- Top of page
- Materials and Methods
- Literature Cited
The need for standardized biodiversity assessment methods to enable biodiversity quality to be measured is outlined. A general approach to sampling is suggested. The results of the use of this sampling process are given for two case studies of different taxonomic groups. The data assessment is made easier by the use of a bespoke computer program. Examples of the program output are presented. The advantage of this standardized measurement of biodiversity compared to species lists and the use of indicator species are outlined in the case studies macrofungi and butterflies. It was shown that the biodiversity quality of sites can be compared by the use of a range of measured biodiversity indices in a way that allows sites to be compared through time or between sites/treatments. In one case (butterflies), data that have been collected systematically in a recording scheme have been analyzed retrospectively and yielded valuable information on changes in biodiversity quality. The uses of this method in establishing baselines in restoration ecology are discussed. The importance of restoration ecology in the conservation of biodiversity could be underlined by the use of the methods presented in this article.
- Top of page
- Materials and Methods
- Literature Cited
Biodiversity is defined in the 1992 Convention on Biological Diversity (CBD 1992) as:
… the variability among living organisms from all sources including, inter alia, terrestrial, marine and other aquatic ecosystems and the ecological complexes of which they are part; this includes diversity within species, between species and of ecosystems.
Like many other politically determined conventions the above paragraph contains a large number of compromises and assumptions. The definition seems to be determined by the apparent need to list species as a means of recording biodiversity and yet the most commonly used “biodiversity indices” are measures of evenness (Shannon–Wiener and Simpson’s) or dominance (Berger–Parker) rather than lists. The list approach does not tell us anything about the balance of individuals of each species, and the “indices” approach can produce the nonsense that a site with 10 species can have a higher “biodiversity index” than a site of 100 species. This is the basis of the comment by Gaston and Spicer (2004)“… it is clear that no single measure of biodiversity will be adequate. Indeed, given its great complexity, it would be foolish to believe that the variety of life in an area, however small or large that area might be, could be captured in a single number.”
The CBD convention also refers to the different scales of measurement of biodiversity from genes to ecosystems. Sensibly, most scientists have chosen to work at the species level (however they are defined; see Haeupler 1999, p. 186) as being the only basis on which work that does not entail impossibly large tasks (describing ecosystems in detail) or the reduction of study to populations of whole organisms so small that they are of dubious reality to the general case (genes). The latter approach also requires a significant investment in technology that would not be possible on a large-scale given current funding. Politically, species are also more familiar and acceptable as a basis for biodiversity estimation.
In this article the author takes the view that the definition of biodiversity needs amendment to allow for the effect of the experience of biodiversity and its values and to go beyond a list approach (see Haeupler 1999, for a devastating review of the numerous articles written about “biodiversity” without the discipline of an agreed definition). Similarly, there are limitations to the pragmatic approach to biodiversity in the use of indicators. Frequently they are chosen for “political” reasons and represent “good politics, poor science” (Akeroyd 1996).
The Convention on Biological Diversity Decision IV/1 and Recommendation III/5 (Handbook of the Convention on Biological Diversity) set the overall target in setting indicators that they should address matters such as:
the way indicators relate to management questions
the ability to show trends
the ability to distinguish between natural and human-induced change
the ability to provide reliable results (i.e., through the establishment of standard methodologies)
the degree to which indicators are amenable to straightforward interpretation
the question of baselines for measurement, in the light of the fact that application of a preindustrial baseline may often prove problematic.
This has been interpreted for an action plan at the Malahide Stakeholders Conference (European Environment Agency Conference, Malahide, 2004) as a list of biodiversity headline indicators as follows:
trends in extent of selected biomes, ecosystems, and habitats
trends in abundance and distribution of selected species
change in status of threatened and/or protected species
trends in genetic diversity of domestic animals, cultivated plants, and fish species of major socioeconomic importance
coverage of protected areas
area of forest, agricultural, fishery, and aquaculture ecosystems under sustainable management
numbers and costs of invasive species
impact of climate change on biodiversity
marine trophic index
connectivity/fragmentation of ecosystems
water quality in aquatic ecosystems
patents (to be developed)
funding to biodiversity.
Biodiversity in this article will be regarded as a quality of a site that can be extrapolated from a number of measured characteristics of the populations studied that are shown to be present at the time of sampling. The use of this approach will address the whole of, or elements of, the headline indicators given above, 1, 2, 3, 4, 7, and 9, and through these all that listed (i–vi) in the CBD. The role of ecological restoration will be underlined by being able to measure biodiversity quality of sites and show changes in quality over time. The starting point for this therefore is to devise a relatively simple standardized sampling method that will work for a wide range of taxonomic groups and yet allow the measurement of a number of shared biodiversity characteristics.
