggested that much of
this material has accumulated along the mantle transition zone between
660 and 410 km, and that the volume of felsic material in this zone could
be six times the volume of continental crust. They accordingly name this
region the “second continent.”In this work (and the many previous
works cited herein) we have shown that not only the continental crust
can be recycled, but large portions of the sub-continental lithospheric
mantle can be returned to the convecting mantle. Thus, our concepts of
continental and lithospheric stability through time have changed, and
are undergoing a stage of metamorphosis to a more dynamic Earth
model in which much more continental material has been extracted
from the mantle than previously thought, but most of it has been taken
back, and is now lying on the mantle transition zone as lost continents.
The Earth's earliest crust may have been anorthositic, but virtually none
of this material is left on the surface. Kawai et al. (2009) speculated that
much of this early crust may now reside along the core-mantle boundary
forming the D″layer, since anorthosite would have a similar density to
mantle at the transition zone, but with phase changes it is denser than
the lower mantle and could founder, sinking to form a “third continent”,
lost along the core–mantle boundary.
This presents a very different picture of the Earth than geoscientists
realized a decade ago. Three of the major boundaries on the planet (at-
mosphere-crust, mantle transition zone, and D″) are marked by large
regions of felsic or continental crust, with large density contrasts across
the boundaries. We know that plate tectonics operates on the surface,
giving rise to the first continent, but have not yet explored whether or
not some form of plate tectonics may operate on the second and third
continents. Just as it took centuries to document plate tectonics on the
first continent, it will be a challenge for the next generation of Earth sci-
entists to determine if plate tectonics operates along all three of the
major Earth interfaces.
7. Conclusions
Analysis of three cratons that have lost parts of their roots leads to a
general model for loss and recycling of sub-continental lithospheric
mantle.
1. Dehydration reactions from flat-lying slabs can significantly hydrate
the overlying mantle, generating melts.
2. Flat-lying slabs in the mantle transition zone are generated largely
by rollback of the subduction zone, not lateral penetration of the
slabs.
3. Rollback causes significant extension and thinning of the overlying
lithosphere, and a rise in the base of the SCLM boundary.
4. Rollback of flat-lying slabs causes huge influxes of fertile mantle to
move into the void created by the rollback.
5. The new fertile and hydrated mantle rises into the space created by
the slab rollback and lithosphere thinning, causing adiabatic melting.
The melts rise to the base of the SCLM causing melt–peridotite
reactions that destroy the roots of cratons.
6. Cratons are not forever; they can be destroyed and recycled back to
the mantle in the orogen–craton–orogen cycle.
7. Models of crustal growth need to be modified with the new recog-
nition that much more continental lithosphere has been recycled
to the mantle than previously thought.
8. Estimates of mantle composition need to take into consideration a
much larger flux of recycled crustal and SCLM material.
11T.M. Kusky et al. / Tectonophysics xxx (2014) xxx–xxx
Please cite this article as: Kusky, T.M., et al., Flat slab subduction, trench suction, and craton destruction: Comparison of the North China,
Wyoming, and Brazilian cratons, Tectonophysics (2014), http://dx.doi.org/10.1016/j.tecto.2014.05.028

Acknowledgements
This study was supported by the grants National Natural Science
Foundation of China (Nos. 91014002 and 40821061) and Ministry of
Education of China (No. B07039). We thank two anonymous reviewers
from Tectonophysics for helpful comments which greatly improved the
paper.
References
Allmendinger,R.W., Jordan, T.E.,Kay, S.M., Isacks, B.L., 1997. The evolution of the Alti plano–
Puna Plateau of the Central Andes. Ann. Rev. Earth Planet. Sci. 25, 139–174.
Artemieva, I.,Mooney, W., Perchuc, E., Thybo,H., 2002. Processes of lithospheric evolution:
new evidence on the structure of the continental crust and uppermost mantle.
Tectonophysics 358, 1–15.
Barazangi, M., Isacks, B., 1976. Spatial distribution of earthquakes and subduction of the
Nazca plate beneath South America. Geology 4, 686–692.
Beck, S.L., Zandt, G., 2002. The nature of orogenic crust in the central Andes. J. Geophys.
Res. 107, B10. http://dx.doi.org/10.1029/2000JB000124.
Becker, T.W., Faccenna, C., 2009. A review of the role of subduction dynamics for regional
and global plate motions. In: Lal lemand, S., Funiciello (Eds.), Su bduction Zone
Geodynamics. Springer, Berlin, pp. 3–34.
Bercovici, D., Karato, S., 2003. Whole-mantle convection and the transition-zone water
filter. Nature 425, 39–44.
Bowring, S.A., Housh, T., 1995. The Earth's early evolution. Science 269, 1535–1540.
