Gondwana Research 24 (2013) 958–968

Contents lists available at ScienceDirect

Gondwana Research

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

Role of lithosphere in intra-continental deformation: Central Australia

B.L.N. Kennett ⁎, G. IaffaldanoResearch School of Earth Sciences, The Australian National University, Canberra, ACT 0200, Australia

⁎ Corresponding author.E-mail address: Brian.Kenneth@anu.edu.au (B.L.N. K

1342-937X/$ – see front matter © 2012 International Ahttp://dx.doi.org/10.1016/j.gr.2012.10.010

a b s t r a c t

a r t i c l e i n f o

Article history:Received 7 February 2012Received in revised form 17 October 2012Accepted 22 October 2012Available online 21 November 2012

Keywords:Central AustraliaLithosphereDeformationStrength

Since the Proterozoic, there has been a set of deformation cycles in central Australia culminating in the AliceSprings Orogeny around 400 Ma. These events occurred away from plate boundaries and involved extensionas well as compression, although their precise history remains difficult to unravel from the geologic record.Much evidence of deformation is left in the central Australian crust, which features significant Moho topog-raphy and an associated gravity signal. In the past, several mechanical models invoked crustal thickening andconsiderable compression to explain these geophysical characteristics. However, it is hard to envisage exten-sive deformation affecting the crust alone, but leaving no deformation record in the sub-crustal lithosphere.In recent seismic tomography studies, there is continuous seismically-fast lithosphere in central Australiabelow depths of about 100 km. In this region, the uppermost lithospheric mantle is seismically slow, but ex-hibits no significant attenuation of seismic waves. These new constraints make simple crustal thickening un-likely to be the main mechanism to generate variations of the Moho depth in central Australia. Here wepropose a mechanical model of deformation that involves the entire lithosphere. Wemake no strong assump-tions about the history of deformation cycles. Our model does not require lithospheric thickening at any stageof the deformation cycle, and results in a present-day scenario compatible with shallow as well as deep con-straints on the lithosphere structure.

© 2012 International Association for Gondwana Research. Published by Elsevier B.V. All rights reserved.

1. Introduction

Since the Proterozoic the central region of the Australian conti-nent has been the stage of a series of orogenic events that have lefta clear imprint on the present-day landscape, yet which occurredseemingly far from major past plate boundaries (e.g. Veevers, 2000;Haines et al., 2001). Modest ranges of hills are quite striking in thegenerally low topography, but are accompanied by major east–westoriented gravity anomalies (Fig. 1). These anomalies are among thelargest in intra-continental settings world-wide outside regions of ac-tive tectonism, yet the last orogenic activity was at around 300 Ma.

TheMusgrave Province experienced three phases ofMesoproterozoicorogeny (Smithies et al., 2011); the 1345–1292 MaMountWestOrogeny,the 1220–1150 Ma Musgrave Orogeny and the 1080–1030 Ma GilesEvent. A distinctive component of the Musgrave Orogeny is a 70 Myrperiod of ultrahigh-temperature (UHT) methamorphism requiring tem-peratures at lower crustal conditions above 900 °C (Smithies et al.,2011). To achieve such a state, the asthenosphere has to rise to close tothe base of the crust, possibly funnelled by configuration of the litho-spheric roots of the cratons following the amalgamation of the Northand West Australian Cratons with the South Australian Craton prior to1290 Ma. Subsequently a cratonic style lithosphere has formed in themantle, but the zone has remained a locus of deformation. During the

ennett).

ssociation for Gondwana Research.

Neoproterozoic the Centralian Superbasin, consisting of the Georgina,Ngalia, Amadeus and Savory basins, formed in the zone where thethree cratonic regions had come together (e.g., Walter et al., 1995).One explanation for the formation of the superbasin is that the litho-sphere was extended and thinned as a consequence of a mantle plumecentred on the eastern edge of the Gawler Craton (Zhao et al., 1994),which gave rise to the Gairdner large igneous province with dykeswarms in central Australia at around 825 Ma. The superbasin was de-formed and subdivided by late Neoproterozoic and Paleozoic orogenicevents (e.g., Paterson, Petermann, and Alice Springs Orogenies). In thePhanerozoic, continued deformation occurred in central Australia, al-though it is far from any active plate boundaries, possibly as a result ofmutual rotation of the cratonic blocks (Roberts and Houseman, 2001).The Petermann Orogeny took place along the southern margin of theAmadeus Basin, while the Alice Springs Orogeny occurred along itsnorthern margin.

The Petermann Orogeny at around 620–520 Ma (e.g., Cawood andKorsch, 2008; Walsh et al., 2013–this volume) has left its imprint onthe Musgrave Block (Fig. 2), with evidence for the presence ofsub-eclogitic rocks indicating crustal thickening (Camacho andMcDougall, 2000). The sediments shed from the orogenic belt nowform the distinctive features of Uluru and Kata Kjuta – also known asAyers Rock and the Olgas – rising from the Amadeus basin (Fig. 2).There is evidence of extension and subsequent compression in Ordovi-cian metamorphic rocks from the Harts Range in the Arunta Inlier(Fig. 2), indicating the presence of sedimentary basins more than

Published by Elsevier B.V. All rights reserved.

Fig. 1. Free-air gravity anomalies in central Australia.

