Evaluating the efficacy of planktonic foraminifer calcite δ18O data for sea surface temperature...

Evaluating the efficacy of planktonic foraminifer calcite d18O datafor sea surface temperature reconstruction for the Late Miocene

Évaluation de l’efficacité du d18O de la calcite des foraminifèresplanctoniques pour la reconstitution des températures de la surface

des mers au Miocène supérieur

Mark Williams a,*, Alan M. Haywood a, Steve P. Taylor b, Paul J. Valdes c,Bruce W. Sellwood d, Claus-Dieter Hillenbrand a

a British Antarctic Survey, Geological Sciences Division, High Cross, Madingley Road, Cambridge CB3 0ET, United Kingdomb Department of Meteorology, University of Reading, Earley Gate, PO Box 243, Reading RG6 6BB, United Kingdom

c School of Geographical Sciences, University of Bristol, University Road, Bristol BS8 1SS, United Kingdomd School of Human and Environmental Sciences, University PO Box 227, Whiteknights, Reading RG6 6AB, United Kingdom

Received 2 September 2004; accepted 22 December 2004

Available online 21 October 2005

Abstract

This study examines the efficacy of published d18O data from the calcite of Late Miocene surface dwelling planktonic foraminifer shells,for sea surface temperature estimates for the pre-Quaternary. The data are from 33 Late Miocene (Messinian) marine sites from a modernlatitudinal gradient of 64°N to 48°S. They give estimates of SSTs in the tropics/subtropics (to 30°N and S) that are mostly cooler than present.Possible causes of this temperature discrepancy are ecological factors (e.g. calcification of shells at levels below the ocean mixed layer),taphonomic effects (e.g. diagenesis or dissolution), inaccurate estimation of Late Miocene seawater oxygen isotope composition, or a realLate Miocene cool climate. The scale of apparent cooling in the tropics suggests that the SST signal of the foraminifer calcite has been reset,at least in part, by early diagenetic calcite with higher d18O, formed in the foraminifer shells in cool sea bottom pore waters, probably coupledwith the effects of calcite formed below the mixed layer during the life of the foraminifera. This hypothesis is supported by the markedlycooler SST estimates from low latitudes—in some cases more than 9 °C cooler than present—where the gradients of temperature and the d18Ocomposition of seawater between sea surface and sea bottom are most marked, and where ocean surface stratification is high. At higherlatitudes, particularly N and S of 30°, the temperature signal is still cooler, though maximum temperature estimates overlap with modern SSTsN and S of 40°. Comparison of SST estimates for the Late Miocene from alkenone unsaturation analysis from the eastern tropical Atlantic atOcean Drilling Program (ODP) Site 958—which suggest a warmer sea surface by 2–4 °C, with estimates from oxygen isotopes at Deep SeaDrilling Project (DSDP) Site 366 and ODP Site 959, indicating cooler than present SSTs, also suggest a significant impact on the d18O signal.Nevertheless, much of the original SST variation is clearly preserved in the primary calcite formed in the mixed layer, and records secular andtemporal oceanographic changes at the sea surface, such as movement of the Antarctic Polar Front in the Southern Ocean. Cooler SSTs in thetropics and sub-tropics are also consistent with the Late Miocene latitude reduction in the coral reef belt and with interrupted reef growth onthe Queensland Plateau of eastern Australia, though it is not possible to quantify absolute SSTs with the existing oxygen isotope data.Reconstruction of an accurate global SST dataset for Neogene time-slices from the existing published DSDP/ODP isotope data, for use ingeneral circulation models, may require a detailed re-assessment of taphonomy at many sites.© 2005 Elsevier SAS. All rights reserved.

* Corresponding author. School of Earth and Environmental Sciences, University of Portsmouth, Burnaby Building, Burnaby Road, Portsmouth PO1 3QL,United Kingdom.

E-mail address: [email protected] (M. Williams).

Geobios 38 (2005) 843–863

http://france.elsevier.com/direct/GEOBIO/

0016-6995/$ - see front matter © 2005 Elsevier SAS. All rights reserved.doi:10.1016/j.geobios.2004.12.001

Résumé

Dans cette étude sont examinées les données publiées du d18O de la calcite des foraminifères planctoniques du Miocène supérieur vivanten surface, ceci afin d’estimer les températures de surface des mers (SST) pour l’antéquaternaire. Les données proviennent de 33 sites marinsdu Miocène supérieur (Messinien) le long d’un gradient latitudinal actuel allant de 64°N à 48°S. Elles indiquent une SST dans la zonetropicale–subtropicale (jusqu’à 30°N et S) globalement plus froide qu’actuellement. Les causes possibles de cette différence de températuresont soit écologiques (ex : calcification des coquilles sous la zone de mélange ( = ocean mixed layer), taphonomiques (ex : diagenèse oudissolution), liées à une estimation imprécise de la composition isotopique de l’oxygène de l’eau de mer au Miocène supérieur ou biennaturelles, reflétant ainsi un réel climat froid au Miocène supérieur. L’échelle du refroidissement apparent au niveau des tropiques suggère quele signal SST de la calcite des foraminifères a été réinitialisé, au moins en partie, par la calcite diagénétique précoce à d18O élevé formée dansles coquilles de foraminifères au sein des eaux interstitielles froides des fonds marins. Ceci est à ajouter aux effets de la calcite formée sous lazone de mélange durant la vie des foraminifères. Cette hypothèse est soutenue par des SST froides très marquées à partir des basseslatitudes—dans certains cas de 9 °C plus froides qu’actuellement—où les gradients de température et la composition en d18O de l’eau de merentre la surface et le fond sont très marqués et où la stratification de la surface océanique est élevée. À plus hautes latitudes, particulièrementau N et au S de 30°, le signal des températures est toujours plus froid bien que les estimations des températures maximales recoupent les SSTactuelles au N et au S de 40°. La comparaison des SST estimées à partir de l’analyse des alkenones dans l’Atlantique Est-tropical (OceanDrilling Program (ODP) Site 958)—analyse qui indique une surface des mers 2 à 4 °C plus froide—et des SST estimées à partir des isotopesde l’oxygène (Deep Sea Drilling Project (DSDP) Site 366 et ODP Site 959)—isotopes indiquant des SST plus froides qu’actuellement—suggèreégalement un impact significatif sur le signal de d18O. Néanmoins, une part importante de la variation de SST d’origine est conservée dans lacalcite primaire formée dans la zone de mélange. Elle enregistre des changements océanographiques séculaires et temporels à la surface deseaux, tels que des mouvements du Front Polaire Antarctique. Des SST plus froides dans les zones océaniques tropicales et subtropicales sontégalement compatibles avec la réduction latitudinale de la ceinture de corail au Miocène supérieur et avec l’interruption de la croissancerécifale sur le Plateau de Queensland (Est de l’Australie). Toutefois il n’est pas possible de quantifier de façon absolue les SST avec lesdonnées isotopiques existantes. La reconstitution précise des SST à l’échelle du globe et pour les tranches de temps du Néogène définies(données isotopiques DSDP/ODP publiées), pourrait nécessiter un re-examen de la taphonomie dans de nombreux sites.© 2005 Elsevier SAS. All rights reserved.

Keywords: Late Miocene; Planktonic foraminifera; Oxygen isotopes; Sea surface temperature; Taphonomy

Mots clés : Miocène supérieur ; Foraminifères planctoniques ; Isotopes de l’oxygène ; Température de surface des mers ; Taphonomie

1. Introduction

The oxygen isotopic composition of fossil planktonic fora-minifer shells has been used as a proxy for ancient sea sur-face temperature (SST), facilitating the reconstruction ofancient climate and oceanography (e.g. Matthews et al., 1980;D’Hondt and Arthur, 1996; Norris et al., 2002). Accurateassessment of secular SST variation is critical for the con-struction of general circulation models (GCMs) used to assesspast climate (e.g. Dowsett et al., 1999). More recently, theefficacy of the oxygen isotope method from planktonic fora-minifer calcite for absolute SST reconstructions has beenquestioned (for a summary of the debate see Zachos et al.,2002; Pearson et al., 2002), though all authors concur thatthe technique is useful for identifying secular and temporaltrends in oceanography. Perhaps most significant is the rec-ognition of early diagenetic calcite formed in planktonic fora-minifer shells in cool sea bottom pore waters, which reset theoriginal temperature signature acquired near the sea surface(Schrag et al., 1995; Pearson et al., 2001). The discrepancybetween ancient SSTs estimated by the oxygen isotopemethod, and modern SSTs, is most marked at shallow burialdepths in the tropics, where the difference in temperature andd18O between surface and sea bottom pore waters is greatest,where surface temperatures may have remained relatively con-stant during the Cenozoic, and where the effects of deeper

burial diagenesis are limited (Schrag et al., 1995; Pearson etal., 2001).

This paper uses a Late Miocene dataset of published oxy-gen isotopes from planktonic foraminifer calcite to identifythe latitudinal distribution of SSTs this produces, and toexplain this in terms of the interplay between taphonomicprocesses (dissolution, diagenesis), oceanography, foramini-fer ecology, and varying estimates of d18O composition forthe oceans of that time. We comment on the feasibility ofreconstructing global SST datasets for Neogene time-slicesusing the existing published Deep Sea Drilling Project(DSDP) and Ocean Drilling Program (ODP) oxygen isotoperecords.

2. Dataset and stratigraphy

The dataset comprises oxygen isotope records from 10 dif-ferent species of planktonic foraminifera at 33 marine loca-tions (DSDP/ODP sites/holes, and some locations detailedby Matthews et al., 1980; Fig. 1). These are from betweenpresent-day latitudes 64°N and 48°S, through a modern meanSST range of 5.3–29.3 °C (Table 1), encompassing much ofthe latitudinal range of SST. The locations display an unevengeographical distribution with most from the Atlantic andIndian oceans (Fig. 1).

844 M. Williams et al. / Geobios 38 (2005) 843–863

The data are sourced from a number of intervals withinthe timeframe 7.2–5.6 Ma (Appendix 1A, B), originally basedon a compilation of isotope data by Dr. Steve P. Taylor in2001, “Late Miocene oxygen isotope data and sea surfacepalaeotemperatures”, at the University of Reading (a copy ofthis compilation is held in the library at the British AntarcticSurvey, http://www.antarctica.ac.uk/resources/library). Thedata are constrained by biochronological, magnetochrono-logical and isotopic markers (Fig. 2), sourced fromODP/DSDP reports and the Ocean Drilling Stratigraphic Net-work (ODSN) website at: http://www.odsn.de/odsn/services/rangecharts/rangecharts.html. Resolution is based on a sub-jective assessment of chronological markers at each site(Appendix 1A): we have not attempted graphic correlation.Chronological control was aided by a major, well-documented, negative excursion in d13C of approximately0.8 per mil (grey area of Fig. 2). The timing of the LateMiocene d13C shift is constrained to 7.6–6.6 Ma (Hodell etal., 2001; Billups, 2002). A number of biochronological indi-ces including the nannofossils Amaurolithus primus, Amau-rolithus amplificus and Discoaster quinqueramus, and theplanktonic foraminifera Globorotalia conomiozea andGloborotalia tumida are also useful (Fig. 2). Age determina-tion at some sites is based on astronomical tuning (e.g. ODPholes 925B, 926A, 926B, 959C). For comparison, our analy-sis includes some locations (selected data for Globiger-inoides sacculifer) from the Atlantic planktonic foramin-iferal isotope dataset of Matthews et al. (1980; = 6 ± 0.5 Ma,according to their chronology). This is correlated with upper-most nannoplankton zone NN11 and lowermost zone NN12,

and probably lies at the upper end of the selected time inter-val (Berggren et al., 1995: Fig. 5).

