[M-IS15] Geophysical particulate gravity current

[MIS15-P05] Prototype of probabilistic hazard map for snow avalanche reflecting uncertainties of snow cover conditions during winter

*Takahiro Tanabe¹, Sojiro Sunako¹, Kouichi Nishimura² (1.National Research Institute for Earth Science and Disaster Resilience, 2.Professor emeritus of Nagoya University)

Keywords:Snow avalanche, Hazard map, Uncertainty

Snow avalanches are the most common gravity mass flows occurring in snowy regions. The most well-known avalanche runout zoning methods consider topographic angles (Bakkehøi et al., 1983). However, these methods are too simplistic to yield quantitative data, such as flow thickness, speed, and pressure. Consequently, recent advancements in numerical modeling and computers have led to the implementation of hazard maps using dynamics models (Issler et al., 2023). In these hazard maps, potential release areas (PRAs) are first estimated based on topographical data, vegetation, and climatic conditions, then the runout area is delineated by a numerical simulation with a typical parameter set (e.g. initial volume, friction coefficient, etc.) for each PRA. These maps showed better results than maps obtained by topographic angles, however, such parameters have uncertainties prior to the occurrence. This study aims to develop a prototype of probabilistic hazard map for snow avalanches, accounting for the uncertainty of the input values.
In this study, PRAs are initially estimated based on topographical conditions. Subsequently, dynamics model calculations are then conducted on these estimated PRAs, incorporating uncertainty model input values. The uncertainty of input values is assumed to be represented by probability density functions. The resulting hazard map, termed a probabilistic hazard map, delineates the runout area along with its associated probability. For instance, a hazard map based on uniform distributions with upper and lower bounds for the snow height and the friction coefficient that can be taken during winter season at a PRA signifies the potential threat of the area. To evaluate uncertainty, polynomial chaos quadrature (PCQ) is used (Dalbey et al., 2008). PCQ approximates numerical results using a polynomial of input values with uncertainty and demonstrates faster convergence compared to the Monte Carlo method when the number of input values are small. The PCQ method enables the mapping of avalanche hazards in specific locations with lower numerical cost. In the presentation, a prototype of probabilistic hazard map for snow avalanches, reflecting uncertainties in snow cover conditions during winter will be presented.