Introduction

During the austral summer of 1963–64, personnel of the University of Wisconsin carried out a program of aeromagnetic observations in Antarctica (Reference BehrendtBehrendt, 1964[b]). In the course of this work barometric and radio-altimeter observations were made at approximately 15 km. intervals, or closer over steep surface slopes, which enabled calculation of snow surface elevation along the flight lines. The radio-altimeter used operated at 4,300 Mc./sec. so snow penetration was negligible. Oversnow traverse tracks were used as control where crossed. As several of these flights were in the relatively little known Filchner Ice Shelf area they enabled an improved surface elevation map to be drawn.

Data from the following oversnow traverses in the area were also used: Sentinel traverse, 1957–58 (Reference Bentley and OstensoBentley and Ostenso, 1961), Ellsworth-“Byrd” traverse, 1958–59 (personal communication from E. A. Bradley), Filchner Ice Shelf traverse, 1957–58 (Reference BehrendtBehrendt, 1962[b]), Commonwealth Trans-Antarctic Expedition, 1957–58 (Reference PrattPratt, 1960), Ellsworth Highland traverse, 1960–61 (personal communication from C. R. Bentley), Antarctic Peninsula traverse, 1961–62 (Reference BehrendtBehrendt, 1964[a], “Byrd”-Pole traverse, 1960–61 (personal communication from F. L. Dowling) and the 1963–64 “Byrd station” traverse (personal communication from M. Hochstein). In addition surface elevation data were used from the 1960–61 aeromagnetic flights (Reference Behrendt and WoldBehrendt and Wold, 1963). The oversnow traverse elevations are accurate for the most part to ±50 m. The adjusted elevations along the flight lines are believed accurate to about ±100 m., although they are probably better over the ice shelf itself.

Discussion

Figure 1 is the snow surface elevation map. Several features shown warrant some discussion. The ice stream draining a large part of the polar plateau is indicated by the contours in the area south of lat. 85° S. between long. 60° and 100° W. These contours are not as steep in one location as those shown by Reference BehrendtBehrendt and others (1962). An error was found in one of the flight lines used in that map; these data were not used here.

Fig. 1.

Snow surface elevation map of the Filchner Ice Shelf area of Antarctica. Elevation data are spaced for the most part at about 6 km. intervals on oversnow traverses and at about 15 km. on flight lines

Reference BehrendtBehrendt (1962[a]) postulated the existence of this ice stream on the basis of the mass discharge of the Filchner Ice Shelf, and calculated that about 100 × 10¹⁵ g. yr.^-1 of mass flowed out of the section of the ice front between Berkner Island and the coast to the east. Reference Giovinetto and MellorGiovinetto (1964), in a discussion of the ice drainage basins of Antarctica, calculated that the mass output in the eastern part of the Filchner Ice Shelf drainage system should amount to about 200 × 10¹⁵ g. yr.^-1 from a determination of the mass input in the entire system. The difference of 100 × 10¹⁵ g. yr.⁻¹ was greater than the observational errors would allow, even neglecting bottom melting beneath the ice shelf. Both authors assumed that the bulk of the drainage from the polar plateau was through the section east of Berkner Island, based on the map of the Filchner Ice Shelf available following the 1957–58 Filchner Ice Shelf traverse (Reference NeuburgNeuburg and others, 1959). This traverse party, of which the author was a member, projected the grounded ice feature crossed at about lat. 80° 30′ S., long. 62° W. north-eastward to connect with Berkner Island. Two parts of one of the flights shown in Figure 1 crossed the ice shelf in this area and showed no evidence of a grounded ice connection. This being the case, it is clear that a large part of the ice entering from the polar plateau is flowing through this section.

A simple calculation was made to determine whether the discrepancy of 100×10¹⁵ g. yr.^-1’ noted by Reference Giovinetto and MellorGiovinetto (1964) could be explained. The distance between Berkner Island and the grounded ice feature to the west-south-west is about 160 km. as shown in Figure 1. Reference BehrendtBehrendt (1962[a]) showed an ice thickness of 500 m. for this part of the ice shelf. The product of these quantities gives the cross-sectional area of the ice flowing through this section, 8×10¹¹ cm.². Since the error estimate cited by Giovinetto is ±57×10¹⁵ g. yr.⁻¹, the velocity of ice-shelf flow required through this section is obtained by dividing the cross-sectional area into the mass flux difference, viz. 1.25±0.71×10⁵ cm. yr.⁻¹ or 1.25+0.71 km. yr.⁻¹. This is a reasonable figure and thus the discrepancy can be accounted for.

In Figure 1 the line indicating the boundary of the grounded ice connects the feature west-south-west of Berkner Island to the higher ice farther west. This was done because surface observations made by the author and other members of the Filchner Ice Shelf traverse party showed that the grounded ice contact in the vicinity of this feature was at an elevation of about 110 m. This contrast with the 200 m. contour border of the ice shelf in other areas is probably the result of the fairly shallow depth to bedrock in the area where this grounded feature was crossed (Reference BehrendtBehrendt, 1962[a]) relative to the parts of the area where the 200 m. contour defines the boundary.

The western area of the ice shelf is still relatively unknown and several islands other than those shown could exist. The precise locations and configurations of these are uncertain as they were only seen from the air, with the exception of the one reached by the Filchner Ice Shelf traverse. The position of the flight line shown in the area of these islands is possibly somewhat in error.

It is likely that the boundary of the ice shelf in the vicinity of lat. 77° S., long. 70° W. may actually be farther to the north and that the mountains(?) in this area may be the same as those studied by the Antarctic Peninsula traverse (Reference BehrendtBehrendt, 1963) in the vicinity of lat. 75° S., long. 70° W. On the flight shown in the latter area, I observed no rock exposures south of those shown and the snow surface appeared to descend continuously to the south. Visual observations on this same flight also showed the apparent ice shelf boundary south of “Eights station” where elevations of less than 200 m. were measured.

The area of ice shelf shown in Figure 1 is 0.43×10⁶ km.², not including the islands, which compares with 0.39×10⁶ km.² from Reference ThielThiel (1962), 0.355×10⁶ km.² from Reference SuyetovaSuyetova (1963) and 0.50×10⁶ km.² from Reference GiovinettoGiovinetto (unpublished). Giovinetto’s value was based on evidence from the Antarctic Peninsula traverse and probably is in agreement with the value obtained from Figure 1 within the experimental error, although it appears somewhat high. Thiel’s and Suyetova’s values are too low and reflect the maps prior to the discovery of the extension of the ice shelf into the area immediately south of “Eights station” as discussed by Reference Giovinetto and BehrendtGiovinetto and Behrendt (1964). The area could be increased to as much as 0.48×10⁶ km.² if the north-western border approaches the mountains mapped by the Antarctic Peninsula traverse as discussed above. These values are comparable to the somewhat larger Ross Ice Shelf of 0.54×10⁶ km.² (Reference Giovinetto and MellorGiovinetto, 1964).

Acknowledgements

Richard Wanous and Per Gjelsvik deserve thanks for collecting the altimeter data in conjunction with their aeromagnetic work. Air Development Squadron 6 (VX6), U.S. Navy, provided the air support for this program. The field work was supported by a grant from the National Science Foundation to the University of Wisconsin and was carried out while the author was employed at that institution.