Machine Translated by Google

Journal of Environmental Management 210 (2018) 349e358

Content lists available at ScienceDirect

Journal of Environmental Management

journal homepage: [Link]/locate/jenvman

Research article

A full-scale comparison of two hybrid constructed wetlands

treating domestic wastewater in Pakistan
Mahwish Ali a, b , Diederik PL Rousseau B, Safia Ahmed a, *

The

Department of Microbiology, Quaid-i-Azam University, 45320 Islamabad, Pakistan b

Department of Industrial Biological Sciences, Ghent University Campus Kortrijk, Graaf Karel de Goedelaan 5, 8500 Kortrijk, Belgium

article info abstract

Articlehistory: Objective of the current work was to monitor the year-round response of full-scale hybrid
Received 4 May 2017 constructed-wetlands (CWs) treating domestic wastewater under variable continuous flow. Two
Received in revised form
systems were evaluated: system-I consisted of an anaerobic baffled reactor (ABR) followed by a
9 January 2018
saturated vertical subsurface-flow (VSSF) CW and a free-water-surface (FWS) CW as a tertiary
Accepted 12 January 2018
treatment; system-II consisted of an ABR followed by a horizontal subsurface-flow (HSSF) CW and
FWS. Maximum reduction of 80 and 78%, 81 and 82%, 63 and 69%, 79 and 89% for chemical oxygen
demand (COD), biological oxygen demand (BOD), total kjeldahl nitrogen (TKN) and total suspended
Keywords:
Anaerobic baffled reactor solids (TSS) was achieved in Systems I and II respectively. There was also effective removal (94%
Constructed wetlands and 93%) of the bacterial population in both systems while more than 94% of pathogenic
kC* model microorganisms were removed. Data from both systems were further used to compute the first
Wastewater treatment order rate constants for the keC* model commonly used in CW design. The treatment performance
Reuse was confirmed to follow a first-order reaction rate, in which the k20 values of chemical oxygen
demand (COD), biological oxygen demand (BOD), total kjeldahl nitrogen (TKN), total phosphorus
(TP) and total suspended solids were calculated as 165, 117, 133, 7.5 and 78 m yr1 respectively for
VSSF and 226, 134, 199, 22 and 73 m yr1 respectively for HSSF. A positive correlation with temperature was d
© 2018 Elsevier Ltd. All rights reserved.

1. Introduction from industrial, agricultural and municipal homes. Under this

legislation different ranges of parameters were set as standard (EPA,
Pakistan, as many developing countries, is facing serious water P, 1997).
quantity and water quality issues due to industrial, agricultural and For smaller discharges, constructed wetlands have been proven
municipal activities. Water availability among others is predicted to to be a cost-effective and sustainable treatment facility, especially
be less than 700 m3 per capita by 2025 which is far below the for developing countries (Zhang et al., 2014; Elzein et al., 2016).
international standard of 1500 m3 per capita (Martin et al., 2006). In Nowadays, treated wetlands are not only used for small systems but
addition, water pollution is causing unfavorable health conditions by are also implemented at large scales (Masi et al., 2017). Con-structed
raising the level of water borne diseases (Azizullah et al., 2011). wetlands are wastewater treatment systems engineered to utilize the
An important contributor to water pollution is municipal sewage, natural processes and interactions of wetland vegetation, bacteria
with an estimated discharge to surface water bodies of 7.57 106 m3 and substrates in a more controlled environment (Vymazal, 2011).
day1 . Less than 10% of this wastewater is currently Most studies reveal that they provide high quality wastewater
being treated in municipal wastewater treatment plants and there are treatment at relatively low CAPEX and OPEX costs compared to
no established provisions for the reuse of treated water (Martin et other conventional technologies (Day et al., 2004).
al., 2006). In 1997, Government of Pakistan set forth National They also have many potential additional benefits, such as effluent
environmental quality standards (NEQs) under Pakistan reuse, biomass production and habitat provision (Rousseau et al.,
environmental protection act (PEPA) to monitor effluent discharge quality 2008). It has also been proven that the flush toilet-esewereconstructed
wetland system is a culturally acceptable alternative for Pakistan
(Nawab et al., 2006).
* Corresponding author. Tel.: þ92 5190643009. Experience with this wastewater treatment technology is still
E-mail address: sahmed@[Link] (S. Ahmed).

