Late Quaternary sediment facies in Prydz Bay, East ...

Antarctic Science 10 (3):236-246 (1998)

Late Quaternary sediment facies in Prydz Bay, East Antarctica and their relationship to glacial advance onto the continental EUGENE DOMACK’, PHIL O’BRIEN2, PETER HARRIS3,FIONA TAYLOR3*‘,PATRICK G. QUILTY4s3, LAURA DE SANTlS5and BENJAMIN RAKER6 ‘Department of Geology, Hamilton College, Clinton, NY 13323, USA ‘Australian Geological Survey Organization (AGSO), GPO Box 3 78, Canberra, ACT 2601, Australia ’Antarctic Cooperative Research Centre (CRC), GPO Box 252-80. Hobart, TAS 7001. Australia ‘Australian Antarctic Division, Channel Highway, Kingston, TAS 7050, Australia ’Osservatorio Geofisico Sperimentale, PO Box 201 I , 34016 Opicina (TS), Italy ‘Department of Earth and Environmental Sciences, Wesleyan University, Middletown. CT 06459. USA

Abstract: A marine survey in Prydz Bay, provides an unparalleled view of glacigenic and marine sedimentation across Prydz Channel and Amery Depression during the Late Quaternary. Gravity cores and a suite of eight radiocarbon dates indicate that the Late Wisconsin Glacial Maximum (LGM) was associated with grounding of a palaeo-ice shelf along the periphery of Prydz Channel. Deposition in front of the grounding line was dominated by ice-rafting. A granulated facies, containing angular clay and diamicton clasts, was produced by a combination of regelation freezing, near to the grounding line, and remelting of this basal debris in the sub-ice shelfsetting. Beneath these LGM marine deposits lie two key beds of diatom ooze that are distinct in size sorting and Pliocene diatoms. These “interstadial” units can be traced across most ofthe Prydz Channel, and are underlain by additional glacial marine units. Debris related to the Lambert Deep is distinct from detritus from eastern Prydz Bay and deposition ofthese two sources within the channel oscillated during the LGM. We suggest that coastal drainage systems contributed to a limited glaciation of the shelf during the LGM, rather than direct outflow via the Lambert /Amery system. It is proposed that shelf-wide glaciation is related to the duration ofglacial sea level lowstands rather than the absolute magnitude of eustatic fall during such episodes. Received 7 January 1998, accepted 4 June 1998

Key words: diatoms, glacial marine, Late Quaternary, Prydz Bay, radiocarbon Introduction

Before linkages can be made between Northern and Southern Hemisphere events during glacial-interglacial cycles, the nature and extent of Antarctic glacial maxima need to be resolved to the same level as our knowledge of the boreal ice sheets (Denton & Hughes 1979, Broecker & Denton 1989). This problem needs attention because palaeo-oceanographic studies (based upon deep sea sediments) are now focused upon the phasing ofglacial cycles in the northern and southern high latitudes (Charles eta[. 1996, Bard et a[. 1997). These studies have produced a terminology of “glacial” and “deglacial” cycles that is based upon isotopic/geochemical events, regardless of what is known about Antarctic ice volume change over short time scales. In other words, while these geochemical events are clearly oceanographic signals of climate change, their linkage to Antarctic glacial advance or retreat remains problematic. Sediment records from the Antarctic continental shelves are key pieces in the puzzle of Quaternary history of the Antarctic ice sheets (Anderson 1989). This study focuses upon one such piece of the puzzle, the continental shelf of Prydz Bay. The Lambert Glacier/Amery Ice Shelf complex drains into

Prydz Bay and is an important component of the Antarctic cryosphere, comprising some 14% of the ice drainage from the East Antarctic Ice Sheet (Fig. 1; Hambrey et al. 1991). Seismic surveys show that the bay is underlain by thick sequences of prograded sediment that were deposited in response to expansion of the East Antarctic ice sheet and attendant advances of the Lambert glacier to the shelf break (Stagg 1985, Cooperetal. 1991). Althoughglacialgrounding line moraines have been recognized and mapped based upon seismic and echo sounding records (Leitchenkov et al. 1994), the age and depositional history of these features and their association with sea-floor deposits is unknown, despite preliminary knowledge ofthe chronology of late PleistoceneHolocene deposits on the shelf (Domack et al. 199 1, Pushina et al. 1997). The goals of Leg 119 of the Ocean Drilling Program were to address glacial events which related to initial glaciation of the East Antarctic subcontinent (Cooper et al. 1991). The five sites drilled in Prydz Bay (Fig. I ; Barron & Larsen 1991), however, did not advance our knowledge of Late Pleistocene glacial events significantly. In this paper, we present the results of a cruise of the

236 Downloaded from https:/www.cambridge.org/core. University of Tasmania Library, on 21 Apr 2017 at 00:31:06, subject to the Cambridge Core terms of use, available at https:/www.cambridge.org/core/terms. https://doi.org/10.1017/S0954102098000339

SEDIMENT FACIES AND GLACIAL ADVANCE IN PRYDZ BAY

I

I

I

Australian icebreaker RSV Aurora Australis, in FebruaryMarch 1997. Gravity cores, GI-gun seismic reflection data, and sidescan sonar data were collected from a number of presumed grounding line features in an attempt to gain a better understanding of the late Quaternary glacial history of Prydz Bay and the character of associated shelf deposits (Fig. 2). We describe the succession of facies observed in the cores and relate them to depositional environments, sedimentation patterns, and radiocarbon analyses. In so doing, we build upon earlier work which has contributed to our understandingof the Antarcticglacial marine environment (Carey&Ahmad 1961, Barretteral. 1991, Alleyetal. 1989, Andersonetal. 1980,1991, Drewry&Cooper 1981,Domack 1982, Kellogg & Kellogg 1988, Futterer & Melles 1990, Hambrey et al. 1991, Pudsey et al. 1994, Licht et al. 1996).

