Applying automated avalanche terrain mapping to wildlife ecology: A case study of Dall's sheep in southcentral Alaska

Avalanche hazard has been studied extensively in the context of human risk, but its role in alpine wildlife ecology remains poorly understood. Long-term telemetry studies have revealed that avalanches are an important source of mortality for some mountain ungulates. Investigating the processes underlying these events is challenging because avalanche activity and snowpack conditions are rarely monitored across the spatial and temporal extent of wildlife studies. Recent advances in automated avalanche terrain mapping provide a new opportunity to address this knowledge gap by quantifying wildlife exposure to modeled avalanche terrain across large landscapes. We demonstrate this application using Dall's sheep (Ovis dalli dalli) in southcentral Alaska, where a 12-year study identified avalanches as the leading known cause of adult mortality. We used GPS locations from 134 adult Dall's sheep (385 sheep-years) collected between 2012 and 2024 across Chugach State Park and the Talkeetna Mountains, Alaska, USA. We characterized seasonal patterns of terrain exposure using the Automated Avalanche Terrain Exposure Scale and complementary metrics representing potential release areas, runout terrain, slope, and aspect. Guided by previous studies of Dall's sheep habitat selection, we developed a priori ecological expectations for seasonal exposure and used descriptive summaries of GPS locations to examine patterns of sheep use. Sheep locations occurred almost exclusively within avalanche terrain throughout the year, with Complex terrain accounting for 61–67% of locations across seasons. Winter and spring locations occurred on slightly steeper slopes while simultaneously having lower potential release area values and lower-value runout terrain than summer and autumn locations. Use of south-facing terrain increased from 64% of locations in summer to 70% in autumn and winter and 83% in spring—coinciding with the highest seasonal concentration of avalanche mortality. These descriptive patterns illustrate how complementary avalanche-terrain metrics can reveal patterns of wildlife exposure that would not be apparent from slope angle and aspect alone. Integrating automated avalanche terrain mapping with wildlife telemetry provides a new approach for identifying patterns of terrain exposure to guide further investigation into factors influencing wildlife avalanche risk.