Opinion on I. Technical Specifications for Classification and Presentation of Ecological Status of Surface Waters, wrc Ref CO 4658/1 (CSTEE/99/4A), II. Technical Specifications for Monitoring of Ecological Status of Surface Waters, NERI J. Nr 171/2-0079 (CSTEE/99/4A-Add.1), III. Proposed EU Water Framework Directive - Development of a Specification for the Intercalibration of Biological Monitoring Methods, WRc CO 4751/1 (CSTEE/99/4A-Add. 3). Note: this report is a second part of document II. It is not a separate study to be reviewed as such but rather to be seen as further specification on possibilities of ensuring comparability of monitoring results across regions and river basins. For this reason the mandate to the CSTEE only alludes to documents I and II above.

Preamble

The CSTEE has already adopted two previous opinions related to the Water Framework Directive. The Committee was in unanimous agreement when considering that the Directive represented an advance over previous legislation in using ecological effects as a basis of control rather than focussing on agents with a potential for adverse effects. The assessment of water quality is therefore in terms of the functioning and structure of ecological systems, rather than chemical contamination only. On the other hand, defining ecological quality, based upon both structure and functioning status of communities and ecosystems, represents a real challenge for the current state of science, and requires a great research input, as quoted in the last comment of the first CSTEE opinion on the WFD:

Finally, taking into account the complexity of the ecological and ecotoxicological problems faced by the Directive, it is the opinion of the Committee that a research programme specifically oriented to provide a scientific basis required for the fully implementation of this Directive must be activated on several key items. For example:

The CSTEE recognises the efforts of DG ENVIRONMENT, DG RESEARCH, and the Member States to provide an appropriate answer to the scientific questions related to the implementation of the WFD and is prepared to offer its views for this scientific and regulatory challenge.

CSTEE MANDATE (submitted by DG Environment of the European Commission)

Terms of reference for the consultation of the CSTEE on:

I) Report on a study on Technical Specifications for Classification and Presentation of Ecological Status of Surface Waters, WRc Ref CO 4658/1, and

II) Report on a study on Technical Specifications for Monitoring of Ecological Status of Surface Waters, NERI J. Nr 171/2-0079

The CSTEE is invited to comment on the two Reports which have both been drawn up in order to explore research needs in relation to classification and monitoring in support in particular of the implementation of Annex II and V of the proposed Water Framework Directive. The Committee's attention is drawn to the responses in the studies to comments and recommendations on classification and monitoring of aquatic ecosystems made by the CSTEE in its Opinion of the Water Framework Directive of 27 May 1998.

Furthermore, The Committee is invited to present any suggestions for research and development activities which could further explore the scientific questions and suggestions raised by the two reports, inter alia in terms of preparation of applications under the 5 ^th Framework Programme for Research. In this context, the attention of the Committee is drawn to the recent publication of a call for tender on research on aquatic ecology under that Programme.

Regarding research needs the opinion of the Committee is further requested on those identified by the two studies on particular aspects of classification and monitoring as well as suggestions to take these issues forward in order to support the practical implementation of the Water Framework Directive.

GENERAL COMMENTS

The general view of the CSTEE is that the consultants` reports provide neither a comprehensive review of the literature relevant to the subject nor a good basis for developing useful technical guidance. Among others, the reports on water-quality monitoring and assessment, which were published by the ministry of the Environment (The Hague, The Netherlands) at the request of Environment DG of the Commission in 1995 ¹ have not been considered in the submitted documents.

A significant part of the documents under review here is devoted to comment on the CSTEE opinions given earlier. But in our view most of the disagreements are related to misunderstandings of the CSTEE proposals, and/or to the focus of the consultants on current operational experience.

The CSTEE believes that the approach to the intercalibration exercise is fundamental to the success of the Water Framework Directive and needs to be more critically evaluated. The committee remains unconvinced that just harmonising the existing national systems (that focus on benthic invertebrates for rivers) will necessarily achieve the desired outcome.

