Introduction

The European Water Framework Directive (WFD; 2000/60/EC) establishes a basis for the protection and improvement of transitional (i.e. estuarine) and coastal waters, amongst other systems. Its final objective is to achieve not less than ‘good ecological status’ for all waters, by 2015 (Directive, 2000; Borja, 2005). Ecological assessment is based upon the status of biological, hydromorphological and physico-chemical quality elements. The biological elements to be considered are phytoplankton, macroalgae, angiosperms, benthic macroinvertebrates and, in transitional waters only, fish.

Ecological status (ES) of a water body is determined by comparing data obtained from monitoring networks (Ferreira et al., 2007) with reference (undisturbed) conditions, thus deriving an ecological quality ratio (EQR). The EQR is expressed as a numerical value lying between 0 and 1; ‘High status’ is represented by values close to 1, whilst ‘Bad status’ values lie close to 0. The range is divided into five ES classes, ‘High’, ‘Good’, ‘Moderate’, ‘Poor’ and ‘Bad’. This legislative demand presents substantial challenges for scientists, who are asked to deal with ‘non-scientific’ concepts. ‘Good’ or ‘Poor’ ES is immediately and clearly understandable by everyone, yet at the same time absolutely vague when they have to be translated into a set of rules or quantified into a numerical value. Therefore, different experts may have a different definition of them.

The normative definitions (Annex V) of the WFD describe the parameters of the biological quality elements that must be included in the ES assessment of a water body. For the marine macroinvertebrate community, these include composition and abundance of invertebrate taxa and the proportion of disturbance-sensitive and tolerant taxa. Following these criteria, several indices and approaches have been published to assess the benthic invertebrate ES (e.g. Borja et al., 2000, 2004a, 2007; Simboura & Zenetos, 2002, Rosenberg et al., 2004; Muxika et al., 2007; Perus et al., 2007; Meyer et al., 2008).

Proposed methodologies should be applicable to the range of types into which the main European eco-regions are divided (see CIS, 2003; Borja et al., 2004a; Heiskanen et al., 2004; European Commission, 2008; GIG, 2008). The purpose of defining these types is to enable type-specific reference conditions to be established, making it possible to assess the ES for different geographical and habitat conditions. These type-specific reference conditions are the basis of the classification schemes, and, as such, impact on all subsequent aspects of the implementation of the WFD (including intercalibration of the quality class boundaries assessed by different methodologies, assessment of the quality status of each of the biological elements, and monitoring, assessment and reporting of the water body status).

The objective of this contribution is to review the current status of the macroinvertebrate methodologies proposed for European waters under the WFD, showing the weaknesses and strengths of the WFD implementation process and proposing future research topics and challenges.

Some of the information in this contribution have been obtained from the European Commission (2008) and the Geographical Intercalibration Group (GIG) reports (2008), available at http://circa.europa.eu/Public/irc/jrc/jrc_eewai/library.

European types

The European maritime area has been divided into four eco-regions (Fig. 1): (i) the Atlantic/North Sea Eco-region Complex comprising the Northeast Atlantic Ocean, North Sea, Norwegian Sea and the Barents Sea; (ii) the Baltic Sea Eco-region; (iii) the Mediterranean Sea Eco-region and (iv) the Black Sea (CIS, 2003; Casazza et al., 2003; Borja, 2005). These eco-regions have been used as the basis for the intercalibration groups (GIG, 2008), with the objective of harmonising the process of water quality assessment across the large variety of marine habitats in Europe.

Fig. 1
figure 1

Map of Europe showing eco-regions and countries applying the WFD (those with names) and those which have approved methodologies for macroinvertebrate quality assessment. Key: W.M.—without method agreed; M.A.IC.—with method agreed and intercalibrated (France has a method agreed and intercalibrated within the Atlantic, but not in the Mediterranean); M.A.N.IC.—with method agreed, but not intercalibrated

Full size image

Under the WFD, each eco-region must be divided into types, which can be defined according to two alternative typology systems. System A uses eco-region, salinity and mean depth as determination factors. System B uses latitude, longitude, tidal range and salinity as obligatory factors; together with current velocity, wave exposure, mean water temperature, mixing characteristics, turbidity, retention time (in the case of enclosed bays), mean substratum composition and water temperature range as optional factors.

