Corina`s thesis: “The Carpet of the Sun”

Transkript

The Carpet of the Sun
On the Quantification of Algal Biomass
Corina Carpentier
Dissertation thesis
MASARYK UNIVERSITY, Faculty of Science,
RECETOX – Research Centre for Environmental Chemistry and
Ecotoxicology, Brno, Czech Republic
Supervisors:
Prof. Blahoslav Maršálek, Ph.D. (Masaryk University Brno)
Prof. dr. ir. Nick van de Giesen (Technical University Delft, Civil
Engineering & Geosciences)
Brno, 2014
ISBN: 978-94-91602-23-8
Druk: Print Service Ede
Layout: Sinds 1961 | grafisch ontwerp
Cover photo: Reflection of algae in oxygen bubbles; by Anja Carpentier
Title: The title of this thesis was taken from the song “Carpet of the Sun”,
from the album “Ashes are Burning” (1973) by the English band
‘Renaissance’. I cannot think of a more appropriate way to describe algae,
as they are a product of the sun’s light and warmth.
Colophon
Bibliographic information
Author:
Corina Carpentier
Title of dissertation:
The Carpet of the Sun On the quantification of algal biomass
Title of dissertation
Koberec slunce - Jak kvantifikovat biomasu
(in Czech):
řas?
Ph.D. study program:
Chemistry
Specialisation:
Ecotoxicology
Supervisors:
Prof. Blahoslav Maršálek, Ph.D.
(Masaryk University) and
Prof. dr. ir. Nick van de Giesen
(Technical University Delft, Civil
Engineering & Geosciences)
Year of defence:
2014
Key words:
phytobenthos, phytoplankton,
cyanobacteria, fluorescence, biomass,
biovolume
Key words (in Czech):
fytobentos, fytoplankton, sinice,
fluorescence, biomasa, biovolume
3
Abstract
Abstract (English)
There are many different ways to determine algal densities or biomass.
There is no such thing as the best method to analyse algal biomass. The
decision on what method to use largely depends on the purpose of the
study performed. The objectives of this study were related to various
different ways of determining algal biomass, such as by algal biovolume
combined with cell counts, measuring in vivo absorption or in vivo
fluorescence.
Algal cell counts and biovolume
A semi-automated method to determine algal biovolume while performing
microscopic cell counts was developed with the help of digital image
processing technology. Firstly, geometric shapes were defined and
assigned to phytoplankton taxa. Complex shapes were converted to
combinations of geometrical shapes and measurable dimensions were
defined, such as for Staurastrum and Ceratium. Additionally, it was
necessary to harmonise the counting strategy with the biovolume
measurements for filamentous and coenobium-forming algae, thereby
introducing the “counting-unit”. Secondly, the database and measurement
routine of the BACCHUS application were developed, including a statistical
evaluation method. The statistical analysis was used for two purposes:
1. to define the required precision of the biovolume results in order to
define a maximum number of measurements to be performed, thus
minimising the effort involved;
2. to enable a statistical evaluation of both biovolume results and
individual dimensions, in order to reduce the number of taxa or
dimensions to be measured.
The statistical analysis made this application a very powerful tool, since it
helped to minimise the amount of time and effort needed to perform a
suitably precise biovolume analysis. Practice in the laboratory with several
technicians showed that plankton counts can be combined with biovolume
measurements very easily and comfortably. It has to be noted that in
routine samples not enough individuals of a taxon may be present in the
sample to achieve the desired precision. However, even if no image-
5
Abstract
analysis system is used to measure biovolumes, the statistical analysis
can be used to reduce the time needed for analysis.
In vivo absorption
The newly developed European Guidance Standard EN 16161 (2012) on in
vivo absorption methods to determine chlorophyll-a concentrations was
compared to other commonly used methods for the determination of
chlorophyll-a, such as ISO 10260 (1992). The decomposition of absorption
spectra by means of Gauss curves as previously shown by numerous other
researchers results in adequate fit curves, although a freely modifiable 3
(for chlorophyll-a absorption around 675 nm) yields better results than a
fixed value at 675 nm. It appeared to be impossible to correlate the EN
16161 results to chlorophyll-a concentrations as determined with the ISO
10260 method. Thus, the applicability of the EN 16161 Guidance Standard
remains questionable until this issue has been resolved.
In vivo fluorescence methods to determine phytoplankton biomass
An Alert Level Framework for harmful cyanobacteria based on microscopic
cell counts has a major disadvantage: microscopic cell counts are
generally time-consuming, and as such valuable time can be lost waiting
for the results, whereas adequate and timely actions are required to
ensure public safety. Online and on-site fluorescence techniques have the
advantage of being able to analyse samples at a higher frequency and
produce results faster. These features are especially important in the
monitoring of drinking water intake points where toxic algae blooms may
appear suddenly, requiring immediate action. In the study, however, this
required a conversion of the Alert Levels units from cell densities to
cyanobacterial chlorophyll-a levels. This thesis shows how the conversion
from cell counts to in vivo fluorescence measurements of chlorophyll-a
was accomplished, including an example of the application of this Alert
Level Framework in a lake near Warsaw, Poland.
In vivo fluorescence methods to determine phytobenthos biomass
Especially in rivers, phytobenthic primary productivity is considered to be
a suitable indicator of stream water quality, since the short-lived
organisms are generally sessile, take up nutrients from the water phase
and respond rapidly to a changing environment. Moreover, benthic algal
6
Abstract
communities are important parameters in ecological investigations since
they support lotic food webs as a food source for grazers and they can
attenuate currents and stabilise sediments, thereby modifying the aquatic
habitat.
Empirical regression models which link algal biomass and water column
nutrients have been widely used in the eutrophication management of
freshwater lakes and reservoirs. Establishing similar relationships for
nutrients and phytobenthos biomass in streams has appeared more
difficult and is thus less advanced, since these links are generally weaker
in streams than in lakes. Problems caused by enrichment – such as
aesthetic degradation, loss of pollution-sensitive invertebrate taxa through
smothering of substrata by algae, clogging of water intake structures and
degradation of water quality (particularly dissolved oxygen and pH)
resulting in fish kills – may be addressed by such relationships, thus
providing an objective framework for stream water quality management.
Additionally, eutrophication models for streams can be extended to predict
ecosystem properties important to lotic food web structures.
This study has taken the first step in this direction by developing a
suitable field method to measure phytobenthos biomass per algal group,
without the problems related to quantitative sampling. Some assessment
results of the first elaborate dataset for the Danube River are shown,
including phytobenthos biomass as a parameter. However, further
datasets of different types of streams and rivers now need to be collected
and evaluated. The field data thus collected can serve as reliable input for
the establishment of a model for the nutrient-phytobenthos biomass
relationship, in order to assess the trophic state of rivers and streams.
Further research is then necessary to actually develop such a model and
identify class boundaries for phytobenthos biomass. Once developed, it
may provide a practical routine method to assess the trophic state of
rivers and streams as part of their ecological status in accordance with the
Water Framework Directive.
7
Abstract
Abstrakt (Czech)
Je známo mnoho různých způsobů jak stanovit množství řas, či jejich
biomasu. Není však jedna nejlepší metoda
použitelná pro analýzu
biomasy řas. Rozhodnutí o tom, která metoda je nejvhodnější závisí
především na účelu, pro který metodu vybíráme. Cíle této práce se
vztahují k různým metodám a způsobům stanovení biomasy řas, jako
například objemové biomase v kombinaci s počty buněk, měření absorpce
in vivo nebo in vivo fluorescence.
Počty buněk a objemová biomasa
Poloautomatická
metoda
pro
stanovení
objemové
biomasy
řas
za
současného počítání buněk pod mikroskopem byla vyvinuta za pomocí
digitální
analýzy
obrazu.
Nejdříve
jsou
vyhodnoceny
geometrické
parametry, které jsou přiřazovány k jednotlivým taxonům fytoplanktonu.
Celkové obrysy
jsou převedeny na geometrické tvary a měřitelné
dimenze jsou definovány jako například Staurastrum a Ceratium. Zaprvé
bylo nutné harmonizovat strategii počítání s měřením objemové biomasy
pro vláknité a cenobiální řasy zavedením “počítací jednotky”. Zadruhé,
statistická vyhodnocovací metoda byla vyvinuta tak, aby bylo možno
použít databázi pro měření a analýzy (aplikace BACCHUS). Statistická
analýza byla použita pro dva účely:
1. Definovat
požadovanou
přesnost
objemové
biomasy
ve
smyslu
definovat minimální počty měření které musí být realizovány a tím
minimalizovat nutné úsilí.
2. Umožnit
statistické
vyhodnocení
objemové
biomasy
a
měření
jednotlivých buněk s cílem redukovat počty taxonů a jejich rozměrů,
které musí být měřeny.
Díky statistickému vyhodnocení se stala tato aplikace velmi silným
nástrojem, který minimalizuje čas a úsilí nutný k získání dostatečně
precizních dat o objemové biomase. Praktické odzkoušení v laboratory s
několika techniky ukázalo, že počty buněk mohou být kombinovány s
objemovou biomasou snadno a komfortně. Je dobré zmínit, že při rutinním
zpracování vzorků není přítomno dostatek jedinců pro získání dostatečné
přesnosti. Ikdyž není využita analýza obrazu pro stanovení objemové
biomasy, statistická analýza šetří čas nutný pro zpracování vzorku.
8
Abstract
In vivo absorpce
Nově vytvořená evropská směrnice European Guidance Standard EN
16161 (2012) pro detekci koncentrace chlorofylu-a in vivo absorpční
metodou byla porovnána s dalšími běžně používanými metodami pro
detekci chlorofylu-a, jako je ISO 10260 (1992). Rozložení absorpčního
spektra pomocí Gausovy křivky, jak již dříve prokázaly výsledky řady
vědců,
dává
dobrou
shodu
pro
parametry
křivky,
i
když
volně
modifikované 3 (pro chlorofyl– a absorbce okolo 675 nm), dává lepší
výsledky než fixní vlnová délka 675nm. Srovnávat výsledky dle metody
EN 16161 s ISO 10260 se zdá nemožné a proto použitelnost EN 16161
Guidance Standard zůstává problematická, dokud tento problém nebude
vyřešen.
Metoda stanovení biomasy fytoplanktonu pomocí in vivo fluorescence
Mikroskopická metoda stanovení počtu buněk toxických sinic, na které je v
legislativě
založena
Signální
koncentrace
představuje
časově
velmi
náročnou metodu, kdy je prodleva ve získání výsledků, zatímco pro
veřejné zdraví je vyžadována rychlá aktivita. Online a in-situ fluorescenční
techniky mají výhodu vysoké frekvence a rychlosti pořízení dat. Tyto
možnosti jsou důležité speciálně pro monitoring pitných vod, především v
povrchových zdrojích, kde se toxické sinice mohou vyskytovat. Předložená
studie přináší konverzní parametry převodu počtu buněk sinic používaných
v signálních úrovních platné legislativy na koncentrace chlorofylu-a.
Dizertační práce ukazuje realizaci převodu počtu buněk na in vivo
fluorescenci kvantifikovanou jako chlorofyl-a, včetně příkladu aplikace
těchto signálních hladin v nádrži v Polsku poblíž Varšavy.
Stanovení biomasy fytobentosu pomocí in vivo fluorescence
Primární produkce fytobentosu je především v případě řek považována za
přijatelný indicator kvality vody, protože mikroorganismy s krátkým
životním cyklem přijímají živiny a rychle reagují na měnící se podmínky.
Navíc společenstva fytobentosu jsou důležitým ekologickým parametrem
podporovaným vodními proudy a závisející na predaci zoobentosem, čímž
významně modifikují vodní prostředí.
Empirické modely spojené s biomasou řas a živinami ve vodném sloupci
byly široce využívány pro management trofie jezer a přehrad. Stanovení
9
Abstract
podobných vztahů řas a živin pro toky se jeví mnohem složitější a proto
méně pokročilé, protože tyto vztahy jsou mnohem slabší než v jezerech.
Problémy vzniklé obohacením vody živinami – estetické, ztráta zoobentosu
citlivého na znečištění, přerůstání substrátů fytobentosem a degradace
kvality vody (například fluktuace
rozpuštěného kyslíku, či zvýšené pH)
máji za následek například úhyny ryb a proto jsou tématem pro
management kvality vody toků. Navíc - modely pro hodnocení eutrofizace
toků by měly být rozšířeny o parametry, které jsou schopny hodnotit
parametry důležité pro potravní sítě lotických ekosystémů.
Tato práce vykonala první krok pro vývoj vhodné terénní metody
kvantifikace celkové biomasy a skupin řas fytobentosu, která není
omezena
problémem
spojeným
s
kvantitativním
vzorkováním.
Do
hodnocení řeky Dunaje byly využity data kvantifikace fytobentosu jako
parametru kvality vody. Tato data budou dále využita ke srovnání a
korelacím s dalšími parametry kvality vody. Reálná data z terénu mohou
být použita k
nastavení modelu trofie toků pomocí vztahu živiny-
fytobentos. Další výzkum může navázat na stanovení skupin fytobentosu a
jejich vztahy v různém trofickém prostředí. Takto vyvinutý systém může
být použit pro hodnocení trofie toků jako část hodnocení ekologického
stavu v souladu s Rámcovou směrnicí o vodách.
10
Contents
Contents
Bibliographic information
3
Abstract (English)
5
Abstrakt (Czech)
8
List of original articles
17
The author’s contribution to the articles
18
Aims of the present dissertation thesis
19
1.
General Introduction
21
1.1. Planktonic and benthic microalgae
21
1.2. Problems caused by algae
22
1.2.1. Planktonic algae
22
1.2.2. Benthic algae
23
1.3. Analysing algae
23
Section 1: Phytoplankton biomass: Biovolume and in vivo
absorption
2.
Introduction to phytoplankton biovolume analysis
3.
Rapid and versatile routine measurements of phytoplankton
29
biovolumes with BACCHUS
33
3.1. Introduction
33
3.2. Materials and methods
34
3.2.1. Hardware
34
3.2.2. Basic and complex shapes
34
3.2.3. Measuring the dimensions of phytoplankton taxa
40
3.2.4. Master tables in the BACCHUS set-up
47
3.2.5. Measurements
48
3.2.6. Statistical analysis of the results
54
3.2.7. Data retrieval
58
3.3. Evaluation and assessment of results
58
3.3.1. Evaluation method
59
3.3.2. Fixed biovolumes
60
3.3.3. Fixed ratios between dimensions
65
3.3.4. Impact of regular evaluation
65
11
Contents
3.4. Discussion and conclusions
4.
66
3.4.1. The BACCHUS application
66
3.4.2. Recent developments
68
European Guidance Standard EN 16161 is incomparable
to ISO 10260 for the determination of phytoplankton
chlorophyll-a
71
4.1. Introduction
71
4.2. Materials and method
72
4.3. Results
76
4.4. Discussion
85
Section 2: Phytoplankton and cyanobacterial biomass: In vivo
fluorescence of chlorophyll-a
5.
Introduction to fluorescence measurements for the
assessment of algal or cyanobacterial chlorophyll-a
91
5.1. Photosynthesis and fluorescence
91
5.2. Algal fingerprints: method to distinguish algal groups on the
basis of their pigment composition
93
5.3. Comparison of fluorescence measurements with other
methods for the determination of algal biomass
95
5.3.1. Fluorescence measurements versus spectrometric
analysis of chlorophyll-a
95
5.3.2. Fluorescence measurements versus biovolume
analysis
5.4. Control samples and sample storage
6.
96
97
5.4.1. Freezing algal cells
98
5.4.2. Diluting algal samples
98
5.4.3. Alternative control samples
99
Establishing an Alert Level Framework for cyanobacteria in
drinking water resources using the Algae Online Analyser
12
for monitoring cyanobacterial chlorophyll-a
101
6.1. Introduction
101
103
Contents
6.3. Results
105
6.3.1. Seasonal changes of phytoplankton biovolume and
community structure
105
6.3.2. Seasonal changes of chlorophyll-a concentration
measured by spectrophotometry
106
6.3.3. Seasonal changes of chlorophyll-a concentration and
phytoplankton community structure measured by the
AOA
6.3.4. Seasonal changes of microcystins concentrations
6.4. Discussion
106
109
109
6.4.1. Relationship between concentrations of chlorophyll-a
measured by fluorescence (AOA) and by
spectrophotometry
109
6.4.2. Relationship between cyanobacterial biovolume and
concentration of chlorophyll-a of cyanobacteria
determined by fluorescence (AOA)
111
6.4.3. Estimation of hazard from cyanobacterial toxins
based on the concentration of cyanobacterial
chlorophyll-a measured by fluorescence (AOA)
7.
113
Cyanobacterial and total chlorophyll-a concentrations as a
method for early-warning water quality monitoring
119
7.1. Introduction
119
120
7.3. Results
122
7.4. Discussion
125
Section 3: Phytobenthos biomass: In vivo fluorescence of
chlorophyll-a
8.
Introduction to in situ and in vivo fluorescence measurements
to determine phytobenthos biomass
129
8.1. From phytoplankton to phytobenthos
129
8.2. Considerations regarding field fluorescence measurements
of phytobenthos biomass
129
8.2.1. Preparations
130
8.2.2. Measurement procedure
132
13
Contents
8.3. Fluorescence measurements versus HPLC-analysis of
chlorophyll-a
133
8.3.1. Sampling or field measurements?
133
8.3.2. Comparison of fluorescence measurements with
HPLC-analysis according to DIN 38412-16
8.4. Spatial variability of phytobenthos
9.
137
139
The influence of hard substratum reflection and calibration
profiles on in situ fluorescence measurements of benthic
algal biomass
145
9.1. Introduction
145
9.2. Materials & Methods
148
9.2.1. Chlorophyll-a fluorescence measurements
148
9.2.2. Dye filter and algae
151
9.2.3. Substrata
152
9.2.4. Field trial
154
9.3. Results
155
9.3.1. Experiment 1: Dye filter measurements on a black
and white background
155
9.3.2. Experiment 2: Dye filter measurements on natural
stones
156
9.3.3. Experiment 3: Algae measurements on natural and
black substrata
157
9.3.4. Experiment 4: Dye filter measurements on natural
substrata revisited
163
9.3.5. Experiment 5: field trial of natural algae on natural
substrata
166
9.4. Discussion
168
9.5. Implications
171
9.6. Conclusions
171
10. Joint Danube Survey: a field example
173
10.1. Introduction
173
10.2. The Danube River
173
10.3. Phytobenthos biomass and abiotic factors
175
10.4. Phytobenthos biomass and (re-)colonisation from tributaries 180
10.5. Phytobenthos biomass and nutrient levels
183
10.6. Phytobenthos biomass and phytoplankton density as total
chlorophyll-a
14
183
Contents
10.7. Phytobenthos biomass and benthic invertebrates (grazing)
184
10.8. Phytobenthos biomass compared to benthic diatoms
186
10.9. Concluding remarks
188
11. General Discussion and Summary
189
11.1. Objective 1: Development of a semi-automated method to
determine plankton biovolume while performing microscopic
cell counts
189
11.2. Objective 2: Assessment of the in vivo absorption method
and comparison to other methods for the determination of
phytoplankton chlorophyll-a
190
11.3. Objective 3: Development of an Alert Level Framework
based on in vivo cyanobacterial fluorescence to protect
drinking water resources
190
11.4. Objective 4: Development of an in situ and in vivo
fluorescence method for the quantification of phytobenthos
per algal group
191
12. References
195
13. Acknowledgements
215
15
Publications
List of original articles
1. Carpentier, C.J., H.A.M. Ketelaars, A.J. Wagenvoort and C.P.R.
Pikaar-Schoonen. 1999. Rapid and versatile routine measurements of
plankton biovolumes with BACCHUS. Journal of Plankton Research 21:
1877-1889.
2. Izydorczyk, K., C. Carpentier, J. Mrowczyński, A. Wagenvoort, T.
Jurczak, and M. Tarczyńska. 2009. Establishment of an Alert Level
Framework for cyanobacteria in drinking water resources by using the
Algae Online Analyser for monitoring cyanobacterial chlorophyll-a.
Water Research 43:989-996.
3. Carpentier, C., A. Dahlhaus, N. van de Giesen and B. Maršálek. 2013.
The influence of hard substratum reflection and calibration profiles on
in situ fluorescence measurements of benthic microalgal biomass as an
indicator of trophic state. Environ. Sci.: Processes Impacts, 15, 783793.
4. Kalaji, H.M., M. Skonieczny, O. Sytar, M. Brestic, K. Bosa, S.
Pietkiewicz and C.J. Carpentier. Cyanobacterial and total chlorophylla concentrations as a method for
early-warning water quality
monitoring. Submitted to CLEAN Soil, Air, Water. 5 April 2013.
5. Moldaenke, C., S. Legell and C.J. Carpentier. European Guidance
Standard
EN
16161
is
incomparable
to
ISO
10260
for
the
determination of phytoplankton chlorophyll-a. Submitted to Journal of
Plankton Research. 6 April 2014.
17
Publications
The author’s contribution to the articles
1. C. Carpentier worked for three years on the development of the
BACCHUS programme, implemented the biovolume measurement
method at the laboratory of the Water Storage Company Brabantse
Biesbosch, evaluated the first datasets and prepared the publication.
2. C. Carpentier supervised a Master’s student from the Univeristy of
Łodz (Poland) for this work and supervised the evaluation of the data
and preparation of the publication.
3. C. Carpentier performed the field work and the data-analysis,
participated in the discussions on the technical improvement of the
fluorometer, and prepared the publication.
4. C. Carpentier performed the data-analysis and assessment in relation
to the Alert Level Framework, and wrote large parts of the publication.
5. C. Carpentier analysed the data, participated in the discussions on the
results and wrote considerable parts of the publication.
18
Aims
Aims of the present dissertation thesis
The overall aim of this thesis was to study various methods for the
quantification of algae, in order to develop or improve analytical tools for
(ecological) assessments of phytoplankton and phytobenthos biomass
developments.
Specific objectives were to:
1. develop a semi-automated method to determine plankton biovolume
while performing microscopic cell counts in order to reduce the time
and effort necessary for analysis;
2. assess the newly developed in vivo absorption method and compare it
to other methods for the determination of phytoplankton chlorophyll-a;
3. develop an Alert Level Framework based on in vivo cyanobacterial
fluorescence rather than microscopic cell counts, in order to protect
drinking water resources from harmful cyanobacteria;
4. develop
an
in
situ
and
in
vivo
fluorescence
method
for
the
quantification of phytobenthos per algal group.
19
General introduction
1.
General Introduction
1.1.
Planktonic and benthic microalgae
Approximately 600 million years ago, marine invertebrate phyla entered
the fossil record (the so-called Cambrian explosion). In the 2,500 million
years before that, approximately 85% of the earth’s geological history,
the only organisms to have populated the world were bacteria and
cyanobacteria (Begon et al., 1990).
Prokaryotic cyanobacteria evolved at least 2,500 million years ago from
photosynthetic bacteria which developed the metabolic machinery to use
water as a source of electrons and hydrogen for fixing carbon dioxide
(Campbell, 1993). These organisms are the evolutionary ancestors of all
modern plants, using chlorophyll-a and two photosystems to convert
sunlight into organic biomass by means of photosynthesis (see Chapter
3). Thus, cyanobacteria were the first organisms in evolutionary history to
produce oxygen (Huisman et al., 2005).
Cyanobacteria and eukaryotic phytoplankton (“true algae”) form the basis
of the aquatic foodweb in the pelagic zone of lakes and large rivers and
streams (Huisman et al., 2005). Although many motile flagellates are
classified as protozoa on the basis of their morphological and reproductive
characteristics,
these
organisms
are
treated
collectively
with
cyanobacteria and eukaryotic phytoplankton since their primary nutritional
source is autotrophic photosynthesis, which results in the production of
new organic matter. Throughout this thesis, these groups together are
referred to as (planktonic) algae or phytoplankton.
Planktonic algae are mostly found in lakes and reservoirs. Generally,
phytoplankton is unable to maintain populations in fast-flowing streams,
since their doubling rates do not exceed downstream losses due to
currents. In slow-moving rivers and backwaters, planktonic algae may
develop a stable community. Benthic algal communities occur on virtually
all substrata which receive light and are found in nearly all running
waters, thus forming an important part of the fluvial food web (Allan &
Castillo, 2007).
21
1.2.
Problems caused by algae
1.2.1. Planktonic algae
The presence of oxygen-producing algae is extremely important for all life
on earth: approximately 70-80% of all oxygen is believed to be produced
by (marine) algae. Moreover, algae produce carbohydrates, oils, proteins,
vitamins and organic minerals and are thus used in food, animal feed,
cosmetics, pharmaceuticals, biofuels, and so on. However, despite these
positive aspects, massive algal growth is an unwanted phenomenon in
waters used as a source for the production of drinking water, since it can
cause several problems during treatment processes (Clasen, 1994;
Petruševski, 1996):
•
algae which enter a treatment plant using chlorine for disinfection
purposes increase chlorine demand and act as precursors for the
formation of trihalomethanes and other halogenated by-products, such
as chloroform;
•
certain phytoplankton species can produce elevated levels of toxins, or
negatively affect the taste and odour of drinking water;
•
algae can disrupt the coagulation process, resulting in a higher
coagulant dose;
•
colony or thread-forming algal species in particular can cause the
clogging of filters;
•
algae which pass filters can contribute to biological aftergrowth in the
distribution network.
The production of toxins or taste and odour compounds predominantly by
cyanobacteria not only poses a threat to drinking water quality but can
cause
problems
for
recreational
waters.
Exposure
to
(planktonic)
cyanobacteria during recreational activities such as swimming may lead to
rashes, eye irritation and other effects such as nausea and stomach
aches. However, the chances of being exposed to a lethal dose of
cyanobacterial toxins while swimming are small, although children run a
higher risk as they may not be put off by “pea-soup green” water (BurgerWiersma & Versteegh, 1994). Cyanobacteria have been known to produce
toxins in concentrations lethal to domestic animals and cattle since the
nineteenth century (Francis, 1878).
22
1.2.2. Benthic algae
The
mass
occurrence
concentrations
of
the
of
benthic
earthy-musty
cyanobacteria
compounds
can
cause
geosmin
and
high
2-
methylisoborneol in water (Wood et al., 1983). Especially in drinking
water reservoirs and aquaculture plants, it is important to keep benthic
cyanobacterial biomass and thus concentrations of these compounds to a
minimum, since their threshold odour concentrations are very low: 1.3
ng/L for geosmin and 6.3 ng/L for 2-methylisoborneol (Young et al.,
1996). This means that the presence of even very low concentrations of
these substances will adversely affect the taste and odour of drinking
water or fish. An estimated 30% of the potential annual revenue of the US
catfish industry is lost as a result of odour contamination (Smith et al.,
2008).
Additionally, the mass occurrence of benthic algae in rivers and streams
indicates elevated nutrient levels (eutrophication), which can result in
negative aesthetic effects (Welch et al., 1988) or the reduced functioning
of the lotic ecosystem (Dodds, 2006).
1.3.
Analysing algae
In summary, for ecological, commercial and public health reasons, insight
into the development of planktonic and benthic algae in freshwaters is
essential. Thus, adequate methods of monitoring and analysis are
indispensable. Despite the large amount of research carried out on
measurement methods for planktonic and benthic microalgae in the past
decades, large voids still exist in our knowledge of microalgal community
developments.
This thesis presents a discussion on 4 practical problems of analysing
algae and their solutions:
1. Phytoplankton cell counts are a common unit to express algal densities
in surface waters. However, the counting of phytoplankton taxa alone
yields insufficient information on plankton dynamics. Phytoplankton
counts combined with biovolume measurements provide a much better
insight into the relative contribution of the taxa to total biomass.
Conventional methods for determining phytoplankton biovolume (using
a microscope and ocular micrometer) are labour-intensive and prone
23
to errors. Therefore, a new, less time-consuming and more accurate
method was developed using semi-automated digital image processing
techniques.
2. In vivo absorption of chlorophyll-a at 675 nm is considered to be
proportional to the total chlorophyll-a concentration in a sample. A
newly developed European Guidance Standard (EN 16161, 2012) uses
this parameter to quantify phytoplankton in marine and freshwaters,
thereby aiming to provide a simpler alternative to the commonly used
ISO 10260 spectrophotometric method for chlorophyll-a determination
of 1992. This thesis describes a comparison between these two
methods to determine whether EN 16161 indeed provides a valid
alternative to ISO 10260.
3. Total chlorophyll-a is a common parameter to express algal densities
in surface waters. However, total chlorophyll-a analysis by means of
spectrometry alone does not provide sufficient information regarding
specific algal groups, such as cyanobacteria. Microscopic analysis to
determine the ratios of different algal groups and to calculate the
contribution of each group to the total chlorophyll-a content is
laborious and time-consuming. In situ fluorescence measurements
allow for online analysis of cyanobaterial chlorophyll-a in combination
with total chlorophyll-a content. This thesis describes a method to
determine alert levels of harmful cyanobacteria to protect drinking
water resources using in situ fluorescence measurements.
4. Measuring fluorescence is a commonly used method to determine
microphytobenthic biomass expressed as chlorophyll-a per square
centimetre. However, this in situ method suffers from reflection
interference by the substratum on which the benthic microalgae grow.
An automatic correction procedure for substratum reflection was
developed which improves the reliability of the results significantly,
without increasing the time necessary to perform the measurements or
complicating the measurement procedure.
Section 1 describes the study of two methods for phytoplankton biomass
determination: microscopic phytoplankton counting in combination with
biovolume analysis, and in vivo absorption of chlorphyll-a at 675 nm.
Section 2 presents in situ fluorescence measurements of cyanobacterial
chlorophyll-a as an early-warning system for drinking water intake
24
protection. Section 3 concerns the in situ fluorescence measurements of
phytobenthic biomass and the correction procedure for substratum
reflection.
Each section contains an introductory chapter describing the background
and considerations of the respective method under investigation.
25
Section 1
Phytoplankton biomass: Biovolume and in vivo
absorption
Based on:
Carpentier, C.J., H.A.M. Ketelaars, A.J. Wagenvoort and C.P.R. PikaarSchoonen. 1999. Rapid and versatile routine measurements of plankton
biovolumes with BACCHUS. Journal of Plankton Research 21: 18771889.
Carpentier, C.J., A.J. Wagenvoort, H.A.M. Ketelaars and R.A. Sperber.
1997. BACCHUS, software programme for interactive measurements of
biovolume of phytoplankton and zooplankton with statistical evaluation
and data-storage. WBB-report hk/cc/aba 97-35014. Werkendam. N.V.
WBB (in Dutch).
Moldaenke, C., S. Legell and C.J. Carpentier. European Guidance
Standard EN 16161 is incomparable to ISO 10260 for the determination
of phytoplankton chlorophyll-a. Submitted to Journal of Plankton
Research. 6 April 2014.
27
Chapter 2: Introduction to biovolume
2.
Introduction to phytoplankton biovolume analysis
There are many different ways to determine densities or biomass of
phytoplankton. There is no such thing as one single best method for all
possible aims. The decision on which method to use largely depends on
the purpose of the study performed. For biological water assessment,
focus is mainly given to species-richness and the saprobic index.
Therefore, any investigation largely focuses on the qualification of the
phytoplankton, in which case the microscopic evaluation of samples may
be sufficient to serve this purpose. However, if the goal is to study
plankton dynamics, succession or algal toxin production, or to evaluate
the results of the manipulation or management of water bodies, it is
essential to not only qualify (i.e. species diversity), but also to quantify
plankton biomass (in terms of chlorophyll-a, biovolume, carbon content,
etc.).
The diameter of phytoplankton cells varies from 1 to >1000 µm (Semina,
1978; Reynolds, 1984). As is obvious from this large range, the
determination of algal density by counting cells is an inadequate
quantitative
measure
of
phytoplankton
biomass
(Lohman,
1908;
Nauwerck, 1963; Bellinger, 1974; Vilicic, 1985). Performing microscopic
counts can result in a misinterpretation of the sample. A combination of
cell counts and biovolume measurements thus provides a more realistic
image of biomass ratios (Bellinger, 1974; Jiménez et al., 1987). Figure
2.1 shows that the density figures indicate the dominance of Rhodomonas.
However, the biovolume results indicate no such dominance. The cause of
this is the very small biovolume of Rhodomonas compared to that of the
other taxa present.
Many researchers have therefore stressed the use of accurate biomass
estimates in phytoplankton studies (e.g. Rott, 1981). High-performance
liquid chromatography (HPLC), which separates and quantifies indicator
pigments
for
different
phylogenetic
algal
groups,
has
often
been
suggested as an alternative to microscopic phytoplankton enumeration
combined with biovolume measurements (Wilhelm et al., 1991.; Millie et
al., 1993). However, the taxonomic resolution of HPLC-analysis is low and
29
Section 1: Phytoplankton biomass
standardisation is difficult to achieve (Millie et al., 1993; Latasa et al.,
1996; Roy et al., 1996). Colour-image-analysed fluorescence microscopy
does not provide any taxonomic information but can only help to separate
the algae from the cyanobacteria).
A.
B.
Figure 2.1
Phytoplankton counting results of samples from De Gijster
reservoir in 1998 (A) and biovolumes of the phytoplankton taxa
from the same samples (B).
30
Chapter 2: Introduction to biovolume
The biovolume of a phytoplankton taxon varies in space and time, thus a
fixed biovolume per taxon cannot be used to determine biovolume
accurately (Ruttner, 1952; Nauwerck, 1963; Sicko-Goad et al., 1977).
Lohman
(1908)
convincingly
showed
that
the
biovolume
of
(phyto-)plankton genera needs to be determined for individual organisms
and therefore he determined biovolume by describing each taxon as a
standard geometric form or combination thereof. Complex forms were
modelled and the biovolume was determined as the displacement volume.
Ruttner (1952) also used this method, and was followed by (among
others) Nauwerck (1963), Bellinger (1974), Ruttner-Kolisko 1977), Kim
(1980), Rott (1981), Pridmore & Hewitt (1984), Reymolds (1984) and
Elser et al. (1986).
31
Chapter 3: BACCHUS biovolume
3.
Rapid and versatile routine measurements of
phytoplankton biovolumes with BACCHUS
3.1.
Introduction
Plankton biovolumes are usually measured using an ocular micrometer.
However, this is a time-consuming method (Wilhelm et al., 1991; Schmid
et al., 1998). In addition, the use of this instrument may cause errors in
measurements due to variations in the angle and the distance between
the eye of the observer and the eyepiece of the microscope. Variations in
the accuracy of measurements can derive from fatigue of the observer
since they may alter the position of the head relative to the eyepiece, and
also from differences among observers (McCauley, 1984). In addition, the
technician must estimate and manually record length data – a very timeconsuming and a motion-intensive process which is subject to errors (Roff
and
Hopcroft,
1986).
In
a
comparison
of
the
precision
of
a
microcomputer-based measuring system with a conventional microscope
eyepiece micrometer method, Roff and Hopcroft (1986) found that the
precision of measurement of the latter is much lower, leading to much
larger errors in biomass calculations due to compounded errors.
In the last few decades, (semi-)automated measuring systems have been
described to overcome some of these drawbacks. Sprules et al. (1981)
presented a microcomputer-based technique which uses a calliper system
linked to a microcomputer through an analogue-digital converter to
measure a variety of biological features. Data processing and flexible
output calculations in these systems are limited, however. Completely
automated
systems
to
determine
biovolumes
and/or
length
were
developed by Jeffries et al. (1980, 1984), Rolke and Lenz (1984) and
Enserink (1995). The great advantage of this type of measuring system is
the speed with which measurements can be performed. However, this
advantage is largely undone by the time-consuming preparations needed
to make a sample suitable for automatic measurement. Enserink (1995)
clearly showed that forming a well-recognisable image is a very laborious
task. An important problem is caused by organisms overlapping each
other. All organisms have to be separated first or else the computer
considers them as one. Enserink (1995) concluded that a completely
automated measuring system can only be useful if at least 50-100 similar33
shaped organisms are to be measured in one sample. Allen et al. (1994)
also concluded that completely automated measuring systems are quick
but unsuitable for gathering accurate information at species level.
Accurate
identifications of
species
can only be
achieved using
a
microscope. With the help of a semi-automated interactive measuring
system, microscopic counts can easily be combined with measurements of
plankton dimensions. According to Allen et al. (1994), this is probably the
fastest and most accurate way to determine biovolume on a routine basis.
A
semi-automated
measuring
system
called
BACCHUS
(Biovolume
Analysis Cum Calculus and High Utility Statistics) based on an imageanalysis programme linked to a database was developed in the late 1990s
to determine biovolumes of phytoplankton taxa while counting routine
samples. In principle, the biovolume of a given species is determined from
a few dimensions based on a geometric approximation to describe its
volume. Online statistical evaluation of the data reduces the number of
necessary measurements to a minimum.
3.2.
Materials and methods
3.2.1. Hardware
Digital image processing and analyses were performed with a QUANTIMET
500IW IBM-compatible computer (300 MHz, 64 Mb internal memory, 17"
monitor; Leica, Cambridge, UK) with an absolute mouse (digitising tablet,
Wacom, Neuss, Germany), and an inverted microscope (Leica, Wetzlar,
Germany) with either a 1 CCD monochrome or a 3 CCD colour video
camera (Sony, Hoofddorp, The Netherlands) fitted to the drawing tube.
Any microscope used for phytoplankton counting can be used provided
that it can be fitted with a drawing tube. The BACCHUS application used
the database programme Microsoft Access and the image-analysis
software QWin Y2.3a (Leica, Cambridge, UK). Measurement routines were
created using the built-in macro software QUIPS (QUANTIMET imageprocessing system).
3.2.2. Basic and complex shapes
Eleven geometric shapes (and combinations or parts thereof) were used to
describe the shape of different phytoplankton taxa and calculate their
34
biovolume: sphere, half sphere, ellipsoid, cone, cylinder with round or
elliptic base, brick, spindle, discus, barrel and ellipsoid segment (Figure
3.1).
Figure 3.1
Schematic drawing of several geometric forms: 1: sphere; 2.
ellipsoid; 3: cone; 4: cylinder (with round or elliptic base); 5:
brick; 6: spindle; 7: discus; 8: barrel; 9: ellipsoid segment (d:
diameter; l: length; b: width; h: height). Original drawing.
In principle, every shape must be described by three dimensions from
which the biovolume can be calculated. In the case of a sphere, only one
dimension (diameter) needs to be measured, but it is raised to the third
power because all three dimensions have the same value. For cones,
barrels, spindles and cylinders, two dimensions need to be measured; the
third dimension equals one of the others. For all other shapes, length,
width and height are necessary to calculate the biovolume. Since only up
to two dimensions can be visualised under the microscope, at least one
dimension has to be derived from one of the others or a fixed value has to
be determined from a number of specimens in a special effort. Therefore,
the BACCHUS application offers the possibility to add a factor or a fixed
35
value. The factor has to be taken from the literature or (if possible) has to
be calculated from one's own measurements. For example, in the
Biesbosch reservoirs of Evides Water Company in the Netherlands, the
ratio between height and diameter of the diatom Actinocyclus normanii
(cylinder) was determined to be 0.82 (n = 67; minimum = 0.60;
maximum = 1.06).
The BACCHUS programme offers the possibility to add the volumes of
different geometric shapes up to a maximum of 10 individually measured
dimensions. This feature makes it possible to describe complex shapes.
The measurements can be simplified for three combinations of shapes:
cylinder + half sphere, cylinder + cone and cone + half sphere. The total
length of the combined form can be measured instead of the individual
dimensions of each shape, thus simplifying the overall procedure (Figure
3.2). These combinations have therefore been added to the program as
standard shapes.
Figure 3.2
Schematic drawings of the combined forms: 1. cylinder + half
sphere; 2: cylinder + cone; 3: cone + half sphere (l: length; d:
diameter). Original drawing.
Some organisms cannot easily be described by a simple geometric shape
because they have many appendices, e.g. Ceratium (Dinophyceae),
Desmodesmus and Staurastrum (Chlorophyceae). Special formulae were
developed to describe the shape of these three taxa.
36
Desmodesmus
Each individual of Desmodesmus is measured as the length of a cell
(dimension ‘l’) and the width of the entire coenobium (dimension ‘b’), see
Figure 3.3. The width per cell is then derived from dimension b and the
number of cells in the coenobium, after which the biovolume of the
(ellipsoid) cells is calculated. This biovolume is multiplied by the number
of cells in the coenobium in order to arrive at the total biovolume per
coenobium. This elaborate method is necessary, since the cells of
Desmodesmus have a very small width. Measuring the width of several
cells together is therefore easier and more accurate than trying to
measure the width of one individual cell. The biovolume of Desmodesmus
is standardised to a coenobium of 4 cells. The use of standardised
coenobia is further explained in Paragraph 3.2.3.
Figure 3.3
Desmodesmus. bcoen: width of the coenobium; lcell: length of a
cell. Drawing by H.A.M. Ketelaars, after Van Essen, 1974.
ISBN: 978-94-91602-23-8
Druk: Print Service Ede
Ceratium
Layout: Sinds 1961 | grafisch ontwerp
The shape of Ceratium consists of the cell’s body in the shape of an
ellipsoid, an apical horn and 3 additional horns which can be described as
cones with a circular base (Anonymus, 1992; Figure 3.4).
The formula describing Ceratium’s biovolume, is:
37
V


