Crystallographic c-axis orientation

c-axes have been plotted and contoured on the lower hemisphere of a Schmidt net. The results are summarized in Table I. Only one sample from the core, at 76 m depth, showed a preferred orientation with a concentration of c-axes greater than 7.5 per cent per per cent of area. Remaining samples showed weak maximum concentrations, ranging between 5.0 and 7.5 per cent per per cent of area. The pole of concentration is centred close to the vertical. The ice sample taken within 2 m of the ice-rock contact showed a maximum of only 5.0 per cent per 1 per cent of area.

Crystal and bubble fabric

Ice crystals in the core are generally equigranular (Fig. 2a) and there is no evidence of interlocking patterns which are common in strained glacier ice. There is a wide variation of crystal size from layer to layer so that layering in the ice is often very marked. Crystal sizes (expressed here as cross-sectional area) range from 0.01 to 21.0 cm².

Fig. 2. Ice samples from the Meighen Ice Cap core.

The lines across the photographs are rubber bands used to hold the sample in place. All the samples are orientated vertically and shown against a centimetre grid.

• a. Basal ice from 121.1–121.2 m.

• b. Moderately coarse ice from 78.7 m. The upper 4–5 cm of the section consists of pond ice.

• c. Ice layering at 40.3 in.

• d. Bubbly layer, i.e. old firn, at 40.4 m.

The mean crystal size of each layer has been determined by counting the number of crystals falling within a certain area of a centimetre grid. Crystals falling across the boundary lines of the measuring area were included only if more than half their area was within the counting grid area. The mean crystal size ( x ) for the samples from each to m depth interval has been calculated according to:

where x ₁ is the mean crystal size for each layer in cm’ and n is the number of layers. The results are plotted in Figure 3.

Fig. 3. Variation of ice-crystal size with depth.

A variance analysis of the to m means of crystal size has also been made. The trend of increasing crystal size between the surface and 55 m, and the decrease of crystal size towards the base of the core are both statistically significant (at the 1 per cent level). However, the variations between 55 and 115 m are not significant and the discussion will consider this interval as consisting of generally coarse ice.

The air-bubble fabric varies in detail from layer to layer. There are many elongate bubbles with the long axis of the bubble normal to the surface, but there are always other spherical bubbles in the same section. No evidence is found in the sampled part of the core of bubble stretching, i.e. where all the bubbles in a single section are elongated with their long axes parallel to each other. In general, coarse ice is very clear, whereas fine ice is bubbly, but the study of the relationship between bubble and crystal fabric has not been pursued.

Stratification is evident in the ice core but it is difficult to interpret. Wherever layers are distinct on the photographs (e.g. Fig. 2c) their thickness has been measured. The mode of layer thickness so observed is 9–12 cm and the mean thickness is 12.6 cm. Present-day total accumulation of snow would form an ice layer 15–18 cm thick (Reference PatersonPaterson, unpublished). As some run-off usually takes place in a superimposed-ice zone, the mean measured layer thickness of 12.6 cm of ice may be accepted as a reasonable estimate of mean annual accumulation in the past. However, as there must also have been some years when the balance was negative, this accumulation rate should be considered a maximum.

Firn layers

The position of dirt and firn layers in the core was recorded by Paterson (personal communication) on extraction of the core. Firn layers occur to a depth of 22 m. Below this no permeable layers occur, but there are occasional very bubbly layers which probably represent old firn (Fig. 2d). Reference BensonBenson (1962) has estimated that the firn–ice transition zone is more than 77 m deep in dry snow, 60–75 m deep in the percolation facies and 0–35 m deep in the soaked facies. Some support of buried firn layers may be effected by ice glands crossing that firn layer, but firn is unlikely to occur below 35 m in a dominantly superimposed-ice zone where the mean density, and hence load, is high. 35 m is therefore an absolute maximum depth for firn to occur on the Meighen Ice Cap. As the balance is generally negative at the core site now (Reference ArnoldArnold, 1965; personal communication from W. S. B. Paterson), it implies that not more than 13 m of ice have ablated from the surface since the deepest firn layer, now at a depth of 22 m, was at a depth of 35 m.

