PrecambrianResearch147(2006)79–99 The Tambien Group, Ethiopia: An early Cryogenian (ca. 800–735Ma) Neoproterozoic sequence in the Arabian–Nubian Shield Mulugeta Alenea, Gawen R.T. Jenkinb,∗, Melanie J. Lengc,d, D.P. Fiona Darbyshirec aDepartmentofEarthSciences,AddisAbabaUniversity,P.O.B.1176,AddisAbaba,Ethiopia bDepartmentofGeology,UniversityofLeicester,Leicester,LE17RH,UK cNERCIsotopeGeosciencesLaboratory,BritishGeologicalSurvey,Nottingham,NG125GG,UK dSchoolofGeography,UniversityofNottingham,Nottingham,NG72RD,UK Received19October2005;receivedinrevisedform2February2006;accepted12February2006 Abstract TheTambienGroupinTigrai,Ethiopia,comprisesanumberofinlierseachcontaining2–3kmthicknessofinterbeddedcarbonate andclasticsediments,cappedinoneinlierbyaglaciogenicdiamictite.Arangeofgeochemicalindicessuggestnear-pristineC-and Sr-isotopevaluesarepreservedandthese,togetherwithlithologicalvariations,allowlocalcorrelationbetweentheseinliersand correlationwiththeglobalNeoproterozoicisotopestratigraphy.AcompositesectionoftheTambienGroupshows(cid:1)13Cincarbonate of∼+6‰atitsbase,decreasingupwardstotwinlowsof∼–4‰separatedbyabriefexcursionbacktopositivevalues,thenrisesagain toaplateauof∼+6‰beforefinallydecreasingsharplyto−2‰beneaththeNegashdiamictiteatitstop.Noglaciogenicsediments areobservedassociatedwiththelowertwinnednegativeanomalies.The87Sr/86Srvaluesincarbonateschangefrom∼0.7063in thelowerunitsto∼0.7067intheupperunits.The(cid:1)13Coforganicmatterchangeslittlethroughthesequence(−24.2±1.3‰), withtheresultthattheC-isotopefractionationbetweencarbonateandorganicmatterdecreasesfrom+26to+21‰upthroughthe secondlowincarbonate(cid:1)13Cbeforeincreasingto∼+29.5‰intherestofthesequence.Togetherwithexistingradiometricage constraints,theSr-isotopedataindicatethattheNegashdiamictiteisSturtianinage,andthelowernegativeC-isotopeanomalies appeartocorrelatewiththenon-glaciogenicBitterSpringsStagerecognisedinAustraliaandSvalbard.Accordingtothiscorrelation theTambienGroupwasdepositedintheintervalca.800–735Ma.OurdatareinforcetheemergingviewthatnotallNeoproterozoic negativeC-isotopeanomaliesareassociatedwithglaciation.ThevariationsinC-isotopefractionationswithinoursequencecontrast withthosefromAustraliaandareattributedtodifferencesinlocalenvironmentalvariables,probablytemperature.Thesedifferences areconsistentwiththerecentproposalthattheBitterSpringsStageanomalyistheresultofapairofInertialInterchangeTruePolar Wanderevents[Halverson,G.P.,Maloof,A.,Schrag,D.,Dudas,F.,Hurtgen,M.,inpress.Stratigraphyandgeochemistryofaca800 ManegativecarbonisotopestageinnortheasternSvalbard.Chem.Geol.][Maloof,A.C.,Halverson,G.P.,Kirschvink,J.L.,Schrag, D.P.,Weiss,B.P.,Hoffman,P.F.,inpress.Combinedpaleomagnetic,isotopicandstratigraphicevidencefortruepolarwanderfrom theNeoproterozoicAkademikerbreenGroup,Svalbard.GSABull.],andfurtherhigh-resolutiondatingandpalaeomagneticstudies oftheTambienGroupshouldallowtestingofthishypothesis. ©2006ElsevierB.V.Allrightsreserved. Keywords: Carbonate;Carbon;Oxygen;Strontium;Isotopes;Sturtianglaciation;BitterSpringsStage;TruePolarWander ∗ Correspondingauthor.Tel.:+441162523934;fax:+441162523918. E-mailaddress:[email protected](G.R.T.Jenkin). 0301-9268/$–seefrontmatter©2006ElsevierB.V.Allrightsreserved. doi:10.1016/j.precamres.2006.02.002 80 M.Aleneetal./PrecambrianResearch147(2006)79–99 1. Introduction extendingdeepbelowtheSturtianglacial,perhapsonly recorded elsewhere in Svalbard/East Greenland, Aus- The Neoproterozoic was a critical phase in Earth tralia and Canada (Hill and Walter, 2000). We present evolution marking the transition from an oxygen defi- new C- and Sr-isotope data that define a more com- cient(possiblysulphidic)oceantoanoxygenatedocean pleteisotopestratigraphyfortheserocks,includingatthe moresimilartothepresentday(AnbarandKnoll,2002; base of the sequence a pronounced negative C-isotope Hurtgenetal.,2005).Thiswasaccompaniedbydrastic anomaly that is not associated with glaciogenic rocks. climatechange,possiblyincludingatleasttwo‘Snow- These data allow us to make an improved correlation ballEarth’globalglaciationseachfollowedbyextreme