A lower to middle Eocene astrochronology for the Mentelle Basin (Australia) and its implications for the geologic time scale

HAL Id: insu-02317396

https://hal-insu.archives-ouvertes.fr/insu-02317396

Submitted on 30 Sep 2020

HAL is a multi-disciplinary open access

archive for the deposit and dissemination of

sci-entific research documents, whether they are

pub-lished or not. The documents may come from

teaching and research institutions in France or

abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est

destinée au dépôt et à la diffusion de documents

scientifiques de niveau recherche, publiés ou non,

émanant des établissements d’enseignement et de

recherche français ou étrangers, des laboratoires

publics ou privés.

A lower to middle Eocene astrochronology for the

Mentelle Basin (Australia) and its implications for the

geologic time scale

Maximilian Vahlenkamp, David de Vleeschouwer, Sietske Batenburg, Kirsty

Edgar, Emma Hanson, Mathieu Martinez, Heiko Pälike, Kenneth Macleod,

Yong-Xiang Li, Carl Richter, et al.

To cite this version:

Maximilian Vahlenkamp, David de Vleeschouwer, Sietske Batenburg, Kirsty Edgar, Emma Hanson, et

al.. A lower to middle Eocene astrochronology for the Mentelle Basin (Australia) and its implications

for the geologic time scale. Earth and Planetary Science Letters, Elsevier, 2020, 529, pp.115865.

�10.1016/j.epsl.2019.115865�. �insu-02317396�

Contents lists available atScienceDirect

Earth

and

Planetary

Science

Letters

www.elsevier.com/locate/epsl

A

lower

to

middle

Eocene

astrochronology

for

the

Mentelle

Basin

(Australia)

and

its

implications

for

the

geologic

time

scale

Maximilian Vahlenkamp

a

,

∗

,

David De Vleeschouwer

a

,

Sietske J. Batenburg

b

,

c

,

Kirsty M. Edgar

d

,

Emma Hanson

d

,

Mathieu Martinez

c

,

Heiko Pälike

a

,

Kenneth G. MacLeod

e

,

Yong-Xiang Li

f

,

Carl Richter

g

,

Kara Bogus

h

,

i

,

Richard

W. Hobbs

j

,

Brian

T. Huber

k

,

Expedition

369

Scientific

Participants

1

a_{MARUM–CenterforMarineandEnvironmentalSciences,UniversityofBremen,KlagenfurterStraße2-4,Bremen,28359,Germany} b_{DepartmentofEarthSciences,UniversityofOxford,SouthParksRoad,Oxford,OX13AN,UnitedKingdom}

c_{GéosciencesRennes,UMR6118,UniversitédeRennes1,35042Rennes,France}

d_{SchoolofGeography,EarthandEnvironmentalSciences,UniversityofBirmingham,AstonWebbBuildingBlockA,EdgbastonB152TT,UnitedKingdom} e_{DepartmentofGeologicalSciences,UniversityofMissouri,Columbia,101GeologicalSciencesBuilding,Columbia,MO65211,USA}

f_{SchoolofEarthSciencesandEngineering,NanjingUniversity,163XianlinAvenue,Nanjing210046,PRChina} g_{SchoolofGeosciences,UniversityofLouisianaatLafayette,611McKinleyStreet,Lafayette,LA70504,USA}

h_{InternationalOceanDiscoveryProgram,TexasA&MUniversity,1000DiscoveryDrive,CollegeStation,TX77845-9547,USA}

i_{CamborneSchoolofMines,CollegeofEngineering,MathematicsandPhysicalSciences,UniversityofExeter,Penryn,Cornwall,TR109FE,UnitedKingdom} j_{DepartmentofEarthSciences,UniversityofDurham,Durham,DH13LE,UnitedKingdom}

k_{NationalMuseumofNaturalHistory,SmithsonianInstitution,MRC:NHB121,10thandConstitutionAvenueNW,Washington,DC20560,USA}

a

r

t

i

c

l

e

i

n

f

o

a

b

s

t

r

a

c

t

Articlehistory:

Received31July2019

Receivedinrevisedform24September 2019

Accepted26September2019 Availableonline9October2019 Editor: I. Halevy Keywords: Eocene astrochronology astronomicalforcing IODP369 MentelleBasin

The geologic time scale for the Cenozoic Era has been notably improved over the last decades by virtue of integrated stratigraphy, combining high-resolution astrochronologies, biostratigraphyand magnetostratigraphy with high-precision radioisotopic dates. However, the middle Eocene remains a weak link.The so-called“Eocene timescale gap” reflects thescarcity ofsuitable studysections with clearastronomically-forcedvariationsincarbonatecontent,primarilybecauselarge partsoftheoceans werestarvedofcarbonateduringtheEocenegreenhouse.InternationalOceanDiscoveryProgram(IODP) Expedition369coredacarbonate-richsedimentarysequenceofEoceneageintheMentelleBasin(Site U1514,offshoresouthwest Australia).Thesequenceconsistsofnannofossilchalkand exhibitsrhythmic claycontentvariability.Here,weshowthatIODPSiteU1514allowsfortheextractionofanastronomical signal and the construction of an Eocene astrochronology, using 3-cm resolution X-Ray fluorescence (XRF)corescans. TheXRF-derivedratiobetweencalciumandironcontent(Ca/Fe)tracksthelithologic variability and serves as thebasis forour U1514astrochronology. We present a16 million-year-long (40-56 Ma) nearly continuous history of Eocene sedimentation with variations paced by eccentricity andobliquity.Wesupplementthehigh-resolutionXRFdatawithlow-resolutionbulkcarbonandoxygen isotopes, recording thelong-term coolingtrend fromthe Paleocene-EoceneThermal Maximum (PETM – ca. 56 Ma) into the middle Eocene (ca. 40 Ma). Our early Eocene astrochronology corroborates existing chronologies based on deep-sea sites and Italian land sections. For the middle Eocene, the sedimentologicalrecordatU1514providesasingle-sitegeochemicalbackboneandthusoffersafurther steptowardsafullyintegratedCenozoicgeologictimescaleatorbitalresolution.

©2019TheAuthor(s).PublishedbyElsevierB.V.ThisisanopenaccessarticleundertheCCBY-NC-ND license(http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction

Eocene climate is characterized by a long-term cooling trend aftertheextreme greenhouseconditionsof the earlyEocene,the

*

Correspondingauthor.

E-mailaddress:mvahlenkamp@marum.de(M. Vahlenkamp).

1 _{https://iodp.tamu.edu/scienceops/precruise/swaustralia/participants.html.}

warmestclimatestateoftheCenozoicEra.Duringthemiddleand late Eocene, globalclimateshifts in astepwise mannerfromthe early Cenozoicgreenhouse into an icehouse world (Zachos etal., 2008). This climate transition was accompanied by decreases in greenhouse gas concentrations,changes in precipitation patterns, carboncycleshifts,andanumberoftransientwarming eventsor “hyperthermals”pacedbyastronomicalforcing(e.g.Cramer etal., 2003; Lourens etal., 2005; Westerhold et al., 2007, 2008, 2017;

https://doi.org/10.1016/j.epsl.2019.115865

0012-821X/©2019TheAuthor(s).PublishedbyElsevierB.V.ThisisanopenaccessarticleundertheCCBY-NC-NDlicense (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Sexton etal., 2011; Turner etal., 2014; Anagnostou etal., 2016; Lauretano et al., 2016; Westerhold et al., 2017). The short-lived hyperthermaleventsare associatedwithnegative

δ

13Cexcursions interpreted to represent theinjection of isotopicallylight carbon intotheatmosphereandoceanandarelinked withlarge partsof the ocean floor beingthe starved ofcarbonate (e.g. Kennett and Stott,1991; Dickens etal., 1995; Zachos etal., 2005; Stap etal., 2010; Littleretal.,2014; Westerholdetal.,2017).

Constrainingtherateandpaceofpastclimatechangeis essen-tialto disentanglethenature anddrivers ofpastclimate change. Over the past decades, the analysis of astronomically-forced

cli-mate cycles recorded in sedimentary sections has advanced our

understanding ofcelestial mechanics andclimatedynamics. Sub-stantial progressin the understanding of celestial mechanics, i.e. thecalculationofastronomicalsolutions backintime(e.g.Berger andLoutre,1991; Laskaretal.,2004,2011;Zeebe,2017),andthe analysis of the expression of the astronomic imprint in marine stratigraphicrecordsledtotheestablishmentofastronomical tun-ing methods asa standard technique forthe Neogenetime scale resultinginNeogenechronostratigraphieswithonlyminor uncer-tainties (Gradstein et al., 2012; Hinnov and Hilgen, 2012). From

here, the astronomical time scale (ATS) was extended into the

Paleogene. An astronomical time scale for the Oligocene (Pälike et al., 2006) and across the Eocene-Oligocene Transition, with consistent results betweendifferent authors (Pälike et al., 2006; Westerhold et al., 2014), stretchesinto the youngest partof the middleEocene.

ForthemiddleEocene,however,alackofhighlyresolved sed-imentary records has led to an “Eocene astronomicaltime scale

gap”, which has hindered the establishment of a comprehensive

astronomicaltimescalefortheentireCenozoicandledto substan-tialdivergencebetweenexisting astrochronologiesforthe middle Eocene hampering their integration into the most recent fully-referencedGeologicTimeScale(GTS2012; Gradstein etal.,2012). GTS2012usesa6th-orderpolynomial fitforChronsC13rthrough

C23r to bridge the gap betweenthe astronomical agemodels of

thelate Eocene(Pälike etal., 2006) and theearly Eocene(Hilgen etal.,2010).

