Introduction
The 1959 and 1960 Expéditions Antarctiques Belges were organized by the Centre National de Recherches Polaires de Belgique. The expeditions’ headquarters, “Base Roi Baudouin” (lat. 70° 26′ S., long. 24° 19′ E.), were constructed by the 1958 Expédition Antarctique Belge. The programme of the expeditions included geological, gravity and geodetic surveys by the authors. The bulk of these investigations deal with the Sør-Rondane, an impressive mountain range approximately 200 km. south of “Base Roi Baudoin” (Fig. 1). In addition to their main programme, the authors undertook movement and ice-thickness measurements on some of the glaciers in the western part of this range. The main results of these measurements are reviewed in this paper. The Campagne d’Été, Expédition Antarctique Belgo-Néerlandaise, 1964–65, was organized by the Comité Antarctique Belgo-Néerlandais.
Fig. 1. Eastern Dronning Maud Land. Belgicafjella were first photographed by the Expédition Antarctique Belge, 1958 (leader G. de Gerlache) and Dronning Fabiolafjella by the Expédition Antarctique Belge, 1960 (leader G. Derom)
The Sør-Rondane
The Sør-Rondane forms a 250 km. long, east–west, wedge-shaped mountain range which acts as a barrier and dams the glacier flow from the Antarctic Ice Sheet (Fig. 2). It seems that the flow in the unobstructed areas to the east and west of the mountain range is far more important than the flow through it. However, about 15 major glaciers drain the ice from the polar plateau and subdivide the range into its main units. In general, these glaciers are orientated south to north but their direction may vary from north-north-west to north-east. Their widths range from 3 to 35 km. and their lengths increase eastwards, reaching a maximum of 80 km. (Byrdbreen).
Fig. 2. Ice-flow and thickness measurements in the Sør-Rondane
In addition to the main drainage glaciers, the Sør-Rondane has its own well-developed glacierization system which is dependent entirely upon local conditions. Alpine, cirque and niche glaciers, and small ice caps occur within the range.
North of the range the confluence of the Sør-Rondane glaciers forms a piedmont zone (inland ice slope) but, as individual geographic features, the glaciers have disappeared. Near “Base Roi Baudouin” the inland ice slope joins the ice shelf.
The Sør-Rondane offers a wealth of evidence to establish the diminution of glacierization. Dry glacial valleys, extensive mammillated surfaces and vast moraine-covered areas are abundant throughout the range. There is considerable reduction in glacierization both on the polar plateau (and its drainage northwards) and locally.
A general account of the range has been given by Reference Van AutenboerVan Autenboer (1962), and more detailed data on glacierization and deglaciation are given by Reference AutenboerVan Autenboer (1964[a]).
A map on a scale of 1 : 250,000, drawn from air photographs taken by the U.S. Navy Antarctic expedition of 1946–47 (Operation “Highjump”) and published by the Norsk Polarinstitutt in 1957, proved to be most valuable during the field work. However, the altitudes have had to be changed materially as a result of the triangulation and barometric profiles by Reference BlaiklockBlaiklock (unpublished).
Observations
Ice-thickness and movement measurements were carried out on the following glaciers: Gunnestadbreen (2 profiles), Gillockbreen, the glacier between Teltet and Utsteinen, and Jenningsbreen. In addition, movement was observed on Gjelbreen and ice thickness was measured on H.E. Hansenbreen, on several smaller glaciers, and on the inland ice slope. The latter measurements are not considered here (Fig. 2).
The profiles are approximately east–west (at right-angles to the main flow direction of the glaciers) with stations at 1,600 m. intervals.
Ice-flow measurements
During 1959 and 1960 a triangulation scheme was observed in the Sør-Rondane. The geodetic and survey work consisted of about 40 triangulation stations covering an area about 200 by 50 km. (Blaiklock, unpublished). The geographical coordinates of Teltet, the primary triangulation station, were determined by astral observations. After adjustment of the scheme, the geographical positions of the individual stations were calculated and they were then transposed into grid coordinates.
