Habitat amount, not habitat configuration, best predicts population genetic structure in fragmented landscapes

Context: Landscape structure shapes the genetic structure of populations by delimiting spatial patterns of dispersal and reproduction across generations. Thus, descriptions of human-altered landscapes can be used to predict demographic and evolutionary outcomes of populations. Effectively measuring landscape structure to predict genetic structure requires that we understand the relative importance of distinct components of landscape structure (e.g., habitat amount and configuration) in creating spatial patterns of genetic variation. Objectives: We thus developed an individual-based simulation model to test predictions about the relative importance of habitat amount and configuration in producing genetic structure. We also investigated the independent relationships between components of landscape structure and the population dynamics that underlie genetic effects. Methods: We ran experiments in which we allowed gene flow and population size to vary as emergent outcomes of the interactions between hypothetical populations and heterogeneous landscapes. Results: We found that the amount of habitat in a landscape is a much better predictor of genetic structure than is habitat configuration. This pattern holds across a range of landscapes and dispersal distances and behaviors. When habitat is non-contiguous (i.e., fragmented), habitat amount mediates production of genetic differentiation by regulating both the size and isolation of habitat patches, which in turn regulate population size and gene flow. Conclusions: These results suggest that habitat amount, a simple measure that is easy to calculate, may often be the best metric for predicting population genetic structure and that when possible, measures of habitat amount and population size should be incorporated into landscape genetic studies