• ROMANIA
  • PHYTOBENTHOS
  • RIVERS

Method name

National (Romanian) Assessment Method for Rivers Ecological Status based on Phytobenthos (Diatoms)

Acronym

RO-AMRP

Scientific development and confirmation of the national method for ecological status assessment of waterbodies (rivers, lakes) based on diatoms (phytobenthos), and completion of the intercalibration exercise.

2: Rivers

Martyn Kelly

Bowburn Consultancy, 11 Monteigne Drive, Bowburn, Durham DH6 5QB

1.Introduction

2.Description of the current national assessment method

3.National data set

4.Typology

5.Performance of national metric and other diatom indices

6.Reference/benchmark conditions

7.Boundary setting procedure

8.Checking of WFD compliance and feasibility

9.IC procedure

10. Normalisation of boundaries

11. Conclusions

12. References

Appendix 1: Rationale from excluding rivers typology RO16 (“watercourses qualitatively influenced by natural causes”: chlorosodic) from intercalibration

Appendix 2: Procedure for computation of ecological status for Romanian rivers using phytobenthos

1.Introduction

Intercalibration of phytobenthos-based methods of ecological assessment of rivers was largely completed as part of the second phase of intecalibration. An exercise for Eastern Continental GIG included methods from Austria, Bulgaria, Czech Republic, Hungary, Slovakiaand Slovenia. The Romanian method was not included in the final report, however, largely due to poor correlations with the Intercalibration Common Metric.

This report describes development of an alternative Romanian phytobenthos assessment system for rivers (excluding large rivers, which are covered by a separate intercalibration exercise) and shows how this is compliant with the WFD Normative Definitions and that the class boundaries correspond with those agreed by the completed intercalibration exercise.

2.Descriptionof thecurrent national assessment method

The national assessment method for phytobenthos uses diatoms as proxies for the entire phytobenthos community. Samples are collected two or three times a year(April-May, July-August, September/October (November));each consists of 5-20 spatial replicates from all representative habitats (mostly stones). These are then digested using hydrogen peroxide and permanent slides are prepared. These are analysed in the laboratory and at least 300 (usually at least 400) diatoms are identified to species and the number of each species counted. The main identification literature used is Krammer and Lange-Bertalot (1986-2004).

The national metric that was originally developed was designed to assess eutrophication, organic pollution, general degradation (unspecified pressures), hydromorphological alteration, alteration of bank habitat etc. It was a multimetric index consisting of four metrics, weighted as follows:

-Saprobic Index (IS)30%

-Taxa number Index (INT) 15%

-Shannon-Wiener Diversity Index (ID) 30%

-Biologic Index for Diatomea/Bacillariophycea (IBD)25%

The index is calculated as follows:

Multimetric index = 0,3*EQRIS+0,15*EQRINT+0,3*EQRID+0,25*EQRIBD

For each metric, the “expected value” for EQR calculations is obtained from Table 1.

Table 1. Guide values for calculation of components of Romanian national metric

Typology / Guide values
Saprobic Index
(max.) / Shannon-Wiener diversity (min) / Taxa number
(min.) / IBD
(min.)
RO01 / 1,25 / 2,8 / 16 / 20
RO02+03 / 1,5 / 2,6 / 18 / 19
RO04+05 / 1,65 / 2,5 / 18 / 19
RO06-11 / 1,8 / 2,4 / 17 / 18
RO12-15 / 1,7 / 2,5 / 18 / 17
RO16 / 1,6 / 1,9 / 13 / 19
RO17 / 1,3 / 2,6 / 15 / 18
RO18-19 / 1,6 / 2,4 / 14 / 17
RO20 / 1,75 / 2,2 / 16 / 17

For ecological status/ecological potentialclassification, the following thresholds are used:

–High ecological status/Maximum ecological potential min. 0.78

–Good ecological status/Good ecological potential min. 0.62

–Moderate ecological status /Moderate cological potential min. 0.39

–Poor ecological status min. 0.28

–Bad ecological status max. 0.28

Where there are several samples from a single water body, classification is based on the average value of the multimetric index.

3.National data set

Diatom count data and associated environmental information are available from 3234 samples representing 466 water bodies in Romania. These data were collected between 2010 and 2013 and are summarised in Table 2.

