Species Difference In Regional Dispersal Ability And Species Abundance Distribution

Biodiversity and species abundance distribution are the essential contents of ecology. Proposition of ecological neutral theory has diverted ecologists' attentions back to this area again. Community neutral theory has achieved a great success in accurately predicting species abundance distribution than traditional niche theory. But because of criticisms and suspicions of apparently counterintuitive fundamental assumptions in neutral model, test and modificatoin of these controversial hypothesises have become the current hot spots. In this paper, we relaxed the strong assumption of the symmetrical dispersal from regional community to local community on the basis of Hubbell's neutral model. We found out that slight difference in dispersal ability among species into neutral model dramatically reduces species richness and result in considerable deviation of species abundance distributions from those predicted by neutral model. Difference in species immigration ability, if there is, can overwhelm the role of random drift and make local community dynamics becomes deterministic. Species with higher regional dispersal ability tends to have higher relative abundance in local communities. It's larger that the effect of slight difference in dispersal among species on relative species abundance and the amount of rare species. By Etienne's method, species relative abundance distribution curves under species differences in immigration rate can be fitted well by neutral model although we found out that the input parameters were very different from the paremeters calculated by Etienne's method, which implies that the neutral pattern does not mean neutral process.Species abundance distribution or relative species abundance distribution is the very important technique in studying community structure and assembling processes. There have been a lot of mathematical models proposed to explain the diverse species abundance distributions observed in natural communities. But a publicly accepted criterion has not been given rise to distinguish the goodness of fit among different mathematical models in fitting different species abundance distributions. Actually, the results of goodness of fit vary distinctly across different fitting methods. Here, we fitted the species abundance distribution randomly sampled in three different habitats on subalpine meadows of the eastern Tibetan Plateau using a great many of ecological mathematical models included in Ulrich's RAD program and found out that the amount of model that fitted well the north-facing slope is the largest, the second is level field, the least is the south-facing slope. Among the models we tested under two different testing methods, we found out that the geometric model fitted best. both determinant and stochastic models fitted the above three kinds of species abundance distribution under two different fitting methods.