Southern Ocean warming and hydrological change during the Paleocene – Eocene thermal maximum

Introduction Conclusions References


Conclusions References
Tables Figures
Traditional stable oxygen isotope (δ 18 O) and Mg/Ca studies on planktonic foraminifera from deep-sea sediments indicate a 5-8 • C surface warming during the PETM (Kennett and Stott, 1991;Thomas et al., 2002;Zachos et al., 2003).Reconstruction of absolute sea surface temperatures (SST) from such sections has been problematic because of post-sedimentary recristallization of planktonic foraminifera (Schrag et al., 1995;Pearson et al., 2001).Additionally, reduced pH may have dampened foraminifer δ 18 O excursions, potentially resulting in too low estimates of PETM warming (Uchikawa and Zeebe, 2010).More recently, the application of organic paleothermometers, such as TEX 86 and MBT/CBT in marginal marine sequences provided estimates of absolute temperature evolution across the PETM and other hyperthermals in the Northern Hemisphere (e.g., Sluijs et al., 2006;Zachos et al., 2006;Weijers et al., 2007).These studies showed exceptionally high Arctic temperatures during this time interval, suggesting very low meridional temperature gradients (Sluijs et al., 2006).
The marginal marine sections used in these studies have also revealed significant increases in river discharge and sediment input (e.g., Crouch et al., 2003;Giusberti et Introduction

Conclusions References
Tables Figures
We have generated geochemical and palynological data through upper Paleocenelower Eocene sediments recovered during Ocean Drilling Program (ODP) Leg 189 at Site 1172 on the East Tasman Plateau, deposited at ∼65 • S paleolatitude (Exon et al., 2004) (Fig. 1).Micropaleontological information from the Southwest Pacific showed that this site was located within the Antarctic-derived, northward flowing Tasman Current throughout the Paleogene, which is consistent with general circulation model experiments (Huber et al., 2004).While in an earlier study, based on initial shipboard samples we suggested that the PETM might not have been recovered at this site (R öhl et al., 2004), detailed post-cruise investigations identified a condensed PETM section, on the basis of the CIE.We perform TEX 86 , dinoflagellate cyst assemblage analyses and X-ray fluorescence (XRF) core scanning, in order to reconstruct paleoenvironmental conditions at southern high latitudes across the PETM.

Material
Sediments of late Paleocene and early Eocene age at Site 1172, Hole 1172D, consist mainly of organic-rich green and gray clay-and siltstones with low abundance of calcareous and siliceous microfossils, but high abundance of palynomorphs (notably dinocysts but also terrestrial pollen and spores).the glauconite grains being irregular and angular, which indicates that glauconite was formed in situ (based on thin sections, not shown).Lithological and palynological information, as well as the general absence of calcareous microfossils, suggests an overall very shallow marine depositional setting, to restricted conditions with marked runoff from the nearby shores (Shipboard Scientific Party, 2001;R öhl et al., 2004).

Methods
The archive halves of Core 189-1172D-15R were subject to XRF Core Scanning.Subsequently, half-splits of these archive halves were sampled on a resolution of 1 to 2 cm, after which samples were freeze-dried.Splits of samples were then taken for palynology and organic geochemistry.

X-ray fluorescence (XRF) core scanning
We measured the elemental composition of sediments from Cores ODP 189-1172D-15R to -17R at the MARUM, Bremen University, Germany, using the XRF core scanner (Richter et al., 2006;Tjallingii et al., 2007).The XRF core scanner acquires bulksediment chemical data from split core surfaces.Although measured elemental intensities are predominantly proportional to concentration, they are also influenced by the energy level of the X-ray source, count time, and physical properties of the sediment (R öhl and Abrams, 2000;Tjallingii et al., 2007).XRF data were collected every cm down-core over a 1 cm 2 area using 30 s count time.We used a generator setting of 20 kV and an X-ray current of 0.15 mA.Introduction

