Provenance history of the Bangemall Supergroup and implications for the Mesoproterozoic paleogeography of the West Australian Craton

This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/copyright

http://www.elsevier.com/copyright

Author's personal copy

Precambrian Research 166 (2008) 93–110

Contents lists available at ScienceDirect

Precambrian Research

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

Provenance history of the Bangemall Supergroup and implications for theMesoproterozoic paleogeography of the West Australian Craton

D. McB. Martina,1, K.N. Sircombeb, A.M. Thornea,∗, P.A. Cawoodc, A.A. Nemchind

a Geological Survey of Western Australia, 100 Plain Street, East Perth 6004, WA, Australiab Onshore Energy and Minerals Division, Geoscience Australia, GPO Box 378, Canberra, ACT 2601, Australiac Tectonics Special Research Centre, School of Earth and Geographical Sciences, University of Western Australia, 35 Stirling Highway, Crawley 6009, WA, Australiad West Australian School of Mines, Curtin University, GPO Box U 1987, Perth 6845, WA, Australia

a r t i c l e i n f o

Article history:Received 8 June 2006Received in revised form 14 June 2007Accepted 25 July 2007

Keywords:Bangemall SupergroupEdmund GroupCollier GroupMesoproterozoicPaleocurrentsProvenanceDetrital zircon geochronologyRodinia

a b s t r a c t

The 4–10 km-thick Bangemall Supergroup, comprising the Edmund and Collier Groups, was depositedbetween 1620 and 1070 Ma in response to intracratonic extensional reactivation of the Paleoproterozoiccompressional Capricorn Orogen. The supergroup can be further divided into six depositional packagesbounded by unconformities or major marine flooding surfaces. U–Pb dating of over 1200 detrital zircongrains from 19 samples representative of each of the major sandstone units within these packages hasfailed to identify any zircon populations attributable to syndepositional magmatism. However, this exten-sive dataset provides a provenance history of the Bangemall Supergroup, which is here integrated withpaleocurrent data which indicates that all source areas were located within the Mesoproterozoic WestAustralian Craton, with the main source area for the northern Bangemall Supergroup being the GascoyneComplex and southern Pilbara Craton. All samples have prominent age modes in the 1850–1600 Ma range,indicating significant contribution from the northern Gascoyne Complex and coeval sedimentary basins.Some samples also display prominent modes in the 2780–2450 Ma range, consistent with derivation fromthe Fortescue and Hamersley Groups of the southern Pilbara Craton. The Edmund Group has age-spectrain which the dominant modes become older upwards, recording unroofing of the underlying basementfrom the Gascoyne Complex to the Archean granites and greenstones of the Pilbara Craton. In contrast,the Collier Group records unroofing of the underlying Edmund Group, with possible additional contri-bution from the Pilbara Craton and Paterson Orogen, and is characterized by age-spectra in which thedominant modes become younger upwards. These data imply that the West Australian Craton remainedintact throughout the Mesoproterozoic assembly of Rodinia, and was the only source of detritus for theBangemall Supergroup.

Crown Copyright © 2008 Published by Elsevier B.V. All rights reserved.

1. Introduction

The Bangemall Supergroup is a Mesoproterozoic successionconsisting of ∼4–10 km of mostly fine-grained siliciclastic andcarbonate sedimentary rocks that unconformably overlie Paleopro-terozoic igneous and metamorphic rocks of the Gascoyne Complex,and Paleoproterozoic sedimentary rocks on the margins of theArchean Yilgarn and Pilbara cratons (Fig. 1). The succession is sub-divided into a lower Edmund Group and an upper, unconformablyoverlying Collier Group (Martin and Thorne, 2004). Deposition ofthe Bangemall Supergroup occurred in response to intracontinental

∗ Corresponding author. Fax: +61 8 222 3633.E-mail address: [email protected] (A.M. Thorne).

1 Current address: BHP Billiton, 225 Georges Terrace, Perth, WA 6000, Australia.

extensional reactivation of structures formed during the Paleopro-terzoic Capricorn and Mangaroon Orogenies (Cawood and Tyler,2004; Sheppard et al., 2005). The Bangemall Supergroup was subse-quently deformed during the Neoproterozoic Edmundian Orogeny(Martin and Thorne, 2004), and is overlain by Neoproterozoic toPhanerozoic strata of the Officer Basin to the east (Perincek, 1996;Williams, 1992), and by the Phanerozoic Carnarvon Basin to thewest (Fig. 1).

The age of the Bangemall Supergroup is poorly constrained,although deposition must postdate the intrusion of c.1620 Ma gran-ites into the unconformably underlying Gascoyne Complex (Martinand Thorne, 2004). The most reliable age constraints are pro-vided by a suite of c.1465 Ma dolerite sills intruded exclusivelyinto the Edmund Group, and a suite of c.1070 Ma dolerite sills thatwere intruded mainly into the Collier Group (Wingate, 2002). Bothdolerite suites show localized evidence of magma interaction with

0301-9268/$ – see front matter. Crown Copyright © 2008 Published by Elsevier B.V. All rights reserved.doi:10.1016/j.precamres.2007.07.027


94 D.McB. Martin et al. / Precambrian Research 166 (2008) 93–110

wet or partially lithified sediments, including soft-sediment defor-mation, fluidization, and quench fragmentation (Martin, 2003;Muhling and Brakel, 1985). The depositional age of the EdmundGroup must therefore be slightly older than 1465 Ma, and the CollierGroup slightly older than 1070 Ma.

During the course of recent regional geological mapping ofthe Bangemall Supergroup, 19 samples of sandstone have beencollected from the major sandstone-dominated units for SHRIMPU–Pb detrital zircon analysis (Figs. 2 and 3). The initial objectiveof this sampling strategy was to constrain the depositional age ofthe Bangemall Supergroup more precisely by attempting to iden-tify penecontemporaneous magmatic zircon populations youngerthan the underlying basement that may have been introduced bysyndepositional volcanic activity. Since all zircon populations iden-tified can be attributed to basement derivation, the detrital zircon

age data were integrated with extensive paleocurrent data witha view to determining the provenance history of the BangemallSupergroup and the Mesoproterozoic paleogeography of the WestAustralian Craton. These results have important implications forunderstanding Rodinia reconstructions by identifying the ages andrelative positions of source areas that contributed detrital zirconsto the Bangemall Supergroup during the assembly of Rodinia.

2. Geological setting

2.1. Regional geology

The study area is situated in the northwestern CapricornOrogen (Fig. 1), where the Bangemall Supergroup uncon-formably overlies Paleoproterozoic igneous and low- to high-grade

Fig. 1. Regional geological setting of the Bangemall Supergroup within the Capricorn Orogen, showing the distribution of the Edmund and Collier basins. Details of the studyarea, indicated by a dashed outline, are presented in Fig. 2. PC = Pilbara Craton, YC = Yilgarn Craton, MS = Mangaroon Syncline, TS = Ti Tree Syncline.


D.McB. Martin et al. / Precambrian Research 166 (2008) 93–110 95

metamorphic rocks of the Gascoyne Complex to the south, and low-grade metasedimentary rocks of the Wyloo and Capricorn Groupsto the north. To the east, the supergroup unconformably overliesArchean to Paleoproterozoic rocks of the Pilbara Craton, Hamers-ley Basin, and Ashburton Basin. Southeast of the study area, thesupergroup is either in faulted contact with the Yilgarn Craton, orunconformably overlies Paleoproterozoic low-grade metasedimen-tary deposits including the Bryah, Yerrida, Padbury, and Earaheedybasins.

The Edmund and Collier Groups (Fig. 2), constitute the Edmundand Collier basins, respectively (Martin and Thorne, 2004). Bothbasins are dominated by fine-grained lithofacies, reflecting theirlargely distal setting with respect to the original basin margins.Shallow marine and fluvial facies are largely restricted to thebasal units of the Edmund Group and the Calyie Formation inthe middle Collier Group (Fig. 2). Facies distributions and lat-eral facies changes in the Edmund Basin are strongly controlledby syndepositional faults, such as the northwest-trending TalgaFault, which forms the southern margin of a significant paleoto-pographic high—the Pingandy Shelf (Martin and Thorne, 2004).The Talga Fault is interpreted as a major reactivated basementstructure that separates the medium- to high-grade GascoyneComplex from the lower grade Wyloo Group to the northeast(Fig. 2). Although this fault was initiated as a synsedimentarynormal fault, downthrown to the southwest, it has also expe-rienced periods of reverse movement during the 1070–750 MaEdmundian Orogeny. Other important syndepositional structureswhich influenced Edmund Basin sedimentation include the faultedsouthwestern limb of the Wanna Syncline, the Lyons River Fault,and the northwest-trending fault system which bounds the Man-garoon and Ti Tree Synclines (Figs. 1 and 2). These basement

faults were less significant during the evolution of the CollierBasin.

