Oceanographic influences on the stability of the Cosgrove Ice Shelf,Antarctica

Abstract

Ferrero Bay, located in eastern Pine Island Bay (PIB) of the Amundsen Sea Embayment, is one of the largest and southernmost fjords yet studied in Antarctica. High-resolution multibeam swath bathymetric data, chirp sonar sub-bottom profiles, and three Kasten cores were collected in Ferrero Bay during the IB Oden Southern Ocean 2009–2010 cruise (OSO0910). Core KC-15 from the inner bay yielded two carbonate ages providing a minimum age for ice sheet recession from this sector of PIB by ~11 cal. kyr BP. In total, seven additional acid insoluble organic (AIO) fraction radiocarbon ages provide a linear age model with an R² of 0.99. Variations in magnetic susceptibility, grain size, total organic carbon (TOC) and nitrogen, diatom abundance, and foraminiferal assemblage and abundance are used to interpret glacial history and paleoceanographic conditions. Grounding line retreat was characterized by advection of planktic foraminifera beneath an ice shelf that may have extended across the middle continental shelf. Following initial deglaciation, the Cosgrove Ice Shelf covered Ferrero Bay, and productivity was virtually absent during the mid-Holocene, while benthic foraminifera indicate periodic incursion of warm Circumpolar Deep Water. The ice shelf persisted until 2.3 cal. kyr BP, when TOC and diatom abundance increased as the bay opened and coastal areas deglaciated. Abundant diatoms demonstrate open marine conditions and seasonal sea ice during the recent open water phase, while high benthic foraminiferal abundance indicates active benthos. The retreat of the Cosgrove Ice Shelf was out of phase with Antarctic Peninsula ice shelves and ice-core proxy temperatures, implying that it did not respond to Holocene climate events but rather to the influence of Circumpolar Deep Water and possibly to internal glacial dynamics.

Keywords

Amundsen Sea Circumpolar Deep Water diatoms foraminifera Pine Island Bay

Introduction

The collapse of several fringing ice shelves in the Antarctic Peninsula (AP) over the last three decades has been linked to poleward shift of isotherms due to atmospheric warming (Figure 1; Cook and Vaughan, 2010; Morris and Vaughan, 2003). Ice shelf collapse was most dramatically illustrated by the Larsen B Ice Shelf which broke up in March 2002, causing outlet glaciers to accelerate, increase ice mass losses, and contribute to sea-level rise (Cook and Vaughn, 2010; De Angelis and Skvarca, 2003; Morris and Vaughan, 2003; Scambos et al., 2003, 2004). Today, the remaining AP ice shelves exist where mean annual atmospheric temperatures (MAATs) are below −9°C; above this threshold, surface melting plays a more active role, which can lead to enhanced crevasse propagation and eventual collapse (Figure 1; Doake and Vaughan, 1991; Hughes, 1981; Morris and Vaughan, 2003).

Figure 1.

Map of the Antarctic Peninsula (AP) and the Amundsen Sea Embayment (ASE), illustrating the climatic limit of ice shelf stability. Ice shelf collapse has followed the southward migration of the −9°C MAAT isotherm (Morris and Vaughan, 2003; Vaughan et al., 2003). The western AP and ASE are also under the influence of relatively warm Circumpolar Deep Water.

Recent hydrographic studies demonstrate that warmer water masses also destabilize ice shelves (e.g. George VI IS; Bentley et al., 2009). The strongest case is in the Amundsen Sea Embayment (ASE), where oceanographic measurements have causally linked melting of the underside of Amundsen ice shelves to Circumpolar Deep Water (CDW) at rates >40 m/yr (Figures 1 and 2; Jacobs et al., 1996, 2011, 2012, 2013; Jenkins et al., 2010, 2012; Rignot and Jacobs, 2002). Furthermore, coupled three-dimensional (3D) ocean-ice shelf models project that Pine Island Glacier will irreversibly retreat after 20 years of warm ocean conditions (De Rydt and Gudmundsson, 2016). Cook et al. (2016) have further suggested that CDW is the primary cause of current ice recession in the AP region.

Figure 2.

(a) Bathymetric map of the Amundsen Sea Embayment modified from Jakobsson et al. (2012); (b) multibeam swath bathymetry collected during the OSO0910 cruise in Ferrero Bay, with core locations of KC-15, KC-16, and KC-17 marked with red stars and furrow orientations with black arrows; and (c) closer view of distal bathymetry, highlighting the megascale glacial lineations (MSGLs) located seaward of crag-and-tail features.

Questions remain as to how ice shelves in higher latitudes will react to continued regional warming and what role oceanographic processes, particularly in response to CDW encroachment onto the continental shelf, will play in their stability. These questions are being addressed through numerical modeling, but geological records of past ice shelf behavior, particularly response to atmospheric and oceanographic influences, can provide critical constraints on ice shelf dynamics. Here, we present results from a multi-proxy analysis of sediment cores from Ferrero Bay from which we interpret the behavior of the Cosgrove Ice Shelf during the Holocene within the context of regional terrestrial and marine data. The Cosgrove Ice Shelf in the northeastern part of Pine Island Bay (PIB) experiences MAAT around −16°C, which is well below the −9°C threshold at which AP ice shelves have collapsed in recent years (Figure 1; Morris and Vaughan, 2003). CDW is also observed within eastern ASE. Ferrero Bay, therefore, presents an important site in which climatic and oceanographic controls on ice shelf stability can be tested.

Background

Geologic setting

Ferrero Bay is a narrow embayment located within eastern PIB of the ASE (Figures 1 and 2). The bathymetry of Ferrero Bay is structurally controlled; it is part of a series of fault-bounded rift basins related to Antarctica’s breakup with New Zealand in the Late Cretaceous (e.g. Dalziel and Elliot, 1982; Gohl, 2012; Storey et al., 1991). The northern perimeter of Ferrero Bay is characterized by water depths up to 1300 m, some of the greatest depths of PIB, due to an east–west-oriented fault on the northern side of the basin that is associated with the diffuse Bellinghausen Plate Boundary (Figure 2; Cochran et al., 2015; Gohl, 2012). This is bounded by the E-W elongate bedrock high, which forms King Peninsula and separates the Cosgrove Ice Shelf from the Abbot Ice Shelf to the north (Figure 2). The southern side of Ferrero Bay averages a depth of ~700 m and is bounded by Canisteo Peninsula. King Peninsula and the inland Hudson Mountains (Figure 2) are composed of olivine basalts and tuffs. By contrast, the Canisteo Peninsula to the south is dominated by basement complex consisting of granites, diorites, and gneisses, likely of Triassic and Cenozoic age (Wade and LaPrade, 1969), which is reflected in the clay mineralogical signature of surface sediments at the mouth of Ferrero Bay (Ehrmann et al., 2011).

Climatic and oceanographic setting

The Cosgrove Ice Shelf is located within the Eights Coast, which is characterized by a polar climate, with −16°C MAAT (Figure 1; Morris and Vaughan, 2003). Precipitation on the Eights Coast is relatively high, however, because of low-pressure systems that travel onshore during the austral winter (Vaughan et al., 2003).

