Paleomagnetic data and U-Pb isotopic age determinations from Coats Land, Antarctica: Implications for late Proterozoic plate reconstructions

JOURNAL OF GEOPHYSICAL

RESEARCH, VOL. 102, NO. B4, PAGES 7887-7902, APRIL 10, 1997

Paleomagnetic data and U-Pb isotopic age determinations from Coats Land, Antarctica: Implications for late Proterozoic plate reconstructions Wulf

A. Gose

Departmentof GeologicalSciencesand Institutefor Geophysics,Universityof Texas at Austin

Mark A. Helper, JamesN. Connelly,Fred E. Hutson Departmentof GeologicalSciences,Universityof Texas at Austin

Ian W. D. Dalziel Departmentof GeologicalSciencesand Institutefor Geophysics,Universityof Texasat Austin

Abstract. Paleomagnetic resultsandisotopicagedeterminations for granophyre andrhyolite from small, isolated nunataksin southernCoats Land, Antarctica, are used to evaluate late

Proterozoicplatereconstructions. U-Pb zircondatesfor the two rocktypesindicatecoeval crystallization at 1112+ 4 Ma. A concordant 1106+ 3 Ma titanitedatefromthegranophyre overlapsthe crystallizationage,implyingrapidcooling,andis consistentwith field and petrographic evidenceof no subsequent penetrativedeformation,metamorphism, or hydrothermal disturbance.The meandirectionof magnetizationof the rhyoliteat LittlewoodNunataksis statistically indistinguishable from themeandirectionsof five sitesin the granophyreandcrosscutting rhyolitedikesat BertrabNunataks.The groupmeanvirtualgeomagnetic poleof 22.9øN, 80.3øE

(N=6, A95=6.8 ø)compares favorablywiththeonlyotherextantPrecambrian paleomagnetic poles for the EastAntarcticcraton,two polesfrom westernDronningMaud Land. The EastAntarctic andLaurentianpolesof 1.1 Ga do not coincideafterrestorationof the continentsto a position suggested by the SWEAT hypothesis juxtaposingthe Pacificmarginsof EastAntarctica-Australia and Laurentia,indicatingeitherthat the hypothesisis incorrector that CoatsLand andpartsof westernDronningMaud Land (hereinthe CMG province)were not part of the East Antarctic cratonat 1.1 Ga. In supportof the latter,thereis reasonableagreementof the 1.1 Ga CMG poles and approximatelycoevalpolesfrom the Kalaharicratonof West Gondwanawhen the CMG is restoredto a positionadjacentthe Kalaharicraton. Sucha reconstruction placesthe CMG in West GondwanaratherthanEast Gondwana,as originallyimpliedin Rodiniareconstructions, and is consistent with previouslyrecognizedlinksbetweenthe geologyof the Kalaharicratonand westernDronningMaud Land. It furtherimpliesthatthe CMG did not becomepart of East Antarcticauntil latestPrecambrianto Cambriantime. A new reconstruction placesa partially assembledWest Gondwanaoff the presentsoutheastern marginof Laurentiaat 1.1 Ga suchthat polesof the CMG andKalaharifall on the Laurentianpolarwanderpathfor thistime period.

all continents, i.e., between circa 1000 and 550 Ma.

Introduction

Until re-

cently, however, the configuration of this supercontinenthas

Paleomagnetic data,Proterozoicand early to mid-Paleozoic ophioliticcomplexes,and late Precambrian rifted marginsall indicate that the Pangea supercontinentreconstructedusing seafloorspreadingdata was not the original configurationof

remainedobscure. Moores [1991], following earlier suggestions by Canadian workers [Eisbacher, 1985; Bell and Jefferson, 1987], juxtaposed the Laurentian and East Antarctic-Australian cratons prior to the opening of the

continents on Earth. Thermal subsidenceanalyses of late Precambrianrifted continentalmargins[Bondet al., 1984] and comparison of apparentpolar wanderpaths[Piper, 1987] sug-

Pacific

gestthat at leastone earlier supercontinent existedduringthe Neoproterozoic Era betweenthe end of the Grenvilleorogeny and the depositionof lower Paleozoicmiogeoclinalstrataon

between the southwest United States. and East Antarctica.

Copyright1997 by the AmericanGeophysicalUnion. Papernumber96JB03595. 0148-0227/97/96JB-03595509.00

7887

Ocean basin near the end of the Precambrian.

He

referred to this as the "SWEAT" hypothesison the basis of intercratonic

connections

he believed

could

be identified

Most critically, Moores [1991] noted that igneous and metamorphicrocks in the ShackletonRange along the Pacific margin of the Precambriancratonof East Antarctica(Figure 1) have an age range of 1.6-1.8 Ga, comparableto those of the Yavapai and Mazatzal provincesof Arizona and New Mexico [Karlstrom and Bowring, 1988; Hoffman, 1989]. He also pointed to the existencein Dronning Maud Land, that part of

7888

GOSEETAL.:PALEOMAGNETISM ANDPROTEROZOIC U-PbAGES,ANTARCTICA

the EastAntarcticcratonsouthof Africa(in present-day coor- the southwestUnited States,providinga major constraintfor dinates;Figure 1), of high-grademetamorphic rockswith an the SWEAT reconstruction.Dalziel [1991, 1992a, b, 1994] age range comparableto that of the Grenville provinceof

and Hoffman [1991] also suggested that the protoAppalachianmargin of Laurentia was juxtaposedwith the furthersuggested that the "Grenvillian"rocksof Dronning proto-Andeanmarginof Amazoniaduringthe Neoproterozoic, Maud Land can be tracedalongstrikeinto southeastern India thusleadingto new reconstructions of the supercontinent that eastern and southern North America, circa 1.2-1.0 Ga.

He

andthe Albany-Fraser belt of southwestern Australia,thereby was assembledduring Grenvillian orogenesis(for a review, see implyingthatEastGondwanawasassembled by theendof the Dalziel [1995]). This supercontinent,which would have exGrenville orogeny. Dalziel [1991] supportedMoores' [1991] hypothesiswith a computer-generatedreconstructionand pointed out the existenceof extremelysmall, isolatedoutcropsof silicic volcanic and hypabyssalrocks of Grenvillian age in southernCoats

istedprior to the openingof both the Pacificand Iapetusocean basins, has been named Rodinia by McMenamin and McMenarnin [1990]. A secondsupercontinentalentity may have existedephemerallyfollowing the openingof the Pacific Ocean and amalgamationof Gondwanalandand prior to the

Land at the head of the Weddell Sea, 300 km from the

opening of Iapetus [Dalziel, 1992a, b, 1994]. This has been named Pannotia [Powell, 1995; Dalziel, 1997]. The paleogeography during the late Mesoproterozoic and Neoproterozoiceras is particularly important to understandingof Earth history. It marks a time of increaseddiversificationand turnover rates in protistan microfossilsand the emergenceof Metazoan life, and immediately preceded the Cambrian "explosion" of marine invertebrates [Knoll, 1994].

