Fairchild, I. J., Bonnand, P., Davies, T., Fleming, E. J., Grassineau, N.,Halverson, G. P., ... Stevenson, C. T. E. (2016). The Late Cryogenian WarmInterval, NE Svalbard: Chemostratigraphy and genesis. PrecambrianResearch, 281, 128-154. https://doi.org/10.1016/j.precamres.2016.05.013

Peer reviewed version

License (if available):CC BY-NC-ND

Link to published version (if available):10.1016/j.precamres.2016.05.013

Link to publication record in Explore Bristol ResearchPDF-document

This is the author accepted manuscript (AAM). The final published version (version of record) is available onlinevia Elsevier at http://www.sciencedirect.com/science/article/pii/S0301926816301577. Please refer to anyapplicable terms of use of the publisher.

University of Bristol - Explore Bristol ResearchGeneral rights

This document is made available in accordance with publisher policies. Please cite only the publishedversion using the reference above. Full terms of use are available:http://www.bristol.ac.uk/pure/about/ebr-terms

1

TheLateCryogenianWarmInterval,NESvalbard:chemostratigraphyandgenesis

IanJ.Fairchilda,1,PierreBonnandb,TesniDaviesa,EdwardJ.Fleminga,c,NathalieGrassineaud,GalenP.Halversong,MichaelJ.Hambreyh,EmilyM.McMillana,ElizabethMcKaya,IanJ.Parkinsoni,CarlT.E.StevensonaaSchool of Geography, Earth and Environmental Sciences, University of Birmingham, Birmingham B15 2TT, UK bDepartment of Earth Sciences, South Parks Road, Oxford OX1 3AN, UK cCASP, West Building, 181A Huntingdon Road, Cambridge, CB3 0DH, UK dDepartment of Earth Sciences, Royal Holloway University of London, Egham, Surrey TW20 0EX, UKgDepartment of Earth and Planetary Sciences/Geotop, McGill University, 3450 University St., Montréal, QC,

H3A 0E8, Canada hDepartment of Geography and Earth Sciences, Aberystwyth University, Aberystwyth, Ceredigion SY23 3DB,

UK iDepartment of Earth Sciences, University of Bristol, Wills Memorial Building, Queen’s Road, Bristol, BS8

1RJ, UK ABSTRACTTheLateCryogenianWarmInterval(LCWI)referstoanon-glacialintervalthatseparatespresumedrepresentativesoftheSturtianandMarinoanpanglaciations.Itsdurationispoorlyconstrainedradiometricallyanditsdepositsarerelativelypoorlyknowninmostgeographicregions.Thispaperaimstoconstraintheduration,palaeoenvironmentsandpetrogenesisofsuchdepositsintheclassicregionofNESpitsbergen,Svalbard.Thesuccessioncomprisesa200-205mdolomiticshale(MacdonaldryggenMember,knownasE3,oftheElbobreenFormation)overlainbyooliticdolomiteSlangenMember(E4),15-25mthick,withlimestonedevelopedattopandbaseofE3inthesouthofthearea.TheassumedagecontextofthesuccessionhasbeenconfirmedbythepresenceofatypicalSturtiancapcarbonateprofileofnegativetopositiveδ13C,andprimarySrisotopecompositionsofbasalE3limestones<0.7072andofupperE3limestonesof0.7076. AtthebaseofE3,interstratificationofcapcarbonatewithice-raftedandredepositedglacialsedimentsoccurs.Earlydiageneticstabilizationofcarbonatemineralogyfromaprecursor,possiblyikaite,tocalciteordolomiteisinferred.E3ispredominantlydolomiticsilt-shale,withsub-millimetrelamination,lackingsandorcurrent-relatedsedimentarystructures.Thinfinelaminaearepartlypyritizedandinterpretedasmicrobialmats.Dolomitecontentis25-50%,withδ13Cvaluesconsistentlyaround+4‰,avalueattributedtobufferingbydissolutionofaprecursormetastablecarbonatephase.Localcalcitecementassociateswithlowδ13Cvalues.Thecarbonatesformsilt-sized,chemicallyzonedrhombiccrystalsfromanenvironmentwithdynamicallychangingFeandMn.Three-dimensionalreconstructionsofcm-scaledisturbancestructuresindicatethattheyrepresenthorizontallydirectedsock-likefolds,developedbyreleaseofoverpressureintothinsurficialsedimentoverlyinganearly-cementedlayer. AshoalingupwardsunitnearthetopofE3displayscalciumsulphatepseudomorphsindolomiteinthenorth,butstorm-dominatedlimestonesinthesouth,bothbeingoverlainbyperitidalooliticdolomites,exposedunderthesucceedingWilsonbreenglacialdeposits.ThereisnoTrezonaδ13Canomaly,possiblyimplyingtop-truncationofthesuccession. Regular0.5m-scalesedimentaryrhythms,reflectingsubtlevariationsinsedimenttextureorcompositionoccurthroughoutE3andareinterpretedasallocyclic.Theyarethoughttobemainlyprimaryinorigin,locallymodifiedslightlyduringearlydiageneticcementation.Rhythmsareproposedtorepresentca.18kyrprecessioncycles,implying6-8Myrdepositionbetweenglaciations.

1CorrespondingauthoratSchoolofGeography,EarthandEnvironmentalSciences,UniversityofBirminghamB152TT,UK.Tel:+441214144181.E-mailaddress:i.j.fairchild@bham.ac.uk(I.J.Fairchild)

2Keywords:Cryogenianshalerhythmicsedimentationchemostratigraphyorbitalforcingdolomite

31. Introduction

DistinctandwidespreadglaciationisakeyphenomenonoftheEarthsystemintheCryogenianPeriod(FairchildandKennedy,2007;Shields-Zhouetal.,2012),butmuchlessattentionhasbeenpaidtothenon-glacialdeposits.Recentradiometricdateshavestrengthenedtheviewoftwowidespreadglaciations,onestartingnearthebeginningoftheCryogenian(redefinedtobe720Ma,IUGS,2014),andonefinishingattheCryogenian-Ediacaranboundary(Rooneyetal.,2015).TheinterveningperiodisreferredtoastheLateCryogenianWarmInterval(LCWI)byShieldsetal.(2012).BasedonnewobservationsfromSvalbard,wedeterminethepalaeoenvironmentsandpetrogenesisoftheshale-dominatedLCWI,placeconstraintsonitsdurationfromcyclostratigraphy,andassessitschemostratigraphiccorrelations. The720MalowerCryogenianboundaryisclosetoaclusterofpreciseU-PbdatesonzirconsrecordingtheonsetofglaciationinNWCanada(716-717Ma,Macdonaldetal.2010),Oman(711Ma,Bowringetal.,2007)andSouthChina(716Ma,Lanetal.,2014).Likewise,thebasal,globallycorrelatedcapcarbonatetotheMarinoanglaciation,whichdefinestheCryogenian-Ediacaranboundary,iswell-constrainedtoca.635MabasedonradiometricagesinChina(Condonetal.,2005;Zhangetal.,2008),Namibia(Hoffmannetal.,2004),Tasmania(Calveretal.,2013)andNWCanada(Re-Osdate;Rooneyetal.,2015).BothearlyandlateCryogenianglaciations,referredtoastheSturtianandMarinoan(basedonlocationsinSouthAustralia),respectively,appeartobegloballydistributed(Lietal.,2013).Insomeregions,thehistoryofoneorbothglaciationsislocallyorregionallycomplex,withdistinctglacialretreatintervalsrecognizedinSouthAustralia(Williamsetal.,2008,LeHeronetal.,2011;Roseetal.,2013),Namibia(Hoffman,2011;LeHeronetal.,2013),Scotland(Spencer,1971;ArnaudandFairchild,2011)andOman(Leatheretal.,2002;Rieuetal.,2007a).However,semi-continuousCryogeniansuccessionsdisplayaclearstratigraphicmotifinwhichtwoglacialunitsboundanunambiguouslynon-glacialinterval,representingamid-Cryogenianinterludeofunknownduration(~5–27My;Rooneyetal.,2014).TheNeoproterozoicsuccessionofNESvalbard,formerlycontiguouswithpresent-dayNortheastGreenland(formerlyknownasEastGreenland;Knolletal.,1986;Fairchild&Hambrey,1995;Hoffmanetal.,2012),preservesthisCryogeniannon-glacialinterval. Chronologicalandchemostratigraphicconstraintsonthemid-Cryogenianingeneral,althoughimproving(Halversonetal.,2010;Rooneyetal.,2014,2015),arecurrentlymorefluidthanthebaseandtopoftheCryogenian(Spenceetal.,2016).AtuffaceousbedjustabovetheSturtianglaciationinSouthChinahasyieldedaU-Pbageof663±4Ma(Zhouetal.,2004).Re-OsdatesfromSturtiangapcarbonateshaveyieldeddatesof662.4±4.6Ma(NWCanada,Rooneyetal.,2014),659.0±4.5Ma(Mongolia,Rooneyetal.,2015)and657±7(AmadeusBasin,Australia,Kendalletal.,2006).ArgumentstodisregardratheryoungerdatesfromtheAdelaideriftbasin(Kendalletal.,2006,2009)werepresentedbyRooneyetal.(2014).DatesfortheonsetofMarinoanglaciationarelessconstrained.U–Pbagesof654.5±3.8Maand636.3±4.9MafromtuffsimmediatelybelowtheNantuoFormation,andwithintheNantuoFormationinSouthChina(Zhangetal.,2008)respectively,providethetightestageconstraintsontheonsetofMarinoanglaciation.Insummary,currentknowledgefromradiometricdatingsuggeststhemid-Cryogenianglacialintervalrepresentsthetimebetweenc.663–659Maandc.654–636Ma,adurationofbetween5and27My. Salientchemostratigraphicfeaturesoftheintervalbetweenglaciationsinclude:1)ariseandfallinweathering-controlledchemicalindexofalteration(CIA;Rieuetal.,2007b);2)highδ34Svaluesinsedimentarypyritesandcarbonate-associatedsulphate(Gorjanetal.,2000;Hurtgenetal.,2002;Lietal.,2012);3)initiallystronglyrising,thenslightlyrising87Sr/86Srpattern(Halversonetal.,2007);4)abasalnegativeδ13Ccarbonateanomaly(Sturtiancapcarbonate),followedbyheavycarbonateδ13Csignaturespunctuatedbythepre-glacial,deepnegativeTrezonaanomaly(Halversonetal.,2002),andpossiblyanolder(Tayshir)anomaly,asseeninthelowerTsagaanOloomFormationinSWMongolia(Macdonaldetal.,2009).Detailedpalaeoenvironmentalinterpretationsofthemid-Cryogenianarepatchy.ThemostcomprehensivestudieshavebeenperformedonthickcarbonatesuccessionsinSouthAustralia(e.g.McKirdyetal.,2001;GiddingsandWallace2009a,b;Roseetal.,2012;Hood&Wallace,2014,2015;Wallaceetal.,2015)andNamibia(Halversonetal.,2002,2005,2007;HoffmanandSchrag,2002;Hurtgenetal.,2002;HoffmanandHalverson,2008;Hoffman,2011),andmixedcarbonate-siliciclasticsedimentsofNWCanada(Hofmannetal.,1990;NarbonneandAitken,1995;Dayetal.,2004). NESvalbardpresentstwoCryogenianglacialunits,presumedtobeSturtianandMarinoanequivalents.Theinterveningmid-Cryogeniannon-glacialintervalisexposedonthemainland(NESpitsbergen)andon

4Nordaustlandet(Fairchild&Hambrey,1984;Halversonetal.,2004;Hoffmanetal.,2012;Riedmanetal.,2014;Tahataetal.,2015;Kunzmannetal.,2015).Thestrataaresteeplydipping,butunmetamorphosedanddetailedgeologicalmapsaregiveninHoffmanetal.(2012)andFleming(2014,p.121-122).Were-describeandtestthecorrelationsofthewell-exposedSpitsbergenoutcrops(Figs.1and2),primarilyusingSrandCisotopechemostratigraphy,andnewevidenceispresentedforthepresenceofaSturtiancapcarbonate.Itisarguedthatpreviouslyundescribedregularsedimentaryrhythmsappeartobeorbitallycontrolled,allowinganestimatetobemadeoftheminimumdurationofthenon-glacialinterval.TherhythmsarefoundwithinaremarkablyuniformdolomiticshalefacieswhichisshowntodisplayevidenceforearlycementationinanenvironmentsubjecttodynamicvariationsinFeandMn,andyetunlikePhanerozoicferruginouscarbonates,maintainspositiveδ13Cvalues.Thisstudycontributestounderstandingofthenon-glacialCryogenianandtheunder-researchedfieldofNeoproterozoicshalesedimentology.ItalsopointstothepotentialimportanceofcyclicshaleunitsintheProterozoic.

2. Methods

2.1Field,magneticsusceptibility,petrographicandreconstructivemethodsFieldsectionsweremeasuredbytapeandAbneylevelandcorrectedtotruethickness,ordirectlymeasuredincliffs;totalthicknesswascheckedbyGPS.Samplesweretaken,primarilyforchemicalanalysisataminimumintervalof5mandsawninhalf.Around40sampleswerestudiedinpolishedorstainedthinsections.Polishedsectionswerestudiedbycold-cathodecathodoluminescence(CL)at15kVandseveralofthesealsobybackscatteredelectronmicroscopy(BSE)onaPhillipsXL30environmentalSEMoperatedat15kVandcombinedwithqualitativeelementalanalysis.One6mfieldsectionwassampledcompletelyat2cmresolution(threehundred10gsamples).Asub-set(120)ofthisintensivesamplesetweregroundtoafinepowderandmeasurementsmadeintriplicateoftotalmagneticsusceptibilityinaKLY-3SKappabridgeinstrument(supplementaryTableS1). Twocomputer-generatedthree-dimensionalreconstructionsofcentimetre-scalestructuresindolomiticshalesweremadefromserialsections.Sawnrocksurfaceswereflattenedandseriallygroundonarotatinglabcoveredwithaplatecoatedwith40µmdiamondpaste,checkingtheamountremoved(tothenearest0.1mm)withverniercalipers.EachfreshsurfacewasphotographedwetinafixedpositionandtheSPIERS(SerialPalaeontologicalImageEditingandRenderingSystem)softwaresuite(Suttonetal.,2012)wasusedtoaligntheimagesandbuilda3-dimensionaltemplate,withtheaidofrepeatedmanualadjustmentofcontrastandbrightnesstospecifysedimentarylaminae.

2.2U-PbdatingofdetritalzirconsZircongrainsfromfourpre-glacialsamples(memberE1)wereprocessedandseparatedusingstandardgravimetricandmagneticseparationtechniquesattheNERCIsotopeGeoscienceLaboratories(UK).Zircongrains(standardandunknowns)wereemplacedinepoxymountsandpolishedtoexposethegrains.PhotomicrographsandCLimageswereusedtocharacterizetheinternalstructuresofeachgraintoallowunzonedsamplestobeselected,andtohelprecordinglaserspotlocations.ThelaserstudywasperformedusingaNewWaveResearchsolid-stateNd:YAGlaserablationsystemcoupledwithaHR-NuinstrumentICP-MSintheNIGLlaboratories.AfurthersamplefromtheWilsonbreenFormationwasanalyzedattheUniversityofAdelaideusingsimilartechniquesinwhichanalysiswascarriedoutusinganAgilent7500csICPMScoupledwithaNewWave213nmNd-YAGlaser.FurtherdetailsofmethodsaregiveninlaboratoriesRobertsetal.(2011)anddataarelistedinSupplementaryTableS4.

2.3Srisotopeanalysis

5Fordeterminationof87Sr/86Sr,40-60mgofsamplepowderwasweighed,andthenleachedwith0.5MHClfor24hoursatlaboratorytemperature.Sampleswerethencentrifuged,acidpipettedoutanddriedat110°C.Srwasseparatedfromthematrixusingpre-cleanedEichromSrspecresin.Approximately150µlofresinwasloadedintoapre-cleanedandfritted1mLpipettetip.Thesampleswereloadedontoacleanedandconditionedresinin1mLof2MHNO3.Thesampleswerethenwashedwith0.4mLofthesameacid,1mLof7MHNO3and0.2mLof2MHNO3.TheSrcutwasthencollectedin1mL0.05MHNO3anddried.Totalprocedureblankisnegligible(~20pg)relativetotheamountofSrprocessedthroughthecolumn(~500ng).Thetotalyieldwascloseto90%. SrisotopeswereanalysedattheOpenUniversityusingtheThermoFisherTritonTIMS.ThesampleswereloadedontodegassedRefilamentfollowingtheproceduredescribedbyCharlieretal.(2006).Samplefilamentswereheatedto~1460°CandtheSrbeamwastunedtoobtainastablesignalof~8Vof88Sr.RawSrisotopesratioswerecorrectedforinstrumentalmassfractionationusingtheexponentiallawanda86Sr/88Srratioof0.1194.Inthisstudy,eachmeasurementconsistsofcollecting240ratiosin24blocksof10cycles,witheachratiorepresenting8.4secondsofintegrationtime.Abaselineismeasuredandtheamplifiersarerotatedaftereachblock.Aninternalprecisionof~10ppm(2s.e.)wasobtainedandtheexternalrepeatabilitywasassessedbymultipleanalysisofastandard(NIST987)andisequalto~15ppm(2s.d.,n=5).TheNIST98787Sr/86Srratiomeasuredinthisstudyis0.710224±0.000011.DataarelistedinSupplementaryTableS1.

