Synthesis of habitat models for management of wolverine (Gulo gulo): Identifying key habitat and snow refugia in the Columbia and Rocky Mountains, Canada

The wolverine is a wide-ranging and frequently at-risk or data-deficient species with dramatic range contractions across the northern hemisphere. Recent reports of low population densities inside and outside protected areas in western North America highlight the need for better conservation practice, policy, and planning across large landscapes. To assess broad habitat needs, we synthesized available wolverine habitat models in the Columbia and Rocky Mountains (63,000 km2), Canada. We used coefficients from four existing models to create spatial predictions over environmental datasets including snow, landcover, and roads. We averaged predictions using the distance-weighted mean of equal-area percentile habitat values and validated the output by comparison to independent data. Because persistent spring snow is tightly correlated with highquality habitat, we assessed 2080 spring snow cover forecasts (under Representative Concentration Pathway (RCP) 8.5 high emissions scenario) to identify potential habitat refugia in British Columbia. Our synthesized habitat model identified high-quality habitat along mountain ranges, notably in the Purcell Mountains and the Columbia Icefield in the Rocky Mountains. Mean habitat value was 0.70 (SD: 0.19) inside protected areas and 0.55 (SD: 0.28) outside protected areas. The British Columbia side of the study area is forecasted to lose 44% of persistent spring snow cover by 2080, with declines identified inside many protected areas. By synthesizing existing habitat research and climate forecasts, we provided new insights at the broad spatial scale needed to conserve wide-ranging species like wolverine and to inform land-use planning for recreational access and the establishment and management of protected and conserved areas.