Geological Society of America Bulletin

Downloaded from gsabulletin.gsapubs.org on December 29, 2012 Geological Society of America Bulletin

Ice-flow switching and East/West Antarctic Ice Sheet roles in glaciation of the western Ross Sea

Sarah L. Greenwood, Richard Gyllencreutz, Martin Jakobsson and John B. Anderson

Geological Society of America Bulletin 2012;124, no. 11-12;1736-1749 doi: 10.1130/B30643.1

Email alerting services click www.gsapubs.org/cgi/alerts to receive free e-mail alerts when new articles cite this article Subscribe click www.gsapubs.org/subscriptions/ to subscribe to Geological Society of America Bulletin Permission request click http://www.geosociety.org/pubs/copyrt.htm#gsa to contact GSA Copyright not claimed on content prepared wholly by U.S. government employees within scope of their employment. Individual scientists are hereby granted permission, without fees or further requests to GSA, to use a single figure, a single table, and/or a brief paragraph of text in subsequent works and to make unlimited copies of items in GSA's journals for noncommercial use in classrooms to further education and science. This file may not be posted to any Web site, but authors may post the abstracts only of their articles on their own or their organization's Web site providing the posting includes a reference to the article's full citation. GSA provides this and other forums for the presentation of diverse opinions and positions by scientists worldwide, regardless of their race, citizenship, gender, religion, or political viewpoint. Opinions presented in this publication do not reflect official positions of the Society.

Notes

Ice-ﬂ ow switching and East/West Antarctic Ice Sheet roles in glaciation of the western Ross Sea

Sarah L. Greenwood1,†, Richard Gyllencreutz1, Martin Jakobsson1, and John B. Anderson2 1Department of Geological Sciences, Stockholm University, 10691 Stockholm, Sweden 2Department of Earth Sciences, Rice University, Houston, Texas 77005, USA

ABSTRACT mum did not all operate synchronously, and extent nor to have been subsumed by ice-sheet exerted different drawdown power at dif- glaciation throughout the Pleistocene (Stuiver The long-term behavior of the East and ferent times. Finally, we conclude that Ross et al., 1981; Sugden et al., 1995b, 1999). The West Antarctic Ice Sheets, and their respec- Island acts as an important pinning point in combined effects of large outlet glaciers drain- tive responses to forcing provide essential the Ross Sea ice-sheet–shelf system, stabilizing East Antarctic ice to the north and south, context for assessment of modern dynamic ing grounding line retreat and encouraging and southwesterly prevailing winds (i.e., from changes in ice-fl ow regimes and ice-sheet lasting ice-shelf development. the ice-sheet interior), yielding a very dry local and shelf margins. The western Ross Sea climate with high sublimation rates, result in discharges ice from both the East and West INTRODUCTION an enclave of Antarctica with minimal incom- Antarctic Ice Sheets, and the paleoglacial ing ice supply and little local nourishment. Ross record from this region is therefore valua- Recent changes in the Antarctic ice sheets, Island and McMurdo Sound lie at the border ble in unraveling their long-term behavior. in particular, their fast-fl owing drainage routes between this ice-starved enclave and a zone of New, high-resolution multibeam bathymetric and their margin transition zones, have received mass ice export from Antarctica. data reveal snapshots of well-preserved gla- much attention (e.g., Joughin et al., 2002; Understanding the glacial dynamics and his- cial landforms on the seafl oor around Ross De Angelis and Skvarca, 2003; Scambos et al., tory of Ross Island and McMurdo Sound is Island and McMurdo Sound. Glacial linea- 2004; Payne et al., 2004; Shepherd et al., 2002; signifi cant for three reasons: (1) proximity to tions, grounding zone wedges, draped reces- Pritchard and Vaughan, 2007; Rignot, 2008). the confl uence zone of the West and East Ant- sional moraines, and meltwater channels However, for understanding of (and predictions arctic Ice Sheets means the local ice dynamics record a series of different ice-fl ow events of) ice-sheet dynamics to be robust, it is neces- may hold clues to the relative roles of the two in the region, contradictions between which sary to assess ice-sheet behavior over a range main ice sheets in mass discharge from the con- require major phases of ice-fl ow reorgani- of time scales, from annual-decadal time scales tinent; (2) Ross Island is a pinning point of the zation. From the glacial geomorphology, we provided by modern observations to multiple modern ice-shelf system and has a crucial role reconstruct a four-stage model of ice-fl ow glacial cycles reconstructed from the geological in regional ice dynamics with implications for evolution for the last glacial cycle, consisting record. It is well-established that during the last ice-shelf (and, consequently, ice-sheet) stabil- of: (1) northeastward fl ow into the Ross Sea glacial cycle, expanded East and West Antarctic ity; and (3) a reconstruction of local-regional ice from McMurdo Sound; (2) westward fl ow Ice Sheets fi lled the continental shelves (Ander- dynamics is necessary to inform interpretation from the Ross Sea, around Ross Island, and son et al., 2002), before retreating to their pres- of deep-marine coring records, which seek to onto the Victoria Land coast and coastal sea- ent confi guration. However, the glacial history reveal the long-term climate history of Antarc- fl oor trough; (3) a deglacial phase of ice-sheet in the region around Ross Island and McMurdo tica (e.g., ANDRILL coring project: Naish et al., thinning, minor shifts in fl ow, and ground- Sound is a curiosity in the wider context of the 2007, 2009; McKay et al., 2009; Talarico et al., ing line retreat into McMurdo Sound; and Antarctic ice sheets, both past and present. 2012). Much prior research has been conducted (4) grounding line pinning on Ross Island Ross Island and McMurdo Sound (~77°S– onshore in the Dry Valleys and nearby blocks of during regional retreat, uncoupling of a rem- 78°S, 163°E–172°E) border the Ross Embay- the Transantarctic Mountains (e.g., Stuiver et al., nant Ross Island ice cap, and local oscillation ment (Fig. 1), one of the largest ice drainage 1981; Denton et al., 1989; Sugden et al., 1991, of Victoria Land outlet glaciers. We fi nd that outlets for all Antarctica and a major confl u- 1995a, 1995b, 1999; Marchant et al., 1993a, East Antarctic Ice Sheet ice discharge had a ence zone of ice from both the East and West 1993b; Brook et al., 1995; Stroeven and Pren- strong infl uence on ice-fl ow geometry in this Antarctic Ice Sheets. Close to this confl uence tice, 1997; Hall et al., 2000; Hall and Denton, part of the Ross Sea during the last glacial zone and enveloped by an expanded, grounded 2000a, 2000b; Atkins et al., 2002; Lewis et al., stage, but that it was not necessarily in phase ice sheet during glacial maxima (Stuiver et al., 2006; Staiger et al., 2006) in order to reconstruct with the behavior of the West Antarctic Ice 1981; Anderson et al., 1992, 2002; Denton and their recent and longer-term glacial and land- Sheet. It is similarly evident that the ice Hughes, 2000), the Royal Society, Dry Valley, scape evolutionary histories. Offshore, seismic streams that drained the Ross Sea over the and Convoy Mountain blocks of the Transant- stratigraphic, acoustic seafl oor and sub bottom continental shelf at the Last Glacial Maxi- arctic Mountain Range are relatively ice starved, data, and a number of ocean cores yield a pic- hosting mainly small valley glaciers, which are ture of the sedimentary environments, broad ice- †E-mail: [email protected] thought to have neither expanded to any great fl ow patterns, and longer-term glacial history in

GSA Bulletin; November/December 2012; v. 124; no. 11/12; p. 1736–1749; doi: 10.1130/B30643.1; 5 ﬁ gures.

140°W 160°W 180° 160°E 140°E 2 C SC e A n t M ra l N o T 80°S u ra n n ta s i a n n s t Ross Ice Shelf ByG a r Figure 1. Ross Sea (Ross Em- c t Marie Byrd (Ross Embayment) i DG c bayment) sector of the Antarc- Land tic ice sheets (West Antarctic RvI MG Ice Sheet [WAIS] and East Ant- arctic Ice Sheet [EAIS]) and SG paleo–ice-ﬂ ow reconstructions. RS (A) Regional reconstruction TD DV from Mosola and Anderson V RI MS i (2006) and Shipp et al. (1999). c

(B) Local ﬂ ow-line reconstruc- o 75°S tion from Denton and Hughes RB MwG r

CB i a

(2000). Abbreviated location/ 75°S DiT feature names: BG—Blue L CrB DT PT a glacier; BI—Beaufort Island; ByG—Byrd glacier; CB—Cen- n JB d tral Basin; CrB—Crary Bank; PB DG—Darwin glacier; DiT— EAIS Drygalski ice tongue; DT— WAIS Drygalski Trough; DV—Dry

Valleys; EP—Evans Piedmont glacier; FG—Ferrar glacier; GH—Granite Harbour; JB— 2 JOIDES Basin; KG—Koettlitz B glacier; MB—Mount Bird; MBf KG MBf—Minna Bluff; ME— N Mount Erebus; MG—Mulock Royal Society 78°S glacier; MkG—Mackay glacier; Range MS—McMurdo Sound; MT— Mount Terror; MTN—Mount Terra Nova; MwG—Mawson glacier; PB—Pennell Bank; BG PT—Pennell Trough; RB— FG Dry Valleys WV Ross Bank; RI—Ross Island; TV RS—Royal Society Range; 78°S RvI—Roosevelt Island; SC— Siple Coast; SG—Skelton Ross Ice S Shelf c glacier; TD—Taylor Dome; o McMurdo t t TV—Taylor Valley; WV— ME MTN Sound C o Wright Valley. Background MT a s is contemporary ice-sheet radar t 77°S image (Radarsat Antarctic Map- MB MkG ping Project [RAMP] data) with seaﬂ oor ETOPO topography. GH Convoy Range EP BI Ross Sea CB DT 77°S 170°E 165°E

