Empirical estimation of habitat suitability for rare plant restoration in an era of ongoing climatic shifts

Abstract Accurate estimates of current and future habitat suitability are needed for species that may require assistance in tracking a shifting climate. Standard species distribution models (SDMs) based on occurrence data are the most common approach for evaluating climatic suitability, but these may suffer from inaccuracies stemming from disequilibrium dynamics and/or an inability to identify suitable climate regions that have no analogues within the current range. An alternative approach is to test performance with experimental introductions, and model suitability from the empirical results. We used this method with the Haleakalā silversword ( Argyroxiphium sandwicense subsp. macrocephalum ), using a network of out-plant plots across the top of Haleakalā volcano, Hawaiʻi. Over a ~ 5-year period, survival varied strongly across this network and was effectively explained by a simple model including mean rainfall and air temperature. We then applied this model to estimate current climatic suitability for restoration or translocation activities, to define trends in suitability over the past three decades, and to project future suitability through 2051. This empirical approach indicated that much of the current range has low suitability for long-term successful restoration, but also identified areas of high climatic suitability in a region where plants do not currently occur. These patterns contrast strongly with projections obtained with a standard SDM, which predicted continued suitability throughout the current range. Under continued climatic shifts, these results caution against the common SDM presumption of equilibrium between species’ distributions and their environment, even for long-established native species.