2.2. Stratigraphy and paleogeography

The lithostratigraphic subdivision of the Bangemall Supergrouphas recently been revised (Martin and Thorne, 2002), and in addi-tion, six depositional packages have also been identified (Martinand Thorne, 2004). These six depositional packages (Figs. 2 and 3)are bounded by unconformities or major marine flooding surfaces,and are broadly similar in style and scale to the second ordersequences of Vail et al. (1977). Packages 1–4 constitute the EdmundGroup, whereas the Collier Group consists of packages 5 and 6. Sincethese packages represent genetically linked intervals of strata, theyprovide a useful framework for investigating the temporal evolu-tion of the Bangemall Supergroup. The details of this stratigraphicframework have been presented by Martin and Thorne (2004), andare summarized below.

Package 1 consists of fluvial to shallow marine siliciclastic andcarbonate facies of the Yilgatherra and Irregully formations (Fig. 3)that were deposited on a regional unconformity in response to ini-tial extension and expansion of the Edmund Basin. In addition,stratigraphic relationships, paleocurrent directions, and detritalzircon geochronology presented below, suggest that the fluvialMount Augustus Sandstone is a correlative of locally preserved flu-vial facies at the base of the Yilgatherra Formation (Martin andThorne, 2004). The Mount Augustus sandstone has previously beencorrelated with the Bresnahan Group (Cooper et al., 1998), whichunconformably overlies the Wyloo Group and is in turn uncon-formably overlain by the Bangemall Supergroup. Deposition ofpackage 1 was strongly controlled by paleotopographic highs and

Fig. 2. Geology of the study area showing the distribution of depositional packages and location of major structural elements (Martin and Thorne, 2004). Extensive doleritesills have been omitted for clarity. Sample locations are indicated by their Geological Survey of Western Australia (GSWA) sample numbers, with the exception of sample148972 which is located 50 km southeast of Wanna homestead in the Mount Augustus Sandstone (a correlative of the basal Yilgatherra Formation). See Table 1 for precisesample locations.



growth faults, which is reflected in the highly variable paleocur-rent directions (Fig. 3), particularly in the Yilgatherra Formationand Mount Augustus Sandstone (Martin and Thorne, 2004, Fig. 4).Paleocurrent trends (Fig. 3) are consistent with bimodal transportnormal to a west-northwesterly trending basin margin, as well asalong the basin axis.

Package 2 overlies a major marine flooding surface, and con-sists of deltaic to deep-marine siliciclastic and carbonate facies ofthe Gooragoora, Blue Billy, and Cheyne Springs formations (Fig. 3).

Deposition of package 2 took place mainly below wave-base, andwas not significantly influenced by growth faulting. Shallowerwater facies were restricted to the Pingandy Shelf, where paleocur-rent directions in the Gooragoora Formation were predominantlyto the south. Paleocurrent trends are unimodal, and consistent withtransport normal to the west-northwesterly trending basin margin(Fig. 3).

A very low angle regional unconformity marks the base of pack-age 3, and cuts down at least as far as the Yilgatherra Formation

Fig. 3. Composite stratigraphic column of the Bangemall Supergroup showing subdivision into groups, formations, and depositional packages (after Martin and Thorne,2004). Post-depositional dolerite sills have been omitted for clarity. Also shown are the relative stratigraphic positions of detrital zircon samples (listed by GSWA number),and paleocurrent roses for all directional and non-directional paleocurrent indicators in each package. For all rose diagrams, the sector angle is 10◦ , interval and vectormaximum are in number of data, and all vector means are unimodal.



Fig. 4. Combined probability density diagrams (PDDs) and binned frequency histograms for package 1 zircon data, shown together with a summary of the relevant paleocurrentdata. PDDs are presented for all data (light grey shading), and for data between −5% and 5% discordance (dark grey shading). Concordant and total data values indicate thenumber of analyses represented in PDDs between 1500 and 3000 Ma. Number of concordant analyses >3000 Ma are also indicated. Histograms are for concordant data onlyand bin width is 25 Myr throughout. Light grey shaded bands correspond to the depositional or magmatic U–Pb zircon age ranges of dated rock units in the likely source areas(data sources are presented in the text). From oldest to youngest these are the Pilbara granite-greenstone terrane (PGG, consisting of the Maitland River, Sisters, Cutinduna,and Split Rock Supersuites), Fortescue Group (FG), Hamersley Group (HG), Cheela Springs Basalt and associated dolerites (CSB), Wooly Dolomite (WD), Ashburton Formationand Moorarie Supersuite (AF+MS), and Durlacher Supersuite (DS). Hr% value is the relative heterogeneity of the probability density distribution; values less than ∼70 indicategreater homogeneity and dominance of single age modes interpreted as localised provenance, values greater than ∼70 indicate greater heterogeneity and a wider range ofages interpreted as broader provenance (Sircombe, 2004). See Fig. 3 for explanation of the paleocurrent rose diagrams.



at the base of package 1. This unconformity cuts down-sectiontowards the southeast, consistent with relative uplift in this area.Package 3 consists of shallow- to deep-marine siliciclastic andcarbonate facies of the Kiangi Creek and Muntharra formations.Deposition was actively controlled by the Talga Fault, and a gen-eral shallowing of facies to the southeast along the Pingandy Shelfprovides further evidence of uplift in this area during depositionof package 3. Paleocurrent directions determined mainly from solemarks are consistently to the southwest, and are sub-parallel topaleocurrent trends inferred to be orthogonal to the basin margins(Fig. 3).

The base of package 4 is marked by a regional marine flood-ing surface, although a subtle unconformity that cuts into the topof package 3 is locally present on the Pingandy Shelf. Package 4consists mainly of deep-water siliciclastic and carbonate facies ofthe Discovery, Devil Creek, Ullawarra, and Coodardoo formations(Fig. 3) that were deposited below storm wave-base. Facies andthickness changes within package 4 indicate that deposition wasintermittently affected by reactivation of the Talga and Lyons Riverfaults (Fig. 2). Paleocurrent data from package 4 indicate a majorchange in paleogeography, with paleocurrent directions predomi-nantly along the basin axis towards the northwest in the UllawarraFormation (Fig. 3). Sparse paleocurrent data from the CoodardooFormation suggest a reversal of paleocurrent directions towardsthe southeast. Paleocurrent trends, determined from sole marksin package 4 are sub-parallel to the northwesterly paleocurrentdirections and also reflect axial paleoflow.

A regional unconformity at the base of package 5, that cuts suc-cessively down-section towards the east, marks the base of theCollier Group. Package 5 consists of an upward shallowing succes-sion of mainly deep-marine to fluvio-deltaic siliciclastic facies ofthe Backdoor and Calyie formations (Fig. 3). Paleocurrent direc-tions change from southwesterly in the Backdoor Formation tonorthwesterly in the Calyie Formation, reflecting a change in pale-ogeography resulting from uplift to the southeast and erosion ofthe underlying Edmund Group. Paleocurrent trends are largely sub-parallel to the directions recorded in each formation.

Package 6 comprises the Ilgarari Formation, which consists ofdeep-marine mudstone that records transgression and drowningof the Calyie Formation delta complex at the top of package 5. Verylittle is known of the paleogeography of package 6, due to the lim-ited exposure in the study area. However, paleocurrent directions insandstone turbidites indicate a paleoslope towards the northwest(Martin and Thorne, 2004).

3. Detrital zircon geochronology

3.1. Methods

Samples have been collected from all of the major sandstoneunits within the Bangemall Supergroup regardless of their compo-sitional or textural maturity. A full sample description, analyticaldetails, and results are presented in the Compilation of Geochronol-ogy Data (Geological Survey of Western Australia, 2007).

The main objective was to achieve stratigraphic and geographicrepresentation of each of the depositional packages (Figs. 2 and 3).Sample co-ordinates are presented in Table 1. Several kilograms ofeach sample were processed using standard crushing, heavy-liquidand magnetic-separation techniques. Careful attention was madeto capture a broad range of material during magnetic-separationas recommended by Sircombe and Stern (2002) to minimise anypotential biasing.

Zircon analyses were performed on SHRIMP-II instruments atthe Research School of Earth Sciences of the Australian National

University, and the Western Australian SHRIMP facilities of theJohn de Laeter Centre of Mass Spectrometry at Curtin Universityof Technology. To ensure a representative sample, grains were ana-lyzed at random with operator selection limited to avoiding obviousdefects in grains. U–Th–Pb ratios and absolute abundances weredetermined relative to the QGNG and CZ3 reference zircons usingoperating procedures similar to those described in Compston et al.(1984), Clauoe-Long et al. (1995) and Nelson (1997). Analytical sum-maries of the isotopic data are given in Table 2. Data reduction andvisualization employed SQUID (Ludwig, 2001a), ISOPLOT (Ludwig,2001b), mixture modelling (Sambridge and Compston, 1994) andAgeDisplay (Sircombe, 2004). Zircon provenance data from deposi-tional packages 1–5 of the Bangemall Supergroup are presented inthe form of combined probability density diagrams and binned fre-quency histograms (Figs. 4–9). Ages discussed in the following textare all based on 207Pb/206Pb measurements with two sigma errorsunless otherwise stated. Analyses were filtered at 5% discordance(as a measure of the ratio of 206Pb/238U age to 207Pb/206Pb age) priorto further interpretation. Both concordant and discordant data areplotted in Figs. 4–9 and indicate that no obvious bias is introducedto the interpretation by discordance filtering. Analytical results forsample numbers 148972, 152954, 148969, 148970, 148971, 156614,156734, 148973, 148974, 148975, 152962, 148976, 156641, 152964,148977, 152968 are presented as Supplementary material. Analyt-ical results for samples 169093, 152956, and 169061 are given inthe Compilation of Geochronology (Geological Survey of WesternAustralia, 2007).

