PtL

Ha

b

c

a

ARRAA

KNGLCP

1

i(hpaB1gd

of

h0

Precambrian Research 252 (2014) 166–179

Contents lists available at ScienceDirect

Precambrian Research

jo ur nal home p ag e: www.elsev ier .com/ locate /precamres

etrogenesis of Gunbarrel magmatic rocks: Homogeneous continentalholeiites associated with extension and rifting of Neoproterozoicaurentia

amish A. Sandemana,∗, Luke Ootesa, Brian Cousensb, Taylor Kilianc

Northwest Territories Geoscience Office, Box 1500, Yellowknife, NT, Canada X1A 2R3Department of Earth Sciences, Carleton University, 1125 Colonel By Drive, Ottawa, ON, Canada K1S 5B6Yale University, 210 Whitney Avenue, New Haven, CT, United States

r t i c l e i n f o

rticle history:eceived 25 October 2012eceived in revised form 30 May 2014ccepted 14 July 2014vailable online 23 July 2014

eywords:eoproterozoicunbarrel eventithospheric extensionontinental tholeiitesetrogenesis

a b s t r a c t

The ca. 780 Ma Gunbarrel Igneous Event of northwest Laurentia consists of spatially discrete suites ofsills, dykes and lavas distributed over a vast area extending from Wyoming in the south to the Wop-may Orogen and the Mackenzie Mountains of Northwest Canada. Thick (≤100 m) sills and rare dykes inWopmay orogen and thinner (≤30 m) sills, dykes and rare lavas in the Mackenzie Mountains are mod-erately evolved, augite + oligoclase–labradorite + ilmenite–magnetite gabbros and amygdaloidal basalts.Systematic petrochemical differences between units reveals that each is likely derived from subtly dis-tinct parental magmas collectively exhibiting mutually consistent element variations. The dataset isremarkably homogenous, in particular, the incompatible trace elements and the Sm–Nd isotopes. Allrocks preserve petrochemical evidence of an enriched MORB-like mantle source, but a small litho-spheric component in the primary magmas resulted in elevated LILE, minor negative HFSE anomaliesand sub-depleted mantle but supra-bulk earth �Nd values. The lithospheric component was slightlyolder, modestly fractionated, Sr-depleted, garnet-free (pyroxenitic?) lower crust or, similar material thatwas previously recycled into the lithospheric mantle. Mineral chemical data for plagioclase and clinopy-

roxene in chill margin samples from a Hottah sheet in Wopmay orogen, indicates rapid and repeatedturbulent mixing of geochemically and thermally similar magmas. These were staged from large, lower-most crust(?) magma chambers centred over an asthenospheric thermochemical anomaly thought to lieto the west of present-day North America. These magmas were then rapidly emplaced across westernLaurentia.

. Introduction

Mafic Large Igneous Provinces (LIPs, Coffin and Eldholm, 1994)ncluding continental flood basalts (CFB) and mafic dyke swarmsMDS) are recognized on many of Earths ancient cratons andave been widely correlated on the basis of similar and overlap-ing precise U–Pb baddeleyite ages (Heaman et al., 1992; Heamannd LeCheminant, 1993; Ernst, 2007; Heaman, 2008; Ernst andleeker, 2010) and paleogeographical data (Buchan et al., 1993,

998; Halls and Davis, 2004; Li et al., 2008; Evans, 2009). Inte-rated geochronological, paleomagnetic, as well as geochemicalata on ancient mafic igneous events can therefore be powerful

∗ Corresponding author. Current address: Geological Survey of NL, Departmentf Natural Resources, Government of Newfoundland and Labrador, St. John’s, New-oundland, Canada A1B 4J6. Tel.: +1 709 727 3721; fax: +1 709 729 4270.

E-mail address: hamishsandeman@gov.nl.ca (H.A. Sandeman).

ttp://dx.doi.org/10.1016/j.precamres.2014.07.007301-9268/© 2014 Elsevier B.V. All rights reserved.

© 2014 Elsevier B.V. All rights reserved.

tools in the reconstruction of previous supercontinents. Recon-struction of Earths ancient cratons, which have utilized mainlypaleomagnetic and geochronological data for mafic dyke swarms,has resulted in the construction of global “geochronological bar-codes” for diverse cratons (e.g., Bleeker and Ernst, 2006; Ernstet al., 2008) establishing a robust tool for supercontinental recon-structions. However, a lack of robust geochronological data formany global large igneous provinces, in particular for Precambrianexamples, obfuscates global correlations (e.g., Wingate et al., 1998).Contributing to the difficulties in correlation, robust lithogeochem-ical and radiogenic tracer isotopes for LIPS are widely lacking,making the testing of constraints problematic. Even in well studied,younger Pangean LIPS of the Atlantic borderlands, a firm under-standing of the temporal, spatial and compositional variations is asyet largely incomplete.

Rodinia, the last Precambrian supercontinent was amalga-mated in the Proterozoic and was fragmented between ca. 1300and 600 Ma (Li et al., 2008; Evans, 2009). During this protracted

brian Research 252 (2014) 166–179 167

bitIat1t(LcLFbBeNl(Ng(2

b21elglaLa(ti7ps0dNadofitc

2p

dondS2btHsa

Fig. 1. Simplified geological map of northwestern North America. Thick black linewith arrows indicates movement on Tintina fault; barbs on Cordilleran frontal thrustare on the hangingwall side. Sill traces under the Phanerozoic platform are fromnew aeromagnetic data in these regions (http://www.nrcan.gc.ca/earth-sciences/products-services/geoscience-data-repository/11824). The dyke trends, plume cen-tre, and locations of the Christmas Lake dyke and other Gunbarrel outcrops are fromPark et al. (1995b) and Harlan et al. (2003). More detailed maps of the sill outcrops inthe northern Mackenzie Mountains and Wopmay orogen are in Park et al. (1995b),Sandeman et al. (2007) and Ootes et al. (2008). AR – Arctic Red River area; KR – Keele

H.A. Sandeman et al. / Precam

reak-up event, mafic dyke swarms, cogenetic sills, ultramaficntrusions and volcanic products were emplaced into continen-al crust undergoing fragmentation and rifting (Ernst et al., 2008).n northwest North America, Neoproterozoic mafic rocks associ-ted with three distinct magmatic events were emplaced duringhis interval (Buchan and Ernst, 2004): the areally extensive ca.270 Ma Mackenzie dyke swarm along with correlative basalts ofhe Coppermine River area and the ultramafic Muskox intrusionDupuy et al., 1992; LeCheminant and Heaman, 1989; Heaman andeCheminant, 1993; Schwab et al., 2004; Mackie et al., 2009); thea. 780 Ma Gunbarrel event (Armstrong et al., 1982; Dudás andustwerk, 1997; Harlan et al., 2003) and; the youngest, ca. 720 Maranklin event incorporating the Coronation Sills and Natkusiakasalts (Heaman et al., 1992; Dupuy et al., 1995; Pehrsson anduchan, 1999; Shellnut et al., 2004). Previously proposed and gen-rally contested correlations of late Neoproterozoic LIPs betweenorth America and other continents has been based largely on geo-

ogical similarities, radiometric ages and paleomagnetic settingse.g., Park et al., 1995b; Wingate et al., 1998; Zhou et al., 2002).umerous other correlations have been suggested using strati-raphic or paleomagnetic data and apparent polar wander pathse.g., Moores, 1991; Sears and Price, 2000; Li et al., 2008; Evans,009).

In western North America, the Gunbarrel sills (Fig. 1) haveeen precisely dated by U–Pb baddeleyite methods (Harlan et al.,003) and have well-constrained paleogeographic data (Park et al.,989, 1995a,b). With few exceptions (Sandeman et al., 2007; Ootest al., 2008; Ernst and Buchan, 2010), however, modern and robustithogeochemical and radiogenic isotopic data for these 780 Maabbroic and basaltic rocks are lacking. Prior to this work, only aimited dataset comprising a few partial geochemical analyses arevailable for the 780 Ma Gunbarrel igneous event (e.g., Dudás andustwerk, 1997; Ernst and Buchan, 2010), making lithogeochemicalnd isotopic comparison with other Neoproterozoic LIPs difficulte.g., Park et al., 1995b; Wingate et al., 1998; Zhou et al., 2002). Inhis contribution, we present a modern and robust lithogeochem-cal dataset including Nd radiogenic isotopic data for some of the80 Ma Gunbarrel rocks in northwestern North America. The sam-les are from two areally separate geological settings of gabbroicills and dykes in the Neoproterozoic Mackenzie Mountains (ca..85 Ga) and a third group of samples from gabbroic sheets andykes in the Paleoproterozoic Wopmay orogen (ca. 1.9 Ga) of theorthwest Territories of Canada. We supplement these data with

single Sm–Nd isotopic analysis of the analogous Christmas Lakeyke (Condie et al., 1969) that cuts the Archaean Wyoming cratonf the western United States. Collectively, the datasets provide therst rigorous lithogeochemical fingerprints on the Gunbarrel eventhat can help to better constrain the proposed correlations withontemporaneous LIPs on other continents.

. Regional setting, U–Pb geochronology andaleomagnetic background

In Canada, Gunbarrel rocks comprise gabbroic to basaltic sills,ykes and lavas in the northern Cordillera (Mackenzie Mountains)f the Northwest Territories, dykes in the Rocky Mountains oforthern British Columbia and sheets (subhorizontal but not bed-ing concordant) and dykes in the western Precambrian Canadianhield of the Northwest Territories (Park et al., 1989; Harlan et al.,003; Ernst and Buchan, 2004; Fig. 1). They also occur as dykes inoth the Archaean Wyoming craton of the Rocky Mountains and in

he Tobacco Root Mountains of Montana (Fig. 1; Condie et al., 1969;arlan et al., 2003). Sills, sheets, less common dykes and a thin

equence of amygdaloidal basalt flows in the Mackenzie Mountainsnd Wopmay orogen are the focus of this study. In the Mackenzie

River area; F – Faber sheet; H – Hardisty dyke; M – Margaret sheet; C – Calder sheet;G – Gunbarrel sheet; CLD – Christmas Lake dyke.

Mountains (Fig. 1), stratiform gabbro sills, locally referred to asTsezotene sills, intrude the Neoproterozoic Mackenzie MountainsSupergroup (<1083 Ma; Jefferson and Parrish, 1989), a clasticsedimentary sequence deposited in an epicratonic sea through topassive margin environment (Turner and Long, 2008). The sillsare restricted to the northern parts of the Cordillera where rocksof the Mackenzie Mountains Supergroup are exposed and wereexamined in a northwestern Arctic Red River transect (AR: Fig. 1)and in a southeast Keele River area (KR: Fig. 1). With the exceptionof the correlative Little Dal basalts (LD: Fig. 1), a locally preservedseries of tholeiitic basalt flows at the top of the Little Dal Groupof the Mackenzie Mountains Supergroup (Narbonne and Aitken,1995; Dudás and Lustwerk, 1997), the sills and correlative dykes

are the only identified Proterozoic magmatic rocks in the region(Ernst and Buchan, 2004). They are most common as 1–50 m thick,bedding concordant sheets in the Tsezotene Formation and as sills

1 brian R

a(FiS

toHeHptssndWW(tWoctp

w1nd2sDlPfCt1tTyGuarf∼tp

3

tgoeNftfil

68 H.A. Sandeman et al. / Precam

nd rare dykes in the overlying Katherine and Little Dal groupsAitken et al., 1982; Armstrong et al., 1982; Ootes et al., 2008).arther to the south in the Mackenzie Mountains of the NWT andn northern British Columbia, the Neoproterozoic Windermereupergroup or younger Paleozoic strata cover Gunbarrel rocks.

