Particles matter - ibed.uva.nlibed.uva.nl/.../map-1/2013-bram-mulling---particles-matter.pdf ·...

Particles matter

Transformation of suspended particles in constructed wetlands

ACADEMISCH PROEFSCHRIFT

ter verkrijging van de graad van doctoraan de Universiteit van Amsterdamop gezag van de Rector Magnificus

prof. dr. D.C. van den Boomten overstaan van een door het college voor promoties ingestelde commissie,

in het openbaar te verdedigen in de Agnietenkapelop woensdag 3 juli 2013, te 14:00 uur

doorBram Theodorus Maria Mulling

geboren te Doetinchem

Promotie commissie

Promotor: prof. dr. W. Admiraal

Co-promotor: dr. H.G. van der Geest

Overige leden: prof. dr. ir. D.P.L. Rousseau prof. dr. J.T.A. Verhoeven prof. dr. G.J. Medema prof. dr. K. Kalbitz prof. dr. W.P. de Voogt

Faculteit der Natuurwetenschappen, Wiskunde en Informatica

This research was supported by Stichting Waternet and the Foundation for Applied Water Research (STOWA), Witteveen+Bos and Wetterskip Fryslân. The study was conducted at the department of Aquatic Ecology and Ecotoxicology (AEE), Institute for Biodiversity and Ecosystem Dynamics (IBED), University of Amsterdam (UvA).

Cover design by Bram T.M. MullingPrinted by Ipskamp Drukkers BV

ISBN: 978-94-91407-10-9Copyright © 2013 by Bram T.M. Mulling

Contents

Chapter 1 General Introduction

Chapter 2 Physical and biological changes of suspended particles in constructed wetlands

Chapter 3 Changes in the planktonic microbial community during residence in a constructed wetland

Chapter 4 Processes removing faecal indicator organisms in constructed wetlands

Chapter 5 Trapping of bacterial cells and latex micro-spheres in natural and cultured phototrophic biofilms

Chapter 6 Suspended particle and pathogen peak discharge buffering by a surface flow constructed wetland

Chapter 7 Concluding remarks

Summary

Samenvatting

Acknowledgements

Curriculum vitae

7

21

41

61

87

103

125

139

145

151

155

General introduction

Chapter 1

8

Chapter 1

Discharge of municipal wastewater is one of the major factors threatening aquatic ecosystems across the globe. Originating from urban areas, municipal wastewater consists of a complex and variable mixture of constituents of anthropogenic origin, including pharmaceuticals, surfactants, emulsions, pesticides, herbicides, self-care residues, drugs, hormones, nutrients, plankton, pathogens and organic and inorganic particles (Kadlec and Wallace 2008; Tchobanoglous 2004). The majority of these constituents are removed by regular wastewater treatment, but discharge of treated municipal wastewater can still have profound impacts on receiving surface waters (Holeton et al. 2011). One of the key factors causing this impact is often overlooked: the discharge of allochthonous suspended matter, changing many processes and functions in the receiving aquatic ecosystems (Holeton et al. 2011; Kalff 2002). Suspended matter in treated wastewater consists of a complex mixture of organic and inorganic particles of anthropogenic origin (see box 1), and is amongst others characterized by a distinct bacterial community (Kadlec and Wallace 2008; Tchobanoglous 2004; Seviour and Nielsen 2010). To mitigate the effects of treated wastewater discharge into receiving surface waters additional polishing is currently debated. The purification capacity of wetlands (see box 2) has long been recognized and man-made constructed wetlands (CWs) are used in a wide variety of applications to improve water quality. Since the first attempts to use CWs for water quality improvements of untreated wastewater in the early 1950s, the development and use of CWs for wastewater treatment has spread across the world (Sundaravadivel and Vigneswaran 2001; Kadlec and Wallace 2008; Vymazal 2005). CWs are designed to optimize several naturally occurring physical, chemical and biological processes, like sedimentation and microbial degradation in order to reduce concentrations of the harmful constituents in (treated) wastewater (Kadlec and Wallace 2008). Many studies have demonstrated that CWs

Box 1: Suspended matter classification

Suspended matter is defined as the filterable matter in the water column (Wotton 1994). In general suspended matter is a complex mixture of organic and inorganic particles with different origins, natures, and properties, including size, form, density, specific surface area, surface charge, binding capacity and chemical composition. Inorganic particles include mineral clays, sand, silt, manmade materials and metal oxides. Organic particles or particulate organic matter (POM) contain both living organisms (plankton), which can be further specified by taxonomic identification, and dead biological matter (detritus) (Fig. 1.3).

0.0001

0.0010

0.0100

0.1000

1.0000

Sand

Silt

Clay

Sand

Silt

Clay

0.0001

0.0010

0.0100

0.1000

1.0000

Mesoplankton

Microplankton

Nanoplankton

Picoplankton

Femtoplankton

Mesoplankton

Microplankton

Nanoplankton

Picoplankton

Femtoplankton0.0001

0.0010

0.0100

0.1000

1.0000

CPOM

FPOM

DOM

CPOM

FPOM

DOM

1mm

100µm

10µm

1µm

0.1µmSu

spen

ded

part

icle

s

Fig. 1.3 The composition of suspended particles and different ways of characterizing suspended particles. DOM: dissolved organic matter; FPOM: fine particular organic matter; CPOM: coarse particular organic matter.


9

indeed significantly reduce concentrations of pollutants like, for example, nitrogen, phosphorus, heavy metals, suspended matter and pathogens (Kadlec and Wallace 2008; Zhang et al. 2011; Vymazal 1996; Vymazal 2007; Fisher and Acreman 2004; Cameron et al. 2003; Reinoso et al. 2008; Vidales-Contreras et al. 2006; Vymazal 2005). However, while some studies report removal efficiencies up to 90% for suspended matter (Kadlec and Wallace 2008), other studies show very limited reduction or even addition of suspended matter (van den Boomen and Kampf 2012; van den Boomen et al. 2012). These large differences seem to be related to particle inflow concentrations, with CWs receiving high inflow concentrations generally showing high removal of suspended particles while CWs with low concentrations of inflowing particles have been observed to produce equal or slightly increased outflow concentrations (Ghermandi et al. 2007). However, most of these studies on suspended matter in wetlands focus on bulk measurements of particle concentrations and the understanding of the underlying processes involved in changing the physical and biological composition of suspended particles during residence in CWs is limited. A more detailed analyses of particle composition and dynamics in CWs is therefore needed to optimize the use of CWs in mitigating environmental effects of suspended matter in (treated) wastewater.

Suspended particles in wetland ecosystems Suspended particles have several important functions in aquatic ecosystems and are involved in many physical, chemical and biological processes (Fig. 1.1). Particles not only have direct effects like a decrease

Box 2: Wetlands and ecosystem services

Wetlands are ecosystems with a water table at or near the land surface and function as transitional zones between terrestrial and aquatic environments (Fig. 1.4a) (Keddy 2010). Wetlands are characterized by the specific and characteristic vegetation adapted to these environmental conditions and are recognized as highly productive ecosystems supporting a high diversity of plants and animal species in salty, brackish and freshwater environments (Keddy 2010). Several types of wetlands are defined including marshes, swamps, bogs, temporal waters and fens. Wetlands provide several ecosystems services including flood control, climate change mitigation and adaption, water quality improvement, cultural and recreation values and are reserves of biodiversity (Fig. 1.4b)(Keddy 2010).

b)

a)

Fig. 1.4 a) Schematic representation of a wetland and some of the provided ecosystems services (http://myweb.rollins.edu) and b) Ecosystem services provided by wetlands at local and global scales (Holcova et al. 2009) .

10

Chapter 1

in light penetration, influencing primary production and phytoplankton community composition (Kalff 2002), but are also involved in the more complex cycling of energy and nutrients. Organic particles contain carbon, nitrogen, phosphorus and other essential nutrients for living organisms and serve as a food source for many heterotrophic organisms in the aquatic food web (Wetzel 2001; Kalff 2002). Especially for heterotrophic bacteria living in the water column, suspended particles are a crucial source of energy and nutrients. In the process of particle degradation carried out by these heterotrophs, nutrients are released from the suspended matter by remineralisation thereby also fuelling primary production (Wotton 1994; Kalff 2002). Since decomposition of organic matter by heterotrophic bacteria is an oxygen demanding process, degradation of excessive amounts of organic particles can also lead to anoxic waters and sediments, resulting in mortality of higher organisms like macro-invertebrates and fish (Kalff 2002). In specific areas of (constructed) wetlands where water movement is low, suspended particles are also subjected to sedimentation providing a source of energy and nutrients for the benthic system where specialized consortia of bacteria, fungi and invertebrates are involved in similar processes (Kalff 2002). The central role of suspended particles in the above mentioned processes emerges from the availability of ecological micro-niches for bacteria and other organisms, thereby making these niches hotspots for structural and functional diversity in the water column (Grossart et al. 1998). In (treated) wastewater it is the formation of aggregates that makes particles function as a transport route for the removal of harmful bacteria and viruses from the water column by sequential adsorption and attachment, and sedimentation of the particles (Karim et al. 2004; Reinoso et al. 2008). In the same way suspended particles can purify the water column from dissolved contaminants like metals, pesticides, phosphorus and PCBs (Kalff 2002).

Suspended particles

Nutrient cycling

Energy

Water clarity

Oxygen concentration

Secondary production

Primary production

Viruses and bacteria removal

Contaminant removal

Light penetrationaggregation and deposition

Biomass deposition,

nutrient adsorptionfood resource

Food source

Fig. 1.1 Schematic overview of the linkage between suspended particles, ecosystem processes and water characteristics in wetlands.


11

There are many processes influencing the concentration, nature and type of suspended particles in wetland ecosystems. In general it is acknowledged that sedimentation, flocculation and resuspension are major processes influencing suspended particle concentrations (Keddy 2010). Sedimentation is the settling of particles onto the sediments by gravity and depends on characteristics of both the particles (size, mass, shape) and environmental conditions (current velocity, turbulence, water density) and is most effective for particles larger than 50µm (Marttinen et al. 2003) while small particles, especially low density organic particles, are often not subjected to substantial sedimentation. Particles can be resuspended from the sediments by (wind induced) turbulence, bioturbation or gas release from the sediment (Wotton 1994; Kalff 2002; Durako et al. 1982). Resuspension can strongly increase the suspended particle concentrations in the water column, depending on several environmental conditions such as water depth, wind speed or presence of aquatic vegetation (Kelderman et al. 2012; Wotton 1994; Kalff 2002). Flocculation is the processes in which coumpounds and particles form flocculates and aggregates and occurs when particles collide and stick together (adhesion) as a result of Brownian motion, shear, differential sedimentation, diffusive capture, surface coagulation, filtration, bacterial motility or biological capture (Wotton 1994). Flocculation increases the size and thereby the settling velocity of particles, and is therefore an important process in the sedimentation of smaller particles (Droppo et al. 2004). Flocculation can also increase the concentration of suspended particles by flocculation of smaller compounds into flocs large enough to be considered particular (Simon et al. 2002). The process of flocculation is dependent on many biological, chemical and physical characteristics of both the water and the particles, including particle concentration, surface properties, ionic concentrations, turbulence and environmental conditions (Droppo 1997). Being the opposite process of flocculation, disaggregation breaks up particles in smaller ones, thereby reducing sedimentation rates and even decreasing concentrations by creating smaller particles that are no longer considered as being part of the suspended particle pool. Dissolved compounds like dissolved organic matter (DOM), metals, phosphorus, calcium and sulphides may contribute to the dynamics of suspended particles in wetland ecosystems. DOM and dissolved inorganic compounds not only adsorb onto suspended particles (Johnson et al. 1994), thereby increasing the size and mass and influencing the settling rate of particles, but may also form precipitates (particles) during changes in environmental conditions including oxygen concentrations, pH, temperature and ion concentrations (Kadlec and Wallace 2008). Besides these primarily physical and chemical processes also several biological driven processes influence the concentration, nature and type of suspended particles. As described above, heterotrophic bacteria break down organic particles and convert particulate organic matter into dissolved organic matter and minerals (Wotton 1994; Kalff 2002), thereby reducing the size of large particles, affecting the settling rate and transforming dead organic matter into living bacterial biomass (Wotton 1994; Kalff 2002). Other heterotrophic bacteria

12

Chapter 1

use DOM rather than particulate organic matter (POM) and by doing so they transform dissolved organic matter (DOM) into bacterial biomass (POM) increasing the suspended matter concentration (Kalff 2002). In addition, autotrophic phytoplankton species use nutrients and solar energy to create biomass and can under favourable conditions strongly increase the suspended matter in aquatic ecosystems (Kalff 2002). Other organisms, like many planktonic zooplankton species such as Rotifera, Cladocera, Amoebazoa and Ciliophora are capable to ingest suspended particles and use them as a food source. These secondary producers transform suspended matter into their own biomass (that can still be considered as suspended matter) and increase particle sedimentation by compacting organic matter into dense faecal pellets (Wotton 1994). Moreover, the benthic components of wetland ecosystems might also play a key role in driving the dynamics of suspended particles in the water column. Attached filter feeders, like bivalves and specialized groups of Ciliophora and insect larvae, remove particles from suspension and convert them into biomass and deposit faecal pellets onto the sediment (Ruckelshaus et al. 1993; Chabaud et al. 2006). Benthic algae, often growing on substrates like for example macrophytes stems and forming dense biofilms (Woodruff et al. 1999; Romani et al. 2004), are known to actively and passively trap suspended particles into the biofilm which can then be used by the heterotrophic bacteria inside the biofilm as a food source (Kadlec and Wallace 2008; Balzer et al., 2010; Chabaud et al. 2006; Drury et al. 1993; Eisenmann, et al. 2001; Stott and Tanner 2005). Besides trapping of particles, biofilms also release particles into the water column, either actively for colonisation or passively by erosion (McDougald et al. 2012).

Sediments

Water

Suspended particles

Suspension-feeders

(biotransformation)Dissolution-absorbtion

Egestion, excretion, biodeposition

Decay, exudation, excretion

Diageneticremobilisation

Desorption, dissolution, deflocculation, degradation

Adsorption, precipitation, flocculation, heterotrophy

Precipitation Absorption, partitioning

Fig. 1.2 Schematic overview of processes influencing the concentration, nature and composition of suspended particles in aquatic ecosystems. (after Turner and Millward 2002)


13

Figure 1.2 summarizes the complexity and dynamics of the many processes influencing the behaviour of suspended particles in wetland ecosystems described above. Based on this schematic overview, it may become evident that CWs may hold great potential to change the nature and type of the anthropogenic particles in (treated) wastewater. Thereby transforming the particles and reducing the impact of wastewater discharge on receiving surface waters. However, this capacity strongly depends on the design of the CW since many of these processes result from complex interactions between the suspended particles and the environmental conditions. The lack of research on particle specific processes in CWs, disregarding particle nature and type, obstructs a detailed analyses of these complex interactions. This study was set out to fill up this knowledge gap by providing mechanistic insight in the functioning of CWs regarding suspended particles. The overall aim of this study was to assess the importance of physical, chemical and biological processes modifying the fate and behaviour of suspended particles in surface flow CWs.

To this purpose the following objectives were formulated:- To identify changes in the nature and composition of suspended particles in treated municipal wastewater during residence in surface flow constructed wetlands.- To weigh the importance of individual environmental processes occurring in wetland ecosystems on the fate, behaviour and dynamics of suspended particles. - To assess the prospect of constructed wetlands in mitigating effects and associated risks of municipal wastewater.

Outline of the thesis The first step in this study required the quantification and identification of suspended matter in a surface flow CW in Grou (see box 3). To this purpose chapter 2 focussed on analysing monitoring data of suspended matter concentrations during multiple years complemented with data from short term monitoring experiments in which we studied changes in physical and biological nature and type of suspended particles. After studying these major changes, chapter 3 focussed on changes in the microbial community during residence in the CW. Changes in bacterial abundance, average bacterial size, community composition, community metabolic activity and community metabolic functional diversity were analysed in treated wastewater, effluent of the CW as well as in reference sites. Pathogenic organisms, present in high numbers in treated wastewater, are of major interest since they pose public health risks and chapter 4 therefore focuses on changes in the numbers of faecal indicator organisms and analyses the contribution of processes involved in the removal of pathogens. Bacterial, viral and protozoan faecal indicator organisms numbers were monitored during a one year period on a monthly basis throughout the CW. From this data removal efficiencies were calculated for the different compartments within the CW and the relative importance of the different removal processes were described based on short term field and laboratory experiments.

14

Chapter 1

The role of biofilms is often considered to be of major importance reducing the suspended particle concentrations in CWs, but the specific effects of biofilms are not well studied and difficult to test in full scale CWs. Chapter 5 therefore describes laboratory experiments in which the effects of biofilm composition on the trapping of particles and bacteria were studied in relation to particle size. To this purpose several field grown natural biofilms and cultured mono-specific biofilms were incubated in laboratory aquaria and trapping of different particles was monitored, enabling us to quantify capture efficiencies for the different biofilm compositions and particle sizes. Chapter 2, 3 and 4 all studied particle dynamics in CWs under normal operating conditions, but peaks are occasionally discharged into CWs and could exceed the purification capacity of the CW. These accidental peak discharges are difficult to predict and chapter 6 discusses the behaviour of an artificially induced suspended matter peak discharge into the CW. After discharge of the peak we monitored the peak at several positions in the CW and analysed a multitude of parameters to identify the behaviour of different components of the suspended particles, the location and possible important processes affecting the peak during residence in the CW. The concluding remarks in chapter 7 discuss the main findings of this thesis regarding the dynamics of suspended particles in wetland ecosystems and changes in the nature and composition of suspended particles during residence in CWs. I will try to weigh the importance of individual environmental processes in wetland ecosystems that determine the fate, behaviour and dynamics of suspended particles in CWs. I will also construct an operational model on suspended particle dynamics in wetland ecosystems. This model may ultimately lead to optimization of CWs for specific applications. Based on the results I will conclude this chapter with a review off the options for applying CWs for achieving good ecological quality of surface waters.

Box 3: Study area

All field studies are carried out in a full scale surface flow constructed wetland (CW) located in Grou, The Netherlands (N53 05.535 E5 49.050). This constructed wetland was chosen for this study, because the system was frequently monitored since its construction (2006) and the absence of plans for major changes or maintenance work during the period of this study to both the CW or the wastwater treatment plant (WWTP). The CW receives secondary-treated municipal (mainly domestic) wastewater at a constant hydraulic loading of 1200 m3 day-1 (Table 1.1). After inflow from a settlement tank, the water flows through the CW consisting of a series of three unvegetated ponds and four parallel reed beds (Fig. 1.5). At the end of the reed beds the water is pumped into an ecological buffer zone which is in open connection with the receiving surface water (channel). The unvegetated ponds have an average depth of 1.35m, volumes between 360 and 440 m3 each and total hydraulic retention time (HRT) of 17.9 h. The reed beds are covered with Phragmites australis. They have an average water depth of 40 cm, approximate volume of 443 m3 and an average HRT of 23.6 h.


15

Box 3: Study area (continues)

1

2

2

2

3 3 3 3

4

5

0 25 50 m

0

a)

b)1 2 2 2

43

Fig. 1.5 a) Schematic representation of the study area with the wastewater treatment plant (0) including a settlement tank (1) which discharges treated municipal wastewater into a constructed wetland consisting of unvegetated ponds (2), reed beds (P. australis) (3) connected to an ecological buffer zone (4) between the constructed wetland and the receiving channel (5), b) schematic cross section of the constructed wetland.

Table .1 Dimensions and hydraulic retention time (HRT) of constructed wetland, Aqualân in Grou, The Netherlands. The length, width and volume of the ponds were manually determined and calculated, the length, width and volume of the reed beds were calculated from the construction blueprints. The HRT’s were determined by a tracer experiment.

Ponds Reed beds Total

1 2 3 Total Bed 1-4 Total

Length (m) 55 55 55 165 110 110

Average width (m) 7.6 8.1 8.1 7.9 11.5 46

Average depth (m) 1.34 1.31 1.43 1.35 0.4 0.4

Surface area (m2) 418 446 446 1304 1265 5060 6364

Volume (m3) 362 388 441 1191 443 1771 3552

Hydraulic loading (m3 day-1) 1200 1200 1200 1200 300 1200 1200

HRT (h) 17.9 23.6 41.5

16

Chapter 1

References

Balze M, Witt N, Flemming H- and Wingender J (2010). Faecal indicator bacteria in river biofilms. Water Science and Technology 61 (5), 1105-1111.

Cameron K, Madramootoo C, Crolla A and Kinsley C (2003). Pollutant removal from municipal sewage lagoon effluents with a free-surface wetland. Water Research 37 (12), 2803-2812.

Chabaud S, Andres Y, Lakel A and Le Cloirec P (2006). Bacteria removal in septic effluent: Influence of biofilm and protozoa. Water Research 40 (16), 3109-3114.

Droppo IG, Leppard GG, Flannigan DT and Liss SN (1997). The freshwater floc: A functional relationship of water and organic and inorganic floc constituents affecting suspended sediment properties. Water Air and Soil Pollution 99 (1-4), 43-53.

Droppo IG (2004). Structural controls on floc strength and transport. Canadian Journal of Civil Engineering 31 (4), 569-578.

Drury WJ, Characklis WG and Stewart PS (1993). Interactions of 1mu-m latex-particles with Pseudomonas-aeruginosa biofilms. Water Research 27 (7), 1119-1126.

Durako MJ, Medlyn RA and Moffler MD (1982). Particulate Matter Resuspension Via Metabolically Produced Gas-Bubbles from Benthic Estuarine Microalgae Communities. Limnology and Oceanography 27 (4), 752-756.

Eisenmann H, Letsiou I, Feuchtinger A, Beisker W, Mannweiler E, Hutzler P and Arnz P (2001). Interception of small particles by flocculent structures, sessile ciliates, and the basic layer of a wastewater biofilm. Applied and Environmental Microbiology 67 (9), 4286-4292.

Fisher J and Acreman MC (2004). Wetland nutrient removal: a review of the evidence. Hydrology and Earth System Sciences 8 (4), 673-685.

Ghermandi A, Bixio D and Thoeye C (2007). The role of free water surface constructed wetlands as polishing step in municipal wastewater reclamation and reuse. Science of The Total Environment 380 (1-3), 247-258.

Grossart H, Berman T, Simon M and Pohlmann K (1998). Occurrence and microbial dynamics of macroscopic organic aggregates (lake snow) in Lake Kinneret, Israel, in fall. Aquatic Microbial Ecology 14 (1), 59-67.

Holcova V, Sima J, Edwards K, Semancikova E, Dusek J and Santruckova H (2009). The effect of macrophytes on retention times in a constructed wetland for wastewater treatment. International Journal of Sustainable Development and World Ecology 16 (5), 362-367.


17

Holeton C, Chambers PA and Grace L (2011). Wastewater release and its impacts on Canadian waters. Canadian Journal of Fisheries and Aquatic Sciences 68 (10), 1836-1859.

Johnson BD, Kranck K and Muschenheim DK (1994). Physico-chemical factors in Particle Aggregation. Chapter 4 in: Wotton RS; The biology of particles in aquatic systems. CRC Press, Inc., Boca Raton, Florida.

Kadlec RH and Wallace SD (2008). Treatment wetlands. CRC Press, Florida.

Kalff J (2002). Limnology. Prentice-Hall, Englewood Cliffs, New Jersey.

Karim MR, Manshadi FD, Karpiscak MM and Gerba CP (2004). The persistance and removal of enteric pathogens in constructed wetlands. Water Research 38 (7) , 1831-1837.

Keddy PA (2010). Wetlands an Overview. In: Keddy PA (ed), Wetland Ecology principles and conservation. Cambridge University Press, New York.

Kelderman P, Ang’weya RO, De Rozari P and Vijverberg T (2012). Sediment characteristics and wind-induced sediment dynamics in shallow Lake Markermeer, the Netherlands. Aquatic Sciences 74 (2), 301-313.

McDougald D, Rice SA, Barraud N, Steinberg PD and Kjelleberg S (2012). Should we stay or should we go: mechanisms and ecological consequences for biofilm dispersal. Nature Reviews Microbiology 10 (1), 39-50.

Marttinen SK, Kettunen RH and Rintala JA (2003). Occurrence and removal of organic pollutants in sewages and landfill leachates. The Science of The Total Environment 301 (1-3), 1-12.

Reinoso R, Torres LA and Becares E (2008). Efficiency of natural systems for removal of bacteria and pathogenic parasites from wastewater. Science of the Total Environment 395 (2-3), 80-86.

Romani A, Giorgi A, Acuna V and Sabater S (2004). The influence of substratum type and nutrient supply on biofilm organic matter utilization in streams. Limnology and Oceanography 49 (5), 1713-1721.

Ruckelshaus M, Wissmar R and Simenstad C (1993). The Importance of Autotroph Distribution to Mussel Growth in a Well-Mixed, Temperate Estuary. Estuaries 16 (4), 898-912.

Seviour RJ and Nielsen PH (2010). Microbial communities in activated sludge plants. In: Seviour RJ and Nielsen PH (eds), Microbial ecology of activated sludge. IWA publishing, London, pp. 95-125.

Simon M, Grossart H, Schweitzer B and Ploug H (2002). Microbial ecology of organic aggregates in aquatic ecosystems. Aquatic Microbial Ecology 28 (2), 175-211.

18

Chapter 1

Stott R and Tanner CC (2005). Influence of biofilm on removal of surrogate faecal microbes in a constructed wetland and maturation pond. Water Science and Technology 51 (9), 315-322.

Sundaravadivel M and Vigneswaran S (2001). Constructed wetlands for wastewater treatment. Critical Reviews in Environmental Science and Technology 31 (4), 351-409.

Tchobanoglous (2004). Wastewater Engineering: Treatment and Reuse. McGraw-Hill, New York.

Turner A and Millward GE (2002). Suspended particles: Their role in estuarine biogeochemical cycles. Estuarine Coastal and Shelf Science 55 (6), 857-883.

van den Boomen R, Kampf R and Mulling BTM (2012). Waterharmonica, onderzoek naar zwevend stof en pathogenen, hoofdrapport. STOWA publicatie 2012-10.

van den Boomen R and Kampf R (2012). Waterharmonica’s in Nederland 1996-2011: van effluent tot bruikbaar oppervlaktewater. STOWA 2012-12.

Vidales-Contreras JA, Gerba CP, Karpiscak MM, Acuna-Askar K and Chaidez-Quiroz C (2006). Transport of coliphage PRD1 in a surface flow constructed wetland. Water Environment Research 78 (11), 2253-2260.

Vymazal J (2007). Removal of nutrients in various types of constructed wetlands. Science of the Total Environment 380 (1-3), 48-65.

Vymazal J (2005). Constructed wetlands for wastewater treatment. Ecological Engineering 25 (3), 475-477.

Vymazal J (1996). Constructed wetlands for wastewater treatment in the Czech Republic the first 5 years experience. Water Science and Technology 34 (11), 159-164.

Wetzel RG (2001). Limnology. Academic Press, London.

Woodruff S, House W, Callow M and Leadbeater B (1999). The effects of biofilms on chemical processes in surficial sediments. Freshwater Biology 41 (1), 73-89.

Wotton RS (1994). The biology of particles in aquatic systems. CRC Press, Inc., Boca Raton, Florida.

Zhang L, Xia X, Zhao Y, Xi B, Yan Y, Guo X, Xiong Y and Zhan J (2011). The ammonium nitrogen oxidation process in horizontal subsurface flow constructed wetlands. Ecological Engineering 37 (11), 1614-1619.

Physical and biological changes of suspended particles in a surface flow constructed wetland

Chapter 2

Under revision for Ecological Engineering: Mulling BTM, van den Boomen RM, Claassen THL, van der Geest HG, Kappelhof JWNM, Admiraal W (2013). Physical and biological changes of suspended particles in a surface flow constructed wetland.

22

Chapter 2

AbstractSuspended particles are considered as contaminants in treated wastewater and can

have profound effects on the biological, physical and chemical properties of receiving aquatic ecosystems, depending on the concentration, type and nature of the suspended particles. Constructed wetlands are known to substantially reduce the concentration of suspended particles in treated wastewater, but hardly anything is known about the changes in the type and nature of these particles. Therefore, the aim of the present study was to investigate the changes in the physical and biological characteristics of suspended particles during residence in the full scale surface flow constructed wetland. The constructed wetland consists of unvegetated ponds and reed beds and receives treated municipal wastewater containing low concentrations of suspended particles. It was found that residence in the unvegetated ponds caused no major changes in particle concentration, but the organic content (53% to 33%) and average size (4.3 µm to 3.7 µm) of the suspended particles did decrease, caused by sedimentation of large organic particles and addition of smaller inorganic particles most likely resulting from shore erosion. The bacterial species originating from the wastewater treatment plant (quantified by indicator organisms) decreased strongly in abundance (>90% reduction), especially during residence in the reed beds. Simultaneously the total abundance of bacteria gradually increased, indicating the replacement of the bacterial species present in the treated wastewater with other species during residence in the constructed wetland. These observations indicate that constructed wetlands change the nature and type of the suspended particles during residence in a constructed wetland and reduce the input of anthropogenic particles into receiving surface waters.

Suspended particles

23

IntroductionSuspended matter is defined as the filterable matter in the water column and consists

of particles of highly variable type and origin (Wotton 1994). Suspended matter is involved in many biological, physical and chemical processes in aquatic ecosystems, like primary production, decomposition, nutrient cycling, energy transfer, contaminant binding, oxygen regulation, light absorption and temperature regulation (Noe et al. 2007; Bilotta and Brazier 2008; Olsen et al. 1982; Dawson and Macklin 1998; Russell et al. 1998; Wetzel 2006; Droppo et al. 1997). The role of suspended matter in this complex set of processes depends on the concentration, nature and type of the suspended particles. In turn, the concentration and type of suspended matter is influenced by both internal processes such as sedimentation, resuspension, aggregation, disaggregation, decomposition, biomass growth, cell lysis and biofiltration, and external processes such as atmospheric deposition, shore erosion, chemical precipitation and land runoff (Wotton 1994; Zufall et al. 1998).

Municipal wastewater is a common anthropogenic source of suspended particles into aquatic ecosystems. Although regular treatment of domestic wastewater strongly reduces the concentration of suspended particles (Tchobanoglous, et al. 2004), suspended particles present in treated domestic wastewater still lead to alterations of the physical, biological and chemical properties of receiving aquatic ecosystems (Holeton et al. 2011; Bilotta and Brazier 2008; Tchobanoglous et al. 2004).

Wetlands are known to reduce the suspended particles concentration (Kadlec and Wallace 2008; Knox et al. 2008) and constructed wetlands (CWs) are widely used for polishing (treated) municipal wastewater (Kadlec and Wallace 2008; Vymazal 2005; Mungasavalli and Viraraghavan 2006; Sundaravadivel and Vigneswaran 2001). Constructed wetlands (CWs) receiving higher inflow concentrations generally show substantial removal of suspended particles up to 90% (Ghermandi et al. 2007; van den Boomen et al. 2012; van den Boomen and Kampf 2012). However, CWs with low background concentrations of inflowing particles have been observed to produce equal or slightly increased outflow concentrations (Ghermandi et al. 2007; van den Boomen and Kampf 2012). Most research on particles in wetlands concentrates on bulk measurements of suspended matter. The understanding of the processes involved in and effects on the physical and biological composition of suspended particle during residence in constructed wetlands, is limited.

The present study investigates biological and physical changes in the concentration, nature and type of suspended particles present in treated wastewater during residence in a full scale surface flow constructed wetland receiving treated domestic wastewater containing low concentrations of suspended particles. An extensive five year monitoring program including suspended particles and faecal indicator organisms concentrations, and physicochemical parameters was carried out together with intensive short term monitoring campaign focusing on particle size distribution, organic content, dissolved element concentrations, sedimentation, biota abundance and microscopy imaging to gain more insight into suspended particle dynamics in wetlands.

24

Chapter 2

Material and methods

Study areaThis study was carried out in a full scale surface flow constructed wetland (CW)

located in Grou, The Netherlands. The CW was built in 2006 and receives treated municipal wastewater with a constant hydraulic load (1200 m3 day-1). The inflow of the constructed wetland leads treated wastewater through a series of three unvegetated ponds and four parallel reed beds, before being pumped into receiving surface water (channel) (Fig. 2.1).

The unvegetated ponds are open water systems with an average depth of 1.35 m, width of ±7.9 m, length of 55 m, volume between 360 and 440 m3 each and total hydraulic residence time (HRT) of 17.9 h (Fig 2.1). The reed beds are covered with Phragmites australis and have an average water depth of 0.40 m, width of ±11 m, length of 110 m, approximate volume of 443 m3 each and each receives a hydraulic load of ±300 m3 day-1 with an average HRT of 23.6 h. The total HRT of the CW was 41.5 h. Hydraulic residence times were calculated from the residence time distribution obtained by a tracer experiment using lithium chloride, performed in 2010. The average hydraulic residence time was determined at 50% passage of the lithium chloride load (Mulling et al. 2013).

WWTP Receiving surface water

1200 m3 d-1

300 m3 d-1

300 m3 d-1

300 m3 d-1

300 m3 d-1

HRT: 17.9 h HRT: 23.6 h

Unvegetated ponds Reed beds

Fig. 2.1 Scheme of the surface flow constructed wetland in Grou, The Netherlands, receiving treated municipal wastewater. The CW consists of three unvegetated ponds followed by four parallel reed beds (Phragmites australis) from which the water is pumped into the receiving surface water. The hydraulic loading (m3 d-1) is controlled and maintained stable at the inflow of the unvegetated ponds and the hydraulic residence times (HRT) were determined by a tracer experiment (Mulling et al. 2013).

Five year monitoringMonitoring of the CW was conducted over a five year period from 2007 until 2011

and included analyses of suspended particles, E. coli and physicochemical parameters. The sample frequency was mostly monthly, but varied between sampling locations and in time over the five year period. Samples were taken at three sampling points, at the in- and out-flow of the unvegetated ponds and at the out-flow of the reed beds (respectively PONDS-IN; PONDS-OUT; REED-BEDS-OUT). Samples were taken as point samples 0.2 m below the

Suspended particles

25

water surface at the middle of the water body using a 10L vessel. Subsamples were taken for individual analyses and stored at 4°C prior to further analyses initiated within 24 h. Suspended particle concentrations were determined by filtration (GF-F; Ø47 mm, pore size 0.7 µm) and sequential drying and weighing of the filters according to standard methods (NEN-EN-872 2005). Samples that were below the detection limit of 0.5 mg L-1 were included in further analyses with a value of halve the detection limit (0.25 mg L-1). E. coli concentrations were determined by membrane filtration according to standard methods (NEN-EN-ISO-9308-1 2000). pH and dissolved oxygen were measured in situ according to standard methods NEN-ISO-10523 (2008) and NEN-EN-ISO-5814 (1993), respectively.

Particle characteristics and composition In the period of 2009 till 2011 three short term sampling campaigns (1 to 14 days)

were conducted to quantify inorganic and organic content of suspended particles, suspended particle size distribution, bacteria, eukaryote, Enterococci and Clostridium perfringens abundance.

Microscopic observations of suspended particles water samples (50 mL) were taken at the three sampling points in the CW and fixed with 37% formaldehyde (10% v/v). After fixation halve of the samples were filtered over 0.2 µm polycarbonate filters (Ø47 mm; Sartorius Stedim Biotech, Göttingen, Germany) and stored at -20°C till further analyses with environmental scanning electron microscopy (ESEM). The other halve of the samples were not filtered and stored at 4°C till analysed with light microscopy. Filters were analysed by ESEM under semi vacuum conditions (Hitachi TM3000, Tokyo, Japan) at a magnification of 500×. The un-filtrated samples were left for particle settling at least 24 h before observation with an inverse light microscope. Settled material was transferred onto microscope slide and pictures of the particles were taken at a magnification of 500×.

