Geological Society of America Bulletin 2013 Cawood 14 32

8/17/2019 Geological Society of America Bulletin 2013 Cawood 14 32

1/19

14

The continental record and thegeneration of continental crust

P.A. Cawood1,2,†

, C.J. Hawkesworth1

, and B. Dhuime1,3

1 Department of Earth Sciences, University of St. Andrews, North Street, St. Andrews KY16 9AL, UK 2School of Earth and Environment, The University of Western Australia, 35 Stirling Highway, Crawley, WA 6009, Australia3 Department of Earth Sciences, University of Bristol, Wills Memorial Building, Queens Road, Bristol BS8 1RJ, UK

ABSTRACT

Continental crust is the archive of Earthhistory. The spatial and temporal distributionof Earth’s record of rock units and events isheterogeneous; for example, ages of igneouscrystallization, metamorphism, continentalmargins, mineralization, and sea water andatmospheric proxies are distributed about aseries of peaks and troughs. This distributionreects the different preservation potential ofrocks generated in different tectonic settings,rather than fundamental pulses of activity,and the peaks of ages are linked to the tim-ing of supercontinent assembly. The physio-chemical resilience of zircons and theirderivation largely from felsic igneous rocksmeans that they are important indicators ofthe crustal record. Furthermore, detrital zir-cons, which sample a range of source rocks,provide a more representative record than

direct analysis of grains in igneous rocks.Analysis of detrital zircons suggests that atleast ~60%–70% of the present volume ofthe continental crust had been generated by3 Ga. Such estimates seek to take account ofthe extent to which the old crustal material isunderrepresented in the sedimentary record ,and they imply that there were greater vol-umes of continental crust in the Archeanthan might be inferred from the composi-tions of detrital zircons and sediments. Thegrowth of continental crust was a continu-ous rather than an episodic process, butthere was a marked decrease in the rate ofcrustal growth at ca. 3 Ga, which may havebeen linked to the onset of signicant crustalrecycling, probably through subduction atconvergent plate margins. The Hadean andEarly Archean continental record is poorlypreserved and characterized by a bimodalTTG (tonalites, trondhjemites, and grano-diorites) and greenstone association that

differs from the younger record that can bemore directly related to a plate-tectonic re-gime. The paucity of this early record has ledto competing and equivocal models invokingplate-tectonic– and mantle-plume–domi-nated processes. The 60%–70% of the pres-ent volume of the continental crust estimatedto have been present at 3 Ga contrasts mark-edly with the


2/19

The continental record and the generation of continental crust

Geological Society of America Bulletin, January/February 2013 15

advanced through developments in stratigraphicanalysis, petrography, paleontology, geochemis-try, geochronology, geophysics, and modeling.Crucially, our understanding of the processes in-volved in the generation and the evolution of thecontinental crust has grown enormously throughthe latter part of the twentieth and beginning of

the twenty-rst centuries following on from thedevelopment and acceptance of plate-tectonictheory. This has focused our research on platemargins, the sites of continental crust forma-tion and stabilization, and it has resulted in afundamental change in the way we approachour interrogation of Earth and its record froma descriptive documentation of units and eventsinto investigation into the processes controllingthese features. A factor critical to determiningthese processes is an understanding of rates ofchange, and this has been facilitated by devel-opments in data collection and analysis. Thisexpansion of knowledge has been particularlyimportant in further understanding not just theexposed surcial rock record, but in gaining in-sight into the composition and development ofthe whole crust. In particular, this has led to newideas into what shaped the record, and how rep-resentative, or unrepresentative, it may be.

SOME FACTS AND TERMINOLOGY

The total area of continental crust is 210.4 ×106 km2, or some 41% of the surface area ofEarth, and the volume is 7.2 × 10 9 km 3, whichconstitutes some 70% of Earth’s crustal vol-

ume (Cogley, 1984). The crust extends ver-tically from the surface to the Mohorovi čić discontinuity (Moho) and laterally to the breakin slope in the continental shelf (Rudnick andGao, 2003). The Moho is dened as the jump inseismic primary waves (P-waves) to greater than~7.6 km s –1. This change in seismic-wave veloc-ity is taken as a rst-order approximation of theboundary between mac lower-crustal rock andultramac mantle peridotite: the crust-mantleboundary. Petrologic studies from exposed sec-tions of ocean oor (ophiolites) and from xeno-liths in continental settings suggest that in somesituations, the crust-mantle boundary may dif-fer from the Moho (Malpas, 1978; Grifn andO’Reilly, 1987).

The mean elevation of the continental crustis around 125 m (Fig. 1), and some 31% ofthe crustal area is below sea level. Continentalcrustal thickness varies from 20 to 70 km, aver-aging around 35–40 km (Fig. 2; Mooney et al.,1998). The crust and underlying mantle con-stitute the lithosphere; the mechanically strongouter layer of Earth that forms the surface plates(Barrell, 1914a, 1914b, 1914c; Daly, 1940;White, 1988). Heat transport in the lithosphere

is conductive, and the base is a rheologicalboundary with the isothermal convecting mantle(Sleep, 2005).

Continents include cratons, areas of stablecrust, orogenic belts, and regions of continentalextension, which form either intracratonic riftsor develop into zones of continental breakupand thermal subsidence (passive margins).Orogens evolve through one or more cycles ofsedimentation, subsidence, and igneous activ-ity punctuated by tectonothermal events (orog-enies), involving deformation, metamorphism,and igneous activity, which result in thickeningand stabilization of the lithosphere (Figs. 3Aand 3B). Cratons are ancient orogens that havegenerally been undeformed and tectonicallystable for long periods of time, often since theArchean, and are divisible into shields, whichare regions of exposed crystalline igneous andmetamorphic rock, and platforms, where theshield is overlain by a relatively undeformedsedimentary succession (Fig. 4).

Geologic and geophysical data show thatthe crust is divisible into a felsic upper crustcomposed largely of sedimentary rock (upperfew kilometers) and granite to granodiorite,

a hetero geneous middle crust assemblage ofortho gneiss and paragneiss at amphibolitefacies to lower granulite facies (Fig. 3C), and alower crust consisting of granulite-facies coun-try rocks and basic intrusive rocks and/or cumu-lates (Rudnick and Fountain, 1995; Wedepohl,1995; Rudnick and Gao, 2003). Thicknesses ofthe three crustal layers vary, but the upper andmiddle crustal sections generally form around30% each of a typical crustal prole, with thelower crust forming the remaining 40% (Fig. 5;Rudnick and Gao, 2003; Hawkesworth andKemp, 2006a).

The bulk composition of the crust is equiva-lent to andesite (Fig. 3D) and requires twostages of formation involving extraction ofmac magmas from the mantle and their dif-ferentiation through either fractional crystalliza-tion or remelting and return of the cumulate orresidue to the mantle (Taylor, 1967; Taylor andMcLennan, 1985; Kay and Kay, 1991; Rudnick,1995; Rudnick and Gao, 2003; Davidson andArculus, 2006; Hacker et al., 2011).

Surface heat ux of Archean cratons is gener-ally low (30–40 mW m –2), Phanerozoic regionsshow higher heat ux values (>60–80 mW m –2)

Figure 2. Contour map of the thickness of Earth’s crust (developed from model CRUST 5.1).The contour interval is 10 km (45 km contour interval is also shown to provide greater detailon the continents). To a rst approximation, the continents and their margins are outlinedby the 30 km contour. Source of gure: http://earthquake.usgs.gov/research/structure/crust

/index.php.


3/19

Cawood et al.

16 Geological Society of America Bulletin, January/February 2013

Figure 3. (A) Cerro Aconcagua (6962 m) inwestern Argentina is the highest mountainoutside Asia and the tallest in the Ameri-cas. It lies within the accretionary Andeanorogen and formed through noncollisionalcoupling between the downgoing Nazca oce-anic plate and the continental lithosphere

of the South American plate. (B) Glencoulthrust on the northern side of Loch Glen-coul, NW Highlands, Scotland. Silurian-agethrust faulting of Mesoarchean and Neo-archean (Lewisian) gneiss over a footwallof Cambrian strata (Eriboll and An-t-SronFormations) that are resting unconform-ably on underlying Lewisian basement.(C) Road cut north of Loch Laxford, NWHighlands, Scotland, showing Mesoarcheanand Neoarchean gray gneiss (Lewisian),cut by Paleoproterozoic mac intrusions(Scourie dikes, now foliated and metamor-phosed within the amphibolite facies), anddiscordant sheets of ca. 1.85 Ga (Laxford-ian) granite and pegmatite. (D) Construc-tion of continental crust through andesiticmagmatism demonstrated by the Nevadosde Payachatas volcanoes: Parinacota (right)and Pomerape (left) in northern Chile. Bothwere constructed within the last 300 k.y.,with Pomerape the older (more eroded) andParinacota the younger, having undergonepostglacial sector collapse to produce a largedebris avalanche (foreground hummocks).(E) The Stones of Calanais insert (Callanish)are a Neolithic monument on the island of

Lewis, Scotland, composed of Lewisian Gneiss. The stones are arranged in a central circle augmented by linear avenues leading off to thepoints of the compass. To the north, the avenue comprises a double row of stones. At the heart of the monument, there is a Neolithic burialmound that was raised some time after the erection of the stones. The precise date of construction of the circle is unclear, but the stones wereerected ~5000 yr ago. The purpose of sites like this (including Stonehenge) remains a mystery, but research suggests that they constitutedimportant central places for the local farming community and may have been aligned to prominent features of both land and sky. Thestones were quarried locally, and their transport and construction would have formed a focal activity at the time. (F) Siccar Point, Berwick-shire, NE England. Hutton’s iconic angular unconformity between gently dipping Devonian Upper Old Red Sandstone on near-verticalSilurian sandstones and shales. Source of images: A–C, Peter Cawood; D, Jon Davidson, University of Durham; E, Caroline Wickham-Jones; F, British Geological Survey digital data bank http://geoscenic.bgs.ac.uk/asset-bank/action/viewHome.