[Link]
0301-4797/© 2018 Elsevier Ltd. All rights reserved.
Machine Translated by Google

350 M. Ali et al. / Journal of Environmental Management 210 (2018) 349e358

very limited in Pakistan. Aslam et al. (2007) presented a treatment system, equally serving as an equalization basin by
successful pilot-scale study with vertical flow constructed moderating incoming wastewater flows. The design was based
wetlands treating refinery effluent. More recently, Ijaz et al. on recommendations regarding hydraulic retention time (HRT),
(2015) and Sehar et al. (2015) presented promising results from sludge retention time (SRT) and methane production rate
microcosm CW studies with domestic wastewater. provided in Droste (1997). The resulting water holding volume
This paper presents for the first time results from a full-scale of the ABR is 92.15 m3 (outer dimensions L 12.8 m W 4.57 m H
CW system in Pakistan treating domestic wastewater, operated 2.75 m), providing an HRT of about 20 h, an organic loading rate
and monitored during a full year. Hybrid CWs where purposely .
of 5 kg BOD m3 day1 ) and an up-flow velocity of 0.9e1.2 m h1 It
chosen because literature (eg Vymazal, 2013 and previous lab- is divided into 4 chambers by three sets of baffles. A sludge
scale and pilot-scale studies (Sehar et al., 2015, 2016) have removal interval of 18e24 months was calculated based on
indicated high removal efficiencies. The study had three major sludge production estimates and recommended SRT. The walls
objectives: (1) to evaluate the effectiveness of CW domestic of the ABR and the baffles were constructed of brick masonry followed by p
wastewater treatment in Pakistan as influenced by seasonal Water from the ABR outlet was discharged into a small settling
variation; (2) to compare a hybrid HSSF-FWS system with a tank where the flow was divided into two equal parts to feed
VSSF-FWS system; and (3) to validate the kC* model (Kadlec both hybrid CW systems. An overflow was also present 10 cm
and Knight, 1996) to allow its further use as a design tool in [Link] both outlets to evacuate excessive rain water flows. Note
that the VSSF-CW thus received a continuous supply of
2. Materials and methods wastewater, and was not batch-fed as is usually the case. In
addition, because of the connection with the adjacent collecting
pond, most of the VSSF-CW was water-saturated, except for the uppermost
2.1. Study site description
All wetland cells were covered with a synthetic liner of 5 mm
thickness to prevent any wastewater infiltration to the soil. Both
The CWs were designed to treat the domestic wastewater
the VSSF-CW and HSSF-CW were filled with porous media:
originating from the residential colony of Quaid-i-Azam University
medium gravel of ø25-40 mm with porosity 0.40, and aggregate
in Islamabad. The site has coordinates 3344043.2800N, 73 7
04.0900E and is located at an elevation of 540 m. Wastewater is crush of ø10-15 mm having porosity 0.35 (Fig. 2). The SSF
wetlands were planted with Typha latifolia, Phragmites australis and Vetiver
collected in an underground sewage line, treated in a CW system
The collecting ponds after both SSF CWs were planted with
and then disposed of into the Rawal lake tributary. Before the
water hyacinth (Eichhornia crassipes) and pennywort (Centella asiatica).
CWs were constructed, a septic tank system was treating the
The FWS-CW were planted with water lettuce (Pistia stratiotes).
same wastewater flow, and flow data were recorded. Data from
All plants were locally collected from open drains.
2013 indicated an average flow rate of 110.1 ± 21.4 m3 day1 .
Taking into account 108 households in the residential colony,
2.3. Sampling and water quality analyzes
with on average 6 persons per household, this translates into
an average wastewater production of 170 liters per capita per day.
Grab samples were taken from the inlet and outlet of the
Climatic conditions of this region are characterized by large
systems every ten days from January 2015 to December 2015
seasonal and daily temperature variations. The location typically
and analyzed on the same day of sampling. APHA Standard
experiences hot dry summers and cold winters with monsoon
Methods (2005) were used to determine Orthophosphate, BOD,
spells. The average monthly temperature ranges from 11.3 C to
TSS, nitrates and sulphates while COD was done by a quick-test
29.9 C, with a lowest average minimum of 4.8 C in January up to
kit of range 25e1500 mg l1 (114541) provided by [Link]. Total
a highest average maximum of 36.5 C in June. Islamabad city
kjeldahl ni-trogen and ammonia 2e150 mg l1 (100683) were also
receives heavy monsoon spells during the summer months and
measured by quick-test kits provided by [Link]. Most probable
low rainfall during winter months. The driest months in this
number method (MPN index. (100 mL1 )) for pathogenic indicator
region are from November through January, although
([Link]) was performed. Samples were stored in MacConkey
occasionally brief thunderstorms produce intense precipitation.
broth at 42 C for 24e48 h using multiple tubes with inverted
The normal annual total precipitation in Islamabad city is 1585 mm.
Durham tubes. Tubes with positive results were selected for
further sub culturing on nutrient agar, MacConkey agar and
2.2. System design mannitol salt agar and put for incubation for 24e48 h at 37 C.
Bacterial isolates showing growth were analyzed through
The overall system lay-out is shown in Fig. 1. The wastewater microscopy and subjected to total count. Serial dilution CFU/ml
is sequentially treated by a screening unit, anaerobic baffled technique was used for the evaluation of bacterial load in
reactor and then split into two equal flows to feed hybrid systems I and II.
wastewater. Up to 101 -1010 dilu-tion were prepared and samples
System-I comprises of a saturated Vertical Subsurface-Flow (0.1 ml) from appropriate dilution were spread on nutrient agar
Constructed Wetland (VSSF-CW) followed by a collecting pond plates and were incubated at 37 C for 24 h and colonies were counted by co
and Free-Water-Surface Constructed Wetland (FWS-CW);
similarly System-II consists of a Horizontal Subsurface-Flow 2.4. Data treatment
Constructed Wetland (HSSF-CW) followed by a collecting pond
and Free-Water-Surface Constructed Wetland (FWS-CW). The Data were processed in MS Excel 2016™ to make box plots
required area for the CWs was estimated by using the first-order of different treatment units. Statistical analysis with SPSS
kC* equation with the parameter values for BOD removal as Statistics 23 was also performed. Data from VSSF-CW and HSSF-
mentioned in Kadlec and Knight (1996). Final dimensions are shown in CW Fig. 1. first examined for normality by using the Shapiro-Wilk
were
A screening unit of 1.83 m in length, 0.91 m in width and 1.52 test. Significant differences between both systems were then
m deep, was constructed as preliminary treatment to separate compared through parametric paired sample T-test for normal
coarse solid particles from raw sewage. It was vertically fitted data distribution while non-parametric related sample Wilcoxon
with two overlapping bar screens. signed rank test T-test for non-normal data was performed.
An anaerobic Baffled Reactor (ABR) was constructed as primary Values less than 0.05 (p < .05) were recorded as significant.
Machine Translated by Google

