Big data for a large clade: Bioregionalization and ancestral range estimation in the daisy family (Asteraceae)

Aim In recent years biogeography has been transformed by the increased availability of large‐scale distributional data, phylogenies, and novel quantitative analysis methods and models. More case studies, however, are needed to test the performance of various approaches, in particular at global scales and in species‐rich groups. In this study, we inferred bioregionalization and estimated ancestral areas for the largest plant family, the Asteraceae. Location Global. Methods We used the Global Compositae Checklist data to infer Asteraceae bioregions with cluster and modularity analysis. We reconstructed a phylogeny of genus‐terminals for the Asteraceae family from a supermatrix of nuclear ribosomal internal transcribed spacer and chloroplast data. Combining areas based on the bioregions from modularity analysis and the phylogeny, we then estimated ancestral ranges across the Asteraceae phylogeny under 12 biogeographic models. Results Cluster analysis resulted in several small bioregions from areas with low taxon numbers and linear and disjunct bioregions between Eurasia and Africa. Modularity analysis produced larger and compact bioregions, and we based downstream analysis on its results. The favoured model for ancestral area estimation was BAYAREALIKE+j+x, demonstrating the importance of long distance dispersal in the biogeographic history of the Asteraceae and a strong distance‐dependence of dispersal. Main conclusions Differences between cluster and modularity analysis suggest that the latter may be more robust to incomplete data and produces less disjunct and thus presumably biologically more realistic bioregions. With few exceptions, results of ancestral area estimation confirmed the results of previous studies, in particular South America as the ancestral area of the family, subsequent dispersal to and a secondary radiation from Africa, and the ancestral areas of individual tribes of the family.