lable at ScienceDirect

Journal of Environmental Management 130 (2013) 324e330

Contents lists avai

Journal of Environmental Management

journal homepage: www.elsevier .com/locate/ jenvman

Comparative performance of free surface and sub-surface flowsystems in the phytoremediation of hydrocarbons using Scirpusgrossus

Israa Abdul Wahab Al-Baldawi a,b,d,*, Siti Rozaimah Sheikh Abdullah a, Fatihah Suja b,Nurina Anuar a, Idris Mushrifah c

aDepartment of Chemical and Process Engineering, Faculty of Engineering and Built Environment, Universiti Kebangsaan Malaysia, Selangor, MalaysiabDepartment of Civil and Structural Engineering, Faculty of Engineering and Built Environment, Universiti Kebangsaan Malaysia, Selangor, Malaysiac Tasik Chini Research Centre, Faculty of Science and Technology, Universiti Kebangsaan Malaysia, Selangor, MalaysiadDepartment of Biochemical Engineering, Al-Khwarizmi College of Engineering, University of Baghdad, Baghdad, Iraq

a r t i c l e i n f o

Article history:Received 18 May 2013Received in revised form15 August 2013Accepted 4 September 2013Available online 8 October 2013

Keywords:PhytotoxicityScirpus grossusFree surface flowSub-surface flowHydrocarbon

* Corresponding author. Department of Chemical aulty of Engineering and Built Environment, UnivSelangor, Malaysia. Tel.: þ60 3 89216407; fax: þ60 3

E-mail addresses: israabal@eng.ukm.my, israauBaldawi), rozaimah@eng.ukm.my (S.R.S. Abdullah).

0301-4797/$ e see front matter � 2013 Elsevier Ltd.http://dx.doi.org/10.1016/j.jenvman.2013.09.010

a b s t r a c t

Two types of flow system, free surface flow (FSF) and sub-surface flow (SSF), were examined to select abetter way to remove total petroleum hydrocarbons (TPH) using diesel as a hydrocarbon model in aphytotoxicity test to Scirpus grossus. The removal efficiencies of TPH for the two flow systems werecompared. Several wastewater parameters, including temperature (T, �C), dissolved oxygen (DO, mgL�1),oxidation-reduction potential (ORP, mV), and pH were recorded during the experimental runs. Inaddition, overall plant lengths, wet weights, and dry weights were also monitored. The phytotoxicity testusing the bulrush plant S. grossus was run for 72 days with different diesel concentrations (1%, 2%, and3%) (Vdiesel/Vwater). A comparison between the two flow systems showed that the SSF system was moreefficient than the FSF system in removing TPH from the synthetic wastewater, with average removalefficiencies of 91.5% and 80.2%, respectively. The SSF system was able to tolerate higher diesel concen-trations than was the FSF system.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

Increasing global population and industrial development havecaused an increase in the amount of wastewater requiring treat-ment. Phytoremediation is a broad term that encompasses the useof plants to reduce contaminants in the environment, such as inwater, soil, or groundwater (Rousseau et al., 2004; Epps, 2006;Vymazal, 2009). The main advantages of phytoremediation overconventional treatment technologies, such as membrane filtration,wet oxidation, agitation in bioreactors, and granular-activatedcarbon adsorption, are that it provides habitats for animals, en-courages biodiversity, and recovers ecosystems that are destroyedby human activity at a site. Phytoremediation can improve the

nd Process Engineering, Fac-ersiti Kebangsaan Malaysia,89216148.km@gmail.com (I.A.W. Al-

All rights reserved.

aesthetics of a treatment site, and provides a cost savings of 50e80% compared to conventional technologies (Ji et al., 2002).Several types of wetlands designs can be used for wastewatertreatment and are classified according to the flow of wastewaterthrough a system. The main types of constructed wetlands are freesurface flow (FSF) and horizontal sub-surface flow (SSF) (USEPA,2000; Fountoulakis et al., 2009a; Homer et al., 2009; Gikas et al.,2013a).

The mechanisms for the removal of organic compounds inconstructed wetlands occur via microbial degradation in therhizosphere, phytovolatilization, sorption, plant uptake, as well asmetabolism, filtration, and sedimentation (Braeckevelt et al., 2011;Imfeld et al., 2009). Both FSF and SSF systems have been used totreat petroleum wastewaters. The first application of FSF to treatpetroleum wastewater was in the early 1970s in the Mandan re-finery, North Dakota, USA (Vymazal, 2010). SSF system wetlandswere initially used in Germany at the Mobil Oil AG terminal inBremen (Wallace et al., 2011). FSF wetlands, also known as openwater wetlands and free water surface wetlands look like natural

Fig. 1. Schematic of the phytotoxicity test: a) Set-up of the aquariums; b) Free surfaceflow system; c) Sub-surface flow system.

