linksbetween short-termvelocity variations and the subglacial … · 2012. 1. 11. · linksbetween...

Links between short-term velocity variations and thesubglacial hydrology of a predominantly cold polythermal

glacier

Luke COPLAND1 Martin J SHARP1 PeterW NIENOW2

1Department of Earth and Atmospheric Sciences University of Alberta Edmonton AlbertaT6G 2E3 CanadaE-mail martinsharpualbertaca

2Department of Geography andTopographic Science University of Glasgow Glasgow G12 8QQ Scotland

ABSTRACT The surface velocity of a predominantly cold polythermal glacier (JohnEvans Glacier Ellesmere Island Canada) varies significantly onboth seasonal and short-er time-scales Seasonal variations reflect the penetration of supraglacial water to theglacier bed through significant thicknesses of cold ice Shorter-term events are associatedwith periods of rapidly increasing water inputs to the subglacial drainage system Early-season short-term events immediately follow the establishment of a drainage connectionbetween glacier surface and glacier bed and coincide with the onset of subglacial outflowat the terminus A mid-season short-term event occurred as surface melting resumed fol-lowing cold weather and may have been facilitated by partial closure of subglacial chan-nels during this cold period There is a close association between the timing and spatialdistribution of horizontal and vertical velocity anomalies the temporal pattern of surfacewater input to the glacier and the formation seasonal evolution and distribution of sub-glacial drainage pathways These factors presumably control the occurrence of high-water-pressure events and water storage at the glacier bed The observed couplingbetween surface water inputs and glacier velocity may allow predominantly cold poly-thermal glaciers to respond rapidly to climate-induced changes in surface melting

1 INTRODUCTION

Studies of temperate and predominantly warm polythermalglaciers (polythermal structure types b and c of Blatter andHutter 1991 fig 1) have clearly demonstrated that there is anintricate coupling between the subglacial hydrology and flowdynamics of such glaciers (Iken1981 Bindschadler1983 Ikenand others 1983 Iken and Bindschadler 1986 Kamb 1987Jansson 1995 1996 Raymond and others 1995 Harbor andothers 1997 Kavanaugh and Clarke 2001 Mair and others20012002)This coupling is especially important when surfacemeltwater is able to penetrate to the glacier bed and it isfundamental to such dynamic phenomena as glacier surgesseasonal velocity variations and short-term high-velocityevents Whether and how this coupling affects the dynamicsof glaciers composed predominantlyof ice at sub-freezing tem-peratures is less clearThe goal of this paper is therefore to in-vestigate the relationships between surface melt subglacialhydrology and the flow of a predominantly cold polythermalglacier in the Canadian High Arctic

Predominantly cold polythermal glaciers (polythermalstructure types d and e of Blatter and Hutter 1991 fig 1) arecharacterized by a thick mantle of cold ice overlying a limitedarea of temperate ice at and immediately above the glacierbed in the ablationarea Examples includeWhite GlacierAxelHeiberg Island Canada (Blatter 1987) and McCall GlacierAlaska USA (Rabus and Echelmeyer1997) A recent review(Hodgkins1997) suggests that penetration of surface-derivedmeltwaters to the beds of such glaciers is limited This is

because (a) water flow along intergranular vein networks islargely absent in cold ice and (b) crevasses and moulinswhich introduce large-scale permeability to glacier ice arerare on predominantly cold glaciers due to low rates of ice de-formation and the refreezing of meltwater that drains into cre-vasses If this were the case the influence of surface melting onthe flow of predominantly cold polythermal glaciers would belimited

Several studies however suggest that the surface velocitiesof such glaciers varyon a seasonal basis (Mulaquo ller and Iken1973Iken1974 Andreasen1985 Rabus and Echelmeyer1997)Thisis also the case on the Greenland ice sheet where melt-inducedseasonal variations in surface velocity have been observed in aregion where cold ice is 41200m thick (Zwally and others2002) These studies imply that surface meltwaters can and dopenetrate to the glacier bed through significant thicknesses ofice at sub-freezing temperatures contrary to the suggestion ofHodgkins (1997) They also suggest that coupling between sur-face melting and the flow of predominantly cold but warm-based ice masses may provide a mechanism by which such icemasses can respond rapidly to changes in surface weather andclimate (Zwally and others 2002) Given model predictionsthat anthropogenic climate warming will be most marked innorthern high latitudes (Manabe and others 1991) and thepotential contribution of Arctic glaciers to global sea levelunderstanding of this coupling is a high scientific priority

11 Study site

The study was conducted at John Evans Glacier a sup1165 km2

Journal of Glaciology Vol 49 No 166 2003

337

predominantly cold polythermal valley glacier on the eastcoast of Ellesmere Island Nunavut Canada (79sup340rsquo N74sup330rsquo W Fig 1) The glacier ranges in elevation from 100 to1500 m asl with the long-term equilibrium line at sup1750^850m asl Ice depths reach a maximum of 400m in theupper ablation area and average 100^250m in the lower

ablation area where the velocity measurements were made(Copland and Sharp 2001) For the period1997^99 the meanannual air temperature at 820m asl was ^152sup3C

High bed reflection powers in radio-echo soundingrecords indicate warm ice at the bed throughout most ofthe lower ablation zone except along the glacier marginsand where the ice is thin (Copland and Sharp 2001) A con-tinuous internal reflecting horizon over the centre of thelower terminus suggests that the warm basal ice reaches anaverage thickness of 20 m there In the accumulation andupper ablation areas low bed reflection powers and 15 mborehole temperatures of ^95 to ^151sup3C suggest that theice is cold throughout

The melt season typically occurs between earlyJune andearly August At the start of the melt season meltwatereither ponds on the glacier surface or is routed directly tothe ice margins and is unable to access the glacier interiorOnce englacial drainage of supraglacially derived melt-waters is initiated however approximately 25 of theglacier surface area drains into moulins in a crevasse fieldat the top of the terminus (Fig 2) Later in the melt seasonmeltwater from up to 40 of the surface area of the glacierdrains to the glacier bed via these moulins and a second cre-vasse field sup16 km further upstream Dye tracer experimentsconfirm the link between these water input locations and amajor drainage portal at the terminus (Bingham andothers 2003) Large increases in the suspended-sedimentcontent and electrical conductivity (EC) of the water as itpasses through the glacier suggest that it is routed sub-glacially (Skidmore and Sharp1999)

Inthe early summer the meltwater is trappedat the glacier

Fig 1 Bed topography (m asl) and location ofJohn EvansGlacier

Fig 2 Landsat 7 image of John Evans Glacier (path 049 row 002 10 July 1999) showing the location of the velocity stakes(black dots) geophones (white dots LG ˆ lower geophone MG ˆ middle geophone UG ˆ upper geophone) surveying basestations (white squares) stream gauging stations (triangles) and artesian fountain observed in 1998 ( ) Crevasse field indi-cates location where most supraglacial meltwater reaches the glacier bed Lower weather station is located next to the middlegeophone Black box indicates horizontal extent of Figures 4 and 7