Table 2. Summary of data available for Romanian riverphytobenthos intercalibration exercise. See Table 3 for a description of river types

Type / Aligns to ... / Number of samples
All / Benchmark + reference
Water bodies / Samples / Water bodies / Samples
RO01 / R-E1a / 231 / 1753 / 14+15 / 100+118
RO02 / R-E1a / 11 / 122 / 2+1 / 18+8
RO03 / R-E1a / 6 / 44 / 2 / 14
RO04 / R-E1b / 54 / 295 / 5 / 135
RO05 / R-E1b / 40 / 301 / 7 / 39
RO06 / R-E3 / 17 / 142 / 1 / 4
RO07 / R-E3 / 4 / 21 / 0 / 0
RO08 / R-E3 / 5 / 48 / 0 / 0
RO09 / R-E3 / 0 / 0 / 0 / 0
RO10 / “large rivers” / 8 / 41 / 3 / 12
RO11 / “large rivers” / 7 / 46 / 2 / 13
RO15 / “large rivers” / 3 / 55 / 2 / 24
RO16* / [R-E3] / 11 / 101 / 2 / 12
RO17* / R-M5 / 17 / 76 / 8+2 / 50+11
RO18* / R-M5 / 36 / 115 / 0 / 0
RO19* / R-M5 / 16 / 74 / 1 / 6
Total / 466 / 3234 / 67 / 564

* see notes below. types.

4.Typology

The current Romanian typology is summarised in Table 3, and links to the Eastern-Continental (EC) and Mediterranean (MED) GIG typologies are shown in Table 4. No samples for RO09 were available. Three Romanian river types (RO17, RO18, RO19) are not covered by the EC typology and correspond more closely to the Mediterranean GIG type R-M5 (temporary water courses). Fig. 1 summarises the alkalinity values recorded for each type.

Four other types (RO12, RO13, RO14, RO15) are represented only by the Danube River and will be considered within the “large rivers” cross-GIG intercalibration exercise. Of these types, only RO15 has phytobenthos data; the other three types are considered to be Heavily Modfied Water Bodies.

The remaining types are split between R-E1a, R-E1b and R-E3, with the exception of RO16 (“watercourses qualitatively influenced by natural causes”) which includes watercourses that fall within one of the following situations:

  • a natural background of high concentrations of heavy metals due to their location within metallogenic zones;
  • receiving springs with chlorosodic waters, located near salt massifs, having specific geology, which leads to very high conductivity and TDS values, and other physicochemical parameters;
  • where mining has been practiced for centuries; the impact is mostly due to acid mine waters and heavy metal loads; trends in the quality of these waterbodies are characterized by significant fluctuations depending on season, weather conditions etc.

This type is not compatible with any of the intercalibration type descriptions and will not be considered further in this exercise. More information is given in the Appendix.

Table 3. The Romanian national river typology. See Tables2 and 4 for correspondence with IC and broad typologies.

Type / Code / Ecoregion / Catchment area
(km2) / Geology / Altitude
(m)
Watercourse located in the mountains, piedmont or highlands area / RO01 / 10 / 10-1000 / a - siliceous
b - calcareous
c - organic / >500
Watercourse reach located in the piedmont or highlands area
Watercourse reach located in the piedmont or highlands area, with endemic species / RO02
RO02*1) / 10 / 1000-10000 / a - siliceous
b - calcareous / >500
Watercourse reach located in mountain depression / RO03 / 10 / >10 / a - siliceous
b - calcareous
c - organic / >500
Watercourse located in the hills or plateau areas / RO04 / 10, 11, 12, 16 / 10-1000 / a - siliceous
b - calcareous
c - organic / 200-500
Watercourse reach located in the hills or plateau areas / RO05 / 10 / 1000-10000 / a - siliceous
b - calcareous
c - organic / 200-500
Watercourse located in lowland areas
Watercourse located in lowland areas, without fish fauna in natural conditions / RO06
RO06*2) / 11, 12, 16
12 / 10-2000 / a - siliceous
b - calcareous
c - organic / <200
Watercourse reach located in lowland areas / RO07 / 11 / 1000-3000 / a - siliceous
b - calcareous
c - organic / <200
Watercourse reach located in lowland areas
Watercourse reach located in lowland areas, without fish fauna in natural conditions / RO08
RO08*3) / 12 / 1000-5000 / a - siliceous
b - calcareous
c - organic / <200
Watercourse reach with wetlands, located in lowland areas / RO09 / 16 / 1000-5000 / a - siliceous
b - calcareous
c - organic / <200
Watercourse reach located in lowland areas
Catchment area > 3000 km2- ECO 11
Catchment area > 5000 km2 - ECO 12, 16*4) / RO10
RO10*4) / 11
12, 16*4) / >3000
>5000*4) / a - siliceous
b - calcareous
c - organic / <200
Watercourse reach with wetlands, located in lowland areas
Catchment area > 3000 km2- ECO 11
Catchment area > 5000 km2 - ECO 12, 16*5) / RO11
RO11*5) / 11
12, 16*5) / >3000
>5000*5) / a - siliceous
b - calcareous
c - organic / <200
Danube River - Cazane area / RO12 / 12 / 570900-574850 / b - calcareous / 100-200
Danube River - lower sector between Cazane and Calarasi / RO13 / 12 / 574850-698000 / a - siliceous / 5-70
Danube River between Calarasi and Isaccea / RO14 / 12 / 698000-780650 / a - siliceous / 5
Isaccea - Danube Delta / RO15 / 12 / 780650-805300 / c - organic / <5
Watercourses qualitatively influenced by natural causes / RO16 / 10-1000
Temporary watercourse located in the mountains, piedmont or highlands area / RO17 / 10-1000 / a - siliceous
b - calcareous / >500
Temporary watercourse located in the hills and plateau areas / RO18 / 10-1000 / a - siliceous
b - calcareous / 200-500
Temporary watercourse located in lowland areas / RO19 / 10-2000 / a - siliceous
b - calcareous / <200

