Relating rapid plate-motion variations to plate-boundary forces in …people.rses.anu.edu.au/iaffaldano_g/CONTENT/ib_tectono... · 2010. 8. 24. · Lithgow-Bertelloni and Richards

Tectonophysics 474 (2009) 393–404

Contents lists available at ScienceDirect

Tectonophysics

j ourna l homepage: www.e lsev ie r.com/ locate / tecto

Relating rapid plate-motion variations to plate-boundary forces in global coupledmodels of the mantle/lithosphere system: Effects of topography and friction

Giampiero Iaffaldano a,⁎, Hans-Peter Bunge b

a Department of Earth and Planetary Sciences, Harvard University, Cambridge, United Statesb Geophysics Section, Department of Earth and Environmental Sciences, Ludwig-Maximilians University, Munich, Germany

⁎ Corresponding author.E-mail address: [email protected] (G. Iaffa

0040-1951/$ – see front matter. Published by Elsevier Bdoi:10.1016/j.tecto.2008.10.035

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

Recent high-resolution mod

Received 12 March 2008Received in revised form 4 October 2008Accepted 23 October 2008Available online 8 November 2008

Keywords:Mantle/lithosphere couplingPlate boundary forcesTopography collapsePlateboundary friction

els of past plate motions and their comparisonwith plate motion models inferredfrom space geodetic techniques reveal a number of short-term variations in global plate velocities over thepast 10 Myrs. Such variations serve as powerful probe into the nature and magnitude of plate boundaryforces, because they are unlikely to originate from changes in mantle buoyancy forces, which evolve onlonger time scales. Here we explore the constraints of the velocity record using a novel coupled modeling-approach of global neo-tectonic simulations combined with realistic plate driving forces obtained frommantle circulation models (MCMs) to arrive at simple global budgets of mantle, lithosphere and plateboundary forces. We focus on three plate boundary systems along the Nazca/South America plate margin,the Aleutian trench and the India/Australia plate boundary to show that gravitational spreading from hightopography in the Andes and Tibet contributes substantially to the global plate tectonic force balance andthat this contribution is sufficient to explain some 35% of recent velocity changes over the Earth's surface,including among others the observed 30% convergence reduction between the Nazca/South America plates.Our models make a number of specific predictions such as significant lateral variations in plate couplingforces along a given margin revealed by trench-parallel gravity and bathymetry anomalies and theoccurrence of large earthquakes, as well as differences by as much as a factor of five from margin to margin.They also support the notion of a relatively young plate boundary separating the India and Australia plates,which has been previously suggested based on independent observations. Importantly, we find that themodeled Nazca/South America convergence reduction explains recent spreading-rate variations in the SouthAtlantic and South Pacific, which points to the importance of far field effects on the adjacent continents inexplaining the spreading record of oceanic basins. Our numerical results demonstrate (a) that detailedbudgets of forces acting upon plates can be obtained and (b) support the notion of strong forcing along weakplate boundaries.

Published by Elsevier B.V.

1. Introduction

Since the advent of the theory of Plate Tectonics geodynamicistshave sought to quantify the forces involved in driving plate motions(Forsyth and Uyeda, 1975). However, despite considerable progressover the past decades, our understanding of the basic forces that driveand resist plate movements remains limited. Although there is a long-standing knowledge of the history of platemotions (Gordon and Jurdy,1986; Lithgow-Bertelloni and Richards, 1998), geodynamicists cannotyet give quantitative answers to questions of fundamental interest togeologists: What resisting forces act along plate boundary faults? — Areplate boundary forces comparable in magnitude to the buoyancy forcesdriving plate motions? — Why do plate motions change suddenly?

The difficulty in answering these questions stems from the factthat the force balance in plate tectonics is poorly known. That plate

ldano).

.V.

motions are driven by convection in the Earth's mantle is widelyagreed (Hager and O'Connel, 1981; Davies and Richards,1992), but therelative magnitudes of other driving and resisting forces – especiallyalong plate margins – remain unclear. Geodynamicists have longknown that a fuller understanding on the dynamics of how platesmove can be derived from reconstructions of past plate motions, andmany studies have exploited the record of plate motions ingeodynamic models. Lithgow-Bertelloni and Richards (1998) sum-marize the kinematics of global plate movements over the past120 Myrs, paying special attention to changes in the character of platemotions and plate-driving forces.

Temporal variations in plate motions (magnitude and direction ofplate velocities) have great potential to constrain the budget of forcesacting upon plates. The ability to study past as well as present platevelocities is crucial, because changes in plate motion are necessarilydriven by changes in one or more driving or resisting forces, whichmay be inferred from independent observations. For small plates,where mantle-related driving forces are expected to be small, such

mailto:[email protected]://dx.doi.org/10.1016/j.tecto.2008.10.035http://www.sciencedirect.com/science/journal/00401951

Fig. 1. Observed oceanic spreading half-rates for the past 180 Myrs after a recent global compilation by Müller et al. (2008). Plate boundaries are in white, continents in dark gray.Abrupt changes in spreading rates reveal short-term variations in global plate motions, particularly visible in the South Atlantic (a) as well as in the Indian Ocean (b). Such rapidvariations are unlikely to originate from changes in mantle driving forces, which occur on a longer time scale on the order of 50 to 100 Myrs as indicated by global mantle circulationmodels. Instead, they are related to short-termvariations in plate boundary forces caused, for example, by rapid growth of surface topography at convergentmargins (see text). Theseobservations point to the first-order importance of plate boundary forces in controlling global plate motions.

Fig. 2. Generalized power-law rheology for a viscous, non-Newtonian fluid. In therelation between stress σ and strain rate ε ̇, viscosity μ depends on temperature T, strainrate and depth z through parameters A, B, C, γ, and n. Different classes of rheology areshown in different colors, depending on the values of the power-law exponent n. Stressremains either large (blue curve) or diverges to infinity (red and black curves) atincreasing strain rate for most values of n. Note that a narrow range of values, between−1 and zero, accommodates large strain rates at low stress levels that rapidly decreaseto zero. Convection models incorporating this so-called pseudo-stick–slip or self-lubrication rheology produce plate like behavior. However, the peculiar power lawexponent is at odds with results from laboratory experiments on olivine, which find n inthe range 2 to 5. This points to the importance of brittle failure and faults in thegeneration of plate tectonics, which cannot be accommodated by fluid rheologies. (Forinterpretation of the references to color in this figure legend, the reader is referred tothe web version of this article.)

394 G. Iaffaldano, H.-P. Bunge / Tectonophysics 474 (2009) 393–404

temporal variations have the potential to shed key-insight into thenature and magnitude of forces that act along plate boundaries.

The recent advent of geodetic techniques in the geosciences(Dixon,1991) provides a powerful new tool to illuminate plate motionchanges, and increasingly accurate estimates of global plate motionsover the past 10 years or so are now available from geodesy (Sellaet al., 2002). These estimates come in addition to paleomagneticreconstructions of past plate motions derived from the record ofmagnetic isochrons of the ocean floor. Since the earlier studies, whichresolved global plate motions at intervals of 10 Myrs or so(Engebretson et al., 1984; Gordon and Jurdy, 1986), the temporalresolution of such reconstructions has increased dramatically. Fig. 1shows ocean floor spreading half-rates over the past 180 Myrs takenfrom the recent global compilation of Müller et al. (2008) on the basisof marine magnetic anomaly identifications and following the tech-niques outlined by Müller et al. (1997) and Torsvik et al. (2008). Themodel provides a temporal resolution on the order of 1 to 2 Myrs andreveals a number of abrupt velocity-changes over time-periods of afew Myrs or less, which are particularly prominent in the SouthernAtlantic (Fig. 1a) and in the Indian Ocean (Fig. 1b). Their shortduration makes it difficult to attribute these changes to variations inthe internal mantle buoyancy forces, which evolve on longer timescales on the order of 50 to 100 Myrs as suggested by mantlecirculation modeling (Bunge et al., 1998). It is on the other hand likelythat they are caused by short-term variations in plate boundary forces(Iaffaldano et al., 2006).

1.1. Geodynamic plate modeling

Models of past plate motions such as the reconstructions byMülleret al. (2008) place important constraints on geodynamic models. Inparticular, they require that high strain rates and low resistive stressescoexist along plate margins, so that local geological processes mayrapidly build enough forcing to result in short-term variations of platemotions. Unfortunately, realistic inclusion of rigid plates in geody-namic models persists as a serious challenge to geodynamicists

because it is difficult to simulate shear failure along plate boundaries.One strategy is tomodel knownplate structures and their influence onmantle flow by specifying regions that move in a plate-like manner(Hager and O'Connel, 1981; Davies, 1989; Ricard and Vigny, 1989;Gable et al., 1991; Lithgow-Bertelloni and Richards, 1995). Such

Fig. 3. Normalized strength/depth profiles for a simplified two-layer continentallithosphere, plotted for three different fault friction coefficients and a strain rate in therange of 10−15 to 10−13 l/s, typical of plate tectonics. Laboratory experimentsperformed mainly on quartzite and olivine (see text) indicate that lithosphere strengthincreases linearly with overburden pressure in the upper brittle part, and decreasesexponentially with increasing temperature in the lower ductile part. Note that thebrittle part contributes a significant portion (70 to 80%) of the total lithosphericstrength, independently of the particular choice of friction coefficient and strain ratevalues. Experimental results suggest a friction coefficient around 0.6 (dashed envelope)for rocks under lithostatic pressure conditions, known as Byerlees law. However, variousindependent evidence for weak faults suggestmuch lower values between 0.01 and 0.15(solid envelopes), so that plate boundaries experience considerably lower stresses evenfor high strain rates.