The need for this approach in restoration ecology results from the need to monitor and assess biodiversity consequences of interventions (or noninterventions) intended to restore sites. The characteristics measured by a standard sampling approach will allow observation of the progress of biodiversity restoration of a site and judgment as to the success of the intervention. In this way the role of ecological restoration in halting the decline of biodiversity by 2010 (European Environment Agency Conference, Malahide, 2004) can be clarified. In the same way different sites with different treatments can also be compared.
The need for some form of unification of biodiversity assessment is acknowledged in the draft Global Earth Observation System of Systems (GEOSS) 10-year Implementation Plan that refers to the aim to unify many disparate biodiversity-observing systems and create a platform to integrate biodiversity data with other types of information (GEOSS 10-year Implementation Plan Draft point 4.1.9). Burd et al. (1990), for example, also state that in the case of sampling methods, “researchers … can simply be overwhelmed by the variety of analytical methods available”; they were only referring to the methods used in marine benthic infauna. New (1998) points out that even the use of the term “sample” is open to at least four interpretations and needs to be clearly defined. There is therefore a need for clarity and definition in estimation and measuring biodiversity.
Some clarification was attempted by Greenwood (1996) where he discusses the factors in designing sampling protocols and lists the following: the need for sampling, replication, ensuring samples are representative, the size and number of sampling units, cluster sampling, multilevel sampling, stratified sampling, and surveillance. Ausden (1996) lists 21 methods for sampling 36 invertebrate groups, and each of these methods in turn presents its own problem of standardization down to the prosaic difference between size of boot used when kick sampling in streams and rivers.
Coddington et al. (1991) list seven points as important when deriving sampling protocols for biodiversity estimates of invertebrates in tropical ecosystems, namely:
Collecting methods used consistently by museum systematists for any groups should be incorporated into inventory procedures with as little modification as possible.
The number of collecting methods should be as few as possible, and divisions of microhabitats should be as simple as possible in order to minimize sampling complexity.
The sampling protocol should work adequately in plotless and plot-based sampling situations. For example, time spent sampling may be a good basis for measuring sampling intensity in many cases simply because it is easy to do it in the field.
The sampling unit should be large enough to yield adequate numbers (individuals and/or species) within samples but small enough that the number of separate samples can be large enough for valid statistical treatment.
Data should be collected in forms in which variations can be estimated and analyzed.
Data on numbers of species taken should be able to be combined to produce species abundance distributions, which can be used to estimate species richness.
If possible, the analysis should produce confidence intervals on the estimates produced.
All of this work underlines the need to produce a methodology that unifies sampling philosophy and yet is flexible enough to cater for the individual characteristics of the taxonomic grouping, which is the subject of biodiversity status assessment.
The research reported here using case study examples attempts to produce this unifying approach and specifically relates to two components; first, a sampling philosophy that can be adapted to various different taxonomic groups and then a consideration of the measurement of biodiversity characteristics for two very different groups of organisms, macrofungi and butterflies.
In a very clear, practical and precise summary of biodiversity assessment methods, Sutherland (1996) gives a list of 20 commonest sins in biodiversity sampling.
Following the guidance given by Sutherland above, a sampling methodology was developed for macrofungi by Feest (1999) because he considered that they probably represented a “worst-case” scenario for biodiversity measurement and the author proposed that if a method could work for macrofungi it should be adaptable to a wide range of other organisms. The fruit bodies of macrofungi (Agarics, Boletes, and Gasteromycetes) are ephemeral and discrete (but these in turn may, or may not, relate to clones of genotypes), are of very great difference in size, have a wide range of ecological roles, and have some difficult taxonomic problems. The background philosophy of the sampling was that it should be objective, repeatable, produce standardized population estimates, and also allow estimation of biodiversity quality.
Sampling should be representative of the population present in the ecosystem/site as a whole.
Sampling should be objective.
Sampling should consist of a number of small identically taken subsamples rather than a few large ones (using a few large-size samples implies that either the subsample sites are chosen or there is a strong probability of the sample sites not being representative).
All sites to be compared through time and between sites should be sampled in the same way.
The delimitation of the ecosystem/site to be sampled needs to be carefully determined before sampling.
Distances between subsamples need to be scaled to allow full site coverage.
