Range-wide population genetic analysis of Seaside Sparrows (Ammospiza maritima) supports at least five distinct population segments that do not align with current subspecies descriptions

As an obligate salt marsh species, Seaside Sparrows (Ammospiza maritima) are vulnerable to numerous threats including climate change, coastal erosion, sea-level rise, and both natural and anthropogenic disasters. Of the 9 recognized subspecies, 2 are extinct and 1 is endangered. Previous genetic analyses of mitochondrial DNA (mtDNA) and microsatellite loci showed that current taxonomy does not accurately reflect underlying genetic diversity, with possible consequences for the distribution of conservation resources. To further inform Seaside Sparrow management, we comprehensively describe genetic structure among 24 range-wide sampling locations that include all extant subspecies. We inferred population structure from several thousand single-nucleotide polymorphisms collected from 272 individuals via restriction-site-associated DNA sequencing. Principal components, pairwise FST values, and clustering approaches suggest that Seaside Sparrows on the Atlantic and Gulf Coasts are distinct and consist of at least 5 genetic clusters: 1 in southern Texas, 1 ranging from Aransas County, Texas, to Mississippi; 1 in western Florida; and 2 or 3 genetic groups intermixed along a gradient on the Atlantic Coast. These genetic clusters are not consistent with current subspecies taxonomy and could be used as distinct population segments (DPSs) to inform the most efficient allocation of resources to Seaside Sparrow conservation. Our results regarding the endangered subspecies, A. m. mirabilis, from southern Florida are inconclusive due to low sample size, but indicate that it is distinct and may represent a sixth DPS. Based on our genetic results, we recommend additional song and morphometric analyses in western Florida and a closer study of the boundary between the breeding distributions of A. m. maritima and A. m. macgillivraii to ensure the proper identification of DPSs.