A model for the uncertainty around the yearly trawl-acoustic estimate of biomass of Barents Sea capelin, Mallotus villosus (Müller)

Abstract

Management of Barents Sea capelin, Mallotus villosus (Müller), is based on the precautionary approach, and probabilistic short-term predictions are used directly as a tool. One important part of the overall uncertainty is the uncertainty in the yearly trawl-acoustic estimate made in September. Harvest-control rules are based on historical time-series for the spawning stock, the uncertainty around which therefore has a direct bearing on management of the species. Accordingly, one must quantify the uncertainty, not only of the September estimate the latest year, but also for the whole time-series. In this paper, a model for the uncertainty around the yearly trawl-acoustic estimate is developed. The uncertainty in the mean integrator value by standard one-by-two-degree rectangles is evaluated with a model for the distribution of the basic five-mile integrator values parameterized with data from many historical surveys. The uncertainty from the biological samples is quantified on the basis of the multinomial distribution. A large number of replicates of the historical time-series of September estimates are produced and stored on file for later use.