26/03/2004 : New compilations of IndVal for WINDOWS are available..
Dufrêne, M. and P. Legendre, 1997. Species assemblages and indicator species: the need for a flexible asymmetrical approach. Ecological Monographs, 67 : 345-366.
IndVal is a new and simple method to find indicator species and species assemblages characterizing groups of samples.
The novelty of our approach lies in the way we combine a species' relative abundance to its relative frequency of occurrence in the various groups of samples. Indicator species are defined as the most characteristic species of each group, found mostly in a single group of the typology and present in the majority of the sites belonging to that group. This duality, which is of ecological interest, is rarely completely exploited in such analyses. In these cases, species occupying only one or two sites in one habitat group and present only in that group (rare species) receive the same indicator value as species occupying all sites of that habitat group and found only in that group. There is an important difference between these two species, however. The first one is an asymmetrical indicator: its presence cannot be predicted in all sites of one habitat, but contributes to the habitat specificity. The second species, on the contrary, is a true symmetrical indicator: its presence contributes to the habitat specificity and one can predict its presence in all sites of the group.
To take into account the second part of the symmetric indicator species definition in the expression of the indicator value, a new index was developed. For each species i in each site group j, we computed the product of Aij, the mean abundance of species i in the sites of group j compared to all groups in the study, by Bij, the relative frequency of occurrence of species i in the sites of group j, as follows:
Bij = Nsitesij/Nsites.j
IndValij = Aij * Bij * 100
where IndVal is the Indicator Value of species i in site cluster j. In the formula for Aij, which is a measure of specificity, Nindividualsij is the mean number of individuals of species i across sites of group j, while Nindividualsi. is the sum of the mean numbers of individuals of species i over all groups. We use the mean number of individuals in each group, instead of summing the individuals, because this removes any effect of the number of sites in the various site groups, and of differences in abundance among sites belonging to the same group. Aij is maximum when species i is only present in cluster j. In the formula for Bij, which is a measure of fidelity, Nsitesij is the number of sites in cluster j where species i is present, while Nsites.j is the total number of sites in that cluster. Bij is maximum when species i is present in all objects of cluster j. Quantities A and B must be combined by multiplication because they represent independent informations about the species distribution. Final multiplication by 100 produces percentages.
The indicator value of a species i for a typology of sites is the largest value of IndValij observed over all groups j of that typology: IndVali = max [INDVALij] This index is maximum (= 100%) when all specimens of a species are found in a single group of samples AND when the species occurs in all samples of that group.
The basic idea is to measure the species indicator value for all the levels of a hierarchical typology. It is assumed that, the IndVal index changes along the hierarchical typology, being high and decreasing after for generalist (core) species or being low and increasing after for more specialized (satellite) species when the sample groups are more numerous. The IndVal index allows also to identify species typical for intermediate level of the clustering history.
The statistical significance of the species indicator values is evaluated using a randomization procedure.
Contrary to Twinspan, our indicator index for a given species is independent of the other species' relative abundances, and there is no need to use pseudospecies.
The new method identifies indicator species for typologies of ecological samples obtained by any hierarchical or non-hierarchical classification procedure; its use is independent of the classification method.
Since indicator species give ecological meaning to groups of samples, this method provides criteria to compare typologies, to identify where to stop dividing clusters into subsets, and to point out the main levels in a hierarchical classification of samples. A new way to present species-sample tables, accounting for the hierarchical relationships among species, is proposed.
See the IndVal 2.0 manual for more details on the proposed methods and recent developments.
McGeoch, M.A. & Chown, S.L., 1998. Scaling up the value of bioindicators. Trends in Ecology and Evolution, 13 (2) : 47-48.
An application is available for Macintosh and PC (see below) as a summarized manual. A Unix version is also running on a SUN server at the UCL.
These programs are provided without any explicit or implicit warranty of correct functioning. They have been developed as part of a university- based research program. If, however, you should encounter problems, I will be happy to help solve them (M.Dufrene@mrw.wallonie.be). Researchers can use these programs for scientific purposes, but the source code remains the property of the author. By downloading these files, you accept the conditions of this agreement.
These documents are the part of a web server dedicated to Biodiversity in Southern Belgium (in French !)
Pour tout renseignement sur ce document : M.Dufrene@mrw.wallonie.be