Assessing the impact of diagenesis on foraminiferal geochemistry from a low latitude, shallow-water drift deposit

Assessing

the

impact

of

diagenesis

on

foraminiferal

geochemistry

from

a

low

latitude,

shallow-water

drift

deposit

Stephanie Stainbank

a

,

∗

,

Silvia Spezzaferri

a

,

Eva De Boever

a

,

1

,

Anne-Sophie Bouvier

b

,

Colin Chilcott

c

,

Erica

S. de Leau

c

,

Anneleen Foubert

a

,

Tereza Kunkelova

c

,

2

,

Laetitia Pichevin

c

,

Jacek Raddatz

d

,

e

,

Andres Rüggeberg

a

,

James

D. Wright

f

,

Siyao

M. Yu

f

,

Manlin Zhang

c

,

3

,

Dick Kroon

c

a_{DepartmentofGeosciences,UniversityofFribourg,CheminduMusée6,1700Fribourg,Switzerland} b_{InstituteofEarthSciences,UniversityofLausanne,Géopolis,1015Lausanne,Switzerland}

c_{SchoolofGeoSciences,UniversityofEdinburgh,GrantInstitute,TheKing’sBuildings,JamesHuttonRoad,EH93FEEdinburgh,UnitedKingdom} d_{InstituteofGeosciences,GoetheUniversityFrankfurt,Altenhöferallee1,60438Frankfurt,Germany}

e_{FrankfurtIsotopeandElementResearchCenter(FIERCE),GoetheUniversityFrankfurt,Altenhöferallee1,60438Frankfurt,Germany}

f_{DepartmentofEarthandPlanetarySciences,Rutgers,theStateUniversityofNewJersey,BushCampus,610TaylorRoad,Piscataway,NJ08854-8066,USA}

Keywords:

foraminifera earlydiagenesis tropicaldriftdeposits stableisotopes Mg/Ca IODP

Duetotheirlargeheatandmoisturestoragecapabilities,thetropicsarefundamentalinmodulatingboth regionalandglobalclimate.Furthermore,theirthermalresponseduringpastextremewarmingperiods, suchassuperinterglacials,isnot fullyresolved.Inthisregard,wepresent high-resolution(analytical) foraminiferalgeochemical(δ18OandMg/Ca)recordsforthelast1800kyrfromtheshallow(487m)Inner SeadriftdepositsoftheMaldivesarchipelagointheequatorialIndianOcean.Consideringthediagenetic susceptibility of these proxies, incarbonate-rich environments, we assess the integrity of a suite of commonlyusedplanktonicandbenthicforaminiferageochemicaldatasets(Globigerinoidesruber (white), Globigerinitaglutinata (withbulla),Pulleniatinaobliquiloculata (with cortex)andCibicidesmabahethi)and theiruseforfuturepaleoceanographicreconstructions.

Using a combination of spot Secondary Ion Mass Spectrometer, Electron Probe Micro-Analyzer and Scanning Electron Microscope image data, it is evident that authigenic overgrowths are present on boththeexternalandinternaltest(shell)surfaces,yetthedegree down-coreaswellastheassociated biasisshowntobevariableacrosstheinvestigatedspeciesandproxies.Giventheelevatedauthigenic overgrowth Mg/Ca(∼12–22 mmol/mol)and δ18_{Ovalues (closerto thebenthic isotopiccompositions)}

the whole-test planktonicG.ruber (w) geochemical recordsare notably impacted beyond ∼627.4 ka (24.7 mcd). Yet, considering the setting (i.e. bottom water location) for overgrowth formation, the benthic foraminifera δ18_{Orecord is markedly less impacted withonly minor diagenetic bias beyond} ∼790.0 ka(28.7mcd).Eventhough onlythetopoftheG.ruber (w)and C.mabahethi records (whole-test data)would besuitablefor paleo-reconstructionsofabsolutevalues (i.e.seasurfacetemperature, salinity, seawater δ18_{O),the long-termcycles, while dampened,appear tobe preserved.Furthermore,}

planktonicspecieswiththicker-tests(i.e.P.obliquiloculata (w/c))mightbebettersuited,incomparison tothinner-testcounter-parts(i.e.G.glutinata (w/b),G.ruber (w)),fortraditionalwhole-testgeochemical studiesinshallow,carbonate-richenvironments.Athickertestequatestoasmalleroverallbiasfromthe authigenicovergrowth.Overall,ifthediageneticimpactisconstrained,asdoneinthisstudy,thesetypes ofdiageneticallyalteredgeochemicalrecordscanstillsignificantlycontributetostudiesrelatingtopast tropicalseawatertemperatures,latitudinalscaleoceancurrentshiftsandSouthAsianMonsoondynamics.

*

Correspondingauthor.

E-mailaddress:stephanie.hayman@unifr.ch(S. Stainbank).

1 _{Presentaddress:TNO,GeologicalSurveyoftheNetherlands,Princetonlaan6,3584CBUtrecht,theNetherlands.}

2 _{Presentaddress:NationalOceanographyCentreSouthampton,UniversityofSouthampton,WaterfrontCampus,EuropeanWay,SO143ZHSouthampton,UnitedKingdom.} 3 _{Presentaddress:FacultyofLifeScience,UniversityofBradford,RichmondRd,BD71DPBradford,UnitedKingdom.}

http://doc.rero.ch

3XEOLVKHGLQ(DUWKDQG3ODQHWDU\6FLHQFH/HWWHUVGRLMHSVO

ZKLFKVKRXOGEHFLWHGWRUHIHUWRWKLVZRUN

1. Introduction

Incomparisontodeep-searecords,high-resolution (both tem-poral and analytical) paleoclimate reconstructions for shallow-intermediate,equatorialregions arelimited.Thisisinpartdueto (i)thedrivetoreconstructandmodelpastglobaloceancirculation patterns,whichpredominantlyusehigh-temporallyresolved deep-seaarchives,andwhicharedifficulttorecoverfromlowerlatitudes and(ii)theincreased susceptibilityofcalcifyingorganismsto di-agenetic overprint in shallow, carbonate-rich environments. The tropicsare anintegral componentofbothregionalandglobal cli-maticsystemsduetotheir heatandmoisturestoragecapabilities (Guptaet al.,2010andreferenceswithin;Hastenrathet al.,1993; Schott et al., 2009). Yet, notably there is a lack of reliable sea surfacetemperature(SST)and

δ

18Orecordsfromtheseequatorial regions, particularly from past extreme warming periods, which wouldfurtherfacilitatetheunderstandingoftropicalclimatic sys-tems.Inthisregard,wepresentanewhigh-resolution(analytical) tropicalrecord fromtheInner Seaofthe Maldivesarchipelagoin theequatorialIndianOcean.Wetestthesuitabilityofthisunique tropicalrecordforpaleo-reconstructionsgivenitsshallowlocation, proximitytocarbonatebanksandtheanticipateddiagenetic alter-ationdown-core.

Innumerable studies have shown that the geochemistry of foraminiferaltests(shells)areinvaluablepaleoceanographic,-climatic and-ecologicalarchives(e.g.Birchet al.,2013;KroonandGanssen, 1989; LisieckiandRaymo,2005;Groeneveldet al., 2008; Raddatz et al.,2017;Stainbank et al., 2019). Yet,manyofthese geochem-ical proxies (e.g. Mg/Ca and

δ

18_{O) are susceptible to diagenetic}

alteration, which calls into question their fidelity (Edgar et al., 2015, 2013; Groeneveld et al., 2008; Panieri et al., 2017). Ac-cordingto Edgar et al. (2015) the foraminifera test wall, texture andmorphologyare prone toatleastthree diageneticprocesses: overgrowth,partialdissolutionandrecrystallization.Allthreehave thepotentialto biastheoriginaltest geochemistry andaffectthe accuracy of paleo-reconstructions. As such, prior to undertaking paleo-reconstructions, it isparamount to establish thediagenetic state of this foraminiferal calcite. This is particularly important within carbonate-rich environments which have high sediment porosity and permeability and thus a higher “diagenetic poten-tial”incomparisontothemore impermeableclay-richsediments (Edgaret al.,2013).Additionally,incomparisontothemorerobust, heavily calcified benthic foraminifera, planktonic tests are more susceptible to alteration (Edgar et al., 2013). Moreover, as there isa largearray of planktonic foraminiferatest wall textures;this influence is not necessarily uniform between species. Bé (1977) correlatedincreaseddiageneticresistanceforspecieswithsmaller pores,thickerwalls andlargertest size. Theseconsiderations are importanttoaidintheselectionoflessdiageneticallypredisposed foraminiferalspecies, particularly fordepth stratified (i.e. studies assessingdiscretelivingandcalcificationdepths withinthewater column)thermoclinereconstructions.

To contribute to an improved understanding of ocean-atmos-phericinteractions within the tropical Indian Ocean,and the as-sociated foraminiferaldiagenetic processes,we presenthigh ana-lyticallyresolvedforaminiferaldatasetsfromtheshallow,Maldives Inner Sea. Specifically, our study addresses the following issues: (i) theoverall preservationstate ofthe individual long-term geo-chemicalrecordsandtheirsuitabilityforfuturepaleoceanographic interpretationsand(ii)thedifferential susceptibilitytodiagenesis forthree planktonic foraminiferalspecies, withdiscrete wall tex-tures,anda benthiccounterpart using a combinationofimaging andelemental/isotopicgeochemicalapproaches.

