The University of Maine DigitalCommons@UMaine

Earth Science Faculty Scholarship Earth Sciences

6-10-1995 Landscape Evolution of the Dry Valleys, : Tectonic Implications David E. Sugden

George H. Denton University of Maine - Main, [email protected]

David R. Marchant

Follow this and additional works at: https://digitalcommons.library.umaine.edu/ers_facpub Part of the Earth Sciences Commons

Repository Citation Sugden, David E.; Denton, George H.; and Marchant, David R., "Landscape Evolution of the Dry Valleys, Transantarctic Mountains: Tectonic Implications" (1995). Earth Science Faculty Scholarship. 55. https://digitalcommons.library.umaine.edu/ers_facpub/55

This Article is brought to you for free and open access by DigitalCommons@UMaine. It has been accepted for inclusion in Earth Science Faculty Scholarship by an authorized administrator of DigitalCommons@UMaine. For more information, please contact [email protected]. JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 100, NO. B7, PAGES 9949-9967, JUNE 10, 1995

Landscapeevolution of the Dry Valleys,Transantarctic Mountains: Tectonicimplications

David E. Sugden Departmentof Geography,University of Edinburgh,Edinburgh, Scotland

GeorgeH. Denton Departmentof GeologicalSciences and Institute for QuaternaryStudies, University of Maine,Orono DavidR. Marchant• Departmentof Geography,University of Edinburgh,Edinburgh, Scotland

Abstract. Thereare differentviews about the amount and timing of surfaceuplift in the TransantarcticMountains and thegeophysical mechanisms involved. Our new interpretationof the landscapeevolution and tectonichistory of the Dry Valleysarea of the Transantarctic Mountainsis basedon geomorphic mapping of anarea of 10,000km 2. Thelandforms are dated mainlyby their associationwith volcanicashes and glaciomarine deposits and thispermits a reconsmactionof the stagesand timing of landscapeevolution. Followinga loweringof baselevel about55 m.y. ago,there was a phaseof rapid denudationassociated with planationand escarpmentretreat, probably under semiarid conditions. Eventually, downcutting by rivers,aided in placesby glaciers,graded valleys to nearpresent sea level. The main valleyswere floodedby the seain the Mioceneduring a phaseof subsidencebefore experiencing a final stageof modest upwarpingnear the coast. There hasbeen remarkably little landformchange under the stable, cold,polar conditions of the last 15 m.y. It is difficult to explainthe SiriusGroup deposits, which occurat high elevationsin the area,if they are Pliocenein age. Overall,denudation may have removeda wedgeof rock with a thicknessof over 4 km at the coastdeclining to 1 km at a point 75 km inland,which is in goodagreement with the resultsof existingapatite fission track analyses. It is suggestedthat denudation reflects the differencesin baselevel causedby high elevationat the time of extensiondue to underplatingand the subsequentrole of thermaluplift andflexural isostasy.Most crustaluplift (2-4 km) is inferredto haveoccurred in the early Cenozoicwith 400 m of subsidencein the Miocenefollowed by 300 m of uplift in the Pliocene.

Introduction debate,we do not considerthe issuein this paper(but seeDenton et al, [ 1993]). The aim of this paperis to relate geomorphologicalevidence Extending acrossAntarctica, the Transantarctic Mountains of landscapeevolution to thetectonic processes in the Dry Valleys rise to over 4000 m in the Royal SocietyRange. The mountains of the Transantarctic Mountains. Like the Drakensberg consistof gentlytilted blocksof sedimentaryrocks (Devonian- Mountains in southernAfrica and the Great Dividing Range in Triassic )overlying Pre-Cambrian-Devonian Australia the Transantarctic Mountains consist of an eroded basementand are intruded and cappedby Jurassicdolerites and plateau edge characteristicof high-elevationpassive continental basalts. The mountainscomprise the uplifted rift flank of the margins [Gilchrist and Summerfield,1990]. Understandingthe West AntarcticRift System[Behrendt et al., 1991], which has landscapeevolution of the mountainsis importantbecause it can been periodicallyactive since the separationof Antarcticaand be used to constrainvarious geophysical models of the tectonic Australia [Tessensohnand Worrier, 1991]. processesassociated withsuch margins. The tectonic history of Thereis considerableuncertainty about the mount and theTransantarctic Mountains isintimately involved ina debatetiming of uplift and the tectonic processes involved. Apafite aboutthe stability ofthe East Antarctic IceSheet. Although our fissiontrack analysis indicates thatthere has been about 5 km of new interpretationof landscapeevolution has implicationsfor the denudationin the Dry Valleys area since the early Cenozoic; denudationrates were high in the early Cenozoicbefore tailing off 1NowatInstitute forQuaternary Studies, University ofMaine, [Gleadowand Fitzgerald, 1987; Fitzgerald, 1992; Brown et al., Orono. 1994]. This implies that significantlocal relief existedaround 55 m.y. ago and that the rift flank was alreadyelevated with respect to base level, either as a consequenceof thermal uplift or Copyright1995 by theAmerican Geophysical Union. underplatingor as a result of the rifting of a preexisting,high- elevation surface. This interpretationagrees with evidenceof Papernumber 94JB02875. 0148-0227/95 / 94JB-02875$05.00 limited recent surface uplift as inferred from the elevation of

9949 9950 SUGDEN ET AL.: DRY VALLEY LANDSCAPE EVOLUTION

undisturbedsubaerial Pliocene volcanic cones in denudationprofiles from the northernsides of Ferrar and Wright [Wilchet al., 1993a]. A differentline of evidencesuggests that valleys[Fitzgerald et al., 1986] imply that the blockhas behaved there has been substantialsurface uplift of parts of the as one unit since the early Tertiary, thus holding constantone TransantarcticMountains of 1-2 km over the past 2-3 m.y. variable in landscapeevolution. Second,the block contains [McKelveyet al., 1991; Webbet al., 1984; Webband Harwood, outcropsof SiriusGroup deposits at high elevations(2650 rn on 1987]. This view is based on the existence of the remains of Mount Feather), critical evidence which must eventually be temperatevegetation of presumedPliocene age (2-3 Ma) in Sirius accommodatedin any explanation. Third, the valleys containa Group depositsat high elevations. There are some 30 Sirius terrestrial record of surface deposits extending back to the Groupoutcrops in theTransantarctic Mountains flanking the Ross Miocene, thus providingan unprecedentedlylong window into Ice Shelfbetween latitudes 86øS and 78øS and further outcrops in mountainevolution. A disadvantageof a studyrestricted to one the highmountains of SouthernVictoria Land in the Dry Valleys blockis that any conclusionscannot necessarily be extrapolatedto and the Prince Albert Mountains [Denton et al., 1993; Verbers other areas of the Transantarctic Mountains. and Van der Wateren,1992]. The datingof the depositsdepends There have been two main phases of interest in the criticallyon the age and origin of reworkeddiatoms thought to geomorphologicalevolution of the Dry Valleys. The first have originatedin marine basinsin East Antarcticaand to have involved geologistsof Scott's and Shackleton'sexpeditions, been subsequentlytransported to the mountainsby ice. The whoseinitial observationson the great age of the landscapeand biostratigraphicage of certainof the diatomshas been confirmed the modestimpact of presentglaciers make relevantreading to by isotopicdating of a volcanicash layer in the offshoreCIROS-2 this day [Ferrar, 1907; David and Priestley,1914; Wright and core [Barrett et al., 1992]. Severalother supportinglines of Priestley,1922; Taylor, 1922]. The secondphase accompanied evidencefor recentsurface uplift have beencollated by Behrendt U.S. and New Zealand activities after the International andCooper [ 1991],including the "youthfulappearance" of the rift GeophysicalYear in 1957. In numerouspapers elaborating the shoulderand the presenceof offshoreHolocene fault scarps. geologicaland glacialhistory of the area, a generalconclusion Thesevarious views were synthesizedby Fitzgerald[1992] who was that the main valleysowed their origin to earlier stagesof suggestedthat crustal uplift of the TransantarcticMountains in glaciationunder more temperateconditions [e.g., Taylor, 1922; the Dry Valleys area has occurredthroughout the Cenozoicbut Bull et al., 1962; Calkin, 1974a; Denton et al., 1971, 1984; with two acceleratedpulses at 55-40 Ma and 10-0 Ma. Nichols,1971; Selbyand Wilson,1971]. This view is consistent In view of the uncertaintyabout the timingof the uplift, it is with the argumentsof Van der Waterenand Verbers[1992] in difficult to relate the tectonic evolution of the mountains to the northernVictoria Land. In addition, there is recognitionof the widerprocesses of platetectonics. At leasttwo periodsof rifting particularrole of cold desertagents of erosion. For example, have been inferred from study of the embayment Selby [1971, 1974] recognized the prevalence of straight [Tessensohnand Worner,1991; Cooperet al., 1991;LeMasurier rectilinearslopes and tentlikeridges which he attributedto salt and Rex, 1991]. The first phasebegan in the Mesozoic(possibly weatheringand wind deflationcharacteristic of this type of frigid, linked to the separationof Antarcticaand Australia) and was arid environment. The characteristicslope angles were also associatedwith widespreadgraben downfaultingand crustal related to rock strengthcharacteristics [Augustinus and Selby, thinning. The secondbegan in the Eoceneand was associated 1990]. All these slopeswere thoughtto be wearing back the with uplift and tilting of the TransantarcticMountains. One initial, steeper,glacial trough sides and cirques. interpretationof the TransantarcticMountains is thatthey formed This paper puts forward a new interpretationof the by asymmetricrifting associatedwith extensionand subsidence in geomorphology.The key conclusionsrelevant to the landscape the adjacentRoss embayment [Fitzgerald, 1992; Fitzgerald et al., evolution are as folows. 1986]. Accordingto this structuralinterpretation the mountains 1. The landscapeis relict, and with the exceptionof a few representan upperplate margin [Lister et al., 1986, 1991]. Al•o, glaciallandforms there has been negligible slope evolution since flexural effects related to the strong lithospherein East the Miocene. ,together with lateral heat flow from the adjacent 2. The landscapeis essentiallyfluvial, with escarpmentsand extendedWest Antarctic lithosphere,may have played a role planafionsurfaces at high elevationsand fiver valleys (the Dry [Sternand Ten Brink, 1989]. Whicheverprocesses are the most Valleys) dissectingthe mountains. important, one particular problem is why, if there has been 3. Glacial modificationof the landscapehas beenrelatively significantlate Cenozoicuplift, it shouldhave occurred >50 m.y. minor and selective. after the late Mesozoic-earlyCenozoic extensional event which This new interpretationgives an insightinto the longhistory formedthe margin and the adjacent Ross embayment. Here is a of baselevel and environmentalchange, which in mm constrains suite of problemsconcerning the tectonicprocesses associated modelsof the tectonicevolution of thispart of the Transantarctic with passivecontinental margins where geomorphologycan help Mountains. constrainthe varioushypotheses. The Dry Valleyscompose one of the manyblocks within the Approach and Assumptions Transantarctic Mountains [Fitzgerald, 1992]. Bounding The crux of our approachwas landformmapping at a scaleof transformfaults underlieMackay Glacier in the north and Ferrar 1:250,000. The resultingmap has beenpublished, together with Glacier in the south(Figure 1). There are severaladvantages in an explanationof the methodsand landforms,by Denton et al., studyingthe Dry Valleys block. First, similar fission track [1993]. A samplecovering Wright Valley, Taylor Valley, and the SUGDEN ET AL.' DRY VALLEY LANDSCAPE EVOLIZHON 9951