The statistical adequacy of detrital analyses is also an impor-tant consideration, especially when trying to achieve a quantitativeunderstanding of the provenance data (Fedo et al., 2003). Calcula-tions of statistical adequacy have been made for all the samplesfollowing the method of Vermeesch (2004), and these are pre-sented in Table 2 based on the number of concordant analysesavailable for interpretation. The number of analyses, and thus thestatistical adequacy, for each sample varies markedly in this studyand interpretations based on smaller datasets should be treatedwith caution.

Constraining the age of deposition is frequently a goal of detri-tal zircon analysis, although it must be noted that there is nocommonly accepted best practice method for deriving the maxi-mum age of deposition. As noted by Cawood and Nemchin (2001),the youngest detrital zircons may approximate the time of sedi-ment accumulation at active plate margins with contemporaneousigneous activity (Andean margin) whereas in intra-plate settingswithout igneous activity (e.g., passive margins) the youngest grainsmay be tens to hundreds of millions of years older than the time ofsediment deposition. The approach here has been to interpret theage of the youngest statistically verifiable group of analyses as themaximum depositional age although in some cases single grains orsmall groups have been interpreted on the basis that similar agesappear elsewhere in the succession.

3.2. Results

3.2.1. Package 13.2.1.1. Mount Augustus Sandstone (GSWA 148972). SampleGSWA148972 was collected from the crest of a small isolatedhill approximately 1.4 km south of Ulna Well, to the west of MountAugustus. The sample is of coarse to very coarse-grained felds-pathic sandstone with well-preserved trough cross-stratification.The relative stratigraphic position of the sample within the MountAugustus Sandstone is unknown because no relationships withunderlying or overlying units are exposed. The sample consists ofwell-rounded quartz (∼60%) and alkali feldspar (35%), as well asrare rounded grains of zircon or monazite and sparse biotite and



Table 1Location details for detrital zircon samples from the Bangemall Supergroup (GPS datum WGS84, UTM Zone 50)

GSWA NO Easting Northing Formation Package 100K Map Sheet

152968 490439 7352318 Calyie 5 Elliott Creek148977 477155 7373550 Backdoor 5 Elliott Creek152964 465511 7367056 Backdoor 5 Elliott Creek156541 460580 7389100 Backdoor 5 Elliott Creek148976 460716 7389232 Backdoor 5 Elliott Creek152962 463182 7367237 Coodardoo 4 Elliott Creek148975 458478 7393233 Coodardoo 4 Elliott Creek148974 458182 7394718 Curran Member 4 Elliott Creek169061 411120 7350230 Kiangi Creek 3 Edmund148973 457872 7397432 Kiangi Creek 3 Elliott Creek156734 401693 7448429 Kiangi Creek 3 Maroonah156614 526961 7367422 Kiangi Creek 3 Kenneth Range148971 400044 7383487 Blue Billy 2 Edmund148970 400408 7383044 Gooragoora 2 Edmund148969 395284 7375788 Irregully 1 Mangaroon152954 363584 7444665 Irregully 1 Maroonah169093 362630 7408600 Yilgatherra 1 Maroonah152956 358630 7453930 Yilgatherra 1 Maroonah148972 437561 7323910 Mount Augustus Sandstone 1 Mount Augustus

riebeckite (both < 1%). The quartz is mainly monocrystalline, withsparse grains of polycrystalline quartz. The paleocurrent vectormean at this site is towards the south, although the predominantpaleocurrent direction for the Mount Augustus Sandstone istowards the southeast (Fig. 3).

Ages are dominated by two clusters at c. 1800 and c. 1680 Mawith a couple of scattered ages up to 2500 Ma (Fig. 4). Thec. 1680 Ma cluster consists of 61 ages and yields a weightedmean of 1679 ± 3 Ma (MSWD: 1.0), which is interpreted as themaximum age of deposition. The c. 1800 Ma component con-sists of 25 ages and yields a weighted mean of 1792 ± 4 Ma(MSWD: 1.3). The overall poor quality of the zircon suggests lim-ited sedimentary transportation because such grains are morelikely to be eliminated the further they are transported. Alongwith the tight statistical constraints on the dominant compo-

nents indicating limited mixing suggests proximal sources forthis sedimentary unit. Grain #67 revealed a significant core/rimage pair of 1796 ± 11 and 1687 ± 8 Ma indicating that material inthe older provenance component was recycled in the youngerevent.

3.2.1.2. Yilgatherra Formation (GSWA 169093). Sample GSWA169093 was taken from a 0.5 m-diameter boulder of Yilgath-erra Formation, 2.5 km southwest of Cundarra Bore (GeologicalSurvey of Western Australia, 2007). The sample is of medium-to coarse-grained quartz sandstone. It consists almost entirely ofmonocrystalline quartz with undulose extinction (96–97%), andaccessory tourmaline (2%).

The age distribution is exclusively Proterozoic with three modesyielding weighted means at 1685 ± 20 Ma (4 analyses; MSWD:

Fig. 5. Combined PDDs and binned frequency histograms for samples within package 2. See Figs. 3 and 4 for explanation of plots and abbreviations.

Author's personal copy10

0D

.McB.M

artinet

al./Precambrian

Research166

(2008)93–110

Table 2Analytical summaries of detrital zircon samples statistical adequacy calculations based on Vermeesch (2004)

†p-max% is the probability that the number of analyses will miss a component that comprises 0.05 of the total distribution. ‡f-act the actual fraction of the total distribution for which there is a 95% chance of being missed bythe number of analyses.



Fig. 6. Combined PDDs and binned frequency histograms for samples within package3. See Figs. 3 and 4 for explanation of plots and abbreviations.

0.60), 1799 ± 11 Ma (19 analyses; MSWD: 0.76) and 1978 ± 13 Ma (6analyses; MSWD: 0.46) (Fig. 4). The youngest mode is interpretedas the maximum age of deposition.

3.2.1.3. Yilgatherra Formation (GSWA 152956). Sample GSWA152956 was collected approximately 2 m above the base of theYilgatherra Formation, at the southern end of a prominent sand-stone ridge, 7 km northeast of Boora Boora Bore (GeologicalSurvey of Western Australia, 2007). The sample is of very coarse-grained, trough cross-stratified sandstone with heavy mineral lagsin hand specimen. Trough cross-stratification at this site indicatesa paleocurrent direction to the southeast. The sample consistspredominantly of quartz (83%), K-feldspar (5–7%), and accessoryminerals (10%), including tourmaline (2–3%), and muscovite (1%),with trace amounts of apatite and zircon. Quartz grains are mostlymonocrystalline with undulose extinction. Minor polycrystallinequartz is derived from chert. The composition of this sample sug-

gests derivation from a predominantly plutonic source area, withlesser contribution from recycled sedimentary rocks.

The age distribution is dominated by a Proterozoic mode of 21analyses, yielding a weighted mean age of 1801 ± 7 Ma (MSWD:1.08), which is interpreted as the maximum age of deposition(Fig. 4). There are scatterings of individual ages up to c. 2800 Ma.

3.2.1.4. Irregully Formation (GSWA 152954). Sample GSWA 152954was taken from a thin, massive to planar-laminated, dolomiticsandstone bed close to base of the Irregully Formation, approx-imately 6 km south-southwest of Minnie Springs. This unit isoverlain by trough cross-stratified dolomitic litharenite with troughaxes aligned approximately north-northeast–south-southwest.The paleocurrent direction at this site is probably similar tothat in the underlying Yilgatherra Formation in this area, whichwas directed to the northeast off a local paleotopographic high(Martin and Thorne, 2004). The sample consists of medium- to



coarse-grained, sub-angular mono- and poly-crystalline quartz(18%), accessory lithic grains and minerals (2%), including K-feldspar and zircon, in a matrix of dololutite and recrystallisedsparry dololutite (80%).

Ages are dominated by a single c. 1800 Ma mode, with 24 analy-ses scattered up to c. 3400 Ma without any obvious modality (Fig. 4).Mixture modelling of the c. 1800 Ma mode indicates model ages at1785 ± 7, 1802 ± 22 and 1822 ± 13 Ma. The youngest of these com-ponents is interpreted as the maximum age of deposition.

3.2.1.5. Irregully Formation (GSWA 148969). Sample GSWA 148969was taken from a very thick bed of coarse-grained, ferruginous,dolomitic quartz sandstone, with large scale cross-stratification,in the middle of the Irregully Formation, approximately 2.2 kmnortheast of Needle Hill Bore. This unit is overlain by troughcross-stratified dolomitic litharenite with trough axes alignedapproximately north-northeast–south-southwest. The paleocur-rent direction at this site is probably similar to that in the underlyingYilgatherra Formation in this area, which was directed to the north-east off a local paleotopographic high (Martin and Thorne, 2004).The sample consists of irregular quartz grains commonly withwell-defined rounded detrital cores (40%), inequigranular carbon-ate as cement and detrital grains (35%), and abundant lithic grains(15%) predominantly of low-grade metasedimentary rock. Acces-sory and trace minerals include metamorphic biotite and chlorite,and detrital tourmaline. The paleocurrent direction, indicated by

large-scale trough cross-stratification at this site, is towards thesoutheast.

A c. 1800 Ma mode forms a major component with a broad scat-tering of ages up to c. 3000 Ma without any prominent clusters(Fig. 4). Mixture modelling of the c. 1800 Ma component indicatesmodel ages at 1767 ± 7, 1802 ± 5 and 1820 ± 7 Ma. The youngestgrain age (#91) yields a concordant age of 1761 ± 10 Ma, but theyoungest mixture model component 1767 ± 7 Ma is considered amore robust estimate of maximum depositional age.

3.2.2. Package 23.2.2.1. Gooragoora Formation (GSWA 148970). Sample GSWA148970 was taken from a unit of very thick-bedded, troughcross-stratified, medium-grained quartz sandstone, approximately1.8 km north-northwest of Carnoby Well. The sample is moder-ately fresh, with feldspars altered to clay minerals. It comprisesmonocrystalline quartz (80–85%), colloform-banded kaolinite(15%) and limonitic clasts (5%). No accessory minerals wereobserved in thin section. There are no paleocurrent indicators pre-served at this site, but the regional paleocurrent vector mean of theGooragoora Formation is towards the south.

Ages are dominated by a single mode c. 1780 Ma, with scat-tered individual ages ranging up to c. 2700 Ma (Fig. 5). Becausethe youngest age (#24.1) has low U content, it provides an unreli-able age of 1664 ± 71 Ma (1�). Mixture modelling of the c. 1780 Macomponent produces model ages at 1733 ± 12, 1775 ± 19, 1787 ± 33




and 1820 ± 13 Ma. The youngest model age of 1733 ± 12 Ma is inter-preted as the maximum depositional age.

3.2.2.2. Blue Billy Formation (GSWA 148971). Sample GSWA 148971was taken approximately 2 km north-northwest of Carnoby Well,within a unit of interbedded, very thick-bedded, medium-grained,lithic quartz sandstone and silicified siltstone at the top of the Blue

Billy Formation. Feldspars have been altered to clay minerals, butthe heavy mineral population and lithic fragments are well pre-served. No paleocurrent indicators are preserved at this site, orelsewhere for the Blue Billy Formation, which consists mainly ofplanar-laminated, carbonaceous siltstone.

Ages are dominated by a single mode c. 1790 Ma, with scat-tered individual ages up to c. 3000 Ma (Fig. 5). The dominant mode




consists of 43 ages that yield a weighted mean age of 1787 ± 4(MSWD: 1.07), which is interpreted as the maximum depositionage. The older scattered individual ages appear to have little modal-ity although a cluster of six analyses yields a weighted mean age of2759 ± 6 Ma (MSWD: 1.60).

3.2.3. Package 33.2.3.1. Kiangi Creek Formation (GSWA 156614). Sample GSWA156614 was taken from medium- to thick-bedded, fine-grainedsandstone in Fords Creek, 1.2 km northeast of the confluence ofFords Creek West and Fords Creek East. The sample site is a fewmetres below a distinctive unit of conophyton stromatolites nearthe base of the Kiangi Creek Formation. No paleocurrent indicatorsare preserved at this site, but the paleocurrent vector mean for theKiangi Creek Formation is towards the southwest.

Fig. 9. Combined probability density distributions (PDD) and binned frequency his-tograms for the lower five depositional packages of the Bangemall Supergroup,showing the stratigraphic variation in detrital zircon provenance. See Fig. 4 forexplanation of plot parameters.

A major component of 35 analyses between 1720 and 1800 Madominate the age distribution (Fig. 6). There are scatterings of eightages from 2200 to 2800 Ma. Mixture modelling of the major agecomponent reveals two model ages at 1759 ± 5 and 1790 ± 11 Ma.The younger of these two model ages is interpreted as the maxi-mum age of deposition.

3.2.3.2. Kiangi Creek Formation (GSWA 156734). Sample GSWA156734 was collected from a 25 m-thick unit of thin to very thick-bedded, amalgamated sandstone in the middle of the Kiangi CreekFormation at the Blue Rocks copper mine, 17.7 km east of WongidaWell. The sample is medium-grained quartz sandstone with mud-stone intraclasts ranging in size from coarse sand to pebbles.Many beds show basal normal coarse-tail grading of the mudstoneintraclasts, overlain by low angle cross-laminated to undulatory-laminated sandstone, and are commonly separated by 1–2 mm ofsiltstone. Flute casts and linear tool marks indicate a paleocurrentdirection towards west, although the regional vector mean for theKiangi Creek Formation is towards the southwest.

A major component of 31 analyses between 1700 and 1850 Madominates the age distribution (Fig. 6) and yield a weighted meanage of 1778 ± 10 Ma (MSWD: 1.07), which is interpreted as the max-imum age of deposition. Analyses are scattered up to 3500 Ma,including six analyses with ages over 3000 Ma—the first occurrencein this study of detrital zircon older than the mid-Archean in theBangemall Supergroup samples. Possible clusters occur at c. 2200and c. 2700 Ma.

3.2.3.3. Kiangi Creek Formation (GSWA 169061). Sample GSWA169061 was taken from a prominent outcrop east of the LyonsRiver—Gifford Creek road crossing, 1.5 km southwest of YangibanaYard. This medium-grained sandstone is representative of theupper part of the Kiangi Creek Formation at this locality, and con-sists mainly of monocrystalline quartz, with minor polycrystallinequartz and authigenic K-feldspar, chlorite, and carbonate, and rarezircon.

The age distribution is dominated by 23 analyses that form a Pro-terozoic mode with a weighted mean age of 1799 ± 11 Ma (MSWD:1.2), which is interpreted as the maximum age of deposition (Fig. 6).A scattering of individual ages ranges up to c. 2700 Ma.

3.2.3.4. Kiangi Creek Formation (GSWA 148973). Sample GSWA148973 was taken from very thick-bedded, silicified, fine- tomedium-grained quartz sandstone at Strama Gap. The sampledhorizon immediately underlies the Discovery Formation on the eastbank of Irregully Creek, 0.7 km northeast of Strama Bore. This sand-stone was previously considered to mark the base of the DiscoveryFormation in this area (Martin and Thorne, 2002), but is now con-sidered to represent the uppermost Kiangi Creek Formation in thisarea. It has no obvious internal stratification or heavy mineral lags,but contains abundant mudstone pebble clasts locally. The sam-ple consists of >99% quartz, commonly with rounded detrital coreswith optically continuous overgrowths.

Based on their anomalously young ages and distinctive sectorzoning, grains #10 and #88 have been identified as contami-nants, believed to be derived from samples from the Musgraveregion of Western Australia that were prepared coevally with thissample. The dominant age component consists of 38 analysesranging from c. 1680 to 1815 Ma. Mixture modelling of these anal-yses yields model components at 1681 ± 20, 1734 ± 8, 1760 ± 12,1782 ± 4 and 1807 ± 6 Ma with the latter two components beingdominant (Fig. 6). The youngest model component is interpretedas the maximum age of deposition. A range of scattered individualages occurs between c. 1900 and 3075 Ma.



3.2.4. Package 43.2.4.1. Curran Member, Ullawarra Formation (GSWA 148974). Sam-ple GSWA 148974 was taken from a unit of interbedded fine-grainedlithic sandstone and siltstone within the Curran Member of theUllawarra Formation, below the type section of the CoodardooFormation at Coodardo Gap, approximately 1.2 km northeast ofCoodardo Well. This fine-grained lithic sandstone has no internalstratification or grading. It is rich in poorly sorted silt to mediumsand-sized detrital grains (35–40%) and has a clay-rich matrix. Thegrains are composed of about 1% each of quartz, plagioclase, andmicrocline, as well as sparse quartz-rich to sericite-rich lithic grainsthat make up 25–75% of some layers. Accessory and trace mineralsinclude muscovite, biotite, chlorite, leucoxene, rutile, anatase, andtourmaline. The paleocurrent vector mean for the Curran Memberis towards the northwest.

The age distribution is evenly spread between Proterozoic andArchean and marks the first appearance of dominant amounts ofArchean-aged detrital zircon in the Bangemall Supergroup, with68% of the analyses >2500 Ma (Fig. 7). Two Proterozoic modes yieldweighted mean ages of 1680 ± 8 Ma (MSWD: 1.0) and 1798 ± 5 Ma(MSWD: 0.75) with the former interpreted as the maximum deposi-tion age. Along with a broad scatter of ages from c. 2000 to 3000 Ma,distinct clusters of Archean ages occur at c. 2400–2600 Ma and c.2650–2800 Ma. These are complex clusters, as demonstrated bymixture modeling. In the c. 2400–2600 Ma cluster the modeledages are 2458 ± 4, 2496 ± 6, 2525 ± 5 and 2553 ± 5 Ma. In the c.2650–2800 Ma cluster the modelled ages are 2685 ± 13, 2694 ± 6and 2711 ± 6 Ma.

