A full-field flowline framework including dynamic recrystallization to predict viscoplastic anisotropy in temperate glacier

Ice is a polycrystalline material whose microstructure can induce strong viscoplastic anisotropy. Ice fabric (i.e  preferred crystal orientations) and ice flow are closely linked: strong anisotropy of a polycristal develops as a result of its deformation history. Strong fabrics have indeed been observed both in nature and in laboratory experiments. In a glacier flow, such anisotropy can modify the directional viscosity of ice, making it locally harder or softer. Preferred crystal orientations may therefore influence glacier flow at large scale.

To better characterize this influence, several models have been developed to predict fabric evolution coupled with glacier flow. However, the impact of fabric under near-melting temperature remains poorly quantified. In such conditions, dynamic recrystallization (DRX) is expected to strongly affect fabric evolution. Moreover, field observations that could constrain and validate fabric-evolution models in warm and highly dynamic flow are still scarce, leaving the role of ice textures in glacier dynamics unconstrained. As a result most large-scale glacier simulations still rely on isotropic rheologies combined with enhancement factors, which cannot adequately represent local anisotropy induced by evolving fabrics.

R3iCe [1] is a full-field model using a finite element method that couples both the mechanical behavior and the texture evolution of polycrystalline ice. It was recently developed to predict the evolution of crystal orientations under constant strain rate or deviatoric stress, driven by viscoplastic deformation and dynamic recrystallization. R3iCe has been validated against different laboratory creep experiments, where it successfully reproduces both texture evolution and the associated mechanical softening during tertiary creep. However, the model remains untested under more complex deformation cases, such as those experienced by ice particles within real glacier flows.

In this contribution, we extend the validated R3iCe model toward glacier-scale applications by constructing a R3iCe Flow Line (RFL) approach. It extracts the deformation history of Lagrangian ice parcels from large-scale glacier flow simulations, such as Elmer/ice, and provides the kinematic inputs required to drive R3iCe along glacier flow lines. The scheme is first validated using torsion experiments, which allow us to quantify the errors involved in predicting fabric evolution along a flow line in this controlled setting.

R3iCe Flow Line is then applied to Argentière Glacier (French Alps), which is a temperate glacier. Near-surface samples collected in the ablation area are assumed to represent end-of-flow line fabrics, and are compared with RFL predictions driven by a transient Elmer/Ice flow simulation.

By combining R3iCe Flow Line, a state-of-the-art Elmer/Ice simulation of Argentière Glacier flow, and field observations, this work aims to demonstrate the importance of accounting for fabric and its evolution under temperate, highly deforming conditions, where DRX is at play.

[1] T.Chauve; M. Montagnat; V. Dansereau; P. Saramito; K. Fourteau; A. Tommasi. Comptes Rendus. Mécanique, Volume 352 (2024), pp. 99-134. doi: 10.5802/crmeca.243