...... 0 10 20 30 ' '"'"'""!'i'!'!:i?i'!':':':':""' • • • I Scale (kin)

N 65 ø

MACKAY

,Shapeless ...... '-'"-'"-'"'"'-'"'""'"'"'-'""-ß

ßß . :.:.:':::::'-:¾::•/•ß . ..-.....:.:.:.:::..'::":'"-'""' • Doorly ,...... -...... •, •.....:.: ' S.....:.:.:::::::::::'"' .. !:!:!:i:!:!:!:!:!:i:i:!:!:!.0.. ::Mt.::::::::::::::• :':':':':':':':': Newall Mr. :::::::::::::::::::::::...... Fleming Labyrir LakeVanda :.!.!.!:!:!::.:-: :1"' ======N •.:

Key ...... :Sessrumnir:.:.:. .. I IGlacier As :.'-'.:.'"...:•IceFree •...... IceFree(over 2000m) LakeBonne '.... :::::::i

GLA C

Quartermain Mountains

Mount Feather

Figure1. Locationmap showing the main valleys and bounding transform faults in the Dry Valleys, Antarctica. The faultsare taken from Fitzgerald [ 1992]. 9952 SUGDEN ET AL.: DRY VALLEY LANDSCAPE EVOLUTION

(7..,0

c)

S H E E T SUGDEN ET AL.: DRY VALLEY LANDSCAPE EVOLUTION 9953

-.

Figure 3. Rectilinearslopes topped by an upper cliff on the flanks of Arena Valley. The patch of remnantregolith mantlein the upperleft containsburied ventifacts and stoneswith desertvarnish. The crosson the slopeto the right showsthe site of an ash fall depositwhich occursin the regolith. The ash (Table 1, site 2) is 8.5 Ma and has survivedon a slopeof 28ø.

QuartermainMountains is reproducedin Figure 2. A key possible that similar slope forms may result from different assumptionin any suchmorphological analysis is that different processes. For example, whereas rectilinear slopes are best slopeforms and associationscan be recognizedand used to infer developedin arid and semiarid environments,there is no reason past environmentalconditions. For example,the implicationis why a glaciershould not maintain an existingslope if it is able to that a landscapeof buttescomprising rectilinear slopes topped by carry away any debristhat falls down the slope. Again, it is easy steepcliffs andrising abruptly above a flat-lyingplain reflectsnot to confusethe arcuate head of a canyon formed under desert only structuralcontrol but also the operationof a particular conditionswith a glacial cirque when, in the case of the Dry assemblageof processes(Figure 3). Rectilinearslopes and cliffs Valleys, glaciersmay partly obscurethe foot of the cliff. Such reflect the balancebetween rock strengthand processeswhich limitations must be borne in mind in the subsequentanalysis. are ableto removeany weatheredmaterial made available [Selby, Nevertheless,each landformis part of a coherentassemblage of 1993]. The surrotmdingplain is the resultof processeswhich are landforms,and this helpsto constrainthe interpretation.Also, in sufficientlypowerful to removeany weathereddebris supplied the Dry Valleys where particular landformsare relatively well from the adjacent slopes. The sharper the angle between dated, the approachcan place firm constraintson the histow of rectilinear slope and plain, the more effective the contrast changesin baselevel and surfaceprocesses. betweenrates of slopeweathering and the transportprocesses on theplain. Theform'of such landscapes is influenced by climate, lithology,and structure. Climatic conditionsmost favorablefor Landforms of the Dry Valleys suchlandforms are semiaridwhere episodic floods and the lack of Severallandform assemblages can be recognized(Figure 4), deep absorbentregoliths mean that transportprocesses dominate and it is useful to summarize the main characteristics. over weathering. In terms of lithology,the presenceof a rock Planation Surfaces such as ,which breaks down into granular material, favorsa sharpcontrast between rates of weatheringand rates of Thereare three main surfaces. The uppersurface, formed on transport.In addition,geological structure can influencedirectly Jurassic dolerim, occurs at altitudes of 2000-2400 m and the formof cliffsand plains; it canalso encourage spring sapping composesthe high groundmost distant from the coast(Figure 5). whichinfluences the locationand clarityof the junctionbetween It surroundsthe headsof the main valleysand is boundedon its valley sideand plain. Slopesin morehumid environments tend to coastwardside by an escarpment500-1200 m high, the prime have more concave and convex elements. This is because of the featureof the landscape(Figure 6). Largeinselbergs, often flat- presenceof more vegetationand a deeperregolith and a whole toppedand boundedby rectilinearslopes, rise 400-700 m above rangeof differentprocesses which occur in the regolith[Carson the surfaceto form the highestsummits in the area, such as and Kirkby, 1972]. Mount Feather(3011 m) andShapeless Mountain (2739 m). The It is important to be aware of the limitations of such an surfacehas a relief of 10-30 m. The lower, intermediatesurface, approach,the most importantof which is equifinality;it is - cut mainlyin Beaconsandstone at altitudesof 1700-1850m, is 9954 SUGDEN ET AL.: DRY VALLEY LANDSCAPE EVOLUTION

$hapeles• Mountain Doorly

Newall Mt.