2.4Elementalanalysis(SupplementaryTablesS1toS3)SamplesforICP-AESanalysisatRoyalHollowaywerepreparedbydrilling4mgofsamplepowderfromfresh,sawnsurfacescoveringwith2mLof1.6M(10%v/v)Aristar-gradeHNO3andallowingtoreactovernightatlaboratorytemperature,beforedilutionto0.3Mandremovalfrominsolubleresidue.Theacidstrengthwasrelativelyhightooptimizedissolutionofferroancarbonatephases.Analyseswerecorrectedusingthreestandardsmatrix-matchedforlimestonesandsimilarlyfordolomites.Ca,Mg,FeandMnanalyseswereconvertedtoequivalentmasscarbonatetoderiveananalyticaltotalfromwhichtheinsolubleresiduewasderivedbydifference.Elementalconcentrationswerethencalculatedfortheseelements(togetherwithtraceelementsSr,BaandZn)assumingthemtolieinthecarbonatefraction.Repeatanalysesfromnewdissolutionsnormallyliewithin5%. X-rayfluorescence(XRF)analysison40consecutivesamplesfromtheintensivelysampledsuitementionedabovewascarriedoutattheUniversityofLeicesteronaPANalyticalAxiosspectrometerusingfusedglassbeadspreparedfromdriedpowdersinaratioof1:5with100%Litetraborateflux.ElementsdeterminedwereSi,TI,Al,Fe,Mn,Mg,Ca,Ma,K,PandS,togetherwithlossonignition;dataarepresentedconventionallyasoxidesandanalyticaltotalswere99.7±0.9%. MicroanalysiswascarriedoutattheEdinburghIonMicroprobeFacilityonaCameca4finstrumentonpolished,goldcoatedthinsectionsusingaprimaryO-beamtogeneratepositivesecondaryions.Abeamcurrentof0.5nAwasused.Thebeamwasfocusedtoananalyticalareaofc.2µmwithanoveralldiameterof5-7µm.Anoffsetof75kVwasusedwithaverticalstepsizeof5µm.Ionsof24Mg,27Al,30Si,42Ca,54Fe,55Mn,88Sr,89Y,138Baand140Cewerecounted.ResultswerestandardisedusinginternalstandardsOkaCarbonatiteandNormanCrossCalcite. ForCO2(carbonatecarbon)analysisatbyActivationLaboratoriesLimited(Ontario),powderedsample(0.2g)wasthermallydecomposedat1000°C,H2Oremovedinamoisturetrap,andCO2determinedbyaninfra-redcell.Organiccarbonanalysesweredeterminedatthesamelaboratoriesusingthedifferencebetweentotalcarbonandcarbonatecarbon,butmostvalueswerebelowtheeffectivelimitofdeterminationof0.2wt.%andarenotreproducible.

2.5CarbonandoxygenisotopeanalysisCarbonandoxygenstableisotopedataarepresentedhereasδ13Candδ18OinpartsperthousandwithrespecttotheVPDBstandard.Dataarelistedinsupplementarytable1andincludeanalysesfromFairchild&Spiro(1987)andHalversonetal.(2004).NewdatawasobtainedattheUniversityofBirminghamusinga

6continuous-flowIsoprimeIRMS,withamultiflowpreparationsystem.Samplesofbetween80-250µgpowderedcarbonatewerereactedwithphosphoricacidat90°Cforatleast90minutes;resultswerecalibratedusingIAEAstandardsNBS-18andNBS-19andrepeatabilityonaninternalstandardwasbetterthan0.1‰forδ13Cand0.15‰forδ18O.DataarereportedinSupplementarydataTableS1.

3. LithostratigraphyandcorrelationsCryogeniansedimentsinSvalbardoccurwithinthePolarisbreenGroupoftheHeklaHoekSupergroupandcomprisetheElbobreen,WilsonbreenandDracoisenformations,subdividedintomembers,whichareabbreviatedbymeansofnumbers(Table1;Hambrey,1982).Anearlierglaciation(memberE2)iscorrelatedwiththeolderCryogenian(Sturtian)glaciation(Halversonetal.,2011),andalaterglaciation(WilsonbreenFormation)istruncatedbyatransgressivesuccessioncorrelatedwiththebasalEdiacaran(Halversonetal.,2004),substantiatingaMarinoanage.Theinterveningnon-glacialintervalwasinterpretedbyFairchild&Hambrey(1984)asasingleshoaling-upwardssequencetorepresentoffshoremarinedolomiticshales(memberE3)shallowingupwardstoadolomiticooidalgrainstonewithanhydritepseudomorphsandperitidaltepeestructures(memberE4)depositedinarestricted,shallow,warmenvironment.TheupperboundaryofE4isasubaerialexposuresurfacewithevidenceforfrost-wedging. Halversonetal.(2004)proposedanalternativemodelinwhichE2toE4andtheWilsonbreenFormationweredepositedwithinasingle(Marinoan)glaciation.Thisaccountedfor:a)thepresenceofanegativecarbonisotopeanomaly(correlatedwiththesub-MarinoanTrezonaanomaly)beneathE2,butnotbeneaththeWilsonbreenFormation,despitenoevidenceofsignificanttruncationatthatsurface,b)thelackofaclearstratigraphicδ13Cprofileakintoothersequencesinthisnon-glacialtimeinterval,andc)thepresenceofpseudomorphswithinE3,interpretedasglendonites(afterikaite).Inthismodel,theuniformlyfine-grainedE3facieswereinterpretedtohavebeendepositedbeneathpermanenticecover.Thelaterdiscoveryofacarbonisotopeanomaly(theIslayanomaly)beneathSturtian-agedsedimentselsewhere,alongwithnewstrontiumisotopedatafromtheRussøya(E1)membertippedthebalanceofevidencebacktothetwo-glaciationmodel(Halverson,2006;Halversonetal.,2007). Riedmanetal.(2014)foundonlydepauperateacritarchassemblagesinE3,dominatedbysimplespheres,butfoundsuchlowdiversityassemblagestobetypicalalsooftheSturtianandpost-SturtianintervalinAustraliansections.Tahataetal.(2015)carriedoutamicroanalyticalFe-isotopestudyofpyriteinPolarisbreenGroupsectionsinNordaustlandettotheNEofourstudyareaandobservedδ56/54FeinE3upto+3.91±0.29‰,indicativeofironoxidationatachemoclineaboveaferruginousocean.Theyalsopresentedlow-resolutionδ13Cdatawhichareconsistentwiththosepresentedherein.Theiron-speciationandtracemetalstudyofKunzmannetal.(2015)whichincludessomeE3samples,likewiseconfirmsananoxic,ferruginousocean. Wehavere-measuredtwonearlycompletesectionsoriginallypresentedbyFairchild&Hambrey(1984)atDracoisenandDitlovtoppen(Fig.2).AnewlocationinSEAndromedafjellet,informallytermedReinsryggen,presentsacontinuoussection,butismuchthinnerbecauseofstratacutoutbyafaultcrossingthesectionjustbelowtheE3-E4boundary.Despitethis,thetopofE3andallofE4areexposedhere,astheyareinseveralincompletesectionstothesouth(Halversonetal.,2004).Wepresentanew,field-excavatedsectionofthebaseofE3fromthesouthBacklundtoppenlocationandinupperE3fromthenorthBacklundtoppenridge(Fig.2).BothofthesesectionscontainlimestonesfromwhichusefulnewSrisotopedatawereobtained.

4. Sedimentology,diagenesisandenvironmentalevolution

4.1TheE2-E3transitionIntroduction

7Fairchild&Hambrey(1984)describedandinterpreteddiamictiteandrhythmitefaciesofE2,recordingaconsistentlysharpboundarytodolomiticshaleatthebaseofE3.However,Halversonetal.(2004)additionallyrecordedatransitionalcontactbetweenglacigenicfaciesinE2andlaminatedlimestonesattheSouthBacklundtoppensection.ThekeyissuehereiswhetherdistinctivefeaturesofaSturtiancapcarbonate(Kaufmanetal.,1997;Kennedyetal.,1998)arepresent.Description E2istypically10-15mthickandnearitstopisdominatedbysharp-basedconglomerates,gradedsilty-sandstoneswithclimbingripplesanddropstones,andmillimetretocentimetre-scaleclasticrhythmites(e.g.Fig.3A,B).TheboundarybetweenthetopofE2tolaminatedshalycarbonatesofMemberE3issharpexceptintheSouthBacklundtoppensection(Fig.2)wherea30cmintervaloflimestonerhythmitesoccursametrebelowthetopofE2(Fig.3B),exhibitingwithin-layerasymmetricfoldsandthrusts(Fig.3D).Observedpetrographically,theserhythmitesconsistofcalcitemicrosparandareseparatedbydolomicritelaminaecontainingsandandsilt-gradedolomiteandsiliciclasticdetritus(Fig.3E).Theyareoverlainbyasharp-based1.2m-thickunitvaryinglaterallyfromconglomerate(Fig.3B)tosilt-to-sand-gradeclasticdolomiteandinturnby0.2mofmillimetre-laminatedrhythmiteswithsmalldropstones(recognizedusingcriteriainFairchild&Hambrey,1984).ThebasalfewcentimetresofE3atSouthBacklundtoppen(Fig.3C)andDitlovtoppen(Fig.3F)consistofdololaminiteswithsignificantdetritus,dropstonesanddiamictitepellets.Accessoryfine-grainedpyriteisnearlyubiquitousinalllithologies.Cathodoluminescencemicroscopydemonstratesthepresenceofanauthigenicdolomitephasewithabrightluminescencezonethatsurroundsdetritalparticleswhichhavevariablecathodoluminescencecharacteristics(Fig.3G,H).FairchildandHambrey(1984)foundthatauthigenicmatrixdolomitewithnegativeδ13C(-1to-4‰)andhighFecontent(>25000ppm)wascharacteristicofE2,surroundingclastswithpositiveδ13C. Halversonetal.(2004)recordedlaminatedlimestonewithhighlyvariablenegativeδ13Csignatures(-3to-16‰)inthebasalmetreofE3,probablyequivalenttolimestonestartingatthe+1mlevelinthisstudy.Thislimestonehas0.2-0.3mm-thickmicrosparlaminaewiththininterveningmicritelaminaeandislocallyaffectedbystylolitization(Fig.4B).Themicrosparlayerslocallyhavemillimetre-wide,shallowupwardconvexitiesthatmayreflectanoriginalprimaryupwardmineralgrowth(Fig.4C).However,cathodoluminescencedemonstratesadistinctivereplacivefabricofzonedcalciterhombs(Fig.4D).Abovethe+5mlevel,thecarbonatesaredolomiteratherthanlimestone,buthaveasimilarlaminationstyle. Intheothersections,limestoneisabsentinthebasalE3sediments,butsimilarfaciescomposedofdolomitearefoundinstead.AconsistentCLzonationofdolomicritetodolomicrosparreplacivemosaicsovergrownbydolomitecementsdemonstratestheearlydiageneticoriginofthisdolomitephase(Fig.5).DolomicrosparlaminaesometimesshowupwardconvexitiessimilartotheBacklundtoppenlimestonelaminitesdescribedabove. NotablesedimentarystructuresatvariouslevelsinthebasalfewmetresoftheBacklundtoppenandReinsryggensectionsarerecumbentfoldsafewmmthickwithroundednoses(Fig.4A,5A).Infoldnoses,thecarbonatefabricbreaksupinto"pseudo-allochemical"areasbybrecciationofthelaminitesandthesearethenovergrownbycarbonatecement(Fig.5B,5F). OxygenandcarbonisotopedatafromthewholeofE3andE4areplottedinFig.6wherethevastmajorityofthenegativeδ13CvaluesarefromtheE2-E3transition.LimestonesatthebaseofE3clusteraround-9‰anddisplayconsistentlymorenegativeδ18Ovaluesthandolomite(mostly-1to-4‰).Thisoffsetisgreaterthanthe3‰expectedfromequilibriumprecipitationfromafluidofthesameδ18Ocompositionatthesametemperature(Land,1980).Strontiumconcentrationsinthebasal30mofE3rangefrom514to705ppmforlimestonesand84to319ppmfordolostones(supplementarydata,TableS1). Highlyvariablecarbonisotopevaluesfromthebasallimestonebed(Fig.7)wereinterpretedbyHalversonetal.(2004)asanindicationofearly,microbiallymediateddiagenesis(cf.Irwinetal.,1977).Whilstthisisvalid,thenewlyenlargeddataset(Fig.7)showsthatsuchvariability,withlimestonestendingtobelighterthandolomites,isrestrictedtothebasal2m(Fig.7).Overallinthebasal25mofE3intheBacklundtoppensection,aconsistentrisingtrendisseenfromaround-4to+3‰,corroboratedbyisolatedsamplesfromothersections,andindependentofFeandsiliciclasticcontent.Thebasallaminateddolomites

8atReinsryggenhaveslightlynegativeδ13Cvalues,anditispossiblethattheyarecoevalwithdolomitesinthe5-10mintervalatBacklundtoppen.Interpretation BasedonthecriteriadevelopedbyFairchild&Hambrey(1984),massivediamictitesrepresentthebulkrain-outfromahighdensityoficebergsinaproximalglacimarineposition,whereasthelensingbedsofmatrix-supportedbrecciarepresentsubaqueoussedimentgravityflows(possiblyclosetothegroundingline).Rhythmitesreflectingcyclictidally-influencedsedimentationfromsedimentplumesemergingfromicecliffsbelowwave-base(cyclopelsandcyclopsams).GlaciallyderivedsedimentdiminishedrapidlyattheE2-E3boundaryandsedimentparticleslargerthancoarsesiltarenotfoundabovethebasalfewdecimetresofmemberE3.Sucharapidchangeinsedimentcharacterisconsistentwithrapiddeglaciation.Thereisnodirectevidenceofchangesinwaterdepth,althoughsignificantsea-levelrisewouldbeexpectedatthecloseofapanglaciation. Thesedimentarydeformationfoundinfinesediments(Fig.4A)issuggestiveofdownslopemovementwithrecumbentfoldsandroundedfoldnosesresemblingsmall-scaleslumpfolds.However,exposuresdonotpermitmeasurementstodemonstrateaconsistentslope-relatedorientation.InmicrobiallaminitesinanequivalentcapcarbonatefromNamibia,Prussetal.(2010)documentedtheoccurrenceofrecumbentfoldsandspiralroll-upstructureswhichareassociatedwithsedimentarydykesindicativeofanoriginrelatedtoliquefactioncausedbyfluidescape.However,nosuchancillaryfeaturesoccurintheE3sections.Nevertheless,inbothcasesthedeformationislikelytohavebeenrestrictedtoasurficial(nomorethandecimetre-thick)unlithified,orweaklylithifiedlayer,reflectingsimilarearlycementationofthelayers. Thelimestonerhythmitesfound1mbelowthetopofE2atBacklundtoppenhaveamicrosparrycharacterlikethosenearthebaseofE3andtheirpresenceimpliesprecipitationofacarbonatephasewithinthewatercolumnoratthesedimentsurface,probablyassistedbythepresenceofmicrobes(BosakandNewman,2003).Therelativelyuniformthicknessofthecarbonatelayers(Fig.3B)mightimplyanannualoriginbycomparisonwithlacustrinelaminitesoftheWilsonbreenFormation(Fairchildetal.,2016),buttheregularityislessmarkedinthinsection(Fig.3F)wherelayerscommonlydisplayacompositestructure.Carbonatevarvesarecommoninlacustrineenvironmentsinvariousclimaticzones(LawrenceandHendy,1985;Shanahanetal.,2008)butarenotcharacteristicofmarineenvironmentstodaydespitethelackofaphysico-chemicalreasonwhytheyshouldnotoccur.Photosynthesisismoreeffectiveinraisingcarbonatesupersaturationinwatersoflowerionicstrength(Fairchild,1991),butthedepositionalwatersatE2-E3boundarytimeswouldhavebeensaturatedforcalciteanywaybecauseoftheabundanceofdetritaldolomiterockflour(FairchildandHambrey,1984).Consequently,seasonalfluctuationintheintensityofphotosynthesisisaviablemechanismfortheirformation.Ontheotherhand,therelativelylightoxygenisotopecompositionofthelimestones(-6to-9‰,Fig.6)couldimplyanorigininseawaterdilutedwithmeltwater.Limestonelayersdonotshowevidenceofphysicalcompaction,implyingthatthereplacementofaprecursorbycalcitehappenedduringearlydiagenesisandhenceelevatedburialtemperaturesarenotrequiredtoexplainlighterδ18Ovalues.Theupwardconvexitiesarereminiscentofabulbousmodeofupwardgrowthoftheprecursorfromthesedimentsurface,buttherearenopreservedinclusionstoprovidefurthercluesastotheprimarystyleofgrowth.LessstablecalciumcarbonateformsthatcouldhavebeentheprecursorareamorphousCaCO3,ikaite,vateriteoraragonite.Ikaiteisanattractiveoptionbecauseitformspreferentiallyatlowtemperatures(Shearman&Smith,1985;Sellecketal.,2007;Oehlerichetal.,2013)suchasthoseexpectedtocoincidewithice-rafting;Fairchildetal.(2016)haveshownsimilarreplaciverhombiccalcitefabricstothoseinbasalE3limestonesinreplacedikaiteoftheWilsonbreenFormation. SamplesstudiedfromthosepartsofmemberE2closetotheboundarywithE3appeartexturallytobedominatedbydetritaldolomite,althoughanauthigeniccontributioncanbeinferredfromthenegativeδ13CandsignatureandhighFecontents(FairchildandHambrey,1984).AuthigenicdolomiteisvisibleasabrightCLzoneinbasalE3dolomite,althoughitismostlydarkasexpectedforahigh-Fephase(Fig.3H).InE3,dolomiteformslaminarcarbonatesthataretexturallyverysimilartothelimestonelaminites,anditappearsthatdolomitizationoccursbyreplacementofapartlylithifiedprecursorphasefoldedbysoft-sedimentdeformationbeforecompactionandcementationbydolomite(Fig.5).DuringtheNeoproterozoic,early

