Distribution/abundance relations in a New Zealand grassland landscape

There are many examples in the literature of a positive correlation between the distribution of a species and its local abundance, i.e., widely occurring species tend to be more abundant locally when they do occur. Such relations have been documented over a wide range of taxa and spatial scales. There are five major hypotheses seeking to explain the relation: Random placement, Sampling error, Niche width, Demography, and Metapopulation dynamics. However, there is little evidence to distinguish between them, especially for plants. In this study, plant species presence in a short tussock grassland in inland Canterbury, New Zealand was sampled at three spatial scales (i.e. quadrat sizes; sizes of spatial grain): 0.2×0.3 m, 0.4×0.6 m, and 1.0×1.5 m, each divided into 0.1×0.15 m sub-quadrats for recording local frequency. The data were also analysed at different ranges of soil pH, water content, organic content and fertility (by bioassay). Significant positive correlations between species distribution and local abundance were found at the two larger spatial grain scales. There was no significant relation for native species, but the relation for exotic species was considerably stronger than that over all species. Reduction of environmental heterogeneity, by restricting the dataset to narrower ranges of soil factors, produced higher correlations at the two smaller grain scales. In contrast to reports in the literature, many species (most of them native) showed narrow regional distributions but high local abundance when they were at a site. It is concluded that the Sampling-error hypothesis cannot be an explanation of the results obtained here. Aspects of the results are inconsistent with the Random-placement, Demography and Niche-width hypotheses. The results are most compatible with the Metapopulation dynamics hypothesis. However, it is likely that several mechanisms are operating simultaneously. Whilst detailed analysis of empirical distributions at different spatial grains and over different ecological ranges can exclude some of the hypotheses, detailed studies of the autecology of species are the next necessary step.