West Antarctic ice streams are separated from interstream ridges by lateral shear margins, which are typically several kilometres wide and heavily crevassed (Vornberger & Whillans, 1990). The presence of these features has been long recognized from field observations and satellite images. However, their significant influence on the force balance and dynamics of ice streams has been realized only within the last decade. Past models have made the assumption, common in glaciology, that the gravitational driving stress acting on ice streams is practically completely balanced by basal resistance to flow (e.g. Alley et al., 1989). A priori, there are at least two good reasons to think that such an assumption is reasonable. Firstly, these ice streams have a very large width-to-thickness ratio (10-100; Raymond, 2000; Tulaczyk et al., 2000b), which should favour a relatively small contribution of lateral margins to force balance. Secondly, ice in the shear margins is experiencing high strain rates and strains, which should lead to strong fabric alignment, shear heating and fracturing. All of these factors are expected to weaken ice, leading to an enhancement factor in the flow law of ice of up to ca. 10 (Echelmeyer et al., 1994). Consequently, most ice-stream models published in the early 1990s made the assumption of local stress balance, i.e. that local gravitational driving stress is fully balanced by local basal shear stress, td = tb (Lingle, 1984; Alley et al., 1987a, 1989, Fastook, 1987; Kamb, 1991; Lingle et al., 1991). Only the ice-shelf model of MacAyeal adapted to ice streams included the effect of shear margins on ice-stream force balance (MacAyeal, 1989a,b, 1992b).
The presumption that ice in the shear margins should exhibit a large enhancement factor has been undermined by work of Jackson & Kamb (1997), who reported a small enhancement factor (ca. 1-2) from laboratory studies on ice samples recovered from the southern shear margin of Whillans Ice Stream (WIS) in West Antarctica. This is in spite of the fact that at this location the expected total strain accumulation in the shear-margin ice is ca. 20 strains. Measurements of ice-temperature profiles across the same shear margin have demonstrated that a significant portion of the gravitational energy dissipated during motion of this ice stream is expanded in marginal shear heating (Harrison et al., 1998). The latter study indicated that over the past 50yr the marginal shear averaged 200 kPa there, in good agreement with the independent estimate of Jackson & Kamb (1997) of 220 ± 30kPa coming from their laboratory tests on ice sampled in the same study area. Other observational studies indicate that marginal shear stresses may support ca. 50-100% of the gravitational driving stress in ice streams and their tributaries (Price et al., 2002; Joughin et al., 2004).
These findings corroborated earlier inferences of Whillans & van der Veen (1993), who calculated that marginal shear stress supports almost all of the driving stress on Whillans Ice Stream. In a subsequent publication, the same authors (Whillans & van der Veen, 2001) analysed the details of stress transfer from the shear margin to a several-kilometres-wide basal zone outside of the ice stream. They have shown that the lateral drag is the highest in the upper 40% of ice thickness, which is colder and more viscous than the lower part of the ice column. The transfer of marginal stress to the ice base outside of the ice stream leads to a peculiar situation, in which ice is propelled forward by stress much greater than the local gravitational driving stress that can be calculated from ice thickness and surface slope. In fact, the magnitude of stress transferred from the marginal shear zone to the bed just outside of the ice stream, ca. 60kPa, exceeds by a factor of about three to five times the magnitude of the local gravitational driving stress at the site analysed by Whillans & van der Veen (2001). The unusual force balance around ice-stream shear margins is of great significance because:
1 a large fraction of ice-stream velocity is attained outside of the highly lubricated bed of an ice stream;
2 this force balance helps control the position of shear margins and, thus, the velocity of ice-stream motion and its changes (Jacobson & Raymond, 1998).
The location of shear margins of ice streams can be either fixed by subglacial conditions external to ice thermodynamics, e.g. topography and geology, or it can evolve with the thermodynamic state of the ice sheet itself. Anandakrishnan et al. (1998) and Bell et al. (1998) published results of geophysical studies indicating that, at least in one location, a tributary of Kamb Ice Stream (KIS) has a margin located over a boundary of a subglacial sedimentary basin. This observation provides evidence for the possibility of geological control over ice-stream geometry. However, extensive geophysical studies in the trunk regions of KIS and WIS indicate that there the boundary between slow and fast moving ice corresponds to a transition between frozen and melted bed conditions (Bentley et al., 1998; Gades et al., 2000; Kamb, 2001). There may be a difference in the geological setting of the main ice-stream trunks and the tributary region, which is in general much more dissected by deeply incised subglacial valleys as compared with the gentle topographic troughs associated with ice-stream trunks (Blankenship et al., 2001; Studinger et al., 2001). Where geological control is not the predominant factor, the position of shear margins may evolve through time in response to changes in force balance and thermal energy dissipation in and around the margins (Jacobson & Raymond, 1998; Raymond, 2000; Whillans & van der Veen, 2001). Measurements indicate that in the recent past (ca. 10-100yr) at least some sections of ice-stream shear margins migrated outward at rates of the order of 1-10 myr-1 (Harrison et al., 1998; Echelmeyer & Harrison, 1999). In one location, inward migration of the southern shear margin of WIS has beeb inferred to have taken place at the rate of ca. 100myr-1 within the past few hundred years (Clarke et al., 2000).
Was this article helpful?