959B.L.N. Kennett, G. Iaffaldano / Gondwana Research 24 (2013) 958–968

20 km thick (Buick et al., 2001). Themost recent deformation eventwasthe Alice Springs Orogeny between 400 and 300 Ma (e.g., Haines et al.,2001), inwhich a substantial part of the crustal shorteningwas taken upby thrust zones such as the Red-bank thrust bounding the Arunta blockto the south (Fig. 2) (Goleby et al., 1989). The legacy of these cycles ofextension and compression can be found in substantial geophysicalsignatures.

Very large east–west trending gravity anomalies – up to1000 μm s2 – align with the grain of the orogenic structures in theArunta and Musgrave Blocks (e.g. Lambeck, 1983; Aitken et al.,2009a, 2009b; Aitken, 2010). These features are caused by large hor-izontal gradients of crustal thickness, with major departures fromisostasy that have been preserved for at least 300 Myr. A sequenceof deployments of short-period seismic recorders traversed the regionof major gravity anomalies in the 1980s and revealed striking variationsover very short distances in the relative arrival times of seismic wavesfrom distant earthquakes. Lambeck et al. (1988) used the seismic obser-vations to develop structural models of the crust based on the patternsof relative arrival times for P waves from sources with differentazimuths.McQueen and Lambeck (1996) later undertook a tomographicinversion of the various data sets that confirmed the need for significantvariations in crustal thickness linked to the gravity anomalies. More re-cently, Aitken et al. (2009a, 2009b) provided a detailed interpretation ofthe gravity anomalies on sections across theMusgrave Block that indeedindicates a wedge shaped zone bounded by inclined faults, with local-ised uplift of the Moho as high as 15 km. Furthermore, a full crustal re-flection survey across the Red-Bank zone (Goleby et al., 1989), in theArunta province, also showed clear evidence for major displacement of

the lower crust towards the surface. To the east, a recent reflection seis-mic profile crossing from the Arunta province into the Georgina basinprovides strong indication for thrusting faults at the full crustal scale, al-though with a rather different Moho configuration (Korsch et al., 2011).The major gravity anomaly in the Musgrave Block has also beencrossed by reflection lines (Korsch and Koscitin, 2010), and the in-ferred Moho uplift of about 15 km fits well with the gravity interpre-tation. Central Australia has undergone significant intra-platetectonic activity for over a billion years. But despite the wealth ofgeophysical observations available, it still remains a challenge to dis-entangle the details of multiple deformation cycles which have oc-curred in different parts of the region through time.

In the past severalmodels of deformation cycles have been proposedin order to explain these geophysical features of the central Australiancrust. Among others, Lambeck (1983) resorted to visco-elastic modelsto show that long-lasting compression of the Australian crust would re-sult in significant uplift between different blocks. Erosion and sedimen-tation are also recognised to enhance such process. Roberts andHouseman (2001) used a thin-sheet model to demonstrate that far-field shear tractions along plate boundaries during the Devonian–Carboniferous may result in crustal thickening within central Australia,up to a ratio of ~1.7. More recently, Aitken et al. (2009a, 2009b,2013–this volume) proposed that deformation cycles initiatingwith ex-tension may have weakened the crust and generated normal faulting,while subsequent compression and crustal thickening have upliftedthe Moho base and strengthened the crust at the same time. Underthese conditions, crustal structures can be preserved out of isostatic bal-ance for longer than 300 Myr.

960 B.L.N. Kennett, G. Iaffaldano / Gondwana Research 24 (2013) 958–968

Common to these models is the requirement that deformation cy-cles initiate with extension followed by a larger amount of compres-sion, which may remain in place for extended periods of time. Moreimportantly, models consistently call upon thickening of the centralAustralian crust as the main mechanism to generate Moho topogra-phy. But while it is difficult to envisage a scenario where far-field tec-tonic forces have acted only upon the Australian crust, leaving thelower lithosphere intact, recent findings from seismic tomography –

which we report below – seemingly show no imprint of multiple de-formation cycles left in the lower part of the lithospheric mantle. Inorder to reconcile these seismological observations with structuralconstraints in central Australia, we propose here a deformationmodel that involves the whole lithosphere, without the need to sig-nificantly thicken the crust. Deformation cycles may or may not initi-ate with extension, but we pose virtually no constraint on amountand duration of subsequent compression.

Yi

Cp

AF

Am

ArCa

Eu

Ki

Mu

Of

PC

GSD

Proterozoic basementArchean sediments Archean basement

-10

-20

-30

-40

120 130

0 500km

Fig. 2. Simplified representation of the main tectonic features of central Australia. Chain-dotthe reinterpretation by Direen and Crawford (2003). Key to marked features: AF— Albany–FOrogen, Cu— Curnamona Craton, Er— Eromanga basin, Eu— Eucla basin, Ga— Gawler cratonMu—Musgrave Block, Of— Officer basin, PC Pine Creek Inlier, T— Tennant Creek block, Yi—

2. The current state of the lithosphere in central Australia

Awide range of seismological studies have been carried out in thecentral Australian region, including refraction surveys, deploymentsof short-period and broadband seismic stations, as well as full crustalreflection profiles. The configuration of seismic stations and shotpoints (Fig. 3) is constrained in many places by the logistics, as fewroads traverse the central Australian deserts. The lines of short-period stations have been employed in local tomographic studies(e.g. McQueen and Lambeck, 1996). The broad-band stations havebeen used for surface wave tomography (e.g. Fishwick et al., 2005;Fichtner et al., 2009), and for receiver-function studies of the litho-sphere in the neighbourhood of the stations (Reading et al., 2007). Re-sults from all the seismological studies across the continent haverecently been brought together to create a new Moho depth map ofthe Australian region (Kennett et al., 2011). We show a portion of this

Cu

Er

Ga

La

MI

Mc

T

SD

HR

Paleozoic basementProterozoic sediments

140

ted lines mark outlines of the major cratons. The Tasman line in dashed grey is based onraser belt, Ar— Arunta Block, Am— Amadeus Basin, Ca— Canning basin, Cp— Capricorn, Ki— Kimberley block, La— Lachlan Orogen, Mc—MacArthur basin, MI—Mt Isa block,Yilgarn craton, HR— Harts Range, SD— Simpson Desert, and GSD— Great Sandy Desert.

