LAKES GIG NORTHERN BENTHIC INVERTEBRATE
Contents
1. Description of national assessment methods
Methods and required BQE parameters
Sampling and data processing
National reference conditions
National boundary setting
2. Results WFD compliance checking
3. Results IC Feasibility checking
4. IC dataset collected
5. Common benchmarking
6. IC approach
7. IC Results
8. Description of high and good status communities
Annex 1 Method descriptions
Annex 2 Comparison of littoral (ASPT) and whole lake (CPET) assessments
Annex 3 Continuation of the IC
1. Description of national assessment methods
MS / Method / StatusLake littoral acidification
SE / MILA: MultimetricInvertebrateLake Acidification index (lake acidification) / Finalized agreed national method
NO / MultiClear: Multimetric Invertebrate Index for ClearLakes (lake acidification) / Intercalibratable finalized national method
UK / LAMM (lake acidification) / Finalized agreed national method
Lake eutrophication
SE / BQI (profundal eutrophication) / Finalized agreed national method
FI / BQI (profundal eutrophication) / Finalized agreed national method
SE / ASPT: Average Score per Taxon (littoral eutrophication) / Finalized agreed national method
UK / CPET: Chironimidae Pupal Exuviae Technique (all-lake eutrophication) / Finalized agreed national method
Methods and required BQE parameters
Not all parameters included for all metrics, see below description and explanation why not all parameters are included.
FI: BQI (diversity not included)
SE: ASPT (relative abundance not included), BQI (diversity not included), MILA includes all parameters.
NO: Multiclear includes all parameters
UK: LAMM (diversity not included), CPET (diversity and abundance not included).
MS / Taxonomic composition / Abundance1 / Disturbance sensitive taxa / DiversityNorthern GIG Macroinvertebrates
FI / Benthic Quality Index (BQI) / RA included / BQI / Not included (see footnote)
SE / Average Score per Taxon / RA not included (see footnote) / Average Score per Taxon / Not included (see footnote)
SE / Benthic Quality Index / RA included / Benthic Quality Index / Not included (see footnote)
SE / MILA: Relative abundance (%) of Ephemeroptera
Relative abundance (%) Diptera
Relative abundance (%) of predators / RA included / British AWIC index / Number of mollusc taxa (Gastropoda)
Number of mayfly taxa
NO / MultiClear: Acidification indicator taxa
AWIC-family index
Adj. Henriksson and Medin’s index / RA included / Acidification indicator taxa
AWIC-family index
Adj. Henriksson and Medin’s index / Number of snail (Gastropoda) Number of mayfly (Ephemeroptera) taxa
UK / Lake Acidification Macroinvertebrate Metric / RA included / Lake Acidification Macroinvertebrate Metric / Not included (see footnote)
UK / CPET / RA not incl / CPET / Not included (see footnote)
1 RA: Relative abundance (abundance of single taxa or groups relative to the total abundance of macroinvertebrates or groups of macroinvertebrates)
Abundance was not directly used in the assessment methods
as this parameter is known to be highly variable in aquatic invertebrate communities (Resh, 1979; Barbour et al., 1992; Resh and Jackson, 1993; Johnson, 1998).
Invertebrate abundance was the least informative of ten metrics tested by Sandin and Johnson (2000) as it had the lowest effect size (a measure of the magnitude of impact) and highest spatial, temporal and sample variability.
Invertebrate abundance is rarely, if ever, used in ecological assessment due to the difficulties associated with detecting anthropogenic change with any degree of confidence (Osenberg et al., 1994)
Diversity : SE and FI
Some of the methods do not explicitly take diversity into account. Regarding eutrophication, the BQI indices of Sweden and Finland both take abundance and number of sensitive and insensitive taxa into account. The reason for not including a measure of diversity is that from a theoretical ecological point of view the diversity at low nutrient levels and medium-high nutrient levels should be similar, whereas at medium nutrient levels the diversity will actually be increased (a hump-shaped relationship). At very high nutrient levels, the diversity will of course decrease. Thus there is not a continuous (always going in the same direction) and/or linear pressure-response relationship and therefore diversity is not suitable for inclusion in a profundal eutrophication metric for benthic macroinvertebrates. See e.g. Tolonen (2005) who found a unimodal relationship between taxa richness and trophic gradient, possibly indicating that intermediate disturbance enhances species richness (Cornell and Lawton, 1992), and unpublished results on Swedish data (McGoff & Sandin). In addition, BQI explains a substantial amount of the whole natural macroinvertebrate community variation in lakeprofundal (see e.g. Jyväsjärvi et al. 2009).
