See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/261918901

Effects of cattail biomass on sulfate removal

and carbon sources competition in subsurface-
flow constructed wetlands...

Article in Water Research April 2014

DOI: 10.1016/j.watres.2014.03.077 Source: PubMed

CITATIONS READS

16 100

7 authors, including:

Yi Chen Jan Vymazal

Chongqing University Czech University of Life Sciences Prague
21 PUBLICATIONS 320 CITATIONS 151 PUBLICATIONS 6,920 CITATIONS

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

nutrient removal in wetland buffer strips View project

Everglades Research View project

All content following this page was uploaded by Yi Chen on 01 July 2014.

The user has requested enhancement of the downloaded file.

 w a t e r r e s e a r c h 5 9 ( 2 0 1 4 ) 1 e1 0

Available online at www.sciencedirect.com

ScienceDirect

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

Effects of cattail biomass on sulfate removal and

carbon sources competition in subsurface-flow
constructed wetlands treating secondary effluent

Yi Chen a,b, Yue Wen a,*, Junwei Zhou a, Zhiru Tang a, Ling Li a, Qi Zhou a,
Jan Vymazal b
a
Key Laboratory of Yangtze Water Environment of Ministry of the State Education, College of Environmental Science
and Engineering, Tongji University, Shanghai 200092, PR China
b
Department of Landscape Ecology, Faculty of Environmental Sciences, Czech University of Life Sciences Prague
16521, Czech Republic

article info abstract

Article history: Sulfate is frequently found in the influent of subsurface-flow constructed wetlands (SSF
Received 18 December 2013 CWs) used as tertiary treatments. To reveal the effects of plants and litters on sulfate
Received in revised form removal, as well as the competition for organic carbon among microorganisms in SSF CWs,
11 March 2014 five laboratory-scale SSF CW microcosms were set up and were operated as a batch system
Accepted 28 March 2014 with HRT 5 d. The results showed that the presence of Typha latifolia had little effect on
Available online 12 April 2014 sulfate removal in CWs, with or without additional carbon sources. Cattail litter addition
greatly improved sulfate removal in SSF CWs. This improvement was linked to the
Keywords: continuous input of labile organic carbon, which lowers the redox level and supplies a
Subsurface flow constructed wet- habitat for sulfate reducing bacteria (SRB). The presence of SRB in cattail litter indicated the
lands (SSF CWs) possibility of sulfate removal around the carbon supplier, but the quantity of microbes in
Sulfate cattail litter was much lower than that in gravel. Stoichiometry calculations showed that
Cattail litter the contribution of SRB to COD removal (21e26%) was less than that of methane-producing
Sulfate reducing bacteria (SRB) bacteria (MPB) (47e61%) during the initial stage but dominated COD removal (42e65%)
Carbon competition during the terminal stage. The contributions of aerobic bacteria (AB) and denitrification
bacteria (DB) to COD removal were always lower than that of SRB. It was also observed that
the variations in COD: S ratio had a great influence on the relative abundance of genes
between SRB and MPB and both of them could be used as good predictors of carbon
competition between SRB and MPB in CWs.
2014 Elsevier Ltd. All rights reserved.

and low implementation costs. CWs are increasingly used as a

1. Introduction tertiary treatment to polish the effluent from wastewater
treatment plants (WWTP) because they can successfully treat
Constructed wetlands (CWs) are widely used to treat munic- wastewaters with very low organic concentrations
ipal and domestic wastewaters, due to their simple operation (Greenway, 2005; Vymazal, 2009).

* Corresponding author. Present address: Room 301, Mingjing Building, School of Environmental Science and Engineering, Tongji
University, Shanghai 200092, PR China. Tel./fax: 86 21 65982697.
E-mail addresses: weny@tongji.edu.cn, wenyue@vip.163.com (Y. Wen).

