JOVR•TALOF GEOPHYSXCALRESEARCH ¾OLV•fE 65, NO. 12 DaCEUSER1960

Oxygen-IsotopeRatios in the Blue Glacier,, , U.S.A.

i•OBERTP. SHARP,SAMUEL EPSTEIN, AND IRENE VIDZIUNAS Cal#ornia Institute ot Technology Pasadena, Calitornia

Abstract. The mean per mil deviation from a standard (average ocean water) in the O•]O •e ratio of 291 specimensof ice, tim, snow, and rain from the Blue Glacier is --12A; extremes are --8.6 and --192. This is consistentwith the moist temperate climatological en- vironment. The 0•/0 •øratio of snowdecreases with declining temperature of precipitation, and it also decreaseswith increasingaltitude at 0.5/100 meters. Analysesof the three principal types of ice, coarse-bubbly,coarse-clear, and fine, composing lower Blue Glacier, show that ratios for coarse-clearice are generally lower and for fine ice they are mostly higher than the ratios for coarse-bubblyice. This indicates that the fine ice representsmasses of firn and snow recently incorporatedinto the glacier by filling of crevasses or by infolding in areas of severedeformation. Coarse-clearice massesmay representfragments of coarse-bubblyice within a brecciaformed in the icefall. Becauseof unfavorable orientation, these fragments could have undergone exceptional recrystallization with reduction in air bubblesand, possibly,a relative decreasein 0 TM. A longitudinal septurn in the lower Blue Glacier is characterized by higher than normal O•8]0•ø ratios. These valuesare consistentwith an origin for this feature involving incorporation of much surficial snow and firn near the base of the icefall. Samples from longitudinal profiles on the ice tongue suggestthat ice closeto the snout comesfrom high parts of the accumulation area. Analyses from the light and dark bands of ogives are compatible with the concept that the dark bands represent greatly modified insets of tim-ice breccia filling icefall crevasses. The range in ratios of materialsis much'greater in the accumulationarea than in the ice tongue. This is attributed to homogenization,much of which takes place during the conversion of snow to glacier ice. This is supported by comparative analyses of snow layers when first depositedand months later after alteration. Refreezing of rain and meltwater percolatinginto underlying cold snow is an important mechanismas shownby analysesof ice layers and lenses in the firn formed in this manner.

Introduction. Somepotential uses of oxygen- area of a glacier acquirespatterns in the dis- isotopedata in glaciologicalresearch have been tribution of 018/010ratios which can be used as illustrated by analysesof samples from the natural tracers. Although the ratios are modi- Saskatchewanand Malaspina glaciers [Epstein fied duringconversion of snowto ice and during and Sharp, 1959]. 'Otheruses will be demon- subsequentflow within the glacier,this doesnot strated by analysesof materials from Green- destroytheir value as tracers.Among the modi- land [Benson,1960; IGY Bull., 1959,pp. 82-83] fying influencesare freezingof meltwater and and Antarctica, to be publishedshortly. The rain, captureof snowin crevasses,and homoge- usefulnessof the stable isotopesof oxygenand nization by other unidentifiedprocesses. Thus hydrogenin glaciologicalresearch rests on the the 0•8/0 •øratios tell somethingabout the origi- fact that their range of abundancein snow is nal conditions of accumulation and reflect the relatively large, far exceedinganalytical errors influenceof modifyingprocesses during a subse- of +--0.1in the ratio values.The value and range quent history. of 0•/0 •ø ratios in glaciersdepend principally Oxygen-isotopestudies of glaciersare still in upon meteorologicalconditions, especially upon a formativestage. The usefulnessof this ap- the temperatureat the time of snowfall.Thus proachvaries with the nature of a glacier,its the ratio varies with the storm, the season,the environment,and the problemschosen for study. elevation,and other factors.The accumulation For example,the 0•/0 'ø ratios in snow on the • Contribution No. 967, Division of Geological Greenlandice sheetdisplay simple relationships Sciences. usefulin stratigraphiccorrelation. In contrast,it 4043 4044 SHARP, EPSTEIN, AND VIDZIUNAS

I 47o49 ' Mercury

• PortAngeles ø •:;.•.'• •,

Eo Eo ' I .'r,.-..q-.•$,•0 5•'" I00 150km • ...... -'• ' ''.' ' "Apollo • .'..' .' ..

• • Do ;,.;'.;...-'•,.:.::....:...... 'i :.'. '.;..... ß, • • ..:5. '"' .:.:..:.' , ,• .• :...... ,

•. ..y... , ß • ::L::'.': '. ß •. '-.• ,;•.'.'..ß . ß. . ,, \ ,9'e • '-'.'..•...:: ...... ß . . /i • ...."::":" ' ' ' ' ' ' ' ' '

.. \ P•tA (hrn) '•,• \ \ N Btizzordoss .".-.'.:- :'.' ß o . \ \ \ -.

ß ß

ß .. ß .

ß

ßß'.'i '.::.;. •-

":':"::""?%L so,, "/,?• •/• ..• • Pit C(firn) •?:e .' ß.v/:•4::• % •.. No,thSosin •?:•:•L .'•.4:::.:•;'•'' ß ß ' ":•f•:•.• %'. •;.:•::•L •:•'::'•ß• ß . . .,'.?,...... • .. ..:.:::?•.•'

• ' ß' ".'::•.. : ...... :.•::::':??j') ;'::.).'0-

.. :.. :L• .-...

