Introduction

Wilkes Station, on Clark Island (Fig. 1, p. 843), is situated in the northern part of the Windmill Islands groupFootnote * (Fig. 2). These islands, on the east side of Vincennes Bay, at the western end of the Budd Coast, lat. 66° S., long. 110° E., were discovered and first photographed from the air in 1947. Australian, Russian and American parties visited them in 1956, but no sustained observations were begun in the area until February 1957, when Wilkes Station was established as part of the United States programme for the I.G.Y. This paper summarizes the glaciological work of the wintering over parties of 1957 and 1958. Glaciological data obtained at Wilkes Station in those two years is being compiled and published by the Ohio State University Research Foundation, and reports as they are issuedReference Cameron ^{1 ,} Reference Robertson ^{2 ,} Reference Hollin ³ are made available through World Data Center A (Glaciology) at the American Geographical Society, New York 32, N.Y. Further analysis of the original data is expected to proceed for several years and will be published both in official reports and individual papers. Distances, heights and other figures in this paper are provisional, and final figures will be found in the reports mentioned.

Fig. 1. Aerial view of ice sheet, shear moraine and Ramp east of Clark Island. Clark Island, approximately 5 km. long, is the rock peninsula in the centre of the photograph. Wilkes I.G.Y. Station lies on the western tip of the peninsula, just outside the photograph. Photograph by C. Cronk, February 1959

Fig. 2. The Windmill Islands area

The Windmill Islands

The Windmill Islands are the easternmost of three “oases” which are the major rock outcrops on sixty degrees of the East Antarctic coast between Prydz Bay in the west and the Adélie Coast in the east. The other two are Bunger Hills (long. 101° E.) and Vestfold Hills (long. 78° E.). These three outcrops have much in common, and have been compared recently by Korotkevich.Reference Korotkevich ⁴ The Windmill Islands occupy a rectangle roughly 15×30 km. and contain about 75 km.² of exposed rock. They were mapped geologically by Robertson in 1958, and comprise chiefly migmatites in the north and an intermediate igneous intrusion in the south.Reference Robertson ² Potassium-argon dating gives ages of between 950 million and 1,115 million years for these rocks.Reference Cameron ⁵ Many of the outcrops labelled islands on maps of the area are actually peninsulas, abutting directly against the ice sheet and providing easy access to it. Clark Island, on which Wilkes Station lies, is one such peninsula. From Wilkes Station the other peninsulas are easily accessible by vehicle, either via the ice sheet, or over the sea ice, which provided poor travelling to the south in 1957, but good travelling to all the islands from June to December 1958. Sea-ice studies during 1958 have been described by Tressler.Reference Tressler ⁶ The islands are rugged, particularly the southern ones, but their total relief is not great, and the highest summit in the group is only 113 m. above sea-level. This small relief contrasts sharply with the submarine topography offshore, where a 2,000 m. deep trench and other features suggest faulting and deep glacial erosion.

The Ice Sheet

In March 1957, an auxiliary station named “S-2” was established on the ice sheet 80 km. inland from Wilkes Station, at an elevation of 1,166 m. A Jamesway hut was allowed to drift over, and tunnels were excavated from it to various rooms hewn from the firn and to a 35 m. deep pit (Fig. 3). Work on the ice sheet in 1957 was concentrated at this inland station and on a December journey over the sea ice to Cape Poinsett. The 1958 work included a 600 km. traverse to the south and west of the Windmill Islands, made with two “Weasels” and one “Sno-cat” from September to November. Work on the S-2 trail and on the actual Windmill Islands continued throughout the two years. The S-2 trail and the traverse route are shown in Figure 2. No rock outcrops were found anywhere in the interior.