Bromiley, G.D. , Nestola, F., Redfe rn, S.A.T., Zhang, M., 2010. Water incorporation in
synthetic and natural MgAl
2
O
4
spinel. Geochim. Cosmochim. Acta 74, 705–718.
Chamberlain, K., Frost, C., Frost, R., 2003. Early Archean to Mesoproterozoic evolution of
the Wyoming Prov ince: Archean var iations in Mesozo ic and Tertiary p lutons of
central Idaho. Contrib. Mineral. Petrol. 90, 291–308.
Christensen, U.R., 1996. The influence of trench migration on slab penetration into the
lower mantle. Earth Planet. Sci. Lett. 140, 27–39.
Clift, P.D., Schouten, H., Vannucchi, P., 2009. Arc–continent collisions, sediment recycling,
and the maintenan ce of the continental crust. In: Cawood, P., Kröner, A. (Ed s.),
Accretionary Orogens inSpace and Time. Geological Society of London, Special Publi-
cations, 318, pp. 75–103.
Condie, K.C., 2005. Earth as an Evolving Planetary System. Elsevier, Amsterdam (447 pp.).
Condie, K.C., Belousova, E., Griffin, W.L., Sircombe, K.N., 2009. Granitoid events in space
and time: constraints from igneous and detrital zircon age spectra. Gondwana Res.
15, 228–242.
Cook, F.A., van der Velden, A.J., Hall, K.W., Roberts, B.J., 1998. Tectonic delamination and
subcrustal imb rication of the Precambrian lithosphere in northwe stern Canada
mapped by LITHOPROBE. Geology 26, 839–842.
Darby, B.J., Davis, G.A., Zhang, X.H., 2004. The newly discovered Waziyu metamorphic
core complex, Yiwulu Shan, western Liaoning Province, Northwest China. Earth Sci.
Front. 11, 143–155.
Davies, G.F., 1995. Penetration of plates and plumes through the mantle transition zone.
Earth Planet. Sci. Lett. 133, 507–516.
De Smet, J.H., van den Berg, A.P., Vlaar, N.J., 1999. The evolution of continental roots in
numerical thermo-chemical mantle convection models including differentiation by
partial melting. Lithos 48, 153–170.
Dokuz, A., 2011. A slab detachment and delamination model for the generation of
Carboniferous high-potassium I-type magmatism in the Eastern Pontides, NE
Turkey: the Köse composite pluton. Gondwana Res. 19, 926–944.
Dueker, K., Yuan, H., Zurek, B., 2001. Thick structured Proterozoic lithosphere of the Rocky
Mountain Region. GSA Today 11, 109.
Durrheim, R.J., Mooney, W.D., 1994. Evolution of the Precambrian lithosphere: seismological
and geochemical constraints. J. Geophys. Res. 98, 15,359–15,374.
Engebretson, D.C., Cox, A., Gordon, R.G., 1985 . Relative motions between oceanic and
continental plates in the Pacific Basin. Geol. Soc. Am. Spec. Pap. 206, 1–58.
English, J.M., Johnston, S.T., 2004. The Laramide orogeny: what were the driving forces?
Int. Geol. Rev. 46, 833–838.
Fan, W., Menzies, M.A., 1992. Contribution of the lithospheric mantle to extension-
related volcanics —geochemical evidence from Cenozoic basaltic rocks in
Hainan Island and Leizhou Peninsula, southern China. In: Ruoxin, L. (Ed.),
Chronology and Geochemistry of Cenozoic Volcanic Rocks in China. Seismology
Press, China, pp. 320–329.
Faul, U.H., Jackson, I., 2005. The seismological signature of temperature and grain size
variations in the upper mantle. Earth Planet. Sci. Lett. 234, 119–134.
Foley, S.F., 2008. Rejuvenation and erosion of the cratonic lithosphere. Nat. Geosci. 1,
503–510.
Foster, D.A., Mueller, P.A., Mogt, D.W., Wooden, J.L., Vogl, J.J., 2006. Proterozoic evolution
of the western margin of the Wyomi ng craton: implications for the tectonic and
magmatic evolution of the northe rn Rocky Mountains. Can. J. Earth Sc i. 43,
1601–1619.
Frost, D.J., 1999. The stability of dense hydrous magnesium silicates in earth's transition
zone and lower mantle. Geochem. Soc. Spec. Publ. 6, 283–297.
Frost, C.D., Fruchey, B.L., Chamberlain, K.R., et al., 2006. Archean crustal growth by lateral
accretion of juvenile supracrustal belts in the south-central Wyoming Province. Can.
J. Earth Sci. 43, 1533–1555.
Gao, S., Luo, T.C., Zhang, B.R., Zhang, H.F., Han, Y.W., Zhao, Z.D., Hu, Y.K., 1998. Chemical
composition of the continental crust as revealed by studies in East China. Geochim.
Cosmochim. Acta 62, 1959–1975.