The ASE is one of the most remote and under-studied regions of Antarctica, due in large part to persistent sea-ice cover (e.g. Jacobs et al., 2012). Sea ice has decreased significantly during recent decades (Parkinson and Cavalieri, 2012), however, and during the austral summer of 2009, the IB Oden encountered open seas, resulting in successful coring and geophysical efforts (Anderson et al., 2011). One previous cruise to outer Ferrero Bay collected Piston cores (Kellogg and Kellogg, 1987a, 1987b), but the cores were never studied and were eventually compromised. Smith et al. (2014) also analyzed two gravity cores recovered directly to the south of King Peninsula that support early-Holocene deglaciation of the area.

Mean summer sea surface temperature in Ferrero Bay was −0.9° C when the cores were collected (Locarnini et al., 2010). Water mass temperatures within PIB typically range between −1.5°C and 0°C. The exception to this is dense, relatively warm (by up to 3.5°C) CDW that impinges into the PIB along troughs that are up to 1000 m deep (Figures 1 and 2; Jacobs et al., 2011, 2013; Walker, et al., 2007). Ferrero Bay is part of a long, narrow, landward sloping trough that connects to the shelf edge in the eastern ASE (Figure 2). Water column profiles from CTD and XBT data collected during the IB Oden cruise in 2010 clearly show impinging CDW, with sharp temperature and salinity increase from 275 to 500 m water depth to ~+1.2°C and ~34.6 ppm, respectively (Figures 3 and 4; Anderson et al., 2010).

Figure 3.

Water column temperature and salinity measured during OSO0910 cruise into Ferrero Bay (taken at site of KC-15, shown in Figure 4). Note temperature and salinity increase below 275 m water depth, indicating the presence of Upper Circumpolar Deep Water.

Figure 4.

(a) OSO0910 sub-bottom profiler Line 98A from the axis of Ferrero Bay, showing the small basin with <2 m sediment where KC-15 was collected. (b) Bathymetric map with cruise tracklines, CTD and XBT sites, and Line 98A highlighted in white.

Productivity in the Amundsen Sea is among the highest in the Southern Ocean, with long-lived phytoplankton blooms related to the Pine Island and Amundsen polynyas (Arrigo et al., 2008; Smith and Comiso, 2008; Thuróczy et al., 2012). In contrast to the well-studied Ross Sea, less is known about modern distributions of phytoplankton communities in the Amundsen Sea. An interesting aspect of the Amundsen Sea is that phytoplankton blooms are dominated by haptophyte Phaeocystis antarctica (Arrigo et al., 2014), which have low preservation potential in seafloor sediments. Recent studies have linked phytoplankton blooms to melt events of Pine Island Glacier (Fragoso and Smith, 2012; Gerringa et al., 2012; Thuróczy et al., 2012). Modified CDW has also been argued to bring iron to the shelf as it travels along deep troughs into PIB, resuspending sediment that may facilitate blooms (Dinniman et al., 2003). Diatoms, which have high preservation potential, make up a major fraction of these blooms and are present in open-ocean and ice-marginal settings (Fragoso and Smith, 2012; Leventer and Dunbar, 1996).

Methods

Geophysical methods and coring

High-resolution multibeam swath bathymetry data were collected using the Kongsberg EM122, 12 kHz, 1° × 1° multibeam system installed in IB Oden (Figure 2). The bathymetry data were processed using Fledermaus and exported as grids that were imported to the GIS software ArcMap 10.1 for further analyses. The hull-mounted Kongsberg SBP120, 3° × 3°, chirp sonar system on IB Oden was used to study the shallow stratigraphy, map sediment distribution, and survey for optimal coring sites. Acquired sub-bottom profiles were processed and displayed using Kongsberg’s SBP/Topas software (Figure 4). The chirp sonar system was operated continuously using a 2.5- to 7-kHz 35-ms-long pulse.

In total, three cores were collected in Ferrero Bay during the IB Oden OSO0910 cruise (Figure 2). KC-15 (73.3603°S, 101.8362°W) was taken in the innermost part of the fjord at 1274 m and recovered 1.3 m of sediment. KC-16 (73.454°S, −102.0792°W; 706 m) was taken on a structural high ~700 m and recovered only 0.4 m of sediment. KC-17 (73.4197°S, 102.8267°W; 855 m) was taken in the outer bay and recovered 1.4 m of sediment. Detailed core descriptions include Munsell soil color determination, preliminary grain size and shape, pebble lithology, sedimentary structures, wet sieving for carbonate material, and magnetic susceptibility. When KC-15 and KC-17 were archived, the core catcher material was appended to the bottom of the cores.

Radiocarbon analysis

In total, two marine carbonate samples were collected from KC-15 for radiocarbon analysis and reported by Kirshner et al. (2012; Table 1). To build a more complete chronology of the core, seven bulk sediment samples were selected from KC-15 and analyzed by University of Tokyo AMS laboratory (YAUT) for AIO fraction ¹⁴C dating (refer to YAUT for procedures; Yokoyama et al., 2007, 2010).

Table 1.

Radiocarbon ages from Ferrero Bay.

Laboratory sample ID	Core	Depth (cmbsf)	d¹³C (‰)	±	Fraction modern	±	¹⁴C age (BP)	±	LCO-corrected age yr BP	Calibrated age (cal. yr BP)	±	Sample type
1. YAUT-015117	KC-15	0–2	−49.4	4.8	76.76	0.55	2124	57	1300	−60	219	Organic carbon
2. YAUT-015118	KC-15	10–12	−41.9	7.2	69.52	0.66	2920	76	2096	758	228	Organic carbon
3. YAUT-015119	KC-15	20–22	−48.4	5.3	64.21	0.5	3559	63	2735	1395	254	Organic carbon
4. YAUT-015120	KC-15	30–32	−53.9	6.3	58.04	0.5	4370	70	3546	2350	321	Organic carbon
5. YAUT-015121	KC-15	50–52	−47.4	6.1	49.31	0.42	5680	69	4856	3956	337	Organic carbon
6. YAUT-015122	KC-15	70–72	−35.7	5.7	40.6	0.34	7241	67	6417	5902	279	Organic carbon
7. YAUT-015123	KC-15	110–112	−57	5.5	20.55	0.19	12,710	73	11,886	12,488	319	Organic carbon
8. UCIAMS-78794^a	KC-15	110–112			28.8154	0.0041	10,000	120		9844	364	Foraminifera
9. UCIAMS-78793^a	KC-15	130–130	1.0	0.1	26.4058	0.0010	10,695	35		10,725	292	Shell fragments

AIO: acid insoluble organic.

The reservoir age for the Southern Ocean is estimated as 1300 ± 100, following local contamination offset (LCO) is 824 years based on the surface AIO age in this study. LCO-corrected ages are reported as the radiocarbon and then calibrated using CALIB program v. 7.1 (Stuiver et al., 1998) with Marine13 curve (Reimer et al., 2013). Ages are reported.

Carbonate-derived ages have been recalibrated from a previous study (Kirshner et al., 2012).

These radiocarbon ages were then calibrated by calculating the local contamination offset (LCO, following Hillenbrand et al., 2010) and subtracting it from the radiocarbon ages. The LCO-corrected ages were entered in Calib 7.1 software with the Marine13 curve, which has a built-in global reservoir correction of 400 years (Reimer et al., 2013; Stuiver et al., 1998, 2005), and a local correction of 900 ± 100 years was applied (totaling 1300 ± 100 yr BP reservoir age for the Southern Ocean following Berkman and Forman, 1996).