M. Ali et al. / Journal of Environmental Management 210 (2018) 349e358 351

Fig. 1. Wastewater Treatment scheme at Quaid-i-Azam University, Islamabad, Pakistan. (I) Anaerobic Baffle reactor (ABR) (II) Vertical Subsurface Flow (VSSF) CW (III)
Free Water Surface (FWS) I (IV) Horizontal subsurface flow (HSSF) CW (V) Free Water Surface II, System-I comprises of (I) (II) (III) and System-II consists of (I) (IV) (V).

Fig. 2. Cross section (not to scale) of Medium Porosity (P) and material distribution, CG (Coarse gravel), MG (Medium gravel) and AC (Aggregated crush) with diameter
in mm in HSSF-CW (a) and VSSF-CW (b).

Furthermore, the temperature-corrected kC* model proposed 3. Results

by Kadlec and Knight (1996) was calibrated with the available data.
Data from the last six months were used to avoid the startup effect. Treatment efficiency of constructed wetlands were compared by
Kadlec and Knight (1996) presented the first order model as: analyzing different wastewater parameters.

* *
¼ and kT=q 3.1. Total suspended solids (TSS)
Ce C Ci C (1)

where: C* ¼ background concentration (mg l1 ), kT ¼areal rate The range of TSS described by Metcalf et al. (2003) in domestic
constant adjusted for temperature (m.d1 ), T ¼ temperature (C), Ci wastewater is 120e400 mg l1 . The influential from this study con-
and Ce ¼ influent and effluent concentrations (mg l1 ) and q ¼ tained TSS above that range, ie 430e480 mg l1 . In the ABR, about
hydraulic loading rate (m.d1 ). half of the TSS were removed (48.7%) resulting in effluent
From this equation, kT can be calculated when concentrations concentrations of 150e400 mg l1 (Fig. 3a). VSSF and HSSF CWs
and hydraulic loading rates are known. By lack of adequate consequently removed 54% and 63.3% TSS with effluent
information, default C* values were used, as recommended by concentrations of 60e200 mg l1 and 60e100 mg l1 respectively (Fig. 3a).
Kadlec and Knight (1996) (Table 1). The effect of temperature was HSSF-CW was proven to be more effective than the VSSF-CW (p <
calculated as follows: .01). Another 21.4% and 50% TSS reduction was achieved in FWS-
I and FWS-II respectively, yielding final effluent concentrations of
kT ¼ k20 QT20 (two) 45e150 mg l1 and 25e60 mg l1 (Fig. 3a). System-I thereby achieved
a total efficiency of 82% and system-II achieved a total efficiency
where: k20 ¼ areal rate constant at 20 C (m d1 ) and Q ¼ of 91%. Both systems met the effluent quality required by the
dimensionless temperature coefficient. NEQS (Table 3).

Table 1
Applied background concentrations for SSF wetlands (adopted from Kadlec and Knight (1996)).
1
Background concentrations COD (mg l1 ) BOD (mg l1 ) TKN (mg l1 ) TSS (mg l1 ) TP (mg l )

SSF-C* 0 14 0 39 0.02
SF-C* 0 7.5 0 29 0.02
Machine Translated by Google

352 M. Ali et al. / Journal of Environmental Management 210 (2018) 349e358

Fig. 3. Boxplot of TSS (a), COD (b) and BOD (c) concentrations in different treatment units; values above boxes indicate mean removal efficiency (as a %) of each unit
compared to the preceding unit.