Organisms that have definitive form can give better information than those that do not because they can be counted as individuals (frequency can be used for some organisms with a nondefinitive form, e.g., bryophytes, but there is a loss of information).
The data collected can be used to calculate a range of indices representing different characteristics of biodiversity of which the following have been used:
Species List. The total number of species recorded for the site, but this information should be accompanied by precise details of sampling input in the form of number of samples taken, who did the sampling and identification, and dates when the samples have been taken. Only in this way will any comparison with other sites be possible and observer bias will be noted.
Species Richness. The number of species recorded in the sampled area or per uniform number of observations. This biodiversity quality measure can be compared with others only if they use the same sampling methodology.
Biodiversity Index. Calculated from the aggregated numbers of individuals or frequency for the whole sample. Evenness can then be calculated using Shannon–Wiener or preferably Simpson's index. The latter is preferable as it has a wider amplitude and therefore greater discrimination. The simple to calculate Berger–Parker dominance index has shown to be equally as valuable in practice.
Density. The number of individuals of the taxonomic group being surveyed per unit area or per sample.
Species Value Index. Calculated as the mean and standard deviation of the arbitrarily assigned values for each species recorded. Normally, this is according to the recognized rarity of the species (Feest 1999), but it could be according to intrinsic value or ecosystem importance. Indicator species could be given high values in this scheme. Red Data Book status could also be used to devise the species value index (SVI).
Biomass. Inferred from the sampled taxa by the simple calculation from the biometrics of the species. Toth and Feest (2003) have shown that the dry weight of a macrofungus is proportional to the cap area of macrofungi. Obviously this is only possible for taxonomic groups with a determinate size.
The derivation of some of these indices (species richness, Shannon–Wiener, Simpson’s, and Berger–Parker) is provided by Magurran (1996).
Materials and Methods
- Top of page
- Materials and Methods
- Literature Cited
The following sampling procedure complies with the above principles and is illustrated in 1Figure 1:
Figure 1. Plan view of sampling methodology showing the route for taking 20 subsamples within the area to be surveyed.
Download figure to PowerPoint
Delimit the sampling area according to the needs that required the estimation of biodiversity quality (this normally relates to a common land management plan) and decide on subsampling distances such that a good cover will be achieved by 20 subsamples.
Following a compass direction, stop at uniform distances and take the subsamples (except in regular/ordered habitats where random numbers should be used to generate the distances).
If a boundary is reached, turn internally through 90° and continue.
Continue until 20 subsamples have been made.
Because sampling is “spot sampling,” global positioning system (GPS) locations are given where possible because this will aid the relocation of rare or interesting species if the need arises.
A further factor to consider was the date of sampling because some organisms can only be practically sampled in certain seasons. For these organisms, comparisons of biodiversity data should be at matching seasons/dates; therefore, dates of sampling were always included.
The subsamples can be pitfall traps for invertebrates, butterfly observations/counts, birdsong counts or suitable other samples. For macrofungi the sample consists of recording each fruit body of each species within a 4-m-radius circle around the central sample point. A 4-m-radius circle describes an area almost exactly 50 m2 and 20 of these samples will cover an area of approximately 1,000 m2; thus, total counts of fruit bodies can also be described as per square meter by dividing the overall count by 1000. In the UK Common Bird Census a similar methodology based on birdsong is used (Marchant 1994; Gibbons et al. 1996) where birds singing (or seen) at a series of stop points are recorded. For organisms with indeterminate form (e.g., bryophytes) the presence and absence data within the circle as described for macrofungi can be used for calculations based on frequency (Martin 2001).
In order to present the data clearly a simple computerized system and database (Fungib) for calculating the various measured characteristics of the taxonomic group was created to provide a biodiversity assessment and to speed these calculations where samples are species rich. Examples of the output of the computer assessment of the sample data are given and assessed.
The two macrofungal case study sites were East End Wood, which is an acidic mature woodland with Oak (Quercus sp.) and Chestnut (Castanea sativa) trees, and Weston Big Wood, which is a mature limestone woodland with Ash (Fraxinus excelsior), Lime (Tilia sp.), and Beech (Fagus sylvatica) trees. The butterflies for the case study were surveyed at a limestone grassland site in the Netherlands.
