Classifying aquatic macrophytes as indicators of eutrophication in European lakes

Aquat Ecol (2008) 42:237–251 DOI 10.1007/s10452-008-9182-y Classifying aquatic macrophytes as indicators of eutrophication in European lakes W. Ellis Penning Æ Marit Mjelde Æ Bernard Dudley Æ Seppo Hellsten Æ Jenica Hanganu Æ Agnieszka Kolada Æ Marcel van den Berg Æ Sandra Poikane Æ Geoff Phillips Æ Nigel Willby Æ Frauke Ecke Published online: 20 April 2008 Ó Springer Science+Business Media B.V. 2008 Abstract Aquatic macrophytes are one of the biological quality elements in the Water Framework Directive (WFD) for which status assessments must be defined. We tested two methods to classify macrophyte species and their response to eutrophication pressure: one based on percentiles of occurrence along a phosphorous gradient and another based on trophic ranking of species using Canonical Correspondence Analyses in the ranking procedure. The methods were tested at Europe-wide, regional and national scale as well as by alkalinity category, using 1,147 lakes from 12 European states. The grouping of species as W. E. Penning (&) Deltares, P.O. Box 177, 2600 MH Delft, The Netherlands e-mail: ellis.penning@deltaRes.nl sensitive, tolerant or indifferent to eutrophication was evaluated for some taxa, such as the sensitive Chara spp. and the large isoetids, by analysing the (nonlinear) response curve along a phosphorous gradient. These thresholds revealed in these response curves can be used to set boundaries among different ecological status classes. In total 48 taxa out of 114 taxa were classified identically regardless of dataset or classification method. These taxa can be considered the most consistent and reliable indicators of sensitivity or tolerance to eutrophication at European scale. Although the general response of well known indicator species seems to hold, there are many species that were evaluated differently according to the database A. Kolada Institute for Environmental Protection, Warswawa, Poland W. E. Penning NIOO-Centre for Limnology, Publication 4296 NIOO-KNAW, P.O. Box 1299, 3600 BG Maarssen, The Netherlands M. van den Berg Rijkswaterstaat RIZA, P.O. Box 17, 8200 AA Lelystad, The Netherlands M. Mjelde NIVA, Gaustadalléen 21, 0349 Oslo, Norway S. Poikane Joint Research Centre, 21020 Ispra, Italy B. Dudley CEH, Edinburgh, Bush Estate, EH26 0QB Penicuik, UK G. Phillips Environment Agency for England and Wales, RG1 8DQ Reading, UK S. Hellsten SYKE, University of Oulu, P.O.Box 413, 90014 Oulu, Finland N. Willby University of Stirling, FK9 4LA Stirling, UK J. Hanganu DDNI, Tulcea, Romania F. Ecke Luleå University of Technology, 971 87 Luleå, Sweden 123 238 selection and classification methods. This hampers a Europe-wide comparison of classified species lists as used for the status assessment within the WFD implementation process. Keywords Aquatic vegetation
Indicators
Species classification
REBECCA
EU Water Framework Directive Introduction For implementing the Water Framework Directive (WFD) it is important to identify how aquatic macrophytes can be used to indicate pressures. The WFD uses aquatic macrophyte species composition and abundance as a biological quality element for ecological assessment. In order to ensure consistency in application of the directive, lakes with comparable macrophyte composition and abundance must be assigned a comparable ecological status. Many previous reports on the use of macrophytes as indicators of eutrophication in lakes are confined to single countries (e.g. Palmer et al. 1992), or to relatively small geographical regions within Europe (e.g. Rørslett 1991). Other approaches focus only on selected groups of macrophytes, such as isoetids (e.g. Murphy et al. 1990). It is clear that macrophytes are confined to specific habitats (e.g. Schaminee et al. 1995; Murphy 2002) and that they react to many different changes in their optimal habitat (e.g. Barko et al. 1986; Van Geest 2005). It is therefore often suggested that macrophytes can be used as indicators for pressures such as eutrophication. In the most simplistic sense, certain macrophyte species are present in waters of good status and absent in bad status (e.g. Moss et al. 2003; Birk et al. 2006). All EU member states are currently undertaking efforts to classify species and creating status assessment methods for lakes based on macrophytes, of which some are already available; Germany (Stelzer et al. 2005; Schaumburg et al. 2004), Denmark (Søndergaard et al. 2005), the Netherlands (Van den Berg 2004), Belgium (Leyssen et al. 2005), Sweden (Ecke 2006), Finland (Leka et al. 2007) and Ireland (Free et al. 2006), or in press; UK (Willby et al. 2006) and Norway (Mjelde 2007, unpublished). In the Netherlands, classifying species is based on expert judgement by a group of ca. 30 experts, where not only species occurrence is classified according to 123 Aquat Ecol (2008) 42:237–251 a general ‘ecosystem disturbance’ for specific lake types but also the role that abundance of species plays in their classification. For example, low abundances of Lemna spp. are not assessed negatively, while high abundances of Lemna spp. are. Germany also bases its species classification system on expert judgement, while UK, Sweden and Norway base it on data analyses along a total phosphorous (TP) gradient. Expert judgement and valuation of species sensitivity have proved to be highly country dependent, as was shown in the WFD intercalibration process (Lyche Solheim 2005). Therefore, we aim to describe the classification of sensitive and tolerant species on a Europe-wide scale. Comparing classification or status assessment methods from individual member states has been completed for rivers (Birk et al. 2006), but it has not yet been carried out on a large scale for lakes, although an initial assessment of individual member state datasets show that there are large differences among resulting status descriptions (Tóth et al. 2008, this issue). One of the main