Within the McMurdo Dry Valleys are a number of perennially ice-covered lakes supporting microbial life both in the liquid water beneath the ice and in association with a sediment layer in the ice itself (Wing and Priscu 1993; Lizotte, Sharp, and Priscu in press; Priscu 1995). A dominant environmental factor, the ice cover transmits visible radiation, absorbs infrared radiation, and shields the underlying liquid from wind-induced convection. The ice covers are 3 to 6 meters (m) thick and are maintained by a 30-centimeter (cm) annual accretion on the bottom, at the liquid/solid interface, coupled with ablation at the top.
Continuous ice and air temperatures for 1993 were recorded at Lake Bonney. Ice cores were examined for morphology and to measure absorption coefficients [12 wavelengths: 410 to 694 nanometers (nm)].
Figure 1 reveals that the entire (approximately 4-m) ice cover was isothermal at the melt temperature (0oC) for almost 3 months, and that once the freezing process began, it was approximately 4 months before the frost reached the 3.5 m thermistor. It would likely take almost another month to freeze to the bottom. A temperature gradient is required at the solid/liquid interface for active freezing, implying that growth at the bottom takes place for approximately 4 months. The average temperature gradients (figure 2) show significant near-surface fluctuations, becoming more stable with depth.
Absorption coefficients for a 30 November 1994 Lake Bonney ice core indicate little variation with wavelength but an order of magnitude difference between the top 30 cm and the region from 90 to 155 cm (figure 3).
Reference is made in the literature on the lake ice to long, vertically oriented C-axis crystals (Wilson 1981); however, we were unable to find a published crystallographic study. Findings from our work concur with the earlier assertion that the ice cover from Lake Bonney is vertically C-axis oriented and may thus be classified as S1 (Michel and Ramseier 1971). This has a slight influence on heat transfer, since the thermal conductivity is about 5 percent greater along the C-axis than it is perpendicular to it (see Hobbs 1974).
It has been suggested (e.g., Wilson 1981) that the vertical C-axis crystals serve as "light pipes." The basis for this assertion is unclear! C is the optic axis in ice; however, studies by Lyons and Stoiber (see Hobbs 1974) indicate that clear polycrystalline ice and singel ice crystals have essentially the same absorption coefficients and that no difference in absorption of the ordinary and extraordinary waves is evident. In the present study, examination of thin sections of Lake Bonney ice under cross polarized light demonstrates that the bubbles do not coincide with grain boundaries to thus channel light through the crystal.
Bubbles are effective in scattering light; hence, the morphology is important to thermal and biologic processes related to the lakes. Bubbles in the upper 20-30 cm of the ice in Lake Bonney were heavily frosted on their interior by early November (when observations began). The influence of the heavily frosted bubbles combined with internal radiation etching is apparent in the absorption coefficients (figure 3). In the ice, from 40 cm extending downward to the sediment layer (2 to 2.5 m), bubbles were distinctly frosted only on the upper surv]face. A strong, persistent temperature gradient (figure 2) across a vapor pore will induce a mass transfer of water vapor by sublimation/condensation from the warmer to the colder surface to produce hoar frost.
Pinnate and tear-drop-shaped bubbles, tapered end downward, appear near and above the sand layer. Tear-drop and pinnate shapes have been observed in second-year arctic lake ice that had gone through a refreeze (Swinzow 1966), and tear-drop bubbles and chains found within the upper half of the antarctic lake ice are also similar in shape and orientation to some of those observed in newly formed laboratory ice produced by downward freezing. Extending from the sediment layer to the bottom, vertically oriented, essentially cylindrical vapor bubbles are the prevailing structure, the product of gas exsolution and occlusion during accretion. Craig, Wharton, and McKay (1992) have analyzed gas in bubbles thought to be formed at the liquid/solid interface and carried to the top surface in the encapsulating ice where it is removed by ablation. Through this region, layers of ice 10 to 25 cm long with a high concentration of cylindrical bubbles are interspersed with relatively clear ice. Size, density, and shape of bubbles are to a large degree a function of ice growth rate. For example, at very slow growth rates into a large reservoir, diffusion of gas into the liquid will keep the supersaturation at the solid/liquid interface below the critical level for bubble formation.
A core taken on 17 November 1994 had liquid in a sand inclusion 250 cm below the ice surface (possibly evidence of radiation-induced melt) but was otherwise dry. Two days later, liquid-saturated ice was encountered approximately 100 cm below a topgraphic depression in the ice. The hole slowly (45 minutes) filled with fluid to the hydrostatic lake level, 25 cm below the surface. On the same day, 7 m away in an ice ridge, liquid water was encountered at a depth of 250 cm and filled only the lower 15 cm of the hole. In all subsequent bores (through 8 December) on Lakes Bonney and Vanda, the liquid-saturated ice was between 70 to 100 cm from the top of the ice and quickly rose to the hydrostatic lake level (20 to 45 cm depending on topography). A early as 1966, Henderson et al. (1966) noted that in the summer, a liquid water table developed within the ice cover. The fact that liquid rises to the hydrostatic level of the lake after the water table is penetrated implies that the water table and main lake fluid are connected. Then, based on the temperature data (figure 1), annual atmospheric cooling causes the water table to drop as the liquid refreezes downward.
It seems likely that liquid water infiltration of the ice would induce melt, fill some cracks and bubbles, and upon freezing, cause exsolution of dissolved gas from the liquid to create new bubbles. The annual infiltration and refreezing of liquid water in the ice cover must play a major role in metamorphism.
This work was supported by National Science Foundation grants OPP 91-17909, OPP 92-11773, and OPP 94-19423 to John C. Priscu.
References
Craig, H., R.A. Wharton, and C.P. McKay. 1992. Oxygen supersaturation in ice covered antarctic lakes: Biological versus physical contributions. Science, 255, 318-321.
Henderson, R.A., W.M. Prebble, R.A. Hoare, K.B. Popplewell, D.A. House, and A.T. Wilson. 1966. Journal of Glaciology, 6, 129-133.
Hobbs, P.V. 1974. Ice physics. Oxford: Oxford University Press.
Lizotte, M.P., T.J. Sharp, and J.C. Priscu. In press. Phytoplankton dynamics in the stratified water column of Lake Bonney, Antarctica: I. Biomass and productivity during the winter-spring transition. Polar Biology.
Michel, B., and R.O. Ramseier. 1971. Classification of river and lake ice. Canadian Geotechnical Journal, 8(36), 36-45.
Priscu, J.C. 1995. Phytoplankton nutrient deficiency in lakes of the McMurdo Dry Valleys, Antarctica. Freshwater Biology, 34, 215-227.
Swinzow, G.K. 1966. Ice cover of an arctic proglacial lake. (U.S. Army Cold Regions Research and Engineering Laboratory, Research Report 155). Hanover, N.H.: USACRREL.
Wilson, A.T. 1981. A review of the geochemistry and lake physics of the antarctic dry areas. In L.D. McGinnis (Ed.), Dry Valley Drilling Project (Antarctic Research Series, Vol. 33). Washington, D.C.: American Geophysical Union.
Wing, K.T., and J.C. Priscu. 1993. Microbial communities in the permanent ice cap of Lake Bonney, Antarctica: Relationships among chlorophyll-a, gravel and nutrients. Antarctic Journal of the U.S., 28(5), 246-249.