Ecological Engineering 95 (2016) 129–137

Home , Constructed wetland

Contents lists available at ScienceDirect

Ecological Engineering

jo urnal homepage: www.elsevier.com/locate/ecoleng

Comparing subsurface ﬂow constructed wetlands with mangrove

plants and freshwater wetland plants for removing nutrients and

toxic pollutants

a,b,1 a,c,1 a,∗

Jonathan Y.S. Leung , Qinhong Cai , Nora F.Y. Tam

Department of Biology and Chemistry, City University of Hong Kong, Tat Chee Avenue, Kowloon, Hong Kong

School of Biological Sciences, The University of Adelaide, South Australia, Australia

The Northern Region Persistent Organic Pollution Control (NRPOP) Laboratory, Faculty of Engineering and Applied Science, Memorial University of

Newfoundland, St. John’s, Newfoundland A1B 3X5, Canada

a r t i c l e i n f o a b s t r a c t

Article history: Constructed wetlands (CWs) have been increasingly used to remove nutrients from wastewater, but

Received 16 February 2016

their effectiveness to treat toxic wastewater remains largely unexplored. This study compared the treat-

Received in revised form 1 May 2016

ment efﬁciency of CWs using mangrove plants (Aegiceras corniculatum and Bruguiera gymnorrhiza) and

Accepted 14 June 2016

non-mangrove plants (Acorus calamus, Canna indica and Phragmites australis) in different cultural arrange-

Available online 2 July 2016 + −

ments (mono-culture vs. mixed-culture) to remove nutrients (TOC, TKN, TP, NH4 and NO3 ) and toxic

pollutants (heavy metals, PAHs and phenol) from wastewater. Additionally, the effect of tidal ﬂushing on

Keywords:

the treatment efﬁciency of the mangrove CWs was examined. The effectiveness of CWs was evaluated

Constructed wetland

based on the health status of plants after 6-month irrigation with toxic wastewater, and the removal per-

Freshwater wetland plant

centage of nutrients and pollutants. Following the experimental period, the mangrove plants remained

Heavy metal

Mangrove healthy, while the non-mangrove plants were impaired by the toxic wastewater (e.g. chlorosis and wilt-

Nutrient ing). In both mangrove and non-mangrove CWs, the presence of plants slightly enhanced the removal of

Organic pollutant nitrogenous compounds, while the pollutants were mostly adsorbed onto the sediment. The mangrove

Wastewater treatment

CWs generally had higher removal percentage of both nutrients and pollutants than the non-mangrove

CWs. In the mangrove CWs, however, tidal ﬂushing was necessary not only to facilitate the removal of

−

nutrients, but also to prevent the production of NO3 . Cultural arrangement had no signiﬁcant effect on

the treatment efﬁciency. We conclude that the mangrove CWs, especially planted with A. corniculatum,

have higher application values than the non-mangrove CWs to treat toxic wastewater on condition that

tidal ﬂushing is provided.

1. Introduction Constructed wetlands are composed of wetland plants, sub-

strates and the associated microbial assemblages to treat

Owing to the rapid urbanization and industrialization, water wastewater through various physical and biological processes.

quality has been deteriorating in many developing countries over While pollutants are primarily removed by adsorption onto sub-

the last few decades (e.g. Singh et al., 2002; Wu et al., 2016). strates (Brix, 1997), absorption by plants, especially nutrients, can

Given the high operating and maintenance cost of conventional be substantial (Thullen et al., 2005; Yang et al., 2007; Yadav et al.,

wastewater treatment facilities, discharge of untreated wastewater 2012; Li et al., 2013; Sehar et al., 2015). Besides, roots provide suit-

into water bodies is commonly observed. In recent years, there- able habitats by releasing oxygen or organic exudates for microbes

fore, constructed wetlands (CWs) have been increasingly used to to degrade nutrients (Stottmeister et al., 2003; Vymazal, 2007;

treat wastewater in view of cost-effectiveness and environmental Peng et al., 2014). Nevertheless, the inﬂuence of plants is species-

friendliness (Kivaisi, 2001). speciﬁc, depending on their growth rate and nutrient uptake rate

(Yang et al., 2007). Cultural arrangement may be associated with

treatment efﬁciency. Mixed-culture, for example, is shown to have

higher removal percentage of nutrients than mono-culture by

∗

Corresponding author at: Department of Biology and Chemistry, City University enhancing the effective distribution of roots and diversity of micro-

of Hong Kong, Tat Chee Avenue, Kowloon, Hong Kong.

bial community (Zhang et al., 2010, 2012; Abou-Elela and Hellal,

E-mail address: [email protected] (N.F.Y. Tam).

1 2012). To maximize the treatment efﬁciency of CWs, using the

These authors contributed equally to this work.

http://dx.doi.org/10.1016/j.ecoleng.2016.06.016

130 J.Y.S. Leung et al. / Ecological Engineering 95 (2016) 129–137

optimal planting arrangement (i.e. optimal plants and cultural arrangement). To investigate the effect of tidal ﬂushing, artiﬁcial

arrangement) is monumental. seawater (salinity: 15 ppt) was added to the CWs (5 cm above sedi-

Freshwater wetland plants, such as Phragmites australis and ment surface) to stimulate high tide, while drained to stimulate low

Acorus calamus, have been widely used in CWs, but it is notewor- tide. The sediment was ﬂooded daily from 6:00 p.m. to 10:00 a.m.

thy that most of the previous studies focused on the removal of (next day). For the non-mangrove CWs, A. calamus (Acc) (ca. 120 cm

nutrients with low level of toxic pollutants (e.g. Li et al., 2014; tall), C. indica (Ci) (ca. 80 cm tall) and P. australis (Pa) (ca. 70 cm tall)

Zhang et al., 2014). In reality, high concentrations of salts, nutri- were used. Seedlings of Acc were grown from seeds under green-

ents and toxic pollutants are often found in wastewater, especially house conditions, while seedlings of Ci were purchased from a local

in the developing countries with less stringent discharge stan- gardening company. Seedlings of Pa were collected from the reed

◦ ◦

dards. As such, freshwater wetland plants are probably not versatile bed in Mai Po Nature Reserve (22 29 33 N, 114 02 08 E), Hong

enough for CWs because they may not be able to tolerate the Kong. There were seven planting arrangements: three monocul-

stress due to salinity and toxic pollutants. For example, toxic pol- tures of Acc, Ci and Pa, three mixed cultures in crossed combination

lutants can directly damage the plants or microbial communities (individual ratio = 1:1), and control. Four healthy individuals were

(Schützendübel and Polle, 2002; Alkio et al., 2005; Wang et al., evenly planted in each tank (n = 3 replicate tanks per planting

2007; Park et al., 2012), thereby reducing the treatment efﬁciency arrangement).

of CWs. In this regard, mangrove plants may be competent for CWs

because of their strong tolerance to pollutants as well as their spe- 2.2. Maintenance of constructed wetlands and record of health

cial adaptations to survive in the highly ﬂuctuating environment status

(e.g. salinity, temperature and anoxic substratum) (Wu et al., 2008).

Since mangrove plants are distributed in the intertidal environ- Following the setup of CWs, the plants were daily irrigated with

ment and subject to periodic flooding, tidal flushing should also be 1 L artificial wastewater and allowed to acclimate under green-

considered in the design of CWs when mangrove plants are used. house conditions for two months prior to experimentation. The

The present study aimed to examine the effectiveness of sub- chemical compositions of the artiﬁcial wastewater, which mimic

surface ﬂow constructed wetlands to treat toxic wastewater using those of the wastewater in Shenzhen River, are shown in Appendix

Aegiceras corniculatum, Bruguiera gymnorrhiza, Acorus calamus, I. After acclimation, the initial and ﬁnal (six months after com-

Canna indica and Phragmites australis in different cultural arrange- mencement) stem height and leaf number of each plant were

ments (mono- and mixed-cultures). The former two are dominant recorded. The growth status of plants, such as sprouting, ﬂowering

mangrove plants, while the rest are common wetland plants which and wilting, was also recorded. During the acclimation and experi-

have good ability to treat wastewater (Konnerup et al., 2009; Xu mental periods, the plants were cleaned regularly to prevent fungal

et al., 2010; Li et al., 2014). These wetland plants are collectively or insect infestation.

denoted as ‘non-mangrove’ in this paper. In addition, the effect of

tidal flushing on treatment efficiency of the mangrove CWs was 2.3. Collection of effluent and sediment samples

evaluated. The chemical compositions of the wastewater used in

this study stimulated those in Shenzhen River, South China, which Efﬂuent samples from the CWs were collected bimonthly. The

suffers from severe water pollution due to discharge of indus- debris in the effluent was filtered and the filtrate was used for

trial sewage. The effectiveness and suitability of CWs to treat toxic chemical analyses. Sediment samples were collected at the end of

wastewater were evaluated by the removal percentage of nutri- the experiment. To do so, each plant was carefully removed with

ents and pollutants, and the health status of plants after long-term the aid of a spade and only the sediment underneath was collected.

irrigation with toxic wastewater. The ﬁndings could provide a more The sediment samples were freeze-dried, ground into powder and

realistic evaluation of CWs for wastewater treatment and shed light passed through a 2 mm sieve prior to chemical analyses.

on the signiﬁcance of planting arrangement for optimization of

CWs. 2.4. Analyses of efﬂuent and sediment samples

The pH of sediment was measured using a pH meter (HI 9025,

2. Materials and methods Hanna Instruments, USA) after mixing with deionized water (1:5,

w/v). Redox potential of sediment was measured in situ by a hand-

2.1. Setup of constructed wetlands held Conductivity-pH-Redox-Temperature Meter (WP81, Science

Essentials, Australia). Total organic carbon (TOC) in the efﬂuent

To optimize the performance of plants, sediments collected from was measured using a TOC analyzer (TOC-5000A, Shimazu, Japan).