237

Fig. 1. Pattern of glacial drainage into Prydz Bay from the surrounding ice sheet and Lambert Glacier. Insert shows location of detailed study area (Fig. 2). Site locations 739-743 for Ocean Drilling Program Leg 119 are also shown. Figure modified after Hambrey et al. (1991).

Methods

Gravity cores 9-22 (Fig. 2) were collected during Cruise 186 of RSV Aurora Australis in February-March 1997 while cores KROCK 22 and 24 (Fig. 2) were collected during AGSO cruise 901 of Aurora Australis in 1992-93. A 3 m, 10 cm (internal diameter) steel corer was used with a PVC liner. Liners were extruded on deck and passed through a Bartington MS-2C core sensor for magnetic susceptibility, which was measured every 2.5 cm. Cores were then split down their lengths, described and sampled. Digital photographs of all cores are archived with the cores at the Antarctic Research Centre (Universityof Tasmania). Samples were taken at selected intervals for smear slide analysis (diatoms) and microscopic examination. Sediment was wet sieved through a 63-micron sieve and dried residues were examined qualitativelyfor foraminifera. Approximately 10g

Downloaded from https:/www.cambridge.org/core. University of Tasmania Library, on 21 Apr 2017 at 00:31:06, subject to the Cambridge Core terms of use, available at https:/www.cambridge.org/core/terms. https://doi.org/10.1017/S0954102098000339

SEDIMENT FACIES AND GLACIAL ADVANCE IN PRYDZ BAY

Downloaded from https:/www.cambridge.org/core. University of Tasmania Library, on 21 Apr 2017 at 00:31:06, subject to the Cambridge Core terms of use, available at https:/www.cambridge.org/core/terms. https://doi.org/10.1017/S0954102098000339

239

EUGENE DOMACK el al.

Downloaded from https:/www.cambridge.org/core. University of Tasmania Library, on 21 Apr 2017 at 00:31:06, subject to the Cambridge Core terms of use, available at https:/www.cambridge.org/core/terms. https://doi.org/10.1017/S0954102098000339

5a

240

238

EUGENE DOMACK ef al.

Fig. 2. Location map and bathymetry of the continental shelf in western Prydz Bay. Contour interval I S 100 rn.Nunibcred core sites shown in black. Major grounding line positions shown as dotted and ticked lines as determined by seismic reflection, sidescan sonar, and sub-bottom profiles. Major bathymetric features include the Lambert Decp, Aniery Depression. I’rydz Channel. Nella Rim and Fram Bank. Grounding line “moraines” are named geographically and in progression from seaward to landward as follows: Eastern Prydz Channel 1 & 2 (EPC-I. EPC-2), Nclla Rim I (NR-I). Lainbert Deep 1 CYC 2 (LD-I. LD-2). and Western Prydz Channel I , 2, & 3 (WPC-I, WPC-2, WPC-3) l h e hachures mark the srecp. seaward slope of the grounding line “moraines”. Position of the front of the Amery Ice Shelf as obscrvcd in 1992 is indicated

ofdriedsedimentwasalso wet sieved from selected diamicton intervals and the coarse sand grains (> 500 microns) were examined for petrologic/mineralogic character. The percentage of clear quartz and iron stained quartz were the distinguishing differences among the samples. Samples for radiocarbon analysis were either kept refrigerated (for organic matter dates) or were wet sieved and picked for calcareous foraminifera. Samples were sent to the RAFTER isotope laboratory at Victoria University in Wellington, New Zealand, for accelerator mass-spectrometer analysis. All ages are reported based upon a Libby half-life and corrected for their 613Ccontent (Table I). No reservoir corrections are applied to the data at this time. During cruise 186, we also collected seismic profiles, 12 kHz echo-sounder records, and sidescan sonar images that revealed several sets ofgrounding line wedges (Fig. 2). Many of these features were previously recognized by Leitchenkov etal. (1994)andare believed to beassociated with stabilization of a submarine grounding line because of their gentle stossside ramp and steep lee (seaward) slopes. Very little evidence for internal foreset bedding was observed in these features and this, along with thier widespread distribution, argues against a “till delta” origin, similar to the model ofAlley el al. (1989). Sediment facies Sediment facies are designated by five general textural types and further sub-divided by association with four groups of colours (using the Munsell scheme, Figs 3 & 4).

Siliceous mud and ooze, SMO

Siliceous mud and ooze (SMO) form three distinct stratigraphical units designated here as SMO- 1 and older SMO-2 and SMO-3. SMO units are defined by an abundance of diatoms, ranging from 20% to over 90% of the total sediment mass. SMO-I is the typical diatomaceous facies of the Antarctic continental shelf (Dunbar et al 1985) and occurs across the entire area of study, within and beyond the Prydz Channel (Figs 3 & 4). It is variable in thickness (a few centimetres to > 150 cm) and has colours of light olive and greyish olive, to moderate olive brown. Radiocarbon ages from core KROCK 24 indicate that SMO-I represents deposition during the entire Holocene epoch, as radiocarbon ages on organic matter range from 25 10 * 65 to 12 680 i 110 yr (Table I, Fig. 4). In places, there is an abundant Table 1. AMS radiocarbon dates determined

Lab no.

