Precambrian Research 229 (2013) 93–104

Contents lists available at SciVerse ScienceDirect

Precambrian Research

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

Archean gravity-driven tectonics on hot and flooded continents: Controls onlong-lived mineralised hydrothermal systems away from continental margins

N. Thébauda,!, P.F. Reyb

a Centre for Exploration Targeting, University of Western Australia, School of Earth and Environment, M006, 35 Stirling Highway, WA 6009, Perth, Australiab Earthbyte Research Group, School of Geosciences, University of Sydney, NSW 2006, Sydney, Australia

a r t i c l e i n f o

Article history:Received 1 February 2011Received in revised form 17 February 2012Accepted 1 March 2012Available online 8 March 2012

Keywords:SagductionNumerical simulationArchean TectonicThermal evolutionGold mineralisation

a b s t r a c t

We present the results of two-dimensional numerical modelling experiments on the thermal evolutionof Archean greenstones as they sink into a less dense, hot and weak felsic crust. We compare this thermalevolution to that obtained via the analysis of isotopic data and fluid inclusion microthermometry dataobtained in the Paleoarchean to Mesoarchean Warrawoona Synform (Eastern Pilbara Craton, WesternAustralia). Our numerical experiments reveal a two-stage evolution. In the first stage, cooling affectszones of downwelling as greenstone belts are advected downward, whereas adjacent domes becomewarmer as deep and hot material is advected upward. We show that this is consistent with stable isotopesdata from the Warrawoona Synform, which reveal an early episode of seafloor-like alteration (90–160 "C)strongly focused along steeply dipping shear zones. In a second long-lived stage, lateral heat exchangesbetween domes and basins dominate the system as domes cool down while downwelling zones becomeincreasingly warmer. In the Warrawoona greenstone belt, stable isotopes in gold-bearing quartz veinspost-dating the sagduction-related vertical fabrics reveal that rock–fluid interaction occurred at muchhigher temperatures (234–372 "C) than seafloor-like alteration. We propose that emplacement of thickand dense continental flood basalts, on flooded hot and weak continental plates, led to conditions partic-ularly favourable to hydrothermal processes and the formation of mineral deposits. We further argue thatsagduction was able to drive crustal-scale deformation in the interior of continents, away from plate mar-gins. On largely flooded continents, sagduction-related shear zones acted as fluid pathways promotinggold mineralisation far away from active plate boundaries, continental rift zones or collisional mountainbelts.

© 2012 Elsevier B.V. All rights reserved.

1. Introduction

Several features made the Paleo- to Mesoarchean landscape(3.6–2.8 Gyr old) fundamentally different than it is today, andparticularly favourable to the formation and preservation of golddeposits. First, most Archean cratons are buried under up to 15 kmthick blanket of autochthonous Continental Flood Basalts (CFBs,e.g. Maurice et al., 2009), which include high-magnesium basaltsand komatiites that have only few equivalents on modern Earth.Second, despite their thicknesses most of these thick volcanic pileswere emplaced below, and remained below, sea level for mostof their Archean history (Arndt, 1999; Flament et al., 2008, 2011;Kump and Barley, 2007). This characteristic is of great importanceas it implies that an infinite fluid reservoir was available to feedhydrothermal circulations in the CBFs. Third, radiogenic heat pro-duction in the continental crust was much higher than it is today

! Corresponding author. Tel.: +61 86188 7139; fax: +61 86488 1178.E-mail address: nicolas.thebaud@uwa.edu.au (N. Thébaud).

which, in combination to the thermal insulation effect of CFBs,provided the heat engine, in the form of a strong thermal gradient,to power hydrothermal cells in the CFBs. Fourth, Paleoarcheanto Mesoarchean cratons throughout the world are characterisedby ovoid granitic domes, 40–100 km in diameter, encircled bygreenstone belts. Greenstone belts form strongly foliated ver-tical sheets connected through vertical triple junctions whereconstrictional fabrics dominate (Bouhallier et al., 1995; Chardonet al., 1996; McGregor, 1951). This dome and basin pattern, whichcharacterises many Archean cratons, is commonly interpreted interms of gravitational sinking (sagduction) of dense greenstonebelts into hot, and therefore weak, felsic crust (Chardon et al., 1996,1998; Collins et al., 1998; Dixon and Summers, 1983; Mareschaland West, 1980; McGregor, 1951). This process is often wronglydescribed as “vertical tectonics”. Indeed, during sagduction hor-izontal and vertical displacements are perfectly coupled. Domescan rise and greenstone keels can sink because horizontal dis-placements provide the necessary space for vertical mass transfer(e.g. Mareschal and West, 1980). This process is also often wronglyopposed to plate boundary driven deformation. Indeed there is

0301-9268/$ – see front matter © 2012 Elsevier B.V. All rights reserved.doi:10.1016/j.precamres.2012.03.001

94 N. Thébaud, P.F. Rey / Precambrian Research 229 (2013) 93–104

no mechanical incompatibility between sagduction and platetectonics processes as both can coexist spatially and temporarily(e.g. Bloem et al., 1997; Rey et al., 2003). While Archean lode golddeposits are interpreted traditionally as hydrothermal systemsdeveloped along convergent margins (Groves et al., 1998), sagduc-tion provided a mechanism to deform the interior of continents,hence greatly enlarging the domain where mineral deposits couldform. Fifth, considerations on the rheology of Archean continentallithospheres suggest that continents in the Archean were unable tosustain elevated mountain belts and high orogenic plateaux (Reyand Houseman, 2006; Rey and Coltice, 2008). Elevations > 3000 mwould have been absent at the surface of the Earth, which wouldhave strongly limited erosion, providing an explanation for theremarkable preservation of Archean supracrustal environments.We suggest that these Archean specificities provide some ratio-nales to explain the exceptional endowment in mineral resourcesin general, and in gold in particular, of most Archean cratons whencompared to younger terranes (Goldfarb et al., 2001).

In this paper, we present results of two-dimensional coupledthermo-mechanical numerical experiments on the sagduction ofgreenstones into a hot and weak felsic crust. We compare the mod-elled thermal evolution of greenstones to that obtained throughstable isotopes analyses and fluid inclusions microthermometryon shear zones from the Warrawoona Synform, in the East Pil-bara Granite-Greenstone Terrane (EPGGT, Fig. 1a). Results point toa protracted two-stage hydrothermal history involving: (1) low-temperature syn-deformation fluid–rock interactions during theearly stage of sagduction, and (2) mid-temperature syn- to post-deformation fluid–rock interactions during thermal relaxation andwarming of sagducted greenstone sheets. We argue that sagduc-tion of thick greenstone piles, with a large and near-permanentseawater reservoir above and a hot and weak basement below,led to efficient, crustal-scale and long-lived plumbing systems thatfavoured fluid–rock interactions and the genesis of gold deposits inthe interior of Paleoarchean to Mesoarchean cratons.

