Abstract

The Antarctic Geological Drilling (ANDRILL) program recovered a 1,138-m-long core, AND-2A, during the austral spring 2007–2008 from the Victoria Land Basin for the Southern McMurdo Sound (SMS) Project. The main objective of the SMS Project was to establish a Neogene (ca. 24–2.5 Ma) history recording ice volume changes from an ice-proximal site, in particular the environmental response of the Antarctic margin during the middle Miocene climate transition.

This study focuses on the foraminifera taxa recovered from AND-2A, providing a paleoenvironmental analysis including paleobathymetry and ice-proximity estimates. Three assemblages, Cibicides-Cassidulinoides, Ammoelphidiella, and Ehrenbergina, as well as two subassemblages within the Ammoelphidiella assemblage, Globocassidulina and Nonionella, were identified in this study using cluster analysis. Calcareous eurytopic taxa dominate every assemblage and display similarities between Fresh Shelf Water assemblages and Ice Edge Biofacies that were previously defined in other studies focused in the Southern Ocean, as well as assemblages recovered from previous drilling projects in the Ross Sea region.

Paleoenvironmental interpretations include an early Miocene record of fluctuating periods of glacial influence from ice-distal, abrupt cooling followed by a continued warming and transition into a more ice-proximal setting in the uppermost middle Miocene. The well-preserved middle Miocene record reflects more-productive stable environmental conditions with significant freshwater input from outlet glaciers and rivers at the East Antarctic margin, and coincides with previously published palynomorph data indicating a short abrupt warming. Furthermore, the persistent appearance of the planktonic species Neogloboquadrina pachyderma and glacial/sea ice assemblages can be related to the onset of progressive cooling at the Antarctic margin, including periods of fluctuating sea ice extent or glacial proximity. Finally, the Pliocene record contains taxa similar to that of Fresh Shelf Water assemblages and Ice Edge Biofacies recovered in the Southern Ocean today. Paleobathymetric estimates range from inner shelf depths (∼50–150 m) in the early Miocene reaching periods of outer-shelf depths (∼400 m) in the uppermost early–middle Miocene and Pliocene.

INTRODUCTION

Studying changes in Antarctica’s ice volume is important due to the recognition that southern high-latitude oscillations in climate have an impact far beyond Antarctica’s continental boundaries (Webb, 1990). The recent retreat of ice shelves in the Antarctic Peninsula region (Doake and Vaughan, 1991; Scambos et al., 2000; Domack et al., 2005), possibly as a result of prolonged warming, as well as predictions in estimated warming by the end of the century due to greenhouse gases (Houghton et al., 2001) makes understanding ice volume changes vital. While a large part of Antarctica’s ice volume history is interpreted through the deep-sea oxygen-isotope record (Miller et al., 1991; Zachos et al., 2001; Billups and Schrag, 2002; Holbourn et al., 2005; Holbourn et al., 2007), studies of material collected from the Antarctic margin provide the best opportunity to reconstruct an ice-proximal paleoenvironmental record (Ishman, 1994). The sedimentologic record of AND-2A (Antarctic Geological Drilling [ANDRILL] program, Southern McMurdo Sound [SMS] Project), through facies associations and motifs, has proven sediment accumulation has been preserved at this location proximal to Antarctica glacial influence, providing near-field ground truth of significant paleoclimatic changes (i.e., Middle Miocene Climatic Optimum events Mi1a, Mi1b, and Mi2) (Fielding et al., 2011; Passchier et al., 2011).

Investigating periods in Earth’s history that appear to have been warmer than today will help modelers better predict the mechanisms responsible for driving ice volume changes in the future. At the center of this understanding is the middle–late Miocene climate transition, marking the final shift from a polythermal ice sheet setting, coinciding with a long-term positive carbon excursion during a period of extensive global warmth and high sea level, to the modern cold polar setting or “ice house” conditions inferred from benthic foraminifera δ18O records reflecting strong glacial-interglacial cyclicity (Miller et al., 1991; Zachos et al., 2001; Holbourn et al., 2004; Holbourn et al., 2005). Previous drilling projects around the Antarctic margin have recovered late Eocene–early Miocene (e.g., Barrett and McKelvey, 1986; Barrett, 1989; Barrett, 2007) and younger late Miocene–Holocene (e.g., Ishman and Webb, 1988; Barrett and Hambrey, 1992; McKay et al., 2009; Naish et al., 2009) sediment records highlighting important changes in ice volume. However, not until ANDRILL’s SMS Project has an early–middle Miocene record been recovered from the Antarctic margin. The SMS core AND-2A recovered early–middle Miocene and Pliocene sediments (Acton et al., 2008–2009) filling the gaps of those previous drilling projects and adding to a more complete Cenozoic record of ice volume changes from the Antarctic margin. Furthermore, these marginal records can then be compared to the deep-sea oxygen-isotope record reflecting global events, which can potentially provide insight regarding the role of and feedbacks related to the Antarctica ice volume during major paleoclimatic episodes in the past. Finally, a more complete Cenozoic record will add insight in the ongoing debate over the stability of the East Antarctic Ice Sheet (EAIS) during the Neogene (Mayewski and Goldthwait, 1973; Webb et al., 1984; Harwood, 1985; Webb et al., 1986; Clapperton and Sugden; 1990; McKelvey et al., 1991; Ishman and Reick, 1992; Sugden et al., 1993; Harwood and Webb, 1998; Hambrey and McKelvey, 2000; Sugden et al., 1999).