3.2. Chemical oxygen demand (COD) However, as time passed by, removal efficiency increased and
a maximum reduction of 45% was obtained by the end of the
For raw wastewater entering the treatment facility, COD was experiment (Fig. 3b). Effluent from the ABR had COD
ranging between 100 and 300 mg l1 so it was considered as .
concentrations of 60e220 mg l1 The COD load varied considerably
diluted or weak strength wastewater (Samorn, 1999). During the between months due to variations in flow rate. During winter
primary treatment in the ABR, initially only 15e20% reduction was achieved.
there was less consumption of water as compared to summer (Table 2).
Machine Translated by Google

M. Ali et al. / Journal of Environmental Management 210 (2018) 349e358 353

Table 2
Measured average monthly flow rates for the year 2015.

Months 1 Jan Feb Sea Apr May June Jul Aug Sep Oct Nov Dec
(m3 d )
flow rate 90.5 80 112.5 145.5 127 132.6 128.5 111.2 118.5 101.7 95.3 79.6

Further COD reductions were achieved through the VSSF and 3.4. Total kjeldahl nitrogen (TKN)
HSSF CWs, yielding additional efficiencies of 63% and 50% on
average respectively (Fig. 3b) and effluent concentrations reaching Total Kjeldahl nitrogen is the sum of ammonia, ammonium and
20e112 mg l1 and 30e200 mg l1 respectively. T-test confirmed organic nitrogen in the chemical composition of wastewater. The
that the effluent concentration of the VSSF-CW system was influent TKN ranges between 25 and 45 mgN l1 and is thus
significantly lower than the effluent of the HSSF-CW system, regarded as medium strength wastewater (Samorn, 1999). The ABR
indicating a better performance of the VSSF system (p < .01). For further yields a 19.4 average % reduction in TKN level (Fig. 4a) resulting in
polishing, wastewater was treated in a final FWS-CW system. FWS-I effluent TKN concentrations between 20 and 35 mgN l1 . VSSF-CW
and FWS-II removed another 7 and 20% COD on average respectively and HSSF-CW had an average reduction of 27.8% and 24.4% with
(Fig. 3b), resulting in final effluent concentration of effluent concentrations of 12e28 mgN l1 and 13e28
18e110 mg l1 and 12e110 mg l1 respectively. Thus, an overall mgN l1 respectively (Fig. 4a). Comparatively, both systems showed
COD reduction of 73.6% was obtained for system-I, and 71.5% for significant reduction of TKN and no significant difference between
system-II. There was no significant difference between the efficiency they were found by T-test (P > .05). The tertiary level treatment
of both hybrid systems (p > .05). Maximum reduction in COD through FWS-I and FWS-II resulted in an additional reduction of
was seen in the months of September and June, i.e. 80 and 81% for 10.8% and 27.7% respectively with final effluent values ranging
System-I and System-II respectively. Effluent from the systems also between 10 and 29 mgN l1 and 8e24 mgN l1 (Fig. 4a). A
met the NEQ standards (Table 3). maximum reduction of 60.9% TKN was shown in the month of June
from system-I and 69% in the month of August by system-II. The T-test
confirms better overall performance of System-II compared to
3.3. Biological oxygen demand (BOD)
System-I (p < .01).

BOD was measured as a suitable indicator for biodegradable

3.5. Nitrates
organic matter in wastewater. The raw wastewater had BOD
concentrations ranging from 60-200 mg l1 depending on the water
utilization by the community each month. This confirmed the Wastewater influent entering the treatment facility contained
nitrates within the range of 0e4.3 mgN l1 . After passing through
classification as diluted or weak strength wastewater (Samorn,
1999). As for COD, the initial BOD removal by the ABR was rather the ABR, there was initially 10e20% reduction which increased up
low, ie 9e20% in the first five months, but then increased until to 53% with time; effluent concentrations from ABR ranged between
reaching values of 47%. ABR effluent BOD concentrations varied 0 and 3.2 mgN l1 (Fig. 4b). Effluent from the ABR then
between 45 and 150 mg l1 (Fig. 3c). VSSF-CW and HSSF-CW passed through the VSSF and HSSF CWs where contrasting be-
haviors were observed: a mean increase in nitrate levels in the
achieved average removal efficiencies of 64.5% and 55.2% (Fig. 3c)
with an effluent range of 10e130 mg l1 and 11e127 mg l1 VSSF-CW up to 24.6% whereas a reduction up to 45.5% was shown
respectively. A T-test revealed that the VSSF-CW system showed in the HSSF-CW (Fig. 4b). Further treatment was achieved
significantly higher removal of BOD than the HSSF-CW (p < .01). through FWS-I and FWS-II, resulting in an average reduction of
Further tertiary treatment was done with FWS-I and FWS-II CWs 24.2% and 43.9% (Fig. 4b) with final effluent concentrations ranging
yielding an extra 7.0% and 20.2% BOD removal (Fig. 3c) and final between 0.23 and 2.26 mgN l1 and 0.02e1.20 mgN l1 respectively.
effluent values of 9e70 mg l1 and 7e70 mg l1 respectively. Overall 15.8% and 74.4% reduction was seen from system-I1
Overall BOD reductions from system-I and system-II were 76.2% and system-II respectively.
and 72.5% respectively. A T-test confirmed no significant difference
between the performances of both hybrid systems (p > .05). 3.6. Ammonia
Maximum reduction was shown in the months of September and
June i.e. 81.4% for system-I and 82.3% for system-II. Both systems Domestic wastewater coming from the residential colony of
fulfilled the NEQs (Table 3). QAU contained ammonia in the range of 19e35 mgN l1 which was

Table 3
Comparison of Effluent quality of both Constructed Wetland systems I and II with NEQs.

