Patterson, Brent R

Animal migration is defined as the seasonal movement from one independent and non-overlapping range to another. Understanding how and why animals migrate is important not only to understand their life history processes but also for informing other important ecological processes such as the spread of wildlife disease and habitat alteration. Animal migrations have been impacted by human activity with instances of complete loss of migrations in human-altered areas. Understanding the drivers of migration can help predict responses to future environmental changes and potentially help conserve these phenomena. Seasonal movements of white-tailed deer (deer; Odocoileus virginianus; Zimmerman, 1780) have been linked to seasonal changes in environmental conditions that impact their ability to find food resources and risk of predation. The human shield hypothesis posits that prey species will select habitat close to people to use predator fear of humans to protect themselves from predation. Using global positioning system (GPS) collars, we examined the onset of deer migrations and assessed how environmental variables including snow, temperature, and plant biomass influenced migration departure dates using time-to-event models. We compared deer locations to data from GPS collared coyotes (Canis latrans; Say, 1823) within the same study area to explore daily space-use differences and examine if deer migrations were food or predation-risk driven using generalized linear mixed effects regression models. We found substantial annual and individual variation in deer migration dates. Snow depth was the strongest and most consistent predictor of deer migration, with individuals departing earlier with greater snow depth. Our regression analyses showed that deer selected for habitats closer to and with greater density of anthropogenic structures than coyotes at all times. After removing the animal locations close to areas with active supplemental feeding, these effects were diminished showing no differences in proximity or density of structures. Overall, we found more support for a food driven migration rather than a predator driven human shield. With the reduction in natural food caused by snow cover, we suggest that supplemental feeding is likely influencing the use of wintering areas by deer. The high proportion of deer migrating to human developed areas with supplemental feeders highlights the need for continued research into the impacts of human activity on animal behaviour.

Author Keywords: coyote, human shield, migration, supplemental feeding, white-tailed deer

Quantifying the impacts of environmental conditions on the abundance of wildlife populations is important for making informed management decisions in the face of increasing environmental threats. Managers require robust tools to estimate abundance and density of wildlife rapidly and with precision. Within the context of studying white-tailed deer, I evaluated the use of camera-traps and a recently developed spatial-mark resight model to estimate deer density and evaluate habitat and land use factors influencing deer density. The study was conducted in central Ontario, Canada on approximately 16 km2 of public land including the protected Peterborough Crown Game Preserve. Telemetry locations from 39 radio-collared deer were used and one hundred camera-traps were deployed for a total of 140 days from January 2022 to May 2022. Using telemetry locations and camera-trap photos I built a two-step spatial-mark resight model to estimate deer density. Deer density varied during the study as a portion of the population migrated to wintering areas outside of the study area. Despite fluctuations in precision, estimates improved towards the end of the study as more data became available and deer space use stabilized. The average deer density during the entire study was 3.0 deer/km2 (95% CI= 0.1, 5.8; SD= 1.7; CV= 55%; N= 238 deer). The lowest mean density was 0.2 deer/km2 (95% CI= 0.1, 0.4; SD= 0.1; CV= 50%; N= 15 deer) from February 26th to March 11th and the highest mean density was 4.8 deer/km2 (95% CI= 3.1, 6.2; SD= 0.8; CV= 17%; N= 378 deer) from May 7th to May 20th. When I incorporated spatial covariates into the model to estimate effects on deer density, higher proportions of mixed forest, deciduous forest, and road and trail density all had negative effects on deer density. While models contained some uncertainty, deer density appeared higher in the portion of the study area protected from licensed hunting. This thesis provides a framework for managers to use camera-traps and the spatial-mark resight model to monitor deer populations and link environmental covariates to spatial variation in density. As environmental threats such as habitat loss and infectious diseases increase in severity, monitoring wildlife population numbers will be vital for informed responses to these threats. The two-step spatial-mark resight model with environmental covariates provides managers with a long-term monitoring tool to evaluate management efforts and population health in forested areas.