%%%%

•,C..•,,' ,• / •/.,,.:.,, ,'...'...':•t' / •.1 •-'-----•Up•, ,• • /' .• •surr•c• .,L•• Rectilinearlandscapes Mt. Feather /• ' F •' .• ..:.•lntermediate ! ,,v.•.. • !:':':':':':! trace ßi•?• Rolling slopes and I•./-• •Y•N.•,..• •co•t• su Valley benches 0 10 20• Mountain "-'-Ml, i•dmoat i i i ervalleys km (• Siriusdeposit ned river valleys

Figure4. Map showingthe distribution of themain landscape types and the relationship to theintegrated pattern of rivervalleys. The uppersurface in thewest and its frontingscarp give way to dissectedmountain landscapes toward the coast. patchyand occursmost clearly in the vicinityof the Asgardand upstandinginselberg of Mount Feather. Also, a remnantof the OlympusRanges, whose summitsare remnantsof the upper intermediatesurface cuts acrossgranite basementand dolerite surface.These ranges comprise inselbergs and buttes which show immediatelywest of Mount Newall. However,structure influences a progressivechange from plateau remnant near the main scarpto the morphologyof the surfaces.For example,the uppersurface is isolatedbuttes farther away. A lessextensive surface fringes the underlainby a prominentdolerite sill. Also, irregularitiesin the coastand rises gently from sea level to altitudesof lessthan 200 surface,for examplesurrounding the buttesof the OlympusRange, m. It is separatedfrom the intermediatesurface by a major are the resultof planarioncrosscutting the gentledip of underlying escarpmentand is cutmainly in granite. Inselbergswith rounded thin beds of sandstone. tops,also cut in granite,rise abovethe surface. Thesurfaces andassociated inselbergs andbuttes areerosional Valleys rather than tectonicfeatures. This can be illustratedin the caseof The valleys are the main landformsof the area and extend the QuartermainMountains where an tininterrupted,layered rock from west to east (Figure 4). There are severalnoteworthy structure extends beneath both the upper surface and the features. SUGDENET AL.: DRY VALLEY LANDSCAPEEVOLUTION 9955

w E m Feather 3000--• Shapeless • Obelisk Mt. Newall

1000-1 - - ' t •

N

QuartermainMts. AsgardMts OlympusMts

Glacier • • wngnt • • Valley- Valley• ...... Taylor • ...... 'mJa • ....McKelvey Barwick .. • Uppersurface k-,x--xx-x•Rolling slopes •,,',.',.',.',.',,',,',.',,'qIntermediate surface .•:'•:'-"•-'-'-q'Valley benches

•,,'/'.,'.,"• Coastalpiedmont Figure 5. SchematicCross section showing the relationshipbetween the variouslandscape types: (top) from the inlandupper planation surface to the seaand (bottom)from southto north acrossthe main valleys.

1. The valleys from an arborescentnetwork indicativeof 3. The valleys are cut down to near, or below, sea level. fluvial action. The best exampleis the Victoria Valley System, The Ferrar, Taylor, Debenharnand Mackay Valley mouthsare but the patternis alsoclear in the caseof the Wright and Ferrar below sealevel, while part of the sinuouscentral Taylor Valley is Valleys. All continueto the coast, the Wright and Victoria 90m belowsea level [MudreyandMcGinnis,1975]. Valleysextending beneath the Wilson Piedmont[Calkin, 1974b]. 4. Several deeper valleys have gently sloping, elevated 2. Several valleys are sinuous. The best examplesare valley benches.The bestexamples occur on eitherside of central centralTaylor Valley (Figure7) andlower Wright Valley. Taylor Valley where the bench is often terminatedby a cliff

Figure6. Theupper planation surface bounded by a prominentescarpment in theOlympus Range. The inselbergs andbuttes of theOlympus Range (foreground) are remnants of theupper surface. Shapeless Mountain (background) risesabove the uppersurface. 9956 SUGDEN ET AL.: DRY VALLEY LANDSCAPE EVOLUTION

Figure 7. CentralTaylor Valley viewedtoward the sea showingthe valleybenches mmcated in placesby glacial cliffs. The sinuousnature of the innervalley is broughtout by the shapeof Lake Bonneywhich extends from the snoutof (foreground). A darkvolcanic cone can be seenon the benchto the left.

(Figure 7). Other benchesflank Ferrar and Mackay Valley to be less steep and to lead down to concaveslopes in the valley systemsand parts of Wright andVictoria Valleys. The benchcut bottom. by the Labyrinthat the headof Wright Valley is a particularly Rolling Topography goodexample (Figure 8). 5. All majorvalleys have high-level tributaries. Someof This landscape type, comprising convex-concaveslopes, theseare gradedto the intermediatesurface, as in the Asgard occurs in two situations. The first is on the flanks of the inner Range,and othersare gradedto the level of the valley benches. Taylor and Victoria Valleys, where it lies below 1800 m and the There is someasyrnmetry in that tributariesform a subparallel intermediate surface and yet above the valley benches. The patternon the southernside of upperWright andTaylor Valleys. secondsituation comprises the low-lying hills between the main 6. Taylor, Wright, and Victoria Valleys have a reverse mountainfront and the coastalplain. slope,and present-day meltwater streams drain inland into lakes. Glacial Landforms In thecase of WrightValley the amplitude of thisreverse slope is at least 147 m betweenthe exposedlower valley and the surface Themajor glacial features are straight, cliffed troughs, such of Lake Vanda (Figure8). as thoseoccupied by the Mackayand Miller glaciersand the inlandflanks of FerrarGlacier. Otherglacial cliffs ramcatethe Rectilinear Slopes valleybenches in Taylor,Wright, and lower Victoria Valleys. All Straight slopescutting bedrock at angles of 26-36ø are are markedby their large scale,association with straightened characteristic and occur in three situations. The first includes the valley sectionscomplete with mmcatedspurs, and the cliffed slopesof inselbergs,buttes, canyons, and the main escarpment terminationof the upperrolling or rectilinearridge landscapes (Figure3). Here,the footof therectilinear slope is markedby a (Figure8). CentralWright Valley, which is straightand cliffed, sharpbreak as it givesway to a flat surfaceor valley floor. A hasa broadU-shaped valley cross profile. Landscapesof areal secondassociation is with dissectedmountain ridges and the scouringwhich have been roughened by glacialerosion [Sugden, mountainfront. Ridgecrests are typicallystraight with angular 1974] occurin three main locations:the upper surface,the valley plan viewsand bounded by rectilinearslopes often leading down benches,and the rolling slopesand lOw-lyingplain near the coast. to glaciersin the interveningvalleys. Tors top some ridges. Cirques and couloirsoccupied by small glaciers occur patchily Whenever remnantsof the intermediatesurface occur, the altitude throughoutthe area. Typically, they form small arcuate basins of the adjacentridge crestsis alwayslower than the surface,for excavatedinto a rock rectilinearslope beneath an upperfree face. examplein the Kukri Hills, easternAsgard Range, and the Saint Impressivechannel systems, sometimes over 50 m deep, occur John'sRange (Figure 1). A thirdassociation of rectilinearslopes throughoutthe area. The anastomosingpattern, up-and-down is with the main valleys in the area. The sides of the Ferrar long profiles,ungraded confluences, and associationwith gigantic Glaciervalley seawardof the pointof flotation,Taylor, Wright, potholes all point to the role of subglacialmeltwater in their and the Victoria/Barwick/McKelveyValleys are rectilinear. The formation[Sugden et al., 1991]. They occurnear the headsof the main differencewith elsewhereis that the rectilinearslopes tend main valleys,for example,the Labyrinthin upperWright Valley SUGDEN ET AL.: DRY VALLEY LANDSCAPE EVOLUTION 9957

Figure 8. View up Wright Valley showingLake Vanda(foreground), the valley benchincised by the channelsof the Labyrinth,and the overdeepenedtrough of Wright Valley. The flanks are boundedby the intermediatesurface, bearingthe inselbergsand buttes of the Olympusand Asgard ranges. U.S. Navy photographTMA 1564.

(Figure8), thehead of McKelveyValley, and the Miller glacial evolution. The agesof the oldestsediments and their trough.They are also associated with cols in theAsgard Range, stratigraphic settings are summarizedin Table1, whilethe notablyinSessrunmir Valley. They are cut into both dolerite and relationshipofthe deposits tothe landformes isshown inFigure 2. sandstone. TheDry Valleys themselves arelong known to have been relict Glacialdeposits mantle many of the gentlerslopes and sinceglacial and marine sediments in their seaward mouths are reflecta remarkablycomplex story of deposition,subsequent Miocene in age.At themouth of FerrarGlacier the base of the erosion,and transport [Marcham etal., 1993b,c]. A particularlyCIROS-2 core revealed glaciomarinesediments dated on significantdeposit is theSirius Group. Within the Dry Valleys biostratigraphic grounds as older than 4.5 Ma [Barrett and areait is knownin fourlocations (Figures 2 and4). Threeof Hambrey,1992]. Microfossilsat the baseof the Dry Valleys theseare restricted to theflanks of inselbergswhich rise above Drilling Project (DVDP 11) core drilled into sedimentsat the theupper surface, namely, Mount Feather [Brady and McKelvey, mouth of Taylor Valley are late Miocenein age [McGinnis, 1979], MountFleming, and Shapeless Mountain [McKelvey, 1981' lshman and Rieck, 1992]. The 87Sr/86Srratios on shellsin 1991].The fourth occurs on Table Mountain, overlooking the the Jasonglaciomarine diamicton in central Wright Valley southside of FerrarGlacier [Barrett and Powell, 1982]. The suggestages of 9 + 1.5 Ma, while in situ shellsin the Prospect depositsconsist mainly of tills typicalof warm-basedglaciers, Mesa gravelsare aged5.5 + 0.4 Ma [Prenticeet al., 1993]. All containingfaceted and striated stones within a fine-grainedthese dates suggest that the valleys were already excavated before matrix. beingflooded by thesea during the Miocene. Age of the Landforms The flanksof the valleyscan alsobe shownto be old. In Taylor Valley there are numeroussmall volcanic cones and Dating of rocks and organicremains associatedwith the associatedscotia deposits which have eruptedon to the valley variouslandforms places constraints on the sequenceof landscape benchesand the lower glaciallymoulded slopes (Figure 7). 9958 SUGDEN ET AL.: DRY VALLEY LANDSCAPEEVOLUTION

