Upper Oligocene to Pleistocene planktonic foraminifera stratigraphy at North Atlantic DSDP Site 407, Reykjanes Ridge: diversity trends and biozonation using modern Neogene taxonomic concepts

Deep Sea Drilling Project (DSDP) Site 407, located near the Reykjanes Ridge (southwest of Iceland) offers a rare and extensive record of Late Cenozoic planktonic foraminifera evolution spanning the Neogene and Quaternary periods. This ca. 300 m sequence provides a nearly continuous record of planktonic foraminifera with mostly good preservation quality, aiding the study of pelagic diversity changes over the past 25 million years as the modern North Atlantic Ocean system evolved. Initially investigated in 1979 by Poore, this study presents a taxonomic reassessment of upper Oligocene to Pleistocene planktonic foraminifera at Site 407, including species range documentation, assemblage analysis, biostratigraphic zonation, and age modelling based on planktonic foraminifera, calcareous nannofossils, and scanning electron microscopy. This study employs modern taxonomic perspectives that integrate morphological and stratophenetic frameworks for fossil species with genetic data for taxa having living representatives. Systematic species counts enable quantitative diversity analysis, with a particular focus on the genus Neogloboquadrina, which becomes increasingly prevalent at Site 407 from the late Neogene to Quaternary. The planktonic foraminifera assemblages at Site 407 exhibit a contraction in diversity and a shift in species dominance, notably around 160 m b.s.f. (metres below seafloor) (ca. 8.9–16.5 Ma) and 56 m b.s.f. (ca. 2–3.4 Ma). The upper Oligocene and lower Miocene include species belonging to the genera Catapsydrax, Globoturborotalita, Dentoglobigerina, and Paragloborotalia. An acme of “Ciperoella” pseudociperoensis (lower and middle Miocene), still of uncertain generic affiliation, may have biostratigraphic use. Well-preserved Turborotalita quinqueloba are relatively common throughout the sequence. In Oligocene and Miocene material, T. quinqueloba is accompanied by Tenuitella spp. From the upper Miocene onwards, neogloboquadrinids including Neogloboquadrina praeatlantica, N. atlantica, N. incompta, and N. pachyderma become increasingly common and dominate Pliocene assemblages, together with Globigerina bulloides. Assemblages with an increasingly high-latitude nature, i.e. where N. pachyderma dominates, take over in the lower Pleistocene. Multiple hiatuses are recorded, of which the largest is ca. 8 million years long, separating the middle and upper Miocene (8.9–16.5 Ma; 158.56–160.06 m b.s.f.). Continuous biozonation at Site 407 is challenged by limited species diversity and the absence of standard low-latitude biozone markers, rendering standard schemes ineffective. Recognizable biozones include the low-latitude O7 and M1 Zones in the late Oligocene and early Miocene, respectively; the high-latitude Neogloboquadrina atlantica sinistral Zone in the late Miocene and Pliocene; the Globoconella inflata Zone in the late Pliocene; and the Neogloboquadrina pachyderma Zone in the Pleistocene. The nannofossil biozonation faces similar challenges. A revised biostratigraphic age model integrates calibrated planktonic foraminifera and nannofossil events, incorporating abundant species like “C.” pseudociperoensis, N. atlantica dextral and sinistral, Globoconella puncticulata, G. inflata, and N. pachyderma. These findings are expected to contribute to the Neogene–Quaternary Middle Atlas of planktonic foraminifera and potentially improve the use of neogloboquadrinids in palaeoceanography and biostratigraphy.

1 Introduction

The Neogene period (23.03 to 2.58 million years ago; Raffi et al., 2020) witnessed significant geological and climatic changes, transitioning from a warm early and middle Miocene climate and a cooler late Miocene period to the onset of Northern Hemisphere glaciation in the Pliocene and Pleistocene (e.g. Zachos et al., 2001; Raffi et al., 2020). The late Neogene climate change, combined with the closing of oceanic connections and the subsidence of tectonic ridges, resulted in steepening latitudinal temperature gradients and changes in ocean circulation and stratification globally (Valentine and Jablonski, 2015). These changes created new thermal and nutrient niche boundaries in the pelagic realm, causing the restructuring of global marine plankton communities (Boscolo-Galazzo et al., 2022; Valentine and Jablonski, 2015; Fenton et al., 2023; Woodhouse et al., 2023). This resulted in the latitudinal provincialisation of planktonic foraminifera and increasing separation into biogeographic provinces comprising specific assemblages referable to polar, subpolar, transitional, subtropical, and tropical categories (Kucera, 2007). Understanding the extent to which species will migrate, go extinct, and change their geographic range in the future in relation to warming poles is unclear. Therefore, understanding the development of marine biodiversity patterns over geological time and the factors that influence them are key to contextualizing current trends (Fenton et al., 2023; Woodhouse et al., 2023).

At the global scale, planktonic foraminifera in general experienced significant diversification during the late Neogene, with species diversity increasing through the early Miocene and reaching a peak in the Quaternary (e.g. Fraass et al., 2015). At the regional scale, however, patterns are varied. High latitudes experienced strong biodiversity contraction, and low latitudes remained stable, while temperate and subpolar latitudes, which experience overlapping communities, tended to retain high species richness (Woodhouse et al., 2023). Increasingly, large-scale syntheses of planktonic foraminifera data over long and shorter geological timescales are being used to explore these questions, utilizing plankton databases such as MARGO (Kucera et al., 2005; Strack et al., 2022), ForCenS (Siccha and Kucera, 2017), FORCIS (Chaabane et al., 2023), and, for longer-term Cenozoic timescales, Triton (Fenton et al., 2021, 2023; Woodhouse et al., 2023). While the Triton database puts effort into harmonizing taxonomy and biochronology across latitudes for the Late Cenozoic, a shortcoming is that the Neogene planktonic foraminifera taxonomy in general is rather out of date and strongly in need of revision, and key aspects are not well aligned with modern frameworks for the Paleogene (Olsson et al., 1999; Pearson et al., 2006; Wade et al., 2018) or modern species (underpinned by genetic studies) (Brummer and Kucera, 2022). Moreover, while renewed work on Neogene planktonic foraminifera from the low latitudes is underway, the high latitudes have received less attention (e.g. Wade et al., 2011; King et al., 2020; Lam and Leckie, 2020). Thus, a return to focused studies of targeted regions and specific localities where planktonic foraminifera distributions can be evaluated and their taxonomy analysed is needed.

Using stratigraphic species analysis and scanning electron microscopy (SEM) imaging, this study contributes to efforts towards refining Neogene planktonic foraminifera taxonomy through a detailed re-evaluation of Deep Sea Drilling Project (DSDP) Site 407. Site 407 represents one of the richest existing planktonic foraminifera sequences from the high-latitude North Atlantic and a more westerly positioned complement to the recently recovered cores from the International Ocean Discovery Program (IODP) Expedition 395 (Fig. 1) (Parnell-Turner et al., 2024). Originally studied by Poore (1979), Site 407 has long been known as one of the few existing carbonate-containing sites from the high-latitude North Atlantic that extends back as far as the Oligocene. While Poore (1979) produced an exemplary analysis of the sequence, with excellent-quality SEM images, aspects of the taxonomy are now out of date at both the species and higher taxonomic levels. Moreover, the biostratigraphic framework needs updating to help place microfossil evolutionary events in a consistent chronological model. Taxonomic re-evaluation is therefore needed to improve constraints on species stratigraphic ranges and align the taxonomy with treatments of Recent (Brummer and Kucera, 2022) and Oligocene species (Wade et al., 2018), which are now all underpinned by concepts of wall texture conservatism and, in the case of living taxa, genetics. Here we present the results of our quantitative species analysis, which includes a systematic taxonomy section using up-to-date principles. We produce planktonic foraminifera and nannofossil biostratigraphic zonation, utilizing previously published low-latitude and temperate North Atlantic zonation schemes and a proposed new zonal scheme with modifications. This work will contribute to identifying evidence of endemism at higher latitudes and testing and improving biostratigraphic schemes, including questions of diachroneity across latitudes. It represents part of wider ongoing efforts by the Neogene and Quaternary Planktonic Foraminifera Working Group to improve and harmonize Late Cenozoic planktonic foraminifera taxonomy.

1.1 Geologic and oceanographic setting

DSDP Site 407 (63°56.32^′ N, 30°34.56^′ W; Fig. 1), with a modern water depth of 2472 m b.s.l. (metres below sea level; Shipboard Scientific Party, 1979), sits on late Eocene-age crust (magnetic anomaly chron C13; ca. 35 Ma; LaBrecque et al., 1977) on the western flank of the Reykjanes Ridge (Shipboard Scientific Party, 1979) (Fig. 1). This region is geologically complex, with an equally complex oceanographic history. At the time of the crust formation of Site 407, Icelandic hotspot volcanism was minimal, and the crust had transitional normal oceanic-/hotspot-type characteristics. The upper crust formed at shallow oceanic depths (ca. 1000 m) and subsided to its present depth as crustal spreading and cooling progressed (Luyendyk and Cann, 1979). Site 407 is situated southwest of Iceland today. Iceland itself did not emerge as an island until the early–middle Miocene (Talwani and Eldholm, 1977), when plume volcanism intensified, and nowadays the whole region is underlain by anomalously thick oceanic crust that reflects plume volcanism (Bott, 1983). The locus of seafloor spreading in the northeastern Atlantic has consistently been over the Reykjanes Ridge throughout the Late Cenozoic, with higher spreading rates compared to other ridges in the region (Le Breton et al., 2012).

Initially, Site 407 and the wider North Atlantic region were separated from the Norwegian Sea by the Greenland–Scotland Ridge (GSR), which extends from eastern Greenland to Iceland and the Faroe Islands across to Scotland. The GSR was formed during the Paleocene (anomaly 24; 56 Ma) by mantle-plume-generated hotspot volcanism (Luyendyk and Cann, 1979; Nilsen, 1983). Over time, subsidence and breaching of the GSR occurred during various phases of the Cenozoic (Miller and Tucholke, 1983), influencing the evolution of North Atlantic oceanography, including the development of the Atlantic Meridional Overturning Circulation (AMOC) (Berggren and Schnitker, 1983; Hutchinson et al., 2019; Stärz et al., 2017; Uenzelmann-Neben and Gruetzner, 2018; Wright and Miller, 1996).

Oceanographically, today Site 407 is located in the northern Irminger Sea (Fig. 1) at the confluence of the anticlockwise Atlantic subpolar gyre (Irminger Current; IC), which recirculates warmer Atlantic waters at the surface, with south-flowing polar surface currents exiting the Arctic (East Greenland Current – EGC), and the subarctic front situated to the southwest of Iceland (Fig. 1) (Hátún et al., 2016). Bottom-water flow at Site 407 is driven by deep-water overflows from the Iceland–Scotland and Denmark Strait sectors of the GSR, namely the Iceland–Scotland and Denmark Strait overflow waters (ISOW and DSOW, respectively). The DSOW represents the largest overflow and is joined by the ISOW on the western flank of Reykjanes Ridge to form the Deep Western Boundary Current (DWBC), which is exported southwards (Swift et al., 1980; Hansen and Osterhus, 2000). Modern planktonic foraminifera assemblages in the water column can be described as sub-polar, comprising Neogloboquadrina pachyderma, N. incompta, Turborotalita quinqueloba, Globoconella inflata, Globigerinita glutinata, G. uvula, and Globigerina bulloides, with occasional lower-latitude taxa (Sahoo et al., 2022).

Figure 1 Schematic view of main circulation pathways in the Irminger Sea and Nordic Seas regions. DSDP Site 407 is indicated with the yellow circle. Sites mentioned in this study that are relevant to Site 407 are DSDP Site 408 (green triangle) and IODP Expedition 395 (purple star). Strength and temperature of the currents are represented by arrow width and colour (red is warmer; blue is colder). Ocean currents redrawn from Daniault et al. (2016).

