Implications for landscape permeability
At broad scales, hyenas in this developing region appear to be selecting for ecological characteristics that reflect their resource selection in other, less developed systems. In contrast, movement choices at a fine scale are more nuanced and influenced by anthropogenic factors. Hyena clan sizes in this region are relatively large (with more than 50 animals per group for clans assessed thus far) despite the small size of the protected areas, which could be influencing the apparent movement of hyenas toward people and likely toward anthropogenic resource subsidies. Other studies in similar environments have estimated hyena carrying capacity to be orders of magnitudes lower than the populations seen in our study site (e.g., Holekamp & Smale 1995; Yirga et al. 2014).
Despite this suspected significant reliance on anthropogenic resources, hyenas showed a general aversion to roads, with different selection strengths depending on scale and land management type, which is contrary to what we expected. Fences also present a semi-permeable barrier for spotted hyenas, which appear to cross them as quickly as possible rather than lingering. Other studies have found that keeping development and subdivision below a certain threshold may allow for sustained carnivore navigation of the landscape between core habitat areas (Smith et al. 2019; Xu et al. 2023). This may also prove true for the spotted hyenas, which appear to have complex relationships with infrastructure within and surrounding the two protected areas. Yet, hyena relationships with fences can also provide information that is helpful for management efforts. We can use fence behavior analyses to determine existing permeable fence segments (Xu et al . 2020) and make ecologically-informed decisions about where carnivore (and other wildlife) corridors in and out of fenced regions will be the most useful, practical, and cost-effective.
Overall, our results imply that anthropogenic factors may influence fine scale decision making differently than landscape-scale selection. Hyenas may be adaptable enough to switch to anthropogenic food sources in regions of depleted natural prey or limited resources, yet their ability to rely on anthropogenic food may be linked to regional tolerance of hyenas (e.g., Yirga et al. 2012). Spatially explicit human acceptance and experience have the potential to predict where wildlife corridors are likely to succeed for certain species or taxa, while also providing insight into how wildlife may be using anthropogenic resources (Behr et al. 2017; Ghoddousi et al. 2021). Coupled with hyena context-specific selection for and against infrastructure characteristics, these results demonstrate that a multiscale and multidisciplinary understanding of social-ecological landscape use and navigation can help to determine where and when this species may thrive in human-dominated landscapes. This approach is essential for a species that is key for removing carcasses and diseases from the environment (Sonawane et al. 2021), and in a location that is becoming increasingly fenced, but the social-ecological approaches used here can also be applied to movements and reintroductions of other controversial wildlife species in other settings (see Ditmer et al. 2022; Manfredo et al . 2021; Vasudev et al. 2023; Williamson et al. 2023; Williamson & Sage 2020). Future research on social-ecological landscape permeability for wildlife should include the incorporation of detailed land cover covariates, in-depth quantification of tolerance and experience as spatial proxies for behaviors, and testing of GPS collar data across RSF- and SSF-informed social-ecological least cost corridor models.