Conclusions References
Tables Figures

Back Close
Full

Palynology
Samples of 1-2 cm stratigraphic thickness were freeze-dried and a known amount of Lycopodium spores was added to ∼4 g of material.Samples were then treated with 30% HCl and twice with 38% HF for carbonate and silicate removal, respectively.Residues were sieved using a 15-µm nylon mesh to remove small particles.To break up clumps of residue, the sample was placed in an ultrasonic bath for a maximum of 5 min, sieved again, and subsequently concentrated into 1 ml of glycerine water, of which 10 µl was mounted on microscope slides.Slides were counted for marine (e.g., dinocysts) and terrestrial palynomorphs (e.g., pollen and spores) to a minimum of 200 dinocysts.Marine and terrestrial palynomorph preservation was excellent for all samples.We generally follow dinocyst taxonomy of Fensome and Williams ( 2004), but adapt Sluijs et al. (2009a) for the various spiny peridinioid taxa.Absolute quantitative numbers were counted using the relative abundance of Lycopodium spores (cf., Stockmarr, 1972).

Organic geochemistry
For stable carbon isotope analyses of total organic carbon (δ 13 C TOC ), freeze-dried samples were powdered, treated with 1 M HCl to remove carbonate, centrifuged and the supernatant decanted, followed by two rinses with demineralized water and freezedried again.Residues were analyzed with a Fison NA 1500 CNS analyzer coupled to a Finnigan Delta Plus isotope ration mass spectrometer.Analytical precision and accuracy were determined by replicate analyses and by comparison with in-house standards, and were better than 0.1‰ and 0.1‰, respectively.For biomarker analyses, freeze-dried sediment samples (∼3.5 g dry mass) were extracted with dichloromethane (DCM)/methanol (2:1) using accelerated solvent extraction (Dionex ACE).The extracts were separated by Al 2 O 3 column chromatography using hexane/DCM (9:1, v/v) and DCM/methanol (1:1, v/v) to yield the apolar and polar fractions, respectively.The polar fractions were analyzed using high performance Introduction

Conclusions References
Tables Figures

Back Close
Full liquid chromatography/atmospheric pressure chemical ionization-mass spectrometry, according to Schouten et al. (2007a).Single ion monitoring was used to quantify the abundance of the crenarchaeotal Glycerol Dialkyl Glycerol Tetraether (GDGT) lipids.
The relative abundance of GDGTs were used to calculate TEX 86 (Schouten et al., 2002).TEX 86 is converted to mean annual SST by means of quasi-global core top calibrations.A new calibration with a logarithmic function was recently published (Kim et al., 2010), which is based on the currently available core-top data and a thorough statistical analyses between GDGTs abundances and SST.An earlier calibration assumes a different logarithmic relation (Liu et al., 2009) that produces particularly different SSTs for high TEX 86 values.We also determined the BIT index, which is a ratio between soil bacteria-derived and marine crenarchaeota-derived membrane lipids, a proxy for the amount of river transported soil organic matter versus marine organic matter (Hopmans et al., 2004).

Core depth adjustments
Based on correlations between physical properties data generated on core material and downhole logging, we have slightly changed the meters below sea floor (mbsf) depth of the core sediments (Bijl et al., 2009); we use revised mbsf (rmbsf) for these revised depths throughout.Relative to mbsf, Core 189-1172D-12R was shifted up by 0.36 m; 13R down by 1.87 m; 14R down by 2.84 m; 15R down by 2.4 m; 16R down by 2.57 m and 17R down by 2.66 m.

Stratigraphy
Between 611.89 and 611.86 rmbsf, the δ 13 C TOC curve shows a ∼3‰ negative step from −26 to −29‰, followed by a ∼20 cm interval of relatively stable values, and a subsequent exponential recovery that reaches background values between 611.2 and Introduction