AND-2A was recovered on the western margin of the Victoria Land Basin (77°45.488′S, 165°16.605′E) (Fig. 1) in ∼383.57 m water depth and with >98% core recovery. The Victoria Land Basin is a structural half graben hinged to the west (Fitzgerald, 1992). Flexure-related subsidence associated with rifting has provided accommodation space for sedimentation in the area that is influenced by the EAIS, the Ross Ice Shelf, Ross embayment sea ice, the West Antarctic Ice Sheet, alpine glaciers (Blue and Koettlitz), outlet glaciers (Ferrar and Taylor) from the EAIS, and volcanics (that are part of the West Antarctic Rift system), as well as the uplift of the Transantarctic Mountains throughout the Cenozoic (Fielding et al., 2007).

This study is focused on the foraminiferal assemblages recovered from AND-2A. The foraminiferal assemblages and subassemblages defined in this study will help add insight on the paleoenvironmental history of McMurdo Sound, Antarctica, for the early–middle Miocene and Pliocene and how they can be related to the global middle–late Miocene climatic transition. Paleobathymetry estimates based on benthic and planktonic taxa may provide further data concerning efforts to extract eustatic sea-level trends.

PREVIOUS STUDIES

Since the 1970s, ice-proximal drilling projects have taken place in the Ross Sea and McMurdo Sound, Antarctica, in order to develop a more complete understanding of the Cenozoic. The Deep Sea Drilling Project (DSDP) Leg 28, Sites 270, 271, 272, and 273, recovered a sediment history from the Ross Sea continental shelf that indicated major continental glaciations in Antarctica beginning in the late Oligocene (ca. 25 Ma), predating that of the Northern Hemisphere (Hayes and Frakes, 1975).

The Dry Valley Drilling Project (DVDP) sites 10 and 11, located on the eastern side of Taylor Valley, Antarctica, recovered an upper Miocene record that suggested a paleofjord setting that was influenced by oscillations in ice cover related to the EAIS (Ishman and Webb, 1988). The upper Miocene and lower Pliocene record provided evidence of periodic ice grounding (Ishman and Webb, 1988). Furthermore, uplift of the Transantarctic Mountains during the late Miocene and early Pliocene has been attributed to a rapid-shallowing sequence (Ishman and Webb, 1988).

The McMurdo Sound Sediment and Tectonic Study (MSSTS) recovered an upper Oligocene–Quaternary record with five glacial events during the late Oligocene (30–25 Ma) containing more glacial ice than present (Barrett and McKelvey, 1986). The MSSTS-1 drill core extended Antarctic glacial history back to ca. 31 Ma, 6 m.y. earlier than DSDP Leg 28 (Harwood, 1986; Leckie and Webb, 1986).

The Cenozoic Investigation of the Ross Sea (CIROS) project recovered two cores, CIROS-1 and CIROS-2. Micropaleontologic data from CIROS-1 indicated a late Eocene–early Miocene progressive shallowing trend (Coccioni and Galeotti, 1997; Webb, 1989), four periods of major ice buildup, and a large unified ice sheet on Antarctica (Webb, 1989).

The Cape Roberts Project (CRP) recovered three cores: CRP-1, CRP-2/2A, and CRP-3. The uppermost Eocene–lower Oligocene contained no direct evidence of an ice-marginal setting (Barrett, 2007). The Oligocene–lower Miocene indicated polythermal glacial conditions (Hambrey et al., 2002) with well-developed cyclicity occurring at Milankovitch frequencies (Naish et al., 2001). Micropaleontologic data from the Pliocene record indicated a glacially influenced embayment isolated from the influence of ocean circulation (Webb and Strong, 2000). The Quaternary record suggested a more extensive ice cover than present (Barrett, 2007).