The inorganic fractions were calculated after ignition (550°C) of filtered and dried suspended particles, according to standard methods (NEN-6499 2005). The organic content was converted to carbon content (mg C L-1) by assuming a carbon content in organic matter of 58% (Schulte and Hopkins 1996). Suspended particle size distributions (2.0 and 100 µm) were analysed in 20 mL water samples using a particle counter (PAMAS Waterviewer; PAMAS GmbH, Rutesheim, Germany; sensor: HCB-50/50). Analyses of the particle size distribution were conducted within one hour after sampling and stored at room temperature prior to analyses.

For determination of the eukaryotes and bacteria abundance, water samples (50 mL) were fixated with 37% formaldehyde (10% v/v) and subsequently filtered over 0.2µm polycarbonate filters (Ø47 mm; Sartorius Stedim Biotech, Göttingen, Germany) and stored at -20°C prior to further analyses. Using fluorescence in situ hybridization according to Glockner et al. (1996) eukaryotes and bacteria were coloured with specific probes, respectively EUKb310 (Baker et al. 2003) and EUB338mix (Daims et al. 1999). After hybridization the filters were mounted on microscope slides and embedded in anti-bleaching media (4:1),

26

Chapter 2

Citifluor (AF1; Citifluor Ltd., Leicester, UK) and Vectashield (Vectorshield Laboratories, Inc., Burlingame, US) and spectral images were taken using a confocal laser scanning microscope (Nikon A1) at 600× magnification. Eukaryotes and bacteria were separated from the spectral images using linear unmixing (Zimmermann 2005) and the abundance was determined by image analyses (ImageJ; http://imagej.nih.gov/ij/). Cell counts were converted in carbon content (mg C L-1) by first calculating the biomass by multiplying the count with the average size, assuming spherical cells. The bacteria were then converted from abundance to carbon content by applying a conversion factor of 0.22 g cm-3 biomass (Bratbak and Dundas 1984). The eukaryotes carbon content was calculated by converting the biomass into dry weight (17%) and sequentially to carbon content (43%) (Omori 1969).

Samples for analysis of Enterococci and C. perfringens abundance were stored at 4°C and analyses were initiated within 24 h after sampling. Concentrations of Enterococci were analysed by membrane filtration according to standardised methodology (NEN-EN-ISO-7899-2 2000). C. perfringens concentrations were measured by membrane filtration using 0.45 µm membrane filters (Cellulose Nitrate; Ø47 mm). Filters were placed on Triptose Sulfite Cyclocerine medium (TSC-medium) under anaerobic conditions for 24 ±2 h at 45 ±1°C. After incubation black colonies were counted as C. perfringens. For conformation several colonies were transferred into Mobility-Nitrate reduction medium (BN-medium) and Lactose-Gelatine medium (LG-medium) and incubated at 37 ±1°C for 24 ±2 h. In BN-medium C. perfringens should produce gas, colour the medium yellow and prevent solidification of the medium, in LG-medium C. perfringens should colonise the complete medium volume and colour the medium red, with the addition of zinc this colour should diminish after ten minutes. This method is based on ISO/TC-147/SC4/WG5 (1995); ISO-6461-2 (1986); NEN-EN-ISO-8199 (2007).

Additional measurementsSedimentation traps were employed in 2010 at twelve locations in the sedimentation

ponds (5, 4 and 3 locations in first, second and third unvegetated pond, respectively). At each location eight sedimentation traps were placed on top of the sediment (total number of traps: 12×8=96). The sedimentation traps were 0.33 m high and 0.05 m wide (volume of 650 mL). After 29, 49, 93 and 168 h from each of the 12 locations two (out of 8) sedimentation traps with a combined volume of 1.3 L were recovered. The content of the traps were well mixed, re-suspending all sedimentated matter, transferred in glass 2 L bottles and stored at 4°C prior to analyses. The sedimentation samples were analysed as suspended matter samples within 24 h after sampling, by filtration and sequential drying and weighing of the filters according to standard methods (NEN-EN-872 2005), by dividing these concentrations (mg L-1) by the total volume of the sedimentation traps the total amount sedimentated mass (mg) was calculated per location. Dividing the total of the sedimentated mass by the opening area of the sedimentation trap yielded the sedimentation per cm2. Per location linear regression was performed over the sedimentation per cm2 over the four time points to derive the

Suspended particles

27

sedimentation flux (mg cm2 d-1).Dissolved elements and nutrients were sampled by in-situ filtration (to minimize

contact with the air) over a 0.2 µm nitrate cellulose membrane filter (Ø25 mm; Whatman, Kent, UK). Samples for element analyses (Al, Ca, Cd, Co, Cr, Cu, Fe, K, Li, Mg, Mn, Na, Ni, P, Pb, S, Se, Si, Sr, Zn) were acidified with nitric acid on location to approximately pH 2 to prevent precipitation and stored at 4°C prior to analysis. The element concentrations were analysed using an ICP-AES (Optima 3000XL; AAnalyst 400, Waltham, USA). Nutrients (NH4, PO4, SO4) were analysed using an auto-analyser (Skalar San++System, Breda, The Netherlands) applying standard settings.

Data analysesData was analysed using Microsoft Excel and PAST (Hammer et al. 2001). Most

data was not normally distributed, thus non-parametric tests were used (Kruskal-Wallis) to analyses the data (PAST). The frequency plots were analysed with a G-test using Microsoft Excel. Regression analyses of the sedimentation rates were conducted in Microsoft Excel.

Results

Suspended particle concentrationsThe input of suspended particles into the ponds was relative stable over the five year

period without significant seasonal or daily differences in suspended particle concentrations (data not shown). The average suspended particle concentration (±s.e.) at PONDS-IN was 3.6 ±0.3 mg L-1 (n=153). 76% of the measurements were below 5 mg L-1, eight measurements (5%) were above 10 mg L-1 and the maximum measured suspended particle concentration was 19.0 mg L-1 (Fig. 2.2). The average suspended particle concentration at PONDS-OUT was 5.6 ±0.73 mg L-1, significantly higher than at PONDS-IN (p<0.002) (Fig. 2.2a). The frequency plots of the measured suspended particle concentrations also showed a significant shift to higher values, with 40% of the samples above 5 mg L-1 (p<0.02) (Fig. 2.2b). The suspended particle concentration at PONDS-OUT also showed higher peak values with a maximum of 26.9 mg L-1. After passing the reed beds the average suspended particle concentration decreased significantly (p<0.05) to 3.9 ±0.44 mg L-1. These suspended particle concentrations were similar to the CW inflow concentrations at PONDS-IN. Besides the average concentration, the distribution of the suspended particles measurements at REED-BEDS-OUT and at PONDS-IN did not differ significantly. 77% of the measurements at REED-BEDS-OUT were below 5 mg L-1 (n=52) and 8% measurements above 10 mg L-1 (maximum concentration of 14.5 mg L-1).

28

Chapter 2

05

1015202530

No.

of o

ccur

ance

Suspended particle concentration (mg L-1)

05

1015202530

No.

of o

ccur

ance


05

1015202530

No.

of o

ccur

ance


a) b)

No.

of o

bser

vatio

ns

PONDS-IN

PONDS-OUT

REED-BEDS-OUT

PONDS-IN PONDS-OUT REED-BEDS-OUT

Fig. 2.2 Boxplots of the suspended particle concentration (a) at the in and out flow of the unvegetated ponds and reed beds and frequency plots (b) of the suspended particle concentrations over the 5 year monitoring period. Boxplots represented with the median as the middle horizontal line and the average concentration as an x. PONDS-IN, n=153; PONDS-OUT, n=49; REED-BEDS-OUT, n=52.

Particle characteristics and compositionBoth ESEM and light microscopy images (Fig. 2.3a) showed relative high abundance

of large particles at PONDS-IN, mostly consisting of organic matter, sand and plastics. Several diatoms were observed, of which most were empty shells. At PONDS-OUT the relative large particles were absent, although smaller particles of similar consistency were still observed (Fig. 2.3a). Several diatoms were also observed at PONDS-OUT, along with numerous small motile green algae. At REED-BEDS-OUT organic matter seemed less dense compared with the other two locations and large particles were dominated by several species of diatoms (Fig. 2.3a).

At the beginning of the CW (PONDS-IN) the average (±s.e.) number of suspended particles with a size between 2.0 and 100 µm was 14.9x106 ±0.2x106 no. L-1 (Fig. 2.3b), and consisted mostly of particles between 2.0 and 3.0 µm (approximately 50%; data not shown). The average size of the particles at the PONDS-IN was 4.34 ±0.03 µm (Fig. 2.3b). At PONDS-OUT the total number of suspended particles with a size between 2.0-100 µm did not differ significantly from the PONDS-IN (15.1x106 ±0.2x106 no. L-1). The average particle size decreased significantly to 3.78 ±0.03 µm (p<0.001) (Fig. 2.3b), caused by a decrease of particle abundance larger than 8.0 µm (40% to 90%; positively correlated with size), while simultaneously particles below 5.0 µm increased in abundance between 10-20%. At REED-BEDS-OUT the total number of suspended particles with a size between 2.0 and 100 µm was significantly lower, 9.5x106 ±0.2x106 no. L-1, compared to the PONDS-IN and PONDS-OUT

Suspended particles

29

(p<0.001). During residence in the reed beds suspended particles below 8.0µm decreased between 30-45%, with only minor changes in suspended particle numbers above 8.0µm, resulting in an average particle size of 3.82 ±0.01 µm (Fig. 2.3b).

On average (±s.e.) the suspended particles at PONDS-IN consisted of 53 ±3% organic matter which equals to an average organic carbon content of 1.10 ±0.08 mg C L-1 (Fig. 2.3c). After residence in the unvegetated ponds (at PONDS-OUT) the average contribution of organic matter to the total suspended particles decreased significantly (p<0.01) to 33 ±6.5%, equalling an average organic carbon content of 1.07 ±0.14 mg C L-1 (Fig. 2.3c). During residence in the reed beds average organic carbon content decreased to 0.59 ±0.07 mg C L-1

(Fig. 2.3c). During the residence in the CW the average bacteria abundance (±s.e.) significantly

increased from 3.2x109 ±0.5x109 no. L-1 at PONDS-IN to, 4.9x109 ±0.2x109 no. L-1 at PONDS-OUT and 7.9x109 ±0.8x109 no. L-1 at REED-BEDS-OUT (p<0.001). The average size of the bacteria at PONDS-IN was 0.59 ±0.19 µm, 0.81 ±0.22 µm at PONDS-OUT and 0.69 ±0.32 µm at REED-BEDS-OUT. Calculated back to organic carbon content this translates to respectively 0.07, 0.30 and 0.29 mg C L-1 at PONDS-IN, PONDS-OUT, REED-BEDS-OUT (Fig. 2.3d). The average abundance (no. L-1; ±s.e.) of eukaryotes did not differ significantly between PONDS-IN (5.0x109 ±0.6x109), PONDS-OUT (4.3x109 ±0.5 4.3x109) and REED-BEDS-OUT (4.9x109 ±0.4x109). The average size of the eukaryotes at PONDS-IN was 0.54 ±0.06 µm, 1.39 ±0.13 µm at PONDS-OUT and 1.08 ±0.12 µm at REED-BEDS-OUT. Calculated back to organic carbon content this translates to respectively 0.03, 0.47 and 0.25 mg C L-1 at PONDS-IN, PONDS-OUT, REED-BEDS-OUT (Fig. 2.3d).

Indicator bacteria E. coli, Enterococci and C. perfringens all significantly decreasing in abundance during residence in the CW (p<0.01), were the decrease was on average 58 ±9% higher during residence in the reed beds compared to the unvegetated ponds, with a total decrease of respectively 1.89, 1.37 and 0.95 log10 CFU L-1. The contribution of these indicator organisms (CFU) to the total abundance of bacteria also decreased during residence in the CW from 0.029%, 0.005% and >0.001%, at respectively PONDS-IN, PONDS-OUT, REED-BEDS-OUT (Fig. 2.3e).

30

Chapter 2

Fig. 2.3 Overview parameters measured at PONDS-IN (left column), POND-OUT (middle column) and REED-BEDS-OUT (right column) of the CW. a) Pictures include scanning electron microcopy (500x) and four revered light microscope images (500x) of suspended particles. b) Bar graphs: average (±s.e.) suspended particle concentration, number of suspended particles and average particle size. c) Pie graphs of the average

OM

Eukaryotes

Bacteria

OM

Eukaryotes

Bacteria

OM

Eukaryotes

Bacteria

0

2

4

6

8

10

12

14

16

18


Susp

ende

d pa

rtic

les (

no. L

-1) x

103

Eukaryota

0

2

4

6

8

10

12

14

16

18


Susp

ende

d pa

rtic

les (

no. L

-1) x

103

Inorganic

0

1

2

3

4

5

6

7


Susp

ende

d pa

rtic

les (

mg

L-1

)

Organic

0

1

2

3

4

5


Ave

rage

par

ticle

size

(µm

)

0

2

4

6

8

10

12

14

16

18


Susp

ende

d pa

rtic

les (

no. L

-1) x

106

0

1

2

3

4

5

6

7


Susp

ende

d pa

rtic

les (

mg

L-1

)

0

1

2

3

4

5

6


Susp

ende

d pa

rtic

les (

mg

L-1

)

0

1

2

3

4

5

6

7


Susp

ende

d pa

rtic

les (

mg

L-1

)

mg L-1

0

2

4

6

8

10

12

14

16

18


Susp

ende

d pa

rtic

les (

no. L

-1) x

103

no. L-1

0

1

2

3

4

5


Ave

rage

par

ticle

size

(µm

)

µm

0.000%

0.005%

0.010%

0.015%

0.020%

0.025%

0.030%

0.035%


Indi

cato

r pat

hoge

ns (

CFU

L-1

)

0

1

2

3

4

5

6


Susp

ende

d pa

rtic

les (

mg

L-1

)

0.000%

0.005%

0.010%

0.015%

0.020%

0.025%

0.030%

0.035%


Indi

cato

r pat

hoge

ns (

CFU

L-1

)PONDS-IN PONDS-OUT REED-BEDS-OUT

100 µm 100 µm 100 µm

Organic

Inorganic

Organic

Inorganic

Organic

Inorganic

0

1

2

3

4

5

6

7


Susp

ende

d pa

rtic

les (

mg

L-1

)

Other POM

0

1

2

3

4

5


Ave

rage

par

ticle

size

(µm

)

Bacteria

0

1

2

3

4

5

6

7


Susp

ende

d pa

rtic

les (

mg

L-1

)

Indicator pathogens

Faec

alin

dica

tor

orga

nism

s

a)

b)

c)

d)

e)

Suspended particles

31

organic to inorganic fractions based on the ash weight. d) Pie graphs of the average distribution of organic matter over bacteria, eukaryotes (zooplankton and phytoplankton) and other particulate organic matter based on the calculated carbon content (mg C L-1). e) Bar graphs on the contribution of faecal indicator organisms (E. coli, Enterococci and C. perfringens) to the total bacteria abundance based of FISH bacterial counting (no. L-1) and indicator bacteria culturing (CFU L-1).

Additional measurementsIn the unvegetated ponds the sedimentation fluxes (±s.e.) ranged between 8.57 ±1.27

mg cm-2 d-1 and 0.39 ±0.04 mg cm-2 d-1 and decreased exponentially (r2= 0.96; p<0.001) with distance from the discharge point (Fig. 2.4).

During residence in the CW dissolved element concentrations of sodium, zinc and magnesium gradually increased by >13%, while calcium, potassium, strontium, and sulphur showed minor increases (Table 2.1). Besides the gradual increase of elements during residence in the CW, concentrations of dissolved (ortho)-phosphorus increased 6% during residence in the unvegetated ponds, at REED-BEDS-OUT the concentration decreased by 17%. Between PONDS-IN and PONDS-OUT the concentration of dissolved iron and ammonium decreased again by respectively 19% and 21%, and continued to decrease during residence in the reed beds by 31% and >97% respectively. The strong decrease in dissolved ammonium during residence in the reed beds is also seen in the dissolved concentrations of silicate and manganese both decreasing by 64% and 82%, respectively. Dissolved sulphate concentrations remained stable during residence in the CW and many dissolved metals concentrations were below the detection limit (Cd, Co, Cr, Ni, Pb, Se, Al and Cu).

The average pH was 7.6 ±0.03, 7.4 ±0.1 and 7.4 ±0.04 at respectively PONDS-IN, PONDS-OUT and REED-BEDS-OUT with a significant decrease between the average pH at PONDS-IN (p<0.001) compared with PONDS-OUT and REED-BEDS-OUT (Fig. 2.5). The average dissolved oxygen (DO) at PONDS-IN, PONDS-OUT and REED-BEDS-OUT were respectively 2.9 ±0.3, 4.8 ±0.5 and 5.3 ±0.7, with a significantly (p=0.02) lower DO concentrations at PONDS-IN than at POND-OUT. The DO concentrations showed high daily fluctuations at POND-OUT probably caused by the very local presence of submerged macrophytes, and daily and seasonal fluctuations at REED-BEDS-OUT were probably caused by temporal duck weed coverage or phytoplankton (data not shown; for detailed description, van den Boomen et al. 2012).

32

Chapter 2

R² = 0.9602

0

2

4

6

8

10

12

0 55 110 165

Sedi

men

tatio

n flu

x (m

g cm

-2d-1

)

Distance from discharge (m)

Pond 1 Pond 2 Pond 3

Fig. 2.4 Sedimentation flux of suspended particles (mg cm-2 d-1; ±s.e.) in the unvegetated ponds with distance from the discharge point (PONDS-IN). Sedimentation rates were fitted with an exponential regression line (r2=0.96; p<0.001).

Table. 2.1 Dissolved element and nutrient concentrations (±s.e.) at PONDS-IN, POND-OUT and REED-BEDS-OUT.

PONDS-IN PONDS-OUT REED-BEDS-OUT PONDS-IN PONDS-OUT REED-BEDS-OUT

(mg L-1) (mg L-1) (mg L-1) (mg L-1) (mg L-1) (mg L-1)

Al* 0.011 (0.002) 0.008 (0.001) 0.014 (0.004) Ni <0.001 <0.001 <0.001

Ca 69.1 (0.4) 73.1 (0.2) 72.3 (0.4) P 0.401 (0.012) 0.424 (0.000) 0.363 (0.003)

Cd <0.001 <0.001 <0.001 Pb <0.005 <0.005 <0.005

Co <0.001 <0.001 <0.001 S 8.39 (0.01) 9.08 (0.01) 9.14 (0.03)

Cr <0.001 <0.001 <0.001 Se <0.0126 <0.0126 <0.0126

Cu* 0.014 (0.007) 0.004 (0.003) 0.017 (0.002) Si 7.46 (0.08) 7.45 (0.01) 2.65 (0.06)

Fe 0.15 (0.006) 0.122 (0.002) 0.084 (0.003) Sr 0.273 (0.000) 0.294 (0.000) 0.299 (0.001)

K 15.7 (0.1) 16.7 (0.1) 16.5 (0.2) Zn 0.008 (0.001) 0.010 (0.000) 0.055 (0.058)

Li 0.011 (0.000) 0.012 (0.000) 0.013 (0.000)

Mg 15.9 (0.1) 17.6 (0.1) 18.4 (0.2) NH4 7.83 (0.09) 6.20 (0.22) <0.180

Mn 0.216 (0.000) 0.215 (0.003) 0.038 (0.000) PO4 1.05 (0.02) 1.19 (0.01) 1.01 (0.01)

Na 150.2 (0.4) 168.6 (2.7) 172.6 (1.1) SO4 26.7 (0.1) 26.3 (0.1) 26.4 (0.7)

* Concentration above detection limit but possibly influenced by large measurements errors.

Suspended particles

33

7.0

7.2

7.4

7.6

7.8

8.0

PONDS-IN PONDS-OUT REED BEDS-OUT

pH_TP

0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

PONDS-IN PONDS-OUT REED BEDS-OUT

O2_TPpH

Diss

olve

d ox

ygen

(mg

L-1

)

Fig. 2.5 Average (±s.e.) pH (n=144; 47; 42) and dissolved oxygen (mg L-1; n=15; 33; 29) at PONDS-IN, POND-OUT and REED-BEDS-OUT.

DiscussionThe removal of suspended particles from treated wastewater is one of the major

functions ascribed to constructed wetlands. In this study, however, we observed that the residence of treated municipal wastewater in a surface flow constructed wetland resulted in a minor net increase in the concentration of suspended particles. This apparent inefficient reduction of suspended particles in this study is in the first place linked to the very low inflow concentrations of suspended particles (3.6 ±0.3 mg L-1) which are similar to levels considered as background values in several other studies (Ghermandi et al. 2007; van den Boomen and Kampf 2012). But, more important, by analysing the nature and type of the suspended matter, we demonstrated that the concentrations of suspended particles originating from the wastewater treatment plant (WWTP) effluent are actually strongly reduced, but by addition of other particles during residence in the CW this reduction is masked when looking only at bulk measurements.

In the unvegetated ponds a sedimentation flux of 0.42 kg d-1 was calculated, which amounts to approximately 10% of the total daily input of suspended particles into the CW. The sedimentation of suspended particles decreased exponentially with distance from the discharge point, indicating that the sedimentation is mainly driven by particle size. This is supported by the significant decrease in average size of the suspended particles and suggests that most of the sedimentated particles are originating from the WWTP. The net increase of suspended particles at PONDS-OUT indicates that these sedimentated particles are replaced by particles of a different origin. In addition, the organic content of the suspended particles decreased during residence in the unvegetated ponds, which was most likely caused by sedimentation, fragmentation and decomposition of organic particles (Droppo et al. 1997; Kadlec and Wallace 2008). The coinciding increase in inorganic particles could have been caused by diatom growth (SiO2), formation of precipitates or external input from the atmosphere or bank erosion. Diatoms were however sparse and no substantial growth was observed, probably caused by the relative short HRT in the unvegetated ponds (17.9 h) in relation to the maximum specific growth rate of phytoplankton (Smith 1980; Sarthou et

34

Chapter 2

al. 2005; Sunda and Huntsman 1997). Precipitates (e.g. FeOOH, FePO4, AL(OH)3, AlPO4, CaCO3, CaSO4, (NH4)2SO4, metal-sulphates) may form when changes in pH and O2 occur (Kadlec and Wallace 2008). However, although such changes in pH and O2 were observed in the CW, only minor changes in the concentrations of dissolved precipitant forming elements and nutrients were observed (Fe, NH4,

SO4) suggesting that the formation of precipitates is of minor importance. Also atmospheric deposition seems to be only a minor contributor to the increase of inorganic particles during residence in the unvegetated ponds: a daily input of 5-180 grams to the CW was calculated based on dry atmospheric mass deposition rates of several studies (Holsen et al. 1993; Zufall et al. 1998; Sweet et al. 1998). Assuming that this input is completely mixed in the CW this would equal a concentration addition of only 4 to 150 µg L-1. However, bank erosion in the unvegetated ponds has been observed over the years of operation and is expected to be the major cause of the increase in inorganic suspended particles observed during residence in the unvegetated ponds. To increase the average concentration of suspended particles with 2 mg L-1, approximately 0.6 m3 of sand or 0.8 m3 of clay has to enter the system, which seems to be plausible with visual observations on bank erosion in the studied CW.

In contrast to the unvegetated ponds, in the reed beds a significant decrease in both the concentration and number of suspended particles was observed. This decrease was mainly caused by the removal of particles smaller than 8.0 µm and it is expected that these small particles are being trapped by biofilms that are known to have the capacity to (both actively and passively) filter suspended particles from water, especially in the size range between 2-8 µm (Kadlec and Wallace 2008; Balzer et al. 2010; Chabaud, et al. 2006; Drury et al. 1993; Eisenmann et al. 2001; Stott and Tanner 2005). The reed stems in the reed beds provide a large surface area for the development of biofilms and the presence of these biofilms is corroborated by the microscopic images that revealed an increase in the abundance of diatoms that are often abundant in biofilms. In addition, the strong decrease of ± 4.8 mg L-1 in dissolved silica during residence in the reed beds also indicates substantial diatom growth (Struyf and Conley 2009). In the reed beds the abundance of bacteria significantly increased, but simultaneously indicator bacteria abundance, a proxy of bacteria originating from the WWTP, decreased and indicated strong changes in the bacterial community composition during residence in the reed beds. In general, during residence in the reed beds the concentration, nature and type of the suspended particles greatly changed, including the reduction of WWTP originating bacterial and replacement of phytoplankton community by biofilm associated diatoms. These changes are in concurrence with studies by Toet et al. (2005) and Hey et al. (1994) who observed similar changes in the suspended particle composition of treated wastewater during residence in constructed wetlands.

Suspended particles

35

ConclusionIt was shown that additional polishing of WWTP effluent by constructed wetlands

may appear to be inefficient regarding the reduction of suspended matter at low inflow concentrations when only the total mass is considered, but by analysing more specific chemical and biological characteristics a strong conversion in the nature and type of the suspended matter was observed. Changes advance through the different functional components of the wetland: sedimentation of large organic particles and external input of small inorganic particles are dominant processes in unvegetated ponds, while in reed beds biological processes mark the transition from pathogenic to ‘natural’ microbial communities and from heterotrophic to phototrophic (diatom dominated) communities. Besides the reduction of high levels of suspended matter in treated wastewater, the capacity of constructed wetlands to change the nature and type of suspended matter is an additional tool to be used in the prevention of anthropological input into receiving surface waters.

Acknowledgments

This work was financed by the Foundation for Applied Water Research (STOWA) and supported by Witteveen+Bos, stichting Waternet, Wetterskip Fryslân and the Centre for Advanced Microscopy of the University of Amsterdam. Special thanks go out to supporting personnel, Hans van Nieuwenhuijzen, Michel Collin and Peter wind, Rinse van der Kooij and Marieke Soeter.

36

Chapter 2

References

Baker BJ, Hugenholtz P, Dawson SC and Banfield JF (2003). Extremely acidophilic protists from acid mine drainage host Rickettsiales-lineage endosymbionts that have intervening sequences in their 16S rRNA genes. Applied and Environmental Microbiology 69 (9), 5512-5518.

Balzer M, Witt N, Flemming HCand Wingender J (2010). Faecal indicator bacteria in river biofilms. Water Science and Technology 61 (5), 1105-1111.

Bilotta GS and Brazier RE (2008). Understanding the influence of suspended solids on water quality and aquatic biota. Water research 42 (12), 2849-2861.

Bratbak G and Dundas I (1984). Bacterial Dry-Matter Content and Biomass Estimations. Applied and Environmental Microbiology 48 (4), 755-757.


Daims H, Bruhl A, Amann R, Schleifer KH and Wagner M (1999). The domain-specific probe EUB338 is insufficient for the detection of all Bacteria: Development and evaluation of a more comprehensive probe set. Systematic and applied microbiology 22 (3), 434-444.

Dawson EJ and Macklin MG (1998). Speciation of heavy metals in floodplain and flood sediments: a reconnaissance survey of the Aire Valley, West Yorkshire, Great Britain. Environmental Geochemistry and Health 20 (2), 67-76.

Droppo IG, Leppard GG, Flannigan DT and Liss SN (1997) The freshwater floc: A functional relationship of water and organic and inorganic floc constituents affecting suspended sediment properties. 43-53.

Drury WJ, Characklis WG and Stewart PS (1993). Interactions of 1 mu-M latex-particles with Pseudomonas-aeruginosa biofilm. Water Research 27 (7), 1119-1126.


Ghermandi A, Bixio D and Thoeye C (2007). The role of free water surface constructed wetlands as polishing step in municipal wastewater reclamation and reuse. Science of The Total Environment 380 (1–3), 247-258.

Suspended particles

37

Glockner FO, Amann R, Alfreider A, Pernthaler J, Psenner R, Trebesius K and Schleifer KH (1996). An in situ hybridization protocol for detection and identification of planktonic bacteria. Systematic and Applied Microbiology 19 (3), 403-406.

Hammer Ø, Harper DAT and Ryan PD (2001). PAST: Paleontological statistics software package for education and data analysis. 4 (1), 1-9.

Hey DL, Kenimer AL and Barrett KR (1994). Water-Quality Improvement by 4 Experimental Wetlands. Ecological Engineering 3 (4), 381-397.


Holsen T, Noll K, Fang G, Lee W, Lin J and Keeler G (1993). Dry Deposition and Particle-Size Distributions Measured during the Lake-Michigan Urban Air Toxics Study. Environmental science & technology 27 (7), 1327-1333.

ISO/TC-147/SC4/WG5 (1995).Microbiological methods: Sulfite reducing clostridium.

ISO-6461-2 (1986). Water quality -- Detection and enumeration of the spores of sulfite-reducing anaerobes (clostridia) -- Part 2: Method by membrane filtration.

Kadlec RH and Wallace SD (2008) Treatment wetlands. CRC Press, Florida, .

Knox AK, Dahgren RA, Tate KW and Atwill ER (2008). Efficacy of natural wetlands to retain nutrient, sediment and microbial pollutants. Journal of Environmental Quality 37 (5), 1837-1846.

Mulling BTM, van den Boomen RM, van der Geest HG, Kappelhof JWNM and Admiraal W (2013). Suspended particle and pathogen peak discharge buffering by a surface-flow constructed wetland. Water research 47 (3), 1091-100.

Mungasavalli DP and Viraraghavan T (2006). Constructed wetlands for stormwater management: A review. Fresenius Environmental Bulletin 15 (11), 1363-1372.

NEN 6499 (2005). Environment - Umbrella Standard For The Determination Of Suspended Solids And Dry Residue And Their Residues On Ignition .

NEN-EN-872 (2005). Water quality - Determination of suspended solids - Method by filtration through glass fibre filters.

38

Chapter 2

NEN-EN-ISO 5814 (1993). Water quality - Determination of dissolved oxygen - Electrochemical probe method.

NEN-EN-ISO-7899-2 (2000). Water quality - Detection and enumeration of intestinal enterococci - Part 2: Membrane filtration method.

NEN-EN-ISO-8199 (2007). Water quality - General guidance on the enumeration of micro-organisms by culture.

NEN-EN-ISO-9308-1 (2000). Water quality - Detection and enumeration of Escherichia coli and coliform bacteria - Part 1: Membrane filtration method.

NEN-ISO-10523 (2008). Water quality - Determination of pH. Noe GB, Harvey JW and Saiers JE (2007). Characterization of suspended particles in Everglades wetlands.

Limnology and Oceanography 52 (3), 1166-1178.

Olsen CR, Cutshall NH, Larsen IL (1982). Pollutant Particle Associations and Dynamics in Coastal Marine Environments - a Review. Marine Chemistry 11 (6), 501-533.

Omori M (1969). Weight and chemical composition of some important oceanic zooplankton in the North Pacific Ocean. Marine Biology 3 (1), 4-10.

Russell MA, Walling DE, Webb BW and Bearne R (1998). The composition of nutrient fluxes from contrasting UK river basins. Hydrological Processes 12 (9), 1461-1482.

Sarthou G, Timmermans K, Blain S and Treguer P (2005). Growth physiology and fate of diatoms in the ocean: a review. Journal of Sea Research 53 (1-2), 25-42.

Schulte EE and Hopkins BG (1996). Estimation of soil organic matter by weight loss on ignition. In: Magdoff FR, Tabatabai MA and Hanlon EA (Eds.). Soil organic matter: Analysis and interpretation. SSSA Spec. Publ., Madison, 21-31.

Smith R (1980). The Theoretical Basis for Estimating Phytoplankton Production and Specific Growth-Rate from Chlorophyll, Light and Temperature Data. Ecological Modelling 10 (3-4), 243-264.


Struyf E and Conley DJ (2009). Silica: An Essential Nutrient in Wetland Biogeochemistry. Frontiers in Ecology and the Environment 7 (2), pp. 88-94.

Suspended particles

39

Sunda W and Huntsman S (1997). Interrelated influence of iron, light and cell size on marine phytoplankton growth. Nature 390 (6658), 389-392.


Sweet C, Weiss A and Vermette S (1998). Atmospheric deposition of trace metals at three sites near the Great Lakes. Water Air and Soil Pollution 103 (1-4), 423-439.

Tchobanoglous G, Burton FL and Stensel HD (2004). Wastewater Engineering: Treatment and Reuse. McGraw-Hill, New York.

Toet S, Van Logtestijn RSP, Schreijer M, Kampf R and Verhoeven JTA (2005). The functioning of a wetland system used for polishing effluent from a sewage treatment plant. Ecological Engineering 25 (1), 101-124.

van den Boomen R and Kampf R (2012). Waterharmonica’s in the Netherlands 1996-2011: from WWTP effluent till usable surface water. ISBN.978.90.5773.559.2 STOWA 2012-12 (in Dutch)

van den Boomen R, Kampf R and Mulling BTM (2012). Research on suspended particles and pathogens in the Waterharmonica (constructed wetland). ISBN.978.90.5773.553.0 STOWA 2012-10 (in Dutch)


Wetzel RG (2006) Wetland ecosystem processes. In: Batzer D.P, S.R.R. (Ed.), Ecology of freshwater and estuarine wetlands. University of California Press, Ltd., Berkeley and Angeles, 285-312.

Wotton RS (1994). The biology of particles in aquatic systems. CRC Press, Inc., Boca Raton, Florida.

Zimmermann T (2005). Spectral imaging and linear unmixing in light microscopy. Microscopy Techniques 95, 245-265.

Zufall M, Davidson C, Caffrey P and Ondov J (1998). Airborne concentrations and dry deposition fluxes of particulate species to surrogate surfaces deployed in Southern Lake Michigan. Environmental science & technology 32 (11), 1623-1628.

Changes in the planktonic microbial community during residence in a constructed wetland

Chapter 3

Submitted as: Mulling BTM, Soeter AM, van der Geest HG, Admiraal W (2013). Changes in the planktonic microbial community during residence in a constructed wetland.

42

Chapter 3

AbstractSuspended particles are a major constituent of municipal wastewater and generally

contain high levels of bacteria, including human pathogens. Discharge of these particles of anthropogenic nature can have profound effects on receiving aquatic ecosystems and mitigation of these effects requires additional polishing of treated municipal wastewater. Previously it was shown that surface flow constructed wetlands are effective in improving water quality by reducing the numbers of faecal indicator organisms. However, faecal indicator organisms represent only a minor fraction of the total planktonic bacterial community and knowledge on the effects of these constructed wetlands on the composition and functioning of the entire planktonic bacterial community is limited. The aim of this descriptive study was therefore to identify changes in the planktonic bacterial community during residence of treated municipal wastewater in a full-scale surface flow constructed wetland. To this purpose water samples were taken in which the bacterial community composition and functioning was analyzed using FISH, DGGE and BIOLOG. Surprisingly, the bacterial abundance at the inflow of the constructed wetland was relatively low compared with receiving surface waters. However, the inflowing bacterial community showed high metabolic activity and functional diversity. During residence in the constructed wetland the bacterial abundance doubled, but decreased in metabolic activity and functional diversity. Shifts in the community composition indicate that these changes are related to turn-over of the bacterial community. The planktonic bacterial community in the effluent of the constructed wetland closely resembled natural bacterial communities in urban and agricultural ditches. Based on these observations we conclude that constructed wetlands are capable to mitigate possible impacts of the particle load in treated wastewaters by transforming the anthropological bacterial community to a bacterial community resembling more “natural” surface waters.

Planktonic bacteria

43

IntroductionDischarges of municipal wastewater is a major source of anthropological impacts on

aquatic ecosystems and therefore it is common to treat municipal wastewater before discharge. Treatment of raw municipal wastewater by wastewater treatment plants (WWTPs) strongly improves water quality, but treated municipal wastewater generally still contains high levels of nutrients, organic matter and high densities of heterotrophic bacteria originating from the WWTP (Tchobanoglous et al. 2004; Seviour and Nielsen 2010). The discharge of treated wastewater can therefore still have a profound impact on receiving surface waters (Seviour and Nielsen 2010; Holeton et al. 2011) and mitigation of this impact requires additional polishing of treated municipal wastewater. To this purpose, several techniques are available, but for the polishing of municipal treated municipal wastewater free surface constructed wetlands (CWs) are often favored (Kadlec and Wallace 2008). Free surface CWs improve the water quality by reducing nutrient concentrations, decreasing the numbers of faecal indicator organisms and improving the oxygen regime (Sundaravadivel and Vigneswaran 2001; Vymazal 2005; Kadlec and Wallace 2008; van den Boomen and Kampf 2012).