As stated by Borja (2005), when scientists are confronted with establishing types, their first reaction is often to account for a high number of very specific types (resulting from the combination of the different geomorphological and hydrodynamic characteristics included in the definition). However, this approach produces an unmanageable situation in subsequent steps of the WFD, e.g. the establishment of the monitoring and management plans for each of the water bodies and types.

One of the most common criticisms of WFD typology implementation arises from the over-simplicity of the approach, from a scientific point of view, because of the heterogeneity of habitats within the different types and water bodies (see Borja, 2005). From the dozens (even hundreds) of potential types across Europe, only 18 coastal types have been intercalibrated (Table 1). These are divided into 6 types in the Baltic Sea; 10 types in the Atlantic Sea; 1 type in the Mediterranean Sea; 1 type in the Black Sea (European Commission, 2008; GIG, 2008).

Table 1 Types intercalibrated in Europe, at each of the eco-regions (seas), characterised using different variables, such as salinity, wave exposure, water depth, tidal range, current velocity, ice cover, water mixing and residence time
Full size table

These types reflect, at a certain level, the main eco-region characteristics. Hence, (i) Baltic Sea types present generally different salinity values, ranging from marine waters to nearly freshwater, are sheltered coasts, and experience some ice cover in winter; (ii) Atlantic Sea types reflect the wide latitudinal range of this eco-region, with very different characteristics and (iii) the Black Sea type reflects the main influence of Danube River upon the coastal salinity.

In turn, within the Mediterranean Sea, the data available did not allow a complete intercalibration among the Member States (MSs) that participated in the IC exercise; only one type (representing shallow sedimentary bottoms, not the prevailing one, among the four initially selected) could be intercalibrated.

Currently, there has been no intercalibration of ES assessment systems for transitional water bodies, as typifying and intercalibration of these waters on a European scale represent huge challenges, due to the heterogeneity of waters which include both estuaries of very different sizes and coastal lagoons with a large variety of connections to the open sea and to continental waters.

On the other hand, hard bottom communities have not yet been intercalibrated, mainly because sampling methods, significance of presence/absence and density of species, and methods for quality assessment are profoundly different from those in use with mobile sedimentary bottoms.

Methodologies in assessing benthic quality status

The benthic invertebrates are well-established components in environmental quality status assessments, with various studies having demonstrated that the benthic macroinvertebrates respond relatively rapidly to anthropogenic and natural stress (Pearson & Rosenberg, 1978; Dauer, 1993). Several authors have reviewed the use of indices for assessing the benthic ‘health’ of a system (Díaz et al., 2004; Pinto et al., 2009). Some (e. g. Washington, 1984) contend that an index is unlikely to be universally applicable, as organisms are not equally sensitive to all types of anthropogenic disturbance and are likely to respond differently to various types of perturbation. Several methods have been published to assess the benthic macroinvertebrates ES in European marine waters (see complete names in Table 2): AMBI (AZTI’s Marine Biotic Index) and M-AMBI (Borja et al., 2000, 2004a; Muxika et al., 2007); Bentix (Simboura & Zenetos, 2002); BQI (Rosenberg et al., 2004). Moreover, many others have been proposed during the WFD intercalibration exercise (Borja et al., 2007), but are only published as annexes to the WFD reports (GIG, 2008) and by the European Commission (2008).

Table 2 Benthic quality assessment methods proposed and/or approved by each Member State of the European Union, including the metrics used in the methodology
Full size table

In total, 12 different methodologies have been officially accepted by European MSs (Table 2). Most of these methods are multimetric, i.e. combining several metrics into an equation (e.g. BQI, DKI, IQI, MarBIT, NQI, etc.), others are multivariate (e.g. M-AMBI and P-BAT) and some are univariate (e.g. MEDOCC and Bentix). Some countries use different methodologies within different eco-regions (e.g. Germany, in Northeast Atlantic and Baltic).

The methodologies vary in their use of the parameters outlined in the WFD normative definitions; some of them utilising all metrics prescribed by the Directive, some only parts thereof and some incorporating different parameters. Hence, for (i) the proportion of disturbance-sensitive taxa, 13 countries (methods) use AMBI as the indicator, 3 the BQI, 2 the Bentix and 2 other methods; (ii) the level of diversity, 11 countries use Shannon’s index, 2 Simpson’s index and 1 ES100 (Hurlbert, 1971) and (iii) abundance of invertebrate taxa, 6 incorporate density, 14 richness, 1 Margalef index, 1 abundance distribution and 2 incorporate biomass (Table 2). In addition, two countries incorporate similarity and one uses feeding guilds.

