Stratigraphic calibration of Oligocene–Miocene organic-walled dinoflagellate cysts from offshore Wilkes Land, East Antarctica, and a zonation proposal

There is growing interest in the scientific community in reconstructing the paleoceanography of the Southern Ocean during the Oligocene–Miocene because these time intervals experienced atmospheric CO2 concentrations with relevance to our future. However, it has remained notoriously difficult to put the sedimentary archives used in these efforts into a temporal framework. This is at least partially due to the fact that the bioevents recorded in organic-walled dinoflagellate cysts (dinocysts), which often represent the only microfossil group preserved, have not yet been calibrated to the international timescale. Here we present dinocyst ranges from Oligocene–Miocene sediments drilled offshore the Wilkes Land continental margin, East Antarctica (Integrated Ocean Drilling Program (IODP) Hole U1356A). In addition, we apply statistical means to test a priori assumptions about whether the recorded taxa were deposited in situ or were reworked from older strata. Moreover, we describe two new dinocyst species, Selenopemphix brinkhuisii sp. nov. and Lejeunecysta adeliensis sp. nov., which are identified as important markers for regional stratigraphic analysis. Finally, we calibrate all identified dinocyst events to the international timescale using independent age control from calcareous nanoplankton and magnetostratigraphy from IODP Hole U1356A, and we propose a provisional dinoflagellate cyst zonation scheme for the Oligocene–Miocene of the Southern Ocean.


Introduction
There is a growing need to better understand the dynamics of the Antarctic cryosphere and the paleoceanography of the Southern Ocean during the Oligocene and Miocene, particularly in view of the apparent similarity between Oligocene and Miocene atmospheric CO 2 concentrations and those of the near future (e.g.Zhang et al., 2013).The vast majority of our understanding of Oligocene-Miocene ice-sheet dynamics and paleoceanography in the high southern latitudes is derived from marine sedimentary archives.A major challenge when employing these archives is to establish accurate age control.Many sedimentary successions from the Southern Ocean lack the calcareous microfossil groups conventionally used for biostratigraphic age control, which stresses the need to utilise the stratigraphic potential of non-calcareous microfossils.Siliceous microfossils have provided an excellent stratigraphic framework for the Neogene of the Southern Ocean (Cody et al., 2008); however, they are prone to dissolution when buried below the diagenetic front of opal, which tends to limit their applicability to deeply buried strata (De-Master, 2003).
For successions from the southern high latitudes, organic microfossils have repeatedly been shown to represent a use-Published by Copernicus Publications on behalf of The Micropalaeontological Society.P. K. Bijl et al.: Oligocene-Miocene Southern Ocean dinocyst stratigraphy ful biostratigraphic tool (e.g.Brinkhuis et al., 2003a, b;Sluijs et al., 2006;Tauxe et al., 2012;Pross et al., 2012;Houben et al., 2011Houben et al., , 2013;;Stocchi et al., 2013).Recently, a calibration of Paleocene and Eocene dinocyst origination and extinction events to the geomagnetic polarity timescale (GPTS) has provided the potential for improved age control in the Southern Ocean for this part of the stratigraphic column (Bijl et al., 2013b(Bijl et al., , 2014)).In contrast, a proper magnetostratigraphic calibration of organic microfossil events in the Southern Ocean for the Oligocene and Miocene epochs is still missing.This may be partially ascribed to the fact that Oligocene-Neogene dinocyst assemblages from the Southern Ocean (i) are of relatively low diversity, (ii) are dominated by stratigraphically long-ranging species (Escutia et al., 2011), (iii) contain a considerable number of formally un-described taxa (Escutia et al., 2011), and (iv) include endemic species that are highly specialised to the (paleo)oceanographic conditions prevailing around Antarctica (Houben et al., 2013).The latter means that the distribution of many dinocyst taxa found in Southern Ocean sediments is restricted to that region (e.g.Zonneveld et al., 2013;Houben et al., 2013).These typical Southern Ocean assemblages were essentially established during the Eocene-Oligocene transition, in conjunction with the onset of major Antarctic cryosphere growth (Houben et al., 2013).As a consequence of the abovementioned characteristics, these dinoflagellate cyst assemblages have yet yielded virtually no bio-events that could be calibrated to the internationally used geologic timescale (Gradstein et al., 2012).
Studies on Oligocene-Neogene dinocyst assemblages have been carried out on sediment cores from Prydz Bay (Hannah, 2006;Warnaar, 2006;Houben et al., 2013), Cape Roberts (Hannah et al., 1998(Hannah et al., , 2000(Hannah et al., , 2001a, b), b), within the Cenozoic Investigation in the Western Ross Sea (CIROS) program (Hannah, 1994(Hannah, , 1997)), the Antarctic Geological Drilling (ANDRILL) program in the Ross Sea (Warny et al., 2009), and Ocean Drilling Program (ODP) Site 696 in the Weddell Sea (Warnaar, 2006;Houben et al., 2013).Further north in the Southern Ocean, dinocyst records have been generated from ODP Site 1168 (Brinkhuis et al., 2003a) and Deep Sea Drilling Project (DSDP) Site 511 (Goodman and Ford Jr., 1983;Houben et al., 2011).Efforts to accurately calibrate the dinocyst events encountered in these studies to existing timescales were not only hindered by taxonomical challenges, but also by other factors such as a lack of independent age control or relatively short and/or incomplete sedimentary archives.A recent revisit of sedimentary records from the Cape Roberts Project (CRP) has yielded formal descriptions of a number of dinocyst species, the ranges of which have been stratigraphically calibrated to some extent (Clowes et al., 2016).
Clearly, a proper stratigraphic calibration of Oligocene-Miocene dinocyst events in the Southern Ocean could provide age constraints for many sedimentary records that have yet remained stratigraphically poorly dated and would also provide an opportunity to date sediments to be recovered  2012) and absolute plate positions of Torsvik et al. (2012).
during future drilling campaigns (e.g.McKay et al., 2016).In 2010, IODP Expedition 318 drilled a series of sedimentary archives off the Wilkes Land margin, East Antarctica (notably Site U1356; Fig. 1).In this paper, we present and describe dinocysts encountered in the Oligocene and Miocene part of the succession at Site U1356, with the aim of tying their distribution to the existing integrated biomagnetostratigraphic age model (Tauxe et al., 2012).We critically evaluate the likelihood that certain species are reworked from older strata or deposited in situ using multivariate statistical analysis.We formally describe two new species, which we consider stratigraphically useful, while we informally report 23 other new taxa.Finally, we establish a first dinocyst zonation scheme for the Oligocene-Miocene of the Southern Ocean.

