Validating graph-based connectivity models with independent presence-absence and genetic data sets.

Area of Science:

• Ecology
• Conservation Biology
• Genetics

Background:

• Habitat connectivity is vital for conservation policy, often modeled using landscape graphs.
• Existing graph models lack empirical validation regarding functional connectivity.
• Gene flow is a key ecoevolutionary process influenced by habitat connectivity.

Purpose of the Study:

• To empirically validate landscape graphs by assessing their correlation with gene flow.
• To compare the ecological relevance of different landscape graph construction methods.
• To inform conservation modeling by evaluating graph construction cost-effectiveness.

Main Methods:

• Modeled habitat network of plumbeous warbler using expert opinion, Jacobs' indices, and SDMs.
• Utilized genetic data from 712 birds across 27 populations to calculate genetic distances and indices.
• Employed maximum-likelihood models and Spearman correlations to relate genetic and graph-based metrics.

Main Results:

• Landscape graphs demonstrated reliability in reflecting connectivity's influence on genetic structure (R² up to 0.30, correlations up to 0.71).
• The most complex method (SDM) did not consistently yield the most ecologically relevant graphs.
• Cross-validation and sensitivity analyses clarified method-specific advantages and limitations.

Conclusions:

• Landscape graphs are relevant tools for conservation planning and modeling.
• The choice of graph construction method should be case-specific, considering cost-effectiveness.
• Further validation across diverse species and landscapes is recommended to enhance connectivity model relevance.