The committee is particularly concerned by the weight given in the documents to the existence of operational experience in the Member States. First, because it does not include the experience of European research laboratories or even countries such as The Netherlands, with a biological monitoring network started on a routine basis in 1992 which covers phyto- and zooplankton; macro-invertebrates, water plants, fish, water birds, and river bank vegetation and fauna ² . Second, because the documents are biased to biotic indices based on the structure of riverine macrobenthic invertebrate communities, without considering other aspects such as the use of functional parameters e.g., primary production for the assessment of lake status. Third, because it is not clear to the CSTEE that trying to harmonise/intercalibrate tools developed several decades ago with a different aim (i.e., to assess the effects related to organic pollution, which was the key problem when these indices were developed) is the best option for the implementation of the Directive.

The CSTEE noted that the consultants' reports are now approximately two years old and that developments within European research laboratories and supported by DG Research mean that considerably more understanding and expertise is available throughout the Union to underpin effective implementation of the WFD.

While the CSTEE continues to support the general aspirations of the WFD it remains concerned about aspects of implementation, especially focussing on selection of reference conditions and the intercalibration process.

The terms of reference for this opinion are directly related to these three documents, and therefore, this opinion is also restricted to the information contained in these consultants' documents. The main conclusion of the CSTEE is that the documents are scientifically weak, have not considered several key publications on these issues, and do not present a realistic view of the current situation.

In the opinion of the CSTEE, the real possibilities for developing scientifically based tools for assessing ecological quality are wider and more achievable than presented in the reports. If DG Environment wanted further opinions on the CSTEE views for the implementation of the WFD, the committee would be pleased to provide it.

SPECIFIC COMMENTS

Technical Specifications for Classification and Presentation of Ecological Status of Surface Waters, WRc Ref CO 4658/1 (CSTEE/99/4A)

This report is apparently written in response to the General Comments in a previous CSTEE Opinion that "This" (the WFD) "should encourage more fundamental research into defining functional sustainability of ecosystems and of specifying the relationship between it and structural characteristics" - something that was prompted by the definition of ecological status under Article 2 and still in as paragraph 21 "an expression of the quality of the structure and functioning of aquatic ecosystems". Annex V, however, focuses on structural attributes and largely ignores functioning.

The main arguments of WRc for ignoring functional parameters are:

However, the CSTEE does not agree with this view and with the arguments expressed in the report. The reasons for considering functional measurements more complex and expensive in comparison with structural measurements are not acceptable. In many aquatic systems functional parameters are more frequently measured than structural ones and are simpler and not too expensive. For example, functional parameters such as primary production, nutrient cycling, etc., are commonly used as quality indicators for lakes.

Understanding the relationships between ecosystem structure and function is of central importance in effective implementation of the Water Framework Directive. In this context ecosystem structure should refer to the species composition of an ecosystem and is estimated by e.g., number of species, evenness or abundance of different species, or by various biodiversity indices. Historically environmental monitoring schemes have largely relied on structural measures, and there is an ongoing debate in the ecological literature on how ecosystem structure (and in particular biodiversity) relates to ecosystem function. Decisions about which structural and/or functional endpoints to employ in biological monitoring programmes require understanding of the relationships between them. This report seems to have missed this point entirely and instead focuses on how the physical and chemical properties of water bodies may be altered by anthropogenic activities and how such are related to functional features of the water bodies. The authors also seem to be confused about which attributes of ecosystems are measures of structure and which of function (see e.g., p. 17, section 3.2, para 2; p. 29, last para).

