The information content of stock-and-recruitment data and its non-parametric classification

Abstract

This paper considers the information content of stock-and-recruitment data and its non-parametric classification. Stock-and-recruitment observations representing several species have been classified as states according to whether they are low stock/low recruitment; low stock/high recruitment; high stock/high recruitment; or high stock/low recruitment. The classification makes it possible to calculate transitions among the states, steady-state probabilities, and first-passage times. If there is reason to believe that there is a “good” stock-and-recruitment relationship, then the parametric approach is better than the non-parametric approach, but if the relationship is poor the non-parametric approach may provide more useful management advice, in addition to taking into accountthe temporal relation among stock-and-recruitment points. The paper emphasizes that a major problem with either the parametric or the non-parametricapproach is that inferencesfor the future from these analyses depend upon the assumptionthat the unidentified mechanisms that caused recruitment to vary in the past will continue to behave in the same way in the future.