In the Canadian Shield, the Gunbarrel magmas intruded intohe Great Bear magmatic zone and Coronation margin of the Pale-proterozoic Wopmay orogen, where they are referred to as theottah sheets (Park et al., 1995a; Harlan et al., 2003; Sandemant al., 2007; Ernst and Buchan, 2004; Jackson et al., 2013). Theottah sheets outcrop as 10 to >50 m thick, typically gently dip-ing sills having an overall southwest strike. From south to northhese include the Faber sheet (labelled as F in Fig. 1), the Margaretheet (M), the Hardisty dyke (H), the Calder sheet (C) and Gunbarrelheet (G) along with a number of other reported dykes that wereot visited during this investigation. All of the sills and the oneyke (Hardisty dyke) identified herein cross the lithospheric-scaleopmay fault zone and intrude the metamorphic internal zone ofopmay orogen which includes reworked late Archaean basement

Fig. 1; Jackson et al., 2013). The sheets cut all units and representhe youngest magmatic event in Wopmay orogen. The west side of

opmay orogen is covered by a Phanerozoic platformal sequencef flat lying Cambrian through Devonian and Cretaceous-Tertiarylastic sedimentary rocks. Regional aeromagnetic data facilitateshe westward tracing of the magnetic Gunbarrel gabbros beneatharts of this platform (Fig. 1).

Initial geochronological work on Gunbarrel rocks yielded Rb–Srhole-rock ages of 766 ± 24 and 769 ± 27 Ma (Armstrong et al.,

982). These have been refined using the U–Pb baddeleyite tech-ique, yielding a more precise pooled 207Pb/206Pb age from allated localities, indicating intrusion at 780.3 ± 1.4 Ma (Harlan et al.,003). Paleomagnetic data exist for a handful of both igneous andedimentary units associated with the ca. 780 Ma Gunbarrel event.ata from the Little Dal basalts (Morris and Aitken, 1982), corre-

ated with the ca. 777 Ma Little Dal quartz diorite (Jefferson andarrish, 1989) are unfortunately, regarded as untrustworthy asaulting associated with the nearby (<4 km) Cretaceous-Tertiaryoates Lake thrust may have rotated many outcrops. Other unitshat have yielded stable paleomagnetic data (Park and Jefferson,991) are either slightly older, such as the >810 Ma Rusty Forma-ion of the Little Dal Group, or in the case of the younger ∼750 Mahundercloud Formation of the Coates Lake Group, likely haveounger overprints (Powell et al., 1993). The overlying Rapitanroup also yielded stable magnetic directions (Park, 1997), but thisnit is thought to be Franklinian (<720 Ma; Macdonald et al., 2010),nd not comparable paleomagnetically to the Gunbarrel event. Theemainder of the Gunbarrel units having paleomagnetic data yieldairly consistent results, even though they are distributed over2000 km2 (Fig. 1), defining a well-averaged paleomagnetic pole

hat indicates a large portion of Laurentia was in a near equatorialosition at 780 Ma (Harlan et al., 2008).

. New petrological results

Herein, mineralogical, lithogeochemical and isotopic data forhe Tsezotene sills and dykes, the Little Dal basalts, the Faber, Mar-aret and Calder sheets and the Hardisty dyke are examined. Manyf these data are presented in Sandeman et al. (2007) and Ootest al. (2008), however, additional whole-rock, mineral chemical andd isotopic data are now incorporated, including Nd isotope data

rom one sample of the Christmas Lake dyke of the Wyoming cra-

on. We also include the robust data of Ernst and Buchan (2010)or the Calder sheet and four samples of “unnamed” dykes exposedn Wopmay orogen. These dyke samples appear to spatially corre-ate with what we have termed the Hardisty dyke, which is most

esearch 252 (2014) 166–179

prominent near Hardisty Lake. Below we summarize our presentstate of knowledge on the mineral chemistry of the Gunbarrel rocks,emphasizing new observations on the well-preserved chill marginsamples of the Faber sheet (Sandeman et al., 2007). All mineralchemical and lithogeochemical data are presented in supplemen-tary data files available for download from the journal website.

Supplementary material related to this article can be found,in the online version, at http://dx.doi.org/10.1016/j.precamres.2014.07.007.

3.1. Petrographic and mineralogical data

Medium- to coarse-grained Tsezotene and Hottah rocks aredominated by equigranular, interlocking, bladed subhedral plagio-clase (0.5–4 mm long axis), subhedral to anhedral clinopyroxene(0.25–3 mm) and subhedral ilmenite grains (<0.1–1 mm – generally∼0.25 mm; Fig. 2A and B). In Tsezotene sills, magnetite replacementof ilmenite is more common than in the Hottah samples. Clinopy-roxene in the Tsezotene samples comprises Mg-rich augite coresthat range to ferroaugite rims (Fig. 3). Clinopyroxene cores in theHottah rocks are dominantly augite whereas rims are FeO-rich andare ferroaugite and locally subcalcic ferroaugite (Fig. 3). Pigeoniteoccurs in the core of one clinopyroxene from a Hottah chill marginsample.

Ilmenite is abundant in the Tsezotene samples and containsequal proportions of Fe and Ti, with little to no Cr, V, or Al. Gen-erally, ilmenite crystals occur as symplectite-like intergrowthswith augite, possibly reminiscent of trapped melt (see Ooteset al., 2008). Like the Tsezotene rocks, ilmenite in the Hottahsheets contains roughly equal proportions of FeO (47.0–54.5 wt.%)and TiO2 (39.1–50.5 wt.%), with minor Cr2O3 (≤0.25 wt.%), V2O5(0.29–0.73 wt.%) and Al2O3 (≤ 1.00 wt.%). Magnetite intergrownwith ilmenite is significantly less titanian (9.9–14.6 wt.% TiO2) andcontains minor Cr2O3 (0.07–0.28 wt.%) and variable but elevatedV2O5 (1.12–2.46 wt.%).

Plagioclase in holocrystalline Hottah sheets exhibits a widerange in composition from An59 in cores to An24 in rims (Fig. 4A),but clear compositional zoning or resorption/dissolution surfaceswere not noted in thin section. In the Tseozotene samples, pla-gioclase exhibited minor optical zoning only in chilled margins(Fig. 4B), that ranged from An59 in cores to An67 in rims. Incontrast to the holocrystalline rocks, chill margin samples fromthe Faber sheet (Fig. 2C) comprise fine-grained (≤1.5 mm), raggedand resorbed plagioclase and clinopyroxene phenocrysts as wellas clinopyroxene–plagioclase glomerocrysts, with rare accom-panying opaque grains (≤0.5 mm). These are set in an aphanitic(≤120 �m) matrix of plagioclase lathes, anhedral clinopyroxeneand subhedral Fe–Ti oxides. The phenocrysts and glomerocrystscollectively comprise ≤15% of the rock (Sandeman et al., 2007)indicating that the magma was weakly porphyritic when emplaced(∼85% melt). Plagioclase and clinopyroxene phenocrysts, bothin glomeroporphyritic clumps and as isolated grains, exhibitragged grain margins indicating late resorption/dissolution of thecrystals. Internally, the clinopyroxene grains show little variationin thin section, although diffuse, brighter Fe-rich rims are typicalin backscatter imagery. In clinopyroxene, abrupt compositionalshifts from augite to Ca-enstatite and then back to augite, suggestsignificant changes in the thermochemical evolution of the hostmagmas. Plagioclase grains in the Hottah chill margins (Fig. 4C–F)exhibit apparently simple, restricted ranges in composition. How-ever, close examination reveals very complex growth-resorptionfeatures as representatively demonstrated in Fig. 5 for plagioclase 1

(sample 06HS8023B). The plagioclase grains from the chill marginhave compositions that are typically more calcic than those in theholocrystalline samples of the Faber sheet, ranging from An67.0 toAn52.5 (Figs. 4 and 5). Two of four examined phenocrysts were noted

H.A. Sandeman et al. / Precambrian Research 252 (2014) 166–179 169

Fig. 2. Representative photomicrographs of: (A) a medium-to-coarse-grained example of the Faber sheet (06HS7017); (B) a medium-grained Tsezotene sill (07LO4b); and( clinopr

tstdcc

E

Fp

C) chill margin sample of the Faber sill (06HS8023B). Key: Act – actinolite; Cpx –

ock.

o have at least four, and possibly up to seven, distinct internalurfaces that cross-cut oscillatory zoning as well as patchy extinc-ion zones that lack oscillatory zoning (Fig. 5). Electron microprobe

ata for plagioclase along selected grain traverses reveal complexompositional patterns. These distinct optical discontinuitiesorrespond to abrupt, as well as gradual shifts in anorthite content

1200°C 1000°C

Wo50

n Fs

Augite Ferroaugite

SubcalcicFerroaugite

Pigeonite

SubcalcicAugite

Tsezotene sills

FS Cpx-1

FS Cpx-2FS Cpx-3

FS Cpx-4 06SB1160

06HS7018

06HS7017

06HS7019

06HS7016

Fabe

r she

et c

oars

e-gr

aine

d sa

mpl

es

Fabe

r she

et c

hill

phen

ocry

sts

KEY:

ig. 3. Clinopyroxene compositions from Tsezotene and Hottah samples in theyroxene quadrilateral.

yroxene; Ilm-Mt – ilmenite–magnetite; Pl – plagioclase; host – marginal country

(Fig. 5). The compositional zoning in these optically defined regionscomprises “reverse” oscillatory zoning, corresponding to steadilyincreasing An content (Fig. 5: zone 1), whereas other optical zonesexhibit “normal”, but very rapid decreases in An content towardsthe outer parts of the zone (Fig. 5: zones 2 and 4). A third variety ofcompositional variation exhibits “spikey” patterns having abrupt,relatively large (up to 7 An units) compositional shifts (Fig. 5: zones5–8) relative to less jagged, both “normal” and “reverse” oscillatoryzoning variations (up to 4 An units). These jagged patterns typicallyoccur at the margins of the phenocrysts, perhaps indicating anincrease in the frequency of recharge by more primitive melts.