 l  (2, 72  b 2  1, 65  b  (g1  g2 )  2  ( g12  g22 )) 
 (g12  f1  g22  f2  g23  f3  g24  f4 )
48
12
(3.1)
where:
g : width of the horn (g1 is the apical horn, g2 is the horn directly opposite the
apical horn, g3 en g4 are the additional horns);
f : length of the horn (f1 is the apical horn, f2 is the horn directly opposite the
apical horn, f3 en f4 are the additional horns);
l
: length of the cell’s body;
b : width of the cell’s body.
A.
Figure 3.4
B.
C.
Habitus (A) and schematic drawings (B, C) of Ceratium. See also
Formula 3.1. Drawings by H.A.M. Ketelaars; A. after Streble &
Krauter (1988); B, C: original drawing.
It is important to measure the horns in the right order, since the width of
the apical horn and the one directly opposite (g1 and g2) also occur in the
first part of the formula. The order of measurement for the two additional
horns (g3 en g4) does not affect the result, but of course these dimensions
have to be measured in combination with the correct length (f3 en f4).
Delivering the dimensions in the correct order is essential for the
BACCHUS application, in order to ensure that the measured values are
38
included in the right place of the formula. To prevent errors, BACCHUS
therefore clearly requests specifically described dimensions, using labels
as indicated in Figure 3.4.
Staurastrum
Staurastrum can be described as two cell bodies, each in the shape of a
truncate pyramid, and a number of prickles in the shape of a circular
cone. Figure 3.5 shows Staurastrum with 6 prickles, as found in the
Biesbosch reservoirs in the Netherlands. However, Staurastrum can also
occur with more than 6 prickles. Therefore, the formula for the calculation
of Staurastrum’s biovolume is given for a variable number of prickles. The
volume of a circular cone can be described as:
V
1
   d2  h
12
(3.2)
where:
V : volume;
d : diameter at the base;
h : height.
The volume of a truncate pyramid is described as:
V
1
 h  (Ag  Ab  Ag  Ab )
3
(3.3)
where:
Ag : surface of the base (equilateral triangle);
Ab : surface of the top (equilateral triangle);
h : height.
The surface area of an equilateral triangle is described as:
A
1 2
l  3
4
(3.4)
where:
A : surface area;
l
: length of the sides of the triangle.
The biovolume of Staurastrum is calculated by adding the volume of two
truncate pyramids (Formula 3.3 and 3.4) to the volume of as many
circular cones as there are prickles (Formula 1.2). This results in the
following formula (see also Figure 3.5 for dimensions):
39
V
2
1
2
2
 h  (  3  ( l mi
 l bo
)
3
4
3 2 2
1
 l mi  l bo ) 
 N    d 2st  l st
16
12
(3.5)
where:
h : heigth of the cell body;
lmi : length of the middle of the cell body;
lbo : length of the top side of the cell body;
lst : length of a prickle;
dst : diameter of a prickle;
N : total number of prickles.
A.
Figure 3.5
B.
Staurastrum; A: view from above; B: side-view. See also Formula
3.5. Ddrawings by H.A.M. Ketelaars; after: Carter (1923).
3.2.3. Measuring the dimensions of phytoplankton taxa
Loricae and mucous layers
Generally, only the cell body is measured when determining biovolume.
Some taxa, such as Dinobryon and Kephyrion, have a lorica; others, such
as Crucigenia, Dictyosphaerium, Radiococcus and Oocystis have a mucous
layer (Figure 3.6). Loricae and mucous layers are not included in the
biovolume calculations.
Microcystis
The biovolume of Microcystis is based on a amorphous colony with a
thickness of three layers of cells (125 cells per square of 50 x 50 µm). The
number of cells on the largest surface area of the colony is counted using
the fine knob on the microscope. Microcystis is not the only taxon which
occurs as amorphous colonies. Other examples are Merismopedia and
Coelosphaerium.
40
3.
2.
1.
Figure 3.6
4.
5.
Dimensions to be measured of several taxa: a. Dinobryon, b.
Crucigenia, c. Dictyosphaerium, d. Radiococcus, e. Oocystis (l:
length; b: width; d: diameter). Drawings by H.A.M. Ketelaars (1,
4: after Streble & Krauter, 1988; 2, 3, 5: after Van Essen, 1974).
Coenobia
Eleven taxa of the Chlorophyceae class are not counted as single cells, but
as a standardised coenobium. Table 3.1 shows these taxa, and indicates
whether single cells or the entire coenobium are measured when
determining biovolume. Taxa for which individual cells are measured do
not pose a problem, since the formula of the geometric shape of the cells
can easily be multiplied by the number of cells in the coenobium to arrive
at the total biovolume of the counting unit.
Of the other six taxa, four have coenobia of a variable number of cells:
Coelastrum, Eudorina, Pandorina and Pediastrum (Figures 3.7, 3.8 and
3.9). The biovolume measurement of Eudorina (as shown in Figure 3.7) is
comparable to Pandorina. The counting units of these two taxa are
coenobia of 16 cells.
41
Table 3.1
Phytoplankton taxa with a coenobium as a counting unit.
taxon
no. of cells/
to be
dimension to be measured
coenobium
measured
Coelastrum
16
coenobium
diameter
Crucigenia
4
coenobium
length (without mucous)
Dictyosphaerium
8
cell
diameter (without mucous)
Eudorina
16
coenobium
diameter (without mucous)
Oocystis
4
cell
length
Pandorina
16
coenobium
length, width (without
mucous)
Pediastrum
16
coenobium
diameter
Radiococcus
4
cell
diameter
Desmodesmus
4
coenobium
cell length, coenobium width
Sphaerocystis
8
cell
diameter
Tetrastrum
4
cell
diameter
Figure 3.7
Eudorina. l: length; b: width. Drawing by H.A.M. Ketelaars after:
Streble & Krauter (1988).
When measuring the coenobium as a unit, a standard counting unit for a
coenobium (4, 8 or 16 cells) is necessary. Together with the actual
number of cells in the coenobium, a conversion factor is created to
convert the real biovolume to a standardised biovolume. This conversion
factor is calculated as the standard number of cells in a coenobium divided
by the actual number of cells in the coenobium, and is necessary to
perform statistical calculations on the results. Because of cell division
42
behaviour, the number of cells in a coenobium is always a power of two.
Calculation of the mean biovolume of two coenobia, one of 16 cells and
one of 32 cells, would result in a meaningless biovolume of a non-existing
coenobium with 24 cells. This problem is solved by converting the
biovolume of coenobia to standardised biovolumes of a fixed number of
cells (a power of two). The actual number of cells is stored in the database
because this information can be useful when real coenobium size is
necessary, e.g. for grazing studies.
The calculation of the standardised biovolume is done as follows. It is
assumed that all cells in a coenobium have the same biovolume, and the
total biovolume thus depends only on the number of cells present in the
coenobium. As a result, the biovolume of a coenobium of 32 cells will be
exactly twice the volume of a coenobium of 16 cells. The formula for
calculating the standardised biovolume is then:
V16 
16
 V32
32
(3.6a)
where:
V16 : volume of a 16-cell coenobium;
V32 : volume of a 32-cell coenobium.
Instead of 32, one can also enter ‘x’ for the number of cells in the
coenobium, thus generalising the formula:
V16 
16
 Vx
x
(3.6b)
If for example the formula for a cylinder is used to calculate the biovolume
(e.g. for Pediastrum), formula 3.6b is changed as follows:
1
16 1
2
   d16
h 
    d232  h
4
32 4
(3.7)
The height of Pediastrum cannot be measured, because the thin
coenobium never settles on its side. By adding a few drops of glycerol to
the sample under the microscope, Pediastrum can be manipulated to
stand on its side, so that the height of the cylinder can be measured
(Figure 3.8). The height of Pediastrum was thus determined to be 6.1 µm
(n = 9; minimum = 4.1 µm maximum = 10.2 µm), regardless of the
number of cells in the coenobium.
43
Figure 3.8
Pediastrum, A: side-view, B: view from above. h: height; d:
diameter. Drawings by H.A.M. Ketelaars.
Simplifying Formula 3.7 while assuming a constant value for h, results in:
2

d16
d
1 2
 d32  d16  32
2
2
(3.8)
or generally speaking:
dx
d16 
x
16
(3.9)
where:
x
: number of cells in the coenobium.
Coelastrum and Eudorina both have spherical shapes. When incorporating
the geometric formula of a sphere in Formula 1.6a, a standardised
diameter for a coenobium of 16 cells can be calculated:
d
1
1 1
1
3
3
   d16
     d332  d16
  d332  d16  3 32
6
2 6
2
2
(3.10)
More generally speaking, the formula is:
d16 
dx
3
x
16
where:
d16 : diameter of a 16-cell coenobium;
44
(3.11)
dx : diameter of a coenobium with x cells;
x
For Pandorina a similar relationship is valid, but since Pandorina’s shape is
an ellipsoid, the length and width of a 16-cell coenobium is derived as:
1
1 1
1
2
2
   l16  b16
     l32  b 232  l16  b16
  l32  b 232
6
2 6
2
(3.12)
where:
l16 : length of a 16-cell coenobium;
b16 : width of a 16-cell coenobium;
l32 : length of a 32-cell coenobium;
b32 : width of a 32-cell coenobium;
It is assumed that the length and width of the ellipsoid increase in size at
a constant ratio. Thus follows:
b16 
l16 
b 32
3
2
l32
3
(3.13a)
(3.13b)
2
or more generally:
bx
b16 
3
x
16
(3.14a)
lx
l16 
3
x
16
(3.14b)
where:
x
With the help of the standardised biovolume (V16) the running mean of the
results can be calculated, including the statistical precision of the results,
as will be explained in Section 3.2.6. Without a standardised biovolume,
this would not be possible.
Crucigenia
Generally, Crucigenia (Figure 3.6b) only occurs as coenobia of four cells.
Very seldomly, other numbers of cells per coenobium are found, but since
these are rare exceptions, these individuals are not included in the
45
biovolume measurements. Only 4-cell coenobia are measured. The
biovolume of this taxon thus does not need to be corrected to a fixed
number of cells.
Coelastrum
For Coelastrum, a specialised formula for calculating its biovolume is used.
Coelastrum forms spherical coenobia of spherical cells (Figure 3.9).
Counting these cells is very difficult, since only a two-dimensional image
of the spherical coenobium is visible under the microscope. By counting
the number of cells in the outer row (light grey cells in Figure 3.9) and
adding twice the number of cells within this outer row (front and back; the
dark grey cells in Figure 3.9), the total number of cells in the coenobium
can be determined. This is always a power of 2.
The biovolume of Coelastrum is calculated by measuring the diameter of
one cell, calculating the biovolume of a sphere and then multiplying this
number by the counted number of cells in the coenobium. The biovolume
of Coelastrum is standardised to 16 cells.
Figure 3.9
Schematic view of a spherical coenobium of Coelastrum. Original
drawing.
Filamentous taxa
Filamentous taxa also have a counting unit which is not a single cell.
These taxa are for example: Oscillatoria, Planktothrix, Aulacoseira,
Melosira varians, and Stephanodiscus. The counting unit for these taxa is
the length of 4 squares of the ocular raster, thus 200 µm. this counting
unit was chosen because the counting of individual cells in a filament is a
46
very laborious task which does not provide much relevant information.
Filaments shorter than 200 µm are counted as parts of 200 µm. The
biovolume of these taxa is calculated by measuring the diameter of the
filaments and entering this value in the geometric formula for a cylinder
with a spherical base. The length of the cylinder is then always 200 µm,
regardless of the number of cells present. The diameter of the filament is
measured only once per filament, as the cells are all similar and their
diameter is assumed to be the same.
Using a filament as a counting unit is meaningless, since a breach of the
filament as a result of for example the sample preparation would then
result in a different number of counting units.
3.2.4. Master tables in the BACCHUS set-up
For a clear understanding of the names of tables, fields and buttons used
in the application, the following layout was used: database tables are in
bold capitals (e.g. T_CURRENT), table fields are in bold (e.g. sample),
buttons are in bold italic (e.g. Data evaluation).
Before
measurements
can
be
performed,
taxon-specific
data
and
information on magnification have to be entered in the database master
table (Figure 3.10).
Figure 3.10
Form in which the taxon-specific data of BACCHUS are entered.
All essential information to calculate the biovolume of each taxon is stored
in the database master table (MEASURE_LIST). For this list, geometric
47
approximations from the literature were evaluated and, if necessary,
improved. In addition, for some taxa, new formulae were derived as
described above. Table 3.2 shows a list of taxa found in the Biesbosch
reservoirs, which were used to develop the BACCHUS application.
One of the master tables also includes all calibration values of the used
microscopes and possible magnifications. These calibration values are
needed to convert the number of pixels into units of length. The
microscope type and magnification are selected before starting the
measurements. Depending on this choice, the image analysis programme
retrieves a corresponding calibration value from the database. Distortion
effects were tested by measuring a preserved specimen of the alga
Cryptomonas (Cryptophyceae) 25 times in the middle and in the top and
bottom left and right corner of the screen. Analysis of variance (Anova)
using the statistical software package Statistica for Windows (Statsoft,
1997) showed no significant difference (P = 0.28).
3.2.5. Measurements
The communication between the image analysis programme and the
database is established through dynamic data exchange (DDE) links. The
database is started automatically from the image analysis programme.
Before the biovolume measurements can be performed however, the
microscope set-up has to be checked to confirm that the conversion factor
between pixels and unit of length is still correct (Figure 3.11). It is
impossible to start measuring without successfully completing this check.
An officially certified object micrometer is used for this purpose. The
technician draws a straight line of predefined length using a digital image
of the object micrometer. The computer verifies whether the length of this
line corresponds with the stored value. In our application, a maximum
deviation of 3% of the predefined length is tolerated. It is impossible to
start measuring without passing the calibration check. After having
entered user and sample identification codes in the appropriate fields
in a table (T_CURRENT), biovolume measurements can be started.
48
Figure 3.11
Schematic presentation of the BACCHUS application.
49
50
:
:
**
*
cell
Actinocyclus
normanii
Aulacoseira
(1/4).π.d².l
cylinder
cylinder
cylinder
ellipsoïde
ellipsoïde
cylinder
(1/4).π.d².h
(1/4).π.w².h
(1/6).π.b².l
(1/6).π.d³
(1/4).π.d².h
(1/6).π.d³
(1/6).π.b².l
(1/12).π.b².l
(1/6).π.d³
sphere
sphere
ellipsoïde
kegel
(1/4).π.d².l
(1/6).π.d³
mathematical
formula*
cylinder
sphere
geometrical
shape
h = 0,82·d
h = 0,35·d
h = 1,00·d
factor
d=6,6
d=6,7
d=3,5
d=4,3
fixed
dimension
(µm)
d=5,8
l=200 µm (4
cylinder
(1/4).π.d².l
squares)
Melosira varians
l=200 µm (4
cylinder
(1/4).π.d².l
squares)
S. binderanus
l=200 µm (4
cylinder
(1/4).π.d².l
squares)
Stephanodiscus
l=200 µm (4
cylinder
(1/4).π.d².l
squares)
d: diameter; l: length; b: width; h: height
Nvol = number of volume measurements; Nf = number of measurements performed to determine a factor
cell
Centrales>25µm
(Centrales)
cell
cell
cell
Bacillariophyceae
Chrysococcus
Dinobryon
Kephyrion
Oscillatoria
Microcystis
ring (21
cells/ring)
l=200 µm (4
squares)
square (125
cells/square)
l=200 µm (4
squares)
cell
cell
cell
counting unit
Mallomonas
Synura
Centrales<=25µm
Chrysophyceae
Anabaena
Cyanophyceae
Aphanizomenon
taxon
**
67
74
47
45
63
43
Nf
221
156
2750
2863
median volume
per counting unit
(µm³)
2163
**
117
47
45
63
43
Nvol
Rott, 1981
Rott, 1981
Rott, 1981
Carpentier et al.,
1997
Rott, 1981
Rott, 1981
Carpentier et al.,
1997
Carpentier et al.,
1997
Carpentier et al.,
1997
Carpentier et al.,
1997
Wetzel & Likens,
1991
Rott, 1981
Reynolds, 1984
Rott, 1981
Reynolds, 1984
Rott, 1981
reference
Phytoplankton taxa with geometrical form, mathematical formula, factor, dimensions to be measured, static volume and number of measured
individuals
algal class
Table 3.2
cell
spindle
brick
brick
(1/12).π
.b².((b/2)+l)
(1/6).π.b².l
(2/15).π.b².l
(2/15).π.b².l
(2/15).π.b².l
(2/15).π.b².l
l.b.h
l.b.h
(1/6).π.w².l
(2/15).π.b².l
l.b.h
l.b.h
l.b.h
w = 0,53.l
w = 0,54.l
w=h
w=h
w = 0,54.l
w=h
w=
0,052.l
w=h
w=
0,060.l
w=h
factor
l=7,9
l=61,0
fixed
dimension
(µm)
l=55,8
**
218
218
106
19
Nf
46
612
median volume
per counting unit
(µm³)
625
π/48.l.(2,72.b²+1,65.b.(g1+g2)+2.(g1²+g2²))+π
/12.(g1².f1+g2².f2+g3².f3+g4².f4)***
mathematical
formula*
Gymnodinium
cell
ellipsoid
(1/6).π.b².l
w=0,48.l
13
Peridinium
cell
sphere
(1/6).π.d³
Euglenophyceae
Euglena
cell
spindle
(2/15).π.b².l
Phacus
cell
ellipsoid
(1/6).π.b.h.l
h =0,65.w
Trachelomonas
cell
ellipsoid
(1/6).π.b².l
*
:
d: diameter; l: length; b: width; h: height
**
:
Nf = number of measurements performed to determine a factor; Nvol = number of volume measurements;
***
:
See Figure 3.4 for description of ‘f’ and ‘g’.
Dinophyceae
cell
cone + ½
sphere
combination
of shapes
cell
Ceratium
ellipsoid
cell
cell
cell
cell
cell
cell
cell
Navicula
Nitzschia
Rhoicosphenia
Surirella
Tabellaria
Pennales nn.
Cryptomonas
<25µm
Cryptomonas
>25µm
Rhodomonas
Cryptophyceae
spindle
spindle
spindle
spindle
brick
brick
ellipsoid
cell
cell
cell
Cymatopleura
Diatoma
Fragilaria
brick
cell
Asterionella
formosa
geometrical
shape
Bacillariophyceae
(Pennales)
counting
unit
taxon
algal class
Tabel 3.2 (continued)
**
180
106
19
Nvol
1997
1997
1997
1997
1997
Rott, 1981
Rott, 1981
Carpentier et al., 1997
Rott, 1981
Rott, 1981
Michielsen & Ketelaars,
1994
Anonymus, 1992
Rott, 1981
Carpentier et al.,
Carpentier et al.,
Carpentier et al.,
Carpentier et al.,
Rott, 1981
Carpentier et al.,
Rott, 1981
Rott, 1981
Rott, 1981
reference
51
52
**
:
:
****
:
*
Schroederia
Sphaerocystis
Staurastrum
spindle
sphere
combination
of shapes
ellipsoid
sphere
ellipsoid
brick
sphere
sphere
sphere
spindle
ellipsoid
sphere
spindle
ellipsoid
ellipsoid
cylinder
sphere
ellipsoid
ellipsoid
sphere
ellipsoid
cylinder
spindle
spindle
sphere
sphere
sphere
spindle
sphere
geometrical
shape
bcel=1/N.bc
b=0,091.l
b=0,71.l
l=b=0,5h
b=0,18.l
factor
**
103
35
9
h=6,1
9
Nf
l=16,1
fixed
dimension
(µm)
14,7
median
volume per
counting unit
(µm³)
8
313
oen
(2/15).π.b².l
b = 0,17.l
l=22,0
35
127
(1/6).π.d³
2/3.h.(1/4.√3.(l²mi+l²bo)+√(3/16.l²mi.l²bo) + 1/2.π.dst².lst
(1/6).π.b².l
l.b.h
(1/6).π.d³
(1/6).π.d³
(1/6).π.d³
(2/15).π.b².l
(1/6).π.b².l
(1/6).π.d³
(2/15).π.b².l
(1/6).π.b².l
(1/6).π.b².l
(1/4).π.d².h
(1/6).π.d³
(1/6).π.b².l
(1/6).π.b².l
(1/6).π.d³
(1/6).π.b².l
(1/4).π.d².h
(2/15).π.b².l
(2/15).π.b².l
(1/6).π.d³
(1/6).π.d³
(1/6).π.d³
(1/6).π.b².l
(1/6).π.d³
mathematical
formula*
Tetraedron
cell
(1/6).π.b².l
Tetrastrum
coen. (4 cls)
(1/6).π.d³
d=5,3
d: diameter; l: length; w: width; h: height
Nvol = number of volume measurements; Nf = number of measurements performed to determine a factor
coen. = coenobium; cls = cells
cell
coen. (8 cls)
cell
Cosmarium
Crucigenia
Dictyosphaerium
Didymocystis
Eudorina
Kirchneriella
Lagerheimia
Micractinium
Monoraphidium
Oöcystis
Pandorina
Pediastrum
Planktosphaeria
Pteromonas
Quadrigula
Radiococcus
Desmodesmus
Actinastrum
Ankistrodesmus
Ankyra
Asterococcus
Carteria
Chlamydomonas
Closterium
Coelastrum
Chlorophyceae
counting unit
cell
cell
cell
cell
cell
cell
cell
coen. (16
cls)****
cell
coen. (4 cls)
coen. (8 cls)
cell
coen. (16 cls)
cell
cell
cell
cell
coen. (4 cls)
coen. (16 cls)
coen. (16 cls)
cell
cell
cell
coen. (4 cln)
coen. (4 cln)
taxon
algal class
Tabel 3.2 (continued)
**
8
35
103
Nvol
Michielsen & Ketelaars, 1994
Rott, 1981
Wetzel & Likens, 1991
Rott, 1981
Rott, 1981
Rott, 1981
Rott, 1981
Rott, 1981
dit rapport
Rott, 1981
Rott, 1981
Wetzel & Likens, 1991
Reynolds, 1984
Rott, 1981
Rott, 1981
Rott, 1981
reference
The next step is to choose a taxon from the list, which contains all taxa
available
in
the
database.
In
addition,
the
used
objective
with
corresponding magnification is selected. This information is also stored in
T_CURRENT along with the user and sample identification code.
According to the chosen magnification, the database finds a corresponding
calibration value which is used to convert pixels into a unit of length. At
this point, a timestamp (day, date, hour, minute, second) is generated in
the database and stored in T_CURRENT as well. The combination of the
timestamp with the user and sample identification code, taxonname and magnification creates a unique record for the database.
As soon as the microscopic image is digitised, a window automatically
appears with the first dimension to be measured (retrieved from
MEASURE_LIST). Measurements are made by drawing a straight line
using the digitiser. In addition, the image-analysis system makes it
possible to follow the curvature of curved organisms (e.g. Closterium and
Monoraphidium). The measured value (number of pixels) is converted to a
unit of length using the calibration value mentioned above. The converted
value is sent to the database and also stored in the record of
T_CURRENT as value1. A second, third, etc., dimension is measured and
stored in the same way, thus generating value2, value3, etc., to a
maximum of 10 values. After measuring all the necessary dimensions for
that taxon, the programme automatically calculates the biovolume of the
organism using the taxon-specific information from MEASURE_LIST. This
calculated biovolume is stored in T_CURRENT and displayed on the
screen. The biovolume is calculated as:
biovolume
mm3 / l  biovolume
µm3/counting unit   density
counting unit/ml   16