9dolomitizationisubiquitousinperitidalmarinerocksbutpatchierandinsubtidalsediments(KnollandSwett,1990).InthecaseoftheE2-E3boundarysediments,thepresenceoforganiccarbonboundtoclaysandsignificantsulphatereduction(asevidencedbypyritecontent)wouldhavefavoureddolomiteformation(Mazzullo,2000;Zhangetal.,2012). Varyingporewaterchemistryduringdiageneticmineralreplacementisimpliedbythecomplexcathodoluminescencezonationobservedinbothcalciteanddolomiteandthepresenceofauthigenicphasessuchaspyrite.Alocalizedroleforbacterialprocessesinmodifyingtheδ13CofDICisclearinthebasallimestone(Halversonetal.,2004).However,thesteadyriseinδ13CvalueswithstratigraphicheightthroughE3agreeswithalow-resolutionstudyshowinganupwardrisingtrendinδ13Cofbothorganiccarbonandcarbonatecarboninthesamestudyarea(Kaufmanetal.,1997)andimpliesthatthecarbonisotopevaluesofE3carbonatesarenotdominatedbylocalbacterialinfluences.Thesecularriseinδ13Coccursinfine-grainedpost-Sturtiancapcarbonatesonmultiplecratonsandimpliesaglobalsignature(HalversonandShields-Zhou,2011),orparallelvariationsinlocalconditionssuchaswaterdepth(GiddingsandWallace,2009a).TheFe-enrichmentanddarkcolour(upto0.7%organiccarbon,Kunzmannetal.,2015)arealsofeaturesofbasalE3whichbearcomparisonwithotherSturtiancaps(Kennedyetal.,1998),whilstfeaturesthatarerestrictedtoMarinoancaps(e.g.barite,tepees,giantwaveripples)areabsent.

4.2DolomiticshalesedimentationinE3IntroductionMostofmemberE3iscomposedofdolomiticsilt-shale,withsub-millimetrescalelaminationandnotractionstructures(FairchildandHambrey,1984).Althoughnooverallchangeinfaciesisapparentfrom20to180maboveitsbase(Figs.2,8A),astrikinglyregularrhythmicorcyclicmodeofsedimentationisrevealedondifferentiallyweatheredoutcropsatDracoisen(Fig.8B)andDitlovtoppen.Themoreresistantlevelswithinrhythms(cycles)commonlyappearonweatheredsurfacestobepurercarbonates(Fig.8C),andlocallydisplayaconcretionaryforminwhichlayerthicknessisdoublethatoutsidetheconcretion(Fig.8F).However,morecommonly,resistantandnon-resistantweatheredpartsofsedimentaryrhythmsaremoredifficulttodistinguishclose-up(Fig.8D,E)thanatadistance(Fig.8B)andthissubtletyprovidesasignificantchallenge.Webuildupthestorythroughsuccessivelydescribingandinterpreting:1)overalllithology,2)characteristicsedimentarydisturbancestructures,3)petrologyandgeochemistryand4)rhythmicity.LithologyThesiliciclasticcomponentofE3isdominatedbyfinetocoarsequartzo-feldspathicsilt;sandisvirtuallyabsent.Typicallaminaeare0.1-0.4mmthickandconsistofathicker,dolomiticsiltsub-laminaandathinnerparting,richerinphyllosilicates,oftenwithenhancedlevelsofframboidalpyrite(Fig.9A,B,D-G).Thesiltysub-laminaearelocallymicro-nodularincharacter(Fig.9A),andatcertainhorizonsthelaminarstructuredisplayslaterallyimpersistentcentimetre-scaledisturbanceswithtightlyfoldedlaminae(Fig.9B,F)whicharedescribedinthenextsection.Above170mstratigraphicheight,finerlaminaearelessnoticeable(Fig.9H)andhydrodynamicsedimentsortingisshowntexturallyabove180mwhen100-200µm-thicklensesofsortedsiltandfaintcross-laminationappearinwhatarenowdolomiticsiltstonesratherthansilt-shales(Fig.9C). Theuniform,fine-grainedlithologyofthemainpartofmemberE3,coupledwithevidenceforconsistentsub-millimetrescalelaminationindicatesaregimeofslowsedimentationwithoutstronghydrodynamicactivity.Inparticular,sharp-basedorgradedunitsareabsentandlaminaearemorecontinuousthanthelensingstructuresthatcanforminshalesfromredepositionofsoftmudclasts(Schieberetal.,2010).Theimplieddepositionalenvironmentwasconsistentlybelowwavebaseuntilthefirstsignsofsortinginlensinglaminae(e.g.Fig.9C)nearthetopofthemember.Thelackofevidenceforsedimentgravityflowsimpliesacontinentalshelf,ratherthanslopeenvironment,andhencewaterdepthsoftheorderof100-300m.Thepresenceoflaminaevaryingingrainsizeindicatessomevariationinsedimentsupplyandaccumulation. Pale,micronodularlaminaeimplyearlycarbonatecementation,andtheselayerswouldhaveresistedcompaction.Thediscrete,thin,andflexiblenatureofthefiner,darkerlaminaeisconsistentwithanoriginas

10microbialmats,whicharewell-knowninProterozoicoffshoresiliciclasticsediment(Schieber,1986).Suchmatstypicallymetabolizedbyanoxicphotosynthesis.Thefocussingoforganicmatterinthesehorizonsisconsistentwiththepreferentialoccurrenceofpyrite(Schieber,1986;Prussetal.,2010).Disturbancestructures Fieldandpetrographicobservationsshowthatthedeformedlaminaenotedabove(Fig.9B,F)formadistinctgeometricclassofsedimentarystructure,referredtohereasdisturbancestructures,whichhavesimilarcharacteristicsthroughoutthemainpartofmemberE3.Thesestructuresarevisibleinthefieldindiscretehorizons(Fig.2),althoughrequireoptimalweatheringandilluminationtobeclearlyseen.Thestructuresformequant,centimetre-scaleconcaveorconvexmarkings(Fig.10A)withcurvedandupturnedlaminae(Fig.10B).Locallytheyformtrainsofseveralstructureswithuptocentimetre-scaleseparation(Fig.10C).Intransversesection,structuresexhibitacoreofconcentricovoidlaminaewithlaminadisturbancedecreasingawayfromthestructure;hencetheyarestratigraphicallyconfinedonacentimetre-scale(Fig.10D)andhencearemuchsmallerthanthestructuresinterpretedaboveasslumpfoldsfromthebaseofmemberE3.Locallytheyformmorecomplexen-echelonarraysinwhichapparentlycontinuous,butdeformedlaminaeseparateeachstructure(Fig.10E). Athree-dimensionalmodelwasconstructedfromthesampleshowninFig.10Ffrom71serialsectionscuttoadepthof9mm.Themodelshowstheconcentriccorelaminae(Fig.10G)tobeconical,flaringintothesectionedslab(upwardsinFig.10H,I).Theterminationsarecomplex(e.g.double-pointedinFig.10H)andtheflattenednatureoftheconesindicatesanoverallsheath-likegeometry.Thecoreissurroundedbycontinuouslaminae(e.g.greenlayerinFig.10J),thenbydisturbedlaminae(purpleandblue,Fig.10J),thedeformationofwhichisdisjointedfromtheinnerpartofthestructure. AsecondmodelwasgeneratedfromthelowerrightoftheslabshowninFig.11Abasedonfrom31sectionscuttoadepthof6.2mm.Thismodelrevealsanen-echelontrainofstructures,eachwithatubularcorethatisovoidincross-section(Fig.11B)andnarrowsintotheslab(Fig.11C).Thelaminationseparatingsomeofthestructuresisclearlycontinuous(e.g.Fig.10A),butlaminaearebentinsuchawayastosuggestareversefaultgeometry.Otherwise,laminaemorethanafewmillimetresaboveadisturbancestructureappeartobeunaffected.Adeeperserialsectioncutthroughthissamplerevealsbrittledeformationalstructures,e.g.thesharpapparentlyreverse-faultoffsetarrowedinFig.10Dwhichappearstocontinueasanormalfaultlowerintheslab. Insummary,thetypicalgeometryoftheE3structurescanbesummarizedashorizontallydirectedprolapses—thatis,sheath-likefoldsflattenedinthehorizontalplane.Inplanview,thedirectionofmovementisunclearasthestructuresareequantandinternallaminationisonlylocallyvisible,butthedirectionofclosureoflaminationcanbediscernedbyserialsectioning.Theavailableevidenceindicatesnopreferredorientationofthestructures. Wecanruleseveralpossibleoriginsforthesestructuresongeometricgrounds.Theunbrokenlaminationrulesoutanoriginastracefossils,whilsttheonlyputativebodyfossilthatshowssimilaritiesisHorodyskia,whichmanifestsasachainofbead-likestructures(cf.Fig.10C)butispreservedaspitsormoulds(Greyetal.,2010),lackinginternallamination.Likewise,thesheath-likecorestotheE3structuresarelaminatedthroughout,unlikethesedimentpackingofanimalburrows.Also,thestructuresdifferfromthefamilyofmicrobialmat-relatedphenomenaknownas“wrinklestructures”(Porada&Bouougri,2007)becausetheE3structuresareirregularlyspaced,ratherthanformingacontinuous,self-organizednetwork. Theclosestsimilarityinexternalmorphologyiswithhorizontaltubularstructurestermedroll-upstructureswhicharefoundinmoderncoastalenvironments,andbothshallowanddeepProterozoiclaminatedsediments(Sarkaretal.,2014).TherearetwoProterozoicexamplesdescribedindetail.SimonsonandCarney(1999)describedcentimetre-highanduptodecimetre-widerecumbentfoldswithroundedlimbsandinconsistentvergencewithinpartiallycarbonate-cementedshalesofthePalaeoproterozoicHamersleyGroupofWesternAustralia.Prussetal.(2010)illustratedavarietyoffoldstructuresupto10cmwideby1cmwide,andcommonlycontainingspirallaminae,foundassociatedwithsedimentarydykesinthinlylaminatedmicrobiallaminitesofSturtiancapcarbonatesofNamibia.Inmoderncoastalenvironmentsrollupsformwhereamicrobialmatbreaksandedgescurlupduetosubaerialexposure(Sarkaretal.,2014),whereastheancientdeep-waterstructuresarelesscompletelyunderstood.

11 AlthoughE3disturbancestructuresdonotcontainspirallinglaminaeandhencetheycannotbetermedrollupstructures,acomparisonisstillinstructive.Thefirstpointisthatroll-upstructuresdemonstratesedimentelasticity,apropertywhicharisesfromtheabundanceofmucilaginousextracellularpolymericsubstances(Beraldi-Campesi&Garcia-Pichel,2011;Chewetal.,2014).Thelackofbreakageoflayers,despitetightbending,isseentoointheE3disturbancestructures,andtheyarealsointerpretedtocontainmats. Thesecondpointofcomparisonwithrollupstructuresisthetriggerfordeformation.IntheirHamersleyexampleSimonson&Carney(1999)couldnotdecidebetweenseveralcandidatetriggeringmechanisms:downslopemovement,currentshear,highpore-fluidpressuresandcyclicdisturbancebyseismicwaves.However,intheRasthofexamples,theassociateddykesappeartohavebeenzonesofactivefluidescapeatthetimeofsedimentation,andPrussetal.(2010)interprettheroll-upstructurestobeassociatedwithlocalliquefactionrelatedtodykeemplacement. TheE3disturbancestructures,occurthroughouttheaffectedsediments,ratherthanbeingfocusedatspecificeventhorizons,anddonothaveapreferredorientation.Hence,thereisnodirectevidenceforcurrentshear,seismicdisturbanceordownslopemovement,buttheverysmallscaleofthedisturbancesisconsistentwithlocalizeddeformationtriggeredbyexcessporepressureinliquefiedsediment.Akeypointisthatthestructuresarenotstronglycompactedandhencesubsequent,near-surfacecementationcanbeinferred(asinthecaseofSimonson&Carney,1999).Thisisalsoconsistentwiththepresenceofundisturbedlaminationafewmillimetresabovedisturbancestructures.Hence,thedeformationisassociatedwithasurficiallayer,capableofductiledeformation,aswouldbeastrongandelasticmat,overlyingprogressivelymorerigid,cementedsediment.Ifearlycementationwasfocusedmoreinsomelaminaethanothers,porefluidpressureswouldberaisedbeneathmorelithifiedhorizons,asphysicalcompactionproceeds.InE3,evidenceoflocalizedfracturingofslightlycementedlayers(Fig.11D)isconsistentwithreleaseofunderlyingfluid.RatherthanformingdykesasintheRasthofFormation(Prussetal.,2010),theporefluidsinE3sedimentsappearstohavebeenalocallyeruptedintoalesscemented,lowerpressurelayerpropellingasmallclotofsedimentsideways,creatingtheprolapsedstructures.Theen-echelonstructurecanbeinterpretedasanobliquefaultzonewithinwhichfluidisreleased,eitheronseveraldistinctoccasions,ormorelikelysimultaneouslyatseveraldifferentlevelswithinthe1.5cm-highstructure,withlittledeformationatinterveningslightlymorecementedhorizons(e.g.2onFig.11A). Itcanbepredictedthatsuchsmall-scalefluid-escapestructures,whicharedifficulttostudyinthefield,maybemuchmorewidespreadinNeoproterozoicshalesthanhavebeendescribeduptonow.Ashasbeenshown,theyprovideusefulevidenceforthenatureofearlydiagenesis.GeochemistryandpetrologyThestratigraphicvariationofδ13Cinthe20-180mintervalisextremelylimited,withdolomitevaluesnearlyallintherangeof+2.5to+4‰(Fig.12).Low,outlyingδ13Cvaluesalmostallcontainsignificantcalciteandhenceweinferthatthecalcitehasalowerδ13Cvaluethandolomiteinthesamesamples.Thecalciticsamplesalsohavelowerinsolubleresiduecontents(Fig.13).Dolomitesshowmorevariationinδ18O(4‰)thanδ13C(2‰,Figs.7,12),andthereisaclearupwardstratigraphictrendinδ18Ofrom40-180minbothDracoisenandDitlovtoppen.Howeverinthe20-40minterval,δ18Ovariesfrom-6toslightlyabove0‰andvariesbetweensections.Theonlyothermeasuredparametershowingastratigraphictrendisacid-solubleFe(plottedasFeCO3),whichdiminishesupwards(Fig.13),andisslightlyhigherinDitlovtoppensamplesthanatDracoisen. ThemineraltexturesandinternalstructuresaremostclearlyrevealedinCLandBSEmicroscopy(Fig.14).Dolomiteformschemicallyzonedcrystals(Fig.14A,B,E)ofsimilarsizetothesiliciclasticsilt.Thezonesarerhombic,buttheexternalcrystalshapeissubhedral,adjacenttosilicates,oranhedralagainstotherdolomitecrystals(Fig.14E,F).Calciteshowslessregularcrystalshapes(Fig.14E)andtendstooccurmorediscretelyaslargerirregularcrystals(Fig.14F).IntheconcretionofFig.8F,calcitecrystalsenclosedetritusanddolomitecrystalsalike(Fig.14D)andhencepost-datesdolomite. IonmicroprobemicroanalysisrevealsthatmostindividualcrystalsdisplayzonationasillustratedforoneofthesamplesinFig.15A.IndividualcrystalseithershowFe-MncovariationorvariationinFealone.Strikingly,bothdolomiteandcalciteshowthesameoverallrangeinconcentration,andthispatternisfound

