Compositing and subsampling to reduce costs and improve power in benthic infaunal monitoring programs

Abstract

Monitoring of soft-sediment infauna can be a useful tool in impact assessment, but sorting of specimens from sediments may be very time consuming and expensive. This can force a reduction in the intended sampling program, leading to low statistical power and increased uncertainty in decision making. We pooled field samples of marine infauna into composites, then exhaustively subsampled the composites to examine the statistical distribution of organisms in the subsamples. Abundances of organisms among subsamples were generally consistent with a random distribution, with deviations from randomness occurring more frequently for standard replicates than for subsamples. For the most abundant taxa, subsamples had much smaller variances than replicates, and in 76% of cases, the type of sample (i.e., subsample or replicate) had no effect on the mean abundance of organisms recorded. As our subsamples appeared generally unbiased, we then considered the cost and power for hypothetical monitoring programs using an MBACI analysis, and found substantial cost savings and increases in power to detect impact by using compositing and subsampling.