Phytobenthos Intercalibration - Rivers
Alpine Geographical Intercalibration Group
Final Report
;
The intercalibration of the boundary values for the Phytobenthos method in the Alpine River GIG was done following the procedure of the Central Baltic GIG (see Kelly et al. 2006) and the results show that a further modification is not necessary. Thus, maximum consistency with the approach in the other Geographical Intercalibration Groups is guaranteed.
1.National Assessment Systems
MS / National metricAustria / Multimetric method consisting of 3 modules/metrics:
- trophic status module (based on TI: Rott et al. 1999)
- saprobic status module (based on SI: Rott et al. 1997)
- reference species module (portion of defined reference and bioregion-specific species in total abundance and species number)
France / IBD (national routine index: AFNOR, 2000)
Germany / Diatom Module: WFD Diatom Index = Average of the sum of abundances of type specific reference species (following Schaumburg et al. 2005) and Trophic Index (Rott et al., 1999) or (in one special case) Saprobic Index (Rott et al., 1997). Additional metrics are available for cases of acidification or salinisation.
Non Diatom Module: WFD Reference species Index depends on type specific taxa and abundances (following Schaumburg et al. 2005)
Macrophyte Module: WFD Reference species Index depends on type specific taxa and abundances (following Schaumburg et al. 2005). Additional metrics are available for cases of mass growth stands of special taxa.
Ecological status is calculated and classified from the average of the three module scores. If a module is absent, status class can be calculated with two modules or, exceptionally, with a single module. For this reason every module is classified separately and can be considered separately for intercalibration purposes. The national classification system needs all modules of the benthic flora occurring in a monitoring section of a water body.
Italy / work in progress (see the Annex 1)
Slovenia / Multimetric method consisting of 2 modules/metrics:
- Saprobic index (Zelinka & Marvan 1961)
- Trophic index (Rott et al. 1997)
worst case approach
Spain / three indexes:
IPS (Coste in Cemagref 1982)
IBD (Prygiel & Coste 2000)
CEE (Descy & Coste 1990, 1991)
(IPS seems to be the most adequate index for the Alpine GIG)
2.Comparison methodology
In general this is similar to the methods that were applied in the invertebrate - intercalibration.
The technical steps of the calculation are:
a)Calculation of the ICM–metrics IPS and TI (done by Michel Coste and Juliette Tison)
b)Calculation of EQR values for the ICM–metrics (reference values is the median of the reference sites)
c)Regression between national method and ICM
d)Converting national boundary values into ICM boundary values (including 95% CL) using the regression formula
3.Calculation of the ICM – metrics:
Similar to the Central Baltic GIG the IPS (indice de polluosensibilité (Cemagref1982)) and the TI (Trophic diatom index Rott et al.(1999)) were used.
The calculation was done in the same way as in the Central Baltic GIG:
IPS:
EQR_IPS = Observed value / reference value
TI:
as this is a trophic index it needs to be adjusted so that high values represent high EQR values (4 is the maximum possible value of the TI):
EQR_TI = (4-observed value) /(4-reference value )
4.Calculation of the final ICM
Two options (similar to the approach in the Central Baltic GIG)we tested:
A) Arithmetic mean of IPS and TI
B) Minimum of IPS and TI (worst case)
Option A resulted in better correlations between the national methods and the ICMi and was therefore selected for the use in the intercalibration procedure. This option was also used by the Central Baltic GIG (Results for option B can be obtained from the data file).
5.Regression between national methods and ICM
Figure 1 and Table 1 show the results of the regression:
Figure 1: Regression: x-axis = national method and y-axis = ICM
Table 1: Data from Regression in Figure 1
number of sites / number of reference sites / R² of national method vs ICM / lowest EQR value in data setType R-A1
Austria / 157 / 18 / 0.96 / 0.39
France / 117 / 66 / 0.56 / 0.43
Germany / 46 / 9 / 0.90 / 0.20
Italy / - / - / - / -
Slovenia / 11 / 4 / 0.68 / 0.60
Type R-A2
Austria / 111 / 17 / 0.92 / 0.27
France / 52 / 26 / 0.82 / 0.43
Spain / 40 / 6 / 0.81 / 0.66
Italy / - / - / - / -
6.Results: Boundary values
Table 2 and Figure 2 show the results from the conversion of the national boundary values into ICM values. Table 3 shows the average ICMi boundary values +/- 5 percent confidence interval (upper and lower limit of the band.