o a:5 a:5 a:5 +1 +1 +1 +1 +1 +1 +1 +1 +1 c• • c• • SUGDEN ET AL.: DRY VALLEY LANDSCAPE EVOLUTION 9959

• o 9960 SUGDEN ET AL.: DRY VALLEY LANDSCAPE EVOLUTION

Fourteeneruptions range in age from 1.5 to 3.89 Ma and demonstratethat the underlyingbedrock slopes are older thanthis [Wilch et al., 1993b]. Details of three eruptionsare listed in Table 1 (sites 16-18). In Wright Valley there are terminal morainesassociated with tributary glaciersforming loops on its southernflank, which are >3.7 Ma in age [Hall et al., 1993]. Their preservation demonstratesthat there has been little morphologicalchange subsequently. The great age of the slopesof the Quartermainand Asgard rangescan be establishedby their associationwith in situ volcanic ashestrapped in frost cracks, tills, regolith and ash avalanche cones[Marchant et al., 1993a, b, c). The asheshave been dated by laserfusion 4øAr/39Ar analysis of singlefeldspar crystals. There is goodevidence that the ashesaccumulated as a result of directairfall or by avalanchingassociated with the settlingof the ash,rather than by desertaeolian processes. This is demonstrated by the physical characteristicsof the ash, notably the

o oc: 5 concentration,coarse grain size, poor sorting (bimodal) and presenceof angularshards, all characteristicsof primary ashfalls with limited transport [Fisher and Schminke,1984]. It is also confirmedby the distinctgeochemistry of individualdeposits and the consistencyof age of individual crystalswithin a deposit [Marchantet al., 1995]. The ashesare remarkablyunweathered and contain less than 10% clay-sized grains, while individual volcaniccrystals show few signsof chemicalweathering. Dating of unweatheredgrains shows that the depositsare Miocene in age. Table 1 lists the detailsof 15 sites and revealsa rangeof ages from 3.9 Ma to 15.25 Ma. The ashesoccur on high-altitudevalley floors, rectilinear slopes,and even in minor lake hollows. One significantsite is shown in Figure 3, where ash has survived in regolith on a rectilinear slope of 28ø since it was depositedat the time of an eruption8.5 m.y. ago. Other ashesare preservedin thin regolith or in ash avalanchecones. The implicationis that the slopesin the Asgard and Quartermainrange had achievedtheir present detailedform by the mid-Mioceneand have remainedessentially unchangedfor 15 m.y. This conclusionis consistentwith the presence of Pliocene ash in the CIROS-2 borehole in Ferrar Glaciervalley, asrecorded by Barrett et al., [ 1992]. It is important to note the associationbetween Pliocene Sirius Group depositsand the higher surfacesand associated inselbergs.There is an implicationthat thesehigh-level surfaces, like the overlyingdeposits, may be Pliocenein age, as arguedby Van der Wateren and Verbers [1992] in northern . This poses an immediateproblem in that they are distinctly youngerthan the depositsat lower altitudesin the Dry Valleys. At this point in the argumentit is worth recallingthat the dating of the Sirius Groupdeposits relies on biostratigraphiccorrelation of diatoms,supported by the datingof volcanicash in one offshore borehole[Barrett et al., 1992]. At presentthere are no published, direct exposure age or ash dates from the deposit itself. Moreover, there is debate about the age and significanceof the Nothofagusfossils in the deposit[Burckle and Pokras, 1991], the significanceof the diatoms [Burckle, 1995] and the climatic conditionsin the areaat the time of volcaniceruptions 3 m.y. ago [Marchantet al., 1993a]. There is limited but useful informationabout the age of SUGDEN ET AL.: DRY VALLEY LANDSCAPE EVOLUTION 9961

offshoresediments which relates to the story of landscapeinterpretation. First, the presenceof escarpmentsprogressively evolution. Seismic surveysof the offshorebasins reveal the consuminghigher surfacesreflects denudation by parallel slope presenceof gently seawarddipping sediments thought to have retreat. Sucha patternof denudationis well known in semiarid been mainly derived from the rising Transantarctic Mountains environmentswhere the key processis the powerof surfacewash [Daveyand Christoffel, 1986; Cooper et al., 1991]. The CIROS-1 in creatingpediplains and removingregolith from upstanding drill hole penetrated700 m into deltaicsediments associated with slopes[Carson and Kirkby, 1972]. Second,the detailed form of the FerrarGlacier. It reachedonly halfway to the basementbut the slopesis characteristicof semiarid conditions,notably the revealedsediments dating from the early Oligocene(36 Ma) to buttesbounded by rectilinearslopes, for examplein the Olympus earlyMiocene. The implicationis thatthe FerrarGlacier valley Range,and the presenceof box canyonsat the head of a dendritic wasin existenceat the time [Barrettet al., 1989]. Presumably, fiver valley system, as, for example, in , the lower, equallythick depositsreflect deposition during the QuartermainMountains and the Asgard and Olympus Ranges. Eocene. There are minimal thicknesses of Miocene and Pliocene Third, buried desertsurface ventifacts occur within the regolith sediments.The presenceof basementgranite from the earliest found on some rectilinear slopes(Figure 3), a finding which is Oligocenehas been interpretedto mean that some 2.5 km of consistentwi.th an origin under semiarid conditions[Marchant et BeaconSupergroup had been eroded from the adjacentmountains al., 1993b,c]. by about36 Ma. Semiaridconditions may alsohave markedthe initial stages Finally,it is worthnoting the oldestdeposits associated with of downcuttingas suggestedby the rectilinear ridge landscapes the presentlandforms. Jurassicbasalts (Kirkpatrick basalt) and rectilinear slopesof the mountain front. Possibly,further eruptedinto a depressionon a land surfacewhich is preserved downcuttingsaw a change to more humid conditions which todayon an inselbergrising above the uppersurface in the Allan allowed the creationof the rolling slopesin the inner reachesof Hills [Ballance and Watters, 1971]. The Allan Hills are the the main valleys and the gentle slopesof the valley benches,but equivalent of inselbergssuch as Mount Feather and Mount this morphologicalchange could be related to lithology as the Flemingand lie on the northernflank of UpperMackay Glacier valleys increasinglycut into basementrocks. The rivers were (Figure 1). able to grade their main valleys to near sea level some distance from the coast. In the caseof Lake Bonney,which today occupies LandscapeEvolution a fiver-cut,s'muous section of TaylorValley, the valley was close The nature, spatial relationships,and age of the various to sealevel some35 km from the coast(Figure 7). landform associationsprovide the basis for reconstructingthe A change to cooler conditionsis indicated by evidence of sequenceof landscapedenudation and the climatic environments glaciationat a wide variety of altitudes. Cold polar conditions involved. We assumethat the high upper and intermediate have existed since the main elementsof the landscapewere planationsurfaces are the oldestelements of the landscape.Since formed. This is implied by the preservationof the relict regolith inselbergsassociated with the intermediatesurface comprise parts on slopesof 26-30ø for millionsof years[Marchant et al., 1995]. of the uppersurface, it is temptingto suggestthat the lower of the It is also demonstratedby the minimal erosionaleffects of valley two surfacesis younger. However,it is quite possiblethat both glaciersflowing into the main valleys. For example,the glaciers surfaces formed contemporaneouslybut with their detailed originatingin the AsgardRange and flowing into Wright Valley morphologyinfluenced by lithology. The valleys and dissected have failed to excavatetroughs, suggesting that their baseshave rectilinearridge landscapesare cut into the two surfaces.This is remainedbelow the pressuremelting point sincetheir formation. illustratedby the way the surfacesare progressivelydissected They are also fringed by relatively small morainesderived from towardthe coastand by the way the remnantinselbergs and peaks valley side debris,and theserepresent the total load over at least lie at comformableor slightly lower altitudesthan the adjacent 3.7 m.y. [Hall et al., 1993]. Suchsmall debrisloads are typicalof surfacefrom which theyhave been eroded. glaciersin a cold polar environment. The downcuttingseems to be fluvial in origin. This is Relationshipto Structure and Tectonics illustrated by the integrated pattern of the tributaries, the sinuosityof the main valleys,the presenceof valley benches,and There is good evidenceof differential denudation,with a the way thevalleys are graded to nearsea level. The main valleys maximum near the mountain front and less inland. This is well and their associatedrolling slopesand valley bencheshave been displayedby the tilt of the basementsurface of Ordovicianor incisedmost deeply, as is illustratedby the way tributaryvalleys Devonianage, known asthe Kukri "peneplain",which is assumed are left hangingabove the main valley floors. Finally, glacierice to havebeen approximatelyhorizontal at the time of its formation has coveredall slopesfrom the highestupland surfaceto valley and is now overlain by Beacon Supergroupsedimentary rock floors and the coastal lowland. Glacial erosion has been [Gunnand Warren, 1962]. It is exposedin the innermostreaches significant in straighteningand deepeningpreexisting sinuous of Taylor and Wright valleyswhere it is now at an altitude of 700 valleys but has been minimal on someof the higher interfluves, m and 900 m respectivelyand dips inland at 2-3ø. It rises in suchas the AsgardRange, where forms due to glacial erosionare discreteblocks towards the coast,reaching altitudes of 1800 m in rare. the vicinity of Mount Newall. The main mountainfront itself is The intermediatesurface reflects scarp retreat, probably downfaulted [Fitzgerald, 1992]. Uniform rectilinear slopes undersemiarid conditions, as arguedby Denton et al., [1993]. It trimcatethe fault lines and are associatedwith valley incision of is useful to summarizethe argumentsin favor of such an the mountainfront. Also, the samefault zone cuts acrossTaylor 9962 SUGDEN ET AL.: DRY VALLEY LANDSCAPE EVOLUTION

