Weighting of acoustic- and trawl-survey indices for the assessment of North Sea herring
Abstract Acoustic surveys are used in 20 stock assessments within the ICES community, and almost all as relative indices of abundance, but little has been done to explore their performance in detail. The North Sea herring acoustic survey started in 1979 and by 1984 had become an internationally-coordinated survey conducted annually in July. Along with trawl- and larvae-survey indices, it has been used to tune a catch-at-age assessment model of North Sea herring. In this article, the precision of the survey is estimated, using data at ICES statistical-rectangle level from 1989 to 2001, and bootstrap-resampling methods modified by geostatistical estimates of the spatial autocorrelation. Similar techniques are applied to the larvae, Methot and trawl surveys that provide the other data on the distribution and abundance of North Sea herring. The comparison of survey performance is also examined using the bootstrap estimates of abundance to give 1000 simulated assessments of North Sea herring using the integrated catch-at-age (ICA) method. The results of these analyses are compared and the annual acoustic survey is shown to provide the most precise estimate of relative abundance for adult North Sea herring each year. The weighting of the various indices within the assessment is investigated. A weighting method is presented that provides a more precise method for estimating the stock. The more precise assessments are compared for retrospective pattern. An assessment is proposed which provides the most precise stock estimates with the best retrospective pattern. This assessment has been reviewed and accepted by the ICES Advisory Committee on Fisheries Management. The importance of the acoustic survey and its contribution to the assessment in relation to the other indices is discussed.