problems in performing a Europe-wide assessment of macrophyte responses to pressures is the lack of a Europe-wide homogenous monitoring methodology, although attempts are being made to harmonise the methodology using a common CEN standard or lake habitat survey (CEN 2006). In this article we use and discuss the results of two different methods to classify aquatic macrophytes according to their response to a eutrophication pressure represented by the summer average total phosphorous concentration. The results are part of a study aimed to determine the usability of macrophytes as indicators of eutrophication pressures throughout Europe, to examine the effects of scale and typology when classifying macrophyte species into eutrophication response classes, and to examine the effects of different methods on the final ecological quality assessment (see also Penning et al. 2008, this issue). Methods Data used Data on macrophyte abundance and water quality were collected for as many European countries as possible. In total, data from 12 countries and 1,147 lakes are included in the REBECCA macrophyte Aquat Ecol (2008) 42:237–251 239 database. For some lakes samples of more years are available, resulting in an overall of 1,442 data points (Table 1; see Moe et al. 2008, this issue for further description on the database). Data providers assigned type and status (reference or non-reference lake) to most of the lakes provided to the database. Only true aquatic macrophyte, of the growth forms isoetids, elodeids, nymphaeids, lemnids and charids were included (Mäkirinta 1978). Helophytes (emergent aquatic plants) were excluded from the analyses as their response to eutrophication is obscured by soil trophic characteristics, exposure, shoreline management and their ability to emerge from the water phase. Species for which it was unclear whether they represented the helophyte or aquatic form (e.g. Juncus bulbosus, Sagittaria sagittifolia and Hippuris vulgaris) were excluded from the analyses. Species occurring in less than 4 lakes were included, but were not given a response-class (i.e. sensitive, tolerant or indifferent to eutrophication). Macrophyte data were available at species level, with some exceptions of poorly identified groups, notably Chara, for which most surveys grouped species as Chara sp. Table 1 Overview of content of the total REBECCA macrophyte database, showing the number of lakes from each country Country Belgium Estonia Total Typed Ref. Chl. a TotP Secchi 2 2 4 0 2 1 2 2 4 4 0 3 Finland 527 345 154 179 Ireland 117 18 40 117 117 106 Latvia 144 144 15 144 137 144 Lithuania Netherlands Norway Poland 396 459 5 5 2 5 4 5 46 46 34 46 46 46 269 60 168 74 269 74 5 5 2 5 5 5 Romania 19 17 8 18 17 17 Sweden 254 0 101 0 250 92 50 33 13 50 50 0 1442 679 537 643 United Kingdom Total 1294 954 Total, total number of lakes; Typed, number of lakes for which there is an assigned intercalibration type; Reference, number of lakes defined as reference lakes for the intercalibration exercise. The number of lakes in which the environmental variables Chlorophyll a (lg l-1), total P (lg l-1) and Secchi depth (m) was available is also shown. Not all data has been used for the reported analyses The main water quality parameters collected were concentrations of total phosphorus and chlorophyll a and Secchi depth. For most data points these are summer-average values, but in some cases the origin of the values is unknown, and these might be single sample values. In the Finnish lakes TP is given as a 10year average. Information on lake typology, specifically size, mean depth, colour, alkalinity and altitude, was also collected for each lake. For some lakes, this information was only available on a nominal scale, conforming to the WFD intercalibration types. The Geographic Intercalibration Group (GIG) typology proposed for the intercalibration exercise (Heiskanen et al. 2004) is based on lake size, alkalinity, depth and, depending on lake type, altitude and humic content. The typology was used in the structuring of the database and subsequent analyses. The REBECCA macrophyte database contains partly incomplete data, as data were provided by various sources within individual member states. Not all data provided to the REBECCA database could therefore be used in the analyses that are here reported. Further discussion on the REBECCA macrophyte database, from a more technical standpoint can be found in Moe et al. (2008, this issue). Classification of species We tested species classifications based on the Percentile method (Mjelde 2007) and the LEAFPACS method (Willby et al. 2006), on Europe-wide, regional and national species lists. For identifying sensitive and tolerant species by Percentiles, we have used species presence or absence along a lake water total phosphorous gradient. This approach uses the 75th percentile of the phosphorous score to distinguish sensitive and tolerant species, and the 25th percentile to distinguish indifferent and tolerant species. The following description of sensitive, tolerant and indifferent species was applied: 1. Sensitive species: species which are most apparent or only appear in reference lakes, often in high abundance. Frequency and abundance decreases (and often the species disappears entirely) with increased eutrophication pressure; the 75th percentile of observations is chosen as the phosphorous limit above which a species is no longer considered sensitive. 123 240 2. 