Fig. 4. Computational grid used in our thin-shell numerical models of global lithospheredynamics. Finite elements are plotted in thin black. Continents are in dark gray, view iscentered on Europe. The global plate configuration is explicitly built into thecomputational grid through contact-elements (bold black lines). Plate boundariesfeature a depth dependent rheology, with a friction coefficient ranging between 0.01and 0.15, and dislocation-creep parameters inferred from laboratory experimentsperformed on olivine. Non-faulted lithosphere also features a depth dependentrheology, with friction coefficients in the range of 0.6 to 0.85, according to Byerlee'slaw for brittle failure of rocks.

395G. Iaffaldano, H.-P. Bunge / Tectonophysics 474 (2009) 393–404

models demonstrate that plate motions are determined by thedistribution of buoyancy forces in the mantle and, conversely, thatthe mantle buoyancy distribution is very strongly influenced byexisting plate geometry (Richards and Engebretson, 1992; Zhang andChristensen, 1993; Steinberger, 2000; Bunge et al., 2002; McNamaraand Zhong, 2005).

An alternative approach is through highly non-linear (non-New-tonian) viscous creep, strain-rate weakening rheologies, and visco-plastic yielding (Weinstein and Olson, 1992; Bercovici, 1993; Zhong etal., 1998; Trompert and Hansen, 1998; Tackley, 2000; Richards et al.,2001). Moresi and Solomatov (1998) explored the effects of stronglytemperature-dependent viscosity combined with a plastic yieldstress: the former causes the cold upper boundary layer (lithosphere)to be strong, while the latter allows the boundary layer to fail locally inregions of high stress. The success of these models, measured througha so-called “plateness”, is evident when rheologies with extremestrain softening are applied (Bercovici, 2003). One such rheology, inwhich both viscosity and stress decrease with increased strain rate,is called pseudo-stick–slip, or self-lubrication (Bercovici, 1995). Theessence of this behavior is summarized in Fig. 2. Stress σ is propor-tional to strain rate ε̇ through a viscosity μ that depends on tempera-ture T, strain rate and depth z; where A, B, C, γ, and n parameterizethe dependence. However, it is evident from Fig. 2 that self-lubricationarises within a narrow band of power-law exponents ranging between−1 and zero. These values are in poor agreement with laboratoryexperiments of ductile deformation performed on olivine, which findn in the range 2 to 5 (Kirby, 1983).

The challenge to develop plate-like behavior in convection modelsreflects the difficulty to account for brittle failure and reactivation ofpre-existing faults in the uppermost cold region of the lithosphere,which are thought to be of first-order importance in controllinglithosphere strength (Gurnis et al., 2000). Fig. 3 shows normalizedstrength envelopes of a simplified two-phase continental lithosphere.The behavior of strength with depth is parameterized through empi-rical laws that are determined from laboratory experiments performedmainly on quartzite, abundant in the upper 20 km of continental

lithosphere, and olivine, which dominates at greater depths. Strengthincreases linearly with overburden pressure in the upper, brittle partof the lithosphere; it then decreases exponentially with increasingtemperature in the lower, ductile part. The ductile deformation istypically parameterized through a simplification of the generalizedpower-law rheology for viscous flow (Kohlstedt et al., 1995). Althoughit is technically difficult to determine the parameter values fromexperiments (due to the practical limitation of operating in the labo-ratory at strain rates in the range between 10−15 and 10−13 l/sappropriate for plate tectonics), extrapolation from measurements athigher strain rates on the order of 10−4 l/s provides some guidance forappropriate ductile flow laws.

The high strength in the upper part of the lithosphere expressesthe resistance of rocks at low temperature to break, or slide past eachother when already faulted. Experimental results indicate a simplelinear relationship to parameterize this behavior (Byerlee, 1978),where shear stress is proportional to the effective normal pressurethrough a friction coefficient on the order of 0.6 (see dashed envelopein Fig. 3). There is, however, mounting evidence for significantly lowervalues (in the range 0.01 to 0.15, see solid envelopes in Fig. 3) alongfaults and, more generally, along plate boundaries (Hickman, 1991;Bird, 1998; Suppe, 2007). Time-dependency of friction caused, forexample, by pore-pressure and temperature variations may in partexplain these differences (Rice, 1992, 2006) and there has recentlybeen much effort to unify under a single, comprehensive theoreticaldescription the frictional behavior of faults over a range of spatial andtemporal scales (Cox, 2002). It is nonetheless important to point outthat the brittle portion of plate boundaries, which exhibits low frictionand contributes 70 to 80% of the total strength, experiences relativelylow stresses independently of the strain rate.

Fig. 5. Temperature distribution in the Earth's mantle from a recent, high-resolution 3-Dglobal circulation model (see Table 1). Blue represents cold, denser material whereasred is hot and buoyant mantle. View is on Africa, coastline is in white, with continentaltopography in transparent green color scale. Present-day plate boundaries are outlinedin blue. More than 100million grid points discretize the Earth's mantle, equivalent to anaverage grid spacing of 20 km or less. Circulation models include radial variations inmantle viscosity (factor 40 increase from the upper to the lower mantle), internal heatgeneration from radioactivity, bottom heating from the core, and a history of subductionspanning the past 120 Myrs. A cold downwelling is visible beneath Tibet where theancient Tethys Ocean subducted under Eurasia; a hot and buoyant upwelling is visibleas well beneath the spreading triple-junction of the Antarctica, Africa and Australiaplates. The circulation model provides a realistic, first-order estimate of internalbuoyancy forces driving global plate motions. (For interpretation of the references tocolor in this figure legend, the reader is referred to the web version of this article.)


1.2. Models of Neo-tectonics and the large-scale circulation of the mantle

Geodynamicists have introduced weak zones at the surface ofmantle convection models in an attempt to account for brittle failurein the lithosphere (Davies, 1989; King and Hager, 1990). The logicaldevelopment of this approach is the inclusion of discontinuitiesdirectly into the computational grid and the representation of faultsthrough contact-elements. This has been done, for example, in themodeling work of Zhong and Gurnis (1995) and the global neo-tectonic model of Kong and Bird (1995). The neo-tectonic models inparticular have reached a high level of maturity allowing them toaccount for surface topography, regional variations of lithospheredensity and thickness according to either Pratt or Airy isostaticcompensation (Lithgow-Bertelloni and Guynn, 2004), thermal regimeof the lithosphere – based on heat flow measurements and crustalradioactive decay – and importantly for realistic plate configurations(Richardson and Coblentz, 1994; Bird, 1998). The models typically usefinite-element formulations to solve the equations of mass andmomentum conservation, and to compute the instantaneous forcebalance and associated plate velocities. In some cases they takeadvantage of the so-called thin-sheet approximation to reduce thecomputational complexity from 3-D to 2-D. The use of finite elements,moreover, makes it feasible to implement empirical, depth-dependentrheologies of the lithosphere that account for ductile as well as brittledeformation. In Fig. 4 we show the computational grid that we adoptin the global lithosphere model of Kong and Bird (1995).

Independent of advances in neo-tectonic models there has beenmuch progress in modeling global mantle flow (Tackley et al., 1994;Bunge et al., 1997; Zhong et al., 2000) and the large-scale circulation ofthe mantle. Mantle circulation models account for the dynamic effectsfrom aweak asthenosphere on the horizontal length-scales of the flow(Bunge et al., 1996), and include internal heat generation fromradioactivity, and a significant amount of heat flow from the core, forwhich there is growing evidence (Bunge, 2005; van der Hilst et al.,2007). Combined with constraints on the history of subduction(Richards and Engebretson, 1992) they map temporal variations oflarge-scale mantle flow and allow us to place first-order estimates onthe internal mantle buoyancy forces that drive plates. Importantly,there is now sufficient spatial resolution, due to advances in paralleland network computing (Oeser et al., 2006), to resolve the vigorousconvective regime of the mantle in these models. Fig. 5 shows thetemperature distribution in the mantle from one recent model withmore than 100million grid points, equivalent to a grid point spacing of20 km and less throughout the model mantle.

MCMs provide first-order estimates of mantle buoyancy forces, butdo not account for the complex processes in the lithosphere such asthe brittle failure. Neo-tectonic models on the other hand includestresses originating within the lithosphere as well as realistic plateconfigurations, but need to make assumptions on the mantle buoy-ancy field to complete the force balance. The logical step is mergingthe two model classes to simulate the coupled mantle convection/plate tectonics system, so that key components of the lithosphericforce-balance may be simultaneously accounted for.

