We are using bobcats (Lynx rufus) as a model organism to examine how roads affect the abundance, distribution, and genetic structure of a wide-ranging carnivore. First, we compared the distribution of bobcat-vehicle collisions to road density and then estimated collision probabilities for specific landscapes using a moving window with road-specific traffic volume. Next, we obtained incidental observations of bobcats from the public, camera-trap detections, and locations of bobcats equipped with GPS collars to examine habitat selection. These data were used to generate a cost-surface map to investigate potential barrier effects of roads. Finally, we have begun an examination of genetic structure of bobcat populations in relation to major road networks. Distribution of vehicle-killed bobcats was correlated with road density, especially state and interstate highways. Collision models suggested that some regions may function as demographic sinks. Simulated movements in the context of the cost-surface map indicated that some major roads may be barriers. These patterns were supported by the genetic structure of bobcats. The sharpest divisions among genetically distinct demes occurred along natural barriers (mountains and large lakes) and in road-dense regions. In conclusion, our study has demonstrated the utility of using bobcats as a model organism to understand the variety of threats that roads pose to a wide-ranging species. Bobcats may also be useful as one of a group of focal species while developing approaches to maintain existing connectivity or mitigate the negative effects of roads.

We examined the utility of bobcat (Lynx rufus) detections by deer hunters in New Hampshire, USA, as an index of relative abundance. To validate the index, we compared the number of hunter outings that detected bobcats per 1,000 outings (hunting seasons 2009 through 2013) with a statewide model of habitat suitability after we partitioned both by wildlife management units (WMUs). We also compared annual hunter-detection rates with tallies of road-killed bobcats. The habitat-suitability model and bobcat road mortalities were correlated with hunter-survey indices. Linear and exponential trends in hunterdetection rates of AE10%/year could be detected with power !99% over 5-year periods. As a result, we believe that major changes in bobcat abundance could be detected by the hunter-survey index. However, without information needed to calibrate the relationship between the index and bobcat populations (e.g., linear or exponential), it was not possible to estimate the extent of change in bobcat abundance by a specific change in detection rate. We consider hunter detections a useful approach to monitor bobcats if applied in concert with other indices (e.g., road-kill tallies, trapper harvests, etc.) to obtain a reliable gauge of population trends.