Glacial retreat patterns and processes determined from integrated sedimentology and geomorphology records

Accepted Manuscript Glacial retreat patterns and processes determined from integrated sedimentology and geomorphology records

Lindsay O. Prothro, Lauren M. Simkins, Wojciech Majewski, John B. Anderson PII: DOI: Reference:

S0025-3227(17)30277-3 doi:10.1016/j.margeo.2017.09.012 MARGO 5694

To appear in:

Marine Geology

Received date: Revised date: Accepted date:

2 June 2017 8 September 2017 24 September 2017

Please cite this article as: Lindsay O. Prothro, Lauren M. Simkins, Wojciech Majewski, John B. Anderson , Glacial retreat patterns and processes determined from integrated sedimentology and geomorphology records. The address for the corresponding author was captured as affiliation for all authors. Please check if appropriate. Margo(2017), doi:10.1016/j.margeo.2017.09.012

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT Glacial retreat patterns and processes determined from integrated sedimentology and geomorphology records Lindsay O. Prothro1, Lauren M. Simkins1, Wojciech Majewski2, and John B. Anderson1 1

US

Corresponding author: Lindsay O. Prothro ([email protected])

CR

IP

T

Department of Earth, Environmental and Planetary Science, Rice University, 6100 Main Street, Houston, Texas 77005, USA; 2 Institute of Paleobiology, Polish Academy of Sciences, Twarda 51/55, 00-818 Warszawa, Poland

AN

Abstract

M

Although hundreds of cores have been collected on the

ED

Antarctic continental shelf over the past five decades, definitive interpretations of depositional environments associated with

PT

marine-based ice advance and retreat are hindered by similarities

CE

in sediment facies and a lack of geomorphic context. The recent use of an advanced multibeam bathymetry system allows for more

AC

detailed ice sheet reconstructions from glacial landforms that were previously not resolved with older generation systems. Here we present results from a recent cruise to the Ross Sea, Antarctica that focuses on integrating sediment facies analyses into a geomorphic framework to confidently determine depositional environments and sedimentological processes since the Last Glacial Maximum. Grain-size analysis, geotechnical properties, and 1

ACCEPTED MANUSCRIPT micropaleontology from targeted sediment cores are used to develop improved criteria for making the distinction between subglacial diamictons and ice-proximal diamictons, which can be applied to geomorphically blind sediment cores. Our consideration

T

of geomorphic context while interpreting sediment facies has also

IP

allowed us to determine that most debris melted out of the base of

CR

paleo-ice shelves in the Ross Sea only 1.2 km seaward of the grounded ice margin. Additionally, we find that agglutinated

US

foraminifera specimens are characteristic of sluggish marine

AN

currents and a resulting accumulation of siliceous detritus rich in organic matter; however, calcareous specimens dominate in

M

higher-energy current settings and in ice-proximal environments

ED

with lesser amounts of siliceous detritus and associated organic material. Overall, sedimentary successions record contrasting

PT

behaviors of the West Antarctic and East Antarctic ice sheets that

CE

extended onto the continental shelf during the Last Glacial Maximum. Not only is this coupled geomorphic and

AC

sedimentologic approach useful for understanding glacimarine processes and ice sheet retreat patterns, but it is essential for selecting appropriate facies transitions for chronological constraints of past ice sheet behavior. Keywords

2

ACCEPTED MANUSCRIPT Antarctica, Grounding line, Sediments, Glacial landforms, Deglaciation

T

1. Introduction

IP

Ice sheets grounded below sea level are particularly

CR

sensitive to climate change due to their direct contact with the changing ocean, prompting a rigorous effort by the scientific

US

community to understand the behavior of retreating ice sheets and

AN

identify processes that destabilize them. The grounding line, the most seaward position at which the ice is coupled to the bed, is

M

understood to be a potential tipping point of ice sheet stability.

ED

However, much of what is known about grounding lines is theoretical (e.g., Alley et al., 1987; Dupont and Alley, 2005;

PT

Schoof, 2007; Vieli and Nick, 2011) or based on satellite and

CE

airborne remote sensing (e.g., De Angelis and Skvarca, 2003; Rignot et al., 2011; Le Brocq et al., 2013) and ice-penetrating

AC

radar surveys (e.g., Anandakrishnan et al., 2007; MacGregor et al., 2011; Christianson et al., 2016). Direct observations of modern grounding lines and the processes that control their behavior are limited. The geological record provides an opportunity to study the factors that influence ice sheet behavior over longer timescales and at higher spatial resolutions than is possible with observations of contemporary ice sheets. 3

ACCEPTED MANUSCRIPT During the Last Glacial Maximum (LGM), an expanded Antarctic Ice Sheet covered the continental shelf (Fig. 1; Bentley et al., 2014). Subsequent deglaciation left a record of ice sheet retreat in the form of ice-marginal landforms (i.e., recessional

T

moraines and grounding zone wedges) that mark the locations of

IP

paleo-grounding lines and record the nature of grounding-line

CR

retreat across the continental shelf. Recessional moraines form either by deformational push processes of existing sediment at the

US

grounded line (Boulton, 1986; Ottesen and Dowdeswell, 2006) or

AN

through subglacial sediment transport to the grounding line (Ottesen et al., 2005; Todd et al., 2007), and are characteristically

M

symmetric and often roughly linear (e.g., Batchelor and

ED

Dowdeswell, 2016; Simkins et al., 2016). Grounding-zone wedges (Fig. 2a) are asymmetric features that, if large enough to be

PT

resolved in seismic profiles, occasionally display seaward-dipping

CE

foresets, indicating landform progradation and growth in the direction of ice flow through subglacial sediment delivery to the

AC

grounding line (Larter and Vanneste, 1995; Anderson, 1999; Dowdeswell and Fugelli, 2012). The asymmetry of the features and downlapping over existing surfaces are indicative of progradation even when internal foreset reflections are not resolved in seismic data, which is likely due to a lack of variability in lithology and, thus, no difference in acoustic properties between

4

ACCEPTED MANUSCRIPT beds (Fig. 3). Studying the morphology and distribution of icemarginal landforms provides constraints on grounding line sediment flux (e.g., Nygård et al., 2007; Jakobsson et al., 2012; Batchelor and Dowdeswell, 2015), ice-sheet retreat styles (e.g.,

T

Dowdeswell et al., 2008; Ó Cofaigh et al., 2008; Graham et al.,

IP

2010; Halberstadt et al., 2016), and processes active at the

CR

grounding line, such as subglacial meltwater expulsion (e.g.,

US

McMullen et al., 2006; Horgan et al., 2013; Simkins et al., 2017). Sediment cores can complement and transcend the

AN

functionality of landform analyses by providing more detailed

M

records of glacial, biological, and oceanographic processes, as well as potential chronostratigraphic control on past marine-based ice

ED

sheet retreat. Glacial and glacimarine sediment facies

PT

interpretations and lithofacies descriptions have improved through time as a result of technological advances in subsurface imaging,

CE

bathymetric mapping, and sedimentological analyses (e.g.,

AC

Anderson et al., 1980; Domack et al., 1999; Anderson, 1999; Evans and Pudsey, 2002; McKay et al., 2009; Passchier et al., 2011, Hillenbrand et al., 2013). To date, most published facies models have been created based on cores for which geomorphic setting is not explicitly known; however, integrating core data with precise geomorphic context is necessary for confidently determining depositional environments, as well as processes acting

5

ACCEPTED MANUSCRIPT at former grounding line positions. Furthermore, proper sediment facies identification is imperative for glacial reconstructions in selecting transitions that constrain the timing of ice retreat from the continental shelf during the last glacial cycle and previous cycles.

IP

T

We present results from a detailed multibeam bathymetric survey and targeted coring mission in the Ross Sea, Antarctica

CR

(Fig. 1), during which core sites were selected primarily from

US

longitudinal transects across grounding-zone wedges deposited during and following the LGM (Fig. 3). We use sedimentological

AN

and micropaleontological analyses to define glacial retreat

M

sediment facies within a geomorphic framework and gather information about grounding line processes, enabling the

ED

development of a standardized facies model that can be

PT

extrapolated to cores, including drill cores, for which precise

CE

context is unknown.

AC

2. Regional setting The Ross Sea Embayment (Fig. 1) receives drainage from

approximately 25% of the continent’s ice, making it the largest single drainage basin in Antarctica. Multiple ice streams from the East Antarctic Ice Sheet (EAIS) and West Antarctic Ice Sheet (WAIS) flow into the embayment to form the extensive (~400,000

6

ACCEPTED MANUSCRIPT km2) Ross Ice Shelf. Subglacial geomorphic features, in particular, glacial lineations (King et al., 2009), provide a record of ice stream paleodrainage on the continental shelf (Greenwood et al., 2012; Anderson et al., 2014; Halberstadt et al., 2016). Geochemical and

T

provenance studies indicate the the EAIS filled the western Ross

IP

Sea and the WAIS dominated the eastern Ross Sea (Anderson et

CR

al., 1984; Licht et al., 2005; Farmer et al., 2006). Lineations extend to the continental shelf margin in many places, marking the

US

maximum extent of grounded ice, beyond which shelf-edge gullies

AN

represent the immediate proglacial environment during the LGM (Gales et al., 2013). However, more isolated patches of glacial

M

lineations, some of which are confined to the topsets of grounding

ED

zone wedges, record changes in ice flow during the post-LGM retreat phase (e.g., Greenwood et al., 2012; Simkins et al., 2016).

PT

Recessional moraines and grounding-zone wedges overprint the

CE

lineations, and mark the path of retreating ice as well as providing information about retreat behavior (Halberstadt et al., 2016).

AC

Large grounding-zone wedges have been identified in the Ross Sea using seismic data (Shipp et al., 1999; Howat and Domack, 2003; Mosola and Anderson, 2006; Bart and Cone, 2012), but many smaller (<10 m in height) grounding-zone wedges and recessional moraines—once only resolvable in high-frequency acoustic (i.e. CHIRP) data or side-scan sonar—are now identified using the

7

ACCEPTED MANUSCRIPT latest generation multibeam swath bathymetry system to collect large datasets that resolve sub-meter-scale features (Halberstadt et al., 2016). Types of retreat features and their configurations clearly

IP

T

differ aross the seafloor, with large grounding-zone wedges widely separated by pristine glacial lineations in the eastern Ross Sea

CR

(formerly occupied by the WAIS), in contrast to the western Ross

US

Sea (formerly occupied by the EAIS) which is characterized by numerous, typically closely spaced, small grounding-zone wedges

AN

and recessional moraines (Shipp et al., 2002; Mosola and

M

Anderson, 2006; Halberstadt et al., 2016). The exception is isolated composite grounding-zone wedges on the outer

ED

continental shelf, which suggest extended pauses in grounding line

PT

retreat (Howat and Domack, 2003; Bart et al., 2017), likely during glacial maxima. By the association of grounding-zone wedge

CE

sediment volume with duration of pause in retreat (Howat and

AC

Domack, 2003), the distribution of landforms in the eastern Ross Sea indicates the grounding line migrated landward 10s to 100s of kilometers, punctuated by long pauses of grounding line position stability (Mosola and Anderson, 2006). However, in the western Ross Sea, retreat occurred through numerous smaller events of generally less than several kilometers of retreat with shorter periods of stability. Though western Ross Sea grounding-zone

8

ACCEPTED MANUSCRIPT wedges are relatively small, there remains considerable variability in their morphologies and distribution, indicating a dynamic range of grounding line behavior during deglaciation. Thus, we largely focus our sediment facies work on the western Ross Sea where the

T

abundance of grounding line features formed by EAIS retreat

CR

US

processes based on sediment properties.

IP

provides greater opportunity for understanding retreat patterns and

AN

3. Materials and Methods

During the austral summer of 2015, we conducted a

M

multibeam survey in the western Ross Sea aboard the RV/IB

ED

Nathaniel B. Palmer using a Kongsberg EM-122 system on cruise NBP1502A. In conjunction with Knudsen CHIRP 3260 high-

PT

frequency acoustic data, the multibeam data was used to target a

CE

variety of glacial geomorphic features for coring. In total, fortynine Kasten cores (abbreviated as KC in core names) and two

AC

jumbo piston cores (JPC) were acquired (Fig. 1), primarily from the topsets, foresets, and toes of grounding zone wedges (Fig. 3). Kasten cores were cleaned, photographed, and described before sampling onboard. TorVane shear strength measurements were obtained at intervals of 5-10 cm. Samples were taken every 5 cm for water content measurements and for grain-size analysis, and every 10 cm for foraminifera assemblage investigation. Remaining 9

ACCEPTED MANUSCRIPT material was archived and shipped to Florida State University’s Antarctic Research Facility (ARF), where the cores were x-rayed and analyzed using a Geotek multi-sensor core logger (MSCL) which conducted linescan imaging, magnetic susceptibility, and

IP

T

gamma density measurements. Grain-size samples were treated with a sodium

CR

hexametaphosphate solution to disaggregate grains. A 500-μm

US

sieve was used to remove larger grains, and the finer fraction composed of sediment matrix material was analyzed using a

AN

Malvern Mastersizer Hydro 2000G laser particle size analyzer.

