Developing methods and applications for the analysis of cetacean social networks

Abstract

Cetaceans, the whales, dolphins, and porpoises, represent a taxon of intense interest for researchers studying non-human social structure. Social network analysis has become a central tool for studying these species, however the collection, analysis, and application of cetacean social network data comes with numerous challenges. In this thesis, I address key research gaps in the study of cetacean social networks, using the well-studied southern resident killer whale populations as my study system. In the first chapter, I present a systematic literature review on cetacean social networks, in order to identify open areas for future research and development. In Chapter 2, I address the question of social complexity and its quantification. Using mixture models, I develop and test measure of social complexity based on relationship diversity that can be derived from association networks. In Chapter 3, I demonstrate that a commonly used statistical procedure for regression in association networks does not specify a proper null hypothesis, and results in high type I error rates. In Chapter 4, I use unmanned aerial systems methods to measure association and interaction networks within a group of southern resident killer whales, finding important differences in the structure of these different networks. In Chapter 5, I use long-term photographic data to model the spread of a novel pathogen over the social network of the endangered southern resident killer whale community to assess overall risk and potential management strategies. In Chapter 6, I use a multi-decade dataset of social associations, survival, and fecundity to test the link between aspects of the social environment and fitness in the southern resident killer whale population. In the final chapter, I provide a general discussion and synthesis of my results, and suggest areas for future research, both generally and within the southern resident population specifically.