Abstract

Systematic acoustic surveys to estimate the Spanish fraction of the Atlantic sardine stock were begun in 1983. Since 1991, with the use of the new SIMRAD EK500 echo-sounder–echo-integrator, the area covered has been extended to the 1000 m depth contour to observe the complete distribution of the main pelagic species. The last three surveys, from spring 1991 to spring 1993, had a range of 20–1000 m depth. They were first analysed using traditional methodology, which does not give any estimation of the variance. The data were therefore analysed using geostatistical techniques and the resulting estimates of relative abundance were compared. Sardine showed high variability between years and zones both in distribution and density, expressed as number of fish per square nautical mile. The two methods of analysis gave different biomass estimates by zone, especially in 1993. Variograms computed over areas of fish presence did not show, in general, a clear spatial structure and sills were reached at 3 nmi of range. The precision of these variograms was low, ranging from 23–40%, expressed as relative standard error. In order to improve precision it is necessary to increase sampling intensity and change the survey grid. For sardine, a systematic parallel survey design with a random start and 6 nmi between transects would be more appropriate.

This content is only available as a PDF.