Late Pleistocene Interactions of East and West Antarctic Ice-Flow Regimes: Evidence from the McMurdo Ice Shelf

We present new in terpretations of deglaciation i n M cM urdo Sound and the western Ross Sea, with observationally based reconstruct ions of interactions between East and West Antarct i c ice at the last glacial maximum ( LGM), 1 6 000, 1 2 000, 8000 and 4000 B P . At the LG M, East Antarct ic ice from M ulock Glacier spli t; one branch turned westward sout h of R oss Island b u t the other branch rounded R oss I sland before flowing southwest in to M cMurdo Sound. This flow regime, constrained by an ice saddle north of Ross I s land, is consisten t wi th the reconst ruc t ion of StuiYer and others (198Ia). After the LGl\1, grounding-l ine retreat was most rapid in areas wit h greatest water depth, especially along the Vic toria Land coas t . By 1 2 000 BP, the ice-flow regime in r..,1cMurdo Sound c hanged to through-flowing tv1u lock Glacier ice, wi th lesser con tr ibutions from Koe t t li tz , Blue and Ferrar G laciers, because the former ice saddle north of Ross Island was replaced by a dome. The modern flow regime was es tablished ",4000 BP. I ce derived from high elevations on the Pola r Plateau but now s tranded on the l\IIclV[urdo lee Shelf, and the pat tern of the Transant arct ic Mountains erratics support our reconstruct ion s of Mulock Glacier ice rounding Minna Bluff but wi t h a l l ice from Skelton G lacier abla ting south of the bluff. They are inconsistent wit h Drewry's ( 1 97 9) LGM reconst ruc t ion that includes Skelton Glacier ice in the McMurdo Sound through-flow. D rewry's ( 1 97 9) model closely approximates our results for 1 2 000-4000 BP. Ice-sheet modeling holds promise for det e rm ining whether deglaciat ion proceeded by groundi ng-line retreat of an ice sheet tha t was largely s tagnanr, because i t ne\'er approached equilibri um flowline profiles a ft er the Ross I ce Shelf grounded, or of a dynamic ice sheet wi th flowline profiles kep t low by active ice s treams that extended northward from present-day outlet glaciers after the Ross Ice S helf grounded. INTRODUCTION glacial geologic observat ions and the observed presen t ­ day ice dynamics of the M cMurdo I ce Shelf (MIS), a part of the Ross I ce S helf. The reconstruct ions aid in elucidat ing modern glaciological processes operating on this ice shelf and suppor t suggestions by Kellogg and others ( 1 990) that some existing M I S featu res a rc remnants from a former ice-flow regime. M I S dat a permi t evaluation o f two differenr publ ished reconstruc­ t ions of the f low rcg ime du ring t h e LG M , a n d development of four n e w reconstructions for the R oss Sea embayment during the deglacial transit ion from thc grounded late Wisconsin R oss Sea ice sheet to the modern Ross I ce Shelf, all of which include ice derived from both East Antarc tica and \Ves t Antarctica. One key objective of the West An tarctic ice sheet (VV A I S) ini t iat i\'e is docu mentation of eusta tic sea-level change resu l ting from the possible dis integration of the \Vest An tarctic ice sheet (Bindschad ler, 1 991 ). This e ffort req uires detailed knowledge of ice-sheet history i n \Vest Antarctica, inclu ding the history of grounding-l ine and calving-margin retreat to the posi t ions now occupied by these feat ures and an accurate dynamic overview of the 'vV es t An tarctic ice sheet, i ncl udi ng presen t and former precipi tation sources, flow regimes and changing i nteract­ ions wi th ice derived from East An tarctica. The develop­ ment of reliable p redictive models for fu ture ice-sheet retreat depends on these data. Yet , so far most VV AIS s tudies have concent rated either on modern ice dynamics in \Vest Antarc tica or on elucidat ing former ice maxima. This paper p resents a series of preliminary observ­ a tionally based reconstructions of interactions bet ween t h e East and West Antarctic ice sheets during and s ince t h e last glacial maximum (LGM). These two ice sheets are separated by a dividing wall ( t he Transan t arct ic Mountains) and i nteract today primarily where through­ flowing outlet glaciers from East Antarctica impinge on the Ross Ice Shelf. Our reconst ruct ions are based on 486 West Antarctica i s a key area for s tudying glacial history and dynamics, b ecause the world's largest extan t marine-based ice shee t is located t here. Of particu lar importance are the Dry Valleys of sout hern Victoria Land, s i tua ted between the seasonally open waters of t h e Ross S e a and McMu rdo Sound on the e a s t and the East Antarct ic ice sheet to the west. These i ce-free valleys are shielded from direct E as t Antarctic ice-sheet influence by the Transantarctic M o untains. I t can be argued that a th ird m ajor factor, in addition to East and \'Vest Antarct i c ice flu c tu ations, opera tes in this area : local glaciers Kellogg and others: Late Pleistocene illteractiolls oJ Eas! alld 11 'est Antarctic ice-j1ow regimes respond to the proximi ty of an open-water moisture source in t h e Ross Sea, This i nfluence is actually controlled by \Vest Antarctic ice extent, be ing insignif­ icant during glacial maxima and reaching