The ANDRILL McMurdo Ice Shelf (MIS) Project recovered core AND-1B that recorded a dynamic late Miocene–Holocene ice sheet history including more than 60 glaciomarine sequences that showed a transition from subpolar, meltwater-influenced ice sheets in the late Miocene to the polar ice sheets and/or ice shelves of the Pleistocene (McKay et al., 2009). During the Pliocene, cycles of subglacial, glaciomarine, and open-marine lithofacies indicate cyclic advance and retreat of the West Antarctic Ice Sheet across the Ross embayment paced by obliquity (∼40 k.y.) Milankovitch cycles (Naish et al., 2009). CIROS-2 contained a discontinuous upper Miocene–Pleistocene record indicating cold-climate subglacial conditions during the Pleistocene, as well as fjordal and lake settings during the Pliocene (Barrett and Hambrey, 1992). The SMS project recovered a high-quality lower–middle Miocene and Pliocene record containing evidence of cyclicity, as well as varying degrees of ice extent over the drill site (Harwood et al., 2008–2009; Fielding et al., 2011; Passchier et al., 2011). A variety of environmental settings related to glacial advance and retreat cycles and variations in water depth are suggested for the SMS core (Harwood et al., 2008–2009). A high-quality palynomorph record indicates a short abrupt warming in Antarctica during the middle Miocene (Warny et al., 2009). This paper will focus on the benthic foraminifera record of AND-2A by defining assemblages and assessing their paleoenvironmental significance, as well as providing paleobathymetric depth estimates.

METHODS

Three-hundred and eighty-five samples from various intervals throughout all sections of core were examined for foraminiferal analysis. Additionally, 20 large thin sections were made commercially in order to specifically examine planktonic species. Sample processing included different disaggregation techniques based on sediment consolidation. The upper ∼120 m below sea floor (mbsf) consisted primarily of loosely consolidated sediment that was manually disaggregated into pebble-size particles or smaller. A mortar and pestle was used for more consolidated samples. Samples were then soaked in a 5% Calgon solution for 24 h.

Samples from 122 mbsf to ∼140 mbsf were manually disaggregated with a mortar and pestle, and samples from ∼140 mbsf were initially fragmented using a Carver Laboratory Press then further reduced to grape-size pieces using a mortar and pestle. The fragmented samples were then soaked in 3% H2O2 (hydrogen peroxide) for 24 h and boiled for 5 min.

After disaggregation, the samples were cooled and sieved using a 63 μm sieve. The 63 μm residue was then dried at <50 °C in order for specimens to be picked manually from dry samples until barren. For each sample, foraminifera preservation as defined by Strong and Webb (2000) was evaluated based on visual observation.

Statistical techniques commonly used in order to determine a detailed population structure had limited use because the number of specimens recovered from various intervals was small. Benthic foraminiferal assemblages in this study are based on cluster analysis of presence/absence data of benthic foraminifera using the Bray-Curtis coefficient in PRIMER software (version 5). Of the 385 sampled intervals, only 150 were used in cluster analysis. Barren intervals, as well as samples containing very few benthic specimens with only a single species appearing, were excluded.

Paleoenvironmental interpretations of foraminiferal assemblages rely on previous studies of modern distributions in the Southern Ocean, as well as previous drilling projects on the Antarctic margin. Depth estimates were formulated from a combination of sources citing modern depth ranges of benthic species in the Southern Ocean (Fig. 2), as well as the ecology and life cycle of the planktonic Globorotalia sp.

In order to be consistent with, and compare foraminiferal assemblages directly to, other paleoenvironmental proxies and interpretations of AND-2A, such as facies distribution (Passchier et al., 2011) and stratigraphic motifs (Fielding et al., 2011), the age model used in this study follows that from Fielding et al. (2011). This age model is compiled from an integration of magnetostratigraphy, Ar-Ar age dates, biostratigraphic indicators, Sr-isotope chronology, and sequence stratigraphy (Acton et al., 2008–2009; SMS Science Team 2010; DiVincenzo et al., 2010; Fielding et al., 2011). Table 1 (modified from Fielding et al., 2011) gives core depth and corresponding age ranges.

RESULTS

The foraminifera recovered from AND-2A are dominated by calcareous benthic taxa. Four planktonic taxa were recovered from middle Miocene and Pliocene intervals: Neogloboquadrina pachyderma, Antarcticella antarctica, Globorotalia sp., and Globorotalia cf. G. miotumida explicationis (Supplemental Table 1). Two agglutinated species were also identified, Bathysiphon sp. and Cystammina pauciloculata. In total, 37 genera represented by 78 species were identified. In some cases preservation was so poor that species identification could only be made to genus level. The most specimens recovered from any one sample was 218.

Three assemblages were identified using the results from the cluster analysis, Cibicides-Cassidulinoides, Ammoelphidiella, and Ehrenbergina (Fig. 3). The Ammoelphidiella Assemblage contains two subassemblages, Globocassidulina and Nonionella, also based on the same cluster analysis (Fig. 3). Names assigned to the assemblages and subassemblages are based on common taxa within the sample intervals. Table 2 displays the presence/absence distribution of taxa within each assemblage and subassemblage.