Author Keywords: camera-trap, chronic wasting disease, landscape ecology, spatial-capture recapture, white-tailed deer, wildlife management

Eastern wolves (Canis lycaon) have hybridized extensively with coyotes (C. latrans) and gray wolves (C. lupus) and are listed as a `species of special concern' in Canada. Previous studies have not linked genetic analysis with field data to investigate the mechanisms underlying Canis hybridization. Accordingly, I studied genetics, morphology, mortality, and behavior of wolves, coyotes, and hybrids in and adjacent to Algonquin Provincial Park (APP), Ontario. I documented 3 genetically distinct Canis types within the APP region that also differed morphologically, corresponding to putative gray wolves, eastern wolves, and coyotes. I also documented a substantial number of hybrids (36%) that exhibited intermediate morphology relative to parental types. I found that individuals with greater wolf ancestry occupied areas of higher moose density and fewer roads. Next, I studied intrinsic and extrinsic factors influencing survival and cause-specific mortality of canids in the hybrid zone. I found that survival was poor and harvest mortality was high for eastern wolves in areas adjacent to APP compared with other sympatric Canis types outside of APP and eastern wolves within APP. Contrary to previous studies of wolves and coyotes elsewhere, I hypothesized that all Canis types exhibit a high degree of spatial segregation in the Ontario hybrid zone. My hypothesis was supported as home range overlap and shared space use between neighboring Canis packs of all ancestry classes were low. Territoriality among Canis may increase the likelihood of eastern wolves joining coyote and hybrid packs and exacerbate hybridization. Canids outside APP modified their use of roads between night and day strongly at high road densities (selecting roads more at night), whereas they responded weakly at lower road densities (generally no selection). Individuals that survived exhibited a highly significant relationship between the difference in their night and day selection of roads and availability of roads, whereas those that died showed a weaker, non-significant response. My results suggest that canids in the unprotected landscape outside APP must balance trade-offs between exploiting benefits associated with secondary roads while mitigating risk of human-caused mortality. Overall, my results suggest that the distinct eastern wolf population of APP is unlikely to expand numerically and/or geographically under current environmental conditions and management regulations. If expansion of the APP eastern wolf population (numerically and in terms of its geographic distribution) is a conservation priority for Canada and Ontario, additional harvest protection in areas outside of APP may be required. If additional harvest protection is enacted, a detailed study within the new areas of protection would be important to document specific effects on eastern wolf population growth.

Author Keywords: Canis, coyotes, eastern wolves, hybridization, resource selection, survival

The primary method used to maintain white-tailed deer (Odocoileus virginianus) populations at densities that are ecologically, economically, socially, and culturally sustainable is hunter harvest. This method considers only the removal of animals from the population (the direct effect) and does not conventionally consider the costs imposed on deer as they adopt hunter avoidance strategies (the risk effect). The impact of risk effects on prey can exceed that of direct effects and there is interest in applying this concept to wildlife management. Deer are potential candidates as they have demonstrated behavioural responses to hunters. I explored the potential of such a management practice by quantifying how human decisions around hunting create a landscape of fear for deer and how deer alter their space use and behaviour in response. I used a social survey to explore the attitudes of rural landowners in southern and eastern Ontario towards deer and deer hunting to understand why landowners limited hunting on their property. I used GPS tracking devices to quantify habitat selection by hunters and hunting dogs (Canis familiaris) to better understand the distribution of hunting effort across the landscape. I used GPS collars to quantify the habitat selection of deer as they responded to this hunting pressure. I used trail cameras to quantify a fine-scale behavioural response, vigilance, by deer in areas with and without hunting. Human actions created a highly heterogeneous landscape of fear for deer. Landowner decisions excluded hunters from over half of the rural and exurban landscape in southern and eastern Ontario, a pattern predicted by landowner hunting participation and not landcover composition. Hunter decisions on whether to hunt with or without dogs resulted in dramatically different distributions of hunting effort across the landscape. Deer showed a high degree of behavioural plasticity and, rather than adopting uniform hunter avoidance strategies, tailored their response to the local conditions. The incorporation of risk effects into white-tailed deer management is feasible and could be done by capitalizing on a better understanding of deer behaviour to improve current management practices or by designing targeted hunting practices to elicit a landscape of fear with specific management objectives.