Gao,S.,Rudnick,R.,Carlson,R.,McDonough,W.,Liu,Y.S.,2002.Re–Os evidence for
replacement of ancient mantle lithosphere beneath the North China craton.
Earth Planet. Sci. Lett. 6135, 1–15.
Gao, S., Rudnick, R.L., Yuan, H.-L., Liu, X.-M., Liu, Y.-S., Xu, W.L., Ling, W.-L., Ayers, J., Wang,
X.-C., Wang, Q.H., 2004. Recycling lower continental crust in the North China craton.
Nature 432, 892–897.
Gao, S., Zhang, J.F., Xu, W.L., 2009. Delamination and destruction of the North China craton.
Chin. Sci. Bull. 54, 3367–3378.
Goodwin, A.M., 1991. Precambrian Geology: The Dynamic Evolution of the Continental
Crust. Academic Press, London (666 pp.).
Griffin, W.L., Zhang, A., O'Reilly, S.Y., Ryan, C.G., 1998. Phanerozoic evol ution of the
lithosphere beneath the Sino-Korean craton: mantle dynamics and plate interactions
in East Asia. Geodynamics 27, 107–126.
Griffin, W.L., O'Reilly, S.Y., Abe, N., Aulbach, S., Davies, R.M., Pearson, N.J., Doyle, B.J., Kivi, K.,
2003a. The origin and evolution of Archean lithospheric mantle. Precambrian Res. 127,
19–41.
Griffin, W.L., O'Reilly, S.Y., Natapov, L.M., Ryan, C.G., 2003b. The evolution of lithos pheric
mantle beneath the Kalahari Craton and its margins. Lithos 71, 215–241.
Griffin, W.L., Kobussen, A.F., Babu, E.V.S.S.K., O'Reilly, S.Y., Norris, R., Sengupta, P., 2009. A
translithospheric suture in the vanished 1-Ga lithospheric root of South India: evi-
dence from contrasting lithospheric sections in the Dharwar Craton. Lithos 1125,
1109–1119.
Griffin, W.L., Begg, G.C., O'Reilly, S.Y., 2013. Continental-root control on the genesis of
magmatic ore deposits. Nat. Geosci. 6, 905–910.
Helmstaedt, H.H., Gurney, J.J., 1995. Geotectonic controls of primary diamond deposits:
implications for area selection. J. Geochem. Explor. 53, 125–144.
Herzberg, C., Rudnick, R., 2012. Formation of cratonic lithosphere: an integrated thermal
and petrological model. Lithos 149, 4–15.
Hoffman, P.F., 1988. United plates of America, the birth of a craton; early Proterozoic
assembly and growth of Laurentia. Ann. Rev. Earth Planet. Sci. 16, 543–603.
Huang, J.L., Zhao, D.P., 2006. High resolution mantle tomography of China and surround-
ing regions. J. Geophys. Res. 111, B09305.
Humphreys, E., Hessler, E. , Dueker, K., Farmer, G.L., Erslev, E., Atwater, T., 2003. How
Laramide-age hydration of North American lithosphere by the Farallon slab con-
trolled subsequent activity in the western United States. Int. Geol. Rev. 4, 575–595.
Ichiki, M., Baba, K., Obayashi, M., Utada, H., 2006. Water content and geotherm in the
upper mantle above the stagnant slab: interpretation of electrical conductivity and
seismic P-wave velocity models. Phys. Earth Planet. Inter. 155, 1–15.
Isozaki, Y., Aoki, K., Nakama, T., Yanai, S., 2010. New insight into a subduction- related
orogen: a reappraisal of the geotectonic framework and evolution of the Japanese
Islands. Gondwana Res. 18, 82–105.
Jagoutz, O., Behn, M.D., 2013. Foundering of lower island-arc crust as an explanation for
the origin of the continental Moho. Nature 504, 131–134.
Jones, C.H., Lang Farmer, G., Unruh, J., 2004. Tectonics of Pliocene removal of lithosphere
of the Sierra Nevada, California. Geol. Soc. Am. Bull. 116, 1408–1422.
Jordan, T.H., 1975. The continental tectosphere. Rev. Geophys. Space Phys. 13 (3), 1–12.
Jordan, T.H., 1981. Continents as a chemical boundary layer. Phil. Trans. Roy. Soc. London
A301,359–373.
Jordan, T.H., 1988. Structure and formation of the continental lithosphere. In: Menzies,
M.A., Cox,K. (Eds.), Oceanic and Continental Lithosphere, Similaritiesand Differences.
Journal of Petrology, Special Volume 1, pp. 11–37.
Kaban, M.K., Schwintzer, P., Artemieva, I.M., Mooney, W.D., 2003. Density of thecontinen-
tal roots:compositional and thermal contributions. Earth Planet. Sci. Lett. 209,53–69.