Geological Society of America Bulletin, November/December 2012 1737 Downloaded from gsabulletin.gsapubs.org on December 29, 2012 Greenwood et al. the wider Ross Sea (e.g., Anderson et al., 1984, long-term paleostability of the East Antarctic Regional pictures of paleoglacial geometries 1992; Shipp et al., 1999; Domack et al., 1999; Ice Sheet, generating much debate in this regard and timing largely bypass the intricacies of ice Mosola and Anderson, 2006; Wellner et al., (e.g., Clapperton and Sugden, 1990; Webb and fl ow around Ross Island and McMurdo. Sea- 2006). New, high-resolution multibeam and Harwood, 1987, 1991; Barrett et al., 1992; Bar- floor macromorphology and high-resolution Chirp sonar data presented here from the sea- rett and Hambrey, 1992; Sugden et al., 1993, multibeam imagery indicate that grounded ice fl oor around Ross Island and McMurdo Sound 1995b, 1999; Marchant et al., 1993a, 1993b; from the Ross Sea fl owed northward and was bridge the regional-local gap in understanding Stroeven and Prentice, 1997; Hindmarsh et al., funneled into the major troughs of the outer the paleo-ice dynamics of this area during the 1998; Denton and Hughes, 2000; Naish et al., shelf (Fig. 1A). Geophysical data and a number last glacial period. Subglacial bedforms and ice- 2007; McKay et al., 2009). Detailed geomor- of cores reveal subglacial tills on the middle marginal landforms around Ross Island reveal phological and sedimentological mapping of and (in parts) outer continental shelf (Anderson the interplay among a local ice cap, an expanded glacial deposits around Ross Island, McMurdo et al., 1984, 1992; Domack et al., 1999; Licht grounded ice sheet, a regional ice-shelf system, Sound, and the ice-free coastal zones has been et al., 1999, 2005). Drumlins, megascale glacial and onshore ice-starved valleys. used to reconstruct the degree to which the most lineations, recessional moraines, and grounding recent glaciation encroached upon the longer- zone wedges occupy each of the shelf troughs, SETTING AND PREVIOUS RESEARCH term landscape (e.g., Brook et al., 1995; Den- which are inferred to have carried fast-fl owing ton and Hughes, 2000; Denton and Marchant, ice streams (Shipp et al., 1999; Domack et al., McMurdo Sound currently hosts an ice 2000; Dochat et al., 2000; Hall et al., 2000; Hall 1999; Mosola and Anderson, 2006). In the more shelf bifurcated from the Ross Ice Shelf. Geo- and Denton, 2000a, 2000b; Staiger et al., 2006). open midshelf regions, different clusters of lin- logical and geomorphological evidence from Rather than an expansion of ice from the Taylor eations are misaligned (Shipp et al., 1999), indi- ice-free enclaves of land around the sound Dome of the East Antarctic Ice Sheet and local cating more complex, shifting fl ow geometries suggest that Ross Sea ice previously fi lled and mountain blocks, an intricate fl ow of regional during the last glacial cycle. However, incom- grounded in the sound, encroaching onshore ice from the Ross Sea is reconstructed to have plete data coverage has hitherto limited further during previous Pleistocene glacial maxima fi lled the McMurdo region, spreading west- assessment of the dynamics of ice fl ow in the (e.g., Stuiver et al., 1981; Denton et al., 1989; ward, and encroached on land at the Last Glacial midshelf region, and the relationships between Denton and Marchant , 2000; Fig. 1). Ross Is- Maxi mum (Fig. 1B; Denton and Hughes, 2000; ice of competing sources/outlets during the last land, to which the Ross Ice Shelf is currently Denton and Marchant, 2000). This interpreta- glacial are not well resolved. pinned, and around which the shelf bifurcates, tion is based upon the distribution and compo- There is currently a scale mismatch between supports local ice caps and fi elds. Mounts Ter- sition of glacial deposits, and the requirements generalized regional models and detailed local ror (3230 m), Erebus (3794 m), Terra Nova of deltas and lake shorelines for ice-margin information. Furthermore, the styles of recon- (2130 m), and Bird (1765 m) attain such ele- positions and damming of proglacial lakes. The struction differ: Denton and colleagues re- vations that the island would have acted as a “Ross Sea Drift,” with a non-volcanic clast com- constructed a static, maximum picture of ice signifi cant block to regional ice during glacial position indicating provenance from the Central fl ow lines and the paleo-ice surface; offshore maxima, and at the same time nourished local Transantarctic Mountains and with shell frag- insights are largely directed toward the styles ice accumulation. The mid-outer continental ments indicating offshore (Ross Sea) grounding, of ice fl ow and yield snapshots of the patterns shelf is dissected by a series of troughs and mantles numerous ice-free slopes of the islands of ice fl ow and retreat. New, offshore data from intertrough banks (Fig. 1) with width scales of and peninsulas around McMurdo Sound. Its oc- close to Ross Island and in McMurdo Sound ~50–100 km and relief typically 50–200 m; the currence and altitudinal distribution have been contribute toward bridging these scales and deepest troughs lie in the western Ross Sea. The well documented (Stuiver et al., 1981; Denton integrating onshore and offshore observations. continental shelf as a whole shallows seaward, a and Marchant, 2000), and its maximum eleva- product of deep erosion on the inner shelf and tions have been used to reconstruct a maximum MARINE DATA AND OBSERVATIONS a seaward-thickening wedge of strata on the outer paleo-ice surface from Minna Bluff, around shelf. High-resolution seismic profi les (Alonso Ross Island, and through McMurdo Sound, to Swedish icebreaker Oden is equipped with a et al., 1992) supported by drill cores (Ander- where westward-transported sediments termi- Kongsberg 12 kHz EM122 1° × 1° multibeam son and Bartek, 1992) show that this wedge nate in the coastal mouths of the Dry Valleys as echo sounder, and Chirp sonar subbottom pro- consists of massive seismic facies that rest on lake shoreline deposits and moraines (e.g., Hall fi ler SBP120 3° × 3°, for seafl oor morphologi- unconformities, the relief of which mirrors the et al., 2000; Hall and Denton, 2000b; Denton cal mapping and subbottom sediment profi ling. modern seafl oor, alternating with acoustically and Marchant, 2000). Elements of the more pre- Three years of transit data around McMurdo layered seismic facies that are interpreted as cise ice-fl ow-line geometry depicted in Figure Sound, Ross Island, and the western Ross glaciomarine deposits (Alonso et al., 1992). 1B derive from the distribution of distinctive Sea (Oden Southern Ocean [OSO] data sets The coastal mountain blocks of southern Vic- kenyte erratics. Kenyte, a rare igneous phono- OSO0708, OSO0910, OSO1011; Fig. 2A) toria Land have received much attention due to lite, crops out extensively as bedrock on the provide high-resolution snapshots of glacial the region’s hyperarid local climate and resul- glacially scoured western fl ank of Ross Island landform assemblages, which were mapped in tant glacial and landscape histories, which are at Capes Royds, Evans, and Barne and at Turks a geographic information system (GIS) comple- arguably anomalous with respect to much of Head, but it is limited to these sites (Denton and mented with three-dimensional (3-D) visuali- the rest of Antarctica. Long-term ice-free zones Marchant , 2000). Its occurrence, and absence, zation software to fully explore landform (Sugden et al., 2006) and local valley glaciers in Ross Sea Drift around McMurdo Sound and morphology, relationships, and form with regard operating out of phase with the continental ice the Scott Coast underpins the fl ow-line recon- to the subbottom data. The Oden data (http:// sheet (Denton et al., 1989; Brook et al., 1995; struction. Ross Island itself is inferred to have oden.geo.su.se) provide valuable new informa- Hall et al., 2000) render these parts of southern supported local, independent ice, confl uent with tion, yet their coverage is limited to the transit Victoria Land a key region in deciphering the the regional grounded ice sheet. lines, and it is useful to consider other avail-

1738 Geological Society of America Bulletin, November/December 2012 Downloaded from gsabulletin.gsapubs.org on December 29, 2012 Glaciation of the western Ross Sea

170°0′0″E 165°0′0″E

Terrestrial A elevation (m) HP 3A

78°0'0"S > 2000 3G

3F 1000 TH Ross Island S CBa/R c o tt 0 m C o a s MS t CC 3B

3H S ″ 0 ′ 3E

CBr 77°0

OSO Figure 2. Data used and asso ciated 07 08 0910 geomorphological mapping and OSO interpretation. (A) Oden multibeam from cruises OSO0708, BI OSO0910, and OSO1011, laid S

″ 3D upon the GeoMapApp multireso- 0 ′ lution bathymetry. Inset boxes 77°0 OSO1011 give locations for images in Fig- ure 3. BI—Beaufort Island; CB— Marine depth (m) CB 3C Central Basin: CBa/R—Cape –200 Barne/Cape Royds; CBr—Cape DT Bird; CC—Cape Crozier; DT— –600 Drygalski Trough; HP—Hut

Point Peninsula; MS—McMurdo –1000 m

Sound; TH—Turks Head. Terres- trial topography is from Radarsat 2 S

″ B 0

Antarctic Mapping Project ′

(RAMP) digital elevation model. 78°0 N (B) Mapped glacial landforms (fi ne/pale lines) and fl ow set/sum- mary interpretations (heavy lines). Lineations/fl ow sets are in gray, rm with arrows showing ice-fl ow direction; moraines/margins are in orange, with arrows indicating fs6

retreat direction; channels are in S ″ 0 blue; ribbed moraine is in green. ′ Order of crosscutting assemblages 77°0 fs3 3a is shown by white disks; center fs7 line shows uppermost. fs5 2a