2. The East Pilbara Craton: an example ofsagduction-related “simple structural complexity”

The East Pilbara Granite Greenstone Terrane (Fig. 1a) is one ofthe best documented examples of dome-and-basin pattern inter-preted in terms of gravitational instability (Collins, 1989; Deloret al., 1991; Hickman, 1983; Teyssier et al., 1990; Thébaud, 2006;Van Kranendonk et al., 2004a), although alternative views exists(Blewett, 2002; Kloppenburg et al., 2001). The East Pilbara GraniteGreenstone Terrane provides a Paleoarchean to Mesoarchean geo-logic backdrop to study tectono-thermal processes and fluid–rockinteractions in a sagduction setting. It preserves a geologic historyoften largely overprinted in many Neoarchean cratons. The bulkof the domes consist of 3324–3300 Ma old syn- to post-kinematicsuites of high-K granitic suites that are derived from, and intru-sive in, older 3460–3430 Ma Tonalite–Trondjhemite–Granodiorite(TTG) gneisses and greenstones of the Warrawoona Group(Hickman, 1983; Hickman and Van Kranendonk, 2004; Smithieset al., 2003; Van Kranendonk et al., 2002, 2007). The domes arethemselves intruded by younger, mainly 3300–3240 Ma old, gran-ites. The older TTG gneisses formed the basement of greenstones,the emplacement of which is associated with two major volcaniccycles starting with the deposition of Warrawoona Group at ca.3490 Ma and followed by the deposition of the Kelly Group at ca.3335 Ma (Fig. 2) (Van Kranendonk et al., 2007). The thicknessesof the greenstone covers show significant lateral variation from8 to 12 km for the Warrawoona Group, and 4–9 km for the KellyGroup (e.g. Hickman, 1983; Van Kranendonk et al., 2007). Con-siderations on both the present crustal thickness of the Pilbara

(35–37 km) and the average erosion level (ca. 7 km) suggest thatthe greenstones accumulated on top of a 30–35 km thick conti-nental crust. Ubiquitous pillow-lava, hydrothermal cherts, VMSmineralisation, epidote-chlorite-Ca–Na-plagioclase-Ca-amphibolesecondary mineral assemblages and intense silicification through-out the greenstone pile imply subaqueous emplacement (Barley,1984; Barley and Pickard, 1999; Buick and Barnes, 1984; DiMarcoand Lowe, 1989; Van Kranendonk, 2006).

The structural complexity of greenstone covers is a function oftheir position with respect to the domes. On the NE flank of theMount Edgar dome, the Marble Bar greenstone belt lies directlyon top of the Mount Edgar granitic complex and is structurallypart of the dome (Fig. 1b). In the Marble Bar greenstone belt, thestratigraphy of the Warrawoona and Kelly Groups is relatively wellpreserved and structurally simple with a monotonous dip of 20–60"

to the NE. In contrast, in the Warrawoona synform (white star inFig. 1b1), greenstones are steeply dipping and strongly deformedagainst the near-vertical southwest margin of the Mount Edgardome (Fig. 3). At this location, the greenstone cover belongs toa basin pinched between the Mount Edgar dome to the northand the Corunna Down dome to the south. Compared to that ofthe Marble Bar greenstone belt, the structure in the Warrawoonasynform is more complex showing multiple phases of folding, aprominent horizontal to vertical stretching lineation, and numer-ous shear zones and quartz vein arrays (e.g. Thébaud et al., 2008).Contrasting, yet predictable structural complexity, with complexstructures above downwelling regions (e.g. Warrawoona syncline)and simpler structures above rising domes (e.g. Marble Bar belt) isa key attribute of sagduction settings (e.g. Bouhallier et al., 1995;Thébaud, 2006).

3. Numerical experiment setup

In order to interpret the thermal history derived from isotopicstudies, and to understand fluid flow in a sagduction setting, wehave performed a series of numerical experiments to documentthe thermal and mechanical history of sagducted greenstones.The process of sagduction has been tested through numerical andphysical experiments (de Bremond d‘Ars et al., 1999; Dixon andSummers, 1983; Mareschal and West, 1980; Robin and Bailey,2009; West and Mareschal, 1979). This process is driven by theneed for minimisation of internal gravitational potential energyof a system involving a density inversion (denser layer above alayer of lower density). It is resisted by the viscosity of the sys-tem, in particular by that of the stronger layer involved in thesagduction. Hence, the timing of sagduction is inversely propor-tional to the density contrast and proportional to the viscosity ofthe stronger layer. The emplacement of greenstones increases thegeothermal gradient (e.g. Rey et al., 2003; Sandiford et al., 2004;West and Mareschal, 1979), which in turn reduces the viscosityand accelerates sagduction. To model this process, we use Ellip-sis, a Lagrangian integration point finite element code capable oftracking time dependent variables in combination with an Eule-rian mesh. This coupled Lagrangian/Eulerian approach allows forthe accurate tracking of density interfaces during large deformation(Moresi et al., 2001, 2002). We use viscoplastic rheologies mim-icking standard rheological profiles for the continental lithosphere(e.g. Brace and Kohlstedt, 1980). Our experiments include realisticgeotherms with self-radiogenic heating and partial melting withfeedback on viscosities and densities.

In our numerical experiments, the greenstone cover is 15 kmthick in average, consistent with the thickness of many Archeangreenstone covers, and in particular the average cumulative thick-ness of the Warrawoona and Kelly Groups in the EPGG (e.g. VanKranendonk et al., 2007) to which we will confront our results.

N. Thébaud, P.F. Rey / Precambrian Research 229 (2013) 93–104 95

Fig. 1. (a) Simplified Geological Map of the North Pilbara Terrain modified after Van Kranendonk et al. (2002). EPGGT, East Pilbara Granite-Greenstone Terrane; CPTZ, CentralPilbara Tectonic Zone; WPGGT, West Pilbara Granite-Greenstone Terrane; MB, Marble bar greenstone belt. Legend: (1) Greenstones, (2) Granitoid complexes, (3) Sedimentarybasins, (4) Hamersley basin, (5) Phanerozoic cover. (b1) Simplified structural sketch of Mount Edgar and Corunna Down granitic domes complexes and surrounding regions.(b2) The first derivative of the topography shows that domes (thick dashed lines) are characterised by gentle topography and include both granitic rocks (crossed region onthe right panel) and greenstone cover (grey region). In contrast a rough topography characterises the basins surrounding the domes. The mining districts of Klondike in theWarrawoona Synform (white star), and the mining district of Bamboo Creek (white hexagon) are both located above downwelling regions.

Small thickness variations are introduced to allow for the initia-tion of sagduction. Gravity modelling in the East Pilbara GraniteGreenstone Terrane (Blewett et al., 2004) constraints the bulkdensity of our CFBs (2840 kg m#3). These CFBs are emplaced inthree successive 2.5 km thick layers at t0, t0 + 30 myr (lower War-rawoona Group), t0 + 60 myr (upper Warrawoona Group), plus a7.5 km thick layer (the Kelly Group) emplaced at t0 + 140 myr.Assuming t0 = 3490 Ma, this emplacement history approximatesthat of the Warrawoona and Kelly Groups (e.g. Bagas et al., 2002;Van Kranendonk et al., 2007). Our model CFBs are deposited on top

of a 30 km thick basement to which we assign a depth-independentdensity of 2720 kg m#3. In comparison to Robin and Bailey (2009),our experiments use a smaller density contrast between greenstoneand basement (120 kg m#3 vs 200 kg m#3), a thicker greenstonecover (15 km vs 10 km), and the emplacement of the greenstonecover follows a stepwise history over 140 myr rather than an instan-taneous emplacement.