Fig. 3. Diagram of auxiliary station S-2

The form of the ice sheet in this part of Antarctica was determined by surveys carried out by the glaciological parties. Figure 4 is a surveyed profile of the trail from Wilkes Station to S-2. The profile is smoothly parabolic and the 80 km. trail unbroken by crevasses. At the upper end, at S-2, the surface slope is in no direction more than 40′. A baseline and movement network roughly 20 km.² in area were surveyed here in March 1957 and resurveyed in January 1959. Down the trail the surface angle increases and averages a little over 2° in the upper part of the ablation zone, approximately 4 km. inland from the islands and 225 m. above sea level. The main topographic feature of the ablation zone of the ice sheet is a shear moraine a few metres wide and 25 km. long, approximately 120 m above sea-level (Fig. 1). This moraine is composed of debris brought to the surface by the upward shearing of the ice sheet as it presses against “The Ramp” (see below) and the rock of the Windmill Islands. It provides a particularly fine example of a moraine type recently described from Thule, Greenland, by BishopReference Bishop ⁷ and Schytt.Reference Schytt ⁸ Between July and December 1958 the movement relative to rock of the ice immediately inland from the moraine was 69 cm. forward and more than 13 cm. upward. Below the shear moraine, and between it and the actual rock peninsulas, is an ice slope of 5° which averages one kilometre in length. This slope, “The Ramp” (Fig. 1), is composed of glacial ice and superimposed ice less active than that above the shear moraine. Down on the peninsulas the ice cover is no longer continuous, but patches of perennial snow and ice in the lee of rock eminences cover much of the area.

Fig. 4. Profile of the ice sheet between Wilkes Station and S-2

The profile described above is typical only of the ice sheet inland from the peninsulas of the Windmill Islands. Inland and south-south-west of the archipelago, in the area of the 1958 traverse, the ice sheet is much more undulating and often crevassed. In the area of “Mt. Long”, a cartographic feature apparently non-existent, the ice sheet forms a series of crevassed domes with a relief of one or two hundred metres. A few kilometres to the north the ice overlooks the west side of Vincennes Bay, and drops to the sea in a slope broken by long parallel crevasses often over 10 m. wide. On the sea coast the forward movement of the ice sheet is much greater than where it abuts against the Windmill Islands. South-west of the islands the Vanderford Glacier,Reference Løken and Cameron ⁹ 15 km. wide, moved forward in its centre more than 2 m. per day between March 1957 and November 1958. This ice stream must drain a large part of the area covered by the 1958 traverse. North-east of the islands, the ice sheet is not so well channelled. Ice movement at the tip of Cape Folger averaged only 15 cm. per day between March 1958 and December 1958.Reference Tressler ⁶

Ice Thickness

The chief geophysical efforts of the United States I.G.Y. programme were concentrated in West Antarctica, and no major traverses were planned from Wilkes Station. However, during his visit to the station in the summer of 1957–58, Mr. J. K. Sparkman of the University of Wisconsin measured the value of gravity at points on rock and on the ice sheet to 64 km. inland. Because in the interior these measurements could not be tied to seismic work or to measurements on a nunatak, the thickness of the ice near its margin has been calculated after extrapolating under the ice sheet the Bouguer anomalies obtained from the measurements on rock outside the ice sheet. In Figure 4 this extrapolation has been made for 9 km. and is probably quite accurate over this short distance.

Over this 9 km. the shear stress τ on the bed of the ice sheet, calculated after Nye Reference Nye ¹⁰ as approximately equal to dgh sin A, varies only between 0.6 and 0.7 bars (d being the density of the ice, g the gravitational acceleration, h the thickness of the ice, and A the surface slope). This relatively low value is consistent with the apparently smooth base of the ice sheet, the small movement described above, and possibly high basal temperatures. It should be stressed, however, that the ice was not at the melting point at its margin, and that away from a few deep lakes the rock peninsulas are frozen to an unknown depth. As has been pointed out by various workers, the Antarctic “oases” are not a geothermal phenomenon.

Observations at the Surface

Stake observations of accumulation and ablation were made at 60 points along the S-2 trail. For the period February 1958 to January 1959, assuming a density of 0.4 g./cm.³, the stakes show an average accumulation over the whole trail of 13.7 cm. of water. This figure actually represents the accumulation left after an unknown amount of original precipitation has been lost by subsequent sublimation or has been reworked and blown out to sea by the wind. On the lower part of the ice sheet the winter accumulation is ablated away each summer and a net loss occurs. Ablation at the coast obviously takes place by calving and probably by bottom melting. But where the ice ends on the Windmill peninsulas, and inland to S-2, the chief processes of ablation appear to be melting and evaporation in the summer and sublimation throughout the year. Large losses by sublimation were recorded in the superimposed ice and ablation zones.

