MEPS 214:103-110 (2001)  -  doi:10.3354/meps214103

ABSTRACT: An unusually large number of replicated macrofaunal samples (70) was taken from the Western Baltic in May 1995 for a ringtest in an ICES/HELCOM intercalibration exercise. This data set was employed in this study in order to investigate the performance of numerical methods used for predicting species richness and to assess the accuracy of the estimates of abundance and diversity currently used in benthic ecology. The results of this study indicate that: (1) more than 10 replicates are required in order to include in the data set more than two-thirds of the species found in 70 replicates, and more than 53 replicates are required in order to include 95% of the species; (2) estimates of average abundance and of average Shannon-Wiener diversity index using 5 replicates could result in less than 40% error; this could be less than 30% for 10 replicates and less than 5% for 70 replicates; (3) both types of species-richness predictions (jackknife estimate and S ₀) increased with increasing number of samples used in the calculations, indicating that their ability to assess overall species richness in the community is rather limited; in particular, it is shown that jackknife overestimates and S₀ slightly underestimates species richness. Different configurations of the S₀ method were tested in order to optimize its performance, and it was found that both truncation and increasing sampling lag result in increased and stabilized estimates of species richness.

KEY WORDS: Macrobenthos sampling methodology · Species-area curves · Abundance · Diversity · Western Baltic