Incorporating range position and temporal stability of projected changes in tree species habitats to assess regional climate change vulnerability

Background/Question/Methods

As we consider the potential impacts of climate change on tree species, we are confronted with the challenge of relating how such global perturbations intersect with a multitude of regional environmental challenges. Many of these near-term factors operate at a more rapid temporal and finer spatial dimension.   To assess climate change vulnerabilities within a regional context, we must place climate change impacts into a more applicable domain.  Here we demonstrate one pathway to increase the utility of coarse-scale species distribution models for regional vulnerability assessments.  Specially, building from our series of 134 tree species distribution models for the eastern US (DISTRIB) and projected habitat responses under a range of GCMs, we address how to create a species distributional boundary which aligns within a focal area to estimate the likely tree species pool; how, considering the rate of projected change over the course of the 21^st century (1990, 2040, 2070, 2100) can characterize the temporal immediacy of the risks; and finally, how the stability of the projected habitat change builds confidence into vulnerably assessments. Our results are from four regions within the eastern US: Northern Wisconsin (WI), Western Maine (ME), Central Pennsylvania (PA), and Coastal South Carolina (SC).

Results/Conclusions

These disparate regions clearly demonstrate how the regional context of assessing vulnerability can influence conclusions regarding potential changes in forest composition through time.  For example, across chestnut oak’s range, we show high consistency in a projected mild increase in habitat by the end of the century, especially within the heart of it distribution (PA).  However, at the periphery of its range, we see marked disagreement in the stability of habitat changes (e.g., in ME the differences between high and low emissions scenario predictions are greater than the magnitude of projected increases under high emissions).  Further, by considering the range position of a species relative to the location of interest we reduce the number of potential candidate species to consider in a vulnerability assessment.  Finally, with the inclusion of projections for each species at 30-year intervals we track the trajectory and speed of when suitable habitat may become available for a species.  For example, black ash in WI shows a potential marked reduction in habitat within the next 30 years, while for other species, the greatest potential for habitat change is delayed until the end of the century.  These results provide a means rapidly comparing regional climate change vulnerabilities for forests.