ShackletonRange (Figure 1). Dalziel [1991, 1992a] used the locationof theseoutcrops,hereinreferredto as the CoatsLand Nunataks,to constrainthe positionof an Antarctic"Grenville Front" (Figure 1) to the ice-coveredregion betweenthese nunataksand the ShackletonRange.The southwestern termi-

nus of this front was suggestedas a piercingpoint to be matchedwith the southernend of its Laurentiancounterpart in

Meerteta'l.[1994]used paleomagnetic poles fromseveral continents to demonstrate that the Rodinian reconstructionof

Dalziel[ 1992a]couldnothaveexistedat circa1250Ma, prior to the GrenvilleOrogeny. However,Powellet al. [1993] showed thatNeoproterozoic polesarecompatible withjuxta-

ic

w

position of the Pacific margins of Laurentia and East Gondwanaduringthe periodcirca900-725Ma. The datafrom East Gondwanausedin their analysis,however,comefrom

Antarc

Australia andIndia,andthetimingof EastGondwana amalgamationis not well established [Grunowet al., 1996]. For example,there may be an early Paleozoicsuturein the Shackleton Range[Marsh,1983;Buggisch et al., 1993;Dalla

Saldaet al., 1992b;Tessensohn, 1996]andalongtheIndian Oceanmarginsouthof India[Shiraishi et al., 1994].A more

thorough testof the SWEAThypothesis therefore requires Neoproterozoic and/orlate Mesoproterozoic paleomagnetic datafromtheEastAntarctic craton. Owingto theextentof the EastAntarcticice sheet,exposed rockssuitablefor suchanal-

1.6-1.8 GaI

ysesarerestricted to the periphery of thecraton,a position that mayrenderthemsuspect andunrepresentative of the cratonproperdue to potentialdisplacements or translations, especiallyfor exposuresalong the Transantarctic Mountains

Coats Land

[e.g.,BorgandDePaolo,1994].Yetthese rocksrepresent the best,andin mostcasestheonly,opportunity for testing the

Nunataks

(Area of Fig. 2)

hypothesis. Evenif suchrocksproveto be allochthonous and

unsuitable fortesting SWEATperse,theymayreveal, asdiscussed below,important insights intothepaleogeography of thisimportant time intervalthatareunobtainable elsewhere. Filchner

Ice Shelf

lOO

o

lOO Km

PolarProjection I

Figure 1. Map of outcropsof Precambrianand lower Paleozoicrocksalongthe WeddellSea marginof the East

Antarctic Precambriancraton (see inset), showingthe suggested continuation of the Grenville Front of Laurentia

Herewereporttheresults of paleomagnetic andgeochronologicstudiesof the circa 1.1 Ga igneousrocksof the Coats LandNunataks, the rockssuggested by Dalziel [1991]to be the most proximal equivalentsin East Antarcticaof the Grenville provinceof the southernUnited States.Thesedata indicate,however,thatpriorto circa550 Ma, CoatsLandwas part of West Gondwanarather than East Gondwana.Thus the "AntarcticGrenville front" cannotbe usedto define the relative positionsof the Laurentian and East Antarctic cratons.

[Dalziel, 1992b],the Maudheimand Grunehogna provinces contribute to a betterunderstanding of [from Moyeset al., 1993], the ShackletonRange,and the Thedatado,however, location of the Coats Land nunataks. Generalized areasof rock

exposure are shown in black.

Abbreviations are Ahl,

Neoproterozoic paleogeography by suggesting proximity of

the Laurentianand Kalaharicratonsat circa 1100 Ma, and

Ahlmannryggen; B, Borgmassivet; DML, DronningMaud

demonstrate that the East African suturebetweenEast andWest

Land; TAM, Transantarctic Mountains.

Gondwanaland mayhaveextended to thePacificmarginof

GOSE ET AL.: PALEOMAGNETISM AND PROTEROZOIC U-Pb AGES, ANTARCTICA

Antarctica as recently suggestedby Grunowet al. [1996] and Tessensohn [1996].

Regional Setting Grenville age rocks (1.2-1.0 Ga) in westernDronningMaud Land and Coats Land are found in the amphibolite- to granulite-faciesMaudheim Province [Groenewaldet al., 1991]

BB5A

.•.•..--,•r......

[ii?"•'">•i•i;•!i:11iii•:iii• 0 20I

LN 1P3 LN 1-3A

B N•'"' •'•:"•:•7•••••:• • g , , 1 NUNATAKS

I

• ":•

,

I BN1P3 •

II Nunaaks

••

78o•[

LITTLEWOOD

BN2P2 0

200 m

...... ppomat:a,

and some 600 km to the southwest in the small, isolated

nunataks of southern Coats Land (Figure 1). Geologic similarities and early Mesozoic continental reconstructions controlledby seafloor spreadingdata [Martin and Hartnady, 1986; Lawyer and $cotese, 1987] suggestthat the Maudheim Province and the bordering Archean basement and Mesoproterozoiccover to the north, the GrunehognaProvince (Figure 1) [Groenewaldet al., 1991], were contiguouswith the Natal/Namaqua/MozambiqueBelts and the Kalahari craton, respectively,in southernAfrica [Moyeset al., 1993;Jacobset al., 1993; Groenewald, 1993] until Mesozoic break up of Gondwanaland.With the exceptionof the stableinteriorof the Kalahari craton, all of these areas underwent varying degrees of tectonic and metamorphicreworkingin latest Proterozoic to Early Cambriantime duringthe Pan African/Brazilideoro-

7889

............ •....'............... •::". ............................................... •::':•o • N

•i•::•:•:•l:•::::... ::::::::::::::::::::::::::::::::::::::: • •:.•}?• •[oS •

Approximate Scale

•

Figure 2. Maps of the Bertrab and Littlewood Nunataks showing the location, geology, and sampling sites. The locationof Figure 2b is shownby the black box in Figure 2a. Figure 2b showsthe location of the Berttab of Littlewo0d Nunataks. Medium shading in Figures 2c and 2d denotes granophyre at the Bertrab Nunataks and rhyolite at the Littlewood

Nunataks.

Within

the Bertrab Nunataks, solid,

black lines are mafic dikes and NE trendingdikes labeled "r" are rhyolite.

geniesof interiorGondwanaland andthe Rossorogenyof the Transantarctic (paleo-Pacific) margin. In contrast, the nunataks of southern Coats Land apparently escaped such overprinting[Aughenbaugh et al., 1965;Marsh and Thomson, 1984; Storey et al., 1994], making them amongthe best can- ciated with hypabyssalor extrusivemodesof emplacement[cf. didatesin this part of the East Antarcticcratonfor paleomag- Barker, 1970]. netic study. Paleomagnetic samples were collected at six sites in the Bertrab Nunataks; four from granophyre (BN1P1, BN1P3, Coats Land

Nunataks

The Coats Land Nunatakscomprisethree, very small, isolated groups of exposuresof igneous rocks, the Moltke, Bertrab,and LittlewoodNunataks(Figure2), of whichonly the latter two are readily accessible;we found the rocks at the Moltke localityto be inaccessible beneatha verticalice fall. Bertrab

