Antarctic sea ice is not confined by land margins but is open to the Southern Ocean. Sea-ice extent contracts and expands on an annual cycle in a roughly concentric zone around Antarctica. The ultimate extent is controlled by a balance of air temperature, leads, wind direction, upper ocean structure, and pycnocline depth. Some of these parameters are controlled in the atmosphere by the relative position of the subpolar trough with respect to the sea ice. In the ocean, variations in the Antarctic Circumpolar Current are important. The extent and thickness of Antarctic sea ice are sensitive to the depth and thermal properties of overlying snow, about which relatively little is known.
A reduction in Antarctic sea ice volume of about 25-45% is predicted for a doubling of CO2, with sea ice retreating fairly evenly around the continent (Gordon and O'Farrell, 1997). This CSIRO model assumes a 1% yr-1 compounding increase of CO2, corresponding to global warming of 2.1°C. Using a similar but modified model that has a higher albedo feedback and predicted global warming of 2.8°C, Wu et al. (1999) calculate a reduction in mean sea-ice extent of nearly two degrees of latitude, corresponding to 45% of sea-ice volume. These estimates do not represent the equilibrium state, and sea ice can be expected to shrink further, even if GHGs are stabilized. Changes in Antarctic sea ice will have little impact on human activity except where they allow shipping (mostly research, fishing, and tourist vessels) to get closer to the Antarctic continent. However, there are important biological and oceanographic impacts derived from reductions in sea ice, as well as significant ecological consequences attributable to changes in the magnitude and timing of seasonal sea-ice advance and retreat (Smith et al., 1999b).
The term permafrost refers to layers of earth materials at relatively shallow
depth that remain below 0°C throughout 2 or more consecutive years, independent
of material composition or ice content. The role of permafrost in the context
of global change is three-fold (Nelson et al., 1993): It records temperature
changes (Lachenbruch and Marshall, 1986), translates changes to other components
of the environment (Osterkamp et al., 2000), and facilitates further climatic
changes through release of trace gases (Rivkin, 1998; Robinson and Moore, 1999).
Permafrost is highly susceptible to long-term warming. Although it is an important
factor in ice-free parts of Antarctica (Bockheim, 1995), permafrost is far more
extensive in the continental areas of the subarctic and Arctic (Brown et al.,
1997; Zhang et al., 1999), and changes there have great potential to affect
human activities adversely. Accordingly, frozen ground activity has been designated
as a "geoindicator" for monitoring and assessing environmental change
(Berger and Iams, 1996).
Most biological and hydrological processes in permafrost terrain are confined
to the active (seasonally thawed) layer, which forms a boundary across which
exchanges of heat, moisture, and gases occur between the atmospheric and terrestrial
systems. Its thickness is a response to a complex interplay among several factors,
including aboveground climate, vegetation type and density, snow-cover properties,
thermal properties of the substrate, and moisture content. If other conditions
remain constant, the thickness of the active layer could be expected to increase
in response to climate warming. Long-term records of active-layer thickness
are rare; those that do exist, however, indicate that variations occur at multiple
temporal scales and are determined by climatic trends and local conditions at
the surface and in the substrate (Nelson et al., 1998b).
Because heat transfer within thick permafrost occurs primarily by conduction,
the shallow earth acts as a low-pass filter and "remembers" past temperatures.
Temperature trends are recorded in the temperature-depth profile over time scales
of a century or more (e.g., Lachenbruch and Marshall, 1986; Clow et al., 1998;
Osterkamp et al., 1998; Taylor and Burgess, 1998). Permafrost is affected primarily
by long-term temperature changes and thus contains a valuable archive of climate
change (Lachenbruch and Marshall, 1986), although changes in snow cover must
be taken into account. In contrast, the temperature regime of the overlying
seasonally thawed active layer is highly complex, owing to nonconductive heat-transfer
processes that operate much of the year (Hinkel et al., 1997).
Multi-decadal increases in permafrost temperature have been reported from many locations in the Arctic, including northern and central Alaska (Lachenbruch and Marshall, 1986; Osterkamp and Romanovsky, 1999), northwestern Canada (Majoriwicz and Skinner, 1997), and Siberia (Pavlov, 1996). Temperature increases are not uniform, however, and recent cooling of permafrost has been reported in northern Quebec (Wang and Allard, 1995). To obtain a more comprehensive picture of the spatial structure and variability of long-term changes in permafrost, the Global Terrestrial Network for Permafrost or (GTNet-P) (Brown et al., 2000; Burgess et al., 2000b) was developed in the 1990s. The program has two components: long-term monitoring of the thermal state of permafrost in a network of boreholes, and monitoring of active-layer thickness and processes at representative locations. The latter networkthe Circumpolar Active Layer Monitoring (CALM) program (Nelson and Brown, 1997)has been in operation since the mid-1990s and incorporates more than 80 stations in the Arctic and Antarctic. The CALM network includes components in which regional mapping (Nelson et al., 1999) and spatial time series analysis of active-layer thickness (Nelson et al., 1998a) are performed.
Other reports in this collection