Who skis where, when?

• Main Author: Toft, Håvard Boutera
• Format: Dissertation
• Language: English
• Subjects: avalanche; backcountry usage; decision-making; exposure; risk

Backcountry skiing has emerged as a so-called adventure sport in recent decades and has had a considerable social and economic impact in Northern Norway. Tourism, Norway's fifth-largest export industry, is experiencing significant growth and is especially important for small communities. However, recreation in avalanche terrain comes at a cost due to the risk of avalanches. Consequently, any change in avalanche risk could directly or indirectly impact a large proportion of the population in these areas. Moreover, many avalanche deaths with frequent search and rescue operations are unsustainable for these small communities. There are currently no precise methods to calculate the risk of backcountry skiing or measure whether the avalanche forecast leads to a behavioral change among skiers, selecting less exposed terrain or avoiding avalanche terrain altogether when the forecasted avalanche hazard is high. Measuring avalanche risk or whether avalanche forecasts influence skiers' terrain choices requires comprehensive data on daily backcountry usage, detailed insights into skiers' locations, and slope-scale avalanche conditions. As a steppingstone, this thesis has developed an automated model (AutoATES) to classify avalanche terrain exposure, comparable in performance to human mappers. Such a model enables large areas to be mapped using a consistent and efficient method. Scalability is essential when we want to compare terrain exposure with backcountry usage in the future. Furthermore, this thesis has attempted to enumerate backcountry usage using two methods. The first approach used signaling data. Unfortunately, there were considerable discrepancies between the estimated positions from the signaling data and our validation data using an independent GPS track. The second approach, using beacon checkers to count backcountry skiers, was far more successful. During two seasons, from December to May from 2021 to 2023, we recorded 56,760 individual trips from 26-29 trailheads within the ~2,600 km2 study area, offering valuable insights into backcountry usage as a function of time of day, week, and month. In the future, it may be possible to estimate the proportion by analyzing the extensive database of GPS tracks submitted to Center for Avalanche Research and Education to measure what percentage of these activities originate at a beacon checkpoint. Once we know the proportion, it could be possible to compare our data with accident and fatality data to estimate the