Variation in preservation and abundance of foraminifera indicates degrees of taphonomic alteration throughout the core. Most specimens show obvious signs of recrystallization and/or corrosion to some degree. Furthermore, acidity down-core (Panter et al., 2008–2009) may have led to the dissolution of biogenic carbonate. However, there are intervals of abundant and well-preserved biogenic carbonate (including foraminifera) extending down into the lower Miocene record. Rapid sedimentation due to the dynamic nature of glacial-interglacial cycles and uplift in the Transantarctic Mountains throughout the Cenozoic may also play a role in the relatively low abundances of foraminifera tests recovered from some intervals in AND-2A.

Cibicides-Cassidulinoides Assemblage

The Cibicides-Cassidulinoides Assemblage occurs throughout the AND-2A record (Fig. 4). This Assemblage contains 19 genera represented by 30 species (Table 2). Notable taxa within this assemblage include Cibicides spp., Cassidulinoides spp., Oolina spp., Eponides bradyi, Ammoelphidiella sp., and Fursenkoina schreibersiana. This assemblage is most common in motif 1 of Fielding et al. (2011), while sparsely occurring within boundary layers of motifs 3 and 4b (Fig. 4).

Ammoelphidiella Assemblage

The Ammoelphidiella Assemblage occurs throughout the AND-2A record (Fig. 4). The increased abundance of Ammoelphidiella sp., Nonionella sp., Globocassidulina spp., and Pullenia sp. helps define this Assemblage. Fursenkoina schreibersiana, Epistominella exigua, and Rosalina globularis are other common species in this Assemblage. Two subassemblages, Globocassidulina and Nonionella, subdivide the Ammoelphidiella Assemblage. The subassemblage Globocassidulina contains 33 genera represented by 59 species (Table 2) and is most dominant through the middle Miocene record of AND-2A (Fig. 4C). Common taxa in this subassemblage include Globocassidulina subglobosa, Globocassidulina sp., Nonionella iridea, Ammoelphidiella sp., Cibicides sp., Pullenia sp., Elphidium magellanicum, Fursenkoina schreibersiana, and Epistominella exigua, which account for ∼80% of the foraminifera recovered from this subassemblage. The planktonic species Globorotalia sp. occurs commonly within sampled intervals corresponding to this subassemblage (Fig. 4B and 4D). Furthermore, this subassemblage commonly occurs in intervals of highest foraminifera abundance (Fig. 4), of best specimen preservation, and within motifs 1, 2, and 3 (Fig. 4).

The subassemblage Nonionella contains 18 genera represented by 30 species (Table 2) and occurs throughout the AND-2A record (Fig. 4). Representative taxa include Cibicides sp., Nonionella iridea, and Ammoelphidiella sp. Additionally, Fursenkoina scheibersiana, Epistominella exigua, and Rosalina globularis occur but in lower abundance than in subassemblage Globocassidulina. Subassemblage Nonionella is distinct from subassemblage Globocassidulina by the decrease in specimen abundance as well as an almost complete absence of Globocassidulina spp. This subassemblage occurs most commonly within motif 1 but also largely at motif boundary layers (Fig. 4).

Ehrenbergina Assemblage

The Ehrenbergina Assemblage occupies predominantly the uppermost middle Miocene and Pliocene record of AND-2A (Fig. 4). Thirty-one genera are represented by 55 species within this Assemblage (Table 2). However, 36 species occur fewer than 5 times and 15 species occur only once. Dominant taxa in this Assemblage are Ehrenbergina glabra, Globocassidulina sp., Globocassidulina subglobosa, Cassidulina porrectus, and Angulogerina earlandi. However, Angulogerina earlandi is almost exclusively in one interval (41.21–41.23 mbsf). Less-common taxa include Eponides bradyi, Ammoelphidiella sp., Cibicides spp., and Quinqueloculina spp., with rare occurrences of Fursenkoina schreibersiana, Epistominella exigua, Cassidulinoides parkerianus, and Nonionella iridea, which also help define this Assemblage. It most commonly occurs within ice-proximal motifs 1 and 4a (Fig. 4).

DISCUSSION

Early Miocene

The early Miocene (ca. 20.02 to ca. 15.9 Ma) record of AND-2A is dominated by the Cibicides-Cassidulinoides Assemblage and the subassemblages of the Ammoelphidiella Assemblage (Fig. 4A–4B). The Cibicides-Cassidulinoides Assemblage contains low-diversity eurytopic (Cassidulinoides, Eponides, Ammoelphidiella, Cibicides, Pullenia, Epistominella, and Fursenkoina), opportunistic (Globocassidulina and Epistominella), and dissolution-resistant (Cibicides, Melonis, Epistominella, and Pullenia) taxa. The subassemblage Globocassidulina represents a greater diversity of taxa and higher abundance of specimens compared to the Cibicides-Cassidulinoides Assemblage and the subassemblage Nonionella, but has similar eurytopic (Nonionella, Cassidulinoides, Eponides, Ammoelphidiella, Cibicides, Pullenia, Elphidium, Fursenkoina and Epistominella), opportunistic (Globocassidulina and Epistominella), and dissolution-resistant (Epistominella, Elphidium, Nonionella, Melonis, Pullenia, and Cibicides) taxa. Subassemblage Nonionella is composed predominantly of low-diversity infaunal species, whereas, subassemblage Globocassidulina contains greater species diversity with a mixture of infaunal and epifaunal species.