1.2 The status of Neogene planktonic foraminifera taxonomy

In contrast to Paleogene and living species, Neogene planktonic foraminifera classification has not been systematically evaluated or refined since the 1980s and remains one of the current challenges in planktonic foraminifera stratigraphic and evolutionary research. A Neogene and Quaternary working group has been formed to work on Miocene to Recent taxa. The Neogene evolution and phylogeny of planktonic foraminifera have been addressed in several notable works, including Cifelli (1969), Kennett and Srinivasan (1983), Bolli and Saunders (1985), Jenkins (1985), Wei and Kennett (1986), Tappan and Loeblich (1988), Bolli et al. (1989), Aze et al. (2011), Ezard et al. (2011), Fraass et al. (2015), and Lowery et al. (2020). One of the major differences between these works is the usage of subgenera, in particular within the genus Globorotalia. This genus is diverse, with a large species diversity (36 recognized species) compared to other Neogene and Quaternary genera. Globorotalia, in its wider sense, consists of several distinct lineages characterized by the presence/absence of a keel, chamber, and test shape (e.g. conicotruncate vs. lenticular) and test size (Bandy, 1972). Many workers have treated these lineages as subgenera (e.g. Bandy 1972; Kennett and Srinivasan, 1983; Aze et al., 2011), whereas others have refrained from using subgenera (e.g. Triton database; Fenton et al., 2021; Woodhouse et al., 2023). Most recently, however, Brummer and Kucera (2022) note that the subdivision of Globorotalia is not necessary because it is a monophyletic genus and that the use of subgenera has little tradition among extant planktonic foraminifera. However, the usage of subgenera could be useful when tracing the phylogeny of the extant species (e.g. Aze et al., 2011). In this study, we acknowledge the various lineages within the genus Globorotalia, based on the phylogenetic trees in Aze et al. (2011), and use the subgenera Globoconella for classifying the Neogene precursors to G. inflata (Brummer and Kucera, 2022), following the taxonomy of Globoconella as listed in Lam and Leckie (2020b).

A limitation of existing Neogene works (including Kennett and Srinivasan, 1983) is that the taxonomic assessments and phylogenetic models are largely based on studies from low latitudes and the southwest Pacific, omitting higher latitudes and including the northern North Atlantic to Arctic sector. In the Atlantic Ocean, the low-latitude tropical region has remained the major centre for the dispersal of planktonic foraminifera throughout the Cenozoic (Bolli, 1957; Banner and Blow, 1965; Bolli and Bermudez, 1965; Blow, 1969; Berggren, 1969, 1972; Lamb and Beard, 1972; Parker, 1973). Late Cenozoic planktonic foraminifera have received attention in the reports of individual DSDP, Ocean Drilling Program (ODP), and IODP expeditions (ODP Leg 105, Aksu and Kaminski, 1989; DSDP Leg 12, Berggren, 1972; DSDP Leg 48, Murray, 1979; DSDP Leg 49, Poore, 1979; DSDP Leg 81, Huddlestun, 1984; DSDP Leg 94, Weaver, 1987; ODP Leg 104, Spiegler and Jansen, 1989; ODP Leg 151, Spiegler, 1996; ODP Leg 152, Spezzaferri, 1998; ODP Leg 162, Flower, 1999; IODP Site 1313, Sierro et al., 2008), but these works have not been integrated. From some of these studies, temperate and high-latitude North Atlantic (Berggren, 1978; Poore and Berggren, 1975; Spiegler and Jansen, 1989; Weaver and Clement, 1986) and northeastern Atlantic (North Sea) (Anthonissen, 2009, 2012) biostratigraphic zonation schemes have been developed. Our reanalysis of Site 407 provides a good opportunity to assess the applicability of these zonations in the Irminger Sea.

2 Material and methods

2.1 Location and lithostratigraphy

A total of 183.31 m (40 % recovery) of Oligocene to Quaternary sediments was recovered at Site 407 by rotary coring (RCB) (Shipboard Scientific Party, 1979). For this study, upper Oligocene to Quaternary sediments were examined (0.57–329.5 m b.s.f.; metres below seafloor), comprising the entire sediment package above basement. The recovered sediments, apparently pelagites, are characterized by the following lithologies (Shipboard Scientific Party, 1979): (1) Unit 1 (0–46.3 m), with calcareous sandy mud with intervals of calcareous and marly calcareous ooze with variable volcanic ash content (up to 20 %); (2) Unit 2 (46.3–160.7 m), with nannofossil ooze (46.3–124.0 m) and nannofossil chalk (124.0–160.7 m); (3) Unit 3 (160.7–272.0 m), with siliceous nannofossil chalk with an interbedded chalk–volcanic ash zone (215–224.5 m); and (4) Unit 4 (272.0–300.5 m), with nannofossil chalk (272.0–280.0 m) and nannofossil chalk basalt pebble gravel and (5) interlayered sediments between basalt, nannofossil chalk (320.3–321.3 m), and foraminiferal nannofossil chalk (329.4–331.2 m). Biostratigraphic studies conducted as part of the initial work on Site 407 identified at least two unconformities separating the middle Miocene from upper Miocene and the upper Pliocene from lower Pleistocene sediments (Shipboard Scientific Party, 1979; Poore, 1979). The initial Site 407 age framework was based on nannofossil and planktonic foraminifera biostratigraphy only, as the rotary cored sediments were unsuitable for palaeomagnetism measurements due to drilling disturbance (Shipboard Scientific Party, 1979). Core sample notation within the text, figures, and Appendix information sections follow the standard International Ocean Discovery Program (IODP) format, with sample names including site, core number, section number, and sample interval in centimetres within a section.

2.2 Planktonic foraminifera sample preparation, assemblage counting, and imaging

A total of 76 samples (20 cc volume) at ca. 1.5 m intervals were obtained from Cores 1R through 36R from sections 1, 2, and 3 in each core from the IODP Core Repository in Bremen. The initial Site 407 age model, indicating sedimentation rates between < 0.5 and > 4 cm kyr⁻¹, implies a sample resolution of a few 100 kyr to over 1 Myr (Shipboard Scientific Party, 1979). All samples were freeze-dried and weighed before disaggregation. Samples were soaked in demineralized water and agitated for 1–6 h in Erlenmeyer flasks. The flasks were immersed in an ultrasonic bath for up to 30 s as a final disaggregation step. Chalky samples from the deepest levels were first manually broken into fragments. The disaggregated material was wet-sieved over a 63 µm mesh sieve and dried overnight at 49 °C before reweighing to obtain the sand fraction dry mass.

For planktonic foraminifera relative abundance variations, the > 125 µm size fraction of each sample was split into aliquots of approximately 300 specimens (mean = 288; Table A1) using a microsplitter and sprinkled onto a black picking tray. Planktonic foraminifera were examined using a ZEISS Stemi 508 binocular light microscope and identified and counted to species level. The > 125 µm size fraction was found to be sufficient to capture the abundance of most species present in the assemblages, including small species such as Globigerinita glutinata, G. uvula, Tenuitella spp., and Turborotalita quinqueloba. However, very small species such as Orcadia riedeli and Tenuitellita fleisheri were only found in the > 63 µm size fraction. After each count, the full > 63 and > 125 µm size fraction residues were sprinkled evenly on a picking tray and inspected for these small and additionally rare species, some of which might have biostratigraphic value. In cases where additional species were identified, these are noted as “R” in the count tables (Appendix Tables A1 and A2). Certain intervals (47.48–158.56 m b.s.f.) were strongly affected by fragmentation. Here the high abundance of fragments meant that it was not always possible to do systematic 300 specimen counts, as the whole specimens were heavily diluted by the fragments. Sample 407-14R-1, 56–58 cm (120.56 m b.s.f.), had such a high degree of fragmentation that it would be too time-consuming to filter through all the fragments to find whole tests; thus, it was omitted. Since the Neogene taxonomy is undergoing a revision, a composite of reference classification frameworks was used for species identification. This included Kennett and Srinivasan (1983), the Atlas of Oligocene Planktonic Foraminifera (Wade et al., 2018), the review by Brummer and Kuèera (2022) of modern species, and the online database Mikrotax (Young et al., 2017), which includes current working species concepts and generic assignments (Young et al., 2017). Additionally, for some taxa, we refer back to the Poore (1979) species diagnosis. Preservation of whole foraminifers was classified as good, moderate, or bad and quantified by a fragmentation index (FI). Specimens missing > 50 % of their test were considered a fragment. Fragment percentages were determined in the same manner as relative abundance, i.e. counting all fragments in a sample (> 125 µm).

For biozonation, we apply the low-latitude planktonic foraminifera biozonation scheme of Wade et al. (2011), as far as possible in the Oligocene and Miocene, based on identification of bioevents including first (FOs) and last occurrences (LOs) of significant taxa (Table A1). For the Pliocene and Pleistocene, the North Atlantic zonation of Weaver and Clement (1986) and Nordic Seas zonation of Spiegler and Jansen (1989) were applied. Additionally, age-calibrated accessory events from Raffi et al. (2020) were used throughout the whole sequence. Coiling ratios were recorded for Neogloboquadrina atlantica, since the change from dominantly dextral to sinistral is an important age marker in the North Atlantic (e.g. Raffi et al., 2020; Weaver and Clement, 1986; Spiegler and Jansen, 1989). Our age–depth model was constructed by plotting FO and LO events using the “Age–depth plot” tool available through Mikrotax (https://www.mikrotax.org/system/age-depth-plot.php, last access: 20 June 2024). This tool also calculated sedimentation rates.

SEM imaging was conducted to document taxonomic concepts, including wall texture characteristics and the presence/absence of spine holes, focusing on well-preserved examples of each species (Plates 1–21). The instrument used was a JSM-7000F floor SEM, housed in the Department of Materials and Environmental Chemistry at Stockholm University (accelerating voltage = 10 kV; working distance = 10 mm). The specimens were mounted on sticky carbon discs and gold-coated before analysis (2×60 s; applied current of 20 mA).

2.3 Planktonic foraminifera diversity measures

Species diversity, both simple diversity and diversity metrics that consider the relative abundance of species, were calculated throughout the upper Oligocene to Quaternary section using the following approaches:

• Species richness (number of species per sample).

• Shannon–Wiener Index (H'), i.e. $H{}^{\prime }=-\sum \left(p_i\right)\cdot \ln \left(p_i\right)$, where p_i is the proportion of the entire community made up of species i. It assumes that individuals are randomly sampled from an independent large population and all the species are represented in the sample (Shannon and Weaver, 1949), where the higher the value, the more diverse the assemblage will be (Dirzo and Mendoza, 2008). It is a useful measure of biodiversity that considers the species richness and evenness of the species present in the assemblage.

• Evenness, where the concept of evenness refers to the extent to which each species is represented among the samples. Values vary between zero and one, representing one species being dominant and all other species being present in very low numbers (close to zero) or all species being represented by equal numbers (close to one) (Dirzo and Mendoza, 2018).

• Diversity and evenness were calculated using the PAST software, version 4.13 (Hammer et al., 2001).

Additionally, semiquantitative estimates of the abundance of lithic grains (mainly quartz) were determined based on of relative abundance in the > 125 µm fraction, using the following abundance categories: rare (1 %–3 %), few (4 %–15 %), and common (16 %–30 %). This was used as a proxy for the onset of Northern Hemisphere glaciation during the Pliocene and Pleistocene, assuming that the lithic grains represent ice-transported material (ice-rafted debris, IRD) that was carried to Site 407 by icebergs calved from marine-terminating ice sheets on Greenland, Iceland, and Svalbard.

2.4 Calcareous nannofossil samples and analysis

For calcareous nannofossil analysis, small amounts of unprocessed sediment from the same 76 samples used for the foraminifera studies were prepared using the Bown and Young (1998) smear slide technique. The smear slides were examined under a ZEISS Axio Scope.A1 polarizing light microscope at ×1000 magnification and imaged (Plates 22 and 23). Species identification followed the taxonomy of Perch Nielsen (1985) and Young (1998), as well as the most up-to-date descriptions and illustrations in the online database Nannotax3 (Young et al., 2022). The focus for this study was identifying nannofossil biostratigraphic markers, and the assemblages are only briefly summarized, in contrast to the planktonic foraminifera, which are the focus of this study and for which the assemblages and taxonomy are considered in detail. Nevertheless, qualitative calcareous nannofossil species relative abundances were recorded and presented as being dominant (D), present (P), reworked (RW), or rare (R) (Fig. S3). The calcareous nannofossil biozonation is based on identification of the first (FOs) and last occurrences (LOs) of significant taxa in the Site 407 samples (Table B1). The standard nannoplankton biozonation scheme of Martini (1971) was applied based on the selected bioevents.

3 Results

3.1 Planktonic foraminiferal assemblage and diversity changes

A total of 84 species was identified in the 76 new samples from Site 407, of which only 1 sample was barren of planktonic foraminifera (Sample 407-24R-2; 56–58 cm; 217.06 m b.s.f.). A summary of the species' downcore stratigraphic distributions is presented in Table 1 (full species ranges shown in Tables A1 and A2), where the most important species distributions are shown by depth against the epochs interpreted from the planktonic foraminifera and calcareous nannofossil biostratigraphy (see Sect. 3.3), whereas panel (B) in Table 1 shows the Neogene and Quaternary climate (benthic δ¹⁸O) and ice volume plotted against age and was used to make rough climatic interpretations based on the planktonic foraminifera assemblages. For 70 species from the > 125 µm size fraction, occurrences are reported as the relative abundances of the counted assemblage. On average, 268 specimens were counted per sample. A further 14 species, which were not found in the > 125 µm 300 specimen counts, but were observed when the whole sample fraction (> 63 µm size fraction) was examined, are recorded as semiquantitative estimates (Tables A1 and A2). For each species recognized, our taxonomic concept is described in the systematic taxonomy section, together with detailed taxonomic notes.

