Effects of year-class strength on age determination

Abstract

Little has been done to assess factors that can affect precision and accuracy in the fish age determination process. In this study, we examine how the presence of strong year-classes (i.e. foreknowledge of these year-classes) can affect the ages that are generated. It has been suggested recently that reader precision (i.e. reader repeatability) places limits on the accuracy of ages and that reader awareness of year-classes could both increase estimated precision and exaggerate the strength of strong year-classes beyond that which could be expected considering reader repeatability. That is, if reader repeatability was 60% at a given age, then a maximum of 60% of any strong year-class of that age could be expected to be aged correctly. However, consciously or subconsciously, age readers can be expected to tend to age toward strong year-classes. In order to examine these hypotheses, an otolith test sample was constructed for walleye pollock (Theragra chalcogramma) by combining samples that were collected in 3 adjacent years and that were known to include the strong 1978 year-class. Thus, the test sample gave the appearance of consisting of three neighbouring age-classes in nearly equal numbers. The age readers involved in this study had results that were better than might be expected, maintaining their reader precision even with this difficult sample. However, the results also appear to indicate that the 'dominant' year-class was spread to other ages in comparison with ages obtained when the age readers were aware of the strong year-class.