Core and depth (cm)

613C

tincorrected Carbon age source’

R21977/2 R21977t3 R.2197714 R21751/7 R24049/1 R21308/18 R21308/19 R2.1308117

186GC20 110-120 186 GC22 26 5-31 5 186 GC22 26 5-31 5 186 GC22100-101 186 GC12 344-346 KROCK24 5-6 KROCK 24 100-101 KROCK 24 138-139

-07 07 03 -23 8 02 -24 3 -24 1 -26 6

33 5 9 0 i 4 4 0 20 230 i 120 20 770 -t 160 22 170-t 180 25 480 f 160

MT

G crasya N pachyderma POC

BY

2510i65 POC 8540 -t 85 POC 12 680 f 110 POC

*Organic carbon source: MF = mixed benthic and planktonic forams, ostracods; POC = particulate organic carbon, acid insoluble, BF= benthic foraminifera

Downloaded from https:/www.cambridge.org/core. University of Tasmania Library, on 21 Apr 2017 at 00:31:06, subject to the Cambridge Core terms of use, available at https:/www.cambridge.org/core/terms. https://doi.org/10.1017/S0954102098000339

SEDIMENT FACIES AND GLACIAL ADVANCE IN PRYDZ BAY agglutinated foraminiferal assemblage in SMO- 1. SMO-2 is a key bed found in an intermediate stratigraphic position. It is limited in lateral extent to the floor ofthe Prydz Channel (some 15 000 km2)and is notably thicker near the seaward end ofthe channel (core 22; Fig. 3). However, unlike the Holocene (SMO-1) unit, SMO-2 shows distinct sizesorting of diatom species. Smear slides of SMO-2 in outer Prydz Channel (core 22) demonstrate a dominance of large robust frustules, primarily those of the genus Eucarnpia, and a relative abundance of Pliocene diatoms, Rouxia spp. and Thalassiosira torokina. Farther into the channel (core 09) SMO-2 is composed primarily of diminutive diatoms such as Fragilariopsis curtu, fragments, and rare Rouxia spp. and T. torokina. The magnetic susceptibility signatures are distinctly higher for SMO-2 than SMO-1, and SMO-2 has colour values in the lighter hues, such as dusky yellow green and pale olive (Fig. 5). A single organic matter radiocarbon analysis on SMO-2 (core 22) provided an uncorrected age of 22 170 & 180 yr BP (Table I). A third SMO unit (SMO-3) is recognized in the base of core KROCK 22 and in core 20 (Fig. 3). Its diatom component appears to be identical to that of SMO-2. Diamicton, D Poorly sorted, terrigenous sediments (diamictons) are the most diverse of the sediment facies recognized in our cores (Figs 3 & 4). We have designated these as D-1 to D-5, based upon their magnetic susceptibility, provenance, and partly upon their stratigraphic order (Figs 3-5). D-1 is a thin, stratified, and widespread diamicton of glacial marine origin, found across most of the Prydz Channel and to the east. It lies

directly beneath SMO- 1 commonly with very thin,transitional silt interbeds. Its unique characteristics include high magnetic susceptibility (1 00-200 x CGS) and a well-preserved and abundant calcareous foraminiferal assemblage. Foraminifera include abundant Globocassidulina crassa, Cibicides refulgens, and Neogloboquadrina pachyderma. Monospecific samples from core 22 (26.5-3 1.5 cm) yielded AMS radiocarbon ages of 20 770 f 160 and 20 230 f 120 yr BP for N. pachyderrna and G. crassa respectively (Table I). D-2 is a distinctive, reddish-brown diamicton associated with the Lambert Deep region, landward ofthe western Prydz Channel grounding line 1 (WPC- 1;Figs 2-4). D-2 underlies a very thin SMO- 1 unit that, in places, is defonned into small lenses of SMO (core 18). This diamicton is quartzose, structureless, barren of fossils, and has a low MS signature (typically less than 100 x loe6CGS; Fig. 5). A gradational decrease in water content downcore is characteristic, best exemplified in cores 16 and 17 (Fig. 3). The uppermost portions of D-2 are easily deformed by shearing while lower portions are considerably more cohesive and resist deformation. Quartz grains consist of a mixture of iron stained and clear quartz (Fig. 5), the former of which is likely to be sourced from red beds of the Lambert Graben. The equivalent to D-2 on the eastern side of the Amery Depression is the dark grey to olive grey diamicton, D-3 (Figs 3-5). D-3 is limited to areas east of the eastern Prydz Channel 1 grounding line (EPC-1) where it is found directly below SMO- 1. Hence, it is pre-Holocene in age or latest Pleistocene (Fig. 4). It has a very high MS (250-350 x lop6 CGS), is rich in Iithic grains, and generally barren of fossils (Fig. 4). Quartz grains are exclusively clear, not iron stained (Fig. 5). D-3 is found in cores 10-14, 21 and KROCK 24

0 5 VR

SV, 5GV L N

i n VR

RED BROWN

BROWN

OLIVE GREY

nu€

OLIVE GREEN

24 I

1ov

PALE VALUE

Fig. 5. Bivariant plot of mean magnetic susceptibility (MS) versus Munsell hue of diamicton units (A = 01-5), siliceous mud/ ooze (Q = SMO-1 and = SMO-2), and silty clays (0 and 0). Each symbol represents a unit from an individual core and is based upon 15 to 50 measurements for each mean value of MS. Note distinct MS and colour signature of each unit. Lambert Deep diamictons (D-2) are characterized by red to brown hues and MS values less than 100, while diamictons from Amery Depression and Prydz Channel (D-1,-3 and -5) are grey to olive grey and have high to moderate MS values. The two SMO units are distinct in both MS (SMO-1, mean < 10; SMO-2, mean = 20-40) and in the lightness of the value in the olive green chroma (SMO-2 has a colour of pale olive or dusky yellow green). Insert illustrates variation in quartz grain types among the

various diamicton units.