M

Outputs of mode, cumulative volume percent, and frequency volume percent are generated by the Malvern software, and

ED

graphical percentiles of cumulative grain-size distributions were

PT

used to calculate grain-size mean, sorting, and skewness statistics according to Folk and Ward (1957). Samples for foraminifera

CE

analysis were wet-sieved through a 63-μm sieve and dried.

AC

Foraminifera were picked from the >125 μm fraction. The finer fraction was not investigated because of strong dilution of foraminifera with abundant terrigenous material. In samples with abundant microfossils, dried samples were divided using a microsplitter prior to picking. Specimens exhibiting strong discoloration, clear signs of abrasion, and/or representing species not found in present day Ross Sea were considered to be reworked. Translucent

10

ACCEPTED MANUSCRIPT specimens showing smooth non-etched tests were considered as potentially being in situ, and they were identified and quantified as separate species. The investigated material is housed at the Institute of Paleobiology of the Polish Academy of Sciences

CR

IP

T

(Warszawa) under the catalog number ZPAL F.65.

4. Results

US

Because glacial source material and continental shelf

AN

physiography vary significantly in the Ross Sea, differences in the composition and texture of seafloor sediment should be evident

M

across various paleo-ice stream troughs (Anderson et al., 1984;

ED

Licht et al., 2005). In addition to source control, oceanographic processes, which are partly controlled by continental shelf margin

PT

physiography and bank/trough configuration, affect biological

CE

productivity and reworking of sediments by marine and biological processes (Dunbar et al., 1985). The potential for iceberg keels to

AC

disturb and rework sediments is also controlled by water depth and oceanography (Barnes and Lien, 1988). While such variability is recognized in this study, areas with high potential for iceberg reworking, such as bank tops and the shelf margin, were avoided when coring to obtain undisturbed sedimentary sections. However, sediments that have been reworked by marine currents are present over a large portion of the continental shelf and provide 11

ACCEPTED MANUSCRIPT oceanographic information, and so are included in this analysis largely based on previous studies (Dunbar et al., 1985; Anderson, 1999). Including these marine-influenced deposits, six primary facies are identified in the Ross Sea, each representing a major

T

glacial environment or process. Distinctions based on physical

IP

description, grain size, geotechnical properties, biogenic

US

brief description of each facies to follow.

CR

properties, and geomorphic setting are detailed in Table 1, with a

4.1. Facies 1: Diamicton

AN

Cores taken from paleo-subglacial geomorphic settings,

M

such as from fields of glacial lineations and topsets of grounding-

ED

zone wedges, typically sample very poorly-sorted (defined as SD=2–4φ), matrix-supported diamicton. X-radiographs indicate

PT

the diamicton is massive and relatively homogeneous (Fig. 4a).

CE

This is reflected by exceptional downcore uniformity in grain-size character and mineralogic composition (inferred from magnetic

AC

susceptibility) of the matrix material (e.g., bottom unit of KC19 in Fig. 5a), which distinguishes it from other lithofacies. Grain-size distributions consistently display very poor-sorting in all samples, but a ~10 μm mode is always dominant. Water content directly correlates with grain-size sorting and inversely with gamma density. There is a general absence of pristine biogenic material.

12

ACCEPTED MANUSCRIPT Foraminifera are sparse and dominated by specimens that are clearly reworked (e.g., bottom unit of KC19 in Fig. 6a). 4.2. Facies 2: Foraminifera-bearing diamicton

T

Diamicton was also recovered in cores taken from the

IP

foresets and toes of grounding-zone wedges. Generally, x-

CR

radiographs of this lithofacies are indistinguishable from Facies 1, showing a massive character with pebbles scattered throughout the

US

diamicton. Additionally, the grain-size distributions of the matrix material show very poor-sorting (defined as SD=2–4φ) with a ~10

AN

μm mode. Unlike Facies 1, degree of sorting and grain-size mode

M

tends to vary downcore and between different cores, as does

ED

magnetic susceptibility (e.g., bottom unit of KC48, Fig. 5b). This variability can be very minor (SD range 2–2.3φ), as demonstrated

PT

in Fig. 4b and 6a, or more extreme, displaying inconsistent sorting

CE

(SD range 1.5–3φ), varying pebble content, and containing multiple grain-size modes as large as medium sand (Fig. 7b).

AC

Though downcore grain size may be variable, we do not observe grading. Regardless of grain-size character, an assemblage of pristine calcareous benthic foraminifera (bottom unit of Fig. 6b) dominated by Globocassidulina subglobosa is present. Seemingly reworked foraminifera are also present, but in minor (<5 specimens per gram of dry sediment) amounts. In some cores, planktonic Neogloboquadrina pachyderma are present in moderate (up to 13

ACCEPTED MANUSCRIPT 30% of assemblage) amounts. We term this diamicton ―foraminifera-bearing‖ because the majority of foraminifera present are in good condition compared to the specimens found in Facies 1 and the benthic foraminifera appear to be in situ.

IP

T

4.3. Facies 3: Foraminifera-bearing pelletized diamicton

CR

A third type of diamicton occurs in several cores taken from grounding-zone wedge foresets and toes. The composition of

US

this diamicton is similar to the facies discussed in Section 4.2, in that no difference is evident in grain-size distributions (212–225

AN

cm of KC48 in Fig. 5b) or foraminifera assemblages (212–225 cm

M

of KC48 in Fig. 6b) from Facies 2. Like Facies 1 and 2, x-

ED

radiographs of Facies 3 show a primarily massive structure, but differ by displaying a somewhat mottled appearance (Fig. 4b, 4d).

PT

This is due to the presence of abundant granule- to pebble-sized

CE

soft sediment pellets. The pellets, which are round or prolate semiconsolidated clots of sediment, are referred to as soft-sediment

AC

clasts in the Mertz Trough of East Antarctica (McMullen et al., 2006), polymicts (Licht et al., 2009) and mud grains (Bonaccorsi et al., 2000) in the Ross Sea, and till pellets in the Ross Sea (Domack et al., 1999; Howat and Domack, 2003), McMurdo Sound (Cowan et al., 2012; 2014), and the Larsen continental shelf of NE Antarctic Peninsula (Evans et al., 2005). Grain-size analysis was first conducted separately on matrix material and chemically 14

ACCEPTED MANUSCRIPT disaggregated isolated pellets to identify any differences between internal pellet grain size and matrix material. Within single intervals, no clear grain-size variability was identified between the pellets and the matrix; thus, the soft sediment clasts are derived

T

from the same material as the hosting matrix. Water content is

Facies 1 and Facies 2 (Table 1).

US

4.4. Facies 4: Relatively well-sorted silt

CR

IP

higher and shear strength is markedly lower in this unit than in

Cores containing relatively well-sorted fine silt with a ~10

AN

μm dominant mode (e.g., 47–217 cm in KC17, Fig. 5c) were

M

recovered across portions of the western Ross Sea. This similarity

ED

of this mode to that of Facies 1–3 implies a common source. This lithofacies is classified as ―poorly-sorted‖ (defined as SD=1–2φ)

PT

according to Folk and Ward (1957), but we refer to it as ―relatively

CE

well-sorted‖ as it is better sorted than any other lithofacies. Its split core surface has a distinctively smooth and waxy appearance.

AC

Facies 4 texture is massive (Fig 4c) to weakly laminated (Fig 4d) in character. Deposits range from 1 to 170 cm in thickness. Stratigraphically, these deposits generally occur within Facies 2, 3, or 5 (discussed below) or at transitions from Facies 2 to Facies 5. When found within Facies 2 and 3 in cores on grounding-zone wedge foresets and toes, this lithofacies typically contains a foraminiferal assemblage of primarily minute (~150 μm) 15

ACCEPTED MANUSCRIPT calcareous benthic foraminifera (e.g., 47–217 cm in KC17 in Fig. 6c) including Globocassidulina. The thickest occurrence of this lithofacies (47–217 cm in KC17) contains a second dominant species, Cassidulina neoteretis, which is not found in great

T

abundance in other cores. The fine silts are distributed across large

IP

portions of the continental shelf, in most cases within tens of

CR

kilometers seaward of subglacial meltwater channels (Simkins et al., 2017), but they have also been recovered as far 75 km seaward

AN

channel (NBP9801 PC55 in Fig. 8).

US

of the continental shelf margin and 250 km from an observed

M

4.5. Facies 5: Diatomaceous sandy silt

ED

Diatomaceous sandy silt is a common surface sediment across the Ross Sea, found at the top of every NBP1502A core.

PT

Our CHIRP data show that this lithofacies drapes the seafloor,

CE

similar to previous observations (Shipp et al., 1999). Regionally, the lithofacies thickens to the west and the north (Fig. 9), based on

AC

cores from cruises DF80, NBP9401, NBP9407, NBP9501, NBP9801, NBP9902, and NBP1502A. Unlike the other facies discussed, this unit is olive in color and characterized by a ~40 μm grain-size mode (topmost unit in each core, Fig. 5a–c). Occasionally the unit displays an additional mode at ~10 μm (e.g., KC19 and KC48, Fig. 5a, 5b). The 40 μm silt mode is composed of diatoms and minor amounts of sponge spicules and radiolarians; 16

ACCEPTED MANUSCRIPT whereas, the 10 μm silt mode is dominated by terrigenous silt with diatom fragments (Fig. 10). Sand content is typically <15% unless concentrated within distinct lenses of gravel and sand (e.g., KC48, Fig. 5b), which are interpreted as ice-rafted debris layers. Grain-

T

size distributions tend to be fine-skewed, unlike the normal

IP

distributions of Facies 1–4, and sorting is significantly better than

CR

in Facies 1–3 but not as strong as Facies 4. Water content is twice as high as in Facies 1–3, directly correlating with grain-size

US

sorting. Ash is located near the base of this lithofacies in nine

AN

cores in the southwestern sector of the Ross Sea (Fig. 8), identified visually and in X-rays as thin millimeter-scale laminations (Fig.

M

4e) and by a magnetic susceptibility signature that is an order of

ED

magnitude higher than in other facies (c.f. Licht et al., 1999). This lithofacies is heavily bioturbated, indicated by lack of sedimentary

PT

structure (with the exception of disturbed ash laminations; Fig 4e)

CE

and the presence of burrows that penetrate into underlying units and interrupt the otherwise sharp lower contact. Agglutinated

AC

foraminifera (e.g., Miliammina arenacea) usually greatly outnumber calcareous forms, which are rarely found in this unit. Sponge spicules are common, often giving the unit a somewhat fibrous appearance and spongy texture. 4.6. Facies 6: Winnowed sand, gravel, bioclastics

17

ACCEPTED MANUSCRIPT Grab samples and the tops of sediment cores from bank tops and the outer continental shelf, collected on previous cruises (DF80 and DF84), contain coarse-skewed sediments with a clear truncation of material smaller than fine sand (Fig. 9; Anderson et

T

al., 1980; 1984, Dunbar et al., 1985; Anderson, 1999). Kasten

IP

coring attempts during cruise NBP1502A of banktops were

CR

unsuccessful, suspected to be due to contact with a coarse,

impenetrable substrate. This lithofacies is in contrast to the finer

US

sediments of Facies 5 that blanket much of the inner continental

AN

shelf. Calcareous material is present in the form of bioclastics and foraminifera, including some LGM-aged carbonates (Taviani et al.,

PT

5. Discussion

ED

M

1993).