Both the Cibicides-Cassidulinoides Assemblage and the Ammoelphidiella Assemblage contain a high abundance of epibenthic species Cibicides spp. and Rosalina globularis. These epibenthic species are important in that they indicate the presence of high bottom-current activity due to their attached mode of life (Murray, 1991; Ishman and Szymcek, 2003). Their occurrence at ∼820–780 mbsf is significant in that this interval is punctuated by barren intervals coinciding with a Miocene deep-sea hiatus (ca. 20.0 to ca. 18.0 Ma), potentially caused by an increased rate of circulation and corrosiveness of bottom waters during the periods of enhanced polar glaciation (Barron and Keller, 1982).

The abundance of Rosalina globularis in the Ammoelphidiella Assemblage at ∼820–780 mbsf and ∼625–550 mbsf gives insight into paleobathymetry. Similar taxa were recovered from surface samples collected in 95–124 m of water depth from the George V–Adelie continental shelf (Milam and Anderson, 1981). Milam and Anderson (1981) associated their assemblage abundant with Rosalina globularis with Fresh Shelf Water. Modern sediments recovered from Explorers Cove, McMurdo Sound, Antarctica, contained Rosalina globularis from depths as shallow as <27 m (Bernhard, 1987). Modern studies from the Antarctic Peninsula region have noted Rosalina globularis appearances in trough assemblages at depths ∼900 m that were also associated with Fresh Shelf Water (Ishman and Szymcek, 2003). Additionally, Elphidium sp. is common to this early Miocene record and has been associated with shallow water (Herman and Hopkins, 1980; Bergsten, 1994; Conradsen, 1995; Gustafsson and Nordberg, 1999; Ishman and Szymcek, 2003) (Fig. 2) and fluctuating- or lower-salinity environments (Jiang and Klingberg, 1996; Hald and Korsun, 1997). The abundance of shallow-water taxa Rosalina globularis and Elphidium sp. in addition to deeper-water benthic species Epistominella exigua and Cassidulina porrectus (Fig. 2) indicates that shallow-water species were living at their depth limits of ∼50 m. Furthermore, these foraminifera intervals correlate with sedimentological evidence of high-latitude temperate glacial regimes (motif 3 and 4b), as well as periods of meltwater incursions (motif 2) (Fig. 4A). Therefore, an estimation of water depth of ∼50–150 m for the early Miocene under the influence of freshwater is a reasonable interpretation (Figs. 2 and 4A).

The broad depth range of Rosalina globularis highlights findings from studies concerning modern foraminifera ecology around the Antarctic that suggest water mass rather than water depth is a key contributor to their distribution, due to an irregular carbonate compensation depth (CCD) (Kennett, 1968; Anderson, 1975; Osterman and Kellogg, 1979). Localities with heavy pack ice do not favor primary productivity resulting in increased dissolved CO2 content in the water and shoaling of the CCD. This type of environment is more acidic and favors agglutinated foraminifera taxa at the expense of calcareous taxa (Osterman and Kellogg, 1979), a situation not observed in the AND-2A assemblages. The oxygenated water mass, Ross Sea Shelf Water, is thought to depress the CCD in the southwestern Ross Sea today allowing for the precipitation and preservation of biogenic carbonate (Osterman and Kellogg, 1979), and includes the foraminifera Ehrenbergina glabra, Globocassidulina subglobosa, and Cassidulinoides porrecta that are very similar to taxa recovered from AND-2A. Fresh Shelf Water forms from the seasonal melt of sea ice during the summer months (Milam and Anderson, 1981), but can also be influenced by fresh water input via glacier meltwater, rivers, streams, and/or an increase in precipitation, a situation also favoring the preservation of biogenic carbonate as seen in modern foraminifera distributions in the eastern Weddell Sea (Anderson, 1975). Many solution-resistant taxa occur within the Cibicides-Cassidulinoides and Ammoelphidiella assemblages, particularly subassemblage Nonionella (i.e., Epistominella exigua, Elphidium magellanicum, Nonionella iridea, Cibicides sp., Melonis sp., and Pullenia sp.) (Kennett, 1968; Leckie and Webb, 1983, 1986), perhaps indicating periods during which preservation of biogenic carbonate was close to its limit. The absence of agglutinated taxa indicates a lower CCD most likely due to primary productivity and/or the mixing of water masses (Osterman and Kellogg, 1979). The preliminary diatom record shows their occurrences sparse below ∼432 mbsf in AND-2A (Taviani et al., 2008–2009), suggesting that the preservation of biogenic carbonate may also be due to the mixing of water masses resulting from buildup and melting of ice along the coastal margin in the McMurdo Sound region of Antarctica rather than primary productivity. Therefore, the early Miocene benthic foraminiferal assemblages as indicated by the preservation of biogenic carbonate and dominance of solution-resistant taxa indicate an environment under the influence of freshwater input.