Site description and lithology of IODP Hole U1356A
This research uses samples from IODP Hole U1356A, which was drilled on the base of the continental rise of the Wilkes Land continental margin (Fig. 1a; present latitude 63 • 18.6 S, 135 • 59.9 E; Escutia et al., 2011).From early Oligocene (∼ 34 Ma) to middle Miocene (∼ 10 Ma) times, the site has moved from a paleolatitude of 59.8 ± 4.8 • to 61.5 ± 3.3 • S respectively (Van Hinsbergen et al., 2015;Fig. 1).Oligocene and Miocene sediments were recovered between 890 and 3 m below sea floor (m b.s.f.; Fig. 2).Core recovery was poor for the uppermost 95 m of Hole U1356A, as a consequence we focus our analyses on the interval between 95.4 and 894 m b.s.f.(Cores 11R to 95R Section 3).
Sediments consist of alternations of laminated and bioturbated siltstones, occasional claystones and coarser beds, the latter of which are occasionally deformed.Such deposits indicate contourite deposition (Escutia et al., 2011).Occasionally, carbonate is so abundantly present that limestone beds were formed.Diatoms are abundant down-core to about 400 m b.s.f.but are not preserved below this level.The presence of calcareous nanoplankton indicates that the carbonates were derived from a pelagic, biogenic source.Outsized clasts are confined to specific depth intervals.These intervals are 890-875, 735-675, 580-450 and 200-150 m b.s.f. (Escutia et al., 2011).

Bio-magnetostratigraphic age model for IODP Hole U1356A
Stratigraphic constraints for the Oligocene-Miocene succession from IODP Hole U1356A comprise calcareous nanoplankton, radiolarian, diatom, sparse dinocyst biostratigraphy and magnetostratigraphy (Tauxe et al., 2012;Fig. 2).The work of Tauxe et al. (2012) was calibrated to the Gradstein et al. (2004) timescale.Here we update the age model of the succession to the GPTS of Gradstein  (2012), in which black is normal polarity, white is reversed, lilac is uncertain polarity and grey is no data.Age model has been (re-)calibrated to GTS2012 of Gradstein et al. (2012); see text and Table 1.
et al. (2012; Table 1).Furthermore, the age model for the Miocene part of the succession had been updated by Crampton et al. (2016) using constrained optimisation of available stratigraphic datums (CONOP; see also Sangiorgi et al., 2017); here we follow this updated age model for the Miocene part of the succession.These revisions suggest that the sediments between 895.4 and 95.4 m b.s.f.(i.e.Cores U1356A-95R-3w, 82 cm to U1356A-11R) were deposited between the earliest Oligocene (33.7 Ma) and early late Miocene (10.7 Ma).A 14 Myr long hiatus at 895.4 m b.s.f.spans the mid-Eocene to earliest Oligocene (46-33.7 Myr).Crampton et al. (2016) identified a hiatus between Cores 14R and 15R that was previously unaccounted for.Tauxe et al. (2012) interpreted a hiatus between Cores 46R and 47R, lasting from ∼ 23.0 to 17.0 Ma.Above Core 45R, high-resolution diatom biostratigraphy suggests that the succession is nearly continuous (Tauxe et al., 2012).However, Cores 45R-47R yield insufficient stratigraphic information to confidently assign an age.Our study presents high-resolution dinocyst data, identification of new stratigraphic markers and a revisit of the existwww.j-micropalaeontol.net//37/105/2018/