Table 2: National boundary values converted into ICM – values; the 95% CL is taken from the regression
National boundary / National boundary ICMiMS / H/G / G/M / H/G / G/M
value / +/-95% CL / value / +/-95% CL
Type R-A1
Austria / 0.87 / 0.56 / 0.92 / 0.003 / 0.72 / 0.006
France / 0.86 / 0.71 / 0.87 / 0.020 / 0.76 / 0.031
Germany / 0.73 / 0.54 / 0.92 / 0.011 / 0.78 / 0.018
Italy / - / - / '- / '- / '- / '-
Slovenia / 0.80 / 0.60 / 0.95 / 0.054 / 0.75 / 0.106
Type R-A2
Austria / 0.87 / 0.56 / 0.95 / 0.008 / 0.71 / 0.012
France / 0.86 / 0.71 / 0.84* / 0.024 / 0.69 / 0.034
Spain / 0.94 / 0.74 / 0.89 / 0.023 / 0.67 / 0.051
Italy / '- / '- / '- / '- / '- / '-
* The French boundary value is only very slightly below the acceptable band (0.00124 point below) and the upper 95% confidence limits lie within the acceptable band. The analyses suggest that the present FR boundary lies within the statistical error of the exercise and that there is not a problem with either the FR reference concept or ecological status definition.The exact values are:
France High-Good Boundary for RA2: 0.8439
Lower limit of the band (= average - 0.05): 0.84514
Figure 2: Comparison of the national boundary values when converted into ICM – values (+/- 95% CL). Green: High/Good boundary, Red: Good/Moderate boundary. The bands show the average boundary values (bold line) +/- 0.05.
Table 3: Harmonised boundary values: average ICMi boundary values +/- 5 percent “confidence interval “(upper and lower limit of the band) expressed as EQR-value of the ICMi
River Type and boundary / average boundary (ICMi) / lower limit(ICMi) / upper limit(ICMi)R-A1 HIGH-GOOD / 0.92 / 0.87 / 0.97
R-A1 GOOD-MODERATE / 0.75 / 0.70 / 0.80
R-A2 HIGH-GOOD / 0.90 / 0.85 / 0.95
R-A2 GOOD-MODERATE / 0.69 / 0.64 / 0.74
7.Literature
CEMAGREF, 1982. Etude des méthodes biologiques d'appréciation quantitative de la qualité des eaux. Rapport Q.E. Lyon. Division Qualité des Eaux - Pêche et Pisciculture, Lyon.
Kelly, M., C. Bennett, M. Coste, F. Delmas, L. Denys, L. Ector, C. Fauville, M. Ferreol, M. Golub, M. Kahlert, J. Lucey, B. Ni Chathain, I. Pardo, P. Pfister, J. Picinska-Faltynowicz, C. Schranz, J. Schaumburg, J. Tison, H. van Dam & S. Vilbaste, 2006. Central/Baltic GIG Phytobenthos Intercalibration Exercise - Draft final report. Bowburn Consultancy, Durham.#
Rott, E., P. Pfister, H. van Dam, E. Pipp, K. Pall, N. Binder & K. Ortler, 1999. Indikationslisten für Aufwuchsalgen in Österreichischen Fliessgewässern, Teil 2: Trophieindikation und autökologische Anmerkungen. BundesministeriumfürLand- und Forstwirtschaft, Wasserwirtschaftskataster, Wien.
8.Data File used for Analysis:
ANNEX 1
Notes on phytobenthos intercalibration of Italian rivers (Alpine GIG)
11th June 2007
Prepared by Mancini L. (ISS), Buffagni A. (CNR-IRSA), Cazzola C. (CNR-IRSA), Erba S. (CNR-IRSA), Marchegiani S. (ISS), Puccinelli C. (ISS)
for the Italian Ministry for the Environment and Territory
This document summarizes the situation in Italy for WFD Rivers Diatoms Intercalibration. Different steps (1-7) will be followed and fulfilled to satisfy intercalibration requirements, starting from data collection and overview on data availability going to boundary derivation and harmonization.
- Data availability
- Reference sites/samples availability
- National method
- Validation through pressure analysis
- Relationship with ICMi
- Boundary derivation
- Harmonization
Data availability
Data have started to be centrally collected and the process is close to an end. The following table reports data availability at this stage.
R-A1 / R-A2Tot samples / 63 / 19
Enough samples have been collected for R-A1 type, possible implementation in the number of samples for R-A2 is expected. When dataset is finalized, all the data will be sent to GIG co-ordinator.
The minimum EQR value for the National method will be provided in order to verify the length of the gradient covered.
Availability of Reference sites/samples
R-A1 / R-A2No. of Reference sites/samples / 4 / 2
Refinements are required for what concerns Reference selection and identification, in order to possibly improve the number of sites/samples available especially for R-A2.