3. Aquat Ecol (2008) 42:237–251 Tolerant species: species occurring at an increased frequency and abundance at higher eutrophication pressure. Often rare or with lower abundance in reference lakes; the 75th percentile of observations lies above the upper phosphorous limit and the 25th percentile lies above the lower phosphorous limit. Indifferent species: species with wide preference, common not only in reference lakes but also in slightly eutrophic lakes; the 75th percentile of observations is more than the upper phosphorous limit and the 25th percentile is less than the lower phosphorous limit. These species disappear in hypertrophic lakes. The LEAFPACS method is a more elaborate statistical approach for classifying species distributions. The method uses multivariate statistical methods to assign response-group classifications to aquatic macrophytes, and was designed especially to provide a relatively objective method for defining biological quality-element classifications for the EU Water Framework Directive. The method consists of two separate parts; the first part assigns a trophic rank to each macrophyte species, based on the global dataset, and the second part assigns a lake-type specific response group classification to each species. The first part of the analysis could easily be replaced by expert judgement based assessment of species trophic status, such as that produced for the UK by Hill and Ellenberg (1999). For the REBECCA project, it was decided that the dataset would best provide this trophic rank because the dataset had such a wide geographical distribution and there was no single expert-judgment system that covered the region known to the authors. The second part of the method followed that of the LEAFPACS project (Willby et al. 2006). In summary, lake trophic rank scores were calculated for each lake as the average of the species trophic ranks of all species occurring in that lake. These lake trophic ranks were then used as the environmental variable in a further CCA analysis. The axis-1 and tolerance values produced by this analysis were used to define the response class of each species as follows: – – as a ‘positive responder’ if its axis-1 score was less than or equal to zero. as a ‘negative responder’ if its axis-1 score was greater than zero, and the result of subtracting its 123 – tolerance from its axis-1 score was less than or equal to zero. as a ‘strongly negative responder’ if the result of subtracting its tolerance from its axis-1 score was greater than zero. The following deviations from the LEAFPACS method should be noted: site scores were calculated as the average of trophic rank (TR) scores of species occurring in each sample, only presence/absence data were used (no abundance data were used for weighting the site scores). We have not assigned any response groups to species occurring in\4 samples in a particular lake type. The percentile method was calculated for the all data points in the whole REBECCA dataset, where TP values were available (n = 1294), for Northern and Central European Geographical Intercalibration Groups (GIG: as defined for the intercalibration exercise according to Heiskanen et al. 2004), lowmoderate alkalinity (\1 meq l-1) and high alkalinity ([1 meq l-1) lakes within the GIGs, in addition to Norway, Finland and Latvia separately. The LEAFPACS classification was calculated separately for lowmoderate and high alkalinity lakes within the two GIGs. Results In total the REBECCA database consists of 114 aquatic macrophyte taxa, comprising 105 species and 9 hybrids. The most common species in the database, found in more than 30% of the lakes, were Nuphar lutea, Potamogeton natans, P. perfoliatus, Myriophyllum alterniflorum, Nymphaea alba, Isoetes lacustris, Ranunculus reptans, Utricularia vulgaris, Sparganium angustifolium, Isoetes echinospora and Eleocharis acicularis. All these species are typical for low-moderate alkalinity, oligo-mesotrophic lakes (e.g. Murphy 2002), which reflect a dataset dominated by lakes from Northern Europe, although some of them are also found in more eutrophic lakes (e.g. Van den Berg 2004) When comparing the whole dataset (percentile method) with the list for the Central GIG, 71% of the species were classified to the same group while 2% were diversely classified as tolerant or sensitive (see Appendix). Twenty-seven percent of the species showed small discrepancy (categorised as Aquat Ecol (2008) 42:237–251 241 the percentiles method and tolerant in the LEAFPACS method. Six of the species classified as being indifferent in the percentile method were included as tolerant species in the LEAFPACS method. For the Central GIG high alkalinity lakes the two methods gave similar results, 62% were identically classified, while 2 species, Najas marina and Nitella flexilis, got opposing classifications. Twelve of the indifferent species were classified as sensitive in the LEAFPACS method, and five as tolerant. For the Northern GIG high alkalinity lakes, 70% of the species received the same classification while 12% had an opposing classification. The remaining 12% were variably classified as indifferent, tolerant or sensitive. For the low-moderate alkalinity lakes only 48% of the species had the same classification, while for 31% the classification was contradictory. Especially many elodeids like numerous Potamogeton species were not classified concordantly. There is a clear positive relationship between total P and Secchi depth for the Northern GIG data (r2 = 0.63), but not for Central GIG data (r2 = 0.24). For the individual countries with sufficient data r2 are indifferent-tolerant or sensitive-indifferent). These were mainly species with marked lower abundance in the Central GIG compared to the Northern GIG lakes. The same comparison with the Northern GIG list showed 77% species with a similar classification, while none had a different classification. The percentage of species classified as sensitive differed among countries: 55%, 48% and 35%, respectively, for Norway, Finland and Latvia, which likely reflects the amount of impacted lakes compared to reference lakes in the datasets from these countries. A comparison between the two different classification methods (percentile and LEAFPACS) were made for Central GIG, high and low-moderate alkalinity lakes, and for Northern GIG, high and low-moderate alkalinity lakes (Fig. 1). For the Central GIG lowmoderate alkalinity lakes, the two tested methods gave similar results and 72% of the species had