Upload
emadhsobhy
View
216
Download
0
Embed Size (px)
Citation preview
7/27/2019 ber12-561
1/9
Boreal environment research 12: 561569 issn 1239-6095
hk 24 ob 2007 2007
ib w pypk y
K vu1), l lp2) d a-l hp3)
1)Section o Ecology, Department o Biology, FI-20014 University o Turku, Finland (e-mail: kristiina.
[email protected])2)Finnish Environment Institute, FI-00250Helsinki, Finland3)Karelian Research Institute, University o Joensuu, and North Karelian Environment Centre, FI-80100
Joensuu, Finland
Received 14 Sep. 2005, accepted 15 Mar. 2007 (Editor in charge o this article: Johanna Mattila)
vu, K., lp, l. & hp, a.-l. 2007: ib w pypk y-
. Boreal Env. Res. 12: 561569.
Phytoplankton is one of the biological quality elements of the Water Framework Directive
of European Union for ecological classication of surface waters included. The reliabil -ity and comparability of phytoplankton data is, therefore, essential in assessment of theclassication. The recently published standard for phytoplankton analysis based on theUtermhl technique supports stringent demands for comparability of data. We tested thecomparability of phytoplankton analyses (Utermhl method) between professional and
non-professional analysts. The coefcient of variation was highly variable, 10%42% forprofessionals and 8%57% for non-professionals. Elimination of the error caused by sub-sampling and the increase in the minimum number of enumerated units decreased the CVespecially within professionals. CV was smaller (4%19%) when samples were counted byone person. In order to minimize the variability caused by differences between subsamples,new methods for comparison testing both for technical enumeration and for species identi-cation should be developed.
IntroductionPhytoplankton communities are sensitive to
changes in their environment and, therefore,phytoplankton total biomass and many phyto-plankton species are used as indicators of waterquality (Jrnefelt 1952, Heinonen 1980, Olrik1994, Reynolds 1997, Heinonen et al. 2000,Reynolds et al. 2002, Brettum and Andersen2005). Phytoplankton communities give moreinformation on changes in water quality than
mere nutrient concentrations or chlorophyll aconcentrations. Standards for chemical labora-tory analyses (e.g. National Board of Waters inFinland 1981 and 1984, SNV Rapport 3075),
which are relatively simple to follow, have beenused since the 1980s. In contrast, quantitativeanalyses of natural phytoplankton multispecies
communities are more complicated to perform.Furthermore, only recently a commonly acceptedstandard for phytoplankton enumeration waspublished (EN 15402 2006).
Currently phytoplankton counting methodsvary from country to country. Several alternativecounting chambers, e.g. Sedgwick-Rafter, Lundand Palmer-Maloney, are used for phytoplankton
analyses if inverted microscopes are not available(Welch 1948, Palmer and Maloney 1954, Lund etal. 1959, Guillard 1978). The Sedgwick-Rafterchamber is best suited for counting large algae
7/27/2019 ber12-561
2/9
562 Vuorio et al. Boreal env. res. v. 12
at low concentrations, and the Palmer-Maloneychamber is of use for nanoplankton counting
(Guillard 1978, EN 15402 2006). The Utermhltechnique (Utermhl 1958), however, is better
suited in quantitative analyses of multispeciesphytoplankton communities, especially in nutri-ent-poor waters. This method requires an invertedmicroscope equipped with high quality optics inorder to obtain reliable results (Olrik et al. 1998,Lepist 1999). Although a uniform methodologyfor phytoplankton counting based on Utermhlmethod was suggested by Olrik et al. (1998) forfreshwater, the recommended method is not fullyfollowed or evaluated. The new European stand-ard for phytoplankton counting and statistical
analyses based on Utermhl technique has beenpublished (EN 15402 2006) and should replaceconicting national standards in 29 Europeancountries at the latest by February 2007.
