Modelling dimensionality in species abundance distributions: description and evaluation of the Gambin model

Abstract

Problem: One of the longest standing problems in ecology has been the identification of an ecologically meaningful model that adequately and parsimoniously describes the relationship between the number of species and the numbers of individuals of each species in a sample. Methods and results: Here, we develop a new species abundance distribution model based on the gamma distribution and termed the Gambin model. In this model, a single variable, α, determines the shape of the gamma distribution and the ‘dimensionality’ of the sampled community. Using 33 existing data sets, we show that the model produces species abundance distributions that can be fitted to a wide range of marine and terrestrial data and that the Gambin model outperforms three prominent alternative models: the log-series, lognormal, and zero sum multinomial models. Conclusions: The Gambin α parameter provides a form of diversity metric, reflecting the complexity of a community’s interactions with its environment.