Earth and Planetary Science Letters 287 (2009) 412–419

Contents lists available at ScienceDirect

Earth and Planetary Science Letters

j ourna l homepage: www.e lsev ie r.com/ locate /eps l

Latest on the absolute age of the Paleocene–Eocene Thermal Maximum (PETM): Newinsights from exact stratigraphic position of key ash layers +19 and −17

Thomas Westerhold a,⁎, Ursula Röhl a, Heather K. McCarren b, James C. Zachos b

a MARUM — Center for Marine Environmental Sciences, University of Bremen, Leobener Strasse, 28359 Bremen, Germanyb Earth and Planetary Sciences, Univ. of California, Santa Cruz, CA 95064, USA

⁎ Corresponding author. Tel.: +49 421 21865672; faxE-mail address: twesterhold@marum.de (T. Westerh

0012-821X/$ – see front matter © 2009 Elsevier B.V. Adoi:10.1016/j.epsl.2009.08.027

a b s t r a c t

a r t i c l e i n f o

Article history:Received 27 February 2009Received in revised form 17 August 2009Accepted 18 August 2009Available online 18 September 2009

Editor: T.M. Harrison

Keywords:EoceneAr/Arash layerPETMGeomagnetic Polarity Time Scaleash −17

We constructed a precise early Eocene orbital cyclostratigraphy for DSDP Site 550 (Leg 80, Goban Spur, NorthAtlantic) utilizing precession related cycles as represented in a high-resolution X-ray fluorescence basedbarium core log. Based on counting of those cycles, we constrain the exact timing of two volcanic ash layers inSite 550 which correlate to ashes +19 and −17 of the Fur Formation in Denmark. The ashes, relative to theonset of the Paleocene–Eocene Thermal Maximum (PETM), are offset by 862 kyr and 672 kyr, respectively.When combined with published absolute ages for ash −17, the absolute age for the onset of the PETM isconsistent with astronomically calibrated ages. Using the current absolute age of 28.02 Ma for the Fish CanyonTuff (FCT) standard for calibrating the absolute age of ash −17 is consistent with tuning option 2 in theastronomically calibrated Paleocene time scale of Westerhold et al. (2008) [Westerhold, T., Röhl, U., Raffi, I.,Fornaciari, E., Monechi, S., Reale, V., Bowles, J., and Evans, H.F., 2008, Astronomical calibration of the Paleocenetime: Palaeogeography, Palaeoclimatology, Palaeoecology, v. 257, p. 377–403]. Using the recently recalibratedabsolute age of 28.201 Ma for the FCT standard is consistent with tuning option 3 in the astronomicallycalibrated Paleocene time scale. The new results do not support the existence of any additional 405-kyr cycle inthe early Paleocene astronomically tuned time scale.

© 2009 Elsevier B.V. All rights reserved.

1. Introduction

During the late Paleocene and early Eocene the opening of the NorthAtlantic was accompanied by enormous effusive and explosivevolcanism (Rea et al., 1990; Eldholm and Thomas, 1993) whichdeposited volcanic ash layers over much of northern Europe (KnoxandMorton, 1988). Single-crystal (sanidine) 40Ar/39Ar dates from earlyEocene ash layers +19 and −17 in the Fur Formation of Denmark(Bøggild, 1918) have played an important role in constructing theGeomagnetic Polarity Time Scale (GPTS) (Cande and Kent, 1992, 1995;Ogg and Smith, 2004) and for deciphering potential triggermechanismsof the transient global warming event known as the Paleocene–EoceneThermal Maximum (PETM) (Storey et al., 2007; Zachos et al., 2008).

Theageof the PETM, for lack of a coincident anddatable ash layer, hasto be derived by interpolating its position relative to the radiometricdated ash −17 within magnetochron C24r (for discussion see Wester-hold et al., 2007). Two factors are critical in this exercise: 1) the exactdistance between the PETM and ash−17, which is not known due to ahiatus in the reference record from DSDP Site 550 (Aubry et al., 1996),and 2) the accuracy of this radiometric age, which depends on thecorrect absolute age of the Fish Canyon Tuff (FCT, ~28.02 Ma) sanidine

: +49 421 2189865672.old).

ll rights reserved.

reference standard (Villeneuve, 2004). Several recent advances shouldhelp address these limitations and improve the absolute dating of thePETM. First, the age of the FCT standard has been recalibrated and theuncertainty of the method was reduced to 0.25% (Kuiper et al., 2008).Second, an astronomically calibrated floating time scale has beendeveloped for the Paleocene based on the identification of the stablelong eccentricity cycle (405-kyr). The tuning yields three options for theabsolute age of the PETM (Westerhold et al., 2008) one ofwhich (option3) would be consistent with the new calibrated age of the K/Pgboundary (Kuiper et al., 2008), but yields a relatively old age of 56.33 Mafor the PETM. This age, however, is in strong conflict with existingradioisotopic constraints. Furthermore, there seems to be a discrepancyin the number of 405-kyr eccentricity cycles in early Paleocene records(Hilgen et al., 2008; Kuiper et al., 2008).

