The impact of anthropogenic threats can be behaviour-specific, and behavioural information could be incorporated into spatial management to improve conservation outcomes. We used multi-sensor biologging tags containing high-resolution movement sensors (e.g., accelerometer, magnetometer) and animal-borne video cameras, combined with supervised machine learning to automatically identify key behaviours for flatback turtles Natator depressus. Subsequently, we evaluated behaviour-specific spatiotemporal habitat use patterns.
Boosted regression trees identified foraging and resting during 7074 dives (AUC >0.9), using features describing locomotory activity, body posture, and three-dimensional dive paths validated by ancillary video footage. Foraging was characterised by dives with longer duration, variable depth, tortuous bottom phases; resting was characterised by dives with decreased locomotory activity and longer duration bottom phases. Foraging and resting utilisation distributions showed minimal spatial segregation. Expected diel patterns of behaviour-specific habitat use were superseded by the tides at the nearshore study site. Resting occurred within larger foraging areas in subtidal waters proximate to intertidal areas, allowing efficient access to intertidal food resources upon inundation at high tides when foraging was ~25% more likely.
Our generalisable approach shows how threat mitigation for flatback turtles can prioritise protection of important locations at pertinent times and facilitates a conceptual advance in dynamic spatial management.