Terrestrial carnivores are among the most imperiled species worldwide, yet some species are resilient and are recovering in human-dominated landscapes after decades or centuries of absence. Bobcat (Lynx rufus) populations were extirpated from much of Midwestern US in the mid-1800’s, and are currently expanding and recolonizing their former range. In this study, we investigated multi-scale habitat selection for Ohio’s expanding bobcat population, and examined habitat connectivity in order to evaluate the conduits for dispersal statewide. We used citizen observations collected between 1978 and 2019 and logistic regression to evaluate population-level habitat selection, and GPS telemetry data for 20 individuals collected between 2012 and 2014 and a distribution-weighted exponential Resource Selection Function to evaluate individual-level habitat selection within home ranges. At the population level, bobcats selected for higher amounts of forest and pasture (at a 50 km2 scale) and herbaceous vegetation (at 15–50 50 km2 scales), thus overall heterogeneous forested habitat. At individual (home range) level, bobcats selected for forested habitats with low road density and farther away from high traffic roads; they also showed weak selection for open habitat at the home range level. Male home ranges were significantly greater than female home ranges. Lastly, we used the population-level spatial outputs (i.e. habitat suitability map) to parameterize habitat connectivity models using circuit theory in the program Circuitscape. We tested three relationships between habitat suitability and resistance to movement and used a subset of data on potential dispersing individuals to evaluate which relationship performed best. All three relationships performed almost equally well, and we calculated a weighted averaged connectivity map as our final map. Habitat was highly permeable to movements between core areas of two genetically distinct subpopulations located in southeastern Ohio. We also identified potential dispersal corridors from the core areas to other regions of Ohio dominated by agriculture and suburban development via forested riparian corridors. Overall, our analysis offers new information on habitat selection and connectivity in a rebounding felid population and offers important ecological information for wildlife management strategies. We recommend that the suitability and connectivity models should be periodically updated until the population reaches an equilibrium, and be integrated with data from neighboring states for a comprehensive assessment of a conservation success story.

No abstract

Bobcats (Lynx rufus) were extirpated from many midwestern states in the mid‐1800s owing to habitat loss and overharvesting. Recently, bobcats have recolonized Ohio, USA, and neighboring states and given their furbearer status elsewhere, there is interest in opening a harvest season; however, demographic factors and viability of this population are currently unknown. We developed a spatial population simulation model to assess the long‐term viability of the bobcat population in Ohio in the face of 2 human‐induced factors limiting population growth: potential harvest and road mortality. We combined habitat suitability and road mortality risk data with vital rates for bobcats from Ohio and other populations to simulate possible scenarios for Ohio's population. Our baseline scenario simulations showed no risk of extinction for Ohio's bobcat population in the next 40 years, but population trajectories were lower and exhibited greater uncertainty when we modeled the population with a lower maximum density of animals per cell. At low harvest intensity (αh = 0.05), the bobcat population also exhibits low risk of extinction. When harvest intensity increases (αh = 0.1, 0.15) or when adults (≥2 yr) are targeted by harvest, simulations show declining populations, greater uncertainty in projections, and possible risk of extirpation. Our models indicated that if harvest and road mortality are additive, then the bobcat population could withstand marginal increases in road mortality (αr = 0.1) at low harvest intensity (αh = 0.05). Future increases in road mortality with higher harvest intensity (αh = 0.1, 0.15) were unsustainable. Our results can be used by wildlife managers to assist with decisions on population‐level management. This simulation model can easily be adapted for other large mammals and can be modified to assess a variety of ecological and anthropogenic influences to wildlife populations.

No abstract