The implementation of the Water FrameworkDirective (European Union 2000) requires fre-quent phytoplankton monitoring and compara-ble phytoplankton results, e.g. the use of equalcounting methods and harmonization of spe-cies identication, at least among members of
the same Geographical Intercalibration Group(GIG). Ecological classication of surface watersincluded in the Water Framework Directive isbased not only on quantities of phytoplankton(biomass) but also on taxonomic compositionand size structures of phytoplankton communi-ties. At the present, phytoplankton analysis is theonly method to produce these taxonomic qualityelements, including e.g. indicator species. Har-monization of enumeration procedures and spe-cies identication should, therefore, be includedin the quality control of phytoplankton analyses.
During the last decades, phytoplanktonresearch quality controls in brackish water
have been introduced among Baltic countries(HELCOM 1997). For freshwaters, however,quality controls of phytoplankton analyses are
still rare, despite the fact that phytoplanktoncommunities have been studied in Nordic coun-tries since the beginning of the 20th century(in Finland e.g. Levander 1900, Blomqvist etal. 1915, Jrnefelt 1937, and in Sweden e.g.Cleve-Euler and Huss 1912). Furthermore, thenation-wide routine phytoplankton monitoringprogrammes have been operated by Finnish and
Swedish authorities since the 1960s (e.g. Hei-nonen 1980, Willn 2001, Lepist 2004). Sincethe 1980s Finnish phytoplankton analysts aswell as the Nordic phytoplankton and periphyton
group (NPPG) have harmonized species identi-cation. Recently, intercalibrations of phytoplank-ton enumerations have also been introduced e.g. in 1989 the Danish National EnvironmentalMonitoring Programme for lakes was launchedincluding routine intercalibrations among activeDanish phytoplankton analysts (Jensen and Sn-dergaard 1994, Lauridsen et al. 2005).
Performed intercalibrations have shown thatthere are differences in phytoplankton resultsbetween laboratories as well as between indi-
vidual phytoplankton counts carried out by dif-ferent persons (Barinova et al. 1980, Rott 1981,Huttunen 1985, Niemi et al. 1985, Jensen andSndergaard 1994, HELCOM 1997). This mustbe taken into account when comparing phyto-
plankton results. Though in Nordic countriesthe Utermhl technique is widely used (Olriket al. 1998), every individual analyst modiesthis technique according to the properties of themicroscope used.
There is thus an urgent need to reconsiderphytoplankton analysis methodology, to trainanalysts, and to perform intercalibrations amongworking analysts to increase the accuracy of
phytoplankton data. Furthermore, for ecologicalclassication there is a need to determine theacceptable level of variation caused by differ-ences in subsamples. In this paper we provideresults from two intercalibrations among Finn-
ish universities and research laboratories, andone joint FinnishSwedishNorwegian inter-calibration. The aim of these intercalibrationswas to evaluate the identication of species,and, more importantly, variability of quantitativeresults. Here we discuss the variability of count-ing results between analysts, and methodologicalproblems related to intercalibrations.
Methods
Three intercalibration sessions were performed in2002, 2003 and 2005. Participants were dividedinto two groups, and the results of phytoplank-ton analyses performed by professional analysts
7/27/2019 ber12-561
3/9
Boreal env. res. v. 12 Intercalibrations o reshwater phytoplankton analyses 563
were, on daily basis, compared with those ofanalysts occasionally counting phytoplankton
samples or students in training to be professionalphytoplankton analysts. A nonparametric Mann-
Whitney U-test was used to compare differencesbetween professional and non-professional phy-toplankton analysts.
First intercalibration session 2002
The rst intercalibration session was also a train-ing session in the methodology, and was thusdeliberately kept simple. The phytoplanktonsample for this session represented a 2-meter-
deep water column of a lake in SW Finland.Samples of 200 ml were preserved with acidLugols solution and combined into one sampleof two litres. This combined sample was mixedand divided into subsamples of 100 ml for eachparticipant. The two taxa to be counted werethe cryptophyte Rhodomonas lacustris and thecolony forming diatom Fragilaria crotonensis.In addition, only the volume to be sedimented(25 ml) was provided. Otherwise, every par-
ticipant used their own modication of the Uter-mhl counting method (Utermhl 1958, Olrik etal. 1998). Participants were instructed to give thecell numbers as units per litre. Altogether 13 ana-lysts participated in the rst intercalibration.
