Past, present and future refugia for Tasmania's palaeoendemic flora

Aim: Refugia under past climates have been important in structuring current patterns in diversity, while refugia under anthropogenic climate change will likely be important in retaining this diversity and shaping new patterns. However, few studies have examined the congruence of past, present and future refugia, or the spatiotemporal connectivity of these refugia. Our aim was to test the extent of overlap of refugia under Last Glacial Maximum (LGM), present (2015) and likely future climates (2100), for Tasmania's palaeoendemic flora. We then aimed to identify areas of high spatiotemporal refugia connectivity, as priority areas for conservation and management. Location: Tasmania, Australia. Methods: We developed and applied a new community-level approach to identifying refugia, based on generalized dissimilarity modelling of compositional turnover and a set of reference sites with known biodiversity value. Using these projections of palaeoendemic plant refugia for past, present and future climates, we developed and applied a second approach to quantify the level of connectivity of these refugia over space and time. Results: Although there was large overlap (85%) between current and future climates in the distribution of the highest value palaeoendemic refugia, the small congruence of these areas with refugia at the LGM resulted in only a small area ( c. km 2 ) of persistent high value refugia over all three time periods. Despite this, our spatiotemporal analysis identified several areas of high connectivity in refugial environments for Tasmania's palaeoendemic flora over time. Main conclusions: The community-level approaches we demonstrate here to quantify refugia and their spatiotemporal connectivity have the potential to advance our understanding of biodiversity dynamics, particularly for taxonomic groups that are species-rich, poorly studied or comprised of many rare species, where species-level approaches are less suitable.