3.2.4.2. Coodardoo Formation (GSWA 148975). Sample GSWA148975 is of medium-grained feldspathic quartz sandstone in theupper part of the Coodardoo Formation, on the northern limb ofthe Wanna Syncline, about 3.7 km south-southeast of Strama Bore.At this locality, the Coodardoo Formation consists of thick-bedded,massive sandstone, with a few laminated siltstone interbeds. Thesample consists mainly of monocrystalline quartz (∼85%) with rareoptically continuous overgrowths. Lithic grains are common (5–7%)and vary from cherty to microgranular and from quartz-rich tosericite-rich. Interstitial limonite-stained clays are common (5–7%).Anatase/rutile and rounded zircon are rare. Paleocurrent indica-tors are sparse in the Coodardoo Formation, but the vector meanazimuth is towards the southeast.

The age distribution is dominated by Archean ages, with 68% ofthe analyses >2500 Ma (Fig. 7). Grain #29 provides an intriguinglyconcordant and relatively young result at c. 1600 Ma, however arepeat analysis is strongly reversely discordant suggesting a prob-lem with either the sample or the analysis. Either way, this result isregarded as unreliable for further interpretation. Two Proterozoicmodes occur, although the first consists of only two analyses onone grain (#44) yielding a weighted mean of 1674 ± 24 Ma (MSWD:1.19). This result can only be treated as the maximum age of depo-sition by the virtue of similarly aged analyses being present inalmost all the other samples being discussed here. The secondProterozoic mode is a little more robust with six analyses yield-ing a maximum depositional age of 1807 ± 7 Ma (MSWD: 1.2). Theage distribution is dominated by a cluster ranging from c. 2450to 2550 Ma. Mixture modelling indicates prominent model agesat 2498 ± 8, 2516 ± 6 and 2539 ± 6 Ma. A subsidiary cluster rangesfrom c. 2600 to 2750 Ma with mixture model ages at 2647 ± 9,2685 ± 8 and 2701 ± 16 Ma. There are also individual ages scatteredup to 3402 ± 7 Ma (#31).

3.2.4.3. Coodardoo Formation (GSWA 152962). Sample GSWA152962 is of coarse-grained sandstone from the upper part of theCoodardoo Formation in the type section at Coodardo Gap, on the

southern limb of the Wanna Syncline. At this locality, sandstonebeds are generally medium- to thick-bedded, massive to normallygraded, and interbedded with thin to medium beds of parallel-planar-laminated siltstone. The sample consists mainly of poorlysorted, sub-angular to well-rounded grains of strained monocrys-talline quartz, with polycrystalline quartz and microquartz eachcomprising <5% of the rock. Plagioclase and alkali feldspar arepresent in trace amounts. The paleocurrent direction is towardsthe southeast.

The age distribution is evenly spread with Archean ages con-sisting 60% of the total (Fig. 7). Two distinctive modes are alsoseen in the Archean with a cluster of ages between 2500 and2600 Ma yielding mixture model age components at 2528 ± 7 and2565 ± 10 Ma. The second Archean cluster yields a weighted meanage of 2705 ± 11 Ma (6 analyses; MSWD: 1.6). Scattered individualages also include the oldest zircon seen in this study, at 3597 ± 7 Ma(#60). Two Proterozoic modes occur with weighted means at1692 ± 15 Ma (9 analyses; MSWD: 1.3) and 1793 ± 13 (8 analyses;MSWD: 0.95). The former is interpreted as the maximum age ofdeposition.

3.2.5. Package 53.2.5.1. Backdoor Formation (GSWA 148976). Sample GSWA 148976is of the lowermost thick-bedded sandstone (0.4 m thick) in theBackdoor Formation on the northern limb of the Wanna Syn-cline, on Irregully Creek, approximately 3 km north-northwest ofDog Pool Bore. The sample consists of fine-grained, parallel strat-ified quartz sandstone that in thin section consists of angular tosubrounded monocrystalline quartz (∼65%) and cherty clasts ofmicrocrystalline quartz (1–2%) in a matrix of interstitial carbon-ate (∼30%). Accessory and trace minerals include muscovite andpartly altered biotite, leucoxene, brown tourmaline, and zircon. Thepaleocurrent direction at this site, determined from sole marks, istowards the south.

The age distribution is well spread, with Archean ages domi-nating (62%) a broad scattering of ages ranging from c. 2400 to3200 Ma (Fig. 8). There are also a few individual ages at the youngend of the distribution with the youngest individual age, grain #29,being 1640 ± 18 Ma. This result can only be treated as the max-imum age of deposition by the virtue of similarly aged analysesbeing present in almost all the other samples being discussed here.A small Proterozoic-aged cluster of nine analyses yields mixturemodel component ages at 1773 ± 30, 1789 ± 20 and 1807 ± 14 Ma.An Archean-aged cluster of 15 analyses yields a complex mix-ture model with significant model age components at 2503 ± 8,2522 ± 17 and 2537 ± 8 Ma. There are also suggestions of statisti-cally indefinable clusters at c. 2600–2650 Ma and c. 2720 Ma.

3.2.5.2. Backdoor Formation (GSWA 156541). Sample GSWA 156541was collected from the same general vicinity as GSWA 148976, butat a slightly higher stratigraphic level. It was taken from thick-bedded, fine-grained quartz sandstone, with a paleocurrent vectormean towards the south.

The age distribution is dominated by a broad range of Archeanand earliest Proterozoic ages (Fig. 8). Two small Proterozoic clus-ters are noted. The first consists of three c. 1600 Ma grains withduplicates of the youngest grain (#13) yielding a weighted meanage of 1613 ± 33 Ma (MSWD: 0.33). Combining this result with theother two ages in this group produces an unreasonably high MSWDvalue invalidating this as a statistically definable group. On thebasis of similar ages appearing in other samples discussed here, theweighted mean age of grain #13 is tentatively taken as the maxi-mum age of deposition. The second Proterozoic cluster containsseven ages, but the youngest (#54) and oldest (#76) must be dis-carded before a meaningful weighted mean age of 1789 ± 10 Ma



(MSWD: 0.34) is obtained. Most ages fall between c. 2300 and3300 Ma. Mixture modelling reveals a plethora of model com-ponent ages at 2451 ± 8, 2492 ± 6, 2525 ± 6, 2571 ± 8, 2707 ± 6,2732 ± 11 and 2815 ± 7 Ma.

3.2.5.3. Backdoor Formation (GSWA 152964). Sample GSWA 152964is of coarse-grained sandstone in the middle of the Backdoor For-mation on the southern limb of the Wanna Syncline, approximately2.2 km east of Coodardo Gap (Fig. 2). This unit is a correlativeof the Jeeaila Sandstone Member as mapped by previous work-ers east of the study area (Muhling and Brakel, 1985), and locallycontains abundant mudstone intraclasts. The sample is a moder-ately sorted, fine- to medium-grained, sub-angular to well-roundedquartz sandstone, with syntaxial quartz cement. Quartz grains aremainly monocrystalline, with minor chert and microquartz grains,and trace amounts of well-rounded tourmaline and zircon. Troughcross-stratification at this site indicates a paleocurrent directiontowards the east-southeast.

Only 20 analyses have suitable concordance, which greatlyreduces the statistical adequacy of the sample (for instance, thereis a 95% chance of missing a component that comprises 1 in 5 in thetotal population). Therefore, interpretations based on these data areonly tentative. The youngest age is a singular 1621 ± 71 Ma (grain#12), but the associated high common-Pb ensures that this age isnot reliable (Fig. 8). Two provenance components can be tentativelyidentified with weighted mean ages at 1758 ± 24 Ma (7 analyses;MSWD: 0.56) and 2439 ± 18 (5 analyses; MSWD: 1.09).

3.2.5.4. Backdoor Formation (GSWA 148977). Sample GSWA 148977was taken from very thick-bedded, medium-grained, trough cross-stratified quartz sandstone in the upper Backdoor Formation,approximately 150 m below the contact with the Calyie Forma-tion. The sample, from 8.4 km southeast of Bend Bore, consists ofmoderately fresh sandstone, with feldspars partially replaced byclay minerals and locally abundant mudstone intraclasts. Thereis abundant quartz with well-defined rounded cores (80%) andoptically continuous overgrowths in thin section, as well as rareclay-rich and quartz-rich microcrystalline clasts (2–3%). Accessoryand trace minerals include greenish tourmaline, partly with palebluish authigenic overgrowths, leucoxene, and zircon. Paleocurrentdirections are hard to determine at this site, but possibly trendnorth to south or northwest to southeast.

One singularly young grain (#59) with two almost identicalanalyses has been identified as a contaminant from Musgraveregion samples prepared coevally with this sample. The age distri-bution is dominated by a c. 1800 Ma mode comprising 34 analysesthat yield mixture model ages of 1785 ± 5 and 1799 ± 8 Ma (Fig. 8).The former age is taken as the maximum age of deposition. Thereis also a broad range of individual ages from c. 1950 to 3500 Ma inwhich there are no prominent clusters.

3.2.5.5. Calyie Formation (GSWA 152968). Sample GSWA 152968 isfrom near the top of the Calyie Formation, in Gregorys Gap onKoorabooka Creek, approximately 0.7 km south-southeast of God-freys Well. This medium-grained, massive to trough cross-stratifiedsandstone consists of moderately sorted, well rounded, fine- tomedium-grained monocrystalline quartz, with minor chert andpolycrystalline quartz. Well-rounded detrital tourmaline is presentin trace amounts, and feldspar is absent. The grains are cementedwith syntaxial quartz cement, and display moderately suturedmargins. Paleocurrent directions, determined from trough cross-stratification, are towards the west.

