GModel PEPI-5250; No.of Pages18 ARTICLE IN PRESS Physics ... · encompassing virtually all subduction zone segments with well-constrained geometry, largely following the segments

G

P

T

Ea

b

c

a

ARRA

GDCTD

EM

KPSFS

1

eTtimvtr

htpsSaa

k

0d

ARTICLE IN PRESSModel

EPI-5250; No. of Pages 18

Physics of the Earth and Planetary Interiors xxx (2010) xxx–xxx

Contents lists available at ScienceDirect

Physics of the Earth and Planetary Interiors

journa l homepage: www.e lsev ier .com/ locate /pepi

he global range of subduction zone thermal models

llen M. Syracusea,∗, Peter E. van Kekenb, Geoffrey A. Abersc

Department of Geoscience, University of Wisconsin, 1215 W. Dayton St., Madison, WI 53706, USADepartment of Geological Sciences, University of Michigan, 1100 North University Avenue, Ann Arbor, MI 48109-1005, USALamont-Doherty Earth Observatory, Columbia University, PO Box 1000, 61 Rte 9W, Palisades, NY 10964, USA

r t i c l e i n f o

rticle history:eceived 31 August 2009eceived in revised form 11 January 2010ccepted 5 February 2010

uest Editorsaisuke Suetsuguraig Binaoru Inoueouglas Wiens

a b s t r a c t

We model 56 segments of subduction zones using kinematically defined slabs based on updated geome-tries from Syracuse and Abers (2006) to obtain a comprehensive suite of thermal models for the globalsubduction system. These two-dimensional thermal models provide insight to the dehydration and melt-ing processes that occur in subduction zones. Despite the wide range of slab geometries, ages, convergencevelocities and upper plates the predicted thermal structures share many common features. All modelsfeature partial coupling between the slab and the overriding plate directly downdip of the thrust zone,invoked to replicate the cold nose observed in measurements of heat flow and seismic attenuation. Wetest four separate assumptions about the causes of the partial coupling: (1) the downdip end of the partialcoupling is at a constant depth, (2) it is at constant distance trenchward from the arc, (3) it is defined bya critical surface slab temperature, or (4) it is adjusted such that the hottest part of the mantle wedge

ditorark Jellinek

eywords:late tectonics

beneath the arc is at a constant temperature for all subduction zones. In all of these models, slabs reachtemperatures where the top of the oceanic crust and sediments dehydrate before they reach subarcdepths, and the overlying mantle wedge is too hot for hydrous minerals to be stable at subarc depths. Bycontrast, the interior of the oceanic crust and underlying mantle within the downgoing plate remains coldenough for hydrous phases to be stable beyond the arc in all but the hottest subduction zones, allowing

d the
ubduction zonesinite-element modelingeismic attenuation
water to be carried beyon

. Introduction

Subduction zones and the volcanic arcs formed above themxtend across approximately 40,000 km of the Earth’s surface.heir thermal structure affects the processes that contribute tohe dehydration of subducting crust and mantle, the generation ofntermediate-depth earthquakes, melt production in the overlying

antle wedge and the formation of arc volcanoes. Given the largeariation in slab input parameters (age, speed), geometry (slab dip,rench advance or roll-back) and overriding plate structure the fullange of these thermal structures is not well understood.

Previous numerical models of subduction zone temperaturesave often focused on comparing idealized geometries and parame-erizations, examining the effects of varying factors such as slab dip,late convergence velocity, or subducting plate age, or modeling a

Please cite this article in press as: Syracuse, E.M., et al., The global rangedoi:10.1016/j.pepi.2010.02.004

mall number of representative subduction zones (e.g., Davies andtevenson, 1992; Peacock, 1996; Kincaid and Sacks, 1997; Peacocknd Wang, 1999; Conder et al., 2002; van Keken et al., 2002; Geryand Yuen, 2003; Kelemen et al., 2003; Abers et al., 2006; Cagnioncle

∗ Corresponding author. Tel.: +608 890 1019.E-mail addresses: [email protected] (E.M. Syracuse),

[email protected] (P.E. van Keken), [email protected] (G.A. Abers).

031-9201/$ – see front matter © 2010 Elsevier B.V. All rights reserved.oi:10.1016/j.pepi.2010.02.004

arc in the slab.© 2010 Elsevier B.V. All rights reserved.

et al., 2007). While these studies have been helpful in illuminatingthe effects that each of these individual parameters has on the ther-mal structure of a subduction zone they do not allow for a detailedcomparison of the entire suite of parameters in combinations foundover the global range of subduction zones. Recently, Wada andWang (2009) discussed models for 15 arc segments with a spe-cific focus on the nature of the cold fore-arc corner, referred to asthe ‘cold nose’ (Kincaid and Sacks, 1997). The low temperature andlow-attenuation nature of this corner is inferred to exist in a grow-ing number of subduction zones from measurements of heat flow(e.g., Furukawa and Uyeda, 1989; Blackwell et al., 1990; Hyndmanand Wang, 1995; Hyndman and Peacock, 2003; Yoshimoto et al.,2006) and seismic attenuation (Hashida, 1989; Takanami et al.,2000; Schurr et al., 2003; Stachnik et al., 2004; Rychert et al., 2008).The low-attenuation and low temperature nature of this regionindicates that it is isolated from the large scale counterflow in themantle wedge, which suggests that decoupling of the viscous inter-action between the slab and overlying mantle wedge is necessaryto a depth significantly below the downdip limit of the seismo-

of subduction zone thermal models. Phys. Earth Planet. In. (2010),

genic zone. Based on heat flow constraints, Wada and Wang (2009)concluded that the depth to decoupling was uniform at ∼80 kmdepth and that this indicates a universal mechanism governing thedecoupling between slab and mantle, without specifying the detailsof this mechanism.

dx.doi.org/10.1016/j.pepi.2010.02.004

dx.doi.org/10.1016/j.pepi.2010.02.004

http://www.sciencedirect.com/science/journal/00319201

http://www.elsevier.com/locate/pepi

mailto:[email protected]



dx.doi.org/10.1016/j.pepi.2010.02.004

ARTICLE IN PRESSG Model

PEPI-5250; No. of Pages 18

2 E.M. Syracuse et al. / Physics of the Earth and Planetary Interiors xxx (2010) xxx–xxx

Table 1Thermodynamic parameters used in thermal modeling.

Symbol Name Value Units

A Scaling factora 3.5 × 1022 s−1

n Stress exponenta 3.5m Grain-size exponenta 0E* Activation energya 540 kJ/molV* Activation volume 0 cm3/mol� Shear modulusa 80 GPa� Density 3300 kg/m3

T� Potential temperature, mantleb 1421.5 ◦C�T Mantle adiabatic gradient 0.5 ◦C/kmcp Specific heat 1250 J/kg/K�crust Thermal conductivity, crust 2.5 W/m/K�crust Thermal conductivity, mantle 2.5 W/m/Kqsurface Surface heat flow 65 mW/m2

qupper crust Heat production, upper crust 1.3 W/m3

qlower crust Heat production, lower crust 0.27 W/m3

o

wscotgggv(zt

2

2

ec

Fig. 2. An example parameterization of a subduction zone. The slab surface is char-acterized by its geometry, convergence velocity, sediment layer thickness and age.Its interface with the overlying plate and mantle are characterized by the depthat which the coupling between to the transitions from fully decoupled to partially

Fu

a Parameters used in Karato and Wu (1993) dislocation creep regime for drylivine.b Based on Stein and Stein (1992).

In this study, we provide a significant extension to previousork and provide a global comparison of the thermal structure of

ubduction zones. We are particularly interested in how the coldorner can be maintained and wish to provide further diagnosticsn the nature of the transition from decoupling to coupling betweenhe slab and overriding mantle in subduction zones. To develop thelobal suite of models we parameterize the two-dimensional (2D)eometry of a large number of cross-sections that evenly space thelobal subduction system using a slightly improved and expandedersion of the global geometry compilation of Syracuse and Abers2006). We obtain a complete and uniform sampling of subductionones and use high-resolution finite-element models to examinehe range of thermal models.

. Methods


.1. Modeling approach

We develop 2D thermal models for 56 sections of volcanic arcncompassing virtually all subduction zone segments with well-onstrained geometry, largely following the segments defined by

ig. 1. A map of the subduction zones used in this study, following groupings of Syracusesed in all following figures to indicate arc segment. Small circles indicate segments of su

coupled (50 km depth for all subduction zones) and the transition from partiallycoupled to fully coupled (see text). The overlying plate is characterized by its typeand thickness or age.

Syracuse and Abers (2006) (Table 1 and Fig. 1). The modelingapproach closely follows that of van Keken et al. (2002) in whichthe slab is defined kinematically by motion along a prescribed slabsurface, and flow in the overlying mantle wedge is computed inresponse to this motion (Figs. 2 and 3). The mantle viscosity isdefined by a non-Newtonian, temperature- and stress-dependentrheology based on the dislocation creep flow law for dry olivineof Karato and Wu (1993), and does not incorporate more complexflow laws such as those in Kneller et al. (2007). The governing equa-tions are provided in the appendix. We assume the mantle to beincompressible and ignore compressible effects such as viscous dis-sipation and adiabatic heating and cooling. We also ignore the roleof secondary convection in the wedge, which simplifies the modelscompared to those that taken into account strong thermal or com-positional buoyancy in the wedge along with low mantle viscosities


(e.g., Billen and Gurnis, 2001; Gerya and Yuen, 2003) The modeledtemperature is effectively the potential temperature and we addan adiabatic gradient of 0.5 K/km a posteriori. Given the relativelyshallow depths of the subduction zone models this provides an

and Abers (2006). The color and shape of symbols on the map are the same as thosebduction zones that are modeled but have no overlying arc.

dx.doi.org/10.1016/j.pepi.2010.02.004

ARTICLE ING Model


E.M. Syracuse et al. / Physics of the Earth and

Fig. 3. Initial 2D thermal model for Alaska showing boundary conditions. Tocean is theslab temperature at the left side of the model, based on either the halfspace coolingmodel or the GDH1 plate-cooling model (Stein and Stein, 1992). The velocity of theslab (vslab) is equal to the plate convergence velocity perpendicular to the dip of theslab (vc), and vupper plate and vwedge are the velocities of the upper plate and wedge,respectively, parallel to the slab surface. L is the distance along the slab surface forwhich the mantle wedge is partially coupled to the slab. The left side and bottom ofthe model are inflow/outflow boundaries. The right side of the model has T(z) = T0

for the upper 95 km, and T(z) = TP below 95 km depth, with a velocity of zero from 0to 50 km depth and a zero normal stress boundary condition for depths greater than5wg

aaAmt(

doesstTtmse

deI

described in Syracuse and Abers (2009). These corrections are less

Fu

0 km. Flow in the wedge is governed by conservation of mass and momentum,here P is the dynamic pressure, � is viscosity and ε̇ is strain rate. An adiabatic

radient of 0.5 ◦C/km is added to all temperatures a posteriori.

cceptable approximation to the fully compressible equations (Leend King, 2009). All associated parameters are shown in Table 1.

mantle potential temperature of 1421.5 ◦C is assumed for allodels, based on Stein and Stein (1992), though the actual poten-

ial temperature may vary up to 200 ◦C between subduction zonesLangmuir et al., 1992; Wiens et al., 2006).

