Using spatial models to identify refugia and guide restoration in response to an invasive plant pathogen.
The global spread of invasive plant pathogens is increasing, putting natural forests and ecosystem services under threat. Spatial data can quantify range non-overlap between invasive pathogens and their hosts to identify existing and potentially restorable refugia to enable 'escape' from the pathogen. In 2017, myrtle rust Austropuccinia psidii was detected in New Zealand. New Zealand has 27 native woody and liana species in the plant family susceptible to the disease (Myrtaceae), many of which are ecosystem dominants and economically important. Spatial methods were used to compare the current New Zealand distribution of myrtle rust with the modelled distributions of its potential native hosts. To guide management and set priorities for conservation of at-risk species, areas of potential refugia where a Myrtaceae species is predicted to occur outside the pathogen range were identified under two myrtle rust distribution scenarios. Myrtle rust will thrive in New Zealand's warmer regions. Many native Myrtaceae are distributed within this area, but several species occur extensively outside the core range of the disease. Species distributed in cooler southern regions will be best placed to persist in refugia. Myrtaceae species with specific habitat requirements and narrow geographical ranges in warmer (northern) areas are likely to require ex situ or active in situ management. Even widely distributed species will benefit from the restoration of suitable habitat that supports multiple species outside the myrtle rust range. Synthesis and applications. Spatial data can be used to identify refugia and restoration opportunities, and thus inform landscape-level management responses to invasive pathogens. This approach can guide decisions over where to implement in situ (e.g. fungicide spraying) versus ex situ (e.g. seed banking, botanic gardens) management efforts.