the same classification. Only Chara globularis and Nitella flexilis had opposing classifications, meaning that one method classified them as sensitive while the other as tolerant. They were both ranked as sensitive species in a b Northern GIG Low-moderate alkalinity lakes Central GIG Low-moderate alkalinity lakes percentiles ISOETIDS no. of species percentiles LEAFPACS 30 ELODEIDS NYMPHAEIDS LEMNIDS ISOETIDS CHARIDS 25 10 20 8 15 6 10 4 5 2 0 0 c Northern GIG High alkalinity lakes LEAFPACS 12 d percentiles ELODEIDS NYMPHAEIDS LEMNIDS CHARIDS Central GIG High alkalinity lakes percentiles LEAFPACS LEAFPACS 25 14 ISOETIDS ELODEIDS NYMPHAEIDS LEMNIDS CHARIDS ISOETIDS ELODEIDS NYMPHAEIDS LEMNIDS CHARIDS 12 20 no. of species 10 15 8 6 10 4 5 2 0 Fig. 1 Comparison between percentile method and LEAFPACS method for Northern GIG lakes (a, c) and Central GIG lakes (b, d). The negative and strongly negative responders (LEAFPACS tol indiff tol sens indiff tol sens indiff tol sens indiff tol sens indiff sens tol sens indiff tol indiff sens tol indiff tol sens indiff tol sens indiff sens 0 method) are merged and treated as sensitive species, while the positive responders are treated as tolerant species 123 242 Aquat Ecol (2008) 42:237–251 0.63, 0.79, and 0.41 for Finland, Norway and Latvia, respectively. Based on all species classification analysis (Appendix) we obtained 48 species with identical classification, which can be regarded as the European list of aquatic macrophyte species that were more robustly defined as sensitive or tolerant against eutrophication (Table 2). The lists include all the large isoetids, and some Chara spp., among the sensitive species, while most lemnids are included among the tolerant species. The remaining species are partly indifferent to eutrophication pressure, but also include species missing or rare in some regions or countries. Table 2 List of overall European sensitive and tolerant species based on the comparison of different methods and datasets to classify species as tolerant or sensitive Overall European sensitive species Overall European tolerant species Eleocharis acicularis Callitriche cophocarpa Isoetes echinospora I. lacustris Ceratophyllum demersum Elodea nuttallii Littorella uniflora Myriophyllum verticillatum Lobelia dortmanna Potamogeton crispus Ranunculus reptans P. obtusifolius Subularia aquatica P. pectinatus Callitriche hamulata P. pusillus Myriophyllum alterniflorum P. trichoides M. sibiricum Ranunculus circinatus Potamogeton filiformis Zannichellia palustris P. polygonifolius Nymphoides peltata P. 9 nitens Sagittaria sagittifolia 9 natans P. 9 zizii Hydrocharis morsus-ranae Ranunculus confervoides Lemna minor R. peltatus Lemna trisulca Utricularia australis Salvinia natans U. intermedia U. minor Spirodela polyrrhiza Stratiotes aloides U. ochroleuca Trapha natans Nuphar lutea 9 pumila Nitellopsis obtusa Sparganium angustifolium S. hyperboreum Chara delicatula C. rudis C. strigosa Tolypella canadensis The species listed are those with the same classification regardless of classification method or dataset 123 The percentiles of observations used to define phosphorus levels separating sensitive, tolerant and indifferent species are based on expert judgement. However, we assume that the reference species Isoetes lacustris and Lobelia dortmanna in low-moderate alkalinity lakes, and Chara spp., in high alkalinity lakes, should have very low abundance or be completely absent in impacted lakes. For example, for Norway this gives a distinction between sensitive and tolerant species based on the 75th percentile at around 20 lg l-1 TP (Fig. 2), while for Latvia the 75th percentile boundary lines around 60 lg l-1 TP (Table 3). These thresholds can be evaluated by visual interpretation (Fig. 3a), or by use of more sophisticated statistical techniques such as quantile regressions (Fig. 3b). In a visual interpretation, the concentration is identified where a specific species or selected group of species, such as large isoetids, notably declines in total abundance along a total phosphorous gradient. Alternatively, this could also be an average value estimated to be the middle point of a gradient of decline. Using quantile regressions, a more objective line defining thresholds based on the quantiles below or above a certain percentile of the observations can be identified. For both these methods of threshold detection a clear drop in abundance of large isoetids in Northern lakes can be seen around 20 lg l-1 TP, and above 50 lg l-1 TP their abundance becomes very low. Discussion In essence, the LEAFPACS method and the percentile method are based on a similar idea of identifying groups of species as tolerant or sensitive to a eutrophication pressure. However, the results of the classification are not easy to compare as the percentile method distinguished a group of indifferent species that is lacking in the LEAFPACS method. Only species that are classified as tolerant and sensitive in both methods can therefore be used for the comparison of the results. Our suggested list of overall European sensitive and tolerant species consists of 48 taxa, which strongly agrees with earlier knowledge and expert judgement (e.g. Palmer et al. 1992; Ecke 2006; Rørslett 1991; Kurimo 1970; Pot 2003; Murphy 2002). The characterisation of the large and perennial isoetids as sensitive to eutrophication in low-moderate Aquat Ecol (2008) 42:237–251 243 1000 total P 100 10 TOLY CAN SPAR HYP UTRI OCH LOBE DOR UTRI INT JUNC BUL LITT UNI ISOE LAC ISOE ECH UTRI MIN SPAR NAT NITE OPA SUBU AQU RANU PEL RANU RPT POTA POL LIMO AQU CRAS AQU CHAR STR RANU CON MYRI SIB MYRI ALT CALL PAL ELEO ACI CALL HER NUPH PUM SPAR ANG CHAR GLO POTAXNIT POTA GRA CHAR CON UTRI VUL CALL HAM POTA FIL POTA PRA CHAR DEL CHAR RUD LYTH POR ELAT HYD POTA LUC CHAR ASP HIPP VUL POTA ALP NUPH LUT SAGI SFO POTA PER NYMP ALB ELOD CAN POTA BER POTA NAT CALL STA PERS AMP POTA FRI RANU AQU POTA CRI POTA PUS POTA RUT ELAT TRI SPAR EME POTA