3.2. Lithogeochemical data

The new geochemical data compilation for the Gunbarrel suiterocks is presented in the Supplementary data. The compilationincludes data for the Faber sheet (Sandeman et al., 2007), theTsezotene sills (Ootes et al., 2008) and also new lithogeochemicaldata for five Little Dal basalts, four samples of the Hardisty dyke,four specimens of the Margaret sheet as well as new radiogenicSm–Nd isotopic data for five Tsezotene sills, one Tsezotene dyke,and eight samples of the Faber sheet. A single sample from theChristmas Lake dyke has been analyzed for its Sm–Nd radiogenicisotopic composition. This sample lacks a corresponding lithogeo-chemical analysis but was collected from the same site as sample

BT-33 in Condie et al. (1969). These data are, where appropriate,supplemented by data from the Tsezotene sills and the associatedamygdaloidal Little Dal basalts in the Mackenzie Mountains (Dudásand Lustwerk, 1997), the Calder sheet, Hardisty dyke, and Faber

170 H.A. Sandeman et al. / Precambrian Research 252 (2014) 166–179

B

An70

Or10

An50

Pl-4

An70An20

Or10

06HS7019 -

06SB1160 - A

Pl-2 Pl-3

Pl-1Pl-Tsezotene

F

D E

Czoning zoning

zoning

zoningzoning

zoning

F rom h( in sam

sBri2t

i2

3

zaeatIibesdtepobp

ig. 4. Ab–Or–An diagrams for plagioclase from Gunbarrel rocks. (A) plagioclase fC)–(F) core to rim analyses across four individual plagioclase grains in a chill marg

heet in Wopmay orogen (Perrier, 1988; Stoffers, 2005; Ernst anduchan, 2010). For the purposes of this contribution, only the moreobust and precise, modern data is emphasized in petrogeneticnterpretations (Stoffers, 2005; Sandeman et al., 2007; Ootes et al.,008; Ernst and Buchan, 2010). Salient petrochemical aspects ofhe Gunbarrel rocks are summarized in Table 1.

Supplementary material related to this article can be found,n the online version, at http://dx.doi.org/10.1016/j.precamres.014.07.007.

.2.1. Comments on element mobility and alterationThe subaerial Little Dal basalts, two samples of fine-grained Tse-

otene sills from the Keele River area (06CL42A, 06CL43A; Fig. 1)nd, one fine-grained quenched offshoot of the Hardisty dykexhibit significantly more variability in their Al2O3, SiO2, K2O, CaOnd Na2O and large ion lithophile elements (LILE: Rb, Sr, Ba, Cs)han the thicker, holocrystalline Tsezotene sills and Hottah sheets.n conjunction with their strongly altered petrographic character,ncluding an abundance of chlorite–calcite amygdales, the Little Dalasalts are extensively altered, resulting in mobilization of theselements (Dudás and Lustwerk, 1997). This alteration is empha-ized by the extreme variability of K2O and Al2O3 and the strongepletion of CaO at elevated Al2O3 (Table 1). The remainder ofhe sills and dykes preserve less variable abundances of the majorlements and LILE and appear to represent primary magmatic com-

ositions. The less mobile major and trace elements, along with allf the high field strength elements (HFSE), show more systematicehaviour for all samples and these provide the firmest basis foretrogenetic interpretation of the rocks.

olocrystalline Hottah sheets; (B) plagioclase from holocrystalline Tsezotene sills;ple of the Faber sheet (06HS8023B). Note the change of scale between A and B–F.

3.2.2. Rock classification and major and trace element variationsConventional major and mobile trace element classification

plots may yield spurious results when applied to altered igneousrocks. Gunbarrel samples are basalts and basaltic andesites, but themajority of the Little Dal basalts have elevated K2O and are trachy-basalts in terms of their total alkalies versus SiO2 contents (notshown, LeBas et al., 1986). The immobile trace element classifica-tion plot of Pearce (1996) demonstrates that all of the Gunbarrelsamples are basalts with the exception of two specimens fromDudás and Lustwerk (1997) that have anomalously high Nb relativeto all samples (Fig. 6) and plot as alkali basalts. The samples exhibitstrong FeOT and TiO2 enrichment trends and low Nb/Y ratios indi-cating that they are subalkaline, tholeiitic basalts. Their tholeiiticnature is further outlined in Fig. 7A where the vast majority plot inthe tholeiite rather than the alkali basalt field. They are continentaltholeiites varying to ocean island basalt in the MgO–FeOT–Al2O3discrimination diagram of Pearce et al. (1977, not shown) and arecontinental basalts in the La–Y–Nb tectonic discrimination diagram(Fig. 7B) of Cabanis and Lecolle (1989).

Selected major, compatible and incompatible trace elements forall Gunbarrel samples are plotted (Fig. 8) against Mg# (molecular[MgO/MgO + FeOT] × 100). Rocks from each of the distinct locali-ties define restricted clusters or arrays in major and trace elementspace, with the exception of the Faber sheet and the two Tse-zotene dykes from the Keele River area. The two Tsezotene dykesare distinctly more primitive (high Mg#) and alkaline (elevated

TiO2, P2O5, Nb) compared to all other Gunbarrel rocks. Relativeto normal mid-ocean ridge basalts (NMORB Lehnert et al., 2000;http://www.petdb.org/; Sun and McDonough, 1989), all rockshave relatively low Mg# (21–52), MgO (1.62–8.45 wt.%) and CaO

H.A

. Sandem

an et

al. /

Precambrian

Research

252 (2014)

166–179

Table 1Field units, petrochemical subdivisions and salient petrochemical indices of Gunbarrel event rocks of western North America.

Unit n Mg# TiO2 (wt.%) (La/Yb)CN (La/Sm)CN (Gd/Yb)CN Eu/Eu* (Th/Nb)CN (Th/La)CN

∑REE �Ndt t = 780 Ma

Hottah sheetsFaber Sill n = 19 20.8–36.2 2.33–3.41 2.47–2.82 1.61–1.82 1.20–1.38 0.86–1.03 2.40–3.61 1.58–2.06 107.4–181.6 1.3–1.7Margaret Sill n = 4 30.3–33.2 2.82–3.03 2.56–2.79 1.59–1.75 1.26–1.46 0.82–0.96 2.36–2.69 1.62–1.80 131.1–155.9 ndHardisty dyke n = 8 29.0–39.0 2.31–2.92 3.17–3.59 1.79–2.14 1.43–1.52 0.80–0.88 2.16–2.79 1.75–1.88 149.1–198.9 ndCalder Sill n = 9 13.6–46.3 2.03–2.80 2.24–2.29 1.36–1.38 1.43–1.46 0.87–0.89 2.05 1.58 120.7–123.3 nd

Tzesotene sillsArctic Red River n = 21 32.1–38.2 2.40–2.74 2.67–3.47 1.62–2.04 1.36–1.64 0.75–0.94 2.61–3.14 1.77–2.16 131.7–156.7 1.4–1.6ARR dyke n = 1 30.5 3.14 4.1 1.94 1.65 0.82 2.58 1.93 230 1.5Keele River n = 5 27.1–36.0 2.36–3.73 2.65–3.46 1.67–1.96 1.37–1.51 0.76–0.89 1.91–2.30 1.69–2.02 142.3–177.7 ndKeele Dyke n = 2 46.3–52.4 2.78–2.87 3.49–4.18 2.01–2.17 1.42–1.50 0.85–1.02 1.88–1.90 1.74–1.91 136.5–145.7 nd

Little dal basaltsGP-A n = 3 26.5–37.4 2.40–2.67 3.85–4.91 1.73–2.96 nd nd nd 1.29–2.20 nd 1.1–1.5GP-B-C n = 11 31.9–48.1 1.76–2.25 2.20–4.15 1.34–2.19 1.11–1.57 0.80–1.08 2.25–3.27 1.14–2.69 93.0–109.9 1.5–1.6GP-G n = 2 41.1–44.1 1.87–2.02 2.64–2.66 1.81–2.17 nd nd nd 1.42–1.54 nd 0.2Christmas Lake dyke n = 1 32.5 2.94 nd nd nd nd nd nd nd 1.5

Note: n = number of samples; Mg# = [molecular MgO/(MgO + FeOT)] × 100; CN = chondrite normalized (after Sun and McDonough, 1989); �Nd calculated. According to the model of DePaolo (1981a); nd = not determined.

Fig. 5.

(A)

Photom

icrograph

of a

pla

the

Faber sh

eet. N

ote th

e op

tical zo

8 are

discu

ssed in

text; (B

) p

lagioclath

e p

hen

ocryst sh

own

above (c–r).

N7

resorption

-dissolu

tion su

rfaces an

dth

e p

lagioclase.

(6.24–9.58 w

t.%),

low bu

t va

erate to

high

TiO2

(1.76–3.7(10.9–20.8

wt.%

) an

d K

2 O (0.36–5.44

wt.%

). Th

e rem

ainin

g m

ajorelem

ents

SiO2 ,

Na

2 O an

d P

2 O5

and

volatile con

tents

(LOI)

are sim

i-lar

to th

ose for

NM

OR

B or

EMO

RB

(enrich

ed m

id-ocean

ridge

basalt;Su

n an

d M

cDon

ough

, 1989).

In gen

eral th

e H

ottah sh

eets exten

d to

.01

.1

11

0

.01 .1 1

Zr/TiO2

Nb

/Y

Basalt

And esite/B

asalt

Rhyolite/D

acite

Alkali B

asalt

Trachy-A

ndesite

Phonolite

Tephri-phonolite

Foidite

Wopm

ayF

ab

er s

he

et

Ca

lde

r sh

ee

t

Ma

rga

ret s

he

et

Ha

rdis

ty d

yke

CordilleraA

rctic

Re

d R

ive

r

Ke

ele

Riv

er

Little

Da

l ba

sa

lt

Arc

tic R

ed

dyke

Ke

ele

Riv

er d

yke

Fig. 6.

Zr/TiO2

versus

Nb/Y

classification

diagram

after Pearce

(1996). O

pen

symbols

are d

ata of

Du

dás

and

Lustw

erk (1997)

and

Perrier (1988).

gioclase p

hen

ocrystn

ation in

the

grain.

