Geoscience Frontiers 6 (2015) 49e59

H O S T E D BY Contents lists available at ScienceDirect

China University of Geosciences (Beijing)

Geoscience Frontiers
journal homepage: www.elsevier.com/locate/gsf

Research paper

Geodynamics of oceanic plateau and plume head accretion and their

role in Phanerozoic orogenic systems of China
Peter G. Betts a, *, Louis Moresi a, b, c, Meghan S. Miller d, David Willis a
a
School of Earth, Atmosphere and Environment, Monash University, Clayton Campus, VIC 3800, Australia
b
School of Mathematical Sciences, Monash University, Clayton Campus, VIC 3800, Australia
c
Department of Earth Sciences, University of Melbourne, Melbourne, VIC 3001, Australia
d
Department of Earth Sciences, University of Southern California, USA

a r t i c l e i n f o a b s t r a c t

Article history: We present three 3D numerical models of deep subduction where buoyant material from an oceanic
Received 31 January 2014 plateau and a plume interact with the overriding plate to assess the inuence on subduction dynamics,
Received in revised form trench geometry, and mechanisms for plateau accretion and continental growth. Transient instabilities of
14 July 2014
the convergent margin are produced, resulting in: contorted trench geometry; trench migration parallel
Accepted 21 July 2014
Available online 9 August 2014
with the plate margin; folding of the subducting slab and orocline development at the convergent
margin; and transfer of the plateau to the overriding plate. The presence of plume material beneath the
oceanic plateau causes at subduction above the plume, resulting in a bowed shaped subducting slab.
Keywords:
Subduction
In plateau-only models, plateau accretion at the edge of the overriding plate results in trench migration
Accretion around the edge of the plateau before subduction is re-established directly behind the trailing edge of the
Plateau plateau. The plateau shortens and some plateau material subducts. The presence of buoyant plume
Plume material beneath the oceanic plateau has a profound inuence on the behaviour of the convergent
Central Asian Orogen margin. In the plateau plume model, plateau accretion causes rapid trench advance. Plate convergence
Geodynamics is accommodated by shearing at the base of the plateau and shortening in the overriding plate. The
trench migrates around the edge of the plateau and subduction is re-established well behind the trailing
edge of the plateau, effectively embedding the plateau into the overriding plate. A slab window forms
beneath the accreted plateau and plume material is transferred from the subducting plate to the over-
riding plate through the window. In all of the models, the subduction zone maintains a relatively stable
conguration away from the buoyancy anomalies within the downgoing plate. The models provide a
dynamic context for plateau and plume accretion in Phanerozoic accretionary orogenic systems such as
the East China Orogen and the Central Asian Orogen (Altiads), which are characterised by accreted
ophiolite complexes with diverse geochemical afnities, and a protracted evolution of accretion of exotic
terranes including oceanic plateau and terranes with plume origins.
2014, China University of Geosciences (Beijing) and Peking University. Production and hosting by
Elsevier B.V. All rights reserved.

1. Introduction
Plume-subduction interactions and oceanic plateau-subduction
interactions allow buoyant oceanic lithosphere to engage with
convergent margins (Cloos, 1993; Murphy et al., 1998). On the
modern Earth the inuence and occurrence of upwelling mantle
hotspots at downgoing subducting oceanic lithosphere have been
* Corresponding author. Tel.: 61 3 9905 4150; fax: 61 3 9905 4903.
E-mail address: peter.betts@monash.edu (P.G. Betts).
Peer-review under responsibility of China University of Geosciences (Beijing).
imaged in seismic tomographic data (Zhao et al., 2007; Miller and
Lee, 2008), and presence of oceanic plateaus, seamounts, and
other aseismic ridges occur along most convergent margin (Vogt,
1973), where they are correlate with regions of signicant trench
and slab modications (Miller et al., 2004, 2005; Mason et al.,
2010). A third scenario exists when a plateau or seamount in-
teracts with the convergent margin whilst a plume is impinging at
the base of the oceanic lithosphere. For example, aseismic ridges
within the Nazca plate may represent incipient stage of plateau
accretion with the plume interacting with the ridges that have not
been subducted (Pilger, 1984; OConnor et al., 2007). Several ex-
amples of plume subduction have been proposed for ancient

