Evolution of Specialization on Re- source Utilization in Structured Metapopulations

We study the evolution of resource utilization in a structured discrete time metapopulation model with an infinite number of patches, prone to local catastrophes that wipe out the entire local population but leave the patch itself habitable. We assume that there are two distinct resources in each patch and the strategy of an individual is the degree of specialization on the usage of these resources. We study the course of evolution using the adaptive dynamics approach. We show that slow local growth and high catastrophe probability favor generalism whereas fast local growth and low catastrophe probability favor specialism. The relation between the evolutionary dynamics of the model and the emigration probability is, however, more complicated. High emigration probability always favors generalism but there exists also parameter combinations such that also low values of emigration probability are more favorable to generalism than intermediate values. Such a non-monotonous relation between emigration probability and evolutionary dynamics has not been observed before. Evolutionary branching (i.e., appearance of two phenotypically distinct lineages) is possible in the model. The parameter domain where it is possible becomes larger when catastrophe probability is increased. For different parameter values and initial states the following endpoints of evolution are possible: monomorphic metapopulation specializing on one of the two resources, monomorphic generalist metapopulation or dimorphic metapopulation consisting of the two specialists. In the model also a trimorphism of the generalist and the two specialists is possible but not reachable in an initially monomorphic population if mutation steps are small and occur only infrequently. Our modeling approach offers a natural way to include more than two patch types into the models. This has not been usually possible in the previous metapopulation models focusing on the evolution of specialization.