Author Keywords: Brownian bridge movement models, hunting, landscape of fear, resource utilization functions, risk effects, white-tailed deer

I investigated if spatio-temporal behaviour of grey wolves (Canis lupus) determined via GPS collar locations could be used to discriminate predation events generally, and among prey species, in Prince Albert National Park during winter, 2013-2017. I used characteristics of spatio-temporal GPS clusters to develop a predictive mixed-effect logistic regression model of which spatial clusters of locations were wolf kill sites. The model suffered a 60 % omission error when tested with reserved data due to the prevalence of deer kills with correspondingly low handling time. Next, I found a multivariate difference in the percentage of habitat classes used by wolves in the 2 hours preceding predation events of different prey species, suggesting that wolf habitat use reflects prey selection at a fine-scale. My results highlight the difficulty and future potential for remoting discriminating wolf predation events via GPS collar locations in multi-prey ecosystems.

Author Keywords: Canis lupus, GPS clusters, GPS collars, grey wolf, habitat use, predation

My objective was to understand how individual variation, in conjunction with variation in habitat, can affect individual and population-level variation in animal space use. I used coyotes (Canis latrans) as a model species to investigate the roles of hybridization, an inherited intrinsic factor, and spatial memory, a learned intrinsic factor, on space use. I used a diversity of methods and approaches, including meta-regression, multiple imputation, simulations, resource selection functions, step selection functions, net-squared displacement analysis, and survival analysis. A major contribution was my investigation of the performance of multiple imputation in a meta-regression framework in Chapter 2. My simulations indicated that multiple imputation performs well in estimating missing data within a meta-regression framework in most situations. In Chapter 3, I used published studies of coyote home range size in a meta-regression analysis with multiple imputation to examine the relative roles of hybridization and environmental variables on coyote home range size across North America. I found that hybridization with Canis species was a leading factor driving variation in coyote space use at a continental scale. In Chapter 4, I used telemetry data for 62 coyotes in Newfoundland, Canada, to investigate the influence of cognitive maps on resource use. I found that resident coyotes used spatial memory of the landscape to select or avoid resources at spatial scales beyond their immediate sensory perception relative to transient coyotes, presumably increasing their fitness. Taken together, my dissertation demonstrates that intrinsic factors, such as genetic ancestry and spatial memory, can have substantial influences on how animals use space at both individual and population levels, and at both a local and a continental scales.

Author Keywords: canis latrans, hybridization, meta-regression, multiple imputation, Newfoundland, spatial memory

Livestock predation by wild predators is a frequent and complicated issue, often cited as a significant factor in the decline of livestock production and justification for killing predators. Coyotes (Canis latrans) are the primary predators of sheep in Ontario. Some farms appear to be more susceptible to predation than others, despite the use of mitigation techniques. I explored land cover in the vicinity of farms as a potential influence on the level of predation, as coyote abundance and wild prey are correlated with certain habitat types. Using model competition, I show that landscape explains little variation in predation levels over all farms, but can explain 27% of variation in the percent of a flock killed. Total forest edge habitat and distance between forest patches were both positively associated with losses, suggesting a reduction in forest cover surrounding a farm puts the flock at greater risk. In addition, I tested four disruptive deterrents for effectiveness at protecting flocks. A matched-pairs analysis did not show a statistically significant benefit of these non-lethal mitigation tools.

Author Keywords: Alternative Prey, Canis latrans, Coyote, Landscape, Predation Deterrents, Sheep