withtheglobalisotopestratigraphy,extendingthisNeo- greenhouseconditions(Hoffmanetal.,1998b;Hoffman proterozoic sequence downwards to at least 800Ma in andSchrag,2002).Whilerecentdatinghasshownasig- age,andprovidesupportforanewmodelfornon-glacial nificanttimegapbetweenthelastinferredglobalglacia- carbonisotopeanomaliesintheNeoproterozoic. tionandthedevelopmentofmulticellularorganismsand theensuingCambrianradiation(Hoffmannetal.,2004; 2. Regionalsetting Condonetal.,2005),itisconceivablethatglobalglacia- tionsmayhavebeencriticalinsettingthesceneforlater The Neoproterozoic basement of central Tigrai, evolutionaryevents(HoffmanandSchrag,2002;Knoll northern Ethiopia (Fig. 1a), is composed of low-grade and Carroll, 1999). However, there is still controversy metavolcanics,slate,carbonatesandplutonics(Arkinet as to the exact number and intensity of glaciations in al.,1971;Beyth,1971;Garland,1972;Alene,1998).It the Neoproterozoic, let alone their causes and effects. is in the south (Nafka terrane) of the Arabian–Nubian Theinferredglobalglaciationsaretemporallyassociated Shield(ANS)thatdevelopedduringtheNeoproterozoic withlargenegativecarbonisotopeanomaliesinthesed- byrifting,arcaccretionandterraneamalgamationpro- imentaryrecord,withtheanomalyeitherstartingatthe cessesintheEastAfricanOrogen(Stern,1994;Meert, top of the glacial deposits, or just beneath them. Inter- 2003; Johnson and Woldehaimanot, 2003). The base- pretations of these carbon isotope anomalies vary, and mentisoverlainbyflat-lyingPhanerozoicsedimentsand they may well be the result of more than one process, volcanics. (HoffmanandSchrag,2002;Schragetal.,2002;Pavlov Two major sequences are recognised in the base- et al., 2005; Kennedy et al., 2001; Jiang et al., 2003; ment(Beyth,1971):theTsalietGroup,anoldermainly HigginsandSchrag,2003).Irrespectiveoftheirorigin, metavolcanic/metavolcanoclasticsequenceandtheTam- thecombinationoftheseanomalieswiththeworldwide bien Group, a younger metasedimentary slate and car- depositionofglacialdepositsprovidesthepotentialfor bonate succession. Beyth (1971) described the bound- globalchemostratigraphiccorrelationintheNeoprotero- ary between the two sequences as ‘probably uncon- zoic(Knolletal.,1986;JacobsenandKaufman,1999; formable’.However,Alene(1998)hasdescribedagra- Walteretal.,2000;Melezhiketal.,2001b;Halversonet dationalcontactbetweentheTsalietGroupandbasalpart al.,2005).However,itisincreasinglybeingrecognised oftheTambienGroup,althoughhigherunitsoftheTam- that some Neoproterozoic sequences contain negative bienGrouponlaptheTsalietGroupindicatingtiltingof carbon isotope anomalies that are not associated with depocentres located on the present inliers during Tam- glaciogenic rocks (e.g. Hill and Walter, 2000; Corsetti bienGroupdeposition.Bothsequenceshavebeensub- and Kaufman, 2003; Zhang et al., 2005; Condon et jectedtotwomajorphasesoffoldingasaresultofN–S al., 2005; Melezhik et al., 2005, Halverson et al., in and E–W regional compression, respectively (Alene, press,Dehleretal.,2005).Theseanomaliescomplicate 1998;AleneandSacchi,2000).D foldedbeddingand 1 chemostratigraphiccorrelationsanddemandalternative producedtightminorfolds(wavelengthofmmtodm), explanationsfortheirorigin. elongationlineationandpervasiveregionalfoliation.D 2 PartoftheNeoproterozoicTambienGroupinnorth- causedlongwavelength(upto8km),upright,openpar- ern Ethiopia was recently remapped by Alene (Alene, allelfoldswithoutproducingasignificantcleavage.D is 2 1998;AleneandSacchi,2000;Aleneetal.,2000).This consideredtobeduetoE–Wdirectedshorteningassoci- Neoproterozoic sequence is one of the least studied in atedwiththeend-phasecollisionbetweenEastandWest Africa,yetliesatacriticalgeotectonicboundarybetween Gondwanaland,andcorrelateswithpost-accretionstruc- the Arabian–Nubian Shield (mostly juvenile crust) to turesdescribedelsewhereintheANS(Abdelsalamand thenorthandMozambiquebelt(mostlyreworkedolder Stern,1996). crust)tothesouth.Existingdata(Milleretal.,2003)sug- Mineral assemblages in the Tsaliet metavol- gesttheTambienGroupcouldrepresentararesequence canics indicate that peak regional metamorphism M.Aleneetal./PrecambrianResearch147(2006)79–99 81 Fig.1. (a)GeologicalmapoftheMaiKenetal–NegashstudyareaincentralTigraimodifiedafterAlene(1998).Insetmapshowsthelocationofthe studyareainnorthernEthiopia.ThePhylliteUnitisthoughtbealateralequivalentoftheTambienGroup.Insomeliteraturethegranitesarenamed astheMarebgranitesuite,butweprefernottousethisnomenclatureasitimpliesapost-tectonictiming,whereasthegranitesherearepre-D2.Age datesareshownonpre-D2granites;sources:NegashandHauzien—Milleretal.