Parameters National Effluent Effluent Concentrations Number of Samples Effluent Concentrations Number of Samples
Quality Standards (mg l1 ) of System I (mg l1 ) above NEQs of System II (mg l1 ) above NEQs

Average S.D. Average S.D.

COD 150 47 7.4 0/36 52 10.6 0/36

BOD5 80 30 4.7 0/36 35 6.7 0/36
TSS 150 84 17.6 1/36 38 7.2 0/36
SO4 600 90 12.8 0/36 90.7 11.9 0/36
PO4 30 0.2 0/36 2.1 0.4 0/36
TKN It is

2 16.2 1.1
Nitrites It is

19 1 0.03 0.02
Nitrates It is

0.02 3.8 0.01 0.2 1.3 0.3

Ammonia 40 12 1.8 0/36 19.9 2.1 0/36
CFU/ml
It is

2.4 104 1.8 104 2.74 104 1.9 104

MPN/100ml
It is

48 7 64 9
Machine Translated by Google

354 M. Ali et al. / Journal of Environmental Management 210 (2018) 349e358

Fig. 4. Boxplot of TKN (a), NO3-N (b) and NH4-N (c) concentrations in different treatment units; values above boxes indicate mean removal efficiency (as a %) of each unit compared
to the preceding unit.
Machine Translated by Google

M. Ali et al. / Journal of Environmental Management 210 (2018) 349e358 355

in the normal range of 12e45 mgN l1 as specified by Metcalf et al. 2e12 mg l1 of orthophosphates, well within the normal range of
(2003). After the primary treatment, minimal reduction (4%) was 4e14 mg l1 given by Metcalf et al. (2003). These concentrations
observed. The VSSF-CW however removed 45% ammonia leading were already below the national standards. After the ABR, the
to effluent concentrations of 7e25 mgN l1, whereas no reduction at concentrations reached 1.3e9.6 mg l1 after an average reduction
all was observed in the HSSF-CW (Fig. 4c). After the tertiary of 29.3% (Fig. 5b). A further 42.6% and 42.4% removal was achieved
treatment from FWS-I and FWS-II, 13% and 22% reduction was through VSSF-CW and HSSF-CW systems with the effluent
observed with final effluent concentrations of 1.5e22 mgN l1 and concentration varying between 0.7 and 9 mg l1 and 0.7e6 mg l1
12e30 mgN l1 respectively (Fig. 4c). Total average reductions of respectively (Fig. 5b). T-test revealed no significant difference
52.8% and 25% were observed in system-I and system-II between both systems (p > .05). During the tertiary treatment, FWS-
respectively. T-test revealed better performance of system-I with I and FWS-II mean reductions of 17.7% and 33% were achieved
lower concentrations of ammonia in the effluent compared to with the final treated water having phosphate concentrations of
system-II (p < .01). After the successive treatments, effluents from 0.5e5 mg l1 and 0.09e4 mg l1 (Fig. 5b). Therefore, system-I
both systems fulfilled the required NEQ standard (Table 3). provided overall 67% removal and system-II provided 72%
reduction in phosphates. No significant difference was found
3.7. Sulphates between the performance of system-I and system-II (p < .01). A
maximum reduction of 78.8% was provided by system-I in the
Sulphates were monitored as a potentially important electron month of September and 88% from system-II in the month of November.
acceptor for anoxic COD/BOD5 degradation. Raw wastewater
contained sulphates in the range of 70e200 , thus already within 3.9. Colony forming unit (CFU)
mg l1 the acceptable range given by the NEQs (Table 3). Only
minimal removal of sulphates, ie 10.7% on average, was achieved Colony forming units was used to calculate the total number of
by the ABR (Fig. 5a). VSSF-CW and HSSF-CW yielded a mean viable bacteria present in the wastewater. Normal range of CFU
reduction of 28.8% and 26.5% with effluent concentrations of present in domestic wastewater is 10-105 ml1 (Metcalf et al., 2003).
35e151 mg l1 and 44e135 mg l1 respectively with no significant differenceThe
(p > .05) of CFU present in the wastewater was 2.78 105e6.8 105
range
(Fig. 5a). FWS-I and FWS-II achieved an additional mean reduction ml1 (log10 5.4e5.8). Hardly any reduction of CFU occurred in the
of 3.86% and 5.84% (Fig. 5a) with the final effluent concentrations ABR (Fig. 6a). However, through the VSSF-CW and HSSF-CW, 82%
reaching 30e125 mg l1 and 41e130 mg l1 respectively (Fig. 5a). and 77.7% removal was observed, with effluent values of 1.09
Maximum sulphates reduction occurred in the month of October 105e2.04 104 ml1 (log10 5.03e4.30) and 1.16 105e2.53 104 ml1
for both systems, and amounted to 53e54%. (log10 5.06e4. 40) (Fig. 6a). T-test revealed no significant difference
in removal of bacterial count in both VSSF and HSSF CWs effluent
3.8. Phosphates (p > .05). FWS-I and FWS-II removed another 50.2% and 67%
respectively with final effluent reaching 7.3 103e1.04 104 ml1
Wastewater entering the treatment facility contained (log10 3.8e4.0) and