◦ ◦

the mangrove in Sai Keng (22 25 12 N, 114 16 06 E) and freshwa- Flow injection analyzer (QuikChem 8000, Lachat Instruments, USA)

◦ ◦ +

ter wetland in Long Valley (22 29 55 N, 114 06 59 E) were used to was used to analyze the concentrations of ammonium (NH4 ) and

−

ﬁll up the purpose-made tanks (66 cm long × 25 cm wide × 37 cm nitrate (NO3 ) in the efﬂuent and sediment. Potassium chloride

+ −

tall, 20 cm sediment depth, Fig. 1) for the mangrove and non- (2 M) extraction was applied to extract NH4 and NO3 in the sed-

mangrove CWs, respectively. The sediment in Sai Keng mangrove iment. Total Kjeldahl nitrogen (TKN) and total phosphorus (TP) in

was sandy (Sand: 78.4%; Silt: 21.6% and Clay: 0%) with hydraulic the efﬂuent and sediment were also determined by ﬂow injection

retention time of 10.6 h, whereas the sediment in Long Valley analyzer, following Kjeldahl acid digestion. Heavy metals in the

wetland was silty (Sand: 30.0%; Silt: 67.4% and Clay: 2.6%) with sediment were extracted using concentrated nitric acid, followed

hydraulic retention time of 12.0 h in the CWs. Individuals of B. gym- by microwave digestion. The concentrations of heavy metals in the

norrhiza (Bg) (ca. 20 cm tall) and A. corniculatum (Ac) (ca. 20 cm tall), extract and efﬂuent, including cadmium (Cd), chromium (Cr), cop-

collected from the mangrove in Sai Keng, were cultured in the man- per (Cu), iron (Fe), manganese (Mn), nickel (Ni), lead (Pb) and zinc

◦

grove CWs under greenhouse conditions (temperature: 24 ± 1 C; (Zn), were analyzed by inductively couple plasma optical emission

relative humidity: 80%) for one month prior to experimenta- spectrometry (Optima 2100 DV, PerkinElmer, USA). A certiﬁed ref-

tion. There were four planting arrangements: monoculture of Ac, erence material (MESS-3, National Research Council, Canada) was

monoculture of Bg, mixed culture of Ac and Bg (individual used to estimate the recoveries of heavy metals, which ranged from

ratio = 1:1), and control (i.e. no plants). Four healthy individuals 73.5 to 101.2% with relative standard deviation (RSD) less than

were evenly planted in each tank (n = 3 replicate tanks per planting 1.78%.

J.Y.S. Leung et al. / Ecological Engineering 95 (2016) 129–137 131

Fig. 1. The design of the subsurface-ﬂow constructed wetland used in the present study.

Liquid-liquid extraction method was applied to extract PAHs in addition to planting arrangement. PERMANOVA was performed

from 100 ml efﬂuent sample, while accelerated solvent extrac- using software PRIMER 6 with PERMANOVA+ add-on.

tor (ASE 200, Dionex, USA) was used to extract PAHs from

3 g sediment sample. Ethyl acetate was used as the extracting 3. Results

solvent for both PAHs and phenol. Before extraction, appropri-

ate amounts of m-terphenyl and ﬂuorophenol were spiked into 3.1. Health status of plants

the efﬂuent and sediment as internal standards for PAHs and

phenol, respectively. The concentrations of phenanthrene (Phe), After the 6-month experimental period, we observed that the

pyrene (Pyr) and benzo(a)pyrene (BaP) were determined by gas health status of plants was unaffected by planting arrangement

chromatography–mass spectrometry (7890A, Agilent Technolo- in both mangrove and non-mangrove CWs, while the effect of

gies, USA), equipped with a mass selective detector (HP 5975, tidal ﬂushing was discernible. The stem height and leaf number

Agilent Technologies, USA) and a HP-5MS fused silica capillary of Ac increased by 40.0% and 56.9%, respectively, when tidal ﬂush-

× × ␮

column (30 m 0.25 mm 0.25 m). Oven temperature was pro- ing was not applied (Table 1). Tidal ﬂushing could facilitate the

◦ ◦ ◦ −1

gramed from 80 C to 180 C (30 C min ) and then ramped to growth of Ac by 23.4% in stem height and 21.0% in leaf number,

◦ ◦ −1

300 C (6 C min ). Ultra-pure helium was used as the carrier and increase survival by 16.6%. Irrespective of tidal ﬂushing, the

−1

␮

gas at a ﬂow rate of 1 ml min . 1 l sample was injected with growth of Bg was not obvious. Both Ac and Bg were healthier with

◦

an initial temperature of 280 C using splitless injection mode. more greenish leaves and fewer shedding leaves when tidal ﬂush-

The mass spectrum was obtained using electron impact ioniza- ing was applied. As for the non-mangrove plants, their stem height

tion mode at 70 eV. SIM scanning was used for the characteristic increased (by 16.2% in Acc; 14.9% in Ci; 9.4% in Pa), but their leaf

ions of 178 (Phe), 202 (Pyr) and 252 (BaP). The concentra- number decreased (by 2.1% in Acc; 10.6% in Ci; 7.2% in Pa). Pa was

tion of phenol was determined by gas chromatography (HP less tolerant to the wastewater than Acc and Ci in view of its higher

5890, Agilent Technologies, USA), equipped with a ﬂame ion- mortality. Wilting and yellowish leaves were observed in Acc and

ization detector and a DB-225MS fused silica capillary column Ci, despite their 100% survival.

(30 m × 0.25 mm × 0.25 ␮m). The temperature was programmed

◦ ◦ ◦ −1 ◦

from 60 C to 110 C (40 C min ), followed by ramping to 150 C

3.2. Removal of nutrients and pollutants in wastewater

◦ −1

(5 C min ). Matrix spike recovery test was conducted to estimate

the recoveries of PAHs and phenol. The recoveries of Phe, Pry, BaP

In general, nutrients were substantially removed by the man-

−

and phenol were 72.6%, 81.3%, 105.7% and 108.2%, respectively,

grove CWs, except NO3 (Table 2). The removal percentage of

with RSD less than 6.15%.

TOC (70.8–85.6%) and TP (100%) was neither affected by planting

arrangement nor tidal ﬂushing. However, the removal percentage

+ −

2.5. Data treatment and statistical analysis of TKN, NH4 and NO3 was signiﬁcantly enhanced by tidal ﬂushing

(Appendix II). As for planting arrangement, the control had lower

The removal percentage of nutrients and pollutants was calcu- removal percentage of TKN than Ac + Bg. Without tidal ﬂushing,

lated based on the following formula: the planted CWs had signiﬁcantly higher removal of NH4 than the

control, except Ac. Tidal ﬂushing was necessary for the removal of

−

Influent concentration Effluent concentration × −

Removal percentage 100 NO3 , whether the CWs were planted or not. With tidal ﬂushing,

Influent concentration −

higher removal of NO3 was generally found in the planted CWs.

−

Permutational analysis of variance (PERMANOVA) was con- Without it, the removal percentage of NO3 was negative in the

ducted to examine the effect of planting arrangement on planted CWs but positive in the control, indicating the production

−

(1) removal percentage of nutrients and pollutants by the of NO3 in the planted CWs. Heavy metals, except Mn (74.4–91.2%)

non-mangrove CWs, and (2) physico-chemical properties of sedi- and Zn (84.1–93.4%), and organic pollutants were substantially

ment after the 6-month experimental period in the non-mangrove removed by the mangrove CWs (>99% for most of them) (Appendix

CWs. As for the mangrove CWs, tidal flushing is another fixed factor III). Tidal flushing enhanced the removal of Mn (Appendix II).