(2003),single-zirconPb/Pbevaporationmethod;MaiKenetal —Beyth(1972),K–Arage.(b)DetailedgeologyoftheMaiKenetal–Tsediaareaandsamplelocations(opensquarein(a)),basedon1:50,000 mappingbyAlene(1998).Way-upcriteriaincludestromatoliticlamination,channel-fillandslumpstructuresinthecarbonates. 82 M.Aleneetal./PrecambrianResearch147(2006)79–99 at pumpellyite–actinolite to lower greenschist facies methodageof854±3Maonlow-grademetavolcanics (∼245–375◦C by chlorite thermometry) was attained (Teklay,1997).AsimilarsequencetotheTsalietGroup coevalwithD (Alene,1998;AleneandSacchi,2000). to the NW in Axum yielded ages of 756–806Ma on 1 Webelievethatmaximumtemperaturesexperiencedby syntectonic granites (Sm–Nd and chemical Th–U–Pb the Tambien Group are at the low end of this range zirconisochronmethods;Tadesseetal.,2000)suggest- ◦ <250 C,onthebasisof(a)O-andC-isotopefractiona- ingTsalietGroupdepositionbeforethattime. tionsbetweencalciteanddolomite(Aleneetal.,1999), (b)closetodepositionalcalcite-organicmatterC-isotope 3. TheTambienGroup fractionations(reportedbelow),and(c)theblackcolour of palynomorphs preserved in some samples (Mullins, TheTambienGroupisexposedinfourinliersformed UniversityofLeicester,personalcommunication). byD synclinoria,fromwesttoeast:MaiKenetal,Tse- 2 Radiometric ages of the granites (Fig. 1a), locally dia, Chehmit and Negash (Beyth, 1971; Fig. 1a). Each intruded between D and D (Alene, 1998; Alene and sequencecontains∼2–3kmtruethicknessofsediment 1 2 Sacchi,2000),constraindepositionandearlymetamor- afterdipsfromD foldsareremoved(Fig.2).D folds 2 1 phism of the Tambien Group to >613Ma. The age of causeonlylocalrepetitionsofstratigraphy the Tsaliet Group, and hence a maximum depositional IntheMaiKenetalsection(Figs.1band2)thelitho- age of the Tambien Group, is less well constrained, logic sequence comprises (from bottom to top): Lower with no local radiometric data available. A metavol- Slate (Werii Slate), Lower Limestone (Assem Lime- canic sequence similar to the Tsaliet Group in Eritrea stone),UpperSlate(TsediaSlate)andUpperLimestone to the north gave a single-zircon Pb/Pb evaporation (MaiKenetalLimestone)—namesinbracketsarethose Fig.2. StratigraphicsequenceoftheTambienGroupmetasedimentsintheMaiKenetal,TsediaandNegashsections,northernEthiopia.Unitnames giveninparenthesesarethoseusedbyBeyth(1971).Therangeofthicknessofunitsineachinlierisgiven;thethicknessesplottedarethoseinthe areainwhichsamplesweretaken. M.Aleneetal./PrecambrianResearch147(2006)79–99 83 used by Beyth (1971). In the Tsedia and Chehmit sec- tions,thelowerthreeunitsarerepresentedbuttheUpper Limestone is not present. Although the correlation of theLowerLimestoneisnotstraightforward,therelative stratigraphicpositionandlithologic(composition,grain size) and textural features (degree of deformation and fabric development) suggest that it is the same unit. In allthreesections,theLowerLimestoneisfoundbetween theUpperandLowerSlates. IntheNegashsection(Fig.2)thesequenceisdiffer- ent and it includes (from bottom to top) Negash Slate, Dolomite and Slate, limestone (Negash Limestone), Slate/Pebbly slate (Diamictite Slate). The composition andgrainsizewithonlyweakfabricdevelopmentsug- gestthattheNegashLimestonewouldbebestcorrelated withtheUpperLimestoneintheMaiKenetalsectionand weprovideisotopedatatosupportthisbelow.Garland (1972)groupedthetwolowermostunitsoftheNegash sequence as the Didikama Formation and the three top unitsastheMatheosFormationanddescribedbothfor- mationsasunitsyoungerthantheTambienGroup.How- ever,Beyth(1971),Alene(1998)andBeythetal.(2003) all consider the Negash section as part of the Tambien Group. 