The Ca and Mg isotope record of the Cryogenian Trezona carbon isotope excursion

Citation for this paper:

Ahm, A. C., Bjerrum, C. J., Hoffman, P. F., Macdonald, F. A., Maloof, A. C., Rose, C. V., … Higgins, J.

A. (2021). The Ca and Mg isotope record of the Cryogenian Trezona carbon isotope excursion. Earth

and Planetary Science Letters, 568, 1-13. https://doi.org/10.1016/j.epsl.2021.117002.

UVicSPACE: Research & Learning Repository

_____________________________________________________________

Faculty of Science

Faculty Publications

_____________________________________________________________

The Ca and Mg isotope record of the Cryogenian Trezona carbon isotope excursion

Anne-Sofie C. Ahm, Christian J. Bjerrum, Paul F. Hoffman, Francis A. Macdonald,

Adam C. Maloof, Catherine V. Rose, … & John A. Higgins

August 2021

© 2021 Anne-Sofie C. Ahm et al. This is an open access article distributed under the terms of

the Creative Commons Attribution License.

https://creativecommons.org/licenses/by/4.0/

This article was originally published at:

Contents lists available atScienceDirect

Earth

and

Planetary

Science

Letters

www.elsevier.com/locate/epsl

The

Ca

and

Mg

isotope

record

of

the

Cryogenian

Trezona

carbon

isotope

excursion

Anne-Sofie

C. Ahm

a,c,

∗

,

Christian

J. Bjerrum

b

,

Paul

F. Hoffman

c

,

Francis

A. Macdonald

d

,

Adam

C. Maloof

a

,

Catherine

V. Rose

e

,

Justin

V. Strauss

f

,

John

A. Higgins

a

a_{PrincetonUniversity,GuyotHall,Princeton,NJ08540,USA}

b_{UniversityofCopenhagen,ØsterVoldgade10,1350CopenhagenK,Denmark} c_{UniversityofVictoria,SchoolofEarthandOceanSciences,BC,Canada}

d_{UniversityofCalifornia,SantaBarbara,DepartmentofEarthScience,CA93106,USA} e_{UniversityofStAndrews,IrvineBuilding,St.Andrews,UnitedKingdom}

f_{DartmouthCollege,FairchildHall,Hanover,NH03755,USA}

a

r

t

i

c

l

e

i

n

f

o

a

b

s

t

r

a

c

t

Articlehistory:

Received28September2020 Receivedinrevisedform4May2021 Accepted10May2021 Availableonlinexxxx Editor:L.Derry Keywords: Trezonaexcursion SnowballEarth CaandMgisotopes diagenesis carboncycle

The Trezonacarbon isotope excursionis recorded onfivedifferent continents inplatform carbonates depositedpriortotheend-CryogenianMarinoanglaciation(>635Ma)andrepresentsachangeincarbon isotopevaluesof16–18‰. Basedonthespatialand temporalreproducibility,theexcursionpreviously hasbeeninterpreted astracking thecarbonisotopiccompositionofdissolved inorganiccarboninthe globaloceanbeforethedescentintoasnowballEarth.However,inmodernrestrictedshallowmarineand freshwatersettings,carbonisotopevalueshaveasimilarlylargerange,whichismostlyindependentfrom openoceanchemistryandinsteadreflectslocalprocesses.Inthisstudy,wecombinecalcium,magnesium, andstrontiumisotopegeochemistrywithanumericalmodelofcarbonatediagenesistodisentanglethe degreetowhichtheTrezonaexcursionreflectschangesinglobalseawaterchemistryversuslocal shallow-waterplatform environments.Ouranalysisdemonstratesthatthe mostextremecarbonisotopevalues (∼-10‰ versus +10‰)are preserved in formerplatform aragonite that was neomorphosed tocalcite duringsediment-bufferedconditions and recordthe primarycarbon isotopecompositionof platform-top surfacewaters.In contrast,the downturn and recoveryof the Trezonaexcursion are recordedin carbonatesthatwerealteredduringearlyfluid-buffereddiagenesisandcommonlyaredolomitized.We alsofindthatthenadiroftheTrezonaexcursionisassociatedwithafractionalincreaseinsiliciclastic sediments,whereastherecoveryfromtheexcursioncorrelateswitharelativeincreaseincarbonate.This relationshipsuggeststhattheextremenegativeisotopicshift inplatformaragoniteoccurredinconcert withperiodsofincreasedinputofsiliciclasticsediments,changesinwaterdepth,andpossiblynutrients toplatformenvironments.Althoughtheprocessforgeneratingextremelynegativecarbonisotopevalues in Neoproterozoic platform carbonates remains enigmatic, we speculate that theseexcursions reflect kineticisotopeeffectsassociatedwithCO2invasioninplatformwatersduringperiodsofintenseprimary

productivity.

©2021TheAuthor(s).PublishedbyElsevierB.V.ThisisanopenaccessarticleundertheCCBYlicense (http://creativecommons.org/licenses/by/4.0/).

1. Introduction

The Cryogenian Period (

∼

720–635 Ma)of the Neoproterozoic Era is characterized by major reorganizations of Earth surface processes and is bracketed by two Snowball Earth events, the olderSturtianandtheyoungerMarinoan glaciation(Hoffmanand Schrag, 2002). However, many important aspects of these

evolu-*

Correspondingauthorat:PrincetonUniversity,GuyotHall,Princeton,NJ08540, USA.

E-mailaddress:aahm@princeton.edu(A.-S.C. Ahm).

tionaryandclimaticchangesremainenigmaticduetothesparsity of radiometric age constraints andthe challenges of interpreting geochemicalrecordsfromancientplatformcarbonates.

Toovercome thelack of bothradiometric ageconstraints and biostratigraphy in Neoproterozoic successions,

δ

13C stratigraphy hasbeenusedasaglobalcorrelationtool. Cryogeniancarbonates arecharacterizedbysignificant variabilityintheisotopicratiosof

13_C/12_{C (}

_δ

13_{C), with excursions of similar shape and magnitude}

recordedinmultiplelocations(Halversonetal., 2005).The repro-ducibilityof

δ

13Cexcursionsacrosscontinentsisbroadlysupported by strontium isotopes (87Sr/86Sr) and radiometric ages (Fig. 1), https://doi.org/10.1016/j.epsl.2021.117002

A.-S.C. Ahm, C.J. Bjerrum, P.F. Hoffman et al. Earth and Planetary Science Letters 568 (2021) 117002

Fig. 1. GeneralizedCryogenianstratafromSouthAustralia,Namibia,andNorthwestCanada.Notechangesinverticalscale.Darkgraybarsindicatethestratigraphicintervals studiedhere.Allsuccessionsarecharacterizedbytwoglacialhorizons–thelowerSturtianandtheupperMarinoanglaciation–whichhavebeencorrelatedbyradiometric dates.However,carbonatesdepositedduringthenon-glacialinterludehavefewradiometricageconstraintsandhavebeencorrelatedbasedoncarbonandstrontiumisotope stratigraphy(published87_Sr/86_{SrratiosfromHalversonetal.,2005,2007).DatesfromSouthAustraliaarefrom:(1)Coxetal.(2018),(2)FanningandLink(2008),(3)Kendall}

etal.(2006),(4)Roseetal.(2013),(5)Preiss(2000),(6)Calveretal.(2013).DatesfromNamibiaarefrom:(7)Praveetal.(2016),(8)Hoffmannetal.(2004),(9)Halverson etal.(2005).DatesfromNorthwestCanadaarefrom:(10)Rooneyetal.(2014),(11)Macdonaldetal.(2010),(12)Baldwinetal.(2016),(13)Rooneyetal.(2015),(14) Macdonaldetal.(2018).

whichhasservedasevidencethattheseexcursionsdirectlyrecord perturbationsintheglobalcarboncycle(e.g.,Kaufmanetal.,1997; HoffmanandSchrag,2002).However,theinterpretationthat

δ

13C inNeoproterozoicplatformcarbonatesrecordstheisotopic compo-sitionofdissolved inorganiccarbon(DIC) ofopen-oceanseawater isatoddswithobservationsfrommoderntoMioceneplatformand periplatformcarbonates.Incontrasttopelagicsedimentsdeposited inthedeepsea,carbonatesthatoriginallyprecipitatedonshallow banks oftendonot reflectthe averagecarbonatesink(Swart and Eberli,2005;Swart,2008).