- Top of page
- Materials and Methods
- Literature Cited
A case study comparing two sites in the United Kingdom are given for macrofungi (using twenty 50-m2 plots) in 23Figures 2 and 3 and one other site in the Netherlands for butterflies (using the Pollard and Yates 1993 survey methodology) in 4Figure 4. The macrofungal indices compiled are compared in 1Table 1 and the butterfly indices in 2Table 2. Other features to note are the identification of each of the 20 subsamples at the top of the chart and the GPS location of these points in the next rows. The species recorded are listed on the left-hand side, numbers of individual fruit bodies etc. in the chart cells, the total number of individual fruit bodies etc. per subsample in the bottom row, and per whole sample on the right-hand side in the sum column with the overall total at the bottom of this column; the SVI for the recorded species is given in the SVI column and the mean value at the bottom of the column; the biomass index is indicated in the final column with the total at the bottom.
Figure 2. Survey results for macrofungi in East End Wood, Hampshire, United Kingdom (26 October 2001). The 20 subsample sites and their GPS location are given across the top of the figure. Species identification is given down the left-hand margin, numbers of individuals in the cells, totals per subsample along the bottom and per species in the third column from the right (and total of all fruit bodies at the bottom of this column), SVI in the second from the right column with the mean at the bottom of the column, and the biomass index in the right-hand column with the total at the bottom. A summary of the calculated results including the various indices is given in the left-hand corner. This is a printout of the Fungib program.
Download figure to PowerPoint
Figure 3. Survey results for macrofungi in Weston Big Wood, Somerset, United Kingdom (24 September 2000). The 20 subsample sites are given across the top of the figure (no GPS). Species identification is given down the left-hand margin, numbers of individuals in the cells, totals per subsample along the bottom and per species in the third column from the right (and total of all fruit bodies at the bottom of this column), SVI in the second from the right column with the mean at the bottom of the column, and the biomass index in the right-hand column with the total at the bottom. A summary of the calculated results including the various indices is given in the left-hand corner. This is a printout of the Fungib program.
Download figure to PowerPoint
Figure 4. Survey results of 20 fixed 50-m survey lengths of a limestone grassland site in the Netherlands. Species identification is given down the left-hand margin, numbers of individuals in the cells, totals per subsample along the bottom and per species in the third column from the right (and the overall total at the bottom of the column), SVI in the second from the right column with the mean at the bottom of the column, and the biomass index in the right-hand column with the total at the bottom. A summary of the calculated results including the various indices is given in the left-hand corner. This is a printout of the Fungib program.
Download figure to PowerPoint
Table 1. Biodiversity quality indices for two sites, East End Wood in Hampshire, United Kingdom, and Weston Big Wood in Somerset, United Kingdom.
|Site||East End Wood||Weston Big Wood|
|Shannon–Wiener index||3.257 (3.393)||2.573 (2.444)|
|Simpson's index||10.813 (18.618)||7.594 (6.640)|
|Berger–Parker index||0.273 (0.148)||0.243 (0.297)|
|Fruit body density||0.689/m2||0.74/m2|
|SVI||3.738 ± 1.900||4.667 ± 3.742|
|Cap area index||26,019 cm2||15,540 cm2|
Table 2. Butterfly biodiversity value indices for a limestone grassland site in the Netherlands sampled in 1994 and 2003.
|Density (population size total catch)||2,102||2,091|
|SVI and standard deviation||4.148 ± 3/689||3.786 ± 2.273|
|Biomass index (based on wingspan in centimeters)||9,228||9,983|
In practice, sampling for the fungi takes 2–3 hours depending on fruiting density with a variable time needed for identification later of the more critical species. Identifying butterflies on the wing is a skilled activity and dependent on weather conditions. The counts are made of the number of butterfly individuals of each species in a 50-m length along a transect of the site (Pollard & Yates 1993) aggregated for a whole calendar year. Figures 2 and 3 show clearly the difference between a site where the approximate number of macrofungal species present (species richness) have been recorded (East End Wood where no new species occur after plot 17) and a site where the biodiversity in the form of the number of species present has not been fully covered in 20 plots (Weston Big Wood where the slope indicated by the addition of new species shows no flattening off at plot 20 and this is indicated in Table 1 by the addition of the plus sign after the species richness index of 45).
In assessing the relative biodiversity quality of the two sites, the data are used to provide two sets of data based on either numbers of fruit bodies or relative biomass to give biodiversity indices (Shannon–Wiener, Simpson’s, and Berger–Parker indices) because macrofungal fruit bodies may either represent individuals or be parts of a clone. The great difference in size also means that basing a biodiversity index on individual fruit bodies might produce misleading results if a small fruit body is in fact a thousand times less in biomass than an individual much larger fruit body. The data from the biomass index have therefore been used to also calculate biodiversity indices in terms of relative biomass, and these figures are given in parentheses after the figure based on fruit body counts. For East End Wood, this has made a considerable difference in the values obtained due to the large numbers of one species (Hebeloma pumilum: Cortinariaceae), which is a small species.