As the data are sourced from intervals between sites thatare, in some cases, different in age by several hundreds tothousands of years (Fig. 3), some inter-site variation in thed18O signature of the foraminifer calcite will relate to orbitaldriven fluctuations in global climate (Hodell et al., 2001;Grützner et al., 2003). This is less of a problem for quantify-ing overall trends (cooler, warmer) at those sites where a rangeof isotopic values is available through this time interval (e.g.ODP sites 926, 959: Fig. 3), as these will yield SST estimatesfrom a period of more than 1 million years. However, it is aproblem for those sites where only one or two isotopic valuesare available from this interval (Fig. 3), as these may sit eitherwithin a glacial or interglacial episode. We have included thesesites in the total database to maximise the latitudinal and geo-graphical spread of data (Fig. 4a, b), but we use them selec-tively in the assessment of the latitudinal patterns of tempera-ture variation produced by this methodology (see Fig. 4c).

3. Calculating SSTs for the Late Miocene

The d18O values from planktonic foraminifer shells pro-vide a record of calcification temperatures, but to relate thesevalues to sea temperatures requires a detailed understandingof the ecology and physiology of individual species and ofthe surface layer of the oceans (King and Howard, 2003).Several studies have demonstrated that effects such as pho-tosynthesis of symbiotic algae (Spero and Lea, 1993), shell

Fig. 1. Location of DSDP/ODP sites/holes that yield oxygen isotope data from Late Miocene planktonic foraminifera. Dashed line delimits the coastline of theLate Miocene (Markwick unpublished, 2003).Fig. 1. Localisation des sites/forages du DSDP/ODP qui fournissent les données isotopiques à partir des foraminifères planctoniques du Miocène supérieur. Leslignes en pointillés délimitent la ligne des côtes au Miocène supérieur (Markwick inédit, 2003).

845M. Williams et al. / Geobios 38 (2005) 843–863

http://www.antarctica.ac.uk/resources/library

http://www.odsn.de/odsn/services/rangecharts/rangecharts.html


calcification and ontogeny (Spero and Lea, 1996), and car-bonate chemistry (Spero et al., 1997) may influence the degreeto which foraminiferal calcite produces equilibrium with sea-water. Furthermore, the temperature sensitivity of foramin-iferal material must be retained from shell chamber forma-tion through the various stages (development ofgametogenetic calcite, effects of dissolution, diagenesis) lead-ing to the time when the material is recovered for analysis(Elderfield and Ganssen, 2000).

Calcification temperatures for 10 species of Late Mioceneplanktonic foraminifera are evaluated here (Appendix 2). Theecology and physiology of most of these species is well con-strained (Bé and Hutson, 1977; Gasperi and Kennett, 1993;Pearson and Shackleton, 1995; Chaisson and Ravelo, 1997;Palmer et al., 1998; Niebler et al., 1999; Ganssen and Kroon,2000; Dekens et al., 2002; Schmidt and Mulitza, 2002; Kingand Howard, 2003). Because of their different ecologies, lifecycles and physiologies, eight palaeotemperature equations

are used to assess the calcification temperatures of the spe-cies (Table 2). These equations are based on laboratory cul-ture experiments of foraminifera (Erez and Luz, 1983; Bemiset al., 1998; Spero and Lea, 1993) and those from the openocean (Mulitza et al., 2003).

4. Constraining seawater isotopic composition (d18Osw)

On the timescale of less than 10 million years, the globalmean oxygen isotope composition of seawater (d18Osw) isdependent primarily on the volume of continental ice-sheets(Bice et al., 2000). At times of greater ice-sheet volume, suchice acts as an effective reservoir for the lighter isotope 16O,resulting in elevated (more positive) d18O levels in the world’soceans. Shackleton and Kennett (1975a) estimated that fortime periods of ice-free conditions, such as the Cretaceousand Early Cenozoic, the mean d18Osw would have been

Table 1Difference between modern SST and Late Miocene SST estimates. The Late Miocene SSTs are based on the maximum and minimum values for SST given inAppendices 2 and 3 (below). These exclude the temperature estimates produced by Erez and Luz (1983) equation (see Table 2), except in those instances wherealternative temperature equations are not available (e.g. for G. conomiozea and D. altispira). Where more than one species is analysed at an individual location,the temperature range is based on the species regarded as being the best proxy (G. sacculifer at lower latitudes and N. pachyderma at higher latitudes). Thesedata are plotted graphically in Fig. 4a–c. Also given is an indication of relative burial depth

Location (mostlyDSDP/ODP)

Latitude Late Miocene (Messinian) Averagedifference

Modern SST Mean burial depth (m)Temp min Temp max 0–150 151–250 250+

408 63,38 1,7 8,9 –3,1 8,4 X646B 58,21 –2,6 11,8 –0,7 5,3 X552A 56,04 0,5 8,5 –6,1 10,6 X610E 53,22 4 10,8 –5 12,4 X410 45,51 4,5 13,1 –6,7 15,5 X335 37,28 18,3 20,7 –0,5 20 X310 36,87 14,5 15,7 –3,7 18,8 X470 28,91 10,8 16,4 –4,4 18 X98 25,53 21 23,3 –4,35 26,5 X502A 11,49 24,6 29,1 –0,75 27,6 X158 6,62 23,4 26,7 –2,75 27,8 X366A 5,68 21,4 22,4 –5,1 27 X758A 5,38 20,2 26 –5,6 28,7 X925 4,12 23,1 24,7 –3,7 27,6 X503A 4,07 22 27,1 –2,45 27 X926B 3,72 19,5 25,4 –5,05 27,5 X926C 3,72 18,5 24,5 –3 27,5 X959C 3,63 15,4 24,3 –7,65 27,5 X216 1,46 15,2 18,8 –11,7 28,7 X289 –0,5 19,7 23,1 –7,9 29,3 X157 –1,75 24 25,6 1,1 23,7 X709C –3,92 14,3 23,6 –9,25 28,2 X237 –7,08 16,9 24,1 –7,1 27,6 X238 –11,15 18,1 22,8 –6,45 26,9 X214 –11,34 18 22,1 –6,45 26,5 X811A –16,52 15,5 22 –7,85 26,6 X817A –18,16 17,8 22,7 –3,55 23,8 X519 –26,14 15,4 16,5 –6,15 22,1 XCH115-86 –30 14,3 16,7 –1,7 17,2 X590 –31,17 15 19,8 –1,6 19 X284 –40,51 7,8 10,3 –0,95 10 X704 –46,88 3,2 13 1,1 7 X281 –48 5,6 9,9 0,75 7 X


approximately –1.0 per mil standard mean oceanic water(SMOW). This value has been applied to studies of the oceansof the Early Tertiary (e.g. Zachos et al., 1994; Bice et al.,2000) but is unsuitable for calculations of the Late Miocene,as there is evidence for the presence of continental ice (Shack-leton and Kennett, 1975b; Ehrmann et al., 1991; Thiede etal., 1998; Stein, 1991; Phillips et al., 2001). In our palaeotem-

perature calculations we have assumed latest Miocene ice-volumes similar to present (dsw of 0.0 per mil SMOW). Learet al. (2000) estimate dsw of about –0.2 to –0.3 per mil for theinterval 7–6 Ma. The effects of factoring in a –0.2 per milSMOW change in the calculations shown inAppendices 2 and3 is to reduce estimated SSTs by about 1 °C (calculations notshown).

Fig. 2. Late Miocene chronology based on Berggren et al. (1995). The grey area in the lower part of figure denotes the interval of the Late Miocene d13Cexcursion. The LAD and FAD for key nannoplankton and planktonic foraminifera are shown.Fig. 2. Chronologie du Miocène supérieur basée sur Berggren et al. (1995). La zone grise dans la partie inférieure de la figure souligne l’intervalle de l’excursiondu d13C au Miocène supérieur. Sont indiqués également les LAD et FAD de foraminifères planctoniques et nanoplancton clés.

Table 2Temperature equations used to calculate SST. dc is the isotopic composition of the calcite sample (VPDB) and dsw is the isotopic composition of the ambientseawater (SMOW, corrected to VPDB scale in our calculations of SST). Eq. (1), applied to foraminifera at all sites, estimates warmer temperatures than thosefrom data for open ocean plankton tow G. sacculifer, Eq. (2) (Mulitza et al., 2003). Small specimens (250–350 µm) of G. sacculifer may form their shells in thethermocline layer under low light conditions, whilst larger specimens inhabit shallower, warmer, high light environments, for which Eq. (3) is calibrated. Bemiset al. (1998), in culture experiments with G. bulloides (between 15 and 24 °C), noted that successive chambers are progressively enriched in 18O relative toearlier ones and therefore larger shells are enriched in 18O. To compensate for this, they produced three equations, but for palaeoenvironmental reconstructionsthey recommend the 12-chamber Eq. (4) as an average. Because of the influence of light in its habitat, Bemis et al. (1998) devised two equations for assessingpalaeothermometry for O. universa (Eqs. (5) and (6)). We use Eq. (5), calibrated for low light. In Appendices 2 and 3 shell calcification temperatures forG. obliquus, D. altispira, G. conomiozea, G. quinqueloba and N. atlantica are sourced from Eq. (1). Eq. (8) is applied to N. atlantica, which is related toN. pachyderma (see Aksu and Kaminski, 1989), and Eq. (7) to G. obliquus, which is related to G. ruber

Foraminifer species Numberof sites

Equationnumbers

Temperature equation Source reference

All species 33 1 T (°C) = 16.998 – 4.52 (dc – dsw) + 0.028 (dc – dsw)2 Erez and Luz 1983G. sacculifer 18 2 T (°C) = –4.35 (dc – dsw) + 14.91 Mulitza et al. 2003G. sacculifer 18 3 T (°C) = –5.67 (dc – dsw) + 12 Spero et al, 2003G. bulloides 6 4 T (°C) = –4.89 (dc – dsw) + 13.20 Bemis et al. 1998Orbulina universa 3 5 T (°C) = –4.8 (dc – dsw) + 16.5 Bemis et al. 1998Orbulina universa 3 6 T (°C) = –4.8 (dc – dsw) + 14.9 Bemis et al. 1998G. ruber 2 7 T (°C) = –4.44 (dc – dsw) + 14.20 Mulitza et al. 2003Globigerinoides obliquus 1 7 T (°C) = –4.44 (dc – dsw) + 14.20 Mulitza et al. 2003N. pachyderma 2 8 T (°C) = –3.55 (dc – dsw) + 12.69 Mulitza et al. 2003Neogloboquadrina atlantica 1 8 T (°C) = –3.55 (dc – dsw) + 12.69 Mulitza et al. 2003


Fig. 3. Stratigraphical source of oxygen isotope data from planktonic foraminifer calcite for the Late Miocene (data from Appendix 1A). A dashed line indicates that the chronology is approximate.Fig. 3. Source stratigraphique des données isotopiques de l’oxygène issues des foraminifères planctoniques pour le Miocène supérieur (données voir Appendice A). Pointillés lorsque la chronologie est approxi-mative.

848M

.William

set

al./G

eobios38

(2005)843–863

The local or regional surface d18Osw is a function of theglobal mean composition as well as geographical variationscaused by the combined effects of changes in evaporation–precipitation patterns, runoff (in coastal regions), and oceancirculation patterns (Zachos et al., 1994; Schmidt, 1998). Inorder to account for geographical variations, we have usedthree different methodologies and datasets to assess surfacewater d18O.

Zachos et al. (1994) proposed an equation relating d18O topalaeolatitude. Using present-day open ocean data collectedas part of the Geochemical Ocean Sections Study (GEO-SECS), they derived the following equation:

(9)dsw = 0.576 + 0.041y − 0.0017y2 + 1.35 × 10−5y3

Here dsw is the isotopic composition of the seawater (d18O,SMOW) and y is the absolute latitude in the range 0–70°. Thelatitudinal correction of Eq. (9) provides a useful first approachto account for local or regional variations in surface waterd18O.

