Contrast across a boundary: Differing histories of two eelpout populations on a continuous continental slope

Abstract

Detecting drivers of population divergence across phylogeographic breaks is a major challenge in marine population genetics, partly due to uncertainties in molecular dating. Although the calibration of demographic transition (CDT) method is the most promising rate calibration technique for population-level events, there is a risk of making the false assumption that post-glacial population expansions coincided with global temperature rise. Further generalization of the technique is required to evaluate population histories of marine species whose abundance may be independent of changes in temperature. We explored the population history of a dwarf eelpout Petroschmidtia teraoi on Japanese continental slopes in the Sea of Japan and Sea of Okhotsk as a model organism for phylogeographic study. For rate calibration, we estimated historical stock sizes from a species distribution model and incorporated this data into the CDT method as a proxy for demographic transition (we term this combined method extended CDT, or eCDT). Our species distribution model showed that the range of P. teraoi is mainly determined by water depth and locality (i.e., latitude and longitude). Reconstructed historical stock sizes revealed contrasting demographic trends in the northern and western Sea of Japan on either side of the Noto peninsula, located at the midpoint of the Japanese mainland. We also found a sharp genetic break in the same location, and inferred contrasting demographic histories of local populations from the molecular data. The striking differences in the demographic histories of local populations even on apparently continuous continental slopes are most likely due to differences in available habitat area. eCDT appeared to work well as a rate calibration method for our temperature-independent study species, and our species distribution model revealed that local populations could have diverged due to discontinuities in available habitat. Finally, we assessed the strengths and weakness of the novel calibration method by analyzing simulated sequences.