Processes at work on North American tree species range edges have implications for the predictability of their movements under climate change

Background/Question/Methods

Predicting how climate change will influence tree species distributions requires understanding abiotic constraints on species as well as biotic interactions such as competition and disease. So far, predictions have been largely based on abiotic conditions because biotic interactions are notoriously difficult to understand and predict, as well as expected to vary across species’ ranges. A classic expectation is that species are limited by climate, namely cold temperatures and/or arid conditions, at some range edges, while at other edges, in the absence of abiotically harsh limits, biotic interactions, such as competition, limits species’ ranges. Insights into the nature of constraining factors at range edges will improve predictions of range shifts under climate change. A range edge mainly caused by harshness of climate is more likely to directly trace shifts in climate, than a range edge mainly caused by biotic interactions. Our goal is to identify the nature of range edges and to highlight differences in resulting predictions for range shifts for models ignoring or incorporating insights on the nature of range edges. We use three different models for accomplishing this goal: Distribution models for the realized niche, Phenofit for the fundamental niche, and MAPSS for the competition landscape.

Results/Conclusions

Most of the 13 modeled tree species show the expected patterns of climate limiting northern edges and competition limiting southern edges. Consequently, predictions for range shifts coincide between different approaches much more closely for the northern range edges, than the southern range edges. Southern range edges are vastly overpredicted by climate envelopes and mechanistic, climate-based models (Phenofit). It is unclear whether better prediction of southern range edges by classical correlational distribution models is related to mechanism, for example via implicit modeling of competitors. Therefore, it is also unclear how reliable such models could predict future distributions. However, clear advances can be made by combining climate-based mechanistic models (Phenofit) addressing the fundamental niche, with general expected competition landscapes (MAPSS) for predicting the future of the notoriously fuzzy southern range edges.