fs4 S ″ 0 ′ 77°0

fs2

010203040 fs1 km

170°0′0″E 165°0′0″E

Geological Society of America Bulletin, November/December 2012 1739 Downloaded from gsabulletin.gsapubs.org on December 29, 2012 Greenwood et al. able data from a wider regional context. The straight and parallel lineations between 285 m tionally indicate some crosscutting of drumlins GeoMap App database (http://www.geomapapp and 3235 m in length (average 982 m) with at the western and southern end of the assem- .org/; a resource hosted at the Lamont-Doherty amplitudes of 2–5 m, trending approximately blage (fs2a). This confi guration of main swarm Earth Observatory at Columbia University) hosts NW slightly oblique to coast-parallel. Closer and crosscutting “tail” is mirrored by the as- and integrates global bathymetry data of the best to shore (i.e., in the lateral, across-ice-fl ow di- semblage sweeping around Cape Bird (fs3, 3a). available resolution for any given region into a rection), the lineations are best described as These assemblages indicate substantial fl ow of single multiresolution product. Multibeam data drumlins; further from Ross Island, the linea- regional, grounded ice to the westernmost Ross were extracted from the GeoMapApp platform, tions become longer and are better described Sea, with some accompanying minor shifts and further landform mapping was conducted to as megascale glacial lineations (Clark, 1993). in fl ow geometry leading to overprinting and complement that derived from the Oden lines. Three to four large sedi ment accumulations are crosscutting of drumlins. The complete landform mapping, together with overprinted by the lineations, lightly smearing interpretative assemblage grouping (fl ow sets), the wedge morphology in the downstream di- Mackay Glacier is shown in Figure 2B; the following sections rection. Wedges are spaced at 2–3 km and have describe the key landform clusters revealed by amplitudes of 20–40 m. Shoreward, where the A well-preserved assemblage of glacial linea- our data sets. lineations are more drumlin-like, the underly- tions and grounding zone wedges (Figs. 2 and ing ridges are smaller and/or more heavily de- 3E) records ice retreat and associated adjust- Hut Point Peninsula formed. Ridge orientation is broadly E-W where ments in fl ow direction from the open conti- they lie north of Ross Island; off Cape Crozier, nental shelf back through a coastal trough into A series of ridges drapes the seafl oor between the moraines and wedges are oriented NE-SW, Granite Harbour (NBP0401, 0301, 9601; fl ow Hut Point Peninsula and Cape Barne, fronting the perpendicular to the overriding lineations (Figs. set 5). Lineations are elongate and highly paral- region into which the Erebus glacier tongue cur- 2B and 3B). Well-defi ned and low-amplitude re- lel, suggesting that this fl ow route likely operated rently protrudes (Figs. 2 and 3A). The ridges are cessional moraines form a pristine drape over the as a feeder to more extensive ice streaming, and oriented along the contours of the falling slope whole assemblage. These have a typical spacing in its decay as a fast-fl owing outlet glacier. Suc- from Ross Island into McMurdo Sound, lying at on the order of 1 km, with amplitudes of 0.5–4 m, cessive retreat positions are each marked with a depths between 350 m and 500 m. These ridges and are oriented perpendicular to the lineations. grounding zone wedge, each with an associated have very clear geomorphological expression yet The whole assemblage indicates an oscillating group of lineations of subtly different orienta- display an average amplitude of only ~1 m; they ice margin with local retreat and readvance cycles tions as the fl ow became increasingly confi ned are arranged with a crest-crest spacing of just during the retreat of the grounded ice sheet. to the local bathymetry. On the lateral (north- over 100 m (based upon fi ve cross-ridge profi les ern) fl ank of the coastal trough, a small group totaling 12.6 km profi le distance and covering Central Basin of lineations indicates earlier fl ow directed to- 91 measurable features). Within the limits of our ward the outer shelf, indicative of contributions data coverage (OSO0910), the assemblage of A 15 km track-line stretch of OSO1011 data to regional grounded ice. Furthermore, a suite ridges covers an area of ~5 km across-ridge, and shows very fi ne sedimentary lineations trending of grounding zone wedges displaying very simi- 12–13 km along their length. The ridges drape a into the Central Basin. Lineations are highly elon- lar morphology to those confi ned to the Granite seafl oor scarp of ~30 m depth and slope angle of gate, parallel, and <5 m in amplitude. This suite Harbour coastal trough is apparent underneath around 10° with little perturbation. They possess is consistent with patches of high-resolution data the Drygalski lineations fl ow set (fs2). These are a slightly curvilinear individual geometry, and embedded within the GeoMapApp bathymetry concave toward the Mackay coast and indicate the assemblage as a whole is faintly concave to- (e.g., Nathanial B. Palmer (NBP) cruises 9902, a phase of grounded ice retreat prior to an in- ward Ross Island. The form and setting of these 9407, 0305A, 9801) that show similarly trend- cursion of regional ice from the Ross Sea with ridges lead us to interpret them as moraines. ing parallel, linear features (Figs. 2 and 3C). This which the Drygalski lineations are associated. The ocean depth is such that these must mark a repeated lineation signature across multiple data marine-terminating ice margin, and we speculate patches records a widespread northeastward ice- McMurdo Sound these are De Geer moraines (Lindén and Möller, fl ow path into the Central Basin on the midconti- 2005; Todd et al., 2007). They indicate steady nental shelf (fl ow set 1). McMurdo Sound itself has a complex geo- retreat of thick ice from McMurdo Embayment morphology, likely shaped by volcanic, glacial, onto Ross Island, and their orientation suggests a Drygalski Trough and marine processes. The western fl ank of local ice source independent of the regional Ross the sound consists of a broad, shallow coastal Sea ice-sheet–shelf system. A striking landform assemblage in the re- shelf dipping steadily into the central sound; gional data (primarily NBP0401; also NBP the deepest parts of the sound are on the eastern Northeastern Ross Island–Cape Crozier cruises 0302, 0801, 0301B, 0702) sweeps west- (Ross Island) fl ank. In the glacial domain, we ward from the Central Basin toward the Vic- tentatively identify both subglacial bedforms Sedimentary lineations are evident in all three toria Land coast (fl ow set 2; Figs. 2 and 3D). and channels, the latter indicating (melt)water Oden multibeam data sets around the north side Straight, elongate, and highly parallel linea- activity. Additionally, we fi nd large, braided of Ross Island. On the NE fl ank of the island, off tions comprise this fl ow pattern, which passes channel systems likely related to postglacial Cape Crozier, a large suite of elongate lineations perpendicular to the trend of the outer-shelf submarine processes. is found in concert with a series of grounding troughs before being drawn toward the head of zone wedges, overprinted moraines, and pris- Drygalski Trough. The GeoMapApp bathym- Subglacial Bedforms tine, draped recessional moraines (Figs. 2 and etry confi rms and extends snapshots of this fl ow In the deepest, central-eastern part of Mc- 3B; fl ow set 7; OSO0708). This assemblage lies set revealed by Oden data sets OSO0708 and Murdo Sound, an irregular seafl oor topography at 700–750 m depth and consists of extremely OSO1011, though, importantly, the latter addi- is overprinted by a small cluster of streamlined

1740 Geological Society of America Bulletin, November/December 2012 Downloaded from gsabulletin.gsapubs.org on December 29, 2012 Glaciation of the western Ross Sea

Distance (m) 2 km A 04008001200 B –470 –480 –490 Depth (m) MSGLs Recessional moraines N

Drumlins

Distance (m) Figure 3. (A) Moraines drape 0 1000 2000 3000 4000 5000 the seafl oor between Hut Point –440 –460 Wedge fronts Peninsula and Cape Barne; 2 km –480 Depth (m) profi les across the assemblage –500 show high-frequency, low-am- C D plitude undulations. (B) Cape Crozier megascale glacial lin- N N GZWs eations (MSGL) overrun till Lineations wedges (fl ow to bottom right) and are themselves draped per- pendicularly by low-amplitude GZW recession moraines. (C) High- resolution track lines show lin- eated seafl oor (toward bottom Lineations left) leading into the Central Basin. (D) Lineations sweep left to right (E to W) toward 2 km 2 km Granite Harbour and the Dry- galski Trough, overprinting E F earlier grounding zone wedges (GZW). (E) Coastal trough ex- N tending from Mackay glacier Ribbed moraine? (Granite Harbour) contains Lineations an assemblage of lineations and grounding zone wedges; (F) ribbed moraines on the Scott GZWs Coast shallow shelf flank a deeply incised channel system; GZW (G) channels with an up-and- Bifurcating channel system down long profi le, coast-par- 2 km allel to Ross Island, incised in GZWs N irregular bedrock topography; Distance (m) 0 200 400 600 800 1000 1200 Ribbed moraine? (H) meandering and braiding –560 G channel system exiting Mc- –580 2 km Murdo Sound, carved in soft Depth (m) –600 sediments with a hummocky Fig.4 topography. 2 km Meandering channel Channel Hummocky topography Cape Bird (Ross Island) Channels Braiding channel system Depth scale

–200

N –600 Channel exit 2 km –1000 (m) N H

Geological Society of America Bulletin, November/December 2012 1741 Downloaded from gsabulletin.gsapubs.org on December 29, 2012 Greenwood et al. features consistent with drumlin morphology channels in Wright Valley: Lewis et al., 2006); reaches, this takes the form of a single primary (OSO1011; fl ow set 6). Fifty-four such linea- (2) drainage of proglacial lakes from the Dry channel; north of ~77°15′S, two or three “main” tions are mapped, trending in a northward direc- Valleys following removal of a damming ice channels pass through a large, shallow, braided tion out of McMurdo Sound. Though this Oden margin in McMurdo Sound; or (3) submarine channel system, weaving north to northeastward track line passes close by those from OSO0708 channels formed by gravity-controlled turbidite through thick sediment cover (Fig. 4). Chirp so- and OSO0910, the drumlin signature is identifi ed fl ow (e.g., Ó Cofaigh et al., 2006; Laberg et al., nar profi les show a continuous, ~1–2 m drape in neither of those data sets, nor the GeoMapApp 2007; Noormets et al., 2009). The geometry and of acoustically laminated sediments parallel- bathymetry. South and west of this assemblage, size of these systems are rather atypical of a ing the seafl oor morphology, in the immediate on the shallow coastal shelf off the Scott Coast, subglacial domain: The systems bifurcate from subsurface (Fig. 4B). This drape is underlain by we observe a cluster of irregular ridges (Figs. 2 a point (rather than a line/area) source and are a 10–15-m-thick, almost acoustically transpar- and 3F). These features are broadly linear, ~400– both larger and considerably less intricate than ent unit, containing a series of convex-upward 1800 m long and ~200–400 m wide, with some the nearby Wright Valley Labyrinth channels weak refl ectors. We suggest that these relate to sinuosity but no obvious concavity. They occur attributed to massive subglacial fl oods in the a shifting channel system (i.e., braiding), and both on interchannel “plateaus” and more exten- mid-Miocene (Lewis et al., 2006). Furthermore, that the draping, parallel uppermost refl ectors sively on either side of a seafl oor channel system albeit large in size, we question their potential indicate that no postdepositional erosion has (see following). The ridges (and inter-ridge de- for preservation given repeated grounding and occurred (i.e., that the present seafl oor mor- pressions) display neither the regularity of sub- retreat of Ross Sea ice throughout the Pliocene– phology formed syndepositionally). We infer marine sand waves nor of water “rill” patterns, Pleistocene (Alonso et al., 1992; Anderson and a braided channel system formed by repeated and their broad shape, lack of linear continuity, Bartek, 1992; Naish et al., 2007, 2009; McKay erosion-deposition events, possibly turbiditic. and irregular pattern are not consistent with gla- et al., 2009; Bart et al., 2011). We also rule out At its exit from the sound, this system comes cial moraines. We interpret a glacial, rather than formation in a single episode of ice-dammed into an area of more explicit (sub)glacial infl u- submarine, imprint, and speculate that these may lake drainage from the “upstream” Taylor Valley, ence, and it is not possible to rule out a glacial be ribbed moraines. where proglacial lakes have been reconstructed origin for parts of the network. We suggest a (Stuiver et al., 1981; Denton et al., 1989; Hall predominantly submarine origin for the central Channels et al., 2000; Wagner et al., 2006). If the simple McMurdo channel system, as for the western The fl oor of McMurdo Sound contains abun- assumption is taken that the discharge rates re- systems, possibly driven by a proglacial supply dant evidence of channel systems (Figs. 2 and constructed for the nearby Labyrinth channels of dense, debris-laden waters. 3F–3H). These can be divided into three groups: (~2 × 106 m3 s–1: Lewis et al., 2006) are an ap- Ice-sheet grounding and retreat, ice-fl ow-line deep incisions in a bedrock landscape toward propriate, ballpark estimate of those required to confl uence from various potential inputs, and the south of the sound; two large systems on the carve the offshore channels, then a proglacial a variety of (sub/proglacial) hydrological sys- western slopes; and an extensive channel com- lake dammed at 350 m elevation in Taylor Valley tems have contributed to a highly irregular and plex carved in sediments on the fl oor of the deep- (Hall et al., 2000; Hall and Denton, 2000a) hummocky seafl oor morphology in McMurdo est parts of McMurdo Sound. Running parallel would drain within ~5.5 h. This is a maximum Sound, at a landform (rather than macro-) scale. to the western Ross Island coastline, off Capes estimate based on Hall et al.’s (2000) maximum This is particularly the case in the northern part Barne/Royds, intricate deep gorges, with typical lake level and using a U.S. Geological Survey of McMurdo Sound, where a largely bedrock– dimensions of ~250 m width and ~20 m depth, light detection and ranging (LiDAR)–derived thin sediment environment transitions to a zone dissect a bedrock-dominated seafl oor at around digital elevation model (DEM) to calculate lake hosting thick sediments with a hummocky and 550–650 m ocean depth (OSO0910 and Geo- volume; a minimum lake level at 78 m would complex surface morphology (e.g., Fig. 4). We MapApp bathymetry). One such feature has an drain its supply within half an hour, making this note that McKay et al. (2008) disregarded core undulating thalweg (a climb of 30 m over 500 m mechanism highly unlikely for the primary for- records from central and northern McMurdo channel length), suggesting these channels were mation of the large channel systems. This is sup- Sound, reporting heavily reworked, “incom- formed subglacially under hydrostatic pressure. ported by research by Wagner et al. (2006), who, plete” facies sequences compared to other On the opposite side of the sound, two large based on the sedimentary record of Fryxell Basin nearby cores and inverted radiocarbon stratigra- systems emanate from the Scott Coast and fl ow in Taylor Valley, found that the proglacial lake phies. Unraveling the sequence of geomorpho- down the submarine coastal slope (GeoMapApp level lowered discontinuously and proceeded via logical and sedimentological domains here will data: NBP0401). These fl ow from an ocean evaporation rather than drainage. The downslope be a challenge for systematic and high-resolu- depth of 250 m to deeper than 600 m, over a dis- bifurcation of these large systems bears strongest tion bathymetry and subbottom mapping. tance of ~15 km. Both systems are distributary, resemblance to turbidite channel systems (e.g., bifurcating toward the center of the sound. The Dowdeswell et al., 2004; Ó Cofaigh et al., 2006; Absent Features main channels are between 650 m and 2000 m Laberg et al., 2007; Noormets et al., 2009), wide, and with depths of 50–70 m. Given the which form when cold, sediment-laden waters Within the limits of available data cover- geometry and orientation of these channel sys- are driven down a topographic gradient. We sug- age, the western Ross Sea, McMurdo Sound, tems, their size (discharge capacity), and argu- gest that these systems are of submarine, rather and the seafl oor around Ross Island are devoid ments for minimal warm-based ice in the coastal than primarily glacial, origin, with the potential of evidence of iceberg grounding or ice-shelf mountain blocks of southern Victoria Land since for reuse and further excavation by meltwater breakup. There are no iceberg scour marks of the mid-Miocene, it is unlikely that these are re- drainage along the Scott Coast. any styles (either chaotic or linear/organized), cent subglacial meltwater systems. Rather, we Finally, the deepest part of McMurdo Sound even at relatively shallow depths less than 500 m consider three possible origins of these chan- carries a channel system from south to north below present sea level, in contrast to the outer nels: (1) long-term preservation of Miocene– through the sound, running approximately par- continental shelf (Shipp et al., 1999) and more Pliocene (glacio)fl uvial channels (cf. Labyrinth allel to the Ross Island coast. In its southern central midshelf parts of the Ross Sea.

