You are on page 1of 6

Ecological Modelling 130 (2000) 151 – 156

A link between ecological diversity indices and measures of
János Izsák a,*, László Papp b
Berzsenyi Dániel College, Department of Zoology, Károlyi G. tér 4, H-9701, Szombathely, Hungary
Hungarian Natural History Museum, Department of Zoology, Baross u. 13, H-1088, Budapest, Hungary


The practice of environmental planning and protection frequently necessitates the quantification of ecological
diversity. Traditional ‘ecological diversity indices’ are based on the abundances of species present. However, such
indices are insensitive to taxonomic or similar differences. With equal species abundances they measure the species
richness (species number) only. Conversely, so-called ‘biodiversity indices’ are based on species differences, but are
insensitive to the abundance conditions. The quadratic entropy index is the only ecological diversity index, the value
of which reflects both the differences ‘and’ abundances of the species. When a species list is given without abundance
data, then, using the quadratic entropy index and postulating equal abundances, one gets the only biodiversity index
derived from a traditional ecological index of diversity. Its extensive form is identical with the sum of differences or
distances between the species present. This index trivially satisfies set monotonicity, an important property for
biodiversity indices. © 2000 Elsevier Science B.V. All rights reserved.

Keywords: Biodiversity measures; Diversity indices; Quadratic entropy; Richness

1. Introduction dance conditions of the species are unessential. As
for the term ‘biological diversity’, it dates back to
Biodiversity is a central idea in the practice of the early 1980s. Perhaps Lovejoy (1980) used it
quantifying the ecological status of different first in the sense of the number of species present.
biotops by known abundances of species. How- The contracted form ‘biodiversity’ was coined by
ever, in large-scale environmental protection, the Rosen in 1985 (cf. Harper and Hawksworth,
species abundances are mostly unknown. In such 1995). As these authors write, for practical pur-
cases we have to use so-called biological diversity poses ‘biodiversity’ can be considered synony-
measures, which are based on taxonomic relations mous with Lovejoy’s ‘biological diversity’. (The
or similar differences of species only. Here abun- history of both expressions is summarized in an
excellent book edited by Hawksworth, 1995). In
the recent past, biodiversity measures other than
* Corresponding author. Tel.: +36-94-313892; fax: + 36-
that based on the number of species present (spe-
E-mail address: (J. Izsák). cies richness) have been introduced. We should

0304-3800/00/$ - see front matter © 2000 Elsevier Science B.V. All rights reserved.
PII: S 0 3 0 4 - 3 8 0 0 ( 0 0 ) 0 0 2 0 3 - 9