Sedimentological proxies

X-radiographs were taken of the archive core halves at the Antarctic Research Facility in Tallahassee, Florida, to identify sedimentary structures, determine pebble abundance, and search for additional carbonate macrofossils. Pebble abundance was determined by counting in situ pebbles (diameter >2 mm) over 5-cm intervals on x-radiographs.

Each core was sampled onboard at 10-cm intervals for grain size, geochemistry, and microfaunal abundance and assemblage prior to archiving. Grain size samples were analyzed at Rice University using a Malvern Mastersizer 2000 Laser Particle Size Analyzer (McCave et al., 1986).

Geochemical proxies

Total organic carbon (TOC), nitrogen, and hydrogen were measured using the Rice Costech Elemental Analyzer after removing carbonate following the vaporous HCl decarbonation method (Komada et al., 2008). Samples were dried, crushed, and weighed into silver capsules following standard procedures. In total, 20 µL of deionized (DI) water were added to each sample, which was then placed in a desiccator adjacent to open beakers of 20–30% HCl for 17 h, and then dried overnight prior to analysis. Calibrations with proline standards yielded errors <0.01% for C and N.

Micropaleontological proxies

Foraminiferal samples were wet-sieved at 63 µm. All specimens were picked and arranged by taxa on micropaleontological slides. Taxonomic classification is that used in Majewski (2013). The collection is hosted at the Institute of Paleobiology, Warsaw, Poland. Total foraminiferal abundances are calculated by dividing the number of all foraminiferal tests by grams of dried sediment (t/gds). Faunal diversities are expressed by the Shannon diversity index,

H = - \sum \frac{n_{i}}{n} \ln (\frac{n_{i}}{n})

where n_i is the number of individuals of species i.

Additionally, samples of 0.05 g mass were analyzed for diatom abundance and prepared following a settling method (Scherer, 1994). Diatoms were counted in each prepared slide at 400× magnification using a transmitted light microscope. Diatom abundance was too low for most of the core sections to conduct a robust statistical analysis on assemblages; instead, qualitative observations of assemblages were noted, and pennate versus centric diatoms were counted in 22-mm transects in each microscope slide. The geochemical, sedimentological, and paleontological data were integrated to interpret depositional environment.

Results

Geophysical results

PIB has a broad, fore-deepened shelf (Figure 2) that is characterized by basement-floored inner shelf and a sediment-floored outer shelf, which is composed of several stratigraphic packages that onlap basement and thicken seaward (e.g. Gohl et al., 2013; Lowe and Anderson, 2002; Wellner et al., 2001). Ferrero Bay is crystalline basement floored (Figure 2). Sub-bottom profiler data show very thin, draping sediment in small pockets within the glacially sculpted basement topography (Figure 4).

The deep, inner basin is characterized by linear glacial features that orient E-W. Crag-and-tail features are prominent in the inner basin and taper to the west. Outer Ferrero Bay is characterized by linear furrows, megaflutes, and cross-cutting lineations that reflect convergence of ice flowing from the bay with ice flowing out of PIB (Figure 2). The southern side of the swath area includes prominent bedrock highs (up to 600 relief) that include networks of linear scours and mini-basins at the crest of the highs.

Radiocarbon results

Shell fragments from the base of KC-15 yielded an age for deglaciation of 10,695 ± 35 radiocarbon yr BP (where present day is AD 1950), which was published by Kirshner et al. (2012) and recalibrated to 10,725 ± 292 cal. yr BP (with 2σ error) in Calib 7.1 (Table 1; Berkman and Forman, 1996). Another age was extracted from planktonic foraminifera in lower units (110 cmbsf), yielding an age of 10,000 ± 120 radiocarbon yr BP, which was calibrated to 9844 ± 364 cal. yr BP (Kirshner et al., 2012). The carbonate-derived ages at the base of the core provide a more robust chronology than is typical for Antarctic marine sediments, where terrigenous flux in proximal facies often creates an exponential increase in radiocarbon age, a phenomenon known as the ‘hockey stick effect’ that results from reworking of old carbon (e.g. Andrews et al., 1999; Yokoyama et al., 2016). Indeed, there is significant increase in the bulk organic fraction radiocarbon age at 110 cmbsf, and it was not included in the age model (Table 1).

The AIO fraction ages were also calibrated using Calib 7.1 (Reimer et al., 2013; Stuiver et al., 1998) with a surface age assumption of −60 cal. yr BP (or AD 2010 when the core was collected). The LCO for the region is 824 years, calculated from the surface AIO fraction age of 2124 ± 57 ¹⁴C yr BP and a Southern Ocean reservoir age of 1300 ± 100 yr BP (Berkman and Forman, 1996; Hillenbrand et al., 2010). The two carbonate-derived radiocarbon ages and six overlying AIO fraction ages from KC-15 provide a linear age model best fit by the line y = 0.0115x + 2.3404 (where y is the depth and x is the age) with remarkable linear regression of R² = 0.99613 (Table 1; Figure 5).

Figure 5.

Age model for Ferrero Bay, derived from Kasten Core 15, which is best fit by the line y = 0.0115x + 2.3404, where y is the depth and x is the age (cal. yr BP). Ages marked with an asterisk are carbonate-derived ages first reported by Kirshner et al., 2012, and are recalibrated with Calib 7.1 for consistency with new AIO fraction ages analyzed in this study. All other ages are derived from the AIO fraction (Table 1). The AIO fraction age demarked by gray text is anomalously old and is not included in the age model.

Kasten core KC-15

In total, four units of different sedimentological, geochemical, and paleontological properties were identified in KC-15. The base of KC-15, unit 1 (137–127 cmbsf), is characterized by a dark brown, clayey sand with abundant pebbles. Magnetic susceptibility is high in unit 1 (Figure 6). The core bottomed out in this unit likely because of its relatively stiff, cohesive properties. Based on the sub-bottom profile of this area, the 137-cm core (including catcher material) appears to capture the entire post-glacial sediment record but not the subglacial till (Figure 4).

Figure 6.

Multi-proxy analysis of the three OSO0910 Kasten cores from Ferrero Bay. Unit distinctions are based on changes in magnetic susceptibility, sedimentology, TOC and TN, and diatom and foraminiferal abundance and assemblages.

The contact of unit 1 and unit 2 is characterized by a transition to light grayish brown sandy mud (Figure 6). Magnetic susceptibility is significantly lower in unit 2 (127–104 cmbsf) but peaks with increasing sand content at 110 cmbsf. TOC and total nitrogen (TN) are near zero in this unit (0–0.2%). Diatoms are virtually absent in unit 2, while foraminifera are abundant. The planktonic species Neogloboquadrina pachyderma sinistral dominate the foraminiferal assemblage of unit 2, and calcareous foraminifera comprise nearly 100% of the assemblage.

Unit 3 (104–30 cmbsf) has a sharp base with color change to greenish bluish gray and significant decrease in grain size (Figure 6). Unit 3 is a silty mud with low magnetic susceptibility, sparse pebbles, and higher TN and TOC than unit 2. Diatoms are present but remain in low abundance throughout unit 3. Foraminifera are sparse but peak in abundance and diversity at 50 and 70 cmbsf. At these intervals, benthic foraminifera dominate, and the CDW-affinity species Bulimina aculeata is present (Figure 6).

Productivity indicators TOC, TN, and diatom abundance are high in unit 4 (30 cmbsf to core top), which is sedimentologically very similar to unit 3 (Figure 6). TOC increases above 0.5–0.75% at the top of the core. Both diatoms and foraminifera abundances reach their highest values in the core-top sediment. Pennate to centric diatom ratios are higher in the top of the section. The upper 2 cm of core sediment has a noticeable color change to orange brown, resembling iron oxidation.