Second intercalibration session 2003
For the second intercalibration, the preparationof the sample was equal to the rst session, butinstructions were given in more detail. The set-tled volume was also 25 ml, and four taxa were tobe counted: the diatoms Aulacoseira subarcticavar. subborealis and Asterionella formosa; thechlorophyte Desmodesmus bicellularis; and thecyanobacterium Radiocystis geminata. A. sub-arctica var. subborealis were counted as singlecells and D. bicellularis as cell pairs, using atotal magnication of 6001000. Asterionellawas also counted as single cells but Radiocystis
as subcolonies containing approximately 150cells using a total magnication of 200250. Atleast 50 units (i.e. cells, cell pairs or subcolonies)of each taxon were counted. Special attention
was paid to the counting of colony forming taxa,therefore no attempt was made to split up the
Radiocystis into single cells. Results were givenas units per one litre. A total of 12 analysts took
part in the second intercalibration.In order to study the reproducibility of phy-toplankton counts, Toini Tikkanen sedimentedthree subsamples and counted each of themusing two different counting methods. The cryp-tophyte Rhodomonas lacustris was counted incrossing diagonals and in 20 elds with a totalmagnication of 500. The diatom Fragilariacrotonensis was counted using the same methodsbut with a magnication of 200. A nonparamet-ric Mann-Whitney U-test was performed to test
the differences between subsamples and betweenthe two different counting methods (SPSS 11.5.1for Windows, SPSS Inc. 19892002).
Third intercalibration session 2005
The joint FinnishSwedishNorwegian intercali-bration in 2005 was conducted during a Nordicphytoplankton workshop. The setup was simpler
than that of a natural sample because only threelaboratory strains were used, a unicellular greenalga and two lamentous cyanobacteria. Twoseparate samples were settled, one containing thechlorophyte Selenastrum sp. and the other con-taining the two strains of lamentous cyanobac-teria Planktothrix sp. and Anabaena sp. Partici-pants were instructed to count 10 elds from onediagonal; Selenastrum sp. was counted as singlecells and the two cyanobacteria as laments.Analysts were also advised to assess the numberof cells ofAnabaena laments and the numberof 25 m long sublaments of Planktothrix.Magnications were 650 for Selenastrum sp.and 400 for the cyanobacterial trichomes. All22 participants counted the sample containingSelenastrum sp., but because of limited timeonly 14 participants counted both samples.
Results
Intercalibrations
During the rst intercalibration, the range of
7/27/2019 ber12-561
4/9
564 Vuorio et al. Boreal env. res. v. 12
cell numbers counted was large: 23253 cellsofRhodomonas lacustris and 4502255 cells ofFragilaria crotonensis. The average cell num-bers per litre are shown in Table 1. Among all
participants, the coefcient of variation (CV) forboth taxa was 21%. Among professionals (N= 8)who count phytoplankton samples daily, the CVwas 19% for both Rhodomonas and Fragilaria.Among other analysts the corresponding valuewas 24%. The differences in phytoplanktoncounts between professionals and non-profes-sionals were not statistically signicant (Mann-Whitney U-test).
During the second intercalibration, the rangeof the cell number counted ofAulacoseira was
129623, and of Desmodesmus 74443. Therange of colonies ofRadiocystis was 55166and ofAsterionella 801109. Table 2 gives the
results converted into number of cells or colo-nies l1. When the cell and colony counts wereconverted into numbers of units l1, the coef-cients of variation were 15% for Aulacoseira,
34% forDesmodesmus, 29% forRadiocystis and35% for Asterionella (Table 2). The variationbetween participants was smaller than in the rstintercalibration attempt: coefcients of variationamong the professionals (N = 78) were 14%for Aulacoseira, 19% for Desmodesmus, 33%for Radiocystis, and 42% for Asterionella; and20, 57%, 23% and 20% for non-professionals,respectively. The phytoplankton counts betweenprofessionals and non-professionals were notstatistically signicant (Mann-Whitney U-test).