I.A.W. Al-Baldawi et al. / Journal of Environmental Management 130 (2013) 324e330 325

marshes or wetland areas with lakes of static water. Constructedwetlands with FSF systems are widely used because of their lowconstruction cost and suitability for wildlife (Fountoulakis et al.,2009a).

There is no general presumption that one type is better thanthe other. However, there have been many studies comparing theperformance of the two systems with different parameters ofchemical oxygen demand (COD), biochemical oxygen demand(BOD5), total suspended solids (TSS), and concentrations of ben-zene, toluene, ethylbenzene, and xylenes, (BTEX). Gikas et al.(2013b) evaluated the purification efficiency of the SSF con-structed wetland treatment process of BTEX in Sternatia di Lecce,Italy. Results show that the percentage removal for BTEX rangingbetween 46% and 55%. Furthermore, Ji et al. (2002) treated heavyoil-produced water from China’s Liaohe Oilfield with an SSF sys-tem. Treatment efficiency was evaluated, and the system demon-strated high mean removal efficiencies of 81%, 89% and 89% forchemical oxygen demand (COD), biochemical oxygen demand(BOD5), and mineral oil, respectively. In addition, Ji et al. (2007)ran reed beds for three years to treat heavy oil-produced waterfrom China’s Liaohe Oilfield water in FSF system-constructedwetlands. Treatment showed mean removal efficiencies of 71%,77%, and 92% for COD, BOD5, and mineral oil, respectively.Comparing these two studies shows that the SSF system was moreefficient than the FSF system in the removal of COD and BOD5,while the two flow systems had the same removal efficiency formineral oil. SSF wetlands are generally used for the treatment ofwastewater in small hamlets as a result of the minimal energyrequired, and since maintenance workers are not needed(Pedescoll et al., 2011; Gessner et al., 2005). The flow in the SSFsystem is horizontal through the supported media and includesaerobic, anoxic, and anaerobic zones. The aerobic zone is in therhizosphere around the roots that provide oxygen to the substratezone (Tee et al., 2012).

A study by the Society of Petroleum Engineers was performedto compare and assess the treatment performances of two full-scale constructed wetlands. The first, in Casper, Wyoming, beganoperations in 2003 and was an SSF system that was designed toremediate BTEX and gasoline-range organics. The removal effi-ciency was 100% with non-detectable concentrations in theeffluent. The second was located in Wellsville, New York and op-erations began operations in 2008. The constructed wetland usedan FSF system, and the removal efficiency was 94% for aniline and93% for nitrobenzene (Wallace et al., 2011). Nazm et al. (2009)compared the two flow systems, and the results showed that thetreatment performance of the horizontal SSF was better than thatof the FSF in regards to the removal of suspended solids and totalCOD, especially at high temperatures. This was because of highalgal growth during the spring months in the FSF, resulting in higheffluent totals of COD and TSS. A comparison between the FSF andSSF systems regarding several factors, including size, cost, andoperability, along with the potential public health issues, wasperformed by Kadlec (2009). Moderate to high removal efficiencyfor BOD, TSS, ammonia, total nitrogen, and phosphorus was foundfor the FSF. There was little to no difference in the area requiredfor the FSF and SSF systems. However, SSF was better than FSF inthe prevention of human health problems, including mosquitocontrol, and minimizing interactions with wildlife. The cost of FSFis 25e50% less than that of SSF; however, there are several factorscan affect this. For example, winter storage or compartmentali-zation, and mosquito and animal control can increase the cost ofan FSF system, but the removal of clogging materials can add tothe cost of an SSF system (USEPA, 2000; Kadlec, 2009).Fountoulakis et al. (2009a) mentioned that the SSF system inconstructed wetland is most widely used in Europe. However, FSF

constructed wetlands are increasingly being preferred becausethey are cheaper to construct and may have higher wildlife habitatvalues. From previous studies, it can be concluded that SSF sys-tems generally perform better. However, there has not been astudy comparing the performance of FSF and SSF systems on TPHremoval. In this study, the phytotoxicity of hydrocarbons on thebulrush plant Scirpus grossus was assessed for the two flow sys-tems: FSF and SSF. The emergent S. grossus bulrush plants wereused to test the ability of this species to enhance the removal ofdiesel contaminants in water through bioremediation (Al-Baldawiet al., 2013a,b). S. grossus is an aquatic species with a high growthrate and is often found growing in large colonies in water (WeedScience Society of America, 2011). The perennial S. grossus plantsare found in tropical and temperate regions, and can grow inwetlands, shallow lakes, and streams (Stottmeister et al., 2003). Itis indigenous to Malaysia and is locally known as Rumput mend-erong (Crop Protection Compendium, 2011). Based on a study byTangahu et al. (2013), S. grossus can phytoremediate wastewatercontaining 50 mgL�1 Pb with a removal efficiency of 99.7%. In thisstudy, a similar species of S. grossus was used to phytoremediatewastewater containing diesel. The objective of this study was tocompare the two flow systems and to determine which systemperforms better in hydrocarbon phytoremediation. The best per-forming system will be implemented in a pilot scale in the futurestudy.

