Differences in performance among four indices used to evaluate diversity in planktonic ecosystems

The performance of four common estimators of diversity are investigated usi ng calanoid copepod data from the Continuous Plankton Recorder (CPR) survey . The region of the North Atlantic and the North Sea was divided into squar es of 400 nautical miles for each 2-month period. For each 144 possible cas es, Pielou's pooled quadrat method was performed with the aims of determini ng asymptotic diversity and investigating the CPR sample-size dependence of diversity estimators. It is shown that the performance of diversity indice s may greatly vary in space and time (at a seasonal scale). This dependence is more pronounced in higher diverse environments and when the sample size is small. Despite results showing that all estimators underestimate the 'a ctual' diversity, comparison of sites remained reliable from a few pooled C PR samples. Using more than one CPR sample, the Gini coefficient appears to be a better diversity estimator than any other indices and spatial or temp oral comparisons are highly satisfactory. In situations where comparative s tudies are needed but only one CPR sample is available, taxonomic richness was the preferred method of estimating diversity. Recommendations are propo sed to maximise the efficiency of diversity estimations with the CPR data. (C) 2001 Ifremer/CNRS/IRD/Editions scientifiques et medicales Elsevier SAS.