ETD

• animal-borne
• feeding behaviour
• loggerhead turtles
• sensors
• videocamera

Understanding fine-scale underwater behaviour of marine turtles is a crucial step to fully comprehend their ecological role, spatio-temporal distribution and use of resources and habitat, and, ultimately, to implement effective conservation and management strategies. However, such information is scarce, due to the difficulties of observing turtles in their natural environment. In particular, feeding is a challenging research topic to address for the lack of adequate tools able to detect and monitor such a fundamental part in the life and ecology of every animal. Recently, the use of animal-borne video cameras, unveiled a new, very powerful method to indagate such fine-scale behaviours. In this study 21 free-ranging juvenile loggerhead turtles (CCL: 63.7 ± 1.1 cm) were equipped with a prototypical animal-borne multi-sensor loggers (named as DIVID) consisting of: i) tri-axial accelerometer, magnetometer and gyroscope, ii) time-depth recorder iii) temperature sensor and iv) an animal-borne video camera and released in the waters around Lampedusa. Through this device, and after the production of a personalized ethogram, the sea turtle underwater behaviour was investigated. Then, both pelagic and benthic feeding were analysed in detail, assessing frequency and duration of feeding events. The duration of pelagic feeding events resulted significantly shorter than benthic feeding events (p < 0.01) likely due to the type of prey consumed. Turtles showed a preference for hard substrate with algal cover over Posidonia meadows for benthic foraging (p < 0.05) for the possible higher abundance or accessibility of preferred benthic prey. GLMMs results evidenced association between feeding events observed with camera and sensors-derived parameters related to turtles’ position in water column (depth) or body movements (changes in dynamic acceleration or rotational movements). The present research evaluated the combined use of animal-borne multi-sensors loggers and cameras, highlighting their role as promising powerful tools to explore animal behaviours at 360 degrees. Moreover, the approach developed could be used in future studies to infer feeding events, as well as other behaviours, only from sensor data, bypassing camera’s limitations and can be applied to study different behaviours in many marine species of relevant interest.