We have also reconstructed the present-day global distri-bution of d18O using GEOSECS data and other data com-piled by Schmidt (1999); Bigg and Rohling (2000) (full

Fig. 4. Difference between present SST and Late Miocene SST estimates. a)The overall pattern (all 33 sites) shows significantly cooler SSTs for thetropics. The presence of a largely uniform temperature gradient from low tohigh latitude indicates that the values are reflecting, at least in part, originaltemperature variation for the oceans at that time. Sites mentioned in the mainbody of the text are labelled. Also shown is the mean burial depth by lati-tude. b) Mean temperature difference from present at each site. The greatertemperature difference from present in the tropics suggests that the isotopicsignature of the foraminifer calcite, at least in part, has been reset by calcitewith higher d18O in sea bottom pore waters much cooler than those of thetropical sea surface. Equatorial DSDP Site 157, where Late Miocene andpresent SSTs are very similar, may have been affected by calcite over-growths with lower d18O (seeAppendix 1B), perhaps formed at greater burialdepth (see Appendix 1A). c) Shows only those sites with isotopic valuesfrom several horizons (Fig. 3). This provides a range of data that removesproblems of an oscillating climate (glacial to interglacial), which could bias

SST estimates from those sites with only one or two isotopic values from thestratigraphical interval. d) SST estimates from 25 sites with core top mate-rial of G. sacculifer (from database: http://www.pangaea.de/home/smulitza/).Estimates are typically 0–3 °C cooler than actual SSTs, though the discre-pancy is greater in tropical regions, possibly indicating the influence of cal-cite formed near the base of, or below the mixed layer during the life of theforaminifera (see Waelbroeck et al., 2004). SST estimates use the equationof Zachos et al. (1994) for the calculation of d18Osw.Fig. 4. Différences entre les SST estimées pour l’Actuel et le Miocène supé-rieur. a) La tendance générale (les 33 sites considérés) est une SST plusfroide pour les tropiques. La présence d’un gradient de température large-ment uniforme depuis les basses vers les hautes latitudes indique que lesvaleurs reflètent, au moins en partie, la variation originale de températurepour les océans à cette période. Les sites mentionnés dans le corps principaldu texte sont indiqués dans la figure. La profondeur d’enfouissement suivantla latitude est également montrée. b) Différence de température moyenneavec l’Actuel pour chaque site. La plus grande différence de températureavec l’Actuel dans les tropiques suggère que la signature isotopique de lacalcite des foraminifères a été réinitialisée au moins en partie par la calciteenrichie en d18O dans les eaux interstitielles des fonds marins, plus froidesque celles des eaux de surface tropicales. Le site DSDP 157 situé au niveaude l’Equateur où les SST de l’Actuel et du Miocène supérieur sont très simi-laires, a pu être affecté par des dépôts de calcite à faible teneur en d18O (voirAppendice 1B), peut-être formés à des profondeurs d’enfouissement plusgrandes (voirAppendice 1A). c) Montre seulement deux sites avec des valeursisotopiques provenant de plusieurs horizons (Fig. 3). Ceci fournit une gammede données qui résout le problème d’un climat oscillant (glaciaire à intergla-ciaire) qui pourrait influencer les estimations de SST de ces sites qui ontseulement 1 ou 2 valeurs isotopiques par intervalle stratigraphique. d) Esti-mations de SST pour 25 sites ayant du matériel de forage contenant G. sac-culifer (tiré de la base de données: http://www.pangaea.de/home/smulitza/).Les estimations donnent des valeurs de 0 à 3 °C plus froides que les SSTactuelles bien que la différence soit toutefois plus forte dans les régions tro-picales, indiquant peut-être l’influence de la calcite formée près de la base,ou en dessous de la zone de mélange pendant la vie du foraminifère (voirWaelbroeck et al., 2004). Les estimations de SST utilisent l’équation deZachos et al. (1994) pour le calcul du d18OSW.


http://www.pangaea.de/home/smulitza/

http://www.pangaea.de/home/smulitza/

dataset available in Schmidt et al., 1999). In order to recon-struct a continuous surface model we employed a spatial inter-polation routine based on grid averaging (Taylor et al., 2004).This highlights the highly variable distribution of surfacewater d18O (by longitude and latitude) and has enabled us tomake a more constrained correction for local seawater varia-tions in our temperature calculations (Appendices 2 and 3).Nevertheless, lack of correlation (r2 = 0.0978) between thevalues generated by Eq. (9) and the dataset of Schmidt et al.(1999) indicate that estimates of d18Osw using these tech-niques still need to be refined.

Both of the above methods also assume local seawatersalinities for the Late Miocene similar to the present. Prior toshoaling and closure of the Panama Isthmus, the Pacific andAtlantic had similar salinities. However, since closure of theisthmus, the Atlantic Ocean has become more saline than thePacific, andAtlantic d18O is also higher, reflected by the 0.5‰enrichment of d18O in planktonic foraminifera in the Carib-bean (Haug and Tiedemann, 1998; Haug et al., 2001). To com-pensate for these differences we produce a third interpreta-tion of Late Miocene surface d18Osw based on a climatemodelling study for the Messinian (Appendix 3). The particu-lars of the atmospheric GCM (AGCM) used within this study(hereafter referred to as HadAM3) have been well-documented (Gordon et al., 2000). The model was initialisedwith a Messinian geography (Paul Markwick, unpublished2003), SSTs based upon a 27 cos (lat) function including asmall seasonal cycle. Atmospheric CO2 concentrations wereset at 390 ppmv which is in line with estimates derived fromthe stomatal density of fossil leaves (Kürschner et al., 1996).Greenland was specified to be ice-free whilst Antarctica wasset with ice coverage similar to present-day. Orbital param-eters were set to modern values and vegetation cover was

specified as shrub everywhere. The model was integrated for13 simulated years with the final 11 years being used to cal-culate the required climatological means.

Model calculated values for the d18O of seawater (d18Osw)are an attempt to capture longitudinal and latitudinal changeas a function of climate, and is based upon precipitation minusevaporation (P – E) estimates derived from the AGCM.Present-day observed d18Osw is calibrated against observedP – E. The resulting formula (see below) is used to predictd18Osw gradients for the Messinian.

(10)d18O sw = 0.1725 − 0.008 (P − E)

The variation in d18Osw produced by the three methodsdocumented above leads to differences in the calcificationtemperatures calculated from the foraminifera (Appendices2 and 3). However, the trend towards cooling or warming atindividual locations is largely unidirectional and the overallpattern is the same for all three reconstructions. The SST esti-mates for the Messinian are compared to modern mean annualSST values at each site. Modern values are based on Rey-nolds and Smith (1995) and Smith and Reynolds (1998) whointegrated in situ (ship and buoy) SST data, satellite retriev-als and sea-ice coverage data to produce a 1° spatial resolu-tion monthly climatology (dataset available at http://www.emc.ncep.noaa.gov/research/cmb/sst_analysis).

5. Estimated latitudinal variation in SSTs for the LateMiocene

The overall pattern of SSTs for the Late Miocene esti-mated by the methodology described above shows much

Fig. 5. Variation in Late Miocene surface seawater d18O estimated for the 33 marine sites used for the SST estimates. The methods for assessing d18Osw

variation are explained in the text. At most sites the difference in estimates is much less than 0.5 per mil SMOW, though at a few sites the estimates diverge byas much as 1.0 per mil SMOW.Fig. 5. Variation de l’eau de mer de surface au Miocène supérieur. d18O estimé pour les 33 sites marins utilisés. Les méthodes pour déterminer les variations ded18O sont expliquées dans le texte. Dans la plupart des sites, la différence est bien moins que 5 pour mille SMOW bien que dans quelques sites, les estimationsdivergent jusqu’à 1,0 pour mille SMOW.


http://www.emc.ncep.noaa.gov/research/cmb/sst_analysis

http://www.emc.ncep.noaa.gov/research/cmb/sst_analysis

cooler tropics (sometimes by more than 9 °C), cooler mid-latitudes (by about 1–6 °C)—though maximum SST esti-mates approach modern at these latitudes, and a range of tem-perature variation at higher latitudes that overlaps the modern(Fig. 4a–c).

In the tropical Indian Ocean and to a lesser extent in thetropical Pacific and Atlantic oceans, the isotopic evidencegives estimates of surface waters several °C cooler thanpresent in the Late Miocene (Fig. 4a–c). In the Atlantic andPacific tropics, SSTs are typically 3–7 °C cooler, but in theIndian Ocean tropics some SST estimates are more than 9 °Ccooler than present (Fig. 4a). None of the tropical IndianOcean sites yield values approaching the modern, whereastwo Atlantic sites (DSDP 502, 503) and one Pacific site(DSDP 157) yield some SST estimates that are similar topresent (Fig. 4a). In the Indian Ocean six different sites(Fig. 1), all lying below modern equatorial water masses, showa consistent pattern of cooler SST estimates than present, withLate Miocene SSTs typically estimated in the range 17–24 °C(modern range 26.5–28.7 °C). At four localities, cooler thanpresent SST is estimated from the isotopic evidence from bothG. sacculifer and Dentoglobigerina altispira. The large dis-crepancy from modern SSTs at these tropical sites may havebeen even greater for the Late Miocene, as SSTs may havebeen a few degrees higher than present in the central IndianOcean prior to closure of the Indonesian Seaway between4 and 3 Ma (Cane and Molnar, 2001).

Locations in the southern hemisphere south of 30° yieldSSTs that approach or encompass the modern range of varia-tion (Fig. 4a–c).At two southerly sites (DSDP 281, ODP 704),the range of variation is from several °C cooler, to several °Cwarmer. In the North Atlantic north of 40°, three locationsshow consistently cooler SSTs to present (DSDP sites 410,552, 610), though maximum estimated temperatures approachthe modern. At a further site (DSDP 408), the temperatureestimates range from markedly cooler (1.7 °C), to just warmer(8.9 °C), relative to present (8.4 °C). Each of these sites yieldsisotopic data from several horizons for the Late Miocene(Fig. 3), so the pattern is consistent through a wide time inter-val. In contrast, one North Atlantic location (DSDP Hole611C; see Keigwin et al., 1987; site not shown on Fig. 1)produces oxygen isotope values from planktonic foraminifercalcite (d18O = –0.84 to 0.23 VPDB, d18O SMOW = 0.41based on Schmidt et al., 1999) that yield SST estimates rang-ing from similar to modern (12.8 °C), to markedly warmer(18 °C; not shown on Fig. 4a–c).

6. Interpreting the Late Miocene estimated SSTpattern

The temperature estimates from the foraminifer shells givea latitudinal variation that is similar to that of modern SSTs,but in the tropics and sub-tropics is generally several °C cooler(Fig. 4a–c). This suggests that the foraminifera are, at least in

part, recording original latitudinal SST variation preservedin primary calcite formed in the mixed layer at or near theocean surface. The cooler SST signals from the oxygen iso-tope data can be interpreted either as a physiological or eco-logical effect on the foraminifera in life, as a problem of estab-lishing seawater d18O composition (a factor in the calculationof SST; see Table 2), as a taphonomic effect (diagenesis ordissolution), or as a real signal, in which case the LateMiocene tropics were much cooler than present.

We do not consider poor chronology to be a major prob-lem with those sites yielding isotopic values from multiplehorizons, as these give a time-averaged trend for the LateMiocene, in some cases over periods greater than 1 millionyears (Fig. 3). Some problems of correlation are introducedby including those sites with isotopic values from only oneor two horizons (Fig. 3), as these may sit either in glacial orinterglacial episodes, and therefore will not yield an averagesignature. Nevertheless, when we exclude these sites (Fig. 4c),the same overall latitudinal temperature variation emerges(with a cool tropics), suggesting that the pattern is robust.