Palynological sample processing
In total, 272 samples were processed and studied for palynology.Sample resolution varies between 20 cm in Core 84R-85R to over 9 m in intervals of low core recovery (Fig. 2).Palynological processing of freeze-dried, crushed and weighed samples (13 g on average, with a standard deviation of 3.4 g) involved decalcification overnight with 30 % hydrochloric acid (HCl), followed by decanting, rinsing with water and centrifuging (2100 rotations per minute (rpm) for 5 min).The decanted residue was further processed by adding 38 % hydrofluoric acid (HF) to remove silicates.After completion of the chemical reaction with HF, samples were placed on a shaker table for about 2 h, and subsequently were allowed to settle overnight.An excess of 30 % HCl was added to remove fluoride gels, after which samples were centrifuged (2100 rpm for 5 min) and decanted.The entire HF step was repeated for full digestion of silicates.Organic residues were sieved over a 250 µm sieve to remove large palynoclasts, and the supernatant was sieved over a 10 µm sieve, with help of an ultrasonic bath to break and clean up organic residue.The remaining residues were mounted on glass microscope slides with glycerine jelly, and cover slips were sealed with nail polish.All palynological residues and microscope slides are stored in the collection of the Laboratory of Palaeobotany and Palynology of Utrecht University, the Netherlands.2 for a species list).Of particular stratigraphic importance are the numerous species and morphotypes of the genus Lejeunecysta encountered, many of which have recently been formally described from Cape Roberts, Ross Sea (Clowes et al., 2016).At least 25 new dinocyst taxa were recognised in our study, two of which we consider key species for stratigraphy and formally describe herein.The remaining 23 species are informally described (Table 3).The nomenclature of dinocyst features follows Evitt (1985).

Statistical approach for defining reworked versus in situ dinocysts
A prominent feature in the palynological associations from the Oligocene and Miocene of IODP Site U1356 is the abundance of dinocysts that were initially assumed reworked from older strata.Reworked Eocene taxa dominate the palynological assemblages in the lowermost 30 m of the Oligocene (Cores 95R to 93R; Houben et al., 2013).In the lower Oligocene, the percentage of reworked elements decreases to around 10-15 % of the total palynomorph count (Houben et al., 2013).Dinocyst studies of Oligocene and younger sediments in the circum-Antarctic area are scarce, which makes the assumption that certain taxa recorded in the Oligocene Species that are a priori assumed to be reworked in are in red; species that are a priori assumed to be in situ are in blue (see Table 2 for species list, which species are assumed reworked, eigenvalues, axes scores and cumulative % of variance explained).Dot sizes indicate the total number of encounters in our analyses and gives an indication of the importance of that species in our data.
strata in our record are reworked uncertain.For instance, at IODP Site U1356 Enneadocysta dictyostila and Deflandrea antarctica are recorded throughout the lower Oligocene (Houben et al., 2013).Published stratigraphic last occurrences of these species in the Southern Ocean are ∼ 32 and ∼ 28 Ma, i.e. 1.5 and 4.5 Myr after the onset of the Oi-1 glaciation respectively (Brinkhuis et al., 2003a, b;Williams et al., 2004;Escutia et al., 2011).This suggests that both species may be deposited in situ in the lower Oligocene of IODP Hole 1356A.However, these last occurrences can be questioned given that rapid glaciation in the early Oligocene led in most places to a substantial sea-level drop (> 75 m equivalent sea level; see e.g.Houben et al., 2012, and references therein).This likely promoted reorganisation of continental shelves, erosion, and therewith a high likelihood of submarine reworking of Eocene shelf sediments into Oligocene deposits (see e.g.Pross et al., 2010).Indeed, these species are highly abundant in Eocene deposits throughout the Southern Ocean (Truswell, 1982;Warnaar et al., 2009;Bijl et al., 2011Bijl et al., , 2013a, b), b), and today Eocene shelf deposits are exposed on the Wilkes Land margin (Escutia et al., 2011).
To critically assess whether any dinocyst taxa recorded in the Oligocene-Miocene strata are reworked or in situ, we have performed a principal component analysis (PCA; ter Braak, 1986) on the quantitative dinocyst distribution data, using the C2 software program (Juggins, 2007), using square-root transformation.The reason why we hypothesise that we can identify reworking with PCA is that we assume reworked species to have co-varying abundances in the record, which divert from the pelagic, in situ species variance.At the particular setting of IODP Site U1356 at the transition between the continental rise and the abyssal plain (Escutia et al., 2011), reworked species are brought to the site with bottom currents originating from the continental shelf.Higher continental shelf erosion would therefore increase the abundance of all reworked species relative to the abundance of in situ species.This would cause the reworked species to cluster together in PCAs.The variance distribution can therefore be used to verify our a priori assumptions of reworking.