http://dx.doi.org/10.1016/j.watres.2014.03.077
0043-1354/ 2014 Elsevier Ltd. All rights reserved.
2 w a t e r r e s e a r c h 5 9 ( 2 0 1 4 ) 1 e1 0
Sulfate is a common contaminant of wastewater effluent
and its removal from CWs has become increasingly important
in recent years because it has various effects on the biogeo-
chemistry of wetlands (Gonzalias et al., 2007; Stein et al., 2007;
Wiessner et al., 2008, 2010; Wu et al., 2011). Bacterial sulfate
reduction (BSR) to sulfide is widely considered to be the
dominant process involved in the sulfate removal from CWs
(Vymazal and Kropfelova, 2008). Many environmental factors
are known to influence BSR in CWs, including: redox potential,
dissolved oxygen, temperature, pH value and type of organic
carbon (Ren et al., 2009). Redox potential and organic carbon
are especially critical for BSR in CWs (Stein et al., 2007). Low
redox levels and sufficient carbon supply could facilitate sul-
fate removal in CWs because the sulfate reducing bacteria
(SRB) responsible for BSR are mainly heterotrophic anaerobes.
In the CWs used for tertiary treatment, organic carbon is
mainly derived from the secondary effluent and plant root
exudates (Wei et al., 2009). Given that the effluent-derived
organic matter is refractory or non-biodegradable (Pinney
et al., 2000; Quanrud et al., 2004), the carbon source from
plant root exudation is likely to be insufficient for significant
sulfate reductions in CWs. Therefore, additional carbon
source is needed to drive the BSR in CWs treating secondary
effluent. Glucose, sucrose and acetate have been used as
liquid carbon sources to drive BSR in CWs (Gonzalias et al.,
2007; Stein et al., 2007; Wiessner et al., 2010). However, they
are relatively expensive, and plant materials may be a po-
tential alternative carbon source because they are low cost,
renewable and widely available. Wetland plant litter has been
reported to accumulate at a rate of 500e2000 g C m2 yr1 in a
mature wetland (Hume et al., 2002), and the cellulose
component could be degraded into available organic carbon
by fungi and bacteria in CWs (Vymazal and Kropfelova, 2008).
Numerous studies have demonstrated the feasibility of using
plant litter to stimulate the nitrate reductive process (Chen
et al., 2011; Hume et al., 2002; Wen et al., 2010), yet very few
studies have reported its effect on sulfate reduction. Apart
from being a carbon supplier, the porous plant litter may also
act as a biofilm carrier for microorganism aggregation.
Fleming-Singer and Horne (2002) demonstrated that the bio-
film was attached to the litter using microscopic images.
However, little is known about the quantitative microor-
ganism distribution in plant litter and gravel in CWs.
Microorganisms compete for available organic carbon
using different metabolic pathways, and the energy release is
directly related to the success of the competition among mi-
croorganisms for different electron acceptors (Rittmann and
McCarty, 2001). Baptista et al. (2003) found that the competi-
tion for organic carbon between SRB and methane-producing
bacteria (MPB) was ubiquitous, and that SRB consumed 25% of
the carbon source in CWs. Stein et al. (2007) further found that
the COD: S ratio is a critical factor for the carbon competition
between SRB and MPB. At a COD: S ratio greater than 2.7, MPB
was the dominant carbon consumer, and, at a COD: S ratio
lower than 1.7, SRB was the dominant carbon consumer.
However, it is not known whether the carbon competition is
related to the abundance or the structure of SRB and MPB
communities present in the CWs. In addition to sulfate, oxy-
gen and nitrate are two other ubiquitous electron acceptors in
secondary effluent. It is well known that aerobic bacteria (AB)
and denitrification bacteria (DB) use oxygen and nitrate,
respectively, as electron acceptors for respiration and are
thermodynamically better than SRB and MPB when it comes
to competing for electron donors. Furthermore, previous
studies have demonstrated that the presence of oxygen and
reduced N-oxides severely inhibit SRB and MPB (Haveman
et al., 2004; Tugtas and Pavlostathis, 2007). Therefore, it is
important to understand the carbon flow in CWs when mul-
tiple electron acceptors are present. However, there have been
few studies concerning the carbon sources competition bal-
ance when CWs are used to treat secondary effluent.
Plants are considered to be indispensable in CWs (Kadlec
and Wallace, 2008). They can either enhance sulfate reduc-
tion by supplying organic carbon with root exudates or inhibit
this process by transferring oxygen to the rhizosphere (Stein
et al., 2007; Wiessner et al., 2010). Numerous studies have re-
ported that plants can inhibit sulfate removal in CWs. How-
ever, the results often vary depending on the plant species
present (Allen et al., 2002; Stein et al., 2007; Wiessner et al.,
2010). Typha spp. is thought to be suitable for sulfate
removal in the low redox conditions (Allen et al., 2002;
Gebremariam and Beutel, 2008). However, it is not known
whether Typha spp. continues to be effective when the CWs
are loaded with litters.
The objectives of this study were: (1) to study the effect of
cattail and cattail litter on sulfate removal; (2) analyze the
organic carbon competition among AB, DB, SRB and MPB in
CWs; (3) quantify the 16S rRNA gene in competing microor-
ganisms for cattail litter and gravel in CWs; and (4) establish
the relationship between the 16S rRNA gene and carbon
competition.
2. Materials and methods
2.1. Source of plant litter and treatment
In this study, cattail (Typha latifolia) litter was used as the
carbon source to drive sulfate reduction and denitrification.
The cattail litters were collected from the same wetland mi-
crocosms in a greenhouse (seen 2.2) and taken into the labo-
ratory in November, 2010. After collection, the cattail litters
were cleaned, washed using neutral detergent, cut into
1e1.5 cm lengths, and dried at 40  C to a constant mass before
being finally preserved in a moisture free container at room
temperature (20  C).
2.2. Design and operation of the SSF CW
Five sequencing batch SSF CW microcosms, each with a bulk
volume of 0.045 m3 (length: 0.3 m, width: 0.3 m, height: 0.5 m)
and a pore volume of 12 L, were set up in this study. Five sets
of systems: unplanted control (W0), litter added microcosms
(W1: 100 g; W2: 200 g), planted microcosms (W3: 22 plants m2)
and planted plus litter added microcosms (W4: 100 g litter,
22 plants m2), were established. All the microcosms were
filled with washed gravel (4 8e13 mm, porosity 0.4) and two
of them (W3 and W4) were planted with cattail (T. latifolia). The
wetland microcosms were located in an air-conditioned
greenhouse at a temperature of 25  1  C since 2005. Details
 w a t e r r e s e a r c h 5 9 ( 2 0 1 4 ) 1 e1 0
of the wetland microcosm design have been given in our
previous studies (Lin et al., 2008; Wen et al., 2010).
Prior to the start of the experiment, the five microcosms
were fed, in batches, with a modified secondary effluent for 6
months in order to establish the plant shoots and microor-
ganisms. The secondary effluent was collected from a neigh-
boring wastewater treatment plant (WWTP) and spiked with
glucose (C6H12O6) and starch (C6H11O5)n, in order to achieve a
final concentration of 50  5 mg COD L1 (The COD ratio of
glucose to starch was 1:1) in W0eW4 microcosms. Cattail litter
was added to the W1, W2 and W4 microcosms after a 6 month
acclimatization period. The cattail litter was homogeneously
mixed with gravel and the details of the litter loading have
been reported in a previous study (Chen et al., 2011). The
experiment began on 5 September, 2011. The wetland micro-
cosms were fed with the secondary effluent from a neigh-
boring WWTP, and the characteristics of the wetland influent
were seen in Table S1. The five microcosms operated as a
batch system with pulse loading, which were filled with
wastewater at the start of each batch and were gravity drained
within 1 h prior to the next batch. All the treatments (W0eW4)
were triplicated, and there were 18 batches (each batch lasted
for 5 d). Solution samples were taken from each microcosm
and each batch at 0, 0.5, 1, 1.5, 2, 4, 8, 12, 22, 28, 36, 46, 52, 60, 70,
76, 84, 96, 100, 108 and 120 h. A 100 mL syringe was used to
collect water samples at a depth of 5, 20 and 30 cm from the
center sampling pipe. Only solution samples taken from 20 cm
depth were analyzed because no vertical gradients in the
water chemistry were observed in the preliminary experiment
and previous experiments with the same microcosm (Wen
et al., 2010). Methane sampling was undertaken in a closed
chamber and the methods for its collection and measurement
were described in the Supporting Information.
2.3. Chemical analysis
Five 50 mL water samples, withdrawn at the appropriate time
intervals, were membrane-filtered (0.22 mm) and immediately
analyzed for dissolved sulfide. Zinc chloride solution was then
added (prior to further analysis for other chemical constitu-
ents) to the filtered samples in order to eliminate any soluble
sulfide by precipitating it as zinc sulfide.
Sulfate and nitrate were detected using a DX ICS-3000 ion
chromatography unit (Dionex Corporation, CA, USA) equipped
with a conductivity detector and a self-regenerating sup-
pressor ASRS-ULTRA II 4-mm (129 mA). The separation was
achieved with a DX IonPac AS11-HC analytical column
(4  250 mm) and an AG11-HC guard column (4  50 mm) both
operating at 35  C. A potassium hydroxide gradient was
applied to the Dionex automatic eluent generator using an
EluGen cartridge (EGC II KOH). The gradient started at 12 mM
KOH, and rose up to 34 mM in 5 min, where it was kept for
3 min. Then it was increased from 34 mM to 52 mM in 1 min
and maintained at that concentration for another 5.5 min.
Finally, it was decreased from 52 mM to 12 mM in 1 min and
kept at that concentration for another 1 min. The total anal-
ysis time was 16.5 min. The injection volume was 1 mL and
the flow rate was 1 mL min1. Elemental sulfur was detected
according to Rethmeier et al. (1997) by extracting the mono-
bromobimane treated samples with chloroform, and then
3
analyzed using HPLC (Agilent 1200, Agilent Technologies, CA,
USA) with a Li-Chrospher 100, RP 18 column equipped with an
UV detector at 263 nm.
Dissolved sulfide and COD were analyzed following stan-
dard methods (APHA, 1998). The analyses of cellulose, hemi-
celluloses, lignin contents and volatile fatty acids (VFAs) were
the same as described in a previous publication (Wen et al.,
2010). The pH and Eh were measured using a portable pH/
mV/temperature meter (HACH, sensION1, USA) fitted with a
gel-filled pH electrode and a combination Ag/AgCl Eh electrode.
2.4. DNA extraction
After batches 2 (10 d), 10 (50 d) and 18 (90 d), approximately 200 g
of gravel and litter were collected from the top (5 cm), middle
(20 cm) and bottom (40 cm) sections of the wetland micro-
cosms. The three samples were combined for DNA extraction.
Before DNA extraction, the gravels/litters were vigorously
shaken at 225 rpm for 3 h in sterile glass bottles in order to
suspend any attached biofilm in the liquid solution. The pre-
cipitate was collected in bottles, after twice centrifuging at
5000 g for 20 min, for further analysis. Total genomic DNA was
extracted from the gravels and litters using E.Z.N.A.  Soil DNA
Kit (OMEGA bio-tek). The quantity and quality of the extracted
DNA were checked by measuring its absorbance at 260 and
280 nm using a UNICO-2100 UV/VIS spectrophotometer.
2.5. Real-time PCR assays
Real-Time PCR assays were carried out in order to quantify the
total bacterial 16S rRNA and key functional genes for meth-
anogenesis (mcrA), sulfate reduction (dsrA) and nitrate
reduction (nirK and nirS) using primers: 338f/518r, Met1f/
Met1r, dsr-1F 50 /dsr-500r 50 , nirK876/nirK1040 and nirSCd3aF/
nirSR3cd, respectively (SI Table S2). QPCR was performed
using a two-step amplification procedure and the thermal
cycling conditions consisted of 2 min at 95  C followed by 40
cycles of 10 s at 95  C, and 40 s at 60  C. All samples were run in
triplicate. A duplicate 10-fold dilution series of standard DNA
was used to calculate a standard curve. The standard curves
for the bacterial and functional genes had R2 values of
0.991e0.999 and the amplification efficiencies were 84e110%.
2.6. High-throughput 16S rRNA gene sequencing and
analysis
High-throughput 454 GS-FLX pyrosequencing of the 16S rRNA
gene was conducted according to standard protocols
(Margulies et al., 2005). Raw sequencing data were deposited
to the NCBI Sequence Read Archive (SRA) with accession no.
SRP034645. BLAST of taxonomic classification down to the
phylum, class and genus level was then achieved using
MOTHUR via SILVA 106 database with a set confidence
threshold of 80%. The details of pyrosequencing and analysis
were supplied in the Supporting Information.
2.7. Calculations and statistical analysis
The competition for carbon sources by different microbes in
wetland microcosms is expressed as the contribution of
4 w a t e r r e s e a r c h 5 9 ( 2 0 1 4 ) 1 e1 0

