Mapping migratory connectivity
Population structure analyses identified five genetically distinct breeding populations with weak genetic differentiation, in contrast to a previous mtDNA analysis that identified only two populations split by a phylogeographic break (Colbeck et al., 2008). Our delineation of the Maritime Provinces breeding population in the far northeast portion of the range corresponds with the Newfoundland population from the mtDNA analysis. Colbeck et al. (2008) hypothesized that the phylogeographic separation of Newfoundland and mainland American Redstart populations was the result of two refugia during Pleistocene glaciations. Our findings of weak genetic differentiation between the Maritime Provinces and other breeding populations suggest that there is ongoing gene flow among these populations. However, the limited admixture of individuals sampled in Newfoundland supports the notion that geographic separation of the island provides some barrier to gene flow, which has been demonstrated in several other avian species . The weakest genetic differentiation was found among the Western Boreal, Northern Temperate, and Southern Temperate breeding populations (FST: 0.004-0.006), suggesting limited barriers to gene flow. The Basin Rockies breeding population had higher genetic differentiation with the eastern breeding populations than the more northern Western Boreal population, which corresponds to the Great Plains functioning as a barrier to gene flow.
Using the five genetically distinct breeding populations allowed us to document at a fine scale more complex migratory patterns than previously identified. At the continental scale, our results broadly correspond to previous stable isotope studies that found eastern breeding American Redstarts overwintered in the eastern nonbreeding range and western breeders overwintered in the west . In several other species of Nearctic–Neotropical migrants, similar patterns of parallel migration have also been observed . However, in contrast to previous isotope analyses in American Redstart , our use of genomic data allowed us to clearly differentiate Maritime Provinces and Northern Temperate breeding birds and revealed that individuals breeding in the Maritime Provinces do not follow the parallel migration pattern. Parallel migration would result in these breeders being found in the far eastern portion of the nonbreeding range (e.g., Lesser Antilles and Trinidad and Tobago). Instead, individuals from the Maritime provinces bypass the Caribbean portion of the nonbreeding range and have a “leap-frog” migratory pattern to eastern Colombia. One explanation for the discordance of the Maritime Provinces migratory connectivity patterns from the rest of the breeding populations is the phylogeographic separation of these regions documented by Colbeck et al. (2008). Migration routes are influenced by the historical separation of Pleistocene glacial refugia and in the American Redstart, an Atlantic Shelf (near the Maritime Provinces) and eastern continental refugia are hypothesized to have caused the observed phylogeographic separation of these regions . A previous mtDNA analysis of American Redstart migratory connectivity only detected several individuals from the Maritime Provinces population in the Caribbean islands of the Dominican Republic and Puerto Rico but lacked samples from South America . Our results suggest that the Maritime Provinces breeding population has the strongest connectivity with eastern Colombia, though the full extent of the population’s nonbreeding distribution throughout South America is unknown.
Our use of genomic data allowed us to characterize migratory connectivity at a fine scale and identify distinct regions on the wintering range that separate breeding populations. In southern Central America, a clear split between the two sampling sites in Costa Rica, a separation of 360 km, occurs where the northern site predominantly has individuals from the Western Boreal breeding population and the southern site has individuals from the Southern Temperate population. This split corresponds with a biogeographic separation of drier broadleaf forest in the northern Pacific side of Costa Rica and moist broadleaf forest in the southern Pacific side (Corrales, Bouroncle, & Zamora 2015). Another geographic split in breeding origin occurs between the Lesser Antilles (Southern Temperate) and the Greater Antilles (Northern Temperate) in the Caribbean. Sampling of American Redstarts in Colombia was limited to the eastern slopes of the East Andes, and may not represent the wider Andes, given that the three chains of the Andes that run through Colombia influence connectivity patterns in the Canada Warbler,Cardellina canadensis Further population assignment studies that include sampling of American Redstarts from the Central and Western Andes, and the Caribbean region of Colombia, may identify the Andes Mountains as another barrier in the nonbreeding region, creating geographic splits in breeding origin for this species.