Subsequently,newandmoredetailedastronomicalcalibrations

ofthe middleEocenetime scale were conductedbasedon

deep-seasedimentcoresrecoveredbytheInternationalOceanDiscovery Program(IODP)andits predecessors.WesterholdandRöhl (2013) andWesterholdetal. (2015) usedtheeccentricitysignalin paleo-climaticrecordsfromOceanDrillingProgram(ODP)Site1260from theequatorialAtlanticandHole702BandSite1263intheSouth

Atlantic to develop a new astrochronology for the poorly

con-strainedintervalofthemiddleEocene(ca.40-48Ma).Boulilaetal. (2018) laterchallengedtheequatorialandsouthAtlantic chronolo-gies and presented a conflicting middle Eocene astrochronology, basedonsedimentsfromthewesternNorthAtlantic(SitesU1408 andU1410). These North Atlantic drift deposits, however, lack a strongeccentricityimprint.Instead,theirastrochronologyisbased onthe stable

∼

173-kyrobliquity amplitude modulationrecorded intheCa/Feratioofsediments.

Thisdifference inthe orbital imprintbetweendifferent strati-graphic sequences (eccentricity vs. obliquity) complicates the in-tegrationofresults.However, Dinarès-Turelletal. (2018) recently constrainedtheageofthebaseofChronC20n to

∼

43.5Ma. This resultisconsistentwiththeageofWesterholdetal. (2015) butin conflictwithBoulila etal. (2018), who placethiseventat

∼

44.4 Ma.Thus,magneticreversalagesandchrondurationsforthe mid-dleEocene“astronomicaltimescalegap”remaincontroversial.The earlyEocene,incontrast,nowfeaturesacoherentastrochronologic framework (Galeotti et al., 2017, 2019; Westerhold et al., 2017; Francesconeetal.,2018).Thisframeworkisbasedonastronomical

tuning of globally distributed oceandrilling sites aswell as land sectionstothe405-kyrcomponentineccentricity.

The tuning process relies on accurate astronomical solutions that provide the precise astronomical configuration of the Earth atspecific momentsin geologicalhistory.The state-of-the-art ec-centricity solutions (Laskaret al., 2011; Zeebe,2017) agree with each other back to

∼

50Ma. Prior to50Ma, 100-kyreccentricity cyclescannotbereliablyreconstructedwiththecurrent astronom-ical models. Moreover, the timing of eccentricity nodes (minima

in very long 2.4 Myr eccentricity) becomes ambiguous prior to

50 Ma,asthey differ betweensolutions.Nonetheless, cyclostrati-graphic techniques and astronomical tuning can still be instru-mentalintheearly EoceneandPaleoceneusingthe405-kyrlong eccentricity component as a tuning target (Lourens et al., 2005; Westerhold et al., 2007, 2008, 2017; Hilgen et al., 2010). The 405-kyreccentricitycycleisextremelycoherentbetweensolutions andstablefurther back intime, makingit the primemetronome forastrochronology insequences olderthan50 Ma(Laskaretal., 2004).

Inthisstudy,wepresentanastrochronologythatspansthe

in-terval fromthePaleoceneEoceneThermalMaximum(PETM,

∼

56

Ma)intothemiddleEocene(

∼

40Ma).Thestudiedinterval strad-dlestheaforementioned“50-Maboundary”forcingustoadopt dif-ferenttuningstrategiesintheearlyandmiddleEocene.Thisstudy uses the carbonate-rich Eocene sequences of recently recovered IODP Site U1514 in the Mentelle Basin(offshore southwest Aus-tralia)topursuethreemajorgoalsthatareaddressedinsuccessive steps:First,weprovideanearlytomiddleEoceneastrochronology for IODPSite U1514 in the MentelleBasin (southwest Australia). Thisastrochronologicframeworkisa valuableassetforfuture

pa-leoceanographic studies in the region. Second, we compare the

sedimentological record at U1514 to other middle Eocene astro-nomicallytunedrecordstoassesstheinconsistenciesbetween ex-isting chronologies for the middle Eocene. Finally, we place the

U1514 lower Eocene sedimentary sequence in the robust global

astrochronological framework. This result will allow for the as-sessment of the fit betweendifferent astronomicalsolutions and geological data. A good fit between a specific solution and geo-logicaldataindicates that thissolutionismore likelyanaccurate reflectionoftheevolutionofsolarsystempriorto50Ma,whenthe uncertainty in astronomical models increases drastically because of the chaotic behavior of our solar system (Laskar etal., 2004; Pälikeetal.,2004; Maetal.,2017).Inotherwords,wereversethe traditional approachandtry toconstrain astronomicalmodelsby extractingtheorbitalsignalfromgeologicaldata.

2. Materialandmethods

2.1. Material

Site U1514 was thenorthernmost (33◦7.2443S, 113◦5.4799E)

and deepest (3850 m water depth) site drilled during IODP

Ex-pedition 369in theMentelleBasin(Fig. 1A) (Huberetal., 2019). The high paleolatitude(

∼

60◦S in at50 Ma)of Site U1514 make thelithologicrecordasensitivemonitorofglobalclimaticchanges and deep-water circulation between midand highlatitudes. Site U1514recovered518.12mofCretaceoustoPleistocenesediments. The lowertomiddleEoceneintervalatSiteU1514 isrecordedby

∼

145 m (140–285 meter composite depth (mcd)) of

carbonate-bearing (12-35 wt%) sediments with well-developed lithological alternationsbetweenlightergreenishgrayclayeynannofossilchalk anddarkernannofossil-richclaystone(Fig.1E;Fig.2)(Huberetal., 2019).Theintervalbetween140mcdand196.06mcdispresentin HoleU1514Aonly.Below196.06mcd,thereisastratigraphic

over-lapbetweenHolesU1514AandU1514Candweusetheshipboard

Fig. 1. (A) LocationofSiteU1514andotherdeep-seasitesusedfortimescalereconstructionintheearlyandmiddleEocene(paleomap∼45Ma,www.odsn.de). (B) Downhole NRM20mTinclinations (after20mTdemagnetization)fromHolesU1514A(red)and U1514C(blue). (C) Magnetostratigraphicinterpretation(white=reversed,black=

normal,gray=undetermined). (D) Shipboard biostratigraphicevents(Huberetal.,2019). (E) Coreimagesstitchedtogether withIODPCoreImage(Druryet al.,2018).

(F) XRF-derivedCa/FeratioaccordingtotheshipboardsplicebetweenHolesU1514A(red)andU1514C(blue).EarlyEocenehyperthermalsfollowthenamingschemeof Crameretal. (2003). (G-H) Bulksedimentcarbonisotopeandoxygenisotopedata.(Forinterpretationofthecolorsinthefigure(s),thereaderisreferredtothewebversion ofthisarticle.)

fortheintervalbetween

∼

196and259.6mcd.Theintervaldeeper than253.82mcdispresentinHoleU1514Conly(Fig.1E).

2.2.Magnetostratigraphy

We used the natural remanent magnetization (NRM)

incli-nations measured on section-halfs after 20 mT demagnetization

treatment, i.e., NRM20mT, to define magnetic polarity (Huber et al., 2019). Paleomagnetic inclinations in the interval of 140-225 mcdaregenerallynoisyandarethusnotinterpretedinthisstudy (Fig. 1B). Towards the base of the section in Hole U1514C in-clination values become less scattered and easier to interpret.

The NRM20mT data allowed us to assign polarities for Chrons

C22r-C24r,aidedbytheshipboardbiostratigraphicconstraintsfrom thebaseofeachcore(Huberetal.,2019).Wefollowed the ship-boardmagnetostratigraphicinterpretationfrom245.87to284mcd. Fortheintervalbetween

∼

230and245.87mcd, wehaverevised theshipboardmagnetostratigraphicinterpretation.This

reinterpre-tationisnecessarytoensurethatthesplicebetweenHolesU1514A

andU1514C and themagnetostratigraphy ofthesetwo holes are

not in conflict. For example, the base of Chron C22r was offset

by

∼

8 m between Hole U1514A and U1514C according to the

shipboard interpretation. This offset is eliminated in our revised interpretation,aspresentedinFig.1.Chronboundarydepthswith differences to the shipboard magnetostratigraphic interpretation aregiveninSupplementaryTable3.

2.3. Biostratigraphy

Samples from Holes U1514A and U1514C core catchers and

selected samples from split-core sections were analyzed for cal-careous nannofossilsandplanktonic foraminifers.Calcareous nan-nofossil (T5 in Huber et al., 2019) and planktonic foraminiferal events(T6inHuberetal.,2019)provideafirst-order chronostrati-graphicframeworkandallowsformagnetostratigraphic interpreta-tions(Huberetal.,2019).

Fig. 2. Detailedrelationshipand cyclostratigraphicinterpretationoflithology,colorimetryand geochemistryat Site U1514intheinterval between147and171mcd. (A) WaveletanalysisoftheCa/Ferecordinthedepthdomain.(B)Coreimages(IODPCoreImage;Druryetal.,2018).(C)Sedimentlightness(L*).(D)XRF-derivedCa/Feratio. (E)CyclostratigraphicinterpretationoftheCa/Fedepth-series.(F)TunedXRF-derivedCa/FedataplottedagainsttheE+0.25TmixfromLa11/La04(Laskaretal.,2004,2011). LightyellowbandsandnumbersinC-Findicate405-kyreccentricitymaxima.SupplementaryFigs. 3-6elucidateourcyclostratigraphicinterpretationsfortheentirestudied interval.