3.1. LowerSlate(WeriiSlate) Ablack,brownorgreenishgrey,well-laminatedand stronglyfoliatedslateintheMaiKenetalandTsediasec- tions(Fig.1b).Itcontainsgreywackeandgraphiticbeds. The basal contact with the underlying Tsaliet Group is gradationalwithnomajordiscontinuitybetweenthetwo groups.Greenishgreyslateandgreywackelayersoccur bothsidesoftheboundary,andthereisnomarkeddefor- mationalormetamorphiccontrastacrossthecontact. 3.2. LowerLimestone(AssemLimestone) In the Mai Kenetal synclinorium this is exposed on the western limb as a sub-vertical layer ∼700m thick Fig.3. FieldphotographsandCLimagesofMaiKenetallimestone. (Fig.1b).Itisgreyorblacklimestonewithstromatolitic (a)StromatoliticlaminationintheLowerLimestoneoftheMaiKene- lamination (Fig. 3a), calcite veins and dark stylolites. talsection;MK72samplelocality.(b)ParasiticD2 foldswithsub- Stromatolitic lamination and slump structures indicate horizontal fold axes in Lower Limestone of the Tsedia section. (c) younging towards the Upper Limestone. It has fine- to CathodoluminescenceimageofsampleMK100(LowerLimestoneof medium-grainedcrystallinetextureandis∼90%calcite; theTsediasection),mostofthesampleshowslittleluminescence,but orangeluminescencepicksoutazoneincalcitemicrospar.Widthof the remainder is dolomite, detrital quartz and feldspar, image1.6mm. andminorpyrite.Incathodeluminescence(CL)itshows predominantlydullbrownanhedraltosubhedralcalcite grains,andlesseramountofredluminescentdolomite. Limestonehereismainlyblacktodarkgrey,commonly In the Tsedia synclinorium the Lower Limestone is intercalated with slate and well bedded and laminated; exposed in both limbs. It contains D folds with NNE insomeplacesapliticsillsintrudeit.Thecompositionis 2 trendingsub-horizontalfoldaxes(Fig.3b).TheLower predominantly crystalline calcite (>90%), with detrital 84 M.Aleneetal./PrecambrianResearch147(2006)79–99 albite and quartz. The texture is similar to that of the notmatrixsupportedanddoesnotcontainanyfeatures Lower Limestone in the Mai Kenetal section. In CL diagnosticofglaciogenicorigin,suchasstriatedpebbles. it shows a largely dark matrix containing orange lumi- nescentandzonedcalcitehavingrhombic,rectangular, 3.6. Dolomite–Slateintercalation andlesscommonlysub-circulargrainshapesoftenwith dark cores (Fig. 3c). Non-luminescent, coarse-grained, A 1200m thick unit comprising white to pink sub-idiomorphicalbiteandquartzshowingpreferredori- dolomite intercalated with dark grey, black and green entationarealsocommon. slateatNegash.Thedolomiteisfinetocoarsegrained, consistingofover90%dolomitewithcalciteanddetri- 3.3. UpperSlate(TsediaSlate) talquartzandalbite.Itdisplaystightfoldandrefolded structuresaswellasstromatoliticlaminations. Alighttodarkgrey,thinlybeddedandwell-laminated andfoliatedslate.AtTsediaitoccursinthecoreofthe 3.7. NegashLimestone syncline overlying the Lower Limestone. The onlap of theUpperSlateontotheTsalietGroupontheeastside Thisisamassiveblacklimestone.Ithasgradational oftheMaiKenetalsynclinorium(Fig.1b)mustrelateto contacts with the underlying dolomite and overlying tilting of the depocentre during deposition of the Tam- slate.Itisfine-grained,laminated,fractured,locallystro- bienGroupand/orfaultingofthecontact. matolitic,andcontainsquartzandcalciteveins.InCLit showsdarktodarkbrownluminescentmicrocrystalline 3.4. UpperLimestone(MaiKenetalLimestone) zones (or bands) strongly speckled with black (non- luminescent) dots alternating with zones without black This forms a sub-vertical NNE trending, massive dots.Thepresenceoftheblackdotsandtheabsenceof black limestone layer in the core of the Mai Kenetal well-recrystallisedcalcitegrainsaresimilarwiththatof synclinorium;youngingisdemonstratedbyslumpstruc- theUpperLimestoneatMaiKenetal. tures.Itcontainsminorsub-verticalfoldaxes,andatthe baseitisintercalatedwiththeUpperSlate.NeartheN–S 3.8. DiamictiteSlate trending fault (Fig. 1b) it is locally brecciated. It con- tainsdarkcarbonaceousseamsandcalcareousfragments Overlying the Negash Limestone is a light to dark probably representing redeposited carbonate flakes. It