The selection, from a list of ecologically relevant candidate metrics to be used within any methodology, should emphasise metrics that are sensitive (respond to anthropogenic action—both degradative and restorative) and representative (able to measure status and trends relative to policy decisions and management actions) (after Borja & Dauer (2008)). In general, the metrics used in the intercalibrated European methodologies are based upon community level characters that represent key community aspects and these characteristics are related to the normative definitions required by the WFD.

Although most of the benthic methods have only been applied to coastal waters, some of them are also being used or are being further developed for use, within transitional waters, such as M-AMBI (Muxika et al., 2007; Borja et al., 2009c), IQI (Prior et al., 2004) or Bentix (Simboura & Reizopoulou, 2008). At the same time, a plethora of new methodologies is being proposed across Europe for water categories (lagoons and coastal and transitional waters), types or habitats; some of which are mentioned below.

  1. (i)

    Multimetric indices based on macroinvertebrates and aquatic flora, for lagoons: the Ecofunctional Quality Index (EQI) (Fano et al., 2003) and the Fuzzy Index of Ecosystem Integrity (FINE) (Mistri et al., 2007, 2008).

  2. (ii)

    Daphne, a multimetric index for coastal waters (Forni & Occhipinti-Ambrogi, 2007).

  3. (iii)

    Brackish Water Benthic Index (BBI), for Baltic Sea (Perus et al., 2007).

  4. (iv)

    Benthic Opportunistic Polychaetes Amphipod index (BOPA) (Gómez Gesteira & Dauvin, 2000; Dauvin & Ruellet, 2007; Dauvin et al., 2007), for coastal and transitional waters.

  5. (v)

    Methods or indices based on taxonomic-free descriptors, such as biological-functional traits or body size distributions, for lagoons (Mouillot et al., 2006; Reizopoulou & Nicolaidou, 2007; Marchini et al., 2008).

  6. (vi)

    The Benthic Index based on Taxonomic Sufficiency (BITS) for non-tidal lagoons (Mistri & Munari, 2008).

Many of the ‘new’ indices or methods have been developed for Mediterranean coastal waters and lagoons in particular, where the peculiar environmental conditions make it difficult to use ‘generalised’ methods, and have stimulated scientists to create ad hoc methods. As a matter of fact, the application of indices mainly developed for the Atlantic coastal waters has achieved controversial results in Mediterranean lagoons or enclosed basins (Ponti & Abbiati, 2004; Marín-Guirao et al., 2005, Labrune et al., 2006, Pranovi et al., 2007; Simboura & Reizopoulou, 2008). On the other hand, the accuracy of methods for evaluating ES is expected to decrease when they are applied at larger spatial scales, because no model (or method or index) can be simple, general and accurate at the same time (Scardi et al., 2008).

The increasing number of proposed methods makes it difficult to assess and compare their suitability for quality assessment. Only in very few cases (see Borja & Muxika, 2005; Borja et al., 2008b) do the authors of the methodologies give guidance in the application of the method, indicate the comparability of their results and discuss the strengths and/or weaknesses of the indices used. This guidance is important for further quality assurance and harmonisation of the methods at the European level and is essential for the implementation of management decisions based on the assessment, which may involve significant economic consequences.

The recommendation of Díaz et al. (2004), on placing greater emphasis on evaluating the suitability of indices that already exist prior to the development of new ones, has not been followed in Europe. Although some authors (e.g. Borja et al., 2004b) have argued for consensus in the use of these new methodologies, it no longer seems possible. This is probably due to the large ecogeographical differences across Europe. However, the selection of multiple methods in assessing benthic quality makes any further validation and intercalibration of the methodologies used in the WFD more difficult. Moreover, the presence of the many, often local, types, to which the methods are applied, requires the definition of the type-specific reference conditions to be used, making more complicate the comparison.

Reference conditions

The reference condition for a biological quality element in a water body type is a description of the biological element relating to undisturbed (=pristine) conditions, i.e. with no, or with only a very minor, impact from human activities (Directive, 2000; Borja et al., 2004a; Muxika et al., 2007). The purpose of setting reference condition standards is to enable the assessment of the biological quality relative to these standards. Also by agreeing on reference conditions for a shared ecosystem type, MSs are giving a first step towards comparability of their assessment methods. Type-specific reference conditions must summarise the range of possibilities and values for the biological quality element, over periods of time and across the geographical extent of the type (CIS, 2003).

