Performance of neutral model analysis in a spatio-temporal series of macrobenthic replicates.

The macrofaunal composition of 240 samples from the continental shelf of Crete, Greece, taken at 4 different depths and in 6 different months of the year, were used in order to test the performance of neutral model analysis in comparison to more traditionally used diversity indices. It was found that neutral model analysis produces less repeatable results in comparison to most of the diversity indices. The deviation statistic V was found to be negative in most of the samples (88%) and significantly negative in 34% of these. However no other indication of disturbance was found at the community level as far as the species composition was concerned. The negative deviation from neutrality was found to be stronger in the samples taken from 100 m depth, negligible in the 70 m depth samples, and intermediate in the deeper ones (130 and 170 m depth). Sampling season was not consistently related to the deviation from neutrality. Sample size was found to affect significantly all the diversity indices, but only in the case of the V statistic were the average values of the 24 data sets not correlated to the pooled sample values.