Abstract
Wildlife conservation requires an understanding of the spatial and temporal distribution of animals. Seasonal environmental change, particularly for migratory species at high latitudes, can have important consequences for the distribution of animals within their known geographic range. This presents a challenge for conservation practitioners because seasonal patterns are shifting due to climate change and individual responses may vary within a population. Here, we modeled seasonal shifts in the behaviour of woodland caribou (Rangifer tarandus caribou, Atihkw) across two of Ontario’s putative ecotypes: eastern migratory and boreal. We used a k-means clustering approach to define seasons using individual caribou behavioural variables including movement speed, straightness, proportional habitat use, and distance from the Hudson and James Bay coasts. We compared how seasonal delineations differ between ecotypes and examined which behaviours are most important for delineating seasonal clusters. We found inter-annual variation in the number of seasons identified with our approach, both within and between ecotypes. Distance to Hudson/James Bay was the most important variable for differentiating seasonal clusters for both ecotypes, and eastern migratory caribou consistently minimized this distance during the calving season. For boreal caribou, distance to Hudson/James Bay did not display a clear seasonal pattern across years, and movement speed was sometimes important for identifying seasonal clusters. Our work quantifies the variability of caribou behaviour in space and time and highlights important considerations for predicting the responses of caribou to ongoing anthropogenic and environmental change.