The models are calculated in a finite-element mesh to 300 kmepth using the Sepran modeling package, which follows the meth-ds described in Cuvelier et al. (1986) (Fig. 4). The size of the linearlements is varied throughout the model, with the smallest dimen-ion of ∼1 km in the regions with the largest temperature- andtress-gradients (such as the mantle wedge corner and the top ofhe slab) and spacing up to 10 km far away from strong gradients.his provides sufficient resolution for reliable temperature predic-ions as is demonstrated in a recent benchmark for subduction zone

odeling (van Keken et al., 2008). Details on the discretization,olution techniques and resolution tests are available in van Kekent al. (2002, 2008).


The models shown here are time-dependent, but the kinematicescription and boundary conditions do not change in time. Anxample of the initial condition for the models is shown in Fig. 3.n most subduction zone models the model evolution is tracked

ig. 4. The grid mesh for Alaska, showing triangular grid elements and grid node of 1 kmpper plate are highlighted by black lines.

PRESSPlanetary Interiors xxx (2010) xxx–xxx 3

for a period of 20 Ma which is sufficient to reach a steady statethermal structure. For a number of subduction zones with low con-vergence velocities we follow the evolution until 40 Ma and for theyoung Nankai subduction zone we establish the thermal structureat 16 Ma.

In order to better correlate with heat flow and seismic attenua-tion measurements, the boundary between the slab and the mantlewedge is separated into three regions, following Kneller et al. (2005,2007): (1) the shallow thrust zone extending to 40–50 km depth,corresponding to the seismogenic zone (Tichelaar and Ruff, 1993)where the subducting plate is fully decoupled from the upper plate(the upper plate is fixed as the subducting lower plate moves freelypast it); (2) the deep subducting plate/mantle wedge interface atwhich the two surfaces are fully coupled and move together, and(3) an intermediate region of partial coupling, in which the mantlewedge moves at 5% of the subducting plate velocity. The 5% cou-pling represents an average of the values tested by Kneller et al.(2007) that reproduce observations of the cold nose in the shallowmantle wedge. Our approach is one example of a number of meth-ods that reproduce the observed cold corner (Wada et al., 2008).Shear heating along the seismogenic zone is not included in currentmodels. The differences between individual 2D thermal model arebased solely on information about the subducting plate (geome-try, velocity, age, subducted and accreted sediment thickness), theupper plate (crustal type, thickness or age, and location of the arc),and the specific location of the boundary between partial and fullcoupling between downgoing plate and overriding mantle wedge(Fig. 2).

2.2. Lower plate parameterization

The geometry of the top of the subducting slab is based theanalysis by Syracuse and Abers (2006) of the global EHB catalog(Engdahl et al., 1998) and the assumption that intermediate-depthseismicity lies within the upper several kilometers of the subduct-ing plate (e.g., Zhao et al., 1992; Widiyantoro and van der Hilst,1996; Bijwaard et al., 1998; Hacker et al., 2003b). For each 500-km-long arc section, ten cross sections parallel to the dip of theslab are taken and the resulting shapes are averaged to obtain aninitial 2D slab geometry. This scheme allows parameterizations formost of the Earth’s subduction zones. These 2D geometries arecorrected for bias in intermediate-depth earthquake location thatoriginate from using a one-dimensional (1D) rather than a three-dimensional (3D) global velocity model for earthquake location, as


than 8 km in depth and 25 km in across-arc direction and result inminor shifts of the slab surface. The slab surface geometries are alsocorrected for the effects of earthquake mislocation on the widen-ing of the Wadati–Benioff zone (WBZ) and associated shifting of the

in the wedge corner and neighboring slab. The slab surface and base of the rigid

dx.doi.org/10.1016/j.pepi.2010.02.004

ING

P

4 h and

ib

rCairFsASis1(

woarvdaSCttmiaSt1tHms

2

(iitPta6adha

2

tfamlmsW

ARTICLEModel


E.M. Syracuse et al. / Physics of the Eart

nferred top of the slab surface into the mantle wedge as describedy Syracuse and Abers (2006).

For Nicaragua and Costa Rica, the slab surfaces are based onecent earthquake relocations from the TUCAN (Tomography Underosta Rica and Nicaragua) array (Syracuse et al., 2008) and takedvantage of the higher-precision locations. The Mexican slabs based on the geometry Perez-Campos et al. (2008) based oneceiver function analysis and tomographic imaging of the slab.or Cascadia, the slab surfaces are based on the seismologicaltudies of Bostock et al. (2002) and Rondenay et al. (2008). Theegean slab surface is based on seismicity from the Internationaleismic Centre (ISC) catalog and is consistent with the seismicmaging of Suckale et al. (2009). The geometry of the Calabrianlab is estimated from the EHB seismic catalog (Engdahl et al.,998). The Nankai slab geometry is adapted from Hirose et al.2007).

Other parameters associated with the subducted plate are dealtith as follows. Plate velocities are taken as the normal component

f plate convergence rates as calculated from magnetic anomalynd geodetic plate velocity models (Syracuse and Abers, 2006 andeferences therein). In areas with back-arc spreading, the full con-ergence velocity includes the back-arc spreading rate so that itescribes motion of the incoming plate with respect to the arc. Thege of the subducting plate is prescribed at the trench and based onyracuse and Abers (2006) and references therein. Slabs from theretaceous Quiet Zone of otherwise unconstrained age are assumedo have an age of 100 Ma. For oceanic plates younger than 65 Mahe halfspace cooling model is used and the GDH1 plate cooling

odel (Stein and Stein, 1992) is used for older plates. The insulat-ng effects of slab sediments are taken into account for erosive andccreting margins as identified by Clift and Vannucchi (2004) andcholl and von Huene (2007). The initial amount of sediment at therench is eroded to a specified thickness when it reaches a depth of5 km. The overlying and subducted sediment thickness is based onhe compilations of Clift and Vannucchi (2004) and Scholl and vonuene (2007) and account for erosion or accretion at the activeargins and the closing of pore spaces in the sediments during

ubduction.

.3. Upper plate parameterization

Each upper plate is described as either continental or oceanicTable 2). If the upper plate is continental, the crustal thicknesss specified based on Clift and Vannucchi (2004). Heat productionn the continental crust is specified to be 1.3 × 10−6 W/m3 withinhe upper 15 km and 2.7 × 10−7 W/m3 in the lower crust, followingeacock and Wang (1999). The continental boundary condition foremperature is consistent with this distribution of heat productionnd adjusted to provide an average continental surface heat flow of5 mW/m2 (Pollack et al., 1993), following the approach of Peacocknd Wang (1999). The upper plate is assumed to be rigid to 50 kmepth. In regions with back arc spreading, the upper plate age isalf the age of the oldest part of the plate to represent an averagege of back arc basin crust.

.4. Partial coupling depth

Partial coupling between the slab and the mantle wedge belowhe thrust zone is necessary to create the cold fore-arc separatedrom mantle wedge flow. Kneller et al. (2005, 2007) and Abers etl. (2006) found that the transition from partial to full coupling


ust occur at least at 70–80 km depth in order to have a cold nosearge enough to agree with constraints from heat flow and seis-

ic attenuation. Note that we maintain partial coupling betweenlab and cold corner, which is different from the interpretation by

ada et al. (2008) and Wada and Wang (2009) who argued that

PRESSPlanetary Interiors xxx (2010) xxx–xxx

the shallow part of the interface needs to be fully decoupled. Yetall studies agree that the transition to full coupling needs to besharp (i.e. occur over a few kilometers). The mechanism causingthis partial or full decoupling is not well understood, but may bedue to regions of ductile shear, stable frictional slip, or transientcreep. The number of arcs with sufficient quality heat flow and/orattenuation measurements is low and not sufficient to determinethe global nature of the structure and mechanism causing the coldfore-arc corner. To further develop a quantitative understandingon the possible controls on the partial to full coupling transitionwe assume four distinct cases for each 500-km subduction zonesection. The first case, referred to as D80, is similar to the geomet-rical constraint mentioned above, in which the boundary betweenpartial and full coupling set at 80 km depth (2.6 GPa) for all sub-duction zones (Table 3). This constant value correlates with theresults from measurements of low seismic attenuation where thecold nose of the subduction zone is largely cut off from mantlewedge flow and persists above the slab to a depth of roughly 80 km(e.g., Alaska: 75 km, Stachnik et al., 2004; Honshu: 80 km, Tsumuraet al., 2000; Central America: 80 km, Rychert et al., 2008; Chile:100 km, Haberland and Rietbrock, 2001).

The second case, referred to as X25, uses a different geometricalconstraint in which the boundary between partial and full couplingoccurs at 25 km trenchward of the arc. Similar to the first case, thisconstant value represents the average offset between the arcwardlimit of the region of low attenuation in the cold nose and thearc (e.g., Honshu: 20–25 km, Tsumura et al., 2000; Central Amer-ica: 20–30 km, Rychert et al., 2008; Chile: 25–35 km, Haberlandand Rietbrock, 2001). It also follows surface heat flow measure-ments from Honshu and Cascadia, where across-arc profiles showan abrupt increase in heat flow 10–30 km trenchward of the arc(Furukawa and Uyeda, 1989; Blackwell et al., 1990; Hyndman andWang, 1995).

The third case, referred to as W1300, provides an attempt toreach uniform temperature below the volcanic arc. In this case theboundary between the partial and full coupling of the slab andoverlying mantle is adjusted such that the maximum temperaturedirectly beneath the arc is 1300 ◦C. This assumes that the limit-ing factor in volcano formation is the point at which the mantlewedge reaches a temperature high enough to allow magma gener-ation (e.g., Peacock, 1996; Schmidt and Poli, 1998) and that the arcforms vertically above the trenchward limit of mantle melting. Wechoose this value for the below-arc temperature as an average esti-mate, but we recognize that the temperature and pressure rangeover which mantle melting occur is strongly dependent on mantlecomposition and water content and that quantitative estimates arehighly dependent on the experimental details (e.g., Kushiro et al.,1968; Kushiro, 1970; Green, 1973; Milhollen et al., 1974; Mysenand Boettcher, 1975; Kinzler and Grove, 1992; Hirth and Kohlstedt,1996; Grove et al., 2006).