In order to more objectively define functional/structural attributes we need to ask: what services do we need from ecosystems (EQOs); what functional attributes are needed to maintain these; what structural attributes go with these? Only in this way shall we be able to objectively define quality criteria and move away from the somewhat ad hoc way of defining quality by reference to "reference sites". If this is accepted it defines a research programme - linking services to functioning, and the latter to structure - that is entirely missed by WRc. The WRc report also ignores/misses a lot of ecological literature on the relationship between structure and functioning; spp. diversity and process. Not all of this relates to aquatic ecosystems - but still is of importance in designing a research programme. In conclusion this seems a somewhat superficial analysis of the issues and not a very sound proposal for a research programme.

Regarding both operational experience and cost effectiveness, the CSTEE considers that according to the available information (see comments on the other reports), the Directive represents a significant movement from the current situation to a more scientifically justified basis for decision making and therefore the level of operational experience currently expected for the Member States is, in any case, expected to be low, and mostly related to pilot projects. Moreover, application of various nationally-developed tools, requiring harmonisation and intercalibration of the different methodologies, is not necessarily beneficial for the implementation of the Directive (see comments on the other reports).

The CSTEE would like to emphasise that both the processes occurring in ecosystems and the species performing them are important for assessing the ecological status of water bodies. Therefore the CSTEE considers that the inclusion of functional parameters in the definition of the ecological status is essential for a scientifically sound implementation, and that the level of development of functional parameters currently available, or easily achievable, in Europe is enough for their inclusion as available tools.

Technical Specifications for Monitoring of Ecological Status of Surface Waters, NERI J. nr 171/2-0079 (CSTEE/99/4A-Add1)

This report is concerned with issues for monitoring ecological quality in transitional and coastal waters and with research proposals for the development of "habitat quality" assessment criteria.

It is something of a " pot pourri" but is written clearly with a comprehensive listing of marine monitoring networks. There is a limited review (given as "examples") of classifications of transitional and coastal water quality; there are others that are not included. The bottom line - defining pristine conditions will be difficult for estuaries - is right. "Hindcasting" is suggested as an alternative but not developed; it should be. The section on habitat quality monitoring assessment is thorough and clear. The bottom line, that physical structure, biodiversity and ecosystem process are closely related, is well made, and the call for more research, teasing out causal relationships should be helpful. However, there is literature on this that should be quoted (e.g. Rivers Handbook ³ ).

Therefore, the report provides a rather thorough overview of current European marine monitoring programmes. An especially useful summary is provided in Table 2.1. However, as in the WRc report evaluated before there seems to be some confusion between the use of 'structure' to refer to the physical properties of water bodies rather than to the biological structure of the ecosystem.

Section 8 presents a general view on Direct Toxicity Assessment (DTA), including its use for effluent assessment which is only indirectly related to the implementation of the WFD. It refers to an "in progress" UK programme that is now complete and on which a report has been supplied to the CSTEE (CSTEE/98/16-Add.21). It also usefully refers to experiences on DTA over a large number of countries, worldwide. The potential of DTA as a regulatory tool for water quality control is clearly presented. However, the report does not present the actual position of the CSTEE proposal on the use of DTA. The CSTEE considers DTA be used mostly as an addition to chemical analysis for those waters where the ecological quality cannot be monitored (physically disturbed areas and groundwater). Several of the problems and disadvantages reported for the implementation of these tools, including those related to the use of concentration-fractionation procedures, are not relevant for these particular uses. Under these circumstances, toxicity must be considered as an additional parameter of the analytical part of the monitoring programme (which includes several physical-chemical parameters, quantification of the priority pollutants and other contaminants relevant for the area, and following the CSTEE recommendation, the toxicity of the water samples and several concentrates). The concentration methods required for the analysis of priority and other pollutants, can be selected to allow, simultaneously, the use of toxicological endpoints, and therefore, if the design is well defined, the only additional cost is for the toxicity tests, not for the concentration/fractionation procedures, and is considered as highly cost/effective. Moreover, the problems related to selectivity of the concentration methods and assessment of the total (not the available fraction) of the toxic chemical(s) present in the sample are considered by the CSTEE as advantages rather than obstacles. In fact, the aim of the DTA is to identify toxic chemicals which can be present in the sample at potentially dangerous levels but are not detected in the chemical analysis (Tarazona, 1998 ⁴ ) because:

Under these circumstances selectivity in the extraction can be helpful to determine the "type" of chemical responsible for the toxicity, and the assessment of "total levels" instead of "bioavailable" levels is perfectly comparable to the chemical results, which mostly measure the total concentration of the chemical.

