A probabilistic-based entrainment erosion model combined with the Eulerian-Lagrangian method to simulate snow distribution

Current snow erosion models predominantly depend on threshold friction velocity. However, considering the wind speed fluctuation and the randomness of the particle position on the snow bed, the force on the snow bed particles is highly randomized. To address this, a probability-based snow particle erosion model is proposed in this study, which accounts for the influence of wind speed fluctuations on snow particle entrainment. By integrating the Eulerian-Lagrangian method, the model predicts the erosion, rebound, and redeposition processes of snow particles without relying on a threshold friction velocity. To validate the model, we employ field measurements and wind tunnel data from established studies, demonstrating excellent agreement between numerical predictions and measured results. At the same time, through the comparison of the predicted results of the cloud map, we found that the snow depth minima are generally located in the local high friction velocity region, while the snow depth maxima are generally located in the region where the friction velocity changes sharply along the space and is close to the low friction velocity. The Discrete Phase Model (DPM) simulations used in this paper elucidate the underlying mechanics, demonstrating how particle-laden flow interactions drive these spatial correlations, which helps to preliminarily estimate the location of extreme snow depths based on friction velocity cloud images. Furthermore, this paper proposes a mesh refinement method based on the simulation approach, enabling its application to full-scale drifting snow dynamics simulations on large-span building roofs.