Spatial snowdrift modelling for an open natural terrain using a physically-based linear particle distribution equation

Snowdrift, which results from deposition of wind transported snow, has been primarily estimated empirically rather than using physically-based modelling since the snow redistribution process is extremely complex. This study demonstrates a practical predictive model for snow redistribution based on the Linear Particle Distribution equation, which consists of snow surface diffusion, snow surface advection, and snow surface erosion components. Here, we focus on numerical model development and implementation for two-dimensional natural terrains at meter-scale resolutions with and without perforated snow fences, which has been difficult to model in a two-dimensional field. First, a selected numerical scheme was implemented in the Snow Movement Over Open Terrain for Hydrology model platform and tested by the exact solutions under a few well-defined boundary conditions. Then, to simulate snowdrifts around the snow detention structures in the middle of the computational domain, an equivalent solid snow fence concept was introduced and tested. The model was applied to several terrains in the Laramie Range, Wyoming, and at two sites on the North Slope of Alaska, where wind-induced snow redistribution plays a major role. Data from Airborne Light Detection and Ranging, Ground Penetrating Radar, and Unmanned Aerial Vehicle photogrammetry were used to calibrate and validate the model. The numerical snow redistribution model effectively reproduces the observed snowdrift distributions when snow densification and snowmelt effects were minimal. The model applications illustrated that the diffusion effect generally dominated snow redistribution with limited contributions of advection and erosion effects for abrupt terrain transition and perforated object, respectively.

Automated summary

This study presents a practical predictive model for snow redistribution based on the Linear Particle Distribution equation, and successfully implements it in two-dimensional natural terrains. The model effectively reproduces observed snowdrift distributions when snow densification and snowmelt effects are minimal, with diffusion effects generally dominating snow redistribution.