http://dx.doi.org/10.1016/j.gsf.2014.07.002
1674-9871/ 2014, China University of Geosciences (Beijing) and Peking University. Production and hosting by Elsevier B.V. All rights reserved.
50 P.G. Betts et al. / Geoscience Frontiers 6 (2015) 49e59
convergent margins (Murphy et al., 1998, 2003; Dalziel et al., 2000;
Murphy and Keppie, 2005; Betts et al., 2009). These examples show
how a plume interacting with a convergent margin can cause the
arc to migrate towards the interior of the overriding plate and result
in the transition from arc- to plume-related magmatism in the
overriding plate (Murphy et al., 1998; Dalziel et al., 2000; Betts
et al., 2009). The latter suggests plume material eventually trans-
fers from the downgoing oceanic plate to the overriding plate,
although the mechanism responsible for this is not fully
understood.
1.1. Accretionary orogenic systems of Central and Eastern Asia
The Phanerozoic geological record of convergent margins is
dominated by long-lived accretionary orogenic systems charac-
terised by signicant crustal growth via lateral accretion of exotic
terranes from the oceanic realm (Cawood et al., 2009). Late Neo-
proterozoic to Phanerozoic crustal growth in Asia is preserved in
two long-lived accretionary orogenic systems: the Central Asian
Orogen and the East Asian Orogen (Fig. 1). The Central Asian Orogen
or Altaids (Windley et al., 2007; Wilhem et al., 2012) records
Neoproterozoic to Permian accretion along the southern margin of
Eurasia (Safonova et al., 2009; Safonova and Santosh, 2014) (Fig. 1)
during the closure of the Paleo-Asian Ocean. The East Asian Orogen
records Devonian to present-day accretion along eastern Asia
(Zhang et al., 2008; Safonova and Santosh, 2014) (Fig. 1). The
geological record of these orogens is complex and involves the
accretion of micro-continental fragments, arc terranes (Safonova
and Santosh, 2014), ocean plateaus and seamounts, as well as the
development of terrane-scale oroclines (Windley et al., 2007;
Wilhem et al., 2012).
The recent synthesis of Ocean Island Basalts (OIB) in the Central
Asian and East Asian orogens by Safonova and Santosh (2014) show
35 occurrences OIB terranes entrained within these orogenic sys-
tems (Fig. 1), suggesting they are signicant contributors to crustal
growth along these orogenic belts. Several of the mapped OIB ter-
ranes are suggested to have formed during super plume activity (Li
and Zhong, 2009). For example, Neoproterozoic OIB rocks within
the Central Asian Orogen are associated with the super plume
linked with Rodinia break-up (Safonova et al., 2009; Safonova and
Santosh, 2014). Cretaceous OIB rocks of the East Asian Orogen are
associated with the Pacic super plume (Safonova et al., 2009;
Safonova and Santosh, 2014). These OIB terranes are interpreted
as far-travelled oceanic plateaus. However, in the central Asian
Orogen, there are several Devonian to Permian ophiolite complexes
characterised by diverse geochemical associations that include
calc-alkaline, OIB, MORB, bonnanite and adakite terranes (e.g.,
Wong et al., 2010; Yang et al., 2012). Several of these terrane contain
ophiolites that are similar in age to, or slightly older than, their
inferred timing of accretion.
1.2. Subduction modelling
Flat slab subduction occurs when the subducting lithosphere
dip is shallow (<30 ) beneath the overriding plate. Shallow
subduction can extend for hundreds to a few thousand kilo-
metres beneath the overriding plate. Numerical models have
illustrated several phenomena that can inuence the dip of a
subduction including: subduction driven ow in the mantle
wedge or suction (Stevenson and Turner, 1977); corner ow in
the mantle wedge (Billen and Hirth, 2007); temperature of the
overriding plate (Rodrguez-Gonzlez et al., 2012); the thickness
and trench-ward motion of the overriding plate (Manea et al.,
2012); variation in the slab thickness as a function of age
(Molnar and Atwater, 1978; Capitanio et al., 2011); and the
presence of buoyancy anomalies entrained in the subducting slab
(e.g. Gutscher et al., 2000; van Hunen et al., 2002). Cloos (1993)
explored the role these buoyancy anomalies played on the dy-
namics of the subducting plate and showed the size and density
of the anomaly determined if subduction locked-up or became
entrained within the subducting slab, causing localised and
transient isostatic uplift.
Two-dimensional numerical simulations of accretionary pro-
cesses have focussed on different aspects of subduction and ac-
cretion of buoyant anomalies at convergent margins. Arrial and
Billen (2013) illustrated that eclogitization of an ocean plateau
inhibits accretion by increasing slab negative buoyancy, although
in some dynamic models, slab break-off leads to vertical accre-
tion of the buoyant plateau beneath the overriding plate.
Tetreault and Buiter (2012) suggested that buoyancy alone was
not a barrier to subduction and showed accretion of a buoyant
ocean plateau was partially dependent on the development of a
basal detachment fault and the depth of this detachment. van
Hunen et al. (2002) showed that collision with buoyant conti-
nental masses in a subduction zone leads to slab break-off and
stalling of convergence. More recently, Vogt and Gerya (2014)
proposed three modes of oceanic plateau accretion. These
include scenarios where plateau subduction occurs resulting in
forearc uplift; frontal plateau accretion where the plateau docks
onto the continental margin, resulting in intense deformation in
the overriding plate and slab break-off and; basal plateau ac-
cretion where the oceanic plateau is scraped from the subducting
slab and a new subduction zone developing behind the accreted
terrane.
3D geodynamic modelling of subducting slabs have provided
insight into the 3D slab geometries (e.g., Stegman et al., 2006, 2010;
van Hunen and Allen, 2011; Li et al., 2013), the motion of a trench
along a convergent margin (Schellart et al., 2007; Capitanio et al.,
2011), mantle dynamics (OzBench et al., 2008; Schellart, 2008),
and more recently the dynamics of an overriding plate (Yamato
et al., 2009; Capitanio et al., 2011; Capitanio and Replumaz, 2013;
Magni et al., 2014; Moresi et al., 2014). Recent 3D numerical
modelling with Underworld (Moresi et al., 2007), which we use in
the models presented here, have addressed accretion of entrained
buoyant material in 3D slab subduction models. Mason et al. (2010)
undertook 3D viscoplastic subduction modelling with a buoyant
plateau entrained in the oceanic lithosphere, showing the potential
for large plateaus to cause signicant modication to trench ge-
ometry and creation of a large slab window. Betts et al. (2012)
showed how plume-heads, and their associated buoyancy,
interact with a convergent margin markedly modify trench ge-
ometry, change trench velocity, shorten the overriding plate, as
well damage the subducting slab by creating slab windows. Moresi
et al. (2014) included an overriding plate in their models and un-
dertook 3D simulations of buoyant micro-continental ribbon ac-
cretion and showed that subduction re-established behind the
accreted ribbon via margin parallel migration of a highly arcuate
trench and as a consequence the ribbon was embedded into the
overriding plate.
In this paper we present the results of 3D numerical simula-
tions of thick oceanic plateau crust and the buoyant plume ma-
terial to examine how these exotic oceanic terranes contribute to
lateral continental growth of long-lived convergent margins. Our
models illustrate how accretion of oceanic plateaus and associated
plume material at convergent margins inuence trench geometry
and architecture of the subducting slab. The model results will
serve as a geodynamic context to interpret complex convergent
margin interaction characterised by ophiolite complexes with
diverse and mixed geochemical responses in long-lived accre-
tionary belts.
 P.G. Betts et al. / Geoscience Frontiers 6 (2015) 49e59
Figure 1. General tectonic map showing the location and ages of OIB-hosting accretionary complexes of Central and East Asia Orogens. The map is modied after Safonova and Santosh (2014).

