Comparing snowpack meteorological inputs to support regional wet snow avalanche forecasting

Abstract

Wet snow avalanches are predicted to increase in frequency with climate change and are often difficult to forecast. Improving our understanding of wet snow avalanche timing will help with current forecasting challenges. The onset of wet snow avalanching is closely tied to the temporal progression of liquid water flow through the seasonal snowpack. Measuring the flow of water through the snowpack in-situ is difficult due to the spatial variability of snow depth and structure. However, physical snowpack models can potentially simulate this process. The accuracy of snowpack models is heavily dependent upon the quality of the meteorological input data. A thorough investigation of model output differences using several different meteorological inputs for forecasting water movement and wet snow avalanches has not yet been thoroughly investigated. Here, we evaluate indicators of regional wet snow avalanches produced by the SNOWPACK model using different meteorological input. We compare the accuracy of SNOWPACK modeled outputs driven by two different numerical weather prediction (NWP) forecast models: the High-Resolution Deterministic Prediction System (HRDPS) and the North American Model (NAMnest). We leverage hourly automated weather station data, daily operational avalanche observations along the Going-to-the-Sun Road in Glacier National Park, Montana, United States, and in-situ snow stratigraphy and wetness profile observations to validate the SNOWPACK modeled outputs. This research is directly applicable to avalanche forecasting operations and future avalanche research as wet snow avalanche timing evolves due to climate change.