Fragmentation intensity (%) (Table 1), which could be useful as an indicator of preservation quality, especially regarding calcite dissolution, is variable downcore. In the lower part of the succession (161.56–329.5 m b.s.f.), the fragmentation intensity fluctuates between 11.7 % (169.56 m b.s.f.) and 67.7 % (196.56 m b.s.f.). At 161.56 m b.s.f., fragmentation increases from 12.4 %–49.5 % (160.06 m b.s.f.) and remains high (62.2 %) up to 47.48 m b.s.f. and up to 96.7 % at 120.56 m b.s.f. In the upper part of the sequence, fragmentation decreases from 77.6 % (47.48 m b.s.f.) to 20.4 % (44.56 m b.s.f.) and remains low (average 3.7 %) up to 0.57 m b.s.f.

The majority of samples yielded abundant planktonic foraminifera, and only one sample (Sample 407-24R-2; 56–58 cm; 217.06 m b.s.f.) was barren. Also barren of calcareous nannofossils, the sample was instead dominated by biosilicous material, suggesting strong calcite dissolution. Starting from the base of the section, in intervals close to basement where sediments are intercalated with basalt (Cores 35R and 36R), planktonic foraminifera are common and moderately well preserved, although often infilled with glauconite. Preservation is poor in samples from Core 29R to 32R, but planktonic foraminifera remain common. Specimens in samples from Cores 28R to 19R are few but well preserved.

Table 1 Stratigraphic occurrence of selected planktonic foraminifera abundances at Site 407, shown as a colour gradient (darker blue is for higher abundance). Additional data include lithology, preservation, lithic fragments, and biosilica. Stratigraphic ages are based on planktonic and calcareous nannofossil events from Table 2. Hiatuses are marked by black (major) and grey (potential) wavy lines. R is for rare, G is for good, and M is for medium. There are lithic fragments and biosilica, where D is for dominant, A is for abundant, F is for few, P is for present, and H is for high. (B) Neogene climatic indicators modified from Steinthorsdottir et al. (2021), with δ¹⁸O from De Vleeschouwer et al. (2017) tracking ocean–climate patterns and Northern Hemisphere ice conditions. MCO is the Miocene Climatic Optimum.

These samples are dominated by biosilicous grains (sponge spicules and radiolaria) and non-diagnostic, possibly juvenile, planktonic foraminifera based on their small size. In the remaining intervals (Cores 1R to 18R), foraminifera are well preserved and common, although the fragmentation intensity is high, especially in samples from Cores 6R to 18R (Table 1). Generally, the Site 407 planktonic foraminifera appear glassy (translucent and reflective) or more whitish (frosty) under the light microscope. SEM images show few signs of diagenetic alteration and dissolution, except around Sample 407-13R-1, 56–58 cm (111.06 m b.s.f.), assigned to the Pliocene, where specimens show signs of diagenetic alteration in the form of recrystallization (e.g. Plate 10; Figs. 11, 12).

We record the relative abundance of 76 planktonic foraminifera species from the > 125 µm size fraction (Table A2). Of these, 11 species record relative abundances > 25 % but with large variation through the sequence (Fig. 2). These common species include “Ciperoella” pseudociperoensis (still of an uncertain generic affiliation), Dentoglobigerina venezuelana, Globigerina bulloides, Globigerinella obesa, Globorotalia praescitula, Neogloboquadrina atlantica sinistral and dextral, N. incompta, N. pachyderma, Turborotalita quinqueloba, and Globigerinita glutinata (Fig. 2). This variability is reported below in more detail.

The lower part of the studied section (171.06–392.5 m b.s.f.) is dominated by Globigerinella obesa (average = 42.5 %) and the small microperforate species Globigerinita glutinata (average = 32.8 %). Accessory species are Catapsydrax unicavus, Globoturborotalita woodi, Globorotaloides suteri, and Tenuitella munda. The genus Paragloborotalia diversifies around 275.56–286.46 m b.s.f. (Table 1), contributing on average 7.1 % to assemblages, with T. angustiumbilicata, T. clemenciae, Turborotalita quinqueloba, Globoturborotalita euapertura, G. connecta, and G. bulloides making up much of the remainder. Rare species are Globigerinoides neoparawoodi, Trilobatus quadrilobatus, Catapsydrax dissimilis, C. cf. indianus, Globigerinoides stainforthi, G. hexagonus, and Globigerinella praesiphonifera. Diversity is highest during this interval, as shown by the species richness counts, which vary between 8 (215.56 m b.s.f.) and 20 (291.25 and 198.06 m b.s.f.), and the Shannon diversity (H^′ index), with values fluctuating between 1.25 and 2.49. Evenness fluctuates between 0.21 and 0.92 (Fig. 3). In this interval, 15 bioevents occur, involving the FO of “Ciperoella” pseudociperoensis (407-24R-3; 56–58 cm; 218.56 m b.s.f.), Globoquadrina dehiscens (407-29R-2; 20–22 cm; 264.2 m b.s.f.), Trilobus quadrilobatus (407-29R-2; 20–22 cm; 264.2 m b.s.f.), Globorotalia zealandica (407-19R-3; 56–58 cm; 171.06 m b.s.f.), Globorotalia praescitula (407-20R-1; 100–102 cm; 178 m b.s.f.), Globoturborotalita connecta (407-31R-4; 46–48 cm; 286.46 m b.s.f.), Paragloborotalia continuosa (407-35R-1; 86–88 cm; 320.36 m b.s.f.), P. acrostoma and P. kugleri (407-30R-2; 56–58 cm; 274.06 m b.s.f.), P. mayeri (407-32R-1; 25–27 cm; 291.25 m b.s.f.), P. birnageae and P. semivera (407-30R-3; 56–58 cm; 275.56 m b.s.f.), Dentoglobigerina tripartita (407-31R-4; 46–48 cm; 268.46 m b.s.f.), D. venezuelana (407-32R-1; 25–27 cm; 291.25 m b.s.f.), and D. altispira (407-35R-1; 86–88 cm; 320.36 m b.s.f.). LOs include Tenuitella munda (407-24R-3; 56–58 cm; 218.56 m b.s.f.), Catapsydrax dissimilis (407-23R-1; 56–58 cm; 206.06 m b.s.f.), Globoturborotalita connecta, Catapsydrax unicavus (407-22R-3; 50–52 cm; 199.39 m b.s.f.), Ciperoella ciperoensis (407-29R-2; 20–22 cm; 264.2 m b.s.f.), Globorotaloides suteri and P. birnageae (407-19R-3; 56–58 cm; 171.06 m b.s.f.), P. kugleri (407-28R-3; 56–58 cm; 256.56 m b.s.f.), P. pseudokugleri (407-29R-3; 20–22 cm; 264.5 m b.s.f.), and D. altispira (407-29R-2; 20–22 cm; 264.2 m b.s.f.) (Table A1).

The short interval between 160.06 and 171.06 m b.s.f. is characterized by the highest species richness (about 17; Fig. 2) and is dominated by “Ciperoella” pseudociperoensis, Globorotalia praescitula, and Paragloborotalia sp., accounting for roughly 69 % of the assemblage (Table 1). Accessory species are Dentoglobigerina venezuelana, Globoturborotalita woodi, and Globigerinita glutinata, accounting for 13 % of the assemblage. Minor amounts of G. uvula, Trilobatus trilobus, T. immaturus, P. acrostoma, P. mayeri, P. continuosa, Neogloboquadrina cf. praeatlantica, D. altispira, D. baroemoenensis, Globigerina falconensis, G. umbilicata, Globorotaloides archeomenardii, G. praemenardii, G. zealandica, G. challengeri, and Tenuitella angustiumbilicata account for the remaining part.

The interval between 47.48 and 158.56 m b.s.f. is dominated by the morphologically plastic species Neogloboquadrina atlantica (sinistral) (ca. 35 %; Fig. 2) (Plate 15; Figs. 1–7). A distinct coiling change from dominantly sinistral (average 35 %) to dextral (average 42 %) N. atlantica occurs at 150.56 m b.s.f. (Table 1 and Fig. 2). Other important taxa in this interval are Globigerina bulloides, N. praeatlantica (in the lower part; 139.56–158.56 m b.s.f.), N. incompta, N. pachyderma, T. quinqueloba, and G. glutinata, of which the latter is well represented in the > 125 µm size fraction. Five species of Neogloboquadrina (N. acostaensis, N. atlantica, N. incompta, N. pachyderma, and N. praeatlantica) appear abruptly in Sample 407-18R-1, 56–58 cm (158.56 m b.s.f.), except for N. humerosa, which appears in 407-12R-1, 56–58 cm (101.56 m b.s.f.). These species account for roughly 45 % of the assemblage. Neogloboquadrina praeatlantica is only present in the lower part of this interval, making up about 7 % of the assemblage (139.56–158.56 m b.s.f.). Accessory species are G. woodi, G. apertura, G. decoraperta, Globoconella puncticulata, G. obesa, G. uvula, and P. continuosa and account for the remaining part of the assemblage. Sphaeroidinellopsis paenedehiscens has a short range between 150.56 and 158.56 m b.s.f. This interval has two FO events, namely G. inflata (65.06 m b.s.f.) and G. puncticulata (95.06 m b.s.f.), and six LO events, namely G. woodi (54.56 m b.s.f.), G. puncticulata (9R-2; 53–55 cm; 73.58 m b.s.f.), G. apertura (7R-1; 56–58 cm; 54.56 m b.s.f.), N. atlantica (sinistral and dextral) (6R-3; 55–58 cm; 47.48 m b.s.f.), and T. clemenciae (18R-1; 56–58 cm; 158.56 m b.s.f.). Between Samples 407-18R-1, 56–58 cm (158.56 m b.s.f.), and 407-18R-2, 56–58 cm (160.06 m b.s.f.), 14 species abruptly disappear (“C.” pseudociperoensis, D. venezuelana, D. baroemoenensis, D. globosa, G. dehiscens, G. praescitula, G. archeomenardii, G. praemenardii, G. zealandica, G. challengeri, P. mayeri, P. acrostoma, T. immaturus, and T. trilobus), and there is the sudden appearance of diverse and common species of the genus Neogloboquadrina, including N. pachyderma. Species richness is about 12, H^′, and the evenness values fluctuate between 1.36–2.22 and 0.36–0.80, respectively.

Figure 2 Summary of planktonic foraminifera metrics, including the relative abundance of 11 key species (i.e. with > 25 % relative abundance), and measures of foraminiferal preservation and assemblage diversity.

There is a progressive decrease in the overall number of species, yet evenness remains relatively stable throughout the zone. Shannon H^′ increases steep between Samples 6R-2, 56–58 cm (46.06 m b.s.f.), and 6R-3, 55–58 cm (47.48 m b.s.f.) (Fig. 2; species diversity).

The upper part of the sequence (0.57–46.06 m b.s.f.) is dominated by Neogloboquadrina pachyderma (about 75 % of the total assemblage; Fig. 2). Within this interval, we recognize five morphotypes of N. pachyderma (Plate 16), which appear to be equivalent to the N. pachyderma morphotypes originally recognized in Quaternary sediments of the central Arctic Ocean and distinguished using a numbering system, Nps-1 through Nps-5 (El Bani Altuna et al., 2018; Eynaud et al., 2009; Eynaud, 2011). The most common of the N. pachyderma morphotypes are Nps-1, Nps-2, and Nps-3 (Plate 16; Figs. 3–8). Accessory species over this interval are Globigerina bulloides, N. incompta, Turborotalita quinqueloba, and Globigerinita glutinata. These accessory species together account for about 23 % of the assemblage. Minor amounts of Globorotalia scitula, G. crassaformis, Globigerinita uvula, Globoconella inflata, N. acostaensis, and N. dutertrei are recognized. Rare Tenuitella fleisheri and Orcadia riedeli occur in the > 63 µm size fraction. Other rare species include Globorotalia truncatulinoides, G. hirsuta, G. cf. cariacoensis, G. umbilicata, N. humerosa, Orbulina universa, O. suturalis, and Globigerinella cf. siphonifera. Two noteworthy LO events are present, namely N. acostaensis and N. humerosa at 19.05 and 17.55 m b.s.f., respectively. Species richness is low in the upper part of the sequence, with about 7 on average (Fig. 2).

