JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 107, NO. B1, 10.1029/2000JB000058, 2002
|
|
[5] The West Antarctic Ice Sheet is a marine ice sheet: The underlying bedrock is below sea level and would probably remain so even after postglacial rebound [Doake, 1987]. The ice sheet drains mainly through the Ross and Filchner-Ronne Ice Shelves, which float on the ocean. The ice shelves are large in areal extent and receive substantially higher precipitation than the Antarctic interior; seawater may freeze and accrete onto the base. Shelf ice can be broadly categorized into three layers:
The top 50 m consists of low-density firn, which is of atmospheric origin. It extends down to where the ice becomes nonpermeable, usually taken to be a density of ~830 kg m-3, the close-off density for pores. Firn ice is cold and by implication brittle and is often heavily crevassed [Rist et al., 1999].
Below the firn layer, there is older bubbly ice of atmospheric origin that has become consolidated with time and depth under the weight of overlying ice. The lower part of this layer can be ice that has flowed off the continental ice sheet and therefore may have undergone very large strains.
Ice of marine origin forms the basal layer (below ~160 m) for a large area of the Ronne Ice Shelf (bounded by the dashed line in Figure 1), accreted onto the bottom of the meteoric ice [Oerter et al., 1992b]. Remarkably, this ice is clear due to the lack of air bubbles and has low salinity compared to sea ice. In the center of the Ronne Ice Shelf, marine ice can make up more than half the ice shelf thickness. Marine ice forms downstream of Korff and Henry ice rises by accretion of frazil ice [Bombosch and Jenkins, 1995] and then gradually melts toward the ice front.
|
|
[6] For our experimental program, shelf ice was gathered from three core-drilling sites on the Ronne Ice Shelf (B13/B14 and B15) drilled in 1989/1990 and 1991/1992, and from Berkner Island (B24) drilled in 1994/1995 (Figure 1 and Table 1). Boreholes B13 and B15 were drilled to depths of 215 and 321 m, respectively, deep enough to sample the marine ice accreted to the base of the shelf. Figure 2 shows a composite profile of physical properties measured mainly on our mechanical test specimens for B13, B14, B15, and B24, together with thin sections and Schmidt net plots of caxis ice fabric diagrams. Our purpose in producing a composite profile is to demonstrate the small spatial variability but the strong depth dependence of physical properties important for a fracture model. The profile also shows that our test specimens were representative samples as the measurements conform closely to other published data on these cores [Oerter et al., 1992a; Eicken et al., 1994; Oerter et al., 1994; Gerland et al., 1994].
[7] The temperature profile shown in Figure 2 was measured through the ice shelf at B15. Temperature in the first few meters is highly dependent on the ambient surface temperature and is not shown. The overall profile is S-shaped, which is typical for a bottom accumulation environment [Oerter et al., 1992a]. Temperature profiles for bottom melting regions, such as at B13, are typically cubic.
[8] Ice density was determined from specimen mass and volume before testing. These measurements conform closely to those taken immediately after coring and to those determined by continuous gamma ray absorption measurements [Oerter et al., 1994; Gerland et al., 1994]. The entire density profile is described by an exponential function:
where d is the depth below the upper surface in meters. Certain scatter in these data (Figure 2) occurs because the firn-ice transition for B24 on Berkner Island is slightly deeper than on the ice shelf.[9] The dynamic elastic modulus, also shown in Figure 2, was derived from acoustic velocity measurements on the cylindrical core specimens before mechanical testing. The elastic modulus profile closely corresponds to the density profile, as density has the largest influence on specimen stiffness.
[10] A selection of specimen thin sections (80 mm diameter) taken at various depths viewed under crossed polarizers are also shown in Figure 2. Grain size was determined from the mean grain area in thin section, assuming spherical grains. In the meteoric layer the thin sections show strongly increasing grain size with depth, decreasing porosity, and change of grain shape from equiangular to irregular and elongated grains. Grain size increases roughly linearly through the meteoric ice layer from ~1 to 9 mm. The associated Schmidt net plots show c axis orientations for ice fabric development from shallow, randomly oriented ice, to deeper ice with some fabric development of preferred crystallographic orientation in the form of a small circle girdle, centered on the vertical axis. This is a consequence of extensional stress along the flow line and little or no compression normal to the flow line [Eicken et al., 1994]. Toward the base of the meteoric ice, there is evidence of dynamic recrystallization. The marine ice exhibits highly varied grain size and strings of particulate inclusions, likely to be from bottom sediments [Oerter et al., 1992b]. The transition from meteoric to marine ice is characterized by a sharp reduction in grain size, and within the marine ice, there is an apparent reverse age/grain size relationship. These can be explained by the thermal history of the marine ice [Eicken et al., 1994]. The Schmidt net plots for marine ice specimens also show some fabric development, although this is not as strong as in the deep meteoric ice.

Citation: Fracture of Antarctic shelf ice, J. Geophys. Res., 107(B1), 10.1029/2000JB000058, 2002.