Uncovering behavioural states from animal activity and site fidelity patterns. Issue 2 (10th October 2016)

• Record Type: Journal Article
• Authors: Mahoney, Peter J.; Young, Julie K.
• Editors: Parrini, Francesca
• Is Part Of: Methods in ecology and evolution. Volume 8:Issue 2(2017)
• Page Start: 174
• Page End: 183
• Publication Date: 2016-10-10
• Subjects: accelerometer -- clustering -- coyote -- denning -- GPS -- home range -- movement -- predation -- puma -- territoriality; Ecology -- Periodicals; Evolution -- Periodicals; 577
• DOI: 10.1111/2041-210X.12658
• Languages: English
• ISSNs: 2041-210X

Abstract

Summary: Space use by animals has important implications for individual fitness. However, resource requirements often vary throughout the course of a lifetime and are a reflection of the demands associated with daily tasks or specific life‐history phases, from food acquisition to reproduction, and emphasize the need to classify resource selection relative to specific behavioural states. Site fidelity is often indicative of behaviours important for individual maintenance (e.g. foraging), species' life history (e.g. seasonal site selection), social communication (e.g. scent‐marking) and species interactions (e.g. predation, competition). Thus, resolving site fidelity patterns associated with key behaviours is essential to accurately quantify behavioural‐dependent resource needs and the fitness consequences of space use. We propose a novel method for identifying site fidelity patterns in animal location data using a convex hull clustering program called R Animal Site Fidelity (rASF). We also provide a means of integrating activity as a measure of behavioural state. We demonstrate the utility of the approach in identifying cougar ( Puma concolor ) predation events, coyote ( Canis latrans ) den and rendezvous sites, and coyote territorial boundaries. We parameterized rASF based on site fidelity characteristics that best characterized the clustering behaviour of interest and estimated behavioural state from either dual‐axial accelerometer data or movement trajectory statistics. When behaviour was used in conjunction with cluster‐specific metrics (duration, proportion of diurnal fixes and landscape composition), we could accurately predict prey species associated with cougar kills and differentiate pup‐rearing from scent‐marking sites in coyotes. Site fidelity patterns and activities associated with animal revisitation will be key to identifying the behavioural motivations behind observed patterns of space use. Our approach provides an efficient, rigorous and repeatable means of identifying site fidelity patterns associated with specific behavioural states without the need for direct observations, which are often impossible to collect at large spatial scales and in dense habitat. As such, this framework has significant potential to inform theory in behavioural ecology while providing managers with better resolution on appropriate management targets associated with key aspects of a species' life history.