Nunataks

The Bertrab Nunataks are composed of red- to grayweathered, fine- to medium-grained, oligoclase-phyric granophyre. The granophyre is cut by 1-7 m wide, flowbandedrhyolite dikes and 1-2 m wide, fine-grainedmafic dikes (Figure 2c). Detailed petrographic descriptions of the granophyreand dikes at the largestnunatak (at the Argentine BelgranoII base) are provided by Toubes Spinelli [1983] and Marsh and Thomson [1984]. Storey et al. [1994] reported a Rb-Sr whole rock isochronage of 1076 + 7 Ma (mean square of weighted deviates (MSDW) = 1.6) for eight samplesof granophyre, rhyolite and diabase from this nunatak and suggest,on the basis of these and other limited geochemical data, that the dikes and granophyreare cogenetic. The rocks of the previously undescribedother two nunataks of the Bertrabgroupare nearly identicalto thoseof the largestexposure, except that the granophyreis locally slightly coarserand/orfiner-grained. The middlenunatakof the groupis cut by an ~2 m wide diabasedike that is in turn cut by a 1 m wide, flow-banded rhyolite dike identical to those at the larger nunatakbut trending 045ø. Granophyrictextureslike those found throughoutthe BertrabNunataksare indicativeof rapid crystallizationand/or devitrificationand are commonlyasso-

BN1P4, and BN2P2) and two from rhyolite dikes (BN1P2, BN2P1) (Figure 2c). Granophyresamplesare composedof a fine-grained,granophyricmatrix of quartz and alkali feldspar, which hosts fine- to medium-grainedphenocrystsof highly zoned plagioclase,and irregularly dispersedconcentrationsof biotite, magnetite, titanite, chlorite, and, rarely, amphibole, apatite, and zircon. The granophyreis extremely fresh and unweatheredin outcrop. Thin section examination reveals moderateto thoroughsericitizationof plagioclaseand microtextures and paragenetic relationships among the mafic minerals and oxides that are indicative of hydrothermaland/or deuteric alteration. Relict amphibole is partially or completely replaced by biotite, chlorite, and magnetite, and biotite is partially replacedby chlorite. Magnetite is rimmed by anhedral, extremely fine-grained, yellow titanite, which together are spatially associatedwith clustersof amphibole and biotite. The yellow titanite is morphologically distinct from coarser-grained, isolated,brown, subhedraltitanite. The brown titanite is not spatially associatedwith mafic mineral clusters,nor with partially reactedmineralphases,and is thus interpreted as primary. The two sampledrhyolite dikes are vertical, flow bandedat their margins,and range from 1 to 2.5 m in width. Both are composed of a microcrystalline groundmassof quartz and feldsparthat containssparsephenocrystsof quartz, microcline microperthite, biotite, and, rarely, plagioclase. The groundmass locally exhibits spherulitic and granophyric textures.

Very fine-grained to fine-grained magnetite is dispersed throughoutthe matrix and feldsparphenocrysts.The rhyolite dikes do not show petrographicevidencefor alteration like that seen in the granophyres.

7890 Littlewood

GOSE ET AL.: PALEOMAGNETISM AND PROTEROZOIC U-Pb AGES, ANTARCFICA Nunataks

U-Pb Geochronology

This secondgroupof outcropscomprisesfive small exposures of red-weathered, densely silicified rhyolite that are locatedabout 10 km eastof the BertrabNunataks(Figure2d). Aughenbaughet al. [1965] providedlithologicdescriptions,a chemicalanalysis,and a wholerock K-Ar ageof 851 + 3 0 Ma (recalculatedfollowing Dalryrnple [1979]) for rhyoliteat the

largestoutcrop of thegroup.Thesignificance of thisageis unclear. Volcanic rocks,particularlythosethat are devitrified, generally do not behave as closed systemswith respect to argon and may yield ages that can be substantiallyyounger than the age of crystallizationor coolingdeterminedby other methods [McDougall and Harrison, 1988]. The only other publishedage determinationfor Littlewoodrhyoliteis a 1001 + 16 Ma

whole

rock Rb-Sr

isochron

for a mixed

suite of

Littlewood and Bertrab samples[Eastin and Faure, 1971]. When recalculatedwith moderndecayconstants and moreconservativeerror estimates,the samedata yield an age of 976 + 35 Ma [Storeyet al., 1994]. Despitedifferencesin isotopic ages, Littlewood rhyolite and Bertrab granophyresand dikes

havebeenpreviously beenregarded ascoeval andcogenetic on the basis of limited isotopic, major and trace element data [Eastin and Faure, 1971' Storey et al., 1994]. Paleomagnetic samples of Littlewood rhyolite were collected at three sites (LN1P1, LN1P2, and LN1P3) in the largest Littlewood outcrop (Figure 2d). Rhyolite at this locality is composedof microcrystallinequartz and feldsparthat contains

sparse, angularfragments of sanidine andquartzandis pervaded by chalcedony-filledor, more rarely, drusyquartz-lined veinlets. Black to gray, angular, volcanic rock fragmentsup to 2 cm in longest dimension are present in some samples. Thin section and scanning electron microscopeexamination showsthat the opaquemineralogyof the rhyolite is dominated by hematiterangingfrom --0.5 mm to 2 mm in largestdimension. Hematite is ubiquitousthroughoutthe microcrystalline quartz and feldspar matrix, occurring both as isolated grains and crystal clusters, som• of the latter having partially or completely replaced biotite. The remaining Littlewood Nunataks,for which there are no previouslypublisheddescriptions,are also composedof redweathered, silicified rhyolite. Unlike the largest nunatak, however, rhyolite in these exposures locally preserves featuresin both outcropand thin sectionthat indicatea pyroclastic origin. Such features include the (1) common occurrence of angular volcanic lithic clasts, (2) angular, ragged, fragmentary appearance of feldspar phenocrysts, (3) rare occurrenceof devitrified glass shards, and (4) presenceof elongate, aligned, irregular cavities interpretedto be voids left by weatheringof partially welded pumice. The textural characteristicsof the rhyolite in all outcropsare similar to thoseof partially welded ash flow tuffs that have beenalteredand silicifled by extensive vapor phase alteration, indicating these rocks underwentrapid depositionand cooling. An indistinct, discontinuouslayering defined by weathered surfaces and alignment of elongate cavities at the sampling sites and in other nunataks of the group has a strike that ranges between 020ø and 028ø and dips 10ø-11ø SE. On the basisof the texturesdescribedabove,we interpretthe layering to be primary and of ash flow origin. Ash flows mantle topography such that layering need not be horizontal. Lacking field and petrographicevidence to indicate subsequentdeformation, we interpretthe 10-11o dip as a primary feature.

