Development of a stock–recruit model for simulating stock dynamics for uncertain situations: the example of Northeast Atlantic mackerel (Scomber scombrus)

• Published in: ICES Journal of Marine Science
• Main Authors: Simmonds, E. John, Campbell, Andrew, Skagen, Dankert, Roel, Beatriz A., Kelly, Ciaran
• Format: Article in Journal/Newspaper
• Language: English
• Published: Oxford University Press (OUP) 2011
• Subjects: Ecology; Aquatic Science; Ecology, Evolution, Behavior and Systematics; Oceanography; Northeast Atlantic
• Online Access: http://dx.doi.org/10.1093/icesjms/fsr014; http://academic.oup.com/icesjms/article-pdf/68/5/848/29138724/fsr014.pdf

Abstract Simmonds, E. J., Campbell, A., Skagen, D., Roel, B. A., and Kelly, C. 2011. Development of a stock–recruit model for simulating stock dynamics for uncertain situations: the example of Northeast Atlantic mackerel (Scomber scombrus). – ICES Journal of Marine Science, 68: 848–859. The assumption of a relationship between recruitment and a spawning stock is the cornerstone of the precautionary approach and may constrain the use of a maximum sustainable yield (MSY) target for fisheries management, because the failure to include such a relationship suggests that providing a measure of stock protection is unnecessary. The implications of fitting different functional forms and stochastic distributions to stock-and-recruit data are investigated. The importance of these considerations is shown by taking a practical example from management: the management plan for Northeast Atlantic mackerel (Scomber scombrus), a fish stock with an average annual catch of 600 000 t. The historical range of spawning-stock biomass is narrow, and historical data from a stock assessment explain only a small proportion of the recruitment variability. We investigate how best to reflect the uncertainty in the stock–recruit relationship. Selecting a single model based on simple statistical criteria can have major consequences for advice and is problematic. Selecting a distribution of models with derived probabilities gives a more complete perception of uncertainty in dynamics. Differences in functional form, distribution of deviations, and variability of coefficients are allowed. The approach appropriately incorporates uncertainty in the stock–recruit relationship for FMSY estimation.