A B

C D

E F

Cambrianstrata

Lewisiangneiss

Lewisiangneiss

MORpassivemargin

oceanic lithosphere

oceanic lithosphere

convergentmargin icr

continental lithosphere

Proterozoic Orogen

shieldplatform

ArcheanOrogen

PhanerozoicOrogen

craton

sub-continental lithospheric mantle

continental crust

asthenosphere

sub-oceanic lithospheric mantle

oceanic crust

Figure 4. Schematic cross sec-tion of types of continental litho-sphere emphasizing the thickstable nature of Precambriancratons. Thickness of litho-sphere beneath Archean regionsis of the order of 200–250 kmand oceanic lithosphere is upto 100 km. Abbreviation: icr—intra cratonic rift; MOR—mid-ocean ridge.


4/19



and Proterozoic regions display intermedi-ate values. This variation appears to correlatewith lithospheric thickness, which ranges fromsome 200–250 km beneath cratons to gener-ally


5/19

Cawood et al.


only empha sizes the enormity of deep Earthtime, but it also provides a counterpoint to thestrict biblical interpretation of the age of Earth.

Uniformitarianism provided a framework inwhich to study Earth, and researchers focusedon understanding the processes that shaped andstabilized continental crust preserved in ancient

mountain belts (e.g., Hutton, 1788; Lyell, 1833;Hall, 1859; Dana, 1873; Suess, 1885–1909;Haug, 1900; see discussion and references inDott, 1974; Şengör, 1982). These early obser-vations on the crust and its origins were basedon direct eld observations and were focusedon Phanerozoic sequences in eastern NorthAmerica and Western Europe, notably theAppa lachian and Alpine orogens, with theircontrasting geology inuencing the ideas andthe theories that were proposed. Early workerson both continents acknowledged that the conti-nental crust in these belts included a very thickaccumulation of deformed sediment. NorthAmerican workers considered these sequencesto be shallow-water deposits that accumulatedin asymmetric troughs at the margins betweencontinents and ocean basins, but which also in-cluded input from an outboard source consistingof a long-established basement high or border-land (Hall, 1859; Dana, 1873; Schuchert, 1910).In contrast, European workers in the Alpineorogen regarded the sediments as being deep-marine deposits that accumulated on an ophi-olitic substrate in a symmetrical basin betweencontinents (Suess, 1885–1909; Haug, 1900;Steinmann, 1906). Time-integrated analysis of

these orogenic belts led researchers to specu-late on the stabilization of continental crustthrough a tectonic cycle involving geosyncli-nal, orogenic, and cratonic stages (Haug, 1900;Krynine, 1948; Aubouin, 1965). On a broader

scale, ideas on a tectonic cycle led to the conceptof continental accretion through a succession ofconcentric orogenic belts (Dana, 1873; Suess,1885–1909; Haug, 1900). This relationship wasmost readily observed in North America, withits cratonic core and enveloping younger Appa-lachian and Cordilleran orogens. These ideas

on the stabilization of continental crust throughmountain-building processes developed in aframework involving the permanency of oceanbasins and xed continents—concepts that inthe twentieth century were being increasinglyquestioned, initially by Western European andSouthern Hemisphere workers (Wegener, 1924;Holmes, 1928–1930; du Toit, 1937; Umbgrove,1947; Carey, 1958), who emphasized the tran-sitory and dynamic nature of Earth’s surface.This view of a dynamic Earth was ultimatelyembraced by the broader geological commu-nity and integrated into plate-tectonic theory(Hess, 1962; Vine and Matthews, 1963; Wilson,1966; Dewey and Bird, 1970). It highlighted theprocesses through which the continental crustwas both generated and destroyed.

Chemical and isotopic data on the composi-tion and age of the continental crust, along withgeophysical data on the internal structure of thecrust and lithosphere, were then integrated withevolving ideas on tectonic processes to providefurther insight into the origin and rate of growthof the crust. Early geochemical data enabled es-timates of the average composition of specicrock types/tectonic units and ultimately led toestimations of the average composition of the

entire crust (Clarke, 1924; Goldschmidt, 1954;Poldervaart, 1955; Taylor, 1964; Ronov andYaroshevsky, 1969). This data set has been in-creasingly rened, as well as integrated with,and fed back into, tectonic models of the crust,

and our understanding of the inferred interrela-tionship between the crust and the complemen-tary mantle reservoir from which it is derived(Hart, 1969; Taylor and McLennan, 1985;Rudnick, 1995; Rudnick and Fountain, 1995;McLennan and Taylor, 1996; Rudnick and Gao,2003). These studies helped to establish (1) that

the overall composition of the continental crustis similar to calc-alkaline andesite, and (2) theconcept that the crust is typically derived in twostages, melting of the mantle to generate macmagma, which undergoes fractional crystalliza-tion, with or without assimilation of preexistingcrust, or crystallization, and then remelting togenerate average crustal compositions.

Age and radiogenic isotopic data on rocksand minerals have led to a range of modelson the rate of growth of the continental crust(Fig. 6). Most models indicate that the conti-nental crust has increased in volume and areathrough time (Hurley et al., 1962; Hurley andRand, 1969; Fyfe, 1978; Veizer and Jansen,1979; Armstrong, 1981; Allègre and Rous-seau, 1984; Taylor and McLennan, 1985; Veizerand Jansen, 1985; Armstrong, 1991; Taylorand McLennan, 1996; Belousova et al., 2010;Dhuime et al., 2012). These studies have tendedto argue that crustal growth has been continu-ous, with early models proposing steady or in-creasing rates of growth through Earth history,and subsequent studies emphasizing an earlierperiod of more rapid crustal growth, typically inthe late Archean or early Proterozoic, followedby decreasing rates of growth to the present day.

The early models were often based on the geo-graphic distribution of Rb-Sr and K-Ar isotopeages (Hurley et al., 1962; Hurley, 1968; Hurleyand Rand, 1969; Goodwin, 1996), but these arebiased by orogenic overprinting from events

Age (Ga)0

50

T a y l o r & M c L e n n a n

1 9 8 5

H u r l e

y & R a

n d 1 9

6 9

01.02.03.04.0

Armstrong 1981

A l l è g

r e & R

o u s s e

a u 1 9

8 4

D h u i m

e e t a

l. 2 0 1

2

100

G o o d w i n

1 9 9 6

Archean Proterozoic Phan.

oldestrock

oldestzircon

C o n d

i e & A

s t e r 2

0 1 0

present day surfaceages distribution

Hadean

B e l o u

s o v a

e t a l

. 2 0 1

0

Figure 6. Crustal growth mod-els of Hurley and Rand (1969),Armstrong (1981), Allègre andRousseau (1984), Taylor and

McLennan (1985), Condie andAster (2010), and Dhuime et al.(2012) compared to age distribu-tion of presently preserved crustfrom Goodwin (1996). Sourcesof images: early Earth—http:// www.universetoday.com/58177/earth-formation/; present-dayEarth—National Aeronauticsand Space Administration.


6/19



younger than the age of crust formation, andolder crustal terrains tend to be less well pre-served than those formed more recently. A keyadvance was in the development of isotope sys-tems, in particular, Nd and Hf isotopes, whichcould be used to constrain model ages to indi-cate when new crust was generated, and which

are, for the most part, unaffected by youngerorogenic events (McCulloch and Wasserburg,1978; Patchett et al., 1982). These isotope sys-tems have relatively long half-lives, and so thevariations in Nd and Hf isotopes reect the gen-eration of crust that is relatively long-lived. Ma-terial had to be in the crust for long enough inorder for signicant amounts of the radiogenicisotope to be generated, typically a few hundredmillion years. New crust that was generated andthen destroyed shortly after would not have hadtime to develop a radiogenic signature, and soit is in effect invisible to these isotope systems.Armstrong (1981, 1991) and Fyfe (1978) pro-posed an early burst of continental growth thatwas followed by steady-state or even decreas-ing crustal volumes. Armstrong (e.g., 1981) ar-gued that constant volume was maintained bycrustal recycling, offsetting any additions fromcrustal generation. He was amongst the rst toemphasize the role of recycling (contrast withMoorbath, 1975, 1977) in both understandingcrustal volumes and in potentially affecting thecomposition of the upper mantle. However, be-cause there is little evidence for signicant vol-umes of Hadean and Early Archean crust fromradiogenic isotope systems, it implies that much

of the proposed recycled crust is too young tohave developed distinctive radiogenic isotopesignatures (cf. Hurley et al., 1962; Hurley andRand, 1969; O’Nions et al., 1979; Veizerand Jansen, 1979).

A second aspect is that some models ofcrustal growth invoke continuous growth, andothers indicate that the growth of the crust wasin some way episodic. The isotopic models ofsteady growth of continental crust (plus or mi-nus recycling) would appear to be at odds withmap patterns involving discordant data sets ofdiscrete age provinces that are abruptly trun-cated at their boundaries (Gastil, 1960a, 1960b;Dott, 1964). More recent models of continentalgrowth tend to show pulses of enhanced growthat specic periods in Earth history (e.g., LateArchean; McCulloch and Bennett, 1994; Taylorand McLennan, 1996; Condie, 1998; Rino et al.,2004). Most of the continental growth modelshave been based on compilations of whole-rockisotopic data, in some cases of ne-grained sedi-ments that may have sampled large source areas.However, Condie (1998) used a limited suite ofzircon ages of juvenile continental crust, andRino et al. (2004) developed a model based on

analysis of detrital zircons of igneous origin inthe modern Amazon, Mackenzie, and Missis-sippi Rivers. Rapidly expanding databases ofigneous and detrital zircon data highlight a non-continuous distribution of crystallization ages(see Fig. 7) and have led to further proposalsthat continental growth has been episodic, cor-

responding with phases of the supercontinentcycle and/or to mantle-plume activity (Condie,1998, 2000, 2004; Campbell and Allen, 2008;Voice et al., 2011). However, Gurnis and Davies(1986) noted that young crust is more elevatedthan old crust and hence more easily eroded (cf.Allègre and Rousseau, 1984), and this will leadto preferential recycling of young crust and canlead to an apparent peak in a continuous crustalgrowth curve at 2–3 Ga.