12inothersamples.OverallFe-Mncovarianceisshownbycrystalmeans(Fig.15B)withuptotwoordersofmagnitudevariationinFeandthreeinMn.ThisextraordinaryheterogeneityisreflectedintheCLcharacteristicsofthecarbonates(Fig.14A-C)wherenocommonpatternofzonationisdiscernable.ComparisonofICPandionprobeanalysesillustratesthatalthoughMnanalysesarecloselycomparable,bulkacid-solubleFeisclosetothemaximumobservedformeansofcrystalsanalyzedbyionmicroprobe(Fig.15B)andhencethattheacidattackisreleasingFefromotherphasesaswellascarbonate. Thepresenceofmicronodularlaminae(Fig.9A),macroscopicuncompactedconcretions(Fig.8F)andsedimentdisturbancestructuresallpointtotheroleofearlycementationbydolomite.Calcitepost-datesdolomitepetrographicallyandthelowinsolubleresidueofcalciticlithologies(Fig.13)impliesthatcalcitefillsremainingporosityathorizonswheredolomitecementationwassufficienttoresistanyfurthercompaction.Calciteanddolomiteprecipitatingfromfluidswiththesameδ18Oatthesametemperatureareexpectedtodifferinδ18Oby3‰(Land,1980),butmostcalciteshaveevenlowerδ18O,consistentwithanoriginathighertemperature. BothcalciteanddolomitescavengeFe2+andMn2+fromsolutionduringgrowth(Veizer,1993;Rimstidtetal.,1998)andsotheirhighlyvariableFe-Mnchemistry(Fig.15)canbeassumedtobeadirectreflectionofthechangingporefluidcompositionoftheseelements.ChangingavailabilityofFeandMneitherreflectsmicrobialreductionofprogressivelymorerefractoryoxidizedsourcesoftheseminerals(Irwinetal.,1977)ordownwarddiffusionfromoverlyingseawaterwithfluctuatingFe-Mnchemistry.Thelackofδ13Cvariabilityindolomitesimpliesbufferingbyapre-existingfine-grainedcarbonatephase.SinceCLzonationshowsthatcrystalsgrewatdifferenttimes,theprecursorislikelytohavedissolvedprogressively,permittingthesustainedgradualprecipitationofdolomite.Theprecursorthenpresumablybecomeexhaustedbythetimethatcalciteprecipitated,leadingtoalowerδ13Cvalueindicativeofapartialsourceofcarbonfrombreakdownoforganicmatter(MozleyandBurns,1993).Conversely,theconsistentδ13Csignatureofthedolomitescanbeusedforchemostratigraphicpurposesasanindicatorofmarinewaterδ13Cvalues.Thisresultisunexpected,giventhehypothesisthatlightδ13Cvalueswouldcharacterizeauthigeniccarbonateindeepermarinesettings(Schragetal.,2013),butisconsistentwiththeloworganiccarboncontentofthesediments(Kunzmannetal.,2015recordvalues<0.3%). Theup-sectionreductioninacid-solubleFe(Fig.13),whichcoincideswithariseinorganiccarbontovaluesof0.5-1%(Kunzmannetal.,2015),mayreflectchangesinbothcarbonateandsilicatefractionsandthisisbeinginvestigatedinaseparatestudyofclaymineralogyandFe-speciation.Iftheup-section3‰increaseinδ18Oofdolomites(Fig.12)werecausedsimplybychangingtemperatureofformation,itwouldimplyadecreaseof>12°C,whichisunlikelygiventheotherwiseconsistentcharacterofthesediments.Otherstudieshaveshownoxygenisotopecompositionsofearlydiageneticcarbonateconcretionstobemorenegativethanexpectedatequilibrium(MozleyandBurns,1993)andthatthisismarkedlysoforsideritesprecipitatedfrommicrobialratherthaninorganicsystems(MortimerandColeman,1997).Hence,kineticeffectslinkedtovaryingavailabilityofFeareinterpretedtoberesponsiblefortheδ18Otrend.RhythmicityTherhythmicbeddinginE3showsupprominentlyincertainfieldphotographs(Fig.8),butitisdifficulttoquantify.Wehaveaccuratelyloggedthestratigraphicpositionofthecentreofeachofthemoreresistanthorizonswhichdefineeachcycle,butthicknessmeasurementsofthesehorizonsareimprecisesincethesebedsusuallydonothavesharpbasesortops.TimepermittedloggingofonlyoneprofileeachatDracoisenandDitlovtoppen;comparablecyclestothosefoundatthesetwolocalitieswereonlylocallyvisibleatReinsryggenandBacklundtoppen. TheDitlovtoppensectionpresentsthegreatestnumberofcycles:270arevisibleinthe12mto195mintervalabovethebaseofE3(Fig.2).Thehistogramofthicknessespresentsamodeintherange0.4-0.6mandapositiveskew(Fig.16C).Overall,thecumulativethicknessplotisfairlylinear(Fig.16A),butwithdiscontinuitiesatstratigraphicheightsofaround50,80and150m(abovethebaseofE2),whereunusuallythickrhythmsarepresent.Lackofclearexposureofsomerhythmspresentsareadyexplanationforthepositiveskewofthethicknesshistogram(Fig.16C)andthemodeistakenasmorerepresentativeoftheirtruethickness.

13 TheDracoisensectionpresents120cyclesbetweenthe45mand170mlevelinthesection,withnostratigraphictrends(Fig.16A)andwithahighermeanvalueandastrongerskewthanatDitlovtoppen(Fig.16B,C).TheDracoisensectionwasloggedintheuppercentreofFig.8B,butthephotographillustratesthatmanycyclesvisibleinthecliffintheforegroundbecomelessclearlaterally.Also,sincethetotalthicknessofE3isvirtuallyidenticalatbothsectionsandtheyareonly10kmapart,thedatafromDracoisenareregardedaslesscomplete.Themodalthickness,between0.8and1.2m,islikelytocontainasignificantproportionofdoublerhythms,andthetailofthedistributionlikelyincludesmultiplerhythms.TheDitlovtoppenhistogram(Fig.16C),withamodeataround0.5m,isregardedasbeingclosertothetruepicture. Sincefieldobservationswerenecessarilyimprecise,laboratorydatawasusedtospecifythelithologicalnatureofrhythmsbasedonacombinationofgeochemicalanalysesandmagneticsusceptibility(MS)measurements.Samplesweretakenfromrepresentativeresistant/lessresistantlithologiesthroughoutE3andonselectedsamplesfromahighresolutionsuiteof300collectedbetween40and45mheightintheDracoisensectionandcoveringsevencycles.Organiccarboncontentcouldnotbeusedsincemostsampleswerebelowdetectionlimits(0.2wt.%)fortotalorganiccarbon. Sincetheratioofcarbonatetosiliciclasticsedimentisthecommonestmodeofvariabilityinoffshorecyclicsediments(Ricken,1986,1996),plotsofCaCO3contentoftotalsedimentweregenerated(Fig.17C,D).AlthoughthebasalcycleshowsaresistanthorizonwithhighCacontentat46.5m,suchacorrelationisnotobvioushigherintheintensivelysampledsection(Fig.17C),noristhereanycleardistinctioninCacontentbetweenfield-described‘dolomites’and‘dolomiticshales’intheprofileasawhole(Fig.17D). ToconfirmwhetherCaCO3contentisthemostappropriatevariable,anobjectivedeterminationofmodesofvariabilitywasmadeusingPrincipalComponentsAnalysisontwodatasets:1)XRFandmagneticsusceptibilityanalyseson40samplescovering80cmofthebasalcycle(Fig.17A)and2)68ICP-AESanalysesbetween50and190minstratigraphicheight,including35oftheabove40samples(Fig.17B). FortheXRFdataset,thefirstprincipalcomponent,accountingfor69%ofdatavariabilitycontrastscarbonate-derivedelements(CaandMn,togetherwithlossonignition,i.e.largelyCO2)withsilicatecomponents(Al,Si,Na,K,Ti,andacid-insolubleresiduederivedfromaseparatealiquot).MagnesiumscaleslargelywiththecarbonatecomponentandFemorewiththesilicatecomponent.Magneticsusceptibilityloadswiththesilicatecomponentbutmoreweaklythantheothervariables.ThesecondprincipalcomponentgroupsFeandS,correspondingtoanindependentvariationinpyritecontent.CalculationofmodalanalysesindicatesthatFeislocatedincarbonates>silicates>>pyrite.FortheICP-AESdataset(referringtoacid-soluble,nottotalchemistry),thefirstprincipalcomponent,accountingfor55%ofvariationcontrastsCa,SrandMnwithFe,Fe/Mn,Mg,Znandinsolubleresidue,implyingthataproportionofFe,MgandZnisleachedfromnon-carbonatephasesduringacidleaching.Bariumdisplaysindependentvariation,beingthemosthighlyloadedelementonthesecondprincipalcomponent(Fig.17B).Minorprincipalcomponentsinbothanalysesdonotshowanystratigraphicallymeaningfulvariation.Incombination,thetwosetsofanalysesindicatethatCaispresentnearlyentirelyincarbonates,andconfirmsthattheCacontent(expressedasCaCO3)ofthetotalsampleshouldprovideaneffectivevariableillustratingcompositionalchange. Takenasawhole,theseobservationsindicatethatthemainmodeofvariationinsedimentcompositionisintermsofvariationinthecarbonatetosilicateratio,butonlylocallyisthereasimplecorrelationbetweenhighcarbonate(Ca)contentandmoreresistantbeds.Insuchcases,enhancedearlycementation,leadingtohighercarbonatecontent,likelycreatedalithologymoreresistanttoweathering.Otherwise,theremustbesubtletexturaldifferencesbetweenmoreandlessweathering-resistantpartsofcycles.Theverysubtletyofthecyclesprovidesreassurancethattheyareindeedofprimaryorigin.InPhanerozoicrhythms,itiscommontofindsignificantdiageneticenhancementofprimarydifferencesincarbonatecontentwhetherbyearlycementation(e.g.Westphaletal.,2000)orviapressuredissolution(Ricken,1986)andinextremecasesrhythmsarelostbyamalgamationofbeds(Hallam,1964).Instead,dolomitecementationispervasiveinE3andappearstodevelopcontinuouslyoverarangeofshallowdepthsinallsedimentarybeds,beingaccompaniedbylocaloverpressuringanddeformationofsurfacesedimentsbyreleasedfluids. Insummary,giventhelimitedrolefordiageneticcarbonateredistributionbetweenlayers,weinferthatthecyclicityismainlyafunctionofaweakprimaryvariationinsedimenttextureand/orcarbonatetosiliciclasticratio.Thelackofclusteringofcyclesorstronggradientsinanydescriptiveparameterwithinthe

14E3sedimentscontrastswithautocyclicsystemswhereinternalprogradationalprocessesgeneratecycleswhosestackingisafunctionprimarilyofaccommodationspaceandwhichthereforecannotcaptureaspectsofexternaldriverssuchashierarchicalstackingofcycles(Pollittetal.2015).Incontrast,E3cyclicsedimentscanbeconfidentlyinterpretedasdisplayingallocyclicity,inwhichasubtlechangeincompositionisimposedonafine-grainedsedimentdepositedfarfromsedimentsourcesRicken(1986),typicallybyclimaticforcing(ElrickandHinnov,2007;Weedon,1993).Wereturntothetimescaleofvariationinthediscussion.

4.3TheE3-E4transitionandtheoriginofmineralpseudomorphsTheoverallsuccessionfromE3toE4(Fig.2)hasbeenpreviouslyinterpretedtobeasingleregressiveunit,culminatinginexposureunderperiglacialconditionscorrespondingtothebaseoftheWilsonbreenFormation(Fairchild&Hambrey,1984).Here,wepresentnewdatamainlyrelatedtotheE3-E4transitionandre-assesstheoriginofpseudomorphsandcrack-fillingsacrosstheintervalofrapidshoaling. Thenorthernthreesectionsshowahomologoussuccessionfromthedolomiticsilt-shalesofthemainpartofE3,toincreasinglywell-sortedsiltstonesarrangedindistinct,thicklaminae(Fig.18A),withthethickestexamplesbeinggradedandsomedisplayingbasalguttercastsorwave-generatedcross-laminae.Locally,desiccationcracksormudclasts(Fig.18C)occur.ThesesedimentspassuptransitionallyintodolarenitesatthebaseofE4,locallyseentocontainparallellaminatedtocross-stratifiedooids.ThistransitionissimilartoathickshallowingupwardsparasequencefromtheEdiacaranofNortheastGreenland(Fairchild&Herrington,1989),wherestorm-dominatedfaciespassupintointermittentlyexposedlagoonalandtidalsandflatfacies.Distinctive,butraresedimentarystructureinbasalE4sedimentsareptygmaticallyfolded,crack-fillings,upto3cmhighand~1mmwide.Theyarefilledwithsedimentfromrelativelycoarsesedimentlayersandcuttingfinerlayers(Fig.18B).Theycanbeclassifiedassub-aqueousshrinkageordiastasiscracks(cf.Fairchild&Herrington,1989;Cowan&James,1992).ThisexplanationcouldalsoapplytothedolomitizedstructuresidentifiedbyHoffmanetal.(2012)fromasimilarhorizoninNordaustlandettothenorthofthestudyareaandidentifiedasmolartoothstructureswhichotherwisearenotfoundafterinitialCryogenianglaciationlocally(Fairchild&Hambrey,1995)orglobally(Kuang,2014;Shields,2002,Shields-Zhouetal.,2012). PseudomorphswererecordedbyFairchild&Hambrey(1984),bothatthetopofE3asnumeroussmallcrystalpseudomorphs,typicallypreservedasmoulds,andwithinE4assomewhatlargernodularstructures,typicallyreplacedbysilica.Anhydriteinclusionswerenotedinthelattercase.Inthecurrentwork,crystalpseudomorphsarefoundtoberestrictedtotopofE3,within10mofthetopatDracoisen,andinanarrowerrangeatDitlovtoppen.Incontrast,Hoffmanetal.(2012)showedthepseudomorphstobedistributedintheupperthirdofE3inalesswell-exposedsectioninNordaustlandettothenorthofourstudyarea.Thepseudomorphsaretypically2-3mmacrossandequant,althoughwitharangefromobtuse,toright-angledandacuteinterfacialangles(Fig.18C),whichHalversonetal.(2004)notedwereconsistentwiththecold-watercarbonateikaite(CaCO3.6H2O),althoughnotdiagnosticofit.Halversonetal.(2004)showedδ18Ointherange-15to-10‰inexamplesfilledbyburialdiageneticcalcitespar.AtReinsryggen,pseudomorphsarealsopreservedlocallyascalcitesparcement,butalsoassociatedwithreplacivesilica(megaquartz),containinginclusionsofanhydrite(Fig.18D).Thecharacteristicallylarger,nodularpseudomorphslocallyfoundinmemberE4consistofamixtureofmegaquartzwithanhydriteinclusions(Fig.18F)andvug-liningferroansaddledolomite(Fig.18E).Thepresenceofanhydriteinclusionsimpliesoriginalgypsum(consistentwithacute-angledpseudomorphs)oranhydrite(consistentwithequantshapeandcastellatedright-angledmarginsofsomepseudomorphs).CalciumsulphateevaporitesareindicativeofamildlyevaporativedepositionalenvironmentconsonantwithaveryshallowwatersettingtransitionaltotheoverlyingintraclasticandooidaldolomiteswithlocalreplacedanhydritenodulesinmemberE4(cf.Fairchild&Harrington,1989).Conversely,thedepositionalsettingisnotonewhereikaitewouldbeexpected(Oehlerichetal.,2013)andcontrastswiththevarvedlimestoneswithice-rafteddebriswhereikaitepseudomorphshavebeenrecognizedintheoverlyingWilsonbreenFormation(Fairchildetal.,2016). Inthesouthernpartofthestudyarea,Halversonetal.(2004)notedtheoccurrenceofalimestoneintervalatleast5mthickatthetopofE3atSlangen(5kmwestofBacklundtoppen)andherewedocumenta20mlimestoneintervalatthenorthBacklundtoppensectionseparatingdolomiticsilt-shalesbelowandooliticdolomitesabove(Fig.2).Sedimentarystructuresinthisintervalincludegutteredbedbases,

15hummockycross-stratification,andtwodistinctmetre-thickstromatolitebiostromehorizons.Thesestructuresareallconsistentwithanenvironmentabovestormwave-baseandtransitionalupwardstoperitidalsediments(e.g.Fairchild&Herrington,1989;Knoll&Swett,1990). MemberE4containsthreedistinctlithologicalunits(Fairchild&Hambrey,1984andFig.2).Thelowest,unitA,consistsofparallel-laminatedandcross-stratifiedooidaldolarenites(A)interpretedastidalsandflatdeposits,andwhichisparticularlythickatBacklundtoppenoverlyingthelimestoneunit.Itisoverlainbytocementedfenestralooidaldolomiteswithtepeestructuresindicativeofsalineartesiangroundwaterdischarge(B)inasupratidalsetting.AtDitlovtoppen,thereisanoverlyingmorevariableunitwithmicritic,locallyshalyandbrecciateddolomites(C),possiblyback-barrierinorigin.Furthernorth,inNordaustlandet,E4istransitionaltoasandstoneunit(theBråvikaSandstone),interpretedasfluvialinoriginbyHoffmanetal.(2012).Depositionofalltheseunitswasfollowedbysubaerialexposureandevidenceofperiglacialconditions:decimetre-scalefoldsinunitCandextensivedevelopmentofparallelfracturenetworksofpreviouslycementeddolomite(Fairchild&Hambrey,1984).Bennetal.(2015)associatethesephenomenawithSnowball-typeglacialconditions,inferringamulti-millionyearhiatus.

