Skip to main content
Dryad logo

Allee effects and the allee-effect zone in northwest atlantic cod

Citation

Perälä, Tommi; Kuparinen, Anna; Hutchings, Jeffrey (2021), Allee effects and the allee-effect zone in northwest atlantic cod, Dryad, Dataset, https://doi.org/10.5061/dryad.r4xgxd2dg

Abstract

According to the theory of compensatory dynamics, depleted populations should recover when the threat responsible for their decline is removed because per capita population growth is assumed to be highest when populations are at their smallest viable sizes. Yet, many seriously depleted fish populations have failed to recover despite threat mitigation. Atlantic cod (Gadus morhua) off Newfoundland despite thirty years of dramatically reduced fishing mortality and numerous fishery closures has not recovered suggesting that drivers other than fishing regulate the growth of collapsed fish populations, inhibiting or preventing their recovery. Here, using Bayesian inference, we show strong evidence of Allee effects in a south Newfoundland cod population, based on data on recruitment and spawning stock biomass. We infer the Allee-effect threshold, below which recovery is impaired. We demonstrate the necessity of data at low population sizes to make inferences about the nature of low-abundance dynamics. Our work indicates that Allee effects are not negligible in commercially exploited fish populations, as commonly projected, and that they represent an inhibitory force that can effectively prevent recovery from overfishing. Our findings contrast with prevailing fisheries management practices that assume compensatory dynamics at low abundances with potential to seriously overestimate the recovery potential of collapsed populations.

Methods

The stock and recruitment data was extracted from 2021 Stock Assessment of NAFO Subdivision 3Ps cod.

Usage Notes

The data used in the codes is in Matlab's mat-file format. For easier access it is also presented in a csv-file. The codes are written in Matlab and Stan. One needs Matlab, Cmdstan, MatlabStan and MatlabProcessManager to run the codes. All analyses are performed and the article figures produced by running main.m script.