J ournal oJ GLaciology, VoL. 42, No . 142, 1996

On Mertz and Ninnis Glaciers, East Antarctica

GERD WENDLER, KRlSTINA AHL:\fAs AND CRAIG S. LH\CLE Geophysical Institute, Unive rsity aJ Alaska- Fairbanks, Fairbanks, Alaska 99775-7320, U.S.A.

ABSTRACT. Two large glacier tongues, which extend substa ntiall y across the coastline of King G eorge V Land in Eas t Antarctica, ha\'e been studied by remote sensing (sy nthetic aperture radar, JERS-I ). The to ngue of Mertz G lacier is in a state of advance, wh ile the Ninnis Glacier tongue is retreating. Some mo re speci fi c points are: The distin ctive surface structure a nd th e form of the glacier tongues indicates th a t they are Ooating. While th e tongue o f Ninnis G lacier has lost abo u t two-thirds of its a rea since 191 3, the M ertz Glacier tongue has ad vanced subs ta n tia ll y and has a bo u t doubled its areal ex tent over the sa me time p eriod.

The annual movement of th e tong ue of ~Vl e rl Z G lacier was determined as about 1.2 km. This is close to th e value of the advance o f the tip of the tong ue since 1963, which was determined as 0.9 km year- I.

INTRODUCTION the freq uency and in tensi ty of th e ka tabatic winds whi ch occ u r in th is area. H e gave a popular accoul1l in his book M ertz and Ninnis G laciers are located in Ki ng George V entitled The /taill e oJ the blizzard (M awson, 191 5). The Land, Eas t An tarctica (Fi g. 1). They are prominen t winds he experi enced - m ean annual speed of 19. 6 m s I features ; th e tongue of M ertz Gl acier, presently the la rger - are the stronges t found anywhere on pl a net Earth close of the two, ex tends a bou t 95 km into the ocean and has a n to sea le ve l. His da ta were doubted b ut more recent average width of a bo ut 40 km . This a rea of Antarctica obse rva ti o ns with au tom atic stations, which reco rd over was first explored by Sir Douglas ~1 awso n during his satellite , have co nfirmed hi s findin gs (Stearns a nd heroic trip between 19 11 and 19 14. M awson es tablished Wendler, 1988). his main cam p a t Cape Den iso n (67.02° S, 142.68° E ) in Cape D eniso n is located close to the South ~I agne ti c Commonwealth Bay. H e and hi s team were surprised by Pole a nd durin g the summer of 1912- 13 two trips were

1400 E 1450 E 150· E

Fig. 1. Area map oJ King GeOlge V Land, East Antarctica.

Downloaded from https://www.cambridge.org/core. 30 Sep 2021 at 01:01:56, subject to the Cambridge Core terms of use. 447 Joumal of Glaciology

