STRENGTH CHANGES OF LAYERS OF FACETED SNOW CRYSTALS IN THE COLUMBIA ANDROCKY MOUNTAIN SNOWPACK CLIMATES IN SOUTHWESTERN CANADA

KEYWORDSsnow crystals, faceted crystals, snow strength, snow climate, avalanche, avalanche forecasting

Greg Johnson8 and Bruce Jamieson 8, b

Department of Civil Engineering, University of Calgary, Calgary, Alberta T2N 1N4, CanadaDepartment of Geology and Geophysics, University of Calgary, Calgary, Albterta T2N 1N4, Canada

2. LITERATURE REVIEW

3. SNOWPACK CLIMATES

In a recent study, Stock et al. (1998)observed numerous layers of faceted crystals atRed Mountain Pass, Colorado from December toMarch. He measured stuffblock scores (Johnsonand Birkeland, 1994), and hand hardnesses onvalley bottoms and north and south-facing slopes.The study found faceted crystals were larger onnorth-facing slopes and slower to gain stability andstrength.

Jamieson and Johnston (1999) used Kendall­Tau correlations and ranked variables associatedwith the rate of shear strength change for surfacehoar layers. They found the predictor variablesthat most significantly affect shear strength areheight of the snowpack, the maximum crystal size,and slab thickness.

In southwestern Canada there are threesnowpack climate zones: coastal, intermountain,and continental (Fitzharris, 1981; Mock, 1995).Each of these climate zones (Figure 1) hasdifferent weather and snowpack conditions. Thecoastal climate of the Coast Mountains producesrelatively warm temperatures and heavy snowfall.The Rocky Mountains have a continental climateassociated with cold temperatures and shallowsnowpacks. The intermountain climate of theColumbia Mountains is due to an overlap betweenthe coastal and continental weather systems. As aresult the intermountain climate has less snowfall

1. INTRODUCTION

In southwest Montana 59 % ofinvestigated avalanches between 1990 and 1996(Birkeland et. aI., 1998) and in Canada 26 % of thefatal avalanche accidents between 1972 and 1992(Jamieson and Johnston, 1992) were a resultoffailures of layers of faceted crystals. Statistics forCanada show the number of avalanche fatalities isgreatest in the continental snowpack climate of theRocky Mountains and least in the coastalsnowpack climate of the Coastal Mountains(Jamieson and Geldsetzer, 1996). To preventavalanche accidents, avalanche forecasters relyon snowpack data, weather forecasts, andprevious avalanche activity. Often weather,snowpack conditions, and time restrict datacollection, increasing reliance on snowpackevolution models. However, the shear strength oflayers of faceted crystals is often poorly predictedby such models (Fierz, 1998).

Various snowpack variables influence thestrength of layers of faceted crystals, but thesevariables have not been widely studied. In thisstudy, snowpack variables and shear strength oflayers of faceted crystals were measured in thecontinental climate of the Rocky Mountains andthe intermountain climate of the ColumbiaMountains of southwestern Canada. Relationshipsbetween shear strength and snowpack variablesare established by using physical arguments andSpearman rank correlations.

ABS~RACT: The strength of layers of faceted crystals is important for forecasting snow stability. Duringthe winters of 1993-2000, in the intermountain snowpack climate of the Columbia Mountains and thecontinental snowpack climate of the Rocky Mountains of southwestern Canada, over 100 strengthmeasurements of 16 layers of faceted crystals were made. Rank correlations are used to relate thestrength of layers of faceted crystals with measured snowpack properties and calculated snowpackvariables. Additionally, the contrast in snowpack properties between snowpack climates allowscomparison between shear strength and snowpack factors. Factors showing the greatest potential forpredicting shear strength include load and slab thickness.

a Corresponding author Tel: 403-220-7479; Fax 403-282-7026; e-mail: gjohnson72@hotmail.com

86

November December January February March

4. FIELD WORK

Ii1l Columbia Mts. :

t--------- • Rocky MIs.

.E 25 1-------------,-====u .~ 20 ~..

Q)

'6~ 15Cl

~ 10:::l-~~ 5E{! 0

snowpack promotes low temperature gradientsthat are associated with rounding of grains andstrengthening of layers.

