Snowpack and avalanche hazard assessment and modelling

Forecasters had expressed concerns about the reliability of snowpack simulations in operational avalanche forecasting. This project systematically assessed how accurately regional-scale snow-pack simulations reflect critical snow layers and avalanche conditions identified by expert fore-casters.

Avalanche forecasters and researchers face challenges in interpreting the vast, complex data from snowpack observations and simulations. This project developed new computational methods to align, compare, and summarize snow profiles, making it easier to extract relevant information for avalanche hazard assessment and decision-making.

Snowpack models are powerful but often too complex to use effectively in daily avalanche forecasting. This project aimed to improve operational value by designing clear and intuitive ways to visualize model output.

Snowpack models in Canada are driven by weather forecasts, which can introduce significant uncertainty. This project aims to evaluate and improve model accuracy by assessing precipitation inputs and comparing modelled snow depth with observations.

Avalanche forecasting in North America follows standards set by the Conceptual Model of Avalanche Hazard and the North American Public Avalanche Danger Scale. This project brings together several studies, using both quantitative and qualitative methods, to examine how these standards are applied and to identify ways to improve the accuracy and consistency of hazard assessments.

The motivation behind developing a physics-based avalanche danger level model is to fully represent the hazard assessment performed by avalanche forecasters using snowpack simulations, thereby supporting forecasters in data analyses and improving the consistency of avalanche forecasts.

Snowpack models in Canada are driven by weather forecasts, which can lead to accumulating errors over the season. This project aimed to improve model accuracy by comparing modelled snow depth with observations and implementing real-time corrections.

Clustering is a useful way to summarize model output by exploring spatial and temporal snowpack patterns. This project developed and refined clustering methods to support flexible forecasting, where forecasters define region boundaries based on daily conditions.

Avalanche Problems are building the entry point of the conceptional model of avalanche hazard. While avalanche hazard assessments are often based on expert judgment, forecasters increasingly refer to snowpack models as an additional source of information to support their decisions. This project aims to evaluate and improve algorithms that identify avalanche problems by comparing snowpack model outputs at specific weather stations and across entire forecast regions to expert assessments, observed avalanche activity, and practical usability in forecasting.