132 J.Y.S. Leung et al. / Ecological Engineering 95 (2016) 129–137

Table 1

The average stem height and number of leaves of plants in the mangrove and non-mangrove constructed wetlands before and after 6-month irrigation with artiﬁcial

wastewater. The individuals from the same species are pooled as the effect of cultural arrangement is indiscernible. The numbers in the parentheses indicate percentage

change.

Before After Remarks

n Stem height Number of n Survival (%) Stem Number of

(cm) leaves height (cm) leaves

Mangrove CWs with tidal ﬂushing

A. corniculatum 18 21.6 11.6 17 94.4 35.3 (63.4%) 21.8 (87.9%) Healthy with greenish leaves

B. gymnorrhiza 18 24.0 8.83 16 88.9 24.9 (3.8%) 9.13 (3.4%) Healthy with greenish leaves

Mangrove CWs without tidal ﬂushing

A. corniculatum 18 23.3 11.6 14 77.8 32.6 (40.0%) 18.2 (56.9%) Healthy with greenish leaves

B. gymnorrhiza 18 22.5 10.4 17 94.4 22.9 (1.7%) 10.8 (3.8%) Some individuals with serious shedding leaves

Non-mangrove CWs

A. calamus 24 134.7 10.0 24 100 156.5 (16.2%) 9.79 (−2.1%) Yellowish leaves

C. indica 24 87.1 10.3 24 100 99.2 (14.9%) 9.21 (−10.6%) Some wilting individuals; yellowish leaves

P. australis 24 77.0 7.25 15 62.5 84.2 (9.4%) 6.73 (−7.2%) Some wilting individuals; yellowish leaves

Table 2

The removal percentage of nutrients by the mangrove CWs with or without tidal ﬂushing (mean ± S.D., n = 9).

Removal percentage (%)

With tidal ﬂushing Without tidal ﬂushing

Ac Bg Ac + Bg Control Ac Bg Ac + Bg Control

± ±

TOC 84.2 11.3 78.2 15.5 84.9 11.2 85.6 ± 11.5 78.1 ± 25.0 79.0 ± 19.7 70.8 ± 25.0 77.7 ± 24.0

TKN 88.3 ± 4.90 84.2 ± 2.91 87.3 ± 4.27 83.6 ± 9.62 77.7 ± 7.72 79.1 ± 5.58 82.4 ± 4.18 73.0 ± 8.78

TP 100 ± 0.00 100 ± 0.00 100 ± 0.00 100 ± 0.00 100 ± 0.00 100 ± 0.00 100 ± 0.00 100 ± 0.00

NH4 99.6 ± 0.38 98.1 ± 2.44 97.9 ± 2.70 93.4 ± 11.0 84.8 ± 9.95 94.7 ± 6.16 95.0 ± 5.71 77.9 ± 10.7

−

NO3 76.5 ± 11.5 68.8 ± 18.8 72.5 ± 15.4 52.1 ± 23.4 −72.4 ± 74.2 −83.4 ± 84.6 −147.1 ± 89.3 21.3 ± 31.5

Table 3

The removal percentage of nutrients and heavy metals by the non-mangrove CWs (mean ± S.D., n = 9). The different superscript letters within each parameter indicate

signiﬁcant difference according to PERMANOVA (p ≤ 0.05).

Removal percentage (%)

Acc Ci Pa Acc + Ci Acc + Pa Ci + Pa Control

Nutrients

TOC 83.1 ± 13.2 74.3 ± 21.7 92.3 ± 2.98 82.1 ± 11.9 84.8 ± 16.3 88.6 ± 4.97 83.6 ± 8.03

TKN 57.2 ± 25.0 56.2 ± 23.4 62.8 ± 16.4 57.4 ± 24.0 60.7 ± 21.4 57.8 ± 26.0 47.5 ± 17.6

TP 100 ± 0.00 100 ± 0.00 100 ± 0.00 99.95 ± 0.16 100 ± 0.00 100 ± 0.00 100 ± 0.00

NH4 68.5 ± 11.7 44.6 ± 24.8 62.4 ± 18.4 63.2 ± 19.3 70.8 ± 15.1 62.5 ± 20.7 50.0 ± 21.2

− ±

NO3 79.3 10.6 82.7 25.3 88.3 ± 10.4 75.1 ± 24.1 81.1 ± 13.3 83.2 ± 16.3 81.5 ± 13.1

Heavy metals

Cd 75.1 ± 15.0 73.5 ± 19.7 87.3 ± 13.5 62.4 ± 33.4 64.1 ± 14.6 80.1 ± 13.8 86.9 ± 15.7

Cr 96.93 ± 7.51 99.72 ± 0.48 98.23 ± 4.97 99.03 ± 0.97 93.02 ± 11.4 97.41 ± 5.13 99.83 ± 0.31

ab a a b b a a

Cu 69.1 ± 21.3 83.8 ± 14.3 89.6 ± 15.0 67.2 ± 29.1 56.2 ± 18.5 77.8 ± 18.4 95.9 ± 4.01

Fe 95.29 ± 6.82 97.18 ± 3.79 97.51 ± 5.26 94.25 ± 8.82 92.91 ± 10.1 87.94 ± 12.7 99.64 ± 0.32

Mn 45.4 ± 22.1 50.3 ± 22.3 57.9 ± 19.2 48.5 ± 19.8 62.9 ± 11.8 62.3 ± 23.2 47.8 ± 23.1

Ni 55.2 ± 19.2 63.0 ± 13.4 70.0 ± 23.7 50.8 ± 27.4 71.9 ± 11.1 73.0 ± 12.3 76.1 ± 18.0

ab a ab ab b ab a

Pb 91.9 ± 14.7 98.8 ± 1.05 95.9 ± 9.57 95.3 ± 4.35 85.2 ± 14.4 94.2 ± 6.49 99.8 ± 0.35

b ab a ab ab a a

Zn 49.1 ± 22.9 59.6 ± 23.7 80.5 ± 14.1 64.2 ± 26.8 62.1 ± 13.2 78.7 ± 14.0 83.4 ± 15.3

Compared to the mangrove CWs, the non-mangrove CWs gener- CWs complied with the discharge standards, except Mn and Zn

−

ally had higher removal percentage of TOC (74.3–92.3%) and NO3 (Appendix V). In contrast, the efﬂuent treated by the non-mangrove

(75.1–88.3%), but lower removal percentage of TKN (47.5–62.8%) CWs failed to comply with the standards for Cd, Mn, Ni and Zn, while

+ +

and NH4 (44.6–70.8%) (Table 3). TP was also totally removed by the NH4 and Fe exceeded the standards in some of the non-mangrove

non-mangrove CWs. Planting arrangement could not signiﬁcantly CWs.

inﬂuence the removal percentage of nutrients probably because of

the large variations. The non-mangrove CWs generally had lower

3.3. Changes in sediment properties

removal percentage of heavy metals than the mangrove CWs, espe-

cially Cd, Cu, Mn, Ni and Zn (Table 3). The removal percentage of

As far as the mangrove CWs are concerned, the pH of sediment

heavy metals was also unaffected by planting arrangement, except

slightly decreased in all CWs after the 6-month experimental period

Cu, Pb and Zn. Higher removal percentage was found in the control

(Table 4). Redox potential was signiﬁcantly higher in the planted

than some of the planted CWs (e.g. Acc+ Pa for Cu and Pb; Acc for

CWs (220–274 mV) than the control (11.3–67.0 mV), except Bg with

Zn). Organic pollutants were almost completely removed (>99%)

tidal ﬂushing (Appendix VI). TKN was slightly reduced, while TP

by the non-mangrove CWs, irrespective of planting arrangement −

was at least doubled in all CWs. NO3 was reduced in all CWs with

(Appendix IV). −

different extents. Without tidal ﬂushing, NO3 was generally lower

Considering the Water Pollution Control Ordinance in Hong

in the planted CWs than the control. Heavy metals (except Fe) were

Kong (Cap. 358, section 21), the efﬂuent treated by the mangrove

enriched with different extents, while the degree of enrichment

J.Y.S. Leung et al. / Ecological Engineering 95 (2016) 129–137 133

Table 4

The physico-chemical properties of sediment in the mangrove CWs before and after 6-month irrigation with artiﬁcial wastewater (mean ± S.D., n = 6, except n = 3 for PAHs

and phenol).