5. ChronologyandSrisotopes

5.1DetritalzirconconstraintsonageHopesofconstrainingdepositionalagethroughdetritalzirconageswereraisedbythepresenceofrarealteredvolcanicclastsinE2(Harlandetal.,1993).Also,avolcaniccontributiontotheE3shalesisindicatedbyelevatedTi/Alratios(Kunzmannetal.,2015).DetritalzirconsfromfoursandstonesfromnearthetopofmemberE1haveagesrangingfromca.2.7Matoca.1.0Ga(supplementaryTableS4).Theagedistributionischaracterizedbyabroaddominantmajorpeakfrom1.0Gato1.7Gaandasubordinatepopulationofdatesfrom2.5Gato2.7Ga.ThespectrumoftheE1samplesiscomparabletospectraobtainedfortheMoineSupergroupinScotlandandthemainpeakat1.2GaisprobablylinkedtotheGrenvilleorogen(Cawoodetal.2007).DetritalzirconsinasamplefromtheWilsonbreenFormationexhibitcomparableagespreadtotheE1samples(supplementaryTableS5).Theyoungeststatisticallysignificantsetofdetritalzircondatesfromthesesamplesisca.1.0Ga;youngerdatesarenotconcordantandlikelyreflectpost-crystallisationPb-loss.

5.2SrisotopechemostratigraphyMarinecarbonatesarethemostwidelyusedmaterialsforchemostratigraphyandtheirprimarySrisotopesignaturehasprovedlessambiguousthanδ13Cinglobalcorrelations(Halversonetal.,2010).Inpractice,theratioofradiogenic87Srtostable86SrtendstoincreaseduringdiagenesisbecauseoftheloweringofSrcontentduringcarbonatestabilizationandtheadditionof87SrfromRb-decayinsilicates.CommonscreeningcriteriaarelowRb/Sr,highSr,andlowMn/Sr(Fairchildetal.,2000;Melezhiketal.,2015),thelatterarisingfromtheassumptioninPhanerozoicstudiesthatprimaryMnwouldbenegligibleinconcentration,butthatitwouldincreaseduringmeteoricdiagenesis(Brand&Veizer,1980).However,significantlevelsofprimaryMncanoccurinNeoproterozoicdepositionalenvironments(e.g.Hood&Wallace,2014)andMncanalsobehighinfluidsduringearlymarinediagenesisashasbeenestablishedaboveforE3dolomiticshales.Inourwork,onlylimestoneshadsufficientlyhighSrconcentrationstobeconsidered.ThesewerefoundinboththebasalE3sectionandtheupperE3sectionintheBacklundtoppenarea(Fig.19). ForthebasalE3limestones,87Sr/86SrdecreaseswithincreasingSrcontent(Fig.19A),butthemaximumSrconcentrationisonly705ppm(SupplementaryInformation,TableS1).Mn/Srshowslesssystematicdistribution(Fig.19B)whichisunsurprisinggiventhepetrographicevidenceforgrowthspikesinMn(brightCL)duringgrowthinearlymarinediagenesis.NeverthelessthesamplewithhighestSrandlowestMn/SryieldsthelowestSrisotoperatioof0.70717.Theprimaryseawatersignaturecanbesafelyinferredtobe<0.7072,butcouldberatherlowergiventherelativelylowSrcontentofthesamplesuite.

16 ThelimestonesnearthetopofE3aremoreconventional(allochemicalandstromatolitic)andformthebaseofaregressivepackageimplyingthattheymayhavebeensubjectedtometeoricdiagenesis.Here,awiderrangeofSrconcentrationswasencountered,andthe87Sr/86SrvaluesapproachanasymptotewithhigherSrconcentrations,withalowof0.70767foundinasamplewith1079ppmSr.Thisleastalteredsample,withover1000ppmSrandMn/Sr<0.1wouldpasstherigorousscreeningcriteriausedinpreviousliterature(Fairchildetal.,2000)andaprimarySrisotopesignatureofcloseto0.70765isimplied. ThesenewdatafromSvalbardshowasimilaroveralltrendtovaluesintheLCWIelsewhere,butareslightlymoreradiogenic.Typicalearlypost-Sturtianstrontiumisotopecompositionsare~0.7067–0.7069,rapidlyrisingtoaplateauof0.7071–0.7073,withasingle(pre-Trezonaanomaly)valueashighas0.70735(Kaufmanetal.,1997;Shieldsetal.,1997;McKirdyetal.,2001;Halversonetal.,2007).EarlyEdiacaran87Sr/86Srvaluesarenearlyidenticaltotypicalpre-Marinoanvalues:~0.7072(Jamesetal.,2001;Halversonetal.,2007),buttheserisesharplytoratiosinexcessof0.7077withinthebasalEdiacarancapcarbonatesequence(Halversonetal.,2010).Hence,whereasthelowerE3valuesresemblebasalEdiacaranvalues,upperE3valuesaretoounradiogenictoreflectmiddleEdiacaranvalues.Therefore,eitherthescreeningcriteriaareinvalidforthebasalE3limestones,ortheelevatedvaluesreflectdepositionatatimewhenlocalseawaterwasheavilyinfluencedbyglacialmeltwater.Thelatterhypothesisisconsistentwiththeoxygenisotopeevidencepresentedabove.TheelevatedupperE3valuesaremoreambiguous.Eithertheycaptureahighlyradiogenicpre-Marinoanseawatercompositionotherwisenotseeninothersuccessions,ortheyreflectrestrictionofthebasinandSrreservoirinfluencedbylocalrunoff.WefavourthelatterhypothesisbasedonthemineralogicalevidenceforincreasingrestrictionofthebasindiscussedaboveandthecoherentcarbonisotopeprofilethroughE3–E4thatimpliesthattheserocksentirelypre-datetheTrezonaδ13Canomaly.

6. Discussion

6.1Palaeoenvironmentalevolution

TheCryogenianperiodisthoughttohavehadlowconcentrationsofatmosphericoxygenandhighlyvariablePCO2(Fairchild&Kennedy,2007;Baoetal.,2009).Availableevidencesuggestsanoxicoceans(largelyferruginousbutlocallyeuxinic)beneathanoxygenatedsurfacemixedlayer(Och&Shields-Zhou,2012;Tahataetal.,2015;Sperlingetal.,2015).Thismodel,basedonredoxproxies,isconsistentwithpetrographicevidencefromtheLCWIinSouthAustralia,whereprimarydolomitecementsinareefcomplexwithsignificantverticalreliefrecordgradientsinFeandMn(Hood&Wallace;2014,2015).OurgeochemicalandpetrographicdatafromthemiddleofE3,whichshowsignificantFeandMnenrichment,areconsistentwithanoxicbottomwaters. ThebaseofMemberE3istransitionalwithmarineglacialdepositsandcontinuesintotheshalycarbonatesthatcharacterizemostofthememberanddisplaythecharacteristicmid-Cryogeniannegativetopositiveδ13Cprofile.ThesefeaturesimplythatthatdepositionatthebaseofE3beginsduringthetransitiontoahighstand,consistentwithSturtiancapcarbonatesglobally(Kennedyetal.,1998;Hoffman&Schrag,2002).Thispatterncontrastssharplywiththefallingδ13CprofilerecordedinatransgressivesystemstractatthebaseofMarinoancapcarbonates(Kennedyetal.,1998;HoffmanandSchrag,2002;Halversonetal.,2004).

Thedurationofcapcarbonatedepositionandthemeaningoftheδ13Canomalyhasbeenactivelydebated(Trindadeetal.,2003;Hoffmanetal.,2007;Kennedy&Christie-Blick,2011),butlimitationsonthehydrologicalcyclerevealedthroughmodellingindicatethathighPCO2conditionswouldhavepersistedforuptomillionsofyearspost-glacially(LeHiretal.,2009).Itiscurrentlyunclearwhetherthechangingδ13Cvaluereflectsbathymetryorchangingoceanchemistry(Kaufmanetal.,1997;Giddings&Wallace,2009a).Nevertheless,adistinctstratificationforcedbymeltwaterinfluxisdistinctlypossibleasisactivecarbonateproductionstimulatedbyphotosyntheticblooms(Shields,2005).InthebaseofE3,likeSturtiancapsingeneral,thereisevidencefortherapidstabilizationofmetastablecarbonatetoeitherdolomiteorcalcite,possiblymicrobiallymediated.

17 CyclicshalefacieslikethoseinSvalbardoccurintheimmediatelypost-SturtianTapleyHillFormation,(Giddingsetal.,2009),butthere,asinNamibia,purercarbonatedepositionsubsequentlydominatestheplatformalstratigraphy.Svalbardisunusualthen,inthatconsistentdepositionalenvironmentandlithologyismaintainedthroughmostoftheLCWI.AconsistentshaleenvironmentisalsofoundintheDatangopFormationofSouthChina(Lietal.,2012),butcyclicityisnotdocumented.ThecarbonisotopepatterninE3issimilartoothersectionsoftheLCWI,butroughly2‰lower.Giventhepotentialforbathymetricvariationinδ13C(Giddings&Wallace,2009b)andacomponentofauthigeniccarbonate,thisdifferenceinδ13Cisnotsignificant. Theupperpartofthesuccessionisawell-developedregressivesequence25-40minthickness,culminatinginsubaerialexposure.Conditionsweremildlyevaporiticinmostofthebasin,leadingtothepreservationofpseudomorphsofcalciumsulphates,includingsomepreviouslyinterpretedasikaite.Moretypicalshallowsubtidal,storm-influencedlimestonescharacterizethebaseoftheshallowingupwardsintervalinthesouthofthestudyareaindicativeofamoreopenmarineconnection.TheSouthAustralianandNWCanadiansectionsbetweenglacialsincludeawell-developedprogradationalupperunitcontainingvariablyhighδ13Cvaluesandadistinctnegative(Trezona)anomalythatisalsofoundinseveralotherregions(HalversonandShields-Zhou,2011).Theabsenceofthisfeaturesuggeststop-truncationofthepre-MarinoansuccessioninSvalbard.Thistruncationcouldhavebeenrelativelyminoriftheregressionitselfreflectedtheinitialbuild-upofMarinoaniceathighlatitudes,buttherearenochronostratigraphicconstraintstoconfirmsuchatiming,andHalversonetal.(2002)arguedthattheonsetoftheTrezonaanomalypre-datedglacieustaticsealevelfallrelatedtoonsetofMarinoanglaciation.

TheuppermostsurfaceofE4atthebaseoftheWilsonbreenFormationhasbeeninterpreted(Bennetal.,2015)asamulti-millionyearhiatus,coincidingwithadeep-frozenSnowballEarthconditionculminatinginperiglaciationandglacifluvialerosion,followedbyiceadvancesandretreatsinaterrestrialsetting.ThisrepresentsalatestageintheglaciationwhenPCO2hadrisentohighlevels(Baoetal.,2009;Bennetal.,2015).

6.2AstronomicalforcingandthetimeintervaldurationoftheLCWIThestrikingrhythmicityofE3isinterpretedaboveasallocyclic.GiventheradiometricconstraintsthatthetotaldurationoftheLCWIis5-27Myr,butthatthesuccessioninSvalbardistop-truncated,itislikelytorepresent3-20Myrofdeposition.Givenalsothattherearenearly300cyclesobservedatDitlovtoppen,afirstpassatestimatingcycledurationisthattheyshouldrepresentbetween10kyrand66kyr,withintheMilankovitchband. CyclescorrespondingtoeachofthemainMilankovitchbandsbetween20and400kyrarewell-preservedinappropriatesettingsinthegeologicalrecord(Hinnov,2013;Hilgenetal.,2015).Studyofsuchcyclicityinpelagicsedimentshasprovidedacalibrationof40Ar/39ArgeochronologyandthedefinitiveageoftheMesozoic-Cenozoicboundary(Kuiperetal.,2008).Oldersuccessionsaresubjecttomoreuncertainty,butthereexistalargenumberofsuccessfulstudiesintheMesozoicandincreasinglyinthePalaeozoic(Hinnov,2013)andeventheMesoproterozoic(Zhangetal.,2015)wheremultiplefrequencieshavebeenidentifiedandassignedtoparticularbands.OnesourceofchangeistheeffectofincreasingEarth-Moondistanceovertime,whichisincompletelyknown.However,Waltham(2015)hascalculatedboththechangingperiodicityandestimatedtheassociateduncertainties. Inbothearlyandlate-Cryogeniantimes,NESvalbardisthoughttohavelaininthetropicsorsub-tropics(Lietal.,2013),latitudeswhichareparticularlysusceptibletotheeffectsofprecessionalforcing(e.g.Wangetal.,2008),incontrasttoobliquityforcingathighlatitudes.HenceprecessionalforcingisourpreferredhypothesisfortheE3cycles.IntheoverlyingWilsonbreenFormation,Bennetal.(2015)haveshownthroughclimatemodellingthatglacialadvancesandretreatsinthelatterstagesofaSnowballEarthiceageareconsistentwitharidity-humidityshiftsonprecessionaltimescales.At650Ma,thefourprecessionalfrequencies,whichtodayare18.95,19.1,22.4and23.7kyr,thenwere16.3,16.4,18.8and19.7kyrwithuncertaintiesbetween1.4and1.9kyr(Waltham,2015).Ageneralizedfigureof18kyrwillbeusedhere. TwoexamplesofstrongstratigraphicsignalsofprecessioninPalaeozoicsedimentillustratetheextremesofsedimentationratesinwhichprecessionalsignalsarelikelytobepreserved.TheclassicAptiancoredsectionatPiobbico,Italy(Huangetal.,2010)whichdisplaysprominentprecessionalcyclesin

18greyscalevariationiscomposedofpelagiccarbonatesedimentsthataccumulatedexceptionallyslowlyataround0.5cm/kyr(post-compactionandcorrespondingtoaround10cmpercycle).ThiscontrastswiththeTriassicrift-relatedNewarkGroupwhereshallowingupwardsprecessionalcyclesaround5mthickinlacustrinelithofacieswerecontrolledbyhumidity-aridityvariations(Olsen&Kent,1996).The0.5mmodalthicknessofE3cyclesiscomfortablywithinthisvariationandimpliesadeep-shelfmudrockaccumulationrateofaround2.8cm/kyr,whichiswithintherangeofpelagicsediments.Theinferredaccumulationrateisslightlylowerthanthatofvariousdeep-waterPalaeozoiccycliccarbonateswithlimestone-marlcouplets4-20cmthickthatareinterpretedasmillennialcycles(ElrickandHinnov,2007). NowweattemptamorecarefulestimateoftimedurationrecordedbytheE3-E4succession.TheuppermostpartofE3andmemberE4representaregressivesequencewhichcouldhaveaccumulatedrelativelyquicklyandsoaredisregarded,leaving200mthicknessofE3.Themaximumnumberofobservedcycleswas270,between12mand195mstratigraphicheightatDitlovtoppen,whichwouldscaleto300over200mthickness,providinganestimateoftheminimumnumberofcycles.Anupperboundof400cyclesissetifonetakesthetruemodalthicknessofcyclestobe0.5monaverage,i.e.assumingthatlargerapparentthicknessesareduetomissedcycles.Hence,thedurationofE3canbeestimatedtobeof5.4to7.2Myr(basedon300to400cyclesof18kaduration)whichisclosetotheminimumestimatefromgeochronologydeducedintheintroduction.HoweverinSouthAustralia,NamibiaandNWCanada,theTrezonacarbonisotopeanomalytakesuptheupper5-10%oftheLCWI(Halversonetal.,2005)andthisismissinginSvalbard,implyingthatthedurationshouldbeestimatedinroundfiguresat6to8Myr,assumingthataccommodationspacehadlimiteddepositiononlyintheimmediatelypre-Marinoanperiod.

7. Conclusions

1. ThepresumedcorrelationofmembersE3andE4withtheLCWIissupportedbythediscoveryofawell-developedδ13CprofileandlithologiestypicalofSturtiancapcarbonatesoverlyingmarineglacialdepositsofmemberE2.PrimarySrisotopevaluesdeterminedfromlimestonesinthesouthofthestudyareaof<0.7072atthebaseofE3,and0.7076atthetopareconsistentwiththis.

2. ThepossibleikaitepseudomorphspreviouslydescribedfromthetopofE3,areshowntohaveoriginallybeencalciumsulphate,consistentwiththeevaporativeenvironmentofoverlyingE4.

3. Thesub-E2δ13Canomaly,correlatedwiththepre-MarinoanTrezonaanomalyinHalversonetal.(2004),isnowunderstoodtobethepre-SturtianIslayanomaly,consistentwithSr-isotopeevidence.ThelackofpreservationofaTrezonaanomalyimpliesthatthesectionbetweentheglaciationsistop-truncated,consistentwiththesupratidalsedimentsatthetopofE4.

4. Thesmoothδ13CprofileinE3,despitetheFe-andMn-richnatureofthecarbonates,pointstoanunexpectedlylimitedrolefororganicallymediateddiagenesisintheseshalysediments.ThereisacontrastbetweencarbonatesatthebaseofE3andinthemainpartofthismember.AtthebaseofE3,allcrystalsofdolomiteorcalciteasonehorizonshowthesameCLzonationreplacingalessstableprecursorphase,probablyprecipitatedinthewatercolumn,andwhichcouldbeikaite.InthemainpartofE3,Fe-andMn-richdolomitesshowmuchmorevariableCLpropertiesindicatingdiachronousgrowth,althoughstillduringearlydiagenesisasdemonstratedbylocalnon-compactedconcretions.Consistentδ13Cimpliesbufferingbypre-existingcarbonatephaseandfitswiththeloworganiccontentofthesediments.Fe-andMn-reductionlikelyoccurredinthesurficialzoneorinthewatercolumn,althoughthepresenceofsome13C-depleted,Fe-andMn-rich,pore-fillingcalciteindicatescontinuedFe-andMn-supplydeeperinthesediment.