2.4. X-rayfluorescence

ThebulkelementalcompositionofU1514sedimentswas mea-suredonthesplit-coresurfacesusingthe third-generation Avaat-echXRFcorescanner,locatedattheXRFCoreScanningFacilityin theGulfCoastRepositoryatTexasA&MUniversity.Measurements weretakenevery3cmatasourceenergyof10kV(nofilter,0.16

mA) anda 6-secondscount time foreach measurement. Element

intensities were obtained by processing raw X-ray spectra using the iterativeleast-square software(bAxil) package fromCanberra Eurisys.In thisstudy,we usetheratiobetweencalciumandiron counts(Ca/Fe)toextracttheimprintofastronomicalforcing from theEocene interval ofthe U1514 sedimentary archive.The Ca/Fe ratiotrackslithologicalvariability,withhigherCa/Feratiosin nan-nofossilchalksandlowerCa/Feratiosinnannofossil-richclaystone.

2.5. Bulksedimentcarbonandoxygenisotopes

Bulksediment samples of

∼

2 cm3 _{were dried in a low}

tem-peratureoven at50◦Covernightandthen groundusingan agate

mortarandpestletoformafinepowder.Thesamplewas

homog-enizedandthen

∼

150-250 μgwere usedforisotope analysis.All bulk sediment stablecarbonandoxygen isotope(

δ

13C and

δ

18O)

measurementsweremadeonaGVIsoprimeIRmassspectrometer

with GV multiflow preparation line in the School of Geography,

Earth and Environmental Sciences at the University of Birming-ham.ValuesarereportedrelativetotheViennaPeeDeeBelemnite (VPDB) standard. Stable isotopeanalyses havea typical analytical precisionof0.06

h

for

δ

13Cand0.10

h

for

δ

18Ovalues.Bulk sedi-mentstablecarbonandoxygenisotope(

δ

13_Cand

_δ

18_O)were

mea-suredinlowresolution,insufficienttocaptureallshiftsincarbon andoxygenisotopevaluesassociatedwithearlyEocene hyperther-mals, butthe high-resolution XRF-derivedCa/Ferecord allows us to observe the local sedimentary expression of hyperthermals in theMentelleBasin.

2.6. Timeseriesanalyses

Wavelettransforms ofdepth andtime-serieswere carriedout using the R-package “biwavelet” (Gouhier et al., 2016), which is

based on the original wavelet program written by Torrence and

Compo (1998). All other spectral analyses were carried out us-ingthemultitapermethod(MTM)withthree2

π

-tapers(Thomson, 1982) and conventionalAR1 rednoise modeling,asimplemented intheR-package “astrochron”(Meyers,2014).Depth-to-time

con-version was carried out using function “tune” from the same

R-package. With the function “bandpass” we employ

frequency-selective filters between 0.002-0.003 cycles/kyr (500-333 kyr) to

isolate and extract the component associated with 405-kyr

Table 1

ComparisonofChronreversalagesinmillionsofyears. Chron

boundary

CK95 GTS2004 GTS2012 New Willwood Atlantic Ypresian ATS Umbria-Marche U1514

CandeandKent (1995)

OggandSmith (2005)

Gradsteinetal. (2012)

TsukuiandClyde (2012) Westerholdetal. (2014,2015) Westerhold etal.(2017) Francescone etal.(2018) This study C22r/C23n.1n 50.778 50.730 51.209 51.030 51.051 50.777 50.767 50.573 C23n.1n/23n.1r 50.946 50.932 51.433 51.190 51.273 50.942 50.996 50.855 C23n.1r/C23n.2n 51.047 51.057 51.505 51.290 51.344 51.025 51.047 50.984 C23n.2n/C23r 51.743 51.901 51.884 51.960 51.721 51.737 51.724 51.751 C23r/C24n.1n 52.364 52.648 52.684 52.530 52.525 52.628 52.540 52.527 C24n.1n/C24n.1r 52.663 53.004 53.074 52.800 52.915 52.941 52.930 52.930 C24n.1r/C24n.2n 52.757 53.116 53.199 52.910 53.037 53.087 53.020 53.043 C24n.2n/C24n.2r 52.801 53.167 53.274 52.960 53.111 53.123 53.120 53.126 C24n.2r/C24n.3n 52.903 53.286 53.416 53.070 53.249 53.403 53.250 53.265 C24n.3n/C24r 53.347 53.808 53.983 53.570 53.806 53.899 53.900 53.699

andsolutionsanditsstatisticalsignificanceweusea nonparamet-ric methodby (Ebisuzaki, 1997). This method isimplemented in the function ‘surrogateCor’ (Baddouh et al., 2016). The data with

higher temporal resolution was resampled on the sample grid

ofthe data with lower temporal resolution,using piecewise lin-earinterpolation.Subsequently,thePearson correlationcoefficient

(

r

)

is calculated.Next, the ‘surrogateCor’ function carries out the samecorrelation analysesfor10,000 MonteCarlosimulations, us-ingphase-randomizedsurrogates.Thesurrogatesaresubjecttothe sameinterpolationprocessandcompensatefortheautocorrelation thatcharacterizebothtime-series.Todeterminea p-valueforthe correlation,thePearson correlation coefficientofthedata isthen compared to the distribution of correlation coefficients obtained fromthesurrogates.

3. Results

TheU1514Ca/Feratiorangesbetween1.5and31inthe stud-ied interval and co-varies withsediment brightness (L*) (Fig. 2).

In the lower Eocene, between

∼

218-282 mcd, Ca/Fe ratios

fluc-tuate around 30, indicating high carbonate content. From

∼

230 to 215mcd, the Ca/Fe ratioshifts to values ranging between10 and15(Fig. 1F).From

∼

240to 195mcd,the low-resolutionbulk sediment

δ

18Odata(Fig. 1H) exhibit the characteristic long-term EocenecoolingtrendfromtheearlyEoceneclimaticoptimuminto

the early middle Eocene (195-237 mcd), andfrom 195 mcd

up-wards,thetrendtowardshigher

δ

18_{Ovaluesceases.Superimposed}

on the long-termtrends of the geochemical records are a series ofabrupt excursions to lower Ca/Fe values,which are associated withintervalsofreducedwt%CaCO3content.These

sedimentolog-icalshiftsare characteristicofearlyEocenehyperthermals,which appearwellexpressedatSiteU1514(Fig.1F).

3.1.Cyclostratigraphicinterpretation

The lithologic variations between 150 and 284 mcd at Site

U1514consistofalternationsbetweenlightgreenish gray, nanno-fossil chalk and clay-rich, darker nannofossil chalk (Huber et al., 2019). Inotherwords,thelithologicalalternationsaremarked by quasi-cyclic variations in lightness and carbonate content. Ship-boardcolorimetricandXRF-derivedCa/Feratios closelytrackthis rhythm in lithological properties (Fig. 2). First-order quasi-cyclic variationsbetweenlighteranddarkersedimentsoccurtypicallyon sub-meter scale (Fig. 2). Very darkbands reoccur in intervalsof

about3 to 5 m. The shipboard age model provides an

approxi-matedurationof15.24Mafortheintervalbetween136.67mcd(T

Chiasmolithussolitus)and281.80mcd(T Fasciculithus spp.)(Huber etal., 2019). This initial time constraint translates to an average sedimentationrateof

∼

0.95cm/kyrthroughouttheentirestudied interval.Hence,thefirst-orderalternationsinlithology(

<

1m)are morelikelytocorrespondtothegeologicalexpressionofobliquity

or 100-kyreccentricity,while the rhythm ofdarker layers (

∼

3-5 m)maycorrespondtotheinfluenceof405-kyreccentricity.Longer cyclesof

∼

10-25mare alsoevident inCa/Feratioandmodulate thedarknessindarkerlayers,hencetheamplitudeofthe405-kyr eccentricitycycles,following ourcyclostratigraphic interpretation. We thus associatethe decameter-scale cycles withthe

∼

2.4-Myr eccentricitycycle. Minima inthe 2.4-Myr cycle, oftenreferred to asnodes, markthe strong expression of 405-kyreccentricity cy-cles.We employ these“nodes” asusefulpinpoints toanchor our floatingcyclostratigraphicframeworkinabsolutetime.

Inadditionto establishingtheconnectionbetweenlithological cyclesandtheastronomicalparametersthatdrivethem, astronom-icaltuningoftheU1514Ca/Ferecordtotheastronomicalsolution requires determination of the phase-relationship between eccen-tricityandtheCa/Feproxy.Wecalculatedthecorrelationbetween detrended log(Ca/Fe)(Lowess smoother; smoothing factor

=

0.1) andlow-resolution bulk

δ

18O. The Monte-Carlo basedcorrelation assessmentoftwoauto-correlated time-seriesindicatesapositive correlationcoefficientofr

=

0

.

186 (p-value0.03).This statistically-significant relationship suggeststhat carbonate-poor intervalsare

associated with warmer climatic conditions (assuming minimal

salinity or ice volume changes at this time and no great differ-ences between lithologies in the extent of diagenetic overprint-ing of original ratios). Warmer globalclimates are favored under high-eccentricityandhigh-obliquityconfigurations(e.g.Haysetal., 1976; DeVleeschouweretal.,2017; Vahlenkampetal.,2018).The abovereasoningimpliesthatminimainCa/FeatU1514shouldbe tunedintomaximaineccentricity(andobliquity).Thisrelationship agreeswiththelargestamplitudeofvariabilityinlithology,as in-dicated by the color reflectance andCa/Fe records (Fig. 2), likely corresponding to the largest amplitude of the orbital forcing pa-rameters.Astrongpositivecorrelation(r

=

0

.

302; p-value:

<

0.01) between

δ

18Oand

δ

13Calsoallows usto linkthecarbonatepoor intervalsduringwarmerclimaticconditionstotheinputof isotopi-callylightcarbonfromotherreservoirsintotheocean.Thisfinding isofparticularinterestasitallowsustoidentifyandglobally cor-relate Eocene hyperthermals over large distance, thus, providing an additional framework to projectour astrochronology onto ex-istingrecords. The astronomically-calibrated magnetostratigraphic ages presented in Table 1 are the result of tuning the inverted Ca/Fe record of U1514 to the eccentricitysolution La10c (Laskar etal.,2011).