Over geologicaltime scales,the netinputofcarbonto the at-mosphere must be balanced by the burial oforganic matter and carbonatesediment,withbulkcarbonate

δ

13_{Cvaluescontrolledby}

the fractionof organicmatter burial globally ( forg,e.g., Kaufman

etal., 1997). However,

δ

13Cvalues ofplatform-derived carbonate aresensitivetorestrictionofmarinecirculation,diurnal productiv-ity(Patterson andWalter, 1994; Geymanand Maloof, 2019), and earlydiageneticalterationdrivenbytheadvectionofbothseawater andmeteoricwaterwithinthesedimentpile(AllanandMatthews,

1982; Melim etal., 2002). During time periods where carbonate burial is concentrated in platform environments, the impact of widespread localchanges of depositionalenvironments on global mass balance needs to be reevaluated. As a majorconstituent in carbonate(CaCO3),calciumisotoperatios(

δ

44/40Ca)providean

in-dependentconstraintontheaveragecarbonatesink.Themainsink forseawaterCa2+ _{istheburialofcarbonatesedimentsanda}

pre-dictionfortheglobalcalciumcycleisthat,onaverage,theisotope composition ofcarbonatethroughtime should equalthat ofBulk SilicateEarth(BSE=-1‰,Skulanetal.,1997;BlättlerandHiggins,

2017). In other words,what comes in must go out, andif sam-plingtheaveragecarbonatesink,

δ

44/40Cavaluesshouldapproach

∼

-1‰whenaveragedoverthickcarbonatesuccessions.Asaresult,

δ

44/40_{Cavaluesprovideatooltoreconcilethedecouplingbetween}

platform environmentsandtheglobalocean,the requirementsof globalmassbalance,andthelackofabsolutetimeconstraintsfrom individualNeoproterozoicsuccessions.

Inthisstudy,weinvestigatethevariabilityofcarbonate

δ

44/40_Ca

values across the Trezona excursion, a large negative

δ

13C ex-cursion recorded on multiple continents prior to the Marinoan glaciation.This

δ

13Cexcursionreachesvaluesof

∼ −

10‰and oc-cursstratigraphicallyaboveaprolongedinterval withmostly high

δ

13_{C values (upto +10‰, Kaufman etal., 1997; Halverson et al.,}

2005).Thetimingoftheexcursionisbroadlyconstrainedtowithin theCryogenian ‘non-glacial’ interlude (between660and 640Ma, Fig.1).Bycombiningmeasurementsof

δ

44/40CawithMgisotopes (

δ

26Mg), 87Sr/86Sr, andSr/Ca andMg/Ca ratios in bulk carbonate frommultiple sectionsin Australia,Namibia,andCanada,we test to what extent the Trezona

δ

13C excursion has been altered by diagenesis. We use the geochemical data to fingerprint samples that have preserved their primary

δ

13C values during sediment-buffereddiagenesisincontrastto

δ

13Cvaluesthathavebeenreset during fluid-buffereddiagenesis.Wefindthat intervals character-ized by fluid-buffered diagenesis have less extreme

δ

13C values, and

δ

44/40Ca closer to BSE, relative to intervals characterized by sediment-buffereddiagenesis. These resultsdemonstrate that the mostextreme

δ

13Cvalues,bothpositiveandnegative,areprimary inoriginandareassociatedwithlocalcontrolson

δ

13_Cin

shallow-wateraragoniteproducingenvironments.

2. Methods

2.1. Stratigraphicalsections

Inthisstudy,weinvestigatethegeochemicalvariabilityofthree different Cryogenian carbonate successions (12 stratigraphic sec-tions),from South Australia, Namibia, and Northwest Canada. At each locality,the sedimentology and

δ

13C chemostratigraphyhas beenextensivelystudiedinpreviouspublications(e.g.,McKirdyet al., 2001; Roseetal., 2012; Klaebe andKennedy, 2019; Hoffman,

2011; Macdonald et al., 2018), and we refer to the supplemen-tarymaterialforadetailedsummaryofthegeologicalsettingwith

Fig. 2. Numericalmodeldemonstratingthechangeinbulkcarbonatechemistryovertimewithincreasingdegreeofdiagenesis,inthiscasemodeledasdolomitization(from 0-100%),inbothfluid- (box1,bluedashedline)andsediment-buffered(boxn,redsolidline)conditions(followingmethodsoutlinedbyAhmetal.,2018). A. Thechange incarbonateδ13Cvaluesovertimefromaprimaryvalueof10‰(bluestar)towardsafluidvalueof-4‰(redstar),accountingforafractionationfactorof+1‰between pore-fluidandthediageneticcarbonateminerals(TableS1). B. Thechangeincarbonateδ44/40Cavaluesfromaprimaryvalueof-1.7‰towardsafluidvalueof-0.65‰,with afractionationfactorof0‰(FantleandDePaolo,2007). C. Thechangeincarbonateδ26_{Mgvaluesfromaprimaryvalueof-2.4‰towardsafluidvalueof-0.35‰,witha}

fractionationfactorof-2‰(HigginsandSchrag,2010).Astherateofdolomitizationdecreasewithtime,theinflectioninthesediment-bufferedδ26_{Mgpathwayrepresents}

thepointwherethesupplyofnewMg2+fromadvectionovercomesthedecreaseofMg2+ duetodolomitization. D. Cross-plotofδ13_Cversusδ44/40_{Causingthemodel}

outputsfromAandB. E. Cross-plotofδ26_Mgversusδ44/40_{CausingthemodeloutputsfromBandC.NotethatthemodelphasespaceinDandEisinsensitivetochanges}

inbothreactionandadvectionrate,aschangesineitheroftheseparameterswould haveasimilarimpactonboththeyandx-axis.Themodelisfittothedatausingthe phasespaceinDandE,wherethegoodnessoffitisevaluatedinpartbytheconsistencyofthemodeltopredictasimilardegreeofalteration(%)forthesamesample.A goodmodelfit(pinksquare)isindicatedbyasimilardegreeofalterationacrossthephasespaces,whereasabadfit(bluecircle)isindicatedbyanoffsetinthepredicted degreeofalterationacrossthephasespacesandbeingoutsidethephasespaceinD.

location maps(Fig.S1–S3).Foreach locality,thestratigraphyand radiometricageconstraintsaresummarizedinFig.1.

2.2. Geochemicalanalyses

All measurements presented in this study are performed on carbonate powdersthat previously havebeen measured for

δ

13C and

δ

18Ovalues(Straussetal., unpublished;Hoffman,2011;Rose etal., 2012; Macdonaldetal.,2018).We refertothe supplemen-tary materialfor adetailedoutline oftheCa, Mg, andSr isotope analysesandmajorandtraceelementanalyses.

Calciumisotopemeasurements arereportedforall samplesas therelativeabundanceof44Carelativeto40Causingstandarddelta notation,normalizedtotheisotopiccompositionofmodern seawa-ter.ForCaisotopes,the externalreproducibilityforSRM915band SRM915arelativetomodernseawateris

−

1

.

19

±

0

.

14‰(2

σ,

N

=

120)and

−

1

.

86

±

0

.

16‰(2

σ,

N

=

24),respectively.

Magnesium isotope ratios are reported as the relative abun-dance of 26Mg versus 24Mg, normalized to DSM-3. For Mg, the long-term external reproducibility forCambridge-1 and seawater is

−

2

.

61

±

0

.

10‰(2

σ

,

N

=

81) and

−

0

.

83

±

0

.

11‰(2

σ

,

N

=

47), respectively.

Asubset ofsampleswas selected forstrontiumisotope analy-ses,performedatbothPrincetonUniversity(samplesfromNamibia andAustralia)andWHOI(samplesfromNorthwestCanada). Stron-tium isotope measurements are reported as the ratio of 87Sr over86_{Sr.AtPrincetonUniversity,thelong-termreproducibilityof}

NBS987is 0

.

710280

±

0

.

000006 (N

=

4).AtWHOI, thelong-term

reproducibility of NBS987 is 0

.

710253

±

0

.

000015 (N

=

12). To reduce theinfluenceofin-situ Rb decay(87_Rbto 87_Sr),

measure-mentswerefilteredforSr/Caratios>1.5mmol/molandreportedas theaveragefiltered valueforeach section, consistentwith meth-odsfromprevious publications(Halversonetal.,2005).Insection F1228,whereSr/Caratiosarelow,87Sr/86Srratiosinsteadwere fil-teredbasedontheleastradiogenicvalues.

