PLEASE SCROLL DOWN FOR ARTICLE

This article was downloaded by: [Tang, Xianqiang]On: 1 May 2010Access details: Access Details: [subscription number 921807495]Publisher Taylor & FrancisInforma Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Environmental TechnologyPublication details, including instructions for authors and subscription information:http://www.informaworld.com/smpp/title~content=t791546829

Nutrient removal as a function of benzene supply within vertical-flowconstructed wetlandsXianqiang Tang a; Miklas Scholz b; Paul Emeka Eke b;Suiliang Huang a

a College of Environmental Science and Engineering, Nankai University, Tianjin, 300071, P.R. China b

Institute for Infrastructure and Environment, School of Engineering and Electronics, The University ofEdinburgh, Edinburgh, Scotland, UK

Online publication date: 30 April 2010

To cite this Article Tang, Xianqiang , Scholz, Miklas , Eke, Paul Emeka andHuang, Suiliang(2010) 'Nutrient removal as afunction of benzene supply within vertical-flow constructed wetlands', Environmental Technology, 31: 6, 681 — 691To link to this Article: DOI: 10.1080/09593330903530793URL: http://dx.doi.org/10.1080/09593330903530793

Full terms and conditions of use: http://www.informaworld.com/terms-and-conditions-of-access.pdf

This article may be used for research, teaching and private study purposes. Any substantial orsystematic reproduction, re-distribution, re-selling, loan or sub-licensing, systematic supply ordistribution in any form to anyone is expressly forbidden.

The publisher does not give any warranty express or implied or make any representation that the contentswill be complete or accurate or up to date. The accuracy of any instructions, formulae and drug dosesshould be independently verified with primary sources. The publisher shall not be liable for any loss,actions, claims, proceedings, demand or costs or damages whatsoever or howsoever caused arising directlyor indirectly in connection with or arising out of the use of this material.

Environmental Technology

Vol. 31, No. 6, May 2010, 681–691

ISSN 0959-3330 print/ISSN 1479-487X online© 2010 Taylor & FrancisDOI: 10.1080/09593330903530793http://www.informaworld.com

Nutrient removal as a function of benzene supply within vertical-flow constructed wetlands

Xianqiang Tang

a

, Miklas Scholz

b

*, Paul Emeka Eke

b

and Suiliang Huang

a

a

College of Environmental Science and Engineering, Nankai University, Tianjin, 300071, P.R. China;

b

Institute for Infrastructure and Environment, School of Engineering and Electronics, The University of Edinburgh, William Rankine Building,

The King’s Buildings, Edinburgh EH9 3JN, Scotland, UK

Taylor and Francis

(

Received 7 May 2009; Accepted 4 December 2009

)

10.1080/09593330903530793

The role of benzene, macrophytes and temperature in terms of nutrient removal within constructed wetlands isunknown. Therefore, a research study over approximately 30 months was conducted to assess the potential ofvertical-flow constructed wetlands to treat nutrients and to examine the effect of benzene concentration, presence of

Phragmites australis

(Cav.) Trin. ex Steud (common reed), and temperature control on nutrient removal.Experimental wetlands removed between 72% and 90% of benzene at an influent concentration of 1000 mg L

−

1

. Astatistical analysis indicated that benzene is linked to increased effluent chemical oxygen demand and biochemicaloxygen demand concentrations. However, there was no significant relationship between benzene treatment and bothnitrogen and phosphorus removal.

Phragmites australis

played a negligible role in organic matter (chemical oxygendemand, biochemical oxygen demand, nitrogen and phosphorus) removal. Control of temperature favouredbiochemical oxygen demand removal. However, no significant difference in chemical oxygen demand, and nitrogenand phosphorus removal was detected. Only the combination of the benzene and temperature variables had asignificant impact on biochemical oxygen demand removal. The effluent biochemical oxygen demand concentrationsin temperature-controlled benzene treatment wetlands were much lower than those located in the naturalenvironment. However, any other combination between benzene,

P. australis

and the environmental control variableshad no significant effect on biochemical oxygen demand, chemical oxygen demand, or nitrogen and phosphorusremoval.

Keywords:

vertical-flow constructed wetland; nutrients; benzene; temperature;

Phragmites australis

Introduction

Constructed wetlands are engineered systems designedand constructed to utilize the natural processes involvingwetland vegetation, soils and their associated microbialassemblages to assist in treating wastewater [1,2]. Thesesystems take advantage of physical, chemical andbiological processes occurring in natural wetlands, butdo so within a semi-controlled environment [3]. Previousliterature has claimed that organic matter represented bychemical oxygen demand (COD) and biochemicaloxygen demand (BOD), suspended solids, nitrogen,phosphorus and bacteria can be treated well by wetlandsystems [1,4–7].

Constructed wetlands were successfully studied forwater purification purposes in the 1960s [8];

Scirpuslacustris

L. (common bulrush) removed organicsubstances such as nutrients and even toxic chemicalssuch as phenols. Thereafter, an increasing number ofconstructed wetlands were applied for the treatment ofmunicipal, industrial and agriculture wastewater[6,9,10]. A better understanding of processes, such as

adsorption, uptake by plants and living organisms,biodegradation and transformation, and burial, led tothe development of more constructed wetlands used toremove organic matter, nitrogen and phosphorus fromeutrophic river and lake water [2,11]. Good treatmentperformances in terms of petroleum hydrocarbonremoval (typically more than 60% reduction of the totalinput) by constructed wetlands have been recorded[12–15].

Petroleum hydrocarbon wastewaters also containpollutants such as nitrogen and phosphorus and mayhave substantial COD and BOD values [13]. However,the major focus of the petroleum industry is on assess-ing the hydrocarbon removal efficiency. The potentialtoxicity of hydrocarbon pollutants may restrain or killmicroorganisms living in the wetland plant root zoneand filter media [13]. Nevertheless, COD and evenBOD removal efficiencies for wetlands treating toxichydrocarbons are comparable to wetlands treating othertypes of wastewater [13,16]. Nitrogen and phosphorus,which are essential nutrients for successful biodegrada-

*Corresponding author. Email: m.scholz@ed.ac.uk

Downloaded By: [Tang, Xianqiang] At: 02:28 1 May 2010

682

X. Tang

et al.

tion of hydrocarbon pollutants [17], are better removedin wetlands treating petroleum industry wastewater incomparison with wetlands treating other wastewatercategories [13,18].