Here we present a new cyclostratigraphy for Site 550 (Leg 80,Goban Spur) based on a high-resolution tracemetal records obtainedby X-ray fluorescence (XRF) core scanning to establish the exactrelative timing between the onset of the PETM and the prominent ashlayers +19 and −17. Both ashes are present in Site 550 and havebeen correlated to the Danish ash series of the Fur Formation (Knox,1984, 1985). Incorporating the latest radiometric dates for ash −17,we have estimated the absolute age range of the PETM to assess theeffect on the GPTS and test the astronomically calibrated Paleocenetime scale.

Fig. 1. Early Eocene (55.0 Ma) paleogeographic reconstruction showing the location ofDSDP Site 550 and ODP Site 1262. Reconstruction was made using Web-based softwareat http://www.odsn.de/odsn/services/paleomap/paleomap.html (Hay et al., 1999).

413T. Westerhold et al. / Earth and Planetary Science Letters 287 (2009) 412–419

2. Material and methods

We analyzed 50m of sediment cores retrieved at the NE AtlanticDSDP Site 550 (Fig. 1) using the non-destructive Super-slit X-ray

Fig. 2. Bulk carbon isotope data (Cramer et al. 2003, Raffi et al. 2005) and XRF intensity datablue, iron in red) of Cores 29X to 34X plotted against depth (mbsf). Transient global warminash +19 and−17 (gray dashed lines), and the unconformity (solid black line marks) are indafter Cramer et al. (2003), the nannofossil biostratigraphy after Raffi et al. (2005), and the pcolor in this figure legend, the reader is referred to the web version of this article.)

Fluorescence (XRF) Core Scanner equipped with an Oxford Neptune X-ray tube and a Canberra X-pips detector atMARUM, Bremen University,supported by the DFG-Leibniz Center for Surface Process and ClimateStudies at Potsdam University. The interval spanning Cores 550-29Xto -34X encompass the characteristic carbon isotope excursions (CIEs)of the early Eocene transient warming events PETM (Cramer et al.,2003) and Elmo (Lourens et al., 2005) (Fig. 2). High resolution latePaleocene to early Eocene records from the South Atlantic show pro-nounced evidence of Milancovitch related orbital cyclicity in calcium,iron, and barium which have been successfully utilized to construct adetailed cyclostratigraphy for magnetochron C24r including the PETM(Röhl et al., 2007; Westerhold et al., 2007). Similar variation in Ca, Fe,and Ba content should be present in sediments recovered at Site 550 aswell and therefore be useful for establishing a cyclostratigraphybetween the PETM and Elmo.

The AVAATECH XRF core scanner acquires bulk-sediment chemicaldata from split core surfaces. Although measured elemental intensitiesare predominantly proportional to concentration, they are also influ-enced by the energy level of the X-ray source, the count time, and thephysical properties of the sediment (Tjallingii et al., 2007). Calcium,

from DSDP Site 550 (zirconium in orange, titanium in green, barium in black, calcium ing events PETM and Elmo (blue bar), the main ash deposits (turquoise bar), positions oficated. The magnetostratigraphy and the lettering of carbon isotope transient decreasesosition of ash +19 and −17 after Knox (1984). (For interpretation of the references to

414 T. Westerhold et al. / Earth and Planetary Science Letters 287 (2009) 412–419

titanium, iron, zirconium, and barium data were collected every 2 cmdown-core over a 0.5 cm2 area. For the lighter elements calcium,titanium, and ironweused a generator settingof 10 kV, anX-ray currentof 0.2 mA, and a sampling time of 20s. For the heavier elementszirconium and barium we used a generator setting of 50 kV, an X-raycurrent of 1.0 mA, and a sampling time of 30s. The complete data setpresented is available online in the WDC-MARE PANGAEA database(http://doi.pangaea.de/10.1594/PANGAEA.726654).

3. Results

XRF elemental data for Site 550 exhibits variability that can bedirectly attributed to lithology (Fig. 2). The dominant lithology isnannofossil chalk with high carbonate contents of up to 70%(Shipboard Scientific Party, 1985) expressed as generally high Cavalues. Lower Ca values are present prior to the PETM reflectingsiliceous marls and dissolution intervals (Aubry et al., 1996) includingthe carbon isotope excursions designated F, G, H1 (equivalent of Elmoor ETM-2, Lourens et al., 2005), H2, and I1 in accord with the systemintroduced by Cramer et al. (2003). The Fe and Ca intensities are anti-correlated with lowest Ca values in the carbonate dissolution intervalfrom 411.7 to 409.8mbsf (Raffi et al., 2005). A stratigraphicunconformity (Shipboard Scientific Party, 1985; Aubry et al., 1996)at 407.38mbsf is characterized by a transition from green to redsediments, but with no change in Fe abundance. The Ba intensity datashow extremely well developed cycles especially in Cores 30X to 33X(365–401mbsf) which encompass both ash layers +19 (392.99mbsf)and−17 (399.76 mbsf). Ti values are generally low except for the ashrich horizons in Cores 32X and 33X (384 to 401 mbsf). Apart fromseveral significant Zr peaks in a few distinct ash layers the acquired Zrdata are below detection limit (Figs. 2 and 3).

