Trait-dependent diversification in island biogeography

Species traits are widely hypothesized to be important drivers shaping diversification rates and biodiversity patterns, but have been neglected in the current island biogeography models. This thesis focuses on providing a new framework to investigate how species traits affect macroevolutionary rates and biodiversity on islands, and advance the use of Bayesian inference methods in both island biogeography and trait-dependent diversification models. First, in chapter 2, I introduce a trait-based simulation model in island biogeography and use a robustness pipeline to test whether ignoring traits in island biogeography models can be problematic in estimating parameters and reconstructing the phylogenies. We find that ignoring traits may not affect model accuracy for most typical islands. However, inference errors occur when fitting phylogenetic data with large clade size variations in an island community, which calls for the development of a trait-dependent inference model. Chapters 3 and 4 are designed to provide key steps towards achieving the development of the new framework. These two chapters highlight the power of approximate Bayesian inference (ABC) methods in estimating diversification rates, and provide a set of informative summary statistics in island biogeography or trait-dependent diversification studies. Through these chapters, I hope to provide an opportunity for a better understanding of the spectacular biodiversity patterns in insular systems.