Both forest composition and configuration influence landscape-scale habitat selection by fishers (Pekania pennanti) in mixed coniferous forests of the Northern Rocky Mountains

• We evaluated habitat selection by fishers as a function of forest pattern.
• Incorporating landscape configuration metrics improved models of habitat selection.
• Home ranges had high proximity of mature forest and low amounts of open areas.
• Metrics from industrial and roadless areas differed from fisher home ranges.
• Roadless and wilderness areas should not be assumed to be good fisher habitat.

Informed forest management and effective conservation planning require an understanding of how forest patterns influence wildlife species. The fisher (Pekania pennanti) is a wide-ranging, mesocarnivore species that occurs at low population densities and uses large tracts of forested lands in the western USA. Loss and fragmentation of forested habitats are considered primary threats to fisher populations, however, these factors influence two different components of landscape pattern: composition and configuration. We used data from 18 fishers fitted with Argos satellite transmitter collars to evaluate habitat selection at the landscape scale (i.e., 50–100 km2) in north-central Idaho. We developed a set of a priori models about how fishers might respond to forest pattern and tested the hypothesis that both forest composition and configuration influence habitat selection by fishers at broad spatial scales. Model selection results indicated that a model incorporating metrics of both forest configuration and forest composition performed significantly better those that with either alone. Fishers selected landscapes for home ranges with larger, more contiguous patches of mature forest and reduced amounts of open areas. Landscapes that had ⩾50% mature forest arranged in connected, complex shapes with few isolated patches, and open areas comprising ⩽5% of the landscape characterized a forest pattern selected by fishers in our study. To evaluate how well different forest management histories in our study region might provide habitat for fishers, we compared metrics of forest composition and configuration within fisher home ranges with metrics from forests managed under three distinct management histories. Landscapes managed primarily for timber production and lands managed as roadless/wilderness had significantly more open areas, less mature forest, and reduced proximity of mature forest patches than occupied fisher home ranges. These results can be used to facilitate effective conservation of fishers through informed forest management planning.