232425262728293031

0 20 40 60 80

T (

C˚)

Day

SSF

1% Diesel

2% Diesel

3% Diesel

3

4

5

6

7

8

0 20 40 60 80

pH

Day

SSF

1% Diesel2% Diesel3% Diesel

2

3

4

5

6

7

8

0 20 40 60 80

DO

( m

g/L

)

Day

SSF

1% Diesel2% Diesel3% Diesel

-60

-20

20

60

100

140

180

220

0 20 40 60 80

OR

P (

mV

)

Day

SSF

1% Diesel

2% Diesel

3% Diesel

232425262728293031

0 20 40 60 80

T(C

º)

Day

FSF

1% Diesel2% Diesel3% Diesel

3

4

5

6

7

8

0 20 40 60 80

pH

Day

FSF

1% Diesel

2% Diesel

3% Diesel

-60-202060

100140180220

0 20 40 60 80

OR

P m

V

Day

FSF

1% Diesel

2% Diesel

3% Diesel

2

3

4

5

6

7

8

0 20 40 60 80

DO

mg/

L

Day

FSF

1% Diesel

2% Diesel

3% Diesel

Fig. 2. Physicochemical parameters of the FSF and SSF systems throughout 72 days of exposure.

I.A.W. Al-Baldawi et al. / Journal of Environmental Management 130 (2013) 324e330326

2. Materials and methods

2.1. Experimental set-up of the aquariums

The experimental set-up was divided into two different types offlow systems, FSF and SSF, which was applied to test the phyto-toxicity of hydrocarbons. This study was conducted in a greenhouseat the Universiti Kebangsaan Malaysia (UKM). All of the aquariumswere operated in a batchwise set-up with a single exposure. Foreach flow system, therewere three aquarium replicates (R1, R2, andR3) for each diesel concentration, 1%, 2%, and 3% (Vdiesel/Vwater)(Fig. 1a). The aquariums’ dimensions were 30 � 30 � 30 cm(L � W � D), and they were each planted with nine healthy S.grossus bulrush plants. The plants had originated from rhizomesone month earlier and were all similar in size at approximately50 cm in height.

Synthetic wastewater was prepared by mixing water withstandard diesel obtained from a local Petronas petrol station atdifferent concentrations, i.e., 1%, 2%, and 3% (Vdiesel/Vwater). The

aquariums for the FSF system were filled with 10 cm of sieved finesand of F 2 mm, were planted with S. grossus and, then, coveredwith synthetic wastewater containing diesel (Fig. 1b). The waterlevel was kept at 10 cm above the sand layer to simulate an FSFsystem that is normally used in a constructed wetland. The supportmedia for the SSF aquariumwas filled from the bottom layer to theupper layer with 8 cm of gravel of size F 10e20 mm, 3 cm of gravelof size F 1e5mm, and 10 cm of sieved fine sand of F 2mm (Fig. 1c).It was then planted with S. grossus and covered with syntheticwastewater containing diesel without supplementation of nutrientor fertilizer. The water level was kept at the upper sand layer tosimulate an SSF system that is normally used in a constructedwetland (EPA, 2002).

2.2. Laboratory tests

The following physicochemical parameters were studied: tem-perature (T, �C), pH, dissolved oxygen (DO, mgL�1), and oxidation-reduction potential (ORP, mV). Values were recorded to observe the

I.A.W. Al-Baldawi et al. / Journal of Environmental Management 130 (2013) 324e330 327

physicochemical changes in the solution. On each sampling day, a100 mL water sample was collected from each aquarium. To mea-sure the physiochemical parameters, an IQ 150 multi-probe (IQScientific Instruments, UK) was used for pH, ORP, and temperaturemeasurements, and a dissolved oxygen sensor (GLI International,Model 63, USA) was also used. Moreover, 100 mL water sampleswere collected on each sampling day (0, 7, 14, 28, 42, and 72 days)and the total petroleum hydrocarbons (TPH) were extracted. Watersamples from the FSF system aquarium were collected from thewater surface using a syringe, and while in the SSF system, thesamples were taken 5 cm from the bottom of the aquarium fromtwo pipes in the aquarium. The TPH concentration in the syntheticwastewater was determined using a liquideliquid extractionmethod and gas chromatography to compare the hydrocarbonremoval efficiency of the two flow systems (Lohi et al., 2008). Thesamples were analysed by a GC flame ionisation detector (GCeFID)using a capillary column (Agilent Technologies, Model 7890A, GCsystem, U.K.) with an HP-5 5% phenyl methyl siloxane column(30 m � 0.32 mm i.d. � 0.25 m) and helium as the carrier gas. Thecolumn temperature was programmed to stay at 50 �C for 1 minand then ramped up by 15 �C per minute to 320 �C for 10 min. Thepercentage of TPH removal on each sampling day was determinedusing Equation (1):

%Removal ¼ TPH0 � TPHt

TPH0� 100 (1)

where:

TPH0 ¼ Total petroleum hydrocarbons on sampling day 0TPHt ¼ Total petroleum hydrocarbons on each sampling day

Table 1Mean values of the physicochemical parameters for the FSF and SSF systems.