51
52 P.G. Betts et al. / Geoscience Frontiers 6 (2015) 49e59

2. Model description
The numerical models in this study were designed to assess the
behaviour of the subducting slab, trench geometry, overriding
plate, and the mechanism for subduction re-initiation when a small
oceanic plateau is accreted at the edge of a convergent margin.
The numerical implementation is the particle-in-cell nite
element code Underworld (Moresi et al., 2007) in which material
properties are carried by Lagrangian integration points and map-
ped to element properties through a local, discrete Voronoi, inte-
gration scheme (Veli c et al., 2008). Models are conned to a 3D
Cartesian domain (Fig. 2) 800 km deep, 8000 km  3000 km hor-
izontally represented with a uniformly spaced grid of 256  96  96
elements, in which we solve the incompressible, isothermal Stokes
equation that balances buoyancy forces with viscous stresses and
pressure gradients. The slab motion is entirely driven by its nega-
tive buoyancy and is initiated by a small perturbation in the ge-
ometry as shown in Fig. 2.
Our models build upon Model 80w of Moresi et al. (2014),
substituting a plume-head and/or a plateau for the continental
ribbon material. The overriding plate parameterisation is identical
to that study. Temperature is mapped to density variations as a
persistent property during the simulation. The oceanic plate is of
uniform thickness (100 km) and 7000 km long. A 125 km radius,
30 km deep cylinder of weaker (constant viscosity of 200 times the
reference value), material with the density of basalt is embedded
into the oceanic lithosphere to represent an oceanic plateau being
advected towards the convergent margin. In the base of the oceanic
lithosphere a 125 km radius, 75 km thick cylinder of buoyant, weak
material represents a spreading-out plume head that replaces the
lowermost part of the oceanic boundary layer and extends a further
50 km into the asthenosphere. The asthenosphere viscosity is set to
3 times the reference value at the base of the lithosphere to account
for the modest increase with pressure over depth in the upper
mantle. The overriding plate has three components: (1) a region of
weaker back-arc crust and mantle lithosphere, (2) a strong, cratonic

Figure 2. (a) Model setup showing the initial distribution of distinct material domains which represent the variations in density and strength which persist during the evolution of
the model. The rheological layering is as described in Moresi et al. (2014). The density and strength structure of the subducting plate is computed assuming a half-space cooling
model to map lithospheric age to temperature and then averaging properties within 25 km thick layers: (b) yield strength in each layer of 80 Myr old ocean based on the lithostatic
pressure gradient, (c) density of normal oceanic lithosphere with a 7 km basaltic crust, (d) relative viscosity of normal oceanic lithosphere assuming the wet-olivine diffusion
creep mechanism of Karato and Wu (1993) for all materials, (e) density prole and (f) viscosity prole through a 20 km thick plateau and a 50 km thick plume head at depth 75 km.
In the viscosity plots, the geotherm is also shown (green line) and viscosity is truncated between 102 and 105.
 P.G. Betts et al. / Geoscience Frontiers 6 (2015) 49e59 53

backstop, and (3) a region of transitional crust and mantle litho-
sphere which lies between the two (Fig. 2). The properties of the
oceanic lithosphere, plateau and plume can be calculated using the
included python script (CalculateModelProperties.py) which was
used to plot Fig. 2bef.
3. Results
We describe three end-member 3D numerical experiments. The
rst experiment (Model 1) has a plume head at the base of a sub-
ducting slab and simulates a situation where a plateau or seamount
has not yet been created. The second model (Model 2) has buoyant
material embedded into the oceanic crust to simulate an older
plateau with no plume material interacting with the slab. This
experiment simulates situations where the age of plateau is sub-
stantially older than the timing of plateau accretion. The third
experiment (Model 3) has a plateau embedded into the oceanic
crust and a large plume head beneath the plateau. This experiment
simulates situations where the plateau age is similar to, or slightly
older than, the timing of accretion. Our simulations suggest that the
plume material beneath the oceanic plateau markedly inuences
the geometry of the subducting slab and the trench. Thermal
weakening by the plume head, or inherent slab heterogeneities
such as transform faults and spreading ridges are not considered in
our models. Movies for each of the simulations can be found in the
supplementary material.
3.1. Model 1: plume only model
Model 1 initially maintains a stable convergent margin charac-
terised by slow trench retreat (trench movement towards the
downgoing plate) (Fig. 3). The dip of the slab remains relatively
constant along the length of the margin (Fig. 4) (see movie 1:
supplementary material). As plume material subducts its positive
buoyancy causes the downgoing slab to shallow and at subduction
ensues (Fig. 5def). This buoyancy effect is sufcient to under-plate

