An animal’s behavioral state influences their movements across the landscape and likely also influences their selection of habitat. Therefore, incorporating behavioral state in habitat selection studies may provide details about how and why the animals are using space and resources that would otherwise be obscured if individual movement and behavior are ignored. We investigated the movements and habitat selection of 22 mountain lions (Puma concolor; 15 females, 7 males) in the Davis Mountains in west Texas between 2011 and 2018 using GPS location data collected every 3 hours. We discerned the number of behavioral states describing the mountain lion movements, estimated the distributions of step-lengths and turning angles that best described these states, and estimated the transition probabilities among states by using hidden Markov models (HMMs). We assigned the most likely state to each animal location with the Viterbi algorithm, then used step-selection functions to investigate habitat selection. For each animal step, we randomly selected alternate unused steps based on the movement parameters of the assigned state and used conditional logistic regression to compare the animals’ choices of land cover at each step with that of the matched unused steps.
Results/Conclusions
On average, female mountain lions exhibited shorter 3-hour step-lengths than males (mean ± SE; females: 723.47 ± 90.16 m; males: 971.64 ± 229.29 m). Both sexes used the shortest step-lengths between 15:00 and 18:00. Both female and male mountain lion movements were best described by three movement states with a gamma distribution for step-length, a wrapped Cauchy distribution for turning-angle, and a covariate of hour of day affecting state transition probabilities. One movement state included movements with short step-lengths (females: 12.30 ± 9.92 m; males: 18.56 ± 18.39 m) and sharper turning angles (females: -178.25 ± 19.42°, males: 173.97 ± 17.35°), the second state included moderate step-lengths (females: 470.27 ± 580.81 m, males: 572.15 ± 670.94 m), and wide turning-angles (females: 3.41 ± 5.25°, males: 4.66 ± 5.76°), and the third state included long step-lengths (females: 1,811.19 ± 1,272.69 m, males: 2,735.32 ± 1,517.74 m), and wide turning angles (females: -1.59 ± 26.90°, males: 3.36 ± 30.80°). Bottomland/Riparian and grassland/savannah were the most and least selected land covers, respectively. Our results provide information about mountain lion movement and habitat selection in west Texas and the framework we used provides a method for incorporating behavior into habitat selection studies.