Conclusions References
Tables Figures

Back Close
Full 611.0 rmbsf (Fig. 2).This excursion is located within magnetochron C24r.Average sedimentation rates of 5.7 m/Myrs for this chron (see Material) imply that this excursion occurred ∼1 Ma after the termination of Chron C25n (Fuller and Touchard, 2004;Bijl et al., 2009) 1999), and ∼2 Ma between the PETM and the onset of Chron C24n (Westerhold et al., 2007).Hence, the overall stratigraphic position of the CIE implies the presence of the PETM in Core 1172D-15R.The thickness of the CIE at Site 1172 is 65-90 cm, depending on the definition of its termination (R öhl et al., 2007).Assuming a 170 kyr duration of the CIE (R öhl et al., 2007;Abdul Aziz et al., 2008), this indicates average PETM sedimentation rates of ∼0.4-0.5 cm/kyr, although sediment accumulation rates were likely highly variable in this pro-deltaic setting.BIT values are low throughout the analyzed interval, indicating that TEX 86 values are not influenced by soil derived GDGTs (Weijers et al., 2006).The BIT record exhibits some scatter, but values during the PETM (0.09±0.02) are on average somewhat lower than before and after the PETM (0.13±0.03), suggesting increased marine production of GDGTs or a decreased supply of soil organic matter.Introduction

Conclusions References
Tables Figures

Back Close
Full

Palynology
Palynological assemblages are rich, well preserved and dominated by dinoflagellate cysts (dinocysts).Terrestrial pollen and spores are common to abundant throughout, with relatively high abundances within the PETM (Figs. 2, 3).Stratigraphically important dinocyst taxa include Apectodinium spp., Eocladopyxis peniculatum, Deflandrea spp., Melitasphaeridium pseudorecurvatum, Muratodinium fimbriatum and the recently described species Florentinia reichartii.(Sluijs and Brinkhuis, 2009).In particular the oldest abundant occurrence (>40% of the assemblage) of Apectodinium in the Southwest Pacific Ocean has been calibrated to the PETM (Crouch, 2001).At Site 1172, however, the First Occurrence (FO) of abundant Apectodinium is at ∼612.6 rmbsf, ∼75 cm below the onset of the CIE (Fig. 3).Apectodinium abundances subsequently decrease to ∼2%, followed by a second abundance maximum starting at the onset of the CIE.
Along with Apectodinium spp., other quantitatively significant taxa in the assemblage mostly comprise cosmopolitan taxa such as Senegalinium spp., Glaphyrocysta spp., Eocladopyxis peniculatum, Cordosphaeridium fibrospinosum, Thalassiphora spp., Kenleyia spp., Fibrocysta spp.(and other members of the Cordosphaeridium fibrospinosum complex) (sensu, Sluijs and Brinkhuis, 2009), Diphyes colligerum, Paucisphaeridium, Deflandrea (and a few related Cerodinium), Membranosphaera (often referred to as Elytrocysta in the Southern Ocean), and Spiniferites spp.Hystrichosphaeridium truswelliae, common in certain intervals, was long thought to have been endemic to the Antarctic Realm, but was recently recorded in uppermost Paleocene and PETM sediments on the New Jersey Shelf (Sluijs and Brinkhuis, 2009).In fact, PETM assemblages as a whole are strikingly similar to those reported from the New Jersey Shelf.Only few aspects of the assemblages are typical for the Antarctic Realm (e.g., Wrenn and Beckmann, 1982;Warnaar et al., 2009), including rare Vozzhennikovia spp., and temporally abundant Pyxidinopsis spp.Introduction

Conclusions References
Tables Figures

Back Close
Full Senegalinium spp.dominate assemblages from the base of the studied section (∼615 rmbsf) up to ∼613 rmbsf, an interval with very stable dinocyst assemblages with common Pyxidinopsis, Spiniferites, and Deflandrea spp.Assemblages are slightly richer above ∼613 rmbsf, with more abundant Pyxidinopsis spp.and common H. truswelliae and Membranosphaera spp.A peak in Glaphyrocysta spp.occurs around 613 rmbsf, directly followed by the first Apectodinium acme.Between 612.2 and 611.9 rmbsf, just below the onset of the CIE, successive transient abundances of Glaphyrocysta, Deflandrea, Pyxidinopsis, and Operculodinium spp., and C. fibrospinosum complex occur.A second acme of Apectodinium is recorded concomitant with the CIE.Within the CIE, transient abundances of Glaphyrocysta spp.and Eocladopyxis peniculatum occur.After the CIE, Senegalinium dominates assemblages again, while Operculodinium spp., H. truswelliae and Membranosphaera spp.are common.