greyslatethatformsthecoreoftheNegashsyncline.At is fine-grained and contains abundant calcite veins. It thebasetheslateisintercalatedwithlimestonebedsand comprises mainly calcite (∼95%) with detrital quartz towardsthetopitbecomesdiamictiticcontainingabun- andalbite,pyriteand±hydratedironoxide.InCL,sam- dant matrix-supported pebbles and cobbles of variable pleMK77showsamicriticcalcitematrix,whileMK78 size (mostly 1–15cm diameter) and shape, sometimes shows‘fragmental’calcitetexturewithmainlybrownto with striated surfaces (Miller et al., 2003). The clasts dark,andtoalesserextentorange,calciteluminescence arepolymict,includingblacklimestone,metavolcanics, occurring in different forms including dark long frag- greywacke,quartzitesandgraniticrocks.Thefeaturesof ments and circular shapes mostly speckled with black thediamictiteareconsistentwithaglacialorigin. dotsandhavingorangerims.Thehighlyveinedsample MK79containssporadicpatchesoforangeluminescent 4. Analyticalmethods calciteaswellasafewfine-grained,circularcalcitesthat exhibitbrightorangeluminescence;calciteveinsbelong 4.1. Sampling toatleastfourgenerations. Because all samples have been metamorphosed and 3.5. NegashSlate some parts of the sequence are veined, we have sam- pled in order to characterise any post-D vein material 1 Adarkgreyandgrey–greenslateoverlyingtheTsaliet present,identifiedasbeingcoarselycrystalline(typically Group in the Negash synclinorium. It contains 1–2m 1–5mm grain size) and cross-cutting the metamorphic thickbedsofspottypurpleslate,well-laminatedbands, fabric, as well as the least altered rock away from the graphiticlayers,quartzite,conglomerateandgreywacke. veins.Thisstrategyenablesustoidentifytheleastaltered Beythetal.(2003)suggestedapolymictconglomerate samples at any horizon by monitoring the direction in in this unit could represent a lower glacial diamictite, which the geochemistry of samples was altered by the butwedonotconcurwiththissuggestionbecauseitis lateveinfluids.Weatheredcrustsandmarkerpensample M.Aleneetal./PrecambrianResearch147(2006)79–99 85 numberswereremovedfromallsamplespriortocrush- eter.Isotopevalues((cid:1)13C,(cid:1)18O)arereportedaspermil ing, ensuring no modern contamination. Samples for (‰)deviationsoftheisotopicratios(13C/12C,18O/16O) analysisofhomogeneousrockswereobtainedbycrush- relative to V-PDB using a within-run laboratory stan- ing small volumes, whereas veined or laminated rocks dardcalibratedagainstNBSstandards.Overallanalyti- were sampled on unweathered cut or broken surfaces calreproducibilityisnormallybetterthan0.1‰for(cid:1)13C withadiamond-tippeddrilltoobtain∼200mgsamples and(cid:1)18O(2σ). ofveinandhostrockmaterial,andanydifferentcoloured laminae.Thesesampleswerethensplitforchemical,and 4.5. Organicmatterδ13C C-,O-andSr-isotopeanalysis. Carbonatewasremovedfromsamplesbyreactionof 4.2. Calcitechemistry between19and∼100gofpowderedrockwith5%HCl Powderedsamples(∼100mg)wereweighedandthen overnight.Theresiduecontainingtheorganicmatterwas washed and dried. In some samples a tiny amount of leachedin1Maceticacidfor2h—thisisbelievedto organic scum was released on dissolution and this was quantitativelydissolvecalcite,butleavedolomiteunre- decantedoff.Lossofthis,probablyinsignificantamount acted.Solutionswerequantitativelypipettedoffforanal- of material, which is likely to have had lower (cid:1)13C ysisandresiduesweredriedandweighedtocalculatethe thanthebulksample,meansthatmeasured(cid:1)13Cvalues weightofsampledissolved.Thisweightwasthenusedto ontheresiduecouldrepresentmaximumvaluesforthe calculateconcentrationsinthedissolvedcalcite.Acetic total organic matter. 