From the mountains to the sea: Rethinking Mediterranean glacial refugia as dynamic entities

Aim: Glacial refugia are areas of primary importance for the evolution and conserva- tion of biodiversity. Yet, their geographic location remains loosely defined even in intensively studied areas, preventing a thorough understanding of their role in the spatiotemporal biodiversity dynamics. With this study, we aim to locate the major glacial refugia within the biodiversity hotspot of the Italian peninsula, to understand the processes that warranted the long-term persistence of biodiversity in the face of climate changes. Location: Italian peninsula. Taxon: Terrestrial vertebrates. Methods: We calibrated species distribution models (SDM) for 22 lineages of ter- restrial vertebrates endemic to the Italian peninsula and projected the SDMs to the last-glacial maximum conditions. Then, we combined single-lineage projections to in- vestigate the location and spatiotemporal dynamics of multi-species glacial refugia. Results: Multi-species refugia were mostly found in coastal areas that have been flooded by the post-glacial marine transgressions, and that are currently below the sea level. Indeed, we identified six major areas acting as glacial refugia, mainly located outside the current coastline in the southern part of the peninsula and along the west- ern coast. These areas were close to previously inferred locations of glacial refugia and genetic diversity hotspots, but none coincided with them. Conclusions: Results from this study outline glacial refugia as highly dynamic units. Most of the identified refugial areas have been lost by post-glacial sea-level rise. Accordingly, species persistence through the Late Pleistocene was not granted by long-term environmental stability, but by the opportunity to shift species' distribu- tions along altitudinal gradients, following changes in climate and habitat suitability. Notably, our results decouple glacial refugia from hotspots of genetic diversity. Thus, the current location of a hotspot should not be taken as evidence for the occurrence of a glacial refugium in that location—even though a refugium is likely to be located somewhere nearby.