This project builds on the insights and datasets developed in the Avalanche Problem Algorithm Project, focusing on the same regions. With validated avalanche problem assignments available, we have all of the necessary components for assessing the overall avalanche hazard. The objective of this project is to develop and validate a physics-based algorithm for forecasting avalanche danger levels according to the North American Public Avalanche Danger Scale based on snowpack simulations and aligned with the Conceptual Model of Avalanche Hazard (CMAH).

This project introduces the first transparent, physically grounded danger rating algorithm grounded in the CMAH and tailored to the North American context. This fully data-driven approach bridges the gap between the CMAH and the North American Public Danger Scale through objective, traceable algorithms. It offers a structured and transparent alternative to black-box machine learning models. By addressing the interaction between multiple avalanche problems and their characteristics, this approach aligns with North American forecaster workflows and supports operational efficiency. It also enhances our understanding of simulated avalanche problem characteristics and their combined influence on overall hazard levels. The algorithm relies solely on SNOWPACK output files, therefore this approach is universally applicable across different avalanche climates, countries, and continents without requiring historical datasets.

• Martin Perfler
• Simon Horton
• Florian Herla
• Vincent Vionnet
• Alec Van Herwijnen
• Pascal Haegeli

• Avalanche Canada
• Others?

• Mitacs
• Avalanche Canada
• SARNIF?