Fixing the number of Naturally. Solow et D. 1/(s + 1). s) denotes the (theoretical) probability of an individual belonging to the ith species. measures.. 5 L(1/(s + 1). indices do not depend on taxonomic relations pertaining to the set of species to be preserved.…. The new ecosystem diversity is defined as the let us mention some specific properties of the sum of species diversity and structural diversity. but Dœ % p 2i ignores species abundances.…. For example. Weitzman (1992) pro. relatively homogeneous Rousseau. Izsák.) Some authors take only the hetero- 1979 and Magurran. 1/(s +1)) . (It would be more tion of individuals among the species. This is because − % pi log pi these indices are monotone increasing functions of i=1 the number of species s: 1where s stands for the number of species and pi L(1/s. To make a link between these diversity concepts. Note that the traditional diversity al. On a similar basis. diversity index is perhaps the Shannon – Wiener The value of a traditional heterogeneity index L index with the formula with equal probabilities in the argument can be s considered a richness measure. respectively. Krajewski biodiversity measures (1994). portant role in the future. 1/s) (i = 1. By definition. cies. index. which takes account of the s species set and distances between species.152 J. it correlates with the expected diversity loss by eliminating a subset of species. Another means of ‘biological diversity’ or ‘biodiversity’. However. who compared some index properties on crane species data. 1975. so-called ‘hetero- taxa depends intuitively both on the number of geneity indices’ reflect both the evenness and rich- occurring species and the evenness of the distribu. Diversity indices will continue to play an im- sity measures. because the heterogeneity indices are diversity. Namely.…. are in use in ecology and in other disciplines sity. species diversity. species diversity) have been introduced not simple functions of richness and evenness in the decades past (for a review see Grassle et al. Numerous accurate to speak of evenness and richness ‘di- measures of ecological diversity (or community mensions’. moths. they are close in 2. A gap between diversity indices and spirit to the index proposed below. Jizhong and Shijun (1991) the fact that one can take the species richness elaborated a new diversity index based on species (number of species) for a species diversity index. Numerous measures of evenness Turning to traditional indices of species diver. ecological diversity indices. 1988). Species distances are i=1 used to define distances between a simple species 2these indices can be formulated as 1/D and 1 − and a set of species. Introducing the quantity posed a measure. For example. Although not a diversity measure. 1991). for concrete applications one should be species. between species. provides a review of biodiver. amplifying the concept of diversity is to consider ecological diversity indices are scarcely mentioned the diversity of functional connections between in the literature of biodiversity. ness components of diversity. Magurran. 1988. abundance and functional relations between spe. Egghe and vascular plants and other. notwithstanding species. Papp / Ecological Modelling 130 (2000) 151–156 like to emphasize that each of these measures Other widely used diversity indices include the implicates a somewhat different definition of bio. L. The most popular geneity indices for genuine diversity indices. 1/s. (Pielou. (1993) introduced a ‘preservation measure’. we recall that the diversity of birds. reciprocal Simpson index and the Gini–Simpson diversity. one arrives at the evenness component of informed about figures on numerous relations. a gap re- All these biodiversity indices are based on new mains between the indices of the old concept and conceptions of species diversity and are not linked the indices connected with the newer notions of to traditional diversity indices. As both of these newer measures disregard species abundances.

if the abundances are equal. (f). in Fig. since they are monotone increasing s s functions of the number of species. relating to a drosophilid assemblage. randomly chosen individuals. herbivorous cludes that none of them are based on the previ. 1995). Let dij cies. where dij = 1 for all i "j and cation of the traditional diversity indices gives no dii = 0. tions. alone does. commensalist (c). dsf. the latter on ‘fea. the latter and neglect the abundance conditions. the abundance conditions and the abundances and insensitive to species differences. J. For example. (h) and of unknown habits (u). it is easy to see that Q is equal to the more information than the number of species Gini–Simpson index 1− D. Conversely. dsu = 4. One can conclude that there is a gap be. biodiversity measures are sensitive to This one-sidedness is explainable by three condi. tween a1 and a4 is 3. That is. For example. register only the presence of the species. rather special case. evenness component are irrelevant with these. and not of today. L. biodiversity measures vation problems. mycophagous (m). (1991) defined the ‘distinctive. focusing on the maintenance and conser. the comparison of We propose here a possible link between the abundances is largely meaningless between sys. between species based on some property. that we introduced long ago and used intensively also are speaking here of differences only. take the following types of resources: sity measures. the diversity values ‘will rank the communities Q expresses the expectancy of the difference be- according to the s number of the occurring spe. On the other hand. tween two. realm of diversity indices and measures of biodi- tematically remote organisms (ants and ele. exists. saprophagous (s). tional diversity indices are sensitive to species rent senses. . then the appli. Izsák. form ness indices. by these conditions the traditional hetero. Often the only valu- able data is the number of species. j =1.s + indices. With a fixed number of species.u = 5\dsf. In these circumstances. In the cies’. Third. the distance axioms are in this mentioned properties of the traditional diversity case not fulfilled. we (not necessarily fulfilling the distance axioms). (i. the abundances (within certain limits) are irrelevant. one con. s) be the ‘differences’ between species Or. index (Rao. As noted with dij = dji. Indeed. The explanation follows from the above distances. com. versity. frugivorous Studying these biodiversity measures. 2. 3. and Faith (1992) introduced the ‘phyloge. The former is Another example can be the introduction of based on taxonomic distances. example. in what amounts to the same thing. the sophisticated A simple way to generate species differences biodiversity measures emerging in the recent past can be based on the following definition of the dij take into account the differences between the taxonomic distance of two species: let dij stand species and neglect the abundance conditions. the tradi. 1 the distance be- ness’. Qœ % % dij pi pj = p%Dp i = 1j = 1 paring the biodiversity of different study areas.…. For 1995). where a common ancestor of species i and j Vane-Wright et al. tween traditional diversity indices and biodiver. netic diversity’ of a taxonomic tree. in many cases the correct data on abundances are unknown. A tentative differ- ously mentioned ecological diversity indices ence matrix is schemed in Fig. We have previously reported on the eco- phants). sap feeder (sf). Note. For instance. diversity index Q (Izsák and Papp. or its estimate. First. for the number of internodes from the species Some of these measures relate to taxonomic trees level to the lowest level of the taxonomic tree containing the species present. Papp / Ecological Modelling 130 (2000) 151–156 153 With regard to ‘biological diversity’ in its cur. we can attribute logical application of the quadratic entropy the same abundance value to all occurring spe. and dii = 0. tentative differences (not necessarily distances) ture mismatches’ (for further details see Faith. A possible link between diversity indices and Second. The quadratic diversity above. 1982) is defined as the quadratic geneity indices can be considered as single rich.