Deformation in our models follows the mechanism that requiresthe least differential stress amongst: frictional faulting (Coulombcriteria), plastic faulting and viscous creep. The frictional faulting is

96 N. Thébaud, P.F. Rey / Precambrian Research 229 (2013) 93–104

Fig. 2. Lithostratigraphy of the Pilbara Supergroup. Geochronology references: a (Buick et al., 1995), b (Thorpe et al., 1992), c (McNaughton et al., 1993) and d (Nelson, 1999,2000, 2001).

Table 1List of parameters used in the models.

Parameters Value(s) Unit

(a) Mechanical parametersg: Acceleration of gravity field 9.81 m s#2

!atm: Air density 2.0 kg m#3

!cfb: CFB density 2840 kg m#3

!cc: Crustal density 2720 kg m#3

!m: Mantle density 3310 kg m#3

Ccfb: CFB cohesion 10 MPaCcc: Crustal cohesion 10 MPaCm: Mantle cohesion 40 MPa"a: Strain weakening factor 0.2"0: Strain from which weakening is maximum 0.5"n: Strain weakening sensitivity to accumulated strain 0.5#cfb: CFBs maximum yield stress 100 MPa#cc: Crustal maximum yield stress 250 MPa#m: Mantle maximum yield stress 400 MPa$cfb: CFBs internal angle of friction 5 "

$cc: Crustal internal angle of friction 15 "

$m: Mantle internal angle of friction 25 "

%air: Air viscosity 5 $ 1018 Pa sAcfb: CFBs pre-exponential constant 5 $ 10#5 MPa#n s#1

Acc: Crustal pre-exponential constant 5 $ 10#6 MPa#n s#1

Am: Mantle pre-exponential constant 7 $ 104 MPa#n s#1

ncfb: CFBs stress exponent 3ncc: Crustal stress exponent 3nm: Mantle stress exponent 3Qcfb: CFBs Activation enthalpy 1.9 $ 105 J mol#1

Qcc: Crustal activation enthalpy 1.9 $ 105 J mol#1

Qm: Mantle activation enthalpy 5.2 $ 105 J mol#1

R: Gas constant 8.3145 J mol#1 K#1

(b) Thermal parameters˛m: Mantle coefficient of thermal expansion 2.8 $ 10#5 K#1

&: Thermal diffusivity 0.9 $ 10#6 m2 s#1

Cp: Heat capacity 1000 J kg#1 K#1

Hcfb: CFBs heat production 1.335 $ 10#7 W m#3

Hcc: Crustal heat production 1.335 $ 10#6 W m#3

Hcc: Mantle heat production 0 W m#3

qm: Basal mantle heat flux 0.025 W m#2

(c) Partial melting parametersLatent heat 250,000 JCFBs: Solidus (P) = 993–1.2 $ 10#7·P + 1.2 $ 10#16·P "CLiquidus (P) = 1493–1.2 $ 10#7·P + 1.2 $ 10#16 aPBasement: Solidus (P) = 983–9.37 $ 10#8·P + 6.32 $ 10#17·P "CLiquidus (P) = 1393–9.37 $ 10#8·P + 6.32 $ 10#17·Pa Non listed values for the CFB are taken equal to crustal values.

N. Thébaud, P.F. Rey / Precambrian Research 229 (2013) 93–104 97

Fig. 3. Geological map of the Warrawoona Synform (after Collins et al., 1998; Kloppenburg et al., 2001), localisation on the bottom right panel in Fig. 1.

described by a yield stress: #yield = (C0 + tan (˚) #n) f("), in which C0is the cohesion at atmospheric pressure; ˚ is the angle of internalfriction, #n is the normal stress. Strain localisation is achieved viastrain weakening function f("), which reduces the yield stress as theaccumulated strain (") increases. The strain weakening function

is given by: f (") =!

1 # (1 # "a)("/"0)"n , " < "0"a, " % "0

where " is the

accumulated strain, taken as the second invariant of the deviatoricplastic strain tensor, and "0 is the “saturation” strain from whichthe yield stress is reduced by a proportion "a. "n modulates thedependency between accumulated strain and strain weakening.The yield stress has an upper limiting value #crit, which describessemi-brittle deformation independent of pressure (Ord and Hobbs,1989). For differential stresses that attain the yield stress, the

material fails and deformation is modelled by an effective viscos-ity: nyield = 'yield/2E in which E is the second invariant of the strainrate tensor. In our model, viscous rheologies are based on bothdislocation creep and diffusion creep (Bürgman and Dresen, 2008).This choice seems justified given that (i) diffusion creep is favouredby relatively small gravitational differential stress such as the oneinitiating sagduction, and (ii) diffusion creep is an important flowmechanism in feldspar-bearing assemblages prevalent in TTG.Therefore, viscous creep in our model crust is modelled usingthe fastest of dislocation creep and diffusion creep mechanisms(Turcotte and Schubert, 1982) following: %eff = (1/%dis + 1/%diff)#1.Both creep regimes can be described by an Arrhenius formulation(Karato and Wu, 1993): % = (N0Exp[Q/(RT)]E1#n)1/n with E thesecond invariant of the strain rate tensor, Q the activation energy

98 N. Thébaud, P.F. Rey / Precambrian Research 229 (2013) 93–104

of the creep mechanism, and n the power law stress exponent.For dislocation creep viscosity: N0dis = 1/(2ndisAdis), with Adis thepre-exponential factor in the power law (parameter values fromBrace and Kohlstedt, 1980, Table 1). In the crust, diffusion creepbecomes important under low differential stress, high tempera-ture, and when partial melt is present (Mecklenburgh and Rutter,2003). For many rocks, parameters Q and A for diffusion creepare unknown. However, they must lead to faster strain rate atdifferential stress lower than a given transition differential stress(here #trs = 30 MPa) at which both mechanisms result in the samestrain rate. In our models, we assume that the activation energiesfor dislocation creep and diffusion creep are the same and we dropthe N0 factor for diffusion creep to a value for which the dislocationand diffusion viscosities are equal at the transition differentialstress. For diffusion creep, viscosity n = 1 and N0 is obtained bymatching both diffusive and dislocation strain rates at the transi-tion differential stress (#trs): N0dif = N0dis = 1/(2ndisAdis)·#trs

1#ndis.We disregard diffusion creep in the mantle because the transitiondifferential stress between dislocation creep and diffusion creep islow (ca. 0.5 MPa, Turcotte and Schubert, 1982). In all experiments,viscosities are clipped beyond a minimum of 5 $ 1018 Pa s and amaximum of 5 $ 1022 Pa s. All parameters are reported in Table 1.

In most CFBs, magmatic mineral assemblages are deeply alteredinto micas and talc-bearing retrogressive assemblages. This alter-ation reduces both the density and strength of mafic rocks (deBremond d‘Ars et al., 1999). Hence, we assume an effective viscos-ity one tenth of that of the basement. This is achieved by imposingthat: N0cfb = 0.1N0Crust (i.e. Acfb = 10ACrust, Table 1).