made towards the eas t. " The eastern coastal party" was resolution of 240 m. U nlike satellite im agery that is under the leadership of C. T. M agid an. H e and hi s two d epend ent on radiation of the \'isible or thermal IR companions went by foot from Cape Denison pulling a range in which the surface can be seen on ly in the absence heavy sled. They crossed the tongue of M ertz Glacier, of clouds, radar images are unaffected by weather and which was difficult because of many crevasses. At the darkness. SAR is therefore suitable for use in the polar point of their crossin g, which was relatively close to th e regions during winter and in regions characterized by coastline, they found a maximum eb'ation of 53 m. This hea"y cloud cover. altitude is in good agreem ent with later measurements. Velocity measurements were made using the ASF The Australian map (SQ55-56) of this area, which was Interactive Image Analysis System (IIAS) , which iss ued in 19 71 and based on data from 1961- 62, gave a provides network access to the Arctic Data Visualization maxim u m al ti tude of 49 m for the same general area. and Analysis L aboratory (ADV AL) and the Arctic After crossing l\IIertz Glacier, they co ntinued over the sea R egion Supercomputing Center (ARSC). Image-proces­ ice a nd crossed the tongue of Ninnis Glacier, where they sing software on the IIAS incl udes the Land Analysis determined a maximum elevation of 55 m. Then they System (LAS), International Imaging System's (IIS) we nt on further to explore the coastal region east of System 600 , PCI and Khoros. Ninnis Glacier and eventua ll y returned to Cape Deniso n. During their trip, they ran low on food and supplemented their supplies with penguins, whi ch are plentiful in this LONG-TERM CHANGES area; they reported them to taste quite good . The second party to the east, but a t a more inland ?\IIawson's team was th e first group to map the King course, was "the far eastern party" under the leadership George V Land coas t, using measurements obtained of M awson. His two compa nions were Dr X. l\IIertz and between 1911 a nd 1914. The general outline of the Lieut. B. E. S. Ninnis. They crossed Mertz G lacier further tongues of Mertz and Ninnis Glaciers is included in his inland a nd also reported very crevassed a reas. After map (Mawson, 19 15, p. 240 ). The Australians published having crossed Ninnis G lacier, in an area about 150 km to maps on a scale of I : I 000000 of this area in 197 1 (SR55- the east, Ninnis and his dog team met th eir deaths in a 56) and in 1974 (SQ55-56) (Australia 1971, 1974) based large crevasse. This was even more tragic as his sled on observations carried ou t in 1961 and 1962. In 1993, carried most of the food supply, whi ch they were unable SAR imagery of this area was acquired by JERS-l. to retrieve. The trip home was a race against starvation; Figure 2 shows the outlines of the glacier tongues from all Mertz perished and Mawson himself barely survi\·ed . three sou rces. Table I shows the areas of the glac ier lvIawson named these two glaciers a ft er hi s fallen tongues sea ward of the coastline. com rades . In the case of Mertz Glacier, the size of the tongue has increased substa ntially during this century. Tt has more th an doubled its size, from a bout 3830km2 in 1913 to SA TELLITE IMAGES about 8100 km2 at present. During the earli er p a rt of this century, the glacier tongue m ostly broadened but, sin ce The Alaska SAR Fac ility (ASF), which is operated under 1962, the glacier tongue has advanced abou t 26 km or contract with NASA by the Geophysical Institute, about 0.9 km. year I . University of Alaska- Fairbanks, is the only satellite sy nthetic aperture radar (SAR) receiving station in th e United States. Data from the first and second European Remote Sensing Satellites (ERS-I and ERS-2) and from the J apanese SAR satellite UERS-I ) are received and processed here. ERS-I and 2 do not have on-board tape recorders, so SAR image d ata mu st be d own-linked in real time to a receiving antenna that is "visi ble" to the 67 S~-----'~~----rT--~~--.~.,----~•• ~------~ satellites; that is, th e satelli tes ca n acq uire data only within the station masks of the receiving anten nas. JERS- l has a n on-board tape record er and can therefore acquire SAR data worldwide, which are then down­ linked when the satellite is within a station mas k. Our area of interest was not within the sta tion mask of any ground station for ERS- I and 2, alth ough it is now within ~ S~------r------~+L--~~~~----~ the station mask of the recently operational ~vI cMu rd o antenna . (Also, a receiving station is now under King George V Land construction at Hobart, Tasmania.) The SAR data employed in this study were acquired ove r E ast Antarctica by JERS-I ; the data were su bseq uen tly down-linked and processed at ASF. The SAR on board

JERS- I is L-band (1.27 GHz; wavelength = 0.235 m ) 144' E 146' E 148 0 E and has a swath width of 75 km , a pixel spacing of 12.5 m in full resolution a nd a resolution of 30 m. Low­ Fig . 2. The positions of the N/ertz and Ainnis Glacier reso lu tion im ages have a pixel spacing of 100111 and a tongues in 1912- J3, 1962 and 1993.