Diversity : UK
The UK methods for Lakes do not include explicit diversity measures as the metrics that are included in the current metrics more than adequately describe the pressures. A diversity metric (or sub metric) would not improve the methods response to, or discrimination of pressure & status. Some testing of this has been done as part of the NGIG intercalibration work, where a large number of indicators were correlated with the pressures pH, and ANC. In these analyses explicit diversity measures very rarely came out with statistical significant relationships to the acidification pressure.
Sampling and data processing
Fin BQI: One occasion per sampling season: September to October. Six replicate samples are taken from the deepest point of lake.
SE BQI: One occasion per sampling season: September to October. Five Ekman samples are collected from a 100 m2 area in the middle of the lake (or over the deepest region).
SE MILA: One occasion per sampling season: September to November. Standardized Kick-sampling, SSEN-27828 (20 s x 1 m; 0,5 mesh; 5 replicates taken in autumn). Substratum is disturbed by kicking for 20 s and moving a distance upstream of 1 m
NO Multiclear: Preferably two occasions per sampling season: April to May and October to November
A sample consists of one or several sampling units taken from preferably one habitat type at the sampling site. The hand-net of 200-300 µm mesh-size is used as 'kick-net'. Sediments must be disturbed to a depth of 15 cm (where possible) depending on substrate compactness.
SE ASPT: One occasion per sampling season: September to November. Wind exposed hard bottom (stony) substrates. Standardized Kick-sampling. Substratum is disturbed by kicking for 20 s and moving a distance upstream of 1 m (20 s x 1 m; 0,5 mesh; 5 replicates taken in autumn).
UK LAMM: 2 samples are taken in each spring survey (March-June). One spring survey is enough for classification. However, 3 years worth of data will reduce uncertainty in classification.
To apply the method, invertebrates should be collected from a stony-bottomed section of the littoral zone of the lake with a depth of ? 75 cm. Two samples should be collected from each location sampled. Sampling should normally be undertaken between March and May. The invertebrates should be collected by disturbing the substratum with the feet ("kick sampling") and passing a hand net (nominal mesh size: 1 mm) through the water above the disturbed area. All habitats in the chosen sampling site should be sampled within a 3-minute period. In addition, a pre-sample sweep to collect surface dwelling invertebrates and a post sample manual search, lasting one minute, should be undertaken during which any invertebrates attached to submerged plant stems, stones, logs or other solid surfaces should be collected by hand and placed in the net.
See Annexes for details
National reference conditions
FI BQI:
Existing near-natural reference sites, 80 sites from the whole Finland. Data has been collected between 1992 and 2006.
Referencc criteria: No point source pollution, percentage of agriculture within catchment less than 15 %.
SE ASPT: Existing near-natural reference sites, ca 300, whole of Sweden.Use of pressure filter to identify reference conditions.
SE BQI: Existing near-natural reference sites, ca 110, whole of Sweden. Use of pressure filter to identify reference conditions.
SE MILA:Existing near-natural reference sites,ca 300, whole Sweden.
Reference conditions for MILA (Multimetric Index for Lake Acidification) indices were established using a pressure filter approach, i.e. lakes and streams judged to be perturbed using catchment land use and water chemistry (Johnson and Goedkoop 2007 (in Swedish)) were removed to isolate the gradient of interest. For example, to calibrate the response of MILA to acidity we excluded sites affected by pressures such as eutrophication, liming, urbanization etc to isolate the “acidity” gradient.
NOR MULTICLEAR: Existing near-natural reference sites, 7 lowland and boreal lakes belonging to the low alkalinity clear lakes in Southern Norway (South coast and Eastern parts), Rural areas in South-Eastern part of Norway (counties: Akershus, Hedmark), Southernmost part of Norway (county: Aust-Agder) and Mid-Norway (county: Sør-Trøndelag). Data from non impacted lakes, 2007-2009.