The fourth case, referred to as T550, provides our attemptat incorporating a temperature-constraint from the assumedbrittle–ductile transition along the slab interface. Here, the tran-sition from partial to full coupling occurs when the slab surfacereaches a temperature of 550 ◦C and the depth of the transitionis updated iteratively for each time-step in the time-dependentmodel. The value of 550 ◦C is based on the temperature of thebrittle–ductile transition assuming a strain rate of 10−17 s−1 anda dry olivine rheology for a range of plate ages (Karato andWu, 1993; Conder, 2005). The use of a higher strain rate wouldimply higher coupling transition temperatures. While choosing a


temperature-dependent coupling transition greatly simplifies thephysics of the brittle–ductile transition (e.g., Kohlstedt et al., 1995),it avoids the difficulties associated with consistently defining strainrate and stress at the coupling transition point in these mixedkinematic–dynamic models.

dx.doi.org/10.1016/j.pepi.2010.02.004

Pleasecite

this

articlein

press

as:Syracu

se,E.M.,et

al.,The

globalran

geof

subd

uction

zone

therm

alm

odels.Ph

ys.EarthPlan

et.In.(2010),

doi:10.1016/j.p

epi.2010.02.004

AR

TIC

LE

IN P

RE

SS

GM

odel

PEPI-5250;N

o.ofPages18

E.M.Syracuse

etal./Physics

oftheEarth

andPlanetary

Interiorsxxx (2010) xxx–xxx

5

Table 2Input parameters for each arc section.

Name Lon (◦) Lat (◦) H (km) Arc—trenchdistance(km)

Slab dip (◦) Vc (km/Ma) Age (Ma) Descent rate(km/Ma)

Thermalparame-ter/100(km)

Sedimentthickness(km)

Subductedsedimentthickness(km)

Upper plate type Upper platethickness(km)

Upper plateage (Ma)

Colombia/Ecuador −77.5 0 155 305 28.0 60.0 15.0 28.2 4.2 1.8 0.30 Continental 50Northern Peru Gap −79 −6.8 10.2 69.5 29.0 12.3 3.6 0.7 0.40 Continental 50Central Peru Gap −76.8 −11.2 13.2 66.7 33.9 15.2 5.2 0.7 0.40 Continental 40Peru −70.85 −16.61 117 270 28.7 65.1 42.1 31.4 13.4 0.7 0.40 Continental 40North Chile −68.63 −20.83 132 298 29.3 79.0 46.2 38.7 17.8 0.3 0.20 Continental 45North-Central Chile −68.53 −24.72 126 309 23.6 77.4 42.8 30.9 13.3 0.3 0.20 Continental 40Central Chile Gap −69.5 −30 13.7 73.8 37.7 17.5 6.6 2.0 1.30 Continental 40Central Chile −70.57 −35.24 118 265 27.5 71.6 32.4 33.1 10.7 2.0 1.30 Continental 45South-Central Chile −72.8 −43.75 100 247 29.5 74.7 23.5 36.8 8.6 2.0 1.30 Continental 40South Chile −71.7 −39.5 114 228 34.2 74.9 10.3 42.1 4.3 3.0 0.60 Continental 32Mexico −99.5 19.1 80 360 59.5 47.0 10.0 40.5 4.0 0.6 0.30 Continental 45Guatemala/El Salvador −90.42 14.15 142 196 56.9 66.9 17.4 55.9 9.7 0.5 0.30 Continental 45Nicaragua −86.7 12.5 169 203 62.2 71.1 17.5 62.8 11.0 0.5 0.30 Continental 30Costa Rica −84.73 10.47 87 131 58.9 75.4 15.8 63.9 10.1 0.5 0.30 Continental 40North Cascadia 121.5 42 100 345 22.4 40.0 10.6 12.1 1.3 2.0 0.40 Continental 35Central Cascadia 122 47 90 305 19.8 30.0 6.9 14.2 1.0 2.0 0.40 Continental 40Alaska −153.55 58.87 97 371 42.0 49.0 47.1 33.0 15.6 2.0 0.40 Continental 35Alaska Peninsula −158.8 56.53 108 280 45.0 59.0 52.2 41.8 22.0 0.8 0.40 Continental 40East Aleutians −165.97 54.13 105 204 45.6 64.2 55.3 45.9 25.4 1.5 0.30 Continental 35Central Aleutians −172.52 52.32 96 180 53.7 63.4 55.9 51.0 28.4 1.5 0.30 Continental 35West Aleut −178.8 51.78 107 177 55.9 50.2 56.1 42.0 23.5 1.5 0.30 Continental 30Kamchatka 159.15 53.59 111 227 50.7 75.1 92.9 58.2 54.1 0.6 0.30 Continental 40North Kurile 155.43 50.25 106 219 50.4 78.3 105.6 60.2 63.6 0.6 0.30 Continental 45South Kurile 151.75 46.83 100 185 46.4 77.4 108.5 56.0 60.7 0.6 0.30 Continental 25Hokkaido 147 44.62 116 217 41.9 74.7 115.2 49.7 57.2 0.6 0.30 Continental 25North Honshu 140.88 40.65 114 319 28.7 81.6 129.3 39.2 50.7 0.6 0.30 Continental 40Central Honshu 140.28 37.62 104 289 33.9 82.7 130.5 46.3 60.4 0.6 0.30 Continental 40Izu 139.77 33.13 134 228 46.1 45.6 135.4 32.6 44.1 0.4 0.20 Oceanic 50Bonin 140.88 27.24 164 216 62.8 31.6 145.3 28.1 40.8 0.4 0.20 Oceanic 50North Marianas 143.44 21.93 185 198 61.5 15.3 147.8 13.4 19.9 0.4 0.20 Oceanic 25South Marianas 145.83 17.6 169 273 56.9 50.0 151.6 41.9 63.5 0.4 0.20 Oceanic 25Nankai 137.3 35.3 31.4 43.0 20.0 22.4 4.5 1.5 0.30 Continental 32Kyushu 130.57 31.22 122 222 53.8 72.0 27.0 57.6 15.6 0.4 0.20 Continental 31Ryukyu 128.25 27.85 87 187 43.5 69.5 43.0 48.1 17.1 1.0 0.20 Continental 31North Philippines 120.55 16.33 142 212 57.4 87.7 32.4 73.1 23.6 1.5 0.30 Oceanic 20South Philippines 125.52 9.59 193 167 64.3 69.0 59.8 62.0 36.3 0.4 0.20 Oceanic 20New Britain 150.9 −5.53 137 157 68.2 100.4 25.0 92.5 23.1 1.0 0.60 Oceanic 2Solomon 154.93 −5.83 155 164 70.4 93.6 31.0 87.5 27.1 0.4 0.20 Oceanic 50North Vanuatu 167.47 −13.8 172 142 71.3 39.6 44.0 37.5 19.4 0.4 0.20 Oceanic 5South Vanuatu 169.83 −20.2 94 104 66.8 112.7 50.0 103.6 51.8 0.4 0.20 Oceanic 5Tonga −174.87 −19.18 123 196 52.4 165.8 109.0 131.3 143.2 0.4 0.20 Oceanic 10Kermadec −178.47 −30.2 171 166 56.1 64.6 105.6 53.5 56.6 0.4 0.20 Oceanic 25New Zealand 176.98 −37.86 153 234 52.0 30.4 100.0 24.2 24.2 2.0 0.40 Oceanic 20North Sumatra 97.6 3.87 132 330 49.1 41.5 48.3 31.2 15.1 2.5 1.40 Continental 30Central Sumatra 99.98 0.08 117 315 46.4 40.0 54.8 29.0 15.9 2.0 0.85 Continental 31South Sumatra 102.37 −3.38 105 296 50.2 48.7 68.9 37.4 25.9 1.5 0.30 Continental 33.5Sunda Strait 106.68 −6.75 105 274 49.3 61.0 85.5 46.3 39.6 1.5 0.30 Continental 33.5Java 111.76 −7.81 168 307 45.8 67.8 111.2 48.6 54.0 1.2 0.30 Continental 33.5Bali/Lombok 116.47 −8.42 186 327 45.9 69.8 134.9 50.1 67.7 1.2 0.30 Continental 33.5West Banda Sea 121.83 −8.76 126 290 54.4 73.6 100.0 59.8 59.8 0.4 0.20 Continental 33.5East Banda Sea 129.5 −6.73 159 273 48.0 25.3 100.0 19.5 19.5 0.4 0.20 Continental 16.5North Lesser Antilles −61.67 16.05 122 212 49.8 17.6 83.0 13.4 11.1 0.5 0.25 Oceanic 50South Lesser Antilles −61.18 13.33 141 235 42.4 17.9 86.6 12.1 10.4 2.5 0.50 Oceanic 50Scotia −26.45 −57.78 118 161 64.6 60.8 59.1 54.5 28.6 0.4 0.20 Oceanic 2Aegean 25.4 36.4 120 250 30.5 15.0 200.0 7.6 15.2 4.0 1.00 Continental 35Calabria 15 38.45 43.0 45.0 190.0 30.7 58.3 5.0 1.00 Oceanic 20

dx.doi.org/10.1016/j.pepi.2010.02.004

Please cite this article in press as: Syracuse, E.M., et al., The global range of subduction zone thermal models. Phys. Earth Planet. In. (2010),doi:10.1016/j.pepi.2010.02.004




Table 3Subduction parameters and descriptions of the coupling transition, slab and mantle wedge temperatures, and dehydration for each arc section in each case.

Name D80

Transitiondepth (km)

TransitionT (◦C)

Slab Tbeneath arc(◦C)

Moho Tbeneath arc(◦C)

Maximummantle Tbeneath arc(◦C)

Depth ofmax Tbeneath arc(km)

Transitionoffset fromarc (km)

Slabsurface T,30 km (◦C)


Minimumslab T, 240km (◦C)

Colombia/Ecuador 80 806 959 823 1451 119 115 267 1022 903Northern Peru Gap 80 748 191 980 905Central Peru Gap 80 795 197 1033 890Peru 80 793 861 622 1436 93 78 164 975 725North Chile 80 808 875 626 1450 106 104 148 956 658North-Central Chile 80 773 861 633 1443 98 112 157 961 711Central Chile Gap 80 716 164 1010 843Central Chile 80 599 784 570 1417 95 77 176 990 732South-Central Chile 80 630 763 523 1387 83 38 203 976 756South Chile 80 773 899 690 1428 92 43 278 1038 887Mexico 80 760 760 676 1410 91 25 334 1081 954Guatemala/El Salvador 80 744 891 656 1444 108 44 224 971 758Nicaragua 80 478 876 646 1457 129 59 163 962 673Costa Rica 80 481 690 363 1231 82 5 159 917 642North Cascadia 80 749 936 811 1321 82 31 387 1094 1029Central Cascadia 80 741 888 781 1291 79 31 397 1141 1043Alaska 80 741 808 530 1353 81 27 221 948 675Alaska Peninsula 80 692 791 476 1407 90 32 176 939 602East Aleut 80 683 774 422 1397 87 23 152 929 571Central Aleutians 80 689 763 381 1356 82 12 149 907 558West Aleutians 80 697 792 461 1371 85 14 164 910 596Kamchatka 80 617 720 306 1419 90 27 62 876 382North Kurile 80 620 713 286 1407 87 23 53 855 388South Kurile 80 612 695 250 1392 84 17 52 890 382Hokkaido 80 611 720 327 1422 93 32 59 892 394North Honshu 80 621 721 360 1413 93 54 62 866 417Central Honshu 80 617 704 292 1402 87 34 67 862 394Izu 80 480 743 462 1377 100 49 60 873 460Bonin 80 510 787 582 1375 107 51 72 858 501North Marianas 80 559 860 731 1430 107 52 107 895 630South Marianas 80 470 790 528 1446 117 68 49 857 428Nankai 80 607 262 1014 898Kyushu 80 698 805 463 1431 94 31 156 900 549Ryukyu 80 709 749 354 1279 81 16 165 946 613North Philippines 80 687 833 534 1450 113 54 164 918 612South Philippines 80 685 852 593 1459 131 43 131 899 520New Britain 80 504 818 488 1452 103 24 164 930 620Solomon 80 480 793 488 1380 112 28 152 901 566North Vanuatu 80 598 882 635 1454 111 30 176 935 615South Vanuatu 80 494 720 257 1390 86 7 114 890 478Tonga 80 366 699 182 1451 100 32 24 871 265Kermadec 80 476 773 463 1376 80 46 42 864 392New Zealand 80 656 804 600 1431 105 54 102 871 530North Sumatra 80 630 815 622 1423 100 58 220 939 684Central Sumatra 80 679 812 587 1389 88 37 218 941 671South Sumatra 80 664 749 404 1368 84 19 127 894 497Sunda Strait 80 644 730 341 1394 86 23 90 863 431Java 80 632 781 495 1457 124 78 70 891 396Bali/Lombok 80 628 786 516 1457 144 96 67 887 382West Banda Sea 80 648 752 372 1436 98 38 78 850 381East Banda Sea 80 700 847 675 1420 103 56 123 886 601North Lesser Antilles 80 542 756 582 1296 88 36 119 892 644South Lesser Antilles 80 528 794 652 1331 94 56 101 891 673Scotia 80 627 839 459 1445 89 20 135 935 563Aegean 80 426 709 610 1275 84 65 102 907 704Calabria 80 349 79 867 499