CE

5.1. Distinguishing till from proximal glacimarine sediments

AC

We recognize three diamicton facies (Facies 1-3), each

with subtle but important differences that reflect depositional processes. Facies 1 occurs in sediment cores from glacial lineations and grounding-zone wedge topsets, which are known subglacial features. The diamicton has pebbles throughout and displays downcore homogeneity in the matrix material, both textural (from grain-size measurements) and compositional

18

ACCEPTED MANUSCRIPT (inferred from magnetic susceptibility) (Fig. 5a), which is consistent with traditional sedimentological descriptions of till (e.g., Anderson et al., 1980; Anderson, 1999). Facies 2 was recovered in cores from foresets and toes of

IP

T

grounding-zone wedges. Though it may display some grain-size variability (Fig. 7b), it is most often similar to Facies 1 with

CR

respect to grain sorting (Fig. 7a) and downcore textural and

US

mineralogic homogeneity (Fig. 5b). Such similarities with till and the occurrence of these deposits on the foreset slope leads us to the

AN

interpretation that these are debris flow deposits that were

M

transported from the grounding-zone wedge crest down the lee slope, building the foresets (Fig. 2c); thus, we interpret Facies 2 as

ED

a grounding-line proximal glacimarine facies. Sedimentological

PT

similarities between this lithofacies and till are precisely why distinguishing the two is difficult in the absence of geomorphic

CE

context (Kurtz and Anderson, 1979). However, grounding-zone

AC

proximal settings are thought to be hospitable to foraminifera, resulting in the presence of pristine, presumably in-situ calcareous benthic foraminifera (Globocassidulina, Trifarina, and Cibicides) found in NBP1502A cores (Fig. 6b) and other identified grounding-zone proximal facies (Anderson, 1972; Majewski and Anderson, 2009; Bart and Cone, 2011; Bart et al., 2016). These foraminifera are assumed to inhabit the environment just seaward

19

ACCEPTED MANUSCRIPT of the grounding line along grounding-zone wedge foresets and toes, and locally incorporated into the debris flows along with noticeably reworked specimens. We find that planktonic Neogloboquadrina pachyderma and seemingly reworked

T

foraminifera are also present in minor amounts; however, the

IP

largely pristine foraminiferal assemblage sets this lithofacies apart

CR

from till, which is instead dominated by reworked specimens (Fig.

US

6a).

Facies 3 is also recovered in cores from foresets and toes of

AN

grounding-zone wedges. Abundant pebbles, a very poorly-sorted

M

(SD=2–4φ) matrix grain-size distribution, and dominance of pristine, seemingly in-situ Globocassidulina subglobosa in the

ED

foraminiferal assemblage leads us to interpret this facies as a

PT

grounding-zone proximal glacimarine facies. However, the presence of pellets suggests deposition via a different process from

CE

Facies 2. Given the similarity of their internal grain-size

AC

distributions to that of till (Facies 1), we refer to the pellets as ―till pellets.‖ They are thought to form when sediment pore water temperature within tills drops below the pressure-melting point and subglacial water within the till is extracted upward, bringing aggregates of sediment with it to accrete on the ice base (Fig. 2b; Christoffersen and Tulaczyk, 2003; Cowan et al., 2012). As accretion of basal debris continues, till pellets can migrate several

20

ACCEPTED MANUSCRIPT meters above the ice-bed interface, evidenced by the ―mud clots‖ recognized as high as 4.83 m above the bed in the 1968 Byrd Station ice core (Gow et al., 1979). The till pellets are transported within the basal debris layer of flowing ice to the grounding line

T

and some distance beyond, at which point the sediment-laden ice

IP

base is exposed to seawater and basal melting causes rain-out of

CR

till pellets and other basal debris (Fig. 2c; Domack et al., 1999; Cowan et al., 2012). Therefore, Facies 3 is interpreted to represent

US

deposition directly from melting of sediment-laden basal ice. The

AN

distribution of this facies is constrained to within 1.2 km (0.8 km on average) of the nearest grounding-zone wedge (Fig. 11),

M

providing an upper limit on the extent of paleo-basal debris

ED

meltout extent from the grounding line. This estimate is similar to the observed sediment rain-out from the contemporary Mackay

PT

Glacier that extends no further than 1.5 km past the grounding line

CE

(Powell et al., 1996) and a basal melt-out zone of ~500 m from the

AC

Whillans Ice Stream grounding line (Christianson et al., 2016). We find that extent of Facies 3 beyond the grounding line

is independent of corresponding grounding-zone wedge geometries (e.g., width and height), but there is a rough correlation between the thickness of the facies and grounding-zone wedge height (Fig. 12). This suggests that landform height may partially manifest in the accumulation of the basal debris meltout deposits. Furthermore,

21

ACCEPTED MANUSCRIPT grounding-zone wedge sediment volume, which is a function of sediment supply to the grounding line (Batchelor and Dowdeswell, 2015), has been suggested to correlate with the duration of the pause in retreat (Howat and Domack, 2003; Ó Cofaigh et al.,

T

2008; Jakobsson et al., 2012). Thereby, the relationship between

IP

the thickness of Facies 3 and grounding-zone wedge size (here

CR

expressed as height) should provide some relative indication of grounding-line position stability, where larger, longer-lived

US

grounding-zone wedges are associated with thicker units of Facies

AN

3 than smaller grounding-zone wedges that likely formed over shorter periods of time. However, we acknowledge that other

M

controls on the thickness of Facies 3 might include basal melt rates

ED

seaward of the grounding line, as well as thermal conditions at the base of the grounded ice which control the amount of debris

CE

PT

entrained within the flowing ice.

AC

5.2. Evidence of subglacial meltwater delivery to paleogrounding lines Multibeam data has revealed anastomosing channels cut into bedrock on the Antarctic continental shelf in, for example, Marguerite Bay (Anderson and Oakes-Fretwell, 2008) and Pine Island Bay (Lowe and Anderson, 2003; Nitsche et al., 2013). In both these regions, conspicuous terrigenous silts have been 22

ACCEPTED MANUSCRIPT observed within glacimarine sediments (Kennedy and Anderson, 1989; Kilfeather et al., 2011; Hillenbrand et al., 2013; Witus et al., 2014). In Pine Island Bay, these silt deposits are relatively wellsorted with a ~10 μm mode and are widely dispersed as a seafloor

T

draping unit. This led to the interpretation that they originate from

IP

buoyant plumes sourced from subglacial meltwater discharge at the

CR

grounding line (Witus et al., 2014), consistent with the observation of suspended sediments (Jacobs et al., 2011) and similar deposits

US

(Smith et al., 2017) beneath the modern ice shelf. Seaward of

AN

subglacial meltwater channels that are incised into sedimentary substrate in the western Ross Sea (Fig. 7; Wellner et al., 2006;

M

Greenwood et al., 2012; Simkins et al., 2017), we observe the same

ED

distinct fine silt deposits (Facies 4; Fig. 5c) bounded by or within grounding-zone proximal (Facies 2-3) and open marine deposits

PT

(Facies 5; discussed in Section 5.3.1). These combined

CE

observations support the interpretation of Facies 4 as meltwater plume deposits, and their discovery within or directly above Facies

AC

2 or 3 suggests meltwater systems were active during groundingline retreat. The prevalent ~10 μm mode we identify in this lithofacies is also conspicuous within the till, supporting a subglacial source for Facies 4. Improved sorting likely occurs due to channelized and/or distributed subglacial meltwater transport and proglacial sediment suspension in buoyant meltwater plumes.

23

ACCEPTED MANUSCRIPT Meltwater deposits recovered near grounding zone wedges reflect proximal glacimarine foraminifera assemblages, dominated by Globocassidulina subglobosa and other calcareous benthics such as Cassidulina neoteritis. However, all foraminifera appear to be

T

small and juvenile. In some cores, this lithofacies is barren of

IP

foraminifera (e.g., 47–110 cm in Fig. 6c), likely due to periods of

CR

intense meltwater activity resulting in rapid sedimentation (Ó

US

Cofaigh and Dowdeswell, 2001).

AN

5.3. Open marine sedimentation reflects oceanographic

M

conditions

ED

5.3.1. Low-energy open marine depositional environments

PT

Facies 5 is composed of primarily terrigenous silt that settled from suspension, with <15% sand interpreted as ice-rafted

CE

debris and a significant (10-30%) diatom component. Equivalent

AC

facies have been referred to as ―compound glacial marine‖ to reflect a blend of both fine, current-derived sediments and coarser, ice-rafted sediments (Anderson et al., 1980); however, we use the term ―open marine‖ to describe this facies from a glacial retreat perspective rather than by composition. In the Ross Sea, pebbles and sand are rare in diatomaceous open marine sediments. Some cores contain discrete lenses of gravel and sand at the base of the

24

ACCEPTED MANUSCRIPT lithofacies (Fig. 5b), which we interpret as rare but intense episodes of ice-rafting during ice-shelf collapse which has been shown to occur at least twice during post-LGM ice sheet retreat in the Ross Sea (i.e., Yokoyama et al., 2016).

T

The majority of Facies 5 has a grain-size distribution that

IP

tends to be fine-skewed with a ~40 μm mode that is composed of

CR

primarily intact diatoms (Fig. 10b). The fine tail of the grain-size

US

distribution is composed of terrigenous silt and diatom fragments (Fig. 10a) and is likely delivered to the area by marine currents.

AN

Modern year-averaged marine currents in the southwestern Ross

M

Sea generally flow to the south and west at 5-9 cm/s (Pillsbury and Jacobs, 1985; Picco et al., 1999). Dunbar et al. (1985) suggested

ED

currents are strong enough to transport fine sediments from the

PT

outer continental shelf and bank tops to the deep troughs and basins of the western Ross Sea. However, Jaeger et al. (1996)

CE

argued that these currents are variable in direction and speed and,

AC

therefore, incapable of significant lateral transport, suggesting much of the siliceous material is locally derived. Westward thickening of Facies 5 (Fig. 9) reflects a

combination of greater primary productivity and regional circulation patterns. An association between the location of open water and high percentages of organic carbon and biogenic silica led Dunbar et al. (1985) to conclude that the persistent Terra Nova 25

ACCEPTED MANUSCRIPT Bay polynya in the western Ross Sea (Kurtz and Bromwich, 1985) drives relatively high primary productivity due to more sunlight exposure and potentially warmer surface waters. Arrigo and van Dijken (2003) observed a significant relationship between polynya

T

size and rate of annual primary production, with larger polynyas

IP

being more productive than smaller polynyas, thus leading to

CR

enhanced biogenic sedimentation. Facies 5 also thickens to the north (Fig. 9), which reflects greater longevity of open marine

US

conditions following the LGM on the outer shelf in western Ross

AN

Sea.

M

The exact source of terrigenous silt in Facies 5 is uncertain. Anderson et al. (1984) favor reworking and winnowing of glacial

ED

sediments on the outer continental shelf and banks and transport by

PT

impinging geostrophic currents as an important source of terrigenous silt on the shelf. While we recognize this may be a

CE

contributor, we also consider delivery by subglacial meltwater in

AC

light of the recent discovery of meltwater channels and distinct meltwater facies (Facies 4). Grain-size distributions of the open marine facies occasionally display an additional dominant mode at ~10 μm (Fig. 5b), analogous to the ~10 μm mode found in subglacial tills and meltwater deposits. We interpret specific intervals of dominance of this mode within the open marine facies (Fig. 5b) to represent episodes of meltwater delivery to the

26

ACCEPTED MANUSCRIPT grounding line during which sediment-laden plumes propagate seaward to open water. The shallow (≤430 m) calcite compensation depth (CCD)

IP

calcareous benthic foraminifera and high proportion of

T

in the Ross Sea today (Kennett, 1968) is reflected in the lack of

agglutinated species in the open marine facies. Sluggish current

CR

conditions and associated accumulations of siliceous detritus on

US

the seafloor facilitates organic decay, which produces CO2 that inhibits CaCO3 stability, thereby elevating the CCD and preventing

AN

preservation of calcareous foraminifera (Kennett, 1968). A similar

M

relationship is found in the Weddell Sea (Anderson, 1975) and the George V-Adelie continental margin (Milam and Anderson, 1981).

ED

These combined observations of fine sediments, abundance of

PT

siliceous biogenic material, and lack of calcareous foraminifera are all reflective of the sluggish current conditions that exist on much

AC

CE

of the Ross Sea continental shelf.

5.3.2. Marine current influence on marine sedimentation Facies 6 occupies a significant portion of the outer continental shelf, slope, and banks (RGM, Fig. 9). Due to the coarse and often impenetrable nature of these sediments, we avoided coring these regions. However, these sediments are

27

ACCEPTED MANUSCRIPT important oceanographic indicators, and so we rely on published work to assess their relationship with other surface sediments discussed above. Coarse-skewed grain-size distributions indicate winnowing of sediments finer than ~125 μm (Anderson et al.,

T

1984; Dunbar et al., 1985). The source material was likely glacial

IP

or glacimarine sediments that were exposed to vigorous marine

CR

currents following ice-sheet retreat, and so we recognize the final product as ―open marine.‖ However, these sediments have clearly

US

been subjected to marine current reworking, leading us to instead

AN

refer to the winnowed sediments as ―residual glacimarine (RGM),‖

M

following Anderson et al. (1980).

