AB 1:195-204 (2007)  -  DOI: https://doi.org/10.3354/ab00017

ABSTRACT: Stochastic niche apportionment models use a simple rule-based approach to simulate the process of community structure. These models were used to seek a general rule of community structure by investigating saltmeadow vegetation from New Zealand, South America and southern Europe. The patterns of relative abundances generated by 5 such models were then compared to those observed in the vegetation at both the community and intercontinental spatial scale. A randomization test was used to compare both the mean and variance of the ranks of the relative abundance distributions generated by Tokeshi’s 5 niche apportionment models with those of the observed data. Three of the 5 models produced patterns that were significantly different to the observed pattern at all sites, scales and resolutions. Higher resolution made some improvement in the discrimination between the models, but did not alter the general trend. The random fraction (RF) model fitted the saltmeadow vegetation at the intercontinental scale and 5 of the 6 saltmeadow sites at the community scale. This suggests that hierarchical structuring and biotic interactions may play a role in the structuring of saltmeadow communities. The consistent results at widely different spatial scales, and from sites with different species composition and history, suggest the possibility of a general rule of community structure. By allowing the mean and the variance generated by a specific model to be compared with that of the observed community, mechanistic niche-apportionment models offer a step forward in our understanding of the processes that may structure communities.

KEY WORDS: Salt marsh · Community structure · Tokeshi · New Zealand · Chile · Corsica · Italy · Randomization test