Structural Evolution During Cyclic Glacier Surges: 2. Numerical Modeling

Abstract

The mesoscale structures of a glacier express the history of flow, temperature, and stress. Thus, in principle, numerical ice dynamics models have sufficient physics to examine the formation and transport of these structures. In this study we use a vertically integrated thermomechanical ice dynamics model to simulate the temporally evolving patterns of surficial moraine, stratification, foliation, and folding of glacier ice, and the density and orientation of traces of former crevasses. The modeled glaciers are simplified versions of Trapridge Glacier in northwest Canada that allow diagnostic modeling of influences on glacier structure and help to clarify the physics and numerics. In the model, surges occur every 50 years in response to a prescribed cyclic change in bed friction. Medial moraine patterns are simulated by tracking the englacial and supraglacial trajectory of debris injected at fixed points in the accumulation region. Stratification is assumed to be associated with isochronal surfaces, and vertical foliation is explained in terms of horizontal flattening of strain ellipsoids. Crevasses form when and where the intensity of tensile stress exceeds a prescribed threshold; crack damage is cumulative so that crevasse traces observed at sampling sites are a superposition of the damage accumulated en route. Folding is parameterized but not resolved. By evaluating the deformation gradient tensor along ice particle trajectories and applying the polar decomposition to this tensor, we isolate the cumulative effects of rotation and stretching by ice flow and calculate strain ellipsoids as well as other practical indicators of deformation. Keypoints: c ©2019 American Geophysical Union. All Rights Reserved. • Ice dynamics models have sufficient physics to explain many of the structural features of surging and non-surging glaciers. • Model calculations of the deformation gradient indicate that convergent ice flow favors development of longitudinal foliation structures. • The propensity for folding can be parameterized using scalar invariants of a rank 3 tensor derived from the deformation gradient. c ©2019 American Geophysical Union. All Rights Reserved.