individual COD consumption to the total COD consumption. In the above set of equations, (1) DcCOD represents the total
Total consumption of COD (DcCOD) was calculated from the COD consumption (mg). (2) DnCH4 represents the methane
sum of the release of COD from plant litter and the difference production (mol). (3) DcSO2 represents the amount of sulfate
4
between influent COD and effluent COD in the liquid phase of removed (mg). (4) DcNO3 represents the amount of nitrate
the wetland, see Eq (1): removed (mg). (5) DcO2 represents the oxygen consumption
(mg). (6) kLa represents the oxygen mass transfer coefficient.
DcCOD initial COD in the plant litter The kLa value used in this study was 0.1 d1 (Wen, 2007). Sosat
 COD remained in the plant litter influent COD represents the saturated dissolved oxygen in water, and the
 residual COD in the liquid: value used in this study was 8.25 g O2 m3. So represents the
(1) initial dissolved oxygen in the influent water. The COD con-
centrations and the number of bacterial 16S rRNA genes were
The contributions made by MPB, SRB and DB to DcCOD were very similar throughout the experiment, so the COD con-
calculated from the electron flow distribution data (Isa et al.,
sumption during bacterial biomass production was not taken
1986) using the following equations:
into consideration in the COD mass balance calculations.
DnCH4  64 In this study, differences in sulfate removal rates were
CODMPB %  100% (2) tested for the effects of non-planted vs. planted CWs by one-
DcCOD
way ANOVA and Tukeys tests. Differences in means of Eh
DcSO2  0:67 values among litter added microcosms were analyzed using a
CODSRB % 4
 100% (3) Friedman ANOVA and Wilcoxon signed rank test. Statistical
DcCOD
analyses were performed using SPSS version 19.0 software,
DcNO3  0:77 and were considered significant at the 0.05 level. The relative
CODDB %  100% (4) importance of factors (relative abundance of genes, COD: S
DcCOD
ratio and thermodynamic conditions) for carbon sources
The AB contribution to DcCOD was obtained as follow:
competition in CWs was identified by Path Analysis using
DcO2 AMOS version 21 (SPSS, IBM).
CODAB %  100% (5)
DcCOD

DcO2 kLa  Sosat  So (6)

The contribution made by other microorganisms to DcCOD
was calculated using the following equation:
CODOthers % 1  CODMPB  CODSRB  CODDB  CODAB
 100%
3. Results and discussion
In this study, sulfate removal rates decreased in litter added
microcosms over time, due to the declined carbon supply
during litter decomposition. Taking W2 as an example, the
(7)
sulfate was completely removed over the first 30 d, but there

Fig. 1 e Sulfate and nitrate concentrations in the different wetland microcosms during the initial stage (a, batch 2), middle
stage (b, batch 10) and terminal stage (c, batch 18). tlag: The time-lag during sulfate removal.
 w a t e r r e s e a r c h 5 9 ( 2 0 1 4 ) 1 e1 0 5