Downloaded448 from https://www.cambridge.org/core. 30 Sep 2021 at 01:01:56, subject to the Cambridge Core terms of use. Wendler and others : On NJert;;. and JVinnis Glaciers, East Antarctica

Table 1. Areas of Mert;;. and Ni1lnis Glacier tongues in dumping it over the cres t on to the proglacial sid e in the 1913, 1962 and 1993 style of a bulldozer . This process can graduall y ad,·ance the terminus out over the length of a rela ti,·ely d eep fjord on a tim e-scale of the order of 1000 years (Meier a nd P ost, 1913 1962 1993 198 7); i. e. at a mean advance rate of roughl y 10- 100 m year 1 km2 % km 2 % km2 % In th e case of the M ertz Glacier tongue, however , Figure 2 shows that the terminus ad vanced a bout 26 km be tween 1964 a nd 1993; i. e. a t a m ean rate of a bo u t M ertz 900 m year- I. Compa ri so n of SAR im ages acquired Glacier 3830 100.0 5920 154.6 8 100 2 11. 3 during 1993 a nd 1994 (Fig. 3) yielded an advan ce rate Ninnis of 1200 m year- I, which suggests tha t the aycrage ad van ce G lacier 6060 100.0 3970 65 .5 2 150 35.4 ra te during 1964-9 3 was less because of modes t, episod ic Both iceberg caking . The rela ti ve ly hig h ra te of ad,·ance glaciers 9890 100.0 9890 100.0 10250 103.6 during both time period s, combined with an a bsen ce of m o rphology a nd crevasse pa tte rns cha racteristi c of compressive Oow a t the calvin g p oint, implies tha t the M ertz Gl acier tong ue is Ooating fo r its entire length a nd The beh avior of th e M ertz Glacier tongue is in sharp tha t th e adva n ce m echa nism is m ore likel y to be contras t to Ninnis Glacie r. H ere, we o bsen ·e a n a na logous to the Erebus Glacie r tongue, Anta rc tica ex traordina ry retreat since 19 13; during the 80 year time (H old sworth, 19 74 ) , tha n to Al aska n yord glaciers. Tha t peri od , the glacier tongue re treated 110 km a nd its a rea is, the Ooating glacier tongue is adva n cing at a ra te eq ual was reduced to a bout one-third of its ori gin a l size. M os t of to the ve locity of the ice stream w here it crosses the the retreat w as obsen ·ed during the earli er p a rt of this grounding line, plus a n additiona l increment of velocity century. During 1913-62, it re treated about 90 km; since due to longitudina l strain ra tes induced by hydrosta tic 1962, retreat has continued. imbala nce a t the ice front, less the calving rate. La rge d ecreases in size can res ult fr o m episodi c cal ving, if a glacier tongue is Ooating. Som e depth measurem ents a re available fr om th e wes t sid e of th e tongue of NI e rtz Gl acier, where a water d epth of a bout 350 m was fo und, while the w a ter at th e tip o f the tongue was onl y a bo ut 250 m deep. A lesse r depth in fr ont of a glacier tong ue is fr equently found. Glaciers tra nsport large qua ntities of sediment, ,,,·hi ch a re pred omina ntly deposited a t the tip of th e glacier tongue, in time ma king this a rea less d eep than a reas cl oser to the coast. The wa ter d epth required for Oota ti on a t th e seaward end of the M ertz Gl acier to ngue can be es tima ted using , ./' th e equa ti ons of Ling le and others (1991 ; equa tions (14)­ (19)) , which g ive density as a fun cti on of d epth a nd mean density between a give n d epth a nd the surface based on c1 ensi ty pro fi les measured in core holes on the R oss Ice Shelf. Their equa ti on (19 ) gives the mean d e nsity of th e ice shelf as a fun ction of to ta l thickness, for a firn- ice

