Social-ecological landscape navigation and fine-scale movement decisions
Hyenas in this region are selecting for different factors at the landscape scale than at a fine scale. Differences in resource selection at different scales were particularly apparent for infrastructure and human experience characteristics, and less apparent for environmental characteristics, selection for which largely remained the same across RSF and SSF results. On a fine scale, hyena navigation appears to be more heavily influenced by roads, fences, and human experiences and acceptance than by environmental characteristics such as NDVI and proximity to water. These scale-dependent affects likely stem from a combination of factors, with fine-scale movements dictated by hyena behavioral flexibility, and landscape scale selection dictated by broader resource availability. While there are changes in the magnitude of effect, at both broad and fine scales hyenas in this region generally select for vegetation greenness, lakes, and roads, and against rivers, boundaries, slopes, and high elevation. In arid environments, spotted hyena hydration is mostly derived from prey rather than water sources (Green et al. 1984), and hyenas can also survive for a week or more without water (Holekamp & Dloniak 2010); thus, selection against rivers could reflect the aridity of the region. Additionally, though the two major lakes in the region are highly alkaline, hyenas have been observed hunting flamingos, wallowing, and using the lakesides’ heavy vegetation as refuge during the heat of the day (authors’ observations). Thus, even undrinkable water sources can serve as critical habitat for hyenas (e.g., Matsumoto-Oda 2021).
At the landscape scale, hyenas are selecting against participatory mapped livestock predation areas and for areas of participatory mapped risk, while the opposite holds true for fine scale selection. The landscape-scale selection for and against these human acceptance covariates may indicate that hyenas are broadly selecting for areas in which they may face hazing or be poisoned (i.e., perceived risks from and low tolerance for hyenas), and against areas that people use for livestock grazing (i.e., participatory mapped livestock predation). People are also likely perceiving the highest risks from hyenas in places that constitute suitable hyena habitat, which supports hyena landscape-scale selection for these areas. In densely populated areas where tolerance is high or there are policies against wildlife killing, carnivore populations may thrive (Athreya et al. 2013; Gebresenbet et al. 2018), yet in areas where tolerance is low (such as in our study site), carnivore populations can be negatively affected by retaliation and other practices (Ripple et al. 2014). Hyenas in this study site may also be avoiding participatory mapped livestock predation areas—which are likely most often used for grazing—as locations with consistent, predictable risks (i.e., through deterrents and higher levels of human activities). This reflects findings from previous studies in regions with less development, in which spotted hyenas reduced their activities in response to livestock grazing (i.e., in the Maasai Mara, Kolowski et al. 2007).
Our results also showed seasonal differences in landscape-scale and fine scale hyena selection for environmental and anthropogenic characteristics. At the landscape scale, in the dry season, hyenas exhibited an increase in selection for vegetation greenness, boundaries, and participatory mapped livestock predation. The latter two support our findings that hyenas may be expanding their ranges and traveling outside of the protected areas during the dry season. When resources are scarce, animals living alongside people may be more likely to choose anthropogenic food sources (Johnson et al. 2015), and hyena predation on livestock or scavenging of livestock and other anthropogenic food sources may be increasing during dry seasons. The boundary navigation result at the landscape level also supports these hyenas’ tendency toward dry season range expansion. Meanwhile, at a fine scale, the dry season showed a minor increase in hyena navigation toward roads, with minimal seasonal differences in selection for environmental and human experience characteristics. While species in other studies have been known to use roads for easier traversal of the landscape (Abrahms et al. 2015; Hill et al. 2020), hyenas in this study area may also use roads in the dry season for dust bathing and access to artificial water points, particularly in the conservancy.
When looking at variation in movement for hyenas living in different management types across both seasons, at the landscape scale roads were more likely to be avoided by Soysambu hyenas than LNNP hyenas, despite human use of roads being extremely low in the conservancy at night. One reason for this could be that despite being active at night, hyenas in Soysambu associate roads with the abundance of human activities that occur during the day on the roads within the conservancy, while in LNNP, there is only one activity happening on roads: tourism. Vehicle speeds in the national park are also heavily regulated. Hyena avoidance of roads in Soysambu stands in contrast to research that has found that animals may select for human infrastructure at night for resources or ease of movement while avoiding it during the day when humans are more active (e.g., Toverud 2019). However, hyenas are generally more skittish in the conservancy than in the national park, possibly due to historical hyena bounties (K. Combes, pers. comm.) before the conservancy’s designation as a wildlife habitat. The anthropogenic activity signature on Soysambu’s roads may thus have a strong enough effect on the spotted hyena’s human-caused “landscape of fear” (Smith et al. 2017; Suraci et al. 2019) as to contribute to nighttime avoidance of roads that are devoid of human activity.
SSFs revealed that hyenas with dens in LNNP are also selecting against boundaries to a much stronger degree than hyenas with dens in Soysambu. However, fine scale selection against vegetation greenness and toward verified livestock predation locations outside of the park, as well as known fence-crossing behaviors by LNNP hyenas (Wilkinson et al. 2021b), point to a lack of sufficient resources or space in the national park.