Introduction

In 1988 a mass balance study was initiated in Dronning Maud Land on East Antarctica (Reference Holmlund, Isaksson and KarlénHolmlund and others, 1989). The project included snow coring for accumulation studies and ice velocity surveys and ice depth soundings for balanced flux calculations. The area chosen for this study is an ice stream between the two mountain ranges, Vestfjella and Heimefrontfjella. Mass flux calculations indicate very little flow through the mountain ranges. As much as 60% of the annual flux through the outlet profile of Veststraumen is balanced by the accumulation of snow between the ranges.

Besides their use in the mass balance study, radio-echo records have been interpreted in order to distinguish between parts of the ice sheet frozen to the ground and areas where the basal ice is at the pressure melting point.

Radio-echo soundings were carried out in a nunatak area in Vestfjella and in an ice covered part of Heimefrontfjella, Dronning Maud Land (Fig. 1). The strength of the registrated radio-echoes varied about 10 dB at similar depths indicating differences in the basal thermal conditions of the ice and/or differences in the absorption of radio waves within the ice.

The strength of a reflected pulse is governed by modulations within the transmitter and/or differences in the dielectric constant, or the relative permittivity in the ice. The difference in dielectric constants between ice and water is about 14 times larger than the difference between constants for ice and rock (Reference Smith and EvansEwan Smith and Evans, 1972). Small fractions of water at an interface influences the phase and strength of the reflecting signal (Reference Walford, Kennett and HolmlundWalford and others, 1986). Different rock types do not have significantly different dielectric constant and thus do not have a primary influence on the strength of the reflected pulse. A frozen till or sediment layer may, however, weaken the echo by making the interface less sharp, smearing the changes in physical properties over more than a wavelength.

In Antarctica, where ice depths generally are large, reflected radio-echo pulses are weak. The absorption of signal strength is on the order of 1 dB 100 m⁻¹ for 150 MHz in pure ice (Ewan Reference Smith and EvansSmith and Evans, 1972). The intense reflection from a subglacial waterspot or water film will show up very clear. However, at high altitudes where temperatures are low and no blue ice areas occur, similar bed echoes may occur where the rate of snow accumulation is low. This cannot be explained by basal melting but rather by a higher permittance of radiowaves.

The radio-echo soundings were made within the frame of the Swedish Antarctic Research Programme (SWE-DARP) in January 1989 and 1992.

Fig. 1. Location map, East Antarctica.

Physical Settings

Eastern Heimefrontfjella

The Heimefrontfjella mountain range is situated about 150 km southeast of the Vestfjella. The range is about 150 km long and reduces the outflow of ice from Amundsenisen to Maudheimvidda situated downstream. The drop in elevation is about 1500 m. The highest mountain, Paalnibba, reaches 2711 m a.s.l. East of the range (75°10′S, 10°30′W and 8°30′W) there are two subglacial mountains, covered by only some hundreds of meters of ice. They are visible at satellite images as an uneven wrinkled surface (IfAG, 1988, 1989). The snow surface is rough with sastrugi up to 1 m. The annual rate of snow accumulation is probably low in this area as the flux rates in the Heimefrontfjella valleys draining the area are low. Accumulation, as measured in a bore hole in a nearby site, gave an annual rate of 104 mm (Reference IsakssonIsaksson, 1992). This area is an extension of the saddle-like ice divide from the Dome F ending at Heimefrontfjella.

Fig. 2. Generalized figure showing A, mass balance conditions, bn, and a simplified curve, T, showing a probable temperature distribution in an ice sheet B, and in a blue ice area C.

At Amundsenisen there are few signs of crevasses, in contrast to the valleys in Heimefrontfjella where the ice is heavily crevassed while dropping about 1000 m in altitude.

The northernmost part of Heimefrontfjella, called Milorgfjella, was first observed during the Schwabenland expedition in 1938/39. The whole range was later mapped and named during the Norwegian-British-Swedish expedition 1949/51 (Reference Gjaever and SchyttGjaever and Schytt, 1963). The range contains four groups of mountains separated by valleys. The most prominent valley is called Kibergdalen. The bed topography in Heimefrontfjella has been mapped by Holmlund (Reference Herzfeld and HolmlundHerzfeld and Holmlund, 1988, 1990;Reference Hoppe and Thyssen Hoppe and Thyssen, 1988).