During the third intercalibration, the rangeof the counts ofSelenastrum cells was 340495(CV = 9%,N= 22). ForAnabaena the number of
Table 1. ru f b: ddu
u u d u (u 1), f
(cv% = dd d pg
u) g p d g
pp. tx w ud Rhodomonas
lacustris (rd ) d Fragilaria crotonensis (Fg
). Pp ub 13 ud Rhodomonas
lacustris w (= 13a d 13B).
cu . rd Fg
01 31200 163626
02 23664 140461
03 27988 139700
04 33600 132700
05 29600 186400
06 30400 68815
07 44020 219360
08 29268 177210
09 21726 14017810 33300 250000
11 30666 180761
12 38500 158000
13 43314 157635
P
m 31180 165637
cv% 19 19
N 8 8
n-p
m 33501 177381
cv% 24 24
N 5 5
am 32073 170530
cv% 21 21
N 13 13
Table 2. ru d b: dd-
u u u d u (u 1), f
(cv% = dd d pg
u) g p d g
pp. tx w ud Aulacoseira
subarctica. subborealis(au u;u), Desmodesmus
bicellularis (D b), Radiocystis geminata (rd
g) d Asterionella ormosa(a ).
cu au u;u D b rd g a
.
01 3540000 6020000 99670 90110
02 3409000 5382000 51360 54640
03 3712000 5184000 79880 25990
04 3470000 4400000 48000 49500
05 3540000 4980000 43200 38100
06 4346100 3951000 55800 52200
07 3588145 6960138 79128 47100
08 2585024 46107 2707209 4072000 3357000 44600 73700
10 3700000 3500000 64125 46170
11 2788378 7068214 80384 54636
12 2710400 1760000 60995 59886
P
m 3523784 5268163 62893 48089
cv% 14 19 33 42
N 8 7 8 8
n-p
m 3317695 3921304 62526 58598
cv% 20 57 23 20
N 4 4 4 4
am 3455087 4778396 62771 51592
cv% 15 34 29 35
N 12 11 12 12
7/27/2019 ber12-561
5/9
Boreal env. res. v. 12 Intercalibrations o reshwater phytoplankton analyses 565
cells counted varied from 297 to 542 (CV = 18%,N= 14) and that of trichomes from 43 to 99 (CV= 25%,N= 13). The range in Planktothrix count-ing units of 25 m was 81175 (CV 25%, N=
14) and that of whole trichomes from 35 to 67(CV = 17%, N= 13). Among professional phy-toplankton analysts, the deviations of CV were10% for Selenastrum cells, 18% for Anabaenacells and 23% forAnabaena trichomes, 20% forPlanktothrix counting units and 13% for Plank-tothrix trichomes (Table 3). For other analyststhe corresponding deviations of CV were 8%,18%, 16%, 35% and 20%, respectively. Again,
the differences in phytoplankton counts betweenprofessionals and non-professionals were notstatistically signicant (Mann-Whitney U-test).