Kasten core KC-16

Collected from a bedrock high, KC-16 is significantly shorter in length than KC-15 and KC-17, which were acquired within small depocenters of Ferrero Bay (Figures 2 and 4). KC-16 is much coarser overall and consists of pebble-rich and sand-rich mud. Magnetic susceptibility is an order of magnitude higher in KC-16 than in KC-15 and KC-17, clearly due to its high pebble concentration (up to 20% volume; Figure 6). High TOC and diatom abundance in this short core, in addition to its striking similarity in color to the top of KC-15, indicate that it may correlate temporally with unit 4 of KC-15 and KC-17 (Figures 6 and 7). The high pebble concentration, however, suggests that KC-16 captures a different subfacies that may be related to its location on a bathymetric high (Figure 2). Live benthic organisms in the upper 2 cm indicate that the seafloor surface was preserved.

Figure 7.

Correlation of sedimentary facies in Kasten cores from Ferrero Bay. Variability in sediment character within facies suggests some lateral variability, likely related to proximity to the grounding line and water depth.

Kasten core KC-17

KC-17 is characterized by a similar facies progression as in KC-15, with some differences in microfaunal assemblage (Figure 6 and 7). Unit 1 (147–130 cmbsf) is a dark greenish pebble-rich clayey sand with high magnetic susceptibility (Figure 6). Unit 1 is characterized by the highest sand content measured in Ferrero Bay (~40%) and virtually absent TOC and TN. Diatoms identified in unit 1 include reworked Denticulopsis spp., as has been described by Kellogg and Kellogg (1987a) in PIB. Foraminifera are present in the top of this unit. They are dominated by planktic species N. pachyderma sinistral.

A decrease in magnetic susceptibility and low foraminifera abundance characterize KC-17 unit 2, a greenish sandy clay that contains layers of light greenish silty clay (130–97 cmbsf; Figure 6). Magnetic susceptibility peaks at ~100 cmbsf. Foraminiferal assemblages not only include N. pachyderma, similar to unit 2 of KC-15, but also include the common agglutinated species Milliamina arenacea. Other parallels to KC-15 unit 2 include virtually absent diatoms, TOC, and TN, while sand concentration is high (~35%; Figure 6).

KC-17 unit 3 (97–30 cmbsf) is a greenish silty clay with few pebbles (Figure 6). Grain size decreases to a silt-dominated population from unit 2 to unit 3, much like the facies transition in KC-15. Magnetic susceptibility decreases and stabilizes to low values of ~70 cmbsf. TOC, TN, and diatoms are low and steadily increase in the upper section. Foraminifera are sparse, and Bulimina aculeata is not present in this core. However, Milliamina arenacea is present at the base of unit 3 (Figure 6).

Unit 4 (30 cmbsf to core top) of KC-17 is a brownish green silty clay, characterized by fine grain size and increasing diatom, TOC, and TN percentages (Figure 6). This is identical in grain size, geochemistry, and diatom abundance to KC-15 unit 4. Foraminifera abundance and diversity are highest in the core top and dominated by M. arenacea. Neogloboquadrina pachyderma is present only in the core top. Pennate to centric diatom are higher in the upper section. All productivity proxies are highest in the core top (Figure 6).

Discussion

Paleo-drainage of the West Antarctic Ice Sheet

The bathymetric survey of Ferrero Bay reveals a paleo-trough includes bedrock crag-and-tail features, megaflutes, and deep furrows, all products of ice-sculptured bedrock (Figure 2). The age of these features is uncertain, but given the very thin and patchy sediment cover in the bay, Ferrero Bay likely channeled grounded ice into PIB during the Last Glacial Maximum, contributing to a major ice stream that occupied the Cosgrove-Abbott Trough, which crosses the continental shelf north of the Abbot Ice Shelf (Figure 2; Gohl, 2012; Jakobsson et al., 2012; Klages et al., 2015; Smith et al., 2014). Paleo-ice flow was directed from east to west, with significant curvature in the deep, inner basin. Overlapping furrows in the outer bay potentially record convergence of ice streaming from Ferrero Trough with ice streaming from the south (Figure 2). Linear features exist up to 1300 m and signify grounding of the Wechsler Adult Intelligence Scale (WAIS) to the deep seafloor (e.g. Gohl, 2012; Larter et al., 2014). Potential meltwater storage basins are observed in the middle basin, where a major (>500 m relief) basement high has a connected network of linear scours that may have been generated by subglacial meltwater drainage when ice was grounded in the bay (Figure 2). The basement highs likely served as important pinning points, where the ice could ground and remain relatively stable during deglaciation and sea-level rise.

Unit 1: Glacial recession from Ferrero Bay

Despite minor lateral variation, unit 1 can be correlated from the proximal core (KC-15) to the distal core (KC-17) with remarkable precision (Figures 6 and 7). The magnetic susceptibility is high at the base of both cores and is followed by a trough, which is tightly correlated along with coarse grain size and color change. KC-15 and KC-17 bottomed out in relatively stiff sandy clay, indicating the basal unit 1 is a proximal glacimarine facies and records deglaciation of the fjord. The calibrated radiocarbon age of 10,725 ± 292 cal. yr BP (Table 1; Figure 5; recalibrated from Kirshner et al., 2012) and a projection of the age model to the base of KC-15 (137 cmbsf) provide the minimum age of 11,033 cal. yr BP for deglaciation of proximal Ferrero Bay. Subglacial till was not sampled, however, and the radiocarbon ages of unit 1 represent an absolute minimum age for deglaciation from inner Ferrero Bay. Based on ages from KC-15 and a deglaciation age of 13.6 cal. kyr BP at the mouth of Ferrero Bay (Smith et al., 2014), Ferrero Bay deglaciated approximately 2 kyr after the middle continental shelf, but both areas were covered by a floating ice shelf for some time after grounding line retreat and lift off (Kirshner et al., 2012). Greater thickness of the expanded WAIS in proximal coastal settings likely allowed ice to remain grounded in the trough of Ferrero Bay until ~11.0 cal. kyr BP, after the ice sheet had receded from the mid-shelf (Figure 8).

Figure 8.

Glacial reconstruction of Ferrero Bay and mid continental shelf area of the eastern Amundsen Sea, showing (a) Ferrero Bay completely filled with grounded ice until 11 cal. kyr BP, following deglaciation of the outer ASE after the Last Glacial Maximum. (b) The grounding line receded rapidly to near its current position, while the Cosgrove Ice Shelf persisted during the mid-Holocene, when CDW may have been present (indicated by the presence of B. aculeata). (c) The Cosgrove Ice Shelf receded to its current position by 2.3 cal. kyr BP. CDW is observed on the shelf today (Jacobs et al., 2011, 2013; Walker, et al., 2007) and clearly makes its way into Ferrero Bay (Figure 3). Core locations are projected onto the profile; KC-16 was collected at a local high of ~700 m, which is to the south of this profile (Figure 2). Profile was built from OSO0910 multibeam swath bathymetry. Dashed lines are inferred subglacial topography, because of limited radar coverage and resolution. Ice thicknesses are best estimates based on inferred bathymetry.

