Linköping Studies in Science and Technology Dissertation No. 1509
Wastewater treatment in constructed wetlands:
Effects of vegetation, hydraulics and data analysis methods
Hristina Bodin
Department of Physics, Chemistry and Biology IFM Biology
Linköping University SE-581 83 Linköping, Sweden
Linköping, May 2013
Linköping Studies in Science and Technology. Dissertation No. 1509
Bodin, H. 2013.
Wastewater treatment in constructed wetlands:
Effects of vegetation, hydraulics and data analysis methods
Linköping Studies in Science and Technology ISBN 978-91-7519-649-7
ISSN 0345-7524
Copyright © Hristina Bodin, unless otherwise stated All rights reserved
Cover photo by Hristina Bodin
Constructed wetland at Chemelil Sugar Factory Ltd., Kenya
Printed by LiU Tryck, Linköping, Sweden, 2013
For Elin and Anton, each equally my pride and joy.
“I often say that when you can measure what you are speaking about, and express it in numbers, you know something about it”
Lord Kelvin
“Who speaks for Earth?”
Carl Sagan
A
BSTRACTDegradation of water resources has become one of the most pressing global concerns currently facing mankind. Constructed Wetlands (CWs) represent a concept to combat deterioration of water resources by acting as buffers between wastewater and receiving water bodies. Still, constructing wetlands for the sole purpose of wastewater treatment is a challenging task. To contribute to this research area, the fundamental question raised in this doctorate thesis was: how do factors such as vegetation and residing water movements (hydraulics) influence wastewater treatment in CWs? Also, effects of different data analysis methods for results of CW hydraulics and wastewater treatment were investigated. Research was focused on phosphorus (P), ammonium- nitrogen (NH4+
-N) and solids (TSS) in wastewater and on P in macrophyte biomass.
Studies were performed in pilot-scale free water surface (FWS) CW systems in Kenya (Chemelil) and Sweden (Halmstad) and as computer simulations.
Results from the Chemelil CWs demonstrated that meeting effluent concentration standards simultaneously for all water quality parameters in one CW was difficult.
Vegetation harvest, and thus nutrient uptake by young growing macrophytes, was important for maintaining low effluents of NH4+
-N and P, especially during dry seasons. On the other hand, mature and dense vegetation growing for at least 4 months secured meeting TSS standards. Phosphorus in above-ground green biomass accounted for almost 1/3 of the total P mass removal, demonstrating high potential for P removal through macrophyte harvest in CWs. Also, results suggested that harvest should be species-specific to achieve high P removal by macrophytes and overall acceptable wastewater treatment in CWs. Still, different methods to estimate evapotranspiration (ET) from the Chemelil CWs showed that water balance calculations greatly impacted estimations of wastewater treatment results.
Hydraulic tracer studies performed in the Chemelil and Halmstad CWs showed that mature and dense emergent vegetation in CWs could reduce effective treatment volumes (e-values), which emphasized the importance of regulating this type of vegetation. Also, it was shown that hydraulic tracer studies with lithium chloride performed in CWs with dense emergent vegetation had problems with low tracer recoveries. This problem could be reduced by promoting the distribution of incoming tracer solution into the CW using a barrier near the CW inlet pipe. Computer simulation results showed that the choice of tracer data analysis method greatly influenced quantifications of CW hydraulics and pollutant removal. The e-value could be 50% higher and the pollutant removal 13% higher depending upon used method. On average, the methods of moments (M) with residence time distribution (RTD) truncation at 3 times the hydraulic residence time (tn) resulted in higher (25%) e- and lower (-31%) N-values (dispersion) compared to using the gamma model or M with RTD truncation at tracer background concentration. Still, differences between the gamma model and the M could be minimized by truncating the RTD at tracer background levels when the latter method was used. Thus, comparing results from two methods is justified and could probably help future methodological refinements. Yet, the most common method in published articles in the last 25 years for estimation of hydraulic parameters was the M with RTD truncation at 3tn.
This indicated that overestimated e- and underestimated N-values could be common in published articles. Moreover, unrealistic e-values (above 100%) in published literature could to some extent be explained by tracer data analysis method. Hence, to obtain more reliable hydraulic data and wastewater treatment results from CWs, more attention should be paid to the choice of tracer data analysis method.
Key words: constructed wetland, free water surface flow, wastewater treatment, Kenya, Sweden, vegetation, harvest, Cyperus papyrus, Echinochloa pyramidalis, mass load, phosphorus, ammonium, suspended solids, pollutant removal, hydraulics, residence time distribution, data analysis methods
S
AMMANFATTNINGKonstruerade våtmarker representerar ett koncept för möjligheten att nå en hållbar vattenresurshantering genom att agera som ”filter” mellan föroreningskälla och viktiga vattenresurser såsom sjöar och hav. Mycket kunskap saknas däremot om hur man konstruerar våtmarker med en optimal och pålitlig vattenreningskapacitet. Den här avhandlingen undersöker därför hur vegetation och vattnets väg genom våtmarken (hydrauliken) påverkar avloppsvattenrening i våtmarker. Dessutom undersöktes hur valet av dataanalysmetod av insamlad data påverkar resultaten. Studier genomfördes i Kenya och Sverige i experimentvåtmarker (ca. 40-60 m2) och inkluderade datainsamling av vattenkvalité, hydraulik (spårämnesexperiment) samt biomassa och fosfor i biomassan av två olika våtmarksväxter. Dessutom genomfördes datorsimuleringar.
Resultaten från Kenya visade att växtskörd och efterföljande näringsupptag av nyskördade växter var viktig för att uppnå låga utgående koncentrationer av fosfor och ammonium i en tropisk våtmark, speciellt under torrsäsongen. Däremot var en välutvecklad och tät vegetation viktig för reningen av partiklar. Fosfor i grön växtbiomassa representerade cirka 1/3 av våtmarkernas totala fosforrening, vilket påvisade potentialen i att genom skörd ta bort fosfor från avloppsvatten m.h.a.
konstruerade våtmarker. Resultaten pekade också på att skörden bör vara art-specifik för att uppnå en hög fosforrening och generellt bra vattenreningsresultat. Dock visade olika beräkningsmetoder att vattenbalansen i en tropisk våtmark markant kan påverka vattenreningsresultaten.
Resultaten från spårämnesexperimenten demonstrerade att den effektiva våtmarksvolymen för vattenrening blev mindre vid hög täthet av övervattensväxter.
Detta pekade på att regelbunden växtskörd var viktig för att uppnå god vattenrening i våtmarker. Experiment med spårämnet litium visade att man kan få felaktiga resultat p.g.a. att en del spårämne fasthålls på botten i våtmarken om denna har mycket övervattensväxter. Därför bör spridningen av spårämnet i sådana våtmarker underlättas m.h.a. en spridningsbarriär nära inloppsröret. Simuleringar visade också att valet av dataanalysmetod av spårämnesdata starkt kan påverka resultaten och därmed också vår tolkning av en våtmarks hydraulik och reningskapacitet. Den effektiva volymen kunde vara 50% högre och reningseffekten 13% högre beroende på vilken metod som användes. Likaså kan valet av dataanalysmetod ha bidragit till överskattade och orealistiska effektiva volymer (över 100%) i artiklar publicerade de senaste 25 åren.