Precisedatingof the crystallizationand alterationhistory of the Littlewood and Bertrab rocks is crucial for accurate interpretationof the paleomagneticdata. Both units and the dikesthatcut the granophyre havepreviouslybeenregardedas cogeneticand coeval,but extantisotopicagesare not precise enoughto adequatelytestthis premise,nor do theyaddress the age of alteration. Zircons in Littlewood rhyolite and Bertrab granophyrepermit U-Pb dating of the crystallizationof these rocks, and titanite in the granophyrecan be usedto date the alteration. Resultsof the U-Pb analysesof thesemineralsare discussed belowandpresented in Table 1. Analyticalmethods are given in the appendix. Littlewood

Rhyolite

Three fractionsof high-qualityzircons were analyzedfrom Littlewoodrhyolite (sampleLN13A, Figure2d). Two fractions of approximately 50 euhedral, colorless, inclusion-free, needle-shapedgrains plot within error on concordiawith an

average2ø7pb/2ø6pb ageof 1112+ 4 Ma (Figure3a andTable 1). Fraction Zg, composed of an equivalent number of stubbier, prismatic crystals with tiny (1-2 mm) clear inclu-

sions,is slightlydiscordant(1.6%), givinga 2ø7pb/2ø6Pb age of 1118 Ma (Figure 3a). The zircon morphology of all fractionsis indicative of an igneousorigin and the average

1112+ 4 Ma 2ø7pb/2ø6Pb age of the two concordant fraction thusdatesthe time of igneouszircon crystallization.We infer this age to be contemporaneous with the age of tuff emplacement. The discordanceof Z3 is not well understoodbut may reflect the combinedeffects of a very minor componentof older, inherited zircon and recent lead loss.

Bertrab

Granophyre

Zirconsfrom Bertrabgranophyre(sampleBB5a, Figure2c) are of poorer quality than those from Littlewood rhyolite. Three fractionsof near-equant,faceted zircons containing small inclusionsand crackswere analyzedand definea discordia line with upperand lower interceptsof 1113 + 18/-6 Ma and 212 + 350 Ma respectively(92% probabilityof fit, where10% representsa MSWD of 2) (Table 1 and Figure 3b). The upper intercept,which overlapsthe age determinedfor Littlewood rhyolite,is interpretedto representthe time of crystallization of the granophyre.The 3-4% discordanceof the bestzirconsis

consistent with the loweraveragequalityof thesegrainscomparedto thosefrom the rhyolite and precludesdetermination of more preciseintercepterrors. Two fractionsof titanite, one of primary brown titanite (T1) and anotherof yellow secondarytitanite (T2), were also analyzed(Table 1). Both fractionsyield analyticallyindistinguishableresultsthat overlapconcordiaand give an average

2ø7pb/2ø6Pb ageof 1106+ 3 Ma (Figure3b). The coincident age of both yellow and brown titanite suggestseither that the texturallyearlier brown titanite was completelyresetby alteration that producedyellow titanite and magnetitefrom Fe-Ti oxidesand silicates,or that this alterationoccurredessentially contemporaneous with crystallizationand is thus deutericin origin. The latter interpretationis favoredby the overlapof the titanite and zircon ages. In either event, the coincidence and concordanceof both titanite analysesdemonstratethat alteration of the granophyreis no youngerthan circa 1106 Ma.

GOSE ET AL.: PALEOMAGNETISM AND PROTEROZOIC U-Pb AGES, ANTARCTICA

•-•

ce.• o

c•1 c•1 c•1 o

,•D

ooo ,-•,•,•

o

o

o,-•,•

o o o

o o o,-•,•

•

0

0

•v

o•

',.•

0

ooo •

ooo

NNN•

0

7891

7892

GOSE ET AL.: PALEOMAGNETISM AND PROTEROZOIC U-Pb AGES, ANTARCTICA

The overlap within errors of the zircon and titanite datesis

1.0

also consistentwith the rapid crystallizationand cooling

inferred from the rock textures. If the titanite age isregarded .•

0.8

ature of titanite [Tucker et al., 1987; Heaman and Parrish,

0.6

asacooling agethen theestimated 500-600øC closure temper-• •

1991 ] requires thatthegranophyre initially cooled toorbelow .õ the magnetiteCurietemperature (580øC)by ca. 1106Ma. The

LN1P3.3

0.4

•

BNIP4.2

concordance of thetitanite analyses andtheabsence of signif- •o

0.2

icant lead losssuggestthat theserockshavenot experienced significantreheatingsince emplacement,an inferenceconsistent with the igneous textures and mineralogies and the absence of deformational

0.0

0

i

i

i

i

i

100

200

300

400

500

zircons from the granophyrecan be attributedto recent lead

lossthatwasenhanced byslightly higher U contents (Table1)

600

700

Temperature ( øC )

fabrics. The 3-4% discordance of

.•

1.0

andtheubiquitous fractures within these grains. Thehighly • 0.8

impreciselower interceptis within the range of Mesozoic

_= o.6

events related to Gondwana breakup, whichelsewhere include .õ o.4

uplift, denudation, anderaplacement offlood basalts (for • 0.2 reviewseeDalziel[1992a]).Anyorallof these events could z have promotedlead lossfrom thesefracturedzircons.

o.o

0

.1

.2

.3

.4

.5

.6

.7

.8

.9

1.0

0

.1

.2

.3

.4

.5

Applied Field (Tesla)

Figure 4. Thermal demagnetizationand IRM acquisition with coercivity spectrafor samplesof Bertrab granophyreand Littlewood rhyolite. The analyses demonstrate that the

lA. LITTLEWOOD RHYOLITE (LN13A) 1120

.1892

1110/•Z1 Z2/•// Zircon

206pb

238 U .1850

magnetizationin Bertrab samplesis carried by magnetiteand that Littlewood rhyolite containshematite as the only mag-

•/

1

_

netic

mineral.

1112+/-4Ma

,,,' 7.3

U-Pb dates from the rhyolite and granophyredemonstrate they are coeval. We considerthe more precise U-Pb date of 1112 + 4 Ma as representativeof the crystallizationage of

1090 ,,,'

this suite.

.1808

•

I

I

I

1.90

207pb 235U

I

2.02

1.96

B.BERTRAB GRANOPHYRE (BB5A) J .1892

Titanite

111•0/•'•

_2ø6pb 1106+/-3 MaT2T••

A concordant

date of 1106 + 3 Ma for titanite from

the granophyreoverlapsthe crystallizationage, impliesrapid coolingandis consistent with field andpetrographic evidence for no subsequent thermalor tectonicdisturbance.

PaleomagneticAnalysis Samplesfor paleomagneticanalysiswere collectedfrom rhyolite at the largest Littlewood nunatak and from granophyreand rhyolite dikes at two of the Bertrabnunataks (Figure 2). Samplingprocedures and analyticalmethodsare describedin the appendix.

Polishedthin sectionand scanningelectronmicroscope examinationsidentified magnetitein Bertrab samplesand hematitein Littlewoodsamples.Theseobservations are con-

.1850

1090•

J

•.•_.

firmedby isothermal remanent magnetization (IRM) acquisition experiments andcoercivityspectrum analyses [Dunlop,

Zircon

_•_Z1 Ui=1113.5+18/-6 Ma 1972]. Bertrab samplesacquiremost of their remanencein

1080/1/"' /,,7/-Z2

appliedfieldsof lessthan0.1 mT andreachsaturation by 0.3 mT (Figure 4).

.1808

IL1 Li=212+/350 -I Ma 1.90

207p b 235

•

I --U I

1.96

2.02

Figure 3. Concordiadiagramsfor (a) zircon from Littlewood rhyolite and (b) zircon and titanite (shaded) from Bertrab granophyre. Zircon and titanite fractions are labeled with

referenceto Table 1. Dashedreferenceline in Figure3a is from the best fit discordia line for Bertrab granophyrezircons shownin Figure3b. Abbreviations areLi, lowerintercept; Ui, upper intercept.