The role of aquatic plants in pollutant removal byconstructed wetlands has been discussed in detail previ-ously [1,6,10,11]. There is sufficient evidence todemonstrate the importance of wetland plants in remov-ing nutrients [2]. However, previous studies indicatedthat the adverse effects of petroleum hydrocarbon onplants ranged from short-term reductions in photosyn-thesis to mortality [19–21]. A recent research study,which assessed the plant growth performance inconstructed wetlands treating heavy oil-produced water,indicated that the height of planted reeds reduces withan increase in hydrocarbon concentration [16]. Insummary, previous research paid attention to theharmful effect of hydrocarbons on microbial and plantbiomass. However, the role of plants, including macro-phytes, in nutrient removal has been neglected.

Another major research area is focussing on theassessment of the uncertainty of temperature andhumidity on nutrient removal in constructed wetlandstreating hydrocarbons. Previous research indicates thatorganic matter and nitrogen removal predominantlydepends on the temperature-sensitive microbial activ-ity within the root zone [22]. In comparison, phospho-rus removal in different wetland systems is onlyindirectly affected by the temperature-sensitiveoxygen availability, which influences the impact ofthe redox potential on phosphorus availability [23–26]. In wetlands treating hydrocarbons, however, theeffect of temperature on nutrient removal is not aspronounced as in other areas of wastewater treat-ment. Therefore, further research is necessary to betterpredict the impact of environmental boundary condi-tions on nutrient removal in hydrocarbon treatmentwetlands.

This study aimed to assess nutrient removal as afunction of benzene supply within treatment wetlands.In light of the above considerations revealed by theliterature study, research was conducted subject to thefollowing objectives:

●

to assess the potential for removal of low molec-ular weight hydrocarbon, using benzene as a repre-sentative example, in vertical-flow constructedwetlands;

●

to compare the nutrient removal performanceswithin vertical-flow constructed wetlands, whichwere different in terms of contamination, plantdevelopment and temperature;

●

to statistically examine the key and interactiveeffects of zero and high benzene loading, pres-ence or absence of

Phragmites australis

(Cav.)

Trin. ex Steud, and natural and controlledtemperature conditions on ammonia-nitrogen,nitrate-nitrogen and ortho-phosphate-phosphorusconcentrations.

Materials and methods

Design of constructed wetlands

From April 2005 to October 2007, 12 analogous verti-cal-flow constructed wetlands [15] were designed,constructed and operated predominantly to assess nutri-ent removal as a function of benzene supply. All exper-imental wetlands were located at The King’s Buildingscampus at The University of Edinburgh and wereconstructed with polyethylene columns of 100 mm indiameter and 750 mm in height. The outlet valves werelocated at the centre of the bottom plate of each wetlandcolumn, and were connected to 12 mm internal diametervinyl tubing. This arrangement was used for manuallyadjusting the flow and collecting outflow sample waterat the same time. Furthermore, ventilation pipes with aninternal diameter of 13 mm reaching down to 10 mmabove the bottom of each wetland were installed toencourage passive aeration.

The 12 experimental constructed wetlands weregrouped into two groups of six wetlands each: onegroup was located in a temperature-, light- and humid-ity-controlled indoor room, while the other group wasoperated at natural outdoor conditions. The formergroup allows for the generation of a data set suitablefor modelling because data variability is low, consider-ing that weather changes do not affect pollutantremoval. In comparison, the outdoor experiment betterrepresents a real industrial application in a temperateclimate [27] relatively few sunshine hours (mean of1406 hours per annum), low precipitation (mean of 676mm), and mild temperature (mean monthly minimumof 5.1

°

C and mean monthly maximum of 12.2

°

C)despite the relatively small column sizes, which are,however, not uncommon in the scientific literature[1,6,15].

The packing orders of the six indoor and six outdoorwetlands, which were operated in the same way, werethe same. Wetlands 1, 2, 3 and 4 were filled with thefollowing five successive layers of aggregates to a depthof 600 mm: stones (37.5–75.0 mm), large gravel (10.0–20.0 mm), medium gravel (5.0–10.0 mm), small gravel(1.2–5.0 mm) and sand (0.6–1.2 mm). The correspond-ing layer thicknesses were 10, 15, 10, 15 and 10 cm,respectively. Detailed information is outlined in a previ-ous study summarizing findings related to a proof ofconcept experiment [15]. Wetlands 5 and 6 weredesigned as controls and were not filled with gravel orsand. The role of

P. australis

in removing benzene, itsdegradation products and nutrients was also assessed.

Downloaded By: [Tang, Xianqiang] At: 02:28 1 May 2010

Environmental Technology

683

Constructed wetlands 1 and 2 were planted with nineindividual plants of similar biomass and equal strength.

Phragmites australis

was obtained from a local supplier(Alba Trees Public, Gladsmuir, East Lothian, UK).

Control of temperature

The role of benzene supply and temperature on nutrientremoval within treatment wetlands was assessed withthe chosen system set-up (see above). Concerning theindoor wetlands, a Denco Local Environmental ControlUnit supplied by Denco Limited (East Kilbride, UK)was used to control the constant temperature of 15

°

Cand humidity of 60%. The simulation of day and nightcycles was conducted with the help of three plant growthlights (Sylvania 15,000 h, 36 W, 1200 mm, T8 GroluxFluorescent Tube; supplied by Lyco Direct Limited,Bletchley, Milton Keynes, UK). In comparison, nocontrol of any environmental boundary conditions wasattempted for the corresponding outdoor wetlands.