Table 1 e Sulfate removal efficiency of microcosm wetlands throughout the period of study.
CWs Initial stageb Middle stagec Terminal staged
a a
Sulfate removal rate t (h) Eff. Sulfate removal rate t (h) Eff. Sulfate removal ratea t (h) Eff.
3 1 3 1 3 1
(g S m d ) (%) (g S m d ) (%) (g S m d ) (%)
W0 0.66 e 25.0 0.00 e 0.0 0.00 e 0.00
W1 2.43 120.00 94.4 0.56 e 26.8 0.00 e 0.00
W2 5.16 56.00 100.0 1.79 e 86.2 0.00 e 0.00
W3 0.00 e 0.0 0.00 e 0.0 0.00 e 0.00
W4 2.35 120.00 95.4 0.88 e 40.5 0.00 e 0.00

t, time required to obtain complete sulfate removal (h). d: incomplete sulfate removal at a given time of 5 d.
W0: unplanted control; W1: 100 g cattail litter; W2: 200 g cattail litter; W3: planted with Typha latifolia, 22 plants/m2; W4: 100 g cattail litter
22 plants/m2.
1
a
The sulfate removal rate was calculated as follows: SO2
4 removal V HRT1$(ci(in)$Vi(in)  ci(out)$Vi(out)), where SO2 2
4 removal eSO4  S removal
3 1 2
rates (g m d ); idthe number of batch in sequence (i 1,2,3 . 18); Ci(in), Ci(out)dThe SO4  S concentrations in influent and effluent of batch i
(mg L1); Vin, VoutdThe volume of pore water in influent and effluent (L); VdThe volume of gravel (m3), V 0.036 m3; HRTdThe retention time
for every batch (d), HRT 5 d.
b
Batch 2.
c
Batch 10.
d
Batch 18.

was incomplete removal of sulfate over the next 30e70 d. secondary effluent, nitrate has been widely recognized to
Finally, the sulfate concentrations increased slightly in the interfere with sulfate removal (Jiang et al., 2009). Inhibition of
outlet after 70 d (Fig. 1). Thus, the experiment was divided into SRB activity by nitrite, competition for electron donors and the
three stages, according to the sulfate removal characteristics: Eh gap have been shown to be the main mechanisms behind
an initial stage (days 1e30, batches 1e6), a middle stage (days the inhibition of sulfate reduction in the presence of nitrate
31e70, batches 7e14) and a terminal stage (days 71e90, (Ren et al., 2009). Nitrite accumulation (<0.1 mg L1) was
batches 15e18). Given that the sulfate removal kinetics fol- negligible in this study, and there was a preference for nitrate
lowed a similar pattern at each stage, we chose batch 2 (B2), 10 removal as the Eh rapidly decreased. These factors might
(B10) and 18 (B18) as typical batches that represented the three explain the time-lag in sulfate removal. Furthermore, the
stages, respectively. sulfate concentrations in the litter added CWs showed an in-
crease in the early stages of the middle and terminal periods
3.1. Sulfate removal efficiency and this was probably due to the periodic re-oxidation of
sulfide and sulfur to sulfate by dissolved oxygen in CWs. Apart
As shown in Table 1, the sulfate removal rates were very from oxygen, nitrate in the influent can also easily drive sul-
different among the five wetland microcosms fide and elemental sulfur oxidation to sulfate in wetlands
(W2 > W1 z W4 > W3 z W0), and insignificant sulfate re- (Krishnakumar and Manilal, 1999; Londry and Suflita, 1999).
movals were observed in the wetland microcosms without Wu et al. (2011) demonstrated the multiple sulfur trans-
cattail litter. This suggested that the cattail litter could act as formations, including sulfide re-oxidation in CWs, using the
carbon sources for sulfate reduction and that influent organic stable isotope approach. In this study, the detection of
matter or root exudates had a marginal impact on sulfate elemental sulfur (seen in Fig. S1) as the intermediate product
reduction. Furthermore, the sulfate removal rates decreased during the sulfide oxidation favored the hypothesis
over time in all the litter added microcosms, and complete mentioned above.
sulfate removal was only achieved during the initial stage. As
the experiment continued, incomplete sulfate removal
3.2. Effect of cattail on sulfate removal
occurred in the middle and terminal stages because of the
reduction in the organic carbon content of the cattail litter. As
Wetland plants can either enhance sulfate reduction by sup-
shown in Table 2, the content of cellulose and hemicelluloses
plying organic carbon as exudates or inhibit this process by
in cattail litter fell by 63e70% and 52e66%, while the lignin
transferring oxygen to the rhizosphere (Stein et al., 2007;
content increased by 68e96% over the period of study. This
suggested that cellulose and hemicelluloses were probably the
main carbon sources for sulfate reduction in the CWs added
Table 2 e Changes in the main composition of cattail
with litters, which was consistent with the results obtained litters at the end of the experiment (%DM).a
previously in a similar system (Wen et al., 2010).
Cattail litters Cellulose Hemicelluloses Lignin
Fig. 1 clearly shows that there was a time-lag during sulfate
Raw 31.50  4.02 12.60  2.87 11.20  3.03
removal and sulfate removal occurred only after the complete
End in W1 11.63  1.12 5.28  0.78 19.61  2.94
removal of the nitrate in wetland microcosms added with
End in W2 9.57  1.54 4.25  0.17 21.91  1.93
cattail litter. Whitmire and Hamilton (2005) also found that End in W4 11.21  2.17 6.03  0.61 18.81  2.14
sulfate was removed from injected ground water in wetlands, a
Values are the means  SD (n  3).
but only after the nitrate was depleted. As a component of the
6 w a t e r r e s e a r c h 5 9 ( 2 0 1 4 ) 1 e1 0
Wiessner et al., 2010). As shown in Fig. 1, sulfate reduction was
marginal in the W3 microcosm, due to the lack of sufficient
available carbon sources. When the cattail litter was added as
the carbon sources, the effect of the plants on sulfate reduc-
tion could be obtained from the difference in sulfate removal
between the W1 and W4 microcosms. As shown in Fig. 1, there
were no significant differences in the sulfate removal rates
between the two microcosms (p > 0.05). This indicated that
the net effect of cattail on sulfate reduction was zero in the
wetland microcosms. The results of this study were incon-
sistent with the study conducted by Stein et al. (2007), who
concluded that sulfate reduction significantly decreased in the
presence of plants. This contradiction could be due to the
following reasons: (1) the average temperature in this study
(25  C) was higher than in the previous study, resulting in an
increased carbon supply from the roots and a lower oxygen
transfer in the wetlands (Callaway and King, 1996). (2) the
continuous release of carbon sources from cattail litter
counterbalanced the oxygen transfer capacity of the plants,
which resulted in a relatively constant SRB activity
(Gebremariam and Beutel, 2008). (3) the plant conditions (i.e.
age and health) and species (Schoenoplectus acutus vs. T. lat-
ifolia) were different between studies. (4) (4) there were not
enough replicates to show significant differences between the
two conditions. As mentioned above (Hypothesis 1e4),
wetland plants might have less negative impact on sulfate
reduction than previously thought if the production of
detritus from plant litter is taken into consideration.
3.3. Effect of cattail litter on sulfate removal
Table 1 shows that sulfate removal rates in the W1 microcosm
were 0e2430 mg S m3 d1, much higher than those in W0,
which indicated that cattail litter could greatly improve sul-
fate removal in SSF CWs. Furthermore, doubling the cattail
litter in the wetlands approximately doubled the sulfate
removal rates during the initial stage and trebled the sulfate
removal rates during the middle stage. As part of the ligno-
cellulosic biomass, cattail litters can be easily degraded to
soluble carbohydrate and further fermented into excellent
electron donors (i.e. acetate and propionate) for sulfate
reduction (Zhao et al., 2009). In this study, both acetate and
propionate were detected as components of VFAs during the
initial stage and acetate was the most dominant VFAs (>60%),
which was consistent with previous studies (Chen et al., 2011;
Wen et al., 2010). During the initial stage, the acetate
(25.3 mg L1) and propionate (15.2 mg L1) reached maximum
concentrations in the W2 microcosm (seen in Fig. S5 in SI),
which was consistent with the highest sulfate removal rates
observed in the W2 microcosm (Table 1). Then, the VFAs
concentrations in each microcosm decreased gradually with
time, and it was not detected in the middle and terminal
stages.
In addition to being an electron-donor supplier, cattail
litter also acted as a redox mediator in the wetlands. Fig. 2
showed that cattail litter could rapidly consume the oxygen
and nitrate and lower the Eh values from 50 mV (iron reduc-
tion) to 100 w 200 mV (sulfate reduction) in the wetland
microcosms. The average Eh values increased gradually be-
tween batch 2 and 18, and Eh values below 100 mV occurred
at 28 h, 48 h and 70 h, respectively, for batches 2, 10 and 18.
The longer times (from 28 h to 70 h) needed to reach a suitable
Eh (<100 mV) for sulfate reduction was consistent with the
increase in the time lag for sulfate reduction as the experi-
ment progressed (Fig. 1). This confirmed that the redox state
plays an important role in sulfate removal in SSF CWs
(Faulwetter et al., 2009).
3.4. Carbon sources consumptions between
microorganisms
SSF CWs are generally considered as a system that lies be-
tween the anoxic and anaerobic conditions, and the microbes
within these similar ecospheres are responsible for the car-
bon, nitrogen and sulfur cycling in the SSF CWs (IWA, 2000).
Organic carbon is a critical factor influencing sulfate reduction
in wetlands via redox condition regulation and carbon
competition between SRB and other microbes (Ren et al., 2009;
Stein et al., 2007). Given that the available electron acceptors
in this study were oxygen, nitrate, sulfate, carbon dioxide and
acetate, only AB, DB, SRB and MPB were analyzed for microbial
consumption of organic carbon in CWs.
Theoretical COD consumptions for AB, DB, SRB and MPB in
W1 and W2 wetland microcosms are shown in Fig. 3a. COD
consumptions for total microbes decreased from 281.3 mg L1
to 74.7 mg L1, and from 352.6 mg L1 to 100.9 mg L1 in the W1
and W2 microcosms, respectively. The maximum COD con-
sumption by total microbes occurred during the initial stage
and COD consumptions by MPB followed a similar pattern to
that consumed by total microbes, i.e. higher in the initial
stage. In contrast, COD consumptions by SRB varied depend-
ing on the microcosm (gradually decreased in W1, but did not
significantly change in W2). The decreased COD consumption
by SRB in W1 unit was probably due to the exhaustion of
organic carbon from the litters in the terminal stage. Finally,
COD consumptions by DB and AB did not vary significantly
with system and time. The results clearly show that the extent
of the decreases in COD consumptions with time was contrary
to the thermodynamic sequence of electron acceptors for the
oxidation of organic carbon. Namely, organic carbon tends to
Fig. 2 e Changes in the Eh and COD concentrations in the
litter added SSF CW microcosms during batches 2, 10
and 18.
 w a t e r r e s e a r c h 5 9 ( 2 0 1 4 ) 1 e1 0 7

