Introduction
Human land uses and development are altering landscapes worldwide and restricting the movements of wide-ranging species such as large carnivores (Crooks et al. 2011; Ripple et al. 2014). One outcome of the expanding human footprint is that carnivores and people are increasingly overlapping with one another. Several tools have been employed in human-dominated landscapes to mitigate carnivore interactions with people, including fencing protected areas or other barriers and policies to separate humans from wildlife (e.g., McInturff et al. 2020). Depending on the relative scale and permeability of these physical structures, such efforts often result in new patterns of wildlife movement and navigation (e.g., Wilkinson et al. 2021b). On coexistence frontiers , where human infrastructure and activity have either recently appeared or are rapidly increasing, carnivores must learn to traverse novel landscapes with novel risks. In such regions, it is critical to identify predictors of landscape navigation for carnivores and determine whether and how these species can adaptively traverse these increasingly human-dominated landscapes (Sanjayan & Crooks 2005).
Understanding carnivore landscape navigation within coexistence frontiers requires social-ecological approaches rather than only considering ecological factors (Lute et al. 2020; O’Neal Campbell, 2014). In general, wildlife sharing space with people must navigate three main elements present on the landscape, namely ecological factors, human infrastructure, and human acceptance, that determine social-ecological landscape permeability (e.g., Ghoddousi et al. 2021). This guiding framework is critical because while certain carnivore species may be able to persist in human-altered habitats (see Athreya et al. 2013; Breck et al. 2019; Chapron et al. 2014; Devens et al. 2019), human intolerance may be a strong enough limiting factor that it can override adaptability for carnivore populations or survival of individuals navigating developing landscapes (Moss, Alldredge & Pauli 2016). Thus, human perceptions are likely to be an important factor in determining how carnivores navigate landscapes (Behr et al. 2017). On coexistence frontiers, people may not have developed tolerance or acceptance for species which they have not encountered before or which they are now encountering more frequently (e.g., Lute & Carter 2020). In this case, intolerance may serve as a proxy for people employing nonlethal hazing or deterrents or as an indicator for tendencies to conduct illicit persecutory actions toward wildlife (Benson et al. 2023; Ditmer et al. 2022; Manfredo et al. 2021), such as poisoning (see Ogada 2014) or habitat destruction (see Ripple et al. 2014). In these social-ecological spaces, people also interact with, respond to, and alter ecological features and processes that influence carnivore movement at different scales, such as vegetation availability (e.g., Bateman & Fleming 2012; Suraci et al. 2020) and climatic season (e.g., through changing livestock movements; Schuette et al. 2013).
Spatial and ecological scales are important for contextualizing human coexistence with carnivores (Carter & Linnell 2016). For instance, anthropogenic development may have community-level effects by pushing some species into limited remaining natural habitats (Parsons et al. 2018), yet at a fine scale, carnivores can exhibit adaptations to anthropogenic development – or even be synanthropic—on an individual level (i.e., Moss et al. 2016; Nisi et al. 2021; Wang et al. 2015; Suraci, Nickel & Wilmers 2020). Through these scale-dependent processes, human-dominated landscapes near protected areas can result in a source-sink dynamic, whereby carnivore populations from protected areas that venture into more densely human-populated regions are more likely to die through anthropogenic causes (Lamb et al. 2020). However, individuals may succeed in these spaces by taking advantage of anthropogenic resources at finer scales, depending on human tolerance (i.e., Moss et al. 2016). Thus, for large carnivores, which are often highly mobile, social-ecological landscape permeability across scales is essential to maintain populations. Carnivore species living in both human-dominated environments and protected environments may avoid anthropogenic features such as roads and fences (e.g., Baker & Leberg 2018; McInturff et al. 2020; Young et al. 2019) or change their activity patterns to adjust for human presence (e.g., Gaynor et al. 2018). Yet carnivores may also either avoid or be attracted to human infrastructure at different scales (Poessel et al. 2014), underpinned by the density of the infrastructure (Morales-Gonzalez et al. 2020; Xu et al . 2023), the infrastructure’s impact on resource availability (i.e., Belton et al. 2016), the species involved (Wilkinson et al. 2021b), or individual animal characteristics, such as life stage (Thorsen et al. 2022).
As a widely reviled (Glickman 1995; Macdonald et al. 2022) and behaviorally plastic apex predator (Holekamp & Dloniak 2010), spotted hyenas (Crocuta crocuta , hereafter “hyenas”) are a model species for understanding the nature of carnivore adaptiveness to human-caused landscape change and negative human perceptions. Hyenas are generally considered one of the most behaviorally flexible carnivore species, yet there have been few empirical conclusions regarding the extent and mechanisms of their adaptiveness to human activities, infrastructure, and tolerance. Green et al. (2018) found that hyena populations in Maasai Mara, Kenya increased in an area of anthropogenic disturbance, possibly linked to increased livestock consumption. In densely populated areas in Ethiopia where native prey is depleted, hyenas have become almost entirely dependent on anthropogenic food (e.g., Yirga et al. 2012). However, other studies have found negative, neutral, or nuanced responses to people. In one study in Kenya, hyena activity shifted in response to livestock grazing and other anthropogenic activities (Kolowski et al. 2007), while a study in South Africa found that hyena propensity to visit anthropogenic sites varied depending on season, age, or individual (Belton et al. 2018). Determining whether a species is surviving or thriving in the presence of humans is complex, as even highly synanthropic species may be exposed to greater levels of stress, toxicants, and disease while living in anthropogenic landscapes (Murray et al. 2016; Murray et al. 2019). While movement is thus solely the broadest scale at which to assess carnivore survival and wellbeing within coexistence frontiers, understanding how hyenas and other behaviorally plastic mammalian carnivores navigate anthropogenic landscape change is fundamental to forecasting the resilience of movements, food webs, and ecosystems in rapidly developing landscapes.
We sought to provide insight into spotted hyena abilities to navigate coexistence frontiers by examining the following questions in and around Lake Nakuru National Park and Soysambu Conservancy, Kenya: 1) How do spotted hyenas navigate anthropogenic infrastructure, human activity, tolerance, and ecological factors across scales, seasons, and management types (i.e., fully protected vs. multi-use), and 2) How many spotted hyenas cross through the region’s conservation fence, and how does spotted hyena fence navigation manifest at a fine scale? We predicted that (1) hyenas would respond differently to ecological factors than they do to anthropogenic infrastructure and activity, (2) hyena selection for and against landscape characteristics would differ at different scales (i.e., fine-scale and landscape scale), (3) hyenas would exhibit different responses to covariates in the rainy vs. the dry season, (4) hyenas with dens in the fully protected national park would be more avoidant of barriers, less avoidant of roads, and more attracted to verified and perceived livestock predation hotspots outside of the park than would hyenas with dens in the multi-use conservancy, and (5) fence crossing would be limited to a few select individuals and hyenas likely perceive fence navigation to be risky. We employed resource selection functions (RSFs) and step-selection functions (SSFs) to determine hyena space use and landscape navigation at the home range and step scales, respectively (Reinking et al. 2019). We then used this information to infer whether and how to consider a suite of social-ecological factors when designing for hyena landscape permeability, and present the implications of these inferences for global human-carnivore coexistence.