2012 INTERCALIBRATION TECHNICAL REPORT

MEDITTERANEAN RIVER GIG – MACROPHYTES

Draft 22 February 2012

1. Introduction

The Medittaranean GIG did not complete intercalibration for macrophytes in the first round.For the second intercalibration round, the Mediterranean GIG has filled this gap and completed intercalibration for seven national methods.

2. National assessment methods and their compliance with the WFD normative definitions

Participating countries in the work of the Mediterranean GIG for the BQE river macrophytes were Portugal, Greece, France, Spain, Cyprus, Slovenia and Italy. Malta, although geographically part of the GIG, has not participated.

An overview of the national methods that were intercalibrated is provided in Table 1 below. Detailed descriptions of the national methods, including the metrics used, the methodology for setting national reference conditions and the methodology for setting class boundaries, are provided in Annex A. They all are finalized methods ready to be intercalibrated; only the methods from Italy, Portugal, Slovenia and Spain were formally agreed national methods at the time the technical work was carried out. Detailed descriptions of the methods from France, Italy an Slovenia are included in the WISER method overview (www.wiser.eu).

Table 1 – Overview of national methodsMS Method Abb.Cyprus Multimetric Macrophyte Index MMI

Biological Macrophytes Index for Rivers IBMRFrance Biological Macrophytes Index for Rivers IBMRGreece Biological Macrophytes Index for Rivers IBMRItaly Biological Macrophytes Index for Rivers IBMRPortugal Biological Macrophytes Index for Rivers IBMRSlovenia River Macrophyte Index RMISpain Biological Macrophytes Index for Rivers IBMR

Using the information reported in Annex A, the national methods were checked for compliance with the WFD normative definitions (Table 2). The GIG concluded that there is yet some work to be done concerning taxonomy and individual response to pressures with macroalgae and bryophytes in Mediterranean regions, contrarily to vascular plants. More basic research has to be conducted in order to have fully operational methods including macroalgae. Macroalgae includes various categories of algae e.g. Charophytes (Chara sp.), red algae (Batrachosperum sp.), green algae and filamentous algae (e.g. Cladophora sp., Chaetomorpha sp.). Macroalgae react differently in time and space comparing to angiosperms and some of them are able to tolerate the drought event. Generally the higher taxonomic plants (angiosperms) react within years to changes in degradation whereas algae can react within days or even hours. This characteristic of the green algae and macroalgae can be especially relevant in the evaluation of the ecological status of temporary Mediterranean rivers (type RM5) where the water course can partially or totally dry out during the warm period. Nonetheless, the presence of vast amounts of green filamentous algae, occupying the habitats of vascular plants only occur in poor and bad conditions, therefore can be assumed irrelevant in the present exercise.There are technical difficulties for determining absolute abundance during sampling, and relative abundances are generally judged adequate, namely in methods already used for many years, such as the IBMR method that most countries are using in the present exercise.

The assessment methods of the MS were considered compliant and ready for intercalibration

2

Table 2 – Check of national methods against the WFD normative definitionsCompliance criteria Compliance checking conclusions1. Ecological status is classified by one of five classes (high,

good, moderate, poor and bad).Cyprus, France, Greece, Italy, Portugal, Slovenia, Spain – yes

2. High, good and moderate ecological status are set in line with the WFD’s normative definitions (Boundary setting procedure)

Cyprus – the median value of Multimetric Macrophyte Index of the unstressed sites (0.880) was set as the numerical boundary between H/G ecological status. The rest of the range was divided into four equal classes; for RM4, IBMR class boundaries were estimated by dividing by four the 25th percentile value of unstressed sites (0.795) using IBMR normalized values; thus the class boundaries range 0.199.

France – Boundaries were defined using a reference dataset collected on reference sites, meeting the "reference condition" criteria given on the EU Guidance. For each group of river national types, the base for defining boundaries is the percentile 25 of reference values for H/G boundary. The G/M boundary is derived from equidistant division of the rest. These values were adjusted from expert judgment to balance the lack of data (few reference data or no data for several river types). Boundaries will be adjusted in the limit of IC results, from the data collected into the reference and monitoring networks.

Greece – the median of the IBMR normalised value of the unstressed sites was used as a boundary for the H/G ecological class (0.75). The other quality classes were obtained by dividing the higher value of the normalized index by 4.

Italy - boundary setting has been identified using data from sites belonging to different quality level assessed on expert judgement, others BQE (macrobenthos), historical series of pressure data (land use, hydrological, morphological and chemical features).The H/G boundary has been identified also by assessing IBMR variability range in references sites for each considered river typology. Boundaries for RM5 are not final.

Slovenia – Reference value was determined as a median value of the index RMI at the reference sites. This value is 0.72. Boundary values for the five classes of the ecological status were determined on the basis of the changing of portion of the frequency of so called “good” and “bad” RMI taxa. Portions were calculated on the basis of the frequency of the taxa.

3

Taxa from the group A and AB were taken as “good” and taxa from the group C and BC as “bad”. Boundary value between high and good ecological status was determined where so called “bad” taxa started to appear. Boundary value between good and moderate status was determined where there was a similar portion of “good” and “bad” taxa. Boundary value between moderate and poor ecological status was determined where the portion of “bad” taxa started to exceed the portion of “good” taxa, and boundary value between poor and bad status where “good” taxa do not appear anymore.

Spain – The H/G boundary has been identified by assessing IBMR variability range in references sites for each river type, the value of the 25th percentile of references was used. The remaining gradient was divided by the other four classes.

Portugal – The H/G boundary has been identified by assessing IBMR variability range in references sites for each river type, the value of the 25th percentile of references was used. The remaining gradient was divided by the other four classes.

3. All relevant parameters indicative of the biological quality element are covered (see Table 1 in the IC Guidance). A combination rule to combine parameter assessment into BQE assessment has to be defined. If parameters are missing, Member States need to demonstrate that the method is sufficiently indicative of the status of the QE as a whole.

Cyprus - taxonomic composition and relative abundance; vascular plants, bryophytes and macroalgae (only genus level)France - taxonomic composition and relative abundance; aquatic vascular plants, macroalgae (most at genus level) and bryophytesGreece - taxonomic composition and relative abundance; vascular plants, bryophytes and macroalgae (only genus level)Italy - taxonomic composition and relative abundance; aquatic vascular plants, macroalgae and bryophytes.Portugal – taxonomic composition and relative abundance; vascular plants, bryophytes and macroalgae (the latter not exhaustively surveyed).Slovenia - taxonomic composition and relative abundance; mainly vascular plants, bryophytes and macroalgae were not exhaustively surveyed.Spain - taxonomic composition and relative abundance; aquatic vascular plants, macroalgae (most at genus level) and bryophytes.

Concerning possible forms of

4

combination between phytobenthos and macrophytes, this was not addressed in the present exercise.

4. Assessment is adapted to intercalibration common types that are defined in line with the typological requirements of the WFD Annex II and approved by WG ECOSTAT

Yes, all assessment methods are compatible with intercalibration common types, though further testing is expected for temporary river systems.The common typology has been adjusted according to data analysis (see below).

5. The water body is assessed against type-specific near-natural reference conditions

Yes, for all participating countries.