taxon i
taxon i 
taxon i
10
(3.15)
3
Biovolume results are reported in mm /l.
The user is able to discard this record in case something goes wrong
during the measurement. Otherwise, the database will contain incorrect
information, which will have to be deleted afterwards. If the value is
accepted, the collected data in T_CURRENT are automatically transferred
to another table (T_DATA) and stored as one record consisting of a
taxon-name, sample identification code, user identification code,
timestamp, value1 to value10 and the calculated biovolume and
53
standardised biovolume. At the end of the measurement, the calculated
biovolume is presented on screen. After two or more measurements, the
running geometric mean of the biovolume of this taxon is displayed as
well, and after 10 or more measurements the half-width of the 95%
confidence interval (CI) and criterion (see Section 3.2.6) are also
displayed. However, it is not necessary to measure all individuals of one
taxon before switching to another taxon, since the information is stored
per taxon and statistical results are calculated and displayed on screen
every time a measurement of that taxon is performed.
3.2.6. Statistical analysis of the results
The statistical analysis is performed in the database, where the table
T_DATA is used as a spreadsheet. Calculation of the biovolume is done
by means of a third-power function, containing the measured dimensions
of a taxon. With every measurement of a dimension, an error is made,
which is then raised to a power of 3. This causes the overall error in the
biovolume to increase with the size of the taxon, in other words: the
variance of the biovolume increases with an increasing mean (positive
correlation). In order to make a statistical evaluation possible, the data
has to be log-transformed. In practice, datasets regarding volume or
weight of organisms generally have to be log-transformed to achieve a
normal distribution (Van Heusden, 1972; Bird & Prairie, 1985; LaBaugh,
1995; Sokal & Rohlf, 1995). However, non-transformed results are usually
published, to allow for comparison with previously published data. (TheilNielsen & Søndergaard, 1998).
The biovolumes of small Cryptomonas from De Gijster Reservoir were
used to determine if a log-transformation of the data was necessary. We
assessed the results of three different weeks (week 7, 13 en 43) in 1997,
for which at least 20 observations were available. Both the logtransformed and the non-transformed data of each week were analysed to
see whether they showed a normal distribution (Figure 3.12). Table 3.3
shows the results of the Lilliefors and Shapiro-Wilk normality tests. A
significant outcome of these tests indicates a deviation from a normal
distribution. From Figure 3.12 and Table 3.3 it was concluded that the logtransformed results show a better normal distribution than the nontransformed data. Therefore, all data were log-transformed for evaluation.
54
non-transformed data
Figuur 3.12
ln-transformed data
Expected normal distribution (black line) of the non-transformed
(left column) and log-transformed results (right column) of small
Cryptomonas in week 7, 13 and 43 of De Gijster Reservoir
(1997).
55
Tabel 3.3
Significance of the outcomes of the Lilliefors and Shapiro-Wilk
normality tests of log-transformed and non-transformed results of
small Cryptomonas
P (Shapiro-Wilk test)
P (Lilliefors test)
not transformed
transformed
not transformed
transformed
week 7
< 0,001
< 0,05
< 0,01
< 0,01
week 13
< 0,01
n.s.
< 0,01
n.s.
week 43
< 0,001
n.s.
n.s.
n.s.
All statistical analyses were performed with ln-transformed results. The
mean, standard error (SE), and upper and lower limit of the 95%
confidence
interval
(CI;
UL95%
and
LL95%)
of
the
ln-transformed
biovolumes were calculated for every taxon, using formulae from Sokal
and Rohlf (1995):
Standard deviation (s):
n
(Y  Y)
2
i
s
i 1
n1
(3.16)
where:
Yi: ith individual result,
Y : mean of i results,
n: total number of results.
Standard error (SE):
SE 
s
n
(3.17)
Upper limit of the 95% CI (UL95%):
(3.18)
where:
t: tα = 0.05; d.f. = n - 1.
Lower limit of the 95% CI (LL95%):
(3.19)
56
In order not to spend too much time on the biovolume analysis of each
sample, it was decided to measure a maximum of 30 counting units of
frequently found taxa in a sample. The value of t in Formula 3.18 and
3.19 varies between 2.26 (d.f. = 9) and 2.05 (d.f. =29) between 10 and
30 measurements. For practical reasons, we used the value of 2.1, which
is the t-value of 20 measurements (d.f. = 19).
Figure 3.13
Running mean biovolume and number of measurements of
Cryptomonas. The 95% CI and allowed width of the interval to
reach the desired precision (criterion) are also indicated. •:
individual result; white line: running mean biovolume; grey area:
95% CI; black lines: criterion (0.2 ln units above/below the
mean). The vertical dotted line indicates the point where the
width of the 95% CI reaches the criterion.
The width of the 95% CI indicates the precision of the calculated mean.
The width of this interval is determined by the number of measurements
and the standard deviation. By performing more measurements, the
interval becomes smaller, which means that the mean becomes more
precise. The extent of desired precision can be defined by setting limits to
the maximum allowed width of the interval. In BACCHUS, the half-width of
the 95% CI and criterion are presented for the first time after 10
measurements. After performing 30 measurements without reaching the
57
criterion, measurements are usually stopped, thereby accepting a less
precise mean biovolume. For our laboratory, it was decided that the limits
of the CI must be <0.2 ln units above or below the mean. Figure 3.13
shows the change in mean biovolume and 95% CI when the results show
considerable variation. Measurements of small Cryptomonas were carried
out and their individual biovolumes varied between 900 and 4100 µm 3;
ln(biovolume) varied between 6.8 and 8.3. As shown in Figure 3.13, the
0.2 limit is reached after 22 measurements. Finally, the back-transformed
mean biovolume was calculated for every taxon (i.e. the geometric mean).
3.2.7. Data retrieval
The BACCHUS application contains a standard report which can be used to
generate reports on all measured taxa, the number of measured
individuals
per
taxon,
geometric
mean,
minimum
and
maximum
biovolume per sample. This report can be printed after each session. For
further data processing, all kinds of combinations of taxa, sampling
locations and dates can be retrieved using the Data evaluation option.
The straightforwardness of the database structure makes it easy to
perform queries for specific evaluations. For example, a simple query was
created to examine the length-width ratio of the genus Mallomonas
(Chrysophyceae). It appeared that these individuals could be separated
into two groups, each with a more or less specific length-width ratio
(Figure 3.14). Further microscopic examination showed that Mallomonas
occurred in ellipsoid and spindle-shaped types, indicating at least two
different species.
3.3.
Evaluation and assessment of results
After performing biovolume measurements with the BACCHUS application
for two years, an evaluation of the data was made. The objective of this
evaluation was to try and reduce the required effort involved in the
biovolume analysis and thus reduce the analysis costs. There are two
ways to achieve this:
1. by determining a fixed biovolume for a certain taxon at a certain
location, so that certain taxa no longer need to be measured;
2. by determining fixed ratios between dimensions, thereby reducing the
number of measurements to be performed.
58
Figure 3.14
Scatter plot of length and width of Mallomonas (Chrysophyceae),
showing clusters of ellipsoid and spindle-shaped individuals.
•: individual result (n = 88).
3.3.1. Evaluation method
The data collected for the Biesbosch reservoirs (the Netherlands) in 1996
and 1997 were used for the evaluation. These values were recorded in
T_DATA of the BACCHUS application. Before the actual evaluation, the
collected data were checked for obvious errors and corrected when
possible. All records for which one or more of the required dimensions
were not measured (resulting in a biovolume of 0), were deleted. Despite
the built-in checks for these cases, this occurred in 0.1% of all biovolume
results. In total, 25,898 biovolume results were evaluated.
Figure 3.15 shows a schematic view of the procedure for evaluation of
biovolume results determined with BACCHUS. An evaluation can only be
performed if sufficient data are available (N>30). Very small taxa (<1000
µm3) for which sufficient data were available were given a fixed biovolume
if the precision of the results expressed as the width of the 95%
confidence interval was <0.4 ln units. When geographical differences were
observed, a fixed biovolume per sampling location was determined.
59
Larger taxa were only given a fixed biovolume if the 95% confidence
interval was <0.2 ln units. If a specific seasonal size pattern was
observed, it was decided to continue measuring the taxon, despite
meeting the criteria for a fixed biovolume. If no fixed biovolume could be
determined, possible relationships between dimensions were investigated.
Figure 3.15
Schematic view of the evaluation procedure of biovolume results
determined with BACCHUS. 95%-CI: 95%-confidence interval;
SL: sampling location; N: number of measurements.
3.3.2. Fixed biovolumes
If many measurements throughout all seasons were performed, and
statistical analysis showed that the biovolume of this particular taxon did
not vary significantly, a fixed value could be used to calculate biovolumes
from counts, so that routine biovolume measurements of this taxon were
no longer necessary. However, it is essential to periodically check whether
this fixed value is still valid.
60
Figure 3.16 shows an example of geographical variation in the results of
biovolume measurements. In this case it was not possible to determine a
fixed
biovolume.
This
illustrates
that
literature
data
based
on
measurements from different locations are not suitable for general use.
Figure 3.16
Size frequency distribution of Melosira varians at 4 different
sampling points: the three Biesbosch reservoirs of Evides Water
Company (Res. 1: De Gijster; Res. 2: Honderd en Dertig; Res. 3:
Petrusplaat) and the river Meuse (Intake). These points are only a
few kilometres separated from each other and yet a clear
distinction is evident.
Figure 3.17 shows the results of plankton biovolume measurements in the
Biesbosch reservoirs from 1997 to 2000, including the taxa for which a
fixed biovolume was determined after evaluation. Table 3.4 shows some
cyanobacterial taxa for which a fixed biovolume was determined after the
evaluation of the biovolume measurements.
61
Tabel 3.4
Biovolume of some cyanobacterial taxa with a fixed biovolume on
the basis of the BACCHUS data evaluation
biovolume (µm3)
taxon
counting
N
geom. lower limit upper limit fixed value
mean
95%-CI
95%-CI
(µm3)
271
45.5
42.1
49.2
956
Aphanizomenon sp. l=200 µm
235
2847
2729
2970
2847
Microcystis sp.
653
65.4
62.8
68.1
8175
69
10.8
9.56
12.1
10.8
145
39.6
37.2
42.2
39.6
unit
Anabaena sp.
cell
cell
Pseudanabaena sp. cell
Snowella sp.
cell
The fixed biovolumes which were determined on the basis of this
evaluation were compared to biovolume values found in literature (Table
3.5), when necessary recalculated to match the same counting unit.
Generally, the order of magnitude is similar, although differences were
observed.
62
Figure 3.17
Biovolume (geometric mean and values between 10- and 90percentile) of the counting unit per taxon of phytoplankton
measured between 1997 and 2000 in the Biesbosch reservoirs,
river intake and end-product (semi-purified water for production
of drinking water). The number of counting units measured is
stated in parentheses for every taxon. ● geometric mean;
fixed biovolume since January 1st 1998;
after evaluation;
:
: fixed biovolume
: continued measuring after evaluation;
vertical dotted line: limit of 1000 µm3.
63
Figure 3.17
64
Continued
Table 3.5
Comparison of fixed biovolumes as determined with BACCHUS
and literature values found.
taxon
fixed
literature value
biovolume
(µm3)
reference
(µm3)
Aphanizomenon
3607
2316-4300
Reynolds (1984)
3920
Ahlgren (1970)
Microcystis
8102
8000
Ahlgren (1970)
Dinobryon
364
250-800
Nauwerck (1963)
600-850
Reynolds (1984)
Aulacoseira
10360
12000-16000
Jørgensen et al. (1991)
Melosira varians
20075;
40000
Nauwerck (1963)
3795
Michielsen & Ketelaars (1994)
2500
Nauwerck (1963)
1000
small cells (Ahlgren, 1970)
3700
medium cells (Ahlgren, 1970)
6800
large cells (Ahlgren, 1970)
2305
Michielsen & Ketelaars (1994)
49003*
Cryptomonas
1608
*: first value for reservoirs, second value for river Meuse.
3.3.3. Fixed ratios between dimensions
The assessment of fixed ratios between different dimensions of a taxon
was performed with a regression analysis. The ratio between the length
and width of Pteromonas was calculated on the basis of 131 observations
(Figure 3.18). The relationship was described as:
b = 0.802 (± 0.05) · l (R2 = 0.67)
(3.20)
3.3.4. Impact of regular evaluation
Using fixed biovolumes or fixed ratios between dimensions (based on
analysis of sufficient data) can help reduce the time of biovolume
measurements and thus also the costs. After all taxa had been measured,
processing a sample took from 0.5 to 3 hours. After evaluation and
determination of fixed biovolumes or ratios of dimensions, sample
processing times were reduced to approximately 20 minutes.
65
Figure 3.18
Length-width ratio for Pteromonas. •: individual result (n = 131);
dotted line: relationship as described in Equation 3.20.
3.4.
Discussion and conclusions
3.4.1. The BACCHUS application
Evident spatial and temporal variations exist in phyto- and zooplankton
biovolumes (Ruttner, 1952; Nauwerck, 1963; Ruttner-Kolisko, 1977;
Yilicic,
1985).
measurements
This
makes
unsuitable
the
for
use
of
accurate
literature
biovolume
data
or
single
determinations
(Nauwerck, 1963). With semi-automated measuring systems (e.g. Sprules
et al., 1981; Allen et al., 1994), biovolumes can be measured much more
rapidly on a routine basis.
Compared to semi-automated systems based on the concept of Sprules et
al. (1981), the advantage of using a digitiser is the ability to measure
curved organisms (Roff and Hopcroft, 1986). In contrast to the system
used by Roff and Hopcroft, which measures curved organisms by summing
up short linear measurements traced along the curvature, a real curved
line can be drawn on screen by the image-analysis system used here.
Another advantage of this system, compared to that of Roff and Hopcroft
66
(1986), is that online statistics are used, which minimise the number of
measurements without losing precision. As far as we know, this is the only
(semi-)automated
measuring
system
which
uses
online
statistics.
However, biological measuring systems may evolve rapidly and are
updated on a regular basis, without the updates being published (Verity
and Sieracki, 1993; N.D. Yan, personal communication, 1997). Other
advantages of the BACCHUS application are a very flexible input of master
data, output of stored data and a user-friendly operating environment.
BACCHUS is very easy to customise because it does not require any
expertise in database programming.
Compared to previously published semi-automated systems, which are
inexpensive, the BACCHUS application only runs on a relatively expensive
image analysis system. It should, however, be realised that the imageanalysis software has many more features than those used by the
BACCHUS application and thus can be used for many different purposes.
For example, the image storage and retrieval application was used to build
a reference collection of phyto- and zooplankton, and made possible many
small applications for specific purposes such as fully automated lengthfrequency distributions and macro-invertebrate biomass determinations.
Other applications are, for example, an automated colony count technique
for the detection and separation of confluent microbial colonies and
colonies of various sizes on Petri dishes (Corkidi et al., 1998), automated
detection of cyanobacteria (Thiel and Wiltshire, 1995) and morphometric
analysis of electron micrographs of microalgal cells (Fisher et al., 1998).
In addition, enumeration of the ecologically and toxicologically important
picoplankton (Wilde and Cody, 1998) can be performed on a routine basis.
Roff and Hopcroft (1986) compared their digitised measuring system with
a conventional microscope eyepiece micrometer and found that the
precision of measurement was much lower using an eyepiece micrometer.
From our own measurements, it also became clear that ocular micrometer
measurements are very time consuming. Measurements of Actinocyclus
normanii (Bacillariophyceae) with BACCHUS were performed in about half
the time necessary to perform the measurements with the ocular
micrometer. Biovolume calculations (including statistical parameters) with
BACCHUS require virtually no extra time at all. When measuring
Desmodesmus (Chlorophyceae), the measurements with the ocular
67
micrometer required even more time since the organism had to be reoriented to be able to measure two dimensions perpendicular to each
other. The orientation and position of organisms are unimportant when
using BACCHUS. Ocular micrometer measurements quickly lead to user
fatigue. Therefore, the more measurements are performed, the less
precise they become. When using BACCHUS, this loss of precision, caused
e.g. by fatigue of the user, does not occur. The practice in our laboratory
with many technicians has shown that plankton counts can very easily and
comfortably be combined with biovolume measurements. We introduced
online statistical evaluation of the data to minimise the number of
measurements necessary to reach a mean biovolume which is sufficiently
precise. It has to be noted that in routine samples not enough individuals
of a taxon can be present in the sample to reach the desired precision.
Even if no image-analysis system is used to measure biovolumes, the
statistical analysis can be used to reduce analysing time.
A disadvantage of image analysis is that only two dimensions can be
measured. Laser and digital confocal microscopy can acquire images in
three dimensions, but are not suitable for routine measurements of
plankton biovolumes, because they are either expensive or require a
highly skilled technician (Verity et al., 1996). In addition, analyses are
very time-consuming. These microscopes can, however, be applied to
occasionally measure the third dimension of certain plankton taxa in order
to determine and check the factors used in BACCHUS to calculate the third
dimension.
In
conclusion,
the
BACCHUS
application
allows
for
quick
routine
measurements of phytoplankton biovolumes. By using online evaluation of
collected data, only a minimum number of measurements are performed
to reach a suitably precise mean biovolume.
3.4.2. Recent developments
Since the publication of this article on BACCHUS in 1999, relatively little
has been published on this subject. Hillebrand et al. (1999) and Sun & Liu
(2002) published articles with lists of species and geormetrical shapes, in
order to start the harmonisation of the combinations of species and
68
shapes. All recent literature (of the past decade) regarding the shapes of
organisms has referred to these articles.
In 2007, Vadrucci et al. (2007) concluded that 3D imaging had become a
promising technique, but still needed improvements before it could be
used to measure all three dimensions of plankton taxa adequately for
biovolume determinations.
One new technique has been developed in the past decade however:
automatic dimension detection to estimate cell volume from twodimensional plankton images (Moberg & Sosik, 2012). This technique was
developed specifically for large datasets, linking flow cytometric data with
biovolume determinations. However, the authors concluded that most
routine laboratories do not use flow cytometry but microscopy for
phytoplankton analysis, so the use of the newly developed technique is
still restricted mostly to research activities. An important conclusion of
Moberg and Sosik (2012) was that the assumption of radial symmetry of
organisms along the appropriate axes remains the best approach to
determining the third dimension from a two-dimensional image. This
approach is also followed in the BACCHUS application.
Rott et al. (2007) questioned the statistical basis of the Utermöhl counting
method (Guidance standard EN 15204). In this article, the authors
elaborately discussed the statistics of the counting procedure, and briefly
touched on the determination of biovolume on the basis of the counting
results. The authors assumed that dimensions of algae are generally
measured using an ocular micrometer. However, regarding the statistics
of the biovolume analysis, the authors indicated that shapes could be
approximated
by
‘a
few’
relationships.
Nowhere
measurements
was
it
indicated
using
constant
however
just
allometric
how
many
measurements should be performed to achieve a suitably precise
biovolume
estimate.
This
is
remarkable,
considering
the
authors’
suggestion to require a minimum amount of effort per routine sample to
reach a particular precision level for the counting results. Notwithstanding
the importance of precision levels for counting results, the impact of
errors in dimension measurements for biovolume estimates is more
dramatic, as these errors are raised to a power of 3.
69
In summary, no new routine methods for biovolume determinations have
become available since the development of the BACCHUS application,
according to a search of literature covering the past decade. A major
advantage of the BACCHUS application is the simultaneous statistical
analysis allowing for the achievement of a certain precision level, in line
with the suggestions made by Rott et al. (2007) for the Utermöhl counting
procedure.
70
Chapter 4. In vivo absorption
4.
European Guidance Standard EN 16161 is
incomparable to ISO 10260 for the determination of
4.1.
Introduction
The occurrence of algae and in particular the biomass distribution of algal
species and classes are important for assessments of primary productivity
or eutrophication. Thus, many marine and freshwaters are sampled
frequently to determine algal densities and species/class composition. The
determination of chlorophyll-a is often used to quantify algal biomass.
The ISO 10260 Guidance Standard (ISO, 1992) is a commonly used
spectrophotometric method to determine chorophyll-a concentrations after
extraction of the pigment using organic solvents (ethanol). This method is
rather complex; care is necessary in sample-taking, storage, transport,
filtration, extraction and determination of the pigment concentration in
order to achieve comparability of results. Due to these influential factors,
unreliable results can occur if adequate quality control measures are not
in place.
Another disadvantage occurs when profiles or long-term studies are
carried out: the number of samples and analyses increases quickly. For
this reason, alternative methods such as fluorescence measurements or
the absorption procedure have been developed. The in vivo absorption
technique has recently been adopted as the standard method in an EUwide guidance standard: EN 16161 “Water quality – Guidance on the use
of
in
vivo
absorption
techniques
for
estimation
of
chlorophyll-a
concentration in marine and freshwater samples” (CEN, 2012). This
method is supposed to be considerably simpler and quicker than the ISO
10260 chlorophyll-a extraction method.
Intracellular chlorophyll-a absorbs light at a wavelength of approximately
675 nm; other substances absorbing light at this wavelength can be
eliminated mathematically. According to the authors of the new guidance
standard,
all
corresponding
absorption
values
chlorophyll-a
can
be
concentrations,
allocated
since
directly
the
to
absorption
the
is
proportional to the chlorophyll-a concentration. The use of a submersible
71
spectrofluorometer
allows
for
a
rapid
in
situ
determination
of
concentrations of chlorophyll-a. This paper describes an investigation into
the applicability of the EN 16161 Guidance Standard for the determination
of chlorophyll-a in algal cultures and natural samples.
4.2.
Materials and method
All algal cultures were obtained from the Experimental Phycology and
Culture Collection of Algae at the University of Göttingen (EPSAG,
Germany). Cultures of two green algae (Chlorella vulgaris; strain number
SAG 211-11b, and Scenedesmus obliquus; strain number 276-3b), one
diatom (Cyclotella meneghiniana; strain number SAG 1020-1a), and three
cyanobacteria (Microcystis aeruginosa; strain number SAG 46.80, Nostoc
Sp.; strain number 25.82, and Synechococcus leopoliensis; strain number
SAG 1402-1) were used in the experiments. These are commonly
occurring species in freshwaters of the temperate zone, and therefore
deemed to be suitable to test the EN 16161 method. The algae were
cultured at 20 µE/cm²/s (Hansatech QRT1 Quantitherm Lightmeter).
Different growth media were used: a BG11 nutrient solution for
cyanobacteria (Allen, 1968), Kuhl’s medium (Kuhl & Lorenzen, 1963) for
green algae, and a modified WC medium (Guillard, 1975) for C.
meneghiniana.
Two
different
instruments
AlgaeLabAnalyser
(ALA)
were
cuvette
used
for
the
fluorometer
experiments:
(bbe
the
Moldaenke,
Schwentinental, Germany) and the USB 4000 spectrometer (Ocean Optics,
Ostfildern, Germany). The ALA uses different excitation wavelengths
(between 370nm and 610nm) in order to determine algal fluorescence at
approximately 680 nm and the chlorophyll-a concentration for a maximum
of five algal groups. Beutler et al. (2002) and Chapter 5 of this thesis
describe the principle of this instrument in detail. The USB 4000
spectrometer was used with a cuvette of 10 cm length. The spectral
resolution was 0.2 nm, measurement values were recorded within the 180
to 880 nm range. The 600-710 nm range was evaluated following
Guidance Standard EN 16161. This guidance standard requires a very high
accuracy in the selection of the wavelength of 675 nm. Before using the
USB 4000 spectrometer, the accuracy of instrument was tested by
performing an absorption measurement of a potassium permanganate
72
solution and a transmission measurement of a filter (Dr. Hugo Anders
Company, Nabburg Schwandorf, Germany).
Unialgal dilution series and mixtures of the above-mentioned algae were
analysed for their chlorophyll-a concentration using three different
methods:

fluorometrically, as described in Chapter 5 using the ALA;

following ISO 10260 (ISO, 1992), using the USB 4000 spectrometer;

via an absorption spectrum in accordance with the method described
in EN 16161, using the USB 4000 spectrometer.
The chlorophyll-a concentration of the sample series was between 90 and
180 µg/l. The analysis results were used to draw conclusions regarding
the applicability of the procedures as described in the EN 16161 Guidance
Standard.
The absorption spectra obtained from the spectrometer measurements
were exported to MS Excel for further processing. Each absorption
spectrum was decomposed using the sum of four Gauss curves (Equation
4.1), following the approach of Hoepffner & Sathyendranath (1991) for
the 610-700 nm range.
This resulted in the following function:
4
f ( ; i ; i ; ai )   ai  e