Fig. 1. Range of alkalinity values associated with Romanian national river types, based on data collected between 2010 and 2013. Vertical lines group Romanian national types into appropriate intercalibration types (from left): R-E1a, R-E1b, R-E3, large rivers, RO16 and R-M5 (see text for more details). Note that the scale on the Y axis is logarithmic.

Table 4. E-C GIG common river types and the corresponding national type in Romania.

Common IC type / Description / Corresponding Romanian type
R-E1aCarpathians: small to medium, mid-altitude / Catchment: 10 - 1,000 km2
Altitude: 500-800 m
Geology: mixed
Channel substrate: / RO01, RO02, RO03
(RO02 also included although catchment area is 1000 - 10000 km2. A few sites are < 500 m altitude but correspond, in other respects, to R-E1a.)
R-E1bCarpathians: small to medium, mid altitude / Catchment: 10 - 1,000 km2
Altitude: 200-500m
Geology: mixed
Channel substrate: / RO04, RO05
(RO05 also included although catchment area is 1000 - 10000 km2.)
R-E2 Plains: medium-sized, lowland / Catchment: 100 - 1,000 km2
Altitude: < 200 m
Geology: mixed
Channel substrate: sand and silt
R-E3 Plains: large, lowland / Catchment: > 1,000 km2
Altitude: < 200 m
Geology: mixed
Channel substrate:sand, silt, gravel / RO06, RO07, RO08, [RO09]
(note that RO06 straddles R-E2 and R-E3, but also includes streams smaller than the lower limits for either.)
R-E4 Plains: medium-sized, mid-altitude / Catchment: 100 - 1,000 km2
Altitude: 200-500 m
Geology: mixed
Channel substrate: sand and gravel / -
R-M5 Temporary rivers / Catchment:
Altitude:
Geology:
Channel substrate: / RO17, RO18, RO19
“large rivers” / Catchment:
Altitude:
Geology:
Channel substrate: / RO10, RO11, RO15

5.Performance of national metric and other diatom indices

5.1 Characteristics of the Romanian river database

The river phytobenthos intercalibration exercises address a combination of nutrient enrichment and “general degradation”. Although, in theory, any nutrient may be limiting, in practice attention focuses on the role of phosphorus and, to a lesser extent, nitrogen. Both of these nutrients span a wide gradient in the dataset, with maximum values exceeding 1 mg l-1 PO4-P and 10 mg l-1 Total Oxidised Nitrogen (TIN; Figs. 2a, 2b)and available (i.e. dissolved) nitrogen and phosphorus are positively correlated (Fig. 2a). The likelihood of nitrogen limiting algal production was evaluated using the “Redfield ratio”, which describes the approximate ratio of N and P in algal cells (originally marine phytoplankton) as 16. If N:P > 16 there is more nitrogen in relation to the P available for algal growth whilst if N:P < 16, N is in short supply relative to P (Tett et al., 1985). This ratio is expressed in moles and translates to 7.24:1 when dealing with mass. N:P tends to decrease (i.e. likelihood of N limitation increases) as the P concentration increases;however, potentially N-limiting conditions can occur even at low P concentrations (Fig. 2b). Figs 2c and 2d show that concentrations of total phosphorus and nitrogen have strong positive correlations with their dissolved fractions. Subsequent graphs will focus on the relationship between variables and PO4-P but these plots are evidence that this variable captures the main gradients in inorganic nutrients within the Romanian dataset.