THE NATURE OF THECONTINENTAL RECORD

The rock record is incomplete, which raisesquestions about the extent to which the conti-nental archive is biased by selective preservationof rocks generated in different tectonic settings,and therefore how best to evaluate temporalchanges from the record. Igneous provinces areby denition restricted in space and time, andthey therefore provide regional snapshots of themagmatic processes that have occurred in dif-ferent tectonic settings. The same is broadly truefor metamorphic provinces, and the constraintson the thermal histories of different areas pro-vided by pressure-temperature-time ( P -T -t )

paths. In contrast, sediments contain materialfrom their source rocks irrespective of the con-ditions under which those rocks were generated.The bulk compositions of detrital sedimentshave therefore been used to estimate the aver-age composition of the upper crust (Taylor andMcLennan, 1981, 1985; Condie, 1993; Taylorand McLennan, 1995; Rudnick and Gao, 2003),and to determine its average Nd isotope ratioand model Nd age (Taylor et al., 1983; Allègreand Rousseau, 1984; Michard et al., 1985).However, the sedimentary record is biased (1) interms of the different lithologies, and, hence forPhanerozoic strata, the fossil communities thatare preserved (e.g., Raup, 1972, 1976; Smithand McGowan, 2007), and (2) because youngersource rocks are thought to be more accessibleto erosion than older source rocks. Thus, oldersource rocks may be underrepresented in thebulk compositions of continental detrital sedi-ments (Allègre and Rousseau, 1984). It is un-clear whether comparable biases also occur inthe igneous and metamorphic rocks of the con-tinental crust.

Data compilations emphasize that the spa-tial and temporal distribution of rock units and

events is heterogeneous; for example, ages ofigneous crystallization, metamorphism, conti-nental margins, mineralization, and seawaterand atmospheric proxies are distributed abouta series of peaks and troughs. These appear, atleast in part, to correspond with the cycle ofsuper continent assembly and dispersal (Fig. 7).

Igneous and detrital zircon U-Pb ages andHf isotopic data have identiable peaks in agesof crystallization (Fig. 7; Condie, 1998, 2000,2004, 2005; Rino et al., 2004; Groves et al.,2005; Hawkesworth and Kemp, 2006b; Kempet al., 2006; Campbell and Allen, 2008; Belou-sova et al., 2010; Voice et al., 2011). Peaks inU-Pb crystallization ages correspond with pe-riods of supercontinent assembly at around2.7–2.4 Ga (Superia and Sclavia), 2.1–1.7 Ga(Nuna), 1.3–0.95 Ga (Rodinia), 0.7–0.5 Ga(Gondwana), and 0.35–0.18 Ga (Pangea). Peaksin Hf isotope ages were recognized in earlyand/or regional studies, but expanding data setssuggest a more continuous distribution (Belou-sova et al., 2010; Hawkesworth et al., 2010;Dhuime et al., 2012). The distribution of theages of high-grade metamorphic rocks is alsoepisodic (Fig. 7). Brown (2007) categorizedhigh-grade orogenic belts into high-, intermedi-ate- and low-pressure–high-temperature belts.He noted that the high-pressure belts were re-stricted to the last 600 m.y., and he concludedthat they reect cold subduction as observed atpresent along convergent margins. Intermedi-ate- to low-pressure–high-temperature rocksare preserved dating back to the Late Archean,

and Kemp et al. (2007b) pointed out that theirages are grouped in clusters similar to the peaksof crust generation illustrated in Figure 7. Theimplication is that periods of granulite-faciesmetamorphism are in some way linked to theprocesses of crust generation, as suggested byKemp et al. (2007b), and/or the peaks of theages of crust generation and granulite metamor-phism are themselves a function of the uneven-ness of the continental record.

The age pattern of ancient passive marginsalso reveals major peaks in the Late Archean,late Paleoproterozoic, and late Neoproterozoicto early Paleozoic, which correspond to timesof supercontinent aggregation (Fig. 7; Bradley,2008). The proportion of modern passive mar-gins is somewhat different, correlating with thebreakup of Pangea and the resultant increasein margin area (Bradley, 2008). Smith andMcGowan (2007) noted that the Phanerozoicdiversity of marine fossils is affected by thesupercontinent cycle, with marine rocks domi-nating during rifting phases of supercontinents.Eriksson and Simpson (1998) highlighted thetemporal concentration of eolianites at 1.8 Gaand its association with breakup and assembly


7/19

Cawood et al.


Age (Ga)

1500

1000

500

0

0.5 1.0 1.5 2.0 2.5 3.0 3.50 4.0 4.5

N u m

b e r

Pangaea

Gondwana

Rodinia Nuna

0

1000

2000

3000

4000Superia/Sclavia

Zircon record (n = 100,445)

change in scale

500

1500

Granulite faciesUHT metamorphism

Eclogite-HP granulitemetamorphismHP-UHPmetamorphism

0%

20%

40%

60%

80%

100%

–15

–10

–5

0

5

10

15

N e w c r u s t 8 7

S r / 8 6

M e a n

ε H f

( t ) i n d e

t r i t a l z i r c o n

Seawater

Zircons

Passive margins

0

5

10

15

20

406080

change in scale

N u m

b e r

Modern passive margins

Figure 7. (A) Histogram of over 100,000 detrital zircon analyses showing several peaks in their U-Pb crystallization ages over the courseof Earth history (Voice et al., 2011), which are very similar to the ages of supercontinents. Also shown is the apparent thermal gradientversus age of peak metamorphism for the three main types of granulite-facies metamorphic belts (Brown, 2007). UHT—ultrahigh tempera-ture; HP—high pressure; UHP—ultrahigh pressure. (B) Histogram of the ages of ancient and modern passive margins (Bradley, 2008).(C) Normalized seawater 87Sr/ 86Sr curve (Shields, 2007) and running mean of initial Hf in ~7000 detrital zircons from recent sediments(Dhuime, Hawkesworth, and Cawood, personal observation). Low 87Sr/ 86Sr values in Archean in part reect the lack of data and the largeproportion of submerged continental crust.


8/19



phases of supercontinents. Bradley (2011) re-cently compiled secular trends in a variety ofrock units and events and noted that carbonatites(Woolley and Kjarsgaard, 2008) and green-stone-belt deformation events (Condie, 1994,1995) also bear the imprint of Precambrian su-percontinent cycles.

Mineral deposits are heterogeneously dis-tributed in both space and time, with variationsrelated to long-term tectonic trends associatedwith the supercontinent cycle and changingenvironmental conditions such as atmosphere-hydrosphere conditions and thermal history(Meyer, 1988; Barley and Groves, 1992; Groveset al., 2005; Groves and Bierlein, 2007; Bier-lein et al., 2009). For example, deposit typesassociated with convergent plate margins (ac-cretionary orogens), such as orogenic gold andvolcanic massive sulde (VMS) deposits, and toa lesser extent porphyry Cu-Au-Mo and Sn-W,and epithermal Cu-Au-Ag deposits, exhibitwell-dened temporal patterns that broadlycorrelate with supercontinent assembly (Bier-lein et al., 2009). However, deposits formedin intracratonic settings and related to mantleprocesses (e.g., platinum group elements (PGE)deposits) lack such a correlation (Cawood andHawkesworth, 2012).

GENERATIONAL ARCHIVE ORPRESERVATIONAL BIAS

The evidence for peaks and troughs across therock record, particularly related to igneous rock

generation (Fig. 7), has been used to argue thatcontinental crust formation has been episodic,and that in some way pulses in the formationof continental crust are linked to the develop-ment of supercontinents (e.g., Condie, 1998,2000, 2004, 2005; Rino et al., 2004; Groveset al., 2005; Hawkesworth and Kemp, 2006b;Kemp et al., 2006; Campbell and Allen, 2008;Voice et al., 2011). Punctuated crustal growthremains difcult to explain by global changesin plate-tectonic regimes, which is viewed asa continuous process (but see discussions byO’Neill et al., 2007; Korenaga, 2008; Silverand Behn, 2008a, 2008b), and so it is typicallylinked to mantle-plume activity (Stein and Hof-mann, 1993; Condie, 1998). However, the ande-sitic composition of continental crust, alongwith evidence that the plate-tectonic mecha-nism has been active for much of Earth historyand is a major driver for continental assemblyand dispersal (Cawood et al., 2006; Shirey andRichardson, 2011), suggests that magmatic arcsare the major source of continental growth (cf.Taylor, 1967; Taylor and McLennan, 1985; Mc-Culloch and Bennett, 1994; Rudnick, 1995;Davidson and Arculus, 2006; Hawkesworth

and Kemp, 2006b). Recently, Stern and Scholl(2009) argued that peaks of ages reect periodsof increased magmatic activity associated withincreases in the volumes of subduction-relatedmagmas that are generated during continen-tal breakup. The peaks of igneous crystalliza-tion ages correspond however, with the time of

maximum supercontinent aggregation, not withtheir breakup. Also the preservation potential ofintra-oceanic arcs, which are largely submarine,is poor (Condie and Kröner, 2012). From a com-parison of U-Pb ages and Hf isotope model agesfor ~5100 detrital zircons, Voice et al. (2011)proposed that the age-frequency distribution ofdetrital grains reects predominantly episodiccrustal recycling (i.e., destruction) rather thancrustal growth. Recycling of crust is indeed animportant issue in evaluating growth models ofcontinental crust (cf. Armstrong, 1981), but itsrole can only be evaluated after the validity ofthe record, and the extent to which it has beeninuenced by preservation processes, has beenestablished.

Analysis of the rock record at modern con-vergent plate margins has established that theyare not only major sites for the generation ofcontinental crust but also for its removal andrecycling back into the mantle (Fig. 8). Globalcompilations of sites of continental additionand removal (Scholl and von Huene, 2007,2009; Clift et al., 2009; Stern, 2011) highlightthat crustal growth rates in continental colli-sion zones, sites of continental aggregation andsuper continent assembly, are low and insuf-

cient to generate the present volume of conti-nental crust over the history of Earth. Although

these compilations reach different conclusionson the values for addition and removal of con-tinental crust for individual tectonic settings,reecting the different data sets and proxiesused, they reach a similar overall conclusion:that, on a global scale, the processes of conti-nental addition along destructive plate margins

are counter balanced by those of continental re-moval, resulting in no net growth in the currentvolume of continental crust.