This behavior is indicative of a mixture of

multi- andsingle-domain magnetite.The thermaldemagnetization of the natural remanentmagnetizations(NRMs) also

showstwo typesof magnetite: one hasunblocking temperaturesrangingfrom 200ø to 400øC,dependingon the specific sample,andis interpreted as multidomainmagnetite; the other has unblockingtemperatures very near the magnetiteCurie temperature(580øC), indicativeof single-domainmagnetite with very little or no titanium.

In contrast,Littlewood sampleshave hematiteas the sole carrier of the magneticremanence. The shapeof the IRM

GOSE Er AL.: PALEOMAGNETISM AND PROTEROZOIC U-Pb AGES, ANTARCTICA

N

•

N

0C95 =31'1ø

•

0C95 =26'6ø

7893

N

./•

0C95= 11.5 ø

Figure5. Equal-area stereographic projections of the directions of magnetization of the NRM anddirections afterthermaldemagnetization at 200ø and575øCfor samples fromsiteBN1P1in theBertrabNunataks.Open (solid)symbols arein the upper(lower)hemisphere. PD is thepresent dipolefield direction, whichhasa negativeinclination in the southern hemisphere. Sameconvention appliesto all similarfigures.

U

N

U

BN1P1.6

BN1P3.1

6.1x 10-2 A/m

2.4x 10-2A/m

W, 7H .... •

,

.'.'. 'BN1P4 5.3 x10-2 AJm/I

•a•

.....•...E,.+H S'"D

N'

BN2PI.1

BN2P2.1

7.2x 10-2A/m

1.6 A/m

S

ß

D

5

.

.

D

Figure 6. Vector componentdiagramsfor Bertrab samples. Open squaresare the projection onto the updown-horizontalplane and crossesrepresentthe projectiononto the N-S-E-W plane. Scale is given below samplenumber. Almost all samplescontaintwo componentof magnetization,a low temperaturecomponent closely alignedwith the presentgeomagneticfield and a well-defined,high-temperaturecomponent. Samples were thermally demagnetizedat 150ø, 200ø, 250ø, 300ø, 350ø, 400ø, 450ø, 500ø, 525ø, 550ø, 560ø, 570ø, 580ø, and 600øC.

7894

GOSEET AL.:PALEOMAGNETISM AND PROTEROZOIC U-PbAGES,ANTARCTICA

acquisitioncurve (Figure 4) indicatesthe presenceof coarsegrained hematite (specularite)as well as very fine-grained hematite. In thermal demagnetizationof their NRMs, these samplesretainedover 80% of their remanenceup to 660øCbut lost most of their remanenceafter heatingto 680øC,confirming hematiteas the solecarrierof the remanentmagnetization. The responseof Bertrab samplesto thermaldemagnetization is well illustratedby the behaviorof the samplesfrom BN1P1 (Figure5). NRM directionsare stronglybiasedtowards the dipole field direction (declination is 0 ø, inclination is -84ø) and move steadily toward a direction with a positive

principalcomponentmethodcan not be applied. The characteristic direction of magnetizationfor each sample has been calculated usingthe Fishermeanof all demagnetization steps up to 660øC; statisticalparametersare listedin Table 2. Interpretation

of

Paleomagnetic

The means from the three Littlewood

Data sites are almost identi-

cal (Figure9), which suggests thatthey are from a singlerhyolite flow and hence their directionsrepresenta single spot readingof the geomagneticfield. The three sitesare thuscombined for a single virtual geomagneticpole (VGP, Table 2). inclination. At the same time, the scatter of the directions Bertrabdirectionsare indistinguishable from Littlewooddirecdecreases,as manifestby a decreasein at95and an increasein tions at the 5% significance level. This is consistent with Fisher'sprecisionparameterk. Upon exceedingthe magnetite Bertrab granophyreand Littlewood rhyolite being coeval, as Curie temperature,the directionsdisperse. demonstratedby their U-Pb ages.The combinedBerttab and Most Bertrab samplescontain two well-defined compoLittlewood data sets (six sites) yield a group-meanVGP at nents of magnetization(Figure 6). One componentis stable 22.9øS,260.3øEwith an A95=6.8ø (Table 2). The angulardisup to -400øC and scattersaroundthe presentmagneticfield persionof the VGPs is 8.1ø with upper and lower 95% confidirection (Figure 7). This magnetizationis interpretedas a dencelimits of 14.2ø and 5.7ø [Cox, 1969]. The expecteddisviscousmagnetizationresidingin multidomainmagnetite.In persion for the paleolatitudeof the sampling area (28ø) is the samplesfrom the rhyolite dike at site BN1P2, this was the about14ø for the presentgeomagnetic field and rangesfrom 9ø only component,and these results are not consideredfurther. to 18ø over the past 200 Myr [McFadden et al., 1991]. The The secondcomponentwas isolated over a demagnetization error limits of these model calculations are about 2 ø. Results range of 400ø to 580øC. The linear decay toward the origin from Keweenawanrocks suggestthat secularvariationat circa implies that no other high-temperature magnetization is 1.1 Ga was similar to current variation [Halls and Pesonen, presentin these samples.Statisticalparametersand resultsfor the Bertrab samplesare summarizedin Table 2. The direction of magnetizationof four samplesdiffered from their site mean by more than two angular standarddeviations;these samples have been deleted from the statisticalanalyses. The hematite-bearingLittlewood samplesrespondedvery uniformly to thermal demagnetization.The directionsretained a tight grouping throughout the entire demagnetizationprocess up to 660øC. Because of the very narrow unblocking temperaturerange, the vector componentdiagrams(Figure 8) do not show linear trendsbut rather a clusterof points and the

N

1982]. We regard the overlap, within the confidencelimits, of the observed and expected dispersionsas an indication that paleosecularvariation has been averaged and that the combined data sets representa time-averagedpole position.We nevertheless

consider

the alternative

view

that these data do

not average paleosecularvariation in discussionsof paleogeographic reconstructionsbelow. Age

of

Magnetization

The high unblocking temperaturesof the characteristic magnetizationsof the rhyolite and granophyrewould require a thermal event of similar temperatureto remagnetizethe samples [Pullaiah et al., 1975]. Isotopic and petrographicstudies clearly imply that such an event has not occurred since the crystallization and cooling of titanite at circa 1106 Ma. Chemical remagnetization by a low-temperature fluid event would require precipitationof two different magneticminerals in the two rock units, with both mineralsprecipitatinga very narrow range of grain size, as indicatedby the narrow unblocking temperature spectra,two requirementsthat seem unlikely to result from a single event. These considerationsimply that the magnetization is primary and, becauseof rapid cooling, has an age that is closely approximatedby the U-Pb titanire age. Further evidence that the primary magnetizationhas been successfullyisolated is the observation that our Coats Land pole is distinctively different from younger poles from East Antarctica(seeFigure 10) [Grunow, 1995]. Structural

disturbance

The very limited outcroparea precludesfully determining

Figure 7. Equal-area projection showing that the directions of magnetizationof the low-temperaturecomponentof Bertrab samplesis strongly biased toward the present field direction (PD).

whether this region has been tectonicallydisturbed. There is, however,no field or microscopicevidencefor deformationand the shallowdips of Littlewood rhyolite layering, as discussed above, are consistent with a depositional orientation. Moreover, the agreement between Littlewood and Bertrab poles suggeststhat disturbanceat the outcrop scale has been minimal and implies structuralcoherenceat a 10 km scale.

