Abstract: Context: The accurate estimation of landscape resistance to movement is important for ecological understanding and conservation applications. Rigorous estimation of resistance requires validation and optimization. One approach uses genetic data for the optimization or validation of resistance models. Objectives We used a genetic dataset of snow leopards from China to evaluate how landscape genetics resistance models varied across genetic distances and spatial scales of analysis. We evaluated whether landscape genetics models were superior to models of resistance derived from habitat suitability or isolation-by-distance.
Methods: We regressed genetically optimized, habitat-based, and isolation-by-distance hypotheses against genetic distances using mixed effect models. We explored all subset combinations of genetically optimized variables to find the most supported resistance scenario for each genetic distance.
Results: Genetically optimized models always out-performed habitat-based and isolation-by-distance hypotheses. The choice of genetic distances influenced the apparent influence of variables, their spatial scales and their functional response shapes, producing divergent resistance scenarios. Gene flow in snow leopards was largely facilitated by areas of intermediate ruggedness at intermediate elevations corresponding to small-to-large valleys within and between the mountain ranges.
Conclusions: This study highlights that landscape genetics models provide superior estimation of functional dispersal than habitat surrogates and suggests that optimization of genetic distance should be included as an optimization routine in landscape genetics, along with variables, scales, effect size and functional response shape. Furthermore, our study provides new insights on the ecological conditions that promote gene flow in snow leopards, which expands ecological knowledge, and we hope will improve conservation planning.