1. Introduction

The thermal regime of the White Glacier (Fig. 1), a medium-sized valley glacier on Axel Heiberg Island, lat. 80° N., in the Canadian Arctic Archipelago, has been studied since 1959 as part of a continuous and still ongoing glaciological programme organized by McGill University. This simple-shaped glacier of 16 km length with a smooth longitudinal profile descending from an elevation of 1 800 m to 60 m above sea-level may be typical for many Arctic glaciers.

Some early results from this study pertaining mainly to the seasonal temperature variations within the first 10 m below the glacier surface in the accumulation and ablation areas and the zone of superimposed ice near the equilibrium line already indicated a much more complicated thermal regime than anticipated for the White Glacier (Müller, Reference Müller and Muller1961, 1963[a],[b]). Still too often the assumption is made that the roughly sinusoidal seasonal changes of the air temperature would penetrate the top layers of a glacier to a depth of some 10 m, and that (he temperature read at this depth could be equated with the mean annual air temperature. In the ease of the White Glacier no such agreement was found. The present paper presents further detail of these “anomalies”. Similarly “anomalous” temperatures within and below the seasonally changing layers of Arctic glaciers have been reported from north-western Greenland (Reference SchyttSchytt, 1955; Reference BensonBenson, 1962), Spitsbergen (Schytt, Reference Schytt1964,1969), Zemlya Frantsa-Iosifa (Reference KrenkeKrenke, 1963), Novaya Zemlya (Reference KhmelevskoyKhmelevskoy, 1968), Alaska (Reference MillerMiller, [1956]; Reference Orvig and MasonOrvig and Mason, 1963; Reference Trabant, Trabant, Harrison, Benson, Weller and BowlingTrabant and others, 1975).

Following on the preliminary findings it was decided to extend the study of the thermal regime of the White Glacier to include a detailed assessment of the temperature variations over many years for as many parts of the glacier as possible and finally to add measurements of temperatures throughout its depth. The White Glacier is particularly suitable for such a detailed investigation as much information—up to 15 years of data—is available on other aspects governing the behaviour of this glacier: geometry, in part based on seismic data (Reference RedpathRedpath, 1965); surface velocities and their fluctuations and the associated hydrologic regime (Reference MaagMaag, 1969; Reference Müller and IkenMüller and Iken, 1973; Reference IkenIken, 1974); accumulation and ablation patterns (Reference MüllerMüller, 1962; Reference AdamsAdams, 1966); and micrometeorological and climatic data including records from three automatic, weather stations.

2. Instrumentation And Observation Sites

The temperatures were recorded with strings of thermistors frozen into the ice and read with a Wheatstone bridge (Tettex type). Platinum resistance elements were used for independent checking at selected points. The thermistors were calibrated to an accuracy of ±0. 1 deg and better; taking into account the instrument error of the bridge an absolute accuracy of ±0.2 deg for all measured values can be claimed. Recalibration of the retrieved thermistors and comparison of the data with those obtained from the more stable platinum resistance elements (operating accuracy better than ±0.1 deg) showed that no significant shift of the calibration values had taken place during the many years of field use. Though the tempera- lure-measuring devices were individually connected to the surface and placed in specially designed casings to protect them against high pressure resulting from the freezing-in process and the differential movement of the ice masses over the years, none of them remained functional for longer than eight years.

The shallow holes were hand-drilled using a 3 inch (7.6 cm) Teflon-coated SIPRE coring auger. The three deep holes (to a maximum depth of 280 m) were drilled during the summer of 1974 with a most efficient new type of open system, hot-water drill, adapted from a design by Kasser, Röthlisberger and Iken. In this equipment, the water (local surface melt water), heated by two propane burners and pumped with a pressure of some 20 atmospheres through a rubber tube insulated with PVC, is released through a 4 nun jet at the end of the 2 m long drill head, which is stabilized with about 3 kg of lead grains. An advance rate of 25 m to 50 m per hour was achieved with the drill in the cold ice of the White Glacier.