Regarding biomarkers (section 8.4), the report points out that further research on biomarkers is needed before such could be used in a regulatory context. It should be noted that the last decade or so has seen numerous studies, workshops and research programmes devoted to biomarkers, and these should be thoroughly reviewed before any further research efforts in this area are even considered. An evaluation of the usefulness of the available biomarkers should be requested regarding their capacity as additional tools for setting reference sites. Additionally, the capability of these tools for setting differences between good and moderate status should be carefully evaluated. In particular, it could be relevant to assess if the present knowledge in the field of biomarkers may allow finding some relationships between measured parameters and the actual consequences on the structure and functions of the biological community. This would allow using biomarkers not only as "exposure" indicators, but also as "effect" indicators.

Proposed EU Water Framework Directive - Development of a Specification for the Intercalibration of Biological Monitoring Methods, WRc CO 4751/1 (CSTEE/99/4A-Add. 3)

This report could represent a very important document, in guiding the first step in defining ecological quality with the WFD. However, if the document were intended as a guidance manual as it stands, it would be totally inadequate. It is neither very precise nor very concise in describing the needs of intercalibration. The methodology, sampling design and methods of analysis need much more detail, clarity and care. There are also critical issues concerning the basic approach. The CSTEE has both conceptual and methodological concerns.

From a conceptual point of view, the CSTEE is concerned about the use, almost exclusively, of the benthic macroinvertebrate community in this exercise. Although it is obvious that the intercalibration can be conducted using a step-by-step approach, the selection of reference sites requires a holistic approach which can never be obtained looking exclusively at a single taxonomic group. In addition, the effectiveness of commonly used structural indices for describing the quality of water bodies should be carefully checked. It should be remembered that most available methods were developed when the most important problem of European rivers was organic pollution. Therefore they are based on a ranking of bioindicator organisms as a function of their sensitivity to organic pollution and oxygen depletion. In the last two decades, organic pollution has substantially decreased in most European countries. Therefore, other pollution problems, previously masked and not always evident (e.g. toxic substances), have become more relevant. As for any other taxa the sensitivity of benthic macroinvertebrates to toxic chemicals is, for each chemical, species specific, even for closely related species (Camargo and Tarazona, 1990; Camargo et al., 1992 ⁵ ). The species sensitivity to oxygen depletion is not necessarily correlated to the sensitivity of the species to the specific chemical pollution problems of each area and therefore it cannot be assumed that indicators of organic pollution are good indicators of toxic chemicals. Therefore, the effectiveness of available methods currently used to describe ecosystem response to pollution in European rivers is, for the time being, practically unknown.

From a methodological point of view, the document is not scientifically sound and the proposals are weak in terms of statistical power. Establishing an effective design for this intercalibration exercise and developing appropriate guidance for carrying it out are absolutely essential. The suggested design is rather inefficient, difficult to follow, and somewhat naïve in terms of its expectations of how the results of the pairwise comparisons are likely to appear (e.g., Figure 4.1). At the end of the exercise a large number of such curves will have been generated, they will undoubtedly vary in degree of curvature, steepness and spread, and it may be very difficult to tease out general patterns for the dataset as a whole.

The design is inefficient because efforts will be spent on selecting and measuring sites at both extremes of the quality spectrum (where assessing quality should be obvious). The CSTEE interprets the objective of the exercise to focus attention on the ability of different methods to distinguish between habitats of 'good' and 'moderate' quality (given that the WFD will aim to increase all habitats to a 'good' status). A more appropriate design may therefore be to select several sites within each Member State that represent 'good' and 'moderate' quality habitats. In this way variability within methods and variability between methods (within sites) can be compared (using an analysis of variance type design).