be oxidized preferentially with oxygen, and then nitrate. Thus,

the COD utilization by AB and NB seldom suffer from a lack of
available carbon sources. Once the nitrate becomes depleted,
obligate anaerobes begin to use sulfate, then carbon dioxide
and acetate as electron acceptors, until finally they produce
methane. Thus, the insufficient release of organic carbon and
preferential utilization by AB, DB and SRB, resulted in a dra-
matic decrease in COD consumptions by MPB during the
initial and terminal stages (Fig. 3a).
The relative contribution of AB, DB, SRB, MPB and other
microbes to the COD consumption is shown in Fig. 3b. During
the initial stage, the relative contributions of AB, DB, SRB, MPB
and other microbes to the COD consumption were 1.0e1.1%,
8.4e9.2%, 20.6e26.0%, 47.3e61.0% and 9.1e16.6%, respectively,
which suggested that MPB were responsible for most of the
COD consumption, whereas aerobic activity was limited, due
to the lack of DO in the influent and very low oxygen transport
rates from air to water. During the terminal stage, the relative
contributions of AB, DB, SRB, MPB and other microbes to the
COD consumption were 3.0e3.2%, 27.9e39.2%, 41.6e65.3%,
3.3e13.1% and 0.5e2.9%, respectively, which showed that SRB
and DB were responsible for most of the COD consumption.

Fig. 4 e The quantity of archaea and bacteria in the W1 and

W2 microcosms revealed by real-time PCR. (a) The
distribution of 16S rRNA genes for MPB between gravel (G)
and litter (L). (b) The distribution of 16S rRNA genes for total
bacteria, sulfate reduction genes (dsrA) and denitrification
genes (nir) between gravel (G) and litter (L).