Identification of reworking based on principal component analysis
Of the 167 dinocyst species identified, 67 have positive scores on the first two PCA axes (Table 2; Fig. 3).Of these 67 species, we have a priori inferred 57 (85 %) of them to be reworked (Table 2 and Fig. 3).We do note that some of these species (e.g.Spiniferites ramosus, Dapsilidinium pastielsii) are actually extant species now found in lowerlatitude environments (Zonneveld et al., 2013;Mertens et al., 2014).Given their relatively high abundance in Eocene sediments at the Wilkes Land margin (Bijl et al., 2013a), and their absence in these Southern Ocean high-latitude settings at present (Prebble et al., 2013;Zonneveld et al., 2013), we consider these species to be reworked from the Eocene.This also makes our inferences of labelling species in situ as conservative as possible.Apectodinium spp., Charlesdowniea spp., Glaphyrocysta pastielsii and Membranophoridium perforatum are stratigraphically confined to the early Eocene (Williams et al., 2004;Bijl et al., 2011Bijl et al., , 2013a, b) , b) and are hence obviously reworked.We observe a strict clustering of taxa that were a priori considered reworked, i.e. with positive scores on both axes, having low covariance with undoubtedly in situ taxa.This allows us to infer that the PCA covariance is to a substantial degree related to reworking of dinocysts.However, we infer 11 species with positive scores on the first two PCA axes (i.e. that cluster together with the reworked species) to be in situ: Batiacasphaera compta, Cleistosphaeridium sp.A, Impagidinium elegans, I. patulum, I. paradoxum, Lejeunecysta adeliensis sp.nov., Palaeocystodinium golzowense, Reticulatosphaera actinocoronata, Stoveracysta kakanuiensis and Batiacasphaera sp.D. Apart from Cleistosphaeridium sp.A and the newly described species Lejeunecysta adeliensis, these taxa have been repeatedly encountered in sediments straddling the Eocene-Oligocene boundary sediments (e.g.Sluijs et al., 2003;Eldrett et al., 2004;Egger et al., 2016).We report these taxa in low abundance (Table 2) in our record, predominantly within the interval characterised by a high degree of reworking in the lowermost Oligocene.This may explain the co-variance with evidently reworked species, as the reworking is most severe in the lower part of the Oligocene.Of the 100 species that do not have positive scores on the first two PCA axes, we would a priori infer 10 (10 %) a priori to be reworked: Alisocysta circumtabulata, Cordosphaeridium funiculatum, Dapsilidinium pastielsii, Deflandrea antarctica, Diphyes colligerum, Enneadocysta dictyostila, E. multicornuta, Histiocysta palla, Manumiella druggii, Odontochitina spp.and Spinidinium schellenbergii.The ranges of Manumiella druggii and Odontochitina spp.are restricted to the Cretaceous or lowermost Paleocene (Williams et al., 2004) and therefore are beyond doubt reworked in these Oligocene strata.Whether Deflandrea antarctica and E. dictyostila/multicornuta are reworked remains questionable because many sites show these species range into the Oligocene.However, the dominance of particularly these two species in Eocene records from the Southern Ocean (Brinkhuis et al., 2003b;Bijl et al., 2010Bijl et al., , 2013b) ) suggests that they are reworked in the entire Oligocene-Miocene sequence at Site U1356.Despite these remaining uncertainties, our statistical analysis provides reasonably solid confirmation of our a priori inferences of reworking in our record.

