Variation in demographic rates across populations and environmental gradients can generate disequilibrium between species occurrence and climate, which presents a key challenge for predicting range shifts. This challenge is particularly pronounced for long-lived plants such as trees, which rely upon successful dispersal, colonization, and establishment over multiple, long generations to shift their ranges. The rate and magnitude of tree range shifts may be especially influenced by range-wide, climate-induced variation in recruitment, which acts as a critical bottleneck in tree population dynamics. Despite the importance of recruitment for understanding species range dynamics, the link between the population-level processes regulating seedling establishment and survival, and species occurrence over broad scales remains poorly understood, and considering this link in range models requires assimilation of data and models extending across multiple spatial scales. We applied an integrated range modeling framework that capitalizes on the flexible nature of hierarchical Bayesian modeling to directly assimilate information on seedling survival, recruitment dynamics, and adult occurrence into a cohesive framework for predicting species’ ranges. We used this approach to predict current and future ranges of five dominant tree species in the interior west region of the United States, and compared predictions to those generated using a standard species distribution modeling approach.
Results/Conclusions
Our comparison of species distribution models and integrated range models revealed few major qualitative difference in response curves or range predictions when accounting for recruitment, but integration of seedling information did improve characterization of model uncertainty. Specifically, our results indicate that range-wide variation in recruitment both now and in the future is most uncertain along the edges of occupied regions, which increases uncertainty in projections of future species occurrence along range margins. Further, recruitment dynamics when considered at this scale appear to be only weakly related to coarse-scale climate, and the manner in which non-climatic factors will alter recruitment dynamics under future climate remains unclear. Overall, our findings suggest that the broad-scale climatic dependence of the regeneration niches of trees in the interior west may be weaker than that of the adult climatic niche, and that this enhances uncertainty in predicting range-wide responses of these species to climate change.