With tidal ﬂushing Without tidal ﬂushing

Initial Ac Bg Ac + Bg Control Ac Bg Ac + Bg Control

Physical properties

pH 7.47 7.26 ± 0.10 7.34 ± 0.12 7.24 ± 0.10 7.32 ± 0.14 7.33 ± 0.14 7.26 ± 0.34 7.16 ± 0.36 7.34 ± 0.22

Redox (mV) – 247 ± 39.3 −31.9 ± 61.1 230 ± 55.5 11.3 ± 40.9 220 ± 131 274 ± 52.4 265 ± 63.6 67.0 ± 91.4 Nutrients

−1

TKN (g kg ) 0.47 0.30 ± 0.13 0.33 ± 0.09 0.42 ± 0.08 0.45 ± 0.12 0.34 ± 0.10 0.38 ± 0.06 0.35 ± 0.07 0.33 ± 0.06

−1

TP (g kg ) 0.14 0.27 ± 0.09 0.36 ± 0.08 0.32 ± 0.02 0.32 ± 0.07 0.30 ± 0.07 0.33 ± 0.03 0.32 ± 0.06 0.31 ± 0.02

+ −1

± ±

NH4 (mg kg ) 3.19 0.80 0.44 0.46 0.10 0.60 0.22 0.60 0.40 0.45 ± 0.18 0.41 ± 0.14 0.60 ± 0.35 0.39 ± 0.11

− −1

NO3 (mg kg ) 1.92 1.29 ± 0.61 1.31 ± 0.60 1.63 ± 0.23 1.23 ± 0.30 0.86 ± 0.40 0.61 ± 0.28 1.30 ± 0.86 1.69 ± 0.22

Heavy metals

−1

Cd (mg kg ) 0.12 0.33 ± 0.15 0.10 ± 0.11 0.53 ± 0.38 0.20 ± 0.24 0.23 ± 0.08 0.15 ± 0.07 0.17 ± 0.12 0.80 ± 0.58

−1

Cr (mg kg ) 7.49 13.6 ± 4.68 12.9 ± 4.58 19.0 ± 10.2 13.7 ± 8.98 13.1 ± 3.22 13.5 ± 3.10 15.0 ± 8.18 17.5 ± 8.65

−1

Cu (mg kg ) 9.97 37.1 ± 12.1 20.2 ± 3.27 39.3 ± 5.06 20.9 ± 4.24 32.2 ± 7.24 27.4 ± 7.78 26.5 ± 7.61 36.9 ± 18.9

−1

Fe (g kg ) 7.75 6.66 ± 2.70 7.94 ± 2.03 9.03 ± 5.19 8.11 ± 5.20 7.34 ± 1.38 6.54 ± 1.15 7.49 ± 1.78 7.42 ± 1.47

−1

Mn (mg kg ) 58.4 92.7 ± 9.07 101 ± 31.8 134 ± 48.3 101 ± 53.6 92.1 ± 15.2 118 ± 55.6 86.4 ± 11.2 124 ± 38.9

−1

Ni (mg kg ) 2.20 12.3 ± 6.85 7.50 ± 2.39 10.2 ± 4.81 6.85 ± 1.78 6.98 ± 0.28 8.49 ± 3.80 6.44 ± 1.24 13.7 ± 5.93

−1

± ±

Pb (mg kg ) 15.1 21.2 2.53 19.2 2.75 25.9 10.9 21.6 ± 3.81 22.0 ± 2.37 19.3 ± 3.55 27.1 ± 14.7 27.1 ± 7.19

−1

Zn (mg kg ) 38.2 66.1 ± 17.5 49.9 ± 6.47 67.9 ± 17.2 49.8 ± 6.51 75.8 ± 36.3 62.8 ± 8.12 66.3 ± 42.2 73.9 ± 11.9

Organic pollutants

−1

Phe (mg kg ) 0.02 0.028 ± 0.001 0.038 ± 0.010 0.031 ± 0.006 0.027 ± 0.004 0.028 ± 0.004 0.029 ± 0.005 0.031 ± 0.005 0.025 ± 0.005

−1

Pyr (mg kg ) 0.009 0.009 ± 0.002 0.013 ± 0.006 0.010 ± 0.004 0.010 ± 0.001 0.011 ± 0.002 0.013 ± 0.006 0.009 ± 0.001 0.009 ± 0.003

−1

BaP (mg kg ) N.D. 0.010 ± 0.003 0.006 ± 0.001 0.007 ± 0.002 0.012 ± 0.004 0.008 ± 0.001 0.011 ± 0.002 0.007 ± 0.001 0.013 ± 0.005

−1

Phenol (mg kg ) N.D. 1.26 ± 0.08 1.08 ± 0.09 1.31 ± 0.07 1.18 ± 0.04 1.09 ± 0.14 1.21 ± 0.28 1.27 ± 0.24 1.14 ± 0.27

was neither affected by planting arrangement nor tidal ﬂushing, were observed in C. indica and P. australis. In general, plants with

except Cd, Cu and Ni (Table 4). Without tidal ﬂushing, Cd and Ni faster growth and higher biomass are more ideal for wastewater

were signiﬁcantly more enriched in the control than the planted treatment (Thullen et al., 2005; Konnerup et al., 2009). Given the

CWs (Appendix VI). As for the organic pollutants, Phe was slightly growth and health status of plants, mangrove plants are more com-

enriched, while Pyr remained more or less the same. BaP and phenol petent than non-mangrove plants, while A. corniculatum is more

became detected, indicating their enrichment. ideal than B. gymnorrhiza for mangrove CWs.

Regarding the non-mangrove CWs, the pH of sediment largely

ﬂuctuated among CWs where higher pH was found in Pa, Ci + Pa

4.2. Removal of nutrients

and control than Acc + Ci and Acc + Pa (Table 5). Higher redox poten-

tial was found in the planted CWs than the control, except Pa and

The removal percentage of TOC was high (70.8–92.3%) in both

Ci + Pa. TKN was enriched in some of the CWs where the highest

mangrove and non-mangrove CWs. Apart from adsorption onto the

concentration was found in the control. TP was enriched by approx-

sediment (Abou-Elela and Hellal, 2012; Zhang et al., 2012), removal

+ −

imately 13 times in all CWs. NH4 and NO3 were signiﬁcantly

of TOC can be enhanced by the aerobic condition which favours

reduced in all CWs. Heavy metals were signiﬁcantly enriched,

degradation of organic compounds by heterotrophic microorgan-

except Cr, Fe and Pb (Table 5). Although the effect of planting

isms (Saeed and Sun, 2012). Despite the plant growth in both

arrangement was not obvious possibly due to the large variations,

mangrove and non-mangrove CWs, the planted CWs had similar

it is commonly observed that heavy metals were more enriched in

removal percentage of TOC as the control probably because the

the control than the planted CWs, such as Cd, Cu and Mn. All of the

input of plant litter offsets the plant uptake (Gu et al., 2006). Com-

organic pollutants were signiﬁcantly enriched in the sediment with

paring the mangrove CWs to the non-mangrove CWs, the latter

different extents.

had slightly higher removal of TOC, attributed at least partly to the

higher adsorption capacity of silty soil. In the mangrove CWs, tidal

4. Discussion ﬂushing slightly enhanced the removal of TOC. We suggest that the

anaerobic condition created by tidal ﬂushing allows denitriﬁcation

4.1. Effect of wastewater on the health status of plants where organic compounds are consumed (Saeed and Sun, 2012).

Nitrogenous compounds can be removed by various processes,

To select suitable plant species for CWs, tolerance to toxic including ammoniﬁcation, ammonia volatilization, nitriﬁcation,

wastewater is a critical criterion. In this study, A. corniculatum con- denitriﬁcation, plant uptake and adsorption (Saeed and Sun, 2012).

tinued to grow with high survival rate throughout the experimental Concerning NH4 , ammonia volatilization is negligible in light of

period, indicating its strong tolerance to the toxic wastewater. the low pH (<8.0) of the sediment (Saeed and Sun, 2012). While the

Previous studies showed that its roots can act as a barrier to pre- slightly higher removal in the planted CWs than the control could

clude translocation of heavy metals to the aerial parts (Wong et al., be due to plant uptake (Konnerup and Brix, 2010), we propose that

1997a), thereby minimizing their toxicity. Tidal ﬂushing can fur- nitriﬁcation is the critical process, given the aerobic condition in

ther enhance the growth of A. corniculatum, probably by providing both mangrove and non-mangrove CWs. The more aerobic condi-

optimal salinity or diluting the wastewater in the CWs (Wu et al., tion (i.e. higher redox potential) created by the plants can enhance

2008; Su et al., 2011). The faster growth with tidal ﬂushing could nitriﬁcation due to proliferation of ammonia-oxidizing bacteria and

partly explain the higher removal percentage of nutrients (see Sec- nitrifying bacteria (Vymazal, 2007; Tao and Wang, 2009; Peng et al.,

tion 4.2 for more discussion). In contrast, B. gymnorrhiza showed 2014). The optimal pH (i.e. slightly above 7) for nitriﬁcation in the

limited growth, especially when tidal ﬂushing was not applied, mangrove CWs could be related to the higher removal of NH4 than

possibly indicating the harmful effect of the toxic wastewater. that in the non-mangrove CWs (Abou-Elela and Hellal, 2012). Tidal

The non-mangrove plants were even more vulnerable to the toxic ﬂushing further facilitated the removal of NH4 in the mangrove

wastewater. For example, wilting, chlorosis and leaf senescence CWs. We reason that the extra removal of NH4 is attributed to

134 J.Y.S. Leung et al. / Ecological Engineering 95 (2016) 129–137

Table 5

The physico-chemical properties of sediment in the non-mangrove CWs before and after 6-month irrigation with artiﬁcial wastewater (mean ± S.D., n = 6, except n = 3 for

PAHs and phenol). The different superscript letters within each parameter indicate signiﬁcant difference according to PERMANOVA (p ≤ 0.05).

