Effect of population structure, sampling strategy and sample size on the estimates of selection parameters for shrimp (Crangon crangon) trawls

In selectivity experiments with shrimp trawls, very high numbers of animals in the catches of single hauls are a common feature and therefore, sub-sampling is inevitable. In order to find an acceptable balance between work-load and accuracy in the estimation of the selection parameters, it is important to have a sensible idea on the minimum numbers of shrimps to be measured in each catch fraction (cover, discards and landings). The present theoretical study tries to answer this question by means of computer simulations of different sampling strategies and sample sizes applied to catches with known size compositions.

The results of the simulations are discussed in relation to population structure, shape of the cod-end selection curve, sampling strategy (as the relative amounts measured from the different catch fractions) and sample size, and the method used to calculate the selection parameters.

Samples of 750 animals provide an acceptable compromise between work-load and reliability of the estimated selection parameters, provided that sufficient numbers of length classes are available. Population structure, selection curve and sample size (in that order) largely determine the reliability of the estimates, whereas the sampling strategy, on the other hand, only has a minor effect.