Currents on the shelf break have mean velocities between

ED

17 and 18 cm/s (Jacobs et al., 1974). These conditions are

PT

reflected in bottom photographs, where current ripples indicate bedload transport of sorted sands (Anderson et al., 1984). Singer

CE

and Anderson (1984) performed laboratory flume experiments in

AC

which a maximum sustained current velocity of ~18 cm/s generated RGM-like sediments from till-like beds that were continuously mixed. Mixing of the sediments results in a reduction in cohesion, otherwise only a thin armored surface is created. This mixing on the Ross Sea floor is likely achieved primarily through bioturbation; bottom photographs show abundant benthic life and bioturbation in these areas. An abundance of randomly-oriented

28

ACCEPTED MANUSCRIPT iceberg furrows atop banks (Fig. 9) indicates that iceberg turbation may also play a role in mixing sediments at shallow water depths. Similar associations between surface sediments and marine currents are recognized in the Weddell Sea and on the Wilkes Land

T

and Pennell Coast continental shelves, where size grading has been

IP

related to variations in measured current velocities (Anderson,

CR

1999). Offshore of the Pennell Coast, volcaniclastic sands derived from Cape Adare have been distributed westward along and onto a

US

wide portion of the continental shelf by strong currents associated

AN

with impinging boundary currents (Rodriguez and Anderson, 2004). The wide spatial and bathymetric distribution of RGM and

M

associated sands suggests that similar deposits may be present, and

ED

perhaps even more pervasive in the pre-LGM stratigraphic record.

PT

Today, Modified Circumpolar Deep Water (MCDW), an extension of Circumpolar Deep Water (CDW), is a relatively warm

CE

water mass that impinges onto the Ross Sea continental shelf at

AC

intermediate and surface levels within the water column (Jacobs et al., 1985; Anderson, 1999). It impinges at high velocities (Jacobs et al., 1974) along the eastern sides of troughs and adjacent bank edges (Orsi and Wiederwohl, 2009; Dinniman et al. 2003, 2011) while colder, denser water is directed offshore along the western sides of troughs (Gordon et al., 2004; Orsi and Wiederwohl, 2009). While some of the winnowed fine material from the

29

ACCEPTED MANUSCRIPT banktops and outer shelf is swept seaward by outflow, we believe that much of the fine material is driven by high current velocities associated with the inflow of MCDW along the eastern sides of troughs into the middle and inner continental shelf where lower

T

current velocities allow it to settle (Pillsbury and Jacobs, 1985;

IP

Picco et al., 1999). This fine material is represented by the fine

CR

silt and diatom fragment component of Facies 5 (Fig. 10a).

US

By sweeping away diatoms and preventing the

accumulation of corrosive siliceous detritus, the strong currents

AN

leave the outer shelf and banks more conducive for calcite

M

preservation than the inner shelf. As a result, calcareous foraminifera are stable in regions of MCDW incursion (Osterman

ED

and Kellogg, 1979; Milam and Anderson, 1981). Additionally, the

PT

nutrients delivered by impinging MCDW allows other carbonate organisms to thrive (Taviani et al., 1993; Elverhoi and Roaldset,

CE

1983; Anderson, 1999). In the northwestern Ross Sea, cold-water

AC

bioclastic carbonates, including bryozoans, barnacles, pelecypods, and corals (Fig. 9) occur on the outer shelf and on banks. Radiocarbon ages for these deposits range back to the LGM, so these banks were refugia for carbonate organisms during more extreme glacial conditions (Reid, 1989; Taviani et al., 1993). In summary, the presence of RGM facies indicates impinging MCDW on the outer shelf, whereas Facies 5 reflects little

30

ACCEPTED MANUSCRIPT influence of marine currents on the inner/middle continental shelf. Because impinging warm currents have been shown to affect modern grounding line stability (Shepherd et al., 2004; Thoma et al., 2008; Jacobs et al., 2011), it is important to be able to

CR

IP

T

recognize signs of their influence in the paleo-record as well.

5.4. Sedimentary record of ice sheet retreat

US

Geomorphic features indicate the ice streams originating

AN

from the WAIS in the eastern Ross Sea experienced large grounding line retreat events of 10s to 100s of kilometers that were

M

punctuated by relatively long periods of stability, while ice streams

ED

of the EAIS in the western Ross Sea generally experienced numerous smaller (100s of meters to several kilometers) grounding

PT

line retreat events, with the exception of larger grounding line

CE

retreat events from LGM positions in Northern Drygalski Trough and Joides Trough (Halberstadt et al., 2016). One of the

AC

objectives of this study was to determine if these differences are manifest in the sedimentary record, to be used for interpreting retreat patterns from cores that lack geomorphic context. The vast majority of the cores collected in Ross Sea have sampled till (Anderson et al, 1980, 1984; Domack et al, 1999; Mosola and Anderson, 2006; McGlannan et al., 2017), according

31

ACCEPTED MANUSCRIPT to the criteria outlined by Anderson (1999). These data were used to argue for expansion of the ice sheet across the continental shelf, an argument that is now substantiated by multibeam data showing shelf-wide occurrence of subglacial features (Anderson et al.,

T

2014). While we initially expected the sedimentological properties

IP

of till to reflect differences in subglacial processes and past ice-

CR

stream behavior across the Ross Sea (e.g., soft and stiff tills of Dowdeswell et al., 2004), distinct patterns do not appear to exist.

US

Ongoing research focuses on variability in till properties, such as

AN

shear strength and water content, in relation to geomorphic features and substrate beneath the till. Classical evidence of structural

M

deformation (shear planes, aligned clasts) is not clear in our core x-

ED

rays as it is in previous studies elsewhere in Antarctica (e.g., Ó Cofaigh et al., 2005, 2007); however, others suggest that

PT

homogenization, a result of thorough deformation, eliminates

CE

macroscopic evidence of deformation (Evans et al., 2006; Reinardy et al., 2011). There has been some success with analysis of

AC

microfabrics in distinguishing different types of diamictons (e.g., Menzies, 2000; Van der Meer et al., 2003; Menzies et al., 2006; Reinardy et al., 2011), and in the future this method may show the best evidence of subglacial transport processes by indicating the nature and magnitude of subglacial stresses.

32

ACCEPTED MANUSCRIPT Within a single-phase grounding line retreat scenario, facies successions are represented by increasingly ice-distal facies upcore. In the Ross Sea, typical grounding-zone wedge topset and glacial lineation facies successions are represented by till sharply

T

overlain by open marine sediments (Fig4f, 5a). Previous studies

IP

have suggested that this is the most common succession across the

CR

Ross Sea (Anderson et al., 1984, 1991), especially in the eastern Ross Sea (Mosola and Anderson, 2006). However, we find that

US

cores collected on grounding-zone wedge foresets and toes contain

AN

grounding-zone proximal sediments sharply overlain by draping

M

open marine deposits (Fig. 5b).

We consider grounding-zone proximal sediments the most

ED

important sediment facies for indicating retreat style. They contain

PT

evidence of multiple sedimentary processes acting at paleogrounding lines. Non-pelletized grounding-zone proximal

CE

sediment (Facies 2) is interpreted to represent debris flows that

AC

construct grounding-zone wedge foresets; therefore, the facies extent should be roughly equal to foreset length (Fig. 2a). Because the foreset length varies amongst grounding-zone wedges, precise proximity to a paleo-grounding line cannot be determined in geomorphically-blind cores that contain this unit. However, the pelletized grounding-zone proximal facies (Facies 3) is interpreted to represent melt-out of basal debris (Fig. 2a, c). Because basal

33

ACCEPTED MANUSCRIPT debris is deposited within ~1.2 km of the grounding line (Fig. 11), the majority of any extensive ice shelf would have been debrisfree, with the exception of englacial eolian material and material entrained along the margins of nunataks, volcanoes, and other

T

bedrock features. This is reflected in the minimal coarse, ice-rafted

IP

material observed in Facies 5, which suggests that calving icebergs

CR

from the ice shelf did not contain significant englacial material.

US

Traditionally, facies models recognize a sub-ice shelf facies usually described as either muddy gravel or pebble-free muds

AN

(Kennedy and Anderson, 1989; Domack et al., 1999, Anderson et

M

al., 1991, McKay et al., 2009). The muddy gravel facies is equivalent to our grounding line proximal facies (Facies 2, and

ED

perhaps Facies 3), whereas pebble-free muds could represent

PT

meltwater deposits (Facies 4; Kennedy and Anderson, 1989; Anderson et al., 1991; McKay et al., 2009) or transport of material

CE

from the grounding line by ocean currents (Domack et al., 1999).

AC

Because basal debris melts out near the grounding line and we see very little evidence of IRD in open marine deposits, we argue that the only significant sedimentation that occurred beneath paleoRoss Sea ice shelves was via meltwater plumite deposition (Fig. 2a), though advection of biogenic material and fine sediments beneath the ice shelf may also occur to a minor degree. Over 50 sediment cores collected through access holes in the modern Ross

34

ACCEPTED MANUSCRIPT Ice Shelf were found to contain only overcompacted till, providing an example of a sub-ice shelf setting where no modern terrigenous sedimentation or advection of diatoms occurs beneath the ice shelf (Kellogg and Kellogg, 1986; 1988). Thus, we do not recognize an

T

exclusive ―sub-ice shelf‖ facies, as it is essentially a zone of non-

IP

deposition in the absence of meltwater or marine current influence,

CR

which are not unique to sub-ice shelf environments. Although a distint facies is not observed (or expected), Yokoyama et al. (2016)

US

demonstrate that coupled analysis of diatom and beryllium-10

M

presence of past ice shelves.

AN

concentrations in sediment cores is useful for identifying the

Sedimentary successions are clearly highly dependent on

ED

geomorphic context. However, the question remains whether these

PT

successions contain information about retreat patterns. The seascape in the western Ross Sea is dominated by many closely

CE

spaced recessional features that record frequent but short-lived

AC

periods of grounding line position stability (Fig. 13a–c; Halberstadt et al., 2016; Simkins et al., 2017). Associated cores display a variety of successions, dependent on proximity to the grounding line, meltwater channels, and volcanic provinces. Therefore, a single core is not useful for determining retreat patterns where geomorphic context is unknown. However, examination of the sedimentary successions in a transect or

35

ACCEPTED MANUSCRIPT network of cores in such a region is more telling. A wide spatial distribution of thin pelletized units (Facies 3) below the open marine facies (Fig. 13d) in cores suggests a western Ross Sea-style retreat. Conversely, the seascape in the eastern Ross Sea is

T

characterized by widely spaced GZWs separated by tens of

IP

kilometer stretches of seafloor with pristine glacial lineations (Fig.

CR

13e–g). This is reflected in the majority of sediment cores from the eastern Ross Sea being composed of till that is directly overlain

US

by open marine sediments (Mosola and Anderson, 2006). Even

AN

though the pelletized facies (Facies 3) can be thick where there are

localized (Fig 13h).

M

large composite wedges in the eastern Ross Sea, it is highly

ED

Strategic coring of specific geomorphic features provides

PT

context for accurate facies interpretations, which in turn has allowed us to match certain facies successions to particular

CE

geomorphic environments. Single cores cannot reveal overall

AC

retreat patterns, but the presence of Facies 3 in a core is indicative of a nearby grounding line. Only the observation of this facies (or lack thereof) in several cores may be used to determine the relative amount (and possibly size, Fig. 12) of grounding-zone wedges, and thus retreat patterns.

5.5. Applications to pre-LGM glacial history 36

ACCEPTED MANUSCRIPT Another motivation of this study is to extrapolate what has been learned from facies analysis in geomorphic context to preLGM sediments. The seismic data that is used to select drill sites for upcoming projects such as IODP Expedition 374: Ross Sea

T

West Antarctic Ice Sheet History (McKay et al., 2017) is not of

IP

sufficient resolution to capture small-scale grounding-zone wedges

CR

and channels. Therefore, facies interpretations of ancient sediments are often done without clear context. The limitations involved with

US

obtaining and studying ancient, pre-LGM glacial deposits almost

AN

necessitates a Lyellian approach of using recent deposits to interpret older sediments and speculate geomorphic settings.