Fig. 1. Location, setting, shape and principal componentsof Blue Glacier. appears that icefalls and temperate conditions to determinethe usefulnessof the oxygen-iso- with copiousmeltwater and rapid exchangeof tope methodon a small temperateglacier with material lead to complicationsdifficult to in- a largematerial budget [LaChapelle, 1959, pp. terpret. In investigationsto date, the applica- 443-446]. bility of oxygen-isotopestudies to all types of The Blue Glacier was selected because it is glaciershas not been satisfactorilydefined. The relativelysmall, geometrically simple, and easily present investigationwas made in an attempt accessible;its constitution and structure are OXYGEN-ISOTOPE RATIOS IN THE BLUE GLACIER 4045 known [Allen, Kamb, Meier, and Sharp, 1960], upglacier from the edge of the tim, separates and associatedglaciological and meteorological lower Blue Glacier from its principalaccumula- studiescontribute to an understandingof oxy- tion basins,termed 'upper Blue Glacier.' The gen-isotoperelationships. The principal items Blue Glacier,supposedly temperate, has a high investigatedare (1) the range and mean value rate of massexchange owing to heavy accumu- of 0•8/0 •6 ratios as related to general environ- lation and strong ablation [LaChapelle, 1959, mental conditionsand to the theory of oxygen- p. 445]. isotopefractionation; (2) the influenceof alti- The climatologicalenvironment is strongly tude and temperature on 0•8/0 •6 ratios; (3) maritime, that is, relatively warm and moist. changesin isotoperatios within firn layers dur- Records from the Snowdome station at 2070 m ing alteration; (4) differencesin isotoperatios on upper Blue Glacier [LaChapelle, 1958, p. 12] of the three principaltypes of ice,coarse-bubbly, for the period August 1, 1957,to July 31, 1958, coarse-clear,and fine, composingthis glacier show a mean annual temperature of 1.6øC and their beating on the origin and history of (34.9øF), a meanfor the coldestmonth (March) these types of ice; (5) differencesin isotope of --6.1øC (q-21.1øF), and a mean minimum ratios of ice coming from various accumulation for March of --8.9øC (16.1øF). The lowesttem- areas; (6) variations in isotope ratios along perature recordedwas -15øC (5øF) and the longitudinal profiles on the surface of the ice highest21.7øC (71øF). Total precipitationwas tonguebelow the firn edgeand their relation to 378 cm (148.9 inches)of water, of which 305 cm flow lines within the glacier; (7) differencesin (119.7 inches), or 80 per cent, fell as snow.This isotoperatios of the materials composingogive period of observationwas unusuallywarm and bands as an aid in understandingthe origin of dry, judging from recordsat other meteorologi- this structure; and (8) changesin oxygen-iso- cal stations in northwesternWashington, and tope ratios,if any, producedby recrystallization, the abovefigures are not representativeof long- changesof state, and other processesrelated to range means. In an average year the mean an- solid flow. nual precipitation on Snowdomemay exceed A general program of glaciologicalresearch 500 cm (200 inches) of water. was begun on Blue Glacier in the summer of Lower Blue Glacier consistsof two major and 1957; it has extendedthrough 1960 and will be three minor ice streams, each originating in continued,with the permissionof the National separateaccumulation areas (Fig. 1). Only ma- Park Service.It was precededby Park Service jor streamsA and B extendto the snout; minor observationsin the 1940'sand early 1950'sand streams C, D, and E terminate along the east glaciologicaland glacio-meteorologicalwork dur- margin. Ice stream B consistsof two currents ing 1955 and 1956 [Hubley, 1957], and it has below the icefall separated by an intensely been accompaniedby glacio-meteorologicalre- foliated, structurally complex zone, the longi- searchon upper Blue Glacierfrom 1957to 1959 tudinal septurn,which is unusually rich in fine [LaChapelle, 1958, 1959]. Samplesfor oxygen- ice and coarse-clearice. This septurnseparates isotopeanalyses were collectedduring the win- two arc-shapedfoliation patterns displayedby ter of 1957-19-58 and in the summers of 1958 compositeice streamsA q- B• and B2 q- C and 1959. (Fig. 1). Ice stream A also displaysa seriesof Physical setting and constitutiono/ the Blue weak ogives of the internal variety. Details of Glacier. The Blue Glacier is a small ice stream these and other structuresare given elsewhere that rises high on the northeasternslope of [Allen, Kamb, Meier, and Sharp, 1960]. Mount Olympus (2413 m) in the heart of the Samplingand analysis. The methodof analy- Olympic Mountains of northwesternWashing- sis has been describedelsewhere [Epstein and ton (Fig. !). This glacier is 4.3 km long and Mayeda, 1953,p. 214] and will not be reviewed 1 km wide at the firn edge; it covers4.3 kmg here. In this paper, the result of an analysisis and descends from a maximum elevation of expressedas a relative, per mil deviation of the 2375 m to a terminus at 1265 m. The firn edge 0•/0 • ratio of the samplefrom the ratio of a has an approximateelevation of 1600 m, and standard--in this instance,mean ocean water. the bare ice tongueextends 2 km farther down This value, termed $, is calculatedin the follow- the valley. A major icefall, 300 m high, 0.8 km ing manner. 4046 SHARP, EPSTEIN, AND VIDZIUNAS TABLE 1. Oxygen-IsotopeRatios in Ice of a Core beento placeenough material in small-mouthed Taken 150 Meters below the Firn Edge in plastic bottles to make 25 to 50 cc of water Ice Stream B, August 22, 1958 whenmelted. The bakelitecaps should be tested for tightnessseveral times after collection.Pos- Depth, •i Value of cm 0•8/0 •6 Ratio sible evaporationcan be observedby reference to lines drawn at water level on the outside of

30.5 - 12.5 the bottles or by indentingthe bottlesslightly 61.0 -12.5 with the fingersbefore capping.If the indenta- 91.5 -12.3 tion remains,the cap is obviouslytight. Samples 122.0 - 12.8 should be transferred to glass bottles in the 152.5 - 12.8 183.0 - 12.8 laboratory if they are to be stored for a con- 213.5 -12.9 siderablelength of time. 244.O -12.5 Oxygen-isotoperatios in relation to the cli- 274.5 - 12.4 matological environment. The mean $ of 291 3O5.0 - 12.5 specimensfrom all parts of the Blue Glacier, Average - 12.6 representingglacier ice, firn, snow,and rain, is --12.4. This is not a truly representativefigure, as it doesnot give weight in proper proportions H20 /H20 (sample) to the different materials; nonethelessit is a 18 reasonable figure. The extremes recorded are $---- I H•.O /H•O 18(mean 16ocean water) -- 1 --8.6 and --19.2. These values confirm the basic X 1000 hypothesis concerningfractionation of oxygen In such an arrangementthe • for mean ooean isotopesin natural precipitation [Epstein and water is r,ero. Sinoe the 0•/0 •ø ratio of all .Mayeda, 1953,p. 220]. The OlympicMountains natural precipitationis lower than that of mean lie in a moist, temperate environmentthat is oceanwater, the $ of such precipitationis al- not exceptionallycold even in winter. As theory ways negativein this arbitrary system. predicts,these conditionsyield only moderately Accurateanalyses of the materialscomposing low $ values as compared with much lower a glacier are of limited value unlessthe signifi- valuesfrom glaciersin truly cold environments. cance of the specimenis fully understoodin Until 1957-1958essentially nothing was known terms of the field relations. A clear understand- from direct measurementof the meteorological ing of structuralrelations and a recognitionof conditionsprevailing on Mount Olympusin win- the differenttypes of ice on the Blue Glacier proved necessaryfor intelligent sampling,and 1 I I I I I I I

