It was recalled that the cyclic loading response of ice could be analyzed by using a dislocation-based model of anelasticity to estimate the effective mobile dislocation density. The cyclic loading response was sensitive enough to monitor dislocation density changes that occurred during creep straining. Cyclic and creep loading experiments could therefore be used to determine the relationship between the dislocation density that evolved during creep straining, and the anelastic and viscous components of strain. Creep and cyclic loading experiments were conducted here on synthetic saline and freshwater ice specimens. The results indicated that, in addition to the stress and strain dependences, the dislocation density that developed during straining also depended upon the temperature and microstructure. Due to this temperature dependence, the apparent activation energy for creep - when dislocation multiplication took place - was higher than that for creep in the absence of dislocation multiplication. Published data were found to support this. The results indicated that both the anelastic strain, and the viscous strain rate, varied with the mobile dislocation density. The first results indicated that the power-law stress exponent, n, was equal to 3 when the applied stress was high enough to cause an increase in dislocation density. However, n was equal to about unity when the dislocation density remained constant during straining. The results showed that previous straining increased the stress level that was associated with the transition from n = 1 to n = 3. This provided support for a glide-controlled mechanism of ice creep. A dislocation glide-based formulation for viscous straining was presented. The model agreed well with experimental data on the pure flow regime.

A Dislocation-Based Analysis of Strain History Effects in Ice. D.M.Cole, G.D.Durell: Philosophical Magazine A, 2001, 81[7], 1849-72