Fitting and comparing competing models of the species abundance distribution: assessment and prospect

Tom Matthews, Robert J. Whittaker

Research output: Contribution to journalReview articlepeer-review

132 Downloads (Pure)


A species abundance distribution (SAD) characterises patterns in the commonness and rarity of all species within an ecological community. As such, the SAD provides the theoretical foundation for a number of other biogeographical and macroecological patterns, such as the species–area relationship, as well as being an interesting pattern in its own right. While there has been resurgence in the study of SADs in the last decade, less focus has been placed on methodology in SAD research, and few attempts have been made to synthesise the vast array of methods which have been employed in SAD model evaluation. As such, our review has two aims. First, we provide a general overview of SADs, including descriptions of the commonly used distributions, plotting methods and issues with evaluating SAD models. Second, we review a number of recent advances in SAD model fitting and comparison. We conclude by providing a list of recommendations for fitting and evaluating SAD models. We argue that it is time for SAD studies to move away from many of the traditional methods available for fitting and evaluating models, such as sole reliance on the visual examination of plots, and embrace statistically rigorous techniques. In particular, we recommend the use of both goodness-of-fit tests and model-comparison analyses because each provides unique information which one can use to draw inferences.
Original languageEnglish
Pages (from-to)67-82
Number of pages16
Journal Frontiers of Biogeography
Issue number2
Publication statusPublished - Jun 2014


  • Bayesian statistics
  • information theory
  • macroecology
  • maximum entropy
  • maximum likelihood
  • rank–abundance plot
  • species abundance distribution


Dive into the research topics of 'Fitting and comparing competing models of the species abundance distribution: assessment and prospect'. Together they form a unique fingerprint.

Cite this