Fig. 1. StudylocationandpositionofIODPExpedition359SiteU1467intheInner SeaoftheMaldives(bluecircle),locatedinthenorthernequatorialIndianOcean (GEBCOCompilationGroup,2019).(Forinterpretationofthecolorsinthefigure(s), thereaderisreferredtothewebversionofthisarticle.)

2. Methods

This study uses samples retrieved during the International OceanDiscoveryProgram(IODP)Expedition359in2015.All sam-plescomefromSiteU1467 (4◦51.0274N,73◦17.0223E)drilledat a water depth of 487 m within the drift deposits of the Inner SeaoftheMaldives archipelagointhenorthern equatorialIndian Ocean (Fig. 1) (Betzler et al., 2017). The base of these drift de-positshavebeenbiostratigraphicallydatedto 12.9-13.0Ma, coin-cidingwiththeestablishmentoftheSouthAsianMonsoon(Betzler et al.,2018,2016).Assuch,thedriftdepositsrepresentan impor-tantpaleoceanographicand-monsoonarchiveduetotheirunique abilitytopreservecontinuoussedimentrecordsbybottomcurrent deposition (Lüdmann et al., 2013). The top 61 meters compos-ite depth(mcd)fromHoleU1467A,B, C,composed ofunlithified foraminifer-richwackestonetopackstone,wereusedinthisstudy. The shipboard splicewas adjusted after theExpedition based on X-rayfluorescencecorescannerdatausingFeintensity(countsper second,cps)(Kunkelovaet al.,2018).

Long-term,high-resolution(analytical)whole-test(whole-shell) stable isotopic (

δ

18_{O) records are compiled for three planktonic}

species, Globigerinoidesruber (white), Globigerinitaglutinata (with

bulla), Pulleniatina obliquiloculata (with cortex) and one benthic representative Cibicidesmabahethi. In a few cases when the tar-getbenthicspeciesisrare,Cibicideswuellerstorfi testsareincluded to ensureenough calcite fortheisotopic measurements. Comple-mentarywhole-testMg/CaratiosaremeasuredforG.ruber (w).

Acombinationofmethodsisutilizedtoconstraintheinfluence andsusceptibilityofdiageneticalterationontheforaminiferaltest geochemistry.Avisual ScanningElectronMicroscope(SEM)check is implemented, on selected samples, for both the target plank-tonicandbenthicspecies.Thisvisualcheckisthenusedtodefine aqualitativeDiagenesisRank(DR)toascertainandtrackthe down-core change in the preservation state of the foraminiferal tests. Supplementary spot-geochemical (isotopic and elemental) mea-surementsareobtainedusinga SecondaryIonMassSpectrometer (SIMS)andElectronProbeMicro-Analyzer(EPMA).

2.1. SiteU1467mineralogyandgeochemistry

SiteU1467sedimentcoreinterstitialwaterchemistryand sedi-mentmineralogyarepresentedinBetzleretal.(2017). Concentra-tionsofCa2+,Mg2+andSr2+weremeasuredfrominterstitial

Fig. 2. IODP Expedition 359 Hole U1467B sediment mineralogy and pore water geochemistry (Betzler et al.,2017).

ter(IW) sampleseither byshipboard Inductively Coupled Plasma Atomic EmissionSpectroscopy(ICP-AES)orshore-based ion chro-matography (see Betzleretal., 2017;Blättler et al.,2019for fur-ther details).Relativeconcentrationsofaragonite, high-Mgcalcite (HMC),low-Mgcalcite(LMC),dolomite,andquartzweremeasured using X-raypowder diffraction (XRD)(seeBetzleretal., 2017for a fulloverviewofthedatacollectionmethods).Forsimplicitywe showonlythedata,reportedasmbsf(metersbelowseafloor),from HoleU1467Basityieldedthemostextensiveandintactrecordsat thissite(Fig.2).

2.2. Speciesselectioncriteria

To explore the susceptibilityof different speciesto diagenetic alteration andthebiasesontheir respectivewhole-test geochem-ical compositions, we assess three shallow/intermediate dwelling planktonic foraminiferal species: G.ruber (w), G.glutinata (w/b)

andP.obliquiloculata (w/c).Thesespecies havereportedapparent

calcification depths for the northern equatorial Indian Ocean of

∼

70–77 m, 81 m and 74–104 m, respectively (Stainbank et al., 2019) andhave discrete wall textures,thicknesses and porosities (SupplementaryMaterial 1).Globigerinoidesruber (w)ischosenas it isthepredominant foraminiferalspeciesusedfor surface/near-surface seawater reconstructions (e.g. Birch et al., 2013; Bunzel et al., 2017; Rebotim et al., 2017; Stainbank et al., 2019).

Glo-bigerinoidesruber (w) is a spinose, shallow-dweller with a

can-cellate wall texture anda porosity of

∼

5 pores/25 x 25 μm(Bé, 1968) (Supplementary Material 1). Globigerinitaglutinata (w/b) is included as a second shallow-dweller, as it is non-spinose with a smooth, microperforate wall texture with a porosity of

∼

112 pores/25 x25 μm (Bé,1968; Birch et al., 2013; Stainbank et al., 2019)(SupplementaryMaterial1).Lastly,theshallow-intermediate dwellerP.obliquiloculata (w/c)isincludedasitisnon-spinose,with a smooth wall texture(Birchet al., 2013;Stainbank et al., 2019). Whiletheoriginaltesthasaporosityof

∼

9pores/25x25 μm,the formation of the smooth, glossy cortex layer drastically reduces the diameterandspherical nature ofthepores resulting ina re-duction inthe overallporosity (Bé, 1968; Steinhardt et al.,2015) (SupplementaryMaterial1). Cibicidesmabahethi, acalcareous hya-lineepibenthicspecies,ischosen asthebenthicrepresentativefor this site as it has been shown to accurately record bottom wa-terconditionsforthefirst

∼

200kyr(timeintervalofthestudyby Bunzelet al.,2017)fortheMaldivesInnerSea(Bunzelet al.,2017).

2.3. Traditionalwhole-testforaminiferalgeochemicalanalysis

Sedimentsampleswereairdried,weighedandwashedthrough a32 μmsievetoremovethefinersiltsandclayfraction.Theywere

then oven-dried at 30◦C for 48 hours, weighed and dry sieved into discrete fractions forforaminiferal picking. For all geochem-icalanalyses, specieswere pickedfrompre-defined, narrow sieve sizefractions,tominimizeintra-specific ontogeneticisotopic frac-tionation effects (Stainbank et al., 2019). Additionally, to remove any adhering particles, and to prevent carbonate contamination, all picked specimens were precleanedin Milli-Qwater fora few secondsbyultrasonication.

In total, 3802 traditional whole-test stable oxygen isotopic (

δ

18_{O) measurements were made (data is available on Pangaea.}

Note: 6 datapoints are from Stainbank et al., 2018). From 0–61 mcda measurement for G.ruber (w: 212–250 μm) was obtained every

∼

3cm.IsotopicanalysesofG.glutinata (w/b:125–150 μm)

andP.obliquiloculata (w/c:355–400 μm), were carriedout across

threeselectedintervalswithdifferentialdiageneticinfluences(see Section 3.2 for more details) at a resolution of 3-12 cm; Inter-val 1 (15.2–20.5 mcd), Interval 2 (31.4–34. 5 mcd) and Interval 3(53.3–56.1mcd).Benthic isotopicdata(C.mabahethi >180 μm) were onlygeneratedata

∼

3cmresolution between0–40.0mcd asbelow this interval they become scarce.The G.ruber (w) and

C.mabahethi

δ

18O measurements for the top 6.7 mcd (n= 441)

wereanalyzedatRutgersUniversityonaMicroMass(FISONS) Op-timaIsotopeRatioMassSpectrometerwithanattachedmulti-prep device.Sampleswerereactedwith100%phosphoricacid(H3PO4)

at90◦C for15min,andtheevolvedCO2 gascollectedina liquid

nitrogencoldfingerandanalyzedcomparedtoareferencegas.The samplesarecorrectedusinganinternallaboratorystandard(RGF1), whichisroutinely runagainstNBS-19andNBS-18.Allother

δ

18_O

data(n =3361)were obtained attheGrant Institute of the Uni-versityofEdinburghonaThermoElectronDelta+AdvantageMass Spectrometer integrated with a Kiel carbonate III automated ex-tractionlineandcorrectedusingthelaboratory’sinternalstandard. Approximately 0.05 mgwasrequired foreachmeasurement with each expressed as parts per mil (



) relative to Vienna Pee Dee Belemnite (VPDB). Replicate measurements ofthe standards give theinstrumentsananalyticalprecision (1

σ

)of

∼

0.05



for

δ

18_O

and

δ

13_{C. The reproducibility, based on inter-laboratory replicate}

measurements(n=6),was

∼

0.20



.

Mg/Caratiosweremeasuredon30pooledG.ruber (w:212–250 μm) testsfromasample every

∼

3cm,forthe interval14.8–31.4 mcd (n = 535; data is available on Pangaea). All samples were cleaned according to the standard oxidative protocol of Barker et al.(2003).Sampleswerethenleachedwithaveryweak0.001N HNO3 acid solution prior to dissolving in 0.075MHNO3.The

re-ductivecleaningstepwasexcludedasithasbeenshowntocause thedissolutionoftheforaminiferaltestsresultinginsubsequently lower Mg/Ca values (Barker et al., 2003; Elderfield et al., 2006).