Figure 3. (a) The velocity of the subducting plate (measured at point (1) shown in sub-gure b); rollback of the trench (at z 0, point (2) in b) and the collisional suture (at z zmax,
point (3) in b) for each of the four models from the initiation of the model to a scaled run time of 85 Myr. (c) Comparison of the plate velocity in the reference frame of the
computational domain to the net subduction rate at either end of the trench. Roll back of the slab contributes to higher ux of subducted material than results from the plate
velocity alone. During collision, the trench advances at roughly the same rate as the plate velocity producing a period from 45 to 60 Myr with no subduction in that region. Variation
in the subduction ux is possible along a margin even when the plate velocity is constant due to different rollback rates, but this is necessarily accompanied by a change in shape.
54 P.G. Betts et al. / Geoscience Frontiers 6 (2015) 49e59
Figure 4. Plan view of each of the three models at 80 Myr. (a) Model 1: Plume only. (b)
Model 2: Plateau only. (c) Model 3: Plume plus plateau. The location of the active
trench is indicated by a red line. The overriding plate is shown in blue with a con-
trasting blue grid that visualizes strain. The continental backstop is the darker region
in the overriding plate. The subducting plate is grey with a yellow strain grid. The
plateau is marked as a small orange region in the incoming plate. The blue dotted lines
in (c) represent the position of the slab window beneath the overriding plate. Both
plates are rendered transparent allowing the location of the plume material to show as
a broad red halo around the plateau.
the subducting slab beneath the lithosphere of the overriding plate,
consistent with previous models of at subduction (e.g. van Hunen
et al., 2002). Near the centre of the plume the mantle wedge is
completely removed and at subduction extends for approximately
500 km beneath the overriding plate at approximately 80 million
years (Fig. 5f). As the effects of the plume buoyancy dissipate along
the convergent margin the dip of the slab steepens away from the
impinging plume head, resulting in a bow-shaped subducting slab
(Fig. 4a). The trench remains in a stable, linear position (Fig. 4).
Traction between the slab and the overriding plate, where the
plume impinges on the subducting slab, results in lithospheric
thickening and shortening, and a slight decrease in the rate of
convergence at 60 million years. There is signicant stretching of
the slab at the leading edge of at subduction (see movie 1: sup-
plementary material). This is attributed to negative buoyancy pull
forces in parts of the slab not affected by the plume material and
the minor decrease in convergence as a result of traction between
the at slab and the overriding plate. Slab stretching is not suf-
cient to create a slab window (Fig. 5f), although at the end of model
a small volume of plume material penetrates the resting slab and
bursts into the distal parts of the mantle wedge (see movie 1:
supplementary material). However, most of the plume material
beneath the subducting plate is recycled into the mantle (Fig. 5f).
Supplementary data related to this article can be found at http://
dx.doi.org/10.1016/j.gsf.2014.07.002.
3.2. Model 2: plateau only model
Model 2 maintains a stable convergent margin where the trench
retreats until the oceanic plateau collides with the trench (Fig. 5a
and Fig. 6) (also see movie 2: supplementary material). At the point
of plateau accretion, the trench advances at a rate of less than 1 cm/
yr (Fig. 3), the plateau is shortened by approximately w50% (Fig. 6),
overriding plate thickens above the mantle wedge, and a short
interval of double-sided subduction ensues (Fig. 6). During accre-
tion, some of the plateau base is entrained in the subducting slab
(Fig. 6). In this model a slab window does not form. Following
collision, the trench maintains a linear geometry, the slab dip re-
mains constant, and the trench neither advances nor retreats at
positions laterally away from the plateau (Fig. 5). However, at the
point of accretion the trench migrates around the side of the
plateau forming an arcuate-shaped trench (Fig. 5b and c). Stable
subduction is re-established immediately behind the accreted
plateau once the trench retreats behind its trailing edge (Fig. 5c and
Fig. 6) (see movie 2: supplementary material). Once subduction is
re-established, roll-back of the trench resumes along the entire
length of the convergent margin (see movie 2: supplementary
material). Re-establishment of stable subduction is accommodated
by the development of a detachment fault at the base of the plateau
allowing the plateau to be stranded at the leading edge of the
overriding plate (Fig. 6).
Supplementary data related to this article can be found at http://
dx.doi.org/10.1016/j.gsf.2014.07.002.
3.3. Model 3: combination of plume and plateau
Model 3 initially maintains a stable convergent margin (see
movies 3 and 4: supplementary material), which is characterised by
slow trench retreat (<2 cm/yr) and consistent slab dip (Fig. 3). The
plume head makes contact with the subduction zone before
plateau accretion occurs (w40 million years) (Fig. 7a). The positive
buoyancy effect of the plume head beneath the subducting slab
decreases the slab dip (Fig. 7a), although prior to plateau accretion,
the plate and trench velocity is similar to Models 1 and 2 (Fig. 3).
Accretion of the plateau results in a period of instability along the
 P.G. Betts et al. / Geoscience Frontiers 6 (2015) 49e59 55

Figure 5. (aec) Oblique view of the Model 2: (a) Model 2 at 47 million years at the onset of plateau accretion; (b) Model 2 at 65 million years showing the trench wrapping around
the edge of the plateau and shearing at the base of the plateau as subduction re-initiates behind the plateau; (c) Model 2 at 90 million years showing subduction re-initiation
immediately behind the plateau. (def) Oblique view of Model 1: (d) Model 1 at 41 million years shortly after plume material begins to subduct; (e) Model 1 at 63 million years
showing early stages of at subduction caused by positive buoyancy of the plume material; (f) Model 1 at 82 million years showing at slab subduction. In all models the dark red
dots show the position of the trench and green dots show the top of the subducting slab. Brown dots represent plateau material and light red dots represent plume material.

convergent margin (Fig. 7b and c). Convergence is accommodated
by shearing at the base of the plateau and shortening of the over-
riding plate rather than by subduction alone. Accretion of the
plateau causes the local subduction rate to slow at 30 million years
to less than 0.5 cm/yr at 50 million years (Fig. 3c). Accretion of the
plateau triggers extreme, localised trench advance in front of the
plateau (>2 cm/yr) for approximately 15 million years until the
trench velocity gradually decreases (Fig. 3a). Trench advance in
front of the plateau creates a localised highly arcuate trench ge-
ometry as it wraps around the plateau edges (Figs. 4 and 7). Similar
to the models by Moresi et al. (2014), an orocline, as dened by a
highly arcuate trench, forms as the trench attempts to migrate
parallel with the convergent margin behind the plateau while the
direction of convergence remains the same (Fig. 7def). The
remainder of the subducting slab maintains a stable linear geom-
etry (Figs. 4 and 7) and this large segment of the trench retreats at a
relatively slow rate between < 0.8 cm/yr (Fig. 3a). Trench advance
in front of the plateau is accompanied by signicant thickening of
the overriding plate and subduction erosion (double-sided sub-
duction) (See movies 3 and 4 in the supplementary les). The ac-
cretion of the plateau results in a marginal decrease in plate
velocity compared with Model 2 (Fig. 3a).
When subduction re-establishes behind the plateau at 105
million years, the plate velocity increases (see supplementary
movies 3 and 4) and the plateau becomes imbedded several
hundred kilometres into the interior of the overriding plate
(Fig. 7f). The buoyant plume head is trapped beneath the sub-
ducting slab which maintains the relatively shallow dip and pre-
venting re-establishment of a linear trench behind the plateau as
shown in Model 2 (Fig. 7cef). In Model 3, the combined positive
buoyancy of the plume and plateau increases the coupling be-
tween the subducting slab and the overriding plate. A slab window
is created as the negatively buoyant part of the downgoing slab in
front of the plateau continues to subduct, resulting in extreme slab
stretching and delamination. The slab window is sufciently large
at w67 million years (Fig. 7c) to evolve into a conduit for buoyant
plume material residing beneath the subducting slab to burst
through and interact with the mantle wedge and overriding plate
(Fig. 7def). Plume material continues to ow into the mantle
wedge until w113 million years, when the subduction zone is re-
established. Then the remaining plume material is subducted
with the downgoing slab. This model is consistent with specula-
tions by Betts et al. (2012), in which they infer slab windows allow
plume material to transfer from the downgoing plate to the
overriding plate.
Supplementary data related to this article can be found at http://
dx.doi.org/10.1016/j.gsf.2014.07.002.
4. Discussion
The ancient geological record of long-lived accretionary
orogenic belts such as the Central Asian Orogen shows evidence for
repeated episodes of OIB oceanic crust accretion (Safonova and
Santosh, 2014). For example, the Central Asian Orogen has
numerous ophiolite terranes that are considerably older than the
estimated time of their accretion (Fig. 1). Therefore, these OIB ter-
ranes are unlikely to have their causative plumes residing beneath
them at the time of accretion (Model 2). However, other OIB
56 P.G. Betts et al. / Geoscience Frontiers 6 (2015) 49e59