Here we couple two of the most advanced numerical models ofmantle flow and lithosphere dynamics. We compute MCMs with theTERRA code (Bunge et al., 1997). Based on an icosahedral finite-element grid, the code exploits a highly efficient multigrid algorithmto solve the elliptic problem arising from the momentum balance.Furthermore, it is fully parallelized and performs well on clusters. Tomodel lithosphere dynamics we take the global thin-sheet modelSHELLS (Kong and Bird, 1995). The code is based on a 2-D triangularglobal grid and provides global plate velocities and associatedequilibrium force-fields. Specifically, coupling between the twocodes is performed by using the asthenospheric flow-field computedwith MCMs as velocity boundary condition in the tectonic models toestimate the basal shear tractions at the bottom of the lithosphere.

Accounting for a simple budget of mantle, lithosphere and plateboundary forces, our models of mantle/lithosphere dynamics thusallow us to evaluate the plate tectonic force-balance and to predictglobal plate velocities, as well as other observables, which can betested explicitly against the geologic record.

In the following, we test predictions of plate velocities with therecord of recent plate motions in the Southern Pacific and theSouthern Atlantic. Furthermore, we predict plate boundary deforma-tion induced by friction variations along the Aleutian trench. Finally,we explore the possible effects of a recent plate-boundary opening inthe Indian Ocean on the convergence of India towards Eurasia.Discussion of the results is provided in the final part of themanuscript.

2. Models and results

2.1. Plate boundary forces along the Nazca/South America margin

The Nazca/South America plate margin is particularly suitable totest plate boundary forces, as paleomagnetic (Gordon and Jurdy,1986;DeMets et al., 1994) and geodetic (Norabuena et al., 1999) data indi-cate a significant decline (by some 30%) in convergence velocity overthe past 10 Myrs. The reduction occurred after a period of rapidlyincreasing plate convergence between 40 and 20 Myrs ago (Somoza,1998), often invoked as a key factor in controlling the initiation ofcrustal shortening (Hindle et al., 2002) and Andean uplift (Pardo-Casas and Molnar, 1987; Sobolev and Babeyko, 2005). The timingof the recent slow down is significant in that it is coeval with majorgrowth of the Andes inferred from a variety of independent data(Gregory-Wodzicki, 2000; Ghosh et al., 2006).

Table 1Parameters used in mantle circulation models (MCMs).

Parameter Value Unit

Outer shell radius 6370 kmInner shell radius 3480 kmNumerical grid point resolution (surface) 20 kmNumerical grid point resolution (CMB) 10 kmTemperature (surface) 300 KTemperature (CMB) 4000 KMantle density (surface) 3500 kg m−3

Mantle density (CMB) 5568 kg m−3

Coefficient of thermal expansion (surface) 4.011⁎10−5 K−1

Coefficient of thermal expansion (CMB) 1.256⁎10−5 K−1

Upper mantle viscosity (UMV) 1.0⁎1021 Pa sLower mantle viscosity 40 UMV Pa sThermal conductivity 6.0 W (m K)−1

Internal heating rate 6.0⁎10−12 W kg−l

Heat capacity 1134 J (kg K)−1

Rayleigh number (based on UMV) 109 Adimensional

Fig. 7. Predicted and observed relative plate motions in the South Atlantic and SouthPacific over the past 10 Myrs for a set of adjacent plate pairs: PA/NZ, NZ/SA, NZ/AN, andSA/AF (abbreviations as in Fig. 6). Black bold segments in the small inset indicatepositions along plate boundaries (thin black) at which relative motions have beencomputed. Observed plate motions (with error bars) inferred from paleomagnetic andgeodetic data are represented by black dots, while empty squares indicate relativemotions predicted from our simulations of the global coupled mantle/lithospheresystem. The models explicitly account for the growth of the Andes over the past10 Myrs, and demonstrate that the relative plate motion record can be entirelyexplained with the history of Andean orogeny. Our simulations thus point to theimportance of far-field effects in plate tectonics, and imply that resisting plate marginforces due to Andean growth account for about 18% of global plate motion changes overthe past 10 Myrs (see Fig. 15).


Iaffaldano et al. (2006) test the effect of topography on plateconvergence by computing plate velocities before and after majorAndean orogeny, assuming mantle shear tractions from a MCM and a(low) fault friction coefficient of 0.03 for the tectonic model. Here wetake the same parameters for the tectonic model, but apply buoyancyforces from a new high-resolution MCM shown in Fig. 5. The finerspatial resolution of MCMs allows us, for the first time, to reach aRayleigh number (based on internal heating) of 109, comparable tothe real dynamic vigor of the Earth mantle, and to improve ourestimate of mantle related buoyancy forces. Detailed parameters ofthe MCM are listed in Table 1.

We perform two distinct simulations of global plate motions toisolate the effect of Andean topography on the Nazca/South Americaconvergence rate. In one simulation we use a finite-element gridfeaturing present-day topography from ETOPO 5 data set (NationalGeophysical Data Center, 1998), as high as 5 km in the central Andes,and compute global plate velocities. We then re-compute velocities ina second simulation accounting for a lower topographic relief ofcontinental South America, based on a reconstruction of Andean

Fig. 6. Predicted plate motions in the Hot spot reference frame from coupled global manpaleotopography 10 Myrs ago (in red) and present-day topography (in blue). Plate boundaAustralia, CA — Caribbean, CO — Cocos, EU — Eurasia, IN — India, NA — North America, NZseparate in the computational grid, based on recent evidence of plate separation (see text andresults in a predicted NZ/SA convergence of 10.1 cm/yr at long 71.5° W, lat 25° S, whereasposition (see text). The rates compare remarkably well with observations inferred from a varito 6.7 cm/yr over the past 10 Myrs. The modeling results suggest that the reduction of NZ/SAload of Andes (see text). Similar numerical models moreover confirm that frictional variationexplain the record of plate motion. (For interpretation of the references to color in this figu

paleo-elevation 10 Myrs ago (Gregory-Wodzicki, 2000). Global platemotions computed for the two scenarios are meant to represent,respectively, conditions 10Myrs ago and at present-day. A comparison

tle convection/lithosphere dynamics simulations, corresponding to assumed Andeanries are in black, continents in gray. AF — Africa, AN — Antarctica, AR — Arabia, AU —— Nazca, PA — Pacific, PH — Philippine, SA — South America. IN and AU are treated asFigs. 13 and 14). Note that a lower paleotopography assumed for the Andes 10Myrs agopresent-day topography results in a predicted convergence of 6.9 cm/yr at the sameety of data, which indicate a 30% reduction of NZ/SA plate convergence from 10.3 cm/yrconvergence is caused by resisting plate-margin forces associated with the topographics along the boundary, arising from variations in trench sediment infill, are insufficient tore legend, the reader is referred to the web version of this article.)

Fig. 8. Lateral variations of plate coupling forces (green line) along the Nazca (NZ) SouthAmerica (SA) plate margin that can be attributed to the growth of the Andes (present-day elevation in gray color scale) over the past 10 Myrs. Plate boundary forcescorresponding to Andean uplift are isolated from two distinct simulations ofmomentum balance in the lithosphere, where we either take present-day topographyas reported in the ETOPO5 data-set and shear tractions fromMCMs, or assume a paleo-reconstruction of Andean topography 10 Myrs ago (see text). Plate boundary forces areobtained as the difference of the two equilibrium force-fields, and represent variationsin the degree of coupling along the margin. Note the strong coupling along the centralportion of the margin as opposed to the northern and southern parts. Red dots indicatelarge (MwN8.0) earthquakes reported since 1555, which tend to occur in regions ofmoderate to low plate coupling. (For interpretation of the references to color in thisfigure legend, the reader is referred to the web version of this article.)

Fig. 9. Plate-coupling force anomalies (blue arrows) exerted by the Nazca (NZ) upon theSouth America (SA) plates. The anomalies are computed by subtracting the trench-average value of plate coupling from the absolute plate coupling force and arise fromstrong along-strike variations in plate coupling following Andean growth. Note the twored arrows where the coupling anomalies change sign from positive to negative values,representing strong, local variations in the effective ability of South America to overridethe Nazca plate. The along-strike variations imply two large torques (in bold red): onecounter-clockwise close to the Bolivian Orocline, the other clockwise located in thesouthern margin. We speculate that these torques contribute to the on-goingdevelopment of the convex profile of the margin, which have been inferred to occursince some 25 Myrs, coeval with the period of crustal shortening and orogeny in theAndes. (For interpretation of the references to color in this figure legend, the reader isreferred to the web version of this article.)


of the computed velocity fields predicts a 30% convergence reductionbetween the Nazca and South America plates (Fig. 6), and agrees wellwith the record of present and past plate motions. Further supportingevidence for the dominant effect of Andean topography on plateboundary forcing along the Nazca/South America margin comes fromstress field measurements. Estimates of the present-day (Heidbach etal., 2007; Zoback, 1992) and paleo-stress (Mercier et al., 1992) fieldalong the Andes both indicate rotation of principal stress axes over theperiod of major uplift. By computing the spatial derivative of thesurface velocity field in our models prior and after topographydevelopment, Heidbach et al. (2008) show that these stress rotationsare well reproduced from our calculations.

It is also evident in Fig. 6 that the total convergence reduction isunevenly partitioned between the Nazca and South America plates.This observation can be understood by recalling that basal drivingshear tractions exerted by the mantle on the lithosphere-base scale tofirst-order with the basal surface area of the plate times its velocity.Because the Nazca plate is smaller than South America, a highervelocity reduction is required in the momentum balance for mantleshear tractions to equilibrate topography-generated plate-boundaryforces along the margin.

