Estimating contemporary early life‐history dispersal in an estuarine fish: integrating molecular and otolith elemental approaches

Abstract

Dispersal during the early life history of the anadromous rainbow smelt, Osmerus mordax, was examined using assignment testing and mixture analysis of multilocus genotypes and otolith elemental composition. Six spawning areas and associated estuarine nurseries were sampled throughout southeastern Newfoundland. Samples of adults and juveniles isolated by > 25 km displayed moderate genetic differentiation (F(ST) ~ 0.05), whereas nearby (< 25 km) spawning and nursery samples displayed low differentiation (F(ST) < 0.01). Self-assignment and mixture analysis of adult spawning samples supported the hypothesis of independence of isolated spawning locations (> 80% self-assignment) with nearby runs self-assigning at rates between 50 % and 70%. Assignment and mixture analysis of juveniles using adult baselines indicated high local recruitment at several locations (70-90%). Nearby (< 25 km) estuaries at the head of St Mary's Bay showed mixtures of individuals (i.e. 20-40% assignment to adjacent spawning location). Laser ablation inductively coupled mass spectrometry transects across otoliths of spawning adults of unknown dispersal history were used to estimate dispersal among estuaries across the first year of life. Single-element trends and multivariate discriminant function analysis (Sr:Ca and Ba:Ca) classified the majority of samples as estuarine suggesting limited movement between estuaries (< 0.5%). The mixtures of juveniles evident in the genetic data at nearby sites and a lack of evidence of straying in the otolith data support a hypothesis of selective mortality of immigrants. If indeed selective mortality of immigrants reduces the survivorship of dispersers, estimates of dispersal in marine environments that neglect survival may significantly overestimate gene flow.