Which biodiversity index to use

However, in practice, the difference between these two approaches is usually so small that the simpler formulas are generally acceptable Magurran Belowground insect larvae were not included in the path models because they were sampled on fewer sites than other groups, and their inclusion would have reduced the sample size below acceptable limits given the complexity of our model.

We ran the same structural model with each of the diversity indices, and we report model fit as chi-square and its associated P -value, with P -values greater than 0.

All analyses were performed with R v2. Structural equation models of links between diversity of organisms or traits measured in and around Plantago lanceolata. Solid lines indicate positive effects, while dashed lines indicate negative effects.

The magnitude of the path coefficient is indicated by line thickness. We then used principal component analysis PCA of correlation matrices to determine which measures of diversity were best able to differentiate sites by calculating importance values IV for each index Wilsey et al.

The IVs synthesize information on the importance of each principal component axis and the score for each diversity index to generate one number representing the overall importance of each diversity index in distinguishing plots based on distances between plots in the ordination.

Indices detecting the greatest number of significant effects were judged to be the most effective, although if land use did not affect diversity in our system these indices would actually be the least effective. Estimates of aboveground arthropod, chemical, and two loci of molecular richness overlapped our observed values, suggesting that these observations are robust. However, the generally excellent fit of most models suggests that all adequately represent the data.

The different number of significant paths in each model therefore highlights the different information emphasized by each metric. The only consistently significant path in all models excluding E was a negative effect of herbaceous plant diversity on chemical diversity of P. No relationships appeared to be driven by changes in abundant species or traits. The path analysis also showed that E represents different information than that captured by the other diversity indices Fig.

E did not correlate in a consistent manner with any other index of diversity. For example, E was positively correlated with S for Plantago chemical diversity, but negatively correlated with S for aboveground arthropod diversity.

Only in the belowground insect larvae did E not correlate with any other diversity index Supporting Information, Table S1. This unpredictability of E demonstrates that it carries information not included in the other measures of diversity and argues for including E when analyzing multiple diversity indices. We used PCA to visually represent these correlations and determine if any metrics were better at differentiating plots, despite the strong correlations between all metrics Fig.

Our results agreed with those of Wilsey et al. Principal component analysis of diversity measures taken in and around Plantago lanceolata , and land use intensity LUI. Numbers indicate importance values for each vector.

Land use intensity generally did not affect diversity, with no evidence for stimulation or suppression of AMF, aboveground arthropod, belowground insect larvae, or P. Furthermore, the magnitude of this effect was similar for all six metrics. We also tested which diversity indices provided the greatest ability to discriminate sites, and whether or not the effect of land use on diversity depended on the diversity index chosen.

The failure of our path model to fit the E data suggests that interactions between diversities in our system are driven primarily by differences in abundance, and not by changes in evenness. This shows that as the abundance of the most abundant plant species increases, the abundance of the most abundant chemical metabolite declines. The presence of a significant path from plant to chemical diversity for all indices except E suggests that changes in both rare and abundant metabolites are negatively affected by changes in rare and abundant plants.

Abundance of rare to moderately rare arthropods was positively affected by abundance of rare to moderately rare plants. This may be due to increased niche availability for specialist insect species as plant diversity increased.

This persistent negative relationship between plant and chemical diversity could be explained by likely reductions in P. In sites with low plant diversity, intraspecific P. Any decreases in P. These further hypothesized interactions between P.

As in other studies Wilsey et al. The failure of E to effectively discriminate sites shows that the synthesis of richness and abundance information is necessary for site discrimination and that the individual components of the compound diversity measures S and E are much less informative when considered independently.

A further strength of compound diversity measures over species richness is their reduced dependence on sampling effort Magurran For these groups, compound diversity indices are expected to be more robust than S, although they are still influenced to some extent by sample size Magurran In contrast, we may have overestimated chemical diversity, as number of peaks is probably higher than the number of real metabolites due to fragments, adducts, and isotopes that may occur in the metabolic fingerprinting approach.

On the other hand, metabolic fingerprinting was only done of polar metabolites, so the overall metabolite number in each sample is in total again higher. We found no effect of land use on diversities of AMF, aboveground arthropods, belowground insect larvae, or P. Aboveground arthropods were identified to order, while belowground insect larvae were identified to family. It is not entirely clear what level of resolution is achieved with the NSAM1 primers used in the AMF analysis, but it is almost certainly higher than species level.

Any effects of LUI may only be apparent at finer taxonomic scales. In this analysis, we focused on species associated with P. At least in this system, molecular and chemical diversity were less sensitive to land use than herbaceous plant species.