specimenshad to be collectedwith specificob- o -8 jectivesin view. The practicehas been to dig to fresh-looking ice 10 to 20 cm beneath the surface before tak- ing a sample,in order to eliminatepossible sur- face effects.The data of Table 1 suggestthat the near surfacesamples are reasonablyrepre- sentativeof ice to a depth of at least 3 m. Data from coresamples to that depth displaya rela- I tively high degreeof homogeneity,the rangein $ valuesbeing only 0.6 (--12.3 to --12.9). This {400 1600 1800 2000 2200 comparesnicely with a detailedsurface traverse Altitude in meters of similar dimensions on the Saskatchewan Fig. 2. Relationship between altitude and Glacier[Epstein and Sharp, 1959, p. 100],where $ values of OxB/Ox6 ratio in snow and firn on the differencewas 0.5 (--20.1 to --20.6). Blue Glacier. 1, snow in crevasse;2, snow on Propercare of samplesis necessary.It is par- glacier; 3, snowbank; 4, averageof 13 samples of 1957-1958snow, pit A; 5, average of 18 ticularlyimportant that evaporationsubsequent samplesof 1958-1959snow, pit B; $, average to collectionbe prevented.Our practice has of 45 samplesof 1958-1959snow, pit C. OXYGEN-ISOTOPERATIOS IN THE BLUE GLACIER 4047 TABLE 2. Changesin Oas/Oa• Ratio with Altitude Ej•ects of altitude. The altitude at which at Various Localities condensationoccurs in cloudsaffects the 0•/0 •' ! ratio of the precipitate[Epstein, 1959, pp. 224- Change in • Value of 228], the $ becominglower with increasingalti- tude. Measurements would best be made on 0•8/0•* Ratio per samplescollected at variouselevations during Latitude and 100-m correspondingstages of a singlestorm. 'On Blue Locality Longitude Altitude Glacierthis wasnot possible,and sampleswere taken from remnants of snow found at various Northwestern elevations.This procedureis opento criticism, Greenland 70-78øN; 40-65øW 0.6 as it does not eliminate seasonal variations. Saskatchewan Glacier, Alberta, Nonetheless,the data (Fig. 2) showa reason- Canada 52008•N; 117'12•W 0.2 ably consistentrelation between8 and altitude. SierraNevada, The value becomeslower with increasingalti- Calif. 39'N; 1200W 0.3 Sierra Madre tude at a rate of 0.5 per 100 meters.As is shown (Mountains), in Table 2, this is a steepergradient than that Calif. 34ø30'N; 118 øW 0.2 recordedin someareas, but it is not as steepas BlueGlacier, Wash. 47ø49'N;123ø42'W 0.5 the one in Greenland.Since the magnitudeof the altitudinaleffect must also be influencedby temperaturegradient, altitude gradient, and the ter. The upperBlue Glacierstudy [LaChapelle, natureof individualstorms, it is hardlylikely to 1958,p. 12] nowprovides such information, and be the samein differentplaces. the oxygen-isotoperatios are compatiblewith Current precipitation. Glacier ice is old in thesemeteorological data. the sensethat it consistslargely of materialac- TABLE 3. Samplesof CurrentPrecipitation

Nature Temperature Value of of Date and Hour at Time of O•s/O• Material Collected Location Collection,øC Comments Ratio

Snow 1/ 7/58, 1400 Snowdome -2 Fresh, wind-blown -10.2 Snow 1/ 8/58, 1000 Snowdome -5 Fresh, wind-blown -11.5 Snow 1/ 8/58, 1600 Snowdome -8 Probably fresh, wind- -13.5 blown Snow 1/ 9/58, 0930 Snowdome -4 Fresh, wind-blown -13.9 Snow 1/10/58, 1000 Snowdome -3 Fresh, wind-blown -13.1 Snow 1/11/58, 1000 Snowdome -6 Wind-blown,probably -17.4 reworked Snow 1/12/58, 0930 Snowdome -4.5 Fresh, wind-blown -15.7 Snow 1/ 13/58, 0930 Snowdome • -7 Wind-blown, reworked -15.3 Snow 1/15/58, 1000 Snowdome -2 Fresh, from surface -9.1 Snow 1/16/58, 0900 Snowdome 0 Fresh, from surface -14.3 Snow 2/ 9/58, 1130 Snowdome -4.4 Fresh, from surface -16.7 Snow 2/17/58, 1430 Snowdome 0 Fresh, from surface -13.9 Snow 3/ 6/58, 1004 Snowdome,east -11.1 Fresh powder snow -13.4 slope Snow 3/ 6/58, 1127 Snowdome,SW of -1.1 Fresh powder snow -11.0 center Snow 3/ 8/58, 0900 Snowdome -10.0 Fresh from surface -14.1 Snow 3/28/58, 1330 Snowdome -1.0 Wind-blown -11.9 Snow 3/30/58, 0830 Snowdome -7.0 From surface,worked -19.2 by wind Snow 4/21/58, 1300 Snowdome -5.5 Fresh, from surface -12.5 Snow ßi/24/58, 1330 Snowdome -3.3 Fresh• from surface -17.2 Rain 8/18/59, 1600 Caltechbase camp -•-7 (est.) -9.1 Rain 8/22/59, 0600 Caltechbase camp -•-3 (est.) -10.6 4048 SHARP, EPSTEIN, AND VIDZIUNAS