Repeatability of phytoplanktoncountings
In order to test the reproducibility of the phyto-plankton countings, two different subsamples ofthe same material were sedimented and countedthree times by the same person, using two dif-ferent methods (20 elds versus crossing diago-
Table 3. ru d b: ddu bu u u d u, f -
(cv% = dd d pg u) g p d g pp.
tx w ud w by wg g: Selenastrum, Anabaenad Planktothrix;
s = Selenastrum, ab1 = Anabaena, ab2 = Anabaena, P1 = Planktothrixu 25 ,
d P2 = Planktothrix.
cu . s ab1 ab2 P1 P2
01 64603027 42989047 8997708 12496816 5898497
02 70461235 37890346 7498090 9697529 5398625
03 70135779 35291008 5898497 8097937 5198675
04 80387645 35790881 6898242 12796740 5698548
05 65253939 47687850 7798013 10297376 3998981
06 55327529 54186195 9897478 16395823 539862507 69973051 32091824 4298905 11497071 5698548
08 64928483 44288716 9297631 12296867 6698293
09 79411277 35690907 7298141 13096663 5098701
10 65742123
11 73715796
12 72576700 35191034 5798523 17495543 5698548
13 72576700 29692435 5098701 9697529 4198930
14 80550373
15 70135779
16 63952114
17 74203980
18 7794672519 69973051
20 77295813 45388436 9597555
21 64277571 39090041 6798268 8997708 3499109
22 63789386 29992359 4798777 8197911 4598828
P
m 69085444 40656308 7542523 11852536 5454166
cv% 10 18 23 20 13
N 11 9 9 9 9
n-p
m 71570745 35870861 5618835 10797249 4498854
cv% 8 18 16 35 20
N 11 5 4 5 4
am 70328094 38947220 6952075 11475648 5160224
cv% 9 18 25 25 17
N 22 14 13 14 13
7/27/2019 ber12-561
6/9
566 Vuorio et al. Boreal env. res. v. 12
nals). For Rhodomonas cells the count from 20elds gave a CV of 13% for both subsamples,and that from crossing diagonals was 12% forthe rst subsample and 10% for the second sub-sample. For Fragilaria cells corresponding CVswere 19% and 14% when enumerated from 20elds and 23% and 5% when enumerated fromcrossing diagonals, respectively. The highest CVwas probably due to uneven sedimentation in
one of the three subsamples sedimented. Thenumbers ofRhodomonas ranged from 165 000to 235 000 units l1 when counted from 20 eldsand from 133 380 to 226 200 when counted fromcrossing diagonals. The corresponding valuesfor Fragilaria cells were 690 0001 100 000and 808 0801 993 400 units l1 (Fig. 1). Thisvariability was thus considerably smaller thanvariability between different analysts count-ing subsamples of the same material. However,we also observed a statistically signicant dif-ference (Mann-Whitney U-test: U = 3.5, P =0.02) between the two sedimented chambers in
Rhodomonas counts. The CV was lower whenmore Fragilaria cells were counted from twodiagonals, however, there were no signicantdifferences (Mann-Whitney U-test) between thetwo methods.
Discussion
During all three intercalibrations, thoroughinstructions were given for sample preparation,enumeration technique, and taxa to be counted.
During the rst intercalibration, participantswere allowed to use their own modication ofthe Utermhl method. To decrease the minimumvariation, participants were given more detailedinstructions, e.g. a minimum number of units tobe counted during the second intercalibration.The maximum variation, however, increasedwithin non-professional analysts but especially
within professional analysts. When differences
arising from subsampling and sedimentationwere excluded in the third intercalibration, thevariation within professional analysts decreasedfurther but the variation also decreased withinnon-professional analysts. Despite this decrease,however, the inexperience of some participantscaused unavoidable remaining variation in theresults. Due to high overall variation within bothprofessional and non-professional analysts, thedifferences between professional and non-pro-fessional analysts were not signicant.
Most previous intercalibrations have har-monized marine phytoplankton analyses (e.g.Hobro and Willn 1977, Barinova et al. 1980,Hllfors and Niemi 1990, HELCOM 1997) andfewer intercalibrations have been concerned withfreshwater phytoplankton (e.g. Rott 1981, Jensenand Sndergaard 1994, Veen et al. 2005). Ourintercalibrations applied the Utermhl technique,considered to be the most precise of existingmethods and a common standard for phytoplank-
ton analyses within EU (EN 15204 2006). Thepresent study is limited in scope as comparedwith previous intercalibrations of freshwater
phytoplankton (Rott 1981, Jensen and Snder-
Rhodomonas lacustris
Subsample
1 2 1 2
No
ofcounted
units(x103
l1)
12
14
16
18
20
22
24
20 fields2 diagonals
Fragilaria crotonensis
Subsample
1 2 1 260
70
80
90
100
110
120
130
Fig. 1. t ub (u 1 sD) Rhodomonas lacustrisd Fragilaria crotonensis w d
ubp dug d b pg pby u by t tkk.
