Nonbreeding assignment
Implicit in the assignment of the nonbreeding individuals to breeding populations is that the breeding origin of these individuals is unknown. Given that assignment to the ESSBPs had an accuracy of 99.4% (168 out of 169 samples) for individuals with sequencing coverage of 0.6X – 2.0X, and accuracy of 97.9% (46 out of 47 individuals) for individuals down sampled to sequencing coverage of 0.01X, we assumed that we could correctly assign nonbreeding individuals (sequencing coverage range: 0.02X – 2.0X, mean 1.3X) with high confidence. Assignment of the 148 nonbreeding individuals resulted in the largest number of individuals being assigned to the Northern Temperate population (n = 64) and the least number of individuals being assigned to the Basin Rockies population (n = 2; Table S5). Of the 148 individuals, 139 individuals had assignment consistency of at least 0.8 for the 10 subsets of data, and these individuals were used to infer migratory connectivity. Testing consistency of assignment on the 47 breeding individuals identified three individuals with assignment consistency of < 0.8. One of these individuals from Minnesota was previously identified as an outlier in the PCA, and the other two individuals were from Pennsylvania, which is on the boundary of the Southern Temperate and Northern Temperate populations.
Mapping of the nonbreeding assignment results revealed patterns of strong migratory connectivity across the breeding range. Notably, the Maritime Provinces breeding population has strong connectivity with eastern Colombia, the Northern Temperate breeding population with the Greater Antilles, the Southern Temperate breeding population with the Lesser Antilles, and the Western Boreal breeding population with Central America and Mexico.