Refraction receivers

Refraction shot

Reflection - explosion Reflection - vibrators

Short-period

Broad-band

-10

-20

-30

-40

120 130 140

0 500km

Fig. 3. Distribution of seismic stations in the central Australian region.

961B.L.N. Kennett, G. Iaffaldano / Gondwana Research 24 (2013) 958–968

Moho map in Fig. 4 for the same zone as the gravity map in Fig. 1. Thedistribution of seismic controls implies that there are stronger north–south constraints than east–west, nonetheless crustal thickness varia-tions associated with the major gravity anomalies are substantial andbecome evident in Fig. 4.

Further information on the nature of the crust is available from am-bient noise tomography (Saygin and Kennett, 2010, 2012). In this pro-cedure stacked cross-correlations of seismic ground-motion at station-pairs provide the result of a virtual experiment featuring a source atone station at the receiver at the other. The surface-wave portion ofthe path response is well recovered at shorter periods, and the groupvelocity of these arrivals can be measured as a function of frequencyand then employed in tomography. The group velocity distribution atthe 5 s period is particularly sensitive to the presence of sediments,and clearly maps these around the Musgrave Block (Saygin andKennett, 2010). In addition, group-velocity anomalies at longer periodprovide some indication of shallow thermal effects often referred to asthermal blanketing to the east. The tomographic results for the middlecrust from 10 to 20 km show little contrast in seismic wavespeeds,and therefore no obvious tie to the long-term history of surface geology(Saygin and Kennett, 2012).

However, strong contrasts emerge at the base of the crust. Fig. 5shows a comparison of the recent reflection results from the GOMAline across the Musgrave Block (Korsch and Koscitin, 2010) (Fig. 5a)with the gravity interpretation by Aitken et al. (2009a, 2009b)(Fig. 5b). A clear jump in the position of the Moho is detected, witha much thinner crust under the main gravity high. Oblique boundingfaults are also mapped through the change in reflection character(Fig. 5a). Such major features are well out of isostatic equilibrium,and represent the product of at least 500 Myr of erosion and subse-quent deformation since the Petermann Orogeny. Similar complexityis seen further north in the Arunta Province at the Red-Bank zone(Goleby et al., 1989) from the later Alice Springs Orogeny.

There is seemingly no record of these deformation cycles in thelower lithosphere, which is imaged by means of surface wave tomog-raphy (Fig. 6) employing longer-period seismic waves from earth-quakes in the belts around Australia recorded at portable as well aspermanent seismic stations (e.g. Yoshizawa and Kennett, 2004;Fishwick et al., 2005; Fichtner et al., 2009). In these seismic models,the shallowest part of the inferred structure can be influenced bythe nature of Earth's crust, but images are reliable below 75 kmdepth. A prominent feature, common to all models, is a region of

962 B.L.N. Kennett, G. Iaffaldano / Gondwana Research 24 (2013) 958–968

reduced seismic wavespeeds in the lower lithosphere, approximately80 km below the region affected by the Alice Springs Orogeny(Fig. 6a). At greater depth wavespeeds rapidly rise to the levelsexpected for cratonic lithosphere (Fig. 6b), and indeed include someof the fastest shearwavespeeds found anywhere beneath the Australiancontinent. Seismically-fast lithosphere continues to at least 200 kmdepth (Fig. 6c–d), and possibly even further. The resolution attainedin such studies is of the order of 200 km horizontally and 25 km verti-cally, so that features produced during the deformation cycles shouldshow some expression.

Independent evidence for reduced shear wavespeeds in the upperlithosphere comes from observations of refracted waves (Kaiho andKennett, 2000) from regional sources, recorded at portable stations. Inparticular, aftershocks of the 1987 Tennant Creek earthquake showslightly slow propagation time to stations in the central part of Australiafor both P and Swaves (Fig. 7b); although at greater distances the traveltimes for body-wave phases are faster than for the ak135 referencemodel (Kennett et al., 1995) (Fig. 7a). Both the travel times of P and Swaves propagating from earthquakes within Australia, as well as thesurface wave tomography from regional sources, require the existenceof a zone in the uppermost lithosphericmantle that is somewhat slowerthan average. Importantly, body-wave arrivals retain short-period Swaves, so that there can only be modest attenuation in this zone(Fishwick and Reading, 2008). Because of this, the layering in theshear wavespeed structure and the strong positive velocity gradientinto the fast wavespeeds at depths below 100 km are unlikely to be re-lated to a locally hotter steady-state continental geotherm. In fact, a

120˚ 130˚

−40˚

−30˚

−20˚

−10˚

0 500km

Fig. 4. Portion of Moho map of Australia (Kennet

possible thermal explanation involves the redistribution of high-heatproducing elements from the upper crust to lower depths, as for in-stance it would result from crustal thickening. Yet this would imply el-evated temperatures in the lower lithosphere that are inconsistentwiththe observed attenuation behaviour. These constraints make prolongedperiods of sustained crustal thickening unlikely.