Our global models not only provide us with a convergence recordof the Nazca/South America plate pair. They also allow us to predictthe history of relative motion for the adjacent (Pacific, Africa,Antarctica) plates, which are shown in Fig. 7. We plot observed andpredicted histories of relative motion for the Pacific/Nazca, Nazca/South America, Nazca/Antarctica, and South America/Africa platepairs over the past 10 Myrs and find them in remarkable agreement,implying that the resisting forces along the Nazca/South Americaplate boundary are responsible for driving recent relative plate motionchanges also in the Southern Atlantic and Southern Pacific regions.

Fig. 8 shows the spatial distribution of plate boundary forcesassociated with gravitational spreading of the Andes along thewestern margin of South America. We isolate these forces bysubtracting from each other the equilibrium force-fields correspond-ing, respectively, to our simulations before and after Andean growth.Forces are as high as 8⁎1012 N/m and provide an estimate of tectoniccoupling along the margin. Iaffaldano and Bunge (2008) show thattheir magnitude and lateral distribution (in addition to predictingthe recent history of Southern Atlantic and Southern Pacific platemotions) are entirely sufficient to explain the peculiar pattern of trench-parallel bathymetry (+/−2000 m) and gravity (+/−100 mGal)anomalies (Smith and Sandwell, 1997; Sandwell and Smith, 1997;Song and Simons, 2003) observed along the margin. The significantalong-strike coupling variations due to Andean gravitational spreadingrepresent along-strike forcing differences exerted by the overridingupon the subducting plate and, by virtue of third Newton's law, vice-versa. Interestingly enough, the spatial distribution of coupling varia-tions – revealed by subtracting the trench-average force from theabsolute value – implies two large torques exerted by the Nazca uponthe South American plate (Fig. 9). We speculate that these torquescontribute to the on-going development of the convex profile of themargin, inferred to occur since some 25 Myrs ago (Allmendinger et al.,2005), coevalwith the period ofmajor Andeanorogeny, and particularlypronounced around 7 to 9 Myrs ago (Rousse et al., 2002).

Some authors have argued for lateral friction variations, due tolateral variations in sediment thickness, as a primary cause for lateralvariations in plate coupling. The idea is that lack of sediment infillwould increase the effective coefficient of friction along a subductionzone (Lamb and Davies, 2003) and thus be responsible for stronger(resisting) plate boundary forcing. Along the Nazca/South Americaplate margin our models do not confirm this hypothesis. Instead, wefind that the gravitational spreading of the Andes overwhelms theeffects from reasonable trench friction variations in the budget of plateboundary forces, and that unrealistically large trench-friction varia-tions would be needed to entirely explain the plate motions record(Iaffaldano and Bunge, 2008).

Fig. 10. (a) Observed trench-parallel gravity anomalies along the Aleutian margin (inblack), where the Pacific (PA) subducts under the North American (NA) plate.Anomalies as high as +/−60 mGal vary rapidly from east to west, and suggest ashallow origin of the gravity signal. Continental topography is also shown. (b) Sedimentthickness along the Aleutian trench ranging between zero (blue) and 1 km or more(red). Infill of trench sediments may affect friction coefficient along plate boundaries(see text). Note the absence of major topographic features near the plate boundary, sothat friction variations most probably dominate the budget of resisting forces along thetrench. (For interpretation of the references to color in this figure legend, the reader isreferred to the web version of this article.)

Fig. 11. Bathymetry anomalies (computed and observed) along the Aleutian trench.Gray dashed curve shows observed anomalies, obtained by subtracting the averagetrench-depth from the bathymetry. Data have been corrected for sediment thickness sothat they represent true variations of ocean-floor depth along the trench. Bathymetryvariations are moderate compared to other active subduction systems, and rangebetween−0.8 and 0.7 km. Black curve shows trench-depth anomalies predicted from asimple elastic analytical model for the deflection of the overriding plate under theaction of frictional forces exerted by the subducting plate at the interface (see inset andtext for details). The lateral variations in the resisting frictional forces along the trencharise from variations of the friction coefficient, which we assume to decrease linearlywith increasing sediment infill. We associate a coefficient of 0.01 or 0.15 with maximumor minimum sediment infill (see text). Note the good agreement between observed andpredicted trench-bathymetry anomalies.


2.2. Frictional plate-boundary forces along the Aleutian trench

It is nevertheless important to explore the effect of frictionalvariations at convergent boundaries where topography is not adominant feature. One suitable geologic system is the Aleutian trench,where the Pacific plate subducts under North America. Continentaltopography is not strongly pronounced near the Aleutian margin, butthe local sediment thickness varies substantially (between zero andabout 1 km) along the trench (Fig. 10b), as inferred from ship surveys(dataset from National Geophysical Data Center). Furthermore,geodetic observations reveal moderately varying trench-parallelgravity anomalies along the margin (Sandwell and Smith, 1997)ranging between −60 and +60 mGal (Fig. 10a). We note that themagnitude of these gravity anomalies is about half of what is observedalong the Nazca/South America margin. They are, however, suffi-ciently large to point to trench-parallel bathymetry variations andassociated variations in plate coupling as the cause for the observedgravity signal.

We test in Fig. 11 whether friction variations associated withdifferent sediment infill can be used to explain the observed variationsof local trench depth from the Aleutian average depth. Note that bydoing so we implicitly assume the latter to be determined by long-term processes, other than sedimentation, associated with thesubduction of the Pacific plate. Following our work on trench parallelgravity anomalies along the Nazca/South America margin (Iaffaldanoand Bunge, 2008), we isolate the resisting plate-boundary forces duelateral friction variations by subtracting from each other twoequilibrium force-fields computed with our mantle/lithospheremodels: one is associated with homogeneous friction along thetrench, while the second includes friction coefficient variations in

proportion to the sediment infill. Specifically, we assign a frictioncoefficient of 0.01 to maximum infill, whereas the minimum infillcorresponds to a value of 0.15 (Hickman, 1991; Bird, 1998; Suppe,2007).

We find that the lateral variations of frictional forces (platecoupling) computed from our models along the Aleutian trench neverexceed 1.4⁎1012 N/m, a moderate value when compared to otherdriving and resisting plate-boundary forces (Forsyth and Uyeda,1975). Hence we apply our computed force variations to the analyticsolution (Turcotte and Schubert, 2002) describing the deflection of athin, semi-infinite overriding plate, tectonically loaded on one side(see inset in Fig. 11). This allows us to predict trench-depth variationsassociated with bending of the overriding plate under frictionalforcing from the subducting plate. A Young modulus of 65 GPa, aPoisson ratio of 0.25, and an elastic thickness in the range between 15and 20 km are assumed, consistent with published estimates(Caldwell et al., 1976; Bry and White, 2007; Perez-Gussinye et al.,2007). It is important to remind that the moderate (on the order of800 m) trench-parallel bathymetry anomalies along the Aleutiansvanish over a distance of about 120 km in the forearc and are thereforeassociated with curvature variations smaller than 6⁎10−8 m−1. It isthus reasonable in our prediction to neglect bending-momentsaturation (McNutt and Menard, 1982), which typically occurs forcurvatures equal to or greater than 5⁎10−7 m−1 (Levitt and Sandwell,1995). Furthermore, we account for gravitational restoring forces fromthe denser underlying asthenosphere as well as from intrinsicbathymetric variations along the trench associated with the lateralage differences of the subducting plate (Müller et al., 2008), followingthe thermal model for oceanic-lithosphere subsidence of Stein andStein (1992).

The predicted bathymetry anomalies agree well with the observa-tion, as seen from Fig. 11. Note that we correct for sediment thicknessthe original bathymetric data of Smith and Sandwell (1997), so thatthey represent true variations of ocean-floor depth along the trench.Based on the good agreement, we map the lack or excess of massexpressed by our computed bathymetry variations into predictions offree-air gravity anomalies along the trench. Specifically, we integrate a

Fig. 12. Observed (dashed gray) and predicted (solid black) profiles of free-air gravityanomaly along the Aleutian trench. Predicted gravity anomalies are computed byintegrating a Bouguer formula for the density contrast of water versus crust to computethe gravity contribution at sea level of a layer with thickness equal to the computedbathymetry anomaly, shallower or deeper than the average trench depth. Predicted andobserved profiles compare well and support the notion that lateral trench frictionvariations due to lateral variations in sediment infill dominate the budget of plateboundary forces along this margin.

Fig. 13. Observed convergence of India (red) and Australia (blue) relative to fixedEurasia over the past 20 Myrs. Convergence rates are computed through rigid-rotationEuler poles at long 86°E, lat 27°N of the India/Eurasia margin. Present-day values(squares) are derived from geodetic techniques while the paleomagnetic record (solidlines) is computed by averaging finite rotations of magnetic anomalies identified alongthe Carlsberg and South-East Indian ridges (labeled respectively as CBr and SEIr in inset.CIr is Central Indian ridge). Note that convergence rates are very similar between 20 and11 Myrs ago, when India and Australia appear to behave as one single plate withpresumably little internal deformation. The convergence relative to Eurasia differs moredistinctly over the last 11 Myrs, when India and Australia slowed down by almost 2 cm/yr and 0.5 cm/yr, respectively. Inset shows locations of identified unconformities ofsediments (magenta dots) as well as great (MsN6.5) earthquakes (green dots)indicating left-lateral strike–slip motion in the northern portion of the Ninety EastRidge (orange contours). Those evidence suggest diffuse deformation in the IndianOcean particularly pronounced during late Miocene, and have been interpreted asseparation between the India and Australia. Plate boundaries are in white, continentaltopography in gray color-scale. (For interpretation of the references to color in thisfigure legend, the reader is referred to the web version of this article.)