w intermediatesurface in the vicinity. At the coastwhere there is 1600 - relief of 900 m in basement rocks, the denudation exceeds 4.0 1400- km. These different methods of calculating amounts of ._. 1200- denudationagree closely. Together,they suggestthat erosionhas •E1000- removedan irregularwedge of rock more than 4 km thick at the z o 800 coastand taperingto lessthan 1 km 75 km inland. < 6OO '" 400' Model of LandscapeEvolution '" 200- It is possible to bring together the various strands of SL- MT TERMINATION evidenceand to presenta hypothesisof the evolutionof the Dry 2000• I BULLPASS • I. I SPIKECAPE Valleys which synthesisesgeomorphology, tectonics and climate. 10O0 f:/!•i•:''' "' ii•MT iiiiil JASON'-•ii•ii::::" MT. THESEUS"'•'""""'""''•1 IMT.DOORLY I We suggest the area begins as one flank of an extensionalrift COAST (Figure 10a). The continentaledge has an altitudeof 1200 m. Figure9. Crosssection of theTrans antarctic Mountains parallel Admittedly,this is arbitrary,but 1200 m seemsa realisticaltitude to WrightValley showing differential uplift since 55 Ma, after for the interior of a large continent,especially in view of the Ten Brink and Stern [1992]. The uplift is representedby a pre- presence of extensive flood basalts which could indicate upliftisochron related to a coastaldatum. The isochron is derived underplating[Lister et al., 1991]. Also, assumingno differential from apatitefission track analysisand representsthe partial tectonicmovement, this figure has been calculatedto have been annealingzone before uplift [Gleadowand Fitzgerald, 1987]. the initial altitude of western South Africa on the basis of the volumeof offshoresediments [Gilchrist and Sutmru•rfield,1994]. The 1200 m altitude is higher than the 500 m assumedby Valley and tnmcatesthe valley benchesof the interior. These Gleadowand Fitzgerald [1987], but this latter value is admitted relationshipssuggest that the faultingoccurred before and during to be poorly constrained;it is based on the view that the low- thephase of valleydowncutting, but not subsequently. gradient floodplain beneath the Kirkpatrick basak would have The overall patternof the differentialcrustal uplift can also been closeto sea level and on analogoussituations in Greenland. be inferredfrom fissiontrack analysisand is shownin Figure9. Such argumentsdo not necessarilycarry weight in escarpment The profile,which runs at right anglesto the coastthrough Mount landscapes.Under suchcircumstances, it is perhapshelpful to Doorly, representsthe distortionof a preuplift, approximately bear in mind both estimates as an indicator of the level of horizontal, isochronin relation to sea level at the coast [Gleadow uncertaintyinvolved in thisinitial assumption. and Fitzgerald, 1987]. The downthrow of the faults at the Following extension, scarp retreat associated with mountainfront is of the orderof 1650 m, approximatelythe same pediplanationextended inland from the new baselevel createdat as the difference in altitude between the intermediate surface and the new coastand alongthe main river valleys(Figure 10b). The the coastalpiedmont. In view of this approximatecoincidence, upperand intermediatesurfaces may be the result of scarpretreat the coastal piedmont could be a downfaultedremnant of the from two differentbase levels, perhaps reflecting the original base intermediate surface. level at the time of rifting and a subsequentbase level fall as a It is possibleto use the tectonicand structuralrelationships resultof someuplift. Alternatively,they may reflectthe dominant to estimate the amount of denudation involved. One estimate role of a major dolerite sill in interruptingthe slope of a single comesfrom the fissiontrack analysis. Sincethe originalisochron phaseof base level change. Either way, the pediplainseroded is thoughtto haveformed in the partial annealingzone at a depth down to granitebasement at the coastand removedat least 3 km of around4-5 km, its exposurealong the crest of the mountain of coverrocks in this location. Modest lowering occurredon the front implies denudationof around4-5 km, with progressively uppersurface, perhaps sufficient to createinverted relief whereby lesseramounts inland [Fitzgerald, 1992]. Anotherestimate comes the Jurassicbasaks, originally extruded in depressions,now fromthe relative elevation of the Kulcripeneplain. Assuming that cappedsome of the inselbergsrising abovethe surface,as in the the Kukri peneplainwas horizontaland covereduniformly with , caseof theAllan Hills today. The contrastbetween the high rates 3.1 km of youngerrocks (2.5 km of the BeaconSupergroup of erosionthat would be expectedalong the scarp and the low [Barren, 1981] and about 600 m of dolerite sills [Hamilton et al., ratesof erosionon flat surfacesis well describedby Gilchrist and 1965;Gleadow and Fitzgerald,1987], thenone can estimatethe Summedeld[1990]. amountof denudationfrom placeto place. In Wright Valley the Figure 10c showsthe situationfollowing downcutting of the Kukri peneplaindeclines from 1650-1800m at the mountainfront major fiver valleys. Valleys have cut into the mountainscausing to 1050 m in the vicinityof Mount Jasonand 700 m in upper most fretting toward the sea. The intermediatesurface is tilted Wright Valley where it disappears beneath dolerite. back and has virtually no seawardslope, while the valleys are Extrapolatingthe samedip would suggestit is at an altitudeof graded to sea level. Apparently, much of the downcutting 100 m beneathMount Fleming,an inselbergrising to 2400 m. If accompaniedfaulting and rock uplift of the mountainfront, which theseapproximate figures are usedand the altitudeof the present would accountfor a furtherfall in relativebase level. It is likely summits is allowed for, then there has been denudation of 0.6 km that glacierscontributed to the widening and straighteningof aboveMount Fleming,1 km from the uppersurface and over 3.1 somemajor valleys. km from above the summits of the mountain front and the There followeda periodof apparentsubsidence sufficient to SUGDEN ET AL.: DRY VALLEY LANDSCAPE EVOLUTION 9963

W

m Kirkpatrick Tectonics Denudation Age 2000] basalt • Sea process (Ma) extension (a) 10000u•er s• and new -1000• --'• -- Kulcri"peneplain" base level

-3000 -

30004_ . _. originals_ur•_a•ce 2000 -- u••m•iint•ermedaaterock uplift, semi-arid 55 (b) -"'uppe; ...... ,..--" 0 faulting? planation

Fleming Obehsk Newall "-x upper m intermediate, 2000- -...... coastal continued valley down- ?->15 (c) rock uplift cuttingby and faulting rivers and glaciers

Fleming Obehsk Newall I'-x upper m intermediate ß 2000 ...... coastal rock fiords, >9 - <3.5 subsidence polar desert, >400m cold-based _,000 ; ½/ I glaciers

Fleming Obehsk Newall renewed polar desert, < 3.5 - 0 uplift no rivers, (e) 2000lOOO cold-based 0 0 glaciers -1000 0 10 20 30 40 50 60 70 km

%XXx. Kukri 'peneplain' /•, sill/ basalt

• fjords Figure10. Empiricalmodel of landscapeevolution in theDry Valleys area of theTransantarctic Mountains showing thesituation (a) afterinitial extension, (b) dlxringplanation with a lowerbase level, (c) aftervalley downcutting, (d) duringa subsequentphase of subsidence,and (e) followingfurther uplift of 300 m at the mountainfront. The possibleages of thestages of evolution,the associated climate, and tectonic environments are also shown.