PEC LEMN TRI CALL COP CHAR TOM POTA OBT MYRI VER CERE DEM MYRI SPI LEMN MIN SPIR POL 1 Fig. 2 The percentile method used for Norwegian lakes in which the grey points represent the 25th and 75th percentile and the red dots the 50th percentile. Sensitive species: 75th percentile less than 20 lg l-1 TP, Tolerant species: 75th percentile more than 20 and 25th percentile more than 10 lg l-1, Indifferent species: 75th percentile more than 20 lg l-1 and 25th percentile less than 10 lg l-1. Species recorded in less than four localities are excluded in the graph. Upper and lower phosphorus limits for the definition of sensitive and tolerant species are based on expert judgement (from Mjelde 2007). Vertical grey lines separate the sensitive, indifferent and tolerant species groups according to these limits alkalinity lakes seems unambiguous. However, only three Chara species are included among the sensitive species. This may be because of the heterogeneous character of the high alkalinity lakes. In addition, many lake surveys included in the REBECCA database included several Chara spp. species merged, which make the discussion about the classification of the Chara species difficult. Merging datasets from different countries results in classifications that differ with those based on countrywise judgement. For example, for Norway and Finland Lobelia dortmanna is characterised as a typical sensitive species, which does not appear in eutrophic waters. This species is missing from all Finnish and Norwegian high alkalinity lakes. In the Swedish dataset, this species exists in eight high alkalinity lakes, with total P varying Table 3 Suggested limits of total P (lg l-1) used for classifying aquatic macrophytes based on 25th and 75th percentiles for the overall REBECCA database, total Central and Nordic GIG; high and low alkalinity within the GIGs and for three individual countries All Central GIG Tot. High Nordic GIG Low Latvia Tot. High Low Norway Finland Sensitive species 75th less than 30 60 60 26 60 30 30 30 20 30 Tolerant species 75th more than 30 60 60 26 60 30 30 30 20 30 25th more than 18 25 25 15 20 15 15 15 10 15 Indifferent species 75th more than 30 60 60 26 60 30 30 30 20 30 25th less than No of lakes 18 25 25 15 20 15 15 15 10 15 1294 231 151 40 137 932 106 751 269 369 High, High alkalinity; Low, Low-moderate alkalinity 123 244 a Large Isoetids 16 14 12 summed abundance Fig. 3 (a) Sudden drop for large isoetids by visual interpretation. The y-axis gives the summed abundance of all large isoetid species occurring within a single lake recording. On the x-axis the summer average total P (lg l-1) recorded for each lake sample. (b) Quantile regression for large isoetids in low-moderate alkalinity Nordic lakes. The y-axis shows a cumulative frequency index, which means sum of the semiquantitative scores of all isoetid species at a given site Aquat Ecol (2008) 42:237–251 10 8 6 4 2 0 1 100 10 1000 Total Phosphorous (ug/l) 14 b 8 6 4 0 2 cumulative frequency index 10 12 0.95 0.9 0.75 0.5 2 5 10 20 50 100 200 total phosphorus, µg/l between 60 and 80 lg l-1. These lakes may be shallow or very shallow allowing sensitive species to survive and grow despite low light conditions. Unfortunately, depth information at the sample locations was not available to test this hypothesis. However, such inequality can give ‘incorrect’ classifications and a list based on the whole GIG dataset will not be transferable to the individual country. Country-based lists provide the most consistent results, as they include data from more homogenous monitoring methods. Cluster analyses showed that potential other explanations for this country based 123 homogeneity of data, such as spatial proximity, were not the main reason, as in that case countries such as Norway and Sweden should be closer together and Romania and the Netherlands further apart (see Lyche Solheim 2006). In some situations the full pressure gradient was not represented within the dataset, especially for specific lake types this was apparent. Extending the dataset with dedicated samples from the full gradient would help to reflect a truer optimum for many species. For this the cooperation between e.g. neighbouring countries might be necessary. Aquat Ecol (2008) 42:237–251 Some of the classified tolerant species (and positive responders) are not able to grow in hypertrophic conditions. However, some tolerant species are considered valuable indicators by experts in Central GIG for specific situations (e.g. Nitellopsis obtusa). Their absence can be indicative of hypertrophic situations (van den Berg 2004). This is the last step before complete disappearance of aquatic macrophytes and a transition to a cyanobacteria dominance state. The here described classification methods do not distinguish such species as sensitive to hypertrophic situation. Ideally, a classification system including all aquatic macrophytes would probably be made separately for each individual lake type, or separated by alkalinity and humic content for each country. However, this could not be tested here since the subsets of data from the overall 1,147 lakes in the REBECCA database were not always sufficiently large ([20) to assess individual lake types. In our dataset, combined data on light measurements and chlorophyll data was lacking for many lakes. Therefore, we used TP as an index for eutrophication (and indirectly for light conditions). Reduced light condition caused by increased phytoplankton biomass is probably the most important effect of eutrophication on aquatic macrophytes (Kirk 1994). Species sensitive to eutrophication are also often species that are less tolerant to poor light conditions which can be not only due to eutrophication but also due to increased water column suspended sediments (e.g. in association with wind induced sediment resuspension) (James et al. 2004). Unfortunately information on exposure, depth at sampling location and maximum depth of colonisation was not available