Understanding the complex biotic and abiotic influences on population persistence following landscape changes remains key to predicting future trends. In the Cascade Mountains of central Oregon, forest edges are encroaching on alpine meadows which harbor charismatic wildflower and pollinator communities. Forest encroachment reduces meadow area and connectivity. We sampled population genetic structure in hummingbird-pollinated Western columbine (Aquilegia formosa)across a range of meadow sizes and connectivity and asked 1) whether genetic structure among plant populations is predicted by hummingbird movements among meadows and 2) if there was evidence for an effect of meadow size and connectivity on genetic diversity. We posited that if hummingbirds primarily move through open habitats, the likelihood of a hummingbird visit to flowers inside forests should be less than that to flowers in meadows, thus increasing levels of inbreeding in small, isolated meadows. Under this hypothesis, we predicted that plant population structure at putatively neutral loci would be largely dictated by hummingbird behavior and show trends consistent with isolation by distance along least-cost paths parameterized by contemporary hummingbird movement.
Results/Conclusions
We sampled hummingbird movements within a systematic grid of RFID readers and used a Bayesian hierarchical logistic regression to test whether pit tagged hummingbirds were more likely to visit readers given differing intervening land-cover types. We genotyped 167 individuals at 70 SNP loci across 30 meadows and 5 meadow complexes. Roughly 34.1% of genetic variation occurs among complexes, 35.5% among meadows within a complex, 18.0% among samples within a meadow, and 12.4% within individuals. We tested for deviations from Hardy-Weinberg proportions (HWP) and found more significant results than predicted at random. This was driven largely by 7 populations with heterozygote deficiency. Pairwise relatedness of individuals yielded evidence for isolation by distance (Mantel test, p<0.001); however, observed genetic structure in neutrally-behaving loci is not best explained by models of contemporary hummingbird movement or related to intervening habitat-type. Contrary to predictions, populations significantly deviating from HWP showed evidence for decreased inbreeding with smaller meadow sizes. Taken together, these results suggest continued potential for gene flow among meadows despite woody encroachment, and the possibility of reduced inbreeding at the local scale as meadows contract.