In our experiments, the radiogenic heat production in the crustis calculated backward in time to 3.5 Ga from the averaged crustalheat production of modern Archean crusts (Taylor and McLennan,1995). The total crustal radiogenic production is partitioned intothe pre-greenstone basement and we assume little heat produc-tion in the greenstone cover and none in the mantle (cf. Table 1).Assuming a basal heat flux of 25 $ 10#3 W m#2, the steady-stateMoho temperature is at 710 "C before the emplacement of the firstmember of the CFB. A free-slip boundary condition is applied tohorizontal and vertical boundaries and no plate boundary force isapplied. Deformation is achieved through gravity-driven processesinitiated by small variations in thickness of the greenstone cover(Fig. 4).

Partial melting impacts significantly on the thermo-mechanicalproperties of the system (temperature, viscosities and densities). Inour numerical experiments, viscosities decrease over three ordersof magnitude as the melt fraction increasing from 0.2 to 0.3.In nature, the viscosity drop reaches many orders of magnitude(Clemens and Petford, 1999), however only the first two or threeorders are likely to have mechanical significance. As the temper-ature increases from the solidus to the liquidus, the density ofthe partially melted region decreases by 13%, which increases itsbuoyancy. Solidus and liquidus parameters are reported in Table 1.

4. Results

Crustal flow: From t0 to t0 + 140 myr, the Moho temperatureincreases from 704 "C to 772 "C leading to onset of partial melt-ing, which is limited to only a few percent of melt at the verybase of the crust (Fig. 4b1). This temperature increase is due tothe thermal insulation of the radiogenic crust progressively buriedunder the greenstone cover (Mareschal and West, 1980; Rey et al.,2003; Sandiford et al., 2004; West and Mareschal, 1979). There isno mechanical disturbance of the greenstone-basement interface(Fig. 4b1), which is consistent with the concordant contact betweenthe Kelly Group and the Warrawoona Group (Van Kranendonket al., 2004b). At t0 + 150 myr, 10 myr after the deposition of the

7.5 km thick Kelly group, partial melting affects a growing portionof the lower crust facilitating the initiation of gravitational insta-bilities which strain the greenstone cover (Fig. 4b2). Convectivemotion affects the deep crust. From t0 + 150 myr to t0 + 158 myr Ma,a couple of gravitational instabilities develop, one at ca. 152 myr(Fig. 4b3), the other at ca. 158 myr (Fig. 4b6). Their length-scale (ca.100 km) and time-scale (10 myr) are compatible with those of thedome and basin pattern of the East Pilbara Granite Greenstone Ter-rane (Collins et al., 1998), and consistent with the results of Robinand Bailey (2009). In the upper crust, above downwelling regions, asmuch as 60% bulk horizontal shortening accommodates sagductiondownward flow. Upward motion in domes and downward motionin basins are coupled through horizontal motions in the upper andlower part of the crust. Notably, enclaves of greenstones are caughtinto the upward flow in domes and exhumed from the base of thecrust into the upper crust.

Thermal evolution: From the onset of sagduction at t0 + 148 myr,downward flow advects cool greenstone rocks in downwellingregions where the temperature at 10 km depth decreases from ca.210 to ca. 90 "C (Fig. 5a). In contrast, in rising domes the temper-ature at 10 km depth increases from ca. 210 to ca. 660 "C (Fig. 5a).In about 12 myr, partial convective overturn leads to the build-up of a long-wavelength (ca. 100 km) profound lateral thermalanomalies >500 "C through fast advective cooling and heating ofdownwelling and upwelling regions respectively (Fig. 5a). The max-imum horizontal temperature gradient reaches 26 "C/km. Fromt0 + 158 myr onward, following the main sagduction stage, tem-perature anomalies, around fully developed domes, decrease. Thissecond post-deformation stage, led to rapid conductive heating ofgreenstone basins and conductive cooling of domes (Fig. 5b). Thetemperature at 10 km depth in the basins increases from 90 to200 "C in ca. 8 myr, whereas the temperature at the same depthin the domes decreases from 660 to 310 "C (Fig. 5b). Our numeri-cal experiments also revealed that, because of the sheer size of thethermal anomalies, and in part because domes are diachronous,deformation and thermal relaxation last over 150–200 myr afterthe bulk of sagduction. This is compatible with the folding of theGorge Creek Group (<3240 Ma) as well as the hornblende Ar–Arages in the EPGGT that spread from 3350 to 3100 Ma (Davids et al.,1997; Kloppenburg et al., 2001; Wijbrans and McDougall, 1987).

5. Comparison to temperature records derived frompetrological and isotopic data

To verify the 2-stage thermal history documented in ournumerical experiments we compare it to that derived from geo-chemical and isotopic data from the Warrawoona Synform (Hustonet al., 2001; Thébaud et al., 2008). In the Warrawoona Synform,hydrothermal alteration has been described as strongly partitionedinto ductile shear zones that accommodated the sagduction of thegreenstones cover (Fig. 3) (Thébaud et al., 2006). The Fielding’sFind Shear Zone, parallel to the northern border of the felsic vol-canic Wyman Formation (Fig. 3), has been recently the focus of adetailed fluid–rock interaction study. A section across this shearzone shows that the increase in strain correlates with enhancedhydrothermal alteration, including strong silicification, coupled toa pronounced zoning of major elements, trace elements and oxygenisotopes values (Thébaud et al., 2008). The bulk rock ı18O val-ues increase towards the shear zone from 19 to 25‰, indicating afluid-buffered system (Fig. 6). This zonation is consistent with theshear zones having acted as fluids pathways. Similar extreme (upto 18‰) 18O-enrichments have been documented in metabasalts ofthe Superior Province (Kerrich et al., 1981), and in association withPaleoarchean to Mesoarchean hydrothermal vents in Barberton (deWit et al., 1982; Knauth and Lowe, 2003; Lowe and Byerly, 1986).

N. Thébaud, P.F. Rey / Precambrian Research 229 (2013) 93–104 99

Fig. 4. Initial settings (top box) and snapshots of the numerical experiment from t0 + 140 to t0 + 158 myr following the emplacement of the Kelly Group at t0 + 140 myr. Thehistory from t0 is t0 + 140 myr is not shown. This period corresponds to the progressive emplacement of the greenstone cover, and to a warming up of the crust. Duringthis time there is with little to no deformation. Thickness variations of the lower Warrawoona greenstone speedup initiation of sagduction. Two circular red markersrecord the horizontal versus vertical motions as well as the magnitude of shortening in the greenstone cover. From t0 + 140 to t0 + 158 myr, gravity-driven shortening abovethe downwelling region is >60%. Blue shading shows post yielding plastic strain. Arrows pointing at passive vertical markers in the basement document the deformationpattern.