Denoting a measure by I. J links differences is a more suitable measure of biodiver- ecological diversity indices and biodiversity sity than the mean of differences.) It is an important condition. s) stands for the difference question. 1. (We enu- merate here all pairs two-fold. . we introduce the measure then. a3.s + 1 (i= 1. L.….4 + d3. counterpart. Concerning index properties. Difference matrix for the different resource types. 2. as opposed We can conclude that the sum of the species to other ecological diversity indices’.4) = (1/8) · (1 + 2 + 3 + 2 + 3 + 3) = 14/8 B J (A) = (1/23) · 2(2+3 + 3) =2.154 J. For difference J is not an ideal biodiversity index.4+d2. 1.…. a4} and species a2 in Fig. measures. the quadratic entropy index will be a func- F: =s 2 · J= % dij tion of the number of species and their taxonomic i.2+d1. set monotonicity is a property generally required for biodiversity measures. the mean Fig. taking the set A œ{a1. relations between species influence J.s + 1\% dij=F(A)(x QA) i. This ensures that the index value will increase by adding a new species to a species set A. Papp / Ecological Modelling 130 (2000) 151–156 Fig. When the same hypothetical or formal abun. To abbreviations see the text. that ‘the of the elements in A and as + 1 œ x. Thus. That is. For example. j relations. change from this intensive measure to its extensive dance values are attributed to each abundance. j i i. The set monotonicity applies to F. being that for 1 any set A and element x J: % di. this is a technical where di. The index J does not satisfy this requirement.3+d1.3+d2. A taxonomic tree with fix levels. this property means that I(A@{x}) \I(A). F(A@{x})=% dij+% di. j tween two randomly chosen ‘species’. J(A @ {a2})=(1/24) · 2(d1. In this special case Q can be written in the form which is the sum of the differences of species. j s 2 i. a5 see the text. j expressing the mean taxonomic distance be. Izsák. For taxonomic distances of species a1.