1742 Geological Society of America Bulletin, November/December 2012 Downloaded from gsabulletin.gsapubs.org on December 29, 2012 Glaciation of the western Ross Sea

A N file (B) ottom pro subb

Depth (mbsl) main channel, continuous and 860 meandering through McMurdo Sound

shallow braided channel system 870

Distance (km) 02 880 B main channel braided channel system

Depth (mbsl) 890

895

900

905

910

915 Distance (km) 0 2

Figure 4. Seafl oor channel systems at the northeastern exit of McMurdo Sound. (A) Multibeam data (OSO0910) reveal a deep incision fl anked by shallow braided channels, which regional GeoMapApp data reveal to be part of a larger meandering and braiding network draining the fl oor of McMurdo Sound (Fig. 3H). (B) Subbottom (Chirp) data along the central profi le of the multibeam swath show a 10-m-deep primary incision north of ~3-m-deep braiding. The braided section consists of a 10–15-m-thick acoustically transparent unit with a few weak convex refl ectors of similar spatial scale to the seafl oor morphology. An acoustically laminated drape (1–2 m) indicates no postdepositional erosion. We infer system formation under a series of repeated, shifting erosion-deposition events, possibly a product of submarine turbid currents driven by postglacial input of cold, dense meltwaters from the ice- shelf system, which mobilize and rework the (glacial) sediment cover in the northern sector of McMurdo Sound. Depth scales are given in meters below sea level (mbsl).

ICE-FLOW AND RETREAT MODEL fl ow confi guration dominated by northeastward- ning and oscillation around Ross Island, and a fl owing ice, requiring fl ow from McMurdo small readvance of Victoria Land outlet glaciers Figure 2B summarizes the landform assem- Sound; (2) a phase of westward fl ow of Ross following the loss of the regionally grounded blages evident around Ross Island–McMurdo Sea ice around Ross Island toward the Victoria Ross Sea ice. Here, we briefl y describe the evi- Sound into discrete units we treat as a basis for Land coast, and into Drygalski Trough; (3) a dence and fl ow geometries associated with each reconstructing the sequence of ice-fl ow and re- likely deglacial phase in which minor shifts in of these stages, before further discussion. treat events in this region. Deriving the simplest the westward fl ow of regional ice accompanied Flow set 1 consists of a coherent group of lin- interpretation of the available data, we identify margin (grounding line) retreat and the opening eations curving northeastward from the direction four stages of glacial history (Fig. 5): (1) an ice- of McMurdo Sound; and (4) grounding line pin- of McMurdo Sound into the Central Basin. To

Geological Society of America Bulletin, November/December 2012 1743 Downloaded from gsabulletin.gsapubs.org on December 29, 2012 Greenwood et al.

170°0′0″E 165°0′0″E 170°0′0″E 165°0′0″E 1 2 78°0'0"S S ″ 0 ′ 77°0 S ″ 0

′

77°0 2

N S ″

0 3 4 ′ 78°0 S ″ 0 ′ 77°0 S ″ 0 ′ 77°0

010203040 km 170°0′0″E 165°0′0″E 170°0′0″E 165°0′0″E

Figure 5. Four stages (1–4) of ice-fl ow evolution are reconstructed in the western Ross Sea. See text for description of fl ow geometries. Arrows indicate ice-fl ow paths: solid lines are based on our offshore data and published onshore drift distributions; dashed lines are infer- ence. Dotted lines indicate ice-margin (grounding) positions. Stippled patches represent ice not signifi cantly contributing to the regional ice sheet (e.g., cold-based ice, local ice cap, sticky spot, terrestrial valley glaciation). Refer to Figure 1 for regional map-view context. drive this fl ow pattern, ice must exit McMurdo fl ow around Beaufort Island, which, together ratics westward onto the Scott Coast (Fig. 1B). Sound to the north. Such a fl ow geometry in turn with central parts of Ross Island, likely were Any Ross Island ice cap would be effectively requires infl ow of ice from the Ross Sea into cold-based or otherwise geomorphologically blocked at this time from contributing to the southern McMurdo Sound and/or signifi cant inactive. Stage 2 reveals a radically different regional ice sheet, though it may have assisted contributions of ice from the outlet glaciers that ice-fl ow geometry, represented by the westward in the transport of Ross Island erratics across penetrate the Victoria Land mountain blocks and ice-fl ow movement of fl ow sets 2 and 3. These McMurdo Sound to the Scott Coast. The shape tap the East Antarctic Ice Sheet (stage 1). Flow set fl ow sets require a much diminished Victoria of fl ow set 2 indicates that westward-spreading 6 and the potential ribbed moraine cluster in Mc- Land infl uence, and strong regional infl ow from regional ice was drawn across the Central Basin Murdo Sound may also fi t this stage, as may fl ow the Ross Sea. This confi guration is most compat- and into Drygalski Trough. Elongate and highly set 4, which indicates north-northeast fl ow from ible with terrestrial evidence reported by Denton parallel lineations reveal the fast-fl owing nature a Ross Island ice cap. Alternative interpretations and Marchant (2000), which indicates regional of this event. for any of these three landform assemblages do Ross Sea ice wrapping around Ross Island, car- The extent of stages 1 and 2 onto the outer not negate the Stage 1 fl ow confi guration, which rying Ross Sea subglacial debris onto the lower continental shelf, and their timing with respect is demanded by fl ow set 1. We infer bifurcating slopes of the island and Ross Island kenyte er- to the maximum stage of glaciation are unknown