6. Assessment results are expressed as EQRs Yes – all methods express assessment results as EQRs.

7. Sampling procedure allows for representative information about water body quality/ ecological status in space and time

Sampling procedures of macrophytes allow representative information of annual cycle. Data concerns one survey per site usually in spring-summer season.Cyprus - one survey from March to May, according to the weather of the sampling year (the driest the year the earlier the survey month)France - one survey from May to July.Greece - one survey in summer season (June-July)Italy - two surveys: one from April to September; the second sample in October and November. Portugal- one survey from May to August, according to the natural yearly cycle of the river (the further south and higher altitude the site, the earlier the survey).Spain – one survey from July to September.Slovenia - one survey from July to September

8. All data relevant for assessing the biological parameters specified in the WFD’s normative definitions are covered by the sampling procedure

Both relevant parameters of BQE are covered (i.e. species composition and abundance). Bryophytes and macroalgae were only diagnostically assessed in some cases. Sampling is performed at the peak growth period thus insuring a good representative image of the site.

9. Selected taxonomic level achieves adequate confidence and precision in classification

Species level is required for most of taxa, which is considered adequate to achieve confidence and precision in classification. Genus level is accepted only for macroalgae due to a limited existence of taxonomic keys in many cases.

5

3. IC feasibility checking

Typology

Table 3 – Overview of common intercalibration types in the Mediterranean rivers GIG and MS sharing the types

The same broad types were used as in the first intercalibration round for the other BQEs, but the borders of the types were redefined for a better adjustment to the ecological reality. The redefinition was based on ordination data treatment, and presented, discussed and agreed in the General Mediterranean GIG meetings.

The biological analysis of macrophytes revealed a poor segregation between types RM1, RM2 and RM4, both for the reference sites (Global R of ANOSIM =0,263; significance level=0,1%) and for the all sites (Global R of ANOSIM =0,218; significance level=0,1%). Therefore these types were treated together throughout the IC process.

Mediterranean temporary rivers revealed large structural and functional differences [Aguiar FC, Cambra J, Chauvin C, Ferreira T, Germ M, Kuhar U, Manolaki P, Minciardi MR, Papastergiadou E. 2010. Floristic and functional gradients of river plant communities: a bio geographical study across the Mediterranean Basin. XV Congress of the Iberian Association of Limnology, 5-7 July 2010, Azores: 55] and will be subject to a separate data treatment. We compared the floristic composition and the functional guilds using the RM5 dataset and the permanent rivers (RM1, RM2 and RM4), and used the classification for reference and impaired sites. Five countries provide data for RM5 and seven countries for the other types. We found a consistent lower species diversity and cover of aquatic macrophytes in temporary rivers than in permanent rivers, accompanied by rather similar species richness across the Mediterranean Basin (around ten species per site). With increasing flow intermittency, a significant increasing number of hygrophyte species, and lower ratio of hydrophytes/helophytes was observed. The number of pteridophytes, mosses and liverworts (including supra-aquatic bryophytes) species is similar between these two river types, and these patterns remain the same in reference and impaired sites, albeit the differences resulting from pressures.

For the three methods included in the intercalibration, there are sufficient possibilities for intercalibration; the RMI method is appropriate for types R-M1, R-M2 and R-M5, the IBMR is appropriate for all types except R-M3 and R-M5 (for Cyprus), and MMI is appropriate for type R-M5. It was concluded that intercalibration is feasible for all common types except R-M3..

Pressures

Table 4 below lists the pressures addressed by the national methods.

Table 4 – Overview of the sensitivity to pressures of the national methodsMethod PressureRMI (Slovenia) Catchment land use, eutrophicationIBMR (Cyprus, Greece, France, Italy, Portugal, Spain)

Cyprus –hydrological, morphological, land-use and physico-chemical degradation.Greece - hydrological, morphological, land use and physico-chemical degradation.France - eutrophication, general degradation, hydromorphological

6

Common IC type Type characteristics MS sharing IC common type RM1 catchment <100 km2; mixed

geology (except non-siliceous); highly seasonal

France, Italy, Portugal, Slovenia, Spain

RM2 catchment 100-1000 km2 ; mixed geology (except non-siliceous); highly seasonal

France, Greece, Italy, Portugal, Slovenia, Spain

RM3 catchment 1000-10000 km2 ; mixed geology (except

siliceous); highly seasonal

Greece, Portugal, Spain

This type cannot be intercalibrated because assessment methods were not fully developed yet.

RM4 non-siliceous streams; highly seasonal

Cyprus, France, Greece, Italy, Spain

RM5 temporary rivers Cyprus, Italy, Portugal, Slovenia, Spain

degradation;Italy –eutrophication, general degradation, pollution by organic matter;Portugal – eutrophication, general degradation.Spain - eutrophication, general degradation.

MMI (Cyprus) Cyprus – hydrological, morphological, land use and physico-chemical degradation.

Additionally, the Mediterranean GIG database was used to test the response of the each metric to: individual types of pressures and to a general degradation gradient (obtained from PCA axes scores). The existence of many gaps in the quantitative data prejudiced the analysis of relations to pressure, especially concerning physico-chemical variables. For RM 1,2,4 type, the PCA based on pressure data of morphological variables and land-use of the MedGIG showed that the PCA axis 1 explain 56.4% of the total variation (100% on the first three axes), and the Pearson correlation coefficient is 0.55 (p<0.000001). Responses to individual gradients are illustrated in Figure 1.

Figure 1 Response of assessment methods (IC types RM1,2,4) to individual pressures

For RM5, the PCA based on pressure data of morphological variables, hydrological variables and pH showed that the PCA axis 1 explain 57.7% of the variation, the Pearson correlation coefficient is 0.58 (p<0.000001). Concerning individual pressures the R (p<0.0001), for “upstream dams” = -0.63, R “water abstraction” =-0.54; Stream hydrology =0.44, habitat alteration= -0.38 were the highest, highlighting the importance of flow alterations and physical disturbances in these rivers.

It was concluded that intercalibration is feasible in terms of pressures.

Assessment ConceptAll methods are based on multihabitat sampling covering the whole channel and inner banks. RMI and IBMR are indicator species based methods, although for RMI the indicator species are mainly spermatophytes and pteridophytes. The indicator value of the species (score) is derived from presence and abundance of species that showed a correlation with pressure gradient. The MMI has a different concept (based on functional plant groups), though sampling has the same conceptual basis, and the index can be applied to other countries.In spite of these differences, the GIG considers the intercalibration feasible in terms of assessment concepts.

Conclusions feasibility check

It was concluded that intercalibration is feasible for the common types R-M1, R-M2, R-M4 and R-M5. However, due to a lack of data availability type R-M5 had to be excluded at this stage (see below).