Fig. 3. Close up of XRF data in the bentonite rich interval at Site 550 against depth (mbsf). ZPosition of ash +19 and −17 are marked by blue bars. (For interpretation of the reference

4. Ash layer identification

A series of over forty ash layers appearing as thin bentonites havebeen identified in Hole 550 (Shipboard Scientific Party, 1985). Theirchemical characteristics and timingmatchwithwidespreadash series innorthwestern Europe, especially the North Sea Basin, in the latestPaleocene and earliest Eocene (Nannofossil-Zone NP9/NP10) (Knox,1984, 1985; Müller, 1985). The tuffaceous intervals in the North Seabasin are contemporaneous with the 2nd phase of the North AtlanticIgneous Province (NAIP) volcanism from 53 to 56 Ma (Saunders et al.,1996) and diagnostic part of the Sele and Balder Formation in drill coresfrom the North Sea (Knox, 1996). Around 180 individual fairly wellpreserved ash layers are exposed in coastal cliffs around Limfjorden(NWDenmark) in the Fur Formation (Larsen et al., 2003) and are iden-tified as fall-out pyroclastics in water (Haaland et al., 2000). The suc-cession has been divided into two visually distinctive parts: the lowerpart with the mostly light colored ash layers of the ‘negative’ series(−39 to−1) and the upper partwith the black basaltic ash layers of the‘positive’ series (+1 to +140) (Bøggild, 1918; Larsen et al., 2003). Thetransition from negative to positive series represents the switch fromeffusive to explosive volcanism (Larsen et al., 2003). Because both ash+19 and ash −17 have a distinctive feldspar composition it waspossible to also identify them in ashes recovered at DSDP Site 550 at392.99 and 399.76mbsf, respectively (Knox, 1984, 1985).

Valuable information with respect to the ash layers can be obtainedfrom analysis of bulk sediment looking at the distribution of immobileelements such as Ti and Zr (Haaland et al., 2000; Larsen et al., 2003).Apart from two rhyolitic layers (+13 and +19) the positive seriesoriginate from evolved Fe–Ti rich tholeiitic basalts (Larsen et al., 2003)and can be identified in the XRF data by their characteristic high Fe, Tiand Zr peaks as well as low Ca peaks (Figs. 2 and 3). The basaltic ashlayers in the XRF record from 385 to 395mbsf represent themain phase

irconium in orange, titanium in green, barium in black, calcium in blue, and iron in red.s to color in this figure legend, the reader is referred to the web version of this article.)

415T. Westerhold et al. / Earth and Planetary Science Letters 287 (2009) 412–419

(Phase 2b of Knox and Morton, 1983) of the ash deposits at Site 550.Within this phase the rhyolitic peralkaline ash +19 is particularlydistinctive because it lies outside the compositional field of basalt in thepositive series (Larsen et al., 2003). In the XRF data this particular ashlayer shows a small peak in Ti and low Fe, Ba, and Zr intensities (Fig. 3)being consistent with the geochemical characterization of Larsen et al.(2003). Based on calculations of the total ash volume ejected ash+19 isthought to be one of the largest basaltic pyroclastic eruptions ingeological history (Egger and Brückl, 2006). Most prominent is thephonolitic to trachytic ash −17 of the negative ash series (Larsenet al., 2003) because it contains crystals of sanidine suitable for highprecision radiometric dating (Storey et al., 2007). The low Ti and highZr composition points to an alkalinemagma type (Larsen et al., 2003).The Danish ash −17 is isochronous with the Skraenterne Formation

Fig. 4. Evolutionary wavelet power spectra of Ba intensity data from DSDP Site 550 against dein the evolutionary wavelet power spectra are normalized linear variances with blue represelines enclose regions of greater than 95% confidence. Cross-hatched regions on either endbands that run across the spectra, which indicate the dominance of precession. The propprecession (21 kyr) are indicated by arrows. Note: the age model in (b) is based on linear i(Westerhold et al., 2007) between both events. (For interpretation of the references to colo

tuff in East Greenland (Storey et al., 2007) and has been proposed tooriginate in the East Greenland Gardiner Complex (Heister et al.,2001; Storey et al., 2007). Ash−17 can be easily identified in the XRFdata from Site 550 due to the prominent Zr peak at 399.76 mbsf(Figs. 2 and 3). A high Ba peak, a small Ti peak aswell as lows in Ca andFe intensities are consistent with geochemical analysis of ash −17(Larsen et al., 2003).

5. Cyclostratigraphy

Given the sedimentation rates, it is clear that theprimary oscillationsin Ba intensity are related to precession cyclicity. Blackman–Tukey andevolutionary wavelet analysis methods were applied using AnalySeries(Paillard et al., 1996) and the wavelet software of Torrence and Compo

pth (a) and relative age with respect to the onset of the PETM (b). The shaded contoursnting low spectral power, and red representing high spectral power. The black contourindicate the cone of influence where edge effects become important. Note the distinctosed position of the long (405 kyr) and short (100 kyr) eccentricity cycle as well asnterpolation between the PETM and Elmo layer given the relative distance of ~1.8 Myrr in this figure legend, the reader is referred to the web version of this article.)