Downloaded from https:/www.cambridge.org/core. University of Tasmania Library, on 21 Apr 2017 at 00:31:06, subject to the Cambridge Core terms of use, available at https:/www.cambridge.org/core/terms. https://doi.org/10.1017/S0954102098000339

242

EUGENE DOMACK et al.

(Figs 2 & 4). D-3 is structureless, compact throughout, and does not demonstrate the gradation in water content characteristic of D-2. Diamictons older than D-2 and D-3 are limited to two units recovered in the lowermost portions of cores 20 and 22 (Fig. 3). Both of these diamictons are distinctive in their colour and MS. At the base of core 20, is a structureless, reddish-brown diamicton that has a colour similar to D-2 but has a higher MS signature (>I00 x lo-(‘CGS). There is a mixture of clear and iron-stained quartz within this diamicton (Fig. 5). Because of this and its low stratigraphic position, it is designed as D-4. Within the base of core 22, a large clast of indurated diamicton (D-5) is found within an iceberg turbate deposit (see later section). It is not in situ but is eroded from arelatively older stratigraphic unit present in the vicinity of core 22. It has a colour and MS similar to D-3, but clearly lies below it. Granulated facies

In many ofthe cores, a peculiar sediment is found that can best be described as a granulated texture. It has previously been recognized in Antarctic marine sediments as a “brecciated” (O’Brien & Harris 1996) or a “pelletized” texture (Domack etal. 1996). The characteristicsofthesegranulatedsediments are as follows: a) grain supported texture, matrix of sandy mud, b) granulesof silty clay, and/or diamicton matrix are lightly consolidated,

c) granules are angular and very loosely compacted, and, d) units lack observable grading but are size sorted. Colours of the granulated beds vary from reddish brown to olive grey, reflecting a difference in source similar to the diamicton units (Figs 3 & 4). The granulated facies is found either interbedded with silty clay and SMO in the Prydz Channel, or above massive diamictons (D-2 and D-3) that lie outside of the Prydz Channel. Similar textures within cores from the Ross Sea were interpreted as a transitional ice shelf-

