Modeling cougar habitat in the Northeastern United States

• I modeled cougar habitat in the Northeastern U.S. using spatial-explicit Habitat Suitability Indices.
• I present results from original and replicated methods, including two ensemble models.
• Models identified between 20,457 km2 and 160,971 km2 of suitable cougar habitat.
• Based solely on geographic area and prey densities, the Northeast may be able to support between 322 and 2535 cougars.
• Because data deficiency limits comprehensive model validation, these are heuristic tools only.

While the U.S. Fish and Wildlife Service has declared the Eastern cougar (Puma concolor couguar) extinct, proposing to remove the subspecies from the Endangered Species List, in the Northeastern United States there are over 2300 eyewitness reports and nearly a dozen confirmed accounts of cougars. This discrepancy between what has been documented by management agencies and what has been perceived by regional residents raises questions about the current and future presence of cougars in the region, yet little work has been done to examine the Northeast's capacity to support this species. I used spatially-explicit Habitat Suitability Indices to model cougar habitat in the six New England states and that portion of New York East of the Hudson River. I present one original model and five models in which I replicated methods originally established by other authors outside of the study region. For each model I identified contiguous habitat parcels capable of supporting viable breeding populations of cougars according to two estimates of population range size. I evaluated model results by comparing the percent forested land cover within viable habitat patches to that associated with historic cougar kills. I also assessed model agreement by generating two ensemble models – one comprised of each individual model output, and one comprised of viable contiguous habitat that was coincident across all models. I found that all individual models and one ensemble model identified viable habitat according to both population range estimates, while the second ensemble model identified viable habitat according to the liberal range estimate only. Individual models identified between 20,457 and 160,971 km2 of top ranking habitat, enough area to theoretically support between 322 and 2535 cougars. Collectively these models provide a set of heuristic tools that shed light on a species that could influence future trophic interactions in the region. In light of my findings and the active expansion of cougar territory into the Midwest, I recommend that regional management begin to educate local residents about the nature of human–cougar interactions, and to consider preliminary management strategies for dispersing Midwestern cougars as resources allow. I recommend also that future modeling efforts integrate human input from regional biologists, and that these models be used to help evaluate cougar sighting reliability.