1744 Geological Society of America Bulletin, November/December 2012 Downloaded from gsabulletin.gsapubs.org on December 29, 2012 Glaciation of the western Ross Sea without further detailed examination of the interpreted advance to a maximum position in suggest that during deglaciation, grounded ice outer continental shelf record. Stages 3 and 4 the central Ross Sea after 13.8 14C ka (though the lingered along the shallow Scott Coast longer are, however, deglacial in timing. Stage 2 must validity of this date is questioned by Mosola and than the more “open” part of the western Ross therefore represent a confi guration either prior Anderson, 2006). Retreat was likely under way Sea, damming the coastal valleys well after the to deglaciation (i.e., Last Glacial Maximum) or, by ca. 11.6 14C ka, with loss of the calving front regional grounding line had retreated southward. given the correspondence of its fl ow geometry from the outer western Ross Sea by 11.1 14C with stage 3, may also be part of the regional ka (Domack et al., 1999), and passing Drygal- DISCUSSION retreat sequence. Flow sets 2a and 3a both cross- ski Tongue ca. 10–9.5 14C ka (Domack et al., cut their larger, more extensive counterparts, 1999; Licht and Andrews, 2002) and north of Ice-fl ow patterns around Ross Island–Mc- with a geometry more closely guided by the Ross Island by 8–6.5 14C ka (Licht et al., 1996; Murdo Sound reveal two prevailing glacial underlying topography: Cape Bird, and the sea- McKay et al., 2008). The grounding line is sug- states in the western Ross Sea: an ice-sheet fl oor topography off Beaufort Island. We sug- gested to have retreated from south of Ross confi guration driven by NE-spreading ice from gest little shift in the regional ice-fl ow direction Island by ca. 10 14C ka (McKay et al., 2008), southern Victoria Land, and one driven by but increased splay at the distal end of the two implying rapid retreat of the regional ground- westward-spreading ice from the Ross Embay- lineation assemblages associated with ice-sheet ing line, during which time an ice shelf devel- ment. The latter is consistent with most recent thinning and proximity to a lobate margin (Fig. oped and the calving front retreated at a steady models for Last Glacial Maximum ice fl ow in 5C: stage 3). The opening of a narrow embay- but more sedentary pace. A large collection of this region (e.g., Denton et al., 1989; Barrett ment from the head of Drygalski Trough toward radiocarbon dates derived from algae and shell and Hambrey, 1992; Denton and Hughes, 2000; McMurdo Sound was conducive to lobate mar- fragments in lacustrine and raised marine de- Hall et al., 2000; Mosola and Anderson, 2006; gin development, while grounded ice remained posits help to constrain a chronology from ter- Talarico et al., 2012), in which, broadly speak- along the Scott Coast, continuing to dam pro- restrial sites around McMurdo Sound (Denton ing, ice from the Ross Embayment envelops glacial lakes (Hall et al., 2000; Hall and Denton, and Marchant, 2000; Dochat et al., 2000; Hall Ross Island and spreads into McMurdo Sound 2000a, 2000b) until after the regional grounding and Denton, 2000a, 2000b). Ice was grounded and the western Ross Sea continental shelf line had passed (see chronology in the follow- across the mouth of Taylor Valley (Hall and troughs. Both offshore and onshore glacial evi- ing). Stage 4 fi nally was marked by regional ice Denton, 2000a) after 23.8 14C ka, depositing dence is in agreement with regards to this stage, uncoupling from a remnant Ross Island ice cap. a moraine sequence from 14.6 to 10.7 14C ka indicating strong infl ow to the region from the Local ice remained grounded off Hut Point Penin- and damming proglacial lakes until 8.3 14C ka. Ross Sea around Ross Island, and minimal sula, forming a sequence of moraines, while the Grounded ice was lost from the valley mouth contribution of local ice; such a confi gura- regional grounding line became temporarily between 8.3 and 6.5 14C ka, with shelf ice also tion is broadly adhered to in three of our four pinned off Cape Crozier, where the landform absent along the Scott Coast by 6–5.5 14C ka. reconstruction stages and may be deemed the record reveals a localized retreat–readvance se- Since our model stages 1 and 2 reconstruct dominant state at or following the Last Glacial quence. Local readvance is additionally recon- complete ice cover of the study area, we are Maximum and through deglaciation. An alterna- structed on the Victoria Land coast (Fig. 5D). not able to assign ages to these confi gurations, tive geometry is, however, demanded by some Deglacial fl ow set 5 cannot be easily reconciled beyond suggesting they occur after 23.8 14C elements of the offshore geomorphology. Linea- as being contemporaneous with the retreating ka when Taylor Valley was fi rst dammed (Hall tion fl ow sets 1 and 4, and potentially much of fl ow sets 2/2a, but it is better accommodated and Denton, 2000a). Alignment of stage 2 with the geomorphology from McMurdo Sound it- as a product of local readvance following with- existing interpretations of terrestrial deposits self (fs 6, ribbed moraine, elements of the chan- drawal of regional ice. Grounding zone wedges and their associated chronology (Denton and nel systems), cannot be accommodated within within and lateral to the Granite Harbour coastal Marchant, 2000) would place stage 2 at the Last a “westward-spreading” model, and point to a trough indicate fi nal retreat both into Mackay Glacial Maximum, and stage 1 would therefore strong northeastward element to ice fl ow. glacier and to the Evans piedmont glacier. represent buildup post–23.8 14C ka. Alterna- A northeastward fl ow trajectory from Mc- Remnant ice grounded on the southern Scott tively, the coherence of ice-fl ow patterns in the Murdo Sound into the Central Basin has certain Coast must have been close to stagnation, with marine landform data between stages 2 and 3 requisites for the regional ice-fl ow geometry, minimal incoming ice supply, and was therefore suggests that stage 2 may only shortly precede none of which has hitherto been explicitly in- likely to undergo rapid collapse as soon as ice the latter, and that stage 2 represents part of the ferred for the last glacial period, and some or became thin enough to fl oat. retreat phase from this part of the western Ross all of which must be fulfi lled: (1) ice fl ow south Sea. These alternative chronology scenarios to north through McMurdo Sound; (2) radial ice CHRONOLOGY place slightly different emphasis upon differ- fl ow from a Ross Island ice cap; and (3) strong ent lines of evidence, but if (rapid) retreat oc- contributions from several southern Victoria Establishing a chronology for glacial events in curred from the outer shelf without interior fl ow Land outlet glaciers. Weak outfl ow from Vic- Antarctic marine environments is a well-known reorganization, they may in fact be compati- toria Land glaciers would most likely be guided challenge, where marine fauna for radiocarbon ble. We tentatively place stages 3 and 4 at ca. by coastal topography and drawn into Dry galski dating are sparse and dating bulk organic sedi- 11–10 14C ka and 10–8 14C ka, respectively. In Trough, contrary to our observations of fl ow ment is fraught with uncertainties of reservoir the absence of a more detailed pattern of both across toward the Central Basin. We therefore correction and reworking of the sediment col- grounding and calving line retreat, and given the suggest that there must have been signifi cant umn (Andrews et al., 1999; Licht and Andrews, inherent uncertainties in radiocarbon dating in fl ow from outlet glaciers such as the Mackay 2002; Anderson et al., 2002). Notwithstanding this environment, it is diffi cult to derive a more and Mawson, from the Skelton, Mulock, Dar- such problems, a limited chronology is provided precise chronology of deglaciation. The pattern- win, and Byrd glaciers, and potentially also by a series of offshore cores and from onshore based reconstruction we present here, however, from Ferrar and Koettlitz glaciers in order to organic deposits. Licht and Andrews (2002) combined with the available radiocarbon dates drive ice fl ow not only north through McMurdo

Geological Society of America Bulletin, November/December 2012 1745 Downloaded from gsabulletin.gsapubs.org on December 29, 2012 Greenwood et al.

Sound but additionally northeast into the Cen- this may have occurred during deglaciation, in lets have been well documented elsewhere, for tral Basin (see Fig. 1 for glacier locations). The which case our NE fl ow stage (i.e., dominant example, on the Siple Coast of West Antarctica major outlet glaciers penetrating the Dry Val- East Antarctic contribution) dominated the Last (e.g., Conway et al., 2002; Joughin et al., 2002), leys and Royal Society Range have been pre- Glacial Maximum in the Ross Sea, and stages in Pine Island Bay (Graham et al., 2010), and viously acknowledged to block the westward 2–4 are all products of a late-glacial readvance between the M’Clure and Amundsen paleo–ice infl ow of regional Ross Sea ice (e.g., Denton and subsequent fi nal retreat. Alternatively, the streams in Arctic Canada (Stokes et al., 2009). and Marchant , 2000; Staiger et al., 2006); we stage 1-2 transition occurred during buildup to On this basis, and the contrasting upstream suggest they not only passively block, but con- the last maximum position, fi rst driven by East geometries with which each trough’s fl ow is tribute to northward ice fl ow through McMurdo Antarctic and Victoria Land contributions be- associated, there cannot be a direct, simple re- Sound. Sediment provenance tracing from the fore a retreat, readvance, and associated fl ow lationship between the ice stream troughs of the ANDRILL AND-1B core (McKay et al., 2009; reorgani za tion led Ross Embayment ice to outer continental shelf and the modern-day Siple Talarico et al., 2012) attests to the contribution domi nate the mid- to outer continental shelf Coast ice streams. of ice from Skelton and Mulock glaciers into fl ow geometry at the Last Glacial Maximum. We In the fi nal stages of our reconstruction, local McMurdo Sound. Sandroni and Talarico (2006) note that on the opposite side of the continent, forces come to dominate over regional forces. additionally documented Pleistocene sediment in the Weddell Sea, East and West Antarctic Moraines on both sides of Ross Island suggest contributions to the CIROS-2 core (fronting advances are signifi cantly out of phase. Recent a remnant ice cap was left after regional degla- Ferrar Valley) from Ferrar glacier, the Blue fi eld data (Anderson and Andrews, 1999; Bent- ciation, with thick ice remaining grounded just and Koettlitz glaciers and from the Ross Sea. ley et al., 2010; Hein et al., 2011) and numeri- off the present-day coastline. Drift signatures Such contributions from Victoria Land outlet cal ice-sheet modeling (Le Brocq et al., 2011) on Ross Island itself have, however, been ex- glaciers are neither incompatible with limited suggest that ice advanced in the eastern sector clusively associated with regional Ross Sea terrestrial glacial activity in the Dry Valleys prior to the Last Glacial Maximum, which itself ice rather than local: Sediments contain far- nor with mapped drift distributions and associ- exhibited little difference in confi guration com- traveled erratics and marine shells (Denton and ated reconstructions ascribed to Ross Sea ice pared to today, whereas in the western Weddell Marchant , 2000; Dochat et al., 2000) and have (e.g., Denton and Marchant, 2000; Denton and Sea, modest ice-sheet thickening and grounding an upper drift elevation distribution consistent Hughes, 2000). Enhanced drawdown through line advance are reconstructed with thinning with lateral grounding of regional ice. Denton select routes can be reconciled with intervening commencing at ca. 15 cal. ka. We do not go so and Hughes (2000) inferred confl uence of local mountain blocks that are terrestrial and alpine far as to directly correlate events observed in and regional ice around the periphery of Ross in landscape character, while the most recent the Ross Sea with those proposed in the Wed- Island, consistent with the regional drift distri- stages of our reconstruction are those that would dell Sea, but we simply note that the East and bution. We suggest that the point of confl uence be expected to dominate the landscape imprint, West Antarctic Ice Sheets have been observed to was dynamic through the glacial cycle, and at and these are consistent with drift distributions behave asynchronously in both of the two main the time of uncoupling was located offshore. around McMurdo Sound. confl uence outlets. In the fi nal stage, the remnant Ross Island ice East Antarctic contributions to the western While the recent ANDRILL coring confi rms cap must therefore have been largely cold-based Ross Sea during recent glacial maxima have a long-term dynamic Antarctic ice sheet with re- and geomorphologically inactive, with no sig- been widely invoked (e.g., Hughes, 1977; Den- peated ice-sheet expansion and grounding in the nifi cant local production of sediments and with ton et al., 1989; Shipp et al., 1999 [see Fig. 1A]; Ross Sea throughout at least the Pliocene (Naish limited reworking of the earlier-deposited re- Licht et al., 2005; Farmer et al., 2006), but in the et al., 2007, 2009; McKay et al., 2009), we show gional material. specifi c area of McMurdo Sound–Ross Island, that within even a single cycle, the McMurdo Regional retreat was accompanied by local the prevailing model is one of westward-spread- Sound–western Ross Sea region experiences oscillations of the grounding line off Cape ing Ross Sea ice. Here, we provide evidence major shifts in (grounded) ice-fl ow sources and Crozier (Ross Island) and along the Mackay of a phase of NE fl ow off Ross Island, into the confi guration. Such a marked shift in ice-fl ow stretch of the Victoria Land coastline. We sug- Central Basin, prior to a series of ice-fl ow stages geometry points to the value of a midshelf/mid– gest that the Mackay outlet glacier and nearby dominated by westward ice movement from ice-sheet setting for discerning fundamental fl ow piedmont lobes responded to the loss of region- the Ross Sea. Given these two contrasting ice reorganizations of an ice sheet (cf. Clark, 1993), ally grounded ice, otherwise blocking their geometries, a signifi cant question concerns the likely over longer periods within a glacial cycle, termini, by making a small readvance onto the shift, or transition between states. Grounding in contrast to minor readjustments and shorter- coastal shallow shelf, in a similar manner to zone wedges beneath the fl ow set 2 lineations term fl ow sensitivity close to the margin. Fur- the contemporary loss of buttressing ice shelves (Fig. 3D) imply that a phase of eastward-fl ow- thermore, investigation at the head of ice stream or ice-sheet sectors elsewhere (e.g., De Angelis ing ice preceded westward-fl owing ice, and that troughs reveals the contrasting behavior of dif- and Skvarca, 2003; Payne et al., 2004). Similar the transition between these phases was one of ferent ice streams: Here, for example, in a NE dynamics are observed off Cape Crozier, where margin (grounding line) retreat–fl ow reorgani- fl ow confi guration, Drygalski Trough does not a sequence of overrun grounding zone wedges, zation–renewed grounding line advance. It is draw down signifi cant ice, while under a west- glacial lineations, and draped recessional simplest to assume that these grounding zone ward fl ow confi guration, the Central Basin is by- moraines (Fig. 3B) likely records uncoupling of wedges relating to eastward fl ow were related passed. The upstream landform record indicates regional from Ross Island ice, and subsequent to the phase of NE fl ow from McMurdo and that fast fl ow in the Drygalski Trough and in the restabilization of the respective margin (ground- Victoria Land, recorded by the lineation fl ow outer troughs fed from the Central Basin likely ing) positions. The fact that the only evidence set toward the Central Basin (fs 1; stage 1). We did not operate in concert, but rather these conti- of regional grounding positions comes off Ross may therefore conclude that a retreat–readvance nental shelf troughs exhibited differing behavior Island points to a stabilizing, pinning role of cycle characterized the transition between our at different times. Such out-of-phase behavior the island upon the Ross Sea ice-sheet–shelf two main ice-sheet geometries. Potentially, and fl ow switching between neighboring out- system. The calving front of the Ross Ice Shelf

