Modelling and observing subglacial processes beneath Antarctic ice streams
Ice sheets are drained by fast flowing glaciers that discharge large quantities of ice into polar oceans. Changes in the motion of fast flowing glaciers, such as ice streams and major outlet glaciers, can therefore affect the mass balance of ice sheets. Glaciological research has shown that the flow of ice streams is controlled by material properties of weak subglacial sediment. Many glaciers, particularly the fast flowing ones, overrides a thin layer of glacially eroded sediment called till. The soil mechanical behaviour of subglacial till is therefore a key glaciological issue.
Figure 1. Velocity map for Ice Stream A-E in West Antarctica (Courtesy of Ian Joughin, Univ. Washington).
Figure 2. Re-fuelling snowmobiles during a traverse from Siple Dome to UpC camp on Kamb Ice Stream.
Even small property changes in the till layer beneath ice streams can have large effects on ice dynamics. Basal melting followed by build-up of pore water pressure can lead to accelerated ice velocities. Contrary, basal freezing can lead to impeded ice flow due to strengthening of subglacial sediments. In fact, cryostatic dewatering of subglacial sediment can cause ice streams to stop entirely because basal freezing can become a run-away process. The best example of abrupt ice stream stoppage is seen in West Antarctica where Kamb Ice Stream (also known as Ice Stream C) no longer flows at the enhanced velocity that characterises ice streams (Figures 1 and 2). The depth of buried crevasses that originated from the period of fast flow indicates that the stoppage took place approximately 150 years ago.
Figure 3. Schematic diagram illustrating the coupled flow of water, heat and solutes in a freezing subglacial till layer.
Dr Christoffersen's research on the dynamic history of Kamb Ice Stream started as a graduate student at the Technical University of Denmark and visiting scholar at University of California, Santa Cruz. Together with Slawek Tulaczyk at UCSC, a numerical ice-flow/till-layer model was developed that provided a quantitative explanation for the stoppage of Kamb Ice Stream. The model used principles from frost-heave theory to simulate the coupled flow of water, heat and solutes induced in the subglacial till layer by depression of the freezing point due to ice-water interfacial effects (Figure 3). The model showed that a till porosity reduction of a few percent may be sufficient to prevent fast ice-streaming flow and that ice streams may shut down after approximately 100 years of basal freezing (Figure 4).
Figure 4. Diagrams with model results showing (left) decrease in ice velocity during the stoppage of an stream, (middle) increase in the freezing rate from loss of frictional heat, and (right) increase in basal shear strength due to dewatering of the till layer.
The distribution and flow of water, heat, and solutes in a subglacial till-layer model are coupled to each other through the use of the Clapeyron’s equation, which makes the modelled ice-water phase-change temperature rely on pressure, solute concentration and interfacial curvature. Having examined the response of subglacial sediments to basal freeze-on, the model was used to evaluate the acquisition of subglacial sediment by fast flowing ice streams. These model results showed that the till layer underlying a freezing ice base experiences characteristic changes in vertical distribution of porosity and pore-water composition (Figure 5).
Figure 4. Depth-time diagrams showing modelled changes in the porosity (%) of a 5-m-thick till layer and formation of segregated ice lenses from downward progression of a freezing front.
A borehole camera system developed at the Jet Propulsion Laboratory was in 2000-01 used to image the composition of ice Kamb Ice Stream, West Antarctica. Hence, we now know that Antarctic ice streams contain up to 15-m-thick layers of accreted basal ice. The debris content of the basal ice layer has a high spatial variability. The upper c. 90 % of the basal ice layer in Kamb Ice Stream is composed of multiple layers of several-m-thick clean accretion ice. The clean ice layers are separated by layers of banded ice facies or ice facies with inclusions of dispersed sediment. The lower c. 10 % is composed by debris-rich basal ice similar to frozen sediment (Figure 6).