7/27/2019 ber12-561
7/9
Boreal env. res. v. 12 Intercalibrations o reshwater phytoplankton analyses 567
gaard 1994) deviating the following ways: First,entire samples from nature were not analysed;instead, only a few easily recognised specieswere considered. Second, calculations of the
species biovolume were not performed; insteadwe enumerated only cell/colony numbers. Last,ultrasonically treated samples of colony formingcyanobacteria were not used, because ultrasoni-cation complicates species identication, andbecause ultrasonication is not included in thenew EU standard (EN 15204 2006). Despitethese deviations our results are congruent withthe earlier studies showing high variability (CVs)between laboratories and individual analysts,especially when natural samples are used (e.g.
Hobro & Willn 1977, Barinova et al. 1980, Rott1981, Niemi et al. 1985, Veen et al. 2005).
The exact quantitative composition of phy-toplankton in a sample is difcult, if not impos-sible, to assess. Estimates of relative uncertaintytherefore must be derived from phytoplanktoncounts by multiple persons. Quantitative uncer-tainty depends partly on the abundance of a spe-cic taxon in a sample and partly on the amountof units counted (Lund 1958, Olrik et al. 1998).
Because the volume of sedimented water is basedon phytoplankton densities in water, changes inabundances affect the number of units enumer-ated (Eloranta 1978). This uncertainty of countsmust be expressed for each taxon separately.
In our comparisons, counts of colonial spe-cies showed highest CVs between analysts.During the second intercalibration, the CV washigh for Desmodesmus suggesting that some of
the participants may have counted this bicellularspecies as single cells although instructions were
given to count two-celled colonies. Moreover,high CV forRadiocystis suggests that the size ofsubcolonies was difcult to determine. Specialattention was thus paid to the colony formingtaxa that were counted, either as subcoloniesestimated by the cell number (such as the cyano-bacteria Radiocystis) or as single cells (the dia-tomsAsterionella and Fragilaria). Colonies con-sisting of several tens or hundreds of cells maygive a distorted picture of the number of colonies
if only the number of cells is enumerated. There-fore, a sufcient number of colonies (e.g.Radio-cystis geminata), not cells, must thus be countedto reduce the CV. Some researchers have tried
to solve the problem by counting ultrasonically
treated samples, where the CV would have beenmuch reduced (Jensen and Sndergaard 1994,Olrik et al. 1998). Even in this method, how-
ever, sufcient numbers of colonies may not becounted as the cells may originate from only afew large colonies, e.g. during HELCOM inter-calibrations systematic errors arouse from low
counts (HELCOM 1997). In the present studythe extent of uncertainty (CV) decreased withincreasing number of units counted. Our resultsconrm that the number of units counted has tobe high enough to minimize the CV.
Even though special attention was paid to theless trained participants, our results reveal that
the primary determinants of variation betweenanalyses are the experience and carefulness ofthe participants as earlier noticed by Barinovaet al. (1980) and Rott (1981). This variation,specically, may originate from the failure toprecisely follow instructions, individual modi-cations of the Utermhl method, or differencesin the optical quality of microscopes. Beforephytoplankton results can be accepted as validin an international context, analysts must thus
undergo thorough training at a professional labo-ratory and intercalibrate their results with thoseof professional analysts.