1746 Geological Society of America Bulletin, November/December 2012 Downloaded from gsabulletin.gsapubs.org on December 29, 2012 Glaciation of the western Ross Sea has been pinned around Ross Island since ca. northeastward from the Victoria Land coast Island particularly signifi cant as a stabilizing 8–9 14C ka (McKay et al., 2008), and the com- and McMurdo Sound into the Central Basin or pinning point for the regional grounding and plete lack of iceberg regrounding features on of the Ross Sea. We must infer, therefore, that calving lines. The glacial and postglacial records the midcontinental shelf area around the island outlet glaciers of the Transantarctic Mountains from such locations are in stark contrast to oth- is likely indicative of lasting ice-shelf develop- were active during the last glacial cycle, in- ers in close proximity, which bear a strong gla- ment with no deep calving close to the ground- cluding through mountain blocks often noted cial landform imprint. McMurdo Sound itself, ing line. This contrasts with an abundance of to be both glaciologically and geomorphologi- in particular, hosts a variety of geomorphologi- iceberg scour features in the outer-western and cally inactive. This activity is invoked both for cal domains, with thick sediments, and evidence more central parts of the Ross Sea, which are in- a period of extensive ice-sheet grounding in the of heavy reworking of material in both glacial terpreted to record abrupt collapse of a fringing Ross Sea (our stage 1) and during deglaciation and postglacial periods. Strong geomorphologi- ice shelf, deep draft calving, and rapid retreat (stage 4), manifested as localized responses cal activity here perversely raises complications (Jakobsson et al., 2011; Shipp et al., 1999). We to the loss of buttressing following retreat of for interpretation of long-term records such as suggest that Ross Island facilitates grounding the regional Ross Sea ice sheet. During other those derived from the ANDRILL cores, due to line restabilization after rapid ice mass loss from stages, we reconstruct a dominant westward the range of infl uences governing the deep sedi- the outer continental shelf, and that the pinning fl ow of ice, inconsistent with strong East Ant- mentary stratigraphy. role of the island encourages renewed ice-shelf arctic outfl ow and suggestive of a greater role We point to the importance of a midcontinen- development followed by steady calving line re- of West Antarctic ice in the Ross Sea. This is tal shelf (and mid–ice sheet) setting for distilling treat to the modern, stable calving line position not to say that West Antarctic ice directly fed the various controls upon mid- to outer-shelf gla- off Ross Island. It would be likely that the island the western Ross Sea at these times, but that ciation. Here, we present a reconstruction from has served such a role through several of the it fi lled central and eastern sectors suffi ciently such a setting that resolves some of the variety most recent cycles recorded in the ANDRILL to defl ect southern Transantarctic Mountain ice of ice-fl ow sources, drawdown patterns, and the cores; conversely, prior to the formation of the around Ross Island and onto the western con- style of retreat. We suggest that the midshelf is a volcanic island in the early Pliocene (see Hor- tinental shelf. We therefore conclude the role valuable target for investigating longer-term and gan et al., 2005; McKay et al., 2009), we may of East Antarctic and West Antarctic ice in the larger-scale patterns of glacial history, which are expect that the stability of the ice-sheet–shelf Ross Sea is time-varying, and potentially not often less apparent in settings distal to the zones system, and its sedimentological imprint, was in phase on a larger scale. Furthermore, our of ice-fl ow confl uence. rather different. observations of ice bypassing drawdown into ACKNOWLEDGMENTS the outer continental shelf troughs at various CONCLUSIONS times negate a simple correspondence between This research was supported by Stockholm Univer- the outer-shelf glacial troughs and the modern sity strategic research grants to Jakobsson. The crew The southwestern Ross Sea, around Ross Siple Coast (West Antarctic) ice stream net- and scientifi c teams for Oden cruises 0708, 0910, and Island and McMurdo Sound, represents a mid- work. The modern confi guration of West Ant- 1011 are thanked, and funding from the Swedish Polar Research Secretariat, Swedish Research Council, and shelf setting at a critical location with respect to arctic ice streams must be a product of interior U.S. National Science Foundation, which supported untangling the varying roles of East and West ice-fl ow reorganization subsequent to the last these expeditions, is gratefully acknowledged. We Antarctic ice in the Ross Sea glacial drainage glacial expansion into the Ross Sea (cf. Licht thank George Denton and an anonymous reviewer, system. Recently acquired, high-resolution, et al., 2005; Farmer et al., 2006). and GSA Bulletin Associate Editor Ben Laabs, for their reviews of the manuscript. multibeam seafl oor data from the icebreaker While entirely enveloped by a regional-conti- Oden are complemented and extended by a nental–scale ice-sheet system, several spots REFERENCES CITED multiresolution composite regional bathymetry within the McMurdo region were/are consider- data set, and together they provide a number ably less geomorphologically active. Terrestrial Alonso, B., Anderson, J.B., Diaz, J.T., and Bartek, L.R., 1992, Pliocene–Pleistocene seismic stratigraphy of the of snapshots of glacial landform systems relat- alpine valley glaciation likely persisted in the Ross Sea: Evidence for multiple ice sheet grounding ing to previous expansion of grounded ice into southern Victoria Land mountain blocks while episodes, in Elliot, D.H., ed., Contributions to Ant- arctic Research III, Antarctic Research Series Volume the Ross Sea. Landforms from subglacial, ice- outlet glaciers drew down the ice sheet around 57: Washington, D.C., American Geophysical Union, marginal, glaciofl uvial, and likely submarine them. Beaufort Island and its associated shallow p. 93–103. domains yield a picture of multiple phases of seafl oor topography exhibit no glacial landform Anderson, J.B., and Andrews, J.T., 1999, Radiocar- bon constraints on ice sheet advance and retreat ice fl ow during the last glacial cycle, associated imprints, in contrast to the immediately adja- in the Weddell Sea, Antarctica: Geology, v. 27, with different ice-fl ow confi gurations. We inter- cent seafl oor; indeed, ice-fl ow patterns seem to p. 179–182, doi:10.1130/0091-7613(1999)027<0179: pret a four-stage ice-fl ow reconstruction con- be diverted around the rise in topography. We RCOISA>2.3.CO;2. Anderson, J.B., and Bartek, L.R., 1992, Cenozoic glacial sisting of early fl ow northeastward into the Ross demonstrate that Ross Island supported an ice history of the Ross Sea revealed by intermediate reso- Sea, two stages of westward-spreading ice from cap likely confl uent with regionally expanded lution seismic refl ection data combined with drill site information, in Kennett, J.P., ed., Antarctic Research the Ross Sea onto the Scott Coast, and a fi nal ice in the Ross Sea, and it was more extensive Series, Volume 56: Washington, D.C., American Geo- deglacial stage, which witnesses uncoupling of than present at the time of deglacial uncoupling, physical Union, p. 231–263. a Ross Island ice cap from regionally retreating yet the onshore glacial geology is dominated by Anderson, J.B., Brake, C.F., and Myers, N.C., 1984, Sedi- mentation on the Ross Sea continental shelf, Antarc- ice and local outlet glacier oscillations. regional Ross Sea drift. We conclude, therefore, tica: Marine Geology, v. 57, p. 295–333, doi:10.1016 The reconstruction we derive from the gla- that during deglaciation, the Ross Island ice cap /0025-3227(84)90203-2. cial landform record demands a contribution was largely cold-based, or minimally geomor- Anderson, J.B., Shipp, S.S., Bartek, L.R., and Reid, D.E., 1992, Evidence for a grounded ice sheet on the Ross of ice into the western Ross Sea from the East phologically active, and incapable of rework- Sea continental shelf during the late Pleistocene and Antarctic Ice Sheet. Such contributions, likely ing earlier deposits. Its physical prominence preliminary paleodrainage reconstruction, in Elliot, D.H., ed., Contributions to Antarctic Research Series from both south of Ross Island and from south- as an obstacle to regional ice fl ow, and perhaps III, Antarctic Research Series Volume 57: Washington, ern Victoria Land, to the west, drove ice fl ow its late-stage, cold-based nature, renders Ross D.C., American Geophysical Union, p. 39–42.

Geological Society of America Bulletin, November/December 2012 1747 Downloaded from gsabulletin.gsapubs.org on December 29, 2012 Greenwood et al.