Abundant foraminifera intervals represented by Cibicides-Cassidulinoides and Ammoelphidiella assemblages coincide with interpreted warming periods in the early Miocene (∼812–789 and ∼648–569 mbsf) and within motifs representing ice-distal environments (motifs 3 and 4b) (Fielding et al., 2011; Passchier et al., 2011) (Fig. 4A). Macropaleontologic evidence of shell fragments suggests episodic influxes of turbid water (Taviani et al., 2008–2009), which correlates well with the presence of epibenthic foraminiferal species that indicate high bottom-current activity, as well as an influence of Fresh Shelf Water enhancing the preservation of biogenic carbonate. However, the presence of pectinids in early Miocene intervals with abundant foraminifera also indicates periods of calm water and low sedimentation, which aid in the preservation and increased abundance of foraminifera. Furthermore, intervals barren of foraminifera occur within glacially dominated facies (Passchier et al., 2011) and motif 2 (Fielding et al., 2011) representing an abrupt cold ice-proximal setting with a grounding line close to the drill site, and corresponding with the Mi1b glaciation at 17.7 Ma (Fielding et al., 2011; Passchier et al., 2011) (Fig. 4A).

Following the abrupt Mi1b cooling period of the lower Miocene (ca. 17.35 to ca. 15.9 Ma) record, AND-2A exhibits a slight transition from a more ice-distal to ice-proximal setting into the uppermost lower Miocene (Fig. 4A and 4B). The uppermost lower Miocene is dominated by the presence of subassemblage Globocassidulina (∼500 to ∼350 mbsf). Subassemblage Globocassidulina co-occurs with pectinids and open-marine diatoms indicating lower sedimentation rates, more open marine conditions, and greater nutrient availability in a less glacially dominated environment (Fig. 4B). Furthermore, the subassemblage Globocassidulina occurs in samples that contain the planktonic foraminifera Globorotalia sp. The life cycle of some Globorotalia taxa is significant in that part of the life cycle includes migrating to water depths >50 m (Lipps, 1979; Norris et al., 1994). When combined with the two subassemblage benthic populations, the presence of the Globorotalia sp. supports water depth estimates between outer-shelf and upper-slope depths (>50 and 400 m) (Figs. 2 and 4B).

As discussed earlier, subassemblage Nonionella contains more infaunal species, as well as lower abundance of foraminifera, compared to subassemblage Globocassidulina, which contains a higher diversity of infaunal and epifaunal species, as well as a greater abundance in specimens. The infaunal and dissolution-resistant species of subassemblage Nonionella and the common association with motif boundary layers indicate stressed environmental conditions such as a lack of nutrient availability, possibly from a combination of influences: perennial sea ice cover or sub–ice-shelf conditions, both resulting in the lack of primary productivity and supported by the paucity of diatoms (Taviani et al., 2008–2009). Furthermore, the emergence of the Ehrenbergina Assemblage with taxa associated with Ice Edge Biofacies follows the appearance of subassemblage Nonionella at the onset of and within motif 2 representing a subpolar glacial regime.

The lower Miocene Ammoelphidiella subassemblages are recovered from a stratified diamictite-mudstone facies indicating fluctuations in the extent of tidewater glaciers draining the East Antarctic Ice Sheet ca. 17.35 to ca. 15.9 Ma with a grounding line south of the modern position (Passchier et al., 2011). These fluctuations during the early Miocene regarding glacier extent, and the onset of progressive cooling in the latest early Miocene, are highlighted by both the cyclostratigraphic record of interbedded biosiliceous mudstones, mudstones with clasts, interlaminated sandstones, stratified diamictites, and conglomerates within motifs 2 and 3 (Fielding et al., 2011; Passchier et al., 2011), as well as the foraminiferal alternations between subassemblages Globocassidulina and Nonionella (Fig. 4A–4B). Additionally, predating the emergence of the Ehrenbergina Assemblage in the uppermost lower Miocene record, the first appearance of the modern polar planktonic species Neogloboquadrina pachyderma is observed (Fig. 4B). The alternation of Ammoelphidiella subassemblages, in addition to the first appearance of Neogloboquadrina pachyderma, and the emergence of the Ehrenbergina Assemblage in the uppermost lower Miocene highlights a dynamic nature of environmental conditions during this part of the early Miocene with the potential onset of progressive cooling (Fig. 4A–4B).