Table 11. Foraminifera identified from 186GC20 110-120 cm ~~~~~~~

~~~

~~

~

~

~

Benthic Foraminifera Cassidulina terefis - most abundant taxa in the sample Ehrenbergina glabra Globocassidulina biora Asbononion echolsi Epistominella exigua Eponides tumidulus Nonionella iridea Melonis afjnis Stainforlhia complanata Texfularia catenala (agglutinated form) Planktonic Foraminifera Neogloboquudrinu pachyderma

rafted facies deposited beneath basal debris zones and above till units (Domack er ul. 1996). The intcrnal texture of the granules is similar to mud clots described from basal debris zones of the Greenland and West Antarctic Ice sheets (Gow et al 1979, Gow & Meese 1996). Silty clay

Silty clay is a very common sediment type within the Prydz Channel and is less cunimon landward of the major moraines that rim the channel (Figs 2-4). It occurs in a variety of colours from reddish brown to olive grey, to moderate yellowish brown. It has a moderate and variable MS, reflecting a mixture of detrital magnetite within the mud. Grey muds have a higher MS than reddish brown or yellowish brown muds (Fig. 5 ) . Characteristic of these muds is a low content of coarse gravel grains and poorly sorted sand (1-5%). The muds contain burrow structures, are mottled, and have gradational contacts with other units. Some of the silty clays from Prydz Channel contain isolated mud clots of a colour distinct from that of the surrounding matrix (Fig. 5). For instance, olive grey muds often have reddish brown mud clots within them and reddish brown muds may have olive grey clots within them, similar to “drop clot” textures reported from Late Quaternary sub-glacial lacustrine sediments in central New York state (Ridge et al. 1991). In places, foraminifera are abundant in the silty clay and consist of a diminutive fauna, including planktonics (N. pachyderma) and very well-preserved Ehrenberginu gfubra (Table 11). A sample of mixed benthic foraminifera from core 20 provided an uncorrected radiocarbon age of 33 590 5 440 yr BP (Table I). Sorted sand

Sorted sand is an uncommon sediment type in the Pryclz Channel and is found as a single, 25 cm thick, normally graded unit in core 09 (Fig. 3). This sand is well sorted, free of mud and grades from coarse to medium grained. It has a high MS (>300 x lO.‘CGS) and exhibits a sharp lower contact and a gradational upper contact with silty clay. Sedimentary structures

Most sedimentary structures within the sections are limited to mottling and/or burrow features in the silty clays and SMO: diamictons are structureless, as the granulated facies appear to be. Laminated beds ofsilty clay are rare and discontinuous. Themost noticeable structures are deformation features found in core 22. Below 125 cm in core 22, small-scale folds and faults are well preserved in a section that contains olive grey silty clay and convoluted beds of SMO-2 and reddish-brown silty clay. Deformation fabric has the appearance of horizontal foliation near the base of the core where a large clast of diamicton (D-5) is surrounded by foliated, sandy silty clay

Downloaded from https:/www.cambridge.org/core. University of Tasmania Library, on 21 Apr 2017 at 00:31:06, subject to the Cambridge Core terms of use, available at https:/www.cambridge.org/core/terms. https://doi.org/10.1017/S0954102098000339

SEDIMENT FACIES AND GLACIAL ADVANCE IN PRYDZ BAY (Fig. 2). Micro-faults are clearly preserved with reverse offset. The structures are compatible with iceberg turbation and are similar to features described by Woodworth-Lynas & Guigne (1990).

Microfossils Siliceousmicrofossils(diatoms)occur in all ofthe sedimentary units although the occurrence varies from very rare (most diamictons and silty clays) to very abundant (SMO). Calcareous microfaunas are less common and are restricted to two stratigraphical intervals. D- 1 contains a well-preserved foraminifera assemblage (G. crassa, C. refulgens, and N pachyderma) best exemplified by samples from core 22. This microfossil assemblage also contains abundant echinoid spines and continues into the lower silty clay for some i0-20 cm. Cores 12 and 20 are the only others with abundant calcareous microfauna. In core 20, the uppermost silty clay (at-1 m depth)contains acalcareousmicrofaunacharacterized by abundant ostracods (articulated valves) and a wellpreserved foraminiferal assemblage consisting dominantly of Ehrenbergina glabra, N . pachyderma, and Globocassidulina spp. (Table 11). A mixed sample of these carbonate microfossils from 110-120 cm yielded an AMS date of 33 590 f 440 yr BP (Fig. 2; Table I). Below 330 cm in core 12, a granulated facies and silty clay contain a unique foraminiferal assemblage characterized by abundant Patellina corrugata and SpiriNina spp., genera not common within modem faunas in theregion (Quilty 1985, Milam &Anderson 1981). A sample consisting of both of these species from 344-346 cm yielded an AMS radiocarbon age of25 480 i 160 yr BP (Fig. 2 &Table I). The diamicton (D-3) above 330 cm in core 12 is essentially barren of microfossils. Discussion and interpretation The most significant contribution of this study is the observation that facies changes observed in gravity cores are coincident with observed grounding line positions across the shelf (Figs 3-5). Structureless to massive diamictons (D-I and D-3) are found only landward of the grounding lines WPC-1 and EPC-1, thus indicating till deposition while the grounding line was occupying these positions around the Prydz Channel (Fig. 6). While till deposition was taking place beneath the ice sheet (ice plain), sub-ice shelf deposition via ice rafting was also occurring within the channel (Fig. 6). This led to a mixture of granulated sediment, representing undermelt of the basal debris zone (direct vertical settling) and lateral advection of suspended silt and clay. Very little of the material along the grounding line wedge seems to be of sediment gravity flow origin; apossible debris flow diamicton occurs in core 18 and a single, thin turbidite occurs in core 09 (Fig. 3). Complex interbedding of the granulated facies and silty clay, as seen in the cores from Prydz Channel (Figs 3 & 4), is

243