Anderson, J.B., Shipp, S.S., Lowe, A.L., Wellner, J.S., and Marine Geology, v. 204, p. 203–214, doi:10.1016 p. 223–226, doi:10.1130/ 0091-7613(1996) 024<0223: Mosola, A.B., 2002, The Antarctic Ice Sheet during /S0025-3227(03)00338-4. COLWIR> 2.3.CO;2. the Last Glacial Maximum and its subsequent retreat Farmer, G.L., Licht, K., Swope, R.J., and Andrews, J., 2006, Licht, K.J., Dunbar, N.W., Andrews, J.T., and Jennings, history: A review: Quaternary Science Reviews, v. 21, Isotopic constraints on the provenance of fi ne-grained A.E., 1999, Distinguishing subglacial till and glacial p. 49–70, doi:10.1016/S0277-3791(01)00083-X. sediment in LGM tills from the Ross Embayment, Ant- marine diamictons in the western Ross Sea, Antarctica: Andrews, J.T., Domack, E.W., Cunningham, W.L., Leventer, arctica: Earth and Planetary Science Letters, v. 249, Implications for a Last Glacial Maximum grounding A., Licht, K.J., Jull, A.J.T., DeMaster, D.J., and Jen- p. 90–107, doi:10.1016/j.epsl.2006.06.044. line: Geological Society of America Bulletin, v. 111, nings, A.E., 1999, Problems and possible solutions Graham, A.G.C., Larter, R.D., Gohl, K., Dowdeswell, J.A., p. 91–103, doi:10.1130/0016-7606(1999)111<0091: concerning radiocarbon dating of surface marine Hillenbrand, C.-D., Smith, J.A., Evans, J., Kuhn, G., DSTAGM>2.3.CO;2. sediments, Ross Sea, Antarctica: Quaternary Research, and Deen, T., 2010, Flow and retreat of the late Qua- Licht, K.J., Lederer, J.R., and Swope, R.J., 2005, Provenance v. 52, p. 206–216, doi:10.1006/qres.1999.2047. ternary Pine Island–Thwaites palaeo–ice stream, West of LGM glacial till (sand fraction) across the Ross Em- Atkins, C.B., Barrett, P.J., and Hicock, S.R., 2002, Cold Antarctica: Journal of Geophysical Research, v. 115, bayment, Antarctica: Quaternary Science Reviews, v. 24, glaciers erode and deposit: Evidence from Allan Hills, p. F03025, doi:10.1029/2009JF001482. p. 1499–1520, doi:10.1016/j.quascirev.2004.10.017. Antarctica: Geology, v. 30, p. 659–662, doi:10.1130 Hall, B.L., and Denton, G.H., 2000a, Radiocarbon chronol- Lindén, M., and Möller, P., 2005, Marginal formation of De /0091-7613(2002)030<0659:CGEADE>2.0.CO;2. ogy of Ross Sea drift, eastern Taylor Valley, Antarctica: Geer moraines and their implications to the dynamics Barrett, P.J., and Hambrey, M.J., 1992, Plio-Pleistocene Evidence for a grounded ice sheet in the Ross Sea at the of grounding-line recession: Journal of Quaternary sedimentation in Ferrar Fiord, Antarctica: Sedimentol- Last Glacial Maximum: Geografi ska Annaler, v. 82A, Science, v. 20, p. 113–133, doi:10.1002/jqs.902. ogy, v. 39, p. 109–123, doi:10.1111/j.1365-3091.1992 p. 305–336, doi:10.1111/1468-0459.00127. Marchant, D.R., Denton, G.H., and Swisher, C.C., III, 1993a, .tb01025.x. Hall, B.L., and Denton, G.H., 2000b, Extent and chronology Miocene-Pliocene-Pleistocene glacial history of Arena Barrett, P.J., Adams, C.J., McIntosh, W.C., Swisher, C.C., of the Ross Sea ice sheet and the Wilson piedmont glacier Valley, Quarternain Mountains, Antarctica: Geograf- III, and Wilson, G.S., 1992, Geochronological evidence along the Scott Coast at and since the Last Glacial Maxi- iska Annaler, v. 75, p. 269–302, doi:10.2307/521204. supporting Antarctic deglaciation three million years mum: Geografi ska Annaler, v. 82A, p. 337–363, doi: Marchant, D.R., Denton, G.H., Sugden, D.E., and Swisher, ago: Nature, v. 359, p. 816–818, doi:10.1038/359816a0. 10.1111/1468-0459.00128. C.C., III, 1993b, Miocene glacial stratigraphy and Bart, P.J., Sjunneskog, C., and Chow, J.M., 2011, Piston- Hall, B.L., Denton, G.H., and Hendy, C.H., 2000, Evidence landscape evolution of the western Asgard Range, Ant- core based biostratigraphic constraints on Pleistocene from Taylor Valley for a grounded ice sheet in the Ross arctica: Geografi ska Annaler, v. 75, p. 303–330, doi: oscillations of the West Antarctic Ice Sheet in west- Sea, Antarctica: Geografi ska Annaler, v. 82A, p. 275– 10.2307/521205. ern Ross Sea between North Basin and AND-1B drill 303, doi:10.1111/1468-0459.00126. McKay, R., Dunbar, G.B., Naish, T.R., Barrett, P.J., Carter, site: Marine Geology, v. 289, p. 86–99, doi:10.1016 Hein, A.S., Fogwill, C.J., Sugden, D.E., and Xu, S., 2011, Gla- L., and Harper, M., 2008, Retreat history of the Ross /j.margeo.2011.09.005. cial/interglacial ice-stream stability in the Weddell Sea Ice Sheet (shelf) since the Last Glacial Maximum from Bentley, M.J., Fogwill, C.J., Le Brocq, A.M., Hubbard, A.L., Embayment, Antarctica: Earth and Planetary Science deep-basin sediment cores around Ross Island: Palaeo- Sugden, D.E., Dunai, T.J., and Freeman, S.P.H.T., 2010, Letters, v. 307, p. 211–221, doi:10.1016/j.epsl.2011 geography, Palaeoclimatology, Palaeoecology, v. 260, Deglacial history of the West Antarctic Ice Sheet in the .04.037. p. 245–261, doi:10.1016/j.palaeo.2007.08.015. Weddell Sea embayment: Constraints on past ice vol- Hindmarsh, R.C.A., van der Wateren, F.M., and Verbers, McKay, R., Browne, G., Carter, L., Cowan, E., Dunbar, G., ume change: Geology, v. 38, p. 411–414, doi:10.1130 A.L.L.M., 1998, Sublimation of ice through sediment in Krissek, L., Naish, T., Powell, R., Reed, J., Talarico, F., /G30754.1. Beacon Valley, Antarctica: Geografi ska Annaler, v. 80, and Wilch, T., 2009, The stratigraphic signature of the Brook, E.J., Kurz, M.D., Ackert, R.P., Raisbeck, G., and Yiou, p. 209–219, doi:10.1111/j.0435-3676.1998.00038.x. late Cenozoic Antarctic ice sheets in the Ross Embay- F., 1995, Cosmogenic nuclide exposure ages and glacial Horgan, H., Naish, T., Bannister, S., Balfour, N., and Wilson , ment: Geological Society of America Bulletin, v. 121, history of late Quaternary Ross Sea drift in McMurdo G., 2005, Seismic stratigraphy of the Plio-Pleisto- p. 1537–1561, doi:10.1130/B26540.1. Sound, Antarctica: Earth and Planetary Science Letters, cene Ross Island fl exural moat-fi ll: A prognosis for Mosola, A.B., and Anderson, J.B., 2006, Expansion and v. 131, p. 41–56, doi:10.1016/0012-821X(95)00006-X. ANDRILL Program drilling beneath McMurdo–Ross rapid retreat of the West Antarctic Ice Sheet in eastern Clapperton, C.M., and Sugden, D.E., 1990, Late Ceno- Ice Shelf: Global and Planetary Change, v. 45, p. 83– Ross Sea: Possible consequences of over-extended ice zoic glacial history of the Ross Embayment, Antarc- 97, doi:10.1016/j.gloplacha.2004.09.014. streams?: Quaternary Science Reviews, v. 25, p. 2177– tica: Quaternary Science Reviews, v. 9, p. 253–272, Hughes, T., 1977, West Antarctic ice streams: Reviews of 2196, doi:10.1016/j.quascirev.2005.12.013. doi:10.1016/0277-3791(90)90021-2. Geophysics and Space Physics, v. 15, p. 1–46, doi: Naish, T., Powell, R., Levy, R., Henrys, S., Krissek, L., Nies- Clark, C.D., 1993, Mega-scale glacial lineations and 10.1029/RG015i001p00001. sen, F., Pompilio, M., Scherer, R., Wilson, G., and the cross-cutting ice-fl ow landforms: Earth Surface Proc- Jakobsson, M., Anderson, J.B., Nitsche, F.O., Dowdeswell, ANDRILL-MIS Science Team, 2007, Synthesis of the esses and Landforms, v. 18, p. 1–29, doi:10.1002 J.A., Gyllencreutz, R., Kirchner, N., Mohammad, initial scientifi c results of the MIS Project (AND-1B /esp.3290180102. R., O’Regan, M., Alley, R.B., Anandakrishnan, S., core), Victoria Land Basin, Antarctica: Terra Antartica, Conway, H., Catania, G., Raymond, C.F., Gades, A.M., Eriksson, B., Kirshner, A., Fernandez, R., Stolldorf, v. 14, p. 317–327. Scambos, T.A., and Engelhardt, H., 2002, Switch T., Minzoni, R., and Majewski, W., 2011, Geological Naish, T., Powell, R., Levy, R., Wilson, G., Scherer, R., of fl ow direction in an Antarctic ice stream: Nature, record of ice shelf break-up and grounding line re- Talarico, F., Krissek, L., Niessen, F., Pompilio, M., v. 419, p. 465–467, doi:10.1038/nature01081. treat, Pine Island Bay, West Antarctica: Geology, v. 39, Wilson , T., Carter, L., DeConto, R., Huybers, P., De Angelis, H., and Skvarca, P., 2003, Glacier surge after p. 691–694, doi:10.1130/G32153.1. McKay, R., Pollard, D., Ross, J., Winter, D., Barrett, ice shelf collapse: Science, v. 299, p. 1560–1562, doi: Joughin, I., Tulaczyk, S., Bindschadler, R., and Price, S.F., 2002, P., Browne, G., Cody, R., Cowan, E., Crampton, J., 10.1126/science.1077987. Changes in West Antarctic ice stream velocities: Obser- Dunbar, G., Dunbar, N., Florindo, F., Gebhardt, C., Denton, G.H., and Hughes, T.J., 2000, Reconstruction of the vation and analysis: Journal of Geophysical Research, Graham, I., Hannah, M., Hansaraj, D., Harwood, D., Ross ice drainage system, Antarctica, at the Last Gla- v. 107, no. B11, p. 2289, doi:10.1029/2001JB001029. Helling, D., Henrys, S., Hinnov, L., Kuhn, G., Kyle, cial Maximum: Geografi ska Annaler, v. 82A, p. 143– Laberg, J.S., Guidard, S., Mienert, J., Vorren, T.O., Haf- P., Läufer, A., Maffi oli, P., Magens, D., Mandernack, 166, doi:10.1111/1468-0459.00120. lidason, H., and Nygård, A., 2007, Morphology and K., McIntosh, W., Millan, C., Morin, R., Ohneiser, C., Denton, G.H., and Marchant, D.R., 2000, The geologic basis morphogenesis of a high-latitude canyon; the Andoya Paulsen, T., Persico, D., Raine, I., Reed, J., Riesselman, for a reconstruction of a grounded ice sheet in McMurdo Canyon, Norwegian Sea: Marine Geology, v. 246, C., Sagnotti , L., Schmitt, D., Sjunneskog, C., Strong, P., Sound, Antarctica, at the Last Glacial Maximum: Geo- p. 68–85, doi:10.1016/j.margeo.2007.01.009. Taviani , M., Vogel, S., Wilch, T., and Williams, T., 2009, grafi ska Annaler, v. 82A, p. 167–211, doi:10.1111 Le Brocq, A.M., Bentley, M.J., Hubbard, A., Fogwill, C.J., Obliquity-paced Pliocene West Antarctic Ice Sheet /1468-0459.00121. Sugden, D.E., and Whitehouse, P.L., 2011, Recon- oscilla tions: Nature, v. 458, p. 322–328, doi:10.1038 Denton, G.H., Bockheim, J.G., Wilson, S.C., and Stuiver, M., structing the Last Glacial Maximum ice sheet in the /nature07867. 1989, Late Wisconsin and early Holocene glacial his- Weddell Sea embayment, Antarctica, using numerical Noormets, R., Dowdeswell, J.A., Larter, R.D., Ó Cofaigh, tory, Inner Ross Embayment, Antarctica: Quaternary modelling constrained by fi eld evidence: Quaternary C., and Evans, J., 2009, Morphology of the upper conti- Research, v. 31, p. 151–182, doi:10.1016/0033-5894 Science Reviews, v. 30, p. 2422–2432, doi:10.1016 nental slope in the Bellingshausen and Amundsen (89)90004-5. /j.quascirev.2011.05.009. Sea—Implications for sedimentary processes at the Dochat, T.M., Marchant, D.R., and Denton, G.H., 2000, Lewis, A.R., Marchant, D.R., Kowalewski, D.E., Baldwin, shelf edge of West Antarctica: Marine Geology, v. 258, Glacial geology of Cape Bird, Ross Island, Antarctica: S.L., and Webb, L.E., 2006, The age and origin of the p. 100–114, doi:10.1016/j.margeo.2008.11.011. Geografi ska Annaler, v. 82A, p. 237–247, doi:10.1111 Labyrinth, western Dry Valleys, Antarctica: Evidence Ó Cofaigh, C., Dowdeswell, J.A., and Kenyon, N.H., /1468-0459.00123. for extensive middle Miocene subglacial fl oods and 2006, Geophysical investigations of a high-latitude Domack, E.W., Jacobson, E.A., Shipp, S., and Anderson, J.B., freshwater discharge to the Southern Ocean: Geology, submarine channel system and associated channel- 1999, Late Pleistocene–Holocene retreat of the West Ant- v. 34, p. 513–516, doi:10.1130/G22145.1. mouth lobe in the Lofoten Basin, polar North Atlan- arctic Ice Sheet system in the Ross Sea: Part 2. Sedimen- Licht, K.J., and Andrews, J.T., 2002, The 14C record of late tic: Marine Geology, v. 226, p. 41–50, doi:10.1016 tologic and stratigraphic signature: Geological Society Pleistocene ice advance and retreat in the central Ross /j.margeo.2005.09.014. of America Bulletin, v. 111, p. 1517–1536, doi:10.1130 Sea, Antarctica: Arctic, Antarctic, and Alpine Re- Payne, A.J., Vieli, A., Shepherd, A.P., Wingham, D.J., and /0016-7606(1999)111<1517:LPHROT>2.3.CO;2. search, v. 34, p. 324–333, doi:10.2307/1552491. Rignot, E., 2004, Recent dramatic thinning of largest Dowdeswell, J.A., Ó Cofaigh, C., and Pudsey, C.J., 2004, Licht, K.J., Jennings, A.E., Andrews, J.T., and Williams, West Antarctic ice stream triggered by oceans: Geo- Continental slope morphology and sedimentary proc- K.M., 1996, Chronology of late Wisconsinan ice retreat physical Research Letters, v. 31, p. L23401, doi:10.1029 esses at the mouth of an Antarctic palaeo–ice stream: from the western Ross Sea, Antarctica: Geology, v. 24, /2004GL021284.

