Fine‐scale knowledge of how anthropogenic effects may alter habitat selection by wolves (Canis lupus) is important to inform conservation management, especially where wolf populations are expanding into more populated areas or where human activity and development are increasingly encroaching on formerly pristine environments.
From 1999 to 2003, we documented rendezvous sites (n = 31) used by wolves (n = 6 packs) in a previously unstudied wolf population in the southern Apennines, Italy (Pollino National Park [PNP]), where conditions are ideal to unveil behavioral adaptations of wolves living in human‐dominated landscapes. We adopted a hierarchical, multi‐scale habitat selection approach by using a set of environmental, topographic, and anthropogenic factors within multi‐grain resource selection functions (MRSFs) at the landscape and the territory extent. Habitat selection by wolves was scale‐dependent and different habitat components affected wolves at different grains. When establishing a territory at the landscape scale, wolves avoided areas featuring high densities of humans, paved roads, and trails, and they preferentially located territories where higher forest cover and rough terrain enhanced concealment and ensured reduced accessibility by humans. Concurrently, wolves also selected open areas and, at coarse grain, areas of high density of dirt roads and trails to enhance hunting and traveling efficiency. When selecting rendezvous sites within territories, wolves still selected for forest cover and lessened avoidance of anthropogenic features apparent at the lower order of selection. When choosing rendezvous sites, wolves also avoided low‐use anthropogenic linear features and rough terrain, unveiling trade‐offs in selection decisions across different spatial and temporal scales. Our results reinforce the notion that occurrence of wolves in human‐modified landscapes is shaped by avoidance of anthropogenic pressure, but they also indicate that rendezvous site selection by wolves conform to a hierarchical decision‐making process that is extent‐ and grain‐dependent. Spatially explicit models of rendezvous site selection such as the one we developed for PNP enhance management effectiveness for strategies such as regulating human access and activity during the pup rearing seasons, preventing human persecution, mitigating livestock depredations, and designing cost‐effective monitoring programs.