Karato, S., 2003. Mapping water content in the upper mantle. In: Eiler, J. (Ed.), Inside the
Subduction Factory. American Geophysical Union Monograph, 138, pp. 135–152.
Kawai, K., Tsuch iya, T., Tsuchiya , J., Maruyama, S., 2 009. Lost primordial continents.
Gondwana Res. 16, 581–586.
Kawai, K., Yamamoto, S., Tsuchiya, T., Maruyama, S., 2013. The second continent: exis-
tence of grantitic continental materials around the bottom of the mantle transition
zone. Geosc i. Front. 4, 1–6.
Kawamoto, T., 2006. Hydrous phases and water transport in the subducting slab. In:
Keppler, H., Smyth, J.R. (Eds.), Water in Nominally Anhydrous Minerals. Reviews in
Mineralogy and Geochemistry, 62. Geochemical Society of America, pp. 273–289.
Kay, R.W., Kay, S.M., 1993. Delamination and delamination magmatism. Tectonophysics
219, 177–189.
Kay, S.M., Maksaev, V., Moscoso, R., Mpodozis, C., Nasi, C., 1987. Probing the evolving
Andean lithosphere: Mid-late Tertiary magmatism in Chile (29°–30°30′S) over the
modern zone of subhorizontal subduction. J. Geophys. Res. 92 (B7), 6173–6189.
Keller, R.G., 2008. Is the Wyoming craton being decratonized? Geol. Soc. Am. Abstr.
Programs 40, 327.
Kincaid, C., Olson, P., 1987. An experimental study of subduction and slab migration. J.
Geophys. Res. 92, 13832–13840.
Kusky, T.M., 1993. Collapse of Archean orogens and the origin of late- to post-kinematic
granitoids. Geology 21, 925–929.
Kusky, T.M., 2010. Encyclopedia of Earth and Space Science. Facts on File, New York 20,
26–35 (two volumes, 960 pp., ISBN-10: 0816070059; SBN-13: 978-0-8160-7005-3).
Kusky, T.M., 2011. Geophysical and geological tests of tectonic models of the North China
Craton. Gondwana Res. 20, 26–35.
Kusky, T.M., Li, J.H., 2003. Paleoproterozoic tectonic evolution of the North China Craton.
J. Asian Earth Sci. 22, 383–397.
Kusky, T.M., Polat, A., 1999. Growth of granite-greenstone terranes at convergentmargins
and stabilization of Archean cratons. In: Marshak, S., van der pluijm, B. (Eds.), Special
Issue on Tectonics of Continental Interiors. Tectonophysics, 305, pp. 43–73.
Kusky, T.M., Santosh, M., 2009. The Columbia Connection in North China. In: Reddy, S.M.,
Mazumder, R., Evans, D., Collins, A.S. (Eds.), Paleoproterozoic Supercontinents and
Global Evolution. Geological Society of London Special Publications, 323, pp. 49–71.
12 T.M. Kusky et al. / Tectonophysics xxx (2014) xxx–xxx
Please cite this article as: Kusky, T.M., et al., Flat slab subduction, trench suction, and craton destruction: Comparison of the North China,
Wyoming, and Brazilian cratons, Tectonophysics (2014), http://dx.doi.org/10.1016/j.tecto.2014.05.028

Kusky, T.M., Windley, B.F., Zhai, M.G., 2007a. Tectonic evolution of the North China Block:
from orogen to craton to orogen. In: Zhai, M.G., Windley, B.F., Kusky, T.M., Meng, Q.R.
(Eds.), Mesozo ic Sub-Continental Lithospheric Thinning under E astern Asia.
Geological Society of London Special Publications, 280, pp. 1–34.
Kusky, T.M., Win dley, B.F., Zhai, M. G., 2007b. Lithos pheric thinning in eastern Asia;
constraints, evolution, and tests of models. In: Zhai, M.G., Windley, B.F., Kusky, T.M.,
Meng, Q.R. (Eds.), Mesozoic Sub-Continental Lithospheric Thinning under Eastern
Asia. Geological Society of London Special Publications, 280, pp. 331–343.
Lallemand, S., 1995. High rates of arc consumption by subduction processes: some
consequences. Geology 23, 551–554.
Larson, R.L., 1991. Latest pulse of the Earth: evidence for a mid-Cretaceous superplume.
Geology 19, 547–550.
Li, S.L., Lai, X.L., Liu, B.F., Wang, Z.S., He, J.Y., Sun, Y., 2011a. Differences in lithospheric
structures between two sides of Taihang Mountain obtained from the Zhucheng–
Yichuan deep seismic sounding profile. Sci. China 54, 871–880.
Li, Y.Y., Yang, Y.S., Kusky, T.M., 2011b. Lithospheric structure in the North China craton
constrained from Gravity Field Model (EGM 2008). J. Earth Sci. 22, 260–272.