Copland and othersVelocity variations and subglacial hydrology of a polythermal glacier

338

bed behind the frozen terminus (Skidmore and Sharp1999)The initial release of subglacially routed meltwater from theglacier typically occurs via an artesian fountain on theglacier surface (Fig 3) andor upwelling through sedimentsa few metres in front of the snout Since observations beganin 1994 dates of initiation of subglacial outflow have variedbetween 22 June and11July (days173^192) As outflow con-tinues the upwelling migrates towards the snout and iseventually replaced by the formation of an ice-walled chan-nel at the glacier bed (the major drainage portal referred toin the previous paragraph) The first water to be releasedhas high total solute concentrations (EC 4 300 ms cm^1) issomewhat turbid and relative to supraglacial runoff con-tains elevated concentrations of ionic species (Na+ K+ andSi) that are products of silicate weathering (Skidmore andSharp 1999) Within a few days the water becomes moredilute although solute concentrations are still much higher(4100 ms cm^1) than for supraglacial water (510 ms cm^1)These observations suggest that the first water to be releasedhas been stored for a long period (possibly over winter)while later waters have been transmitted more rapidlythrough the englacial and subglacial drainage system

2 MEASUREMENTS

21 Surface velocity

Twenty-one velocity stakes were established over the lower-most 2 km of John Evans Glacier in 1998 An additional 13stakes were added in 1999 to extend the network 1kmfurther up-glacier (Fig 2) The terminus region was chosenfor study as it is easily accessible there are surrounding cliffsthat allow survey stations to be located onbedrock and sub-glacial water flow appears to be present (Skidmore andSharp 1999 Copland and Sharp 2001) Reflecting prismswere mounted on 3 m long stakes drilled and subsequentlyfrozen into the ice surfaceThese were periodically redrilledor replaced before surface melting made them unstable

The position of each prism was measured daily (weatherpermitting) in the summers of 1998 and 1999 with a Geo-dimeter 540 total station theodolite To reduce errors thedata were subsampled to determine stake displacementsover periods of 2 days or longer The time between surveys

was used to convert the displacements to velocities in cm d^1

(24 hours) Velocity patterns were interpolated over theentire terminus region from the point measurements at thestakesVelocities were set to 0 around the glacier edge whichis reasonable given that the glacier appears to be frozen to itsbed at the margins (Copland and Sharp 2001) All interpo-lations were completed with the ` v4rsquorsquo interpolation routinein Matlab which is based on biharmonic spline interpola-tion (Sandwell 1987) Interpretations of velocity patternsare made only for areas where stakes are present For thepurposes of our discussion the lower terminus is defined asthe area south of the upper geophone (UG in Fig 2) and theupper terminus is defined as the area to the north andnorthwest of this geophone

22 Surface velocity errors

From the Geodimeter technical specifications uncertaintiesin distance measurements amount to sect (2 mm Dagger 3 ppm)This equates to sect5 mm over a typical survey distance of1km Angle measurements were made to a resolution of 2rsquorsquowhich equates to a maximum potential error of sect48 mmover a distance of 1km Instrument drift during surveyswas corrected for by resurveying the position of referencemarkers after every seven or fewer stake measurements andassuming that drift was linear between surveys In the fol-lowing discussion three types of error are evaluated

(i) Position error (sect cm) this is the error in determining thelocation of a stake during a single survey

(ii) Displacement error (sect cm) this is the error in determin-ing the displacement of a stake between two surveys Itis calculated by summing the position errors from thetwo surveys andor from two or more independentmeasurements of the displacement of a stake

(iii) Velocity error (sect cm d^1) this is the displacement erroradjusted for the time between surveys As the measure-ment period increases the velocity error decreases asthe displacement error is divided by a greater time in-terval

In 1998 all stakes were surveyed independently fromeach of two stations located sup1100 m apart For each meas-urement period this provided two measures of the displace-ment of each stake and the displacement error wascalculated from the difference between these and the meandisplacement of the stake for that survey Separate velocityerror calculations were then made for each stake for each2 day measurement period These are displayed in thevelocity figures discussed later For the entire1998 measure-ment period mean errors for all stakes are sect11cm d^1 inhorizontal velocity and sect06 cm d^1 in vertical velocity

In 1999 the displacement of each stake was determinedonce for each measurement period from surveys at a singlestation To evaluate errors the location of each stake wasmeasured at least twice from the same station during eachsurvey in July and August On a given day the differencebetween the two or more measurements of a stake and itsmean location provides a position errorThe position errorsfrom the surveys at the start and end of a measurementperiod then provide the displacement error which in turnprovides the velocity error The velocity errors were calcu-lated separately for each stake for each 2 day measurementperiod For summer 1999 the mean errors for all stakes were

Fig 3 Artesian fountain observed on the lower terminusbetween days 180 and 186 1998 see Figure 2 for location

339


sect06 cm d^1 in horizontal velocity and sect02 cm d^1 in verti-cal velocity

The following discussions focus mainly on the 1999measurements due to the lower velocity errors improvedspatial coverage and better records from supporting instru-ments In these discussions the stated start and end times ofvelocity `eventsrsquorsquoare approximate due to the 2 day resolutionof the measurements This likely leads to underestimation ofthe true magnitude of velocity changes since events will likelyspan more than one measurement period or be shorter thana measurement period

23 Geophones

In 1997 three geophones were installed at 5 m depth along thecentre of the terminus a lower geophone sup103 km up-glacierfrom the snout a middle geophone sup11km from the snout andan upper geophone sup12 km from the snout (Fig 2) The45 Hz geophones were interfaced to a Campbell Scientificdata logger which counted the number of seismic eventsabove a defined threshold and output a total every hour (in1998) or 2 hours (in 1999) Geophone sensitivity was adjustedto a level that provided a good compromise between sensitiv-ity and noise rejection in studies at Trapridge GlacierCanada (Kavanaugh and Clarke 2001) The geophonerecords did not allow for accurate location of individual seis-mic events because it was not possible to distinguish a weaklocal event from a strong distant one although it was possibleto determine general positioning by noting which geophonesrecorded activity

Given the lack of volcanic activity and of active faults inthis area it is assumed that most detected events had aglacial origin Previous studies at other glaciers have shownstrong correlations between ice velocity water dischargeand seismicity (Iken and Bindschadler 1986 Raymondand others 1995 Kavanaugh and Clarke 2001) This isbecause ice fracturing commonly occurs during periods ofrapid ice velocity due to changes in the internal stress distri-bution In addition basal sliding may produce seismicevents as the ice moves across the glacier bed