XRF
We present high-resolution XRF core scanner data for Core 1172A-15R across the carbon isotope excursion (CIE) of the PETM (Fig. 2).Fe and Ca intensities exhibit a characteristic variability that can be directly attributed to lithology (Fig. 2).The Ca in these sediments is related to carbonate (R öhl et al., 2004).The dominant lithology, organic-rich green and gray clay-and siltstones with low carbonate contents (Shipboard Scientific Party, 2001) is expressed as generally low Ca values in cores below Core 15R (R öhl et al., 2004).Ca values are higher in the lower part of Core 15R, just below the PETM (Fig. 2).The Ca intensities in this interval of Core 15R show regular fluctuations: about four cycles below the onset of the PETM, which are also visible in the BIT index, the TEX 86 sea surface temperature data, and reversely in the percentage of terrestrial palynomorphs.Assuming an average sedimentation rate of 5.7 m/Myr these cycles could represent the low eccentricity frequency of the Milankovitch orbital band (100 kyr cycles).Ca values exhibit a peak during the warming of the PETM, followed by the lowest Ca values in the interval (611.56-610.57rmbsf) (Fig. 2).The Fe Introduction

Conclusions References
Tables Figures

Back Close
Full

The carbon isotope excursion
The ∼3‰ CIE recorded at Site 1172 is slightly smaller in amplitude than the 4-5‰ often recorded in δ 13 C TOC at other marine sites (Bolle et al., 2000;Steurbaut et al., 2003;Sluijs et al., 2006).Since the lowest δ 13 C values for most carbon isotope records across the PETM are located close to the onset of the event (e.g., Bowen et al., 2001;Thomas et al., 2002;Sluijs et al., 2007a), this could imply that the earliest part of the PETM is not represented in our record.However, globally, δ 13 C curves from the PETM indicate a rapid decrease (<10 000 yr), followed by ∼80 kyr of stable carbon isotope values and subsequent recovery, often referred to as the "body" of the CIE (e.g., Thomas and Shackleton, 1996;Bowen et al., 2001;Thomas et al., 2002;Sluijs et al., 2007a;McCarren et al., 2009).Our record clearly shows an interval of stable values of around −29‰ between ∼611.9-611.7 rmbsf, implying that at least part of the stable peak phase is represented in the record.In fact, the magnitude of the CIE in our record is very close to the −3 to −3.5‰ that is generally assumed to have been the excursion in the global exogenic carbon (Zachos et al., 2007;McCarren et al., 2009).Hence, the record at Site 1172 appears to contain at least part of the "body" of the CIE as well as the recovery, allowing comparison to other PETM sites.

Sea surface temperature evolution
The range of SST estimates based on TEX 86 is large due to differences in calibrations; LIU2009 gives relatively low temperature estimates and implies a very low sensitivity Introduction

Conclusions References
Tables Figures

Back Close
Full for TEX 86 values at temperatures >30 • C. In contrast, KIM2010 implies a higher sensitivity and absolute temperature estimates (Fig. 2).In the New Jersey PETM records for example, the KIM2010 is most consistent with mixed layer planktonic foraminifer stable oxygen isotope (δ 18 O) paleothermometry (Kim et al., 2010), also regarding the magnitude of warming.In this upper range of TEX 86 values, we therefore prefer the magnitude of warming implied by the KIM2010 calibration.
A warming of ∼7 • C is similar to or slightly less than the only other Southern Ocean estimates from the Weddell Sea (Sites 689 and 690 at Maud Rise), based on the δ 18 O excursion in the surface dwelling foraminifer Acarinina (Thomas et al., 2002;Zachos et al., 2007).The magnitude of the SST rise is also similar to that recorded at marginal marine sites on the New Jersey Shelf based on foraminiferal δ 18 O and TEX 86 (Zachos et al., 2006;Sluijs et al., 2007b;John et al., 2008).However, the magnitude of warming was smaller in open-ocean and continental settings, and in the Arctic (e.g., Thomas and Shackleton, 1996;Zachos et al., 2003;Tripati and Elderfield, 2005;Wing et al., 2005;Sluijs et al., 2006;Weijers et al., 2007).This suggests that, while the Arctic warmed with a magnitude comparable to the global average (Sluijs et al., 2006), some marginal marine regions warmed slightly more and some polar amplification may have occurred in the Southern Hemisphere.If so, this amplification may have been caused by three mechanisms.First, the melting of small ice sheets on high mountains in Antarctica may have reduced albedo and thus amplified Antarctic warming.This would be consistent with the reconstructed PETM eustatic rise (Sluijs et al., 2008a).Secondly, an increase in atmospheric heat transport may have occurred.Indeed, increased precipitation in Southern Hemisphere PETM records would suggest more latent heat transport from tropical regions to the Antarctic (Robert and Kennett, 1994;Crouch et al., 2003).However, also Arctic sections exhibit evidence for intensified regional hydrology (Pagani et al., 2006), but no amplification of warming is recorded there (Sluijs et al., 2006).Third, a change in ocean circulation during the PETM resulted in regionally enhanced warming in the Southwest Pacific and Weddell Sea.Absolute TEX 86 temperatures are surprisingly high for this latitude, even with the Introduction

