Identifying potentially suitable and accessible refugia to mitigate impacts of an emerging disease on a rare tree.

Identifying refugia from emerging threats is vital to ensure the persistence of rare and threatened species, but modeling habitat distribution for these species is challenging and the role of people in refuge management is rarely considered. Myrtle rust is an emerging infectious disease that represents a grave threat to the rare wetland tree species maire tawake (Syzygium maire) in Aotearoa New Zealand. We combined high-resolution hydrological modeling with integrated species distribution modeling of new and existing S. maire records to identify the extent of habitat in the capital city region available for conservation management. We mapped 2 myrtle rust infection risk scenarios throughout the region to identify areas of relatively low disease risk and used distance of S. maire habitat to the nearest road as a proxy for human accessibility to the area. We identified 1230 km² of S. maire habitat (waterlogged areas) in the region. In these areas, 1-52 km² were the most feasible for conservation because they were predicted to support high relative abundances of S. maire, were accessible by road, and offered lower disease risk. However, protecting trees only in low-risk or accessible refugia was predicted by the species distribution model (SDM) to be insufficient to maintain the regional population as the myrtle rust pandemic proceeds. Our highly local approach to refugia modeling enabled rapid collection of new records of a rare species for species distribution modeling and access to high-resolution topographical data for hydrological modeling. However, limitations to understanding the biophysical limits of myrtle rust and S. maire included model-based constraints on inference, poor spatial precision of historical species records, insufficient information on groundwater drainage, and uncertainty in quantifying disease risk. The success of regional conservation efforts for this species will likely depend on human intervention to increase S. maire occupancy in low-risk habitats and to manage myrtle rust. We therefore recommend leveraging human-nature interactions in areas to create, expand, and protect habitat for rare species in a rapidly changing world.