Zse

An

values

for a

coote

the

arrows

ind

ic th

e corresp

ond

ing

v

riable A

l2 O3

(103

wt.%

) an

d,

high

171

from th

e ch

ill m

argin of

ones

labelled 1

throu

ghre

to rim

traverse across

ating

at least

4 an

d u

p to

ariation in

An

conten

t of

.7–15.2 w

t.%),

mod

- bu

t variable

FeOT

172 H.A. Sandeman et al. / Precambrian R

B

A

00..0000 00..0055 00..1100 00..115500

11

22

33

44

TiO

2

4Zr/(P O *10 )2 5

AAllkkaalliibbaassaalltt

TThhoolleeiiiittiiccbbaassaalltt

La/10 Nb/8

Y/15

CCaallcc--aallkkaallii CCoonnttiinneennttaall

AAllkkaalliinneeiinntteerrccoonnttiinneennttaall

rriiffttss

BBaacckk aarrccbbaassiinn

VVAATT

NNMMOORRBB

EEMMOORRBB

WopmayFaber sheet

Calder sheet

Margaret sheet

Hardisty dyke

CordilleraArctic Red River

Keele River

Little Dal basalt

Arctic Red dyke

Keele River dyke

Fig. 7. (A) TiO2 versus Zr/(P2O5 × 104) discrimination diagram (Winchester andFloyd, 1977) demonstrating that the majority of the Gunbarrel rocks are moderate-to-high-TiO2 tholeiitic, transitional to alkali basalts. (B) Ternary La/10-Y/15-Nb/8pt

hs(sdcLel

s

western coast of present-day North America (Fig. 1; Park et al.,

aleotectonic discrimination diagram (Cabanis and Lecolle, 1989) showing the con-inental basaltic compositions of the Gunbarrel rocks.

igher FeOT and slightly lower Mg# than the Tsezotene sills. Allamples exhibit relatively uniform and low Cr (8–109 ppm) and Ni3–82 ppm), with Co (25–278 ppm) and Sc (27–46 ppm) contentsimilar to NMORB (Fig. 8). Vanadium abundances in the Faber sillecrease dramatically with fractionation (V = 717–172 ppm) indi-ating clinopyroxene and/or ilmenite-magnetite fractionation. Theittle Dal basalts and the Keele River dykes are the most primitivexamples of Gunbarrel magmas, having the highest Mg#’s and the

owest Zr and total rare earth element (REE) concentrations.

The incompatible trace and REE abundances of the Gunbarrelamples similarly fall within narrow ranges for each suite (Table 1;

esearch 252 (2014) 166–179

Figs. 8 and 9) and all are significantly higher than average NMORBwith the exception of Sr. Regardless of field unit or location, therocks exhibit remarkable coherency in their incompatible trace ele-ment geochemistry, including the most mobile elements (Sr, Ba,and Rb). The multi-element profiles exhibit negative slopes withnotable Sr and P troughs, variably developed minor negative Ba, Ta,Nb, and Zr–Hf anomalies and prominent Th enrichment relative toLa ([Th/La]CN = 1.57–2.16). The rocks are mildly light-REE enriched([La/Yb]CN = 2.24–4.18) and have a modest negative slope frommiddle- to the heavy-REE ([Gd/Yb]CN = 1.11–1.65). With the excep-tion of the Little Dal basalts which have minor positive Eu anomalies([Eu/Eu*] = 1.11–1.29, mean = 1.20), the remainder exhibit minornegative Eu anomalies ([Eu/Eu*] = 0.76–1.03, mean = 0.88). The Tse-zotene sills from the Keele River area differ systematically fromthose exposed at the Arctic Red area (Figs. 1 and 9A). The lone Tse-zotene dyke from the Arctic Red River area has higher abundancesof incompatible trace elements, a modest negative Ti anomaly andconcentrations of Sr, Yb, and Lu comparable to the Arctic Red sills. Itis light-REE enriched [La/Yb]CN (4.10) relative to all samples, but hascomparable Eu/Eu* (0.82) to those sills. The Little Dal basalts havethe lowest abundances of all of the incompatible trace elements, buthave similar multi-element profiles (Table 1 and Fig. 9B) that paral-lel those of the Faber sheet. Each distinct Hottah intrusion appearsto exhibit notable, minor but systematic differences in composition(Fig. 9C). The Margaret sheet overlaps extensively with the fieldfor the Faber sheet whereas the Hardisty dyke is slightly enrichedin the light-REE and LILE relative to the remainder of the Hottahsheets. The Hardisty dyke has incompatible trace element profilesand abundances essentially identical to the Tsezotene sills from theKeele River area (Fig. 9).

3.2.3. Nd isotopic dataFifteen specimens were analyzed for their Samarium–

Neodymium isotopic compositions (Table 2) including: fourmedium-grained Tsezotene sills; one Tsezotene sill chill margin;one Tsezotene dyke; seven medium-grained Faber sheet samples;one Faber sheet chill margin and; a sample from the ChristmasLake dyke of Wyoming. The data are presented as time corrected(t = 780 Ma) 143Nd/144Nd ratios and are also expressed as �Ndt

values (Fig. 10). These are supplemented with the radiogenicdata of Dudás and Lustwerk (1997). All yield remarkably coherentpresent-day 144Nd/143Nd ratios ranging from 0.512474 to 0.512565and corresponding �Ndt of +1.3 to +1.7 (mean = +1.5: stdev = 0.1).All of the analyses overlap, within error (±0.5 �Nd units), arehigher than the bulk earth at 780 Ma, but are significantly lowerthan that for contemporaneous depleted mantle (�Ndt DM = +6.3:DePaolo, 1981a; Fig. 10). Although the rocks come from geologicalunits that are separated by up to ∼2500 km (Fig. 1), the �Ndt valuesdefine a remarkably homogeneous dataset (Fig. 10). 147Sm/144Ndvalues for the Gunbarrel rocks are >0.15, precluding the calculationof depleted mantle model ages.

4. Discussion

The ca. 780 Ma age of a number of widespread examples ofGunbarrel event rocks has been well constrained by precise U–Pbdating of baddeleyite (Harlan et al., 2003) and their latitudinal posi-tion during intrusion is constrained to near equatorial (Park et al.,1995a; Harlan et al., 2008). These data, along with the present dayorientations of the Gunbarrel dykes and sills suggest that they werecollectively fed from an upwelling centre (plume) located off the

1995b).The Hottah sheets exposed in the Wopmay orogen of the north-

west Canadian Shield are thicker than the Tsezotene sills exposed

H.A. Sandeman et al. / Precambrian Research 252 (2014) 166–179 173

00

11

22

33

44

55

66

Cs

00

1100

2200

3300

4400

Nb

00

11

22

33

44

55

66

77

88

99

Th

1100 2200 3300 4400 5500 660000

11

22

33

44

55

66

77

Yb

MMgg##

4400

4455

5500

5555

6600

00

11

22

33

44

TiO2SiO2

00..00

00..22

00..44

00..66

00..88 P O2 5

2200 3300 4400 5500 6600

MMgg##

00

110000

220000

330000

Ni

1100

00

220000

440000

660000

880000 V

2200 3300 4400 5500 6600

MMgg##

1100

NMORB

N

N

N

N

N

N

N

NN

O O

E

OE

OE

O

E

O

E

in-situ UCcrystal fractionation(Faber sill)

E

Fig. 8. Selected major and trace elements versus Mg# for Gunbarrel rocks. Shown for comparison are data for global, normal mid-ocean ridge basalts (dashed field: Lehnertet al., 2000: avg NMORB (N), avg EMORB (E), avg OIB (O) from Sun and McDonough, 1989).

Table 2Nd isotopic data for 15 representative Gunbarrel rocks.

Sample Note Field unit Sm (ppm) Nd (ppm) 143Nd/144Ndpresent-day

147Sm/144Ndmeasured

143Nd/144Ndinitial

uncertainty (±) � Nd (CHUR)T TDMa (Ma)

07LO2a Sill Arctic Red River 6.94 26.10 0.512529 0.1608 0.511707 0.000014 1.5 147207LO2b Sill Arctic Red River 6.88 25.87 0.512529 0.1608 0.511706 0.000014 1.5 147507LO2d Sill Arctic Red River 7.26 27.39 0.512533 0.1603 0.511713 0.000014 1.6 144807LO3a Dyke Arctic Red River 11.10 44.93 0.512474 0.1493 0.511711 0.000013 1.5 134307LO6d Sill Arctic Red River 6.91 25.97 0.512524 0.1608 0.511702 0.000014 1.4 148707LO23a1 Chill Arctic Red River 7.29 27.63 0.512521 0.1595 0.511705 0.000014 1.4 146106HS7012 Sill Faber sheet 10.12 37.60 0.512549 0.1627 0.511717 0.000014 1.7 146906HS7018 Sill Faber sheet 5.79 21.01 0.512564 0.1665 0.511712 0.000014 1.6 153806HS7019 Sill Faber sheet 7.54 27.97 0.512544 0.1629 0.511711 0.000014 1.5 149106HS8023b Chill Faber sheet 5.80 21.09 0.512565 0.1661 0.511716 0.000014 1.6 152204AS5003 Sill Faber sheet 5.26 19.07 0.512549 0.1667 0.511696 0.000016 1.3 159204AS5003B Sill Faber sheet 8.22 30.80 0.512540 0.1613 0.511715 0.000014 1.6 145604AS5004 Sill Faber sheet 6.76 24.82 0.512554 0.1647 0.511712 0.000014 1.6 151204AS5010 Sill Faber sheet 7.91 29.45 0.512548 0.1624 0.511717 0.000015 1.7 1463T09-BT15 Dyke Christmas Lake 7.96 30.77 0.512510 0.1564 0.511710 0.000008 1.5 1413

Note: CHUR – 143Nd/144Nd = 0.512638: CHUR – 147Sm/144Nd = 0.1967: � = 0.00000000000654.a After DePaolo (1981a).

174 H.A. Sandeman et al. / Precambrian Research 252 (2014) 166–179

Ro

ck

/NM

OR

B

SrRb

BaTh

TaNb

LaCe

PrP

NdZr

HfSm

Eu Ti

GdTb

DyHo

YYb

Lu

Ro

ck

/NM

OR

BR

oc

k/N

MO

RB

Faber sheet

AR dyke

KR sills

KR dykes

1

10

100

1

10

100

1

10

100

A

B

C

Tsezotene sills

Little Dal basalts

Hottah sheets

Calder sheet

Margaret sheet

Hardisty dyke

Faber sheet

ARsills

OIB

EMORB

LC

Fig. 9. NMORB-normalized (Sun and McDonough, 1989) multi-element plots for theGunbarrel rocks. (A) A field for 19 samples of Tsezotene sills from Arctic Red Riverarea (AR sills) compared to the Arctic Red River dyke and, sills and dykes from theKeele River area (KR). (B) Little Dal basalts compared to a field for the Faber Sill aswell as representative EMORB, OIB and lower crust (LC: Rudnick and Gao, 2003). (C)A field for the Faber sheet (n = 17) is compared to the Margaret sheet, the Hardistyd

igimomhetMzmrha

DM UDM

RL

CLD

0.10 0.15 0.20 0.25 0.30 0.35-10

-8

-6

-4

-2

0

2

4

6

8

ε Nd t=

780 M

a

147 144Sm/ Nd

CHUR

2σ

Faber sheet

Tsezotene sill (ARR)

Little Dal basalt

Tsezotene dyke (ARR)

Archean Slave and Proterozoic Wopmay crust (-11.5 to -29.6)

Fig. 10. �Nd780Ma versus 147Sm/144Nd diagram for Gunbarrel rocks. DM–depletedmantle field; UDM – ultradepleted mantle field; RL – recycled lithosphere field;CHUR – chondritic uniform reservoir; CLD – Christmas Lake dyke. Open symbols arehistorical data of Dudás and Lustwerk (1997). Note that the granitoids of the west-ern Slave craton and the Paleoproterozoic Wopmay orogen yield highly negative

experienced repeated thermochemical replenishment accompa-

yke and the Calder sheet.

n the Mackenzie Mountains of the North American Cordillera, sug-esting that the former may have experienced more substantialn situ crystal fractionation. Both sets of sills have porphyritic chill

argins that, in the case of the Hottah sheets, provide evidencef their pre-emplacement, weakly porphyritic character, extensiveineral textural and chemical variation and rapid quenching. In the

olocrystalline portions of all sills, both plagioclase and clinopyrox-ne are unremarkable in terms of internal zoning and resorptionextures and show rim ward decreases in anorthite content and

g#, respectively. Plagioclase phenocryst compositions in a Tse-otene sill chill margin exhibit a similar range to those in the chillargin of the Faber sheet and show clear, reverse, core (An59)-to-

im (An67) zoning (Ootes et al., 2008). Plagioclase grains in the thick,

olocrystalline portions of the Faber and Margaret sheets preserve

much wider range in composition varying from An59.2 in grain

�Nd(t=780Ma) values of −11.5 to −29.6.

