3.1 Temperature

The mean annual temperature was ₋2.9°C at Bellingshausen Station between 1968 and 1983 (Reference Jacka,, Christon, and Cook,Jacka and others, 1984). and ₋2.7°C at the Great Wall Station on King-George Island between 1985 and 1988. From these and, assuming an adiabatic lapse rate between 0.7° and 1.0°C/100 m, it is estimated that the annual mean air temperature is between ₋5°C and ₋6°C at the summit of the Nelson Island ice cap. During the 1987-88 austral summer, a layer of water-saturated firn was encountered at a depth of 25-26 m in a 30 m core drilled at the summit (GW). A similar layer was found at a depth of 15 m in a 20 m core obtained at site N30. Ihe temperature measured immediately after drilling was 0°C at all depths in these boreholes. Stratigraphic profiles demonstrated that the water layer was at the base of the firn in both cases.

Fig. 2

Some, features of the ice cap on. Nelson Island along transect N. a. Ice-surface and bed elevations. b. Surface-velocity components: u is the horizontal component, positive toward the margin, and w is the vertical component. positive upwards. c. Ratio of basal velocity (u_b) to surface velocity (u_s). d. Isolines of effective shear stress (T_e).

In the 1985-86 summer, three boreholes were drilled, especially for temperature measurements, by using a steam drill. The holes were at 90, 140 and 160 m a.s.l., and were 13, 11.8 and 12.5 m deep, respectively. The average temperature at 10m depth was ₋1.5°C at 90 m, ₋0.7°C: at 140 m and ₋0.1°C at 160 m (Reference Jiawen,Ren, 1990). Obviously, the temperature regime is controlled by the percolation of water from the surface, mainly from surface melting but also from liquid precipitation during the summer. Where the snow-firn layer is thicker, water penetrates deeper and the 10m temperature is higher. Conversely, near the edge of the ice cap, where the snow-firn layer is thinner or absent, most water is lost by runoff. The temperature at depth is undoubtedly at the melting point. This temperature distribution is in striking contrast to that of cold glaciers in polar regions and of glaciers in middle or low latitudes, but is well known from glaciers in similar climates.

3.2 Surface velocity and dynamics

From January 1988 to February 1989, coordinates of 40 stakes, set along transects N and E, were measured three times and the surface velocity at each site was calculated (Reference Chen, Dahe, Xiaojun, and SonglinLiu and others. 1991). The velocity increases toward the margin with a summer maximum of 15.5 m a^₋1 and an annual maximum of 15.3 m a^₋1 at N50. (Figure 2b) illustrates the horizontal (u) and the vertical (W) components along transect N. Both horizontal and vertical velocities increase sharply at about 4 km from GW. This is consistent with the change in surface slope mentioned above. Near the margin, the horizontal velocity decreases but the vertical velocity continues to increase (becomes more negative).

Many studies of the dynamics of large ice sheets have been reviewed by Reference Paterson,Paterson (1981). In previous studies, one of the basic assumptions is often that, on a large scale, the bedrock is horizontal. Locally, on an ice sheet or small ice cap,this assumption is invalid.

Reference Van der Veem, and Whillans,Van der Veen and Whillans 1989 force-budget technique was used to develop a two-dimensional dynamic model along transect N. Using this method, forces acting on an ice mass are divided into the driving stress and the resisting stresses, and the velocity field is coupled to the stress field by a constitutive relation (ice-flow law). Thus, an analysis of stresses within an ice mass can be undertaken only if the surface velocity and the geometry of the ice mass are known. (Figure 2c) illustrates the resulting computed ratio of basal (U_b) and surface velocities U_s). Except for a few sites, basal velocity is greater than 50% of the surface velocity, implying that most of the ice movement is by basal sliding.

Figure 2d shows isolines of effective shear stress τ_e in the section down-glacier from the 3 km (from GW) position. Between GW and the 3 km position, τ_e varies very little and is less than 40 kPa. However, it changes significantly over a short distance near the 4 km position, reaching a maximum in excess of 400 kPa. Although this is the region of steepest surface slope, so large a value is questionable. Transect N does not coincide with a flow line, especially between N30 and N50, where there is a noticeable lateral velocity, but our two-dimensional model assumes the lateral velocity is zero. This certainly causes some error in the result.

In our calculation, the Glen-Nye flow law

was used as the constitutive relation. Here, is the effective strain rate, and A and n are constants. Substitution of the measured surface velocities in the model yields values of n and A of 1.1 and 5.01 × 10⁻⁵kPa^−1.1 a⁻¹, respectively. The low n and high A imply that the ice. although almost linear in its response to stress, has a rather high viscosity.

4.1 Stratigraphy

During the summers of 1985-86, 1987-88 and 1988-89, 54 shallow pits, up to 2 m deep were dug for measurement of stratigraphic profiles. Most pits Were on transect Ν and in some cases pits were dug at the same site in two or more years. Unfortunately, the profiles do not show clear seasonal features sufficiently consistent to be useful.

Figure 3 shows stratigraphic, stable-isotope and density profiles from three cores drilled at N50, Ν30 and GW (from Reference Dahe, Zielinski,, Germain,, Jiawen,, Xiaoxiang, and Wenti,Qin and others, 1994). There are no noticeable seasonal variations in the isotope profiles; this is certainly a consequence of the intense percolation of surface water. The stratigraphic and density profiles indicate that the firn-ice transition is at a depth of 25-26 m at GW and at about 15 m at N30. This is a little deeper than reported by Reference Orheim, and Govorukha,Orheim and Govorukha (1982) on King George Island.

4.2 Accumulation rate

At most measurement sites, we have snow-accumulation data spanning at most 2 months. This is because of the loss of stakes during the winter. Data spanning one mass-balance year (February 1988-February 1989) at N30, N40 and N50 yield accumulation rates of 790, 1050 and 110 kg m⁻² a⁻¹, respectively. However, a clear ash layer was found in all three cores (Fig.3. Element-composition analysis confirmed that all three ash layers are from the 1970 eruption of the Deception Island volcano (Reference Dahe, Zielinski,, Germain,, Jiawen,, Xiaoxiang, and Wenti,Qin and others, 1994). From this, we obtained average accumulation rates between 1970 and 1988 of 1200 kg m⁻² a⁻¹ at GW, 700 m⁻² a⁻¹ at N30 and 6 kg m m⁻² a⁻¹ at N50, or less than indicated by the stake data. The relatively high accumulation at N40 (stake data) is a result of snowdrift in a location where the surface is somewhat concave. Three other 10-15 m cores were taken in the eastern part of the ice cap (at E30, E55 and E95) but no ash layer was found in them.

The equilibrium-line elevation was determined from pit and stake data to be 110 m a.s.l. between 1985 and 1989, with a relatively high value in 1985-86 and a low value in 1988-89. Several snow pits dug in the summer of 1985-86 show that the equilibrium-line elevation is about 150 m on King George Island (Reference JiawenRen, 1988). This is consistent with equilibrium-line elevations in the early 1970s reported by Reference Orheim, and Govorukha,Orheim and Govorukha (1982). That the 10 m temperature at 150)m a.s.l. on King George Island is close to that Of 90m a.s.l. on the Nelson Island ice cap (Reference Jiawen,Ren, 1990) also implies a higher elevation of the on King George Island.