Conclusions References
Tables Figures

Back Close
Full • C cooler than on the New Jersey Shelf at a paleolatitude of ∼35-40 • N (Zachos et al., 2006;Sluijs et al., 2007b), and ∼8-12 • C warmer than those in the Arctic (Sluijs et al., 2006) depending on the applied calibration.Post-PETM SSTs are ∼5 • C lower than earliest Eocene estimates from Tanzania at ∼17 • S (Pearson et al., 2007).Although no TEX 86 information is available from equatorial regions for this time period, this supports previous observations (Bijl et al., 2009;Hollis et al., 2009) of a significantly reduced temperature gradient between the Southwest Pacific and at least the subtropics.As indicated for the Northern Hemisphere data (Sluijs et al., 2006(Sluijs et al., , 2007b;;Zachos et al., 2006), the difference in SST of merely 10 • C between Tanzania, New Zealand and Site 1172 cannot be simulated by current-generation fully coupled models (Shellito et al., 2003;Huber and Nof, 2006).However, oxygen isotopes of reputedly Eocene mollusks from coastal Antarctica indicate significantly cooler conditions on the Antarctic shelf (Ivany et al., 2008).This either implies that a very large temperature difference existed between the East Tasman Plateau and coastal Antarctica, or that TEX 86 and mollusk δ 18 O yield significantly different paleotemperature estimates.Interestingly, regardless of the calibration, peak PETM temperatures are similar to those recorded for the Early Eocene Climatic Optimum (EECO, ∼52-50 Ma) at Site 1172 (Bijl et al., 2009).This suggests that atmospheric greenhouse gas levels were comparable during the peak of the PETM and long-term warmth of the EECO.In fact, although regional differences exist, peak PETM temperatures were similar to those during ETM2 (Sluijs et al., 2009b;Stap et al., 2010).If so, one may speculate that the long-term late Paleocene-early Eocene warming trend that culminated in the EECO as well as the superimposed hyperthermals, were caused by carbon injection from the same reservoir.Such a scenario requires a source that slowly added carbon to the global exogenic pool during the long-term trend resulting in the EECO.During the hyperthermals, it must have released carbon catastrophically, perhaps when an orbital (Lourens et al., 2005) threshold was surpassed, followed by partial recharge.One reservoir that may behave like this is the methane hydrate reservoir (Dickens, 2003).Introduction

Conclusions References
Tables Figures

Back Close
Full Although several potential problems exist with methane hydrates as the only source for 13 C-depleted carbon, a long-term net leakage from hydrates during late Paleoceneearly Eocene warming is consistent with a concomitant decrease in deep ocean δ 13 C as observed in benthic foraminiferal calcite.