Dinocyst taxonomy: formal description of two new species
Numerous previously unknown dinocyst species were encountered (Tables 2, 3; Fig. 4).Here we formally describe the new species Lejeunecysta adeliensis sp.nov., which we compare with the other species of Lejeunecysta (Table 4, Fig. 5) that have recently been described from the CRP drill cores (Clowes et al., 2016) and were also recognised at Site U1356 (Escutia et al., 2011; see Fig. 5 in this paper).Here we also formally describe a stratigraphically important new species: Selenopemphix brinkhuisii sp nov., the first occurrence of which defines the base of one of the dinocyst zones proposed here.The range top of this species is close to the top of magnetochron C12n.We also provide key diagnostic features for 23 other taxa (Table 3); the formal description of these other species will be published at a later stage.

Systematic paleontology
Division Dinoflagellata Bütschli 1885 (Fensome et al., 1993) Subdivision Description.An acavate cyst with a strong pentagonal outline.Both the epicyst and the hypocyst widen strongly towards the cingulum.The phragm is thin (< 1 µm) and psilate.The apex has straight lateral sides extending to a solid apical horn tip.The antapex is characterised by two separate, relatively long, slender horns that distally taper into a pointed tip.The wall of these pointed tips is thicker, resulting in solid ends of the horns.The archeopyle involves an iso-to slightly lati-deltaform second intercalary plate (2a) that is often adnate.The tabulation is only indicated by the archeopyle.The cingulum is marked by a set of parallel flanges along the lateral sides of the cyst, yet is absent on the dorsal and ventral sides.The cyst widens strongly from the anterior and posterior ends towards the cingulum, almost to form lateral horns.The sulcus is indiscernible on the specimens observed.Geographic distribution.To date, Lejeunecysta adeliensis sp.nov.has only been recognised in the type section.
Affinities.Lejeunecysta adeliensis sp.nov. is different from all other Lejeunecysta species found at Site U1356 (see, e.g.Fig. 5).It most closely resembles L. katatonos, but differs from that species by having antapical horns that are completely separated rather than closely connected as in L. katatonos.The antapical horns of L. adeliensis sp.nov.are similar to those of L. pulchra, but the taxa differ in Depth (m b.s.f.)  Diagnosis.Species of Selenopemphix featuring crests on the longitudinal cingular sutures.On these septa as well as on some other sutures, numerous short, slender, conical spines arise, with pointed tips.The sutures between the precingular and cingular plates as well as those between the cingulum and the postcingular plates bear short (2-4 µm) crests along which the spines are aligned.These septa seem interrupted at least, but not exclusively at the cingular plate boundaries.The sulcus is indented, giving the cyst outline in polar view the shape of a bent ellipse.
Stratigraphic distribution.In the type section, Selenopemphix brinkhuisii occurs between mid-C13n and mid-C12n.This equates to between 33.4 and 30.8Ma.
Geographic distribution.To date, Selenopemphix brinkhuisii sp.nov.has only been recognised in the type section.

Improved bio-magnetostratigraphic age model
Our high-resolution dinocyst biostratigraphy (Fig. 4) generally confirms the published age model for the Oligocene and Miocene interval of Hole U1356A (Tauxe et al., 2012); we have refined the age control around the Oligocene-Miocene boundary.New dinoflagellate cyst events recognised in this study enabled the identification of the position and constrained the duration of the hiatus around the Oligocene-Miocene boundary.Specifically, the presence of Edwardsiella sexispinosa in samples from Cores 47R and 46R (Fig. 4) confirms that these cores are close to the Oligocene-Miocene boundary: at nearby Ocean Drilling Program (ODP) Site 1168, the first stratigraphic occurrence (FO) of E. sexispinosa is around 22. 5 Ma (Brinkhuis et al., 2003a).The FO of E. sexispinosa was detected in sediments that still contain late Oligocene calcareous microfossils (Tauxe et al., 2012): the last stratigraphic occurrence (LO) of Reticulofenestra bisecta in Core 46R (22.97 Ma) and Globigerina euapertura (23.0 Ma) in Core 45R suggest that E. sexispinosa has an older FO on the Antarctic margin than in southern Australia.In turn, we suggest that the reversed magnetic polarities in Cores 45-46R correlate to the youngest Oligocene reversed intervals: C6Cn.2r and C6Cn.1r (Brinkhuis et al., 2003a).The normal polarity at the top of Core 46R (Fig. 4) then correlates to C6Cn.3n.We then extrapolate this to Core 45R, which has normal magnetic polarity and, based on the absence of typical Miocene marker species, should correlate to C6Cn.2n: the Oligocene-Miocene boundary.Our interpretation slightly differs from the interpretation published in Tauxe et al. (2012), which infers the onset of C6Cn.2n in Core 46R and the hiatus to be somewhere between 432 and 402 m b.s.f.(Cores 45R-43R; Tauxe et al., 2012).Our correlation of high-resolution dinocyst biostratigraphy and the available paleomagnetic signal indicates that the hiatus lies between Cores 44R and 45R, and Core 45R falls within the Mi-1 glaciation (Fig. 2).