Bouguer formula for the density contrast of water versus crust tocompute the free-air gravity contribution at sea level of a layer withthickness equal to the computed bathymetry anomaly, shallower ordeeper than the average trench depth. That is, we reconstruct thealong-strike free-air gravity contribution from the deflected oceanfloor. Predictions of gravity anomalies are shown in Fig. 12 andcompare well with the geodetic observations of Sandwell and Smith(1997). The result suggests that lateral friction variations are capableindeed of dominating the budget of resisting forces along platemargins when no large topography is present.

2.3. Effects of an intra-plate boundary within the India/Australia plate

While our instantaneous calculations cannot be taken tomodel thetemporal evolution of plate boundaries, they do allow us to test theeffects of variations in plate geometry, and in particular the creation ofnew plate boundaries on global plate motions. The principle of inertiaimplies that any such event would invariably trigger plate motionchanges, because the budget of basal drag and plate boundary forceswould be repartitioned. A recent episode is thought to have occurredin the Indian Ocean, where a variety of evidence have been interpretedas the generation of a diffuse boundary between the India andAustralia plates, dated between 8 (Wiens et al., 1985) and 20 (Gordonet al., 1998) Myrs ago. Weissel et al. (1980) document ocean-floordeformation at about 8 Myrs from buckling of marine sediments.Ongoing deformation in the Indian Ocean is further supported bypronounced (MsN6.5) and localized seismicity (Stein and Okal, 1978)suggestive of left-lateral strike–slip motion along the northern portionof the Ninety East Ridge (see inset in Fig. 13).

We show in Fig. 13 the observed convergence history of India andAustralia relative to Eurasia since early Miocene based on geodetic(Sella et al., 2002) as well as paleomagnetic (Gordon and Jurdy, 1986;DeMets et al., 1994;Cande and Stock, 2004; Merkouriev and DeMets,2006) data collected along the Carlsberg and South East Indian ridges(labeled respectively as CBr and SEIr in inset of Fig. 13). Convergencerates are almost indistinguishable (within error-bars) between 20 and11 Myrs ago, suggesting that India and Australia acted as one singleplate with presumably little deformation occurring in between. Overthe past 11 Myrs however their convergence to Eurasia differsdistinctly. While India slowed down by almost 2 cm/yr, convergenceof Australia to Eurasia remained almost steady, with only some 0.5 cm/yr of reduction. Timing of the India/Eurasia plate-motion change

coincides reasonably well with the occurrence of diffuse deformationin the Indian Ocean. More relevant is the fact that Tibet had attainedmost of its current elevation (Tapponnier et al., 2001) prior to the slowdown of the Indian plate and prior also to the presumed formation ofthe India/Australia plate boundary, implying that resistive plateboundary forces arising from the gravitational load of Tibet werealready in place to act against convergence.

In Fig. 14 we test explicitly whether plate-boundary forces fromhigh Tibet are sufficient to explain the observed reduction of India/Eurasia plate convergence, once the former is separated from Australiaby an additional plate boundary. We perform two distinct simulationsof global platemotions, onewith India and Australia as one single plateand the other with two plates built into the computational finite-element grid. Note that the two simulations differ fromeach other onlyin the presence/absence of the plate boundary between India andAustralia, while all other parameters such as lithosphere structure,asthenospheric mantle flow as well as present-day topography havebeen kept identical. A single India/Australia plate results in a predictedconvergence of 5.2 cm/yr at long 86°E, lat 27°N (Fig. 14a), whereasIndia being separated from Australia implies a convergence of 3.5 cm/yr at the same position (Fig. 14b). The latter prediction comparesremarkably well with the geodetic estimate (see Fig. 13).

In order to verify whether the modeled velocity reduction arisingfrom India/Australia separation is indeed affected by plate boundaryforcing from high Tibet topography, we repeat the same above-mentioned simulations assuming a low (b800 m) topography forTibet. As expected, we find very little convergence reduction (fewmm/yr) upon introduction of the India/Australia plate boundary inthis case. Thus our results suggest that the pronounced convergencereduction of India/Eurasia is best explained by the separation of India

Fig. 14. Predicted India (IN) plate motion relative to Eurasia (EU) from two distinct simulations with India and Australia (AU) acting, respectively, (a) as one single and (b) as twoseparate plates, where plate boundaries in our computational mesh are shown in bold white and finite elements in thin white. Plate motions are computed at long 86°E, lat 27°N.Abbreviations of plate names as in Fig. 6. Note that a single India/Australia plate results in a predicted convergence of 5.2 cm/yr relative to EU, incompatible with the geodeticestimate (see Fig. 12). Two separate plates result in a convergence of 3.5 cm/yr of IN relative to EU, similar to the present-day observation. In the latter scenario, resisting forces fromthe gravitational load of Tibet act only against the smaller India plate, and are thus more effective in slowing the convergent motion.


and Australia occurring at a time when resisting forces from highTibet, possibly exceeding 1⁎1013 N/m, were already in place. Finally, itis worth mentioning that our simulations predict an increasedconvergence between India and Australia, concentrated in the IndianOcean, compatible with the aforementioned geologic and geodeticrecord.

3. Discussion

Our results are interesting because they indicate that joint modelsof the mantle/lithosphere system are beginning to achieve a level ofmaturity that allows them to explicitly test a range of hypotheses onthe force balance in plate tectonics, and to identify key controllingparameters. While buoyancy forces from mantle convection con-tribute significantly to the dynamics of plate motion, our resultsdemonstrate that plate boundary forces are of sufficient magnituderelative to these driving forces to affect plate motions and platedeformation, and to initiate rapid plate motion changes.

One key controlling parameter in regulating plate velocity is thetopography of large mountain belts, because their topographic loadconsumes a considerable amount of the driving forces available inplate tectonics. Along the Nazca/South America plate boundary theseforces are sufficient to reduce the convergence rate over the past10 Myrs by some 30%. This reduction is, however, not an isolatedepisode of a rapid plate motion slow down. Instead many suchvariations are documented from the global compilation of Müller et al.(2008), which points to the importance of topography in the globaltectonic system (Cloetingh et al., 2007). Thus it is reasonable toexplore the spreading record of oceanic basins, in order to testpotential correlations with topographic events as well as withinferences of paleo-altitudes in mountain belts of the surroundingcontinents (Iaffaldano et al., 2007).

The fact that ourmodels accurately predict the spreading history ofthe Pacific/Nazca, Nazca/South America, Nazca/Antarctica, and SouthAmerica/Africa plate boundaries is of equal interest. The result is notentirely surprising. It arises in part from the kinematic constraints ofplate tectonics on the sphere. But on a more fundamental level itreflects the elliptic nature of the momentum equation and the

physical regime of creeping flow, which inherently accommodates far-field effects in tectonic settings. Our results indicate that these far-field effects cannot be ignored in the geologic record at least in somecases.

The strong influence of mountain belts on the plate tectonic forcebalance could have important implications. In an influential paperRaymo and Ruddiman (1992) advanced the notion that Cenozoicclimate change may have been caused by the uplift of Tibet. In otherwords the rise of large mountain plateaus may act as a tectonic forceon climate. Low erosion rates have been implicated as a pre-requisitefor the creation of large mountain plateaus (Sobel et al., 2003), so ourresults suggest conversely that climate may act – through largetopography – as a force in plate tectonics.

It is interesting to note how the distribution of large earthquakesrelates to plate coupling variations, often referred to as asperities(Kanamori, 1986). Song and Simons (2003) pointed to a statisticallysignificant correlation of lateral coupling variations with theoccurrence of major seismic events. Following those efforts, Fulleret al. (2006) showed that forearc-basin sedimentation in the over-riding plate may also affect the occurrence or large subduction zoneearthquakes. Llenos andMcGuire (2007) have noted recently that therupture area of large (Mw 7.5 to 8.7) earthquakes in circum-Pacificsubduction zones coincides with positive gradients of trench parallelgravity anomalies. In Fig. 8 we showed that great (MwN8) earth-quakes reported since 1555 (from the NOAA Significant EarthquakesDatabase and the Global Centroid Moment Tensor catalog) along theNazca/South America margin fall into regions where our modelspredict a moderate to low amount of plate coupling. Moreover, ourcalculations indicate that plate boundary forces vary not onlylaterally along a given margin, but differ also strongly from marginto margin. For instance, plate coupling along the Nazca/SouthAmerica margin due to large Andean topography is on average fivetimes stronger than friction-related coupling along the Aleutiantrench. We thus speculate that strong plate coupling forces, althoughthey would favor frequent seismicity along a given subduction zone,could potentially act to inhibit large earthquakes; possibly due to thefact that it becomes too difficult to rupture over large coherent areasin such regions.


