Microbes are believed to be at the core of the wastewater treatment processes in constructed wetlands (CWs). The aim of this study was to assess the microbial biomass carbon (MBC) and Shannon's diversity index (SDI) in the substrate of CWs planted with Phragmites australis, Hymenocallis littoralis, Canna indica and Cyperus flabelliformis, and to relate MBC and SDI to the pollutant removal in the systems. Significant higher MBC was observed in CWs with H. littoralis and C. indica than in CWs with P. australis, and the MBC differed with season and substrate depth. The microbial community in the wetlands included four phyla: Cyanobacteria, Proteobacteria, Chloroflexi, and Acidobacteria, with a more diverse community structure in wetlands with C. flabelliformis. The MBC in the substrate and the SDI of the 15–20 cm depth correlated with the removal of biochemical oxygen demand, NH4-N and NO3-N. Our results indicate that substrate SDI and MBC can both be regarded as bioindicators of the pollutant removal ability in CWs.

INTRODUCTION

Constructed wetlands (CWs) are widespread as a decentralized wastewater treatment option because of their high economic benefits, environmental friendliness and high pollutant removal efficiency. Microorganisms play important roles in the wastewater treatment processes in CWs, as most removal processes are mediated by microbial processes (Stottmeister et al. 2003).

Faulwetter et al. (2009) reported that the removal of a particular pollutant is typically correlated with a specific microbial functional group. Therefore, deepening the understanding of the microbial community structure is helpful to reveal the mechanisms of pollutant removal in CWs. Polymerase chain reaction-denaturing gradient gel electrophoresis (PCR DGGE) is a useful tool for assessing microbial community structure diversity (Ibekwe et al. 2003). This method was first applied to microbial ecology by Muyzer et al. (1993), and since then the method has been used to explore the community structure of bacteria in CWs (Adrados et al. 2014). The Shannon diversity index (SDI) can be obtained from DGGE finger printing profiles to quantify the diversity levels in different samples. Another important parameter is microbial biomass carbon (MBC) which is usually used to evaluate microbial biomass. MBC is an indirect measurement of microbial density and has been suggested to be a useful parameter for evaluating the pollutant removal efficiency in CWs (Truu et al. 2009).

Hence, the MBC and SDI can provide profound information about the microorganisms in CWs. Nevertheless, most of the studies on MBC and SDI have been focused on their relation to environmental factors, such as land utilization, plant species and seasons (Calheiros et al. 2009). We are not aware of any research aimed at studying the correlation between MBC, SDI and pollutant removal efficiencies in CWs. A comprehensive understanding of the microbial community and its relation to the pollutant removal efficiencies in CWs can lead to a better understanding of degradation processes and pathways in CWs. The aims of the present study were: (i) to assess the pollutant removal efficiency, the MBC and the microbial community structure in the substrate of CWs planted with four plant species; (ii) to analyze the relationships between MBC, SDI, and pollutant removal efficiency in the systems; and finally (iii) to deepen our understanding of the roles of the microbial communities in the pollutant removal processes in CWs.

METHODS

Experimental setup

The experiment was conducted in Guangzhou, China, where the annual average temperature is 21 °C, and the annual precipitation varies from 1,230 to 2,491 mm, 85% of which falls during the wet season (from April to September). Sixteen CWs (2.0 m × 1.0 m × 0.7 m; length × width × depth) were constructed with a substrate consisting of fresh river sand and crushed granite (diameter approximately 5 mm) with substrate depth of 0.55 m. The water level was kept 0.15 m above the substrate. Four wetland plant species, Phragmites australis (Cav.) Trin ex Steudel, Hymenocallis littoralis (Jacq.) Salisb., Canna indica Linn. and Cyperus flabelliformis Rottb., were planted as mono-cultures in the beds at a density of four plants per square metre (n = 4) in March 2011. The CWs were batch-loaded (one loading per 7 days) with effluent from a septic tank treating the domestic sewage from a university dormitory (300 population equivalent). The influent was loaded with a pump above the substrate level at one end of the beds, and the effluent was controlled with a drainage pipe placed at the other end of the beds 15 cm above the substrate. The hydraulic loading rate was 28.6 mm d−1 resulting in a nominal hydraulic retention time of 7 days.

Pollutant removal

Five sampling campaigns were conducted during spring (March), summer (May, July) and winter (October, December) of 2011, where each replicate (n = 4) of the planted beds was sampled (n = 20). Composite samples (100 mL) of both influent and effluent water were collected per bed, per batch load when feeding and emptying. The water samples were transported to the laboratory and stored at 4 °C for no more than 4 h until analysis. The water samples were analyzed for the conventional water quality parameters following the procedures described by Wei et al. (2002). In brief, the biochemical oxygen demand (BOD5) was measured by the dilution and inoculation method, the chemical oxygen demand (COD) by the potassium dichromate method, total phosphorus (TP) by the potassium persulfate oxidation–molybdenum blue colorimetric method, soluble phosphorus (SP) by the molybdenum blue colorimetric method, total nitrogen (TN) by the alkaline potassium persulfate decomposition–UV spectrophotometry method, ammonium-nitrogen (NH4-N) by the Nessler's reagent spectrophotometry method, and nitrate-nitrogen (NO3-N) by the UV spectrophotometry method. Removal efficiencies were calculated on the basis of the difference between influent and effluent concentration at each sampling time.