Influent and nutrient supply

Of all petroleum hydrocarbons, benzene is consideredto be the most problematic because of its high toxicityand relatively high water solubility [28], which wasaddressed in this study by thorough mixing of theinflow water for several minutes before introduction tothe wetland rigs. Benzene (BDH analytical reagent,C

6

H

6

(99.7%), supplied by VWR International Limited,Lutterworth, UK) was chosen as a representative targetpollutant to assess the removal of low molecular weighthydrocarbons. Two types of influent were used in abatch flow mode: tap water and tap water spiked withdissolved benzene. Twice per week, wetlands 2, 4 and 6received tap water and nutrients (see below), whilewetlands 1, 3 and 5 received tap water contaminatedwith 1000 mg L

−

1

benzene and nutrients (see below).The selected benzene concentration compares well withthose concentrations quoted by the petroleum industryfor produced waters [12–17].

The well-balanced slow-releasing nitrogen (24%),phosphorus (8%) and potassium (14%) Miracle-Grofertilizer (formerly Osmocote, produced by Scot EuropeB. V., The Netherlands) was used as the predominantsource of nutrient supply to enhance plant and microbialgrowth, and to improve the benzene treatment effi-ciency of the wetland systems. Approximately 8 g of thefertilizer was added directly to all wetlands every twoweeks until 29 May 2006, when the concentration wasincreased to 30 g to assess the effect of increase in nutri-ent concentrations on hydrocarbon removal. From 26June 2006, the amount was lowered to 15 g, which wasseen as a more realistic nutrient concentration for opti-mum benzene removal.

Water sampling and analysis

Since April 2005, all indoor and outdoor wetlands werefully saturated and flooded to a depth of 10 cm abovethe top level of the packing media. The wetlands werefilled with water (i.e. inflow) directly after being fullydrained manually within a short period of less than 10minutes twice per week. A cycle of filling and emptyingof each wetland was performed twice per week. Watersamples were taken for benzene analysis once permonth until January 2007 and then twice per monththereafter as the systems fully matured.

During the entire experimental period, watersamples were collected and analysed twice per week forfive days at 20

°

C nitrogen-allythiourea BOD, COD,ammonia-nitrogen, nitrate-nitrogen, ortho-phosphorus-phosphate, pH, dissolved oxygen (DO) and redoxpotential. Benzene was determined with a PerkinElmer(Beaconsfield, UK) gas chromatograph FID Model9700 and a corresponding headspace sampler HS-101.American standard methods [18] were used for allanalytical work.

Statistical analyses

All statistical tests were performed using the software

Statistical Package for the Social Sciences

[29]. In allcases, significance was defined as

p

< 0.05, if not statedotherwise. One-way analyses of variance (ANOVA)and Duncan’s multiple range tests [11] were carried outto assess the differences between means of benzeneremoval efficiency and effluent nutrient concentrationsin different constructed wetlands. Three-way ANOVAswere applied to examine the influences of benzeneconcentration, plant presence and environmental bound-ary condition control and their interactions with eachother on the nutrient removal efficiencies in the testedwetlands. For all ANOVAs, it was checked that thetested variables were normally distributed. Otherwise,the variables were log

10

-transformed, which was themost suitable transformation function to bring the vari-ance closer to the mean [30].

Results

Benzene removal

The benzene removal for selected indoor and outdoorwetlands, which received 1000 mg L

−

1

benzene in theinflow water, is summarized in Figure 1. The meanreductions for the indoor wetlands were approximately90% (equates to 900 mg L

−

1

or 49 g m

2

d

−

1

), whereasmuch lower removal efficiencies of between roughly72% and 80% (between 720 and 800 mg L

−

1

or between39 and 43 g m

2

d

−

1

) were observed for the outdoorconstructed wetlands (Figure 1).

Downloaded By: [Tang, Xianqiang] At: 02:28 1 May 2010

684

X. Tang

et al.

Figure 1. Mean benzene removal efficiency for experimental constructed wetlands spiked with benzene. Standard deviation bars sharing different letters are significantly different from each other at

p

≤

0.05 according to Duncan’s multiple range tests.

The benzene distribution curves overlapped for theindoor wetlands. The outdoor wetlands 1 and 3 had simi-lar distribution patterns (Figure 2). Furthermore, between20% and 40% of the total benzene effluent concentrationvalues were approximately 0 mg L

−

1

for both the indoorand outdoor wetlands (Figure 2), which indicates thatapproximately 100% removal of benzene can beachieved at a probability of between 20% and 40%.

Figure 2. Distribution of the effluent benzene concentration proportions for outdoor and indoor experimental constructed wetlands spiked with benzene.

Organic matter removal

Effluent concentrations of organic matter in terms ofBOD and COD are shown in Figure 3. Mean effluentorganic matter concentrations in benzene treatmentwetlands were high, and varied between 34.0 and 43.8mg L

−

1

for BOD, and between 199.1 and 453.7 mg L

−

1

for COD. A three-way ANOVA indicated that benzene

treatment is independent of the presence of

P. australis

(Tables 1 and 2). The control of environmental bound-ary conditions was significant for BOD removal at

p

=0.011. Furthermore, the relationship between highbenzene concentrations and the control of environmen-tal boundary conditions on BOD removal was signifi-cant at

p

= 0.043 (Table 1).

Figure 3. Mean effluent (A) biochemical oxygen demand and (B) chemical oxygen demand (COD) concentrations for all experimental constructed wetlands. Standard deviation bars sharing different letters are significantly different from each other at

p

≤

0.05 according to Duncan’s multiplerange tests.

Nitrogen removal

The mean effluent nitrogen concentrations werebetween 26.8 and 48.2 mg L

−

1

for ammonia-nitrogenand between 19.1 and 53.8 mg L

−

1

for nitrate-nitrogen(Figure 4). There was no significant difference in nitro-gen removal based on findings from the one-wayANOVA and Duncan’s multiple range tests (Figure 4).Results obtained by the three-way ANOVA showed no

Figure 1. Mean benzene removal efficiency for experimen-tal constructed wetlands spiked with benzene. Standard devi-ation bars sharing different letters are significantly differentfrom each other at p ≤ 0.05 according to Duncan’s multiplerange tests.

Figure 2. Distribution of the effluent benzene concentrationproportions for all outdoor and indoor experimental con-structed wetlands spiked with benzene.

