Forecasting species distributions: coping with uncertainties
In collaboration with Isabelle Chuine, Sylvain Delzon, Mark Kirkpatrick, Antoine Kremer, François Massol, Ophélie Ronce, Niklaus Zimmermann.
Climate change will strongly affect the demography and geographic range of numerous species. Different types of models have so far investigated how species distributional ranges could evolve when faced to climate change. From the most general to the most species-centered models, these can be classified into three broad categories:
1. theoretical evolutionary models assume that a species' persistence in a given environment depends on the matching between a quantitative trait and an optimal phenotype. These models offer quite general results; however they oversimplify species' responses to spatial heterogeneity.
2. correlative species distribution models, based on the current relationships between bioclimatic variables, predict how species ranges can evolve in future climates. These models, however, are based on purely statistical relationships which are not related to the species biology, and cannot be disentangled from historical contingencies.
3. process-based species distribution models are based on the responses of biological processes important to the species life cycle (e.g. growth, respiration, offspring production). These models require a good knowledge of the species studied.
While evolutionary processes are implemented in theoretical model, they currently lack from data-driven models. The contributions of microevolution and phénotypic plasticity to species adaptation needs further exploration. I am interested in deriving theoretical models whose assumptions are closer to observations, and in limiting uncertainties of the more data-driven models.