MBC

Two sampling campaigns were conducted in summer (July) and winter (December) of 2011, respectively, for microbial analysis of the substrate in all beds. Prior to the sampling, the surface water in the beds was drained. Substrate samples were then taken from the surface layer (0–5 cm depth) and the deeper (15–20 cm depth) layer with a core sampler. Great effort was applied when collecting the deeper layer samples, to avoid the disturbance from interstitial water and the detachment of microbes from the substrate. The samples (approximately 150 g) were sieved through a 2 mm diameter mesh-size sieve to remove plant material and coarse sand and to homogenize the samples before analysis of MBC using the improved chloroform fumigation extraction method (Witt et al. 2000).

Microbial community structure

DNA extraction

In summer (July) and winter (December) of 2011, samples of the substrate were taken as described above for community structure analysis from all systems. A composite sample from the 4 replicates of each monoculture in summer (July) and winter (December) and both depths (0–5 and 15–20 cm) was further analyzed. The DNA of the bacteria in the substrate was extracted directly following the procedure described in the manual of the Power Soil DNA Isolation Kit (Mo Bio, USA). The extractions were stored at −20 °C until further analysis.

PCR amplification

PCR amplification of the bacteria 16S rRNA target fragments was performed in a BioRad PCR cycler (PCT200, BioRad, USA) using the universal bacterial primers GC-968f (5′-CGCCCGCCGC GCGCGGCGGG CGGGGCGGGG GCACGGGGGG AACGCGAAGA ACCTTAC-3′) and 1401R (5′-CGGTGTGTACAAGGCCCGGGAACG-3′). The amplification fragment was about 440 bp. Each 50 μL of PCR mixture contained 1 μL of each forward and reverse primer (10 pmol each), 1 μL of DNA template, 25 μL of Taq-Mix (Dongsheng, China), and 22 μL of sterile deionized water. The amplification program was described by Evans et al. (2004). The PCR products were checked in a 1% agarose gel with Gel Red (Biotium, USA) under UV light (Muyzer et al. 1993).

DGGE

DGGE was carried out using a 6% (w/v) polyacrylamide gel with a denaturant gradient from 40 to 60%, where the 100% denaturant was defined as 7 mol L−1 urea and 40% formamide (v/v). Twenty microlitres of the PCR samples were added into the denaturing gradient gel holes. After running in 1× TAE (Tris-Acetate-EDTA) buffer for 15 h at constant conditions of 60 °C and 80 V within the D-Code universal mutation detection system (Bio-Rad, USA), the gel was checked with Gel Red (Biotium, USA) under UV light within a molecular imager (Bio-Rad, USA).

DNA sequencing and DGGE analysis

The target bands were carefully excised from the DGGE gel and rinsed twice with 20 μL of sterilized distilled water in sterilized centrifuge tubes and then eluted overnight at −20 °C with 20 μL of sterilized distilled water. The PCR system described above (the primer pair 968f–1401r without GC clamp) was used to amplify the template. The target products were purified using an agarose gel extraction kit (Sigma, USA), subsequently cloned into pMD18-T vector (TaKaRa, Dalian) and finally transformed into the competent Escherichia coli cells (TaKaRa, Dalian, China). The positive clones were identified, and the cultured bacteria solutions sequenced by Shanghai Bio-Technologies Co., Ltd (Shanghai, China).

The banding patterns of the DGGE profile were analyzed by the Quantity One software (Bio-Rad, USA). The unweighted pair group method with arithmetic mean was used to generate the cluster chart. Sequences were compared with the GenBank database.

The SDI was calculated by the peak density value obtained by the Quantity One with the biodap software (Thomas & Clay 2000). The SDI was calculated as: 
formula
1
where ni represents the peak intensity of each individual band in the gel, and N is the sum of the peak intensities of all the bands.

Statistical analysis

Statistical analysis was performed using the SPSS statistical software (IBM SPSS statistics 20.0, USA). Repeated measures ANOVA (analysis of variance) was conducted to compare pollutant removal efficiencies between CWs planted with different plant species along the experiment. By doing this, repeated measures were selected from general linear model function in SPSS. Five different months were chosen as within-subject factors and different planted CWs were chosen as subject factors. The difference in MBC in the substrate of beds with different plant species was evaluated using two-way ANOVA. One-way ANOVA and post hoc Tukey's HSD (honest significant difference) tests were used to compare the depth and season influence on the MBC in each CW at the 95% confidence level. Spearman correlation analysis was used to evaluate the relationships between the SDI, MBC and pollutant removal efficiency in July and December. Prior to statistical analysis, all data were tested for homogeneity of variance by Levene's test.