explained by fluctuating ice discharge and/or flow directions. Two distinct sources for sediment are recognized and these could have fluctuated as ice discharge was dominated by flow from the Lambert Deep (LD, red sediment) or by flow from the east, Eastern Prydz Channel (EPC, olive grey sediment). The basal debris zone in Fig. 6 is illustrated in order to show three-dimensional changes in debris source, but it may be that different levels of debris in the ice also reflect different sources. It is interesting to note that the granulated facies along WPC- 1 is clearly of LD provenance whereas the silty clay is of EPC provenance (Figs 3-5). This clearly indicates the separation of coarse, vertically settling, detritus from the advected fines that are transported underthe influence ofsubice shelfcurrents. Silty clays ofLD provenance only dominate the outermost portions of the channel, as at the site of core 22 (Fig. 3). This suggests aclockwise palaeo-circulation beneath the ice shelf in Prydz Channel. While the red mud was released near to the junction of the Lambert Deep and Prydz Channel, the inferred circulation would have advected this fine material out toward the north, the site of core 22 (Fig. 2). The origin of the mud clots or granules can be related to regelation processes taking place upstream of the grounding line. Regelation of a mud-rich slurry at the base of the ice could produce clots of frozen clay that would be incorporated into the basal debris zone. This process has been observed at the base of debris zones in polar ice sheets (Gow & Meese 1996). We believe that deposition of granulated and silty clay facies within the Prydz Channel is dependent upon grounding of an ice sheethce shelf system along the periphery of the channel. As the grounding line receded from LGM positions, at WPC-1 and EPC-1, siliciclastic sediment supply to the channel was cut off, as the grounding line (sediment source) was now divorced from the Prydz Channel deposystem, because the grounding zone deepens toward the continent. Recession of the calving line across the channel led to the development of fossiliferous diamicton (D- 1) and eventually toopen marine SMO-I, which reflects seasonally open marine environments similar to the present. D-1 is essentially an iceberg rafted sediment; its variable magnetic susceptibility and fossil content reflect a variety of sources and depositional settings most consistent with concentrated iceberg rafting along the calving line of a receding ice shelf (Domack et al. 1996, Domack & Harris 1998). Some contribution from eolian sources (Dunbar et al. 1989, Barrett et af. 199 1) could also have occurred. The age for the transition from grounded ice to open marine environments east of EPC- 1 is constrained to about 11 000 radiocarbon yr BP (O’Brien & Harris 1996; Fig. 3). Hence, the transition into SMO-I across the entire study area is believed to be related to recession following the Late Wisconsinan glacial maximum or LGM. The age for D- 1 is further constrained to the LGM by radiocarbon ages of 20.2-20.7 ka yr BP on sandy mud foundjust below D- 1 in core 22 (Table I &Fig. 3). The close agreement of these two ages is a good indication that the assemblages have not been

Downloaded from https:/www.cambridge.org/core. University of Tasmania Library, on 21 Apr 2017 at 00:31:06, subject to the Cambridge Core terms of use, available at https:/www.cambridge.org/core/terms. https://doi.org/10.1017/S0954102098000339

244

EUGENE DOMACK ef al.

Fig. 6. Cartoon of sub-ice shelf sedimentation within the Prydz Channel and Amery Depression. Undermclt beneath the basal debris zone produces a granulated facies from direct settling of the coarse fraction (mud clots) and a distal, silty clay from the lateral advection of suspended fines. Debris sources may vary across the Amery Depression from debris derived from the Lambert Deep drainage (reddish brown sediments) and/or from the eastern side of the Arnery Depression (grey sediments). Debris sources are also distinguished based upon local (low level basal debris) and distal (high level basal debris) sources along a single flow path. Advection of diatoms into the Pry& Channel I s also shown. Variations in facies boundaries between time T,, and T, are suggestive of varying flow velociities or ice volume drainage from the south-east or south-west. Downward pointing triangle = coarse debris from WPC source, diamond = coarse debris from EPC source, circles = suspended fines.

reworked significantly. The duration of ice shelf conditions across the Prydz Channel is difficult to determine except in relative terms. SMO-2 is a key bed that is restricted to the Prydz Channel and that we interpret as indicating a suppression of siliciclastic input into the channel at a time prior to the LGM (about 22 000 yr BP; Fig. 2 ) . Deposition of SMO-2 is therefore related to a retreat of the grounding line from W C - 1 and EPC- 1 positions. Subsequent readvanceat the LGM lowstand removed SMO-2 in areas landward of WPC- 1 and EPC- 1. Deposition of SMO-2 was also contemporaneous with a lower stand of sea level. This is because it was synchronous with iceberg or ice-shelf turbation of the substrate at the site of core 22, which today is well below the depth of iceberg grounding (at 660 m; Barnes & Lien 1988, Keys era[. 1990). The distribution and character of SMO-2 also suggest a subice shelf setting for its deposition. Its diatom flora is not the same as modem SMO- I assemblages and it has a lower C/Sibl0 ratio than SMO- 1. SMO-2 is dominated at any one site by diatoms of a given size. For instance, in core 22, large and

robust forms of Eucampia and reworked Pliocene fornis dominate while in the inner reaches of Prydz Channel, small fragments and F. curta dominate the assemblage. T h i s suggests that significant current sorting of the diatoms has taken place and that the source of reworked frustules is the outer reaches of the Prydz Channel. In fact, we observe an abundance of Pliocene diatoms in near-surface sediments of the Prydz Trough Mouth Fan. Circulation within the Prydz Channel at this time was likely in the form of a clockwise gyre, as previously suggested for LGM ice shelf conditions and as is presently found in Prydz Bay (Nunes Vaz & Lennon 1996). SMO-3 likelyrepresents asimilar palaeoenvironmental setting as SMO-2 but at an earlier time, prior to D-4. In addition, there wasan earlierperiodofice-sheetgrounding prior to SMO-2 along the periphery of Prydz Channel, as shown by granulated and silty clay facies that are older than SMO-2. Only in core 20 is a diamicton (D-4) similar in character to basal till recognized within the main axis of the channel itself. This unit lies at the base of the ice-shelf facies and predates both SMO-2 and SMO-3 deposits. The age of