Physicochemicalparameters

SSF SSF

Mean S.D. Mean S.D.

T (�C) 27.96 1.30 25.96 1.02pH 5.27 0.79 6.50 0.50DO (mg/L) 4.88 1.02 4.84 1.08ORP (mV) 102.28 48.11 32.18 28.82

2.3. Plant growth with diesel contaminant

The growth of S. grossus in the aquariums with different dieselconcentrations (1%, 2%, and 3%) was observed over 72 days. Oneplant from each aquarium was collected on each sampling day (7,14, 28, 42, and 72 days). The plant was fully rinsed with tap water,thewater was absorbed using a paper towel, and the length and thewet weight were recorded. Then, the dry weight was determinedby drying the entire plant sample in an oven (Memmert, Germany)at 70 �C for 72 h until a constant weight was reached (Peng et al.,2009).

2.4. Statistical analysis

All experimental data were subjected to an analysis of varianceusing SPSS, version 16.0 (SPSS Inc., U.S.A.) Differences in physico-chemical parameters and TPH removal between the FSF and SSFsystems in constructed wetlands were determined byindependent-samples t-test. To evaluate the interaction betweenfactors of the TPH removal efficiency from water, physicochemicalparameters and the growth of the plants, the two-way analysis ofvariance (ANOVA) was usedwith a post hocmultiple comparison ofmeans using the Duncan method at a 95% confidence level orP � 0.05. In addition, the correlations between wet and dry weightwere analysed by Pearson correlations.

3. Results and discussion

3.1. Monitoring physical environmental conditions

The wastewater physicochemical parameters, T, pH, DO, andORP, are depicted in Fig. 2. The differences between the

physicochemical parameters in the FSF and SSF systems wereanalysed using a t-test on the independent samples. The results ofP ¼ 0.000, 0.000, 0.841, and 0.000 for T, pH, DO, and ORP, respec-tively, indicate that the physicochemical parameters between thetwo flow systems were statistically significant (P< 0.05), except forDO.Mean values were found to be significantly different for the twoflow systems for T, pH, and ORP (Table 1). Throughout the 72 days,the wastewaters’ mean temperatures were 27.96 �C and 25.96 �Cfor the FSF and SSF systems, respectively. The low temperature inthe SSF system was due to treatment occurring in the substrate;while in the FSF system, treatment occurred at the surface and wasexpected to vary with air temperature (Akratos and Tsihrintzis,2007). The mean pH levels were 5.27 and 6.5 for the FSF and SSFsystems, respectively. The lowest observed pH was in the FSF sys-tem due to the microbial utilization of straight-chain hydrocarbonsor death of microbes (Chan, 2011). Towards the end of exposure, itwas found that for the two flow systems and for all treatments, thepH gradually increased.

Nivala et al. (2012) indicated that mechanisms such as waveaction and wind-induced mixing that contribute to surface re-aeration in FSF wetlands are not possible in SSF wetlands. Inaddition, the available oxygen in SSF treatment wetlands is derivedfrom atmospheric diffusion, from the plant via the aerenchyma, andfrom the convective air flow within the porous media (Kadlec andWallace, 2009; Tanner and Kadlec, 2003). Subsequently, there is animpact on oxygen availability in SSF wetlands. In this comparativestudy, the DO values of the water from each aquarium showed nosignificant differences between the two systems, and the DOmeanswere 4.88 and 4.84 mgL�1 for the FSF and SSF systems, respectively.The FSF system showed significantly higher oxidation-reductionpotentials (ORP) than did the SSF system (P < 0.05). The ORPoscillated betweenþ15.57 andþ173.6 mV for the FSF system, whilethe ORP oscillated between�33.3 andþ66.7mV for the SSF system.Thus, both of the systems’ environments were in the aerobic andanoxic range. For the two flow systems and for all of the treatments,there were indications of intense microbial processes and thepresence of oxygen in the systems. According to Lin andMendelssohn (2009), diesel in the substrate reduces the sub-strate’s oxidation-reduction potential, indicating that the substrateis becoming more anaerobic.

3.2. Adaptation of S. grossus to diesel

Plant growth parameters, including wet weight, dry weight,total length, and visual observation of withered or dead plants,were monitored and recorded. The statistical analysis showed sig-nificant differences between the sampling day versus the wet anddry weight for the SSF systems, and significant differences betweenthe concentrations versus the wet and dry weight for the FSF sys-tems (P < 0.05), as shown in Table 2. The wet and dry weights of S.grossus in both of the systems are depicted in Fig. 3. The Pearsoncorrelation coefficients between the wet and dry weight of S.grossus were r ¼ 0.558 and P < 0.01, and r ¼ 0.843, P < 0.01 for theFSF and SSF systems, respectively. It was obvious that S. grossusgrew better in the SSF system than in the FSF system. This is due to

Table 2Statistical analyses of Scirpus grossus wet and dry weight.