24Weather stations

Three automatic weather stations were installed on JohnEvans Glacier in May1996The lower station is located nextto the middle geophone in the centre of the terminus at sup1200m asl (Fig 2) Hourly measurements of air temperature andsurface albedo from this station were used to quantifychanges in weather and ice surface conditions Rates of sur-face lowering recorded by an ultrasonic depth gauge(UDG) provide an indication of relative rates of meltwaterproduction These records have not been corrected for dif-ferences in surface density between ice and snow so provideonly a general measure of surface melt rates

25 Discharge records

To provide an indication of the meltwater inputs to the sub-glacial drainage system in 1999 pressure sensors wereinstalled upstream of the main crevasse field in an ice-mar-ginal lake and at two locations in a supraglacial streamdraining from the lake (Fig 2) To monitor the flow of waterfrom the glacier a pressure sensor was installed in the mainsubglacial stream shortly after it exited the snoutThis pro-glacial record is approximate due to frequent bed aggrada-

tion and erosion around the sensor and the need to move thesensor because of channel migration The sensors were notcalibrated to discharge so each record was rendered dimen-sionless by scaling it between the highest and lowest waterlevels over the period of interest

26 Subglacial flow routing

Copland and Sharp (2000) reconstructed subglacial flowrouting at John Evans Glacier from the form of the sub-glacial hydraulic potential surface using the method ofShreve (1972) Reconstructions were made for assumed basalwater pressures at 0^100 of ice overburden pressure butno significant differences were found in predicted flow rout-ing This indicates that subglacial topography rather thanvariations in ice thickness provides the dominant controlonbasal water flow atJohn Evans GlacierThe drainagepat-tern presented here is for basal water pressures at 100 ofice overburden pressure

27 Residual vertical velocity (cavity opening)

To determine the importance of basal processes in account-ing for the measured vertical velocities the residual verticalvelocity (wc) was calculated for each survey This is the ver-tical velocity remaining after vertical strain and the hori-zontal movement along a sloping bed have been accountedfor and is commonly attributed to basal cavity formation(Iken and others 1983 Hooke and others 1989) In realityit may be caused by any unaccounted-forbasal process suchas a change in the dilatancy of subglacial till wc was calcu-lated from (Hooke and others1989)

wc ˆ ws iexcl ub tan shy iexcl _zzh hellip1daggerwhere ws is measured Lagrangian vertical velocity at thesurface ub is horizontal basal velocity shy is bed slope _zz isdepth-averaged vertical strain rate and h is ice thicknessFor these calculations measured values of us (horizontalsurface velocity) interpolated to a regular 100 m grid wereused instead of ub since basal velocity was not measureddirectly Since ub micro us in general this will produce a min-imum estimate of the rate of cavity opening Ice thicknesswas determined by radio-echo sounding (Copland andSharp 2001) and interpolated to the same 100 m grid Bedslope was calculated from the change in bed elevation acrossthe four gridpoints to the north east south and west of thepoint of interest

In Equation (1) _zzjs (the measured vertical strain rate atthe surface) was initially used as an approximation to _zz Byassuming incompressibility _zzjs was determined fromthe sur-face horizontal strain rates ( _zzjs ˆ iexcl _xxjs iexcl _yyjs) which werecalculated from the gradients in surface horizontal velocitybetween adjacent gridpoints In reality _zzjs is unlikely to bethe same as _zz at depth (Balise and Raymond 1985 Hookeand others1989) so it is necessary to assess how vertical vari-ations in _zz would affect calculations of wc

If it is assumed that there is no cavity opening (wc ˆ 0)and that 0 5 ub 5 us then for a given ice thickness it is pos-sible to determine whether and how _zz must vary withdepth (Hooke and others 1989 Mair and others 2002) Fornegative values of shy (ie when the glacier is flowing down-hill) three cases are possible

(a) hellipwsdagger micro hellip _zzjshdagger micro hellipws iexcl us tan shy dagger _zz is not requiredto change with depth relative to _zzjs


340

(b) hellip _zzjshdagger gt hellipws iexcl us tan shy dagger _zz is required to decreasewith depth relative to _zzjs

(c) hellip _zzjshdagger lt hellipwsdagger _zz is required to increase with depthrelative to _zzjs

These equations are reversed in the more unusual casewhen shy is positive (ie when the glacier is flowing uphill)Basal vertical displacement is unlikely to have occurred ifthe surface vertical velocity can be explained by a combin-ation of basal sliding and vertical strain rates that are eitherconstant or decreasing with depth (ie scenarios a and babove) Basal vertical displacement is most likely to haveoccurred if the surface vertical velocity is higher than canbe plausibly explained by an increase in _zz with depth rela-tive to _zzjs (scenario c above) Classification of the glacierinto areas where each of these cases holds for each velocityevent therefore helps in defining areas where basal verticaldisplacement is likely to have occurred Hooke and others(1989) argued that a value of _zzjsh that is sup13 mm d^1 lessthan ws (when shy is negative) is likely indicative of cavityopening under ice sup1100 m thick To ensure that areas ofbasal vertical displacement are conservatively identified atJohn Evans Glacier discussion is limited to areas where theinferred rate of cavity opening is at least 1cm d^1

3 RESULTS

31 Long-term velocities

Velocities were calculated for winter 199899 (day 2051998to day 146 1999) and 19992000 (day 214 1999 to day 1612000) to provide a measure of the ` backgroundrsquorsquo velocitiesagainst which short-term variations can be compared Onlythe winter 19992000 velocities are presented here as therecords cover a larger area and are very similar to the199899 records (Fig 4) For the stakes that were measuredin both winters differences average 014 cm d^1 in horizontalvelocity and 057sup3 in horizontal direction

The mean horizontal velocity for all stakes was35 cm d^1 for winter 19992000 compared to a mean hori-zontal velocity of 53 cm d^1 for summer 1999 (days 17550^204881999) (Fig 5) Horizontal velocities for summer 1998were also sup150 higher than winter values which suggeststhat basal sliding is an important process during at least thesummer at John Evans Glacier This seasonal increase in

Fig 4 (a) Winter 19992000 horizontal velocity contoursEach arrow indicates the direction and velocity of a stakeSquares indicate stakes grouped into factor 1 in principal com-ponent analysis (PCA see end of section 31 for explanation)circles indicate stakes grouped into factor 2 (b) Winter19992000 vertical velocity contours All velocities in cm d 1

Fig 5 Summer 1999 mean horizontal velocities (solid blackline) winter 19992000 mean horizontal velocities (dashedline) errors in horizontal velocity (vertical bars) and standarddeviation in horizontal velocity (grey lines) (a) for all stakes(b) for stakes grouped by factor 1in PCA(ie lower terminus)and (c) for stakes grouped by factor 2 in PCA (ie upper termi-nus) (d) Factor scores for the two factors (ie the standardizedPCA scores on the factors over the measurement period)