Integration of data on the rates and sites ofcontinental generation and recycling (Fig. 8)with the observed punctuated rock record(Fig. 7) suggests that peaks in age data maynot represent episodic growth but instead re-ect the greater preservation potential of rocksformed during the latter stages of ocean closureand collision, and that the record is thereforebiased by the construction of supercontinents(Hawkesworth et al., 2009, 2010; Condie et al.,2011). Thus, the observed rock record of igne-ous crystallization ages is the integration of thevolumes of magma generated during the threephases of the supercontinent cycle (subduction,collision, and breakup), and their likely pres-ervation potential within each of these phases.This is illustrated schematically in Figure 9:Magma volumes are high in subduction set-tings but low during continental collision andbreakup (Fig. 8). In contrast, the preservationpotential of rocks in convergent and breakupsettings is poor, whereas the preservation poten-tial of late-stage subduction prior to collision,and collisional settings is high (Fig. 8). Peaks in

crystallization ages that are preserved (shadedarea under the curves in Fig. 9) would then

collisional orogen

extensional continental

marginsediment shed

from orogenic weltfuture

accretionary orogen

retreating plate margin

mafic underplate

MOR

crustalmelts

0.1 2.5

2.5

0.2

delaminatedlithospheric root

0.7

0.2

hot-spot

0.1

Collisional ExtensionalConvergent

Figure 8. Schematic cross section of convergent, collisional, and extensional plate bound-aries associated with supercontinent cycle showing estimated amounts (in km 3 yr –1) of conti-nental addition (numbers in blue above Earth surface) and removal (numbers in red belowsurface). Data are from Scholl and von Huene (2007, 2009). The volume of continental crustadded through time via juvenile magma addition is approximately compensated by the re-turn of continental and island-arc crust to the mantle, implying that there is no net growthof continental crust at the present day. Data compilations from Clift et al. (2009) and Stern(2011) calculated values of crustal recycling that are greater than the volume of juvenilecrustal addition, requiring a decrease in present-day continental volumes. MOR—mid-ocean ridge.


9/19

Cawood et al.


reect the balance between the magma volumesgenerated in the three stages of supercontinent

evolution and their preservation potential. Assuch, they would be unrelated to any underlyingvariation in the rate of crust generation. Notethat the resultant peak in Figure 9 correspondsto the collisional phase of the supercontinentcycle, even though this is not a major phase ofcrustal generation (compare with data in Figs. 7and 8), rather than periods of subduction- andextension-related magmatism (Hawkesworthet al., 2009). In detail, magmas generated insubduction-zone settings crystallize less zir-con per volume of magma than those in colli-sion settings (e.g., Moecher and Samson, 2006;Dickinson, 2008). Nonetheless, given the dif-ferences in the volumes of magma generated,the number of zircons crystallized in the for-mer is many orders of magnitude higher thanthose generated during the collision stage (Fig.10). Thus, the greater abundance of zircons forwhich ages correspond with the time of super-continent assembly (Fig. 7) is only possible dueto their higher preservation potential and cannotbe simply related to the volumes generated (Fig.10). Furthermore, the ux of sediment from thesource region to depositional basin reectsthe inuence and feedback among relief, cli-

mate, and tectonic setting. High runoff in zonesof crustal thickening and uplift will result inrapid exhumation, erosion, and high sedimentux (Koons, 1995; Clift, 2010). Basins adjacentto zones of continental collision or convergentAndean margins that receive high orographic/ monsoonal rainfall feed high-ux river sys-

tems. The Yellow, Amazon, and BrahmaputraRivers are the top three sediment-producingrivers in the world, whereas rivers draining low-relief, arid environments have low sedimentux (Summereld and Hulton, 1994). Thus,the high sediment ux within collision zonesis likely to further accentuate the preservationbias–induced , episodic zircon record.

Preservation bias also explains other seculartrends related to the supercontinent cycle. Thepeaks in passive margin ages at around 2.5 Ga,2.0 Ga, and 0.5 Ga are consistent with selec-tive preservation. If passive-margin distributionwere related to the time at which they were de-veloped, they should follow a predictable pat-tern related to changes in area of continentalmargins through time, with a minimum numberof margins corresponding to the peak in super-continent aggregation when continental marginarea is reduced relative to the area of the indi-vidual constituent continents. In detail, their dis-tribution during a supercontinent cycle shouldbe characterized by: (1) a decrease in globalpopulation of passive margins during supercon-tinent assembly; (2) few passive margins whenthe supercontinent is fully assembled; and (3) an

increase in number of passive margins dur-ing supercontinent breakup as surface area ofcontinental margins increases (Bradley, 2008).This is not what is typically observed, and onlythe most recent supercontinent, Pangea, and itssubsequent breakup record, represented by thedistribution of modern margins, appear to fol-

low this trend. The difference in passive-margindistribution associated with Pangea breakuprelative to those of earlier supercontinents canbe explained by the fact that the next supercon-tinent after Pangea has not yet formed (termedAmasia by Hoffman, 1992), and hence anypreservation bias in the record will not be appar-ent until then. Unlike the relationship betweenpeaks in passive-margin ages that correspond tothe Superia/Sclavia, Nuna, and Gondwana super-continents, there is no peak associated withRodinia. A possible explanation is that closureof the ocean related to Rodinia assembly did notinvolve passive margins draining older sourceregions, but rather was bounded by convergentplate margins (e.g., like the current circum-Pacic “Ring of Fire”).

The time interval corresponding with theRodinia supercontinent also lacks anomalies inthe 87Sr/ 86Sr ratio seawater record or in the aver-age ε Hf values from large zircon compilations(Fig. 7). The Sr isotope ratio is different from in-ferred proxies of continental growth, such as theU-Pb detrital zircon record, in that it is unlikelyto have been inuenced by preservation bias inthe geological record. It is taken as a measure of

super-continent cycle

subduction collision breakup

magmavolumes

preservation potential

Figure 9. The volumes of magma generated(blue line), and their likely preservation po-tential (red line) based on relations outlined

in Figure 8, vary through the three stagesassociated with the convergence, assembly,and breakup of a supercontinent. Peaks inigneous crystallization ages that are pre-served (shaded area) reect the balance be-tween the magma volumes generated in thethree stages and their preservation poten-tial, and will result in an episodic distribu-tion of ages in the rock record.

0

1

2

3

4

subduction collision breakup

V o

l u m e o

f m a g m a

( k m

N um

b er

of z i r c on s / M

a / k m

3 / a )

volume of magma

number of zircons

5.0.E+20

1.0.E+21

1.5.E+21

2.0.E+21

Figure 10. Volumes of magma generated (in km 3 yr –1) during the three stages of the super-continent cycle (from Scholl and von Huene, 2007, 2009; see also Fig. 8) compared to es-timates of number of zircons likely to crystallize for each setting (per m.y., per km). Thevolume of zircon generated per million years was calculated from the average Zr contentin magmas, using the relationship: vol% zircon = 1.15 wt% Zr (Dickinson, 2008); and Zr =150 ppm, 520 ppm, and 375 ppm for subduction, collision, and anorogenic magmas, respec-tively (Dickin son, 2008). Average zircon dimensions of 150 × 60 × 60 µm were used to convertzircon volumes into number of zircons. The number of zircons that crystallized per millionyears is normalized to the total length of convergent margins (42,000 km; Scholl and vonHuene, 2009; Stern, 2011), in order to obtain zircon generation rates in million years per km.


10/19



the relative amount of continental versus mantleinput, and positive excursions, such as thosecorresponding with the Gondwana and Nunasupercontinents, are often considered indica-tive of uplift and erosion of continental base-ment during continental collision (e.g., Bartleyet al., 2001; Prokoph et al., 2008). Similarly, the

negative values ofε

Hf during these time intervalsare taken to indicate greater crustal reworkingasso ciated with collision-related crustal thicken-ing. For Rodinia time, the absence of preservedpassive margins, of an enriched, low ε Hf signal,and a spike in seawater 87Sr/ 86Sr ratios suggeststhat margins juxtaposed during supercontinentassembly consisted of largely juvenile mate-rial, perhaps analogous to those rimming themodern Pacic.

EROSION AND THECONTINENTAL RECORD

The clastic sedimentary record, which sam-ples a range of source rocks, and may providethe only record of a source that has been lost dueto erosion, dismemberment, or overprinting, iswidely thought to provide a more representativerecord of the evolution of the continental crustthan the present outcrop of igneous and meta-morphic rocks (Taylor and McLennan, 1995).The sedimentary record is however biased; platemargins are dominated by young rocks, whichare therefore more prone to erosion than olderrocks. Thus, the bulk compositions of sedimentsare biased toward the younger material in the

source terrains (Allègre and Rousseau, 1984).Erosion-induced bias is expressed throughthe erosion factor K , as dened by Allègre andRousseau (1984). K contrasts the relative pro-portions of rocks of different ages in the catch-ment area with the proportion of those sourcerocks present in the sediments analyzed (Fig.11). This has proven difcult to measure innatural systems, and values ranging between 2and 3 have been commonly assumed in previousstudies (Garrels and Mackenzie, 1971; Allègreand Rousseau, 1984; Goldstein and Jacobsen,1988; Jacobsen, 1988; Kramers and Tolstikhin,1997; Kramers, 2002). These values for K wouldsuggest that ~25%–30% of the present volumeof the continental crust had been generated by3 Ga (Fig. 12). Dhuime et al. (2011b) developedan approach to measure K in a modern riversystem using Hf isotopes in detrital zircons andNd isotopes in ne-grained sediments from theFrankland River in southwest Australia. Theydemonstrated that K was variable (4–17), andthat it increased downstream with water volumeand with topographic relief.