100 N. Thébaud, P.F. Rey / Precambrian Research 229 (2013) 93–104

0 100 200 3000

200

400

600

800

t0 + 140 myrt0 + 148 myr

t0 + 158 myr

T (ºC

)

Distance (km)

Upwelling Upwelling Upwelling

Downwelling Downwelling

0 100 200 3000

200

400

600

800

t0 + 158myr

t0 + 162myr

t0 + 166 myr

T (ºC

)

Distance (km)

a) Temperature at 10 km depth, from t0+140 to t0+158 myr, building up of a thermal anomaly

b) Temperature at 10 km depth, from t0+158 to t0+166 myr, relaxation of the thermal anomaly

26 ºC/km

26 ºC/km

Fig. 5. Thermal evolution at 10 km depth. Upper graph shows the evolution from t0 + 140 to t0 + 158 myr following the emplacement of the Kelly Group. During this stage,the temperature in the basin decreases from 210 to 90 "C, whereas the temperature in the dome rises from 210 "C to 660 "C. This leads to a horizontal temperature differenceof ca. 570 "C over ca. 65 km distance, with a horizontal gradients up to 26 "C per km. From t0 + 158 to t0 + 166 myr (lower graph), the temperature evolution changes withcooling of the granitic domes from 660 to 310 "C and heating in the greenstone basin from 90 to 200 "C.

18O-enrichment up to ca 14‰ have been also described in ophiolitesand modern-day seafloor alteration zones and is interpreted as theproduct of hydrothermal convection cells involving seawater at lowtemperature (e.g. Putlitz et al., 2001). In Archean sagduction sett-ings, seafloor-like alteration would have been strongly enhanced bydeformation and the development of faults and shear zones aboveand within downwelling regions. In the Warrawoona Synform, syn-deformation alteration process was driven by interaction with largevolumes of low temperatures (&90–160 "C) seawater-derived flu-ids during the early stage of sagduction process (Thébaud et al.,2008). This is consistent with our numerical modelling showingthat downward advection of greenstones maintains a low temper-ature environment (<100 "C) in the top 10 km of the downwellinggreenstones.

In the Warrawoona Synform, ductile shear zones also host gold-bearing quartz ± calcite ± sulphide ± ankerite veins (Huston et al.,2001). These veins do not exist in the nearby less deformed MarbleBar Greenstone Belt, but they do in other downwelling regions suchas the Bamboo Creek mining district on the NE margin of MountEdgar (white hexagon in Fig. 1b1). These gold-bearing quartz veinsdevelop largely in association with shear zones in downwellingregions and have a syn- to post-shearing origin (Thébaud et al.,2006). Across the Warrawoona Synform, quartz veins display arestricted range of ı18O values with a mean value of +13.2 ± 2‰(Fig. 6). This suggests that quartz precipitated from a homogeneous

fluid under near-isothermal conditions. These veins are there-fore in isotopic disequilibrium with their silicified host, whichimplies an emplacement postdating the silicification (Thébaudet al., 2008). Quartz-hosted CO2–NaCl–H2O–CH4 fluid inclusionsyielded homogenisation temperatures between 234 and 372 "C(Huston et al., 2001; Thébaud et al., 2006). Using Zheng (1993)quartz/water fractionation equation, Thébaud et al. (2008) calcu-lated that the ı18O of the water from which these quartz veinsprecipitated is in the range of +2.4 to +12.1‰. This oxygen iso-topic composition, together with the significant enrichment inpotassium and base metals revealed by Synchrotron radiation X-ray fluorescence fluid inclusion analyses, support a metamorphicand/or magmatic source (Thébaud et al., 2006). These hotter fluidswere mobilised at a later stage of the sagduction process, possiblyduring the devolatilisation of the greenstones keels and/or duringthe large production of potassic melt (Thébaud et al., 2008).

In summary, field observation, isotopic and fluid inclusionsmicrothermometry data, and our numerical investigation pre-sented here, document a long-lived two-stage hydrothermalhistory involving: (1) low-temperature (90–160 "C) syn-deformation seafloor-like fluid–rock interactions strongly focusedalong steeply dipping ductile shear zones during the early stage ofsagduction, and (2) moderate-temperature (234–372 "C) syn- topost-deformation fluid–rock interactions, strongly focused alongshear zones, leading to the formation of gold-bearing quartz veins

N. Thébaud, P.F. Rey / Precambrian Research 229 (2013) 93–104 101

Fig. 6. Whole rock ı18O and wt% SiO2 and quartz vein ı18O plotted against a synthetic lithological section of the Fielding’s Find shear zone from the Warrawoona Synform(modified after Thébaud et al., 2008). Whole Rock analyses across the Fielding’s Find Shear Zone reveal that he bulk rock ı18O values together with silica content, from bothsides of the shear zone, increase progressively towards its centre. Such enrichment in heavy oxygen isotopes and silicification was proposed to be driven by interaction withlarge volumes of low temperature (&90–160 "C) hydrothermal fluids (sea-water) (Thébaud et al., 2008). The quartz veins are in isotopic disequilibrium with their alteredhost, which suggest that a fluid circulation event followed the silicification around the Fielding’s Find Shear Zone. Vein quartz display a relatively restricted range of ı18Ovalues across the Fielding’s Find shear zone, with a mean value of +13.2 ± 2‰. Together with fluid inclusion analyses the quartz veins are regarded as the signature of acomposition involving magmatic and/or metamorphic fluid sources (Thébaud et al., 2006).

post-dating the vertical fabrics. The early alteration assemblagescan be linked to the early stage of greenstone sagduction and theco-development of (i) crustal-scale thermal anomalies, duringwhich cooling affected the downwelling greenstone, and (ii)vertical shear zones which acted as efficient fluid pathways forsupracrustal porous flow (Fig. 7a). In the early stage of sagduc-tion, large volumes of fluid, trapped in hydrated minerals duringseafloor alteration, would have been transported deep into thecrust. The second hydrothermal stage can be linked to the releaseof metamorphic and/or magmatic fluid in association to thedevolatisation of sagducted greenstone covers, and crystallisationof felsic magmas (Fig. 7b). These fluids would have been releasedat a time when sagduction-related lateral thermal anomaliesentered the relaxation phase during which the temperature inthe greenstone keels rapidly increased. In the strongly foliated

steeply dipping greenstone sheets, these hot crustal fluids wouldhave been efficiently channelled toward the surface through thewell-established network of shear zones formed at an early stage.This is consistent with field observations showing that quartzveins are restricted to downwelling regions. The slow conductiverelaxation of the thermal anomaly may have helped to maintainthis plumbing system for several 10 s of myr.

6. Discussion and implication for Paleoarchean toMesoarchean gold mineralisation

To form a gold deposit, the average crustal gold content of1.3 ppb must be concentrated by one thousand to one hundredthousand times (e.g. Walshe and Cleverley, 2009). Temperaturegradients above shallow magmatic intrusions typically drive

102 N. Thébaud, P.F. Rey / Precambrian Research 229 (2013) 93–104

Fig. 7. Conceptual tectonograms illustrating Archean intracontinental mineral systems associated to gravity driven tectonics and fluid flow during sagduction of subaqueousgreenstone covers (green envelops) into their hot and weak basement (red envelops). During the early stage of sagduction (upper panel), early deformation enhances low-temperature seafloor-like alteration along downwelling regions. Sagduction transfers cooler supracrustal rocks and large volumes of hydrated minerals deep into the crust.As shown in our numerical experiments, dome regions become a few hundred degrees hotter than the intervening downwelling regions. These lateral thermal gradientscontribute to focus fluid circulations into downwelling regions. In a second stage (lower panel), the relaxation of this lateral temperature anomaly led to the rapid heatingand devolatisation of hydrated steeply dipping greenstone sheets. Released metamorphic and magmatic fluids focus into pre-existing shear zones, which act as crustal-scalefluid pathways to the surface where mineral deposits form. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of thearticle.)

efficient hydrothermal systems in which fluid–rock interactionis the main concentration mechanism (Eldursi et al., 2009). Itis therefore not surprising that shallow magmatic activities inback-arc settings, as well as sub-seafloor volcanism associatedwith continental rift systems, are the main settings of gold depositson the modern Earth (Goldfarb et al., 2001). In other words, miner-alised hydrothermal systems are strongly partitioned onto oceanfloors and active continental margins. However, it is likely thatmost Archean ocean floors and Archean continental plate marginsdid not survive the past 3 Gyr. This suggests that the bulk of thepreserved Archean gold deposits may have formed in the interiorof continents, away from plate margins. On the modern Earth,this setting is the least favourable for hydrothermal systems, butperhaps not so in the Archean.