1 2 3 4

habita t a ltera tion

0 ,0

0 ,4

0 ,8

1 ,2

1 ,6

1 2 3 4

urbanisation

0 ,0

0 ,4

0 ,8

1 ,2

1 ,6

1 2 3 4

channelisation

0,0

0 ,4

0 ,8

1 ,2

1 ,6

assessment m

ethod

0,0 22,5 45,0 67,5 90,0

extensive agriculture (% )

0 ,0

0 ,4

0 ,8

1 ,2

1 ,6

asse

ssm

ent m

etho

d

3,5 27,6 51,8 75,9 100,0

sem i-natural areas (% )

0,0

0,4

0,8

1,2

1,6

asse

ssm

ent m

etho

d

1 2 3 4

agriculture

0,0

0 ,4

0 ,8

1 ,2

1 ,6

assessment m

ethod

7

4. IC option applied

A combination of Option 1 and Option 3 was applied.For R-M1, 2, 4 we performed the IC Option 1 for six countries of the Mediterranean rivers GIG macrophytes (Cyprus, France, Greece, Italy, Portugal, Spain) that use the same assessment method (IBMR, Biological Macrophyte Index for Rivers), same data acquisition and same numerical evaluation. Then the IC Option 3 - similar data acquisition, but different numerical evaluation (BQE sampling and data processing generally similar, so that all national assessment methods can reasonably be applied to the data of other countries) was performed for the national method of Slovenia (River Macrophyte Index), using the fixed Median value of the H/G boundary of the countries intercalibrated previously. This alternative was followed after discarding the sites where RMI calculations were questionable, due to low number of indicator taxa, which specially occurred for the database of France, Cyprus and Spain. The regression of RMI and the IBMR was made using sites from Greece, Portugal and Italy (n=103).The correlation coefficient of the RMI and the IBMR was r=0.6770; p=0.000001.

For the type RM5:Italy, Cyprus, Spain, Slovenia, and Portugal share this intercalibration type. We considered that only Portugal, Cyprus and Italy had a representative number of sites to intercalibrate (Spain and Slovenia have only 4 sites). When Option 3 was performed, the r value of MMI (Cyprus national method) with the other national methods (IBMR) was very low (see table below) and did not allow the comparison. The bad correlation between MMI and IBMR was expected at least in the Cyprus case, as the IBMR had been shown not applicable to Cyprus R-M5 using CY data. Member State/Method R pCyprus/ MMI vs. IBMR (IT, CY, PT) 0.045 0.6824Cyprus/ MMI vs. IBMR (CY, PT) 0.001 0.9376Cyprus/ MMI vs. IBMR (IT, CY) 0.071 0.6008

We proceed with the Option 1 to intercalibrate. The intercalibration procedures described for Option 1 of RM1,2,4 were applied for Italy and Portugal, and results are presented in Annex 2, however it was agreed during the 7th Mediterranean meeting (Madrid, 10-11 October) that the intercalibration results were not reliable due the small number of sites of the database (30 for Portugal and 13 for Italy) and the low number of reference sites (3 reference sites, all from Portugal). It was concluded that intercalibration of RM5 is not feasible for the moment, however it is the Med GIG opinion that macrophytes are a reliable BQE to be used in the assessment of the ecological status for the WFD. See Annex 2.6 for further explanations.

5. IC dataset collected and benchmarking

Data set

Data from 479 samples were used; an overview split up by MS and common type is provided in Table 5 below.

Table 5 – Overview of biological, physicochemical and pressure data compiled for the Mediterranean Macrophytes intercalibrationIC type Member

StateBiological data and classification

Physico- chemical data

Other pressure data

RM1 France 12 10 12

Italy 25 25 25

Portugal 30 30 30

Slovenia 6 6 6

Spain 10 10 10

Sum 83 81 83

RM2 Greece 31 34 35

Italy 9 9 9

Portugal 30 30 30

Slovenia 15 9 15

Spain 10 11 11

8

Sum 95 93 100

RM3 Greece 7 7 7

Portugal 30 30 30

Spain 20 20 20

Sum 57 57 57

RM4 Cyprus 14 14 14

France 25 23 25

Greece 1 1 1

Italy 36 36 36

Spain 64 64 64

Sum 140 138 140

RM5 Cyprus 43 43 43

Italy 13 13 13

Portugal 30 30 30

Slovenia 4 4 4

Spain 4 9 9

Sum 94 99 99

Data sets were checked against data acceptance criteria, with no problems identified (see table 6) and Figure 2).

Table 8 – Data acceptance criteria

Data acceptance criteria Data acceptance checking

Data requirements (obligatory and optional)

All countries delivered geographical location and river typology, as well as biological and environmental data, and also pressure data. The data was organized under the same agreed frame. Checking included the assessment of the “mediterranicity” of the sites, that is, the degree of water availability during summer.

Reference dataset for most of countries is reasonably complete in pressure and environmental data.

There were however gaps in the database, notably in physico-chemical data. Also for some sites in most countries, the physico-chemical data is qualitative or if quantitative, the value is referred just as below or above the LD; gaps on pressures vary in the type of parameter between countries and within countries.

The sampling and analytical methodology

Sampling methods are similar for most countries, however in some surveys, the inner bank have not be included fully, as well as

9

bryophytes and macroalgae.

Level of taxonomic precision required and taxalists with codes; harmonisation procedure

All vascular plants were identified at the species level. However, the number of species related to water is very small when compared to helophytes and other waterlogged species, depending on the country and on the temporality of the fluvial systems.

All the biological information and lists of species provided by the MS had to be compared, harmonized and the final result agreed by all MS in the GIG meetings.

The procedure was as followed:

All MS delivered taxa lists with similar taxonomic precision (species level for most of the taxa), codes and synonymy were defined at the coordination level. Preparatory steps included the harmonization of the database, especially concerning problems in synonym, and the harmonization of the cover scale (three different cover scales for the seven countries). All countries assigned a floristic group for each species (algae, bryophyte moss, bryophyte hepatic, pteridophyte, spermatophyte, lichen, heterotrophic organisms -fungi & bacteria- and attributed an “aquaticity” level for all taxa recorded [C. Chauvin in Birk S., N Willby, C Chauvin, HC Coops, L. Denis, D. Galoux, A. Kolada, K. Pall, I. Pardo, R. Pot, D. Stelzer. 2007. Report on the Central Baltic River GIG Macrophyte Intercalibration Exercise, June 2007].

MedGIG macrophyte coordination harmonized the classification done by MS for common species. For data treatment only species with aquaticity level 5 or lower were accepted, in order to achieve comparable biological information. From the 736 species gathered in the original MedGIG, around 40% were eliminated (woody riparian species, brackish water or salty marsh species, terrestrial ruderals and grasses). Species codes were attributed using the CEMAGREF codes in order to be included in the CEN database. New codes were assigned and suggested; the information was delivered to CEMAGREF-Bordeaux.

The minimum number of sites / samples per intercalibration type

Minimum of 15 sites per IC type are available.

Sufficient covering of all relevant quality classes per type

All 5 quality classes are represented – see Figure 2. However, the database included few sites classified as poor and bad quality classes, especially for the RM1,2,4 type.

10

Figure 2 - Distribution of MedGIG macrophyte database samples by quality class (5 classes) taking into account national EQR values per type. RM1, RM2, RM4 are shown together

National reference sites exist in the common database for all IC types, except for RM3 (Table 9); the reference sites

were those non or minimally disturbed found in the Mediterranean region. The common reference conditions approach

was used for the intercalibration exercise.