Initial Acc Ci Pa Acc + Ci Acc + Pa Ci + Pa Control

Physical properties

bc ab a c c a a

pH 5.99 5.81 ± 0.27 6.17 ± 0.34 6.38 ± 0.24 5.53 ± 0.29 5.50 ± 0.12 6.18 ± 0.09 6.37 ± 0.25

a a b a a ab b

Redox (mV) – 329 ± 7.75 374 ± 7.22 253 ± 63.4 385 ± 15.8 359 ± 8.84 299 ± 103 232 ± 12.7 Nutrients

−1 ab b b b ab ab a

TKN (g kg ) 0.59 0.69 ± 0.07 0.61 ± 0.08 0.60 ± 0.15 0.60 ± 0.04 0.69 ± 0.07 0.70 ± 0.07 0.83 ± 0.05

−1

TP (g kg ) 0.26 3.80 ± 0.38 3.54 ± 0.50 3.60 ± 0.57 3.34 ± 0.40 3.95 ± 1.32 3.85 ± 0.46 3.79 ± 0.31

+ −1

NH4 (mg kg ) 4.37 0.90 ± 0.44 0.54 ± 0.25 0.63 ± 0.30 0.90 ± 0.40 0.61 ± 0.45 0.54 ± 0.06 0.65 ± 0.27

− −1

NO3 (mg kg ) 31.6 1.81 ± 0.53 1.46 ± 0.25 1.74 ± 0.38 0.99 ± 0.46 0.98 ± 0.53 1.53 ± 0.46 1.09 ± 0.65

Heavy metals

−1 ab b a ab a ab ab

Cd (mg kg ) 0.26 0.57 ± 0.22 0.35 ± 0.18 0.75 ± 0.26 0.49 ± 0.22 0.88 ± 0.19 0.67 ± 0.14 0.71 ± 0.29

−1

Cr (mg kg ) 21.8 19.2 ± 2.16 17.0 ± 4.27 21.1 ± 7.55 18.1 ± 1.90 20.2 ± 2.94 20.0 ± 1.92 22.0 ± 6.78

−1 ab b ab b ab ab a

Cu (mg kg ) 17.0 41.7 ± 14.8 29.8 ± 10.1 45.7 ± 12.5 33.1 ± 4.36 48.0 ± 7.19 39.5 ± 3.72 53.6 ± 14.6

−1

± ± ± ±

Fe (g kg ) 9.10 8.13 0.98 7.59 0.98 10.2 2.22 7.28 0.81 10.6 2.95 9.97 ± 2.00 10.2 ± 2.76

−1

Mn (mg kg ) 47.0 111 ± 15.7 91.5 ± 24.3 137 ± 63.4 94.3 ± 10.5 115 ± 13.9 111 ± 13.7 136 ± 58.2

−1

Ni (mg kg ) 3.10 13.7 ± 2.54 10.1 ± 2.77 17.5 ± 7.91 12.4 ± 2.27 14.2 ± 1.09 13.0 ± 1.62 17.9 ± 7.49

−1

Pb (mg kg ) 28.6 24.6 ± 1.30 24.3 ± 4.14 32.7 ± 10.9 25.4 ± 3.64 33.6 ± 6.05 29.5 ± 3.03 36.0 ± 17.3

−1

Zn (mg kg ) 43.8 152 37.7 92.4 21.3 134 ± 51.7 93.9 ± 11.6 140 ± 17.0 136 ± 36.5 127 ± 53.6

Organic pollutants

−1

Phe (mg kg ) 0.03 0.096 ± 0.044 0.131 ± 0.081 0.059 ± 0.021 0.079 ± 0.001 0.075 ± 0.023 0.038 ± 0.003 0.064 ± 0.014

−1

Pyr (mg kg ) 0.010 0.022 ± 0.005 0.075 ± 0.090 0.016 ± 0.003 0.017 ± 0.003 0.020 ± 0.003 0.013 ± 0.002 0.033 ± 0.011

−1

BaP (mg kg ) N.D. 0.018 ± 0.002 0.019 ± 0.011 0.016 ± 0.002 0.013 ± 0.001 0.014 ± 0.002 0.011 ± 0.003 0.013 ± 0.003

−1

± ± ±

Phenol (mg kg ) N.D. 1.07 0.31 1.18 0.16 1.69 0.80 1.30 ± 0.05 1.43 ± 0.37 1.83 ± 0.11 1.43 ± 0.54

anaerobic ammonium oxidation (anammox) where NH4 is oxi- possible reason, given the higher mobility of heavy metals under

dized to nitrogen gas by nitrite under anaerobic condition created acidic and oxidized conditions (Ho et al., 2012). The removal per-

by tidal ﬂushing (Saeed and Sun, 2012). Further investigation is centage in the planted non-mangrove CWs was similar to, or even

required to evaluate the contribution by this process. lower than, that in the control. This not only implies that plant

In the mangrove CWs, plant uptake is responsible for the higher uptake was not discernible, but also the roots of non-mangrove

−

removal of NO3 . Nevertheless, the remarkable effect of tidal ﬂush- plants could curtail the removal of heavy metals in the CWs. Pre-

−

ing indicates that denitriﬁcation is the key process to remove NO3 , vious studies showed that roots can penetrate the soil and open

−

where denitrifying bacteria utilize NO3 as the terminal electron new pore channels and preferential pathways in the sediment, cul-

acceptor and organic carbon as electron donor (Vymazal, 2007; minating in better ﬂow and higher effective porosity (Brix, 1997;

Saeed and Sun, 2012). Unlike nitrifying bacteria, the growth of deni- Knowles et al., 2011; Hua et al., 2014). This proposition is substan-

trifying bacteria is promoted under anaerobic condition (Peng et al., tiated by the slightly lower removal percentage in the CWs planted

−

2014). Without tidal ﬂushing, NO3 accumulated in the efﬂuent with A. calamus which has higher root biomass than P. australis and

due to excessive nitriﬁcation (Maltais-Landry et al., 2009), indi- C. indica.

cated by the negative removal percentage. Regular tidal ﬂushing is, PAHs and phenol were effectively removed (>99% for most

therefore, necessary to allow both nitriﬁcation and denitriﬁcation samples) by both mangrove and non-mangrove CWs. Adsorption

to occur by creating an intermittent aerobic and anaerobic condi- onto the substrate or organic matter is also the key pathway to

tions in the sediment (Vymazal, 2007). Interestingly, the removal of remove these organic pollutants due to their high adsorption afﬁn-

−

NO3 in the non-mangrove CWs was high (>75.1%), despite the aer- ity (Polprasert et al., 1996; Cottin and Merlin, 2008). The adsorption

obic condition. The reason is still unknown, but it may be ascribed of organic pollutants can be reﬂected by their enrichment in the

to the microbial activity in the sediment since the plants had indis- sediment. Interestingly, the enrichment of Phe and Pyr in the non-

cernible effect. Overall, the removal of nitrogenous compounds, mangrove CWs was higher than the mangrove CWs, despite their

indicated by TKN, was more efﬁcient in the mangrove CWs than the complete removal by the CWs. This suggests that a certain portion

non-mangrove CWs, but tidal ﬂushing was necessary for removal of Phe and Pyr were lost in the mangrove CWs. Microbial degrada-

−

of NO3 in the mangrove CWs. tion, which is a prominent process in mangrove sediment (Yu et al.,

Phosphorus was effectively removed by both mangrove and 2005), is a possible reason for the loss. Yet, further investigation is

non-mangrove CWs. Removal of phosphorus is primarily associ- needed.

ated with physical processes, such as adsorption, precipitation and

accumulation in the organic matter or substrate (Vymazal, 2007).