Furthermore, Ferrero Bay is characterized by a low Holocene accumulation rate of 0.011 mm/yr, an order of magnitude less than northern AP fjords (e.g. Allen et al., 2010; Majewski et al., 2012; Michalchuk et al., 2009; Milliken et al., 2009; Totten et al., 2015). This may be due to latitudinal decrease in erosion rate associated with increasing polar conditions and the freezing of the glaciers to their beds (Fernandez et al., 2016), but basin-wide volume calculations are needed to test this idea.

Unit 2: Advection beneath the Cosgrove Ice Shelf

As observed in both KC-15 and KC-17, unit 2 is interpreted as a proximal glacimarine facies with coarse, poorly sorted grain size, and low productivity (Figures 6 and 7). Unit 2 represents a second phase of grounding line recession in Ferrero Bay from 10.6 to 9.6 cal. kyr BP, resulting in a more distal sub-ice shelf setting at the core sites.

Presence of planktonic foraminifera in unit 2 (Figure 6) may indicate ocean circulation into the innermost bay beneath an ice shelf (Figure 8). Abundant, well-preserved planktonic foraminifera were attributed to landward advection beneath the Amery Ice Shelf (Hemer et al., 2007); alternatively, Majewski et al. (2012) demonstrated that an abundance of planktonic foraminifera in fjord settings can result from stronger inflow of open-ocean water when turbidity is low. Here, we rule out the second case because of low productivity during this interval. Abundance of planktonic foraminifera N. pachyderma in the sub-ice shelf facies associated with a deglaciation event from 12.3 to 10.6 cal. kyr BP was argued by Kirshner et al. (2012) to be evidence for offshore currents advecting under an extensive Amundsen Sea ice shelf. Moreover, the abundance of N. pachyderma in both transitional sediments in KC-15 and KC-17 in Ferrero Bay as well as KC-19 (Kirshner et al., 2012) imply that ocean currents were accessing the grounding line at this time.

Unit 3: Stable Cosgrove Ice Shelf during the mid-Holocene

Unit 3 represents a significant shift to a distal sub-ice shelf setting during the mid-Holocene (Figures 6 –8). Unit 3, observed in both KC-15 and KC-17, is composed of silty mud and lacks pebbles and sand that might indicate ice rafting or proximity to the grounding line. Low TOC, TN, and diatom abundance suggest long-term stability of the Cosgrove Ice Shelf from 9.6 to 2.3 cal. kyr BP. A progressive increase in these productivity indicators indicates gradual recession of the ice margin. Sparse but present diatoms were likely transported under the ice shelf by advection, as they are light and easily suspended (e.g. Anderson et al., 1991; Domack et al., 2005; Evans and Pudsey, 2002; Hemer et al., 2007).

While N. pachyderma is not present in unit 3, the benthic foraminifer B. aculeata is present at 70 and 50 cmbsf in KC-15 (Figure 6). It may suggest advection of warm deep water onto the shelf during the mid-Holocene, as this foraminifer is strongly associated with the presence of CDW (Ishman and Domack 1994; Majewski et al., 2016). Continued warm deep incursion would have been channeled underneath the Cosgrove Ice Shelf during deposition of unit 3 (Figure 8), much like what is observed below the Pine Island Glacier’s ice shelf today (Jenkins et al., 2010, 2012). Fine grain size, low productivity, and lack of biogenic particles indicate sub-ice shelf conditions persisted throughout the early- and mid-Holocene.

Unit 4: Late-Holocene recession of the Cosgrove Ice Shelf

Unit 4 marks recession of the Cosgrove Ice Shelf and the opening of Ferrero Bay to its present configuration by 2.3 cal. kyr BP (Figure 8). High TOC, TN, and diatom abundance show increased productivity in absence of the Cosgrove Ice Shelf (Figure 6). Moreover, pennate diatoms are abundant near the core top, indicating sea ice was common after the bay opened. This is consistent with historic observations of heavy seasonal sea-ice cover (e.g. Jacobs et al., 2012; Parkinson and Cavalieri, 2012). Moreover, subfossil/modern diatom assemblages from the tops of KC-15 and KC-16 are described by Świło et al. (2016) and have lower abundance and higher diversity than AP fjords, characterized by a Fragilariopsis curta assemblage that is associated with sea-ice persistence and low primary production.

The uppermost 2 cm in KC-15 and KC-17 are high in TOC (Figure 6) and hosted benthic organisms and modern biological activity when they were collected. KC-15 also shows evidence of iron oxidation and oxygenated bottom water that promotes benthic fauna. This indicates that the seafloor surface is well preserved in these cores.

Ice Shelf history

Ferrero Bay contains a complete Holocene sedimentary record with no evidence of a hiatus and provides a minimum age of 11 cal. kyr BP for glacial recession of the eastern ASE coast. Cosmogenic exposure ages of the Hudson Mountains and the Kohler Range (northeast of Pine Island Glacier; Figure 2) record ice surface elevation decrease in the early-Holocene (Johnson et al., 2014; Lindow et al., 2014). A second grounding line recession from 10.6 to 9.6 cal. kyr BP in Ferrero Bay may broadly relate to ice surface lowering to near modern levels by 8.0 cal. kyr BP as recorded by ¹⁰Be exposure ages in the Hudson Mountains (Johnson et al., 2014).

The Cosgrove Ice Shelf existed throughout the early- to mid-Holocene and remained stable in outer Ferrero Bay until ~2.3 cal. kyr BP, when it receded to its current configuration (Figures 6 –8). Cosmogenic ages of boulders near sea level on an island seaward of the Canisteo Peninsula on the southern boundary of Ferrero Bay record exposure at 2.2 ± 0.2 kyr (Figure 2; Johnson et al., 2008). While these results alone do not indicate whether the exposure was due to emergence and isostatic rebound or due to local ice recession on the island, the minimum age of ice shelf loss in inner Ferrero Bay supports the scenario of ice recession. This would imply that the Cosgrove Ice Shelf completely filled Ferrero Bay, and ice remained stable at sites seaward of Canisteo Peninsula, likely providing further buttress to an advanced Cosgrove Ice Shelf, until the late-Holocene.

The Prince Gustav Ice Shelf of the northern AP destabilized and collapsed by 5.0 cal. kyr BP (Figures 1 and 9; Pudsey and Evans, 2001) and the Larsen A Ice Shelf collapsed by 3.8 cal. kyr BP (Figures 1 and 9; Brachfield et al., 2003). Both reformed by 2.0 and 1.8 cal. kyr BP, respectively, during climatic cooling observed in the James Ross Island and EPICA Dome C ice cores (Figures 1 and 9; Brachfield et al., 2003; Monnin et al., 2001; Mulvaney et al., 2012; Pudsey and Evans, 2001). By contrast, the Larsen B Ice Shelf was present but progressively thinning throughout the Holocene and only recently collapsed in 2002 (Figures 1 and 9; Domack et al., 2005). The George VI Ice Shelf collapsed by the early-Holocene ~9 to 7.8 cal. kyr BP, following a period of rapid climate warming and impinging warm CDW onto the inner shelf (Figures 1 and 9; Allen et al., 2010; Bentley et al., 2009; Smith et al., 2007). Thus, while AP ice shelf collapse was not strictly synchronous, ice shelves generally receded during the warm intervals of the early- or mid-Holocene and reformed during the late-Holocene Neoglacial.

Figure 9.

