Abstract. Validation of the age determination procedure using otoliths of European anchovy in the Bay of Biscay was achieved by monitoring very strong year-classes in successive spring catches and surveys, as well as the seasonal occurrence of edge types. Historical corroboration of the ageing method was obtained by cross-correlation between successive age groups by year-classes in catches and surveys . Summary annual growth in length is also presented. Yearly annuli consist of a hyaline zone (either single or composite) and a wide opaque zone, disrupted occasionally by some typical checks (mainly at age-0 and age-1 at peak spawning time). Age determination, given a date of capture, requires knowledge of the typical annual growth pattern of otoliths, their seasonal edge formation by ages and the most typical checks. Most opaque growth occurs in summer and is minimal (translucent) in winter. Opaque zone formation begins earlier in younger fish (in spring), and this helps distinguish age-1 from age-2þ.

The distance between successive echosounder transmissions or “pings” is the main parameter determining the resolution of sampling along a transect in acoustic surveys. An increase in either ping interval or vessel speed increases the inter-ping distance, hence decreasing the effective resolution. In this study, we analysed whether a loss in along-transect resolution affects the mean acoustic backscattering energy, leading to uncertainty and/or bias in abundance estimates. To this end, we echo-integrated a real acoustic survey, followed by the application of a systematic resampling scheme to simulate a decrease in pinging resolution. For each transect, the mean NASC calculated at each resolution was compared with the mean NASC at the original resolution. Transects were characterised according to their heterogeneity and spatial autocorrelation to investigate their effect on the relationship between abundance error and sampling resolution. Uncertainty was seen to increase with decreasing resolution, with higher heterogeneity and lower spatial autocorrelation accelerating the rise in imprecision. Although the mean bias across replicates was zero, the asymmetry of the bias distributions increased with decreasing resolution, leading to an increasing probability and magnitude of underestimation (https://aztigps.shinyapps.io/PingRateStudio/).