Fig. 5. Boxplot of Sulphate (a) and Phosphate (b) concentrations in different treatment units; values above boxes indicate mean removal efficiency (as a %) of each unit
compared to the preceding unit.
Machine Translated by Google

356 M. Ali et al. / Journal of Environmental Management 210 (2018) 349e358

3 103e3.73 104 ml1 (log10 3.47e4.47) respectively (Fig. 6a). 3.11. First-order model calibration
System-I gives a total removal of 94.5% and system-II gives
93.8%. T-test demonstrated no significant difference between From equations (1) and (2), and taking into account the
both systems' performance. default C* values from Table 3, the k20 and Q values of different
variables were calculated for the very first time according to
3.10. Most probable number (MPN) climatic conditions for the capital of Pakistan and the results
are shown in Table 4.
The most probable number test was performed to quantify
the concentration of pathogenic microorganisms in wastewater. [Link]
According to Metcalf et al. (2003), pathogens present in domestic
wastewater typically range between 106 -1010. (100 ml1 ). Data 4.1. Role of the anaerobic baffled reactor
from the domestic wastewater to be treated showed values
above 1100. (100 ml1 ) (Applied limit of quantification, LOQ). Apart from the not-abnormal (Bodkhe, 2009) lower removal
Effluent from the ABR still contained viable microorganisms efficiencies during the start-up phase, the anaerobic baffled
above the LOQ (Fig. 6b). However, after both the VSSF-CW and reactor appeared to be an efficient pretreatment unit under the
HSSF-CW systems, considerably lower numbers were observed, given circumstances. ABR effectively removed solids and
with comparable (p > .05) effluent concentrations in the range organics, with TSS reductions was reaching up to 68%, and
of 30e293 (100 ml1 ) and 90e273 (100 ml1 ) respectively (Fig. BOD5 and COD removals reaching up to 45% at the end of the
6b ). FWS-I and FWS-II showed an additional 65% and 62% experiment. This ensured reasonable organics and solids
removal respectively. Final effluent concentrations of system-I loading rates on the subsequent HSSF-CW and VSSF-CW, and
should thus
contained MPN index of 9e105. (100 ml1 ) and system-II achieved to 70e135. be able
(100 ml1 ) of to
MPNprevent clog-aging in the longer term
after treatment. A T-test showed no difference between the final (Knowles et al., 2011). Note that the ABR alone did not allow to
effluent concentrations of system-I and system-II (p > .05). consistently reach NEQs for TSS, BOD5 and COD.

Fig. 6. Boxplot of CFU (a) and MPN (b) concentrations in different treatment units; values above boxes indicate mean removal efficiency (as a %) of each unit compared
to the preceding unit.
Machine Translated by Google

M. Ali et al. / Journal of Environmental Management 210 (2018) 349e358 357

Table 4
Q values and k20 values for Vertical and Horizontal subsurface flow constructed wetlands.

Units of CW Q Values k20 (m year1 )

COD BOD TSS TP TKN COD BOD TSS TP TKN

VSSF-CW 1.068 1.096 1.00 0.97 1.09 164.7 116.5 133 7.5 78
HSSF-CW 1.11 1.11 0.97 0.96 1.06 226 134 199 22 73
FWS-I 1.02 1.03 0.99 0.98 1.09 53.6 63.7 101.3 26.07 80.3
FWS-II 1.01 1.04 1.00 1.00 1.07 38.09 44.33 330 49.3 45.6

