Characterization of the snow microstructural bonding system through the minimum cut density

• The minimum cut density characterizes the snow microstructural bonding system.
• Good correlation between this indicator and thermal and elastic properties is found.
• The minimum cut density well accounts for structural anisotropy.
• Snow is shown to present a very low connectivity.

Snow density is commonly used to phenomenologically parameterize other snow properties such as strength or thermal conductivity. However, density is insufficient to fully characterize the variety of microstructures, as revealed by the existing scatter in the parameterizations. This can be explained by the role of bonds which are almost as important as grains to describe the macroscopic properties of snow. A quantification of the narrow constrictions or bonds between snow grains is thus essential to accurately parameterize snow properties. In order to characterize the reduced thickness of the ice matrix at bonds, we introduce a new microstructural indicator, the minimum cut density, ρmc. This variable quantifies, on three-dimensional (3D) microtomographic images of snow, the minimal effective density of a surface that disconnects two opposite faces of the sample. The obtained minimum cut density values are surprisingly low, in the range [0.2, 35] kg m−3 for the tested samples with density in the range [100, 350] kg m−3. This reveals the high variability and weak connectivity of the bonding system in snow. Structural anisotropy of faceted crystals and depth hoar is also well characterized by the minimum cut density. To evaluate the physical and mechanical relevance of this microstructural indicator, we estimate the thermal conductivity and the Young's modulus of the studied snow samples through microstructure-based simulations. An excellent correlation is found between the Young's modulus and the minimum cut density (R2 = 0.97). In particular, the minimum cut density well accounts for the anisotropy of the Young's modulus in faceted snow types. The correlation of thermal conductivity and minimum cut density is also highly significant (R2 = 0.88) but not better than the parameterization with density (R2 = 0.92). We show that this difference is mainly due to the role played by the thermal conduction of air.