Integrated species distribution models predict motorized and non-motorized outdoor recreation across seasons

Outdoor recreation is occurring at unprecedented levels, yet our knowledge of how to best monitor and analyse outdoor recreation is outdated. Inaccurate and unreliable recreation information affects everything from designing sustainable recreation plans to developing conservation strategies for sensitive or migratory species. We asked: (1) what are patterns of recreation intensity and how do they differ by activity type (motorized, non-motorized) and season; and (2) which variables best explain recreation intensity across space and time and how do relationships with variables differ by activity type and season. In western Canada, a hub for outdoor recreation in all seasons, we collected five unique datasets on recreation use and occurrence (Strava, systematic and incidental aerial surveys, trail counters, cameras) to predict motorized and non-motorized recreation intensity across six seasons. We fit integrated species distribution models (iSDM) to multiple data types including count, presence/absence, and presence-only data to predict the distribution of recreation. We assessed the effects of terrain, vegetation, accessibility, infrastructure, and snow and climate to understand patterns and predict recreation intensity across six seasons: early winter (Nov–Dec), mid-winter (Jan–Feb), late winter (Mar–Apr), spring (May–Jun), summer (Jul–Aug), and fall (Sep–Oct). We found substantial variation of recreation intensity across the study area and seasons, for both activity types. We also found human access (e.g., trails) explained patterns in recreation activities across seasons. Winter recreation occurred at higher elevation relative to non-snow seasons, whereas summer recreation occurred in more steep terrain. By using iSDMs we were able to leverage multiple datasets with disparate spatial or temporal coverages. This work fills important research and knowledge gaps for people managing land-use, conservation and recreation, specifically those working beyond protected and conserved areas where few open data sources exist. Our research suggests recreation intensity is dynamic and static proxies for recreation commonly used in research and monitoring may be insufficient.