The WFD identifies four options for deriving reference conditions: (i) comparison with an existing ‘pristine’/undisturbed site (or a site with very minor disturbance); (ii) using historical data and information; (iii) using models or (iv) using expert judgement. Borja et al. (2004a) have stated that one of the problems in deriving reference conditions for some types arises from the absence of un-impacted areas or lack of pre-industrial historical data. The use of ‘virtual’ reference locations as an ‘expert judgement’ approach has therefore been defined and proposed by Borja et al. (2004a) and Bald et al. (2005) and has been used successfully in macroinvertebrate status assessment (Rosenberg et al., 2004; Muxika et al., 2007; Borja et al., 2009a). Modelling the reference conditions using autecological data has also been used (Meyer et al., 2008). In some eco-regions (e.g. the Mediterranean Sea), marine reserves have been proposed as possible reference areas, as they provide the best ecological conditions within the eco-region (Casazza et al., 2004).

The methods used to determine reference conditions varies greatly between MSs: (i) using data from marine reserves (Greece), (ii) deriving them from the autecology of all macrozoobenthos species reported (Germany); (iii) using the highest (or lowest, depending on the metric) values of each of the metrics found in the data material classified to be at least in good status (Denmark); (iv) using the median of the 10% highest values of the method applied (Finland) or (v) using a mixture of expert judgment, historical data and modelling (Spain).

However, it is very difficult to confirm the values used as reference conditions, and, to our knowledge, complete reference conditions have only been published for the Spanish method (Muxika et al., 2007; Borja et al., 2009b) and, partially, for the British, Irish, Danish and Norwegian methods (Borja et al., 2007). More information, although still only partial, can be found in GIG (2008). This lack of detailed, yet essential, information makes comparison of some methods extremely difficult. Proposers of new indices should provide all the necessary details.

Validation of methodologies

Following Borja & Dauer (2008), validation of the methodologies should ideally include: (1) testing of the index using an independent data set different to the index development data set (calibration data set), (2) setting a priori correct classification criteria and/or (3) presentation of a strong a posteriori justification for use, based upon best professional judgment (see also Duel et al., 2007). Moreover, it should be demonstrated that the methodologies respond to anthropogenic pressures. Independent validation, by scientists other than those proposing the methodology, should be done. However, this is not the case for several of the European methodologies.

Some degree of validation has been undertaken for some of the indices, as shown below.

  • AMBI and M-AMBI have been validated using many different anthropogenic pressures, on a worldwide basis (e.g. Borja et al., 2003; Solís-Weiss et al., 2004; Chenery & Mudge, 2005; Muniz et al., 2005; Muxika et al., 2005; Carvalho et al., 2006; Quintino et al., 2006; Dauvin et al., 2007; Fleischer et al., 2007; Cheung et al., 2008). Following these investigations, AMBI has been found to respond to different drivers and pressures, such as: hypoxia and eutrophication processes; oil platform discharges; engineering works; dredging; fish aquaculture; etc.; however, not to some hydro-morphological pressures, such as sand extraction (Muxika et al., 2005). M-AMBI has also been validated against several pressures, including hydromorphological ones (Muxika et al., 2007; Bigot et al., 2008; Borja et al., 2009a).

  • Bentix has been checked using anthropogenic pressures, such as those coming from aquaculture (Aguado-Giménez et al., 2007), mining debris (Simboura et al., 2007) or eutrophication (Simboura et al., 2005).

  • Several different pressures, sometimes unspecific, have been used in the validation of BQI (e.g. Rosenberg et al., 2004; Reiss & Kröncke, 2005; Labrune et al., 2006; Dauvin et al., 2007; Zettler et al., 2007).

  • For other methods, the literature is less clear (see Prior et al., 2004; Josefson et al., 2008).

In a very few cases, methods used in Europe have been compared (even intercalibrated) with others used in the USA (Borja et al., 2008a; Chainho et al., 2008). The results from these exercises indicate that, although the approaches in both geographical areas are different, the conclusions regarding the benthic health can be very similar.