Table 9. Total number of national reference sites provided by Member States:Member State RM1 RM2 RM4 RM5 TotalCyprus - - 3 8 11France 10 - 17  - 27Greece - 13 1 - 14Italy 5  0 12 0 17Portugal 10 10 - 9 29Slovenia 0 0 - 0 0Spain 10 1 28 0 39Sum 35 24 61 17 137

Benchmarking

The Mediterranean GIG has developed common benchmark conditions. A proposal for common benchmark conditions thresholds was brought to the General GIG meetings, and intensively discussed till a final collective decision was made. The same and common thresholds were used for the BQEs macroinvertebrates, phytobenthos and macrophytes (see Table 9). The selection of IC benchmark sites was done through a 3 steps procedure. Steps 1 and 2, up to the establishment of thresholds, were performed with the information for MS reference sites provided to the MedGIG database. We used sites supplied to the MedGIG for invertebrates, diatoms and macrophytes databases. The global database was composed of a total of 919 member states reference sites distributed through the 4 IC river types (RM1, RM2, RM4, RM5) and 7 MS (CY, FR, GR, IT, PT, SI, SP).Step 1. Original national reference sites are selected only if all their categorical variables have class 1, no impact or minimal impact.Step 2. Benchmark thresholds are calculated for numerical pressure variables using the common database, based on sites selected in Step 1 and for each IC type. Extreme values for each pressure variable and IC type were previously excluded after histograms and boxplots inspection. These observed ranges characterize the pressure levels existent in the minimally impacted sites, for each IC type, in the Mediterranean region. A unique value for each pressure variable (benchmark thresholds) was afterwards calculated for all IC types, corresponding to the maximum pressure acceptable overall Mediterranean types, in order to reach a common tolerance level. However, for RM5, the temporary rivers, different ranges for water oxygenation were established for low water periods, assuming a higher organism tolerance to harsher thermal conditions.The following table 10 describes the thresholds established and applied for each pressure variable and used by the MEDGIG macrophytes for IC. The values are common to all IC types except for O2, with a different limit for RM5.

Table 10 – Criteria for identifying benchmarking sites for the MED GIGPressure variables References are accepted if

11

RM1, RM2 and RM4 RM5General Morphology (Classes 1-3)General Hydrology (Classes 1-3) ≤ 2Riparian Vegetation (Classes 1-3)DO (mg/L) 1 6,39-13,70O2 (%) 1 73,72-127,92 60,34-127,92N-NH4+ (mg/L) ≤0,09N-NO3

- (mg/L) ≤1,15P-Total (mg/L) ≤0,07P-PO4

3- (mg/L) ≤0,06% Artificial areas (catchm) ≤1% Intensive agriculture (catchm) ≤11% Extensive agriculture (catchm) ≤32% Semi-natural areas (catchm) ≥68% Urbanisation (reach) 2 ≤1% Land use (reach) 2 ≤20% Agriculture (reach) 2 ≤201 for macrophytes only, instead of O2 (%)2 for diatoms only, instead of land use in the catchment

Step 3. Final abiotic screening of benchmarks. Potential benchmarks, left out in Step I, are rescreened and

those with categorical variables in class 2 (any number) but simultaneously with numerical variables values

within the thresholds defined in Step 2 are chosen and added to the reference set of Step 1.

Finally, in order to recover surveys for the macrophyte reference database, the entire database was

rescreened for all MS and all types. If samples passed the thresholds for an IC reference they were also

included in the set of benchmarks.

Table 11 - Number of IC benchmark sites selected abiotically using the described procedure, for intercalibration for

macrophytes from entire macrophyte data base and percentage of sites retained in relation to the total reference sites

provided by the MS.

Member State RM1 RM2 RM4 RM5 Total

Cyprus - - - - 3 100% 5 63% 8 73%France 6 60% - - 10 59% - - 16 59%Greece - - 9 69% 1 100% - - 10 71%Italy 4 80% - - 11 92% - - 15 88%Portugal 8 80% 5 50% - - 3 33% 16 55%Slovenia 0 - 0 - - - 0 - - -Spain 5 50% 0 0% 16 57% 0 0% 21 54%Sum 23 66% 14 58% 41 67% 8 47% 86 63%

River types RM1, RM2 and RM4 will be combined for the intercalibration, and a sufficient number of reference sites

are available to make a statistically reliable estimate (Table 11). For RM5 the number of available benchmark sites may

be insufficient (8 sites for the overall database, Table 11).

In order to be sure that the benchmark sites selected were not influenced by pressures we ran Spearman rank

correlations between pressure variables and EQR values for benchmark sites, assuming that these correlations should be

low. Further screening data proceed. The high percentage of observations that indicates no or low alterations in

12

morphological and hydrological pressures, and low values for eutrophication variables (Spearman rho <0.18, p<0.001)

indicates that reference sites are true and were well selected (see illustrative examples for types RM 1,2,4 in Fig. 3).

13

Fig. 3. Number of observations (sites) of types IC RM1,2,4 along the individual pressure gradients.

For RM5 the analyses are not statistically steadfast for the MedGIG benchmark dataset. Fig. 4 illustrates the

number of benchmark sites in each category of pressure variables.

14

Fig. 4. Number of observations (sites) of types IC RM5 along the individual pressure gradients.

When carrying out the intercalibration calculations, EQR site values were obtained by dividing the absolute values of national assessment method (IBMR) by the median of the IBMR values of MedGIG benchmark sites (previously screened with MedGIG criteria; see before). Since there are no benchmark sites for Slovenia in the MedGIG macrophytes database, we used the median of benchmark sites of the MedGIG dataset (only used for IC of Slovenia).

Biological communities at the reference sitesIn the reference sites of RM1, 2 and 4 a low cover of alien species (Arundo donax, Azolla sp., Bidens sp., Paspalum distichum) and a high diversity of bryophytes was observed. The cover of Rhynchostegium riparioides, Fontinalis antipyretica, Fontinalis squamosa, Fissidens crassipes are higher in the reference sites when compared with the borderline communities of G/M classes. Some species were not observed in these sites, or had a very low frequency of occurrence and/or abundance. This is the case for instance of Potamogeton pectinatus, Iris peseudacorus, Enteromorpha sp. and Schoenoplectus lacustris. For RM5, the low number of benchmark sites did not allow analyses of the macrophyte communities.

15

6. Comparison of methods and boundaries

Procedures for the IC of types RM1,2,4

We performed the following steps for translating the national boundaries, defining a harmonisation guideline, evaluating the level of boundary bias, harmonising class boundaries and analysing the class agreement.

1. Boundary harmonization using Option 1: all countries except SI, IC types RM1, 2, 4. 1.1 We calculated the EQR site values for Mediterranean GIG (MedGIG EQR), using the IBMR absolute

values divided by the median of the IBMR values of MedGIG benchmark sites. 1.2 We converted boundaries to IC EQR using regression analyses per MS: MedGIG EQR vs. National EQR.1.3 The boundary bias criterion was used: boundary bias should be less than a quarter of the width of a class.

We used the median of the boundaries to compute the boundary bias. We changed iteratively the values of the boundaries that did not meet the criteria until the median of the boundaries was included within the quarter of the class (High or Good).

1.4 We converted all harmonized boundaries at the scale of IC EQR IBMR into the absolute IBMR values (product between IC EQR and the median value of IBMR at MedGIG benchmark sites).

2. Boundary harmonization using option 3 2.1 Whenever possible, RMI (Slovenia National Method) was computed for the entire database. The IBMR

was also computed for the Slovenian sites.2.2 The IBMR was regressed against RMI in order to check the relatedness of indices and to convert RMI

boundaries into the IBMR scale. Only sites from Greece, Italy, Portugal and Slovenia were used in the regression (Spain, Cyprus and France were not considered because the RMI was computed for a very small number of sites).