In the following subsections (Sections 6.1–6.4) we exam-ine the possibilities of physiology and ecology contributingto cooler SST estimates, misinterpreted d18Osw affecting thecalculation of SST, taphonomic resetting of the temperaturesignal, or a real “cool Late Miocene” effect.

6.1. Vital effects and ecology

There are fewer problems associated with the palaeoeco-logical interpretation of Late Miocene foraminifera than thereare, for example, with Late Cretaceous or Early Tertiary forms,as many Late Miocene species are extant. Nevertheless, it ispossible that some environmental factors may have biasedthe results towards cooler SST estimates. Application of theeight temperature equations given in Table 2 (see explana-tion therein) should compensate for many of the vital effectsassociated with the species used to estimate Miocene SSTs.For example, recalibrating some of the temperature valueswith the Globigerina bulloides temperature equation for13-chambered (i.e. larger, more 18O-enriched) specimens(Bemis et al., 1998) records shifts of only 0.4–0.6 °C, so thatsize cannot be the key influence on the cooler temperaturesfrom the North Atlantic.

Although typically inhabiting the mixed layer, during theirlife cycle Globigerinoides ruber, G. sacculifer, Globigerinabulloides and Neogloboquadrina pachyderma produce game-togenic calcite, formed within the thermocline (Duplessy etal., 1981; Erez and Luz, 1983; Niebler et al., 1999), so thatcalcification temperatures for these species may be averagedtowards cooler values (see also Bauch et al., 1997; Schmidtand Mulitza, 2002; Mulitza et al., 2004; Waelbroeck et al.,2004). We have checked original accounts for the presence ofgametogenic calcite, including sac-like structures in G. sac-culifer, which might distort the original temperature signa-ture towards cooler estimates of SST (see Appendix 1B). Theaccretion of secondary gametogenic calcite may have skewed


some values towards cooler estimates of SST, particularlyfrom G. sacculifer and N. pachyderma (see Schmidt andMulitza, 2002; Mulitza et al., 2004). Analysis of G. sacculi-fer from modern core top sediments, mostly from the Atlan-tic Ocean (Fig. 4d), indicate that processes influencing a coolertemperature signal than the sea surface, occur at an early (pre-fossil) stage, particularly in the tropics. This supports the argu-ment that this involves calcite formed near the base of, orbelow the ocean mixed layer during the life cycle of the fora-minifer, perhaps related to the depth of the chlorophyll maxi-mum zone (see Waelbroeck et al., 2004). During periods ofstratification of the upper water column, temperatures at thechlorophyll maximum may differ by several °C from surfacetemperatures (Fairbanks and Wiebe, 1980). The temperaturedifference in the core top material, typically 0–3 °C cooler, isusually much less than in the fossil data from the LateMiocene (contrast Fig. 4a, d), so that these effects do not fullyexplain the temperature discrepancy in the latter material(where estimates are sometimes more than 10 °C cooler;Fig. 4a). However, in the equatorial regions, the core top mate-rial sometimes yield SST estimates 5 °C cooler than actual(see Fig. 4d), so that calcite formed in the foraminifer shellsat depths near the base of, or below the mixed layer, mayhave had a significant influence on the temperature signal ofthe foraminifera at some Late Miocene sites.

G. bulloides is a spring bloom species associated withenhanced up-welling and productivity, reaching its highestabundance at the depth of the chlorophyll maximum zone(Ganssen and Kroon, 2000; King and Howard, 2003). In thenorthern hemisphere it is a species typical of the spring bloomand records SSTs for February and March between 30° and40°N (Ganssen and Kroon, 2000; Elderfield and Ganssen,2000). Through numerous analyses at North Atlantic sites,the Late Miocene SST estimates are cooler (typically by3–6.6 °C) than modern mean annual SSTs, so that data fromDSDP sites 408, 410, 552 and 610 may have been biased forSpring temperatures.

6.2. Accuracy of estimating d18Osw

The variation in surface seawater isotopic composition(d18Osw) produced by the methods documented above (Fig. 5),inevitably leads to differences in the calcification tempera-tures calculated for the foraminifer shells from the equationsgiven in Table 2 (Appendices 2 and 3). Difficulties estimat-ing d18Osw have been cited as a reason for discrepanciesbetween SST estimates from fossil assemblages and thosefrom oxygen isotopes for the Last Glacial Maximum at 18 ka(Wolff et al., 1998). For example, at temperatures between24 and 28 °C, a change in d18O of 0.22 per mil SMOW isequivalent to a temperature difference of 1 °C (Wolff et al.,1998). At DSDP Site 408, the range of Late Miocene SSTestimates is from negative to positive relative to present, andis, in part, influenced by estimates of local d18Osw that varyfrom –0.22 per mil to 1 per mil SMOW (Appendix 2). Nev-

ertheless, at most sites, the range of variation in estimatedd18Osw is much less than this (Fig. 5), and the trend of coolerSST estimates is largely unidirectional and the overall pat-tern is the same (Fig. 4a; Appendices 2 and 3). It is the isoto-pic composition of the foraminifer calcite, rather than the esti-mated d18Osw, that is the key component influencing the SSTestimates. Hence, this methodology appears to produce a pat-tern of latitudinal temperature variation that is robust (Fig. 4a).

6.3. Taphonomy—assessing the importance of a secondarycalcite overprint

Masking of the primary oceanographic signal of foramini-fer calcite may occur because of diagenetic overprinting. Bothcalcite overgrowths and dissolution of the foraminifer shellsduring early burial tend to bias results to isotopically heaviervalues and “cooler” temperature estimates (Savin and Dou-glas, 1973; Killingley, 1983; Isern et al., 1996). Pearson et al.(2001) have suggested that early diagenetic alteration, in theform of fine-scale recrystallisation, may be more pervasivethan originally recognised. Deeper burial, with recrystallisa-tion in pore waters that have lower d18O, will result in theopposite effect, with “warmer” temperature estimates (Schraget al., 1995). A further variable is the rate of recrystallisation,which appears to be faster for bulk carbonates at low lati-tudes, though determining this for foraminifer calcite is farmore difficult (Schrag et al., 1995: pp. 2267 and referencestherein).

The isotopic data evaluated from the Late Miocene plank-tonic foraminifer shells suggest the presence of an early sec-ondary calcite with higher d18O. This can explain the muchcooler SST estimates for the Late Miocene at tropical sites(Fig. 4a–c), where the contrast between the oxygen isotopecomposition of surface waters and sea bottom pore waters isgreatest, particularly at sites with shallow burial, where theeffects of deeper burial diagenesis are minimal.

Pearson et al. (2001) noted a marked discrepancy betweenthe Greenhouse climate of the Late Cretaceous and Early Ter-tiary, and SST estimates in the tropics of only 15–23 °C basedon the d18O composition of planktonic foraminifer calcite, asimilar pattern to our Late Miocene dataset from the IndianOcean (Fig. 4a). They identify resetting of the temperaturesignature of the foraminifer calcite by early diagenetic cal-cite formed in cool sea bottom pore water conditions. Thus,the temperature signature of the Late Miocene foraminifercalcite may also have been reset, at least in part, by earlydiagenesis. Perhaps most telling, is the much greater discrep-ancy between present SSTs and Late Miocene SST estimatesin the tropics relative to mid- to high latitudes (Fig. 4a, b),where the contrast between temperature and oxygen isotopiccomposition between the sea surface and seabed is greatest(Schrag et al., 1995). At low latitudes, bottom water and porefluid temperatures are much cooler than surface water tem-peratures and secondary calcite formed in the first few hun-dred metres of burial may have higher d18O than the primarycalcite (Stoll and Schrag, 2000: pp. 313). In the Late Miocene


dataset the maximum SST difference from present is recordedin the equatorial Indian Ocean (e.g. at DSDP Site 216, whichis on average 11.7 °C cooler, with minimum SST estimatesmore than 14 °C cooler than present). This suggests that earlydiagenetic calcite, formed in cool sea bottom pore water con-ditions with higher d18O, is a factor resetting the temperaturesignal, in this instance resulting in SST estimates that are halfthat of modern values (Table 1).

It is possible to estimate the amount of potential diage-netic calcite using a simple mass balance equation:

(11)d18Ototal = X*d18Odiagenetic + Y*d18Oprimary

Where d18Ototal is the d18O of the primary foraminifer shelland the diagenetic calcite; d18Odiagenetic is the d18O of thediagenetic carbonate; d18Oprimary is the original d18O of theforaminifer shell calcite; X is the amount of diagenetic cal-cite (with 0 ≤ X ≤ 0.5; the probable range of diagenetic over-print is up to 50% according to Pearson et al., 2001, 2002);and Y is the amount of primary foraminifer calcite.

WithX + Y = 1we can rewrite Eq. (11):

d18Oprimary =d18Ototal − X*d18Odiagenetic

1 − X

Thus, at Indian Ocean ODP hole 709C, an extreme case interms of temperature estimates, specimens of G. sacculiferyielding d18O values of –0.45 per mil (which equates to SSTestimates as low as 14.3 °C), require a correction for one-third diagenetic calcite to produce a SST estimate close tomodern at this site. This assumes calcite formed in bottomwaters with a value of ca. 3.3 per mil (based on analyses ofLate Miocene benthic foraminifera at this site by Woodruffet al., 1987; the d18O values for the benthic foraminifer car-bonate are corrected using the adjustment factor of Shackle-ton et al., 1984). However, the actual degree of diageneticoverprint may be somewhat smaller than this (and specimensat this site with d18O of –1.31 per mil, require a correction foronly about 25% diagenetic calcite), as some of the tempera-ture offset will be due to calcite formed deep within, or belowthe mixed layer during the life of the foraminifer (see Section6.1, above): many specimens at Site 709C possess sac-likefinal chambers.

At mid to high latitudes the range of temperature esti-mates approach and sometimes exceed the modern (e.g. DSDPsites 281, 408). Primary d18O values for calcite at these sites,formed in the mixed layer near the ocean surface, were closerto the equilibrium values with cold pore water at the sea bot-tom (Schrag et al., 1995: pp. 2271). For example, at ODPSite 410, specimens of G. bulloides have an isotopic compo-sition of 1.69 per mil (Appendix 2). Here, a correction forone-third diagenetic calcite would shift the Late Miocene SSTestimate only to a few °C warmer (assuming carbonate form-ing at the sea bottom with values of ca. 2.5 per mil; see Keig-win et al., 1987 for data; corrected by adjustment factor of

Shackleton et al., 1984). Other factors, probably ecology (seeSection 6.1), may be more important in recording cooler SSTat this site.

Supporting evidence for SSTs similar to modern at south-ern high latitudes is provided by data for opal deposition docu-mented by Hillenbrand and Fütterer (2001). This suggestsslightly warmer conditions in the Bellingshausen Sea duringthe Late Miocene relative to the Quaternary. At ODP Site704, foraminifer-based temperature estimates vary from 3.2 to13 °C for this interval, compared with a modern mean SST of7 °C. This variation may reflect changes in the position of theAntarctic Polar Front during the Late Miocene (Müller et al.,1991: pp. 465).

Other taphonomic processes appear to have influenced thetemperature signal of the foraminifera minimally. Mixing ofassemblages from warm and cold periods through bioturba-tion, noted in modern core top assemblages (Mulitza et al.,2004), should be less of a problem in the time-averaged val-ues produced from the Late Miocene dataset, which in manycases cover periods in excess of 1 million years (Fig. 3). Also,there is no simplistic correlation between greater modernwater depth (we assume no substantial changes from LateMiocene bathymetry) and cooler estimated Late Miocene shellcalcification temperature (Fig. 6a), as might be expected ifdissolution was a major factor resetting the temperature sig-nature of the foraminifer calcite (Mulitza et al., 2004). Nev-ertheless, data from modern core top specimens of G. saccu-lifer suggest that dissolution might exert a role at somelocalities, though the temperature discrepancy is not as greatas in the fossil material.