The shift from MPB-dominated COD consumption to SRB-

Fig. 3 e Relative contributions of sulfate reduction bacteria
(SRB), denitrification bacteria (DB), aerobic bacteria (AB) and
methanogenic production bacteria (MPB) to the total COD
consumptions in the W1 and W2 microcosms during
batches 2 and 18. (a) COD consumptions of SRB, DB, AB and
MPB, (b) Individual COD consumption expressed as a
percentage of the total COD consumption.
dominated COD consumption was likely due to the decrease
in the COD: S ratio. A low COD: S ratio appeared to facilitate
COD consumption by SRB, which have a high affinity for
various electron donors (i.e. acetate and hydrogen), as shown
by a relatively low half saturation constant (Km 0.001 mM
hydrogen; Km 0.02 mM acetic acid) (Ren et al., 2009). Choi and
Rim (1991) found that, at a COD: SO2 4 ratio greater than 2.7
MPB are dominant, and at a COD: SO2 4 ratio lower than 1.7,
SRB are dominant. During the initial and terminal stages, the
calculated COD: SO2
4 ratio was 2.5 and 0.9 in W1, and 3.9 and
1.2 in W2, respectively, which showed that MPB might be
dominant during the initial stage and SRB might be dominant
during the terminal stage, which was consistent with microbe
dependent COD consumption. Stein et al. (2007) found that
about 40% of the COD removal was attributable to SRB at 24  C
in unplanted wetland, which was in the range (20.6e65.3%)
observed in this study under similar conditions. Garcia et al.
(2004) also found that sulfate reduction was most respon-
sible for the removal of organic matter (36e100% of the total
removal) in SSF CWs with a depth of 0.5 m.
8 w a t e r r e s e a r c h 5 9 ( 2 0 1 4 ) 1 e1 0
3.5. Quantification of different microbial groups
attached to gravel and cattail litter
Real-time PCR was used to compare the numbers of copies of
16S rRNA genes attached to gravel and cattail litter in the W1
and W2 microcosms. Three functional groups that have a
critical role in carbon competition: d DB (using the sum of the
P
nitrite reductase genes as functional markers, nir (nirK
nirS)), SRB (using dsrA as the functional marker) and MPB
(using mcrA as the functional marker), were investigated in
this study.
Fig. 4a and b shows that the numbers of copies of the 16S
rRNA gene per gram litter were higher than those per gram
gravel, which indicated that the cattail litter was also a good
biofilm carrier for microbe aggregation, probably due to its
large surface area, porous structure and continuous carbon
supply (Fleming-Singer and Horne, 2002). However, on a vol-
ume basis (genes copies m3), the numbers of copies of the
genes in cattail litter were 2e4 orders of magnitude lower than
those on gravel (seen in Table S2). Thus, it was expected that
microorganisms on gravel provided a major contribution to
carbon removal in CWs compared to the ones on litter.
As shown in Fig. 4a and b, the numbers of copies of the 16S
rRNA gene in the W1 microcosm was 0.6e3.3  104 copies g1
gravel (mcrA), 6.8e9.9  106 copies g1 gravel (dsrA) and
3.0e3.9  106 copies g1 gravel (nir), which accounted for
0.03e0.14% (mcrA), 30.2e40.8% (dsrA) and 12.5e17.4% (nir) of
the total number of bacterial 16S rRNA genes, respectively.
The significant number of dsrA observed in the wetland
gravels agreed with previous studies showing that the SRB
were the dominant group in SSF CWs (Baptista et al., 2008;
Krasnits et al., 2009). Compared with the W1 microcosm, the
numbers of copies of mcrA, dsrA and nir genes per gram
gravel increased 100e123%, 90e124% and 51e83%, respec-
tively in the W2 microcosm. This suggested that the increased
carbon supply from the cattail litter could facilitate the growth
of MPB, SRB and DB on wetland gravel. Previous studies also
found that microbial biomass and activity was correlated with
organic carbon and that the decomposition products from
woody and herbaceous materials could be utilized by micro-
bial communities in CWs (Nguyen, 2000; Salomo et al., 2009).
Furthermore, as the organic carbon from cattail litter dropped
over time, the quantity of MPB dramatically decreased, but
there was only a small decrease in other microbes, which
indicated that the MPB might be the most affected by carbon
supply in CWs.
3.6. The relationship between COD consumption and the
relative abundance of DB, SRB and MPB
Given that the plant litter in CWs was continuously decom-
posing and the dry weight decreased during the experiment,
we compared the numbers of copies of mcrA, dsrA and nir
genes per m3 of wetland (seen in Table S3), rather than
numbers per gram, with the COD consumption by each
microbe in order to understand whether the carbon sources
competition was related to the quantity of microbes. As
shown in Fig 5, the COD consumption (MPB) : COD consumption
(SRB) ratio was well correlated with the mcrA gene copies : dsrA
gene copies ratio (R2 0.93), suggesting that competition
between carbon sources occurred and the relative abundance
of MPB and SRB significantly influenced carbon flow in the
wetlands. The similar thermodynamic characteristics and the
overlapping ecological niches for MPB and SRB resulted in
carbon competition in the biological system (Ren et al., 2009).
Thus, the utilization rate of substrates (i.e. acetate and
hydrogen), which is dependent on the quantity and activity of
microbes, dominated the COD consumption. However, a sig-
nificant correlation was not found between the COD con-
sumption ratio and the gene copies ratio for nir and dsrA. This
indicated that some other factors, other than the relative
abundance of the microbes, mainly influenced the COD
competition between DB and SRB. Faulwetter et al. (2009)
suggested that the nitrate reduction (DG0 118.8 kJ mol of
electrons1) could yield much more energy than sulfate
reduction (DG0 25.4 kJ mol of electrons1) and methano-
genesis (DG0 23.2 kJ mol of electrons1), which indicated
that the COD was preferentially consumed by DB. Further-
more, the presence of nitrate in influent could also inhibit SRB
and MPB (Haveman et al., 2004; Tugtas and Pavlostathis, 2007)
but may facilitate nitrate reduction at the beginning of each
batch. As mentioned above, the thermodynamic sequence
and the influent nitrate co-determined a preferential utiliza-
tion of COD for nitrate reduction, regardless of the relative
abundance of DB, but relative abundance was important for
the subsequent COD competition between SRB and MPB.
3.7. Carbon sources competition between SRB and MPB
3.7.1. Diversity and structure of SRB and MPB revealed by
high-throughput sequencing