Most of the familiar surface features of ice sheets were observed. In the superimposed ice and ablation zones, radiation hollows initiated in snow were gradually metamorphosed in late summer into ice which then took the form of ridges a few centimetres high facing the mid-day sun. These ridges were so ubiquitous and so strictly aligned that they were successfully used as directional guides for night-time “Weasel” journeys involving up to 40 km. of navigation over the lower ice sheet. In the firn zone, and particularly in the area of the 1958 traverse, surface forms could be grouped basically into those constructed by the blizzard winds and those eroded by katabatic winds, as at Mawson.Reference Mellor ¹¹ The predominant lineation of the former was 105°, and of the latter 150°. Sastrugi were not well developed in the area of S-2, but in domed and windy regions of the traverse they reached heights of 1 m. Whalebacked firn ridges, formed downwind from crevasse bridges, had a relief up to 10 m., and in poor light were often mistaken for nunataks.

Observations below the Surface

Observations below the surface were made at 60 locations by means of pit and core studies. Pits were normally 1 to 2 m. deep. Cores, 75 mm. in diameter, obtained with a SIPRE auger, were normally taken to a depth of 10 m. The deep pit at S-2 was dug to 35 m., and cores for an additional 27 m. to a total depth of 62 m. were obtained with the auger.Reference Cameron ¹² In most of these locations studies were made of wetness (summer only), hardness, density, grain size, crystal size and fabric, and electrical conductivity.

Densities recorded included 0.05 g./cm.³ (one of the few snowfalls in which crystals or grains were not rounded by wind action), 0.37 g./cm.³ (average value for winter snow on the lower ice sheet), and 0.48 g./cm.3 (wind-packed new snow in some lee locations). On the inland traverse unusually high densities of 0.6 g./cm.³ at the surface and 0.8 g./cm.³ at 10 m. were measured in an especially windy area of the dry firn zone.Reference Hollin ¹³ At S-2 the complete transformation from snow (0.43 g./cm.³ at the surface) to ice (0.91 g./cm.³ at 61 m.) was plotted by means of 544 density determinations. Density profiles from both the windy area and S-2 are shown in Figure 5. Grain diameters in dry firn areas were normally less than 0.5 mm. in the case of new snow, but increased to more than 0.5 mm. after one summer and to the order of 1.0 mm. at a depth of 10 m. Crystal cross-sectional areas ranged from less than 0.1 mm.² in new snow to 3 mm.² at 37 m. in the deep pit, 6 mm.² at 61 m. in the pit, 30 mm.² in the shearing area of the ablation zone (much smaller in the actual shear planes) and 50 mm.² at Cape Folger. Several individual crystals from the superimposed ice and ablation zones exceeded 7 cm. in length. Crystal fabric studies were made at 12 locations. Growth fabrics were generally random, but deformation fabrics included both girdles and isolated maxima. The electrical conductivity of melted samples of ice and snow was measured. The results fit into a definite pattern, in which snow and ice which has obviously originated near the coast and which presumably has a relatively high salt content has a much higher conductivity than that with an interior origin. This phenomenon proved a useful tool in the study of the local glacial history.

