Commentary: Do we have a consistent terminology for species diversity? Back to basics and toward a unifying framework
- First Online:
- Cite this article as:
- Moreno, C.E. & Rodríguez, P. Oecologia (2011) 167: 889. doi:10.1007/s00442-011-2125-7
- 403 Views
After decades of misusing the term diversity in community ecology, over the last 5 years some papers have offered important advances toward developing a more rigorous mathematical background, which allows us to achieve more clarity in the terminology for the vast range of biological phenomena that have been placed under the umbrella of this term. Some points have been clearly stated in previous papers of this Views and Comments section, and new terms have even been proposed for specific cases, but other issues, such as the need for the prefix true have not been discussed. Our aim is to clarify some of the terms and concepts, the proper use of which appears still to remain unclear, and to provide biologists with a simplified version of the general framework resulting from recent contributions, with an emphasis on identifying points of consensus in the field. We also comment on the possibility of extending the basics of this general framework to other facets of the broad term biodiversity, such as functional or phylogenetic diversity.
KeywordsAlpha diversityBeta diversityGamma diversitySpecies richnessSpecies turnover
Recent contributions have renewed interest in species diversity and have moved this line of research to the forefront in high-impact ecological journals. In Oecologia, we (Moreno and Rodríguez 2010; hereafter referred to as MR) opened a series of Views and Comments by challenging some of the points in the proposal made by Jurasinski et al. (2009; hereafter referred to as JRB) about the proper terminology for quantifying species diversity in a consistent manner, i.e. that would stand the test of time. Tuomisto (2010a; hereafter referred to as HT) then built upon our arguments, agreeing with us in that (1) alpha and gamma diversities are properties of two different levels of biological organization, and thus each must have a particular framework for its analysis, (2) we should differentiate the terms of species diversity analysis based on presence–absence data (species richness) from those of species diversity analysis that are based on richness plus abundance data, and (3) we must clearly distinguish between terms, concepts, available methods for a mathematical analysis, and the biological processes they relate to, as stated by Hill (1973): “… Diversities are mere numbers and should be distinguished from the theories which they support”.
HT expands on these issues and describes a logical terminology for diversity by using analogies to explain the concepts previously described by other authors, including herself (Tuomisto 2010b, c). She also proposes new, quite specific terms to separately address each variant that a diversity concept may have, and in doing so, partially modifies the terminology she had proposed in previous publications (see below). In general, we agree with HT’s conclusion that a terminology for phenomena related to species diversity is already available. Here, we aim to contribute to the discussion by clarifying the meaning of the terms, and help to construct a simple, unified version of a general framework that synthesizes the essentials of species diversity concepts. We hope this leads to the framework being easily understood and efficiently applied in ecology, biogeography and conservation biology. We will also comment on the difference between species diversity and the broad concept of biodiversity, highlighting how recent proposals extend this framework to other facets such as functional or phylogenetic diversity.
We will not address here the great confusion around terms, or the lack of clarity in the concepts that has prevailed for decades about alpha and beta diversities. This confusion triggered the analysis presented by JRB, and at that time MR appropriately concluded their comment by saying that “A better approach is needed … to systematize the identification of the many facets of species diversity”. Fortunately, almost simultaneously, other papers appeared to shed light on ways to identify the different response variables that are included under the term diversity. One example is the valuable work of Tuomisto (2010b, c) in which she took on the huge task of compiling, organizing and explaining everything that has previously been referred to as “beta diversity”. Though one may not agree with how she analyzed and summarized the different variants of beta diversity and their related terms, there is no doubt that this titanic effort—with the painstaking attention to detail that was applied—was absolutely necessary. Thanks to these and other relatively recent contributions (e.g., Jost 2007; HT), further steps have been taken and we can continue moving toward a theory of species diversity based on a more solid foundation. There is, however, the risk of misinterpretation because these meticulous and indispensable analyses—which emphasize the differences among the variants—may prevent us from seeing the unifying characteristics of the general framework. When we look so closely at the tree, we might lose sight of the forest, and those who read HT might be overwhelmed by the detail. This is why we will now go back to the basic biological definitions that are well supported by recent mathematical advances.