13C/12C analyses were performed acidsolutionsweredrieddownandredissolvedin10% by combustion in a Carlo Erba NA1500 on-line to a HClforanalysiswithaJobinYvonHoribaUltima2opti- VGTripleTrapandOptimadual-inletmassspectrometer, calemissionICP.Precisionandaccuracyareestimated at±5%. with(cid:1)13CvaluescalculatedtotheV-PDBscaleusinga within-runlaboratorystandardscalibratedagainstNBS- 19 and -22. Replicate analysis of well-mixed samples 4.3. Calcite87Sr/86Sr indicatedaprecisionof±<0.1‰(1S.D.). Approximately 20mg of powdered sample were treatedwithRomil©ultrapure1Maceticacid.Aftercen- 5. Preservationofprimaryisotopesignatures trifuging, leachates were dried down and the residue dissolvedin2.5MHCl.Strontiumwasseparatedbycon- Because our samples have experienced deformation ventionalcationexchangetechniquesusingBiorad©AG and low-grade metamorphism it is imperative that we 50W-X8ionexchangeresin.Strontiumwasloadedonto demonstratewearerecordingdepositionalisotopesigna- rhenium (Re) filaments together with a tantalum oxide turesratherthanalteredvaluesimposedpost-deposition. (TaO) activator following the method of Birck (1986). ThisproblemappliestoallNeoproterozoicsamplesbut Isotope ratios were measured on a Finnegan MAT Tri- isoftenperceivedtobemoreacuteforsamplesthathave ton multicollector mass spectrometer operated in static beenmetamorphosed.However,thismaynotnecessarily mode. Nineteen analyses of NBS 987 measured dur- bethecasebecause,oncetheyhaverecrystallisedduring ing the period of study yielded a mean 87Sr/86Sr of theearlieststagesofmetamorphism,itiswidelyagreed 0.7102542±0.0000092 (2σ). For consistency all the that limestones and marbles exhibit some of the low- measuredratiosarenormalisedrelativetotheaccepted est permeabilities of any metamorphic rocks (Holness value of 0.710248 for NBS 987. Replicate determina- and Graham, 1995) and are often metamorphosed as tions (n=11) of the North Atlantic seawater standard closed systems. Even when infiltrated by fluids along yielded0.7091753±0.0000062(2σ). veinsthedegreeoffluid–rockinteractionandexchange maybesmallandlocalisedsothatisotopesystemsinthe 4.4. Calciteδ13Candδ18O rockarerelativelyresistanttoexchange.Decarbonation reactionsbetweensilicatesandcarbonatesareknownto Samples were ground in agate and equivalent of causedecreasesin(cid:1)13Cand(cid:1)18Oinhighergrademeta- 10mg of carbonate was reacted with anhydrous phos- morphicrocks(Schidlowski,1987;Nabelek,1991),but ◦ phoricacidinvacuoovernightataconstant25 C.The providedthatlimestonescontainlowconcentrationsof liberated CO was cryogenically separated from water silicatematerialand/orhavenotbeenheatedtotemper- 2 vapour under vacuum and collected for analysis. Mea- aturesatwhichdecarbonationtakeplace,thentheyare surementsweremadeonaVGOptimamassspectrom- likely to retain isotope signatures from the diagenetic 86 M.Aleneetal./PrecambrianResearch147(2006)79–99 realm. Indeed, Melezhik et al. (2001a) detailed exam- tinevalues.Theselowvaluesareduetointeractionwith ples of amphibolite grade marbles that have retained low (cid:1)18O meteoric or metamorphic fluids that formed closetopristineC-andSr-isotopesignatures.Therefore, post-D vein samples MK72v and MK79v (Table 1). 1 the end result may be that most of the processes that Two of the lowest (cid:1)18O samples in the Upper Lime- couldhavealteredtheisotopecompositionsinourrocks stones(MK79and80)areclosetotheprominentpost-D 2 probablyoccurduringdiagenesis—incommonwithall N–Sfault(Fig.1b)suggestingthatthiswasaconduitfor Neoproterozoiccarbonates.Manyofthevariablescon- low(cid:1)18Ofluidsandprovidingafurtherconstraintonthe trolling the degree of alteration of isotope ratios, such timingoffluidalteration.The(cid:1)18OvalueofMK72vcal- as effective fluid/rock ratios, are the same whether in cite of –20.6‰ V-PDB (+9.68‰ V-SMOW) would be themetamorphicordiageneticrealms,andmonitorsof inequilibriumwithfluidof∼0‰V-SMOWat∼200◦C alteration that are proxies for these variables, such as (Zheng,1999),butveintemperatureswouldhavetohave ◦ Mn/Sr,shouldbeexpectedtoidentifyalteredisotopeval- beenaslowas100 Cforthecalcitetobedepositedin uesirrespectiveoftheirtimings.Theoreticaloverviews equilibriumwithafluidof−8‰V-SMOW—theupper of the relative robustness of carbonate isotope systems limitofNeoproterozoicmeteoricwaterifseawaterwas duringfluid–rockinteraction(BannerandHanson,1990; −8‰.ThuseventhefluidthatformedMK72vwasprob- JacobsenandKaufman,1999)showthat,onthebasisof ably not dominated by seawater or meteoric water, but