GOSEET AL.:PALEOMAGNETISM ANDPROTEROZOIC U-PbAGES,ANTARCTICA U

7895

N

LN1P1.8

LN 1 P2.8

0.29 A/m

0.59 A/m

W,:.H.........

S

D

LN1P3.3

LN 1 P3.9

0.32 A/m

0.38 A/m

......

S

D

S

, , •E,•H

D

Figure 8. Vectorcomponent diagramsfor samplesof Littlewoodrhyolite. Most samplesretained >80% of

theirNRM intensityup to 660øC. Samples werethermallydemagnetized at 150ø, 200ø, 250ø, 300ø, 350ø, 400o, 450ø, 500ø, 525ø, 550ø, 560ø, 570ø, 580ø, 600ø, 620ø, 640ø, 660ø, and680øC.Conventions as in Fig. 6.

To evaluatestructuraldisturbancesat a larger scalerequires

N

comparison with otherPrecambrian polesfrom Antarctica,of whichthereare only two. Hodgkinson[1989] sampled44 sites in intrusionsof the Borgmassivetin the Grunehogna Provinceof westernDronning Maud Land (Figure 1), some 1200 km to the northeastof our samplingarea.The sampling

sitesweredistributed overapproximately 2000km2. Because of the uniformresponseof pilot samplesto stepwisealternating field demagnetization, bulk demagnetization levelswere chosenfor the remainingsamples.No magneticreversalswere observed.A meanpoleof 233øE,6øS,A95=7øwas determined. Common directionsfrom intrusionsshowingclearly different

relativeagesin the field led Hodgkinson to suggestthat the rockswere remagnetized by a low temperature metamorphic D eventactingovera 10 to 100 m.y. time span. A 1200Ma age W of magnetizationwas arrived at by comparisonwith the 30 ø 60 ø Africanapparent polarwanderpath,in particularwith thepole positionfrom the UmkondoLavas.Hargraveset al. [1994] Figure 9. Site mean directions of Littlewood (squares) and reviewed the Umkondo and other pole positionsfrom the Kalahari Craton and their age information, arriving at an

averageageof 1100 Ma for all of theseAfricanrocks. We notethatMoyeset al. [ 1995] conclude, on the basisof Rb-Sr

Bertrab (circles) siteswith their circlesof 95% confidence. PD is the present dipole field direction in the sampling area. Bertrab directionsare indistinguishablefrom Littlewood directions at the 5% significance level. The results from site

and Sm-Nd whole rock data, that the Borgmassivetintrusions

BN1P2

have been deleted from the statistics.

7896

GOSE ET AL.: PALEOMAGNETISM AND PROTEROZOIC U-Pb AGES, ANTARCTICA

Table 2. Statistical Parametersof PaleomagneticResults

N BN1P1 BN1P3 BN1P4 BN2P1 BN2P2 LN1P1 LN1P2 LN1P3

LN Sample Mean

R

D

I

k

0t95

Lat

Long

dp

dm

9(1) 9(1) 9(1) 9(1) 4(0) 10(0) 10(0) 10(0)

8.94 8.89 8.53 8.54 3.92 9.98 9.98 9.96

285.7 285.7 286.2 288.5 304.8 284.6 288.0 288.0

52.6 44.2 35.5 43.4 48.7 50.1 48.9 50.6

128.3 73.2 17.0 17.6 35.7 367.1 472.3 201.5

4.6 6.1 12.9 12.6 15.6 2.5 2.2 3.4

29.1 22.1 15.8 21.0 22.3 27.2 25.5 26.9

78.1 76.2 75.2 78.7 94.9 76.7 79.6 70.0

4.3 4.8 8.6 9.8 13.5 2.3 1.9 3.1

6.3 7.6 14.9 15.7 20.5 3.4 2.9 4.6

30

29.91

286.9

49.9

LN Site Mean

3

3.00

286.9

49.9

BN+LN POLES

6

5.95

10

9.22

BN1P2

340.1

1.5

26.5

78.8

1.3

2.0

2782.3

304.6

2.3

26.5

78.8

3.1

3.1

98.8

6.8

22.9

80.3

11.5

14.9

57.1

299.6

19.0

23.8

-64.1

N is thenumberof samples included (deleted) in thestatistics; R is theresultant vector;D andI aretheeastdeclination and inclination; k is Fisher's precision parameter; a95is theradiusof the95%circleof confidence; Lat andLongarethelatitudeand eastlongitude of thepoleposition transferred inothenorthern hemisphere; dpanddrnarethesemiaxes of theellipses of 95% confidence. The BN (Bertrab)sitesarelocatedat 77ø 52.4'S,34ø 38.1'W,theLN (Littlewood)sitesat 77ø 53.0'S,34ø 19.3'W.

have an age of circa 1000 Ma. It is unclearwhetherMoyes et al. [1995] datedany of the rocksusedby Hodgkinson[1989]. In the Ahlmannryggen,alsoin the GrunehognaProvinceof westernDronningMaud Land (Figure 1), Peters [1989] collectedsamplesfrom 11 flows and sills. He performeddetailed geochemical, isotopic, petrographic, and rock magnetic analyses.Paleomagneticsampleswere subjectedto progres-

ANT

sive demagnetization,mainly by alternatingfield, and characteristic directions were obtained by analyzing vector compo-

nent diagrams. All sites were of the same polarity, which Peters [1989] interpretedto indicatethat the sitesare approximately the same age. A mean pole position of 239.9øE, 8.8øS,A95=5.5ø wasdetermined.Peters[1989]considered a KAr ageof 1183+ 33 Ma from onedike to be representative for

475

ANT 515

60 ø

120 ø

Figure 10. Comparisonof mean paleomagnetic pole positionof--1100 Myr old rocksfrom the Kalahari cratonof Africa (KAL) with polesfromthe CoatsLand(COATS),Borgmassivet (BORG), andAhlmannryggen (AHL). The polesfrom Antarcticahavebeenrotatedinto the Africancoordinate system(seeTable 3). Also shownarethe meanAntarcticapolesfor 515 and475 Ma [Grunow,1995]in Africancoordinates. The traceof crossesshows,in 10ø increments,the pole positionof Coats Land if the samplingsites are subjectedto vertical axis rotation. Stereographicprojection.

