NEA GIG – MACROZOOBENTHOS

1. Methods and required BQE parameters

Member State / Full BQE method / Taxonomic composition / Abundance / Sensitive taxa / Tolerant taxa / Diversity / Combination rule of metrics
Portugal / yes / Not strictly – only as groups of different sensitivity / Total abundance (in Margalef index) and relative abundance of different sensitivity groups (in AMBI) / Yes (in AMBI) / Yes (in AMBI) / yes (as Shannon-Wiener diversity index; and Margalef index for richness) / Is a multivariate analysis (Factor Analysis) which calculates vectorial distances to reference conditions
Spain (Andalusia) / yes / not strictly – only as groups of different sensitivity (in BO2A) / total abundance (in Margalef index) and relative abundance of different sensitivity groups (in B02A) /
Where:
fOA = frequency of opportunistic Annelida (abundance of opportunistic annelids divided by the total abundance); OA were identified according to the AMBI ecological groups, i.e., all the annelids ascribed to EG IV and V in AMBI;
fA = frequency of amphipods (sensitive taxa) except genus Jassa (abundance of amphipods except g. Jassa divided by total abundance) / Margalef’s richness index at the family level
dfam = (Sfam – 1) / ln N,
where Sfam is the number of families and taxa higher than family and N is the total abundance / obtained by linear regression:
TaSBeM = 0.56 BO2A (EQR) + 0.44 dfam (EQR)
Spain
(Cantabria) / yes / Yes (as Bray-Curtis index) / Yes, as total abundance / yes / yes / As richness / Average metric scores
Spain (Basque Country
M-AMBI) / Yes / Not strictly – only as groups of different sensitivity, but also including richness of all species / Not strictly – only relative abundance of different sensitivity groups (all species) / Yes (in AMBI) / Yes (in AMBI) / Yes (Shannon Wiener) / Is a multivariate analysis (Factor Analysis) which calculates vectorial distances to reference conditions. Values are calculated at sample level; then, the value at water body level is calculated by weighting each location within the WB (see methodology in Borja et al., 2009b)
France / Yes / Yes / Yes (total abundance, % Bivalvia and % Polychaeta) / Yes (AMBI and % sensitive amphipods) / Yes (AMBI) / Yes (Shannon Wiener and infauna evenness index) / Average of the scores (0 ; 0,5 or 1) obtained by the 10 metrics involved in the calculation of MISS-ETW
Germany (AeTV) / yes / yes / yes / yes / yes / Yes as mean species number and alpha diversity / Tolerant and sensitive species, abundance, species number is in the AeTI index included; diversity is a Co-Parameter
Germany
(M-AMBI) / yes / Not strictly – only as groups of different sensitivity, but also including richness of all species / Not strictly – only relative abundance of different sensitivity groups (all species) / yes / yes / yes / Is a multivariate analysis (Factor Analysis) which calculates vectorial distances to reference conditions
The Netherlands
(BEQI 2) / yes / Yes, but not strictly – only as 5 species groups of different sensitivity / Not strictly – only relative abundance of (a) different sensitivity groups in AMBI and (b) relative abundances of all species in Shannon / yes / yes / Yes, as Species richness and Shannon index / Average of 3 univariately calibrated indicator EQR scores
Belgium
(BEQI) / yes / yes / Yes, as total density / yes / yes / Yes, as number of species / average metric scores
United Kingdom / yes / Not strictly – only as groups of different sensitivity, taxa number and evenness / Not strictly – only relative abundance of different sensitivity groups / yes / yes / Yes, as taxa number / Weighted average metric scores
Ireland / yes / Not strictly – only as groups of different sensitivity, taxa number and evenness / Not strictly – only relative abundance of different sensitivity groups / yes / yes / Yes, as taxa number / Weighted average metric scores
Sweden / yes / Not strictly – only as groups of different sensitivity, and species number / Not strictly – only relative abundance of different sensitivity groups / yes / yes / Yes, as species number / It is one metric