The running mean ε Hf value for zircons rang-ing in age from Hadean to Cenozoic is ~0 (e.g.,

Fig. 7; Belousova et al., 2010; Roberts, 2012),

which indicates that there was a signicantcrustal component in the magmas from whichthose zircons crystallized. Crustal melting is inturn often associated with crustal thickening,and hence with areas of high relief. Denudationrates increase with increasing average relief(Summereld and Hulton, 1994), and so thebias in the sedimentary record may be domi-nated by erosion in areas of high relief. If so,and this has yet to be tested rigorously, valuesof K = 10–15 may be more appropriate for ero-sion and deposition of sediments that dominatethe geological record. It follows that mod-els for the evolution of the continental crustbased on analysis of sedimentary rocks shouldbe based on the values of K that characterizeareas of crustal thickening and hence markedtopographic relief. Assuming that they are ap-plicable back into the Archean, such values of K would suggest that 60%–70% of the continentalcrust had been generated by 3 Ga (Fig. 12). In-voking a lower value of K for the Archean (e.g.,K ~2), to reect possible lower topographic re-lief for the major areas of continental erosion,but maintaining a high value for post-Archeanunits ( K ~15), generates a similar growth curve

to that for a single-stage model with a uniform

value of K ~15. This is because the age contrastbetween new crust and preexisting crust in thebinary model of Allègre and Rousseau (1984) issmall in the Archean.

UNRAVELING THE RECORD OFCONTINENTAL GROWTH

Zircons yield high-precision U-Pb crystal-lization ages. Zircons can also be analyzed forHf and O isotopes, and for trace elements, andthey contain silicate inclusions, which can beused to constrain the nature of the host magmafrom which they crystallized (e.g., Jenningset al., 2011).

Hafnium isotope studies of magmatic anddetrital zircons have been utilized to explorethe petrogenesis of granitic rocks (Belousovaet al., 2006; Kemp et al., 2007a) and to unravelcrustal evolution (Patchett et al., 1982; Amelinet al., 1999; Vervoort et al., 1999; Grifn et al.,2004). Hf is concentrated in zircon (~1 wt%),and so zircons have very low Lu/Hf, and theirmeasured 176Hf/ 177Hf ratio approximates that ofthe host magma at its time of generation. TheHf isotope ratios are a measure of the crustal

OLD YOUNG

[x][1 – x]

[y][1 – y]

Continent (source of sediments)

Continental sediments

no preferential erosion(x = y and K = 1)

preferential erosionof young segment(y > x and K > 1)

Figure 11. Schematic repre-sentation of the “preferentialerosion” model used by Allègreand Rousseau (1984) to modelthe growth of the continental

crust through time. The “ero-sion factor,” K , is defined by K = ( y /[1 – y])/( x /[1 – x]).

Time since present (Ga)

V o

l u m e o

f c o n

t i n e n

t a l c r u s t

( % )

0

25

50

7

100

5

0 1 2 3 4

Nd isotopes in shales(from Allègre & Rousseau, 1984)

K = 5 0

K = 2

K = 3

K = 6 K = 4

K = 1 0

K = 1 5

Figure 12. Continental growthcurves for the Gondwana super-continent, calculated from theNd isotope data for Aus tralianshales (Allègre and Rousseau,

1984). The variation of the ero-sion factor K has a dramaticinuence on the shape of thegrowth curves. If K = 1 (i.e., nopreferential erosion of the dif-ferent lithologies producing thesediment), then 30% of the conti-nental crust was generated bythe end of the Archean, but thisincreases to 75% if K = 15.


11/19

Cawood et al.


residence age, i.e., the average time since thesources of the igneous rocks from which the zir-cons crystallized were extracted from the mantle(the Hf “model” age). The calculations of Hfmodel ages from zircon are more difcult to tiedown than, for example, model Nd ages fromNd isotope ratios in sediments. This is primarily

because the Lu/Hf ratio of zircon is much lowerthan that in the host magma, or more criticallyof that in the crustal precursor to the graniticmelts from which the zircon crystallized. TheLu/Hf ratio of the crustal precursor thereforeeither has to be inferred from trends of initialHf isotope ratios versus age, or assumed usingaverage Lu/Hf ratios for mac rocks or the bulkcontinental crust. This inevitably introducesuncertainty, and our approach is to plot distri-butions of calculated model ages so that the rela-tive variations can be evaluated.

The second major assumption involved inthe calculation of model ages is that of the Hfisotope ratios of new continental crust. The de-pleted mantle has been regarded as the comple-mentary reservoir to the continental crust, andbecause the depleted mantle is thought to resideat relatively shallow levels in Earth’s mantle,it has also been regarded as the source of thebasaltic magma involved in the generation ofnew continental crust (Jacobsen and Wasser-burg, 1979; O’Nions et al., 1980; Allègre et al.,1983). Depleted mantle compositions (as mea-sured in mid-ocean-ridge basalt [MORB]) havetherefore typically been used in model age cal-culations, but new continental crust is mostly

generated along destructive plate margins (e.g.,Taylor, 1967; Rudnick, 1995; Davidson andArculus, 2006; Scholl and von Huene, 2007,2009). Such magmas are more enriched iso-topically than MORB, since they tend to containa contribution from recycled sediments (Whiteand Patchett, 1984; Plank, 2005; Chauvel et al.,2008). The implication is that new crust, i.e.,the magmas that cross the Moho, typically havelower Nd and Hf radiogenic isotope ratios thanmagmas generated from the depleted mantle(DePaolo, 1981; Vervoort and Blichert-Toft,1999; Dhuime et al., 2011a). The weightedmean of ε Hf in modern island-arc basaltic lavasis 13.2 ± 1.1 (Dhuime et al., 2011a), lower by~3–4 ε units than the average ε Hf of contempo-rary MORB (e.g., Salters and Stracke, 2004;Workman and Hart, 2005). If model ages arecalculated using the new continental crust evo-lution line, anchored by the present day ε Hf valueof 13.2, they are up to 300 m.y. younger thanthose calculated using depleted mantle compo-sitions (Dhuime et al., 2011a).

A key aspect in developing improved modelsfor the generation and evolution of the continen-tal crust is the ability to interrogate the informa-

tion now available in the hundreds of thousandsof zircon analyses. One difculty is that manycrustally derived magmas contain a contributionfrom sedimentary source rocks, and such sedi-ments typically include material from a numberof different source rocks. Thus, the Hf isotoperatios and the model Hf ages of such magmas,

and the zircons that crystallized from them, arelikely to yield “mixed” source ages that do notprovide direct evidence for when new crust wasgenerated from the mantle. It is therefore neces-sary to nd ways to strip off the zircons that maycontain a contribution from sedimentary sourcerocks; this is most easily done using oxygen iso-topes, and then crust generation models can beconstructed on the basis of the zircon record thatmay more faithfully record when new crust wasgenerated.

Hf isotope analyses of zircons are increas-ingly being combined with oxygen isotopes(i.e., 18O/ 16O, expressed as δ 18O relative to theVienna standard mean ocean water [VSMOW ]standard). The latter are fractionated bysurcial processes, and so the δ 18O value ofmantle-derived magmas (5.37‰–5.81‰ infresh MORB glass; Eiler et al., 2000) contrastswith those from rocks that have experienced asedimentary cycle, which have generally higherδ 18O values. This is reected in the high δ 18Ovalue of the crystallizing zircons, and it is a“fingerprint” for a sedimentary componentin granite genesis, and thus for the reworkingof older crust. The in situ measurement of Oisotope ratios from the same zircon zones ana-

lyzed for Lu-Hf is therefore used to distinguishthose zircons from magmas that include a sedi-mentary component, which might in turn havehybrid model ages. In practice, there are largenumbers of zircon analyses that are not accom-panied by O isotope data. Thus, Dhuime et al.(2012) recently explored the extent to whichthe variations between Hf isotopes and δ 18O inzircons might be generalized in order to evalu-ate the changes in the proportions of reworkedand new crustal material in zircons of differentages. The changing proportions of model agesthought to represent the generation of new crust(new crust formation ages) and of hybrid modelages (those with high δ 18O) are represented bythe black curve in Figure 13A. This was thenused to recalculate the distribution of new crustformation ages (Fig. 13B, green curve) from thedistribution of model ages for ~7000 detritalzircons for which O isotope data are not avail-able (Fig. 13B, black histogram). The shape ofthe green curve suggests that new continentalcrust formation (crust generation) is a continu-ous process. A new model for the evolution ofthe continental crust was then established fromthe changes in the proportions of new and re-

worked crust calculated from the Hf and Oisotope data (Fig. 13B, green and orange histo-grams, respectively). This model suggests that~65% of the present-day volume of the conti-nental crust was already established by 3 Ga(Figs. 6 and 13B, inset), and it is striking thatthis gure of ~65% is similar to that indepen-

dently estimated from Nd isotopes in sedimentsif K is ~15 (Fig. 12). The average growth rateof the continental crust during the rst ~1.5 b.y.of Earth’s history is estimated to be ~3 km 3 ofcrust added to the continental mass each year(Fig. 13B, inset, stage 1). Intriguingly, this issimilar to the rates at which new crust is gener-ated (and destroyed) at the present time (Scholland von Huene, 2007, 2009). There was thena reduction in the net rates of growth of thecontinental crust at ca. 3 Ga, and subsequentlythe rate of crustal growth has been calculatedat ~0.8 km 3 of new crust added each year (Fig.13B, inset, stage 2). This reduction in the aver-age growth rate may primarily reect an in-crease in the rates at which continental crust isdestroyed (recycled), linked perhaps to the sug-gested onset of subduction at ca. 3 Ga (Cawoodet al., 2006; Condie and Kröner, 2008; Shireyand Richardson, 2011; Dhuime et al., 2012).

TECTONIC SETTING ANDCRUSTAL RECORD

The distribution of U-Pb crystallization agesin detrital sediments varies with respect tobasin type, and, like the modal abundances of

the clastic ll (cf. Dickinson and Suczek, 1979;Garzanti et al., 2007), it varies in response totectonic setting (Cawood et al., 2012). The ob-served zircon record within a basinal succes-sion represents the summation of two majorvariables: the presence or absence of synsedi-mentation magmatic activity and the overallspread of ages recorded. For those basins thatcontain igneous zircons with ages close to thetime of sedi ment accumulation, this also re-ects the setting of the magmatic activity, forexample, forearc, trench, and backarc basins atconvergent plate margins. Older grains reectthe prehistory of the basin’s distributive prov-ince and will likely show an episodic patternof peaks and troughs reecting preservationalbias within the supercontinent cycle (Fig. 9;Hawkesworth et al., 2009; Condie et al., 2011).These variables can be represented graphicallyby plotting the distribution of the differencebetween the measured crystallization age for adetrital zircon grain and the depositional age ofthe succession in which it occurs (Fig. 14).