Provisional dinocyst zonation scheme
We have identified 13 key (regional) dinocyst extinction or origination events (bold in Figs. 4, 6) from all the in situ dinocysts (Fig. 4).We prioritised those dinocyst events that had relatively high abundances, had sharp appearances or extinctions and can be calibrated well to magneto-subchrons.These index events form the basis for the Oligocene-Miocene Southern Ocean dinocyst zonation proposed here (see Table 5 for an overview of zone boundary definitions), which comprises 10 zones for the Oligocene and 3 zones for the Miocene.We name the zones as Southern Ocean Oligocene Dinocyst Zone (SODZ) or Southern Ocean Miocene Dinocyst Zone (SMDZ).We acknowledge that the zonation is incomplete due to a hiatus covering the early  and one encompassing the mid-Miocene (13-10.8Ma) and the occasional low core recovery.We calibrate the ages of the zone boundaries to the GPTS by indicating foremost where in existing magnetostratigraphic chrons the zone boundaries occur.This is consistently indicated as a percentage into a certain chron, measured from the bottom, and by assuming linear sedimentation rates between reversals in our record.

Zone SODZ1
Base definition.The base of this zone is defined by the FO of Malvinia escutiana.
Top definition.The top of this zone is defined by the base of SODZ2, i.e. the FO of Selenopemphix brinkhuisi, sp.nov.

Rupelian
Calibration.The base of this zone might be included in a 14 Myr hiatus at the type locality.At DSDP Site 511, the FO of Malvinia escutiana is correlative with the Oi-1 isotope excursion (Houben et al., 2011), thus occurring in the earliest Oligocene (e.g.Coxall and Wilson, 2011).On the basis of this calibration, we calibrate the base of Zone SODZ1 to the Oi-1 isotope excursion.The top of the zone is calibrated to ∼ 95 % in C13n.
Age. Remarks.This dinocyst zone covers the onset and first 1 million years of glaciation of Antarctica.The abundance of reworked (middle to late) Eocene dinocysts in the Oligocene sequence suggests that the time interval represented in this dinocyst zone is characterised by a relatively rapid (∼ 200 kyr) glacial advance followed by a gradual retreat of the ice sheet.

Zone SODZ2
Base definition.The base of this zone is defined by the first stratigraphic occurrence of Selenopemphix brinkhuisii sp.nov.
Top definition.The top of this zone is defined by the base of SODZ3, i.e. the FO of Lejeunecysta attenuata.
Calibration.The base of this zone is calibrated to 95 % in C13n; the top is calibrated to the 50 % in C12r.
Characteristic species.This interval is characterised by a dominance of Lejeunecysta spp.and Brigantedinium spp., while Selenopemphix spp.and Malvinia escutiana decline in abundance.
Significant events.This zone includes the FO of Operculodinium eirikianum.
Remarks.This zone correlates to a time interval of further warming and Antarctic ice retreat, as interpreted from the benthic foraminiferal oxygen isotope compilation of Pälike et al. (2006).Of particular interest is the near disappearance Selenopemphix antarctica because today high abundances of this species are strongly tied to the sea-ice ecosystem of coastal Antarctica (Prebble et al., 2013).

Zone SODZ3
Base definition.The base of this zone is defined by the first stratigraphic occurrence of Lejeunecysta attenuata.
Top definition.The top of this zone is defined by the base of SODZ4, i.e. the FO of Corrudinium labradori.
Calibration.The base is calibrated to 50 % in C12r; the top of the zone is calibrated to the C12r-C12n boundary.
Characteristic species.This zone is characterised by the abundance of Lejeunecysta spp.Furthermore, Selenopemphix brinkhuisii sp.nov. is common in this interval.
Significant events.This zone sees the FO of several new Batiacasphaera and Protoperidinioid species, Hystrichokolpoma bullatum, and Operculodinium sp.cf.eirikianum.