Genom att fokusera mer på valet av dataanalysmetod och t.ex. jämföra resultaten från två olika metoder kan man minimera risken för bristfälliga resultat och därmed felaktiga slutsatser om en våtmarks vattenreningskapacitet.
Nyckelord: konstgjorda våtmarker, avloppsvatten, vattenrening, fosfor, ammonium, partiklar, Kenya, Sverige, växter, Cyperus papyrus, Echinochloa pyramidalis, skörd, hydraulik, dataanalysmetod
L
IST OFP
APERSThis doctorate thesis is comprised of the following papers, which are referred in the text by their Roman numerals.
I Bojcevska, H. & Tonderski, K. (2007). Impact of loads, season, and plant species on the performance of a tropical constructed wetland polishing effluent from sugar factory stabilization ponds. Ecological Engineering 29, 66–76.
II Bojcevska, H., Raburu, P.O. & Tonderski, K.S. (2006). Free water surface constructed wetlands for polishing sugar factory effluent in western Kenya - macrophyte phosphorus recovery and treatment results. In: Dias, V., Vymazal, J.
(eds.) Proceedings of the 10th International Conference on Wetland Systems for Water Pollution Control, 23-29 September 2006; Ministério de Ambiente, do Ordenamento do Territóri e do Desenvolvimento Regional (MAOTDR) and IWA: Lisbon, Portugal, pp. 709–718.
III Bodin, H. & Persson, J. (2012). Hydraulic performance of small free water surface constructed wetlands treating sugar factory effluent in western Kenya.
Hydrology Research 43, 476
–
488.IV Bodin, H., Mietto, A., Ehde, P.M., Persson, J. & Weisner, S.E.B. (2012). Tracer behaviour and analysis of hydraulics in experimental free water surface wetlands. Ecological Engineering 49, 201–211.
V Bodin, H., Persson, J., Englund, J.E. & Milberg, P. (2013). Diluting the evidence? How residence time analyses can influence your results. Submitted manuscript1.
Paper I, III and IV are reprinted with kind permission from the copyright holders.
1 Submitted to Journal of Hydrology
Note: the author H. Bodin has also formerly been known as H. Bojcevska.
M
Y CONTRIBUTION TO THE PAPERSI I planned the study togheter with co-author. I conducted the field work with assistans from B. Odhiambo Owour and J. Odenge. I carried out the laboratory analysis. I performed the statistical analyses of the aquired data. I wrote the paper with contribution from co-author.
II I planned the study togheter with co-authors. I conducted the field work with assistans from B. Odhiambo Owour and J. Odenge. P.O. Raburu, was in charged of the macrophyte harvesting events. I carried out the water quality related laboratory analysis. I performed the statistical analyses of the aquired data. I wrote the paper with contribution from co-authors.
III I planned the study togheter with co-author. I conducted the field work with assistans from B. Odhiambo Owour. I carried out the water quality related laboratory analysis. I performed the statistical analyses of the aquired data. I wrote the paper togheter with co-author.
IV Stefan Weisner, Jesper Persson and Anna Mietto planned the study. Per Magnus Ehde, Anna Mietto and Stefan Weisner conducted the field work. Per Magnus Ehde carried out the laboratory analysis. I conducted the statistical data analysis with support from Stefan Weisner. I wrote the paper with assistance from co- authors.
V I planned the study together with co-authors. I carried out the literature review and computer simulations. I conducted the analysis of the aquired data. I wrote the paper with contribution from co-authors.
To all above papers, I have had main responsibility for coordination of the journal submission processes and therein related responses to reviewers and correspondence with editors.
T
ABLE OFC
ONTENTSAbstract ____________________________________________________________ i Sammanfattning ____________________________________________________ iii List of Papers ______________________________________________________ v My contribution to the papers _______________________________________ vi Table of Contents ___________________________________________________ vii 1. Introduction ______________________________________________________ 1 2. Literature review __________________________________________________ 2 2.1 Constructed Wetlands (CWs) for wastewater treatment _______________ 2 2.1.1 Tropical versus temperate constructed wetlands ____________________ 3 2.2 Pollutant removal processes in constructed wetlands ________________ 4 2.2.1 Nitrogen removal in constructed wetlands _________________________ 6 2.2.2 Phosphorus removal in constructed wetlands ______________________ 7 2.2.3 Suspended solids removal in constructed wetlands ________________ 10 2.3 Critical factors for pollutant removal _____________________________ 11 2.3.1 The role of hydraulic- and pollutant load _________________________ 11 2.3.2 The role of wetland vegetation _________________________________ 12 2.3.3 The role of wetland hydraulics _________________________________ 16 3. Objectives ______________________________________________________ 22 4. Methods ________________________________________________________ 23 4.1 Study sites and experimental designs ____________________________ 23 4.2 Sampling programme for water flow and quality ____________________ 23 4.3 Methods for hydraulic tracer studies _____________________________ 24 4.4 Above-ground macrophyte harvest and nutrient analyses ___________ 24 4.5 Water balance estimations ______________________________________ 25 4.6 Calculation of pollutant mass balances ___________________________ 26 4.7 Scientific work in developing countries ___________________________ 26 5. Main results and discussion _______________________________________ 27 5.1 The Chemelil constructed wetland system ________________________ 27 5.1.1 Water balance _____________________________________________ 27 5.1.2 Wastewater treatment in the Chemelil CW _______________________ 27 5.1.3 Effects of macrophytes ______________________________________ 31 5.1.4 Macrophytes as nutrient traps _________________________________ 32 5.2 The hydraulic tracer studies ____________________________________ 34 5.2.1 Effects of tracer data analysis method ___________________________ 36 6. Conclusions ____________________________________________________ 39 7. Future studies ___________________________________________________ 40 8. Acknowledgements ______________________________________________ 41 References _______________________________________________________ 42 Appendix A
Appendix B Appendix C Appendix D Paper I Paper II Paper III Paper IV Paper V
1. I
NTRODUCTIONWater of sound quality is the key for vital socio-economic functions on Earth.
However, the two recent centuries of industrial and agricultural expansion has, in combination with weak regulatory mechanisms, led to widespread degradation of water resources (European Environment Agency 1995; Muyodi et al. 2010; Entrekin et al. 2011; Goulden 2011; Gerbens-Leenes and Hoekstra 2012). Moreover, in the last five decades, the degradation has been amplified by intensive use of fertilizers to maximize crop yields on arable land (Walsh 1991a; Matson et al. 1997; Sophocleous 2004) and simultaneous drainage of numerous natural wetlands (Mitsch and Gosselink 2000; Kingsford and Thomas 2004; Muyodi et al. 2010). The drainage trend has led to less contact time between these natural water purification systems and water-borne contaminants originating from various anthropogenic sources. Consequently, this has meant extensive losses of water-borne particles and nutrients from land to important water resources (Walsh 1991b; Hopkinson and Vallino 1995; Matson et al. 1997;
Falkowski et al. 2000).
It is well known that nutrients, such as nitrogen (N) and phosphorus (P), limit primary production in aquatic ecosystems (Wetzel 2001; Elser et al. 2007). Not surprising, excess of these nutrients in aquatic ecosystems has been observed to cause algal blooms (Smith 2003) and associated eutrophication problems such as oxygen depletion, toxic effects on non-target organisms and overall altered ecosystems functions (Glasgow and Burkholder 2000; Boesch et al. 2001; Dudgeon et al. 2006).
Ultimately, such changes have also meant loss of services that these ecosystems provide and reduced quality of life for human populations depending on them (Postel and Carpenter 1997; Kivaisi 2001; Lung’ayia et al. 2001; Beeton 2002). Accordingly, costs for water resource degradations are estimated to be high and thus benefits of managing water resources adequately would be large (World Bank 2010). Evidently, there is a need for effective management efforts, where one possible action is to focus on minimizing nutrient losses from pollutant-producing catchments to water resource areas. One way is to construct wetlands to mitigate the water purification needs for upstream contaminated wastewater from various anthropogenic sources. However, key questions that need to be answered are, (1) what factors regulate wastewater treatment in constructed wetlands (CWs) and (2) which is the most accurate way to analyze wastewater treatment data from CWs? These are the fundamental questions in focus in this thesis. The scope of this thesis is limited to examining how P (both dissolved and particulate P fractions), N (as ammonium-N) and total suspended solids (TSS) can be managed effectively in CWs.
2. L
ITERATURE REVIEW2.1 C
ONSTRUCTEDW
ETLANDS(CW
S)
FOR WASTEWATER TREATMENTIn the last 3 decades, accumulating research has demonstrated that wetlands may provide water quality improvement (Nichols 1983; Richardson 1985; Knight et al.
1993; Mashauri et al. 2000; Blahnik and Day 2000; Kadlec 2003; Kadlec 2009) at lower capital costs compared to other water treatment methods (Brix 1999; Ko et al.
2012). This may not be surprising, since year-round occurrence of water stimulates vegetation growth and microbiological activity which make wetlands one of the most productive ecosystems on Earth (Kadlec and Wallace 2009). Through high microbiological productivity, aided by sunlight, wind and soil, wetlands can transform a variety of pollutants into less harmful by-products or life-supporting nutrients.
Hence, using wetlands, it is possible to meet water quality standards with negligible use of fossil-based energy and chemicals (Gearheart 1992; Martin et al. 1999; Scholz et al. 2010). These benefits have created a global interest in constructing wetlands for the main purpose of wastewater treatment (Kadlec and Wallace 2009; Isosaari et al.
2010; Vymazal 2011).
Based on differences in the water flow, two main groups of CWs can be distinguished:
free water surface (FWS) CWs and subsurface flow (SSF) CWs. In FWS CWs the water surface is exposed to the atmosphere and flows horizontally over the soil surface. The mean water depth is usually less than 0.4 m, and thus, FWS CWs are frequently dominated by rooted emergent or submersed, or floating vegetation depending on the water depth. In SSF CWs, the water surface is kept below the surface of the substrate, which may support different types of rooted emergent vegetation. The SSF CW type may be further divided into vertical and horizontal flow systems (Kadlec and Wallace 2009). The focus in this thesis is on FWS CWs receiving relatively constant point-source wastewater.
According to Wallace and Knight (2006), in 2006 there were more than 400 small- scale wastewater treatment CWs (water flow < 2000 m3 per day) in operation in North America. Also, since CWs have relatively low construction, operational and maintenance costs, there is a growing interest in implementing them in developing countries (Kivaisi 2001). In Sweden, mainly as a result of the Swedish Environmental Quality Objectives implemented in 1999, the construction of at least 12 000 hectares of wetlands were stated as a numeric objective (Government Bill 1997/98:145;
Government Bill 2000/01:130). A study by Strand and Weisner (2013) reported that the wetland construction programme in Sweden has been a cost-effective method for decreasing transport of diffuse pollution from arable land. However, despite increasing use of CWs for wastewater treatment, several studies have demonstrated that the technique may have problems to meet local water quality standards (Newman and Clausen 1997; Li et al. 2008; Rodriguez and Lougheed 2010; Forbes et al. 2011;
Kantawanichkul and Duangjaisak 2011; Brix et al. 2011). Unpredictable and undesirable treatment behaviour reveals that we have insufficient knowledge to be able to optimize CWs for wastewater treatment. Thus, more studies are needed to find biological and physical features that regulate the complex pollutant removal processes in CWs (Gottschall et al. 2007).
Generally, the key questions that remain to be answered involve functions of vegetation and hydraulic residence time (Mitsch et al. 2005; Díaz et al. 2009; Erler et al. 2011; García-Lledóa et al. 2011) in the CW treatment process. Especially, more knowledge is needed to bridge the gap between CW hydraulic performance (water flow patterns) and wastewater treatment performance, where only few empirical studies have been published so far (Dierberg et al. 2005; Kusin et al. 2010).
Regarding CWs in developing countries, the starting point for optimized performance is collection of long-term water quality data. This is important since it can be assumed that CW performance in tropical regions (where majority of developing countries reside) is different from that of temperate ones due to different climatic conditions and vegetation cycles. Thus, these differences indicate that design, operational and maintenance strategies used for CWs in temperate regions probably cannot be directly applied to tropical ones.
2.1.1TROPICAL VERSUS TEMPERATE CONSTRUCTED WETLANDS
Wastewater treatment in CWs is highly affected by climatic conditions since these directly regulate abiotic factors such as solar radiation, temperature, precipitation and evapotranspiration (ET), i.e. the combined effect of water loss via water surfaces and transpiration from wetland vegetation (Wittgren and Mæhlum 1997; Kadlec and Wallace 2009). The abiotic factors in turn affect the biotic factors which involve microbiological activity and vegetation dynamics. Some studies have indicated higher treatment efficiency of tropical CWs and attributed this to higher temperatures which stimulate year-round vegetation growth and microbiological activity and thus higher nutrient uptake (Kivaisi 2001; Kaseva 2004; Diemont 2006; Katsenovich et al. 2009).
In fact, turnover rates of above-ground vegetation biomass in tropical areas are rapid, and are reported to occur at 1 to 3 months intervals, as opposed to once a year in temperate regions (Reddy et al. 1999; Kadlec 2005a). Moreover, the breakdown of large amounts of vegetation litter may lead to more rapid development of anoxic conditions, which can cause release of phosphorus (P) from wetland sediments. Hence, the impact of vegetation tissue losses to senescence and decomposition on CW treatment performance can be important (DeBusk and Ryther 1987) by causing lower net removal of nutrients in tropical CWs compared to those in colder regions.
The functioning of a CW is considerably influenced by the water balance since this can alter its hydrology and thus pollutant removal capacity (Kadlec and Knight 1996).
Especially, tropical CWs generally experience periods of extreme rain or drought which can affect their treatment performance considerably (Lim et al. 2001;
Kyambadde et al. 2005; Diemont 2006; Katsenovich et al. 2009). For tropical CWs in particular, high ET may impact the water balance by reducing water outflow rates resulting in higher hydraulic residence times (HRTs) and through condensation increased ouflow pollutant concentrations. Katsenovich et al. (2009) noted higher relative removal (as % of inflow concentration) but also higher outflow concentrations of total dissolved P (TDP) during the dry season compared to the wet season.