Figure 3. Mean effluent (A) biochemical oxygen demand(BOD) and (B) chemical oxygen demand (COD) concentra-tions for all experimental constructed wetlands. Standard de-viation bars sharing different letters are significantly differentfrom each other at p ≤ 0.05 according to Duncan’s multiplerange tests.

Downloaded By: [Tang, Xianqiang] At: 02:28 1 May 2010

Environmental Technology

685

significant effect of benzene, plant presence and envi-ronmental boundary condition control, as well as theirinteractions with each other, on either ammonia-nitro-gen or nitrate-nitrogen removal (Tables 3 and 4).

Figure 4. Mean effluent (A) ammonia-nitrogen (NH

4

-N) and (B) nitrate-nitrogen (NO

3

-N) concentrations for all experimental constructed wetlands.

Phosphorus removal

Except for the high mean effluent ortho-phosphate-phosphorus concentration of 23.7 mg L

−

1

obtained forthe outdoor wetland 2, all other wetlands had similarmean effluent phosphorus concentration ranges ofbetween 14.4 and 17.9 mg L

−

1

(Figure 5). Results fromthe three-way ANOVA (Table 5) indicated that benzenewith plant presence, benzene with environmentalboundary condition control, as well as their interactionswith each other, had no significant effect on ortho-phos-phate-phosphorus removal.

Figure 5. Mean effluent ortho-phosphate-phosphorus (PO

43

−

-P) concentrations of all experimental constructed wetlands.

Other water quality variables

Mean values of effluent DO, pH and redox potentialdiffered significantly (

p

< 0.05) between indoor and

outdoor wetlands (Table 6). Multiple comparisonsdetected significantly higher DO concentrations inoutdoor wetlands. Those wetlands treating benzene hadsignificantly lower DO concentrations compared withnon-benzene treatment wetland controls (Table 6).However, the DO values were obtained from samplestaken from the wetland columns and were not directlymeasured within the columns. It is therefore likely thatall DO concentrations provided in Table 6 are slightlyhigher, as a result of contact with the atmosphere, thanthe ones that would be present within the actualwetlands.

The pH values were significantly higher in wetlands1, 3 and 5 than in wetlands 2, 4 and 6. Higher values forthe redox potential occurred in outdoor non-benzenetreatment wetlands.

Discussion

The results obtained in this study indicated that vertical-flow constructed wetlands could remove relativelyhigh benzene concentrations effectively. However, the

Table 1. Results of three-way analyses of variance examining the role of the presence of

Phragmites australis

(Cav.) Trin. exSteud (common reed) (Plant), environmental boundary condition (temperature, light and humidity) control (ENV) and benzeneconcentration on the effluent biochemical oxygen demand.

Source Sum of squares df F-ratio

p

Plant 196.049 1 0.893 0.345ENV 1458.880 1 6.649 0.011Benzene 53411.130 1 243.422 <0.001Plant and ENV 11.541 1 0.053 0.819Plant and benzene 62.501 1 0.285 0.594ENV and benzene 907.864 1 4.138 0.043Plant, ENV and benzene 96.251 1 0.439 0.508

Error 52660.339 240 n/a n/a

df: degree of freedom; F-ratio: test statistic used to decide whether the sample means are within the sampling variability of each other;

p

:

p

-valuein the analysis of variance table, which gives an overall confidence for the fit to be rejected; n/a: not applicable.

Table 2. Results of three-way analyses of variance examining the role of the presence of

Phragmites australis

(Cav.) Trin. exSteud (common reed) (Plant), environmental boundary condition (temperature, light and humidity) control (ENV) and benzeneconcentration on the effluent chemical oxygen demand.

Source Sum of squares df F-ratio

p

Plant 0.065 1 0.065 0.694ENV 0.888 1 0.888 0.145Benzene 44.625 1 44.625 <0.001Plant and ENV 0.003 1 0.003 0.937Plant and benzene 0.068 1 0.068 0.685ENV and benzene 0.126 1 0.126 0.581Plant, ENV and benzene 0.053 1 0.053 0.720

Error 99.560 240 n/a n/a

df: degree of freedom; F-ratio: test statistic used to decide whether the sample means are within the sampling variability of each other;

p

:

p

-valuein the analysis of variance table, which gives an overall confidence for the fit to be rejected; n/a, not applicable.

Downloaded By: [Tang, Xianqiang] At: 02:28 1 May 2010

686

X. Tang

et al.

concentration-based removal efficiencies of between72% and 90% were lower than the 90% to 93% rangereported elsewhere [12,16]. This might be due to the

relatively high influent benzene concentration of 1000mg L

−

1

in this study compared with ranges between 20and 60 mg L

−

1

reported in the referenced literaturesources. Nevertheless, a direct comparison is not possi-ble because information on loading rates, climaticconditions and operational details are partly missing.

Results obtained from multiple comparisons impliedthat the control of temperature significantly influencedhydrocarbon removal (Figure 1). Similar observationswere also reported elsewhere. For example, the hydro-carbon treatment efficiency was significantly lower inwinter than in spring and summer [12]. Approximately15

°

C is the lower temperature threshold for effectiveorganic compound removal [1].

This study determined the factors affecting nutrientremoval in benzene treatment wetlands. The applicationof a slow releasing fertilizer (see above) contributed todetectable small additional amounts of effluent BODand COD in wetlands 2, 4 and 6. In contrast, benzeneresidual and its biodegradation products were the mostimportant sources of carbon contributing to highconcentrations of organic matter in the treatmentwetlands (Figure 3), considering that the tap water wasvirtually free of BOD. The effluent DO and BODconcentrations in benzene treatment wetlands weresignificantly lower and higher, respectively, than thosein non-benzene treatment wetland controls (Figure 1).Consumption of DO and a corresponding increase inBOD effluent concentrations compared with non-benzene treatment wetlands indirectly indicated biodeg-radation of benzene. This finding was confirmed by anobservation made previously [12], showing that biodeg-radation was responsible for nearly 80% of hydrocarbonreduction, if the corresponding influent concentrationwas <100 mg L

−

1

.Unlike organic matter removal, the effect of

benzene treatment on nitrogen and phosphorus removalwas not significant in the wetlands governed by anaero-bic and aerobic processes (Tables 3 to 5). Aerobic

Figure 4. Mean effluent (A) ammonia-nitrogen (NH4-N)and (B) nitrate-nitrogen (NO3-N) concentrations for all exper-imental constructed wetlands.