Analyses were conducted at the Institute of Geosciences of the Goethe University of Frankfurt andat the School of GeoSciences at the University of Edinburgh. The former measurements (n = 232)werecarriedoutbyInductivelyCoupledPlasmaOptical Emis-sionSpectrometry(ICP-OES)onaThermoscientificiCap6300(dual viewing) andthe latter (n = 303) on a Vista Pro ICP-OES. Inter-laboratoryreplicates (n=11)were run toassess potential differ-encesandthereproducibilitywhichwas determined tobe better than0.30mmol/mol(meandifference).InFrankfurt,thefinal cen-trifugedsamplesolutionsweredilutedwithanyttriumsolution(1 mg/l)priorto measurementto allowforthecorrection ofmatrix effects.Whereas,inEdinburghasetofsixstandardsolutions(with increasingCaconcentrationsfrom0to100%)wereusedformatrix andbackgroundcorrection(Sadekovet al.,2013).Atbothfacilities, fivecalibrationsolutionsweremeasuredbeforeeachanalysisto al-lowforintensityratiocalibrations.Element/Cameasurementswere standardizedusinganinternal consistencystandard(ECRM752-1, 3.761mmol/molMg/Ca).ReplicatesoftheECRMattheEdinburgh (n=61)andFrankfurt(n=45)facilitiesyielded2

σ

Mg/Ca uncer-tainties (SD) of0.15 and 0.17mmol/mol, respectively. During all Mg/Cameasurements,theelementsAl,Fe,andMnwerescreened tocheckforMn-Feoxidecoatingsandclaymineralcontamination. Proceduralblankswereroutinelyruntomonitorforpotential con-taminationduringthecleaningprocess.

2.4.ElectronProbeMicro-Analyzer(EPMA)analysis

Samplesforboth EPMA andSIMS (Section 2.5) analyses were selectedtoproviderepresentatives fromtheentirelength ofcore includingalteredversus non-alteredportions (see Section 3.2 for samplelocations).Polishedcross-sectionsofG.ruber (w)(n=36)

andP.obliquiloculata (w/c)(n=6)testswere preparedby

embed-dingthetestsinStruersEpothinepoxy.Briefly,selectedtestswere (i)cleanedaccordingtothesameoxidativeprotocol(Barkeret al., 2003)asforthetraditionalwhole-testMg/Cageochemicalanalyses and(ii)placedumbilicalside-downonadhesivefilmwithin1-inch Struers mounting cups, then (iii) embedded in Epothin and de-gassedunderhighvacuumfor5minutesandonceset(iv)polished usingasuccessivelyfinerStruerswater-baseddiamondsuspension from9 μmdownto0.25 μmtoexposeacross-sectionofeachtest (SupplementaryMaterial 2). All EPMAsamples were then coated in5nmofcarbon.

QuantitativeMgandCaspotanalysesandsemi-quantitative el-emental maps were measured using a JEOL JXA-8530F Electron Probe Micro-Analyzer, equipped with a Schottky field emission gun, atthe Faculty of Geosciences and Environment of the Uni-versityofLausanne.Calcite,CaCO3(Ca=39.98wt%)anddolomite,

CaMg(CO3)2 (Mg=13.18wt%)wereusedasstandardsforCa and

Mg, respectively. Spot measurements were acquired using a 10 kVaccelerating voltageandprobe size of1–2 μm. Overview ele-mentalmaps were carried out witha 0.50 μm step size,an

im-ageresolution of720

×

600pixels, 7kVaccelerating voltage,an electron beamcurrent of10 nAand a dwelltime of50 ms. Ad-ditional higher resolution elemental maps were obtained witha 0.23–0.29 μmstepsizeandanimageresolutionof256

×

192pixels.

2.5. SecondaryIonMassSpectrometer(SIMS)analysis

Polished cross sections, of selected G.ruber individuals (n = 27) were prepared for SIMS analyses following the same proto-col asfor theEPMA samples(Section 2.4), with theexclusion of the initialoxidative cleaning step. Insitu oxygen isotopeanalyses werecarriedoutusingaCamecaIMS1280-HRSecondaryIonMass Spectrometer (SIMS) at the SwissSIMS ion probe facility within theUniversityofLausanne(SupplementaryMaterial 3).Aprimary Cs+ ion beam with an intensity of 1.2nA was focused to a size of

∼

10 μm. Thesecondary16_Oand18_{Owere detected}

simultane-ouslyontwofaradaycups,whichwerecalibratedatthebeginning of the session. The mass resolving power was set to 2400 (en-trance slit set at121 μm and multicollection slit1), in orderto fullyremovepossible17_{OHinterferenceonmass18.Asingle}

anal-ysis took

∼

4 min, including30 sec. of pre-sputtering, automatic centering ofsecondarydeflectors and20cyclesof5sec. of mea-surements. Calibration was made using an in-housepure calcite standard (UNIL-C1). Six analyses of the standard were made at the beginning of the measurements foreach mount, after which UNIL-C1 was measured once every four analyses. Reproducibility ofthestandard(spottospot)was 0.4and0.3



(2

σ

)onthetwo mounts,respectively.Internaluncertainty(singleanalysis)was0.3 to0.4



(2

σ

).

2.6. ScanningElectronMicroscopy(SEM)

Thirty-five samplesfromthetop 61mcdofthe coreareused to constrain the preservation state of the target planktonic and benthic foraminiferal species. Samples are positioned at regular intervals down-coreand includerepresentatives from glacial and interglacialpeaks(seeSection3.2forsamplelocations).Fromeach sample, threeto fourG.ruber (w), G.glutinata (w/b), P.

obliquiloc-ulata (w/c) andC.mabahethi/wuellerstorfi tests were picked from

thesamesizefractionsasusedforthegeochemicalanalyses.Tests were cleanedforafew secondsby ultrasonicationinMilli-Q wa-ter,crackedopenandmountedoncarbonbasedadhesivediscson standard SEMstubs.Stubs werecoated, undervacuum, in40nm of goldusing a Bal-tec SCD 050sputter coater andimagedon a Thermo Fisher SEM FEIXL30SFEG. The preservation state of each speciesforeach sampleisascertained,usingSEMimages, accord-ing to three categories (i) external test surface, (ii) internal test surface and(iii)test walltexture(Table 1). Aranking isassigned foreachcategory andthesum(total outof7) usedtodefine the overall qualitative DiagenesisRank (DR) foreach species, with 0 defining‘pristinetests’and7themost‘diageneticallyaltered’. Table 1

DiagenesisRankcategoriesandclassifications. Ranking Category

1. External surface 2. Internal surface 3. Wall texture 0 Pristineunalteredsurfacetexturewithno

authigenicovergrowthcrystalspresentboth onthetestsurfaceorwithinthepores

Pristinemicrogranularwalltexturewith well-definedpores

1 Smallovergrowthspresent(upto∼0.5 μm) withapatchy-uniformdistribution

Visiblyalteredwalltexturewith authigeniccrystalswithinthewallcross sectionseitherduetoovergrowth precipitationorpartialrecrystallization 2 Mostlymediumsizedovergrowthspresent

(∼0.5-2 μm)withauniformdistribution 3 Mostlylargesizedovergrowthspresent

(>2 μm)withauniformdistribution

Using the MATLABv2018a software,a color-mapis generated foreachspeciestovisualizethedown-coreevolutionin preserva-tionstate(i.e.thechangeinDR).The35samplesareusedtodefine aninitialmatrix,andtherankingsinterpolatedbetweeneachpoint usingalinearfunction.

2.7. Agemodel

The stratigraphic framework for the upper 61 mcd of Site U1467 is established by graphically correlating both our plank-tonic G.ruber (w) and benthic C.mabahethi

δ

18_{Orecords to the}

stacked reference curve of Ahn et al. (2017) (Prob-stack). Prob-stack was selectedas it isbased on a compilation of 180global benthic isotope records and includes the datasets from the fre-quently used LR04

δ

18Ostack ofLisiecki andRaymo(2005). The age model for our datasets is obtained using Analyseries v2.0.4 (Paillard et al.,1996).To facilitatethetuning process, initiallyall datasets were linearly detrendedandour planktonic andbenthic recordssmoothed(3pt.runningaverage)toeliminatesomeofthe noise.Themainglacialandinterglacialpeaksareusedastheinitial tie-points,withadditionalprominentmaximaandminima subse-quentlycorrelatedtooptimizethevisualalignmentofthedatasets (SupplementaryMaterial4).

3. Results

3.1. Carbonatemineralogy

AtIODPExpedition359HoleU1467B,LMCisubiquitous, gener-allycontributedbyplanktonicforaminiferaandcoccoliths,making up the bulkof thesediment mineralogy fromthe core-topdown to 700mbsf(Fig. 2,Betzleretal.,2017).Aragonite iscontributed mainlybyHalimeda fragments,coralfragmentsandpteropods,and decreasesfrom

∼

46%between0–10mbsftoonly9%at200mbsf. From 200 to 500 mbsf aragonite still has a small, variable con-tribution (0–9 %). However, a marked increase in the aragonite contribution isnotedfrom500–600 mbsf(up to23%)withagain adecreasedownto700mbsf.HMC(withMg2+concentrations>4 mol%; Morse, 2003), which is contributed predominantly by red corallinealgae, benthicforaminifera,bryozoans, andechinoderms, isonlypresent(4–7%)inthesedimentsinthetop10mbsf.Minor peaksindolomitemainlyoccurwithinthetop400mbsf.The pore-fluid species (Ca2+, Mg2+ and Sr2+) are visibly correlated with the aragoniteconcentrations(Betzleretal.,2017). BothCa2+ _and

Mg2+ _{respectively,increaseanddecreaseconcomitantlywitha}

re-ductioninaragonitepercentcontribution.Sr2+_{increasesrapidlyto}

peakconcentrationsaround100mbsfwhereitplateausuntil

∼

500 mbsfbelowwhichthereisagradualdecline.

