Specific surface area evolution during dry snow metamorphism: insights from interface growth velocity computed on 4D tomographic data

Dry snow microstructure refers to the complex three-dimensional arrangement of ice and air at the sub-millimeter scale. This microstructure undergoes constant shape transformations known as snow metamorphism. These transformations are driven by variations in equilibrium vapor pressure at the ice-air interface, which depend on the local curvature and temperature gradient. A key descriptor of snow microstructure is the specific surface area (SSA), which is the surface area of the ice and air interfaces normalized per ice volume or mass. This metric is commonly used to quantify the average grain size in snowpack models. Moreover, SSA affects important physical properties of the snowpack, including the spectral albedo of the surface and fluid permeability. Consequently, accurately representing SSA evolution in snowpack models is crucial. Overall, snow SSA decays over time, except in specific conditions where SSA increases, such as high temperature gradients. Current descriptions of SSA in snowpack models, such as CROCUS or SNOWPACK, are not fully satisfying, especially they fail to reproduce SSA increase. It restricts the model’s ability to represent processes under high temperature gradients, as typically occurring in Arctic regions. Recent efforts have been made to derive theoretical relations between SSA and microstructural and growth parameters, but have been applied to a limited number of snow evolution experiments.

In this work, we build upon these previous studies and investigate the physical mechanisms driving SSA evolution for numerous dry snow metamorphism scenarios. We re-derive a relationship between the SSA temporal evolution, the local interface growth velocity, and the local mean curvature. To examine the implications of this relation on different snow microstructures, we acquired 20 time series of 3D X-ray tomographic images of dry snow metamorphism at high temporal and spatial resolution during cold-lab experiments. These experiments span a wide range of thermal boundary conditions and initial snow types. Using this data set, we compute local properties on the grain surface, including interface growth velocity, mean curvature, and temperature gradients. Focusing on a subset of experiments, we present SSA evolution for temperature gradients ranging from 10 to 100 K/m. In particular, we investigate the mechanisms responsible for SSA increase at high temperature gradients. We aim to disentangle the respective contributions of local microstructural shape and local temperature gradients to the overall SSA evolution. A more comprehensive understanding of the mechanisms at stake in the SSA evolution will help develop a robust representation of SSA in snowpack models.