(  i ) 2
2 i 2
i 1
(4.1)
where:
λ
= current wavelength of curve to be fitted
λi
= peak wavelength of Gauss curve i in nm
σi
= variance of Gauss curve i in nm
ai
= amplitude of Gauss curve i
The different Gauss curves represented the absorbing substances in the
solutions (Figure 4.1): Gauss Curve 1 for humic substances, Gauss Curve
2 for phycocyanin (blue-green algae) and chlorophyll-c (Cyclotella
meneghiniana), and Gauss Curve 3 for chlorophyll-a. Gauss Curve 4, in
73
which green algae and Cyclotella meneghiniana could be found and which
had a peak wavelength of ca. 740 nm, was of unknown origin.
0,5
0,08
0,45
0,07
Absorption (m-1)
0,4
0,06
0,35
0,3
0,05
0,25
0,04
Gauss curve 1
Gauss curve 2
Gauss curve 3
Gauss curve 4
0,2
0,15
0,1
0,03
0,02
0,01
0,05
0
0
400
450
500
550
600
650
700
750
Wavelength (nm)
Figure 4.1
Four Gauss curves used to decompose the absorption signal
within the 600-710 nm range (grey area). Gauss curve 1: humic
substances; Gauss curve 2: phycocyanin (blue-green algae);
Gauss curve 3: chlorophyll-a; Gauss curve 4: a.o. green algae
and Cyclotella meneghiniana.
The fit curves were created by using the initial 4 Gauss curves as a
starting point for an iterative process during which the peak wavelengths
λi and amplitudes ai of the Gauss curves were varied in order to achieve
the closest fit to the measured curves. To this end, each spectrometric
measurement
value
(5
values
per
nm)
was
compared
to
the
corresponding value of the sum of the decomposed curves (fit curve). The
deviations between the norm and fit curves in the individual measurement
values were squared and added. Following the approach as indicated in
EN16161, a minimisation routine in Excel was used (Solver, according to
the sum of the smallest error square) in order to achieve the best possible
correspondence between the norm and fit curves. The fit procedure was
carried out for the 600 to 750 nm wavelength range. According to
74
EN16161, “fitting” within the wavelength range of 650 to 710nm is
sufficient. In order to take this into account accordingly, the error square
during fitting within this range was more strongly weighted (factor 20)
than the error square outside this range (Factor 1).
Bidigare et al. (1990) determined the half-width of peaks of chlorophyll-a
to be 22 nm; this value was used for the fits (Figure 4.2), following the
description in Annex B of the EN16161 Guidance Standard.
Figure 4.2
In vivo absorption spectrum of chlorophyll-a, with the halfwidth of
the 675 nm peak indicated as a grey dotted line (after: Bidigare
et al., 1990 and CEN, 2012).
The chlorophyll-a concentration as determined using the absorption
method in accordance with EN16161 corresponds to the absorption coefficient a3 (Equation 4.1). EN16161 stipulates that the corresponding λ3
be kept constant at 675 nm. However, it was hypothesised that an
absorption maximum slightly deviant from 675 nm gives better results.
Therefore, two series of tests were performed. In the first test, the
corresponding λ3 was kept constant at 675nm in accordance with
EN16161 (Table 4.1, ‘unmodifiable λ3’). In the second series of tests, λ3
was left freely modifiable (Table 4.1, ‘modifiable λ3’). The results of both
75
tests were compared and the linearity was tested by using a series of
dilutions for each algal culture. The conditions as shown in Table 4.1 were
applied.
Table 4.1
Secondary conditions to the fit of the absorption measurements
for chlorophyll-a.
variable
conditions
λ1
370-400nm
λ2
625-640nm
λ3
unmodifiable: 675nm
modifiable: 650-700nm
4.3.
λ4
730-750nm
a1,2,3,4
freely modifiable, ≥ 0
σ1,2,4
freely modifiable, ≥ 0
σ3
22 nm
Results
The results of the measurements to check the accuracy of the USB 4000
spectrophotometer are shown in Figures 4.3 and 4.4. Figure 4.3 shows the
typical absorption spectrum of a 0.1 mmol/L potassium permanganate
solution, as also determined by others (e.g. Francis et al., 2011; Giokas et
al.,
2009).
From
these
figures,
it
was
concluded
that
the
spectrophotometer used can measure down to exactly 1 nm in the
important measurement range. They also show that the USB 4000
produces little noise.
Figure 4.5 shows the measured and reconstructed absorption curve
between 600 and 710 nm with an unmodifiable λ3 of 675nm and an
absorption half-width of 22 nm for cultures of Chlorella vulgaris and
Microcystis
aeruginosa.
The
curves
for
Scenedesmus
vulgaris,
Synechococcus leopoliensis, Nostoc Sp. and Cyclotella meneghiniana are
not shown. The description of the individual Gauss curves for all algal
species investigated can be found in Table 4.2.
76
Figure 4.3
Absorption spectrum of a 0.1 mmol/L potassium permanganate
solution measured using the USB 4000 Spectrophotometer (after:
Giokas et al., 2009).
Figure 4.4
Transmission spectrum of the filters with the manufacturer’s
values “Peak wavelength 702.3 nm, 10.7 nm half-width”.
77
Generally, the measured and fitted curves in Figure 4.5 are in good
agreement. This indicates that the decomposition of the chlorophyll-a
absorption spectrum using four Gauss curves in the spectral range of 650710 nm is essentially correct.
A.
B.
Figure 4.5
Measured (solid green line) and reconstructed (solid red line)
absorption curve of Chlorella vulgaris (A.), Microcystis aeruginosa
(B.) with an unmodifiable λ3 of 675nm. Dotted lines: Blue: Gauss
Curve 1; Red: Gauss Curve 2; Green: Gauss Curve 3; Black:
Gauss Curve 4 (see Fig. 4.1 and Hoepffner & Sathyendranath,
1991).
78
Chlorella
vulgaris
0,000
0,150
0,147
Amplitude a3 of Gauss Curve 3 at 675nm.
Nostoc
sp.
0,038
0,127
0,156
0.890
0.045
0.038
2.67E-04
0.279
418.6
23.0
22.0
153.0
430.0
635.7
675.0
730.0
1.221
Nostoc sp.
Cyclotella
meneghiniana
0,016
0,102
0,083
Cyclotella
meneghiniana
0.867
0.038
0.016
1.43E-04
0.241
641.2
36.3
22.0
59.6
413.5
625.0
675.0
733.2
0.246
Synechococcus
leopoliensis
0,034
0,062
0,082
Synechococcus
leopoliensis
0.413
0.038
0.034
4.90E-04
0.088
547.6
23.1
22.0
167.6
430.0
632.9
675.0
730.0
0.873
Microcystis
aeruginosa
0,057
0,112
0,135
Microcystis
aeruginosa
0.910
0.071
0.057
4.50E-04
0.255
409.9
23.5
22.0
148.1
430.0
635.7
675.0
730.0
2.173
Scenedesmus
vulgaris
0.242
0.040
0.000
0.00E+00
0.201
467.3
32.5
22.0
212.2
429.8
625.0
675.0
748.9
0.157
Amplitude, half-width and peak wavelength of the four Gauss
curves for an unmodifiable λ3 of 675 nm and a half-width of 22
Scenedesmus
vulgaris
0,000
0,128
0,150
nm.
amplitude Gauss Curve 3
value according to ISO 10260
value according to Fluorometric method
Table 4.3
Chlorella
vulgaris
0.424
0.036
0.000
0.00E+00
0.189
550.6
28.7
22.0
80.4
427.0
640.0
675.0
730.0
0.250
Amplitude, half-width and peak wavelength of the four Gauss curves for an unmodifiable λ3 of 675 nm and a halfwidth of 22 nm.
Table 4.2
amplitude Gauss Curve 1 (a1)
calculated absorption coefficient
half width curve 1 (nm)
peak wavelength curve 1 (λ1; nm)
χ2
Table 4.2
79
The amplitude of Gauss Curve 3 is supposed to be the same for all algal
species (Bidigare et al., 1990) and to correspond to 1 µg/L, i.e. a3= 8.8 x
10-4 m-1. For the green algae Chlorella vulgaris and Scenedesmus vulgaris,
the result of the EN 16161 method deviates strongly from the value
determined by the ISO 10260 and fluorometric methods, since Gauss
Curve 3 with the peak wavelength of 675 nm does not contribute to the fit
curve (Fig. 4.5A and Table 4.3).
Table 4.3
Amplitude a3 of Gauss Curve 3 at 675nm, compared to the values
according to ISO 10260 and the fluorometric method.
Chlorella
Cyclotella
Nostoc Synechococcus Microcystis Scenedesmus
vulgaris meneghiniana
sp.
leopoliensis
aeruginosa
vulgaris
a3
0,000
0,016
0,038
0,034
0,057
0,000
ISO 10260
0,150
0,102
0,127
0,062
0,112
0,128
fluorometry
0,147
0,083
0,156
0,082
0,135
0,150
Table 4.3 shows the measurement results of three compared methods.
The chlorophyll-a concentrations of the values determined fluorometrically
and by ISO 10260 were converted to absorption signals using a specific
absorption co-efficient of chlorophyll-a with a value of 88.15 L/g.cm
(Bidigare et al., 1990). Figures 4.6A, B and C present a graphic
representation of the (lack of) correlation between the methods shown in
the table. They show that the proposed EN 16161 method does not
correlate with the ISO method, especially as a result of the errors
produced for the green algae. The ISO 10260 and the fluorometric
ISO 10260
methods show a better correlation.
A.
80
0,16
0,14
0,12
0,10
0,08
0,06
0,04
0,02
0,00
0,00
0,01
0,02
0,03
EN 16161
0,04
0,05
0,06
Fluorometric method
0,18
0,16
0,14
0,12
0,10
0,08
0,06
0,04
0,02
0,00
0,00
0,01
0,02
0,03
0,04
0,05
0,06
EN 16161
ISO 10260
B.
0,16
0,14
0,12
0,10
0,08
0,06
0,04
0,02
0,00
0,00
C.
Figure 4.6
0,05
0,10
0,15
0,20
Fluorometric method
Correlation between EN 16161 and ISO 10260 (A), between EN
16161 and the fluorometric method (B), and between the
fluorometric method and ISO 10260 (C) with an unmodifiable λ3
of 675nm. : green algae; : cyanobacteria; ●: diatoms;
solid line: trend line y = ax.
Table 4.4 shows that the dilution series were linear within the measured
range for all three methods under investigation.
In the second series of experiments, a slight variation in the peak
wavelength around 675 nm (λ3) was allowed in an attempt to improve the
results. Figure 4.7 shows the measured and reconstructed absorption
curves within the 600-710 nm range with a modifiable λ3 between 650
and 700 nm for cultures of Chlorella vulgaris and Microcystis aeruginosa.
The curves for Scenedesmus vulgaris, Synechococcus leopoliensis, Nostoc
81
Sp. and Cyclotella meneghiniana are not shown. The description of the
individual Gauss curves for all algal species investigated is presented in
Table 4.5. Peak wavelengths between 683 nm and 697 nm were set
during the fitting procedure using a random
wavelength 3. The
appropriate wavelength varied from species to species.
Table 4.4
Results of a dilution series of Chlorella vulgaris, analysed in
accordance with ISO 10260 1992 (E), fluorometrically (ALA) and
spectrophotometrically (USB 4000).
dilution factor
method
chlorophyll-a concentration
1
2
4
8
16
R2
235
105
62
28
18
0.995
116
53
27
13
6.5
0.997
247
130
70
32
14
0.997
according to ISO 10260 (µg/L)
converted absorption according to
EN 16161
fluorometrically determined
chlorophyll-a (µg/L)
The calculated absorption co-efficient is the value which is produced when
a3 is divided by the chlorophyll-a content determined by the ISO 10260
method, i.e. corresponding to 1 µg/L chlorophyll-a. According to Bidigare
et al., it should be at 8.8 x 10-4, independent of the algal species used.
The variance (σ) was converted to a half-width by multiplication using a
factor of 2.3.
82
A.
B.
Figure 4.7
Measured (solid green line) and reconstructed (solid red line)
absorption curve of Chlorella vulgaris (A.) and Microcystis
aeruginosa (B.) with a modifiable λ3 between 650 and 700 nm.
Dotted lines: Blue: Gauss Curve 1; Red: Gauss Curve 2; Green:
Gauss Curve 3; Black: Gauss Curve 4 (see Fig. 4.1 and Hoepffner
& Sathyendranath, 1991).
The fit curves in Figure 4.7 correspond better to the measured values than
the fit curves in Figure 4.3. The χ2 values indicating the goodness of fit
shown in Table 4.2 are higher by a factor of 8-86 than when 3 was
allowed to vary (Table 4.5). Lower values for χ2 indicate a better fit.
83
84
Chlorella
vulgaris
0.327
0.045
0.037
2.58E-04
0.130
854.8
66.8
22.0
70.9
430.0
626.0
698.0
734.1
0.029
Cyclotella
meneghiniana
1.287
0.019
0.026
2.23E-04
0.579
382.0
37.3
22.0
256.9
373.7
625.0
684.5
750.0
0.024
0.640
0.023
0.056
3.90E-04
0.122
654.7
28.0
22.0
272.3
430.0
631.4
685.2
730.0
0.028
Nostoc sp.
Synechococcus
leopoliensis
0.418
0.020
0.048
6.87E-04
0.029
675.0
22.2
22.0
306.0
411.0
635.1
683.6
730.0
0.015
Microcystis
aeruginosa
0.682
0.043
0.076
6.04E-04
0.066
676.3
27.1
22.0
314.3
416.9
633.8
684.1
730.0
0.025
Scenedesmus
vulgaris
0.356
0.013
0.015
1.01E-04
0.138
509.8
33.6
22.0
131.1
430.0
640.0
694.2
747.1
0.006
Amplitude, half-width and peak wavelength of the four Gauss curves with a modifiable λ3 between 650 and 700 nm
and a half-width of 22 nm.
Table 4.5
calculated absorption coefficient
χ2
Table 4.5
Amplitude, half-width and peak wavelength of the four Gauss
curves with a modifiable λ3 between 650 and 700 nm and a half-
width of 22 nm.
Table 4.6 shows the measurement results of the three methods compared.
The chlorophyll-a concentrations determined using the ISO 10260 and
fluorometric methods were converted to absorption signals using a specific
absorption co-efficient of chlorophyll-a with a value of 88.15 L/g.cm
(Bidigare et al. 1990). Allowing 3 to vary resulted in a contribution of
Gauss Curve 3 to the fit curve for the green algae Chlorella vulgaris and
Scenedesmus vulgaris.
Table 4.6
Amplitude a3 of Gauss Curve 3, with 3 varying between 683 and
698 nm (see Table 4.5 for peak wavelengths per species),
compared to the values according to ISO 10260 and the
fluorometric method.
Chlorella
Cyclotella
vulgaris meneghiniana
Nostoc Synechococcus Microcystis Scenedesmus
sp.
leopoliensis
aeruginosa
vulgaris
a3
0.037
0.026
0.056
0.048
0.076
0.015
ISO 10260
0.150
0.102
0.127
0.062
0.112
0.128
Fluorometry
0.147
0.083
0.156
0.082
0.135
0.150
Figures 4.8A and B show that even when 3 was allowed to vary the
correlation between the ISO 10260 and the fluorometric methods on the
one hand and the EN 16161 method on the other did not become any
better. The absorption co-efficient does not seem to be proportional to the
actual chlorophyll-a content.
ISO 10260
0,18
0,15
0,12
0,09
0,06
0,03
0
0,00
0,02
0,04
0,06
0,08
EN 16161
A.
85
Fluorometric method
0,2
0,16
0,12
0,08
0,04
0
0,00
0,02
0,04
0,06
0,08
EN 16161
B.
Figure 4.8
Correlation of EN 16161 (with a modifiable λ3) with ISO 10260
(A), and with the fluorometric method (B). : green algae; :
cyanobacteria; ●: diatoms; solid line: trend line y = ax. For the
comparison of the ISO 10260 and fluorescence method, see
Figure 4.6C.
4.4.
Discussion
The investigation carried out has raised considerable doubt as to the
general applicability of the proposed method. With a selection of six
common algal species from three different groups, no correlation between
the chlorophyll-a concentration as determined by the ISO 10260 (based
on the extraction of chlorophyll-a using ethanol) or fluorometric methods
and the EN 16161 method (based on the in vivo absorption of chlorophylla) could be produced. This is consistent with findings of Prieur &
Sathyendranath (1981), who concluded that the relationship between
chlorophyll-a concentrations and absorption coefficient depends on the
nature of the waters considered. They suggested that the change in
absorption per unit of chlorophyll-a is gradual rather than abrupt.
Cleveland (1995) found a logarithmic instead of a linear relationship
between fluorometrically
determined
chlorophyll-a and the
in vivo
absorption coefficient at 675 nm for the temperate zone, whereas Bricaud
et al. (1995) were unable to establish any relationship between the in vivo
absorption
coefficient
and
concentrations
of
chlorophyll-a.
Factors
affecting the relationship between concentrations of chlorophyll-a and the
86
absorption coefficient at 675nm could be pigment composition and/or the
package effect (Bricaud et al., 1995; Barlow et al., 2002), cell size (Ciotti
et al., 2002; Fujiki & Taguchi, 2002) and irradiance (Fujiki & Taguchi,
2002).
The EN 16161 Guidance Standard presents two datasets in Annex D. The
first one shows a linear relationship between in vivo photometric
absorption and laboratory absorption of extracts. The second dataset is a
comparison between the in vivo absorption method as described in EN
16161 and “lab chlorophyll-a”. It is unclear, however, whether the
comparison method used is the ISO 10260 method or another method,
since in both cases no indication of the exact method used could be found.
Generally, the results presented in this chapter confirm that the
decomposition of absorption spectra by means of Gauss curves as
previously shown by numerous others (e.g. Morrison, 2003; Hoepffner &
Sathyendranath, 1991) results in adequate fit curves, although a freely
modifiable 3 yields better results than a fixed value at 675 nm. The
linearity of the EN 16161 method within a concentration range of 10-25
µg/L was confirmed. Gower et al. (1999) showed linearity between
absorption peaks and chlorophyll-a concentrations within a 1-20 µg/L
range,
which
is
sufficient
for
expected
oceanic
chlorophyll-a
concentrations (Bricaud et al., 1995). However, since no relationship was
observed between the EN 16161 method and the ISO 10260 method,
further investigations are necessary to determine the general applicability
of the EN 16161 method.
87
Section 2
Phytoplankton and cyanobacterial biomass:
In vivo fluorescence of chlorophyll-a
Based on:
Izydorczyk, K., C. Carpentier, J. Mrowczyński, A. Wagenvoort, T.
Jurczak, and M. Tarczyńska. 2009. Establishment of an Alert Level
Algae Online Analyser for monitoring cyanobacterial chlorophyll-a.
Water Research 43: 989-996.
Kalaji, H.M., M. Skonieczny, O. Sytar, M. Brestic, K. Bosa, S. Pietkiewicz
and
C.J.
Carpentier.
Cyanobacterial
and
total
chlorophyll-a
concentrations as a method for early-warning water quality monitoring.
Submitted to CLEAN Soil, Air, Water. April 2013.
89
Chapter 5. Introduction to fluorescence
5.
Introduction to fluorescence measurements for the
assessment of algal or cyanobacterial chlorophyll-a
5.1.
Photosynthesis and fluorescence
Luminescence (a generic word for all types of light emission) was
discovered in 1565 by a Spanish medical doctor and botanist (Nicolas
Monardes) studying a plant extract to cure a kidney ailment (Govindjee,
1995). A major breakthrough followed in 1931, when H. Kautsky linked
the fluorescence of chlorophyll-a to photosynthesis (Govindjee, 1995).
Later, Kautsky and co-workers discovered that an interruption of the
photosynthetic pathway results in increased fluorescence (Kautsky et al.,
1960 in Maxwell & Johnson, 2000). Since then, many scientists have
found ways to use chlorophyll-a fluorescence to learn more about the
photosynthesis process and to use it to quantify primary productivity or
plant/algal biomass, as described in detail by e.g. Govindjee (1995, 2004)
and Krause & Weis (1991).
The following (nett) reaction is a simplification of the metabolic pathway
of photosynthesis (Campbell, 1993):
This process is highly efficient: around 98% of the energy taken up is
converted to chemical energy in the form of carbohydrates. In the
equation, the carbohydrate glucose is shown to simplify the reaction,
although
the
main
products
of
photosynthesis
are
acutally
other
carbohydrates (Campbell, 1993). Some of the energy is lost as heat and
about 1% is released in the form of fluorescence (Figure 5.1).
Chlorophyll-a is the primary photosynthetic pigment of all oxygen-evolving
photosynthetic organisms and is present in all plants, algae and
cyanobacteria.
Light
energy
absorbed
by
other
pigments
(e.g.
carotenoids, biliproteins and chlorophyll-b) is transferred to chlorophyll-a,
resulting
in
the
excitation
of
these
chlorophyll-a
molecules.
The
combination of chlorophyll-a and other light-capturing pigments is called
the peripheral antenna (Figure 5.1). Absorption of a photon by the
91
Section 2: Cyanobacterial biomass
chlorophyll-a molecule at the reaction centre of the peripheral antenna
causes the chlorophyll-a molecule to change from its ground state to its
excited state. The light energy is used to boost an electron to an orbital
where it has more potential energy, from where it is released to the
primary acceptor. The complete antenna complex with its chlorophyll-a
reaction centre and the primary electron acceptor is called a photosystem,
and it is located in the thylakoid membranes in the chloroplasts of plants
and true algae (Campbell, 1993). Cyanobacteria are prokaryotic cells and
thus do not have chloroplasts. The phycobiliproteins are aggragated in
phycobilisomes and attached to the thylakoid membranes (Brock et al.,
1997). Unlike the pigments in true algae, the phycobilisomes diffuse
rapidly on the cyanobacterium’s membrane surface and do not form stable
complexes with the chlorophyll-a reaction centres (Sarcina et al., 2001).
Figure 5.1
Schematic view of the photosynthesis process: light is captured
by the peripheral antenna, transferred to chlorophyll-a and
converted to carbohydrates. Some energy is lost as heat and
about 1% is released as fluorescence (picture: courtesy of bbe
Moldaenke, Schwentinental, Germany).
92
Two photosystem types are known: I and II. The main difference between
these photosystems is the absorption peak of the chlorophyll molecules at
their reaction centres as a result of their association with different
proteins. Although both photsystems absorb light best in the red part of
the spectrum, Photosystem I has its absorption peak at 700 nm (therefore
the reaction centre is called P700), and Photosystem II at 680 nm
(therefore, P680; Campbell, 1993). In this chapter, the main focus will be
on Photosystem II since most fluorometers use the fluorescence of this
photosystem to calculate total chlorophyll-a content. A detailed description
of the structure and functioning of Photosystem II was written by
Shinkarev (2004).
When a pigment molecule returns from its unstable excited state to its
ground state, excess energy is released as heat and light (fluorescence).
In Photosystem II, the fluorescence has a wavelength of approximately
680 nm, which is the result of the amount of energy being released by the
electron falling back to its ground state (Campbell, 1993). Since
Photosystem II produces fluorescence at one particular wavelength, it can
be measured easily using a photomultiplier and an optical band pass filter,
as described in detail by Beutler et al. (2002).
5.2.
Algal fingerprints: method to distinguish algal groups on the
basis of their pigment composition
In vivo chlorophyll-a fluorescence measurements have been carried out
for a variety of purposes in both marine and freshwater environments for
many years. In situ fluorometry for phytoplankton is performed for the
purpose of biomass determination (Heaney, 1978), depth profiling
(Schimanski et al., 2006), photosynthetic process characterisation (Kolber
et al., 1998) or as an early-warning system for the detection of
cyanobacterial blooms (Gregor et al., 2007; Izydorczyk et al., 2005,
2009). For these purposes, a submersible probe is used to determine the
chlorophyll-a fluorescence of algae in the water column.
When light reaches a surface of any kind, it may be absorbed, reflected or
transmitted. Pigments absorb visible light; different pigments absorb light
at different wavelengths. Chlorophyll is green because it absorbs red and
blue light while reflecting and transmitting green light (Campbell, 1993).
93
The pigments present in the peripheral antenna determine which
wavelengths can be absorbed and used to fuel the photosynthetic process.
Figure 5.2 shows an overview of algal groups, the pigments present in the
peripheral antenna and the spectral range at which these pigments absorb
light best.
Figure 5.2
Composition
of
peripheral antenna
and
associated
spectral
absorption range for different algal groups (picture: courtesy of
bbe Moldaenke, Schwentinental, Germany).
Most commercially available fluorometers are Pulse-Amplitude Modulated
(PAM) fluorometers. Light pulses of different wavelengths (e.g. at 470,
525, 570, 590 and 610 nm) are emitted, and the fluorescence response of
the algal assemblage is recorded at approximately 685 nm. An optical
filter prevents excitation light from reaching the detector, which would
otherwise cause an offset. The fluorescence signal is proportional to the
pigment density and distribution of the algal groups. A spectral fingerprint
determines the fluorescence intensity for each algal class at every
wavelength for a given density (Figure 5.3). After calibrating the
instrument, the fluorescence signals can be deconvoluted to algal biomass
as the chlorophyll-a concentration with the help of these spectral
fingerprints. An additional measurement at 370 nm is used to correct for
any effects of the presence of fluorescent coloured dissolved organic
material (CDOM).
94
Green Algae
Cyanobacteria
Diatoms
Cryptophyta
CDOM
2
Fluorescence [r.u.]
1,8
1,6
1,4
1,2
1
0,8
0,6
0,4
0,2
0
370
420
470
520
570
Wavelength [nm]
Figure 5.3
Spectral fingerprints of different algal classes (Beutler et al.,
2001). CDOM: Coloured Dissolved Organic Matter.
5.3.
Comparison of fluorescence measurements with other
methods for the determination of algal biomass
5.3.1. Fluorescence measurements versus spectrometric analysis
of chlorophyll-a
There is a distinct difference between spectrometric chlorophyll-a analysis
(e.g. ISO 10260 (ISO, 1992)) and in vivo fluorescence measurements of
chlorophyll-a. The spectrometric method is based on the analysis of total
chlorophyll-a,
chlorophyll-a
molecules
extracted
using
concentration
produce
in
warm
µg/L.
fluorescence
ethanol,
Since
light,
an
resulting
only
in
in
a
total
active
chlorophyll-
vivo
fluorescence
measurement will determine not the total but the active chlorophyll-a
concentration in µg/L. At low concentrations, the results of both methods
will be similar, however, at higher concentrations, a (self-)shading effect
will cause the fluorescence method to yield lower results. If chlorophyll-a
molecules are present in the sample but the excitation light cannot reach
them, they will not produce a fluorescence response. Figure 5.4 illustrates
the relation between total chlorophyll-a concentrations determined using
spectrometry and active chlorophyll-a concentrations measured with a
fluorometer (AlgaeOnlineAnalyser (AOA), bbe Moldaenke, Schwentinental,
95
Germany).
The
difference
between
total
chlorophyll-a
and
active
chlorophyll-a increases with the concentration, thus the measured values
deviate more from the line y = x (Figure 5.4). Similar results were found
by Gregor & Maršálek (2004).
Total chlorophyll-a (µg.L-1), AOA
50
40
30
20
10
0
0
20
40
60
80
Total chlorophyll-a (µg.L-1), spectrophotometric method
Figure 5.4
Relationship
between
chlorophyll-a
concentrations
measured
using the spectrometric and fluorescence (AOA) methods (n=51).
The various shapes of the dots indicate that samples were taken
at different sampling locations in Sulejów Reservoir (Poland). The
dotted line indicates y = x; the solid line indicates the trend line y
= 1.0656 x0.8116 (R2= 0.889) (After: Mrówczyński, 2004).
5.3.2. Fluorescence measurements versus biovolume analysis
Although both biovolume (see Chapters 1 and 2) and chlorophyll-a are
indicators of algal biomass, these parameters are not readily comparable.
Felip
and
Catalan
(2000)
showed
that
the
relationship
between
chlorophyll-a and biovolume of phytoplankton varied with the season and
the species composition. Additionally, they showed that a chlorophyll-a
maximum does not necessarily coincide with a biovolume maximum in a
lake. A biovolume maximum appeared to be related to an improvement in
growth conditions (mainly nutrient levels), whereas a chlorophyll-a
96
maximum was related to changes in cell pigment content as a result of
spatial or successional trends in species dominance.
Van der Oost (2010) compared cyanobacterial biomass as chlorophyll-a
(in vivo fluorescence) with both cell counts and biovolume analyses, and
found very similar relations. However, the dominant cyanobacterial
species in the samples compared were Microcystis and Anabaena, which
have similar biovolumes (around 50 µm3). This explains why both the cell
counts and the biovolume analyses showed a similar relation to the
chlorophyll-a analyses. In this case, a good relation was found for both
the cell counts and the biovolume analyses as compared to the
chlorophyll-a concentrations (R2 = 0.81 and 0.78 respectively). Van der
Oost (2010) showed that, generally, biovolume shows a better relation to
chlorophyll-a than cell counts. The reasons for this are explained in
Chapter 2. Section 5.4.2 shows an example of the comparison of
fluorometrically
determined
chlorophyll-a
and
biovolume
for
cyanobacteria.
5.4.
Control samples and sample storage
Control samples are essential for quality assurance and quality control
procedures regarding analytical methods. However, control samples for (in
vivo) fluorescence analysis are difficult to prepare and store for a long
time, as chlorophyll-a is unstable outside a living cell. In some
laboratories, suspensions of Chlorella powder are used as a control
sample. Chlorella powder can be obtained easily as it is commercially sold
as a food additive for e.g. horses. Another commonly used method is to
freeze and store a culture or natural sample of algae, and use this as a
control sample. Whenever a control sample is needed, it is thawed and
diluted. However, whenever algal cells are frozen, thawed and/or diluted,
they are exposed to a series of chemical and physical stresses (Taylor &
Fletcher, 1999) and these processes can severely affect algal cell
morphology and/or physiology (Holm-Hansen, 1963). It is therefore
important to take this into account when using frozen control samples for
algal fluorescence measurement methods.
97
5.4.1. Freezing algal cells
Many
researchers
(dry-)freezing,
have
preparing
tried
serial
to
preserve
subcultures,
microorganisms
storage
under
oil
by
or
lyophilisation (Day et al., 1997). However, these methods are not always
successful. In their experiments, Day et al. (1997) showed that virutally
no Chlorella cells survived freeze-drying or air-drying. Cryopreservation
seemed to be the most promising technique, as 70% of the Chlorella cells
survived the treatment in their experiments. Morris (1976) found a
recovery of only 38% of Chlorella cells after cryopreservation.
The exact procedure for cryopreservation of algae is difficult to determine
and varies per algal species. Taylor and Fletcher (1999) presented an
overview of storage and preservation methods for algae as published over
many years. The formation of ice crystals during the freezing process is
generally the main cause of cell damage. How and where ice crystals are
formed depends largely on the freezing procedure followed, more
particularly the rate of cooling (e.g. applying a 2-step procedure from
room temperature to -30/-50°C to -196°C), and the use of cryoprotective
additivies such as glycerol or DMSO (dimethyl sulphoxide).
Although not as important as the freezing method, the thawing procedure
can also affect algal cell survival (Taylor & Fletcher, 1999). Any
cryoadditivies added during the freezing procedure can become cytotoxic
during thawing, and thus the algal sample needs to be diluted straight
away.
5.4.2. Diluting algal samples
Diluting an algal sample can lead to osmotic stress. Cultured algae are
generally grown in a medium containing > 20 osmol/m3, and dilution with
water (2-2.5 osmol/m3) can increase the osmotic pressure on the cells
considerably. A difference of 18 osmol/m3 is similar to a pressure of 44
kPa (4.5 m water pressure). Coping with such changes in osmotic
pressure requires energy which cannot be used for the photosynthetic or
metabolic process. Papageorgiu and Stamatakis (2004) showed that
changes
in
osmotic
pressure
within
fluorescence response of cyanobacteria.
98
the
cells
directly
affect
the
In the laboratory, algal samples can be diluted with e.g. culture medium
of similar osmolarity in order to avoid osmotic stress. If, for some reason,
samples needed to be diluted in the field, the best possible option would
be to use filtered water from the lake or river under investigation. Of
course, this requires the presence of mobile filtering equipment.
5.4.3. Alternative control samples
Instead of using control samples of real algae, it is also possible to use (a
mixture of) fluorescent dyes, such as rhodamine. Rhodamine exhibits
fluorescent features similar to the phycoerythrin pigment (Nguyen et al.,
1987). A disadvantage, however, is the fact that Rhodamine B poses a
health risk and thus has to be handled with care.
A.
Figure 5.5
B.
Test cuvette for a laboratory fluorometer containing a laser-dyecoated glass prism (A) and a performance attachment for a field
fluorometer
(B),
both
produced
by
bbe
Moldaenke
(Schwentinental, Germany).
Alternatively, it is possible to use a cuvette holding a prism with a coating
(Figure 5.5A) containing a mixture of fluorescent laser dyes which absorb
light in the visible range (400-700 nm). These dyes are stable for a long
period of time and can thus be used to check the configuration of the
fluorometer. The results can be recorded in a control chart (see example
99
in Figure 5.6). The advantages of using these dyes as a coating on a
(glass) surface are that there is no need to prepare solutions of the dyes,
thus avoiding having to handle potentially toxic substances, and the
coated surfaces can (in another form) also be used for field fluorometers
(Figure 5.5B). A disadvantage of this approach is that any errors related
Signal performance attachment /
test cuvet
to sample preparation cannot be observed.
0
Figure 5.6
5
10
15
n
25
30
Example of a chart for control sample results. •: fluorescence
measurement of dye;
deviation;
: plus or minus 3 times standard
: plus or minus 2 times standard deviation;
: average.
100
20
Chapter 6. Cyanobacteria Alert Level Framework
6.
Establishing an Alert Level Framework for
cyanobacteria in drinking water resources using the
Algae Online Analyser for monitoring
cyanobacterial chlorophyll-a
6.1.
Introduction
The occurrence of toxic cyanobacterial blooms is an adverse effect of
eutrophication. Microcystins (toxins produced by Microcystis and other
cyanobacterial genera) are possibly carcinogenic, have tumour-promoting
properties and have been associated with primary liver cancer (Falconer,
2005). Cyanobacterial blooms may therefore limit ecosystem services,
including the provision of drinking water, due to difficulties and the
increased costs of water treatment (Jurczak et al., 2005). Because of its
potential hazard, the World Health Organisation established a guideline
value of 1 μg/L for microcystin-LR in drinking water (WHO, 1996). Since
January 2007, this regulation has become more stringent since the
standard has been extended to total microcystins. However, this guideline
still only covers one of the many different cyanobacterial toxins. Given the
fact that under natural circumstances cyanobacterial blooms are not unialgal, there is a significant risk that mixtures of toxins will occur. Since
conventional drinking water treatment processes are not always fully
effective in toxin removal, humans may risk exposure to toxic substances.
Therefore, it is important to adapt the treatment process to the actual
quality of raw water.
Bartram et al. (1999) proposed an Alert Level Framework (ALF) for
cyanobacteria. This framework provides a monitoring and management
action sequence for water treatment plant operators and managers. The
framework assesses the development of a potentially toxic cyanobacterial
bloom through a monitoring programme with associated actions in four
stages linked to ‘Alert Levels’. The actions accompanying each level cover
categories such as additional sampling and testing, operational measures,
consultation with health authorities and other agencies, and media
releases (House et al., 2004). The ALF is based on cyanobacterial cell
counts and chlorophyll-a concentrations. The threshold levels presented
101
by Bartram et al. (1999) are a continuation of those determined by Burch
(1993) and modified by House et al. (2004) (Table 6.1).
Table 6.1
Comparison of threshold definitions for Alert Levels (AL) for
surface waters proposed by different authors.
Burch
level
(1993)
(cls*/mL)
Bartram et al. (1999)
(cls/mL)
(mm3/L)
House et al. (2004)
(μg chl-a
/L)
(cls/mL)
(mm3/L)
detection
level
-
200
-
0.1
500
-
AL 1
500
20,000
2.0
10.0
2,000
AL 2
2,000
100,000
10.0
50.0
5,000
1
AL 3
15,000
-
-
-
50,000
10
*: cls: cells
Cyanobacterial densities can rapidly increase under favourable conditions,
so continuous monitoring is essential to be able to adequately respond to
changes in water quality. Currently available methods for cyanobacterial
cell counts and chlorophyll-a analysis are time-consuming and unsuitable
for online monitoring. Additionally, the chlorophyll-a analysis method does
not differentiate eucaryotic phytoplankton from cyanobacteria, which is
especially important for the selective detection of cyanobacteria in mixed
phytoplankton assemblages. Phycocyanin is an acknowledged indicator of
cyanobacteria (Osutki et al., 1994; Lee et al., 1994; 1995; Ahn et al.,
2002 and Vincent et al., 2004) and the application of a modern analytical
technique which enables its quick and easy detection is an important tool
for the monitoring of cyanobacterial blooms. Izydorczyk et al. (2005) and
Gregor et al. (2007) indicated that the measurement of phycocyanin
fluorescence is an effective early-warning system for cyanobacteria in
reservoir intake water.
The AlgaeOnlineAnalyser (AOA, bbe Moldaenke GmbH, Schwentinental,
Germany) is a fluorometer which measures chlorophyll-a fluorescence of
algae in water, and allows for separate determination of Chlorophyceae
(green algae), Bacillariophyceae (diatoms, dinoflagellates), Cyanophyceae
(blue-green algae) and Cryptophyceae. The operation procedure is the
same as that of the FluoroProbe (bbe Moldaenke GmbH, Schwentinental,
102
Germany) used in the field, but the AOA is an online laboratory/station
fluorometer. It measures the fluorescence of different pigments, each
being characteristic of a certain phytoplankton group. The AOA contains
five light-emitting diodes (LEDs) which excite the pigments by emitting
pulsated light at wavelengths of 450, 525, 570, 590 and 610 nm. For
further information on this instrument and the measuring principle, see
Beutler et al. (2002).
The main goal of the presented research was to show the application of
cyanobacterial chlorophyll-a measurement, using the AOA, to detect
potentially toxic cyanobacteria in raw water at the drinking water study
site of Sulejow Reservoir. Additionally, concentrations of cyanobacterial
chlorophyll-a were used to assess the hazard of cyanobacterial toxins in
the raw water, and to define thresholds for an Alert Level Framework for
this location.
6.2.
The Sulejow Reservoir is a shallow, lowland dam reservoir situated in
central Poland in the middle course of the Pilica River. At full capacity, the
reservoir covers 23 km2, has an average depth of 3.3 m and a volume of
75 x 106 m3. The mean retention time is about 30 days. The dominant
cyanobacterial species is Microcystis aeruginosa (Zalewski et al., 2000;
Izydorczyk et al., 2005). Analysis of cyanobacterial bloom samples
confirmed their highly hepatotoxic character (Tarczynska et al., 2001;
Jurczak et al., 2004; Mankiewicz-Boczek et al., 2006 and Izydorczyk et
al., 2008). The Sulejow Reservoir is an important freshwater resource for
the city of Łódź. The drinking water intake point is located in the central
part of the reservoir at the end of a narrow bay called Bronisławow (Figure
6.1).
Online fluorescence measurements were carried out at the drinking water
intake point (Sampling Station V in Figure 6.1). The AOA was used as an
online sensor from April 17 to June 26, 2003, and from August 19 to
October 9, 2003. During these periods, fluorescence data were collected
every 15 minutes and recorded by the AOA. Once a week, samples were
analysed
for
biovolume
assessment,
chlorophyll-a
and
toxin
concentrations.
103
Figure 6.1
Location of the sample stations in the Sulejow Reservoir
For calibration purposes, additional samples were taken at four sampling
points in the Sulejow Reservoir (Figure 3.1). Sampling Station I was
situated in the lower, lacustrine part of the reservoir; II in the central,
transitional part, and III in the upper, riverine part. Sampling point IV was
located in the narrow bay directly in front of the water intake point. All
water samples were collected five times during the spring 2003 (April 30;
May 7, 21, 28; June 11) and five times during the summer 2003 (August
20; September 3, 10, 17, 24). At these sampling points, fluorescence
measurements,
phytoplankton
biovolume
assessments,
chlorophyll-a
analyses and toxin analyses were performed. The measurements by AOA
were performed within two hours after sampling in the water intake.
Chlorophyll-a concentrations were analysed by a method based on
acetone extraction and determination by spectrophotometry (Lawton et
al., 1999).
Water samples for the determination of phytoplankton biovolume were
preserved
in
Lugol’s
solution
and
sedimented
in
the
laboratory.
Phytoplankton was counted in concentrated samples, using a FuschRosenthal
counting
chamber
and
a
Nikon
Optiphot-2,
102
(400x
magnification). Phytoplankton biovolume (mm3/L) was determined based
on a volumetric analysis of cells using geometric approximation.
104
Microcystins were analysed in two forms – dissolved in water (extracellular) and cell-bound in suspended matter (intra-cellular). For the
HPLC-analysis, 1 litre water samples containing cyanobacterial material
were filtered immediately after sampling, using Whatman GF/C filters
(pore size 0.45 μm). The preparation of cyanobacterial material and
determination of microcystins by HPLC-DAD were performed according to
descriptions in Jurczak et al. (2005) and Meriluoto and Codd (2005).
Microcystins were identified by their characteristic absorption spectra and
retention times against the following microcystin standards: MC-LR, MCRR, demethyl(dm)-MC-RR, MC-YR, MC-LY, MC-LW, MC-LF. Microcystin
concentrations were expressed as the sum of all varieties. Statistical
analyses were performed using the Statistica® software (Statsoft Inc.,
Tulsa, USA).
6.3.
Results
6.3.1. Seasonal changes of phytoplankton biovolume and
community structure
During spring, the phytoplankton community, as found at almost all
sampling
stations,
was
dominated
by
diatoms,
namely,
Cyclotella
comensis, C. radiosa, Fragilaria capucina, F. crotonensis and Asterionella
formosa (Figure 6.2A). The highest diatom biovolume which varied
between 20 and 58 mm3/L was observed at Station III. Diatoms also
dominated at Station II, where total phytoplankton biovolume oscillated
between 1.6 and 34 mm3/L. Low phytoplankton biovolume was observed
at Stations IV and V. Unfortunately, the April 30 sample from Station IV
was not analysed, although high biovolume was expected. At Station I,
the diatoms and cryptophytes (Cryptomonas ovata, Campylomonas
marssonii, C. platyuris) dominated in early spring, while cyanobacterial
densities increased in late spring.
During
summer,
the
phytoplankton
community
was
dominated
by
cyanobacteria, mainly Microcystis aeruginosa (Figure 6.3A). The maximum
cyanobacterial biovolume was observed at Station IV and amounted to
485 mm3/L on September 10. Biovolume levels at Station I were in the
same range as Station II, varying between 4.2 and 31 mm3/L. Lower
values were observed at Station V, where the maximum phytoplankton
105
biovolume reached 5.7 mm3/L. At Station III, domination of cryptophytes
(max. 18.5 mm3/L) followed by diatoms (max. 17.8 mm3/L) was
observed.
6.3.2. Seasonal changes of chlorophyll-a concentration measured
by spectrophotometry
During spring, a gradual increase of total chlorophyll-a concentrations was
observed at Stations I, II, and III (Figure 6.2B) where values ranged from
2.5 to 71 μg/L. The spring maximum chlorophyll-a concentration was 157
μg/L, observed at Station IV. Lower values ranging from 0.3 to 11.7 μg/L
were observed at Station V.
During late summer, similar chlorophyll-a levels varied between 3.5 and
58 μg/L at Stations I and II (Figure 6.3B). Higher values were found at
Station III (13.8 - 52 μg/L). The summer maximum was observed at
Station IV where values ranged from 4.6 to 251 μg/L, whereas the lowest
chlorophyll-a concentration was found at Station V (1.6 μg/L).
6.3.3. Seasonal changes of chlorophyll-a concentration and
phytoplankton community structure measured by the AOA
During spring, the maximum concentration of total chlorophyll-a, 76 μg/L,
was
observed
at
Station
IV
(Figure
6.2C).
The
chlorophyll-a
concentrations at Station I, II, and III were within a similar range, varying
between 5.5 and 43.7 μg/L. The lowest value was observed at Station V
(0.9 and 9.6 μg/L; Figure 6.2C). During spring, the total chlorophyll-a
originated mostly from Bacillariophyceae and Cryptophyceae. In late
spring, an increased contribution to total chlorophyll-a levels from green
and blue-green algae was observed.
During late summer, increased chlorophyll-a levels from cyanobacteria
were observed at all sampling stations, with the exception of Station III
(Figure 6.3C). At this station, chlorophyll-a from Cryptophyceae and
Bacillariophyceae dominated. The summer maximum concentration of
total chlorophyll-a amounting to 214 μg/L (including 198 μg/L chlorophylla from cyanobacteria) was observed at Station IV. Total chlorophyll-a
concentrations at other stations varied between 2.7 and 40 μg/L.
106
Figure 6.2
Seasonal
dynamics
of
the
phytoplankton
biovolume
(A.),
chlorophyll-a concentration determined by spectrophotometric
method (B.), chlorophyll-a concentration measured by AOA (C.)
and microcystins concentrations (D.) at five sampling stations in
the Sulejow Reservoir and Water Intake during spring 2003 (n.a.
- no analysis).
107
Figure 6.3
Seasonal
dynamics
of
the
phytoplankton
biovolume
(A.),
chlorophyll-a concentration determined by spectrophotometric
method (B.), chlorophyll-a concentration measured by AOA (C.)
and microcystins concentrations (D.) at five sampling stations in
the Sulejow Reservoir and Water Intake during late summer
2003.
108
6.3.4. Seasonal changes of microcystins concentrations
During spring, the maximum concentration of intracellular microcystins
(17.5 μg/L) was observed at Station IV (Figure 6.2D). The spring
maximum concentrations at Stations I, II, III, and V amounted to 2.86,
0.85, 0.40, and 0.45 μg/L, respectively. Microcystins dissolved in water
amounting to 0.2 μg/L were only observed at Station I.
During late summer, microcystins dissolved in water were observed at