GOSE ET AL.: PALEOMAGNETISM AND PROTEROZOIC U-Pb AGES, ANTARCTICA

all sampledrocks. The Borgmassivetand Ahlmannrhyggen poles are similar to ours (Figure 10), although large uncertaintiesin their age do not permit rigorouscomparison. Tectonicdisturbancesby undetectedlocal vertical axis rotation are a potential source of the 30ø angular difference betweenthe Coats Land and Dronning Maud Land poles. The rocks nearest the Littlewood

and Bertrab

nunataks

in the

southern Maudheim Province are transected by a major Grenville dextral shear zone [Jacobs, 1991; Jacobs et al., 1993]. If the shear zone extends southwardbeneath the ice, the Coats Land Nunataks would lie close to its projectedtrace. Strike-slipmotionon this fault couldhave potentiallyresulted in local block rotations affecting the nunataks. Calculated

pole positionsfor different amountsof hypotheticalvertical axis rotations(Figure 10) do not, however, place the Coats Land pole appreciablycloserto the otherAntarcticpoles,or

7897

Discussion

In lightof •stablished geologic tiesbetween Dronning Maud Land and southernAfrica [Groenewaldet al., 1991, Groenewald 1993; Jacobs et al., 1993], it is informative to

comparethe polesfrom Antarcticawith polesof similarage from Africa after restoringthe two cratonsto their well-determinedGondwananconfiguration[Lawyeret al., 1992]. There are five poleswith an age of about1100 Ma from the Kalahari craton that yield a mean pole at 65.4øN, 36.1øE and an A95=6.5 ø [Hargraves et ai., 1994]. In this reconstruction (Figure 10), the Coats Land pole differs from the mean

Kalaharipoleby 30ø. The twopolesarestatistically distinct, but this may be due to poor age constraintsfor the Kalahari poles. The Umkondo dolerites have a Rb/Sr whole rock-biotite

isochronage of 1080 +140/-25 Ma [Allsoppet al., 1989], the

to Africanpoles(seebelow)for thesametimeperiod(Figure age of the post-Watersbergdoleritesis derived from one sill 10). Undetectedtilt could also be a factor but restorationof

datedby Rb/Sr at 1090 + 15 Ma [Allsoppet al., 1989], andthe

the i0-11ø dip at Littlewood Nunataks by rotationaround the Timbavatigabbrohas4øAr/39Aragesof 1074+ 4 Ma and strikeof the layeringmovesthe CoatsLand pole 2ø away from

1123 __+ 5 Ma [Burger and Walravern,1979, 1980]. Thus the

theDronning MaudLandpolesandresults in a slightlypoorer differencebetweenthe CoatsLand andthe Kalaharipolesmay groupingof the combinedLittlewoodandBerttabdatasets. To be due to their differentagesand platemotion.In supportof achieve overlap with the other Antarctic poles by tilting this possibility,we note that eight polesfrom North America requiresa minimumtilt of 31o aroundan axisnearlyperpendic- with U-Pb agesbetween 1087 Ma to 1109 Ma (see Table 3) ular to the inferred trace of the shear zone. There is no field differ by up to 30ø. We thusview the discrepancy betweenthe two data sets as being largely due to poor age control and a evidenceto supportsucha deformation. We concludefrom the above analysisthat within the admit- limited data base and concludethat Coats Land, the tedly rather poor constraintsof available data, large-scale Grunehogna ProvinceandKalaharicratoncouldhavebeenpart tectonicdisturbanceis probablynot the sourceof the disparity of the same plate during this time interval. Failure of the betweenthe Coats Land and DronningMaud Land poles. We CoatsLand pole to averagepaleosecularvariationcould also thusinfer that the differencesare principallya consequence of account for thediscrepancy of thetwodatasetsbut,asargued

differingagesof magnetization.Failureof the CoatsLand above, we believe this is not the case. Figure 11 showsthe.apparentpolar wanderpath(APWP) for pole to averagepaleosecularvariationcould also be a factor coordinates for thetimefrom1200 but, as noted above, the overlap of predicted and observed NorthAmericain present angulardispersionsof the Coats Land VGPs implies this is to 1000 Ma basedon the compilationsof Dunlop and Stirling not the case. [1985] and Costanzo-Alvarez et al. [1993]. The pathshownis

unrotated poles

Sampling Sites 1200

Euler Pole

Ma

90øW

90øE

rotated poles

Euler Pole

COATS

1000

Ma

Figure 11. Generalizedtrend of the North American apparentpolar wander path from ---1200to ---1000Ma and the MesoproterozoicAntarctica poles (Coats Land (COATS), Ahlmannryggen(AHL), Borgmassivet (BORG)) in presentcoordinatesand after rotation around the Euler pole for the SWEAT reconstruction. Antarctica and Australia are shown in the SWEAT configuration[Dalziel, 1991]. Dark shadingindicatesthe Grenville ProvinCe of North America and its proposedcontinuationinto East Antarctica [Dalziel, 1991].

Mollweideprojection.

7898

GOSE ET AL.: PALEOMAGNETISM AND PROTEROZOIC U-Pb AGES, ANTARCTICA

a schematic,free-handfit meantonly to showthe generalage

discrepancy;the angular distance between our pole and the

trend;detailed,annotated pathsarepresented by Harlanet al. North Americanpolesof the sameage is far in excessof what [1994]andPark and Gower[1996].Alsoplottedis ourmean couldbe ascril•edto suchan error. The age of magnetizationfor the Borgmassivetand Ahlpole positionfrom southernCoatsLand and the two Proterozoicpolesfrom westernDronningMaud Land in mannryggenp61esis not well enoughdeterminedto definipresent-day coordinates, as well as in theirpositions after tively argue how their pole positionsrelate in detail to the juxtaposing EastAntarctica withLaurentia as proposed by North AmericanAPWP, but at face value, they alsodo not fall

Dalziel[1991]. Therotatedpolepositions arederivedfromhis SWEAT reconstruction, which is basedon the matchof the

on the APWP where they should accordingto the SWEAT reconstruction. These data thus seem to imply either that the

piercingpointsformedby the apparent truncation of the SWEAT hypothesisis incorrector that one major constraint for the reconstruction,the Grenville front correlation,is in structiondefinesan Euler pole at 13øS,300øEanda rotation error. Given the wide rangeof supportingdatathat have been

Grenville belts in Laurentia and East Antarctica. The recon-

angleof f2 =-135ø. In present-day coordinates, theAntarctic offered sincethe hypothesiswas first suggested[Rosset al. polesareverydifferent fromtheNorthAmerican data,whereas 1992; Stump,1992;Powell et al., 1993;Borg and DePaolo, 1994; Ernst et al., 1995; Park et al., 1995; Young, 1995] and

in the SWEAT reconstruction, they fall on the 1000 Ma segment of the North AmericanAPWP. Althoughthe overlapof

the wealth of global observationsthat juxtapositionof

theCoatsLandp01ewiththeNorthAmerican APWPappears to

Laurentia and East Antarctica-Australia seemsto explain (see

be a positive result,its position is morethan60ø awayfrom summaryby Dalziel [1995]), we deemthe latter possibility thecirca1100Ma partof thepath,wherethepoleshouldfall more likely and exploreit below. A principal tenet of the piercing point correlationof on the basisof the isotopicagesreportedabove. Neither verticalaxis rotationof the samplingareanor any reasonable Dalziel [1991] was that the circa 1.1 Ga rocks of the Coats tilt canbringthe CoatsLandpolesignificantly closerto the Land Nunataks and those of the Maudheim and Grunehogna 1100Ma polesof NorthAmerica in theSWEATfit. Likewise, provincesto the northin westernDronningMaud Land were

failureto average secularvariation cannotaloneexplainthe part of the EastAntarcticcratonby the endof Grenvillianoro-