The study interval (0 to

∼

61 mbsf), corresponds to Unit I (0–110 mbsf) of the logged lithostratigraphic units (a total of 6 units wereidentified)atSite U1467 (Betzleretal., 2017). It con-sists of an unlithified foraminifer-rich wackestone to packstone (veryfine tofinegrained) withacalcareous oozematrix(Betzler etal., 2017).As noted, theabundance ofaragoniteishighin the study intervalwithHMCpresentinthefirst10mbsf.Thespecies Ca2+ andSr2+ are presentin their lowest concentrations (mean Ca2+_:10.61mM,Sr2+_{111.42mM)whereas,Mg}2+_{isinitshighest}

concentration(meanMg2+_{:53.58mM)(Betzleretal.,2017).}

3.2. Traditionalwhole-testforaminiferalgeochemistryanddiagenesis

ranks

All geochemicalrecords(

δ

18_{OandMg/Ca)show clearcyclicity}

down-core (Figs.3–4). Based on the alignmentof the stable iso-tope stratigraphies, the planktonic

δ

18OG.ruber (w)record from

0–61.0mcdrepresentsthetimeinterval0–1800.0kyrcovering Ma-rineIsotope Stage(MIS)1-64(Figs.3a,4a-b). Thebenthicrecord, spanning 0–40.0 mcd, encompasses the interval 0 to 1179.0 kyr (Figs.3c,4a,c),MIS1-35.Thesedimentationratevariesdown-core between0.4–16.3cm/kyr(Fig.4d).

SEM observations reveal diagenetic alterations are primarily duetotheprecipitationofauthigeniccalcite(overgrowth)(Fig.5). Furthermore,thediageneticsusceptibilityofthefour investigated species,G.ruber (w),G.glutinata (w/b),P.obliquiloculata (w/c) and

C.mabahethi, differs down-core (Figs. 5–6). The resultant

Diage-nesis Rank (DR) values reflect a gradual decline, in the overall preservationstatefrom0to1800.0kyr(0–61.0mcd)(sample loca-tionsnotedonFigs.3,6).Whilethisdown-coredeclineisvariable acrossallspecies,allhavepristine,unalteredtestsinthecore-top samples(359-U1467A,B-Mudline)(Fig.5).Furtherdown-core, au-thigenicovergrowths occurbothon theinternalandexternaltest surfacesandwithinpores,resultinginareductioninpore-size un-til complete infilling is achieved (Fig. 5, Supplementary Material 1).Thesizeoftheseauthigenicovergrowthsvariesfromsmalland patchytolarge(>2 μm)well-developed,euhedralcrystalsgrowing perpendicular to the test wall andcovering the entire inner and outertestsurfaces(SupplementaryMaterial1).

Pulleniatinaobliquiloculata (w/c)appearstobethemost

suscep-tible speciesto authigenic overgrowth withmedium sized, well-developed crystalsalready observed at

∼

274.5 ka (11.6 mcd) on theexternal testsurfaces.In contrast,thesearefirst observedon

G.ruber (w)at

∼

790.0ka(28.7mcd),G.glutinata (w/b)at

∼

558.3

ka(22.2mcd) andthebenthictestsfurtherdown coreat

∼

807.9 ka (29.1 mcd). Large well-developed authigenic crystals are evi-dent on tests of G.ruber (w), G.glutinata (w/b), P.obliquiloculata

(w/c)and C.mabahethi from

∼

1190.9 ka(40.6 mcd),

∼

1015.1 ka (36.0 mcd),

∼

864.7 ka(31.3 mcd) and

∼

872.0 ka (31.7), respec-tively (SupplementaryMaterial 1). Inaddition tosize, differences inovergrowthdistributionsarenotedbetweentheplanktonicand benthicspecies. While all species show roughly similar develop-ment ofovergrowth crystals across their outer test surface, with concentrationsofovergrowthsinthepores,distinctdifferencesare observed ontheir inner test surfaces.For the planktonic species, authigenicovergrowthsare moreorlesshomogeneous acrossthe surfacewhereasforthebenthicspecies,theyaregenerally concen-tratedwithintheporesandalongthesutureswithpatchygrowth acrosstheinnersurfaces(SupplementaryMaterial1).

Intervals 1-3 were specifically selected, withreference to the

G.ruber (w) DR data, as representative segments with

differen-tial diagenetic influences (Fig. 6). Variability between the three planktonic

δ

18Orecordsdoesoccuracrosstheintervals(Fig.6). In-terval1,whichcoversMIS10-13(354.0–505.0kyr),hasthelowest DRvaluesforallspecies(0–3).Interval 2,whichcoversMIS21-26 (866.7–974.0 kyr), is located in the section with a notable in-creaseinthe respectiveDRvalues(2–4). Finally,Interval3covers MIS54-59(1573.8–1690.9 kyr)inthemostaltered portionofthe records withthe highest DRvalues (> 4). The glacial-interglacial cycles are more prominent across all records in Interval 1. Both

G.ruber (w) andP.obliquiloculata (w/c)display a large amplitude

changefromtheglacialpeakMIS12acrossthedeglaciationtothe interglacialpeak MIS11 (



∼

2.20



and2.50



,respectively).

Globigerinitaglutinata (w/b)showsa smallerchangeacrossMIS12

toMIS11(



1.41



),andlessprominentglacial-interglacialpeaks inInterval2comparedto G.ruber (w)andP.obliquiloculata (w/c).

Interval 3 shows notably higher isotopic valueswith the glacial-interglacialpeaksmarkedly reduced, although still clearly visible, forallspecies.Despitethevariationandextentofdiagenetic alter-ation, Pleistocene glacial-interglacialcycles can be readily identi-fiedinalllong-term

δ

18_{Orecords(Fig.6).However,beyond}

_∼

₁₃₀₀

ka (46.1 mcd), theG.ruber (w) data showsa notable down-core trendtowardshigher

δ

18Ovalues(Fig.6).

Fig. 3. Long-termgeochemicalrecordsfromIODPExpedition359SiteU1467againstcoredepthwith(a)theG.ruber (w)δ18_{O,(b)Mg/Carecordsand(c)theC.mabahethi} δ18_{Orecord.VisualSEMcheck(greycircles)andEPMA/SIMS(redandblackcircles,respectively)samplesareshownforreference.A3-ptmovingaveragesmoothingwas}

appliedtoourdatasets(thicklines)tohighlightthecycles.

3.3.Comparisonofbiogeniccalciteandauthigenicovergrowth

geochemistry

SIMS spot analyses (

∼

10 μm diameter) are similar to the G.

ruber (w)test wall thickness (

∼

9–12 μm)such that a large

pro-portion of the measurements partly overlapped onto epoxy. The location of these compromisedanalyses (n =35) was confirmed by SEM observation, and only those measurements which were located solely on the foraminifera test were included in the re-ported analyses (Fig. 7, Supplementary Material 3). Overall, 16 SIMS and 36 EPMAspot measurements from the F-0 (final), F-1 (penultimate)andF-2(antepenultimate)chambersweremeasured on seven samples for G.ruber (w), (Fig. 7, Supplementary Mate-rials2-3). An additional five EPMA spot analyses of large authi-genic overgrowths were measured (Fig. 7, Supplementary Mate-rial 2). Limited SIMS and EPMA sample sets were obtained due tothe limitationsof spotsize (for SIMS)andepoxy stability (for EPMA).Yet,thedatasetsarestilldeemedsufficienttoassess over-alltrendsandtheapplicationofthesemicro-analyticaltechniques inforaminiferaldiageneticstudies.Meanvaluesandstandard de-viationsof the measurements along with sample locations (with referenceonthe

δ

18OG.ruber (w)curve)areshowninFig.7.

