EGG PRODUCTION ESTIMATES OF ANCHOVY BIOMASS IN THE SOUTHERN BENGUELA SYSTEM

Abstract

The spawning biomass of the southern Benguela anchovy (Engruulis cupensis) stock has been estimated annually since 1984 by means of acoustics and the egg production method (EPM), both techniques being integrated into one survey. In 1985 and 1986 the survey area was stratified on the basis of expected fish distribution patterns, and the transects were randomized within limits to allow rigorous investigation of the survey variance of both the EPM and acoustic estimates. The acoustic estimates of relative fish density in the vicinity of trawl stations were incorporated to weight individual trawl parameters for the egg production method, and within-stratum variances were estimated from the weighted mean parameter values for each transect. In 1986 the acoustic and EPM estimates agreed within 15%, but in 1985 the acoustic estimate was about 30% greater than the EPM estimate. High coefficients of variation of 0.35 to 0.41 were associated with the EPM estimates, mainly reflecting imprecision in the estimation of egg mortality rate and female spawning fraction. This paper briefly presents the results of the surveys, and provides a complete description of the statistical methods appropriate to a randomized, stratified design for egg production surveys, incorporating the necessary weighting factors for trawl parameters.