A first multi-sensor gear-based characterisation of the behaviour of target and bycatch species during pelagic longline capture

Little is known about the behaviour and stress responses of animals captured by longline gear from hooking until retrieval at the vessel, despite the importance of this process for bycatch mitigation and post-release survival. Most existing studies have relied on physiological sampling at haulback, which provides only a snapshot of stress at retrieval and requires invasive handling. This study used a custom multi-sensor data logger, the SMARTSNAP device, attached directly to the longline to continuously record the behaviour of individuals captured on commercial pelagic longlines in the Gulf of Lion and around Réunion Island between 2022 and 2023. A total of 40 individuals across 15 species, including target tuna species and elasmobranch bycatch, were recorded from hooking until retrieval. Within active capture phases, individual dive sequences were identified and characterised through metrics describing dive duration, depth, descent speed, and directional movement. Atlantic bluefin tuna showed a marked decline in fight intensity over successive dives, whereas blue shark and pelagic stingray displayed more stable patterns. Principal component analysis revealed that individuals which ultimately died following capture were characterised by elevated movement and depth metrics, while surviving individuals showed comparatively lower values. Similar mortality-associated behaviours were observed in other tuna species. These findings support the hypothesis that behavioural dynamics can serve as a proxy of capture-induced exhaustion, demonstrating the potential of gear-based instrumentation to directly assess animal condition during capture. Further work is needed to determine whether these observations can be translated into practical applications for bycatch mitigation, rapid release strategies, and the management of capture-related stress, with potential implications for animal welfare and flesh quality.