The distribution of the temperature measurement sites on and in front of the White Glacier is shown in Figure 1. So far only one site—Beaver Camp—is permanently located in the accumulation, area. Most sites are near or just below the mean equilibrium line. The annual equilibrium line changes its position widely from year to year; in 1962 it. was found at an elevation of 1 300 m while in 1964 it fell to only 400 m above sea level. The three temperature measurement sites in the Anniversary profile and the cluster of sites on the tongue are clearly in the ablation area.

3. The Data

The main purpose of the present study was to assess differences in temperatures at the 10 m depth over the entire White Glacier and to check if the findings varied over the years. To permit comparison between the different parts of the glacier and over many years it was necessary to take the temperature readings more or less simultaneously and always at the same time of the year, e.g. at the end of the accumulation season in late August. Furthermore, the shift of the zero level of the profile due to accumulation and ablation had to be taken into account when plotting the data for these comparisons. Figures 2, 3, 4 and 5 present the near- surfacc temperature comparison over a number of years from the highest to the lowest site on the glacier. The differences from year to year are largest at Reaver Camp (accumulation area), are already considerably less at Moraine Camp (site C) and are small at the two ablation area sites, Anniversary profile and Lower Ice.

Fig. 1. Location map, showing all englacinal temperature measurement sites on the While Glacier to date and their names and symbols. The positions of the three automatic weather stations on and near the glacier and the site of a temperature profile in the permafrost in front of the glacier are also indicated

Fig. 2. Late-summer near-surface temperatures at B eaver Camp I 424 m

Figure 4 also shows the comparison of shallow-depth temperatures from sites on the eastern (A3), central (A6) and western (A9) portion of the Anniversary cross-profile. The ice near the centre line of the glacier is coldest, while the ice along the steep basaltic rock wall on the east side is almost 4 deg warmer. Equally large temperature differences were observed across the Moraine profile (Table I), with the coldest ice being found near the eastern margin of the glacier (M2 and M4) and the warmest on the western side (M1o); in August 1963 the difference amounted to 4.5 deg. During the budget year 1962/63—as during many others— the western portion of the “Moraine” profile remained firn covered, resulting from a relatively thick winter snow pack and little ablation during the summer, whereas in the eastern half of the profile bare ice was already exposed in spring and possibly during much of the winter. Both firn line and slush limit are usually situated higher on the eastern side of the glacier than on the western parts.

During the summer 1974 additional shallow-depth temperatures were obtained from three deep-drilling sites: Site I at a short distance above the mean equilibrium line, and sites II and III respectively 2 km and 4 km below it (Fig. 1). The data collected from these sites (Fig. 6 and Table I) in no way simplified the pattern of temperature distribution at the 10 m depth. Site II, expected to show low temperatures, proved to contain warmer ice than either of the other two (this situation existed not only near the surface but extended to a depth of some 160 m). Sites I III were thought to be close enough to a central flow line to provide information on the longitudinal development of the temperature pattern. It now seems that the slight deviation of site II to the west brought it into the realm of the warmer ice descending on the western side of the “Moraine” profile, while the displacement of site III by an amount of 200 m to the east of the centre line pul its profile into the stream of colder ice corning from the M4 area. The surrounding of site II frequently forms an accumulation outlier, whereas site 111 lies in an area of strong ablation and reduced winter snow due to wind abrasion.

Fig. 3. Late-summer near-surface temperatures at Moraine Camp C 832 m

To summarize all the 10 m depth temperature data (including one value from a site in the permafrost area in front of the glacier), all values are plotted against altitude and put in juxtaposition with the mean annual air temperature over the White Glacier area in Figure 7. The air-temperature profile was calculated from the 1951-60 surface and aerological (950 and 850 mbar level) records of the weather stations at Eureka and Isachsen used in conjunction with the data collected by the three automatic weather stations on Axel Heiberg Island at the Base Camp, on the glacier near the Moraine profile and near Beaver Camp. A statistical relationship of the form T = a+b T in was applied, where T is the temperature over the glacier surface computed front the large-scale atmospheric temperatures T(in obtained by interpolation between Eureka and Isachsen records. Thirty months of automatic weather-station data from the White Glacier area was available for this calculation. A strong mean annual inversion to an elevation of about 200 m and from there an almost isothermal air temperature profile over the glacier to an elevation of at least 1 500 rn was found.