341


horizontal velocity is not distributed evenly however but isconcentrated during 2^4 day high-velocity events whichprovide the focus of the discussion belowVertical velocitiesare low during the winter reaching a maximum of approxi-mately ^05 cm d^1 in an area of steep surface slopes betweenthe upper and lower terminus (Fig 4b) Mean verticalvelocities generally vary little between winter and summeralthough significant variations do occur during some high-horizontal-velocity events (Fig6)

To assess the spatial scale of forcing mechanisms andhow these change over time principal component analysis(PCA) was used to define groups of stakes with coherentpatterns of horizontal velocity variation (Johnston 1978)From analysis of all 2 day horizontal velocity data over days16471^211981999 PCA identified eight principal compon-ent factors with eigenvalues 41 (Table 1) When the 34velocity stakes are grouped based on their loadings on thesefactors two main clusters are identified factor 1 groupsstakes from the lower terminus while factor 2 groups stakesfrom the upper terminus (Fig 4a) Grouping using morethan the first two factors subdivides these two clusters intosmaller groups that are not as spatially contiguous In addi-tion the percentage of variance explained by the other fac-tors is generally low (Table 1) Since most of the variance inhorizontal velocity is explained by the first two factors(639) and the loadings on these factors cluster stakes intopopulations that act differently during high-velocity events(see below and Fig 5d) these groupings are used in discus-sion of the 1999 short-term velocity data PCA was not per-

formed on the 1998 velocity data due to the relatively smallnumber of stakes in that year

32 1999 short-term velocity events

321 Event 199 days 18460^18904This event encompasses two measurement periods days18460^18694 and days 18694^18904 during which meanhorizontal velocity was almost 100 above winter levels at67 cm d^1 (Fig 5a) There was little change in either meanhorizontal or mean vertical velocity between these periods(Figs 5 and 6) so the mean velocities from the two periodscombined are presented in Figure 7a^d

The relative increases in horizontal velocity during event199 were spatially non-uniform the mean velocity of stakesin the lower terminus increased by 110 relative to winterlevels while that of stakes in the upper terminus increasedby 75 (Fig 5b and c) Compared to winter the centralregion of highest velocities broadened and there was a con-sequent narrowing of the marginal shear zones particularlyover the lower terminus and close to the snout (Figs 4a and7a and b) The largest absolute velocity anomalies of up to6 cm d^1 occurred over the upper terminus close to wherewinter velocities were highest and close to the predicted sub-glacial drainage pathway there (Fig7b) Over the centre ofthe lower terminus horizontal velocities were an almost con-stant 3 cm d^1 above winter levels (Fig 7b) and there was amarked rotation of the surface velocity vectors (relative tothe winter vectors) towards the west (Fig 7a cf Fig 4a)The rotation of the velocity vectors increased towards theglacier snout where it reached 445sup3 at some stakes

There was little change in vertical velocity from winterlevels with most vertical velocities close to 0 cm d^1 (Figs 6and 7c) Residual vertical velocity was also generally lowwith only a limited area over the lower terminus wherecavity opening is suggested by the apparent need for _zz toincrease with depth relative to _zzjs (Fig7d)

Event199 occurred at the end of the first week of intensesummer melting (Fig 8a) As indicated by the albedo andUDG records from the lower weather station surface melt-ing had been occurring for approximately 1month prior toevent 199 (Fig 8b and c) although much of this melt mayhave refrozen in the cold snowpack By the start of the eventthe albedo had fallen to 40 (Fig 8b) which defines theapproximate boundary between a snow- and ice-coveredsurface (Paterson 1994 p59) The surface lowering ratewas continuously positive during the event (Fig 8d) The

Fig 6 Summer 1999 mean vertical velocities (solid black line)winter 19992000 mean vertical velocities (dashed line) errorsin vertical velocity (vertical bars) and standard deviation invertical velocity (grey lines) (a) for all stakes (b) for stakesgrouped by factor 1 in PCA (ie lower terminus) and (c) forstakes grouped by factor 2 in PCA (ie upper terminus)

Table 1 Eigenvalues and explained variance for the principalcomponent factors with an eigenvalue 41 from a PCA of2 day horizontal velocities over days 16471^21198 1999

Factor Eigenvalue Explained variance

1 1713 5042 460 1353 243 714 210 625 171 506 146 437 132 398 107 32


342

ice-marginal lake drained rapidly during event 199 (Fig 9)(although it is not known when this drainage event started)and the first major discharges from the snout were observedduring the overnight period between days 185 and 186 Thelargest geophone activity of the summer was recorded at themiddle and lower geophones on day 186 (Fig 8f and g) Afurther three smaller periods of activity were recorded ateach of these geophones before the end of event 199 whileno unusual counts were recorded at the upper geophoneduring this time (Fig 8e)

322 Event 299 days 19660^19888Event 299 saw large horizontal velocity increases at moststakes relative to both the preceding measurement period(days 19392^19660) and winter levels (Figs 4a 5 and 7e fi and j) The horizontal velocity increases were highly spa-tially variable as the increase in velocity of the stakes wasmuch higher over the upper than over the lower terminus(Fig 5b and c) On a local scale the horizontal velocity in-creases over the lower terminus were focused on the easternside with velocities only a little above winter levels on the

western side (Fig 7j) Over the upper terminus the largestvelocity increases of the year occurred on the western sidewith velocity anomalies up to 18 cm d^1 or sup1400 abovewinter levels (Fig7j)These anomalies were localized abovethe predicted areas of subglacial water flow and decreasedrapidly towards the glacier margins In addition there wasa strong rotation in horizontal velocity vectors towards theglacier margin in this area compared to both the winter andpreceding measurement periods (Figs 4a and 7e and i) Therotation was greatest for stakes closest to the glacier edgewhich in this region consists of unsupported ice walls tensof metres high

There were rapid and dramatic changes in verticalvelocity during and prior to event 299 particularly overthe upper terminus (Fig 6) There was strong vertical upliftat most stakes during the event compared to significant ver-tical lowering in the preceding measurement period (Figs 6and 7g and k) For the lower-terminus stakes verticalvelocities averaged ^14 cm d^1 in the preceding period and+10 cm d^1 during event 299 (Fig6b) For the upper-termi-nus stakes mean vertical velocities were ^11cm d^1 in the

Fig 7 Velocities in cm d 1for event 199 (days18460^18904 1999) pre-event 299 (days19392^19660 1999) event 299 (days19660^19888 1999) and event 198 (days 18270^18470 1998) respectively(a e i m) horizontal (b f j n) horizontal anomalies (ie dif-ference from winter 19992000 black lines mark reconstructed subglacial drainage pathways) (c g k o) vertical (d h l p) residualvertical Cavity opening is only likely where the assumption that wc ˆ 0 requires _zz to increase with depth relative to _zzjs

343


preceding period and Dagger35 cm d^1 during the event (Fig6c)As with horizontal velocity these patterns are highly spa-tially variable with the largest changes occurring over thewestern part of the upper terminus (Fig 7g and k) Atseveral stakes in this region the change in rate of verticalvelocity relative to the preceding period was 410 cm d^1