1748 Geological Society of America Bulletin, November/December 2012 Downloaded from gsabulletin.gsapubs.org on December 29, 2012 Glaciation of the western Ross Sea

Pritchard, H.D., and Vaughan, D.G., 2007, Widespread ac- Stroeven, A.P., and Prentice, M.L., 1997, A case for Sirius basement clasts and grain detrital modes in ANDRILL celeration of tidewater glaciers on the Antarctic Penin- Group alpine glaciation at Mount Fleming, South Vic- core AND-1B: Global and Planetary Change sula: Journal of Geophysical Research–Earth Surface, toria Land, Antarctica: A case against East Antarctic (in press). v. 112, F03S29, doi:10.1029/2006JF000597. Ice Sheet reduction: Geological Society of America Todd, B.J., Valentine, P.C., Longva, O., and Shaw, J., 2007, Rignot, E., 2008, Changes in West Antarctic ice stream Bulletin, v. 109, p. 825–840, doi:10.1130/0016-7606 Glacial landforms on German Bank, Scotian Shelf: dynamics observed with ALOS PALSAR data: Geo- (1997)109<0825:ACFSGA>2.3.CO;2. Evidence for Late Wisconsinan ice-sheet dynamics and physical Research Letters, v. 35, p. L12505, doi: Stuiver, M., Denton, G.H., Hughes, T.J., and Fastook, J.L., implications for the formation of De Geer moraines: 10.1029/2008GL033365. 1981, History of the marine ice sheet in West Antarc- Boreas, v. 36, p. 148–169, doi:10.1111/j.1502-3885.2007 Sandroni, S., and Talarico, F.M., 2006, Analysis of clast tica during the last glaciation: A working hypothesis, .tb01189.x. lithologies from CIROS-2 core, New Harbour, Ant- in Denton, G.H., and Hughes, T.J., eds., The Last Great Wagner, B., Melles, M., Doran, P.T., Kenig, F., Forman, S.L., arctica—Implications for ice fl ow directions during Ice Sheets: New York, Wiley, p. 319–436. Pierau, R., and Allen, P., 2006, Glacial and postgla- Plio-Pleistocene time: Palaeogeography, Palaeoclima- Sugden, D.E., Denton, G.H., and Marchant, D.R., 1991, cial sedimentation in the Fryxell basin, Taylor Valley, tology, Palaeoecology, v. 231, p. 215–232, doi:10.1016 Subglacial meltwater channel systems and ice sheet southern Victoria Land, Antarctica: Palaeogeography, /j.palaeo.2005.07.031. overriding, Asgard Range, Antarctica: Geografiska Palaeoclimatology, Palaeoecology, v. 241, p. 320–337, Scambos, T.A., Bohlander, J.A., Shuman, C.A., and Skvarca, Annaler , v. 73, p. 109–121, doi:10.2307/520986. doi:10.1016/j.palaeo.2006.04.003. P., 2004, Glacier acceleration and thinning after ice Sugden, D.E., Marchant, D.R., and Denton, G.H., 1993, Webb, P.-N., and Harwood, D.M., 1987, Terrestrial fl ora of shelf collapse in the Larsen B embayment, Antarc- The case for a stable East Antarctic Ice Sheet: The the Sirius formation: Its signifi cance for late Cenozoic tica: Geophysical Research Letters, v. 31, p. L18402, background: Geografi ska Annaler, v. 75, p. 151–154, glacial history: Antarctic Journal of the United States, doi:10.1029/2004GL020670. doi:10.2307/521199. v. 22, p. 7–11. Shepherd, A.P., Wingham, D.J., and Mansley, J.A.D., 2002, Sugden, D.E., Denton, G.H., and Marchant, D.R., 1995a, Webb, P.-N., and Harwood, D.M., 1991, Late Cenozoic gla- Inland thinning of the Amundsen Sea sector, West Ant- Landscape evolution of the Dry Valleys, Transantarctic cial history of the Ross Embayment, Antarctica: Quater- arctica: Geophysical Research Letters, v. 29, p. 1364, Mountains: Tectonic implications: Journal of Geophysi- nary Science Reviews, v. 10, p. 215–223, doi:10.1016 doi:10.1029/2001GL014183. cal Research, v. 100, no. B7, p. 9949–9967. /0277-3791(91)90020-U. Shipp, S., Anderson, J., and Domack, E., 1999, Late Pleisto- Sugden, D.E., Marchant, D.R., Potter, N., Souchez, R.A., Wellner, J.S., Heroy, D.C., and Anderson, J.B., 2006, The cene–Holocene retreat of the West Antarctic Ice Sheet Denton, G.H., Swisher, C.C., III, and Tison, J.-L., 1995b, death mask of the Antarctic ice sheet: Comparison system in the Ross Sea: Part 1. Geophysical results: Preservation of Miocene glacier ice in East Antarctica: of glacial geomorphic features across the continental Geological Society of America Bulletin, v. 111, Nature, v. 376, p. 412–414, doi:10.1038/376412a0. shelf: Geomorphology, v. 75, p. 157–171, doi:10.1016 p. 1486–1516, doi:10.1130/0016-7606(1999)111<1486: Sugden, D.E., Summerfi eld, M.A., Denton, G.H., Wilch, T.I., /j.geomorph.2005.05.015. LPHROT>2.3.CO;2. McIntosh, W.C., Marchant, D.R., and Rutford, R.H., Staiger, J.W., Marchant, D.R., Schaefer, J.M., Oberholzer, 1999, Landscape development in the Royal Society P., Johnson, J.V., Lewis, A.R., and Swanger, K.M., Range, southern Victoria Land, Antarctica: Stability SCIENCE EDITOR: NANCY RIGGS 2006, Plio-Pleistocene history of Ferrar glacier, Ant- since the mid-Miocene: Geomorphology, v. 28, p. 181– ASSOCIATE EDITOR: BENJAMIN J.C. LAABS arctica: Implications for climate and ice sheet stability: 200, doi:10.1016/S0169-555X(98)00108-1. Earth and Planetary Science Letters, v. 243, p. 489– Sugden, D.E., Bentley, M.J., and Ó Cofaigh, C., 2006, Geo- 503, doi:10.1016/j.epsl.2006.01.037. logical and geomorphological insights into Antarctic MANUSCRIPT RECEIVED 21 NOVEMBER 2011 Stokes, C.R., Clark, C.D., and Storrar, R., 2009, Major ice sheet evolution: Philosophical Transactions of the REVISED MANUSCRIPT RECEIVED 10 FEBRUARY 2012 changes in ice stream dynamics during deglaciation Royal Society, ser. A, v. 364, p. 1607–1625. MANUSCRIPT ACCEPTED 25 FEBRUARY 2012 of the north-western margin of the Laurentide Ice Talarico, F.M., McKay, R.M., Powell, R.D., Sandroni, S., Sheet: Quaternary Science Reviews, v. 28, p. 721–738, and Naish, T., 2012, Late Cenozoic oscillations doi:10.1016/j.quascirev.2008.07.019. of Antarctic ice sheets revealed by provenance of Printed in the USA

Geological Society of America Bulletin, November/December 2012 1749