Ice shelf history and paleo-temperature proxies from EPICA Dome C and James Ross Island ice cores (Monnin et al., 2001; Mulvaney et al., 2012; modified from Hodgson, 2011; PNAS). White bars represent grounded ice at the seabed, blue bars represent times of an extended ice shelf in the embayments, and red bars represent periods of open water. The Cosgrove Ice Shelf adds a polar end member to the dynamic history of ice shelves in the Antarctic Peninsula (Brachfield et al., 2003; Domack et al., 2005; Pudsey and Evans, 2001; Smith et al., 2007) and the Arctic (Ward Hunt Ice Shelf; Antoniades et al., 2011). The Cosgrove Ice Shelf does not appear to be sensitive to atmospheric temperature changes and retreated when most Antarctic Peninsula ice shelves were either stable or reforming.

Interestingly, the largest Arctic ice shelf, the Ward Hunt Ice Shelf, was absent throughout the early- and mid-Holocene warm periods and then formed 4.0 cal. kyr BP and collapsed 1.4 cal. kyr BP (Figure 9; Antoniades et al., 2011). The ice shelf reformed 800 years ago during the ‘Little Ice Age’ (Antoniades et al., 2011), broadly coinciding with a small advance of the Müller Ice Shelf in the AP (Figure 1; Shevenell et al., 1996; Taylor et al., 2001).

The Cosgrove Ice Shelf retreated when most AP ice shelves were either stable or reforming during late-Holocene climatic cooling (Figure 9). The out-of-phase behavior of the Cosgrove Ice Shelf indicates that it was either insensitive to atmospheric temperature changes or that climatic shifts were less pronounced in the more polar climate of the Amundsen Sea coast than in the AP. In either case, ocean circulation, especially impingement of warm CDW, appears to have been the dominant control on glacial stability of the Cosgrove glacial system throughout the Holocene. Similarly, the George VI Ice Shelf was largely out of phase with northern AP ice shelves because of the influence of CDW (Figures 1 and 9; Smith et al., 2007). The asynchronous global pattern of ice shelf retreat and regrowth during the Holocene contrasts with the synchronous nature of historic ice shelf loss (Figure 9), in which 28,000 km² of AP ice shelf area has been irreparably lost to the sea since the 1980s (Cook and Vaughan, 2010; Hodgson et al., 2006, 2011). Recent ice shelf loss in the western AP may be largely driven by the influence of CDW as well (Cook et al., 2016).

Conclusion

To date, Ferrero Bay provides the southernmost and high-resolution Holocene record of an outlet glacier-fed ice shelf in Antarctica. Following liftoff and grounding line retreat ~11 cal. kyr BP, the Cosgrove glacial system sustained an ice shelf that covered Ferrero Bay throughout the early- and mid-Holocene warm periods, when many AP ice shelves collapsed. The presence of foraminifera that are associated with CDW, which upwells into Ferrero Bay today, demonstrates that persistent upwelling of relatively warm water during the Holocene caused under-melting of the Cosgrove Ice Shelf and likely forced its eventual collapse. The Cosgrove Ice Shelf did not recede to its current position until 2.3 cal. kyr BP, when climate was cooling. This coincides with deglaciation of the southern perimeter of Ferrero Bay from a near-sea-level cosmogenic exposure age of 2.2 ± 0.2 kyr. Thus, the Cosgrove Ice Shelf was out of phase with other Antarctic ice shelves and did not respond to climate variability but rather to oceanographic forcing during the Holocene. Amundsen Sea and western AP ice shelves will likely experience further recession under the influence of CDW, impacting velocity and stability of outlet glaciers in the ASE, which comprise the third-largest drainage system of the West Antarctic Ice Sheet today.

Footnotes

Acknowledgements

We thank the captain and crew of the IB Oden, the onboard scientific party, and the Antarctic Research Facility at Florida State University, and Catherine Ross for assistance with sample collection and preparation. We thank the reviewers for their constructive comments that helped improve the paper.

Funding

This research was funded by the US National Science Foundation Office of Polar Programs NSF/ARRA ANT-0837925 to J.B. Anderson. The OSO0910 expedition was a collaboration between the National Science Foundation, the Swedish Polar Research Secretariat, and Swedish Research Council.

References

Allen

Oakes-Fretwell

Anderson

, et al. (2010) A record of Holocene glacial and oceanographic variability in Neny Fjord, Antarctic Peninsula. The Holocene 20: 551–564.

Anderson

Jakobsson

Party

(2010) Oden Southern Ocean 0910 OSO0910 Cruise Report. Stockholm: University of Stockholm.

Anderson

Kennedy

Smith

, et al. (eds) (1991) Sedimentary Facies Associated with Antarctica’s Floating Ice Masses (Glacial marine sedimentation; paleoclimatic significance, special paper 261). Boulder, CO: The Geological Society of America.

Andrews

Domack

Cunningham

, et al. (1999) Problems and possible solutions concerning radiocarbon dating of surface marine sediments, Ross Sea, Antarctica. Quaternary Research 52: 206–216.

Antoniades

Francus

Pienitz

, et al. (2011) Holocene dynamics of the Arctic’s largest ice shelf. Proceedings of the National Academy of Sciences of the United States of America 108: 18899–18904.

Arrigo

Brown

Mills

(2014) Sea ice algal biomass and physiology in the Amundsen Sea, Antarctica. Elementa: Science of the Anthropocene 2(1): 000028.

Arrigo

Van Dijken

Bushinsky

(2008) Primary production in the Southern Ocean, 1997–2006. Journal of Geophysical Research 113: C08004. DOI: 10.1029/2007JC004551.

Bentley

Hodgson

Smith

, et al. (2009) Mechanisms of Holocene palaeoenvironmental change in the Antarctic Peninsula region. The Holocene 19(1): 51–69.

Berkman

Forman

(1996) Calcareous marine species in the Southern Ocean. Geophysical Research Letters 23(4): 363–366.

10.

Brachfield

Domack

Kissel

, et al. (2003) Holocene history of the Larsen-A Ice Shelf constrained by geomagnetic paleointensity dating. Geology 31: 749–752.

11.

Cochran

Tinto

Bell

(2015) Abbot Ice Shelf, structure of the Amundsen Sea continental margin and the southern boundary of the Bellingshausen Plate seaward of West Antarctica. Geochemistry, Geophysics, Geosystems 16: 1421–1438.

12.

Cook

Vaughan

(2010) Overview of areal changes of the ice shelves on the Antarctic Peninsula over the past 50 years. The Cryosphere 4: 77–98.

13.

Cook

Holland

Meredith

, et al. (2016) Ocean forcing of glacier retreat in the western Antarctic Peninsula. Science 353(6296): 283–286.

14.

Dalziel

Elliot

(1982) West Antarctica: Problem child of Gondwanaland. Tectonics 1(1): 3–19.

15.

De Angelis

Skvarca

(2003) Glacier surge after ice shelf collapse. Science 299: 1560–1562.

16.

De Rydt

Gudmundsson

(2016) Coupled ice shelf-ocean modeling and complex grounding line retreat from a seabed ridge. Journal of Geophysical Research: Earth Surface 121(5): 865–880.

17.

Dinniman

Klinck

Smith

Jr (2003) Cross-shelf exchange in a model of the Ross Sea circulation and biogeochemistry. Deep Sea Research Part II: Topical Studies in Oceanography 50: 3103–3120.

18.

