Re-examination of stock–recruitment relationships: a meta-analysis

Abstract

Abstract Estimating stock–recruitment (S–R) relationships is a fundamental challenge in fisheries management. However, there is little consensus among fishery scientists on whether a significant quantitative relationship exists between spawning stock biomass (SSB) and recruitment, and whether an optimal model is available to represent this S–R relationship. In this study, we conducted a meta-analysis to determine the relationship between SSB and ln-transformed recruitment per spawning stock (RPS) using data from 92 stocks worldwide by applying four conventional S–R models (the Beverton–Holt, Ricker, Deriso–Schnute, and Pella–Tomlinson models). We also examined the residuals from standard model (SM) and measurement error model (MEM) regressions, which consider only the error in the y-variable and that in both the x- and y-variables, respectively, by integrating them after standardization. We showed that 90.2% of the stocks had significant negative correlations (i.e. density effects) between SSB and ln(RPS). The Pella–Tomlinson S–RPS model exhibited the most stable and high performance in both SM and MEM regressions. Our results suggest the Pella–Tomlinson S–RPS model as the most probable candidate to assess the dynamics of several stocks.