typicalconcentrationsofC,OandSrinrocksandfluids, waspositivelyshiftedbyfluidrockexchange;thissug- itistobeexpectedthat(cid:1)13Cshouldbemorerobustthan gestsrelativelylowfluid/rockratios.NotethattheMK72 87Sr/86Sr,whichisinturnmorerobustthan(cid:1)18O.Thus rockhasretainedahigher(cid:1)18Ovaluethantheveinsin if(cid:1)18Ocanbeshowntobepristine,thenitislikelythat the same rock, indicating that either water/rock ratios Sr and C isotopes will also be, etcetera; therefore, we were low (<1; Jacobsen and Kaufman, 1999), or that dealwith(cid:1)18OvaluesfirstandthenproceedtoSrandC exchangewaskineticallylimited,orsomecombination isotopes. ofthetwo.SinceSr-isotopeexchangerequiressimilaror largerwater/rockratioscomparedtoO-isotopeexchange 5.1. δ18O to become significant (Jacobsen and Kaufman, 1999), carb andSr-exchangekineticsarelikelytobethesamerate Rock (cid:1)18O values vary between −4.7 and orslowerthanoxygenisotopeexchange,thislendssup- carb −13.9‰ with an average of −9.6‰ (Table 1). Vein porttopreservationofdepositionalSr-isotopesignatures (cid:1)18O values are lower on average (−12.6‰), rang- inthesesamples. carb ing between −7.9 and −20.6‰. Rock (cid:1)18O values carb are distinctly lower than in modern marine limestones 5.2. 87Sr/86Sr (∼0‰), but both measurements on pristine carbonates throughthePhanerozoic(Veizeretal.,1999)andmod- Two Upper Limestone samples ∼600m apart ellingstudies(Wallmann,2001)areincreasinglyindicat- (∼225mstratigraphically)havestrontiumisotoperatios ingthatNeoproterozoicseawaterwassignificantly18O- thatareidentical(mean:0.706732±0.000002;Table1) depleted compared to modern seawater. It is suggested withinanalyticalerror.TwosamplesoftheLowerLime- that pristine Neoproterozoic carbonates may have had stone at Mai Kenetal and Tsedia have a mean ratio (cid:1)18Oaslowas−8‰,withvariationsaroundthisvalue of 0.7062±0.0001; one vein in the Lower Limestone resulting from local depositional environment, temper- showsaslightlylowervalue(0.70603). ature and glaciation. Certainly Hoffman and Schrag Several criteria have been used to assess the preser- (2002) and Kasemann et al. (2005) present data from vation of Sr-isotope ratios in Neoproterozoic carbon- carbonates around the Ghaub glacial in Namibia that ates (e.g. Asmerom et al., 1991; Kaufman et al., 1993; show parallel secular trends in (cid:1)18O and (cid:1)13C, where Kaufman and Knoll, 1995; Jacobsen and Kaufman, the(cid:1)18Odataareinterpretableintermsofdepositional 1999; Fairchild et al., 2000; Melezhik et al., 2001b; temperature variations and thus could be primary, and Thomasetal.,2004).Low(cid:1)18O valuesmayindicate carb thelatterauthorsarguethatthisisthecase.These(cid:1)18O exchangewithmeteoricormetamorphicwater,although valuesareintherange−2to−11‰,similartoestimates asdiscussedabove,theinitialstartingpoint,andthere- ofMesoproterozoicprimarymarinecalcite(cid:1)18Oof−6 fore what represents a ‘low’ value is still debatable. to−10‰(FrankandLyons,2000;Kah,2000).Thusitis Furthermore,meteoricwaterwillnotcontainsignificant possiblethatsomeofourhigher(cid:1)18Ovaluesarecloseto Srunlessithasinteractedwithrocksearlierinitsflow- primaryvalues,butmanyarebelowtypicalcut-offs(e.g. path,andthenitsacquired87Sr/86Srratiomayormaynot >−11‰;Kaufmanetal.,1993)usedtodistinguishpris- bedifferentfromtherocksthatweareexamining.Hence Table1 IsotopedataandelementconcentrationsinaceticacidsolublefractionofTigraiNeoproterozoiccarbonates SampleNo. Unit Description Heighta (cid:1)13Ccarb (cid:1)18Ocarb (cid:1)13Corg εTOCc TOCd 87Sr/86Sr Ca Fe Mg Mn Sr Mn/Sr Mg/Ca Residuee (m) (‰b) (‰b) (‰b) (‰) (%) (wt.