4.4. Applicability and improvement of the CWs

These processes also explain why the sediment was signiﬁcantly

enriched with phosphorus. Microbial activity and plant uptake can

Despite the high removal of pollutants, removal of nutrients

also contribute to the removal of phosphorus (Vymazal, 2007).

(except TP) by the CWs was not very satisfactory. In addition to the

aforementioned reasons, the HRT in the CWs (about 0.5 day) may

4.3. Removal of pollutants not be long enough to maximize the removal of nutrients. Previous

studies showed that the removal percentage of nutrients increases

The removal percentage of heavy metals was very high, espe- with HRT because more time is allowed for the plants, microbes and

cially in the mangrove CWs. Adsorption is regarded as the main substrates to interact with the nutrients (e.g. uptake, decomposi-

pathway to remove heavy metals (Marchand et al., 2010). As such, tion, adsorption, denitriﬁcation, etc.) (Cui et al., 2010; Ghosh and

the higher removal percentage in the mangrove CWs than the Gopal, 2010; Su et al., 2011; Zhang et al., 2012; Sehar et al., 2015).

non-mangrove CWs can be explained by the higher degree of Technically, a deeper substrate bed should be used because it can

enrichment in the sediment. On the contrary, the lower removal increase HRT, make the sediment less aerobic to facilitate denitri-

percentage in the non-mangrove CWs suggests that the sediment ﬁcation, and minimize the land area required for constructing CWs

cannot substantially retain the heavy metals. Leaching could be a with longer HRT (Yadav et al., 2012). Nonetheless, HRT should not

J.Y.S. Leung et al. / Ecological Engineering 95 (2016) 129–137 135

Appendix I. The concentrations of nutrients, heavy metals

and organic pollutants in the artiﬁcial wastewater in the

present study.

−1 −1 −1

Nutrients Concentration (mg L ) Heavy metals Concentration (mg L ) Organic pollutants Concentration (mg L )

Dissolved organic carbon (DOC) 60 Cadmium (Cd) 0.1 Phenanthrene (Phe) 1

Total Kjeldahl nitrogen (TKN) 45 Chromium (Cr) 0.5 Pyrene (Pyr) 0.5

Ammonium (NH4 ) 25 Copper (Cu) 2 Benzo(a)pyrene (BaP) 0.1

−

Nitrate (NO3 ) 0.5 Iron (Fe) 30 Phenol 10

3−

Phosphate (PO4 ) 5 Lead (Pb) 1

Manganese (Mn) 5

Nickel (Ni) 1

Zinc (Zn) 5

be too long as larger land area is required and the treatment efﬁ-

ciency does is necessarily enhanced (Toet et al., 2005). We suggest

that HRT of 1.5–3 days should be adequate to enhance the removal

Appendix II. PERMANOVA table showing the effect of

of nutrients without remarkable increase in land usage.

planting arrangement and tidal ﬂushing on the removal

In this study, cultural arrangement had no signiﬁcant effect on

+ −

percentage of TKN, NH , NO and Mn in the mangrove

the treatment efﬁciency of CWs (see also Fraser et al., 2004; Chang 4 3

CWs. The different superscript letters in the pairwise

et al., 2012), but plant species should still be considered for the

comparison indicate signiﬁcant difference. TF: Tidal

design of CWs. In order to maximize the cost-effectiveness, the

ﬂushing; N.S.: Not signiﬁcant.

plant used in CWs should be perennial, tolerant to pollutants and

have long life span to allow long-term removal of pollutants with-

out regular harvest (Yang et al., 2008). The non-mangrove plants df MS Pseudo-F p Pairwise comparison

are rather susceptible to pollutants, meaning that they cannot not TKN

a ab b

be used for long-term removal of pollutants. In contrast, mangrove Arrangement 3 141 2.98 0.026 Ac + Bg Ac Bg

Controlb

plants can meet these requirements due to their strong tolerance,

a b

Flushing 1 1059 22.3 0.001 TF no TF

high biomass, high productivity and extensive root systems which

Arrangement 3 63.7 1.34 0.26

can facilitate the removal of pollutants (Wong et al., 1997b). Nev- × Flushing

ertheless, tidal ﬂushing must be provided to avoid the production NH4

−

of NO3 from the mangrove CWs. We recommend using gei wai Arrangement 3 237 5.53 0.001 Within TF: N.S.

Within no TF: Ac + Bg

design (Cha et al., 1997) or tidal operated CWs (Wu et al., 2015)

Bga Acb Controlb

to create the intermittent aerobic and anaerobic conditions in the a b

Flushing 1 1755 41.0 0.001 Within Ac: TF no TF

CWs so that the removal of both nutrients and pollutants can be

Within Bg: N.S.

maximized. Within Ac + Bg: N.S.

Within Control: TF no TFb

5. Conclusion

Arrangement 3 294 6.87 0.002

× Flushing

−

As CWs are increasingly used to treat wastewater particularly in NO3

Arrangement 3 17176 5.20 0.002 Within TF: Ac

the developing countries, it is important to maximize their treat-

ab ab b

Ac + Bg Bg Control

ment efﬁciency by optimizing the design. We revealed that the

Within no TF: Control

mangrove plants, especially A. corniculatum, had stronger tolerance b b b

Ac Bg Ac + Bg

5 a b

than the non-mangrove plants to toxic wastewater. The presence Flushing 1 3.54 × 10 107 0.001 Within Ac: TF no TF

a b

of plants can only slightly enhance the removal of nutrients in both Within Bg: TF no TF

Within Ac + Bg: TF no

mangrove and non-mangrove CWs, whereas the cultural arrange-

TFb

ment had no signiﬁcant effect. Pollutants were primarily removed a

Within Control: TF no

by adsorption onto the sediment. Overall, the mangrove CWs had TFb

higher removal percentage of both nutrients and pollutants than Arrangement 3 28256 8.57 0.001 −

the non-mangrove CWs, except NO3 . To make the mangrove CWs Flushing

− Mn

more versatile, tidal ﬂushing is needed not only to remove NO3 ,

+ Arrangement 3 40.5 0.14 0.936

but also to facilitate the removal of TOC, TKN and NH4 . Given a b

Flushing 1 1230 4.24 0.05 TF no TF

the health status of plants, treatment efﬁciency and compliance Arrangement 3 497 1.71 0.176

with discharge standards, we conclude that mangrove CWs have a Flushing

higher application value than the non-mangrove CWs to treat toxic

wastewater on condition that tidal ﬂushing is provided.

Acknowledgement

This work was supported by the Innovation and Technology

Fund of Hong Kong SAR (Project No.: GHP/004/08SZ).

Appendix III. The removal percentage of heavy metals and

organic pollutants by the mangrove CWs with or without

tidal ﬂushing (mean ± S.D., n = 9).

136 J.Y.S. Leung et al. / Ecological Engineering 95 (2016) 129–137

Removal percentage (%)

With tidal ﬂushing Without tidal ﬂushing

Ac Bg Ac + Bg Control Ac Bg Ac + Bg Control

Cd 98.98 ± 0.64 98.28 ± 1.02 98.93 ± 0.82 97.94 ± 2.37 99.80 ± 0.32 99.37 ± 0.95 97.23 ± 4.63 99.59 ± 0.97

± ±

Cr 99.95 0.11 99.95 0.15 99.89 ± 0.28 99.65 ± 1.05 99.88 ± 0.33 99.97 ± 0.08 99.76 ± 0.72 99.98 ± 0.05

Cu 98.20 2.60 97.17 ± 4.32 97.97 ± 3.02 94.76 ± 5.63 96.03 ± 6.22 96.74 ± 5.00 95.54 ± 8.42 96.79 ± 4.98

Fe 99.52 ± 1.02 99.16 ± 1.73 99.71 ± 0.36 99.79 ± 0.22 99.75 ± 0.31 99.62 ± 0.65 98.98 ± 2.60 99.37 ± 0.79

Mn 90.9 ± 7.20 81.8 ± 6.52 91.2 ± 6.34 86.4 ± 12.0 74.4 ± 33.5 88.1 ± 7.35 77.0 ± 22.9 77.1 ± 16.6

Ni 99.89 ± 0.19 99.77 ± 0.42 99.05 ± 2.80 99.57 ± 0.52 99.89 ± 0.22 98.52 ± 3.47 97.45 ± 4.20 98.07 ± 4.12

Pb 99.95 ± 0.09 99.93 ± 0.10 99.93 ± 0.09 99.97 ± 0.06 99.90 ± 0.14 99.83 ± 0.22 99.39 ± 1.54 99.70 ± 0.59

± ±

Zn 87.1 7.42 84.1 7.31 85.3 ± 7.16 85.6 ± 12.9 88.1 ± 11.0 86.0 ± 10.3 86.5 ± 11.8 93.4 ± 11.3

Phe 99.98 ± 0.01 99.98 ± 0.01 99.98 ± 0.02 99.98 ± 0.01 99.96 ± 0.03 99.97 ± 0.01 99.97 ± 0.02 99.97 ± 0.02

Pyr 99.96 ± 0.02 99.96 ± 0.02 99.97 ± 0.02 99.96 ± 0.02 99.94 ± 0.04 99.96 ± 0.01 99.94 ± 0.04 99.95 ± 0.02

BaP 99.91 ± 0.15 99.91 ± 0.12 99.94 ± 0.09 99.92 ± 0.17 99.81 ± 0.31 99.94 ± 0.09 99.80 ± 0.28 99.91 ± 0.11

Phenol 100 ± 0.00 99.99 ± 0.01 99.99 ± 0.01 99.99 ± 0.01 99.54 ± 1.24 99.94 ± 0.07 99.54 ± 1.14 99.94 ± 0.07

Appendix IV. The removal percentage of organic pollutants by the non-mangrove CWs (mean ± S.D., n = 9).

