Wet-snow avalanches can be difficult to forecast. However, recent studies have shown that an index (LWCindex) related to the mean liquid water content of the entire snowpack can be used to predict the onset of periods with high wet-snow avalanche activity. Nevertheless, this index has not yet been verified. We therefore compared modelled and measured liquid water content to wet-snow avalanche activity for four winter seasons at the Dorfberg test site, above Davos, Switzerland. Using the 1-D snow cover model SNOWPACK, we simulated snow stratigraphy, the mean liquid water content and water infiltration within the snowpack. Simultaneously, we used an upward-looking ground penetrating radar (upGPR) to monitor mean liquid water content and changes in percolation depth. Measurement and simulations agreed well and showed that increased wet-snow avalanche activity started when the mean liquid water content of the snowpack reached 0.6-1% by volume. Concurrently, at the onset, a significant diurnal increase in liquid water content was also observed. In three out of four melt seasons, the first arrival of water at the bottom of the snowpack coincided with the onset of high wet-snow avalanche activity. Overall, these results show that the index could improve prediction of wet-snow avalanche activity. The model approach might be particularly helpful for narrowing down the period of temporary avalanche mitigation measures (e.g., preventive closures, artificial release by explosives), since conditions favoring wet-snow avalanches usually persist only for a short period of time. Combined with a numerical weather prediction model, this approach may allow for effective wet-snow avalanche forecasting.
EPIC Alfred Wegener Institut