What is and what is not alpha diversity?
From the point of view of community ecology, species diversity (not biodiversity) is essentially related to the structure of communities, the reciprocal of mean proportional abundances (Hill 1973). If all the species in a community have exactly the same abundance (i.e. they are equally common, or equivalent in number), then diversity is maximized and proportional to the number of species (species richness). This is what Jost (2006) called true diversity, a term proposed to refer only to those quantities that measure this concept and behave in accordance with some properties that are intuitively assumed in the concept of diversity as it is used by ecologists (though they may not always be aware of doing so), such as the replication principle (Jost 2007, 2010). Also, these quantities meet the mathematical requirements for framing his proposal (Jost 2006, 2007) and match standard rules of inference; one of the central concerns in ecology and conservation biology (Jost 2009). These are the reasons the term true diversity was proposed.
As HT explains in detail, the measures of true diversity may consider to a greater or lesser extent the proportional abundances of species, thus allowing us to select them according to the biological question we have in mind. On the one hand, diversity of order zero (0D) is only the number of species (S) because it does not take proportional abundances into account at all. Thus, S is a measure of the species richness concept and should be referred to as such. This distinction moved HT to propose new terms to precisely distinguish the notation used in her previous papers (Tuomisto 2010b, c). On the other hand, diversity of orders greater than zero (e.g., 1D, 2D, with the q parameter being any non-negative number, not just an integer) do include, in addition to species richness, proportional abundances. Thus, under this framework, 0D is at the extreme of a continuum of diversity measures that are expressed in units that have been called “equally common species” (MacArthur 1965), “effective number of species” (Hill 1973; Jost 2006; Tuomisto 2010b, c) or “equivalent number of species” (Jost 2006, 2007). As Ellison (2010), the editor of a recently published forum, states: “if the interest is in describing the diversity of a single assemblage, the numbers equivalent, not the entropy, should be the diversity measure of choice”. We agree with this conclusion and, in the interests of establishing a consistent terminology, we propose using only the word effective to refer to these measurement units of diversity (e.g., effective number of species, effective number of communities) given that it is beginning to be used in current literature. Further, the term effective is also used with the same meaning it has in physics, economics and other sciences. For example, in political sciences, when comparing electoral systems across countries, we can use the effective number of electoral parties by counting the parties and weighting this number by the relative number of votes they receive, and in genetics, the effective population number has long been used as the number of an idealized population in which each individual has an equal expectation of progeny.
Another point that we would like to highlight in this process of constructing a unifying theory is that ambiguity in the terminology must be strictly avoided, and the concepts must be named according to their meaning. For example, other measures that are undoubtedly related to the concept of diversity, but that do not comply with the properties mentioned above for this concept, such as the Shannon index of entropy, the probability expressed by the Simpson index, or variance-based methods and indices, should be referred to using proper adjectives such as entropy-based indices, probability-based indices or variance-based indices. Of course, these and many other measures are extremely useful in ecology, but they should be referred to based on their meaning rather than by the umbrella term diversity. Plurality is obviously welcome in the measures to analyze different ecological issues, but clarity and consistency in the use of the terms is essential.
Some thoughts on the term true
We have explained why Jost (2006) coined the term true diversity, which helped HT to explain her ideas and to avoid the conceptual confusions and misunderstandings previously discussed in this series of Views and Comments. The semantics of the term true should not be the focus of our attention, because even when it is used to refer to a real and effective concept, in accordance with the mathematical principles assumed, it does not in any way imply monism. This term does not deny the existence of any other parameter, but expresses the need for clear terminology. As Jost (2009) concludes “It is not very important what name we give to the core concept of … diversity…We could call it canonical diversity or mathematical diversity or neutral diversity or anything else”. We believe that, eventually, we will be able to eliminate the adjective true and use only the term diversity to refer to a mathematically well-defined concept. But until this happens, it will be necessary to use the word true to make clear which terms are being used in ecology, conservation biology and genetics.
What is and what is not beta diversity?