Concentration FSF SSF

Wet weight Dry weight Wet weight Dry weight

1% 10.42 � 1.89 3.57 � 0.88 16.43 � 5.32 3.03 � 1.412% 11.89 � 1.41 2.89 � 1.22 16.00 � 5.01 2.92 � 1.263% 8.56 � 1.44 2.00 � 0.58 14.00 � 3.9 2.62 � 1.02Significant

with concentration0 0 0.276 0.599

Significant with days 0.708 0.73 0 0Pearson correlation

(wetedry)0.558 0.843

All values are the means (n ¼ 3) � SD in (g).

Fig. 4. Percentage increase in length of Scirpus grossus throughout the 72 days of thediesel exposure period. A: significant increase in length between the two systems; a:the increase in length between two systems was not statistically significant (P < 0.05).

I.A.W. Al-Baldawi et al. / Journal of Environmental Management 130 (2013) 324e330328

hydrocarbon compounds that can move through plant cell mem-brane producing toxic effects, in addition to their hydrophobicity,which can prevent water infiltration and aeration essential for theplant growth (Fernández et al., 2011).

For S. grossus growth over the 72 days with 1% diesel, some ofthe plants were visibly withered in the FSF system, while in the SSFsystem, the plants showed no visible differences in appearance andno plant withering was recorded. With 2% diesel contaminatedwater, plants grew normally in the SSF system, while in the FSFsystem, plants showed signs of phytotoxicity, such as yellowing ofthe stems and withering, and biomass values decreased after day42. Plants in the SSF system with 3% diesel contaminated watershowed steady growth after 42 days, while in the FSF system, all ofthe plants showed a high level of phytotoxicity that inhibited thepropagation of plant roots and rhizosphere microbes (Wang et al.,2011), with the death of many plants occurring after day 7. Thisfinding is consistent with other studies by Liu et al. (2011) and

0

5

10

15

20

25

30

FSF SSF FSF

%2leseiD%1

Wet

bio

mas

s (g

)

0

1

2

3

4

5

6

7

8

9

10

FSF SSF FSF S

%2leseiD%1

Dry

bio

mas

s (g

)

Fig. 3. Measurements of the wet weight and dry weight of Scirpus

Zhang et al. (2012), who examined the growth of Scirpus triqueter insoils with a diesel concentration of 20,000 mg kg�1. These authorsconcluded that the degradation ratios would decline due to thetoxicity of diesel to the plants and microorganisms at higher dieselconcentrations (Liu et al., 2011; Zhang et al., 2012).

Based on the results of this study, for both the FSF and SSFsystemswith 1% diesel, the increase in plant length was similar, andthere was no significant difference between the two systems, asshown in Fig. 4. In the SSF systemwith 2% diesel, the plants showedbetter growth than in the FSF system. Comparing the SSF with FSFsystems with 3% diesel, the plants continued to grow in the SSFsystem, while they died in the FSF system. Overall, S. grossus grew

SSF FSF SSF

leseiD%3leseiD

Day 7

Day 14

Day 28

Day 42

Day 72

SF FSF SSF

leseiD%3leseiD

Day 7

Day 14

Day 28

Day 42

Day 72

grossus throughout the 72 days of the diesel exposure period.

0

10

20

30

40

50

60

70

80

90

100

7day

s

14da

ys

28da

ys

42 d

ays

72 d

ays

7day

s

14da

ys

28da

ys

42 d

ays

72 d

ays

7day

s

14da

ys

28da

ys

42 d

ays

72 d

ays

1% Diesel 2% Diesel 3% Diesel

% R

emov

al o

f T

PH

FSF

SSF

Fig. 5. Percentage of TPH removal by Scirpus grossus after 72 days of diesel exposure. Bars indicate the standard error of the three replicates (n ¼ 3). A: Difference in diesel removalfrom water between the FSF and SSF flow systems was statistically significant. a: No statistical significance (P < 0.05).

Table 3Results of the two-way ANOVA (F-ratios) used to test between-subject effects(concentration, day, system) and their interaction.

Independent variables F Sig.

Concentration 41.5641 0.0000Day 94.3783 0.0000System 65.7319 0.0000Concentration*day 4.21042 0.0005Concentration*system 29.2748 0.0000Day*system 5.71653 0.0006Concentration*day*system 1.90462 0.0760

I.A.W. Al-Baldawi et al. / Journal of Environmental Management 130 (2013) 324e330 329

well and tolerated diesel up to 1% in the FSF system and 2% in theSSF system. The tolerance of SSF system was up to 2% diesel due tosubstrate bed that enhances biodegradation as the major processesin the removal of the organic compound loads, sedimentation andsorption that were not available in FSF system.

