Rain-on-snow (ROS) events can cause the roof snow to be heavier with uneven distribution of water retention due to slope and snowpack structure. Current load codes apply an additional uniform weight of the water that is not accounted for in the ground snow load, assuming uniform load distribution. A multi-point flux approximation based three-dimensional model is employed to resolve water transport in snow as a porous medium for ROS load estimation. It captures distinct non-uniform saturated and unsaturated layers, forming a triangular highly saturated water storage zone at the eaves. The water content and pressure at different levels near the outflow boundary showed steepest gradients near the bottom boundary, which stabilized once equilibrium between inflow and outflow was reached. The rainfall intensity-duration analysis showed the maximum surcharge load at 48 mm/h intensity for 1 h duration, exceeding current ROS load standards. Although most cases are effectively resolved, the model exhibits hydraulic stagnation near the seepage boundary, causing peak load overestimation (24%) for low slope (2°)-high depth configurations. The improved model with a local dynamic grain-growth function based on wet-snow metamorphism is integrated to update hydraulic conductivity locally. This refinement reduced the peak load error to 5.6% for high-snow depth low-sloped case, while temporal outflow was also closer to the experiments. The parametric analysis with the improved model under varying span, depth, roof slope, and rainfall intensity identified the roof slope as the primary determinant for load intensification, with rainfall intensity and snow depth exerting non-linear compounding effects on water retention.
Bajracharya et al. (2026) studied this question.