Despite clear divisions between inter­continental and continental snowpack climatesvariability of snowpack depth within each climateis common (Mock, 1995). Typically the westernslopes of the Columbia and Rocky Mountainshave thicker snowpacks and are slightly warmerthan the eastern slopes. The eastern slopes of theC?lumbia Mountains have a transitional snowpackclimate between the intermountain and continentalclimates. As a result the snowpack structureincludes more layers of faceted crystals on theeastern slopes than on the western slopes.

Figure 2. Snowpack temperature gradients inthe Columbia and Rocky Mountains ofsouthwestern Canada.(Taken from Johnson, in preparation)

During the winters of 1993-2000, in theColumbia Mountains and the Rocky Mountains ofsouthwestern Canada, over 100 strengthmeasurements of 16 layers of faceted crystalswere made. The shear strength of layers offaceted crystals was measured with the shearframe. If the faceted layer was deeper than 75 cmand older than two weeks it was only tested onceper week, otherwise twice a week. Measurementprocedures for air temperature; layer thickness,layer hardness, snowpack depth, crystal type,crystal size, and water content are described inCM (1995).

4.1 Shear frame testThe shear'frame test was used to

measure the shear strength of the facet layers.The shear strength of the facet layer wascalculated by dividing the recorded minimum forceby the area of the frame. The shear strength is

Fig~re 1. Snowpack climates of the westernUnited States and southwestern Canada(After Mock, 1995) .

• Coastal

~Intennountain

~Continental

than the Coast Mountains, but more than theRocky Mountains (Armstrong and Armstrong1987). Typical mid-winter snowpack depths at treeline are 2 m to 4 m in the Columbia Mountains and1 m to 1.5 m in the Rocky Mountains.

In southwestern Canada there has notbeen a comprehensive study defining differentsnowpack climates. Average snowpack andweather data collected from two sites in the RockyMountains of Canada and three sites from theColumbia Mountains are similar to Armstrong andArmstrong's (1987) and Mock and Birkeland's(1999) results for the western United States.

The continental snowpack of theCanadian Rockies is known for its layers offaceted crystals and depth hoar. Typically thetemperature gradient averaged over the snowpackis greater than the faceting threshold of 100C/m(Akitaya, 1974) from the first snowfall until the endof February (Figure 2) (Johnson, In preparation).The result is layers of depth hoar and facetedcrystals. By March, facets and depth hoar deep inthe snowpack are warm, under a low temperaturegradient, and start to round. However, near­surface faceting still occurs due to localizedtemperature gradients caused by diurnalfluctuations of air temperature and radiation.

In general, the snowpack temperaturegradient never reaches the threshold value of1Q°C/m for faceting in the Columbia Mountains(Figure 2). Most of the faceting occurs due tonear-surface temperature gradients (Birkeland,1998). After layers of facets are buried, a thick

\

87

Table 1. List of predictor factors and response variables.PREDICTOR VARIABLES

~ Shear strength of layers of faceted crystals measured with the shear frame (kPa)~jnter Shear strength of layers of faceted crystals measured with the shear frame in inter-continent

snowpack climates (kPa)~cont Shear strength of layers of faceted crystals measured with the shear frame in continental

snowpack climates (kPa)

5 RESULTS 'AND DISCUSSION5.1 Rank correlations

4.3 Snowpack temperatures

The shear strengths of layers of facetedcrystals are a result of the interaction with varioussnowpack variables. These are termed predictorvariables because they might be useful forpredicting shear strength (Table 1).

The response variable is the shearstrength ~. To assess snowpack factors that areassociated with changes in shear strength thedata were rank correlated in three climaticcategories:

1) continental snowpack climate2) intermountain snowpack climate3) both snowpack climates.

The temperature profile of the snowpawas measured with digital thermometers in 10intervals. Temperatures were also measured 5above, below, and in the faceted layers.