Leads and lags
The genus Apectodinium originated close to the Danian-Selandian boundary (Brinkhuis, 1994;Guasti et al., 2006) but abundant occurrences were restricted to low latitudes until the PETM (Bujak and Brinkhuis, 1998;Iakovleva et al., 2001).On a global scale, Apectodinium dominates dinoflagellate assemblages described from the PETM (Heilmann-Clausen, 1985;Bujak and Brinkhuis, 1998;Egger et al., 2000;Crouch et al., 2001;Steurbaut et al., 2003;Sluijs et al., 2006Sluijs et al., , 2007a)).At Site 1172, however, the lowermost acme starts approximately 70 cm below the CIE.Average sedimentation rates of 5.7 m/Myr suggest that this acme leads the CIE by some 100 kyr, although sedimentation rates likely varied significantly in this pro-deltaic setting.An influx of abundant Apectodinium has also been shown to lead the CIE on the New Jersey Shelf, the Central North Sea and, perhaps, New Zealand (Sluijs et al., 2007b), by approximately 5 kyr (perhaps slightly longer if sedimentation rates decreased in response to sea level rise (see, Sluijs et al., 2008a).The early Apectodinium acme recorded at Site 1172 can be interpreted in two ways.First, if the uppermost Paleocene record at Site 1172 is relatively expanded, the early acme may actually correlate to the early onset recorded at other sites.Latest Paleocene sedimentation rates of ∼10 cm/kyr are required to support this hypothesis, which is significantly higher than the average across this part of the section at Site 1172, but quite common for marginal marine settings.If so, the rise in TEX 86 around 612 rmbsf might comprise the warming recorded in New Jersey between the onset of the Apectodinium acme and the CIE (Sluijs et al., 2007b), although the latter records do not show a subsequent cooling immediately prior to the onset of the CIE.We cannot exclude this hypothesis because of the poor constraints on sedimentation rates in the uppermost Paleocene part of the Introduction

Conclusions References
Tables Figures

Back Close
Full section.Second, the early Apectodinium acme may imply that conditions at Site 1172 locally became similar to low latitude equatorial environments ∼100 kyr prior to the CIE, but unrelated to the PETM.If so, this would imply extremely anomalous environmental change on the East Tasman Plateau, associated with the first and mass occurrence of a typical low latitude dinoflagellate in the Southern Ocean, which is not accompanied by significant change in other proxies and, critically, not recorded in nearby sections at lower latitudes in New Zealand (Crouch et al., 2001(Crouch et al., , 2003;;Crouch and Brinkhuis, 2005).Although, this hypothesis seems inconsistent with the strictly low-latitude biogeography of abundant Paleocene Apectodinium, we cannot exclude it with the present data.

Depositional environment, hydrology, sea level and productivity
Low sedimentary Ca values are in line with a shallow marine depositional environment, dominated by siliciclastic input.The Ca record shows a peak during the lower part of the PETM.The persistence of carbonate accumulation in this interval indicates that the carbonate saturation state of the bottom waters at this shelf site did not drop below values that would prevent carbonate preservation, as recorded at other shelf locations (e.g., Bolle et al., 2000;John et al., 2008).Hence, the calcite saturation depth and calcite compensation depths in the deep ocean (Zachos et al., 2005), remained deeper than the shelf in the Southwest Pacific Ocean.Representatives of the genus Senegalinium dominate dinocyst assemblages for most of the studied interval (Fig. 3).Dinocysts assignable to this genus have been shown to tolerate very low salinities (Brinkhuis et al., 2006).Furthermore, high Senegalinium abundances have been associated with salinity stratification on the New Jersey Shelf during the PETM (Sluijs and Brinkhuis, 2009).Moreover, Senegalinium likely represents heterotrophic dinoflagellates, thereby thriving in relatively nutrientrich waters (Sluijs et al., 2005).Accordingly, consistent with lithological information, we interpret high abundances of Senegalinium spp.prior to the PETM as to indicate relatively high productive shelf settings, with sustained fresh-water runoff from Introduction

