Tracking cohorts: Analysis of migration in the early life stages of an estuarine fish

Larvae and juveniles are often most abundant at some distance from where spawning occurs. Such apparent distribution shifts can be the result of larval migration, seasonal shifts in the distribution of spawners, or spatial heterogeneity in mortality. In the present study, a cohort-based method was adopted to test for larval migration of bay anchovy (Anchoa mitchilli) and to estimate its rate and timing. A spatially and temporally extensive series of ichthyoplank-ton samples was collected hi the lowest 122 km of the Hudson River in summer 1998. Bay anchovy eggs and young-of-year individuals were abundant (mean density =1.6 m^-3 and 0.95 m^-3, respectively). Daily age used to define cohort membership was determined from otoliths. Age and length were strongly related (R² = 0.74). The age-length relationship varied over the season such that early-season bay anchovy hatched in June grew more rapidly than later-season anchovy hatched in August. Cohorts were defined as week-long birth-date classes, and net migration patterns of 11 cohorts were analyzed. Mean location shifted upriver as cohorts increased in age to 6 wk, at a net rate of about 0.6 km d^-1. Cohort abundance usually decreased over time in downriver regions, but usually increased in upriver regions. The cohort analysis confirmed that the upriver shift in distribution of early-stage anchovy is not entirely due to changes in adult spawning behavior or to clines in larval mortality, and must be partly the result of larval migration.