Removal percentage (%)

Acc Ci Pa Acc + Ci Acc + Pa Ci + Pa Control

Phe 99.94 ± 0.08 99.95 ± 0.03 99.96 ± 0.03 99.99 ± 0.02 99.96 ± 0.03 99.99 ± 0.2 99.95 ± 0.03

Pyr 99.83 ± 0.36 99.95 ± 0.03 93.93 ± 0.04 99.92 ± 0.06 99.88 ± 0.14 99.9 ± 0.07 99.93 ± 0.04

BaP 99.18 ± 2.04 99.72 ± 0.30 99.50 ± 0.57 99.81 ± 0.18 98.72 ± 2.58 97.32 ± 4.61 99.50 ± 0.57

± ±

Phenol 99.97 0.03 99.97 0.03 99.96 0.04 99.95 0.07 99.84 0.41 98.5 ± 3.03 99.94 ± 0.08

Appendix V. Comparisons of efﬂuent quality with the HK guideline values based on the Water Pollution Control Ordinance (Cap

358, section 21), which allows recreation in inland waters.

+ −

NH4 NO3 TKN TP Cd Cr Cu Fe Mn Ni Pb Zn Phenol

−1

HK guideline values (mg L ) 20 50 50 10 0.1 1 1 10 1 1 1 1 0.4

Mangrove CWs with tidal ﬂushing

√√√√√√√ √ √ √ √ × × Ac √√√√√√√√ √√ √ × ×

Bg √√

√ √ √ √ √ √ √ √ √

Ac + Bg × ×

Mangrove CWs without tidal ﬂushing

√ √ √ √ √ √ √ √ √ √ √ √

Ac ×

√ √ √ √ √ √ √ √ √ √ √

Bg × ×

√ √ √ √ √ √ √ √ √ √ √

Ac + Bg × ×

Non-mangrove CWs √√√ √ √ √

× × × × × × ×

Acc √

√ √ √ √ √ √

× × × × ×

Ci √ √ √ √ √ √ √

Pa × × × × × ×

√ √ √ √ √ √ √ √

× × × × ×

Acc + Ci √ √ √ √ √ √

Acc + Pa × × × × × × ×

√ √ √ √ √ √ √

Ci + Pa × × × × × ×

Appendix VI. PERMANOVA table showing the effect of

planting arrangement and tidal ﬂushing on redox potential,

−

NO3 , Cd, Cu and Ni in the sediment of the mangrove CWs.

Flushing 1 1.10 4.38 0.041 Within Ac: N.S.

a b

The different superscript letters indicate signiﬁcant Within Bg: TF no TF

difference. TF: Tidal ﬂushing; N.S.: Not signiﬁcant. Within Ac + Bg: N.S.

Within Control: no TF

TFb

df MS Pseudo-F p Pairwise comparison

Arrangement 3 1.03 4.10 0.012

Redox × Flushing potential Cd

5 a a Arrangement 3 0.45 4.96 0.008 Within TF: Ac + Bg

Arrangement 3 1.84 × 10 31.5 0.001 Within TF: Ac Ac + Bg ab bc c Controlb Bgb Ac Control Bg

a Within no TF: Control

Within no TF: Bg

b b b

a a b Ac Ac + Bg Bg

Ac + Bg Ac Control

5 Flushing 1 0.05 0.53 0.50 Within Ac: N.S.

Flushing 1 1.50 × 10 25.7 0.001 Within Ac: N.S.

a b Within Bg: N.S.

Within Bg: no TF TF

Within Ac + Bg: TF no

Within Ac + Bg: N.S.

TFb

Within Control: N.S.

Within Control: no TF

Arrangement 3 93573 16.0 0.001 TFb × Flushing

−

NO3 Arrangement 3 2.36 8.72 0.001

a × Flushing

Arrangement 3 1.28 5.09 0.003 Within TF: Ac + Bg Bgab Acab Controlb Cu

a a

a Arrangement 3 419 4.05 0.013 Within TF: Ac + Bg Ac

Within no TF: Control

ab bc c Bgb Controlb

Ac + Bg Ac Bg

Within no TF: N.S.

J.Y.S. Leung et al. / Ecological Engineering 95 (2016) 129–137 137

Flushing 1 41.0 0.40 0.54 Within Ac: N.S. Park, J.S., Brown, M.T., Han, T., 2012. Phenol toxicity to the aquatic macrophyte

a b

Within Bg: no TF TF Lemna paucicostata. Aquat. Toxicol. 106, 182–188.

Within Ac + Bg: TF no Peng, L., Hua, Y., Cai, J., Zhao, J., Zhou, W., Zhu, D., 2014. Effects of plants and

TF temperature on nitrogen removal and microbiology in a pilot-scale integrated

a vertical-ﬂow wetland treating primary domestic wastewater. Ecol. Eng. 64,

Within Control: no TF

TFb 285–290.

Polprasert, C., Dan, N., Thayalakumaran, N., 1996. Application of constructed

Arrangement 3 797 7.70 0.001

wetlands to treat some toxic wastewaters under tropical conditions. Water Sci.

× Flushing

Technol. 34, 165–171.

a Saeed, T., Sun, G., 2012. A review on nitrogen and organics removal mechanisms in

Arrangement 3 20.4 1.01 0.413 Within TF: Ac

subsurface ﬂow constructed wetlands: dependency on environmental

ab ab b

Ac + Bg Bg Control

parameters, operating conditions and supporting media. J. Environ. Manage.

Within no TF: Control 112, 429–448.

ab b b

Bg Ac Ac + Bg

Schützendübel, A., Polle, A., 2002. Plant responses to abiotic stresses: heavy

Flushing 1 4.18 0.21 0.672 Within Ac: N.S. metal-induced oxidative stress and protection by mycorrhization. J. Exp. Bot.

Within Bg: N.S. 53, 1351–1365.

Within Ac + Bg: N.S. Sehar, S., Sumera Naeem, S., Perveen, I., Ali, N., Ahmed, S., 2015. A comparative

Within Control: no TF study of macrophytes inﬂuence on wastewater treatment through subsurface

TF ﬂow hybrid constructed wetland. Ecol. Eng. 81, 62–69.

Singh, M., Müller, G., Singh, I.B., 2002. Heavy metals in freshly deposited stream

Arrangement 3 129 6.36 0.001

sediments of rivers associated with urbanisation of the Ganga Plain, India.

× Flushing

Water Air Soil Pollut. 141, 35–54.

Stottmeister, U., Wießner, A., Kuschk, P., Kappelmeyer, U., Kästner, M., Bederski, O.,

Müller, R., Moormann, H., 2003. Effects of plants and microorganisms in

References constructed wetlands for wastewater treatment. Biotechnol. Adv. 22, 93–117.

Su, Y.M., Lin, Y.F., Jing, S.R., Hou, P.C., 2011. Plant growth and the performance of

mangrove wetland microcosms for mariculture efﬂuent depuration. Mar.

Abou-Elela, S.I., Hellal, M.S., 2012. Municipal wastewater treatment using vertical

Pollut. Bull. 62, 1455–1463.

ﬂow constructed wetlands planted with Canna, Phragmites and Cyprus. Ecol.

Tao, W., Wang, J., 2009. Effects of vegetation, limestone and aeration on nitritation:

Eng. 47, 209–213.

anammox and denitriﬁcation in wetland treatment systems. Ecol. Eng. 35,

Alkio, M., Tabuchi, T.M., Wang, X., Colon-Carmona, A., 2005. Stress responses to

836–842.

polycyclic aromatic hydrocarbons in Arabidopsis include growth inhibition and

Thullen, J.S., Sartoris, J.J., Nelson, S.M., 2005. Managing vegetation in surface-ﬂow

hypersensitive response-like symptoms. J. Exp. Bot. 56, 2983–2994.

wastewater-treatment wetlands for optimal treatment performance. Ecol. Eng.