Whole-test

δ

18_{OandMg/CavaluesarecomparabletoSIMSand}

EPMAspotmeasurements forcore-topSample 1at0ka (0mcd) andtoEPMAdataforSample2at409.0ka(17.4mcd).From Sam-ple3(=864.7 ka,31.3mcd)toSample 7(=1743.8ka,58.7mcd) thewhole-testG.ruber geochemicalvaluesareconsistentlyhigher than theSIMS (z =1.5-5.2; p =0.9-1)andEPMA (z =2.7-8;p = 1) data obtainedfrom individual chambers(Fig. 7a-b). Moreover, thedifferencevarieswiththehighestvaluesandlargestoffsetsin thedeepest,oldest sample.Interestingly,one F-1SIMSdatapoint forSample 7hitsubstantialdiageneticovergrowthsandhasa rel-atively high

δ

18_{O value of -1.24}

_

_{(Supplementary Material 3).}

TheEPMAovergrowthdata(Samples6and7)revealsaHMC com-position with 3–6 times higher Mg/Ca values than original test compositions.Samples3–7covertheintervalcorrespondingtoan observed changein theslope of thelong-termG.ruber (w)

δ

18_O

record with a down-coretrend towards highervalues. A marked increase inthe slope ofthelong-term G.ruber (w)Mg/Ca record is,however, alreadynotedprior to Sample 3 (Fig. 7d). Similar to the EPMAspotanalyses (Fig. 7b), the elementalmaps ofboth G.

ruber (w)andP.obliquiloculata (w/c)showclearMgenrichmentin

theauthigenicovergrowths (Fig.8). Bothspecieshavea LMCtest (Figs.8.a1,8.b1),whereasP.obliquiloculata (w/c)hasnaturalHMC growthbandsindicatedonFig.8(yellowarrows).Overgrowths,on

Fig. 4. (a)Thebenthicforaminiferaδ18_{Ostack(prob-stack)ofAhnet al.(2017) isusedtoobtaintheagemodelfor(b)IODPExpedition359SiteU1467G.ruber (w)and}

(c)C.mabahethiδ18_{Orecords.A3-ptmovingaveragesmoothingwasappliedtoourdatasets(b,c:thicklines)tohighlightthecycles.Sedimentationrates(Sed.Rate)are}

referencedascm/kyrin(d).MarineIsotopeStages(MIS)areshown(verticaldashedlines)andformthemaintie-pointsfortheagemodel(SupplementaryMaterial4).

boththeinnerandoutertestsurfaces,have4–10timeshigherMg intensitiesthanthebiogeniccalcite(Figs.8.a2,8.b2).

4. Discussion

4.1. Preservationstateoftheindividuallong-termgeochemicalrecords

The stable oxygen isotope composition (

δ

18O) of the original foraminiferal test isa function of the ambient seawater temper-ature and

δ

18_{Ocomposition during its initial precipitation}

(Pear-son, 2012). Similarly, biogenic test Mg/Ca ratios reflect ambient seawater temperatures (e.g. Raddatz et al., 2017; Rippert et al., 2016).Authigenicovergrowthisanticipatedtooccurnearthe sedi-ment/bottom-waterinterfaceanddown-core(e.g.Regenberget al., 2007). Assuch, its oxygenisotope composition iscontrolled by a varietyoffactorsincludingthermodynamicfractionation,the

δ

18_O

andpHoftheprecipitatingfluidaswellasmineralogyandkinetic effectsduringtheprecipitation(KimandO’Neil,1997;O’Neilet al., 1969; Swart,2015).The elemental ratiocompositionsare related tositemineralogyandporewatergeochemistry.

From 0 to

∼

200 kyr (0 to

∼

7.3 mcd), the cycles and ampli-tudes of IODPExpedition 359 Site U1467

δ

18OG.ruber (w) and

C.mabahethi records,are comparable tothe records produced by

Bunzel et al.(2017) (Fig. 6). Bunzel et al.(2017) compiled stable isotopicdatasets fromCoreSO236-052-4 retrievedfromthe Mal-divesInner Sea(03◦55.09N;73◦08.48E)atawaterdepthof382 m (Note: the slight amplitudinaldifference betweenthe benthic datasetsisdueto differencesinstudy sitedepths).Notably, Bun-zel et al. (2017) used a lower analytical resolution and a larger test size (250–350 μm) for their G.ruber (w) record, yet visu-allythetwodatasetsarenearly identical.Thisatteststothegood preservation state of Site U1467 samples for the first

∼

200 kyr (7.3 mcd). Furthermore, we can consider the G.ruber (w)

δ

18_O

andMg/Carecordsfrom0to

∼

627.4 kyr(0to

∼

24.7mcd) tobe pristine/near-pristine. This is reflected by the low DR values (0-2)andismanifestinthecoherence betweenthe SIMSandEPMA spot analyses data (Samples 1 and 2) and the whole-test geo-chemicaldatasets.Inparticular,EPMAdataforSample 2(Fig.7b) illustrates that at leastup until 409.0 ka (17.4 mcd) the whole-test Mg/Ca values closely reflect the original test geochemistry. While small, patchy HMC overgrowth is present on the inter-nal surfaces(DR = 1), it doesnot appear tobe sufficientto bias thewhole-test geochemistry. Calculated SSTsare also notable for being within the modern, seasonal seawater temperature range (Fig. 7d) (Krahmann and Krüger, 2018; Quadfasel, 2017; Reolid et al.,2017).

Fig. 5. ScanningElectronMicroscope(SEM)imagesof(a)G.ruber (w),(b)G.glutinata (w/b),(c)P.obliquiloculata (w/c)and(d)C.mabahethi/C.wuellerstorfi showingtheouter (1,2and4)andinnertestsurfaces(3,5).ImagesofpristineIODPExpedition359SiteU1467testsfromthecore-top(Sampleat0ka:A,B-Mudline)areshownin1-3and imagesofdiageneticallyalteredtests(Sampleat1743.8ka,58.7mcd:B-7H-3,75-76cm)areshownin4-5.Representativeexamplesofovergrowthsizeanddistributionare shownfortheinnertestsurfacesofG.ruber (w)in(e-h)(Samples:e.B-Mudline;f.B-3H-4,99-100cm;g.B-4H-5,36-37cm;h.C-6H-5,57-58cm).Whitebarsshowtest wallthicknessesandredarrowsindicaterepresentativeexamplesofovergrowthcrystals.Whole-testscalebars=100 μmandclose-upimagescalebars=20 μm.

Thedegreeofinfluenceoftheauthigenicovergrowthon whole-testgeochemicalrecordscanvarysignificantlyandisnot necessar-ilyconsistentacrossthevariousproxies(i.e.

δ

18OversusMg/Ca).A change inthe preservation state of the G.ruber (w) tests occurs between

∼

627.4 to 790.0 kyr (24.7–28.7 mcd) onwards (DR val-ues>2). Discrepanciesare alsorecorded by theSIMS and EPMA datafromindividualchamberswhichare significantlylower than respectivewhole-testgeochemicalvaluesfromSample 3(



∼

1.1



and

∼

1.5mmol/mol,respectively)toSample7(



∼

1.8



and

∼

4.1mmol/mol,respectively)(Fig.7a,b).Thereisanotable down-core increase in the slope of the

δ

18O G.ruber (w) record. This iscaptured by SIMS Samples3–5,which show a

∼

1



increase inthewhole-testvaluesandSample7,themostdiagenetically al-tered, which shows a

∼

2



increase. The compositional trends observedinthe

δ

18_{Odataarenotably amplifiedinthe long-term}

Mg/Ca record, with a steeper down-core slope increase. Consid-ering the original test geochemistry is usually < 6 mmol/mol, onlyalimitedamountofHMCovergrowth(x =

¯

12–22 mmol/mol) wouldbe required to bias the whole-test Mg/Ca towards signifi-cantlyhigher values(Fig. 7). This is reflected in EPMASample 3 (= 864.7 ka, 31.3 mcd), which has minimal overgrowth present onthe externaltest surface (external categoryDiagenesis Rank= 1),yetsmall-mediumwell-developed euhedralcrystalsacrossthe entireinner test surface (internalcategory Diagenesis Rank= 2). Thisovergrowth,althoughsmall,issufficienttoelevatethe

whole-test Mg/Ca values and increase the SST reconstruction by

∼

2◦C. Thisis furthersubstantiated by EPMASamples4,6and7,which have higher DR values dueto poorerpreservation states andan increased presence and/or size of overgrowth crystals. All have lower spotF-0 and F-1 chamber Mg/Ca values in comparison to thewhole-testdata,resultinginprogressivelylargerdiscrepancies inSSTestimatesof

∼

4–7◦C.

TheprecipitationofHMCovergrowths,asopposedtothemore stableLMC,could beafunctionofboththecompositionand con-tentofthebulksedimentaswellastheporewatergeochemistry (e.g. Panieri et al., 2017; Regenberg et al., 2007). At Site U1467, HMC is only present in the top 10 mbsf (Fig. 2). Where HMC in thesediment hasdissolved, Mg2+ _{isreleased andmay be}

in-corporated(elsewhere)incrystallineovergrowthsonbiogenic car-bonates.Aragonitedissolutionandactivesulphate reductionwere notedbelow50mbsfatthissitebyBetzleretal.(2017).The for-merhadbeennotedearlierinapteropodstudybySreevidyaet al. (2019) who reported intense dissolution of aragonitic pteropods beyond

∼

800 ka. As such, an increased saturation state of pore waters,withrespecttoMg2+_{,incombinationwithactivesulphate}

reduction may have favored HMC over LMC authigenic precipi-tates(Panieriet al.,2017;Swart,2015).Higherresolutionsediment mineralogytogether withpetrographyandporewaterprofilesare neededtoconcludeastotheexactprocessesbehindthedepth oc-currence andMg-content ofthe HMC overgrowths. However, the

Fig. 6. Long-termstableisotope(δ18_{O)recordsforIODPExpedition359SiteU1467for(a)theplanktonicspeciesG.ruber (w)withintervals1-3comparingG.glutinata (w/b)}

andP.obliquiloculata (w/c)records(Note:SelectedMarineIsotopeStages(MIS)areshownforreference);DiagenesisRank(DR)colormapsareshownfor(b)G.ruber (w),(c)