B

A

WEST

u.rE

GONDWANA

GONDWANA I

Kalahari 30ø/ CMG

•

,Antarctica i. i

Kalahari

• •' C/ 30ø• •

Figure 12. (a) Rodinia reconstructionas proposedby Dalziel [1991, 1992a]. The Coats Land-MaudheimGrunehogna(CMG) province, the Ellsworth-Whitmore Mountain Block (E), and the Kalahari craton of Africa are part of East Gondwana(stippled). Stereographic projection. (b) Proposedplate reconstruction for-1100 Ma that makes the Coats Land-Maudheim-Grunehogna(CMG) province, the Ellsworth-Whitmore Mountain Block (E), and the Kalahari cratonpart of West Gondwana.Geologicaljustificationfor this reconstructionis given by Dalziel [1992b, 1994]. In this scenario,the CoatsLand, Ahlmannryggen,Borgmassivetpolesfrom Antarcticaand the mean Kalahari pole fall preciselyonto the Laurentianapparentpolar wander path of the

correspondingage. The continentsare shownwith their presentcoastlines for ease of recognitiononly. Stereographic projection. (c) Identification of paleomagnetic poles which are shown with their 95% confidenceerrors. COATS, CoatsLand; AHL, Ahlmannryggen;BORG, Borgmassivet; KAL, Kalahari; LS, Logan Sills; CLD, Coldwell; O1, lower Osler; Ou, upper Osler; PL, Portage Lake; MI, Michipicoten; MA, MamainsePoint; CH, CopperHarbor.SeeTable 3 for referencesand ages.

GOSE ET AL.: PALEOMAGNETISM AND PROTEROZOIC U-Pb AGES, ANTARCTICA

genesis. On the basis of studiesthat linked the geologyof western Dronning Maud Land to that of southernAfrica [Groenewald et al., 1991; Moyes et al., 1993; Jacobset al., 1993], this in turn implied that the Kalahari cratonof Africa was part of East Gondwana,as the Archeanand Proterozoic rocks of the GrunehognaProvinceon the continentalmargin of East Antarctica were clearly part of that craton prior to the Mesozoic opening of the southwestIndian Ocean. Rodinia reconstructions of Dalziel [1992a, b, 1995] (Figure 12a) reflect this premise. Our new data suggestthat it is incorrect. The structural relationship of the Coats Land Nunataks to the Grunehogna Province cannot be directly established,as the two are separatedby over 1000 km of the East Antarcticice sheet(Figure 1). The generalagreementof the CoatsLand pole with the poles from the Borgmassivetand Ahlmannryggenis permissive of the idea that within present-day Antarctica, Coats Land and western Dronning Maud Land were part of a single tectonicunit along the margin of the presentWeddell Sea by about 1.1 Ga. We refer to this terrain herein as the CoatsLand-Maudheim-Grunehogna (CMG) province. The lack of overlapof the CMG poleswith polesof the sameage on the North American

APWP

after restoration to the SWEAT

recon-

struction (Figure 11) suggests, however, that the CMG provincemay not have been part of the East Antarcticcraton during the same time period. This implies that the contact between the CMG

and older rocks of the East Antarctic craton

is younger than 1.1 Ga and not Grenvillian in age, an inferencesupportedby recentsuggestions of a continuationof the Neoproterozoicto Cambrian(?)East African orogenicbelt into East Antarcticaand the Shackletonrange [Shiraishiet al., 1994; Grunow et al., 1996; Tessensohn,1996]. This oro-

genicbeltis believedto markthesuturebetweenEastandWest Gondwana that formed during the final amalgamation of Gondwanaland[Dalziel, 1991; Hoffman, 1991; Stern, 1994]. Given the establishedgeologic ties between the Kalahari craton and the CMG, the proposedcontinuationof the suture

7899

intothe Shackleton rangeoffersthepossibility thattheCMG was part of West Gondwana,not East Gondwana,prior to suturing with theEastAntarcticcratonin latestPrecambrian to Cambrian

time.

Figure 12b showsa late Proterozoicplate reconstruction that couldexplainour paleomagnetic datafrom CoatsLand. This reconstructionis broadly based on the suggestionof Dalziel [ 1991] andHoffman [1991] that Laurentiawas located in Rodinia between East and West Gondwana during Neopro-

terozoic times following Grenville orogenesis.Specifically, it shows the Kalahari craton in the position with respect to

Laurentiathat was suggested by Dalziel [1992b,1994]for the hypothetical latestNeoproterozoic supercontinent, Pannotia, followingopeningof the PacificOcez. p andamalgamation of Gondwanaland alongtheEastAfricansuture.The CoatsLandMaudheim-Grunehogna(CMG) province and EllsworthWhitmore Mountains block of Antarctica are attached to the

Kalaharicratonas part of WestGondwana, ratherthanbeing attached to the East Antarctic craton and hence part of East

Gondwana.We highlightthe Kalaharicratonandthe CMG terranein the reconstruction,leaving the remainderof West Gondwana undifferentiated. This is because the timing of

suturing alongthelatestPrecambrian to Cambrian Pan-African and Brazilide orogenicbelts of West Gondwanaremains

problematic andis certainlyin partyounger than1100Ma [see,e.g., Renneet al., 1990' Hoffman, 1991' Dalziel, 1992b; Grunow et al., 1996]. Also shown in Figure 12b are eight poles from North America ranging in age from 1087 to 1109 Ma as well as the three Antarctic poles and the mean Kalahari pole, the Antarctic and Kalahari poles having been rotated correspondingto the paleogeographic reconstruction (Table 3). In this scenario, the Coats Land pole falls precisely on the North American apparent polar wander path and the Borgmassivet, Ahlmannryggen,and Kalahari poles lie on the same path displaced toward slightly younger ages. This observationis sup-

Table 3. MesoproterozoicPaleomagneticData From Antarctica, Africa, and Laurentia Reference

Rock Unit

Antarctica

Lat Coats Land

Ahlmanryggen Borgmassivet

Long

Frame

Africa

Lat

Laurentia

Long

Lat

80.3øE

68.0øN 116.3øE

8.8øN 6.0øN

59.9øE 53.0øE

64.7øN 61.4øN

54. IøE 42.1øN 188.4øE 39.4øE 35.2øN 185.2øE

Peters[1989] Hodgkinson[1989]

65.4øN

36.1øE

Hargraves et al. [1994]

34.7øN

216.6øE

Reference

22.9øN

Kalahari Mean

58.7øN

Long

190.4øE

this study

Logan Sills

49.4øN 219.9øE

Halls and Pesonen[1982]

Coldwell

50.4øN 204.5øE

Harlan [1993]b

UpperOsler

33.9øN 177.9øE

Halls [1974]

Lower Osier

43.1øN

Halls [1974]

PortageLake Michipicoten

26.5øN 181.2øE 26øN 175øE

Mamainse Point

CopperHarbor

194.5øE

38øN

186øE

22.2øN 180.8øE

Age, Ma a

Halls and Pesonen[1982] Palmerand Davis [1987] Palmer and Davis [ 1987]

Diehl and Haig [1994]

1109 1108