PhD defence Ornela Dehayem Nanwou
Ornela Dehayem Nanwou (TRÊS)
Promotores: Prof. R.S. Etienne, Prof. F. Lens (LU), Prof. L.M. Lima Valente (TRÊS/Naturalis)
Diversification in isolation
Traits, sampling, and inference in island biogeography
Species richness is heterogeneously distributed across the tree of life, for both islands and continents. Trait-associated differences in diversification rates are frequently proposed as the main drivers of these disparities among clades. State-dependent Speciation and Extinction (SSE) models have been developed to test this hypothesis across the tree of life; however, this has not yet been tested on island communities, which have a different data structure to those required by standard SSE models.
In this thesis, I developed a framework tailored to island communities to investigate how species traits influence colonisation and diversification, with the aim of explaining disparities in species richness observed among lineages in island communities. I applied this framework applied to the Canary Islands angiosperm community to test whether angiosperm diversification rates (speciation and extinction) and/or transition dynamics depend on the growth form (herbaceous vs woody) or life span (annual vs perennial) or both.
We found that woody vs herbaceous lineages and perennial vs annual lineages follow different evolutionary trajectories on the Canaries. Woody lineages exhibit higher cladogenesis rates and lower extinction rates than herbaceous lineages, an overall faster diversification. In addition, life span influences both speciation and extinction dynamics among herbaceous lineages: perennial lineages generally show higher cladogenesis rates than annual lineages, while extinction rates are lowest in perennial herbaceous groups. However, a concealed trait dependent model (CTD model) is preferred over the examined trait dependent model (ETD model). This indicates that growth form and life span do not fully explain diversification patterns, pointing to an important role for unobserved factors.