Downloaded from https:/www.cambridge.org/core. University of Tasmania Library, on 21 Apr 2017 at 00:31:06, subject to the Cambridge Core terms of use, available at https:/www.cambridge.org/core/terms. https://doi.org/10.1017/S0954102098000339

SEDIMENT FACIES AND GLACIAL ADVANCE IN PRYDZ BAY D-4 is older than 33.6 ka yr BP (Fig. 3) and since it reflects grounding of ice within the channel, it is the one till most likely correlative to shelf-edge grounding ofthe East Antarctic Ice Sheet across Prydz Bay and sediment transport to the continental slope. Therefore, the LGM (isotope stage 2) in this portion of Antarctica was not the most extensive of the Late Quaternary glaciations and the term “maximum” is a regional misnomer. The question that remains unanswered is why did the most extensive glaciation of the Prydz Bay shelf pre-date the Northern Hemisphere LGM? Since stage 2 is widely regarded as one of the most extreme of the Quaternary isotopic enrichment events, we are left with the possibility that the duration of the glacial stage is more important than the magnitude of the ice volume andor sea level change. For instance, stage6 (the penultimateglaciation) was nearlytwice the duration of the stage 2 low stand, using a threshold value ofabout-lOOm(Martinsonetal. 1987, Chappell& Shackleton 1986, Imbrie et a1 1984). Long periods of lower sea level might be more conducive to growth ofthe ice sheet across the continental shelf since response times vary between outlet systems feeding into the Lambert Graben (Amery Ice Shelf) and the ice sheet itself. We therefore suggest that D-4 may represent the penultimate glaciation and widespread glaciation out to the shelf break within the Prydz Channel. In contrast, the LGM glaciation ofPrydz Bay clearly reflected an advance ofoutlet systems and subsystems (i.e., LambertDeep drainage) rather than wholesale advance ofthe East Antarctic Ice Sheet. Thus, the short duration ofthe LGM may not have allowed the entire ice sheet to expand. Alternatively, we must ask if it is possible that a random oscillator, which may determine advances ofthe West Antarctic Ice Sheet (MacAyeal 1992), also applies to the Lambert Glacier system of the East Antarctic Ice Sheet. These ideas remain speculative, however, until more precise chronology can be provided for the sedimentary successions discussed above, specifically the interstadial units, SMO-2 and SMO-3. Conclusions As a result of this study, we have gained considerably in our understanding of the Late Quaternary stratigraphy of the Pry& Bay region. Specifically, we conclude the following: a) The Last Glacial Maximum (oxygen isotope stage 2) was not a shelf-wide glaciation and the Prydz Channel remained free of grounded ice during this interval, and ice shelf conditions existed across the channel during the LGM. b) The Pry& Channel contains a relatively condensed section of glacial marine and biogenic sediments that are uninterrupted by till layers, at least as far back as 30 000 yrs BP. This stratigraphy is continuous across the channel and contains three key beds of siliceous mud and ooze.

245

c) Major changes in lithofacies such as from till to glacial marine (ice shelf) deposits are associated with the welldefined grounding line “moraines” that rim the Prydz Channel. d) Glacial flow into the Prydz Bay shelf during glaciation is complicated and involves separate ice sources marked by distinct petrologic provenance of both tills and iceshelf sediments. Acknowledgements The authors would like to thank the crew and technical assistants whose help and patience during voyage 186 of the RSV Aurora Australis made the cruise most successful. The assistance of Lisette Robertson in preparing samples for radiocarbon analysis (University of Tasmania, CRC) was greatly appreciated. Dr Scott Ishman of the US Geological Survey (Reston) kindly provided a species list for AMS sample R21977/2. Discussions with Dr Will Howard (University of Tasmania) were most helpful as were the comments of the reviewers, J. Dowdeswell and A . Solheim. References ALLEY,R.B., BLANKENSHIP, D.D.,ROONEY, S.T. & BENTLEY, C.R 1989. Sedimentation beneath ice shelves - the view from ice stream B. Marine Geology, 85, 101-120. J.B. 1989. Glacial-marine sedimenration. Washington, ANDERSON, DC: American Geophysical Union, Short Course in Geology, vol 9, 127 pp. M.J. & DOMACK, E . W . 199 1. ANDERSON, J.B., KENNEDY,D.S., SMITH, Sedimentary facies associated with Antarctica’s floating ice masses. Geological Society of America, Special Paper, No. 261, 1-26. ANDERSON, J.B., KURTZ,D.D., DOMACK, E.W. & BALSHAW, K.M. I980 Glacial and glacial marine sediments from the Antarctic continental shelf. Journal ofGeofogy, 88, 399-414. F. & SONZOGNI, C. 1997. Interhemispheric synchrony BARD,E., ROSTEK, ofthe last deglaciation inferred from alkenone palaeothermometery. Nature, 385, 707-710. BARNES, P.W.& LIEN,R. 1988. Icebergs rework shelf sediments to 500 m offAntarctica. Geology, 16, 1130-1 133. P.J., HAMBREY, M.J. & ROBINSON, P . R . 1991 Cenozoic BARRETT, glacial and tectonic history from CIROS-I, McMurdo Sound. In THOMSON, M.R.A., CRAME, J.A. & THOMSON, J.W., eds Geological evolution of Aniarctica. Cambridge: Cambridge University Press. 65 1-656. BARRON J.E., LARSEN,B. et al. 1991. Proceedings oJ the Ocean Drilling Program Scientific Results, 119, 1003 pp. BROECKER, W.S. & DENTON. G.H. 1989. The role ofocean-atmosphere reorganisations in glacial cycles. Geochemica e / Cosrnochirnica Acta, 53, 2465-2501. S . W . & AHMAD, N. 1961. Glacial marine sedimentation. In CAREY, RAASCH,G.O.,ed. Geology ofthe Arctic, Vol. 2. Toronto: University of Toronto Press, 865-894. J . & SHACKLETON, N.J. 1986. Oxygen isotopes and sea level. CHAPPELL, Nature, 324, 137-140. J., NINNEMANN, U.S., FAIRBANKS, CHARLES, C.D., LYNCH-STIEGLITZ, R.G. 1996. Climate connections between the hemispheres revealed by deep sea sediment corelice core correlations. EarthandPlanetary Science Letters, 142, 19-21.