416 T. Westerhold et al. / Earth and Planetary Science Letters 287 (2009) 412–419

(1998) (http://paos.colorado.edu/research/wavelets), respectively. Weuse the well developed cycles in the Ba intensity record of Site 550 todevelop a cyclostratigraphy. Ba is positively correlatedwith Ca and anti-correlated with Fe and Ti. This suggests that Ba in Site 550 sediments isnot of terrigenous origin. Therefore, Ba variations in Site 550 areinterpreted to predominately reflect changes in marine paleoproduc-tivity (Paytan and Griffith, 2007) and thus the cyclic response of thecarbon cycle to climate forcing. Spectral analysis and evolutionarywavelets of Ba intensities without ash layers reveal a distinct peak at0.65 m (Fig. 4a). Age constraints for the distance PETMrecovery to Elmo(Westerhold et al. 2007) imply that the 0.65 m cycle has a mean periodof ~25 kyr which corresponds to earth's orbital precession cycle(Fig. 4b). The absence of a strong shift in the period-band related toprecession (21 kyr, Fig. 4) suggests that stable and uniform sedimen-tation rates prevailed from the unconformity to the Elmo event. AGaussian filter was used to extract the dominant cyclicity (Fig. 5). Wethen assigned theprecession cycle numbers as given inWesterhold et al.(2007) to the extracted cycles using the revised nannofossil biostratig-raphy of Site 550 (Raffi et al., 2005) and Site 1262 (Agnini et al., 2007).This way a correlation of cycles between Site 550 andWalvis Ridge sites1262 and 1263 was established (Appendix A). Based on the fact that Baintensity maxima correspond to Fe intensity maxima and Ca intensityminima at Walvis Ridge sites (Röhl et al., 2007), we assume that Baintensity maxima at Site 550 correspond to Ba cycle minima at WalvisRidge.

The accuracy was checked by comparing the bulk δ13C record(Cramer et al., 2003; Raffi et al., 2005) with the orbitally tuned high-resolution bulk δ13C record of ODP Site 1262 (McCarren and Zachos,personal communication). The two δ13C curves match (Fig. 6) andverify that the cyclostratigraphy of Site 550 is robust. Between theunconformity (407.38mbsf) and the PETM carbon isotope values atSite 550 are stretched because of the lack of reliable tie points within acondensed interval in the center of the CIE. The relative distance ofashes+19 and−17 to the onset of the PETM and the relative positionwith respect to magnetochron C24r based on the new cyclostrati-graphy is shown in Table 1. We estimate that the error of thecyclostratigraphy is less than two precession cycles.

6. Discussion

6.1. Relative distance of ash −17 to the onset of the PETM

On the basis of a 40Ar/39Ar date for ash−17 and its relative distanceto the PETM at Site 550 an age of 55.0 Mawas assigned to the NP9/NP10zonal nannofossil boundary (Swisher and Knox, 1991). Both Cande andKent (1992, 1995) and Berggren et al. (1995) utilized this age as a tiepoint for their time scales. Because this calibration point was deemedflawed (Aubry et al., 1996), Ogg and Smith (2004) placed ash −17midway through C24r based on the magnetostratigraphy of Site 550,and then applied the age of 55.07±0.5 Ma derived by J. D. Obradovich(see postscript of Berggren et al., 1995) to themiddle of C24r in the newGPTS2004. All these assumptions, however, are void because of a criticalunconformity in the Site 550 record between the ash layers and thePETM. Wei (1995) concluded that the exact magnetostratigraphicposition of ash −17 and the NP9/NP10 boundary could not bedetermined because of hiatuses. Nonetheless, an attempt was made toorbitally calibrate a stacked composite bulk carbon isotope record forfourDSDP andODP Sites. This effort, yielded relative ages of 440 kyr and240 kyr post the CIE for ashes+19and−17, respectively (Cramer et al.,2003).

With benefit of the new cyclostratigraphy for the Walvis Ridge sites(Röhl et al., 2007;Westerhold et al., 2007) the revised Site 550 chronologyplaces ashes +19 and −17 at 862 and 672 kyr, respectively, after theonset of the CIE (Table 1). Moreover, ash −17 is located at a point 57%through C24R. This confirms that hiatuses at Site 550 compromised theastronomical calibration of the carbon isotope stack as proposed by

Cramer et al. (2003), although the relative distance between ash layers+19 and −17 is similar (~200 kyr). The revised relative stratigraphynowalso suggests that thepositive ash series (+1 to+140) inDenmarkwas deposited within ~300 kyr instead of the proposed ~600 kyr byEgger and Brückl (2006).