Heights in the Sør-Rondane were all related through the triangulation observations to Teltet “1920”. Teltet itself was tied to “Base Roi Baudouin” and sea-level by aneroid barometer traverse. The mean of four main traverses observed during 1959–61 was used. On each traverse three aneroids were read. The standard error of the mean is 5 m. The positions and heights of the markers of the glaciological–gravimetric traverses were fixed by theodolite resection from the main triangulation scheme. Observations at the glacier stakes were repeated as frequently as possible.
At the resected glaciological stations, two sets of horizontal angles and one set of vertical angles were observed with a Wild T2 theodolite. Usually five to seven points that had been given coordinates previously were chosen. Where possible, they were triangulation stations themselves or well-fixed peaks. At stations in the middle of glaciers it was easier to choose well-placed peaks around the horizon, but some of the gravity stations on rock were at the bottom of ridges, so that visibility was restricted to one-half of the horizon.
The positions of the stations were calculated by a semi-graphical method. The coordinates of the triangulation sections used were the preliminary ones calculated during the winter at “Base Roi Baudouin” and not those finally computed at the Institut Géographique Militaire in Brussels. However, the final coordinates differ from the preliminary ones by only very small amounts and it was considered unnecessary to recalculate the resected positions, inasmuch as only differences between successive positions of the glacier stakes are considered.
For the glaciological stakes, the actual movement is of more interest than the true position. At stake 11 on Gunnestadbreen four resections were observed over a period of 32 months (Fig. 3). The direction and amplitude of the movement did not vary appreciably for the three intermediate periods. The coordinates of any resection are considered to be accurate to less than 0.5 m. relative to one another.
Fig. 3. Diagrams illustrating the semi-graphical resection of stake 11 on Gunnestadbreen at four observation dates, and the calculation of its velocity and direction
The results of ice-flow measurements for the 1959–61 period are given in Table I, and those for the 1960–65 period are given in Table II. A comparison between velocity figures for the same marker stakes during these two periods shows remarkably good agreement.
Table I. Ice-Flow Measurements in the Sør-Rondane, 1956–61*
Table II. Ice-Flow Measurements in the Sør-Rondane, 1960–65*
Gravity measurements
The gravimetric work of the 1959 and 1960 Expéditions Antarctiques Belges included relative measurements of gravity and an airborne survey, which was intended as a contribution to regional gravity mapping (Reference Autenboer and AdieVan Autenboer, 1964[b]). For glaciological purposes only differences in gravity were measured. The stations were the same as those used for the velocity measurements, with an additional station on rock at each end of the profile. The profiles were measured twice in each direction and the mean of the measured differences was taken. Correction for drift was only occasionally possible. A Worden geodetic-type gravimeter (No. 453) was used, and it had been checked on the Belgian and international (Oslo–Bodø) calibration bases. The calibration curve supplied by the manufacturers was found to be satisfactory and it was used for the computations. Transport for the glaciological gravity measurements was mainly by dog sledge and the gravimeter was suspended in a framework rigged on the back of the sledge. Using a thoroughly trained dog team pulling a well-loaded sledge and with some care to avoid steep blue-ice areas, this method of transport proved to be satisfactory. A preliminary reconnaissance to select a smooth part of the glacier and to find easy access to rock on both sides of the glacier was often necessary. The elevation and latitude determinations for the stations have been described above.
Reference MartinMartin ([1949]) was the first person to use gravity measurements for the determination of ice thickness. In his method, as it has been applied here, the observed gravity value at a glaciological station is reduced to the elevation and latitude of a reference station on rock. The density of ice (0.92 g./cm3) is used in the combined (free-air and Bouguer) elevation correction, implying the existence beneath the glaciological station of an infinite slab of ice, the thickness of which is equal to the difference in elevation between the two stations. This assumption would be correct when the value at the reference station and the reduced value of the glaciological station are equal. A difference between these two values is considered as an anomaly and requires further interpretation.