Table 3. Results of three-way analyses of variance examining the role of the presence of

Phragmites australis

(Cav.) Trin. exSteud (common reed) (Plant), environmental boundary condition (temperature, light and humidity) control (ENV) and benzeneconcentration on the effluent ammonia-nitrogen.

Source Sum of squares df F-ratio

p

Plant 0.142 1 0.458 0.499ENV 0.002 1 0.006 0.936Benzene 0.143 1 0.460 0.498Plant and ENV 0.010 1 0.031 0.860Plant and benzene 0.061 1 0.198 0.657ENV and benzene 0.036 1 0.115 0.735Plant, ENV and benzene 0.003 1 0.011 0.918

Error 57.187 240 n/a n/a

df: degree of freedom; F-ratio: test statistic used to decide whether the sample means are within the sampling variability of each other;

p

:

p

-valuein the analysis of variance table, which gives an overall confidence for the fit to be rejected; n/a, not applicable.

Downloaded By: [Tang, Xianqiang] At: 02:28 1 May 2010

Environmental Technology

687

conditions are likely to occur in the top water zone, andduring and directly after draining of the wetlands. Thisfinding is in good agreement with previous studies [13],but does conflict with biodegradation research [31]

showing that nitrate-nitrogen was a favourable electronacceptor for benzene reduction.

An inhibitory effect of nitrate on hydrocarbon degra-dation might have occurred. Adequate fertilizer concen-trations increase the biodegradation rate, whereasexcessive fertilization often has a negative effect [1,23].However, the nitrogen compounds within the wetlandcolumns cannot be qualified and quantified because theyare directly related to the fertilizer used. The release rateof nitrogen from the fertilizer is also unknown becauseit is a function of the specific environmental conditionswithin each column.

The relative unimportance of benzene on nitrogenand phosphorus removal in this study could mean thefollowing: the effective removal of hydrocarbon mightreduce its adverse effect on nutrient removal. Further-more, negligible nitrogen and phosphorus concentra-tions were consumed during the process of benzenebiodegradation.

The role of aquatic plants and their effect on nutrientremoval in hydrocarbon treatment wetlands need to bediscussed. One-way and three-way ANOVA found that

Table 4. Results of three-way analyses of variance examining the role of the presence of

Phragmites australis

(Cav.) Trin. exSteud (common reed) (Plant), environmental boundary condition (temperature, light and humidity) control (ENV) and benzeneconcentration on the effluent nitrate-nitrogen.

Source Sum of squares df F-ratio

p

Plant 0.623 1 1.013 0.316ENV 0.127 1 0.207 0.650Benzene 0.195 1 0.317 0.574Plant and ENV 0.323 1 0.525 0.470Plant and benzene 0.102 1 0.166 0.316ENV and benzene 0.006 1 0.010 0.920Plant, ENV and benzene 0.005 1 0.008 0.927

Error 113.213 240 n/a n/a

df: degree of freedom; F-ratio: test statistic used to decide whether the sample means are within the sampling variability of each other;

p

:

p

-valuein the analysis of variance table, which gives an overall confidence for the fit to be rejected; n/a, not applicable.

Table 5. Results of three-way analyses of variance examining the role of the presence of

Phragmites australis

(Cav.) Trin. exSteud (common reed) (Plant), environmental boundary condition (temperature, light and humidity) control (ENV) and benzeneconcentration on the effluent ortho-phosphate-phosphorus.

Source Sum of squares df F-ratio

p

Plant 0.446 1 3.653 0.058ENV 0.039 1 0.322 0.571Benzene 0.018 1 0.146 0.703Plant and ENV 0.072 1 0.593 0.442Plant and benzene 0.052 1 0.424 0.516ENV and benzene <0.001 1 0.000 0.985Plant, ENV and benzene 0.008 1 0.069 0.794

Error 22.459 240 n/a n/a

df: degree of freedom; F-ratio: test statistic used to decide whether the sample means are within the sampling variability of each other;

p

:

p

-valuein the analysis of variance table, which gives an overall confidence for the fit to be rejected; n/a: not applicable.

Figure 5. Mean effluent ortho-phosphate-phosphorus(PO4

3−-P) concentrations of all experimental constructed wet-lands.

Downloaded By: [Tang, Xianqiang] At: 02:28 1 May 2010

688

X. Tang

et al.

P. australis

had no significant effect on organic matterremoval (Figure 3, Tables 1 and 2). Similar removalfigures for BOD and COD were reported for plantedand unplanted wetlands treating other wastewater types[32,33].

A statistical analysis indicated that

P. australis

hadno great importance on nitrogen and phosphorusremoval in benzene treatment wetlands (Figures 4 and5, Tables 3 to 5). This observation was in agreement withother findings [34] but in contrast to results reportedelsewhere [2,33], which indicated that wetland plantsplayed a significant role in nitrogen and phosphorusremoval. Studies verified that nutrient uptake by plantsis only significant under low nutrient loading conditions[34]. Plant biomass accumulation can account forbetween 5.1% and 26.2% of total nitrogen, and between40.5% and 80.9% of the total phosphorus removal withinfluent mean concentrations of 4.8 mg L−1 for totalnitrogen and 0.2 mg L−1 for total phosphorus [11].Furthermore, the nutrient uptake efficiency for bothabove-ground and below-ground tissues decreases withgreater nutrient availability [35]. In this study, highnitrogen and phosphorus loadings (Figures 4 and 5) wereresponsible for the relative unimportance of the roleplayed by plants in nutrient removal. It is likely that, thehigher the inflow load, the lower the percentage of nutri-ents that can be found within the mature plant biomass.

Environmental boundary condition control wasunimportant for COD removal (Table 2). The removalof COD within constructed wetlands can occur via aero-bic and anaerobic biological processes, as well as by avariety of physical processes including adsorption andfiltration [1]. No significant (p = 0.145) environmental(temperature, light and humidity) sensitivity mightimply that physical rather than biological processeswere responsible for COD removal in benzene treat-ment wetlands. However, statistical analyses indicatedsignificant effects of environmental boundary conditioncontrol on BOD removal (Table 1). Microbial degrada-tion plays the dominant role in BOD removal for

benzene treatment wetlands and other types of waste-water treatment systems [1,33,36]. Reducing the vari-ability of temperature, light and humidity resulted inimproved BOD treatment performances in this study(Figure 3).

However, conflicting opinions concerning thetemperature dependency of BOD removal within treat-ment wetlands can be found in the literature [1]. Thegeneral temperature dependency of BOD with respectto various biological treatment processes has beenreported elsewhere [37]. Similar temperature dependentrelationships have also been outlined for wetlands[22,38]. In contrast, several other studies suggested anegligible effect of temperature on BOD removal inwetlands [13,23,39].

The redox potential has been closely linked to lightconditions, and sufficient light leads to higher corre-sponding redox values [40] because a higher photosyn-thetic rate leads to more oxygen being transported bythe macrophytes into the rhizosphere. Compared withthe indoor wetlands, higher values for the redox poten-tial in the outdoor wetlands (Table 6) indicated thatwetlands operated in natural environments experiencemuch better light conditions than those operated underplant growth lights. Lower effluent BOD concentrationsand redox potential values were recorded for indoorbenzene treatment wetlands, especially for wetlands 1and 3, if compared with the corresponding outdoorwetlands (Figure 3 and Table 6). Low effluent BODconcentrations correlated well with low redox potentialvalues. This confirmed findings showing that highremoval efficiencies for BOD are related to less reduc-ing conditions [29].

For nitrogen, results from a three-way ANOVAindicated no significant dependence on environmentalboundary condition control (Tables 3 and 4). Neverthe-less, many researchers [2,25,40] reported on the sensi-tivity of environmental boundary condition controlwith respect to nitrogen removal within constructedwetlands. Processes such as ammonification,

Table 6. Mean ± standard deviation for effluent dissolved oxygen (DO), pH and redox potential (redox) concerning the indoor(in) and outdoor (out) experimental constructed wetlands.

Variable 1 in 2 in 3 in 4 in 5 in 6 in

DO (mg L−1) 2.7a±1.64 3.1b±1.78 2.7a±1.64 3.3b±1.54 3.2b±1.39 4.9f±1.57pH (-) 6.7d±0.43 6.0b±0.57 6.4c±0.49 5.6a±1.02 6.4c±0.58 6.4c±0.59Redox (mV) 165.4b,c±58.59 166.3b,c±52.08 143.0a±56.21 177.1c,d±55.72 164.6b±53.92 175.8b,c,d±51.92

Variable 1 out 2 out 3 out 4 out 5 out 6 out

DO (mg L−1) 3.9c±2.00 4.5d,e±2.03 4.3c,d±1.80 4.7e,f±2.01 6.2g±2.44 6.3g ±2.36pH (-) 6.7d±0.42 5.6a±1.18 6.7d±0.43 6.0b±0.85 6.8d±0.47 6.6d±0.65Redox (mV) 172.3b,c,d±59.60 191.1e±51.58 168.6b,c±45.87 181.3d,e±48.35 163.5b±51.48 164.5b±53.20

Means marked with different letters are significantly different from each other at p ≤ 0.05 according to the Duncan’s multiple range test. Thesample number for all variables and filters was 193.

Downloaded By: [Tang, Xianqiang] At: 02:28 1 May 2010

Environmental Technology 689

nitrification and denitrification are characterized astemperature-dependent in wetlands [41]. Nitrificationrates are reported to become inhibited at water temper-atures of around 10 °C, and rates drop rapidly to zerobelow approximately 6 °C [42]. Denitrificationprocesses have been observed at temperatures as lowas 5 °C [43]. Compared with the temperature depen-dency, however, previous studies indicate that oxygenavailability is the main variable regulating theprocesses of nitrogen removal [41,44]. Low and insuf-ficient oxygen transport by diffusion and by plantsthrough their aerenchyma (airy tissue allowingexchange of gases between shoots and roots) mayresult in the absence of environmental boundary condi-tion dependency for nitrogen removal (Table 4) [45].

No significant effect of environmental boundarycondition control on phosphorus removal was noted (seealso above). These findings are in disagreement withthose observations made elsewhere [23,25], indicatingthat phosphorus removal was indirectly affected bytemperature through its effects on redox potential levels.Under high phosphorus loading conditions (Figure 5),similar redox potential ranges (143.0 to 177.1 mV and163.5 to 191.1 mV for indoor and outdoor wetlands,respectively) could be responsible for the lack of rela-tionship between environmental boundary conditioncontrol and phosphorus removal [46].

The interactions between benzene, P. australis andenvironmental boundary conditions on nutrientremoval were also assessed in this study. The combi-nation of high benzene concentration and the presenceof environmental boundary condition control had onlya significant effect on BOD removal (Tables 1 and 2).Previous work [13] showed that BOD removal rangesare between 48% and 73% in hydrocarbon treatmentwetlands, and many hydrocarbons are not toxic tomicroorganisms except at high doses. In this study,low BOD effluent concentrations indicate that thehigh benzene concentration of 1000 mg L−1 did nothinder the overall microbial community in degradingBOD.

Conclusions and recommendations for further research

Experimental vertical-flow constructed wetlands can bean effective option for the treatment of low molecularweight hydrocarbons such as benzene. The control ofenvironmental boundary conditions such as tempera-ture, light and humidity resulted in an additional 15%benzene reduction.

No significant effect of benzene treatment on nitro-gen and phosphorus removal was observed. However,this does not mean that these nutrients do not contributeto microbial benzene degradation.