National method
At the moment Italy does not have an official WFD compliant method for Diatoms. Usually, where diatoms samples are collected many different Indices are calculated, mainly derived from other European countries (e.g. IPS: Coste, 1987). One of the most commonly calculated indices at the national level is the Pollution index based on diatoms (EPI_D: Dell’Uomo, 1996), but its suitability at the national scale has still to be demonstrated.
For this reasons, we are now calculating a relatively large set of Diatoms metrics, so that a proper choice for a WFD-compliant, appropriate metric or multi-metric index can be done based on field data.
Validation through pressure analysis
Italy is now in the stage of testing different indices against nutrients concentrations in order to select the best performing index as the possible National index. This process will also include the testing of the Intercalibration Common Index selected for Alpine, Central and Mediterranean GIG.
Relationship with ICMi
The regression between the National method, once selected, and the Intercalibration Common Metric index (average of IPS_EQR and TID_EQR) will be provided. Because at this stage a National method is not yet selected we are hereafter providing R2 values between the Diatoms indices presently calculated for R-A1 and R-A2 types (Table 2, 3).
Table 1 reports the indices that have been calculated.
Table1. Calculated indices in R-A1 and R-A2 Italian datasets.
Abbreviation / Full name / ReferenceIDAP / Indice Diatomique Artois Picardie / (Lecointe et al., 2003)
IBD / Biological Diatom Index / (Prygiel & Coste, 1999)
SHE / Steinberg & Schiefele trophic index / (Steinberg & Schiefele, 1988)
SID / Ind. saprobique / (Rott et al., 1997)
WAT / Watanabe et al pollution index / (Lecointe et al., 2003)
SLAD / Sládecek's pollution index / (Sladecek, 1986)
DESCY / Descy's pollution index / (Descy, 1979)
L&M / Leclercq & Maquet's pollution index / (Leclercq & Maquet, 1987)
IDG / Generic Diatom Index / (Lecointe et al., 2003)
CEE / Commission for Economical Community index / (Descy & Coste, 1991)
LOBO / Lobo & al. Bresil (2003) / (Lobo & al. Bresil, 2003)
IDP / Pampean Diatom Index / (Gomez N. Licursi M., 2001)
DI_CH / Hurlimann Suisse (2002) / (Hurlimann Suisse, 2002)
TDI / Trophic Diatom index / (Kelly & Whitton, 1995)
EPI_D / Pollution index based on diatoms / (Dell'Uomo, 1996)
IPS / Specific Pollution Sensitivity Index / (Coste, 1987)
TID / Trophic index / (Rott et al., 1999)
Table 2.R2 values between different diatoms index in R-A1 dataset (Italy).
Idap / Ibd / She / Sid / Wat / Sla / Descy / L&m / Idg / Cee / Lobo / Idp / Di-ch / Tdi / Epi-d / Ips / Tid/4Ibd / 0.37
She / 0.36 / 0.43
Sid / 0.32 / 0.35 / 0.74
Wat / 0.45 / 0.61 / 0.78 / 0.56
Sla / 0.54 / 0.52 / 0.59 / 0.53 / 0.65
Descy / 0.09 / 0.12 / 0.56 / 0.45 / 0.38 / 0.19
L&m / 0.49 / 0.50 / 0.69 / 0.70 / 0.67 / 0.79 / 0.44
Idg / 0.55 / 0.32 / 0.60 / 0.49 / 0.52 / 0.39 / 0.26 / 0.44
Cee / 0.62 / 0.65 / 0.78 / 0.66 / 0.85 / 0.72 / 0.43 / 0.82 / 0.62
Lobo / 0.43 / 0.41 / 0.41 / 0.22 / 0.49 / 0.25 / 0.13 / 0.23 / 0.25 / 0.49
Idp / 0.06 / 0.13 / 0.45 / 0.19 / 0.31 / 0.17 / 0.42 / 0.23 / 0.17 / 0.25 / 0.16
Di-ch / 0.13 / 0.19 / 0.10 / 0.14 / 0.14 / 0.28 / 0.06 / 0.29 / 0.14 / 0.23 / 0.02 / 0.02
Tdi / 0.60 / 0.74 / 0.38 / 0.34 / 0.55 / 0.61 / 0.02 / 0.44 / 0.40 / 0.62 / 0.48 / 0.06 / 0.14
Epi-d / 0.65 / 0.75 / 0.67 / 0.55 / 0.81 / 0.69 / 0.22 / 0.67 / 0.51 / 0.85 / 0.63 / 0.20 / 0.13 / 0.79
Ips / 0.52 / 0.67 / 0.83 / 0.67 / 0.83 / 0.73 / 0.45 / 0.81 / 0.62 / 0.92 / 0.40 / 0.35 / 0.21 / 0.56 / 0.81
Tid/4 / 0.65 / 0.67 / 0.53 / 0.56 / 0.61 / 0.68 / 0.15 / 0.66 / 0.49 / 0.76 / 0.40 / 0.16 / 0.16 / 0.78 / 0.75 / 0.72
ICMi / 0.64 / 0.72 / 0.69 / 0.65 / 0.75 / 0.75 / 0.27 / 0.77 / 0.58 / 0.88 / 0.43 / 0.24 / 0.19 / 0.74 / 0.84 / 0.89 / 0.95
Table 3.R2 values between different diatoms index in R-A2 dataset (Italy).