5. Centimetre-scaledisturbancestructuresinmid-E3arereconstructedtobehorizontalprolapses(sheath-likefolds)ofmicrobiallyboundandunlithifiedsurficialsediments.Theymaybecharacteristicoffluidescapefromunderlyingsedimentsundergoingactivecementation.

6. Subtlevariationsincarbonateandsiliciclasticcontentgiverisetotheallocyclicsedimentaryrhythmsthatarethedominantfeatureofwell-weatheredoutcropsofE3.Wheremostcompletelyexposed,modalcyclethicknessis0.5m.Ifthecyclesareforcedbyprecession,asisconsistentwiththelowpalaeolatitude,thepreservedthicknessoftheintervalbetweenglaciationsrepresentsatimeperiodof6-8Myr.

19AcknowledgementsThisworkwassupportedbytheNERC-fundedprojectGR3/NE/H004963/1GlacialActivityinNeoproterozoicSvalbard(GAINS).Logisticalunderpinningofourhelicopter-supportedfieldworkwasprovidedbytheUniversityCentreinSvalbardandtheSvalbardGovernor’sOffice,facilitatedbyourcollaboratorDougBenn.Weacknowledgethefollowingfortheiressentialhelp:DanCondon(zirconworkatNIGL),ImranRahman(SPIERSsoftware),DavidLimmerandPaulaBoomer(samplepreparation),IanBoomer(stableisotopeanalysis),NickMarsh(XRFanalysis),PaulStanley(SEM)andJohnCraven(ionmicroprobe).PaulHoffmanandNickChristie-Blickarethankedfortheirreferees’comments.ReferencesArnaud,E.&Fairchild,I.J.2011ThePortAskaigFormation,DalradianSupergroup,ScotlandIn:Arnaud,E.,

Halverson,G.P.&Shields-Zhou,G.(eds)TheGeologicalRecordofNeoproterozoicglaciations.GeologicalSocietyofLondon,Memoir,36,635-642.

Bao,H.,Fairchild,I.J.,Wynn,P.M.,Spötl,C.,2009.Stretchingtheenvelopeofpastsurfaceenvironments:NeoproterozoicglaciallakesfromSvalbard.Science,323,119-122.

Benn,D.I.,LeHir,G.,Bao,H.,Donnadieu,Y.,Dumas,C.,Fleming,E.J.,Hambrey,M.J.,McMillan,E.A.,Petronis,M.S.,Ramstein,G.,Stevenson,C.T.E.,Wynn,P.M.,Fairchild,I.J.,2015.OrbitallyforcedIcesheetfluctuationsattheendoftheMarinoanSnowballEarthglaciation.NatureGeoscience,DOI:10.1038/NGEO2502.

Beraldi-Campesi,H.,Farcia-Pichel,F.,2011.Thebiogenicityofmodernterrestrialroll-upstructuresandsignificanceforancientlifeonland.Geobiology,9,10-23.

Bosak,T.,Newman,D.K.2003.MicrobialnucleationofcalciumcarbonateinthePrecambrian.Geology,31,577-580.

Bowring,S.A.,Grotzinger,J.P.,Condon,D.J.,Ramezani,J.,Newall,M.J.,Allen,P.A.,2007.GeochronologicconstraintsonthechronostratigraphicframeworkoftheNeoproterozoicHuqfSupergroupSultanateofOman.AmericanJournalofScience,307,1097–1145.

Brand,U.,Veizer,J.,1980.Chemicaldiagenesisofamulticomponentcarbonatesystem.1.Traceelements.JournalofSedimentaryPetrology,50,1219-1236.

Calver,C.R.,Crowley,J.L.,Wingate,M.T.D.,Evans,D.A.D.,Raub,T.D.,Schmitz,M.D.,2013.GloballysynchronousMarinoandeglaciationindicatedbyU-PbgeochronologyoftheCottonsBreccia,Tasmania,Australia.Geology,10,1127-1130.

Cawood,P.A.,Nemchin,A.A.,Strachan,R.,Prave,T.,Krabbendam,M.,2007.SedimentarybasinanddetritalzirconrecordalongEastLaurentiaandBalticaduringassemblyandbreakupofRodinia.JournaloftheGeologicalSociety,London,164,257-275.

Charlier,B.L.,Ginibre,C.,Morgan,D.,Nowell,D.,Pearson,D.G.,Davidson,J.P.,Ottley,C.J.2006.Methodsforthemicrosamplingandhigh-precisionanalysisofstrontiumandrubidiumisotopesatsinglecrystalforpetrologicalandgeochronologicalapplications.ChemicalGeology,232,114-133.

Chew,S.C.,Kundukad,B.,Seviour,T.,vanderMaarel,J.R.C.,Yang,L,Rice,S.A.,Doyle,P.,Kjelleberg,S.,2014.Dynamicremodellingofmicrobialbiofilmsbyfunctionallydistinctexopolysaccharides.mBio,5(4):e01536-14.doi:10.1128/mBio.01536-14.

Condon,D.,Zhu,M.,Bowring,S.,Wang,W.,Yang,A.,Jin,Y.2005.U-PbagesfromtheNeoproterozoicDoushantuoFormation,China.Science,308,95-98.

Cowan,C.A.andJames,N.P.,1992.Diastasiscracks:mechanicallygeneratedsynaeresis-likecracksinUpperCambrianshallowwaterooliteandribboncarbonates.Sedimentology,39,1101-1118.

Day,E.S.,James,N.P.,Narbonne,G.M.,Dalrymple,R.W.,2004.AsedimentarypreludetoMarinoanglaciation,Cryogenian(MiddleNeoproterozoic)KeeleFormation,MackenzieMountains,northwesternCanada.PrecambrianResearch,133,223-247.

Elrick,M.,Hinnov,L.A.,2007.Millennial-scalepalaeoclimatecyclesrecordedinwidespreaddeeperwaterrhythmitesofNorthAmerica.Palaeogeography,Palaeoclimatology,Palaeoecology,243,348-372.

20Fairchild,I.J.,1991.OriginsofcarbonateinNeoproterozoicstromatolitesandtheidentificationofmodern

analogues.PrecambrianResearch,53,281-299.Fairchild,I.J.,Hambrey,M.J.,1984.TheVendiansuccessionofnortheasternSpitsbergen–petrogenesisofa

dolomite-tilliteassociation.PrecambrianResearch,26,111-167.Fairchild,I.J.,Hambrey,M.J.,1995.VendianbasinevolutioninEastGreenlandandNESvalbard.Precambrian

Research,73,217-233.Fairchild,I.J.,Herrington,P.M.,1989Atempestite-evaporite-stromatoliteassociation(lateVendian,East

Greenland):ashoreface-lagoonmodel.PrecambrianResearch,43,101-127Fairchild,I.J.,Spiro,B.1987.PetrologicalandisotopicimplicationsofsomecontrastingPrecambrian

carbonates,NESpitsbergen.Sedimentology,34,973-989.Fairchild,I.J.,Kennedy,M.J.,2007.NeoproterozoicglaciationintheEarthSystem.JournaloftheGeological

Society,London,164,895-921.Fairchild,I.J.,Spiro,B.,Herrington,P.M.,Song,T.2000.ControlsonSrandCisotopecompositionsof

NeoproterozoicSr-richlimestonesofEGreenlandandNChina,In:Grotzinger,J.P.,James,N.P.(eds.)CarbonateSedimentationandDiagenesisintheEvolvingPrecambrianWorldSEPMSpecialPublication67,297-313.

Fairchild,I.J.,Fleming,E.J.,Bao,H.,Benn,D.I.,Boomer,I.,Dublyansky,Y.V.,Halverson,G.P.,Hambrey,M.J.,Hendy,C.,McMillan,E.A.,Spötl,C.,Stevenson,C.T.E.,Wynn,P.M.,2016ContinentalcarbonatefaciesofaNeoproterozoicpanglaciation,NESvalbard.InrevisionforSedimentology.

Fleming,E.J.,2014.Magnetic,structuralandsedimentologicalanalysisofglacialsediments:insightsfrommodern,QuaternaryandNeoproterozoicenvironments.UniversityofBirmingham,unpubl.PhD.thesis,http://etheses.bham.ac.uk/5136/

Giddings,J.A.,Wallace,M.W.,2009a.SedimentologyandC-isotopegeochemistryofthe‘Sturtian’capcarbonate,SouthAustralia.SedimentaryGeology,216,1-14.

Giddings,J.A.,Wallace,M.W.,2009b.Faciesdependentδ13CvariationfromaCryogenianplatformmargin,SouthAustralia:EvidenceforstratifiedNeoproterozoicoceans?Palaeogeography,Palaeoclimatology,Palaeoecology,271,196-214.

Giddings,J.A.,Wallace,M.W.,Woon,E.M.S.,2009.InterglacialcarbonatesoftheCryogenianUmberatanaGroup,northernFlindersRanges,SouthAustralia.AustralianJournalofEarthSciences56,907–925.

Gorjan,P.,Veevers,J.J.,Walter,M.R.,2000.Neoproterozoicsulfur-isotopevariationinAustraliaandglobalimplications.PrecambrianResearch,100,151–179.

Grey,K.,Yochelson,E.L.,Fedonkin,M.A.,Martin,D.McB.,2010.Horodyskiawilliamsiinewspecies,aMesoproterozoicmacrofossilfromwesternAustralia.PrecambrianResearch,180,1-17.

Hallam,A.,1964.Originofthelimestone-shalerhythmsintheBlueLiasofEngland:Acompositetheory.JournalofGeology,72,157-168.

Halverson,G.P.,2006.ANeoproterozoicChronologyIn:Xiao,S.,Kaufman,A.J.(Eds.)NeoproterozoicGeobiologyandPaleobiology,Springer,NewYork,231-271.

Halverson,G.P.,2011.GlacialsedimentsandassociatedstrataofthePolarisbreenGroup,northeasternSvalbard.In:Arnaud,E.,Halverson,G.P.&Shields-Zhou,G.(eds),TheGeologicalRecordofNeoproterozoicGlaciations.GeologicalSociety,London,Memoirs,36,571-579.

Halverson,G.P.,Shields-Zhou,G.,2011.ChemostratigraphyandtheNeoproterozoicglaciations.In:Arnaud,E.,Halverson,G.P.&Shields-Zhou,G.(eds),TheGeologicalRecordofNeoproterozoicGlaciations.GeologicalSociety,London,Memoirs,36,51-66.

Halverson,G.P.,Hoffman,P.F.,Schrag,D.P.,Kaufman,A.J.,2002.AmajorperturbationofthecarboncyclebeforetheGhuabglaciation(Neoproterozoic)inNamibia:PreludetosnowballEarth?Geochemistry,Geophysics,Geosystems,3,1035.doi:10.1029/2001GC000244.

Halverson,G.P.,Maloof,A.C.,Hoffman,P.F.,2004.TheMarinoanglaciation(Neoproterozoic)innortheastSvalbard.BasinResearch,16,297-324.

Halverson,G.P.,Hoffman,P.F.,Schrag,D.P.,Maloof,A.C.,Rice,A.H.N.,2005.TowardsaNeoproterozoiccompositecarbon-isotoperecord.GeologicalSocietyofAmericaBulletin,117,1181-1207.

Halverson,G.P.,Dudas,F.O.,Maloof,A.C.,Bowring,S.A.,2007.Evolutionofthe87Sr/86SrcompositionofNeoproterozoicseawater.Palaeogeography,Palaeoclimatology,Palaeoecology256,103-129.

21Halverson,G.P.,Wade,B.P.,Hurtgen,M.T.,Barovich,K.M.,2010.Neoproterozoicchemostratigraphy.

PrecambrianResearch,182,337-350.Hambrey,M.J.,1982.LatePrecambriandiamictitesofnortheasternSvalbard.GeologicalMagazine,119,

527-551.Harland,W.B.,Hambrey,M.J.,Waddams,P.,1993.VendiangeologyofSvalbard.NorskPolaristituttSkrifter,

193.Hilgen,F.J.,Hinnov,L.A.,Aziz,H.A.,Abels,H.A.,Batenbrug,S.,Bosmans,J.H.C.,deBoer,B.,Hüsing,S.K.,

Kuiper,K.F.,Lourens,L.J.,Rivera,T.,Ruenter,E.,vandeWal,R.S.W.,Wotzlaw,J.-F.,Zeeden,C.2015Stratigraphiccontinuityandfragmentarysedimentation:thesuccessofcyclostratigraphyaspartofintegratedstratigraphy.In:Smith,D.G.,Bailey,R.J.,Burgess,P.M.,Fraser,A.J.(eds)StrataandTime:ProbingtheGapsinOurUnderstanding.GeologicalSociety,London,SpecialPublications,404,157-197.

Hinnov,L.A.,2013.Cyclostratigraphyanditsrevolutionizingapplicationsintheearthandplanetarysciences.GeologicalSocietyofAmericaBulletin,125,1703-1734.

Hoffman,P.F.,2011.Strangebedfellows:glacialdiamictiteandcapcarbonatefromtheMarinoan(635Ma)glaciationsinNamibia.Sedimentology,58,57-119.

Hoffman,P.F.,Schrag,D.P.,2002.ThesnowballEarthhypothesis:testingthelimitsofglobalchange.TerraNova,14,129-155.

Hoffman,P.F.,Halverson,G.P.2008.OtaviGroupofthewesternNorthernPlatform,theEasternKaokoZoneandthewesternNorthernMarginZone.In:Miller,R.McG.(ed.)TheGeologyofNamibia,vol.2.GeologicalSurveyofNamibia,Windhoek,Namibia,13-69–13-136.

Hoffman,P.F.,Halverson,G.P.,Domack,E.W.,Husson,J.M.,Higgins,J.A.andSchrag,D.P.,2007.ArebasalEdiacarn(635Ma)post-glacial“capdolostones”diachronous?EarthandPlanetaryScienceLetters,258,114-131.

Hoffman,P.F.,Halverson,G.P.,Domack,E.W.,Maloof,A.C.,Swanson-Hysell,N.L.,Cox,G.M.,2012.CryogenianglaciationsonthesoutherntropicalpaleomarginofLaurentia(NESvalbardandEastGreenland),andaprimaryoriginfortheupperRussøya(Islay)carbonisotopeexcursion.PrecambrianRes.,206-207,137-158.

Hoffmann,K.-H.,Condon,D.J.,Bowring,S.A.,Crowley,J.L.2004.U-PbzircondatefromtheNeoproterozoicGhaubFormation,Namibia:ConstraintsonMarinoanglaciation.Geology,32,817-820.

Hofmann,H.J.,Narbonne,G.M.,Aitken,J.D.1990.EdiacarianremainsfromintertillitebedsinnorthwesternCanada.Geology,18,1199–1203.

Hood,A.V.S.,Wallace,M.W.,2014Marinecementsrevealthestructureofananoxic,ferruginousNeoproterozoicocean.JournaloftheGeologicalSociety,London,171,741-744.

Hood,A.V.S,Wallace,M.W.,2015.ExtremeoceananoxiaduringtheLateCroygenianrecordedinreefalcarbonatesofSouthernAustralia.PrecambrianResearch,261,96-111.

Huang,C.,Hinnov,L.,Fischer,A.G.,Grippo,A.,Herbert,T.,2010.AstronomicaltuningoftheAptianStagefromItalianreferencesections.Geology,38,899-902.

Hurtgen,M.T.,Arthur,M.A.,Suits,N.,Kaufman,A.J.,2002.ThesulfurisotopiccompositionofNeoproterozoicseawatersulfate:implicationsforsnowballEarth?EarthandPlanetaryScienceLetters203,413–429.

IUGS(InternationalUnionofGeologicalSciences)2014InternationalCommissiononStratigraphy,AnnualReport2014.http://iugs.org/uploads/ICS%202014.pdf.Accessed16thMarch2015.

Irwin,H.,Curtis,C.,Coleman,M.,1977.Isotopicevidenceforsourceofdiageneticcarbonatesformedduringburialoforganic-richsediments.Nature,269,209-213.

James,N.P.,Narbonne,G.M.,Kyser,T.K.,2001.LateNeoproterozoiccapcarbonates:MackenzieMountains,northwesternCanada:precipitationandglobalglacialmeltdown.CanadianJournalofEarthSciences,38,1229-1262.

Kaufman,A.J.,Knoll,A.H.,Narbonne,G.M.1997.Isotopes,iceages,andterminalProterozoicearthhistory.ProceedingsoftheNationalAcademyofSciences,USA¸94,6600-6605.

Kendall,B.,Creaser,R.A.,Selby,D.2006.Re-Osgeochronologyofpost-glacialblackshalesinAustralia:Constraintsontimingof“Sturtian”glaciation.Geology,34,729-732.