cumulatedtens, hundreds, or eventhousands of -8 years ago. By way of comparisonit is worth lookingat oxygen-isotoperatios in current pre- -io C B A cipitation on the Blue Glacier,chiefly in the form of new-fallensnow, collected during the winter of 1957-1958by the groupon Snowdome. The • for rain samplesfrom the Caltechbase campare higher (Table 3) than thosefor snow on Snowdome.This is to be expectedbecause the elevation of the Caltech camp is 445 m lower and becausethe samplesfrom the camp --Average -2o ß were collected in the summer and those from Snowdome were collected in the winter. Values Fig. 3. Range of $ values in materialsat for the snow samplesdecrease with declining following sites: S, specimensof fresh snow temperatureof precipitation.It is alsoapparent collected on Snowdome, winter of 1957- 1958.A, pit in firn on lower Blue Glacierat (Table 3) that a number of the snow layers elevation 1654 m as sampled August 7, 1958. deposited during windy periods have lower B, pit in firn on Snowdomeat elevation2045 valuesthan snownot blown by wind. The mean m as sampledAugust, 18, 1958.C, pit in South • of ten wind-blown samples is --14.2, com- Basin(Fig. 1) at elevation2205 m as sampled on August 9 and 18, 1958. D, 3-meter core paredwith --13.5 for eightnoneolian specimens. hole in ice 150 m below firn edge as collected Even thoughthis differenceis small,it may be August22, 1958.E, all samplesfrom ice tongue significant.Temperature does not seemto be below edge of the tim. Black bar indicates the controllinginfluence, as it wasslightly higher average $ value. duringaccumulation of the wind-blownmaterial (4.8øC as comparedwith 4.6øC). There is no is complicatedby the fact that the older ice basis in theory for thinking that evaporation probablyoriginated at a higheraltitude and its associated with wind action would reduce the $ values should be somewhat lower for that relative amount of 0 •, and condensationcould reason.Changes occurring within the firn are either raise or lower the value dependingupon probablya morelikely causeof relativeenrich- the nature of the condensingmoisture. The true ment in 0 TMand alsoof much of the homogeniza- explanationmay lie in basicdifferences in windy tion. It seemsunlikely that the differencesin storms and in their histories of precipitation average $ values of samplesfrom Snowdome prior to their arrival on Mount Olympus. and from the ice tonguereflect a secularclimatic The average • for snow from Snowdomeis change,in view of the relativelysmall secular somewhatlower (--13.9) than the average for changein temperaturein this regionover past all samplesfrom the ice tongue (--12.1). Fur- decades[Hubley, 1956,p. 673; Landsberg,1960, thermore, the range of values for the snow p. 1520]. (--9.1 to --19.2) is much greater than for the One possiblemechanism of homogenization materials of the ice tongue (--10.3 to --14.1), and 0 • enrichmentis the refreezingof the melt- as is shownby bars S and E in Figure 3. These water and rain water that percolateinto cold, data suggest a considerablehomogenization, underlyingsnow or tim. In an annual layer of much of which occursduring conversionof snow snow,before the melting season,the part that to glacier ice and which probably continuesat a accumulatesin winter has the greatestreservoir reduced rate during the subsequenthistory of of cold and, on the average,the lower • values. ice in the glacier (compareD of Fig. 3 with A, This winter snowis initially overlainby spring B, and C). A moderate relative enrichmentin and possiblyearly-summer snows which pre- 0 •8also takes place. This couldbe broughtabout cipitated under warmer conditionsand conse- either by adding 0 • or by removing 0 •e. The quently have high • values. Meltwater that analysesdo not indicate that old, far-traveled percolatesinto the cold winter layers comes ice has been enrichedin 0 •, as comparedwith from these surfacesnows of high • values.The young,less-traveled ice (Fig. 5). Actually, they differencein • is probablyeven more markedin suggestthe reverse,although the interpretation the instance of rain water, although much of OXYGEN-ISOTOPE RATIOS IN THE BLUE GLACIER 4049 TABLE 4. Changesin 0'8/0 '6 Ratio of Specific slow process,and it has not been evaluatedby Snow Layers with Time for Samples actual measurements. from Snowdome An attempt to measurethe rate of homogeni- zation was made through cooperationof the • Value of Sample Date 0'8/0 '6 Snowdomegroup by samplinglayers of snow Number Collected Description Ratio when they first accumulatedand resampling them months later after they had undergone Series I considerablealteration. The results (Table 4), 1 2/9/58 Surfaceof a deeplayer - 16.4 though interesting,are not entirely consistent of new snow, or compelling.For example,the snowlayer that marked for recovery accumulatedon February 9, 1958, had a de- 2 6/22/58 Collectedfrom the -12.8 cidedly higher • when resampledon June 22, same layer as sample 1 now re- 1958. Part of this may have beendue to its free- exposedat surface water contentin June, which couldhave beenin (from open snow the neighborhoodof 10 per cent by weight (La- surface) Chapelle, personal communication),but most 3 --12.9 6/22/58 Samelayer as sample of the differencewas presumably due to the 1, protected by a tin can refreezingof water percolatingdown from the 4 6/22/58 Same layer as sample -13.1 surface in late spring. An attempt to evaluate 1, protectedby 1 sq the effectsof percolationwas made by protect- ft of aluminum foil ing part of the same snow layer by means of Series II impermeable coverings. The difference in the 5 2/17/58 Surface of new snow -13.9 samples so protected is consistentwith a re- layer 6/16/58 Samelayer as sample -13.2 ducedpercolation, but it is not of great magni- 5, from open ex- tude. It may be that the percolatingwaters posedsurface gained considerableaccess to the covered snow 6/16/58 Samelayer as sample - 14.6 through lateral capillary channelsand that the 5, protected by tin can protectionwas only partly effective. 6/1(}/58 Samelayer as sample -14.3 The $ values for the snowlayers of February 5, protected by 17 and March 8 (Table 4) had also become aluminum foil higher when resampledon June 16 and June 5, Series III respectively,but the changewas much lessthan 9 3/8/58 Surface of new snow --14.1 that in the snow layer of February 9. The layer shieldedpart of the March 8 layer showeda 10 6/5/58 Same layer as sample -13.5 9, not protected smaller change than the unshieldedmaterial, 11 6/5/58 Samelayer as sample -13.9 which would be expectedif it had receivedless 9, protected by tin meltwater. However, the shieldedsamples of can the February 17 layer had lower • values than the original snow.This is an unexpectedresult for which no satisfactory explanation has yet the cold reservein the snowhas probablybeen been found.The samplingprocedure may be at eliminatedbefore much rain falls. Refreezing of fault. It is known that the • for ice layers and percolatingmeltwater, and to a limited degree lenses in firn sections (Table 5) is generally of rain, in the cold winter snowsis believedto higher than that in the adjacent firn layers. be an important factor in the homogenization Sincethe ice bodiesare formed by refreezingof and relative enrichment in 0 TMwithin the snow percolatedmeltwater, they indicatethat relative that survives the ablation season. This is the enrichmentof 0 •8by this mechanismdoes occur. material, of course,that ultimately makes up Oxygen-isotoperatios in tim. Knowledgeof the ice tongue. Homogenizationmay continue oxygen-isotoperatios within the firn of the ac- after the firn is raisedto the freezingtempera- cumulationarea is valuable,as this is the source ture through exchangeof oxygenbetween the of the material composingthe ice tongue of the firn and percolating water. This should be a glacier. In August 1958, samples were taken 4050 SHARP, EPSTEIN, AND VIDZIUNAS

TABLE 5. Oxygen-IsotopeRatios in 1957-1958 TABLE 5. Continued Firn from Pits on Blue Glacier Depth, • Value of Depth, • Value of cm 018/016Ratio cm O'8/0 •6Ratio

30.5- 38.0 --13.5 Pit A, elevation 1654 m, 45.5- 53.5 --12.1 sampledAugust 7, 1958 61.0- 68.5 --13.1 0 - 5.5 -11.8 63.5 ice layer -- 11.9 5.5- 11.0 --11.9 68.0- 76.0 --10.9 11.0- 17.0 -11.3 73.5 ice layer -- 11.2 17.0- 23.0 -11.2 76.0- 84.0 --13.2 20.0 0.5-cm ice layer - 11.2 91.5- 99.0 • --15.4 23.0- 28.5 -11.7 106.5-114.5 --15.6 28.5- 34.5 --11.3 122.0-129.5 -- 14.9 34.5- 40.0 --11.7 137.0-145.0 --14.3 40.0- 45.5 --12.0 152.5-160.0 -- 13.6 4O.5 0.5-cm ice layer - 10.9 167.5-175.0 --12.7 45.5- 51.5 -12.3 183.0-190.5 -- 11.2 51.5- 57.0 -12.0 198.0-205.5 -- 13.7 67.0 old, water-saturatedfirn - 11.5 213.5-221.0 -- 13.9 228.5-236.0 -- 13.8 Average - 11.6 244.0-251.5 -- 13.5 Max. range 1.4 259.0-267.0 -- 13.9 274.5-282.0 -- 12.9 Pit B, elevation 2045 m, 289.5-300.0 - 13.4 sampledAugust 18, 1958 307.5-315.0 -- 13.9 0 - 7.5 -12.1 322.5-330.0 -- 13.6 7.5- 15.0 -12.4 338.0-345.5 -- 13.6 15.0- 23.0 -12.4 353.0-361.0 -- 13.2 23.0- 30.5 -12.2 368.5-376.0 -- 13.4 30.5- 38.0 -12.3 376.0-383.5 -- 13.0 38.0- 45.5 -12.5 386 1-cm•icelayer -- 10.9 45.5- 53.5 -13.5 383.5-391.0 -- 11.8 53.5- 61.0 -13.3 391.0-399.0 -- 12.6 61.0- 68.5 -13.0 406.5-414.0 -- 12.8 68.5- 76.0 -12.3 421.5-429.0 -- 13.0 76.0- 84.0 -12.1 437.0-444.5 -- 12.6 84.0- 91.5 -12.2 452.0-460.0 -- 12.9 89.O 1-cm ice layer - 11.1 467.5-475.0 -- 12.3 91.5- 99.O -12.7 482.5-490.0 -- 12.2 99.0-106.5 -12.0 498.0-505.5 -- 12.5 101.5 5-cm ice layer - 13.1 513.0-520.5 -- 12.5 106.5-114.5 -11.3 528.5-536.0 -- 12.5 114.5-122.0 -11.3 543.5-551.0 -- 12.4 559.0-566.5 -- 12.8 Average - 12.3 566.5 2.5-cm ice layer -- 13.0 Max. range 2.4 566.5-574.0 --12.5

