Bioresource Technology 234 (2017) 243–252

Contents lists available at ScienceDirect

Bioresource Technology
journal homepage: www.elsevier.com/locate/biortech

Nitrogen removal by the enhanced floating treatment wetlands from the

secondary effluent
Lei Gao, Weili Zhou, Jungchen Huang, Shengbing He ⇑, Yijia Yan, Wenying Zhu, Suqing Wu, Xu Zhang
School of Environmental Science and Engineering, Shanghai Jiaotong University, Shanghai 200240, PR China

h i g h l i g h t s

Thiosulfate-driven autotrophic denitrification was firstly introduced into EFTW.

AEFTW and HEFTW could effectively remove nitrogen from the secondary effluent.

Addition of electron donors reduced the N2O emission.

The electron donors induced distinct shifts in microbial structures.

Mixotrophic denitrification occurred in HEFTW and AEFTW.

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

Article history: Three novel floating treatment wetlands, including autotrophic enhanced floating treatment wetland
Received 18 January 2017 (AEFTW), heterotrophic enhanced floating treatment wetland (HEFTW) and enhanced floating treatment
Received in revised form 2 March 2017 wetland (EFTW) were developed to remove nitrogen from the secondary effluent. Results showed that
Accepted 5 March 2017
the analogously excellent nitrogen removal performance was achieved in AEFTW and HEFTW. About
Available online 9 March 2017
89.4% of the total nitrogen (TN) was removed from AEFTW at a low S/N of 0.9 and 88.5% from HEFTW
at a low C/N of 3.5 when the hydraulic retention time (HRT) was 1 d in summer. Higher nitrification
Keywords:
and denitrification performance were achieved in AEFTW. Addition of electron donors effectively reduced
Secondary effluent
Enhanced floating treatment wetland
the N2O emission, especially in summer and autumn. High-throughput sequencing analysis revealed that
Autotrophic and heterotrophic the electron donors distinctly induced the microbial shifts. Dechloromonas, Thiobacillus and Nitrospira
denitrification became the most predominant genus in HEFTW, AEFTW and EFTW. And autotrophic and heterotrophic
N2O emission denitrification could simultaneously occur in HEFTW and AEFTW.
Microbial community Ó 2017 Elsevier Ltd. All rights reserved.

1. Introduction (GB3838-2002, 2002)). Therefore, improving TN removal from sec-

The excessive N, mainly from fertilizer use, crop N fixation and
industrial activities, causes negative environmental problems such
as eutrophication of terrestrial and aquatic systems, global acidifi-
cation (Gruber and Galloway, 2008). China now faces serious prob-
lems of surface water contamination. 11.5% of major freshwater
lakes and reservoirs in China were rated as severely polluted and
26.2% as slightly polluted (Li et al., 2015).
The secondary effluent from the municipal sewage treatment
plants entering the surface water will cause water quality deterio-
ration and eutrophication of rivers and lakes. As the total nitrogen
(TN) concentration (12–15 mg L1) in the secondary effluent is still
much higher than the surface water criteria (2.0 mg L1 for grade V
in Chinese National Surface Water Environmental Quality Standard
⇑ Corresponding author.
E-mail addresses: heshengbing@sjtu.edu.cn, shengbing_he@163.com (S. He).
ondary effluent has become an urgent need all over China.
The floating treatment wetlands (FTWs), as an innovative vari-
ant of constructed wetlands, has been well applied for water qual-
ity improvement of stormwater, polluted river water, eutrophic
water and secondary effluent due to its merits of low cost, occupy-
ing no additional land area and good suitability for either new con-
struction or retrofit installation on the existing ponds and lakes
(Borne et al., 2014; Karine et al., 2013). The enhanced FTWs
(EFTWs), composed of the bio-carriers and emergent plants, pro-
viding additional substrate surface for the attachment of microor-
ganisms, has been introduced to improve nutrient removal
efficiency (Wu et al., 2016; Zhang et al., 2015). However, the
nitrate removal performance of EFTWs is still unsatisfactory in
treating the secondary effluent, mainly due to the following rea-
sons: 1) the treatment performance is limited by the growth rate
of hydrophytes, especially at low temperature; 2) the secondary
effluent is lack of biodegradable organic carbon and the endoge-

