Habitat complexity mediates the predator–prey space race

Abstract

The spatial relationship between predator and prey is often conceptualized as a behavioral response race, in which prey avoid predators while predators track prey. Limiting habitat types can create spatial anchors for prey or predators, influencing the likelihood that the predator or prey response will dominate. Joint spatial anchors emerge when predator and prey occupy similar feeding habitat domains and risk and reward become spatially conflated, confusing predictions of which player will win the space race. These spatial dynamics of risk‐foraging trade‐offs are often obscured by habitat heterogeneity and community complexity in large vertebrate systems, fueling ambiguity regarding the generality of predictions from predator–prey theory. To test how habitat distribution influences the predator–prey space race, we examine correlation in puma and vicuña habitat selection and space use at two sites, one of which generates a distinct risk–foraging trade‐off at a joint spatial anchor. The distribution of vegetation, which serves as both forage for vicuñas and stalking cover for pumas, differs between the sites; the llano contains a single central meadow that acts as a joint spatial anchor, while the canyon is characterized by more heterogeneous vegetation. Puma–vicuña habitat selection correlation was positive in the llano and negative in the canyon, and similarly, utilization distributions were more strongly correlated in the llano than the canyon. Vicuña locations occurred at higher values of puma habitat selection and utilization in the llano than in the canyon. Similarly, puma locations in the llano occurred at higher values of vicuña habitat selection and utilization than in the canyon. Although pumas consistently selected for and utilized vegetative and topographic cover regardless of habitat distribution, vicuñas only selected against vegetation in the heterogeneous canyon site, reducing spatial correlation with pumas. Our work suggests a joint spatial anchor favors pumas in the space race due to the inability for vicuñas to avoid crucial foraging habitat. The outcome of the predator–prey space race appears to be strongly informed by the distribution of habitat, whereby corresponding predictability of predator and prey favors predators in the spatial game.

Where and when to hunt? Decomposing predation success of an ambush carnivore

“…Meadows (high‐risk, high‐reward) had a greater proportion of kills than expected at all times of day, and interestingly, only in meadows did high encounter rates counteract intrinsically lower catchability during the day. In essence, the nature of meadows to serve as a joint spatial anchor for pumas and vicuñas (Smith et al 2019 b ) promotes the trade‐off between encounter and capture probabilities by stimulating vicuñas to select for this habitat when they have greater detection capacity during the day but avoid it at night (Smith et al 2019 a ).…”

“…Due to the extreme variation in intrinsic vulnerability (i.e., catchability) among highly delineated habitat types in our system, vicuñas select for safe refuge habitats at night (Smith et al 2019 a ), forcing pumas in San Guillermo National Park (which are dependent on vicuñas as their primary food source) to hunt during the day more often than other puma populations. Vicuñas also avoid canyons (Smith et al 2019 b ), reducing encounter probability in what would otherwise be preferred hunting habitat. Vicuña diel migration therefore limits when and where pumas can encounter prey in habitats with sufficient stalking cover for hunting.…”

“…We chose habitat covariates based on association with the habitat classes and previous knowledge of their available stalking cover and selection by vicuñas. TRI is a proxy for topographic stalking cover, avoided by vicuñas (Smith et al 2019 a , b ), and positively associated with canyon habitat. NDVI is a proxy for vegetative stalking cover, selected by vicuñas (Smith et al 2019 a , b ), and strongly predictive of meadow habitat.…”

“…Vicuñas avoid meadows (i.e., use them less than expected given their availability) at night, but return to feed in them during the day (Smith et al 2019 a ). Although vicuñas do use individual and group vigilance to mitigate predation risk in meadows (Donadio and Buskirk 2016), they are ultimately anchored to these predictable feeding sites where detection and evasion of stalking predators are low (Smith et al 2019 b ). At night, vicuñas select for open plains habitats, which have lower intrinsic risk than meadow and canyon habitats due to higher detectability of stalking predators (Smith et al 2019 a ).…”

“…Under the temporal prey‐catchability hypothesis, we predict that the frequency of puma kills will be concentrated at night (when prey vulnerability is highest due to reduced detection capacity, resulting in high conditional capture probability) as in other puma populations (Beier et al 1995, Smith et al 2015), whereas the temporal prey‐abundance hypothesis predicts that kill frequency will follow temporal patterns of encounter rates with vicuñas (predicted to be higher during the day; Smith et al 2019 a ). Similarly, the spatial prey‐catchability hypothesis predicts that puma kills will be concentrated in habitats with stalking cover (where vicuñas have the highest vigilance rates due to greater predation risk; Donadio and Buskirk 2016), whereas the spatial prey‐abundance hypothesis predicts that pumas will disproportionately kill vicuñas where encounter rates are highest (predicted to be in habitats that contain both stalking cover and high‐quality forage; Smith et al 2019 a , b ). We hypothesize that kill probability is best predicted by an interaction between spatial and temporal patterns of encounter and catchability that are related to the dynamic nature of vicuña antipredator behavior (Smith et al 2019 a ), and that there may be a trade‐off between encounter and capture probability that results in equalization of captures (Michel and Adams 2009).…”