Beta diversity refers to the differences in species composition between two or more communities, and this between-group component was rigorously partitioned from gamma or regional diversity by Jost (2006, 2007) to quantify the total amount of compositional differentiation in a region. True beta diversity is the effective number of distinct communities in the region (sensu Jost 2007), a concept that was later referred to by Tuomisto (2010b) as the number of compositional units, but see also the difference between “compositional units”, (CU), and “effective compositional units” (CUE) proposed by HT.
Many other biological phenomena of great importance in biology differ from this main concept of beta diversity, and should be distinguished with precise terms. An extensive compilation of them may be found in the papers of Tuomisto (2010b, c). Surprisingly however, the concepts of turnover, compositional similarity, overlap, and others have all been found (Jost 2007) to be mathematically related to and derivable from true beta diversity, encompassing a larger framework that unifies, relates and generalizes seemingly disconnected quantities that are naturally related to beta diversity (although other measures, such as the widely used Bray–Curtis index, are not directly related to true beta). These contributions provide the rigorous mathematical background for the terms and quantities used to assess beta diversity. From another point of view, more recently, Anderson et al. (2011) call for a primary distinction between the concepts of turnover (as a directional change in community structure along a spatial, temporal or environmental gradient) and non-directional variation in community structure. As previously noted (Vellend 2001; MR; Tuomisto 2010b), turnover is a biological phenomenon that is clearly different from beta diversity. Species turnover has its own place in ecological and biogeographical studies. However, even though “beta diversity” is included in the title of their article, Anderson et al. (2011) fail to recognize the value of this term and they prefer to use “variation”, a very broad and rather ambiguous term. In accordance with MR and HT, they stress the importance of distinguishing between binary (based on presence–absence data) and quantitative measures of differences (abundance data). They focus on classifying the different sampling designs that biologists employ in studies related to beta diversity, which facilitates understanding their different variants. We agree with Anderson et al. (2011) on recognizing the plethora of important biological phenomena housed under the term “beta diversity”, all of which deserve to be studied in depth. However, it is necessary to differentiate between them in order to avoid confusion. Unfortunately, if this is not done correctly, their recommendation of applying more than one measure of differentiation to reveal the underlying ecological processes that drive patterns in beta diversity runs the risk of promoting the uninformed use and abuse of the vast menu of available methods (many of which are not rigorously connected to the qualities they supposedly measure!). As HT says “Making an informed choice of index for a particular study necessitates an understanding of which aspect of the data the indices quantify, and which of them corresponds to what is needed to answer the ecological questions of interest.”
Why is this general framework advisable?
An advantage of expressing alpha diversity in units of effective species numbers is that this quantity allow us to easily compare the magnitude of the differences in diversity between two or more communities, regardless of whether there are any statistical differences between them (Jost 2006). With regard to partitioning gamma diversity into its alpha and beta components, Ellison (2010) expresses the agreement reached by different authors that “… numbers equivalents instead of the classical diversity indices … should be used in any diversity partitioning.” Furthermore, this framework may be extended toward other aspects encompassed by the broad term “biodiversity”. In their conclusion, MR said “there are other facets of biodiversity that have yet to be included in the discussion, such as taxonomic, phylogenetic, and functional diversity. These could be partitioned into components similar to those explained…. For example, we could measure alpha, beta, gamma, or turnover considering the phylogenetic relatedness among species.” The framework described here is species neutral, but just recently, Chao et al. (2010) have derived a general class of diversity measures to take into account both species abundances and species evolutionary relatedness (taxonomic or phylogenetic differences), and these measures possess all of the mathematical properties mentioned above. For example, they adapt and refine the replication principle for including evolutionary information, which is a huge step because previous measures of phylogenetic diversity may not meet the theoretical expectations of biologists. Using this new proposal, we can assess the effective number of maximally distinct lineages over an interval of time, and this provides a robust non-neutral measure of biodiversity. We hope we soon have a valid and more complete framework that encompasses fascinating and varied facets of biodiversity.
CEM thanks colleagues and students for spending time in fruitful discussions about species diversity, and Bianca Delfosse for improving the English. This contribution is a result of projects 95828 FOMIX CONACYT–HIDALGO and 84127 SEP–CONACYT Basic Science. We declare that our study complies with the current laws of Mexico, and that no experiments were performed.