%) (ppm) (ppm) (ppm) (ppm) (%) MaiKenetal MK77 U.Lst. Whole-rockblack 2550 +5.8 −8.5 −23.7 +30.2 0.23 0.7067329±24 40.23 2045 2229 160 2450 0.065 0.006 7.2 limestone MK76v U.Lst. Veincontaining 2325 +6.0 −7.9 38.91 4363 3008 235 5362 0.044 0.008 30.7 calciteandquartz MK76r U.Lst. Whole-rockblack 2325 +5.8 −9.6 −22.1 +28.4 0.14 0.7067306±24 40.60 3621 2838 283 2958 0.096 0.007 11.7 limestone MK78v U.Lst. Veincontaining 2250 +5.0 −8.7 38.40 2967 3272 327 7087 0.046 0.009 24.0 calciteandquartz MK78r U.Lst. Whole-rockblack 2250 +5.2 −9.4 −23.8 +29.7 0.06 39.07 2125 2039 529 2752 0.192 0.005 14.2 limestone M MK80 U.Lst. Whole-rock 2025 +5.3 −11.4 . A brecciatedblack le limestone ne MK79vl U.Lst. Latercalciteveinin 1950 +4.8 −14.4 36.79 1204 2992 93 2360 0.039 0.008 2.2 et MK79ev U.Lst. sEaamrlpielrecMalKci7te9veinin 1950 +4.5 −14.1 41.39 1620 3758 109 3996 0.027 0.009 0.5 al./ P sampleMK79 re MK79r U.Lst. Whole-rockblack 1950 +5.7 −13.9 −23.2 +29.6 0.08 39.02 1996 7769 104 2578 0.041 0.020 11.1 ca m limestone b MK75gl L.Lst. Whole-rocklight-grey 1250 −0.8 −9.2 40.46 354 2470 84 189 0.443 0.006 1.0 ria limestone n MK75b L.Lst. Blacklaminain 1250 −1.0 −9.7 Re s sampleMK75 ea MK74 L.Lst. Whole-rockgrey 1075 −1.0 −9.5 0.003 41.11 227 2924 90 160 0.565 0.007 1.7 rc h limestone 1 MK73 L.Lst. Whole-rockgrey 1040 −0.9 −8.1 −24.7 +24.5 0.01 0.7061116±25 41.31 346 3598 80 165 0.483 0.009 3.0 47 limestone (2 MK72v L.Lst. Calciteveininsample 1005 −1.9 −20.6 0.7060252±27 42.44 206 1762 245 295 0.829 0.004 2.4 00 6 MK72 ) MK72r L.Lst. Whole-rockgrey 1005 −1.0 −11.5 38.13 450 2042 148 156 0.949 0.005 3.2 79 – limestone 9 9 MK71 L.Lst. Whole-rockblack 900 −2.0 −8.9 −27.0 +25.7 0.09 38.25 4486 89190 122 229 0.532 0.233 19.6 limestone Tsedia MK97 L.Lst. Whole-rockblack 1329 −3.0 −11.3 −24.6 +22.2 0.04 39.24 1527 3373 181 2733 0.066 0.009 9.6 limestone MK98b L.Lst. Blacklaminain 1280 −4.3 −11.9 −24.9 +21.0 0.10 37.66 2394 3939 1015 2052 0.495 0.010 52.6 sampleMK98 MK98g L.Lst. Greylaminainsample 1280 −4.5 −11.7 MK98 MK99 L.Lst. Whole-rockgrey 1159 −2.9 −8.1 40.01 1739 2585 725 1050 0.691 0.006 5.4 limestone MK96 L.Lst. Whole-rockblack 1086 −3.9 −11.1 limestone MK100 L.Lst. Whole-rockblack 964 −0.7 −12.2 −24.1 +24.0 0.09 0.7063091±20 40.65 1086 3678 87 2034 0.043 0.009 2.3 limestone 8 7 88 M.Aleneetal./PrecambrianResearch147(2006)79–99 eue d Resi(%) 1.9 4.1 9). 9 Ca 4 6 19 Mg/ 0.00 0.00 al., et Sr 1 6 es Mn/ 0.02 0.02 Hay ( m) 5 3 Corg Sr(pp 279 267 (cid:1)13 − Mnppm) 59 69 Ccarb ( (cid:1)13 ≈ m) 9 1 } Mg(pp 170 224 −1 0)] Fe(ppm) 662 770 +100 Ca(wt.%) 41.66 36.51 (cid:1)130)/(Corg 0 0 1 + 8786Sr/Sr {(cid:1)13[(Ccarb Ftiiogn.a4l.dCatraos(ssmplaoltlsoopfecnarsbyomnbatoels(cid:1))18frOomvsM.(ail)le8r7Sert/a86l.S(r2,0(b0)3(cid:1))1a3nCd.AAldednie- dOC%) 3=10 etal.(1999). T( 1) − oxygenisotopeevidenceofinteractionwithmeteoricor εcTOC(‰) α(TOC imneStar-misoortpohpiecrwataiotesr.iOsunrortoackprsiaomripelveisdfernocmeMofaailKteeranteitoanl 3 (cid:1)13Corgb(‰) er=10 aitneddTrasnegdeiaothfa(cid:1)t1w8Oerevamlueeass,uwreidthfoornSlyr-itshoetoTpseedhiaavesaamlipmle- (cid:1)18Ocarbb(‰) −10.0 −4.7 −9.7 −7.5 −8.8 −6.8 −6.6 ganicmatt b(f∼reoi0mn.g0M0<0a−0i91K)1e‰nleest(saFlri(agMd.i4Koag7)e2.nvTi)chhetahhsaiang8h7MlSyKr1/8876O3S,-rdtrheapetiloevteseildnig-vhferteilnye (cid:1)13Ccarbb(‰) +4.0 +6.8 +3.7 +5.3 +4.9 +6.6 +6.3 eandor rinogckthsaatmthpelevjeuisntflaubiodvceoiutldinhtahveesptrraetviigoruasplhyy,instuergagcetsetd- nat withthelow87Sr/86SrvolcanicsoftheunderlyingTsaliet aHeight(m) 2360 2320 2240 2240 2200 2200 2120 group. encarbo aGlr.o(2u0p0.N3)ehgaavsheLraitmheersthoingehesarm(cid:1)1p8lOestahnaanlyosuerdsbaymMplielslearnedt en we 87Sr/86SrvaluesoverlapwithourUpperLimestonesam- ContinuedTable1() SampleNo.UnitDescription NegashNW121NegashWhole-rockblackLst.limestoneNW122NegashWhole-rockblackLst.limestoneNW123vNegashCalciteveinLst.NW123rNegashWhole-rockblackLst.limestoneNW124fNegashFine-grainedblackLst.limestoneNW124cNegashCoarse-grainedblackLst.limestoneNW125NegashWhole-rockblackLst.limestone aCalculatedheightabovebaseofTambibRelativetoV-PDBstandard.εcistheC-isotopefractionationbetTOCdTotalorganiccarboninrocksample.eAceticacidinsolubleresidue. 8hpsaciamiss7asoneloeeSmyiuuftIgstrloenafo,p/hdlmp8lnllbdti6ieerkeumsoSeoseprtu.rlfraelepahnytsfsarhiaiattssotoitrvleco8aslnaes7tonoerbiSawegrsdfercrelitc/arpaths8xoyhpr6ihecnggsihSynhtctneyralaaariltldifinorh(lnrboiFgsceasoiehaeinalrgsnniitagSgc.iantneoahrt4oganetiwaettRen,()scilJhb.ooto8eihe/fT7rwtSnnrSohhikercsrueitlied/geanri8premht6iLoeetnSdr8ptsowcior7seatwrSaliirleoobt.laerahv,ml/trpse2e8utihL6irh0rnttaSeiehe0agmssra1aessttd)measahi.smuntmesoThtrdouiuiepShlnncsinalirnhgeees-rt