Fig. 4. Comparison of late-summer near-surface temperatures acroofS the glacier at Anniversary profile 370 tn

Discussion Of Results

4.1. The 10 m temperatures as upper boundary input

Figure 7 clearly shows that there is no simple relationship between the mean annual air temperature and the near-surface, i.e. the 10 m, temperature of the White Glacier. The magnitude of difference between the two quantities is given by the horizontal distance between the 10 m temperature points and the air temperature curve on the left-hand side of the figure. Deviations of up to 10 deg are found. The large spread of these deviations though made more obvious by plotting both regional and time variations on one graph—indicates a complexity of causes. Attempts to connect related to m temperature points did not yield satisfactory results. Instead two strongly different situations are shown schematically. The solid line starts with the Beaver Camp site being pari of the lower percolation zone, while the dashed line commences at a Beaver Camp which belongs to the upper percolation zone. The solid line thus depicts a situation in which the equilibrium line on the White Glacier would be situated for some years at an elevation of 800 to 1 000 m a.s.l. The dashed line shows the thermal “surface” situation of a White Glacier with the equilibrium line for a number of years at an elevation of 400 m to 600 m a.s.l., i.e. as it existed in 1964 and 1965.

Fig. 5. Late-summer near-surface temperatures at Lower Ice 205 m

The above interpretation of the data is elucidated by Figure 8 which shows qualitatively and schematically the thermal influence of the different accumulation zones. A description of the characteristics of the various zones is given in Reference MüllerMüller (1962). The input of energy into the glacier through percolating melt water and the release of latent heat during refreezing at depth is strongest in the lower percolation zone. On the White Glacier this process has been found to be active to a depth of up to 6 m below the surface, which can represent as much as eight years of accumulation. In one case, in 1966, 25 cm of wet firn were found at a depth of 7.2 m below almost 3 m of solid ice in a hole made in the lower percolation zone with a SIPRE coring auger. The amount of heat input by the percolation process decreases drastically in the slush zone. In the superimposed ice zone and the zone of transitional superimposed ice, situated immediately below the equilibrium line, some heat is provided to the glacier by the refreezing process. Only in the dry snow zone and the lower parts of the ablation area does the heat input through glacioiogical process become negligible.

Table I. 10 m Temperatures in the white glacier,, axel heiberg island, n.w.t.

If as a result of climatic changes the different accumulation zones are shifted to higher or lower elevations, the resulting effect for a given area on the glacier can be contrary to what would be expected. Figure 8 demonstrates that in the case of a negative climatic change the slush zone may become a lower percolation zone and thus receive more energy input, while a positive climatic change may provide less heat to the area of the original lower percolation zone.

Fig. 6. Englacial temperatures, White Glacier, 25 August ' 974

4.2. Influences on the 10 m temperature pattern

The influence of the differential energy input at the surface on the thermal regime of the glacier is qualitatively depicted in Figure 9. The large irregularity in the near-surface isotherms in the accumulation area is dissipated downwards through conduction and is buried to even greater depth by the downward component of the ice movement, while at the same time being advanced towards the glacier terminus. In the ablation area the upward component of the ice movement shifts the isotherms towards the glacier surface.

A more realistic picture of the temperature distribution in the White Glacier is drawn in Figure 10. Irregularities in the frictional and geothermal heating are not considered in (his presentation; the emphasis is placed on the distribution of the near-surface temperatures. As borne out by observation, the coldest temperatures are shown near the mean equilibrium line, while the warmest are to be found near the lower end of the lower percolation zone. Consequently it is suggested that the large heating in this area may result in the onset of basal sliding at approximately this point or somewhat down-glacier of it.