Name X25


TransitionT (◦C)







SlabSurface T,240 km (◦C)

Minimumslab T,240 km (◦C)

Colombia/Ecuador 136 767 974 726 1446 138 25 233 1084 883Peru 106 583 832 479 1431 104 25 139 1001 714North Chile 119 610 860 470 1448 120 25 124 980 635North-Central Chile 116 582 846 482 1439 114 25 132 996 696Central Chile 106 498 731 507 1403 106 25 177 1017 783South-Central Chile 86 442 704 468 1379 87 25 178 988 752South Chile 91 588 872 633 1423 97 25 246 1066 881

dx.doi.org/10.1016/j.pepi.2010.02.004




E.M. Syracuse et al. / Physics of the Earth and Planetary Interiors xxx (2010) xxx–xxx 7

Table 3 (Continued )

Name X25


TransitionT (◦C)







SlabSurface T,240 km (◦C)


Mexico 80 760 760 655 1410 91 25 334 1104 954Guatemala/El Salvador 103 567 875 590 1442 118 25 200 988 743Nicaragua 129 567 869 550 1459 145 25 163 985 641Costa Rica 55 406 758 413 1388 68 25 159 909 658North Cascadia 83 888 954 800 1316 84 25 438 1138 1030Central Cascadia 67 801 900 801 1306 74 25 435 1103 1027Alaska 81 510 788 486 1351 81 25 192 971 675Alaska Peninsula 86 470 772 420 1403 93 25 153 942 598East Aleutians 78 434 761 397 1398 85 25 132 927 572Central Aleutians 69 432 751 387 1387 75 25 129 897 565West Aleutians 67 449 778 469 1390 78 25 140 907 603Kamchatka 82 359 709 272 1418 91 25 47 855 381North Kurile 79 343 702 259 1408 87 25 35 831 388South Kurile 72 317 690 242 1401 81 25 37 861 387Hokkaido 88 355 703 267 1418 96 25 43 873 390North Honshu 98 341 674 255 1404 100 25 44 849 407Central Honshu 86 328 675 227 1398 90 25 49 837 390Izu 105 395 711 369 1374 111 25 54 848 448Bonin 113 587 815 556 1373 124 25 67 896 491North Marianas 129 709 902 734 1433 135 25 102 954 733South Marianas 124 481 782 439 1448 136 25 47 843 423NankaiKyushu 86 453 789 413 1430 97 25 134 893 545Ryukyu 72 433 742 346 1372 72 25 143 942 618North Philippines 113 505 797 435 1449 126 25 143 916 589South Philippines 124 523 846 525 1460 150 25 113 904 495New Britain 79 448 817 485 1452 103 25 158 931 621Solomon 87 434 789 477 1381 115 25 148 890 562North Vanuatu 95 559 883 615 1456 117 25 169 949 608South Vanuatu 97 487 721 249 1400 85 25 114 881 479Tonga 87 307 689 166 1451 102 25 22 824 263Kermadec 127 458 778 395 1291 76 25 38 868 371New Zealand 112 563 829 553 1426 121 25 79 918 520North Sumatra 108 495 758 506 1414 112 25 191 958 673Central Sumatra 90 527 801 533 1381 93 25 186 974 667South Sumatra 74 399 722 378 1376 81 25 98 879 500Sunda Strait 78 382 718 315 1397 85 25 72 848 432Java 134 439 777 382 1457 143 25 55 904 369Bali/Lombok 157 515 791 378 1456 166 25 51 914 344West Banda Sea 93 387 735 309 1434 103 25 63 832 374East Banda Sea 114 685 887 656 1414 118 25 92 966 596North Lesser Antilles 92 537 762 560 1329 94 25 115 912 740South Lesser Antilles 108 436 819 629 1362 109 25 97 934 801Scotia 73 530 838 470 1445 85 25 129 946 566Aegean 104 542 730 554 1283 97 25 102 968 1214

Name W1300


TransitionT (◦C)









Colombia/Ecuador 50 968 987 1003 1300 148 7 267 1094 1493Peru 115 823 841 869 1300 112 8 164 1008 1493North Chile 130 853 861 884 1300 128 6 148 986 1493North-Central Chile 124 848 860 886 1300 121 8 157 1002 1493Central Chile 113 677 730 767 1300 112 12 176 994 715South-Central Chile 93 635 717 748 1300 93 12 203 992 1493South Chile 101 786 858 880 1300 100 8 276 1073 1493Mexico 102 945 700 715 1300 9 376 1117 1493Guatemala/El Salvador 133 810 858 870 1300 136 7 224 997 1491Nicaragua 162 797 842 853 1300 164 7 183 995 1387Costa Rica 78 699 753 764 1300 80 6 180 920 1485North Cascadia 88 898 956 967 1300 87 17 438 1144 1490Central Cascadia 74 839 912 925 1300 77 43 435 1121 1493Alaska 89 755 801 819 1300 88 10 221 981 1493Alaska Peninsula 101 698 758 778 1300 102 10 174 951 1493East Aleutians 95 676 752 767 1300 95 9 151 934 1492Central Aleutians 84 676 753 766 1300 85 8 149 906 1490West Aleutians 88 677 782 791 1300 91 8 159 920 1486Kamchatka 99 614 691 702 1300 101 9 62 860 1493

dx.doi.org/10.1016/j.pepi.2010.02.004






Name W1300


TransitionT (◦C)









North Kurile 95 627 686 701 1300 97 8 53 835 1493South Kurile 89 566 665 682 1300 90 8 46 868 1493Hokkaido 105 566 670 689 1300 106 9 54 879 1493North Honshu 107 627 681 706 1300 107 10 62 856 1493Central Honshu 97 610 664 687 1300 97 9 67 841 1493Izu 120 562 742 755 1300 121 10 60 899 1493Bonin 141 680 832 835 1300 144 10 72 921 1406North Marianas 157 786 916 920 1300 160 12 107 974 877South Marianas 151 573 759 764 1300 155 9 49 848 1493Kyushu 107 690 773 787 1300 110 8 153 901 1491Ryukyu 79 691 746 764 1300 79 9 162 946 1493North Philippines 136 707 764 781 1300 139 7 161 922 1489South Philippines 178 749 824 835 1300 182 6 129 913 465New Britain 123 571 755 766 1300 127 7 164 934 1424Solomon 134 554 760 760 1300 141 8 152 901 1252North Vanuatu 154 694 879 885 1300 158 6 176 974 579South Vanuatu 86 492 691 707 1300 89 5 114 881 1295Tonga 110 402 626 642 1300 114 8 24 825 1493Kermadec 162 570 739 756 1300 164 7 49 879 535New Zealand 139 780 840 853 1300 140 9 99 943 1483North Sumatra 124 694 779 802 1300 124 9 216 1001 1493Central Sumatra 102 701 804 819 1300 102 11 213 986 1493South Sumatra 89 656 742 753 1300 91 10 124 902 1493Sunda Strait 93 634 711 725 1300 94 9 89 856 1493Java 155 693 775 788 1300 156 7 69 921 1493Bali/Lombok 179 722 773 788 1300 180 7 65 928 1493West Banda Sea 114 625 714 728 1300 116 9 77 838 1493East Banda Sea 134 812 914 919 1300 134 9 105 992 1493North Lesser Antilles 69 499 743 750 1300 82 48 120 886 1493South Lesser Antilles 113 302 823 832 1300 114 20 101 939 1493Scotia 104 655 814 819 1300 107 6 135 957 1417Aegean 50 308 682 695 1300 79 122 122 876 1493

Name T550


TransitionT (◦C)









Colombia/Ecuador 58 550 929 822 1446 138 169 233 1022 883Northern Peru Gap 73 550 167 980 887Central Peru Gap 64 550 169 1033 886Peru 71 550 839 597 1436 92 99 139 975 730North Chile 79 550 863 592 1450 106 106 124 956 660North-Central Chile 69 550 849 627 1443 96 142 132 961 722Central Chile Gap 70 550 142 1010 844Central Chile 67 550 806 629 1417 92 107 177 990 807South-Central Chile 67 550 771 554 1379 87 71 178 976 752South Chile 56 550 879 728 1432 86 98 246 1038 902Mexico 59 550 899 718 1417 83 233 334 1081 956Guatemala/El Salvador 63 550 875 672 1444 102 69 200 971 773Nicaragua 63 550 873 654 1457 124 89 163 962 687Costa Rica 58 550 763 388 1383 71 17 159 917 656North Cascadia 57 550 855 778 1312 72 120 397 1071 1009Central Cascadia 51 550 881 723 1285 81 60 397 1141 1046Alaska 62 550 792 569 1369 73 142 192 948 688Alaska Peninsula 67 550 792 499 1413 85 51 153 933 614East Aleutians 71 550 781 434 1406 82 33 132 929 582Central Aleutians 71 550 775 391 1390 76 22 129 907 570West Aleutians 65 550 787 484 1395 76 28 140 910 608Kamchatka 90 550 741 262 1418 94 17 47 876 385North Kurile 72 550 757 238 1405 91 32 45 855 391South Kurile 94 550 673 182 1201 92 4 37 890 383Hokkaido 92 550 744 277 1422 96 22 43 892 397North Honshu 90 550 759 323 1413 94 37 44 866 422Central Honshu 93 550 732 223 1398 92 16 49 862 397Izu 78 550 766 486 1376 99 51 56 873 470Bonin 73 550 791 597 1375 103 59 67 858 512North Marianas 68 550 849 733 1429 101 62 104 895 638

dx.doi.org/10.1016/j.pepi.2010.02.004





Name T550


TransitionT (◦C)









