Integrating processes, time, and patterns: The event-based approach to historical biogeography
2005 (English)In: The Systematics Association 5th Biennial Meeting, 2005Conference paper (Other scientific)
Traditional cladistic biogeographic methods search for concordance between phylogenetic and distribution patterns of different organisms as evidence of a common history of geological isolation events (vicariance). However, these methods are flawed because they only focus on “topological congruence”, overlooking other important sources of information. Biogeographic processes such as dispersal and extinction are not incorporated directly into the analysis, which makes cladistic results very difficult to interpret. Similarly, integration of timing of species divergences is needed if we want to discriminate between alternative biogeographic scenarios, or to exclude “pseudo-congruence”, i.e., congruent patterns that originated at different times. Finally, cladistic methods do not directly attempt to test the statistical significance of results. In this paper, I review recent developments on biogeographic analysis designed to overcome these problems, and illustrate them with case studies. Event-based parsimony methods, such as dispersal-vicariance analysis or parsimony-based tree fitting, are derived from explicit process models and reconstruct both ancestral distributions and biogeographic events. Each process is associated with a cost that should be inversely related to its likelihood, with the most parsimonious explanation of the data being the one of minimum cost. Using large data sets of phylogenies and permutation tests, event-based methods have been used, for example, to study directions of dispersal and test alternative diversification hypotheses in the Holarctic fauna, or to identify patterns of concerted dispersal in the Southern Hemisphere. Other more probabilistic methods include the application of Bayesian Inference to biogeographic analysis, in which dispersal and extinction can be introduced as random parameters in the model. This method is currently under development and would be applied to infer colonization patterns in the Canary Archipelago.
Place, publisher, year, edition, pages
IdentifiersURN: urn:nbn:se:uu:diva-77278OAI: oai:DiVA.org:uu-77278DiVA: diva2:105190