3.2 Calcareous nannofossil assemblage

Calcareous nannofossils are moderately well preserved to well preserved, and the average species richness is highest in the middle Miocene (160.06–171.06 m b.s.f.) and lowest in the Quaternary (0.57–25.55 m b.s.f.) (Table B1). Coccolithus pelagicus dominates the interval between Samples 407-6R-1, 56–58 cm, and 407-36R-1, 50-52 cm (44.56–329.5 m b.s.f.); Reticulofenestra spp. dominates in the interval between 407-18R-1, 56–58 cm, and 407-36R-1, 50–52 cm and (158.56–329.5 m b.s.f.); Gephyrocapsa spp. dominates in the interval between 407-4R-1, 55–58 cm, and 407-10R-1, 56–58 cm (25.55–82.56 m b.s.f.); and Gephyrocapsa huxleyi dominates in 407-1R-1, 57–60 cm, to 407-2R-3, 56–58 cm (0.57–9.56 m b.s.f.). Accessory taxa, such as Discoaster spp. and Helicosphaera spp., are quite abundant during the Oligocene and Miocene but become few to rare after 158.56 m b.s.f. Sample 407-24R-2, 56–58 cm (217.06 m b.s.f.), is barren of calcareous microfossils, including calcareous nannofossils. Biosiliceous material, including radiolaria fragments and diatoms, is common in most smear slides. The most intense Paleogene reworking was recorded in Samples 407-18R-3, 56–58 cm, to 407-30R-3, 56–58 cm (161.56–275.56 m b.s.f.), as identified by the occurrence of Zygrhablithus bijugatus (Paleocene–Oligocene), Chiasmolithus altus (Eocene–Oligocene), and large Reticulofenestra spp. (Eocene–Oligocene?). In samples from 407-1R-1, 57–60 cm, to 407-5R-1, 56–58 cm (0.57–35.06 m b.s.f.), reworking of upper Cretaceous taxa, such as Eiffellithus eximius, Prediscosphaera cretacea, Watznaueria spp., Microrhabdulus undosus, Micula staurophora, Arkhangelskiella spp., and Tranolithus orionatus is common. Reworked Paleogene species are also present in this interval but less noticeable (mostly Reticulofenestra spp.).

3.3 Age–depth model and integrated planktonic foraminifera and calcareous nannofossil biostratigraphy

The lack of a palaeomagnetic record at Site 407 results in an age–depth model solely consisting of calibrated planktonic foraminifera and calcareous nannofossil bioevents for age control for the past approximately 25 million years. The stratigraphic range data of identified planktonic foraminifera species underpinning the biozonation are presented in Tables 4, A1, and A2, and the stratigraphic range data of calcareous nannofossils are presented in Table B1. Planktonic foraminifera and nannofossil evolutionary events that are important for the biozonation (Table 2) and establishing the age–depth control points were identified and compiled (Table 3; Appendix Tables C1 and C2). The planktonic foraminifera biostratigraphic zonation (Table 2) mainly follows the events described in Wade et al. (2011) and Raffi et al. (2020). As planktonic foraminifera diversity decreases in the upper part of the studied section, increasingly fewer low-latitude bioevents are recognized. Therefore, we also apply the local high-latitude North Atlantic biostratigraphic frameworks of Weaver and Clement (1987) and Spiegler and Jansen (1989). The studied planktonic foraminifera assemblages include 8 age diagnostic marker taxa and 19 recognized (calibrated) taxa from the late Oligocene to the Quaternary (Table C1). Marker taxa are highlighted in bold; all ages are included in GTS2020, except the ages given by Weaver and Clement (1986). The calcareous nannofossil biostratigraphic zonation follows the events described by Martini (1971) and Raffi et al. (2006, 2020). The studied calcareous nannofossil assemblages include 3 age diagnostic marker taxa and 14 recognized (calibrated) taxa from the early Miocene to the Quaternary (Table C2).

The distribution of calibrated bioevents reveals that sedimentation at Site 407 has been interrupted by numerous hiatuses (Fig. 3; Table 3), with the largest hiatus lasting from about 8.9–16.5 Ma between Samples 18R-2, 56–58 cm (160.06 m b.s.f.), and 18R-1, 56–58 cm (158.56 m b.s.f.), separating the late early Miocene from the late Miocene. Other likely hiatuses occur in the Pleistocene (0.36–1.49 Ma), Pleistocene–Pliocene boundary (1.98–3.39 Ma), late Miocene (5.56–8.16 Ma), and early Miocene (21.43–22.71 Ma). Sediment accumulation rates are around 20 m Myr⁻¹. in the early Miocene (H–I interval) and late Oligocene (J–K interval) (Fig. 3). A lower sedimentation rate of ca. 11 m Myr⁻¹ is recorded in the late Miocene (F–G interval). Rates increase to ca. 44 m Myr⁻¹ in the late Miocene-Pliocene (D–E interval), reaching the highest sedimentation rate at Site 407 in the Pleistocene (ca. 78 m Myr⁻¹.; B–C interval).

Figure 3 Age–depth plot based on our revised Site 407 planktonic foraminifera and nannofossil biostratigraphy, with core numbers and lithologic units (left column). Depth is given in metres below seafloor (m b.s.f.). Plt is for Pleistocene, and Q is for Quaternary. Age–depth control points used to calculate sedimentation rates are listed in Table 3. All ages are on GTS2020. Planktonic foraminifera and calcareous nannofossil events are indicated by the numbers 1–39 (Appendix Tables C1 and C2).

Table 2 Planktonic foraminifera (PF) and calcareous nannofossil (NN) bioevents and zonation schemes at Site 407. Zonation schemes used for PF (grey) are Wade et al. (2011), Weaver and Clement (1986), Spiegler and Jansen (1989), and Poore (1979). Zonation schemes used for NN (grey) are Martini (1979) and Steinmertz (1979). The dark green zonation is a proposed PF zonation based on the zones from Spiegler and Jansen (1989), Wade et al. (2011), and Weaver and Clement (1986). New zones are the G. inflata partial range zone (PRZ) between the FO of G. inflata and the FO of “modern” N. pachyderma, the G. puncticulata Zone based on the FO of G. puncticulata and G. inflata, the N. praeatlantica Zone spanning the FO and LO of N. praeatlantica, and the “C.” pseudociperoensis Zone based on the FO and LO of this species. The stratigraphic age (epochs and periods) is based on the PF and NN events from Fig. 3.

Table 3 DSDP Site 407 age–depth control points, sedimentation rates, and hiatus lengths. Calcareous nannofossil events are denoted by “N” and planktonic foraminifera events by “F”.

4 Discussion

4.1 Planktonic foraminifera assemblages and diversity at Site 407 during the Neogene to Quaternary

The assemblage analysis presented here, which is a development after the study by Poore (1979), allows a more quantitative analysis of Neogene plankton diversity patterns in the Site 407 for the northern North Atlantic record. Planktonic diversity and species richness were calculated to constrain diversity increases and decreases in response to late Oligocene to Quaternary climate events. Planktonic foraminifera are present in all samples (except Sample 407-24R-2; 56–58 cm) and show highly variable temporal changes in assemblage composition and relative abundance throughout the past 25 million years (Tables 1 and A1). A distinct faunal change occurred after the middle Miocene, involving increasing species losses and a general diversity decrease, suggestive of strong environmental forcing. These patterns are further detailed at the taxon level below.

Oligocene and early Miocene assemblages are diverse (average $H^{\prime }=\mathrm{1.9}$) and relatively even (average evenness of 0.49), with higher abundances of G. glutinata and G. obesa (Fig. 2; Table B1). In the middle Miocene (160.06–171.06 m b.s.f.), notable changes in species composition and fluctuations in diversity occur, including the temporary appearance of low-latitude species (T. trilobus, T. immaturus, T. bisphericus, G. ruber, G. praemenardii, and G. archeomenardii), albeit in low numbers (Tables 2 and A1) between 16–17 Ma (Fig. 3). This is most likely related to the Miocene Climatic Optimum (MCO), as indicated in the δ¹⁸O record shown in Table 1. The shift in evenness reflects the increasing dominance of a few species through this interval, including “Ciperoella” pseudociperoensis, still of uncertain generic affiliation, which jumps to contributing up to 51 % of the assemblages between 160.06 and 171.06 m b.s.f. (Fig. 2). Globorotalia praescitula also appears during this interval. At 160.06 m b.s.f., evenness is 0.2 and caused by a high abundance of G. praescitula making up 60 % of the assemblage.

From the late Miocene onwards, the pattern of Shannon H^′ diversity (Fig. 2), which records decreased species richness and evenness from 158.56 m b.s.f., suggests that planktonic foraminifera were possibly experiencing stress related to global cooling (Table 1). Neogloboquadrinids increasingly dominate assemblages from this point, with up to 63 % of Pliocene assemblages represented by N. atlantica (sinistral and dextral) and up to 96 % of assemblages represented by N. pachyderma in the Pleistocene. The first IRD, in the form of rare quartz grains, is tentatively placed at 152.03 m b.s.f. and assigned to the late Miocene (Table 1), suggesting an initial but temporary export of clast-carrying icebergs from early Greenlandic glaciers during the middle to late Miocene (St. John and Krissek, 2002). IRD becomes abundant throughout the Pleistocene around 1.77 Ma (from 46.06 m b.s.f. upwards), likely coincident with the persistent ice sheet cover on Greenland from the early Pleistocene.

Currently, and throughout the Quaternary period, both northern and southern Atlantic high-latitude regions exhibit comparable planktonic foraminifera communities, including Neogloboquadrina pachyderma, Globigerina bulloides, and Turborotalita quinqueloba (e.g. Bé, 1960a; Kennett, 1968; Vilks, 1970; Tolderlund and Bé, 1971; Hemleben et al., 1989; Eynaud, 2011). However, during the late Neogene, these high-latitude regions displayed different assemblages, particularly concerning the neogloboquadrinids. The Neogene North Atlantic region predominantly featured various neogloboquadrinids, with assemblages being dominated by the relatively large and morphologically variable morphospecies N. atlantica, whereas only N. acostaensis and N. pachyderma are documented in the Atlantic sector ((sub)Antarctic) of the Southern Ocean (55–90° S; e.g. DSDP Leg 28, Kaneps, 1975; DSDP Leg 35, Rögl, 1976; DSDP Leg 71, Krasheninnikov and Basov, 1983; ODP Leg 114, Pujol and Bourrouilh, 1991; ODP Site 1123, Crundwell and Nelson, 2007). To our knowledge, N. atlantica has not been recorded in the southern Indian Ocean or South Pacific (Berggren, 1992; Martin Crundwell, personal communication, 2023), although it has been recorded in the Pacific on the Californian margin (Dowsett and Poore, 2000). Further work is required to establish the distribution of Neogloboquadrina praeatlantica and Neogloboquadrina atlantica and the extent of possible endemism in the North Atlantic and Mediterranean region (e.g. Hilgen et al., 2000; Lirer et al., 2019).

4.2 Integrated planktonic foraminifera and calcareous nannofossil biostratigraphy and chronology

The planktonic foraminifera zonation schemes applied here are based on age diagnostic events described in Wade et al. (2011), Raffi et al. (2020), Weaver and Clement (1987), and Spiegler and Jansen (1989) (Tables 4 and C1). Application of a zone was only permitted when primary zonal markers defining the base and top of a zone were recorded. The nannofossil events are based on Martini (1971) (Tables 2 and C2). This section will describe the integrated biostratigraphy per epoch.

4.2.1 Late Oligocene and early Miocene

The base of the Site 407 sequence (329.5 m b.s.f.) recorded one planktonic foraminifera age marker, namely the presence of Paragloborotalia pseudokugleri. Other late Oligocene markers such as P. opima and Chiloguembelina cubensis (both Zone O5) are absent, placing the base in Zone O7 (Wade et al., 2011). The Oligocene/Miocene boundary is placed at the FO of Paragloborotalia kugleri (Sample 407-30R-2; 56–58 cm; 274.06 m b.s.f.), marking the base of Zone M1. The earliest Miocene Zone M1 extends to 28R-3, 56–58 cm (256.56 m b.s.f.), based on the LO of P. kugleri. The FO of Globoquadrina dehiscens at 264.2 m b.s.f. marks the boundary between zones M1a and M1b. The LO of Ciperoella ciperoensis is a M1a subzone marker (Wade et al., 2011), but at Site 407, the LO of C. ciperoensis (264.2 m b.s.f.) is recorded in the same depth as the FO of G. dehiscens (264.2 m b.s.f.). Zone M2 is commonly based on the LO of P. kugleri and the FO of Globigerinatella sp. (Wade et al., 2011); however, Globigerinatella sp. is absent at Site 407. Therefore, the boundary between zones M2 and M3 cannot be determined. Accessory datums in this interval are the LO of Tenuitella munda (407-24R-3; 56–58 cm; 218.6 m b.s.f.), the FO of Globorotalia zealandica at 171.06 m b.s.f. (19R-3; 56–58 cm), and the LO of Catapsydrax dissimilis at 206.1 m b.s.f. (407-23R-1; 56–58 cm). This interval is possibly interrupted by a minor hiatus (I–J; Fig. 3) of just over 1 Myr (21.43–22.71 Ma), as indicated by the LO of C. ciperoensis and FO of G. dehiscens recorded at the same depth. Common age diagnostic species such as Fohsella spp. and Praeorbulina spp., as well as the base of Orbulina suturalis, are missing at Site 407, making it difficult to assign any of the early Miocene zones of Wade et al. (2011).

