Polar bears (Ursus maritimus) enter a period of intensified feeding in the spring, which allows for the accumulation of energy stores critical to surviving the open water season. Study on polar bear predation has been limited by sample size and spatial extent, and hypotheses on the demographic composition of seal kills and the spatial distribution of polar bears and seals were incongruent. In this thesis, I used a long-term dataset (1985-2011) of seals killed by polar bears (n = 650) and predation attempts at ringed seal (Pusa hispida) subnivean lairs (n = 1396) in the Beaufort Sea, Canada, to link the habitats polar bears use and the seals that polar bears kill during hyperphagia. Using DNA and field observations, I determined that polar bears primarily killed ringed seals, but that bearded seals (Erignathus barbatus) contributed a significant portion of kill biomass. An increase in seal kill frequency was observed temporally over the spring, associated with the onset of ringed seal whelping. The influence of ringed seal whelping was also observable at inter-annual scales, with total kill frequency positively correlated to years of high ringed seal natality, while adults were killed in higher proportion in years when natality was low. Employing locations of seal kills and attempted hunts at ringed seal subnivean lairs, I examined the habitats in which polar bears hunt and ringed seals whelp. Polar bears selected for active areas of sea ice near the floe edge when hunting seals. Ringed seal whelping areas were located over a range of habitats, and the distribution was correlated with natality. In years of low natality pup kills were observed primarily in shorefast ice close to land, but during years of high natality the distribution widened, and pup kills were observed farther from land and more frequently near active ice areas. Results suggest that during periods of high natality, the habitats in which ringed seals whelp overlaps with areas preferred by polar bears for hunting. The spatial overlap between polar bears and whelping ringed seals likely influences a change in the age-class proportions of kills, as polar bears respond to the availability of vulnerable pups. Finally, I explored the assumptions of common analytical modelling approaches in ecology. I established that including biologically relevant measures, such as the size of kills, provided significant improvement to the models in both fit and interpretation. Measuring only the occurrence of an ecological event, whether temporally or spatially, was found to be insufficient when validated against independent data. The empirical analyses within this dissertation suggest that strong assumptions of ecological models may not always hold. Collecting biologically relevant data in the field, beyond simply recording events, can test model assumptions and validate results, increasing model portability and the relevance of the findings.

The ecological response of climate change is expected to be especially pronounced across the circumpolar Arctic. Predicted declines in sea ice extent and seasonal duration are expected to affect the foraging ecology of marine species, particularly polar bears (Ursus maritimus) that rely on stable sea ice patterns for various aspects of their life history. The goal of this thesis was to quantify body condition and characterize diet composition of polar bears in western Hudson Bay over a broad temporal scale (2004-2014), and identify potential environmental factors that may influence dietary shifts in one of the southernmost subpopulations of this species. Body condition was higher in adult and subadult females than males, consistent with energetic demands of gestation and lactation. Body condition also declined over time in adult and subadult males and females and was influenced by sea ice breakup and freeze-up dates. These trends suggest that the historical climate-driven declines in polar bear body condition documented in western Hudson Bay have continued. Variation in diet composition and dietary niche breadth across age, sex and reproductive groups suggest foraging behaviour is structured by energetic demands, intraspecific competition and sexual body-size dimorphism. Specifically, variation in diet and niche breadth across females was influenced by age (experience), energetic state and avoidance behaviours. Variation in diet composition and niche breadth between male and female bears, however, was more likely structured by body size, whereby capture of larger prey types and a broader range of prey species occurred with increasing body size. Body condition was positively related to niche breadth in adult males but negatively related to niche breadth in females with dependents, suggesting that less-selective foraging (scavenging) does not benefit body condition among reproductive females. Inter-annual fluctuations in diet composition reflected shifts in local prey availability during the study period, and sea ice breakup date influenced the diets of subadults and family groups, suggesting an increased sensitivity to sea ice conditions. Therefore, inter-annual variability in diet and declines in polar bear body condition likely reflect contemporaneous changes in sea ice availability and population demography, and have implications for the long-term conservation of this subpopulation.

Climate change has led to abrupt declines in sea ice over the past three decades. Polar bears (Ursus maritimus) rely on sea ice as their primary habitat to hunt marine mammal prey. Due to their position at the top of the Arctic marine food web, polar bear foraging patterns can provide insights on ecosystem structure and function both spatially and temporally. This thesis used quantitative fatty acid signature analysis (QFASA) to estimate the diets, and adipose tissue lipid content to estimate the body condition, of polar bears in three western Canadian Arctic subpopulations: Northern Beaufort Sea, Southern Beaufort Sea, and Viscount Melville Sound. This thesis identified spatial, temporal, and intraspecific variation in the relationships between sea ice conditions, and polar bear diet and body condition. Polar bears with the greatest ecological constraints on diet composition may be most vulnerable to climate-related changes in ice conditions and prey availability.

Recent declines in Arctic sea ice have been linked to reductions in body condition, survival, and population size of polar bears (Ursus maritimus ) in some regions of the Arctic. Reduced foraging opportunity is hypothesized as the primary mechanism of sea-ice linked declines, but increased energy expenditure, as a result of changing sea ice composition and extent may be an additional underlying mechanistic factor. Field metabolic rates of polar bears have been previously unknown and limited information has existed on foraging or other behaviors in polar bears while on the sea ice. Such data are increasingly needed to better understand polar bear habitat use and improve projections of the effects of climate change on global polar bear populations. The primary objectives of this work were to: 1) develop metrics to remotely quantify wild polar bear foraging rates and behaviors; 2) measure the energetic costs of resting, walking, and swimming in captive polar bears, to better estimate the energy demands of free-ranging polar bears; 3) measure the field metabolic rates, behaviors, and foraging rates of free-ranging polar bears on the spring sea ice; and 4) evaluate the ability of tri-axial accelerometers to measure the energy expenditure of free-ranging polar bears. In this dissertation, I develop methods using tri-axial accelerometers to remotely discriminate the behaviors (Chapter 1) and index the energy expenditure (Chapter 5) of free-ranging polar bears. I further find that polar bears have energetic costs of walking that are commensurate with other quadrupedal mammals (Chapter 2), but appear to have high energetic costs of swimming similar to other semi-aquatic mammals (Chapter 3). Additionally, similar to other large carnivores, polar bears have high resting metabolic rates and high field metabolic rates (Chapter 4). Of nine polar bears monitored on the spring sea ice of the Beaufort Sea, five of these bears had energy deficits as a result of their high energy demands and low foraging success (Chapter 4). These findings highlight the physiological constraints of these large apex carnivores and reinforce their reliance on energy-dense pagophilic seal prey. Furthermore, these findings elucidate the implications of increases in the frequency of spring fasting on polar bear body composition. Given their large body size and hypercarnivory, polar bears appear to be poorly suited to cope with rapid environmental change. The metrics developed in this dissertation provide a framework to monitor polar bear behaviors and energy expenditure to aid conservation and management of the species.