Pit C, elevation 2205 m, Average - 13.0 sampledAugust 9 and 18, 1958 Max. range 4.7 0- 7.5 -12.1 15.0- 23.0 -14.7

from pits in the 1957-1958annual firn layer at do not representa completeannual layer, as three sites spanning the maximum possible much material had been lost at the top through vertical (550 m) and horizontal (1.4 km) range. ablation.The sectionat pit C (altitude 2205m) Continuous channel sampling was used, each wasby far the thickestand mostcomplete (Fig. specimenrepresenting a thicknessbetween 5.5 4) and had unquestionablyundergone the least and 7.5 cm. Ice layers and lenseswithin the firn alteration. were separatelysampled. The sectionssampled The data (Table 5) confirm the trend and OXYGEN-ISOTOPE RATIOS IN THE BLUE GLACIER 4051

• value0•8/0 •6 ratio richmentin 0 TMwith depth is indicated.A com- -16 -15 -14 -15 -12 -II -I0 parison of 'new' firn (--11.2) and 'old' firn EZ• ..... •'nsurface (--10.õ) at the firn edgeshows the sametrend. ---1__ on9 Aug1958 Ice layersand lensesin firn are formedby the refreezingof water that percolatesdown from

.! the surfaceand spreadsout alonga particularly stratigraphic layer [Sharp, 1951, p. 613; Ben- son, 1960, pp. 38-40]. For reasonsalready dis- cussed,these ice bodiesshould have • values higher than those in the adjacent tim. This -200 provedto be the casefor sixout of eightice layers _ • '•r•h•notedtop1957 F•Pn and lensesin the firn pits of the Blue Glacier. - 250 The exceptionsconstitute an unsolvedproblem similarto that encounteredin pits on Saskatche- - 300 wan Glacier [Epstein and Sharp, 1959, p. 94]. _ It may be that the originalsnow layers had such - 350 low • values that permeationwith meltwater

_ 1 did not bring the value up to that of the ad- jacent layers.

_ Types of ice. The tongueof lower Blue Gla- - 450 cier consistsprincipally of three types of ice, © - Ice layers arbitrarily identified as coarsebubbly, coarse - 500 clear,and fine? The fine ice is alsobubbly, and _ all three types are describedin more detail else- - 550 .J where [Kamb, 1959, p. 1893; Allen, Kamb, _ Meier, and Sharp, 1960]. In many instances, - 600 closelyassociated specimens of these types of ice display significantdifferences in 8 values (Fig. 6, Table 6) which may reflectdifferences Fig. 4. 0•/0 •ø ratios in firn and ice layers of pit C, elevation 2205 m. in genesisand history. In most placescoarse- bubblyice constitutesat least90 per cent of the generalmagnitude of the altitudinalinfluence on exposedmaterial. It is consideredto be the 0•'/0 •ø ratios (Fig. 3). The range in • values 'normal' or 'average'ice, and its 8 providesa also increaseswith altitude, being 1.4 at pit A datumfor comparisonwith othertypes of ice. (1õ54 m); 2.4 at pit B (2045 m), and 4.8 at Coarse-clear ice occurs in close association pit C (2205 m). Sincethe range of valuesin with both coarse-bubblyand fine ice. In some new-fallen snow collectednear the s•te of pit B placesit is in sharplydefined masses, and in is still greater,10.1 (Table 5), the conclusionis others the transition into coarse-bubblyice is permissiblethat the decreasein rangeat lower gradual. At most localitiesthe • values for altitudesis at ]east in part a matter of increas- coarse-clearice are distinctly lower than the ing homogenization.The firn at lower elevations valuesobtained for adjacentcoarse-bubbly ice. has simply been permeatedby greater amounts The mean differencein the values for 30 closely of meltwater over a longerinterval of time. Five associatedcoarse-bubbly and coarse-clearice of the specimensfrom pit C havea lower• than pairs is --0.4, and the maximumis --2.0. In 6 any samplefrom the ice tongue,and it seems out of the 30 pairs the 8 of coarse-clearice was that duringthe processof homogenizationa rela- higher,but in 8 of thesethe differencewas only tive increaseof O" in the materials of the tongue +0.1. The 8 of coarse-clearice is also lower has occurred.These relationshipsare borne out than that of associatedfine ice in nearly all in- graphicallyin Figure 4, which showsthat the stances,although one striking exception was frequency and range of variations within the a This or similar material has been called gran- 1958 layer clearly exceedthose in the underly- ulated ice [Epstein and Sharp, 1959,p. 991 or fica ing tim. Furthermore,a small but steady en- ice (Firneis) [Klebelsberg, 1948, p. 39-401. 4'052 SHARP,EPSTEIN, AND VIDZIUNAS

o

,o ß

ß . ß .. o ß

ß

ß

,..

ß z ,.

ß

ß

ß

ß .. .o ø %0.

ß ..

•.o.'•.?:: •.:::• Oi [ 0o**o ß...:.. ß... ,,o0ø ß ß e'o •.-.•../...... :o00. ß ,.'.....' •,, ß . , .. ß...'.'? ß ...... ß ß .. ß ß .. OXYGEN-ISOTOPE RATIOS IN THE BLUE GLACIER 4053

E

-IOO ß

- ii o .../ ...... ,e-...... /'"' ...... e..,.