2.3.Numericaldiageneticmodel

Toanalyzethegeochemicaldata,weuseanumericalmodelof earlycarbonatediagenesis.Thismodelpreviouslyhasbeenusedto simulatediageneticchangesinbothBahamianandNeoproterozoic carbonate(e.g.,Ahmetal.,2018,2019).

Themodelcomputesthegeochemicalchanges thatoccur dur-ing early diagenesis asmetastable carbonate minerals(e.g., arag-onite)are dissolved andmore stablephases (e.g.,dolomite) pre-cipitate. The conceptual model framework is a simplification of the complex geometry of fluid flow anddiagenesis in carbonate sediments,whichisaffected bylocal differencesinporosity, per-meability,andreactionrates.Toaccountforthesedifferences, we presentthemodelresults as2-dimensional phase-spaces(Fig. 2). These model cross-plots represent the geochemical changes that occuratdifferentstagesofdiageneticalteration(from0-100%)and fluidevolutionalongtheflowpath(fluid- tosediment-buffered).In otherwords,eachcross-plotcomprisesthetotalgeochemical vari-abilitythat can beproduced fromreactions betweena carbonate rockandadiageneticfluidwithaprescribedcomposition.

A.-S.C. Ahm, C.J. Bjerrum, P.F. Hoffman et al. Earth and Planetary Science Letters 568 (2021) 117002

Fig. 3. SouthAustralia: TheTrezonaFormationiscorrelatedacrossthebasinbasedoncarbonisotopesandthestratigraphicpositionoftheEnoramaShalebelowandthe glacigenicElatinaFormationabove(Roseetal.,2012,2013).SectionsC215,N424,andA1043arefromtheCentralFlindersRanges,whilesectionR2isfromtheNorthern FlindersRanges(forlocationmapseeFig.S1).

Weusethemodelresultstoevaluate towhatdegreethe geo-chemical signalsacross theTrezonaexcursionare products of di-agenesis. By fitting the model phase space to envelop the data, we can estimate the composition of diagenetic fluidsand differ-ent primary sedimentary end-members (Fig. 2D-E). We refer to thesupplementarymaterialforadetaileddescriptionofthemodel setup and evaluation ofspecific model fits using a bootstrap re-samplingtechnique.

3. Results

3.1. SouthAustralia

In South Australia, the upper part of the Cryogenian succes-sion contains the Trezona excursion (McKirdy et al., 2001; Rose et al., 2012). Stratigraphically below the excursion, the shallow-water limestone of the Etina Formation is characterized by high

δ

13Cvalues

∼

+10‰,relativelyconstant

δ

44/40_{Cavaluesranging}

be-tween-1.6and-1.2‰,andSr/Caratiosthataverage0.86mmol/mol (rangingbetween0.12–2.2mmol/mol,Fig.3A-B).ThemostSr-rich sampleshave87Sr/86Srratiosbetween

∼

0.7075–0.708.

ThedownturnoftheTrezonaexcursionisnotobserved,dueto alackofcarbonateintheEnoramaShale(Fig.1).Thenadirofthe excursionisrecordedinthelower partoftheTrezonaFormation, characterizedbyahighfractionoffine-grainedsiliciclasticmaterial relative tocarbonate, with

δ

13_{C valuesat}

_∼

_{-10‰ (McKirdyetal.,}

2001; Roseetal., 2012). Generally, thisinterval also recordslow

δ

44/40Ca valuesdownto

∼

-2‰, low

δ

18Ovalues(downto-15‰), andhighSr/Caratios averaging

∼

1.7mmol/mol.The mostSr-rich sampleshave87Sr/86Srratiosof

∼

0.7073.

The recovery ofthe excursion spans

∼

250 m, witha gradual return tohigher

δ

13_{C values approaching}

_∼

_{0‰ atthe topof the}

succession. Thisintervalalso recordsbroadlyincreasing

δ

18O val-ues (towards -8‰), increasing

δ

44/40Ca values up to -1.2‰, and decreasing Sr/Ca ratios (Fig. 3). While the depositional

environ-mentinthenadiroftheexcursionis characterizedbysiliciclastic sediments,the increase in

δ

13C valuesup section coincides with abroad increase inrelative carbonateabundance(McKirdy etal.,

2001;Roseetal.,2012).Thiscoarseningupwardssuccession previ-ouslyhasbeeninterpretedasrepresentingashallowing(McKirdy et al., 2001; Rose et al., 2012), although this interpretation has beenchallenged(KlaebeandKennedy,2019).

3.2.Namibia

Namibia is the only location that has a progressive record of theTrezonaexcursiondownturn(Fig.4).TheTrezonaexcursionis foundintheupperOmbaatjieFormationandrecordsasteady de-cline in

δ

13C values downfrom

∼

+7 to-7‰ over astratigraphic intervalof40–80m.Thisintervalalsorecordsdecreasing

δ

44/40_Ca

values(from-0.7to -1.6‰), decreasing

δ

18Ovalues(from -1to -8‰),andincreasing

δ

26Mgvalues (from-2.1to -1.0‰). However, thechangesin

δ

44/40_Ca,

_δ

18_O,and

_δ

26_{Mgvaluesbegin}

stratigraph-ically after the downturn of the

δ

13C excursion has reached 0‰ (Fig.4B-C).

Whilethedownturnoftheexcursionishostedinthe shallow-waterdolostone ofthe Ombaatjie platform,the nadir of the ex-cursioncoincideswithachangefromcarbonateto aunique fine-grained siliciclastic unit (parasequence b8, Hoffman, 2011). The Ombaatjie platform has a well-defined southern limit, beyond whichisaforeslopewithredepositedcarbonates(Hoffman,2011). On the foreslope, the siliciclastic siltstone unit (the Narachaams Member)isthicker(>100m)thanontheplatform,andthe down-turnoftheTrezonaexcursionisnotrecorded.Abovethesiltstone, carbonatedepositionreturnsintheformofa100-m-thick coarsen-ingupwardsuccession(Franni-ausMember)thatrecordsthenadir andasomewhatmorecompleterecoveryfromtheTrezona excur-sionthan ispreservedon the platform(from -7towards 0‰).In parallel,thisintervalalsorecordsincreasing

δ

18Ovalues(from-13 to-2‰),increasing

δ

44/40_{Cavalues(from}

_∼

_{-1.6to-0.8‰),}

decreas-Fig. 4. Namibia: SectionP4006,P7500,andP6503areshallow-watercarbonatesfromtheOmbaatjieFormationthatrecordthedownturnoftheTrezonaexcursion.Section P9500isfromthemoredistalFranni-ausMemberandonlyrecordstherecoveryoftheTrezonaexcursion.ThenadiroftheTrezonaexcursioncoincides withdepositionof aregionalsiltstoneunitwithvaryingthickness,termedtheNarachaamsMemberontheforeslope.Theplatformsectionsarecorrelated(dashedlines)basedonobserved parasequences(previouslytermedb4-b8),whiletheforeslopesection(P9500)isrelatedtotheplatformusingδ13_{CchemostratigraphyHoffmanforlocationmapseeFigS2,}

2011).

ing Sr/Caratios,and

δ

26_{Mgvaluesbetween-2and-1.5‰(Fig. 4).}

TheSr-richlimestonepreserving87Sr/86Srratiosof

∼

0.7074.

3.3. NorthwestCanada

In the Mackenzie Mountains (sections J1132/33), the Trezona excursion is recorded in the upper Keele Formation. Stratigraph-ically below the excursion, the shallow-water carbonate of the lower KeeleFormation ischaracterized by aplateauof high

δ

13C values(the‘Keelepeak’,

∼

+10‰,Kaufman etal., 1997; Dayetal.,

2004). The limestone and dolostone found in this interval have

δ

44/40_{Ca values between -1.2 and -1.0‰, high}

_δ

26_{Mg values up}

to

∼

-0.6‰, andSr/Caratios averaging0.7mmol/mol(Fig. 5). The moreSr-richsampleshave87Sr/86Srvaluesof

∼

0.7073.

