Land‐use change structures carnivore communities in remaining tallgrass prairie

Abstract

The Flint Hills ecoregion is the largest remaining tract of native tallgrass prairie in North America. Contemporary landscape change (e.g., urbanization, agricultural production) in this region is likely affecting native biodiversity; however, we have a limited understanding of how these changes might affect carnivores. We used camera traps distributed across urban‐rural land‐use gradients, and multiseason occupancy models, to investigate the influence of landscape structure and composition on habitat occupancy dynamics of 3 native carnivores (coyote [Canis latrans]; bobcat [Lynx rufus]; and striped skunk [Mephitis mephitis]) and 1 nonnative carnivore (domestic cat) in the Flint Hills, Kansas, USA, during 2016–2017. Additionally, we assessed the potential for coyotes, the apex predator in the Flint Hills, to mediate habitat occupancy of domestic cats. We also examined the relative influence of landscape factors on native carnivore species richness and diversity. As predicted, coyotes were less likely to occupy and colonize sites, and more likely to leave sites, surrounded by urban landcover. Habitat occupancy by bobcats was positively associated with forest landcover and edge densities; however, bobcats seemed insensitive to urban landcover. Additionally, bobcats were more likely to colonize sites with more grassland and row‐crop agriculture landcover. Surprisingly, striped skunk occupancy and colonization rates at sites were negatively related to urban landcover. As expected, domestic cats were more likely to occur at and colonize sites with increased urban landcover and less likely at sites with high coyote occupancy probabilities. Carnivore species diversity and richness were negatively related to the urban landcover. Our results suggest that urban landcover may limit the spatial distributions of some native carnivores in the Flint Hills and contribute to reduced carnivore diversity. Moreover, urban landcover in the Flint Hills may provide domestic cats with refugia from intraguild pressures by coyotes. Our results provide wildlife managers with insights into how contemporary landscape change in this imperiled ecosystem may affect biodiversity conservation and maintenance. © 2018 The Wildlife Society.

Mismatched spatial scales can limit the utility of citizen science data for estimating wildlife‐habitat relationships

“…However, the 95% CI around the occupancy estimates for each WMU overlapped, making it difficult to ascertain the factors responsible for geographic variation in participant‐level occupancy. Our participant‐level occupancy estimates are similar to those reported for coyotes elsewhere in the midwestern USA using camera traps (Berry, Schooley, & Ward, 2017; Lesmeister et al, 2015; Wait et al, 2018). We suggest that analyzing observational data from citizen scientists within a hierarchical modeling framework that accounts for imperfect detection is a better option for modeling the distribution of cryptic wildlife species compared to estimates of observation per unit effort or area sampled.…”

“…Due to the hierarchical nature of our sampling design (participants with multiple visits nested within counties), we initially modeled our data using a single-season multi-scale occupancy model, which estimates occupancy at both the sample unit scale (the county level) and at the sample station scale (the participant level) (Nichols et al, 2008). However, coyotes were detected in all but two counties in Illinois and preliminary analyses showed that estimated county-level occupancy was 1.00, making coyotes ubiquitous throughout Illinois at the county T A B L E 1 Select land cover proportions and weather data (mean and range) by wildlife management unit (WMU) (Lesmeister, Nielsen, Schauber, & Hellgren, 2015;Wait, Ricketts, & Ahlers, 2018) so will be more likely to occupy areas with high patch density.…”

“…Agricultural cover % county that is agriculture − Occupancy Coyote abundance is negatively associated with agriculturally-dominated landscapes (Cherry et al, 2016) Urbanization % county that is urban environment − Occupancy While coyotes can persist near urban areas, they will be less abundant and less likely to occupy these areas (Kays et al, 2008;Leflore, Fuller, Finn, DeStefano, & Organ, 2019;Lesmeister et al, 2015;Randa & Yunger, 2006;Wait et al, 2018;Wang et al, 2015).…”

Habitat selection and spatiotemporal interactions of a reintroduced mesocarnivore

“…We placed cameras on trees at approximately 0.5 m above the ground. We attached bait and lure at approximately 0.5–1 m above the ground on a second tree 2–4 m away from the camera (Wait et al ). We placed bait at this height to keep the bait in the frame while having a substantial portion of foreground visible.…”

Scales of Effect for Urban Mammals Vary Among Species and Environmental Attributes

“…Instead, a conventional modelling approach estimates all model covariates using a single buffer area surrounding study sites (typically denoted by the radius used, e.g., 100 m). For example, for medium to large-bodied urban mammals, 500 or 1000 m buffers are commonly employed (Fidino et al 2016, Wait et al 2018, Monterroso et al 2020. Such a choice is often rooted in expert-knowledge, for example, assuming that a 500 m buffer represents an average home range across many mesocarnivores based on past experience with these species , Moreira-Arce et al 2016).…”