-130 - • / •g /

-14.0 - • - iCE STREAM a iCE STREAM b

o

E

West -• East ,o .__

I( Coarse-bubbly ice

.i Coarse-clear ice ...... Fine ice

.1•

,I / - ICE STREAM A ICE STREAM B

o 50 IOO 200 $oo 400 meters I I I I I I I I" I Scale in meters

, Fig. 6. Plots of oxygen-isotopevariation in samplesof coarse-bubbly,coarse-clear, an(] •ine ice alongtrausverse profiles across lower Blue Glacier; Fig. 5 for location. 4054 SHARP, EPSTEIN, AND VIDZIUNAS TABLE 6. The it Valuesfor Coarse-Bubbly,Coarse-Clear, and Fine Ice

Coarse- Coarse- Difference Clear Fine Bubbly Location Ice Ice Ice CCI-CBI FI-CBI

At base of western icefall --11.6 --11.8 +0.2 6 m downglacierfrom icefall --13.3 --12.4 --0.9 225 m downglacierfrom icefall --13.5 --11.4 --12.1 --1.4 +0.7 Dark ogiveband oppositebase camp --13.1 --11.3 --12.3 --0.8 +1.0 White ogiveband oppositebase camp --12.5 --13.0 --12.2 --0.3 --0.8 Dark ogiveband oppositebase camp --14.1 --11.3 --13.7 --0.4 +2.4 White ogiveband oppositebase camp --10.5 --11.7 --12.1 +1.6 +0.4 Dark ogive band oppositebase camp --12.7 --11.5 --12.7 0 +1.2 Longitudinalsepturn opposite base camp --10.7 --12.5 +1.8 6 m from west rock wall below TP-5 (Fig. 5) --9.8 --12.1 +2.3 '--12.0 --12.4 --12.1 +.1 --0.3 --13.5 --11.9 --12.0 --1.5 +0.1 --12.4 --12.0 --11.7 --0.7 --0.3 --13.0 --10.5 --12.3 --0.7 +1.8 --11.8 --11.1 --11.8 0 +0.7 Upper transverseprofile XX' (Fig. 5) --11.3 --10.5 --11.0 --0.3 +0.5 --12.4 --11.3 --11.2 --1.2 --0.1 --12.5 --10.5 --12.3 --0.2 +1.8 --14.1 --11.0 --12.1 --2.0 +1.1 --12.0 --11.9 --11.3 --0.7 --0.6 --12.2 --10.9 --1.3 Center of ice streamA oppositeTP-5 (Fig. 5) --13.0 --12.4 --12.1 --0.9 --0.3 East edge of stream B oppositeTP-5 --13.4 --11.0 --14.0 •0.6 -]-3.0 Center of stream A oppositeTP-4 (Fig. 5) --12.6 --12.0 --12.4 --0.2 -{-0.4 West part of stream B oppositeTP-4 --12.3 --11.1 --11.4 --0.9 East part of streamB oppositeTP-4 --13.0 --11.1 --12.8 --0.2 -{-1.7 --12.4 --12.3 --12.5 -]-0.1 -]-0.2 --12.4 --11.4 --12.5 -]-0.1 -]-1.1 --13.5 --13.0 --13.0 --0.5 0 Lower transverseprofile YY' (Fig. 5) --13.0 --12.2 --12.8 --0.2 -]-0.6 --12.3 --11.5 --11.6 --0.7 -t-0.1 --11.6 --10.0 --11.4 --0.2 -]-1.4 --12.5 --11.5 --12.5 0 -]-1.0 --13.4 --12.8 --13.8 -]-0.4 +1.0 West medial moraine 300 m above terminus --11.7 --12.6 -]-0.9 East margin 200 m above terminus --12.2 --12.6 -]-0.4 found in a pod of coarse-clearice with a $ of relatively abundant, are so readily recognized -8.6 completelysurrounded by fine ice with a by their distinctiveform and crystal structure ratio of --9.8. that there is little chance of confusingthem On the basisof $ values and field relations, with the masses of coarse-clear ice under con- it is suggestedthat the massesof coarse-clearice sideration. (2) Bodies of coarse-clearice with may originatein at least two ways: (1) Coarse- $ values lower than those of the associated clear ice with a • higherthan that of the associ- coarse-bubblyice may represent chunks of ated ice may representlocal bodiesof material deeper and older ice which becamemixed with that gradually recrystallizedwhile soakedwith surficial material in the icefall. Ice breccias con- water. Since most meltwater and rain water are taining fragmentsof coarse-grainedbluish ice in richer in 0 •s than most of the underlyingsnow, a matrix of firn and fragmentedice have been firn, or ice, incorporationof such water by re- observed near the base of the fall. If the bluish crystallizationwould producea higher •. Refer- ice of the brecciarepresents deeper material, it enceis not madehere to refrozenpools of water should have a lower $ because it comes from filling moulins,crevasses, or other depressions. higher in the accumulationbasin. Bodiesof ice formed in this manner, although The massesof coarse-clearice in the ice tongue OXYGEN-ISOTOPE RATIOS IN THE BLUE GLACIER 4055 are much less bubbly than the breccia frag- a lower $ than the adjacentice. Thus, even fine ments. This could be due to an unusual degree ice with $ values lower than those in the asso- of recrystallizationexperienced by these frag- ciated coarse-bubblyice may originate as insets ments becausethe orientation of their crystals of snow.The same explanationmight also hold was not well suited to the direction of stress at for the one exampleof fine ice with a $ lower the base of the icefall. Recrystallization is than the coarse-clear ice mass it enclosed. the principal meansby which ice grains in the Transverseprofiles. Variations in oxygen- breccia fragmentscould be reorientedand the isotope ratios along profiles extending trans- clarity of the ice increasedby reduction of air versely across a glacier should reveal differ- bubbles.This interpretationis supportedby the ences in the site of accumulation of the material fact that nearly all the ice near the glacier composingindividual ice streams.Samples were snout, which presumablyhas experiencedmuch taken along two principal transverseprofiles recrystallization,is much clearer than most ice and alongseveral shorter traverses across parts farther upglacier.Ice along lateral margins of of the ice tongue of lower Blue Glacier. The the glacier,where large crystalssuggest consider- uppermostprofile (XX', Fig. 5) is 150 to 250 able recrystallization,is also relatively clear. metersbelow the firn edge.It starts at the west Most fine ice has higher • values than the wall and crossesice streamsA, B, and C (Fig. associatedcoarse-bubbly ice, but in a few in- 1). An apron of firn coversmost of ice streams stances the values are the same or even lower D and E, so they were not sampled.The lower (Table 6). The mean differencein 35 coarse- profile (YY', Fig. 5) is 450 m above the snout bubbly and fine ice pairs is +0.7, and the maxi- and extendscompletely across the glacier from mum is +3.0. Statistical analysis shows that wall to wall. It involvesonly ice streamsA and this difference is not due to chance. In 8 out of B, the other streamshaving terminated farther the 35 pairs, the $ of fine ice is lower. upglacier.Data from the miscellaneousshorter At some locations it is clear from direct ob- traversesare not discussed,but they are con- servation that the fine ice representsinsets of sistent with results from the longer profiles. snow or firn filling crevasses.However, it is not Separate samplesof coarse-bubbly,coarse- clear from the field relationshipsthat all fine ice clear, and fine ice were taken at each collection in lower Blue Glacier originatedin this manner. site alongthe principal transverseprofiles. The Some of the fine-icebodies are lensesor pods, $ valuesof thesesamples are plottedin Figure6. others are highly irregular in shape,and many One relationship immediately apparent is the are thin folia intimately associatedwith the differencein $ values of the three types of ice other types of ice. If all the fine ice represents alreadydiscussed. If fine ice actually represents inset bodies of firn, the geometricalrelations insetsof tim, the differencein $ valuesbetween have beenso greatly modifiedand the fine ice so it and the accompanyingcoarse-bubbly ice closelyincorporated into the prevailingfoliation shouldbe greateralong the lowerprofile because structure of the glacier that the inset mode of the coarseice of the lower profile comesfrom origin is no longerobvious. We have entertained higherin the accumulationbasin. This proves the thought that some of the fine ice may be to be the case. ground-up or recrystallized[Kamb, 1959, pp. The differencein the average $ of coarse- 1896-1900] coarse-bubblyice, but the $ values bubbly and coarse-clearice is lesson the lower do not generally support these ideas. The $ profile than on the upper profile, --0.2 com- values of fine ice in thin folia are roughly the pared with --0.7. This is consistentwith the same as the values for firn fillings in crevasses. greater similarity in appearanceof these two In general, snow filling a crevassewould be typesof ice in the lower reachof the glacierand expectedto have a higher $ than the ice of the with the generalevidence of homogenization, crevassewalls, which representsmaterial that but the reasonsfor the relative changesin $ accumulatedat higherelevations. However, vari- values are not known. ations in the 'O•8/O•eratio among individual Useful comparisonsof the differentice streams snowstorms are considerable(Table 3), and it can be made on the basis of $ values in coarse- is possiblethat snowaccumulating in a crevasse bubbly ice alone.Along profileXX • thesevalues may occasionallyhave about the same or even have only a small rangewithin ice streamA and 4056 SHARP, EPSTEIN, AND VIDZIUNAS are about the same in both ice streams A and ]3. gins of fine and coarse-clearice, this is perhaps This is to be expected,as both streamsoriginate not surprising.However, the samplesof coarse- in accumulation basins having a similar mor- bubbly ice do show a somewhatirregular but phologyand essentiallythe sameelevation. The unmistakabledecrease in • downglacier(Fig. 7). only marked departureis within ice stream13 at Irregularities in these curves probably reflect the crossingof the longitudinal septurn where local inhomogeneitieswithin the ice left over the • values are distinctly higher than average from the firn. Althoughphysical aspects of indi- (Fig. 6). This is consistentwith the preferred vidual firn layers may be obscuredwithin the hypothesisof origin for this feature involving glacier tongue,it is hardly likely that the con- the incorporationof a large amount of snowand siderable differences in 8 values for individual firn in and at the base of the icefall [Allen, firn layers (Fig. 4) are completely eliminated. Kamb, Meier, and Sharp, 1960]. Furthermore, variations related to secular cli- Samplesfrom the lower profile show essen- matic changesmay exist and must involve large tially the samefeatures as the upper profile,in- masses of ice. If allowances are made for such cludinghigher 3 values in the longitudinalsep- inhomogeneities,the casefor a modest decrease turn. The low 3 at the east end of the lower in 8 is acceptable. profile (YY•, Fig. 6) may be due to excessive A more reliableevaluation is perhapsafforded marginal ablation, which exposes relatively by comparisonof the mean 8 for all samplesof deeperice. Lack of a correspondinglylow value coarse-bubblyice taken along the two trans- at the west end of the profile couldbe due to a verse profiles,XX' and YY' (Fig. 6). The mean much lower ablation related to shadedexposure value along the lower profile (--12.5) is lower and to protection afforded by residual snow than the mean value along the higher profile banks. In general,ice along the marginswould (--11.9). Other 8 values for coarse-bubblyice be expectedto have somewhatlower 3 values on the surfaceof the glacier,but not locatedon becauseof the slowervelocity, which gives op- the longitudinalprofiles (Fig. 5), confirmin gen- portunity for exposureof deeperice by greater eral the trend toward lower values downglacier. melting. Thus, data from Blue Glacier offer modestsup- Lon(jitudinalprofiles. Deductionsconcerning port to the deductionsof Reid [1896, p. 919] longitudinallines of flow in a valley glaciersug- concerningflow lines in a valley glacier. gest that ice appearingon the surfaceat posi- Oxyglen-isotoperatios in oglives. The ogives tions progressivelyfarther below the firn edge of lower Blue Glacierappear as lunate, alternate came from successivelyhigher parts of the ac- white and darker bands on the surface of ice cumulationarea. If this is correct,the • values stream A below the firn edge.The white bands should become progressivelylower from firn average roughly 25 m in width, the darker edge to glacier terminus. bands 5 m. These bands represent the outcrop To explorethis relation,a seriesof composite traces of layers of material within the glacier. sampless was collectedin 1958 alongthe center The white bands are 90 to 95 per cent coarse- flow lines of ice streams A and 13. In 1959, bubbly ice, and the darker bands are a more samplesof coarse-bubblyice alone were taken heterogeneousmixture of coarse-bubbly,fine alongthe center flow line of ice streamB. The (up to 35 per cent), and coarse-clearice (up to 8 values of these samplesare shown on the 10 per cent). The origin of these ogivesis a map (Fig. 5), and the values for the coarse- matter of speculation,but onehypothesis [Allen, bubbly ice aloneare plotted in Figure 7. Kamb, Meier, and Sharp, 1960] is that the The compositesamples collected along the darkerbands represent insets of ice brecciathat longitudinalprofiles show no consistenttrend in accumulatedin icefall crevasses.It is interesting 8 values (Fig. 5). Consideringthe possibleori- to seewhat light the O•s/O• ratios can throw on the origin of these structures. • A compositesample consistsof small chips of The samplesanalysed came from two different ice taken at ten separatespots distributed over an but closelyassociated sets of ogives,represent- area not exceeding10 to 12 m in radius.The three ing two white and three darker bands.The • commontypes of ice are included in roughly the estimated proportionsexposed on the surface at values for fine and coarse-clearice display wide the sampling site. variations (Table 7), presumablyfor reasons OXYGEN-ISOTOPE RATIOS IN THE BLUE GLACIER 4057