tra nsiti on a t a d epth of 65 m . If ~I e rt z Glacie r is " ba rely fooIl0(0------50 km --~)I Ooating" in wa ter 250 m d eep at th e calving front, a surface elevati on there of 4 7 m would correspond to a total ice-shelf thickn ess of297 m. (The mean d ensity of th e Fig . 3. A sketch ojthe advance of the AlIat;;. Glacier tongue ice shelf wo uld th en be 864.7 kg m 3. A d e nsi ty of jor the /9 month period June 1993- December /994. This 1028 kg m 3 is assumed for sea water. ) If the surface sketch was obtained b)! subtracting the two SAR images elevati on a t the calving front is as hi gh as 50 m , howeHr, jrom each other . th en a wa te r d epth of 270 m would be required for Oota ti on, corres ponding to a total ice-shelf thickness of 320 m. (The mean density of the ice shelf would then be In Figure 4, a drawing is hown indi cating the 868.2 kg m 3.) A height of 53 m was the m aximum hypothes ized rela tionship between the crevasse p a tte rn elevati on m easured on th e glacier tongue but the mean on the Mertz Glacier tongue a nd stress, associa ted with height is a ro und 40 m. This indicates th a t M ertz Glacier the tra nsition from m a rgin al shear to longitudina l stress is not gro unded a t its seawa rd end. a nd orthogonal normal stresses o n th e tongue. The Alaskan glaciers grounded in deep yords cha racter­ indicated ve locity profile (das hed vectors) on th e ground­ isti call y ad vance by maintaining th eir termini in stable ed ice stream w as hypothes ized , n o t measured , a nd is pos iti ons in sha ll ow wa ter, on th e cres t of a morain e shown onl y to indicate th e proba ble res training influe nce where the wa ter depth is of the order of 10- 100 m (M eier of shear along the m a rgin s. The c revasse pa ttern o n the a nd Post, 198 7). The terminus adva nces by a brading Ooating ice tong ue is hypothes ized due to th e longitudina l debris from the up-glacier side of th e mora ine a nd stress regim e a nd is in good agreem e nt with the crevasse

Downloaded from https://www.cambridge.org/core. 30 Sep 2021 at 01:01:56, subject to the Cambridge Core terms of use. 449 Joumal qf Glaciology

sum of the area of the two glaciers, hardl y any losses or gains have been realized since 191 3; the observed 3.6% increase ill areal extent is within the accuracy of the

fT·-"";' ln~ measurements. The behavior of these two glaciers, lying in' tunu.ui,' h l:3u'iE::d h\ n.,,: . n~ In:( ... wnding ice next to each other in the same climatic zone, again shows cliff.... lfounlJ ~dgc., how difficult it is to rela te th e beh avior of advances or retreats of floating glacier tongues to climatic change (e.g. Sturm and others, 1991 ). In this case, opponents or proponents of globa l warming will each find one glacier to support their ideas. Generally speaking, there has been OCl

SHORT-TERM CHANGES

W e have obtained one SAR UERS-l ) image of Ninnis Glacier (Fig. 6), which shows this glacier in a state of Cl \CIFR break-up and dramatic retreat. There a re two images of M ertz Glacier; one image was taken close to mid-winter (30 M ay 1993), while the second was obtained close to mid-summer (23 December 1994), being about 19 months Fig. 4. D rawing indicating hypothesi:::ed relationship apart in time. Hence, short-term ch a nges could only be between crevasse pattern on the Mert::: Glaczer tongue and analyzed for the la tter glacier. \I\'hen comparing the two stress, associated with the transitionji-om marginal shear to images (Fig. 7), the most amazing fact is th e similari ty longitudinal stress and ortlzogonal normal stresses on the between them, and a close r examina tion is needed to see glacier tongue. the differences . The L-band SAR on board JERS- l proba bly "sees through" the dry winter snow. The sea ice is more pronounced in the winter image, an observation pa ttern obsen'ed on the glacier tongue (see a lso Fig. 8). to be ex pected. \rVhil e the hi gh ice concentration is When the glacier is viewed in visible li gh t (Fig. 5), exp ected in winter, there is a large amount of sea ice for remarkably no surface features a re detected. This results the mid-summer scene. Howeve r, the summer of 1994-95 from the fact that the visible part of the spectrum reDects was one of the most severe in the region. The French ha d light at the surface, while microwave radi a tion can par­ m ajor problems in resupplying their main station, tially penetrate the snow to a certain extent. Dumont d'Urville, a nd by this time of the year they Regarding advance ve rsus retreat and considering th e had been unable to reach their sta tion by supply ship. \rV e were in the area a few days aft er this image was obtained and, even with the powerful USCG Cutter PoLar Star we stayed about 30 miles offshore and fl ew in by helicopter on 28 December 1994, visiting Cape Webb, located on an ice-free outcrop between Mertz and Ninnis Glaciers. In the winter scene, less sea ice can be see n on the west side of the glacier tongue; "streaks" of newly form ed ice can be observed and, from the direc tion of these streaks, a southeasterl y wind direction can be d educed. This wind direction was confirmed from our coastal automatic weather stations. Other differences to be obse rved in the SAR Images are: There is a different number of sm all icebergs in front of the glacier tongue. These ice bergs break off periodi­ call y from the glacier. Especially, when the water is ice-free, th ey can be a spectacular sight. Fig. 5. Aerial plwt ograjJh oJ the floating Mert.;:: Glacier tongue taken in December 1994 showing the height oJ the One of th e two large fl oating icebergs to the eas t of the terminus as ajJproximately 40 m. M ertz Glacier to ngue, originating from Ninnis