Given that the dependent variables will be ratios (from 0-1) there will be some statistical issues that need addressing (e.g., transformation of the ratios so that assumptions of data normality are not violated). It would be helpful at this early stage to consider issues of statistical power in order to determine the minimum difference between sites that it is desirable to detect, the level of type II error (false negative) considered acceptable, etc.

In the present sampling scheme, pairs of Member States will each take their own samples at a particular site and perform an assessment of its quality on the basis of their own classification system. This does not permit distinguishing between variability due to sampling technique and variability due to analysis/classification method. Presumably, if differences appeared in quality status between the methods used by different Member States, it would be important to determine the source of the variability so that it could be eliminated or corrected for. One way to separate these sources of variability would be to allow different Member States to classify a site on the basis of the same sample, and to allow a single Member State to classify a site on the basis of samples taken by other Member States. In addition, for the specific "intercalibration" or in more proper terms "comparison" of the different indices and approaches for the assessment of the benthic macroinvertebrate community, the evaluation of existing data should be done prior to the field exercises.

With regard to implementing the Water Framework Directive, it is unclear how the ecological status ratios will be used in practice. Are they to be used as a rough system for prioritisation and relative ranking of sites or as specific triggers for management or legal action? There may be statistical issues to consider in this context as well. For example, if 'good' quality is defined by a ratio between 0.6 - 0.7, and a particular site receives a rating of 0.58, would this be considered as less than 'good' and therefore requiring attention? If only one rating is available, this question cannot be addressed statistically; if more than one rating is available (giving for example a yearly average), then issues of statistical power, type I and type II error rates will have to be considered.

It would seem important to involve a biostatistician prior to any measurements being taken so that details of the sampling design can be optimised. Likewise detailed technical guidance on, for example, the data units to be used, presentation and reporting format to be employed, etc. will be essential for the exercise to be a success.

Finally, the proposed intercalibration grouping should cover the whole range of possibilities, giving special attention to the differences. For example, assuming that Alpine conditions represent mountain conditions (not only the Alps) the inclusion of other mountain systems (such as the Pyrenees, including France and Spain) should be required.

----------------------------------------

¹ Monitoring water quality in the future- Executive summary

Monitoring water quality in the future- Chemical monitoring
Monitoring water quality in the future- Mixture toxicity parameters
Monitoring water quality in the future- Biomonitoring
Monitoring water quality in the future- Monitoring strategies for complex mixtures
Monitoring water quality in the future- Organisational aspects

² R.M.A. Breukel. (2000) Nation-wide water monitoring: the experience of the Netherlands. RIZA document 2000.051X, Lelystad, March 2000

³ Calow, P. & Petts, G. E. (1992) The Rivers Handbook, 2 vols, Blackwell Scientific Publications, Oxford.

⁴ Tarazona J.V. (1998) Scientific concepts and uncertainties in the identification of ecotoxicological thresholds of acceptability and danger. The role of biological routes. In D.M. Pugh and J.V. Tarazona (editors) Regulation for Chemical Safety in Europe: Analysis, comment and criticism. Kluwer Academic Publishers. Dordrecht/Boston/London, pp 41-63.

⁵ Camargo J.A. and Tarazona J.V. (1990) Acute toxicity to freshwater benthic macroinvertebrates of fluoride ion (F-) in soft water. Bull. Environ. Contam. Toxicol. 45: 883-887., 1990.

Camargo, J., Garcia de Jalon, D., Muñoz, M.J. and Tarazona, J.V. (1992) Sublethal effects of fluoride ion (F-) in freshwater insect larvae (Hidropsyche bulbifera, H. exocellata, H. pellucidula and Ch. marginata) exposed to sodium fluoride. Aquatic Insects. 14: 23-30.