On this basis, detrital zircon data can begrouped into three main tectonic settings:convergent, collisional, and extensional. The


12/19



detrital zircon patterns plotted in this way showa spectrum of results with overlap between thethree broad basin types, particularly for increas-ing age of detritus with respect to depositionalage (Fig. 14). There is, however, a signicantchange in the proportion of ages asso ciatedwith the youngest and older magmatic eventsbetween settings. Convergent margins set-tings are dominated by detrital zircon agesclose to the depositional age of the sediment,and some arc-trench basins display unimodalage spectra. Zircons from collisional basin set-tings contain a lower proportion of grains withcrystallization ages approaching the deposi-tional ages, but they still contain a signicantproportion of grains with ages within 150 m.y.of the host sediment. This pattern reects theinput of material from the magmatic arc thatexisted during ocean closure prior to collisionand the variable amounts of syncollision mag-matism, along with a spectrum of older ages

from cratonic blocks caught with the collisionzone. Detrital zircon age patterns from exten-sional basins are dominated by grains that aresignicantly older than the depositional age ofthe basin. Syndepositional magmatism in ex-tensional settings, such as volcanic rifted mar-gins, is largely of mac composition with a lowyield of zircon (e.g., Fig. 10). The potential ofthis approach is that it can now be applied tosedimentary sequences in old terrains to evalu-ate tectonic settings in areas and periods wherethey are not well constrained. This may be ofparticular use in studies of the Archean, wheretectonic events may have masked originalsettings. Initial studies indicate that the earlyArchean sedimentary sequences at Isua Green-land accumulated at a convergent plate margin,whereas the ancient zircons at Jack Hills Aus-tralia that occur in Late Archean and youngerstrata developed in extensional environments(Cawood et al., 2012).

TECTONIC CONTROLS ONTHE GENERATION OF NEWCONTINENTAL CRUST

Present-day continental crust is predomi-nantly generated through plate tectonics at con-vergent plate margins (Taylor, 1967; Rudnick,1995; Davidson and Arculus, 2006), stabilizedthrough orogenesis (Cawood and Buchan, 2007;Cawood et al., 2009, 2011), and preferentiallypreserved in the long-term geological archivethrough the supercontinent cycle (Hawkesworthet al., 2009, 2010). It is this combination ofcrust generation and subsequent processes thatdetermines the preservation of the crust and isresponsible for the episodic rock record of thecontinental archive.

Analysis of Hf isotopic data from detrital zir-con data sets (Belousova et al., 2010; Dhuimeet al., 2012) suggests that new continental crusthas been generated continuously through time,

New crustformation ages

Hybrid ages

0.0

0.2

0.4

0.6

0.8

1.0

N u m

b e r

P r o

p.

of n

ew

c r u s t f or m

a t i on

a g e s 0

25

50

75

100

Hf model ages

Calculated new crust

Reworkedcrust

N u m

b e r

0

200

400

600

Time since present (Ga)0 1 2 3 4

Stage 1Stage 2

Volume ofcontinental crust(%)

0

25

50

75

100

0 1 2 3 4

A

B

Figure 13. (A) Distribution ofHf model ages in 1376 detritaland inherited zircons sampledworldwide, from which O iso-topes have been measured (fromDhuime et al., 2012, and refer-ences therein). O isotopes were

used by Dhuime et al. (2012)to screen “hybrid model ages”(gray bins) and model ages thatrepresent true periods of crustgeneration (green bins) in theglobal model age distribution.The proportion of new crustformation ages versus hybridmodel ages is represented bythe red dots. These dots dene asystematic variation with time,represented by the black curve.(B) The systematic relationshipdened by the black curve in Awas used by Dhuime et al. (2012)to calculate the distribution ofnew crust formation ages (greenhistogram) from the distributionof model ages in ~7000 detritalzircons worldwide. From thevariations in the proportions ofthe new crust (green histogram)and the reworked crust (orangehistogram, which representsthe distribution of the crystal-lization ages of zircons with Hfmodel ages greater than their

crystallization ages), a continen-tal growth curve has been calcu-lated (blue curve, inset).


13/19

Cawood et al.


with the bulk of the continental crust gener-ated prior to the Proterozoic, and a progressivedecrease in the rate of continual growth sincethe Archean. The volumes of continental crustestimated by this approach are minimum val-ues, as some crust is rapidly recycled such thatit does not leave any isotopic record (cf. Arm-strong, 1981), whereas other crust generates fewzircons (e.g., mac crust). Upper limits on the

volume of crust on early Earth are assumed tobe at or near the current volume (Fyfe, 1978;Armstrong, 1981; see also Fig. 6).

Differences between the predicted minimumvolumes of continental crust and actual pre-served volumes through time are signicant(Fig. 6). Dhuime et al. (2012) argued that some~65% of the current-day crustal volume waspresent by the end Archean, yet compilationsof present age distributions equate to less than5% of the current volume at that time (Goodwin,1996). The striking implication is that much ofthe crust that was generated in the Archean, andsince then, has been destroyed and recycledback into the mantle.

The overall calc-alkaline andesitic composi-tion of continental crust suggests that most ofthe crust was generated by processes similar tomodern-day convergent plate margins (e.g., Rud-nick, 1995; Davidson and Arculus, 2006). Ourcalculations on rates of generation and growthof continental crust of up to ~3 km 3 per annum(Dhuime et al., 2012) are comparable to rates ofmagmatic addition at convergent plate margins(e.g., Scholl and von Huene, 2007, 2009; Cliftet al., 2009). The plate-tectonic mechanism is a

response to the thermal state of the planet. Estab-lishing how long plate tectonics have been themodus operandi of continental growth remainsdifcult to tie down, with suggestions rangingfrom not long after initial formation of the litho-sphere in the Hadean (e.g., Sleep, 2007; Harrisonet al., 2008) to the Neoproterozoic (e.g., Stern,2005). Increasing evidence from a variety ofsources, including geological, paleomagnetic,

geochemical, geophysical, and mineralizationpatterns, suggests that plate tectonics have beenactive since at least the Late Archean (Cawoodet al., 2006; Condie and Kröner, 2008; Rey andColtice, 2008; Sizova et al., 2010). Extrapola-tion of the role of plate tectonics further backinto the Archean or into the Hadean is hinderedby the paucity of the rock record, apart from afew regional remnants (e.g., Isua greenstone belt,Acasta Gneiss, Nuvvuagittuq greenstone belt),along with mineral fragments (detrital zirconsand their inclusions) of appropriate age (e.g.,Jack Hills of the Narryer Gneiss terrane).

Models of the tectonic setting(s) in whichearly continental crust was generated must con-sider its complementary subcontinental litho-spheric mantle, which also shows an episodicand for the most part coeval age distribution(Pearson, 1999; Pearson et al., 2007), albeitbased on a small data set. This implies that thelithosphere as a whole, not just the crust, is af-fected by postgenerational tectonic processes.The buoyancy and strength of Late Archeancratonic lithosphere (Lee et al., 2011), assum-ing that the currently preserved crust is repre-sentative of crust generated at that time, suggest

that the paucity of an earlier record may reectsecular changes in generation and destructionprocesses, probably linked to thermal evolutionof the mantle.

Figure 15 presents a temporal evolution ofearly Earth lithosphere (S. Foley, 2012, personalcommun.). The section is deliberately generic,

avoiding explicit representation of plate-tectonicprocesses, although these are implicit in theproposed development of accretionary and thencollisional orogens after 3.0 Ga. The diagramhighlights one of the critical features of the pre-served early crust, which differentiates it frommore modern crust, the change from an initialbimodal association of TTG (tonalites, trond-hjemites, and granodiorites) and greenstone intoa regime producing continental crust of ande-sitic composition through orogenic cycles. Thisresults in Archean crust having marked bimodaldistributions in silica (Martin, 1993; Rollinson,2010; Van Kranendonk, 2010). Such bimodaldistributions are a feature of intraplate continen-tal ood basalt provinces (e.g., Mahoney andCofn, 1997), and this has led to suggestionsthat much of the Archean crust was generated inintraplate settings, which in many instances havebeen linked to mantle plumes (e.g., Smithieset al., 2005; Bédard, 2006; Bédard et al., 2012).Archean crust and younger crust both havemarked negative Nb anomalies, and these aretypically taken as a strong indication of magmasgenerated in subduction-zone settings (Pearce,1982; Wilson, 1989; Pearce and Peate, 1995).However, such trace-element discriminants have

been established for mac rocks, and they workbest in rocks that can be related reasonably di-rectly to their mac precursors—as in moderndestructive plate-margin settings (Wilson, 1989;Macdonald et al., 2000). In association with amarked silica gap, negative Nb anomalies couldhave been developed during melting of themac source rocks, or they could be a featureof those source rocks, and so they may not bea reliable indicator of the setting in which thehigh-silica rocks were generated. The negativeNb anomalies of the TTG, and hence the aver-age Archean crust, have been attributed to smallamounts of residual rutile (Rollinson, 2010), andthe presence of amphibole (Foley et al., 2002),during partial melting of hydrated source rocks,and as such they do not require a subduction-related setting. A greater challenge, however, isthe presence of water in the source rocks for theTTG (Rapp, 1997), and their marked positivePb anomalies (Rollinson, 2010), and how thesemight be introduced in an intraplate setting.These equivocal geochemical signatures haveresulted in subduction-related and plume-relatedinterpretations, often for similar regions/periods(cf. Bédard, 2006; Wyman, 2012).