causethe main valleysto become'inundated by the sea, even in Subsequently,there was a phaseof reneweduplift alongthe sinuous valley sections devoid of significant glacial mountainfront (Figure 10e). This was sufficientto drain the overdeepening,such as inner Taylor Valley (Figure 10d). In fjords and to create reverseslopes in both Wright and Taylor Wright Valley, morainemorphology suggests that the inundation Valleys. Continuity of slopes across mountain front faults was300 m higherthan sea level at present[Hall et al., 1993]. If suggeststhat thisrecent uplift hasnot reactivatedany faults. one allowsfor the marine sedimentinfill in fluvial partsof Taylor The age of the various stagesof landscapeevolution is Valley, thenthe subsidenceamounted to >400 m. constrainedon the onehand by apatitefission track evidence that 9964 SUGDEN ET AL.: DRY VALLEY LANDSCAPE EVOLUTION rapid denudationbegan 55 m.y. ago [Fitzgerald,1992; Gleadow explainthe minimumthickness of late Mioceneand Pliocene and Fitzgerald, 1987]. This implies that the planationsurfaces sediments in the CIROS-1 drill hole immediately offshore andvalleys formed after this date. On the otherhand, denudation [Barrett et al., 1989]. was largely completeby 15 Mal as indicatedby the age of the surficialdeposits on valley-sideslopes (Figure 10c). The timing Wider Implications of the subsidenceof the Dry Valleys block to an intervalbetween Althoughthe model of landscapeevolution is tentative and >9 and <3.5 m.y. ago is constrainedby the age of marine incomplete,it has the merit of integrating the sequenceof sedimentsin the mouthsof Wright, Taylor, and Ferrar Valleys landscapeevolution with tectonic evidence and producing a (Figure 10d). combinedinterpretation which is consistent. There are several The ages attributed to the stagesof landscapeevolution in importantimplications for our understandingof the evolutionof the model raise one major difficulty, namely, the Sirius Group passivecontinental margins. depositsthought to have been emplaced less than 2-3 m.y.ago by The landform evidencesuggests that since continental the mainice sheet. If so, thentheir preservation only on the separation,denudation has removed a wedgeof rockover 4 km oldestrelict elements of thelandscape implies that the planation thickat thecoast and 1 km thick75 km inland.At thesame time, and valley incisionhas occurredsubsequently. However, the assumingan initial elevation of 1200 m, the land surface presenceof olderdeposits on theselower dissected slopes and elevationhas been lowered by 1 km in the downfaultedcoastal valleyspresents a difficulty;it impliesthat there has beenno zoneand has beenraised by 0.8 km on the rift shoulderof the significantlandscape modification in the last 15 m.y. So it is mountainfront andby 1 km at the headof the valleys75 km necessaryto find an alternativeexplanation for the age and inland.If theinitial elevation is correct,then crustal uplift since locationof the SiriusGroup deposits. At presentwe haveno riftinghas been at least3 km at the downfaultedcoast, 4 km at satisfactoryexplanation, but it is usefulto highlighttwo further themountain front, and 2 km75 km inland.These latter figures argumentssuggesting that the SiriusGroup deposits in the Dry agreewith the view that uplift has been greateston the rift Valleys cannot be Pliocenein age. First, we have shown shoulderand has declinedprogressively inland [Gleadowand elsewherethat it is possibleto fix the limit of thoseEast Antarctic Fitzgerald,1987; Fitzgerald, 1992; Ten Brink and Stern, 1992]. ice sheetoutlet glaciers flowing into TaylorValley duringthe The smaller magnitudeof our figures is the result of the Pliocene[Denton et al., 1993;Marchant et al., 1994]. Well- difference in assumed initial elevation. constrainedreconstructed pi'ofiles based on theselimits do not Our reconstructionof landscapeevolution places some permitice to be sufficientlythick to coverthe adjacentSirius constraintson the tectonicprocesses involved. Rates of Groupdeposit on Mount Feather. Thus it is difficultto envisage denudationwere high in the early Cenozoicand tailed off to the glaciologicalconditions that couldexplain this particular virtuallynothing by 15Ma. Thechange from high to lowrates of SiriusGroup deposit. Second, the lack of weatheringin Miocene denudationduring the Cenozoic is mostsimply explained by a volcanicashes, their associationwith cold polar soil wedges, and progressivelydecreasing fall in baselevel followingan initial their preservationin sensitivelocations point to continuedcold change.The initial rift allowedplanation to extendinland from desertconditions since the mid-Miocene [Marchant et al., 1993c]. the new coast. Subsequentvalley downcutting(and some Thiswould seem to precludethe possibility of themild conditions planation?)was associated with crustal uplift and faulting at the necessaryto explainthe remainsof temperatevegetation in the mountainfront. Uplift latergave way to subsidence,as reflected SiriusGroup deposits. by the incursion of the Miocene/Pliocenesea into the lower The climatic environmentassociated with landscapevalleys to levels300 m abovethose at present.The survivalof evolution in our model is based on the dominant landform ancientriver valleys graded close to present sea level in thispart associations.We have arguedthat planationand at least the of theTransantarctic Mountains suggests that subsequent uplift initial valley cuttingprobably occurred in semiaridconditions hasbeen limited. This latter conclusion agrees with Wilch et al., withscanty regolith and vegetation and yet sufficient flash floods [1993a],who used the presence of volcaniccones in TaylorValley to favor pediplanation[Denton et al., 1993]. Although to suggestthat surface uplift canhave been no morethan 300 m in downcuttingmay have been wholly related to baselevel changes thelast 2.54 m.y. asa resultof uplift,it is possiblethat it couldbe partially caused Thispattern of denudationand uplift is consistentwith the bythe onset of wetterconditions. In sucha casethicker, regolith evolutionof an upperplate passive mountain range. The initial wouldproduce convex-concave slopes and favor incision of fiver Cenozoicbase level change agrees with the viewof Fitzgerald valleys,aided perhaps by intermittentglaciation. A morehumid [1992]that the Transantarctic Mountains form the upthrown block climateduring the early Oligocene to earlyMiocene is consistent of an asymmetricrift. Suchasymmetry explains the existence of with the resultsof the nearbyCIROS-1 drill hole whichfound anuplifted flank which creates a largeinitial base level change leavesof Nothofagusof Oligoceneage [Barrett et al., 1989]. The and permitsscarp retreat from the new coast. The evidenceof coldpolar desert conditions of the last 15 m.y. associatedwith crustaluplift soon after rifting, followed by subsidence,points to little signof runningwater help to explainthe preservation and therole of thermalprocesses. Typically, thermal uplift by lateral association of volcanic ashes with cold climate landforms heatconduction or secondaryconvection would be expectedto [Marchantet al., 1993b].The lack of waterwould also explain leadto initialuplift of 500-1500m [Summe•eld,1988]. Such whythe main rivers were unable to excavate their valleys in spite upliftcould follow rifting after a lagof 10-20m.y. and decay with of modestrates of surfaceuplift in thelast 15 m.y. It wouldalso time,eventually leading to subsidence[Steckler, 1985; Fleitout et SUGDEN ET AL.: DRY VALLEY LANDSCAPE EVOLUTION 9965 al., 1986]. Such a processwould explain the crustal uplift Craddock,pp. 1059-1067, Universityof WisconsinPress, Madison, associatedwith faulting in the early Cenozoicand the subsidence 1982. of several hundred meters in the Miocene. Barrett,P.J., M.J. Hambrey, D.M. Harwood, A.R. Pyne and P.N. Webb, Anotherprocess likely to haveplayed a role is flexural uplift Synthesis,in Antarctic Cenozoic History From the CIROS-1 in responseto isostaticunloading. Gilchrist and Summerfield Drillhole, McMurdo Sound, editedby P.J. Barrett, DSIR Bull. N2., [1990] notehow this canbe expectedto lead to long-termsurface 245,241-251, 1989. uplift, initially at the coast but migrating inland with the Barrett,P.J., C.J. Adams,W.C. Mcintosh,C.C. Swisher,and G.S. Wilson, progressionof the escarpment.Such an effecthas been calculated Geochronologicalevidence supporting Antarctic deglaciation three to accountfor about 600 m of uplift along part of the western million yearsago, Nature, 359,816-818, 1992. marginof southernAfrica. In the Dry Valleys, it could account Behrendt,J.C. and A. Cooper,Evidence of rapid Cenozoicuplift of the for a comparableamount of uplift in the early Cenozoic. Yet shoulderescarpment of the CenozoicWest Antarcticrift systemand a anotherprocess of uplift could be glacio-isostaticand could be speculationon possibleclimate forcing, Geology,19, 315-319, 1991. relatedto a forebulgebeyond the marginof the East Antarcticice Behrendt,J.C., W.E. Lemasuder,A.K. Cooper,F. Tessonsohn,A. Trehu, sheet. Althoughthe amplitudeof suchan effectis lessthan about and D. Damaske,Geophysical studies of the West Antarcticrift 100 m [Stern and Ten Brink, 1989], it could have been system, Tectonics,10, 1257-1273, 1991. contributedto someof the uplift sincethe Miocene. Brady,H., Late Cenozoichistory of Taylor and Wright Valleysand McMurdoSound inferred from diatomsin Dry Valleys Drilling Conclusion ProjectCores, in AntarcticGeoscience Symposium on Antarctic Geologyand Geophysics,edited by C. Craddock,pp. 1123-1131, This view of modestmountain uplift since the Plioceneand Universityof WisconsinPress, Madison, 1982. limited geomorphicactivity sincethe mid-Miocenediffers from Brady,H., and B. McKelvey, The interpretationof a Tertiarytillire at the alternativeview of 1-2 km of uplift since the Pliocene and Mount Feather, southern Victoria Land, Antarctica, J. Glaciol., rather dynamicchanges in glaciationand vegetation,as suggested 29(102), 189-193, 1979. elsewhere [Webb and Harwood, 1987; Webb et al., 1984; Brown,R.W., M. A. Summerfield,and A. J. W. Gleadow,Apatite fission McKelveyet al., 1991; Barrett et al., 1992]. The apparentyouth trackanalysis: Its potentialfor the estimationof denudationrams and of the Siriusgroup deposits is the onepiece of evidencethat does implicationsfor modelsof long-termlandscape development, in not fit our geomorphologicalreconstruction. We haveno accepted ProcessModels and TheoreticalGeornorphology, edited by M. J. explanationfor the apparentage of the Sirius Group deposits. Kirkby,pp. 23-53,John Wiley, New York, 1994. We would simply note that our geomorphicreconstruction is Bull, C.B.B.,B.C. McKelvey,and P.N. Webb,Quaternary glaciations in underpinnedby a considerablenumber of absolutedates and that SouthernVictoria Land, Antarctica, J. Glaciol.,4(31), 63-78, 1962. it agreeswell with all otherknown geophysical evidence. While Burckle,L.H. A criticalreview of themicropalaeontological evidence used we recognize that we have studied only one block of the to infera majordrawdown of theEast Anatarctic Ice Sheetduring the TransantarcticMountains, the conclusionsreached here may have earlyPliocene, in Climateand Evolution,edited by E.S. Vrba, Yale wider implications. UniversityPress, New Haven,Conn., in press,1995. Burckle,L.H., and C.M. Pokras,Implications of a Pliocenestand of Acknowledgments. The researchwas supportedby the Divisionof Polar Nothofagus(Southem Beech) within 500 kilometres of the South Programsof theNational Science Foundation, the U.K. NaturalEnvironment Pole,Antarct. Sci., 3, 389-403, 1991. ResearchCouncil, the RoyalSociety, and the Royal Society of Edinburgh. Calkin,P.E., Processesin the ice-freevalleys of southernVictoria Land, We are very gratefulto Paul Fitzgeraldand Michael Surmnerfieldfor valuablecomments on an earlydraft, and to four refereeswhose constructive Antarctica,in Researchin Polar andAlpine Geornorphology, edited commentsimproved the manuscript markedly. by B.D. Faheyand R.D. Thompson,pp. 167-186, Departmentof Geography,University of Guelph,Ontario, Canada, 1974a. References Calkin,P.E., Subglacialgeomorphology surrounding the ice-freevalleys of southernVictoria Land, Antarctica,J. Glaciol.,13(69), 415-429, Augustinus,P.C., and M.J. Selby, Rock slopedevelopment in McMurdo 1974b. Oasis,Antarctica, and implications for interpretationsof glacial history, Carson,M.A., and M.J. Kirkby,Hillslope Form and Process,375 pp., Geogr.Annlr. Stockholm,72A (1), 55-62. 1990. CambridgeUniversity Press, New York, 1972. Ballance, P.F., and W.A. Watters, The Mawson diamictite and the Cooper,A.K., F.J. Davey, and K. Hinz, Cmstalextension and originof Carapacesandstone. Formations of the FerrarGroup at Allan Hills sedimentarybasins beneath the Ross Sea and Ross Ice Sheif, and Carapace Nunatak, Victoria Land, Antarctica, N.Z.J. Geol. Antarctica,in GeologicalEvolution of Antarctica,edited by M.R.A. Geophys.,14, 512-527, 1971. Thomson,J.A. Crane.and J.W. Thomson,pp. 285-291, Cambridge Barrett,P.J., Historyof the RossSea region during the depositionof the UniversityPress, New York, 1991. BeaconSupergroup 400-180 million yearsago, J. R. Soc.N. Z., 11, Davey,F.J., and D.A. Christoffel,Correlation of seismicreflection dam, in 447-458, 1981. AntarcticCenozoic History From the MSSTS-1Drillhole, McMurdo Barrett,P.J., and M.J. Hambrey,Plio-Pleistocene sedimentation in Ferrar Sound,edited by P.J. Barrett,DSIR Bull.N.Z., 237, 405-412, 1986. Fiord,Antarctica, Sedimentology, 39, 109-123,1992. David,T.W.E., and R.E. Priestley,Glaciology, physiography, stratigraphy Barrett,P.J., and R.D. Powell, Middle Cenozoicglacial beds at Table and tectonicgeology of South Victoria Land, in British Antarctic Mountain, Southern Victoria Land, in Antarctic Geoscience Expedition1907-09, Reports on the ScientificInvestigations, Vol. 1, Symposiumon AntarcticGeology and Geophysics,edited by C. Geology,319 pp, Heinemann,London, 1914. 9966 SUGDEN ET AL.: DRY VALLEY LANDSCAPE EVOLUTION