Fig. 5. Firn density curves at S-2 and in windy area of traverse

These sub-surface observations had various objectives, and in the fine zone the chief of these was the determination of past accumulation. At S-2 this was determined by Cameron chiefly on the basis of density and grain size differences and crust occurrences. One hundred and seventy-four annual layers were traced between the surface and 35 m., so that the firn at that depth must have been deposited in approximately 1783. The average annual accumulation since then equals 13.3 cm. of water. Checks on this value may be provided by pit data from 1958, ¹⁸O/¹⁶O studies, the application of Sorge’s law,Reference Bader ¹⁴ and the deformation recorded by the relative movement network mentioned earlier. Compaction of the deep pit (roofed over and laddered) is being measured on a wire from the top to the bottom and by pegs inserted in the walls. At the 30 m. level of the pit a horizontal deformation tunnel 2 m. in diameter and 6 m. long was completed in October 1957. The closure of this tunnel and its movement relative to the pit shaft is being measured by pegs at various distances inside and by a deformation grid on the end wall. Up to January 1959 the closure at the tunnel midpoint was 2.9 cm. vertically, 2.2 cm. diagonally and 1.6 cm. horizontally.

Meteorological Observations and Temperatures in the Upper Ice

For future studies of the heat balance on the ice sheet, meteorological observations are available from three stations. At the main base a full meteorological programme which included twice daily upper air soundings was conducted from the beginning of 1957 by aerologists of the U.S. Navy. At S-1 (Fig. 4), 7.5 km. inland from the base and 263 m. above sea-level, a thermograph was operated throughout the two years and wind speed observations were made during December 1958 and January 1959. At S-2 three-hourly observations of temperature, cloud cover, wind speed and direction and snow surface conditions were made continuously from May 1957 to January 1958 and occasionally afterwards. Intermittent radiation measurements were made at both Wilkes Station and S-2.

Because summer there normally includes one or two months with mean temperatures above 0° C., the base and its immediate area fit into the tundra category of most climatic classifications. At the base the highest temperature recorded was 8° C. in January 1958 and the lowest −38° C. in July 1958. Temperatures were generally lower in 1958 than they were in 1957, and the snowline in the summer of 1958–59 was 350 m. lower than in 1957–58. The mean wind speed at Wilkes, 18 km./hr. between March 1957 and January 1958, was the lowest recorded at any East Antarctic station. However, the generally calm weather of Clark Island was regularly interrupted by storms, and gusts of at least 95 km./hr. were experienced during every month between March 1957 and December 1958. In April 1958 the wind mast was blown down whilst recording a gust of 214 km./hr. During one of these storms rain fell as late in the year as 17 May in 1958. At S-2 the highest temperature recorded was 4° C. in December 1957 and the lowest −48° C. in June 1957. Unlike the base, with its usual calm and occasional high winds, S-2 was normally windy, but never experienced such extreme winds as the base. The mean wind speed at S-2 between March and December 1957 was 40 km./hr. Mean wind speeds in the area of the 1958 traverse are probably even higher than at S-2. On the traverse the average of twice daily wind observations for 61 days was 47 km./hr.

The firn temperatures at a depth of 10 m. (where the range of the annual temperature wave is reduced to less than 1° C.) were obtained whenever cores were taken from the ice sheet. Ten-metre temperatures on the S-2 trail are plotted against altitude in Figure 6. Above 375 m., in the firn zone, the straight line of the plot probably represents a direct relationship between air temperature and firn temperature. Below 375 m., in the superimposed ice and ablation zones, factors such as melting and differing albedos alter this relationship and, as can be seen from Figure 6, have a net warming effect. At S-1 ice temperatures at depths of

Fig. 6. Altitude vs. temperature at 10 m. depth

The Regimen

The sub-surface observations on the trail from S-2 to the main base provide a typical section of a polar glacier and its accumulation and ablation zones.Reference Hollin ¹³ In the firn zone of accumulation, S-2, 1,166 m. above sea-level, is probably slightly below the upper limit of summer melt infiltration on the ice sheet. Down-slope from S-2 ice laminae and layers produced by summer melting increase rapidly in thickness. For example, at 850 m. elevation continuous ice layers observed in pits are frequently 4 cm. thick. Still further downhill, at approximately 375 m. elevation, nearly all of each year’s net accumulation is transformed into ice by summer melting and subsequent refreezing; and this elevation represents the upper limit of the superimposed ice zone of accumulation. From this level the superimposed ice zone extends for 5.5 km. down the trail and provides a strikingly smooth and regular surface for travel. Tentative criteria have been established for distinguishing annual layers in this zone. Features of the superimposed ice were groups of extremely thin, hexagonal, plate-like cavities, usually 1 or 2 mm. across, and always parallel to the basal plane of the containing ice crystal. These cavities differ from true vapour figures in that they are filled with gas, but they do appear to be formed by solar radiation. They may, therefore, be of stratigraphic value since they must form primarily near the top of any given layer of superimposed ice. At the lower limit of the superimposed ice zone at approximately 225 m. elevation summer melting and run-off become greater than accumulation, and the ablation zone is entered. At a little below this elevation, shallow melt stream gullies appear.

