Cougar genetic variation and gene flow in a heterogeneous landscape

Abstract

Management of game species requires an understanding not just of population abundance, but also the structure of and connections between populations. Like other large-bodied carnivores, the cougar (Puma concolor) exhibits density –dependent dispersal and is capable of long-distance movement; in the absence of barriers to movement, these traits should lead to high connectivity between individuals and a lack of genetic differentiation across areas of continuous habitat. Previous research has suggested that cougar movement may be influenced by landscape variables such as forest cover, elevation, human population density, and highways. I assessed the population structure of cougars (Puma concolor) in Washington and southern British Columbia by examining patterns of genetic variation in 17 microsatellite loci, and the contribution of landscape variables to this genetic variation. I evaluated population structure using genetic clustering algorithms and spatial principal components analysis. I quantified the effect of distance on genetic variation by calculating the correlation between the genetic distance and geographic distance between every pair of individuals, as well as the spatial autocorrelation of genetic distances. To compare the observed pattern of genetic differentiation with that which would arise solely from isolation by distance, I simulated allele frequencies across the study area where the cost to movement between individuals was proportional to the distance between them. I also evaluated the support for evidence of male-biased dispersal in allele frequencies. Bayesian clustering analyses identified four populations in the study area, corresponding to the Olympic Peninsula, Cascade Mountains, northeastern Washington and Blue Mountains; these clusters were supported by patterns of genetic differentiation revealed with spatial PCA. v Although I found a significant relationship between the geographic and genetic distance between individuals, simulated allele frequencies displayed no meaningful spatial pattern of differentiation, suggesting that male dispersal would be adequate within the scale of the study area to prevent genetic isolation from occurring if the only factor to affect dispersal was geographic distance. While cougars are capable of long-distance dispersal movements, dispersal in heterogeneous landscapes may be mediated by the resistance of the landscape to movement. I derived resistance surfaces for forest canopy cover, elevation, human population density and highways based on GIS data and estimated the landscape resistance between pairs of individuals using circuit theory. I quantified the effect of the resistance to movement due to each landscape factor on genetic distance using multiple regression on distance matrices and boosted regression tree analysis. Both models indicated that only forest canopy cover and the geographic distance between individuals had an effect on genetic distance, with forest cover exhibiting the greatest relative influence. The boundaries between the genetic clusters I found largely corresponded with breaks in forest cover, showing agreement between population structure and landscape variable selection. The greater relative influence of forest cover may also explain why a significant relationship was found between geographic and genetic distance, yet geographic distance alone could not explain the observed pattern of allele frequencies. While cougars inhabit unforested areas in other parts of their range, forested corridors appear to be important for maintaining population connectivity in the northwest.