http://dx.doi.org/10.1016/j.biortech.2017.03.036
0960-8524/Ó 2017 Elsevier Ltd. All rights reserved.
244 L. Gao et al. / Bioresource Technology 234 (2017) 243–252
Nomenclature
FTW floating treatment wetland
EFTW enhanced floating treatment wetland
HEFTW heterotrophic enhanced floating treatment wetland
AEFTW autotrophic enhanced floating treatment wetland
AD autotrophic denitrification
TN total nitrogen
DO dissolved oxygen
nous organic carbon in EFTWs from plant exudates, plant litter and
microbial decomposition, which does not meet the need of com-
plete denitrification. Therefore, when treating the secondary efflu-
ent, the external electron donors are required. Organic electron
donors, such as glucose, sodium acetate, methanol and ethanol
are the most frequently used. However, they may cause the sec-
ondary contamination, and add an extra operation cost on sludge
disposal (Zhu and Getting, 2012). Recently, inorganic electron
donors, such as the reductive sulfur compounds, have attracted
significant attention, as they require no external organic carbon
source and produce lower amount of excess sludge (Chung et al.,
2014). In addition, thiosulfate (S2O2 3 ) has exhibited higher
bioavailability and better denitrification rate, compared with ele-
mental sulfur and sulfide (Sahinkaya and Dursun, 2015; Zhou
et al., 2016). Thiosulfate-driven AD occurs according to the stoi-
chiometric equation (Mora et al., 2014):
  þ
S2 O2
3 þ 1:16NO3 þ 0:035CO2 þ 0:519HCO3 þ 0:110NH4
þ 0:124H2 ! 0:110C 5 H7 O2 N þ 0:578N2 þ 0:435Hþ þ 2SO2
4
ð1Þ
To our knowledge, so far there is no case introducing
thiosulfate-driven AD into EFTWs.
Nitrous oxide (N2O) is a powerful greenhouse gas with 298
times global warming potential of CO2 and is now increasing glob-
ally at a rate of 0.2–0.3% per year (IPCC, 2007; Ravishankara et al.,
2009). N2O emissions have been extensively studied in riparian
wetlands and constructed wetlands (García-Lledó et al., 2011;
Mander et al., 2015), but rarely in EFTWs.
For EFTWs, microbial activities play a vital role in N removal.
The microbial community structure provides an approach to reveal
the biological nitrogen removal mechanism in EFTWs. Different
electron donors may influence the microbial composition, leading
to different macroscopic nitrogen removal efficiency. However,
detailed information of functional microorganisms, microbial com-
position and variation in EFTWs has not yet been reported.
In this study, three EFTWs, including AEFTW, HEFTW and EFTW,
were introduced to remove nitrogen from the secondary effluent.
All three EFTWs tanks were identical. EFTW served as a control
reactor with no addition of external electron donor, while AEFTW
was added with thiosulfate and HEFTW with acetate as the elec-
tron donor. The objectives of this study are 1) to compare the nitro-
gen removal performance in AEFTW, HEFTW and EFTW; 2) to
evaluate N2O emission in EFTWs and explore the influence of dif-
ferent electron donors on N2O emission; 3) to investigate the
microbial community and biological nitrogen removal mecha-
nisms in AEFTW, HEFTW and EFTW.
2. Materials and methods
2.1. Experimental setup and operation
The experiment was operated in the botanic garden
(121°260 3100 N, 31°20 500 E) of Shanghai Jiao Tong University, City of
TP total phosphorus
CW constructed wetland
HD heterotrophic denitrification
HRT hydraulic retention time
COD chemical oxygen demand
RDA redundancy analysis
Shanghai, China. An artificial river flows through the botanic gar-
den. Three identical EFTWs tanks (L  W  H = 0.85 m 
0.40 m  0.90 m), each consisting of an inlet zone and a reaction
zone (Fig. 1), were installed in the botanic garden near the river.
The operational water depth was 0.8 m and the effective volume
of reaction zone was 224 L. A stainless steel frame (0.6  0.38 m,
L  W) with nine holes was fixed to the tank to support the
plants. Hydroponic pots with inner diameter of 9.5 cm were placed
in the frame. Iris pseudacorus were transplanted into the pots with
a plant density of 18 rhizomes m2. The plants used in the
experiment were of similar height and weight. Six strings of
spherical polypropylene bio-carriers were hung under the stainless
steel frame, every string was made up of five spherical fillers
with the diameter of 15 cm and the specific surface area of
380 m2 m3.
The simulated secondary effluent was prepared from the river
water of botanic garden, supplemented with sodium nitrate
(KNO3), calcium nitrate tetrahydrate (Ca(NO3)24H2O), ammonium
chloride (NH4Cl) and monosodium phosphate (NaH2PO4). The
influent concentration of total nitrogen (TN), nitrate nitrogen
(NO +
3 -N), ammonia nitrogen (NH4-N), total phosphorus (TP) and
sulfate (SO4 ) were 15.13 ± 1.1 mg N L1, 11.25 ± 1.23 mg N L1,
2
3.28 ± 0.5 mg N L1, 0.5 ± 0.2 mg P L1 and 77.01 ± 10.16 mg L1,
respectively. Moreover, sodium acetate was added into the influent
of HEFTW as electron donor for the heterotrophic denitrification
while sodium thiosulfate into AEFTW for the autotrophic denitrifi-
cation. The ratio of external COD/TN (C/N, W/W) was set as 3.5 and
5 and the S2O2 3 /TN (S/N, M/M) 0.9, 1. The EFTWs experiment was
operated in a continual flow mode for 330 days from March 15,
2015 to February 1, 2016.
2.2. Water sampling and chemical analysis
Water samples were taken once every three days from the three
EFTWs. COD, TN, NO  +
3 -N, nitrite (NO2 -N) and NH4-N of water sam-
ple were measured according to the standard method (APHA,
2005). Sulfate (SO2 2
4 ) and thiosulphate (S2O3 ) were analyzed by
ion chromatography (Metrohom 883, Metrohom, Switzerland).
DO, pH and temperature were detected by DO meter (HQ30d,
HACH) and pH meter (HQ11d, HACH), respectively. All measure-
ments of each sample were determined in three replicates.
2.3. Sludge sampling and molecular analyses
In order to investigate the variation of microbial community
composition in the three EFTWs, the sludge samples H1, H2; A1,
A2; E1, E2 were collected from plant rhizosphere of HEFTW, AEFTW,
EFTW tanks on May 29, 2015 and January 1, 2016, respectively, rep-
resenting the initial and final microbial sample of experiment,
respectively. The bacterial DNA was extracted using QIAamp DNA
stool Mini Kit. (QIAGEN, Germany) following the manufacturer’s
protocol. The extracted DNA was amplified with the primers 338F
(ACTCCTACGGGAGGCAGCAG) and 533R (TTACCGCGGCTGCTGG
CAC) in the V3 domain of bacterial 16S rRNA genes. PCR amplifica-
 L. Gao et al. / Bioresource Technology 234 (2017) 243–252
Fig. 1. Schematic diagram of the enhanced floating treatment wetland.
tion program included initial denaturation at 98 °C for 2 min, fol-