M

Combined seismic data and drill cores indicate marine ice sheet

ED

expansion in the Ross Sea as early as late Oligocene time (see Anderson et al., in press for recent review). Prior results also

PT

indicate that glacial conditions were very different prior to the

CE

Pliocene; diatomaceous sediments were not prevalent and diamictons interpreted as till occurred less frequently (Barrett,

AC

1975; Balshaw, 1981). However, closer inspection of late Oligocene and Miocene stratigraphy reveals some similarities to the post-LGM sediments discussed in this study. Within the late Oligocene-early Miocene section sampled in Site 270, there are diverse calcareous foraminiferal assemblages present and two intervals of soft sediment clasts (Hayes and Frakes, 1975; Leckie

37

ACCEPTED MANUSCRIPT et al., 1983). The clasts are also found within a similar interval in Site 272 (Hayes and Frakes, 1975). Previous researchers speculated the sediment clasts form by slumping of semi-lithified muds (Hayes and Frakes, 1975; Leckie and Webb, 1983);

T

however, Hayes and Frakes (1975) note that the units themselves

IP

are undeformed. Additionally, Leckie and Webb (1983)

CR

discovered recycled Late Cretaceous and Paleogene foraminifera within the soft sediment clasts. These observations are consistent

US

with the Cowan et al. (2012) model of basal freeze-on of till

AN

material which rains out as pellets at the grounding line, and with our descriptions of pelletized grounding-zone proximal

M

glacimarine facies. Therefore, we interpret the sediment clasts as

ED

evidence of nearby marine-based grounded ice, and perhaps no more than ~1.2 kilometers from the grounding line if the ice shelf

PT

basal temperature regime and volume of entrained debris were

CE

similar to the post-LGM deglaciation.

AC

During the Miocene, foraminifera underwent a distinct shift

from diverse calcareous benthic assemblages dominated by Globocassidulina spp. to a 60% agglutinated assemblage (Leckie and Webb, 1983). This change could reflect gradual cooling and more saline waters or increased diatom concentrations, leading to an elevated CCD and poor calcite preservation (Leckie and Webb, 1983). Grain-size distributions (Barrett, 1975) show no evidence

38

ACCEPTED MANUSCRIPT of winnowing, indicating currents were sluggish during this time, facilitating organic decay and further inhibiting calcite preservation (Milam and Anderson, 1981; Dunbar et al., 1985). The presence of Globocassidulina is consistent with the post-LGM

T

grounding-zone proximal facies, whereas diatom abundance and a

IP

shift from calcareous to agglutinated foraminifera assemblages is

CR

consistent with the post-LGM open marine facies, suggesting largely quiescent marine currents. Furthermore, some samples

US

within the agglutinated zone even show bimodality with strong

AN

coarse silt and fine silt components (Barrett, 1975), similar to the open marine facies in this study. If the same processes control

M

these modes as in the post-LGM record, diatoms and meltwater

ED

processes may both have been significant contributors to deposition during the early Miocene, though large outwash

PT

channels are not seen in seismic records until the middle Miocene

CE

(Anderson and Bartek, 1992). These similarities to post-LGM sediments suggest that the post-LGM record may be our best

AC

resource for interpreting more ancient stratigraphic records.

6. Conclusions 1. The Ross Sea is dominated by six primary sediment facies, though variations may be present due to localized processes. Till (Facies 1) is characterized by downcore homogeneity in 39

ACCEPTED MANUSCRIPT mineralogy and texture, is very poorly-sorted and may contain reworked foraminifera, if any. Two primary grounding-zone proximal processes create unique glacimarine facies. Debris flow deposits (Facies 2) consist of very poorly-sorted diamicton and

T

basal debris meltout deposits (Facies 3) consist of very poorly-

IP

sorted pelletized diamicton. Both deposits may exhibit downcore

CR

variability and, along with some clearly reworked specimens, primarily contain a diverse in-situ calcareous foraminifera

US

assemblage that distinguishes them from till. Subglacial meltwater

AN

deposits (Facies 4) are relatively well-sorted fine silts with a dominant ~10 μm mode. This facies is widespread and occurs

M

within grounding-zone proximal and open marine deposits. In

ED

grounding-zone proximal cores it typically marks the transition from grounding-zone proximal to open marine. Thus, meltwater

PT

discharge was active during grounding line retreat. Open marine

CE

deposits (Facies 5) are diatomaceous and have an agglutinated foraminifera assemblage due to an elevated CCD and sluggish

AC

marine currents. They are often fine-skewed, demonstrating partial origin as the winnowed product of other sediments. The residual glacimarine sediments (Facies 6) from which they are partially derived are located on the outer continental shelf and banks and are coarse-skewed with truncations that exclude the finer sediments

40

ACCEPTED MANUSCRIPT that have been swept away by strong currents. They contain calcareous foraminifera and macrofossils. 2. The basal debris meltout zone (Facies 3) extends no further than ~1.2 km from the grounding line, and averages ~750 m. Thus, the

IP

T

presence of this facies in a geomorphically-blind core indicates a nearby grounding zone. 3. Foraminifera data integrated with

CR

sedimentology reveals that calcareous foraminifera occur in

US

grounding-zone proximal glacimarine deposits and residual glacimarine sediments. The latter represent high energy open

AN

marine settings where currents deliver nutrients and sweep away

M

siliceous material.

ED

4. Small scale grounding line events characterize the western Ross Sea, and are reflected in sediment cores by the prevalence of the

PT

pelletized grounding-zone proximal glacimarine facies, which is

CE

the best-constrained indication of nearby grounding-zones. In contrast, the pelletized facies is rarely found in the eastern Ross

AC

Sea, where retreat episodes are marked by widely separated grounding zone wedges indicative of episodic retreat and longlived pauses of the grounding line, and sedimentary successions are dominated by till overlain by open marine facies with rare proximal glacimarine facies. Thus, a single core is not indicative of overall retreat patterns, but a transect or network of cores may be useful. 41

ACCEPTED MANUSCRIPT 5. The results presented here are applicable through the PlioPleistocene, much of the Miocene, and perhaps portions of the late Oligocene. Miocene sediments appear to demonstrate the same relationships between foraminifera composition and diatom

T

presence that we observe in post-LGM sediments. Grain-size

IP

analyses for pre-LGM sediments are limited but those that have

CR

been performed suggest sediment facies that are similar to those of the post-LGM. Soft sediment clasts that appear to have been

US

created in the same manner as our grounding-zone proximal

AN

pelletized facies may be useful in identifying paleo-grounding line locations. A coarse-skewed grain-size distribution with clear

M

evidence of winnowing (RGM) is indicative of strong currents,

ED

such as those associated with MCDW today. Its presence in cores

CE

PT

from earlier glacial cycles may be indicative of similar phenomena.

AC

Acknowledgments We thank the crew of the RV/IB Nathaniel B. Palmer and

the Antarctic Support Contract personnel for their efforts toward a successful expedition. We are grateful to Philip Bart and LSU students for shipboard assistance, Sarah Greenwood for help with shipboard data acquisition and post-cruise discussions, Adlai Fonseca for laboratory assistance, and Charlotte Sjunneskog and Steven Petrushak of the FSU Antarctic Research Facility for 42

ACCEPTED MANUSCRIPT geotechnical assistance. This work was funded by the National Science Foundation [NSF-PLR 1246353 to JBA].

7. References

IP

T

Alley, R.B., Blankenship, D.D., Bentley, C.R., Rooney, S.T., 1987.

CR

Till beneath Ice Stream B: 3. Till deformation: Evidence and implications. Journal of Geophysical Research 92,

US

8921-8929.

AN

Anandakrishnan, S., Catania, G.A., Alley, R.B., Horgan, H.J., 2007. Discovery of till deposition at the grounding line of

M

Whillans Ice Stream. Science, 315(5820), 1835-1838.

ED

Anderson, J. B., 1972. Nearshore glacial-marine deposition from modern sediments of the Weddell Sea. Nature, 240(104),

PT

189-192.

CE

Anderson, J. B., 1975. Ecology and distribution of foraminifera in

AC

the Weddell Sea of Antarctica. Micropaleontology, 69-96.

Anderson, J. B., 1999. Antarctic Marine Geology. Cambridge University Press.

Anderson, J. B., Kurtz, D. D., Domack, E. W., Balshaw, K. M., 1980. Glacial and glacial marine sediments of the Antarctic continental shelf. Journal of Geology 88, 399-414.

43

ACCEPTED MANUSCRIPT Anderson, J. B., Brake, C. F., Myers, N. C., 1984. Sedimentation on the Ross Sea continental shelf, Antarctica. Marine Geology 57, 295-333. Anderson, J. B., Kennedy, D. S., Smith, M. J., Domack, E. W.,

IP

T

1991. Sedimentary facies associated with Antarctica’s

floating ice masses. Geological Society of America Special

CR

Papers, 261, 1-26.

US

Anderson, J. B., Bartek, L. R., 1992. Cenozoic glacial history of the Ross Sea revealed by intermediate resolution seismic

AN

reflection data combined with drill site information. In

M

Kennett, J.P., and Warnke, D.A. (eds) The Antarctic

ED

Paleoenvironment: A Perspective on Global Change: Part One, 231-264.

PT

Anderson, J. B., Oakes-Fretwell, L., 2008. Geomorphology of the

CE

onset area of a paleo‐ice stream, Marguerite Bay, Antarctic

AC

Peninsula. Earth Surface Processes and Landforms 33(4), 503-512.

Anderson, J.B., Conway, H., Bart, P.J., Witus, A.E., Greenwood, S.L., McKay, R.M., Hall, B.L., Ackert, R.P., Licht, K., Jakobsson, M., Stone, J.O., 2014. Ross Sea paleo-ice sheet drainage and deglacial history during and since the LGM. Quaternary Science Reviews, 100, 31-54.

44

ACCEPTED MANUSCRIPT Arrigo, K. R., Van Dijken, G. L., 2003. Phytoplankton dynamics within 37 Antarctic coastal polynya systems. Journal of Geophysical Research: Oceans 108(C8). Balshaw-Biddle, K. M., 1981. Antarctic glacial chronology

IP

T

reflected in the Oligocene through Pliocene sedimentary section in the Ross Sea (Doctoral dissertation, Rice

CR

University).

US

Barnes, P. W., Lien, R., 1988. Icebergs rework shelf sediments to

AN

500 m off Antarctica. Geology, 16(12), 1130-1133. Barrett, P. J., 1975. 22. Textural characteristics of Cenozoic

ED

Antarctica.

M

preglacial and glacial sediments at Site 270, Ross Sea,

PT

Bart, P. J., Cone, A. N., 2012. Early stall of West Antarctic Ice Sheet advance on the eastern Ross Sea middle shelf

CE

followed by retreat at 27,500 14 Cyr BP. Palaeogeography,

AC

Palaeoclimatology, Palaeoecology 335, 52-60.

Bart, P. J., Coquereau, L., Warny, S., Majewski, W., 2016. In situ foraminifera in grounding zone diamict: a working hypothesis. Antarctic Science, 28(04), 313-321. Bart, P.J., Krogmeier, B.J., Bart, M.P. and Tulaczyk, S., 2017. The paradox of a long grounding during West Antarctic Ice Sheet retreat in Ross Sea. Scientific reports, 7(1), p.1262. 45

ACCEPTED MANUSCRIPT Batchelor, C.L., Dowdeswell, J.A., 2015. Ice-sheet grounding-zone wedges (GZWs) on high-latitude continental margins. Marine Geology 363, 65-92. Batchelor, C.L. and Dowdeswell, J.A., 2016. Lateral shear-

IP

T

moraines and lateral marginal-moraines of palaeo-ice

CR

streams. Quaternary Science Reviews, 151, 1-26.

Bentley M.J., et al., 2014. A community-based geological

US

reconstruction of Antarctic Ice Sheet deglaciation since the Last Glacial Maximum. Quaternary Science Reviews 100,

AN

1–9.

M

Bonaccorsi, R., Finocchiaro, F., Quaia, T. and Brambati, A., 2000.

ED

From Basal Till to Open-Water Sedimentation: Late Pleistocene to Holocene Changes in Core ANTA95-77C2

PT

(Glomar Challenger Basin, Ross Sea). In Proceedings of

CE

the Workshop Palaeoclimatic Reconstructions from Marine

AC

Sediments of the Ross Sea (Antarctica) and Southern Ocean: Maps (No. 4, p. 185). Terra Antarctica Publication.