2. National reference conditions

Member State / Methodology used to derive the reference conditions / Number of reference sites (if they exist) / Location of reference sites (if they exist) / Criteria use for selection of reference sites (if they exist)
Portugal / Expert knowledge, historical and monitoring data (95 percentile of best sites), and modelling (extrapolating model results) / No real reference sites, just historical data and present monitoring data. / No real reference sites; but data from main habitats in all major TW systems in Portugal used to derive reference values. / Not applicable because TW in Portugal are always subject to some level of anthropogenic impact. However, least disturbed samples at distinct habitats were selected based on expert judgment after information on pressures in systems at sampling occasion. Also, present monitoring data from Portugal used to model the natural variation of BAT sub-metrics across TW subtidal habitats.
Spain (Andalusia) / Alternative benchmark conditions were derived by averaging the values of the TaSBeM components at the alternative nechmark sites. / No true reference sites were available, so four alternative benchmarking sites (from the official regional monitoring network) were set on four out of the nine subtypes (ecotopes) defined for intercalibration common type D. These were located in Andalusia (southern Spain) at (regional codes in brackets): Guadiana river salt marshes (40T0060) and estuary (40T1020), Piedras river estuary (62T0050) and mouth of the Guadalquivir estuary (51T0120). / aprox. coordinates (DD ETRS89, long. / lat.):
st. 40T0060: -7,4066 / 37,1891
st. 40T1020: -7,324 / 37,2081
st. 62T0050: -7,0685 / 37,2103
st. 51T0120: -6,3431 / 36,8002 / Not applicable because all the transitional waters in Andalusia are subject to anthropogenic impacts. Alternative benchmarking sites were selected on the basis of least disturbed conditions and pressure information (screening for lower urban and industrial sewage and lower agriculture land use sites), historical data and expert knowledge.
Pressure criteria: 1. Transitional LUSI ≤ 4; 2. Concentration of soluble P in sediment < 10mk/kg
Spain (Cantabria) / Historical data from least disturbed conditions / 40 (not true reference but least disturbed sites) / Cantabria Region. North Coast of Spain. / Least disturbed sites, mainly regarding the absence of sewage discharges.
Spain (Basque Country)
M-AMBI / Expert knowledge, Historical data, Modelling (extrapolating model results) (see Muxika et al., 2007) / Not true but low disturbed areas (see Muxika et al., 2007) / ‘virtual’ reference locations, based on historical data and expert judgment (see Muxika et al., 2007) / In order to determine high reference AMBI values, average densities of species, for the period 1995–2003, were obtained. Subsequently, opportunistic species (EG IV and V) were removed from the species list and the AMBI was derived for each typology, following the guidelines in Borja and Muxika (2005). See also Muxika et al. (2007)
France
(MISS-ETW) / Existing near-natural reference sites, Expert knowledge, Least Disturbed Conditions / Least impacted sites (expert knowledge) among a campaign performed on 11 estuaries of the Channel-Atlantic coastline, in autumn 2008. / 11 estuaries of the French Channel & Atlantic coastline (Somme, Seine, Orne, Mont Saint Michel Bay, Trieux, Belon, Loire, Aiguillon Cove, Charente, Gironde, Bidassoa) / A reference conditions dataset was built, including the stations (238) of all sites that clearly corresponded to a given set of environmental variables defining the benthic habitat.
For each defined habitat, reference conditions datasets were analysed to obtain percentiles 5%, 25%, 75% and 95% for each metric.
This permitted to capture the range of variations of a given metric.
Germany (AeTV) / Expert knowledge, Historical data, Least Disturbed Conditions / Reference species list, no reference sites / Reference species list, no reference sites / Reference species list, no reference sites
Germany
(M-AMBI) / Expert knowledge, Least Disturbed Conditions / not true reference but least disturbed sites / Weser Estuary / Lowest AMBI, highest Shannon Diversity and Species number were chosen from dataset. Different reference values were determined for different habitats (salinity tranches, sediments and sub- and intertidal)
The Netherlands
(BEQI-2) / (a) Historical data for 1990-2005; (b) theoretically correct reference value for AMBI(ref) = 0; theoretically correct bad values: S(bad) = 0; H’(bad) = 0; AMBI(bad) = 6. (c) Statistical modelling for S(ref) and H’(ref): 99 percentile of S and H’ for complete ecotope dataset (highest indicator value which is robustly not an outlier). The use of a low percentile for bad conditions has been abandoned since (a) we now recognize that the theoretical bad values used in the NEAGIG are correct and (b) the EQR-values obtained with this did not meet the expert opinions. / Not true reference sites, but least disturbed sites can be selected if necessary. / Data from the Eems Dollard (not officially part of the IC process, period 1991 – 2007) were used to check the BEQI-2 design and sensitivity. / Not applicable because TW in The Netherlands are always subject to at least some level of anthropogenic impact. However, least disturbed samples from distinct sampling locations can be selected based on expert judgment using information on pressures at the sampling locations.
Belgium / Expert knowledge, Historical data, Least Disturbed Conditions / No reference sites / Reference condition based on samples, reflecting the spatial and temporal variation under least disturbed conditions. The approach to do this is by expert judgment.
United Kingdom / Expert knowledge, Historical data, Least Disturbed Conditions, Modelling (extrapolating model results) / All available data from UK and Ireland used to model IQI sub-metric reference condition values. / Approach does not require selection of specific reference sites / Reference condition samples were identified as being from least disturbed conditions, selected on the basis of a) expert judgement and b) from impact gradient study control sites. Reference condition values for AMBI, Simpsons and taxa number were identified from the data. Data was used from sites with low levels of natural disturbance and outliers (e.g. those with anomalously high taxa numbers in contrast to the remaining data) were identified according to expert judgement and excluded.
Ireland
Sweden / Expert knowledge, Least Disturbed Conditions / unknown / unknown / unknown