The Oligocene/Miocene boundary cannot be precisely identified in the nannofossil record due to the reworking of Paleogene specimens into Miocene sediments. Specifically, Zygrhablithus bijugatus (from the Paleocene–Oligocene) and Chiasmolithus altus (from the Eocene–Oligocene) have been found in Miocene layers up to 161.56 m b.s.f. This reworking obscures the true LO of Helicosphaera recta, which marks the base of NN1 (early Miocene). Additionally, Sphenolithus ciperoensis, another early Miocene marker, is absent in all samples. The FO of the Helicosphaera ampliaperta group at 275.56 m b.s.f. (Sample 407-30R-3; 56–58 cm) indicates the earliest Miocene (Raffi et al., 2020). However, it is important to note that in the northwestern Atlantic the FO of H. ampliaperta is not considered a reliable datum (Fabbrini et al., 2019), corresponding with it being an outlier in our Site 407 age–depth model (Fig. 3). The late Oligocene and early Miocene sediments lack clear nannofossil age markers but contain a few accessory markers. The FO of Discoaster druggii (referred to as Discoaster cf. druggii in this study) at 263.06 m b.s.f. (Sample 407-29R-1; 56–58 cm) marks the base of zone NN2. Triquetrorhabdulus carinatus, which indicates the top of NN2, is extremely scarce and occurs sporadically throughout the sequence. The boundaries between NN2 and NN3 and NN3 and NN4 cannot be determined due to the absence of key age diagnostic events, such as the FO of Sphenolithus belemnos. However, Sphenolithus heteromorphus, an additional marker for NN3, is present but rare and scattered.

4.2.2 Middle Miocene (160.06–171.06 m b.s.f.)

No primary age diagnostic planktonic foraminifera are recorded in the middle Miocene. However, multiple accessory events are present. Globorotalia archeomenardii/praemenardii, G. praescitula, Paragloborotalia acrostoma, and P. mayeri are recorded at 160.06 m b.s.f. Globorotalia archeomenardii/praemenardii has its FO in the middle Miocene Zone M5b/M6, and G. praescitula has its LO in Zone M7 (Raffi et al., 2020). The range of Paragloborotalia acrostoma extends into Zone M5 and P. mayeri into Zone M10 (Leckie et al., 2018). Thus, this interval is tentatively placed in the middle Miocene.

Likewise, the nannofossil data did not record any age diagnostic species in the middle Miocene. The presence of Sphenolithus heteromorphus and Helicosphaera magnifica at 171.06 m b.s.f., which are known globally to range from the middle part of zone NN2 to the lower part of NN4 (Boesiger et al., 2017), provides some stratigraphic control. The last recorded occurrences of Sphenolithus puniceus, Cyclocargolitus floridanus, Helicosphaera ampliaperta, and S. heteromorphus all occur at 160.06 m b.s.f., the same horizon at which a clustering of planktonic foraminifera disappearances occurs, which implies a large hiatus between (Table 3), which our age model suggests spans approximately 8 million years (8.88–16.54 Ma). The large hiatus indicates either sediment removal or a period of non-deposition at Site 407.

4.2.3 Pliocene and late Miocene (47.48–158.56 m b.s.f.)

The Pliocene and late Miocene sediments contain few age diagnostic species, with planktonic foraminifera assemblages being dominated by neogloboquadrinids. The temperate/subpolar and Nordic Seas events by Weaver and Clement (1986) and Spiegler and Jansen (1989) are somewhat applicable to this interval (Fig. 5). The Weaver and Clement (1987) Pliocene zonation is divided into five zones, of which only one is applicable on the material from Site 407, namely the Globorotalia inflata partial range zone (PRZ). This is the zone between the FOs of Globoconella inflata (1.90–2.18 Ma) and sinistrally coiled encrusted Neogloboquadrina pachyderma (1.66–1.77 Ma) (in our study referred to as “modern” N. pachyderma). At Site 407, G. inflata has its FO at 65.06 m b.s.f. (407-8R-2; 56–58 cm) and N. pachyderma at 44.56 m b.s.f. (407-6R-1; 56–58 cm). Therefore, the interval between 44.56 and 65.06 m b.s.f. is placed in the G. inflata PRZ (Fig. 2), corresponding to the late Pliocene Zone PL6 by Raffi et al. (2020).

The coiling change from dominantly dextral to sinistral of N. atlantica at 152.03 m b.s.f. (407-17R-3; 53–55 cm), which is an age-calibrated (6.8 Ma) accessory marker retained in the latest planktonic foraminifera biozonation (Wade et al., 2011; originally calibrated in the North Atlantic, Berggren et al., 1995) occurs in Zone M13b (Wade et al., 2011; Raffi et al., 2020). In the Spiegler and Jansen (1989) Nordic Seas (ODP Leg 104) zonation scheme, this change in coiling forms the boundary between the N. atlantica (dextral, hereafter dex.) and N. atlantica (sinistral, hereafter sin.) Zones and is assumed to be between 6 and 6.2 Ma. The top of their N. atlantica (sin.) Zone falls within PL6 (Raffi et al., 2020). In the North Atlantic N. atlantica's LO is at 2.05–2.36 Ma (Weaver and Clement, 1987). This event was not used by Wade et al. (2011) on the grounds that it is likely a local event and in need of further evaluation. At Site 407, N. atlantica persists up to 47.48 m b.s.f., just above the Plio-Pleistocene boundary and above the C–D hiatus recognized in our data, which would imply a similar yet slightly younger age for this LO of ca. 1.8 Ma. This could be within error in the previous findings, given the uncertainty in both our sedimentation model and the position of this C–D hiatus around this horizon (1.98 to 3.39 Ma; Fig. 3 and Table 3). The N. atlantica (sin.) Zone thus spans 47.48–152 m b.s.f., and the Miocene/Pliocene boundary falls within the lower part of this zone (Fig. 2). Spiegler and Jansen (1989) recorded a second N. atlantica coiling change involving a shift to dominantly dextral forms around 2.3 Ma. This N. atlantica (dex.) Zone, which overlays the N. atlantica (sin.) Zone at ODP Hole 644A, however, is relatively short and spans roughly 10 m of sediment. The top of this zone is placed at 1.84 Ma. A change from dominantly sinistral to dextral-coiling N. atlantica in the early Pleistocene is not observed in our Site 407 samples, since both sinistral and dextral N. atlantica go extinct when the coiling is still dominantly sinistral (Table 1). The latest Pliocene N. atlantica (dex.) Zone has thus far only been recorded at Hole 664A; however, the age falls within the recorded LO of N. atlantica in the work by Weaver and Clement (1986) (2.05–2.36 Ma), and when combined with the short duration of the Zone, it could be overlooked or not sampled at other North Atlantic sites. The presence of a hiatus at this horizon is further supported by a sharp lithological change, marking the contact between lithological Units 1 (calcareous sandy mud) and 2 (nannofossil ooze) between 46.06 and 47.48 m b.s.f. (Sample 407-6R-3; 55–58 cm). The hiatus may have removed the interval of abundant dextral-coiling N. atlantica, with the result producing an apparent simultaneous disappearance of both N. atlantica (sin. and dex.) in our Site 407 records. The presence of a latest Pliocene N. atlantica (dex.) Zone needs further investigation as it has the possibility of being a useful marker elsewhere in the high-latitude North Atlantic.

Another notable older neogloboquadrinid event at Site 407 is the abrupt and simultaneous appearance at 158.56 m b.s.f. of various Neogloboquadrina species, including N. acostaensis, N. atlantica (dex.), N. atlantica (sin.), N. incompta, and N. praeatlantica, together with the well-known modern polar specialist N. pachyderma, indicating a late Miocene age. Neogloboquadrina acostaensis (Wade et al., 2011) and N. pachyderma have their evolutionary first occurrences in Zone M13a (Raffi et al., 2020) and N. humerosa in Zone M13b (Raffi et al., 2020). Neogloboquadrina humerosa is rare and first recorded at 101.56 m b.s.f. (407-12R-1; 56–58 cm). Because of the clustering of calibrated nannofossil and planktonic foraminifera bioevents between 160.06–158.56 m b.s.f. close to these neogloboquadrinid FOs in our age–depth model (Table 3, Fig. 5), we infer that this pattern does not represent true first appearances but the presence of a major hiatus. It should be noted that the FO of the N. pachyderma sinistral age marker utilized in our age model is 9.37 Ma, which was calibrated in the southern Indian Ocean (Kerguelen Plateau) (Berggren et al., 1995). This date might not be accurate in the high-latitude North Atlantic due to diachroneity between latitudes (see Lam and Leckie, 2020a). It is not possible to investigate the possibility of cross-latitude diachroneity of the N. pachyderma FO datum at Site 407, as the true FO of this species is lost due to the hiatus. Age calibration of the FO of N. pachyderma in stratigraphically complete North Atlantic Neogene sequences is thus still needed. IODP Expedition (Exp.) 395 records (Parnell-Turner et al., 2024) promise new opportunities to do this. Weaver and Clement (1986) and Spiegler and Jansen (1989) made other late Miocene observations. In the temperate North Atlantic (ODP Leg 94), Weaver and Clement (1986) recorded a change from dominantly sinistral to dextral N. pachyderma at 5 Ma, with their “N. pachyderma” remaining dominantly dextral until the latest Pliocene. With current knowledge, we can assume that the Weaver and Clement (1986) N. pachyderma dextral is N. incompta, based on modern DNA perspectives of both species (Darling et al., 2006). Nevertheless, at Site 407, N. incompta is also more abundant than N. pachyderma in the interval between 139.56-158.56 m b.s.f., after which N. pachyderma reaches higher abundances. Spiegler and Jansen (1989) recorded a N. acostaensis Zone between the FOs of N. acostaensis and N. atlantica. At Site 407, both species appear simultaneously.

In the Pliocene and late Miocene of Site 407, multiple-accessory-calibrated planktonic foraminifera bioevents exist. This includes the latest Pliocene LO of Globoturborotalita woodi (Wade et al., 2011; Raffi et al., 2020), which in Site 407 occurs at 54.56 m b.s.f. (7R-1; 56–58 cm), and the FO of Globoconella puncticulata at 95.06 m b.s.f. (11R-3; 56–58 cm). Globoconella puncticulata is a common marker in sediments recovered during DSDP Leg 94 around 4.06–4.30 Ma (Weaver and Clement, 1987), corresponding with PL1-2 (Raffi et al., 2020), and is used as the base of the Weaver and Clement (1987) Globorotalia puncticulata PRZ. The top of this PRZ is marked by the LO of Globorotalia cf. crassula (sensu Weaver), which was not recorded at Site 407. The LO of Paragloborotalia continuosa occurs at 142.4 m b.s.f. (Sample 407-16R-3; 56–58 cm). Although not formally calibrated, this event is reported to occur in late Miocene Zone M13, based on global data sets, including those from the North Atlantic (Leckie et al., 2018). N. praeatlantica (Plate 17) and N. cf. acostaensis (Plate 14; Figs. 4–6) disappear at 139.56 m b.s.f. (407-16R-1; 56–58 cm). The range of N. praeatlantica in the North Atlantic is still unknown. In the Mediterranean, the FO is recorded in Zone M10 (Hilgen et al., 2000a; 2005) and extends into the Tortonian (Foresi et al., 2002). The sudden disappearance of N. praeatlantica and N. cf. acostaensis could suggest a second hiatus. However, no sediment was recovered in Core 15 (Shipboard Scientific Party, 1979), and no noteworthy events are recorded in the nannofossil record, fragmentation intensity, or lithology supporting this.

The Pliocene–late Miocene nannofossils are similarly limited in terms of their biostratigraphic utility. Only one biozone, NN11, was recorded based on the presence of primary zonal markers. Zone NN11 is a fairly long total range zone comprising the interval between the FO and LO of Discoaster quinqueramus (150.56 and 149.06 m b.s.f., respectively), with an age calibration of about 8.10–5.53 Ma (Raffi et al., 2020). This corresponds broadly with the shift from dominantly dextral to sinistral coiling in N. atlantica calibrated to 6.8 Ma at 150.56 m b.s.f. Multiple calibrated accessory calcareous nannofossil events are recorded (Table B1), which contribute to the age model. Discoaster hamatus, the marker for the base and top of NN9; Catinaster coalitus, the marker for the base of NN8; and Discoaster kugleri, the marker for the base of NN7 all have a short-term range in the late Miocene that were not found in this study because the base of the late Miocene sequence is most likely younger than the ranges of these species. Discoaster asymmetricus and Ceratolithus rugosus, the marker species for the top of NN13 and NN12, respectively, were not recorded in any samples, most likely because they are low-latitude taxa.