Collisional ExtensionalConvergent

b b ca b

c

b

a

0 500 1000 1500 2000 2500 3000

Crystallization age – deposition age (Ma)

50

100

0

C u m u

l a t i v e p r o p o r t

i o n

( % )

A

B

Figure 14. (A) Summary plotof variation of the differencebetween the measured crystal-lization age for a detrital zircongrain and the depositional age

of the succession in which it oc-curs based on cumulative pro-portion curves, and displayedas a function of three maintectonic settings: convergentsetting (a), collisional setting(b), and extensional setting (c).(B) Schematic cross section ofconvergent (a), collisional (b),and extensional (c) plate bound-aries associated with supercon-tinent cycle showing simpliedbasinal settings for accumula-tion of detrital zircons.


14/19



The Jack Hills zircons have received a greatdeal of attention, and yet it is unclear how rep-resentative they are of magmatic process in theHadean and Early Archean (Compston and Pid-geon, 1986; Kober et al., 1989; Mojzsis et al.,2001; Wilde et al., 2001). Kemp et al. (2010)highlighted how different the Hf isotope-crys-

tallization age trends were compared with themagmatic records of destructive plate-marginassociations. Instead, they invoked an enrichedmac crust that might have formed by terrestrialmagma ocean solidication, and interactionwith the nascent hydrosphere at low tempera-tures. Foundering of such a hydrated basalticshell to deeper crustal levels would in turn resultin partial melting and the generation of the TTG(Kamber et al., 2005) and crystallization of thezircons now found in the Jack Hills sediments.Hadean and Early Archean detrital zircons fromWyoming show similar features that are relatedto anhydrous melting of primitive mantle in aplume-like setting (Mueller and Wooden, 2012).

Models of the evolution of continental crust,particularly early ones, focused on processes ofcrustal generation. Armstrong (1991, and refer-ences therein) referred to this as the myth ofcrustal growth, and he was one of the rst toemphasize that crustal growth depends on thebalance between the rates at which new crust isgenerated and the rates at which it is destroyedby weathering and erosion, and ultimately re-turned to the mantle. Such processes are clearlyobserved at the present day as plate tectonicsinvolve the generation and recycling at conver-

gent plate margins through arc magmatism andsediment recycling through sediment subduc-tion and subduction erosion (von Huene andScholl, 1991; Scholl and von Huene, 2007,2009; Stern, 2011). Residual mac material isrecycled back into the mantle in most modelsfor the generation of the relatively evolved bulkcrustal compositions, and such processes havepresumably been active for as long as aver-age continental crust has been generated, i.e.,ca. 3.5 Ga. However, it is much more difcultto establish how long signicant volumes ofcrustal material have been destroyed by erosionand subduction, which in turn require an activeplate-tectonic process.

The period from 3 Ga to the end of the Ar-chean is increasingly viewed as a time ofmarked change in the dominant global tectonicregime(s) (Fig. 15). Shirey and Richardson(2011) noted that eclogitic mineral inclusionsin diamonds, which come from subcontinentalmantle lithospheric keels, are only present after3.0 Ga, and they proposed that this reects theonset of subduction and continental collision,in ways comparable to that at the present day.Some models of crustal growth also indicate a

change in the average rate of continental crustalgrowth ca. 3 Ga (Fig. 6; Taylor and McLennan,1985; Dhuime et al., 2012), as do models thatinvolve erosion factors ( K ) of ~15 in the inter-pretation of the Nd isotope ratios in shales (Fig.12). Figure 15 illustrates the possible changes inthe nature of continental crust and lithospherethat may have taken place toward the end of theArchean in response to an evolving thermal re-gime (S. Foley, 2012, personal commun.). Anearly lithosphere consisting of a bimodal asso-ciation of TTG and greenstone evolved throughdevelopment of accretionary and collisionalorogenic processes into a regime producing

ande sitic crust by at least 3.0 Ga. A compila-tion of δ 18O versus age for zircons shows uni-form values for Archean and older grains but anoverall increasing range of values for youngergrains, which Valley et al. (2005) related tocrustal reworking. This is consistent with data,largely from sedimentation patterns, for increas-ing continental emergence in the Late Archeanand early Paleoproterozoic (Reddy and Evans,2009, and references therein).

CONCLUSION

The continental crust is the archive of Earthhistory, and the present record shows an episodicdistribution of rock units and events. There is in-creasing evidence that this distribution is not aprimary feature reecting processes of genera-tion, as has been assumed by many, but is a con-sequence of secondary processes in which platetectonics resulted in a biased preservational rec-ord. The importance of understanding the roleof tectonics in unraveling the time-integratedhistory of Earth’s continental growth is that itprovides a mechanism for the ongoing continu-ous generation of average continental crustal

compositions of calc-alkaline andesite, and thedestruction and recycling of crust on both short-and long-term time scales through convergentplate interaction and delamination of gravita-tionally unstable crust (cf. Armstrong, 1981).

The history of Earth is primarily read throughthe continental record, particularly prior to thatpreserved in the oceans. Advances in analyticaltechniques are allowing unprecedented interro-gation of the record, but many uncertainties andexciting challenges remain in our understandingof the generation of the continental archive, par-ticularly for early Earth, including:

(1) Differentiating primary and secondary sig-

nals in the rock record. It remains a high prior-ity to distinguish those signals that are sensitiveto the biases introduced by tectonic processesfrom those that are not; whether the episodicage distribution is a generational or preserva-tional feature (Fig. 7) is a rst-order example ofthis issue. Components of the record that pre-serve a temporally related frequency distribu-tion, such as igneous crystallization ages, agesof metamorphism, or of passive margins, arehere regarded as secondary signals modied bythe proportions of rocks and minerals of a spe-cic age that are preserved. Such preservation-related biases are most pronounced for thosecomponents of the record related to top-down,plate-margin–driven processes such as thosecontrolled by the supercontinent cycle (e.g.,Fig. 9). Frequency data driven by bottom-up,deep Earth processes may be more independentof such secondary controls and tend to display aprimary signal. For example, PGE mineral de-posits related to layered intrusion occur withinstable continental crust generally removed fromplate margins, and hence their distribution isless likely to have been modied by the super-continent cycle. Similarly, data sets dened by

Precratonic processes Accretionary orogens

CratonizationStable craton

>3500 Ma 3000–2500 Ma


15/19

Cawood et al.


the rst occurrence of a phase or form (e.g., evo-lution of life), or temporal changes independentof the number of data points, such as those re-lated to seawater and atmospheric proxies (e.g.,Sr in seawater; Fig. 7), are also more likely topreserve primary signals.

(2) The mac record. Current models of

continental volumes (Belousova et al., 2010;Dhuime et al., 2012) are based largely on prox-ies related to the generation of crust with anoverall felsic composition (e.g., zircons) and,hence, are minimum volumes. These models donot take into account volumes of crust unrelatedto such proxy calculations, even though part ofthe resultant volume is dependent on productionof an intermediate mac phase. Calculating thevolumes of mac and ultramac crust generatedwithin the continental record may be possibleby tracking the proportion of U-bearing mineralphases, such as baddeleyite and zirconolite, thatoccur within such lithologies or their derivedsediments (Bodet and Schärer, 2000; Rasmus-sen and Fletcher, 2004; Heaman, 2009; Voiceet al., 2011). As with zircon, these minerals canbe analyzed for both U-Pb crystallization agesand Lu-Hf isotopes to constrain their time ofextraction from the mantle. However, their ap-plicability to the problem of continental crustgeneration is yet to be evaluated and may belimited by their general paucity both in igneousrocks and also in derived sediments due to theirlower physio-chemical resilience than zirconthrough the rock cycle.

(3) Processes involved in generation of early

continental crust. Tectonic models for the gen-eration of continental crust on early Earth re-main controversial, and data from geochemicalanalyses and geodynamic modeling remainequivocal. This reects the signicant differ-ences between the character of the early andcurrent crusts of Earth, which are largely tiedto the Earth’s evolving thermal state, and theresultant lack of modern analogues for compar-ing process and constraining assumptions. Basicquestions to be resolved include when and howdid the continental crustal composition evolvefrom a bimodal to a more continuous distribu-tion with respect to silica and other elements?The distribution pattern of U-Pb crystallizationages in detrital sediments, which can differenti-ate convergent, collisional, and extensional ba-sin settings (Fig. 14), may provide a new wayto constrain tectonic environment and process ofgeneration of Archean greenstone sequences. Itsapplicability will be dependent on sequences inwhich the depositional age of the basin is rea-sonably well constrained, such as the Isua belt,which is rich in volcanic detritus with zirconU-Pb ages close to the depositional age of thebasin and consistent with a convergent plate-

margin setting (Nutman et al., 2009; Cawoodet al., 2012).

(4) Processes of crustal recycling. The strik-ing contrast in crustal volumes between thepreserved exposed record and those predictedby the latest models on crustal growth (Fig. 6)requires that crustal recycling has been active

throughout Earth history (Armstrong, 1981).Recycling on current Earth takes place viasediment subduction and subduction erosion atsites of ongoing convergent plate inter action,through slab breakoff (which may contain acomponent of continental lithosphere) duringthe early stages of continental collision, anddelamination during collision and possibly dur-ing crustal differentiation. Were these processesalso the method of crustal recycling in Archeanand older rocks and what controlled the rates atwhich crustal recycling took place?

ACKNOWLEDGMENTS

We thank the University of St. Andrews for fundingsupport. Steve Foley kindly provided a draft version ofFigure 15, and Jon Davidson and Caroline Wickham-Jones provided photos for Figure 3. Comments anddiscussion from Brendan Murphy, editor of the GSA

Bulletin 125th anniversary celebration articles, andGeorge Gehrels and an anonymous reviewer, alongwith those of Cherry Lewis, Walter Mooney, and DaveScholl, are gratefully acknowledged.

REFERENCES CITED

Allègre, C.J., and Rousseau, D., 1984, The growth of the con-tinent through geological time studied by Nd isotopeanalysis of shales: Earth and Planetary Science Letters,v. 67, p. 19–34, doi:10.1016/0012-821X(84)90035-9.

Allègre, C.J., Hart, S.R., and Minster, J.F., 1983, Chemicalstructure and evolution of the mantle and continentsdetermined by inversion of Nd and Sr isotopic data:I. Theoretical methods: Earth and Planetary ScienceLetters, v. 66, p. 177–190, doi:10.1016/0012-821X(83)90135-8.

