2026/07/21 by Tim L. Hiller, Julia A. Nawrocki, Brandon M. Bernhardt +4
Environmental Science · Veterinary · #Wildlife Ecology and Conservation #Wildlife-Road Interactions and Conservation #Animal Behavior and Welfare Studies
paper · doi:10.1002/wmon.70019
Abstract Several indices and studies have indicated declining populations of gray foxes ( Urocyon cinereoargenteus ) in several Midwest states in the United States. Causes of these declines are unknown and may vary by region but could be linked to habitat loss and fragmentation, changing mesocarnivore communities, disease, and other complex factors. During 2020–2023, we collected data (e.g., location, biological samples, reproductive tracts [females only]) from 26 global positioning system (GPS)‐collared gray foxes in central and southern Indiana, and data (e.g., biological samples) from 74 gray fox carcasses statewide. We estimated seasonal space use, cover selection, and resource selection of radiomarked gray foxes based on reproductive (Jan–Jun) and nonreproductive (Jul–Dec) seasons. To assess space use, we used a biased random bridge approach to estimate 95% utilization distributions (UDs), 30% intensity distributions (IDs; an index of the intensity of resource exploitation of areas), and 30% recursive distributions (RDs; rate at which resources are exploited) based on reproductive ( n = 21 foxes) and nonreproductive ( n = 23) seasons. Patterns of seasonal space use consistently included larger areas during the reproductive season (UD = 871.2 ha vs. 822.2 ha), although 95% confidence limits (CLs) overlapped. We used a use‐availability approach to assess second‐ (landscape scale) and third‐order (home‐range scale) cover selection. Calculations of selection ratios (proportional use/proportional availability) by season suggested weak patterns for landscape‐level selection, including no selection for any individual cover type and selection against agriculture‐row crop and open water, regardless of season. Patterns were somewhat stronger for home‐range‐level selection; frequency of visitation and relative time of residence were both positively associated with several cover types (e.g., upland shrubland or forest, developed‐low intensity, developed‐roads) and negatively associated with agriculture‐row crop. To assess resource selection, we used infinitely weighted logistic regression, constructed models using anthropogenic (e.g., distance to nearest developed cover type and building, density of buildings, traffic volume) and environmental (e.g., cover type, landscape connectivity, weather) covariates, and used Bayesian Information Criterion (BIC) to assess model performance. Probability of use was greatest in areas near agricultural cover (strongest relative effect), buildings, and roads, regardless of season, and probabilities substantially decreased >100 m from each, suggesting the potential importance of habitat edge to meet life requisites associated with cover and forage. For assessments of genetics, we extracted DNA from 102 individuals and genotyped each at approximately 1,700 single nucleotide polymorphisms (SNPs) using 2bRADseq. We identified 10 pairs of first‐order relatives (parent‐offspring or full siblings), 7 of which had precise location data, with inter‐individual distances ranging from 0.4 km to 13.2 km (mean = 6.1 km), consistent with the known dispersal distances of gray foxes. We assessed population genetic structure using clustering (STRUCTURE) and principal components analysis (PCA). The STRUCTURE analyses indicated weak genetic structuring, with statistical support for the optimal number of genetic clusters ( K ) = 4; however, individuals largely clustered into 2 groups with substantial admixture, and PCA revealed considerable overlap, suggesting only subtle differentiation rather than clearly distinct populations. To evaluate potential landscape barriers to gene flow, we generated landscape‐resistance surfaces based on the 3 best‐performing species distribution models (forest, refugia, and foraging) constructed using a maximum entropy approach and presence‐only data from community‐based observations analyzed during a different component of our study. We analyzed the relationship between genetic distance and geographic or resistance distances using linear mixed‐effects models with a maximum likelihood population effects structure. Model comparisons indicated that geographic distance best explained variation in genetic distance, and no landscape‐resistance model outperformed isolation by distance. These results suggest minimal barriers to gene flow within Indiana, although our sample distribution was heavily skewed toward south‐central Indiana. We classified mortalities from 13 GPS‐marked foxes as disease (46%), vehicle collision (15%), legal harvest (8%), potential poisoning (e.g., rodenticide, oxalate nephrosis due to ethylene glycol; 15%), and unknown (15%). Based on proportional‐hazards models, survival did not differ by sex, and mean annual survival probability was 0.61 ( n = 26). We used motion‐sensing cameras to monitor natal dens ( n = 6) of radiomarked foxes to confirm presence of pups and date of den abandonment, after which we measured den‐site characteristics. Two dens were closely associated with human‐built structures, and 4 dens were underground burrows. Compared to random locations, gray foxes selected den sites with a greater amount of horizontal cover at heights 1–3 m aboveground. Lastly, our examination of placental scars in reproductive tracts of females resulted in estimated mean litter sizes of 2.00 for females 1 to <2 years old and 3.33 for females ≥2 years old. Our results provide evidence that gray foxes in Indiana are an edge species with relatively large home ranges, relatively low reproductive rates, and good genetic health with no identifiable genetic barriers to dispersal, but they are also subject to relatively high exposure to disease, specifically canine distemper virus. Risk associated with increased movements may be somewhat mitigated through use of habitat edges for cover (e.g., predator avoidance, escape cover), but large movements may also increase exposure to disease vectors (e.g., ticks). Management considerations may include providing small patches designed to serve as buffer zones to increase amount of edge for foraging, and potentially also for use as travel corridors, particularly in areas with extensive agriculture. It may be possible to create edge as an ecological homologue to roads but with a lower potential risk of mortality via vehicle collisions. Conversely, risk could increase if use of such areas increases interactions with certain other wildlife species (e.g., coyotes [ Canis latrans ], northern raccoons [ Procyon lotor ]) that may increase levels of interspecific killing or transmission of diseases.