lowed by 28 cycles of 98 °C for 15 s, annealing at 50 °C for 30 s,
and extension at 72 °C for 30 s, with a final elongation at 72 °C for
5 min (Zhou et al., 2017). Finally, High-throughput sequencing
was conducted with Illumina Hiseq 2000 by Shanghai Personal
Biotechnology Co., Ltd (Shanghai, China). Questionable sequences
and chimeric sequences were examined and removed by calling
USEARCH (v5.2.236, http://www.drive5.com/usearch/) via the
Qiime software (v1.9.0, http://qiime.org/). High-quality sequences
were clustered into the operational taxonomic units (OTU) by set-
ting a 97% Sequence similarity using UCLUST, and the OTU with
the abundance value less than 0.001% were eliminated (Bokulich
et al., 2013). The taxonomic sequences classification was conducted
using the Greengenes reference database at an 80% confidence
threshold. Rarefaction curves, bacterial community richness (Chao
1 and Ace) and diversity (Shannon and Simpson) were obtained
using the Mothur software package 1.9.0.
2.4. Gas sampling and N2O gas emission rate measurement
Gas sampling was carried out once a month from April 2015 to
January 2016. The static chamber technique was used to estimate
the N2O emissions in three different EFTWs (Wu et al., 2009). The
closed chamber was made of polymethyl methacrylate with a total
volume of 340 L. Battery-driven fans were operated inside the
chamber to ensure that the gas was well mixed during sampling.
Five gas samples were collected at 0, 30, 60, 90, 120 min after
enclosure between 8:00 and 12:00 am. The gas samples were
indrawn into 500 mL gas sampling bags (Eler Inc., Shanghai, China)
using a gas sampling pump. The N2O concentration was deter-
mined using an Agilent 6890 gas chromatograph (Agilent Tech-
nologies Inc., USA) with an electron capture detector (lECD) and
a Poropak Q column (2 m), using 15 mL min1 pure nitrogen
(99.999%) as the carrier gas. The temperatures for the column,
injector, and detector were set at 70, 100, and 280 °C, respectively.
The N2O flux (mg m2 d1) calculated from the slope of the linear
plot of the concentration versus time by taking the chamber’s sur-
face area into account (Uggetti et al., 2012).
245
2.5. Statistical analysis
All statistical analyses were conducted with SPSS version 20.0
software (SPSS Inc., Chicago, USA) and were considered significant
at 0.05 level. The differences in TN, NO + 
3 -N, NH4-N, NO2 -N, pH, DO
in EFTWs were tested by one-way ANOVA with Duncan’s post hoc
test. Variables that were not normally distributed were transform-
ing by the Box-cox transformation prior analyses. The Pearson cor-
relation and Spearman rank-order correlation were used to
characterize the relationship between temperature (T), HRTs and
N in effluents. The relationships between bacterial community
and physicochemical variables were analyzed using redundancy
analysis (RDA) in Canoco version 4.5 for Windows.
2.6. Calculation of the nitrogen removal efficiency, contribution of the
microbial transformation
The TN removal efficiency was calculated using the flowing
equation:
TNRE ð%Þ ¼ ðTNinf  TNeff Þ=TNinf  100% ð2Þ
where TNRE is the TN removal efficiency, TNinf represents the influ-
ent TN concentration (mg L1) and TNeff, the effluent TN concentra-
tion (mg L1).
For the calculation of the contribution of the microbial transfor-
mation to total nitrogen removal in HEFTW and AEFTW, the fol-
lowing assumptions were made:
1) The external electron donors only increased the microbial
transformation.
2) Nitrogen removal by other routes was assumed to be the
same in 3 EFTWs.
Therefore, it could be expressed as follows:
Microbial transformation contribution of AEFTW or HEFTW
¼ ðTNREAEFTW or HEFTW  TNREEFTW Þ=TNREAEFTW or HEFTW  100%
ð3Þ
246 L. Gao et al. / Bioresource Technology 234 (2017) 243–252
where TNREAEFTW or HEFTW is the TN removal efficiency of AEFTW or
HEFTW, TNREEFTW stands for the TN removal efficiency of EFTW.
3. Results and discussion
3.1. Nitrogen removal performance
Fig. 2 showed the concentration of different forms of N in the
influent and effluent throughout the experiment period. The aver-
age effluent TN of HEFTW (2.80 mg L1, b) and AEFTW
(3.63 mg L1, b) were significantly lower than that of EFTW
(12.04 mg L1, a). During the start-up stage (0–75 days in spring),
C/N, S/N were set as 5 and 1 for HEFTW and AEFTW and HRT
was 3 days. TN concentration in HEFTW decreased immediately
from 5.3 to 2.25 mg L1 in first 30 days and then remained steady
at approximately 2 mg L1 till day 220. Differently, effluent of
AEFTW showed a gradual decrease from 8.59 to 4.02 mg L1 for
2 months. The phenomenon might be due to the fact that the het-
erotrophic denitrifying bacteria proliferated faster than the auto-
trophic one. The optimal effluent TN of HEFTW, AEFTW
decreased to 1.23, 0.80 mg/L at low C/N (3.5), S/N (0.9) and HRT
(1d) in August while peak nitrogen removal efficiency (39%) of
EFTW was obtained in July at the HRT of 3 d. It was believed that
the high temperature in summer accelerated the denitrifying rate,
promoted the plant growing rate, leading to more uptake of nitro-
gen, and also promoted the production of endogenous organic mat-
ters from root exudates and died plant detritus (Borne et al., 2013).
Decrease of TN removal efficiency was not found in HEFTW and
Fig. 2. TN, NO + 
3 -N, NH4-N and NO2 -N concentration variations of influent and effluent and temperature change in HEFTW, AEFTW and EFTW.
AEFTW when HRT was shortened at the temperature above
20 °C, but indeed existed in EFTW. TN effluents increased to 5.40,
6.22 and 14.6 mg L1 in HEFTW, AEFTW and EFTW, when the tem-
perature dropped to 10 °C and below, The reason probably was
that the low temperature slowed down the microbial activity (Xu
et al., 2016). To improve nitrogen removal performance in cold
weather, C/N and S/N were increased to 5 and 1, same as the
start-up stage. Then TN decreased to 3.34 and 3.59 mg L1 in
HEFTW and AEFTW, suggesting that the TN removal efficiency
could be improved effectively at low temperature by increasing
the electron donor supply for microorganism utilization. TN in
the effluents correlated with temperature (for EFTW, r = 0.821,
p < 0.001; HEFTW, r = 0.779, p < 0.001; and AEFTW, r = 0.764,
p < 0.001). With the temperature dropped from 30 to 10 °C, the
TN removal efficiency decreased by 34.3%, 11.2% and 16.5% in
EFTW, HEFTW and AEFTW, respectively. And this indicated that
the external electron donors reduced the dependence of TN
removal on temperature.
The effluent NO3 -N showed a similar trend as TN (Fig. 2C) for
the reason that the influent was rich in nitrate. Nitrate removal