The differing sensitivities of diversity indices to LUI in our analysis were largely driven by our need to correct for multiple comparisons. In analyses using only one diversity index, similar significant effects of LUI would have been detected using any of the indices we included, except E. Thus, when conducting simple statistical analyses of a specific effect of disturbance on diversity, the choice of index does not appear particularly important.

We included two trait-based measures of diversity in order to assess their performance relative to species diversity. Chemical and plant diversity except evenness were consistently negatively correlated, but there were no other correlations between chemical and molecular diversity and diversity of any other organism.

This shows that trait-based diversity measures can capture unique information and can be useful when considered along with species diversity. For example, based on the significant relationships identified in our path analysis between chemical, plant, and arthropod diversity, we were able to formulate new hypotheses about regulation of defense induction in our system that would not have been apparent without chemical diversity in the model.

This suggests that, at least very generally, relationships between richness and evenness of these traits are similar to those seen in species diversity. Also, estimates of total richness were much closer to observed richness values for molecular and chemical diversity than for most organism groups. This suggests that it is easier to thoroughly sample at least some traits than it is to sample species.

Given the dependence of many indices on sampling effort, this is a clear benefit of trait-based diversity measures.

Overall, the trait-based measures of diversity performed very well and potentially have a place in other biodiversity studies as thorough measures of richness that capture information largely missed by organismal species diversity. At the very least, analyses such as this should precede selection of such indicator taxa to ensure that nonindicator taxa are in fact behaving as expected.

Recall that a taxon is any of our levels of classification. One would presume that more species equals more diversity. However, comparing two areas of equal species richness may show that they are not equally diverse. For example, lets consider a list of tree species in two forest ecosystems:.

Community A Community B. Water Oak Water Oak. Post Oak Post Oak. Blackjack Oak Hickory. Live Oak Pine. Bur Oak Cedar Elm. Pin Oak Pecan. Hickory Black Walnut. Although each community has seven different species, in A all are within the same genus and thus the same family, order, class, and division , whereas in B we have representatives of six genera, four families, two orders, two classes, and two divisions.

Clearly B is more diverse. Richness tends to increase over area. In other words, a larger area will harbor more different species, probably because of a larger variety of microhabitats and resources. Additionally, sampling over a larger area increases the chance of finding rare species. So, how large an area or how many samples is necessary in order to have collected all the species present?

Several mathematical methods are used to determine this. All are based upon the collector's curve. The negative exponential distribution is not often found in nature; it describes a fairly even distribution of individuals over species Fig. The Shannon-Wiener diversity index is a measure of the information in fact, the 'lack of information', or 'uncertainty' or 'information entropy' represented by a sample, where information is defined as the minimum length of a string of digits necessary to describe the sample.

The minimum length of a string of binary digits to describe a number is proportional to the logarithm of this number. If all the individuals belong to the same species there are no distinct equivalent permutations, i.

The conversion of the species-abundance distribution Eq. Log in. Page Discussion. Read View source View history. Jump to: navigation , search. Article reviewed by. Job Dronkers See the discussion page. An introduction to numerical classification. London: Academic Express. Evolution of species diversity in land communities. Evolutionary Biol. The mathematical theory of communication.

Ecological diversity. New York: Wiley Interscience. An introduction to mathematical ecology. New York: Wiley. Measurement of diversity. Nature The measurement of diversity in different types of biological collection. Diversity and evenness: a unifying notation and its consequences.

Practical measures of marine biodiversity based on relatedness of species. Functional diversity FD , species richness and community composition. Both samples have the same richness 3 species and the same total number of individuals However, the first sample has more evenness than the second.

This is because the total number of individuals in the sample is quite evenly distributed between the three species.

In the second sample, most of the individuals are buttercups, with only a few daisies and dandelions present. Sample 2 is therefore considered to be less diverse than sample 1. A community dominated by one or two species is considered to be less diverse than one in which several different species have a similar abundance. As species richness and evenness increase, so diversity increases.

Simpson's Diversity Index is a measure of diversity which takes into account both richness and evenness. Simpson's Diversity Indices. The term 'Simpson's Diversity Index' can actually refer to any one of 3 closely related indices. Simpson's Index D measures the probability that two individuals randomly selected from a sample will belong to the same species or some category other than species.

There are two versions of the formula for calculating D. Either is acceptable, but be consistent. With this index, 0 represents infinite diversity and 1, no diversity. That is, the bigger the value of D, the lower the diversity.

This is neither intuitive nor logical, so to get over this problem, D is often subtracted from 1 to give:. Simpson's Index of Diversity 1 - D.