Step-selection functions to infer active search behavior
Of the four carnivore species, none made movements such that they encountered parturient mule deer (within a 200-meter proximity) in the first 30 days post parturition more often than the random movements generated in the step-selection function (Figure 3a; Figure 4a; Tables S1–S2) suggesting they were not actively searching for mule deer neonates. We documented no GPS-collared cougars or bobcats within 200 meters of a GPS-collared mule deer at the time a simultaneous fix was taken during the 30 days following parturition so models could not be fit for these species. For elk, only two carnivore species exhibited movements that positioned them in proximity of elk post-parturition more often than did hypothetical movements, cougars (βelk presence = 0.94, P = 0.001; Figure 3b; Table S3) and bears, although for bears the result was marginally significant (βelk presence = 0.30, P = 0.059; Figure 3b; Table S3). Sex-specific analysis revealed that male bears were more likely to encounter elk than expected by chance (βelk presence = 0.56, P = 0.004, N = 7 bears; Figure 4b; Table S4) while there was no evidence supporting this effect for female bears (βelk presence = -0.45, P = 0.24, N = 4 bears; Figure 4b, Table S4). Projecting model predictions onto the landscape spatially and overlaying GPS locations of post-parturient female elk in the 30 days post birth further revealed similarities between areas used by neonatal elk and areas selected for by male bears and cougars as documented by the high degree of spatial overlap (Figure 5a,c). Female bears, however, generally did not select for areas that overlapped with elk neonates (Figure 5b), suggesting their space use decisions were driven more heavily by other resources that were unrelated to elk.