Denton,G.H., R.L. Armstrong,and M. Stuiver,The Late Cenozoicglacial Res.Ser., vol. 56, editedby J.P. Kennettand D. A. Wamke,pp. 327- historyof Antarctica,in The Late CenozoicGlacial Ages,edited by 347, Washington,D.C., 1992. ICIC Turekian, pp. 267-306, Yale University Press, New Haven, Lemasurier,W.E., and D.C. Rex, The Marie Byrd Land volcanicprovince Conn.,1971. and its relationto the CainozoicWest Antarcticrift system,in The Denton,G.H., M.L. Prentice,D.E. Kellogg,and T.B. Kellogg,Late Tertiary Geologyof Antarctica, edited by R. J. Tingey,pp. 249-284, Oxford historyof the Antarcticice sheet:Evidence from the Dry Valleys, UniversityPress, New York, 1991. Geology,12,263-267, 1984. Lister,G.S., M.A. Etheridge,and P.A. Symonds,Detachment faulting and Denton,G.H., M.L. Prentice,and L.H. Burckle,Cainozoic history of the the evolutionof passivecontinental margins, Geology, 14, 46-250. Antarcticice sheet, in The Geologyof Antarctica,edited by R. 1986. Tingey,pp. 365-443,Oxford University Press, New York, 1990. Lister,G.S., M.A. Etheridge,and P.A. Symonds,Detachment models for Denton,G.H., D.E. Sugden,D.R. Marchant,B.L. Hall and T.I. Wilch, East the formationof passivecontinental margins, Tectonics,10, 1038- AntarcticIce Sheetsensitivity to Plioceneclimatic change from a Dry 1064, 1991. Valleys perspective, Geogr. Annlr. Stockholm,75A(4), 155-204, McGinnis,L.D. (Fxl.),Dry ValleysDrilling Project,Antarct. Res. Ser., vol. 1993. 33,765 pp.,AGU, Washington,D.C., 1981. Ferrar,H.T., Reporton thefield geologyof the regionexplored during the Marchant,D.R., C.C. SwisherIII, D.R. Lux, D.P., West, Jr. and G.H. 'Discovery' Antarctic Expedition, 1901-4, National Antarctic Denton,Pliocene palcoclimate and East Antarctic ice-sheet history from Expedition 1901-4, Natural History, vol. 1, Geology, British surficialash deposits, Science, 260, 667-670, 1993a. Museum, London, 1907. Marchant, D.R., G.H. Denton, and C.C. Swisher III, Miocene-Pliocene- Fisher,R.V. and H.U. Schminke,Pyroclastic Rocks, Springer-Verlag, New Pleistoceneglacial history of ArenaValley, QuartennainMountains, York, 1984. Antarctica,Geogr. Annlr, Stockholm, 75A (4), 269-302, 1993b. Fitzgerald,P.G., The TransantarcticMountains of SouthernVictoria Land: Marchant,D.R., G.H. Denton,D.E. Sugden,and C.C. SwisherIII, Miocene The applicationof fission track analysisto a rift shoulderuplift, glacialstratigraphy and landscapeevolution of the WesternAsgard Tectonics,11,634-662, 1992. Range, Antarctica,Geogr. Annlr. Stockholm,75A (4), 303-330, Fitzgerald, P.G., M. Sandiford, P.J. Barrett, and A.J.W. Gleadow, 1993c. Asymmetricextension association with uplift and subsidenceof the Marchant, D.R., G.H. Denton, J.G. Bockheim, S.C. Wilson, and A.R. Kerr, TransantarcticMountains and RossEmbayment, Earth Planet. Sci. Quaternarychanges in level of upper Taylor Glacier, Antarctica: Lett., 81, 67-78, 1986. Implicationsfor palaeoclimateand East Antarctic Ice Sheetdynamics, Fleitout,L., C. Froidevaux,and D. Yuen, Active lithosphericthinning, Boreas,25, 29-43, 1994. Tectonophysics,132,271-278, 1986. Marchant, D.R., C.C. Swisher III, N. Potter, and G.H. Demon, Antarctic Gilchrist, A.R., and M.A. Summeffield, Differential denudation and palcoclimatereconstructed from volcanicashes in the Dry Valleys flexural isostasyin the formationof rifted-marginupwarps, Nature, region,southern Victoria Land, Geol. Soc. Am. Bull., in press,1995. 346, 6286, 739-742, 1990. McKelvey, B.C., The Cainozoicglacial record in southVictoria Land: A Gilchrist,A.R., andM.A. Summeffield,Tectonic models of passivemargin geologicalevaluation of the McMurdoSound drilling projects, in The evolutionand their implicationsfor theoriesof long-termlandscape Geologyof Antarctica,edited by R.J. Tingey,pp. 434-449, Oxford development,in ProcessModels and TheoreticalGeomorphology, UniversityPress, New York, 1991. editedby M.J. Kirkby,pp. 55-84,John Wiley, New York, 1994. McKelvey, B.C., P.N. Webb, D.M. Harwood and M.C.G. Mabin, The Gleadow,A.J.W., and P.G. Fitzgerald,Uplift historyand structureof the DominionRange Sirius Group: A recordof the late Pliocene-early TransantarcticMountains and new evidencefrom fissiontrack dating Pleistocene Beardmore Glacier, in Geological Evolution of of basementapatites in the Dry Valleysarea, Southern Victoria Land, Antarctica, edited by M.R.A. Thomson,J.A. Crane, and J.W. Earth Planet. Sci. Lett., 82, 1-14, 1987. Thomson,pp. 675-682, CambridgeUniversity Press, New York, Gunn, B.M., and G. Warren, Geology of Victoria Land between the 1991. Mawson and Mulock Glaciers,Antarctica, Bull. N. Z. Geol. Surv. 71, Mudrey,M.G., and L.D. McGinnis,Dry Valleys Drilling Project,Bull. 5, 157pp., 1962. North. Ill. Univ., De Kalb, 1975. Hall, B.L., G.H. Denton,D.R• Lux and J.G. Bockheim,Late Tertiary Nichols,R.L., Glacialgeology of the Wright Valley, McMurdoSound, in Antarcticpaleoclimate and ice-sheetdynamics inferred from sufficial Researchin the Antarctic, edited by L.O. Quam, pp. 293-340, depositsin WrightValley, Geogr.Ann!r. Stockholm, 75A(4), 239- American Association for the Advancement of Science, 268, 1993. Washington,D.C., 1971. Hamilton, W., P.J. Hayes, R. Calvert, V.C. Smith, S.D. Elmore, P.R. Prentice,M.L., J.G. Bockheim, S.C. Wilson, L.H. Burekle,D.A. Hodell, C. Barnett,and N. Conklin,Diabase sheets of theTaylor Glacier region, Schluchter,and D.E. Kellogg,Late NeogeneAntarctic glacial history: VictoriaLand, Antarctica, U.S. Geol.Surv. Prof Pap.,456B, 71 pp., Evidence from central Wright Valley, in The Antarctic 1965. Paleoenvironment:A Perspectiveon Global Changes,2, Antarc.Res. Harwood, D.M., Siliceousmicrofossils, in AntarcticCenozoic History Ser.,vol. 60, editedby J.P. Kennettand D. A. Wamke,pp. 207-250, Fromthe CIROS-1 Drillhole,McMurdo Sound, edited by P.J.Barrett, AGU, Washington,D C., 1993. DSIR Bull. N. Z., 245, 67-97, 1989. Selby, M.J., Slopesand their developmentin an ice-free, arid area of Ishman,S.E. and H.J. Rieck, A Late NeogeneAntarctic glacio-eustatic Antarctica,Geogr. Annlr. Stockholm,53A(3-4), 235-245, 1971. record, Victoria Land Basin margin, Antarctica,in The Antarctic Selby,M.J., Slopeevolution in an Antarcticoasis, N. Z. Geogr., 30(1), Palcoenvironment:A Perspectiveon GlobalChange, Part 1, Antarct. 18-34, 1974. SUGDEN ET AL.: DRY VALLEY LANDSCAPEEVOLUTION 9967