a5} in Fig. J does Ecol... D. In: J. Chapman and Hall. diet or . 1980.. Estimation. W. While writing this paper. D.). Biodiversity.. measures and the traditional diversity indices. substance may ‘choose’ a species regardless of the Izsák. one can obtain a variety of biodiversity functions of the number of species. Biol. establishing different dij insofar as they are ‘not’ monotone increasing values.F. 24 – 43. pp. . unified approach. former are insensitive to abundance conditions. 1982. London. Changes in biological diversity. Theor.L. Chapman and Hall. Papp / Ecological Modelling 130 (2000) 151–156 155 Remarkably. Conclusion Numerous measures of biodiversity are in use. measures in the modern senses. The above counter-example on set monotonicity for J is also applicable in the Acknowledgements present case. London. J. Instead. 1991. Measurement and or a set of species. 213 – 224. one arrives at logical Diversity in Theory and Practice. a specific toxic and Estimation. References satisfying different advantageous statistical prop- Egghe.. 1. The Global 2000 Report to the Presi- types of diversity measures.L. L.. diversity indices with equal species abundances (see Section 2). 2... Phylogenetic pattern and the quantification of organismal biodiversity. neither J nor F is a richness index other properties. Croom Helm.E. Ecological Diversity and Its Measure- As demonstrated above. in general. M. Phylogenetic measures of biodiversity: a which is essential for biodiversity measures. Pop. Izsák contains more species than D. only exception is the quadratic entropy index Q. G. the first author was a holder of the Hun- garian Széchenyi Scholarship. regardless of abundance condi. 1991. (Ed. Thus. International the mean distance of species in the set in question. 59. J. Inf.). OTKA No. The assumption of random choice of a species Hawksworth. 1995. Transfer principles and a erties. pp. 1994. In: Hawksworth. Biol. 1. Har- the requirement of set monotonicity. A counter-example for F: take C = {a1. Environ. a gap remains between these classification of concentration measures. it satisfies dent. and T 16 892 for L. tween ecological diversity indices and biodiversity Jizhong. A. Biodiversity.P. The Faith. Izsák. 2 (The Technical Report). 1979. New not necessarily taxonomic distances. (Ed. or differences between species or other taxa. An index of ecosystem diversity. However. Moreover. Preface..L. 1 – 10. the dij differences are ment.. D. Stat. T.. Index J bridges the gap be. 1988.P. Penguin Books.). Faith. E. C. Shijun. The Sci. 4. Biol.. Harper. Chapman and which is a function both of abundances and taxo.E. 1992. 1995. L.R. The comparison and critique. 61. Lovejoy. Krajewski. Then. Pielou. Papp. Conservation evaluation and phylogenetic the latter do not take into account the distances diversity. (Ed. Measurement and Estimation. Wiley. Am.. favor. 5 – 12. For example.L. C.. Hawksworth. R. This contrasts with the properties of versity research. Hall. the more appropriate biodiversity measure. Ecol. Fairland. 21. Chap. 1995.L. 33 – 39. tions. Mod. Magurran. D. it will be possi- by the species number and the differences between ble to gain insight into their usefulness in biodi- the species. They Rao.P.). Taillie. Papp. Z. However. Biodiversity. D. 479 – 489. J. Diversity and dissimilarity coefficients: a may represent differences in life style.. nomic distances or differences. J.C. Applying the measure F and similar ably. 42. a2. that is taking equal abundances is in some Hawksworth. J. Rousseau. measures. London. Research supported by the Hungarian National F(C) = 10BF(D)= 12. Taking equal Grassle. L. latter’s abundance. Vol. notwithstanding that C Scientific Fund. 1995.. or distances York. London.O. G. 151 – 163. (Ed. 327 – 332. Co-operative House. F is mondsworth. Soc. Ch. Application of the quadratic entropy indices for diversity studies of drosophilid assemblages.. their values are determined simultaneously measures in concrete field studies. not satisfy the property of set monotonicity. D. Smith. Thus. Measurement cases reasonable. satisfying the distance axioms.. 45 – 58. Conserv. MD. a3} and D ={a1. 1975. 69. pp. In: extensive version F is also a link between the two Barney.. Conserv. Patil. Ecological Diversity. Eco- abundance values for the species. T 017 027 for J.

L.. J.I.. J. Polasky. On diversity. 235 – 254. 24. On the measurement What to protect: systematics and the agony of choice..J. 406. 68. R. 1991. Broadus.156 J. Econ. 55. Humphries. 1993. A. 60– Conserv. Izsák. .L. 363 – Vane-Wright. S... 1992. Williams. M. . Papp / Ecological Modelling 130 (2000) 151–156 Solow. Env.. J. of biological diversity. C. Biol. 107. Manag. Econ. P.. Weitzman. Q.M.