Genomic data recover previously undetectable fragmentation effects in an endangered amphibian

A critical consideration when using molecular ecological methods to detect trends and parameterize models at very fine spatial and temporal scales has always been the technical limits of resolution. Key landscape features, including most anthropogenic modifications, can cause biologically important, but very recent changes in gene flow that require substantial statistical power to detect. The problem is one of temporal scale: Human change is rapid and recent, while genetic changes accumulate slowly. We generated SNPs from thousands of nuclear loci to characterize the population structure of New York-endangered eastern tiger salamanders (Ambystoma tigrinum) on Long Island and quantify the impacts of roads on population fragmentation. In stark contrast to a recent microsatellite study, we uncovered highly structured populations over an extremely small spatial scale (approximately 40 km² ) in an increasingly human-modified landscape. Geographic distance and the presence of roads between ponds were both strong predictors of genetic divergence, suggesting that both natural and anthropogenic factors contribute to the observed patterns of genetic variation. All ponds supported small to modest effective breeding populations, and pond surface area showed a strong positive correlation with population size. None of these patterns emerged in an earlier study of the same system using microsatellite loci, and we determined that at least 300-400 SNPs were needed to recover the fine-scale population structure present in this system. Conservation assessments using earlier genetic techniques in other species may similarly lack the statistical power for small-scale inferences and benefit from reassessments using genomic tools.