Idap / Ibd / She / Sid / Wat / Sla / Descy / L&m / Idg / Cee / Lobo / Idp / Di-ch / Tdi / Epi-d / Ips / Tid/4Ibd / 0.10
She / 0.16 / 0.28
Sid / 0.40 / 0.40 / 0.80
Wat / 0.33 / 0.18 / 0.39 / 0.58
Sla / 0.72 / 0.01 / 0.28 / 0.44 / 0.46
Descy / 0.21 / 0.12 / 0.45 / 0.49 / 0.26 / 0.13
L&m / 0.75 / 0.04 / 0.09 / 0.33 / 0.43 / 0.65 / 0.21
Idg / 0.15 / 0.15 / 0.06 / 0.10 / 0.07 / 0.11 / 0.07 / 0.01
Cee / 0.76 / 0.30 / 0.48 / 0.77 / 0.47 / 0.61 / 0.43 / 0.58 / 0.11
Lobo / 0.42 / 0.00 / 0.04 / 0.04 / 0.08 / 0.35 / 0.14 / 0.29 / 0.00 / 0.19
Idp / 0.04 / 0.36 / 0.65 / 0.59 / 0.17 / 0.07 / 0.53 / 0.03 / 0.00 / 0.29 / 0.00
Di-ch / 0.28 / 0.12 / 0.02 / 0.00 / 0.06 / 0.23 / 0.00 / 0.19 / 0.03 / 0.06 / 0.41 / 0.18
Tdi / 0.62 / 0.42 / 0.30 / 0.54 / 0.33 / 0.52 / 0.12 / 0.48 / 0.27 / 0.73 / 0.13 / 0.21 / 0.01
Epi-d / 0.62 / 0.27 / 0.11 / 0.37 / 0.29 / 0.49 / 0.02 / 0.39 / 0.44 / 0.56 / 0.07 / 0.05 / 0.03 / 0.75
Ips / 0.53 / 0.49 / 0.68 / 0.86 / 0.43 / 0.45 / 0.34 / 0.31 / 0.28 / 0.79 / 0.07 / 0.46 / 0.00 / 0.71 / 0.57
Tid/4 / 0.59 / 0.56 / 0.37 / 0.59 / 0.28 / 0.31 / 0.27 / 0.32 / 0.26 / 0.70 / 0.13 / 0.31 / 0.00 / 0.75 / 0.64 / 0.85
ICMi / 0.59 / 0.56 / 0.47 / 0.69 / 0.33 / 0.36 / 0.30 / 0.32 / 0.28 / 0.75 / 0.12 / 0.36 / 0.00 / 0.76 / 0.64 / 0.92 / 0.99
Alpine GIG – Rivers– Phytobenthos Intercalibration - Comparison – Final Report2007-05-31
Boundary derivation
The H/G and G/M boundaries will be possibly derived on the basis of the so called REFCOND approach as previously done for invertebrates. This means that the H/G boundary will be presumably fixed at the 25th percentile value of Reference samples and the G/M one will than be derived as the 25th percentile of reference values multiplied by 0.75. The boundaries for the National method will than be converted via regression into values of the Intercalibration Common Metric index.
Harmonization
If the National boundaries expressed in terms of Intercalibration Common Metric index do not fall within the band proposed by the GIG, they will be shifted up in order to fall within the band.
ANNEX 2
List of GIG-Contacts:
MS / Contact / emailAustria / Gisela Ofenböck
Franz Wagner /
France / Juliette Tison
Michel Coste /
Germany / Ilona Schlösser
Folker Fischer /
Italy / Laura Mancini
Andrea Buffagni /
Slovenia / Gorazd Urbanic
Bernarda Rotar /
Spain / Miriam Pardos
Concha Duran /
1