2.3 The boundary bias of RMI (after conversion to the IC EQR IBMR scale using the regression equation) was computed using the median values after Option 1 boundary harmonization. In case the RMI boundaries did not meet the criteria (boundary bias lower than a quarter of a class), these were moved until the median of the boundaries was included within a quarter of a class.

2.4 We converted the harmonized boundaries at the IC EQR IBMR scale into RMI values using the inverse function of the regression equation.

3. Class agreement was computed for all pair combination of national methods. Only sites that were classified as H, G or M according to both national methods were used to compute class agreement. First, a piecewise transformation of the IC EQR IBMR values was performed using the formula:

MinT – ((X – Min)*0.2) / (Max – Min)

where MinT – Minimum of the new transformed class (0.6 for G and 0.8 for H), X – index value, Min –theoretical index minimum, Max – theoretical index maximum.

Class agreement was computed as the mean absolute difference between the index values after piecewise transformation divided by 0.2 (the width of each class after piecewise transformation). Class agreement should be less than 1 (meaning that the mean differences should be less than the width of 1 class). The percentage of classifications differing by half a quality class was also calculated.

Results IC Types RM 1, RM 2 and RM4

Following the Annex V, the result of H/G and G/M boundaries comparison is included in tables and illustrated in the figures of the Annex. Tables 12-13 illustrate the calculations necessary for boundary bias checking and the initial result. The figures show the location of MS boundaries ± quarter of band width, and the median of the Mediterranean GIG boundary; abbreviations: CY –Cyprus; FR – France; GR – Greece; IT – Italy; SP –Spain; PT – Portugal; SI –Slovenia; MedGIG – Mediterranean GIG. IBMR – Biological Macrophyte Index for Rivers. RMI – River Macrophyte Index.

Table 12. Regression analyses per MS used to convert the national boundaries into comparable boundaries (step 2)MS Regression equations

CY y = 0.5275x + 0.4837FR y = 0.7939x + 0.1794GR y = 0.6316x + 0.5124

16

IT y = 0.9889x - 0.0395PT y = 1.0833x - 4E-05SP y = 0.9614x - 0.0056

Table 13. IC results Option 1 with original boundaries for H/G and G/M boundary (values of original boundaries were regressed using MedGIG reference data values for each MS). Max - maximum of national EQR.

CY FR GR IT PT SP

Max 1.011 1.393 1.144 1.226 1.443 1.386MedGIG_H/G 0.903 0.918 0.986 0.851 0.997 0.908MedGIG_G/M 0.798 0.807 0.866 0.752 0.747 0.677MedGIG_M/P 0.693 0.703 0.746 0.603 0.498 0.446MedGIG_P/B 0.588 0.592 0.632 0.455 0.249 0.225

H width to Max 0.11 0.48 0.16 0.38 0.45 0.48G width 0.10 0.11 0.12 0.10 0.25 0.23M width 0.10 0.10 0.12 0.15 0.25 0.23H/G bias -0.01 0.00 0.07 -0.06 0.08 0.00G/M bias 0.02 0.03 0.09 -0.02 -0.03 -0.10

H/G bias_CW -0.09 0.04 0.61 -0.17 0.34 -0.01G/M bias_CW 0.22 0.31 0.76 -0.23 -0.11 -0.42

Table 14. Median value of the MedGIG MS using IBMR as national method for H/G and G/M boundaries Median MedGIG

H/G 0.9127

G/M 0.7749

17

Fig. 5 Illustrative results of boundary comparison a) H/G original boundaries; b) G/M original boundaries (converted using MedGIG benchmark data)

Table 15. IC results Option 1 with harmonized boundaries for H/G and G/M boundary (values of original boundaries were regressed using MedGIG reference data values for each MS). Max - maximum of national EQR.

CY FR GR IT PT SPMax 1.011 1.393 1.144 1.226 1.443 1.386MedGIG_H/G 0.903 0.918 0.928 0.851 0.967 0.908MedGIG_G/M 0.774 0.771 0.757 0.752 0.747 0.709MedGIG_M/P 0.693 0.703 0.746 0.603 0.498 0.446MedGIG_P/B 0.588 0.592 0.632 0.455 0.249 0.225

H width to Max 0.108 0.476 0.216 0.376 0.476 0.478G width 0.129 0.147 0.171 0.099 0.220 0.199M width 0.081 0.068 0.011 0.148 0.249 0.263H/G bias -0.010 0.005 0.015 -0.062 0.054 -0.005G/M bias 0.020 0.017 0.003 -0.003 -0.007 -0.045

18

H/G bias_CW -0.09 0.03 0.09 -0.17 0.25 -0.01G/M bias_CW 0.24 0.25 0.25 -0.03 -0.03 -0.23

Table 16. Median value of the MedGIG MS using IBMR as national method for H/G and G/M boundaries

19

Median MedGIGH/G 0.9127

G/M 0.7543

Fig. 6 Illustrative results of boundary comparison Option 1 a) H/G harmonized Boundaries; b) G/M harmonized Boundaries (converted using MedGIG benchmark data)

20

Table 17 Conversion to the national boundary values (using the regression equations). CY: x=( y-0.4837)/0.5275); FR: x = (y-0.1794)/0.7939; GR: x = (y-0.5124)/0.6316; PT: x= (y+ 4E-05)/1.0833; SP: x=(y+0.0056)/0.9614.

CY FR GR IT PT SPHarmonized boundaries (MedGIG scale)

H/G 0.9031 0.9177 0.9861 0.8505 0.9966 0.9077G/M 0.7981 0.8066 0.8661 0.7516 0.7474 0.6770

Harmonized boundaries (national scale)

H/G 0.795 0.93 066 0.90 0.89 0.95G/M 0.550 0.75 0.39 0.80 0.69 0.74

Boundary comparison and harmonization is summarized in Table **, as well as a summary of results. For the intercalibration of national method of Slovenia for the BQE macrophytes (River Macrophyte Index, RMI), Option 3 was used. Figure 7 shows the regression of EQR values of IBMR calculated with MedGIG reference data (EQR_IBMR_MedGIG) and EQR values of RMI calculated with MedGIG reference data (EQR_RMI_MedGIG). Sites values used are from Greece, Italy, Portugal and Slovenia; n=103.

Fig. 7 Scatterplot showing the regression of EQR values of IBMR calculated with MedGIG reference data (EQR_IBMR_MedGIG) and EQR values of RMI calculated with MedGIG reference data (EQR_RMI_MedGIG).

Table 18 Boundary values used for boundary harmonization; original national boundary values are in Anex 1. RMI_BS (benchmark standardized - RMI boundary values calculated with the median values of the MedGIG reference data. RMI_reg IBMR are the boundary values converted to the IBMR scale)

 RMI RMI_BS RMI_reg IBMR

Max 1.1925H/G 1.4454 0.8814G/M 0.9412 0.7363M/P 0.7059 0.5910

f(x) = 0.62 x + 0.3R² = 0.458192071109817

Regress_EQR_IBMR MedGIGLinear (Regress_EQR_IBMR MedGIG)

EQR_RMI_MedGIG

EQ

R_I

BM

R_M

edG

IG

21

P/B 0.4706 0.4459Min 0.2353 0.3007

Table 19 IC results Option 3 with harmonized boundaries for H/G and G/M boundary (values of original boundaries were regressed using MedGIG reference data values for each MS). ax - maximum of national EQR. The harmonized median value H/G and G/M (see table 17) was fixed to intercalibrate RMI.

