Redefining the climate niche of plant species: A novel approach for realistic predictions of species distribution under climate change

Abstract

Climate change is increasingly affecting plant species distributions, in ways that need to be predicted. Here, in a novel prediction approach, we developed the relevant climate niche (RCN) of plants, based on thorough selection of climate variables and implementation of a non-parametric Bayesian network for climate simulations. The RCN was conditionalized to project the fate of Silene acaulis in North America under moderate (Representative Concentration Pathway 4.5; RCP4.5) and extreme (RCP8.5) short-term (2011–2040) climate scenarios. We identified a three-variable climate hypervolume for S. acaulis. Within 20 years >50% of current locations of the species will be outside the defined climate hypervolume. It could compensate for climate change in 2011–2040 through a poleward shift of 0.97 °C or an upshift of 138 m in the RCP4.5 scenario, and 1.29 °C or 184 m in the RCP8.5 scenario. These results demonstrate the benefits of redefining the climate niche of plant species in the form of a user-defined, data-validated, hierarchical network comprising only variables that are consistent with species distribution. Advantages include realism and interpretability in niche modeling, and new opportunities for predicting future species distributions under climate change.