Corollary evidence for the foraminifer SST signature beingreset in the tropics is indicated by alkenone unsaturation evi-dence at eastern Atlantic ODP Site 958, off the west coast ofAfrica. Here SSTs of 23.6–26.9 °C are estimated for the LateMiocene (Herbert and Schuffert, 1998), compared with mod-ern SST at this site of 21.5 °C. In contrast SSTs estimatedfrom planktonic foraminifera in the eastern tropical Atlantic(DSDP Site 366, ODP Site 959), at lower latitudes, suggesttemperatures cooler than present, in the range 15.4–24.3 °C(Appendices 2 and 3), where modern SSTs are 27–27.5 °C.

In our dataset, planktonic foraminifera are sourced from arange of burial depths between ca. 36 and 600 m, though mostare from above 300 m (Figs. 4a and 6b). Greater depth ofburial may also influence the temperature signal of the plank-tonic foraminifer calcite. At DSDP Hole 611C Keigwin et al.(1987: pp. 949) noted that d18O values of foraminifer calcitehave been lowered by 1.0 per mil VPDB by carbonate pre-cipitation at elevated in situ temperatures. They noted that18O depletion increases with depth at this site. This results ind18O values in foraminifer calcite that provide SST estimatesthat are as much as 7 °C warmer than present (see above).Burial effects may also have influenced the ‘warmer thanpresent temperature signal’ at East Pacific DSDP Site 157(Fig. 4a, b), where the foraminifera are sourced from a nan-nofossil chalk at depth 325 m, and are associated with calciteovergrowths (Appendix 1A, B), though modern SST at this


Fig. 6. a) Modern water depth (metres) plotted against mean temperature difference between Late Miocene and present. Assuming no substantial changes inwater depth, there is no simple correlation between depth and temperature difference, though greater dissolution (causing 18O-enrichment)—shifting tempe-rature estimates towards cooler values, might be expected in deeper water. Note that different species have varying resistance to dissolution, with N. pachy-derma resistant and G. bulloides very unstable. a1) is the same plot simply for specimens of G. sacculifer, a species regarded as dissolution susceptible. b)Mean temperature difference from present, plotted against mean burial depth (metres). There is a weak suggestion for SSTs less negative than present, at greaterburial depth, that might identify secondary calcite precipitated during deeper burial with lower d18O. c) Temperature difference from present for Late Miocenesites plotted against latitude and modern SSTs, for those sites with mean burial depth of less than 150 mbsf (see also Table 1).Fig. 6. a) Diagramme de la profondeur de l’eau actuelle (en m) en fonction de la différence de température moyenne entre le Miocène supérieur et l’Actuel.Supposant l’absence de changements substantiels dans la profondeur de l’eau, il n’y pas de corrélation simple entre profondeur et différence de températurebien qu’une plus grande dissolution (provoquant un enrichissement en d18O) déplaçant les estimations de température vers des valeurs plus froides, pourraientêtre attendue dans les eaux plus profondes. A noter que les espèces ont une résistance variable à la dissolution, (ex : N. pachyderma résistant et G. bulloides trèsinstable). a1) Est le même point simplement pour les spécimens de G. sacculifer, une espèce considérée comme pouvant se dissoudre facilement. b) Différencede température moyenne pour l’Actuel en fonction de la profondeur d’enfouissement moyenne (mètres). Ceci suggère des SST moins négatives que dansl’Actuel, à des profondeurs d’enfouissement plus grandes, pouvant indiquer la précipitation d’une calcite secondaire au cours d’un enfouissement plus profond(d18O plus faible). c) Différence de température entre l’Actuel et le Miocène supérieur en fonction de la latitude et des SST actuelles, pour les sites dont laprofondeur d’enfouissement moyenne est inférieure à 150 mbsf (voir aussi Table 1)


site is relatively cool for the tropics (Fig. 4a), and the posi-tion of modern cool water currents may partly have influ-enced the temperature difference.

Schmidt et al. (2004) noted that the isotopic compositionof re-precipitated diagenetic calcite can be influenced by porewaters that are isotopically light from the weathering of vol-canic rocks: ash-rich layers and volcanic rocks are commonin DSDP/ODP cores (see also Lawrence, 1991). Stoll andSchrag (2000) noted the same effect, with pore fluids beingisotopically light from low temperature alteration of the under-lying basement rock. Thus, more deeply buried foraminifermaterial may have been influenced by pore water that is iso-topically light (see Fig. 6b). In the Late Miocene dataset thetotal range of burial depths is similar for the tropics and higherlatitudes (Fig. 4a), though overall, tropical sites are less deeplyburied. This might have influenced the more negative LateMiocene SST estimates for the tropics, as these may havebeen more subject to secondary calcite with higher d18O val-ues precipitated in cold pore waters, rather than secondarycalcite precipitated in isotopically light pore waters at greaterburial depths. Plotting temperature difference from presentfor those sites at mean burial depth of less than 150 m hints atthis (Fig. 6c).

Pearson et al. (2001: pp. 485) inferred that stable isotopedata from planktonic foraminifera preserved in carbonateoozes and chalks may have been more susceptible to diage-netic alteration, whilst material from clay-rich sites may bebetter preserved. Most of the sites studied here are sourcedfrom carbonate oozes (Appendix 1B), which may haveencouraged diagenetic alteration. Nevertheless, sites withclay-rich lithologies also yield SST estimates lower thanpresent (e.g. DSDP sites 470, 502A), so this does not appearto form a consistent pattern in the Miocene dataset that wehave examined.

6.4. Cooler Late Miocene oceans?

The possibility of a taphonomic overprint in the tropicshighlights the difficulty using the existing publishedDSDP/ODP oxygen isotope data from planktonic foramini-fer shells for accurate (i.e. near absolute) spatial reconstruc-tions of Late Miocene SSTs, without a thorough re-assessmentof the preservation of the foraminifer shells at each site. Nev-ertheless, the isotopic data still reflect real oceanographictrends, such as latitudinal variation in SST from the tropics tohigh latitudes for the Late Miocene (Fig. 4a, c), whichresembles the modern latitudinal gradient. Also, enough ofthe primary temperature signal is preserved for this data toidentify secular and temporal trends in oceanography and cli-mate, particularly at higher latitudes with lower sedimenta-tion rates (e.g. see Schrag et al., 1995: Fig. 7). Thus, LateMiocene changes in SST at southerly high latitude ODP Site704 probably reflect changes in the position of the AntarcticPolar Front (Müller et al., 1991) during several hundred thou-sand years (Figs. 3 and 4a). In the southwest Pacific, the steeplatitudinal gradient of d18O preserved in foraminifer calcite

between ODP Site 811 (–0.4 to –1.17 per mil VPDB) andDSDP Site 281 (1.75 per mil VPDB) is consistent with thesteep latitudinal temperature gradient associated with themodern subtropical convergence in that area. Savin et al.(1985) noted a similar latitudinal gradient in their d18O recordsat 8 Ma. Cooler SSTs in the southwest Pacific may be con-sistent with Late Miocene reduced coral growth on the Queen-sland Plateau off eastern Australia (Isern et al., 1993, 1996).Temperatures below 24 °C induce stress in coral growth.Cooler tropical SSTs are also consistent with a latitudinalreduction in the global coral reef belt and lower taxonomicdiversity of corals during the latest Miocene (Perrin, 2002and references therein), though it is not possible to quantifyabsolute SSTs with the existing oxygen isotope dataset forthe tropics. This highlights the need for studies of additionalSST proxies, such as foraminifer assemblages, Ca/Mg ratiosand alkenones.

7. Conclusions

Oxygen isotopic signatures from the calcite of 10 speciesof planktonic foraminifera at 33 marine sites covering a lati-tudinal range from 64°N to 48°S are used in spatial estimatesof SST for the Late Miocene. Localities in the tropics havecooler than present SSTs, the degree of cooling (more than9 °C at some sites) suggests the influence of early diageneticcalcite, with d18O higher than the primary calcite, precipi-tated in foraminifer shells in cold sea bottom/pore waters.This was probably coupled with the effects of calcite formednear the base of, or below the mixed layer, during the life ofthe foraminifera. This hypothesis is supported by the greaterdifference in SST estimates between present and Late Mi-ocene for the tropics, relative to mid to high latitudes. At thelatter, primary d18O values for calcite formed near the oceansurface will have been similar to the equilibrium values withcold sea bottom pore water. It is also consistent with higherthan present SST estimated in the Atlantic Ocean tropics fromalkenone unsaturation data at ODP Site 958, contrasted withcooler than present estimates from foraminifer oxygen iso-tope data at DSDP Site 366 and ODP Site 959. Despite thetemperature discrepancy, original SST variation from theocean mixed layer is preserved in primary calcite, and recordssecular and temporal oceanographic changes at or near thesea surface. Cooler SSTs in the tropics and sub-tropics alsoconcur with the Late Miocene latitude reduction in the coralreef belt, though it is not possible to quantify absolute SSTs.We conclude that reconstruction of an accurate global SSTdataset for Neogene time-slices with the existing publishedDSDP/ODP isotope data, for use in GCMs, requires a detailedre-assessment of taphonomy at many sites.

Acknowledgments

Mark Williams thanks the Université Claude BernardLyon-1, Jean Vannier and staff of the UFR des Sciences de la


Terre for their hospitality during a Visiting Professorship(May, 2004). We thank Paul Markwick (GETECH, Leeds)for the palaeogeography of Fig. 1, and Daniela Schmidt (Bris-tol), Frédéric Quillévéré (Lyon), and Alistair Crame and JohnSmellie (BAS), for commenting on earlier versions of themanuscript. Paul Pearson (Cardiff) and Stefan Mulitza (Bre-men) provided valuable and constructive reviews of this paper.This work is part of the GEACEP programme at the BritishAntarctic Survey (BAS, NERC). Palaeoclimate modelling wasfacilitated by the provision of high performance computingtime through NERC and CSAR. We publish with the permis-sion of the Director, BAS (NERC), Professor C.G. Rapley.

Appendix A

DSDP/ODP sites and holes where oxygen isotope data forLate Miocene planktonic foraminifera are recorded. Age datain a) include information from ODSN at http:www.odsn.de/odsn/services/rangecharts/rangecharts.html (given in ital-ics). In a), depths quoted for sample horizons are boreholedepths from ODP/DSDP reports. These are given as metresbelow sea floor (mbsf) and metres below sub-bottom depth(m). For biochronological indices, FAD = first appearancedatum; LAD = last appearance datum. In b), the nature of theforaminifer material analysed (size, diagenesis, dissolution)and the source sediment are indicated.