An alternative explanation, suggested by Fishwick and Reading(2008), calls upon the presence in the upper part of the lithosphericmantle of significant amounts of seismically-slow materials, such ashydrated minerals like paragastic amphibole derived from the uppermantle. We note that this would imply some mechanism of transportupwards of these minerals, from the upper mantle into the subcrustallithosphere.

Seismic tomography beneath central Australia presents a situationwhere there is clear evidence of major crustal deformation, which oc-curredmore than 1000 km away from past plate boundaries. In the up-permost mantle the lithosphere features slightly lowered shearwavespeeds, which sharply increase to cratonic levels from about100 km depth. This unusual scenario occurs directly beneath the zonewhere the Australian cratons amalgamated in the Proterozoic (Bettsand Giles, 2006; Cawood and Korsch, 2008) and where, subsequently,there have been periods of intraplate tectonic activity; suggesting a cor-relation between the prolonged history of deformation and a highlyunusual lithospheric structure. An intriguing feature of this part ofAustralia is that it is somewhat distinctive in the nature of the transitionfrom the lithosphere to asthenosphere. Not only is there evidencefor rather thick lithosphere (see, e.g., Kennett and Salmon, 2012,

140˚

101520253035404550556065

Moh

o D

epth

(km

)

t et al., 2011) in same configuration as Fig. 1.

Fig. 5. (a) Portion of reflection profile across the Musgrave ranges showing the distinct uplift of the Moho (based on Korsch et al., 2011). The CDP points are a nominal 20 m apart sothat 1000 cdp points corresponds to 20 km. The section is at approximately true scale (based on an EMS crustal velocity of 6 km/s). (b) Model from Aitken et al (2009a) based ongravity field modelling, at a slightly smaller horizontal scale. WL corresponds to the Wintiginna Fault and WT to the Woodroffe Thrust in (a).

963B.L.N. Kennett, G. Iaffaldano / Gondwana Research 24 (2013) 958–968

Fig. 9), but also the transition seems rather sharper than its surround-ings (Yoshizawa and Kennett, 2012). Indeed the edges of the main cra-tonic blocks can be identified by a thinning of the lithosphere–asthenosphere transition.

It is difficult to imagine that the deformation cycles involved onlythe crust. However, if the entire lithosphere underwent multiple cy-cles of extension and compression, with the surface relief being con-tinuously re-weathered, why is there seemingly no record in itspresent-day deeper structure?

3. Mechanical deformation in central Australia

We propose here a simple model for the evolution of the centralAustralian lithosphere under cycles of deformation. Fig. 8a shows a

cross section of the mechanical lithosphere under no particular condi-tion of compression or extension. We note that the initial architecturemay well be more complex than in our sketch, as a consequence ofprior deformation. Nevertheless, the reader will see that our infer-ences do not have a strong dependence on the initial structural condi-tions. We concentrate on the behaviour of the whole lithosphere, butrecognise that within the crust there will need to be ductile deforma-tion to accommodate the cycles of compression and extension. In themodel in Fig. 8: green lines delimit the brittle crust; oblique segmentswithin the crust represent weakness planes possibly inherited fromthe amalgamation of cratons since the Proterozoic.

We make the weak assumption that the base of the mechanicallithosphere is initially isothermal, shown in red in Fig. 8a. We canimagine this red line to be associated with the temperature at about

964 B.L.N. Kennett, G. Iaffaldano / Gondwana Research 24 (2013) 958–968

100 km depth. Thin black segments mark lengths, such as the dis-tance between the closest weakness planes within the crust, L1, aswell as the initial thicknesses of the crust and lithosphere, Hc1 andHm. The blue square traces minerals – for instance, amphiboles –

originally located at the base of the lithosphere. We imagine thatthe deformation cycle begins with extension and stretching. As a re-sult, crust and lithosphere become progressively thinner. Fig. 8bsketches the scenario associated with maximum extension. Becauseof the brittle nature of the crust, we can envisage the occurrence of

4.11

120 130 140-40

-35

-30

-25

-20

-15

-10

-4

-3

-3

-2

-2

-1

-1

120 130 140-40

-35

-30

-25

-20

-15

-10

-4

-3

-3

-2

-2

-1

-1

0 500km

0 500km

175 km

75 km

Fig. 6. Images of the shear wavespeed distribution beneath central Australia from the model o

normal faulting within it shown by the green oblique segments, L2 in-dicates the stretched distance between the two closest normal faults.Mass conservation requires that

L1Hc1ρc1 ¼ L2Hc2ρc2; ð1Þ

where ρc1 and ρc2 indicate the crustal densities before and after ex-tension, while Hc2 is the minimum thickness of the stretched crust.We assume the density to be constant with depth for the sake of

4.96Wavespeed [km/s]

120 130 1400

5

0

5

0

5

0

120 130 1400

5

0

5

0

5

0

0 500km

0 500km

125 km

225 km

f Fishwick et al. (2008), at depths of (a) 75 km, (b) 125 km, (c) 175 km, and (d) 225 km.

-20

-30

130 140 130 140

-18 s 18 srelative to ak135

-5.0 5.0Perturbation [%]

0 500km 0 500km

(a) (b)

Fig. 7. Comparison of the shear wavespeed distribution for the central Australian region: (a) observations of the time of passage of refracted waves in the mantle lithosphere,updated from Kaiho and Kennett (2000) and (b) SV wavespeeds from the model of Fishwick et al. (2008). There is a close correspondence in the position of the slower zonebased on independent observations. In each case the reference model is ak135 (Kennett et al., 1995).