3.3. Assessment of TPH removal efficiency between FSF and SSF

The removal efficiency of TPH from water contaminated withdifferent diesel concentrations (1%, 2%, and 3%) for 72 days ofexposure in the two flow systems was illustrated in Fig. 5. Based on

Table 4Mean values of TPH removal percentage form wastewater for the FSF and SSF sys-tems for different diesel concentrations.

Dieselconcentration

Day %TPH removal FSF %TPH removal SSF

Mean S.D. Mean S.D.

1% 7 45.33A 5.03 31.85A 10.2314 55a 5 57.22a 8.5428 59.33a 6.03 62.00a 4.5142 60a 6.56 65.13a 9.5672 70a 5 80.08a 5.55

2% 7 24.67a 9.29 35.33a 4.5114 53.67a 10.26 47.56a 8.4228 62.45a 7.50 70.63a 8.0742 69.67a 6.11 79.44a 7.1772 80A 4.36 91.55A 1.77

3% 7 23a 8.72 32.16a 7.0814 33.33A 5.69 55.83A 5.4528 32.33A 9.45 56.65A 5.5442 32.67A 4.93 73.29A 6.4872 42.33A 6.11 81.12A 3.56

A, a: In each column for each system, mean values followed by (A) are significantlydifferent (P < 0.05) and (a) indicates that they are not significantly different(P > 0.05).

the t-test analysis, The P-value was 0.004; therefore, the differencesin removal between the two flow systems were statistically signifi-cant. The TPH removal percentages in the FSF and SSF systems wereanalysed using a two-way ANOVA test to show the interaction be-tween factors. TPH removal differed significantly with concentration,days of exposure, and treatment at the P < 0.005 level. In addition,there was a statistically significant interaction at the P < 0.005 levelbetween (concentration*day), (concentration*system), and (day*-system), but there was a non-significant interaction between (con-centration*day*system) at the P ¼ 0.076 level, as shown by the two-way ANOVA test (Table 3) and illustrated in Fig. 5.

The TPH mean removal values are listed in Table 4, and showthat there were no significant differences between the FSF and SSFflow systems at the 1% and 2% diesel concentrations, but there wasa significant difference between the FSF and SSF flow systems at the3% diesel concentration (P< 0.05). Among all of the concentrations,the maximum removal degradation efficiency percentage after 72days of diesel exposure was 80.2% and 91.5% by the FSF and SSFsystems, respectively, with a 2% diesel concentration. Hydrocarbondegradation is believed to occur through a rhizosphere effect(Imfeld et al., 2009; Gikas et al., 2013b). As plants exude organiccompounds with their roots, the density, diversity, and activity ofspecific microorganisms in the surrounding rhizosphere increases,and this, in turn, degrades hydrocarbons. Therefore, the SSF systemwas more reliable than the FSF system, which was also noted by theUSEPA (2000), because in an SSF system, water must remain belowthe media surface to minimize human contact. Furthermore, thehigh removal efficiency observed in the SSF systemwas most likelybecause of the gravel matrix that isolates water from the pollutantsand works as a filter bed (Brovelli et al., 2011). Regarding an FSFsystem, volatilization is expected to be more apparent in FSF wet-lands, as water remains in direct contact with the atmosphere(Imfeld et al., 2009). A study by Fountoulakis et al. (2009b) foundthat the SSF system performed better in removing pollutants. Inparticular, the average removal of polycyclic aromatic hydrocarbons(PAHs) and linear alkylbenzene sulfonates (LAS) were 79.2% and55.5%, respectively, for the SSF constructed wetland, and 68.2% and30.0%, respectively, for the FSF constructed wetland.

4. Conclusions

FSF and SSF systems used for the phytoremediation of waste-water contaminated with diesel using S. grossus were operated for72 days. During this period, with the exception of 3% diesel con-centration in the FSF system, all treatments used in this study

I.A.W. Al-Baldawi et al. / Journal of Environmental Management 130 (2013) 324e330330

efficiently removed TPH from wastewater. The most appropriateflow systemwas determined by comparing the growth of the plantS. grossus and the level of TPH removal from wastewater. Aftercomparing the two systems, the FSF system had a greater efficiencyand performance in the removal of lower diesel concentrations ofup to 1%, while the SSF system had the same efficiency and per-formance for removal of higher diesel concentrations of up to 2%.Thus, it was determined that both the FSF and SSF systems had theability to remove TPH, but the greatest removal efficiency at highdiesel concentrations was observed in the SSF system. Other rea-sons for preferring the SSF system over the FFS system in this studyare the robust growth of plants, mosquito control, and the mini-mization of the impact of humanewildlife interactions.

Acknowledgements

The authors thank Universiti Kebangsaan Malaysia (UKM-KK-03-FRGS0119-2010) and the Tasik Chini Research Centre for sup-porting this research project. They also acknowledgewith thanks tothe Iraqi Ministry of Higher Education and Scientific Research forproviding a doctoral scholarship for the first author.

