Identifying key physical processes in snow compaction at different strain rates

The mechanical behavior of snow is complex, as it depends on a variety of physical processes occurring at different scales, from the microstructure (sintering, bond breakage, etc.) to the scale of the snowpack and entire slopes. In particular, snow mechanical behavior is highly dependent on strain rate, with a ductile-to-brittle transition occurring at strain rates of about 10^-4-10^-3 s^-1. It is important to develop comprehensive snow mechanical models accounting for this complexity for applications such as avalanche hazard evaluation, snowpack compaction or hydrological studies.

In this work, our objective is to build a continuous numerical model in a finite strain framework, that captures the key mechanical behavior of snow in a large range of strain rates. In particular, this model should be capable of properly retrieving the various deformation patterns observed in experiments, from quasi-homogeneous deformation in the ductile regime to the emergence of unstable localization patterns, such as compaction bands or cracks, typically observed in the brittle regime.

The model is based on an elasto-viscoplastic constitutive law, inspired by the Modified Cam Clay model, which is characterized by an elliptical yield surface. Two specific effects are included in the evolution of this yield surface throughout the deformation process: a hardening effect due to the compaction of the snow, and a viscous effect due to the competition between bond breakage and sintering of the microstructure. This law has been implemented in the software Matter [1] based on the Material Point Method (MPM). This method combines Lagrangian integration points and a fixed background mesh, which allows for computations of large deformations.

We performed 2D simulations of centimeter-scale samples (15mm x 15mm), undergoing uniaxial displacement-controlled compaction, at different strain rates between 1.8x10^-6 and 7.5x10^-3 s^-1. These simulations are meant to reproduce the laboratory experiments of Bernard et al. [2], which were carried out in an X-ray microtomograph, providing reconstructions of the snow microstructure and deformation throughout the compression. Detailed comparisons between numerical and experimental results will be presented to evaluate the robustness of the numerical model.

In addition, a systematic sensitivity analysis was conducted to investigate the impact of the various physical processes considered in the constitutive law on the observed compaction patterns. Of particular interest is the role of sintering on the emergence and propagation speeds of localization bands. Finally, future adaptations of the model to investigate the propagation of instabilities in heterogeneous snowpacks will be discussed.

 

[1] Blatny, L. and Gaume, J.: Matter (v1): An open-source MPM solver for  granular matter, Geosci. Model Dev., 18, 9149–9166, https://doi.org/10.5194/gmd-18-9149-2025, 2025.

[2] Antoine Bernard. Etude multiéchelle de la transition ductile-fragile dans la neige. Science des matériaux [cond-mat.mtrl-sci]. Université Grenoble Alpes, 2023. Français. ⟨NNT : 2023GRALI027⟩. ⟨tel-04145610⟩