Liegeois, J.-P., Abdelsalam, M.G., Ennih, N., Ouabadi, A., 2013. Metacraton: nature, genesis,
and behavior. Gondwana Res. 23, 220–237.
Lin, W., Wang, Q.C., Wang, J., Wang, F., Chu, Y., Chen, K., 2011. Late Mesozoic extensional
tectonics of the Liaodong Peninsula massif: response of crus t to continental
lithosphere destruction of the North China craton. Sci. China 54, 843–857.
Lithgow-Bertelloni, C., Richards, M.A., 1998.The dynamics of Cenozoic and Mesozoicplate
motions. Rev. Geophys. 36, 27–78.
Liu,J.G.,Rudnick,R.L.,Walker,R.J.,Gao,S.,Wu,F.Y.,Piccoli,P.M.,Yuan,H.L.,Xu,W.L.,Xu,Y.G.,
2011a. Mapping lithospheric boundaries using Os isotopes of mantle xenoliths:an
example from the North China Craton. Geochem. Cosmochem. Acta 75,
3881–3902.
Liu, J.L., Ji, M., Shen, L., Guan, H.M., Davis, G.A., 2011b. Early Cretaceous extensional struc-
tures in the Liaodong Peninsula: structural associations, geochronological constraints
and regional tectonic implications. Sci. China 54, 823–842.
Liu, L., Spasojevic, S., Gurnis, M., 2008. Reconsctucting Farallon plate subduction beneath
North America back to the Late Cretaceous. Science 322, 934–938.
Machetel, P., Humler, E., 2003. High mantle temperature during Cretaceous avalanche.
Earth Planet. Sci. Lett. 208, 125–133.
Manea, V.C., Perez-Gussinye, M., Manea, M., 2012. Chilean flat slab subduction controlled
by overriding plate thickness and trench rollback. Geology 40, 35–38.
Maruyama, S., Okamoto,K., 2007. Water transportation from the subducting slab into the
mantle transition zone. Gondwana Res. 11, 148–165.
Maruyama, S., Yuen, D.A., Windley, B.F., 2007a. Dynamics of plumes and
superplumes through time. In: Yuen, D.A., Maruyama, S., Karato, S.I., Windley,
B.F. (Eds.), Superplumes: Beyond Plate Tectonics. Springer, Dordrecht, pp.
441–502.
Maruyama, S., Santosh, M., Zhao, D. , 2007b. Superplume, supercontinent, and post-
perovskite: mantle dynamics and anti-plate tectonics on the core-mantle boundary.
Gondwana Res. 11, 7–37.
Menzies, M., Fan, W.M., Zhang, M., 1993. Paleozoic and Cenozoic lithoprobes and loss
of N120 km of Archean lithosphere, Sino-Korean craton, China. In: Prichard, H.M.,
Alabaster, T., Harris, N.B.W., Neary, C.R. (Eds.), Magmatic Processes and Plate Tectonics.
Geological Society of London Special Publications, 76, pp. 71–81.
Mueller, P.A., Frost, C.D., 2006. The Wyoming Province: a distinctive Archean craton in
Laurentian North America. Can. J. Earth Sci. 43, 1391–1397.
Niu, Y.L., 2005. Generation and evolution of basaltic magmas: some basic concepts and a
new view on the origin of Mesozoic–Cenozoic basaltic volcanism in eastern China.
Geol. J. China Univ. 11, 9–46.
Norton, I.O., 2007. Speculations on Cretaceous tectonic history of the northwest Pacific
and a tectonic origin for the Hawaii hotspot. In: Foulger, G.R., Jurdy, D.M. (Eds.),
Plates, Plumes , and Planetary Pr ocesses. Geological Society of Ame rica Special
Paper, 430, pp. 451–470.
O'Reilly, S.Y., Griffin, W.L., Poudjom, Djomani, Y.H., Morgan, p., 2001. Are lithospheres
forever? Tracking changes in subcontinental lithospheric mantle through time. GSA
Today 11, 4–10.
O'Reilly, S.Y., Zhang, M., Griffin, W.L., Begg, G., Hronsky, J., 2009. Ultradeep continental
roots and their oceanic remnants: a solution to the geochemical “mantle reservoir”
problem? Lithos 112, 1043–1054.
Pavlis, G.L., 2011. Three-dimensional wave-field imaging of data from the US Array: new
constraints on the geometry of the Farallon slab. Geosphere 7, 785–801.
Peacock, S.M., 1993. Large-scale hydration of the lithosphere above subducting slabs.
Chem. Geol. 08, 49–59.
Peacock, S.M., 2003. Thermal structure and metamorphic evolution of subducting slabs.
In: Eiler, J. (Ed .), Inside the Subduction Factory. American Geophysical Union
Monograph, 138, pp. 7–22.
