Bioregionalization, graph theory and simulated worlds: revisiting foundational goals of biogeography with 21st century tools

CGL2017-86926-P., MINECO, 2018-2020.

The delineation of bioregions (geographic areas differentiated by their taxonomic composition) has been always central to biogeography. This has to do with their use as spatial reference units in comparative analyses, with their potential to be used in conservation projects, and with the increasing availability of distribution data and their effect of stimulating revisions of bioregional schemes.

Which bioregions we recognize is closely determined by what methods and which data we use to define them, something not explicitly tested before. In particular, we intend to classify these artificial data with two main types of techniques: (1) classical multivariate methods that classify the units of analysis based on compositional similarity matrices. And (2) flexible and powerful techniques derived from graph theory that allow classifying with or without hierarchy and/or overlap between groups, and some of which even allow including additional information (metadata) referred to, for instance, phylogenetic relationships, and/or historical memberships of analysis units to ancient land masses or paleoclimates. Along with this, the most favored method by these tests will be used to define new bioregional schemes based on current distributions and phylogenies of different taxonomic groups (vertebrates, butterflies, trees and vascular plants in general) and at different spatial extents (Iberian Peninsula, Europe, the US, and global). Some datasets will be Some databases will be made or updated as part of this project.