4.2.4 Pleistocene (0.57–46.06 m b.s.f.)

The Pleistocene interval lacks any age diagnostic planktonic foraminifera species that delineate established low-latitude biozones. One event that is considered to be a useful marker in high northern latitudes is the first common occurrence (FCO) of “modern” N. pachyderma (Plate 16; Figs. 3–12), which are larger and more encrusted than early morphotypes (Plate 16; Figs. 1, 2). In Site 407, this occurs at 44.56 m b.s.f. (407-6R-1; 56–58 cm), after which the species dominates assemblages to the top of the sequence. In the high-latitude North Atlantic, the FCO of N. pachyderma is around 1.7 Ma (subchron Olduvai at DSDP Site 611; Weaver and Clement, 1987). The low- to mid-latitude marker for the base of Zone PT1a (also the Pliocene/Pleistocene boundary), Globigerinoides fistulosus, is not present at Site 407; thus, this zone cannot be recognized (Wade et al., 2011). We can, however, approximate the start of the Pleistocene by the first occurrence of abundant lithic clasts at 46.06 m b.s.f. (Sample 407-6R-2; 56–58 cm), implying the arrival of ice rafted debris (IRD) that is indicative of extensive Northern Hemisphere glaciation. This is represented by a downcore lithological transition from calcareous sandy mud to nannofossil ooze at 46.3 m b.s.f. (Shipboard Scientific Party, 1979). The IRD persists to the top of the sequence. A secondary event in the Pleistocene is the LO of N. acostaensis at 19.05 m b.s.f. (407-3R-3, 55–58 cm) that occurs in PT1a based on age-calibrated data from Raffi et al. (2020). However, the boundary between Zone PT1a and PT1b cannot be determined because the marker Globorotalia tosaensis is missing. The Arctic Zone 1 (AZ1) corresponds to the N. pachyderma (sin.). Zones of Weaver and Clement (1986) and Spiegler and Jansen (1989) span the entire Quaternary.

The calcareous nannofossil record divides the AZ1 into the early and late Pleistocene based on two datums, namely the LO of Pseudoemiliania lacunosa at 17.55 m b.s.f. and the FO of Gephyrocapsa huxleyi at 9.56 m b.s.f. (Table B1). A possible erosional unconformity (A–B; Fig. 3) in the late Pleistocene caused a hiatus of about 1 Ma (0.36–1.49 Ma), based on the LOs of P. lacunosa and N. acostaensis. A similar erosional unconformity is found at ODP Site 918, causing a hiatus of about 320 kyr (1.39–1.71 Ma) (St. John and Krissek, 2002). The presence of G. huxleyi in the interval from 9.56 m b.s.f. (407-2R-3; 56–58 cm) to the top of the core (407-1R-1; 57–60 cm) represents Zone NN20/21 (Martini, 1971), where G. huxleyi and Gephyrocapsa spp. can be found in high numbers (Fig. S3). The early Pleistocene is based on Zone NN19/20 (Martini, 1971), based on the interval between the LO of Discoaster brouweri at 56.06 m b.s.f. (407-7R-2; 56–58 cm) and the LO of P. lacunosa at 17.55 m b.s.f. (407-3R-2; 55–58 cm). An accessory event in AZ1 is the LO of Helicosphaera sellii at 44.56 m b.s.f. (407-6R-1; 56–58 cm) calibrated to 1.24 Ma (Raffi et al., 2020).

4.2.5 Sedimentation rates and identified hiatuses

Sedimentation rates at Site 407 varied considerably over time; in the late Oligocene and early Miocene, rates were around 20 m Myr⁻¹, decreasing to 11 m Myr⁻¹ in the late Miocene, then increasing to 44 m Myr⁻¹ in the late Miocene–Pliocene, and peaking at 78 m Myr⁻¹ in the Pleistocene (Fig. 3). Initial rates calculated by Shor and Poore (1979) show 15 m Myr⁻¹ in the upper Oligocene to lower Miocene and 45 m Myr⁻¹ in the upper Miocene/Pliocene at Site 407, which are comparable to the rates from this study. At the nearby Site 408 (see Fig. 1), the rates were 17 and 35 m Myr⁻¹, respectively. On the eastern flank of the Reykjanes Ridge, the average sediment accumulation rates in the Pleistocene were higher compared to the western flank up to 120 m Myr⁻¹ (DSDP Site 114; Laughton and the Exp. 12 Scientific Party, 1972). This difference is attributed to sediment drift formation (e.g. Björn Drift) occurring along rises on the western sides of the basins related to bottom-water circulation and the Coriolis force (see Wold, 1999, for more details), whereas the sediment drift on the western flank of Reykjanes Ridge (Snorri Drift) has an indistinct morphology possibly related to low sediment supply (Wold, 1999). West of Site 407, on the Greenland continental rise (ODP Site 918), sedimentation rates were 22 m Myr⁻¹ in the Miocene, reaching 199 m Myr⁻¹ in the Pliocene and then decreasing to approximately 80 m Myr⁻¹ from the mid-Pliocene to present (Larsen et al., 1994a). Sedimentation on the Greenland continental rise is influenced by regional factors (Rasmussen et al., 2003). Overall, increased sediment supply after the mid-Pliocene resulted from the uplifted GSR and turbidity currents from Iceland, with terrigenous input from Iceland increasing during the Pleistocene (Wold, 1999).

The hiatuses identified at Site 407 likely reflect periods when bottom waters were vigorously eroding sediments, as reflected by the sudden increase in fragmentation in the middle Miocene, which may represent winnowed transported material (49.5 %; 160.06 m b.s.f.) (Fig. 3). The large middle Miocene hiatus is widespread in the region of Site 407 (DSDP Sites 406, 408, and 112; Shor and Poore, 1979) and in the Iceland–Faroe Ridge (IFR) (DSDP Sites 336 and 352; Talwani and Udintsev, 1976), reflecting the initiation of overflow water through the IFR as a proto-ISOW as the GSR became deeper and wider (Uenzelmann and Neben, 2018) (Fig. 1).

In the late Miocene, deposition restarted at Site 407, as well as at other sites along the path of the overflow waters (Sites 403–406, 408, 114), so the renewal of the deposition might be related to the increasing depth of the Reykjanes Ridge crest (Shor and Poore, 1979). Strong bottom waters occurred throughout the late Miocene up to the late Pliocene, as reflected in the high fragmentation throughout this period (about 62 %). Planktonic foraminifera tests are well preserved in this interval and do not show signs of corrosion. The Pliocene–Pleistocene hiatus only occurred on the western flank of the Reykjanes Ridge (Sites 407 and 408) and may be correlated with the uplift of the ridge crest (Shor and Poore, 1979) and an increase in ISOW and DSOW. That only minor test fragmentation recorded in the Pleistocene could be interpreted to suggest that the bottom waters were not as erosive then as during the Miocene and Pliocene. On the other hand, the dominance of N. pachyderma in the Pleistocene, which has a very robust dissolution resilience test, could explain the reduced fragmentation.

Changes in GSR sill depth have influenced the bottom-water exchange between the North Atlantic and the Nordic Seas throughout the Late Cenozoic, with different segments of the GSR deepening at different times (Wright and Millers, 1996; Uenzelmann-Neben and Gruetzner, 2018). An important step in this process was subsidence of the Iceland–Faroe Ridge below sea level sometime in the Oligocene; this allowed the initiation of strong bottom waters via a proto-ISOW during the Neogene (Fig. 1) that entrained large quantities of sediment, eroding and depositing it as hiatuses/sediment drifts in the North Atlantic (Nilsen, 1983; Uenzelmann-Neben and Gruetzner, 2018). This continued periodically during the late Oligocene and Miocene (Uenzelmann-Neben and Gruetzner, 2018).

4.2.6 Suggestions for a revised planktonic foraminifera zonation for the northeastern Atlantic

Since the majority of biostratigraphic frameworks are based on low-latitude taxa, their applicability at Site 407 does not work beyond Zone M1, after which all the zonal markers from Wade et al. (2011) are absent. This section will discuss a revised planktonic foraminifera zonation for the northeastern Atlantic based on the zonation schemes by Wade et al. (2011), Weaver and Clement (1986), and Spiegler and Jansen (1989). A summary of these schemes is shown in Fig. 3.

The Neogene planktonic foraminiferal zonation has stayed relatively stable over the past decades, as several key markers initially introduced by Bolli (1957) and Banner and Blow (1965) remain in current zonal schemes such as the low-latitude scheme by Wade et al. (2011). The standard set of low-latitude zones has changed from Paleogene (“P”) and the Neogene (“N”) zones, applied by Poore (Blow, 1969; Kennett and Srinivasan, 1983), to the “O”, “M”, “Pl”, and “PT” zones for the Oligocene, Miocene, Pliocene, and Pleistocene, respectively (Berggren et al., 1995; Wade et al., 2011). Since the Poore (1979) Site 407 study, biostratigraphic frameworks have developed considerably. Moreover, the stratigraphic range of various species have been extended, and new age-calibrated marker species have been identified, based on new observations (Olsson et al., 1999; Pearson et al., 2006; Wade et al., 2018).

The application of the Weaver and Clement (1987) zonation based on the DSDP Leg 94 results (North Atlantic; ca. 38–52° N), has improved age control for the late Miocene and Pliocene by applying several additional palaeomagnetically calibrated events, e.g. the change in coiling direction (dextral to sinistral) in N. atlantica, the LO of G. puncticulata, the FO of G. inflata, and the FCO of N. pachyderma. Nonetheless, only two of the Weaver and Clement (1986) zones work at Site 407 (Fig. 2); these include the N. pachyderma Zone and the G. inflata PRZ, as the zonal marker Globorotalia cf. crassula that delineates other zonal boundaries is absent at Site 407. The Spiegler and Jansen (1989) alternative zonation scheme, which was based on records from the Nordic Seas (Vøring Plateau; ODP Leg 104), relies more heavily on neogloboquadrinids, and from this, we can apply the N. atlantica (dex.), N. atlantica (sin.), and N. pachyderma Zones, spanning the late Miocene to Quaternary, to Site 407 sediments. A downside of the Spiegler and Jansen zonation is the age resolution; the lower diversity and longer evolutionary ranges of species at higher latitudes, including N. atlantica (sin.), results in zones of long time duration.

A proposed zonation for Site 407 involves a combination of the zones by Spiegler and Jansen (1986), Wade et al. (2011), and Weaver and Clement (1986), with three additional local events, i.e. Globoconella puncticulata Zone, Neogloboquadrina praeatlantica Zone, and the “Ciperoella” pseudociperoensis Zone (Table 2). These zones are defined as the biostratigraphic interval characterized by the FO and LO of the nominate taxa, based on their distinct stratigraphic range. In particular, the G. puncticulata and N. praeatlantica Zones increase the resolution of the N. atlantica sinistral “superzone”. Additionally, a fourth zone could be added between the LO of Tenuitella munda and the FO of “C.” pseudociperoensis, covering the whole sequence. However, the large middle/late Miocene hiatus and the missing sediments from Core 15R impact the extent of the zones; i.e. the true LO of “C.” pseudociperoensis and true FO of N. praeatlantica fall within the large hiatus, and the true LO of N. praeatlantica is most likely in Core 15R. Furthermore, both species need more investigation as their geographic and stratigraphic extent are not known yet (see the “Taxonomic notes” for “C.” pseudociperoensis and Sect. 4.2 for N. praeatlantica).

Other aspects of Poore's biostratigraphy, however, remain the same after our work, including the inability to differentiate planktonic foraminifera Zones N5 and N6 (here updated to Zones M2 and M3, respectively; Raffi et al., 2020) due to the absence of Globigerinatella sp. at Site 407. The explanation for the decreasing biostratigraphic resolution is a general decrease in faunal diversity (Figs. 4 and 5) and increasing endemism particularly related to N. pachyderma dominance forced by Neogene climate change. This likely explains the absence or extreme scarcity of key low-latitude marker species such as the Fohsella lineage (middle Miocene Zones M7–M10), Globigerinatella sp. (base M3), Praeorbulina (Zone M5; Wade et al., 2011), and Orbulina. All the Pliocene and Pleistocene markers are missing, namely FO Globorotalia tumida, LO Globorotalia margaritae, LO Sphaeroidinellopsis seminulina, LO Dentoglobigerina altispira, LO Globorotalia miocenica, LO Globigerinoidesella fistulosa, and LO Globorotalia tosaensis (Wade et al., 2011). Other species that might have been present could be missing in the large hiatuses, e.g. the FO of Globoturborotalita nepenthes (base of M11; Wade et al., 2011). This species has been recorded at other sites in the North Atlantic. At DSDP Site 408 and Holes 410 and 410A, their abundance is sparse to few (Poore, 1979). It is more common on the Rockall Plateau at ODP Hole 982B (Flower, 1999), DSDP Sites 404 and 406 (Murray, 1979), and DSDP Leg 81 (Huddlestun, 1984). It seems more likely that G. nepenthes is absent at higher North Atlantic latitudes because it is a warm-water species (Huddlestun, 1984), confirmed by its absence in the East Greenland Margin (ODP Leg 152; Spezzaferri, 1998), where the middle Miocene record is continuous (compared to Site 407). We largely agree with the Poore (1979) species level assignment based on his SEM images; however, many changes have occurred on the genus level as will be discussed in the next section.