Remarks:

ES (AN)

1. Reference conditions are still under development. As new data will be incorporated into the regional datasets in the following years, reference conditions will be updated accordingly and are expected to be finished in time for their application in the RBMP of 2015.

2. Description of benchmark conditions:

Frequency of opportunistic polychaetes inferior to 0.10 and frequency of amphipods higher than 0.30, which correspond to BO2A values ~ 0.030. Margalef’s richness to the family level not lower than 0.95.

Benthic communities with high abundance of individuals from genus Corophium and Nephtys.


3. National boundary setting

Member State / Methodology used to set H/G boundary / Methodology used to set G/M boundary
Portugal / For TW, boundaries were taken from the coastal phase I intercalibration exercise and will be harmonized over the intercalibration exercise. Boundaries compliance with normative definitions will be verified after harmonized over the intercalibration exercise. / For TW, boundaries were taken from the coastal phase I intercalibration exercise and will be harmonized over the intercalibration exercise.
Boundaries compliance with normative definitions will be verified after harmonized over the intercalibration exercise.
Spain (Andalusia) / H/G boundary obtained by calculating the 25th percentile of TaSBeM values for alternative benchmark samples. / G/M boundary derived from discontinuities in the relationship between TaSBeM and the transitional-LUSI multipressure index: given by the 25th percentile of TaSBeM values for all samples with t-LUSI≤3 (excluding reference samples).
Spain (Cantabria) / Equidistant division of the EQR gradient / Equidistant division of the EQR gradient
Spain (Basque Country) / Boundaries taken over from the coastal intercalibration exercise, and harmonized during the second phase for TW intercalibration / Boundaries taken over from the coastal intercalibration exercise, and harmonized during the second phase for TW intercalibration
France / Calibrated against pre-classified sampling sites, statistical approach and expert judgment
It is considered that no pristine condition occurs in the large majority of European estuarine systems. Therefore no ‘High’ ecological quality status was defined. / Calibrated against pre-classified sampling sites, statistical approach and expert judgment
Good : all metrics values are higher than percentile 75% or lower than 25% but no more than 2 metrics (over 10) displaying values higher than percentile 95% or lower than percentile 5%
Moderate : Idem GOOD + more than 3 but less than 5 metrics display values higher than percentile 95% or lower than percentile 5%
Germany (AeTV) / Calibrated against pre-classified sampling sites / Calibrated against pre-classified sampling sites
Germany (M-AMBI) / German boundaries taken over from the intercalibration exercise phase 1 / German boundaries taken over from the intercalibration exercise phase 1
The Netherlands (BEQI-2) / Equidistant division of the EQR gradient, resulting in a H/G boundary of 0.8. We think that this H/G boundary is theoretically correct, if (a) the metric design is logical, (b) the reference and bad values are set (if possible) theoretically correctly and (c) a linear human pressure-response curve of the indicators in the metric is assumed. / Equidistant division of the EQR gradient, resulting in a G/M boundary of 0.6. We think that this H/G boundary is theoretically correct.
Belgium / High-good boundary set as a percentile value (25, 75 or 50%, depending on parameter) of the parameters value distribution (bootstrapping) in the reference dataset / Good/Moderate boundary set at the 5 percentile value of the parameters value distribution (bootstrapping) in the reference dataset
Equidistant division of the EQR gradient for the other classes (M/P; P/B).
Ireland / Boundaries taken over from the CW intercalibration exercise and applied to TW assessments. / Boundaries taken over from the CW intercalibration exercise and applied to TW assessments.
United Kingdom / Boundaries taken over from the CW intercalibration exercise and applied to TW assessments. / Boundaries taken over from the CW intercalibration exercise and applied to TW assessments.
Sweden / a mixture of statistical aspects such as percentiles, statistical tests, and expert judgments / a mixture of statistical aspects such as percentiles, statistical tests, and expert judgments


4. Pressures-response relationships