Relative merits of using numbers and biomass in fish community studies

Abstract

In descriptions of fish communities, the question of which measure of abundance should be used, whether numbers or biomass, has never been addressed. While both measures are often available, the common practice is to use numerical abundance for such descriptions, without any explicit justification for this choice. In order to contribute to the clarification of this issue, we have compared correspondence analyses/TWIA-classifications performed on both the numerical density and the biomass of the same trawl catches from a region off the western coast of Africa. The situation in which the quantitative aspect of the samples is disregarded, represented by presence/absence analyses, is illustrated for comparison. As it appeared likely that the length of the ecological or biogeographical 'gradient' would be of importance in how different the results of the 2 abundance measures would emerge, a series of subsamples of the total material, representing shorter ecological gradients, was also analysed. The analyses show that in most situations the choice between numbers and biomass matters little. However, in the case of short ecological 'gradients', when all or the large majority of species are common to all samples studied, and the individual fish sizes are different, there may be a marked difference in the patterns shown by the alternative measures of abundance.