Downloaded450 from https://www.cambridge.org/core. 30 Sep 2021 at 01:01:56, subject to the Cambridge Core terms of use. ,, 'endler alld oth ers: 0 11 Jlerl;:; all d Aillllis Glaciers, East An/arc /iea

Fig . 6. SA R (JERS-l) image of. \ illll is (;1(lcier. 1993.

G lacier, has d isappeared in th e second im age. id entifying clearl y th e same point in bo th images. S ubtrac ti on o f' the dig ita l \'ersions o f the two images The la rge tl oa tin g iceb e rg to the cast o f the glacier to ng u e yi eld ed a " diflc rence image" Cr om whic h c hanges could is some\\·ha t ro ta ted . I t has a size o [ 45 km by a b o ut be seen , esp ecia ll y in fr o nt o f the glacier to ng ue (Fig. 3). If 18 km a nd it a ls o o ri gina ted [l'om Ninnis Glacier. no calving had occurred , these \'a lues re present the .\J a tura ll y. \\'c \\'cre a lso interes ted in the movement of the ach a nce o f' the glacie r to ng ue. The ach-a nce is fa irl y g lacic r. For this, som e " fi xed " points h ad to b e unifo rm a lo ng the wh ole [i-o nt of the glacie r , which might es ta blished. W c fo und th ese a t Buch a na n Bav to the be ta ke n as a n incli cati o n tha t caking pla \'ed onl y a \\Tst o f~l crtz G lacier a nd a t Cape Hurley a t Fisher's Bay minor ro le during th c 19 m onth peri od. In a ny case, to the cast o f' the g lacie r. Flicking back a nd forth o n the \'alll es d educed in this way represe nt a minimum in cam pu ter sc ree n b e t weell th e t\\'o i m ages, th e mO\'em en t mo\'e m e nl. Again , valu es a ro und 1 km year 1 we re found. oCt he glacier could be seen, as there is sufli cient struc ture Perha p s the most striking aspect oC the surface is the to sce spcc ifi c points. " 'c determined the speed a t three orthogon a l p a ttern of in tc rsccLing cre\'asses (Fig . 8). This cross-sections or the g lacier: lin e A close to th e coastline. structurc is not found fa rthe r inland o n the glacier. line B a bo ut 50 km [rom the coastline a nclline C towa rds Inla nd, p a r a ll el cre\'asses p e rpendicul a r to the direction the tip of th e ice to n g u e (Fig. 8) . The speed was quite of mO\'em e nt a re dominant. This is the structure to be unifo rm and no sta tis ti call y \'alid dilTerences could be expected a nd fo und on o the r glaciers (P a terson, 1994 ) . d etected (o r th e th ree lin es . The a bsence of a \'eloci t y The ccll s o n M ert z Glacie r , which a rc quite vi sibl e on the d ecrease toward the tip, w here th e ice is thinner, is a g ain glacie r to n g ue, hm'e a w iclth of 1- 2 km. a n indicati on tha t the entire glacier tong ue is fl oating . In Knowing the glacie r thic kness close to the coastline, a ll cases, the 1ll 0 \T m ent (o r the 19 m o nth time peri od th e glacie r speed a nd the w id th of th e glacie r tongue, the \'a ri ed bet\\Tcn 1500 a nd 2000 m . A m ean \'alue of moti o n tota l exp o rt o f ire fr om l\le n z Glacier can b e estima ted as or 10 I 0 m \'ca r 1 was determined. This is in lair agreem e nt 12.7 km :l w a ter \'ca r I . The to tal a rea o f l\Ie rtz Glacier is w ith the ach- a nce of the g lacier tong ue, [or whi ch \\T h a d a bo ut 6 0000kl~1 2 This estima te is based on altitude (o und a mean \'alue o f 900 m a nnua ll y. It sho uld b e eln'ali o n s o btained in dig ita l format fro m the Scott Pola r pointed o ut th a t measuring the displacement is no t as Research Institute. H e n ce, o n a\'erage, ~I e rt z Glacie r simple as mi g ht be exp ected . This is no t ca used b y the ex ports a n a nnua l 2 12 mm w .e . acc umula tion of' snow and resolution: th c pixcl size is 12.5 m , w hieh res ults in a ice. The o nl y ma nned sta tio n a t Dumont d'Un'ille does n o mina l resolutio n o f a bo u t 30 m but ra th e r b y not m easure prec ipita ti o n , as the times with strong winds