A.1. Appendix

Site/hole number Latitude (°N) Longitude (°E) Stratigraphical calibration (within the Late Miocene)98 25,53 –77,32 Equivalent to uppermost foram zone N17 , core 3–3, 14–16 cmCH 115-86 –30 –35,57 Equivalent to uppermost foram zone N17, 554 cm157 –1,75 –85,9 Core 36-1, at depth 325 m - isotopic signature from average of 2 values, above local FAD of A.

amplificus158 6,62 –85,24 Zone N17(?), depths 88 to 98.5 m, cores 9 to 11: FAD of G. tumida is within core 10 (at depth

96.03 m)214 –11,34 88,72 95 to 130 m (cores 11–14): 95–115 m= 5.3 to 6.6 Ma (to top d13C excursion), below 130 m for base

of zone CN9216A 1,46 90,21 Zone CN9a, about 7 Ma at depth 109.32mbsf (core 1A)237 –7,08 58,13 Just above top d13C excursion at 95.82mbsf, nannofossil zone CN9, foram zone N17, depths

100.32 to 83.32 m (cores 10, 11) (FAD G. tumida at 85.01238 –11,15 70,53 Top d13C excursion at ca. 175.5mbsf, nannofossil zone CN9, foram zone N17, depths 140 to

180mbsf (cores 16–20) (upper level just below281 –48 147,8 Foram zone Mt10, FAD G. conomiozea between samples 5,CC and 6-2, 45 cm (intra-Kapitean, poor

stratigraphical resolution); sample 5-2, 90 (within core284 -40,51 167,7 ‘Kapitean Stage’, FAD G. conomiozea between 284-17-6, 20 cm and 284-17,CC (intra-Kapitean,

values between 147.9 and 154.4mbsf) [159.4 m, core 17]289 –0,5 158,5 Zone N17-depth ca. 160–200 m (SSP 1975): isotopic values from core 23-1, 104–106, depth

210.04mbsf = ’6.02 Ma’ according to Gasperi and Kennett310 36,87 176,9 Top d13C excursion, core 8-4, at depth 66.6 m to core 7-CC at depth 60.5 m [FAD of G. tumida at

58.51 m, core 7]335 37,28 –35,2 Equivalent to uppermost foram zone N17, core 4-3, 79–81 cm [at circa 318.4 m, level with D. quin-

queramus]366A 5,68 –19,85 Equivalent to uppermost foram zone N17, core 12-1, 99-101 cm [at ca. 102 m, well below FAD of

G. tumida]408 63,38 –28,91 Above d13C excursion - at depth 107 m, foram zone N17 (poorly constrained stratigraphy), LAD of

D. quinqueramus at 108 m [FAD G. conomiozea at410 45,51 –29,31 FAD A. primus 222.5 m, FAD G. conomiozea 216.5 m, upper d13 shift 210–200 m, LAD D. quin-

queramus ca. 160 m470 28,91 –117,5 Foram zone N17 and A. primus nannofossil zone (core 8-1, depth 35–39 cm) (poor stratigraphical

resolution) [FAD of G. conomiozea at 70.24 m, core 8]502A 11,49 –79,38 Top d13C excursion at depth 200.3 m, FAD G. tumida at 145.14 m503A 4,07 –95,64 Above d13C excursion at depth ca. 200.81 m, FAD A. primus below 200 m, FAD G. tumida at

139.27 m (Core No. 33)519 –26,14 –11,67 Above d13C excursion, close to FAD and LAD of A. amplificus, FAD A. primus and LAD D. quin-

queramus (depths 120.9 to 125.3 m)552A 56,04 –23,23 Above d13C excursion at depth 139.8 m, FAD A. primus at 152.25 m, to approximate LAD D. quin-

queramus at ca. 103 m, Chrons C4N (partim) to C3AR590A –31,17 163,4 Depths sampled 160–260 m, conomiozea+ nepenthes zones, A. primus+ D. quinqueramus in cores

18-6 to 27,CC [FAD G. tumida – 160.19 m, FAD G.610E 53,22 –18,89 FAD Amaurolithus at 297 m, LAD D. quinqueramus at ca. 247 m [FAD G. conomiozea at

288.66 m]646B 58,21 –48,37 Top Chron 7–630mbsf, nannofossil zone NN11, FAD of A. primus 574 to 574.5mbsf, LAD

D. quinqueramus 497.1 to 506.9mbsf, Miocene–Pliocene704 –46,88 7,42 Magnetostratigraphy, C3An 240-225mbsf, C3Ar to 260mbsf, C4n to 275mbsf709C –3,92 60,56 FAD of A. primus at ca. 94.25mbsf + or –2.65, LAD of D. quinqueramus below 60mbsf in hole 709A

(cores 8H to 11H in 709C): interval ca. 5.6 to 6.6 Ma758A 5,38 90,36 Foram zone N17, nannofossil zone CN9: 5.9 Ma = 76.4mbsf, 6.57 Ma = 84.3mbsf,

7.09 Ma = 92.1mbsf (see Dickens and Owen, 1999 age model)811A –16,52 148,2 FAD G. tumida at 101 mbsf, LAD D. quinqueramus at 91.6 mbsf; Messinian approximates depths

98.11–120.6 mbsf817A –18,16 149,8 FAD A. primus at ca. 200mbsf, FAD A. amplificus above 195mbsf, LAD A. amplificus just below

186mbsf, LAD D.quinqueramus just below 180 mbsf925B 4,12 –43,3 Astronomical tuning to 6.14 Ma at depth 166.65mbsf - core 19H-1, 65–67, and 5.92 Ma at depth

160.16mbsf - core 18H-3, 66-68926B 3,72 –42,91 Astronomical tuning, 5.6 to 6.59 Ma, depths 169.4mbsf to 186.7mbsf926C 3,72 –42,91 Astronomical tuning, 5.6 to 6.69 Ma, depths 167.2mbsf to 188.9mbsf959C 3,63 –2,74 Astronomical tuning, 6.69–5.6 Ma, calibrated with Berggren et al. 1995, depths 85.71mbsf to

97.41mbsf


A.2. Appendix

Site/holenumber

Latitude (°N) Longitude(°E)

Type of material analysed and source sediment Primary source reference(s)

98 25,53 –77,32 G. sacculifer 550 µm (large) (diagenesis/dissolution) Matthews et al. 1980CH 115-86 –30 –35,57 G. sacculifer greater than 500 µm (large) (diagenesis/dissolution) Matthews et al. 1980157 –1,75 –85,9 O. universa greater than 175 µm (calcite overgrowths) - nannofossil

chalkKeigwin 1979

158 6,62 –85,24 G. sacculifer, 175–295 µm size range (small) - nannofossil chalk/ooze Keigwin 1979; Barrera et al.1985

214 –11,34 88,72 G. sacculifer specimens from 350 to 420 µm (large) - foram/nannoooze

Vincent et al. 1985

216A 1,46 90,21 G. sacculifer specimens from 350 to 420 µm (large) - clay foram-richnanno ooze

Vincent et al. 1985

237 –7,08 58,13 D. altispira from 350 to 420 µm - foram/nanno chalk Vincent et al. 1985238 –11,15 70,53 G. sacculifer from 350 to 420 µm (large) - nannofossil ooze Vincent et al. 1985281 –48 147,76 G. conomiozea - foram/nannofossil ooze Shackleton and Kennett 1975a;

Shipboard Scientific Party 1975a284 –40,51 167,68 N. pachyderma – chalk Kennett et al. 1979; Kennett and

Vella 1975; Shipboard ScientificParty 1975b

289 –0,5 158,51 G. sacculifer specimens from greater than 250 µm fraction -nanno/foram chalk/ooze

Shipboard Scientific Party 1975c;Gasperi and Kennett 1993

310 36,87 176,9 O. universa greater than 175 µm - clayey siliceous/radiolarian ooze Keigwin 1979335 37,28 –35,2 G. sacculifer 300-600 µm - nannofossil ooze Matthews et al. 1980366A 5,68 –19,85 G. sacculifer 500-800 µm (large) - nanno/marl ooze Matthews et al. 1980408 63,38 –28,91 G. bulloides - nannofossil ooze Keigwin et al. 1987 and referen-

ces therein410 45,51 –29,31 G. bulloides - nannofossil ooze Keigwin et al. 1987 and referen-

ces therein470 28,91 –117,52 G. bulloides specimens from 150 to 250 µm – claystone Barrera et al. 1985502A 11,49 –79,38 G. sacculifer specimens from 175 to 295 µm (small) - calcareous clay Keigwin, 1982b, 1982a; Ship-

board Scientific Party 1982503A 4,07 –95,64 G. sacculifer specimens from 175 to 295 µm (small) - siliceous nanno-

fossil oozeKeigwin, 1982b, 1982a; Ship-board Scientific Party, 1982

519 –26,14 –11,67 O. universa - sediment dissolution noted - nannofossil ooze McKenzie et al. 1984; ShipboardScientific Party 1984

552A 56,04 –23,23 G. bulloides - nannofossil ooze Keigwin et al. 1987 and referen-ces therein

590A –31,17 163,36 G. sacculifer from 295 to 412 µm range (small and large) - nannofossilooze

Kennett 1984; Shipboard Scienti-fic Party 1986

610E 53,22 –18,89 G. bulloides - nannofossil ooze Keigwin et al. 1987 and referen-ces therein

646B 58,21 –48,37 N. atlantica, 250–500 specimens above 63 µm mesh - nannoooze/mudstone

Aksu and Hillaire-Marcel 1989;Aksu and Kaminski 1989

704 –46,88 7,42 N. pachyderma from 150 to 250 µm - pelagic ooze Hailwood and Clement 1991;Müller et al. 1991

709C –3,92 60,56 G. sacculifer specimens from 300 to 350 µm (small) - calcareous ooze Vincent et al. 1991; Dickens andOwen 1999

758A 5,38 90,36 G. sacculifer specimens from 300 to 350 µm (small) - calcareous ooze Vincent et al. 1991; Dickens andOwen 1999

811A –16,52 148,16 G. ruber specimens from 180 to 250 µm - foram/nannofossil ooze Isern et al. 1993 1996817A –18,16 149,76 G. ruber specimens from 180 to 250 µm – pelagic ooze Isern et al. 1993925B 4,12 –43,3 G. sacculifer specimens from 355 to 425 µm (large) -

foram/nannofossil oozeChaisson and Ravelo 1997

926B 3,72 –42,91 G. sacculifer range of sizes due to sparse Late Miocene material -nanno ooze

Shackleton and Hall 1997

926C 3,72 –42,91 G. sacculifer range of sizes due to sparse Late Miocene material -nanno ooze

Shackleton and Hall 1997

959C 3,63 –2,74 G. sacculifer, 5 specimens from 300 to 355 µm (small) - nanno chalkwith clay

Shipboard Scientific Party 1996;Norris 1998


Appendix B

Late Miocene compared to present-day mean annual SSTvalues (after Reynolds and Smith, 1995). Isotopic values areto VPDB standard: for conversion of d18Osw SMOW valuesto VPDB scale, a correction factor of –0.22 per mil is appliedfor Eq. (1), and for other Eqs. (2)–(8), a correction factor of

–0.27 per mil. In the column for Bemis et al. (1998) and Mulitzaet al. (2003), the former provides the temperature equationsfor G. bulloides and O. universa, whilst the latter is for G. sac-culifer, G. ruber and N. pachyderma. Temperature estimatesusing the Zachos et al. (1994) calculation for d18Osw utilisethose equations used to calculate temperature values given inbold in the columns based on Schmidt et al. (1999).


Appendix C

Late Miocene SSTs calculated using d18Osw SMOW data from a modelling study (see text for model details). For techniquessee Appendix 2.


References

Aksu, A.E., Kaminski, M.A., 1989. Neogene and Quaternary planktonicforaminifer biostratigraphy and biochronology in Baffin Bay and theLabrador Sea. In: Srivastava, S.P., Arthur, M.A., Clement, B. (Eds.),Proceedings of the Ocean Drilling Program, Scientific Results 105,College Station, TX, pp. 287–304.

Aksu, A.E., Hillaire-Marcel, C., 1989. Upper Miocene to Holocene oxygenand carbon isotopic stratigraphy of Sites 646 and 647, Labrador Sea. In:Srivastava, S.P., Arthur, M.A., Clement, B., et al. (Eds.), ProceedingOcean Drilling Program, Scientific Results, 105. College Station, TX,pp. 689–703.

Barrera, E., Keller, G., Savin, S.M., 1985. Evolution of the Miocene ocean inthe eastern North Pacific as inferred from oxygen and carbon isotopicratios of foraminifera. In: Kennett, J.P. (Ed.), The Miocene Ocean:Paleoceanography and Biogeography. Memoir Geological Society ofAmerican 163, pp. 83–102.