Station IV (0.83 μg/L) and Station V. (0.10 μg/L; Figure 6.3D). The
maximum summer concentration of microcystins in cells (6.49 μg/L) was
observed at Station IV. The summer maximum concentration at Stations
I, II, and III amounted to 1.72, 0.78 and 0.04 μg/L, respectively. At
Station V, the intracellular microcystin concentrations varied between 0
and 0.28 μg/L.
6.4.
Discussion
6.4.1. Relationship between concentrations of chlorophyll-a
measured by fluorescence (AOA) and by spectrophotometry
Total chlorophyll-a concentration is used as an indicator for phytoplankton
total biomass. However, the spectrophotometric and HPLC methods used
to determine total chlorophyll-a concentrations are generally timeconsuming. The application of a real-time and simple fluorometric method
such as that used in the AOA can be a valuable tool for monitoring
cyanobacterial blooms.
A
correlation
was
found
between
chlorophyll-a
measured
by
spectrophotometric analysis (Chlspectro) and by fluorescence measurements
using the AOA (ChlAOA), and is described as Equation 6.1:
ChlAOA [μg/L] = 1.350935 * (Chlspectro [μg/L])0.756400
(6.1)
(Analysis of Variance: F=139.8, df=1.43, p<0.00001; Figure 6.4)
109
80
Chlorophyll-a [ µg/l],
Spectrophotometry
70
60
50
40
30
20
10
0
0
10
20
30
40
50
Total chlorophyll-a [µg/l], AOA
Figure 6.4
Relationship between chlorophyll-a concentrations measured by
fluorescence (AOA) and by spectrophotometry. Solid line: y =
1.3716 x1.0111 (R2 = 0.765).
Two higher values of chlorophyll-a were not included in the correlation
analysis because of the statistically insignificant number of days of
intensive blooms. Figure 6.4 shows two points which deviate from the
correlation line (data: August 20, Station I and September 10, Station II;
Figure 6.3). Without these points, a correlation between chlorophyll-a
measured
by
spectrophotometric
analysis
and
by
fluorescence
measurements using the AOA is described as Equation 6.2:
ChlAOA [μg/L] = 1.108796 * (Chlspectro [μg/L])0.805641
(6.2)
(Analysis of Variance: F=380.6, df=1.41, p<0.00001)
Figure 6.4 shows that the relationship between total chlorophyll-a
concentration
measured
by
AOA
and
measured
through
spectrophotometry does not follow the linear correlation y = x. The values
obtained from the fluorescence method are lower than those of the
110
spectrophotometric method. This deviation was expected, because of the
two different methods of measuring chlorophyll-a. The spectrophotometric
method analyses total chlorophyll-a after the destruction of cells by the
extraction step. The fluorescence method, however, analyses only active
chlorophyll-a in the living cells. Inactive chlorophyll-a from dying cells or
cells shaded by others does not show fluorescence and remains therefore
undetected. Because of shading, this effect is more prominent at higher
phytoplankton densities.
Leboulanger et al. (2003) showed a high correlation between total
chlorophyll-a measured by a submersible fluorescence probe (based on
the same measuring principle as the AOA and FluoroProbe; bbe Moldaenke
GmbH, Schwentinental, Germany) and chlorophyll-a measured by a
spectrophotometric method during a bloom of Planktothrix rubescens in
Bourget lake (r = 0,77; n = 55; p < 0,01). Gregor and Maršálek (2004)
and Gregor et al. (2005) obtained similar results while testing this same
submersible probe for the measurement of total chlorophyll-a in a
reservoir and several rivers. They also showed a high correlation between
total chlorophyll-a measured by the FluoroProbe and chlorophyll-a
measured by the spectrophotometric method (r=0.97, n=18 and r=0.95,
n=96, respectively).
Gregor and Maršálek (2004) suggested that the maximum detection limit
for Microcystis blooms by field monitoring of fluorescence is about 50-60
μg/L. At higher densities, the increased shading effect will reduce field
measurement accuracy. During our own research, algal blooms with
densities in this range were observed only twice. Beutler et al. (2002)
estimated that this shading effect can cause a 10% deviation in the
results.
6.4.2. Relationship between cyanobacterial biovolume and
concentration of chlorophyll-a of cyanobacteria determined
by fluorescence (AOA)
The total chlorophyll-a measurements do not describe the composition of
the
phytoplankton
chlorophyll-a
diatoms,
community.
concentration
blue-green
algae
per
and
The
AOA,
which
phytoplankton
cryptophytes),
group
determines
(green
enables
the
algae,
preliminary
111
taxonomic determination. This is especially useful for the selective
detection of cyanobacteria in mixed phytoplankton assemblages. The
cyanobacterial pigments are characterised by maximal excitation at 610
nm, caused by the photosynthetic antenna pigment phycocyanin. This
appears to be a useful indicator to determine the density of cyanobacteria
in water samples (Watras & Baker, 1988).
Cyanobacterial biomass [mg/l]
40
30
20
10
0
0
5
10
15
20
Cyanobacterial chlorophyll-a [µg/l],
Figure 6.5
Relationship between cyanobacterial biovolume determined by
cell counts and concentration of cyanobacterial chlorophyll-a
measured by AOA.
Solid line: y = 0.7656 x1.0981 (R2 = 0.536).
This study demonstrates a positive correlation between cyanobacterial
biovolume and concentration of cyanobacterial chlorophyll-a measured by
the AOA (r = 0.62; n = 46; p < 0.05) (Figure 6.5). This correlation was
demonstrated for conditions when cyanobacterial biovolume was below 40
mm3/L and the phytoplankton was dominated by Microcystis aeruginosa.
Values of biovolume higher than 40 mm3/L were not included in the
correlation analyses because of the statistically insignificant number of
days of intensive blooms.
112
The precision of the counting method for cyanobacterial cells and/or
filaments should be considered when interpreting the results.The precision
of the colonial cyanobacteria count of e.g. Microcystis aeruginosa reaches
more than 50% (House et al., 2004). However, methods in which the
mucilage of colonies of chroococcal cyanobacteria is hydrolysed and a
homogeneous cell suspension is counted, produce much better results
since the uncertainty in counting decreases significantly (Hoogenboezem
et al., 2004).
Additionally, the variation of data points in this correlation (Figure 6.5)
resulted
from
differences
in
the
phycocyanin
concentration
per
cyanobacterial cell. This could be due to either nitrogen levels (Rapala,
1998)
or
light
intensities
(Grossman
et
al.,
1994)
affecting
the
phycocyanin content per cell.
6.4.3. Estimation of hazard from cyanobacterial toxins based on
the concentration of cyanobacterial chlorophyll-a measured
by fluorescence (AOA)
The Alert Level Framework is based on the occurrence of cyanobacterial
cells. All cyanobacteria should be treated with caution, until either testing
confirms the absence of toxicity or past local knowledge indicates the
absence of hazard (House et al., 2004). For this reason, this research also
investigated the estimation of hazard from cyanobacterial toxins based on
levels of chlorophyll-a from cyanobacteria (measured by fluorescence:
AOA).
This study demonstrated a positive correlation between the chlorophyll-a
concentration of cyanobacteria (measured by AOA) and the concentration
of intracellular microcystins (r=70, n=46, p < 0.05). Therefore, it should
be presumed that most cyanobacterial cells in the Sulejow reservoir are
toxic (Figure 6.6).
The trend line shows average values of toxins in cells from all results;
however, in early-warning systems, averages are only of limited value. It
is better to use the maximum value (meaning, the maximum value of
toxins in cells) or ‘the worst-case scenario’ (defined as the 90th percentile
of results, to exclude outliers).
113
Intracellular microcystins [ug/l]
4
3
2
1
0
0
5
10
15
20
Cyanobacterial chlorophyll-a [µg/l], AOA
Figure 6.6
Relationship
between
concentrations
of
cyanobacterial
chlorophyll-a and intracellular microcystins. Solid line: linear
correlation (y = 0.1 x; R2 = 0.495); dotted line: worst-case
scenario.
First, the ratio between the concentrations of intracellular microcystin and
chlorophyll-a of cyanobacteria measured by AOA was calculated. Next, the
90th percentile value was determined in order to indicate the worst-case
scenario (Table 6.2). The 90th percentile proportion was obtained from our
data: 1 μg/L cyanobacterial chlorophyll-a corresponded to 0.2 μg/L
intracellular microcystins; this means that 4.94 μg/L chlorophyll-a from
cyanobacteria corresponded to 1 μg/L intracellular microcystins. To
achieve a more realistic maximum value than one based on a single data
point, these values are based on the total dataset.
114
Table 6.2
Comparison of intracellular microcystins in cyanobacterial cells in
a worst-case scenario defined as the 90th percentile of results, to
exclude outliers using the data of Figure 6.6.
μg microcystins per μg cyanobacterial
chlorophyll-a measured by AOA
maximum
0.28
90 percentile
0.20
average
0.08
Table 6.3 shows threshold definitions for the Alert Levels proposed by
House et al. (2004) based on cyanobacterial cell counts. Thresholds for
microcystin concentration were also included using the above-mentioned
conversion factors, as proposed by Falconer (1994). He proposed a safety
value of 5,000 cells/mL, which corresponds to 0.2 pg toxins/cell. This is
equal to a dose of 1 µg toxins per litre.
Table 6.3
Thresholds
for
cyanobacterial
chlorophyll-a
concentrations,
measured by the AOA in the Alert Levels Framework for Sulejow
Water Intake (Columns 2 and 3 are based on House et al. (2004)
and Falconer (1994), respectively).
levels
cyanobacterial
cyanobacterial cells
microcystin-LR
[cells/mL]
[μg/L]
[μg/L]
500
0.1
0.49
Alert Level 1
2,000
0.4
1.87
Alert Level 2
5,000
1.0
4.94
Alert Level 3
50,000
10.0
49.43
Detection Level
chlorophyll-a, AOA
The advantages of the AOA are its potential for adjusting sampling
frequency, as well as the fact that all measurements are performed online.
These features are especially important for monitoring the drinking water
intake point, where toxic algae blooms may appear suddenly, requiring
immediate
action.
The
dynamics
of
daily
average
cyanobacterial
chlorophyll-a concentrations determined by the AOA confirmed the
importance of continuous measurements (Figure 6.7). Fluctuations in this
graph are the result of e.g. wind speed and direction, entrainment,
115
vertical migration of the algae or zooplankton predation. Until this time,
the drinking water intake point had been monitored once or twice per day
using
the
chlorophyll-a
spectrophotometric
method
and
biovolume
assessment. Since daily cycles had not been taken into account, this
provided insufficient information about changes in plankton abundance
throughout the day. An additional important advantage of using the AOA
for monitoring is the short time necessary to obtain results (due to online
measurements). This allows for the maximum possible time to take the
necessary actions (increase in treatment substances dosage, start of
additional treatment processes or closure of the intake). Additionally, the
AOA is a small instrument which can be installed almost anywhere. If
desired, it could be installed after each treatment step in order to
determine its effectiveness in the removal of algae.
Daily average cyanobacterial
15
10
ALERT LEVEL
2
5
17-04
24-04
01-05
08-05
15-05
22-05
29-05
05-06
12-06
19-06
26-06
03-07
10-07
17-07
24-07
31-07
07-08
14-08
21-08
28-08
04-09
11-09
18-09
25-09
02-10
09-10
0
ALERT LEVEL
1
DETECTION
LEVEL
Date
Figure 6.7
Daily average concentrations of cyanobacterial chlorophyll-a
determined by AOA with indicated thresholds of specific Alert
Levels for Sulejow Water Intake. No data was collected between
early July and half August.
According to our results and experience, the AOA is a useful tool for the
monitoring of potentially toxic cyanobacterial blooms at the drinking water
intake
116
point.
We
defined
thresholds
for
Alert
Levels
based
on
concentrations of cyanobacterial chlorophyll-a, specifically for the case of
surface water used in the production of drinking water. In this study,
these thresholds were defined specifically for our drinking water intake,
however the procedure could easily be adapted to other locations.
117
Chapter 7. Application of Alert Level Framework
7.
Cyanobacterial and total chlorophyll-a
concentrations as a method for early-warning
water quality monitoring
7.1.
Introduction
Water quality is directly affected by discharges resulting from human
activities, and changing climatic conditions. Climate change scenarios
predict that rivers, lakes, and reservoirs will experience increased
temperatures, more intense and longer periods of thermal stratification,
modified hydrology, and altered nutrient loading (Newcombe, 2012; Pearl
& Paul, 2012). These environmental changes will have substantial effects
on freshwater phytoplankton species composition and biomass, potentially
favouring cyanobacteria over other phytoplankton (Cayelan et al., 2012).
Cyanobacteria cause discoloration and reduced visibility in recreational
waters and drinking water resources (De Julio et al., 2010)].
The adverse effects of cyanobacteria have raised public health concerns in
many countries in recent years for a variety of reasons. These include
cases of poisoning attributed to toxic cyanobacteria and the contamination
of water resources (especially lakes) resulting in increased cyanobacterial
growth. The development of strains containing toxins is a common
experience in polluted inland water systems all over the world, as well as
in some coastal waters. Thus cyanobacterial toxins, or "cyanotoxins", have
become a concern for human and animal health (Marian et al., 2007;
Cook et al., 2004; Gregor & Maršálek, 2005; Stewart et al., 2006).
Organisms responsible for cyanobacterial poisoning (CTP) include an
estimated 40 genera, but the main ones are Anabaena, Aphanizomenon,
Cylindrospermopsis,
Lyngbya,
Microcystis,
Nostoc
and
Oscillatoria
(Planktothrix) (Carmichael, 2001; Pavlova et al., 2006). Toxin production
by cyanobacteria appears highly variable, both within and between
blooms. Toxin production and potency can also vary with time for an
individual bloom (Duy et al., 2000). Globally, the most frequently found
cyanobacterial toxins in blooms from fresh and brackish waters are the
cyclic peptide toxins of the microcystin and nodularin families (Cook et al.,
2004; Pavlova et al., 2006; Blahova et al., 2007). Thus, the monitoring of
119
cyanobacteria is important to determine their dynamic development
throughout the seasons.
Several researchers have identified alert levels of potentially dangerous
cyanobacterial densities, and defined monitoring and management actions
to reduce adverse effects on public health (e.g. Bartram et al., 1999).
These Alert Level Frameworks are based on cyanobacterial cell counts and
total chlorophyll-a concentrations. These alert levels were translated into
cyanobacterial chlorophyll-a levels for a Polish lake by Izydorczyk et al.
(2009).
This study focused on applying the Alert Level Framework as developed by
Izydorczyk et al. (2009) to a seasonal assessment of water quality in a
lake in Warsaw City. Maintaining and preserving water quality in this lake
has been a long-standing goal of residents and other interest groups.
Given trends in regional land use, the Predictive Season-Based Water
Quality Model for Lake Powsinkowskie (Dobrzańska & Hewelke, 2010) has
shown that the most immediate threat to the water quality of Lake
Powsinkowskie is nutrient enrichment due to an increased population, and
associated development near the lake. Information regarding the current
status of the water quality in Lake Powsinkowskie has been limited. The
timing and duration of the cyanobacterial bloom season depend largely on
the climatic conditions of the region. In temperate zones (such as Poland),
blooms of cyanobacteria are most prominent during the late summer and
early fall and may last for 2–4 months (Chorus & Bartram, 1999). The aim
of this research was to assess the severity of cyanobacterial development
in the Polish lake.
7.2.
Lake Powsinkowskie (52°9'27"N, 21°5'55"E) is situated in the Wilanуw
region, in Warsaw, Poland. The Vistula river enters the lake on the eastern
side. Just north of Lake Powsinkowskie is another lake: Lake Wilanowskie
(Krauze-Tomczyk & Ostrowski, 2006). On one side of this lake, there is a
residential area, and on the other side are fields owned by the Warsaw
University of Life Sciences (SGGW, Poland) (Figure 7.1). For many years,
these fields have been used for experiments using fertilizers and/or
pesticides at high and normal concentrations to increase agricultural
120
production, and – more recently – to develop more efficient water
resource management concepts in relation to agricultural activities under
changing climatic conditions (Romanowska-Duda et al., 2003; 2010).
Figure 7.1
Experimental
area
of
Lake
Powsinkowskie,
including
the
residential area and the fields in use by SGGW.
Measurements of cyanobacterial chlorophyll-a in Lake Powsinkowskie were
performed from May to July 2012. The weather forecast of Warsaw
reported that during these three months the temperatures in Warsaw
were between 12 and 34°C (data from Meteorological Institute).
Cyanobacterial chlorophyll-a levels were measured using the AlgaeTorch
(bbe Moldaenke GmbH, Schwentinental, Germany). The AlgaeTorch is a
fluorescence measurement device specifically developed for performing
field measurements of total chlorophyll-a and cyanobacterial chlorophyll-a
(Gregor & Maršálek, 2005; Pobel et al., 2011). This instrument is based
on the same principle as the bbe FluoroProbe (Beutler et al., 2002) and
provides a rapid measurement of the concentrations of cyanobacterial and
total chlorophyll-a in the water. Due to the fact that algae of the same
division contain a similar quantity and quality of photosynthetic pigments,
their fluorescence excitation spectrum (with a fixed emission wavelength
at 680nm) is significant. Thus, it is possible to differentiate divisions of
algae by their fluorescence excitation spectrum. The bbe AlgaeTorch for
algae differentiation uses 7 LEDs for fluorescence excitation. These LEDs
121
emit light at 3 selected wavelengths (470nm, 525nm, 610nm; 2 LEDs
each). An additional LED is used for turbidity estimations based on the
reflected light of any particles in the water. The turbidity estimations are
used to indicate the reliability of the chlorophyll-a measurements. At high
turbidity levels, the chlorophyll-a measurements should be interpreted
with care, since artifacts such as shading may occur, resulting in
measurement errors. The turbidity estimations of the AlgaeTorch are not
accurate enough for the determination of turbidity as a water quality
parameter.
In this study, the torch was immersed to a depth of 20 cm at five
sampling points once an hour, and triplicate measurements were
performed at each point. The measurement series was performed once
per week on both sides of the lake. 10 measurements were performed at
different places, to obtain average chlorophyll-a levels for each side of the
lake. The spatial concentration of cyanobacterial and total chlorophyll-a
results in the lake were represented using Surfer (v. 7.0, Golden Software
Inc.), and statistical analyses were performed using SPSS (IBM, SPSS
Inc.; version 19.0.0).
7.3.
Results
Figure 7.2 shows the results of cyanobacterial and total chlorophyll-a for
both the (western) residential side and the (eastern) SGGW side of the
lake (called “fields”). The results show that cyanobacteria comprised
generally only a small portion of the total algal densities during the time of
the investigations. The variability in the results was higher for the
measurements performed on the residential side of the lake as shown by
the larger bars indicating the 95% confidence intervals of the results in
Figure 7.2B. This indicates a more patchy algal distribution.
122
Figure 7.2
Results of cyanobacterial (solid line) and total chlorophyll-a
(dotted line) for the SGGW side (“fields”; A) and the residential
side (B) of Lake Powsinkowskie.
Alert Level 1 indicates that monitoring frequencies should be increased,
and the lake should be checked regularly for scum-forming. Alert Level 2
was not exceeded during the investigations (Figure 7.3); this would have
triggered a further increase in sampling frequencies, including microscopic
analysis to determine species composition and, if necessary, toxin
analysis. Figure 7.3 shows the average cyanobacterial chlorophyll-a levels
123
for both the residential and the SGGW sides again, this time in relation to
Alert Levels 1 and 2 according to Izydorczyk et al., (2009). Figure 7.3B
shows that the average cyanobacterial biomass exceeded Alert Level 1
twice at the residential side of the lake during the time of the
investigations. The average cyanobacterial chlorophyll-a levels did not
exceed Alert Level 2 on the SGGW side of the lake (Figure 7.3A). The
upper limits of the 95% confidence interval can be regarded as a worstcase situation. These upper limits exceeded Alert Level 1 on 5 dates on
each side of the lake. Alert Level 2 was exceeded by one upper limit on
14th June on the residential side of the lake, thus indicating that
cyanobacterial biomass has the potential to form harmful scums on at
least one side of the lake.
Figure 7.3
Results of average cyanobacterial chlorophyll-a related to Alert
Level 1 (1.9 µg/L cyanobacterial chl.-a, yellow line) and Alert
Level 2 (4.9 µg/L cyanobacterial chlorophyll-a, red line) for the
SGGW side (“fields”; A) and the residential side (B) of Lake
Powsinkowskie.
124
7.4.
Discussion
Although the results of Lake Powsinkowskie in 2012 do not indicate an
immediate cause for concern, further investigations are recommended to
identify the nature of cyanobacterial developments in the lake. The
investigations were performed from May until the end of July. However,
the highest cyanobacterial densities would have been expected towards
the end of the summer. Unfortunately, it was impossible to continue the
measurements after July 2012, so the highest densities were probably not
observed and may have occurred in August – September 2012. The
highest algal densities were expected at the SGGW side of the lake due to
the use of fertilizers on the fields belonging to the University. However,
the residential side showed slightly higher algal densities. Algal densities
vary in space and time, and the predominantly western winds in the
summer period would have concentrated algal densities on the SGGW side
of the lake; this was not the case here however. It seems that the
conditions on the residential side of the lake were more favourable for
algal growth than on the SGGW side.
It is known that in Central/Eastern European countries and developing
countries many people do not connect their houses to the wastewater
network of the city but use their own cesspool. Some of these cesspools
may be linked to the lake, indeed some pipes were found from the
residential area leading towards the lake. It is also known that poorly
functioning septic systems are a common source of nutrients that can also
stimulate algal blooms (Backer, 2002). Additionally, in many urban areas,
leaching of nutrients from land to water is a significant and growing
problem (Stehl-Delbanco et al., 2003). Since dense cyanobacterial
growths depend on high amounts of nutrients, they are favoured by
nutrient input from land, thus leading to the increasing intensity and
frequency of blooms in lakes and estuarine waters (Pearl, 1988). These
factors may have caused the increase in total chlorophyll-a as well as the
elevated cyanobacterial chlorophyll-a levels on the residential side of the
lake.
The monitoring of cyanobacteria is of growing interest because of its
potential public health risk, as was also shown by Pavlova et al. (2006).
Real-time monitoring of cyanobacteria using field fluorometers can serve
125
as an early-warning system for potentially hazardous conditions. Different
types of fluorescence instruments used to detect the in vivo fluorescence
of cyanobacterial pigments in natural waters have been developed and are
used for monitoring potentially toxic cyanobacteria (Zamyadi et al., 2012;
Jeffrey et al., 1997). The Alert Level Framework based on cyanobacterial
chlorophyll-a
levels
is
a
rapid
and
easily
applicable
method
for
determining the severity of cyanobacterial developments using in situ
fluorometry.
126
Section 3
Phytobenthos biomass:
In vivo fluorescence of chlorophyll-a
Based on:
Carpentier, C., A. Dahlhaus, N. Van de Giesen and B. Maršálek. 2013.
The influence of hard substratum reflection and calibration profiles on in
situ fluorescence measurements of benthic microalgal biomass as an
indicator of trophic state. Environ. Sci.: Processes Impacts, 15, 783793.
127
Chapter 8. Introduction to phytobenthos
8.
Introduction to in situ and in vivo fluorescence
measurements to determine phytobenthos biomass
8.1.
From phytoplankton to phytobenthos
In vivo fluorescence measurements as an indicator of algal biomass have
been performed as in situ measurement for many years (Lorenzen, 1966;
Heaney, 1978; Gregor and Maršálek, 2004; Izydorczyk et al., 2005,
2009),
however
this
method
has
mostly
been
used
to
analyse
phytoplankton. With this method, the main algal groups (mainly green
algae,
diatoms,
cryptophytes
and
cyanobacteria)
are
quantitatively
differentiated by their specific fluorescence excitation spectrum (see
Chapter 4).
The fluorescence method for analysing phytoplankton biomass is not
directly applicable to phytobenthos analysis. Two major aspects need to
be taken into account when performing fluorescence measurements on
phytobenthos:
1. The algae are not suspended in the water column but grow in layers
on a substratum, which can cause a self-shading effect. This is
important as the in vivo fluorescence method only detects active
chlorophyll-a, i.e. chlorophyll-a showing a fluorescence response as a
result of the uptake of light energy.
2. The substratum on which the algae grow influences the measurement
results through reflection and transmission effects.
These effects are described in detail and quantified in Chapter 8, which is
based on Carpentier et al. (2013).
8.2.
Considerations regarding field fluorescence measurements
of phytobenthos biomass
When performing fluorescence measurements on phytobenthos in the
field, it is important to follow a standard procedure to allow for
comparison of the results. The text below is an excerpt of a standard
operating procedure (SOP) developed on the basis of practical experience
gained from the field work performed to collect the results described in
Chapters 8, 9 and 10. The motivation was to assist other biologists or field
129
Section 3: Phytobenthos biomass
workers in carrying out fluorescence measurements of phytobenthos
biomass in the field.
8.2.1. Preparations
1 Site selection
The selection of a suitable sampling/measurement site is most often a
compromise between the most ideal location with regard to the survey’s
objectives and practical restrictions in the field. When selecting a proper
field location for sampling or performing in situ measurements, a number
of considerations should be taken into account.
For the broad scale (a sampling location within a catchment), an existing
sampling site is preferred over a new site. This provides the opportunity to
use (historical) environmental information from previous surveys.
The choice of an appropriate stretch will be determined partly by the
choice of substratum. However, it is important to select a stretch which
best serves the objectives of the study. The selection of the precise area
within a reach is directly related to the substratum, and differs for fastand slow-flowing rivers.
1.1 Fast-flowing rivers
 Cobbles and boulders are the easiest to sample/measure.
 Avoid pools and ponded areas, so as not to include loosely attached
taxa.
 Ensure the measurements are performed in a representative way,
rather than randomly1. The use of transects is advised.
 Select areas affected by a similar light regime, for example by bankside
shading.
 The minimum depth of the substratum in the water is insignificant, as
long as the substratum is completely submerged and not exposed to the
air.
 The maximum depth of the substratum in the water is insignificant, as
long as it is still in the euphotic zone.
1
In a random sample, each stone has the same chance of being selected for a measurement. In a
representative sample, specific stones are selected to represent algal biomass under specfic
conditions (e.g. light regime).
130
1.2 Slow-flowing rivers
 In slow-flowing rivers, it may be impossible to sample cobbles or
boulders, because of the deposition of silt. In this case, vertical hard
surfaces, such as quays and bridge stanchions, can be used.
 Always choose an inert substratum such as stone or concrete, and avoid
wood.
1.3 Description of measuring location
In all cases, it is important to make detailed notes of the sampling
location, in order to aid data interpretation later on. When making these
notes, keep the following in mind:
 A detailed description of the site is required on the first visit; on
subsequent visits notes of unusual occurrences must be made at the
very least.
 Describe the area in clear details, including sketch maps and/or photos.
 Use a standard format for the site description, containing at least:
o Name of the sampler;
o Date and time of sampling;
o Name of river sampled and a location identification description;
o Measurements of the channel width and depth;
o Estimates of substratum composition;
o Cover of filamentous algae and other macrophytes;
o Extent of bankside shading;
o Weather conditions during sampling;
o Time since last spate (flood or storm).
2 Sampled substrata
The favoured substratum for studying the benthic algae community is
rocks/hard surfaces. This substratum is widely available in many rivers.
Although
geology
has
a
pronounced
influence
on
benthic
algal
communities, the type of stone can usually be discounted.
When using artificial substrata, several additional criteria for the selection
of suitable substrata are important. Firstly, the substratum itself may not
possess fluorescent characteristics, such as may be the case for certain
types of unglazed tiles or bricks. Reed stems cannot be used, as they
contain fluorescent chlorophyll-a which can result in an overestimation of
131
phytobenthic biomass. When using glass slides, the slides must be
covered with a non-transparent material on one side in order to avoid
measurement errors as a result of the fact that the phytobenthic biofilm
develops on both sides of the glass slide. Nutrient leaching substrata such
as wood may influence biomass development considerably. Additionally,
the substratum has to be chosen in such a way that it does not attract the
attention of passers-by. Ropes hanging from the shore are susceptible to
deliberate removal, which can disrupt the entire experiment. For artificial
substrata,
a
constant
exposure
time
is
necessary
for
meaningful
comparisons. The minimum advised period is 4 weeks, longer if the water
is oligotrophic, temperature is low or if there is heavy shade.
8.2.2. Measurement procedure
1 Cobbles and boulders
Benthic algae show patchiness in growth and therefore it is important to
carefully select the number of cobbles or boulders measured in one
stretch and the number of measurements on one single stone (see Section
8.4
for
further
considerations
regarding
spatial
variability).
All
measurements should be performed in the same way, preferably while
lying in the water. Do not remove the stones from the water unless it is
too dangerous to perform the measurements in the stream itself or if
there is too much interference from other factors such as turbidity. In
case of bright sunlight and measurements just below the water surface, it
may be necessary to allow for a dark adaptation time of approximately 2
minutes in order to avoid interference of the measurements by photoinhibition.
Possible sources of variability in the measurement results are e.g.:
 the presence of mud or silt, causing the sample to be covered by
particles;
 saturation by thick chlorophyll layers;
 the natural diversity of the algae;
 the presence of filamentous algae
 photo-inhibition by high sunlight intensity
 the migration of algae (on soft sediment only)
132
When performing the measurement, ensure the fluorometer is placed on
the stone in a precisely vertical position. Measure at least five different
stones per stretch and allow for at least ten consecutive measurements
per stone. If large fluctuations in the results are shown, more than ten
measurement values are necessary. Clean the fluorometer between
measurements on different stones by rinsing it with drinking water and
drying it with a towel, to avoid cross-contamination.
After measuring the required number of stones from one stretch, decide
which stones to sample for taxonomic analysis (if applicable). Choose
stones with a sufficiently high density of benthic algae, and preferably a
variety of algal groups.
2 Artificial substrata
Perform the measurements on the substratum in the same way as the
boulders and cobbles. Select different parts of the substrate in order to
achieve a random result.
PE foils, PVC, ceramic or glass tiles can be taken to the lab for
measurements and taxonomic analysis, but in this case the foils should be
transported while kept in stream water, so as not to influcence the algae
too much. When performing the measurements in the lab, keep the foils in
the water.
8.3.
Fluorescence measurements versus HPLC-analysis of
chlorophyll-a
8.3.1. Sampling or field measurements?
Reliable field data are at the basis of a successful (ecological) study of any
kind. In 2009, a new CEN Guidance Standard was introduced for the
sampling and analysis of phytobenthos (i.e. not just diatoms) in shallow
running waters (EN 15708: Water quality - Guidance standard for the
surveying, sampling and laboratory analysis of phytobenthos in shallow
running water). This method, as well as Guidance Standard EN 13946
(2003) on diatom sampling, is based on taking samples of (preferably)
epilithic algae from a number of cobble stones using a scalpel or a
toothbrush. The resulting suspension is collected in a bottle and then
taken to the laboratory where samples are prepared for microscopic
analysis. In the case of diatoms, permanent slides are prepared. However,
133
a major disadvantage of these methods is that they were initially
developed for qualitative biodiversity research rather than quantitative
analysis.
The sampling method as described in EN 13946 for diatoms and EN 15708
for all phytobenthic algal groups on hard substrata was evaluated in the
field by performing in situ fluorescence measurements before and after
sampling. For this experiment, five cobble stones (approximately 10-15
cm in diameter) were collected from the Odense River at the Skallebanke
Wetland on Fyn Island, Denmark (latitude N 55° 16.176′; longitude E 10°
13.643′). In situ fluorescence measurements were performed using a
BenthoTorch
fluorometer
(bbe
Moldaenke
GmbH,
Schwentinental,
Germany). The BenthoTorch uses three different LEDs for fluorescence
excitation of different algal pigments (Table 8.1), and one LED to
determine any reflection effects from the substratum beneath the algal
biofilm (see Chapter 9).
Table 8.1
LED
Excitation wavelengths of the BenthoTorch and corresponding
pigments and algal groups.
excitation
associated pigment
corresponding algal group
wavelength
(nm)
1
470
chlorophyll-a+b
Chlorophyceae
2
525
xanthophyll /
Bacillariophyceae (diatoms),
fucoxanthin, peridin
Dinophyceae
3
610
phycocyanin
Cyanobacteria
4
700
no pigment
correction of background
reflection
After the initial fluorescence measurements, the stones were brushed with
a clean toothbrush and/or scraped with a scalpel to remove as much of
the biofilm as possible. The size of the brushed surface area was carefully
measured for the purpose of biomass calculation. After brushing, both the
stone’s surface and the toothbrush were visually examined for the
presence of algae. Finally, another fluorescence measurement was
performed to check whether the entire biofilm had been removed from the
stone’s surface.
134
Table 8.2 shows the results of the fluorescence measurements before and
after sampling. The results show that not all of the biofilm had been
removed after sampling. The removal rates found on different stones
varied per stone and per algal group.
Table 8.2
stone
Fluorescence measurements before and after sampling according
to EN 13946, and the percentage of algal biomass removed.
algal group
biomass before
biomass after
sampling
sampling
-2
removal (%)
(µg.cm )
(µg.cm-2)
cyanobacteria
0.51
0.16
68
diatoms
0.23
0.18
23
green algae
0.14
0.02
86
cyanobacteria
0.16
0.03
81
diatoms
0.32
0.07
78
cyanobacteria
1.40
1.07
24
diatoms
0.08
0.07
13
cyanobacteria
0.21
0.02
91
diatoms
0.22
0.05
77
cyanobacteria
2.55
1.65
35
diatoms
0.10
0.04
57
green algae
1
2
green algae
3
green algae
4
green algae
5
Figure 8.1 shows the removal rates per stone for the cyanobacteria. On
none of the stones were the cyanobacteria completely removed. Stones 3
and 5 especially show a considerable residue of cyanobacteria after taking
the sample. A higher cyanobacterial biomass seemed to result in the
lowest removal rates (compare Stones 3 and 5 with Stones 1, 2 and 4 in
Table 8.2), which is consistent with the findings of Biggs et al. (1998a).
135
Chlorophyll-a (µg/cm2)
3
2
1
0
1
Figure 8.1
2
3
Stone
4
5
Cyanobacterial chlorophyll-a before (black) and after (grey)
sampling of the five cobble stones of Table 8.2.
A visual inspection of the toothbrush after sampling showed that at least
some of the sample had been left behind in the hairs of the brush, as
these were slightly green. The visual inspection of the stones after
sampling showed a thin green layer still visible where the sample had
been taken, indicating that not all algae had been removed from the
stone. This is consistent with the fluorescence measurements after
sampling. It was, however, impossible to remove the entire biofilm by
toothbrush or scalpel because of the uneven surfaces of the stones.
The results of this experiment show that not all of the biofilm could be
removed by brushing and scraping. These findings are consistent with
those of Cattaneo and Roberge (1991), who found that in streams and
rivers in which tightly attached algae were predominant only a quarter of
the phytobenthos was removed by the scraping or brushing method.
Davies and Gee (1993) compared the results of the scraping or brushing
method with a phytobenthos sampler made of two disposable syringes as
described by Loeb (1981) and with their own sampler which used scouring
pad discs to remove phytobenthos from uneven stones. The scouring pad
disc method removed significantly more phytobenthos biomass than the
other methods, although the results showed a larger standard error.
136
Considering that one part of the sample had been left behind in the
toothbrush and another on the stone, the overall observation was that the
sampling procedure as described in EN 13946 is not suitable for
quantitative analysis of phytobenthos. The fact that removal rates varied
per stone and per algal group increased the discrepancy between the
sample taken and the reality of the situation in the field.
8.3.2. Comparison of fluorescence measurements with HPLCanalysis according to DIN 38412-16
As the results of the experiment in Section 8.3.1 had shown that sampling
natural stones introduces a considerable error in the analysis results of
phytobenthos biomass, the in situ fluorescence measurements and HPLCanalysis of algal pigments were compared using flat tiles of artificial
substrata, which can be sampled more easily than natural stones. It was
assumed that in this way, the sampling procedure in itself would not affect
any correlation between the two methods.
This experiment was conducted at a small peninsula in the Westensee, a
lake near Kiel, Northern Germany. The depth of the lake at the
experimental location was approximately eighty centimetres and the water
was transparent to the bottom. This experiment was performed in a lake
rather than a stream, in order to facilitate the use of artificial substrata.
The chosen location for this experiment had a sandy bottom with little
vegetation, thus avoiding shading of the samples and the mechanical
destruction of the biofilm on the artificial substratum.
A set of twelve black and ten grey PVC tiles cut into squares of 4x8 cm
and twelve glass slides were used as substrata for the phytobenthos. The
PVC tiles were mounted together with two frames for the object slides on
a wire on four poles at a depth of 25cm below the water surface, just
beyond the reed belt of the lake. Two samples of each substratum were
taken and the substrata were replaced immediately after sampling every
ten days between April and August 2011. Each tile was measured in five
places with the BenthoTorch before taking it to the laboratory. The
remaining substrata were measured in two places and left at the sampling
site. At the laboratory, the samples were scraped according to CEN
Guidance Standard EN 13946 (2003), and the biofilm was suspended in
137
water quantitatively. The suspension was filtrated onto GF/C glass fibre
filters (Whatman), pre-treated and analysed with HPLC according to DIN
38412-16 (1985).
The comparison of in situ fluorescence measurements and HPLC-analysis
of algal pigments is shown in Figure 8.2. The graph shows a linear
relationship between the two methods within the investigated density
range of 0 – 6 µg chlorophyll-a per cm2. The significance of the regression
line was calculated according to Fowler et al. (1998), and the linear
relationship between the BenthoTorch measurements and the DIN-method
was found to be significant (t = 1.9997; df = 69; P = 0.05).
Figure 8.2
138
Comparison between in situ fluorescence measurements using the
bbe BenthoTorch and HPLC-analysis of algal pigments on two
types of artificial substrata. The trend line is shown as a dotted
line (R2 = 0.834; n = 71); the line y = x is shown as a solid line.
The correlation between in situ fluorescence measurements and HPLCanalysis of algal pigments suggests that reliable field data of phytobenthic
chlorophyll-a as an indicator of benthic algal biomass can be obtained
when sampling errors have been eliminated. This suggests that in situ
fluorescence measurements could be a better alternative for biomass
quantification than sampling.
8.4.
Spatial variability of phytobenthos
The patchy distribution of phytobenthos requires a careful selection of
sampling sites, and some thought should be given as to the number of
measurements to be performed in order to achieve a representative
result. Guidance Standard EN 15708 (2009) indicates that at least five
randomly selected cobble stones should be used for Single Habitat
Sampling (SHS), however, the standard does not provide a (statistical)
foundation for this number.
The purpose of the study largely determines what is accepted as a
sufficiently representative result. An example is given below of how the
number of measurements necessary to obtain a suitably representative
result can be determined. The statistics used are identical to the method
for determining the number of analyses to be performed to determine a
suitably precise mean biovolume of a phytoplankton taxon, as described in
Chapter 3.
The data to determine the minimum number of measurements necessary
for a representative result and illustrating the patchy distribution of
phytobenthos were collected in 2007 as part of a larger project, the
Second Joint Danube Survey (JDS2), organised by the International
Commission for the Protection of the Danube River (ICPDR). A detailed
description of the survey, its sampling locations, analysis procedures and
results can be found in Liška et al. (2008).
139
A.
140
B.
Figure 8.3
Results of the fluorescence measurements performed on 32
cobble stones on the left bank (A.) and 20 cobble stones on the
right bank (B.) of the River Danube, at Bratislava, Slovak
Republic.
green algae
cyanobacteria
diatoms
For this experiment, 52 cobble stones were collected for fluorescence
measurements on the left bank (at 1869 km) and right bank (at 1865 km)
of the River Danube at Bratislava, Slovak Republic. Five measurements
per stone were performed at the same sampling spot, and the results
were averaged. The measurements per stone took about 30 seconds; a
measurement was made once every four seconds. The results of the
fluorescence measurements are shown in Figure 8.3 and illustrate the
differences between sampling spots only a few centimetres away from
each other.
The differences found can be explained by the patchy distribution of the
phytobenthos, caused by small differences in light conditions, flow
velocity, grazing or other small-scale influences. The results from both
banks show a considerable difference in the ratios between algal groups,
141
especially for green algae and cyanobacteria (Figures 8.3A and B). The
ratios of algal groups for the total biomass at the left and right banks are
presented in Table 8.3.
Table 8.3
Average biomass and ratios of green algae, cyanobacteria and
diatoms for the left and right banks of the Danube at Bratislava.
total
green algae
2
left bank
right bank
total
cyanobacteria
2
diatoms
2
chlorophyll-a
(µg/cm )
(µg/cm )
(µg/cm )
(µg/cm2)
0.29 ± 0.078*
0.44 ± 0.066
0.08 ± 0.025
0.81 ± 0.111
(36%**)
(54%)
(10%)
0.59 ± 0.077
0.61 ± 0.164
0.14 ± 0.039
(44%)
(46%)
(10%)
0.40 ± 0.067
0.51 ± 0.074
0.10 ± 0.022
(40%)
(50%)
(10%)
1.34 ± 0.217
1.01 ± 0.124
*: 95% Confidence interval of mean value.
Left Bank: n = 32; t = 1.96; Right Bank: n = 20; t = 2.093; Total: n = 52;
t= 1.96.
**: Percentage of total amount given in parentheses.
A Mann-Whitney U test was performed to determine the null-hypothesis,
i.e. there is no difference in algal densities (per group or in total) between
the left and the right banks (n1=32; n2=20).
This test showed that:

there is no significant difference in cyanobacterial biomass between
the left and right banks (P=0.114);

there is a significant difference in diatom biomass between the left and
right banks (P<0.05);

there is a highly significant difference in both green algal and total
biomass between the left and right banks (P<0.01).
These differences could have been caused by local differences in e.g. flow
velocity, light regime, nutrient availability and/or grazing pressure.
The results for total biomass, expressed as chlorophyll-a per square
centimetre, are presented in Figure 8.4. The width of the 95% confidence
interval (CI95%) was calculated using formulae taken from Sokal and Rohlf
142
(1995), shown in the graph as a grey area. The width of this interval
indicates the precision of the calculated mean and is determined by the
number of measurements and the standard deviation. By performing more
measurements, the interval becomes smaller, which means that the mean
becomes more precise. Figure 8.4 shows that a 94% reduction of the
width of the CI95% was achieved between the second and the 25th
measurement
(from
7.14
to
1.46
µg.cm-2).
Between
25
and
50
measurements, the width of the CI95% changed from 0.41 to 0.26 µg.cm-2,
a reduction of 37%. This shows that after a certain number of
measurements, the contribution of each extra measurement to the
precision of the mean becomes less important. The running mean and
standard deviation at 5, 25 and 50 measurements are given in Table 8.4.
Table 8.4
n
Running mean, standard deviation and t-values for the 95%
confidence interval of total chlorophyll-a at 5, 25 and 50
measurements (see also Figure 8.4).
running mean
standard deviation
t (P = 0.05)
5
0.74 ± 0.59
1.46
2,5706
25
0.96 ± 0.20
0.41
2,0595
50
1.01 ± 0.13
0.26
2,0086
The results in Table 8.4 show that the 95% confidence interval at five
measurements covers a range of 80% above or below the running mean.
This range is 20% at 25 measurements and 13% at 50 measurements.
Depending on the aim and scope of the investigation, a well-chosen limit
value for the width of the CI95% minimises the efforts and ensures that
enough measurements are performed to obtain suitably representative
results. The extent of the desired precision can be defined by setting limits
for the maximum allowed width of the interval. For the Joint Danube
Survey of 2007, a maximum range of the 95% confidence interval of 20%
above or below the mean was considered sufficiently precise, and this
criterium
was
reached
after
25
measurements.
This
number
is
considerably higher than the minimum of five cobble stones which
Guidance Standard EN 15708 requires for sampling.
143
5
Total chlorophyll-a (µg/cm2)
4
3
2
1
0
-1
-2
-3
-4
0
5
10
15
20
25
30
35
40
45
50
Number of measurements
Figure 8.4
144
Total chlorophyll-a as determined by fluorescence measurements
for the River Danube at Bratislava (•: values of individual
measurements; −: running mean). The grey area indicates the
width of the 95% confidence interval, as a measure for the
precision of the mean.
Chapter 9. Substratum reflection
9.
The influence of hard substratum reflection and
calibration profiles on in situ fluorescence
measurements of benthic algal biomass
9.1.
Introduction
In vivo chlorophyll-a fluorescence measurements of phytoplankton have
been carried out for a variety of purposes in both marine and freshwater
environments for many years. The basic principle of in vivo fluorescence
measurements has been described in detail by Beutler et al. (2002). In
situ chlorophyll-a fluorescence measurements have also been carried out
in coastal and marine systems for several years in phytobenthos research
(Serôdio et al., 1997; Kromkamp et al., 1998), although most recent
publications deal with biofilms on soft sediments (Serôdio et al., 2009;
Perkins et al., 2011). This chapter discusses results obtained from
experiments on hard substrates only.
Until now, phytobenthos research in freshwater systems has mostly been
focused on diatom biodiversity rather than the biomass of all algal groups.
However, since the implementation of the EU Water Framework Directive
(2000/60/EC; WFD), insight into the biomass of the total community of
benthic algae has become more important. Phytobenthos is considered to
be a suitable parameter to determine the impact of nutrient pollution
especially in rivers, since the organisms are sessile, take up nutrients
mostly from the water phase and therefore represent the trophic state at
the location sampled in the form of assimilated nutrients (i.e. nutrients
converted to biomass).
One of the important differences between fluorescence measurements on
phytobenthos and phytoplankton is the presence of a substratum. The
general set-up for phytobenthos fluorescence measurements is to expose
a small surface of a substratum to light of varying wavelengths and record
the chlorophyll-a fluorescence response of the benthic algae present.
However, when exposing the benthic algae to light, the substratum
beneath the algae will be inevitably included in the measurement. In such
a case, this substratum partially reflects the light coming from the
fluorometer and can thus affect the measurement results. Obviously, such
145
an effect does not occur when performing fluorescence measurements in
the water column. Therefore, a fluorescence measurement method used
for phytoplankton monitoring cannot simply be applied to phytobenthos,
but needs to be adapted to take into account the influence of the
substratum. If the biofilm is relatively thick, the algae themselves act as a
substratum, causing similar reflection effects. In these cases, the effect of
the underlying substratum may be negligible.
To quantify measurement errors due to substratum effects, some
researchers have performed measurements on the (sterilised) substratum
alone, without algae (Kromkamp et al., 1998; Serôdio, 2003). When
studying algae on soft substrata, it is possible to take a small sample of
the
sediment
to
the
laboratory,
sterilise
it
and
perform
blank
measurements. In the case of hard substrata such as stones, there is a
much higher heterogeneity in surface reflection. It is thus impractical to
take each sampled stone back to the laboratory for sterilisation and blank
measurements.
Figure 9.1 shows several different possibilities for the excitation light and
fluorescence to travel through the biofilm. When the light enters the
phytobenthic biofilm from the top, it is partially caught by the algae in its
path, resulting in a chlorophyll-a fluorescence response (Figure 9.1;
Arrows 2). Some of the excitation light is not absorbed by the microalgae
at all and is partially reflected (Figure 9.1, Arrows 1). This light does not
interfere with the measurement results, since all known fluorometers
contain an optical filter which prevents the unchanged reflected excitation
light from reaching the detector of the instrument. The remaining part of
the excitation light travels all the way through the biofilm, to be (partially)
reflected by the substratum. On its way back, another part of the reflected
excitation light is absorbed by the algae, resulting in an additional
fluorescence response (Figure 9.1, Arrow 3).
146
Figure 9.1
Schematic view of the general set-up to measure phytobenthos
biomass with a fluorometer (the phytobenthos layer has been
separated from the substratum layer for clarification purposes).
The wide vertical arrows indicate the excitation light beam coming
from the fluorometer. The light can either be reflected by the
biofilm or the substratum without altering it (indicated as the
dotted arrows of Option 1), or can be absorbed by the algae and
be partly be re-emitted as fluorescence light (wavy arrows of
Options 2 and 3). The incoming light can either be absorbed
directly by the algae (Option 2), or when the light re-enters the
algal mat from the bottom, after reflection by the substratum
(Option 3).
Depending on the colour and texture of the substratum (e.g. sand or
different types of stones), this additional fluorescence response is
hypothesised to result in a significant overestimation of phytobenthos
biomass.
In this study, four laboratory experiments were performed in order to
quantify the effect of reflection on the chlorophyll-a fluorescence response
of algae on different substrata, and to determine whether this effect can
be compensated for by using a substratum-specific correction procedure.
Several
different
types
of
hard
substratum
were
used
in
these
experiments. Soft sediments were not tested. In the first experiment, a
comparison of chlorophyll-a fluorescence measurements of a dye filter on
black and white artificial substrata was made in order to quantify the
147
influence of reflection on the result. Next, a similar experiment was
performed with the dye filter on natural substrata (stones) of different
colour and texture. In Experiment 3, the chlorophyll-a fluorescence of
natural algae on artificial and natural substrata was quantified and the
fluorescence response was corrected for substratum-specific reflection.
Experiment 4 was a repetition of the second experiment, but this time it
included the correction procedure of the third experiment to determine
whether the procedure as defined in Experiment 3 resulted in more
accurate
biomass
measurements.
Subsequently,
a
field
trial
was
conducted (Experiment 5) with natural assemblages of algae on natural
stony substrata, and the results were compared to conventional laboratory
analysis methods for chlorophyll-a.
This article describes the set-up and results of these experiments and
their implications for in situ chlorophyll-a fluorescence measurements of
phytobenthos biomass.
9.2.
Materials & Methods
9.2.1. Chlorophyll-a fluorescence measurements
Chlorophyll-a fluorescence measurements were performed using the
BenthoFluor and BenthoTorch fluorometers developed by bbe Moldaenke
GmbH
(Schwentinental,
Germany).
The
technical
details
of
this
instrument, including its calibration and application in the field, have been
described by Aberle et al., 2006. The BenthoFluor and BenthoTorch are
Pulse-Amplitude Modulated (PAM) fluorometers, as described in Section
3.2. The BenthoFluor emits light pulses of five different wavelengths (470,
525, 570, 590 and 610 nm), and the fluorescence response of the algal
assemblage is recorded at 690 nm.
An increase in the density of the phytobenthos biomass (i.e. a thicker
biofilm) causes a non-linear, self-shading effect on the fluorescence
response, as is shown in the example of the 470 nm LED in Figure 9.2.
This graph also shows the R2 values for linear regression and a second
order polynomial, showing a better result for the second order polynomial.
A third order polynomial does not improve the regression line further;
thus R2 increases from 0.9882 to 0.9888 (data not shown).
148
20
LED 470 nm
fluorescence (r.u.)
18
R² = 0,9176
linear regression
16
2nd order polynomial
14
R² = 0,9882
12
10
8
6
4
2
0
0
5
10
15
20
microalgal density (µg chlorophyll-a/cm²)
Figure 9.2
Example of increased self-shading effect with increased algal
density of the fluorescence signal after exposure to light with a
wavelength of 470 nm. The measurements were performed using
the BenthoFluor.
The BenthoFluor mathematically linearises the raw measurement data to
correct for this effect and to enable data analysis, thereby using Equation
9.1:
yi = y’i·(1+·ai·y’i)
(9.1)
where:
yi = measured fluorescence signal at wavelength i;
y’i = theoretical fluorescence signal at wavelength i, taking into account only the
reflection effect of the substratum;
ai = wavelength-dependent empirical factor to correct for self-shading effects.
This procedure is similar for both phytoplankton and phytobenthos
measurements. However, for micro–phytobenthos, an additional factor for
substratum reflection (factor ‘b’) needs to be added to the formula, in
order to correct for the substratum’s influence on the fluorescence signal
(Equation 9.2).
149
y’i·= b·xi·
(9.2)
where:
xi = theoretical fluorescence signal at wavelength i, not taking into account the
self-shading and reflection effects of the substratum (linear model);
b = factor expressing the reflection properties of the substratum.
Combining these two equations results in the following mathematical
approach, from which the corrected fluorescence signal is calculated:
yi = b·xi·(1+b·ai·xi)
(9.3)
Default sets of the factors a and b in Equation 9.3 were originally derived
empirically by bbe Moldaenke GmbH (Schwentinental, Germany) for the
following substrata:

sand/stone: for measurements on stony or sandy substratum;