Webb and Strong (2006) suggested the upper Oligocene and lower Miocene foraminifera record of CRP-2/2A may be related to cyclical events. The CRP-2/2A foraminifera record contains similar eurytopic and opportunistic species to that of the subassemblage Globocassidulina. In CRP-2/2A eurytopic and opportunistic species were noted to dominate the continental shelf during interglacial periods (Webb and Strong, 2006). Periods of both rapid (Passchier et al., 2011) and low sedimentation rates indicated by the presence of pectinids (Taviani et al., 2008–2009) co-occurring with intervals of abundant foraminifera further supports the argument that the early Miocene record of AND-2A within the Ammoelphidiella Assemblage was cyclic, fluctuating between sea ice and/or strongbottom-water currents during glacials and open-marine conditions in interglacials.

Middle Miocene

The middle Miocene record of AND-2A (∼305–135 mbsf) is dominated by subassemblage Globocassidulina (Fig. 4C). The sample intervals of middle Miocene age contain some of the best foraminifera preservation and the highest specimen abundances. The good preservation of calcareous specimens is likely the result of freshwater input from streams, rivers, lakes, precipitation, and/or melting ice, as well as primary productivity. The high abundance is associated with food availability as Epistominella exigua and Globocassidulina subglobosa are opportunistic species commonly associated with high nutrient flux in deep-sea environments (Gooday, 1988). The middle Miocene record of foraminifera co-occurs with palynomorph occurrences, suggesting a short abrupt period of warming with increased freshwater input potentially aiding in the preservation of biogenic carbonate as well as nutrient availability (Warny et al., 2009).

The Ehrenbergina Assemblage begins to appear consistently in the middle Miocene of AND-2A, as well as the planktonic species Neogloboquadrina pachyderma (Fig. 4C). The onset of the appearance of the Ehrenbergina Assemblage is significant in that species associated with this assemblage (i.e., Ehrenbergina glabra) have been associated with ice-edge conditions. Anderson (1975) noted Ehrenbergina glabra and Globocassidulina subglobosa with Fresh Shelf Water in the eastern Weddell Sea. Osterman and Kellogg (1979) related the appearance of Ehrenbergina glabra, Globocassidulina subglobosa, Cassidulina porrectus, and Angulogerina earlandi in Pleistocene sediments to shallow areas of the Ross Sea, Antarctica and Ross Sea Shelf Water. Ishman and Szymcek (2003) associated Ehrenbergina glabra with Ice Edge Biofacies in the Antarctic Peninsula region. The low-diversity, eurytopic, epibenthic, solution-resistant taxa of the Cibicides-Cassidulinoides Assemblage also indicates glacial influence with periods of turbid water conditions, perhaps from subglacial melting events rather than sea ice, as sea-ice diatoms are absent from the uppermost middle Miocene record dominated by motif 1, which represents a cold subpolar/polar glacial regime (Fig. 4C). Furthermore, in conjunction with the appearance of the Ehrenbergina Assemblage, the planktonic foraminifera Neogloboquadrina pachyderma within this interval indicates the onset of true polar conditions influenced by ice (Kennett, 1968; Lipps, 1979; Ishman and Szymcek, 2003). This interpretation is consistent with both sedimentologic facies associations and motifs, as the Ehrenbergina and Cibicides-Cassidulinnoides assemblages occur within motif 1 indicating further progressive cooling since the latest early Miocene (ca. 16.4 Ma) (Fig. 4C).

Pliocene

The Pliocene record contains all three assemblages, as well as the three planktonic species Neogloboquadrina pachyderma, Globorotalia sp., and Antarcticella Antarctica (Fig. 4D), and is most prevalent to motif 1 representing a cold, subpolar/polar glacier regime. The dominance of the Ehrenbergina and Cibicides-Cassidulinoides assemblages, in addition to other paleoenvironmental proxies (i.e., sedimentology facies, motifs, and diatoms) strongly suggest an environment under the influence of a polar glacial regime. Taxa representing Ice Edge Biofacies of the Ehrenbergina Assemblage, and the abundance of the planktonic species Neogloboquadrina pachyderma as well as epibenthic taxa of the Cibicides-Cassidulinoides Assemblage indicating strong bottom-current activity co-occurring with sea-ice diatoms, provides insight into a Pliocene record dominated by a strong glacial influence. While this is an incomplete Pliocene record (Fielding et al., 2011), AND-2A provides significant insight concerning the onset of a cold polar glacial regime potentially transitioning in the latest middle Miocene (ca. 14.5 Ma) and being fully reached by the Pliocene (ca. 5.0 Ma) for McMurdo Sound. The upper Miocene record, while absent from AND-2A, is captured in the base of the AND-1B record and exhibits cold polar conditions (motif 1 in McKay et al., 2009) preceding the subpolar conditions in the latest Miocene (<13.4 Ma). Pliocene benthic and planktonic foraminifera species of AND-2A indicate increased glacial conditions including persistent ice cover in a shelf to upper-slope setting with water depths up to ∼400 m (Figs. 2 and 4D).