Figure 6. Cross sections of Model 2 where the plateau accretes onto the overriding plate in the absence of plume buoyancy beneath the downgoing plate. The green line indicates
the top of the subducting plate. The plateau is highlighted with orange shading. (a) Model 2 at 35 million years before the plateau has interacted with the trench. (b) Model 2 at 45
million years following plateau accretion. The plateau becomes partially subducted and material is entrained by the slab (see movie 2: Supplementary material) and is drawn into a
cylindrical tendril. The remainder of the plateau has accreted to the edge of the overriding plate and has shortened approximately 50%. (c) Model 2 at 55 million years. The plateau
and overriding plate have shortened and thickened to accommodate convergence. The subducted slab begins to stretch and attenuate. Abrasion of the overriding plate has resulted
in double-sided subduction, which is indicated by the coupling of overriding plate lithosphere with the subducting slab. (d) Model 2 at 65 million years. Double-sided subduction
continues and the slab continues to stretch and attenuate. (e) Model 2 at 70 million years. The plateau has started to disaggregate and the slab is attenuated to the extent that the
upper crust (green line) has also been disaggregated. Subduction has begun to re-establish behind the plateau and detachment has formed underneath the plateau. (f) Model 2 at 80
million years. Subduction has completely re-established behind the plateau, which is now part of the overriding plate. The plateau entrained in the subducting slab has
disaggregated.

provinces appear to have been incorporated within the accre-
tionary orogen shortly after they were emplaced based on their
ages and time of accretion (Fig 1). In these cases, the OIB terranes
may have accreted when a plume was residing beneath them
(Model 3). A third scenario, where a plume interacts with the
convergent margin (Model 1), will have a more cryptic signal in the
accretionary margin geological record depending on whether or
not a slab window opens and plume material is emplaced in the
overriding plate.
We have investigated these three scenarios to assess the varia-
tions in the geodynamic and geological response of OIB plateau
accretion and plume-modied orogenesis (Murphy et al., 1998).
Models 2 and 3 show that accretion creates a relatively localised
zone of intense crustal deformation characterised by trench
advance, arcuate trench geometry, transfer of parts of the plateau
from the downgoing plate to the overriding plate, and eventually
the re-establishment of subduction behind the accreted plateau
(Fig. 4). Subduction along the rest of the active convergent margin
 P.G. Betts et al. / Geoscience Frontiers 6 (2015) 49e59 57

Figure 7. Oblique view of the Model 3 with the same colour key as Fig. 4. In addition, the green dots highlight the location of the edge of the subducting plate. (a) Model 3 at 40
million years, immediately before the plateau accretes. The positive buoyancy plume material in front of the plateau has decreased slab dip. (b) Model 3 at 50 million years showing
the onset of plateau accretion. (c) Model 3 at 67 million years. Rapid trench advance occurs in front of the plateau and the trench migrates behind the plateau. The slab window has
begun to open. (d) Model 3 at 82 million years. The plume material is transferred through the slab window to the mantle wedge. Shearing beneath the plateau occurs as subduction
begins to re-establish behind the plateau. (e) Model 3 at 105 million years. Subduction has re-established behind the plateau and the plateau is embedded into the overriding plate.
(f) Model 3 at 113 million years. After subduction has re-established behind the accreted plateau, some of the plume material has transferred into the mantle wedge whereas the
plume material behind the plateau is trapped beneath the subducting slab. The black arrows represent the trajectories of plume material ow.