GFC: for measurements on a GF/C filter (Whatman GF/C, 47 mm, Cat
No 1822-047);

black: for measurements on a black background.
Table 9.1 shows the self-shading (a) and background correction (b)
factors that the BenthoFluor generally uses as default values for its
mathematical conversion. To obtain these factors, only a very limited
number of stones were used to determine factor ‘b’ for the substrata
‘stone’. It was an arbitrary choice to normalise all other results to a stone
background, hence the background correction factor of ‘1’ for this type of
substratum.
After performing the first two experiments, the BenthoFluor was adjusted:
the 570nm and 590nm LEDs were removed and an extra LED, emitting
light at a wavelength of 700 nm, was included. The 700 nm LED was
added to allow for the correction of the substratum reflection. The
instrument was then re-designed and a prototype of this new instrument,
the BenthoTorch, was used to perform measurements in Experiments 3, 4
and 5.
150
Table 9.1
Default self-shading and background correction factors for three
substrata, as derived empirically by bbe Moldaenke.
calibration
profile
self-shading factors (ai)
background
per wavelength (nm)
correction
470
525
570
590
610
factor (b)
sand/stone
0.0012
0.0005
0.01
0.0032
0.0027
1
GF/C
0.0012
0.0005
0.01
0.0032
0.0027
0.75
black
0.0012
0.0005
0.01
0.0032
0.0027
2.1
9.2.2. Dye filter and algae
In
Experiments
1
and
2,
a
dye
filter
(bbe
Moldaenke
GmbH,
Schwentinental, Germany; product number F30174) was used instead of
algal cultures to allow the substrata to be exchanged easily. The dye filter
consisted of a combination of laser dyes which fluoresce within a
comparatively narrow spectral range (typically 30 nm); a mixture of dyes
was used in order to cover the same part of the (visible) spectral range as
benthic cyanobacteria. The mixture of dyes was embedded in synthetic
resin, which facilitated its use on different substrata. In this way, the
fluorescence response of benthic cyanobacteria could be imitated, with the
advantage of having a stable, unchanging signal over a long period of
time.
Experiment 3 was conducted using uni-algal cultures of Chlorella vulgaris
(green algae; strain number SAG 211-11b), Cyclotella meneghiniana
(diatoms; strain number SAG 1020-1a), Synechococcus leopoliensis
(cyanobacteria; strain number SAG 1402-1) and Microcystis aeruginosa
(cyanobacteria; strain number SAG 46.80). The cultures were grown using
cool fluorescent lamps (Philips TLD 58W/84 and TLD 58W/25). The
diatoms were grown at a light intensity of 82 µmol.m-2.s-1; the green
algae were grown at 190 µmol. m-2.s-1 and the cyanobacteria at 11 µmol.
m-2.s-1.
A
light:dark
cycle
was
not
applied.
All
light
intensity
measurements were performed with a Hansatech QRT1 Quantitherm
Lightmeter.
Solutions of five different densities of the three algal groups (green algae,
cyanobacteria and diatoms) were prepared for the measurements with
151
cultured algae. Small amounts of each culture were embedded in a layer
of agar (1.5 g agar per 100 ml water). In total, 5 suspensions of
increasing densities per algal group were each mixed with the agar
solution in equal amounts (50:50), and after the mixture had set, circular
slices of 1 cm2 and 1-2 mm thick were cut out. These slices were prepared
so that the substrata could be exchanged more easily and measurements
performed with the same algal densities on all substrata. The agar plates
with embedded algae were kept under humid and low-light conditions (3
µmol.m-2.s-1),
similar
to
the
circumstances
under
which
all
other
experiments were performed. A light:dark cycle was not applied. The
fluorescence signal of the algae in the agar was measured every five
minutes. After two hours, the signal had decreased by less than 5%. The
agar itself did not influence the fluorescence measurements, as was
concluded
from
measurements
performed
on
agar
plates
without
embedded algae. From this experiment, it was concluded that the duration
of the experiments performed with algae embedded in agar should not
exceed 2 hours, to avoid artefacts due to decreasing signal strength. All
experiments with real algae were performed well within this time limit,
and usually took no longer than 30 minutes. No dark adaptation period
was applied before any of the measurements.
9.2.3. Substrata
The order of magnitude of the substratum’s reflective properties was
determined using both artificial and natural substrata. As reflection is not
only determined by the substratum’s colour, but also by its texture, two
different types of artificial black substratum were used in Experiment 1. A
GFC filter was used as a white substratum. Most white surfaces, such as
paper, produce fluorescence and thus cannot be used in such an
experiment. All substrata were moistened with a few drops of water
before the measurements, in order to imitate field conditions as closely as
possible.
Before the experiments, the natural substrata were cleaned in an
ultrasonic bath to remove any biofilm. Next, measurements were
performed to confirm that no fluorescence could be detected after
cleaning.
152
The following materials were used as (artificial) substrata:
1. Black plastic discs (PVC, matt), used in the calibration procedures for
the BenthoFluor and BenthoTorch;
2. Black microfiber cloth, glossy;
3. White GF/C filters, used for filtering water samples and collecting
phytoplankton biomass;
4. Seven cobble stones of different shades and textures, ranging from
light to dark, see Table 9.2 for descriptions;
5. Sand (standard construction sand, sterilised and ultrasonically cleaned,
see Table 9.2 for description).
Table 9.2
Description of the stones and sand used as substrata for the
experiments. The colour of the substrate is indicated using the
Munsell soil-colour notation (Munsell, 2009).
stone* feature
1
type
metamorph
description
Gneiss, with layers of pink feldspar, biotite and quartz
colour (dry)
80% light red 7.5R 5.5/6; 20% light gray N 7.5
colour (moist) 90% light red 7.5R 6.5/6; 10% gray N 6
3
type
metamorph
description
schist, with small amounts of biotite, muscovite, quartz
and feldspar; layers less pronounced
colour (dry)
dark gray 7.5YR 4/1
colour (moist) dark brown 7.5YR 3/1.5
5
type
(meta-)granite
description
slightly metamorphous granite, with feldspar, muscovite
and quartz
colour (dry)
light brown 7.5 YR 6/3.5
colour (moist) 80% brown 7.5YR 4/2; 10% reddish yellow 7.5YR 7/8;
10% red 7.5R 5/7
6
type
(meta-)granite
description
heavily eroded, slightly metamorphous granite, with
feldspar, quartz, muscovite and biotite
colour (dry)
reddish brown 2.5YR 5/4
colour (moist) 80% red 2.5YR 4.5/6; 20% black 7.5YR 2.5/1
*: Where a sub-number has been assigned to a stone number, it denotes a
different position on the same stone.
153
Table 9.2 (continued)
stone* feature
7.1
type
metamorph
description
gneiss, with layers of pink feldspar, biotite and quartz,
more eroded (rougher) than Stone 1
colour (dry)
light reddish brown 2.5YR 6.5/4
colour (moist) 70% red 10R 4.5/6; 30% reddish gray 2.5YR 5/1
7.2
type
metamorph
description
gneiss, with layers of pink feldspar, biotite and quartz,
more eroded (rougher) than Stone 1
colour (dry)
80% reddish brown 2.5YR 5.5/4; 20% grayish brown
2.5Y 5/2
colour (moist) 85% red 10R 5/6; 15% dark grayish brown 2.5Y 3.5/2
8.3
type
granite
description
igneous
rock,
fine
crystalline
plutonic
rock
with
hornblende, feldspar; possibly diorite (calcium-richer
than granite)
colour (dry)
dark grayish brown 2.5Y 3.5/1.5
colour (moist) black 2.5Y 2.5/1
8.4
type
granite
description
igneous
rock,
fine
crystalline
plutonic
rock
with
hornblende, feldspar; possibly diorite (calcium-richer
than granite)
colour (dry)
very dark grayish brown 2.5Y 3/1.5
colour (moist) black 2.5Y 2.5/1
11
type
chert
description
fine crystalline quartz; chalcedony
colour (dry)
white N 8
colour (moist) light gray N 7.5
-
type
sand
description
rich in quartz with some feldspar; variable grain size
colour (dry)
pale brown 10YR 6.5/3
colour (moist) brown 10YR 4/3
*: Where a sub-number has been assigned to a stone number, it denotes a
different position on the same stone.
9.2.4. Field trial
For the field trial (Experiment 5), five cobble stones (approximately 10-15
cm in diameter) were collected from the Odense River at the Skallebanke
154
Wetland on Fyn, Denmark (latitude N 55° 16.176′; longitude E 10°
13.643′). First, fluorescence measurements were performed in situ, using
the bbe BenthoTorch (see description above).
Next, the stones were brushed with a clean toothbrush and/or scraped
with a scalpel, to remove as much of the biofilm as possible, following the
sampling procedure for hard substrata as described in CEN Guidance
Standard EN 15708. The size of the brushed surface area was carefully
measured for biomass calculation purposes. After brushing, another
fluorescence measurement was performed to check if any of the biofilm
had been left behind on the stone’s surface. The samples were collected in
small vials and taken to the laboratory for analysis. Total chlorophyll-a
analyses were performed according to two different methods: the German
standard DIN 38412-16 (DIN, 1985), which is almost identifical to ISO
Standard 10260 (ISO, 1992) and the method of the HELCOM-CombineManual, Annex C-4 (HELCOM, 2008). The difference between these
methods is in the extraction: The DIN and ISO standards describe a hot
extraction
with
90%
ethanol,
which
enables
the
photometric
determination of chlorophyll-a and of phaeopigments before and after
acidification
with
hydrochloric
acid.
The
HELCOM-Combine-Manual
describes a cold extraction method using 96% ethanol. This method
allows for the determination of phaeopigments only by fluorometric
evaluations. The results of the in situ fluorescence measurements for total
chlorophyll-a were compared to the results of the laboratory analyses.
9.3.
Results
9.3.1. Experiment 1: Dye filter measurements on a black and white
background
The dye filter was measured on all three artificial substrata (Table 9.3).
The results show that the reflection of the white substratum was much
higher than the black substrata, leading to an overestimation of the ‘algal
biomass’ by a factor of almost 2. The black synthetic microfibre cloth
reflected more light than the black PVC, increasing the fluorescence signal
by more than 20%. Both the colour and the texture of the substratum
influenced
the
measurement
results
considerably,
thus
making
it
necessary to compensate for this effect in order to determine algal
biomass more reliably.
155
Table 9.3
Results of the dye filter measurements on three different
substrata (n=10). The results were calculated using the
background correction factor b of Equation 3 for the black and
GF/C calibration profiles (see also Table 9.1).
measurement result
background correction
(µg.cm-2)
factor (b)
10.47 ± 0.021*
2.1
black microfibre cloth
12.80 ± 0.037
2.1
white GF/C filter
19.03 ± 0.025
0.75
substratum
black PVC
*: 95% confidence interval
9.3.2. Experiment 2: Dye filter measurements on natural stones
In the second experiment, natural stones were used as substrata (Table
9.4). The dye filter of Experiment 1 was measured against two different
backgrounds: a (light) reddish brown gneiss (Stone 7 in Table 9.2) and
(very) dark grayish brown igneous rock (Stone 8 in Table 9.2).
Table 9.4
Results of the dye filter measurements on two natural stones
(n=10). The results were calculated using the background
correction factor b of Equation 3 for the black and the stone/sand
calibration profile (see also Table 9.1). The percentages in
parentheses indicate the differences between the measurements
made on the stones and the black PVC.
measurement result (µg.cm-2)
substratum
bblack = 2.1
bstone = 1
black PVC
10.47± 0.021*
Stone 7 P1
19.29 ± 0.049 (+84 %)
7.38 ± 0.017 (-29 %)
Stone 7 P2
17.10 ± 0.020 (+63 %)
6.61 ± 0.008 (-37 %)
Stone 8 P3
11.89 ± 0.026 (+13.5 %)
4.78 ± 0.011 (-54 %)
Stone 8 P4
11.23 ± 0.033 (+7.3 %)
4.57 ± 0.013 (-56 %)
*: 95% confidence interval
When using the calibration profile for a black background (see Table 9.1),
the results of Stone 8 (dark) were much closer to the black value of 10.47
µg.cm-2 than Stone 7 (light) (see Tables 9.3 and 9.4). However, when the
calibration profile for sand/stone (see Table 9.1) was used to calculate the
results of the dye filter, the outcome was an underestimation of ‘algal
biomass’ by over 50% on Stone 8 (Table 9.4). In this case, the results of
156
the light-coloured Stone 7 were closer to the values measured on black
PVC than the results of Stone 8. All results were compared to the value
measured on black PVC, as this background is used during calibration
procedures, and therefore represents the correct value.
9.3.3. Experiment 3: Algae measurements on natural and black
substrata
The spectral fingerprints (see Figure 3.3) of three algal cultures (see
Section 9.2.2) were measured using the prototype of the BenthoTorch
against a black PVC background. Figure 9.3 shows the non-linear effects
with increased algal density, as a result of self-shading. The spectral
fingerprints of the three algal groups were calculated from the results in
fluorescence (r.u. x 1000)
Figure 9.3, resulting in Figure 9.4.
900
470nm LED
800
610nm LED
700
525nm LED
600
500
400
300
200
100
0
0
1
2
3
4
5
6
7
8
9
10
11
algal density (µg/cm²)
A.
157
900
470nm LED
800
610nm LED
700
525nm LED
600
500
400
300
200
100
0
0
1
2
3
4
5
6
7
8
9
10
11
8
9
10
11
B.
900
470nm LED
800
610nm LED
700
525nm LED
600
500
400
300
200
100
0
0
1
2
3
4
5
6
7
C.
Figure 9.3
158
Fluorescence signal of green algae (A), cyanobacteria (B) and
diatoms (C) at a density range of 0-10 µg.cm-2. The data were
fitted to a second order polynomial curve. R2 values ranged from
0.9966 to 0.9993.
120
cyanobacteria
green algae
100
signal (r.u. x 1,000)
diatoms
80
60
40
20
0
450
500
550
600
650
700
750
wavelength (nm)
Figure 9.4
Spectral fingerprints of the algal cultures of green algae,
cyanobacteria and diatoms against a black background, as
measured by the BenthoTorch prototype.
Figure 9.5 shows the measurements of the 700 nm signal (without algae)
for ten substrata: black PVC, white GF/C filter, sand and seven cobble
stones, all with different reflective properties due to their natural shade
and texture. The results of the 700 nm measurements were different for
each substratum (Figure 9.5), thus illustrating the need for a substratumspecific correction factor. Sand grains can act as small mirrors reflecting
the light to a larger extent than stones. This explains why the reflection
signal of sand is higher than that of most of the stones used in the
experiment, although some of them had a similar or even lighter colour
than the sand.
Next, the fluorescence signal of the cultures of green algae (Figure 9.6),
cyanobacteria and diatoms was determined on the substrata under
investigation. The difference in signal strength and curve was caused by
the reflective properties of the substratum, thus making this signal
suitable for use in calculating a substratum-specific reflection factor.
159
Figure 9.5
700nm signals for ten different moistened substrata (2 stones
were measured on opposite sides), arranged from low to high
reflective properties.
Figure 9.6
Fluorescence signal after excitation at 470 nm of green algae on
seven substrata. The graphs for cyanobacteria and diatoms (not
shown) were very similar.
160
A.
B.
161
C.
Figure 9.7
Relationship between the 700 nm signal and the reflection factor
at 470 nm (A), 610 nm (B) and 525 nm (C) for green algae,
cyanobacteria and diatoms at various densities.
Figure 9.7 shows the relationship of the signal at 700 nm and the
reflection factor for all three algal groups and at various densities (0-10
µg.cm-1). The significance of the regression lines was calculated according
to Fowler et al. (1998) and all three relationships were found to be highly
significant (t470 = 34.582, t525 = 39.728, t610 = 24.304; df = 120; all tvalues exceeded 2.617 at P=0.01). The graphs show the relationship
between the 700 nm signal and the reflection factor for all three algal
groups and at various densities as an exponential function:
y’i = xi ∙ e(c ∙ z)
(9.4)
where:
z = 700 nm signal;
c = correlation factor.
Comparing Equation 2 and Equation 9.4, it was concluded that the
background factor ‘b’ is not a fixed constant, but a function of the 700 nm
signal:
b = e(c ∙ z)
162
(9.5)
Table 9.5 shows the correlation factor ‘c’ for each individual algal group
and for all algal groups together, at 470 nm, 525 nm and 610 nm. Thus, it
was possible to calculate the density of algae on the substratum by
measuring the signals at 470 nm, 525 nm, 610 nm and 700 nm, including
the correction for the substratum-specific reflection.
Table 9.5
Correlation factor ‘c’ at 470 nm, 525 nm and 610 nm for green
algae, cyanobacteria and diatoms and for all algal groups
combined. The coefficient of determination R2 indicates the
proportion of variability in the dataset for the calculation of ‘c’.
λ (nm)
correlation factor c
R2
all algal groups
470
3,88∙10-6
0,930
610
4,19∙10
-6
0,952
4,06∙10
-6
0,934
525
green algae
470
3,97∙10-6
0,936
610
4,22∙10
-6
0,959
525
4,21∙10-6
0,952
cyanobacteria
470
3,83∙10-6
0,975
610
4,35∙10-6
0,878
525
-6
0,964
3,95∙10
diatoms
470
3,68∙10-6
0,907
610
4,14∙10
-6
0,939
3,91∙10
-6
0,929
525
9.3.4. Experiment 4: Dye filter measurements on natural substrata
revisited
Experiment 4 was a repetition of the second experiment with the dye filter
on Stones 7 and 8. However, this time the 700 nm signals measured by
the BenthoTorch in Experiment 3 and Equation 5 were used to calculate
the substratum-dependent reflection factors. The recalculated results are
shown in Table 9.6.
163
Table 9.6
substratum
black PVC
Stone 7, P1
Stone 7, P2
Stone 8, P3
Stone 8, P4
Comparison of the dye filter measurements on two natural stones
(n=10), as originally measured with the BenthoFluor (cf. Table
9.2), and after correction using the 700 nm value and Equation 5
(Experiment 3). The percentages in parentheses indicate the
differences between the measurements made on the stones and
the black PVC.
measurement result (µg.cm-2)
bblack = 2.1
bstone = 1
b = e(c ∙ z)
10.47± 0.0212*
19.29 ± 0.0485
7.38 ± 0.0165
9.90 ± 0.0248
(+84 %)
(-29 %)
(-5.5%)
17.10 ± 0.0197
6.61 ± 0.0081
10.02 ± 0.0029
(+63 %)
(-37 %)
(-4.3%)
11.89 ± 0.0259 (+13.5
4.78 ± 0.0108
11.28 ± 0.0212
%)
(-54 %)
(+7.7%)
11.23 ± 0.0327
4.57 ± 0.0126
10.96 ± 0.0107
(+7.3 %)
(-56 %)
(+4.7%)
*: size of 95% confidence interval
The results after correction using the 700 nm value (Column ‘b = e (c ∙ z)’ in
Table 9.6) were much closer to each other and to the original value
measured on a black background than the results calculated with the
correction factors for a stone and black background (Figure 9.9A). Figure
9.9B shows the difference of the means for all results (all four samples
together) using the background correction factors for stone and black
background, compared to that of the corrected results using Equation 5.
Figure 9.8 shows that the function describing factor b corrects for both
over- and underestimation of the effects caused by the substratum.
164
A.
difference (µg/cm²)
6
5
4
3
2
1
0
b-stone
b-black
c
B.
Figure 9.8
Mean results per sample for each calibration procedure (A) and
difference of the means of all results (B) as shown in Table 9.4.
The error bars (B) indicate the 95% confidence interval of the
means of the difference between the datasets (‘c’ indicates the
results when caclulating factor b as e(c ∙ z)).
165
9.3.5. Experiment 5: field trial of natural algae on natural
substrata
The results of the in situ fluorescence measurements on five cobble stones
from the River Odense in Denmark before and after sampling are shown in
Table 9.7. It became clear from the fluorescence measurements after
sampling that not all algae had been removed from the surfaces of the
stones. The sampling method as described in CEN Guidance Standard EN
15708 was originally designed for qualitative analysis of benthic diatoms,
and is not representative enough for quantitative sampling.
Table 9.7
In situ measurements of phytobenthic biomass as total
chlorophyll-a on eight cobble stones from River Odense
(Denmark) before and after sampling, including the surface of the
sampled area on each stone.
biomass as total chlorophyll-a (µg.cm-2)
surface area
before sampling
after sampling
sampled (cm2)
1
1.08
0.84
30
2
0.50
0.32
64
3
0.43
0.17
48
4
0.54
0.17
25
5
0.88
0.10
20
6
3.24
0.84
20
7
1.28
0.31
15
8
0.95
0.89
10
stone
However, the samples taken from the substrata contained on average
38% less chlorophyll-a than the amounts detected by the in situ
fluorescence method before sampling (see Table 9.7). Table 9.8 shows the
results of the chlorophyll-a analysis of the samples according to the DIN
and HELCOM methods. The results were recalculated as µg.cm-2 based on
the surface area sampled. These results show that it is virtually impossible
to remove all phytobenthic biomass from a natural surface, thus
illustrating the problem with the standard sampling methods currently in
use. When the amounts of chlorophyll-a left behind on the stones (as
determined by in situ fluorescence measurements after sampling) were
subtracted from the original results of the fluorescence measurements (3rd
column in Table 9.8), the fluorescence results of the in situ measurements
166
were generally closer to the analytical results found. However, it is unclear
what caused the discrepancy between the different methods for Stones 1
and 8.
Table 9.8
Results of the analysis of chlorophyll-a concentrations according
to DIN 38412 (1985) and HELCOM (2008) compared to results of
the fluorescence measurements for the part of the algae
effectively removed from the stone.
biomass as total chlorophyll-a (µg.cm-2)
in situ fluorescence
stone
DIN
HELCOM
1
0.65
0.69
0.24
2
0.22
0.24
0.18
3
0.25
0.23
0.26
4
0.36
0.36
0.37
5
0.68
0.69
0.78
6
-
2.17
2.40
7
-
1.02
0.97
8
-
0.31
0.06
0.43
0.44
0.37
average
(before – after)
Figure 9.9 shows the correlation between the HELCOM and DIN methods
on the one hand, and the BenthoTorch measurements on the other. In
order to exclude the sampling error of the HELCOM/DIN methods, only the
corrected in situ measurements (before – after; see Table 9.8) were used.
A good correlation was found (R2 = 0.9208).
167
chlorophyll-a [µg/cm²]
3,0
2,5
R² = 0,9208
2,0
1,5
1,0
0,5
0,0
0,0
Figure 9.9
9.4.
0,5
1,0
1,5
2,0
chlorophyll-a [µg/cm²], HELCOM/DIN
2,5
Correlation between concentrations of chlorophyll-a measured
according to HELCOM (triangles) and DIN (squares) and by using
the corrected BenthoTorch (BT) results as shown in Table 9.8.
Discussion
The results of Experiment 2 show that algal biomass cannot be calculated
reliably from fluorescence measurements using a constant background
correction factor for all types of stones. In contrast to the original
background correction factor as determined by the manufacturer (Table
9.1), this reflection factor is not the same for all stony substrata. A
substratum-specific factor was needed for each type of stony substratum
in order to determine algal biomass accurately. However, for practical
reasons, this factor should preferably be determined while performing the
measurement of the phytobenthos, so as not to introduce an additional
calibration procedure before each measurement. Also, the determination
of the substratum’s reflective properties in the field should not be
hampered by the presence of algae.
At 700 nm, the fluorescence signal of all three algal groups is very small
(see Figure 9.4). The signal at 700 nm was therefore selected to
determine the reflective properties of the substratum, as any interference
from algae at this wavelength is negligible. By measuring the reflection
factor
168
simultaneously
during
each
fluorescence
measurement,
a
substratum-specific correction factor is determined which instantaneously
corrects the measurement result, without interference from the algae on
the substratum. The black PVC background generally used in calibration
procedures for the fluorometer was chosen as the standard background to
which all signals were normalised.
The quantification of reflection effects of stony substrata as described in
Section 9.3 (Results) is a distinctly different procedure from performing an
auto-zero measurement to determine the offset of a fluorometer. An
offset signal is created when a part of the light pulse emitted by the
fluorometer itself is reflected by the substratum and reaches the detector
unchanged. Fluorescence detectors generally use an optical filter at
approximately 690 nm. However, in reality, the optical filter allows light of
slightly lower and higher wavelengths to reach the detector as well. Other
PAM fluorometers, such as the Phyto-PAM Fluorometer (Walz, Effeltrich,
Germany) use a maximum excitation wavelength of 665 nm, which causes
a slight offset, since this wavelength is close enough to the detector’s
wavelength to pass through the optical filter. This offset can be quantified
by performing an auto-zero measurement on sterile sediment or in the air
(personal communication J. Kromkamp, August 2011). The BenthoFluor
uses a maximum excitation wavelength of 610 nm, which does not reach
the detector and therefore does not create an offset signal. If the offset
were to be included in the fluorescence calculations, Equation 3 would
have to be re-written as follows:
yi = b·(xi·- offset) (1+b·ai·(xi – offset))
(9.6)
The reflection factor b, now replaced by the function as described in
Equation 9.5, is the result of the presence of the substratum. However, it
does not correct for the offset, but for the extra fluorescence signal
caused by the excitation light on its way back through the biofilm (see
also Figure 9.1). No references have been found in the literature to
indicate the use of any correction factor other than the auto-zero
measurement for PAM fluorometers.
The in situ fluorescence method described above has several advantages
and disadvantages. On the one hand, the field measurements are
169
relatively easy to perform and problems with the sampling of stony
substrata can be avoided. As compared to total chlorophyll-a analysis, the
fluorescence
method
provides
more
detailed
information
on
the
contribution of different algal groups with regard to total phytobenthic
biomass. On the other hand, for very thick biofilms (> 20 µg/cm2) the
quantification of biomass per algal group becomes more difficult, and the
self-shading effect can no longer be corrected with a simple mathematical
correction. The light from the fluorometer may not reach the lowest layers
of the biofilm, thus total algal biomass will be underestimated. The
presence of macroalgae may also cause an error in the results. Therefore,
measurements on stones covered in macroalgae should be avoided. The
method cannot be used for epiphytic algae, as the chlorophyll-containing
macrophyte-substratum may interfere with the measurement results.
In some cases, the application of a dark-adaptation period prior to the
fluorescence measurements may be necessary (Perkins et al., 2011;
Schreiber, 2004). On soft sediments, the use of relatively simple aids
makes it possible to apply a dark-adaptation period or a period of low light
(Jesus et al., 2006). On stony substrata, this is more complicated, since
the surface of the stones is rarely flat, and attaching a box or cover to the
stones cannot be performed easily. A tool which allows for the application
of a dark or low-light adaptation period for stony substrata has not yet
been developed. Therefore, measurements performed on stony substrata
under high-light conditions should be interpreted with care. However,
under normal conditions and if sampling locations are carefully selected,
no problems with the interpretation of the results are expected.
Experiment 5 shows that it is difficult to compare the in situ fluorescence
method to conventional chlorophyll-a analysis methods, as the sampling
procedure introduces a considerable error in the results. Several efforts
were made to develop a sufficiently representative sampling method for
micoalgae on hard substrata (Loeb, 1981; Davies and Gee, 1993).
However, no suitable quantitative sampling method has yet been
developed. The problems with the sampling methods as described by
Davies and Gee (1993) suggest that in situ fluorescence measurements
are a better alternative to the laboratory analysis of biomass. Additionally,
one should keep in mind that the in situ fluorescence method measures
170
active chlorophyll-a from live cells only, whereas the ISO/DIN and
HELCOM
methods
determine
total
chlorophyll-a.
This
may
cause
differences in the results.
9.5.
Implications
The results have implications for phytobenthos biomass measurements in
the field. Especially when performing measurements on highly reflective
substrata, it should be noted that the reflected light from the substratum
will excite the algal pigments on its way back through the biofilm and thus
cause
an
increase
in
the
fluorescence
signal,
resulting
in
an
overestimation of algal biomass. In contrast, if the substratum is less
reflective than the substratum used in the calibration procedures, the
algal biomass will be underestimated.
For fluorometers which do not have a built-in 700 nm LED to correct for
the reflection of the substratum, this procedure should be performed
manually. For soft sediments with a uniform colour and texture in a
certain area of investigation, the results may not even differ considerably,
and taking a representative sample to the laboratory to determine the 700
nm signal could be a relatively simple procedure. However, for hard
substrata (stones), determining the 700 nm signal could be a problem,
unless it can be performed directly in the field. If this is impossible, a
limited number of specific natural stones, whose reflective properties are
known, could be used in the same way as artificial substrata.
The substratum-specific correction factor makes it possible to compare
datasets over time, and eliminates the need to use artificial substrata
when suitable natural substrata are available. Furthermore, the results of
measurements on different substrata with varying reflective properties, for
example sandy and muddy substrata, or stony substrata and concrete
bridge pillars, can be compared more easily.
9.6.
Conclusions
In situ biomass measurements using a fluorometer are recommended for
the collection of relevant information about benthic algal communities at
algal group level. The method provides accurate biomass information that
cannot be extracted from the commonly used qualitative analysis of
species
composition
and
abundance.
The
correction
procedure
for
171
reflection effects improves the reliability of the results significantly,
without increasing the time necessary to perform the measurements or
complicating the measurement procedure. In the case of extremely thick
biofilms or very high light conditions, interpreting the results should be
carried out with caution.
172
Chapter 10. JDS2 field example
10.
Joint Danube Survey: a field example
10.1. Introduction
The Second Joint Danube Survey (JDS2), organised by the International
Commission for the Protection of the Danube River (ICPDR) from 13
August to 26 September 2007, was - according to our best knowledge the first survey ever where phytobenthos biomass was analysed on a
large scale in combination with many other parameters, such as nutrients
(N and P), phytoplankton (competition for light) and benthic invertebrates
(grazing).
The aim of this chapter is not to quantify the relationships between these
parameters, but to illustrate how a comparison of relevant parameters can
provide insight into the (ecological) development of the river system and
how phytobenthos biomass fits into this assessment.
The JDSs Final Scientific Report (Liška et al., 2008) describes in detail how
and where samples were taken and which parameters were analysed with
which methods. The parameters assessed in this chapter include:
phytobenthos
biomass,
flow
velocity,
sediment
composition,
water
transparency, bank shading, re-colonisation potential from tributaries,
nutrients (N and P), phytoplankton densities, benthic invertebrate grazer
densities and benthic diatoms index.
10.2. The Danube River
The Danube River is 2,857 km long and flows through 13 countries
(ICPDR, 2002). Altogether, the entire Danube river basin is spread over
17 countries. In order to be able to interpret the collected data, the
Danube
River
geographical
was
subdivided
sections,
into
according
several
to
different
geological
‘homogenous’
structure,
slope,
geomorphology and hydrology. Generally, two division methods were used
to
analyse
the
phytobenthos
data
collected
during
the
JDS
(Sommerhäuser et al., 2003):
173
A. The Upper, Middle and Lower Danube concept, splitting the Danube
River into three parts (Table 10.1):
 The Upper Danube Basin covers the area from the source tributaries
in the Black Forest down to the Devin Gate east of Vienna. Major
tributaries are the Lech, Isar, Inn, Traun, Enns from the right2 and
Morava from the left.
 The Middle Danube Basin comprises the largest portion of the
catchment and covers the part from Devin Gate to the Iron Gate
dams (Serbia Montenegro/Romania). Major tributaries in this region
are (from the left) Váh, Hron, Ipel and Tisa, and from the right
Raba, Sio, Drava, Sava and Velika Morava.
 The Lower Danube Basin is formed by the Romanian-Bulgarian
lowland and its upland plateaus and mountains. The important
tributaries in this region are Timok, Iskar and Jantra from the right,
and Jiu, Olt, Arges, Ialomita, Siret and Prut from the left.
B. A more elaborate division into 10 section types, validated using
benthic invertebrate datasets (Table 10.2).
Table 10.1
JDS2 sampling point division for the Upper, Middle and Lower
Danube (Sommerhäuser et al., 2003).
section
sampling points
Upper Danube
JDS1 - JDS16
Middle Danube
JDS17 - JDS66
Lower Danube
JDS67 – JDS95
Phytobenthos biomass measurements could not be performed at the
sampling points in Sections 9 and 10, since the sediment was too fine and
soft to allow for adequate sampling or field measurements.
2
The banks are defined as left and right when facing the river in downstream
direction.
174
Table 10.2
Danube
JDS2 sampling point division over 10 sections according to
geographical region (Sommerhäuser et al., 2003).
Danube section type
stream kilometre
JDS2 sampling
section
points
1
Upper Course of the Danube
2786 - 2581
1
2
Western Alpine Foothills
2581 - 2225
2-6
2225 - 2001
7-10
2001 - 1791/1790
11-19
Danube
3
Eastern Alpine Foothills
Danube
4
Lower Alpine Foothills
Danube
5
Hungarian Danube Bend
1791/1790 - 1497
20-37
6
Pannonian Plain Danube
1497 - 1071
38-59
7
Iron Gate Danube
1071 - 931
60-62
8
Western Pontic Danube
931 - 378
63-86
9
Eastern Wallachian Danube
378-100
87-92
10
Danube Delta
100-0
93-96
10.3. Phytobenthos biomass and abiotic factors
The flow velocity was estimated (recorded as slow, medium or fast) at
each sampling location and the composition of the substratum and
transparency of the water column were recorded. The average results for
these parameters for the Upper, Middle and Lower Danube are presented
in Table 10.3.
Table 10.3
Average results
sampling.
for
some
sediment composition
section
Upper
abiotic
factors
recorded
flow
during
transparency
%
%
% fine
slow medium fast low medium high
stone
sand
sediment
84
15
1
10
16
74
15
32
53
40
30
30
71
17
12
26
39
35
21
27
52
45
55
0
45
55
0
Danube
Middle
Danube
Lower
Danube
175
As expected, the upstream section of the Danube River is characterised by
generally stony substrata and a fast-flowing river with high transparency
(low turbidity). The Middle Danube shows a more diverse substratum, a
decreased flow velocity and lower transparency. The Lower Danube is
characterised by more fine sediment, a slow to medium flow and generally
low transperency (high turbidity).
Despite these differences, phytobenthic biomass did not show any
significant differences between the three sections, neither as total
biomass, nor per algal group (Figure 10.1 A, B, C, D).
A.
B.
C.
D.
Figure 10.1
176
Average total phytobenthic (A), green algal (B), cyanobacterial
(C) and diatom (D) biomass for the Upper (n=9), Middle (n=44)
and Lower Danube (n=13).
When phytobenthic biomass was individually compared to flow velocity,
transparency and substratum composition, no significant differences were
observed (Figure 10.2, 10.3 and 10.4). Regarding the composition of the
substratum, there seems to be a slight increase in diatoms on stony
substrata as compared to sand and fine sediments, however, this is not
significant. It could be the result of a sampling artefact, since diatoms
may migrate more easily through sand and fine sediment, thus avoiding
sampling on these substrata.
Figure 10.2
Average biomass per algal group of
medium (n=21) and slow (n=58)
The error bars indicate the 95%
green algae; cyano: cyanobacteria;
samples taken at fast (n=23),
flow conditions of the water.
confidence intervals (green:
diatom: diatoms).
177
Figure 10.3
Average biomass per algal group of samples taken at high (ht;
n=35), medium (mt; n=40) and low (lt; n=27) transparency of
the water. The error bars indicate the 95% confidence intervals
(green: green algae; cyano: cyanobacteria; diatom: diatoms).
Figure 10.4
Average biomass per algal group of samples taken from substrata
consisting of more than 60% cobbles (stone; n=40), more than
55% sand (sand; n=18) and of more than 50% fine sediments
such as silt and mud (fine; n=28). The error bars indicate the
95% confidence intervals (green: green algae; cyano:
cyanobacteria; diatom: diatoms).
178
Besides these three abiotic factors, the percentage shading of the river
banks was recorded where samples were taken. When comparing the
average biomass per algal group as determined on substrata which were
less or more than 50% shaded, no significant differences were observed
(Figure 10.5).
Figure 10.5
Average biomass per algal group of samples taken at less than
50% shading (<50%; n=84) and more tha 50% shading (>50%;
n=18). The error bars indicate the 95% confidence intervals.
From these
observations
it was
concluded
that
these
factors
by
themselves do not have a significant influence on the development of
phytobenthic biomass. Flow velocity could influence phytobentjos biomass
development (as shown by Biggs, 1996 and Biggs et al., 1998a), however,
in order to determine any such relationship in the Danube River, flow
velocity would have to be measured more accurately, rather than
estimating which of three categories applies. An additional important
factor for phytobenthos biomass is the time between spates (storms or
floods), as shown by Biggs et al. (1999).
179
10.4. Phytobenthos biomass and (re-)colonisation from
tributaries
The Danube River is joined by 18 tributaries on its 2,500 km stretch from
the source to the Black Sea (Table 10.4). These tributaries are a starting
point for phytobenthos re-colonisation after scouring flood incidents in the
Danube River. During the Second Joint Danube Survey, phytobenthos field
measurements were performed not only in the Danube River but also in
many of these tributaries. Thus, it was possible to make an evaluation of
the phytobenthos biofilm composition at the confluence of the tributaries
(slightly upstream of the tributary itself), and compare it with the results
of sampling points both upstream and downstream of the confluence and
on the left and right banks of the river. As an example, the results for the
Inn River are shown below. The Inn River joins the Danube River on the
right bank near the German/Austrian border. Figure 10.6 shows the total
biomass per algal group and composition of the Inn tributary and the
upstream and downstream sampling points in the Danube River.
Upstream of the confluence of the Inn River, green algae comprise only a
small part of the phytobenthic community. Further downstream the green
algae form more than half of the total phytobenthic biomass. This may be
the result of colonisation of green algae originating from the Inn River,
although no definite conclusions can be drawn on the basis of this onetime result.
180
JDS81
JDS84
RS
RS
RS
RS/BG
BG
RO
BG
BG
RO
Tisa
Sava
Velika
Morava
Timok
Iskâr
Olt
Jantra
Russenski
Lom
Arges
RO
RO/MD
JDS66
SK
SK
SK/HU
HU
HR/RS
Váh
Hron
Ipoly
Sió
Drava
Siret
Prut
JDS56
DE/AT
SK
Inn
Morava
JDS90
JDS91
JDS71
JDS74
JDS78
JDS51
JDS49
JDS21
JDS24
JDS25
JDS37
JDS42
JDS6
JDS15
country
1.0
1.0
?
?
0.3
0.4
1.0
0.2
?
7
1.0
0.8
0.5
0.7
1.0
1.4
4.3
0.1
tributary
sample
distance to
code
confluence
(km)
Left
Left
Left
Right
Right
Left
Right
Right
Right
Right
Left
Left
Left
Left
Right
Right
Danube
bank side
of
confluence
Right
Left
Braila
Braila
Upstream Arges
Upstream Timok
(Rudujevac/Gruia)
Upstream Iskâr (Bajkal)
Upstream Olt
Downstream
Zimnicea/Svishtov
Upstream Ruse
Upstream Tisa (Stari
Slankamen)
Downstream Tisa/
Upstream Sava (Belegis)
Upstream Velika Morava
JDS89
JDS89
JDS83
JDS80
JDS70
JDS73
JDS77
JDS65
JDS55
JDS50
JDS48
JDS20
JDS23
JDS23
JDS36
JDS41
JDS5
JDS14
13
32
2
2
3
1
13
4
4
15
0
2
3
11
36
5
53
1
previous sampling point JDS
sample distance to
code
tributary
(km)
Niederalteich
Upstram Morava
(Hainburg)
Komrano/Komarom
Sturovo/Esztergom
Sturovo/Esztergom
Paks
Upstream Drava
name
Danube and its tributaries sampled during JDS2.
name
Table 10.4
Downstream
Ruse/Giurgiu
Downstream
Arges/Oltenita
Reni
Reni
Iza/Szony
Szob
Szob
Baja
Downstream Drava
(Erdut/Bogojevo)
Downstream Tisa/
Upstream Sava (Belegis)
Upstream Pancevo /
Downstream Sava
Downstream Velika
Morava
Pristol/Novo Selo
Harbour
Downstream Iskâr
Downstream Olt
Downstream Jantra
JDS92
JDS92
JDS85
JDS82
JDS72
JDS75
JDS79
JDS67
JDS57
JDS52
JDS50
JDS22
JDS26
JDS26
JDS38
JDS43
JDS7
JDS16
24
5
3
10
8
3
22
11
6
11
15
5
9
1
16
12
20
11
following sampling point jds
code
distance from
tributary
(km)
Jochenstein
Bratislava
name
181
A.
Microphytobenthic
chlorophyll-a (µg/cm²)
1,4
1,2
1,0
0,8
0,6
0,4
0,2
0,0
5l
5r
6
7l
7r
Sampling location
B.
Figure 10.6
Composition of phytobenthic biofilm at the confluence of
the Inn River (sample location JDS6) in comparison to the
upstream point JDS5 and downstream point JDS7 (A) and total
phytobenthic biomass per algal group at each sampling point
(B). green: green algae; cyano: cyanobacteria; diatom:
diatoms).
182
10.5. Phytobenthos biomass and nutrient levels
The comparison of phytobenthos biomass and total phosphorous/total
nitrogen is shown in Figure 10.7 A and B. The data were labelled
according to the sampling location (Upper, Middle or Lower Danube). No
significant relationship between these nutrients and phytobenthos biomass
was found.
A.
Figure 10.7
B.
Comparison of phytobenthic biomass and ortho-phosphate (A)
and total nitrogen (B) in the river water (•: Upper Danube; •:
Middle Danube; •: Lower Danube).
10.6. Phytobenthos biomass and phytoplankton density as total
chlorophyll-a
Figure 10.8 shows the comparison of phytoplankton and phytobenthos
biomass in relation to orthophosphate levels in the Danube River water.
Phytobenthos biomass could only reach higher densities (>1 µg/cm) when
phytoplankton
biomass
remained
low
(generally
<10
µg/L).
The
competition for light is most likely the cause of this pattern. Higher orthophosphate concentrations in the water do not seem to affect phytobenthos
biomass development directly, or favour either planktonic or benthic
algae.
183
Figure 10.8
Comparison of phytoplankton densities (expressed as total
chlorophyll-a per litre) and average phytobenthos biomass in the
Danube River (n=66). Ortho-phosphate labels: •: 0.000-0.024
mg P/L; •: 0.025-0.039 mg P/L;•: 0.040-0.049 mg P/L; •: 0.0500.099 mg P/L; •: ≥ 0.100 mg P/L.
10.7. Phytobenthos biomass and benthic invertebrates (grazing)
Benthic invertebrates were sampled and analysed at all sampling locations
in the Danube River. Species of the Theodoxus and Viviparus genera were
chosen as indicators for grazing pressure for two reasons:
1. these genera are known grazers;
2. many individuals of these two genera were found at almost all JDS2
sampling points.
As the Theodoxus and Viviparus genera are not the only grazers in the
Danube River, a more elaborate investigation would need to be performed
to include all grazers in the comparison, which may change the findings
given below. Figure 10.9 shows the relationship between invertebrate
grazers and phytobenthos biomass for all sampling points where 40 or
more grazing individuals per square metre were observed. A low grazing
pressure seems to stimulate phytobenthos biomass development at first,
as it reduces the competition among the benthic algae for food and light.
At higher densities (>100 individuals/m2), the grazing pressure reduces
184
phytobenthos biomass. At very low grazer densities (< 40 ind./m²), no
correlation was found; phytobenthos biomass is most likely affected by
2,0
Phytobenthos biomass
(µg chl.-a/cm²)
(µg chl.-a/cm²)
other factors than grazing.
1,5
1,0
0,5
0,0
0
100
200
300
2,0
1,8
1,6
1,4
1,2
1,0
0,8
0,6
0,4
0,2
0,0
0
Invertebrate grazers (ind./m²)
A.
Figure 10.9
100
200
300
B.
Relationship of benthic invertebrate grazer densities of the genera
Theodoxus and Viviparus, and phytobenthos biomass. A:
sampling points ≥ 40 ind./m² (R² = 0.415); B. sampling points ≥
100 ind./m² (R² = 0.746).
The highest densities of grazers were found in Section 5 of the Danube
River (Hungarian Danube Bend; see Table 10.2). In this section, the
inverted relationship between grazers and phytobenthos biomass was
even more pronounced (Figure 10.10).
(µg/cm²)
1,6
1,4
1,2
1
0,8
0,6
0,4
0,2
0
0
50
100
150
200
250
300
Figure 10.10 Relationship of benthic invertebrate grazer densities of the genera
Theodoxus and Viviparus, and phytobenthos biomass for sampling
points in Danube Section 5 (R² = 0.925).
185
10.8. Phytobenthos biomass compared to benthic diatoms
The commonly used IPS index (L’Indice de Polluosensibilité Spécifique
(CEMAGREF, 1982 in Descy & Coste, 1991) values were calculated for the
JDS2 results of benthic diatom analyses, and classified using the reference
values and class boundaries as adopted in the Slovak classification
system. Figure 10.11 shows IPS index values aggregated for the Upper,
Middle and Lower Danube River, including the class boundaries. Figure
10.12 shows the phytobenthos biomass results found for each class on the
basis of the IPS index values for the entire Danube stretch and separately
for the Upper, Middle and Lower Danube River. These figures show that
phytobenthos biomass cannot be related directly to the IPS index values,
as a wide range of biomass results were found for each IPS value.
Figure 10.11 IPS index values for benthic diatoms and class boundaries
according to the Slovak classification system. Class boundaries:
between high and good ecological status
between good and moderate ecological
between moderate
status and poor ecological
between poor and
bad ecological status
status
186
A.
B.
Figure 10.12 Comparison of phytobenthos biomass and IPS index values for
benthic diatoms for the entire Danube stretch (A) and separated
for the Upper, Middle and Lower Danube (B).
187
10.9. Concluding remarks
Many factors influencing biomass and species composition of stream
benthic algae have been researched, in many different combinations:

