Inference of leadership networks of marine megafauna from acoustic data

The study of social interactions within species has been a captivating subject in biology. These interactions dominate over various aspects of an organism’s life and profoundly influence the
ecosystems they inhabit.A particular significance lies in how social animals often engage in cooperative behaviors, such as collective hunting, communal parenting, and group defense, with direct implications for their survival and reproduction.
While sharks have historically been perceived as solitary predators of the oceans, recent marine research has challenged this notion, revealing intricate social behaviors in select species. Notably, the blacktip reef shark (Carcharhinus melanopterus) displays hierarchical social structures influenced by factors such as sex and size, particularly in the coral reefs of French Polynesia. Similarly, the grey reef shark (Carcharhinus amblyrhynchos) exhibits dominance-submission dynamics, especially concerning their habitat.

To investigate these complex social interactions, complex systems theory is employed, exploring systems with numerous interconnected elements that give rise to collective emergent behaviors. The research focuses on two shark species, constructing networks to delineate leader-follower dynamics across multiple observation stations using acoustic data from the integrated Marine Observing System (IMOS). Aggregated networks reveal a structure closely resembling a hierarchical Directed Acyclic Graph (DAG), achieved through the systematic removal of weak interactions. Comparative analysis against randomized networks highlights the observed networks’ proximity to a perfect hierarchy.

Results encompass event data analysis, revealing appearance probabilities and circadian patterns, network inference, including leadership networks, leadership and participation indices, and aggregated leadership networks. The study also investigates directed acyclic graphs (DAGs), strategies for breaking cycles, and the comparison of observed networks to random configuration networks. This comprehensive research successfully unveils the hierarchical dynamics within social interactions among shark species, particularly focusing on the blacktip reef shark and the grey reef shark. The study not only achieves an understanding of these hierarchies but also lays the groundwork for further exploration, offering valuable insights into the intricate social behavior of these marine creatures.