Since no information on other causes affecting the gravity values is available, the observed anomaly is explained by assuming that the thickness of the ice, and consequently the height of the rock basement, differs from the assumption by an amount h 2.
It is therefore equal to the difference between the Bouguer corrections for density, d 1 (ice) and d 2 (rock), applied to an infinite slab of height, h 2, which equals the difference in elevation between the rock basement and the reference level, i.e. h 2 = Anomaly/L(d 2−d 1). This difference is negative when the subglacial topography lies below the reference level and positive when it lies above it. However, Martin allowed no correction for the regional variation of gravity which is contained in the observed gravity (Reference LittlewoodLittlewood, 1952). On longer traverses this might introduce a large error and especially in Antarctica where a large anomaly gradient can be expected, as there is a trend towards isostatic equilibrium of the crust under the load of the Antarctic Ice Sheet.
Reference Bull and HardyBull and Hardy (1956), in their work on Austerdalsbreen in Norway, interpolated a curve of regional anomalies based upon the anomalies at the rock stations at each end of the profile. The shape of the curve is based upon the anomaly curves measured in the lower ice-free parts of the glacier valley.
For the present interpretation there is no information regarding the curve of regional anomalies and therefore the gradient has been assumed to be linear between the rock stations at each end of the traverses. The difference between the two values at the rock stations, after reduction to the same height and latitude, has been distributed over the profile. Further calculations were carried out using Martin’s method and in Figures 4–8 the calculated cross-profiles are represented by a dotted line.
In the calculations using Martin’s method the presence of an infinite slab of ice beneath the stations is implied in the Bouguer correction. This assumption can be applied to gravity stations on a large continental ice sheet if the subglacial bedrock relief is simple. However, it is bound to introduce a serious error when a valley glacier with a U-shaped cross-profile and well-confined boundaries is considered. Since this was true in the case of the Sør-Rondane glaciers, a check was necessary. This was kindly done by Ir. L. Jones, Géographe à l’Institut Géographique Militaire in Brussels, who suggested the method and carried out the computations. In this procedure the Bouguer anomalies were first computedFootnote * for a two-dimensional model of a glacial valley having the cross-profile estimated by Martin’s method. Two of the dimensions of this model, representing the cross-section of the glacier, are situated within the vertical plane of the gravimetric profile. The third one, perpendicular to this plane, is assumed to be infinite and corresponds to the long axis of the glacier. Comparisons of the anomalies calculated by this method and the observed anomalies show a serious discrepancy (or remanent anomaly) (Figs. 4–8). This discrepancy is greatest in the central part of the glacier where the gravity effect of the basement rising towards the boundaries of the glacier has to be accounted for. The thicknesses calculated by Martin’s method apparently do not correspond with the true ones.
Fig. 4. H.E. Hansenbreen. Cross-profile and Bouguer anomalies. Profile A is calculated using Martin’s method. Profile B is the adopted profile
Fig. 5. Glacier between Teltet and Utsteinen. Cross-profile and Bouguer anomalies. Profile A is calculated using Martin’s method. Profile B is the adopted profile
Fig. 6. Gunnestadbreen. Cross-profile and Bouguer anomalies. Profile A is calculated using Martin’s method. Profile B is the adopted profile
Fig. 7. Gillockbreen. Cross-profile and Bouguer anomalies. Profile A is calculated using Martin’s method. Profile B is the adopted profile
Fig. 8. Jenningsbreen. Cross-profile and Bouguer anomalies. Profile A is calculated using Martin’s method. Profile B is the adopted profile
In order to achieve a better approximation to the subglacial profiles, further models were successively chosen, their effects calculated using Hubbert’s method and compared with the curve of observed anomalies. When the calculated curve of anomalies coincides with the observed one, the model apparently represents the cross-section of the glacier, excluding the effects of possible geological accidents and of changes in densities (e.g. moraines contained within the ice). For practical purposes the results are considered satisfactory when the two curves are within a few milligals of each other. The adopted ice thickn given in Table III and are represented in Figures 2 and 4–8.