22

 SI

Max 1.1925MedGIG_H/G 0.8814MedGIG_G/M 0.7363MedGIG_M/P 0.5910MedGIG_P/B 0.4459

H width to Max 0.311G width 0.145M width 0.145H/G bias -0.031G/M bias -0.018

H/G bias_CW -0.101G/M bias_CW -0.125

Fig. 8.4 Illustrative results of boundary comparison Option1. 8.4a) H/G harmonized boundaries; 8.14) G/M harmonized Boundaries (converted using MedGIG benchmark data)

Some national methods did not comply with the boundary bias (see Figure 5). Boundary bias is exceeded by the methods of:

Greece: H/G boundary (too stringent) Portugal: H/G boundary (too stringent) Cyprus: G/M boundary (too stringent) Greece: G/M boundary (too stringent) France: G/M boundary (too stringent) Spain: G/M boundary (too relaxed)

The boundary bias criterion was used: boundary bias should be less than a quarter of the width of a class. We used the median of the boundaries to compute the boundary bias. We changed iteratively the values of the boundaries that did not meet the criteria until the median of the boundaries was included within the quarter of the class (High or Good).

The required boundary adjustments are specified in Table 20. Spain agreed to raise the H/G boundary to comply with the comparability criteria. According to the harmonization results Cyprus, Greece, Portugal and France could lower the national boundaries if they wish to do so, but all except France preferred to maintain the original national values before the harmonization. The average absolute class difference was calculated after boundary adjustment (Table 21, Annex B); the percentage of agreement after harmonization is 74%, and 100% of classification values are within one class. The percentage of classifications differing by half of a quality class is small (16.37%).

23

Table 20 National class boundaries and boundaries bias adjustment (adjusted/harmonized boundaries if bias >|0.25|). Proposed adjustments: ↑ boundary to be raised, ↓ boundary should be lowered. Final accepted boundaries for MS for types RM1,2,4

RM1,2,4 Original Harmonized Final agreed national boundaries

H/G G/M H/G G/M H/G G/M

CyprusBoundary 0.795 0.596 0.550 ↓ 0.795 0.596

Bias CW -0.09 0.22 0.24

FranceBoundary 0.93 0.79 0.745 ↓ 0.93 0.745

Bias CW 0.04 0.31 0.25

GreeceBoundary 0.75 0.56 0.66 ↓ 0.39 ↓ 0.75 0.56

Bias CW 0.61 0.76 0.09 0.25

ItalyBoundary 0.90 0.80 0.90 0.80

Bias CW -0.17 -0.23

PortugalBoundary 0.92 0.69 0.89 ↓ 0.92 0.69

Bias CW 0.34 -0.11 0.25

SpainBoundary 0.95 0.71 0.74 ↑ 0.95 0.74

Bias CW -0.01 -0.42 -0.23

SloveniaBoundary 0.80 0.60 0.80 0.60

Bias CW -0.101 -0.125

Table 21 Average absolute class differences (AACD) of national methods after boundary harmonisation and percentage of classifications differing by half of a quality class (PCHC) (0.1 after piecewise transformation).

MS AACD PCHC

Cyprus 0.27 20.18France 0.23 14.01Greece 0.34 23.74Italy 0.26 17.57Portugal 0.25 13.17Spain 0.24 12.81Slovenia 0.22 13.09Average 0.26 16.37

Correspondence common types vs national types The harmonized boundaries were converted to the IC EQR IBMR original boundaries using the regression equation obtained. Final boundaries were computed for the MS that have boundary adjustments to comply with the harmonization criteria.

At the MS level, the harmonized values will be adapted to the original national assessment methods.

24

Remaining gaps in the current intercalibration: what still needs to be doneAll goals for intercalibration work programme 2008-2011 dated from 24th September 2008, were achieved, namely translation of IC results into national systems, refinement of criteria for setting reference conditions between MS, intercalibration of national methods, intercalibration of most relevant pressures, evaluation of the relevance of macrophytes as an ecological indicator in Mediterranean rivers in temporary rivers.

However, large rivers (RM3) and temporary rivers (RM5) were not intercalibrated. See sections 4.1 and 7.1, respectively. The combination of phytobenthos and macrophytes was not addressed in the present exercise.

7. Ecological characteristics of reference conditions and/or status classes

RM1, RM2 & RM4In these rivers types, the major changes in the floristic communities between the good and the moderate ecological status are associated with the decrease of species richness (median in G sites =12; M sites=7), and an increase in cover and frequency of pondweed taxa, such as Potamogeton pectinatus and P. nodosus, macroalgae (e.g. Enteromorpha sp., Cladophora sp.), and other hydrophyes (for instance Lemna gibba), and of some emergent species, such as Schoenoplectus lacustris. In opposition, there is a loss and/or decrease in cover of bryophytes (both mosses and liverworts), mainly Rhynchostegium riparioides, Fontinalis antipyretica, Fissidens crassipes, Eurhynchium praelongum, Lunularia cruciata and Amblistegium riparium, and of some amphibious and hygrophyte species (Lotus pedunculatus, Carex elata, Carex pendula).Some infrequent species in the data base, bryophytes (Bryum sp.), isoetids, Juncus sp., Myosotis sp. were only observed in sites classified in Good ecological status. Some alien invasive species also raise their abundance due the raise of pressures; this is the case of Azolla sp. RM5It was not possible to have a reliable description of the IC type-specific biological communities due to the small dataset.

25

ANNEX A – Descriptions of National assessment systems for macroinvertebrates

MS National method Reference conditions setting at National level

Cyprus Biological Macrophyte Index for Rivers (IBMR)

Described in:

Haury, J., Peltre M.-C., Trémolières M., Barbe J., Thiebaut, G., Bernez, I., Daniel, H., Chatenet, P., Haan-Archipof, P., Mulller, S., Dutartre, A., Laplace-Treyture, C., Cazaubon, A. & Lambert-Servien, E. (2006) A new method for assess water trophy and organic pollution – The Macrophyte Biological Index for Rivers (IBMR) : its application to different types of rivers and pollution. Hydrobiologia 570: 153-158.

Web page describing the national method: https://hydrobio-dce.cemagref.fr/

Metrics:

K = abundance (translated in 5 classes), CS = trophic score (0-20), E = stenoecy coefficient (1-3)

IBMR = Σ(K.CS.E)/Σ(K.E)

A pressure gradient based on hydromorphological variables and land use was obtained from seven variables: Channel profile (cross section alteration), Channel morphological alteration, Stream hydrology, Dam influence, Water abstraction, Corine landcover/land use, Urban and industrial areas in the immediate vicinity of sites. Multivariate PCA was used to determine the national reference sites.