maintains a stable conguration and appears to be relatively un-
affected by plateau accretion (Fig. 4).
The presence of plume material beneath the plateau changes the
behaviour of accretion signicantly. In Model 2 accretion of a small
oceanic plateau is relatively simple. Plateau accretion at the edge of
the overriding plate (Figs. 4 and 5) causes modest trench advance in
front of the plateau (<1 cm/yr) (Fig. 3), which is insufcient to drive
a major episode of orogenesis in the overriding plate (Schellart
2008). The transition from stable subduction before and after ac-
cretion is characterised by migration of the trench around the side
of the plateau before stable subduction is restored immediately
behind the plateau. Subduction re-initiation is accommodated by
the development of a detachment shear zone (Fig. 5) at the base of
the plateau, which has been observed in other 2D numerical
models (e.g. Tetreault and Buiter, 2012). Model 3 illustrates the
buoyancy effects of plume material beneath the downgoing slab
and how the accretion characteristics change markedly. The plume
material beneath the subducting slab promotes at slab subduc-
tion, which prevents the plateau from subducting and drives the
plateau towards the interior of the overriding plate resulting in
relatively rapid trench advance in front of the plateau. The rate of
trench advance is sufcient to drive localised orogenesis and crustal
shortening in the overriding plate (Fig. 3). Compared with Model 2,
most of the plateau is accreted to the overriding plate. The transi-
tion between stable subduction before and after accretion is also
more complex. The trench migrates around the side of the plateau
and begins to retreat parallel to the convergent margin behind the
trailing edge of the plateau before stalling and creating an orocline
in front of the re-established trench (Fig. 4). Margin parallel trench
retreat has been simulated behind an accreted buoyant continental
ribbon (Moresi et al., 2014), however, unlike in this model, trench
retreat did not pause. In our Model 3 the buoyancy effects of the
plume material prevent the margin parallel retreat behind the
entire length of the accreted plateau.
Previous 3D numerical models suggest that slab windows can
open in response to plume head interaction with a subduction zone
(Betts et al., 2012). Our models show that subduction of plume
material has little impact on the integrity of the subducting slab.
58 P.G. Betts et al. / Geoscience Frontiers 6 (2015) 49e59
Rather, the buoyancy of the plume material causes localised at
subduction that transitions back to a normal dip along the length
of the convergent margin (Fig. 4). However, a slab window is
created when plume material resides beneath the plateau at the
time of accretion. This scenario is likely to occur when a plume
interacts with ocean lithosphere proximal to the convergent
margin such as that occurring in the present day Carnegie and Coco
ridges (Herner et al., 2003). Betts et al. (2012) speculated that the
transfer of a plume from the downgoing plate to the overriding
plate occurs via a slab window created when the subduction stalls
or slows in response to the plume head interacting with a
convergent margin. Our models also suggest that plume material
transfers from the downgoing plate to the overriding plate through
slab windows (Fig. 6). These results conrm the hypothesis that
plume-modied orogenesis, as proposed by Murphy et al. (1998),
requires that accretion of a plateau and plume collision occurs
nearly simultaneously at the trench. This scenario is most likely to
occur when the plateau age is similar to the timing of the plateau
accretion.
Plume-modied subduction models also have implications for
the magmatic evolution of the accretionary orogens. Flat slab
subduction caused by plume material residing beneath the down-
going slab should result in arc migration towards the interior of the
overriding plate, forming a curvilinear arc along the length of the
convergent margin. Adakite magmatism (Defant and Drummond,
1990) is associated with opening of slab windows, which allows
upper mantle beneath the subducting plate to interact with the
mantle wedge (Abratis and Wrner, 2001; Gorring and Kay, 2001)
or by melting of slab edge (Falloon et al., 2008). Previous in-
terpretations for adakite genesis have proposed that slab windows
form by either ridge subduction (Abratis and Wrner, 2001) or at
subduction-transform intersections (Falloon et al., 2008). Our
modelling shows that plume and plateau accretion creates a slab
window, and therefore has the potential to create adakite melts
along with calc-alkaline and boninite melts typical of convergent
margins (e.g. Lee and Lim, 2014).
In the ancient geological record there are numerous examples of
ophiolite complexes in accretionary orogens with diverse
geochemical signatures. For example the Central Asian Orogen is
characterised by Devonian to Permian ophiolite complexes that
display diverse geochemical associations characterised by calc-
alkaline, OIB, MORB, boninite and adakite magmas (e.g. Wong
et al., 2010; Yang et al., 2012). Specic examples include the Sar-
tuohai ophiolitic mlange (Fig. 1) with its plume-type geochemical
afnity, which was accreted onto eastern West Junggar Craton
during closure of the Junggar Ocean (Yang et al., 2012). Other
ophiolite mlanges have been interpreted in the context of ridge
subduction and slab windowing because of their adakite signatures
(Zhang et al., 2008; Wong et al., 2010). Although such complex
associations could also be from subduction of a plume head
impinging at the base of subducting oceanic lithosphere (e.g.,
Wyman and Kerrich, 2009; Gazel et al., 2011; Betts et al., 2012).
Other examples, show that ultra-mac and mac rocks are char-
acterised by mixed subduction and plume geochemical signatures
such as the Xiangshanzhong maceultramac intrusion in eastern
Tianshan (Tang et al., 2013) (Fig. 1). The interpreted tectonic setting
for this ophiolite is complex and may be explained by a plume
interacting and modifying a congested subduction zone.
Finally, our models illustrate that complex geochemical associa-
tions for magmas (e.g. with plate margin, adakitic, and plume-related
signatures) can form in relatively simple convergent margin settings
that are transiently and locally disrupted by plateau accretionary
events. They do not require a complicated subduction zone setting
involving multiple subduction zones or rapid switches in subduction
polarity as are often depicted in the geological literature.
5. Conclusions
We have modelled three end member subduction scenarios in
three-dimensions to explore the inuence of a plateau and a plume
head beneath the converging plate and the resulting change in slab
dynamics. Plume-only subduction results in at slab morphology
creating a subduction zone with a bowed geometry. Flat slab
subduction results in removal of the mantle wedge above the