Amelin, Y., Lee, D.-C., Halliday, A.N., and Pidgeon, R.T.,1999, Nature of the Earth’s earliest crust from haf-nium isotopes in single detrital zircons: Nature, v. 399,p. 252–255, doi:10.1038/20426.

Armstrong, R.L., 1981, Radiogenic isotopes: The case forcrustal recycling on a steady-state no-continental-growth Earth: Philosophical Transactions of the RoyalSociety of London, ser. A, Mathematical and PhysicalSciences, v. 301, p. 443–472, doi:10.1098/rsta.1981.0122.

Armstrong, R.L., 1991, The persistent myth of crustalgrowth: Australian Journal of Earth Sciences, v. 38,p. 613–630, doi:10.1080/08120099108727995.

Aubouin, J., 1965, Geosynclines: Amsterdam, Elsevier, 335 p.Barley, M.E., and Groves, D.I., 1992, Supercontinentcycle and the distribution of metal deposits throughtime: Geology, v. 20, p. 291–294, doi:10.1130/0091-7613(1992)0202.3.CO;2.

Barrell, J., 1914a, The strength of the Earth’s crust: The Jour-nal of Geology, v. 22, p. 289–314, doi:10.1086/622155.

Barrell, J., 1914b, The strength of the Earth’s crust: The Jour-nal of Geology, v. 22, p. 441–468, doi:10.1086/622163.

Barrell, J., 1914c, The strength of the Earth’s crust: The Jour-nal of Geology, v. 22, p. 655–683, doi:10.1086/622181.

Bartley, J.K., Semikhatov, M.A., Kaufman, A.J., Knoll,A.H., Pope, M.C., and Jacobsen, S.B., 2001, Globalevents across the Mesoproterozoic–Neoproterozoicboundary: C and Sr isotopic evidence from Siberia:Precambrian Research, v. 111, p. 165–202, doi:10.1016

/S0301-9268(01)00160-7.

Bédard, J.H., 2006, A catalytic delamination-driven modelfor coupled genesis of Archaean crust and sub-con-tinental lithospheric mantle: Geochimica et Cosmo-chimica Acta, v. 70, p. 1188–1214, doi:10.1016/j.gca.2005.11.008.

Bédard, J.H., Harris, L.B., and Thurston, P., 2012, The hunt-ing of the snArc: Precambrian Research (in press).

Begg, G.C., Grifn, W.L., Natapov, L.M., O’Reilly, S.Y.,Grand, S.P., O’Neill, C.J., Hronsky, J.M.A., Djomani,Y.P., Swain, C.J., Deen, T., and Bowden, P., 2009, The

lithospheric architecture of Africa: Seismic tomogra-phy, mantle petrology, and tectonic evolution: Geo-sphere, v. 5, p. 23–50, doi:10.1130/GES00179.1.

Belousova, E.A., Grifn, W.L., and O’Reilly, S.Y., 2006, Zir-con crystal morphology, trace element signatures andHf isotope composition as a tool for petrogenetic mod-elling: Examples from eastern Australian granitoids:Journal of Petrology, v. 47, p. 329–353, doi:10.1093

/petrology/egi077.Belousova, E.A., Kostitsyn, Y.A., Grifn, W.L., Begg, G.C.,

O’Reilly, S.Y., and Pearson, N.J., 2010, The growthof the continental crust: Constraints from zircon Hf-isotope data: Lithos, v. 119, p. 457–466, doi:10.1016

/j.lithos.2010.07.024.Bierlein, F.P., Groves, D.I., and Cawood, P.A., 2009, Metal-

logeny of accretionary orogens—The connection be-tween lithospheric processes and metal endowment:Ore Geology Reviews, v. 36, p. 282–292, doi:10.1016

/j.oregeorev.2009.04.002.Bleeker, W., 2003, The Late Archean record: A puzzle in

ca. 35 pieces: Lithos, v. 71, p. 99–134, doi:10.1016 /j.lithos.2003.07.003.

Bodet, F., and Schärer, U., 2000, Evolution of the SE-Asiancontinent from U-Pb and Hf isotopes in single grainsof zircon and baddeleyite from large rivers: Geo-chimica et Cosmochimica Acta, v. 64, p. 2067–2091,doi:10.1016/S0016-7037(00)00352-5.

Boyd, F.R., 1989, Compositional distinction between oce-anic and cratonic lithosphere: Earth and Planetary Sci-ence Letters, v. 96, p. 15–26, doi:10.1016/0012-821X(89)90120-9.

Boyd, F.R., Pokhilenko, N.P., Pearson, D.G., Mertzman,S.A., Sobolev, N.V., and Finger, L.W., 1997, Com-position of the Siberian cratonic mantle: Evidencefrom Udachnaya peridotite xenoliths: Contributionsto Mineralogy and Petrology, v. 128, p. 228–246, doi:10.1007/s004100050305.

Bradley, D.C., 2008, Passive margins through Earth history:Earth-Science Reviews, v. 91, p. 1–26, doi:10.1016

/j.earscirev.2008.08.001.Bradley, D.C., 2011, Secular trends in the geologic record

and the supercontinent cycle: Earth-Science Reviews,v. 108, p. 16–33, doi:10.1016/j.earscirev.2011.05.003.

Brown, M., 2007, Metamorphic conditions in orogenicbelts: A record of secular change: International Geol-ogy Review, v. 49, p. 193–234, doi:10.2747/0020-6814.49.3.193.

Campbell, I.H., and Allen, C.M., 2008, Formation of super-continents linked to increases in atmospheric oxygen:Nature Geoscience, v. 1, p. 554–558, doi:10.1038

/ngeo259.Carey, S.W., 1958, A tectonic approach to continental drift,

in Carey, S.W., ed., Continental Drift: A Symposium:Hobart, University of Tasmania, p. 177–355.

Cawood, P.A., and Buchan, C., 2007, Linking accretion-ary orogenesis with supercontinent assembly: Earth-Science Reviews, v. 82, p. 217–256, doi:10.1016

/j.earscirev.2007.03.003.Cawood, P.A., and Hawkesworth, C.J., 2012, Temporal re-

lations between mineral deposits and global tectoniccycles: Implications for prospectivity, in Jenkins, G.,ed., Ore Deposits and Earth History: Geological Soci-ety of London Special Publication (in press).

Cawood, P.A., Kröner, A., and Pisarevsky, S., 2006, Pre-cambrian plate tectonics: Criteria and evidence: GSAToday , no. 7, v. 16, p. 4–11, doi:10.1130/GSAT01607.1.

Cawood, P.A., Kröner, A., Collins, W.J., Kusky, T.M.,Mooney, W.D., and Windley, B.F., 2009, Accretionaryorogens through Earth history, in Cawood, P.A., andKröner, A., eds., Earth Accretionary Systems in Spaceand Time: Geological Society of London Special Pub-lication 318, p. 1–36.


16/19



Cawood, P.A., Leitch, E.C., Merle, R.E., and Nemchin,A.A., 2011, Orogenesis without collision: Stabilizingthe Terra Australis accretionary orogen, eastern Aus-tralia: Geological Society of America Bulletin, v. 123,p. 2240–2255, doi:10.1130/B30415.1.

Cawood, P.A., Hawkesworth, C.J., and Dhuime, B., 2012,Detrital zircon record and tectonic setting: Geology,v. 40, p. 875–878, doi:10.1130/G32945.1.

Chauvel, C., Lewin, E., Carpentier, M., Arndt, N.T., andMarini, J.-C., 2008, Role of recycled oceanic basalt and

sediment in generating the Hf-Nd mantle array: NatureGeoscience, v. 1, p. 64–67, doi:10.1038/ngeo.2007.51.

Clarke, F.W., 1924, The Data of Geochemistry: U.S. Geo-logical Survey Bulletin 770, 841 p.

Clift, P.D., 2010, Enhanced global continental erosion andexhumation driven by Oligo-Miocene climate change:Geophysical Research Letters, v. 37, p. L09402, doi:10.1029/2010GL043067.

Clift, P.D., Vannucchi, P., and Morgan, J.P., 2009, Crustalredistribution, crust–mantle recycling and Phanerozoicevolution of the continental crust: Earth-Science Re-views, v. 97, p. 80–104, doi:10.1016/j.earscirev.2009.10.003.

Cogley, J.G., 1984, Continental margins and the extent andnumber of the continents: Reviews of Geophysics, v. 22,p. 101–122, doi:10.1029/RG022i002p00101.

Collins, A.S., and Pisarevsky, S.A., 2005, Amalgamatingeastern Gondwana: The evolution of the circum-Indianorogens: Earth-Science Reviews, v. 71, p. 229–270,doi:10.1016/j.earscirev.2005.02.004.

Compston, W., and Pidgeon, R.T., 1986, Jack Hills, evidenceof more very old detrital zircons in Western Australia:Nature, v. 321, p. 766–769, doi:10.1038/321766a0.

Condie, K.C., 1993, Chemical composition and evolution ofthe upper continental crust: Contrasting results fromsurface samples and shales: Chemical Geology, v. 104,p. 1–37, doi:10.1016/0009-2541(93)90140-E.

Condie, K.C., 1994, Greenstones through time, in Condie,K.C., ed., Archean Crustal Evolution: Developmentsin Precambrian Geology: Amsterdam, Elsevier,p. 85–120.

Condie, K.C., 1995, Episodic ages of greenstones: A key tomantle dynamics?: Geophysical Research Letters, v. 22,p. 2215–2218, doi:10.1029/95GL01804.

Condie, K.C., 1998, Episodic continental growth and super-continents: A mantle avalanche connection?: Earth andPlanetary Science Letters, v. 163, p. 97–108, doi:10.1016

/S0012-821X(98)00178-2.Condie, K.C., 2000, Episodic continental growth models:Afterthoughts and extensions: Tectonophysics, v. 322,p. 153–162, doi:10.1016/S0040-1951(00)00061-5.

Condie, K.C., 2004, Supercontinents and superplume events:Distinguishing signals

Geological Society of America Bulletin 2013 Cawood 14 32

Documents

Transcript of Geological Society of America Bulletin 2013 Cawood 14 32