Temperature Distribution in Ice Sheets

Most of the East Antarctic ice sheet has a positive surface mass balance, giving higher velocities near the coast. This implies a generally downward vertical flux (Fig. 2A). This transportation of cold ice and snow influences the temperature distribution within the ice in the sense that the thermal gradient is flatter than the geothermal gradient in the bedrock beneath. The gradient is, however, positive and steep close to the bottom because of heat released by internal deformation and because of the influx of heat from the underlying bedrock (Fig 2B).

If the mass balance is negative, as it is on blue ice fields in nunatak areas, the temperature distribution is significantly different (Fig. 2A and C). The net loss of mass at the surface is compensated by a vertical movement upwards in order to preserve a balanced state. The gradient of the temperature profile is steep close to the surface and more gently sloping close to the bottom (Fig. 2C), as on a subpolar glacier (Reference Hooke, Gould and BrzozowskiHooke and others, 1983). The governing variables are the rate of ablation and the air temperature. The base of ice sheets at low altitudes can thus probably reach the pressure melting point beneath blue ice areas even if the ice depth is moderate.

Field Surveys

Ground based surveys

The equipment was installed in the front cabin of a Hägglunds snow vehicle and the 150 MHz dipole antennaes with reflectors were mounted on a 6 m long pole fixed to the roof of the vessel. The dipoles were thus situated about 1 m above the ground. The sampling interval was generally 10 s (30 m). Positions were logged using the Trimble Pathfinder Professional system with time spacing equal to the sampling speed of the echo sounder. This constellation of equipment was used only in the Heimefrontfjella range for logistic reasons.

Fig. 3. Radio-echo record showing a subglacial mountain range southeast from Heimefrontfjella. The length of the profile is about 20 km and ice depths are given in metres. Note that internal layers are lacking. The profile was surveyed in January 1992.

Results

In both areas studied the radar recordings received show relatively strong echoes from the base of the ice sheet and few or no internal echoes. The signals received were generally strong and clear and at one site, beneath a supraglacial lake, the signals were very strong indicating the existence of melt water at the base. This supraglacial lake close to the nunatak Skansen has been described by Jonsson (1988) and a more full description is in preparation by Jonsson. Snow layering was not surveyed at Vestfjella but the blue ice clearly indicates a net loss of mass from the surface.

Eastern Heimefrontfjella was visited and surveyed in January 1992. The bed topography is mirrored by the wrinkles that can be observed in the satellite images (IfAG, 1988). The reflected signal was 10–20 dB stronger at this site in comparison with nearby areas to the east and to the south. East and south of the area internal layers were surveyed down to a depth of approximately 500 m. These internal layers were not visible in the study area (Fig. 3).

Discussion

In order to use radio-echo technique to distinguish wet-based areas from frozen-bed areas attention must be paid to the rate of scattering within the ice. Variation in the ability for radiowaves to penetrate the ice may give misleading information. There is a significant risk to overestimate the extension of wet based areas. However, data suggest changes of 10–20 dB in strength at 1500–2000 m to be caused by changes in the permittance of radiowaves within the ice.

In eastern Heimefrontfjella there is no reason to believe that the ice sheet is wet-based as the movement is slow and the temperature is low. Thus, relative strong bed echoes may indicate basal melting condition, low accumulation rate or both. In addition to accumulation measurements, radio-echo soundings may give us quantitative information on variations in accumulation rates if internal layers can be identified and mapped over large distances. Radio-echo soundings may also give qualitative views of the dynamics in the area studied as ice divides may show up as areas with low rates of internal scattering.

Conclusion

The areas with strongly reflecting spots, or strong reflection in general, occur beneath blue ice areas or areas with little mass turnover. A negative mass balance influences the flow and the thermal distribution within the ice. Blue ice at low altitudes, with high ablation rates may cause a significant warm-up of the ice body because of the vertical up-going movement.

However, at high altitudes where no nunataks are visible above the snow surface and where there is no blue ice, but a low rate of mass turnover, the echoes also get strong in comparison to neighbouring areas with higher accumulation rates. To differentiate wet-based areas from cold-based areas in radio-echo records true changes in the dielectric constant must be distinguished. In order to do this special attention must be paid to variations in the relative permittity within the ice.

Acknowledgements

This project was financed by the Swedish Natural Science Research Council (NFR) and the Department of Physical Geography at Stockholm University. The logistics were kindly supplied by the Swedish Polar Research Secretariat. I am very grateful to the Norwegian Council for Scientific and Industrial Research (NTNF-PFM) for stimulating collaboration and help with radar hardware and software. Thanks are also due to the Swedish Authority for Spent Nuclear Waste (SKN), Nord-Optik AB (NOAB), IBM and Verbatim.