In the Archean, basaltic piles up to 15 km thick emplaced onolder felsic continental crusts were common (de Wit and Ashwal,1997), and most of them emplaced onto flooded continents (Arndt,1999; Flament et al., 2008; Kump and Barley, 2007). Higher sealevel in the Archean would have maintained an infinite reservoirof water above the continental basalts, whereas higher surfaceheat flow, due to enhanced radiogenic heat production in the crustand due to CFBs thermal insulation, would have powered shallowhydrothermal cells at the interface between continent and sea-water, accounting for widespread low temperature alteration and

silicification observed in greenstone belts. Numerical and phys-ical experiments have shown that density inversions related tothe emplacement of thick continental basalts onto less dense andweak basement are large enough to drive the foundering of basalticpiles into the continental crust (Chardon et al., 1998; de Bremondd‘Ars et al., 1999; Dixon and Summers, 1983; Robin and Bailey,2009; West and Mareschal, 1979; this study). Our numerical exper-iments show that sagduction is capable of promoting significanthorizontal temperature gradients up to 26 "C/km. These horizontalgradients may have been large enough to force fluids, released dur-ing devolatisation of greenstones and crystallisation of domes, backinto the cooler greenstone keels, where steeply dipping shear zoneswould have channelised these hotter fluids toward the surface.Importantly, sagduction is a mechanism that explains crustal-scaledeformation within the inner-part of continental plates, away fromtheir margins. Therefore in the Archean, the association of volcanicrocks above a felsic crust but below sea level, in tectonically activeregions with high heat flow – critical to formation of gold deposits –was not restricted to continental rifting or active continental mar-gins. Sagduction on flooded continents would have promoted theformation intracontinental hydrothermal systems expanding con-siderably the prospectivity of Archean cratons, compared to that ofmodern continents whose prospectivity is limited to active platemargins and continental rifts.

N. Thébaud, P.F. Rey / Precambrian Research 229 (2013) 93–104 103

7. Conclusions

Using realistic rheologic and thermal parameters, our coupledthermo-mechanical numerical experiments confirm that, in theArchean, sagduction was a natural response to density inversionsand thermal weakening due to the emplacement of thick continen-tal flood basalts onto radiogenic crusts. Our experiments show that,during sagduction, the advection of cold rocks into downwellingbasins and advection of hot rocks into rising domes lead to crustal-scale, horizontal, thermal anomalies in excess to 500 "C over anhorizontal distance of ca. 30 km. Large temperature gradients up to26 "C/km and the availability of large volumes of fluids (i.e. (i) sea-water above flooded continents, (ii) metamorphic fluids releasedduring dehydration of sagducting supracrustal greenstones, and(iii) magmatic fluids released during crystallisation of magma indomes) would have powered crustal-scale hydrothermal systems.The formation around granitic domes of networks of sagduction-related steeply dipping greenstone sheets and vertical shear zonesconnected through vertical triple junctions, would have facilitatedcrustal-scale transfer of fluids, matter and heat. Fluid–rock inter-actions in these tectonically active plumbing systems would haveoverprinted low-temperature fluid–rock interactions that occurredduring the emplacement of CFBs onto flooded continents and dur-ing the early stage of sagduction. This is consistent with structural,petrological and isotopic data. Based on our numerical experi-ments, field observations and existing petrological and isotopicdata, we propose that in the Archean, the sagduction of thickgreenstone piles with a large and near-permanent seawater reser-voir above, and a hot and weak basement below, allowed for thedevelopment of efficient and long-lived plumbing systems. Thisplumbing systems favoured crustal-scale fluid–rock interactionsand the formation of craton-wide gold deposits in the interior ofcontinents, far away from their margins, a context very differentfrom the paleoproterozoic and Phanerozoic, where orogenic andmagmatic orebodies are closely associated with active margin sett-ings.

Acknowledgements

This work was supported by the Australian Research Council’sDiscovery funding scheme (ARC DP 0342933) and AUSCOPE-NCRIS and Computational Infrastructure for Geodynamics softwareinfrastructure. This paper has benefited from the reviews of JeanBébard and Laurent Aillères, and from countless discussions withN. Flament, N. Coltice, P. Philippot, F. Wellmann, M. Fiorentini, K.Gessner, M. Van Kranendonk and many other colleagues from theUniversity of Sydney, the University of Paris 6 and the Centre forExploration Targeting in Perth.

References

Arndt, N., 1999. Why was flood volcanism on submerged continental platforms socommon in the Precambrian? Precambrian Research 97, 155–164.

Bagas, L., Smithies, R.H., Champion, D.C., 2002. Geochemistry of the Corunna DownsGranitoid Complex, East Pilbara Granite/Greenstone Terrane, Western Australia.Geological Survey of Western Australia - Annual Review Technical Papers2001–02, 61–69.

Barley, M.E., 1984. Volcanism and hydrothermal alteration, Warrawoona Group, EastPilbara. In: Muhling, J.R., Groves, D.I., Blake, T.S. (Eds.), Archaean and ProterozoicBasins of the Pilbara, Western Australia: Evolution and Mineralization Poten-tial. University of Western Australia, Geology Department, University ExtensionSpecial Publication, pp. 23–36.

Barley, M.E., Pickard, A.L., 1999. An extensive, crustally-derived, 3325 to 3310 Masilicic volcanoplutonic suite in the eastern Pilbara Craton: evidence from theKelly Belt, McPhee Dome and Corunna Downs Batholith. Precambrian Research60, 185–241.

Blewett, R.S., 2002. Archaean tectonic processes: a case for horizontal shortening inthe North Pilbara Granite-Greenstone Terrane, Western Australia. PrecambrianResearch 113, 87–120.

Blewett, R.S., Shevchenko, S., Bell, B., 2004. The North Pole Dome: a non-diapiricdome in the Archaean Pilbara Craton, Western Australia. Precambrian Research5, 105–120.

Bloem, E.J.M., Dalstra, H.J., Ridley, J.R., Groves, D.I., 1997. Granitoid diapirism duringprotracted tectonism in an Archaean granitoid-greenstone belt, Yilgarn Block,Western Australia. Precambrian Research 85, 147–171.

