The influence of late Pleistocene climatic cycles on divergence and distribution of Great Horned Owl (Bubo virginianus) populations

Although the evolution of many North American species is often thought to have been affected by past glacial cycles, the importance of glacial refugia in contributing to the diversification of North American species is debated. Within temperate North America, Bubo virginianus (Great Horned Owls) are widely distributed and found across many habitats. North American B. virginianus is divided into nine subspecies based on phenotypic diversity. Previous work has failed to identify any genetically differentiated populations, but genetic and geographic sampling were limited. Here, we aim to characterize the population structure of B. virginianus in the United States and Canada and test alternative historical factors that influenced this structure. We tested three hypotheses: isolation by distance (IBD), isolation by environment (IBE), and historical isolation by habitat fragmentation during Pleistocene climatic cycles. If B. virginianus populations were affected by glacial cycles, we would expect 1) increased genetic diversity in forested refugial areas and decreased genetic diversity in historically unsuitable areas, 2) less admixture between identified populations in areas of known refugia, 3) reduced gene flow between historically isolated populations, and 4) late Pleistocene population divergences. We used reduced-representation genomic sequencing of 114 specimen-vouchered samples. We identified three populations that were only weakly differentiated and geographically correspond to eastern, northwestern and southwestern North America, a pattern that is spatially congruent with many Nearctic taxa. We did not detect any correlation between allelic variation and environmental variables thus rejecting IBE. We also identified an east-west and north-south reduction in effective migration that corresponded to the contact zones between the three populations and thus rejected expectations of IBD. Our results were overall consistent with expectations of habitat fragmentation due to the Pleistocene climatic cycles. We identified areas of increased effective genetic diversity and individuals with less admixture in areas corresponding to Pleistocene putative forest refugia and decreased effective genetic diversity in other areas. Our results also recovered relatively low levels of gene flow between the three populations. Demographic models identified the timing of the most recent population divergence as occurring fewer than 8,000 generations ago, roughly corresponding to the last glacial maximum (LGM) with a much deeper divergence between the southwest population and the eastern + northwestern populations during the Pleistocene approximately 55,000 generations ago. Despite debate surrounding the effects of glaciation on avian diversification in North America as a whole, we argue that recent glacial cycles, including the LGM, have profoundly impacted the evolutionary history of B. virginianus.