Multi-morph eco-evolutionary dynamics in structured populations

Abstract

Our understanding of the evolution of quantitative traits in nature is still limited by the challenge of including realistic trait distributions in the context of frequency-dependent selection and ecological feedbacks. We develop a theoretical framework to analyse the dynamics of populations composed of several morphs and structured into distinct classes (e.g. age, size, habitats, infection status, species…). Our approach extends to class-structured populations a recently introduced “oligomorphic approximation” which bridges the gap between adaptive dynamics and quantitative genetics approaches and allows for the joint description of the dynamics of ecological variables and of the moments of multimodal trait distributions. We also introduce a new approximation to simplify the eco-evolutionary dynamics using reproductive values. This effectively extends Lande’s univariate theorem not only to frequency- and density-dependent selection but also to multimodal trait distributions. We illustrate the effectiveness of this approach by applying it to the important conceptual case of two-habitat migration-selection models. In particular, we use our approach to predict the equilibrium trait distributions in a local adaptation model with asymmetric migration and habitat-specific mutational variance. We discuss the theoretical and practical implications of our results and sketch perspectives for future work.