To a lesser degree, monitoring results also confirmed nutrient patterns and therefore contact times between wastewater and
removal, with average TKN reductions of 19.4% and average matrix materials must be different. A tracer test could elucidate this
further.
phos-phate reductions of 29.3%. The total construction cost of this ABR
amounted to PKR 1.35 million (z12,920 US$) or about PKR
2093.PE1 (z20 US$.PE1 ). Weekly inspections were required for 4.3. FWS-CW as polishing units
cleaning of screens and cleaning of inlet & outlet pipes. Combined
with sludge removal costs, this amounted to about PKR 0.09 million As shown before, NEQs were not consistently met in all cases by
per year Operational &Maintenance costs (861 US$.yr1 ). the VSSF-CW and HSSF-CW, thus corroborating the use of
additional polishing units, here in the form of FWS-CWs with floating
macrophytes. The FWS-CW do what they are designed to do, with
4.2. Comparison of the HSSF-CW and the VSSF-CW
effluent reaching the NEQs for all parameters and samples, except
for one single TSS outlier (Table 3).
Generally speaking, removal efficiencies of most parameters are
The FWS CWs from this study, compared to the VSSF-CW and
lower than what is typically mentioned in literature (Vymazal,
HSSF-CW systems, do not have any substrate bounded rhizosphere.
2002; 2006, 2007), which is most probably due to the very short
Plant roots are freely hanging in the water and they form a floating
hydraulic retention times applied in these systems (less than one
mat so they are helpful in removing organic matter through the
day), and because of the fact that the VSSF-CW was operated as a
processes of filtration, entrapment through roots and conversion in
continuously-fed saturated down flow system, instead of the more
the attached biofilms (Tanner and Headley, 2011). Taking into ac-
typical unsaturated, intermittently fed mode of operation.
count the relatively high removal efficiencies of TSS, and the
No significant differences were found between both effluents for
relatively low BOD and COD removal efficiencies in the FWS CWs, it
the wastewater components TKN, SO4, PO4, CFU and MPN. it was
appears that filtration and entrapment may be the dominant
however statistically confirmed that the VSSF-CW performed
processes. This is further corroborated by the rather low temperature
better than the HSSF-CW when looking at the effluent concentrations
coefficients (Q close to 1), which seems to point more to non-
of COD, BOD5 and NH4-N, and therefore consistently fulfilled the
biological processes.
NEQs. The drip irrigation and unsaturated top layer may have
Assuming that plant uptake of nutrients would be influenced by
enhanced oxygen transfer to the wastewater (Paing et al., 2015),
temperature, TKN and phosphates show a contrasting behavior,
which seems to be corroborated by the absence of nitrification in
with TKN clearly improving at higher temperatures, whereas
the HSSF-CW in contrast to an obvious nitrification rate in the
Phosphates are not affected by temperature. This indicates that
VSSF-CW. Apart from aerobic processes, the removal of sulphates in
other processes are important in removing TKN. As for phosphorus,
both wetland systems (~26e29%) prove that anaerobic processes
the temperature coefficients close to 1 seem to contradict a
also contribute to organics removal (Stein et al., 2007).
biological removal process and it remains to be seen therefore if
However, the VSSF-CW has a larger size and thus longer HRT
oc-casional harvesting of the floating plants could optimize P-removal
than the HSSF-CW (22.54 h versus 15.05 h). Taking this factor into
(Vymazal et al., 2006).
account, the HSSF-CW seems more effective in organic matter
removal on a unit area basis, as derived by the higher k20 values
5. Conclusions
(Table 4). These k20 values are in the higher range of observed
literature values (S1). The continuous feeding pattern of the
VSSF-CW may explain this rather unusual observation. Anaerobic baffle reactor as primary treatment was significant
Total suspended solids removal is an exception, in the sense that for the removal of solids and degradation of organic material. Both
statistics confirm that the HSSF-CW effluent concentrations are VSSF-CW and HSSF-CW proved to be efficient for the degradation of
lower than those of the VSSF-CW, despite its lower HRT and surface contaminants in wastewater. Free water surface on tertiary level as
area. The HSSF-CW therefore consistently meets the NEQs, whereas polishing treatment was shown effective for the removal of all the
the VSSF-CW exhibits a few exceedances. contaminants. Efficiency of the systems were affected by seasonality
Both systems furthermore showed a temperature-dependent variation with higher efficiency seen during summer. KC* model
removal for BOD5, COD and TKN (Q > 1; Table 4), which is not implemented was also proven to be useful for Asian climate. Both
surprising since removal of those parameters is dominated by systems meet the national standards for water discharge.
microbiological processes. Physical processes such as
sedimentation and filtration of TSS and adsorption of phosphates seem Acknowledgments
not
correlated with temperature (Q z 1; Table 4).
Another important observation is the fact that both systems This study was supported by funding from Higher education
significantly removed phosphates, but with the HSSF-CW having an Commission of Pakistan under the project “Small Scale Sewage
about 3-fold higher k20 value than the VSSF-CW (Table 4; 22 versus Treatment and Wastewater Reuse System for Pakistan” in
7.5m year1 respectively). Given that matrix materials are the same collaboration with Quaid-i-Azam University, Islamabad Pakistan and Georg
in both cases, and that the different area is taken into account in the Washington University, USA. We are also thankful to HEC for
kC* model, this leads to the assumption that hydraulic flow providing scholarship for six months under international research
Machine Translated by Google

