Empirically optimized landscape resistance reveals barriers and facilitators of gene flow in woodland caribou

Abstract

Understanding how landscape features shape genetic connectivity is critical for conserving wide-ranging species like woodland caribou (Rangifer tarandus caribou), particularly in managed landscapes where resource extraction and anthropogenic disturbance occur alongside natural disturbance. We used microsatellite genotypes from 244 individuals and four genetic distance metrics (Fij, Dps, PCA10, PCA64) to optimize resistance surfaces with ResistanceGA and assess spatial patterns of genetic structure in the Churchill Range, northwestern Ontario. Despite moderate genetic diversity (mean He = 0.68) and extremely low pairwise FST (< 0.003), spatial autocorrelation and clustering analyses indicated weak but significant genetic structure consistent with isolation by distance. Resistance modeling showed that isolation by distance alone provided a poor fit relative to models incorporating landscape features. Wetlands consistently emerged as the dominant predictor across three allele frequency–based metrics (DPS, PCA10, and PCA64), highlighting their role in sustaining broad-scale gene flow. In contrast, coniferous forest paired with water was the top predictor for kinship-based Fij, suggesting that conifer habitats influence genetic structure through more recent demographic processes such as philopatry and site fidelity. These results establish a genetic baseline for caribou in the Churchill Range and demonstrate that connectivity is shaped primarily by wetlands, with conifer habitats also contributing to more recent genetic structuring. Keywords: woodland caribou, landscape genetics, gene flow, isolation by distance (IBD), isolation by resistance (IBR), population structure, habitat connectivity, wetlands, coniferous forest