centre of the plume, which continues along the base of the over-
riding plate lithosphere up to 500 km from the trench. Plateau
collision without an underlying plume results in docking or ac-
cretion of the plateau at the edge of the overriding plate. The
overriding plate and the plateau are then shortened and a portion
of the plateau subducts with the downgoing slab. Subduction re-
establishes immediately behind the plateau and is accommodated
by development of zones of intense shearing at the base of the
plateau. A localised arcuate-shaped trench is created in response to
the trench migration around the edge of the plateau. Away from the
region of accretion, the trench geometry remains stable and the
downgoing slab geometry does not change. A third model with
plume material directly beneath a plateau profoundly inuences
the geometry of the convergent margin during accretion. Upon
plateau accretion, the trench in front of the plateau advances
rapidly. Convergence is then accommodated by intense shortening
in the overriding plate and shearing at the base of the plateau. The
trench migrates around the edge of the plateau and then begins to
roll-back parallel to the plate margin behind the plateau before
stalling. Subduction re-establishes behind the trailing edge of the
plateau, effectively embedding the plateau within the overriding
plate. The buoyancy of the plume precludes signicant recycling of
the plateau with the subducted slab. In this model a large slab
window is created forming a conduit that allows plume material to
transfer from the downgoing slab into the mantle wedge beneath
the overriding plate. Subduction away from the point of accretion
remains relatively stable. The modelling shows accretion of young
plateaus, with their plumes still residing beneath them, has the
greatest impact in modifying plate dynamics, deformation in the
overriding plate, and modifying the trench and slab geometry. The
latter model may also provide an explanation as to how plume
material can transfer from the subducting plate into the overriding
plate.
Acknowledgements
Peter Betts acknowledges Monash Research Acceleration Pro-
gram, which funded part of the research. Louis Moresi acknowl-
edges support from the Australian Research Councils Discovery
Projects funding scheme (projects Nos. DP130101946 and
DP110101697) and use of the NCI National Facility in Canberra,
Australia, which is supported by the Australian Commonwealth
Government. Meghan Miller acknowledges funding from NSF-
CAREER award EAR-1054638. David Willis holds an Australian
Postgraduate Award scholarship (APA).
Appendix A. Supplementary data
Supplementary data related to this article can be found online at
http://dx.doi.org/10.1016/j.gsf.2014.07.002.
References
Abratis, M., Wrner, G., 2001. Ridge collision, slab-window formation, and the ux
of Pacic asthenosphere into the Caribbean realm. Geology 29, 127e130.
Arrial, P.-A., Billen, M.I., 2013. Inuence of geometry and eclogitization on oceanic
plateau subduction. Earth and Planetary Science Letters 363, 34e43.
 P.G. Betts et al. / Geoscience Frontiers 6 (2015) 49e59
Betts, P.G., Mason, W.G., Moresi, L., 2012. The inuence of a mantle plume head on
the dynamics of are treating subduction zone. Geology 40, 739e742.
Betts, P.G., Giles, D., Foden, J., Schaefer, B.F., Mark, G., Pankhurst, M.J., Forbes, C.J.,
Williams, H.A., Chalmers, N.C., Hills, Q., 2009. Mesoproterozoic plume-modied
orogenesis in eastern Precambrian Australia. Tectonics 28. http://dx.doi.org/
10.1029/2008TC002325. TC3006.
Billen, M.I., Hirth, G., 2007. Rheologic controls on slab dynamics. Geochemistry,
Geophysics, Geosystems 8 (8), Q08012.
Capitanio, F.A., Faccenna, C., Zlotnik, S., Stegman, D.R., 2011. Subduction dynamic
and the origin of Andean orogeny and the Bolivian orocline. Nature 480, 83e86.
Capitanio, F.A., Replumaz, A., 2013. Subduction and slab breakoff controls on Asian
indentation tectonics and Himalayan western syntaxis formation. Geochem-
istry, Geophysics, Geosystems 14, 3515e3531.
Cawood, P.A., Krner, A., Collins, W.J., Kusky, T.M., Mooney, W.D., Windley, B.F.,
2009. Accretionary orogens through Earth history. In: Cawood, P.A., Krner, A.
(Eds.), Earth Accretionary Systems in Space and Time, 318. Geological Society of
London Special Publications, pp. 1e36.
Cloos, M., 1993. Lithospheric buoyancy and collisional orogenesis: subduction of
oceanic plateaus, continental margins, island arcs, spreading ridges, and sea-
mounts. Geological Society of America Bulletin 105, 715e737.
Dalziel, I.W.D., Lawver, L.A., Murphy, J.B., 2000. Plumes, orogenesis, and super-
continental fragmentation. Earth and Planetary Science Letters 178, 1e11.
Defant, M.J., Drummond, M.S., 1990. Derivation of some modern arc magmas by
melting of young subducted lithosphere. Nature 347, 662e665.
Falloon, T.J., Danyushevsky, L.V., Crawford, A.J., Meffre, S., Woodhead, J.D.,
Bloomer, S.H., 2008. Boninites and adakites from the northern termination of
the Tonga Trench: implications for adakite petrogenesis. Journal of Petrology
49, 697e715.
Gazel, E., Hoernle, K., Carr, M.J., Herzberg, C., Saginor, I., den Bogaard, P. v., Hauff, F.,
Feigenson, M., Swisher III, C., 2011. Plumeesubduction interaction in southern
Central America: mantle upwelling and slab melting. Lithos 121, 117e134.
Gorring, M.L., Kay, S.M., 2001. Mantle processes and sources of neogene slab win-
dow magmas from Southern Patagonia, Argentina. Journal of Petrology 42,
1067e1094.
Gutscher, M.-A., Spakman, W., Bijwaard, H., Engdahl, E.R., 2000. Geodynamics of at
subduction: seismicity and tomographic constraints from the Andean margin.
Tectonics 19, 814e833.
Herner, R., Hoernle, K., Barckhausen, U., Hauff, F., 2003. Geodynamic evolution of
the Galpagos hot spot system (Central East Pacic) over the past 20 m.y.:
constraints from morphology, geochemistry, and magnetic anomalies.
Geochemistry, Geophysics, Geosystems 4, 1108.
Karato, S., Wu, P., 1993. Rheology of the upper mantle: a synthesis. Science 260,
771e778.
Lee, C., Lim, C., 2014. Short-term and localized plume-slab interaction explains the
genesis of Abukuma adakite in Northeastern Japan. Earth and Planetary Science
Letters 396, 116e124.
Li, Z.-H., Xu, Z., Gerya, T., Burg, J.-P., 2013. Collision of continental corner from 3-D
numerical modelling. Earth and Planetary Science Letters 380, 98e111.
Li, Z.-X., Zhong, S., 2009. Supercontinent-superplume coupling, true polar wander
and plume mobility: plate dominance in whole-mantle tectonics. Physics of the
Earth and Planetary Interiors 176, 143e156.
Magni, V., Faccenna, C., van Hunen, J., Funiciello, F., 2014. How collision triggers
backarc extension: insight into mediterranean style of extension from 3-
d numerical models. Geology 42, 511e514.
Manea, V.C., Prez-Gussiny, M., Manea, M., 2012. Chilean at slab subduction
controlled by overriding plate thickness and trench rollback. Geology 40,
35e38.