However, the cited study, reported that during the wet season, a three-fold increase in mass load of TDP resulted in a doubling of the area-specific mass removal rate (g TDP per m2 wetland) of TDP compared to the situation in the dry season.
Despite the fact that the water balance can have a significant impact on the treatment capacity of tropical CWs, many researchers have evaluated pollutant removal based only on differences between inflow and outflow pollutant concentrations, hence omitting effects of the water balance (Juwarkar et al. 1995; Perfler et al. 1999; da Motta et al. 2000; Meutia 2001; Kyambadde et al. 2004). Thus, underestimations of CW treatment results may be numerous (Goulet et al. 2001; Lin et al. 2002; Jing et al.
2002; Diemont 2006). Consequently, including waterbalance effects and evaluating treatment performance of tropical CWs based on mass loading analyses might lead to more correct evaluations of these systems (Katsenovich et al. 2009).
2.2 P
OLLUTANT REMOVAL PROCESSES IN CONSTRUCTED WETLANDSGenerally, key processes responsible for pollutant removal in CWs are sedimentation, chemical precipitation and adsorption, microbial activities and macrophyte uptake (Vymazal 2001; Kadlec and Wallace 2009). Still, numerous factors affect the removal processes and a pollutant is generally removed as a result of several interconnected processes. In Table 1, the most common removal mechanisms for P, N and TSS are listed. However, the removal mechamisms in CWs remain an active research area (Gottschall et al. 2007; Kadlec and Wallace 2009), and therefore, the information enclosed in Table 1 should not be taken as definate. The removal processes in wetlands are very complex and which processes dominate depends to a great extent on vegetation and water flow, factors that are relatively unexpensive and easy to manipulate. Thus, more knowledge about how vegetation and water flow affects pollutant removal may enable the design of low cost CWs that are optimized for wastewater treatment.
Nitrogen removal in CWs is highly affected by both temperature and the level of oxygen, whereas it is usually agreed that TSS and P removal processes, due to their physical (sedimentation) and chemical (precipitation and adsorption) characteristics, are not as sensitive to temperature effects. However, P removal can be significanly affected by the oxygen status in the CW (Soto-Jimenez et al. 2003; Kadlec and Wallace 2009). Also, on a seasonal level, differences in the removal of P can be seen due to higher P uptake by vegetation during summer and lower during fall and winter.
Also, high returns of P from the decomposing vegetation litter during the fall may counterbalance the high uptake rates during summer (Kadlec and Knight 1996).
Table 1. Summary of removal processes for total suspended soilds (TSS), phosphorus (P) and nitrogen (N) in free water surface constructed wetlands receiving wastewater (adapted with modification from Stowell et al. 1981). p = primary process;
s = secondary process; i = incidental process (a process as a result of another process); c = contributory process.
Removal process TSS P N Description Physical
Absorption p Gas transfer to and from water surface
Filtration p
Particulates filtered mechanically as water flows through substrate, plant litter and roots
Sedimentation p i i Gravitational settling of larger particles and contaminants
Volatilization p Gas absorption with a net flux out of the water surface
Diffusion p/s p/s
Movement of matter from a volume of high concentration to a volume of low concentration of that element
Chemical
Adsorption/Desorption s p
Intermolecular attractive force (van der Waals force)
Precipitation p Formation of co-precipitates with
insoluble compounds
Redox reactions i p/s p/s Oxidation: electron loss Reduction: electron gain Biological
Algal assimilation p/s p/s Algal uptake of nutrients in wastewater
Vegetation assimilation c p/s p/s Uptake and metabolism of nutrients in wastewater by higher plants
Bacterial assimilation Aerobic Anaerobic
s/i s/i
i i
i i
Immobilization of colloidal solids and soluble inorganics and organics by suspended, benthic and plant supported bacteria
2.2.1NITROGEN REMOVAL IN CONSTRUCTED WETLANDS
In CWs, nitrogen is removed via several pathways that may start with the decomposition of organic nitrogen present in wastewater by heterotrophic bacteria and fungi to ammonium (NH4+), a process called ammonification (Fig. 1). This process may occur both in aerobic and anaerobic environments, but is much faster under the former condition. Also, ammonium may be temporarily removed from the water column by binding to negatively charged sites on soil particles in the CW sediment through adsorption (Mitsch and Gosselink 2000). However, changes in water chemistry or hydrology can release loosely bound NH4+ back to the water column through desorption (Reddy and Patrick 1984). Still, removal of NH4+ in CWs is dominated by an aerobic process called nitrification, in which NH4+ is oxidized to nitrite (NO2-) and further to nitrate (NO3-) by bacteria (Fig. 1). The nitrification rates in a CW will be favored by oxic conditions, the availability of inorganic carbon and NH4+, as well as temperature and pH ranges of 30–40°C and 7.5–8, respectively (Kadlec and Knight 1996). Nitrification may still occur at dissolved oxygen levels down to about 0.3 mg L-1 (Reddy and Patrick 1984). However, more recent studies have indicated that oxygen levels of 1–2 mg L-1 in CW water columns can decrease NH4+ removal through lower nitrification rates (Hammer and Knight 1994; Okurut et al. 1999). Ammonium may also be lost to the atmosphere as ammonia gas (NH3) through a process called ammonia volatilization (Table 1; Fig. 1). However, at pH of below 8 and quiescent flow conditions, which are typical for CWs, volatilization is normally insignificant as a process for NH4+ removal (Kadlec and Knight 1996).
Figure 1. A simplified illustration of the nitrogen cycle in free water surface constructed wetlands (modified from Kadlec and Knight 1996).
Diffusion may transfer NO3- from the CW water phase to the sediments and vice versa (Fig. 1; Reddy and Patrick 1984). Nitrate is transformed by bacteria through a process called denitrification (Fig. 1) to nitrogen gas (N2), which diffuses from the CW water surface and thus returns nitrogen to the atmosphere.
NH4+
floc or sediment/soil
bottom level NH4+
NH4+
NO3-
NO3-
NO3-
NO2-
NO3-
NO
4
4+ 3-
NO2 NO NO- NO3-
roots Organic N
roots Organic N Organic N
roots Organic N
NH NH
NH3 (gas) Air
water level N2 + N2O (gases)
Organic N Organic N
NH3(gas)
The denitrifying bacteria are located on epiphytic biofilms that coat submerged CW surfaces, which mainly are sediments, plant parts and litter (Bastviken et al. 2003).
Denitrification is favored by anoxic conditions, high availability of organic carbon, high temperatures (optimum 60–75°C) and pH ranges of 6–8.5 (Reddy and Patrick 1984). The denitrification process means a one-way loss of nitrogen from a CW system (Fig. 1).
The transfer of di-nitrogen gas (N2) from the atmosphere to nitrogen in the water phase (a form of absorption) is a bacterial process called nitrogen fixation (Fig. 1). However, nitrogen fixation is normally not significant in CWs receiving wastewater (Kadlec and Knight 1996).
Assimilation of nitrogen, i.e. uptake of inorganic nitrogen and transformation to organic nitrogen in living cells and tissues of plants, algae and microorganisms (Table 1), is considered to be insignificant in high-load CWs (Brix 1997; Kadlec 2005b).
Wetland macrophytes favor NH4+
-N as the nutrient form of nitrogen since it is more energy efficient as a building stone for amino acids, proteins and other nitrogenous organic molecules (Kadlec and Knight 1996; Wetzel 2001). Still, macrophytes can also take up NO3-, but the assimilation rate is determined by the NH4+-N availability in the CW (Martin and Reddy 1997). However, when vegetation dies, substantial amounts of N are returned to the water and soil of the CW through decomposition and mineralization (Nichols 1983; Howard-Williams 1985). Thus, CW vegetation is only considered to be a temporary N sink, except if the biomass is harvested (Brix 1997).
Generally, FWS CWs with emergent vegetation treating wastewater of inlet concentrations in the range 7.4–19 mg NH4+
-N L-1 can achieve relative removal of incoming concentration in the order of 28–60% and an absolute removal of 0.18–1.51 g NH4+-N m-2 day-1 (Kadlec and Knight 1996; Kyambadde et al. 2005; Vymazal 2007).
2.2.2PHOSPHORUS REMOVAL IN CONSTRUCTED WETLANDS
The key P fractions in a wastewater receiving CW are normally particulate phosphorus (PP) and dissolved phosphorus (DP; Fig. 2), together comprising total P (TP).
Dissolved P, normally as phosphate-phosphorus (PO43--P), is available for living organisms, whereas PP generally must go through transformations to become available and further transformed to organic P (OP). Both abiotic and biotic processes control the relative P fraction sizes and the transformation rates of P fractions within the CW water column and sediment/soil. Abiotic processes include sedimentation, adsorption to sediment/soils, precipitation, and exchange processes between soil/sediment and the overlying water column. Biotic processes include assimilation by vegetation, plankton, periphyton and microorganisms (Reddy et al. 1999; Table 1). Many researchers consider accretion, i.e. the creation of new stable residuals, as the major long-term P storage process in wetlands (Nichols 1983; Howard-Williams 1985; Richardson and Marshall 1986; Reddy et al. 1999; Kadlec and Wallace 2009). The accretion process is the sum of chemical P precipitation, adsorption and accumulation of detritus in the CW sediment/soil.
As wastewater enters a CW, organic and inorganic particles and associated P start to form a new sediment layer through the accumulation of a low density material called floc. The P content in the floc in CWs receiving secondary wastewater of approximately 3 mg P L-1 is reported be around 0.1–0.4% per dry weight (DW). Still, the thickness of the floc layer can be 20–30 cm and hold 3–50 g P m-2 CW (Kadlec and Wallace 2009). The floc may be removed from the CW via vacuuming or other suitable method. Still, if left untouched with time, the floc decomposes and forms a new soil layer, and thus incorporates a part of the P that has entered the CW through wastewater. However, under certain conditions P accumulated in the CW floc or soil can resuspend back to the water column as PP and later desorb to DP (Fig. 2).
Figure 2. A simplified illustration of the phosphorus cycle in free water surface constructed wetlands. PP is particulate phosphorus, DP is dissolved phosphorus and OP is organic phosphorus.
The key processes that regulate the P content in the floc or the CW sediment are adsorption/desorption and precipitation/solubilization (Fig. 2). The first process involves intermolecular attractive forces between PO43--P and inorganic and organic particles in the CW sediment (Table 1; Fig. 2). The second process is the formation of complexes between PO43--P and metal minerals present on edges of inorganic and organic particles or free cations in the water column. The bonds in a P mineral complex are less reversible than the adsorption bonds between P and inorganic or organic particles (Reddy et al. 1999). Eventually, these P fractions may be incorporated in the CW soil through the accretion process (Kadlec and Wallace 2009).
Still, PO43--P removal by CWs may depend on the P concentration in the pore water (water filled voids between soil particles) (Reddy et al. 1999; Kadlec and Wallace 2009), where a reduction in the pore water P concentration, for example as a result of changes in incoming wastewater composition, can lead to release of P from the CW sediment (Nichols 1983) or floc (Kadlec and Wallace 2009).
Wastewater inflow
Water outflow
floc or sediment/soil Immobilization
PP
OP
DP
Mineralization
Mineralization
Decomposition Fragmentation
Immobilization
Ca-P P Fe-P P Al-P PP DP
PP DP
Adsorption
OP Desorption
Diffusion
PP Decomposition
Fragmentation Resuspension Soil accumulation Resuspension
Soil accumulation Adsorption Desorption
Precipitation Desorption
Solubilization Desorption
water level
bottom level roots
Air
Also, adsorption/desorption and precipitation/solubilization are controlled by factors such as pH, redox potential and the amount of P and metal minerals in the sediment (Kadlec and Wallace 2009). In CW environments with 6 < pH < 8, PO43--P can be adsorbed to, and precipitate with, minerals of iron and aluminum present in the CW sediment/soil (Richardson 1985; Gale et al. 1994; Lindstrom and White 2011). Still, for iron rich sediments, aerobic conditions are also required for the P precipitation to occur (Kadlec and Wallace 2009). The complex between PO43--P and calcium is not sensitive to anoxic conditions (Kadlec and Wallace 2009), but requires CW environments with pH > 8 for the precipitation to occur and remain stable (Nichols 1983; Reddy and D´Angelo 1994; Reddy et al. 1999).
Several studies have declared that P removal by CWs normally decreases with time and discussed that this could be because most sorption sites on the metal minerals present in the soil have become saturated (Richardson 1985; Okurut et al. 1999;
Kadlec 2005a). This behaviour may be referred to as the “aging phenomena” in wetlands (Kadlec 1984). As a consequence of the “aging phenomena”, the sediment/soil can start to leak P (Nichols 1983; Richardson 1985; Kadlec and Knight 1996). However, leakage of P from newly created CWs has also been observed due to lower P concentrations in incoming water to the CW than historical P concentration on the site (Kadlec and Wallace 2009). Also, a change from oxic to predominantly anoxic surroundings in the CW sediment can cause solubilization of P bound to iron minerals in the wetland soil (Fig. 2; Gale et al. 1994; Vymazal 2001; Soto-Jimenez et al. 2003;
Kadlec and Wallace 2009). Thus, adsorption and precipitation of P by CW sediment/soils is not necessary stable, but instead reversible. The only mechanism to ultimately remove P from the CW is via dredging of the sediment/soil for ultimate disposal (Stowell et al. 1981) or by harvesting the CW macrophytes (Brix 1997).
However, Lindstrom and White (2011) suggested that aluminum sulfate (alum) addition could be a more cost effective method to improve P removal in aging CWs.
Still, studies have suggested that P precipitated with aluminum chemicals is of low P fertilization value and could, if applied to agricultural fields, result in P accumulations in soils with potential environmental risks associated to surface runoff of P (Ippolito et al. 2002; Krogstad et al. 2005).
Phosphorus is a limiting element for growth of algae and macrophytes in CWs, and thus, these organisms effectively assimilate P from wastewater present in CWs.
However, a part of the assimilated P is returned to the water phase upon senescence and succeeding decomposition, but may also be incorporated into new CW sediments and finally accreted (Kadlec and Wallace 2009). The net effect of macrophytes on the water phase P concentration may depend on macrophyte age and climatic conditions, but also on the CW operation and maintenance (Kadlec 2005a).
Literature shows that FWS CWs with emergent vegetation treating wastewater with inlet concentrations in the range 3.7–24 mg TP L-1 achieve variable relative removal of incoming concentration in the order of 9–62% and an absolute removal ranging between 0.02–2.14 g TP m-2 day-1 (Kadlec and Knight 1996; Greenway and Woolley 1999; Kyambadde et al. 2005; Diemont 2006; Vymazal 2007).
2.2.3SUSPENDED SOLIDS REMOVAL IN CONSTRUCTED WETLANDS
Removing suspended solids (SS) in wastewater is important since it prevents silting and also removes nutrients masses (attached to solids) reaching down-stream water bodies. Generally, removal of SS is not affected by temperature (Kadlec 2003; Kadlec and Wallace 2009) indicating that abiotic factors dominate the removal processes. The key variables for SS removal in CWs are adequate time for settling combined with trapping in the litter, sediment and soil. The relatively slow flow in CWs is often enough to give time for physical settling of SS, and at a constant flow, macrophytes may effectively contribute to increased sedimentation (Braskerud 2001).
Macrophytes may also limit resuspension of the sedimented particles by trapping these in the litter layer (Kadlec and Wallace 2009). Still, fragmentation of detritus from macrophytes and algae can produce particulate matter and thus increase SS concentration in a CW. Also, bioturbation by fish and mammals, water shear and gas flotation by oxygen or methane can resuspend solids into the CW water column. In addition, a weak seasonal dynamic in TSS removal may occur in CWs through algae and macrophytic turnover rates (Kadlec and Wallace 2009).
Based on numerous CW studies, Kadlec and Knight (1996) proposed a rule of thumb which stated that about 75% of incoming TSS is removed if inflow concentrations of TSS are above 20 mg L-1. The absolute removal of TSS in FWS CWs with emergent vegetation may be in the order of 2–10 g TSS m-2 day-1 depending on the TSS mass loading rates (Kadlec and Wallace 2009).
2.3 C
RITICAL FACTORS FOR POLLUTANT REMOVALGenerally, factors that are critical for pollutant removal in CWs may be studied from a process-, wetland- or landscape scale (Groffman et al. 1988; Trepel and Palmeri 2002;
Fig. 3). The process scale may involve effects of pH, temperature, nutrient availability and redox conditions, whereas the landscape scale may look at effects of upstream activities, land use, geology, climate and season. The wetland scale, which was in focus in this thesis, lies intermittently between these two, and may involve studying effects of hydraulic- and pollutant load, wetland vegetation and hydraulics on the removal of pollutants in wastewater. Of course, all the scales are highly interconnected and affect each other though complex ecological sequences.
Figure 3. Critical factors for removal of wastewater pollutants in constructed wetlands (modified from Trepel and Palmeri 2002).
2.3.1THE ROLE OF HYDRAULIC- AND POLLUTANT LOAD
An increase in hydraulic loading rate (HLR), which is the water volume entering the CW divided by the CW surface area, typically also increases the amount of pollutants passing through the system. Thus, pollutant accessibility to microorganisms and the overall CW media increases, resulting in high absolute mass removal per unit time.
However, at the same time the contact time between wastewater pollutants and the critical removal components of the CW is shortened (compared to a lower HLR), resulting in a lower relative pollutant removal of incoming concentration or mass. In fact, numerous studies have demonstrated that an increase in HLRs, and thus higher pollutant mass loading rates, usually give higher area-specific mass removal rates in CWs (Greenway and Woolley 1999; Lin et al. 2002; Jing et al. 2002; Kadlec 2005a;
Kadlec 2005b), however at the expence of higher effluent concentrations (Diemont 2006; Kadlec et al. 2010).
Pollutant removal Process scale
geology climate
wetland vegetation wetland hydraulics
hydraulic & pollutant load pH
redox status
season
land use
nutrient availability Wetland scale
upstream activity Landscape scale
Lin et al. (2002) described increased absolute mass removal of P ranging from 0.06 to 0.14 g m-2 day-1 and for NH4+
-N from 0.04 to 0.09 g m-2 day-1 with increasing HLRs (18–68 mm day-1) in a planted FWS CW receiving aquaculture wastewater.
Nevertheless, at the highest HLR (135 mm day-1) the removal rates decreased to 0.12 g P m-2 day-1 and 0.08 g NH4+
-N m-2 day-1, proposing that there could be an optimal level of HLR to achieve maximum pollutant removal. Also, Jing et al. (2002) reported increased absolute mass removal rate of PO4-3 and NH4+-N with increasing mass loading rates for CWs operating at a hydraulic residence time (HRT) of 2 to 4 days.
Still, at a HRT of 1 day (HLR of 120 mm day-1) mass removal rates decreased noticeably, which the authors concluded was an effect of insufficient time for removal processes to occur. However, Lin et al. (2002) and Jing et al. (2002) applied the different HLRs to the same CW system subsequently, thus possibly introducing interpretation problems due to effects such as the “aging phenomena” as described by Kadlec (1984), which may cause exhaustion of the CW sediment P and NH4+
-N adsorption capacity. Also, the development of anaerobic conditions in the CW sediment/soil due to accumulation of dead plant material during the experiment could cause solubilization of P precipitates and lower nitrification rates. Also, the studied CWs were microcosms, thus, increasing the risk for not adequately imitating the realistic environmental conditions. Hence, incorrect conclusions about the most favorable HLR giving maximum mass removal in the CWs could have been made by Lin et al. (2002) and Jing et al. (2002). Still, Martín et al. (2013) studied the effect of different HLRs ranging from 19 to 130 mm day-1, which were applied subsequently to a large FWS CWs (9 ha) over a period of 2 years. Results from the cited study showed that removal of TP (0.0031–0.025 g m-2 day-1) and TSS (0.344–2.41 g m-2 day-1) increased with increasing HLR, without any noted decrease at the highest HLR (which was applied last). Still, the TP mass loading rates were somewhat lower than in Lin et al. (2002) and Jing et al. (2002) which could have affected the level of achieved
“aging”. Nevertheless, despite the large body of research, at present a broad generalization of finding optimum HLR for achieving maximum pollutant removal in CWs is not possible. Ultimately, the CW designer has to decide whether certain concentration criteria are to be met or a high mass removal is desired. Moreover, nutrient mass load to a CW may also strongly affect macrophyte nutrient uptake rates and thus regulate the amount of nutrients that can be removed via macrophyte harvesting (Kadlec and Wallace 2009). In more detail, nutrient concentration of several emergent macrophyte species, show increasing tissue P with increasing P load, stretching from 0.05–0.5% P per DW biomass (Kadlec and Wallace 2009). Also, normally, macrophyte standing crop (total amount of dead or living biomass found at any given moment in a CW) increases with increasing P load, extending from about 1000 g DW biomass m-2 (low nutrient availability) to 6000 g DW biomass m-2 (high nutrient availability; Kadlec 2005a).
2.3.2THE ROLE OF WETLAND VEGETATION
Generally, rooted emergent macrophytes assist pollutant removal in CWs through both indirect and direct mechanisms. The indirect mechanisms are related to physiological, morphological and biogeochemical factors aided by macrophytes to the CW matrix.
More specifically, macrophytes may indirectly facilitate removal processes by providing surfaces for active and diverse microbial communities (Eriksson and Weisner 1996; Eriksson and Andersson 1999; Bastviken et al. 2003; Rossmann et al.
2012), by preventing resuspension of sedimented pollutants (Braskerud 2001) and supplying oxygen to the CW soil (Reddy et al. 1989; Sorrell and Armstrong 1994; Brix 1997). The latter process occurs via aerenchymatous stem tissue which transports oxygen to roots rhizomes from where the oxygen may eventually diffuse to the CW sediment and pore water. Thus, it is not surprising that numerous studies performed on the CW scale, that have compared non-vegetated with vegetated CWs, report significantly higher removal of wastewater pollutants in the latter systems, and that even macrophyte species can be of significance (Klomjek and Nitisoravut 2005;
Diemont 2006; Yang et al. 2007; Brisson and Chazarenc 2009; Rossmann et al. 2012;
Gagnon et al. 2012).
The direct pollutant removal mechanism provided by macrophytes is the assimilation of nutrients from wastewater into the macrophytic biomass. Looking at P, Kadlec (2005a) states that only 10–20% of the assimilate P is permanently stored as residuals from decomposition processes while the rest is returned to the system upon vegetation senescence (Kadlec 2005a). Thus, the only way to ultimately remove P trapped in the macrophytic biomass is through harvesting. For temperate climates, Kadlec (2005a) stated that macrophyte uptake in emergent FWS CWs was between 0.0014 and 0.055 g P m-2 day-1 and suggested that 20% of the uptake could be considered as a net removal achieved through above-ground biomass harvesting. This gives a net removal of P via harvesting of less than 0.011 g P m-2 day-1 for emergent FWS CWs in temperate climates. However, higher macrophyte uptake rates have been reported for tropical CWs receiving wastewater. Okurut (2001) reported that Cyperus papyrus L. growing in CWs receiving 0.26 g P m-2 day-1 had an uptake rate of 0.024 g P m-2 day-1. Kyambadde et al. (2004) stated uptake rates by C. papyrus of 0.090 g P m-2 day-1 at a load of 0.62 g PO43--P m-2 day-1. Later, Kyambadde et al. (2005) demonstrated even higher P uptake rates by C. papyrus in the order of 0.18 g P m-2 day-1 at a load of 2.2 g PO43--P m-2 day-1. Thus, the amount of area-specific P removed via harvesting of tropical CWs could be considerably higher than that reported by Kadlec (2005a). This is not surprising since the above-ground biomass of various macrophytes grow and die at quicker cycles in tropical climates compared the single annual growing season in northern climates (Reddy et al. 1999). Also, the seasonal translocations of nutrients from above-ground to below-ground macrophyte parts are less pronounced in tropical regions (Vymazal 2007) which creates the possibility to implement CW management strategies that include multiple harvests per year. This type of management technique in CWs can possibly play a significant role for the removal and reclamation route for nutrients in wastewater.
Nevertheless, the literature contains conflicting data about the significance of emergent macrophytes as nutrient traps in wastewater receiving CWs. Generally, harvesting macrophytes for nutrient removal is not valued as significant due to the poor nutrient uptake rates by macrophytes when related to the high nutrient loads normally found in wastewater treating CWs (Brix 1997; Kadlec 2005a; Kadlec 2005b). Looking at P, generally, less than 10% of the P loaded to CWs can be found in harvested biomass (Kim and Geary 2001; Okurut 2001; Toet et al. 2005; Kyambadde et al. 2005). Still, when relating the P found in macrophytic biomass to the total removal of P by a FWS CW a different picture emerges.
Studies from tropical CWs have reported that the fraction of total P removal by emergent macrophyte can be in the order of 35–90% of the total P removed by a CW if the macrophytes are at an exponential growth stage (Okurut 2001; Kyambadde et al.
2004). Still, some explanations for the high macrophyte uptake rates reported by Kyambadde et al. (2005) and Kyambadde et al. (2004), other than the tropical conditions, could be the substrate-free design and the fact that the plants were only 7 to 8 months old at the end of the studies.
The lower macrophyte uptake rates reported by Okurut (2001) compared to the ones obtained by Kyambadde et al. (2005) were most likely related to lower mass load into the CWs in the former study. Also, in the study by Okurut (2001) higher macrophyte age and thus the senescent stage of the macrophytes at the end of the 18 months long study period may have resulted in lower macrophyte P uptake. Still, Martín et al.
(2013) reported that P accumulated in above-ground biomass of macrophytes growing in FWS CWs in Spain represented around 20% of the total P removed by the system during at the end of a 2 year study period. In the cited study, macrophyte P uptake after a 2 year growing period was 0.019 g P m-2 day-1. Thus, studies show that macrophyte P uptake may be a major component of total P removal by CWs and that macrophyte growth stage (age) may be an important factor for the assessment of macrophyte nutrient uptake in relation to overall CW mass load and removal.
Correct vegetation management is critical for achieving sustainable optimal water treatment functions in CWs (Thullen et al. 2005). This may not be surprising since macrophyte turnover rates in aquatic ecosystems can augment bioavailable P to the water column (Hansson and Granéli 1984) and thereby influence overall nutrient removal in CWs in a negative way (Lü et al. 2012). Many researchers have signified the importance of regular macrophyte harvesting in CWs to achieve efficient removal of nutrients from wastewater (DeBusk and Ryther 1987; Okurut 2001; Kyambadde et al. 2005; Bojcevska 2007; Vymazal 2007; Borin and Salvato 2012; Hoffmann et al.
2012). Still, Martín et al. (2013) reported a period of degraded water quality after macrophyte harvest. Nevertheless, regular and proper harvesting methods may keep macrophytes at optimum growth stage and thus ensure optimal nutrient uptake (Vymazal 2007), since P concentrations in above-ground macrophyte biomass declines with increasing growing time. This decline can be substantial, i.e. for Phragmites australis (Cav.) Steud., per cent P of DW biomass can be 0.47% at day 30 and decrease to 0.17% at day 180 of growth (Kadlec and Wallace 2009). Thus, an efficient harvesting frequency may also aid in the maintenance of a high quality standing crop of young and actively growing macrophyte shoots. This way, nutrient uptake and storage by macrophytes growing in CWs can be nearly continuous and represent a significant route for nutrient removal from wastewater. Also, through harvest, vital nutrients such as phosphorus may be recycled as fertilizers, decreasing the need for mined phosphorus, which is not a sustainable source of this vital nutrient when looking at long-term geological, economic and geopolitical aspects (Elser 2012). Thus, for developing countries, the harvested biomass could represent a valuable resource for local communities (Abila 1998; Gichuki et al. 2001).
In a review of 35 studies on effects of macrophyte species on removal efficiency in CWs, Brisson and Chazarenc (2009) concluded that species identity does matter, but that making well-founded recommendations for species selection was difficult.
Okurut et al. (1999) and Kyambadde et al. (2005) reported that CWs planted with C.
papyrus displayed significantly higher area-specific mass removal rates of NH4+
-N than those planted with other macrophyte species.
