Conceptual and statistical problems associated with the use of diversity indices in ecology

Abstract

Diversity indices, particularly the Shannon-Wiener index, have extensively been used in analyzing patterns of diversity at different geographic and ecological scales. These indices have serious conceptual and statistical problems which make comparisons of species richness or species abundances across communities nearly impossible. There is often no a single statistical method that retains all information needed to answer even a simple question. However, multivariate analyses could be used instead of diversity indices, such as cluster analyses or multiple regressions. More complex multivariate analyses, such as Canonical Correspondence Analysis, provide very valuable information on environmental variables associated to the presence and abundance of the species in a community. in addition, particular hypotheses associated to changes in species richness across localities, or change in abundance of one, or a group of species can be tested using univariate, bivariate, and/or rarefaction statistical tests. The rarefaction method has proved to be robust to standardize all samples to a common size. Even the simplest method as reporting the number of species per taxonomic category possibly provides more information than a diversity index value. Rev. Biol.

Alternative ways to measure impacts of dam closure to the structure of fish communities of a neotropical river

“…While diversity indices have been widely applied in impact assessment, the comparative analysis presented here highlights the limited information they can convey (Barrantes and Sandoval, ). As diversity indices respond to both richness and evenness, when both parameters change (with a disturbance) a diversity index may fail to detect degradation because they summarize ecological information in a single dimension (represented by an interval‐a number‐on a finite or infinite scale).…”

Monitoring Fungicide Sensitivity Levels and Mycelial Compatibility Groupings of Sclerotium rolfsii Isolates from Florida Peanut Fields

“…Simpon's index indicates that the likelihood of randomly selecting two individuals from the same MCG is low for Florida peanut fields by county. It is important to note that all diversity indices are affected by sample size, which was low in this study (Gotelli and Ellison, 2013;Barrantes and Sandoval, 2009;Soetaert and Help, 1990). However, both the Shannon and Simpson indices are widely used for discerning population diversity and their agreement supports the conclusion that high level of diversity was present.…”

Fish Species Richness in Lowland Streams with a Geothermal Input Phenomenon, Sarapiquí de Heredia, Costa Rica

“…To try to obtain more information from biodiversity indices, in particular alpha diversity index poses both a conceptual problem and a statistical problem [18]. Measures that combine species composition and abundance, such as Shannon-Wiener index, have already been pointed out to have no meaning by themselves, because an index value of a set of data is a reduction of information that makes interpretation difficult [18].…”

“…Measures that combine species composition and abundance, such as Shannon-Wiener index, have already been pointed out to have no meaning by themselves, because an index value of a set of data is a reduction of information that makes interpretation difficult [18]. Beta diversity index, such as Sørensen similarity index, and species richness information allowed to make proper comparisons of sites that evidently showed differences in species richness and abundance.…”