Communities are affected there by harsher conditions (e.g., strong wave energy and currents, and less suitable light and temperature regimes), which may ...
Marine Pollution Bulletin 50 (2005) 787–789 www.elsevier.com/locate/marpolbul
Correspondence Guidelines for the use of AMBI (AZTIÕs Marine Biotic Index) in the assessment of the benthic ecological quality 1. Introduction Recently, interest on benthic indicators has increased dramatically, with a long list of new indicators proposed (see Diaz et al., 2004, for a revision). One such indicator, the AMBI, was designed to establish the ecological quality of European coasts, analysing the response of softbottom communities to natural and man-induced changes in water and sediment quality (Borja et al., 2000, 2003a). The increasing use of this tool has led to a public debate about some of the doubts encountered in using such indices (Simboura, 2004; Borja et al., 2004b; Dauvin, 2005); and has caused us to receive a large number of e-mails trying to adapt AMBI to their own requirements. This has led us to develop the present guidelines, in order to make any AMBI results comparable.
2. AMBI basis and applications The AMBI offers a Ôpollution or disturbance classificationÕ of a particular site, representing the benthic community ÔhealthÕ (sensu Grall and Gle´marec, 1997). The AMBI is based upon previous ecological models, such as those of Gle´marec and Hily (1981) and Hily (1984). The theoretical basis is that of the ecological adaptative strategies of the r, k and T (McArthur and Wilson, 1967; Pianka, 1970; and Gray, 1979) and the ecological succession in stressed environments (Bellan, 1967; Pearson and Rosenberg, 1978; and Salen-Picard, 1983). Most of the concepts within the AMBI are based upon previous proposals, for example: (i) the species should be classified into five ecological groups (EG) (following several authors, such as Leppa¨koski, 1975; Gle´marec and Hily, 1981 and Grall and Gle´marec, 1997); and (ii) with a scale introduced, from 0 to 7, based upon Hily (1984), Hily et al. (1986) and Majeed (1987). However, the most novel contribution of the AMBI was the formula permitting the derivation of a series of continuous values (Borja et al., 2000). Hence, taking into account the final objective of the proposal, several thresholds in the scale of the AMBI were estab0025-326X/$ - see front matter 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.marpolbul.2005.04.040
lished; those were based upon the proportions amongst the five EG (see Fig. 2, in Borja et al., 2000). These thresholds are coincident with the Ôbenthic community healthÕ proposed by Grall and Gle´marec (1997) (see Table 1, in Borja et al., 2000), whose sources can be found in Reish (1959), Bellan (1967) and Pearson and Rosenberg (1976). Further, the AMBI has been applied in the assessment of the ÔEcological StatusÕ, under the European Water Framework Directive (see Borja et al., 2003b, 2004a,b). In this particular case, these authors recommend the use of AMBI only as a part of a set of measures and indices (a multimetric approach), such as diversity and richness, in order to minimise misclassification problems in the assessment of the ÔEcological StatusÕ (Borja et al., 2004a). The AMBI has been verified successfully in relation to a very large set of environmental impact sources, including drill cutting discharges, submarine outfalls, harbour and dyke construction, heavy metal inputs, eutrophication, engineering works, diffuse pollutant inputs, recovery in polluted systems under the impact of sewerage schemes, dredging processes, mud disposal, sand extraction, oil spills, fish farming, etc (Muxika et al., 2005). The geographical areas where it has been applied extend over the Atlantic Ocean, Baltic Sea, Mediterranean Sea, North Sea, and Norwegian Sea, all in Europe; similarly, also in Hong Kong, Uruguay and Brazil (Muxika et al., 2005; Muniz et al., in press).
3. Guidelines in using AMBI Although the AMBI is particularly useful in detecting temporal and spatial impact gradients, its robustness could be reduced when only a very low number of taxa (1–3) and/or individuals (10, remove insecta. • Remove juveniles, when the species are not identified. • Remove non-soft sediment taxa (e.g. Nudibranchia). • Remove epifaunal taxa (e.g. Bryozoa). • Remove planktonic taxa (e.g. Crangonidae). • Certain taxa should be grouped together (e.g. genus types a–d). • Never use high taxonomic levels (e.g. Bivalvia, Gastropoda), except those included in the species list (e.g. Nemertea, etc.). Normally, we have found that less than 10% (usually less than 5%) of the individuals per sample are not assigned. When the percentage of taxa that are not assigned is high (>20%), the results should be evaluated with care, because there may be subsequent problems in the interpretation. When the percentage of taxa not assigned is >50%, the AMBI should not be used. In order to avoid ambiguous results, it is preferable to calculate the AMBI values for each of the replicates,
then to derive the mean value. The absence of fauna in some of the replicates (following confirmation that it is not a sampling artefact) should be considered as a sign of high disturbance; consequently, this is much better accounted for, through the use of the mean value of AMBI from the replicates. Hence, all of the replicates should be included in the initial data file. The assignation of a taxa to one of the five EG, together with problems associated with taxonomy (synonyms, etc.), could lead to misclassification. Some authors attempt to adapt the EG to their own listings, in order to obtain better assessments. We do not support the idea of building specific EG lists for different geographical areas or different impact sources, as this can lead to an unmanageable tool use; it does not allow any comparison of the results between different areas, or impacts. The assignation of the species to EG requires some consensus between the scientific community (Borja et al., 2004b). As such, we attempt to improve and update continuously the species list, working together with benthic experts from the UK Environment Agency and others; similarly, responding to the suggestions of several authors, requesting changes in some species EG or including new species, when it is justified. Hence, it is necessary to use always the latest version of the species list. The different thresholds in the site pollution classification, or even in the ecological status, depend upon the thresholds established within the AMBI scale values. Changing the thresholds would alter the final classification (as in other methods). Probably, in some of the cases, the range of values applicable to different habitats (estuaries, coastal areas, etc.) should be changed. For example, in some estuaries it is very difficult to reach AMBI values