Intercalibration of methodologies

Prior to the implementation of WFD assessment, any proposed methodology must be intercalibrated between the MSs within an eco-region (Directive, 2000; Borja et al., 2007). Each MS shall divide the EQR scale for their monitoring system into the abovementioned five ES classes, by assigning a numerical value to each of the class boundaries. The value for the ‘High/Good’ and the ‘Good/Moderate’ class boundaries should be established through the intercalibration exercise. This is to ensure that the established class boundaries are consistent with the normative definitions of the WFD, and the different methodologies used are comparable between MSs.

Currently, there are very few peer-reviewed publications concerning the intercalibration of macroinvertebrates (Borja et al., 2007; Simboura & Reizopoulou, 2008; Occhipinti-Ambrogi et al., accepted), although many of the coastal types have been already intercalibrated within the WFD intercalibration groups (European Commission, 2008; GIG, 2008).

From the Baltic Sea, only Finland, Sweden, Germany and Denmark have intercalibrated their methodologies within five of the European types (Table 3, Fig. 1). The other Baltic countries (Estonia, Latvia, Lithuania and Poland) are at different stages of developing assessment methods and have only been able to contribute partially to the intercalibration exercise (Fig. 1). It is interesting to note the important differences in the boundaries across and within the types, indicating that the biogeographical and hydrographical characteristics (mainly salinity) of the types requires type-specific boundaries to make the results from different methods comparable. Hence, the boundaries for high-good classes range between 0.69 and 0.99 of the EQR, whilst the boundaries for good-moderate range between 0.31 and 0.60 (Table 3).

Table 3 Types intercalibrated between different Member States, setting the boundaries between High-Good and Good-Moderate macroinvertebrate quality status
Full size table

In the case of the North East Atlantic Sea, all countries have participated in the intercalibration, assigning boundaries for 12 types or sub-types (Table 3, Fig. 1). In this eco-region, the ranges of the boundaries are narrower (even it encompasses a much broader latitudinal range than previous geographic group), with EQRs extending from 0.67 to 0.92, in the case of high-good boundary, and 0.53 to 0.81, in the good-moderate boundary (Table 3).

Within the Mediterranean Sea, only four out of six MSs (Cyprus, Greece, Slovenia, and Spain) have agreed their boundaries, within one type (Table 3, Fig. 1). The EQR values of the boundaries range between 0.73 and 0.83 for high-good and from 0.47 to 0.62 for good-moderate classes (Table 3). Italy and France have not yet completed the process. Italy is working towards evaluating actual reference sites, which take into account the Italian coastal variability.

Within the Black Sea, Bulgaria and Romania have agreed EQR boundaries for each of the metrics or methods they use (Fig. 1), ranging between 0.83 and 0.89 for high-good and from 0.53 to 0.69 for good-moderate classes (Table 3).

When comparing median, mean and standard deviation values for the boundaries between status classes, it can be seen that mean high-good boundary lies around 0.8 in all eco-regions, whilst good-moderate boundary lies around 0.6, except in the Baltic Sea where the mean value is 0.41 (Fig. 2). Median values are close to mean values (≤0.02 EQR), excepting in the Baltic Sea, in which the differences are of 0.05 EQR, within both quality classes (Fig. 2). As mentioned above, this is due to the special characteristics of this sea, where the strong salinity gradient, together with a general dominance of euryoecious and tolerant species, may make it difficult to distinguish between natural stress and anthropogenic pressures. This problem has been discussed elsewhere (Dauvin, 2007; Elliott & Quintino, 2007), and can cause difficulties in implementing the WFD within transitional waters, especially in oligohaline and tidal freshwater stretches of the estuaries.

Fig. 2
figure 2

Median, mean and standard deviation values for the boundaries between High-Good and Good-Moderate status, calculated within each European eco-region, using data in Table 3

Full size image

Future research

Indices used to evaluate ES are often perceived as an objective procedure. In reality, they involve many steps that are based on subjective expert judgement, which is the key for any environmental assessment, evaluation or diagnosis (Scardi et al., 2008). All the proposed macroinvertebrate indices are based on some subjective choices, such as: metrics selection, classification of species into ecological groups, identification of coefficients in the mathematical formulas, definition of thresholds, etc. The use of expert judgement has not always been acknowledged and made evident by authors proposing the indices, thus leaving room for discussion and further refinement (Weisberg et al., 2008). The reduction of subjective steps, towards a more objective methodology, is necessary to reach the required scientific agreement for the translation of benthic indices into national legislation. Advanced computational techniques, such as fuzzy logic or neural networks, could provide help in dealing with the sources of uncertainty that affect the process of index development (Scardi et al., 2008; Marchini et al., 2009).