Kendall,B.,Creaser,R.A.,Calver,C.R.,Ruab,T.D.,Evans,D.A.D.,2009.CorrelationofSturtiandiamictitesuccessionsinsouthernAustraliaandnorthwesternAustraliabyRe-Osblackshalegeochronologyand

22

theambiguityof“Sturtian”-typediamictite-capcarbonatepairsaschronostratigraphicmarkerhorizons.PrecambrianResearch,172,301-310.

Kennedy,M.J.,Christie-Blick,N.,2011.CondensationoriginforNeoproterozoiccapcarbonatesduringdeglaciation.Geology,39,319-322.

Kennedy,M.J.,Runnegar,B.,Prave,A.R.,Hoffman,K.-H.,Arthur,M.A.1998.TwoorfourNeoproterozoicglaciations?Geology,26,1059-1063.

Knoll,A.H.,Swett,K.,1990.CarbonatedepositionduringtheLateProterozoicEra:anexamplefromSpitsbergen.AmericanJournalofScience,290-A,104-132.

Knoll,A.H.,Hayes,J.M.,Kaufman,A.J.,Swett,K.,Lambert,I.B.1986.SecularvariationincarbonisotoperatiosfromUpperProterozoicsuccessionsofSvalbardandEastGreenland.Nature,321,832-838.

Kuang,H.-W.,2014Reviewofmolartoothstructureresearch.JournalofPalaeogeography,3,359-383.Kuiper,K.F.,Deino,A.,Hilgen,F.J.,Krigsman,W.,Renne,P.R.,Wijbrans,J.R.,2008.Synchronizingrockclocks

ofEarthhistory.Science,320,500-504.Kunzmann,M.,Halverson,G.P.,Scott,C.,Minarik,W.G.,Wing,B.A.,2015.GeochemistryofNeoproterozoic

blackshalesfromSvalbard:ImplicationsforoceanicredoxconditionsspanningCryogenianglaciations.ChemicalGeology,417,383-393.

Lan,Z.,Li,X.,Zhu,M.,Chen,Z.-Q.,Zhang,Q.,Li,Q.,Lu,D.,Liu,Y.,Tang,G.,2014.ArapidandsynchronousinitiationofthewidespreadCryogenianglaciations.PrecambrianResearch,255,401-411.

Land,L.S.,1980.Theisotopicandtraceelementgeochemistryofdolomite:thestateoftheart.InZenger,D.H.,Dunham,J.B.&Ethington,R.L.(eds.),ConceptsandModelsofDolomitization,SEPMSpecialPublication,28,87-110.

Lawrence,M.J.F.,Hendy,C.H.,1985.WatercolumnandsedimentcharacteristicsofLakeFryxell,TaylorValley,Antarctica.NewZealandJournalofGeologyandGeophysics,28,543-552.

LeHeron,D.P.,Cox,G.M.,Trundley,A.E.,Collins,A.,2011.Sea-icefreeconditionsduringtheearlyCryogenian(Sturt)glaciation,SouthAustralia.Geology,39,31–34.

LeHeron,D.P.,Busfield,M.,Kamona,F.,2013.AninterglacialonSnowballEarth?DynamicicebehaviourrevealedintheChuosFormation,Namibia.Sedimentology,60,411-427.

LeHir,G.,Donnadieu,Y.,Goddéris,Y.,Pierrehumbert,R.T.,Halverson,G.P.,Macouin,M.,Nédélec,A.,Ramstein,G.,2009.TheSnowballEarthaftermath:Exploringthelimitsofcontinentalweatheringprocesses.EarthandPlanetaryScienceLetters,277,453-463.

Leather,J.,Allen,P.A.,Brasier,M.D.,Cozzi,A.2002.NeoproterozoicsnowballEarthunderscrutiny:EvidenceformtheFiqglaciationofOman.Geology,30,891-894.

Li,C.,Love,G.D.,Lyons,T.W.,Scott,C.T.,Feng,L.,Huang,J.,Chang,H.,Zhang,Q.,Chu,X.,2012.EvidenceforaredoxstratifiedCryogenianmarinebasin,DatangpoFormation,SouthChina.EarthandPlanetaryScienceLetters,331-332,246-256.

Li,Z.-X.,Evans,D.A.D.,Halverson,G.P.,2013.NeoproterozoicglaciationsinarevisedglobalpalaeogeographyfromthebreakupofRodiniatotheassemblyofGondwanaland.SedimentaryGeology,294,219-232.

Macdonald,F.A.,Jones,D.S.,Schrag,D.P.,2009.Stratigraphicandtectonicimplicationsofanewlydiscoveredglacialdiamictite-capcarbonatecoupletinsouthwesternMongolia.Geology,37,123-126.

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.CalibratingtheCryogenian.Science327,1241–1243.

Mazzullo,S.J.,2000.Organogenicdolomitizationinperitidaltodeep-seasediments.JournalofSedimentaryResearch,70A,10-23.

McKirdy,D.M.,Burgess,J.M.,Lemon,N.M.,Yu,X.,Cooper,A.M.,Gostin,V.A.,Jenkins,R.J.F.,Both,R.A.2001.AchemostratigraphicoverviewofthelateCryogenianinterglacialsequenceintheAdelaideFold-ThrustBelt,SouthAustralia.PrecambianResearch,106,149-186.

Melezhik,V.A.,Ihlen,P.M.,Kuznetsov,A.B.,Gjelle,S.,Solli,A.,Gorokhov,I.M.,Fallick,A.E.,Sandstad,J.S.,Bjerkgård,T.,2015.Pre-Sturtian(800-730Ma)depositionalageofcarbonatesinsedimentarysequenceshostingstratiformironoresintheUppermostAllochthonoftheNorwegianCaledonides:Achemostratigraphicapproach.PrecambrianResearch,261,272-299.

Mortimer,R.J.G.,Coleman,M.L.,1997.Microbialinfluenceontheoxygenisotopiccompositionofdiageneticsiderite.GeochimicaCosmochimicaActa,61,1705-1711.

23Mozley,P.S.,Burns,S.J.,1993.Oxygenandcarbonisotopiccompositionofmarinecarbonateconcretions:

Anoverview.JournalofSedimentaryPetrology,63,73-83.Narbonne,G.M.,Aitken,J.D.,1995.NeoproterozoicoftheMackenzieMountains,northwesternCanada.

PrecambrianResearch,73,101–121.Och,L.M.,Shields-Zhou,G.A.,2012.TheNeoproterozoicoxygenationevent:Environmentalperturbations

andbiogeochemicalcycling.Earth-ScienceReviews,110,26-57.Oehlerich,M.,Mayr,C.,Griesshaber,E.,Lücke,A.,Oeckler,O.M.,Ohlendorf,C.,Schmahl,W.W.,Zolitschka,

B.,2013.Ikaiteprecipitationinalacustrineenvironment–implicationsforpalaeoclimaticstudiesusingcarbonatesfromLagunaPotrokAIke(Patagonia,Argentina).QuaternaryScienceReviews,71,46-53.

Olsen,P.E.,Kent,D.V.,1996.MilankovitchclimateforcinginthetropicsofPangaeaduringthelateTriassic.Palaeogeography,Palaeoclimatology,Palaeoecology,122,1-26.

Pollitt,D.A.,Burgess,P.M.,Wright,V.P.,2015.Investigatingtheoccurrenceofhierarchiesofcyclicityinplatformcarbonates.In:Smith,D.G.,Bailey,R.J.,Burgess,P.M.,Fraser,A.J.(eds)StrataandTime:ProbingtheGapsinOurUnderstanding.GeologicalSociety,London,SpecialPublications,404,123-150.

Porada,H.,Bouougri,ElH.,2007.Wrinklestructures–acriticalreview.Earth-ScienceReviews,81,199-215.Pruss,S.B.,Bosak,T.,Macdonald,F.A.,McLane,M.,Hoffman,P.F.,2010.MicrobialfaciesinaSturtiancap

carbonate,theRasthofFormation,OtaviGroup,northernNamibia.PrecambrianResearch,181,187-198.

Ricken,W.,1986.DiageneticBedding.AmodelforMarl-LimestoneAlternations.LectureNotesinEarthSciences,6.Springer-Verlag,Berlin.

RIcken,W.,1996.Beddingrhythmsandcyclicsequencesasdocumentedinorganiccarbon-carbonatepatterns,UpperCretaceous,WesternInterior,U.S.SedimentaryGeology,102,131-154.

Riedman,L.A.,Porter,S.M.,Halverson,G.P.,Hurtgen,M.T.,Junium,C.K.,2014.Organic-walledmicrofossilassemblagesfromglacialandinterglacialNeoproterozoicunitsofAustraliaandSvalbard.Geology,42,1011-1014.

Rieu,R.,Allen,P.A.,Cozzi,A.,Kosler,J.,Bussy,F.,2007a.AcompositestratigraphyfortheNeoproterozoicHuqfSupergroupofOman:integratingnewlitho-,chemo-andchronostratigraphydataoftheMirbatarea,southernOman.JournaloftheGeologicalSociety,London,164,997-1009.

Rieu,R.,Allen,P.A.,Plotze,M.,Pettke,T.,2007b.CompositionalandmineralogicalvariationsinaNeoproterozoicglaciallyinfluencedsuccession,Mirbatarea,southOman:Implicationsforpaleoweatheringconditions.PrecambrianResearch,154,248-265.

Rimstidt,J.D.,Balog,A.,Webb,J.,1998.Distributionoftraceelementsbetweencarbonatemineralsandaqueoussolutions.GeochimicaetCosmochimicaActa,62,1851-1863.

Roberts,N.M.,Parrish,R.R.,Horstwood,M.S.,Brewer,T.S.,2011.The1.23GaFjellhovdanerhyolite,Grøssæ-Totak;anewagewithintheTelemarksupracrustals,southernNorway.NorwegianJournalofGeology,91,239-246.Rooney,A.D.,Macdonald,F.A.,Strauss,J.V.,Dudás,F.O.,Hallmann,C.,Selby,D.,2014.Re-OsgeochronologyandcoupledOs-SrisotopeconstraintsontheSturtiansnowballEarth.ProceedingsoftheNationalAcademyofSciences,111,51-56.

Rooney,A.D.,Strauss,J.V.,Brandon,A.D.,Macdonald,F.A.,2015.ACryogenianchronology:Twolong-lastingsynchronousNeoproterozoicglaciations.Geology,43,459-462.

Rose,C.V.,Swanson-Hysell,N.L.,Husson,J.M.,Poppick,L.N.,Cottle,J.M.,Schoene,Maloof,A.C.,2012.ConstraintsontheoriginandrelativetimingoftheTrezonaδ13Canomalybelowtheend-Cryogenianglaciation.EarthandPlanetaryScienceLetters,319-320,241-250.

Rose,C.V.,Maloof,A.C.,Schoene,B.,Ewing,R.C.,Linnemann,U.,Hofmann,M.,Cottle,J.M.,2013.Theend-CryogenianglaciationofSouthAustralia.GeoscienceCanada,40,256-293.

Sarkar,S.,Banerjee,S.,Smanta,P.,Chakraborty,N.,Chakraborty,P.P.,Mukhopaghyay,S.,Singh,A.K.,2014.Microbialmatrecordsinsiliciclasticrocks:ExamplesfromfourIndianProterozoicbasinsandtheirmodernequivalentsinGulfofCambay.JournalofAsianEarthSciences,91,362-377.

Schieber,J.,1986.ThepossibleroleofbenthicmicrobialmatsduringtheformationofcarbonaceousshalesinshallowMid-Proterozoicbasins.Sedimentology,33,521-536.

Schieber,J.,Southard,J.B.,Schimmelmann,A.,2010.Lenticularshalefabricsresultingfromintermittenterosionofwater-richmuds–interpretingtherockrecordinthelightofrecentflumeexperiments.JournalofSedimentaryResearch,80,119-128.

24Schrag,D.P.,Higgins,J.A.,Macdonald,F.A.,Johnston,D.T.,2013.Authigeniccarbonateandthehistoryof

theglobalcarboncycle.Science,339,540543.Selleck,B.W.,Carr,P.F.,Jones,B.G.,2007.Areviewandsynthesisofglendonites(pseudomorphsafter

ikaite)withnewdata:assessingapplicabilityasrecordersofancientcoldwaterconditions.JournalofSedimentaryResearch,77,980-991.

Shanahan,T.M.,Overpeck,J.T.,Beck,J.W.,Wheeler,C.W.,Peck,J.A.,King,J.W.,Scholz,C.A.,2008.TheformationofbiogeochemicallaminationsinLakeBosumtwi,Ghana,andtheirusefulnessasindicatorsofpastenvironmentalchanges.JournalofPaleolimnology,40,339-355.

Shearman,D.J.,Smith,A.J.,1985.Ikaite,theparentmineralofjarrowite-typepseudomorphs.ProceedingsoftheGeologistsAssociation,96,305-314.

Shields,G.A.2002.‘Molar-toothmicrospar’:achemicalexplanationforitsdisappearance~750Ma.TerraNova,14,108-113.

Shields,G.A.2005.Neoproterozoiccapcarbonates:acriticalappraisalofexistingmodelsandtheplumeworldhypothesis.TerraNova,17,299-310.

Shields,G.A.,Stille,P.,Brasier,M.D.,Atudorei,N.-V.,1997.StratifiedoceansandoxygenationofthelatePrecambrianenvironments:ApostglacialgeochemicalrecordfromtheNeoproterozoic.TerraNova,9,218-222.

Shields-Zhou,G.A.,Hill,A.C.,Macgabhann,B.A.,2012.Chapter17—TheCryogenianPeriod.In:Gradstein,F.M.,Ogg,J.G.,Schmitz,M.D.,Ogg,G.M.,(eds).TheGeologicTimeScale.Boston:Elsevier,393-411.

Simonson,B.M.,Carney,K.E.,1999.Roll-upstructures:EvidenceofinsitumicrobialmatsinLateArcheandeepshelfenvironments.Palaios,14,13-24.

Spence,G.H.,LeHeron,D.P.,Fairchild,I.J.,2016.Sedimentologicalperspectivesonclimatic,atmosphericandenvironmentalchangeintheNeoproterozoicEra.Sedimentology,

Spencer,A.M.,1971.LatePre-CambrianGlaciationinScotland.GeologicalSociety,London,Memoirs,6.Sperling,E.A.,Wolock,C.J.,Morgan,A.S.,Gill,B.C.,Kunzmann,M.,Halverson,G.P.,Macdonald,F.A.,Knoll,

A.H.,Johnston,D.T.,2015.StatisticalanalysisofirongeochemicaldatasuggestslimitedlateProterozoicoxygenation.Nature,523,451-454.

Sutton,M.D.,Garwood,R.J.,Siveter,DavidJ.,Siveter,Derek,J.,2012.SpiersandVAXML;Asoftwaretoolkitfortomographicvisualisation,andaformatforvirtualspecimeninterchange.PalaeontologiaElectronicaArticle:15.2.5T.

Tahata,M.,Sawaki,Y.,Yoshiya,K.,Nishizawa,M.,Komiya,T.,Hirata,T.,Yoshida,N.,Maruyama,S.,Windley,B.F.,2015.ThemarineenvironmentsencompassingtheNeoproterozoicglaciations:EvidencefromC,SrandFeisotoperatiosintheHeclaHoekSupergroupinSvalbard.PrecambrianResearch,263,19-42.

Trindade,R.I.F.,Font,E.,D’Agrella-Filho,A.S.,Nogueira,A.C.R.,Riccomini,C.2003.Low-latitudeandmultiplegeomagneticreversalsintheNeoproterozoicPugacapcarbonate,Amazoncraton.TerraNova,15,441-446.

Veizer,J.,1993.Traceelementsandstableisotopesinsedimentarycarbonates.In:Reeder,R.J.(ed.)Carbonates:MineralogyandChemistry.MineralogicalSocietyofAmerica,Blacksburg,265-299.

Wallace,M.W.,Hood,A.V.S.,Woon,E.M.S.,Giddings,J.A.,Fromhld,T.A.,2015.TheCryogenianBalcanoonareefcomplexesoftheNorthernFlindersRanges:ImplicationsforNeoproterozoicoceanchemistry.Palaeogeography,Palaeoclimatology,Palaeoecology,417,320-336.

Wang,Y.,Cheng,H.,Edwards,R.L.,Kong,X.,Shao,X.,Chen,S.,Wu,J.,Jiang,X.,Wang,X.andAn,Z.,2008.Millennial-andorbital-scalechangesintheEastAsianmonsoonoverthepast224,000years.Nature,451,1090-1093.

Waltham,D.,2015.Milankovitchperioduncertaintiesandtheirimpactoncyclostratigraphy.JournalofSedimentaryResearch85,990–998.

Weedon,G.,2003.Time-SeriesAnalysisandCyclostratigraphy.CUP.Westphal,H.,Head,M.J.,Munnecke,A.,2000.Differentialdiagenesisofrhythmiclimestonealternations

supportedbypalynologicalevidence.JournalofSedimentaryResearch70,715–725.Williams,G.E.,Gostin,V.A.,McKirdy,D.M.,Preiss,W.V.,2008.TheElatinaglaciations(MarinoanEpoch),

SouthAustralis:Sedimentaryfaciesandpalaeoenvironments.PrecambrianResearch,163,307-331.Zhang,F.,Xu,H.,Konishi,H.,Kemp,J.M.,Roden,E.E.,Shen,Z.,2012.Dissolvedsulphide-catalyzed

precipitationofdisordereddolomite:Implicationsfortheformationmechanismofsedimentary

25

dolomite.GeochimicaetCosmochimicaActa,97,148-165.Zhang,S.,Jiang,G.,Han,Y.,2008.TheageoftheNantuoFormationandNantuoglaciationinSouthChina.