The ablation zone above is not confined to the area of the Windmill Islands only, but extends along the coast on either side of the archipelago. Strongly preferred bubble and crystal orientations in ice from the Vanderford Glacier and Cape Folger show that net ablation is occurring in those places also. In other words the ablation here is not merely a local warming effect produced by the low albedo of the Windmill Islands. Climatically and glaciologicalIy, therefore, it is misleading to call this area an “oasis”. However, the term is certainly accurate in a biological sense and, because it is convenient and expressive, it may survive. It should be stressed that the regimen described here is not common to all of the Antarctic coast in the same latitude, and that in areas such as the West Ice Shelf (long. 85° E.) closely interrelated topographic, meteorological and oceanic factors can bring the accumulation zone down to sea-level.

Glacial Geology

Glacial geological observations included studies of the small amount of fluvial erosion in the area, of lake sediments, and of physical and chemical weathering phenomena. The main feature of this area is the 25 km. long shear moraine previously described. End moraines are rare, and most of the drift has the form of a thin and patchy cover of ground moraine or till. Below a well defined line identified by LøkenReference Cameron ¹ at approximately 30 m. elevation this cover has been swept away or into hollows by the sea, which must have previously been higher relative to the land. The vegetation of the area has been described by Llano,Reference Cameron ¹⁵ and the marine life by Tressler.Reference Davis and Schaeffer ¹⁶

Historically, little evidence of multiple glacial stages can be expected from such a small and low-lying area. However, a sequence of four events has been traced. (1) The occurrence of striae and erratics everywhere on the islands shows that they have been completely covered by ice at some time, and the amount of weathering suggests that the ice last retreated across, the islands at least several thousand years ago. (2) Subsequently, probably isostatically, the land has been uplifted 30 m. in relation to the sea. (3) Since the uplift the ice may have. readvanced. This is suggested by the position of the ice sheet only a few metres away from completely intact raised shoreline sequences (wave-washed rocks from 30 m. down to sea-level) on the Balaena Islets and Lewis Islet (long. 134° E.). (4.) Most recently a thinning of perennial, snow and ice patches by at least 1.5 m., probably in the last few score years, is shown by a well marked lichen and moss trim-line on Clark Island.

In an attempt to provide a chronology for the events above, various samples are being dated by different means. The period of deglaciation on the islands may be determined by the ³⁶Cl method,Reference Davis and Schaeffer ¹⁶ despite the complication of contamination by marine chlorine. Specimens. collected for ¹⁴C dating of post-glacial events include coralline algae, marine shells, penguin bones from long abandoned rookeries, bones from the mummified bodies of elephant seals and whalebones. On the raised beaches search was made for both driftwood and pumice but neither was found. As dates are supplied, and as field work is further extended and correlated, it is hoped that our insight into the history and process of glaciation in Antarctica will develop rapidly.

Acknowledgements

This paper is based on the field work and ideas of all members of the glaciological parties. These comprised in 1957 R. Cameron (Chief Glaciologist), O. Løken and J. Molholm, and in 1958 J. Hollin (Chief Glaciologist), C. Cronk and R. Robertson. All were employed for field work by the Arctic Institute of North America and for data reduction by the Ohio. State University Research Foundation. The logistic and personal help of all members of the U.S. Navy and civilian wintering over and summer support groups and of the Australian National Antarctic Research Expedition is acknowledged with gratitude. The authors express their thanks to Dr. A. P. Crary and Dr. R. P. Goldthwait for reading and commenting on the manuscript of this paper.