In this study, 454 high-throughput sequencing was applied to
analyze the 16S rRNA gene of bacteria in CWs. Results showed
that Desulfobacter (619 sequences) and Desulfovibrio (380 se-
quences) were the dominant SRB (Fig. S3), and Methanosaeta
(3772 sequences) and Methanoregula (4436 sequences) were the
dominant MPB at the genus level of CWs (Table S4). Previous
Fig. 5 e The relationship between the COD consumption
ratio and the 16S rRNA gene copies ratio among different
microbial communities in wetland microcosms added with
litter.
 w a t e r r e s e a r c h 5 9 ( 2 0 1 4 ) 1 e1 0
literature demonstrate that Desulfobacter and Methanosaeta are
respectively aceticlastic SRB and MPB (utilization of acetate),
and Desulfovibrio and Methanoregula are hydrogenotrophic SRB
and MPB (utilization of H2) (Ren et al., 2009). Therefore, it is
possible that the SRB (i.e. Desulfobacter) and MPB (i.e. Meth-
anosaeta) could compete for the carbon sources in CWs
because they share the same substrate (acetate). The pro-
duction and competition of acetate by SRB and MPB can be
seen in Fig. S4. Given that acetate was only accumulated in the
initial stage but it was not observed in the middle and terminal
stages (Fig S5 in SI), the carbon competition between SRB and
MPB was expected to be significant during the middle and
terminal stages when there was a lack of acetate in CWs.
3.7.2. Causal model of carbon sources competition between
SRB and MPB
Path analysis was used to test different causal models
explaining the variation in COD consumption between SRB
and MPB using predictive variables (COD: S ratio, relative
abundance of genes and thermodynamic conditions). Thus,
the direct and indirect factors for carbon competition can be
identified and visualized. As shown in Fig. S6, for carbon
sources competition between MPB and SRB, the most impor-
tant predictive variables were relative abundance of genes
(R2 0.49) and COD: S ratio (R2 0.50). Furthermore, relative
abundance of genes was positively correlated with COD: S
ratio (R2 0.97). This indicated that the COD: S ratio could be
the driven force for the carbon sources competition between
MPB and SRB, because the COD: S ratio can either directly or
indirectly (through change the relative abundance of genes)
influence the carbon sources competition in CWs. Further-
more, thermodynamic conditions had no significant direct
impact on carbon sources competition between MPB and SRB.
The less important role of thermodynamic conditions in car-
bon sources competition in CWs was likely due to the similar
thermodynamic characteristic and overlapping ecological
niches of MPB and SRB (Faulwetter et al., 2009).
According to the carbon competition between MPB and
SRB, it is recommended to harvest the senescent cattail litters
in CWs and add them as a carbon source evenly throughout a
year, in order to achieve a relatively low COD: S ratio that
could facilitate the sulfate reduction and inhibit the CH4 for-
mation simultaneously in CWs.
4. Conclusion
The effects of T. latifolia and its litter on sulfate removal, as
well as the organic carbon competition in laboratory-scale SSF
CWs, were investigated in this study. The results showed that
the presence of cattail plants had a marginal effect on sulfate
reduction in CWs, probably due to the balance of positive
(supply of organic carbon) and negative (release oxygen from
roots) effects. The addition of cattail litter could greatly
improve the sulfate removal in SSF CWs through the contin-
uous input of labile organic carbon, lowering the redox level
and supplying habitats for SRB. Stoichiometry calculations
suggested that the contribution of SRB to COD consumption
(21e26%) was less than that of MPB (47e61%) during the initial
stage but dominated COD consumption (42e65%) during the
9
terminal stage. Path analysis illustrated that both COD: S ratio
and the relative abundance of 16S rRNA genes could be used
as the good predictors of carbon competition between SRB and
MPB in CWs.
Acknowledgments
This work was financially supported by the National Natural
Science Foundation of China (51378372), the Fundamental
Research Funds for the Central Universities (20113139), and
Czech University of Life Sciences and ESF & MEYS (CZ.1.07/
2.3.00/30.0040).
Appendix A. Supplementary data
Supplementary data related to this article can be found at
http://dx.doi.org/10.1016/j.watres.2014.03.077.
references
Allen, W.C., Hook, P.B., Biederman, J.A., Stein, O.R., 2002.
Temperature and wetland plant species effects on wastewater
treatment and root zone oxidation. J. Environ. Qual. 31 (3),
1010e1016.
APHA, 1998. Standard Methods for the Examination of Water and
Wastewater. American Public Health Association,
Washington, DC.
Baptista, J.D., Davenport, R.J., Donnelly, T., Curtis, T.R., 2008. The
microbial diversity of laboratory-scale wetlands appears to be
randomly assembled. Water Res. 42 (12), 3182e3190.
Baptista, J.D.C., Donnelly, T., Rayne, D., Davenport, R.J., 2003.
Microbial mechanisms of carbon removal in subsurface flow
wetlands. Water Sci. Technol. 48 (5), 127e134.
Callaway, R.M., King, L., 1996. Temperature-driven variation in
substrate oxygenation and the balance of competition and
facilitation. Ecology 77 (4), 1189e1195.
Chen, Y., Wen, Y., Cheng, J., Xue, C.H., Yang, D.H., Zhou, Q., 2011.
Effects of dissolved oxygen on extracellular enzymes activities
and transformation of carbon sources from plant biomass:
implications for denitrification in constructed wetlands.
Bioresour. Technol. 102 (3), 2433e2440.
Choi, E., Rim, J.M., 1991. Competition and inhibition of sulfate
reducers and methane reducers in anaerobic treatment.
Water Sci. Technol. 23, 1259e1264.
Faulwetter, J.L., Gagnon, V., Sundberg, C., Chazarenc, F.,
Burr, M.D., Brisson, J., Camper, A.K., Stein, O.R., 2009. Microbial
processes influencing performance of treatment wetlands: a
review. Ecol. Eng. 35 (6), 987e1004.
Fleming-Singer, M.S., Horne, A.J., 2002. Enhanced nitrate removal
efficiency in wetland microcosms using an episediment layer
for denitrification. Environ. Sci. Technol. 36 (6), 1231e1237.
Garcia, J., Aguirre, P., Mujeriego, R., Huang, Y.M., Ortiz, L.,
Bayona, J.M., 2004. Initial contaminant removal performance
factors in horizontal flow reed beds used for treating urban
wastewater. Water Res. 38 (7), 1669e1678.
Gebremariam, S.Y., Beutel, M.W., 2008. Nitrate removal and DO
levels in batch wetland mesocosms: cattail (Typha spp.)
versus bulrush (Scirpus spp.). Ecol. Eng. 34 (1), 1e6.
Gonzalias, A.E., Kuschk, P., Wiessner, A., Jank, M., Kastner, M.,