Data from Yamashita et al. (1999) and Bowring and Podosek (1989).

cores to An23.6 in rims (Fig. 4) indicating that the Hottah sheets aregenerally more evolved than the Tsezotene sills.

The chill margin of the Faber sheet preserves sparse (≤15 vol.%)clinopyroxene and plagioclase phenocrysts (≤1.5 mm) and glom-erocrysts in a very-fine-grained microlitic groundmass of pla-gioclase + clinopyroxene + ilmenite–magnetite. These chill marginplagioclase phenocrysts preserve incontrovertible textural andchemical evidence of abrupt and substantial variations in internalcomposition. Many grains exhibit well-developed internal oscilla-tory zoning (normal and reverse), and also exhibit distinct, internalcross-cutting surfaces (Fig. 5). Optical observations and detailedelectron microprobe traverses across four chill margin plagioclasegrains indicates complex growth-resorption histories, presumablycaused by repeated thermochemical perturbation of the phe-nocrysts in the host magma. Changes in An values across the grainsindicate that they were abruptly removed from equilibrium condi-tions and underwent resorption/dissolution followed by renewedgrowth on at least four and likely greater than eight occasionsduring their magmatic residence. Abrupt increases in An valuesacross optical discontinuities indicate episodes of recharge of themagma chamber by hotter, but chemically similar less-fractionatedbasaltic magmas. In contrast, low amplitude, steady increase and/ordecrease in the An content of the plagioclase reflects periods of pro-gressive oscillatory growth zoning in quasi-equilibrium settings.Numerous resorption-dissolution and growth surfaces indicatenumerous episodes of vigorous and repetitive recharge and mixingof thermochemically similar batches of basaltic magma. Resorbed-dissolved rims of phenocrystic clinopyroxene and plagioclase in theaphanitic matrix indicate that these grains were not in equilibriumwith their final host magma and are in fact antecrysts. The liq-uidus phase assemblage of clinopyroxene + plagioclase + ilmenitealong with their quartz normative compositions indicates thatthese are basaltic magmas that ultimately equilibrated at rela-tively shallow lithospheric depths where plagioclase is stable. TheGunbarrel magmas were, therefore, hot, dry, reduced, weakly por-phyritic mafic magmas derived from at least one large and possiblymany, turbulently convecting dynamic magma chamber(s) that

nying upwelling of deeper, mantle-derived magmas The primarystaging chamber(s) are speculated to have been located at the

H.A. Sandeman et al. / Precambrian R

1100

110000

SSrrRRbb

BBaaTThh

TTaaNNbb

LLaaCCee

PPrrPP

NNddZZrr

HHffSSmm

EEuuTTii

GGddTTbb

DDyyHHoo

YYYYbb

LLuu

11

Ro

ck

/NM

OR

B

RRaayylleeiigghh ffrraaccttiioonnaallccrryyssttaalllliizzaattiioonn

50%

10%

SSoouurrccee::0066HHSS77001188MMgg##==3366..22

MMoosstt eevvoollvveedd::0066HHSS77001122MMgg##==2200..88

Fig. 11. Calculated model liquids for Rayleigh fractional crystallization of the Fabersheet using the most primitive sample 06HS7018 as the source. Fine dashed blacklines are liquid compositions representing 10–50% crystallization. Mineral modesat(

MiTaphbl

4

gtPaepNrNRlLtE

4

iatcwapartcltt

re 2% olivine – 36% clinopyroxene – 60% plagioclase – 2% magnetite. Applied dis-ribution coefficients are from Rollinson (1993) and the GERM Partition CoefficientKd) Database (http://earthref.org/KDD/e:15/?&sort=mineral).

ohorovicic discontinuity (MOHO; Herzberg et al., 1983) as thiss likely the most prominent density filter in the lithosphere.hese chambers would have rapidly developed and evolved over

central asthenospheric upwelling locus (plume?) located west ofresent-day North America (Fig. 1). The derivative magmas mayave been staged in intermediary, shallow level magma chamber(s)ut were ultimately injected into their present-day, relatively shal-

ow crustal exposure levels.

.1. Implications of lithogeochemistry

The Gunbarrel magmas are subalkaline tholeiitic basalts andabbros that have hallmarks of both low- and high-TiO2 continen-al tholeiites (Turner and Hawkesworth, 1995; Gibson et al., 1996;eate and Hawkesworth, 1996; Farmer, 2003). Relative to NMORB,ll Gunbarrel rocks have low Mg#’s and appear to representvolved, FeOT-, K2O- and TiO2-rich, relatively CaO- and Al2O3-oor basaltic rocks that exhibit uniform and low Cr (8–70 ppm),i (<20–60 ppm), but high contents of the LILE and REE. These

ocks are, therefore, not primary melts of a shallow, depleted,MORB-like asthenospheric mantle (e.g., Roeder and Emslie, 1970;ingwood, 1975). Instead, their elemental abundances and interre-

ationships indicate that, with the exception of enrichment of theILE and depletion of the HFSE (Ti, P, Nb, Ta, Zr, Hf), the most primi-ive examples of the Gunbarrel rocks most closely resemble evolvedMORB (Fig. 9).

.1.1. Upper crustal emplacement and fractionationMineral modes and chemistry, along with bulk-rock chem-

cal constraints indicate that fractional crystallization of thessemblage plagioclase + clinopyroxene + ilmenite ± olivine con-rolled the final petrological evolution of the Hottah sheets. In thease of the Tsezotene sills, in situ upper crustal crystal fractionationas minimal, as their mineral chemical variations are restricted,

nd their bulk compositions form tight clusters in all geochemicallots (Figs. 6–9). The Hottah sheets are typically thicker intrusionsnd the most extensively sampled Faber sheet exhibits the widestange in composition (Figs. 6–9). Cooling of the Hottah sheets inhe upper crust was, therefore, accompanied by more substantial

rystallization. In situ equilibrium and Rayleigh fractional crystal-ization models for the Faber sill was tested using the incompatiblerace elements (Fig. 11). The modelled liquid compositions for bothypes of crystallization roughly match those of the Faber sill and

esearch 252 (2014) 166–179 175

are not greatly different. The most primitive example of the Fabersill (06HS7018) was used as a starting composition. A fractionat-ing mode of 60% plagioclase: 36% clinopyroxene: 2% olivine: 2%magnetite was assumed and the distribution coefficients given inRollinson (1993) and the GERM Partition Coefficient (Kd) Database(http://earthref.org/KDD/e:15/?&sort=mineral) were applied. Themodel trace element patterns determined for liquids after 10, 20,30, 40 and 50% Rayleigh fractional crystallization are shown inFig. 11. The modelled liquids diverge slightly from the natural rocks,particularly for Th, Eu, Ti and the HREE. The pattern for the mostevolved specimen of the Faber sheet (06HS7012) can be approx-imated by ∼50% Rayleigh fractional crystallization of the startingcomposition (06HS7018: Fig. 11). This suggests that the most prim-itive example of each distinct suite of Gunbarrel rocks is a closeestimate of the primary magma composition released from deepcrustal/upper mantle magma chambers, into their present settingin the upper crust.

4.1.2. Constraints on the origin of Gunbarrel magmasModelling of the possible sources of continental basalts is dif-

ficult for many reasons, not least of which is the diversity ofpossible contaminants that the primary mantle-sourced magmasmay interact with upon their ascent. Moreover, the involvementof the continental lithosphere (crust or mantle) in mantle-derivedmafic magmatism is challenging to assess, owing to our limitedknowledge of its structure, composition, and thermal charac-ter, information derived mainly via geophysical observations andsampling and analysis of rare mantle xenoliths and xenocrystsentrained in alkali basalts and kimberlites. As the magmatic source(plume?) of the Gunbarrel magmas is no longer preserved, we haveto rely on incompatible trace element ratios from the Gunbarrelrocks to help determine the nature of the mantle source(s) forthese mafic igneous rocks (e.g., Condie, 1997; Pearce, 2008). Here,the ratios of Th/Ta, Th/Nb and Ta/Yb are used as proxies for litho-spheric contributions to mafic magmas. Four trace element ratiodiagrams are discussed in order to help constrain the origin of theGunbarrel magmas (Fig. 12) and help determine the relative role ofpossible trace element contributions from a number of recognizedmantle and crustal end-members that may be involved in the gen-esis of basaltic suites. A Th/Ta versus La/Yb diagram (modified afterCondie, 1997; Fig. 12A) demonstrates that the Gunbarrel rocks havehigh Th/Ta ratios, lying well above those compositions defined byglobally recognized, geochemically defined asthenospheric mantlecomponents (Sun and McDonough, 1989). Gunbarrel rocks have,however, La/Yb ratios similar to those for enriched MORB mantle(EMM). Bulk continental crust (BC: Rudnick and Gao, 2003) has aTh/Ta ratio similar to those for the Gunbarrel rocks, however, upper(UC) and middle (MC) continental crust have higher Th/Ta. With theexception of lower crust (LC), the crustal values for La/Yb are signif-icantly higher than those for the Gunbarrel samples. The Gunbarrelrocks therefore appear to contain a Th-enriched and/or Ta-depletedlithospheric component and cannot be easily derived from depletedMORB mantle (DMM), EMM, or, intraplate deep asthenosphericmantle (viz. ocean island basalt: OIB Fig. 12A). AFC of crustal mate-rial might account for some of the variations in the petrochemistryof the Gunbarrel rocks, however, their parental magmas must havestarted with Th/Ta elevated relative to that of DMM, EMM, or OIBmelts.