358 M. Ali et al. / Journal of Environmental Management 210 (2018) 349e358

support initiative program. We are also obliged to Ghent Uni- Strategic Country Environmental Assessment. South Asia Region, World Bank 1.
Metcalf, Eddy, Burton, FL, Stensel, HD, Tchobanoglous, G., 2003. Wastewater En-
versity, Department of Industrial Biological Sciences, Belgium, gineering: Treatment and Reuse. McGraw Hill.
for providing working research opportunities. Nawab, B., Nyborg, IL, Esser, KB, Jenssen, PD, 2006. Cultural preferences in
designing ecological sanitation systems in North West Frontier Province,
Pakistan. J. Environ. Psychol. 26, 236e246.
Appendix A. Supplementary data Paing, J., Guilbert, A., Gagnon, V., Chazarenc, F., 2015. Effect of climate, wastewater
composition, loading rates, system age and design on performances of French
Supplementary data related to this article can be found at vertical flow constructed wetlands: a survey based on 169 full scale systems.
Ecol. Eng. 80, 46e52.
[Link]
Rousseau, D., Lesage, E., Story, A., Vanrolleghem, PA, De Pauw, N., 2008. Con-
structed wetlands for water reclamation. Desalination 218, 181e189.
References Samorn, M., 23 February 1999. Environmental Engineering Program, Asian Institute
of Technology. Bangkok, Thailand Available online.
Aslam, M.M., Malik, M., Baig, M., Qazi, I., Iqbal, J., 2007. Treatment performances of Sehar, S., Naeem, S., Perveen, I., Ali, N., Ahmed, S., 2015. A comparative study of
compost-based and gravel-based vertical flow wetlands operated identically for macrophytes influence on wastewater treatment through subsurface flow hybrid
refinery wastewater treatment in Pakistan. Ecol. Eng. 30, 34e42. constructed wetland. Ecol. Eng. 81, 62e69.
Azizullah, A., Khattak, MNK, Richter, P., H€ader, D.-P., 2011. Water pollution in Sehar, S., Naz, I., Khan, S., Naeem, S., Perveen, I., Ali, N., Ahmed, S., 2016.
Pakistan and its impact on public healthda review. Environ. Int. 37, 479e497. Performance evaluation of integrated constructed wetland for domestic
Bodkhe, S., 2009. A modified anaerobic baffled reactor for municipal wastewater wastewater treatment. Water Environment. Res. 88, 280e287.
treatment. J. Environ. Manage. 90, 2488e2493. Stein, OR, Borden-Stewart, DJ, Hook, PB, Jones, WL, 2007. Seasonal influence on
Day, J.W., Ko, J.-Y., Rybczyk, J., Sabins, D., Bean, R., Berthelot, G., Brantley, C., sulfate reduction and zinc sequestration in subsurface treatment wetlands.
Cardoch, L., Conner, W., Day, J. ., 2004. The use of wetlands in the Mississippi Water Res. 41, 3440e3448.
Delta for wastewater assimilation: a review. Ocean Coast. Manage. 47, 671e691. Tanner, CC, Headley, TR, 2011. Components of floating emerging macrophyte
Droste, R.L., 1997. Theory and Practice of Water and Wastewater Treatment. John treatment wetlands influencing removal of stormwater pollutants. Ecol. Eng. 37,
Wiley & Sons Incorporated. 474e486.
Elzein, Z., Abdou, A., elgawad, IA, 2016. Constructed wetlands as a sustainable Vymazal, J., 2002. The use of sub-surface constructed wetlands for wastewater
wastewater treatment method in communities. Procedure Environ. Sci. 34, treatment in the Czech Republic: 10 years' experience. Ecol. Eng. 18, 633e646.
605e617. Vymazal, J., 2007. Removal of nutrients in various types of constructed wetlands.
EPA, P, 1997. Pakistan Environmental Protection Act, 1997. Government of Pakistan, Sci. Total Environ. 380, 48e65.
Ministry of Environment, 25pp. Vymazal, J., 2011. Plants used in constructed wetlands with horizontal subsurface
Ijaz, A., Shabir, G., Khan, Q.M., Afzal, M., 2015. Enhanced remediation of sewage flow: a review. Hydrobiologia 674, 133e156.
effluent by endophyte-assisted floating treatment wetlands. Ecol. Eng. 84, 58e66. Vymazal, J., 2013. The use of hybrid constructed wetlands for wastewater treatment
with special attention to nitrogen removal: a review of a recent development.
Kadlec, R., Knight, R., 1996. Treatment Wetlands, 1996. New York. Water Res. 47, 4795e4811.
Knowles, P., Dotro, G., Nivala, J., García, J., 2011. Clogging in subsurface-flow Vymazal, J., Greenway, M., Tonderski, K., Brix, H., Mander, Ü., 2006. Constructed
treat-ment wetlands: occurrence and contributing factors. Ecol. Eng. 37, 99e112. Wetlands for Wastewater Treatment, Wetlands and Natural Resource Man-
Masi, F., Bresciani, R., Martinuzzi, N., Cigarini, G., Rizzo, A., 2017. Large scale agement. Springer, pp. 69e96.
application of French reed beds: municipal wastewater treatment for a 20,000 Zhang, DQ, Jinadasa, K., Gersberg, RM, Liu, Y., Ng, WJ, Tan, SK, 2014. Application
inhabitant's town in Moldova. Water Sci. Technol. [Link] of constructed wetlands for wastewater treatment in developing countries and
wst.2017.188. review of recent developments (2000e2013). J. Environ. Manage. 141, 116e131.
Martin, P., Nishida, J., Afzal, J., Akbar, S., Damania, R., Hanrahan, D., 2006. Pakistan