Is pre-breeding prospecting behaviour affected by snow cover in the irruptive snowy owl? A test using state-space modelling and environmental data annotated via Movebank
Résumé
Background: Tracking individual animals using satellite telemetry has improved our understanding of animal
movements considerably. Nonetheless, thorough statistical treatment of Argos datasets is often jeopardized by their
coarse temporal resolution. State-space modelling can circumvent some of the inherent limitations of Argos datasets,
such as the limited temporal resolution of locations and the lack of information pertaining to the behavioural state of
the tracked individuals at each location. We coupled state-space modelling with environmental characterisation of
modelled locations on a 3-year Argos dataset of 9 breeding snowy owls to assess whether searching behaviour for
breeding sites was affected by snow cover and depth in an arctic predator that shows a lack of breeding site fidelity.
Results: The state-space modelling approach allowed the discrimination of two behavioural states (searching and
moving) during pre-breeding movements. Tracked snowy owls constantly switched from moving to searching behaviour
during pre-breeding movements from mid-March to early June. Searching events were more likely where snow cover
and depth was low. This suggests that snowy owls adapt their searching effort to environmental conditions encountered
along their path.
Conclusions: This modelling technique increases our understanding of movement ecology and behavioural decisions of
individual animals both locally and globally according to environmental variables.