G.glutinata (w/b),(d)P.obliquiloculata (w/c)and(e)C.mabahethi with(f)thelong-termδ18_{OrecordforthebenthicspeciesC.mabahethi.Thesamplesusedforthevisual}

SEMcheckarealsoshownasgreycircleson(a).Redandtealcurvesshowδ18_{OG.ruber (w)andC.mabahethi data,respectivelyfromBunzelet al.(2017).A3-ptmoving}

averagesmoothingwasappliedtothedatasets(thicklines)tohighlightthecycles.

combined influence of HMC with naturally elevated

δ

18_{O values}

(Tarutani et al., 1969), with thermodynamic fraction (



∼

19◦C between the surface seawater and at 487 m depth) and related

fluidgeochemistry changeswouldfurtherjustifythe precipitation of authigenicovergrowths with elevated

δ

18Ovalues. These pro-cesses could thus explain the notable biases (stable isotopic and

Fig. 7. (a)SIMSand(b)EPMAspotdataversuswhole-testICP-MS/ICP-OESmeasurementsforG.ruber (w)withthelocationofthesamples(circleswithnumberID:black= SIMSandred=EPMAsamples)referencedon(c)thelong-termG.ruber (w)δ18_{Ocurve;(d)G.ruber (w)whole-testMg/Cadataand(e)theG.ruber (w)DiagenesisRank}

(DR)colormap.IODPExpedition359SiteU1467sampleIDs:(1)A,B-Mudline;(2)B-3H-3,9-10cm;(3)B-4H-5,36-37cm;(4)B-6H-2,75-76cm;(5)B-6H-3,75-76cm;(6) B-6H-4,75-76cm;(7)B-7H-3,75-76cm.Seasurfacetemperature(SST)calculationsarebasedontheG.ruber (w)equationofAnandet al.(2003).Redcurveshowsδ18_OG. ruber (w)datafromBunzelet al.(2017).A3-ptmovingaveragesmoothingwasappliedtoalldatasets(thicklines)tohighlightthecycles.Note:F-0(final),F-1(penultimate) andF-2(antepenultimate)chambers,WT=whole-testandOG=overgrowthmeasurements.WTdataina)andb)forA,B-MudlineSample1isfromStainbanket al.(2019).

elemental) on the planktonic whole-test geochemical data from SIMS/EPMA Sample 3 (= 864.7 ka, 31.3mcd) to 7 (= 1743.8 ka, 58.7mcd).

AlthoughnoSIMSorEPMAdatawascollectedforthebenthicC.

mabahethi species,achangein preservationstate (DR>2) occurs

between

∼

790.0 to 807.9 kyr (28.7–29.1mcd). As such, we can surmisethe recordsbeyondthisinterval to havesome degreeof diageneticoverprint(Figs.6e-f).Thisissupportedbyobservations oflarge,well-developedovergrowthontheinternaltestsurfacesat

872.0 ka(31.7 mcd).Eventhough minimal overgrowthispresent on the external surfaces, it is presumed that the large internal crystals would be sufficient to affect the whole-test geochemical records.However, althoughthe Mg/Cabiaseswouldbe similar to

that on the G.ruber (w) records, the

δ

18O influences should be

minorconsidering diageneticprecipitation isanticipated tooccur near the sediment/bottom water interface in close proximity to wheretheoriginalbiogeniccalciteprecipitated(Edgaret al.,2013). Thiswouldimplysimilar seawatertemperatures; however,

Fig. 8. Mg(intensity)EPMAmaps.ExamplesofIODPExpedition359HoleU1467B pristinespecimens(Mudline)areshownfor(a1)G.ruber (w)and(b1)P. obliquiloc-ulata (w/c)withdiageneticallyalteredexamplesshownfor(a2)G.ruber (w)(7H-3, 75-76cm)and(b2)P.obliquiloculata (w/c)(6H-4,75-76cm).T=test,OG= over-growth,EE=edgeeffect.Referencetosamplelocations(ID1,6and7)areshownon Fig.7.Notethehigh-Mg(yellowarrows)andlow-MgbandingintheP. obliquilocu-lata (w/c)tests.Scalebars=10 μm.

encesin pH,mineralogyandkinetic effectsduring theauthigenic overgrowth precipitation cannot be excluded. This hypothesis is supported by the prominentglacial-interglacial cycles withtheir amplitudes seeminglywell preservedforthe entiretyofthe ben-thic

δ

18Orecord(0-1179.0kyr)(Figs.3–4,6).

4.2. Diageneticsusceptibilityofindividualplanktonicspecies

We found distinct differencesin the preservation state of the three investigatedplanktonicspecies. Assuch, thediageneticbias on their respective whole-test geochemical compositions varies. This is illustrated in Intervals 1-3 in Fig. 6, with a variable preservation of the glacial-interglacial amplitudinal differences. Thewhole-testG.ruber (w),G.glutinata (w/b)andP.obliquiloculata

(w/c)

δ

18Orecordsall reflect clear,large amplitudecycleswithin Interval 1.However,relativetothethicker-walledP.obliquiloculata

(w/c), the former two speciesrecord dampened signalswith the peaksnotablylesspronouncedacrossInterval2.AcrossInterval3, inthemostdiageneticallyalteredportionoftherecord,alldatasets are notablyaffectedwithreducedglacial-interglacialmaximaand minimaandall

δ

18Orecordsconvergebecauseofincreasingwhole test

δ

18Ovalues.

Itmightseemcontradictorythatthespecies,whichappearsto be most diagenetically susceptible, would have the best whole-test geochemical preservation across Interval 2. However, as P.

obliquiloculata (w/c)hasthethickesttest(

∼

24 μm,Figs.5,9),the

rationaleisthatathickertestwouldequatetomoreofthe whole-testcompositionbeingweightedbythebiogeniccalcite.Therefore, in proportion to the primary signal, the authigenic overgrowth would contribute a smaller percentage and result in a smaller bias compared to G.ruber (w) andG.glutinata (w/b)which have significantly thinner tests (

∼

9–12 and

∼

6–10 μm, respectively;

Fig. 9. Interpretationsofdiageneticsusceptibilitiesforthewhole-testgeochemistry ofthethreeinvestigatedplanktonicspecies.InterpretationsarebasedonDiagenesis Rank(DR)values,spotEPMA/SIMSandwhole-testgeochemistrydatafromIODP Ex-pedition359SiteU1467.InterpretationsofdissolutionsusceptibilitiesafterKucera (2007).

Figs. 5, 9). Thisis in accordance withobservations by Bé (1977) andKucera(2007).Bé(1977) correlatedincreaseddiagenetic resis-tanceforspecieswithsmallerpores, thickerwallsandlargertest size,whileKucera(2007) reports anincrease indissolution resis-tanceforspecieswithalowerporosity(Fig.9).

5. Conclusions

Foraminiferalgeochemicalrecordsfromshallow,carbonate plat-formsareintegralforunderstandingglobalpaleoceanographicand -climatic trends but are subject to diagenetic influences. This is the case at IODP Expedition 359 Site U1467, which represents a shallow-intermediate water archive from the equatorial Indian Ocean of importance for South Asian Monsoon (SAM) climatic studiesandlatitudinalscale shiftsofoceancurrents.Prior to uti-lizingourhigh-analyticallyresolvedgeochemicalrecordsfor paleo-reconstructions,itisimportanttoconstraintheinfluenceof diage-neticalterationonboththepreservationoftheoriginal geochemi-calcompositionsandlong-termcyclicity.

The presence and influence ofauthigenic overgrowths on the insideand/or outsideof theforaminiferaltest aswell asthrough thepartial orfull recrystallization ofthe testwall has longbeen recognised. Importantly, this authigenic calcite is not ubiquitous or uniform in terms of presence, size, distribution or geochemi-cal composition andassuchshould be assessedon an individual studysitebasis.Forsimplicity,easeofuseandtofacilitate repro-ducibilityinother studies,thequalitativeDRdetailedandusedin thisstudy onlyconstrains thepresenceofauthigenicovergrowths onthetestsurfacesaswell aswithinthewallstructure.However, itisnotedthatboththeprimaryandauthigeniccalcitecanbe sus-ceptibletodissolution,whichwouldfurtheralterthegeochemical compositions.Thesedissolutionfeatures,whichcanincludeetched surfacesaswellaslattice-likecalcitestructures, aswell aspartial recrystallizationofthewalltexturearenotedinthedeepest,oldest studiedsamples.

The degree of influence of the authigenic overgrowth on the whole-testgeochemistryvariesacrosstheforaminiferalspecies(G.

ruber (w), G.glutinata (w/b), P.obliquiloculata (w/c),C.mabahethi)

andproxies (i.e.