965B.L.N. Kennett, G. Iaffaldano / Gondwana Research 24 (2013) 958–968

simplicity (this is equivalent to considering the average density over,e.g., the actual range of variability for the crust). Eq. (1) can be alter-natively written as

L1L2

¼ Hc2ρc2

Hc1ρc1: ð2Þ

Mass conservation also implies that the base of the mechanicallithosphere bulges upwards, allowing hotter mantle to fill the void.At the typical depth of the lithosphere base, mantle material is how-ever as dense (we will refer to its density as ρm) as the lower litho-sphere, as inferred from the seismological reference model ak135(Kennett et al., 1995).

The pressure at the base of the mechanical lithosphere varies be-tween the two scenarios. Let P1 and P2 be the pressure before andafter extension, at the position of the empty dot in Fig. 8a–b.

P1 ¼ g ρc1Hc1 þ ρmHm½ �; ð3Þ

P2 ¼ g ρmHm þ ρmHc1−Hc2

2

� �þ ρc2Hc2

� �; ð4Þ

where g is the gravitational acceleration. Note that in Eq. (4) we ne-glect the weight of the material within the newly formed depressionof the crust. The pressure difference therefore is

P1−P2 ¼ g ρc1Hc1−ρc2Hc2−ρm

2Hc1−Hc2ð Þ

h i; ð5Þ

or alternatively

P1−P2

g¼ Hc1ρc1 1−Hc2ρc2

Hc1ρc1

� �−ρmHc1

21−Hc2

Hc1

� �: ð6Þ

We can simplify using Eq. (2)

P1−P2

g¼ Hc1ρc1 1− L1

L2

� �−ρmHc1

21− L1ρc1

L2ρc2

� �: ð7Þ

Because under extension ρc1≥ρc2, we can write

P1−P2

g≥Hc1ρc1 1− L1

L2

� �−ρmHc1

21− L1

L2

� �; ð8Þ

or, in a simpler form,

P1−P2

g≥Hc1 1− L1

L2

� �ρc1−

ρm

2

� �: ð9Þ

We take ρc1≈2700 kg m−3 and ρc2≈3300 kg m−3. Now sinceL1≤L2, after extension a pressure deficit is established at the base ofthe lithosphere. This will cause the mechanical lithosphere to re-bound upwards, so that the pressure balance is restored. The dynam-ics of such process may be inferred by requiring that viscous stressesin the lower lithosphere counter-balance the stress deficiency causedby the extension. From Fig. 8b we can see that the pressure deficitgenerated by the crustal depression is

τd ≈Hc2gρc1: ð10Þ

L1

Hc1Hm

IsothermalL2

Hc2Hm

Hc1−Hc2

(a)

(b)

(c)

(d)

(e)

(f)

Fig. 8. Proposed mechanical model of deformation for the Central Australian lithosphere.(a) Initial tectonic setting. Grey oblique segments within the crust represent pre-existingweakness planes, inherited from previous deformation cycles, which might have beenpresent prior to extension. (b) Thinning of the crust and upward bulgingof the lithospher-ic mantle following extension. (c) Upward rebound of crust and lithospheric mantle, dueto isostatic re-equilibration. (d) Thermal diffusion of the heat-content displaced upwardsas a result of lithospheric mantle bulging. (e) Relief of crust surface and bulging of litho-spheric mantle following compression. (f) Weathering of crustal relief and thermal diffu-sion of the lithospheric mantle heat-content. See Section 3 for a detailed description.

966 B.L.N. Kennett, G. Iaffaldano / Gondwana Research 24 (2013) 958–968

This must be equal to the viscous resistive stress arising within theductile lower mechanical lithosphere

τr ¼ −μ _�≈−μvΔx

≈−μ1L2

dHc2

dt; ð11Þ

where μ is the effective viscosity of the lower lithosphere, _� is thestrain rate, v is the typical velocity of induced flow, t is time, and Δxis the typical distance over which such flow decays to zero. The neg-ative sign accounts for the fact that Eq. (11) represents a resistivestress. As a result, we obtain

1Hc2

dHc2

dt¼ − gρc1L2

μ: ð12Þ

Eq. (12) has an exact solution. In fact, Hc2 varies exponentiallywith a typical time equal to the inverse of the right-hand side ofEq. (12). Using μ=1021 Pa s (Mitrovica and Forte, 2004) and L2=100 km, we find that it takes ~12,000 yr for the lithosphere to re-bound to the condition of mechanical equilibrium, where the sum ofthe forces is null (Fig. 8c). By mass conservation, such rebound willonce again produce an increase of the upward bulge at the litho-sphere base to about Hc1−Hc2.

Under these circumstances, the lithosphere is inmechanical, but notthermodynamic nor chemical equilibrium. In fact, the hotter-mantlebulge will diffuse its heat into the surroundings. Because the bulgesize is much smaller than the mantle itself, we may consider the latteras a heat source. Therefore thermodynamic equilibrium – where tem-perature gradients do not change through time – will be achievedonce the temperature gradient with depth returns to be the same asin the initial state in Fig. 8a. To that end, heat-transfer must occur pre-dominantly along the local horizontal direction, shown by the red ar-rows in Fig. 8c and e. The governing equation for the diffusive process is

dTdt

¼ kd2Tdx2

; ð13Þ

where T is temperature and k is the thermal diffusivity of the uppermantle. We simplify Eq. (13) using finite increments, and obtain

ΔTΔt

¼ kΔTΔx2

: ð14Þ

From Eq. (14) we may compute the time it takes to restore ther-modynamic equilibrium (Fig. 8d) by heat diffusion

Δt ¼ L22k: ð15Þ

Using again L2=100 km and the thermal diffusivity of harzburgitek=2×10−6 m2 s−1 (Waples and Waples, 2004), we obtain the esti-mate Δt≈160 Myr.