TABLE 7. Oxygen-IsotopeRatios of Materials being 0.5/100 m. This behavior is not unique ComposingOgive Bandsof Blue Glacier to snow,as it characterizesmany types of pre- cipitation [Epstein, 1956]. The Blue Glacier Typeof Natureof $Valu e data also show that homogenizationof oxygen Band Material O•8/O•6 isotopesbegins shortly after snowaccumulates, Ogive set I Composite* - 12.2 and significanteffects are evident within a few Coarse-bubblyice - 12.1 months.It appearsthat much homogenizationis White (lw) Fine ice - 11.7 effectedin the accumulationarea through re- Coarse-clear ice - 10.5 freezing of downwardpercolating water derived from melting of surface snow and from rain. Composite* - 12.4 Coarse-bubblyice - 12.7 The effects of vapor transfer and diffusion of Dark (ld) Fine ice - 11.5 oxygen are matters of speculation. Further Coarse-clear ice - 12.7 homogenizationoccurs within the ice tongue during flow, but the processescausing it are not Ogive set 2 Coarse-bubblyice - 12.3 Dark (2d) Fine ice - 11.3 known. They may involve recrystallization, Coarse-clear ice - 13.1 changesof state, and diffusion.The homogeniza- tion is of local extent and doesnot destroylarge- Coarse-bubblyice - 12.2 scale heterogeneitiesin • values which can be White (2w) Fine ice - 13.0 Coarse-clear ice - 12.5 used to study and interpret glacier structure and behavior. Coarse-bubblyice - 13.7 Differencesin 0•8/0 '6 ratios proved to be a Dark (2d) Fine ice -- 11.3 useful aid in understandingsome of the struc- Coarse-clear ice - 14.1 tures within the ice tongue of the Blue Glacier. They indicate that thin layers of fine ice in the * A compositesample consists of small chipsof foliation pattern represent greatly drawn out ice taken at ten separatelocations within a radius of 10 to 12 meters,representing the three types of massesof firn incorporatedinto the glacier by ice in approximatelytheir estimatedabundance. infolding or insetting,largely within and at the base of the icefall. The 0•/0 '6 ratios support already discussed.The • of coarse-clearice in the interpretation that creation of the longi- white band l w is higher and the • of fine ice in tudinal septum, a major structural feature, oc- white band 2w is lower than would normally be curs near the base of the icefall and involves the expected.The significanceof this is not readily incorporationof large amountsof snow and firn. apparent.The • of coarse-bubblyice is lower in The hypothesisthat ogive dark bandsrepresent the darker bands than in the white bands,the greatly modifiedinsets of tim-ice breccia filling averageddifference being 0.7. A possibleexpla- icefall crevassesis supportedby the oxygen- nation of this is that the insets of ice breccia isotope data. Variations in 8 values also attest contain considerabledeep ice which has lower to the probable validity of deductionsconcern- • values. Possibly,this could come about by ing longitudinalflow lines within a valley glacier avalanchingand crevasse-wallcalving in the by showingthat ice near the terminusprobably icefall. comesfrom the higher parts of the accumula- Summary and conclusions. The oxygen-iso- tion area. tope data from Blue Glacierlead to somerela- On the other hand, the presentstudies do not tively straightforward,definite conclusionsand contribute significantlyto an understandingof lend support to certain interpretations and the fundamental control of '0'8/0 '6 ratios exer- speculations,but they alsoraise many questions cised by various aspectsof the meteorological to which there are as yet no clear answers.This environment.Processes of homogenization,par- is not unexpectedin view of the complexitiesof ticularly within the ice tongue, remain largely the Blue Glacier environment and the formative unknown. Study of other glaciers in different stage of oxygen-isotopestudies of glacier ma- climatologicalenvironments is in order. They terials. should be of simple geometry with the least The Blue Glacier analysesshow the usual de- number of complicating influences.Sampling crease in • with increasing altitude, the rate constitutesa major problem which cannot be 4058 SHARP,EPSTEIN, AND VIDZIUNAS