efficiency was approximately 100% from April to December as a
result of high denitrifying efficiency in HEFTW and AEFTW. How-
ever, the effluent NO 3 -N in AEFTW slightly increased from day
178. The nitrate in EFTW effluent exceeded the influent from day
0 to 45 and day 231 to 330, revealing the loss of microbial capacity
of denitrifiers in EFTW at low temperature, probably due to the
lack of biodegradable organics. And the effluents NO 3 -N increased
with water temperature decreasing (r = 0.842, p < 0.001,
r = 0.758, p < 0.001 and r = 0.444, p = 0.044 for EFTW, AEFTW
 L. Gao et al. / Bioresource Technology 234 (2017) 243–252
and HEFTW, respectively). This result illustrated that the denitrifi-
cation rate was strongly affected by temperature and that the het-
erotrophic denitrifiers in HEFTW was most robust to overcome the
adverse effects of temperature.
As shown in Fig. 2B, from day 0 to 200, the effluents NH+4-N in
HEFTW, AEFTW and EFTW showed no significant difference
(p > 0.05) from each other with the average concentration of
0.63, 0.75 and 0.43 mg L1. However, the average effluent NH+4-N
of HEFTW (2.01 mg L1, a) turned to be much higher than that of
AEFTW (0.89 mg L1, b) and EFTW (0.54 mg L1, b) after day 200
with the highest NH+4-N concentration reaching 2.70 mg L1 and
the lowest NH+4-N removal efficiency of 24%. The most likely reason
was that the heterotrophic denitrifiers outcompete the nitrifiers
rivals to thrive in the system after 2 months of operation at a short
HRT (1 d) in the anoxic environment (as seen in Fig. S2). During the
last stage of the experiment, the effluent nitrate decreased remark-
ably with the addition of acetate in HEFTW, suggesting that the
acetate addition stimulated the thriving of heterotrophic denitri-
fiers, leading to much higher denitrify rate. Additionally, the efflu-
ent NH+4-N increased when HRT was shortened to 1 d, probably due
to the fact that the autotrophic nitrifying bacteria required a long
reproductive time to accomplish the nitrification process in EFTWs.
The average concentrations of effluent NO 2 -N of 3 EFTWs were
as follows: AEFTW (0.23 mg L1, b) >HEFTW (0.14 mg L1, ab)
>EFTW (0.10 mg L1, a), which proved that the autotrophic denitri-
fiers were more effective in converting NO 
3 to NO2 than the het-
erotrophic denitrification (Chen et al., 2009; Reyes-Avila et al.,
2004). The concentration of NO 2 in AEFTW increased when HRT
was decreased to 1 d, due to the slower denitrifying rate of AD than
HD. Nevertheless, NO 2 -N was resumed to lower concentration
after the autotrophic denitrifiers adapted to the short HRT.
As seen in Fig. 3, TN removal efficiencies were 80.5, 88.5, 84.3,
74.6% in HEFTW and 59.9, 89.4, 84.1, 71.8% in AEFTW in spring,
summer, autumn and winter, significantly higher than 17.5, 28.9,
20.4, 11.2% in EFTW. It could be roughly calculated that the micro-
bial transformation contributed about 67.3–85.0% to the total
nitrogen removal in HEFTW and 67.7–84.4% in AEFTW according
to Eqs. (2) and (3). This finding was consistent with the previous
reports that the microbial conversion was the dominant pathway
to nitrogen removal (Reinhardt et al., 2006; Søvik and Mørkved,
2008). The nitrogen removal performance in HEFTW and AEFTW
were vastly superior to a hybrid floating treatment bed using plant
and periphyton (30% in 3 days) (Liu et al., 2016) and an enhanced
ecological floating bed of only 15.3% (Wu et al., 2016).
Fig. 3. Concentration of N species in effluent and TN removal efficiency of HEFTW (A), AEFTW (B), and EFTW (C) in different seasons. (spr, sum, aut and win represent the
seasons of spring, summer, autumn and winter).
247
In spring, the average effluent TN of AEFTW and HEFTW was
5.44 mg L1 and 2.23 mg L1. And the extra N in AEFTW was
mainly NO 3 -N, implying that the autotrophic denitrifies required
longer time to adapt to a new water environment. Nonetheless,
the effluent TN of HEFTW and AEFTW were extremely close in
autumn and winter. NH+4-N accounted for the majority of TN in
HEFTW whereas NO 3 -N was overwhelming majority in AEFTW.
Higher NH+4-N concentration in HEFTW indicated that nitrification
was most likely the rate-limiting step, this was in accordance with
the low DO in HEFTW (as discussed in Section 3.4) and the obser-
vation of low abundance of nitrifying bacteria in the biomass (as
discussed in Section 3.3). The higher NO 3 -N in AEFTW might be
due to the decline of the microbial activity of the autotrophic den-
itrifiers at lower temperature. And the effluent NO 3 -N was effec-
tively reduced after the S/N increased to 1 from day 290
(Fig. 2C). The proportion of NO 3 -N in EFTW effluent exceeded
90% in one year operation, which illustrated that the denitrification
was limited because of the shortage of labile organic compounds
from the influent and plant generation.
No significant difference was observed among the effluent COD
of EFTW (28.8 ± 7.10 mg L1), HEFTW (30.6 ± 7.91 mg L1) and
AEFTW (29.5 ± 7.22 mg L1), suggesting that the added acetate
was completely consumed and the amount of external carbon
would not cause the secondary contamination. The TP contents
in the effluent of EFTW, HEFTW and AEFTW were 0.19 ± 0.07,
0.20 ± 0.09 and 0.17 ± 0.09 mg L1, respectively. Meanwhile, the
effluent S2O2 3 was undetected in the most time except some
unstable running days and the average concentration was
0.67 ± 2.24 mg L1, and the effluent SO2 4 in AEFTW were
232.4 ± 34.9 mg L1, proving the AEFTW with thiosulfate-driven
AD a promising process for treating the secondary effluent.
3.2. N2O fluxes in different seasons
The mean N2O fluxes were 2.64, 1.19, 2.44, 3.67 mg N m2 d1
in HEFTW; 5.35, 1.14, 5.72, 6.48 mg N m2 d1 in AEFTW and
5.96, 5.72, 5.60, 6.61 mg N m2 d1 in EFTW from spring to winter
(Fig. 4). Comparing with the previous studies which reported the
rates ranging from 1.27 to 636 mg N m2 d1 in constructed wet-
lands (CWs) (Sims et al., 2013; Uggetti et al., 2012), the fluxes in
this study were little higher than the median N2O flux
(3.12 mg N m2 h1) from the CWs (Mander et al., 2014). The pos-
sible reasons are as follows: 1) the influent in this study was the
simulated secondary effluent which was rich of nitrate; 2) the
248 L. Gao et al. / Bioresource Technology 234 (2017) 243–252
Fig. 4. N2O fluxes in HEFTW and EFTW in different seasons (Spr, sum, aut and win
represent spring, summer, autumn and winter).
plant roots in EFTW were directly growing in water column rather
than in substrate as in CWs, thus N2O can be transported from root
zone immediately to atmosphere via vascular plants (Yavitt and
Knapp, 1998). In CWs, wetland soil or substrate may cause N2O
