Linking climate history and ice crystalline fabric evolution in polar ice sheets

Abstract

An ice sheet consists of an unfathomable number of ice crystallites (grains) that typically have a preferred orientation of the crystalline lattices, termed fabric. At the surface of ice sheets, the microstructural processes that control the grain structure and fabric evolution are influenced by climate variables. Layers of firn, in different climate regimes, may have an observable variation in fabric which can persist deep into the ice sheet; fabric may have 'memory' of these past climate regimes. To model the evolution of a subtle variation in fabric below the firn-ice transition, we have developed and released an open-source Fabric Evolution with Recrystallization (FEvoR) model. FEvoR is an anisotropic stress model that distributes stresses through explicit nearest-neighbor interaction. The model includes parameterizations of grain growth, rotation recrystallization and migration recrystallization which account for the major recrystallization processes that affect the macroscopic grain structure and fabric evolution. Using this model, we explore the evolution of a subtle variation in near-surface fabric using both constant applied stress and a stress-temperature history based on data from Taylor Dome, East Antarctica. Our results show that a subtle fabric variation will be preserved for ~200ka in compressive stress regimes with temperatures typical of polar ice-sheets. The addition of shear to compressive stress regimes preserves fabric variations longer than in compression-only regimes because shear drives a positive feedback between crystal rotation and deformation. We find that temperature affects how long the fabric variation is preserved, but does not affect the strain-integrated fabric evolution profile except when crossing the thermal-activation-energy threshold (~-10°C). Even at high temperatures, migration recrystallization does not rid the fabric of its memory under most conditions. High levels of nearest-neighbor interactions between grains will rid the fabric of its memory more quickly than low levels of nearest-neighbor interactions. Because FEvoR does not compute flow, an integrated fabric-flow model is needed to investigate the flow-fabric feedbacks that arise in ice sheets. Using the open-source Parallel Ice Sheet Model (PISM) and FEvoR, we develop a combined flow-fabric model (PISM-FEvoR). We provide the first integrated flow-fabric model that explicitly computes the fabric evolution and includes all three major recrystallization processes. We show that PISM-FEvoR is able to capture the flow enhancement due to fabric by modeling a slab-on-slope glacier, initialized with a variety of fabric profiles. We also show that the entire integrated fabric-flow history affects the final simulated flow. This provides a further, independent validation of using an integrated fabric-flow model over a constant enhancement factor in ice-sheet models.