RESPONSE VARIABLES

Weight of the overlying snow per unit horizontal area (kPa)Averaged slab density of snow above layer of faceted crystals (kg/m3

)

Thickness of the slab, the depth of snow over the layer of faceted crystals (cm)Hardness of the faceted layers measured with the hand hardness test (CAA, 1995)Depth of snowpack (cm)Number of days since the layer of faceted crystals formedTemperature gradient measured across layers of faceted crystals. Measurements are taken5 cm above and 5 cm below the layer (oC/m)Temperature of the layer of faceted crystals (0C)Thickness of the layer of faceted crystals (cm)Temperature gradient averaged over the snowpack (oC/m)Maximum length of faceted crystals (mm)Minimum length of faceted crystals (mm)Air temperature (0C)

PslabHRW1

HSAgeTG

Load

TwlLTAlHSEmaxEminTA

then adjusted for the size effects of the shearframe (Sommerfield 1980, Fohn 1987). For adetailed description of the shear frame test seeJamieson and Johnston (In press).

where P is density and h is the layer thickness.The second method is known as a "core

load". Vertical core samples of snow to the depthof a desired layer are extracted from the snowpackand weighed. A cylinder with a cross sectionalarea of 28 cm2 is inserted down into the snowpackuntil it reached the faceted layer or is full. The loadis calculated by dividing the weight of the samplesby the cylinder area and the number of cores.

4.2 Slab load

The slab weight per unit area is oftenreferred to as the load on weak layers. Each timeshear frame tests were performed, slab loadswere measured by two methods. The first methodrequires all snowpack layers above faceted layersbe identified. A cylindrical 100 cm3 sample of snowfrom each layer is taken. The slab load iscalculated by taking the product of gravity and thesum of layer thickness and densities.

88

Correlations with shear strength {Leont} forfaceted layers within continentalsnowpacks.

PS~b

H

RwtHSAgeTGTwlLTAlHSEmaxEminTA

N101

101

101

97101101101

999280999998

R p-Ievel

0.71 7.5E-17

0.68 3.7E-15

0.68 6.5E-15

0.61 4.8E-110.57 4.7E-100.44 5.0E-060.39 5.8E-050.28 4.9E-03-0.23 3.0E-020.23 3.7E-020.17 9.4E-020.05 5.9E-01-0.05 6.1E-01

H

Load

Age

Pslab

TwlHSRW1

TGEmaxEminLTATAIHS

N48

48

48

48474845484646434535

R p-Ievel

0.82 1.0E-12

0.79 2.1E-11

0.76 2.8E-10

0.71 2.2E-080.68 1.3E-070.64 1.2E-060.64 1.8E-06

0.33 2.2E-020.32 3.2E-020.27 6.6E-02-0.27 7.6E-02-0.10 5.1E-010.03 8.5E-01

Pslab

Load

RwI

HAgeEmaxTAHSTAlHSLTGTwlEmin

N53

53

52

53535353534549535253

R p-Ievel0.52 5.7E-05

0.42 1.9E-03

0.40 3.2E-03

0.34 1.3E-020.32 2.1E-020.29 3.5E-020.25 7.2E-020.23 9.8E-020.24 1.1 E-01-0.16 2.6E-010.16 2.6E-010.12 3.8E-010.02 8.7E-01

The response variable Lcont represents shearstrength in continental snowpacks, Linter forintermountain snowpacks, and L for shear strengthin both climates.

The distribution of all shear strengthmeasurements shown in Figure 3 failed theKologorov-Smornov test of normality (d = 0.258, P< 0.01). Since the response variable is notnormally distributed, Spearman rank correlationsare used. This correlation technique only requiresthe data to be at least on an ordinal scale. TheSpearman statistic R ranges from -1 to 1, with -1and 1 being a perfect correlation and 0 indicatingno correlation. The significance level p representsthe reliability of a correlation. A p-Ievel of 0.05indicates there is a 5% probability of the analysisrevealing a correlation in uncorrelated data(Statistica, 1999).

Serial correlations are a measure of the

III 40 Ic:0..III~ 30CDIII.ca 20~

0..CD 10 ;.cE~z

o-0.5 0.5 1.5 2.5 3.5 4.5 5.5 6.5

Shear Strength (kPa)

Figure 3. Distribution of shear strength andexpected normal curve

89

relationship between past observations andpresent observations of a variable (Chatfield,1980, pg 23-32). They are present in the data andcause overestimates of significance levels forcorrelations between response and predictorvariables.

5.2 Results of rank correlations

Snowpack properties that correlated withshear strength of faceted snowpack layers arelisted in terms of their statistical significance inTable 2 and cross-correlations between thepredictors are listed in Table 3. In continentalsnowpack climates shear strength Leont

significantly correlated with 6 snowpack variables(p < 0.05). In intermountain climates shearstrength Linter significantly correlated with 9snowpack properties. When the data were notpartitioned into a climatic region, shear strength Lcorrelated with 10 snowpack properties. Thevariables that positively correlated with shearstrength in all three categories are load, age,thickness of the slab, density of the slab, and handhardness. In addition to these properties themaximum crystal size correlated with shearstrength in the continental Leont and intermountainLinter climates.