Brix, H., 1997. Do macrophytes play a role in constructed treatment wetland?

25, 583–593.

Water Sci. Technol. 35, 11–17.

Toet, S., Van Logtestijn, R.S.P., Kampf, R., Schreijer, M., Verhoeven, J.T.A., 2005. The

Cha, M.W., Young, L., Wong, K.M., 1997. The future of trational extensive (gei wai)

effect of hydraulic retention time on the removal of pollutants from sewage

shrimp farming at the Mai Po Marshes Nature Reserve, Hong Kong.

treatment plant efﬂuent in a surface-ﬂow wetland system. Wetlands 25,

Hydrobiologia 352, 295–303.

375–391.

Chang, J.J., Wu, S.Q., Dai, Y.R., Liang, W., Wu, Z.B., 2012. Treatment performance of

Vymazal, J., 2007. Removal of nutrients in various types of constructed wetlands.

integrated vertical-ﬂow constructed wetland plots for domestic wastewater.

Sci. Total Environ. 380, 48–65.

Ecol. Eng. 44, 152–159.

Wang, Y., Shi, J., Wang, H., Lin, Q., Chen, X., Chen, Y., 2007. The inﬂuence of soil

Cottin, N., Merlin, G., 2008. Removal of PAHs from laboratory columns simulating

heavy metals pollution on soil microbial biomass, enzyme activity, and

the humus upper layer of vertical ﬂow constructed wetlands. Chemosphere 73,

711–716. community composition near a copper smelter. Ecotoxicol. Environ. Saf. 67,

75–81.

Cui, L., Ouyang, Y., Lou, Q., Yang, F., Chen, Y., Zhu, W., Luo, S., 2010. Removal of

Wong, Y.S., Tam, N.F.Y., Chen, G.Z., Ma, H., 1997a. Response of Aegiceras

nutrients from wastewater with Canna indica L. under different vertical-ﬂow

corniculatum to synthetic sewage under simulated tidal conditions.

constructed wetland conditions. Ecol. Eng. 36, 1083–1088.

Hydrobiologia 352, 89–96.

Fraser, L.H., Carty, S.M., Steer, D., 2004. A test of four plant species to reduce total

Wong, Y.S., Tam, N.F.Y., Lan, C.Y., 1997b. Mangrove wetlands as wastewater

nitrogen and total phosphorus from soil leachate in subsurface wetland

treatment facility: a ﬁeld trial. Hydrobiologia 352, 49–59.

microcosms. Bioresour. Technol. 94, 185–192.

Wu, Y., Tam, N.F.Y., Wong, M.H., 2008. Effects of salinity on treatment of municipal

Ghosh, D., Gopal, B., 2010. Effect of hydraulic retention time on the treatment of

wastewater by constructed mangrove wetland microcosms. Mar. Pollut. Bull.

secondary efﬂuent in a subsurface ﬂow constructed wetland. Ecol. Eng. 36,

1044–1051. 57, 727–734.

Wu, S., Dong, X., Chang, Y., Carvalho, P.N., Pang, C., Chen, L., Dong, R., 2015.

Gu, B., Chimney, M.J., Newman, J., Nungesser, M.K., 2006. Limnological

Response of a tidal operated constructed wetland to sudden organic and

characteristics of a subtropical constructed wetland in south Florida (USA).

ammonium loading changes in treating high strength artiﬁcial wastewater.

Ecol. Eng. 27, 345–360.

Ecol. Eng. 82, 643–648.

Ho, H.H., Swennen, R., Cappuyns, V., Vassilieva, E., Van Gerven, T., Tran, T.V., 2012.

Wu, Q., Zhou, H., Tam, N.F.Y., Tian, Y., Tan, Y., Zhou, S., Li, Q., Chen, Y., Leung, J.Y.S.,

Potential release of selected trace elements (As, Cd, Cu, Mn, Pb and Zn) from

2016. Contamination, toxicity and speciation of heavy metals in an

sediments in Cam River-mouth (Vietnam) under inﬂuence of pH and oxidation.

industrialized urban river: implication for the dispersal of heavy metals. Mar.

Sci. Total Environ. 435, 487–498.

Pollut. Bull. 104, 153–161.

Hua, G.F., Zhao, Z.W., Kong, J., Guo, R., Zeng, Y.T., Zhao, L.F., Zhu, Q.D., 2014. Effects

Xu, J., Zhang, J., Xie, H., Li, C., Bao, N., Zhang, C., Shi, Q., 2010. Physiological

of plant roots on the hydraulic performance during the clogging process in

responses of Phragmites australis to wastewater with different chemical

mesocosm vertical ﬂow constructed wetlands. Environ. Sci. Pollut. Res. 21,

13017–13026. oxygen demands. Ecol. Eng. 36, 1341–1347.

Yadav, A.K., Abbassi, R., Kumar, N., Satya, S., Sreekrishnan, T.R., Mishra, B.K., 2012.

Kivaisi, A.K., 2001. The potential for constructed wetlands for wastewater

The removal of heavy metals in wetland microcosms: effects of bed depth,

treatment and reuse in developing countries: a review. Ecol. Eng. 16, 545–560.

plant species, and metal mobility. Chem. Eng. J. 211–212, 501–507.

Knowles, P., Dotro, G., Nivala, J., García, J., 2011. Clogging in subsurface-ﬂow

Yang, Q., Chen, Z.H., Zhao, J.G., Gu, B.H., 2007. Contaminant removal of domestic

treatment wetlands: occurrence and contributing factors. Ecol. Eng. 37,

99–112. wastewater by constructed wetlands: effects of plant species. J. Integr. Plant

Biol. 49, 437–446.

Konnerup, D., Brix, H., 2010. Nitrogen nutrition of Canna indica Effects of

Yang, Q., Tam, N.F.Y., Wong, Y.S., Luan, T.G., Su, W.S., Lan, C.Y., Shin, P.K.S., Cheung,

ammonium versus nitrate on growth, biomass allocation, photosynthesis,

S.G., 2008. Potential use of mangroves as constructed wetland for municipal

nitrate reductase activity and N uptake rates. Aquat. Bot. 92, 142–148.

sewage treatment in Futian, Shenzhen, China. Mar. Pollut. Bull. 57, 735–743.

Konnerup, D., Koottatep, T., Brix, H., 2009. Treatment of domestic wastewater in

Yu, S.H., Ke, L., Wong, Y.S., Tam, N.F.Y., 2005. Degradation of polycyclic aromatic

tropical, subsurface ﬂow constructed wetlands planted with Canna and

hydrocarbons by a bacterial consortium enriched from mangrove sediments.

Heliconia. Ecol. Eng. 35, 248–257.

Environ. Int. 31, 149–154.

Li, L., Yang, Y., Tam, N.F.Y., Yang, L., Mei, X.Q., Yang, F.J., 2013. Growth

Zhang, C.B., Wang, J., Liu, W.L., Zhu, S.X., Ge, H.L., Chang, S.X., Chang, J., Ge, Y., 2010.

characteristics of six wetland plants and their inﬂuences on domestic

Effects of plant diversity on microbial biomass and community metabolic

wastewater treatment efﬁciency. Ecol. Eng. 60, 382–392.

proﬁles in a full-scale constructed wetland. Ecol. Eng. 36, 62–68.

Li, Z., Xiao, H., Cheng, S., Zhang, L., Xie, X., Wu, Z., 2014. A comparison on the

Zhang, C.B., Liu, W.L., Wang, J., Ge, Y., Gu, B.H., Chang, J., 2012. Effects of plant

phytoremediation ability of triazophos by different macrophytes. J. Environ.

diversity and hydraulic retention time on pollutant removals in vertical ﬂow

Sci. 26, 315–322.

constructed wetland mesocosms. Ecol. Eng. 49, 244–248.

Maltais-Landry, G., Maranger, R., Brisson, J., 2009. Effect of artiﬁcial aeration and

Zhang, D.Q., Jinadasa, K.B.S.N., Gersberg, R.M., Liu, Y., Ng, W.J., Tan, S.K., 2014.

macrophyte species on nitrogen cycling and gas ﬂux in constructed wetlands.

Application of constructed wetlands for wastewater treatment in developing

Ecol. Eng. 35, 221–229.

countries—a review of recent developments (2000–2013). J. Environ. Manage.

Marchand, L., Mench, M., Jacob, D., Otte, M., 2010. Metal and metalloid removal in

141, 116–131.

constructed wetlands, with emphasis on the importance of plants and

standardized measurements: a review. Environ. Pollut. 158, 3447–3461.