TerraNova,20,289-294.Zhang,S.,Wang,X.,Hammarlund,E.U.,Wang,H.,Costa,M.M.,BjerrumC.J.,Connelly,J.N.,Zhang,B.,Bian,

L.,Canfield,D.E.,2015.Orbitalforcingofclimate1.4billionyearsago.ProceedingsoftheNationalAcademyofSciences,112,E1406–E1413,doi:10.1073/pnas.1502239112.

Zhou,C.,Tucker,R.,Xiao,S.,Peng,Z.,Yuan,X.,Chen,Z.,2004.NewconstraintsontheagesofNeoproterozoicglaciationsinsouthChina.Geology,32,437-440.

26Table1:SummarystratigraphyandpalaeoenvironmentsofthePolarisbreenGroup,afterHambrey(1982),Halverson(2011)andBennetal.(2015).GlacigenicunitsareshowninredandtheE3-E4interval,thesubjectofthispaper,isshowninbold.GeologicalSystem Group Formation Member Thickness

(m) Lithologies Interpretedenvironment

Ediacaran

Polarisbreen

Dracoisen D1toD7 465

Capcarbonatepassingupintoshalethensandstoneand

mudstonealternationswithacentralandacapping

dolomiteunit

Transgressivecoastaltooffshore,becomingplayalake,thencoastal

Cryogenian

Wilson

breen

W1toW3 130-180

Diamictitesandsandstoneswithlimestoneanddolomite

unitsintheW2andW3overlyingsandstonesover

brecciateddolomite

Basalperiglaciatedsurface,locally

succeededbyfluvialdeposits,then

glacilacustrineandglacifluvialdeposits

Elbobreen

E4(Slangen) 20-30 Ooliticdolomite Regressiveperitidal

E3(Macdonaldryggen) 200

Finelylaminateddolomiticsiltyshale Offshoremarine

E2(Petrovbreen) 10-20Dolomiticdiamictites,

rhythmitesandconglomerates

Glacimarine

baseat720Malikelytolieinthismember

E1(Russøya) 75-170

Dolomitesoverlainbylimestonewithmolartoothstructure,blackshaleand

dolomite

Shallowmarine

27Figurecaptions12Figure1.(left)LocationoftheSvalbardarchipelagowiththestudyregionshownasarectangle(enlargedon3right)onthemainisland,Spitsbergen.Intherightdiagramrockoutcrops(nunataks,grey)aresurroundedbyice4andsnow(white).AbbreviationsDRA(Dracoisen),DIT(Ditlovtoppen),AND(Andromedafjellet),REIN(informal5name,Reinsryggen),BACN(Backlundtoppen-KvitfjellaridgewhereasectionofupperE3andE4isexposed),6BACS(SouthBacklundtoppen,whereasectionofE2andbasalE3isexposed).78Figure2.StratigraphicprofilesthroughElbobreenFormation,E3(MacdonaldryggenMember)andE4(Slangen9Member)atlocationsshowninFigure1.1011Figure3.E2-E3transitionalfaciescontainingglacigenicsedimentfromDitlovtoppen(F)andsouth12Backlundtoppen(allothers).A.Thinsectionofgradedrhythmites,4mbelowthetopofE2.B.Limestone13rhythmitessharplyoverlainbyaneventbedwithsandybase,becomingpebblyupwards(1mbelowtopofE2).14C.Polishedhandspecimenexhibitingprominentdiamictite(till)pellet(upperright)andice-raftedsediment15withinlaminitescontainingauthigenicdolomite(5cmabovebaseofE3).D.Limestonelaminitesshowing16sedimentarydeformation(samehorizonasB.)E.Thinsectionoflimestonerhythmites(asB,D)showing17microsparlaminaewithinterveningdolomicriticlaminaecontainsedimentgrains.Prominentcentralstylolite.F.18Precipitateddolomitelaminiteswithdropstones,5cmabovebaseofE3.G,H.SamesampleasC.inthinsection19undertransmittedlight(G)andcathodoluminescence(H).Detritusisdolomitesiltandsandwithvariable20luminescencecharacteristics,quartz(black)andfeldspar(blue)silt.Authigenicdolomiteisabundantinthe21matrixandanearlybrightcementzonesurroundsclasts.2223Figure4.LimestonesfromSouthBacklundtoppen,1mabovethebaseofE3.A.Fieldphotographwithrecumbent24soft-sedimentfolds(examplesoffoldnosesarearrowed).B.Thinsectiondisplayingpalemicrosparlaminaeand25recumbentfoldnoses(arrows).C.andD.transmittedlightandcathodoluminescencerespectively.Microspar26laminaewithslightlybulboustopsconsistofmosaicsofrhombiccalcitecrystalswithaconsistent27cathodoluminescencezonationindicatingsimultaneousreplacementofaprecursorphase.28

Figure5.DolomitizedbasalE3facies,2.2mabovebaseofE3,Reinsryggen.A.Thinsectiondisplayingpale29dolomicrosparlayersalternatingwithdolomicritewithsoft-sedimentfolds.B.Photomicrographintransmitted30lightshowingbrecciationinto“pseudo-allochems”thatoccursintheregionoffoldnosesandwhichcreatespore31spaces(white)laterfilledbycement.C.andD.pairedtransmittedlightandcathodoluminescence32photomicrographsshowinggrowthofreplacivedolomicrosparcrystalswithconsistentluminescencezonation.E.33andF.pairedtransmittedlightandcathodoluminescencephotomicrographsshowingreplacivedolomicrosparof34“pseudo-allochems”passingintoclear,non-luminescentdolomitecement.35

Figure6.OxygenandcarbonateisotopeplotofE3andE4samples.36

Figure7.StratigraphicvariationinsedimentchemistryaroundtheE2-E3transitionandlowerE3.Carbonisotopes37showasteadyrisefromthebaseoftheformation.Extremevariabilityisrestrictedtothebasalbed.Reinsryggen38dolomiteshavelessnegativevaluesthanothersitesatthesamestratigraphicposition.InsolubleresidueandFe39displayparalleltrends(mirroredasplotted)withaninitialdecreasecorrespondingtopurercarbonates,followed40laterbyanincreasetothemixedcarbonate-siliciclasticsedimentcharacteristicofthebulkofE3.41

Figure8.RhythmicsedimentationinmemberE3.A.Dracoisensectionviewedfromthesouth,startingwith42beginningofexposureapproximately40mabovethebaseofE3.Doubleyellowlinesshowdirectionofbedding.43Flat-toppedhilltorightisunconformableCarboniferouscover.B.Enlargementofboxedareain(A)illustrating44variabledegreeofexpressionofsedimentaryrhythms.C.Twoorange-weatheringdolostonehorizons,theupper45oneofwhichwedgesouttorightofthe35cmhammer(ca.150mabovebaseofDitlovtoppensection).D.Non-46resistantpartofrhythm(93.8m,Dracoisen);heightofview45cm.E.Resistantpartofrhythm(97.4m,47Dracoisen);heightofview45cm.F.Distinctlyconcretionarydolostonehorizoncontainingsedimentarylayers48thatthinbyafactoroftwolaterallyintodolomiticshale(42m;Dracoisensection);ruler19cmlong.49

28Figure9.Transmittedlightpetrographyofdolomiticsilt-shalesfromthemainpartofE3intheDracoisensection.50A-C.scaleisinmillimetres.A.Samplewithrelativelyhigh(60%)carbonatecontentformingdolomite-andcalcite-51cementedconcretionat42.75m).Laminationisparallel,butmicro-nodularonsub-millimetrescale.B.Sample52with50%carbonate(dolomite>>calcite)consistingofcarbonatedomainswithfinesiltseparatedbyclay-rich53laminaethatshowcomplexdeformation(disturbancestructures)at67m.C.Samplecontaining70%dolomite54andalsorichinquartzo-feldspathicsilt(187m)showingincipientcross-laminationandverythinlensesofcoarser55silt;suchsilt-richsedimentsareonlyfoundintheE3-E4transitionzone.DandE.Samesampleasin(A)56illustratingdiscontinuousclay-pyriteflaserlaminaeseparatedbydomainsoffinesiltwithcarbonate.F.Same57sampleas(B)illustratingdeformedlaminae.G.Samplewith35-40%carbonate(dolomite)withmuchsiliciclastic58siltanddistinctdeformedclay-richlaminae,106.9m.H.Samplewith30%carbonateasdolomitewithabundant59fine-mediumsiliciclasticsilt,173m.60

Figure10.Disturbancestructures.A-C.Beddingplaneoutcropviews.A.Typicalequantstructures1-2cmacross,61Dracoisen65m.B.Close-upoftwostructureswithdeformedlaminaevisible(widthofphoto3cm),62Ditlovtoppen,40m.C.Lineararrayofstructures,Dracoisen,65m.D-F.Verticalcross-sections.D.Smoothly63weatheredoutcropwithdisturbancestructureboundedaboveandbelowbyundisturbedlamination,Klofjellet,64exactstratigraphicpositionunknown.E.Weatheredoutcropshowingseveralstructuresverticallyoffsetfrom65eachother(mmscaleatlowerleft),Dracoisen,45m.F.Polishedsurfaceofsampleusedforserialsections,66Dracoisen,53m.G-J.Computermodelfromserialsectioningofsampleshownin(F);letteringindicatessnapshot67frommovieinsupplementaryinformation.G.Conicalcoretostructureseeninverticalsection.H.As(G)rotated68through90°anddisplayingdouble-pointedtermination.I.As(H),butalsoshowingenclosing(red)lamina.J.69Completemodelseenfromreversesidewithsheath-likelaminaeinthecoreandcontinuousoverlyinggreen70laminaoverlainbyfurtherdeformedlaminae.71

Figure11.TrainsofdisturbancestructuresfromKlofjellet(exacthorizonunknown).A.Cuthandspecimen72displayingcontinuouslaminationinterruptedbyvariousstructuresincludingatrainclimbingfrom1to3.Apale73laminaat2isnearlyundisturbedacrossthetrain.B-C.Modelfrom31serialsections(totalling0.62mmdepth)of74thelowerrightportionofthesample.B.Viewofmodelfromfrontofslabshowingtubularnatureofstructures75(cf.A).C.Transverseviewofmodelwithfrontatarrowheadtoright,illustratingnarrowingofstructuresintothe76slabonleftofimage.D.Anothercutfromsamesampleillustratingbothdisturbancestructuresandbrittlefailure77(e.g.arrowed).Examplesoflocalvariationsinlaminathicknessarenumbered.Diagonallinesaresawmarks.78

Figure12.StratigraphicvariationwithinE3ofcarbonatestableisotopevalues.KeyasinFig.6.79

Figure13.StratigraphicvariationwithinE3of%insolubleresidue(non-carbonate)andacid-solubleFe,expressed80asFeCO3.KeyasforFig.6.81

Figure14.E3samplesfromDracoisensectionviewunderCL(A-C)andbyBSE(D-F).A.Sampleat48.7mwith40-8245%carbonate(dolomite>>calcite).Siliciclasticsiltincludesquartz(dark),feldspar(blue)andapatite(green).83Zonationisinconsistentbetweendolomitecrystals,i.e.theydidnotgrowsimultaneously.B.Sampleat46.55m84with45%carbonate(dolomite).Similarto(A)butwithmoresegregationofquartzo-feldspathicdetritusand85dolomiteintolaminae.C.Sampleat68.55mwith50%carbonate(subequaldolomiteandcalcite).Finertexture86than(A)and(B)withfine-to-mediumquartzsiltandheterogenouszonationincarbonatecrystals.D.Sampleat8742.75mwith60%carbonate(subequaldolomiteandcalcite),visiblyconcretionaryinthefield(Fig.8F).Dolomite88crystalstendtobeisolatedandenclosedbycalcite.E.andF.Sampleat177mwith30%carbonate(dolomite).89Dolomiteformssubhedralcrystalswithbrighter(moreFe-rich)rims;calcitetendstoshowmoreirregularor90enclosingshapes;pyriteframboidsareconspicuous.C=calcite,D=dolomite,F=feldspar,P=pyrite,Q=quartz.91

Figure15.Ionmicroprobedata(fulldatainSupplementaryinformation,TableS2)fromindividualcarbonate92crystalsperformedbydownwardablationofcarbonatecrystalsto100µmdepthin5µmsteps.Analysesare93thusbiasedtowardscrystalcentresandnoinformationisavailableaboutrelativeageofzones.A.Exampleof94sampleF7163(42.75m,Dracoisen)wherepointsjoinedwithlinesaredepthtraversesofsinglecrystals.Overall95thereisoverlapinchemistryofdolomiteandcalcitewithordersofmagnitudevariationinabundance,and96

29tendencyforvariationinFeorbothFe&Mnwithinindividualcrystals.B.comparisonofmeansofdolomiteand97calcitecrystalsfromionmicroprobeanalysiswithICPdataforacid-solubleFeandMn,bothexpressedaswt.%98carbonate.ICPmeansarehighinFeduetosomeleachingofFefromsilicatephases.SamplesarefromDracoisen99sectionatthefollowingstratigraphicpositions:109(48.3m,seealsoFigs.9A,D,E,14D),F7199(106.9m,see100alsoFig.9G),F7182(68.55m,seealsoFig.14C),F7163(42.75m),F7266(177.0m,seealsoFig.14E,F).101

Figure16.SedimentaryrhythmswithinmemberE3.A.Comparisonofcumulativethicknessofindividualcyclesat102theDracoisenandDitlovtoppenlocalities.B.andC.HistogramsofcyclethicknessattheDracoisenand103Ditlovtoppenlocations(numbersbeneathbinsindicateupperendofbinrange).104

Figure17.GeochemistryofE3rhythmicsedimentsfromtheDracoisensection.A.andB.showresultsfrom105PrincipalComponentsAnalysisofXRFandICPdatarespectively.Thediagramsdisplaytheweightingonthefirst106principalcomponent(F1)onthex-axis,togetherwiththe%ofdatavariation,plottedagainstF2onthey-axis.In107bothcaseCaprovidesthebestindicatorofsamplevariation,mappingnearlyperfectlyontoF1.C.VariationofCa108intotalsample(ICPanalysis)fromadetailedsamplingsectionthroughsevenfield-definedrhythms(XRFdataare109notplotted,butareverysimilar).Theintensivelysampledharderbedataround45.5misenrichedinCa,butat110higherlevelsthereisalackofclearrelationshipwithrhythms.D.VariationofCaintotalsample(ICPanalysis)111fromhigherhorizonsinE3intheDracoisensectionAtmosthorizonsthereisalackofsystematicdifferenceinCa112contentbetweenthefield-definedlithologies.113

Figure18.PhenomenaaroundtheE3-E4transition.A.Polishedhandspecimenillustratingdolomiticsiltstones114withdistinctcoarserandfiner(dark)laminaeandcommon2-3mmvugsrepresentingcrystalpseudomorphs.115Scalebarincm;9mbelowtopofE3,Dracoisen.B.Siltylaminateddolareniteswithripple-laminatedthingraded116unitsandptygmaticallyfoldedsediment-filledcontractioncracks.2mabovebaseofE4,Reinsryggen.C.117Dolomiticsiltstonewithdark-colouredelongatedmudclastsandabundantcrystalpseudomorphs,someopen,118somefilledbycalcitewithminorquartz.0.5mbelowthetopofE3,Reinsryggen.D.Samesampleas(C)119illustratingdiageneticquartzcontaininganhydriteinclusions(e.g.arrowed).E.Fine-graineddolarenites120containingnodularvuggypseudomorphsnowcomposedofquartz(white)orferroandolomite(orange).Scale121barincm,baseofE4,Ditlovtoppen.F.Samesampleas(E)illustratinganhydriteinclusions(e.g.arrowed)in122diageneticquartz.123

Figure19.StrontiumisotopedatafromE3limestonesinBacklundtoppensections.A.87Sr/86SrversusSr.B.12487Sr/86SrversusMn/Sr.125

Supplementarydata126

SupplementarydataTableS1.Stableisotope,Srisotope,ICPandmagneticsusceptibilitydata127

SupplementarydataTableS2.Ionmicroprobedata128

SupplementarydataTableS3.X-rayfluorescencedata129

SupplementarydataTableS4.ZircondatafromNIGL.130

SupplementarydataTableS5.ZircondatafromAdelaidelaboratory.131

SupplementaryVideo1:“3loops.avi”fromthree-dimensionalcomputermodelderivedfromsampleillustrated132inFigure11.133

SupplementaryVideo2:“core.avi”fromthree-dimensionalcomputermodelderivedfromsampleillustratedin134Figure11.135

Figure1

Figure2

Figure3

Figure4

Figure5

Figure6

Figure7

Figure8

Figure9

Figure10

Figure11

Figure12

Figure13

Figure14

Figure15

Figure16

Figure17

Figure18

Figure19