Sensitivity of the Effective Viscosity of Temperate Ice to its Water Content

Abstract:

Speeds of ice streams depend on the shearing resistance of their margins. In West Antarctica large fractions of these margins consist of temperate ice with significant intergranular water. A set of shear experiments on ice at its melting temperature by Duval (1977, IASH, 118, 29-33) indicated that for water contents of ~0.01 to 0.8%, effective viscosity decreased by a factor of ~3. Recent modeling indicates that this sensitivity of viscosity to water content localizes strain in shear margins, causing shear heating that increases meltwater available for enhancing basal slip (Haseloff et al., 2019, JGR-ES, 124, 2521-2541).

To add to scant data on the effect of water on ice rheology, we are conducting experiments with a large ring-shear device modified to act as a viscometer. A ring of ice, with inner and outer diameters of 0.5 and 0.9 m, respectively, and a thickness of ~0.2 m, is sheared in confined compression. The ice ring is made by mixing deionized water and snow (crystal sizes of 1-5 mm) and maintained at the pressure-melting temperature by a temperature-controlled bath (0.01°C precision). The ice water content is varied by imposing various combinations of confining pressure (700-1500 kPa) and shear strain rate (7.1 x 10^-9 to 1.2 x 10^-7 s^-1). Experiments are conducted at a constant strain rate and only until peak shear stresses are attained to avoid the complicating influence of fabric development in tertiary creep. Water content is measured by recording the speed of an induced cold wave and solving the relevant Stefan problem.

Results indicate sensitivity of effective viscosity to water content three times larger than that measured by Duval, based on regression across the full range of water contents explored to date: 0.31-1.22%. This higher sensitivity may reflect, to some extent, that Duval’s experiments were conducted to tertiary creep. At water contents and strain rates below ~0.6% and 2.5 x 10^-8 s^-1, respectively, the viscosity sensitivity to water content is larger by a factor of nine than at larger water contents and strain rates. This effect was absent in Duval’s results, which involved strain rates only higher than 5 x 10^-8 s^-1, and may be significant because lower strain rates better reflect those of ice-stream margins. This effect may also point to a switch in deformation mechanisms in temperate ice at lower strain rates.