6.2. Implications for the absolute age of the PETM

With more precise constraints on the timing of ash −17 relative tothe onset of the PETM, we can now constrain the absolute age of thePETM, and consider the implications for the orbitally calibratedPaleocene time scale. An absolute age calibration for the PETM is stillunattainable because of uncertainties in both radiometric dating andorbital solutions (Westerhold et al., 2007, 2008). The primary sources ofuncertainty include: (1) a 2.5% error propagation in Ar/Ar geochronol-ogy because of uncertainties in the ages of standards and radioactivedecay rates (also see Renne et al., 1998a,b; Min et al., 2000), and (2) thechaotic diffusion in the orbital solutions beyond ~42 Ma (Laskar et al.,2004) which does not allow for precise computation of minima in thevery long eccentricity cycle needed for accurate orbital tuning of theearly Cenozoic. However, because of its overall stability, the long eccen-tricity cycle (405 kyr) can be used for orbital tuning (Varadi et al., 2003;Laskar et al., 2004) todevelop afloating time scale (e.g., HinnovandOgg,2007). This approach was used to successfully develop a cycle basedchronology for the entire Paleocene epoch (Westerhold et al., 2008).This cyclostratigraphy presents three different tuning options offset byone long eccentricity cycle resulting into three possible absolute ageestimates of 55.53, 55.93 and 56.33 Ma, respectively for the PETM(Fig. 7).

In order to obtain an absolute age for the PETM from the new cyclerecord of Site 550we need a precise radiometric age for ash+19 and/orash −17. The most precise age has been obtained from multiple — aswell as single grain 40Ar/39Ar laser-fusion determinations on sanidinesfrom ash−17 (Storey et al., 2007). According to this study ash−17 hasan absolute age of 55.12±0.12 Ma calibrated to a standardmonitor ageof 28.02 Ma for the FCT (Renne et al., 1998b; Villeneuve, 2004).Assuming a 672 kyr duration between ash −17 and the onset PETM,and an error of ±21 kyr in the cyclostratigraphic position of ash −17,coupledwith a±120 kyr uncertainty in the radiometric age of ash−17,weobtain anage range for the PETMof55.651 to55.933 Ma (Fig. 7). Thisage would be consistent with the option 2 age (55.93 Ma) derived fromthe astronomical calibration (Westerhold et al., 2008).

Alternatively, with the new astronomically recalibrated age for theFCT of 28.201 Ma (Kuiper et al., 2008), the ash −17 age of Storey et al.(2007) must be corrected to 55.48±0.12 Ma. In this case the age rangefor the PETMwould be from 56.011 to 56.293 Ma between option 2 andoption 3 of Westerhold et al. (2008) (Fig. 7). Because the age range iscloser to option 3 we suggest that the radiometric age of the ash −17might be slightly older, about 55.15 Ma, as already shown in Fig. S1 ofStorey et al. (2007). If so, the age of the K/Pg boundary of 66.08 Maderived from astronomically calibrated stratigraphic framework for thePaleocene would be consistent with the revised astronomical age of65.95 Ma (Kuiper et al., 2008) finally eliminating the dating dilemmabetween cyclostratigraphy and radiometric dating (Westerhold et al.,2008).

Still a discrepancy exists in the number of 405-kyr cycles in theolder part of the Paleocene in the Zumaia record in Spain (Kuiper et al.,2008)which, if verified,would prevent the construction of a tuned andstable Paleocene time scale (Hilgen, 2008). One additional 405-kyrcycle was found in Zumaia for the interval between the K/Pg boundaryand the top of the normal polarity interval of C28n compared toWalvisRidge (South Atlantic) and Shatsky Rise (NW Pacific) records (Wester-hold et al., 2008). One additional 405 kyr cycle, would require an age of55.93 Ma for the PETM in order to be consistent with the recalibratedage of the K/Pg boundary of 65.95 Ma. However, this estimate wouldbe inconsistent with the calibrated age of the PETM using the relative

Fig. 5. Cyclostratigraphy for Cores DSDP 550-30X and -31X(a), -32X and -33X (b), and -34X (c) using high-resolution XRF barium (black) and calcium (blue) intensity data. Theprecession-related cycles have been extracted by Gaussian filtering. The filters have been adjusted according to the proposed change in cycle thickness by detailed wavelet analysis.Numbers indicate the number of precession cycles relative to the onset of the PETM as in Westerhold et al. (2007). Position of ashes +19 and −17 according to Knox (1984);position of nannofossil markers after Raffi et al. (2005). [Precession filter details: a, 365 to 374 mbsf at 1.61±0.483 cycle/m (c/m). 375 to 383 mbsf at 1.46±0.438c/m. b, 384 to393 mbsf at 1.55±0.465 c/m. 394 to 401 mbsf at 1.53±0.459 c/m. c, 403 to 408 mbsf at 1.46±0.438 c/m.] (For interpretation of the references to color in this figure legend, thereader is referred to the web version of this article.)