Fig. 2. Relationships between inorganic nutrients in the Romanian river phytobenthos dataset: a. PO4-P versus Total Inorganic Nitrogen (NO3-N + NO2-N + NH4-N); b. PO4-P versus N:P ratio (calculated as TIN:PO4-P; horizontal line indicates N:P = 7.2); c. Total Phosphorus (TP) versus PO4-P; d. Total Nitrogen (TN) versus TIN.

5.2 Performance of the Romanian national metric

The current Romanian phytobenthos metric has a significant relationships with total inorganic nitrogen (TIN: NO3-N + NO2-N + NH4-N) and biochemical oxygen demand (BOD)but not with alkalinity (summarising geological differences between rivers) or PO4-P (Fig. 3); however the predictive power these relationships are both low (Table 5) and it would be unwise to predict ecological status class boundaries from these.

The strongest relationship, of those shown in Fig. 3, was for TIN and Fig. 4 shows the performance of the four component metrics against TIN separately, in order to understand why the national metric is not performing well. Both of the diversity measures (number of taxa and Shannon-Wiener index) along with the saprobic index show a flat response across the pressure gradient, with the widest range of values recorded at intermediate levels of pressure. IBD had a stronger relationship with TIN but still only explained a small part of the total variation (Table 5).

Inclusion of diversity metrics in ecological status assessments should, in theory, provide a measure of the condition of the community that lacks the inherent circularity associated with most pressure metrics. However, the relationship between diversity of diatom assemblages and water quality is not straightforward (van Dam, 1981; Blanco et al., 2012). This is partly because the diatom assemblage is only part of a larger community (DeNicola & Kelly, 2014) whose diversity is, in turn, shaped by a variety of other forces, many natural. Approaches to sampling may also contribute to the problem: a single diatom sample collected strictly according to EN13948 (CEN, 2014), for example, will only represent a part of alpha diversity at a site, ignoring seasonal and microhabitat variations.

Fig. 3. Relationship between Romanian national phytobenthos metric and key environmental variables: a. alkalinity; b. dissolved phosphorus (“PO4-P”); c. total oxidised nitrogen (“TIN”); and, d. 5-day biochemical oxygen demand (“BOD5”).

Table 5. Regression parameters for the relationship between Romanian national metric and key environmental variables, and for component metrics of the Romanian national metric individually against TIN . *: p < 0.05; **: p < 0.01; ***: p < 0.001; N.S.: not significant.

Variable (log10 transformed) / Equation / F / R2
Romanian metric versus environmental parameters
alkalinity / 0.004x + 1.056 / 0.269 N.S. / -0.0003
PO4-P / 0.005x + 1.058 / 1.177 N.S. / < 0.0001
TIN / -0.017x + 1.048 / 10.62 ** / 0.004
5 day BOD / 0.018x + 1.045 / 5.961 * / 0.002
Component metrics against TIN
Saprobic index / 0.001x + 1.397 / 0.13 N.S. / -0.0004
Number of taxa / -0.073x + 1.119 / 14.09 ** / 0.005
Shannon_Wiener index / 0.030x + 0.894 / 13.64 ** / 0.005
IBD / -0.06x + 0.773 / 160.4 *** / 0.06

Fig. 4. Performance of the four component metrics of the Romanian national phytobenthos metric versus TIN (the variable with the strongest association with the complete multimetric).

