INTRODUCTION
Although the distinction between active search and incidental encounter of prey is key to understanding predator-prey dynamics, whether, and under what conditions predators actively search for particular prey species is poorly understood. Diet breadth theory suggests that predators seeking to maximize their rate of energy acquisition should have a more generalist prey profile until the encounter rate with the most profitable prey item crosses a critical threshold that makes it suboptimal to include lower-profitability prey in their diet (MacArthur & Pianka 1966; Figure 1; Supporting Information Text S1). Diet specialization on abundant prey is necessarily associated with active search for that prey, but as the density of the focal prey species declines, additional species enter the diet which reduces the rate of acquisition of the focal prey. A switch from active search for a prey species at high density to incidental encounter of the same species at low density can have stabilizing effects on prey population dynamics (Murdoch 1969) and results in the canonical sigmoid Type III functional response that is associated with many of the most interesting dynamics in predator-prey systems including alternative stable states, bifurcations, and predator pits (Holling 1959; Messier 1994; Leviet al. 2015).
The density-dependent response of predators that is evidence for active search can be found by observing disproportionate predation in years in which a given prey species is at high density. However, in many systems prey densities vary predictably within a year due to a birth pulse that provides a uniquely vulnerable life stage for predators. A classic case of this phenomenon is the seasonal pulse of songbird nests (Schmidt, Goheen & Naumann 2001; Vigallon & Marzluff 2005), and both active search (Pelech, Smith & Boutin 2010) and incidental encounter (Schmidt 2004) strategies have been documented as a function of nest density, vulnerability, and maternal vigilance or defense (Schmidt 1999). In addition to songbirds, large herbivore systems worldwide feature a predictable birth pulse of neonates that are vulnerable to predators across a wide body size gradient. Although mortality of ungulate neonates has been studied extensively (Linnell, Aanes & Andersen 1995), it is still largely unknown whether, and under what conditions predators engage in active search behavior during the birth pulse, which has important implications for prey population dynamics.
The propensity of carnivores to target neonates likely depends on factors intrinsic to both the predator and prey. For predators, this may vary by body size, hunting proficiency or mode, and ability to cope with maternal defense by the large herbivore. For example, ungulate neonates are highly vulnerable to predation immediately following parturition from a suite of carnivores, but quickly become sufficiently vagile to elude those that are less predaceous or of smaller body size. Thus, species such as bears and many mesopredators experience a particularly short resource pulse of neonates (Linnell, Aanes & Andersen 1995; Zager & Beecham 2006; Griffin et al. 2011) (Figure 1a, b). In contrast, the birth pulse may be less consequential to the largest felids and canids that can efficiently capture large-bodied ungulates throughout the first year of life and beyond (Figure 1c). Even within a species, the response of predators to the ungulate birth pulse may vary by individual or sex as has been shown in bears (Jacoby et al.1999; Zager & Beecham 2006; Rayl et al. 2015).
Because predatory behavior during the ungulate birth pulse is influenced by idiosyncratic combinations of factors intrinsic to both predators and prey, studies involving multiple carnivore and multiple ungulate species will be needed to elucidate the generality of search behavior. Previous research has largely focused on identifying spatial shifts in habitat use by bears (but see Bastille-Rousseau et al . 2016; Svobodaet al . 2019) toward birthing grounds or areas on the landscape more likely to contain neonates, which has resulted in different conclusions wherein both active search (Rayl et al . 2018) and incidental encounter (Bastille‐Rousseau et al. 2011; Bowersocket al. 2021) were inferred. Much stronger inference is possible using analytical methods that utilize spatiotemporal encounters between individual predators and individual prey from contemporaneous GPS telemetry data. Further, identifying general conditions associated with incidental encounter and active search requires relocation data on multiple species of predators and prey across gradients in body size, abundance, and life histories.
Here we use contemporaneous GPS tracking data at the level of encounters between individual predators and individual prey to assess whether carnivores actively searched for or incidentally encountered ungulate neonates in a multi-predator, multi-prey system. Each carnivore species (cougar [Puma concolor ], coyote [Canis latrans ], black bear [Ursus americanus ], bobcat [Lynx rufus ]) varied in size, life history, and predatory ability, ranging from large-bodied obligate predators to smaller-bodied omnivores, while prey species (mule deer [Odocoileus hemionus ] and elk [Cervus canadensis ]) varied dramatically in abundance. Our primary objective was to determine whether predators encountered parturient female ungulates more often than expected by chance, indicating active search, while controlling for their habitat preferences. If predators did indeed exhibit targeted search behavior, a secondary objective was to determine whether a shift in space use toward parturition habitat tracked the phenology of the birth pulse consistent with an effort to maximize detections of neonates. We hypothesized that the magnitude of response by predators to the birth pulse would be greater toward elk than mule deer for all carnivore species because elk were approximately 5 times more abundant than mule deer and could thus cross the profitability threshold associated with specialization under diet-breadth theory. We expected the carnivore species with a more generalist diet profile (bears and coyotes) would alter their foraging behaviors more strongly than cougars or bobcats because generalist consumers are more fluid in their response to changing resources (Ostfeld & Keesing 2000; Yang et al. 2008). Bears are the least carnivorous of these taxa, and previous research suggests that male bears are disproportionately predaceous (Rode, Robbins & Shipley 2001, Boertje et al. 1988; Jacoby et al. 1999), so we additionally hypothesized that male bears would be more likely to exhibit active search behavior. Further, cougars kill mule deer and elk of all age classes year-round (Clark et al. 2014), so we expected that the ungulate birth pulse may be less consequential to cougars given their adeptness at killing larger prey such that they may not exhibit a change in behavior during the earliest neonatal period. Finally, we hypothesized that bobcats would show the weakest response, since they rarely consume elk and mule deer in our study area (Ruprecht et al. 2021b).