Additionally, characteristics in macrophyte structures could influence nitrification rates and thus removal of NH4+-N in CWs. Eriksson and Andersson (1999) reported that the biofilm nitrification varied greatly between litter of different macrophytes and suggested that the difference in biofilm nitrification among the species was a result of differences in the physical and chemical characteristics of the litter. In addition, substances released from the decomposing macrophytes could have an inhibiting effect on the nitrifying bacteria and their activity which in turn may promote more or less biofilm development and hence also nitrifying bacteria (Bastviken et al. 2003).
2.3.3THE ROLE OF WETLAND HYDRAULICS
Theory. Hydraulics, which describes water flow movements in both time and space, is recognized as one of the key factors regulating water quality improvements in CWs (Headley and Kadlec 2007; Kadlec and Wallace 2009; Kusin et al. 2010). Also, a more comprehensive understanding of CW hydraulics is required before we accurately can predict treatment performance. The significance of CW hydraulics becomes obvious when considering that the starting point for many pollutant removal processes is the contact between the wastewater and the surfaces in the CW, i.e. vegetation, detritus and substrate. Therefore, understanding the characteristics of hydraulics is the key to designing optimized CWs with respect to both efficient wastewater treatment and land utilization. One way to study CW hydraulics is through empirical tracer studies. The fundamental idea of tracer studies is that an inert tracer, working as an indicator of water movement, is instantaneously injected at the CW inlet. After injection, measurements of the tracer concentration are made at the CW outlet at certain time intervals. By plotting the measured outlet tracer concentration against time when taken, the water residence time distributions (RTD) of the CW can be illustrated (Fig.
4) and specific hydraulic parameters may be quantified using different data analysis methods.
Figure 4. A residence time distribution (RTD) curve from a hydraulic tracer experiment performed in a constructed wetland. Dots show measured tracer concentration data at wetland outlet and the line shows a fit of the measured data using a mathematical model.
0.0 0.2 0.4 0.6 0.8 1.0 1.2
0.0 1.0 2.0 3.0 4.0
tracer concentration (mg L-1)
time (h)
The basic hydraulic parameters encountered in the CW literature are the mean (tm) and variance (2) of the RTD. The tm parameter indicates the average residence time of water in the CW and 2 proposes the degree of water dispersion (Fogler 2006).
The tm-value may be related to the theoretical or nominal residence time, tn-value (defined as the theoretical CW volume divide by the flow) to quantify the effective or
“active” CW volume (Thackston et al. 1987) as in Eq. (1)
sys effective n
m
V V t
et (1)
where e = effective CW volume ratio (-), Vsys = theoretical CW volume (m3) and Veffective = total effective CW volume (m3) derived via tracer data. If e = 1 this indicates that the entire CW volume is active in pollutant removal, i.e. 100% effective volume.
However, numerous studies show that e-values of CWs usually are in the range of 0.20 to 0.98 with a typical mean value of 0.82 (±0.8 SD) for FWS CWs (Kadlec and Wallace 2009). Interestingly, e-values higher than 1 have also been reported in the literature and have mainly been associated to flaws in water flow measurements during the tracer experiment or estimations of the theoretical CW volume (Kadlec 1994a;
Kjellin et al. 2007).
The 2 parameter may be related to the tm parameter to give a dimensionless characterization of water dispersion within a CW. The resulting parameter is called the number of tanks (N) in the first-order tanks-in-series (TIS) model and may be defined as (2a; Kadlec and Wallace 2009)
2 2
mN t
(2a) where N = number of tanks in TIS model, tm2
= squared mean hydraulic residence time (h2) and 2 = variance of RTD (h2).
Also, another expression of the number of tanks (N) as in Eq. (2b) may be found in the literature (Kadlec and Knight 1996; Persson et al. 1999; Kadlec and Wallace 2009).
However, Kadlec and Knight (1996) stated that Eq. (2b) should not be used to calculate N if the RTD is characterized by a broad crest, since this normally means that the tp-value, which is the time at which the peak tracer concentration occurs, is hard to define correctly.
p m
m
t t N t
(2b)If N
, this is indicative of plug-flow conditions, which is considered to be optimal for wastewater treatment in CWs. Under plug-flow conditions, water parcels entering at the same time will move through the entire CW volume at the same velocity and reach the outlet simultaneously.However, it has been well established that actual CWs experience large deviation from plug-flow conditions, as indicated by N-values ranging between 0.3 to 11 with a typical mean value of 4.1 (±0.4 SD) for FWS CWs (Kadlec and Wallace 2009).
In addition, to provide a more general measure of hydrodynamic conditions in CWs, Persson et al. (1999) proposed the hydraulic efficiency parameter () which includes both the effective volume factor (e-value) and the dispersion factor (N-value) as in Eq.
(3)
n p m
p m n m
t t t
t t t t
e N
1 1
1 (3)
where = hydraulic efficiency; tp = time at which the peak tracer concentration occurs [h] and tn = theoretical residence time of CW. The foremost advantage of using
is that it can be determined directly from the peak value of the RTD curve, thus omitting uncertainties related to RTD truncation method in order to estimate tm
(Persson et al. 1999). Despite the simplicity of calculating the parameter, it is not as widely used as the e- or N-parameter to describe CW hydrodynamics. Nevertheless, reported values of are in the range of 0.08 to 0.76 (Holland et al. 2004; Min and Wise 2009; Lange et al. 2011; Kusin et al. 2012). In addition, the term hydraulic efficiency has in some studies been used to describe the effective volume (Martinez and Wise 2003; Wang et al. 2006), thus, adding unnecessary confusion to the terminology used to describe CW hydraulics.
Methodological aspects of tracer studies. An important aspect related to the success of a tracer study is that the tracer concentration after injection is above the detection limit of the analytical method used for estimation of the tracer concentration. This is very important since, the tracer concentration measurements together with measurements of water outflow will yield the tracer mass recovery. High tracer mass recoveries (80–100% relative to injected tracer mass) serve as indicators of successful CW hydraulic tracer studies (Kadlec and Wallace 2009). Still, one major obstacle for obtaining a high tracer recovery may be related to the density of the tracer solution that is dosed into the CW. Usually, the tracer solution, with a certain Cdose, will have a slightly higher density than the wastewater residing in the CW. However, since most tracer tests in CWs are conducted with the pulse injection method (which implies that tracer solution is poured into or near the CW inlet all at once), a high density of the tracer solution may cause the tracer pulse to sink to the CW bottom. As a result, a top- to-bottom density stratification between the heavier tracer solution and the CW water may trap a portion of tracer mass at the CW bottom. This type of scenario may then result in retardation of the tracer flow and/or stagnation of the tracer in depressions zones and could thus have a distorting effect on results obtained from tracer studies (Schmid et al. 2004; Headley and Kadlec 2007; Kadlec and Wallace 2009). It is therefore recommended that the density of the tracer solution should be within 1% of the density of the CW water. Still, methodological problems such as those caused by density effects remain particularly problematic at low flow conditions such as those in CWs (Headley and Kadlec 2007), and thus, more studies are needed on how the density stratification may be minimized. For CW with point discharges, Headley and Kadlec (2007) recommend that the tracer is injected at or near the inlet pipe where there is some turbulence.
