SEARCH

SEARCH BY CITATION

Keywords:

  • metacommunities;
  • species sorting;
  • nestedness analysis;
  • simulations;
  • Type 1 error

Abstract

Nestedness analysis is a popular tool for inferring spatial species distributions, and therefore has management and conservation relevance. Ecologists frequently compute nestedness and subsequently use Spearman rank correlations for inferring relationships between the observed nested ranks of sites with biogeographic and environmental variables. Using temporary pond microcrustaceans hatched from microcosms as a case study, this paper shows that the application of this method can be problematic. While the overall degree and significance of nestedness was robust against a statistical error, the results obtained from randomly generated matrices, in which community structure from the original microcrustacean incidence matrix was maintained (fixed rows –fixed columns constraints), showed that rank correlations of observed nested patterns can be vulnerable to a Type 1 error (detecting an effect when there is none). Using expected nestedness patterns derived from rarefied original matrices to control for sample size effects did not change this result. This problem may have arisen as a result of a quantitative bias related to the disproportionate impact of rank positions of individual ponds in the analysis. Future extensive simulations studies, involving different community structures, should help identify the general reliability of rank correlation results in nestedness analyses. (© 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim)