Corina`s thesis: “The Carpet of the Sun”
Transkript
Corina`s thesis: “The Carpet of the Sun”
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 6.2. Materials and methods 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 7.2. Materials and methods 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 General introduction 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 General introduction 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 General introduction 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 General introduction 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 Section 1: Phytoplankton biomass 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 Chapter 3: BACCHUS biovolume 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 Section 1: Phytoplankton biomass 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 Chapter 3: BACCHUS biovolume 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 Section 1: Phytoplankton biomass 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 Chapter 3: BACCHUS biovolume 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 Section 1: Phytoplankton biomass 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 Chapter 3: BACCHUS biovolume 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 Section 1: Phytoplankton biomass 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 Chapter 3: BACCHUS biovolume 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 Section 1: Phytoplankton biomass 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) Chapter 3: BACCHUS biovolume dx : diameter of a coenobium with x cells; x : number of cells in the coenobium. 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 : number of cells in the coenobium. 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 Section 1: Phytoplankton biomass 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 Chapter 3: BACCHUS biovolume 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 Section 1: Phytoplankton biomass 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 Chapter 3: BACCHUS biovolume 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 Section 1: Phytoplankton biomass 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 Carpentier et al., 1997 Carpentier et al., 1997 Rott, 1981 Rott, 1981 reference Chapter 3: BACCHUS biovolume 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 Michielsen & Ketelaars, 1994 Carpentier et al., 1997 Michielsen & Ketelaars, 1994 Carpentier et al., 1997 Rott, 1981 Wetzel & Likens, 1991 Rott, 1981 Rott, 1981 Rott, 1981 Carpentier et al., 1997 Rott, 1981 Rott, 1981 dit rapport Rott, 1981 Rott, 1981 Michielsen & Ketelaars, 1994 Carpentier et al., 1997 Michielsen & Ketelaars, 1994 Michielsen & Ketelaars, 1994 Michielsen & Ketelaars, 1994 Carpentier et al., 1997 Carpentier et al., 1997 Wetzel & Likens, 1991 Reynolds, 1984 Rott, 1981 Rott, 1981 Rott, 1981 Michielsen & Ketelaars, 1994 Michielsen & Ketelaars, 1994 reference Section 1: Phytoplankton biomass Chapter 3: BACCHUS biovolume 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 Section 1: Phytoplankton biomass 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 Chapter 3: BACCHUS biovolume 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 Section 1: Phytoplankton biomass 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 n1 (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 Chapter 3: BACCHUS biovolume 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 Section 1: Phytoplankton biomass 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 Chapter 3: BACCHUS biovolume 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 Section 1: Phytoplankton biomass 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 Chapter 3: BACCHUS biovolume 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 Section 1: Phytoplankton biomass 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 Chapter 3: BACCHUS biovolume 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 Section 1: Phytoplankton biomass Figure 3.17 64 Continued Chapter 3: BACCHUS biovolume 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 Section 1: Phytoplankton biomass 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 Chapter 3: BACCHUS biovolume (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 Section 1: Phytoplankton biomass 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 Chapter 3: BACCHUS biovolume 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 Section 1: Phytoplankton biomass 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 phytoplankton chlorophyll-a 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 Section 1: Phytoplankton biomass 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 Chapter 4. In vivo absorption 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 Section 1: Phytoplankton biomass 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 Chapter 4. In vivo absorption 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 Section 1: Phytoplankton biomass 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 Chapter 4. In vivo absorption 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 Section 1: Phytoplankton biomass 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) amplitude Gauss Curve 2 (a2) amplitude Gauss Curve 3 (a3) calculated absorption coefficient amplitude Gauss Curve 4 (a4) half width curve 1 (nm) half width curve 2 (nm) half width curve 3 (nm) half width curve 4 (nm) peak wavelength curve 1 (λ1; nm) peak wavelength curve 2 (λ2; nm) peak wavelength curve 3 (λ3; nm) peak wavelength curve 4 (λ4; nm) χ2 Table 4.2 Chapter 4. In vivo absorption 79 Section 1: Phytoplankton biomass 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 Chapter 4. In vivo absorption 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 Section 1: Phytoplankton biomass 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 Chapter 4. In vivo absorption 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 amplitude Gauss Curve 1 (a1) amplitude Gauss Curve 2 (a2) amplitude Gauss Curve 3 (a3) calculated absorption coefficient amplitude Gauss Curve 4 (a4) half width curve 1 (nm) half width curve 2 (nm) half width curve 3 (nm) half width curve 4 (nm) peak wavelength curve 1 (λ1; nm) peak wavelength curve 2 (λ2; nm) peak wavelength curve 3 (λ3; nm) peak wavelength curve 4 (λ4; nm) χ2 Table 4.5 Section 1: Phytoplankton biomass 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. Chapter 4. In vivo absorption 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 Section 1: Phytoplankton biomass 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 Chapter 4. In vivo absorption 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 Framework for cyanobacteria in drinking water resources by using the 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 Chapter 5. Introduction to fluorescence 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 Section 2: Cyanobacterial biomass 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 Chapter 5. Introduction to fluorescence 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 Section 2: Cyanobacterial biomass 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 Chapter 5. Introduction to fluorescence 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 Section 2: Cyanobacterial biomass 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 Chapter 5. Introduction to fluorescence 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 Section 2: Cyanobacterial biomass 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 Section 2: Cyanobacterial biomass 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 Chapter 6. Cyanobacteria Alert Level Framework 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. Materials and methods 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 Section 2: Cyanobacterial biomass 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 Chapter 6. Cyanobacteria Alert Level Framework 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 Section 2: Cyanobacterial biomass 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 Chapter 6. Cyanobacteria Alert Level Framework 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 Section 2: Cyanobacterial biomass 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 Chapter 6. Cyanobacteria Alert Level Framework 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 Section 2: Cyanobacterial biomass 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 Chapter 6. Cyanobacteria Alert Level Framework 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 Section 2: Cyanobacterial biomass 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 Chapter 6. Cyanobacteria Alert Level Framework 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 Section 2: Cyanobacterial biomass 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 Chapter 6. Cyanobacteria Alert Level Framework 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 Section 2: Cyanobacterial biomass 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 Chapter 6. Cyanobacteria Alert Level Framework 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 Section 2: Cyanobacterial biomass 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. Materials and methods 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 Chapter 7. Application of Alert Level Framework 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 Section 2: Cyanobacterial biomass 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 Chapter 7. Application of Alert Level Framework 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 Section 2: Cyanobacterial biomass 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 Chapter 7. Application of Alert Level Framework 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 Section 2: Cyanobacterial biomass 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 Chapter 8. Introduction to phytobenthos 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 Section 3: Phytobenthos biomass 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 Chapter 8. Introduction to phytobenthos 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 Section 3: Phytobenthos biomass 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 Chapter 8. Introduction to phytobenthos 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) Section 3: Phytobenthos biomass 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 Chapter 8. Introduction to phytobenthos 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 Section 3: Phytobenthos biomass 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. Chapter 8. Introduction to phytobenthos 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 Section 3: Phytobenthos biomass A. 140 Chapter 8. Introduction to phytobenthos 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 Section 3: Phytobenthos biomass 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 Chapter 8. Introduction to phytobenthos (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 Section 3: Phytobenthos biomass 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 Section 3: Phytobenthos biomass 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 Chapter 9. Substratum reflection 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 Section 3: Phytobenthos biomass 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 Chapter 9. Substratum reflection 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 Section 3: Phytobenthos biomass 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 Chapter 9. Substratum reflection 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 Section 3: Phytobenthos biomass 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 Chapter 9. Substratum reflection 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 Section 3: Phytobenthos biomass 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 Chapter 9. Substratum reflection 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 Section 3: Phytobenthos biomass 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 Chapter 9. Substratum reflection 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 Section 3: Phytobenthos biomass fluorescence (r.u. x 1000) 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 algal density (µg/cm²) B. fluorescence (r.u. x 1000) 900 470nm LED 800 610nm LED 700 525nm LED 600 500 400 300 200 100 0 0 1 2 3 4 5 6 7 algal density (µg/cm²) 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. Chapter 9. Substratum reflection 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 Section 3: Phytobenthos biomass 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 Chapter 9. Substratum reflection A. B. 161 Section 3: Phytobenthos biomass 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) Chapter 9. Substratum reflection 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 Section 3: Phytobenthos biomass 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 Chapter 9. Substratum reflection 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 Section 3: Phytobenthos biomass 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 Chapter 9. Substratum reflection 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 Section 3: Phytobenthos biomass 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 Chapter 9. Substratum reflection 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 Section 3: Phytobenthos biomass 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 Chapter 9. Substratum reflection 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 Section 3: Phytobenthos biomass 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 Section 3: Phytobenthos biomass 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 Chapter 10. JDS2 field example 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 Section 3: Phytobenthos biomass 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). Chapter 10. JDS2 field example 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 Section 3: Phytobenthos biomass 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 Chapter 10. JDS2 field example 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 Section 3: Phytobenthos biomass 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 Chapter 10. JDS2 field example 181 Section 3: Phytobenthos biomass 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 Chapter 10. JDS2 field example 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 Section 3: Phytobenthos biomass 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 Chapter 10. JDS2 field example 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) Phytobenthos biomass (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 Invertebrate grazers (ind./m²) 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). Phytobenthos biomass (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 Invertebrate grazers (ind./m²) 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 Section 3: Phytobenthos biomass 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 Chapter 10. JDS2 field example 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 Section 3: Phytobenthos biomass 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 General discussion & conclusions 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 phytoplankton chlorophyll-a 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 General discussion & conclusions 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 General discussion & conclusions 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 General discussion & conclusions 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 Framework for cyanobacteria in drinking water resources by using the 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. Werkendam. (In Dutch). 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. Werkendam. (In Dutch). WBB. 1997b. Kwantitatieve bepaling van de fytoplanktonsamenstelling in water. Intern voorschrift K‑WKB-W 3005, versie 5. N.V. WBB. Werkendam. (In Dutch). 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