References

Al-Baldawi, I.A., Sheikh Abdullah, S.R., Suja’, F., Anuar, N., Idris, M., 2013a. Phyto-toxicity test of Scirpus grossus on diesel-contaminated water using a subsurfaceflow system. Ecol. Eng. 54, 49e56.

Al-Baldawi, I.A., Sheikh Abdullah, S.R., Suja’, F., Anuar, N., Idris, M., 2013b.A phytotoxicity test of bulrush (Scirpus grossus) grown with diesel contami-nation in a free-flow reed bed system. J. Hazard. Mater. 252e253, 64e69.

Akratos, C.S., Tsihrintzis, V.A., 2007. Effect of temperature, HRT, vegetation andporous media on removal efficiency of pilot-scale horizontal subsurface flowconstructed wetlands. Ecol. Eng. 29, 173e191.

Braeckevelt, M., Reiche, N., Trapp, S., Wiessner, A., Paschke, H., Kuschk, P.,Kaestner, M., 2011. Chlorobenzene removal efficiencies and removal processesin a pilot-scale constructed wetland treating contaminated groundwater. Ecol.Eng. 37, 903e913.

Brovelli, A., Carranza-Diaz, O., Rossi, L., Barry, D.A., 2011. Design methodology ac-counting for the effects of porous medium heterogeneity on hydraulic residencetime and biodegradation in horizontal subsurface flow constructed wetlands.Ecol. Eng. 37, 758e770.

Chan, H., 2011. Biodegradation of petroleum oil achieved by bacteria and nematodesin contaminated water. Sep. Purif. Technol. 80, 459e466.

Crop Protection Compendium, 2011. http://www.cabicompendium.org/NamesLists/CPC/Full/SCIRPS.htm (27.07.11.).

EPA 832-F-00-023. 2000. United States Environmental Protection Agency, Office ofwater Washington, D.C. Available from: http://water.epa.gov/scitech/wastetech/upload/2002_06_28_mtb_wetlands-subsurface_flow.pdf (03.05.12.).

Epps, A.V., 2006. Phytoremediation of Petroleum Hydrocarbons. U.S. EnvironmentalProtection Agency Office of Solid Waste and Emergency Response Office ofSuperfund Remediation and Technology Innovation, Washington, DC. Availablefrom: www.epa.gov www.clu-in.org (11.05.12.).

Fernández, M.D., Pro, J., Alonso, C., Aragonese, P., Tarazona, J.V., 2011. Terrestrialmicrocosms in a feasibility study on the remediation of diesel-contaminatedsoils. Ecotoxicol. Environ. Saf. 74, 2133e2140.

Fountoulakis, M.S., Terzakis, S., Chatzinotas, A., Brix, H., Kalogerakis, N., Manios, T.,2009a. Pilot-scale comparison of constructed wetlands operated under highhydraulic loading rates and attached biofilm reactors for domestic wastewatertreatment. Sci. Tot. Environ. 407, 2996e3003.

Fountoulakis, M.S., Terzakis, S., Kalogerakis, N., Manios, T., 2009b. Removal ofpolycyclic aromatic hydrocarbons and linear alkylbenzene sulphonates fromdomestic wastewater in pilot constructed wetlands and a gravel filter. Ecol. Eng.35, 1702e1709.

Gessner, T.P., Kadlec, R.H., Reaves, R.P., 2005. Wetland remediation of cyanide andhydrocarbons. Ecol. Eng. 25, 457e469.

Gikas, P., Ranieri, E., Tchobanoglous, G., 2013a. Removal of iron, chromium and leadfrom waste water by horizontal subsurface flow constructed wetlands. J. Chem.Technol. Biotechnol. 88, 1906e1912.

Gikas, P., Ranieri, E., Tchobanoglous, G., 2013b. BTEX removal in pilot-scale hori-zontal subsurface flow constructed wetlands. Desalin. Water Treat. 51, 3032e3039.

Homer, J., Glassmeyer, C., Sauer, N., Powell, J., 2009. Alternative wastewater treat-ment: on-site biotreatment wetlands at the Fernald Preserve visitors centre e

9434. In: WM2009 Conference, March 1e5, Phoenix AZ. http://www.wmsym.org/archives/2009/pdfs/9434.pdf (07.07.12.).

Imfeld, G., Braeckevelt, M., Kuschk, P., Richnow, H., 2009. Review: monitoring andassessing processes of organic chemicals removal in constructed wetlands.Chemosphere 74, 349e362.

Ji, G., Sun, T., Zhou, Q., Sui, X., Chang, S., Peijun, L.P., 2002. Constructed subsurfaceflow wetland for treating heavy oil-produced water of the Liaohe Oilfield inChina. Ecol. Eng. 18, 459e465.

Ji, G.D., Sun, T.H., Ni, J.R., 2007. Surface flow constructed wetland for heavy oil-produced water treatment. Bioresour. Technol. 98, 436e441.