The carbonates containing the ‘Keele peak’ are succeeded by siliciclastic dominated strata (the ‘Keele clastic wedge’, Aitken,

1991;Dayetal.,2004).OverlyingtheKeeleclasticwedge,the Tre-zona nadir is found within subtidal limestone andcontains

δ

13C valuesdownto-10‰,

δ

44/40Cavaluesdownto-1.9‰,Sr/Cavalues upto4mmol/mol,and87_Sr/86_Srratiosof

_∼

_{0.7074(Fig.5).}

Follow-ing the nadir of the excursion,the uppermost stata of the Keele Formation record increasing

δ

13_{C valuesbefore the deposition of}

glacialsediments.Acrossthisinterval,

δ

44/40_{Cavaluesincrease}

to-wards-1‰andSr/Caratiosdecreaseto<1mmol/mol.

IntheWerneckeMountains,theTrezonaexcursionisrecorded in the Durkan Memberof the Ice Brook Formation. Stratigraphi-cally below the Durkan Member, is the Mount Profeit dolostone recording

δ

13C values between

∼

0 to +5‰ (Macdonald et al.,

2018). Thisinterval has

δ

44/40Ca valuesbetween

∼

-1.5and-1‰,

δ

18Ovalues between

∼

-11 to-8‰, andSr/Caratios between0.6 and1.5mmol/mol.Sampleswithlow

δ

44/40Ca andhighSr/Ca ra-tiosrecord87_Sr/86_Srratiosof

_∼

_{0.7073–0.7076.}

TheDurkanMemberconsistsofsiltstoneandshalewith thin-beddedlimestoneandmultiplehorizonsofcarbonateclastbreccia (sectionF1228,Macdonaldetal.,2018). Thisinterval containsthe nadirandrecoveryoftheTrezonaexcursion,recordingvaluesthat increasefrom-10‰ towards 0‰across

∼

250m.The recoveryof theexcursion coincides with an increase in carbonatedeposition relative tosiliciclastics andan increase in

δ

44/40Ca valuesfrom -1.8to-1.0‰. Thisinterval alsorecordshighSr/Caratiosupto2.5 mmol/mol,andrelativelyscattered

δ

18Ovalues(from-6to-12‰, Fig.5).

4. Discussion

4.1. TheTrezonaexcursion–globalandlocalcontrols

Despite the fact that the Trezona downturn is followed by sometimeshundredsofmetersof

δ

13Crecoverytowards0‰, pre-viousmodels havefocused oninterpreting thedeclinein

δ

13C as associated with global changes in sea water chemistry and the onset of Snowball Earth (e.g., Hoffman and Schrag, 2002). Here, however,we highlightthree features ofthe relationship between carbon andcalcium isotopes across the entire Trezonaexcursion thatareinconsistentwiththisinterpretation.

First,we considertowhat degreethe

δ

44/40_{Cavaluesthat are}

observed across the Trezona excursion may reflect steady state changes in the global calcium cycle. The main sink for seawater Ca2+_{istheburialofcarbonate.Apredictionfortheglobalcalcium}

cycleisthat, onaverage,the calciumisotope compositionof car-bonatesedimentsthroughtimeshouldequalthatofBSE(

∼ −

1‰, on time-scales >106 years,Skulan etal., 1997; Blättler and Hig-gins,2017).However,themajorityoftheCryogeniansampleshave

δ

44/40_{Ca <-1‰(Fig. 6), both before theTrezona excursionandin}

A.-S.C. Ahm, C.J. Bjerrum, P.F. Hoffman et al. Earth and Planetary Science Letters 568 (2021) 117002

Fig. 5. NorthwestCanada: StratigraphiccolumnsandisotopicmeasurementsfromNorthwestCanada.SectionF1228isfromtheWerneckeMountainsinYukonwhilesections J1132andJ1133arefromtheMackenzieMountainsintheNorthwestTerritories(forlocationmapsseeFig.S3).

Fig. 6. Cross-plotsofgeochemicaldatafromSouthAustralia,Namibia,andNorthwestCanada,incomparisontpplatformandperiplatformcarbonatesfromtheBahamasand authigenicdolostonefromtheNeogeneMontereyFormation(graysymbols,Blättleretal.,2015;Higginsetal.,2018;Ahmetal.,2018). A ThecorrelationbetweenhighSr/Ca ratiosandlowδ44/40_{Cavaluessuggeststhat theTrezona-bearingstrataoriginatedasaragonitethatwereneomorphosedduringsediment-bufferedconditions,preserving}

muchoftheoriginalgeochemistry,includingδ13_{Cvalues.Incontrast,highδ}44/40_{CavaluesandlowSr/Caratiosindicateneomorphismordolomitizationunderfluid-buffered}

conditionswherethecarbonategeochemistrywasresettowardthevalueofthediageneticfluid. B Thecorrelationbetweenhighδ26_Mgandlowδ44/40_{Cavaluesindicates}

sediment-buffereddolomitizationwhereaslowδ26_Mgandhighδ44/40_{Cavaluesindicatefluid-buffereddolomitization. C Thecorrelationbetweenδ}13_Cvaluesandδ44/40_Ca

showsalinkbetweenfluid-buffereddiagenesisandlessextremeδ13Cvaluesversussediment-buffereddiagenesisofplatformaragoniteandextremeδ13Cvalues. D The correlationbetweenδ18Oandδ44/40Cavaluesisconsistentwithamixingtrendbetweentwoend-members:(1)samplesthathaverecrystallizedduringearlydiagenesis wherebothδ18_Oandδ44/40_{Cavaluesareresettowardsseawater,and(2)samplesthathaverecrystallizedduringlate-stagediagenesisathigherburialtemperatureswhere} δ44/40_{Cavaluesarepreserved(sediment-buffered)andδ}18_{Ovaluesarereset.}

theTrezonanadir,andmassbalancerequiresthatthesesediments cannotrepresenttheaveragecarbonatesinkoverlongtimescales. Second, we consider to what degree the synchronicity of the carbonandcalciumisotopeexcursionsareconsistent witha tran-sientchangeinglobalcarbonateburial.OnthesurfaceofEarth,the residence time forcalcium isan order ofmagnitude longerthan forcarbon (

∼

106 _versus

_∼

₁₀5 _{years,e.g., Gussoneetal., 2020).}

Therefore,anyperturbationtocarbonateburialshouldnotresultin asynchronousstratigraphicchangeincarbonate

δ

13Cand

δ

44/40Ca values,ifweassumethattherelativeresidencetimeofthecalcium andcarboncycleswere scaledsimilarlytotoday,(Holmden etal.,

2012b). Ifthe Trezonaexcursion represents a transient perturba-tion,the

δ

44/40Caexcursionshouldtakelongertorecoverthanthe

δ

13_{Cexcursion.However,theTrezonadatarecordscovariationwith}

nolagbetweencarbonandcalciumisotopes(Fig.3–5).

Third, we consider to what degree the magnitudeof the Tre-zona

δ

44/40Cadownturnandrecoveryisconsistentwithatransient perturbation in global seawater (<1 Myrs). There are three main mechanismsthat cancausea negative calciumisotopeexcursion: (1) increasedweathering,(2) ocean acidification, or(3) a change fromcalcitetoaragoniteseas(e.g.,Gussoneetal.,2020). Numeri-calmodelsthatincludethecoupledcalcium-carboncycle,andthe inherentlinkstocarbonatesaturationandprecipitationrates,have demonstrated thatthe maximum

δ

44/40_{Ca excursion that can}

oc-curfromthecombinedeffectsofincreasedweatheringandocean acidificationis

∼ −

0

.

3‰ (KomarandZeebe, 2016). Inaddition,a switch inthedominantmineralogy oftheaveragecarbonatesink fromcalcitetoaragonitecouldcauseatransientnegativeexcursion incarbonate

δ

44/40Cavalues,asaragoniteis-0.5‰ moredepleted in 44_{Ca than calcite (Gussone etal., 2005).In thisscenario, once}

a new steadystate is reached (<1myrs after),seawater

δ

44/40_Ca

will be 0.5‰ higher and

δ

44/40Ca values of the average carbon-atesink,nowaragonite,willagainreflectBSE(e.g.,Gussoneetal.,

2020).Adding thecombined effectsfrom(1)–(3), theupperlimit for changing

δ

44/40Ca values as a result of transient global per-turbation is -0.8‰, which would imply that the duration of the excursion was <106 _{years,andincludedtheunlikely combination}

ofincreasedweathering,oceanacidification, anda changefroma calcite to aragonitesea.For comparison, therecoveryof the Tre-zona excursion records a change in

δ

44/40_{Ca values of}

_∼

₀

_.