δ

18Oversus Mg/Ca). Considering thesite miner-alogy,whole-test Mg/Cacompositionsare notably moreimpacted than

δ

18Orecords.However,takingintoconsiderationour individ-ualspotanalyses (SIMS andEPMA)andDR data,thetop

∼

627.4

kyroftheG.ruber (w)Mg/Caand

δ

18_{Orecordsarewellpreserved}

and largely reflect the primary geochemical compositions. Given

the setting ofovergrowth precipitation, the benthic C.mabahethi

δ

18_{Orecordfrom0–790.0kyr(DR=0–2),ifnottheentirebenthic}

record,isconsidered tohavepristine/near-pristinepreservationof the primary

δ

18O signal. As such, the top portions of both the

G. ruber (w) (0 to

∼

627.4 kyr; 0-24.7 mcd) and C.mabahethi (0

to

∼

790.0 kyr;0-28.7mcd) records(whole-testdata)aresuitable for paleoceanographic reconstructions of absolute seawater tem-perature,salinityand

δ

18_{Ovalues.Notwithstandingthedown-core}

diageneticinfluences,theglacial-interglacialcyclicity,while damp-ened,ismaintainedasattestedbythecoherenceofboththeG.

ru-ber (w)andC.mabahethi

δ

18_{OrecordswithProb-stack.Assuch,the}

completegeochemicalrecordscould stillbe usedfor cyclostratig-raphytoassesslong-termorbitalscaleproperties.

Overall, we show foraminiferal species (e.g. P.obliquiloculata)

with thicker tests are better suited for whole-test geochemical analyses, in diagenetically susceptible settings, as they have a higher proportionof biogenic calcite relative to authigenic over-growth. Moreover, while only a portion of our whole-test geo-chemicalrecordsreflecttheprimarycompositions,SIMSandEPMA spotmeasurements can be utilized to overcomebiases in diage-neticallyimpacted intervals. Consideringthe importance ofthese typesofrecordsfortropicaltemperaturestudies,itisvitalto em-ployan approach,whether itbe spotversus whole-test analyses, the utilization of specimens with thicker tests or a visual DR, which will allow the preservation state to be assessed to obtain moreaccurateandreliablepaleo-reconstructions.

Declarationofcompetinginterest

Theauthorsdeclarethattheyhavenoknowncompeting finan-cialinterestsorpersonalrelationshipsthatcouldhaveappearedto influencetheworkreportedinthispaper.

Acknowledgements

Funding: This work was funded by the Swiss National Sci-enceFoundation (SNSF) through grant number 200021_165852/1 awardedto SSp.DK acknowledgesfundingsupport fromthe Nat-ural Environment Research Council (NERC) through grant num-ber NERC-NE/N012739/1. Additionally, EDB was funded through theSwissNationalScienceFoundation projectQuantiCarb@Spring (projectnumber:154810).

The authorswould like to thankthe International Ocean Dis-covery Program (IODP) for supplying the samples used in this study.FromtheUniversityofFribourg,wethankAlexSalzmannfor hishelp withembeddingthe EPMAandSIMS samples,Christoph Neururerforhishelp withtheSEM imagingandSandraBorderie for her help with Fig. 1, which was drawn using the free soft-warepackageGMT(TheGenericMappingTools;WesselandSmith, 1991).MartinRobyristhankedforhisassistanceduringtheEPMA analysisat the University of Lausanne. FIERCEis financially sup-ported by the Wilhelm and Else Heraeus Foundation, which is gratefullyacknowledged.ThisisFIERCEcontributionNo.19.Lastly, weacknowledgetheconstructivecommentsoftwoanonymous re-viewersthathelpedtorefinethemanuscript.

Appendix A. Supplementarymaterial

Supplementary material related to this article can be found onlineathttps://doi.org/10.1016/j.epsl.2020.116390.The rawdata presentedinthisarticle canbe found at:https://doi.org/10.1594/ PANGAEA.914883,https://doi.org/10.1594/PANGAEA.914882.

References

Ahn, S., Khider, D., Lisiecki,L.E., Lawrence, C.E.,2017. A probabilisticPliocene– Pleistocenestackofbenthicδ18_{OusingaprofilehiddenMarkovmodel.Dyn.}

Stat.Clim.Syst. 2,1–16.https://doi.org/10.1093/climsys/dzx002.

Anand,P.,Elderfield,H.,Conte,M.H.,2003.CalibrationofMg/Cathermometryin planktonicforaminiferafromasedimenttraptimeseries.Paleoceanography 18, 1050.https://doi.org/10.1029/2002PA000846.

Barker,S., Greaves,M.,Elderfield,H.,2003. Astudyofcleaningproceduresused forforaminiferalMg/Capaleothermometry.Geochem.Geophys.Geosyst. 4,8407.

https://doi.org/10.1029/2003GC000559.

Bé, A.W.H., 1977. An ecological,zoogeographic and taxonomic review ofrecent planktonicforaminifera.In:Ramsay,A.T.S.(Ed.),OceanicMicropaleontology1. AcademicPress,London,pp. 1–100.

Bé,A.W.H.,1968.Shellporosityofrecentplanktonicforaminiferaasaclimaticindex. Science 80(161),881–884.

Betzler,C.,Eberli,G.P.,AlvarezZarikian,C.A.,theExpedition359Scientists,2017. MaldivesMonsoonand SeaLevel.In:ProceedingsoftheInternational Ocean DiscoveryProgram,359.CollegeStation,TX(InternationalOceanDiscovery Pro-gram).

Betzler, C., Eberli, G.P., Kroon, D., Wright, J.D., Swart, P.K., Nath, B.N., Alvarez-Zarikian,C.A.,Alonso-García,M.,Bialik,O.M.,Blättler,C.L.,Guo,J.A.,Haffen,S., Horozal, S., Inoue,M., Jovane,L.,Lanci, L.,Laya,J.C., LingHuiMee,A., Lüd-mann,T., Nakakuni, M., Niino, K.,Petruny,L.M., Pratiwi, S.D.,Reijmer,J.J.G., Reolid,J.,Slagle,A.L.,Sloss,C.R.,Su,X.,Yao,Z.,Young,J.R.,2016.Theabrupt onsetofthemodernSouthAsianMonsoonwinds.Sci.Rep. 6,29838.https:// doi.org/10.1038/srep29838.

Betzler,C.,Eberli,G.P.,Lüdmann,T.,Reolid,J.,Kroon,D.,Reijmer,J.J.G.,Swart,P.K., Wright,J.,Young,J.R.,Alvarez-Zarikian,C.,Alonso-García,M.,Bialik,O.M., Blät-tler,C.L.,Guo,J.A.,Haffen,S.,Horozal,S.,Inoue,M.,Jovane,L.,Lanci,L.,Laya, J.C.,HuiMee,A.L.,Nakakuni,M.,Nath,B.N.,Niino,K.,Petruny,L.M.,Pratiwi,S.D., Slagle,A.L.,Sloss,C.R.,Su,X.,Yao,Z.,2018.RefinementofMiocenesealeveland monsooneventsfromthesedimentaryarchiveoftheMaldives(IndianOcean). Prog.EarthPlanet.Sci. 5,5.https://doi.org/10.1186/s40645-018-0165-x. Birch, H., Coxall, H.K., Pearson, P.N., Kroon, D., O’Regan, M., 2013. Planktonic

foraminifera stable isotopes and water column structure: disentangling eco-logical signals. Mar. Micropaleontol. 101, 127–145. https://doi.org/10.1016/j. marmicro.2013.02.002.

Blättler, C.L.,Higgins, J.A.,Swart,P.K., 2019.Advectedglacial seawaterpreserved in the subsurface of the Maldives carbonate edifice. Geochim. Cosmochim. Acta 257,80–95.https://doi.org/10.1016/j.gca.2019.04.030.

Bunzel,D.,Schmiedl,G.,Lindhorst,S.,Mackensen,A.,Reolid,J.,Romahn,S.,Betzler, C.,2017.Amulti-proxyanalysisofLateQuaternaryoceanandclimatevariability fortheMaldives,InnerSea.Clim.Past 13,1791–1813.https://doi.org/10.5194/ cp-13-1791-2017.

Edgar,K.M.,Anagnostou,E.,Pearson,P.N.,Foster,G.L.,2015.Assessingtheimpact of diagenesison δ11_{B, δ}13_{C, δ}18_{O, Sr/Ca and B/Cavalues in fossilplanktic}

foraminiferalcalcite.Geochim.Cosmochim.Acta 166,189–209.https://doi.org/ 10.1016/j.gca.2015.06.018.

Edgar,K.M.,Pälike,H.,Wilson,P.A.,2013.Testingtheimpactofdiagenesisonthe

δ18_{Oand δ}13_{Cofbenthic foraminiferalcalcitefromasedimentburialdepth}

transectintheequatorialPacific.Paleoceanography 28,468–480.https://doi.org/ 10.1002/palo.20045.

Elderfield,H.,Yu,J.,Anand,P.,Kiefer,T.,Nyland,B.,2006.Calibrationsforbenthic foraminiferalMg/Capaleothermometryandthecarbonateionhypothesis.Earth Planet.Sci.Lett. 250,633–649.https://doi.org/10.1016/j.epsl.2006.07.041. GEBCO Compilation Group, 2019. GEBCO 2019 Grid. https://doi.org/10.5285/

836f016a-33be-6ddc-e053-6c86abc0788e.

Groeneveld, J., Nürnberg, D., Tiedemann,R., Reichart, G.J.,Steph, S., Reuning, L., Crudeli, D., Mason,P., 2008. Foraminiferal Mg/Ca increase inthe Caribbean during the Pliocene: Western Atlantic Warm Pool formation, salinity influ-ence,ordiageneticoverprint?Geochem.Geophys.Geosyst. 9,Q01P23.https:// doi.org/10.1029/2006GC001564.

Gupta,A.K.,Sarkar,S.,De,S.,Clemens,S.C.,Velu,A.,2010.Mid-brunhes strength-ening of the Indian Ocean dipole caused increased equatorial East African and decreased Australasian rainfall. Geophys. Res. Lett. 37, L06706. https:// doi.org/10.1029/2009GL042225.

