A Monte Carlo Approach to Approximating the Effects of Pore Geometry on the Phase Behavior of Soil Freezing

Freezing in porous media is associated with a host of dynamic phenomena that stem from the presence and mobility of premelted liquid at subzero temperatures. Accurate assessments of the progressive liquid-ice phase transition is required for predictive models of frost damage, glacier-till coupling, and many other cold regions processes, as well as for evaluating the capacity for water storage in near-surface extraterrestrial environments. We use a Monte Carlo approach to sample the pore space in a synthetic 3D packing of poly-dispersed spherical particles and evaluate local geometrical constraints that allow us to assess changes in the relative proportions of pore fluid and ice. By approximating the phase boundary geometry in fine-grained pores while considering both the curvature of the liquid-ice interface and wetting interactions with matrix particles, our model predicts changes in phase equilibrium in granular media over a broad temperature range, where present accounting for the colligative effects of chloride and perchlorate solutes. In addition to formulating the constitutive behavior needed to better understand properties and processes in frozen soils, our results also provide insight into other aspects of phase equilibria in porous media, including the formation of methane hydrates in permafrost and marine sediments, and the partitioning between liquid water and vapor in the vadose zone.