Doake

CSM

Vaughan

(1991) Rapid disintegration of the Wordie Ice Shelf in response to atmospheric warming. Nature 350: 328–330.

19.

Domack

Duran

Leventer

, et al. (2005) Stability of the Larsen B ice shelf on the Antarctic Peninsula during the Holocene epoch. Nature 436: 681–685.

20.

Ehrmann

Hillenbrand

Smith

, et al. (2011) Provenance changes between recent and glacial-time sediments in the Amundsen Sea embayment, West Antarctica: Clay mineral assemblage evidence. Antarctic Science 23(5): 471–486.

21.

Evans

Pudsey

(2002) Sedimentation associated with Antarctic Peninsula ice shelves: Implications for palaeoenvironmental reconstructions of glacimarine sediments. Journal of the Geological Society 159: 233–237.

22.

Fernandez

Anderson

Wellner

, et al. (2016) Latitudinal variation in glacial erosion rates from Patagonia and the Antarctic Peninsula (46°-65°S). GSA Bulletin 128(5–6): 1000.

23.

Fragoso

Smith

(2012) Influence of hydrography on phytoplankton distribution in the Amundsen and Ross Seas, Antarctica. Journal of Marine Systems 89(1): 19–29.

24.

Gerringa

Alderkamp

Laan

, et al. (2012) Iron from melting glaciers fuels the phytoplankton blooms in Amundsen Sea (Southern Ocean): Iron biogeochemistry. Deep Sea Research Part II: Topical Studies in Oceanography 71: 16–31.

25.

Gohl

(2012) Basement control on past ice sheet dynamics in the Amundsen Sea Embayment. West Antarctica. Palaeogeography, Palaeoclimatology, Palaeoecology 335: 35–41.

26.

Gohl

Uenzelmann-Neben

Larter

, et al. (2013) Seismic stratigraphic record of the Amundsen Sea Embayment shelf from pre-glacial to recent times: Evidence for a dynamic West Antarctic ice sheet. Marine Geology 344: 115–131.

27.

Hemer

Post

O’Brien

, et al. (2007) Sedimentological signatures of the sub-Amery Ice Shelf circulation. Antarctic Science 19: 497–506.

28.

Hillenbrand

Smith

Kuhn

, et al. (2010) Age assignment of a diatomaceous ooze deposited in the western Amundsen Sea Embayment after the Last Glacial Maximum. Journal of Quaternary Science 25(3): 280–295.

29.

Hodgson

(2011) First synchronous retreat of ice shelves marks a new phase of polar deglaciation. Proceedings of the National Academy of Sciences 108(47): 18859–18860.

30.

Hodgson

Bentley

Roberts

, et al. (2006) Examining Holocene stability of Antarctic Peninsula ice shelves. Eos, Transactions American Geophysical Union 87(31): 305–308.

31.

Hofmann

Klinck

Lascara

, et al. (1996) Water mass distribution and circulation west of the Antarctic Peninsula and including Bransfield Strait. In: Ross

Hofmann

Quetin

(eds) Foundations for Ecological Research West of the Antarctic Peninsula (Vol. 70, Antarctic Research Series). Washington, DC: American Geophysical Union (AGU), pp. 81–104.

32.

Hughes

(1981) The weak underbelly of the West Antarctic ice sheet. Journal of Glaciology 27(97): 518–525.

33.

Ishman

Domack

(1994) Oceanographic controls on benthic foraminifers from the Bellingshausen margin of the Antarctic Peninsula. Marine Micropaleontology 24(2): 119–155.

34.

Jacobs

Giulivi

Dutrieux

, et al. (2013) Getz Ice Shelf melting response to changes in ocean forcing. Journal of Geophysical Research: Oceans 118: 4152–4168.

35.

Jacobs

Jenkins

Hellmer

, et al. (2012) The Amundsen Sea and the Antarctic ice sheet. Oceanography 25: 154–163.

36.

Jacobs

Hellmer

Jenkins

(1996) Antarctic ice sheet melting in the Southeast Pacific. Geophysical Research Letters 23(9): 957–960.

37.

Jacobs

Jenkins

Giulivi

, et al. (2011) Stronger ocean circulation and increased melting under Pine Island Glacier ice shelf. Nature Geoscience 4(8): 519–523.

38.

Jakobsson

Anderson

Nitsche

, et al. (2012) Ice sheet dynamics inferred from glacial morphology of the Central Pine Island Bay Trough, West Antarctica. Quaternary Science Reviews 38: 1–10.

39.

Jenkins

Dutrieux

Jacobs

, et al. (2012) Autonomous underwater vehicle exploration of the ocean cavity beneath an Antarctic ice shelf. Oceanography 25(3): 202–203.

40.

Jenkins

Dutrieux

Jacobs

, et al. (2010) Observations beneath Pine Island Glacier in West Antarctica and implications for its retreat. Nature Geoscience 3: 468–472.

41.

Johnson

Bentley

Gohl

(2008) First exposure ages from the Amundsen Sea embayment, west Antarctica: The late quaternary context for recent thinning of Pine Island, Smith, and Pope Glaciers. Geology 36(3): 223–226.

42.

Johnson

Bentley

Smith

, et al. (2014) Rapid thinning of Pine Island Glacier in the early Holocene. Science 343(6174): 999–1001.

43.

Kellogg

(1987a) Microfossil distributions in modern Amundsen Sea sediments. Marine Micropaleontology 12: 203–222.

44.

Kellogg

(1987b) Recent glacial history and rapid ice stream retreat in the Amundsen sea. Journal of Geophysical Research 92: 8859–8864.

45.

Kirshner

Anderson

Jakobsson

, et al. (2012) Post-LGM deglaciation in Pine Island Bay, West Antarctica. Quaternary Science Reviews 38(30): 11–26.

46.

Klages

Kuhn

Graham

, et al. (2015) Palaeo-ice stream pathways and retreat style in the easternmost Amundsen Sea Embayment, West Antarctica, revealed by combined multibeam bathymetric and seismic data. Geomorphology 245: 207–222.

47.

Komada

Anderson

Dorfmeier

(2008) Carbonate removal from coastal sediments for the determination of organic carbon and its isotopic signatures, δ13C and Δ14C: Comparison of fumigation and direct acidification by hydrochloric acid. Limnology and Oceanography: Methods 6(6): 254–262.

48.

Larter

Anderson

Graham

, et al. (2014) Reconstruction of changes in the Amundsen Sea and Bellingshausen sea sector of the West Antarctic ice sheet since the last glacial maximum. Quaternary Science Reviews 100: 55–86.

49.

Leventer

Dunbar

(1996) Factors influencing the distribution of diatoms and other algae in the Ross Sea. Journal of Geophysical Research: Oceans 101(C8): 18489–18500.

50.

Lindow

Castex

Wittmann

, et al. (2014) Glacial retreat in the Amundsen Sea sector, West Antarctica–First cosmogenic evidence from central Pine Island Bay and the Kohler Range. Quaternary Science Reviews 98: 166–173.

51.

Locarnini

Mishonov

Antonov

, et al. (2010) NOAA Atlas NESDIS 68: World Ocean Atlas 2009. Volume 1: Temperature (ed S Levitus ). Washington, DC: U.S. Government Printing Office, 184 pp.

52.

Lowe

Anderson

(2002) Reconstruction of the West Antarctic Ice Sheet in Pine Island Bay during the last glacial maximum and its subsequent retreat history. Quaternary Science Reviews 21(16–17): 1879–1897.