Hastenrath,S.,Nicklis,A.,Greischar,L.,1993.Atmospheric-hydrosphericmechanisms ofclimateanomaliesinthewesternequatorialIndianOcean.J.Geophys.Res. 98, 20,219–20,235.https://doi.org/10.1029/2004JD004981.

Kim,S.-T.,O’Neil,J.R.,1997.Equilibriumandnonequilibriumoxygenisotopeeffects insyntheticcarbonates.Geochim.Cosmochim.Acta 61,3461–3475.https://doi. org/10.1016/S0016-7037(97)00169-5.

Krahmann,G.,Krüger,K.,2018.PhysicaloceanographyduringSONNEcruiseSO235. PANGAEA.https://doi.org/10.1594/PANGAEA.887805.

Kroon,D.,Ganssen,G.,1989.NorthernIndianOceanupwellingcellsandthe sta-ble isotope compositionoflivingplanktonic foraminifers.Deep-Sea Res. 36, 1219–1236.https://doi.org/10.1016/0198-0149(89)90102-7.

Kucera,M.,2007.Planktonicforaminiferaastracersofpastoceanicenvironments. In:Hillaire-Marcel,C.,deVernal,A.(Eds.),ProxiesinLateCenozoic paleoceanog-raphy.Elsevier,pp. 213–262.

Kunkelova,T.,Jung,S.J.A.,deLeau,E.S.,Odling,N.,Thomas,A.L.,Betzler,C.,Eberli, G.P.,Alvarez-Zarikian,C.A.,Alonso-García,M.,Bialik,O.M.,Blättler,C.L.,Guo,J.A., Haffen,S.,Horozal,S., LingHuiMee,A.,Inoue,M.,Jovane,L.,Lanci,L.,Laya, J.C.,Lüdmann,T.,Bejugam,N.N.,Nakakuni,M.,Niino,K.,Petruny,L.M.,Pratiwi, S.D.,Reijmer,J.J.G.,Reolid,J.,Slagle,A.L.,Sloss,C.R.,Su,X.,Swart,P.K.,Wright, J.D.,Yao,Z.,Young,J.R.,Lindhorst,S.,Stainbank,S.,Rueggeberg,A.,Spezzaferri, S., Carrasqueira,I.,Yu,S., Kroon,D.,2018.Atwomillionyearrecordof low-latitudeariditylinkedtocontinentalweatheringfromtheMaldives.Prog.Earth Planet.Sci. 5,86.https://doi.org/10.1186/s40645-018-0238-x.

Lisiecki,L.E.,Raymo, M.E.,2005. APliocene-Pleistocene stackof57globally dis-tributedbenthicδ18_{Orecords.Paleoceanography 20,PA1003.https://doi.org/10.}

1029/2004PA001071.

Lüdmann,T.,Kalvelage,C.,Betzler,C.,Fürstenau,J.,Hübscher,C.,2013.TheMaldives, agiant isolatedcarbonateplatformdominatedbybottomcurrents.Mar.Pet. Geol. 43.https://doi.org/10.1016/j.marpetgeo.2013.01.004.

O’Neil,J.R.,Clayton,R.N.,Mayeda,T.K.,1969.Oxygenisotopefractionationin diva-lentmetalcarbonates.J.Chem.Phys. 51,5547–5558.https://doi.org/10.1063/1. 1671982.

Paillard, D., Labeyrie,L., Yiou,P.,1996. Macintoshprogramperformstime-series analysis.Eos 77,379.https://doi.org/10.1029/96eo00259.

Panieri,G.,Lepland,A.,Whitehouse,M.J.,Wirth,R.,Raanes,M.P.,James,R.H.,Graves, C.A., Crémière,A., Schneider,A.,2017. Diagenetic Mg-calciteovergrowths on foraminiferaltestsinthevicinityofmethaneseeps.EarthPlanet.Sci.Lett. 458, 203–212.https://doi.org/10.1016/j.epsl.2016.10.024.

Pearson,P.N.,2012.Oxygenisotopesinforaminifera:overviewandhistoricalreview. In:Ivany,L.C.,Huber,B.T.(Eds.),ReconstructingEarth’sDeep-TimeClimate-The StateoftheArtin2012,PaleontologicalSocietyShortCourse,November3.The PaleontologicalSocietyPapers,pp. 1–38.

Quadfasel,D.,2017.PhysicaloceanographyduringRVSONNEcruiseSO127.17 De-cember1997to7January1998fromPortKlangtoMalé.PANGAEA.https:// doi.org/10.1594/PANGAEA.881513.

Raddatz,J.,Nürnberg,D.,Tiedemann,R., Rippert,N.,2017.Southeasternmarginal West Pacific Warm Pool sea-surface and thermocline dynamics during the Pleistocene(2.5–0.5Ma).Palaeogeogr.Palaeoclimatol.Palaeoecol. 471,144–156.

https://doi.org/10.1016/j.palaeo.2017.01.024.

Rebotim,A.,Voelker,A.H.L.,Jonkers,L.,Waniek,J.J.,Meggers,H.,Schiebel,R.,Fraile, I.,Schulz,M.,Kucera,M.,2017.Factorscontrollingthedepthhabitatof plank-tonicforaminiferainthesubtropicaleasternNorthAtlantic.Biogeosciences 14, 827–859.https://doi.org/10.5194/bg-14-827-2017.

Regenberg,M.,Nürnberg,D.,Schönfeld,J.,Reichart,G.-J.,2007.Earlydiagenetic over-printinCaribbeansedimentcoresanditseffectonthegeochemicalcomposition

ofplanktonicforaminifera.Biogeosciences 4,957–973.https://doi.org/10.5194/ bg-4-957-2007.

Reolid,J., Reolid,M., Betzler, C., Lindhorst,S., Wiesner, M.G., Lahajnar,N., 2017. UpperPleistocenecold-watercoralsfromtheInnerSeaoftheMaldives: taphon-omyandenvironment.Facies 63,8.https://doi.org/10.1007/s10347-016-0491-7. Rippert,N.,Nürnberg,D.,Raddatz,J.,Maier,E.,Hathorne,E.,Bijma,J.,Tiedemann, R.,2016.ConstrainingforaminiferalcalcificationdepthsinthewesternPacific warmpool.Mar.Micropaleontol. 128,14–27.https://doi.org/10.1016/j.marmicro. 2016.08.004.

Sadekov,A.Y.,Ganeshram,R.,Pichevin,L.,Berdin,R.,McClymont,E.,Elderfield,H., Tudhope,A.W.,2013.Palaeoclimatereconstructionsrevealastronglinkbetween ElNiño-SouthernOscillationandTropicalPacificmeanstate.Nat.Commun. 4, 2692.https://doi.org/10.1038/ncomms3692.

Schott,F.A.,Xie,S.-P.,McCrearyJr,J.P.,2009.IndianOceancirculationandclimate variability.Rev.Geophys. 47,RG1002.https://doi.org/10.1029/2007RG000245. Sreevidya,E.,Sijinkumar,A.V.,Nath,B.N.,2019.Aragonitepteropodabundanceand

preservationrecordsfromtheMaldives,equatorialIndianOcean:inferenceson pastoceaniccarbonatesaturationanddissolutionevents.Palaeogeogr. Palaeo-climatol.Palaeoecol. 534,109313.https://doi.org/10.1016/j.palaeo.2019.109313. Stainbank,S.,Spezzaferri,S.,Kroon,D.,deLeau,E.S.,Rüggeberg,A.,2018.The Plank-tonicforaminifera Globigerinoideseoconglobatus n.sp. ina glacial–interglacial context:IODP359SitesU1467and U1468.SwissJ.Geosci. 111(3),511–522.

https://doi.org/10.1007/s00015-018-0304-9.

Stainbank,S.,Kroon,D.,Rüggeberg,A.,Raddatz,J.,deLeau,E.S.,Zhang,M., Spezza-ferri,S.,2019.Controlsonplanktonicforaminiferaapparentcalcificationdepths for the northern equatorial Indian Ocean. PLoS ONE 14, e0222299.https:// doi.org/10.1371/journal.pone.0222299.

Steinhardt,J.,Cléroux,C.,DeNooijer,L.J.,Brummer,G.-J.,Zahn,R.,Ganssen,G., Re-ichart,G.-J.,2015.Reconcilingsingle-chamberMg/Cawithwhole-shellδ18_Oin

surfacetodeep-dwellingplanktonicforaminiferafromtheMozambique Chan-nel.Biogeosciences 12,2411–2429.https://doi.org/10.5194/bg-12-2411-2015. Swart,P.K.,2015.Thegeochemistryofcarbonatediagenesis:thepast,presentand

future.Sedimentology 62,1233–1304.https://doi.org/10.1111/sed.12205. Tarutani,T.,Clayton,R.N.,Mayeda,T.K.,1969.Theeffectofpolymorphismand

mag-nesiumsubstitutiononoxygenisotopefractionationbetweencalcium carbon-ateandwater.Geochim.Cosmochim.Acta 33,987–996.https://doi.org/10.1016/ 0016-7037(69)90108-2.

Wessel,P.,Smith,W.H.,1991.Freesoftwarehelpsmapanddisplaydata.EosTrans. AGU 72(41),441–446.