Cavity opening would have been possible over only alimited area in the period prior to event 299 (Fig 7h) butcould have occurred over the entire lower terminus and thewestern part of the upper terminus during the event (Fig7l) In particular the residual vertical velocity of415 cm d^1 along the axis of predicted subglacial drainageover the western upper terminus is much higher than canbeexplained by changes in vertical strain rate with depth andmakes it likely that cavity opening occurred in this area

During the 2 days prior to event 299 a storm systemreduced air temperatures to below freezing and producednew snowfall (Fig 8a) The albedo sensor records theincrease in albedo due to the new snowfall (Fig 8b) andthe UDG records net accumulation during this period(Fig 8c and d) Temperatures increased rapidly on day 197and stayed well above freezing for the remainder of theevent (Fig 8a) Surface melting resumed as the weatherwarmed and the highest sustained surface lowering ratesof the summer occurred throughout event 299 (Fig 8d)

All the water-level sensors showed a clear response tothese weather changes (Fig 9) Water levels were low at alllocations during the cool period prior to event 299 and diur-nal variability disappeared due to the reduction in meltwatersupply As air temperature and surface melt increased at theonset of event 299 water levels rose Surface lowering ratespeaked on day 1981 water levels in the ice-marginal lakeand supraglacial streams peaked between days 1987 and1988 while the water level in the proglacial stream peakedon day 1991 (Figs 8d and 9) The upper geophone recordedthe largest number of events of the summer on day198 duringevent 299 (Fig 8e) Smaller numbers of events with countswell below those during event 199 were also recorded at themiddle and lower geophones (Fig 8f and g)

33 1998 short-term velocity event

331 Event 198 days 18071^18471Only one clear event was recorded in summer 1998 (Fig10)

Fig 8 (a^d) Summer 1999 lower-weather-station records(a) air temperature (b) albedo (c) surface height measuredfrom the start of the melt season and (d) surface melt ratePositive values indicate ablation negative values indicate accu-mulation (e^g) Summer 1999 geophone records (e) uppergeophone (f) middle geophone and (g) lowergeophone Notethat only the relative magnitude of events is comparable betweengeophones not the actual number of counts

Fig 9Water-pressure records for summer 1999 (see Fig 2 forsensor locations)The measurements are standardized betweenthe highest and lowest values from each sensor over the period ofinterestThe proglacial sensor is repositioned during breaks


344

This event covers two measurement periods from days18071^18271 and 18271^18471 with the largest and mostsignificant horizontal velocity increases during the secondhalf of the event (Fig 10a) Consequently only the velocitypatterns from this second period are plotted (Fig7m^p)

The mean horizontal velocity for the 15 stakes presentduring event 198 was 50 cm d^1 during the first half of theevent and 82 cm d^1 during the second half (Fig10a) Thiscompares to a mean winter velocity for these stakes of33 cm d^1 The high standard deviation in horizontalvelocity (36 cm d^1) during event 198 implies either thatthe increases were not spatially uniform or that there islarge spatial variability in measurement errors (Fig 10a)The highest velocity increases (up to 14 cm d^1) occurredover the eastern part of the lower terminus (Fig 7n) Thisarea of high horizontal velocities is defined by at least fourstakes and the velocity increases are well above error esti-mates As in event 199 there was a rotation in the velocityvectors towards the west compared to winter patterns andthis was greatest close to the glacier snout (Figs 4a and 7m)

Vertical velocities varied little and were close to wintervalues during event 198 (Figs 7o and 10b) Of the 15 stakesmeasured during the event most had vertical velocities of 0to ^1cm d^1 The residual vertical velocity averagedsup10 cm d^1 and there were few areas where cavity openingmay have occurred (Fig7p)

As with event 199 event 198 occurred during a periodof rapid melting approximately 30 days after the start of themelt season Air temperatures were higher than at any pre-vious time in 1998 and remained well above freezingthroughout (Fig 11a) Approximately 95 cm of surface low-ering occurred in the month prior to event 198 as the sur-face snow cover melted (Fig 11c) and melt rates reachedtheir highest sustained levels of the summer during theevent (Fig11d)

Event 198 coincided with the initiation of subglacial out-flow from the terminus This occurred primarily via anartesian fountain on the lower eastern terminus on day 180(Fig 3) which reached its peak height during days 182^184before declining around day 186 The fountain reachedheights of 5 m andbrought large volumes of relatively turbidhigh-EC water to the glacier surface through an ice thicknessof sup170 m (measured by radio-echo sounding Copland andSharp 2001) This suggests a subglacial origin for the waterand indicates that basal water pressures were sup1120 of ice

Fig 10 Summer 1998 horizontal (a) and vertical (b) meanvelocity (solid black line) winter 19992000 mean velocity(dashed line) errors in velocity (vertical bars) and standarddeviation in velocity (grey lines) Six more stakes were addedto the original network of 15 from day 185 onwards

Fig 11 Same as Figure 8 but for summer 1998

345


overburden pressure over at least part of the lower terminusLongitudinal fracturing of the ice surface was observedaround the artesian fountain with most fracturing occurringwhenthe fountain first emergedThis fracturing was recordedby a small number of events at the middle geophone aroundday 18025 and by a large number of events at the lower geo-phone around day18046 (Fig11f and g) No significant activ-ity was recorded at the upper geophone at this time (Fig11e)

4 DISCUSSION

From these measurements it is evident that short-termvelocity events are associated with periods of rapidly increas-ing meltwater inputs to the subglacial drainage systemSimilar relationships have also been observed during short-term high-velocity events on temperate and predominantlywarm polythermal glaciers (eg Iken and Bindschadler1986 Jansson 1995 Mair and others 2001) and on theGreenland ice sheet (Zwally and others 2002) At JohnEvans Glacier the first velocity event of the melt seasontypically occurs about1month after the onset of melt whenthe connection between the supraglacial and subglacialdrainage systems is first established and subglacial outflowbegins (events 198 and 199) A subsequent event occurredlater in the melt season when melt resumed after a pro-longed cold spell (event 299)

41 Early-season events

Rising water inputs during early-season events are due to acombination of warm weather high rates of surface meltandthe drainage of supraglacialand ice-marginal lakes with-in and at the upstream ends of the supraglacial channel sys-tems (Figs 8a and d 9 and11a and d) Outflow of subglaciallyrouted waters at the glacier terminus typically begins within24 hours of surface waters starting to drain into the glaciervia the crevasse field at the top of the terminus This is prob-ablybecause the new water input pressurizes the existing sub-glacial reservoir beneath the lower terminus driving outletdevelopment at the snout Evidence for pressures above iceoverburden levels is providedby the formation of the artesianfountainwhile the high EC of the first waters to be released isindicative of long subglacial residence times Fracturing asso-ciated with outlet development likely contributes to the highcounts recorded at the lower geophone during events199 and198 (Figs 8g and11g)

