In the conterminous United States, wolverines (Gulo gulo) occupy semi-isolated patches of subalpine habitats at naturally low densities. Despite previous work modeling wolverine habitat, a consensus has yet to be established on what factors most strongly influence wolverine habitat use. In this analysis, we aimed to determine the relative importance of human land use, snow water equivalent, high-elevation talus, landforms, elevation, and vegetation community type on wolverine habitat quality for animals collared around the Greater Yellowstone Ecosystem. We conducted this analysis at a second- and third-order selection scale to examine the importance of scale in habitat selection. We also compared output from parametric and nonparametric methods, including standardized logistic regression coefficient, logistic regression variable importance, pseudo-R-squared output, machine learning variable importance, and random forest machine learning mean decrease gini, to determine how commonly used approaches with different assumptions produce different results. We ran these analyses separately for male and female wolverines, as each sex has vastly different home range sizes and selection pressures driving habitat selection.
Results/Conclusions
and analysis methods. Snow water equivalent, distance to high-elevation talus, and latitude-adjusted elevation were the driving selective forces for wolverines across the Greater Yellowstone Ecosystem (second-order). At the same time, more nuanced landform types increased in importance for movement within-home ranges (third-order). Overall, our results indicate that wolverine habitat selection is, in large part, driven by high-elevation structural features. However, we also found that snow water equivalent, which is subject to change because of anthropogenic impacts, is an important feature for habitat selection as well.