Pollack, H.N., 1986. Cratonization and thermal evolution of the mantle. Earth Planet. Sci.
Lett. 80, 175–182.
Prodehl, C., Mooney, W.D., 2012. Exploring the Earth's crust —history and res ults of
controlled-source seismology. Geol. Soc. Am. Mem. 208, 764.
Ramos, V.A., Folguera,A., 2009. Andean flat-slab subduction through time. Geol. Soc.Lond.
Spec. Publ. 327, 31–54.
Richard, G., Bercovici, D., Karato, S., 2006. Slab dehydration in the Earth's mantle transi-
tion zone. Earth Planet. Sci. Lett. 251, 156–167.
Rino, S., Komiya, T., Windley, B.F., Katayama, I., Motoki, A.,Hirata, T., 2004. Majorepisodic
increases of continental crustal growth determined from zircon ages of river sands;
implications fo r mantle overturns in the Early Precambrian. Phys. Earth Planet.
Inter. 146, 369–394.
Rollinson, H., 2010. Coupled evolution of Archean continental crust and sub-continental
lithospheric mantle. Geology 38, 1083–1086.
Rudnick, R.L., 1995. Making continental crust. Nature 378, 571–578.
Schellart, W.P. , Freeman, J., Stegma n, D.R., Moresi, L., May, D., 2007. Evoluti on and
diversity of subduction zones controlled by slab width. Nature 446, 308–311.
Schellart, W.P., Stegman, D.R., Freeman, J., 2008. Global trench migration velocities
and slab migration induced upper mantle volume fluxes: constraints to find an
Earth reference frame based on minimizing viscous dissipation. Earth Sci. Rev.
88, 118–144.
Schmidt,M.W., 1995. Lawsonite;upper pressure stability and formation of higher-density
hydrous phases. Am. Mineral. 80, 1286–1292.
Scholl, D.W., von Huene, R., 2010. Subduction zone recycling processes and the rock
record of crustal suture zones. Can. J. Earth Sci. 47, 633–654.
Seton, M., Müller, R.D.,Zahirovic, S., Gaina, C., Torsvik, T., Shephard, G., Talsma, A., Gurnis,
M., Turner, M., M aus, S., Chandler , M., 2012. Global continental a nd ocean basin
reconstructions since 200 Ma. Earth Sci. Rev. 113, 212–270.
Smith, A.D., 2003. A reappraisal of stress field and convective roll models for the origin
and distribution of Cretaceous to recent intraplate volcanism in the Pacific Basin.
Int. Geol. Rev. 45, 287–302.
Smyth, J.R., Frost, D.J., 2002. The effect of water on the 410-km discontinuity: an experimen-
tal study. Geophys. Res. Lett. 29, 1485–1489. http://dx.doi.org/10.1029/2001GL014418.
Smyth, J.R., Holl, C.M., Frost, D.J., Jacobsen, S.D., Langenhorst, F., McCammon, C.A., 2003.
Structural systematics of hydrous ringwoodite in Earth's interior. Am. Mineral. 88,
1402–1407.
Snelson, C.M., He nstock, T.J., Kel ler, R.G., Miller, K.C., Levander, A. , 1998. Crustal and
uppermost mantle structure along the Deep Probe seismic profile. Rocky Mt. Geol.
33, 181–198.
Stachel, T., Viljoen, K.S., Brey, G., Harris, J.W., 1998. Metasomatic processes in lherzolitic
and harzburgitic domains of diamondiferous lithospheric mantle: REE in garnets
from xenoliths and inclusions in diamonds. Earth Planet. Sci. Lett. 159, 1–12.
Stachnik, J.C., Dueker, K., Schutt, D., Yuan, H., 2008. Imaging Yellowstone plume–
lithosphere interactions from inversion of ballistic and diffusive Rayleigh
wave dispersion and crustal thickness data. Geochem. Geophys. Geosyst. 267.
http://dx.doi.org/10.1029/2007GC001901.
Steinberger, B., 2008. Reconstructing Earth history in three dimensions. Science 322,
866–868.
Stern, C.R., 2011. Subduction erosion: rates, mechanisms, and its role in arc magmatism
and the evolution of the continental crust and mantle. Gondwana Res. 20, 284–308.
Taylor, S.R., McLennan, S.M., 1995. The geochemical evolution of the continental crust.
Rev. Geophys. 33, 241–265.
Tian, Y.,Zhao, D.P., 2013. Reactivationand mantle dynamicsof North China craton:insight
from P-wave anisotropy tomography. Geophys. J. Int. 195, 1796–1810.
Tolstikhin, I., Kramers, J., 2008. The Evolution of Matter: From the Big Bang to the Present
Day Earth. Cambridge University Press, Cambridge, UK (532 pp.).