High horizontal velocities at this time are presumablylinked to high basal water pressures (and possibly increasingsubglacial water storage) which reduce basal friction andenhance basal velocityThe PCA scores for the factor1 stakesreached their highest values during event199 (Fig 5d) as thelargest horizontal velocity changes of the year occurred overthe lower terminus (Fig 7a and b) This suggests that thefrozen snout providedthe greatest hydraulic resistance to out-flow of subglacial waters at this stage of the melt seasonTheincreasingly large westward rotation of the horizontalvelocity towards the snout compared to winter patterns (Figs4a and 7a and m) suggests that the frozen margin alsoimpeded the flow of the warm-based ice upstream Thisresulted in compression and left-lateral shearing in the regionof the warm^cold transition at the glacier bed Within thewarm-based region there was a close association betweenthe area of highest horizontal velocity anomalies during event198 and the locations of both the predicted subglacial water

flow and the artesian fountain (Fig 7n see Fig 2 for locationof artesian fountain) This suggests that the hydrological for-cing responsible for enhanced basal velocity may have beenstrongest in the vicinity of major subglacial drainage axes

42 Mid-season event

Event 299 occurred after a prolonged cold period duringwhich there was snowfall surface melt ceased (Figs 8a band d) and water levels dropped in all monitored streams(Fig 9) It is likely that subglacial water pressures dropped atthis time as meltwater inputs to the glacier interior decreasedDepending on the subglacial drainage configuration thiswould have allowed closure of basal cavities or subglacialchannels by ice andor basal sediment deformation andorcompaction of the glacier bed due to a reduction in the poros-ity of basal sediment All of these processes would result innet surface lowering as was observed in areas close to thesubglacial drainage pathways in the period prior to event 299 (Fig 7g) The highest rates of lowering during this periodoccurred in the northwest part of the upper terminus wherethe ice thicknesses are greatest (200^250m) (Copland andSharp 2001) The inferred reduction in basal water pressurewould have also increased basal drag accounting for therelatively low horizontal velocities at this time (Fig7f)

Melt resumed as the weather warmed at the start of event299 (Fig 8a and d) with a rapid rise in water levels at allmonitoring locations (Fig 9) At this time water levels in thesupraglacial streams reached their highest levels since event199 indicating that a large meltwater pulse entered the sub-glacial drainage system via the crevasse field at the top ofthe terminus Event 299 was characterized by extremelyhigh horizontal and vertical velocities in the area immedi-ately above the predicted location of the subglacial drainageaxis beneath the western part of the upper terminus (Fig7i^k) The large number of counts registered at the uppergeophone at this time (Fig 8e) together with the high PCAfactor 2 scores (Fig5d) supports the fact that event 299 wascentred over the upper terminus

The very pronounced velocity response in the northwestregion of the terminus at this time suggests that basal waterpressures may have risen to very high levels in this area whenmeltwaters were reintroduced to the bed following the coldperiodThe inferred rates of cavity opening in this area sug-gest that extensive ice^bed decoupling may have resulted(Fig7l) This behaviour was likely a response to the contrac-tion of drainage channels beneath the thick ice in this areaduring the preceding cold period which impeded meltwaterdrainage across this section of the glacier bed The strongsouthward rotation of the velocity vectors in this area (Fig7i cf Fig 7e) implies that normal glacier flow towards theeast-northeast was impeded at this time This can beexplainedby a downflow increase in basal drag arising froma downstream collapse or reduction in size of the subglacialdrainage channels

Over the lower terminus horizontal and verticalvelocities during event 299 were lower than in the westernpart of the upper terminus but they reached peak valuesabove the eastern drainage axis (Fig 7j and k) Inferredrates of basal cavity opening were also highest in this areaalthough cavity opening appears to have occurred acrossthe whole lower terminus region (Fig 7l) This is somewhatproblematic given the argument above that the largevelocity response in the upper terminus region is attribut-


346

able to closure of major drainage channels downstreamfrom this region Such closure if complete would presum-ably have prevented transfer of subglacial water to the lowerterminus There are two possible resolutions to this (i)incomplete closure allowed some basal water flow to thelower terminus but still caused backing up and high waterpressure below the upper terminus (ii) the reopening ofchannels led to a delayed water input to the lower terminuscausing the velocity event there to lag the one in the upperterminus Unfortunately the resolution of the velocitymeasurements is too coarse to show whether the velocityevents in the upper and lower terminus regions were exactlysynchronous

43 Comparisons between events

A comparison of the horizontal and vertical velocities asso-ciated with events 198 and 199 reveals some interesting dif-ferences In 1998 the largest horizontal velocity anomalieswere associated with the eastern subglacial drainage axis(Fig7n) while in 1999 they were less pronounced and morewidely distributed (Fig7b)Vertical velocities during events198 and 199 were generally low (Fig7c and o) with cavityopening restricted to a small area on the western side of thelower terminus during event 199 (Fig7d and p) The smal-ler magnitude of the event 199 horizontal velocity anoma-lies suggests that basal water pressures did not rise as high asin event 198 which is consistent with the observation thatno artesian fountain formed on the glacier surface in 1999One possible explanation is that large subglacial channelsformed during the exceptionally warm summer of 1998 didnot close completely during winter 199899 and that theyallowed more efficient drainage of the initial water inputsto the glacier bed in early summer 1999This could explainthe limited area of cavity opening during event 199although the generally small vertical velocity responseduring the early-season events makes it likely that sub-glacial drainage was mainly distributed at this time

The form of the horizontal velocity anomalies over thelower terminus during event 299 provides some evidencefor the reorganization of subglacial drainage (Fig 7j) Theanomalies suggest that the hydrological forcing for thisevent was associated with the eastern drainage axis andthus different from the more widespread forcing that seemsto have been associated with event 199 (Fig 7b) As dis-cussed above event 198 also appears to have been forcedfrom the eastern drainage axis (Fig7n) although there aresignificant differences in the ice velocity responses associatedwith the two events in the lower terminus region Horizontalvelocity anomalies were higherand more widespread in event198 (Fig 7j and n) whereas cavity opening was more wide-spread during event 299 with a peak along the easterndrainage axis (Fig 7l and p) These differences may reflectdifferences in the amounts of water entering the glacier andin the character of the subglacial drainage system betweenthe two events Event198 was associated with the first inputof surface water to the glacier bed during summer 1998 andoccurred when much of the glacier upstream from the cre-vasse field was still snow-covered Event 299 by contrastoccurred when glacier ice was exposed over much of thisarea Given that air temperatures reached similar highvalues during both events (Figs 8a and11a) melt productionin the area draining into the crevasse field was likely higherduring event 299