The standard wetland plant Phragmites australis(Cav.) Trin. ex Steud (common reed) played a negligiblerole in organic matter, nitrogen and phosphorus removal.The presence of P. australis and its interactions withbenzene and environmental boundary conditions (seeabove) were not significant in terms of nutrient removal.It follows that a filter filled with aggregates is likely tobe the preferred unplanted wetland option for benzeneremoval, if sufficient nutrients are present.

Concerning the presence of environmental boundarycondition control, nitrogen and phosphorus removalwas not significantly affected by the presence ofbenzene. Benzene is generally toxic to microbial biom-ass responsible for respiration. The BOD is likely to berepresentative for a mature system, where microbesadjusted to the presence of benzene dominate, as wasthe case in this study.

The research highlights the great potential forwetlands to be used in removing wastewaters from theoil industry. This is particularly relevant for countrieswhere sufficient land is available, and where a low-costand low-technology treatment solution is preferred overpotentially unsustainable traditional technologies.Developing countries such as Nigeria may also chooseconstructed wetlands to clean up contaminated river andlake waters.

Further research on nitrogen and phosphorussequestration in plant biomass is encouraged. It wouldalso be valuable to assess the impact of differentbenzene concentrations on plants, and their correspond-ing efficiencies in treating benzene.

AcknowledgementsMr. Tang received a Ph.D. scholarship from the ChinaScholarship Council, National Natural Science Foundation ofChina (No. 50479034). The authors wish to acknowledgefurther funding received from the first and second ExecutiveGovernor of Ebonyi State (Dr Sam Egwu), Prof. Ozo N. Ozoand from the Petroleum Technology Development Fund(Nigeria) for Mr. Eke’s Ph.D. study. The technical assistanceof various visiting researchers and final year project studentsis acknowledged.

References[1] [1] M. Scholz, Wetland Systems to Control Urban Runoff,

Elsevier, Amsterdam, 2006.[2] [2] J. Vymazal, Removal of nutrients in various types of

constructed wetlands, Sci. Total Environ. 380 (2007),pp. 48–65.

[3] [3] D.A. Hammer and R.K Bastian, Wetlands ecosystems:Natural water purifiers? in Constructed Wetlands forWastewater Treatment, D.A. Hammer, ed., LewisPublishers, Chelsea, MI, 1989, pp. 5–19.

[4] [4] K. Sakadevan and H.J. Bavor, Phosphate adsorptioncharacteristics of soils, slags and zeolite to be used assubstrates in constructed wetland systems, Water Res.32 (1998), pp. 393–399.

[5]

Downloaded By: [Tang, Xianqiang] At: 02:28 1 May 2010

690 X. Tang et al.

[5] J. Vymazal, Removal of BOD5 in constructed wetlandswith horizontal subsurface flow: Czech experience,Water Sci. Technol. 40 (1999), pp. 133–138.

[6] [6] M. Scholz, P. Höhn, and R. Minall, Mature experimen-tal constructed wetlands treating urban water receivinghigh metal loads, Biotechnol. Progr. 18 (2002), pp.1257–1264.

[7] [7] H. Wand, G. Vacca, P. Kuschk, M. Krüger, and M.Kästner, Removal of bacteria by filtration in plantedand non-planted sand columns, Water Res. 41 (2007),pp. 159–167.

[8] [8] K. Seidel, Reinigung von Gerwässern durch höherePflanzen, Deut. Naturwiss. 12 (1966), pp. 297–298 (InGerman).

[9] [9] B.C. Braskerud, Factors affecting phosphorus retentionin small constructed wetlands treating agricultural non-point source pollution, Ecol. Eng. 19 (2002), pp. 41–61.

[10][10] L.H. Fraser, S.M. Carty, and D. Steer, A test of fourplant species to reduce total nitrogen and total phos-phorus from soil leachate in subsurface wetland micro-cosms, Bioresour. Technol. 94 (2004), pp. 185–192.

[11][11] L.F. Li, Y.H. Li, D.K. Biswas, Y.G. Nian, and G.M.Jiang, Potential of constructed wetlands in treating theeutrophic water: Evidence from Taihu Lake of China,Bioresour. Technol. 99 (2008), pp. 1656–1663.

[12][12] C. Salmon, J.L. Crabos, J.P. Sambuco, J.M. Bessiere,A. Brasseres, P. Caumette, and J.C. Baccou, Artificialwetland performances in the purification efficiency ofhydrocarbon wastewater, Water Air Soil Pollut. 104(1998), pp. 313–329.

[13][13] R.L. Knight, R.H. Kadlec, and H.M. Ohlendorf, Theuse of treatment wetlands for petroleum industry efflu-ents, Environ. Sci. Technol. 33 (1999), pp. 973–980.

[14][14] S.H. Keefe, L.B. Barber, R.L. Runkel, and J.N. Ryan,Fate of volatile organic compounds in constructedwastewater treatment wetlands, Environ. Sci. Technol.38 (2004), pp. 2209–2216.

[15][15] P.E. Eke and M. Scholz, Benzene removal with verti-cal-flow constructed treatment wetlands, J. Chem.Technol. Biotechnol. 83 (2008), pp. 55–63.

[16][16] G.D. Ji, T.H. Sun, and J.R. Ni, Surface flow constructedwetland for heavy oil-produced water treatment,Bioresour. Technol. 98 (2007), pp. 436–441.

[17][17] J. Cooney, The fate of petroleum pollutants in freshwa-ter ecosystems, in Petroleum Microbiology, R.M. Atlas,ed., Macmillan, New York, 1984, pp. 399–433.

[18][18] APHA, AWWA, and WEF, Standard Methods forthe Examination of Water and Wastewater, 20th ed.,American Public Health Association, Washington,DC, 1998.

[19][19] S.W. Lu, Y.S. Xu, and M.J. Shen, Plant Science,Higher Education Press, Beijing, 1987, pp. 137–161.

[20][20] M.D. Scrimshaw and J.N. Lester, The occurrence andeffects of organic micropollutants in salt marsh sedi-ments with reference to the UK Essex coast: A review,Environ. Technol. 17 (1996), pp. 951–964.

[21][21] S.R. Pezeshki, M.W. Hester, Q. Lin, and J.A. Nyman,The effects of oil spill and clean-up on dominant USGulf coast marsh macrophytes: A review, Environ.Pollut. 108 (2000), pp. 129–139.

