Behavioural responses to resource availability by northern ungulates
This thesis is an investigation of ungulate distribution at multiple levels of ecological organization as related to nutritional tradeoffs made between forage quantity and quality and to movement processes, specifically by wood bison ('Bison bison athabascae') and caribou (' Rangifier tarandus'). First, I examined the effect of vegetation structure on short-term feeding rates of bison grazing 'Carex atherodes'. Sward height affected bison selectivity, but not their functional response. In swards composed of both stem and leaf, intake rates were significantly depressed at low biomass, compared with those obtained on simpler leaf-only swards. Secondly, long-term intake of sedge by bison declined with sedge digestibility and thus decreased with sedge biomass. Models based on dual food intake constraints predicted that the optimal sward biomass for time minimizing and energy maximizing strategies was 10 g/m2, and 279 g/m2, respectively. Observations of captive animals grazing in artificially constructed mosaics of vegetation patches indicated that bison behaved as time minimizers. Thirdly, I examined the relationship between multi-scale habitat selection by bison and food intake constraints. Patch use at three spatial scales was non-random. At intermediate spatial scales, patch use was strongly correlated with mean patch digestibility, but not with expected cropping rate. At the smallest scale, the same was true for digestibility, but cropping rate was negatively correlated with patch use. Thus, it appears that digestive constraints have a stronger influence than do cropping constraints on patch use by free ranging bison. Lastly, I examined the utility of a correlated random walk model in describing landscape level movements of caribou. For paths recorded through one annual cycle, the model overpredicted net displacement of caribou through time. For shorter seasonal paths. there was excellent correspondence between model predictions and observations. When predictions and observations were not in agreement, autocorrelational analysis of turn directions showed them to be non-random in a manner expected from behavioural observations during the periods in question. Such a model of movement could be linked with resource use models in the development of more sophisticated portrayals of plant-herbivore interactions.