_85‰

on all three continents across

∼

250m of stratigraphy,while the downturnrecordsachangeof

∼

-1‰inNamibiaacross40–80m.

The three features discussed above indicate that the

δ

44/40Ca excursionisunlikely toreflectaglobalCa-cycleperturbation.The implication forthe carbonisotope record isthat the Trezona ex-cursionmaypredominantlyreflectchanges inlocalplatform con-ditionsanddiagenesis,witharagoniteburialinsome depositional settings, while other possibly larger carbonate sinks were cal-cite or dolomite. Alternatively, the processes driving changes in

δ

44/40_Caand

_δ

13_{Cvalueswouldhavetobedecoupled,forexample,}

with

δ

44/40Cavaluesrecordingchangesindiagenesisand mineral-ogy while

δ

13C isotopes are recordingglobal seawaterchemistry. This latterscenario cannot be ruled out, butcarbonate

δ

13_{C and}

δ

44/40_{Ca aremodified atbroadlysimilar fluid-to-rockratios(Ahm}

etal., 2018), anditisunlikely that carbonate

δ

44/40Ca valuesare alteredwhile

δ

13C valuesare preserved. Belowwe usea numer-ical modelofdiagenesis toevaluate towhatdegree thesesignals may instead be influenced by changes in mineralogy, diagenesis (fluid- versussedimentbuffereddiagenesis), andthelocal deposi-tionalenvironment.

4.2. Effectsofmineralogy,diagenesis,anddolomitization

Inmodern platform settings,carbonate

δ

13Cand

δ

44/40Ca val-uesdo nottrackglobalseawaterchemistry (Swart,2008;Higgins etal.,2018).Carbonisotopevaluesspanfrom

∼

-11to+8‰

(Gey-manandMaloof,2021),and

δ

44/40_{Cavaluesvarybetween-1.6and}

-0.2‰ (Higginset al., 2018). Today we accept that thesemodern platform carbonatesdo not representthe average carbonatesink andthat thegeochemicalchanges largely reflectlocal changesin mineralogy and diagenesis. To what degree are Cryogenian plat-formcarbonatesaffectedbysimilarcommonlocalprocesses?

4.2.1. Cryogenianplatformaragonite

The carbonates hosting the Trezona excursion record covaria-tion between Sr/Ca and

δ

44/40Ca values that is consistent with alteration of platform aragonite across a continuum of fluid- to sediment-buffered conditions(Fig. 6A). Modernplatform settings showasimilar correlationthatreflectsmixingbetweenthree dif-ferent carbonate end-members (Higgins et al., 2018): (1) plat-form aragonite with high Sr/Ca ratios (

∼

10-12 mmol/mol) and low

δ

44/40Cavalues(

∼

-1.5‰),(2)fluid-bufferedneomorphosedor dolomitized carbonate with low Sr/Ca ratios and high

δ

44/40Ca that approach modern seawater (0‰), and (3)sediment-buffered neomorphosed calcite (former aragonite), that has retained low

δ

44/40Ca values (

∼

-1.5‰) and relatively high Sr/Ca values (

∼

2-4 mmol/mol).

Calciumisotopes and Sr/Ca ratios are fingerprints of different diageneticend-membersbecausetheCaisotopefractionation fac-torandSr partitioningare sensitive toboth mineralogy and pre-cipitation rate(Tang et al., 2008; Gussoneet al., 2005). Primary aragoniteismoredepletedin44_{CaandenrichedinSr(-1.5‰and}

10mmol/mol)relativetoprimarycalcite(-1‰and

∼

1mmol/mol). DiageneticcalciteordolomiteischaracterizedbylowerSrcontents (<1 mmol/mol) and less fractionated

δ

44/40Ca values, approach-ing

∼

0‰atequilibriumwiththepore-fluids(FantleandDePaolo,

2007).

InthecarbonatesthatrecordtheTrezonaexcursion,noprimary aragonite is preserved, and instead we interpret the correlation between Sr/Ca ratios and

δ

44/40_{Ca as reflecting mixing between}

fluid- and sediment-buffered end-members. The combination of high Sr/Ca ratios and low

δ

44/40Ca values excludes the possibil-ity of diagenetic alteration from either marine (would increase

δ

44/40Ca towards seawater values,Higgins et al., 2018) or mete-oric fluids (would lower Sr/C ratios, Allan and Matthews, 1982), andinsteadindicatessediment-buffereddiagenesisofformer arag-onite (Fig. 6A). The interpretation of precursor aragonite also is supported by petrographic observations ofaragonitic ooid fabrics intheTrezonaFormationinSouthAustralia(Singh,1987).

Acrosstheexcursion,twoseparatestratigraphicintervalsrecord sediment-buffered neomorphism of former aragonite. First, pre-excursion carbonatewithhigh

δ

13C of up to +10‰ (Keele peak), and second, in the nadir of the excursion (Fig. 6C). It is likely that these extreme

δ

13C values of the Keele peak and Trezona nadirrecord thechemistry ofthe environment where the arago-nitesedimentsoriginallyprecipitated.Incontrast,theintermediate stratathatrecordthedownturnandrecoveryoftheTrezona excur-sionlikelywerealteredduringearlyfluid-buffereddiagenesis,and recordthechemistryoftheearlydiageneticpore-waters(Fig.6C).

Inadditiontoevidenceforearlydiagenesis,thereis geochemi-calevidenceforlate-stagealteration.While

δ

44/40Caand

δ

13C val-uesarealteredatbroadlysimilar fluid-to-rockratios,

δ

18Ovalues remainsensitivetodiagenesis insettingswherebothcalciumand carbonare sediment-buffered(Ahmet al.,2018). Inother words, itispossibletohavecarbonatethathasretainedprimary

δ

44/40Ca and

δ

13C values, while

δ

18Ovalues have beenreset during late-stagediagenesis.Forexample,aragonitethatwasdepositedinthe Trezonanadirwasnotsignificantlyalteredduringearlymarineor meteoricdiagenesis, andtherefore preservedlow

δ

44/40_{Ca values}

andhighSr/Ca ratios.This aragonite, however, eventually recrys-tallized to low-Mg calcite during burial. At this stage, the pore-fluidswere sediment-buffered withrespect to calcium, but

δ

18_O

A.-S.C. Ahm, C.J. Bjerrum, P.F. Hoffman et al. Earth and Planetary Science Letters 568 (2021) 117002

Table 1

Comparisonofdiageneticmodelresults.Modelfitsaregivenasbestfitwithminimumandmaximumuncertaintybounds. Downturn fluid Uncertainty (σ) Recovery fluid Uncertainty (σ) Modern Seawater

δ26_{Mg (‰) -0.35 [-0.4, -0.1] 0.10 [-0.05, 0.25] -0.82} δ44/40_{Ca (‰) -0.65 [-0.68, -0.55] -0.50 [-0.65, -0.40] 0.00} δ13C (‰) -4.0 [-6.5, -2.7] 7.0 [4.0, 12.5] 0.0–1.5 DIC/Ca2+(mol/mol) 0.05 [0.04, 0.19] 0.4 [0.2, 0.8] 0.2 Mg2+/Ca2+(mol/mol) 0.6 [0.4, 1.1] 1.1 [1.0, 2.0] 5.1

values were reset to lower and more variable values(down to -15‰, Fig.6D)duetotheincrease inburialtemperature(

∼

100°C, if assuming fluid values of 0‰ V-SMOW, Kim and O’Neil, 1997). In contrast,carbonatethat was dolomitized andstabilizedduring earlymarinediagenesiswasmoreresistanttolate-stagealteration and

δ

18O valuesare less altered(values up to -1‰, Fig. 6D). Al-though it is possibleto alter

δ

13C valuesduring late-stageburial (Derry,2010),thecorrelationbetweenverydepleted

δ

44/40Caand

δ

13Cvaluesindicatesthat thedeepburialpore-fluids werehighly bufferedbycarbonatedissolution,andsediment-bufferedwith re-specttobothcalciumandcarbon.

