Effective dispersal and genetic structure of a small mammal in an intensively managed agricultural landscape: is there any barrier to movement?

Abstract

Context: Dispersal and gene flow are key processes for the persistence of populations through the maintenance of large effective sizes and recolonization of empty patches. Therefore, assessing how dispersal is modulated by landscape characteristics and the species’ perception of them is are crucial in a context of anthropogenic landscape change and habitat fragmentation. Objectives: We used the rodent Calomys venustus as a model to study genetic connectivity in highly transformed agricultural landscape, to understand how species cope with changes in human modified environments. Methods: We placed sixty five trap lines within an area of approximately 1500 km2 in an agroecosystem in central Argentina, separated by different landscape elements (crop fields, dirt roads, paved roads, and water streams). To assess their influence on dispersal, we implemented two Bayesian clustering methods, and several methodologies based on the pattern of genetic similarity between pairs of individuals relative to their geographic separation. Results: Genetic autocorrelation analysis showed a clinal pattern, with a genetic patch size of approximately 4 km. Bayesian clustering suggested the presence of 7 randomly mating populations with highly variable geographic extent that followed an isolation by distance pattern. Their boundaries did not coincide with any apparent barrier to dispersal. An Estimated Effective Migration Surfaces analysis confirmed that dispersal in this species would not be impeded by landscape features. Conclusions: The strong habitat preferences of C. venustus would not be related to dispersion. Population limits were not defined by apparent barriers to gene flow. We propose that the perceived range of landscape complexity in C. venustus and the grain of resources at particular sites in the landscape would determine the boundaries of genetic clusters.