Forecasting for deep slab avalanches

Avalanches released by deep weak layers are known to be difficult to forecast. This study considers the predictive merit of weather and snowpack data for avalanches that released throughout two winters on layers of faceted crystals. These layers formed above rain crusts in November 1996 and in November 1997 in the Columbia Mountains of western Canada. This study focuses on the first winter, during which the facet layer released an estimated 500,000 tonnes of snow in 700 dry slab avalanches. The facet-crust combinations were the result of a cold air mass cooling a layer of dry snow on top of a rain-wetted layer. The resulting temperature gradient in the dry snow formed a relatively weak layer of facets on top of a hard crust. By early January 1997, the faceted layer from the first winter was buried 1-2.5 in below the surface in many starting zones in the North Columbia Mountains. Most avalanches occurred during or within 2 days of loading by snowfall or wind transported snow. Increases in air temperature over 4-5 days correlated with increased avalanche activity. The effects of warming and cooling on slab stability are discussed, but for thick slabs, current theories do not explain observations of decreased stability. We argue that the fractures that release natural deep slab avalanches may be initiated where the slab is locally thin. Based on rank correlations, the highest ranked predictors of natural avalanches include previous avalanche activity, accumulated snowfall over several days, changes in air temperature over 4-5 days, snowpack properties including a shear frame stability index, and the difference in hardness between the facet layer and the crust. (C) 2001 Elsevier Science B.V. All rights reserved.