The removal of particular matter from the water is another purification process occurring is these systems, but is only profound in CWs receiving relative high concentrations of suspended particles while CWs receiving low concentrations of suspended particles are seemingly ineffective in removing suspended particles (Ghermandi et al. 2007; van den Boomen and Kampf 2012). However, all studies on particle dynamics in CW analyze bulk concentrations of suspended matter only, while suspended particles consist of particles of highly variable type and origin (Wotton 1994) and includes not only dead organic and inorganic particles, but also zooplankton, phytoplankton and bacterioplankton. Especially bacteria are regarded as important players in many biological processes and biochemical cycles in aquatic ecosystems (Kalff 2002) and the anthropogenic nature of the bacterial community in treated municipal wastewater may therefore influence the functioning of receiving aquatic ecosystems. It is expected that residence in the different functional compartments of a CW could strongly change the nature and composition of the planktonic bacterial community. However, besides detailed observations on the removal of bacterial faecal indicator organisms (Kadlec and Wallace 2008; Vymazal 2005; Reinoso et al. 2011; Moleda et al. 2008; Karim et al. 2004; Ghermandi et al. 2007, Diaz et al. 2010), knowledge about changes in the planktonic bacterial community during residence in CWs is virtually none existing. Therefore the aim of this descriptive study was to identify changes in the planktonic bacterial community during residence in a surface flow CW.

To meet this aim, we set up a sampling campaign in a full scale surface flow CW consisting of unvegetated ponds and reed beds, that receives municipal wastewater with low concentrations of suspended particles (3.6 mg L-1; Mulling et al. 2013; van den Boomen and Kampf 2012; Ghermandi et al. 2007) and (Phragmites australis). Water samples were taken at five points in the CW and structural and functional changes in the bacterial community were described using FISH, DGGE and BIOLOG. In addition, we analyzed water samples from six

44

Chapter 3

different types of surface waters for comparisons between the bacterial communities in the CW and more natural bacterial communities.


Site descriptions and samplingThis study was conducted at seven locations in The Netherlands (Fig. 3.1). The main

location was a full scale surface flow constructed wetland (CW) located in Grou that was built in 2006 and receives a constant hydraulic loading of 1200 m3 day-1 of treated primarily municipal wastewater. After inflow into the CW, the treated wastewater flows through three in series connected unvegetated ponds, four parallel reed beds and a collection pond before being pumped into receiving surface water (Fig. 3.1). The ponds are open water systems without vegetation with an average depth of 1.35 m, a volume between 360 and 440 m3 each and a combined hydraulic retention time (HRT) of 17.9 h (Fig. 3.1). The reed beds are covered with Phragmites australis and have an average water depth of 40 cm, an approximate volume of 443 m3, and each receive a hydraulic loading of ±300 m3 day-1 with an average HRT of 23.6 h. The total HRT of the CW is 41.5 h (Mulling et al. 2013).

Samples were taken in June 2010 as point samples 10-20 cm below the water surface at five different locations in the CW: at the in- and outflow of the unvegetated ponds and at the outflow of the reed beds (further referred to as PONDS-IN; PONDS-OUT; REED-BEDS-OUT respectively) and in the middle of these compartments (further referred to as PONDS and REED-BEDS respectively) (Fig. 3.1). The six reference surface waters used for comparison differed in water quality and typology. Locations could be described as an artificial fen created by sand excavation and fed by rain and groundwater, an agricultural ditch next to grassland with dairy cows, an urban ditch located next to an apartment building in Amstelveen (The Netherlands), an excavated recreational peat lake, an urban river flowing into the city of Amsterdam and a canal in the centre of Amsterdam (Fig. 3.1).

Planktonic bacteria

45

Fig. 3.1 Left: Sampling locations in the Netherlands. Right: Map of the WWTP in Grou, the Netherlands (0) with a sedimentation tank (1) which discharges treated municipal wastewater into a CW consisting of unvegetated ponds (2), reed beds (3) and an ecological buffer zone (4) which is in open connection the receiving channel (5). Sampling points were located at; PONDS-IN (a), PONDS (b), PONDS-OUT (c), REED-BEDS (d), REED-BEDS-OUT (e).

Measurements and analyses: Bacterial abundance and sizeFor the determination of bacterial abundance, water samples (3×50 mL) were fixated

with 37% formaldehyde (end concentration: 10% v/v) within 30 min after sampling. After one hour of incubation, the samples were filtered over 0.2 µm polycarbonate filters (Ø47mm; Sartorius Stedim Biotech, Göttingen, Gremany) and stored at -20°C till further analyses. Bacteria were labelled using fluorescence in situ hybridization (FISH) according to Glockner et al. (1996), with a general bacteria probe mix, EUB338mix (Daims et al. 1999). After hybridization the filters were mounted on microscope slides and imbedded in anti-bleaching media (4:1), Citifluor (AF1; Citifluor Ltd., Leicester, UK) and Vectashield (Vectorshield Laboratories, Inc., Burlingame, USA). From each filter 5 - 10 randomly chosen images, each composed from the average of 16 frames, were taken using a confocal laser scanning microscopy (Nikon A1, Tokyo, Japan) at 600× magnification. The images were converted to binary images, after which the bacteria were sized and counted automatically using ImageJ (Collins 2007) with a diameter size threshold of 0.2 µm. Bacterial counts and average size are grouped per sampling location, resulting in 30 technical replicas for the CW samples and 15 for the surface waters samples. The data of several of the sites were not normally divided, therefore the data was statistically tested for differences between sites with Kruskal-Wallis tests using PAST (Hammer et al. 2001).

46

Chapter 3

Measurements and analyses: Community-level physiological profiling (CLPP)For the functional diversity and bacterial activity BIOLOG GN2 plates (Biolog, Inc.,

Hayward CA, USA) were used, a cell culturing method in which the microbial community is tested for the capacity to utilize 95 different carbon substrates. From each location water samples (3×50 mL) were taken and inoculation of the BIOLOG plates occurred within 8 h. Prior to inoculation, samples were filtered over a pre-washed nitrate-cellulose filter with pore size 5.0 µm (Ø47 mm; Sartorius Stedim Biotech, Göttingen, Gremany) and sonicated for 2 min. This pretreatment was performed to remove large grazers and acquire a homogeneous sample. The BIOLOG plates were incubated at 15°C with a light/dark cycle of 14/10 h. Following the manufactures protocol a plate reader (VersaMax™ Microplate reader, Molecular devices, Sunnyvale, USA) was used to determine the utilization of each carbon source every 24 h over a four days incubation period. The average well color development (AWCD) was calculated by subtracting the absorption of the blank from the signal in each substrate absorption before averaging all 95 different substrates. Differences between the AWCD after 96 h of incubation were analyzed with an ONEWAY-ANOVA using PAST (Hammer et al. 2001). For the community metabolic diversity, the absorption from the blank was subtracted from the absorption of each substrate, than substrates with a signal higher absorption than 0.2 in at least two of the three replicas per site were defined as being utilized. Data was converted to binary data representing that the carbon source was used or not. Principal component analyses were performed on the community metabolic diversity profiles from the different locations using PAST (Hammer et al. 2001).

Measurements and analyses: Bacterial community compositionSpecies composition was analyzed by DGGE for the general bacterial community

and methane oxidizers as an example of a specific functional group. This functional group was chosen as an example because generally methane is abundant is WWTP effluent and large community shifts were expected caused by changes in oxygen levels throughout the CW (Kadlec and Wallace 2008; Oremland 1988; Conrad 2007).

For determination of the total DNA content samples were filtrated on site over 0.2 µm cellulose nitrate membrane filters (Ø25mm; Whatman NC 20) and stored at -20°C prior to DNA extraction. DNA extraction was performed using PowerWater® DNA Isolation Kit (MO BIO laboratories Inc., Carlsbad, USA) according to the manufacturer’s instructions. For analyses of the total bacterial community general bacterial primers (F357GC and R518) were used which amplify the variable V3 region of 16S rDNA (Muyzer et al. 1993). PCR was conducted with the following cycling conditions: Initial denaturation: 94°C, 5 min Cycling step: 94°C, 30 sec, 54°C, 30 sec, 72°C, 1 min; 35 cycles; Final elongation 72°C, 8 min Reaction volumes were 50 µL containing: 1 µL of template DNA (5 ng µL-1), 8.75 µL PCR-grade water (Applichem, Darmstadt, Germany) 25 µL 2× premixture (Epicentre, Madison, USA), 5 µL primer GCA189 (5 pmol µL-1), 5 µL primer 661_nd (5 pmol µL-1), 5 µL BSA (1:5, Biolabs, Ipswich, USA) and 0.25 µL taq-polymerase (Invitrogen, New York, USA). Methanotroph

Planktonic bacteria

47

specific pmoA were amplified using a nested design which has been previously described (Steenbergh et al. 2010; Lin et al. 2005). The nested PCR was performed to obtain enough pmoA product for the analysis and to make sure only pmoA of MOB was amplified since also ammonium oxidizers have a similar gene which potentially could be amplified by the first primer-set (ammonia monooxygenase-α subunit) (Holmes et al. 1995). The first PCR amplification used the primers A189 and A682 (Holmes et al. 1995) and consisted of 35 cycles: Initial denaturation: 94°C, 5 min Cycling step: 94°C, 1 min, 56°C, 1 min, 72°C, 1 min; 35 cycles; Final elongation 72°C, 6 min. Reaction volumes of the first round were 25 µL containing: 1 µL of template DNA (5 ng µL-1), 3.88 µL PCR-grade Water (Applichem, Darmstadt, Germany) 12.5 µL 2× premixture (Epicentre, Madison, USA), 2.5 µL primer a682 (5 pmol µL-1), 2.5 µL primer a189 (5pmol µL-1), 2.5 µL BSA (1:5, Biolabs, Ipswich, USA) and 0.125 µL taq-polymerase (Invitrogen, New York, USA). Before the nested PCR was performed samples were diluted 5 till 200 times depending on the intensity of the bands on a 1% agarose gel. For the nested PCR primer set GCA189 and mb661_nd (Lin et al. 2005) was used and consisted of 25 cycles: Initial denaturation: 94°C, 5 min Cycling step: 94°C, 30 sec, 56°C, 30 sec, 72°C, 30 sec; 25 cycles; Final elongation 72°C, 5.5 min. Reaction volumes of the nested PCR were 50 µL containing: 1 µL of (diluted) template DNA from the first round, 8.75 µL PCR-grade Water (Applichem, Darmstadt, Germany) 25 µL 2× premixture (Epicentre, Madison, USA), 5 µL primer GCA189 (5 pmol µL-1) , 5 µL primer 661_nd (5 pmol µL-1), 5 µL BSA (1:5, Biolabs, Ipswich, USA) and 0.25 µL taq-polymerase (Invitrogen, New York, USA). All PCRs were performed in a MBS 0.2 S thermocycler (ThermoHybaid, Ashfort, UK).

After DNA extraction and amplification DGGE analysis was performed according to (Bodelier et al. 2005), with minor adjustments. The PCR products were separated on a 0.6 mm thick vertical gel consisting of 6% (w/v) polyacrylamide (37.5:1 acrylamide:bisacrylamide) and a linear gradient of the denaturants urea and formamide, increasing from 30% at the top of the gel till 70% at the bottom of the gel. The 100% denaturant was defined as 7 M urea with 40% v/v formamide. The gels were loaded with 45 µL of PCR product and 0.20 µL loading dye per µL PCR product. In total three clone ladders described by (Steenbergh et al. 2010) as mix nr.2 were loaded, one at each side of the gel and one in the middle to be able to calibrate the different gels. Electrophoresis was performed in a buffer containing 40 mM Tris, 40 mM acetic acid, 1 mM EDTA (pH 7.6) (0.5× Tris- acetate-EDTA buffer) for 17 h at 100 V (temp. 60°C). Gels were stained for 1 h in 0.1µL mL-1 Ethidiumbromide in 0.5× TAE-buffer. The bands were visualized by UV-light and then photographed. Images were analyzed using Phoretics 1D Advanced (5.20, Biosystematica, Llandysul, UK). The lanes were manually detected and 90% of the lane width was analyzed. Bands were manually detected using the pixel intensity graph. The retardant factor (rf) was calibrated using the ladder. The bands were converted to binary data reflecting if bands were either present or not present. From the DGGE band data dendrograms were constructed based on the Euclidean distance using PAST (Hammer et al. 2001).

48

Chapter 3

Measurements and analyses: Carbon and BOD5 measurementsTotal carbon (TC), inorganic carbon (IC) and total organic carbon (TOC)

concentrations were analysed using a total organic carbon analyser (Schimadzu, TOC-Vcph; Kyoto, Japan). To determine the concentrations of dissolved organic carbon (DOC) water samples were vacuum filtered over pre-washed 0.2 µm cellulose nitrate membrane filters (Whatman NC 20) and analysed with the same method as described for the TOC. Biochemical oxygen demand (BOD5) was analysed according to standardized methods (NEN-EN-1899-1 1998)

Results

Bacterial abundanceThe planktonic bacterial abundance at all sites ranged between 106 and 107 cells mL-1

(Fig. 3.2). At the inflow of the constructed wetland, the average bacterial abundance (±s.e.) was 3.2×106 ±4.8×105 cells mL-1 (Fig. 3.2). Between PONDS-IN and PONDS the abundance remained the same, but in the second halve of the unvegetated ponds (between PONDS and PONDS-OUT) the bacterial abundance increase to an average of 4.9×106 ±2.8×105 cells mL-1 (p<0.001) (Fig. 3.2). The same pattern was observed during residence in the reed beds, with no significant change in the bacterial abundance in the first halve of the reed beds, but an significant increase in the second halve of the reed beds to 7.9×106 ±7.8×105 cells mL-1 (p<0.001) (Fig. 3.2). The average bacterial abundance in the surface waters (comparison sites) showed high variation between sites, ranging from 2.2×106 ±6.3×105 cells mL-1 in the canal, to 1.2×107 ±6.7×105 cells mL-1 in the artificial fen (Fig. 3.2).

0,E+00

3,E+06

5,E+06

8,E+06

1,E+07

1,E+07

2,E+07

Bact

eria

l abu

ndan

ce

(no.

mL-1

)

Fig. 3.2 Average (±s.e.) bacterial cell abundance (illuminated by FISH) in the CW and various types of surface waters (comparison sites).

Planktonic bacteria

49

Metabolic activity and functional diversityResidence in the unvegetated ponds had no effect on the high Average Well Color

Development (AWCD) of 1.6 observed at PONDS-IN (Fig. 3.3), but during residence in the reed beds a significant decrease in maximum AWCD (to 1.2) was observed (p<0.001) (Fig. 3.3). The AWCD at the outflow of the reed beds was comparable the AWCD of the urban ditch samples, which were relatively high when compared to the other surface waters, ranging from 1.2 and 0.7 after 96 h incubation.

After 96 h incubation, all three samples in the unvegetated ponds showed a high functional diversity, with 93, 92 and 91 of the 95 available substrates being utilized at PONDS-IN, PONDS and PONDS-OUT respectively (Fig. 3.3). At REED BEDS and REED-BEDS-OUT the number of utilized carbon substrates decreased to 85 and 87 out of 95 respectively. The surface waters all showed lower diversity in carbon substrate utilization compared with the CW, ranging from 81 out of 95 in the urban ditch to 64 out of 95 in the peat lake. The principal component analyses (PCA) of the functional diversities showed separation of several locations: the unvegetated ponds, reed beds, the flowing water surface waters and the stagnant waters like the artificial fen and peat lake (Fig. 3.4). The first principal component explained 41% of the variance in functional diversity of carbon utilization and was mainly determined by the utilization of carboxylic acids and amino acids (Table 3.1). The second principal component explained 14% of the variance and was composed of a variety of carbon source types (Table 3.1).

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

0 24 48 72 96

AW

CD

Time (h-1)

Ponds in Ponds Ponds out

Reed beds Reed beds out

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

0 24 48 72 96

AW

CD

Time (h-1)

Urban ditch Agricultural ditch Canal

River Artificial fen Peat lake

0102030405060708090

100

0 24 48 72 96

Com

mun

ity m

etab

olic

div

ersi

ty

Time (h-1)



0102030405060708090

100

0 24 48 72 96

Com

mun

ity m

etab

olic

div

ersi

ty

Time (h-1)



0102030405060708090

100

0 24 48 72 96

Com

mun

ity m

etab

olic

div

ersi

ty

Times (h-1)



0102030405060708090

100

0 24 48 72 96

Com

mun

ity m

etab

olic

div

ersi

ty

Times (h-1)



Fig. 3.3 Community level physiological profiling (Biolog) average well color development (AWCD) and community metabolic diversity in the CW (PONDS-IN – REED-BEDS-OUT) and varies types of surface waters.

50

Chapter 3

Ponds inPonds

Ponds out

Reed bedsReed beds out

Artificial fen

River

Canal

Agricultural ditch

Peat lake

Urban ditch

-2

-1

0

1

2

-4 -3 -2 -1 0 1 2 3

Prin

cipa

l Com

pone

nt 2

(14%

)

Principal Component 1 (41%)

Fig. 3.4 Principal component analysis of the functional carbon utilization diversity in the unvegetated ponds (black triangles), reed beds (black circles) and comparison sites (open squares).

Table 3.1 Carbon source loads determining the principal component axis. Carbon source PC 1 PC 2 Carbon source PC 1 PC 2

Carbohydrates Amino acids

i-Erythritol 0.22 L-Threonine 0.25

Lactulose 0.20 0.35 L-phenylalanine 0.23

Carboxylic acids L-Ornithine 0.21 0.20

Sebacic Acid 0.24 Amides

α-Keto Valeric Acid 0.23 Glucuronamide 0.21 -0.26

α-Keto Butyric Acid 0.21 0.20 L-Alaninamide 0.20

p-Hydroxy-Phenylacetic Acid 0.21 0.20 Amines

Itaconic Acid 0.21 0.20 Phenyethylamine -0.29

D-Glucosaminic Acid 0.20 0.35 Methyl ester

Dicarboxylic acids Succinic Acid Mono-methyl Ester -0.40

Succinamic Acid 0.24 -0.26

Planktonic bacteria

51

Carbon and BOD5The total carbon (TC) concentration in the CW ranged between 60 to 100 mg L-1 and

did not show changes during residence in the CW (Fig. 3.5). The average inorganic fraction in all samples was 54 ±3 mg L-1 (Fig. 3.5). The organic carbon fraction was approximately 20 mg L-1 throughout the CW and mainly consisted on dissolved organic carbon (DOC), with a marginal fraction of carbon present in the form of particulate organic carbon (POC) (Fig. 3.5). The average biological oxygen demand (BOD5) at RED-BEDS-OUT (1.55 ±0.12 mg L-1) was significantly lower than at PONDS-IN (3.20 ±0.60 mg L-1) (p<0.05). The ratio between the biological oxygen demand (BOD5) and the DOC (BOD5/DOC) decreased from 0.14 to 0.08 during residence in the unvegetated ponds and was 0.07 at REED-BEDS-OUT.

0

20

40

60

80

100

120

Car

bon

(mg

C L

-1)

TC

TOC

DOC

Fig. 3.5 Average concentration of total carbon (TC), organic carbon (TOC) and dissolved organic carbon (DOC) in different locations in the CW. N=3-5; ±s.d.

Bacterial community compositionThe denaturant gradient gel electrophoresis (DGGE) of general bacteria resulted in

22 bands at PONDS-IN (Fig. 3.6). This relative high number of bands was reduced to 11 at PONDS and to 10 at PONDS-OUT. At the end of the reed beds the number of bands decreased further to 7. Although four bands at PONDS IN were also present at REED-BEDS-OUT, only one band present at PONDS-IN remained present at all locations in the CW. The number of bands observed in most of the surface waters ranged between four (agricultural ditch, canal) and six (peat lake), with the artificial fen as the only exception with 13 bands. Cluster analyses of the band patterns showed a strong separation of PONDS-IN and the artificial fen from all other sites. The surface waters (except the artificial fen) clustered closely together and also include the REED-BEDS-OUT. The other sites of the CW (ponds, PONDS-OUT and REED-BEDS) clustered together and showed a lower Euclidean distance from the comparison sites compared to PONDS-IN (Fig. 3.6).

52

Chapter 3

The CH4 oxidizing bacteria showed similar numbers of dominant bands throughout the CW ranging from 4 (reed beds) to 7 (PONDS-IN) (Fig. 3.7). The band composition was however highly variable: dominant bands at PONDS-IN were, for example, all replaced by other dominant bands at PONDS. Similar shifts in band patterns were observed between the other locations in the CW. Besides the agricultural ditch (1 dominant band) the surface waters showed similar numbers of dominant bands ranging between 3 (urban ditch) and 7 (artificial fen), but showed few mutual bands between different surface waters. This high diversity of dominant bands between sites was reflected in the cluster analyses of the CH4 oxidizing bacteria band patterns (Fig. 3.7).

Ponds inArtificial fen

Peat lakeRiverCanal

Agricultural ditchUrban ditch

Reed beds outReed bedsPonds out

Ponds

5

4,54

3,53

2,52

1,51

0,5

Euclidean distance

PondsPonds outReed bedsReed beds outUrban ditchAgricultural ditchCanalRiverPeat lakeArtificial fenPonds in

Euclidean distance0.0 4.54.03.53.02.52.01.51.00.5

Fig. 3.6 Euclidean distance dendrogram of the general bacteria DGGE in the CW and various types of surface waters (comparison sites).

43,63,22,82,421,61,20,80,40

Distance

River

Reed beds

Agricultural ditch

Artificial fen

Urban ditch

Reed beds out

Ponds in

Canal

Peat lake

Ponds

Ponds out

Euclidean distance0.0 4.03.53.02.52.01.51.00.5

Ponds in

Artificial fen

Peat lake

River

Canal

Agricultural ditch

Urban ditchReed beds out

Reed beds

Ponds outPonds

Fig. 3.7 Euclidean distance dendrogram of CH4 oxidizing bacteria in the CW and various types of surface waters (comparison sites).

Planktonic bacteria

53

Discussion

Bacterial community at the inflow of the CWThe bacterial abundances observed were all in range of previously determined

bacterial abundances in surface waters (Sanders et al. 1992; Glöckner et al. 1999), but the bacterial abundance at the inflow of the CW was on the low side in comparison with the comparison surface waters, while a relative high bacterial abundance was expected in the treated wastewater. Usually a large fraction of planktonic bacteria is associated with particles (Grossart et al. 1998) and the low inflow of suspended particles probably causes a relative low number of bacteria to enter the CW. Nonetheless, the metabolic activity and functional diversity of the bacterial community at the inflow of the CW was very high in comparison with the surface water samples, indicating that the bacterial community at the inflow of the CW consisted of a high percentage of active bacteria. The presence of a metabolic active and functional divers microbial community is important for the functioning of WWTPs (Tchobanoglous et al. 2004) and is therefore well maintained by providing favorable conditions (oxygen, temperature, resources). Previous research showed that 80% of the bacteria in activated sludge from WWTPs is metabolic active (Nielsen and Nielsen 2002) against 8 to 47% in the water column of lakes (Haglund et al. 2002). This high metabolic active fraction in activated sludge was indeed reflected in the high activity of the bacterial community at the inflow of the CW.

Unvegetated pondsHigh concentrations of DOC at PONDS-IN provide a large pool of substrates for

the heterotrophic bacterial community in the CW. The concentration of DOC remained the same during residence in the unvegetated ponds, but the ratio between the biological oxygen demand (BOD5) and the DOC (BOD5/DOC) decreased from 0.14 to 0.08. This indicates that the composition of the DOC changed and decreased in quality which has been often observed in other constructed wetlands (Kadlec and Wallace 2008), probably caused by a combination of removal of labile carbon substrates by metabolic uptake and abortion to POM, the formation of new DOC compounds in the system by degradation of POM and photochemical degradation of DOC (Lindell et al. 1995). However, the bacterial metabolic activity and functional diversity of the bacterial communities remained the same during residence in the unvegetated ponds and did not reflect the observed decrease in substrate quality. Moreover, the abundance of bacteria increased, indicating relative decrease in metabolic activity per cell. The DGGE analyses showed that the number of bacterial taxa decreased during residence in the unvegetated ponds, indication either partial turnover of bacterial community of shifts in the dominance between the original occurring taxa during residence in the unvegetated ponds.

54

Chapter 3

These dynamics between bacterial species were well illustrated by the changes in the composition of methane oxidizing bacterial community in the unvegetated ponds, showing a shift towards a strong dominance of a type-Ia taxa which is regarded as a fast growing, but bad competitor (r-type) group of organisms (Steenbergh et al. 2010), indicating favorable conditions (Tanner et al. 1997). This primarily includes sufficient supply of organic matter (Conrad 2007) and low oxygen concentrations (Oremland 1988). Earlier research in this constructed wetland (van den Boomen et al. 2012) showed that the oxygen levels increase from almost anoxic at the PONDS-IN to 3 to 4 mg L-1 at PONDS-OUT, combined with high sedimentation fluxes of organic particles in the first part of the constructed wetland (van den Boomen et al. 2012).

Reed bedsIn the reed beds, the bacterial metabolic activity and functional diversity decreased

to levels comparable with the communities of the urban and agricultural ditches. Similar as in the unvegetated ponds, this decrease in activity coincided with an increase in bacterial abundance, even further reducing the relative bacterial metabolic activity per cell. The DGGE patterns showed that these changes are probably related to changes in the bacterial community composition. This is supported by the changes observed in the metabolic diversity were the utilization of carbon sources is reduced, especially (di)carboxylic acids and amino acids. The decreasing capacity of the bacterial community to utilize these type of carbon course is also the mean difference between CW influent and the comparison surface waters. The PCA analyses also showed that residence in the CW changes the bacterial functional diversity towards the urban and agricultural ditch systems. These two ditch systems closely resembled wetland ecosystems with relative high abundance of macrophytes, which indicates that the bacterial community at the outflow of the CW may largely consists of bacteria from originating from the last compartment, the reed beds. Surface areas in the reed beds, including reed stems, plant litter and sediment are generally colonized by biofilms which are not only known to retain high amounts of suspended particles and bacteria from the water column (Balzer et al. 2010; Chabaud et al. 2006; Stott and Tanner 2005; Eisenmann et al. 2001), but simultaneously also release large numbers of bacteria into the surrounding water column (McDougald et al. 2012; Picioreanu et al. 2001). The large shift in the planktonic bacterial community characteristics during residence in the reed beds is therefore expected to be caused by strong exchange between water and solid surfaces. Although it is notable that there were small differences between the different measurements, both the general taxa composition and the metabolic activity of the bacterial community supports this strong shift in community composition during residence in the reed beds to resemble similar natural ecosystems.

Planktonic bacteria

55

ConclusionThis study showed that constructed wetlands receive a metabolic highly active bacterial

community together with a high load of carbon substrates from the wastewater treatment plant. In the different functional compartments of the CW, changes in various biological, chemical and physical conditions occur, inducing many changes in the characteristics of the bacterial community. The planktonic bacterial community flowing out of the CW resembles communities of physically similar natural ecosystems. Constructed wetlands are therefore suitable to reduce the input and impact of anthropogenic bacterial communities discharged by wastewater treatment plants into receiving surface waters.

AcknowledgmentsThis work was financed by the Stichting Waternet and supported by Foundation

for Applied Water Research (STOWA), Witteveen+Bos, Wetterskip Fryslân and NIOO-KNAW. Special thanks go out to the people that supported this research, Richard Soeter, Paul Bodelier, Maxine Bogaert en Rinse van der Kooij.

56

Chapter 3

References

Balzer M, Witt N, Flemming H–C and Wingender J (2010). Faecal indicator bacteria in river biofilms. Water Science and Technology 61 (5), 1105-1111.

Bodelier PLE, Meima-Franke M, Zwart G and Laanbroek HJ (2005). New DGGE strategies for the analyses of methanotrophic microbial communities using different combinations of existing 16S rRNA-based primers. FEMS microbiology ecology 52 (2).


Collins TJ (2007). ImageJ for microscopy. BioTechniques 43 (1), 25-30.

Conrad R (2007). Microbial ecology of methanogens and methanotrophs. Advances in Agronomy 96, 1-63.

Daims H, Bruhl A, Amann R, Schleifer KH and Wagner M (1999). The domain-specific probe EUB338 is insufficient for the detection of all Bacteria: Development and evaluation of a more comprehensive probe set. Systematic and applied microbiology 22 (3), 434-444.

Diaz FJ, O’Geen AT and Dahlgren RA (2010). Efficacy of constructed wetlands for removal of bacterial contamination from agricultural return flows. Agricultural Water Management 97 (11), 1813-1821.



Glockner FO, Fuchs BM and Amann R (1999). Bacterioplankton compositions of lakes and oceans: a first comparison based on fluorescence in situ hybridization. Applied and Environmental Microbiology 65 (8), 3721-3726.


Grossart H, Berman T, Simon M and Pohlmann K (1998). Occurrence and microbial dynamics of macroscopic organic aggregates (lake snow) in Lake Kinneret, Israel, in fall. Aquatic Microbial Ecology 14 (1), 59-67.

Planktonic bacteria

57

Haglund A, Tornblom E, Bostrom B and Tranvik L (2002). Large differences in the fraction of active bacteria in plankton, sediments, and biofilm. Microbial ecology 43 (2), 232-241.



Holmes AJ, Costello A, Lidstrom ME and Murrell JC (1995). Evidence that Particulate Methane Monooxygenase and Ammonia Monooxygenase may be Evolutionarily Related. FEMS microbiology letters 132 (3).

Kadlec RH and Wallace SD (2008) Treatment wetlands. CRC Press, Florida.

Kalff J (2002). The Bacteria. In: Kalff J. Limnology. Prentice-Hall, Inc., Englewood Cliffs, 349-375.

Karim MR, Manshadi FD, Karpiscak MM and Gerba CP (2004). The persistance and removal of enteric pathogens in constructed wetlands. Water Research 38 (7), 1831-1837.

Lin JL, Joye SB, Scholten JCM, Schafer H, McDonald IR and Murrell JC (2005). Analysis of methane monooxygenase genes in mono lake suggests that increased methane oxidation activity may correlate with a change in methanotroph community structure. Applied and Environmental Microbiology 71 (10).

Lindell M, Grane´li W, Tranvik LJ (1995). Enhanced bacterial growth in response to photochemical transformation of dissolved organic matter. Limnology and Oceanography 40 (1), 195–199.

McDougald D, Rice SA, Barraud N, Steinberg PD and Kjelleberg S (2012). Should we stay or should we go: mechanisms and ecological consequences for biofilm dispersal. Nature Reviews Microbiology 10 (1), 39-50.

Molleda P, Blanco I, Ansola G and de Luis E (2008). Removal of wastewater pathogen indicators in a constructed wetland in Leon, Spain. Ecological Engineering 33 (3-4), 252-257.

Mulling BTM, van den Boomen RM, van der Geest HG, Kappelhof JWNM and Admiraal W (2013). Suspended particle and pathogen peak discharge buffering by a surface-flow constructed wetland. Water research 47 (3), 1091-100.

Muyzer G, de Waal EC and Uitterlinden AG (1993). Profiling of complex microbial-populations by denaturing gradient gel-electrophoresis analysis of polymerase chain reaction-amplified genes-coding for 16S ribosomal-RNA. Applied and Environmental Microbiology 59 (3), 695-700.

58

Chapter 3

NEN-EN-1899-1 (1998). Water quality - Determination of biochemical oxygen demand after n days (BODn) - Part 1: Dilution and seeding method with allylthiourea addition.

Nielsen J and Nielsen P (2002). Quantification of functional groups in activated sludge by microautoradiography. Water Science and Technology 46 (1-2), 389-395.

Oremland RS (1998). Biogeochemistry of methanogenic bacteria. In: Zehnder AJB (ed.). Biology of anaerobic microorganisms. John Wiley & Sons, New York, 641–705.

Picioreanu C, van Loosdrecht M and Heijnen J (2001). Two-dimensional model of biofilm detachment caused by internal stress from liquid flow. Biotechnology and bioengineering 72 (2), 205-218.


Sanders R, Caron D, Berninger U (1991). Relationships between Bacteria and Heterotrophic Nanoplankton in Marine and Fresh Waters - an Inter-Ecosystem Comparison. Marine Ecology Progress Series 86 (1), 1-14.

Seviour RJ and Nielsen PH (2010) Microbial communities in activated sludge plants. In: Seviour RJ and Nielsen PH (Eds.). Microbial ecology of activated sludge. IWA publishing, London, 95-125.

Steenbergh AK, Meima MM, Kamst M and Bodelier PLE (2010). Biphasic kinetics of a methanotrophic community is a combination of growth and increased activity per cell. FEMS microbiology ecology 71 (1).



Tanner C, Adams D and Downes M (1997). Methane emissions from constructed wetlands treating agricultural wastewaters. Journal of environmental quality 26 (4), 1056-1062.

Tchobanoglous G, Burton FL and Stensel HD (2004) Wastewater Engineering: Treatment and Reuse. McGraw-Hill, New York.

van den Boomen R, Kampf R (2012). Waterharmonica’s in the Netherlands 1996-2011: from WWTP effluent till usable surface water. STOWA publication 2012-12. (in Dutch)

Planktonic bacteria

59

van den Boomen R, Kampf R and Mulling BTM(2012). Research on suspended particles and pathogens in the Waterharmonica (constructed wetland).. STOWA publication 2012-11. (in Dutch)


Wotton RS (1994) The biology of particles in aquatic systems. CRC Press, Inc., Boca Raton.

Processes removing faecal indicator organisms in constructed wetlands

Chapter 4

Submitted as: Mulling BTM, van der Oost R, van der Wielen PWJJ, van der Geest HG, Admiraal W (2013). Processes removing faecal indicator organisms in constructed wetlands.

62

Chapter 4

Abstract Treatment of municipal wastewater by wastewater treatment plants strongly reduces the numbers of faecal indicator organisms and pathogens, but in discharged treated wastewater numbers are still high enough to be a major anthropogenic input of faecal indicator organisms and pathogens into receiving surface waters, potentially threatening recreational purposes and drinking water production. The natural purification capacity of wetlands has long been recognized and man-made constructed wetlands are widely used to improve the water quality of treated wastewater before discharge into receiving surface waters. Although the removal efficiencies of constructed wetlands are well-studied, the importance of different processes involved in the removal of faecal indicator organisms and pathogens are often not analysed. In this study we monitored the monthly numbers and removal of several bacterial, protozoan and viral faecal indicator organisms in a full scale constructed wetland receiving treated municipal wastewater. In addition we performed small scale experiments to estimate the importance of individual processes influencing pathogen dynamics including sedimentation, predation, UV irradiance, mortality and external input. The results showed substantial removal of faecal indicator organisms during residence in the constructed wetland, with total removal efficiencies ranging between 96 and 99% for bacterial faecal indicator, 89% for a faecal protozoan indicator and 73 and 88% for the viral indicators. Zooplankton predation and biofilm retainment appeared to be the most important processes governing the removal of faecal indicator organism removal in the constructed wetland, whereas the effects of sedimentation, inactivation and UV-radiation are of minor importance. No significant reintroduction of faecal indicator organisms by waterfowl or other warm blooded animals was observed, although very high abundance of waterfowl were present in the constructed wetland during freezing periods.

Faecal indicator organisms

63

Introduction Wastewater discharge is one of the major anthropogenic inputs of faecal pathogens into surface waters (Arnone and Walling 2007; Tchobanoglous et al. 2004). Although treatment of raw wastewater by wastewater treatment plants (WWTPs) strongly improves the hygienic water quality (Wen et al. 2009; Tchobanoglous et al. 2004; Wery et al. 2008) treated wastewater still contains high numbers of faecal pathogenic organisms (Seviour and Nielsen 2010), compromising the hygienic quality of receiving surface waters (Holeton et al. 2011; Tchobanoglous et al. 2004). To reduce the faecal pathogen load of surface water, additional polishing of treated wastewater is often desirable. Several technical methods have been developed to remove pathogenic organisms, including UV treatment, slow sand filtration, membrane filtration and addition of reactive oxygen species (Gomez et al. 2006). These technical methods are, however, relatively expensive and energy consuming, and can produce harmful transformation products that are discharged into receiving surface waters (Watson et al. 2012; Tchobanoglous et al. 2004). Alternatively, constructed wetlands (CWs) are widely used biological systems for polishing treated wastewater and it has been demonstrated that these wetlands are sustainable and low cost systems for improving the hygienic quality of treated wastewaters (Molleda et al. 2008; Vymazal 2005; Kadlec and Wallace 2008). Although the effects of CWs on removing faecal indicators have been studied extensively (Kadlec and Wallace 2008; Vymazal, 2005 Reinoso et al. 2011; Moleda et al. 2008; Karim et al. 2004; Ghermandi et al. 2007, Diaz et al. 2010), the majority of these studies focus only on the net removal efficiencies, without quantifying the different processes involved. The processes involved in the removal of faecal indicator or pathogenic organisms from treated wastewater during residence in CWs include sedimentation, predation, natural mortality, biofilm retainment, UV inactivation, resuspension and population growth (Brookes et al. 2004; Chabaud et al. 2006; Stott and Tanner 2005; Kadlec and Wallace 2008). The aim of this study was to quantify the effects of these different processes on the removal of faecal bacterial and viral indicators in constructed wetlands. To meet this aim, monthly water samples were colected from different compartments of a full-scale surface flow constructed wetland (unvegetated ponds and reed beds) receiving treated municipal wastewater. Numbers of faecal indicators for bacterial pathogens, Escherichia coli, Enterococci and Bacteroidales (Gerba 2000), indicators for protozoa pathogens, Clostridium perfringens, and indicators for viral pathogens (Payment and Franco 1993; Gerba 2000), Bacteroides phages (Ebdon et al. 2012; Gerba 2000), were analyzed. Furthermore, samples were also analyzed for zooplankton abundance and species composition to estimate the species specific predation capacity. Finally, additional laboratory and field experiments were conducted to estimate the effect of temperature, total predation, UV radiation and sedimentation.

64

Chapter 4


Treatment plant and constructed wetland This study was carried out in a full scale surface flow constructed wetland (CW) located in Grou, The Netherlands. The CW was built in 2006 and primarily receives a constant hydraulic loading of 1200 m3 day-1 treated municipal wastewater. The inflow of the constructed wetland leads treated waste water through a series of three unvegetated ponds and four parallel reed beds, before being pumped into receiving surface water (channel) (Fig. 4.1). The unvegetated ponds are open water systems without vegetation with an average depth of 1.35 m, volume between 360 and 440 m3 each and total hydraulic retention time (HRT) of 17.9 h (Fig 4.1; Table 4.1). The four reed beds each have an average water depth of 40 cm, approximate volume of 443 m3, are covered with Phragmites australis and receive ca. 300 m3 treated wastewater day-1 with an average HRT of 23.6 h. The total HRT of the CW was 41.5 h. HRTs were calculated from the residence time distribution obtained by a tracer experiment using lithium chloride (van den Boomen et al. 2012). The average HRT was determined at 50% passage of the lithium chloride load.

Table 4.1 Dimensions and hydraulic retention time (HRT) of constructed wetland, Aqualân in Grou, The Netherlands. The length, width and volume of the ponds were manually determined and calculated, the length, width and volume of the reed beds were calculated from the construction blueprints. The HRT’s were determined by a tracer experiment.



Length (m) 55 55 55 165 110 110

Average width (m) 7.6 8.1 8.1 7.9 11.5 46

Average depth (m) 1.34 1.31 1.43 1.35 0.4 0.4

Surface area (m2) 418 446 446 1304 1265 5060 6364

Volume (m3) 362 388 441 1191 443 1771 3552


HRT (h) 17.9 23.6 41.5

Sampling Water samples were taken monthly from January till December 2010, between 09.00 and 11.00 am at the in- and outflow of each compartment (unvegetated ponds and reed beds; Fig. 4.1). Sampling moments for the wetland sampling locations were not corrected for the hydraulic residence time which could in combination with possible daily variation in faecal indicator organism concentrations influence the removal efficiency calculations. In the middle of the unvegetated ponds and reed beds samples were taken for zooplankton


65

analyses (Fig. 4.1). For zooplankton identification and counting, two types of water samples were taken from April until December 2010. For larger zooplankton taxa water samples (10 L) were filtered over a 30 µm filter. For Ciliophora and Amoebozoa smaller than 30 µm water samples (3 L) were fixated with Lugol and left for settling for 5 days. In addition, sedimentation traps were placed at twelve locations in the unvegetated ponds (5, 4 and 3 locations in first, second and third unvegetated pond respectively (Fig. 4.1)). At each location eight sedimentation traps with a height of 33 cm, a diameter of 5cm and a volume of 650 mL, were placed on top of the sediment. After 29, 49, 93 and 168 h two sedimentation traps (total volume of 1.3 L) from each of the twelve locations were retrieved, the content was well mixed and transferred in 2 L glass bottles. Samples were taken from these bottles to determine the concentration of E. coli in the same manner as normal water samples (paragraph 2.3). Prior to further analysis all samples were stored at 4°C.

1

2

2

2

3 3 3 3

4

5

a

b

c

0 25 50 m

0

Fig. 4.1 Map of the WWTP in Grou, the Netherlands (0) with a sedimentation tank (1) which discharges treated municipal wastewater into a constructed wetland consisting of unvegetated ponds (2), reed beds (3) and an ecological buffer zone (4) which is in open connection with the receiving channel (5). Sampling points were located at; PONDS-IN (a), PONDS-OUT (b), REED-BEDS-OUT (c). At twelve locations in the unvegetated ponds (2) sedimentation traps were placed indicated with white dots. Zooplankton samples were collected in the middle of the unvegetated ponds and reed beds indicated by the open circles.

66

Chapter 4

Faecal indicator organisms E. coli and enterococci numbers were determined by membrane filtration according to standardized procedures NEN-EN-ISO-9308-1 (2000) and NEN-EN-ISO-7899-2 (2000), respectively. C. perfringens numbers were determined by placing filtered samples (Cellulose Nitrate; Ø47mm) on Triptose Sulfite Cyclocerine medium (TSC-medium) under anaerobic conditions for 24 ±2 h at 45 ±1°C sequentially for confirmation onto Mobility-Nitrate reduction medium (BN-medium) and Lactose-Gelatine medium (LG-medium) and incubated at 37 ±1°C for 24 ±2 h. This method is based on standardized procedures ISO/TC-147/SC4/WG5 (1995), ISO-6461-2 (1986) and NEN-EN-ISO-8199 (2007). Bacteroidales were quantified by qPCR (van der Wielen and Medema, 2010). In short, 100 ml of a water sample was filtered over a 25 mm polycarbonate filter (0.22 µm pore size). DNA was isolated using a FastDNA spin kit for soil (Qbiogene, US) according to the supplier’s protocol. Subsequently, primer sets Allbac 296F and Allbac 412R were used with the Taqman probe Allbac375Bhqr to quantitatively determine the 16S rRNA gene copy numbers of Bacteroidales spp in the water samples using a real-time PCR apparatus. For determination of the plaque forming units of Bacteroides phages, plates with Bacteroides host cells were incubated with water samples according to standardized procedure (ISO-10705-4 2001). Two different host strains of Bacteroides were used to enable differentiation between general and human specific Bacteroides phages. For the concentration of general Bacteroides phages Bacteroides fragilis was used as host (ISO-10705-4 2001), whereas for the human specific Bacteroides phages Bacteroides strain HB13 was used as host (Payan et al. 2005).

Zooplankton The abundance of larger zooplankton taxa (>30 µm) was determined by a two stepped inverted light microscope (Olympus; IX70-71; minimal magnification of 60×) counting procedure. First samples are left to settle for 15 min before counting zooplankton. Subsequently, samples were fixated with lugol, left for another 15 min to settle and recounted. The difference between the first and second enumeration was interpreted as the number of living organisms. For counting the number of Ciliophora and Amoebozoa smaller than 30 µm, water samples (3 L) were fixated with Lugol and left to settle for 5 days. Overlying water was carefully removed and the remaining sample (4-5 mL) was transferred into a smaller container. Numbers were determined at a magnification of 400-600× using an inverted light microscope.

Laboratory incubation 20 L water samples were collected at the inflow of the CW (PONDS-IN) of which halve was filtered through 5 µm polycarbonate membranes filters (Ø47mmm; Nuclepore™, California USA) on location to remove the zooplankton. Three 1 L samples of both the filtered and unfiltered samples were incubated at 10 and 20°C in presence and absence of UV radiation (Exoterra Repti Glo 40 Watt; 57% PAR, 33% UVA, 10% UVB; 5 cm distance to


67

water surface). All samples were incubated under a 16/8 light/dark-cycle. All 24 h samples were taken from each incubation and transferred onto Coliscan® S Easygel® (Micrology Laboratories, Goshen, IN, USA) plates, according to the manufacturer’s instructions. The plates were incubated at 36°C, and after 20 h of incubation, plates were photographed. The number of colony forming units (CFU) of E. coli were counted with ImageJ 1.44 software (http://rsb.info.nih.gov/ij/) using the Cell Counter plugin.

Data analyses To obtain monthly removal rates for both the unvegetated ponds as reeds beds, the outflow numbers (log10) were subtracted from the inflow numbers (log10). The removal data was normally distributed and to test if average removal of the indicator organisms differed significantly from zero one sample t-tests were performed (PAST; Hammer et al. 2001). To test differences in the removal of indicator organisms between the ponds and reed beds paired t-test were performed (PAST; Hammer et al. 2001).Sedimentation fluxes (±s.e.) were calculated by linear regression analyses over the four sampling moments at each of the twelve sampling points using SPSS (version 17.0). Conversion factors between abundance and predation capacity were gathered from literature (Table 4.2). Several low abundance zooplankton species of groups including several Amoebozoa species, Ceriodaphnia and Rotifera sp. are expected bacteriovores (Knight and Waller 1992; Abrantes and Goncalves 2003), but the predation capacity for these zooplankton species is not described in literature and were therefore estimated using taxonomically related species (Table 4.2). Correlations between the predation capacity of zooplankton species and log removal of faecal indicators were calculated using SPSS (version 17.0). For normally distributed data a Pearson’s correlation test was performed, otherwise a Spearman’s rho test was performed.

Results

Removal of faecal indicators in constructed wetlands The average number of faecal bacterial indicators ranged between 103 and 106 CFU L-1 at PONDS-IN (Fig 4.2a-c). The number of C. perfringens at PONDS-IN showed no large monthly changes throughout the year, but a trend of lower numbers in summer compared with winter was observed (Fig 2c). E. coli and enterococci at PONDS-IN showed strong variations between consecutive months throughout the year (Fig 4.2a,b). 16S rRNA gene copies of Bacteroidales at PONDS-IN ranged between 108 and 1011 gene copies L-1 (Fig 4.2d) and similar to the other bacterial indicators Bacteroidales showed lower numbers in summer compared with winter (Fig 42d). The number of the faecal viral indicator B. fragillis phages at PONDS-IN was around 103 plaque forming units (PFU) L-1, whereas Bacteroides strain HB13 phages ranged between 101 and 103 PFU L-1 (Fig 4.2e,f).

68

Chapter 4

0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

B. H

B13

pha

ges

(log 1

0PF

U L

-1)

Ponds in Ponds out Reed beds out

0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0


B. fr

agili

s pha

ges

(log 1

0PF

U L

-1)


0.0

2.0

4.0

6.0

8.0

10.0

12.0


Bact

erio

dale

s (lo

g 10

gene

cop

ies L

-1)


0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0


C. p

erfr

inge

ns(lo

g 10

CFU

L-1

)


0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0


Ente

roco

cci (

log 1

0C

FU L

-1)


0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0


E. c

oli(

log 1

0C

FU L

-1)


a)

b)

c)

d)

e)

f)

0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0


E.co

li(lo

g 10

CFU

L-1

)


Fig. 4.2 Numbers of fecal indicator organisms in colony (a-c) and plague (e-f) forming units and copies (d) per liter in the constructed wetland.

Both bacteriophages showed relative high numbers in summer (July/August). Most faecal indicators also showed relative stable outflow numbers (REED-BEDS-OUT) throughout the year, with similar seasonal fluctuations as seen at the inflow (PONDS-IN) (Fig 4.2). However, high numbers of E. coli and enterococci at the outflow were found in January, February, August and September. During residence in the unvegetated ponds, numbers of most faecal indicators significantly decreased (p<0.05), with average log removal efficiencies of 0.29 ±0.09, 0.48 ±0.17 and 0.30 ±0.08 log10 CFU for respectively E. coli, enterococci and C. perfringens, 0.33 ±0.06 log10 gene copies of Bacteroidales, and 0.54 ±0.16 log10 PFU of B. fragillis phages (Fig. 4.3). With an average log removal of 0.17 ±0.14 log10 PFU, numbers of Bacteroides HB13 phages did not decrease significantly during residence in the unvegetated ponds (Fig. 4.3). Residence in the reed beds caused significant reduction in numbers of all faecal indicators


69

(p<0.05), with an average log removal of 1.19 ±0.29, 0.88 ±0.20 and 0.64 ±0.10 log10 CFU of E. coli, enterococci and C. perfringens, respectively, 1.67 ±0.08 log10 gene copies of Bacteroidales, and 0.63 ±0.10 and 0.49 ±0.12 log10 PFU of B. fragillis phages and Bacteroides HB13 phages, respectively (Fig. 4.3). Although for all faecal indicator organisms a trend of higher removal in the reed beds compared with the unvegetated ponds was observed, this difference was only significant for E. coli, C. perfringens and Bacteroidales (p<0.05; Fig 4.3). Over the total constructed wetland the average removal for bacterial faecal indicators ranged between 96 and 99%, for protozoan faecal indicator was 89% and viral indicator between 73 and 88%.

0.00.20.40.60.81.01.21.41.61.82.0

E. Coli(CFU)

Enterococci(CFU)

C. perfringens(CFU)

Bacteroides(gene copies)

B.fragilis-phages(PFU)

B.HB13-phages(PFU)

Indi

cato

r or

gani

sm r

emov

al (

log 1

0)

Ponds Reed beds

E. coli(CFU)

0.00.20.40.60.81.01.21.41.61.82.0

E. Coli(CFU)

Enterococci(CFU)

C. perfringens(CFU)

Bacteroides(gene copies)

B.fragilis-phages(PFU)

B.HB13-phages(PFU)

Indi

cato

r or

gani

sm r

emov

al (

log 1

0)

Ponds Reed beds

Enterococci(CFU)

C. perfrigens(CFU)

Bacteroidales(gene copies)

B. fragilis phages(PFU)

B. HB13 phages(PFU)

Fig. 4.3 Average removal of faecal indicator organisms (±s.e.; n=6-12) in the unvegetated ponds (black bars) and the reed beds (white bars).

-20000

-15000

-10000

-5000

0

5000

10000

15000

0 55 110 165

Sedi

men

tatio

n flu

x (C

FU c

m-2

d-1)

Distance from discharge (m)

Pond 1 Pond 2 Pond 3

Fig. 4.4 The sedimentation fluxes of E. coli in the unvegetated ponds. Error bars represent uncertainty in calculated (linear regression) sedimentation fluxes.

70

Chapter 4

Sedimentation E. coli numbers in the sedimentation traps varied strongly between sampling moments and resulted in large error margins in the calculated sedimentation fluxes of E. coli, especially in the first unvegetated pond (Fig. 4.4). No significant sedimentation of E. coli was observed in the unvegetated ponds. Significant sedimentation of suspended matter (accounting for 10% of total inflow) was observed in the same sedimentation traps (van den Boomen et al. 2012), indicating proper functioning of the sedimentation traps.

Bacteroides phages The ratio between bacteriophages infecting B. fragillis and Bacteroides HB13 was on average (±s.e.) 0.61 ±0.04, 0.61 ±0.04 and 0.53 ±0.07 at respectively PONDS-IN, PONDS-OUT and REED-BEDS-OUT. The ratio between bacteriophages infecting B. fragillis and Bacteroides HB13 showed no significant changes during residence in the unvegetated ponds or the reed beds, which indicates no input of faecal matter from warm blooded animals (including waterfowl) into the CW and a comparable removal rate of bacteriophages infecting B. fragilis and Bacteroides phages infecting Bacteroides strain HB13.

Zooplankton abundance and predation capacity In the unvegetated ponds the average abundance of zooplankton (bacteriovores) was 2.2×105 ±8.2×104 individuals L-1 (±s.e.), the average abundance in the reed beds was similar with 2.1×105 ±6.1×104 individuals L-1 (Table 4.2). At both locations in the CW, >99% of the total zooplankton individuals identified were classified as bacteriovores. The zooplankton community showed seasonal variation with peak abundances in May, August and November and were observed in both the unvegetated ponds and reed beds (data not shown). The most abundant group of organisms in the unvegetated ponds and reed beds were Amoebozoa smaller than 30µm, followed by the Ciliophora smaller than 30 µm (Table 4.2). The community of zooplankton taxa larger than 30µm was dominated by Ciliophora species including Vorticella (Table 4.2). On average the abundance of Ciliophora as well as the bacteriovore species richness was higher in the reed beds compared with the unvegetated ponds (Table 4.2). The estimated average predation capacity of the total zooplankton community in the unvegetated ponds was 1.3×107 ±4.9×106 bacteria h-1 L-1 (±s.e.), which was not significantly different in the reed beds where an average predation capacity of 2.1×107 ±8.0×106 bacteria h-1 L-1 was estimated (Table 4.2). In both the unvegetated ponds and the reed beds the Ciliophora were in general the main contributors to the zooplankton community predation capacity (>90%) with the largest group being small Ciliophora (>73%). The small Amoebozoa were a stable but minor contributor (4%) to the community predation capacity (Table 4.2). The predation capacity of the two most abundant zooplankton groups in the unvegetated ponds, small Ciliophora and Amoebozoa showed seasonal variation, with periods of low (June, July, Sep, Dec) and periods of high (August, November) predation capacity corresponding with the dynamics of E. coli removal (Fig 4.5). The predation capacity


71

of small Amoebozoa showed a significant positive correlation with the log removal of E. coli in the unvegetated ponds (0.88; p<0.05). Some of the other zooplankton species, with lower contributions to the community predation capacity, did also show significant correlations between predation capacity and the log removal of certain specific faecal indicators as well. In the unvegetated ponds Nauplius-larvae correlated positively with the removal of C. perfringens (0.86; p<0.05) and Bacteroides HB13 phages (0.89; p<0.05), while Vorticella correlated negatively with the log removal of E. coli (-0.78; p=<0.05). In the reed beds, the correlations were generally weaker compared with the unvegetated ponds, with no significant correlations with the removal of bacterial or protozoan faecal indicators. Significant negative correlations were observed between the predation capacity of Ciliophora sp. (-0.84; p=<0.05) and Chydorus (0.96; p<0.05) with the removal of B. fragilis phages. Significant positive correlations were observed in the reed beds between Vorticella and Ceriodaphnia correlations and the removal of Bacteroides HB13 phages (-0.82; p=<0.05).

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

-0.1

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

Jun Jul Aug Sep Oct Nov Dec

Am

oebo

zoa

<30µ

m

E.c

olir

emov

al (L

og10

CFU

)

E.coli Amoebozoa <30µm (x104)

0

10

20

30

40

50

60

70

80

-0.1

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9


Pred

atio

n ca

paci

ty (

Bac

teri

a h-1

L-1

)

Faec

al in

dica

tor

rem

oval

(Log

10)

E.coli (CFU) Ciliophora <30µm (x106)

0

10

20

30

40

50

60

70

80

-0.1

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9


Pred

atio

n ca

paci

ty (

Bac

teri

a h-1

L-1

) x10

6

Faec

al in

dica

tor

rem

oval

(Log

10)

E.coli (CFU) Ciliophora <30µm (x106)E. Coli (CFU)

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

-0.1

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9


Am

oebo

zoa

<30µ

m

Faec

al in

dica

tor

rem

oval

(Log

10)

E.coli (CFU) Amoebozoa <30µm (x104)Amoebozoa <30µm ( 104)

0

10

20

30

40

50

60

70

80

-0.1

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

Jun Jul Aug Sep Oct Nov DecPr

edat

ion

capa

city

(B

acte

ria

h-1L

-1) x

106

Faec

al in

dica

tor

rem

oval

(Log

10)

E.coli (CFU) Ciliophora <30µm (x106)Ciliophora <30µm ( 106)

Fig. 4.5 Monthly E. coli (log10 CFU) removal and predation capacity of Ciliophora and Amoebozoa smaller than 30 µm.

72

Chapter 4

Table 4.2 Bacteriovore zooplankton abundance (s.e.) in the unvegetated ponds and reed beds, and the calculated predation capacity.

Organism Conversion factor Unvegetated ponds

Predation capacity Abundance Estimate predation capacity

(Bacteria ind.-1 h-1 L-1) Reference (ind. L-1) (Bacteria h-1 L-1) % of total

Amoebozoa <30 µm 2 [1] 187942 (78849) 375885 (157697) 4 (1)

Arcellinida Arcella 200 [1] † 139 (67) 27751 (13313) 0 (0)

Arcellinida Difflugia 284 [1] † 3 (2) 905 (658) 0 (0)

Arcellinida Euglypha 164 [1] † 13 (9) 2187 (1403) 0 (0)

Gymnamoebae 405 [2] 4 (2) 1564 (977) 0 (0)

Ciliophora <30 µm 388 [3,4] 32039 (12513) 12430938 (4854850) 89 (4)

Ciliophora sp. 950 [4,5,6,7,8,9,10,11] 255 (101) 242203 (95977) 3 (1)

Vorticella 386 [10] 119 (87) 45765 (33617) 1 (0)

Cladocera Ceriodaphnia 1500 ‡ <1 (0) 658 (658) 0 (0)

Chydorus 1561 [12] 0 (0) 0 (0) 0 (0)

Daphnia sp. 4440 [12] 6 (5) 24555 (20022) 3 (3)

Daphnia ambigua 4550 [9] 0 (0) 0 (0) 0 (0)

Copepoda Harpacticoida 42 [8] 0 (0) 0 (0) 0 (0)

Nauplius-larvae 33 [8] 2 (1) 50 (19) 0 (0)

Nematoda 21 [8] 7 (2) 139 (50) 0 (0)

Rotifera Brachionus 880 [12] 0 (0) 0 (0) 0 (0)

Conochilus 11600 [12] 1 (1) 6052 (6052) 0 (0)

Keratella quadrata 280 [13] <1 (0) 85 (85) 0 (0)

Polyarthra 125 [13] <1 (0) 16 (16) 0 (0)

Pompholyx 525 [14] 0 (0) 0 (0) 0 (0)

Rotifera sp. 200 ‡ 27 (16) 5371 (3210) 0 (0)

Synchaeta 260 [13] 1 (1) 188 (168) 0 (0)

Total 220645 (81668) 13164311 (4863663)

[1] Rogerson et al. 1996, [2] Pickup et al. 2007, [3] Kuppardt et al. 2010, [4] Ayo et al. 2001, [5] Sherr et al. 1987, [6] Decamp et al. 2000, [7] Decamp & Warren 1998, [8] Epstein & Shiaris 1992, [9] King et al. 1991, [10] Sanders et al. 1989, [11] Albright et al. 1987, [12] Oomswilms et al. 1995, [13] Ronneberger 1998, [14] Agasild & Noges 2005. † estimated on Rogerson et al. 1996, ‡ estimated on size/predation capacity of organisms in same phylum.


73

Table 4.2 Continuous

Organism Reed beds

Abundance Estimated predation capacity

(ind. L-1) (Bacteria h-1) % of total

Amoebozoa <30 µm 160450 (54941) 320900 (109882) 4 (3)

Arcellinida Arcella 107 (52) 21328 (10389) 0 (0)

Arcellinida Difflugia 21 (10) 6077 (2744) 0 (0)

Arcellinida Euglypha 3 (2) 500 (357) 0 (0)

Gymnamoebae 7 (4) 2719 (1472) 0 (0)

Ciliophora <30 µm 49507 (21325) 19208619 (8273956) 73 (13)

Ciliophora sp. 910 (420) 864225 (399434) 16 (10)

Vorticella 661 (621) 255162 (239594) 3 (3)

Cladocera Ceriodaphnia 42 (41) 62504 (61786) 1 (1)

Chydorus 10 (3) 15951 (5338) 0 (0)

Daphnia sp. 20 (12) 87363 (52708) 2 (1)

Daphnia ambigua 10 (10) 45500 (45500) 1 (1)

Copepoda Harpacticoida <1 (0) 11 (8) 0 (0)

Nauplius-larvae 216 (156) 7117 (5164) 0 (0)

Nematoda 28 (13) 581 (277) 0 (0)

Rotifera Brachionus 6 (6) 5500 (5500) 0 (0)

Conochilus 0 (0) 0 (0) 0 (0)

Keratella quadrata 0 (0) 0 (0) 0 (0)

Polyarthra 1 (1) 179 (179) 0 (0)

Pompholyx <1 (0) 211 (211) 0 (0)

Rotifera sp. 68 (27) 13571 (5370) 0 (0)

Synchaeta 1 (1) 325 (325) 0 (0)

Total 212367 (60890) 20918343 (8023309)

74

Chapter 4

Laboratory incubations Average E. coli numbers at the start of the lab incubations were 213±5 and 233±7 CFU mL-1 in respectively the unfiltered and filtered (5 µm) samples (Fig 4.6). The different treatments caused no significant difference between treatments in the pH (8.69 ±0.05) and dissolved oxygen (7.76 ±0.11 mg L-1) until the end of the incubation period (data not shown). Incubation of the samples under field conditions (Fig. 4.6a) resulted in a constant decrease of E. coli numbers over time resulting in 72 ±6 CFU mL-1 after 69 h incubation. Removal of predation pressure by filtration of the samples (5 µm) before incubation (Fig 4.6c) resulted in an increase of E. coli numbers. Exposure to UV radiation of filtered samples (Fig 4.6b) caused a decrease in E. coli numbers to 35 ±19 CFU mL-1 after 69 h incubation. Incubation of filtered samples at a higher temperature (20°C; Fig 4.6d) resulted in an initial increase of E. coli to 311 ±7 CFU mL-1, but was followed by a steady decrease in the following days to 144 ±27 CFU mL-1 after 69 h of incubation.

0

50

100

150

200

250

300

350

400

450

0 20 40 60 80

E.c

oli(

CFU

mL

-1)

Time (h-1)

10 C Zooplankton [+] UVR [-]

0

50

100

150

200

250

300

350

400

450

0 20 40 60 80

E.c

oli(

CFU

mL

-1)

Time (h-1)

20 C Zooplankton [-] UVR [-]

0

50

100

150

200

250

300

350

400

450

0 20 40 60 80

E.c

oli(

CFU

mL

-1)

Time (h-1)

10 C Zooplankton [-] UVR [+]

0

50

100

150

200

250

300

350

400

450

0 20 40 60 80

E.c

oli(

CFU

mL

-1)

Time (h-1)

10 C Zooplankton [-] UVR [-]a)

b)

c)

d)

Fig. 4.6 E. coli numbers (CFU; ±s.e.; n=3) in incubated WWTP effluent (Grou) over time. Zooplankton [+]: unfiltered water; Zooplankton [-]: 5µm filter water; UVR [-]: only white light; UVR [+]: white light and UV exposure; temperature: 10 or 20°C.


75

Discussion

Removal of faecal indicator organisms The removal efficiencies for E. coli and enterococci observed in this study were relatively low compared to other studies (Molleda et al. 2008; Reinoso et al. 2008; van den Boomen and Kampf 2012). Bacterial pathogens were more efficiently removed than protozoa and viral pathogens, which seem to be related to the differences in inflow concentrations. However, the low numbers of faecal bacterial, protozoan and viral indicators in the effluent of the constructed wetland still demonstrated that the constructed wetland is a suitable tool to reduce the potential threat of treated wastewater for water recreation and drinking water production. For all faecal indicator organisms the average removal tended to be higher in the reed beds than in the unvegetated ponds, which can only partially be explained by the larger hydraulic retention time in the reed beds. In the next paragraphs we try to explain the difference in removal efficiency of the indicator organisms between the unvegetated ponds and reed beds by tentatively comparing the relative importance of sedimentation, predation, UV exposure, natural mortality and biofilm trapping for the removal efficiency.

Sedimentation Observations with the sedimentation traps showed that sedimentation of E. coli in the unvegetated ponds is an unimportant process in the removal of E. coli, despite 10% sedimentation of the inflowing suspended matter (van den Boomen et al. 2012). The low sedimentation rate of E. coli is in concurrence with a study of Boutilier et al. (2009), who concluded that sedimentation contributed less than 3% of E. coli removal in wastewater treatment wetlands. The importance of sedimentation is mainly dependent on the association of microorganisms to suspended matter as “free living” microorganisms exhibit very low sedimentation rates (Brookes et al. 2004; Medema et al. 1998; Boutilier et al. 2009; Jin et al. 2004). In our study, the laboratory experiments showed that the filtration of WWTP effluent over a 5µm filter did not influence the E. coli concentration, indicating that the majority of E. coli cells was indeed either free living or associated with particles smaller than 5µm. The constructed wetland in this study receives very low concentrations (3.6 mg L-1) of suspended matter (Ghermandi et al. 2007). Low concentrations of inflowing suspended matter probably prevent substantial aggregation of faecal organisms like E. coli with suspended matter, thereby reducing the importance of sedimentation for the removal of faecal microorganisms in this constructed wetland. Due to the dimensions of the reed beds no sedimentation traps could be placed and a no good estimation of sedimentation rates E. coli in the reed beds could be made. The inflow of suspended matter is however low and a substantial increase is sedimentation of faecal indicator organisms is unlikely.

76

Chapter 4

Predation Predation on the microbial consortia by zooplankton was dominated by Ciliophora (and to lesser extent Amoebozoa) in both the unvegetated ponds and the reed beds. The importance of Ciliophora to the total predation rate on particles, including microorganisms, has been shown by several studies (Albright et al. 1987; Sanders et al. 1989; Vaque et al. 1994). Ciliophora comprise a large and diverse group of organisms which differ in size, behaviour and occurrence (Brusca and Brusca 2003). Ciliophora prefer to ingest freely suspended rather than attached bacteria (Albright et al. 1987). They can be either non-selective or selective grazers, some preferring rod-shaped coliforms (incl. E. coli) over coccus-shaped enterococci (Epstein and Shiaris 1992). Several other studies have indicated that bacterial grazing by zooplankton is depending on the size of both bacteria and grazers (Porter et al. 1983; Knoechel and Holtby 1986; Hart and Jarvis 1993). The correlations between the removal of an individual faecal indicator and specific zooplankton species as shown in this research seem to support selective predation of zooplankton species on faecal microorganisms. The high correlations between the removal of faecal indicators and the abundance of the most numerous zooplankton species suggest that predation by zooplankton is an important contributor to the removal of the faecal indicator organisms in the unvegetated ponds. These findings in the field were supported by laboratory experiments which demonstrated that E. coli removal is strongly reduced when the entire zooplankton community is removed. Removal of faecal indicator organisms is used to predict the behavior of faecal pathogens like Campylobacter, Norovirus, Enterovirus and Cryptosporidium. Because zooplankton species might selectively predate on certain microbial species, it remains uncertain whether faecal pathogens are removed by the zooplankton community to the same extent as the faecal indicators measured in this study. Additional laboratory experiments are required to demonstrate the effect of predation on other faecal pathogen species. Although the abundance of zooplankton was similar in the unvegetated ponds and the reed beds throughout the year, substantial less correlations were found in the reed beds between faecal indicator removal and zooplankton abundance. This does not directly mean that the predation on faecal organisms by zooplankton is not occurring in the reed beds, but the contribution of predation on the total removal of faecal microorganisms in the reed beds is probably lower, as lower numbers of inflowing faecal indicator organisms make them a small source of food and other processes including biofilm retainment may be more important. In general, the predation capacity of the zooplankton community in both the unvegetated ponds and reed beds exceeded the total number of faecal indicator organisms (excluding the gene copies of Bacteroidales) with two orders of magnitude, indicating that the predation capacity of the zooplankton community is a driving force for the removal of faecal indicator organisms and pathogens in constructed wetlands.


77

UV exposure The influence of UV radiation on the removal of the faecal indicator organisms in treated wastewater during residence in constructed wetlands was shown by the laboratory experiments to be potentially strong and has been proven to effectively decrease the concentration of bacterial and viral organisms in surface waters, wastewater and other media (Garvey et al. 1998; Sommer et al. 2000; Craik et al. 2001; Brookes et al. 2004; Tchobanoglous et al. 2004; Whitman et al. 2004; Oteiza et al. 2005). Sinton et al. (2002) found that the inactivation of bacteria was ten times higher in sunlight than in darkness and concluded that this effect is mainly caused by UV-B radiation. However, we do not expect this to be an important factor in the constructed wetland itself: based on the calculated light attenuation coefficient estimated from the average DOC content of the water (Morris et al. 1995) at PONDS-IN, the penetration of UV radiation ranges from 15 to 35 cm for UV-A (380 nm) and from 3 to 6 cm for UV-C (250 nm). Because of this low penetration of especially UV radiation of low wavelength (UV-C), which is most harmful for organisms, only a minor contribution to the disinfection of faecal microorganisms is expected for UV radiation in the unvegetated ponds (average depth 2 m). The reed beds are covered by a thick layer of vegetation most of the year which will lower UV irradiance strongly (Boutilier et al. 2009) and reduce the disinfection contribution of UV radiation in the reed beds, even with a shallow average water depth of 30 cm.

Inactivation of faecal indicators The inactivation of microorganisms is another factor contributing to the reduction in numbers of faecal indicator organisms in constructed wetlands. Several studies (Boutilier et al. 2009; Darakas 2002) show that E. coli can survive twice as long at temperatures around 10°C compared with 20°C and that 10°C is actually the optimal survival temperature for E. coli (Darakas 2002). Similar results were observed in our study where incubation at 20°C caused a substantial reduction in E. coli numbers, whereas incubation at 10°C did not result in a decrease of E. coli. The contribution of inactivation to the removal of faecal indicator organisms in constructed wetlands is however dependent on individual indicator organisms (E. coli, enterococci, Bacteroides-phages, etc), temperature, organic matter, redox conditions, etc and is therefore difficult to estimate. The temperature dependence on the natural mortality may contribute to higher removal efficiencies in summer compared with winter. Khatiwada and Polprasert (1999) estimated a contribution to coliform removal by temperature modulated death in pond systems be relatively minor (6.5% at 20°C).

External input The number of Bacteroides phages indicated that there was no significant external input from June till December by warm blooded animals like birds. Several studies (Kadlec et al. 2010; Knowlton et al. 2002) have observed the attraction of large numbers of waterfowl to constructed wetlands during winter, especially in freezing periods. In this study generally

78

Chapter 4

around five water birds were observed in the unvegetated ponds of the constructed wetland, but in January and February around 100-200 birds were counted in the ponds. The faeces of waterfowl can contain high numbers of faecal indicator organisms like E. coli and enterococci but also pathogens like Campylobacter (Moriart et al. 2011; Moriart et al. 2012; Meerburg et al. 2011; Benskin et al. 2009). On a daily base each individual bird can add 106-1010 CFU of both E. coli and Enterococci to the water (Moriart et al. 2011). Based on these numbers and assuming total mixing of the faeces, the waterfowl population approximately in the unvegetated ponds in wintertime could increase the abundance of E. coli and enterococci by 80-170 CFU L-1. This addition of 101 to 102 CFU L-1 is very minor in comparison with the average removal of 104 CFU L-1 and external input by waterfowl is negligible in the faecal indicator dynamics.

Trapping in biofilm The trapping of bacteria and viruses in biofilms is another factor that could contribute to the removal of faecal microorganisms in reed beds, but was not addressed in this study. Biofilms are known to retain large quantities of small sized particles in the size range of viruses (0.1 µm), bacteria (1 µm) and parasitic protozoa (4.5 µm) (Stott and Tanner 2005; Balzer et al. 2010; Drury et al. 1993a; Drury et al. 1993b; Eisenmann et al. 2001). The reed beds in our study were more efficient in removing faecal indicator organisms from the water, which could only partially be explained by a difference in HRT. The submerged parts of the reed stems in the reeds beds in the CW form a large surface area for biofilm development and it is therefore expected that particle trapping by these biofilms is an important factor causing the difference in removal efficiencies between unvegetated ponds and reed beds.

Conclusion This study showed that constructed wetlands significantly reduce the number of bacterial, protozoan and viral faecal indicators. As these organisms are used as an indication for removal of faecal pathogens it can be concluded that treatment of WWTP effluent by a constructed wetland results in lower numbers of faecal pathogens, effectively reducing the discharge of pathogens into receiving surface waters. At the inflow of the constructed wetland E. coli cells were observed to be mostly planktonic (not associated with suspended matter) together with low concentrations of suspended matter for aggregation. Consequently, the contribution of sedimentation to the removal of these faecal indicators was low for E. coli and probably also for other faecal indicator organisms and pathogens. Planktonic bacterial cells and virus particles are however favorable for zooplankton predation which was observed to be an important process regulating the removal of faecal indicators in the unvegetated ponds. However, indications of selective grazing by individual zooplankton groups and species make the contribution of predation to pathogen removal variable and species specific and may therefore differ between faecal


79

indicator organisms and other pathogens. UV light was shown to have a strong potential impact on the removing of faecal indicator organisms in a laboratory experiment, but the dimensions and characteristics of the CW combined with high DOC concentrations in the treated wastewater, hamper penetration of UV light into the water column and thereby minimize the effects of UV light of the removal of faecal indicators. Although inactivation of faecal indicator organisms and pathogens in the CW was difficult to quantify, literature suggests a low contribution of inactivation to the total removal of faecal indicator organisms. Significant reintroduction of faecal indicator organisms by defecation by waterfowl and other warm blooded animals was not observed, but could be periodically significant in systems with low flow rates and a high surface area. In the reed bed systems, removal rates of faecal indicator organism were higher than in the unvegetated ponds. Here it is argued that particle trapping by biofilms growing on the reed stems is an important factor causing the difference in removal efficiencies between the two functional compartments in the CW. Based on these observations it can be concluded that the biological components of CW support the driving forces determining the capacity to remove “free living” faecal indicator organisms from treated wastewater.

Acknowledgments Sampling and analyses were conducted by “Het Waterlaboratorium” (HWL), Haarlem, The Netherlands. Special thanks go out to Daan Mes, Anne de Valença and Hans Breeuwer for the support to this study.

80

Chapter 4

References

Abrantes N and Goncalves F (2003). The dynamics of Ceriodaphnia pulchella (Cladocera) in laboratory. Acta Oecologica-International Journal of Ecology, 24 (1), 245-249.

Arnone RD and Walling JP (2007). Waterborne pathogens in urban watersheds. Journal of Water and Health 5 (1), 149-162.

Albright LJ, Sherr EB, Sherr BF and Fallon RD (1987). Grazing of ciliated protozoa on free and particle-attached bacteria. Marine Ecology-Progress Series 38 (2), 125-129.

Balzer M, Witt N, Flemming H-C and Wingender J (2010). Faecal indicator bacteria in river biofilms. Water Science and Technology 61 (5), 1105-1111.

Benskin CMH, Wilson K, Jones K and Hartley IR (2009). Bacterial pathogens in wild birds: a review of the frequency and effects of infection. Biological Reviews 84 (3), 349-373.

Brookes JD, Antenucci J, Hipsey M, Burch MD, Ashbolt NJ and Ferguson C (2004). Fate and transport of pathogens in lakes and reservoirs. Environment International 30 (5), 741-759.

Brusca RC and Brusca GJ (2003). Invertebrates; second edition. Sinauer Associates, Inc., Sunderland.

Boutilier L, Jamieson R, Gordon R, Lake C and Hart W (2009). Adsorption, sedimentation, and inactivation of E. coli within wastewater treatment wetlands. Water Research 43 (17), 4370-4380.


Craik SA, Weldon D, Finch GR, Bolton JR and Belosevic M (2001). Inactivation of Cryptosporidium parvum oocysts using medium- and low-pressure ultraviolet radiation. Water Research 35 (6), 1387-1398.

Darakas E (2002). E-coli kinetics - Effect of temperature on the maintenance and respectively the decay phase. Environmental Monitoring and Assessment 78 (2), 101-110.

Diaz FJ, O’Geen AT and Dahlgren RA (2010). Efficacy of constructed wetlands for removal of bacterial contamination from agricultural return flows. Agricultural Water Management 97 (11), 1813-1821.

Drury WJ, Characklis WG and Stewart PS (1993). Interactions of 1 mu-m latex-particles with Psuedomonas-aeruginosa biofilms. Water Research 27 (7), 1119-1126.


81

Drury WJ, Stewart PS and Characklis WG (1993). Transport of 1-mu-m latex-particles Psuedomonas-aeruginosa biofilms. Biotechnology and Bioengineering 42 (1), 111-117.

Ebdon JE, Sellwood J, Shore J and Taylor HD (2012). Phages of Bacteroides (GB-124): A Novel Tool for Viral Waterborne Disease Control? Environmental science & technology 46 (2), 1163-1169.

Epstein SS and Shiaris MP (1992). Rates of microbenthic and meiobenthic bacterivory in a temperate muddy tidal flat community. Applied and Environmental Microbiology 58 (8), 2426-2431.


Garvey E Tobiason JE, Hayes M, Wolfram E, Reckhow DA, Male JW (1998). Coliform transport in a pristine reservoir: Modeling and field studies. Water Science and Technology 37 (2), 137-144.

Gerba CP (2000). Indicator Microorganisms. In: Maier RM, Pepper IL, Gerba CP (Eds.), Environmental Microbiology. Academic Press, San Diego, 491-503.


Gomez M, de la Rua A, Garralon G, Plaza F, Hontoria E and Gomez M (2006). Urban wastewater disinfection by filtration technologies. Desalination 190 (1-3), 16-28.


Hart RC and Jarvis AC (1993). Insitu determinations of bacterial selectivity and filtration-rates by 5 Cladoceran zooplankters in a hypertrophic subtropical reservoir. Journal of Plankton Research 15 (3), 295-315.


ISO/TC-147/SC4/WG5 (1995). Microbiological methods: Sulfite reducing clostridium.

ISO-10705-4 (2001). Water quality -- Detection and enumeration of bacteriophages -- Part 4: Enumeration of bacteriophages infecting Bacteroides fragilis.

82

Chapter 4

ISO-6461-2 (1986). Water quality -- Detection and enumeration of the spores of sulfite-reducing anaerobes (clostridia) -- Part 2: Method by membrane filtration.

Jin G, Englande J, Bradford H and Jeng H, (2004). Comparison of E-coli, Enterococci, and fecal coliform as indicators for brackish water quality assessment. Water Environment Research 76 (3), 245-255.

Karim M.R., Manshadi F.D., Karpiscak M.M., Gerba C.P., 2004. The persistance and removal of enteric pathogens in constructed wetlands. Water Research, 38, 1831-1837.

Kadlec RH, Cuvellier C and Stober T (2010). Performance of the Columbia, Missouri, treatment wetland. Ecological Engineering 36 (5), 672-684.

Kadlec RH and Wallace SD (2008). Treatment wetlands. CRC Press, Florida, USA, pp1016.

Khatiwada NR and Polprasert C (1999). Kinetics of fecal coliform removal in constructed wetlands. (3) 109-116.

Knight J and Waller W (1992). Influence of the Addition of Cerophyl(r) on the Selenastrum-Capricornutum Diet of the Cladoceran Ceriodaphnia-Dubia. Environmental Toxicology and Chemistry, 11 (4), 521-534.

Knoechel R. and Holtby LB (1986). Cladoceran filtering rate – body length relationships for bacterial and large algal particles. Limnology and Oceanography 31 (1), 195-200.

Knowlton M, Cuvellier C and Jones J (2002). Initial performance of a high capacity surface-flow treatment wetland. Wetlands 22 (3), 522-527.

Medema G, Schets F, Teunis P and Havelaar A (1998). Sedimentation of free and attached Cryptosporidium oocysts and Giardia cysts in water. Applied and Environmental Microbiology 64 (11), 4460-4466.

Meerburg BG, Koene MGJ and Kleijn D (2011). Escherichia coli Concentrations in Feces of Geese, Coots, and Gulls Residing on Recreational Water in The Netherlands. Vector-Borne and Zoonotic Diseases 11 (6), 601-603.

Molleda P, Blanco I, Ansola G and de Luis E (2008). Removal of wastewater pathogen indicators in a constructed wetland in Leon, Spain. Ecological Engineering 33 (3-4), 252-257.

Moriarty EM, Weaver L, Sinton LW and Gilpin B (2012). Survival of Escherichia coli, Enterococci and Campylobacter jejuni in Canada Goose Faeces on Pasture. Zoonoses and Public Health 59 (7), 490-497.


83

Moriarty EM, Karki N, Mackenzie M, Sinton LW, Wood DR and Gilpin BJ (2011). Faecal indicators and pathogens in selected New Zealand waterfowl. New Zealand Journal of Marine and Freshwater Research 45 (4), 679-688.

Morris DP, Zagarese H, Williamson CE, Balseiro EG, Hargreaves BR, Modenutti B, Moeller R and Queimalinos C (1995). The attentuation of solar UV radiation in lakes and the role of dissolved organic carbon. Limnology and Oceanography 40 (8), 1381-1391.

NEN-EN-ISO-7899-2 (2000). Water quality - Detection and enumeration of intestinal Enterococci - Part 2: Membrane filtration method.

NEN-EN-ISO-8199 (2007). Water quality - General guidance on the enumeration of micro-organisms by culture.


Oteiza JM, Peltzer M, Gannuzzi L and Zaritzky N (2005). Antimicrobial efficacy of UV radiation on Escherichia coli O157 : H7 (EDL 933) in fruit juices of different absorptivities. Journal of Food Protection 68 (1), 49-58.

Payan A, Ebdon J, Taylor H, Gantzer C, Ottoson J, Papageorgiou GT, Blanch AR, Lucena F, Jofre J and Muniesa M (2005). Method for isolation of Bacteroides bacteriophage host strains suitable for tracking sources of fecal pollution in water. Applied and Environmental Microbiology 71 (9), 5659-5662.

Payment P and Franco E (1993). Clostridium-Perfringens and Somatic Coliphages as Indicators of the Efficiency of Drinking-Water Treatment for Viruses and Protozoan Cysts. Applied and Environmental Microbiology 59 (8), 2418-2424.

Porter KG, Feig YS and Vetter EF (1983). Morphology, flow regimes, and filtering rates of Daphnia, Ceriodaphniam and Bosmina fed natural bacteria. Oecologia 58 (2), 156-163.


Sanders RW, Porter KG, Bennett SJ and Debiase AE (1989). Seasonal patterns of bacterivory by Ciliates, Rotifers, and Cladocerans in a fresh-water planktonic. Limnology and Oceanography 34 (4), 673-687.

Seviour RJ and Nielsen PH (2010). Microbial communities in activated sludge plants. In: Seviour RJ and Nielsen PH (Eds.), Microbial ecology of activated sludge. IWA publishing, London, 95-125.

Sinton LW, Hall CH, Lynch PA and Davies-Colley RJ (2002). Sunlight inactivation of fecal indicator bacteria and

84

Chapter 4

bacteriophages from waste stabilization pond effluent in fresh and saline waters. Applied and Environmental Microbiology 68 (3), 1122-1131.

Sommer R, Lhotsky M, Haider T and Cabaj A (2000). UV inactivation, liquid-holding recovery, and photoreactivation of Escherichia coli O157 and other pathogenic Escherichia coli strains in water. Journal of Food Protection 63 (8), 1015-1020.



van den Boomen R and Kampf R (2012). Waterharmonica’s in the Netherlands 1996-2011: from WWTP effluent till usable surface water. ISBN.978.90.5773.559.2 STOWA 2012-12

van den Boomen R, Kampf R and Mulling BTM A (2012). Research on suspended particles and pathogens in the Waterharmonica (constructed wetland). ISBN.978.90.5773.553.0 STOWA 2012-10

van der Wielen PWJJ and Medema GJ (2010). Unsuitability of quantitative Bacteroidales 16S rRNA gene assays for discerning fecal contamination of drinking water. Applied and Environmental Microbiology 76 (14), 4876-4881.

Vaque D, Gasol JM and Marrase C (1994). Grazing rates on bacteria – the significance of methodology and ecological factors. Marine Ecology-Progress Series 109 (2-3), 263-274.

Vymazal J (2005). Removal of enteric bacteria in constructed treatment wetlands with emergent macrophytes: A review. Journal of Environmental Science and Health Part a-Toxic/Hazardous Substances & Environmental Engineering 40 (6-7), 1355-1367.

Watson K, Shaw G, Leusch FDL and Knight NL (2012). Chlorine disinfection by-products in wastewater effluent: Bioassay-based assessment of toxicological impact. Water research 46 (18), 6069-6083.

Wen QX, Tutuka C, Keegan A and Jin B (2009). Fate of pathogenic microorganisms and indicators in secondary activated sludge wastewater treatment plants. Journal of Environmental Management 90 (3), 1442-1447.

Wery N, Lhoutellier C, Ducray F, Delgenes JP and Godon JJ (2008). Behaviour of pathogenic and indicator bacteria during urban wastewater treatment and sludge composting, as revealed by quantitative PCR. Water Research 42 (1-2), 53-62.


85

Whitman RL, Nevers MB, Korinek GC and Byappanahalli MN (2004). Solar and temporal effects on Escherichia coli concentration at a lake Michigan swimming beach. Applied and Environmental Microbiology 70 (7), 4276-4285.

Trapping of bacterial cells and latex micro-spheres in natural and cultured phototrophic biofilms

Chapter 5

Manuscript: Mulling BTM, Admiraal W, van Beusekom SAM, Bichebois S, Schwartz T, van der Geest HG (2013). Trapping of bacterial cells and latex micro-spheres in natural and cultured phototrophic biofilms.

88

Chapter 5

AbstractPhototrophic biofilms play a vital role in aquatic ecosystems by trapping and processing

abiotic and biotic particles. To study the effects of both particle and biofilm characteristics on particle trapping, incubation experiments were conducted using both natural and cultured mono-culture phototrophic biofilms and several different traceable particles. In a first experiment we observed substantial trapping of particles smaller than 15 µm, with highest trapping efficiencies for particles in the size range 5.0 - 8.0 µm (49% in 150 minutes). In a second experiment field grown natural biofilms and Achnantes lanceolata (diatom) biofilms rapidly accumulated micro-spheres and P. putida cells, while Nitzschia perminuta (another diatom species) and the cyanobacterial biofilms trapped only few particles. Trapping of P. putida was up to one order of magnitude higher than the trapping of micro-spheres, which indicates selective trapping of particles. These results showed that the trapping of particles by phototrophic biofilms in aquatic ecosystems can be substantial, but is a highly variable process strongly dependent on both species composition of the biofilms and the size and nature of particles.

Biofilm trapping

89

IntroductionPhototrophic biofilms cover almost all illuminated surfaces in aquatic systems and consist

of consortia of phototrophic and associated heterotrophic micro-organisms mechanically stabilized to the substratum by a mucous matrix of extracellular polymeric substances (EPS) (Stewart and Franklin 2008). The microbial community in biofilms play an important role in biochemical cycles of aquatic ecosystems (Battin et al. 2007; Woodruff et al. 1999; Costerton et al. 1995; Sutherland 2001; Hall-Stoodley et al. 2004) and the abundance of phototrophic and heterotrophic micro-organisms make biofilms hotspots of biological activity (Romani et al. 2004; Vadeboncoeur et al. 2008; Heck and Hesslein 1995). One of the mechanisms through which biofilms influence biochemical cycles is the adhesion and trapping of particles from the water column (Bouwer 1987), which can then be retained for substantial periods of time, serve as food for the heterotrophic community or become part of the biofilm microbial community. The capacity of biofilms to trap particles is considered sufficiently high to determine water clarity in aquatic ecosystems, even in running waters (Kadlec and Wallace 2008; Jowett and Biggs 1997; Gantzer et al. 1988). This purifying role of biofilms has long been recognized and is utilized in rotating biological contactors (Alleman et al. 1982), sequencing batch reactors (Wilderer and McSwain 2004), other technical reactors and constructed wetlands for the removal of organic matter and pathogens from polluted waters (Kadlec and Wallace 2008). Several studies have investigated the trapping of particles by phototrophic biofilm and have shown that the trapping of particles by phototrophic biofilms is affected by particle properties and biofilm structure (Okabe and Kuroda et al. 1998; Stott and Tanner 2005; Eisenmann et al. 2001; Chabaud et al. 2006), but effects of species composition of phototrophic biofilms on the trapping of particles with different characteristics is not well understood.

Therefore the aims of this study were to 1) compare phototrophic biofilm trapping efficiency of different particle size classes, 2) compare the trapping capacity of biofilms dominated by different species of diatoms and cyanobacteria and 3) compare the trapping capacity of biofilm between model particles (latex micro-spheres) and bacteria (with Pseudomonas putida as a model organism). For this purpose laboratory incubation experiments were conducted using both natural mixed species biofilms and mono-culture biofilms of diatoms and cyanobacteria. Microspheres and Pseudomonas putida cells were added as suspensions and particle concentrations and Pseudomonas putida abundance were monitored over time in the water column and in the biofilm.

Materials and methods

Experiment I: Particle size; Phototrophic biofilmsNatural biofilms were collected from an experimental rainwater pond at the University of

Amsterdam, Science Park, The Netherlands (N52 21.286 E4 57.471). Following Barranguet et al. (2004) 1.5 cm2 glass discs were mounted in vertical racks suspended in the water for six weeks and mature biofilms (dominated by green algae) were collected and incubated on clean support racks in cone shaped vessels (1.7 L) in WC Medium (Guillard and Lorenzen 1972). The biofilms were kept at

90

Chapter 5

18°C and left to acclimate for 2 days before the start of the experiments. Water in the cones was mixed by pumping water from the bottom of the cones back to top, creating a vertical water flow to minimize particle loss by sedimentation.

Experiment I: Particle size; Biofilm, particle incubationsAfter 2 days of acclimatisation, medium was drained from the cones to remove any

accumulated particles and replaced with particle-free WC medium. Three cones containing glass slides with attached biofilm were used as experimental vessels. To quantify the particles present in the biofilm or released into the water by the biofilm a control cone contained biofilms, but no added particles. To quantify the loss of particles over time not induced by the presence of biofilm another control cone was used which contained no biofilm, but with added particles. The experiment started one hour after refreshment of the medium, by addition of a mixture of particles with difference sizes. The mixture consisted of micro-spheres (Polysciences Inc., Polybead® Polystyrene, Germany) of five different nominal sizes (2 μm; 4,5 μm; 6 μm; 10 μm and 25 μm). The slight negative surface charge of the micro-spheres correspond to the surface charge of many microbial organisms found in aquatic environments with a pH above 7 (Stevik et al. 2004; Ongerth and Pecoraro 1996; Gerba 1984). Water samples were collected on several time points during incubation to quantify the concentration of particles in the water in the size ranges of 1.5 - 3.0 µm, 3.0 - 5.0 µm, 5.0 - 8.0 µm, 8.0 – 15 µm and 15 - 35 µm. To this purpose, 2 mL water samples (three technical replicates) were taken from each cone and stored in 10 mL tubes for further analyses. Water samples were then digested to remove organic matter by adding 1 mL 10% HCl, 1 mL 10% HNO3 and 4 drops of H2O2 for 24 h. Next, samples were diluted with MilliQ water to a total volume of 30 mL and particles in these samples were counted using a particle counter (PAMAS® WaterViewer; Sensor HCB-LD-50/50, Germany). The experiment lasted for 150 min and at that time the glass slides with attached biofilm (six per cone) were taken from each aquarium and transferred into tubes (three replicates per cone) and sequentially digested and counted as described above.

Data was not normally distributed for all samples and differences between treatments were therefore analysed with Kruskal-Wallis tests using PAST (Hammeret et al. 2001). Particle removal rates (linear regression) and first order removal rates (kd; exponential decrease) were calculated with MS Excel. The kd values determined using:

(1)

Where N0 is the particle concentration after addition of the particles and Nt the particle concentration at time point t (h).

Experiment II: Biofilm composition and particle type; Phototrophic biofilmsTwo types of phototrophic biofilms were used. First, natural biofilms were collected from

lakes in the infiltration dunes of Leiduinen, The Netherlands (N52 20.726 E4 31.653) as described in

𝑘𝑘𝑑𝑑 = −ln �𝑁𝑁𝑡𝑡𝑁𝑁0

�

𝑡𝑡 1

Biofilm trapping

91

Barranguet et al. (2004). Glass discs were pre-colonized in the field for 7 days at ca. 10°C at two sites (further referred to as Natural biofilm I and II). Although both sites received the same water from the dune aquifer, these biofilms, both diatom dominated, developed slightly different (Barranguet et al. 2005). Secondly, mono-culture biofilms of diatoms (Achnantes lanceolata and Nitzschia perminuta) and cyanobacteria (Cylindrospermum stagnale and Leptolyngbya foveolarum), isolated from floodplains from the River Rhine (van der Grinten et al. 2004), were grown by incubation of glass discs in mono-cultures of these different species in 2 L aquaria with WC medium (Guillard and Lorenzen 1972) for 10 days at 20°C. After colonization, the biofilms were transferred to 8 L aquaria with WC medium that was mixed by aeration, and left for three days to acclimatize at 20°C.

Experiment II: Biofilm composition and particle type; Biofilm, particle incubationsAfter acclimatisation of the biofilm, the experiment started by addition of particles to the 8

L aquaria. Two types of particles were used: 1) fluorescent micro-spheres with a nominal diameter of 1.0 μm (with a slightly negative surface charge; Polysciences Inc., Fluoresbrite® BB Carboxylate Microspheres) and 2) living bacteria (P. putida, which have rod shaped cells of approximately 1µm wide and 3µm long). The stock solution of micro-spheres contained 5.4×109 particles mL-1 and for a final concentration of 107 particles mL-1, 15 mL of stock solution was added to the aquaria. P. putida was cultured on liquid LB medium to a concentration of 5×108 cells mL-1 and concentrated by centrifugation to a concentration of 1.2×109 cells mL-1. From this stock of P. putida, 6.7 mL was added to the aquaria resulting in a final concentration of 107 cells mL-1 in the aquaria. After the addition of micro-spheres and bacteria, samples were taken at several time points during incubation to monitor the concentration of particles (micro-spheres and P. putida) in the water and the amount of particles trapped in the biofilms. For determination of particles in the water 4×10 mL of water was sampled from each aquarium, fixated by addition of 10 mL 96% EtOH and stored at 4°C prior to further analyses. For determination of trapped particles in the biofilms, four glass discs (surface area 1.5 cm2

per disc) were removed from each aquarium; the biofilm was removed from the glass slide using a razor blade and transferred into four separate tubes. The biofilm samples were fixated with 10 mL 50% EtOH, homogenised with an ultra-blender and stored prior to further analyses. Samples were divided for the analyses of micro-spheres and P. putida, which resulted in duplicate measurements of each aquarium for both micro-spheres and P. putida. Fluorescence micro-spheres concentrations in the water and in the biofilm were determined using flow-cytometry (Epics Elite instrument, Coulter Corporation, Hialeah, FL). For P. putida concentrations in the water and in the biofilm samples were filtered over 0.2 µm poly carbonate membranefilters (Whatman, Ø25 mm), the filters were labelled with a P. putida specific fluorescence rRNA probes (DuTeau et al. 1998) using fluorescence in situ hybridization (FISH) according to Glockner et al. (1996). After labelling the filters were mounted on microscope slides and P. putida cells were counted using an inverted fluorescence microscopy (Olympus). First order removal rates (kd; exponential decrease) were calculated using MS Excel (eq.1).

92

Chapter 5

Results

Experiment I: Particle sizeThe control treatment with only particles showed no significant (p>0.05) decrease in

concentrations of particles for all sizes during the experimental period (Fig. 5.1; dotted grey lines). The treatment with biofilm but without addition of particles showed release of particles by the biofilms into the water, but the concentration of released particles remained the same over the experimental period for all particle sizes (data not shown). This resulted in higher concentrations of suspended particles in the treatment with both biofilm and micro-spheres compared with the control treatment with micro-spheres only. Suspended particle concentrations released by the biofilm were on average 19 ±3% lower than the concentrations of added micro-spheres for each particle size. Average removal rates over 150 min of incubation were of 140 ±25, 39 ±16, 18 ±4, 8 ±2 and 2 ±1 particles mL-1 min-1 (± s.e.) for particles size classes of 1.5 - 3.0 µm, 3.0 - 5.0 µm, 5.0 - 8.0 µm, 8.0 - 15 µm, and 15 - 35 µm respectively (Fig. 5.1a-e). The corresponding first order removal rates (Kd constants) were 0.188 h-1 (1.5 - 3.0 µm), 0.225 h-1 (3.0 - 5.0 µm), 0.246 h-1 (5.0 - 8.0 µm), 0.244 h-1 (8.0 - 15 µm) and 0.231 h-1 (15 - 35 µm) (Fig. 5.6).

In the biofilm, a significant (p<0.05) increase in particle concentrations was observed after 150 min for the size classes 1.5 - 3.0 µm, 3.0 - 5.0 µm, and 5.0 - 8.0 µm. A similar, but not significant, increase was observed for particles of 8.0 - 15 µm (Fig. 5.2). The trapping efficiency differed strongly between the different size classes: the percentage of trapped particles was highest for particles in the size class of 5.0 - 8.0 µm, (49 ±12% of the particles was trapped after 150 min) while less than 5% trapping was observed for the smallest and biggest size classes (Fig. 5.3).

Biofilm trapping

93

0

10000

20000

30000

40000

50000

60000

70000

0 30 60 90 120 150

0

5000

10000

15000

20000

0 30 60 90 120 150

010002000300040005000600070008000

0 30 60 90 120 150

0

500

1000

1500

2000

2500

3000

3500

0 30 60 90 120 150

0100200300400500600700800

0 30 60 90 120 150

1.5 – 3.0 µm

3.0 – 5.0 µm

5.0 – 8.0 µm

8.0 - 15 µm

15 - 35 µmPart

icle

s in

wat

er (n

o. m

L-1

)

Time (m) 0

200

400

600

800

0 30 60 90 120 150

Without biofilm With biofilm

a)

b)

d)

c)

e)

Fig. 5.1 Experiment 1; Average particle concentrations (particles mL-1) in solution for several particle size classes (a-e), over an incubation period of 150 minutes, after addition of a mixture of micro-spheres at t=0, in the presence (solid black line with open circles; n=8-9) or absence (dotted grey line with solid squares; n=3) of a natural mature biofilm.

94

Chapter 5

1000

10000

100000

1000000

10000000

1.5 - 3.0 µm 3.0 - 5.0 µm 5.0 - 8.0 µm 8.0 - 15 µm 15 - 35 µm

Only biofilm Particles added

Part

icle

s in

bio

film

(no.

cm

-2) *

**

Particle size1000

10000

100000

1000000

10000000

2.0 µm 4.5 µm 6.0 µm 10 µm 25 µm

Only biofilm Particles added

Fig. 5.2 Experiment 1; Average particle concentrations (particles cm-2) in natural mature biofilm for several particle size classes, after an incubation period of 150 minutes with (white bars; n=8) or without (grey bars; n=3) addition of a mixture of micro-spheres differing in sizes at t=0. Error bars represent the standard error and stars indicate significant difference (P<0.05).

-20%

0%

20%

40%

60%

80%

100%

1.5 - 3.0 µm 3.0 - 5.0 µm 5.0 - 8.0 µm 8.0 - 15 µm 15 - 35 µm

Add

ed p

artic

les

capt

ured

by

biof

ilm (

%)

Particle size

b

bc cd

aad

Fig. 5.3 Experiment 1; Average calculated trapping efficiency of micro-spheres by natural mature biofilms after an incubation period of 150 minutes. Error bars represent the standard error (n=8) and letters indicate similarity (P<0.05).

Experiment II: Biofilm composition and particle type Fifteen minutes after addition of the particles, the average concentration of micro-spheres

in the water was 1.2×107 ±1.9×106 particles mL-1 (± s.d.) (Fig. 5.4). In the blank (aquarium with glass slides without attached biofilm) the micro-spheres concentration remained similar to the starting concentration for the first 8 h (Fig. 5.4a). After 8 h the concentration of micro-spheres in water of the blank treatment steadily decreased to a concentration of 2.0×106 ±6.6×104 particles mL-1 after 72 h.

Biofilm trapping

95

During the same period of time the number of particles found in the biofilms increased to 1.9×105

±0.3×104 particles cm-2. The decrease of micro-spheres from the water was lower in the presence of cyanobacterial biofilms (Fig. 5.4b,c) and after 72 h micro-spheres concentration were 2.2×106 ±1.2×105 and 7.8×106 ±1.9×105 particles mL-1 respectively. The trapping of particles by L. foveolarum after 72 h was high compared with the blank (Fig. 5.4b; 6.8×106 ±4.6×103), whereas particle numbers in C. stagnale biofilm were similar to the blank treatment (Fig. 5.4c;1.5×105 ±2.5×103). The natural biofilms and cultured diatom biofilms all showed a stronger decrease in micro-sphere concentrations in the water compared with the blank treatment with concentrations of 2.0×106 ±2.3×103, 2.4×106 ±4.4×104, 1.3×106 ±4.7×104 and 9.2×104 ±4.0×101 particles mL-1 after 72 h for N. perminuta, A. lanceolata, Natural biofilm I and Natural biofilm II, respectively (Fig. 5.4d-g). Simultaneously the number of particles in the biofilm increased ranging from 6.1×105 ±1.2×103 particles cm-2 (Fig. 5.4d; N. perminuta) to 1.8×106 ±7.0×103 particles cm-2 (Fig. 5.4f; Natural biofilm I) after 72 h. The first order removal constants (Kd) were 0.025 h-1, 0.025 h-1, 0.006 h-1, 0.019 h-1, 0.023 h-1, 0.032 h-1 and 0.068 h-1 for the blank, L. foveolarum, C. stagnale, N. perminuta, A. lanceolata, Natural Biofilm I and Natural Biofilm II respectively (Fig. 5.6).

In the blank treatment (no biofilm present), P. putida had a maximum abundance in the water of 1.9×106 ±4.3×105 cells mL-1 after 30 min of incubation and then gradually decreased to 9.9×102 ±4.0×103 cells mL-1 after 72 h (Fig. 5.5a). A small fraction of P. putida cells attached to the empty glass slides during the first 8 h of incubation and decreased to zero again after 48 h incubation. In biofilms consisting of cyanobacteria (L. foveolarum and C. stagnale) numbers of P. putida cells also decreased to zero after 72 h (Fig. 5.5b,c). The P. putida abundance found in the C. stagnale biofilm did however show substantial increase of P. putida abundance in the biofilm after 24 h of incubation (Fig. 5.5c). Incubation with biofilms composed of N. perminuta removed P. putida cells in the water phase almost completely within 48 h of incubation, although this decrease did not result in a substantial increase in P. putida cells in the N. perminuta biofilm (Fig. 5.5d). Biofilms of A. lanceolata caused a very rapid decline of P. putida cells in the water phase within the first 24 h of incubation, which coincided with a strong increase of P. putida cells in the biofilm (Fig. 5.5e). Both natural biofilms, especially Natural biofilm I, showed a strong decrease of P. putida cells in the water phase, which completely disappeared within the first 48 and 24 h of incubation with respectively Natural biofilm II and Natural biofilm I (Fig. 5.2f,g). In Natural biofilm II an increase of P. putida cells in the biofilm were observed short after addition of P. putida, but rapidly decreased again within the 24 h (Fig. 5.5g). This rapid disappearance of P. putida cells in the biofilm was also observed in Natural biofilm I were P. putida almost directly disappeared (Fig. 5.5f).

96

Chapter 5

Fig. 5.4 Experiment 2; development of micro-sphere concentration (1µm; particles mL-1) in suspension (solid black line with open circles) and number of micro-spheres in biofilm (particles cm-2; dotted grey line with solid squares) over an incubation period of 72 h after addition of micro-spheres at t=0 h.

L. foveolarum (cyanobacteria)

N. perminuta (diatomeae)

Time (h)

Part

icle

sno

. mL

-1(w

ater

)

Blank

Natural biofilm I Natural biofilm II

C. stagnale (cyanobacteria)

A. lanceolata (diatomeae)

0

500000

1000000

1500000

2000000

2500000

0 24 48 72

Water Biofilm

a)

b)

d)

f)

c)

e)

g)Pa

rtic

les

no. c

m-2

(bio

film

)

0

500000

1000000

1500000

2000000

2500000

0

2000000

4000000

6000000

8000000

10000000

12000000

14000000

16000000

0 24 48 72

Water Biofilm

0

500000

1000000

1500000

2000000

2500000

0

2000000

4000000

6000000

8000000

10000000

12000000

14000000

16000000

0 24 48 72

Water Biofilm

0

500000

1000000

1500000

2000000

2500000

0

2000000

4000000

6000000

8000000

10000000

12000000

14000000

16000000

0 24 48 72

Water Biofilm

0

500000

1000000

1500000

2000000

2500000

0

2000000

4000000

6000000

8000000

10000000

12000000

14000000

16000000

0 24 48 72

Water Biofilm

0

500000

1000000

1500000

2000000

2500000

0

2000000

4000000

6000000

8000000

10000000

12000000

14000000

16000000

0 24 48 72

Water Biofilm

0

500000

1000000

1500000

2000000

2500000

0

2000000

4000000

6000000

8000000

10000000

12000000

14000000

16000000

0 24 48 72

Water Biofilm

0

500000

1000000

1500000

2000000

2500000

0

2000000

4000000

6000000

8000000

10000000

12000000

14000000

16000000

0 24 48 72

Water Biofilm

Biofilm trapping

97

0

500000

1000000

1500000

2000000

2500000

0100002000030000400005000060000700008000090000

100000

0 24 48 72

Biofilm Water

0

500000

1000000

1500000

2000000

2500000

0100002000030000400005000060000700008000090000

100000

0 24 48 72

Biofilm Water

0

500000

1000000

1500000

2000000

2500000

0100002000030000400005000060000700008000090000

100000

0 24 48 72

Biofilm Water

Time (h)

Pseu

dom

onas

put

ida

in b

iofil

m (n

o. c

m-2

)

Pseu

dom

onas

put

ida

in w

ater

(no.

mL

-1)

Blank

Natural biofilm I Natural biofilm II

0

500000

1000000

1500000

2000000

2500000

0100002000030000400005000060000700008000090000

100000

0 24 48 72

Biofilm Water

a)

f)0

500000

1000000

1500000

2000000

2500000

0100002000030000400005000060000700008000090000

100000

0 24 48 72

Biofilm Water

0

500000

1000000

1500000

2000000

2500000

0100002000030000400005000060000700008000090000

100000

0 24 48 72

Biofilm Water

N. perminuta (diatomeae) A. lanceolata (diatomeae)d) e)0

500000

1000000

1500000

2000000

2500000

0100002000030000400005000060000700008000090000

100000

0 24 48 72

Biofilm Water

0

500000

1000000

1500000

2000000

2500000

0100002000030000400005000060000700008000090000

100000

0 24 48 72

Biofilm Water

L. foveolarum (cyanobacteria) C. stagnale (cyanobacteria)b) c)

g)

Fig. 5.5 Experiment 2; development of P. putida abundance (cells mL-1) in suspension (solid black line with open circles) and numbers of P. putida in biofilm (cells cm-2; dotted grey line with solid squares; attachment to glass disk in blank) over an incubation period of 72 h after addition of micro-spheres at t=0 h.

98

Chapter 5

0.000

0.050

0.100

0.150

0.200

0.250

0.300

0.350Fi

rst o

rder

rem

oval

rate

(kd)

Micro-spheres P. putida

Fig. 5.6 The first order removal rates of particles from suspension in the first (left) and second (right) experiment for microspheres (grey bars) and pseudomonas putida (white bars). Microspheres used in the first experiment had a nominal size of 1 µm, in the second experiment size classes of particles were measured, but microspheres added had a nominal size of 2.0 µm, 4.5 µm, 6.0 µm, 10 µm and 25 µm (left to right).

DiscussionSubstantial trapping of particles was observed for particles smaller than 15 µm, while bigger

(>15 µm) particles were hardly removed from suspension by the biofilms present. The efficiency of particle trapping by phototrophic biofilms was, however, found to be strongly dependent on the size of the suspended particles and an optimum particle trapping efficiency (almost 50% removal from the suspension in a short period of time) was observed for particles between 5.0-8.0 µm diameter. These findings of substantial particle trapping partially corroborate a study of Stott and Tanner (2005), who described increasing trapping efficiency by natural mixed species biofilm grown in constructed wetlands with increasing particle sizes ranging from viruses (0.1 µm), bacteria (1 µm) to parasitic protozoa (4.5 µm). The trapping of particles does not only occur at the surface of the biofilm, but also deeper in the biofilm depending on the transport of particles via water channels in biofilms (Okabe, Yasuda and Watanabe, 1997). The abundance and dimensions of these water channels may affect the penetration, collision chance, surface area and consequently affect the possibility for particle attachment in biofilms. Trapping of large particles may be hindered by lack of penetration possibilities, whereas small particles may have a relative lower chance of coming into contact with the biofilm, resulting in an optimum trapping efficiency of medium sized particles.

The removal rates (kd) of the 1 µm diameter micro-spheres by the two natural biofilms (0.32 h-1 and 0.68 h-1) observed in experiment 2 of this study are comparable to previously observed removal rates (Stott and Tanner 2005). In our first experiment, however, much higher removal rates (ranging from 0.188 h-1 and 0.246 h-1) were observed. This difference could be explained by differences in the composition and maturity of the biofilms (resulting from differences in the colonization period) and

Biofilm trapping

99

by differences in the experimental setup (mainly the acclimatization period in experiment 2 could have cleaned the biofilms before the addition of particles). Although the removal rates are based on the initial trapping of particles in clean biofilms, and are therefore probably an overestimation of long-term trapping rates, in general the biofilms showed a high capacity to strongly reduce the suspended particle concentrations in the water.

The second experiment showed that the composition of the phototrophic biofilm, which can differ greatly in growth form, mucus production and associated heterotrophic bacteria, is an important factor in the trapping of suspended particles. Diatom biofilms were more effective in trapping micro-spheres from the surrounding water compared to cyanobacteria biofilms which virtually trapped no particles. Stott and Tanner (2005) also showed differences in particle trapping depending on the species composition, where phototrophic biofilms removed more particles compared to heterotrophic biofilm.

Biofilms that were capable of trapping high numbers of micro-spheres also showed significant trapping of P. putida cells with removal rates of P. putida cells being even up to one order of magnitude higher than removal rates of the micro-spheres. This difference between micro-spheres and P. putida cells could indicate the occurrence of selective particle trapping by the biofilms, but since P. putida generally prefers a sessile lifestyle and actively attach to any solid surface, it cannot be excluded that this difference could also (partly) be caused by attachment to the walls of the aquaria. In addition, the biofilms seemingly kill off the majority of trapped P. putida cells. In the natural biofilms, the rapid depletion of suspended P. putida cells was accompanied with only a relative small increase and sequential rapid decrease of P. putida cells in the biofilm. This disappearance of P. putida cells was probably caused by protozoa commonly associated with natural biofilm (Curds 1982; Fenchel 1986). In a study of Chabaud et al. 2006 the capacity of removal of particles from septic effluent was subscribed for 60% to biofilm associated protozoa. Eisenmann et al. (2001) also subscribed a substantial contribution to the trapping of suspended particles from wastewater by ciliates and feed activity by ciliates in biofilm have been suggested to increase particles trapping (Okabe et al. 1997; Eisenmann et al. 2001). Moreover, certain phototrophic species inhabiting the biofilms could express antibiotic effects and this may also contribute to different killing rates of P. Putida.

These experiments have shown that the trapping of particles from suspension by phototrophic biofilms in aquatic ecosystems can be very substantial for particles smaller than 15 µm. The efficiency of particle trapping is however highly variable and is dependent on both the size and nature of the particles as the species composition of the phototrophic biofilm.

AcknowledgementsThis study is supported by the European Commission under contract no. EVK1-1999-00001

and the Foundation for Applied Water Research (STOWA). The Netherlands Centre for Limnology (KNAW-NIOO) supported the flow-cytometric counting of micro-spheres.

100

Chapter 5

References

Alleman JE, Veil JA and Canaday JT (1982). Scanning Electron-Microscope Evaluation of Rotating Biological Contactor Biofilm. Water research 16 (5), 543-550.

Barranguet C, van Beusekom SAM, Veuger B, Neu TR, Manders EMM, Sinke JJ and Admiraal W (2004). Studying undisturbed autotrophic biofilms: still a technical challenge. Aquatic Microbial Ecology 34 (1), 1-9.

Barranguet C, Veuger B, Van Beusekom SAM, Marvan P, Sinke JJ and Admiraal W (2005). Divergent composition of algal-bacterial biofilms developing under various external factors. European Journal of Phycology 40 (1), 1-8.

Battin TJ, Sloan WT, Kjelleberg S, Daims H, Head IM, Curtis TP and Eberl L (2007). Microbial landscapes: new paths to biofilm research. Nature Reviews Microbiology 5 (1), 76–81.

Bouwer EJ (1987). Theoretical investigation of particle deposition in biofilm systems. Water Research 21 (12), 1489–1498.


Costerton JW, Lewandowski Z, Caldwell DE, Korber DR and Lappin-Scott HM (1995). Microbial biofilms. Annual Review of Microbiology 49, 711–745.

Curds CR (1982). The ecology and role of protozoa in aerobic sewage treatment processes. Annual Review of Microbiology 36, 27–46.

DuTeau N, Rogers J, Bartholomay C and Reardon K (1998). Species-specific oligonucleotides for enumeration of Pseudomonas putida F1, Burkholderia sp. strain JS150, and Bacillus subtilis ATCC 7003 in biodegradation experiments. Applied and Environmental Microbiology 64 (12), 4994-4999.


Fenchel T (1986). Protozoan filter feeding. In: Corliss JO and Patterson Dj, Progress in protistology. Biopress Ltd., Bristol, 65–113.

Gantzer CJ, Rittmann BE and Herricks EE (1988). Mass transport of biodegradable materials to streambed biofilms. Water Research 22 (6), 709–723.

Biofilm trapping

101

Gerba CP (1984). Applied and Theoretical Aspects of Virus Adsorption to Surfaces. Advances in Applied Microbiology 30, 133-168.


Guillard RRL and Lorenzen CJ (1972). Yellow-green algae with chlorophyllide c. Journal of Phycology 8 (1), 10-14.

Hall-Stoodley L, Costerton JW and Stoodley P (2004). Bacterial biofilms: from the natural environment to infectious diseases. Nature Reviews Microbiology 2 (2), 95–108.

Hammer Ø, Harper DAT and Ryan PD (2001). PAST: Paleontological statistics software package for education and data analysis. Palaeontologia Electronica 4 (1), 1-9.

Hecky RE and Hesslein RH (1995). Contributions of benthic algae to lake food webs as revealed by stable isotope analysis. Journal of the North American Benthological Society 14 (4), 631–653

Jowett I and Biggs B (1997). Flood and velocity effects on periphyton and silt accumulation in two New Zealand rivers. New Zealand Journal of Marine and Freshwater Research 31 (3), 287-300.


Okabe S, Yasuda T and Watanabe Y (1997). Uptake and release of inert fluorescence particles by mixed population biofilms. Biotechnology and bioengineering 53 (5), 459-469.

Okabe S, Kuroda H and Watanabe Y (1998). Significance of biofilm structure on transport of inert particulates into biofilms. Water Science and Technology 38 (8-9), 163-170.

Ongerth JE and Pecoraro JP (1996). Electrophoretic mobility of Cryptosporidium oocysts and Giardia cysts. Journal of Environmental Engineering-Asce 122 (3), 228-231.

Romani A, Giorgi A, Acuna V and Sabater S (2004). The influence of substratum type and nutrient supply on biofilm organic matter utilization in streams. Limnology and Oceanography 49 (5), 1713-1721.


102

Chapter 5

Stevik TK, Aa K, Ausland G and Hanssen JF (2004). Retention and removal of pathogenic bacteria in wastewater percolating through porous media: a review. Water research 38 (6), 1355-1367.

Stewart PS and Franklin MJ (2008). Physiological heterogeneity in biofilms. Nature Reviews Microbiology 6 (3), 199–210

Sutherland IW (2001). The biofilm matrix—an immobilized but dynamic microbial environment. Trends in Microbiology 9 (5), 222–227.

van der Grinten E, Janssen M, Simis S, Barranguet C and Admiraal W (2004). Phosphate regime structures species composition in cultured phototrophic biofilms. Freshwater Biology 49 (4), 369-381.

Wilderer P and McSwain B (2004). The SBR and its biofilm application potentials. Water Science and Technology 50 (10), 1-10.

Woodruff S, House W, Callow M and Leadbeater B (1999). The effects of biofilms on chemical processes in surficial sediments. Freshwater Biology 41 (1), 73-89.

Vadeboncoeur Y, Peterson G, Vander Zanden MJ and Kalff J (2008). Benthic algal production across lake size gradients: interactions among morphometry, nutrients, and light. Ecology 89 (9), 2542–2552.

Suspended particle and faecal indicator organism peak discharge buffering by a surface flow constructed wetland

Chapter 6

Based on: Mulling BTM, van den Boomen RM, van der Geest HG, Kappelhof JWNM, Admiraal W (2013). Suspended particle and pathogen peak discharge buffering by a surface-flow constructed wetland. Water research 47, 1091-1100.

104

Chapter 6

Abstract Constructed wetlands (CWs) have been shown to improve the water quality of treated wastewater. The capacity of CWs to reduce nutrients, pathogens and organic matter and restore oxygen regime under normal operating conditions cannot be extrapolated to periods of incidental peak discharges. The buffering capacity of CWs during peak discharges is potentially a key factor for water quality in the receiving waters. Therefore, the aim of the present study was to investigate the behaviour of peak discharges of suspended particles, (associated) physicochemical parameters and pathogenic organisms from a wastewater treatment plant (WWTP) in a full scale constructed wetland (CW). By mixing clarified water and sludge rich water from the settlement tank of the WWTP, the suspended particle concentration was increased for eight hours from ± 3.5 to ± 230 mg L-1, and discharged into a full scale surface flow constructed wetland. An increase of suspended particle concentration following the peak discharge concurred with increases in turbidity and oxygen demand, total nutrient load (nitrogen, phosphorus and carbon) and pathogens (E. coli and Enterococci). Temperature, pH, conductivity and dissolved nutrient concentrations (nitrogen, phosphorus and carbon) were however unaffected by the initial peak discharge. After retention in the unvegetated ponds (the first CW compartment) the applied suspended particle peak with a total load of 86.2kg was reduced by >99%. Similar peak buffering was observed for the turbidity, oxygen demand and settable volume. Simultaneously dissolved nutrient concentrations increased, indicating partial mineralization of the suspended particles during retention in the unvegetated ponds. The peak buffering of pathogens was lower (40-84%), indicating differences in removal processes between other suspended particles and pathogens. The results indicated that the suspended particles were probably mostly removed by sedimentation and mineralization, where pathogens were more likely buffered by biofilm retainment, mortality and predation, mainly in reed ditches. After passing through the total CW the residuals of the suspended particle peak discharge were temporal increased concentrations of inorganic carbon (IC), NH4 and E. coli (respectively 11%, 17% and 160% higher than steady state concentrations). The observations support the positive role of CWs for effective buffering of wastewater discharge peaks.

Peak discharge

105

Introduction Constructed wetlands (CWs) are man-made ecological systems that are used in a wide variety of applications to improve water quality. Since the first attempts to use CWs for water quality improvements of untreated wastewater in the early 1950s, the development and use of CWs for wastewater treatment has spread across the world (Sundaravadivel and Vigneswaran 2001). Nowadays, CWs are often used as an additional polishing step to reduce the potential impact of treated wastewater on receiving surface waters (Kadlec and Wallace 2008; Vymazal 2005a). CWs are designed to utilize several occurring physical, chemical and biological processes, like sedimentation and microbial degradation to reduce the negative impact of various constituents in (treated) wastewater. It is demonstrated that concentrations of nutrients (especially nitrogen and phosphorus), organic compounds, suspended particles, pathogens, heavy metals and hormones (Kadlec and Wallace 2008) are significantly reduced by CWs: under normal operating conditions removal efficiencies for organic compounds and suspended particles range between 60-95%, while nutrient removal efficiencies are generally below 60% although higher efficiencies up to 90% have been reported (Zhang et al. 2011; Zurita et al. 2009; Vymazal 1996; Vymazal 2007; Fisher and Acreman 2004; Cameron et al. 2003). Removal of pathogens (including bacteria, viruses and parasites) is generally high, with removal efficiencies ranging between 80 to >99% (Foekema 2012; Reinoso et al. 2008; Vidales-Contreras et al. 2006; Vymazal 2005b). The removal of hormones like estrogens generally ranges between 35-95% (Cai et al. 2012; Song et al. 2011; Shappell et al. 2007; Gray and Sedlay 2005). In the Netherlands, several CWs are operational for more than 10 years. From monitoring studies it is demonstrated that in these CWs concentrations of nitrogen and phosphorous in the effluent of tertiary treatment WWTP are reduced with 10-25% and 2-40% respectively (van den Boomen and Kampf 2012), pathogens with a log 2.0 to 2.5 (>99%) (Foekema 2012; van den Boomen et al. 2012) while organic and suspended matter reduction is very limited (van den Boomen and Kampf 2012; Van den Boomen et al. 2012). The removal efficiencies mentioned above are mainly based on measurements during normal operating conditions. The removal efficiency is however affected by the hydraulic and pollutant loading, with increasing loadings causing decreasing efficiencies (Vymazal 2007; Fisher et al. 2009; Toet et al. 2005). Present knowledge so far is thus based on relative constant loading levels and monthly monitoring data, and little is known about removal efficiencies during incidental peak discharges. Kruit et al. (2009) showed that incidental peak discharges carry elevated loads of suspended particles (sludge) and occur a few times a year at several WWTPs in The Netherlands, primarily caused by storm water inflow and malfunctioning of the wastewater treatment plant. The present study aims to investigate the behaviour of incidental peaks in concentrations of suspended particles in a constructed wetland system and discuss the major processes affecting peak behaviour in individual CW compartments. A suspended particle peak was induced and discharged into a full scale horizontal flow constructed

106

Chapter 6

wetland receiving secondary-treated municipal wastewater. The suspended particle peak was monitored in the CW by analysing selected physicochemical parameters, concentrations of suspended particles, organic matter, nutrients and faecal indicator organisms at the in- and outflow of the two different CW compartments, unvegetated ponds and reed beds (Phragmites australis).

Materials and methods

Design of the constructed wetland This study was carried out in a full scale surface flow constructed wetland (CW) located in Grou, The Netherlands (N53 05.535 E5 49.050). The CW was constructed in 2006 and receives secondary-treated municipal (mainly domestic) wastewater with a constant hydraulic loading of 1200 m3 day-1. After inflow of treated municipal wastewater from a settlement tank, the water flows through the CW consisting of a series of three unvegetated ponds and four parallel reed beds (Fig. 6.1). At the end of the reed beds the water is pumped into an ecological buffer zone which is in open connection with the receiving surface water (channel) (Fig. 6.1). The unvegetated ponds have an average depth of 1.35m, volumes between 360 and 440 m3 each and total hydraulic retention time (HRT) of 17.9 h (Table 1). The reed beds are covered with Phragmites australis. They have an average water depth of 40 cm, approximate volume of 443 m3 and an average HRT of 23.6 h. The total HRT of the complete CW is 41.5 h. The hydraulic retention times were calculated from the retention time distribution obtained from a tracer experiment using lithium chloride (Paragraph 6.2.3).

Table 6.1 Dimensions and hydraulic retention time (HRT) of CW Aqualân Grou. The length, width and volume of the ponds were manually determined and calculated, the length, width and volume of the reed beds were calculated from the construction blueprints. The HRT’s were determined by a tracer experiment.



Length (m) 55 55 55 165 110 110

Average width (m) 7.6 8.1 8.1 7.9 11.5 46

Average depth (m) 1.34 1.31 1.43 1.35 0.4 0.4

Surface area (m2) 418 446 446 1304 1265 5060 6364

Volume (m3) 362 388 441 1191 443 1771 3552


HRT (h) 17.9 23.6 41.5

Peak discharge

107

1

2

2

2

3 3 3 3

4

5

a

b

c

0 25 50 m0 25 50 m

0

Fig. 6.1 Map of the WWTP in Grou, the Netherlands (0) with a settlement tank (1) which discharges treated municipal wastewater into a CW consisting of unvegetated ponds (2), reed beds (3) and an ecological buffer zone (4) which is in open connection with the receiving channel (5). Sampling points were located at; PONDS-IN (a), PONDS-OUT (b), REED-BEDS-OUT (c).

Suspended particle peak discharge To experimentally induce a peak of suspended particles, the normal inflow of treated wastewater into the CW was temporally replaced by an external pump connected to two inflow points located in the settlement tank of the WWTP. By mixing water from near the water surface of the settlement tank and particle rich water from deeper layers of the settlement tank, a controlled suspended particle peak was discharged during 8h into the CW while maintaining normal hydraulic loading.

Conservative tracer One hour before the suspended particle peak discharge, an artificial conservative hydraulic tracer (Leibundgut et al. 2009) in the form of lithium chloride (10 kg LiCl; 1.6 kg Li+; 16 g Li+ L-1) was discharged into the CW as a pulse directly above the inflow point of the CW and monitored using a Li+-selective electrode (Mettler Tolledo, DX207) to 1) determine the retention time distribution and amount of dead volumes present in the CW, 2) estimate the sampling times for the suspended particle peak at each individual sampling point and 3) describe the peak behaviour of a conservative dissolved substance that can serve as a reference for the other peak measurements. The average hydraulic retention times were determined at 50% passage of the lithium chloride load.

108

Chapter 6

Sampling The experiment was conducted in November 2009. Sampling stations were located at the inflow of the unvegetated ponds (PONDS-IN; a), between the unvegetated ponds and the reed beds (PONDS-OUT; b) and at the outflow of the reed beds (REED-BEDS-OUT; c) (Fig. 6.1; a-c), To determine the background levels of each parameter during normal operation, water samples were taken one week before, two hours before and two weeks after the peak discharge. To monitor the suspended particle peak during retention in the CW, four water samples were taken during passing of the peak at each sampling point. At each sampling point the first water samples were taken between one to two hours after observing the maximum lithium concentration, the next three samples were taken with an interval between 2 (at the outflow of the unvegetated ponds) and 4 h (at the outflow of the reed beds). Water samples were taken 10-20 cm below the water surface using a 10 L vessel from which subsamples were taken and stored appropriately prior to analysis.

Measurements and analysis To accurately describe the behaviour of the suspended particle peak in the CW, and quantify changes in the associated biological and physicochemical parameters, a set of measurements was performed. At each sampling station turbidity was measured continuously (10 min interval) using HACH 1720C model NTU meters. At the same frequency temperature, conductivity, redox potential, pH and dissolved oxygen (DO) was measured using HANNA HI9828 multimeters. Biochemical and chemical oxygen demand (BOD and COD) were analysed according to NEN-EN-1899-1 (1998) and NEN-6633 (2006). Suspended particles, settable volume, E. coli, Enterococci and chlorophyll-a concentrations were analysed according to standardized methods (respectively NEN-EN-872 2005; NEN-6623 2005; NEN-EN-ISO-9308-1 2000; NEN-EN-ISO-7899-2 2000; NEN-6520 2006). For determination of the total DNA (as a proxy for the total amount of organisms) content samples were filtered on site over 0.2 µm cellulose nitrate membrane filters (Whatman NC 20) and stored at -20°C prior to DNA extraction. DNA extraction was performed using DNA extraction kits (MIOBIO, Powerwater) according to the manufacturer’s instructions and total DNA content was determined using absorption at 260 nm (NanoDrop 1000 spectrophotometer; Thermo Scientific). Water samples were analysed for total nitrogen (TN; NEN-6643 2003), nitrogen kjeldahl (Nkj; NEN-6646 2006), nitrite (NO2; NEN-EN-ISO-13395 1997), nitrate (NO3; NEN-EN-ISO-13395 1997), ammonium (NH4; NEN-6646 2006), total phosphorus (TP) and orthophosphate (PO4; NEN-EN-ISO-15681-2 2005) concentrations. Total carbon (TC), inorganic carbon (IC) and total organic carbon (TOC) concentrations were analysed using a total organic carbon analyser (Schimadzu, TOC-Vcph). To determine the concentrations of dissolved organic carbon (DOC) water samples were vacuum filtered over pre-washed 0.2 µm cellulose nitrate membrane filters (Whatman NC 20) and analysed with the same method as described for the TOC. The particulate organic carbon concentration (POC) was calculated by subtracting DOC from TOC.

Peak discharge

109

Analysis of peak behaviour To determine if any of the measured parameters were affected by the induced suspended particle peak discharge, values measured during passing of the suspended particle peak were compared with maximum background values. These maximum background values were estimated for each sampling point individually from the average of four measurements taken before and after the peak discharge (eq. 1; Fig. 6.2). Measured values were classified as a peak when:

( )).(.)( bbbmp XdsXAverageXX +=> (1)

Where Xp is the value of parameter X measured during passing of the suspended particle peak, Xb the average background values of parameter X and Xbm the estimated maximum background values. Peak values less than 5% higher than Xbm were not regarded as peaks (determination limit). Using the same maximum background values (Xbm), peak loadings were determined by calculation of the peak area using linear interpolation of Xp.

Peak maxim

um

Peak load

Time

Conc

entr

atio

n

Determination limit

Background + s.d.

Background average

Fig. 6.2 Identification of the background values, determination limits, peak maximum (as concentration) and peak loads (as time integrated transport) were calculated for each individual parameter. The determination limits were considered as the average background, the standard deviation (see eq. 1) plus an additional 5%.

110

Chapter 6

Results

Conservative tracer The maximum concentration of lithium chloride measured in CW influent (discharge point) was 1.7 mg Li+ L-1 (Table 6.2). During retention in the unvegetated ponds the Li+ peak broadened and decreased 34% in maximum concentration to 1.1 mg Li+ L-1 (Fig. 6.2). Retention in the reed beds decreased the maximum concentration of lithium another 33% to 0.7 mg Li+ L-1 and caused a broad peak consisting of two peaks (Fig. 6.3). The double peak was caused by differences in hydraulic loading and consequent HRT between the four individual reed beds. From the 1.6 kg of Li+ discharged 1.2 kg (75%) was recovered at PONDS-OUT and 1.1 kg (70%) li+ at REED-BEDS-OUT.

Background values Background values of water temperature, pH, conductivity, COD, TC, IC, TOC and total DNA were relative stable at PONDS-IN with respectively variation (s.d.) less than 20% of the average values (Table 2). The suspended particle concentration, BOD, E. coli, Enterococci, turbidity, POC and Nkj, nitrogen, phosphorus, DOC, redox potential and DO were relative variable (s.d. ≥35% of the mean) at PONDS-IN (Table 6.2). On average total carbon consist of 70% inorganic carbon, 18% dissolved organic carbon and 12% particulate organic carbon. A large portion of nitrogen and phosphorus was present as dissolved fraction; ammonium (67%) and orthophosphate (67%). Background values of halve of the parameters including most nutrients, OD, turbidity, redox potential and POC, increased in stability (>10% smaller s.d.) during retention in the CW (Table 6.2). The s.d. of the pH, DO, conductivity, suspended particles and IC remained similar during retention in the CW (less than 10% change in s.d.). The variability of OC, TC, faecal indicator organisms, total DNA and temperature increased (>10% larger s.d.) during retention in the CW.

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0

Lith

ium

con

cent

ratio

n (m

g L

i+L

-1)

Time (days)a) PONDS-IN b) PONDS-OUT c) REED-BEDS-OUT

Fig. 6.3 Pulse response of the Li+ concentration at a) PONDS-IN, b) PONDS-OUT and c) REED-BEDS-OUT after addition of tracer at t=0. Values are displayed without subtraction of the background/interference value of 0.1 mg Li+ L-1.

Peak discharge

111

Description of peak discharge With measured suspended particle peak maximum of 230 mg L-1 and total load of 86.2 kg at PONDS-IN (Table 6.2), an average incidental suspended particle peak as described by Kruit et al. (2009), was successfully induced. The suspended particle peak had no influence on the temperature, pH, conductivity and concentrations of PO4, NO2, NO3, NH4, IC at PONDS-IN. Small increases of dissolved oxygen (6%) and redox potential (9%) were observed, which were probably caused by the positioning of the temporal discharge point, which was just above the water surface (for visual observations of the turbidity). The maximum suspended particle concentration was 3003% higher than the maximum background value and caused an increase in turbidity of 1709%. The largest increase was observed in the settable volume which was normally below the detection limit of 0.1 mL L-1 at PONDS-IN and increased at least 25900% (compared with the detection limit) to a concentration of 26.0 mL L-1. OD, faecal indicator organisms, total DNA content, OC, Nkj and TP also increased significantly compared with maximum background values (Table 6.2).

0

50

100

150

200

250

-1.0 -0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0

Suse

pned

e pa

rtic

le c

once

ntra

tion

(mg

L-1

)

Time (days)

a) PONDS-IN b) PONDS-OUT c) REED-BEDS-OUT

a)

b)

0

5

10

15

20

25

-1.0 -0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0

Tot

al n

itrog

en (m

g N

L-1

)

Time (days)

a) PONDS-IN b) PONDS-OUT c) REED-BEDS-OUT

Fig. 6.4 Suspended particle concentration (a) and total nitrogen (b) at a) PONDS-IN, b) PONDS-OUT and c) REED-BEDS-OUT during a 5 day period. Grey bar represents the period of suspended particles peak discharge. Background measurements before and after the peak discharge (t=<-1.0 >4.0) are not displayed for illustrational purposes.

112

Chapter 6

Table 6.2 Background values (±s.d.), maximum measured peak height and loading for the measured parameters at a) PONDS-IN, b) PONDS-OUT and c) REED-BEDS-OUT.

Average background value (s.d.) Maximum value (%)***

PONDS-IN PONDS-OUTREED-

BEDS-OUT PONDS-IN PONDS-OUTREED-

BEDS-OUT

Lithium (mg Li L-1) 0.1 (±0.0) 0.1 (±0.0) 0.1 (±0.0) 1.7 (1564%) 1.1 (1067%) 0.7 (588%)

Temperature (°C) 11.7 (±0.2) 9.9 (±1.0) 9.1 (±1.5) 11.9 (0%) 10.6 (0%) 10.3 (0%)

pH 7.4 (±0.2) 7.6 (±0.3) 7.6 (±0.1) 7.6 (0%) 7.8 (0%) 7.7 (0%)

Redox potential (mV) 196 (±146) 421 (±68) 367 (±185) 371 (9%) 449 (0%) 482 (0%)

Dissolved Oxygen (mg L-1) 0.4 (±0.5) 4.0 (±1.4) 0.9 (±1.1) 0.9 (6%) 5.1 (0%) 1.2 (0%)

Conductivity (µS cm-1) 856 (±135) 949 (±164) 1074 (±138) 945 (0%) 1005 (0%) 1176 (0%)

Turbidity (NTU) 1.1 (±0.5) 1.8 (±0.5) 1.3 (±0.3) 28.4 (1709%) 3.8 (63%) 1.3 (0%)

BOD (mg L-1) 2.3 (±1.5) 1.3 (±0.6) 1.7 (±0.6) 72.0 (1765%) 3.0 (57%) 2.0 (0%)

COD (mg L-1) 40.0 (±7.0) 37.3 (±1.5) 35.3 (±2.3) 260.0 (453%) 43.0 (11%) 37.0 (0%)

Suspended particles (mg L-1) 5.1 (±2.3) 5.7 (±1.6) 4.2 (±2.0) 230.0 (3003%) 8.7 (19%) 3.9 (0%)

Settable volume (ml L-1) <0.1 (-)* <0.1 (-)* <0.1 (-)* 26.0 (≥25900%) <0.1 (-)* <0.1 (-)*

E. Coli (CFU mL-1) 180.7 (±126.6) 58.3 (±19.6) 0.6 (±0.5) 4260.0 (1290%) 1040.0 (1240%) 2.7 (160%)

Enterococci (CFU mL-1) 21.3 (±10.1) 6.2 (±2.3) 3.4 (±2.3) 480.0 (1430%) 45.0 (430%) 1.8 (0%)

Total DNA (ng L-1) 38.3 (±4.9) 11.7 (±5.6) 15.3 (±4.6) 513.1 (1090%) 38.5 (122%) 37.3 (88%)

Chlorophyll-a (µg L-1) <7 (-)* <7 (-)* <7 (-)* <7 (-)* <7 (-)* <7 (-)*

TC (mg C L-1) 69.7 (±5.2) 51.8 (±5.9) 73.0 (±27.2) 105.0 (40%) 71.0 (37%) 68.6 (0%)

IC (mg C L-1) 48.9 (±4.8) 35.8 (±2.4) 38.7 (±3.5) 47.1 (0%) 49.3 (38%) 47.4 (12%)

TOC (mg C L-1) 20.9 (±0.4) 16.2 (±3.3) 34.2 (±23.7) 81.1 (281%) 22.9 (42%) 21.1 (0%)

DOC (mg C L-1) 12.3 (±3.7) 14.9 (±3.5) 29.1 (±23.7) 27.0 (69%) 21.6 (45%) 21.8 (0%)

POC (mg C L-1) 8.6 (±3.3) 1.3 (±0.1) 5.2 (±0.0) 54.1 (355%) 2.2 (68%) 2.8 (0%)

TN (mg N L-1) 6.8 (±3.0) 6.2 (±0.6) 5.7 (±0.7) 23.0 (135%) 9.0 (34%) 7.5 (17%)

Nkj (mg N L-1) 6.2 (±3.7) 5.3 (±1.4) 4.5 (±1.2) 22.8 (131%) 8.5 (27%) 6.3 (11%)

NO2 (mg N L-1) 0.1 (±0.1) 0.1 (±0.1) 0.1 (±0.0) 0.1 (0%) 0.1 (0%) 0.1 (0%)

NO3 (mg N L-1) 0.6 (±0.7) 0.8 (±0.9) 1.2 (±0.8) 0.8 (0%) 0.7 (0%) 1.2 (0%)

NH4 (mg N L-1) 4.7 (±3.5) 4.1 (±1.4) 3.5 (±1.1) 8.3 (0%) 7.3 (34%) 5.1 (11%)

TP (mg P L-1) 0.3 (±0.2) 0.2 (±0.1) 0.2 (±0.1) 5.2 (924%) 0.4 (21%) 0.3 (15%)

PO4 (mg P L-1) 0.2 (±0.2) 0.2 (±0.1) 0.2 (±0.0) 0.2 (0%) 0.3 (46%) 0.2 (0%)

*: Detection limit**: A combination of carbon exclusion in the analyses and high background variability of the individual carbon fraction resulted in divergent values at PONDS-IN.***: Peak height in percentage compared with the background boundary (eq. 1; Fig. 2) at each specific location.****: Added load

Peak discharge

113

Table 6.2 Continuous. (note difference in parameter units)

Peak load

PONDS-IN PONDS-OUTREED-BEDS-

OUT

Lithium (kg Li) 1.6**** 1.2 1.1

Temperature - - -

pH - - -

Redox potential - - -

Dissolved Oxygen (g) 1.9 - -

Conductivity - - -

Turbidity - - -

BOD (kg) 21.3 1.2 -

COD (kg) 90.5 0.2 -

Suspended particles (kg) 86.2 0.2 -

Settable volume (L) x103 9.6 - -

E. Coli (CFU) x109 1355 817.6 0.4

Enterococci (CFU) x109 264.4 43.2 -

Total DNA (g) 283.9 29.1 5.3

Chlorophyll-a - - -

TC (kg C)** 3.6 15.8 -

IC (kg C) - 10 0.9

TOC (kg C)** 7.8 5.7 -

DOC (kg C) 5.5 7.3 -

POC (kg C)** 4.9 0.2 -

TN (kg N) 5.9 2.1 0.4

Nkj (kg N) 1.5 0.1 0.1

NO2 - - -

NO3 - - -

NH4 (kg N) - 1.7 0.9

TP (g P) 2197.7 13.6 14.7

PO4 (g P) - 31.7 -

114

Chapter 6

Peak buffering in the constructed wetland Fig. 6.4 shows the suspended particle and TN concentration measured at the three sampling locations in the CW before, during and after discharge of the suspended particle peak discharge, as two examples for parameters directly influenced by the suspended particle peak discharge. Peaks were high and narrow at the inflow of the CW (matching with the discharge period) and were observed to broaden and decrease in height during retention in the CW. The peak maximum of suspended particles decreased strongly from 230.0 mg L-1 to 8.7 mg L-1 from the inflow of the CW to the outflow of the ponds (Table 6.2). Although the maximum suspended particle concentration observed at the outflow of the ponds is still 19% higher compared to the background concentration, the load of the suspended particle peak was almost completely buffered (>99%) during retention in the unvegetated ponds with a reduction of 86 kg of the total 86.2 kg discharged (Table 2; Fig. 6.4). The strong buffering in both terms of peak maximum and load of the suspended particle peak during retention in the unvegetated ponds was also observed for the settable volume, COD, BOD, total DNA, POC, TP and turbidity. TN and Nkj were buffered to a lesser extent during retention in the unvegetated ponds, with a 83% and 86% decrease in peak maximum and 64% and 90% of peak loading respectively (Table 6.2; Fig 6.4). At the outflow of the ponds, peak concentrations of PO4, NH4 and IC were respectively 46%, 34% and 38% higher than the maximum background concentration. After retention in the reed beds, most of the parameters that were raised by the suspended particle peak discharge, returned to background levels (Table 6.2). Although the peak maximum of TP at the outflow of the reed beds was still 15% higher than the maximum background value, the TP load was strongly buffered during retention in the CW, with only 14.7 g P of a total peak load of 2197.7 g P remaining (Table 6.2; Fig. 6.5). The buffering of TN during retention in the reeds beds was lower compared to the unvegetated ponds. With TN decreasing from 5.9 to 0.4 kg N, 7% of the TN peak load was still present (mainly as NH4) at the outflow of the CW (Table 6.2; Fig. 6.4). The maximum peak concentration of E. coli decreased by 76% during retention in the unvegetated ponds, which is low in comparison to the reduction of the Enterococci peak maximum of 92% and suspended particles (99%). During retention in the reed beds E. coli peak load decreased considerably stronger compared to reduction during retention in the unvegetated ponds (Fig. 6.5). In contrast with Enterococci, which showed no difference from the maximum background value, the maximum peak concentration of E. coli at the outflow of the CW was still 160% higher than the maximum background value. Although the additional 1355 × 109 CFU E. coli during the peak discharge was buffered for > 99%, the peak discharge caused an additional amount of 0.4 × 109 CFU E. coli discharged into receiving surface waters (Table 6.2; Fig. 6.5). Similarly, also the total DNA peak was almost completely buffered (98%) but still, 5.3 g of additional DNA was discharged into receiving surface waters, being 88% higher than background levels (Table 6.2; Fig 6.5).

Peak discharge

115

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%

Settable volume (m3)

COD (kg)

Suspended particles (kg)

POC (kg C)

BOD (g)

Turbidity (NTU) x106

Enterococci (CFU) x109

TOC (kg C)

E.Coli (CFU) x109

most aof

Total DNA (g)

TN (kg N)

Nkj (kg N)

Lithium (kg)

Suspended particles (mg L-1)

Ponds Reed beds


Lithium

Nkj

TN

Total DNA

TP

E.Coli

TOC

Enterococci

Turbidity

BOD

POC

Suspended particles

COD

Settable volume

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%


COD (kg)


POC (kg C)

BOD (g)



TOC (kg C)

E.Coli (CFU) x109

most aof

Total DNA (g)

TN (kg N)

Nkj (kg N)

Lithium (kg)


Ponds Reed bedsUnvegetated ponds

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%


COD (kg)


POC (kg C)

BOD (g)



TOC (kg C)

E.Coli (CFU) x109

most aof

Total DNA (g)

TN (kg N)

Nkj (kg N)

Lithium (kg)


Ponds Reed bedsReed beds

Fig. 6.5 The total reduction of the peak loadings and the relative contribution of the two different CW compartments; unvegetated ponds (grey) and reed beds (white) to this reduction.

Discussion

Experimental considerations This study was based on the analysis of an experimentally induced peak discharge of suspended particles into a constructed wetland. Kruit et al. (2009) identified suspended particles as the major component of accidental peak discharges, but comprehensive information on particle peak discharges is currently lacking. The experimental approach used in study allowed us to make detailed observations on the behaviour of the suspended particles during retention in the CW. It can be expected that suspended particles discharged during actual WWTP malfunctioning are similar in composition (e.g. OM, metals, sand, micro-organisms) compared to suspended particles discharged during normal operating conditions. During heavy rain fall, however, WWTPs additionally receive relative high quantities of urban runoff, originating from roads, roofs, vehicles and gardens, containing relative high levels of heavy metals, oil and different micro-organisms of which the majority is associated with suspended particles (Milukaite et al. 2010; Davies and Bavor 2000; Makepeace et al. 1995). This specific type of peak discharge is, however, not addressed in this study. The large loss of lithium during retention in the unvegetated ponds was probably caused by incomplete mixing of the tracer with the WWTP effluent during the initial discharge. Mixing of the tracer prior to discharge in the CW (i.e. in the settling tank of the WWTP) may reduce the change of incomplete mixing. The minor loss of lithium during retention in the reed beds indicated however that lithium behaved as a conservative tracer and can be used as a reference for the peak behaviour of dissolved substances and showed to

116

Chapter 6

be effective to determine the time of sampling at each sampling point. To generate a reference peak more specific for the behaviour of suspended particles future experiments should incorporate an artificial conservative particulate tracer like fluorescent beats (Leibundgut et al. 2009). The calculation of the peak load of each parameter was based on relative few measurements during and around the peak discharge (Leibundgut et al. 2009), but the large differences between the peak heights in relation to the background values in this study made reliable estimations of the buffering capacities possible. In general the method of generating a peak discharge was successful and enables the possibility for controlled inducing peak discharges of desirable suspended particle concentrations, hydraulic loading and duration, and showed to be a good tool to 1) test the buffering of suspended particles and faecal indicator organisms during peak discharge by a CW, and 2) gain insight into the underlying purification mechanisms in the different functional compartments of a CW.

Peak behaviour: Mixing and diffusion During retention in the CW the changes in height and width of the lithium peak indicates the occurrence of mixing, diffusion and dispersion. In addition, tailing effect indicates the presence of dead zones and preference streams in the CW (Leibundgut et al. 2009). These observations are in agreement with the large differences observed between the theoretical HRT (when calculated from the volume of the CW and the hydraulic loading), and the HRT determined using the lithium tracer. The presence of dead zones could positively affect the buffering of peak discharges by promoting lateral mixing and delaying the movement of small water volumes to move through the CW. On the other hand the dead zones reduce the effective volume of the compartments, thereby increasing the water velocity and consequently decreasing sedimentation rates and (interaction) time in the CW.The decrease in peak maximum and loading of most parameters measured in this study was stronger in comparison with lithium, suggesting that the buffering of these parameters was only partially caused by mixing and diffusion and indicates the involvement of other processes.

Peak behaviour: Particles The strong buffering of the suspended particle peak during retention in the ponds was most likely primarily caused by sedimentation which has previously been identified as an important process for suspended particle removal in CWs (Sundaravadivel and Vigneswaran 2001; Kadlec and Wallace 2008). Since sedimentation is dependent on the size, density and shape of suspended particles (Droppo et al. 1997; Dietrich 1982), the composition of the suspended particles influences the suspended particle peak behaviour. By using suspended particles from the deep layers of the WWTP settlement tank and mixing them with water from the top of the settlement tank it can be expected that the average particle size in the

Peak discharge

117

suspended particle peak was larger compared to the suspended particles discharged during normal operating condition. In addition, the increased concentrations of suspended particles have caused increased rates of flocculation in the unvegetated ponds and consequentially increase the average size and weight of the suspended particles (Droppo et al. 1997; Chang et al. 2006). Large particles are subjected to higher sedimentation rates (Droppo et al. 1997), consequentially increasing the importance of sedimentation in the removal of suspended particles during the peak discharge compared to normal situations. Besides sedimentation, increases in NH4, PO4 and IC concentrations at the outflow of the ponds indicate partial mineralization of the organic particles (Bridgham et al. 1998). However, the importance of mineralisation for the peak buffering is still unclear, because it was not possible to determine if the mineralisation occurred in the pelagic zone or in the benthic zone after sedimentation. Although no direct evidence is provided by this experiment, defragmentation into dissolved substances (Droppo et al. 1997) and ingestion by zooplankton (Kadlec and Wallace 2008; Decamp and Warren 1998) are other processes which can be expected to influence the buffering of suspended particle peak discharges during retention in the unvegetated ponds, but are probably of minor importance in comparison to sedimentation (Sundaravadivel and Vigneswaran 2001; Kadlec and Wallace 2008). The small remainder of the suspended particle peak (<1%) at the outflow of the ponds was buffered during retention in the reed beds. In these reed beds, the strong interaction between water and solid surfaces (reed stems and sediment), could cause retainment of suspended particles by e.g. biofilms growing on these surfaces (Stott and Tanner 2005; Eisenmann et al. 2001; Polprasert et al. 1998) thereby stimulating additional processes in the buffering of peak discharges.

Peak behaviour: Faecal indicator organisms The buffering of the E. coli and Enterococci peaks during retention in the unvegetated ponds was lower compared to the buffering of other suspended particles, indicating lower sedimentation rates. This observation is supported by previous research demonstrating that 50-90% of E. coli in domestic and dairy wastewater is “free floating” or associated with small particles (<5µm) and that removal of E. coli in CWs by sedimentation is of minor importance (Davies and Bavor 2000; Boutilier et al. 2009). Although it is previously shown that the removal efficiency of faecal indicator organisms is positively correlated with inflow concentration (Vymazal 2005b), the removal efficiency of the E. coli during our experimentally induced peak discharge was not higher than removal efficiencies under normal operating conditions with lower loads. A possible explanation for this inconsistency could be a decrease in predation pressure on E. coli by zooplankton (previously recognized as an important pathogen removal process in CWs (Kadlec and Wallace 2008; Brookes et al. 2004) by an increase in additional food sources for the zooplankton species provided by the suspended particle peak. Reduced mortality or even regrowth of E. coli are other possible mechanisms (Kadlec and Wallace 2008). For Enterococci¸ however, removal efficiency during retention in the unvegetated

118

Chapter 6

ponds was higher in comparison to the background removal efficiency, indicating that the buffering of peak discharges of pathogens is a species specific process. Differences in the hydrology and subsequent effects on sedimentation and mixing processes, could partially explain the large difference in E. coli removal efficiency between unvegetated ponds and reed beds. Another difference between the unvegetated ponds and reed beds is the presence of large surfaces on the reed stems that can be colonized by biofilms that are known for their capacity to retain pathogens (Stott and Tanner 2005; Balzer et al. 2010; Flood and Ashbolt 1999). Trapping of pathogens by biofilms is determined by the physical characteristics (size, shape, charge) of the pathogens (Eisenmann et al. 2001), indicating a pathogen specific retainment by biofilm. Although the buffering of E. coli was strong, concentrations at the outflow of the CW were still slightly increased by the peak discharge and elevated E. coli concentrations were discharge into the receiving surface water, posing a minor temporal increased health risk.

Conclusion The method used in this study successfully induced a real-time peak discharge, temporally increasing the suspended particle concentration and associated parameters, without affecting physicochemical parameters and dissolved nutrient concentrations. Difference between the relative low buffering of the faecal indicator organism peaks and strong buffering of the other suspended particles peak indicated differences in importance of individual removal processes between suspended particle types. Suspended particles were probably mostly removed by sedimentation and mineralization, where pathogens were more likely buffered by biofilm retainment, mortality and predation. Differences between the buffering of the E. coli and Enterococci peaks further indicate particle specific buffering of peak discharges. In general it has been shown that while maintaining a constant hydraulic loading surface flow constructed wetlands can strongly buffer accidental suspended particle peaks, and can henceforth be used as a tool to strongly reduce the impact of accidental suspended particle peaks to receiving surface waters.

Acknowledgments This work was financed by the Foundation for Applied Water Research (STOWA) and supported by Witteveen+Bos, stichting Waternet and Wetterskip Fryslân. Special thanks go out to supporting personnel, Hans van Nieuwenhuijzen, Michel Collin and Peter wind, Rinse van der Kooij, Marieke Soeter and Hilde Terlouw.

Peak discharge

119

References

Balzer M, Witt N, Flemming H-C and Wingender J (2010). Faecal indicator bacteria in river biofilms. Water Science and Technology 61 (5), 1105-1111.

Boutilier L, Jamieson R, Gordon R, Lake C and Hart W (2009). Adsorption, sedimentation, and inactivation of E. coli within wastewater treatment wetlands. Water Research 43 (17), 4370-4380.

Bridgham SD, Updegraff K and Pastor J (1998). Carbon, nitrogen, and phosphorus mineralization in northern wetlands. Ecology 79 (5), 1545-1561.

Brookes JD, Antenucci J, Hipsey M, Burch MD, Ashbolt NJ and Ferguson C (2004). Fate and transport of pathogens in lakes and reservoirs. Environment International 30 (5), 741-759.

Cai K, Elliott CT, Phillips DH, Scippo M, Muller M and Connolly L (2012). Treatment of estrogens and androgens in dairy wastewater by a constructed wetland system. Water research 46 (7), 2333-2343.

Cameron K, Madramootoo C, Crolla A and Kinsley C (2003). Pollutant removal from municipal sewage lagoon effluents with a free-surface wetland. Water Research 37 (12), 2803-2812.

Chang TS, Joerdel O, Flemming BW and Bartholomae A (2006). The role of particle aggregation/disaggregation in muddy sediment dynamics and seasonal sediment turnover in a back-barrier tidal basin, East Frisian Wadden Sea, southern North Sea. Marine Geology 235 (1-4), 49-61.

Davies CM and Bavor HJ (2000). The fate of stormwater-associated bacteria in constructed wetland and water pollution control pond systems. Journal of Applied Microbiology 89 (2), 349-360.

Decamp O and Warren A (1998). Bacterivory in ciliates isolated from constructed wetlands (reed beds) used for wastewater treatment. Water Research 32 (7), 1989-1996.

Dietrich WE (1982). Settling velocity of natural particles. Water Resources Research 18 (6), 1615-1626.

Droppo IG, Leppard GG, Flannigan DT and Liss SN (1997). The freshwater floc: A functional relationship of water and organic and inorganic floc constituents affecting suspended sediment properties. Water, Air, and Soil Pollution 99 (1-4), 43-53.


120

Chapter 6

Fisher J, Stratford CJ and Buckton S (2009). Variation in nutrient removal in three wetland blocks in relation to vegetation composition, inflow nutrient concentration and hydraulic loading. Ecological engineering 35 (10), 1387-1394.

Fisher J and Acreman MC (2004). Wetland nutrient removal: a review of the evidence. Hydrology and Earth System Sciences 8 (4), 673-685.

Flood JA and Ashbolt NJ (1999). Virus-sized particles can be entrapped and concentrated one hundred fold within wetland biofilms (Reprinted from Advances in Environmental Research, (3), 403-411, (2000). Advances in Environmental Research 3 (4).

Foekema EM (Ed) (2012). The influence of constructed wetlands on the quality of WWTP effluent and recommendations for improvements. IMARES rapport C005/12.

Gray JL and Sedlak DL (2005). The fate of estrogenic hormones in an engineered treatment wetland with dense macrophytes. Water Environment Research 77 (1), 24-31.


Kruit J, van Leusden M and Uijterlinde C (2009). Prevention of WWTP sludge washout events demands good dry matter management. H2O (10), 44-48.

Leibundgut C, Maloszewski P and Külls C (2009). Tracers in Hydrology. Wiley-Blackwell, Chichester.

Makepeace DK, Smith DW and Stanley SJ (1995). Urban stormwater quality- summary of contaminant data. Critical Reviews in Environmental Science and Technology 25 (2), 93-139.

Milukaite A, Sakalys J, Kvietkus K, Vosyliene MZ, Kazlauskiene N and Karlaviciene V (2010). Physico-Chemical and Ecotoxicological Characterizations of Urban Storm Water Runoff. Polish Journal of Environmental Studies 19 (6), 1279-1285.

NEN-6520 (2006). Water - Spectrophotometric determination of chlorophyll-a content.

NEN-6623 (2005). Water - Determination of the amount of settleable solids.

NEN-6633 (2006). Water and sludge - Determination of chemical oxygen demand (COD).

NEN-6643 (2003). Water - Determination of nitrogen content after UV destruction by continuous flow analysis (CFA) and spectrometric detection.

Peak discharge

121

NEN-6646 (2006). Water - Photometric determination of the content of ammonium nitrogen and of the sum of the contents of ammoniacal and organically bound nitrogen according to Kjeldahl by continuous flow analysis.

NEN-EN-1899-1 (1998). Water quality - Determination of biochemical oxygen demand after n days (BODn) - Part 1: Dilution and seeding method with allylthiourea addition.

NEN-EN-872 (2005). Water quality - Determination of suspended solids - Method by filtration through glass fibre filters.

NEN-EN-ISO-13395 (1997). Water quality - Determination of nitrate nitrogen and nitrite nitrogen and the sum of both by flow analysis (CFA en FIA) and spectrometric detection.

NEN-EN-ISO-15681-2 (2005). Water quality - Determination of orthophosphate and total phosphorus contents by flow analysis (FIA and CFA) - Part 2: Method by continuous flow analysis (CFA).

NEN-EN-ISO-7899-2 (2000). Water quality - Detection and enumeration of intestinal enterococci - Part 2: Membrane filtration method.


Polprasert C, Khatiwada NR and Bhurtel J (1998). A model for organic matter removal in free water surface constructed wetlands. Water Science and Technology 38 (1), 369-377.


Shappell NW, Billey LO, Forbes D, Matheny TA, Poach ME, Reddy GB and Hunt PG (2007). Estrogenic activity and steroid hormones in swine wastewater through a lagoon constructed-wetland system. Environmental science & technology 41 (2), 444-450.

Song H, Yang X, Nakano K, Nomura M, Nishimura O and Li X (2011). Elimination of estrogens and estrogenic activity from sewage treatment works effluents in subsurface and surface flow constructed wetlands. International journal of environmental analytical chemistry 91 (7-8), 600-614.


122

Chapter 6


Toet S, Van Logtestijn RSP, Kampf R, Schreijer M, Verhoeven JTA (2005). The effect of hydraulic retention time on the removal of pollutants from sewage treatment plant effluent in a surface-flow wetland system. Wetlands 25 (2), 375-391.

van den Boomen R and Kampf R (2012). Waterharmonica’s in the Netherlands 1996-2011: from WWTP effluent till usable surface water. ISBN.978.90.5773.559.2 STOWA 2012-12

van den Boomen R, Kampf R and Mulling BTM (2012). Research on suspended particles and pathogens in the Waterharmonica (constructed wetland). ISBN.978.90.5773.553.0 STOWA 2012-10

Vidales-Contreras JA, Gerba CP, Karpiscak MM, Acuna-Askar K and Chaidez-Quiroz C (2006). Transport of coliphage PRD1 in a surface flow constructed wetland. Water Environment Research 78 (11), 2253-2260.

Vymazal J (2007). Removal of nutrients in various types of constructed wetlands. Science of the Total Environment 380 (1-3), 48-65.

Vymazal J (2005a). Constructed wetlands for wastewater treatment. Ecological Engineering 25 (5), 475-477.

Vymazal J (2005b). Removal of enteric bacteria in constructed treatment wetlands with emergent macrophytes: A review. Journal of Environmental Science and Health, Part A- Toxic/Hazardous Substances and Environmental Engineering 40 (6-7), 1355-1367.

Vymazal J (1996). Constructed wetlands for wastewater treatment in the Czech Republic the first 5 years experience. Water Science and Technology 34 (11), 159-164.

Zhang L, Xia X, Zhao Y, Xi B, Yan Y, Guo X, Xiong Y and Zhan J (2011). The ammonium nitrogen oxidation process in horizontal subsurface flow constructed wetlands. Ecological Engineering 37 (11), 1614-1619.

Zurita F, De Anda J and Belmont MA (2009). Treatment of domestic wastewater and production of commercial flowers in vertical and horizontal subsurface-flow constructed wetlands. Ecological Engineering 35 (5), 861-869.

Concluding remarks

Chapter 7

126

Chapter 7

This thesis has shown that constructed wetlands (CWs) are capable of transforming the nature and composition of anthropogenic suspended particles discharged by a wastewater treatment plant. It was demonstrated that CWs buffer the input of particles and that a short residence time is sufficient to alter virtually all chemical, physical and biological characteristics of the particles. In the next paragraphs I will firstly want to focus on the identification of the nature and composition of suspended particles during residence in CWs. Many different approaches have been developed to analyse the fate of particles and based on the findings in this thesis I will try to link up these different approaches that were originally developed for different purposes. Consequently an attempt will be made to weigh the importance of individual environmental processes in wetland ecosystems that determine the fate, behaviour and dynamics of suspended particles and present an operational model taking earlier models into account. The model provides a detailed overview on suspended particle constituents and processes influencing the concentration and composition of particles in wetland ecosystems. Thirdly, the present study demonstrated the ability of CWs to clear treated wastewater from faecal indicator organisms, re-establish a microbial community similar to that in ‘natural’ surface water. Based on these observations, this chapter will be concluded with a review of the options for applying CWs for safeguarding the high quality standards such as imposed by the European Water Framework Directive, swimming water regulations or other water quality regulations for urban and natural landscapes, followed by the general conclusions of the thesis.

Suspended particles: an important topic for all water experts The term suspended particles is unspecific as it lumps a spectrum of different particles. Suspended particles include a large variety of inorganic and organic particles that differ in nature and composition with numerous functions and roles in aquatic ecosystems. Taking into account the linkage between suspended particles and many fundamental processes that determine the functioning of aquatic ecosystems, it is not surprising that suspended particles are of interest to experts from many different professional disciplines including science (ecologists, biogeochemists), technology (wastewater managers, drinking water producers) and government (policy makers, surface water managers). Although these experts share an interest in suspended particles, they also have a different notion of the definitions, characteristics, origins, functioning and dynamics of these particles, leading to differences in the approach and methodologies that they select to investigate these particles in aquatic systems. The general approach of governmental institutes and water authorities is based on a set of defined rules regarding water quality indicators and use a subset of parameters which can be used to assess the trophic state of aquatic ecosystems (phytoplankton community, organic matter content; Dodds et al. 2002) and public health risks (abundance of pathogenic organisms and indicator organisms; Directive 2006/7/EC). The approach of technological disciplines is often focused on assessing the level of contamination in drinking and wastewater including organic matter,

Concluding remarks

127

inorganic matter, nutrients and pathogenic organisms (e.g. bacteria, viruses, parasites) in order to develop and optimize methods that are used to ensure drinking water quality or mitigate effects of wastewater discharges into receiving surface waters (Kadlec and Wallace 2008; Tchobanoglous et al. 2004). In contrast to the approaches that are based on target concentrations for defined types of suspended particles as described above, scientists often select a specific set of parameters for characterizing particles that are linked to one specific research interest. These scientific approaches include the analysis of community dynamics and food web interactions of plankton, quantifying energy, nutrient and mass fluxes and determining ecosystem functioning (Kalff 2002). All the approaches described above provide important and useful information about suspended particles. However, it is essential to cross the disciplinary borders and to link the observations from these studies to understand the complex dynamics of suspended particles in aquatic ecosystems. By combining standard particle measurements with more detailed particle characterization including various microscopic identifications, chemical analyses, microbial techniques and metabolic assessments, chapters 2, 3 and 4 have shown that removal of particles occurs in CWs receiving low particle inflow concentration, but removal is counterbalanced by the generation of other particles. No net removal of particles takes place, but the nature and composition of the particles is transformed. Additional analyses have provided information on the relative importance of individual processes inside CWs responsible for the removal and transformation of particles (chapter 2 to 6).

Suspended particle dynamics in wetlands ecosystems: an operational model Many conceptual models have been developed to analyze the fluxes of solutes, particles and pathogens in wetlands. The removal of dissolved nitrogen (Kadlec 2012; Kadlec and Wallace 2008) and phosphorus (Walker and Kadlec 2011; Kadlec and Wallace 2008) has been conceived as a dynamic process driven by residence time and biological transformation. Hogan et al. (2012) and Shen et al. (2008) have analysed the decay rate of viral and bacterial pathogens in wetlands, while Chrysikopoulos et al. (2010) and Johnson et al. (1994) quantified the transformation and sedimentation process of biocolloids in wetlands. Kadlec and Wallace (2008) describe the major removal processes of suspended particles in CWs and include processes like sedimentation, resuspension, chemical precipitation and the influence of macrophytes. Another conceptual model by Johnson et al. (1994) focusses especially on the physical interaction of particles in aquatic environments, including processes like sedimentation, aggregation, precipitation and diffusion, but largely excluded interactions with sediments and macrophytes in the pilot wetlands. Most of the studies mentioned above are dedicated to components and processes of interests and have not been designed to describe local ecosystems in a comprehensive way. This is in contrast to the ecological model PC Lake for lakes which includes sediments and riparian zones, using carbon as a key currency and other elements such as phosphate as steering factors (Janse 2005). The latter model has explicit state variables expressed in carbon units in the same way as most

128

Chapter 7

ecosystem models. On the other hand, the organic components of suspended particles in food webs structures and biomass fluxes have not been specified and the co-action and joint transformation of organic and inorganic fractions of suspended particles have been ignored. The diversity of models describing wetland processes indicates the difficulty of defining an operational basis for models encompassing the many aspects of particles in (constructed) wetlands ecosystems. In this paragraph I will attempt to combine findings in the present thesis with elements of existing models to construct an operational model for suspended particles and the physical, chemical and biological processes affecting particle dynamics in wetland ecosystems (Fig. 7.1; Table 7.1). The identifiable groups of suspended particles used in this operational model include inorganic particles, dead particulate organic matter (POM), phytoplankton, heterotrophic bacteria, viruses and zooplankton. The multitude of processes and interactions between these particle groups, dissolved substances and the wetland benthos impairs the use of a general currency in the operational model, although large sections of the operational model could be expressed in (bio)mass or carbon. Overall this operational model is meant to show the variety of suspended particles in both constructed and natural wetlands, and illustrate the complexity and interdependence of processes influencing suspended particle dynamics in wetland ecosystems. Table 7.1 supports the operational model by labeling the individual processes and summing up the main drivers of these processes. Within the presented operational model three groups of processes can be distinguished 1) processes operating from outside the wetland (Fig. 7.1; # in white circles), 2) interactions between particles, the wetland and dissolved compounds in the water (Fig. 7.1; # in grey circles) and 3) processes involved in interactions between different groups of particles (Fig. 7.1; # in black circles). The first group of processes includes the in- and out-flow of particles (Fig. 7.1 #1, 2), external import of particles from the atmosphere or surrounding land surfaces (Fig. 7.1. #5) and export of matter from the ecosystem (Fig. 7.1 #6, 7). The importance of the in- and out-flow (Fig. 7.1 #1, 2) is dependent on amount of inflowing water relative to the total volume of the wetland and the particle concentration in the inflowing water (Table 7.1). Although most import processes often occur irregularly or even incidental (Fig. 7.1 #5), import of particles can occur in several different ways and include atmospheric decomposition, defecation by waterfowl and other higher organisms, and runoff from adjacent land surfaces (Table 7.1). Within wetlands sedimentation and resuspension of particles are known as major processes in particle dynamics (Fig. 7.1 #9, 10; Kadlec and Wallace 2008; Spray and McGlothlin 2004) and the effects of these processes on individual particles is dependent on hydrodynamics of the water (velocity, turbulence, viscosity) and particle characteristics (size, shape, mass, density) (Wotton et al. 1994; Table 7.1). Particles that are normally not subjected to sedimentation can form aggregates with other particles, effectively increasing the size and sedimentation rate (Wotton 1994; Droppo et al. 1997). This flocculation process is dependent on the characteristics and concentration of the particles, were high particle

Concluding remarks

129

concentrations increases the occurrence of flocculation of particles (Fig. 7.1 #11; Table 7.1). After sedimentation, particles can accumulate in the sediments, serve as food for benthic macro-invertebrates and heterotrophic bacteria (Fig. 7.1) or may be resuspended (Fig. 7.1 #9). Benthic community processes including ingestion, excretion and decomposition (Fig. 7.1 #13-16) of particles change the composition and properties of particles and can influence the susceptibility of particles to resuspension (Fig. 7.1 #9). Other processes that can lower particle concentrations in wetlands include deflocculation of aggregates (Fig. 7.1 #12), excretion of dissolved substances by plankton (Fig. 7.1 #13), respiration by plankton (Fig. 7.1 #28), ingestion by higher organisms (Fig. 7.1 #26) and trapping of particles by biofilms (Fig. 7.1 #24; chapter 5). On the other hand addition of particles can occur by growth of the phyto- and bacterio-plankton population (Fig. 7.1 #18, 21), litterfall of macrophytes (Fig. 7.1 #20), precipitation of dissolved substances (Fig. 7.1 #25), excretion by higher trophic organisms (Fig. 7.1 #22), detachment of biofilms (Fig. 7.1 #23) and erosion of wetland shores. The importance of these processes can differ between wetlands and have substantial temporal variations, especially the biological processes (Table 7.1; Kadlec and Wallace 2008). Interactions between groups of particles often result in transformation of particles with no direct effect on particle concentrations (Fig. 7.1; # in black circles). These processes are mostly related to food web interactions between different groups of planktonic organisms and particular organic matter and include excretion and defecation (Fig. 7.1 #29), ingestion (Fig. 7.1 #30; chapter 4), infection (Fig. 7.1 #31) and decomposition (Fig. 7.1 #32). As most of these processes are biological, substantial temporal variations can also be expected for these processes (Table 7.1; Kadlec and Wallace 2008).

Application of the operational model to surface flow constructed wetlands The operational model illustrates the multitude of processes that affect the concentration and composition of particles during residence in CWs, but the relative importance of these individual processes is likely to differ between wetlands. Commonly the inflow of particles is the main driver of particles dynamics in CWs used for polishing treated wastewater (Fig. 7.1 #1). The discharge of particles from wastewater treatment plants into CWs is generally both high and continuous which prevents the establishment of a stable particle composition shaped by the wetland ecosystems (Fig. 7.1). The generally high flow rate in CWs therefore drives a continuous transformation process in the particle composition by the inflow and removal of particles and the addition and consequent outflow of particles generated within the wetland ecosystem (Fig. 7.1 #2; chapter 2, 3, 4). The present study has shown that even with a relative short hydraulic residence time (42 h) the combination of these removal and addition processes can cause a strong turnover in the particle composition (chapter 2-5). It has been shown that sedimentation is a major process driving particle removal in CWs (Ghermandi et al. 2007; Kadlec and Wallace 2008). However this is only the case in CWs receiving relative high inflow concentrations of particles, whereas CWs

130

Chapter 7

Susp

ende

dpa

rticles

DO

M

ICCO

2

depo

sits

Inor

gani

c pa

rtic

les

Phyt

o-pl

ankt

on

Bact

erio

-pl

ankt

on

Zoo-

plan

kton

POM

Viru

ses

Larg

eco

nsum

ers

Mac

roph

ytes

Biof

ilm

Bent

hic

cons

umer

sHe

tero

trop

hic

bact

eria

12

34

56

7

910

13

11

1514

1716

18

198

25

20

20

21

2324

2226

27

2613

13

28

22

28

28

2928

2929

29

30

30

30

3031

3131

3233

21

12

22

2728

2719

22

Fig.

7.1

Ope

ratio

nal m

odel

of s

uspe

nded

par

ticle

dyn

amic

s in

wet

land

eco

syst

ems.

Arr

ows

are

expl

aine

d in

Tab

le 7

.1, r

epre

sent

pro

cess

es o

pera

ting

from

ou

tsid

e th

e w

etla

nd (w

hite

circ

les;

#1-8

), in

tera

ctio

ns b

etw

een

part

icle

s an

d w

etla

nd a

nd d

issol

ved

com

poun

ds in

the

wat

er (g

rey

circ

les;

#9-2

8) a

nd

proc

esse

s inv

olve

d in

inte

ract

ions

bet

wee

n di

ffere

nt g

roup

s of p

artic

les (

in b

lack

circ

les;

#29-

33).A

bbre

viat

ions

use

d in

the fi

gure

; car

bon

diox

ide (

CO

2), in

orga

nic c

arbo

n (I

C),

diss

olve

d or

gani

c car

bon

(DO

M) a

nd p

artic

ular

org

anic

car

bon

(PO

M).

Concluding remarks

131

#Pr

oces

sD

river

s

1In

flow

of p

artic

les

Wat

er fl

ow; p

artic

le co

ncen

trat

ion

2O

utflo

w o

f par

ticle

sW

ater

flow

; par

ticle

conc

entr

atio

n

3G

as d

iffus

ion

Resp

iratio

n; at

mos

pher

ic C

O2 co

ncen

trat

ions

; pH

4G

as d

iffus

ion

Resp

iratio

n; d

issol

ved

CO

2 conc

entr

atio

ns; p

H

5Ex

tern

al in

put

Run

off; d

efec

atio

n; at

mos

pher

ic d

epos

ition

6Pr

edat

ion;

har

vest

ing

Pred

ator

, pre

y po

pula

tion

dens

ities

; man

agem

ent

7G

razi

ng; h

arve

stin

gG

raze

r den

sitie

s; m

anag

emen

t

8U

ptak

e; p

hoto

synt

hesis

; m

acro

phyt

e gr

owth

Irra

dian

ce; t

empe

ratu

re

9Se

dim

enta

tion

Part

icle

char

acte

ristic

s; tu

rbul

ence

10Re

susp

ensio

nPa

rtic

le ch

arac

teris

tics;

turb

ulen

ce

11Fl

occu

latio

n; a

ggre

gatio

n;

abso

rptio

nPa

rtic

le co

ncen

trat

ion;

par

ticle

char

acte

ristic

s; bi

olog

ical

act

ivity

12D

efloc

cula

ting;

disa

ggre

gatio

n;

deso

rptio

nPa

rtic

le co

ncen

trat

ion;

par

ticle

char

acte

ristic

s; bi

olog

ical

act

ivity

13Ex

cret

ion;

def

ecat

ion

Reso

urce

qua

lity

and

quan

tity;

bio

logi

cal a

ctiv

ity;

tem

pera

ture

; oxy

gen

conc

entr

atio

n

14D

ecom

posit

ion;

ass

imila

tion

Reso

urce

qua

lity

and

quan

tity;

bac

teria

l act

ivity

; te

mpe

ratu

re; o

xyge

n co

ncen

trat

ion

15Ex

cret

ion;

mor

talit

yRe

sour

ce q

ualit

y an

d qu

antit

y; b

acte

rial a

ctiv

ity;

tem

pera

ture

; oxy

gen

conc

entr

atio

n

16In

gest

ion;

ass

imila

tion

Reso

urce

qua

lity

and

quan

tity;

bio

logi

cal a

ctiv

ity;

tem

pera

ture

; oxy

gen

conc

entr

atio

n

17Ex

cret

ion;

mor

talit

yRe

sour

ce q

ualit

y an

d qu

antit

y; b

iolo

gica

l act

ivity

; te

mpe

ratu

re; o

xyge

n co

ncen

trat

ion

#Pr

oces

sD

river

s

18U

ptak

e; p

hoto

synt

hesis

; ph

ytop

lank

ton

grow

thIr

radi

ance

; wat

er cl

arity

; nut

rient

avai

labi

lity

19U

ptak

e; p

hoto

synt

hesis

; m

acro

phyt

e gr

owth

irrad

ianc

e; w

ater

clar

ity; n

utrie

nt av

aila

bilit

y

20M

acro

phyt

e lit

ter

Plan

t den

sity;

tem

pera

ture

; irr

adia

nce;

gra

zing

; ha

rves

ting

21U

ptak

e; a

ssim

ilatio

nSu

bstr

ate

qual

ity; b

acte

rial d

ensit

y, te

mpe

ratu

re

22Ex

cret

ion;

def

ecat

ion

Con

sum

er d

ensit

y; re

sour

ce q

uant

ity; b

iolo

gica

l ac

tivity

; pre

datio

n; h

arve

stin

g

23Er

osio

n; d

isper

sal

Wat

er v

eloc

ity; p

opul

atio

n gr

owth

24Tr

appi

ng; c

olon

izat

ion

Biofi

lm co

mpo

sitio

n; p

artic

le co

ncen

trat

ion

and

char

acte

ristic

s; re

lativ

e su

rfac

e ar

ea

25Pr

ecip

itatio

n; a

bsor

ptio

nSo

lute

conc

entr

atio

n; p

H, d

issol

ved

oxyg

en;

cond

uctiv

ity

26In

gest

ion;

ass

imila

tion

Pred

ator

den

sity;

reso

urce

qua

lity

and

quan

tity;

ox

ygen

conc

entr

atio

n; h

arve

stin

g in

tens

ity

27G

razi

ng; i

nges

tion;

ass

imila

tion

Gra

zer d

ensit

y; re

sour

ce q

ualit

y an

d qu

antit

y;

oxyg

en co

ncen

trat

ion;

tem

pera

ture

28Re

spira

tion

Biol

ogic

al a

ctiv

ity; t

empe

ratu

re; o

xyge

n co

ncen

trat

ion

29Ex

cret

ion;

def

ecat

ion;

mor

talit

yRe

sour

ce q

ualit

y an

d qu

antit

y; b

iolo

gica

l act

ivity

; te

mpe

ratu

re; o

xyge

n co

ncen

trat

ion

30In

gest

ion

Reso

urce

qua

lity

and

quan

tity;

bio

logi

cal a

ctiv

ity;

tem

pera

ture

; oxy

gen

conc

entr

atio

n

31In

fect

ion;

ass

imila

tion

Hos

t den

sity;

tem

pera

ture

32D

ecom

posit

ion;

ass

imila

tion

Reso

urce

qua

lity

and

quan

tity;

bio

logi

cal a

ctiv

ity;

tem

pera

ture

; oxy

gen

conc

entr

atio

n

33Fo

rmat

ion

of b

iolo

gica

l in

orga

nics

Reso

urce

qua

lity

and

quan

tity;

bio

logi

cal a

ctiv

ity;

tem

pera

ture

; oxy

gen

conc

entr

atio

n

Tabl

e 7.1

Pro

cess

es in

fluen

cing

the

susp

ende

d pa

rtic

les d

ynam

ics i

n w

etla

nd e

cosy

stem

s. N

umbe

rs co

rres

pond

with

arr

ows p

rese

nted

in F

ig. 7

.1.