417T. Westerhold et al. / Earth and Planetary Science Letters 287 (2009) 412–419

Fig. 6. Bulk carbon isotope data fromODP Site 1262 (green; McCarren and Zachos, personal communication) and DSDP Site 550 (black; Cramer et al., 2003; Raffi et al., 2005), and XRFdata from Site 550 (Ca in blue, Ba in black, Ti in green, Zr in orange) plotted relative to the onset of the PETM. The agemodel for 550 is based on the cyclostratigraphy developed in thisstudy, the age model for 1262 is fromWesterhold et al. (2007). Data from ash layers have been removed from the Ca and Ba data. Blue bars mark the transient global warming eventsPETM and Elmo, the turquoise bar marks themain ash deposits, the gray dashed lines mark the positions of ash layers +19 and−17, and the solid black line the unconformity at Site550. Note that the interval prior to the unconformity at Site 550 is condensed. The lettering of carbon isotope transient events is after Cramer et al (2003). (For interpretation of thereferences to color in this figure legend, the reader is referred to the web version of this article.)

418 T. Westerhold et al. / Earth and Planetary Science Letters 287 (2009) 412–419

distance of ash −17 and a FCT standard age of 28.201 Ma. Because thetuning of Westerhold et al. (2008) provides a solution which isconsistent with both the age of the ashes and the relative durations ofthe Paleocene section in relation to the PETM and the K/Pg boundary,ourfindings suggest the tuningof Zumaiamight be in error. These issueswill be resolved when precise ages for +19 ash are available and thedetailed ash −17 correlation might get confirmed in other sections.

7. Conclusion and perspectives

High-resolutionXRFdata fromDSDPSite550 enabledus to constructa reliable cyclostratigraphy for the early Eocene. The establishedcyclostratigraphy is consistent with the ODP Leg 208 Walvis Ridgecyclostratigraphy (Westerhold et al., 2007) and for the first timeprovides the exact relative position of ash layers +19 and −17 withinmagnetochron C24r. The combination of the most recent radiometricage estimates for ash −17 and the here within newly constraineddistance between ash −17 and the onset of the PETM results in newpossible absolute age ranges for thePETMwith respect toabsolute age of

Table 1Position of ash +19 and−17 in DSDP Site 550 with respect to magnetochron C24r andthe onset of the PETM.

Ash DSDP Site 550 depth(mbsf)

Relative positionin C24r

Relative distance to onset PETM(kyr)

+19 392.99 C24r.37a 862±21−17 399.76 C24r.43a 672±21

a inverted stratigraphic placement relative to the present (see text).

the FCT sanidine standard age. The new results suggest that therecalibrated FCT standard age of 28.201 Ma (Kuiper et al., 2008) isconsistent with tuning option 3 in the astronomically calibrated

Fig. 7. Absolute age range for the onset of the PETM (gray bars) based on the age (Storeyet al., 2007) and relative distance (this study) of ash−17 with respect to the age of theFish Canyon Tuff (FCT) sanidine standard of 28.02 Ma (Villeneuve, 2004) and28.201 Ma (Kuiper et al., 2008). Horizontal black lines mark the three possible optionsof the age range for the onset of the PETM based on the astronomically calibratedPaleocene time scale (Westerhold et al., 2008).

419T. Westerhold et al. / Earth and Planetary Science Letters 287 (2009) 412–419

Paleocene time scale (Westerhold et al., 2008) but do not support theexistence of any additional 405-kyr cycle in the early Paleocene record.

Acknowledgments

We are grateful to Dennis Kent and an anonymous referee for theirthorough reviews. This research was supported by the DeutscheForschungsgemeinschaft (DFG), the DFG-Leibniz Center for SurfaceProcess and Climate Studies at the University of Potsdam, and theNational Science Foundation (NSF Grant EAR-0628719). We areindebted to the IODP Bremen Core Repository (BCR) staff for corehandling, and V. Lukies for her work in the XRF Core Scanner Lab atMARUM. This research used samples and data provided by the IntegratedOcean Drilling Program (IODP).

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at 10.1016/j.epsl.2009.08.027.

References

Agnini, C., Fornaciari, E., Raffi, I., Rio, D., Röhl, U., Westerhold, T., 2007. High-resolutionnannofossil biochronology of middle Paleocene to early Eocene at ODP Site 1262:implications for calcareous nannoplankton evolution. Mar. Micropaleontol. 64,215–248.

Aubry,M.-P., Berggren,W.A., Stott, L.D., Sinha,A., 1996. Theupper Paleocene–lowerEocenestratigraphic record and the Paleocene–Eocene boundary carbon isotope excursion:implications for geochronology. In: Knox, R.W.O.B., Corfield, R.M., Dunay, R.E. (Eds.),Correlation of the Early Paleogene in Northwest Europe: Geological Society SpecialPublication, vol. 101, pp. 353–380.

Berggren, W.A., Kent, D.V., Swisher III, C.C., Aubry, M.P., 1995. A revised Cenozoicgeochronology and chronostratigraphy. In: Berggren, W.A., Kent, D.V., Aubry, M.P.,Hardenbol, J. (Eds.), Geochronology, Time Scales and Global Stratigraphic Correlation:SEPM, Spec. Publ., vol. 54, pp. 129–212.

Bøggild, O.B., 1918. Den vulkanske Aske i Moleret samt en Oversigt over Danmarks ældreTertiærbjærgarter.

Cande, S.C., Kent, D.V., 1992. A new geomagnetic polarity time scale for the late Cretaceousand Cenozoic. J. Geophys. Res. 97, 13,917–13,951.