Pressure criteria: < 10% intensive agriculture, <1% artificial land use, < 10 p.e./km2 pop.dens., no acid load exceedance. Chemical criteria: ANC (Acid Neutralizing Capacity) > 30 µeq/L, pH > 6. Biological criteria
UK LAMM : Existing near-natural reference sites, Expert knowledge, Historical data, Modelling (extrapolating model results)
Number of sites: 8 sites for clear-water lakes, 6 for humic-water lakes
Geographical coverage: Representative lakes throughout UK at risk from acidification
Location of sites: Representative lakes throughout UK at risk from acidification
Data time period: Historical data from 2005-2008
Reference sites screened using the Damage matrix. See table 6.1 in /'Macroinvertebrate Classification Diagnostic Tool Development/' SNIFFER Report WFD60. This matrix assesses sites based on their Acid Neutralising Capacity (ANC) in relation to Ca content.
National Boundary setting
The description of national boundary setting procedures in detail, graphs showing dose-relationships and description of high, good and moderate communities are to be found in Annex 1.
FI BQI:Boundaries are derived as follows: H/G = 0.75, G/M = 0.60, M/P = 0.30, and P/B = 0.10
SE ASPT:Equidistant division of the EQR gradient.
SE BQI:Equidistant division of the EQR gradient.
SE MILA:
-The reference value was defined as the median MILA index value of unperturbed sites stratified by type (here defined simply by ecoregion).
-EQR values for the reference population were calculated as observed value divided by reference (established by typology) value.
-The borderline between High and Good status was defined as the 25th percentile of the distribution of the reference data.
-A threshold approach was used for setting the Good/Moderate boundary. EQR MILA values normalized for ecoregion differences were regressed against mean annual pH and the intercept at pH 5.6 was used as the borderline between Good and Moderate quality. A pH value of 5.6 was selected since many previous studies have shown marked changes in fish and invertebrate assemblages at this threshold (e.g. Johnson et al. 2007). In addition, variability in the regression supports this threshold; variance appears to collapse (funnel shaped response) at around pH below 6.0 (Fig. 1).
-The remaining class boundaries were set using equidistance.
Figure. 1. EQR values of MILA regressed against mean annual pH. The different colors reflect the three main ecoregions (regions14, 22 and 20), the different symbols show reference (crosses) and putative acidified (circles).
NOR MULTICLEAR
The H/G boundary: assigned to represent the lower 5th percentile of scores for all reference sites (i.e. that 95 % of all sites identified as reference sites are assigned to high ecological status). At present this value represent all reference sites since the number of reference sites are so few (N=7).
The H/G boundary value on the MultiClear scale has been set to 4.0 (absolute value) representing an EQR = 0.95 (table 3).
The G/M boundary and the subsequent boundaries: The boundaries G/M, M/P and P/B are based on the exponential relationship between MultiClear and AcidIndex1 (Forsuringsindeks 1; see which in turn represent an equidistant division of the subsequent EQR gradient (from Good to Bad). The reason for this approach is that 1) the relationship between AcidIndex1 and MultiClear is clear and strong (R2 = 0.95), 2) the relationship between acidification and changes in AcidIndex1 are also clear and strong; R2 = 0.6 with pH as the predictor variable (table 2), 3) AcidIndex1 has been widely used in Norway for more than 20 years and proven reliable, and 4) the borders between the categories based on AcidIndex1 are easily defined and based on changes in ecosystem structure in accordance with the normative definitions by the Water Framework Directive (WFD; 2000/60/EC). AcidIndex1 is a very simple index based on the presence or absence of selected indicator taxa (see annex1 in Veileder 01:2009 from Direktoratsgruppa Vanndirektivet; ) assigned as very tolerant (score=0), slightly sensitive (score=0.25), moderately sensitive (score=0.5) and highly sensitive to acidification (score=1). The value of the index varies between 0 and 1. A value of zero means that (slightly, moderately or highly) acid sensitive macroinvertebrates are absent. A value of one means that at least one specimen of the most (highly) acid-sensitive taxa is present. However, a score of one based on a single individual from one sample may constitute an unreliable result. Therefore, in the Norwegian assessment method, assigning the ecological status is based on mean index values calculated from at least four, preferably more samples, including both spring and autumn samples.
For AcidIndex1 the G/M boundary has been set to 0.75 (absolute value). Based on the relationship between AcidIndex1 and MultiClear, the G/M boundary value on the MultiClear scale has been set to 3.13 (absolute value) representing an EQR = 0.74 (table 3).
Figure 2. MultiClear vs AcidIndex1 (RADDUM1) for Norwegian low alkalinity, clear lakes (non-linear regression, N=15).