Dirt layers

At many levels in the core there are several 1–5 mm pockets of dirt giving the appearance of dirt layers (Fig. 4). The number of dirt layers occurring in each ro m is shown in Figure 5. The dirt particles are generally more than 0.1 mm in diameter and have been blown on to the ice cap in summer from the surrounding ice-free land. The variation of dirt content with depth is largely a reflection of the variation of the area of ice-free land on the island.

Fig. 4. Close-up of dirt layer. centimetre grid, 100.6 m depth. The arrow paints. inn the vertical direction.

Fig. 5. Distribution of dirt layers with depth.

Crystal size and depth

Firn- and ice-fabric studies above the firn line in Greenland (Reference LangwayLangway, 1962) and Antarctica (Reference GowGow, 1963) have shown that there is a gradual increase of crystal size with depth. In the Meighen Ice Cap core, an increase in crystal size from 0.5 to 4.0 cm² occurs between the surface and 55 m (Fig. 3), but there is no further significant increase below this. For an explanation of the variation of crystal size with depth in this core one must consider the structure and process of formation of superimposed ice.

In a study of the formation of superimposed ice on the Devon Island ice Cap, Koerner (unpublished) found that the crystal size of superimposed ice forming in summer is related, first, to the degree of saturation of the snow cover, and secondly, to the temperature of the ice under the snow. When the snow pack is mainly dry and the underlying ice is very cold, some melt water from the surface percolates down and refreczes rapidly on the ice to form fine-grained superimposed ice. As the snow cover becomes increasingly saturated, melt water stands on the, by now, warmer ice and refreezes slowly to form much coarser ice. In cold summers, saturation of the snow proceeds slowly and mainly fine ice forms and some firn may be left at the end of summer. In warmer summers saturation of the snow cover is complete and water may stand on the surface to form very coarse ice when it refreezes. Dating of layers is not possible as two or more layers of different textures may form in one summer. However, the gross variations of crystal size with depth may be explained in terms of varying summer temperatures. Using this interpretation, the coarse ice between 55 and 115 m depth was deposited in warmer summers than the ice above or below it.

There are two objections to this interpretation. First, the extreme coarseness of many of the ice layers between 55 and 115 m depth cannot be simply related to relatively high summer temperatures. Ice of similar coarseness has not been found in any of a series of eight 5-m cores from the superimposed-ice zone of the Devon Island Ice Cap, and it should form at altitudes just above the equilibrium line if the crystal size of superimposed ice is solely related to the summer temperature. Secondly, coarse ice layers were not found near the top of the core. During the transition from cold summers to the present condition of negative balance at the core site, there should have been a period when coarse ice accumulated if the crystal size—summer temperature relationship is a close one. It is possible that such coarse ice Iayers have ablated during the present period of negative balance, in which case this objection no longer applies.

However, accumulation and ablation conditions on an ice cap may be altered by a change of surface slope without any climatic change. Gentle slopes on an ice cap are usually associated with poor drainage so that slush fields and ponds may develop during the summer melt period. Refreezing at the end of the melt season adds a layer of coarse ice to the surface as slush and ponded water refreeze relatively slowly. Conditions like these may have existed on the Meighen Ice Cap while it built up from a thickness of about 10 to 70 m. By the time the ice cap was 70 m thick at the core site, a slope had developed to allow drainage of melt water and thereby prohibit the development of slush fields and ponds during the summer. During both conditions of steep and gentle slopes, fine ice will form early in the melt season at the base of the snow cover (Reference SchyttSchytt, 1955). Whereas poor drainage retards ablation and often allows large thicknesses of coarse ice to form in relatively short summers, steep slopes, by promoting more efficient drainage of melt water, accelerate ablation. In this case only the fine ice formed in early summer may remain at the end of the melt season.

A complete explanation of the change of crystal size with depth must include both the topographic and climatic arguments.

Relic ablation surface

There is a significant hiatus in the distribution of dirt layers with depth in the 40–50 m interval (Fig. 5). This is at a similar level to the hiatus in the change of crystal size with depth (Fig. 3). It has already been suggested that the distribution of dirt layers in the core reflects variations of the area of ice-covered land. The hiatus may, therefore, represent a relic ablation surface formed after a period of negative balance when the ice cap not only thinned at the core site but also decreased in area.