Tonegawa, T., Hirahara , K., Shibutani, T., Iwamori , H., Kanamori, H., Shiomi, K., 2008.
Water flow to the mantle transition zone inferred from a receiver function image
of the Pacific slab. Earth Planet. Sci. Lett. 274, 346–354. http://dx.doi.org/10.1016/j.
epsl.2008.07.046.
Trubitsyn, V.P. , Mooney, W.D., Ab bott, D.H., 2003. Cold cratonic roots and thermal
blankets: how continents affect mantle convection. Int. Geol. Rev. 45, 479–496.
Ulmer, P., Trommsdorff, V., 1995. Serpentine stability to mantle depths and subduction-
related magmatism. Science 268, 858–861.
Utsonomiya, A., Ota, T., Windley, B.F., Suzuki, N., Uchio, Y., Munekata, K., Maruyama, S.,
2007. History of the Pa cific superplume: imp lications for Pacific pal eogeography
since the late Proterozoic. In: Yuen, D.A., Maruyama, S., Karato, S.I., Windley, B.F.
(Eds.), Superplumes: Beyond Plate Tectonics. Springer, Dordrecht, pp. 363–408.
Van der Meer, D.G., Torsvik, T.H., Spakman, W., van Hinsbergen, D.J.J., Amaru, M.L., 2012.
Intra-Panthalassa Ocean subduction zones revealed by fossil arcs and mantle structure.
Nat. Geosci. 5, 215–219.
Van Hunen, J., van den Berg, A.P., Vlaar, N.J., 2000. A thermo-mechanical model of hori-
zontal subduction below and overriding plate. Earth Planet. Sci. Lett. 182, 157–169.
van Hunen, J., van der Berg, A.P., Vlaar, N.J., 2004. Various mechanisms to induce present-
day flat subduction and implications for the younger Earth: a numerical parameter
study. Phys. Earth Planet. Inter. 179–194. http://dx.doi.org/10.1016/j.pepi.2003.07.027.
Vlaar, N.J., 1983. Thermal anomalies and magmatism due to lithospheric doubling and
shifting. Earth Planet. Sci. Lett. 65, 322–330.
von Huene, R., Scholl, D.W., 1991. Observations at convergent margins concerning sedi-
ment subduction, subduction erosion , and the growth of contin ental crust. Rev.
Geophys. 29, 279–316.
Wang, J., Zhao, D.P., Yao, Z.X., 2013 . Crustal and upper most mantle structure and
seismotectonics of North China craton. Tectonophysics 582, 177–187.
Whitmeyer, S.J., Karlstrom, K.E., 2007. Tectonic model for theProterozoic growthof North
America. Geosphere 3, 220–259.
Wilde, S.A., Zhou, X.-H., Nemchin, A.A., Sun, M., 2003. Mesozoic crust–mantle interaction
beneath theNorth China craton—a consequence of the dispersalof Gonwanalandand
accretion of Asia. Geology 31, 817–820.
Williams, Q., Hemley, R.J., 2001. Hydrogen in the deep earth. Ann. Rev. Earth Planet. Sci.
29, 365–418.
Windley, B.F., 1995. The Evolving Continents, 3rd ed. Wiley, New York (526 pp.).
Windley, B.F., Maruyama, S., Xiao, W.J., 2010. Delamination/thinning of sub-continental
lithosphereic mantle under eastern China: the role of water and multiple subduction.
Am. J. Sci. 910, 1250–1298.
Wooden, J.L., Mueller, P.A., 1988. Pb, Sr, and Nd isotopic compositions of a suite of Late
Archean igneous rocks, eastern Beartooth Mountains: implications for crust–mantle
evolution. Earth Planet. Sci. Lett. 87, 59–72.
Wu, F.Y., Walker, R.J., Ren, X.W., Sun, D.Y., Zhou, X.H., 2003. Osmium isotope constraints
on the age of the lithospheric mantle bene ath northeastern Ch ina. Chem. Geol.
14156, 1–23.
13T.M. Kusky et al. / Tectonophysics xxx (2014) xxx–xxx
Please cite this article as: Kusky, T.M., et al., Flat slab subduction, trench suction, and craton destruction: Comparison of the North China,
Wyoming, and Brazilian cratons, Tectonophysics (2014), http://dx.doi.org/10.1016/j.tecto.2014.05.028

Wu, F.Y., Xu, Y.G., Gao, S., 2008. Controversial on the studies of lithospheric thinning and
craton destruction of North China. Acta Petrol. Sin. 24, 1145–1174.
Xu, Y.G., Huang, X.L., Ma, J.l., Wang, Y.B.,Iizuka, Y., Xu, J.F.,Wang, Q., Wu, X.Y., 2004.Crust–
mantle interac tion during the tec tono-thermal reactivation of the No rth China
Craton: constra ints from SHRIMP zir con U–Pb chronology and geochemistr y of
Mesozoic plutons from western Shangdong. Contrib. Mineral. Petrol. 147, 750–767.