5.3 Performance of other diatom-based metrics

The phytobenthos intercalibration exercises addressed the effect of eutrophication and general degradation on benthic algal assemblages. Although, in theory, any nutrient may be limiting, in practice attention focuses on the role of phosphorus although preliminary analyses (Fig. 2b) have shown that nitrogen, too, may be limiting in some circumstances. In the following section, “PO4-P” (“dissolved phosphorus”) will be used as a primary stressor gradient against which the performance of candidate metrics can be judged. These candidate metrics were calculated using the Omnidia package (version 5.3; Lecointe et al., 1993) and the outcomes plotted against PO4-P (Fig. 5). Note that the version of the TDI in this package is not the same as the metric currently used by UK and Ireland for assessing ecological status (Kelly et al., 2008). Linear regressions between log10 TP and log10 Alkalinity indicate the ability of the metric to capture the pressure gradient, and the scale of interference from type / geological factors (Table 6). In all cases, both “pressure” and “type” were significant, though R2 was very low in the case of Sladacek’s index (INDSLA). The scatterplots for this index (Fig. 5m,n) show a group of sites with very low values for this index, which sit outside the main trend. However, as this index is not used elsewhere in Europe, the cause of these outliers was not investigated in detail.

The other metrics all had noisy relationships with pressure, explaining no more than 20% of the total variation, with a further 3-9% of variation explained by the interaction with alkalinity (as a proxy for underlying geology). The best relationship with pressure was obtained using EPI-D; however, this metric also had the strongest interaction with alkalinity.

The final decision was to use the phytobenthos Intercalibration Common Metric (pICM), the average of IPS and Rott’s TI (Rott et al., 1999) for the following reasons:

  • This combination of metrics has already been used widely around Europe for intercalibrating diatom assessment systems for rivers (Birk & Hering, 2008; Kelly et al., 2009; Almeida et al., 2013). These exercises have included rivers in both Eastern-Continental and Mediterranean GIGs, into which Romanian streams have been placed. By using pICM as the Romanian national metric, the intercalibration process will be more straightforward, as there will be no additional errors introduced by regressing the national metric against the ICM in order to convert Romanian data to the common scale.
  • The pICM combines the IPS, a metric of “general degradation”, that is effective across a wide range of water quality, and Rott’s TI, a “trophic” metric originally designed for rivers, that is most sensitive at relatively low levels of pressure (optimised for elevated inorganic, rather than organic, enrichment). The two metrics are, therefore, complementary and allow more sensitive analyses of pressure-response relationships than would be possible using either alone

However, as calculation of pICM requires the metrics to be converted to EQRs, and as the “expected” values vary between stream types, it is not possible, at this stage, to include plots showing the performance of the pICM relative to raw metrics.

Fig.5. Scatterplots showing the relationship between candidate diatom metrics and total phosphorus (TP, left) and alkalinity (right). Note that the scale for TDI, Rott’s TI and Rott’s SI have all been inverted by Omnidia to ease comparisons with other metrics. Both X axes are displayed on a logarithmic scale.

Fig. 5 (cont.)

Table 6. Regression parameters between diatom index values and TP alone and in combination with alkalinity. Both TP and alkalinity were significant in all cases; “difference” indicates the increase in R2 through addition of Alkalinity to the equation.

Index / TP / TP + Alkalinity
R2 / Significance / R2 / Difference
IBD / 0.163 / *** / 0.237 / 0.074
TDI / 0.134 / *** / 0.224 / 0.090
IPS / 0.148 / *** / 0.219 / 0.071
Rott’s TI / 0.154 / *** / 0.215 / 0.061
Rott’s SI / 0.150 / *** / 0.183 / 0.033
EPI-D / 0.204 / *** / 0.294 / 0.090
Sladacek’s index / 0.015 / *** / 0.050 / 0.035

6.Reference/benchmark conditions

6.1 Principles of intercalibration and the importance of the reference condition

The principle of ecological assessment adopted by the Water Framework Directive is that the ecological state of a water body is compared with that expected in the absence of anthropogenic conditions. This can be expressed as a ratio: observed/expected which, in turn, assumes that all Member States have examples of water bodies in the ”expected” condition (”reference state”). ECOSTAT developed a set of criteria that could be applied to all water bodies when selecting reference sites; these, in turn, were ”tuned” by each GIG to account for biogeographical differences between regions (Table 7). If all participants in an intercalibration exercise have valid reference sites, then the comparison of boundaries can proceed using these as a shared benchmark against which national methods can be calibrated.

For the present exercise, reference sites were abundant for RO01 (Table 9); however, there was only one reference site for RO02 and two for RO17, and none for any other types, based on Romanian national criteria. However, a lack of reference sites elsewhere in EC GIG precluded use of this approach. Sufficient reference sites were, however, available for R-M5 rivers to allow this approach to be adopted for the intercalibration of RO17-RO19.