Fig. 7. 10 m depth temperatures, White Glacier. The left-hand line depicts the mean annual air temperatures; the solid curve approximates the upper boundary temperature condition for A White Glacier with a mean equilibrium line situated at an elevation of about goo m a.s.l.; the dashed curve represents the glacier temperature situation associated with an equilibrium line elevation of about poo m a.s.l. (N,B. These lines do not attempt to approximate the array of measured temperature points)

4.3. Deep temperatures and position of the pressure melting point

None of the deep temperature profiles obtained in the White Glacier so far have reached bedrock. However, the lower portions of these profiles (Fig. 6) provide sufficient evidence to support the above-made suggestion that the White Glacier is able to slide over large portions of its bed. Extension of the almost linear lower parts of the profiles from sites II, III and L I meet the pressure melting point temperature at or above the approximately known glacier depth. At site II where the deepest thermistor at 268 m showed a temperature of 3-5°C the extrapolated profile reaches the pressure melting point of —o.2°C between 300 m and 310 m depth. At site III, where the glacier may be slightly thicker, the profile also points to a somewhat lower position of the pressure melting point. The temperature gradient in the lowermost part of these profiles amounting to about —0.08 deg/m is considerably larger than the geothermal heat flux would indicate.

Fig. 8. Schematic presentation of thermal influence of the different accumulation zones (6: zone of transitional superimposed ice; 7: ablation zone s.st.)

Fig. 9. Conceptual model showing thermal influence of the different accumulation and ablation zones (2—7) on a valley glacier, with and without glacier movement

Fig. 10. Qualitative pattem of temperature distribution in an Arctic valley glacier (1- 5 : accumulation zones )

5. Acknowledgements

The author is most indebted to many members of the Axel Heiberg Island Expedition for helpful assistance in the field and in the office. Heinz Blatter carried out the laborious task of deep drilling at sites I, II .and III. Roger Braithwaite calculated the mean annual air temperature profile shown on Figure 7. Atsumu Ohmura provided many valuable contributions to the discussion. Logistic support was generously made available by the Polar Continental Shelf Project of the Canadian Department of Energy, Mines and Resources. The project received financial support from the National Research Council of Canada.

Fig. 11. Comparison of measured and calculated (1) and (2) mean temperatures 8 m to 3o m depth below ice surface at Lower Ice 205 m

Discussion

M. M. Miller: The seemingly anomalous thermal profile in the accumulation zone which you explain by percolation perhaps has another explanation. For example, in Alaskan glaciers we have found a possibly comparable situation—i.e. colder ice near the equilibrium line (névé-line in these more temperate cases), at least in early winter profiles. We have considered this to relate to heavy snowfall in the autumn, at which time accumulation is maximum at higher elevation, yet often with bare ice remaining exposed below the névé-line. This new snow-pack (with low diffusivity) insulates the higher reaches against penetration of the early winter cold wave. As further snowfall occurs, this ultimately lowers the point of origin of the annual temperature wave. In contrast at the névé-line, the much larger diffusivity of bare ice and the presence of exposed crevasses permit deeper penetration of the early season cold. In other words, the blanketing snow at higher levels tends to increase the amplitude and lag effects of the cold wave and effectively impedes its penetration. I wonder whether this might be a causal factor in the development of your out-of-phase thermal profiles, or at least a supplement to the percolation mechanism which you describe?

F. MÜller: The insulating effect of the early winter snow cover may also play a considerable role in the near-surface temperature regime of the White Glacier; it would supplement and strengthen the percolation effect. However, it seems to be of lesser importance on the White Glacier for the following reasons: (i) The early winter snow cover seldom amounts to more than one metre and is often less at Beaver Camp (accumulation area); and (2) The early spring temperature profiles for the first five or six metres below the surface show temperatures which are little different from those near the equilibrium line; the drastic warming in the accumulation area occurs during the warm part of the summer, after the o°C line has reached this altitude.

W. D. Harrison: HOW did you know that the lower part of the glacier is sliding?

MÜller: Reference Müller and IkenMüller and Iken (1973) have shown that there are large variations in the velocity during the summer that can only be explained by sliding.

Harrison: IS there a finite layer of isothermal ice at the base?

MÜller: We do not know yet because we do not know whether any of our deep holes have penetrated to bedrock.

R. Leb, Hooke: TO what do you attribute the zero lapse rate?

M>Üller: The mean annual zero lapse rate results from the strong inversion during the winter and spring months, as revealed by the radio-sonde data of Eureka and Isachsen and the automatic weather station data from the White Glacier area.