Xu, W.W., Zheng, T.Y., Zhao, Li, 2011. Mantle dynamics of the reactivating North China
craton: constraintsfrom the topographies of the 410-km and 660-km discontinuities.
Sci. China 54, 881–887.
Yang, W.C.,2003. Flat mantle reflectors in EasternChina: possible evidence of lithospheric
thinning. Tectonophysics 369, 219–230.
Yang, Y.T., 2013. An unrecognized major collision of the Okhotomorsk Block with East
Asia during the Late Cretaceous, constraints on the plate reorganization of the North-
west Pacific. Earth Sci. Rev. 126, 96–115.
Yuan, X.C.,1996. Chief compiler, Atlas of Geophysics in China, Publication 201 ofthe Inter-
national Lithosphere Program. Geological Publishing House, Beijing.
Yuan, H., Dueker, K., 2005. Upper mantle tomographic Vp and Vs images f the Rocky
Mountains in Wyoming,Colorado and New Mexico: Evidence for thick laterally
heterogeneous lithosphere.In:Keller,R.G.,Karlstrom,K.E.(Eds.),TheRocky
Mountain region: an evolving lithosphere: Tectonics, Geochemistry, and
Geophysics. 154 . American Geophysical Union Geophysical Monograph, Boulder,
CO, pp. 325–345.
Yuan, H., Dueker, K.G., 2010. Relict slab and young plume: seismic view of the present
time Wyoming lith osphere. Earth Sci. Fron t. 17, 127–138.
Zhai, M., Liu,W.J., 2003. Paleoproterozoic tectonic history of the North China craton: a re-
view. Precambrian Res. 122, 183–199.
Zhai, M.G., Santosh, M., 2011. Theearly Precambrian odyssey of the North China Craton: a
synoptic overview. Gondwana Res. 20, 6–25.
Zhang,C.H.,Li,C.M.,Deng,H.L.,Liu,Y.,Liu,L.,Wei,B.,Li,H.B.,Liu,Z.,2011.Mesozoic
contraction deformation in the Yanshan and northern Taihang mountains and
its implications to the destruction of the North China craton. Sci. China 54,
798–823.
Zhang,J.B.,Ling,W.L.,Liu,Y.S.,Duan,R.C.,Gao,S.,Wu,Y.B.,Yang,H.M.,Qiu,X.F.,Zhang,Y.Q.,
2014. Episodic Mesozoic thickening and reworking of the North China Archean lower
crust correlated to the fast-spreading Pacificplate.J.AsianEarthSci.80,63–74.
Zhao, D.P., Ohtani, E., 2009. Deep slab subduction and dehydration and their geodynamic
consequences: evidence from seismology and mineral physics. Gondwana Res. 16,
401–413.
Zhao, G.C., Zhai, M.G., 2013. Lithotectonic elements of Precambrian basement in the North
China craton: review and tectonic implications. Gondwana Res. 23, 1207–1240.
Zhao, G.C., Wilde, S.A., Cawood, P.A., Sun, M., 2001. Archean blocks and their boundaries
in the North China Craton: lithological, geochemical, structural and P–T path con-
straints and tectonic evolution. Precambrian Res. 107, 45–73.
Zhao, D.P., Maruyama, S., Omori, S., 2007. Mantle dynamics of Western Pacific and East
Asia: insight from sei smic tomography and mineral physics. Gondwan a Res. 11,
120–131.
Zheng, Y.F., Wu, F.Y., 2009. Growth and reworking of cratonic lithosphere. Chin. Sci. Bull.
54, 3347–3353.
Zheng, J., Griffin, W.L., O'Reilly, S.Y., Liou, J.G., Zhang, R.Y., Lu, F., 2005. Late Mesozo-
ic–Eocene mantle replacement beneath the eastern North China Craton: evi-
dence from the Paleozoic and Cenozoic peridotite xenoliths. Int. Geol. Rev. 47,
457–472.
Zhu, R.X., Chen, L., Wu, F.Y., Liu, J.L., 2011. Timing, scale and mechanism of the destruction
of the North China Craton. Sci. China 54, 789–797.
Zhu, R., Xu, Y.G., Zhu, G., Zhang, H.F., Xia, Q.K., Zheng, T.Y., 2012. Destruction of the North
China Craton. Sci. China 55, 1565.
14 T.M. Kusky et al. / Tectonophysics xxx (2014) xxx–xxx
Please cite this article as: Kusky, T.M., et al., Flat slab subduction, trench suction, and craton destruction: Comparison of the North China,
Wyoming, and Brazilian cratons, Tectonophysics (2014), http://dx.doi.org/10.1016/j.tecto.2014.05.028