Our models use a history of subduction (Lithgow-Bertelloni andRichards, 1998) to constrain the budget of global mantle buoyancyforces. Since the arrival of global tomographic models more than twodecades ago (Dziewonski, 1984) the ability of seismologists to imagethe 3-D mantle structure has increased dramatically and a variety ofhigh resolution mantle heterogeneity models are now available(Grand et al., 1997; Masters et al., 2000; Montelli et al., 2004; vander Hilst et al., 2007). Simultaneous progress has been achievedrecently in high-pressure mineral physics to model complex thermo-dynamic phase equilibria of mantle mineral assemblages under theelevated temperature and pressure conditions of the deep earth(Stixrude and Lithgow-Bertelloni, 2007; Piazzoni et al., 2007). Thelatter development permits, in principle, to map seismic hetero-geneity directly into temperature and compositional effects (Mataset al., 2007). When combined with seismic normal mode studies onthe mantle density structure (Ishii and Tromp, 2004) these develop-ments would make it possible to derive independent constraints onthe distribution of mantle buoyancy forces that could be tested intectonic studies.

Overall, a significant portion of recent changes in global platemotions can be attributed to topography-related forcing along plateboundaries rather than to mantle buoyancy. We summarize thesefindings by plotting the relative plate motion changes observedglobally over the past 10Myrs in Fig.15. Green and red bars (about 35%of the total change) show variations in plate motion that are related,respectively, to the growth of the high Andes or to the presumedrecent separation between India and Australia. The physical causes fora separation of India and Australia remain to assess, although evidenceof its occurrence are available. Cloetingh andWortel (1985) compute abalance of forces acting along the boundaries of the India/Australiaplate and show that large intra-plate stresses focus in the Indian

Fig. 15. Observed variations of adjacent-plates motions over the past 10 Myrs.Abbreviations of plate names as in Fig. 6. For each couple of adjacent plates, variationsare computed as magnitude of difference between relative rotation poles at 10 Myrs,derived from paleomagnetic reconstruction, and at present-day, obtained throughgeodetic techniques (GPS). The Cocos oceanic plate is not considered, since a geodeticestimate for its rotation pole is not available. Variations of AF/SA, AN/NZ, NZ/SA, andNZ/PA adjacent-plates systems (green bars) can be entirely explained through theeffect of Andean growth (see Figs. 6 and 7). They amount to about 18% of the globalrelative motions changes over the past 10 Myrs. Variations of IN/EU as well as IN/AUrelative motions (red bars) can be entirely explained through the effect of separationbetween India and Australia (see Fig. 14). They amount to about 17% of the globalrelative motions changes over the past 10 Myrs. Thus our models of mantle/lithospheredynamics explicitly predict about 35% of the global plate motion changes observed overthe past 10 Myrs from two well-identified tectonic variations. (For interpretation of thereferences to color in this figure legend, the reader is referred to the web version of thisarticle.)

Ocean in proximity of the Ninety East Ridge. This is further supportedby the record of seismicity showing strike–slip mechanisms in thatregion (Stein and Okal, 1978). Deformation occurring east of theNinety East Ridge is also evident from magnetic anomalies (Gordonet al., 1998) identified along the Central Indian Ridge (labeled as CIr ininset of Fig. 13) as well as seismic reflection profiles showingunconformities of sediment layers in the Bengal Fan (Weissel et al.,1980). Finally, it is worth mentioning that the Wharton Basin, locatedwest of 90°E, hosts a fossil ridge that ceased to spread some 45 Myrsago (Liu et al., 1983) and represent at present-day the youngest andthus thinnest feature in the region. Results from our numericalsimulations are broadly consistent with those findings. We aretempted to speculate that the addition of plate boundary forcesrelated to gravitational spreading of Tibet might have triggered failurein the presumably already-weak Indian Ocean. The notion of a yieldstress in plate tectonics, inherent in this interpretation is born out bygeodynamic models that combine an asthenosphere, temperaturedependent viscosity, and viscoplastic yielding (Richards et al., 2001).

The main shortcoming of our modeling approach is the instanta-neous nature of the calculations; that is we compute the global forcebalance only for a given instant in time. The approach, however, maybe justified given our current limited understanding of the complexprocesses involved in global lithosphere dynamics. Some of theseprocesses include erosion exerted by climate and sedimentation,friction and dip-angle variations along faults as well as plate boundarymigration. While progress is being made to introduce these processesinto 2-D and 3-D geodynamic models (Moresi et al., 2007), it isinstructive to test some simple hypotheses of the plate tectonic forcebalance in the global models studied here.

4. Conclusions

We have introduced a novel modeling approach to globallithosphere dynamics, where neo-tectonic simulations are coupledto shear driving forces obtained from 3-D MCMs to compute self-consistent budgets of forces acting upon tectonic plates. By takingadvantage of the geologic record of plate-motion histories andobserved gravity field, we have computed the force balance forthree different convergent systems. We find that gravitationalspreading of the high Andes along the Nazca/South America marginis entirely sufficient to explain the recent history of plate motions inthe Southern Pacific and Southern Atlantic, which points to theimportance of far-field effects and the influence of continentaltopography on rapid spreading variations in the ocean basins. Ourresults also confirm that strong lateral variations in resisting forces,which represent along-strike differences in plate coupling, are fullycapable to induce the observed large trench-parallel gravity andbathymetry anomalies along the South American margin. Conversely,along the Aleutian trench where high topography is not a majorfeature, the budget of resisting forces is controlled primarily bytrench-friction variations related to sediment infill. In this case,coupling along the margin is 5-fold lower than along the Andes, yetsufficient to deform the margin and generate observable trench-parallel gravity and bathymetry anomalies. Finally, we find that therecent history of India/Eurasia convergence can be linked explicitly tothe separation of the Indian and Australian plates at a time when hightopography in Tibet was already in place. By acting directly against theIndian plate large resisting forces of Tibet are fully capable ofexplaining the recent slowdown of India/Eurasia plate convergence.Overall, the gravitational spreading from high topography in theAndes and Tibet quantitatively explains some 35% of the global platemotion changes recorded since the late Miocene. This remarkablefirst-order result clearly demonstrates the ability of plate boundaryforces to affect the global plate velocity field. The level of maturityachieved by neo-tectonic simulations coupled with 3-D MCMs thusallows geodynamicists to make explicit predictions of the plate


tectonic force balance that can be tested against the geologic record ofpresent and past plate motions.

Acknowledgments

We thank Dietmar Mueller and an anonymous reviewer for carefuland constructive comments. This work was supported by the Elitenet-work of Bavaria and a Reginald A. Daly Postdoctoral Fellowship (G.I.).

References

Allmendinger, R.W., Smalley Jr., R., Bevis, M., Caprio, H., Brooks, B., 2005. Bending theBolivian orocline in real time. Geology 33, 905–908.

Bercovici, D., 1993. A simple model of plate generation frommantle flow. Geophys. J. Int.114, 635–650.

Bercovici, D., 1995. A source-sink model of the generation of plate tectonics from non-Newtonian mantle flow. J. Geophys. Res. 100, 2013–2030.

Bercovici, D., 2003. The generation of plate tectonics from mantle convection, EarthPlanet. Sci. Lett. (Frontiers) 205, 107–121.

Bird, P., 1998. Testing hypotheses on plate-driving mechanisms with global lithospheremodels including topography, thermal structure, and faults. J. Geophys. Res. 103,10115–10129.

Bry, M., White, N., 2007. Reappraising elastic thickness variations at oceanic trenches. J.Geophys. Res. 112, B08414.

Bunge, H.-P., 2005. Low plume excess temperature and high core heat flux inferred fromnon-adiabatic geotherms in internally heated mantle circulation models. Phys.Earth Planet. Inter. 153, 3–10.

Bunge, H.-P., Richards, M.A., Baumgardner, J.R., 1996. The effect of depth dependentviscosity on the planform of mantle convection. Nature 379, 436–438.

Bunge, H.-P., Richards, M.A., Baumgardner, J.R., 1997. A sensitivity study of 3-Dspherical mantle convection at 108 Rayleigh number: effects of depth dependentviscosity, heating mode and an endothermic phase change. J. Geophys. Res. 102,11991–12007.

Bunge, H.-P., Richards, M.A., Lithgow-Bertelloni, C., Baumgardner, J.R., Grand, S.P.,Romanowicz, B.A., 1998. Time scales and heterogeneous structure in geodynamicearth models. Science 280, 91–95.

Bunge, H.-P., Richards, M.A., Baumgardner, J.R., 2002. Mantle-circulation models withsequential data assimilation: inferring present-day mantle structure from plate-motion histories. Philos. Trans. R. Soc. Lond. Ser. A 360, 2545–2567.

Byerlee, J.D., 1978. Friction of rocks. Pure Appl. Geophys. 116, 1189–1198.Caldwell, J.G., Haxby, W.F., Karig, D.E., Turcotte, D.L., 1976. On the applicability of a

universal elastic trench profile. Earth Planet. Sci. Lett. 31, 239–246.Cande, S.C., Stock, J.M., 2004. Pacific–Antarctic–Australia motion and the formation of

the Macquarie plate. Geophys. J. Int. 157, 399–414.Cloetingh, S.A.P.L., Wortel, R., 1985. Regional stress field in the Indian plate. Geophys.