132

Chapter 7

receiving low particle concentrations generally show no net removal or even addition of particles (Ghermandi et al. 2007; van den Boomen and Kampf 2012). The CW investigated in the present study generally receives low concentrations of particles (±3.6 mg L-1; chapter 2) and under those conditions sedimentation of particles accounted for the net 10% removal of inflowing particles (chapter 2; Fig. 7.1 #9). This modest importance of sedimentation of particle composition under normal situations is however highly increased during elevated inflowing concentration of particles (±230 mg L-1; chapter 6). Under these high particle inflow conditions, sedimentation of more than 99% of the inflowing particles was observed (chapter 6). The dominance of sedimentation of particles was probably enhanced by substantial aggregate formation (Fig. 7.1 #11), as large flocs were seen in the CWs after a peak discharge (chapter 6; unpublished observation). Although the net removal of faecal indicator organisms was also shown to increase (chapter 6), compared with particles the relative removal of faecal indicator organisms was lower. This indicates sedimentation to be a selective process and maybe of less importance for hygienic water quality improvements compared to the removal of particles in general (chapter 4, 6). This was confirmed in chapter 4 where it was shown that E. coli cells were primarily “free living”, not associated with large particles, and no sedimentation of E. coli cells was observed. Nevertheless concentrations of E.coli (97%) and several other faecal indicator organisms (73-99%) strongly declined during transport through the wetland (chapter 4). Organic and inorganic components of particulate matter was shown to develop slightly different in the wetlands studied (chapter 2). Particular organic matter (POM) was subjected to sedimentation indicated by a shift in organic content from 53% to 33% during residence in the unvegetated ponds. In contrast, inorganic particles increased, primarily due to shore erosion and to a minor extent due to atmospheric deposition, chemical precipitation and diatom growth (chapter 2; Fig. 7.1 #5, 25). Thus, some processes (e.g. sedimentation) may be shared by organic and inorganic particles, whereas others (chemical precipitation and erosion) apply only to one component, i.e. inorganic particles. Certainly the grazing of zooplankton on suspended bacteria and organic particles covered by bacteria constitute an important decay of inflowing bacterial consortia composed of faecal bacteria and heteroptrophic bacteria from the wastewater treatment process. However, the development of zooplankton depends on the residence time of the water, providing a window of opportunity for rapidly growing species. Van den Boomen and Kampf (2012) referred to the unvegetated ponds of CWs as ‘daphnid ponds’, probably focusing on ponds that allowed the larger water fleas to reproduce at a sufficient rate to compensate for losses via the outflow. In the present study the water residence time in the ponds was so short that only smaller zooplankton could sustain dense populations that were, however demonstrated to exert a strong grazing pressure (chapter 4). Biofilms have been demonstrated in experiments to trap biotic and abiotic particles in a selective way (chapter 5; Fig. 7.1 #24). These observations accord with those by Stott and Tanner (2005) and demonstrate a robust mechanism that traps and consolidates particles in

Concluding remarks

133

CWs. On the other hand, biofilms were also shown to expel particles to overlying water in experiments (chapter 5) and epiphytic diatom species were observed in the water of the CW, suggesting exchange of particles rather than net trapping. Vegetation and especially reed have been applied as agents for retaining particles, a capacity that relies for a substantial part on the cover of biofilms on water plants and the root system of emergent plants such as reed. For that reason densely vegetated ponds have been utilized to filter effluents (Kadlec and Wallace 2008, Vymazal 1995, Sundaravadivel and Vigneswaran 2001; van den Boomen and Kampf 2012). Evidently the extent of water plant stands in the CW is a vital driver for retention of particles directly by the plant structure or via its biofilm cover. How filtering of effluents by wetlands proceeds with time is not immediately evident from Fig. 7.1. Yet this thesis has shown a conversion of particles during residence in the CW. This conversion was most evident in observations on the composition of bacterial consortia at early and late stages of the treatment. Wastewater treatment plant effluent was characterized by active bacteria probably originating from the treatment plant and faecal bacteria, while at the outflow less active consortia resembling natural surface water communities were observed (chapter 3).

The considerations discussed above indicate that the relative importance of processes in (constructed) wetlands cannot be derived immediately from the model structure (Fig. 7.1). Yet the operational model allows some predictions based on the fundamentals of its components such as the planktonic food web , the physical structure and biofilm development around water plants and the ‘laws’ ruling co-agulation, precipitation and sedimentation. Keeping these diverse mechanism in mind and considering their interactions such as presented in the operational model, it may be possible to design CWs optimized for certain applications. Whatever the component of interest is, faecal bacteria, nutrients, or toxicants, the transformation of one component is tightly bound to the interconnected processes described in Fig. 7.1 and Table 7.1. Constructed wetlands in a changing world Ever increasing urbanization of delta areas decreases the occurrence of natural wetland ecosystems by land reclamation and dredging, while canalization and hydrological changes progressively reduce riparian zones (Duha et al. 2008). At the same time the inflow of particles into aquatic ecosystems increases through runoff caused by intensified land-use and (in some areas) deforestation. Also atmospheric deposition of fine dust is increasing and the discharge or spreading of man-made particles, such as soot from combustion processes, micro-plastics from degrading polymeric materials and road run-off is likely to intensify. On the other hand the water cycle of cities is demanding a large fraction of natural waters and the reflux of used water loaded with pathogenic micro-organisms to surface water creates public health risks via recreational activities or via the preparation of drinking water (Mills

134

Chapter 7

2007; Duha et al. 2008). The combination of all these changes pose a challenge to water management that is already facing turbid surface waters with unstable sediments harboring decimated stocks of water plants and as a result non-compliance with European rules on ecological quality (Lammens et al. 2008; Kelderman et al. 2012). Thus there is a need for an advanced policy on water, incorporating the role of abiotic and biotic particles in water treatment and the management of surface water. The present thesis has shown that CWs have several capacities to process suspended particles and pathogens, which could be utilized to intercept particle and pathogen fluxes before these reach surface waters. Chapter 2 and 3 showed removal of effluent particles and replacing these by a particles resembling those in natural surface waters, chapter 4 showed strong improvements on the hygienic water quality to a level generally below the European water quality standards for swimming water (Directive 2006/7/EC), and chapter 6 illustrated the effectiveness of CWs to buffer accidental peak discharges of particles. These services are a result of the combined effect of residence in unvegetated ponds and reed beds. The unvegetated ponds were shown to have a major contribution to the buffer capacity of the CW when faced with an accidental peak discharge (chapter 6), whereas the reed beds were more important in changing the nature and composition of particles and the reduction of pathogens (chapter 2, 3, 4). CWs are evidently effective in reducing the environmental impacts of the discharge of man-made particles in wastewater into receiving surface waters. The water quality of polished treated wastewater (after residence in a CW) is generally even of better quality in comparison with receiving surface waters. In this way CWs, polishing treated wastewater, could be utilized on a larger scale. So far a few CWs treat only a small fraction of the total municipal effluents in The Netherlands. The Dutch organization of water authorities fosters the experimental application of CWs in several modifications and so far the perspective for further application seem positive. One potential obstacle for applying CWs on all municipal wastewater plants is the substantial surface area needed. Such claims could pose problems in densely populated urban areas. Some of the attributes of CWs may also be offered by natural wetlands. As mentioned before the disappearance of wetland systems is a major driver of current water turbidity problems (especially in The Netherlands) and re-establishment of riparian zones or wetlands around surface waters could restore the natural capacity of aquatic ecosystems to balance excess particles in surface waters. This option potentially provides a tool of great value in achieving good water quality standards set by the European Water Framework Directive (Directive 2000/60/EC) and the swimming water regulations (Directive 2006/7/EC).

Concluding remarks

135

Conclusions This thesis has shown that constructed wetlands transform suspended particles in (treated) municipal wastewater through selective precipitation in ponds, biological filtering by plankton communities and physical and biological retention in reed beds. These processes effectively remove faecal indicator bacteria and viruses, while in situ production generates suspended particles including bacterial consortia that closely resemble that of natural shallow water systems. This transformation is driven by a complex interaction of physical, chemical and biological processes occurring in the wetland ecosystem. Based on the observations in this thesis an operational model was constructed describing these interacting processes in relation to changes in the concentrations, nature and type of suspended particles in wetland ecosystems. This model provides a scientific basis for the design of constructed wetlands optimized to protect our natural surface waters from harmful impacts of the increasing levels of suspended particles from different sources.

136

Chapter 7

References

Chrysikopoulos CV, Masciopinto C, La Mantia R and Manariotis ID (2010). Removal of Biocolloids Suspended in Reclaimed Wastewater by Injection into a Fractured Aquifer Model. Environmental science & technology 44 (3), 971-977.

Directive 2000/60/EC of the European parliament and of the council (2000). Establishing a framework for Community action in the field of water policy.

Directive 2006/7/EC of the European Parliament and of the Council (2006). Concerning the management of bathing water quality and repealing Directive 76/160/EEC

Dodds WK (2002). Freshwater ecology: concepts and environmental applications. Academic Press, San Diego, California.

Droppo IG, Leppard GG, Flannigan DT and Liss SN (1997) The freshwater floc: A functional relationship of water and organic and inorganic floc constituents affecting suspended sediment properties. Water Air Soil Pollution 99, 43-53.


Hogan JN, Daniels ME, Watson FG, Conrad PA, Oates SC, Miller MA, Hardin D, Byrne BA, Dominik C, Melli A, Jessup DA and Miller WA (2012). Longitudinal Poisson Regression To Evaluate the Epidemiology of Cryptosporidium, Giardia, and Fecal Indicator Bacteria in Coastal California Wetlands. Applied and Environmental Microbiology 78 (10), 3606-3613.

Janse JH (2005). Model studies on the eutrophication of shallow lakes and ditches. Ph.D. Thesis, Wageningen University, ISBN 90-8504-214-3

Johnson BD, Kranck K and Muschenheim DK (1994). Physico-chemical factors in Particle Aggregation. Chapter 4 in: Wotton RS; The biology of particles in aquatic systems. CRC Press, Inc., Boca Raton, Florida.

Kadlec RH (2012). Constructed Marshes for Nitrate Removal. Critical Reviews in Environmental Science and Technology 42 (9), 934-1005.

Kadlec RH and Wallace SD (2008) Treatment wetlands. CRC Press, Florida.

Kalff J (2002) Limnology. Prentice-Hall, Englewood Cliffs, New Jersey.

Concluding remarks

137

Shen C, Phanikumar MS, Fong TT, Aslam I, McElmurry SP, Molloy SL and Rose JB (2008). Evaluating bacteriophage P22 as a tracer in a complex surface water system: The Grand River, Michigan. Environmental science & technology 42 (7), 2426-2431.

Spray S and McGlothlin KL (2004). Chapter 2: Soils and sediments. In: Spray S and McGlothlin KL, Wetlands. Rowman and Littlefield publishers Inc., Lanham, Maryland, 30-54

Sundaravadivel M and Vigneswaran S (2001). Constructed wetlands for wastewater treatment. Critical Reviews in Environmental Science and Technology 31, 351-409.


van den Boomen R and Kampf R (2012). Waterharmonica’s in Nederland 1996-2011: van effluent tot bruikbaar oppervlaktewater. STOWA 2012-12.

Vymazal J (2005). Constructed wetlands for wastewater treatment. Ecological Engeneering 25 (3), 475-477.

Walker WW jr. and Kadlec RH (2011). Modeling Phosphorus Dynamics in Everglades Wetlands and Stormwater Treatment Areas. Critical Reviews in Environmental Science and Technology 41 (1), 430-446.

Wotton RS (1994) The biology of particles in aquatic systems. CRC Press, Inc., Boca Raton, Florida.

Summary

140

Summary

Suspended matter represents a complex mixture of particles of various nature and origin (living, dead and lifeless), and plays a major role in wetland processes. Discharges of treated municipal wastewater releases considerable amounts of anthropogenic particles into surface waters, even when the majority of particles in municipal wastewater is removed by regular wastewater treatment. Constructed wetlands (CWs) are widely used for polishing treated municipal wastewater to mitigate the effects of these effluents on receiving surface waters. Most research on suspended matter in CWs is however focused on bulk measurements of suspended matter disregarding particle composition, and this obstructs analyses of the turn-over of particles during residence in CWs. Therefore, the overall aim of this study was to assess the importance of physical, chemical and biological processes modifying the fate and behaviour of suspended particles in surface flow constructed wetlands.

To this purpose the following objectives were formulated:- To identify changes in the nature and composition of suspended particles in treated municipal wastewater during residence in surface flow constructed wetlands.- To weigh the importance of individual environmental processes occurring in wetland ecosystems on the fate, behaviour and dynamics of suspended particles. - To assess the prospect of constructed wetlands in mitigating effects and associated risks of municipal wastewater.

In chapter 2 the changes in the physical and biological characteristics of suspended particles were analysed during residence in a full scale surface flow CW. We found that residence in the unvegetated ponds caused no major changes in particle concentration, but decreases the organic content (from 53% to 33%) and average particle size (from 4.3µm to 3.7µm) of the suspended particles. This was most likely caused by sedimentation of large organic particles and addition of smaller inorganic particles resulting from shore erosion. The bacterial species originating from the wastewater treatment plant (quantified by analysis of faecal indicator organisms) decreased strongly in abundance (>90% reduction), especially during residence in the reed beds. Simultaneously the total abundance of bacteria gradually increased, indicating that bacterial species present in the treated wastewater are being replaced by other species during residence in the CW. These observations indicate that CWs change the nature and type of the suspended particles.

It was demonstrated that surface flow CWs effectively reduce the numbers of faecal indicator organisms (chapter 2). However, faecal indicator organisms represent only a minor fraction of the total planktonic bacterial community. The aim of chapter 3 was therefore to identify changes in the planktonic bacterial community in treated municipal wastewater during residence in a CW. To this purpose water samples were taken at various locations in- and outside the CW and the bacterial community composition and functioning was analyzed using FISH, DGGE and BIOLOG. Surprisingly, the bacterial abundance at the inflow of the

Summary

141

CW was relatively low compared to receiving surface waters. However, the inflowing bacterial community showed high metabolic activity and functional diversity. During residence in the CW the bacterial abundance doubled, but decreased in metabolic activity and functional diversity. Shifts in the community composition revealed by DGGE indicated that these changes are related to a turn-over of the bacterial community. The planktonic bacterial community in the effluent of the constructed wetland closely resembled natural bacterial communities in urban and agricultural ditches. Based on these observations we conclude that constructed wetlands are capable to mitigate possible impacts of treated wastewaters by transforming the anthropogenic bacterial community into a bacterial community resembling more “natural” surface waters.

Although the removal efficiencies of constructed wetlands are well-studied, the importance of different processes involved in the removal of faecal indicator organisms and pathogens are often not analysed. In chapter 4 we monitored removal of bacterial, protozoan and viral faecal indicator organisms in different functional compartments of a full scale CW receiving treated municipal wastewater and, in addition, we performed small scale experiments to estimate the importance of individual processes underlying the dynamics of pathogen concentrations, including sedimentation, predation, UV irradiance, mortality and external input. The results showed substantial removal of faecal indicator organisms during residence in the constructed wetland, with total removal efficiencies ranging between 96 and 99% for bacterial faecal indicators, 89% for the faecal protozoan indicator and 73 and 88% for the viral indicators. Zooplankton predation and biofilm trapping appeared to be the most important processes governing the removal of faecal indicator organism in the constructed wetland, whereas the effects of sedimentation, inactivation and UV-radiation are of minor importance. No significant reintroduction of faecal indicator organisms by waterfowl or other warm blooded animals was observed, although waterfowl was present in very high abundance in the constructed wetland during freezing periods.

In the previous chapters we found indications that benthic organisms may play a key role in the turn-over of particles in constructed wetlands. To elaborate on these observations, in chapter 5 the particle trapping efficiency of phototrophic biofilms, present in the CW on large surface areas in the reed beds, was quantified. In a first experiment, substantial trapping of particles smaller than 15 µm by mature natural biofilms was observed, with an optimum trapping efficiency for particles in the size range 5.0 - 8.0 µm of which 49% was trapping within 150 minutes. In a second experiment natural biofilms and mono-specific Achnantes lanceolata biofilms rapidly accumulated both artificial (micro-spheres) and bacterial particles (Pseudomonas putida cells), while mono-specific biofilms composed of Nitzschia perminuta or cyanobacteria exhibited a very low trapping efficiency. Trapping of P. putida cells was up to one order of magnitude higher than the trapping of micro-spheres, which indicates selective trapping of various particles. With these experiments we showed that the trapping of particles

142

Summary

by phototrophic biofilms in wetland ecosystems can be substantial, but is strongly dependent on both species composition of the biofilms and the characteristics of the particles.

Most research on particle removal in constructed wetlands is performed under normal operating conditions, but the capacity of CWs to buffer incidental peak discharges caused by heavy rain fall or wastewater treatment plant malfunctioning is not well studied. In chapter 6 we investigated the changes in particle concentrations, (associated) physiochemical parameters and abundances of faecal indicator organisms in the different functional compartments of the CW during an peak discharge, by manually raising particle concentrations from ±3.5 mg L-1 to ±230 mg L-1 in the inflow of the CW for a duration of 8 hours. After residence in the unvegetated ponds (the first CW compartment) the particle peak was reduced by >99%. Similar buffering was observed for the turbidity, oxygen demand and settable volume. Simultaneously dissolved nutrient concentrations increased, indicating partial mineralization of the suspended particles during residence in the unvegetated ponds. The peak buffering of faecal indicator organisms was lower (40-84%), indicating differences in removal processes between faecal indicator organisms and other particles. The results indicated that large organic particles were probably mostly removed by sedimentation and mineralization, whereas smaller pathogens were mainly in removed reed ditches as a result of biofilm trapping, mortality and predation. After passing through the entire CW the residuals of the suspended particle peak discharge were temporarily increased concentrations of inorganic carbon, NH4 and Escherichia coli (respectively 11%, 17% and 160% higher than under normal operating conditions). These observations illustrate that CWs are effective systems to buffer wastewater peak discharges.

This thesis has shown that constructed wetlands transform suspended particles in (treated) municipal wastewater through selective precipitation in ponds, biological filtering by plankton communities and physical and biological retention in reed beds. These processes effectively remove faecal indicator bacteria and viruses, while in situ production generates suspended particles including bacterial consortia that closely resemble that of natural shallow water systems. This transformation is driven by a complex interaction of physical, chemical and biological processes occurring in the wetland ecosystem. Based on the observations in this thesis I have constructed an operational model describing these interacting processes in relation to changes in the concentrations, nature and type of suspended particles in wetland ecosystems (chapter 7). This model provides a scientific basis for the design of constructed wetlands optimized to protect our natural surface waters from harmful impacts of the increasing levels of suspended particles from different sources.

Samenvatting

146

Samenvatting

Zwevende stof vormt een complex mengsel van allerlei verschillende deeltjes, en speelt een belangrijke rol in processen die plaats vinden in ondiepe watersystemen. Lozingen van afvalwater brengen aanzienlijke hoeveelheden deeltjes in het oppervlaktewater die daar van nature niet thuishoren, zelfs als de grootste hoeveelheid deeltjes uit het afvalwater wordt verwijderd door reguliere afvalwaterbehandeling. Om de effecten van die lozingen te beperken worden vaak zuiveringsmoerassen gebruikt om het afvalwater verder na te zuiveren. Onderzoek naar de effectiviteit waarmee dergelijke zuiveringsmoerassen de zwevende deeltjes uit het water verwijderen is echter voornamelijk gericht op bulkmetingen van zwevende stof, waarbij de exacte samenstelling van de deeltjes buiten beschouwing wordt gelaten. Hierdoor weten we niet wat voor veranderingen er optreden in de samenstelling van zwevende stof tijdens verblijf in zuiveringsmoerassen, noch welke processen hieraan ten grondslag liggen. Het doel van dit proefschrift is daarom om vast te stellen hoe verschillende fysische, chemische en biologische processen bijdragen aan de lotgevallen van de zwevende stofdeeltjes in zuiveringsmoerassen.

Om dit doel te bereiken zijn de volgende specifieke doelen geformuleerd:- Het vaststellen van veranderingen in de aard en samenstelling van zwevende stofdeeltjes in behandeld afvalwater tijdens het verblijf in een zuiveringsmoeras.- Het schatten van de relatieve bijdrage van individuele processen in een zuiveringsmoeras op het gedrag, de dynamiek en het lot van zwevende stofdeeltjes.- Het beoordelen van de mogelijkheden om met behulp van zuiveringsmoerassen schadelijke effecten van afvalwaterlozingen te voorkomen.

In hoofdstuk 2 hebben we de veranderingen in de fysische en biologische eigenschappen van zwevende stofdeeltjes in het effluent van een rioolwaterzuiveringsinstallatie tijdens het verblijf in een zuiveringsmoeras geanalyseerd. In de onbeplante vijvers (het eerste compartiment van het zuiveringsmoeras) trad geen grote wijziging op in de concentratie van de zwevende stofdeeltjes, maar de fractie organische deeltjes werd wel minder (van 53% naar 33%) en de gemiddelde grootte van de deeltjes werd ook kleiner (van 4.3μm tot 3.7μm). Dit is werd waarschijnlijk veroorzaakt door sedimentatie van grote organische deeltjes en gelijktijdige toevoeging van kleinere anorganische deeltjes als gevolg van oevererosie. Het aantal bacteriën uit de rioolwaterzuiveringsinstallatie (bepaald aan de hand van fecale indicatororganismen) daalde sterk (>90%) in de rietvelden (het tweede compartiment van het zuiveringsmoeras). Echter, het totale aantal bacteriën steeg geleidelijk, wat aangeeft dat de bacteriën die afkomstig zijn van het gezuiverde afvalwater worden vervangen door andere bacteriën tijdens het verblijf in het zuiveringsmoeras. Deze waarnemingen geven aan dat zuiveringsmoerassen grote invloed hebben op de aard van de zwevende stofdeeltjes.

Samenvatting

147

In hoofdstuk 2 is aangetoond dat zuiveringsmoerassen het aantal fecale indicatororganismen verminderden. Echter, fecale indicatororganismen vormen slechts een kleine fractie van de totale bacteriële gemeenschap die in de waterkolom aanwezig is. Het doel van hoofdstuk 3 was dan ook om veranderingen in de bacteriële gemeenschap in behandeld afvalwater te identificeren tijdens het verblijf in een zuiveringsmoeras. Om dit te onderzoeken werden op verschillende locaties in het zuiveringsmoeras watermonsters genomen, waarin de samenstelling en het functioneren van de bacteriële gemeenschap werd geanalyseerd. Om te kijken in hoeverre deze gemeenschappen lijken op natuurlijke gemeenschappen zijn ook watermonsters genomen in een aantal oppervlaktewateren in Nederland. Verrassend is dat het aantal bacteriën bij het instroompunt van het zuiveringsmoeras relatief laag was in vergelijking met natuurlijk oppervlaktewater. Echter, de bacteriële gemeenschap bij het instroompunt vertoonde een hoge metabolische activiteit en functionele diversiteit. Tijdens het verblijf in het zuiveringsmoeras verdubbelde het aantal bacteriën, maar daalde de metabolische activiteit en de functionele diversiteit. Deze veranderingen gingen gepaard met verschuivingen in de samenstelling van de bacteriële gemeenschap. Bij de uitstroom van het zuiveringsmoeras lijkt de bacteriële gemeenschap erg op natuurlijke bacteriële gemeenschappen die voorkomen in meer natuurlijke oppervlaktewateren. Op basis van deze observaties concludeer we dat zuiveringsmoerassen in staat zijn om mogelijke risico’s van ongewenste micro-organismen in behandeld afvalwater te verminderen.

Hoewel de verwijderingsefficiëntie van zuiveringsmoerassen goed is onderzocht, wordt de relatieve bijdrage van de verschillende processen die betrokken zijn bij de verwijdering van fecale indicator organismen en ziektekiemen vaak niet geanalyseerd. In hoofdstuk 4 hebben we de verwijdering van bacteriële, protozoaire en virale fecale indicatororganismen in een zuiveringsmoeras gedurende een periode van 1 jaar gemeten. Daarnaast hebben we experimenten uitgevoerd waarin we de bijdrage van individuele processen die invloed hebben op de dynamiek van ziektekiemen, (waaronder sedimentatie, predatie, UV straling, sterfte en externe toevoeging) hebben onderzocht. In het zuiveringsmoeras vond aanzienlijke verwijdering van fecale indicatororganismen plaats, met verwijderingsefficiënties variërend tussen de 96 en 99% voor de bacteriële fecale indicatoren, 89% voor een fecale protozoaire indicator en 73 en 88% voor de virale indicatoren. Predatie door het aanwezige zoöplankton en retentie in de biofilms op rietstengels bleken de belangrijkste processen te zijn in de verwijdering van fecale indicator organismen. Sedimentatie, inactivatie en sterfte ten gevolge van UV-straling bleken weinig invloed te hebben op de verwijdering. Ook werd er geen significante herintroductie van fecale indicator organismen door watervogels of andere warmbloedige dieren waargenomen, hoewel zeer hoge aantallen watervogels aanwezig waren in het zuiveringsmoeras tijdens vorstperiodes.

148

Samenvatting

Op basis van de aanwijzingen in hoofdstuk 4 dat biofilms op de rietstengels in het zuiveringsmoeras een belangrijke rol spelen in de dynamiek van zwevende stofdeeltjes, is in hoofdstuk 5 dieper ingegaan op het invangen van verschillende deeltjes door fototrofe biofilms. In een eerste experiment werd aangetoond dat volgroeide natuurlijke biofilms aanzienlijke hoeveelheden kleine deeltjes (<15 µm) kunnen invangen. Van alle deeltjes tussen de 5 en 8 µm werd bijna de helft ingevangen binnen 150 minuten. In een tweede experiment is gekeken naar verschillen tussen verschillende biofilms en naar verschillen tussen deeltjes. Het bleek dat zowel plastic bolletjes als levende bacteriën (Pseudomonas putida cellen) snel ingevangen worden, niet alleen door natuurlijke biofilms maar ook door biofilms die slechts bestaan uit 1 soort alg. Invanging van P. putida cellen was wel tot één orde grootte sneller dan het invangen van de kunstmatige deeltjes. Ook was er verschil tussen verschillende algensoorten: Achnantes lanceolata biofilms vertoonden een grote invang capaciteit, terwijl biofilms met Nitzschia perminuta of cyanobacteriën zeer lage invanging vertoonden. Met deze experimenten is aangetoond dat het invangen van deeltjes door fototrofe biofilms in moerassystemen aanzienlijk kan zijn, maar dat dit proces sterk afhankelijk is van zowel de soortsamenstelling van de biofilms als van de eigenschappen van de deeltjes.

Het meeste onderzoek naar zwevende stofdeeltjes in zuiveringsmoerassen wordt uitgevoerd onder normale operationele omstandigheden. Echter, de capaciteit van zuiveringsmoerassen om incidentele piekbelastingen (bijvoorbeeld ten gevolge van zware regenval of storing in een afvalwaterzuiveringsinstallatie) op te vangen is niet goed onderzocht. In hoofdstuk 6 hebben we deze bufferingscapaciteit onderzocht door te kijken naar veranderingen in deeltjes concentraties, en alle daarbij behorende parameters, in een zuiveringsmoeras gedurende een piekbelasting. De piekbelasting is kunstmatig veroorzaakt door de deeltjes concentratie in het instroomde water te verhogen van ± 3,5 mg L-1 tot ± 230 mg L-1 gedurende 8 uur. Na verblijf in de onbeplante vijvers (eerste compartiment) was de deeltjespiek verlaagd met meer dan 99%. Een vergelijkbare buffering werd waargenomen voor de troebelheid, zuurstofverbruik en bezinkbaar volume. Tegelijkertijd namen de concentraties aan opgeloste nutriënten echter toe, wat verklaard kan worden door gedeeltelijke mineralisatie van de zwevende stofdeeltjes tijdens het verblijf in de onbeplante vijvers. De piek buffering van fecale indicatororganismen was lager (40-84%), wat aangeeft dat verschillende processen van belang zijn bij de verwijdering van micro-organismen en van andere deeltjes. De resultaten gaven aan dat deeltjes voornamelijk worden verwijderd door sedimentatie en mineralisatie, terwijl fecale indicatororganismen vooral worden verwijderd als gevolg van invanging door biofilms en predatie. Bij de uitstroom van het gehele zuiveringsmoeras werden nog maar geringe restanten van de opgewekte piekbelasting waargenomen in de vorm van tijdelijk verhoogde concentraties aan anorganische koolstof, ammonium en Escherichia coli (respectievelijk 11%, 17% en 160% hoger dan de normale concentraties). Dit experiment toonde aan dat zuiveringsmoerassen effectieve systemen zijn om piekbelastingen vanuit waterzuiveringsinstallaties te bufferen en daarmee oppervlaktewateren te beschermen.

Samenvatting

149

Dit proefschrift heeft laten zien dat zuiveringsmoerassen zwevende stofdeeltjes in (behandeld) afvalwater transformeren door middel van selectieve precipitatie in onbeplante vijvers, biologische filtering door plankton gemeenschappen en fysische en biologische retentie in rietvelden. Deze processen zijn effectief in het verwijderen van fecale indicator bacteriën en virussen, terwijl in situ productie ervoor zorgt dat de zwevende stofdeeltjes, grotendeels gevormd door bacteriële consortia, aan het einde van het zuiveringsmoeras sterk lijken op die van natuurlijk ondiep water systemen. Deze transformatie wordt aangedreven door een complexe interactie tussen fysische, chemische en biologische processen die zich voordoen in het moerassysteem. Op basis van de waarnemingen in dit proefschrift heb ik een operationeel model geconstrueerd waarin deze interacterende processen worden weergeven in relatie tot veranderingen in de concentraties, aard en type van zwevende stofdeeltjes in moerassystemen. Dit model biedt een wetenschappelijke basis voor het ontwerpen en optimaliseren van zuiveringsmoerassen die gebruikt kunnen worden om natuurlijk oppervlaktewater te beschermen tegen schadelijke effecten van de toenemende concentraties van zwevende stofdeeltjes uit verschillende bronnen waaronder afvalwater.

Acknowledgements

152

Acknowledgements

I would like to start with thanking the people that were directly involved in this project for the last four years. Harm I always enjoyed our meetings in which you were able to take care of the finishing touch, I could count on you to have my back in difficult situations and I will always have good memories of our great trips to Barcelona and Prague (budete se dostat zpět do člunu). Wim, your door was always open with an attentive ear, although I think I sometimes talked a little bit too much for you. I would like to thank you for all the good suggestions and remarks on all the papers. Joost, besides our chats about outdoor activities, our meetings were always very productive and with your eye for detail and different view on results, you often found additional matters of interest. Rob, with all your experience in constructed wetlands and your network you were always a good source of information and I especially enjoyed arranging the large field experiments together in the beginning of the project. All together I am grateful for having a good advisory committee, I have learned a lot from each of you and I want to thank you all for the support and input during this Ph.D. project. Secondly I would like to thank my family who has always supported me. Marieke I have been lucky that you are also knowledgeble in ecological research, which allowed you not only to be an immensely important physical, mental and emotional support in my life, but also work together on research topics. My greatest gratitudes are for you my love and I hope that we will be able to support each other for long times to come. Mom and dad, I want to thank you for always supporting me in all the things I wanted to do, even do that path was not always clear. Joke and Richard I also want to thank you for your support and for the many nights I was able to stay over whenever it was needed. I also want to thank other family members for their support and the countless times they provided some much needed distraction from all the work. Of course I want to thank all my colleagues I had during the past four years. Ellard and Marino you two were the only two Ph.D. students working in the department when I started and we certainly had some good times (even though you two needed to get use to a non-beer drinker). Ale, Merrin en Raul, you were always in for drinks and good relaxing conversations and I’m glad that we have great times together both it in Holland and in Belgium. Susanne and Sasha I have enjoyed the many tea breaks, walks, drinks and accompanying conversations, and I hope I did not complain too much to you at the end of the project. Ciska and Brittany, you two came into the group a little bit later, but I certainly enjoyed your company. To all my Ph.D. colleagues I wish all the best in all you are doing and hope we will have many future opportunities to raise the glass together. Helen you were my first roommate and I could not have had a better one, and I hope you will quickly find a good place that suits your wishes and of course you are always welcome to come over a try some of good whiskeys. Michiel, I will never forgot how many stories you can tell and I will miss them. I wish you the best and hope calmer weather is on the horizon for you. Bas and Jasper, you two were also always good conversation partners and I wish you all the best. Arie, you started at working at the department just after I moved to Germany and we did not see each

Acknowledgements

153

other that much, but nonetheless we already had very good times and I certainly hope you will enjoy being a ‘paranimf ’. Leo, many thanks for all the help and advice you have provided over the course of this project. My field experiments would not have been successful without massive help and support from a lot of people. Rinse, Michel and Peter many thanks for your commitment and all the good times we had. Special thanks go out for Hans, you stood by me from before dawn till after dusk, I will commemorate your energy and drive. Ron and Yolanda, I also want to thank you for the pleasant collaboration and good times during the WIPE project. During this project I have had the pleasure of working with several students. Christine, you were my first student and although it wasn’t always rosy, I wish you luck in your future endeavors. Anne and Daan, we worked only relatively short time together, but I thoroughly enjoyed it. Maxine, Simon and Zoi, thank you for all the work and wish you all good luck in your career. The list of people to thank almost seem endless, but there are still several people I would like to thank. Jenny, Caterina (there is enough whiskey for you too), Angelica, Kristen and everybody else at IGB, I thank you for the shamefully short, but very good period in Germany. I wish you all the best and I’m sure we will see each other again. Karlijn and Miranda I want to thank you both for all the support you have given me especially during the last part of this thesis. Both the mental reflections and the insides from a person that recently went through same ordeal were very important. David, thank you for all the support you provided during the past four years, including all the moving from house to house and the many good times of distraction (not sure if I have ‘enough’ whiskey for you). I hope your new career path will work out!

Curriculum Vitae

155

Curriculum vitaeBram T.M. Mulling was born on the 3rd of December 1983 in Doetinchem, The Netherlands. After following secondary school at the Ulenhof in Doetinchem he studied Laboratory Sciences at the University of Professional Education Larenstein in Velp. He graduated in 2005 with a major in (environmental) Toxicology. He continued with M.Sc. study in Biological Sciences with a major in Limnology and Oceanography at the University of Amsterdam (UvA) and graduated in 2008. During his education he performed research projects at the Netherlands Institute for Sea Research (NIOZ), the Netherlands Institute for Ecology (NIOO), the University of Amsterdam (UvA) and the University of Waikato (New Zealand). In these projects he studied aquatic food webs, population dynamics and the effects of environmental problems such as eutrophication and raising CO2 levels on phytoplankton communities. After graduation he started the Ph.D. research described in the thesis. The work was conducted at the University of Amsterdam (UvA), department of Aquatic Ecology and Ecotoxicology (AEE) in collaboration with Stichting Waternet, the Foundation for Applied Water Research (STOWA) and Witteveen+Bos. This project was supervised by prof. dr. Wim Admiraal (UvA) and dr. Harm G. van der Geest (UvA) and resulted in several publications and this thesis. Momentarily Bram ([email protected]) is looking for a new challenge in the field of aquatic ecology or environmental science as researcher or project manager.

PublicationsMulling BTM, van den Boomen RM, van der Geest HG, Kappelhof JWNM, Admiraal W (2013). Suspended

particle and pathogen peak discharge buffering by a surface-flow constructed wetland. Water Research 47 (3), 1091-1100.

Mulling BTM, van den Boomen RM, Claassen THL, van der Geest HG, Kappelhof JWNM, Admiraal W (2013). Physical and biological changes of suspended particles in a surface flow constructed wetlands. Under revisions for Ecological Engineering.

Mulling BTM, van der Oost R, van der Wielen PWJJ, van der Geest HG, Admiraal W (2013). Processes removing faecal indicator organisms in constructed wetlands. Submitted

Mulling BTM, Soeter AM, van der Geest HG, Admiraal W (2013). Changes in the planktonic microbial community during residence in a constructed wetland. Submitted

Mulling BTM, Admiraal W, van Beusekom SAM, Bichebois S, Schwartz T, van der Geest HG (2013). Retention of bacterial cells and latex beads in natural and cultured phototrophic biofilms. Manuscript

Foekema EM (Ed). (2012). The influence of constructed wetlands on the quality of WWTP effluent and recommendations for improvements. IMARES report C005/12 (Dutch)

van den Boomen RM, Kampf R, Mulling BTM (2012). Research on suspended particles and pathogens in the Waterharmonica (constructed wetland). STOWA publication 2012-10, ISBN.978.90.5773.553.0. (Dutch)

van der Stap I, Vos M, Kooi BW, Mulling BTM, van Donk E, Mooij WM (2009). Algal defenses, population stability, and the risk of herbivore extinctions: a chemostat model and experiment. Ecological Research 24 (5), 1145-1153.

Particles matter - ibed.uva.nlibed.uva.nl/.../map-1/2013-bram-mulling---particles-matter.pdf ·...

Documents

Transcript of Particles matter - ibed.uva.nlibed.uva.nl/.../map-1/2013-bram-mulling---particles-matter.pdf ·...