Conclusions References
Tables Figures

Back Close
Full nearby rivers.This interpretation is corroborated by relatively abundant terrestrial palynomorphs (Fig. 3).Although the low BIT values suggest a relatively low contribution of soil derived organic matter, these most likely rather reflect the high burial fluxes of marine organic matter and isoprenoid GDGTs.A decrease in Senegalinium abundance at 613.3 rmbsf is accompanied by a peak in Glaphyrocysta spp., a taxon of which abundant occurrences are often associated with transgressive system trackts and sea level rise (Iakovleva et al., 2001;Pross and Brinkhuis, 2005), suggesting uppermost Paleocene transgression at Site 1172.Generally increasing abundances of other normal marine taxa, such as Pyxidinopsis spp., C. fibrospinosum cpx., Spiniferites spp., and Operculodinium spp.support this interpretation.A second peak in Glaphyrocysta spp. at the onset of the CIE also suggests renewed sea level rise during the early stages of the PETM, supported by dropping BIT index values and a second drop in Senegalinium abundances.This transgression is seen along continental margins on a global scale and has, hence, shown to represent eustatic rise (Sluijs et al., 2008a).
The dinocyst trends across the PETM in the Tawanui Section (New Zealand) are similar to the succession at Site 1172 (Crouch et al., 2001;Crouch and Brinkhuis, 2005).Main differences include the lower numbers of the more nearshore taxa and runoff indicators such as Senegalinium spp., and slightly higher numbers of more offshore taxa such as Spiniferites and Impagidinium spp. at the slope setting of Tawanui.At Tawanui, the combined proxy records where interpreted to reflect increased weathering and continental runoff (Crouch et al., 2003), while at the much more proximal Site 1172, PETM eustatic rise appears to have compensated the possible coeval runoff increase as reflected in the dinocyst assemblages.
A peak of Eocladopyxis spp., a member of the extant family Goniodomidae that mainly inhabits polysaline, lagoonal environments (Wall et al., 1977), occurs within the PETM.Potentially, SSTs were only warm enough for this species to thrive in the Southern Ocean during peak PETM warmth.The ecology of extant Goniodomids provides additional boundary conditions for the local environment.Abundant representatives Introduction

Conclusions References
Tables Figures

Back Close
Full of related species in the modern ocean (e.g., the harmful species Pyrodinium bahamense) have been related to hypersaline conditions (Reichart et al., 2004).Critically, however, in many regions, seasonal storm activity appears important to resuspend dormant cysts into the water column to hatch and fulfill their life cycle (Villanoy et al., 1996(Villanoy et al., , 2006)).In such systems, the subsequent bloom initiates when salinities drop due to increased river run off, and when turbulence is minimal (e.g., Dale, 2001;Siringan et al., 2008).Increased river run off and surface ocean stratification might be induced by tropical storms.Hence, storm activity and seasonal river input might have increased in the Southwest Pacific region during the PETM, consistent with increased abundances of terrestrial palynomorphs.In any case, a peak in Goniodomids has at Site 1172 only been recorded during the PETM (Fig. 3) and the EECO (Brinkhuis et al., 2003;Sluijs et al., 2003;Bijl, 2007), indicating a very particular environment for this region, likely associated with a change in seasonality of regional hydrology, maximum temperatures and perhaps storm activity.
The dinocyst record suggests relatively stable conditions through the latest Paleocene and some more variation close to the onset of the CIE and within the PETM, with short-lived abundances of Glaphyrocysta, Eocladopyxis, Pyxidinopsis, Cordosphaeridium fibrospinosum complex, Spiniferites, Operculodinium, and Membranosphaera.Such intra-PETM variability has also been recorded in continental deposits from Wyoming (Bowen et al., 2004;Wing et al., 2005;Kraus and Riggins, 2007) and on the New Jersey Shelf (Sluijs and Brinkhuis, 2009).Although at the moment the cause of these variations are unknown, they do suggest that climate during the PETM may have been dynamic, perhaps on a global scale.Full   (Kim et al., 2010) and LIU2009 (Liu et al., 2009); BIT index; Apectodinium percentage of the dinocyst assemblage; abundance of terrestrially derived palynomorphs as a percentage of total palynomorphs; and XRF intensity data for iron and calcium.
Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | al.
Glauconite and accessory minerals are recorded in varying abundance (Shipboard Scientific Party, 2001), with Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | 86 and BIT Late Paleocene SSTs average ∼26 • C (1σ=0.9) based on TEX 86 , regardless of the applied calibration (Fig. 2).Concomitantly with the onset of the CIE, SSTs rise to average PETM values of ∼31 • C (1σ=0.7) following KIM2010, or ∼29 • C for LIU2009 with peak values of almost 33 • C and 31 • C for the two calibrations, respectively, at 611.70 rmbsf.The magnitude of PETM warming was thus ∼7 • C with KIM2010 and 4 • C with LIU2009.SSTs returned to pre-excursion values during the recovery of the CIE.The warming and the CIE are preceded by two samples with relatively low temperatures (∼25 • C).
Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | and Ca intensities are overall anti-correlated.The Fe record exhibits maximum values (broad peak) in the upper part of the CIE, where the Ca values are low and the terrestrial palynomorphs show maximum values.
Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | conservative LIU2009.Pre-PETM SSTs of ∼26-27 • C and PETM SSTs of ∼30-34 • C are approximately 3-6 Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | complete PETM record was identified in sediments recovered from the East Tasman Plateau during ODP Leg 189, deposited at a paleolatitude of ∼65 • S. Sediments are almost devoid of biogenic calcite but yield rich organic microfossil Discussion Paper | Discussion Paper | Discussion Paper |