Some reworking of older sediment is indicated in the Pliocene record based on the occurrence of the planktonic species Antarcticella sp. Leckie and Webb (1985) suggested Antarcticella antarctica (Candeina antarctica) occurred from late Oligocene to middle Miocene. However, Antarcticella antarctica appears only within the Pliocene intervals of AND-2A. Reworking of older material is expected in an environment under the influence of glaciers. That being said, we consider the reworking of middle Miocene planktonic species to add possible further evidence that a cold polar glacial regime was in place by the middle Miocene. Reworking is considered to be minor and does not affect our paleoenvironmental interpretations by and large, due to the agreement with other studies concerning Pliocene foraminiferal assemblages recovered in the Ross Sea region (i.e., Ishman and Webb, 1988), in addition to AND-2A sequences stratigraphy, sedimentology facies associations and paleontologic findings (Fielding et al., 2011; Passchier et al., 2011; Warny et al., 2009), as well as micropaleontologic evidence of diatoms indicating a sea-ice influence (Taviani et al., 2008–2009).

CONCLUSIONS

The three benthic foraminiferal assemblages recovered from ANDRILL’s SMS project core AND-2A provide evidence for a fluctuating Neogene glacial history for the McMurdo Sound, Antarctica. Interpretations based on assemblages are consistent with sedimentologic (Fielding et al., 2011; Passchier et al., 2011), palynomorph (Warny et al., 2009), and paleontologic (Taviani et al., 2008–2009) proxies.

The lower Miocene foraminifera record of AND-2A indicates fluctuating periods of glacial influence coinciding with changes in sedimentary motifs and facies (Fielding et al., 2011; Passchier et al., 2011). The common occurrence of dissolution-resistant taxa of subassemblage Nonionella and epibenthic species of the Cibicides-Cassidulina assemblage indicates the influence of freshwater input and turbid water conditions during warming events in an ice-distal environment of shelfal water depth (50–150 m) (Fig. 4A). The dominance of eurytopic taxa of subassemblage Globocassidulina from ca. 16.7 to ca. 16.2 Ma, in addition to other paleotonologic proxies (Taviani et al., 2008–2009), indicates episodes of less glacially dominated environments of outer-shelf to upper-slope (>50–400 m) depth (Fig. 4B). The middle Miocene record contains some of the best-preserved and abundant intervals of foraminifera, followed by the persistent occurrence of the Ehrenbergina Assemblage, indicating the onset of ice-proximal conditions after 15.9 Ma. While the Pliocene record is incomplete, taxa representing Ice Edge Biofacies (i.e., Ehrenbergina glabra, Globocassidulina subglobosa, Cassidulina porrectus, and Anguloberina earlandi), strong bottom-water current activity (i.e., Cibicides-Cassidulina Assemblage), and recycled middle Miocene planktonic species co-occurring with sea-ice diatoms provide insight concerning the onset of a cold polar glacier regime, with further support from sedimentologic facies (Passchier et al., 2011) and motifs (Fielding et al., 2011).

This material is based upon work supported by the National Science Foundation under Cooperative Agreement No. 0342484 through subawards administered and issued by the ANDRILL Science Management Office at the University of Nebraska–Lincoln, as part of the ANDRILL U.S. Science Support Program. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation.

The ANDRILL Program is a multinational collaboration between the Antarctic programs of Germany, Italy, New Zealand, and the United States. Antarctica New Zealand is the project operator, and has developed the drilling system in collaboration with Alex Pyne at Victoria University of Wellington and Webster Drilling and Enterprises Ltd. Scientific studies are jointly supported by the U.S. National Science Foundation, New Zealand Foundation for Research Science and Technology, Royal Society of New Zealand Marsden Fund, the Italian Antarctic Research Programme, the German Research Foundation (DFG), and the Alfred Wegener Institute for Polar and Marine Research (Helmholtz Association of German Research Centres). Antarctica New Zealand supported the drilling team at Scott Base; Raytheon Polar Services supported the science team at McMurdo Station and the Crary Science and Engineering Laboratory. The ANDRILL Science Management Office at the University of Nebraska–Lincoln provided science planning and operational support.

The authors would like to thank reviewers Mark Leckie and Hugh Morgans for their valuable insights and recommendations, which made this manuscript better and more comprehensive.

1Supplemental Table. Excel file of supplemental data. If you are viewing the PDF of this paper or reading it offline, please visit http://dx.doi.org/10.1130/GES00771.S1 or the full-text article on www.gsapubs.org to view the Supplemental Table.