3.2. AstronomicaltuningofU1514

WegeneratedtheU1514astrochronologictimescalebysplitting the record in two separate intervals: the middle Eocene interval

between

∼

142 and203 mcd andthe lower Eocene interval

be-tween

∼

203and284mcd.ForthetuningofthemiddleEocene in-terval,weusedthefulleccentricitysolution.Forthetuningofthe lowerEocenesequence,weinitiallysolelyusedthestable405-kyr

eccentricity cycles as a tuning target. Indeed, the 405-kyr com-ponent iscoherent inall available eccentricitysolutions (Supple-mentary Fig. 1,Laskaretal.,2011).Wethencalculatedcorrelation

coefficients betweenthe 405-kyr tuned log(Ca/Fe) atSite U1514

and different eccentricity solutions (La10a, La10b, La10c, La10d, La11). We found the strongest negative correlation

(

r

= −

0

.

178

)

withLa10c.Subsequently,werefinedthelowerEocenetuning us-ing thefull eccentricity La10csolution asa target.Tie points for boththeinitialandtherefinedtuningsare shownin Supplemen-taryFig. 2andareprovidedinSupplementaryTable 4.

3.2.1. EarlyEocene

WeusedtheC23n.2n/C23rchronboundary(245.87mcd)asan anchor for our early Eocene tuning. This choice is motivated by four elements: (1) the C23n.2n/C23r magnetic reversal is sharp and well expressed at Site U1514; (2) The interpretation of this magneticreversalis identicalinthe shipboardinterpretation and our revised interpretation; (3) The reversal occurs in a 405-kyr

eccentricity minimum atU1514, in the Atlantic and in the

Um-briaMarche sections; (4)The ageestimates of Westerhold et al. (2017) (51.737Ma)andFrancescone etal. (2018) (51.724 Ma)for thisreversalareidenticalwithinuncertainty.Wetunedthe expres-sion of the 405-kyr eccentricity minimum in Ca/Fe close to this

magnetostratigraphic boundary into the same 405-kyr minimum

asWesterholdetal. (2017) andFrancesconeetal. (2018).This ap-proachresultsinamagnetochronageof51.751MaatU1514.From this anchor point, we adopted a 405-kyr eccentricity tuning

ap-proach up- and downward to obtain an astrochronology for the

earlyEocene.Between204and284mcd,weidentified20405-kyr eccentricity cycles of 3-5 m thickness, which we tuned into cy-cles ecc120 to ecc139, referring to the number of 405-kyr cycles

backintime,startingfromthepresent.Unfortunately,wemissthe sedimentaryexpression ofcycleecc135becauseofarecoverygap

between

∼

266and 269 mcd. The age-depth model ofthe lower

Eocene using the full La10c eccentricity solution is constrained by 39tie-points.The sedimentation rateduring the earlyEocene

(203–284mcd) varied between

∼

0.5and2 cm/kyrandaveraged

around1 cm/kyr (Fig.3), butwasremarkablyconsistent between

∼

221and281 mcd.

3.2.2. MiddleEocene

Our tuning approach in the middle Eocene exploits the sta-bilityof the currentastronomicalsolutions over the past50 Ma. In particular, the timing of the nodes in the

∼

2.4-Myr compo-nentofeccentricityisconsistent betweenthedifferent astronom-ical solutions in the middle Eocene. These nodes are minima in the

∼

2.4-Myr eccentricity cycle and characterized by strong ex-pressionofthe405-kyreccentricitycycleandweakexpression of short100-kyreccentricity.Theyareparticularlywell-expressed in

the U1514 sedimentary record. Thus, they can be identified

vi-sually as long-lasting maxima in Ca/Fe that extend over several meters. In the wavelet transform, these intervals are marked by increasedpowerinthe 405-kyreccentricityband andlowpower inthe100-kyreccentricityband (Fig.4).Weusedboth indicators toidentifythestratigraphicpositionofnodesinthe

∼

2.4Myr ec-centricitycycle:

∼

150,

∼

170,

∼

180and

∼

202mcd.

BuildingonourtuningforthelowerEocene,wetiedthesefour nodesintothose oftheastronomicalsolution,beginning withthe nodecenteredat202mcdintotheoldeststableeccentricitynode

around ecc119 (47.8 Ma). We subsequently tuned the following

nodesat180,170and150mcdintotherespectivenodesatcycles

ecc112, ecc107 and ecc100 in the astronomical solution. We then

refinedtheagemodelbyfirstintroducingadditionaltiepointson the405-kyreccentricitylevelandthenonthe100-kyreccentricity levelbetween40and48.5Ma.According tothistuning, intervals ofextremeeccentricitycorrespondtothedepositionof

carbonate-poorsedimentsatSite U1514.OurmiddleEoceneagemodel con-sists of 59 age-depth ties in the interval between 142 and 203 mcd.The lithologicalalternationsinthe middleEoceneofU1514 also reflect an importantobliquity component, inparticular near nodesinthe

∼

2.4Myreccentricitycycle, whentheexpressionof 100-kyreccentricityisweak.Asaresultoftherecorded obliquity-eccentricity interactions we findthat the U1514 Ca/Fe signal re-sembles the astronomic signature of an eccentricity

+

(0.25*tilt) mix between40 and 48 Ma (Fig. 2). Sedimentationrates during themiddleEocenevariedbetween

∼

0.3and1.8 cm/kyrand aver-agedaround0.7cm/kyr.

4. Discussion

This work provides a detailed chronostratigraphic framework for future Eocene paleoclimatic studies at Site U1514. However, in thiswork, we also use Site U1514 to assess the

inconsisten-cies between existing middle Eocene chronologies and to place

the U1514 lower Eocene sedimentary sequence in a global

as-trochronologicalframework. Toachievethesegoals,anintegration ofourresultswithexistingchronologiesiscrucial. Magnetostratig-raphy providesindependenttemporalcontrolbetween234.73and 261.56mcdbutisofambiguousqualityoutsidethisinterval. Nev-ertheless, we can correlate the lithological variations at U1514

with sedimentological and geochemical data from the Atlantic

using chemostratigraphy. U1514 Ca/Fe data clearly exhibit early Eocene hyperthermals (Fig. 5) that we identified based on their stratigraphic position relative to magnetochron boundaries, their interrelation(bundlingandamplitude)andtheir characteristic oc-currence patterns (Figs. 5 and 6). The pattern of hyperthermals matchesquitewellbetweenU1514,theAtlanticdeep-seasitesand Italian land sections (Westerhold etal., 2017; Francescone et al., 2018). Using our integrated approach, we were able to identify

all hyperthermals between the PETM (55.93 Ma) and

hyperther-mal “W”(49.58 Ma). Ourchemostratigraphiccorrelation respects themagnetostratigraphicframeworkandfollowsthehyperthermal namingschemeofCrameretal. (2003) (Figs.1,5,6).Interestingly, compared to theAtlantic sites,theMentelle Basinhyperthermals between245and282 mcd(E2-L)havearelativelyweakexpression

whereasthosebetween225and245 mcd(M-W)havearelatively

strongexpression(Fig.5).

InthemiddleEocene,distinctfeatures,suchastheshapeofthe nodeat48Ma,distinctmaximainCa/FeatU151446.9Ma,aswell as theposition ofthe nodesandthe lowest Ca/Feallow a direct comparison of the U1514 Ca/Fe time-series to the proxyrecords oftheAtlanticsites(Fig.5).ThesimilaritybetweenSouthAtlantic shiftsinstablecarbonisotopesandthesedimentologicalandstable carbonisotoperecord intheIndianOcean hintsatthe global na-tureof theclimateperturbations onastronomicaltimescalesthat typifytheearlyandmiddleEocene(Fig.5).

Biostratigraphy didnot play a major role inthe lower Eocene inter-basin correlations described above because the shipboard biostratigraphiceventsbetween232and285mcdhave uncertain-tiesof

∼

1Myrandinternal inconsistencies ofup to

∼

2Myr(i.e. 20 m in the depth-domain) (Huber et al., 2019). However, this intervaliswell-constrainedbymagnetostratigraphyand/orthe hy-perthermalchemostratigraphy.

Thelastoccurrence(LO)ofbenthicforaminiferaStensioeina

bec-cariformis is found in U1514C-11R-CC. This position is in

accor-dancewiththemagnetostratigraphy(ChronC24r)andtherelative position of the PETM slightly higherin the sediment column. In contrast, the first occurrence (FO) of the planktonic foraminifera

Globanomalinaaustraliformis occursinU1514C-7R-CC,37.5mabove

the LO of S.beccariformis, yet globally the FO G.australiformis is

Fig. 3. EarlyEoceneastrochronologyforIODPSiteU1514.Weshowtheintervalbetween∼46.5and56 Ma.(A)InvertedCa/FedatafromSiteU1514inthedepthdomain,(B) waveletanalysisofU1514inthedepthdomain.(C)OrbitalsolutionsLa10a,La10b,La10c,La10dandLa11(Laskaretal.,2011).(D)Eccentricitybandpassfilteroftunedtime seriesofU1514(black)overtheastronomicalsolutionLa10c.(E)tunedCa/FedatafromSiteU1514(F)sedimentationrateinU1514,(G)waveletanalysisofthetunedrecord fromU1514.

thesporadicandrareoccurrenceof thisspeciesatSite U1514in theearlyEocenemakingtheFOartificiallyhigh.

The LO of Subbotina velascoensis and Morozovella aequa in

U1514C-7R-CCandU1514C-5R-CC, respectively, occur inC23n or

youngerinU1514.However,thepositionoftheseeventsiswithin C24raccordingtoGTS2012.

