Stream and canopy cover promote population connectivity in two endemic Pacific northwestern salamander species, Rhyacotriton kezeri and R. variegatus

Background/Question/Methods

Landscape genetic methods can be used to identify the most effective conservation measures to maintain connectivity among populations. Analyses of habitat factors that facilitate or restrict gene flow among populations are particularly useful for species with specific habitat requirements and limited dispersal abilities. The Columbia torrent salamander (Rhyacotriton kezeri) and the southern torrent salamander (R. variegatus) are congeners with low desiccation tolerance and restricted geographic ranges, endemic to the Pacific northwestern United States. Both species are closely associated with fast moving headwater streams in old-growth forests, yet are currently not protected at the state or federal level despite the fragmentation of their habitat due to timber harvesting. However, R. kezeri is currently under consideration for listing under the Endangered Species Act. Given their specific habitat needs, we predicted that genetic distance in both species would be positively correlated with climate and landscape variables that increase the risk of desiccation, and tested this prediction using two landscape genetic models: least-cost paths and Circuitscape. Furthermore, we predicted that R. kezeri would show stronger evidence of population declines related to habitat loss than R. variegatus, due to its more highly restricted range.

Results/Conclusions

Two genetic distance measures, pairwise F_ST and proportion of shared alleles (D_PS), suggested that gene flow was low among sampling localities and genetic structure was high overall in both species. The level of genetic structure and other population genetic measures did not differ greatly between the species. Using both least-cost path and Circuitscape models of landscape resistance, we found that high stream cover, low canopy cover, and high heat-load index all restricted gene flow among populations of both species. Slope was generally positively correlated with genetic distance. Growing season precipitation was present in the top models for R. variegatus. We suggest that the conservation status of both species be revisited given this evidence of high genetic structure, the level of habitat fragmentation in their ranges, and their reliance on dense canopy cover for dispersal. Maintaining stream corridors with buffers of dense canopy cover may maximize connectivity despite the pressures of timber harvest and urbanization.