Elastoviscoplastic micromechanical modeling of the transient creep of ice

Abstract : A salient feature of the rheology of isotropic polycrystalline ices is the decrease of the strain rate by more than 2 orders of magnitude during transient creep tests to reach a secondary creep regime at a strain which is systematically of ∼1%. We use a recent (so-called “affine”) version of the self-consistent mean-field theory to model the elastoviscoplastic behavior of ice. The model aims at bridging scales between the rheology of single grain and the one of polycrystals by evaluating the intergranular interactions. It takes into account the long-term memory effects, which manifests itself by the fact that local stress and strain rate in grains depend on the whole mechanical history of the polycrystal. It is shown that the strong hardening amplitude during the transient creep is entirely explained by the stress redistribution within the specimen, from an almost uniform stress distribution upon instantaneous loading (purely elastic response) to strong interphase and intraphase heterogeneities in the stationary regime (purely viscoplastic response). The experimental hardening kinetic is much too slow to be explained by the same process; it is attributed to the hardening of hard glide slip systems (prismatic slip) in the transient regime. Moreover, the model very well reproduces the permanent creep rate of several highly anisotropic specimens of the Greenland Ice Core Project ice core (pronounced crystallographic textures), when accounting for a single-grain rheology that well matches the experimental one. Our results are consistent with recent findings concerning dislocation dynamics in ice.