[22][22] C. Ouellet-Plamondon, F. Chazarenc, Y. Comeau, andJ. Brisson, Artificial aeration to increase pollutantremoval efficiency of constructed wetlands in coldclimate, Ecol. Eng. 27 (2006), pp. 258–264.

[23][23] R.H. Kadlec and R.L. Knight, Treatment Wetlands,Lewis Publishers, Boca Raton, FL, 1996.

[24][24] H.B. Wittgren and T. Maehlum, Wastewater treatmentwetlands in cold climates, Water Sci. Technol. 35(1997), pp. 45–53.

[25][25] P.A.M. Bachand and A.J. Horne, Denitrification inconstructed free-water surface wetlands II: Effects ofvegetation and temperature, Ecol. Eng. 14 (2000), pp.17–32.

[26][26] V.A. Tsihrintzis, C.S. Akratos, G.D. Gikas, D.Karamouz, and A.N. Angelakis, Performance, and costcomparison of a FWS and a VSF constructed wetlandsystem, Environ. Technol. 28 (2007), pp. 621–628.

[44][27] MET Office, Edinburgh, 1971–2000. http://www.metof-fice.gov.uk/climate/uk/averages/19712000/sites/edin-burgh.html (Accessed 23 February 2010).

[27][28] S.J. Johnson, K.J. Woolhouse, H. Prommer, D.A.Barry, and N. Christofi, Contribution of anaerobicmicrobial activity to natural attenuation of benzene ingroundwater, Eng. Geol. 70 (2003), pp. 343–349.

[28][29] SPSS: Statistical Package for the Social Sciences, SPSSInc., Chicago, IL, 2003.

[29][30] J. García, P. Aguirre, R. Mujeriego, Y.M. Huang, L.Ortiz, and J.M. Bayona, Initial contaminant removalperformance factors in horizontal flow reed beds usedfor treating urban wastewater, Water Res. 38 (2004),pp. 1669–1678.

[30][31] Z.F. Hu, J.F. Dou, X. Liu, X.L. Zheng, and D. Deng,Anaerobic biodegradation of benzene series compoundsby mixed cultures based on optional electronic accep-tors, J. Environ. Sci. 19 (2007), pp. 1049–1054.

[31][32] K.R. Hench, G.K. Bissonnette, A.J. Sexstone, J.G.Coleman, K. Garbutt, and J.G. Skousen, Fate of physi-cal, chemical, and microbial contaminants in domesticwastewater following treatment by small constructedwetlands, Water Res. 37 (2003), pp. 921–927.

[32][33] M.P. Ciria, M.L. Solano, and P. Soriano, Role ofmacrophyte Typha latifolia in a constructed wetlandfor wastewater treatment and assessment of its poten-tial as a biomass fuel, Biosyst. Eng. 92 (2005), pp.535–544.

[33][34] H. Brix, Do macrophytes play a role in constructedtreatment wetlands? Water Sci. Technol. 35 (1997), pp.11–17.

[34][35] G.R. Shaver and J.M. Melillo, Nutrient budgets ofmarsh plants: Efficiency concepts and relation to avail-ability, Ecology 65 (1984), pp. 1491–1510.

[35][36] P.E. Lim, T.F. Wong, and D.V. Lim, Oxygen demand,nitrogen and copper removal by free-water-surface andsubsurface-flow constructed wetlands under tropicalconditions, Environ. Int. 26 (2001), pp. 425–431.

[36][37] Metcalf & Eddy Inc., Wastewater Engineering:Treatment, Disposal, and Reuse, 3rd ed., McGraw-Hill,New York, 1991.

[37][38] S.C. Reed and D.S. Brown, Subsurface flow wetlands –a performance evaluation, Water Environ. Res. 67(1995), pp. 244–248.

[38][39] S. Gerke, L. Baker, and Y. Xu, Nitrogen transforma-tions in a wetland receiving lagoon effluent: Sequentialmodel and implications for water reuse, Water Res. 35(2001), pp. 3857–3866.

[39][40] A. Wießner, U. Kappelmeyer, P. Kuschk, and M.Kästner, Influence of the redox condition dynamics onthe removal efficiency of a laboratory-scale constructedwetland, Water Res. 39 (2005), pp. 248–256.

[40][41] A.G. Werker, J.M. Dougherty, J.L. McHenry and W.A.Van Loon, Treatment variability for wetland wastewa-

Downloaded By: [Tang, Xianqiang] At: 02:28 1 May 2010

Environmental Technology 691

ter treatment design in cold climates, Ecol. Eng. 19(2002), pp. 1–11.

[41][42] J. Herskowitz, S. Black, and W. Sewandowski, Listowelartificial marsh treatment project, in Aquatic Plants forWater Treatment and Resource Recovery, K.R. Reddyand W.H. Smith, eds., Magnolia Publishing, Orlando,FL, 1987, pp. 247–254.

[42][43] S.J. Brodrick, P. Cullen, and W. Maher, Denitrificationin a natural wetland receiving secondary treated efflu-ent, Water Res. 22 (1988), pp. 431–439.

[43][44] A.N. Bezbaruah and T.C. Zhang, Performance of aconstructed wetland with a sulfur/limestone denitrifica-

tion section for wastewater nitrogen removal, Environ.Sci. Technol. 37 (2003), pp. 1690–1697.

[44][45] P. Kuschk, A. Wießner, U. Kappelmeyer, E. Weißbrodt,M. Kästner, and U. Stottmeister, Annual cycle of nitrogenremoval by a pilot-scale subsurface horizontal flow in aconstructed wetland under moderate climate, Water Res.37 (2003), pp. 4236–4242.

[45][46] A. Luo, J. Zhu, and P.M. Ndegwa, SE – structures andenvironment: Removal of carbon, nitrogen, and phos-phorus in pig manure by continuous and intermittentaeration at low redox potentials, Biosyst. Eng. 82(2002), pp. 209–215.

Downloaded By: [Tang, Xianqiang] At: 02:28 1 May 2010