An Evaluation of Techniques Used to Index Recruitment Variation and Year-Class Strength

Recruitment is one of the most important factors influencing fish population dynamics, and although long-term data are probably the best source of information for explaining recruitment dynamics, such data are often unavailable due to logistic or monetary constraints. Consequently, a number of indices have been developed to index recruitment variability and year-class strength from a single sample. The purpose of this study was to compare two indices, the recruitment variability index (RVI) and recruitment coefficient of determination (RCD), with empirical estimates of walleye Sander vitreus recruitment variability in eight Kansas reservoirs. In addition, Studentized residuals from catch curves were compared with empirical estimates of year-class strength. Empirical estimates of recruitment variability were not related to RCD or RVI values from a single sample, and Studentized residuals from catch curves did not adequately index walleye recruitment for any of the study populations. Sampling bias was probably not the cause for the observed patterns, but examination of age-specific mortality rates indicated that mortality differed among ages (e.g., the mortality of age-2 fish was less than that of age-3 and older fish) and that mortality of a given age differed through time (i.e., the mortality of age-3 fish varied among years). These results suggest that angler behavior, in response to strong and weak year-classes, may affect the ability of indices to adequately assess recruitment dynamics. Mean RVI values estimated from multiple years of age structure data were significantly correlated with empirical estimates of walleye recruitment variability, suggesting that the RVI may be useful for indexing recruitment when data are available for more than 1 year. Although the indices examined in this study provide adequate assessments of recruitment dynamics for many fish populations, these results illustrate the importance of differential mortality among age-groups on index values and suggest that some form of index validation is necessary prior to their use.