Bouhallier, H., Chardon, D., Choukroune, P., 1995. Strain patterns in Archaean dome-and-basin patterns: the Dharwar craton (Karnataka, India). Earth and PlanetaryScience Letters 135, 57–75.

Brace, W.F., Kohlstedt, D.L., 1980. Limits on lithospheric stress imposed by laboratoryexperiments. Journal of Geophysical Research 85, 6248–6252.

Buick, R., Barnes, K.R., 1984. Cherts in the Warrawoona Group: Early Archean Silici-fied Sediments Deposited in Shallow Water Environments, vol. 9. University ofWestern Australia, Geology Department, University Extension Special Publica-tion, pp. 37–53.

Buick, R., Thornett, J.R., McNaughton, N.J., Smith, J.B., Barley, M.E., Savage, M., 1995.Record of emergent continental crust &3.5 billion years ago in the Pilbara Cratonof Australia. Nature 375, 574–577.

Bürgman, R., Dresen, G., 2008. Rheology of the Lower Crust and Upper Mantle: evi-dence from rock mechanics, geodesy, and field observations. Annual Review ofEarth and Planetary Sciences 36, 531–567.

Chardon, D., Choukroune, P., Jayananda, M., 1996. Strain, patterns, decollement andincipient sagducted greenstone terrains in the Archaean Dharwar craton (southIndia). Journal of Structural Geology 18, 991–1004.

Chardon, D., Choukroune, P., Jayananda, M., 1998. Sinking of the Dharwar Basin(South India): implications for Archaean tectonics. Precambrian Research 91,15–39.

Clemens, J.D., Petford, N., 1999. Granitic melt viscosity and silicic magma dynamics incontrasting tectonic settings. Journal of the Geological Society 156, 1057–1060.

Collins, W.J., 1989. Polydiapirism of the Archean Mount Edgar Batholith, PilbaraBlock, Western Australia. Precambrian Research 43, 41–62.

Collins, W.J., Van Kranendonk, M.J., Teyssier, C., 1998. Partial convective over-turn of Archaean crust in the east Pilbara Craton, Western Australia: drivingmechanisms and tectonic implications. Journal of Structural Geology 20,1405–1424.

Davids, C., Wijbrans, J.R., White, S.H., 1997. 40Ar/39Ar laser probe ages of meta-morphic hornblendes from the Coongan Belt, Pilbara, Western Australia.Precambrian Research 83, 221–242.

de Bremond d‘Ars, J., Lecuyer, C., Chardon, D., 1999. Hydrothermalism and diapirismin the Archean: gravitational instability constraints. Tectonophysics 304, 29–39.

de Wit, M.J., Ashwal, M.D., 1997. Greenstone Belts (Oxford).de Wit, M.J., Hart, R., Martin, A., Abbot, P., 1982. Archaean abiogenetic and probable

biogenetic structures associated with mineralized hydrothermal vent systemsand regional metassomatism, with implications for the greenstone belt studies.Economic Geology and the Bulletin of the Society of Economic Geologists 77,1783–1802.

Delor, C.A., Burg, J.P., Clarke, G., 1991. Relations diapirisme-metamorphisme dans laProvince du Pilbara (Australie Occidentale): implications pour les régimes ther-miques et tectoniques à l‘Archéen. Comptes Rendus de l‘Académie des Sciencesde Paris 312, 257–263.

DiMarco, M.J., Lowe, D.R., 1989. Petrography and provenance of silicified earlyArchaean volcaniclastic sandstones, eastern Pilbara Block, Western Australia.Sedimentology 36, 821–836.

Dixon, J.M., Summers, J.M., 1983. Patterns of total and incremental strain in subsidingtroughs: experimental centrifuged models of inter-diapir synclines. CanadianJournal of Earth Sciences 20, 1843–1861.

Eldursi, K., Branquet, Y., Guillou-Frottier, L., Marcoux, E., 2009. Numerical investiga-tion of transient hydrothermal processes around intrusions: heat-transfer andfluid-circulation controlled mineralization patterns. Earth and Planetary ScienceLetters 288, 70–80.

Flament, N., Coltice, N., Rey, P.F., 2008. A case for late-Archaean continental emer-gence from thermal evolution models and hypsometry. Earth and PlanetaryScience Letters 275, 326–336.

Flament, N., Rey, P.F.C.N., Dromart, G., Olivier, N., 2011. Lower crustal flow keptArchean continental flood basalts at sea level. Geology 39, 1159–1162.

Goldfarb, R., Groves, D., Gardoll, S., 2001. Orogenic gold and geologic time: a globalsynthesis. Ore Geology Reviews 18, 1–75.

Groves, D.I., Goldfarb, R.J., Gebre-Mariam, M., Hagemann, S.G., Robert, F., 1998. Oro-genic gold deposits: a proposed classification in the context of their crustaldistribution and relationship to other gold deposit types. Ore Geology Reviews13, 7–27.

Hickman, A.H., 1983. Geology of the Pilbara Block and its Environs. 1981/36 (Perth).Hickman, A.H., Van Kranendonk, M.J., 2004. Diapiric processes in the formation of

Archaean continental crust, East Pilbara Granite-Greenstone Terrane, Australia.In: The Precambrian Earth: Tempos and Events. Elsevier.

Huston, D., Blewett, R., Mernaugh, T., Sun, S.-S., Kamprad, J., 2001. Gold Deposits ofthe Pilbara Craton: Results of AGSO Research, 1998–2000.

Karato, S., Wu, P., 1993. Rheology of the upper mantle a synthesis. Science 260,771–776.

Kerrich, R., Fryer, B., Milner, K.J., Peirce, M.G., 1981. The geochemistry of gold-bearingchemical sediments, Dickenson mine, Red Lake, Ontario: a reconnaissance study.Canadian Journal of Earth Sciences 18, 624–637.

Kloppenburg, A., White, S.H., Zegers, T.E., 2001. Structural evolution of the War-rawoona Greenstone Belt and adjoining granitoid complexes, Pilbara Craton,Australia: implications for Archaean tectonic processes. Precambrian Research112, 107–147.

104 N. Thébaud, P.F. Rey / Precambrian Research 229 (2013) 93–104

Knauth, L.P., Lowe, D.R., 2003. High Archaean climatic temperature inferred fromoxygen isotope geochemistry of cherts in the 3.5 Ga Swaziland Supergroup,South Africa. Geological Society of America Bulletin 115, 566–580.

Kump, L.R., Barley, M.E., 2007. Increased subaerial volcanism and the rise of atmo-spheric oxygen 2.5 billion years ago. Nature 448, 1033–1036.

Lowe, D.R., Byerly, G.R., 1986. Archaean flow-top alteration zones formed initiallyin a low-temperature sulphate-rich environment. Nature 324, 245–248.

Mareschal, J.C., West, G.F., 1980. A model for Archean tectonism. Part 2. Numeri-cal models of vertical tectonism in greenstone belts. Canadian Journal of EarthSciences 17, 60–71.

Maurice, C., David, J., Bédard, J.H., Francis, D., 2009. Evidence for a widespread maficcover sequence and its implications for continental growth in the NortheasternSuperior Province. Precambrian Research 169, 45–65.