Koser, H., 2007. Treatment of an artificial sulphide containing
10 w a t e r r e s e a r c h 5 9 ( 2 0 1 4 ) 1 e1 0
wastewater in subsurface horizontal flow laboratory-scale
constructed wetlands. Ecol. Eng. 31 (4), 259e268.
Greenway, M., 2005. The role of constructed wetlands in
secondary effluent treatment and water reuse in subtropical
and arid Australia. Ecol. Eng. 25 (5), 501e509.
Haveman, S.A., Greene, E.A., Stilwell, C.P., Voordouw, J.K.,
Voordouw, G., 2004. Physiological and gene expression
analysis of inhibition of Desulfovibrio vulgaris Hildenborough
by nitrite. J. Bacteriol. 186 (23), 7944e7950.
Hume, N.P., Fleming, M.S., Horne, A.J., 2002. Plant carbohydrate
limitation on nitrate reduction in wetland microcosms. Water
Res. 36 (3), 577e584.
Isa, Z., Grusenmeyer, S., Verstraete, W., 1986. Sulfate reduction
relative to methane production in high-rate anaerobic
digestion: technical aspects. Appl. Environ. Microbiol. 51 (3),
572e579.
IWA, 2000. Constructed Wetlands for Pollution Control. Processes,
Performance, Design and Operation. International Water
Association, London.
Jiang, G.M., Sharma, K.R., Guisasola, A., Keller, J., Yuan, Z.G., 2009.
Sulfur transformation in rising main sewers receiving nitrate
dosage. Water Res. 43 (17), 4430e4440.
Kadlec, R.H., Wallace, S.D., 2008. Treatment Wetlands, second ed.
CRC Press, Boca Raton, FL.
Krasnits, E., Friedler, E., Sabbah, I., Beliavski, M., Tarre, S.,
Green, M., 2009. Spatial distribution of major microbial groups
in a well established constructed wetland treating municipal
wastewater. Ecol. Eng. 35 (7), 1085e1089.
Krishnakumar, B., Manilal, V.B., 1999. Bacterial oxidation of
sulphide under denitrifying conditions. Biotechnol. Lett. 21 (5),
437e440.
Lin, T., Wen, Y., Jiang, L.G., Li, J.B., Yang, S.L., Zhou, Q., 2008. Study
of atrazine degradation in subsurface flow constructed
wetland under different salinity. Chemosphere 72 (1),
122e128.
Londry, K.L., Suflita, J.M., 1999. Use of nitrate to control sulfide
generation by sulfate-reducing bacteria associated with oily
waste. J. Indus. Microbiol. Biotechnol. 22 (6), 582e589.
Margulies, M., Egholm, M., Altman, W.E., Attiya, S., Bader, J.S.,
Bemben, L.A., Berka, J., Braverman, M.S., Chen, Y.J., Chen, Z.T.,
Dewell, S.B., Du, L., Fierro, J.M., Gomes, X.V., Godwin, B.C.,
He, W., Helgesen, S., Ho, C.H., Irzyk, G.P., Jando, S.C.,
Alenquer, M.L.I., Jarvie, T.P., Jirage, K.B., Kim, J.B., Knight, J.R.,
Lanza, J.R., Leamon, J.H., Lefkowitz, S.M., Lei, M., Li, J.,
Lohman, K.L., Lu, H., Makhijani, V.B., McDade, K.E.,
McKenna, M.P., Myers, E.W., Nickerson, E., Nobile, J.R.,
Plant, R., Puc, B.P., Ronan, M.T., Roth, G.T., Sarkis, G.J.,
Simons, J.F., Simpson, J.W., Srinivasan, M., Tartaro, K.R.,
Tomasz, A., Vogt, K.A., Volkmer, G.A., Wang, S.H., Wang, Y.,
Weiner, M.P., Yu, P.G., Begley, R.F., Rothberg, J.M., 2005.
Genome sequencing in microfabricated high-density picolitre
reactors. Nature 437 (7057), 376e380.
Nguyen, L.M., 2000. Organic matter composition, microbial
biomass and microbial activity in gravel-bed constructed
wetlands treating farm dairy wastewaters. Ecol. Eng. 16 (2),
199e221.
Pinney, M.L., Westerhoff, P.K., Baker, L., 2000. Transformations in
dissolved organic carbon through constructed wetlands.
Water Res. 34 (6), 1897e1911.
Quanrud, D.M., Karpiscak, M.M., Lansey, K.E., Arnold, R.G., 2004.
Transformation of effluent organic matter during subsurface
View publication stats
wetland treatment in the Sonoran Desert. Chemosphere 54 (6),
777e788.
Ren, N.Q., Wang, A.J., Zhao, Y.G., 2009. Ecology of Sulfate-
reducing Bacteria in Anaerobic Biotreatment Processes.
Sciencep, China.
Rethmeier, J., Rabenstein, A., Langer, M., Fischer, U., 1997.
Detection of traces of oxidized and reduced sulfur compounds
in small samples by combination of different high-
performance liquid chromatography methods. J. Chromatogr.
760 (2), 295e302.
Rittmann, B.E., McCarty, P.L., 2001. Environmental Biotechnology:
Principles and Applications. McGraw Hill, New York.
Salomo, S., Munch, C., Roske, I., 2009. Evaluation of the metabolic
diversity of microbial communities in four different filter
layers of a constructed wetland with vertical flow by Biolog
(TM) analysis. Water Res. 43 (18), 4569e4578.
Stein, O.R., Borden-Stewart, D.J., Hook, P.B., Jones, W.L., 2007.
Seasonal influence on sulfate reduction and zinc
sequestration in subsurface treatment wetlands. Water Res.
41 (15), 3440e3448.
Tugtas, A.E., Pavlostathis, S.G., 2007. Inhibitory effects of nitrogen
oxides on a mixed methanogenic culture. Biotechnol. Bioeng.
96 (3), 444e455.
Vymazal, J., 2009. The use constructed wetlands with horizontal
sub-surface flow for various types of wastewater. Ecol. Eng. 35
(1), 1e17.
Vymazal, J., Kropfelova, L., 2008. Wastewater Treatment in
Constructed Wetlands with Horizontal Sub-surface Flow.
Springer, Dordrecht.
Wei, L.L., Zhao, Q.L., Xue, S., Jia, T., Tang, F., You, P.Y., 2009.
Behavior and characteristics of DOM during a laboratory-scale
horizontal subsurface flow wetland treatment: effect of DOM
derived from leaves and roots. Ecol. Eng. 35 (10), 1405e1414.
Wen, Y., Chen, Y., Zheng, N., Yang, D.H., Zhou, Q., 2010. Effects of
plant biomass on nitrate removal and transformation of
carbon sources in subsurface-flow constructed wetlands.
Bioresour. Technol. 101 (19), 7286e7292.
Wen, Y., 2007. A Study on Treatment of Polluted Surface Water by
Using Horizontal Subsurface-flow Constructed Wetlands.
Ph.D. Dissertation. Tongji University, China.
Whitmire, S.L., Hamilton, S.K., 2005. Rapid removal of nitrate and
sulfate in freshwater wetland sediments. J. Environ. Qual. 34
(6), 2062e2071.
Wiessner, A., Gonzalias, A.E., Kastner, M., Kuschk, P., 2008. Effects
of sulphur cycle processes on ammonia removal in a
laboratory-scale constructed wetland planted with Juncus
effusus. Ecol. Eng. 34 (2), 162e167.
Wiessner, A., Rahman, K.Z., Kuschk, P., Kastner, M., Jechorek, M.,
2010. Dynamics of sulphur compounds in horizontal sub-
surface flow laboratory-scale constructed wetlands treating
artificial sewage. Water Res. 44 (20), 6175e6185.
Wu, S.B., Jeschke, C., Dong, R.J., Paschke, H., Kuschk, P.,
Knoller, K., 2011. Sulfur transformations in pilot-scale
constructed wetland treating high sulfate-containing
contaminated groundwater: a stable isotope assessment.
Water Res. 45 (20), 6688e6698.
Zhao, B.H., Yue, Z.B., Ni, B.J., Mu, Y., Yu, H.Q., Harada, H., 2009.
Modeling anaerobic digestion of aquatic plants by rumen
cultures: cattail as an example. Water Res. 43 (7), 2047e2055.