i ts greatest importance during interglacials when shelfand sea-ice limits are least extensive, The availabili ty of a local moisture source is important for this paper primarily with respect to fluctuat ions of Koett l i tz and Skelton Glaciers. A detailed la t e Quaternary paleoclimatic record for the Dry Valleys region (Denton and others, 1 97 1 , 1 989; St uiver and others, 1 98 1 a) shows that t h e dominant features were periodic advances of the 'Vest An tarctic ice sheet, which th ickened and grounded in the R oss Sea and M cM u rdo Sound, pushing lobes of grounded ice into the Dry Valleys and leaving a record of ice-dam med lakes, marginal and recessional moraines and raised beaches. Numerous conven t ional 14e dates on mari ne shells and terrestrial fresh-water algae s uggest a late Qua ternary age of ",1 8-20 BP for the most recent adl'ance of West Antarc t ic ice (Ross Sea i ce sheet) . D rewry ( 1 979) suggested t ha t grounded ice extent was cons iderably less extensive t han portrayed by S tuiver and others ( 1 98 I a) and hence tha t their ice-flow t rajectories were u nlikely. In particular, Drewry ( 1 979) called for north ward flo\l· of grounded ice past Ross Island in Mc:'lurdo Sound, with no westward flow into the Dry Valleys. The mechanism and h istory of deglaciat ion following the late VVisconsin ice maxim um in West Antarc t ica is a significant glacial geologic problem, i,wolving ice thin­ ning and grounding-line ret reat of the R oss Sea ice sheet in the central and outer R oss Sra, and flot at ion to form the Ross I c e Shelf in the central and southern R oss Sea . MclVr u rdo Sound is of particular importance for u nder­ standing th i s t ransition, because apparent remnants of former grounded ice are presen·ed as part of the ;'lIS (Kellogg and others, 1 990). GLACIAL GEOLOGY OF THE McMURDO ICE SHELF The \I I S is an extension of the Ross Ice S helf that occupies sou thern McMurdo Sound (Stuar t and Bull, 1 963). I t i s bounded on the sou th by Minna Bluff, on the north by R oss Island and the seasonally open waters of Mc;'I urdo Sound, on the west by !\fount Discol'ery and the Dry V alleys of southern Victoria Land, and on the east by the R oss Ice Shelf and White Island (Figs I and 2). Koett l i t z Glacier flows into the MIS west of Brown Peninsula . The eastern part of the �I IS is dominated by net annual surface accumulat ion, like the remainder of the Ross Ice Shelf, but the western part is ch aracterized by surface ablat ion (Swithinbank, 1 970) . Prominent bands and patches of sediment debris man tle the MIS surface in th is area. The following extended discussion concentra t es on the \\·estern ablation area and is based largely on Kellogg and others ( 1 990). The MI S debris bands comprise sediment (ranging in size from clay to boulders bu t dominated by sand and gravel), m arine biotic remains (including: s i li ceous sponges, m olluscs, barnacles, corals, bryozoans, Foram­ inifera, dia toms and fish), and non-marine algae and diatoms t h a t li\"e in seasonal melt ponds of the M I S surface. Debenham ( 1 9 1 9) proposed t h a t marine sed<J Coulmon I. t :, ROSS SEA � ",finD Nova OOl/id GJ \ .. jJL. Boy : 1" ;" " : DrYc;lolski Ice Tongue Mowson GI. Cope _ Franklin I. a: Hiekey eeQu'ort I. <f c • '" I-\ Fig. i. index //Zap showing locations mentiolled 1I1 text. See Figure 2 Jor details oJ the M/S. S)lmbols: B = Blue Glacier; CB = Cape Bird: eR = Cape Ro)'ds; F = Fenar Glacier: A" = A"oel/lit::. Glacier: MP = J\Jarb/e Point. i ment and biota were incorporated into the l\US by basal adfreezing, to ('merge el·entually on the ;'US su rface because of net a n nual surface a blation. Subseq uent workers, who have reported biotic remains on t h e MJS, include: Debenha m , 1 949, 1 96 1 ; Swi th inbank and others, 1 96 1 ; Speden, 1 962; Gow and o thers , 1 965; Kellogg and others, 1 977 ; Brady, 1 978; H ayward and Taylor, 1 984; Kellogg, 1 98 7 ; H oward-\\' illiams and others, 1 989 , 1 990; H art, 1 990; S uren, 1 990. GOvlI and Epstein ( 1 9 72), Stuiver and ot he rs ( 1 98 1 b) and Kellogg and o thers ( 1 990) used oxygen-isotopic analyscs of l\I!S ice to show tha t lowermost Koettli tz Glacier and most ice i n thl' surface-ablat ion z one formed from frozen sea water, supporting Debenham's hypothesis . I ce-thickness d a t a obtained by rad io echo-soun ding (Swithinbank, 1 9 70) and drilling (Gow and Epstein, 1 972) are shown in Figure 3. Thicknesses between IV I ount Disco\"ery and Black Island and north of \i(inna Bluff exceed 1 00 m. F u rt h er north, most th icknesses range from 20 to 50 m. Th i s i nformation, t ogether wi th observed surface-ablation r a tes (0 .51 .0 m year I; Gow, 1 967 : personal communicat ion from A. L. DeVries, 1 98 7 ), suggests that ice in the northern part of the l\i(I S Illust be less than 1 00 years in age b u t the presence of fossil shells (He ages :::;7 750 BP; all dates in this paper are uncorrected) suggests that basal freezing and surface a blat ion must be nearly in e q u ilibrium. T h e 1\IlS probably did not disintegrate completely during the H olocene because 14e dated shells have prog,"essively older ages towards the north along debris bands.