The ablation surface is not marked by a very dirty layer as observations on the Devon Island Ice Cap by the author have shown that dirt accumulating at the surface during a period of negative balance may be washed off periodically in warm overcast weather. A decrease in both the area and thickness of the ice cap, before the uppermost 40–50 m of ice accumulated, would account for the development of a slope at the core site, which is now only about 2 km from the ice-cap edge (Fig. 1). It would also account for the sudden increase in the number of dirt layers above the 40–50 m depth interval as there was a greater area of ice-free land surrounding the ice cap once positive balance conditions restarted. Thus by the time the balance changed from negative to positive again at the core site, a slope had developed thereby altering the accumulation conditions so that superimposed ice, finer-grained than before, formed.

The relic ablation surface may be approximately dated if the previously determined maximum accumulation rate of 12.6 cm of ice/year is used. 40–50 m of ice now lie over the ablation surface and a further maximum of 13 m of ice have ablated in the present warm period. This much ice would have taken at least 400–500 years to accumulate to which the period of negative balance now existing at the core site must be added. From the present depth of the deepest fern layer it has been postulated that a maximum of 13 m of ice have ablated from the surface. The mean specific balance at the core site between 1960 and 1966 was −15 cm of ice/year (Reference ArnoldArnold, 1965; personal communication from W. S. B. Paterson). At this rate, negative-balance conditions have continued at the core site for at most 80 years. Paterson (unpublished), in a study of the temperature profile in the core hole, concluded that the mean annual surface temperature on the Meighen Ice Cap began increasing about 85 years ago, attained a maximum about 25 years ago, and has been decreasing since. This is in general agreement with the conclusion of the present study which makes the ablation surface at least 500–600 years old.

On the northern Ellesmere Ice Shelf, Reference MarshallMarshall (1955) found a disconformity at a depth of 25 m, above which the ice consisted of granular iced firn. As Reference MarshallMarshall (1955) estimated a lower accumulation rate there (7.6 cm of ice/year) and probably more than t o m of ice have ablated this century, the disconformity may be contemporaneous with the relic ablation surface on the Meighen Ice Cap. However, Reference CraryCrary’s (1960) work on the same ice shelf and on the ice island T-3 gives a quite different time scale of positive and negative balance conditions. He concluded that the present period of negative balance has continued for the last approximately 400 years and was preceded by a period of positive balance about 1 200 years long.

Origin and age of the ice cap

The 5 m of ice from the bottom of the core consist of relatively fine ice with a lower dirt content than the 40 m of ice above it (Figs. 2a and 5). Some of the layers in the bottom 5 m of ice may have developed from firn but other layers are composed of equidimensional crystals which are characteristic of superimposed ice. The fabric and the low dirt content suggest that for each of several years snow remained over most of Meighen Island and that the annual snow accumulation was largely changed to ice by melting and refreezing. Once a snow or ice cover remained throughout the year, the surface albedo in summer was so changed that even a return to slightly warmer conditions was insufficient to produce a negative balance. The occurrence of dirt layers throughout the entire core (Fig. 5) suggests that the ice cap has not covered the entire island for more than a few years at any time throughout its history.

At a rate of growth of 12.6 cm of ice/year the ice cap would take approximately 1000 years to reach a maximum thickness of 125 m. As it is almost certain that many years of negative balance occurred during its growth, the ice cap must be considerably older than this. However, as all the evidence indicates that it has never been very much thicker than at present it can hardly have survived the climatic optimum, which it almost certainly postdates. In this respect it forms a contrast with the ice sheets on Ellesmere Island (Reference Hattersley-SmithHattersley-Smith, 1961) and Axel Heiberg Island (Reference Müller and FletcherMüller, 1966), and the Barnes Ice Cap (Reference IvesIves, 1962), which have survived from the Wisconsin glaciation. The Ellesmere Ice Shelf, like the Meighen Ice Cap, has developed in sub-Atlantic times (Reference CraryCrary, 1960).