consumption exceeding N2O production and reduce the diffusion
rate of N2O (Jørgensen et al., 2012). As shown in Fig. 4, the N2O
Fluxes varied seasonally in HEFTW with a higher mean flux in win-
ter than in other seasons. The possible explanation was that the
low temperature in winter slowed down the overall rate of denitri-
fication, resulting in N2O accumulation in the end products (Søvik
et al., 2006).
In general, EFTW had the highest fluxes in four seasons without
significant change and HEFTW had the lowest fluxes in spring,
autumn and winter with one exception of the lowest flux of AEFTW
in summer. External carbon source and sulfur source have dramat-
ically enhanced the nitrogen removal efficiency without negative
impact on N2O emissions. Carbon (acetate) addition significantly
reduced the N2O fluxes, because high C/N promoted complete den-
itrification in transforming NO
3 to N2 and then declined the accu-
mulation of intermediate product. The Previous studies have
demonstrated that the successful treatment performance and low
N2O emission in CWs could be obtained at COD/N ratio of 5 (Wu
et al., 2009). Meanwhile, high S/N ratio plays an important role
in reducing N2O emission in sulfur-oxidizing autotrophic denitrifi-
cation system (Yang et al., 2016). The low N2O emissions of AEFTW
and HEFTW were accompanied by the high TN removal efficiency
and low NO
x -N concentration in summer (Fig. 3). The increased
NO
x -N concentration and the decreased TN removal efficiency
led to more N2O emission. As described above, the major nitrogen
removal process in HEFTW and AEFTW was denitrification, N2O
was most likely generated from the incomplete AD and HD. The
effluent nitrite was higher in AEFTW than HEFTW (Fig. 2D) in
spring, autumn and winter, explaining the higher N2O flux in
AEFTW. Although TN removal efficiency was similar in these two
processes, high nitrite concentration during denitrification led to
the accumulation of NO and N2O.
3.3. Microbial community analysis
As Table 1 shows, after removing the chimera sequences and
the ineffective sequences, Illumina MiSeq analysis of the samples
from 3 EFTWs yielded a total of 379,315 sequences. The lowest
ratio of high quality sequences was 90.47%, suggesting the results
were reliable and representative. 8535–122553 OTUs were clus-
tered with 97% similarity by performing the alignment and filter-
ing. Chao 1 and ACE richness estimators increased in HEFTW,
AEFTW and EFTW implying that all microbial community became
prosperous in the experiment. Comparing with E1, E2 richness
(Chao1: 2439, ACE: 3391) increased greatly and was greater than
the richness indexes of all the other samples. Plants in EFTW pro-
moted the bacterial growth on fillers (Wu et al., 2016). Chen et al.
(2015) found that planted CWs were of greater microbial richness
than the unplanted CWs. The external electron donors in HEFTW
and AEFTW promoted the growth of some specific microorganisms,
thus the quantity of minority microbial species relatively reduced,
which made the Chao1 and ACE richness index less than E2. The
Simpson and Shannon estimators decreased from 0.9931, 8.49 to
0.9814, 8.26 in HEFTW, declaring the reduction of microbial com-
munity diversity. This was most because that the organic carbon
sources supported the growth of heterotrophic denitrifiers, and
other bacteria were outcompeted by the heterotrophic bacteria.
On the other side, the higher Simpson and Shannon indexes in
EFTW showed that the microbial diversity in EFTW was the highest
among these three wetlands in the experiment. Simpson estima-
tors were almost equal in A1 and A2, implying that the thiosulfate
addition might produce less adverse impact on the microbial diver-
sity than the acetate addition.
In order to understand the phylogenetic diversity of bacterial
communities HEFTW, AEFTW and EFTW systems, qualified reads
were assigned to phyla, class (Fig. S1) and genus level (Fig. 5). Pro-
teobacteria, Actinobacteria, Chloroflexi and Bacteroidetes
accounted for approximately 80% in all samples. Proteobacteria
was the dominant phylum in all samples accounting for 46.4,
65.2, 47.5, 62.0, 34.5 and 44.3% in H1, H2, A1, A2, E1 and E2,
respectively (Fig. S1A). The Proteobacteria phylum contained many
species of bacteria related to global carbon, nitrogen and sulfur
cycling, which were widely found in natural and constructed wet-
lands (Ansola et al., 2014). The dominated classes were Betapro-
teobacteria, Alphaproteobacteria and Gammaproteobacteria in H1
and A1, while E1 was dominated by Anaerolineae, Alphaproteobac-
teria, Actinobacteria and Betaproteobacteria (Fig. S1B). After eight
months running, Betaproteobacteria increased to 36.4 and 34.1%
in H2 and A2. The proportion of Alphaproteobacteria increased
from 15.1% in E1 to 20.8% in E2 and become the largest class, illus-
trating that the external carbon and sulfur source altered the struc-
ture of microbial community in EFTWs.
Fig. 5 shows the relative abundance of the most dominant bac-
teria genera for the HEFTW, AEFTW and EFTW after a long period of
operation. The most abundant genera for HEFTW were Dechloro-
monas, C39, Desulfobacter, Thiothrix, Sulfuricurvum, Flavobacterium
and Thauera. Dechloromonas and Thauera are two denitrifying bac-
teria belonging to family Rhodocyclaceae which are capable of uti-
lizing acetate under anoxic conditions (Ginige et al., 2005).
Desulfobacter is a representative genus of sulfate-reducing bacteria
in wetlands which could oxidize acetate into CO2 via tricarboxylic
acid or acetyl-CoA (Chen et al., 2016). Although there were nearly
no documented knowledge of the genus C39, it was also abundant
in both H1 (2.10%) and H2 (2.89%), indicating that some species of
C39 may be associated with the acetate in denitrification process.
For AEFTW, the dominant genera were Thiobacillus, Sulfuricurvum,
Sulfurimonas, Nitrospira, Thiothrix, Flavobacterium, Dechloromonas.
Genera related to sulfur-oxidizing AD accounted for 19.8% of total
bacterial community. This finding further verified that AEFTW
were rich in autotrophic denitrifers and coincided to the excellent
TN removal efficiency in AEFTW. Thiobacillus and Sulfurimonas
were the most commonly reported autotrophic denitrifiers in
CWs (Shao et al., 2010) and up-flow biofilter (Zhou et al., 2017).
In EFTW, Nitrospira, Flavobacterium, Nitratireductor and devosia
were the most abundant genera. Nitrospira are well-known nitrite
oxidizing bacteria which transform nitrite into nitrate. Nitrospira
accelerated the consumption of nitrite and then indirectly
 L. Gao et al. / Bioresource Technology 234 (2017) 243–252 249