A plot of Th/Yb versus Ta/Yb (Pearce, 1982) underscores the ele-vated Th and moderate Ta/Yb of the Gunbarrel rocks (Fig. 12B),which lie above the mantle array. Again, they have significantlyhigher Th/Yb at equivalent Ta/Yb than asthenosphere-derived

basaltic rocks derived from DMM, EMM and OIB (Fig. 12B), but theydo plot directly above EMM melts (i.e., Th-enriched EMM). Theyhave similar Th/Yb but lower Ta/Yb ratios than lower continentalcrust (Fig. 12B; Rudnick and Gao, 2003). The Gunbarrel rocks exhibit

176 H.A. Sandeman et al. / Precambrian Research 252 (2014) 166–179

RFC

AFC(BC)

PM

DMM

UC

OIB

ASV

SSZ

CLM

MC

BC

LC

EMM

MORB

SHO

CA

CA

TH

oceanicarcs

continentalarcs

OIB

EMMPM

DMM

UC

ASV

SSZ

CLM

BCMC

LC

AFC(BC)

RFC

CLM

PM

EMM

OIB

LC

UC

DMM

ASV

SSZ

AFC(BC)

RFC

BC

OIB

DMM

EMM

PM

LC UC

ASV

SSZ

20%CLM

BC

AFC (BC)

RFC

A B

D

11 1100 110000 ..0011 ..11 11 1100TTaa//YYbb

11

1100

110000TT

hh//TT

aa

..0011

..11

11

1100

TThh

//YYbb

00..00 00..22 00..44 00..66 00..88 11..00 11..22 11..4400

11

22

33

44

55

66

77

LLaa//SS

mm

00..00 00..22 00..44 00..66 00..88 11..00

11

22

33

44

GGdd

//YYbb

C

LLaa//YYbb

TThh//NNbbTThh//NNbb

WopmayFaber sheet

Calder sheet

Margaret sheet

Hardisty dyke

CordilleraArctic Red River

Keele River

Little Dal basalt

Arctic Red dyke

Keele River dyke

LC

Fig. 12. Incompatible element plots for the Gunbarrel rocks. (A) Th/Ta vs. La/Yb (adapted after Condie, 1997). (B) Th/Yb vs. Ta/Yb (adapted from Pearce, 1982). (C) La/Smvs. Th/Nb. (D) Gd/Yb vs. Th/Nb. Dashed field is for the ensimatic Mariana Arc (Elliott et al., 1997) showing supra-subduction zone effects. Curves labelled AFC representtrajectories and envelopes for assimilation fractional crystallization (DePaolo, 1981b) of bulk continental crust (BC), applying r-values of 0.1, 1.0 and 2, and using sample06HS7018 as the parent and bulk continental crust as the assimilant. Dots represent F-values (F = mass assimilant/mass melt) from 0 to 5. Curves labelled RFC representRayleigh fractional crystallization of the assemblage Pl + Cpx + Ol + Mt (60:36:2:2) from sample 06HS7018. White dots represent % residual liquid remaining in gradations of10%. Solid lines joining DMM-OIB-CLM (calc-alkaline lamprophyre: Rock, 1991) are simple binary mixtures of these mantle end members. Shaded arrows show the effectsr enricho r crus

ca(ofitTttcrdf(wtocadFfao

esulting from asthenospheric source variations (ASV) and suprasubduction zone

cean island basalt; PM – primitive mantle (Sun and McDonough, 1989). UC – uppe

ompositions similar to continental arc basalts (Pearce, 2008)nd are comparable to many Proterozoic gabbroic dyke swarmsCondie, 1997). Fig. 12C shows the elevated La/Sm and Th/Nb ratiosf the Gunbarrel rocks and further emphasizes that they are distinctrom asthenosphere derived DMM, EMM, or OIB. They exhibit sim-lar elemental ratios, however, to the lower crust and similar La/Smo EMM. This figure also serves to highlight two distinct groups ofsezotene sills having mildly differing La/Sm ratios. It is unlikelyhat AFC of any average crustal composition can easily account forhe inter- or intra-suite variation in La/Sm. Moreover, these pro-esses cannot readily generate the variation observed in the Th/Nbatios, and in particular, the inversely correlated linear trends evi-ent for the Hardisty dyke versus the positively correlated trendor the Faber sheet (Fig. 12C). A further plot of Gd/Yb versus Th/NbFig. 12D) illustrates the modest enrichment of the MREE/HREE, asell as the strong Th/Nb enrichment of the Gunbarrel rocks rela-

ive to the asthenospheric array. The weakly elevated Gd/Yb valuesf the Gunbarrel rocks suggest that their parental magmas likelyontained a small proportion of melt derived from a deep, OIB-likesthenospheric mantle source (Pearce, 2008; Niu et al., 2011). Thisiagram also outlines two distinct magma types comprising the

aber sheet characterized by mildly differing Gd/Yb ratios. Rayleighractional crystallization cannot account for the range in the ratios,lthough AFC of average continental crust may account for somef these variations. Collectively these diagrams demonstrate that

ment (SSZ). DMM – depleted morb mantle; EMM – enriched morb mantle; OIB –t; MC – middle crust; BC – bulk crust; LC – lower crust (Rudnick and Gao, 2003).

Rayleigh fractional crystallization alone cannot fully account for thepetrological evolution of the Faber sheet, or any of the remainingGunbarrel rocks.

The Gunbarrel rocks must represent fractionated basalts derivedfrom more primitive mantle-derived melts that likely containeda high proportion of EMM, with a small proportion of a litho-spheric end-member having high Th and LREE/HREE. Such a Th-and LREE-enriched lithospheric component would most simply belower crust. Incorporation of this material would have occurredeither during ascent of the Gunbarrel magmas through the crustat the magmas origin, i.e., where the plume met the MOHO, orpossibly through contamination of the mantle source via earlierincorporation of metasomatized lithospheric material (subduction-derived?).

4.1.3. Implications of Nd isotopic dataThe Gunbarrel rocks are very homogeneous in terms of

their incompatible trace element and Nd isotopic compositions(Figs. 8–12 and Tables 1 and 2). The minor, but reproduciblevariations in the high quality data for sills intruded into threegeologically distinct terranes, the Neoproterozoic Mackenzie

Mountains Supergroup, the Paleoproterozoic-Archaean Wopmayorogen, and the Archaean Wyoming craton, indicate that theircontamination via assimilation of present-day local mid-uppercrustal materials is improbable. This is because the Archaean and

brian R

PosLntoaNc(tmiRpl

4

rsAtTteAfctdpbr

S(SaWCh2Niuoficewlch

5

oataR

H.A. Sandeman et al. / Precam

aleoproterozoic host rocks in the Slave craton and Wopmayrogen are dominated by mid-crustal, potassic granitoids withtrongly fractionated multi-element profiles, strongly elevatedILE, LREE and Gd/Yb, very prominent HFSE troughs and, stronglyon-radiogenic 143Nd/144Nd ratios at 780 Ma (Fig. 10). Such con-aminants would have more profoundly modified the compositionsf the Gunbarrel magmas, even with low volumes of assimilantnd low rates of assimilation. This is particularly evident in thed isotopic compositions of the Gunbarrel magmas. Deep crustalontamination above the reconstructed asthenospheric upwellingplume?) centre during staging of the Gunbarrel magmas prior toheir final emplacement is much more plausible. As that environ-

ent was located off western North America (current coordinates)t has either been moved to another continent during rifting ofodinia, or it destroyed during excessive Phanerozoic tectonic over-rinting. As such it is not possible to ascertain the nature of that

ower crust.

.2. Constraints for paleogeographic reconstructions

Park et al. (1995b) used paleogeographical data and matchedadiating dyke swarms in both western Laurentia (Gunbarrel) andouthern Australia (Gairdner) to recommend a connection withustralia at 780 Ma. Subsequently, Wingate et al. (1998) tested

his hypothesis by determining the age of a single Gairdner dyke.he ca. 830 Ma age (Wingate et al., 1998) is distinctly older thanhe 780 Ma age of the Gunbarrel event in North America (Harlant al., 2003) and, as no ca. 830 Ma events are recognized in Northmerica, the Australia-North America connection in Rodinia has

allen out of favour. McPhie et al. (2010), however, presented aa. 770 ± 10 Ma U–Pb LA-ICP-MS titanite age for a Gairdner dykehat intrudes the ca. 1590 Ma Olympic Dam iron-oxide copper goldeposit in South Australia (Creaser and Cooper, 1993). It remainsossible, therefore, that some of the “Gairdner dykes” may in facte Australian correlatives of the North American Gunbarrel eventocks.

Numerous 830–780 Ma mafic igneous ages are reported fromouth China (e.g., Zhou et al., 2002; Li et al., 2008). Li et al.2008) places Australia adjacent to Laurentia at 780 Ma, withouth China lying between. Evans (2009) departs from this model,nd using well-constrained apparent polar wander paths, placesest Africa adjacent to northwest Laurentia and removes South

hina and Australia from the connection entirely. East Antarcticaas been suggested as a Laurentia neighbour (Goodge et al.,008) and Siberia has been implicated as a neighbour in theeoproterozoic (Sears and Price, 2000); these are also removed

n the reconstruction of Evans (2009). Regardless of the cratonnder consideration (West Africa, South China, Australia, Siberiar Antarctica), our dataset provides a first pass lithogeochemicallter to further critically test reconstruction hypotheses. Futureomparison and correlation of widespread LIP assemblages, how-ver, must utilize lithogeochemical and tracer isotopic data alongith paleomagnetic and precise, high resolution U–Pb geochrono-

ogical studies. This contribution indicates that direct, pre-riftorrelatives of the Gunbarrel suite will similarly be regionallyomogenous.

. Conclusions

It is widely accepted that the generation and emplacementf the Gunbarrel magmas was in response to distal upwelling of

mantle plume associated with, and potentially responsible for,he protracted Late Proterozoic rifting of the western margin ofncestral North America and the break-up of the supercontinentodinia (Park et al., 1995b; Harlan et al., 2003). The mineralogy

esearch 252 (2014) 166–179 177

and mineral chemical data of the Gunbarrel rocks from numerouslocalities in western North America demonstrate that these are pla-gioclase + clinopyroxene + ilmenite–magnetite – bearing gabbrosand basalts. Chill margin samples from the Faber sheet and a Tse-zotene sill demonstrate that the parental magmas for these unitswere weakly plagioclase + clinopyroxene porphyritic (≤15% crys-tals) basaltic magmas at their time of emplacement into the uppercrust. These crystals represent antecrysts that formed earlier in thepetrogenetic history of these magmas, likely in deep seated, lowercrustal magma chambers. The plagioclase and clinopyroxene ante-crysts in the chill margins preserve many internal thermochemicalresorption–dissolution surfaces as well as their final crystal faces,which indicate the crystals were rarely in equilibrium with theirhost magma. The parental Gunbarrel magmas likely ponded at theMOHO in large stratiform magma chambers that were repeatedlyreplenished over short intervals during which dynamic, turbulentmixing produced remarkably homogenous magmas. Accompany-ing lithospheric extension, tapping of these large magma chambersresulted in the widespread emplacement of the Gunbarrel rocksacross northwest Laurentia.