Cande, S.C., Kent, D.V., 1995. Revised calibration of the geomagnetic polarity timescale forthe Late Cretaceous and Cenozoic: J. Geophys. Res. 100, 6093–6095.

Cramer, B.S., Wright, J.D., Kent, D.V., Aubry, M.-P., 2003. Orbital climate forcing of δ13Cexcursions in the late Paleocene–Eocene (chrons C24n–C25n). Paleoceanography 18,1097. doi:10.1029/2003PA000909.

Egger, H., Brückl, E., 2006. Gigantic volcanic eruptions and climate change in the earlyEocene. Int. J. Earth Sci. (Geol. Rundsch.) 95, 1065–1070.

Eldholm, O., Thomas, E., 1993. Environmental impact of volcanic margin formation.Earth Planet. Sci. Lett. 117, 319–329.

Haaland, H.J., Furnes, H., Martinsen, O.J., 2000. Paleogene tuffaceous intervals, Grane Field(Block 25/11), Norwegian North Sea: their depositional, petrographical, geochemicalcharacter and regional implications. Mar. Pet. Geol. 17, 101–118.

Hay,W.W.,DeConto, R.,Wold, C.N.,Wilson, K.M.,Voigt, S., Schulz,M.,Wold-Rossby,A., Dullo,W.-C., Ronov, A.B., Balukhovsky, A.N., Soeding, E., 1999. Alternative global Cretaceouspaleogeography. In: Barrera, E., Johnson, C. (Eds.), The Evolution of Cretaceous Ocean/Climate Systems: Geological Society of America Special Paper, vol. 332, pp. 1–47.

Heister, L.E., O'Day, P.A., Brooks, C.K., Neuhoff, P.S., Bird, D.K., 2001. Pyroclastic depositswithin the East Greenland Tertiary flood basalts. J. Geol. Soc. (Lond.) 158, 269–284.

Hilgen, F.J., 2008. Recent progress in the standardization and calibration of the CenozoicTime Scale. Newsl. Stratigr. 43, 15–22.

Hilgen, F., Aubry, M.-P., Berggren, B., van Couvering, J., McGowran, B., Steininger, F.,2008. The case for the original Neogene. Newsl. Stratigr. 43, 23–32.

Hinnov, L.A., Ogg, J.G., 2007. Cyclostratigraphy and the astronomical time scale. Stratigraphy4, 239–251.

Knox, R.W.O.B., 1984. Nannoplankton zonation and the Palaeocene/Eocene boundarybeds of NW Europe: an indirect correlation by means of volcanic ash layers. J. Geol.Soc. 141, 993–999.

Knox, R.W.O.B., 1985. Stratigraphic significance of volcanic ash in Paleocene and Eocenesediments at Sites 549 and 550. In: Graciansky, P.C.D., Poag, C.W., et al. (Eds.), InitialReports DSDP, vol. 80. U.S. Government Printing Office, Washington, pp. 845–850.

Knox, R.W.O.B., 1996. Correlation of the early Paleogene in northwest Europe: an overview.Spec. Publ. — Geol. Soc. Lond. 101, 1–11.

Knox, R.W.O.B., Morton, A.C., 1983. Stratigraphical distribution of early Palaeogenepyroclastic deposits in the North Sea Basin: Proceedings of the Yorkshire GeologicalSociety, vol. 44, pp. 355–363.

Knox, R.W.O.B., Morton, A.C., 1988. The record of early Tertiary N Atlantic volcanism insediments of the North Sea Basin. In: Morton, A.C., Parson, L.M. (Eds.), Early TertiaryVolcanism and the Opening of the NE Atlantic: Special Publications: London,Geological Society, vol. 39, pp. 407–419.

Kuiper, K.F., Deino, A., Hilgen, F.J., Krijgsman, W., Renne, P.R., Wijbrans, J.R., 2008.Synchronizing rock clocks of Earth history. Science 320, 500–504.

Larsen, L.M., Fitton, J.G., Pedersen, A.K., 2003. Paleogene volcanic ash layers in the DanishBasin: compositions and sourceareas in theNorthAtlantic Igneous Province. Lithos 71,47–80.

Laskar, J., Robutel, P., Joutel, F., Gastineau, M., Correia, A., Levrard, B., 2004. A long-termnumerical solution for the insolation quantities of the Earth. Astron. Astrophys. 428,261–285.

Lourens, L.J., Sluijs, A., Kroon, D., Zachos, J.C., Thomas, E., Röhl, U., Bowles, J., Raffi, I., 2005.Astronomical pacing of late Palaeocene to early Eocene globalwarming events. Nature435, 1083–1087.

Min, K., Mundil, R., Renne, P.R., Ludwig, K.R., 2000. A test for systematic errors in 40Ar/39Ar geochronology through comparison with U/Pb analysis of a 1.1-Ga rhyolite.Geochim. Cosmochim. Acta 64, 73–98.