Nannofossil events in the early Eocene are more consistent withthemagnetostratigraphyand hyperthermalsthan planktonic foraminiferaeventsbutarestilloffsetfromGTS2012agesbyupto

1Myr.TheLOofTribrachiatusorthostylus inU1514A-28X-CCinour

tuningbetween49.87and50.84isinexcellentagreementwiththe assignedage of50.5Ma in the GTS2012. However, there ispoor agreementbetweentheU1514 ages fortheFOofBlackitesinflata

andtheLOofDiscoastersublodoensis whichappearolder(between

48.85 and 49.69 Ma) than in the GTS2012 (events at 47.84 and

46.21 Ma, resp.). Despite this discrepancy, we have good

confi-denceinourearlyEoceneagemodel,asitiswellconstrainedby the recognitionof hyperthermal W based on thestrong negative excursion in Ca/Fe ratios at 226.02 mcd. For the middle Eocene, thecalculatedageofallthreenannofossilevents(FOChiasmolithus

gigas, LO Chiasmolithus gigas and FO Reticulofenestraumbilica) at

Site U1514 are offset by at least 1 Myrs from the GTS2012. On

theotherhand,themiddleEoceneplanktonicforaminiferaldatum,

LOGlobigerinathekaindex occursatthe expectedageaccordingto

GTS2012.

Despite mismatches between the numerical ages of

biostrati-graphic eventsinourU1514 agemodelandGTS2012, we donot

doubt the magneto-, chemo-, and cyclostratigraphypresented in thisstudy.Indeed,thelatterthreestratigraphictechniquesare in-ternally andmutually consistent.The low resolution ofthe ship-boardbiostratigraphic data (increasingthe potential forsampling errorsto havelarge effects),low diversityplanktonicforaminifera

Fig. 4. MiddleEocenecyclostratigraphyforIODPSiteU1514.Weshowtheintervalbetween∼39.5and48.5 Ma.(A)InvertedCa/FedatafromSiteU1514inthedepthdomain, (B)waveletU1514inthedepthdomain,(C)EccentricitybandpassfilteroftunedtimeseriesofU1514(black)overorbitaleccentricity+0.25tilt(grey)fromLa11(eccentricity) andLa04(obliquity)solution(Laskaretal.,2004,2011).EccentricitybandpassfilteroftunedtimeseriesofU1514(black).Longeccentricity(405-kyr)cyclenumbersrelate tonodesinverylongeccentricity(∼2.4-Myr).(D)tunedCa/FedatafromSiteU1514(E)sedimentationrateinU1514.(F)Waveletanalysisoftheeccentricity+0.25Tiltmix fromLa11(eccentricity)andLa04(obliquity)and(G)waveletanalysisofthetunedrecordfromU1514.

assemblages, known and unrecognized diachroneity in the

bios-tratigraphic eventsbetweenbasins, and/or errors inthe GTS2012 ages(Gradsteinetal.,2012) likelyexplainexistingmismatches.The strongindependentagemodelatthissiteispromising forfuture high-resolutionbioeventcalibration.

Weassesstheuncertaintyofourastrochronologybycomparing

the ages of the early Eocene hyperthermals between our U1514

record, the Atlantic deep-sea sites (Westerhold et al., 2017) and theUmbria/Marche sections(Francesconeetal., 2018).We finda

maximumoffsetof

∼

100kyr athyperthermal M.This maximum

offsetisagoodmeasurefortheuncertaintyonourtuning.

4.1. ImplicationsfortheearlyEoceneGPTS

FiveAtlanticdeep-seasites(Sites1258, 1262,1263, 1265, and 1267)withdetailedcyclo-,bio-, chemo- andmagnetostratigraphy constitutethegeomagneticpolaritytimescale (GPTS)fortheearly Eocene(Westerholdetal.,2017).Francesconeetal. (2018) recently confirmedthisgeochronology onthelevelof405-kyreccentricity cyclesusingsectionsfromcentralItaly(Table1).

TheU1514 astrochronology presentedheresuggestsvery

sim-ilar durations for magnetochrons C23n.1n through C24n.3n

(Ta-ble2), as thosereported by Westerhold et al. (2017) for the At-lanticcomposite.Inconsistenciesdonotexceed200kyr.The com-bineddurationofChronsC23n.1n,C23n.1r.andC23n.2nis220kyr

longer at U1514 compared to the Atlantic composite. Moreover,

the age of hyperthermal O andof the C23n.2n/C23r reversal ac-cord well betweenSitesU1514 and1263. Ournumericalagesfor

hyperthermalsK(orETM3)toH1(orETM2)matchwiththe

tun-ing ofWesterhold etal. (2017). Therefore,weattributetheminor inconsistencies betweenthe duration estimates ofthe subchrons within C24ntosmalluncertaintiesinthegeomagnetic interpreta-tionsatU1514and1263.

Also, froma chemostratigraphic perspective, our early Eocene astrochronology agrees with previously published timescales. In-deed,agedifferencesforthePETMandotherearly Eocene hyper-thermalsbetweenU1514,ontheone hand,andWesterholdetal. (2017) or Francescone et al. (2018), onthe other hand,are gen-erally less than 100 kyr (Table 3). It should be noted that we inferredallhyperthermalsbasedontheircarbonate-poor sedimen-tology,exceptforthePETM,forwhichtheidentificationisdirectly basedonitsnegative

δ

13Cexcursion(Fig.1G).ThePETMoccursat 55.87 Ma atSite U1514, inexcellent agreement withtheage re-portedbyWesterholdetal.,2017.WethusconfirmthatthePETM is notlinked tohigheccentricityforcing,unlike theother Eocene hyperthermals.

Thetimingofthehyperthermalsindicatesthatminor inconsis-tencies in the age of magnetic reversals betweenour studyand Westerhold et al. (2017) could be explained by discrepancies in

Fig. 5. GlobalastrochronologicframeworkfortheearlytomiddleEocene.ComparisonoftheastrochronologyatU1514withSites1260(Ironcountsgray,precessionamplitude black;WesterholdandRöhl,2013),and1258(δ13_{Cbenthic;Sextonetal.,2011)fromtheequatorialAtlanticandSites1263(δ}13_{Cbulk(Westerholdetal.,2015),δ}13_Cbenthic

(Stapetal.,2010;Lauretanoetal.,2015,2016)),and1262(δ13Cbenthic;Stapetal.,2010; Littleretal.,2014)fromWalvisRidge.AllagemodelsfortheAtlanticsitesare accordingtoWesterholdetal. (2017).

Fig. 6. ChemostratigraphiccorrelationbetweenAtlanticSites1263,1267,1262fromWalvisRidgeandSiteU1514intheMentelleBasinfortheintervalbetweenhyperthermal “N”andhyperthermal“W”.Hyperthermalsarecharacterizedbydarkerbandsandincreasedironcountsatallsites.SplicedcoreimageswerecreatedusingIODPCoreImage (Druryetal.,2018).

themagnetostratigraphicinterpretation oflower Eocenedeep-sea sites(Fig. 3). The magnetostratigraphic framework of Westerhold et al. (2017) and Francescone et al. (2018) is based on multi-plesitesandlandsectionsandisthereforearguably morerobust thanoursingle-sitemagnetostratigraphicinterpretation ofChrons C22rthrough C24r.Furthermore,theseauthors agemodels result inmoregradualvariationsinmid-oceanridgespreadingrates.

4.2.TowardsaconsistentmiddleEoceneGPTS

ThemiddleEocenetimescale(betweenC18r-C21n)iscurrently characterizedbylargeuncertaintiesarisingfromlessconsistent

ra-dioisotopicdatingandastrochronologiesthantherestofthe Ceno-zoic.Aparticularobstaclefortheconstructionofacomplete Ceno-zoic astronomical timescale is the considerable discrepancy be-tweentworecentastronomicalagemodelsforthemiddleEocene (Westerholdetal.,2013,2015;Boulilaetal.,2018).

Site U1514 doesnot currentlyhave a detailed andconsistent

magneto- or biostratigraphic framework for the middle Eocene.

Nevertheless, U1514 yields the potential to evaluate the respec-tivequalityofcompetingchronologiesthathavebeenproposedas

the backbone for the middle Eocene timescale. The resemblance

oftheCa/FerecordatSite U1514withsedimentologicaland geo-chemical data from the Atlantic (Fig. 5 and Fig. 7) allows

com-Fig. 7. DetailedcomparisonoftheastrochronologyatU1514intheSEIndianOceanwithpublishedastrochronologiesfromtheAtlantic(Westerholdetal.,2013,2015,2017) inintervalsnotconstrainedbymagnetostratigraphyatU1514.(A)LowerEoceneintervalbetween48and51 Ma.NotethesimilaritiesinparticularbetweenU1514Ca/Feand 1258benthicδ13_{C.(B)MiddleEoceneintervalbetween44and46 Ma.TheredrectanglemarkstheintervalwhereBoulilaetal.,2018suggestedtwoinsteadofthree405-kyr}

cycles.NotethatthelowestvaluesinFecoincidewiththecoolestorbitalconfigurationandhighestFevalueswiththewarmestastronomicalconfigurationinbothU1514 and1263.(C)MiddleEoceneintervalbetween41and43 Ma.TheredsquaremarkstheintervalcharacterizedbyahiatusaccordingtoBoulilaetal. (2018).Notethatthe highestFevaluesinU1514and1260occurduringtheC19rEventat41.5Maandthesimilarityoftheshapeofindividual100-kyrcyclese.g.thecycle42.6-42.75Ma.

Table 2

ComparisonofChrondurationsinmillionsofyears.