to adequately assess the response of macrophytes to Secchi depth. There is a large variation around the average response to eutrophication expressed as a change in total P concentration. If inorganic or humic substances are the main factors for poor light conditions, it is less obvious that total phosphorus can be used as an indicator for the effect of eutrophication. In humic lakes, total P is partly bound to organic particles unavailable for biomass production, resulting in a non-linear relationship between total P and light (Lamers et al. 2002). In lakes highly influenced by inorganic substances resuspended from the sediment the relationship between total P and light conditions will also be non-linear. In addition, many very shallow lakes, such as the Dutch border lakes and various lakes from the Danube delta (having large surface areas, and\3 m deep) can have acceptable light 245 conditions despite relatively high phosphorus levels (Coops et al. 2007; Covaliov et al. 2003). Based on this, classification systems and ecological status assessments based on species occurrence along the phosphorus gradient alone will have limited value in the very shallow lakes such as many of the high alkaline lakes in Central GIG. However, the Northern GIG dataset is dominated by deeper lakes of low colour where variation in light conditions depends mainly on changes in phytoplankton biomass. Therefore, classification systems based on sensitive or tolerant species seem highly valuable for the lakes in the Northern GIG. The data collected within the REBECCA project macrophyte database showed general trends in response to eutrophication pressure, but variation of responses of individual macrophyte species with respect to their classification in line with the pressure gradient throughout Europe and throughout lake types is high. This might be partly due to the fact that the total P data used now represent on average a summeraverage value in most cases, but not in all (e.g. in Finnish lakes a 10-year summer average is used) and also seasonal changes in values are not available in this dataset. Gibson et al. (1996) indicate that these seasonal variations can be an important factor in the analyses of the interannual trends in trophic status in shallow lakes. Although the general response of well known indicator species seems to hold, there are many species that are classified differently according to the database selection and classification methods. How these different classifications affect the outcome of selected indices assessing ecological status is discussed in another article (see Penning et al. 2008, this issue). Acknowledgements REBECCA was funded by the European Commission under the 6th Framework Program, Contract No.: SSP1-CT-2003-502158—REBECCA. The authors thank all intercalibration representatives who contributed to the realisation of the database and the formulation of ideas and concepts during discussions in REBECCA and GIG meetings, specifically Laszlo Tóth (JRC) and Deirdre Tierney (EPA, Ireland). We are grateful to those who supplied data to the REBECCA dataset of European macrophyte data: Heikki Toivonen (SYKE, Finland); Tapio Rintanen (Finland); Helle Mäemets (Centre for Limnology, Estonia), Luc DeNeijs (Institute of Nature Conservation, Belgium); Vaida Olsauskyte (Lithuanian Environmental Protection Agency, Lithuania); Hanna Soszka (Institute of Environmental Protection, Poland); Arie Naber (Institute for Inland Water Management and Waste Water Treatment, the Netherlands). Gary Free (EPA, Ireland) and Eddy Lammens (Institute for Inland Water Management and Waste Water Treatment, the Netherlands) provided valuable comments on earlier versions of this manuscript. 123 246 123 Appendix List of species per species group included in REBECCA macrophyte study, with results of classifications. Column headings ‘P’ and ‘L’ represent methods based on percentiles or on LEAFPACS, respectively, and ‘n’ is the number of lakes in which a species occurred Species All P n N-GIG C-GIG CH P P P n I Baldellia ranunculoides o 8 + 5 I I Crassula aquatica Elatine hexandra o - 24 20 o o 24 8 I Elatine hydropiper - 131 - 128 I Elatine orthosperma I Elatine triandra 3 n CL L P n NH L 1 - n 1 5 1 1 1 + + NL P L + + n 57 o 57 - 372 I Eleocharis acicularis - 394 I Eriocaulon aquaticum - 20 I Isoetes echinospora - 380 - 378 I Isoetes lacustris - 487 - 424 I Limosella aquatica - 23 o 23 - 18 - - - 13 Littorella uniflora - 207 - 130 - 16 Lobelia dortmanna - 292 - 260 - 4 Lythrum portula o 13 - 12 - 462 - 462 I Subularia aquatica - 323 - 319 E Callitriche cophocarpa o 25 + 25 E E Callitriche hamulata Callitriche hermaphroditica o 93 101 - 78 79 11 2 - -- 7 - - 7 + + o o 5 13 o - 1 8 o 224 - 223 1 1 o 15 o 8 o 5 3 E Ceratophyllum demersum + 184 + 90 + 87 o 107 o 78 o 2 + 15 + - 24 - -- 11 o 237 + 20 E Hydrilla verticillata E Myriophyllum alterniflorum - 482 - 419 E Myriophyllum sibiricum - 45 - 45 E Myriophyllum spicatum + 184 + 68 - 9 - 7 24 8 - 8 3 o 8 - + 124 o 13 - 74 1 31 3 + + 56 + 5 o 41 - - 328 - 56 - 189 2 + 90 2 - - 365 - 101 - 186 3 - - 405 - 117 - 197 1 - - 22 - 11 o 4 - - 9 + + 12 - - 118 - 55 - 16 - 4 + - 6 - - 251 - 81 - 102 1 o + 11 o 7 + - 21 - - 419 - 88 - 256 3 2 - - 302 - 75 - 157 3 + + 21 + 9 + 14 4 5 - + 76 58 - 63 27 - 1 24 o 112 + + + - 2 19 o - 4 2 - - 210 - 46 o + 8 o 6 83 + + 4 + + 33 + + 53 + 17 o 30 + 61 - 60 + + 18 + + 30 - + 75 o 23 o 39 o 49 + 14 - 12 o 56 + 2 o 11 1 2 2 o + 79 2 o + 13 11 o -- 22 - - 387 - 115 - 185 - - 6 - + 37 - 15 - 30 + + 44 + + 24 + 9 Aquat Ecol (2008) 42:237–251 Callitriche palustris Elodea canadensis n n - 3 Callitriche stagnalis Elodea nuttallii - 1 E E - 1 E E P n 1 I Ranunculus reptans P + + n 3 I I P o o 1 I Latvia L 3 o Finland P 4 4 Norway Species All P n N-GIG C-GIG CH P P P L + + n E Myriophyllum verticillatum + 79 + 53 E Najas flexilis o 11 + 7 E Najas marina + 24 E Najas tenuissima - 3 E Potamogeton acutifolius + + n 26 24 + CL - P n NH L 24 n 2 NL P L n P L + + 13 + + 38 3 + + 4 n Norway Finland Latvia P P P n + 16 + n 5 + 2 1 3 2 28 24 3 2 n 3 2 E Potamogeton alpinus - 311 - 296 - 8 3 - - 5 - -- 29 - + 255 o 90 o 144 E Potamogeton berchtoldii - 322 o 284 - 8 2 o + 6 + - 27 - + 245 o 83 - 138 E Potamogeton compressus o 83 o 65 + 18 o + 18 + + 12 o + 51 2 - 30 o 6 E E Potamogeton crispus Potamogeton filiformis + - 94 82 + - 52 63 o - 31 11 + - + -- 28 9 3 2 + - + - 24 30 o - + - 28 33 + - 7 38 + 5 - 4 11 E Potamogeton friesii o 61 o 45 o 14 o - 14 + + 36 - + 9 + 21 E Potamogeton gramineus - 317 - 285 - 20 o - 11 9 o - 33 - + 245 - 66 - 128 + 8 o 74 - - 2 E Potamogeton lucens o 181 + 64 o 93 o - 84 + + 9 + + 37 - + 26 + 11 - 5 E Potamogeton obtusifolius o 177 + 151 + 11 + + 7 + + 4 + + 28 + + 120 + 30 + 55 3 + + 30 o + 15 + 10 1 o 9 o + 16 o - 56 - + 394 o 73 - 239 - 75 - -- 18 - 16 o + 6 o - 43 - + 153 - 47 - 18 E Potamogeton pectinatus + 128 + 45 + 71 + + 68 E Potamogeton perfoliatus o 627 - 470 o 115 o - 99 E Potamogeton polygonifolius - 23 - 18 E Potamogeton praelongus - 224 - 199 - 22 - - 16 - 62 E Potamogeton pusillus + 56 o 19 + 30 + + 29 1 + + 7 - + 12 + 4 + 6 o 17 o 5 3 2 + + 5 o + 12 + 9 + 4 + 10 + + 10 o 6 o - 4 2 + - 14 - + 17 - 12 - - 4 3 1 3 3 Potamogeton rutilus o 23 Potamogeton trichoides + 10 E Potamogeton vaginatus E Potamogeton gramineus 9 perfoliatus o 39 o 31 E Potamogeton gramineus 9 natans - 4 - 4 E Potamogeton filiformis 9 pectinatus E Potamogeton gramineus 9 lucens - 25 - 24 E Ranunculus aquatilis o 58 o 25 E Ranunculus baudoti 2 2 3 3 1 1 3 1 2 - 31 2 - - 29 2 2 - - 10 - + 14 + + 10 - + 15 1 o 12 1 1 1 1 - 24 247 123 E E 1 Aquat Ecol (2008) 42:237–251 Appendix continued 248 123 Appendix continued Species All P n N-GIG C-GIG CH P P P L + + n n P n NH L 9 P n 1 NL P L Norway Finland Latvia P P P L n 20 + + 5 9 - - 34 - 32 - 4 - 222 - 29 - 133 - 7 38 o 57 o 24 n n E Ranunculus circinatus + 36 + 25 + + E Ranunculus confervoides - 43 - 43 - -- E Ranunculus peltatus - 249 - 249 + - 18 - E Ranunculus penicillatus E Utricularia australis - 7 - 7 - + 7 E Utricularia intermedia - 130 - 127 - - 7 - - 114 - E Utricularia minor - 85 - 83 - - 4 - - 77 - 50 E Utricularia ochroleuca - 30 - 30 - -- 30 - 27 E E Utricularia vulgaris Zannichellia palustris o + 418 36 + 364 15 + + + 308 7 - 53 2 + 10 CL n n 1 1 o + 38 20 1 o + + 34 17 o + 4 3 + + + 41 8 + + 73 - + 473 o 111 o 248 + 118 3 - + 165 - 19 o 117 + 24 - - 154 - 134 56 - + 350 o 109 o 163 - 11 3 - + 27 - 50 o + 23 o 107 N Nuphar lutea o 766 o 572 o 154 o - 122 + + 32 N Nuphar pumila o 206 - 178 o 25 o - 14 + + 11 o 26 o + 22 o + 4 o 56 o + 46 + + 10 N Nuphar lutea 9 pumila - 174 - 172 N Nymphaea alba o 454 o 416 N Nymphaea candida 9 tetragona N Nymphaea candida o 86 o 30 N Nymphaea tetragona o 24 o 24 N Nymphaea alba 9 candida 2 2 1 1 - - 2 o 2 3 3 173 o 3 + 4 + 4 + + 4 o 208 o 184 - 13 - - 8 + + 5 + + 43 o + 140 o 35 o 63 N N Potamogeton natans Sagittaria natans o o 669 68 o o 513 68 - 129 o - 102 + + 27 - - 83 o o + + 422 66 o 133 o o 189 53 + + 7 + 8 2 - -- 15 - - 348 - 147 - 196 201 138 Sagittaria sagittifolia 9 natans + 8 + 8 - 396 - 388 2 N Sparganium gramineum o 205 o 205 2 o + 1 o N Sparganium hyperboreum - 27 - 27 1 - -- 26 - 20 - 6 N Sparganium natans o 99 o 97 5 o + 88 - 12 o 54 N Sparganium angustifolium 9 gramineum - 21 - 21 - + 21 + 76 + 44 + + 37 + + 48 2 o 5 o + 5 L Hydrocharis morsus-ranae + 120 L Lemna gibba + 7 - + + 7 + - + 25 1 1 2 + 30 + 25 1 Aquat Ecol (2008) 42:237–251 Nymphoides peltata Persicaria amphibia Sparganium angustifolium 25 24 N N - 2 N N 1 3 Species All P n N-GIG C-GIG CH P n P P n CL L n L Lemna minor + 264 + 215 + 47 + + 42 L Lemna trisulca + 146 + 78 + 36 + + 33 L Salvinia natans + 7 + 7 + + 7 L Spirodela polyrhiza + 84 + 58 + 26 + + 24 L Stratiotes aloides + 60 + 56 + 4 + + 4 + 5 + + o 49 - 23 - -- - 15 - -- 15 o - 22 12 - 25 3 - -- 24 3 L Trapa natans + 5 C Chara aspera o 76 33 - - 26 C Chara hispida o 22 o 7 - 14 - -- 14 C Chara intermedia C Chara rudis o 20 - 15 - -- C Chara strigosa - 7 - 6 C Chara tomentosa o 24 + 10 C Chara vulgaris o 12 C Nitella confervacea C Nitella flexilis o 48 Nitella wahlbergiana Nitellopsis obtuse C Tolypella canadensis C Tolypella glomerata + 18 23 o 14 2 + + 22 + + 34 + 13 + 19 + 9 + + 26 o + 28 o 13 - 11 2 1 - + 7 o -- 14 10 - - 10 o 10 + - 6 o + o + 9 - + 63 2 + 6 132 o 6 2 + 3 41 3 1 - 5 - 1 40 + + 40 1 1 - - 14 1 - 11 2 3 3 - 6 9 1 + 4 3 - - o 9 2 - + 4 - + 27 - - 121 1 2 3 o - 6 1 - 25 - 51 2 - 67 2 2 1 o - 2 4 4 3 + - - 1 5 6 25 o 6 3 o - 30 2 - 5 - 2 29 - o 1 14 1 1 5 - 3 1 1 o - 3 3 81 + 40 - C C + 4 - 97 Nitella translucens 56 + 5 10 - C + 51 - 135 8 158 6 3 o 142 + + + Chara globularis o + + o C - 53 26 16 9 10 Nitella mucronata + + - o Nitella opaca + + o - 10 n 5 1 o n 3 12 Chara fragilis C P o Chara filiformis C P n 20 C 2 P + C 1 n - 15 5 L 28 47 15 1 P - + - n o o - 1 + L Latvia 1 Chara connivens 2 + P n Finland 2 Chara contraria Chara delicatula 1 L Norway 4 C 1 P NL 21 C C 2 NH Aquat Ecol (2008) 42:237–251 Appendix continued -- 5 - 10 5 2 249 123 Classifications in the percentiles method are Sensitive (-), Tolerant (+) and Indifferent (o). LEAFPACS classifications are positive responder (+), negative responder (-), or strongly negative responder (- -).C = Central GIG, N = Northern GIG, H = High alkalinity, L = Low-moderate alkalinity. Species groups: I—isoetids, E—elodeids, N— nymphaeids, L—lemnids, C—charids 250 References Barko JW, Adams MS, Clesceri NL (1986) Environmental factors and their consideration in the management of submersed aquatic vegetation—a review. J Aquat Plan Manag 24:1–10 Birk S, Korte T, Hering D (2006) Intercalibration of assessment methods for macrophytes in lowland streams: direct comparison and analysis of common metrics. Hydrobiologia 566:417–430 CEN (2006) Water quality—guidance standard for the surveying of macrophytes in lakes. prEn 15460 Coops H, Kerkum FCM, van den Berg MS, van Splunder I (2007) Submerged macrophyte vegetation and the European Water Framework Directive: assessment of status and trends in shallow, alkaline lakes in the Netherlands. Hydrobiologia 584:395–402 Covaliov S, van Geest G, Hanganu J, Hulea O, Torok L, Coops H (2003) Seasonality of macrophyte dominance in floodpulsed lakes of the Danube Delta. Hydrobiologia 506(1– 3):651–656 Ecke F (2006) Kompletterande utredningar för revideringen av bedömningsgrunder för makrofyter i sjöar. Report, Institutionen för tillämpad kemi och geovetenskap, Luleå tekniska universitet, 28 pp Free G, Little R, Tierney D, Donnelly K, Caroni R (2006) A reference based typology and ecological assessment system for Irish lakes-preliminary investigations. Environmental Protection Agency, Wexford. www.epa.ie Gibson CE, Foy RH, Bailey-Watts AE (1996) An analysis of the total phosphorus cycle in some temperate lakes: the response to enrichment. Freshw Biol 35(3):525–532 Heiskanen AS, Van der Bund WJ, Cardoso AC, Noges P (2004) Towards good ecological status of surface waters in Europe—interpretation and harmonisation of the concept. Water Sci Technol 49(7):169–177 Hill MO, Ellenberg H (1999) Ellenberg’s indicator values for British plants: technical annex. ECOFACT research report S. Stationery Office Books. ISBN: 1870393481 James FJ, Barko JW, Butler MG (2004) Shear stress and sediment resuspension in relation to macrophyte biomass. Hydrobiologia 515:181–191 Kirk JTO (1994) Light and photosynthesis in aquatic ecosystems, 2nd edn. Cambridge University Press, 509 pp Kurimo U (1970) Effect of pollution on the aquatic macroflora of the Varkaus area, Finnish Lake District. Ann Bot Fenn 7:213–254 Lamers LPM, Smolders AJP, Roelofs JGM (2002) The restoration of fens in the Netherlands. In: Nienhuis PH, Gulati R (eds) The ecological restoration of wetlands in the Netherlands. Kluwer, Amsterdam. Also in Hydrobiologia 478:107–130 Leka J, Toivonen H, Leikola N, Hellsten S (2007) Makrofyytit Suomen järvien ekologisen tilan ilmentäjinä. Valtakunnallisen makrofyyttiaineiston käyttö ekologisen tilaluokittelun kehittämisessä. Suomen ympäristö, 42 p. + app Leyssen A, Adriaens P, Denys L, Packet J, Schneiders A, Van Looy K, Vanhecke L (2005) Toepassing van verschillende biologische beoordelingssystemen op Vlaamse potentiele interkalibratielocaties overeenkomstig de Europese 123 Aquat Ecol (2008) 42:237–251 Kaderrichtlijn Water – Partim ‘Macrofyten’. Rapport van het Instituut voor Natuurbehoud IN.R. 2005.05 in opdracht van VMM, Brussel Lyche Solheim A (ed) (2005) Reference conditions of European lakes. Indicators and methods for the Water Framework Directive Assessment of Reference Conditions. REBECCA report D7. www.rbm-toolbox.net/docstore/docs/3.1713. D7-uusi.pdf Lyche Solheim A (ed) (2006) Dose-response relationships between biological and chemical elements in different lake types. REBECCA report D11. www.rbm-toolbox. net/docstore/docs/3.1713.D11.pdf Mäkirinta U (1978) Ein neues ökomorphologisches Lebensformen-System der aquatischen Makrophyten. Phytocoenologia 4:446–470 Moe SJ, Dudley B, Ptacnik R (2008) REBECCA databases: experiences from compilation and analyses of monitoring data from 5,000 lakes in 20 European countries. Aquat Ecol. doi:10.1007/s10452-008-9190-y Mjelde M (2007) Macrophytes and eutrophication in lakes (unpublished) Moss B, Stephen D, Alvarezn C, Becares E, Van der Bund W, Collings SE et al (2003) The determination of ecological status in shallow lakes—a tested system (ECOFRAME) for implementation of the European Water Framework Directive. Aquatic Conserv Mar Freshw Ecosyst 13:507– 549 Murphy KJ (2002) Plant communities and plant diversity in softwater lakes of Northern Europe. Aquat Bot 73(4):287– 324 Murphy KJ, Rørslett B, Springuel I (1990) Strategy analysis of submerged lake macrophyte communities: an International example. Aquat Bot 36:303–323 Palmer MA, Bell SL, Butterfield I (1992) A botanical classification of standing waters in Britain: applications for conservation and monitoring. Aquatic Conserv Mar Freshw Ecosyst 2:125–143 Penning WE, Dudley B, Mjelde M, Hellsten S, Hanganu J, Kolada A, Van den Berg M, Poikane S, Phillips G, Willby N, Ecke F (2008) Using aquatic macrophyte community indices to define the ecological status of European lakes. Aquatic Ecol. doi:10.1007/s10452-008-9183-x Pot R (2003) Veldgids nr. 17: Veldgids Water- en oeverplanten, -352. KNNV Uitgeverij & STOWA, Utrecht Rørslett B (1991) Principal determinants of aquatic macrophyte richness in northern European lakes. Aquat Bot 39:173– 193 Schaminee JHJ, Weeda EJ, Westhof V (1995) De vegetatie van Nederland. Deel 2. Plantengemeenschappen van wateren, moerassen en natte heiden. Opulus Press, Uppsala, Leiden Schaumburg J, Schranz C, Hofmann G, Stelzer D, Schneider S, Schmedtje U (2004) Macrophytes and phytobenthos as indicators of ecological status in German lakes—a contribution to the implementation of the Water Framework Directive. Limnologica 34:302–314 Søndergaard M, Jeppesen E, Peder JJ, Lildal SA (2005) Water Framework Directive: ecological classification of Danish lakes. J Appl Ecol 42(4):616–629 Stelzer D, Schneider S, Melzer A (2005) Macrophyte based assessment of lakes—a contribution to the implementation Aquat Ecol (2008) 42:237–251 of the European Water Framework Directive in Germany. Int Rev Hydrobiol 90(2):223–237 Tóth LG, Poikane S, Penning WE, Free G, Mäemets H, Kolada A (2008) First steps of the central-baltic intercalibration exercise: from where do we start? Aquatic Ecol. doi: 10.1007/s10452-008-9184-9 Van den Berg MS (2004) Achtergrondrapportage referenties en maatlatten waterflora. Rapportage van de expertgroepen macrofyten en fytoplankton. STOWA report 251 Van Geest G (2005) Macrophyte succession in floodplain lakes. Spatio-temporal patterns in relation to hydrology, lake morphology and management. Thesis, Wageningen University, The Netherlands Willby N, Pitt J, Phillips G (2006) Summary of approach used in LEAFPACS for defining ecological quality of rivers and lakes using macrophyte composition. Draft Report January 2006 123