Depensation in fish stocks: a hierarchic Bayesian meta-analysis

Abstract

The probability of different levels of depensation within four taxonomic groups was calculated using a Bayesian technique called hierarchical modeling. With this method we combined spawner-recruit data from many stocks within a taxon to estimate the distribution describing the variability of depensation within that taxon. The spawner-recruit model we use allows for both depensation (lower than expected recruits at low population levels) and hypercompensation (where recruits are higher than expected at low population levels). The end product of our analysis is a probability distribution that can be used as a Bayesian prior when analyzing a new data set. We examined four taxonomic groups (the salmonids, gadiforms, clupeiforms and pleuronectiforms) and found that, for all of the taxa, the most likely values fell close to or within the range of no depensation. However, because the distributions were very broad we suggest that analysis of stock recruitment data should incorporate spawner-recruit curves that include the possibility of depensation and hypercompensation.