Conservation efforts in isolated archipelagos such as Hawaii often focus on habitat-based conservation and restoration efforts that benefit multiple species. Unfortunately, identifying locations where such efforts are safer from climatic shifts is still challenging. We aimed to provide a method to approximate these potential habitat shifts for similar data- and research-limited contexts. We modeled the relationship between climate and the potential distribution of native biomes across the Hawaiian archipelago to provide a first approximation of potential native biome shifts under end-of-century projected climate. Our correlative model circumvents the lack of data necessary for the parameterization of mechanistic vegetation models in isolated and data-poor islands. We identified locations consistently expected to remain the same in terms of the native biome compatibility by the end of the century with a robust evaluation of sources of uncertainty in our projections. Our results show that, despite large differences in climate projections considered, 35% of the areas considered are consistently projected to maintain their current compatibility to native biomes. By integrating our native biome compatibility projections with maps of current actual cover, we identified areas ideal for long-term habitat conservation and restoration. Our modeling approach can be used with relatively simple data; offers multiple forms of projection confidence estimates, model calibration, and variable selection routines; and is compatible with ensemble projections. This method is not only applicable to potential native cover, as done in this study, but to any set of vegetation classes that are related to environmental predictors available for modeling.
|Publication Subtype||Journal Article|
|Title||Identifying opportunities for long-lasting habitat conservation and restoration in Hawaii’s shifting climate|
|Series title||Regional Environmental Change|
|Contributing office(s)||Pacific Island Ecosystems Research Center|
|Google Analytic Metrics||Metrics page|