Putting the health in hidden Markov models: incorporating allostatic load indices into movement ecology analyses

Individual animal health assessments are a key consideration for conservation initiatives. Environmental shifts associated with climate change, such as documented rises in pathogen emergence, predation pressures and human activities, create an increasingly stressful world for many species and have been linked with marked changes in movement behaviour. Even in healthy individuals, variations in allostatic load, the cumulative effects of long-term stress, may alter behavioural priorities over time. Here, we aimed to build links between animal health assessment information and movement ecology, using narwhals in the Canadian Arctic as a case study. A composite stress index was developed to incorporate multiple available health (e.g. health assessments), stress (e.g. hormones) and body condition metrics from clinically healthy individuals, and applied within the framework of widely used hidden Markov modelling of animal movement data. Individuals with a higher composite stress index tended to prioritize behaviours indicative of a stress response, including increasing the probability of transitioning to transiting behaviour as compared to those with a lower stress index. By incorporating a composite stress index that synthesizes multiple health indices in a flexible framework, we highlight that including information indicative of allostatic load may be important in explaining variation in behaviour, even for seemingly healthy animals. The modelling framework presented here highlights a flexible approach to incorporate health assessment information and provides an approach that is widely applicable to existing and future work on a range of species.