Boulton, G.S., 1986. Push‐moraines and glacier‐contact fans in marine and terrestrial environments. Sedimentology, 33(5), 677-698. Christianson, K., Jacobel, R. W., Horgan, H. J., Alley, R. B., Anandakrishnan, S., Holland, D. M., DallaSanta, K. J., 46

ACCEPTED MANUSCRIPT 2016. Basal conditions at the grounding zone of Whillans Ice Stream, West Antarctica, from ice‐penetrating radar. Journal of Geophysical Research: Earth Surface 121, 19541983.

IP

T

Christoffersen, P., Tulaczyk, S., 2003. Signature of palaeo‐ice‐

freeze‐on. Boreas 32, 114-129.

CR

stream stagnation: Till consolidation induced by basal

US

Cowan, E. A., Christoffersen, P., Powell, R. D., 2012.

AN

Sedimentological signature of a deformable bed preserved beneath an ice stream in a Late Pleistocene glacial

M

sequence, Ross Sea, Antarctica. Journal of Sedimentary

ED

Research 82, 270-282.

Cowan, E.A., Christoffersen, P., Powell, R.D. and Talarico, F.M.,

PT

2014. Dynamics of the late Plio–Pleistocene West Antarctic

CE

Ice Sheet documented in subglacial diamictites, AND-1B

AC

drill core. Global and Planetary Change, 119, 56-70.

Cunningham, W. L., Leventer, A., Andrews, J. T., Jennings, A. E., Licht, K. J., 1999. Late Pleistocene-Holocene marine conditions in the Ross Sea, Antarctica: evidence from the diatom record. The Holocene 9, 129-139. De Angelis, H. and Skvarca, P., 2003. Glacier surge after ice shelf collapse. Science, 299(5612), 1560-1562. 47

ACCEPTED MANUSCRIPT Dinniman, M.S., Klinck, J.M. and Smith, W.O., 2003. Cross-shelf exchange in a model of the Ross Sea circulation and biogeochemistry. Deep Sea Research Part II: Topical Studies in Oceanography, 50(22), 3103-3120.

IP

T

Dinniman, M.S., Klinck, J.M. and Smith, W.O., 2011. A model

study of Circumpolar Deep Water on the West Antarctic

CR

Peninsula and Ross Sea continental shelves. Deep Sea

US

Research Part II: Topical Studies in Oceanography, 58(13), 1508-1523.

AN

Domack, E.W., Jacobson, E.A., Shipp, S., Anderson, J.B., 1999.

M

Late Pleistocene-Holocene retreat of the West Antarctic

ED

Ice-Sheet system in the Ross Sea: Part 2—Sedimentologic and stratigraphic signature. GSA Bull. 111, 1517-1536.

PT

Dowdeswell, J.A., Cofaigh, C.Ó. and Pudsey, C.J., 2004.

CE

Thickness and extent of the subglacial till layer beneath an

AC

Antarctic paleo–ice stream. Geology, 32(1), 13-16.

Dowdeswell, J. A., Ottesen, D., Evans, J., Cofaigh, C. Ó., Anderson, J. B., 2008. Submarine glacial landforms and rates of ice-stream collapse. Geology, 36(10), 819-822. Dowdeswell, J.A. and Fugelli, E.M.G., 2012. The seismic architecture and geometry of grounding-zone wedges

48

ACCEPTED MANUSCRIPT formed at the marine margins of past ice sheets. Geological Society of America Bulletin, 124(11-12), 1750-1761. Dunbar, R. B., Anderson, J. B., Domack, E. W., Jacobs, S. S., 1985. Oceanographic influences on sedimentation along the

CR

of the Antarctic Continental Shelf 291-312.

IP

T

Antarctic continental shelf in Jacobs, S.S. (ed) Oceanology

Dupont, T.K. and Alley, R.B., 2005. Assessment of the importance

US

of ice‐shelf buttressing to ice‐sheet flow. Geophysical

AN

Research Letters, 32(4).

Elverhoi, A, Roaldset, E., 1983. Glaciomarine sediments and

M

suspended particulate matter, Weddell Sea shelf,

ED

Antarctica: Polar Research, v. 1, p. 1-21.

PT

Evans, J., Pudsey, C.J., 2002. Sedimentation associated with Antarctic Peninsula ice shelves: implications for

CE

palaeoenvironmental reconstructions of glacimarine

AC

sediments. Journal of the Geological Society, London 159, 233-237.

Evans, J., Pudsey, C. J., ÓCofaigh, C., Morris, P., Domack, E., 2005. Late Quaternary glacial history, flow dynamics and sedimentation along the eastern margin of the Antarctic Peninsula Ice Sheet. Quaternary Science Reviews 24, 741774. 49

ACCEPTED MANUSCRIPT Evans, D.J.A., Phillips, E.R., Hiemstra, J.F. and Auton, C.A., 2006. Subglacial till: formation, sedimentary characteristics and classification. Earth-Science Reviews, 78(1), 115-176.

IP

significance of grain size parameters. Journal of

T

Folk, R. L., Ward, W. C., 1957. Brazos River bar: a study in the

CR

Sedimentary Research 27, 3-26.

Gales, J.A., Larter, R.D., Mitchell, N.C. and Dowdeswell, J.A.,

US

2013. Geomorphic signature of Antarctic submarine gullies: implications for continental slope processes.

AN

Marine Geology, 337, pp.112-124.

M

Gordon, A.L., Zambianchi, E., Orsi, A., Visbeck, M., Giulivi, C.F.,

ED

Whitworth, T. and Spezie, G., 2004. Energetic plumes over the western Ross Sea continental slope. Geophysical

PT

Research Letters, 31(21).

CE

Gow, A. J., Epstein, S., Sheehy, W., 1979. On the origin of

AC

stratified debris in ice cores from the bottom of the Antarctic ice sheet. Journal of Glaciology, 23(89), 185-192.

Graham, A.G., Larter, R.D., Gohl, K., Dowdeswell, J.A., Hillenbrand, C.D., Smith, J.A., Evans, J., Kuhn, G. and Deen, T., 2010. Flow and retreat of the Late Quaternary Pine Island‐Thwaites palaeo‐ice stream, West Antarctica. Journal of Geophysical Research: Earth Surface, 115(F3). 50

ACCEPTED MANUSCRIPT Greenwood, S.L., Gyllencreutz, R., Jakobsson, M. and Anderson, J.B., 2012. Ice-flow switching and East/West Antarctic Ice Sheet roles in glaciation of the western Ross Sea. Geological Society of America Bulletin, 124(11-12), 1736-

IP

T

1749. Halberstadt, A.R.W., Simkins, L.M., Greenwood, S.L., Anderson,

CR

J.B., 2016. Paleo-ice sheet behaviour: retreat scenarios and

US

changing controls in the Ross Sea, Antarctica. Cryosphere 10, 1003-1020.

AN

Hayes, D. E., Frakes, L. A., 1975. General synthesis, deep sea

M

drilling project leg 28. Initial Reports of the Deep Sea

ED

Drilling Project, 28, 919-942. Hillenbrand, C.D., Kuhn, G., Smith, J.A., Gohl, K., Graham, A.G.,

PT

Larter, R.D., Klages, J.P., Downey, R., Moreton, S.G.,

CE

Forwick, M. and Vaughan, D.G., 2013. Grounding-line

AC

retreat of the west Antarctic ice sheet from inner Pine island Bay. Geology, 41(1), 35-38.

Horgan, H.J., Christianson, K., Jacobel, R.W., Anandakrishnan, S. and Alley, R.B., 2013. Sediment deposition at the modern grounding zone of Whillans Ice Stream, West Antarctica. Geophysical Research Letters, 40(15), 3934-3939.

51

ACCEPTED MANUSCRIPT Howat, I. M., Domack, E. W., 2003. Reconstructions of western Ross Sea palaeo‐ice‐stream grounding zones from high‐ resolution acoustic stratigraphy. Boreas, 32, 56-75. Jacobs, S. S., Bauer, E. B., Bruchhausen, P. M., Gordon, A. L.,

IP

T

Root, T. F., Rosselot, F. L., 1974. Eltanin Report. Cruises 47–50, 1971; 52–55, 1972. Lamont-Doherty Geological

CR

Observatory of Columbia University Technical Report CU-

US

2-74.

Jacobs, S. S., Fairbanks, R. G., Horibe, Y., 1985. Origin and

AN

evolution of water masses near the Antarctic continental

M

margin: evidence from H218O/H216O ratios in seawater.

ED

Oceanology of the Antarctic Continental Shelf, 59-85. Jacobs, S. S., Jenkins, A., Giulivi, C. F. & Dutrieux, P., 2011.

PT

Stronger ocean circulation and increased melting under

CE

Pine Island Glacier ice shelf. Nature Geoscience 4, 519–

AC

523.

Jakobsson, M., Anderson, J.B., Nitsche, F.O., Gyllencreutz, R., Kirshner, A.E., Kirchner, N., O'Regan, M., Mohammad, R. and Eriksson, B., 2012. Ice sheet retreat dynamics inferred from glacial morphology of the central Pine Island Bay Trough, West Antarctica. Quaternary Science Reviews, 38, 1-10.

52

ACCEPTED MANUSCRIPT Jaeger, J.M., Nittrouer, C.A., DeMaster, D.J., Kelchner, C. and Dunbar, R.B., 1996. Lateral transport of settling particles in the Ross Sea and implications for the fate of biogenic material. Journal of Geophysical Research: Oceans,

IP

T

101(C8), 18479-18488. Kellogg, D.E. and Kellogg, T.B., 1986. Diatom biostratigraphy of

CR

sediment cores from beneath the Ross Ice Shelf.

US

Micropaleontology, 74-94.

Kellogg, T.B. and Kellogg, D.E., 1988. Antarctic cryogenic

AN

sediments: biotic and inorganic facies of ice shelf and

M

marine-based ice sheet environments. Palaeogeography,

ED

Palaeoclimatology, Palaeoecology 67(1-2), 51-74. Kennedy, D.S., Anderson, J.B., 1989. Glacial-marine

PT

sedimentation and Quaternary glacial history of Marguerite

CE

Bay, Antarctic Peninsula. Quaternary Research 31, 255-

AC

276.

Kennett, J. P., 1968. The Fauna of the Ross Sea: Part 6: Ecology and Distribution of Foraminifera. Department of Scientific and Industrial Research Bulletin 186, 1-47. Kilfeather, A.A., Cofaigh, C.Ó., Lloyd, J.M., Dowdeswell, J.A., Xu, S. and Moreton, S.G., 2011. Ice-stream retreat and iceshelf history in Marguerite Trough, Antarctic Peninsula: 53

ACCEPTED MANUSCRIPT Sedimentological and foraminiferal signatures. Geological Society of America Bulletin, 123(5-6), 997-1015. King, E.C., Hindmarsh, R.C. Stokes, C.R., 2009. Formation of mega-scale glacial lineations observed beneath a West

IP

T

Antarctic ice stream. Nature Geoscience, 2(8), 585-588.

CR

Kurtz, D.D. and Anderson, J.B., 1979, Recognition and sedimentologic description of recent debris flow deposits from

AN

Petrology, v. 49, p. 1159-1170.

US

the Ross and Weddell Seas, Antarctica: Journal of Sedimentary

Kurtz, D. D., & Bromwich, D. H., 1985. A recurring,

M

atmospherically forced polynya in Terra Nova Bay in

ED

Jacobs, S.S. (ed) Oceanology of the Antarctic Continental

PT

Shelf 177-201.

Larter, R.D. and Vanneste, L.E., 1995. Relict subglacial deltas on

CE

the Antarctic Peninsula outer shelf. Geology, 23(1), 33-36.

AC

Le Brocq, A.M., Ross, N., Griggs, J.A., Bingham, R.G., Corr, H.F., Ferraccioli, F., Jenkins, A., Jordan, T.A., Payne, A.J., Rippin, D.M. and Siegert, M.J., 2013. Evidence from ice shelves for channelized meltwater flow beneath the Antarctic Ice Sheet. Nature Geoscience, 6(11), 945-948.

54

ACCEPTED MANUSCRIPT Leckie, R. M., Webb, P. N., 1983. Late Oligocene–early Miocene glacial record of the Ross Sea, Antarctica: Evidence from DSDP site 270. Geology, 11(10), 578-582. Licht, K. J., Dunbar, N. W., Andrews, J. T., Jennings, A. E., 1999.