5.3 Age

Age positively correlated with shearstrength in all three correlation categories(Table 2). However, when strength is plotted overtime it reaches maximum values that depend onsnowpack climate (Figure 4).

16 -.--~~--~~~"'""I-----

5.4 Load

Hand hardness roughly measures theresistance to penetration. It is an index of strengThe positive correlation (p = 4.8E-11, Table 2)between hand hardness and shear strength L isexpected.

In this study faceted crystals in continensnowpacks reached maximum hand hardnessesof approximately 1F+, but rounding facets inintermountain snowpacks sometimes reached P.The difference in hand hardnesses for eachsnowpack climate is primarily due to load(p = 6.2E-8, Table 3); load is associated with boformation and growth.

The temperature gradient weakly butpositively correlated (p = 2.1E-2, Table 2) with thshear strength Linter of layers of faceted crystalS ithe intermountain snowpack of the ColumbiaMountains. However, an effect of temperaturegradient on shear strength is not expected sincemost of the temperature gradient data hasmagnitudes less than 1aOClm.

5.6 Slab density (Ps/ab)

5.7 Hand hardness (RwJ

5.9 Temperature gradient (TG)

The temperature of the weak layer (TwI)positively correlated (p = 1.3E-7, Table 2) withshear strength in intermountain snowpacks, butdid not for continental snowpacks. In the deepsnowpacks of the Columbia Mountains,temperatures gradually increase over the winter,promoting rounding of the faceted crystals andbond growth. The temperature of the weak layercorrelated with slab thickness (p = 1.2E-4,Table 3) and age (p = 4.8E-15) suggesting aslayers age they are buriedly deep and wellinsulated from cold air.

In continental climate of the RockyMountains the temperatures of weak layers slowarm throughout the winter. However, thinoverlying slabs reduce insulation from cold air. Asa result the temperature of weak layers undergofluctuations that may reduce correlations.

5.8 Temperature of the weak layer (Twl)

Shear strength positively correlated withslab density Pslab (Table 2) in all three strengthcategories. This is not surprising sinceslab density is a function of slab thickness(p = 1.6E -12, Table 3) and load (p = 1.5E-22).

20 40 60 80 100 120 140

Age (days)

Figure 4. Shear strength and load overtime inthe continental and intermountain snowpackclimates.

Slab thickness strongly correlated withstrength Linter for intermountain (p = 1.0E-12,Table 2) and for continental snowpacks LeonI (p =1.3E-2). Slab thickness is expected to affectstrength because thicker slabs are associated withlarger loads (p = 1E-27, Table 3), warmertemperatures (p = 1.2E-4) , and low temperaturegradients (p = 5.0E -9).

5.5 Slab thickness (H)

Load positively correlated with the shearstrength of facet layers Linler in the intermountainsnowpack (p = 2.1E-11, Table 2) more significantlythan the continental snowpack LeonI (p = 1.9E-3,Table 2). Load is also more significantly correlated(p = 7. 5E-17, Table 2) to shear strength L thanany other snowpack property. These threecorrelations show the strong dependence of shearstrength on load, until maximum shear strengthwas reached.

In each climate, maximum shear strengthswere reached after different periods of time.Maximum shear strength was reached afterapproximately 90 days within the continentalclimate and approximately 60 days within theintermountain climate (Figure 4). After 60 days theload in the intermountain climate averagedapproximately 5.25 kPa. After 90 days the load incontinental climates averaged approximately 1.75kPa (Figure 4). Beyond these points, loadscontinued to increase, but other factors (such aslayer temperature, previous temperature gradientand crystal size) may prevent further increases instrength.

