Stochastic dynamic population model of North Sea sandeels, and its application to precautionary management procedures

Abstract

Sandeels (sand lance) are important prey for fish, bird and mammal predators. In the North Sea, they are also the target of the region's largest 'single-species' fishery, with annual yields of ca. 0.8 million tonnes, mostly (>90%) comprising the species Ammodytes marinus. This species is relatively short-lived, and has large, naturally occurring, population fluctuations that confound the use of traditional methods of fishery forecasting for their management. Here we present a stochastic age-structured model that describes the population dynamics of North Sea sandeels. Recruitment to the population occurs according to a stochastic, autocorrelated Ricker function. Uncertainty in the recruitment parameters was quantified by bootstrap analysis, leading to 3 different versions of the model that allowed a range of potentially viable recruitment scenarios to be explored. Monte Carlo simulations were performed under different fishing mortality levels, and the responses of the population assessed. The model is used to quantify recruitment and the likelihood of the sandeel spawning stock biomass falling below 600 000 and 430 000 t, which represent the precautionary and biomass limit levels, respectively, imposed by managers upon the stock.