Table 1
High quality sequences, richness and diversity estimators of the sludge samples.

Sample Effective sequences High quality sequences Ratio (%) OTUs Chao1 ACE Simpson Shannon
H1 78242 72852 93.1 11477 1603 2105 0.9931 8.49
H2 59016 54381 92.1 11011 1967 2739 0.9814 8.26
A1 67920 62335 91.8 10668 1517 2091 0.9843 8.01
A2 62552 57667 92.2 10995 1854 2509 0.9854 8.25
E1 51291 46403 90.2 8535 1346 1911 0.9780 7.24
E2 60294 55579 92.2 12553 2439 3391 0.9946 9.33

Fig. 5. Taxonomic classification of bacterial reads retrieved from (A) H2, (B) A2, and (C) E2 at the genus level representative of matured bacterial community of HEFTW,
AEFTW and EFTW. Genera making up less than 0.3% of total composition are defined as ‘‘others”.

enhanced the transformation of ammonia into nitrite in EFTW.
Flavobacterium is a denitrification-related genus frequently
detected in denitrification systems (Wang et al., 2014). Nitratire-
ductor could reduce nitrate to nitrite. Devosia was related to
nitrogen-fixing root-nodule symbiosis with aquatic plants (Rivas
et al., 2002).
Interestingly, Dechloromonas, Flavobacterium, Thiothrix and Sul-
furicurvum were found abundant in HEFTW and AEFTW, indicating
that HD and AD might proceed simultaneously in these two
EFTWs. Desulfobacter was the second largest genus in HEFTW,
which could utilize acetate as electron donor and transform sulfate
into sulfide or other sulfur compound. Then, sulfur compound pro-
moted the growth of Thiothrix and Sulfuricurvum for denitrification.
The autotrophic denitrifiers coexisted with heterotrophic denitri-
fiers in EFTWs, which is beneficial to the nitrogen removal, as it
offers a possibility to use both acetate and thiosulfate in one EFTW
250 L. Gao et al. / Bioresource Technology 234 (2017) 243–252

to form the mixotrophic denitrification. Zhou et al. (2017) found

that Thiobacillus and Dechloromonas existed in both autotrophic
denitrification reactors and heterotrophic reactors. Sulfate-
reducing bacteria, like Desulfococcus and Desulfovibrio also coex-
isted with Thiobacillus in the heterotrophic system, which further
proved the occurrence of mixotrophic denitrification.
Table 2 exhibited the classification of functional genera related
to heterotrophic denitrification, facultative autotrophic denitrifica-
tion, autotrophic denitrification and Nitrification. Composition of
denitrifies dramatically increased from 10.6% to 19.7% and 11.3%
to and 24.2% in HEFTW and AEFTW, meanwhile the denitrifies in
EFTW remained almost unchanged from 6.62 to 6.57%. This could
explain why the HEFTW and AEFTW have analogously excellent
nitrogen removal performance compared with EFTW. In contrast
to the denitifcation-related genera, nitrification-related genera
were more abundant in EFTW (4.65%) at the final stage of experi-
ment. The proportion of nitrifiers decreased from 1.07 to 0.58%
as the consequence of depressed growth of nitrifiers by tremen-
dous amplification of heterotrophic denitrifies. On the contrary,
nitrification-related genera accounted for 2.92–2.55% of total bac-
teria, demonstrating that the autotrophic denitrifies coexisted with
the nitrifiers. This discovery revealed the reason why thiosulfate-
driven AD could achieve higher nitrification performance than Fig. 6. Redundancy analysis triplot to explore the correlation between environ-
mental factors and bacterial community based on genus level structure.
HD in autumn and winter.
RDA analysis, a kind of constrained ordination, was used to ana-
lyze the relationship between the relative abundance of predomi-
nant bacteria (genus level) and the environmental variables. As Desulfobacter were positively correlated with C/N, pH and NH+4-N
displayed in Fig. 6, the first and second axis represented 47.0% concentration. Thiobacillus, Sulfurimonas, Sulfuricurvum, Thiothrix,
and 40.6% of variation, respectively. It showed that the 6 samples Dechloromonas and C39 were negatively correlated with TN.
could cluster into 3 distinct groups including group I (A2), group Flavobacterium and nitrospira were negatively correlated with
II (H2) and group III (A1, H1, E1 and E2). This further proved that C/N, pH and NH+4-N concentration. In summary, the addition of car-
different electron donors significantly altered the microbial com- bon and sulfur all contributed to the decrease of DO. Then the low
positions in AEFTW and HEFTW. The microbial composition in DO favored the reproduction of autotrophic and heterotrophic den-
EFTW without external electron donor was similar during the itrifers, leading to higher nitrate and TN removal. Thiosulfate addi-
experiment. The angles of the arrows depicted in RDA indicated tion accelerated the nitrification and denitrification by stimulating
that Thiobacillus, Sulfurimonas, Sulfuricurvum, Thiothrix, Flavobac- the growth of nitrifiers (Nitrospira) and autotrophic denitrifers
terium and nitrospira were positively correlated with S/N and (Thiobacillus, Sulfurimonas, Sulfuricurvum and Thiothrix). Acetate
NO 2 -N concentration in the effluent. Dechloromonas, C39 and addition decreased the nitrification since the growth of nitrifiers

Table 2
Relative abundance of important functional genera.

