Matrix dominance and landscape resistance affect genetic variability and differentiation of an Atlantic Forest pioneer tree

Abstract

Context: Despite the importance of secondary forests for the maintenance of biodiversity, the impact of pioneer trees on habitat loss and fragmentation is poorly understood. Objectives: We analyzed the effects of landscape structure on genetic variability (node level analyses) and genetic differentiation (link level analyses) of the pioneer tree Cecropia hololeuca in the Atlantic Forest of southeastern Brazil. At the node level, we analyzed the effects of landscape structure (forest and pasture amount, compositional heterogeneity, number and shape of patches, and matrix dominance), topography (slope and elevation), and effective population size (Ne) on allelic richness (AR), genetic diversity (He) and inbreeding coefficient (f). At the link level, we analyzed the effect of four resistance surfaces (isolation by geographic distance, land cover, elevation, and slope) on genetic differentiation between populations (FST, G’ST, Jost’s D). Methods: We genotyped 257 individuals of C. hololeuca using eight microsatellite loci. At the node level, we calculated landscape and topographic variables at six spatial scales. At the link level, we optimized the resistance surfaces using ResistanceGA. We used a model selection approach to select the most parsimonious models. Results: At the node level, matrices dominated by pastures had higher AR than landscapes dominated by Eucalyptus at the 1500 m spatial scale. At the link level, FST was best explained by slope, with lower values imposing more resistance. All resistance surfaces explained G’ST. Conclusions: In addition to conserving and restoring habitats, management practices that improve matrix permeability should be adopted to favor the movement of dispersers, consequently improving natural regeneration, increasing allelic richness, and decreasing genetic differentiation of natural populations.