A coherent biogeographical framework for Old World Neogene and Pleistocene mammals

In order to understand mammalian evolution and compute a wide range of biodiversity indices, we commonly use the 'bioregion', a spatial division adapted to ecological and evolutionary constraints. While commonly conducted by neontologists, the establishment of bioregions in palaeontology is generally a secondary analysis, shaped on subjective time scales and areas specific to the investigated questions and groups. This heterogeneity, coupled with the scale-dependency of biodiversity indices, prevents the clear identification of macroecological and macroevolutionary trends for large taxonomic groups like extinct mammals. Here we tackle this issue by providing a coherent framework for Neogene and Pleistocene mammals of the Old World following two steps: (1) a temporal scale adapted to mammalian evolutionary history (i.e. evolutionary fauna) is defined by poly-cohort analysis; (2) bioregions are then computed for each evolutionary fauna by clustering, ordination and intermediate approaches at multiples spatial scales (i.e. continental to regional) for Eurasia and Africa. Additionally, providing a coherent framework for a wide range of mammalian datasets, our results show: (1) the synchronous emergence and fall of five mammalian evolutionary faunas identified at chronological scales varying from the epoch to the geological stage; (2) a transition from a longitudinal to a latitudinal biogeographical structuring between the Miocene and Pliocene, especially in Europe; (3) the long-term affinity of southern Asian with African faunas, in sharp contrast with the modern Palaearctic bioregion extension; and (4) the establishment of a vast Mediterranean bioregion from fragmented areas in the Late Miocene to its full extent in the Pleistocene.