Measuring the effect of changes in catchability on the variance of marine survey abundance indices

Abstract

The average catch per tow by a research vessel survey is often used as an index of abundance. An estimate of the variance of such indices that is based only on the between station variability in catch may underestimate the true variance if catchability varies over time. In this paper, the survey index variance is estimated indirectly by cross-calibrating time series of virtual population analysis estimates and trawl survey indices of abundance. The method is applied to surveys of some fish stocks on Georges Bank, in southern New England, and in the Barents Sea. For these surveys, it appears that the true variance of the survey indices is approximately twice as large as the usual estimates based on the within survey variance. As an application, a time series technique, which requires an estimate of the survey index variance, is used to generate a more precise index of abundance. The results indicate that for the surveys examined the variance of the estimated abundance index is 30–40% smaller than the original (average catch per tow) index.