Zone SODZ4
Base definition.The base of this zone is defined by the FO of Corrudinium labradori.Calibration.The base of the zone is calibrated to the C12n-C12r boundary, while the top is calibrated to 50 % between the top of C12n and the base of C11n.1n.
Characteristic species.The top of the zone sees a rapid increase in various morphotypes of protoperidioid dinocysts that cannot be placed within Brigantedinium spp., Lejeunecysta spp. or Selenopemphix spp.Almost throughout the zone, many specimens of Brigantedinium spp.have a discontinuous, transparent, perforated second outer wall layer (Plate 3G-I).
Significant events.This zone includes the range of Oligokolpoma galeottii, the LO of Phthanoperidinium amoenum, Protoperidinium sp.C, and Corrudinium sp.A, and the FO of Corrudinium sp.A, C. devernaliae and cavate Brigantedinium.

Zone SODZ5
Base definition.The base of this zone is defined by the LO of Hystrichokolpoma bullatum.
Top definition.The top of this zone is defined by the base of SODZ6, i.e. the LO of Lejeunecysta katatonos.
Calibration.The base of this zone is calibrated to 50 % between the top of C12n and the base of C11n.1n, the top to the base of C11n.1n.
Characteristic species.This zone sees the first pulse of dominance of new Protoperidinium taxa that cannot be placed in the more common existing genera Lejeunecysta, Selenopemphix spp.and Brigantedinium spp.
Remarks.Tectatodinium pellitum is considered a thermophilous species in the North Sea area in the Pliocene (Head, 1994).Calibration.The base of this zone is calibrated to 80 % in C6Cn.2r.The top of this zone is calibrated to 80 % in C5Cr.
Age. 23.0-17.1 Ma.The time interval represented in the hiatus is discussed above, under Sect.4.2.We chose to have this zone encompass the early Miocene hiatus in the type section.We hope that other sedimentary records can provide a more complete early Miocene sedimentary record by which zonations for this time interval can be proposed in order to improve the dinocyst zonation scheme.
Characteristic species.Brigantedinium spp., Operculodinium spp., Batiacasphaera spp., Nematosphaeropsis labyrinthus and Impagidinium spp.Calibration.The base of this zone is calibrated to 75 % in C5Bn.2n.The top of this zone is calibrated to 33 % between the FO of the diatom species Actinomma golowini and the FO of Denticulopsis praedimorpha, in the absence of a proper paleomagnetic signal in the core (Tauxe et al., 2012).
Age. 15.1 Ma to an undefined end age.
Characteristic species.This zone is characterised by alternating high abundances of the combination of Brigantedinium spp. with Selenopemphix spp.and Nematosphaeropsis labyrinthus with Operculodinium spp.Towards the top of this zone, a high abundance of Brigantedinium simplex, B. pynei and Selenopemphix nephroides is observed.

Correlation of stratigraphic ranges of selected non-endemic dinocysts
Although the majority of dinocyst taxa we find in our record comprise species that are endemic to the Southern Ocean, or are long-ranging and cosmopolitan, some species have wellcalibrated stratigraphic ranges elsewhere.We list a selection of these species in Table 6.The consistency between our dinocyst stratigraphy and those of others lends some support to our dinocyst zonation and ties our zonation at least to some extent to other dinocyst stratigraphies.However, we are unable to tie our dinocyst events very precisely to other stratigraphies because many magnetic reversals in the record at U1356 fall within core gaps, which make the interpretation of magnetic reversals uncertain, and therewith, the position of the dinocyst events with respect to the magnetostratigraphy.Moreover, comparison of dinocyst events across latitudes has proven complicated because dinocysts tend to be very sensitive to environmental conditions, which may not necessarily change in tandem over the globe through time.Therefore, dinocysts generally tend to have asynchronous originations and extinctions across different latitudinal bands over the globe (e.g.Williams et al., 2004).Nonetheless, with an error bar of about 200 kyr, some dinocyst events (Table 6) seem to be consistent between Site U1356 and a variety of locations over the globe, which adds to their significance as a biostratigraphic marker, and may help to stratigraphically date sediments from other parts of the Southern Ocean as well.

Conclusions
We formally describe two new dinoflagellate cyst species and present a first zonation scheme for Oligocene-Miocene dinocysts of the Southern Ocean.Temporal resolution of the proposed zonation is in some intervals sufficiently high to allow accurate (i.e.ca.500 kyr resolution) stratigraphic calibration.However, in some intervals, poor core recovery, two hiatuses covering the early Miocene (∼ 23.0-17.1 Ma) and the middle Miocene  and/or low species diversity cause zones with an undesirably long duration.We acknowledge that the zonation might not be applicable to all dinocyst-bearing Southern Ocean sediments due to the strong (latitudinal) provincialism of the dinocyst biogeography in the Southern Ocean today.