Nutrients or nutrient availability (Hillebrand and Sommer, 2000; Hilton
et al., 2006; McCormick et al., 2001);

Nutrients and light conditions or canopy cover (Fuller et al., 2008; Hill
and Fanta, 2008; Schiller et al., 2007);

Flood disturbance (Biggs et al., 1998a; Cardinale et al., 2005; 2006);

Nutrients and flood disturbance (Biggs and Smith, 2002);

Nutrients and grazing (Rosemond et al., 1993; Stevenson et al.,
2006);

Nutrients, light and grazing (Liess and Kahlert, 2007; Rosemond,
1993; Rosemond et al., 2000).
Nutrients, light conditions (including canopy cover), flood disturbance (as
stream velocity) and grazing are all factors which largely determine the
phytobenthos community structure and taxonomic richness. However,
additional
anthropogenic
disturbances
such
as
sewage
treatment
discharges (Mainstone and Parr, 2002), toxic mining waste (Prat et al.,
1999), antifouling biocide (Alsterberg et al., 2007), or atrazine and zinc
(Guasch et al., 2003) may cause more severe changes in community
structure, leading to loss of biodiversity.
Phytobenthos density at a particular location is thus the result of many
factors. Some of these factors contribute to phytobenthic algal growth
(e.g. light, nutrients), others such as scouring floods or grazing reduce
phytobenthic biomass (Biggs, 1996). In the analysis as desribed above,
each parameter was assessed individually against phytobenthos biomass
results. In order to gain a better insight into phytobenthic biomass
developments in the Danube River, more data needs to be collected and
assessed
in
a
multi-parameter
analysis
of
all
factors
influencing
phytobenthos biomass. If anything, the analysis shows that further
research is necessary to develop an indicator of ecological or trophic state
out of phytobenthos biomass.
188
General discussion & conclusions
11.
General Discussion and Summary
There are many different ways to determine algal densities or biomass.
There is no such thing as the best method. Thus, the decision on what is
the best suitable method largely depends on the purpose of the study
performed. The objectives of this study were related to various different
methods to determine algal biomass, as stated in the Chapter “Aims of the
Present Dissertation”. Below, the results of each of these objectives are
discussed and summarised.
11.1. Objective 1: Development of a semi-automated method to
determine plankton biovolume while performing microscopic
cell counts
A semi-automated method to determine algal biovolume while performing
microscopic cell counts was developed with the help of digital image
processing technology as described in Chapter 3.
Firstly, geometric shapes were defined and assigned to phytoplankton
taxa. Complex shapes were converted to combinations of geometrical
shapes and measurable dimensions were defined, such as for Staurastrum
and Ceratium. Additionally, it was necessary to harmonise the counting
strategy
with
the
biovolume
measurements
for
filamentous
and
coenobium-forming algae, thereby introducing the “counting-unit”.
Secondly, the database and measurement routine of the BACCHUS
application were developed, including a statistical evaluation method. The
statistical analysis was used for two purposes:
1. to define the required precision of the biovolume results in order to
define a maximum number of measurements to be performed, thus
minimising the effort involved;
2. to enable statistical evaluation of both biovolume results and individual
dimensions, in order to reduce the number of taxa or dimensions to be
measured.
The statistical analysis made the BACCHUS application a very powerful
tool, since it helped to minimise the amount of time and effort needed to
perform a suitably precise biovolume analysis. Practice in the laboratory
with several technicians showed that plankton counts can be combined
with biovolume measurements very easily and comfortably. It has to be
189
noted that in routine samples not enough individuals of a taxon may be
present in the sample to reach the desired precision. However, even if no
image-analysis system is used to measure biovolumes, the statistical
analysis can be used to reduce the time needed for analysis.
11.2. Objective 2: Assessment of the in vivo absorption method
and comparison to other methods for the determination of
The newly developed European Guidance Standard EN 16161 (2012) was
compared to other commonly used methods for the determination of
chlorophyll-a, such as ISO 10260 (1992), in Chapter 4. The decomposition
of absorption spectra by means of Gauss curves as previously shown by
numerous others (e.g. Morrison, 2003; Hoepffner & Sathyendranath,
1991) results in adequate fit curves, although a freely modifiable 3 (for
chlorophyll-a absorption around 675 nm) yields better results than a fixed
value at 675 nm. It appeared to be impossible to correlate the EN 16161
results to chlorophyll-a concentrations as determined with the ISO 10260
method. Thus, the applicability of the EN 16161 Guidance Standard
remains questionable until this issue has been resolved.
11.3. Objective 3: Development of an Alert Level Framework based
on in vivo cyanobacterial fluorescence to protect drinking
water resources
Bartram et al. (1999) proposed an Alert Level Framework for harmful
cyanobacteria based on microscopic cell counts. A disadvantage of this
method is that microscopic cell counts are generally time-consuming, and
as such valuable time can be lost waiting for the results, whereas
adequate and timely actions are required to ensure public safety. Online
and on-site fluorescence techniques have the advantage of being able to
analyse samples at a higher frequency and produce results faster. These
features are especially important in the monitoring of drinking water
intake points where toxic algae blooms may appear suddenly, requiring
immediate action. In order to use in vivo fluorescence methods, it was
necessary to convert the Alert Levels units from cell densities to
cyanobacterial chlorophyll-a levels, as described in Chapter 6. Chapter 7
showed an example of the application of this Alert Level Framework in a
Polish lake near Warsaw, Poland.
190
The dynamics of daily average cyanobacterial chlorophyll-a concentrations
determined by the AlgaeOnlineAnalyser confirmed the importance of
continuous measurements by making daily cycles in algal densities visible.
The short time necessary to produce measurement results allowed for the
maximum possible time to take the necessary actions (increase of
treatment substances dosage, starting additional treatment processes or
closing the intake).
11.4. Objective 4: Development of an in situ and in vivo
fluorescence method for the quantification of phytobenthos
per algal group
The abundance and community structure of primary producers such as
(micro-)algae
and
macrophytes
are
indicators
of
the
level
of
eutrophication, one of the major anthropogenic pressures identified by the
European Water Framework Directive (2000/60/EC). Especially in rivers,
phytobenthic primary productivity is considered to be a suitable indicator
of stream water quality, since the short-lived organisms are generally
sessile, take up nutrients from the water phase (thus forming an indicator
for assimilated nutrients) and respond rapidly to a changing environment.
Moreover,
benthic
algal
communities
are
important
parameters
in
ecological investigations since they support lotic food webs as a food
source for grazers and they can attenuate currents and stabilise
sediments, thereby modifying the aquatic habitat (Allan & Castillo, 2007).
Dodds et al. (2002) stated that it is important to understand the
relationship between nutrient concentration and (micro-)phytobenthic
biomass when assessing a stream’s trophic state. Empirical regression
models which link algal biomass and water column nutrients have been
widely used in the eutrophication management of freshwater lakes and
reservoirs (Cooke et al., 1993; Smith 1998; Smith et al. 1999).
Establishing similar relationships for nutrients and phytobenthos biomass
in streams has appeared more difficult and is thus less advanced (e.g.
Biggs 2000), since these links are generally weaker in streams than in
lakes (Dodds et al. 2002). Problems caused by enrichment – e.g. aesthetic
degradation,
loss
of
pollution-sensitive
invertebrate
taxa
through
smothering of substrata by algae, clogging of water intake structures and
degradation of water quality (particularly dissolved oxygen and pH)
191
resulting in fish kills (Biggs et al. 2000) – could be addressed by such
relationships, thus providing an objective framework for stream water
quality management. Additionally, eutrophication models for streams can
be extended to predict ecosystem properties important to lotic food web
structures (Dodds et al. 2002).
Biomass development of (micro)phytobenthos is affected by numerous
factors other than nutrients, such as: (1) current velocity, (2) frequency
of biomass-scouring floods, (3) suspended sediment, (4) shading, (5)
substratum type, and (6) grazing (Welch et al., 1992). Although the
extent to which these factors influence the relationship between nutrient
levels and phytobenthos biomass may vary considerably, it seems that
nutrients may be the dominant factor in natural streams (Welch et al.,
1992). Similar to the approach in lakes, where successful statistical
models are based only on average nutrient loading and average residence
time, it may be possible to develop a stream model based on only a few
variables. Biggs et al. (1998b, 1998c) suggested that growth, as a
function of nutrient supply, and losses, as a function of the frequency of
flood disturbance, primarily determine the differences in mean monthly
benthic algal biomass among unshaded, temperate streams. In streams
with moderate to low nutrient supply and low flood frequency, losses from
grazing animals may also be significant (e.g., Rosemond 1994, Biggs et
al. 1998b, 1999). Biggs et al. (1999) explained 78.8-88.6% of the
variance in mean monthly chlorophyll-a in various studies of phytobenthos
in New Zealand streams using combinations of nutrient variables and
measures of flood disturbance. This suggests that it may very well be
possible to develop relatively simple models for the relationships between
nutrients and phytobenthos biomass in streams and rivers, which do not
exist at the present time.
This study has taken the first step in this direction by developing a
suitable field method to measure phytobenthos biomass per algal group,
without the problems related to quantitative sampling (see Chapter 9).
Chapter 10 shows some assessment results of a first elaborate dataset for
the Danube River, including phytobenthos biomass as a parameter.
However, further datasets of different types of streams and rivers now
need to be collected and evaluated. The field data thus collected can serve
192
as reliable input for the establishment of a model for the nutrientphytobenthos biomass relationship, in order to assess the trophic state of
rivers and streams.
Further research is then necessary to actually develop such a model and
identify class boundaries for phytobenthos biomass. Once developed, it
may provide a practical routine method to assess the trophic state of
rivers and streams as part of their ecological status in accordance with the
Water Framework Directive.
193
References
12.
References
Aberle, N., Beutler, M., Moldaenke, C. and Wiltshire, K.H. 2006. 'Spectral
fingerprinting' for specific algal groups on sediments in situ: a new
sensor. Arch. Hydrobiol. 167, 575-592.
Ahn, Ch-Y., Chung, A-S. and Oh, H-M. 2002. Rainfall, phycocyanin and
N:P ratios related to cyanobacterial blooms in a Korean large
reservoir. Hydrobiologia 474, 117-124.
Allan, J.D. and Castillo, M.M. 2007. Stream ecology: Structure and
function of running waters. 2nd edition. Chapman and Hall, New York.
Allen, M.M. 1968. Simple conditions for growth of unicellular blue-green
algae on plates. J. Phycol. 4, 1–4.
Alsterberg, C., Sundbäck, K. and Larson, F. 2007. Direct and indirect
effects of an antifouling biocide on benthic microalgae and meiofauna.
J. Exp. Mar. Biol. Ecol. 351, 56-72.
Anonymus. 1992. ATT Loseblattsammlung. Ceratium hirundinella (O.F.
Müller) Schrank . Dinophyceae. No. 525 310 108.
Backer, L.C. 2002. Cyanobacterial harmful algal blooms (CyanoHABs):
Developing a public health response. Lake Reservoir Manage. 18 (1),
20-31.
Barlow, R.G., Aiken, J., Holligan, P.M., Cummings, D.G., Maritorena, S
and
Hooker,
S.
2002.
Phytoplankton
pigment
and
absorption
characteristics along meridional transects in the Atlantic Ocean. DeepSea Res. I 47, 637-660.
Bartram, J., Burch, M., Falconer, I.R., Jonem, G. and Kuiper-Goodman, T.
1999. Situation assessment, planning and management. In: I.
Chorus, J. Bartram, (Eds), Toxic cyanobacteria in water. A guide to
their public health consequences, monitoring and management, E&FN
Spon on behalf of the World Health Organization, London, p. 179–209.
Begon, M., J.L. Harper & C.R. Townsend. 1990. Ecology: Individuals,
Populations
and
Communities.
Blackwell
Scientific
Publications,
Oxford, pp. 945.
Bellinger, E.G. 1974. A note on the use of algal sizes in estimates of
population standing crops. Br. Phycol. J. 9, 157-161.
195
References
Beutler, M., K.H. Wiltshire, C. Lüring, C. Moldaenke, and Lohse, D. 2001.
Fluorometric
depth-profiling
of
chlorophyll
corrected
for
yellow
substances. In Proceedings of the Auqaculture, Environment and
Marine Phytoplankton Symposium, 21-23 May 2001, Brest (France),
pp 231-238.
Beutler, M., Wiltshire, K.H., Meyer, B., Moldaenke, C., Lüring, C.,
Meyerhöfer, M., Hansen, U.P. and Dau, H. 2002. A fluorometric
method for the differentiation of algal populations in vivo and in situ.
Photosynth. Res. 72, 39-53.
Bidigare, R.R., M.E. Ondrusek, J.H. Morrow and D.A. Kiefer. 1990. In vivo
absorption properties of algal pigments. Proc. Int. Soc. Optical Eng.
1302, 290-302.
Biggs, B.J.F. 1995. The contribution of flood disturbance, catchment
geology and land use to the habitat template of periphyton in stream
ecosystems. Freshwater Biol. 33, 419-438.
Biggs, B.J.F. 2000. Eutrophication of streams and rivers: dissolved
nutrient-chlorophyll relationships for benthic algae. J. North Am.
Benthol. Soc., 19(1), 17-31.
Biggs, B.J.F. and Smith, R.A. 2002. Taxonomic richness of stream benthic
algae: effects of flood disturbance and nutrients. Limnol. Oceanogr.
47(4), 1175-1186.
Biggs, B.J.F., Goring, D.G. and Nikora, V.I. 1998a. Subsidy and stress
responses of stream periphyton to gradients in water velocity as a
function of community growth form. J. of Phycol. 34, 598-607.
Biggs, B.J.F., C. Kilroy, and Lowe, R.L. 1998b. Periphyton development in
three valley segments of a New Zealand grassland river: Test of a
habitat matrix conceptual model within a catchment. Arch. Hydrobiol.
143, 147-177.
Biggs, B.J.F., Smith, R.A. and Duncan, M.J. 1999. Velocity and sediment
disturbance of periphyton in headwater streams: biomass and
metabolism. J. North Am. Benthol. Soc. 18(2), 222-241.
Biggs, B.J.F., R.J. Stevenson, and Lowe, R.L. 1998c. A habitat matrix
conceptual model for stream periphyton. Arch. Hydrobiol. 143, 2556.
196
References
Blahova, L., Babica, P., Maršalkova, E., Maršálek, B. and Blaha, L. 2007.
Concentrations and seasonal trends of extracellular microcystins in
freshwaters of Czech Republic – results of the National monitoring
Program, CLEAN 35(4), 348-354.
Breitig, G & Von Tümpling, W. 1982. Ausgewählte Methoden der
Wasseruntersuchung. Band II. 2. Auflage. Gustav Fischer Verlag.
Jena.
Bricaud, A., Babin, M., Morel, A. and Claustre, H. 1995. Variability in the
chlorophyll-specific absorption coefficients of natural phytoplankton:
Analysis and parameterization. J. Geophys. Res. 100(C7), 13,32113,332.
Brock, T.D., Madigan, M.T., Martinko, J.M. and Parker, J. 1997. Biology of
microorganisms, 8th edition. Prentice Hall International Inc., USA, p.
573.
Burch, M.D. 1993. The Development of an Alert Levels and Response
Framework for the Management of Blue-Green Algae Blooms. In
Proceedings of Australian Centre for Water Quality Research and the
University of Adelaide Symposium on Blue-Green Algae Blooms – New
Developments in Research and Management, February 17, Adelaide,
South Australia.
Burger-Wiersma, T. and Versteegh, J.F.M. 1994. Toxische cyanobacteriën
in
recreatiewateren.
In
opdracht
van:
RIVM,
Bilthoven.
Rapportnummer: 609021003. (In Dutch).
Burgis, M.J. 1975. The biomass of cyclopoids in two Thames Valley
reservoirs. Verh. Internat. Verrein. Limnol. 19, 1834-1840.
Campbell, N.A. 1993. Biology, 3rd edition. NY: The Benjamin/Cummings
Publishing Company Inc., USA, pp. 1190.
Cardinale, B.J., Hillebrand, H. and Charles, D.F. 2006 Geographic
patterns of diversity in streams are predicted by a multivariate model
of disturbance and productivity. J. Ecol. 94, 609-618.
Cardinale, B.J., Palmer, M.A., Ives, A.R. and Brooks, S.S. 2005.
Diversity-productivity relationships in streams vary as a function of
the natural disturbance regime. Ecology 86(3), 716-726.
Carmichael, W.W. 2001. Health effects of toxin-producing cyanobacteria:
“The CyanoHabs”, Hum. Ecol. Risk Assess. 7, 1393-1407.
197
References
Cattaneo, A. and Roberge, G. 1991. Efficiency of a brush sampler to
measure periphyton in streams and lakes. Can. J. Fish Aquat. Sci. 48,
1877-1881.
Cayelan, C.C., Ibelings, B.W., Hoffmannd, E.P., Hamiltone, D.P. and
Brookes,
J.D.
2012.
Eco-physiological
adaptations
that
favour
freshwater cyanobacteria in a changing climate, Water Res. 46 (5),
1394-1407.
CEN. 2003. EN 13946: Water quality - Guidance standard for the routine
sampling and pre-treatment of benthic diatoms from rivers.
CEN. 2009. EN 15708: Water quality - Guidance standard for the
surveying, sampling and laboratory analysis of microphytobenthos in
shallow running water.
CEN. 2012. EN 16161: Water Quality - Guidance on the use of in vivo
absorption
techniques
for
the
estimation
of
chlorophyll-a
concentration in marine and fresh water samples.
Chorus, I. and Bartram, J. (Eds). 1999. Toxic Cyanobacteria in Water, A
Guide
to
Their
Public
Health
Consequences,
Monitoring
and
Management, Published by E&FN Spon on behalf of the World Health
Organization, London.
Chorus, I., Falconer, I.R., Salas, H.J. and Bartram, J. 2000. Health risks
caused by freshwater cyanobacteria in recreational waters. J. Toxicol.
Environ. Health B Crit. Rev. 3:4, 323-347.
Ciotti, A.M., Lewis, M.R. and Cullen, J.J. 2002. Assessment of the
relationships between dominant cell size in natural phytoplankton
communities and the spectral shape of the absorption coefficient.
Limnol. Oceangr. 47(2), 404-417.
Clasen, J. 1994. Die Bedeutung biologischer Untersuchungen für die
Steuerung
und
Überwachung
von
Aufbereitungsanlagen
an
Trinkwassertalsperren. In: Ketelaars, H.A.M., Nienhüser, A.E. &
Hoehn,
E.
(Eds),
Die
Biologie
der
Trinkwasserversorgung
aus
Talsperren, ATT Information, Academic Book Centre, De Lier, p. 8598.
Cleveland, J.S. 1995. Regional models for phytoplankton absorption as a
function of chlrophyll a concentration. J. Geophys. Res. 100(C7),
13,333-13,344.
198
References
Cook, C.M., Vardaka, E. and Lanaras, T. 2004. Toxic cyanobacteria in
Greek freshwaters, 1987–2000: occurrence, toxicity, and impacts in
the Mediterranean region, Acta Hydroch. Hydrob. 32(2), 107-124.
Cooke, G.D., Welch, E.B., Martin, A.B., Fulmer, D.G., Hyde, J.B. and
Schrieve, G.D. 1993. Effectiveness of Al, Ca, and Fe salts for control
of
internal
phosphorus
loading
in
shallow
and
deep
lakes.
Hydrobiologia 253, 323-335.
Danuhyarso-Novemsia, S. 1996. Voorbereidende werkzaamheden voor
het opzetten van een routinematige methode voor het bepalen van
het biovolume. WBB-rapportnummer hk/no 96-35133. N.V. WBB.
Werkendam. (In Dutch).
Davies, A.L. and Gee, J.H.R. 1993. A simple periphyton sampler for algal
biomass estimates in streams. Freshwater Biol. 30, 47-51.
Day, J.G., Watanabe, M.M., Morris, G.J., Fleck, R.A. and McLellan, M.R.
1997. Long-term viability of preserved eukaryotic algae. J. Appl.
Phycol. 9, 121-127.
De Julio, M., Fioravante, D.A., De Julio, T.S., Oroski, F.I. and Graham,
N.J.D.
2010.
A
methodology
for
optimising
the
removal
of
cyanobacteria cells from a brazilian eutrophic water, Braz. J. Chem.
Eng. 27(01), 113-126.
Descy J.-P. and Coste, M. 1991. A test of methods for assessing water
quality based on diatoms. Verh. Internat. Verein. Limnol. 24, 21122116.
DIN. 1985. DIN 38412 Teil 16. Deutsche Einheitsverfahren zur Wasser-,
Abwasser-
und
Schlammuntersuchung;
Testverfahren
mit
Wasserorganismen (Gruppe L) – Bestimmung des Chlorophyll-aGehaltes von Oberflächenwasser (L42).
Dobrzańska, J. and Hewelke, P. 2010. Environmental evaluation of
chosen oxbow lakes of the Vistula in Warsaw. Infrastruktura i Ekologia
Terenów Wiejskich 09, 65–75.
Dodds, W.K. 2006. Eutrophication and trophic state in rivers and
streams. Limnol. Ocanogr. 51(1-2), 671-680.
Dodds,
W.K.,
Smith,
V.H.
and
Lohman,
K.
2002.
Nitorgen
and
phosphorous relationships to benthic algal biomass in temperate
streams. Can. J. Fish. Aquat. Sci. 59, 865-874.
199
References
Dumont, H.J., Van de Velde, I. and Dumont, S. 1975. The dry weight
estimate of biomass in a selection of Cladocera, Copepoda and
Rotifera from the plankton, periphyton and benthos of continental
waters. Oecologia 19, 75-97.
Duy, T.N., Lam, P.K.S., Shaw, G. and Connell, D.W. 2000. Toxicology
and risk assessment of freshwater cyanobacterial (bluegreen algal)
toxins in water, Rev. Environ. Contam. Toxicol. 163, 113-186.
Elser, J.J., Elser, M.M. and Carpenter, S.R. 1986. Size fractionation of
algal
chlorophyll,
carbon
fixation
and
phosphatase
activity:
relationships with species-specific size distributions and zooplankton
community structure. J. Plankton Res. 8, 365-383.
Falconer, I.R. 2005. Cyanobacterial Toxins of Drinking Water Supplies:
Cylindrospermopsins and Microcystins. Boca Raton: CRC Press.
Falconer, I.R. 1994. Health problems from exposure to cyanobacteria and
proposed safety guidelines for drinking and recreational water. In:
Codd, G., Jefferies, T.M., Keevil, C.W. and Potter, E. (Eds). Detection
methods for cyanobacterial toxins. Cambridge: The Royal Society of
Chemistry, p. 3-10.
Felip, M. and Catalan, J. 2000. The relationship between phytoplankton
biovolume and chlorophyll in a deep oligotrophic lake: decoupling in
their spatial and temporal maxima. J. Plankton Res. 22(1), 91–105.
Fowler, J., Cohen, L. and Jarvis, P. 1998. Practical statistics for field
biology. 2nd Edition. Wiley, Chichester, United Kingdom.
Francis, G. 1878. Poisonous Australian Lake. Nature 18, 11-12.
Francis, P.S., Hindson, C.M., Terry, J.M., Smith, Z.M., Slezak, T., Adcock,
J.L., Fox, B.L. and Barnett, N.W. 2011. Enhanced permanganate
chemiluminescence. Analyst 136, 64-66.
Fuller, R.L., LaFave, C., Anastasi, M., Molina, J., Salcedo, H. and Ward, S.
2008. The role of canopy cover on the recovery of periphyton and
macroinvertebrate
communities
after
a
month-long
flood.
Hydrobiologia 598, 47-57.
Fujiki, T. and Taguchi, S. 2002. Variability in chlorophyll a specific
absorption coefficient in marine phytoplankton as a function of cell
size and irradiance. J. Plankton Res. 24(9), 859-874.
200
References
Giokas,
D.L.,
Tsogas,
derivatization
G.Z.
and
coupled
Vlessidis,
to
flow
A.G.
2009.
injection
On-line
permanganate
chemiluminscence detection of total carbonyl compounds in natural
waters and drinking water. Anal. Chim. Acta 651, 188-195.
Guillard, R.R.L. 1975. Culture of phytoplankton for feeding marine
invertebrates. In: Smith, W.L. and Chantey, M.H. (Eds), Culture of
Marine Invertebrate Animals. Plenum Publishers, New York, p. 29–60.
Govindjee.
1995.
Sixty-three
years
since
Kautsky:
Chlorophyll
a
fluorescence. Aust. J. Plant Physiol. 22, 131–160.
Govindjee. 2004. Chlorophyll a fluorescence: A bit of basics and history.
In:
Papageorgiou,
fluorescence:
A
G.C
and
signature
Govindjee
of
(Eds),
photosynthesis.
Chlorophyll
Springer,
a
The
Netherlands, pp. 1-42.
Gower, J.F.R., Doerffer, R. and Borstad, G.A. 1999. Interpretation of the
685
nm
peak
in
water-leaving
radiance
spectra
in
terms
of
flurescence, absorption and scattering, and its observation by MERIS.
Int. J. Remote Sensing 20(9), 1771-1786.
Gregor J. and Maršálek B. 2004. Freshwater phytoplankton quantification
by chlorophyll a: a comparative study of in vitro, in vivo and in situ
methods. Water Res. 38(3), 517-522.
Gregor, J. and Maršálek, B. 2005. A simple in vivo fluorescence method
for
the
selective
detection
and
quantification
of
freshwater
cyanobacteria and eukaryotic algae, Acta Hydroch. Hydrob. 33(2),
142-148.
Gregor, J., Geris, R., Maršálek, B., Hetesa, J. and Marvan, P. 2005. In
situ quantification of phytoplankton in reservoirs using a submersible
spectrofluorometer. Hydrobiologia 548, 141-151.
Gregor, J., Maršálek, B. and Šipková H. 2007. Detection and estimation
of potentially toxic cyanobacterial in raw water at the drinking water
treatment plant by in vivo fluorescence method. Water Res. 41, 228234.
Grossman, A.R., Schaefer, M.R., Chiang, G.G. and Collier, J.L. 1994. The
responses of cyanobacteria to environmental conditions: light and
nutrients.
In:
Bryant,
D.A.
(Ed.).
The
molecular
biology
of
cyanobacteria. Kluwer Academic Publishers, Dordrecht, pp. 641-675.
201
References
Guasch, H., Admiraal, W. and Sabater, S. 2003. Contrasting effects of
organic and inorganic toxicants on freshwater periphyton. Aquat.
Toxicol. 64, 165-175.
Haney, J.F. and Hall, D.J. 1973. Sugar-coated Daphnia: a preservation
technique for Cladocera. Limnol. Oceanogr. 18, 331-333.
Heaney S.I. 1978. Some observations on the use of the in vivo
fluorescence
technique
to
determine
chlorophyll-a
in
natural
populations and cultures of freshwater phytoplankton. Freshwater
Biol., 8, 115-126.
HELCOM.
2008.
Manual
for
marine
monitoring
in
the
COMBINE
Programme of HELCOM, PART C. Programme for monitoring of
eutrophication and its effects, Annex C-4 Phytoplankton chlorophyll-a:
http://www.helcom.fi/groups/monas/CombineManual/AnnexesC/en_G
B/annex4/.
Hill, W.R. and Fanta, S.E. 2008. Phosphorus and light colimit periphyton
growth at subsaturating irradiances. Freshwater Biol. 53: 215-225.
Hillebrand, H. and Sommer, U. 2000. Diversity of benthic microalgae in
response to colonization time and eutrophication. Aquat. Bot. 67,
221-236.
Hillebrand, H., Dürselen, C.-D., Kirschtel, D., Pollingher, U. and Zohary,
T. 1999. Biovolume calculation for pelagic and benthic microalgae. J.
Phycol. 35, 403-424.
Hilton, J., O'Hare, M., Bowes, M.J. and Jones, J.I. 2006. How green is my
river? A new paradigm of eutrophication in rivers. Sci. Total Environ.
365, 66-83.
Hoehn, E., Clasen, J., Scharf, W., Ketelaars, H.A.M., Nienhüser, A.E.,
Horn, H. Kersken, H. and Ewig, B. 1998. Erfassung und Bewertung
von Planktonorganismen. ATT Technische Informationen 7 . R.
Oldenbourg Verlag München, Siegburg, pp. 151.
Hoepffner, N. and Sathenyendranath, S. 1991. Effect of pigment
composition on absorption properties of phytoplankton. Mar. Ecol.
Prog. Ser. 73, 11-23.
Holm-Hansen, O. 1963. Viability of blue-green and green algae after
freezing. Physiol. Plant. 16, 530-540.
202
References
Hoogenboezem, W., Wagenvoort, A.J. and Blaauboer, K. 2004. The
occurrence of toxic cyanobacteria in some Dutch surface waters used
for the production of drinking water. RIWA-report, ISBN: 90-6683108-1, Nieuwegein. To be downloaded from www.riwa.org under
Publications.
Horn, W. 1978. Der Anteil des Crustaceenplanktons an der Verringerung
des Phytoplanktongehalts in der Saidenbachtalsperre. Ph. D. thesis.,
University of Technology, Dresden.
Hornig, U. 1983. Analyse der Struktur und Dynamik der wichtigsten
Zooplanktonpopulationen in der Talsperre Bautzen in Beziehung zur
Änderung der Fischbesiedlung. Dipl. Arb. TU Dresden, Sektion
Wasserwesen.
House, J., Ho, L., Newcombe, G. and Burch, M. 2004. Management
strategies for toxic blue-green algae: Literature Survey. A report for
the CRC for Water Quality and Treatment Project 2.0.2.4.1.2.
Huisman J., Mattijs, H.C.P. and Visser, P.M. (Eds). 2005. Harmful
Cyanobacteria. Springer, Dordrecht, the Netherlands, pp. 241.
Hütte. 1924. "Hütte". Des Ingenieurs Taschenbuch. 1. Band. 24. Auflage.
Akademischen Verein Hütte, Berlin.
ICPDR. 2002. Joint Danube Survey. Technical Report of the International
Commission for the Protection of the Danube River. ICPDR, Vienna.
ISO.
1992.
ISO
measurement
10260:
of
Guidance
biochemical
Standard
parameters
on
water
–
quality:
spectrometric
determination of the chlorophyll-a concentration.
Izydorczyk,
K.
and
Tarczynska,
M.
2005.
Influence
of
wind
on
cyanobacterial bloom development in shallow, lowland reservoir.
Ecohydrol. Hydrobiol. 5, 255-263.
Izydorczyk, K. Tarczynska, M. Jurczak, T. Mrówczyński, J. and Zalewski,
M. 2005. Measurement of phycocyanin fluorescence as an online early
warning system for cyanobacteria in reservoir intake water. Environ.
Toxicol. 20, 425-430.
Izydorczyk, K., Carpentier, C., Mrowczyński, J., Wagenvoort, A., Jurczak,
T. and Tarczyńska, M. 2009. Establishment of an Alert Level
Algae Online Analyser for monitoring cyanobacterial chlorophyll-a,
Water Res. 43, 989-996.
203
References
Izydorczyk,
K.;
Jurczak,
T.;
Wojtal-Frankiewicz
A.,
Skowron,
A.,
Mankiewicz-Boczek, J., Tarczyńska, M. 2008. Influence of abiotic and
biotic factors on microcystin content in Microcystis aeruginosa cells in
a eutrophic temperate reservoir. J. Plank. Res. 30(4), 393-400.
Jeffrey, S.W., Mantoura, R.F.C. and Wright, S.W. 1997. Phytoplankton
pigments in oceanography: Guidelines to modern methods, UNESCO
Publishing, Paris, pp. 661.
Jesus, B., Perkins, R.G., Mendes, C.R., Brotas, V. and Paterson, D.M.
2006. Chlorophyll fluorescence as a proxy for microphytobenthic
biomass: alternatives to the current methodology. Mar. Biol. 150, 17–
28.
Jiménez, F. Rodríguez, J. Bautista, B and Rodríguez, V. 1987. Relations
between chlorophyll, phytoplankton cell abundance and biovolume
during a winter bloom in Mediterranean coastal water. J. Exp. Mar.
Biol. Ecol. 105, 161-173.
Jones, J.G. 1978. A guide to methods for estimating microbial numbers
and biomass in fresh water. Freshwater Biological Association. Scient.
Publ. 39.
Jurczak, T., Tarczynska, M., Izydorczyk, K., Mankiewicz, J., Zalewski, M.
and Meriluoto, J. 2005. Elimination of microcystins by water treatment
process – examples from Sulejow Reservoir, Poland. Water Res. 39,
2394-2406.
Jurczak, T., Tarczynska, M., Karlsson, K. and Meriluoto, J. 2004.
Characterization
and
diversity
of
cyanobacterial
hepatotoxins
(microcystins) in blooms from Polish freshwaters identified by liquid
chromatography-electrospray
ionisation
mass
spectrometry.
Chromatographia 59, 571-578.
Kiefer, F. 1960. Ruderfusskrebse (Copepoden). Kosmos, Stuttgart.
Kim, K. 1980. Contribution a l’étude de l’écosysteme pelagique dans les
parages de Carry-le-Rouet. Tethys 9, 317-344.
Kolber, Z.S., Prasil, O. and Falkowski, P.G. 1998. Measurements of
variable chlorophyll fluorescence using fast repetition rate techniques:
defining methodology and experimental protocols. Biochim. Biophys.
Acta 1367, 88-106.
Koste, W. 1978. Rotatoria. Die Rädertiere Mitteleuropas begründet von
Max Voigt. 2. Tafelband. Gebrüder Borntraeger. Berlin.
204
References
Krause,
G.H.
and
Weis,
E.
1991.
Chlorophyll
fluorescence
and
photosynthesis: the basics. Annu. Rev. Plant Physiol. Plant Mol. Biol.
42, 313-349.
Krauze-Tomczyk, I. and Ostrowski, J. (Eds.) 2006. Komisja Nazw
Miejscowości i Obiektów Fizjograficznych [Commission Names of
Places
and
Physiographic
Objects]:
Hydronimy,
Wody
stojące
[Standing waters], Główny Urząd Geodezji i Kartografii [Head Office of
Geodesy and Cartography], Warszawa.
Kromkamp, J.C., Barranguet, C. and Peene, J. 1998. Determination of
microphytobenthos
PSI1
quantum
efficiency
and
photosynthetic
activity by means of variable chlorophyll fluorescence. Mar. Ecol.–
Prog. Ser. 162, 45-55.
Kuhl, A. and Lorenzen, A. 1963. Handling and culturing of Chlorella. In :
Prescot, D.M. (Ed.). Methods in cell physiology. Acad Press N.Y., pp 1183.
Latasa, M., Bidigare, R.R., Ondrusek, M.E,. and Kennicutt, M.C. 1996.
HPLC analysis of algal pigments: a comparison exercise among
laboratories
and
recommendations
for
improved
analytical
performance. Mar. Chem. 5l-, 315-324.
Lawton, L.; Maršálek, B.; Padisak, J.; Chorus, I. 1999. Determination of
cyanobacteria in the laboratory In Toxic cyanobacteria in water. A
guide
to
their
public
health
consequences,
monitoring
and
management. Chorus, I.; Bartram, J., Eds, E&FN Spon, pp. 347-367.
Leboulanger, Ch.; Dorigo, U.; Jacquet, S.; Le Berre, B.; Paolini, G.;
Humbert, J-F. 2002. Application of a submersible spectrofluorometer
for rapid monitoring of freshwater cyanobacterial blooms: a case
study. Aquat. Microb. Ecol. 30, 83-89.
Lee, T.; Tsuzuki, M.; Takeuchi, T.; Yokoyama, K.; Karube, I. 1994. In
vivo
fluorometric
method for
early
detection
of cyanobacterial
waterblooms. J. Appl. Phycol. 6, 489-495.
Lee, T.; Tsuzuki, M.; Takeuchi, T.; Yokoyama, K.; Karube, I. 1995.
Quantitative determination of cyanobacteria in mixed phytoplankton
assemblages by an in vivo fluorimetric method. Anal. Chim. Acta 302,
81-87.
205
References
Liess, A. and Kahlert, M. 2007. Gastropod grazers and nutrients, but not
light, interact in determining periphytic algal diversity. Oecologia 152,
101-111.
Liška, I., Wagner, F. and Slobodník, J. (Eds). 2008. Joint Danube Survey
2, Final Scientific Report, 1st edn, ICPDR - International Commission
for the Protection of the Danube River.
http://www.icpdr.org/jds/files/ICPDR_Technical_Report_for_web_low_
corrected.pdf.
Loeb, S.L. 1981. An in situ method for measuring the primary
productivity and standing crop of the epilithic periphyton community
in lentic systems. Limnology and Oceanography, 26, 394-399.
Lohman,H. 1908. Untersuchungen zur Feststellung des vollständigen
Gehaltes des Meeres an Plankton. Wiss. Meeresunters. 10, 131-370.
Lorenzen, C.J. 1966. A method for the continuous measurement of in
vivo chlorophyll concentration. Deep Sea Research, 13, 223-227.
Mainstone, C.P. and Parr, W. 2002. Phosphorus in rivers - ecology and
management. The Science of The Total Environment 282-283, 25-47.
Mankiewicz-Boczek, J.; Izydorczyk, K.; Romanowska-Duda, Z.; Jurczak
T.; Stefaniak, K.; Kokociński, M. 2006. Detection and monitoring
toxigenicity of cyanobacteria by application of molecular methods.
Environ. Toxicol., 21, 380–387.
Maxwell, K. and Johnson, G.N. 2000. Chlorophyll fluorescence – a
practical guide. Journal of Experimental Botany, 51 (345), 659-668.
McCormick, P.V., O’Dell, M.B., III, R.B.E.S., Backus, J.G. and Kennedy,
W.C.
2001.
Periphyton
responses
to
experimental
phosphorus
enrichment in a subtropical wetland. Aquatic Botany 71, 119-139.
Meriluoto, J. and Codd, G.A. (Eds.). 2005. TOXIC. Cyanobacterial
monitoring and cyanotoxin analysis. Abo, Abo Akademi University
Press.
Michielsen,
BF
&
Ketelaars,
biomassaparameter
HAM.
1994.
Het
celvolume
als
voor het fytoplankton. WBB rapportnummer
hk/mvw 94-27095. N.V. WBB. Werkendam. (In Dutch).
Millie, D.F., Paerl, H. and Hurley, J.P. 1993. Microalgal pigment
assessments
using
High-Performance
Liquid
Chromatography:
a
synopsis of organismal and ecological applications. Can. J. Fish.
Aquat. Sci. 50, 2513-2527.
206
References
Moberg, E.A and Sosik, H.M. 2012. Distance maps to estimate cell
volume
from
two-dimensional
plankton
images.
Limnology
and
Oceanography: Methods 10, 278-288.
Morris, G.J. 1963. The Cryopreservation of Chlorella. Arch. Microbiol.
107, 309-312.
Morrison, J.R. 2003. In situ determination of the quantum yield of
phytoplankton
chlorophyll
a
fluorescence:
A
simple
algorithm,
observations, and a model. Limnol. Oceanogr. 48 (2), 618-631.
Mrówczyński, J. 2004. Application of fluorescence measurements as an
early warning system in detecting cyanobacterial blooms. University of
Lodz, Faculty of Biology and Environmental Protection, Poland, pp. 91.
Munsell Color X-Rite. 2009. Munsell Soil Color Charts. Grand Rapids, MI,
USA.
Nauwerck,
A.
1963.
Phytoplankton
im
Die
See
Beziehungen
Erken.
A.-B.
zwischen
Zooplankton
Lundequistska
und
Bokhandeln,
Uppsala.
Newcombe, G. 2012. Cyanobacteria: Impacts of climate change on
occurrence, toxicity and water quality management, Water Research
46 (5), 1347-1348.
Nguyen, D.C., Keller, R.A., Jett, J.H. and Martin, J.C. 1987. Detection of
single molecules of phycoerythrin in hydrodynamically focused flows
by laser-induced fluorescence. Analytical Chemistry 59, 2158-2161.
Nogrady, T, Wallace, RL & Snell, TW. 1993. Rotifera. Biology, ecology and
systematics. Guides to the identification of the microinvertebrates of
the continental waters of the world. Volume 1. SPB Academic
Publishing bv. The Hague.
Osutki, A.; Omi, T.; Hashimoto, S.; Lizaki, M.; Takamura, N. 1994. HPLC
fluorometric determination of natural phytoplankton phycocyanin and
its usefulness as cyanobacterial biovolume in highly eutrophic shallow
lake. Water Air Soil Poll. 76, 383-396.
Paerl, H.W. 1988. Nuisance phytoplankton blooms in coastal, estuarine,
and inland waters, Limnology and Oceanography, 33 (4, part 2), 823847.
Paerl, H.W. and Paul, V.J. 2012. Climate change: Links to global
expansion of harmful cyanobacteria, Water research, 46 (5), 13491363.
207
References
Papageorgiou, G.C. and Stamatakis, K. 2004. Water and Solute transport
in
cyanobacteria
as
probed
by
chlorophyll
fluorescence.
In:
Papageorgiou, G.C. and Govindjee (Eds), Chlorophyll a Fluorescence:
A Signature of Photosynthesis. Springer, The Netherlands, pp. 663678.
Pavlova, V., Babica, P. Todorova, D., Bratanova, Z. and Maršálek, B.
2006. Contamination of some reservoirs and lakes in Republic of
Bulgaria by microcystins, Acta Hydrochimica et Hydrobiologica, 34,
437-441.
Perkins, R.G., J.C. Kromkamp, J. Serôdio, J. Lavaud, B. Jesus, J-L.
Mouget, S. Lefebvre and R.M. Forster. 2011. The application of
variable chlorophyll fluorescence to microphytobenthic biofilms. In D.
Suggett, O. Prasil and M. Borowitzka [eds], Chlorophyll a fluorescence
in aquatic sciences: methods and applications, Developments in
Applied Phycology. Vol 4: 237-275. Springer Science+Business Media
B.V.
Petruševski, B. 1996. Algae and particle removal in direct filtration of
Biesbosch water. Influence of algal characteristics, oxidation, and
other pre-treatment conditions. Thesis Technical University, Delft.
Pobel, D., Robin, J. and Humbert, J.-F. 2011. Inﬂuence of sampling
strategies on the monitoring of cyanobacteria in shallow lakes:
Lessons from a case study in France, Water Research, 45, 1005-1014.
Prat, N., Toja, J., Solà, C., Burgos, M.D., Plans, M. and Rieradevall, M.
1999. Effect of dumping and cleaning activities on the aquatic
ecosystems of the Guadiamar River following a toxic flood. The
Science of The Total Environment 242, 231-248.
Pridmore, R.D. and Hewitt, J.E. 1984. Chlorophyll a as an indicator of
phytoplankton cell volume in 12 lakes, North Island, New Zealand.
New Zealand Journal of Botany 22, 295-301.
Prieur, L. and Sathyendranath, S. 1981. An optical classification of
coastal and oceanic waters based on the specific spectral absorption
curves of phytoplankton pigments, dissolved organic matter, and
other particulate materials. Limnol. Oceanogr. 26 (4), 671-689.
Rapala, J. 1998. Toxins production by freshwater cyanobacteria: effects
of environmental factors. Ph. D. thesis. University of Helsinki.
208
References
Reynolds,
C.S.
1984.
The
ecology
of
freshwater
phytoplankton.
Cambridge University Press. Cambridge.
Romanowska-Duda, Z.B., Grzesik, M. and Kalaji, H.M. 2010. Physiological
activity of energy plants fertilized with sevage sludge and usefulness
of
the
Phytotoxkit
test
in
practice.
Environment
Protection
Engineering, 36 (1), 73-81.
Romanowska-Duda, Z.B., Kalaji, H.M. and Strasser, R. 2003. Chlorophyll
fluorescence as marker for heavy metals effects on water plants
(Spirodela
oligorrhiza),
Polish
Journal
of
Natural
Sciences,
1,
supplement, 213-214.
Rosemond,
A.D.
1993.
Interactions
among
irradiance,
nutrients,
constrain a stream algal community. Oecologia 94, 585-594.
Rosemond, A.D. 1994. Multiple factors limit seasonal variation in
periphyton in a forest stream. Journal of the North American
Benthological Society 13, 333-344.
Rosemond, A.D., Mulholland, P.J. and Brawley, S.H. 2000. Seasonally
shifting limitation of stream periphyton: response of algal populations
and assemblage biomass and productivity to variation in light,
nutrients, and herbivores. Canadian Journal of Fisheries and Aquatic
Sciences 57, 66-75.
Rosemond, A.D., Mulholland, P.J. and Elwood, J.W. 1993. Top-Down and
Bottom-Up Control of Stream Periphyton: Effects of Nutrients and
Herbivores. Ecology 74(4), 1264-1280.
Rott,
E.
1981.
Some
results
from
phytoplankton
counting
intercalibrations. Schweiz. Z. Hydrol. 43, 34-61.
Rott, E., Salmaso, N. and Hoehn, E. 2007. Quality control of Utermöhlbased phytoplankton counting and biovolume estimates – an easy
task or a Gordian knot? Hydrobiologia 578, 141-146.
Roy, S., Chanut, J.-P., Gosselin, M. and Sime-Ngando, T. 1996.
Characterization of phytoplankton communities in the lower St.
Lawrence Estuary using HPlC-detected pigments and cell microscopy.
Mar. Ecol. Prog. Ser. 142, 55-73.
Ruttner, F. 1952. Planktonstudien der Deutschen Limnologischen SundaExpedition. Arch. Hydrobiol. Suppl. 21, 1-274.
Ruttner-Kolisko, A. 1977. Suggestions for biomass calculation of plankton
rotifers. Arch. Hydrobiol. Beih. 8, 71-76.
209
References
Sarcina,
M.,
Tobin,
M.J.
and
Mullineaux,
W.
2001.
Diffusion
of
phycobilisomes on the thylakoid membranes of the cyanobacterium
Synechococcus 7942. Journal of Biological Chemistry 276 (50),
46830-46834.
Sars, G.O. 1993. On the freshwater crustaceans occurring in the vicinity
of Christiania. Translation of the handwritten manuscript (1861).
Marit, E., Eie, J.A., Halvorsen, G., Hobæk, A. and Larsson, P (Eds).
University of Bergen. Bergen.
Schiller, D.V., Martí, E., Riera, J.L. and Sabater, F. 2007. Effects of
nutrients and light on periphyton biomass and nitrogen uptake in
Mediterranean streams with contrasting land uses. Freshwater Biology
52, 891-906.
Schimanski, J., M. Beutler, C. Moldaenke, and U.-P. Hansen. 2006. A
model for correcting the fluorescence signal from a free-falling depth
profiler. Water Res. 40, 1616–1626.
Schreiber, U. 2004. Pulse-Amplitude-Modulation (PAM) Fluorometry and
Saturation Pulse Method: An overview. In G.C. Papageorgiou and
Govindjee
[eds],
Chlorophyll
Fluorescence:
A
Signature
of
Photosynthesis. Kluwer Academic Publishers.
Serôdio, J. 2003. A chlorophyll fluorescence index to estimate short-term
rates of photosynthesis by intertidal microphytobenthos. J. Phycol.
39, 33–46.
Serôdio, J., J.M. da Silva and F. Catarino. 1997. Nondestructive tracing of
migratory rhythms of intertidal benthic microalgae using in vivo
chlorophyll a fluorescence. J. Phycol. 33, 542-553.
Serôdio, J., P. Cartaxana, H. Coelho and S. Vieira. 2009. Effects of
chlorophyll fluorescence on the estimation of microphytobenthos
biomass using spectral reflectance indices. Remote Sensing of
Environment 113, 1760-1768.
Shinkarev, V. 2004. Photosystem II: Oxygen Evolution and Chlorophyll a
Fluorescence Induced by Multiple Flashes. In: Papageorgiou, G.C and
Govindjee
(Eds),
Chlorophyll
a
fluorescence:
A
signature
of
photosynthesis. Advances in Photosynthesis and Respiration, vol 19.
Springer, The Netherlands, pp. 197-229.
210
References
Sicko-Goad, L., Stoermer, E.F. and Ladewski, B.G. 1977. A morphometric
method
for
correcting
phytoplankton
cell
volume
estimates.
Protoplasma 93, 147-163.
Smith, J.L., G.L. Boyer & P.V. Zimba 2008. A review of cyanobacterial
odorous
and
bioactive
metabolites:
Impacts
and
management
alternatives in aquaculture. Aquaculture 280, 5-20.
Smith, V.H. 1998. Cultural eutrophication of inland, estuarine, and
coastal waters. In: Pace, M.L, Groffman, P.M. (Eds.), Successes,
Limitations and Frontiers in Ecosystem Science. Springer, New York,
pp. 7-49.
Smith, V.H., Tilman, G.D. and Nekola, J.C. 1999. Eutrophication: impacts
of excess nutrient inputs on freshwater, marine, and terrestrial
ecosystems. Environmental Pollution 100, 179-196.
Sokal, R.R. and Rohlf, F. 1995. Biometry: the principles and practice of
statistics in biological research, 3rd edn, W.H. Freeman, New York,
USA.
Sommer U., Gliwicz, Z.M., Lampert, W. and Duncan, A. 1986. The PEGmodel of seasonal succession of planktonic events in fresh waters.
Arch. Hydrobiol. 106, 433-471.
Sommerhäuser, M., Robert, S., Birk, S., Hering, D., Moog, O., Stubauer,
I. and Ofenböck, T. 2003. Activity 1.1.6 “developing the typology of
surface waters and defining the relevant reference conditions”. Final
Report. UNDP/GEF Danube Regional Project. http://www.undpdrp.org/pdf/1.1_River%20Basin%20Management%20%20Phase%201/1.1_UNDP-DRP_Typology%20of%20SW_116_fr.pdf.
Stevenson, R.J., Rier, S.T., Riseng, C.M., Schultz, R.E. and Wiley, M.J.
2006. Comparing effects of nutrients on algal biomass in streams in
two regions with different disturbance regimes and with applications
for developing nutrient criteria. Hydrobiologia 561(149–165).
Stewart, I., Webb, P.M., Schluter, P.J. and Shaw, G.R. 2006. Recreational
and occupational field exposure to freshwater cyanobacteria – a
review of anecdotal and case reports, epidemiological studies and the
challenges for epidemiologic assessment. Environmental Health: A
Global Access Science Source, 5 (6), 11-13.
211
References
Stone, D. and Bress, W. 2007. Addressing public health risks of
cyanobacteria in recreational freshwaters: The Oregon and Vermont
framework, Integrated
Environmental Assessment and Management
3, 137-143.
Stеhl-Delbanco, A., Hansson, L.-A. and Gyllstram, M. 2003. Recruitment
of resting stages may induce blooms of Microcystis at low N:P ratios,
Journal of Plankton Research 25 (9), 1099-1106.
Sun, J. and Liu, D. 2002. Geometric models for calculating cell biovolume
and surface area for phytoplankton. Journal of Plankton Research 25
(11), 1331-1346.
Tarczynska, M.; Romanowska-Duda, Z.; Jurczak, T.; Zalewski, M. 2001.
Toxic cyanobacterial blooms in drinking water reservoir – causes,
consequences and management strategy. Water Sci. Technol. Wat.
Sup. 1, 237-246.
Taylor, R. & Fletcher, R. 1999. Cryopreservation of eukaryotic algae – a
review of methodologies. Journal of Applied Phycology 10, 481-501.
Utermöhl,
H.
1958.
Zur
Vervollkommung
der
quantitativen
Phytoplankton-Methodik. Mitt. Int. Ver. Limnol. 9, 1-39.
Vadrucci, M.R, Cabrini, M and Basset, A. 2007. Biovolume determination
of
phytoplankton
guilds
in
transitional
water
ecosystems
of
Mediterranean Ecoregion. Transitional Water Bulletin 2, 83-102.
Vincent, R.K.; Qin, X.; Mc Kay, R.M.L.; Miner, J.; Czajkowski, K.; Savino,
J.; Bridgeman, T. 2004. Phycocyanin detection from LANDSAT TM
data for mapping cyanobacterial blooms in Lake Erie. Remote Sens.
Environ. 89, 381-392.
Wagenvoort, A.J. 2000. Biovolume of phytoplankton and zooplankton,
Evaluation
zooplankton
of
biovolume
from
1997
measurements
to
2000
and
of
the
phytoplankton
connection
and
between
BACCHUS and (Eco)Lims. WBB-report aw/aba 2001-35118. N.V.
Waterwinningbedrijf Brabantse Biesbosch, Werkendam. (in Dutch).
Wagenvoort, AJ & Lambregts-van de Clundert, FE. 1993. Efficiëntie van
de bovenstaande vloeistof bij de kwantitatieve fytoplankton analyse,
volgens K-WKB-W 3005. WBB rapportnummer hk/aw/sta 93-078. N.V.
WBB. Werkendam. (In Dutch).
212
References
Wagenvoort, AJ, Carpentier, CJ & Ketelaars, HAM. 1997. Eerste evaluatie
van de routinematige fytoplankton biovolumebepaling met behulp van
BACCHUS.
WBB
rapportnummer
aw/aba
97-35104.
N.V.
WBB.
Watras, C.J.; Baker, A.L. 1988. Detection of planktonic cyanobacteria by
tandem in vivo fluorometry. Hydrobiol. 169, 77-84.
WBB. 1997a. Kwantitatieve bepaling van de zoöplanktonsamenstelling in
water. Intern voorschrift K‑WKB-W 3004, versie 6. N.V. WBB.
WBB. 1997b. Kwantitatieve bepaling van de fytoplanktonsamenstelling in
water. Intern voorschrift K‑WKB-W 3005, versie 5. N.V. WBB.
Welch, E.B., Jacoby, J.M., Horner, R.R. and Seeley, M.R. 1988. Nuisance
bioamss levels of periphytioc algae in streams. Hydrobiologia 157,
161-168.
Welch, E.B., Quinn, J.M. and Hickey, C.W. 1992. Periphyton biomass
related to point-source nutrient enrichment in seven new zealand
streams. Wat. Res. 26(5), 669-675.
Wetzel, R.G. 2001. Limnology: Lake and River Ecosystems, 3 rd edition.
Academic Press, USA, pp. 142-150.
WHO. 1996. Guidelines for Drinking Water Quality, vol. 2, second ed.
Health Criteria and Other Supporting Information. World Health
Organization, Geneva, Switzerland, pp. 940–949 and pp. 281–283 in
1998 addendum to Volume 2.
Wilhelm, C., Rudolph, I. and Renner, W. 1991. A quantitative method
based on HPLC-aided pigment analysis to monitor structure and
dynamics of the phytoplankton assemblage - A study from Lake
Meerfelder Maar (Eifel, Germany). Arch. Hydrobiol. 121, 21-35.
Wood, S., Williams, S.T., White, W.R. & Jones, F. 1983. Factors
influencing
geosmin
production
by
a
streptomycete
and
their
relevance to the occurrence of earthy taints in reservoirs. Wat. Sci.
Tech. 15, 191-198.
Young, W.F., Horth, H., Crane, R., Ogden, T. and Arnott, M. 1996. Taste
and odour
threshold concentrations
of potential potable
water
contaminants. Water Research 30 (2), 331-340.
213
References
Zalewski,
M.;
Wagner-Lotkowska,
I.;
Tarczynska,
M.
2000.
Ecohydrological approaches to the elimination of toxic algal blooms in
the a lowland reservoir. Verh. Int. Ver. Theor. Angew. Limnol. 27,
3176-3183.
Zamyadi, A., McQuaid, N., Prйvost, M. and Dorner, S. 2012. Monitoring
of potentially toxic cyanobacteria using an online multi-probe in
drinking water sources, Journal of Environmental Monitoring, 14 (2),
579-588.
214
Acknowledgements
13.
Acknowledgements
In this section, I would like to express my gratitude to the people who
have supported and supervised me throughout this undertaking. I started
my PhD when working at Kiwa Water Research (now KWR Watercycle
Research Institute), the Netherlands. Ron van Megen, the director of the
institute, asked me about my plans...; my answer to do a PhD even
surprised myself. Adriana Hulsmann needed only a very short time to find
me a supervisor at Masaryk University of Brno, a place where I would not
have thought to look myself for PhD-support.
However, internal supervision of this project within KWR turned out to be
less easily organised. Arthur Meuleman started out as my supervisor, but
then decided to change jobs and leave the institute; thus he passed this
task to Marc Balemans. When Marc left as well, Peter Hesen took over for
a short time, until he also took another job outside the institute. By the
time Gertjan Zwolsman took over, support for my PhD research topic had
dwindled within the institute, and thus I decided to leave the institute
myself and continue as self-employed/an entrepreneur.
I thank you all for your efforts to support me, however short a time it may
have been.
Many colleagues at KWR came to my aid when I needed them most.
Willem Koerselman organised a brainstorm session with me to come up
with alternative methods for sampling benthic algae. He mobilised some
25 colleagues to spend their lunch break and some free time to come up
with fantastic ideas to improve sampling procedures. Thank you all so
much!
Additionally, many of you were always interested to hear how I was doing,
and supported me when I needed it. It felt like a warm bath after a long
and cold day. Especially after my husband died, you all helped me get
back on my feet and cotinue the work. In that respect, I particularly want
to thank Jan Willem Kooiman and Sonja de Goede: you were there for me,
listened and offered advice, and helped make my return to work much
easier than I thought it was going to be.
The first chapters of my thesis cover some of my earlier work at the Water
Storage Company Brabantse Biesbosch. I had a great time at the biology
215
Acknowledgements
laboratory, and the development of the biovolume analysis method would
not have been so successful without the contributions of Karina PikaarSchoonen,
Ronald
Sperber,
Sandra
Danuhyarso-Novemsia,
Francien
Lambregts-van de Clundert, Ger-An de Jonge-Pinkster, Meta Frank, Sonja
de Goede and Arco Wagenvoort, and the support of Angelique Baardse
and all the others in the lab. In later years, Arco also helped me
tremendously with practical advice and graphics skills.
Henk Ketelaars was one of the very few persons who has been with me
from the beginning (to the end). Supervisors have come and gone but he
has always been there. He has always been available for discussions to
help come up with solutions when I was completely stuck. All the pieces of
the puzzle were right in front of me, but I did not see the whole picture
until he suggested I change the title of my thesis. With that, he showed
that the simplest solutions are usually the best, since only then do all the
pieces fall into place. I cannot thank you enough!
After being self-employed for two years, Joep van den Broeke and I
teamed up to establish Benten Water Solutions. Joep not only was a
business partner and colleague, he also provided comments to some of
my articles and support to finalise my thesis.
Jarek Mrowczynski’s internship at KWR ultimately resulted in the article on
which Chapter 6 of this thesis is based. If it hadn’t been for the Polish
taxi-driver in Kiel (Germany), we probably would not have met at bbe’s
Aqualife workshop, and this article would not have been part of my thesis
now. Katarzyna Izydorczyk and colleagues at the University of Łodz,
Poland, did a considerable amount of work to get the article published.
Another person I met at the bbe workshops was Hazem Kalaji of Warsaw
University of Life Sciences, Poland. Whether plant leaves or algae, the
photosynthetic principles are the same, and I very much enjoyed the talks
and discussions, especially the one where he compared measurements on
the photosynthetic process to buying a second-hand car! I look forward to
continuing to work with you outside the context of this thesis.
216
Acknowledgements
Christian Moldaenke of bbe Moldaenke, Germany, kindly provided me with
instrumentation and support to perform the field experiments as described
in Chapters
8, 9
and 10. The
discussions regarding chlorophyll-
fluorescence in general and the development of the BenthoTorch
instrument in particular helped me to shape my thoughts on chlorophyll
analysis using fluorescence-techniques. And although we do not always
seem to understand each other, I have learnt a lot from these discussions,
and I hope I was able to give you some helpful ideas, too. Many people at
bbe worked together to develop the first BenthoTorch prototype, and I am
proud that I was allowed to use it for my field experiments. Anna
Dahlhaus accompanied me to Denmark to perform field tests with it in the
Odense River, as part of the Aquarehab Project, and also Hanno Dahlhaus
assisted with the laboratory tests and interpretation of the results in Kiel. I
also thank other bbe staff for advice and support.
The results presented in Chapter 10 could not have been collected without
the help and support of Jaroslav Slobodník and Jarmila Makovinská.
Jaroslav made it possible to add phytobenthos biomass analyses to the
parameter list for the second Joint Danube Survey, and Jarmila was kind
enough to perform the measurements during the six-week trip covering a
3,000 km stretch of the Danube River. This resulted in an overwhelming
amount of data on phytobenthos biomass in relation to every imaginable
biotic and abiotic parameter. The results of the third Joint Danube Survey
of 2013 could not be included in the thesis, but these datasets have
provided me with a lot of insight into the factors which influence
phytobenthic biomass development.
Being a PhD student at the University of Brno while at the same time
having a job at a research institute in the Netherlands or being selfemployed was not easy to combine. The one person to experience this
first hand was my supervisor Blahos Maršálek. I remember us meeting at
the offices of the immigration police in Brno, to efficiently use the waiting
hours
by
having
scientific
discussions
on
the
experiments
to
be
performed. You and your wife Eliška Maršálková tried your very best to
make me feel at home in Brno, and help me find my way around the
university and the city. I realise I must have been a difficult student to
supervise, especially since I was not around most of the time. Thank you
217
Acknowledgements
for taking me on as your student, despite the complicated conditions. I
very much appreciate your support after my husband died, to allow me to
continue my PhD when I felt ready for it. And although I was not there for
any length of time, the colleagues at the Department of Experimental
Phycology kindly welcomed me into their group. I very much enjoyed the
field work and cooperation with Jakub Gregor, Katka Sukácová, Markéta
Kozáková and Jana Veselá. I also thank Ludek Blaha for his logistical
support.
My second supervisor, Nick van de Giesen, joined when my PhD study had
been underway for quite a few years already. Jan Vreeburg pointed me in
his direction, and with Nick I found support for my scientific work a bit
closer to home. And although we did not meet on a regular basis, the
discussions we had regarding my work helped me very much to focus on
the important issues, and not lose myself in details so I could work
towards a successful completion. Thank you for this push in the right
direction.
I am fortunate to be the daughter of a couple of wonderful parents, who
have always stimulated me to use my talents to the best of my abilities.
Starting from this solid basis, I was able to broaden my horizons and
become the person I am today. After I had left the nest, you continued to
support me in everything I did, including this PhD study. My father also
helped me in a more practical way, by welding the monopods I needed for
my field experiments in the Odense River.
My dear sister Anja, I am proud that your photograph is on the cover of
my thesis. Aside from my study subject, it illustrates our bond as sisters,
despite our different interests.
For many years, I have been fortunate to be able to share many good
times with very good friends. Esther, Sonja, Ivana, Niels, Müge, Marco
and Resh, you guys have been around since forever, and I have many
fond memories of our getting together, very often to share food and wine.
And when Colin entered my life, he brought with him new friends,
including Dennis & Laura, Tommy & Sabs, who very soon found a place in
my heart as well. Thank you for listening to and supporting me throughout
this adventure.
218
Acknowledgements
And last but not least, my beloved Colin. Without you, I would not be
where I am today. You helped me through probably the most difficult
period in my life, and stayed to share in the good times that followed. I
look forward to continuing this journey together with you. We went
through my thesis word for word while you corrected the English, and I
never knew how important a comma can be for the meaning of a
sentence!
Corina Carpentier
April 2014
219

Corina`s thesis: “The Carpet of the Sun”

Transkript

Podobné dokumenty

chlorophyll fluorescence as a bioindicator of effects on growth in

earproblems

Discussion Paper

Krill - Center for Integrated Marine Technology