Kadlec, R.H., 2009. Comparison of free water and horizontal subsurface treatmentwetlands. Ecol. Eng. 35, 159e174.

Kadlec, R.H., Wallace, S.D., 2009. Treatment Wetlands, second ed. CRC Press, BocaRaton, FL.

Lin, Q., Mendelssohn, I., 2009. Potential of restoration and phytoremediation withJuncus roemerianus for diesel-contaminated coastal wetlands. Ecol. Eng. 35, 85e91.

Liu, X., Wang, Z., Zhang, X., Wang, J., Xu, G., Cao, Z., Zhong, C., Su, P., 2011. Degra-dation of diesel-originated pollutants in wetlands by Scirpus triqueter and mi-croorganisms. Ecotoxicol. Environ. Saf. 74, 1967e1972.

Lohi, A., Cuenca, M., Anania, G., Upreti, R., Wan, L., 2008. Biodegradation of dieselfuel-contaminated wastewater using a three-phase fluidised bed reactor.J. Hazard. Mater. 154, 105e111.

Nazm, M., Uyanik, S., Yesilnacar, M.I., Sahinkaya, E., 2009. Side-by-side comparisonof horizontal subsurface flow and free water surface flow constructed wetlandsand artificial neural network (ANN) modelling approach. Ecol. Eng. 35, 1255e1263.

Nivala, J., Wallace, S., Headley, T., Kassa, K., Brix, H., Afferden, M., Müller, R., 2012.Oxygen transfer and consumption in subsurface flow treatment wetlands. Ecol.Eng. (in press) http://dx.doi.org/10.1016/j.ecoleng.2012.08.028.

Pedescoll, A., Corzo, A., Alvarez, E., Garcıa, J., Puigagut, J., 2011. The effect of primarytreatment and flow regime on clogging development in horizontal subsurfaceflow constructed wetlands: an experimental evaluation. Water Res. 45, 3579e3589.

Peng, S., Zhou, Q., Cai, Z., Zhang, Z., 2009. Phytoremediation of petroleumcontaminated soils by Mirabilis jalapa L. in a greenhouse plot experiment.J. Hazard. Mater. 168, 1490e1496.

Rousseau, D.P.L., Vanrolleghem, P.A., Pauw, N.D., 2004. Model-based design ofhorizontal subsurface flow constructed treatment wetlands: a review. WaterRes. 38, 1484e1493.

Stottmeister, U., Wießner, A., Kuschk, P., Kappelmeyer, U., Ka€stner, M., Bederski, O.,Mu€ller, R.A., Moormann, H., 2003. Effects of plants and microorganisms inconstructed wetlands for wastewater treatment. Biotechnol. Adv. 22, 93e117.

Tangahu, B.V., Abdullah, S.R.S., Basri, H., Idris, M., Anuar, N., Mukhlisin, M., 2013.Phytoremediation of wastewater containing lead (Pb) in pilot reed bed usingScirpus grossus. Int. J. Phytorem. 15, 663e676.

Tanner, C.C., Kadlec, R.H., 2003. Oxygen flux implications of observed nitrogenremoval rates in subsurface-flow treatment wetlands. Water Sci. Technol. 48,191e198.

Tee, H.C., Lim, P.E., Seng, C.E., Nawi, M.A., 2012. Newly developed baffled subsurface-flow constructed wetland for the enhancement of nitrogen removal. Bioresour.Technol. 104, 235e242.

USEPA, October 2000. Manual: Constructed Wetlands Treatment of MunicipalWastewaters. EPA/625/R-99/010. USEPA Office of Research and Development,Cincinnati, OH, p. 154.

Vymazal, J., 2009. Review: the use constructed wetlands with horizontal sub-surface flow for various types of wastewater. Ecol. Eng. 35, 1e17.

Vymazal, J., 2010. Review: constructed wetlands for wastewater treatment. Water 2,530e549.

Wallace, S., Schmidt, M., Larson, E., 2011. Atlantic Richfield company long-termhydrocarbon removal using treatment wetlands. In: The SPE Annual TechnicalConference and Exhibition Held in Denver, Colorado, USA, 30 Octobere2November 2011. http://naturallywallace.com/docs/108_SPE%20145797%20Wallace%202011.pdf (14.10.12.).

Wang, J., Liu, X., Zhang, X., Liang, X., Zhang, W., 2011. Growth response and phy-toremediation ability of reed for diesel contaminant. Procedia Environ. Sci. 8,68e74.

Weed Science Society of America (WSSA), 2011. http://www.wssa.net/Weeds/Invasive/FactSheets/Actinoscirpus%20grossus.pdf. (17.01.13.).

Zhang, X., Wang, Z., Liu, X., Hu, X., Liang, X., Hu, Y., 2012. Degradation of dieselpollutants in Huangpu-Yangtze River estuary wetland using plant-microbesystems. Int. Biodeterior. Biodegrad. 76, 71e75.