Abstract
Many ice flow models treat polar ice as chemically pure and mechanically isotropic, while in reality it can have pre-existing crystallographic preferred orientations, soluble impurities, and mechanical heterogeneities which affect its behaviour. As a result, our predictions of ice flow behaviour can be inaccurate. We performed fixed-rate uniaxial compression experiments on both natural Antarctic ice and laboratory-made ice containing natural soluble impurities, alongside samples of pure, initially isotropic ice for comparison. In addition, we performed shear experiments on pure laboratory ice, and developed an entirely novel method of analysing changing local shear directions across a plane of a sample. All deformed samples were analysed using cryo- Electron Backscatter Diffraction to produce robust microstructural data.
The results show that all ice containing soluble impurities, regardless of microstructure, behaves more weakly than pure ice, with a flow law stress exponent of n≈4 rather than the commonly assumed value of n≈3. Additionally, in natural ice strain is accommodated heterogeneously, as some grains fully recrystallise within strains of <0.2, while others do not recrystallise and show little evidence of strain accumulation. This heterogeneity is only seen in natural ice, and is therefore assumed to be a result of its larger grain size or irregular impurity distribution. Kinematic analysis of sheared pure ice samples shows that local shear directions can vary significantly relative to the imposed shear direction, without microstructure being significantly affected. The analytic methods developed for kinematic analysis may be applied to other materials.