We now consider the effect of subsequent compression acting uponthe lithosphere (Fig. 8e). As a result of mass conservation, crustal blocksare displaced upwards, generatingMoho-depth variations and associatedgravity anomalies. The lithosphere basewill also bulgeupwards, howeverit will take ~200 Myr to return to thermodynamic equilibrium (Fig. 8f),as can be seen from Eq. (15). Depending on the age of the last compres-sion event, the present-day lithosphere base might still show somesignature of ongoing cooling (grey area in Fig. 8f). Such an inferencewould be in line with the seismologically-observed sharp transitionfrom lithosphere–asthenosphere in central Australia (Kennett andSalmon, 2012). Regardless of the uncertainty on the age of the latestcompression, this time is significantly longer than the typical timeassociated with erosive processes and, more generally, withweathering. Therefore erosion can certainly reshape the centralAustralian surface to its present-day flat land.

4. Discussion

The result of our model is a scenario (Fig. 8f), which is indeedcompatible with the geophysical characteristics of the crust and lith-osphere in central Australia. Crustal blocks are displaced upwards,therefore generating significant variations of the Moho depth, aswell as a strong gravity signature associated with them. At the sametime, the middle lithosphere may have been populated with mineralsoriginally located at the base of the mechanical lithosphere (bluesquare in Fig. 8f). This may help to provide an explanation for the ob-served low velocity but weak attenuation of seismic waves. On theother hand, the lower lithosphere has returned to a steady thermalstate typical of cratons, because sufficient time has passed for heatto diffuse even within highly-refractory environments as, for in-stance, harzburgite. Because lithosphere processes occur on typical

Isothermal

(a)

(d)

(e)

(f)

(b)

(c)

Fig. 9. Alternative deformation cycle initiating with compression, followed by exten-sion. (a) Initial tectonic setting. Grey oblique segments within the crust representpre-existing weakness planes, inherited from previous deformation cycles, whichmight have been present prior to compression. (b) Relief of crust surface and bulgingof lithospheric mantle following compression. (c) Weathering of crustal relief andthermal diffusion of the lithospheric mantle heat-content. (d) Thinning of the crustand upward bulging of the lithospheric mantle following extension. (e) Upward re-bound of crust and lithospheric mantle, due to isostatic re-equilibration. (f) Thermaldiffusion of the lithospheric mantle heat-content.

967B.L.N. Kennett, G. Iaffaldano / Gondwana Research 24 (2013) 958–968

time-scales of millions of years, rebound upwards of the stretchedlithosphere (Fig. 8c) is achieved virtually instantaneously — in a geo-logical sense. We also note that the typical diffusion time associatedwith our model, ~200 Myr, is shorter than the documented intervalsof extension/compression cycles in central Australia (e.g. Veevers,2000). Therefore the lithosphere will have had sufficient time toreach each of the stages depicted here. Thus after a cycle of deforma-tion the lithosphere is back close to the condition in which it started,but with a record left in the crust of the specifics of the orogeniccycle. As external conditions change, this zone of contrast in litho-spheric properties can be reactivated in a further cycle of deforma-tion. Thus the Alice Springs Orogeny occurs close enough in time tothe Petermann Orogeny for residual thermal effects to be present,with a slightly weaker lithosphere predisposed to deform.

One process we did not consider in our model is sedimentation ofthe crustal depression (Fig. 8b). The major effect of this of coursewould be to reduce the total rebound, and therefore decrease thesize of the lithospheric bulge (Fig. 8c). We can also envisage a locallyhotter geotherm, due to thermal blanketing from the sediments.However, this should have little or virtually no effect on diffusionof the heat-content from the bulge into the surroundings, becausein our approximation its typical time-scale does not depend on thetemperature difference. Thus, under these circumstances the lowerlithosphere may still return to thermodynamic equilibrium prior tothe compression stage (Fig. 8e).

Contrary to most of the models previously proposed, our scenariodoes not require extension to pre-date compression. In fact, shouldthe lithosphere undergo compression first, uplift of crustal blocksmay still occur along the crustal weakness planes inherited from theamalgamation of cratons since the Proterozoic; this case is sketchedin Fig. 9. The dynamic and thermodynamic processes associatedwith each step are the same as in Fig. 8, but occur at different stagesof the deformation cycle. Note, that the final structure of the litho-sphere is the same in Figs. 8 and 9. But, the scenario in Fig. 9 implicitlycalls for more extension than compression; if that was not the casethen some relief would have been present at stage (e) in Fig. 9. How-ever, one may again invoke weathering to reconcile our deformationmodel with the present-day flat relief of central Australia. We alsonote that at no stage the crust undergoes major thickening.

The sole intrinsic constraint our model poses on compression con-cerns its magnitude. For shortening to occur, far-field forces actingupon the central Australian lithosphere must be sufficient to over-come the strength of crustal faults, which is mainly hosted withintheir brittle portion. This effectively represents a lower limit for com-pression in our model. To first-order, considering a maximum faultfriction of 0.2 (e.g. Suppe, 2007) and brittle crust depths of even asmuch as 20–30 km, the total strength of the brittle crust would be~2 TN m−1. However, this is well within the range of tectonic forcesinferred, for instance, from numerical modelling of plate boundaryprocesses (e.g. Iaffaldano et al., 2007, 2011; Capitanio et al., 2011).

