PERMEABILITY, WETNESS AND CREEP OF TEMPERATE ICE

Abstract
Ice is said to be temperate when it is in local equilibrium with water veins and water inclusions (1,2). Temperate ice remains a poorly understood material offering a large field for investigation. Pressure solution of air in percolating water explains why, in temperate glaciers, ice becomes progressively bubble-free between 100 and 200 m in depth (3). Emptying of large air bubbles (formed by coalescence of several hundred of small ones) through capillary veins is not excluded, however. Permeability of glacier ice suffering strain, grain boundary migration and recrystallization does not follow Darcy's law, and evolves with time, as shown by laboratory and field observations (4,5). Moreover, the temperature field depends on the local stress, rather than on Fourier's law. Therefore, the classical theory of glacier sliding (6), which assumes ice to be dry and impermeable, is unrealistic. Most water veins and lenses should be more or less normal to the principal direction corresponding to the highest compressive stress. With this model, tentative governing equations for temperature and for water flux are suggested. They yield surprising results (7) ; if ice were Newtonian viscous, the flux of cold reaching the interface would be much higher than needed for refreezing all the exuding water. Field measurements have shown that the liquid water content in a temperate Alpine glacier fluctuates widely at the decimetre scale, without general trend from surface to bottom (8). The variations are completely uncorrelated with grain size or salinity. An explanation is offered, which considers water lenses at grain boundaries emptying into the water veins when they have grown enough. Tertiary creep of temperate ice, with a multimaxima fabric, is assumed to follow an isotropic third-power viscous law, even when it is wet. The lowering of viscosity for an increasing water content (9) is understandable if processes at the grain boundaries are thought to be the main factor controlling creep, contrarily to usual theory. Grain boundaries with water lenses should no longer control creep, the adjustment of strains in neighbouring crystals being facilitated by melting-refreezing. As a test for the theory, it should be investigated whether transient reversible creep (Duval's pseudo-elasticity), which is mainly due to the misfit of grains, diminishes when the water content increases.