Significant differences in the length–weight relationships of neighbouring stocks can result in biased biomass estimates: Examples of haddock (Melanogrammus aeglefinus, L.) and whiting (Merlangius merlangus, L.)

Length–weight relationships of fish are often used to estimate the biomass of length distributions or to obtain indices of condition. Although large-scale spatial trends are known to exist, it is often assumed that length–weight relationships do not vary significantly within stocks or between neighbouring stocks. The present study examined length–weight relationships of 1334 haddock (Melanogrammus aeglefinus) and 1186 whiting (Merlangius merlangus) collected on a groundfish survey in the waters around Ireland in 2004. Additionally, condition indices were estimated for individual fish and for length-frequency samples in order to summarise the length–weight relationships. The length–weight regression showed a significant area effect but no significant differences existed between the sexes. The condition indices showed a moderate spatial structure for both species: around 25% of the variability could be explained by the location of the samples, the rest of the variability was due to other sources. Length–weight relationships did not appear to vary significantly within stocks, however differences between stocks were significant. In the present case, a bias of up to 10% could occur in biomass estimates as a result of applying length–weight relationships of one stock to length data of a neighbouring stock.