Mason, W.G., Moresi, L., Betts, P.G., Miller, M.S., 2010. Three-dimensional numerical
models of the inuence of a buoyant oceanic plateau on subduction zones.
Tectonophysics 483, 71e79.
Miller, M.S., Lee, C.-T.A., 2008. Possible chemical modication of oceanic lithosphere
by hotspot magmatism: seismic evidence from the junction of Ninety-east
Ridge and the Sumatra-Andaman arc. Earth and Planetary Science Letters
265, 386e395.
Miller, M.S., Kennett, B.L.N., Lister, G.S., 2004. Imaging changes in morphology,
geometry, and physical properties of the subducting Pacic plate along the
IzueBonineMariana arc. Earth and Planetary Science Letters 224, 363e370.
Miller, M.S., Gorbatov, A., Kennett, B.L.N., 2005. Heterogeneity within the sub-
ducting Pacic slab beneath the Izu-Bonin-Mariana arc: evidence from to-
mography using 3D ray-tracing inversion techniques. Earth and Planetary
Science Letters 235 (1e2), 331e342.
Molnar, P., Atwater, T., 1978. Interarc spreading and cordilleran tectonics as alter-
nates related to the age of subducted oceanic lithosphere. Earth and Planetary
Science Letters 41, 330e340.
Moresi, L., Quenette, S., Lemiale, V., Mriaux, C., Appelbe, B., Mhlhaus, H.B., 2007.
Computational approaches to studying non-linear dynamics of the crust and
mantle. Physics of the Earth and Planetary Interiors 163, 69e82.
Moresi, L.N., Betts, P.G., Miller, M.S., Cayley, R., 2014. Dynamics of continental ac-
cretion. Nature 508, 245e248.
Murphy, B.J., Oppliger, G.L., Brimhall Jr., G.H., Hynes, A., 1998. Plume-modied
orogeny: an example from the western United States. Geology 26, 731e734.
59
Murphy, J.B., Hynes, A.J., Johnston, S.T., Keppie, J.D., 2003. Reconstructing the
ancestral Yellowstone plume from accreted seamounts and its relationship to
at-slab subduction. Tectonophysics 365, 185e194.
Murphy, J.B., Keppie, J.D., 2005. The Acadian orogeny in the northern Appalachians.
International Geology Review 47, 663e687.
OConnor, J.M., Stoffers, P., Wijbrans, J.R., Worthington, T.J., 2007. Migration of
widespread long-lived volcanism across the Galpagos Volcanic Province: evi-
dence for a broad hotspot melting anomaly? Earth and Planetary Science Let-
ters 263, 339e354.
OzBench, M., Regenauer-Lieb, K., Stegman, D.R., Morra, G., Farrington, R., Hale, A.,
May, D.A., Freeman, J., Bourgouin, L., Mhlhaus, H., Moresi, L., 2008. A model
comparison study of large-scale mantle-lithosphere dynamics driven by sub-
duction. Physics of the Earth and Planetary Interiors 171, 224e234.
Pilger, R.H., 1984. Cenozoic plate kinematics, subduction and magmatism: South
American Andes. Geological Society of London Journal 141, 794e802.
Rodrguez-Gonzlez, J., Negredo, A.M., Billen, M.I., 2012. The role of the overriding
plate thermal state on slab dip variability and on the occurrence of at sub-
duction. Geochemistry, Geophysics, Geosystems 13, Q01002.
Safonova, I.Y., Santosh, M., 2014. Accretionary complexes in the Asia-Pacic region:
tracing archives of ocean plate stratigraphy and tracking mantle plumes.
Gondwana Research 25, 126e158.
Safonova, I.Y., Utsunomiya, A., Kojima, S., Nakae, S., Tomurtogoo, O., Filippov, A.N.,
Koizumi, K., 2009. Pacic superplume-related oceanic basalts hosted by
accretionary complexes of Central Asia, Russian Far East and Japan. Gondwana
Research 16, 587e608.
Schellart, W.P., 2008. Overriding plate shortening and extension above subduction
zones: a parametric study to explain formation of the Andes Mountains.
Geological Society of America Bulletin 120, 1441e1454.
Schellart, W.P., Freeman, J., Stegman, D.R., Moresi, L., May, D., 2007. Evolution and
diversity of subduction zones controlled by slab width. Nature 446, 308e311.
Stegman, D.R., Freeman, J., Schellart, W.P., Moresi, L., May, D., 2006. Inuence of
trench width on subduction hinge retreat rates in 3-D models of slab rollback.
Geochemistry, Geophysics, Geosystems 7. http://dx.doi.org/10.1029/
2005GC001056).
Stegman, D.R., Farrington, R., Capitanio, F.A., Schellart, W.P., 2010. A regime diagram
for subduction styles from 3-D numerical models of free subduction. Tectono-
physics 483, 29e45.
Stevenson, D.J., Turner, J.S., 1977. Angle of subduction. Nature 270 (5635), 334e336.
Tang, D., Qin, K., Su, B., Sakyi, P.A., Liu, Y., Mao, Q., Santosh, M., Ma, Y., 2013. Magma
source and tectonics of the Xiangshanzhong mac-ultramac intrusion in the
Central Asian Orogenic Belt, NW China, traced from geochemical and isotopic
signatures. Lithos 170e171, 144e163.
Tetreault, J.L., Buiter, S.J.H., 2012. Geodynamic models of terrane accretion: testing
the fate of island arcs, oceanic plateaus, and continental fragments in subduc-
tion zones. Journal of Geophysical Research 117, B08403.
van Hunen, J., Van Den Berg, A.P., Vlaar, N.J., 2002. On the role of subducting oceanic
plateaus in the development of shallow at subduction. Tectonophysics 352,
317e333.
van Hunen, J., Allen, M.B., 2011. Continental collision and slab break-off: a com-
parison of 3-D numerical models with observations. Earth and Planetary
Science Letters 302 (1e2), 27e37.
Veli
c, M., May, D., Moresi, L., 2008. A fast robust algorithm for computing discrete
voronoi diagrams. Journal of Mathematical Modelling and Algorithms 8,
343e355.
Vogt, P.R., 1973. Subduction and aseismic ridges. Nature 241 (5386), 189e191.
Vogt, K., Gerya, T.V., 2014. From oceanic plateaus to allochthonous terranes:
numerical modelling. Gondwana Research 25, 494e508.
Wilhem, C., Windley, B.F., Stampi, G.M., 2012. The Altaids of Central Asia: a tec-
tonic and evolutionary innovative review. Earth-Science Reviews 113, 303e341.
Windley, B.F., Alexeiev, D., Xiao, W., Krner, A., Badarch, G., 2007. Tectonic models
for accretion of the Central Asian Orogenic Belt. Journal of the Geological So-
ciety 164, 31e47.
Wong, K., Sun, M., Zhao, G., Yuan, C., Xiao, W., 2010. Geochemical and geochrono-
logical studies of the AlegedayiOphiolitic Complex and its implication for the
evolution of the Chinese Altai. Gondwana Research 18, 438e454.
Wyman, D., Kerrich, R., 2009. Plume and arc magmatism in the Abitibi subprovince:
implications for the origin of Archean continental lithospheric mantle. Pre-
cambrian Research 168, 4e22.
Yamato, P., Husson, L., Braun, J., Loiselet, C., Thieulot, C., 2009. Inuence of sur-
rounding plates on 3D subduction dynamics. Geophysical Research Letters 36,
L07303.
Yang, G., Li, Y., Santosh, M., Yang, B., Yan, J., Zhang, B., Tong, L., 2012. Geochronology
and geochemistry of basaltic rocks from the Sartuohaiophiolitic mlange, NW
China: implications for a Devonian mantle plume within the Junggar Ocean.
Journal of Asian Earth Sciences 59, 141e155.
Zhang, Q., Wang, C.Y., Liu, D., Jian, P., Qian, Q., Zhou, G., Robinson, P.T., 2008. A brief
review of ophiolites in China. Journal of Asian Earth Sciences 32, 308e324.
Zhao, D., Maruyama, S., Omori, S., 2007. Mantle dynamics of Western Pacic and
East Asia: Insight from seismic tomography and mineral physics. Gondwana
Research 11, 120e131.