Müller, C., 1985. Biostratigraphic and paleoenvironmental interpretation of the GobanSpur regionbasedona studyof calcareousnannoplankton. In:Graciansky, P.C.D., Poag,C.W., et al. (Eds.), Initial Reports DSDP, vol. 80. U.S. Government Printing Office,Washington, pp. 573–599.

Ogg, J.G., Smith, A.G., 2004. The geomagnetic polarity time scale. In: Gradstein, F., Ogg, J.,Smith, A. (Eds.), A Geological Timescale 2004. InCambridge University Press, pp. 63–86.

Paillard, D., Labeyrie, L., Yiou, P., 1996. Macintosh program performs time-seriesanalysis. Eos Trans. AGU 77 http://www.agu.org/eos_elec/96097e.html.

Paytan, A., Griffith, E.M., 2007. Marine barite: recorder of variations in ocean exportproductivity. Deep-Sea Res., Part 2, Top. Stud. Oceanogr. 54, 687–705.

Raffi, I., Backman, J., Pälike, H., 2005. Changes in calcareous nannofossil assemblagesacross the Paleocene/Eocene transition from the paleo-equatorial Pacific Ocean.Palaeogeogr. Palaeoclimatol. Palaeoecol. 226, 93–126.

Rea, D.K., Zachos, J.C., Owen, R.M., Gingerich, P.D., 1990. Global change at thePaleocene–Eocene boundary: climatic and evolutionary consequences of tectonicevents:. Palaeogeogr. Palaeoclimatol. Palaeoecol. 79, 117–128.

Renne, P.R., Karner, D.B., Ludwig, K.R., 1998a. Radioisotope dating: enhanced: absoluteages aren't exactly. Science 282, 1840–1841.

Renne, P.R., Swisher, C.C., Deino, A.L., Karner, D.B., Owens, T.L., DePaolo, D.J., 1998b.Intercalibration of standards, absolute ages and uncertainties in 40Ar/39Ar dating:.Chem. Geol. 145, 117–152.

Röhl, U.,Westerhold, T., Bralower, T.J., Zachos, J.C., 2007. On the duration of the Paleocene–Eocene thermal maximum (PETM). Geochem. Geophys. Geosyst. 8. doi:10.1029/2007GC001784.

Saunders, A.D., Fitton, J.G., Kerr, A.C., Norry, M.J., Kent, R.W., 1996. The North Atlanticigneous province. In: Mahoney, J.J., Coffin, M.F. (Eds.), Large Igneous Provinces:Continental, Oceanic, and Planetary Flood Volcanism: Geophysical MonographSeries, vol. 100.

Shipboard Scientific Party, 1985. Site 550. In: Graciansky, P.C.D., Poag, C.W., et al. (Eds.),Initial Reports DSDP, vol. 80. U.S. Government Printing Office, Washington, pp.251–355.

Storey, M., Duncan, R.A., Swisher III, C.C., 2007. Paleocene–Eocene thermal maximumand the opening of the Northeast Atlantic. Science 316, 587–589.

Swisher III, C.C., Knox, R.W.O.B., 1991. The age of the Paleocene/Eocene boundary: 40Ar/39Ar dating of the lower part of NP10, North Sea Basin and Denmark (Abstract),IGCP 308 (Paleocene/Eocene boundary events). International Annual meeting andField Conference, 2–6 December 1991. Brussels, p. 16.

Tjallingii, R., Röhl, U., Kölling, M., Bickert, T., 2007. Influence of the water content on X-ray fluorescence core scanning measurements in soft marine sediments. Geochem.Geophys. Geosyst. Tech. Brief 8. doi:10.1029/2006GC001393.

Torrence, C., Compo, G.P., 1998. A practical guide to wavelet analysis. Bull. Am. Meteorol.Soc. 79, 61–78.

Varadi, F., Runnegar, B., Ghil, M., 2003. Successive refinements in long-term integrations ofplanetary orbits. Astrophys. J. 592, 620–630.

Villeneuve, M., 2004. Radiogenic isotope geochronology. In: Gradstein, F., Ogg, J., Smith,A. (Eds.), A Geological Timescale 2004. InCambridge University Press, pp. 87–95.

Wei, W., 1995. A short note on the Paleocene–Eocene transition in DSDP Hole 550.Earth Planet. Sci. Lett. 131, 423–425.

Westerhold, T., Röhl, U., Laskar, J., Bowles, J., Raffi, I., Lourens, L.J., Zachos, J.C., 2007. Onthe duration of magnetochrons C24r and C25n and the timing of early Eoceneglobal warming events: implications from the Ocean Drilling Program Leg 208Walvis Ridge depth transect. Paleoceanography 22. doi:10.1029/2006PA001322.

Westerhold, T., Röhl, U., Raffi, I., Fornaciari, E., Monechi, S., Reale, V., Bowles, J., Evans, H.F., 2008. Astronomical calibration of the Paleocene time. Palaeogeogr. Palaeocli-matol. Palaeoecol. 257, 377–403. doi:10.1016/j.palaeo.2007.09.016.

Zachos, J.C., Dickens, G.R., Zeebe, R.E., 2008. An early Cenozoic perspective ongreenhouse warming and carbon-cycle dynamics. Nature 451, 279–283.