Abstract
Behavioral ecology aims at characterizing animal behavior in relation to their physical and biological environment. A key aspect of this field is assessing the causation and consequences of behaviors. When manipulative experiments are technically challenging to implement, alternative approaches must be developed to investigate these aspects. In this study, we evaluate the effectiveness of mechanistic modelling, combined with correlative approaches on empirical data, to determine behavior causation and consequences. Many pelagic fish species, such as tropical tunas, display an associative behavior with floating objects. Although several studies suggest that that tunas have a lower relative condition when associated with DFADs, the causal link between the two remains undetermined. We develop a behavioral model to investigate the relationship between the associative dynamics of tropical tunas with DFADs and their physiological condition. We consider two hypotheses: H1 that tuna physiological condition decreases when they are associated with DFADs (condition as a consequence of associative behavior), and H2 that tuna tend to associate more when they are in lower condition (condition as a causation of associative behavior). Using bio-electrical impedance analysis data of associated yellowfin tuna at different DFAD densities in the western Indian Ocean, we then show that the lower condition observed for this species is a consequence of its associative behavior. This study demonstrates the relevance of combining mechanistic modelling with correlative approaches when studying behavior in cases where experiments are hard to implement. The use of such approaches rooted in conceptual frameworks allows a better characterization of animal behavior causes and consequences at the relevant time scales.