IP

T

Distinguishing subglacial till and glacial marine diamictons in the western Ross Sea, Antarctica: Implications for a last

CR

glacial maximum grounding line. GSA Bulletin 111, 91-

US

103.

Licht, K.J., Lederer, J.R., Swope, J., 2005. Provenance of LGM

AN

glacial till (sand fraction) across the Ross embayment,

ED

M

Antarctica. Quaternary Science

Licht, K., Palmer, E., Kramer, K., Swope, J., 2009. The polymict

PT

puzzle (abstract). Geological Society of America, Abstracts

CE

with Programs, v. 41, 59.

AC

Lowe, A. L., Anderson, J. B., 2003. Evidence for abundant subglacial meltwater beneath the paleo-ice sheet in Pine Island Bay, Antarctica. Journal of Glaciology, 49(164), 125-138. MacGregor, J.A., Anandakrishnan, S., Catania, G.A., Winebrenner, D.P., 2011. The grounding zone of the Ross

55

ACCEPTED MANUSCRIPT Ice Shelf, West Antarctica, from ice-penetrating radar. Journal of Glaciology, 57(205), 917-928. Majewski, W., Anderson, J. B., 2009. Holocene foraminiferal assemblages from Firth of Tay, Antarctic Peninsula:

IP

T

paleoclimate implications. Marine Micropaleontology 733,

CR

135-147.

McKay, R., Browne, G., Carter, L., Cowan, E., Dunbar, G.,

US

Krissek, L., Naish, T., Powell, R., Reed, J., Talarico, R., Wilch, T., 2009. The stratigraphic signature of the late

AN

Cenozoic Antarctic Ice Sheets in the Ross Embayment.

M

GSA Bulletin 121, 1537-1561.

ED

McKay, R.M., De Santis, L., and Kulhanek, D.K., 2017, Expedition 374 Scientific Prospectus: Ross Sea West

PT

Antarctic Ice Sheet History. International Ocean Discovery

CE

Program. https://doi.org/10.14379/iodp.sp.374.2017.

AC

McGlannan, A.J., Bart, P.J., Chow, J.M. and DeCesare, M., 2017. On the influence of post-LGM ice shelf loss and grounding zone sedimentation on West Antarctic ice sheet stability. Marine Geology. McMullen, K., Domack, E., Leventer, A., Olson, C., Dunbar, R., Brachfeld, S., 2006. Glacial morphology and sediment formation in the Mertz Trough,

56

ACCEPTED MANUSCRIPT East Antarctica. Palaeogeography, Palaeoclimatology, Palaeoecology, 231(1), 169-180. Menzies, J., 2000. Micromorphological analyses of microfabrics and microstructures indicative of deformation processes in

IP

T

glacial sediments. Geological Society, London, Special

CR

Publications 176(1), 245-257.

Menzies, J., van der Meer, J.J. and Rose, J., 2006. Till—as a

US

glacial ―tectomict‖, its internal architecture, and the development of a ―typing‖ method for till differentiation.

AN

Geomorphology 75(1), 172-200.

M

Milam, R. W., Anderson, J. B., 1981. Distribution and ecology of

ED

recent benthonic foraminifera of the Adelie-George V continental shelf and slope, Antarctica. Marine

PT

Micropaleontology 6(3), 297-325.

CE

Mosola, A. B., Anderson, J. B., 2006. Expansion and rapid retreat

AC

of the West Antarctic Ice Sheet in eastern Ross Sea: possible consequence of over-extended ice streams? Quaternary Science Reviews 25(17), 2177-2196.

Nitsche, F. O., Gohl, K., Larter, R. D., Hillenbrand, C. D., Kuhn, G., Smith, J. A., Jacobs, S., Anderson, J. B., Jakobsson, M. 2013. Paleo ice flow and subglacial meltwater dynamics in

57

ACCEPTED MANUSCRIPT Pine Island Bay, West Antarctica. The Cryosphere, 7(1), 249-262. Nygård, A., Sejrup, H.P., Haflidason, H., Lekens, W.A.H., Clark, C.D. and Bigg, G.R., 2007. Extreme sediment and ice

IP

T

discharge from marine-based ice streams: New evidence

CR

from the North Sea. Geology, 35(5), 395-398.

Ó Cofaigh, C. and Dowdeswell, J.A., 2001. Laminated sediments

US

in glacimarine environments: diagnostic criteria for their interpretation. Quaternary Science Reviews, 20(13), 1411-

AN

1436.

M

Ó Cofaigh, C., Dowdeswell, J.A., Allen, C.S., Hiemstra, J.F.,

ED

Pudsey, C.J., Evans, J. and Evans, D.J., 2005. Flow dynamics and till genesis associated with a marine-based

PT

Antarctic palaeo-ice stream. Quaternary Science Reviews,

CE

24(5), 709-740.

AC

Ó Cofaigh, C., Evans, J., Dowdeswell, J.A. and Larter, R.D., 2007. Till characteristics, genesis and transport beneath Antarctic paleo‐ice streams. Journal of Geophysical Research: Earth Surface, 112(F3). Ó Cofaigh, C., Dowdeswell, J.A., Evans, J. and Larter, R.D., 2008. Geological constraints on Antarctic palaeo‐ice‐stream

58

ACCEPTED MANUSCRIPT retreat. Earth Surface Processes and Landforms, 33(4), 513525. Orsi, A.H. and Wiederwohl, C.L., 2009. A recount of Ross Sea waters. Deep Sea Research Part II: Topical Studies in

IP

T

Oceanography 56(13), 778-795.

CR

Osterman, L.E. and Kellogg, T.B., 1979. Recent benthic

foraminiferal distributions from the Ross Sea, Antarctica;

US

relation to ecologic and oceanographic conditions. The

AN

Journal of Foraminiferal Research, 9(3), 250-269. Ottesen, D., Dowdeswell, J.A., Rise, L., 2005. Submarine

M

landforms and the reconstruction of fast-flowing ice

ED

streams within a large Quaternary ice sheet: The 2500-kmlong Norwegian-Svalbard margin (57–80 N). Geological

PT

Society of America Bulletin, 117(7-8), 1033-1050.

CE

Ottesen, D. and Dowdeswell, J.A., 2006. Assemblages of

AC

submarine landforms produced by tidewater glaciers in Svalbard. Journal of Geophysical Research: Earth Surface, 111(F1).

Passchier, S., Browne, G., Field, B., Fielding, C.R., Krissek, L.A., Panter, K., Pekar, S.F., and ANDRILL-SMS Science Team, 2011. Early and middle Miocene Antarctic glacial history from the sedimentary facies distribution in the AND-2A 59

ACCEPTED MANUSCRIPT drill hole, Ross Sea, Antarctica. GSA Bulletin 123, 23522365. Picco, P., Amici, L., Meloni, R., Langone, L. and Ravaioli, M., 1999. Temporal variability of currents in the Ross Sea

IP

T

(Antarctica). In Spezie G., Manzella G.M.R. (eds)

CR

Oceanography of the Ross Sea Antarctica, 103-117.

Pillsbury, R.D. and Jacobs, S.S., 1985. Preliminary observations

US

from long‐term current meter moorings near the Ross Ice Shelf, Antarctica. Oceanology of the Antarctic continental

AN

shelf, 87-107.

M

Powell, R. D., Dawber, M., Mcinnes, J. N., Pyne, A. R., 1996.

ED

Observations of the grounding-line area at a floating glacier

PT

terminus. Annals of Glaciology, 22(1), 217-223. Reid, D.E., 1989. Late Cenozoic glacial-marine, carbonate, and

CE

turbidite sedimentation in the northwestern Ross Sea,

AC

Antarctica [M.A. thesis]: Rice University, 179 p.

Reinardy, B.T., Hiemstra, J.F., Murray, T., Hillenbrand, C.D. and Larter, R.D., 2011. Till genesis at the bed of an Antarctic Peninsula palaeo‐ice stream as indicated by micromorphological analysis. Boreas, 40(3), pp.498-517.

60

ACCEPTED MANUSCRIPT Rignot, E., Mouginot, J. and Scheuchl, B., 2011. Antarctic grounding line mapping from differential satellite radar interferometry. Geophysical Research Letters, 38(10). Rodriguez, A. B., Anderson, J. B., 2004. Contourite origin for

IP

T

shelf and upper slope sand sheet, offshore Antarctica.

CR

Sedimentology, 51(4), 699-711.

Schoof, C., 2007. Ice sheet grounding line dynamics: Steady states,

US

stability, and hysteresis. Journal of Geophysical Research:

AN

Earth Surface, 112(F3).

Shepherd, A., D. Wingham, and Rignot, E., 2004. Warm ocean is

M

eroding West Antarctic Ice Sheet, Geophysical Research

ED

Letters 31, L23402

PT

Shipp, S., Anderson, J. B., Domack, E. W., 1999. Seismic signature of the Late Pleistocene fluctuation of the West

CE

Antarctic Ice Sheet system in Ross Sea: a new perspective,

AC

Part I. GSA Bulletin 111, 1486-1516.

Shipp, S.S., Wellner, J.S., Anderson, J.B., 2002. Retreat signature of a polar ice stream: sub-glacial geomorphic features and sediments from the Ross Sea, Antarctica. Geological Society, London, Special Publications, 203(1), 277-304. Simkins, L.M., Anderson, J.B. and Greenwood, S.L., 2016. Glacial landform assemblage reveals complex retreat of grounded 61

ACCEPTED MANUSCRIPT ice in the Ross Sea, Antarctica. Geological Society, London, Memoirs, 46(1), 353-356. Simkins, L.M., Anderson, J.B., Greenwood, S.L., Gonnermann, H.M., Prothro, L.O., Halberstadt, A.R.W., Stearns, L.A.,

IP

T

Pollard, D. and DeConto, R.M., 2017. Anatomy of a

meltwater drainage system beneath the ancestral East

CR

Antarctic ice sheet. Nature Geoscience, DOI:

US

10.1038/NGEO3012.

Singer, J.K. and Anderson, J.B., 1984. Use of total grain-size

AN

distributions to define bed erosion and transport for poorly

M

sorted sediment undergoing simulated bioturbation. Marine

ED

Geology, 57(1-4), 335-359. Smith, J.A., Andersen, T.J., Shortt, M., Gaffney, A.M., Truffer,

PT

M., Stanton, T.P., Bindschadler, R., Dutrieux, P., Jenkins,

CE

A., Hillenbrand, C.D. and Ehrmann, W., 2017. Sub-ice-

AC

shelf sediments record history of twentieth-century retreat of Pine Island Glacier. Nature 541(7635): 77-80.

Taviani, M., Reid, D. E., Anderson, J. B., 1993. Skeletal and isotopic composition and paleoclimatic significance of late Pleistocene carbonates, Ross Sea, Antarctica. J. Sed. Res. 63, 84-90.

62

ACCEPTED MANUSCRIPT Thoma, M., A. Jenkins, D. Holland, Jacobs, S., 2008. Modelling Circumpolar Deep Water intrusions on the Amundsen Sea continental shelf, Antarctica, Geophysical Research Letters 35, L18602

IP

T

Todd, B.J., Valentine, P.C., Longva, O., Shaw, J., 2007. Glacial landforms on German Bank, Scotian Shelf: evidence for

CR

Late Wisconsinan ice‐sheet dynamics and implications for

US

the formation of De Geer moraines. Boreas, 36(2), 148169.

AN

Van der Meer, J.J., Menzies, J. and Rose, J., 2003. Subglacial till:

M

the deforming glacier bed. Quaternary Science Reviews,

ED

22(15), 1659-1685.

Vieli, A. and Nick, F.M., 2011. Understanding and modelling rapid

PT

dynamic changes of tidewater outlet glaciers: issues and

CE

implications. Surveys in Geophysics, 32(4-5), pp.437-458.

AC

Wellner, J.S., Heroy, D.C. and Anderson, J.B., 2006. The death mask of the Antarctic ice sheet: comparison of glacial geomorphic features across the continental shelf. Geomorphology, 75(1), 157-171. Witus, A.E. Branecky, C.M., Anderson, J.B., Szczucinski, W., Schroeder, D.M., Blankenship, D.D., Jakobsson, M., 2014. Meltwater intensive glacial retreat in polar environments 63

ACCEPTED MANUSCRIPT and investigation of associated sediments: example from Pine Island Bay, West Antarctica. Quaternary Science Reviews 85, 99-118. Yokoyama, Y., Anderson, J.B., Yamane, M., Simkins, L.M.,

IP

T

Miyairi, Y., Yamazaki, T., Koizumi, M., Suga, H.,

Kusahara, K., Prothro, L., Hasumi, H., Southon, F.R.,

CR

Ohkouchi, N., 2016. Widespread collapse of the Ross Ice

US

Shelf during the late Holocene. Proceedings of the

AC

CE

PT

ED

M

AN

National Academy of Sciences 113, 2354-2359.