4.3 Taxonomic and biostratigraphic developments since Poore's 1979 study

Several aspects of planktonic foraminifera taxonomy and systematics, and their use in biostratigraphy, have changed in the 50 years since the Poore (1979) study (Schiebel et al., 2018). First, following developments presented in the Paleocene, Eocene, and Oligocene taxonomic atlases (Olsson et al., 1999; Pearson et al., 2006; Wade et al., 2018), taxonomic frameworks are now underpinned by the principle of quasi-conservativism in wall textures (Pearson, 2018), i.e. that spinose and non-spinose groups have independent origins that can be traced forward from the early Paleocene (Morard et al., 2022) and that microperforates may include multiple separate origins from benthic ancestors (Leckie 2009; Pearson 2018). Species that have remained within the same genus for the past 50 years are Neogloboquadrina spp., Turborotalita quinqueloba, Globigerinita glutinata, G. uvula, Globorotalia scitula, G. praescitula, G. archeomenardii/praemenardii, G. hirsuta, G. truncatulinoides, G. crassaformis, and G. zealandica (Kennett and Srinivasan, 1983; Brummer and Kucera, 2022).

In the original Site 407 study, many species were classified in either Globigerina or Globorotalia. The biggest change concerns the genus Globigerina (Table 4). In the original study, many species were classified as Globigerina; e.g. G. apertura, G. ciperoensis, G. connecta, G. decoraperta, G. euapertura, G. galavisi, G. obesa, G. ouachitaensis, G. pseudociperoensis, G. tripartita, and G. woodi were initially classed as Globigerina. Nowadays, these species are placed in Ciperoella (C. ciperoensis), Globoturborotalita (G. connecta, G. decoraperta, G. euapertura, G. ouachitaensis, and G. woodi), Dentoglobigerina (D. galavisi and D. tripartita), and Globigerinella (G. obesa). Globigerina pseudociperoensis, still of uncertain generic affiliation, is tentatively placed in the genus Ciperoella (see the “Taxonomic notes” and Plate 2) and is still an ongoing topic for the upcoming Neogene atlas. Globigerina itself has now been reduced to a restricted set of species that share a pseudocancellate spinose wall texture forming ridges and a high-arched umbilical aperture bordered by an imperforate rim (Plate 5) (Spezzaferri et al., 2018; Brummer and Kucera, 2022; Fabbrini et al., 2023), while Miocene species referred by Poore to Globoquadrina (i.e. G. altispira, G. baroemoenensis, G. larmeui, and G. venezuelana) are now placed in Dentoglobigerina (Wade et al., 2018b) (Plates 3 and 4).

The genus Globorotalia has also undergone major changes. In the original study, Globorotalia acrostoma, G. birnageae, G. continuosa, G. inflata, G. kugleri, G. mayeri, G. miotumida, G. minutissima, G. munda, G. nana, G. pseudokugleri, G. puncticulata, and G. siakensis were classed as Globorotalia by Poore and are now placed in Globoconella (G. inflata, G. miotumida, and G. puncticulata), Paragloborotalia (P. acrostoma, P. birnageae, P. continuosa, P. kugleri, P. mayeri, P. nana, P. pseudokugleri, and P. siakensis), and Tenuitella (G. minutissima = T. clemenciae) (Table 5).

Table 4 Compilation figure showing the taxonomy, samples, and stratigraphic ranges of species identified in this study (dark blue) and by Poore (1979) (grey) throughout the upper Oligocene to Pleistocene. Hiatuses are indicated with the wavy black line. Stratigraphic ranges of Catapsydrax–Globorotaloides.

Table 5 Compilation figure showing the taxonomy, samples, and stratigraphic ranges of species identified in this study (dark blue) and by Poore (1979) (grey) throughout the upper Oligocene to Pleistocene. Hiatuses are indicated with the wavy black line. Stratigraphic ranges of Globoturborotalita–Tenuitellita. Neogloboquadrina du/pac is for dutertrei/pachyderma intergrade.

The genus Globigerinoides has also undergone considerable revision. After fossil and genetic evidence identified polyphyly in the group, species are now assigned to either Globigerinoides (G. ruber) or Trilobatus (T. bisphericus, T. immaturus, T. quadrilobatus, and T. triloba) (Spezzaferri et al., 2015).

Making a direct comparison between the Poore (1979) study and this study is not straightforward as both studies used a slightly different methodology, i.e. in the different size fractions observed (> 149 µm in the original and > 125 µm, in addition to 63–125 µm, in this study) and in determining species abundance (estimates vs. relative abundance counts, respectively), which means Poore's study recorded fewer small species such as Tenuitellita fleisheri. The differences between taxonomic assignment in the two studies are described in more detail below and illustrated in Table 4.

Species recognized in the Poore (1979) Site 407 range chart (Table 4) that were not found in this study are Paragloborotalia opima, Globigerinoides primordius, Globoturborotalita druryi, Globorotalia margaritae, G. ventriosa, G. exserta, Globoconella miozea, Globigerina bulbosa, Beella praedigitata, Globigerina martini, Sphaeroidinellopsis seminulina, and Hastigerina pelagica. We agree with the identification of G. primordius, G. druryi, G. margaritae, G. exserta, G. miozea, S. seminulina, and H. pelagica based on their holotype images and the images from Poore (1979) but disagree with the identification of G. ventriosa and G. bulbosa. The holotypes of both G. ventriosa and G. bulbosa are sketches, making it difficult evaluate the utility of these species. Globorotalia ventriosa seems to have more curved sutures (Ogniben, 1958) than the specimens from Poore (1979; Plate 4; Figs. 10–12), and the Globigerina bulbosa holotype's (LeRoy, 1944) final chamber is very bulbous and ovate, comprising the bulk (> 50 %) of the test, whereas the Poore (1979) G. bulbosa final chamber is more globular, making up roughly 50 % of the test. Furthermore, no images of P. opima and B. praedigitata are presented in Poore (1979), making it impossible to determine whether we agree or disagree with his identification.

Species recognized in this study but not found in material from Poore (1979) are T. fleisheri (Plate 21; Figs. 9–12), C. cf. indianus (Plate 1; Figs. 2, 3), G. archeomenardii/G. praemenardii plexus (Plate 7; Figs. 1–5), G. truncatulinoides (Plate 9; Figs. 4–6), G. challengeri (Plate 7; Figs. 6–10), G. hexagonus (Plate 10; Fig. 1), P. semivera (Plate 13; Fig. 6), P. siakensis (Plate 13; Fig. 7), N. incompta (Plate 15; Figs. 10–12), G. neoparawoodi, G. ruber (Plate 18; Fig. 1), T. quadrilobatus, T. bisphericus (Plate 18; Fig. 8), T. immaturus (Plate 18; Fig. 9), and T. trilobus (Plate 18; Fig. 10). An interesting observation is the absence of T. primordius in this study and the lack of G. neoparawoodi in the Poore (1979) work. We acknowledge that the two species can easily be confused due to their similar features and overlapping stratigraphic ranges, and we concur with the Poore (1979) identification of T. primordius. A more detailed comparison between the two species is provided in the “Taxonomic notes” section. Compared to the Poore (1979) work, this study employs a higher sampling resolution, and in addition to counting the > 125 µm size fraction, the 63–125 µm size fraction was scanned for rare species, where T. fleisheri, G. truncatulinoides, G. hexagonus, Globigerinoides spp., and Trilobatus spp. were found in low abundances, thus explaining the difference between both studies. The practice of making standardized assemblage counts also forced us to identify most specimens in the assemblage, which can also impact the species identification and diversity. One the most significant developments, relates to refined perspectives on species that have living counterparts for which genetic phylogenies exist (e.g. Morard et al., 2015). In the case of Site 407, this involves the species Globigerina bulloides, Globigerinella cf. siphonifera, Globoconella inflata, Globorotalia scitula, G. hirsuta, G. crassaformis, G. truncatulinoides, Neogloboquadrina pachyderma, N. incompta, N. dutertrei, Orbulina universa, and Turborotalita quinqueloba. Of particular importance is the case of genus Neogloboquadrina, which is explored in the following section.

4.4 North Atlantic Neogloboquadrina

Our results show that neogloboquadrinids have been a key component of high-latitude North Atlantic planktonic foraminifera assemblages since the late Miocene. Despite the 5 Myr long mid-Miocene hiatus at Site 407, in which the first appearance of the genus Neogloboquadrina is lost, the late Miocene to Pleistocene provides insights into the evolution of the genus, including the transition from assemblages dominated by N. atlantica in the Pliocene and Miocene to N. pachyderma domination in the Pleistocene. With abundances often exceeding 90 % of assemblages in many high-southern-latitude and high-northern-latitude regions, N. pachyderma is considered a polar extremist, and the only true polar species that can tolerate intensive sea ice conditions (Bé, 1977; Schiebel et al., 2017). To this day, Neogloboquadrina remains an enigmatic genus. Neogloboquadrinids have a strong preferential coiling direction, reaching 100 %, compared to other genera, where the coiling direction is 50 %, e.g. in Turborotalita quinqueloba (Darling et al., 2006). For decades, it was believed that the two distinct coiling forms of N. pachyderma were related to sea surface temperatures, and therefore they were used as a palaeoceanography proxy (e.g. Ericson, 1959; Kennett, 1968); i.e. an increase in N. pachyderma sinistral was indicative of cooler conditions, and vice versa, for N. pachyderma dextral. However, in the landmark study by Darling et al. (2006) it was shown that both coiling forms are in fact genetically distinct and should be considered distinct species. The right-coiling variant was recognized as N. incompta, and the left-coiling variant as N. pachyderma. Below, we identify five aspects of neogloboquadrinid classification that would benefit from future research in more complete sedimentary sequences.

• The origin of the of genus Neogloboquadrina remains uncertain. The current view is that Paragloborotalia continuosa (Blow 1959) (Plate 12; Figs. 8–11) is ancestral, with Neogloboquadrina acostaensis as the assumed first neogloboquadrinid (Kennett and Srinivasan, 1983). This is based on the (1) general similarity of wall textures (highly cancellate/rugose); (2) coiling pattern, i.e. slight terminal reduction in the chamber numbers per whorl; (3) similarity in test size, apertural size, orientation, and lip; and (4) overlap in stratigraphic ranges in the upper Miocene (Jenkins, 1967). An alternative phylogenetic model suggests that N. praeatlantica was the first species, having evolved from Catapsydrax (Foresi et al., 2002), a presumed non-spinose genus. Currently too few appropriate sections have been studied to test these alternatives, and N. praeatlantica has not been routinely recorded. From a genetic viewpoint, the modern genus Neogloboquadrina appears to constitute a monophyletic clade (Darling et al., 2006). Unfortunately, the interval potentially containing the Neogloboquadrina origin is missing from Site 407, falling within the major middle Miocene–late Miocene hiatus. Observations from Site 407 are consistent with the P. continuosa ancestry hypothesis in that all early neogloboquadrinids overlap in stratigraphic ranges with P. continuosa. We note that N. praeatlantica is closely comparable to P. continuosa in that it has 4 chambers, a similar test size, and the presence of a lip. In this study, P. continuosa was distinguished from N. praeatlantica in that it has a comma-shaped aperture, whereas N. praeatlantica has a higher arched aperture (e.g. Plate 17; Fig. 3). We note that N. praeatlantica had not yet been described at the time of Poore's study. Both N. praeatlantica and P. continuosa have a reticulate wall texture, often characterized by heavy gametogenic calcification. A possible argument against Paragloborotalia being the ancestor to all neogloboquadrinids is that Paragloborotalia is assumed to be sparsely to weakly spinose (Leckie et al., 2018), whereas Neogloboquadrina is non-spinose. Paragloborotalia spinosity, however, is controversial, as discussed by Leckie et al. (2018). Several species have been shown to have features interpreted as spine bases, including P. pseudokugleri, P. kugleri, P. semivera, and P. siakensis (Leckie et al., 2018), yet such evidence is rare and sometimes ambiguous. Paragloborotalia continuosa is thought to be sparsely spinose (Leckie et al., 2018), but due to heavy gametogenic calcification, it is difficult to see spine bases on SEM-imaged specimens. Also, no wall texture images of P. continuosa are included in the Oligocene atlas. In this case, the non-spinose Catapsydrax sp.–N. praeatlantica relationship might be worth further exploration.