Variability between natural phytoplanktonsamples can be high and further increase thevariability between analysts (Hakala et al. 2002).In this study, the variability between naturalsubsamples counted by one analyst was as highas 20%, and even higher, up to 57%, betweendifferent analysts. This variability, thus, has tobe considered when comparing data with thevalid ecological class boundaries. The reliabilityand comparability of quantitative and qualita-tive phytoplankton data in assessment of theecological classication of lakes and coastalwaters depends as well on the quality of thework on phytoplankton and also very much onthe person that works up the material. In future,this variability challenges the implementation of
ecological classication of waters because datafrom human impacted waters analyzed by vari-
ous laboratories is compared with reference dataanalyzed by environmental authorities.
The results of these intercalibration testsshowed that it is necessary to elaborate a test
7/27/2019 ber12-561
8/9
568 Vuorio et al. Boreal env. res. v. 12
which excludes the variation between subsam-ples of natural phytoplankton to reliably compare
true variability between analysts. Furthermore, agood knowledge of phytoplankton taxonomy is
essential in order to correctly identify species forecological classication. Therefore, in additionto the technical performance of phytoplankton
enumeration, comparison tests for species identi-cation are also needed.
Acknowledgements: We acknowledge especially Lic. Phil.Toini Tikkanen, the grand old lady of phytoplankton work inFinland, who made the extra tests, and all the phytoplanktonanalysts who took part the intercalibrations. We also thankProfessor Jouko Sarvala for his valuable comments on themanuscript. Mark Higgins proofread the manuscript.
References
Barinova S.P., Forsskhl M., Kukk E., Melnikova L., Mel-vasalo T., Niemi ., Piirsoo K. & Viljamaa H. 1980.Phytoplankton.Meri 8: 1123.
Blomqvist E., Levander K.L. & Witting R. 1915. Arbetsut-skottet fr underskning af de nska insjarnas vattenoch plankton. Fennia 39: 195.
Brettum P. & Andersen T. 2005. The use of phytoplankton
as indicators of water quality. NIVA-report SNO 4818-2004.Eloranta P. 1978. Effects of size of sample counted in
phytoplankton analyses. Annales Botanici Fennici 15:169176.
Cleve-Euler A. & Huss H. 1912. Vattnet i sjar och vatten-drag inom Stockholm och dess omgivningar. II. Plank-tonunderskningar. Stockholms stads hlsovrdsnmdsrsberttelse 1911.
EN 15204 2006. Water quality. Guidance standard on theenumeration of phytoplankton using inverted micro-
scope (Utermhl technique).European Union 2000. Water Framework Directive2000/60/
EC.Guillard R.R.L. 1978. Counting slides. In: Sournia A. (ed.),
Phytoplankton manual, United Nations Educational, Sci-entic and Cultural Organisation, Paris, pp. 182189.
Hakala J., Vuoristo H. & Lepist L. 2002. Kasviplankton-tutkimusten laadunvarmistuksen ongelmista. Vesitalous3/2002: 4749.
Heinonen P. 1980. Quantity and composition of phytoplank-ton in Finnish inland waters. Publications of WaterResearch Institute 27.
Heinonen P., Ziglio G. & van der Beken A. 2000.Hydrologi-cal and limnological aspects of lake monitoring. John
Wiley and Sons Ltd.HELCOM 1997. Phytoplankton species composition, abun-dance and biomass. Manual for marine monitoring inthe combine programme of HELCOM, EC 8/97, AnnexC-6.
Hobro R. & Willn E. 1977. Phytoplankton countings. Inter-calibration results and recommendations for routinework.Int. Revue ges. Hydrobiol. 62: 805811.
Huttunen M. 1985. Kasviplanktonin interkalibrointi vuonna1979. Vesihallituksen Monistesarja 384: 127.
Hllfors G. & Niemi . 1990. Proposal for standardizationof the way of presenting phytoplankton results. FinnishMarine Research 257: 2936.
Jensen J.P. & Sndergaard M. 1994.Interkalibrering af plan-teplanktonundersgelser I ser. Danmarks Miljunder-sgelser, Teknisk anvisning fra DMU, nr. 8.