Res. Lett. 12, 77–80.Cloetingh, S.A.P.L., Ziegler, P.A., Bogaard, P.J.F., Andriessen, P.A.M., Artemieva, I.M., Bada,

G., van Balen, R.T., Ben-Avraham, Z., Brun, J.-P., Bunge, H.-P., Burov, E.B., Carbonell, R.,Facenna, C., Gallart, J., Green, A.G., Heidbach, O., Jones, A.G., Matenco, L., Mosar, J.,Oncken, O., Pascal, C., Peters, G., Sli-aupa, S., Soesoo, A., Spakman, W.R., S., Thybo, H.,Torsvik, T.H., de Vicente, G., Wenzel, F., Wortel, M.J.R., the TOPO-EUROPE WorkingGroup, 2007. TOPO-EUROPE: the geoscience of coupled deep earth-surfaceprocesses. Glob. Planet. Change 58, 1–118.

Cox, S.F., 2002. Fluid flow in mid- to deep crustal shear systems: experimentalconstraints, observations on exhumed high fluid flux shear systems, andimplications for seismogenic processes. Earth planets space 54, 1121–1125.

Davies, G.F., 1989. Mantle convection model with a dynamic plate-topography, heat-flow and gravity anomalies. Geophys. J. Int. 98, 461–464.

Davies, G.F., Richards, M.A., 1992. Mantle convection. J. Geol. 100, 151–206.DeMets, C., Gordon, R.G., Argus, D.F., Stein, S., 1994. Effect of recent revisions to the

geomagnetic reversal time scale on estimates of current plate motions. Geophys.Res. Lett. 21, 2191–2194.

Dixon, T.H., 1991. An introduction to the Global Positioning System and some geologicalapplications. Rev. Geophys. 29, 249–276.

Dziewonski, A.M., 1984. Mapping the lower mantle: determination of lateralheterogeneity in P-velocity up to degree and order 6. J. Geophys. Res. 89,5929–5952.

Engebretson, D.C., Cox, A., Gordon, R.G., 1984. Relative motions between oceanic platesof the Pacific basin. J. Geophys. Res. 89, 291–310.

Forsyth, D.W., Uyeda, S., 1975. Relative importance of driving forces of plate motion.Geophys. J. R. Astron. Soc. 43, 163–200.

Fuller, C.W., Willett, S.D., Brandon, M.T., 2006. Formation of forearc basins and theirinfluence on subduction zone earthquakes. Geology 34, 65–68.

Gable, C.W., O'Connell, R.J., Travis, B.J., 1991. Convection in 3 dimensions with surfaceplates — generation of toroidal flow. J. Geophys. Res. 96, 8391–8405.

Ghosh, P., Garzione, C.N., Eiler, J.M., 2006. Rapid uplift of the Altiplano revealed through13C–18O bonds in paleosol carbonates. Science 311, 511–515.

Gordon, R.G., Jurdy, D.M., 1986. Cenozoic global plate motions. J. Geophys. Res. 91,12389–12406.

Gordon, R.G., DeMets, C., Royer, J.-Y., 1998. Evidence for long-term diffuse deformationof the lithosphere of the equatorial Indian Ocean. Nature 395, 370–374.

Grand, S.P., van der Hilst, R.D., Widiyantoro, S., 1997. Global seismic tomography: asnapshot of convection in the Earth. GSA Today 7, 1–7.

Gregory-Wodzicki, K.M., 2000. Uplift history of the central and northern Andes: areview. Geol. Soc. Am. Bull. 112, 1091–1105.

Gurnis, M., Zhong, S., Toth, J., 2000. On the competing roles of fault reactivation andbrittle failure in generating plate tectonics frommantle convection. The History andDynamics of Global Plate Motions Monograph, pp. 73–94.

Hager, B.F., O, ' Connel, R.J., 1981. A simple global-model of plate dynamics and mantleconvection. J. Geophys. Res. 86, 4843–4867.

Heidbach, O., Reinecker, J., Tingay, M., Müller, B., Sperner, B., Fuchs, K., Wenzel, F., 2007.Plate boundary forces are not enough: second- and third-order stress patternshighlighted in the World Stress Map database. Tectonics 26, TC6014.

Heidbach, O., Iaffaldano, G., Bunge, H.-P., 2008. Topography growth drives stress rotationsin the central Andes — observations and models. Geophys. Res. Lett. 35, L08301.

Hickman, S.H., 1991. Stress in the lithosphere and the strength of active faults. Rev.Geophys. 29, 759–775.

Hindle, D., Kley, J., Klosko, E., Stein, S., Dixon, T.H., Norabuena, E.O., 2002. Consistency ofgeologic and geodetic displacements during Andean orogenesis. Geophys. Res. Lett.29, 1–4.

Iaffaldano, G., Bunge, H.-P., Dixon, T.H., 2006. Feedback between mountain belt growthand plate convergence. Geology 34, 893–896.

Iaffaldano, G., Bunge, H.-P., Buecker, M., 2007. Mountain belt growth inferred fromhistories of past plate convergence: a new tectonic inverse problem. Earth Planet.Sci. Lett. 260, 516–523.

Iaffaldano, G., Bunge, H.-P., 2008. Strong plate coupling along the Nazca/SouthAmericaconvergent margin. Geology 36, 443–446.

Ishii, M., Tromp, J., 2004. Constraining large-scale mantle heterogeneity using mantleand inner-core sensitive normal modes. Phys. Earth Planet. Inter. 146, 113–124.

Kanamori, H., 1986. Rupture process of sibduction-zone earthquakes. Ann. Rev. EarthPlanet. Sci. 14, 293–322.

King, S.D., Hager, B.H., 1990. The relationship between plate velocity and trenchviscosity in Newtonian and powerlaw subduction calculations. Geophys. Res. Lett.17, 2409–2412.

Kirby, S.H., 1983. Rheology of the lithosphere. Rev. Geophys. 21, 1458–1487.Kohlstedt, D.L., Evans, B., Mackwell, S.J., 1995. Strength of the lithosphere: constraints

imposed by laboratory experiments. J. Geophys. Res. 100, 17587–17602.Kong, X., Bird, P., 1995. SHELLS: A thin-shell program for modeling neotectonics of

regional or global lithosphere with faults. J. Geophys. Res. 100, 22129–22132.Lamb, S., Davies, P., 2003. Cenozoic climate change as a possible cause for the rise of the

Andes. Nature 425, 792–797.Levitt, D.A., Sandwell, D.T., 1995. Lithospheric bending at subduction zones based on

depth soundings and satellite gravity. J. Geophys. Res. 100, 379–400.Lithgow-Bertelloni, C., Richards, M.A., 1995. Cenozoic plate driving forces. Geophys Res.

Lett. 22, 1317–1320.Lithgow-Bertelloni, C., Richards, M.A., 1998. Dynamics of Cenozoic and Mesozoic plate

motions. Rev. Geophys. 36, 27–78.Lithgow-Bertelloni, C., Guynn, J.H., 2004. Origin of the lithospheric stress field, J. Geophys.

Res. 109, B01408.Liu, C.-S., Curray, J.R., McDonald, J.M., 1983. New constraints on the tectonic evolution of

the eastern Indian Ocean. Earth Planet. Sci. Lett. 65, 331–342.Llenos, A.L., McGuire, J.J., 2007. Influence of fore-arc structure on the extent of great

subduction zone earthquakes. J. Geophys. Res. 112, B09301.Masters, G., Laske, G., Bolton, H., Dziewonski, A., 2000. The relative behavior of shear

velocity, bulk sound speed, and compressional velocity in the mantle: implicationsfor chemical and thermal structure. In Earth's Deep Interior, AGU Monograph 117,AGU, Washington D.C.

Matas, J., Bass, J.D., Ricard, Y., Mattern, E., Bukowinski, M.S.T., 2007. On the lower mantletemperature and composition: predictions and limitations from generalizedinversion of radial seismic profiles. Geophys. J. Int. 170, 764–780.

McNamara, A.K., Zhong, S., 2005. Thermochemical structures beneath Africa and thePacific Ocean. Nature 437, 1136–1139.

McNutt, M.K., Menard, H.W.,1982. Constraints on yield strength in the oceanic lithospherederived from observations of flexure. Geophys. J. R. Asrton. Soc. 71, 363–394.

Mercier, J.L., Sebrier , M., Lavenu , A., Cabrera , J., Bellier , O., Dumont , J.-F., Machare , J.,1992. Changes in the tectonic regime above a subduction zone of Andean type: theAndes of Peru and Bolivia during the Pliocene–Pleistocene. J. Geophys. Res. 97,11945–11982.

Merkouriev, S., DeMets, C., 2006. Constraints on India plate motion since 20 Ma fromdense Russian magnetic data: implications for Indian plate dynamics. Geochem.Geophys. Geosyst. 7, Q02002.

Montelli, R., Nolet, G., Dahlen, F.A., 2004. Finite-frequency tomography reveals a varietyof plumes in the mantle. Science 303, 338–343.

Moresi, L., Solomatov, V., 1998. Mantle convection with a brittle lithosphere:thoughts on the global tectonic styles of the Earth and Venus. Geophys. J. Int. 133,669–682.

Moresi, L., Quenette, S., Lemiale, V., Meriaux, C., Appelbe, B., Muehlhaus, H.-B., 2007.Computational approaches to studying non-linear dynamics of the crust andmantle. Phys. Earth Planet. Inter. 163, 69–82.

