When, where, and how prey pair antipredator behaviors to natural and anthropogenic mortality risks

Background Behavioral responses of prey to predation risk have ecological impacts that can be as great as direct mortality. Risk response involves either behavioral changes or spatial avoidance, but it is not clear how prey decide between these strategies. Theory often suggests that prey pair responses to risks based on the hunting mode of the prey (hunting mode hypothesis), but prey may ignore hunting mode to prioritize responding to the most lethal predators (lethality hypothesis). Furthermore, prey may respond to the spatial distribution of these risks (risky places hypothesis) or respond only during the periods of highest risk (risky times hypothesis). Methods To test these hypotheses, we evaluated the behavioral responses of white-tailed deer (Odocoileus virginianus) to risks from two natural mesopredators and human sources of mortality. Specifically, we determined, for each source of risk, whether deer responded with behavioral state changes or spatial avoidance and whether this behavior changed with time (diurnally and annually). We collared and tracked 40 female and 29 male deer. To determine the response of deer to risk, we collected data on the distribution of coyotes (Canis latrans), bobcats (Lynx rufus), human modification, hunters, and roads. We used hidden Markov models (HMM) to determine whether each covariate impacted the probability of transitioning between behavioral states and selection functions (SSF) to determine whether deer spatially avoided each covariate. Results Generally, deer changed behavioral state in response to both mesopredators but avoided human modification. In response to mesopredators, deer consistently shifted to slower movement behavioral states. Spatial responses to human modification varied depending on the time of day. During daylight hours, deer selected for human modification, but during the crepuscular and nighttime period, deer avoided human modification. Conclusions Space use was most strongly related to more lethal humans, providing support for the lethality hypothesis. Despite prioritizing humans, mesopredators impacted behavioral state, suggesting that mesopredators still have important impacts on prey behavior. Finally, temporal patterns of avoidance align with other studies that indicate avoidance of predators is time-dependent, but further highlight the complex push-pull relationship of human modified areas on wildlife.