Sampling biological characteristics of krill: Effect of heterogeneous nature of swarms

An effect of the very patchy distribution of Antarctic krill (Euphausia superba Dana) in various forms of aggregation is that a single swarm may not provide an unbiased estimate of population parameters such as mean length of krill in the local area. Here, we analyse the number of samples required to estimate the characteristics of a local population as precisely as if there were no differences between krill swarms in terms of their biological composition. Krill were intensively sampled over different spatial and temporal scales around South Georgia in 1981 and 1982, and in the Bransfield Strait in 1985. These varied from replicate hauls at a single station over 24 h and repeat sampling in restricted areas over periods of 6 to 14 d to regional surveys around South Georgia and in the Bransfield Strait. Various biological characteristics were measured such as length, maturity, moult stage and feeding state. Depending upon the biological characteristic examined and the area covered by the sampling programme, the number of samples needed to obtain the same degree of precision as would be found in the absence of heterogeneity varied from 3 to > 80 samples. This has important implications for the design of net-sampling programme for monitoring krill populations.