South Marianas 89 550 817 534 1447 120 55 47 857 430Nankai 58 550 262 1014 881Kyushu 72 550 812 471 1434 91 39 136 900 559Ryukyu 70 550 770 366 1372 73 28 143 946 624North Philippines 73 550 837 537 1450 111 62 143 918 623South Philippines 71 550 853 594 1458 129 49 114 899 529New Britain 55 550 827 524 1449 94 41 158 930 635Solomon 81 550 810 499 1380 112 27 148 901 573North Vanuatu 54 550 872 649 1450 100 39 170 935 628South Vanuatu 68 550 776 278 1439 76 12 109 890 490Tonga 129 550 301 144 1350 60 -6 22 871 253Kermadec 82 550 796 482 1387 80 45 39 864 400New Zealand 71 550 799 603 1432 101 67 79 871 540North Sumatra 61 550 797 629 1424 95 89 193 939 693Central Sumatra 61 550 793 600 1395 82 67 186 941 680South Sumatra 70 550 759 422 1384 79 30 102 894 508Sunda Strait 80 550 755 330 1402 84 23 72 863 439Java 86 550 801 497 1458 125 71 55 891 401Bali/Lombok 87 550 808 519 1458 144 88 51 887 387West Banda Sea 86 550 778 354 1437 99 32 63 850 385East Banda Sea 66 550 818 662 1421 98 76 92 886 604North Lesser Antilles 72 550 761 601 1302 83 44 116 892 652South Lesser Antilles 72 550 793 665 1333 90 69 97 891 681Scotia 56 550 833 511 1443 78 42 129 935 575

1295

3

a(cw

3

3

atttcmsmswtttsdg

3

dtbs

Aegean 78 550 726 607Calabria 75 550

. Results

A major contribution from this work is the calculation andnalysis of a suite of thermal models for each subduction zoneTable 3 and Fig. 5). Despite the differing assumptions of the fourases, many similarities arise between results regardless of the case,hich are discussed below.

.1. Case descriptions

.1.1. D80Pressure–temperature paths of slab surfaces are shown in Fig. 6

nd Supplementary Materials for all subduction zones in each ofhe four cases. The dominant feature in the D80 case is the steepemperature gradient at 80 km depth, which occurs at the transi-ion from partial to full coupling. The mean slab temperature at theoupling transition is 621 ◦C (Figs. 7 and 8), and the mean maximumantle wedge temperature beneath the arc is 1400 ◦C. However, it

hould be noted that mantle temperatures are closely tied to theantle potential temperature of 1421.5 ◦C used in their calculation,

o that a lower mantle potential temperature would yield loweredge temperatures overall. Small perturbations in the depth of

he coupling transition result in similar P–T paths above and belowhe transition, with the steep gradient following the depth of theransition. Decreasing the transition depth would result in lowerlab surface temperatures at the end of partial coupling and pro-uce higher maximum wedge temperatures beneath the arc ineneral.

.1.2. X25


In X25, there is a greater variety of P–T paths than in D80ue to the variability of the coupling transition depth. Despitehe increased variation, the P–T paths more than 25 km above andelow the transition depths are similar between cases. The meanlab temperature at the coupling transition is 502 ◦C and the mean

81 67 102 907 70779 867 511

maximum mantle wedge temperature beneath the arc is 1404 ◦C.Increasing the offset between the coupling transition and the arcwould result in lower slab surface temperatures at the end of cou-pling, higher maximum wedge temperatures beneath the arc, anda similar variation in transition depths.

3.1.3. W1300The maximum mantle wedge temperature beneath the arc is

held fixed at 1300 ◦C in W1300. Here, the coupling transition depthhas a mean of 112 km. The mean slab surface temperature at thetransition is 676 ◦C. If a temperature other than 1300 ◦C were usedas the maximum mantle wedge beneath the arc, the coupling tran-sition would move at an average rate of about 1 km in depth per30 ◦C. Increasing the mantle wedge temperature would decreaseslab surface temperatures at the end of coupling.

It should be noted, however, that for the arcs with coupling tran-sition depths significantly greater than 80 km or steep slab dips, twopeaks in mantle wedge temperature beneath the arc often form(Fig. 9). The shallower peak occurs at depths of 50–110 km andis generally smaller than the second, deeper peak in temperature5–15 km above the slab surface. In these cases, the deeper tem-perature maxima are pinned below cooler mantle being draggeddown due to a large partial coupling depth. Since it is not clear atwhat depth melts are most likely to form, an alternative is to adjustthe partial coupling depth such that the shallower mantle maxi-mum reaches 1300 ◦C. However, beneath many arcs, the shallowermaxima cannot reach 1300 ◦C, and decreasing the partial couplingdepth to eliminate the double maxima yields subarc mantle wedgetemperatures hotter than the goal of 1300 ◦C (Fig. 9). We note thatany double maxima disappear if the slab is assumed to be fully


decoupled below the cold corner.

3.1.4. T550In this test, the transition from partial to full coupling occurs

where the slab surface reaches a temperature of 550 ◦C. This

dx.doi.org/10.1016/j.pepi.2010.02.004




Fig. 5. A comparison of thermal models for the Alaskan Peninsula and central Honshu from each case. (A) D80; (B) X25; (C) W1300; (D) T550 for the Alaska Peninsula. (E)D ed byb belol ranger

aiHasioo(diansdmt

3

3

s

80; (F) X25; (G) W1300; (H) T550 for central Honshu. The volcanic front is denotlack dotted lines indicate the slab surface and the slab Moho, assumed to be 7 km

ocation of cross section (black line), trench (black toothed line), and volcanoes (oeader is referred to the web version of the article.)

pproach leads to a range in transition depths from 51 km (1.5 GPa)n the Cascades to 129 km (4.2 GPa) in Tonga, with a mean of 72 km.owever for Tonga, the coupling transition occurs beneath the backrc, resulting in the arc being located above the cold nose. This caseeems unrealistic, as all seismic attenuation and heat flow stud-es thus far have shown the cold nose being localized trenchwardf the arc. Reducing the slab temperature at which the transitionccurs to 500 ◦C and adding a reasonable amount of shear heating30 mW/m2) to the to upper 50 km of the slab interface reduces theepth of the coupling transition only 3 km, not enough to bring it

nto the fore-arc (Fig. 10). Additionally, if the same 500 ◦C temper-ture limit and shear heating is extended to other arcs, the coldose is virtually eliminated in some arcs where attenuation mea-urements image a cold nose. While an element of temperatureependence in determining the depth of the coupling transitionay be a reasonable assumption, it not likely to be the only con-

rolling factor.


.2. Slab surface temperature

.2.1. Coupling transitionDespite the various assumptions about partial coupling depth,

ome features of the resulting models are consistent among differ-

the orange triangle. Solid black lines are temperature contours at 100 ◦C interval,w the slab surface, orange triangles show the location of the arc. Inset maps showtriangles). (For interpretation of the references to color in this figure legend, the

ent cases. The average temperature for all arc segments at the endof partial coupling is similar between the three cases in which it isa free variable, ranging from 502 ◦C to 676 ◦C. This shows that theuse of 550 ◦C as the slab temperature at the end of partial couplingin the T550 case is consistent with other tests; the temperature iswell within the range of slab surface temperatures at the end ofcoupling found in other models.

3.2.2. Thrust zone temperature and temperature at depthWhile the temperature of the slab beneath the arc and

at the coupling transition varies some between cases, thetemperature of the slab surface at shallower depths is con-sistent between cases. Fig. 11 compares the temperatures ofthe slab surface at 30 km depth among cases D80, X25, andW1300 for each arc section, and shows that the assumptionsabout the coupling transition depth do not affect slab surfacetemperatures in the thrust zone. This is expected, since thecoupling transition is always deeper than 30 km and the advec-


tive heat transport of subduction dominates conductive heatloss.

The temperature far below the coupling transition is also largelyindependent of the coupling transition depth. At 240 km depth, allcases show similar temperatures at every subduction zone (Fig. 11).

dx.doi.org/10.1016/j.pepi.2010.02.004




F case,c ed MOb e solidd

Tlf

tg1tttimdoc

T

wlt

T

ig. 6. Pressure–temperature paths of the slab surface for all arc segments in eacholors of the P–T paths and symbols are shown in Fig. 1. Phase diagram for hydratound in minerals. The water-saturated solidus for MORB to ∼3 GPa is shown as thashed white line (Vielzeuf and Schmidt, 2001; Schmidt et al., 2004).

he Pierson’s correlation coefficient, or R value, shows good corre-ation between the slab surface temperatures at 240 km for modelsrom D80 to the X25 case (0.99) and the W1300 case (0.97).

The temperature of slab interior is often described by thehermal parameter, ˚, which is the product of plate age (A), conver-ence velocity (V), and the sine of the slab dip angle (ı) (Kirby et al.,996). Comparing thermal parameter to the slab surface tempera-ure at 30 km and 240 km depth shows that this can also be usedo predict the temperature of the slab surface, but is a poor of slabemperature directly beneath the arc, due to the global variabilityn slab depth beneath arcs (Fig. 12) (Syracuse and Abers, 2006). A

ore accurate predictor of the slab surface temperature at a givenepth than ˚, however, is a closely related value which is basedn the balance between the cooling of the slab as it ages and itsonductive heating from the overlying plate:

slab surface∼ 1√1 + (˚/z)

,

here z is the depth of the slab at which temperature is calcu-ated (Molnar and England, 1995). A comparison to slab surfaceemperature at 30 km depth yields an R value of 0.89 (Fig. 11).


At greater depth, plate reheating must be taken into account.he equation:

Tinterior

Ta= 1 − 2

�e−((�/2.32)2z/˚),

with P–T beneath the arc highlighted. (A) D80; (B) X25; (C) W1300; (D) T550. TheRB (Hacker, 2008) is shown beneath, with shade indicating maximum wt% waterwhite line and the water-saturated solidus for oceanic sediments is shown as the

where Ta is the temperature of the asthenosphere and z is depth,describes the coldest part of the slab interior as it is heated bythe overlying wedge (McKenzie, 1969). We compare the tempera-ture of coldest part of the slab interior (Tinterior) at 240 km depthto the right-hand side of the equation (Fig. 12E); this relationhas an R value of 0.94 and a slope of 1245 ◦C as the estimatedasthenospheric temperature at 240 km depth. Two importantend-members are Cascadia, where the thermal parameter is onequarter that of the next smallest value due to the extremelyyoung age of the slab, and Tonga, where the thermal parame-ter is more than twice that of the next largest value due to theextremely fast convergence rate. If we exclude these two subduc-tion zones we find Ta = 1338 ◦C, since both Cascadia and Tonga haveanomalously cool slab interiors in comparison to their thermalparameters.

