Abstract
The fisher is warranted for protection under the Endangered Species Act in the western United States and, as such, it is especially important that conservation and management actions are based on sound scientific information. We developed a landscape-scale suitability model for interior northern California to predict the probability of detecting fishers and to identify areas of important fisher habitat. Previous models have been extrapolated to this region, but our model was developed from the results of strategically planned detection surveys within the study area. We used generalized additive modeling to create a model that best distinguished detection (n = 55) from non-detection (n = 90) locations on the basis of environmental covariates. Four models were averaged to create a final model including the following variables: Amount of Dense Forest, Percent Hardwood, Medium & Large Trees, Structurally Complex Forest, Adjusted Elevation, Insolation Index and Predicted Abundance of Mammalian Prey. This model was well calibrated and correctly classified fisher detections 83.6% of the time and absences (non-detections) 70.0%. Independent test data were classified less well; 76.2% and 53.0%, respectively, perhaps a result of differences in the spatial and temporal characteristics of the data used to build versus test the model. The model is the first comprehensive portrayal of the distribution and configuration of habitat suitability in this region and provides managers a tool to monitor habitat change over time and to plan vegetation treatments. It also represents an example for the development of similar models for dispersal-limited mammals with large area needs, as well as other species associated with late-successional forests in northern California.
Research highlights
Fishers are carnivores whose conservation is advanced using landscape habitat models. Our model predicts probability of detecting fishers in interior northern California. Predictors include dense forest, tree size, elevation, insolation and mammal prey. We describe a variety of uses for this model in a management context.
Keywords: Fisher; Martes pennanti; California; Klamath; Habitat suitability; Landscape ecology; Habitat modeling