4.2.2. Dolomitizationandfluid-buffereddiagenesis

Intervals of early fluid-buffered diagenesis are observed dur-ing the downturn and recovery of the Trezona excursion, and in Namibia these intervalsare associated with dolomitization. In addition to high

δ

44/40_{Ca values and low Sr/Ca ratios, the}

fluid-buffereddolostoneischaracterizedbyrelativelylow

δ

26Mgvalues. Dolostonethatformedduringearlydiagenesisunderfluid-buffered conditionstendtohavelowerandlessvariable

δ

26_Mgvaluesthan

dolostoneforminginsediment-bufferedconditions(Blättleretal.,

2015). Thesetrendsare theproduct ofa

∼

-2‰isotopic fraction-ationassociatedwiththeincorporationofMgintodolomite (Hig-ginsandSchrag,2010).Forexample,fluid-buffereddolostonefrom the Bahamas platform andNeogeneMonterey Formation are off-setfrommodernseawaterby-2‰(valuesof-2.8‰,Blättleretal.,

2015; Higgins et al., 2018). In contrast, sediment-buffered dolo-stoneis enrichedin 26_{MgduetoRayleigh-type distillation ofthe}

pore-fluidinmoreclosedsystemsettings(Blättleretal.,2015). Across the Trezonaexcursion in Namibia, two separate strati-graphic intervals record evidence of fluid-buffered diagenesis (Fig.8).First,dolostonerecordingthedownturnoftheTrezona ex-cursionhavehigh

δ

44/40Cavaluesandlow

δ

26Mgvalues(Fig.6B). Second, high

δ

44/40Ca valuesandlow

δ

26Mg valuesare recorded in dolostone during the recovery of the Trezona excursion prior to deposition of Marinoan glacial deposits. The relationship be-tween

δ

26_{Mg and}

_δ

44/40_{Ca across these two intervals is similar}

to Bahamian dolostone, but offset towards lower

δ

44/40Ca val-ues(between -1.4and-0.6‰)andhigher

δ

26Mgvalues(between -2.5 and -0.5‰), suggesting that the dolomitizing fluid was en-richedin 26Mganddepletedin44Carelative tomodern seawater (Fig.6B).Furthermore,thesefluid-bufferedintervalsrecordless ex-treme

δ

13Cvalues(betweenapproximately-5and+5‰,Fig.6C)in comparisontotheintervalscharacterizedbysediment-buffered di-agenesis(-10and+10‰).

We useanumericalmodelof earlydiagenesis toestimate the composition and origin of the diagenetic fluids that dolomitized the OmbaatjieFormation,recordingthe downturnof theTrezona excursion (Fig. 7). Model results indicate that primary platform aragonitewithhigh

δ

13Cvalues(estimatedat+10‰)were dolomi-tizedby afluid with

δ

13Cvaluesof-4‰(uncertaintyfrom-6.5to -2.7‰).Inaddition,themodelestimatesfluid

δ

44/40_{Ca valuesof}

-0.65‰and

δ

26Mgvaluesof-0.35‰(uncertaintyof-0.68to-0.55‰ and-0.4to-0.1‰,respectively,Table1).

Platform fluids can be modified from their original seawater compositions due to reactions in the subsurface and subsequent mixing withfreshwater. Forexample, the low

δ

44/40_{Ca values of}

the downturn fluid (Fig. 7F) are consistent with modern obser-vations from restricted platform settings influenced by subma-rine groundwater discharge and subsurface carbonatedissolution (

∼

-1.0 to -0.4‰, Holmden et al., 2012a; Shao etal., 2018). Sim-ilarly, subsurface reactions can modify fluid

δ

26Mg values and Mg2+_/Ca2+ _{ratios. Generally, submarine groundwater discharge} andcarbonatedissolutionleadtolowerMgisotopevalues (Jacob-sonetal., 2010;Shirokova etal.,2013),while high

δ

26Mgvalues are associated with lagoonal and hypersaline environments that aredominatedbyevaporationanddolomitization(Shirokovaetal.,

2013).Themodelresultsindicatethatthediageneticfluid respon-siblefordolomitizationintheOmbaatjieFormationhadboth low Mg2+/Ca2+ratiosandrelativelow

δ

26Mgvalues(Table1), consis-tent with platform waters influenced by submarine groundwater dischargeenrichedin40_Caand24_{Mgduetocarbonatedissolution}

(Fig.7).

The fluid-buffered interval recording the recovery of the Tre-zonaexcursionisobservedonthreeseparatecontinents(Table1). Acrossthisinterval, modelresults indicatethat primary platform aragonitewithlow

δ

13C values(estimated at-10‰)was dolomi-tizedand/or neomorphosed by a fluid with

δ

13C valuesof +7‰, withanuncertaintyrangeof+4to+12.5‰.Inaddition,themodel estimates fluid

δ

44/40Ca values of -0.5‰ and

δ

26Mg values of +0.1‰ (uncertainty of -0.65 to -0.40‰, and -0.05 to +0.25‰, re-spectively,Table1).

The fact that a single diagenetic fluid mayexplain the trends observed during the recovery of the Trezona excursion on three continentscould indicateapossiblelateCryogenian seawater ori-gin (Fig. 7). Fluid

δ

44/40Ca values of -0.5‰ would suggest more

44_{Ca-depleted seawaterrelativetotoday,whichisconsistent with}

estimates of pre mid-Mesozoic seawater, prior to a deep-marine carbonate reservoir (e.g., Akhtar et al., 2020). However, the low fluid

δ

44/40_{Ca values also are within range of modern}

observa-tions of restricted platform waters that are offset from modern seawater (Holmden et al., 2012a; Shao et al., 2018). Our results thereforecannot showconclusivelythatthe‘recoveryfluid’ repre-sentsopen-oceanseawater,buttherelativelyhigh

δ

44/40Ca values suggesta somewhatmodified seawaterorigin. Ifthe

δ

26Mgvalue and Mg2+/Ca2+ ratio of the recovery fluid provide a close esti-mateofCryogenianseawater(0.1‰and1.1mol/mol,respectively, Fig.7G),thenseawaterwassignificantlyenrichedin26_Mgandhad

lowerMg2+/Ca2+ratiosrelativetothemodern(modernvaluesof -0.8‰and5.1mol/mol,respectively,e.g.,Higginsetal.,2018).

4.3.TheTrezonaexcursionasalocalphenomenonlinkedtosiliciclastic input

In thissection, we explore the link between the geochemical signals,ourdiageneticinterpretation,andtheobservedchangesin siliciclastic relative to carbonateinput across the Trezona excur-sion.Ingeneral,oncontinentfringingplatforms,mixeddeposition of carbonateand siliciclastic sediments is influenced by regional climate and local factors such as sediment and nutrient influx fromriverdrainagesystemsthatmeander,avulse,andreorganize. Changinginputsofsiliciclasticmaterialcandrivechangesin plat-formwater-depth, affectingplatform fluid circulation,dueto dif-ferencesinpermeability,porosity,andinthechangingsizeofthe

Fig. 7. NumericaldiageneticmodelfitstotheTrezonadata.Toppanels(A,B)depictthebestmodelfittodatafromtherecoveringlimboftheTrezonaexcursion(recovery fluid).Themiddlepanels(C,D)showthebestmodelfittodatafromthedownturnoftheTrezonaexcursion(thedownturnfluid).Datapointsarecoloredbytheirstandard deviationofthemodeledpercentalteration.Inotherwords,reddatapointsarenotwellexplainedbythediageneticmodel.Thebottompanels(E–I),showtheresultsforthe modeloptimizationwiththebestfitvalue.Thegray areashowstheuncertaintylevelofthemodelcostwhenaccountingfortheuncertaintyofthedata(66thpercentile). Thesignificancelevelrepresentsthe95thpercentileofmodelfitstorandomlygenerateddata(seemethodsfordetails).Forcomparison,modernseawater(e.g.,Higginset al.,2018)andre-calculatedmodelresultsforpost-Marinoanseawateralsoareplotted(basalEdiacarancapcarbonates,Ahmetal.,2019).

freshwaterlens.Forexample,anincreaseinsiliciclasticinputoften suffocatescarbonateproduction,leadingtoadecreaseinsediment accumulation raterelative tosubsidence rate, andincreases plat-form water-depth(Schlager, 1989). An increase inlocal sea-level candriveupwellingofdeep-seatedgroundwaterfromtheplatform interior as thermally-driven circulation increases and pore-fluids arepushedupwards(Kohoutetal.,1977).Incontrast,reducingthe supplyofsiliciclasticmaterialoftenallowscarbonateaccumulation tokeepupwithsubsidencerate, decreasingplatformwater-depth and expanding the freshwater lens in platform interiors. As the lessdensefreshwater flows seawards,it drivesan increase inthe deeper compensatingflow ofseawaterinto platforms(Henderson etal.,1999).Reorganizationofriverscanchange,notonlythe plat-formwaterdepthandsubsurfacefluidflow,butalsothechemical composition of the coastal surface waters (inputs of siliciclastics, nutrients, alkalinity,andremineralizedorganiccarbon)and there-forethe

δ

13C ofcarbonateprecipitatedfromthem(Pattersonand Walter,1994).