-<2;- Continental Shear Strength10 14~ -& Continental Load-12'g .. Intermountain Load

.3 10 -+- Intermountain Shear Strength

] 8 -l================='~~---.s=0,c~

en~

ellGl

.s=(/) 0 +.---r-'----r-----.----r-----,-----,-------r-

o

90

1.1E-042.1E-041.7E-07

4.7E-054.5E-011.8E-03

9.0E-032.9E-013.1 E-013.9E-06

3.9E-075.0E-091.2E-049.6E-031.6E-12

1.0E-276.2E-082.1 E-083.5E-069.7E-041.5E-22

3.1 E-146.1 E-202.0E-061.3E-055.9E-017.4E-012.0E-04

HSLoadHRWI

TGTwlEmax

~Iab

Table 3. p-values for cross correlations between snowpack factors. Correlations are considered

significant if p < 5.0E-2.- Age HS Load H RW1 TG Twl Emax

3.1 E-018.9E-122.7E-071.9E-028.2E-044.8E-158.0E-149.8E-19

The temperature gradient did not correlatewith shear strength in continental snowpacks LeonI'

Temperature gradients in continental climatesfluctuate and may remain greater than 10°C/m formonths. As a result, weak layers of well developedfaceted crystals form. After temperature gradientsdissipate weak faceted layers are slow to gainstrength because the crystals are large and theyare under little load.

5.10 Maximum Crystal Size (Emax)

The maximum crystal size weaklycorrelated (p = 3.2E-2, Table 2) with strength-inboth snowpack climates, but field workers regard itas an important property that affects shearstrength. The positive correlation indicates thatlarger crystals have higher strengths. Both fieldand cold laboratory observations show increasesin crystal sizes when faceted crystals are rounding(Johnson In preparation). This suggests growth ofrounding faceted crystals is associated withstrengthening.

5.11 Faceted crystal type

Faceted crystals are classified as solidfaceted particles (4a) or faceted particles withrounding of facets (4c) based on the appearanceof rounding (Colbeck et a!., 1990). Field workersusually expect increases in strength when facetedcrystals show signs of rounding (4c), not furtherfaceting (4a). In each of the layers of facetedcrystals from the Columbia Mountains, roundingw~s observed from mid-December throughout thewinter. The layers from continental snowpackareas were not reported as rounding facetedcrystals until mid-to-Iate March.

6.0 CONCLUSIONS

Load and thickness of the slab correlatedmost significantly with shear strength of layers of

faceted crystals in intermountain and continentalsnowpack climates. These snowpack climatesprovide contrasts in load and thickness of slabs.

The slab thickness is associated withstrengthening of layers of faceted crystals throughits association with load and low temperaturegradients. Thick slabs insulate weak layers offaceted crystals from cold air temperatures,

. causing warm weak layer temperatures and lowtemperature gradients.

Load had a greater correlation with shearstrength in the intermountain climate than in thecontinental climate. The difference in correlations(Figure 4) shows that large loads cause facetedlayers to gain strength quickly. Physically, moreload pushes faceted crystals closer togetherresulting in increased density, number of contacts,and bond size.

ACKNOWLEDGEMENTS

For their careful field work we are gratefUl to JillHughes, Leanne Allison, Ken Black, ThomasChalmers, Joe Filippone, Michelle Gagnon,Torsten Geldsetzer, Jeff Goodrich, Phil Hein, NickIrving, Colin Johnston, Crane Johnson, AlanJones, Kalle Kronholm, Paul Langevin, Bill Mark,Bruce McMahon, Andrew Rippington, JordyShepherd and Adrian Wilson.

For their assistance with field studies, we thankthe avalanche control section of Glacier NationalPark including Dave Skjonsberg and BruceMcMahon, Mike Wiegele Helicopter Skiing,Canadian Mountain Holidays, and the BC Ministryof Transportation and Highways.

This study was funded by the Natural Sciencesand Engineering Research Council of Canada,Canada West Ski Areas Association, theCanadian Avalanche Association, IntrawestCorporation and the BC Helicopter and SnowcatSkiing Operators Association (BCHSSOA). The

91

supporting members of the BCHSSOA includeCanadian Mountain Holidays, Cat Powder Skiing,Crescent Spur Helicopter Holidays, GreatCanadian Helicopter Skiing, Great Northern SnowCat Skiing, Island Lake Lodge, Klondike Heli­Skiing, Last Frontier Heliskiing, Mike WiegeleHelicopter Skiing, Monashee Powder Adventures,Peace Reach Adventures, Purcell HelicopterSkiing, R.K. Heli-Skiing, Retallack AlpineAdventures, Robson Heli-Magic, Selkirk TangiersHe/i-Skiing, Selkirk Wilderness Skiing, Sno MuchFun Cat Skiing, TLH Heliskiing, Whistler Heli­Skiing and White GrizZly Adventures. Thesupporting members of Canada -West Ski AreasAssociation include Apex Mountain Resort, BanffMt. Norquay, Big White Ski Resort, Hemlock SkiResort, Mt. Washington Alpine Resort, Silver StarMountain Resorts, Ski Marmot Basin, Sun PeaksResort, Sunshine Village, Whistler Blackcomb,Whitewater Ski Resort, and the Resorts of theCanadian Rockies including Skiing Louise,Nakiska, Kimberley Alpine Resort, FortressMountain and Fernie Alpine Resort