EFTW HEFTW AEFTW EFTW

Sample H1 H2 A1 A2 E1 E2
Heterotrophic denitrification related genera (Zhong et al., 2015) Bacillus 0.11 0.02 0.08 0.03 0.05 0.1
Flavobacterium 1.79 1.10 2.07 1.81 0.63 2.14
Hyphomicrobium 0.27 0.52 0.29 0.33 0.44 0.7
Rhodobacter 3.64 0.65 1.27 0.19 3.96 0.61
Comamonas 0.01 0.01 0.02 0 0.01 0.01
Hydrogenophaga 0.23 0.41 0.18 0.20 0.15 0.48
Rubrivivax 0.02 0.03 0.02 0.03 0.01 0.27
Azospira 0 0 0 0.01 0 0.01
Dechloromonas 1.4 10.75 0.48 1.19 0.31 0.68
Thauera 0.12 0.97 2.11 0.08 0.09 0.11
Facultative autotrophic denitrification genera (Lee et al., 2013) Acinetobacter 1.52 0.22 0.08 0.1 0.09 0.05
Pseudomonas 0.44 0.09 0.06 0.06 0.08 0.16
Agrobacterium 0.13 0.17 0.13 0.09 0.09 0.36
Paracoccus 0.15 0.09 0.06 0.03 0.07 0.13
Autotrophic denitrification related genera (Shao et al., 2010) Thiobacillus 0.15 0.24 2.97 7.77 0.17 0.2
Sulfurimonas 0.31 0.24 0.33 3.00 0.21 0.24
Sulfuricurvum 0.08 1.38 0.17 6.93 0.06 0.10
Thiovirga 0.06 0.12 0.08 0.11 0.1 0.09
Thiothrix 0.04 1.86 0.89 2.13 0.05 0.08
Arcobacter 0.15 0.84 0.03 0.11 0.05 0.05
Sum of the denitrification related genera 10.6 19.7 11.3 24.2 6.62 6.57
Nitrification related genera Nitrosomonas 0.03 0.00 0.03 0.01 0.00 0.05
Nitrospira 1.04 0.58 2.89 2.54 0.51 4.60
Sum of the nitrification related genera 1.07 0.58 2.92 2.55 0.51 4.65
 L. Gao et al. / Bioresource Technology 234 (2017) 243–252
(Nitrospira) was depressed. Thiothirx was positively correlated with
S/N and C/N, suggesting that the Thiothirx might be an indicator of
mixotrophic denitrification.
3.4. Physicochemical responses
During the long period study, pH and dissolved oxygen were
monitored (Fig. S2). The pH in EFTW was 6.99 ± 0.17, significantly
lower than the influent (7.56 ± 0.15) (P < 0.05). Lower pH in
planted FTWs was observed in previous FTWs experiments
(White and Cousins, 2013). This change is possibly due to the
release of CO2 and acidic exudate from the plant root system respi-
ration (Iamchaturapatr et al., 2007), or by the degradation of
organic matter in the release of organic acids (Ijaz et al., 2015).
High nitrification effect existing in EFTW may also reduce pH as
described in Section 3.1. The pH in AEFTW (6.79 ± 0.18) was much
lower than EFTW, because the thiosulfate-driven AD generated
massive protons when reduing nitrate into N2 as described in
equation (1). The pH in HEFTW (7.64 ± 0.26) was little higher than
the influent. The alkaline pH could be explained that the acetate-
driven HD produced sufficient alkalinity.
DO in EFTW declined from 4.45 ± 1.68 (influent) to
1.48 ± 1.10 mg L1 (P < 0.05), and the lower oxygen concentration
in EFTW was affected by many factors. First, the pervasive root sys-
tem and spherical media provided huge surface for microbial pop-
ulations to colonize. And the respiration of living plant roots, dead
plant roots and microbial communities contributed to considerable
oxygen consumption (Tanner and Headley, 2011). Further,
although plants could transport the atmospheric oxygen into their
rhizosphere, the EFTW obstructed the air/water interface which
prevented the exchange of DO with atmospheric oxygen. DO in
HEFTW (0.38 ± 0.26 mg L1) and AEFTW (0.81 ± 0.69 mg L1) were
significantly lower than EFTW, which were typically anoxic condi-
tions. The external organic acetate and inorganic thiosulfate con-
sumed DO in water and facilitated the microbial denitrification
in AEFTW and HEFTW.
4. Conclusions
The carbon and sulfur sources addition effectively improved TN
removal efficiency of HEFTW and AEFTW. AEFTW was more suit-
able for the simultaneous removal of NH+4-N and NO 3 -N. The N2O
emissions reduced considerably in HEFTW and AEFTW, especially
in summer. The electron donors induced distinct shifts in microbial
community structures. Dechloromonas, Thiobacillus and Nitrospira
became the largest genera in HEFTW, AEFTW and EFTW, respec-
tively. Autotrophic and heterotrophic denitrification could simul-
taneously occur in HEFTW and AEFTW.
Acknowledgements
The study was supported by the National Natural Science Foun-
dation of China (51378306, 51678356, 51478262). The authors are
grateful to Shanghai Personal Biotechnology Co., Ltd. for the tech-
nological support on high-throughput sequencing and analysis.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.biortech.2017.03.
036.
251
References
Ansola, G., Arroyo, P., de Miera, L.E., 2014. Characterisation of the soil bacterial
community structure and composition of natural and constructed wetlands. Sci.
Total Environ. 473, 63–71.
APHA, 2005. Standard Methods for Water and Wastewater Examination. American
Public Health Association, Washington, D.C..
Bokulich, N.A., Subramanian, S., Faith, J.J., Gevers, D., 2013. Quality-filtering vastly
improves diversity estimates from Illumina amplicon sequencing. Nat. Methods
10 (1), 57–59.
Borne, K.E., Fassman-Beck, E.A., Tanner, C.C., 2014. Floating Treatment Wetland
influences on the fate of metals in road runoff retention ponds. Water Res. 48,
430–442.
Borne, K.E., Tanner, C.C., Fassman-Beck, E.A., 2013. Stormwater nitrogen removal
performance of a floating treatment wetland. Water Sci. Technol. 68 (7), 1657–
1664.
Chen, C., Wang, A., Ren, N., Deng, X., Lee, D.-J., 2009. Optimal process pattern for
simultaneous sulfur, nitrogen and carbon removal. Water Sci. Technol. 59 (4),
833–837.
Chen, Y., Wen, Y., Tang, Z., Huang, J., Zhou, Q., 2015. Effects of plant biomass on
bacterial community structure in constructed wetlands used for tertiary
wastewater treatment. Ecol. Eng. 84, 38–45.
Chen, Y., Wen, Y., Zhou, Q., Huang, J., Vymazal, J., Kuschk, P., 2016. Sulfate removal
and sulfur transformation in constructed wetlands: the roles of filling material
and plant biomass. Water Res. 102, 572–581.
Chung, J., Amin, K., Kim, S., Yoon, S., Kwon, K., Bae, W., 2014. Autotrophic
denitrification of nitrate and nitrite using thiosulfate as an electron donor.
Water Res. 58, 169–178.
García-Lledó, A., Vilar-Sanz, A., Trias, R., Hallin, S., Bañeras, L., 2011. Genetic
potential for N2O emissions from the sediment of a free water surface
constructed wetland. Water Res. 45 (17), 5621–5632.
GB3838-2002, 2002. Chinese National Environmental Quality Standards for Surface
Water. Ministry of Environmental Protection, China.
Ginige, M.P., Keller, J., Blackall, L.L., 2005. Investigation of an acetate-fed denitrifying
microbial community by stable isotope probing, full-cycle rRNA analysis, and
fluorescent in situ hybridization-microautoradiography. Appl. Environ.
Microbiol. 71 (12), 8683–8691.
Gruber, N., Galloway, J.N., 2008. An Earth-system perspective of the global nitrogen
cycle. Nature 451 (7176), 293–296.
Iamchaturapatr, J., Yi, S.W., Rhee, J.S., 2007. Nutrient removals by 21 aquatic plants
for vertical free surface-flow (VFS) constructed wetland. Ecol. Eng. 29 (3), 287–