3.3. Slab Moho temperature

Pressure–temperature profiles of the slab Moho show a muchlarger range in temperatures beneath the arc than profiles of theslab surface (Fig. 13, Supplementary Materials). The slab Moho


is taken to be 7 km beneath the slab surface, and temperaturesbeneath the arc are calculated normal to the point directly beneaththe arc on the slab surface. These temperatures range from 144 ◦Cto 823 ◦C, with the middle 90% of temperatures between 230 ◦C and734 ◦C, with a mean of 483 ◦C and a standard deviation of 157 ◦C.

dx.doi.org/10.1016/j.pepi.2010.02.004




F e coupe the laf on is b

TtbDos

ig. 7. Histograms of the coupling transition depth, slab surface temperature at thach set of assumptions, with mean values and standard deviation indicated. Noteor T550 corresponds to the segments with cooler slabs where the coupling transiti

his clearly indicates that we should not expect a constant Moho


emperature beneath arcs. A comparison of Moho temperatureseneath the arc for different cases produces R values of at least 0.92.espite differences in Moho temperature for an arc section basedn case, the relative temperature differences between various arcections are similar throughout all cases.

ling transition, and the maximum mantle wedge temperature beneath the arc forrge range in coupling transition depth and maximum mantle wedge temperatureeneath the back arc in this case.

3.4. Mantle wedge temperature


Maximum mantle wedge temperatures vertically beneath thearc vary from 1201 ◦C to 1460 ◦C, with the largest range of maxi-mum temperatures for T550 (Fig. 5). In all cases, the mantle wedgetemperatures beneath the arc are within the range of possible

dx.doi.org/10.1016/j.pepi.2010.02.004





Fig. 8. Pressure–temperature plots for (A) the slab surface directly beneath the arc;(B) the slab Moho beneath the arc; (C) the slab surface at the coupling transition.Red circles indicate models from D80, green triangles indicate models from X25, bluesquares indicate models from W1300, and black inverted triangles indicate modelsfrom T550. A and C are overlain on the hydrated MORB phase diagram (Hacker,2008) with contours showing water content, and B is overlain on the hydratedharzburgite phase diagram (Hacker et al., 2003a). Major dehydration reactions forhydrated harzburgite are: (1) dehydration of serpentine/chlorite/brucite (14.8 wt%water; lowest temperature), (2) serpentine/chlorite/dunite (6.2 wt% water), (3) chlo-rite/harzburgite (1.4 wt% water; highest temperature), and (4) phase A (6.8 wt%water, above 6 GPa). (For interpretation of the references to color in this figurelegend, the reader is referred to the web version of the article.)

Fig. 9. An example from the southern Philippines showing the development of adual-maximum in subarc mantle temperature with a reduction in partial couplingdepth. The P–T profiles represent a variety of partial coupling depths, from 95 km(black) to 120 km (light gray). Solid black lines are temperature contours at 100 ◦Cinterval, black dotted lines indicate the slab surface and the slab Moho, assumedto be 7 km below the slab surface, orange triangles show the location of the arc.Temperature colorscale is the same as in Fig. 5. (For interpretation of the referencesto color in this figure legend, the reader is referred to the web version of the article.)

Fig. 10. Comparisons of partial coupling depths for Tonga. (A) D80 case, partialcoupling transition at 80 km depth. Note the cold nose beneath the forearc. (B) T550case, partial coupling transition at a slab surface temperature of 550 ◦C, which occursbeneath the back-arc. (C) T500 case, with 30 mW/m2 shear heating, partial couplingtransition at a slab surface temperature of 500 ◦C, which occurs in the subarc region.

dx.doi.org/10.1016/j.pepi.2010.02.004


ARTICLE ING Model


14 E.M. Syracuse et al. / Physics of the Earth and

Fig. 11. Slab surface temperatures at 30 km and 240 km from D80 compared to X25

Fig. 12. Comparisons of slab temperatures and thermal parameter-related temperature pr(˚ = VA sin ı), which is considered a proxy for slab temperature; (B) SST versus ˚ at 240the balance of cooling with age and conductive heating at 30 km depth (Molnar and Englproxy for temperature based on plate reheating (McKenzie, 1969); (F) SST directly beneagood estimate of slab surface and interior temperature at a given depth, except near the cbeneath the arc due to the variability in depth to slab. Symbols and colors are shown in F



and W1300, showing the consistency at a given depth among different cases.

oxies. (A) Slab surface temperature (SST) at 30 km for D80 versus thermal parameterkm depth; (C) SST versus ˚ at 100 km depth; (D) SST versus a parameter based onand, 1995); (E) the coldest part of the slab interior at 240 km depth compared to ath the arc versus ˚. Thermal parameter and related temperature proxies provide aoupling transition depth. Thermal parameter is a poor estimate of slab temperatureig. 1.

dx.doi.org/10.1016/j.pepi.2010.02.004




Fig. 13. Pressure–temperature paths of the slab Moho (7 km beneath the slab surface) for all arc segments in each case, with P–T beneath the arc highlighted. (A) D80; (B)X Fig. 1e st tem( .

vKB1

4

4

fitstutttchwTfi

25; (C) W1300; (D) T550. The colors of the P–T paths and symbols are shown int al., 2003a): (1) dehydration of serpentine/chlorite/brucite (14.8 wt% water; lowe1.4 wt% water; highest temperature), and (4) phase A (6.8 wt% water, above 6 GPa)

alues for the melting of hydrated mantle (e.g., Kushiro et al., 1968;ushiro, 1970; Green, 1973; Milhollen et al., 1974; Mysen andoettcher, 1975; Kinzler and Grove, 1992; Hirth and Kohlstedt,996; Grove et al., 2006).

. Discussion

.1. Controls on slab-mantle decoupling

The effect of the location of the transition between partial andull coupling between the slab and the mantle wedge was testedn each of the four cases. The temperature of the slab surface athe transition has a range of over 600 ◦C among different models,trongly suggesting that this transition is not based on tempera-ure alone, and therefore does not follow a simple flow law. Thenrealistic models in the T550 case where the coupling transi-ion is beneath the back arc further emphasizes this point dueo their inconsistencies with seismic attenuation, heat flow, andemperatures necessary for mantle melting. Shear heating was notonsidered in this work, so it is possible that slab temperaturesave been slightly underestimated. For these arcs, its inclusion


ould reduce transition depth, possibly to more realistic values.he amount of shear heating cannot be large given the constraintsrom heat flow and the existence of the cold corner. In general its found that the shear heating along the seismogenic zone cannot

. Black lines show major dehydration reactions for hydrated harzburgite (Hackerperature), (2) serpentine/chlorite/dunite (6.2 wt% water), (3) chlorite/harzburgite

exceed 30 mW/m2 (Peacock and Wang, 1999). In all cases, how-ever, subarc temperatures exceed 1200 ◦C indicating that very lowtemperature (800 ◦C) water-saturated melting does not control thelocation of the volcanic front.

4.2. Subarc wedge temperature

In case X25, the location of the coupling transition is basedon the location of the arc. Of the three cases where mantlewedge temperature is a free parameter (D80, X25, and T550),the maximum mantle wedge temperature beneath the arc hasthe smallest standard deviation for the X25 case, though thedistribution of maximum mantle wedge temperature is also sim-ilar for the D80 and T550 cases. This suggests that there is arange of maximum subarc wedge temperatures, as other geo-physical and geochemical studies have suggested (Leeman et al.,2005; Wiens et al., 2006; Rychert et al., 2008; Cooper et al., inrevision) as opposed to a single critical mantle wedge tempera-ture above which the arc forms (Schmidt and Poli, 1998; Peacock,1996).


4.3. Slab crust and sediment dehydration

Insight to the role of water in subduction zones can be gained bycomparing P–T paths to models of slab petrology. Slab surface tem-

dx.doi.org/10.1016/j.pepi.2010.02.004

ING

P

1 h and

pfDtfewTtTa

ioKtash2aetdIomgpt

masbn2(dfbslfittadi(tcsci

4

p2f(ta

ARTICLEModel


6 E.M. Syracuse et al. / Physics of the Eart

eratures are compared to estimates of maximum water contentsor hydrated MORB, from Hacker (2008) (Fig. 6). For the models in80, X25, W1300, and those in T550, where the coupling transi-

ions occur trenchward of the arc, the slab crust has fully or nearlyully dehydrated by the time it is beneath the arc. On average for thentire set of models, the top of the slab crust has fully dehydratedithin a horizontal distance of 20 km from the coupling transition.

he only model with a slab surface temperature beneath the archat is low enough to have more than 1 wt% water is Tonga in the550 case, which has already been deemed unrealistic due to thenomalously deep coupling transition.

Previous studies have discussed the possibility of melting sed-ments and altered oceanic crust beneath the arc as a mechanismf adding subducted slab material into the mantle wedge (vaneken et al., 2002; Kelemen et al., 2003), following the observa-

ions of subducted component signatures in arc lavas (e.g., Johnsonnd Plank, 1999). The slab surface temperatures obtained in thistudy are close to the solidus for the melting of hydrated MORB andydrated sediments beneath the arc (Fig. 6) (Vielzeuf and Schmidt,001; Schmidt et al., 2004). However, since MORB and sedimentsre likely dehydrated by this point, other scenarios may also be rel-vant. Though the temperatures of the slab crust are lower thanhe dry solidus of MORB beneath most arcs (Yasuda et al., 1994),ry sediments may be hot enough to melt (Nichols et al., 1994).

f water is reintroduced to the slab crust and sediments, meltingf the slab crust may also be possible. Additionally, recent esti-ates from slab fluid samples and water-to-solute ratios from a

lobal range of subduction zones indicate subarc slab surface tem-eratures above the solidi of saturated sediment or basalt, and theransport of slab-derived material to the surface (Plank et al., 2009).

One proposed mechanism of water transport into the deeperantle is the formation of serpentine in the mantle wedge directly

bove its location of release from the slab (Iwamori, 2007) with sub-equent transport by wedge flow to larger depths. This is suggestedy seismic images showing a reduced-velocity layer above theorthern Honshu subduction zone (e.g., Kawakatsu and Watada,007). Serpentine is stable to roughly 600 ◦C between 2 and 5 GPaFig. 13). The slab surface is clearly cooler than the mantle wedgeirectly above it, and given the temperatures present along slab sur-ace, serpentine in the overlying mantle wedge will not be stableeyond 2.5–5 GPa for most arcs. Since the arc lies above where thelab is between 3 and 6 GPa (Figs. 8 and 12), serpentine in the over-ying mantle will not be stable past the arc for any arc in any of theour cases with reasonable coupling depths. This conclusion is sim-lar to that of Wada et al. (2008). Chlorite holds roughly one-tenthhe water of serpentine and is stable to approximate 200 ◦C higheremperatures at subarc pressures (Hacker et al., 2003a). For somercs, chlorite is stable at the pressures and temperatures foundirectly above the slab surface, allowing hydrated mantle overly-

ng the slab to carry less than a few wt% water to subarc depthsGrove et al., 2009). However, the P–T paths for several slabs areoo warm for chlorite stability beneath the arcs, and those that areool enough will cross the maximum pressure at which chlorite istable by 8 GPa, or about 250 km. Therefore, both serpentine andhlorite in the mantle wedge are unlikely hosts for water transportnto the deep mantle.