UK LAMM :
Using discontinuities in the relationship of anthropogenic pressure and the biological response.
Where discontinuities could not be found then partitioning based on the Damage Matrix was used.
Detailed description of boundary setting procedure, pressure-response relationship and communities at high, good and moderate status is given in McFarland et al. (2009).
Distinct discontinuities along the ANC pressure gradient were only found at humic sites at ANC 23 µeq/l to derive a good-moderate boundary. These were consistent using pressure metrics (e.g. LAMM), diversity measures (e.g. Shannon) and functional groups (e.g. grazers). Where no consistent breakpoints/step-changes were found, sites were grouped by the damage matrix according to class. The mean LAMM scores of the two adjacent classes were then added together and divided by two to form the boundary.
Conclusions
Finland:
- Based on deviation from reference condition (H/G-boundary = 75 % of reference value)
- statistical not equidistant division of the EQR gradient.
Norway MultiClear:
- The HG boundary was identified as the lower 5th percentile of scores of all reference sites (due to low number of reference sites this equals to the whole reference population; small adjustments may be necessary when more data is available).
- GM = the point where only 50 % of the samples from a site contain very sensitive taxa and the remaining 50 % of the samples contain moderately sensitive taxa.
Sweden: Based on a statistical division of the EQR gradient (equidistant).
UK: CPET: Using paired metrics (relative abundance of sensitive and tolerant taxa) that respond in different ways to the influence of the pressure
UK: LAMM
- Using discontinuities in the relationship of anthropogenic pressure (ANC) and the biological response (LAMM, diversity measures and functional groups).
- Where discontinuities could not be found then partitioning based on the Damage Matrix was used.
2. Results WFD compliance checking
Lake littoral acidification:
SE, UK, NO national assessment methods comply with requirements of WFD.
IE and FI have little acidification pressure / data and do not have national methods. In FI, humic lakes can be acidic, but this is a natural phenomenon in boreal peatlands.
Summary:
-Three countries compliant (SE, UK, NO).
-Two countries have no data (IE, FI).
-Several of the methods do not include either the parameter abundance, or the parameter diversity (explanations above, considered complianr).
Lake eutrophication – profundal:
SE has compliant national methods using profundal (BQI) invertebrates.
FI have a compliant method using profundal invertebrates (BQI).
The BQI methods do not include the parameter diversity per se. (explanations above).
Lake eutrophication – littoral:
SE has compliant national methods using littoral (ASPT) invertebrates.
Lake eutrophication – the whole lake:
UK has a compliant method using chironomid exuviae (CPET).
IE and NO do not have any national method for assessment of lake eutrophication.
Summary:
-Three countries compliant - SE (ASPT and BQI), FI (BQI), UK (CPET).
-IE has no method. NO has no method / data.
Compliance criteria / Compliance checking conclusions- Ecological status is classified by one of five classes (high, good, moderate, poor and bad).
UK; LAMM for clear waters has 4 classes, poor/bad combined. LAMM for humic waters has three classes, moderate, poor, bad combined. WFD-AWICsp and CPET have all 5 classes.
Finland; Yes, BQI has 5 classes.
Norway: Yes, all metrics have 5 classes.
- High, good and moderate ecological status are set in line with the WFD’s normative definitions (Boundary setting procedure)
- All relevant parameters indicative of the biological quality element are covered (see Table 1 in the IC Guidance). A combination rule to combine para-meter 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.
- 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
UK: This is true for LAMM. For WFD-AWICsp the typology is based on Scottish humic and clear waters (cutoff at 10 mg/l) and a Welsh/English typology. CPET is site specific.
FI: yes
NO: yes
- The water body is assessed against type-specific near-natural reference conditions
UK: yes
FI: Yes. Lakes are assessed against near-natural reference conditions where expected (reference) values for BQI are derived with a regression model for each site.
NO: yes
- Assessment results are expressed as EQRs
UK: yes
FI: yes
NO: yes
- Sampling procedure allows for representative information about water body quality/ ecological status in space and time
UK: yes
FI: yes
NO: yes
- All data relevant for assessing the biological parameters specified in the WFD’s normative definitions are covered by the sampling procedure
UK: yes
FI: yes
NO: yes
- Selected taxonomic level achieves adequate confidence and precision in classification
UK: yes
FI: yes
NO: yes
2. Results IC Feasibility checking