• The species N. atlantica is currently poorly integrated into taxonomic and phylogenetic frameworks. The most recent taxonomic treatment of neogloboquadrinids (Kennett and Srinivasan, 1983) acknowledges the existence of N. atlantica but does not incorporate it into a phylogenetic model. As evidenced in this study, N. atlantica is one of the most common species in late Miocene and Pliocene mid- to high-latitude North Atlantic assemblages. It has value as an age marker in the late Miocene, based on a coiling shift from dominantly dextral to sinistral and calibrated to 6.97 Ma (Berggren et al., 1995; Wade et al., 2011; Raffi et al., 2020), as well as a local change back to dominantly dextral coiling at 2.3 Ma (Spiegler and Jansen, 1989; Kaminski and Berggren, 2021). Like N. pachyderma, N. atlantica also shows great morphological plasticity (see “Taxonomic notes“””; Plate 14 and Figs. 7–12; Plate 15 and Figs. 1–7). No attempts have yet been made to categorize morphotypes, and/or evaluate whether there is palaeoecological or stratigraphic value in doing so. It has been speculated that N. atlantica may be closely related to the N. acostaensis–N. humerosa–N. dutertrei lineage, based on morphological similarities within these species (Berggren, 1972). However, it is difficult to discern how N. atlantica would be ancestral to this lineage, since N. acostaensis shows more similarities with the purported neogloboquadrinid ancestor P. continuosa in being smaller and more compact than N. atlantica. Due to the middle–late Miocene hiatus, it is not possible to discern phylogenetic trends within this group, as the majority of neogloboquadrinids appear simultaneously just after the hiatus. Neogloboquadrina atlantica appears to be rather common and has been recognized in the northwest Pacific (Lam and Leckie, 2020b), although not apparently at high southern latitudes (see discussion above). Morphotypes originally described as Globoquadrina asanoi and G. kagaensis from the Pliocene of Japan (Maiya et al., 1976) may be conspecific with N. atlantica (Lam and Leckie, 2020b), as well as Neogloboquadrina inglei from the lower Pleistocene of the California margin (Kucera and Kennett, 2000), but this also needs further attention, and no direct comparisons with the Atlantic occurrences have yet been made.

• N. praeatlantica, which is the purported ancestor of N. atlantica and potentially the first neogloboquadrinid, is not included into the Neogene atlas (Kennett and Srinivasa, 1983) since it was only formally described in 2002 from the late middle Miocene Mediterranean of the Italian Tremiti Islands (Foresi et al., 2002). The first possible mention of N. praeatlantica in the literature is “N. atlantica primitive form” at Sites 407 and 408 (Poore, 1979). So far, N. praeatlantica has not been recorded outside of Site 407 and its Mediterranean-type locality (Hilgen et al., 2000a; Iaccarino et al., 2001; Foresi et al., 2002) and Sicily (DiStefano et al., 2002). We note that N. praeatlantica can be difficult to distinguish from P. continuosa, since both have a small-to-medium low-trochospiral test with 4 chambers in the final whorl, possibly explaining why it has not been recognized in other Neogene planktonic foraminifera studies. An alternative hypothesis argues that N. praeatlantica originated near Iceland (based on the work done by Poore, 1979) and migrated southwards to reach the Mediterranean and subtropical Atlantic (DSDP Site 397) but never got further south, as suggested by the absence of the species in the equatorial Atlantic (ODP Site 926) (Lirer and Iaccarino, 2005).

• The morphological plasticity of N. pachyderma (Plate 16; Figs. 1-12), currently acknowledged within a five-morphotype system used to describe modern and Pleistocene forms (Eynaud et al., 2009; Eynaud, 2011; El Bani Altuna et al., 2018), was not used in the past; therefore, it is unclear when this variability evolved and what the controls are. Multiple previous studies from the Pliocene to the modern have recognized plasticity in N. pachyderma test morphology, and there is a long history of discussing and naming of different morphotypes (e.g. Keller, 1978; Kennett, 1968; Kennett et al., 2000; Kennett and Srinivasan, 1980; Stehman, 1972; Vilks, 1975). Based on estimated genetic divergence times of the N. pachyderma genotypes, the genotype that is found in the central Arctic ocean and Nordic Seas (Type 1) diverged between 1.8–1.5 Ma from the Southern Hemisphere genotypes (Darling et al., 2004). This divergence coincided with strengthening of Northern Hemisphere cooling, suggesting that this genetic split marks the transformation of N. pachyderma into a true polar specialist; thus, its evolution and speciation was coupled to climate change (Darling et al., 2007). In the central Arctic Ocean, five morphotypes are recognized from studies of core-top samples in the Canadian Arctic archipelago (El Bani Altuna et al., 2018; Eynaud, 2011; Eynaud et al., 2009). However, how these fit in with the early evolution of N. pachyderma, including how it should be distinguished from ancestors that look rather similar to some of the modern N. pachyderma morphotypes, is less clear and something that needs attention. This study is the first to record the five “modern” morphotypes back to the early Pleistocene. Better age constraints are still needed to assess whether the split at 1.8–1.5 Ma gave rise to the five morphotypes in the North Atlantic and Arctic oceans.

  Contrary to their Quaternary and Pliocene counterparts, Miocene N. pachyderma morphotypes have typically been lumped together in census studies. This shortcoming might be because characterizing the morphotypes is time-consuming, and the majority of Neogene planktonic foraminifera assemblages are published as DSDP/ODP/IODP data reports that mostly preceded the work by Eynaud et al. (2009). Furthermore, in the 1970s, the surface ultrastructural characteristics and porosity in Late Miocene to Recent specimen were considered to be phenetic changes as a consequence of environmental and palaeoceanographic changes, whereas modern DNA perspectives can also sometimes distinguish genetic differences (e.g. Darling et al., 2006; Huber et al., 1997).

• Where exactly N. incompta (Plate 15; Figs. 10–12) fits with the Neogloboquadrina phylogenetic history is uncertain. Few studies have recorded the FO of N. incompta since it was initially considered a dextral-coiling N. pachyderma, until genetics showed otherwise (Darling et al., 2006). According to some studies, N. incompta evolved before N. pachyderma, with a FO of 12 Ma (Kucera and Schönfeld, 2007), whereas the N. pachyderma FO is 9.37 Ma (Raffi et al., 2020).

• In Poore's work, Quaternary right-coiling neogloboquadrinids with 4–5 chambers in the final whorl are arbitrarily categorized as Neogloboquadrina du/pac (originally described as the P–D intergrade category by Kipp, 1976) and are considered to be intergrades between N. incompta and N. dutertrei (Plate 17; Figs. 9–11, 12; Poore, 1979). This study did not attempt to apply this category, as these specimens closely resemble N. incompta, i.e. compacted test and closed umbilicus, unlike N. dutertrei, which typically has a more globose test with an open aperture. Additionally, this term has been used to categorize Pliocene specimens that are transitional between N. incompta and N. acostaensis or N. atlantica (e.g. Dowsett et al., 1988; Dowsett and Poore, 1990). So, there is no common consensus on the usage of this term for four- to five-chambered right-coiling neogloboquadrinids, indicating a need for further evaluation.

In summary, the phylogeny of Neogloboquadrina needs revisiting. Currently, among age-calibrated planktonic foraminifera events, N. acostaensis is the first neogloboquadrinid to appear in the geological record, followed by N. pachyderma, N. humerosa, and N. dutertrei (Raffi et al., 2020). Neogloboquadrina incompta, N. atlantica, and N. praeatlantica still need to be incorporated into the phylogenetic tree of Neogloboquadrina. Dedicated work on well-dated and continuous northern North Atlantic records is needed to resolve these questions.

5 Conclusions

The present study documents high-latitude North Atlantic planktonic foraminifera biostratigraphy, taxonomy, and species composition in the Irminger Sea over roughly the past 25 million years, integrating these findings with calcareous nannofossil biostratigraphy as complementary tie points. The DSDP Site 407 record suggests quasi-continuous sedimentation with two hiatuses, namely a large hiatus spanning approximately 8 million years in the middle to late Miocene and a shorter hiatus separating the Pliocene and Pleistocene. Both hiatuses have been linked to the initiation of bottom-water currents in the Neogene, which eroded sediments at the site location.

DSDP Site 407 shows a steady decline in species richness and diversity from the late Oligocene to Pleistocene combined with a change in assemblages to a more (sub)polar assemblage as reflected in the high abundance of neogloboquadrinids in the late Miocene to Pleistocene. The highest species richness was recorded in the middle Miocene, most likely coinciding with the MCO, when low-latitude species such as Globigerinoides ruber and Trilobatus spp. appear in the assemblages.

The majority of standard biostratigraphic zonation schemes are based on low-latitude taxa, and since climate deterioration affected the species richness at Site 407, the application of low-latitude zonation schemes was only possible in the lower part of the sequence spanning the late Oligocene and early Miocene Zones O7 and M1. North Atlantic age-calibrated events such as the change in dominantly dextral to sinistral coiling in Neogloboquadrina atlantica in the late Miocene, the last occurrence of Globoconella puncticulata and the first occurrence of Globoconella inflata in the Pliocene, and the first common occurrence of Neogloboquadrina pachyderma in the Quaternary have improved the age constraint at Site 407. This study reports three proposed local biostratigraphic zones in order to improve the age constraint in the early Miocene, late Miocene, and Pliocene. These local zones are (1) the early–middle Miocene “Ciperoella” pseudociperoensis Zone, (2) the late Miocene Neogloboquadrina praeatlantica Zone, and (2) the Pliocene Globoconella puncticulata Zone. All zones cover the biostratigraphic interval between the first and last occurrence of their nominate taxa. However, whether these zones are applicable in other high-latitude North Atlantic studies needs to be investigated in the future. Unfortunately, the LO of “C.” pseudociperoensis and the FO and LO of N. praeatlantica fall within a hiatus limiting the possible extent of their true first and last occurrences.

Despite the large hiatus, Site 407 provides a useful record of pre-Pleistocene neogloboquadrinids, i.e. N. praeatlantica, N. atlantica (dex.), N. atlantica (sin.), N. acostaensis, N. humerosa, N. dutertrei, N. incompta, and N. pachyderma. This study has highlighted outstanding questions relating to neogloboquadrinids, such as how N. praeatlantica, N. atlantica, and N. incompta fit into the Neogloboquadrina phylogeny and the large morphological plasticity in N. atlantica and N. pachyderma. The origin and ancestry of N. pachyderma remains unsolved, and improved constraints on the timing of its first occurrence in the high-latitude North Atlantic are needed, as both events fall within the middle to late Miocene hiatus at Site 407.

Although uncertainties regarding high-latitude biostratigraphic zonation schemes and species ranges remain, this work provides new insights in North Atlantic planktonic foraminifera biostratigraphy, taxonomy, and assemblages. A novel aspect of this study is that it is the only study that has recorded planktonic foraminifera qualitative relative abundance over the past 25 million years in the high-latitude North Atlantic compared to the DSDP/ODP/IODP reports in which assemblage relative abundances are based on estimates. Furthermore, this study recognizes the multiple central Arctic N. pachyderma morphotypes when N. pachyderma became abundant in the Pleistocene. This study provides a framework for the revision of the taxonomy and phylogeny of the Neogene and Quaternary planktonic foraminifera.

Appendix A: Planktonic foraminifera raw counts and relative abundance at DSDP Site 407

Table A1 Planktonic foraminifera assemblage species raw counts from Site 407. The raw counts are made on the > 125 µm size fraction. Species identified outside the counting area are made on the > 63 µm size fraction and are indicated with “R” (rare).

Table A2 Planktonic foraminifera assemblage species relative abundance (%) from Site 407. The raw counts are made on the > 125 µm size fraction. Species identified outside the counting area are made on the > 63 µm size fraction and are indicated with R (rare).

Appendix B: Calcareous nannofossil occurrence at Site DSDP 407

Table B1 Quantitative assessment of calcareous nannofossil occurrences and other constituents (biosilica) at Site 407. D is for dominant, P is for present, RW is for reworked, and R is for rare.

Appendix C: Chronologic ages of planktonic foraminifera and calcareous nannofossil events at DSDP Site 407

Table C1 Chronologic ages of planktonic foraminifera events in Site 407, based on calibrated ages from Raffi et al. (2020) (R20), Wade et al. (2011) (W11), and Weaver and Clement (1986) (WC86). The zonation is based on Raffi et al. (2020) and Wade et al. (2011). Primary zonal markers are indicated in bold. All species used in the age–depth model (Fig. 3) are indicated with a number. X indicates dominating coil change.

Table C2 Chronologic ages of calcareous nannofossil events in Site 407, based on calibrated ages from Raffi et al. (2020) (R20), Wade et al. (2011) (W11), and Weaver and Clement (1986 (WC86). The zonation is based on Raffi et al. (2020) and Wade et al. (2011). Primary zonal markers are indicated in bold. All species used in the age–depth model (Fig. 3) are indicated with a number.