Isolation by resistance analysis reveals major barrier effect imposed by the Tsinling Mountains on the Chinese wood frog

Abstract

AbstractAmphibians are excellent model systems for studying how heterogeneous landscapes drive population differentiation. Though mountain ridges have been widely implicated as barriers to amphibian dispersal, a landscape genetic study by Zhan et al. (BMC Genet, 10, 2009, 17) indicated that the Tsinling Mountains of northwestern China did not significantly impede gene flow of the Chinese wood frog (Rana chensinensis). Using their published genetic data, we re‐assessed the impact of elevation, land cover, and roads on the Chinese wood frog using an isolation by resistance (IBR) approach, which allows for the modeling of complex and heterogeneous landscapes unavailable in classical population genetic analysis. We developed a novel method for optimization of IBR analysis using Circuitscape involving initial permutation‐based optimization of the correlation coefficient obtained through partial Mantel tests, followed by subsequent sensitivity analysis to improve upon the model’s biological relevance. Our results indicated that mountain ridges indeed functioned as barriers to dispersal of Chinese wood frogs, and that previous conclusions were based on methodological limitations. Furthermore, a prominent threshold effect at elevations from 1500–2000 m was evident, with elevations below this range minimally impeding gene flow and higher elevations having a significant barrier effect. We suggest that a combination of expert opinion, sensitivity analysis, and permutation‐based optimization procedures in landscape genetic studies will likely generate models that are both highly explanatory and biologically relevant.