RESUMO
The behavioral mechanisms by which predators encounter prey are poorly resolved. In particular, the extent to which predators engage in active search for prey versus incidentally encountering them has not been well studied in many systems and particularly not for neonate prey during the birth pulse. Parturition of many large herbivores occurs during a short and predictable temporal window in which young are highly vulnerable to predation. Our study aims to determine how a suite of carnivores responds to the seasonal pulse of newborn ungulates using contemporaneous global positioning system (GPS) locations of four species of predators and two species of prey. We used step-selection functions to assess whether coyotes, cougars, black bears, and bobcats encountered parturient adult female ungulates more often than expected by chance in a low-density population of mule deer and a high-density population of elk. We then assessed whether the carnivore species that encountered parturient prey more often than expected by chance did so by shifting their habitat use toward areas with a high probability of encountering neonates. None of the four carnivore species encountered GPS-collared parturient mule deer more often than expected by chance. By contrast, we determined that cougar and male bear movements positioned them in the proximity of GPS-collared parturient elk more often than expected by chance which may provide evidence of searching behavior. Although both male bears and cougars exhibited behavior consistent with active search for neonates, only male bears used elk parturition habitat in a way that dynamically tracked the phenology of the elk birth pulse suggesting that maximizing encounters with juvenile elk was a motivation when selecting resources. Our results suggest that there is high interspecific and intersexual variability in foraging strategies among large mammalian predators and their prey.
RESUMO
Resource selection functions (RSFs) are tremendously valuable for ecologists and resource managers because they quantify spatial patterns in resource utilization by wildlife, thereby facilitating identification of critical habitat areas and characterizing specific habitat features that are selected or avoided. RSFs discriminate between known-use resource units (e.g., telemetry locations) and available (or randomly selected) resource units based on an array of environmental features, and in their standard form are performed using logistic regression. As generalized linear models, standard RSFs have some notable limitations, such as difficulties in accommodating nonlinear (e.g., humped or threshold) relationships and complex interactions. Increasingly, ecologists are using flexible machine-learning methods (e.g., random forests, neural networks) to overcome these limitations. Herein, we investigate the seasonal resource selection patterns of mule deer (Odocoileus hemionus) by comparing a logistic regression framework with random forest (RF), a popular machine-learning algorithm. Random forest (RF) models detected nonlinear relationships (e.g., optimal ranges for slope and elevation) and complex interactions which would have been very challenging to discover and characterize using standard model-based approaches. Compared with standard RSF models, RF models exhibited improved predictive skill, provided novel insights about resource selection patterns of mule deer, and, when projected across a relevant geographic space, manifested notable differences in predicted habitat suitability. We recommend that wildlife researchers harness the strengths of machine-learning tools like RF in addition to "classical" tools (e.g., mixed-effects logistic regression) for evaluating resource selection, especially in cases where extensive telemetry data sets are available.
