Abstract

Habitat connectivity is a key objective of current conservation policies and is commonly modeled by landscape graphs (i.e., sets of habitat patches [nodes] connected by potential dispersal paths [links]). These graphs are often built based on expert opinion or species distribution models (SDMs) and therefore lack empirical validation from data more closely reflecting functional connectivity. Accordingly, we tested whether landscape graphs reflect how habitat connectivity influences gene flow, which is one of the main ecoevolutionary processes. To that purpose, we modeled the habitat network of a forest bird (plumbeous warbler [Setophaga plumbea]) on Guadeloupe with graphs based on expert opinion, Jacobs' specialization indices, and an SDM. We used genetic data (712 birds from 27 populations) to compute local genetic indices and pairwise genetic distances. Finally, we assessed the relationships between genetic distances or indices and cost distances or connectivity metrics with maximum-likelihood population-effects distance models and Spearman correlations between metrics. Overall, the landscape graphs reliably reflected the influence of connectivity on population genetic structure; validation R2 was up to 0.30 and correlation coefficients were up to 0.71. Yet, the relationship among graph ecological relevance, data requirements, and construction and analysis methods was not straightforward because the graph based on the most complex construction method (species distribution modeling) sometimes had less ecological relevance than the others. Cross-validation methods and sensitivity analyzes allowed us to make the advantages and limitations of each construction method spatially explicit. We confirmed the relevance of landscape graphs for conservation modeling but recommend a case-specific consideration of the cost-effectiveness of their construction methods. We hope the replication of independent validation approaches across species and landscapes will strengthen the ecological relevance of connectivity models.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call