From these observations it appears that there is bothinterannual and intra-annual variability in the pathway bywhich meltwater which penetrates to the glacier bed drainsto the glacier terminus It appears that water draining sub-glacially from the upper terminus can at different timesconnect to either or both of the eastern and western drain-age axes beneath the lower terminus The likely location ofthe connection between the two axes is apparent from theform of the contours in Figure 7j

5 CONCLUSIONS

Measurements over 2 years demonstrate that in summer sur-face meltwater penetrates significant thicknesses (4200m) ofcold ice to reach the glacier bed and affect the horizontal andvertical surface velocities of John Evans Glacier Horizontalsurface velocities are higher in summer than in winter acrossthe entire glacier terminus High-velocity events lasting2^4 days occur during periods of rapidly increasing meltwaterinput to the glacier interior Horizontal velocities during theseevents reach up to 400 of winter velocities

Early-summer velocity events are linked to the initiationof supraglacial water inputs to the glacier bed These initialinputs include the drainage of significant volumes of waterpreviously stored in supraglacial and ice-marginal lakes Amid-summer event occurred during a period when melt ratesincreased rapidly following cold weather and new snowfallAll events are probably a result of high and rapidly risingsubglacialwater pressures andor water storageThe horizon-tal and vertical velocity anomalies that result differ signifi-cantly between events This is probably a result of seasonaland interannual changes in the routing of meltwater acrossthe bed and in the nature of the subglacial drainage systemEarly-season events are initiated within a predominantly dis-tributed subglacial drainage system while the mid-summerevent appears to have been initiated within a channelizedsubglacial drainage system that had undergone significantclosure immediately prior to the event during a cold periodwith limited meltwater drainage to the bed

The results of this study indicate a strong couplingbetween the surface velocity of a predominantly cold poly-thermal glacier the magnitude and distribution of surfacewater inputs to the englacialsubglacial drainage systemand the distribution formation and seasonal evolution ofsubglacial drainage pathways The surface velocity of theglacier is extremely sensitive to rapid increases in the deliv-ery of surface waters to the glacier bed This behaviour isqualitatively comparable to that seen on both temperateand warm polythermal glaciers (Iken and Bindschadler1986 Jansson 1995 Kavanaugh and Clarke 2001 Mairand others 2001) and on the Greenland ice sheet (Zwallyand others 2002) A number of characteristics that may betypical of predominantly cold polythermal glaciers influ-ence the details of the behaviour observed These includethe thermal barrier to the flow of ice and basal water at thesnout and margins of the glacier the sudden onset of sub-glacial drainage in the early summer and the importanceof water inputs from ice-marginal and supraglacial lakes

ACKNOWLEDGEMENTS

Funding was provided by the Natural Sciences and Engin-eering Research Council of Canada the UK Natural

347


Environment Research Council ARCICE programme theGeological Society of America the University of AlbertaIzaak Walton Killam Memorial Scholarship and theCanadian Circumpolar Institute of the University of AlbertaLogistic support was provided by the Polar Continental ShelfProject Natural Resources Canada (contribution No023-01)We thank the Nunavut Research Institute and the commu-nities of Grise Fiord and Resolute Bay for permission to workatJohn Evans Glacier andW Davis D Glowacki A ArendtT Wohlleben R Bingham and J Davis for help in the fieldC S Hvidberg (Scientific Editor) J Kavanaugh and PJans-son provided useful comments on the manuscript

REFERENCES

Andreasen J-O 1985 Seasonal surface-velocity variations on a sub-polarglacier inWest Greenland J Glaciol 31(109) 319^323

Balise M J and C F Raymond1985Transfer of basal sliding variations tothe surface of a linearly viscous glacier J Glaciol 31(109) 308^318

Bindschadler R 1983 The importance of pressurized subglacial water inseparation and sliding at the glacier bed J Glaciol 29(101) 3^19

Bingham R G PW Nienowand M J Sharp 2003Intra-annual and intra-seasonal flow dynamics of a High Arctic polythermal valley glacier AnnGlaciol 37181^188

Blatter H 1987 On the thermal regime of an Arctic valley glacier a studyof White Glacier Axel Heiberg Island NWT Canada J Glaciol33(114) 200^211

Blatter H and K Hutter1991 Polythermal conditions in Arctic glaciers JGlaciol 37(126) 261^269

Copland L and M Sharp2000 Radio-echo sounding determinationof poly-thermal glacierhydrology In Noon D G F Stickleyand D Longstaff edsGPR 2000 Eighth International Conference on Ground Penetrating Radar 23^26May 2000 Gold Coast Australia Bellingham WA International Society ofPhoto-optical Instrumentation Engineers 59^64 (SPIE Proceedings4084)

Copland L and M Sharp 2001 Mapping thermal and hydrological condi-tions beneath a polythermal glacier with radio-echo sounding J Glaciol47(157) 232^242

Harbor J M Sharp L Copland B Hubbard P Nienow and D Mair1997The influence of subglacial drainage conditions on the velocity dis-tribution within a glacier cross section Geology 25(8)739^742

Hodgkins R1997 Glacier hydrology in Svalbard NorwegianHigh ArcticQuat Sci Rev 16(9) 957^973

HookeR LeB P Calla P Holmlund M Nilssonand A Stroeven1989A3 year record of seasonal variations in surface velocity Storglacialaquo renSweden J Glaciol 35(120) 235^247

Iken A1974Velocity fluctuationsofanArctic valleyglacier a studyof theWhite GlacierAxel Heiberg Island CanadianArctic Archipelago Montreal Que McGill Uni-versity (Axel Heiberg Island Research Reports Glaciology 5)

Iken A1981The effect of the subglacialwater pressure on the sliding velocityof a glacier in an idealized numerical model J Glaciol 27(97)407^421

Iken A and RA Bindschadler1986 Combined measurements of subglacialwater pressure and surface velocity of Findelengletscher Switzerlandconclusions about drainage system and sliding mechanism J Glaciol32(110)101^119

Iken A H Rolaquo thlisberger A Flotron and W Haeberli 1983The uplift ofUnteraargletscher at the beginning of the melt season ouml a consequenceof water storage at the bed J Glaciol 29(101) 28^47

Jansson P1995Water pressure and basal sliding on Storglacialaquo ren northernSweden J Glaciol 41(138) 232^240

Jansson P 1996 Dynamics and hydrology of a small polythermal valleyglacier Geogr Ann 78A(2^3)171^180

Johnston R J1978Multivariate statistical analysis ingeography Harlow LongmanKamb B 1987 Glacier surge mechanism based on linked cavity configura-

tion of the basal water conduit system J Geophys Res 92(B9)9083^9100Kavanaugh J L and G K C Clarke2001 Abrupt glacier motion and reor-

ganization of basal shear stress following the establishment of a con-nected drainage system J Glaciol 47(158) 472^480