293.
Ijaz, A., Shabir, G., Khan, Q.M., Afzal, M., 2015. Enhanced remediation of sewage
effluent by endophyte-assisted floating treatment wetlands. Ecol. Eng. 84, 58–
66.
IPCC, 2007. Climate Change 2007: The Physical Science Basis, Contribution of
Working Group I to the Fourth Assessment Report of the Intergovernmental
Panel on Climate Change. Cambridge University Press, Cambridge, United
Kingdom and New York, NY, USA.
Jørgensen, C., Struwe, S., Elberling, B., 2012. Temporal trends in N2O flux dynamics
in a Danish wetland – effects of plant-mediated gas transport of N2O and O2
following changes in water level and soil mineral-N availability. Glob. Change
Biol. 18 (1), 210–222.
Karine, E.B., Elizabeth, A.F., Chris, C.T., 2013. Floating treatment wetland retrofit to
improve stormwater pond performance for suspended solids, copper and zinc.
Ecol. Eng., 54
Lee, D.-J.J., Pan, X., Wang, A., Ho, K.-L.L., 2013. Facultative autotrophic denitrifiers in
denitrifying sulfide removal granules. Bioresour. Technol. 132, 356–360.
Li, C., Dong, Y., Lei, Y., Wu, D., Xu, P., 2015. Removal of low concentration nutrients
in hydroponic wetlands integrated with zeolite and calcium silicate hydrate
functional substrates. Ecol. Eng. 82, 442–450.
Liu, J., Wang, F., Liu, W., Tang, C., Wu, C., Wu, Y., 2016. Nutrient removal by up-
scaling a hybrid floating treatment bed (HFTB) using plant and periphyton:
from laboratory tank to polluted river. Bioresour. Technol. 207, 142–149.
Mander, Ü., Dotro, G., Ebie, Y., Towprayoon, S., Chiemchaisri, C., Nogueira, S.,
Jamsranjav, B., Kasak, K., Truu, J., Tournebize, J., Mitsch, W.J., 2014. Greenhouse
gas emission in constructed wetlands for wastewater treatment: a review. Ecol.
Eng. 66, 19–35.
Mander, Ü., Maddison, M., Soosaar, K., Koger, H., Teemusk, A., Truu, J., Well, R.,
Sebilo, M., 2015. The impact of a pulsing water table on wastewater purification
and greenhouse gas emission in a horizontal subsurface flow constructed
wetland. Ecol. Eng. 80, 69–78.
Mora, M., López, L.R., Gamisans, X., Gabriel, D., 2014. Coupling respirometry and
titrimetry for the characterization of the biological activity of a SO-NR
consortium. Chem. Eng. J. 251, 111–115.
Ravishankara, A.R., Daniel, J.S., Portmann, R.W., 2009. Nitrous oxide (N2O): the
dominant ozone-depleting substance emitted in the 21st century. Science 326
(5949), 123–125.
Reinhardt, M., Müller, B., Gächter, R., Wehrli, B., 2006. Nitrogen removal in a small
constructed wetland: an isotope mass balance approach. Environ. Sci. Technol.
40 (10), 3313–3319.
Reyes-Avila, J., Razo-Flores, E.a., Gomez, J., 2004. Simultaneous biological removal of
nitrogen, carbon and sulfur by denitrification. Water Res. 38 (14–15), 3313–
3321.
252 L. Gao et al. / Bioresource Technology 234 (2017) 243–252
Rivas, R., Velázquez, E., Willems, A., Vizcaíno, N., Subba-Rao, N.S., Mateos, P.F., Gillis,
M., Dazzo, F.B., Martínez-Molina, E., 2002. A new species of Devosia that forms a
unique nitrogen-fixing root-nodule symbiosis with the aquatic legume
Neptunia natans (Lf) Druce. Appl. Environ. Microbiol. 68 (11), 5217–5222.
Søvik, A.K., Augustin, J., Heikkinen, K., Huttunen, J.T., Necki, J.M., Karjalainen, S.M.,
Kløve, B., Liikanen, A., Mander, U., Puustinen, M., Teiter, S., Wachniew, P., 2006.
Emission of the greenhouse gases nitrous oxide and methane from constructed
wetlands in Europe. J. Environ. Qual. 35 (6), 2360–2373.
Søvik, A.K., Mørkved, P.T., 2008. Use of stable nitrogen isotope fractionation to
estimate denitrification in small constructed wetlands treating agricultural
runoff. Sci. Tot. Environ. 392 (1), 157–165.
Sahinkaya, E., Dursun, N., 2015. Use of elemental sulfur and thiosulfate as electron
sources for water denitrification. Bioprocess Biosyst. Eng. 38 (3), 531–541.
Shao, M.-F.F., Zhang, T., Fang, H.H., 2010. Sulfur-driven autotrophic denitrification:
diversity, biochemistry, and engineering applications. Appl. Microbiol.
Biotechnol. 88 (5), 1027–1042.
Sims, A., Gajaraj, S., Hu, Z., 2013. Nutrient removal and greenhouse gas emissions in
duckweed treatment ponds. Water Res. 47 (3), 1390–1398.
Tanner, C.C., Headley, T.R., 2011. Components of floating emergent macrophyte
treatment wetlands influencing removal of stormwater pollutants. Ecol. Eng. 37
(3), 474–486.
Uggetti, E., García, J., Lind, S.E., Martikainen, P.J., Ferrer, I., 2012. Quantification of
greenhouse gas emissions from sludge treatment wetlands. Water Res. 46 (6),
1755–1762.
Wang, R., Zheng, P., Xing, Y.-J., Zhang, M., Ghulam, A., Zhao, Z.-Q., Li, W., Wang, L.,
2014. Anaerobic ferrous oxidation by heterotrophic denitrifying enriched
culture. J. Ind. Microbiol. Biotechnol. 41 (5), 803–809.
White, S.A., Cousins, M.M., 2013. Floating treatment wetland aided remediation of
nitrogen and phosphorus from simulated stormwater runoff. Ecol. Eng. 61, 207–
215.
Wu, J., Zhang, J., Jia, W., Xie, H., Gu, R.R., Li, C., Gao, B., 2009. Impact of COD/N ratio
on nitrous oxide emission from microcosm wetlands and their performance in
removing nitrogen from wastewater. Bioresour. Technol. 100 (12), 2910–2917.
Wu, Q., Hu, Y., Li, S., Peng, S., Zhao, H., 2016. Microbial mechanisms of using
enhanced ecological floating beds for eutrophic water improvement. Bioresour.
Technol. 211, 451–456.
Xu, J.-H.H., He, S.-B.B., Wu, S.-Q.Q., Huang, J.-C.C., Zhou, W.-L.L., Chen, X.-C.C., 2016.
Effects of HRT and water temperature on nitrogen removal in autotrophic gravel
filter. Chemosphere 147, 203–209.
Yang, W., Lu, H., Khanal, S.K., Zhao, Q., Meng, L., Chen, G.-H., 2016. Granulation of
sulfur-oxidizing bacteria for autotrophic denitrification. Water Res. 104, 507–
519.
Yavitt, J.B., Knapp, A.K., 1998. Aspects of methane flow from sediment through
emergent cattail (Typha latifolia) plants. New Phytol. 139 (3), 495–503.
Zhang, L., Zhao, J., Cui, N., Dai, Y., Kong, L., Wu, J., Cheng, S., 2015. Enhancing the
water purification efficiency of a floating treatment wetland using a biofilm
carrier. Environ. Sci. Pollut. Res. 23 (8), 7437–7443.
Zhong, F., Wu, J., Dai, Y., Yang, L., Zhang, Z., Cheng, S., Zhang, Q., 2015. Bacterial
community analysis by PCR-DGGE and 454-pyrosequencing of horizontal
subsurface flow constructed wetlands with front aeration. Appl. Microbiol.
Biotechnol. 99 (3), 1499–1512.
Zhou, W., Li, Y., Liu, X., He, S., Huang, J.C., 2017. Comparison of microbial
communities in different sulfur-based autotrophic denitrification reactors.
Appl. Microbiol. Biotechnol. 101, 447–453.
Zhou, W., Liu, X., Dong, X., Wang, Z., Yuan, Y., 2016. Sulfur-based autotrophic
denitrification from the micro-polluted water. J. Environ. Sci. 44, 180–188.
Zhu, I., Getting, T., 2012. A review of nitrate reduction using inorganic materials.
Environ. Technol. Rev. 1 (1), 46–58.