Cyprus Multimetric Macrophyte Index (MMI)

Metrics: Number of nitrophyllous species, log (Number of Gramineae), Number of Helophytes_herb, species richness, mean cover of green algae (categorical)

Information from:Papastergiadou E., P. Manolaki. 2011. MedGIG Report: R-M5 IC river types of Cyprus. Developing an Assessment system of R-M5 river types for Cyprus. Patras University, Greece, 21 pp. (not published)

France Biological Macrophyte Index for Rivers (IBMR)

Described in:

Haury, J., Peltre M.-C., Trémolières M., Barbe J., Thiebaut, G., Bernez, I., Daniel, H., Chatenet, P., Haan-Archipof, P., Mulller, S., Dutartre, A., Laplace-Treyture, C., Cazaubon, A. & Lambert-Servien, E. (2006) A new method for assess water trophy and organic pollution – The Macrophyte Biological Index for Rivers (IBMR) : its application to different types of rivers and pollution. Hydrobiologia 570: 153-158.

Web page describing the national method: https://hydrobio-dce.cemagref.fr/

Metrics:

K = abundance (translated in 5 classes), CS = trophic score (0-20), E = stenoecy coefficient (1-3)

IBMR = Σ(K.CS.E)/Σ(K.E)

Based on existing near-natural reference sites, expert knowledge, Least Disturbed Conditions.

1) base: national reference network (first reference conditions guidance criteria)2) qualitative criteria's list at the basin, reach, site scale evaluated by local experts.3) GIS criteria based on Corine Land Cover data (artificial, intensive agriculture, agriculture in the watershed).

MS National method Reference conditions setting at National level

Greece Biological Macrophyte Index for Rivers (IBMR)

Described in:

Haury, J., Peltre M.-C., Trémolières M., Barbe J., Thiebaut, G., Bernez, I., Daniel, H., Chatenet, P., Haan-Archipof, P., Mulller, S., Dutartre, A., Laplace-Treyture, C., Cazaubon, A. & Lambert-Servien, E. (2006) A new method for assess water trophy and organic pollution – The Macrophyte Biological Index for Rivers (IBMR) : its application to different types of rivers and pollution. Hydrobiologia 570: 153-158.

Web page describing the national method: https://hydrobio-dce.cemagref.fr/

Metrics:

K = abundance (translated in 5 classes), CS = trophic score (0-20), E = stenoecy coefficient (1-3)

IBMR = Σ(K.CS.E)/Σ(K.E)

A pressure gradient based on hydromorphological data was obtained using PCA. Unstressed sites were chosen using the site position on the PCA axes.

Italy Biological Macrophyte Index for Rivers (IBMR)

Described in:

Haury, J., Peltre M.-C., Trémolières M., Barbe J., Thiebaut, G., Bernez, I., Daniel, H., Chatenet, P., Haan-Archipof, P., Mulller, S., Dutartre, A., Laplace-Treyture, C., Cazaubon, A. & Lambert-Servien, E. (2006) A new method for assess water trophy and organic pollution – The Macrophyte Biological Index for Rivers (IBMR) : its application to different types of rivers and pollution. Hydrobiologia 570: 153-158.

Web page describing the national method: https://hydrobio-dce.cemagref.fr/

Metrics:

K = abundance (translated in 5 classes), CS = trophic score (0-20), E = stenoecy coefficient (1-3)

IBMR = Σ(K.CS.E)/Σ(K.E)

The references sites have been selected on the basis of pressures analysis (land use, hydro dynamism, morphological alteration, physical and chemical features) at site, water body and catchment scales; in sampling sites also macrophythe communities have been detected to evaluate structural likeness with references type-specific communities and to evaluate presence and abundance of alien species, intensity and presence of natural disturbances.

MS National method Reference conditions setting at National level

Portugal Biological Macrophyte Index for Rivers (IBMR)

Described in:

Haury, J., Peltre M.-C., Trémolières M., Barbe J., Thiebaut, G., Bernez, I., Daniel, H., Chatenet, P., Haan-Archipof, P., Mulller, S., Dutartre, A., Laplace-Treyture, C., Cazaubon, A. & Lambert-Servien, E. (2006) A new method for assess water trophy and organic pollution – The Macrophyte Biological Index for Rivers (IBMR) : its application to different types of rivers and pollution. Hydrobiologia 570: 153-158.

Web page describing the national method: https://hydrobio-dce.cemagref.fr/

Metrics:

K = abundance (translated in 5 classes), CS = trophic score (0-20), E = stenoecy coefficient (1-3)

IBMR = Σ(K.CS.E)/Σ(K.E)

Reference conditions followed the guidelines and pressure screening criteria provided by the Working Group 2.3 – REFCOND and described on CIS WFD Guidance Document Nº 10 - Rivers and Lakes – Typology, Reference Conditions and Classification Systems. The applied methodology included spatial analysis, historical data analysis and expert judgment. Semi-quantitative analysis was used in order to assess the magnitude of 9 pressure variables (Land Use, Riparian Zone, Sediment Load, Hydrological Regime, Acidification and Toxicity, Morphological Condition, Organic Matter Contamination and Nutrient Enrichment, River Continuity) a procedure adapted from European Project FAME - Development, Evaluation and Implementation of a Fish-based Assessment Method for the Ecological Status of European Rivers. A Contribution to the Water Framework Directive (Contract EVK1-CT-2001-00094). This procedure was applied according to the specificities of the different river types and lack of true reference sites in some river types lead to the selection of “best available sites”. A final biological screening was also made.

Slovenia River Macrophyte Index (RMI)

Described in:Kuhar, U.,  Germ, M.,  Gaberščik,A., Urbanič, G. 2011. Development of a River Macrophyte Index (RMI) for assessing river ecological status. Limnologica 41:235-243.

Web page describing the national method: http://www.mop.gov.si/si/delovna_podrocja/direktorat_za_okolje/sektor_za_vode/ekolosko_stanje_povrsinskih_vod a/

RMI was calculated using to the following equation:

where QAi = abundance of the taxa i from the group A, QABi = abundance of the taxa i from the group AB, QBCi = abundance of the taxa i from the group BC, QCi = abundance of the taxa i from the group C, QSi = abundance of taxa i from all groups (group A, AB, B, BC, C; taxa from the group ABC are not considered), nA = total number of taxa in group A, nAB = total number of taxa in group AB, nBC = total number of taxa in group BC, nC = total number of taxa in group C, nS = total number of taxa in all groups (group A, AB, B, BC, C; taxa from the group ABC are not considered).

The criteria for the selection of the potential reference sites in the rivers include hydromorphological and physico-chemical condition of the site, riparian vegetation, floodplain and land use properties, saprobic index values, and some pressures presence. Potential reference sites were defined without considering the criteria of biotic pressures that includes allochthonous species and fishery management.

S

A CBCAB

n

iSi

n

i

n

iCi

n

iBCi

n

iABiAi

Q

QQQQRMI

1

1 111 21

21

MS National method Reference conditions setting at National level

Spain Biological Macrophyte Index for Rivers (IBMR)

Described in:

Haury, J., Peltre M.-C., Trémolières M., Barbe J., Thiebaut, G., Bernez, I., Daniel, H., Chatenet, P., Haan-Archipof, P., Mulller, S., Dutartre, A., Laplace-Treyture, C., Cazaubon, A. & Lambert-Servien, E. (2006) A new method for assess water trophy and organic pollution – The Macrophyte Biological Index for Rivers (IBMR) : its application to different types of rivers and pollution. Hydrobiologia 570 : 153-158.