Jrnefelt H. 1937. Ein kleiner Beitrag zur Limnologie desTuusulanjrvi. Acta Societas Fauna et Flora Fennica60: 502515.
Jrnefelt H. 1952. Plankton als Indikator der Trophiegruppender Seen.Annales Acadimiae Scientiarum Fennicae AIV18: 129.
Lauridsen T., Sndergaard M., Jensen J.P. & Jeppesen E.
2005. Overvgningsprogram for ser. Danmarks Milj-undersgelser, Teknisk anvisning fra DMU.
Lund J.W.G., Kipling C. & Le Cren E.D. 1959. The invertedmicroscope method of estimating algal numbers and thestatistical basis of estimations by counting. Hydrobio-logia 11: 143170.
Lepist L. 1999: Phytoplankton assemblages reecting theecological status of lakes in Finland. Monogr. Bor. Env.Res. 16: 143.
Lepist L., Vuorio K. & Holopainen A.-L. 2006. Kasviplank-tonin tutkimusmenetelmt. Vesitalous 1: 1621.
Lepist L. 2004. Kasviplankton. Suomen ymprist 682:
1013.Levander K.M. 1900. Zur Kenntnis der Fauna und Florannischer Binnenseen. Acta Societas Fauna et FloraFennica 19: 155.
National Board of Waters 1981. Vesihallinnon analyysimene-telmt. Rapport 213. Vesihallinto, Helsinki.
National Board of Waters 1984. Vesiviranomaisten kyttmtvesisttutkimuksen nytteenottomenetelmt. 2. Reviewof Publications of National Board of Waters 40, Hel-sinki.
Niemi ., Melvasalo T. & Heinonen P. 1985. Phytoplanktoncounting techniques and primary production measure-ments comments on the results of intercalibration.Aqua Fennica 15: 89103.
Olrik K., Blomqvist P., Brettum P., Cronberg G. & ElorantaP. 1998. Methods for quantitative assessment of phy-toplankton in freshwaters, part I. Naturvrdsverket,Svensk miljvervakning, Rapport 4860.
Palmer C.M. & Maloney T.E. 1954. A new counting slidefor nannoplankton.American Society for Limnology andOceanography, Special Publication 21: 16.
Reynolds C., Huszar V., Kruk C., Naselli-Flores L. & MeloS. 2002. Towards a functional classication of the fresh-water phytoplankton.J. Plankton Res. 24: 417428.
Reynolds C. 1997. Vegatation procecces in the pelagic. A
model for ecosystem theory. In: Kinne O. (ed.), Excel-lence in ecology, Ecology Institute, Oldendorf/Luhe,Germany, pp. 1371.
Rott E. 1981. Some results from phytoplankton countingintercalibrations. Schweiz. Z. Hydrol. 43: 3462.
7/27/2019 ber12-561
9/9
Boreal env. res. v. 12 Intercalibrations o reshwater phytoplankton analyses 569
SNV Rapport 3075.Recipientkontroll vatten, metodunderlag.Avsnitt 4. Vattenkemi. Naturvrdsverket 1985.
Utermhl H. 1958. Zur Vervollkommnung der quantitativenPhytoplankton-Methodik. Mitt. Int. Verein. Limnol. 9:138.
Veen A., Boekhoud G., Vos W.G. & de la Paz H.J. 2005.Laboratoriumevaluerend onderzoek; Project 341
Fytoplankton in oppervlaktewater 2 november 2005.
Rijkswaterstaat, Rijksinstituut voor Integraal Zoetwater-beheer en Afvalwaterbehandeling RIZA.
Welch P.S. 1948.Limnological methods. The Blakiston Com-pany, Philadelphia & Toronto.
Willn E. 2001. Phytoplankton and water quality charac-terization: experiences from the Swedish large lakesMalaren, Hjalmaren, Vattern and Vanern. Ambio 30:529537.