.4. Slab mantle dehydration

Serpentine in the subducting mantle has been proposed to trans-ort volatiles into the mantle wedge (Ruepke et al., 2002; Peacock,


003). For most arcs in most model cases, the slab Moho is too coldor serpentine to dehydrate by the time it is beneath the arc (Fig. 13)Hacker et al., 2003a). For some warmer subduction zones such ashe Cascades and Mexico, or arcs with a depth to slab beneath therc larger than ∼140 km such as the Marianas and parts of Central


America, the slab Moho may be hot enough to have partially dehy-drated by the time it has reached beneath the arc. Of course, thisdehydration relies on there being enough water in the slab mantleat shallower depths for serpentinite to have formed before beingsubduction, as has been inferred in Nicaragua (Ranero et al., 2003;Syracuse et al., 2008; Ivandic et al., 2008). In the warmer subduc-tion zones, serpentine in the slab mantle may serve as an additionalwater source to melting in the mantle wedge. In the cooler subduc-tion zones, serpentine stability may entrain water into the deepermantle, at least to 250 km depth, beyond which dense hydrousmagnesium silicates may carry water deeper (Thompson, 1992).

5. Conclusions

Two-dimensional thermal modeling of the global suite of sub-duction zones, with a range of assumptions about coupling betweenthe upper and lower plates, has shown several similarities betweendifferent subduction zones. Cases in which the transition from par-tial to full coupling at the plate interface is either set as a constantdepth, a constant distance from the arc, adjusted such that the max-imum mantle wedge temperature beneath arc is constant amongdifferent arcs, and ones in which the transition is located wherethe slab is at a temperature of 550 ◦C yield a similar range of slabsurface and slab Moho temperatures beneath the arc. However, thetemperature-dependent case results in the cold nose of the mantlewedge to extend to the subarc region for some arcs, indicating thatthe location of the transition cannot be based solely on tempera-ture.

These models provide important insight into the possiblesources of fluid and melt in subduction zones. In general, the slabcrust dehydrates before reaching beneath the arc while the slabmantle is still hydrated, indicating a likely crustal and sedimentsource for volatiles in arc magmas. The mantle wedge is too hotto carry water into the deep mantle via serpentine, The coolerdescending slab mantle can carry water below 200 km depth viaphase A in some subduction zones, carrying several wt% water tothe deep mantle, whereas the overlying mantle wedge is too hot totransport water into the deep mantle.

This study is a first step in a future of more detailed modeling ofthe global subduction system. The modeled thermal structures willbe used in modeling the hydration and water budgets, as well as theflow paths of volatiles and melt throughout subduction zones, lead-ing to a more thorough understanding of relation between inputsand outputs of subduction zones.

Acknowledgements

This research was supported by a collaborative NSF-MARGINSgrant to PvK (OCE-0646757) and GA (OCE-0646632). We thank JunKimura and Junichi Nakajima for discussions and providing the geo-metrical constraints for the Nankai subduction zone, and reviewersJohn Hernlund and Kelin Wang for their helpful comments.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at doi:10.1016/j.pepi.2010.02.004.

References

Abers, G.A., van Keken, P.E., Kneller, E.A., Ferris, A., Stachnik, J.C., 2006. The thermal


structure of subduction zones constrained by seismic imaging: implications forslab dehydration and wedge flow. Earth Planet. Sci. Lett. 241, 387–397.

Bijwaard, H., Spakman, W., Engdahl, E.R., 1998. Closing the gap between regionaland global travel time tomography. J. Geophys. Res. 103, 30055–30078.

Billen, M.I., Gurnis, M., 2001. A low viscosity wedge in subduction zones. Earth Planet.Sci. Lett. 193, 227–236.

dx.doi.org/10.1016/j.pepi.2010.02.004

http://dx.doi.org/10.1016/j.pepi.2010.02.004

ING

P

h and

B

B

C

C

C

C

C

C

D

E

F

G

G

G

G

H

H

H

H

H

H

H

H

H

I

I

J

K

K

K

K

K

K

ARTICLEModel


E.M. Syracuse et al. / Physics of the Eart

lackwell, D.D., Steele, J.L., Frohme, M.K., Murphey, C.F., Priest, G.R., Black, G.L., 1990.Heat-flow in the Oregon Cascade Range and its correlation with regional gravity,Curie-Point depths, and geology. J. Geophys. Res. 95, 19475–19493.

ostock, M.G., Hyndman, R.D., Rondenay, S., Peacock, S.M., 2002. An invertedcontinental Moho and serpentinization of the forearc mantle. Nature 417,536–538.

agnioncle, A.M., Parmentier, E.M., Elkins-Tanton, L.T., 2007. Effect of solid flowabove a subducting slab on water distribution and melting at convergent plateboundaries. J. Geophys. Res. 112, B09402, doi:10.1029/2007JB004934.

lift, P., Vannucchi, P., 2004. Controls on tectonic accretion versus erosion in sub-duction zones: implications for the origin and recycling of the continental crust.Rev. Geophys. 42, RG2001, doi:10.1029/2003RG000127.

onder, J.A., Wiens, D.A., Morris, J., 2002. On the decompression melting struc-ture at volcanic arcs and back-arc spreading centers. Geophys. Res. Lett. 29,doi:10.1029/2002GL015390.

onder, J.A., 2005. A case for hot slab surface temperatures in numerical viscousflow models of subduction zones with an improved fault zone parameterization.Phys. Earth Planet. Inter. 149, 155–164.

ooper, L.B., Plank, T., Arculus, R.J., Hauri, R.J., Hall, P.S., Parman, S.W., in revision,High-Ca boninites from the active Tonga Arc. J. Geophys. Res.

uvelier, C., Segal, A., van Steenhoven, A.A., 1986. Finite Element Methods andNavier–Stokes Equations, Reidel Publishing, Dordrecht, Holland, 438 pp.

avies, J.H., Stevenson, D.J., 1992. Physical model of source region of subductionzone volcanics. J. Geophys. Res. 97, 2037–2070.

ngdahl, E.R., van der Hilst, R., Buland, R., 1998. Global teleseismic earthquake relo-cation with improved travel times and procedures for depth determination. Bull.Seismo. Soc. Am. 88, 722–743.

urukawa, Y., Uyeda, S., 1989. Thermal state under the Tohoko arc with considerationof crustal heat-generation. Tectonophysics 164 (2–4), 175–187.

erya, T.V., Yuen, D.A., 2003. Rayleigh–Taylor instabilities from hydration and melt-ing propel ‘cold plumes’ at subduction zones. Earth Planet. Sci. Lett. 212, 47–62.

reen, D.H., 1973. Experimental melting studies on a model upper mantle com-position at high pressure under water-saturation and water-undersaturatedconditions. Earth Planet. Sci. Lett. 19 (1), 37–53.

rove, T.L., Chatterjee, N., Parman, S.W., Medard, E., 2006. The influence of H2O onmantle wedge melting. Earth Planet. Sci. Lett. 249, 74–89.

rove, T.L., Till, C.B., Lev, E., Chatterjee, N., Medard, E., 2009. Kinematic variables andwater transport control the formation and location of arc volcanoes. Nature 459,694–697.

aberland, C., Rietbrock, A., 2001. Attenuation tomography in the western centralAndes: a detailed insight into the structure of a magmatic arc. J. Geophys. Res.106, 11151–11167.

acker, B.R., Abers, G.A., Peacock, S.M., 2003a. Subduction factory. 1. Theoreticalmineralogy, densities, seismic wave speeds, and H2O contents. J. Geophys. Res.108, 2029, doi:10.1029/2001JB001127.

acker, B.R., Peacock, S.M., Abers, G.A., Holloway, S.D., 2003b. Subduction factory. 2.Are intermediate-depth earthquakes in subducting slabs linked to metamorphicdehydration reactions? J. Geophys. Res. 108, 2030, doi:10.1029/2001JB001129.

acker, B., 2008. H2O subduction beyond arcs. Geochem. Geophys. Geosyst. 9 (3),Q03001, doi:10.1029/2007GC001707.

ashida, T., 1989. 3-Dimensional seismic attenuation structure beneath the Japaneseislands and its tectonic and thermal implications. Tectonophysics 159, 163–180.

irose, F., Nakajima, J., Hasegawa, A., 2007. Three-dimensional seismic velocitystructure and configuration of the Philippine Sea slab in southwestern Japanestimated by double-difference tomography. J. Geophys. Res. 113, B09315,doi:10.1029/2007JB005274.

irth, G., Kohlstedt, D.L., 1996. Water in the oceanic upper mantle: implications forrheology, melt extraction and the evolution of the lithosphere. Earth Planet. Sci.Lett. 144, 93–108.

yndman, R.D., Wang, K., 1995. The rupture zone of Cascadia great earthquakes fromcurrent deformation and the thermal regime. J. Geophys. Res. 100, 22133–22154.

yndman, R.D., Peacock, S.M., 2003. Serpentinization of the forearc mantle. EarthPlanet. Sci. Lett. 212, 417–432.

wamori, H., 2007. Transportation of H2O beneath the Japan arcs and its implicationsfor global water circulation. Chemical Geology 239, 182–198.

vandic, M., Grevemeyer, I., Berhorst, A., Flueh, E.R., McIntosh, K.D., 2008. Impact ofbending related faulting on the seismic properties of the incoming oceanic plateoffshore of Nicaragua. J. Geophys. Res. 113, B05410, doi:10.1029/2007JB005291.

ohnson, M.C., Plank, T., 1999. Dehydration and melting experiments con-strain the fate of subducted sediments. Geochem. Geophys. Geosyst. 1,doi:10.1029/1999GC000014.

arato, S., Wu, P., 1993. Rheology of the upper mantle—a synthesis. Science 260,771–778.

awakatsu, H., Watada, S., 2007. Seismic evidence for deep-water transportation inthe mantle. Science 316, 1468–1471.

elemen, P.B., Rilling, J.L., Parmentier, E.M., Mehl, L., Hacker, B.R., 2003. Thermalstructure due to solid-state flow in the mantle wedge beneath arcs. In: Eiler,J. (Ed.), Inside the Subduction Factory. Amer. Geophys. Union, Washington, DC,pp. 293–311.

incaid, C., Sacks, I.S., 1997. Thermal and dynamical evolution of the upper mantle


in subduction zones. J. Geophys. Res. 102, 12295–12315.inzler, R.J., Grove, T.L., 1992. Primary magmas of mid-ocean ridge basalts. 1. Exper-

iments and methods. J. Geophys. Res. 97, 6885–6906.irby, S.H., Stein, S., Okal, E.A., Rubie, D.C., 1996. Metastable mantle phase transfor-

mations and deep earthquakes in subducting oceanic lithosphere. Rev. Geophys.34, 261–306.

PRESSPlanetary Interiors xxx (2010) xxx–xxx 17

Kneller, E.A., van Keken, P.E., Karato, S., Park, J., 2005. B-type olivine fabric in themantle wedge: Insights from high-resolution non-Newtonian subduction zonemodels. Earth Planet. Sci. Lett. 237, 781–797.