Downloaded from https:/www.cambridge.org/core. University of Tasmania Library, on 21 Apr 2017 at 00:31:06, subject to the Cambridge Core terms of use, available at https:/www.cambridge.org/core/terms. https://doi.org/10.1017/S0954102098000339

246

EUGENE DOMACK ef al.

COOPER, A , STACG,H. & GEIST,E. 1991. Seismic stratigraphy and structure of Prydz Bay, Antarctica: implications from Leg 119. In BARRON, J., LARSEN, B. et al.. eds Proceedings of the Ocean Drilling Program, Scientific result^, 1 19. 5-23. G.H. & HUGHES, T.J., eds. 1979. The Last Great Ice Sheets. DENTON, New York: John Wiley, 484 pp, DOMACK, E.W. 1982. Sedimentology of glacial and glacial-marine deposits on the George V-Adelie continental shelf, East Antarctica. Boreas, 11, 79-97. DOMACK, E.W., JULL,A.J.T. & NAKAO,S . 1991. Advance of East Antarctic outlet glaciers during the Hypsithermal: implications for the volume state of the Antarctic ice sheet under global warming. Geology, 19, 1059-1062. DOMACK, E.W., SHIPP,S. & JACOBSEN, E. 1996. Finally a realistic model for ice shelf sedimentation: new cores collected from the Ross Sea, Antarctica. Geological Society of America. Annual Meeting, Abstracts, 28, A-506. DOMACK, E.W. & HARRIS, P.T. i n press. A new depostional model for ice shelves based upon sediment cores from the Ross Sea and Mac.Robertson shelf, Antarctica. Annals of Glucrology. 27 DREWRY, D.J. & COOPER,A.P.R. 1981. Processes and models of Antarctic glaciomarine sedimentation. Annals of Glaciology, 2. 117-122. DUNBAR, R.B., ANDERSON, J.B., DOMACK, E.W. & JACOBS,S.S 1985 Oceanographic influences on sedimentation along the Antarctic continental shelf. Anturctic Research Series, 43, 29 1-3 12. R.B., LEVENTER, A.R. & SfOCKTON. W.L. 1989. Biogenic DUNBAR, sedimentation in McMurdo Sound, Antarctica. Marine Geology, 85, 155-179. FUTTERER, D.K. & MELLES, M. 1990. Sediment patterns in the southern Weddell Sea: Filchner Shelf and Filchner depression. In BLEIL,IJ. & THIEDE, J., eds. Geologic history of the polar oceans: Arctic versus Antarctic. Dordrecht: Kluwer, 38 1-434. Gow, A.J. & MEESE,D.A. 1996. Nature of basal debris i n the GISP2 and Byrd ice cores and its relevance to bed processes. Annals oj Glaciology, 22, 134-140. Gow, A.J., EPSTIEN, S. & SHEEHY, W. 1979. On the origin ofstratified debris in ice cores from the bottom of the Antarctic ice sheet. Journal of Glaciology, 23, 185-192. HAMBREY, M.J., ERHMANN, W.I. & LARSEN, B. 1991. Cenozoic glacial record from the Prydz Bay continental shelf. East Antarctica. Proceedings of Ocean Drilling Program ScientiJic Results, 119, 77-1 32. HAMBREY, M.J. & DOWDESWELL, J.A. 1994. Flow regime ofthe Lambert Glacier-Amery Ice Shelf system, Antarctica: structural evidence from Landsat imagery. Annals of Glaciology, 20, 401-406. D.G., MCINTYRE, A,, MIX, A., IMBRIE, J., HAYS,J.D., MARTINSON, J.J., PIS AS,^. & SHACKLETON, N.J. 1984. The orbital theory MORLEY, ofthe ice Pleistocene climate: support from arevised chronology of the marine 6 1 8 0 record. I n BERCER,A,, IMBKIE,J., HAYS,J.D., KUKLA, G. & SALTZMAN, B., eds. Milankovitch and climate, Part I Hingham, MA: Reidel, 269-305.

KELLOGG, T.B. & KELLOGG, D.E 1988 Antarctic cryogenic sediments, biotic and inorganic facies of ice shelf and marine-based ice sheet environments. Palaeogeography. Palaeoclimaiology, . Palaeoecology, 67. 51-74. KFYS, H.J.D., JACOBS, S.S. & BARNETT, D. 1990. The calving and drift of iceberg B-9 in the Ross Sea. Antarctica. Antarctic Science. 2. 243-257. G , STAGG, I.I., GANDJUKHIN. v.,COOPER, A.K., TANAHASHI, LEITCHENKOV, M. &O’BRIEN. P. 1994. CenozoicseismicstratigraphyofPrydzBay (Antarctica). Terra Antarctica. 1, 395-397. A.E., ANDREW, J.T. & WILLIAMS, K.M. 1996. LICHT,K.I., JENNINGS, Chronology of late Wisconsin ice retreat from the western Ross Sea, Antarctica. Geology, 24, 223-226. MACAYEAL, D.R. 1992. Irregular oscillations ofthe West Antarctic ice sheet. Nature, 359, 29-32. T C. & MARTINSON, D.G., PISIAS. N.G.. HAYS.J.D.. IMBRIE, J., MOORE, SHACKLETON. N.J. 1987. Age dating the orbital theory of ice agesd e v e l o p m e n t of a h i g h - r e s o l u t i o n 0 to 300,000-year chronostratigraphy. Quaternary Research, 27, 1-29. MILAM, R.W. & ANDERSON, J . R . 1981. Distribution and ecology of recent benthonic foraminifera of the Adelie-George V continental shelfand slope, Antarctica. Marine Micropaieontoiogy, 6,297-325. NUNESVAZ,R.A. & LENNON. G . W . 1996. Physical oceanography o f the Prydz Bay region of Antarctic waters Deep-Sea Research. 43, 603-64 I . O’BRIEN,P.E. & HARRIS. P.T. 1996. Patterns of glacial erosion and deposition in Prydz Bay and the past behaviour of the Lambert Glacier. Papers Rr Proceedings ufRoyul Sociery of Tasmania, 130, 79-85. O’BKIEN,P.E. & LEiTcHENKov, G . 1997. Deglaciation of Prydz Bay, East Antarctica, based on echo sounder and topographic features Antarctic Research Series, 71, 109-126. PUDSEY, C.J., BARKER, P.J. & LARTER, R.D. 1994 Ice sheet retreat from the Antarctic Peninsula continental shclf. Continental .They Research, 14, 1647-1675. PUSHINA, Z.V., KOLOBOV, D.D. & DRUZHINNA, N.I. 1997. Biostratigraphy and paleoecology o f the bottom sediments in Prydz Bay In Riccl, C , ed. The Antarctic region. geological evolution and processes. Siena: Museo Nazionale dell’ Antartide, 869-874. QUILTY, P. 1985. Distribution of foraminiferids in sediments of PrydL Bay and Mac. Robertson shelf. South Australian Department of Mines & Energy, Special Publication, No 5, 329-340. D.A. & MULLER, E € f . 1991. Late Wisconsinan, RIDGE,J.C., FRANZI, pre-Valley Heads glaciation in the western Mohawk Valley, central New York and its regional implications. Geological Society OJ America Bulletin, 103, 1032-1048. STAGG,H.M.J. 1985. The structure and origin of Prydz Bay and Mac.Robertson shelf, East Antarctica. Tectonophysics, 114,3 15-340. WOODWORTH-LYNAS, C.M.T. & GUIGNE, J.Y. 1990. Iceberg scours in the geological record: examples from glacial Lake Agassiz / n DOWDESWELL, J.A. & SCOURSE, J.D.,eds. Glacimarine environments. processes and sediments. Geological Society Special Publication. NO. 53, 217-223.

Downloaded from https:/www.cambridge.org/core. University of Tasmania Library, on 21 Apr 2017 at 00:31:06, subject to the Cambridge Core terms of use, available at https:/www.cambridge.org/core/terms. https://doi.org/10.1017/S0954102098000339