Relations between Stock and Environmental Variables, and an Index of Abundance, for Juvenile Alewives and Blueback Herring

Abstract

The maximum of a series of catch-per-unit-effort (CPUE) estimates was postulated to be less biased as a juvenile abundance index (JAI) by the seasonal downstream migration of juvenile anadromous alewives Alosa pseudoharengus and blueback herring A. aestivalis than a timeaveraged or summed CPUE. The JAI for blueback herring was adjusted for outlier CPUE values occurring in response to high nighttime illumination (moon phase moderated by cloud cover). The JAI was not significantly related to spawning stock abundance, population fecundity, or year-class size even though spawning stock size was correlated with year-class size and future adult returns and juvenile densities were estimated with relatively good precision (±7–15% error). Dischargeinduced bias in the JAI masked intermediate-stage stock-recruitment relations. The validity and usefulness of an index of abundance for juvenile Alosa spp., as obtained by push-net survey methods, are doubtful. Small sample sizes and high variability of biological variables gave wide confidence intervals about correlation coefficients between variables and low power of tests of significance for correlation when the null hypothesis was not rejected. For both juvenile alewives and blueback herring, the JAI decreased with mean July–August discharge, whereas instantaneous daily mortality rate (Z) increased, after adjustment by partial correlation for spawning population fecundity. Instantaneous growth rate decreased with mean July–August water temperature for alewives and increased for blueback herring after adjustment for juvenile density. Differences in availability to the sampling gear between species, high CPUE variance, and biased estimates of the JAI, coupled with the long time series necessary to evaluate a stock–recruitment relation, limit the use of juvenile abundance indices to circumstances in which all requirements of validity, accuracy, precision, and time series length can be adequately met.