Only 18 analyses have suitable concordance, which greatlyreduces the statistical adequacy of the sample (for instance, thereis a 95% chance of missing a component that comprises 1 in 5 in the

total population). Therefore, interpretations based on these dataare only tentative. The age distribution is dominated by a Protero-zoic mode of seven analyses that yield a weighted mean age of1740 ± 18 Ma (MSWD: 1.03) (Fig. 8). The remainder of the concor-dant ages scatter up to c. 3500 Ma with no significant modes.

3.3. Age of potential source areas

Paleocurrent directions within the Bangemall Supergroup indi-cate that the most significant source area lies to the northwestand northeast of the study area, with less important source areaslying to the southwest and southeast (Fig. 3). This suggests that theprincipal source areas were the northern Gascoyne Complex andthe Pilbara Craton. It is also possible that some detritus may havebeen derived from the Rudall Complex in the Paterson Orogen, tothe east-northeast of the study area, which contains rocks of simi-lar age to the Gascoyne Complex. There is no compelling evidenceto suggest that the southern Gascoyne Complex or Yilgarn Cratonprovided first-cycle sedimentary detritus to either the Edmund orCollier basins within the study area. However, detrital zircon popu-lations in some of the potential source rocks, such as the AshburtonFormation, were probably originally derived from the Yilgarn Cra-ton during the Paleoproterozoic assembly of the West AustralianCraton (Sircombe, 2002; Thorne and Seymour, 1991).

The northern margin of the Bangemall Supergroup coincidesapproximately with the boundary between the Ashburton Basinand the Gascoyne Complex (Fig. 1). Potential source rocks exposedsouth of this boundary, immediately to the west and southwest ofthe study area, belong to the Boora Boora and Mangaroon zones ofthe Gascoyne Complex (Martin et al., 2005; Sheppard et al., 2005).The Boora Boora Zone consists of medium-grade metasedimentaryrocks intruded by metamorphosed granites of the Moorarie Super-suite that were later intruded by unmetamorphosed granites ofthe Durlacher Supersuite. The depositional age of the metasedi-mentary rocks is unknown, but is probably similar to that of theAshburton Formation to the north (Sheppard et al., 2005). Granitesof the Moorarie Supersuite were intruded into the entire GascoyneComplex, as well as the Wyloo Group and northern margin of theYilgarn Craton, between 1830 and 1780 Ma (Cawood and Tyler,2004; Martin et al., 2005). The Mangaroon Zone is dominated bymedium- to high-grade metasedimentary rocks and minor meta-igneous rocks of the Pooranoo Metamorphics that were intrudedby granites of the 1680–1620 Ma Durlacher Supersuite (Pearsonet al., 1996; Sheppard et al., 2005). The maximum depositionalage of the Pooranoo Metamorphics is c. 1680 Ma (Sheppard et al.,2005), although the total range of detrital zircon populations is c.2700–1680 Ma (Geological Survey of Western Australia, 2007).

North of the study area, the Bangemall Supergroup is uncon-formably underlain by the Capricorn Group and the 12 km thickWyloo Group (Martin et al., 2005; Thorne and Seymour, 1991).Given the predominance of southerly paleocurrent directions(Fig. 3), these units are likely to have provided a significant amountof detritus, particularly to the lower parts of the Bangemall Super-group. The depositional age of the Capricorn Group, and theAshburton Formation in the upper Wyloo Group (Fig. 2), is c.1800 Ma (Evans et al., 2003; Hall et al., 2001) which is broadlycoeval with the intrusion of granites of the Moorarie Supersuiteinto the Gascoyne Complex and upper Wyloo Group (Cawood andTyler, 2004; Krapez and McNaughton, 1999; Martin et al., 2005;Occhipinti et al., 1998). Detrital zircon populations in the CapricornGroup range from c. 3300 to 1800 Ma (Geological Survey of WesternAustralia, 2007), and the Ashburton Formation from c. 2860 to1780 Ma (K. Sircombe, unpublished data). The lower Wyloo Groupwas deposited between c. 2209 Ma (Martin et al., 1998) and 2031 Ma(Muller et al., 2005). Detrital zircon populations in the lower Wyloo



Group, notably the Mount McGrath Formation and Beasley RiverQuartzite, range in age from c. 3550 to 2450 Ma (Geological Surveyof Western Australia, 2007). The ages of detrital zircon populationsin the Wyloo Group reflect derivation mainly from the underly-ing Mount Bruce Supergroup, with minor contribution from theArchean Pilbara Craton (Fig. 1). The depositional age and detritalzircon age distribution of the Turee Creek Group (Fig. 1) is unknown,but it is older than c. 2208 Ma dolerite sills that intrude it (Mulleret al., 2005), and younger than the c. 2450 Ma Woongara Rhyolite(Barley et al., 1997; Trendall et al., 2004) in the underlying Hamer-sley Group. The Hamersley Group (Fig. 1) has a depositional ageranging from c. 2629 to 2449 Ma, and the underlying FortescueGroup from c. 2780 to 2629 Ma (Thorne and Trendall, 2001; Trendallet al., 2004). Detrital zircons in the Bangemall Supergroup olderthan c. 2800 Ma were most likely derived from the Archean PilbaraCraton, which ranges in age from c. 3550 to 2850 (Van Kranendonket al., 2002).

4. Discussion and conclusions

The combined detrital zircon age data for each depositionalpackage within the Bangemall Supergroup is characterised by poly-modal distributions (Fig. 9), with a distinct difference in provenanceevident between the lower (packages 1–3) and upper (packages4–5) parts of the stratigraphy. Packages 1–3 are characterisedby prominent late Paleoproterozoic modes, with less significantinput from early Paleoproterozoic and Archean sources. The notablechange in provenance between packages 1–3 and 4–5 can alsobe quantified using functional estimation (Sircombe and Hazelton,2004) to calculate the ‘distance’ between the age distributions asillustrated in the dendrogram of Fig. 10. Package 4 and 5 sam-ples form a distinct cluster from the remaining package 1, 2 and3 samples. The exceptions are GSWA 152964, perhaps reflectingthe statistically poor quantity of data available in this sample, andGSWA 148977 which is a possible indication of a return to theProterozoic-dominated provenance near the top of the sampledsequence.

The dominant mode in the Bangemall Supergroup is c. 1800 Ma,with the exception of package 1 which has an additional latePaleoproterozoic mode at c. 1675 Ma. In contrast, packages 4–5are characterised by polymodal zircon age distributions withprominent late Archean to early Paleoproterozoic and late Pale-oproterozoic modes and typically higher relative heterogeneityvalues. With the exception of package 1, there are very few detritalzircons in the c. 2400–1850 Ma age range. There are also relativelyfew detrital zircons older than 3000 Ma, although their abundanceincreases upwards (Fig. 9). In order to assess the significance ofthese changes in provenance, it is necessary to look at the strati-graphic distribution of samples within each package, and theirrelationship to the known paleogeography.

Package 1 is characterised by a progressive upward increasein the number of detrital zircon modes from a strongly bimodaldistribution in the Mount Augustus Sandstone, to a strongly poly-modal distribution in the upper Irregully Formation (Fig. 4). Theprominent mode at c. 1800 Ma probably reflects erosion of sourceareas dominated by the Ashburton Formation and coeval MoorarieSupersuite, or younger rocks derived from these sources (Fig. 11).The slightly younger mode at c. 1680 Ma, in samples GSWA 148972and GSWA 169093 (Fig. 4), can be directly related via paleocurrentdirections to erosion of the Durlacher Supersuite, which also uncon-formably underlies these sample sites (Fig. 11). The origin of detritalzircons in the 2400–1850 Ma range is enigmatic because there arevery few magmatic source rocks of this age in the potential sourceregion. However, there are detrital zircon populations that occupy

this age range in the Ashburton Formation and Pooranoo Meta-morphics that were probably originally derived from the southernGascoyne Complex (Bertibubba and Dalgaringa Supersuites) andsupracrustal rocks on the Yilgarn Craton. Paleocurrent directionswithin the upper parts of package 1 suggest that late Archean andearly Paleoproterozoic detrital zircon populations were probablyderived from erosion of the Hamersley and Fortescue Groups onthe southern margin of the Pilbara Craton (Fig. 11).

The detrital zircon distributions of the two samples within pack-age 2 are similar to those that characterize the upper parts ofpackage 1 (Fig. 5). There is a dominant c. 1800 Ma mode, with a num-ber of significantly smaller modes extending back to c. 2800 Ma. Theyounger mode is most likely derived from the Ashburton Formationand Moorarie Supersuite, with older modes being derived from theArchean to Paleoproterozoic supracrustal successions on the south-ern margin of the Pilbara Craton (Fig. 11). There may, however, havebeen some contribution to these modes through minor reworkingof package 1.