References

Akitaya, E. 1974, Studies on depth hoar.Contributions from the Institute of LowTemperature Science, Series A, 26, 1-67.

Armstrong, R.L and Armstrong, B.R. 1987, Snowand avalanche climates of the western UnitedStates. Avalanche Foramation, Movement, andEffects. IAHS Pub. 162 Wallingford, Oxen:International Association of HydrologicalSciences. 281-294

Birkeland, K. 1998, Terminology and predominantprocesses associated with the formation ofweak Layers of near-surface faceted crystals inthe mountain snowpack. Arctic and AlpineResearch. 2(30) 193-199.

Birkeland, K. Johnson, R., Schmidt, D. 1998,Near-surface faceted crystals formed by diurnalrecrystallization: A case study of weak layerformation in the mountain snowpack and itscontribution to snow avalanches. Arctic andAlpine Research 2(30) 200-204.

Canadian Avalanche Association. 1995,Observation guidelines and recordingstandards for weather, snowpack and

92

avalanches. Canadian Avalanche AssOciatiP.O. Box 2759, Revelstoke, BC, Canada, 98pp.

Colbeck, S, E. Akitaya, R. Armstrong, H. GUblerJ.Lafeuille, K. Lied, D. McClung and E. Mo :1990, International classification for seasonsnow on the ground. International Commissifor Snow and Ice (IAHS), World Data Centerfor Glaciology, U. of Colorado, Boulder, CO,USA.

Fierz, C. 1998, Field observation and modelling 0weak-layer evolution. Annals of Glaciology 26,7-13.

Fitzharris B.B. 1981, Frequency and climatologymajor avalanches at Rogers Pass, 1909-19n.Ottawa National Research Council, CanadaAssociation Committee on GeotechnicalResearch, DBR Paper No. 956.

Fohn, P.M.B. 1987, The stability index and varioutriggering mechanisms. In: B. Salm and H.Gubler, eds., Avalanche Formation, Movementand Effects. International Association ofHydrological Sciences, Publication No. 162,195-211.

Jamieson, J.B. and C.D. Johnston. 1992,Snowpack characteristics associated withavalanche accidents. Canadian GeotechnicalJournal 29, 862-866.

Jamieson, J.B. 1995, Avalanche prediction forpersistent snow slabs. PhD Thesis, Dept. ofCivil Engineering, University of Calgary, 275pp.

Jamieson, J.B. and T. Geldsetzer, 1996.Avalanche accidents in Canada Volume 41984­1996. Canadian Avalanche Association,Revelstoke, B. C Canada.

Jamieson, J.B. and C.D. Johnston. 1999,Snowpack factors associated with strengthchanges of buried surface hoar layers. ColdRegions Science and Technology. 30, 19-34

Jamieson, J.B. and C.D. Johnston. (In Press),Evaluation of the shear frame test for weaksnowpack layers Annals of Glaciology 32.

Johnson, G. (In preparation), Observations offaceted snow crystals. MSc. Thesis, Dept ofCivil Engineering, University of Calgary.

Johnson, R. and K. Birkeland. 1994, Thestuffblock: A simple and effective snowpackstability test. International Snow ScienceWorkshop. PO Box 49 Snowbird, Ut.

Mock C.J., 1995, Avalanche climatology of thecontinental zone of the southern RockyMountains USA. Physical Geography 16(3),165-187.

Mock C. and K. Birkeland. 1999, Avalancheclimate regions of the western United States:an update. Avalanche Review 17(3).

Sommerfeld, R.A. 1980, Statistical models of snowstrength. Journal of Glaciology 26(94),217­223.

Statistica, 1999, Electronic Manual.

Stock, J. K. Elder, and K. Birkeland. 1998, Thetemporal variations of near-surface facetedcrystals, Red Mountain Pass CorridorColorado. ISSW 1998 .'

93