Ourmodel provides some insight into the typical time-scales associ-atedwith healing of continentalmargins. At the present-day the centralAustralian lithosphere appears intact, despite the successive phases ofdeformation. The very high seismic wavespeeds suggest the presenceof harzburgites in regions that have had much thinner lithosphere inthe deep geological past. Our mechanical model requires deformationthrough the lithosphere. Yet when we consider the last event at~300 Ma we see no obvious deformation. From this one may concludethat it took less than 300 Myr for deformation planes to heal, posingan upper limit to the time-scale of such a process.

5. Conclusions

Since the Proterozoic multiple cycles of deformation have shapedthe Australian continent and left significant evidence in its present-day crust. The signature of these processes can be inferred from

reflection seismology, gravity surveys, and sedimentation history.In addition, we report constraints from seismic tomography on thestructure of the lower lithosphere in central Australia, which showno apparent correlation with the history of surface geology. In thepast several models have been proposed to explain the mechanicsassociated with these cycles. While these models focus mainly onthe dynamics of the thickened crust, it is difficult to imagine howthe whole lithosphere could have not been involved in these defor-mation cycles. We proposed a mechanical model where extension/compression cycles act upon the entire lithosphere in central Australia.In ourmodel there is no need for the crust to thicken significantly at anystage, nor to make assumptions about the history of deformation. Theresulting present-day structure of the crust and lithosphere is in goodagreement with shallow as well as deep constraints from centralAustralia.

968 B.L.N. Kennett, G. Iaffaldano / Gondwana Research 24 (2013) 958–968

References

Aitken, A.R.A., 2010. Moho geometry gravity inversion experiment (MoGGIE): a refinedmodel of the Australian Moho and its tectonic and isostatic implications. Earth andPlanetary Science Letters 297, 71–83.

Aitken, A.R.A., Betts, P.G., Weinberg, R.F., Gray, D., 2009a. Constrained potential fieldmodeling of the crustal architecture of the Musgrave Province in central Australia:evidence for lithospheric strengthening due to crust–mantle boundary uplift. Journalof Geophysical Research 111, B12405. http://dx.doi.org/10.1029/2008JB006194.

Aitken, A.R.A., Betts, P.G., Ailleres, L., 2009b. The architecture, kinematics, and litho-spheric processes of a compressional intraplate orogen occurring under Gondwanaassembly: the Petermann Orogeny, central Australia. Lithosphere 1, 343–357.

Aitken, A.R.A., Smithies, R.H., Dentith, M.C., Joly, A., Evans, S., Howard, H.M., 2013.Magmatism-dominated intracontinental rifting in the Mesoproterozoic: theNgaanyatjarra Rift, central Australia. Gondwana Research 24, 886–901 (this volume).

Betts, P.G., Giles, D., 2006. The 1800–1100 Ma tectonic evolution of Australia. Precam-brian Research 144, 92–125.

Buick, I.S., Miller, J.A., Williams, I.S., Cartwright, I., 2001. Ordovician high-grade meta-morphism of a newly recognised late Neoproterozoic terrane in the northernHarts Range, central Australia. Journal of Metamorphic Geology 19, 373–394.

Camacho, A., McDougall, I., 2000. Intracratonic, strike-slip partitioned transpressionand the formation and exhumation of eclogite facies rocks: an example from theMusgrave Block, central Australia. Tectonics 19, 978–996.

Capitanio, F.A., Faccenna, C., Zlotnik, F., Stegman, D.R., 2011. Subduction dynamics andthe origin of Andean orogeny and the Bolivian orocline. Nature 480, 83–86.

Cawood, P.A., Korsch, R.J., 2008. Assembling Australia: Proterozoic building of a conti-nent. Precambrian Research 166, 1–38.

Direen, N.G., Crawford, A.J., 2003. The Tasman Line: where is it, what is it, and is itAustralia's Rodinian breakup boundary? Australian Journal of Earth Sciences 50,491–502.

Fichtner, A., Kennett, B.L.N., Igel, H., Bunge, H.-P., 2009. Full seismicwaveform tomographyfor upper-mantle structure in the Australasian region using adjoint methods. Geo-physical Journal International 179, 1703–1725.

Fishwick, S., Reading, A.M., 2008. Anomalous lithosphere beneath the Proterozoic ofwestern and central Australia: a record of continental collision and intraplate de-formation? Precambrian Research 166, 111–121.

Fishwick, S., Kennett, B.L.N., Reading, A.M., 2005. Contrasts in lithospheric structurewithin the Australian Craton. Earth and Planetary Science Letters 231, 163–176.

Fishwick, S., Heintz, M., Kennett, B.L.N., Reading, A.M., Yoshizawa, K., 2008. Steps inlithospheric thickness within eastern Australia, evidence from surface wave to-mography. Tectonics 27, TC0049. http://dx.doi.org/10.1029/2007TC002116.

Goleby, B.R., Shaw, R.S., Wright, C., Kennett, B.L.N., Lambeck, K., 1989. Geophysical ev-idence for ‘thick-skinned’ crustal deformation in central Australia. Nature 337,325–330.

Haines, P.W., Hand, M., Sandiford, M., 2001. Palaeozoic synorogenic sedimentation in cen-tral and northern Australia: a review of distribution and timing with implications forthe evolution of intracontinental orogens. Australian Journal of Earth Sciences 48,911–928.