m

9•0/810 enlOA • OXYGEN-ISOTOPE RATIOS IN THE BLUE GLACIER 4059 intelligentlyhandled without a thoroughunder- fresh water and ice, Publ. ,•00• U. $. Natl. Acad. standing of the structure and constitutionof Sci., 20-28, 1956. Epstein, Samuel, The variation of the 0•/0 •6 ratio the glacier.Analyses of samplesfrom properly in nature and some geologic implications, Re- situateddeep coreholes in valley glacierscould searchesin Geochemistry, John Wiley & Sons, be significant.Ultimately, a classificationof New York, 217-240, 1959. glaciersin terms of the value and rangeof 0•8/ Epstein, Samuel, and T. Mayeda, Variation of 0 •6 0 • ratios shouldbe possible.This is something content of watersfrom natural sources,Geochim. et Cosmochim.Acta, •, 213-224, 1953. for the future, and a number of glaciersin dif- Epstein, Samuel,and R. P. Sharp, Oxygen-isotope ferent environments must be studied first. variations in the Malaspina and Saskatchewan glaciers,J. Geol., 67, 88-102, 1959. Acknowledgments. The work of 1957 and 1958 Hubley, R. C., Glaciers of the WashingtonCas- on Blue Glacier was an International Geophysical cades and Olympic Mountains; their present Year activity under the auspicesof the U.S. Na- activity and its relationto local climatictrends, tional Committee. The investigation was contin- J. Glaciol., ,•, 669-673, 1956. ued in 1959 with support from the National Sci- Hubley, R. C., Glacierresearch on Mt. Olympus, ence Foundation. The mass spectrometer used , Washington,Arctic In- for oxygen-isotopeanalyses is the property of the stitute of North America,mimeographed, 12 pp., Atomic Energy Commission. The National Park 1957. Service graciously granted permission to make the IGY Bulletin, Oxygenisotope studies, Trans. Am. study and aided in many ways. Suppliesand equip- Geophys.Union, ,•0, 81-84, 1959. ment were carried to the glacier by the U• S. Air Kamb, W. B., Ice petrofabric observationsfrom Force and the U.S. Coast Guard. Equipment was Blue Glacier, Washington,in relation to theor•r loaned by the Snow, Ice and Permafrost Research and experiment,J. Geophys.Research, 6•, 1891- Establishment of the U.S. Army Engineers and 1909, 1959. by the Ot/ice of Naval Research (contract Klebelsberg,R. v., Handbuch der Gletscherkunde N-1896-00). Personnel of the Snowdomeproject on und Glacialgeologie,1, Springer,Vienna, 403 upper Blue Glacier collected samplesof snow dur- pp., 1948. ing the winter of 1957-1958.Field colleaguesC. R. LaChapelle, E. R., Blue Glacier, preliminary re- Allen and J. C. Savage aided in the collection of port on the scientificinvestigations, USNC-IGY other samples.William R. Fairchild of the Angeles Project •2, Dept. Meteorol. Climatol., Univ. Flying Service provided superb logistical support. Washington,mimeographed, 28 pp., 1958. John Nye, Henri Bader, and Edward Anders have LaChapelle, E. R., Annual mass and energy ex- kindly offered critical and much appreciated com- change on the Blue Glacier, J. Geophys. Re- ments on the manuscript, but they do not neces- search,6•, 443449, 1959. sarily endorse any statements made in this paper. Landsberg, H. E., Note on the recent climatic REFERENCES fluctuation in the United States, J. Geophys. Research,65, 1519-1525,1960. Allen, C. R., W. B. Kamb, M. F. Meier, and R. P. Reid, II. F., The mechanicsof glaciers,J. Geol., Sharp, Structure of the lower Blue Glacier, •;, 912-928, 1896. Washington, in press, 1960. Sharp,R. P., Featuresof the firn on upper Seward Benson,C. S., Stratigraphy in snowand firn of the Glacier, St. Elias Mountains, Canada, J. Geol., Greenland ice sheet, Ph.D. thesis, Calif. Inst. of 59, 599-621, 1951. Technology, 213 pp., 1960. Epstein, Samuel, Variation oœthe 0•8/0 •6 ratio of (Manuscript received July 9, 1960.)