Robust lithogeochemical and Nd isotopic data for these rocksdemonstrates remarkable compositional homogeneity. They aresubalkaline, tholeiitic continental basalts that have elevated LILE,LREE and HFSE abundances relative to NMORB. In situ upper crustalfractionation has apparently affected only the thicker, more volu-minous Hottah sheets of Wopmay orogen, whereas all other dykesand sills record minor lithogeochemical and mineral chemicalvariation. The Hottah sheets have experienced modest low-P frac-tional crystallization in the upper crust, whereas the Gunbarrelrocks of the Mackenzie Mountains (Tsezotene sills and Little Dalbasalts) represent compositionally discrete basaltic suites exhibit-ing only minor intersuite and intersample variation. Extendedmulti-element plots and incompatible trace element ratio dia-grams indicate that the most primitive Gunbarrel rocks are likelyfractionated EMORB-like basalts that contain a minor lithosphericcomponent. The lithospheric component must be somewhat olderthan the 780 Ma rocks, shifting their �Nd values to ∼1.5, slightlygreater than the 780 Ma bulk earth but significantly lower than con-temporaneous depleted mantle. Minor, pervasive negative HFSEanomalies suggest contamination by a relatively unevolved litho-spheric component. The modest La/Sm, and La/Yb but low Gd/Ybratios along with strong depletion in Sr but negligible Eu depletionsuggests that neither the source nor the lithospheric contaminantcontained garnet or plagioclase. These observations suggest a prim-itive, lower crust or lithospheric mantle contaminant with modestLILE and LREE enrichment, well-developed HFSE anomalies but lowGd/Yb ratios such as a pyroxenitic, lower crust – upper mantlecumulate residue. The new robust dataset for the Gunbarrel rockswill facilitate proposed correlations with other global mafic eventsat 780 Ma.

Acknowledgements

This research was supported by the Northwest Territories Geo-science Office through both the Diamond Portfolio of Indian andNorthern Affairs Canada and the Government of the NorthwestTerritories. Access to outcrops in the Mackenzie Mountains wasfacilitated by the Peel Plateau and Plain and Sekwi Mountainprojects and in the Wopmay orogen by the South Wopmay bedrockmapping project. We therefore thank project leaders AdrienneJones, Edith Martel and Valerie Jackson for aircraft and field sup-

port. Contributions from Chris Leslie and Yvon Lemieux made thisproject possible. An anonymous reviewer, Richard Ernst and editorRandall Parrish provided helpful reviews of the manuscript. This isNorthwest Territories Geoscience Office contribution #0062.

1 brian R

R

A

A

B

B

B

B

B

C

C

C

C

C

D

D

D

D

D

E

E

E

E

E

E

E

F

G

G

H

78 H.A. Sandeman et al. / Precam

eferences

itken, J.D., Cook, D.G., Yorath, C.J., 1982. Upper Ramparts River (106G) and SansSault Rapids (106H) map Areas, District of Mackenzie. In: Geological Survey ofCanada, Memoir 388, 48 pp.

rmstrong, R.L., Eisbacher, G.L., Evans, P.D., 1982. Age and stratigraphic–tectonicsignificance of Proterozoic diabase sheets, Mackenzie Mountains, northwesternCanada. Can. J. Earth Sci. 19, 316–323.

leeker, W., Ernst, R., 2006. Short-lived mantle generated magmatic events and theirdyke swarms: the key unlocking Earth’s paleogeographic record back to 2.6 Ga.In: Hanski, E., Mertanen, S., Rämö, T., Vuollo, J. (Eds.), Dyke Swarms – TimeMarkers of Crustal Evolution. Taylor and Francis/Balkema, London, pp. 3–26.

owring, S.A., Podosek, F.A., 1989. Nd isotopic evidence from Wopmay orogenfor 2.0–2.4 Ga crust in western North America. Earth Planet. Sci. Lett. 94,217–230.

uchan, K.L., Mortensen, J.K., Card, K.D., 1993. Northeast-trending Early Proterozoicdykes of southern Superior Province: multiple episodes of emplacement recog-nized from integrated paleomagnetism and U–Pb geochronology. Can. J. EarthSci. 30, 1286–1296.

uchan, K.L., Ernst, R.E., 2004. Diabase dyke swarms and related units of Canada andadjacent regions. Geological Survey of Canada Map 2022A, scale 1:5,000,000with accompanying catalogue.

uchan, K.L., Mortensen, J.K., Card, K.D., Percival, J.A., 1998. Paleomagnetism andU–Pb geochronology of diabase dyke swarms on Minto Block, Superior Province,Quebec, Canada. Can. J. Earth Sci. 35, 1054–1069.

abanis, B., Lecolle, M., 1989. Le diagramme La/10-Y/15-Nb/8: un outil pour la dis-crimination des series volcaniques et la mise en evidence des processus demelange et/ou de contamination crustale. Comptes Rendus de l’Academie desSciences, Serie 2, Mecanique, Physique, Chimie, Sciences de l’Univers, Sciencesde la Terre 309, 2023–2029.

offin, M.F., Eldholm, O., 1994. Large igneous provinces: crustal structure, dimen-sions, and external consequences. Rev. Geophys. 32, 1–36.

ondie, K.C., Barsky, C.K., Mueller, P.A., 1969. Geochemistry of Precambrian diabasedikes from Wyoming. Geochem. Cosmochim. Acta 33, 1371–1388.

ondie, K., 1997. Sources of Proterozoic mafic dyke swarms: constraints from Th/Taand La/Yb ratios. Precambrian Res. 81, 3–14.

reaser, R.A., Cooper, J.A., 1993. U–Pb geochronology of middle Proterozoic felsicmagmatism surrounding the Olympic Dam Cu–U–Au–Ag and Moonta Cu–Au–Agdeposits, South Australia. Econ. Geol. 88, 186–197.

ePaolo, D.J., 1981a. Neodymium isotopes in the Colorado Front Range andcrust–mantle evolution in the Proterozoic. Nature 291, 193–196.

ePaolo, D.J., 1981b. Trace element and isotopic effects of combined wallrock assim-ilation and fractional crystallization. Earth Planet. Sci. Lett. 53, 189–202.

udás, F.Ö., Lustwerk, R.L., 1997. Geochemistry of the Little Dal basalts: continentaltholeiites from the Mackenzie Mountains, Northwest Territories, Canada. Can. J.Earth Sci. 34, 50–58.

upuy, C., Michard, A., Dostal, J., Dautel, D., Baragar, W.R.A., 1992. Proterozoic floodbasalts from the Coppermine River area, Northwest Territories: isotope and traceelement geochemistry. Can. J. Earth Sci. 29, 1937–1943.

upuy, C., Michard, A., Dostal, J., Dautel, D., Baragar, W.R.A., 1995. Isotope andtrace-element geochemistry of Proterozoic Natkusiak flood basalts from thenorthwestern Canadian Shield. Chem. Geol. 120, 15–25.

lliott, T., Plank, T., Zindler, A., White, W., Bourdon, B., 1997. Element transport fromslab to volcanic front at the Mariana arc. J. Geophys. Res. 102, 14991–15019.

rnst, R.E., 2007. Mafic-ultramafic Large Igneous Provinces (LIPs): Importance of thepre-Mesozoic record. Episodes 30, 108–114.

rnst, R.E., Buchan, K.L., 2004. Igneous rock associations in Canada 3. Large IgneousProvinces (LIPs) in Canada and adjacent regions: 3 Ga to present. Geosci. Canada31, 103–126.

rnst, R.E., Buchan, K.L., 2010. Geochemical database of Proterozoic intraplate maficmagmatism in Canada. In: Geological Survey of Canada, Open File 6016, 1 CD-ROM.

rnst, R.E., Wingate, M.T.D., Buchan, K.L., Li, Z.-X., 2008. Global record of1600–700 Ma Large Igneous Provinces (LIPs): implications for the reconstructionof the proposed Nuna (Columbia) and Rodinia supercontinents. PrecambrianRes. 160, 159–178.

rnst, R.E., Bleeker, W., 2010. Large Igneous Provinces (LIPs), giant dyke swarms,and mantle plumes: significance for breakup events within Canada and adjacentregions from 2.5 Ga to present. Can. J. Earth Sci. 47, 695–739.

vans, D.A.D., 2009. The palaeomagnetically viable, long-lived and all-inclusiveRodinia supercontinent reconstruction. In: Murphy, J.B., Keppie, J.D., Hynes,A.J. (Eds.), Ancient Orogens and Modern analogues: Geological Society. SpecialPublication 327, London, pp. 371–404.

armer, G.L., 2003. Continental basaltic rocks. In: Holland, H.D., Turekian, K.K.(Eds.), Treatise on Geochemistry. Rudnick, R.L. (Ed.), The Crust, vol. 3, Elsevier-Permagon, Oxford, pp. 85–122.

ibson, S.A., Thompson, R.N., Dickin, A.P., Leonardos, O.H., 1996. Erratum to high-Tiand low-Ti mafic potassic magmas: key to plume lithosphere interaction andcontinental flood basalt genesis. Earth Planet. Sci. Lett. 141, 325–341.

oodge, J.W., Vervoort, J.D., Fanning, C.M., Brecke, D.M., Farmer, G.L., Williams, I.S.,Myrow, P.M., DePaolo, D.J., 2008. A positive test of East Antarctica–Laurentia

juxtaposition within the Rodinia Supercontinent. Science 321,235–240.

alls, H.C., Davis, D.W., 2004. Paleomagnetism and U–Pb geochronology of the2.17 Ga Biscotasing dyke swarm, Ontario, Canada: evidence for vertical-axiscrustal rotation across the Kapuskasing Zone. Can. J. Earth Sci. 41, 255–269.

esearch 252 (2014) 166–179

Harlan, S.S., Heaman, L.M., LeCheminant, A.N., Premo, W.R., 2003. Gunbarrel maficmagmatic event: a key 780 Ma time marker for Rodinia plate reconstructions.Geology 31, 1053–1056.

Harlan, S.S., Geissman, J.W., Snee, L.W., 2008. Paleomagnetism of Proterozoic maficdikes from the Tobacco Root Mountains, southwest Montana. Precambrian Res.163, 239–264.

Heaman, L.M., 2008. Precambrian large igneous provinces: an overview ofgeochronology, origins and impact on earth evolution. In: Srivastava, R.K.,Dhmar, T. (Eds.), Precambrian Mafic Magmatism in the Indian Shield. J. Geol.Soc. India 72, 15–34.

Heaman, L.M., LeCheminant, A.N., 1993. Paragenesis and U–Pb systematics of bad-deleyite (ZrO2). Chem. Geol. 110, 95–126.

Heaman, L.M., LeCheminant, A.N., Rainbird, R.H., 1992. Nature and timing of Franklinigneous events, Canada: implications for a Late Proterozoic mantle plume andthe break-up of Laurentia. Earth Planet. Sci. Lett. 109, 117–131.