Iaffaldano, G., Bunge, H.-P., Bruecker, M., 2007. Mountain belt growth inferred fromhistories of past plate convergence: a new tectonic inverse problem. Earth andPlanetary Science Letters 260, 516–523.

Iaffaldano, G., Husson, L., Bunge, H.-P., 2011. Monsoon speeds up Indian plate motion.Earth and Planetary Science Letters 304, 503–510.

Kaiho, Y., Kennett, B.L.N., 2000. Three-dimensional seismic structure beneath theAustralian region from refracted wave observations. Geophysical Journal Interna-tional 142, 651–668.

Kennett, B.L.N., Salmon, M., 2012. AuSREM: the Australian seismological referencemodel. Australian Journal of Earth Sciences 27. http://dx.doi.org/10.1080/08120099.2012.736406.

Kennett, B.L.N., Engdahl, E.R., Buland, R., 1995. Constraints on seismic velocities in theEarth from travel times. Geophysical Journal International 122, 108–124.

Kennett, B.L.N., Salmon,M., Saygin, E., AusMohoWorking Group, 2011. AusMoho: the var-iation of Moho depth in Australia. Geophysical Journal International 187, 946–958.

Korsch, R., Koscitin, N., 2010. GOMA (Gawler Craton–Officer Basin–Musgrave Province–Amadeus Basin) seismic and MT workshop 2010 Extended Abstracts. GeoscienceAustralia Record 2010/039.

Korsch, R., Blewett, R.S., Close, D.F., Scrimgeour, I.R., Huston, D.L., Kositcin, N., Whelan,J.A., Carr, L.K., Duan, J., 2011. Geological interpretation and geodynamic implica-tions of the deep seismic reflection and MT line 09GA-GA1: Georgina Basin-Arunta Region, in Northern Territory Geological Survey Record 2011-003.

Lambeck, K., 1983. Structure and evolution of the intracratonic basins of central Australia.Geophysical Journal of the Royal Astronomical Society 74, 843–886.

Lambeck, K., Burgess, G., Shaw, R.D., 1988. Teleseismic travel-time anomalies and deepcrustal structure in central Australia. Geophysical Journal International 94, 105–124.

McQueen, H.W.S., Lambeck, K., 1996. Determination of crustal structure in central Australiaby inversion of traveltime residuals. Geophysical Journal International 126, 645–662.

Mitrovica, J.X., Forte, A.M., 2004. A new inference of mantle viscosity based upon jointinversion of convection and glacial isostatic adjustment. Earth and Planetary Sci-ence Letters 255, 177–189.

Reading, A.M., Kennett, B.L.N., Goleby, B., 2007. New constraints on the seismic struc-ture of West Australia: Evidence for terrane stabilization prior to the assembly ofan ancient continent? Geology 35, 379–382.

Roberts, E.A., Houseman, G.A., 2001. Geodynamics of central Australia during the intra-plate Alice Springs Orogeny: thin viscous sheet models. In: Miller, J.A., Holdworth,R.E., Buick, I.S., Hand, M. (Eds.), Continental Reactivation and Reworking. The Geo-logical Society of London, pp. 139–164.

Saygin, E., Kennett, B.L.N., 2010. Ambient seismic noise tomography of Australian con-tinent. Tectonophysics 481, 116–125.

Saygin, E., Kennett, B.L.N., 2012. Crustal structure of Australia from ambient seismic noisetomography. Journal of Geophysical Research117, B01304. http://dx.doi.org/10.1029/2011JB008403.

Smithies, R.H., Howard, H.M., Evins, P.M., Kirkland, C.L., Kelsey, D.E., Hand, M., Wingate,M.T.D., Collins, A.S., Belousova, E., 2011. High-temperature granite magnatism,crust–mantle interaction and the Mesoproterozoic intracontinental evolution ofthe Musgrave Province, Central Australia. Journal of Petrology 52, 931–958.

Suppe, J., 2007. Absolute fault and crustal strength from wedge tapers. Geology 35,1127–1130.

Veevers, J.J., 2000. Billion-year EarthHistoryof Australia andNeighbours in Gondwanaland.Gemoc Press.

Walsh, A.K., Raimondo, T., Kelsey, D.E., Hand, M., Pfitzner, H.L., Clark, C., 2013. Durationof high-pressure metamorphism and cooling during the intraplate PetermannOrogeny. Gondwana Research 24, 969–983 (this volume).

Walter, M.R., Veevers, J.J., Calver, C.R., Grey, K., 1995. Neoproterozoic stratigraphy ofthe Centralian Superbasin, Australia. Precambrian Research 73, 173–195.

Waples, D.W., Waples, J.S., 2004. A review and evaluation of specific heat capacities ofrocks, minerals, and subsurface fluids. Part 1: minerals and nonporous rocks. NaturalResources Research 13, 97–122.

Yoshizawa, K., Kennett, B.L.N., 2004. Multi-mode surface wave tomography for theAustralian region using a 3-stage approach incorporating finite frequency effects.Journal of Geophysical Research 109, B02310. http://dx.doi.org/10.1029/2002JB002254.

Yoshizawa, K., Kennett, B.L.N., 2012. Mapping the Upper Mantle with Multi-modeSurface Waves: Application to the Australian Region, Abstract 2300. InternationalGeological Convention, Brisbane.

Zhao, J.-X., McCulloch, M.T., Korsch, R.J., 1994. Characterisation of a plume-related 800Ma magmatic event and its implications for basin formation in central southernAustralia. Earth and Planetary Science Letters 121, 349–367.