Healthy, dense forests growing in avalanche terrain reduce the likelihood of slab avalanche release by inhibiting the formation of continuous snow layers and weaknesses in the snowpack. Driven by climate change, trends towards more frequent and severe bark beetle disturbances may reduce the protective capacity of mountain forests against avalanches. We examined the spatial variability in snow stratigraphy, i.e., the characteristic layering of the snowpack, by repeatedly measuring vertical hardness profiles with a digital snow micro penetrometer (SMP) in a spruce beetle infested Engelmann spruce forest in Utah, USA. Four study plots were selected: non-infested/green, infested > 3 years ago/gray stage, salvage-logged, and non-forested/meadow. Based on our SMP measurements and a layer matching algorithm, we quantified the spatial variability in snow stratigraphy, and tested which forest, snow and/or meteorological conditions influenced differences between our plots using linear mixed effects models. Spatial variability in snow stratigraphy was best explained by the percentage of canopy cover (R2 = 0.71, p < 0.001), and only 14% of the variance was explained by the stage within the outbreak cycle. That is, differences between green and gray stage stands did not depend on the reduction in needle mass per se. At both study plots, a more heterogeneous snow stratigraphy developed, which translates to disrupted and discontinuous snow layers and therefore reduced avalanche formation. In contrast, salvage logging that reduced the canopy cover to ~25%, led to a spatially less variable and similar snow stratigraphy as observed in the meadow. At these two study plots, a homogeneous snow stratigraphy consisting of distinct vertical and continuous slopeparallel soft and hard snow layers had formed, a condition which is generally more prone to avalanche release. Our findings therefore emphasize advantages of leaving dead trees in place, especially in protection forests where bark beetle populations have reached epidemic levels.