In comparing the biodiversity of the two case study sites, it can be seen that:
East End Wood has a higher species richness but the Weston Big Wood has a species richness that has not been completely measured (Table 1
East End Wood has higher evenness biodiversity indices and lower dominance index.
The two sites have similar fruit body densities.
Weston Big Wood has a higher SVI (in this case rarity value), and this is shown by comparing the standard deviation values that are more susceptible to the presence of rare species especially when the species richness is high and dilutes the impact of a rare species on the mean value.
East End Wood has a much higher biomass of macrofungi.
In comparing the macrofungi of the two sites it might be said that the biodiversity quality of the East End Wood site is of a site with many species, evenness, and higher biomass whereas the Weston Big Wood site has a biodiversity quality identified by the presence of rare species as identified by the higher mean SVI and standard deviation.
The data in Figure 4 are for a sampling of butterflies using the Pollard and Yates (1993) methodology. It has been possible to use these data retrospectively because the methodology is well defined and repeatable. It can be seen that the biomass/numbers of butterflies present in 1994 is dominated by two species, and this results in low biodiversity indices (Shannon–Wiener, Simpson’s, and Berger–Parker indices), but the value of the site for biodiversity is also illustrated by the presence of two rare species and one very rare species resulting in a SVI and standard deviation of 4.148 ± 3.689. Notice that these rare species are confined to either one or two survey lengths.
Table 2 shows the results of the above survey compared to another taken in precisely the same way 10 years later.
- Top of page
- Materials and Methods
- Literature Cited
A simple methodology has been described that produces data indicating that it is possible to derive a series of indices that can be used to describe the biodiversity value of defined taxonomic groups. This methodology has been used by the author for over 5 years and for over 150 surveys of macrofungi, and this practical experience shows that only 20 subsamples are needed to give a good indication of the total biodiversity of a site. These indices have further use in allowing the biodiversity of different sites to be compared and that changes over time can also be recorded. These elements are essential if the conservation of biodiversity, the subject of several different international conventions, directives, and agreements, is to be achieved. This recording of indices is particularly important if the effect of site management on biodiversity is to be recorded such that recommended practices can be given for desired outcomes. An example of this is in the management of roadside verges in the United Kingdom. It is proposed to award regional road management contracts to private companies for a duration of 20 years in the first place. These contracts will be subject to performance targets and one of these is proposed to be that “there will be no net loss of biodiversity.” It is difficult to see how, given the evolution and progression of ecosystems in time, the above condition could be complied with over 20 years other than by the comparison of indices recorded in a standardized way.
In comparing this methodology with other possible methodologies, the usual sampling of fungi is by “wandering about” and almost never is a structured survey encountered (but see, e.g., a series of different studies almost all of which create their own new methodology starting with Haas 1932; Wilkins & Patrick 1939; Frei-Sulzer 1943; Richardson 1970; Arnolds 1976; Ohenoja 1983; and ending with Vogt et al. 1992). This lack of structured surveying produces a highly stochastic recording of species encountered that may or may not indicate the quality of the biodiversity of the macrofungal fruit bodies present (this is, for example, the basis of the British Mycological Society's fungal database). The sampling by wandering about does not allow any standardization of effort or information and also is subject to observer bias because there are a lot of fungal fruit bodies that come into the “small brown jobs” (SBJs) category that are easily missed except when thoroughly recording each fruit body in a defined area. These SBJs are often of interest and are, in practice, not so difficult to identify as may be expected. They can represent a significant part of the biodiversity. Hebeloma pumilum (Cortinariaceae) (see above) would be considered an SBJ by some mycologists.
On a few occasions, permanent large plots have been used for biodiversity assessment of macrofungi (Straatsma et al. 2001) and these have provided information on changes within plots, but it becomes theoretically problematic to use this information in other situations. Baar and Kuyper (1998) showed in a plot experiment that macrofungal species fruiting can be induced to recur by habitat manipulation, but the experiment was based on the occurrence of fruit bodies of rare species thought to have become extinct on a site. How much more information would have been possible in their plot experiment if they had taken the approach of this article?