McGregor, A.M., 1951. Some milestones in the Precambrian of southern Africa.Proceedings of the Geological Society of South Africa 54, 27–71.

McNaughton, N.J., Compston, W., Barley, M.E., 1993. Constraints on the age ofthe Warrawoona Group, eastern Pilbara Block, Western Australia. PrecambrianResearch 60, 69–98.

Mecklenburgh, J., Rutter, E.H., 2003. On the rheology of partially molten syntheticgranite. Journal of Structural Geology 25, 1575–1585.

Moresi, L., Dufour, F., Muhlhaus, H.B., 2002. Mantle convection modeling withviscoelastic/brittle lithosphere: numerical methodology and plate tectonic mod-eling. Pure and Applied Geophysics, 2335–2356.

Moresi, L., Muelhaus, H., Dufour, F., 2001. Particle in cell solution for creeping viscousflows with internal interfaces. In: Muelhaus, H., Dyskin, A., Pasternak, E. (Eds.),Proceedings of the 5th International Workshop on Bifurcation and Localisation.Balkema, Perth.

Nelson, D.R., 1999. Compilation of geochronology data, 1998, Geological Survey ofWestern Australia Record 1999/2, p. 222.

Nelson, D.R., 2000. Compilation of geochronology data, 2000, Geological Survey ofWestern Australia Record 2001/2, p. 205.

Nelson, D.R., 2001. Compilation of geochronology data, 2000, Geological Survey ofWestern Australia Record 2002/2, p. 205.

Ord, A., Hobbs, B.E., 1989. The strength of the continental crust, detachment zonesand the development of plastic instabilities. Tectonophysics 158, 269–289.

Putlitz, B., Katzir, Y., Matthews, A., Valley, J.W., 2001. Oceanic and orogenic fluid–rockinteraction in 18O/16O-enriched metagabbros of an ophiolite (Tinos, Cyclades).Earth and Planetary Science Letters 193, 99–113.

Rey, P., Houseman, G., 2006. Lithospheric scale gravitational flow: the impact ofbody forces on orogenic processes from Archaean to Phanerozoic. In: Buiter,S.J.H., Schreurs, G. (Eds.), Analogue and Numerical Modelling of Crustal-ScaleProcesses. , pp. 153–167.

Rey, P., Philippot, P., Thébaud, N., 2003. Contribution of mantle plumes, crustal thick-ening and greenstone blanketing to the 2.75–2.65 Ga global crisis. PrecambrianResearch 127, 43–60.

Rey, P.F., Coltice, N., 2008. Neoarchean lithospheric strengthening and the couplingof Earth’s geochemical reservoirs. Geology 36, 635–638.

Robin, C.M.I., Bailey, R.C., 2009. Simultaneous generation of Archean crust and sub-cratonic roots by vertical tectonics. Geology 37, 523–526.

Sandiford, M., Kranendonk, M.V., Bodorkos, S., 2004. Conductive incubation andthe origin of Granite-Greenstone dome-and-keel structure: the eastern PilbaraCraton, Australia. Tectonics 23.

Smithies, R.H., Champion, D.C., Cassidy, K.F., 2003. Formation of Earth’s earlyArchaean continental crust. Precambrian Research 127, 89–101.

Taylor, S., McLennan, S., 1995. The geochemical evolution of the continental crust.Reviews of Geophysics 33, 241–265.

Teyssier, C., Collins, W.J., Van Kranendonk, M.J., 1990. Strain and kinematics dur-ing the emplacement of the Mount Edgar Batholith and Warrawoona Syncline,Pilbara Block, Western Australia. In: Geoconferences (W.A.), Perth, WesternAustralia, pp. 481–483.

Thébaud, N., 2006. From Synchroton Characterisation of Single Fluid Inclusions toArchaean Geodynamics: A Study on Fluid–Rock Interaction in a Triple FoliationJunction (The warrawoona Syncline, Western Australia). Paris VI University andThe University of Sydney, 276 pp.

Thébaud, N., Philippot, P., Rey, P., Brugger, J., Van Kranendonk, M.J., Grassineau, N.,2008. Protracted fluid–rock interaction in the mid-Archaean and implication forgold mineralization: example from the Warrawoona Syncline (WA). Earth andPlanetary Science Letters 272, 639–655.

Thébaud, N., Philippot, P., Rey, P., Cauzid, J., 2006. Composition and origin of fluidsassociated with lode gold deposits in a mid-Archaean greenstone belt (Warra-woona Syncline, Pilbara, WA) using Synchrotron radiation X-Ray Fluorescence.Contributions to Mineralogy and Petrology 152, 485–503.

Thorpe, R.I., Hickman, A.H., Davis, D.W., Mortensen, J.K., Trendall, A.F., 1992. U–Pbzircon geochronology of Archaean felsic units in the Marble Bar region, PilbaraCraton, Western Australia. Precambrian Research 56, 169–189.

Turcotte, D.L., Schubert, G., 1982. Geodynamics: Applications of Continuum Physicsto Geological Problems. Wiley, New York.

Van Kranendonk, M.J., 2006. Volcanic degassing, hydrothermal circulation and theflourishing of early life on Earth: a review of the evidence from c. 3490–3240 Marocks of the Pilbara Supergroup, Pilbara Craton, Western Australia. Earth-ScienceReviews 74 (February), 197–240.

Van Kranendonk, M.J., Collins, W., Hickman, A., Pawley, M.J., 2004a. Critical tests ofvertical vs. horizontal tectonic models for the Archaean East Pilbara Granite-Greenstone Terrane, Pilbara Craton, Western Australia. Precambrian Research131, 173–211.

Van Kranendonk, M.J., Hickman, A.H., Smithies, R.H., Nelson, D.R., Pike, G., 2002.Geology and tectonic evolution of the Archaean North Pilbara Terrain, PilbaraCraton, Western Australia. Economic Geology and the Bulletin of the Society ofEconomic Geologists 97, 695–732.

Van Kranendonk, M.J., Smithies, R., Hickman, A., Champion, C., 2007. Review: seculartectonic evolution of Archean continental crust: interplay between horizontaland vertical processes in the formation of the Pilbara Craton, Australia. TerraNova 19, 1–38.

Van Kranendonk, M.J., Smithies, R.H., Hickman, A.H., Bagas, L., Williams, I.R.,Farrell, T.R., 2004. Event stratigraphy applied to 700 million years ofArchaean crustal evolution, previous term Pilbara next term Craton, West-ern Australia, Technical Papers Geological Survey of Western Australia, Perth,pp. 49–61.

Walshe, J.L., Cleverley, J.S., 2009. Gold deposits; where, when and why. Elements 5,288.

West, G.F., Mareschal, J.C., 1979. A model for Archean tectonism. Part 1. The thermalconditions. Canadian Journal of Earth Sciences 16, 1942–1950.

Wijbrans, J.R., McDougall, I., 1987. On the metamorphic history of anArchaean granite-greenstone terrain, East Pilbara, Western Australia usingthe 40Ar/39Ar spectrum technique. Earth and Planetary Science Letters 84,226–242.

Zheng, Y.F., 1993. Calculation of oxygen isotope fractionation in anhydrous silicateminerals. Geochimica et Cosmochimica Acta 57, 1079–1091.