Bauch, D., Carstens, J., Wefer, G., 1997. Oxygen isotope composition ofliving Neogloboquadrina pachyderma (sin.) in the Arctic Ocean. Earthand Planetary Science Letters 146, 47–58.

Bé, A.W.H., Hutson, W.H., 1977. Ecology of planktonic foraminifera andbiogeographic patterns of life and fossil assemblages in the IndianOcean. Micropalaeontology 23, 369–414.

Bemis, B.E., Spero, H.J., Bijma, J., Lea, D.W., 1998. Re-evaluation of theoxygen isotope composition of planktonic foraminifera: experimentalresults and revised paleotemperature equations. Paleoceanography 13,150–160.

Berggren, W.A., Kent, D.V., Swisher III, C.C., Aubry, M.-P., 1995. Geochro-nology, time scales and global stratigraphic correlation. In: Berggren,et al. (Eds.), Geochronology, Time Scales and Global StratigraphicCorrelation. SEPM Special Publication 54, pp. 129–212.

Bice, K.L., Sloan, S.C., Barron, E.J., 2000. Comparison of early Eoceneisotopic palaeotemperatures and the three-dimensional OGCM field: thepotential for use of model derived surface water d18O. In: Huber, B.T.,MacLeod, K.G., Wing, S.L. (Eds.), Warm Climates in Earth History.Cambridge University Press, New York, pp. 79–132.

Bigg, G.R., Rohling, E.J., 2000. An oxygen isotope data set for marinewater. Journal of Geophysical Research 105, 8527–8535.

Billups, K., 2002. Late Miocene through early Pliocene deep water circula-tion and climate change viewed from the sub-Antarctic South Atlantic.Palaeogeography, Palaeoclimatology, Palaeoecology 185, 287–307.

Cane, M.A., Molnar, P., 2001. Closing the Indonesian seaway as a precursorto east African aridification around 3–4 million years ago. Nature 211,157–162.

Chaisson, W.P., Ravelo, A.C., 1997. Changes in upper water-column struc-ture at site 925, late Miocene–Pleistocene: planktonic foraminiferaassemblage and isotopic evidence. In: Shackleton, N.J., Curry, W.B.,Richter, C., Bralower, T.J. (Eds.), Proceedings Ocean Drilling Program,Scientific Results 154, College Station, TX, pp. 255–268.

Dekens, P.S., Lea, D.W., Pak, D.K., Spero, H.J., 2002. Core top calibrationof Mg/Ca in tropical foraminifera: refining paleotemperature estimation.Geochemistry, Geophysics, Geosystems 3 (4). (doi: 10.1029/2001GC000200).

D’Hondt, S., Arthur, M.A., 1996. Late Cretaceous oceans and the cooltropical paradox. Science 271, 1838–1841.

Dickens, G.R., Owen, R.M., 1999. The Latest Miocene–Early Pliocenebiogenic bloom: a revised Indian Ocean perspective. Marine Geology161, 75–91.

Dowsett, H.J., Barron, J.A., Poore, R.Z., Thompson, R.S., Cronin, T.M.,Ishman, S.E., Willard, D.A., 1999. Middle Pliocene palaeoenvironmen-tal reconstruction: PRISM2. US Geological Survey Open File Report,99-535. Available at: http://pubs.usgs.giv/openfile/of99-535/.

Duplessy, J.-C., Blanc, P.L., Bé, A.W.H., 1981. Oxygen-18 enrichment ofplanktonic foraminifera due to gametogenic calcification below theeuphotic zone. Science 213, 1247–1250.

Ehrmann, W.U., Grobe, H., Futterer, D.K., 1991. Late Miocene to Holoceneglacial history of East Antarctica revealed by sediments from sites745 and 746. Proceedings Ocean Drilling Program, Scientific Results,119, College Station, TX, pp. 239–251.

Elderfield, H., Ganssen, G., 2000. Past temperature and d18O of surfaceocean waters inferred from foraminiferal Mg/Ca ratios. Nature 405,442–445.

Erez, J., Luz, B., 1983. Experimental palaeotemperature equation for plank-tonic foraminifera. Geochimica et Cosmochimica Acta 47, 1025–1031.

Fairbanks, R.G., Wiebe, P.H., 1980. Foraminifera and chlorophyllmaximum: vertical distribution, seasonal succession, and paleoceano-graphic significance. Science 209, 1524–1526.

Ganssen, G.M., Kroon, D., 2000. The isotopic signature of planktonicforaminifera from NE Atlantic surface sediments: implications for thereconstruction of past oceanic conditions. Journal of the GeologicalSociety of London 157, 693–699.

Gasperi, J.T., Kennett, J.P., 1993. Miocene planktonic foraminifers at DSDPSite 289: depth stratification using isotopic differences. In: Berger, W.H.,Kroenke, L.W., Mayer, L.A., et al. (Eds.), Proceedings of the OceanDrilling Program, Scientific Results 130, College Station, TX, pp. 323–332.

Gordon, C., Cooper, C., Senior, C.A., Banks, H., Gregory, J.M., Johns, T.C.,Mitchell, J.F.B., Wood, R.A., 2000. The simulation of SST, sea iceextents and ocean heat transports in a version of the Hadley Centrecoupled model without flux adjustments. Climate Dynamics 16, 147–168.

Grützner, J., Rebesco, M.A., Cooper, A.K., Forsberg, C.F., Kryc, K.A.,Wefer, G., 2003. Evidence for orbitally controlled size variations of theEast Antarctic Ice Sheet during the late Miocene. Geology 31, 777–780.

Hailwood, E.A., Clement, B.M., 1991. Magnetostratigraphy of ODP sites703 and 704, Meteor Rise, southeastern South Atlantic. In: Ciesiel-ski, P.F., Kristoffersen,Y., et al. (Eds.), Proceedings of the Ocean DrillingProgram, Scientific Results 114, College Station, TX, pp. 367–386.

Haug, G.H., Tiedemann, R., 1998. Effect of the Isthmus of Panama onAtlantic Ocean thermohaline circulation. Nature 393, 673–676.

Haug, G.H., Tiedemann, R., Zahn, R., Ravelo, A.C., 2001. Role of Panamauplift on oceanic freshwater balance. Geology 29, 207–210.

Herbert, T.D., Schuffert, J.D., 1998. Alkenone unsaturation estimates ofLate Miocene through Late Pliocene sea-surface temperatures at Site958. In: Firth, J.V. (Ed.), Proceedings of the Ocean Drilling Program,Scientific Results 159, College Station, TX, pp. 17–21.

Hillenbrand, C.-D., Fütterer, D.K., 2001. Neogene to Quaternary depositionof opal on the continental rise west of the Antarctic Peninsula, ODP leg178, sites 1095, 1096 and 1101. In: Barker, P.F., Camerlenghi, A.,Acton, G.D., Ramsay, A.T.S. (Eds.), Proceedings of the Ocean DrillingProgram, Scientific Results 178, College Station, TX, pp. 1–33.

Hodell, D.A., Curtis, H., Sierro, F.J., Raymo, M.E., 2001. Correlation of lateMiocene to early Pliocene sequences between the Mediterranean andNorth Atlantic. Paleoceanography 16, 164–178.

Isern, A.R., McKenzie, J.A., Müller, D.W., 1993. Palaeoceanographicchanges and reef growth off the northeastern Australian margin: stableisotopic data from Leg 133, Sites 811 and 817, and Leg 21, Site 209. In:McKenzie, J.A., Davies, P.J., Palmer-Julson, A., et al. (Eds.), Proceed-ings of the Ocean Drilling Program, Scientific Results 133, CollegeStation, TX, pp. 263–280.

Isern, A.R., McKenzie, J.A., Feary, D.A., 1996. The role of sea-surfacetemperature as a control on carbonate platform development in thewestern Coral Sea. Palaeogeography, Palaeoclimatology, Palaeoecology124, 247–272.

Keigwin Jr., L.D., 1979. Late Cenozoic stable isotope stratigraphy andpaleoceanography of DSDP sites from the east equatorial and centralNorth Pacific Ocean. Earth and Planetary Science Letters 45, 361–382.

Keigwin Jr., L.D., 1982a. Isotopic paleoceanography of the Caribbean andEast Pacific: role of Panama uplift in late Neogene time. Science 217,350–353.


Keigwin Jr., L.D., 1982b. Stable isotope stratigraphy and paleoceanographyof sites 502 and 503. In: Prell, W.L., Gardner, J.V., et al. (Eds.), InitialReport of Deep Sea Drilling Project, US Government Printing Office 68.pp. 445–453.

Keigwin Jr., L.D., Aubry, M.-P., Kent, D.V., 1987. North Atlantic LateMiocene stable-isotope stratigraphy, biostratigraphy, and magneto-stratigraphy. Initial Report of Deep Sea Drilling Project 94, Part 2,935–956.

Kennett, J.P., 1984. Miocene to early Pliocene oxygen and carbon isotopestratigraphy in the southwest Pacific. Deep Sea Drilling Project Leg 90.Initial Report of Deep Sea Drilling Project 90, Part 2, 1383–1406.

Kennett, J.P., Shackleton, N.J., Margolis, S.V., Goodney, D.E., Dud-ley, W.C., Kroopnick, P.M., 1979. Late Cenozoic oxygen and carbonisotopic history and volcanic ash stratigraphy: DSDP Site 284, southPacific. American Journal of Science 279, 52–69.

Kennett, J.P., Vella, P., 1975. Late Cenozoic foraminifera and paleoceanog-raphy at DSDP site 284 in the cool subtropical south Pacific. InitialReport of Deep Sea Drilling Project 29, 769–799.

Killingley, J.S., 1983. Effects of diagenetic recrystallisation on 18O/16Ovalues of deep-sea sediments. Nature 301, 594–597.

King, A.L., Howard, W.R., 2003. Planktonic foraminiferal flux seasonalityin subantarctic sediment traps: a test for paleoclimate reconstructions.Paleoceanography 18, 1019 (doi: 10.1029/2002PA000839).

Kürschner, W.M., Van der Burgh, J., Visscher, H., Dilcher, D.L., 1996. Oakleaves as biosensors of late Neogene and early Pleistocene paleoatmo-spheric CO2 concentrations. Marine Micropaleontology 27, 299–312.

Lawrence, J.R., 1991. Stable isotopic composition of pore waters and calciteveins. In: Weissel, J., Pierce, J., Taylor, E., Alt, J., et al. (Eds.), Proceed-ings of the Ocean Drilling Program, Scientific Results 121, CollegeStation, TX, pp. 447–455.

Lear, C.H., Elderfield, H., Wilson, P.A., 2000. Cenozoic deep-sea tempera-tures and global ice volumes from Mg/Ca in benthic foraminiferalcalcite. Science 287, 269–272.

Matthews, R.K., Curry, W.B., Lohmann, K.C., Sommer, M.A., 1980. LateMiocene palaeo-oceanography of the Atlantic: oxygen isotope data onplanktonic foraminifera. Nature 283, 555–557.

McKenzie, J.A., Weissert, H., Poore, R.Z., Wright, R.C., Percival Jr., S.F.,Oberhänsli, H., Casey, M., 1984. Paleoceanographic implications ofstable isotope data from Upper Miocene–Lower Pliocene sedimentsfrom the southeast Atlantic (Deep Sea Drilling Project Site 519). InitialReport of Deep Sea Drilling Project 73, 717–724.

Mulitza, S., Boltovskoy, D., Donner, B., Meggers, H., André, P., Wefer, G.,2003. Temperature: d18O relationships of planktonic foraminifera col-lected from surface waters. Palaeogeography, Palaeoclimatology,Palaeoecology 3215, 1–10.

