Sampling errors in anchovy egg abundance estimates using the PAIROVET net

A replicate haul experiment was carried out using the verticallytowed PAIROVET net in an area of high abundance of anchovy eggsin the Bay of Biscay. The main aim of the experiment was tocompare the performance of ships of different size working inthe same area. In addition, we tested the performance of thePAIROVET net and the deployment methodology used to sample anchovyeggs for biomass assessment purposes. We concluded that underthe experimental conditions experienced in this cruise, differentships yield similar performances of the vertically towed PAIROVETnet if the methodology is kept similar. The ANOVA suggests thatthe systematic central sampling (SCS) scheme currently in usefor the Bay of Biscay anchovy biomass estimations (one stationout at 3 miles) adequately represents the surrounding area alongthe transect line. The among-station component of variance withinthe unit sampling area of the current SCS turned out to be comparableand, in some cases, smaller than the within-station variance.Consequently, the spatial design of the SCS adopted during routinesurveys appears to be a reliable strategy for sampling anchovyeggs in areas of high egg abundance, as in this experiment.In addition, it ensures fairly independent data units. In thearea of the experiment, the most disperse egg stages (like 2-and 3-day-old eggs) show a larger within-station component ofvariance than the spatial component of variance within the rectanglesof the current SCS. This suggests that for these stages, thesampling can still be improved cheaply by increasing the precisionof egg abundance estimates per station with larger sample sizes,without changing the spatial sampling scheme or the number ofstations. However, the benefits of this strategy will hardlyaffect the precision of the 1-day-old egg estimates becauseof their patchy distribution.