Downloaded from https://www.cambridge.org/core. 30 Sep 2021 at 01:01:56, subject to the Cambridge Core terms of use. 45 1 ]oumal of GLaciolog)!

~ MAY 93 23 DEC 94

Fig . 7. SA R (]ERS-l) images qf M ert;:: GLacier taken about 19 months apart from each other. T rue north is 32° to the right .

are too frequent to obta in meaningful values . Further­ is based are a pproxim a ti ons: more, it is sometim es impossible to disting uish between "fa lling" a nd "blowing" snow. (a) The size of the glacier basin is diffi cult to d etermine Giovinetto (1964) gave es timates of precipitati on over in the upper areas, where the slope a ng les are very the Anta rcti c ice shee t. A value of 250 mm year- 1 w.e. small. might be deduced from hi s maps as the mean for the (b) The precipitati on \'alues for th e basin a re onl y M ertz Glacier basin. H ence, th e glacier apparen tly es tima tes; there is a fa irly steep gradie nt in the exports a bout 85% of the snow and ice accumula ted precipita ti on in th e coasta l a rea. over the catchment area. The res t may b e accounted for by snow blowing over the coastline into th e ocean (Loewe, 1972; Gi ovinetto and others, 1992 ), if these Nevertheless, this ex port of 15% of the precIpIta ti on glaciers a re in equilibrium. due to blowing snow appears to be a reasona ble valu e for I t should be poin ted ou t that this is a very rough this area in which very hig h wind speeds a re obse rved es tim a te, as two of the three values on which this number (Wendler a nd oth ers, 1994).

Downloaded from https://www.cambridge.org/core. 30 Sep 2021 at 01:01:56, subject to the Cambridge Core terms of use. 452 ll 'endler and others: On J'viert ::: and. rimlis GLaciers, East Anlarctica

Fig. 8. ,In en/argflllen l oJ lite tongue of lliertz Glacier . . \'ote the jJarliCII/ar cellu/ar slmc/ure 011 the glacier slllJace . Tl'1Ie norlh is 32 to Ihe right. A, Band Care Ihe three cross-sw/ions along It'hieh the lIlo velllml of Ih e glacier was determined.

ACKNOWLEDGEMENTS H olds\\'orth. G. 197+. Erc hus G lacier to n g ue, .\1c.\ l urdo Sound , Antarctica. J. (;I({ciol .. 13 (671, 27 35. Lingle, C. ' .. D. H. Schilling, J. L. Fa'took, \\'. S. B. Paterso n and TJ. W e shou ld like to thank thc fo ll owing: the personnel or Bro\\'n. 1991. A fl o\\' band model of thc R oss Ice ShelL Antarc tica; the Alas ka SAR Facility \\'ho wcre \'Cry helpfu l and B. response to CO ~ -inducl'd cli matic \\'a rming. J. Grol)/~I'>. Res., 961B 4 , ~ l oo re \\'ho d id the co mputer \\·o rk. Financial support 68+9-6871. Loc\\'c. f . 1972. The la nd "rstorms. , ,'eallier, 27. 11 0-121. was obtained ri'om the United States National Science .\la\\·son. D. 1915. Thl' /tollle 'if Ihe bk~ard: bl'ing Ihe .\10/)' o/Ihe .Iwlralasiall Foundation, O PP , grant 94-13879. .llIlarclic EI/mlilivlI, 191 1 I.9N. London, H cincmann. i\ leicr. '\L. F. and A. Post. 1987. Fast lide\\'ater g laciers. J. (;eo/)/~I '\' Re\., 92, B9 . 9051-9058. REFERENCES Paterso n, \I'. S. B. 1994. The jJ/~vjic> ,!/g/acien. Third edilioo. O xford, etc., Ebc\·ier. Pcri a rd , C. and P. Pe ttrc. 1993. Some aspec ts of th e climatology of Australi a. 197 1. .\ Ia/) uI Ihe .l/erl::. (;Iacifl. I: 1.000.000. Canhnra Dumol1t d't.,;n·illc. Addic Land, .\ l1t arcti ca. }ol. J. Climalol .. 13. 3 13 327. Departmcnt of' \ Jincrais a nd Encrg,. Di \' ision of :\'3 ti onal Stearns. C. alld G. \\·eneller. 1988. Resea rc h rcsults from ,\ ntarctic i\ lopping. SR55-56. 1 a utoma ti c \\'ca ther statio ns. Rn'. (;ev/)/~)', . , 26( 1),45 6 1. ,\ ustrali". 197+ . . Il ({/) oI j,'illg (;eorge Lalld. I: 1.000.000. Canbcn'a Sturm, \1.. D. K. Ha ll , C. S. Benson a nel \\'. O. Field. 1991. ;\1 0 11- Dcpartment or 1\ 1in era!' and Energy. Di\'i, ion o f :\'a ti o na l eliIllatic control of g lacier-terminus nuc tuations in th e \I'rangell a nd i\ lapping. (SQ55-56. ) Chugach \ Iountains, Alaska. U.s..-\. J. Giaciol .. 371127 1, 348 356. GiOl·inello . .\ l. B. 196+. Thc drainage s, stCIllS of .\ntarctica: accumula­ \\·end ler. G. and D . Pritcharcl . 1991. T cmperature increase obsclYccI in tion. III :'I lcll or , '\ 1. , ed .. llIlarclic JIIVlI' ({lid iCl' ,II/die;. \ \'ashinglOn , DC. AeI (' li c Land . East Al1la rCli ca: :Iolarcl. J. C.S ., 26,5), 28 1 28-L American Geophysical Union. 127 155. Antarctic R esearc h Series \ \·enel ler. G .. J. C .. \ ndre, P. Pelll'l', J. Gosink and T. Parish . 199+. 2 K a tahatic \\'inds in Ael c lic Land. 10 Bro Ill wich, D. H. and C. R. GiOl'in ctto, i\1. B .. D. H . Brom\\'ich a nd G. \ \·c ndlcr. 1992. AtIllosphnie Slearns. eds. :llIlarclic lIIeleorviog), aod dill/olviog)': sludie, balftl UII aulomatic nct tra nsport o f' \\'atcr \'apor a nd latc nt heal across 70 S. J. (;t'(J/)/~)'.\. ll'mlher III/Iioo,. \\'ashing ton , DC, Amcri can Gcoplwsieal Cnion. 23 Ih, 97t DI ),9 17930. ~ 5. (,\ ntarctie R esearch Series 6 1. )

.'vIS reeeilled 21 D ecember 1995 and accep ied 21 April 1996

Downloaded from https://www.cambridge.org/core. 30 Sep 2021 at 01:01:56, subject to the Cambridge Core terms of use. 453