Web page describing the national method: https://hydrobio-dce.cemagref.fr/

Metrics:

K = abundance (translated in 5 classes), CS = trophic score (0-20), E = stenoecy coefficient (1-3)

IBMR = Σ(K.CS.E)/Σ(K.E)

Based on near-natural reference sites. Selected by expert knowledge. The selection followed the REFCOND guidance and GIG criteria from the 1st phase of the IC exercise).

Annex B – Supplementary material for IC of RM1, RM2 and RM4

Annex B.1 Original national boundaries of MS. IBMR- Biological Macrophytes Index for Rivers; RMI – River Macrophyte Index

Method IBMR RMIMS CY FR GR IT PT SP SI Max 1 1.5292 1 1.2800 1.3325 1.4474 1H/G 0.795 0.93 0.75 0.90 0.92 0.95 0.80G/M 0.596 0.79 0.56 0.80 0.69 0.71 0.60M/P 0.397 0.66 0.37 0.65 0.46 0.47 0.40P/B 0.198 0.52 0.19 0.50 0.23 0.24 0.20

Annex B.2 Average absolute class differences of national methods after boundary harmonisation and percentage of classifications differing by half of a quality class (0.1 after piecewise transformation). All pair of MS are shown.

MS Class disagreement % of classifications differing by half of a quality class

Cyprus vs. France 0.21 12.46Cyprus vs. Greece 0.27 15.66Cyprus vs. Italy 0.31 26.69Cyprus vs. Portugal 0.35 24.56Cyprus vs. Spain 0.33 21.00France vs. Greece 0.25 19.22France vs. Italy 0.25 16.08France vs. Portugal 0.19 7.45France vs. Spain 0.17 9.02Greece vs. Italy 0.44 33.10Greece vs. Portugal 0.34 25.62Greece vs. Spain 0.41 22.42Italy vs. Portugal 0.26 11.63Italy vs. Spain 0.19 10.47Portugal vs. Spain 0.16 0.70France vs. Cyprus 0.18 9.80Greece vs. Cyprus 0.27 16.01Greece vs. France 0.29 22.42Italy vs. Cyprus 0.25 22.09Italy vs. France 0.25 17.05Italy vs. Greece 0.37 29.84Portugal vs. Cyprus 0.34 25.70Portugal vs. France 0.26 15.14Portugal vs. Greece 0.32 26.41Portugal vs. Italy 0.28 16.20Spain vs. Cyprus 0.31 21.83Spain vs. France 0.26 16.90Spain vs. Greece 0.38 23.24Spain vs. Italy 0.23 16.90Spain vs. Portugal 0.16 0.70Slovenia vs. Cyprus 0.24 23.96Slovenia vs. France 0.30 15.97Slovenia vs. Greece 0.36 26.39Slovenia vs. Italy 0.21 14.58Slovenia vs. Portugal 0.15 0.00Slovenia vs. Spain 0.13 3.47Cyprus vs. Slovenia 0.25 22.42France vs. Slovenia 0.18 6.67Greece vs. Slovenia 0.37 24.56Italy vs. Slovenia 0.14 8.14Portugal vs. Slovenia 0.17 3.87Spain vs. Slovenia 0.13 7.04

Annex C – Results IC Option 1 RM5

Annex C.1 Boundaries for countries that share RM5 IC type.

Method MMI IBMR RMI MS CY IT PT SP SIH/G 0.880 0.90 0.93 0.92 0.80G/M 0.660 0.80 0.70 0.69 0.60M/P 0.440 0.65 0.46 0.43 0.40P/B 0.220 0.50 0.23 0.23 0.20

Annex C.2 Original boundaries (national scale and regressed).Regression analyses per MS used to convert the national boundaries into comparable boundaries. IT: y=0.8674x+0.1655; PT: y=0.9999+0.0002

Boundaries

Original Original MedGIG

IT PT IT PTMax 1.1240 1.1538 1.1405 1.1535H/G 0.90 0.93 0.946 0.930G/M 0.80 0.70 0.859 0.700M/P 0.65 0.46 0.729 0.460P/B 0.50 0.23 0.599 0.230

Annex C.3. IC results Option 1 for RM5 (values of original boundaries were regressed using MedGIG benchmark data values for each MS).

Original HarmonizedIT PT IT PT

Max 1.1404 1.1535 1.1404 1.1535MedGIG_H/G 0.946 0.930 0.946 0.930MedGIG_G/M 0.859 0.700 0.726 0.728MedGIG_M/P 0.729 0.460 0.729 0.460MedGIG_P/B 0.599 0.230 0.599 0.230Median H/G MedGIG 0.938 0.938Median G/M MedGIG 0.780 0.727

H width to Max 0.194 0.224 0.194 0.224G width 0.087 0.230 0.220 0.202M width 0.130 0.240 -0.003 0.268H/G bias 0.008 -0.008 0.008 -0.008G/M bias 0.080 -0.080 -0.001 0.001

H/G bias_CW 0.092 -0.036 0.036 -0.036G/M bias_CW 0.612 -0.346 -0.005 0.004

Annex 2.4Illustration of IC results Option 1 for RM5 for the H/G(1st graph) and G/M boundaries (2nd and 3rd graph).

Annex 2.4National class boundaries and boundaries bias adjustment for RM5 IC type (adjusted boundaries if bias >|0.25|). Proposed adjustments: ↑ boundary to be raised, ↓ boundary should be lowered.

RM5 Original Harmonized Final agreed national boundaries

H/G G/M H/G G/M H/G G/M

ItalyBoundary 0,946 0,859 0.726 ↓ Not accepted due to the small

common databaseBias CW 0.092 0.612 -0.005

PortugalBoundary 0,930 0,700 0.728 ↑ Not accepted due to the small

common databaseBias CW -0.036 -0.346 0.004

Annex 2.5Conversion to the national boundary values (using the regression equations).

IT PTHarmonized boundaries (MedGIG scale)

H/G 0.946 0.929G/M 0.726 0.728

Harmonized boundaries (national scale)

H/G 0.90 0.93G/M 0.65 0.73

Annex 2.6 Remarks concerning temporary rivers (RM5)Mediterranean temporary rivers are characterized by high annual hydrological variability, they have alternating wet and dry periods every year, and usually dry out in a predictable way on an annual basis; the flooding can last for several months depending on the degree of harshness of the region. Many of these rivers have high plant biomass and cover in the drought season derived mainly by the increasing cover of amphibious species, and other species that are adapted to wet substrates and that do not require aquatic environments to complete their growth cycle. With increasing eutrophication, the biomass of nythrophyllous species (e.g. Oenanthe crocata, Apium nodiflorum, Typha latifolia) and of macroalgae (e.g. green algae) increase, along with a decrease of cover and occurrence of some aquatic bryophytes (e.g. Fontinalis antypyretica, Bryum pseudotriquetrum). In some impaired sites, free-flowing hydrophytes form dense monoespecific communities (e.g. Lemna sp., Azolla sp.). It is the Med GIG opinion that macrophytes are a reliable BQE to be used in the assessment of the ecological status for the WFD.Diverse classifications of temporary rivers can be found in literature in relation to increasing intermittency and to the predictability of the occurrence of low flows and drying (e.g. near-permanent, episodic, intermittent, seasonal). In addition, there are high hydrological disturbances caused by water diversion, regulation and by a high frequency of extreme climate events (long-term droughts and unexpected floods events) in temporary river systems. Thus these differences between countries should be taken into account in further intercalibration exercise of national methods.