4.3.1. Cryogenian13_{Cenrichedplatformaragonite}

Onallthreecontinents,theinterval precedingtheTrezona ex-cursion is characterized by a high fraction of carbonate relative to siliciclasticmaterial,withhighcarbonate

δ

13C values(

∼

+10‰, Fig. 8, panel 1). Calcium isotopes and Sr/Ca ratios indicate that these stata originally were primary aragonite. High

δ

13C values in Cryogenian platformsurface waters mayreflect generallyhigh Cryogenian seawater

δ

13C values,but we speculate that at least partofthisenrichmentisrelatedtoacombinationofelevated

pri-maryproductivityanddiurnalcyclinginmicrobialmatdominated platformenvironments (GeymanandMaloof,2019).The combina-tionoflow

δ

44/40_{Ca(<-1‰)andhigh}

_δ

13_{Cvaluesimplythatthese} 13_{C-enrichedintervalscannotrepresenttheaveragecarbonatesink}

andneedtobebalancedbytheburialof44_{Caenriched(and}

pos-siblymore13_{Cdepleted)carbonateinotherlocalities.}

4.3.2. TheTrezonaexcursion

The downturn of the Trezona excursion only is recorded in dolostoneinNamibia,whileinotherlocalitiespre-Trezona carbon-ate gives way to siliciclastic strata (Fig. 8, panel 2). In Namibia, we hypothesize that progressive platform drowning is responsi-blefor a temporary increase inplatform fluid-flow that resulted in dolomitization of the underlying carbonate (parasequence b7, Fig.4). Thisphase of dolomitizationdeclines in intensityin con-certwiththedeclinein

δ

13Cvaluestowardsthenadir.Themodel results suggest that the pore-fluids generated from re-circulating platform waters had

δ

13C values

∼

-4‰ (Table 1), likely due to subsurface mixing between platform surface waters (more 13

C-depleted, see below) and upwelling seawater (open-ocean, less

13_C-depleted).

Although the downturn of the Trezona excursion is observed onlyinNamibia,low

δ

13Cvalues(between-7and-10‰)are ob-served informer aragonitein thenadir ofthe Trezonaexcursion worldwide.Theseobservations suggestthat

δ

13C valuesofDICin platform water, where aragonite likely precipitated, shifted from hightolowin concertwithincreasingsiliciclastic relativeto car-bonate input. The nadir of the Trezonaexcursion is recorded in

A.-S.C. Ahm, C.J. Bjerrum, P.F. Hoffman et al. Earth and Planetary Science Letters 568 (2021) 117002

Fig. 8. A. Interpretationofprimaryanddiageneticend-membersobservedacrosstheTrezonaexcursion.Cryogeniancarbonatesrecordchangesinthediageneticregime betweenfluid- tosediment-bufferedthatconsistentlycorrelatewithprominentstratigraphicmarkersandchangesinδ13_{Cvalues.(1)TheintervalprecedingtheTrezona}

excursionischaracterizedbyshallow-watercarbonatewithδ13Cvalues∼+10‰informeraragoniticsediments.(2)ThedownturnoftheTrezonaexcursionisrecordedbya decreaseinδ13CvaluesandisonlyobservedinNamibia.Thedownturnisinterpretedasaresultoffluid-buffereddolomitizationoftheunderlyingsedimentassociatedwith drowningoftheOmbaatjieplatform.(3)Thenadiroftheexcursionismarkedbydepositionofsiliciclasticstataandareductioninfluid-flow,resultinginthepreservationof formerplatformaragonitewithδ13_{Cvalues∼-10‰.(4)Therecoveryoftheexcursionisassociatedwithanincreaseincarbonatedepositionrelativetosiliciclastic,correlating}

withincreasingδ13_{Cvalues.Thisintervalalsorecordsincreasedfluid-buffereddiagenesisofplatformsediments. B.δ}13_{Cvaluesand C.δ}44/40_{Cavaluesarecoloredbythe}

predicteddegreeofalterationfromthediageneticmodel.

settings where aragonite is interbedded with fine-grained silici-clastic sediments,decreasing permeabilityandpossiblyprotecting thearagonitefromearlydiagenesis(Fig.8,panel3).

Thedrivingmechanismforgeneratingextremelydepleted

δ

13C values in Cryogenian surface waters remains enigmatic, but our analysisindicates a linkto siliciclasticinput andastrong diurnal engineinplatformsurfacewaters(withbackground

δ

13Cvaluesof >+7‰). We speculate that an increase inthe input of siliciclastic sedimentsfromriverdrainagemaybeassociatedwithaninfluxof nutrients, whichcouldlead toalocalizedburstinproductivityof microbialmats.Increasedmicrobialproductivityprovidesa possi-blelinktotheextremelydepleted

δ

13Cvaluesobservedinthe Tre-zona nadir.Inmodern hypersalinepondsdominatedbymicrobial mats, periods of intense productivity have been associated with low

δ

13CvaluesofDIC (<-10‰, LazarandErez,1992).Theselow values are a product of kinetic isotopic fractionation during CO2

hydration (estimated kinetic fractionation factors

∼

-11‰, Zeebe andWolf-Gladrow,2001).Inother words,whilesurfacewaterDIC concentrationsare significantly depleteddueto highproductivity rates,CO2 isreplenished bytherelatively slowinvasionfromthe

atmosphere, whichisassociatedwithalargenegativecarbon iso-topefractionation(LazarandErez,1992).

Therearesomeimportantdifferencesbetweenthemodern hy-persaline ponds(Lazar andErez,1992) and the Cryogenian plat-formenvironments.First,whilethemodernpondsarehypersaline, thereisnoevidenceforevaporitesintheCryogeniansuccessions. Second, in the modern ponds, carbonate precipitation is limited during the most extreme periods of disequilibrium, because the invasion of CO2 decreases carbonate saturation (Lazar and Erez,

1992).However,CO2invasionhasbeenconnectedtorapid

carbon-ate precipitation in alkaline environments witha highsupply of Ca2+ _{(Clark et al., 1992). In these environments, subsurface} flu-ids with very low DIC/Ca2+ _{ratios come into contact with the}

atmosphere,resultinginrapidinvasionofCO2andprecipitationof

carbonateswith

δ

13_{Cvaluesdownto-25‰(Clarketal.,1992).We}

imaginethatCryogenian platformwaterswerepoisedsomewhere in-between these modern end-members, consistent with model resultsoflowDIC/Ca2+ratiosinthe‘downturnfluid’(Table1). Ad-ditionally,whilehypersalinitypromotesdisequilibriumduetovery slow rates of air-sea gas exchange, disequilibriumalso has been observed in freshwater lakes during intense algal blooms where rates of productivity exceed rates of CO2 invasion (Herczeg and

Fairbanks,1987).WethereforespeculatethatCryogenianplatform environments dominated by microbial mats and associated with highbackgrounddiurnalproductivity(as evidencebybackground

δ

13C valuesofup to+10‰), maybe poisedcloseto a disequilib-riumthresholdwherethekinetic effectsofCO2 invasioncouldbe

readilyexpressed.Inthisworld,theTrezonaexcursioncapturesa periodwhereanincreaseinsiliciclasticinfluxdroveplatform sur-facewateracrossthisthreshold,resultingin

δ

13Cvaluesof

∼

-10‰.

4.3.3. TheTrezonarecovery

FollowingthenadiroftheTrezonaexcursion,platform carbon-ate

δ

13C eventually recovers before the onset of the Marinoan glaciation. On all three continents, this interval is characterized byanincreaseincarbonatedepositionrelativetosiliciclastics,and coincides with increasing fluid-buffered diagenesis (Fig. 8, panel 4).Ascarbonateinputincreasedandaccommodationspace dimin-ished, platform water-depth decreased, the fresh-water lens ex-panded,andseawaterre-circulationinplatformsintensified(Fig.8, panel4).Anincreaseinfluid-flowacrossthiscoarseningupwards succession is consistent with observations of fluid-flow patterns fromtheBahamaswhereperiodsofrelativesea-levelfallare char-acterizedbyincreasedaragoniteneomorphism anddolomitization (Vahrenkampetal.,1991;Melimetal.,2002).