Selby,M.J., Hillslope Materials and Processes, 451pp., Oxford University Van Der Wateren, F.M., and A.L.L.M. Verbers, Cenozoic glacial geology Press,New York, 1993. and mountainuplift in northernVictoria Land, Antarctica,in Recent Selby,M.J., and A.T. Wilson,Possible Tertiary age for someAntarctic Progressin Earth Science,edited by Y. Yoshidaet al., pp. 707-714, cirques,Nature, 229,623.24, 1971. TerraScientific, Tokyo, 1992. Steckler,M.S., Uplift and extensionat the Gulf of Suez;indications of Verbers, A.L.L.M., and F.M. Van Der Wateren, A glacio-geological inducedmantle convection, Nature, 317, 135-139,19•85. reconnaissanceof the southern Prince Albert Mountains, Victoria Stem,T.A., and U.S. Ten Brink,Flexural uplift of the Transantarctic Land, Antarctica,in RecentProgress in Earth Science,edited by Y. Mountains,J. Geophys.Res., 94, 10315-10330,1989. Yoshidaet al., pp. 715-719. TerraScientific, Tokyo, 1992. Sugden,D.E., Landscapesof glacial erosion in Greenlandand their Webb, P.N., and D.M. Harwood, The Sirius formation of the Beardmore relationshiptoice, topographic andbedrock conditions, in Progress region, Antarct.J. U.S.,22, 8-12. 1987. in Geomorphoiogy,edited by R.$. Waters and E.H. Brown, Spec. Webb, P.N., D.M., Harwood,B.C. McKelvey, J.H. Mercer, and L.D. Stott, Publ.Inst. Brit. Geogr.,7, 177-195,1974. Cenozoic marine sedimentation and ice volume variation on the East Sugden,D.E., G.H. Dentonand D.R. Marchant,Subglacial meltwater Antarcticcraton, Geology, 12,287-291, 1984. channelsystems and ice sheet overriding, Asgard Range, Antarctica, Wilch, T.L., D. IL Lux, G.H. Denton, and W.C. Mcintosh, Minimal Plio- Geogr.Annlr. Stockholm, 73A(2), 109-121, 1991. Peistocene surface uplift in the Dry Valleys sector of the Sugden,D.E., D.R. Marchant, N. Potter,R. Souchez,G.H. Denton, C.C. TransantarcticMountains, Geology, 21,841-44, 1993a. Swisher,III, andJ-L. Tison, Miocene glacier ice in BeaconValley, Wilch, T.L., G.H. Denton, D.R. Lux, and Mcintosh, W.C., Limited Pliocene Antarctica,Nature, 1995. glacierextent and surfaceuplift in middle Taylor Valley, Antarctica, Summerfield,M.A., Globaltectonics and landform development, Prog. Geogr.Annlr. Stockholm,75A(4), 331-351, 1993b. Phys.Geogr., 12,389-404, 1988. Wright,C.S., andR.E. Priestley,Glaciology, British Antarctic (Terra Nova) Taylor,G., ThePhysiography of McMurdo Sound and Granite Harbour Expedition1910-13, vol. 2, 581 pp., Harrison,London, 1922. Region, BritishAntarctic (Terra Nova) Expedition, 1910-13, Harrison,London, 1922. G.H. Denton and D. R. Marchant,Institute for Quaternary Ten Brink,U., andT. Stem,Rift flankupfiifts and hinterland basins: Studies,University of Maine, Orono,ME 04469. Comparisonof the TransantarcticMountains with the great D.E. Sugden, Departmentof Geography,University of escarpmentof southem Africa, J. Geophys.Res., 97, 569-585, 1992. Edinburgh,Drummond Street, Edinburgh, EH8 9XP, Scotland.(e- Tessensohn,F.,and G. Womer,The Ross Sea rift system, in Geological mail: [email protected]) Evolutionof Antarctica,edited by M.R.A.Thomson, J.A. Crane, and J.W.Thomson, pp. 273-277. Cambridge University Press, New York, (ReceivedNovember 2, 1993; revisedOctober 26. 1994; 1991. acceptedNovember 2, 1994.)