Space use and resource selection of bobcats in the Appalachian Mountains of western Virginia

Bobcats are an apex predator and a species of socio-cultural importance in the central Appalachian Mountains. Despite their importance, knowledge of bobcat spatial ecology in the region is sparse. We examined space use and resource selection of bobcats in the Appalachian Mountains of western Virginia during 3 biological seasons: breeding (January-March), kitten-rearing (April-September), and dispersal (October-December). We observed sex effects on all space use metrics, with male seasonal areas of use (SAU) approximately 3 times larger than female SAUs and male movement rates 1.5 times higher than females during all seasons. We found no seasonal effect on SAU size for either sex. Female movement rates increased during the kitten-rearing season, and male movement rates increased during the dispersal season. We examined seasonal bobcat resource selection at 2 hierarchical scales, selection of home ranges within the landscape (2nd order) and selection of locations within home ranges (3rd order). Female bobcats exhibited 2nd order selection for higher elevations and deciduous forest and avoidance of fields. Males exhibited 2nd order selection for higher elevations and fields. Male 2nd order selection appears to be driven largely by the spatial distribution of females, which is mediated through the valley and ridge topography of the study area. Sample size precluded 3rd order analysis for females, however males exhibited 3rd order selection for higher elevations, fields, and deciduous forest. Resource selection patterns varied seasonally for both sexes, possibly driven by seasonal shifts in prey availability. Our findings highlight the importance of forested ridges to bobcats in the region. Our findings also illustrate the differences in space use between sexes, which future research efforts should consider. Further research should investigate seasonal shifts in bobcat prey selection, which may further explain the seasonal resource selection shifts we observed, and highlight potential implications for prey species.

9 174 incorporating the movement of animals through an autocorrelation function derived from 175 movement models fit to the data [39]. Furthermore, AKDE reduces to a conventional kernel 176 density estimator when locations are truly independent, and can correct for missing locations and 177 irregular sampling schedules through an optimal weighting method [40]. We estimated 95% 178 annual home ranges for bobcats with at least 4 months of relocation data, during at least 2 179 seasons. We estimated 95% seasonal areas of use (SAU) for bobcats with locations collected for 180 at least 1 month in a given season. We fit linear mixed effects models for each season using 181 restricted maximum likelihood, with area of 95% SAU as the response variables. We used a 182 natural logarithm transformation for home range sizes to meet assumptions of normality. We 183 included both the interaction and main effects of sex and season as predictors, and treated 184 animal-specific intercepts as random effects. We fit models in the package lme4 [42] and 185 assessed the significance of factors and degrees of freedom using Satterthwaite's method for 186 approximating degrees of freedom in the lmerTest R package [43].
187 Movement Analysis 188 We estimated each bobcat's movement rates in meters moved per hour, calculated as the 189 straight-line step length between successive locations divided by the time lag. Seasonal 190 movement rates were only examined for bobcats that were monitored for at least one month in a

274
, Figure 2). The home ranges of the 2 dispersing males were 84.8 and 88.5 respectively. km 2 km 2 275 Male SAUs were larger than female SAUs during all seasons and there was no effect of season 276 on SAU size (Table 1, Figure 2). Average SAU size of females (11.8 ± 1.2 ) and average km 2 277 SAU size of males (32.8 ± 2.0 ) did not differ from average annual home range size of each km 2 278 sex respectively, indicating that SAUs do not shift location extensively throughout the year. We estimated annual movement rates for all bobcats (n=20), and seasonal movement 284 rates for bobcats with at least 1 month of relocations within a given season. Average male 285 movement rates (232.3 ± 12.0 meters/hour) were approximately 1.5 times higher than average 286 female movement rates (154.4 ± 8.9 meters/hour). Male movement rates were higher than female 287 movement rates during all seasons, and female movement rates were significantly higher during 288 the kitten-rearing season while male movement rates were significantly higher during the 289 dispersal season (Table 2, Figure 3).
290 We conducted 2 nd order resource selection for all resident bobcats (12 males, 6 females), 295 which excluded 2 dispersing males. For females distance to deciduous forest, distance to fields, 296 and elevation were the strongest predictors of 2 nd order resource selection (Table 3, Figure 4).
297 During all seasons, females selected SAUs that were at higher elevations, closer to deciduous 298 forest, and farther from fields than expected (Table 3, Figure 4). Females exhibited strongest 2 nd 15 299 order selection for deciduous forest during the kitten-rearing season (Table 3, Figure 4). Females 300 exhibited strongest 2 nd order avoidance of fields during the kitten-rearing season, weaker 301 avoidance of fields during the dispersal season, and weakest avoidance of fields during the 302 breeding season (Table 3, Figure 4). Females exhibited strongest 2 nd order selection for higher 303 elevations during the breeding season, less strong selection for high elevations during the 304 dispersal season, and weakest selection for high elevations during the kitten-rearing season 305 (Table 3, Figure 4). Females exhibited strongest 2 nd order selection for mixed forest during the 306 dispersal and breeding seasons, but did not select or avoid mixed forest during the kitten-rearing 307 season (Table 3, Figure 4). Females exhibited 2 nd order selection for steeper slopes during the 308 dispersal season, but exhibited 2 nd order selection for more gentle slopes during the breeding and 309 kitten-rearing seasons (Table 3, Figure 4).

310
For males, distance to fields and elevation were the strongest predictors of 2 nd order 311 resource selection (Table 3, Figure 4). During all seasons, males selected SAUs that were closer 312 to fields and at higher elevations than expected (Table 3, Figure 5). Males exhibited strongest 2 nd 313 order selection for fields during the kitten-rearing season and weakest selection for fields during 314 the breeding season (Table 3, Figure 5). Males exhibited weakest 2 nd order selection for high 315 elevations during the breeding season compared to dispersal and kitten-rearing seasons (Table 3, 316 Figure 5). Males exhibited 2 nd order selection for mixed forest during all seasons, but this 317 selection was weakest during the kitten-rearing season, following a similar pattern as females 318 (Table 3, Figure 5). Males exhibited 2 nd order selection for deciduous forest during the dispersal 319 season, but avoided deciduous forest during breeding and kitten-rearing seasons (Table 3, Figure   320 5). Slope was not a significant predictor of male 2 nd order resource selection (Table 3, Figure 5).
16 321  Figure 6). Male bobcats 330 exhibited 3 rd order selection for deciduous forest during all seasons (Table 4, Figure 6). Male 331 bobcats exhibited 3 rd order selection for fields during all seasons, but this selection was stronger 332 during the kitten-rearing season (Table 4, Figure 6). Male bobcats exhibited 3 rd order selection 333 for high elevations during all seasons, with the strongest selection for high elevations during the 334 dispersal season and weakest selection during the breeding season (Table 4, Figure 6). Lastly, 335 male bobcats exhibited 3 rd order selection for gentle slopes during the kitten-rearing and 336 dispersal seasons, but selected for steeper slopes during the breeding season (Table 4, Figure 6).