Thelate-Cryogenian ‘recoveryfluid’responsibleforwidespread diagenetic resetting of platform carbonate had a

δ

13C value of

∼

+7‰,andmostlikelyreflectsthecompositionofplatformsurface waters that have returned to pre-Trezona conditions(from being

13_C-depletedto13_{C-enriched,Fig.8,panel4).}

4.4. Localplatformsignalsandglobalmassbalance

TheprevalenceoflocalgeochemicalsignalsinCryogenian car-bonatesuccessionshasimplications fortheglobalmassbalanceof both carbon and calcium, andthe degree to which the platform carbonatesrepresenttheaveragecarbonatesink.Animportant ob-servationisthatthemajorityofcarbonatesmeasuredinthisstudy have

δ

44/40_{Ca valuesthat are lowerthan BSE (-1‰,Skulanetal.,}

1997), and mass balance requires that these low

δ

44/40_Ca

val-uesarebalanced bytheprecipitationof44_{Ca enrichedcarbonates}

elsewhere (Blättler and Higgins, 2017). As the geological record is incomplete, it isnot possible to show conclusively where this sink should be. One possibility is authigenic carbonate and ce-ments, which tend to have high

δ

44/40Ca values (Blättler et al.,

2015). Another possibility is the precipitation ofcarbonate veins during hydrothermal alteration of basalt (Bjerrum and Canfield,

2004).Athirdpossibilityistoreexaminethe extenttowhichthe TrezonaexcursionisobservedinCryogenianstrataglobally.For ex-ample,theexcursionisnotobservedintheCryogenian carbonate-dominated successions of Mongolia and Panamint Range, Death Valley (Bold et al., 2020; Nelson et al., 2021), which previously hasbeeninterpreted asaresultofsub-glacialerosion(Boldetal.,

2016).Alternatively,it ispossiblethattheseplatforms simplydid not reachadisequilibriumthresholdwhere

δ

13Cvaluesswitch to -10‰, and dueto relatively less siliciclasticinput, also remained moresusceptibleto earlydiagenesis andfluid-bufferedalteration. Mass balance could be achieved if fluid-buffered intervals char-acterized by high

δ

44/40_{Ca values (and less extreme}

_δ

13_{C values}

closerto

∼

0‰)were correlatedwithsediment-buffered intervals characterized by low

δ

44/40Ca (and more extreme

δ

13C values,

∼

+10and-10‰).Theseintervalswouldnot havebeencorrelated previouslyduetothepracticeofusingcarbonisotopestratigraphy inCryogeniansuccessions.

Another feature ofour hypothesis isthat it suggeststhat the correlation between siliciclastic input and negative

δ

13C excur-sions in platform aragonite could be a common Neoproterozoic phenomenonduringperiodsofhighbackgrounddiurnalcarbon cy-cling (baseline

δ

13C>+5‰). Indeed, other excursions such as the Shuram-Wonoka, the Taishir, and the Islay anomaly are associ-ated with a change from 13_{C-enriched to} 13_C-depleted

carbon-ate, coinciding with a fractional increase in siliciclastic material (e.g., Husson et al., 2015; Bold et al., 2016; Park et al., 2019). It maybe possible to changeglobalsiliciclastic fluxesthough ei-ther global climatechange or rapid glacio-eustatic sea-level rise, although currentlythereis noevidence forland-basedice-sheets inthe Cryogenian ‘non-glacialinterlude’. Alternatively,inmodern mixed platform environments, the influx of siliciclastic material largely iscontrolledby stochastic fluvial-deltaicandtectonic pro-cesses that operate on timescales from thousands to millions of years.IfsimilarlocalprocessesareoperatinginCryogenianmixed carbonate-siliciclastic platforms andcontributingtothe switch to negative

δ

13C values,itwould be consistentwiththe lackof the Trezonaexcursioninsome Cryogeniansuccessions(e.g.,Mongolia andPanamintRange)andthelackoftheTaishir anomalyin oth-ers(e.g.,Australia, Namibia,Northwest Canada). Moreover,in the localitieswheretheexcursionisfound,itisnotyetclearifit cor-relates across continents on time-scalesthat are relevantfor the globalcarbonandcalciumcycle(<105

−

106 years).Strontium iso-toperatiosandglacialdiamictite-capcarbonatelithostratigraphical correlationsbroadlyconstrain theTrezonaexcursions within

∼

10

Myr (Fig. 1), but these correlation tools do not provide the res-olutionto confirm that theseexcursions are coeval. Instead,it is possiblethattheswitchfrom13_C-enrichedto13_C-depleted

carbon-ateonlocalplatforms occurredseparatelyacrossabroaderperiod (e.g.,between1–5Myrs),whereglobalclimateandtectonics con-tributedtoalocallyvariable,increasedfluxofsiliciclasticmaterial toplatformenvironments.

5. Conclusions

Thisstudy demonstrates that stratigraphic changes in Cryoge-nianplatformcarbonate

δ

13Cvaluesarecharacterizedbyintervals offluid-versussediment-buffered diagenesis linked tochanges in the relative input of carbonate and siliciclastic sediments. First, the interval preceding the Trezona excursion is characterized by shallow-water conditions with

δ

13C values of

∼

+10‰ in former aragonitesediments.Second,inNamibiathedownturnofthe Tre-zona excursioncorrelates withan increase inplatform fluid-flow rates,dolomitizationoftheunderlyingsediments,andsubsequent platform drowning by siliciclastic material.Diagenetic fluids par-tiallyaresourcedfromplatformsurfacewaterswithexceptionally low

δ

13C andlow

δ

44/40Ca values. Inthe nadir of theexcursion, fluid-flowisreducedduetothedecreaseinpermeabilityfrom in-terbeddedsiliciclastics,resultinginsediment-bufferedpreservation offormerplatform aragonitewith

δ

13C valuesof

∼

-10‰.Finally, therecoveryoftheTrezonaexcursioncorrelateswithan increase incarbonateinputs,anincreaseinplatformfluid-flow,and increas-ing

δ

13Cvaluesofplatformsediments.Thisrelationshipsuggestsa mechanistic link between intervalsof siliciclastic input and

δ

13C excursionsinCryogenianplatformenvironments.

CRediTauthorshipcontributionstatement

Anne-SofieC.Ahm: Conceptualization,Formalanalysis, Method-ology,Visualization, Writing–originaldraft. ChristianJ.Bjerrum:

Supervision, Writing – review & editing. Paul F.Hoffman:

Re-sources,Visualization,Writing–review&editing. FrancisA. Mac-donald: Resources,Visualization,Writing–review&editing. Adam

C.Maloof: Conceptualization,Resources,Writing– review& edit-ing. CatherineV.Rose: Resources,Visualization,Writing–review& editing. JustinV.Strauss: Resources,Visualization,Writing–review &editing. JohnA.Higgins: Conceptualization,Resources,Writing– review&editing.

Declarationofcompetinginterest

Theauthorsdeclarethattheyhavenoknowncompeting finan-cialinterestsorpersonalrelationshipsthatcouldhaveappearedto influencetheworkreportedinthispaper.

Acknowledgements

We thank Louis Derry for editorial handling. In addition,this study greatly benefited from reviews by Ashleigh Hood, two anonymousreviewers,anddiscussionswithJamesBusch,Laurence Coogan, Blake Dyer, Emily Geyman, and Jon Husson. We would liketothankEbenBlakeHodginandAlanRooneyforprovidingSr isotope datafor sectionsfrom Northwest Canada.We alsothank ElizabethLundstromandNicolasSlaterforassistanceinthelabat PrincetonUniversity.Thisworkwassupportedbyagrantfromthe SimonsFoundation (SCOL611878, ASCA)andtheCarlsberg Foun-dationtoASCA.ASCAandCJBalsoacknowledgesupportfromthe Danish National Research Foundation (Grant No. DNRF53). ACM andCVRacknowledgesupportfromNSF(EAR-0842946)for fund-ing fieldwork on the Trezona Formation in South Australia. JAH acknowledgessupportfromNSF(IES-1410317)andfromNSFOCE CAREERGrant(1654571).