53.

McCave

Bryant

Cook

, et al. (1986) Evaluation of a laser-diffraction-size analyzer for use with natural sediments. Journal of Sedimentary Petrology 56(4): 561–564.

54.

Majewski

(2013) Benthic foraminifera from Pine Island and Ferrero bays, Amundsen Sea. Polish Polar Research 34(2): 169–200.

55.

Majewski

Wellner

Anderson

(2016) Environmental connotations of benthic foraminiferal assemblages from coastal West Antarctica. Marine Micropaleontology 124, 1–15.

56.

Majewski

Wellner

Szczuciński

, et al. (2012) Holocene oceanographic and glacial changes recorded in Maxwell Bay, West Antarctica. Marine Geology 326–328, 67–79.

57.

Martinson

Stammerjohn

Iannuzzi

, et al. (2008) Western Antarctic Peninsula physical oceanography and spatio-temporal variability. Deep Sea Research Part II: Topical Studies in Oceanography 55, 1964–1987.

58.

Michalchuk

Anderson

Wellner

, et al. (2009) Holocene climate and glacial history of the northeastern Antarctic Peninsula: The marine sedimentary record from a long SHALDRIL core. Quaternary Science Reviews 28, 3049–3065.

59.

Milliken

Anderson

Wellner

, et al. (2009) High resolution climate record from Maxwell Bay, South Shetland Islands. Antarctic. Geological Society of America Bulletin 121: 1711–1725.

60.

Monnin

Indermühle

Dällenbach

, et al. (2001) Atmospheric CO2 concentrations over the last glacial termination. Science 291, 112–114.

61.

Morris

Vaughan

(2003) Spatial and temporal variation of surface temperature on the Antarctic Peninsula and the limit of viability of ice shelves. Antarctic Research Series 79: 61–68.

62.

Mulvaney

Abram

Hindmarsh

RCA

, et al. (2012) Recent Antarctic peninsula warming relative to Holocene climate and ice-shelf history. Nature 489(7414): 141–144.

63.

Parkinson

Cavalieri

(2012) Antarctic sea ice variability and trends, 1979–2010. The Cryosphere 6: 871–880.

64.

Pudsey

Evans

(2001) First survey of Antarctic subice shelf sediments reveals mid-Holocene ice shelf retreat. Geology 29: 787–790.

65.

Reimer

Bard

Bayliss

, et al. (2013) IntCal13 and MARINE13 radiocarbon age calibration curves 0–50000 years calBP. Radiocarbon 55(4): 1869–1887.

66.

Rignot

Jacobs

(2002) Rapid bottom melting widespread near Antarctic ice sheet grounding lines. Science 296(5575): 2020–2023.

67.

Scambos

Hulbe

Fahnestock

(2003) Climate-induced ice shelf disintegration in the Antarctic Peninsula. Antarctic Research Series 79: 79–92.

68.

Scambos

Bohlander

Shuman

, et al. (2004) Glacier acceleration and thinning after ice shelf collapse in the Larsen B embayment, Antarctica. Geophysical Research Letters 31: 18.

69.

Scherer

(1994) A new method for the determination of absolute abundance of diatoms and other silt-sized sedimentary particles. Journal of Paleolimnology 12(2): 171–179.

70.

Shevenell

Domack

Kernan

(1996) Record of Holocene palaeoclimate change along the Antarctic Peninsula: Evidence from glacial marine sediments, Lallemand Fjord. Papers and Proceedings of the Royal Society of Tasmania 130(2): 55–64.

71.

Smith

Hofmann

Klinck

, et al (1999) Hydrography and circulation of the west Antarctic Peninsula continental shelf. Deep-Sea Research Part I: Oceanographic Research Papers 46: 925–949.

72.

Smith

Bentley

Hodgson

, et al. (2007) Oceanic and atmospheric forcing of early Holocene ice shelf retreat, George VI Ice Shelf, Antarctica Peninsula. Quaternary Science Reviews 26, 500–516.

73.

Smith

Hillenbrand

Kuhn

, et al. (2014) New constraints on the timing of West Antarctic Ice Sheet retreat in the eastern Amundsen Sea since the Last Glacial Maximum. Global and Planetary Change 122: 224–237.

74.

Smith

Jr Comiso

(2008) Influence of sea ice on primary production in the Southern Ocean: A satellite perspective. Journal of Geophysical Research 113: C05S93. DOI: 10.1029/2007JC004251.

75.

Storey

Pankhurst

Millar

, et al. (1991) A new look at the geology of Thurston Island. In: Thomson

MRA

Crame

Thomson

(eds) Geological Evolution of Antarctica. Cambridge: Cambridge University Press, pp. 399–403.

76.

Stuiver

Reimer

(2005) CALIB 7.1: www program and documentation. Available at: http://calib.qub.ac.uk/calib.

77.

Stuiver

Reimer

Bard

, et al. (1998) INTCAL98 radiocarbon age calibration, 24,000-0 cal BP. Radiocarbon 40: 1041–1083.

78.

Świło

Majewski

Totten

, et al. (2016) Diatom assemblages from coastal settings of West Antarctica. Marine Micropaleontology 125: 95–109.

79.

Taylor

Whitehead

Domack

(2001) Holocene paleoclimate change in the Antarctic Peninsula: Evidence from the diatom, sedimentary and geochemical record. Marine Micropaleontology 41: 25–43.

80.

Thuróczy

Alderkamp

Laan

, et al. (2012) Key role of organic complexation of iron in sustaining phytoplankton blooms in the Pine Island and Amundsen Polynyas (Southern Ocean). Deep Sea Research Part II: Topical Studies in Oceanography 71: 49–60.

81.

Totten

Anderson

Fernandez

, et al. (2015) Holocene climate, ocean, and cryosphere interactions in the eastern Antarctic Peninsula: High-resolution, multi-proxy investigation of Herbert-Croft Fjord, James Ross Island. Quaternary Science Reviews 129: 239–259.

82.

Vaughan

Marshall

Connolley

, et al. (2003) Recent rapid regional climate warming on the Antarctic Peninsula. Climatic Change 60: 243–274.

83.

Wade

LaPrade

(1969) Geology of King Peninsula, Canisteo Peninsula, and Hudson Mountains areas, Ellsworth Land, Antarctica. Antarctic Journal of the United States 4(4): 92.

84.

Walker

Brandon

Jenkins

, et al. (2007) Oceanic heat transport onto the Amundsen Sea shelf through a submarine glacial trough. Geophysical Research Letters 34: L02602.

85.

Wellner

Lowe

Shipp

, et al. (2001) Distribution of glacial geomorphic features on the Antarctic continental shelf and correlation with substrate: Implications for ice behavior. Journal of Glaciology 47: 397–411.

86.

Yokoyama

Anderson

Yamane

, et al. (2016) Widespread collapse of the Ross Ice Shelf during the Late-Holocene. PNAS 113: 2354–2359.

87.

Yokoyama

Koizumi

Matsuzaki

, et al. (2010) Developing ultra small-scale radiocarbon sample measurement at the University of Tokyo. Radiocarbon 52: 310–318.

88.

Yokoyama

Miyairi

Matsuzaki

, et al. (2007) Relation between acid dissolution time in the vacuum test tube and time required for graphitization for AMS target preparation. Nuclear Instruments and Methods in Physics Research Section B 259: 330–334.