64

ACCEPTED MANUSCRIPT Figure captions: Figure 1: Location map of the Ross Sea, Antarctica with Last Glacial Maximum flow lines and grounded ice extent (modified from Halberstadt et al., 2016). Cores from cruise NBP1502A are

T

shown, as well as selected DSDP Leg 28 sites and all legacy cores

IP

stored at NSF’s Antarctic Research Facility. Cores discussed in

CR

Figures , 5, 6, and 7 are labeled.

US

Figure 2: (a) Conceptual diagram of a grounding-zone wedge (GZW) and proglacial environment, with associated glacial and

AN

sedimentary processes. Definitions of terms for buoyancy

M

equation: Hi = ice thickness, Hw = water depth, ρi = density of ice

ED

(917 kg m-3), ρw = density of seawater (~1025 kg m-3—may vary). Terrigenous input from meltwater plumes (level in water column

PT

unknown) is observed as far as 250 km from subglacial meltwater

CE

channels in the Ross Sea. (b) Formation of till pellets. (c) Deposition of basal meltout debris (limited to within 1.2 km of the

AC

grounding line) and debris flows (restricted to foreset length). (d) Open marine sedimentation dominated by rainout of organic detritus. (e) Reworking of glacial and glacimarine sediments by marine currents on banktops and the shelf margin, facilitated by bioturbation or iceberg turbation. Figure 3: Example of coring transect demonstrating targeted coring along a grounding-zone wedge using both (a) multibeam 65

ACCEPTED MANUSCRIPT swath bathymetry and (b) CHIRP data. Core locations are shown in multibeam context in Figure 3a, seismic context in Figure 3b, and regional context in Figure 1. Figure 4: X-radiographs of key facies and features. (a) Facies 1,

IP

T

displaying massive texture and large scattered pebbles (dark/dense masses) within a fine-grained matrix. (b) Facies 3, showing

CR

massive texture and scattered pebbles, with some large, prominent

US

till pellets outlined. Overall mottled texture is a result of smaller pellets comprising much of the matrix material. (c) Facies 4 with

AN

massive structure and no visible coarse material. (d) Facies 4 as a

M

laminated lens within Facies 3. (e) Ash laminations just above the Facies 1/5 contact. Burrows within Facies 5 are marked where they

ED

disrupt the laminations. (f) Sharp contact between Facies 1 and 5,

PT

most evident in x-rays as a change in granule/pebble abundance. Note: Air pockets are a consequence of the Kasten core archiving

CE

process but have no effect on surrounding structures.

AC

Figure 5: Representive downcore sedimentary successions for (a) GZW topset, (b) GZW foreset/toe, and (c) core with meltwater deposits, showing grain-size mode, shear strength, water content, and magnetic susceptibility, as well as representative grain-size frequency distributions for each sample grouped according to facies (Wentworth grain-size classifications are denoted above). Multibeam bathymetry for each core location are displayed to the 66

ACCEPTED MANUSCRIPT right of grain-size distributions for geomorphic context. 100 kPa=1 kg cm-2. Figure 6: Representive downcore foraminiferal successions for (a) GZW topset, (b) GZW foreset/toe, and (c) core with meltwater

IP

T

deposits. Note different scales for graphs showing numbers of

specimens per gram of dry sediment. Abundances of planktonic

CR

(black), benthic (light gray), and seemingly reworked (line pattern)

US

foraminifera.

Figure 7: Core photographs and whole-unit particle size

AN

distributions demonstrating (a) typical homogenous Facies 2

M

(debris flow deposits), and (b) more variable example of Facies 2

ED

with a discrete layer of meltwater deposits (Facies 4). Contacts are typically sharp, as shown in Figure 7b. The particle size

PT

distributions represent samples from every 5 cm within the entire

CE

Facies 2 unit, a sufficient interval to capture much of the variability (if any) in grain size which is evident even in a 5-cm photograph.

AC

VPS=very poorly sorted, PS=poorly sorted, according to graphical standard deviation calculations set forth by Folk and Ward, 1957. Figure 8: Map showing distribution of cores that sampled meltwater deposits and channel locations. Locations of ash recovery are superimposed and offshore volcanic islands and seamounts are denoted, as well as currently active Mount Erebus on Ross Island. 67

ACCEPTED MANUSCRIPT Figure 9: Surface processes on the deglaciated Ross Sea continental shelf and slope. Areas with potential coarse residual glacial marine (RGM) sediments that might inhibit coring were avoided on NBP1502A. Locations for LGM grounding line extent

T

and shelf-edge gullies are from Halberstadt et al., 2016.

IP

Sedimentary zones from Taviani et al., 1993, Dunbar et al., 1985,

CR

Anderson, 1999, and this study. Diatomaceous mud thickness is interpreted from cruise reports of DF80, NBP9401, NBP9407,

US

NBP9501, NBP9801, NBP9902, and NBP1502A cores. Exact

AN

values for diatomaceous open marine mud thickness should be regarded with caution, as many of the reported values are from

ED

trends are still valid.

M

piston cores which tend to lose some surface material, but general

PT

Fig. 10: Example of modal components of open marine grain-size distribution. Photomicrographs of smear slides show (a) the fine

CE

mode is dominated by terrigenous silt and diatom fragments, and

AC

(b) the coarse mode is dominated by intact diatoms. Qtz=quartz, PPL=plane polarized light, XPL=cross polarized light. Plane polarized light best displays opaline diatom tests. Diatoms are not detectable in cross polarized light, which therefore allows an uninhibited view of sediment grains. Figure 11: Distance of cores with and without pellets from nearest landward GZW-crest. Stippled pattern represents cores that 68

ACCEPTED MANUSCRIPT contain the pelletized grounding-zone proximal sediments representative of pure basal melt, and hatched pattern represents cores that do not. The maximum extent of the basal debris meltout occurs within 1.2 km from the grounding line, and is compared to

IP

T

other constraints on basal debris extent. Figure 12: Correlation between Facies 3 thickness and GZW

CR

height suggests that the pelletized unit thickness can be used to

US

estimate the relative stability of a paleo-grounding line. Figure 13: Western Ross Sea (EAIS) retreat scenario (a)

AN

Grounded ice extends out to mid-shelf at the LGM; (b) Ice shelf

M

collapses during initial retreat, forming deep iceberg furrows. Ice

ED

begins to retreat in small steps in the upstream direction and onto prominent banks (Yokoyama et al., 2016); (c) Continuous retreat,

PT

forming small GZWs and moraines during short pauses. Ice shelf

CE

may reform during brief periods of stability; (d) Present-day deglaciated continental shelf, with hypothetical core locations.

AC

Eastern Ross Sea (WAIS) retreat scenario (e) Grounded ice extends to the outer continental shelf at the LGM creating shelfedge gullies; (f) Ice remains stable for long periods of time, forming large composite GZWs wedges; (g) Ice lifts off the bed and retreats hundreds of kilometers upstream, leaving pristine MSGLs; (h) Present-day deglaciated continental shelf, with hypothetical core locations; (i) Expected sedimentary succession 69

ACCEPTED MANUSCRIPT for a core obtained near a paleo-grounding line, represented in (d) and (h); (j) Expected sedimentary succession for a core obtained on the topset of a GZW or within an MSGL field, represented in (d)

AC

CE

PT

ED

M

AN

US

CR

IP

T

and (h).

70

US

CR

IP

T

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

M

AN

Fig. 1

71

CR

IP

T

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

M

AN

US

Fig. 2

72

AN

US

CR

IP

T

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

M

Fig. 3

73

US

CR

IP

T

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

M

AN

Fig. 4

74

AC

CE

Fig. 5

PT

ED

M

AN

US

CR

IP

T

ACCEPTED MANUSCRIPT

75

AC

CE

PT

ED

M

AN

US

CR

IP

T

ACCEPTED MANUSCRIPT

Fig. 6

76

AC

CE

Fig. 7

PT

ED

M

AN

US

CR

IP

T

ACCEPTED MANUSCRIPT

77

US

CR

IP

T

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

M

AN

Fig. 8

78

US

CR

IP

T

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

M

AN

Fig. 9

79

ED

M

AN

US

CR

IP

T

ACCEPTED MANUSCRIPT

AC

CE

PT

Fig. 10

80

US

CR

IP

T

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

M

AN

Fig. 11

81

M

AN

US

CR

IP

T

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

Fig. 12

82

AC

CE

PT

ED

M

AN

US

CR

IP

T

ACCEPTED MANUSCRIPT

Fig. 13

83

ACCEPTED MANUSCRIPT Table 1: Ross Sea sediment facies characteristics. Diatom and biogenic silica data is from Cunningham et al. (1999) using core NBP9501 KC39 (same location as core NBP1502A KC48). Open marine and residual glacimarine (RGM) facies descriptions are from Anderson et al. (1980, 1984). Physical description

Geotechnical properties

Biogenic properties

Facies 1

Diamicton with sandy silt matrix. Very dark gray [5Y 3/1]; massive; occasional till pellets; pebbles common, very poor matrix grain size sorting [SD >2]

Homogeneous magnetic susceptibility downcore; water content ~15–30%; shear strength ~10–60 kPa

Calcareous benthic forams, some planktic, abundant reworked, altered foraminifera; 0–4% biogenic silica

Subglacial, on topset of grounding-zone wedge or in lineation field

Till

Facies 2

Diamicton with sandy silt matrix. Very dark gray [5Y 3/1]; massive; occasional till pellets; pebbles common, very poor matrix grain size sorting [SD >2]

Homogeneous magnetic susceptibility downcore; water content ~15–30%; shear strength ~5–40 kPa

Calcareous benthic forams, few planktic, some reworked; 10– 30% biogenic silica

Proglacial, on foreset of grounding-zone wedge

Grounding zone proximal, debris flow

Facies 3

Pelletized diamicton with sandy silt matrix. Dark olive gray [5Y 3/2]; massive or weakly stratified, occasional bioturbation; abundant till pellets; pebbles common, very poor matrix grain size sorting [SD 2–2.5]

Variable magnetic susceptibility downcore; water content ~30–40%; shear strength ~0–20 kPa

Calcareous benthic forams, some planktic; 15–30% biogenic silica, few clearly reworked, altered foraminifera

Proglacial, near grounding line

Grounding zone proximal, basal debris meltout

Facies 4

Dark gray [5Y 4/1] silt with ~10 μm mode; massive or weakly laminated; sparse to no pebbles; poor grain size sorting [SD 1.2– 1.8]

Homogeneous magnetic susceptibility downcore; water content ~40–50%; shear strength ~0–10 kPa

If present, minute calcareous benthic forams and few planktic

Proglacial, seaward of known subglacial meltwater channels

Meltwater deposit

Facies 5

Sandy silt. Olive [5Y 4/3]; massive, bioturbated structure; sparse to no pebbles; occasional sandy, gravelly lenses; matrix with poor grain size sorting [SD 1.5–2]. Fine-skewed distribution

Variable magnetic susceptibility downcore; water content ~30–70%; shear strength ~0–10 kPa

Agglutinated benthic forams, low diversity; ≥25% biogenic silica. 10-30% pristine diatoms and abundant fragments

Draping upper unit, ubiquitous

Open marine

Facies 6

Coarse-skewed grain-size distribution with truncation at ~125 μm indicating winnowing of fine material. Coarse material is primarily ice-rafted lithics but may also be bioclastic carbonates. Crudely to well-stratified

Not measured—Coring not successful.

Primarily calcareous benthic foraminifera; few, if any, diatoms

Continental shelf edge, banktop

Residual Glacimarine (RGM)

IP

CR

US

AN

M

ED

PT

CE

AC

84

Geomorphic context

T

Lithology

Interpretation

ACCEPTED MANUSCRIPT Highlights

 

AC

CE

PT

ED

M

AN

US

CR

IP



Multi-proxy approach distinguishes subglacial and iceproximal facies Meltout of ice shelf basal debris is highly restricted in extent Foraminifera composition demonstrates ice proximity and marine current energy Sedimentary successions reflect ice-sheet retreat style

T



85