Species ranges are dynamic due to disturbances and global environmental changes. Although climate change is expected to lead to range expansions for many species, habitat loss may function as barriers leading to both lowered colonization success and species persistence at the range margins. Correlative species distribution models (SDMs) are commonly applied to estimate potential future risk of climate changes on species ranges where change in range size is used as a proxy for extinction risk, however, they do not model processes directly related to extinction risk and may present biased estimates. Thus approaches that account for multiple threats, their potential interaction, and population-level processes that drive species distributions and abundance more explicitly are required.
We linked a population model with species distribution models (SDMs) to investigate predictions of range margin dynamics at the landscape scale under the influence of potential future climate changes and changes in suitable forest habitat on the persistence of a forest-dependent species, the Hooded Warbler (Setophaga citrina). SDMs were developed using three different global change models (GCMs) to estimate potential range shifts across the entire breeding distribution. To quantify the influence of uncertainty on predictions, we applied a global sensitivity analysis across the population model parameters, evaluated alternative patterns and rates of habitat loss, and assessed the uncertainty associated with different climate model projections used in the integrated SDMs-population model to investigate how error propagation occurs through the whole model process.
Results/Conclusions
Our approach provides a method for assessing species vulnerability to multiple environmental changes that influence species over different scales. We found that at the scale of the entire breeding distribution, potential future climate changes differed across the different GCMs with spatially structured variability occurring at both latitudinal range margins. At the landscape scale, managing for suitable habitats presents a viable option for securing persistent populations, however, even under increasingly favourable climatic conditions, limited suitable habitat leads to declines in predicted extinction risks and variability in the location of the range margins. Importantly, we find that although predictions of extinction risk were sensitive to the GCM used, measures of forested habitat loss were more influential, highlighting general strategies for management actions. This work underscores the importance of using multi-model approaches to evaluate environmental changes that occur over different scales on species persistence and the use of sensitivity analyses to rank the relative influence of uncertain factors.