Chron CK95 GTS 2004 GTS2012 New Willwood Atlantic Ypresian ATS Umbria-Marche U1514

CandeandKent (1995)

OggandSmith (2005)

Gradsteinet al. (2012)

TsukuiandClyde (2012) Westerholdet al. (2014,2015) Westerhold et al.(2017) Francescone et al.(2018) This study C23n.1n 0.168 0.202 0.207 0.160 0.222 0.165 0.229 0.282 C23n.1r 0.101 0.125 0.126 0.100 0.071 0.083 0.051 0.129 C23n.2n 0.696 0.844 0.872 0.670 0.377 0.712 0.677 0.767 C23r 0.621 0.747 0.787 0.570 0.804 0.891 0.816 0.776 C24n.1n 0.299 0.356 0.454 0.270 0.390 0.313 0.390 0.403 C24n.1r 0.094 0.112 0.125 0.110 0.122 0.146 0.090 0.113 C24n.2n 0.044 0.051 0.075 0.050 0.074 0.036 0.100 0.083 C24n.2r 0.102 0.119 0.142 0.110 0.138 0.280 0.130 0.139 C24n.3n 0.444 0.522 0.567 0.500 0.557 0.496 0.650 0.434 Table 3

Overviewofnamesandages(Ma)ofhyperthermaleventsidentifiedinthisstudy. Hyperthermal Ypresian ATS Umbria-Marche U1514

Westerholdetal. (2017) Francesconeetal. (2018) This study PETM 55.93±0.06 56.10 55.87±0.06 E1 55.65 – – E2 55.56 – – F 55.17 – 55.19 H1/ETM2 54.05 54.09 54.06 H2 53.95 – 53.96 I1 53.67 – 53.67 I2 53.55 – 53.55 J 53.26 – 53.22 K/ETM3 52.84 52.80 52.84 L 52.46 52.44 52.39 M 51.97 52.00 51.90 N 51.55 51.63 51.55 O 51.23 51.21 51.25 P 50.86 50.85 50.87 Q 50.76 50.76 50.73 R 50.67 50.62 50.60 S 50.48 50.40 50.47 T 50.37 50.29 50.28 U 49.95 49.96 49.95 V 49.68 49.67 49.63 W 49.58 49.53 49.51

parisononthe405-kyrandeven

∼

100-kyreccentricitylevel.The big advantage ofthe middle Eocene section atU1514 is the ex-cellent expression of the mix between obliquity andeccentricity andparticularly the clearly expressed nodes in

∼

2.4-Myr eccen-tricity(Fig.2).Theinterferencebetweeneccentricityandobliquity is quite characteristic, both in the astronomical solutions and in the sedimentological record (Fig. 2). This distinctness is particu-larlyevident duringnodes,whichincreasesourconfidenceinthe U1514astrochronology.

Boulilaet al. (2018) suggestedan alternativeinterpretation to theastrochronologyofWesterholdandRöhl (2013) andWesterhold etal. (2015).TheyreinterpretedthecyclostratigraphyatSite1263 andHole702BbyshorteningChronC20rbyone405-kyrcycleand proposed a 700-kyr-longhiatus in ChronsC19n andC19r at Site

1260. We, though, find a good correspondence between our

as-tronomicallytunedU1514 Ca/Feandthe

δ

13Ctime-series ofSites 1263 and 702 between cycle ecc107 to cycle ecc119. This

inter-basinal agreement between Sites U1514, 1263 and702 suggests

thatthe Westerholdetal. (2015) chronologyisnotcompromised by data gaps and misinterpretations (Fig. 7B). Similar variability

in proxy records of U1514 and 1260 in Chrons C19n-C19r

pro-videsadditionalsupportforthecompletenessofthestratigraphic record at 1260. Notwithstanding a lack of independent chronos-tratigraphic markers at U1514, the close correspondence of or-bitallypacesedimentary patternsbetweenSites U1514and1260 strongly suggeststhat stratigraphic problems with the North At-lanticSitesU1408andU1410lieatthebasisofthedivergence

be-tweenthecompetingmiddleEocenechronologies.Theproblemsat theNorthAtlanticSitesarelikelyduetothepresenceofhiatuses and slumps in these contourite drift deposits. Contourite drifts in general and the Newfoundland drifts in particular are associ-atedwitherosionalfeaturesandslumping(e.g.Shanmugam,2017; Van Peeretal., 2017). Furthermore,thesedrifts arecharacterized bysubstantialinter-holevariability,complicatingcyclostratigraphic work(VanPeeretal.,2017; Vahlenkampetal.,2018).

4.3. Constraintsforastronomicsolutions

The sedimentological record of U1514 supports previous con-clusionsbyWesterholdetal. (2017) suggestingthattheLa2010b/c eccentricity solutions in which the position of astronomical ob-jects is based on the INPOP08 ephemeris (Fienga et al., 2009; Laskar et al., 2011) aremost consistent with thegeological data

between50-56Ma. However,froman astronomer’spointofview

the La2010b/c solutions are considered less reliable than those

which use the more accurate ephemeris INPOP10 (Fienga et al.,

2011; Laskar et al., 2011). To evaluate the fit of the astronomi-cal solutionstogeological data,wecalculated Pearsoncorrelation coefficientsbetweentheU1514log(Ca/Fe)record andallavailable astronomicalsolutions,fortheintervalbetween56and50Ma.The log(Ca/Fe)recordcorrelatesbetterwithLa2010b

(

r

= −

0

.

173

)

and La2010c

(

r

= −

0

.

178

)

than withLa2010a

(

r

= −

0

.

154

)

, La2010d

(

r

= −

0

.

118

)

,La2011

(

r

= −

0

.

102

)

ortheZeebe (2017) eccentric-itysolutions(r rangesbetween

−

0.126and

−

0.154).

Weemphasizethatthesecorrelationcoefficientshavebeen cal-culated aftertuning to the 405-kyr eccentricity componentonly. The differences in correlation coefficient are therefore related to differences in the nature of the 100-kyr and 2.4-Myr eccentric-ity cycles in the different eccentricity solutions for thisinterval. Asexpected, thecorrelationbetweentheU1514log(Ca/Fe)record and the full La2010c eccentricity solution increases significantly

(

r

= −

0

.

302

)

after direct tuning. The correlation forthe interval

between56 and50 Ma remains weaker thanthe correlation

be-tween Ca/Fe anddifferent astronomicalsolutions forthe interval after50Ma (r

= −

0

.

370 forLa10c).Exploringthe entiresolution spaceforeccentricityin theearly Paleogenemayprovide further improvementorconstraintsforeccentricitysolutions intheearly Eocene.

4.4. AstronomicallyforcedpaleoenvironmentalchangeintheMentelle BasinduringtheEocene

The rapiddropin Ca/Fefromthe lowerto themiddle Eocene around49Maislinkedwithadecreaseinthesedimentationrate from

∼

1cm/kyrto

∼

0.7cm/kyr,exceptbetween47.8and49 Ma, when the sedimentationrate averages

∼

1.6 cm/kyr (Fig. 3). The linkbetweendecreasingCa/Feratiosandsedimentationrates indi-catesreducedcarbonateproductivityorshallowingofthe carbon-atecompensationdepth(CCD).However,thelatterseems

implau-sible,astheIndianOceanCCDdeepensacrosstheearlytomiddle Eocene transition (Van Andel, 1975). Hence, we explain the de-creaseinCaCO3fromthelowertothemiddleEoceneasadecrease

incarbonateproductivityassociatedwithcoolingsurfacewaters. TheU1514bulksediment

δ

18Orecordsupportscoolingof

sur-face waters in the Mentelle Basin from 50.8 Ma onwards,

con-currentwithSouthernHemispheredeep-watercoolingatSite 702 (KatzandMiller,1991).Thisisnotablyearlierthanthemajorityof other deep andsurface temperaturerecords,which show the in-ceptionofcooling at

∼

49Ma(Bijletal., 2009,2013; Hollisetal., 2012; Mudelseeetal.,2014; Inglisetal.,2015).Thus,localsurface temperaturesin theMentelleBasinappear tobe decoupled from globaloceantemperaturetrendsthroughoutthelong-termmiddle Eocenecooling(Fig.1H).ThecoolingatU1514ceasesat

∼

47Ma, muchearlierthantheglobalEocenedeep-seacooling.One possi-bilitytoexplaintherelativelyearlycessationofthecoolingsignal observedatSiteU1514istheinitiationoftheproto-Leeuwin Cur-rent during the middle Eocene (McGowran et al., 1997; Jackson etal.,2019) transportingwarmwaterfromlowlatitudesintothe MentelleBasin.

Anotherintriguingcorrelation isthat thetiming ofdecreasing carbonatedepositionintheMentelleBasinat49Macoincideswith the earliest estimates for a southern opening of the Tasmanian Gateway(i.e., southof theSouth Tasman Rise)(Bijl etal., 2013), whichwouldhavecausedregionaltoglobalscalereorganizationof ocean circulationthat certainlywouldhave affectedtheMentelle Basin.Thisworkwillactasthesolidgroundforfuturestudiesthat aimatunravelingtheprecisepaleoclimaticandpaleoceanographic historyofthebasin.

5. Conclusions

The U1514 sedimentary archive from the Mentelle Basin off-shore southwestAustraliaincludes a 16-million-yearlong history thatilluminatestheeffectsofastronomicalforcingthroughoutthe Eocene inthe region.We used thesedeep-sea sedimentsforthe construction of a detailed astrochronology for the early to mid-dle Eocene. The astronomical forcing signature is expressed in lithologicalvariability betweencarbonate-poor nannofossilchalks andnannofossil-rich claystone,andconsistsofamixedimprintof obliquityandeccentricity.Thestrongexpressionofminimainthe 2.4-Myreccentricitycycle(nodes)atSite U1514allowedfora ro-bustanchoringofourmiddleEocenecyclostratigraphicframework inabsolutetime.