Mulitza, S., Donner, B., Fischer, G., Paul, A., Pätzold, J., Rühlemann, C.,Segl, M., 2004. The South Atlantic oxygen isotope record of plankticforaminifera. In: Wefer, G., Mulitza, S., Ratmeyer, V. (Eds.), The SouthAtlantic in the Late Quaternary. Springer Verlag, Berlin, Heidelberg,pp. 121–142.

Müller, D.W., Hodell, D.A., Ciesielski, P.F., 1991. Late Miocene to earliestPliocene (9.8-4.5 Ma) paleoceanography of the subantarctic southeastAtlantic: stable isotopic, sedimentologic, and microfossil evidence. In:Ciesielski, P.F., Kristoffersen, Y., et al. (Eds.), Proceedings of the OceanDrilling Program, Scientific Results 144, College Station, TX, pp. 459–480.

Niebler, H.-S., Hubberton, H.-W., Gersonde, R., 1999. Oxygen isotopevalues of planktic foraminifera: a tool for the reconstruction of surfacewater stratification. In: Fischer, G., Wefer, G. (Eds.), Use of Proxies inPaleoceanography: Examples from the South Atlantic. Springer Verlag,Berlin, Heidelberg, pp. 165–189.

Norris, R.D., 1998. Miocene–Pliocene surface-water hydrography of theeastern equatorial Atlantic. In: Mascle, J., Lohmann, G.P., Moullade, M.(Eds.), Proceedings of the Ocean Drilling Program, Scientific Results159, College Station, TX, pp. 539–555.

Norris, R.D., Bice, K.L., Magno, E.A., Wilson, P.A., 2002. Jiggling thetropical thermostat in the Cretaceous hothouse. Geology 30, 299–302.

Palmer, M.R., Pearson, P.N., Cobb, S.J., 1998. Reconstructing past oceanpH-Depth profiles. Science 282, 1468–1471.

Pearson, P.N., Ditchfield, P., Shackleton, N.J., 2002. Tropical temperaturesin greenhouse episodes. Nature 419, 898.

Pearson, P.N., Ditchfield, P.W., Singano, J., Harcourt-Brown, K.G., Nicho-las, C.J., Olsson, R.K., Shackleton, N.J., Hall, M.A., 2001. Warm tropi-cal sea surface temperatures in the Late Cretaceous and Eocene epochs.Nature 413, 481–487.

Pearson, P.N., Shackleton, N.J., 1995. Neogene multispecies planktonicforaminifer stable isotope record, Site 871, Limalok Guyot. Proceedingsof the Ocean Drilling Program, Scientific Results 144, College Station,TX, pp. 401–410.

Perrin, C., 2002. Tertiary: the emergence of modern reef ecosystems. In:Kiessling, W., et al. (Eds.), Phanerozoic Reef Patterns, SEPM SpecialPublication 72. Society of Economic Paleontologists and Mineralogists,Tulsa, pp. 587–621.

Phillips, R.L., Ager, T.A., Barron, J.A., Starratt, S., 2001. Diamictite fromthe Chukchi Sea: evidence of northern glaciation during the late Mioceneand early Pliocene. American Geophysical Union, Fall Meeting 2001,abstract #PP52A-0565.

Reynolds, R.W., Smith, T.M., 1995. A high-resolution global sea-surfacetemperature climatology. Journal of Climate 7, 1571–1583.

Savin, S.M., Abel, L., Barrera, E., Hodell, D., Keller, G., Kennett, J.P.,Killingley, J., Murphy, M., Vincent, E., 1985. The evolution of Miocenesurface and near-surface marine temperatures: oxygen isotopic evidence.In: Kennett, J.P. (Ed.), The Miocene Ocean: Paleoceanography andBiogeography. Geological Society of America Bulletin 163, pp. 49–82.

Savin, S.M., Douglas, R.G., 1973. Stable isotope and magnesium geochem-istry of recent planktonic foraminifera from the south Pacific. GeologicalSociety of America Bulletin 84, 2327–2342.

Schmidt, G.A., 1998. Oxygen-18 variations in a global ocean model. Geo-physical Research Letters 25, 1201–1204.

Schmidt, G.A., 1999. Forward modeling of carbonate proxy data fromplanktonic foraminifera using oxygen isotope tracers in a global model.Paleoceanography 14, 482–497.

Schmidt, G.A., Bigg, G.R., Rohling, E.J., 1999. Global seawater oxygen-18 database. (Available at: http://www.giss.nasa.gov/data/o18data).

Schmidt, G.A., Mulitza, S., 2002. Global calibration of ecological modelsfor planktic foraminifera from coretop carbonate oxygen-18. MarineMicropaleontology 44, 125–140.

Schmidt, D.N., Thierstein, H.R., Bollmann, J., Schiebel, R., 2004. AbioticForcing of Plankton Evolution in the Cenozoic. Science, 303, 207–210[DOI: 10.1126/science.1090592] supplementary material at: http://www.sciencemag.org/cgi/content/full/303/5655/207/DC1.

Schrag, D.P., DePaolo, D.J., Richter, F.M., 1995. Reconstructing past seasurface temperatures: correcting for diagenesis of bulk marine carbon-ate. Geochimica et Cosmochimica Acta 59, 2265–2278.

Shackleton, N.J., Hall, M.A., 1997. The Late Miocene stable isotope record,Site 926. In: Shackleton, N.J., Curry, W.B., Richter, C., Bralower, T.J.(Eds.), Proceedings of the Ocean Drilling Program, Scientific Results154, College Station, TX, pp. 367–373.

Shackleton, N.J., Hall, M.A., Boersma, A., 1984. Oxygen and carbon iso-tope data from Leg 74 foraminifers. Initial Reports of the Deep SeaDrilling Project 74, 599–612.

Shackleton, N.J., Kennett, J.P., 1975a. Palaeotemperature history of theCenozoic and the initiation of Antarctic glaciation: oxygen and carbonisotope analyses in DSDP sites 277, 279 and 281. Initial Report of DeepSea Drilling Project 29, 743–755.

Shackleton, N.J., Kennett, J.P., 1975b. Late Cenozoic oxygen and carbonisotopic changes at DSDP site 284: implication for glacial history of theNorthern Hemisphere and Antarctica. Initial Report of Deep Sea DrillingProject 29, 801–807.

Shipboard Scientific Party, 1975a. Site 281. In: Kennett, J.P., Houtz, R.E.,et al. (Eds.), Initial Report of Deep Sea Drilling Project 29. pp. 271–315.

Shipboard Scientific Party, 1975b. Site 284. In: Kennett, J.P., Houtz, R.E.,et al. (Eds.), Initial Report of Deep Sea Drilling Project 29. pp. 403–445.


Shipboard Scientific Party, 1975c. Site 289. In: Andrews, J.E., Packham, G.,et al. (Eds.), Initial Report of Deep Sea Drilling Project 30. pp. 231–398.

Shipboard Scientific Party, 1982. Site 502: Colombia Basin, western Carib-bean Sea. In: Prell, W.L., Gardner, J.V., et al. (Eds.), Initial Report ofDeep Sea Drilling Project 68. pp. 15–162.

Shipboard Scientific Party, 1984. Site 519. In: Hsü, K.J., LaBrecque, J.L.,et al. (Eds.), Initial Report of Deep Sea Drilling Project 73. pp. 27–93.

Shipboard Scientific Party, 1986. Site 590: Lord Howe Rise, 31°S. In:Kennett, J.P., von der Borch, C.C., et al. (Eds.), Initial Report of DeepSea Drilling Project 90, Part 1, pp. 263–376.

Shipboard Scientific Party, 1996. Site 959. In: Mascle, J., Lohmann, G.P.,Clift, P.D., et al. (Eds.), Proceedings of the Ocean Drilling Program,Initial Reports, Project 159. pp. 65–150.

Smith, T.M., Reynolds, R.W., 1998. A high-resolution global sea-surfacetemperature climatology for the 1961–1990 base period. Journal ofClimate 11, 3320–3323.

Spero, H.J., Bijma, J., Lea, D.W., Bemis, B.B., 1997. Effect of seawatercarbonate concentration on foraminiferal carbon and oxygen isotopes.Nature 390, 497–500.

Spero, H.J., Lea, D.W., 1993. Intraspecific stable isotope variability in theplanktic foraminifera Globigerinoides sacculifer: results from labora-tory experiments. Marine Micropaleontology 22, 221–234.

Spero, H.J., Lea, D.W., 1996. Experimental determination of stable isotopevariability in Globigerina bulloides: implications for paleoceanographicreconstructions. Marine Micropaleontology 28, 231–246.

Spero, H.J., Mielke, K.M., Kalve, E.M., Lea, D.W., Pak, D.K., 2003. Mul-tispecies approach to reconstructing eastern equatorial pacific ther-mocline hydrography during the past 360 kyr. Paleoceanography 18 (1),1022, doi: 10.1029/2002PA000814.

Stein, R., 1991. Organic carbon accumulation in Baffin Bay and palaeoen-vironment in high northern latitudes during the past 20 my. Geology 19,356–359.

Stoll, H.M., Schrag, D.P., 2000. High-resolution stable isotope records fromthe Upper Cretaceous rocks of Italy and Spain: glacial episodes in agreenhouse climate? GSA Bulletin 112, 308–319.

Taylor, S.P., Haywood, A.M., Valdes, P.J., Sellwood, B.W., 2004. An evalu-ation of spatial interpolation techniques in global sea-surface tempera-ture reconstructions: last glacial maximum and Pliocene case studies.Quaternary Science Reviews 23, 1041–1051.

Thiede, J., Winkler, A., Wolf-Welling, T., Eldholm, O., Myhre, A.M., Bau-mann, K.-H., Henrich, R., Stein, R., 1998. Late Cenozoic history of thepolar North Atlantic: results from ocean drilling. Quaternary ScienceReviews 17, 185–208.

Vincent, E., Killingley, J.S., Berger, W.H., 1985. Miocene oxygen andcarbon isotope stratigraphy of the tropical Indian Ocean. In: Kennett, J.P.(Ed.), The Miocene Ocean: Paleoceanography and Biogeography. Geo-logical Society of America Memoir 163, pp. 103–130.

Vincent, E., Shackleton, N.J., Hall, M.A., 1991. Miocene oxygen and carbonisotope stratigraphy of planktonic foraminifers at sites 709 and 758,Tropical Indian Ocean. In: Weissel, J., Peirce, J., Taylor, E., Alt, J., et al.(Eds.), Proceedings of the Ocean Drilling Program, Scientific Results121, College Station, TX, pp. 241–252.

Waelbroeck, C., Mulitza, S., Spero, H., Dokken, T., Kiefer, T, Cortijo, E.,2004. A global compilation of late Holocene planktonic foraminiferald18O: relationship between surface water temperature and d18O. Quater-nary Science Reviews, 24, doi:10.1016/j.quascirev.2003.10.014.

Wolff, T., Mulitza, S., Arz, H., Patzold, J., Wefer, G., 1998. Oxygen isotopesversus CLIMAP (18 ka) temperatures: A comparison from the tropicalAtlantic. Geology 26, 675–678.

Woodruff, F., Savin, S.M., Abel, L., 1987. Miocene benthic foraminiferoxygen and carbon isotopes, site 709, Indian Ocean. Proceedings of theOcean Drilling Program, Scientific Results 115, College Station, TX, pp.519–528.

Zachos, J.C., Arthur, M.A., Bralower, T.J., Spero, H.J., 2002. Tropicaltemperatures in greenhouse episodes. Nature 419, 897–898.

Zachos, J.C., Stott, L.D., Lohmann, K.C., 1994. Evolution of early Cenozoicmarine temperatures. Paleoceanography 9, 353–387.


Evaluating the efficacy of planktonic foraminifer calcite δ18O data for sea surface temperature...

Documents

Transcript of Evaluating the efficacy of planktonic foraminifer calcite δ18O data for sea surface temperature...