Figure 1 .
Figure 1.Paleogeography of the southwestern Pacific Ocean and position of IODP Site U1356 (red star) at (a) 0 Ma, (b) 10 Ma, (c) 20 Ma and (d) 30 Ma. Figures adapted from G plates, with plate circuit from Seton et al. (2012) and absolute plate positions of Torsvik et al. (2012).

Figure 2 .
Figure 2. Integrated Ocean Drilling Program Hole U1356A.Core recovery, lithostratigraphic units, age-depth plot and position of samples taken for palynology.Palaeomagnetic data were obtained from Tauxe et al. (2012), in which black is normal polarity, white is reversed, lilac is uncertain polarity and grey is no data.Age model has been (re-)calibrated to GTS2012 ofGradstein et al. (2012); see text and Table1.

3. 2
Taxonomy and nomenclature Dinoflagellate cyst taxonomy follows that cited in Williams et al. (2017; available through AASP -The Palynological Society at http://www.palynology.organd as a website at http://dinoflaj.smu.ca/dinoflaj3/), with the exception of the taxonomy of the subfamily Wetzelielloideae, for which we follow Bijl et al. (2016; see Table

Figure 3 .
Figure3.Principal component analysis plot showing species scores of the first two axes.Species that are a priori assumed to be reworked in are in red; species that are a priori assumed to be in situ are in blue (see Table2for species list, which species are assumed reworked, eigenvalues, axes scores and cumulative % of variance explained).Dot sizes indicate the total number of encounters in our analyses and gives an indication of the importance of that species in our data.

Table 1 .
Tauxe et al. (2012) the Oligocene-Miocene of Hole U1356A.Ages printed in bold are adjusted relative to the information given inTauxe et al. (2012).

Table 2 .
Williams et al. (2017)es and their PCA scores.(a)Speciesassumed in situ; (b) species assumed reworked.Taxonomy according to that cited inWilliams et al. (2017).Eigenvalues and cumulative % explained by the axes are indicated.
(b) Assumed reworked species name n PCA axis 1 PCA axis 2 PCA axis 3 PCA axis 4 Positive on first two axes

Table 2 .
Continued.Assumed reworked species name n PCA axis 1 PCA axis 2 PCA axis 3 PCA axis 4 Positive on first two axes (b) Oligocene-Miocene boundary interval.Although we interpolate sedimentation rates linearly between our stratigraphic tie points, the depositional setting and the incomplete recovery for Hole U1356A dictates caution in inferring continuous sedimentation throughout the record.

Table 3 .
Informal diagnoses of new dinocyst species found and reference to corresponding images in the plates.
www.j-micropalaeontol.net//37/105/2018/P. K. Bijl et al.: Oligocene-Miocene Southern Ocean dinocyst stratigraphy Diagnosis.Psilate species of Lejeunecysta with two distinct, isolated, slender, distally tapering antapical horns.The cyst has an exaggerated pentagonal outline due to the strong widening towards the cingulum of both the epicyst and the hypocyst.The cingulum itself is only expressed at both lateral sides, while barely visible on the dorsal and ventral sides.

Table 4 .
List of Lejeunecysta species encountered in Hole U1356A.Spiniferites with convex septa connecting the gonal processes, occasionally bearing sutural processes Spiniferites sp.B Species of Spiniferites that resembles the gonal protrusions of Impagidinium aculeatum Spiniferites sp.C Species within the Spiniferites complex, thick gonal processes like Achomosphaera alcicornu, but with sutural septa.Some specimens bear traberculae connecting some processes, resembling those of Cannosphaeropsis (i) their overall outline, which is rounded in L. pulchra and more angular in L. adeliensis sp.nov., and (ii) the presence of parallel folds marking the cingulum on the lateral ends, which is typical for L. adeliensis sp.nov., but does not occur in L. pulchra.
Summary of the dinocyst zones proposed in this paper for the Oligocene-Miocene Southern Ocean.See Table5for a summary of the depth and age data.M.esc: Malvinia escutiana; S.bri: 33.7-33.2Ma.Characteristic species.At the type locality, this dinocyst zone is dominated by the abundance of what we infer are reworked late Eocene dinocysts; species that are considered in situ include Malvinia escutiana, Lejeunecysta fallax, L. acuminata, L. rotunda, abundant Selenopemphix nephroides and S. antarctica, Reticulatosphaera actinocoronata, Impagidinium cf.aculeatum, I. velorum, and I. paradoxum.Significant events.Several Impagidinium species have (regional) first occurrences in this zone, such as I. patulum and I. cf.aculeatum.Phthanoperidinium amoenum has an FO in this zone.

Table 6 .
Selected dinocyst species found in the present study, which have well-calibrated events elsewhere that are broadly consistent with their range at U1356.