Herzberg, C.T., Fyfe, W.S., Carr, M.J., 1983. Density constraints on the formation ofthe continental Moho and crust. Contrib. Mineral. Petrol. 84, 1–5.

Jackson, V.A., van Breemen, O., Ootes, L., Bleeker, W., Bennett, V., Davis, W.J.,Ketchum, J., Smar, L., 2013. U–Pb zircon ages and field relationships of Archeanbasement and Paleoproterozoic intrusions, south-central Wopmay Orogen,NWT: implications for tectonic assignments. Can. J. Earth Sci. 50, 979–1006.

Jefferson, C.W., Parrish, R.R., 1989. Late Proterozoic stratigraphy, U–Pb zircon ages,and rift tectonics, Mackenzie Mountains, northwestern Canada. Can. J. Earth Sci.26, 1784–1801.

Lehnert, K., Su, Y., Langmuir, C., Sarbas, B., Nohl, U., 2000. A global geochemicaldatabase structure for rocks. In: Geochemistry Geophysics Geosystems 1, PaperNumber 1999GC000026.

LeBas, M.J., LeMaitre, R.W., Streckeisen, A., Zanettin, B., 1986. A chemical classifi-cation of volcanic rocks based on the total alkali silica diagram. J. Petrol. 27,745–750.

LeCheminant, A.N., Heaman, L.M., 1989. Mackenzie igneous events, Canada: MiddleProterozoic hotspot magmatism associated with ocean opening. Earth Planet.Sci. Lett. 96, 38–48.

Li, Z.-X., Bogdanova, S.V., Collins, A.S., Davidson, A., De Waele, B., Ernst, R.E., Fitzsi-mons, I.C.W., Fuck, R.A., Gladkochub, D.P., Jacobs, J., Karlstrom, K.E., Lu, S.,Natapov, L.M., Pease, V., Pisarevsky, S.A., Thrane, K., Vernikovsky, V., 2008.Assembly, configuration, and break-up history of Rodinia: a synthesis. Precam-brian Res. 160, 179–210.

Macdonald, F.A., Schmitz, M.D., Crowley, J.L., Roots, C.F., Jones, D.S., Maloof, A.C.,Strauss, J.V., Cohen, P.A., Johnston, D.T., Schrag, D.P., 2010. Calibrating the Cryo-genian. Science 327, 1241–1243.

Mackie, R.A., Scoates, J.S., Weis, D., 2009. Age and Nd–Hf isotopic constraints on theorigin of marginal rocks from the Muskox layered intrusion (Nunavut, Canada)and implications for the evolution of the 1.27 Ga Mackenzie large igneousprovince. Precambrian Res. 172, 46–66.

McPhie, J., Kamenetsky, V., Chambefort, I., Ehrig, K., Green, N., 2010. The origin ofOlympic Dam: a revolutionary new view. In: Giant Ore Deposits Downunder,IAGOD Adelaide, Proceedings, April 6–9, pp. 76–77.

Moores, E.M., 1991. Southwest U.S.-East Antarctic (SWEAT) connection; a hypothe-sis. Geology 19, 425–428.

Morris, W.A., Aitken, J.D., 1982. Paleomagnetism of the Little Dal lavas, MackenzieMountains, Northwest Territories, Canada. Can. J. Earth Sci. 19, 2020–2027.

Narbonne, G.M., Aitken, J.D., 1995. Neoproterozoic of the Mackenzie Mountains,northwestern Canada. Precambrian Res. 73, 101–121.

Niu, Y., Wilson, M., Hunphreys, E.R., O’Hara, M.J., 2011. The origin of intra-plate oceanisland basalts (OIB): the lid effect and its geodynamic implications. J. Petrol. 52,1443–1468.

Ootes, L., Sandeman, H., Lemieux, Y., Leslie, C., 2008. The 780 Ma Tsezotene sills,Mackenzie Mountains: a field, petrographical, and geochemical study. In: North-west Territories Geoscience Office, NWT Open Report 2008-011, 21 pp. andappendices. www.nwtgeoscience.ca/gateway

Park, J.K., 1997. Paleomagnetic evidence for low-latitude glaciation during depo-sition of the Neoproterozoic Rapitan Group, Mackenzie Mountains, N.W.T.,Canada. Can. J. Earth Sci. 34, 34–49.

Park, J.K., Jefferson, C.W., 1991. Magnetic and tectonic history of the Late Proterozoicupper Little Dal and Coates Lake Groups of northwestern Canada. PrecambrianRes. 52, 1–35.

Park, J.K., Buchan, K.L., Gandhi, S.S., 1995a. Paleomagnetism of the 779 Ma Hottahgabbro sheets of the Wopmay Orogen, Northwest Territories. Geol. Surv. CanadaCurr. Res. 1995-C, 195–200.

Park, J.K., Buchan, K.L., Harlan, S.S., 1995b. A proposed giant radiating dyke swarmfragmented by the separation of Laurentia and Australia based on paleomag-netism of ca 780 Ma mafic intrusions in western North America. Earth Planet.Sci. Lett. 132, 129–139.

Park, J.K., Norris, D.K., Larochelle, A., 1989. Paleomagnetism and the ori-gin of the Mackenzie Arc of northwestern Canada. Can. J. Earth Sci. 26,2194–2203.

Pearce, J.A., 1982. Trace element characteristics of lavas from destructiveplate boundaries;. In: Thorpe, R.S. (Ed.), Andesites. John Wiley and Sons,pp. 525–548.

Pearce, J.A.,1996. A user’s guide to basalt discrimination diagrams. In: Trace Element

Geochemistry of Volcanic Rocks; Applications for Massive Sulphide Exploration.Short Course Notes, Geological Association of Canada 12, pp. 79–113.

Pearce, J.A., 2008. Geochemical fingerprinting of oceanic basalts with applicationsto ophiolite classification and the search for Archean oceanic crust. Lithos 100,14–48.

brian R

P

P

P

P

P

R

RR

R

R

S

S

Province, Canada: constraints for granitoid origin and crustal isotopic structure.Can. J. Earth Sci. 36, 1131–1147.

H.A. Sandeman et al. / Precam

earce, T.H., Gorman, B.E., Birkett, T.C., 1977. The relationship between major ele-ment chemistry and tectonic environment of basic and intermediate volcanicrock. Earth Planet. Sci. Lett. 36, 121–132.

eate, D.W., Hawkesworth, C.J., 1996. Lithospheric to asthenospheric transition inlow-Ti flood basalts from southern Paraná, Brazil. Chem. Geol. 127, 1–24.

ehrsson, S.J., Buchan, K.L., 1999. Borden dykes on Baffin Island, Northwest Terri-tories: a Franklin U–Pb baddeleyite age and a paleomagnetic reinterpretation.Can. J. Earth Sci. 36, 65–73.

errier, C., (Unpublished BSc. thesis) 1988. Etude comparative de deux intrusionsgabbroiques dans la zone magmatique du Grand Lake de L’ours, province deL’ours, Territoires du Nord-Ouest. Department of Geology, Universite de Mon-treal, pp. 46.

owell, C.M., Li, Z.X., McElhinny, M.W., Meert, J.G., Park, J.K., 1993. Paleomagneticconstraints on timing of the Neoproterozoic breakup of Rodinia and the Cam-brian formation of Gondwana. Geology 21, 889–892.

ingwood, A.E., 1975. Composition and Petrology of the Earth’s Mantle. McGraw-HillBook Co., New York, NY, USA, 618 pp.

ock, N.M.S., 1991. Lamprophyres. Van Nostrand Reinhold, New York, 285 pp.ollinson, H.R., 1993. Using Geochemical Data: Evaluation, Presentation, Interpre-

tation. Longman Scientific & Technical Ltd., Harlow, UK, 352 pp.oeder, P.L., Emslie, R.F., 1970. Olivine-liquid equilibrium. Contrib. Mineral. Petrol.

29, 275–289.udnick, R.L., Gao, S., 2003. Composition of the continental crust. In: Holland, H.D.,

Turekian, K.K. (Eds.), Treatise on Geochemistry. Rudnick, R.L. (Ed.), The Crust,vol. 3, Elsevier-Permagon, Oxford, pp. 1–64.

andeman, H.A.I., Ootes, L., Jackson, V.A., 2007. Field, petrographic, and petro-chemical data for the Faber Sill: insights into the petrogenesis of aGunbarrel event intrusion in the Wopmay Orogen, NWT, Canada. In: North-

west Territories Geoscience Office, NWT Open Report 2007-007, 25 pp.,www.nwtgeoscience.ca/gateway

chwab, D.L., Thorkelson, D.J., Mortensen, J.K., Creaser, R.A., Abbot, J.G., 2004. TheBear River dykes (1265–1269 Ma); westward continuation of the Mackenziedyke swarm into Yukon, Canada. Precambrian Res. 133, 175–186.

esearch 252 (2014) 166–179 179

Sears, J.W., Price, R.A., 2000. New look at the Siberian connection: no SWEAT. Geology28, 423–426.

Shellnut, J.G., Dostal, J., Keppie, J.D., 2004. Petrogenesis of the 723 Ma Coronationsills, Amundsen basin, Arctic Canada; implications for the break-up of Rodinia.Precambrian Res. 129, 309–324.

Stoffers, A., (Unpublished B.Sc. thesis) 2005. Geochemistry and Petrography of MaficDykes from the Great Bear Magmatic Zone. Carleton University, Ottawa, ON, 41pp.

Sun, S.-S., McDonough, W.F., 1989. Chemical and isotopic systematics of oceanicbasalts: implications for mantle composition and processes. In: Saunders, A.D.,Norry, M.J. (Eds.), Magmatism in the Ocean Basins, vol. 42. Geological SocietySpecial Publication, pp. 313–345.

Turner, S., Hawkesworth, C., 1995. The nature of the sub-continental mantle; con-straints from the major-element composition of continental flood basalts. Chem.Geol. 120, 295–314.

Turner, E.C., Long, D.G.F., 2008. Basin architecture and syndepositional fault activ-ity during deposition of the Neoproterozoic Mackenzie Mountains supergroup,Northwest Territories, Canada. Can. J. Earth Sci. 45, 1159–1184.

Winchester, J.A., Floyd, P.A., 1977. Geochemical discrimination of different magmaseries and their differentiation products using immobile elements. Chem. Geol.20, 325–343.

Wingate, M.T.D., Campbell, I.H., Compston, W., Gibson, G.M., 1998. Ion microprobeU–Pb ages for Neoproterozoic basaltic magmatism in south-central Australiaand implications for the breakup of Rodinia. Precambrian Res. 87, 135–159.

Yamashita, K., Creaser, R.A., Stemler, J.U., Zimaro, T.W., 1999. Geochemical andNd–Pb isotopic systematics of late Archean granitoids, southwestern Slave

Zhou, M.-F., Kennedy, A.K., Sun, M., Malpas, J., Lesher, C.M., 2002. Neoproterozoic arc-related mafic intrusions along the northern margin of south China: implicationsfor the accretion of Rodinia. J. Geol. 110, 611–618.