Site U1514 manifests a comprehensive series of early Eocene hyperthermals, which we correlated with Atlantic deep-seasites andItalianlandsections.Inotherwords,U1514holdsasingle-site chemostratigraphicbackboneforthelowertomiddleEocene,well suitedforglobalcorrelation.

We apply Site U1514 asa thirdopinion in thediscussion

be-tween two competing timescales for the middle Eocene

(WesterholdandRöhl,2013; Westerholdetal.,2015; Boulilaetal., 2018). ThecomparisonbetweenIndian OceanSiteU1514 and At-lanticSites702,1260,1263suggestsnosignofhiatusesatthe lat-tersites.Hence,wecannotendorseBoulilaetal.’s(2018) hypothe-sestoexplainthedifferencesbetweenthecompetingtimescales.

Finally,SiteU1514demonstratesaregimeshiftincarbonate de-positionaround49Ma,withadecreaseincarbonateproductivity and a decrease in average sedimentationrates. Interestingly, the timing of this shift in depositional characteristics coincides with the oldest estimates (

∼

49-50 Ma) for the southern opening of the Tasmanian Gateway. Future paleoceanographic studies of the U1514 sedimentary archivewill further explore the possiblelink betweenoceangatewayopeningandoceanographicchangesinthe MentelleBasin.

Acknowledgements

This research used samples and data provided by the

Inter-national Ocean Discovery Program (IODP). M.V., DDV. and H.P.

were fundedby ERCConsolidatorGrant“EarthSequencing”(Grant

Agreement No. 617462). S.J.B. was funded by NERC UK-IODP

grant NE/R012350/1. K.M.E. andE.H. were funded by NERC grant

NE/R012490/1. M.M.was fundedby CNRS.K.B.acknowledges

sup-portfromtheNationalScienceFoundation(OCE–1326927).

Appendix A. Supplementarymaterial

Supplementarymaterialrelatedtothisarticlecanbefound on-lineathttps://doi .org /10 .1016 /j .epsl .2019 .115865.

References

Anagnostou,E.,John,E.H.,Edgar,K.M.,Foster,G.L.,Ridgwell,A.,Inglis,G.N.,Pancost, R.D., Lunt,D.J.,Pearson,P.N., 2016.Changingatmospheric CO2 concentration

wastheprimarydriverofearlyCenozoicclimate.Nature 533,380–384. Baddouh,M.B.,Meyers,S.R.,Carroll,A.R.,Beard,B.L.,Johnson,C.M.,2016.Lacustrine

87

Sr/86SrasatracertoreconstructMilankovitchforcingoftheEocene hydro-logiccycle.EarthPlanet.Sci.Lett. 448,62–68.

Berger,A.,Loutre,M.-F.,1991.Insolationvaluesfortheclimateofthelast10million years.Quat.Sci.Rev. 10,297–317.

Bijl,P.K.,Bendle,J.A.P.,Bohaty,S.M.,Pross,J.,Schouten,S.,Tauxe,L.,Stickley,C.E., McKay,R.M.,Röhl,U.,Olney,M.,Sluijs,A.,Escutia,C.,Brinkhuis,H.,Scientists, E.,2013. Eocenecoolinglinkedtoearlyflow acrosstheTasmanianGateway. Proc.Natl.Acad.Sci.USA 110,9645–9650.

Bijl,P.K.,Schouten,S.,Sluijs,A.,Reichart,G.-J.,Zachos,J.C.,Brinkhuis,H.,2009.Early palaeogenetemperatureevolutionofthesouthwestPacificOcean.Nature 461, 776–779.

Boulila,S.,Vahlenkamp,M.,DeVleeschouwer,D.,Laskar,J.,Yamamoto,Y.,Pälike,H., Turner,S.K.,Sexton,P.F.,Westerhold,T.,Röhl,U.,2018.Towardsarobustand consistentmiddleEoceneastronomicaltimescale.EarthPlanet.Sci.Lett. 486, 94–107.

Cande, S.C., Kent, D.V., 1995. Revised calibration of the geomagnetic polarity timescalefortheLateCretaceousandCenozoic.J.Geophys.Res.,SolidEarth 100 (B4),6093–6095.

Cramer,B.S.,Wright, J.D.,Kent,D.V.,Aubry,M.P.,2003.Orbital climateforcingof

δ13CexcursionsinthelatePaleocene–earlyEocene(chronsC24n–C25n). Paleo-ceanography 18.

De Vleeschouwer, D., Vahlenkamp,M., Crucifix, M., Pälike,H.,2017. Alternating Southern andNorthernHemisphere climateresponsetoastronomicalforcing duringthepast35m.y.Geology 45,375–378.

Dickens, G.R., O’Neil, J.R., Rea, D.K., Owen, R.M., 1995. Dissociation of oceanic methanehydrateasacauseofthecarbonisotopeexcursionattheendofthe Paleocene.Paleoceanogr.Paleoclimatol. 10,965–971.

Dinarès-Turell,J.,Martínez-Braceras,N.,Payros,A.,2018.High-resolutionintegrated cyclostratigraphy fromthe Oyambresection(Cantabria,NIberian Peninsula): constraintsfororbitaltuningandcorrelationofmiddleEoceneAtlantic Deep-SeaRecords.Geochem.Geophys.Geosyst. 19,787–806.

Drury,A.J.,Kotov,S.,Rochholz,F.,Wilkens,R.,Pälike,H.,2018.Uniform,Rapid,Open AccessDatabaseforShipboardIODP/ODP/DSDPImages.AGU,Washington,DC. Ebisuzaki,W.,1997.Amethodtoestimatethestatisticalsignificanceofacorrelation

whenthedataareseriallycorrelated.J.Climate 10,2147–2153.

Fienga,A.,Laskar,J.,Kuchynka,P.,LePoncin-Lafitte,C.,Manche,H.,Gastineau,M., 2009.GravitytestswithINPOPplanetaryephemerides.In:Proceedingsofthe InternationalAstronomicalUnion,vol.5,pp. 159–169.

Fienga,A.,Laskar,J.,Kuchynka,P.,Manche,H.,Desvignes,G.,Gastineau,M.,Cognard, I.,Theureau,G.,2011.TheINPOP10aplanetaryephemerisanditsapplicationsin fundamentalphysics.Celest.Mech.Dyn.Astron. 111(3),363.

Francescone,F.,Lauretano,V.,Bouligand,C.,Moretti,M.,Sabatino,N.,Schrader,C., Catanzariti,R., Hilgen,F.,Lanci,L.,Turtù, A.,2018.A9million-year-long as-trochronological recordof theearly–middle Eocene corroboratedbyseafloor spreadingrates.Geol.Soc.Am.Bull. 131,499–520.

Galeotti,S.,Moretti,M., Sabatino,N.,Sprovieri,M.,Ceccatelli,M.,Francescone,F., Lanci,L.,Lauretano,V.,Monechi,S.,2017.CyclochronologyoftheEarlyEocene carbon isotope record from a composite Contessa Road-Bottaccione section (Gubbio,centralItaly).Newsl.Stratigr. 50,231–244.

Galeotti,S.,Sprovieri,M.,Rio,D.,Moretti,M.,Francescone,F.,Sabatino,N.,Fornaciari, E.,Giusberti,L.,Lanci,L.,2019.StratigraphyofearlytomiddleEocene hyperther-malsfromPossagno(SouthernAlps,Italy)andcomparisonwithglobalcarbon isotoperecords.Palaeogeogr.Palaeoclimatol.Palaeoecol. 527,39–52.

Gouhier,T.,Grinstead,A.,Simko,V.,2016.biwavelet:conductunivariateand bivari-atewaveletanalyses(Version0.20.10).

Gradstein,F.M.,Ogg,J.G.,Schmitz,M.,Ogg,G.,2012.TheGeologicTimeScale2012. Elsevier.

Hays,J.,Imbrie,J.,Shackleton,N.,1976.VariationsintheEarth’sorbit:pacemaker oftheiceages.Science 194,1121–1132.

Hilgen,F.J.,Kuiper,K.F.,Lourens,L.J.,2010.Evaluationoftheastronomicaltimescale forthePaleoceneandearliestEocene.EarthPlanet.Sci.Lett. 300,139–151. Hinnov,L.,Hilgen,F.,2012.Cyclostratigraphyandastrochronology.In:TheGeologic

TimeScale.Elsevier,pp. 63–83.

Hollis,C.J.,Taylor,K.W.R.,Handley,L.,Pancost,R.D.,Huber,M.,Creech,J.B.,Hines, B.R.,Crouch,E.M.,Morgans,H.E.G.,Crampton,J.S.,Gibbs,S.,Pearson,P.N.,Zachos, J.C.,2012.EarlyPaleogenetemperaturehistoryoftheSouthwestPacificOcean: reconcilingproxiesandmodels.EarthPlanet.Sci.Lett. 349–350,53–66. Huber,B.T., Hobbs, R.W.,Bogus, K.A., Batenburg, S.J., Brumsack, H.-J.,Do Monte

Guerra,R.E.,Edgar,K.M.,Edvardsen,T.,Harry,D.L.,Hasegawa,T.,Haynes, S.J., Jiang,T.,Jones,M.M.,Kuroda,J.,Lee,E.Y.,Li,Y.-X.,MacLeod,K.G.,Maritati,A., Martinez,M.,O’Connor,L.K.,Petrizzo,M.R.,Quan,T.M.,Richter,C.,Riquier,L., Tagliaro,G.T.,GarciaTejada,M.L.,Wainman,C.C.,Watkins,D.K.,White,L.T., Wolf-gring,E.,Xu,Z.,2019.SiteU1514.In:Hobbs,R.W.,Huber,B.T.,Bogus,K.A.,the Expedition369Scientists(Eds.),AustraliaCretaceousClimateandTectonics. Pro-ceedingsoftheInternationalOceanDiscoveryProgram,vol. 369.CollegeStation, TX(InternationalOceanDiscoveryProgram).