Figure 5. Borehole camera images of the basal ice layer in Kamb Ice Stream. From left to right: (a) a banded sequence of clean ice and bubbly ice, (b) laminated basal ice, (c) massive basal ice with dispersed debris, and (d) solid debris-rich basal ice. Images are approximately 4 cm in the vertical.
To better understand the ice accretion processes, Dr Christoffersen developed a numerical model of the sediment entrainment process. The model can be used to predict accretion ice facies associated with different subglacial hydrological conditions. Clear accretion ice with little or no debris forms when inflow of subglacial water equals the freezing rate. Fine bands of debris regelate into basal ice if the supply of subglacial water becomes restricted. Alternatively, the debris-bearing basal ice develops a loose framework of uniformly distributed sediment. Debris-rich basal ice develops when the supply of subglacial water cannot satisfy the basal heat budget. Freezing becomes a run-away process that relies on continuous extraction of till pore water. Ice lenses and debris in stratified arrangement develop in fine-grained subglacial sediments when run-away freezing has been triggered (Figure 7).
Figure 7. Schematic diagrams illustrating (left) arrangement of debris in different basal ice facies observed beneath the West Antarctic Ice Sheet, and (right) subglacial conditions predicted by numerical modelling. Roman numerals (I-VI) and letters (a-d) correspond to the image series shown in Figure 5.
Subglacial accretion processes are an important because the distribution and variability of debris in basal ice contain a history of pre-existing subglacial conditions. A better knowledge of subglacial hydrology is crucial with respect to understanding ice stream dynamics and ice sheet evolution.
Figure 8. GPS measurement of stake on Kamb Ice Stream
- Bougamont, M., S. Tulaczyk, and I. Joughin, 2003a, Response of subglacial sediments to basal freeze-on: Application in numerical modelling of the recent stoppage of Ice Stream C, West Antarctica, Journal of Geophysical Research, 108(B4), 2223, doi:10.1029/2002JB001936.
- Bougamont, M., S. Tulaczyk, and I. Joughin, 2003b, Numerical investigations of the slow-down of Whillans Ice Stream, West Antarctica: is it shutting down like Ice Stream C?, Annals of Glaciology, 37, 239-246.
- Christoffersen, P., S. Tulaczyk, F. Carsey, and A. Behar, 2006. A quantitative framework for interpretation of basal ice facies formed by ice accretion over subglacial sediment, Journal of Geophysical Research, F01017, doi:10.1029/2005JF000363.
- Christoffersen, P., and S. Tulaczyk, 2003a. Response of subglacial sediments to basal freeze-on: I. Theory and comparison to observations from beneath the West Antarctic Ice Sheet, Journal of Geophysical Research, 108(B4), 2222, doi:10.1029/2002JB001935.
- Christoffersen, P., and S. Tulaczyk, 2003b. Thermodynamics of basal freeze-on: predicting basal and subglacial signatures beneath stopped ice streams and interstream ridges, Annals of Glaciology, 36, 233-243.
- Christoffersen, P., and S. Tulaczyk, 2003c. Signature of palaeo-ice stream stoppage: till consolidation induced by basal freeze-on, Boreas, 32(1), 114-129.
- Tulaczyk, S., Kamb, W. B. and Engelhardt, H. F., 2000a, Basal mechanics of Ice Stream B, West Antarctica 1. Till mechanics, Journal of Geophysical Research, 105(B1), 463-481.
- Tulaczyk, S., Kamb, W. B. and Engelhardt, H. F., 2000b, Basal mechanics of Ice Stream B, West Antarctica 2. Undrained plastic bed model, Journal of Geophysical Research, 105(B1), 483-494.
- Vogel, S.W.. S. Tulaczyk, B. Kamb, H. Engelhardt, F. D. Carsey, A. E. Behar, L. Lane, and I. Joughin, 2005, Subglacial conditions during and after stoppage of an Antarctic Ice Stream: Is reactivation imminent?, Geophysical Research Letters, 32, L14502, doi:10.1029/2005GL022563.