Mair D P Nienow IWillis and M Sharp 2001 Spatial patterns of glaciermotion during a high-velocity event Haut Glacier drsquoArolla SwitzerlandJ Glaciol 47(156) 9^20

Mair DW F M J Sharp and I CWillis 2002 Evidence for basal cavityopening from analysis of surface uplift during a high-velocity eventHaut Glacier drsquoArolla Switzerland J Glaciol 48(161) 208^216

Manabe S R J Stouffer M J Spelman and K Bryan 1991 Transientresponse of a coupled ocean^atmosphere model to gradual changes ofatmospheric CO2 Part I Annual mean response J Climate 4(8)785^818

Mulaquo ller F and A Iken1973Velocity fluctuations and water regime of Arcticvalley glaciers International Association of Scientific Hydrology Publication 95(Symposium at Cambridge1969ouml Hydrology of Glaciers)165^182

PatersonW S B1994The physics of glaciersThird edition Oxford etc ElsevierRabus BT and K A Echelmeyer1997The flow of a polythermal glacier

McCall Glacier Alaska USA J Glaciol 43(145) 522^536Raymond C F R J Benedict W D Harrison K A Echelmeyer and M

Sturm1995 Hydrologicaldischarges and motion of Fels and Black RapidsGlaciers Alaska USA implications for the structure of their drainagesystems J Glaciol 41(138) 290^304

Sandwell DT1987 Biharmonic spline interpolationof GEOS-3 and SEASATaltimeter data Geophys Res Lett14(2)139^142

Shreve R L 1972 Movement of water in glaciers J Glaciol 11(62) 205^214

Skidmore M L and M J Sharp 1999 Drainage system behaviour of aHigh-Arctic polythermal glacier Ann Glaciol 28 209^215

Zwally H J W Abdalati T Herring K Larson J Saba and K Steffen2002 Surface melt-induced acceleration of Greenland ice-sheet flowScience 297(5579) 218^222

MS received 21 August 2001 and accepted in revised form 31 March 2003


348

predominantly cold polythermal valley glacier on the eastcoast of Ellesmere Island Nunavut Canada (79sup340rsquo N74sup330rsquo W Fig 1) The glacier ranges in elevation from 100 to1500 m asl with the long-term equilibrium line at sup1750^850m asl Ice depths reach a maximum of 400m in theupper ablation area and average 100^250m in the lower

ablation area where the velocity measurements were made(Copland and Sharp 2001) For the period1997^99 the meanannual air temperature at 820m asl was ^152sup3C

High bed reflection powers in radio-echo soundingrecords indicate warm ice at the bed throughout most ofthe lower ablation zone except along the glacier marginsand where the ice is thin (Copland and Sharp 2001) A con-tinuous internal reflecting horizon over the centre of thelower terminus suggests that the warm basal ice reaches anaverage thickness of 20 m there In the accumulation andupper ablation areas low bed reflection powers and 15 mborehole temperatures of ^95 to ^151sup3C suggest that theice is cold throughout

The melt season typically occurs between earlyJune andearly August At the start of the melt season meltwatereither ponds on the glacier surface or is routed directly tothe ice margins and is unable to access the glacier interiorOnce englacial drainage of supraglacially derived melt-waters is initiated however approximately 25 of theglacier surface area drains into moulins in a crevasse fieldat the top of the terminus (Fig 2) Later in the melt seasonmeltwater from up to 40 of the surface area of the glacierdrains to the glacier bed via these moulins and a second cre-vasse field sup16 km further upstream Dye tracer experimentsconfirm the link between these water input locations and amajor drainage portal at the terminus (Bingham andothers 2003) Large increases in the suspended-sedimentcontent and electrical conductivity (EC) of the water as itpasses through the glacier suggest that it is routed sub-glacially (Skidmore and Sharp1999)

Inthe early summer the meltwater is trappedat the glacier

Fig 1 Bed topography (m asl) and location ofJohn EvansGlacier

Fig 2 Landsat 7 image of John Evans Glacier (path 049 row 002 10 July 1999) showing the location of the velocity stakes(black dots) geophones (white dots LG ˆ lower geophone MG ˆ middle geophone UG ˆ upper geophone) surveying basestations (white squares) stream gauging stations (triangles) and artesian fountain observed in 1998 ( ) Crevasse field indi-cates location where most supraglacial meltwater reaches the glacier bed Lower weather station is located next to the middlegeophone Black box indicates horizontal extent of Figures 4 and 7


338


2 MEASUREMENTS

21 Surface velocity













339




23 Geophones



24Weather stations














340




3 RESULTS






341













1 1713 5042 460 1353 243 714 210 625 171 506 146 437 132 398 107 32


342






343











344








345



4 DISCUSSION






42 Mid-season event






346






5 CONCLUSIONS




ACKNOWLEDGEMENTS


347



REFERENCES


































348


2 MEASUREMENTS

21 Surface velocity













339




23 Geophones



24Weather stations














340




3 RESULTS






341













1 1713 5042 460 1353 243 714 210 625 171 506 146 437 132 398 107 32


342






343











344








345



4 DISCUSSION






42 Mid-season event






346






5 CONCLUSIONS




ACKNOWLEDGEMENTS


347



REFERENCES


































348



23 Geophones



24Weather stations














340




3 RESULTS






341













1 1713 5042 460 1353 243 714 210 625 171 506 146 437 132 398 107 32


342






343











344








345



4 DISCUSSION






42 Mid-season event






346






5 CONCLUSIONS




ACKNOWLEDGEMENTS


347



REFERENCES


































348




3 RESULTS






341













1 1713 5042 460 1353 243 714 210 625 171 506 146 437 132 398 107 32


342






343











344








345



4 DISCUSSION






42 Mid-season event






346






5 CONCLUSIONS




ACKNOWLEDGEMENTS


347



REFERENCES


































348












1 1713 5042 460 1353 243 714 210 625 171 506 146 437 132 398 107 32


342






343











344








345



4 DISCUSSION






42 Mid-season event






346






5 CONCLUSIONS




ACKNOWLEDGEMENTS


347



REFERENCES


































348






343











344








345



4 DISCUSSION






42 Mid-season event






346






5 CONCLUSIONS




ACKNOWLEDGEMENTS


347



REFERENCES


































348










344








345



4 DISCUSSION






42 Mid-season event






346






5 CONCLUSIONS




ACKNOWLEDGEMENTS


347



REFERENCES


































348








345



4 DISCUSSION






42 Mid-season event






346






5 CONCLUSIONS




ACKNOWLEDGEMENTS


347



REFERENCES


































348


4 DISCUSSION






42 Mid-season event






346






5 CONCLUSIONS




ACKNOWLEDGEMENTS


347



REFERENCES


































348






5 CONCLUSIONS




ACKNOWLEDGEMENTS


347



REFERENCES


































348


REFERENCES


































348

linksbetween short-termvelocity variations and the subglacial … · 2012. 1. 11. · linksbetween...

Documents