Predicting the distributions of species and communities under altered climatic conditions is a central challenge in ecology. To date, efforts have been limited by disparate data and models that represent different spatiotemporal scales and modeling approaches. For example, ecological niche models typically utilize fine grain species distribution data and climate information, whereas dynamic vegetation models operate at coarser grains, with broad functional types rather than individual species, but incorporate more functional information on biophysical processes like fire and carbon dynamics while also providing outputs highly relevant to species distribution models like canopy characteristics.
To provide more accurate predictions and maximize the utility of each modeling framework, we present an approach that integrates ecological niche models and dynamic vegetation models. With case studies in two disparate biomes, the Neotropical Amazonian and Nearctic Boreal forests, we evaluate the added value of this approach and explore its utility for modeling Pleistocene and near-future plant communities. In the former, we test longstanding questions in paleobiogeography, such as the opening of a dry corridor of savannah in the eastern Amazon, and the connectivity of Amazonian and Atlantic forests. In the latter, we assess the prevalence of predicted no-analog community composition and community structure.
Results/Conclusions
We demonstrate that ecological niche models and dynamic vegetation models offer complementary information that can be integrated to offer more accurate projections of spatial shifts in species and community under climate change. Our model results reveal the extent to which Amazonian forest canopy was intact at the Last Glacial Maximum, and conversely the savannah formation at the forest margins and drier regions. The results lend only partial support to the Amazonian Dry Corridor or the Forest Refugia Hypotheses, however large areas in the eastern Amazon do appear to have been converted to open savannah, and the species composition reflects drier conditions. We show the integrated framework can also be applied to projections of species and community shifts in eastern boreal forests, including where major community types may shift with climate change and to clarify how the composition and structural properties of those communities will likely be altered. Further, niche models underline uncertainties and limitations in shifting climatic envelopes to novel edaphic conditions. We show that the integration of dynamic vegetation models and ecological niche models is relevant for applications as diverse as paleoecological reconstruction and forest management planning.