Table III. Ice Thicknesses Determined from Gravity Measurements
Discussion
Flow rates, directions of flow and ice thicknesses measured in the Sør-Rondane are represented in Figure 2. The annual discharge of four of the Sør-Rondane glaciers can now be roughly estimated. In this calculation the measured surface velocities were applied to corresponding vertical slabs of the cross-profile, drawn as a function of the determined ice thicknesses. As the vertical distribution of the rate of movement is not constant and diminishes downwards (Reference MathewsMathews, 1959), these figures are maximum estimates.Footnote *
No velocity measurements were made on H.E. Hansenbreen. However, other observations (Reference AutenboerVan Autenboer, 1964[a]) indicate that a greater rate of movement can be expected in this glacier. The cross-section of H.E. Hansenbreen is approximately 1.85 times the cross-section of Gunnestadbreen. By estimating its velocity as being at least twice the measured one for Gunnestadbreen, an estimate of 302 × 106 tons/yr. is reached. A comparison of the velocities of Gunnestadbreen and Gjelbreen indicates that the discharge of the latter should not exceed that of Gunnestadbreen.
A study of available air photographstFootnote † and field observations indicates the main sources of nourishment of these glaciers (Reference AutenboerVan Autenboer, 1964[a]).
It is likely that the discharge of H.E. Hansenbreen exceeds the total discharge over a large part of the western part of the range. The Sør-Rondane generally acts as a barrier and dams the flow of ice from the polar plateau. The drainage in the areas to the east and the west of the range, which are free from obstruction, is far more important than the discharge through it. H.E. Hansenbreen is situated at the western limit of the range and its discharge clearly indicates the importance of this deflected flow.
Gunnestadbreen is clearly the most important of the glaciers within the range for which an estimate is available. This glacier is actively fed from the polar plateau. From the 1959 Norwegian air photographs, it appears that to the south there is a trough which funnels the plateau ice supply into the narrow confines of the lower reaches of the glacier. Local glaciers form an important contribution in addition to the plateau supply. These are small glaciers on the southern slopes of Widerøefjellet which discharge southwards and south-eastwards. Their local flow is deflected northwards when they join the main ice discharge from the polar plateau.
Quite a different situation exists east of Gunnestadbreen and south of Gillockbreen, Ellisbreen and Jenningsbreen. Here the Antarctic Ice Sheet approaches a local rise, as can be assumed from the nunataks projecting through it. The ice sheet glides evenly over this bedrock highland area as a smooth, featureless and crevasse-free plain, the ice of which is probably thinner than that supplying Gunnestadbreen. As it reaches the northern edge of the highland it cascades down into the three small glaciers.
The glacier between Utsteinen and Teltet has at present no direct supply from the polar plateau, although evidence of its former extension indicates that this was previously so. It is therefore a local glacier only because its ice supply has been cut off, and its estimated discharge is believed to reflect local nourishment.
There is clearly a direct relationship between the discharge of the Sør-Rondane drainage glaciers and their supply from the plateau.
Acknowledgements
The authors wish to thank the members of the expeditions, especially G. Derom, J. Dubois, L. Goossens, J. Verheyc and J. J. Derwael, who assisted with the field work. The advice and assistance of Ir. L. Jones, Géographe á l’Institut Géographique Militaire in Brussels, who introduced one of us (T.V.A.) to the use of a gravimeter, is gratefully acknowledged. He also carried out the second part of the calculation of the gravity measurements. Professor P. de Bethune (Université de Louvain) and Dr. J. Behrendt (Geophysical and Polar Research Center, University of Wisconsin) read the manuscript and contributed constructive criticism. Dr. R. J. Adie (British Antarctic Survey, Department of Geology, University of Birmingham) kindly revised the manuscript.