Package 3 is dominated by the c. 1800 Ma mode, with individ-ual samples containing relatively small older modes (generally nomore than 2–3 zircons) extending back to c. 2900 Ma (Fig. 6). Whencombined with paleocurrent data and the presence of a significantunconformity at the base of package 3 that cuts down as far as thebase of package 1, these distributions suggest that a significant pro-portion of the detrital zircons in package 3 were probably derivedthrough reworking of underlying packages. However, the presenceof detrital zircons older than 3000 Ma (Fig. 6), which are largelyabsent from the Ashburton Formation (Sircombe, 2002), suggeststhat the Archean Pilbara Craton or material deposited during anearlier exposure of the craton was also being eroded at this time(Fig. 11).

The precise stratigraphic location of the significant change inprovenance between package 3 and package 4 (Fig. 9) is unclearbecause all the samples from package 4 come from the two upper-most formations (Fig. 3). This change is marked by a noticeablereduction in the size of the c. 1800 Ma mode and an equally signif-icant increase in the size of modes in the 2800–2400 Ma age range(Fig. 7). Also, all samples in package 4 have discrete modes in the1800–1600 Ma age ranges that are uncommon in the underlyingpackages. These changes in zircon provenance are accompanied bya change in paleocurrent direction in the preserved Edmund Basinfrom transverse to axial (Fig. 3). The abundance of zircons withmodes in the 2800–2400 Ma age range is consistent with deriva-tion from erosion of the Hamersley and Fortescue Groups, whichwere probably exposed close to the Edmund Basin. This impliesthat the detritus deposited within the study area was introducedinto the Edmund Basin by transverse paleocurrents further to theeast (Fig. 11). Suppression of the younger Paleoproterozoic modessuggests that erosion of source areas containing zircons of this agewas reduced, perhaps due to the significant deepening and expan-sion of the Edmund Basin at the base of package 4 (Martin andThorne, 2004).

The strong modes in the 2800–2400 Ma age range, establishedin package 4, extend into the base of package 5 (Fig. 8). Younger andless significant detrital modes are also present in the 1800–1600 Maage range. Higher up in package 5, above the level of the JeeailaSandstone Member of the Backdoor Formation, the detrital zircondistribution is more typical of packages 2 and 3, with a prominent1800 Ma mode (Fig. 8). This change in detrital zircon distribu-tion is accompanied by a change in paleocurrent direction fromsouth-southwesterly in the lower Backdoor Formation, to south-easterly in the middle Backdoor Formation (GSWA 152964), towest-northwesterly in the Calyie Formation. The detrital zircon dis-tribution at the base of package 5 can be interpreted as due to eithercontinued derivation from early Paleoproterozoic supracrustal



successions on the Pilbara Craton, or to reworking of package 4on the regional unconformity at the base of package 5 (Fig. 11).Both sources are equally likely and are consistent with paleo-geographic constraints. The upper parts of package 5 were mostprobably derived from reworking of older packages below pack-age 4, since the package thickens in a southeasterly directionand c. 1800 Ma basement was probably not widely exposed inthis area. Source rocks of this age are present in the RudallComplex of the Paterson Orogen (Bagas, 2004). However, uncer-tainty in timing of emplacement and relative position to thestudy area diminishes their likely contribution to the CollierBasin.

The provenance history of the Bangemall Supergroup can bedivided into two discrete stages that reflect changes in paleo-geography between the Edmund and Collier Groups. The EdmundGroup records erosion of the Gascoyne Complex and progressiveunroofing of early to late Paleoproterozoic supracrustal rocks onthe southern margin of the Pilbara Craton. During this time, thepaleogeography of the Edmund Basin evolved from a horst-and-graben terrain with significant detrital contribution from localizedsources, to an extensive east–west trending intra-cratonic basinwith the Pilbara Craton to the north providing the main sourceof detritus. Early stages of extension during package 1 may have

been due to orogenic collapse of the Capricorn Orogen shortly afterthe 1680–1620 Ma Mangaroon Orogeny (Sheppard et al., 2005).However, the details of the timing and tectonic drivers of later pale-ogeographic changes in packages 2–4 are unclear, and some of thesepackages may be significantly younger than package 1. These uncer-tainties cannot be resolved until the ages of packages 1–4 are betterconstrained.

In contrast, the Collier Group records unroofing and rework-ing of successively lower levels of the underlying Edmund Group,with possible additional contribution from the Pilbara Cratonand Paterson Orogen. This change in provenance is accompa-nied by the formation of a regional unconformity at the baseof package 5, and a progressive change in paleocurrent direc-tion from south-southwesterly to west-northwesterly, consistentwith uplift to the east of the Collier Basin. This uplift maybe due either to reactivation of the Paterson Orogen prior tothe Miles Orogeny, or to intrusion of the regionally extensiveWarakurna large igneous province above a mantle-plume headlocated beneath central Australia (Martin, 2003; Wingate et al.,2004).

The integration of detrital zircon age-spectra with paleocurrentdata shows that the present distribution of rocks in the Gas-coyne Complex and Pilbara Craton was already in existence by

Fig. 10. Dendrogram illustrating the distance and groupings among the age distributions based on kernel functional estimation (Sircombe and Hazelton, 2004). Circlednumbers on the nodes of the dendrogram indicate the package to which each sample belongs.



the Mesoproterozoic, and can adequately explain the provenancehistory of the Bangemall Supergroup. Furthermore, this historyis consistent with current models for the amalgamation of Pro-terozoic Australia (e.g., Myers et al., 1996; Wingate and Evans,2003), and its position within Rodinia (e.g., Pisarevsky et al., 2003;Wingate et al., 2002). In particular, the data presented here pro-vide no evidence for the existence of exotic source areas marginalto the preserved West Australian Craton, such as those requiredto explain the provenance of the coeval Belt Supergroup in NorthAmerica (e.g., Ross and Villeneuve, 2003). This implies that theMesoproterozoic West Australian Craton either had a greater north-ern and western extent than presently preserved, or was located

along an external passive margin prior to the amalgamation ofRodinia. The former interpretation is favored by the existenceof the c. 755 Ma Mundine Well dyke swarm, which is alignedsub-parallel to the northwestern margin of the craton and is inter-preted to have been emplaced during separation of an unknowncontinental fragment from the West Australian Craton (Wingateand Giddings, 2000). This fragment is currently considered to bethe amalgamated Kalahari-Dronning Maud Land Craton, whichcollided obliquely with the western margin of the West Aus-tralian Craton at 1100–1000 Ma (Fitzsimons, 2002; Pisarevsky etal., 2003) after deposition of the bulk of the Bangemall Super-group.

Fig. 11. Schematic representation of the evolution of source areas and dispersal paths for packages 1–5 of the Bangemall Supergroup, relative to the known distribution ofpotential source rocks within the West Australian Craton. Dashed lines denote uncertain source areas and dispersal paths. Abbreviations: P, Pilbara Craton; Y, Yilgarn Craton;WAC, West Australian Craton; NAC, North Australian Craton; SAC, South Australian Craton; G, Gawler Craton. See text for discussion.



Acknowledgements

David Martin and Alan Thorne publish with the permissionof the Director of the Geological Survey of Western Australia.Suzanne Dowsett is thanked for her help with the preparationof the figures. John Williamson and staff at the Carlisle Labora-tory are especially thanked for their pain-staking zircon separationand mount-making. Paleocurrent roses and statistics were pro-duced using GEOrient 9.2 by R.J. Holcombe. This is Tectonics SpecialResearch Centre Publication Number 411.

Appendix A. Supplementary data

Supplementary data associated with this article can be found,in the online version, at doi:10.1016/j.precamres.2007.07.027.

References

Bagas, L., 2004. Proterozoic evolution and tectonic setting of the northwest PatersonOrogen, Western Australia. Precambrian Res. 128, 475–496.

Barley, M.E., Pickard, A.L., Sylvester, P.J., 1997. Emplacement of a large igneousprovince as a possible cause of banded iron formation 2.45 billion years ago.Nature 385, 55–58.

Cawood, P.A., Nemchin, A.A., 2001. Source regions for Laurentian margin sediments:constraints from U/Pb dating of detrital zircon in the Newfoundland Appalachi-ans. Geol. Soc. Am. Bull. 113, 1234–1246.

Cawood, P.A., Tyler, I.M., 2004. Assembling and reactivating the Proterozoic Capri-corn Orogen: lithotectonic elements, orogenies, and significance. PrecambrianRes. 128, 201–218.

Clauoe-Long, J.C., Compston, W., Roberts, J., Fanning, C.M., 1995. Two Carbonifer-ous ages: a comparison of SHRIMP zircon ages with conventional zircon agesand 40Ar/39Ar analysis. In: Berggren, W.A., Kent, D.V., Aubrey, M.P., Hardenbol,J. (Eds.), Geochronology, Time Scales and Global Stratigraphic Correlation, pp.3–21, SEPM Special Publication 54.

Compston, W., Williams, I.S., Meyer, C., 1984. U–Pb geochronology of zircons fromlunar breccia 73217 using a sensitive high mass-resolution ion microprobe. J.Geophys. Res. Suppl. 89, B525–B534.

Cooper, R.W., Langford, R.L. and Pirajno, F., 1998. Mineral occurrences and explo-ration potential of the Bangemall Basin, Report 64. Western Australia GeologicalSurvey, 42 pp.