In a special issue of Marine Pollution Bulletin, focussing on marine integrated assessment Borja et al. (2009c) identified four areas from which scientific agreement was needed to satisfy future macroinvertebrate quality management: (i) reduction of the present bewildering array of available indices by identifying the index approaches, components and formulations that are most widely successful (Index Format); (ii) establish minimum criteria for index validation processes that demonstrate index accuracy and reliability (Index Validation); (iii) compare and intercalibrate methods to achieve uniform assessment scales across sites and habitats (Index Intercalibration) and (iv) integrate indices across media and ecosystem elements (Index Integration). In relation to Index Format, the challenge for the next decade is to accomplish sufficient index performance comparisons to reach scientific consensus on preferred index approaches for macroinvertebrates. The influence of local and regional natural variabilities (Bonsdorff et al., 2003), and the nature of the habitat or ecosystem element on index accuracy and stability must also be investigated (de Paz et al., 2008).

Index validation is the part of index development that critically evaluates the accuracy and precision of an index (Borja et al., 2009c). The credibility of an index or index combination, among scientists and managers, and its acceptability for evaluating condition at the geographic scale and level of habitat heterogeneity of interest depend on the demonstration of index reliability in a meaningful validation process (see Borja & Dauer, 2008). Any new methodology or index without a reliable and replicable validation, using independent sources, should be discarded or used with extreme caution (Borja et al., 2009c).

When validating an index for the WFD, the use of clear gradients of disturbance and/or recovery, both spatial and temporal, is essential (see Teixeira et al., 2008b; Borja et al., 2009a). Any mismatch between the biological and physical–chemical monitoring results, which may be an indication that the biological methods used are not sensitive to the effects of anthropogenic changes, should also be investigated (Chainho et al., 2008). In this way and prior to any intercalibration exercise, clear and reliable reference conditions for each of the types (even habitat: Blanchet et al., 2008) should be determined. Until now, some interesting approaches have been published (Nielsen et al., 2003; Borja et al., 2004a; Muxika et al., 2007; Teixeira et al., 2008b), but more investigation is needed.

Although many indices were successfully validated during the last decade, the challenge for the coming years is the intercalibration of methods within transitional water types (some attempt to compare indices have been already undertaken, such as in Pinto et al. (2009)). Scientific agreement regarding an approach in transitional waters, which takes into account the high number of natural patterns responsible for general trends regarding benthic diversity variation, is urgently needed (Teixeira et al., 2008a).

Other issues, related to the integration of indices at the water body level, have been undertaken in some of the published studies (Borja et al., 2008c, 2009a, b); however, more research in this topic is needed.

As stated by Borja et al. (2009c), the proliferation of indices adds an element of confusion back into what environmental managers had hoped to simplify. Some of the confusion arises as a result of the different processes used for developing, calibrating and validating indices in different regions. This in turn leads to inconsistencies in assessment across regions. Whereas, the last decade was characterised by an explosion of indices, the coming years should be those of consolidation and agreement (Borja et al., 2009c).

Some of these challenges, together with others such as the influence of climate change on reference conditions, will be undertaken in a new European research project named ‘Water bodies in Europe: Integrative Systems to assess Ecological status and Recovery’ (WISER), coordinated by Daniel Hering (University of Duisburg-Essen, Germany). One of the modules, focusing on marine and transitional waters, will provide (i) new validation of some indicators; (ii) investigation of the response of indicators to different pressures, such as hydromorphological pressures, eutrophication, pollution (metals and organic compounds), etc.; (iii) investigation of macroinvertebrate good ecological potential assessment, within Heavily Modified Water Bodies (see Borja & Elliott, 2007); (iv) an error estimation exercise of the uncertainty in assessing the ES, including a wide range of geographical regions and types; (v) different ways of combining single metrics into multimetrics or holistic assessment; (vi) a sensitivity analysis indicating which parts of the metrics are redundant, if there is duplication or double-counting, and which proportion of the end result is dependent on the individual metrics; (vii) correlations between the sensitive metrics/indicators developed above and (viii) data generated using low-cost monitoring methods will be tested.