Müller, R.D., Roest, W.R., Royer, J.-Y., Gahagan, L.M., Sclater, J.G., 1997. Digital isochrons ofthe world's ocean floor. J. Geophys. Res. 102, 3211–3214.

Müller, R.D., Sdrolias, M., Gaina, C., Roest, W.R., 2008. Age, spreading and spreadingasymmetry of the world's ocean crust. Geochem. Geophys. Geosyst. 9, Q04006.

National Geophysical Data Center, Total Sediment Thickness of the World's Oceans &Marginal Seas (Divins, D.L.), Retrieved date goes here, http://www.ngdc.noaa.gov/mgg/sedthick/sedthick.html.

National Geophysical Data Center, 1998. Digital relief of the surface of the Earth:Boulder, Colorado. National Oceanic and Atmospheric Administration DataAnnouncement 88-MGG-02.

http://www.ngdc.noaa.gov/mgg/sedthick/sedthick.htmlhttp://www.ngdc.noaa.gov/mgg/sedthick/sedthick.html


Norabuena, E.O., Dixon, T.H., Stein, S., Harrison, C.G.A., 1999. Decelerating Nazca–South America and Nazca–Pacific plate motions. Geophys. Res. Lett. 26,3405–3408.

Oeser, J., Bunge, H.-P., Mohr, M., 2006. Cluster design in the Earth Sciences: TETHYS, highperformance computing and communications — Second International Conference,HPCC 2006, Munich, Germany. Lect. Notes Comput. Sci. 4208, 31–40.

Pardo-Casas, F., Molnar, P., 1987. Relative motion of the Nazca (Farallon) and South-American plates since the late cretaceous time. Tectonics 6, 233–248.

Perez-Gussinye, M., Lowry, A.R., Watts, A.B., 2007. Effective elastic thickness of SouthAmerica and its implications for intracontinental deformation. Geochem. Geophys.Geosyst. 8, Q05009.

Piazzoni, A.S., Steinle-Neumann, G., Bunge, H.-P., Dolejš, D., 2007. Amineralogical modelfor density and elasticity of the Earth's mantle. Geochem. Geophys. Geosyst. 8,Q11010.

Raymo, M.E., Ruddiman, W.F., 1992. Tectonic forcing of late Cenozoic climate. Nature359, 117–122.

Ricard, Y., Vigny, C., 1989. Mantle dynamics with induced plate tectonics. J. Geophys. Res.94, 17543–17559.

Rice, J.R., 1992. Fault stress state, pore pressure distributions, and the weakness of theSan Andreas Fault. In Fault Mechanics and Transport Properties of Rocks. AcademicPress.

Rice, J.R., 2006. Heating and weakening of faults during earthquake slip. J. Geophys. Res.111, B05311.

Richards, M.A., Engebretson, D.C., 1992. Large-scale mantle convection and the historyof subduction. Nature 355, 437–440.

Richards, M.A., Yang, W.S., Baumgardner, J.R., Bunge, H.P., 2001. Role of a low-viscosityzone in stabilizing plate tectonics: implications for comparative terrestrialplanetology. Geochem. Geophys. Geosyst. 2.

Richardson, R.M., Coblentz, D.D., 1994. Stress modeling in the Andes: constraints on theSouth America intraplate stress magnitudes. J. Geophys. Res. 99, 22015–22025.

Rousse, S., Gilder, S., Farber, D., McNulty, B., Torres, V.R., 2002. Paleomagnetic evidencefor rapid vertical-axis rotation in the Peruvian Cordillera ca. 8 Ma. Geology 30,75–78.

Sandwell, D.T., Smith, W.H.F., 1997. Marine gravity anomalies from Geosat and ERS 1satellite altimetry. J. Geophys. Res. 102, 10039–10054.

Sella, G.F., Dixon, T.H., Mao, A., 2002. REVEL: A model for recent plate velocities fromspace geodesy. J. Geophys. Res. 107, 2081–2111.

Smith, W.H.F., Sandwell, D.T., 1997. Global sea floor topography from satellite altimetryand ship depth soundings. Science 277, 1956–1962.

Sobel, E.R., Hilley, G.E., Strecker, M.R., 2003. Formation of internally drainedcontractional basins by aridity-limited bedrock incision. J. Geophys. Res. 108, 2344.

Sobolev, S.V., Babeyko, A.Y., 2005. What drives orogeny in the Andes? Geology 33,617–620.

Somoza, R., 1998. Updated Nazca (Farallon)–South America relative motions during thelast 40My: implications formountain building in the central Andean region. J. SouthAm. Earth Sci. 11, 211–215.

Song, T.R.A., Simons, M., 2003. Large trench-parallel gravity variations predictseismogenic behavior in subduction zones. Science 301, 630–633.

Stein, S., Okal, E., 1978. Seismicity and tectonics of the Ninetyeast Ridge area: evidencefor internal deformation of the Indian plate. J. Geophys. Res. 83, 2233–2246.

Stein, C.A., Stein, S., 1992. A model for the global variation in oceanic depth and heatflow with lithospheric age. Nature 359, 123–129.

Steinberger, B., 2000. Slabs in the lower mantle — results of dynamic modellingcompared with tomographic images and the geoid. Phys. Earth Planet. Inter. 118,241–257.

Stixrude, L., Lithgow-Bertelloni, C., 2007. Influence of phase transformations on lateralheterogeneity and dynamics in Earth's mantle. Earth Planet. Sci. Lett. 263, 45–55.

Suppe, J., 2007. Absolute fault and crustal strength fromwedge tapers. Geology 35 (12),1127–1130.

Tackley, P.J., 2000. Self-consistent generation of tectonic plates in time-dependent,three-dimensional mantle convection simulations, part 1: pseudoplastic yielding.Geochem. Geophys. Geosyst. 1.

Tackley, P.J., Stevenson, D.J., Glatzmaier, G.A., Schubert, G., 1994. Effects of multiplephase transitions in a three-dimensional spherical model of convection in Earth'smantle. J. Geophys. Res. 99, 15877–15902.

Tapponnier, P., Zhiquin, X., Roger, F., Meyer, B., Arnaud, N.,Wittlinger, G., Jingsui, Y., 2001.Oblique stepwise rise and growth of the Tibet plateau. Science 294, 1671–1677.

Torsvik, T.H., Müller, R.D., Van der Voo, R., Steinberger, B., Gaina, C., 2008. Global platemotion frames: Toward a unified model. Rev. Geophys. 46, RG3004.

Trompert, R., Hansen, U., 1998. Mantle convection simulations with rheologies thatgenerate plate-like behaviour. Nature 395, 686–689.

Turcotte, D.L., Schubert, G., 2002. Geodynamics. Cambridge University Press, Cam-bridge, England. (456 pages).

van der Hilst, R.D., de Hoop, M.V., Wang, P., Shim, S.-H., Ma, P., Tenorio, L., 2007.Seismostratigraphy and thermal structure of Earth's core–mantle boundary region.Science 315, 1813–1817.

Weinstein, S.A., Olson, P.L., 1992. Thermal-convection with non-Newtonian plates.Geophys. J. Int. 111, 515–530.

Weissel, J.K., Anderson, R.N., Geller, C.A., 1980. Deformation of the Indo-Australian plate.Nature 287, 284–291.

Wiens, D.A., DeMets, C., Gordon, R.G., Stein, S., Argus, D., Engeln, J.F., Lundgren, P.,Quible, D., Stein, C., Weinstein, S., Woods, D.F., 1985. A diffuse plate boundary modelfor Indian Ocean tectonics. Geophys. Res. Lett. 12, 429–432.

Zhang, S., Christensen, U., 1993. Some effects of lateral viscosity variations on geoid andsurface velocities induced by density anomalies in the mantle. Geophys. J. Int. 114,531–547.

Zhong, S.J., Gurnis, M., 1995. Mantle convection with plates and mobile, faulted platemargins. Science 267, 838–843.

Zhong, S.J., Gurnis, M., Moresi, L., 1998. The role of faults, nonlinear rheology, and viscositystructure ingeneratingplates from instantaneousmantleflowmodels. J. Geophys. Res.103,15255–15268.

Zhong, S., Zuber,M.T.,Moresi, L., Gurnis,M., 2000.Role of temperature-dependent viscosityand surface plates in spherical shellmodels ofmantle convection. J. Geophys. Res.105,11063–11082.

Zoback, M.L., 1992. First and second order patterns of stress in the lithosphere: TheWorld Stress Map Project. J. Geophys. Res. 97, 11703–11728.

Relating rapid plate-motion variations to plate-boundary forces in global coupled models of the.....IntroductionGeodynamic plate modelingModels of Neo-tectonics and the large-scale circulation of the mantle

Models and resultsPlate boundary forces along the Nazca/South America marginFrictional plate-boundary forces along the Aleutian trenchEffects of an intra-plate boundary within the India/Australia plate

DiscussionConclusionsAcknowledgmentsReferences

Documents

Relating rapid plate-motion variations to plate-boundary forces in …people.rses.anu.edu.au/iaffaldano_g/CONTENT/ib_tectono... · 2010. 8. 24. · Lithgow-Bertelloni and Richards