Figure 1 .
Figure 1.Site location and surface currents.Early Eocene Paleogeographic reconstruction for the Antarctic.Surface circulation (Huber et al., 2004) indicates the Antarctic-derived Tasman Current over the East Tasman Plateau, which is supported by biogeographical data.Paleogeographic chart obtained from the Ocean Drilling Stratigraphic Network (ODSN) and was modi ed from Bijl et al. (2009).The dashed red arrow around New Zealand indicates potential mixing of low-latitude surface waters (from the East Australian Current) with the Tasman Current.

Fig. 1 .Figure 2 .
Fig. 1.Site location and surface currents.Early Eocene Paleogeographic reconstruction for the Antarctic.Surface circulation (Huber et al., 2004) indicates the Antarctic-derived Tasman Current over the East Tasman Plateau, which is supported by biogeographical data.Paleogeographic chart obtained from the Ocean Drilling Stratigraphic Network (ODSN) and was modified from Bijl et al. (2009).The dashed red arrow around New Zealand indicates potential mixing of low-latitude surface waters (from the East Australian Current) with the Tasman Current.

Fig. 2 .Figure 3 .Fig. 3 .
Fig. 2. Organic geochemical, palynological and XRF results across the PETM of Site 1172.From left to right: stable carbon isotope (δ 13 C) values of total organic carbon (TOC) relative to the Vienna Pee Dee Belemnite standard.Error bars reflect duplicate-based standard deviations and three grey data points are considered outliers because duplicate analyses indicated values consistent with surrounding samples; sea surface temperature (SST) based on TEX 86 following the calibrations KIM2010(Kim et al., 2010) and LIU2009(Liu et al., 2009); BIT index; Apectodinium percentage of the dinocyst assemblage; abundance of terrestrially derived palynomorphs as a percentage of total palynomorphs; and XRF intensity data for iron and calcium.
and ∼2 Ma between this CIE and the onset of Chron C24n.The orbitally based age model from Blake Nose (ODP Leg 171B) and Walvis Ridge (ODP Leg 208) also indicates ∼1 Ma between the top of Chron C25n and the PETM (Norris and R öhl, assemblages.TEX 86 paleothermometry indicates that SSTs warmed from ∼27 to 33 • C during the PETM, with a magnitude similar to or slightly larger than the global estimate of warming.Such surprisingly warm SSTs for this latitude support that meridional temperature gradients were very low across the Paleocene-Eocene transition.Maximum temperatures were similar to those during the EECO, perhaps implying similar greenhouse gas concentrations.If so, one may speculate that long-term late Paleocene to early Eocene warming, carbon isotope trends and superimposed hyperthermals, were associated with carbon release from the same reservoir, perhaps methane hydrates.The globally recorded acme of the taxon Apectodinium leads the CIE at Site 1172, which may represent the same early onset as recorded on the New Jersey Shelf and the North Sea.A decrease in the abundance of the fresh water-tolerant dinoflagellate cyst Senegalinium suggest a decrease in the influence of river run off at the core site during the PETM, possibly in concert with eustatic rise.However, a unique abundance of the euryhaline taxon Eocladopyxis may indicate a change in the seasonality of the regional hydrological system and an increase in storm activity.Finally, significant variation in assemblages within the PETM suggests that the climate state was relatively dynamic during this event.Introduction