Evans, D.A.D., Sircombe, K.N., Wingate, M.T.D., Doyle, M., McCarthy, M., Pidgeon,R.T., Van Niekerk, H.S., 2003. Revised geochronology of magmatism in the west-ern Capricorn Orogen at 1805–1785 Ma: diachroneity of the Pilbara–Yilgarncollision. Aust. J. Earth Sci. 50, 853–864.

Fedo, C.M., Sircombe, K.N., Rainbird, R.H., 2003. Detrital zircon analysis of the sedi-mentary record: Zircon. Rev. Mineral. Geochem. 53, 277–303.

Fitzsimons, I.C.W., 2002. Comparison of detrital zircon ages in the Pinjarra Orogen(WA) and Maud Province (Antarctica): evidence for collision of Western Aus-tralia with Southern Africa at 1100 Ma. In: Preiss, V.P. (Ed.), Geoscience 2002:Expanding Horizons. Geological Society of Australia, Adelaide, p. 228, AbstractsNo. 67.

Geological Survey of Western Australia, 2007. Compilation of geochronology data,June 2007 update. Western Australia Geological Survey.

Hall, C.E., Powell, C.McA., Bryant, J., 2001. Basin setting and age of the Late Palaeopro-terozoic Capricorn Formation, Western Australia. Aust. J. Earth Sci. 48, 731–744.

Krapez, B., McNaughton, N.J., 1999. SHRIMP zircon U–Pb age and tectonic significanceof the Palaeoproterozoic Boolaloo Granodiortite in the Ashburton Province,Western Australia. Aust. J. Earth Sci. 46, 283–287.

Ludwig, K. R., 2001a. SQUID 1.02: a user’s manual, Berkeley Geochronology CenterSpecial Publication 2.

Ludwig, K. R., 2001b. Using Isoplot/Ex, Rev. 2.49, A geochronological Toolkit forMicrosoft Excel, Berkeley Geochronology Center Special Publication 1a.

Martin, D.McB., 2003. Peperite in the Backdoor Formation and its significance to theage and tectonic evolution of the Bangemall Supergroup. West. Aust. Geol. Surv.,53–59.

Martin, D.McB., Li, Z.X., Nemchin, A.A., Powell, C.M., 1998. A pre-2.2 Ga age for gianthematite ores of the Hamersley Province, Australia? Econ. Geol. 93, 1084–1090.

Martin, D.McB., Sheppard, S., Thorne, A.M., 2005. Geology of the Maroonah,Ullawarra, Capricorn, Mangaroon, Edmund, and Elliott Creek 1:100 000 sheets.1:100 000 Geological series explanatory notes. West. Aust. Geol. Surv., 65.

Martin, D.McB., Thorne, A.M., 2002. Revised lithostratigraphy of the Mesoprotero-zoic Bangemall Supergroup on the Edmund and Turee Creek 1:250 000 mapsheets, Record 2002/15. West. Aust. Geol. Surv., 27 pp.

Martin, D.McB., Thorne, A.M., 2004. Tectonic setting and basin evolution of theBangemall Supergroup in the northwestern Capricorn Orogen. Precambrian Res.128, 385–409.

Muhling, P.C., Brakel, A.T., 1985. Geology of the Bangemall Group—the evolution ofan intracratonic Proterozoic basin, Bulletin 128. West. Aust. Geol. Surv., 266.

Muller, S.G., Krapez, B., Barley, M.E., Fletcher, I.R., 2005. Giant iron-ore deposits ofthe Hamersley province related to the breakup of Paleoproterozoic Australia:new insights from in situ SHRIMP dating of baddeleyite from mafic intrusions.Geology 33, 577–580.

Myers, J.S., Shaw, R.D., Tyler, I.M., 1996. Tectonic evolution of Proterozoic Australia.Tectonics 15, 1431–1446.

Nelson, D., 1997. Compilation of SHRIMP U–Pb Zircon Geochronology Data. 1996.Geological Survey of Western Australia. Record. 1997/2.

Occhipinti, S.A., Sheppard, S., Nelson, D.R., Myers, J.S., Tyler, I.M., 1998. Syntectonicgranite in the southern margin of the Palaeoproterozoic Capricorn Orogen, West-ern Australia. Aust. J Earth Sci. 45, 509–512.

Pearson, J.M., Taylor, W.R., Barley, M.E., 1996. Geology of the alkaline Gifford CreekComplex, Gascoyne Complex, Western Australia. Aust. J Earth Sci. 43, 299–309.

Perincek, D., 1996. The stratigraphic and structural development of the Officer Basin,Western Australia. West. Aust. Geol. Surv., 135–148.

Pisarevsky, S.A., Wingate, M.T.D., Powell, C.McA., Johnson, S., Evans, D.A.D., 2003.Models of Rodinia assembly and fragmentation. In: Yoshida, M., Windley, B.F.,Dasgupta, S. (Eds.), Proterozoic East Gondwana: Supercontinent Assembly andBreakup. Geol. Soc., London, pp. 35–55.

Ross, G.M., Villeneuve, M., 2003. Provenance of the Mesoproterozoic (1.45 Ga) Beltbasin (western North America): another piece in the pre-Rodinia paleogeo-graphic puzzle. Geol. Soc. Am. Bull. 115, 1191–1217.

Sambridge, M.S., Compston, W., 1994. Mixture modelling of multi-component datasets with application to ion-probe zircon ages. Earth Planet. Sci. Lett. 128,373–390.

Sheppard, S., Occhipinti, S.A., Nelson, D.R., 2005. Intracontinental reworking in theCapricorn Orogen, Western Australia: the 1680–1620 Ma Mangaroon Orogeny.Aust. J. Earth Sci. 52, 443–460.

Sircombe, K., 2002. Reconnaissance detrital zircon geochronology provenance of thePalaeoproterozoic Ashburton Formation: implications for Pilbara and Yilgarnamalgamation. In: 16th Australian Geological Convention, Geological Society ofAustralia, Adelaide, p. 147.

Sircombe, K.N., 2004. AgeDisplay: an EXCEL workbook to evaluate and displayunivariate geochronological data using binned frequency histograms and prob-ability density distributions. Comput. Geosci. 30, 21–31.

Sircombe, K.N., Hazelton, M.L., 2004. Comparison of detrital zircon age distributionsby kernel functional estimation. Sediment. Geol. 171, 91–111.

Sircombe, K., Stern, R., 2002. An investigation of artificial biasing in detrital zir-con U–Pb geochronology due to magnetic separation in sample preparation.Geochim. Cosmochim. Acta 66, 2379–2397.

Thorne, A.M., Seymour, D.B., 1991. Geology of the Ashburton Basin Western Australia,Bulletin 139. Western Australia Geological Survey, 141 pp.

Thorne, A.M., Trendall, A.F., 2001. Geology of the Fortescue Group, Pilbara Craton,Western Australia, Bulletin 144. Western Australia Geological Survey, 249 pp.

Trendall, A.F., Compston, W., Nelson, D.R., De Laeter, J.R., Bennett, V.C., 2004. SHRIMPzircon ages constraining the depositional chronology of the Hamersley GroupWestern Australia. Aust. J. Earth Sci. 51, 621–644.

Vail, P.R., Mitchum Jr., R.M., Todd, R.G., Widmier, J.M., Thomson III, S., Sangree, J.B.,Bubb, J.N., Hatleilid, W.G., 1977. Seismic stratigraphy and global changes of sealevel. In: Payton, C.E. (Ed.), Seismic Stratigraphy—Applications to HydrocarbonExploration. American Association of Petroleum Geologists, pp. 49–212.

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 Terrane, PilbaraCraton, Western Australia. Econ. Geol. 97, 695–732.

Vermeesch, P., 2004. How many grains are needed for a provenance study? EarthPlanet. Sci. Lett. 224, 441–451.

Williams, I.R., 1992. Geology of the Savory Basin, Western Australia, Bulletin 141.Western Australia Geological Survey, 115 pp.

Wingate, M.T.D., 2002. Age and palaeomagnetism of dolerite sills of the Bange-mall Supergroup on the Edmund 1:250000 map sheet W.A. Western AustraliaGeological Survey, pp. 48.

Wingate, M.T.D., Evans, D.A.D., 2003. Palaeomagnetic constraints on the Proterozoictectonic evolution of Australia. In: Yoshida, M., Windley, B.F., Dasgupta, S. (Eds.),Proterozoic East Gondwana: Supercontinent Assembly and Breakup. GeologicalSociety, London, pp. 77–91.

Wingate, M.T.D., Giddings, J.W., 2000. Age and paleomagnetism of the MundineWell dyke swarm, Western Australia: implications for an Australia–Laurentiaconnection at 755 Ma. Precambrian Res. 100, 335–357.

Wingate, M.T.D., Pirajno, F., Morris, P.A., 2004. Warakurna large igneous province: anew Mesoproterozoic large igneous province in west-central Australia. Geology32, 105–108.

Wingate, M.T.D., Pisarevsky, S.A., Evans, D.A.D., 2002. Rodinia connections betweenAustralia and Laurentia: no SWEAT, no AUSWUS? Terra Nova 14, 121–128.

Documents

Provenance history of the Bangemall Supergroup and implications for the Mesoproterozoic paleogeography of the West Australian Craton