4.11 Indicator Species
4.11.1 Introduction
The search for indicator or characteristic species in groups of sites is a particularly
important question in applied ecology, where conservation and management specialists look for methods to quickly assess or predict the type of environment
characterizing a given location, to assess the type of living community (e.g.,
vegetation type) in a given area, or to monitor temporal changes in some ecological
variables (e.g., mean annual temperature). In fundamental ecology, one often looks
for ecological preferences of species among a set of site groups, either to gain
knowledge about the species themselves or to improve the knowledge of the habitats
sampled in the groups of sites. These approaches are complementary, and different
methods exist to address them. De Cáceres and Legendre (2009) presented the two
complementary approaches and their related indices, classified into indicator value
and correlation indices, respectively.
Indicator value indices are based on the concepts of specificity (highest when the
species is present in the target group but not elsewhere) and fidelity (highest when the
species is present in all sites of the target group). A high indicator value is obtained
by a combination of high specificity and fidelity.
Correlation indices rely on the computation of statistics related to Pearson’s
r correlation coefficient, measuring the relationship between a vector indicating the
presence or abundance of a species at different sites and a vector describing to which
predefined group each site belongs.
De Cáceres and Legendre (2009) describe 12 indices belonging to the two groups.
They state that indicator value indices are more useful for assessing the species
predictive values as bioindicators, e.g. for field determination of community types or
for ecological monitoring, whereas correlation indices should be used to determine
the ecological preference of a given species among a set of alternative site groups
(De Cáceres and Legendre 2009, p. 3573).
4.11.2 IndVal: Species Indicator Values
The Dufrêne and Legendre (1997) IndVal index is computed as the product of the
specificity of a species to the targeted group by its fidelity to the targeted group.
Specificity is defined by the mean abundance of the species within the targeted group
compared to its mean abundance across the groups; fidelity is the proportion of sites
of the targeted group where the species is present. Dave Roberts, the author of the
package labdsv used hereunder, summarized the concept as follows (pers.
comm.): “The indval approach looks for species that are both necessary and
sufficient, i.e. if you find that species you should be in that type, and if you are in
that type you should find that species”.
120
4 Cluster Analysis
4.11.1 Introduction
The search for indicator or characteristic species in groups of sites is a particularly
important question in applied ecology, where conservation and management specialists look for methods to quickly assess or predict the type of environment
characterizing a given location, to assess the type of living community (e.g.,
vegetation type) in a given area, or to monitor temporal changes in some ecological
variables (e.g., mean annual temperature). In fundamental ecology, one often looks
for ecological preferences of species among a set of site groups, either to gain
knowledge about the species themselves or to improve the knowledge of the habitats
sampled in the groups of sites. These approaches are complementary, and different
methods exist to address them. De Cáceres and Legendre (2009) presented the two
complementary approaches and their related indices, classified into indicator value
and correlation indices, respectively.
Indicator value indices are based on the concepts of specificity (highest when the
species is present in the target group but not elsewhere) and fidelity (highest when the
species is present in all sites of the target group). A high indicator value is obtained
by a combination of high specificity and fidelity.
Correlation indices rely on the computation of statistics related to Pearson’s
r correlation coefficient, measuring the relationship between a vector indicating the
presence or abundance of a species at different sites and a vector describing to which
predefined group each site belongs.
De Cáceres and Legendre (2009) describe 12 indices belonging to the two groups.
They state that indicator value indices are more useful for assessing the species
predictive values as bioindicators, e.g. for field determination of community types or
for ecological monitoring, whereas correlation indices should be used to determine
the ecological preference of a given species among a set of alternative site groups
(De Cáceres and Legendre 2009, p. 3573).
4.11.2 IndVal: Species Indicator Values
The Dufrêne and Legendre (1997) IndVal index is computed as the product of the
specificity of a species to the targeted group by its fidelity to the targeted group.
Specificity is defined by the mean abundance of the species within the targeted group
compared to its mean abundance across the groups; fidelity is the proportion of sites
of the targeted group where the species is present. Dave Roberts, the author of the
package labdsv used hereunder, summarized the concept as follows (pers.
comm.): “The indval approach looks for species that are both necessary and
sufficient, i.e. if you find that species you should be in that type, and if you are in
that type you should find that species”.
120
4 Cluster Analysis