Description of the snow microstructure as a 3D assembly of grains

International audienceThe concept of snow grain is commonly used in the snow community, either in situ to identify the snow type or in numerical models to reproduce the physical and mechanical behaviour of the snowpack. Among all possible definitions of grains (optical, crystalline and mechanical), we consider mechanical grains as particles separated by potential stress concentrators and mechanical weakness zones. To detect these grains in the snow 3D microstructure captured by X-ray tomography, we use a geometrical algorithm based on curvature and constriction. The algorithm is parameterized with a local contiguity indicator, which defines the segmentation scale. The representation of the microstructure with grains enables the investigation of the bonding system, which is determinant for mechanical properties. In this work, we propose to quantitatively characterise this bonding system with specific grain contact area, contiguity and a new parameter, the minimal cut area defined as the minimal surface of ice separating two opposite faces of a cubic sample. The analysis conducted on 3D images of snow evolving with isothermal metamorphism shows a strengthening of the bonding system, confirming earlier observations on 2D slices of snow. However, the classical description of the bonding system with specific grain contact area and contiguity suffers from a high dependence to the scale of the grain segmentation. On the contrary, the minimal cut area does not depend on the segmentation scale and also measures the connectivity of the bonding system in a given direction