Kneller, E.A., van Keken, P.E., Katayama, I., Karato, S., 2007. Stress, strain, andB-type olivine fabric in the fore-arc mantle: sensitivity tests using high-resolution steady-state subduction zone models. J. Geophys. Res. 112, B04406,doi:10.1029/2006JB004544.

Kushiro, I., Syono, Y., Akimoto, S., 1968. Melting of a peridotite nodule at high pres-sures and high water pressures. J. Geophys. Res. 73, 6023–6029.

Kushiro, I., 1970. Stability of amphibole and phlogopite in the upper mantle. In:Carnegie Institute of Washington Yearbook, pp. 245–247.

Langmuir, C., Klein, E.M., Plank, T., 1992, Petrological systematics of mid-ocean ridgebasalts: constraints on melt generation beneath ocean ridges, In: Phipps Morgan,Blackman, Sinton (Eds.), Mantle Flow and Melt Generation at Mid-Ocean Ridges,vol. 71, AGU Monograph, pp. 183–280.

Lee, C.Y., King, S.D., 2009. Effect of mantle compressibility on the thermal and flowstructures of the subduction zones. Geochem. Geophys. Geosyst. 10, Q01006.

Leeman, W.P., Lewis, J.F., Evarts, R.C., Conrey, R.M., Streck, M.J., 2005. Petrologicconstraints on the thermal structure of the Cascades arc. J. Volcanol. Geotherm.Res. 140, 67–105.

Kohlstedt, D.L., Evans, B., Mackwell, S.J., 1995. Strength of the lithosphere:Constraints imposed by laboratory experiments. J. Geophys. Res. 100 (B9),17587–17602.

McKenzie, D., 1969. Speculations on the consequences and causes of plate motion.Geophys. J. Roy. Astr. S. 18, 1–32.

Milhollen, G.L., Irving, A.J., Wyllie, P.J., 1974. Melting interval of peridotite with 5.7percent water to 30 kilobars. J. Geol. 62, 575–587.

Molnar, P., England, P., 1995. Temperatures in zones of steady-state underthrustingof young oceanic lithosphere. Earth Planet. Sci. Lett. 131, 57–70.

Mysen, B.O., Boettcher, A.L., 1975. Melting of a hydrous mantle. I. Phase relations ofnatural peridotite at high pressures and temperatures with controlled activitiesof water, carbon dioxide, and hydrogen. J. Petrol. 16, 520–548.

Nichols, G.T., Wyllie, P.J., Stern, C.R., 1994. Subduction zone-melting of pelagic sed-iments constrained by melting experiments. Nature 371, 785–788.

Peacock, S.M., 1996. Thermal and petrologic structure of subduction zones(overview). In: Bebout, G.E., Scholl, D.W., Kirby, S.H., Platt, J.P. (Eds.),Subduction Top to Bottom. Am. Geophys. Union, Washington, DC,pp. 119–134.

Peacock, S.M., Wang, K., 1999. Seismic consequences of warm versus cool subductionmetamorphism: examples from southwest and northeast Japan. Science 286,937–939.

Peacock, S.M., 2003. Thermal structure and metamorphic evolution of subduct-ing slabs. In: Eiler, J. (Ed.), Inside the Subduction Factory. Am. Geophys. Union,Washington, DC, pp. 293–311.

Perez-Campos, X., Kim, Y., Husker, A., Davis, P.M., Clayton, R.W., Iglesias, A., Pacheco,J.F., Singh, S.K., Manea, V.C., Gurnis, M., 2008. Horizontal subduction and trunca-tion of the Cocos Plate beneath central Mexico. Geophys. Res. Lett. 35, L18303,doi:10.1029/2008GL035127.

Plank, T., Cooper, L.B., Manning, C.E., 2009. Emerging geothermometers for estimat-ing slab surface temperatures. Nat. Geosci. 2, 611–615.

Pollack, H.N., Hurter, S.J., Johnson, J.R., 1993. Heat flow from the Earth’s interior:analysis of the global data set. Rev. Geophys. 31, 267–280.

Ranero, C.R., Morgan, J.P., McIntosh, K., Reichert, C., 2003. Bending-related fault-ing and mantle serpentinization at the Middle America trench. Nature 425,367–373.

Rondenay, S., Abers, G.A., van Keken, P.E., 2008. Seismic imaging of subduction zonemetamorphism. Geology 36, 275–278, doi:10.1130/G24112A.1.

Ruepke, L.H., Morgan, J.P., Hort, M., Connolly, J.A.D., 2002. Are the regional variationsin Central American arc lavas due to differing basaltic versus peridotitic slabsources of fluids? Geology 30, 1035–1038.

Rychert, C., Fischer, K.M., Abers, G.A., Plank, T., Syracuse, E., Protti, J.M., Gonzalez,V., Strauch, W., 2008. Stong along-arc variation in attenuation in the man-tle wedge beneath Costa Rica and Nicaragua. Geochem. Geophys. Geosyst. 9,Q10S10, doi:10.1029/2008GC002040.

Schmidt, M.W., Poli, S., 1998. Experimentally based water budgets for dehydratingslabs and consequences for arc magma generation. Earth Planet. Sci. Lett. 163,361–379.

Schmidt, M.W., Vielzeuf, D., Auzanneau, E., 2004. Melting and dissolution of sub-ducting crust at high pressures: the key role of white mica. Earth Planet. Sci.Lett. 228, 65–84.

Scholl, D.W., von Huene, R., 2007. Crustal recycling at modern subduction zonesapplied to the past—issues of growth and preservation of continental basementcrust, mantle geochemistry, and supercontinent reconstruction. In: Hatcher,Carlson, McBride, Martinez Catalán (Eds.), 4D Framework of Continental Crust,vol. 200. Geol. Soc. Monograph, pp. 9–33.

Schurr, B., Asch, G., Rietbrock, A., Trumbull, R., Haberland, C., 2003. Complex patternsof fluid and melt transport in the central Andean subduction zone revealed byattenuation tomography. Earth Planet. Sci. Lett. 215, 105–119.

Stachnik, J.C., Abers, G.A., Christensen, D.H., 2004. Seismic attenuation and man-tle wedge temperatures in the Alaska subduction zone. J. Geophys. Res. 109,


B10304, doi:10.1029/2004JB003018.Stein, C.A., Stein, S., 1992. A model for the global variation in oceanic depth and

heat-flow with lithospheric age. Nature 359, 123–129.Suckale, J., Rondenay, S., Sachpazi, M., Charalampakis, M., Hosa, A., Royden, L.H.,

2009. High-resolution seismic imaging of the western Hellenic subduction zoneusing teleseismic scattered waves. Geophys. J. Int. 178, 775–791.

dx.doi.org/10.1016/j.pepi.2010.02.004

ING

P

1 h and

S

S

S

T

TT

T

v

v

Yoshimoto, K., Wegler, U., Korn, A., 2006. A volcanic front as a boundary of seismic-

ARTICLEModel


8 E.M. Syracuse et al. / Physics of the Eart

yracuse, E.M., Abers, G.A., 2006. Global compilation of variations in slab depthbeneath arc volcanoes and implications. Geochem. Geophys. Geosyst. 7, Q05017,doi:10.1029/2005GC001045.

yracuse, E.M., Abers, G.A., Fischer, K.M., MacKenzie, L., Rychert, C., Protti, M., Gon-zalez, V., Strauch, W., 2008. Seismic tomography and earthquake locations inthe Nicaraguan and Costa Rican upper mantle. Geochem. Geophys. Geosyst. 9,Q07S08, doi:10.1029/2008GC001963.

yracuse, E.M., Abers, G.A., 2009. Systematic biases in subduction zone hypocenters.Geophys. Res. Lett. 36, L10303, doi:10.1029/2009GL037487.

akanami, T., Sacks, I.S., Hasegawa, A., 2000. Attenuation structure beneath the vol-canic front in northeastern Japan from broad-band seismograms. Phys. EarthPlanet. Inter. 121, 339–357.

hompson, A.B., 1992. Water in the Earth’s upper mantle. Nature 358, 295–302.ichelaar, B.W., Ruff, L.J., 1993. Depth of seismic coupling along subduction zones. J.

Geophys. Res. 98, 2017–2037.sumura, N., Matsumoto, S., Horiuchi, S., Hasegawa, A., 2000. Three-dimensional

attenuation structure beneath the northeastern Japan arc estimated from spec-tra of small earthquakes. Tectonophysics 319, 241–260.

an Keken, P.E., Kiefer, B., Peacock, S.M., 2002. High-resolution models of


subduction zones: Implications for mineral dehydration reactions and the trans-port of water into the deep mantle. Geochem. Geophys. Geosyst. 3, 1056,doi:10.1029/2001GC000256.

an Keken, P.E., Currie, C., King, S.D., Behn, M.D., Cagnioncle, A., He, J., Katz, R.F., Lin,S.-C., Spiegelman, M., Wang, K., 2008. A community benchmark for subductionzone modeling. Phys. Earth Planet. Inter. 171, 187–197.


Vielzeuf, D., Schmidt, M.W., 2001. Melting relations in hydrous systems revisited:application to metapelites, metagreywackes and metabasalts. Contrib. Mineral.Petrol. 141, 251–267.

Wada, I., Wang, K., 2009. Common depth of slab-mantle decoupling: reconcilingdiversity and uniformity of subduction zones. Geochem. Geophys. Geosyst. 10,Q10009, doi:10.1029/2009GC002570.

Wada, I., Wang, K.L., He, J.G., Hyndman, R.D., 2008. Weakening of the sub-duction interface and its effects on surface heatflow, slab dehydration, andmantle wedge serpentinization. J. Geophys. Res. 113, B04402, doi:10.1029/2007JB005190.

Widiyantoro, S., van der Hilst, R., 1996. Structure and evolution of lithospheric slabbeneath the Sunda Arc, Indonesia. Science 271, 1566–1570.

Wiens, D.A., Kelley, K.A., Plank, T., 2006. Mantle temperature variationsbeneath back-arc spreading centers inferred from seismology, petrology, andbathymetry. Earth Planet. Sci. Lett. 248, 30–42.

Yasuda, A., Fujii, T., Kurita, K., 1994. Melting phase-relations of an anhydrous mido-cean ridge basalt from 3 to 20 GPa—implications for the behavior of subductedoceanic-crust in the mantle. J. Geophys. Res. 99, 9401–9414.


attenuation structures in northeastern Honshu, Japan. Bull. Seismo. Soc. Am. 96,637–646.

Zhao, D.P., Hasegawa, A., Horiuchi, S., 1992. Tomographic imaging of P-wave andS-wave velocity structure beneath northeastern Japan. J. Geophys. Res. 97,19909–19928.

dx.doi.org/10.1016/j.pepi.2010.02.004

Documents

GModel PEPI-5250; No.of Pages18 ARTICLE IN PRESS Physics ... · encompassing virtually all subduction zone segments with well-constrained geometry, largely following the segments