Inglis,G.N., Farnsworth, A.,Lunt, D., Foster, G.L., Hollis,C.J., Pagani,M., Jardine, P.E.,Pearson, P.N.,Markwick,P., Galsworthy,A.M., 2015.Descent towardthe icehouse:EoceneseasurfacecoolinginferredfromGDGTdistributions. Paleo-ceanography 30,1000–1020.

Jackson,C.A.,Magee,C.,Hunt-Stewart,E.R.,2019.Cenozoiccontouritesinthe east-ernGreat AustralianBight, offshore southern Australia: implications for the onsetoftheLeeuwinCurrent.J.Sediment.Res. 89,199–206.

Katz,M.E.,Miller,K.G.,1991.EarlyPaleogenebenthicforaminiferalassemblageand stableisotopecompositioninthesouthernocean.In:Proc.OceanDrill.Program Sci.Results,vol. 114,pp. 481–516.

Kennett, J.P., Stott, L.D., 1991. Abrupt deep-sea warming, palaeoceanographic changesandbenthicextinctionsattheendofthePalaeocene.Nature 353,225. Laskar,J.,Fienga,A.,Gastineau,M.,Manche,H.,2011.La2010:aneworbitalsolution

forthelong-termmotionoftheEarth.Astron.Astrophys. 532,A89.

Laskar,J.,Robutel,P.,Joutel,F.,Gastineau,M.,Correia,A.,Levrard,B.,2004.A long-termnumericalsolutionfortheinsolationquantitiesoftheEarth.Astron. As-trophys. 428,261–285.

Lauretano,V., Hilgen,F., Zachos, J., Lourens, L., 2016. Astronomically tunedage modelfortheearlyEocenecarbonisotopeevents:anewhigh-resolutionδ13_C

benthicrecordofODPSite1263between∼49and∼54Ma.Newsl.Stratigr. 49, 383–400.

Lauretano,V.,Littler,K.,Polling,M.,Zachos,J.,Lourens,L.,2015.Frequency, mag-nitudeandcharacterofhyperthermaleventsattheonsetoftheEarlyEocene ClimaticOptimum.Clim.Past 11,1313–1324.

Littler, K., Röhl, U., Westerhold, T., Zachos, J.C., 2014. A high-resolution ben-thicstable-isotoperecordfortheSouthAtlantic:implicationsfororbital-scale changes in Late Paleocene–Early Eocene climate and carbon cycling. Earth Planet.Sci.Lett. 401,18–30.

Lourens,L.J.,Sluijs,A.,Kroon,D.,Zachos,J.C.,Thomas,E.,Röhl,U.,Bowles,J.,Raffi, I.,2005.AstronomicalpacingoflatePalaeocenetoearlyEoceneglobalwarming events.Nature 435,1083–1087.

Ma,C.,Meyers,S.R.,Sageman,B.B.,2017.Theoryofchaoticorbitalvariations con-firmedbyCretaceousgeologicalevidence.Nature 542,468.

McGowran,B.,Li,Q.,Cann, J.,Padley,D.,McKirdy,D.M.,Shafik,S., 1997. Biogeo-graphicimpact ofthe Leeuwin Currentin southern Australiasince the late middleEocene.Palaeogeogr.Palaeoclimatol.Palaeoecol. 136,19–40.

Meyers, S.R.,2014. astrochron: AnR Package for Astrochronology, http://cran .r-project.org/package=astrochron.

Mudelsee,M.,Bickert,T.,Lear,C.H.,Lohmann,G.,2014.Cenozoicclimatechanges: areviewbased ontimeseries analysisofmarinebenthic δ18_{Orecords.Rev.}

Geophys. 52,333–374.

Ogg,J.,Smith,A.,2005.Thegeomagneticpolaritytimescale.In:Gradstein,F.,Ogg,J., Smith,A.(Eds.),AGeologicTimeScale2004.CambridgeUniversityPress, Cam-bridge,pp. 63–86.

Pälike,H.,Laskar,J.,Shackleton,N.J.,2004.Geologicconstraintsonthechaotic dif-fusionofthesolarsystem.Geology 32,929–932.

Pälike,H.,Norris,R.D.,Herrle,J.O.,Wilson,P.A.,Coxall,H.K.,Lear,C.H.,Shackleton, N.J.,Tripati,A.K.,Wade,B.S.,2006.Theheartbeatoftheoligoceneclimate sys-tem.Science 314,1894–1898.

Sexton,P.F.,Norris,R.D.,Wilson,P.A.,Pälike,H.,Westerhold,T.,Röhl,U.,Bolton,C.T., Gibbs,S.,2011.Eoceneglobalwarmingeventsdrivenbyventilationofoceanic dissolvedorganiccarbon.Nature 471,349.

Shanmugam,G.,2017.Thecontouriteproblem.In:SedimentProvenance.Elsevier, pp. 183–254.

Stap, L., Lourens,L.J., Thomas, E.,Sluijs, A., Bohaty, S., Zachos, J.C., 2010. High-resolutiondeep-seacarbonandoxygenisotoperecordsofEoceneThermal Max-imum2andH2.Geology 38,607–610.

Thomson,D.J.,1982. Spectrum estimationand harmonic analysis.Proc. IEEE 70, 1055–1096.

Torrence,C.,Compo,G.P.,1998.Apracticalguidetowaveletanalysis.Bull.Am. Me-teorol.Soc. 79,61–78.

Tsukui, K.,Clyde,W.C.,2012. Fine-tuningthe calibration oftheearly tomiddle Eocene geomagneticpolarity timescale: paleomagnetismofradioisotopically datedtuffsfromLaramideforelandbasins.Bulletin 124(5–6),870–885. Turner,S.K.,Sexton,P.F.,Charles,C.D.,Norris,R.D.,2014.Persistenceofcarbon

re-leaseeventsthroughthepeakofearlyEoceneglobalwarmth.Nat.Geosci. 7, 748–751.

Vahlenkamp,M.,Niezgodzki,I.,DeVleeschouwer,D.,Bickert,T.,Harper,D.,Kirtland Turner,S.,Lohmann,G.,Sexton,P.,Zachos,J.,Pälike,H.,2018.Astronomically pacedchangesindeep-watercirculationinthewesternNorthAtlanticduring themiddleEocene.EarthPlanet.Sci.Lett. 484,329–340.

VanAndel,T.H.,1975.Mesozoic/Cenozoiccalcitecompensationdepthandtheglobal distributionofcalcareoussediments.EarthPlanet.Sci.Lett. 26,187–194. VanPeer, T.E.,Xuan,C.,Lippert,P.C., Liebrand,D.,Agnini,C., Wilson, P.A.,2017.

Extractingadetailedmagnetostratigraphyfromweaklymagnetized,Oligocene toearlyMiocenesedimentdriftsrecoveredatIODPSiteU1406(Newfoundland margin,northwestAtlanticOcean).Geochem.Geophys.Geosyst. 18,3910–3928. Westerhold,T.,Röhl,U.,2013.OrbitalpacingofEoceneclimateduringtheMiddle EoceneClimateOptimumandthechronC19revent:missinglinkfoundinthe tropicalwesternAtlantic.Geochem.Geophys.Geosyst. 14,4811–4825. Westerhold, T., Röhl,U., Frederichs,T., Agnini, C., Raffi, I., Zachos, J.C., Wilkens,

R.H.,2017.AstronomicalcalibrationoftheYpresiantimescale:implicationsfor seafloorspreadingratesandthechaotic behaviourofthesolarsystem.Clim. Past 13,1129–1152.

Westerhold,T.,Röhl,U.,Frederichs,T.,Bohaty,S.,Zachos,J.,2015.Astronomical cali-brationofthegeologicaltimescale:closingthemiddleEocenegap.Clim.Past 11, 1181.

Westerhold,T.,Röhl,U.,Laskar,J.,Raffi,I.,Bowles,J.,Lourens,L.J.,Zachos,J.C.,2007. Onthe duration ofmagnetochronsC24r and C25nand the timing ofearly Eoceneglobalwarmingevents:implicationsfromtheOceanDrillingProgram Leg208WalvisRidgedepthtransect.Paleoceanography 22.

Westerhold, T., Röhl,U., Pälike,H.,Wilkens, R., Wilson, P., Acton,G., 2014. Or-bitallytunedtimescaleandastronomicalforcinginthemiddleEocenetoearly Oligocene.Clim.Past 10,955–973.

Westerhold,T., Röhl,U.,Raffi, I., Fornaciari,E.,Monechi, S., Reale,V.,Bowles, J., Evans,H.F.,2008.AstronomicalcalibrationofthePaleocenetime.Palaeogeogr. Palaeoclimatol.Palaeoecol. 257,377–403.

Zachos,J.C.,Dickens,G.R.,Zeebe,R.E.,2008.AnearlyCenozoicperspectiveon green-housewarmingandcarbon-cycledynamics.Nature 451,279–283.

Zachos,J.C.,Röhl,U.,Schellenberg,S.A.,Sluijs,A.,Hodell,D.A.,Kelly,D.C.,Thomas, E.,Nicolo,M.,Raffi,I.,Lourens,L.J.,2005.Rapidacidificationoftheoceanduring thePaleocene-Eocenethermalmaximum.Science 308,1611–1615.

Zeebe,R.E.,2017.Numericalsolutionsforthe orbitalmotionoftheSolarSystem overthepast100Myr:limitsandnewresults.Astron.J. 154,193.