Proof of the value of using a defined methodology of surveying is clearly demonstrated in the butterfly case study. The use of the Pollard and Yates (1993) methodology has allowed a retrospective calculation of the biodiversity indices and a comparison of the changes in biodiversity quality. Whereas the species richness, population, and biomass index have remained very similar, changes have occurred in the other indices. The species populations have become more even (Shannon–Wiener, Simpson’s, and Berger–Parker indices), and the rarity of the species has declined. This latter change has resulted from the loss of one very rare species (SVI 20), which has been replaced by one rare species (SVI 10). This information is obviously of great utility in considering the management of a site and is the information needed for the determination of the effectiveness of ecological restorations.
Martin (2001) used the same sampling philosophy but recorded presence and absence data with 50-cm2 circles to measure the biodiversity of bryophytes for a series of defined areas (compartments) of a single site.
One important improvement that is consequent on using a standardized derivation of measured biodiversity indices is that differences between sites can be measured statistically and the data presented here were subject to further statistical analysis such as recommended by Greenwood (1996). The two macrofungal survey sites reported here (Weston Big Wood and East End Wood) were compared by t test and F test (3Table 3).
Table 3. The t test and F test results showing differences in biodiversity quality between East End Wood and Weston Big Wood.
| ||t Test||F Test|
|Fruit body density||p= 0.353||p= 0.002**|
|SVI||p= 0.135||p≤ 0.001***|
|Biomass index||p= 0.710||p= 0.028*|
From this it can be seen that whereas the mean values do not differ significantly (t test), the variance of the mean value does (F test) in each case. There is a greater variance in the fruit body density and SVI for Weston Big Wood and a greater variance in the biomass index for East End Coppice. This adds further to the understanding of biodiversity quality of the two sites. The same statistical approach could be used to compare the changes in butterfly biodiversity quality illustrated in Table 2.
The basic idea of using the quality of the biodiversity of a taxonomic group to classify a site or establish a baseline dataset is not new. This is the basis of many biological water quality assessment processes and is well accepted. The RIVPACS method (Wright et al. 2000) even uses an “average score per taxon” test, which is the same as the SVI described here (Wright et al. 2000). Danovaro et al. (2004) used a methodology similar to the one reported here to assess the biodiversity quality of nematodes in sediments in deep waters of the Mediterranean. They used this information to follow changes in nematode biodiversity in response to the warming of waters following climate change. That a wholly different study can adopt such similar methods to provide similar information to that reported here indicates the validity of the general approach. The existence and acceptance of these methods indicates that the principles of methodology described here should find acceptance once promulgated.
Of more theoretical importance is the progression of understanding of the nature of biodiversity from a simple list approach, through an indicator species approach to an understanding of the biodiversity quality approach based on standardized indices. The list approach can sometimes produce apparently nonsense results. Tofts and Orton (1998) recorded the macrofungi of a Highland pine woodland in the United Kingdom over a period of 25 years. At the end of this time they were recording new species to the site list at the same rate as when they started. The conclusion from this might be that given infinite time there would be recorded an infinite number of macrofungal species present on the site; it might just be that the asymptote has not yet been reached, but when will it be approached? Further problems of the list approach are that once a species has been recorded on a list, there does not seem to be a mechanism for removing it if it does not occur again, so lists just grow and grow. The real nature of the list approach to biodiversity is that the list is an historical record but is not able to predict future occurrence (Feest 2000). This is a severe disadvantage because it is the expected occurrence of a species on a site that is the rationale for consulting a list.
The indicator approach is also receiving considerable support and is used extensively in conventions to monitor and protect biodiversity (e.g., the European Environment Agency Conference, Malahide, 2004 lists 14 headline indicators but analysis of these shows that they in turn require a species-based approach). The rationale for the use of indicators is that the species recorded are part of a biocenosis that includes a range of other species. What is frequently lacking in this approach is any attempt to validate this assumption in terms of probability or reliability. Very often the species are high–trophic level predatory species that are easy to observe and have considerable intrinsic value to the public and politicians. These species frequently have the ability to switch prey (indeed they often have to do this to survive) and are more reliant on the structure of a habitat than on its component species.
Using the approach described in this article should avoid the disadvantages of the list approach and the indicator (species) approach. Fortunately, use of the biodiversity quality approach based on indices does not preclude the use of other methods because the appropriate information will still be collected and the author suggests that these three approaches should be integrated into any biodiversity quality assessment and in doing so each is reinforced by the other two. The use of the methods described here can form the basis of such an approach, but there needs to be an international agreement on methods before this can be a reality.
The advantage of this approach for restoration ecology is that reliable information is generated that can show the progress of a restoration and demonstrate its value in the conservation of biodiversity.