RESULTS AND DISCUSSION

Pollutant removal efficiency

The pollutant removal efficiencies were stable along the experimental period (March to December 2011), even when taking into consideration that March was the start-up period (Table 1). It should be noted that rain and evapotranspiration, which can influence pollutant concentrations, were not measured. The average removal efficiencies of COD, BOD5, TP, TN and NO3-N were about 70% for all CWs, with no statistically significant differences among the different plant species.

Table 1

Average pollutant removal efficiencies (%) during March to December 2011 in CWs planted with four different wetland species (mean ± SD, n = 20)

  BOD5 COD TP SP TN NH4-N NO3-N 
C. flabelliformis 68 ± 9 72 ± 9 75 ± 11 70 ± 11 70 ± 3 67 ± 10 67 ± 6 
H. littoralis 70 ± 8 60 ± 6 72 ± 10 76 ± 6 67 ± 4 63 ± 9 66 ± 6 
C. indica 64 ± 12 68 ± 13 67 ± 14 72 ± 9 66 ± 6 61 ± 6 64 ± 9 
P. australis 67 ± 4 66 ± 9 73 ± 10 73 ± 7 70 ± 5 65 ± 10 67 ± 2 
  BOD5 COD TP SP TN NH4-N NO3-N 
C. flabelliformis 68 ± 9 72 ± 9 75 ± 11 70 ± 11 70 ± 3 67 ± 10 67 ± 6 
H. littoralis 70 ± 8 60 ± 6 72 ± 10 76 ± 6 67 ± 4 63 ± 9 66 ± 6 
C. indica 64 ± 12 68 ± 13 67 ± 14 72 ± 9 66 ± 6 61 ± 6 64 ± 9 
P. australis 67 ± 4 66 ± 9 73 ± 10 73 ± 7 70 ± 5 65 ± 10 67 ± 2 

Wetland plants play an important function as ecosystem engineers in CWs, although the direct uptake of nutrients by the plants has been shown to contribute only a small fraction of the amount of nutrients removed (Tanner 2001). An important factor influencing the contaminant removal is the plant species used, as the species' abilities to promote aerobic over anaerobic microbial processes are significative (Taylor et al. 2011). However, in the present study, the removal efficiency of the different pollutants did not differ between the plant species.

MBC

Figure 1 shows the variation in MBC with substrate depth and season in the CWs. For all plant species, the MBC was significantly (P < 0.05) higher in the surface 0–5 cm depth substrate than in the deeper 15–20 cm depth substrate, whereas there was no difference between the substrate depths in winter. The MBC in the surface layer was significantly (P < 0.05) higher in summer than in winter in all systems, except the systems planted with C. indica, whereas the MBC in the deeper substrate layer did not differ between seasons.
Figure 1

Variations of MBC in CWs planted with four different wetland species. Different letters above the columns indicate significant differences at the 95% probability level.

Figure 1

Variations of MBC in CWs planted with four different wetland species. Different letters above the columns indicate significant differences at the 95% probability level.

The presence of plants is known to enhance the microbial density in CWs (Gagnon et al. 2007). Kong et al. (2009) and Windham (2001) found that the highest proportion of the root biomass occurred in the 0–10 cm depth for the plant species studied. In our study, the MBC in the 0–5 cm depth was significantly higher than the MBC in the 15–20 cm depth layer in all beds in summer, probably because the dense root systems in this depth provide habitat and nutrients for the microorganisms and stimulate their growth and reproduction. During winter, the MBC in the surface layer of all beds was lower, probably reflecting the fact that the growth and activity of plants were low because of the low temperatures. This is in agreement with results previously published that showed that bacterial community density increased with the temperature (Xiong et al. 2013). The higher temperatures in summer activate soil enzyme activities, resulting in more available nutrients for microorganisms and consequently promoting microorganisms' reproduction and growth. In contrast, the low temperatures in winter decrease the soil enzyme activities, leading to less available nutrients for microorganisms. Meanwhile, the microorganisms' activities also decrease at low temperature, which results in slow growth of microorganisms and subsequently less MBC. The 15–20 cm depth layer of the substrate was not exposed directly to the air temperature, which justifies the lack of seasonal difference in MBC in this layer. Furthermore, along the time of the experiment it was observed that both analyzed layers had roots, with higher presence in the upper layer than in the deeper layer. For future work, the introduction of a deeper layer outside the root zone could help to explain the relevance of the plant/root presence on MBC.

When comparing the MBC among the plant species, it is seen that the substrates in beds planted with H. littoralis and C. indica had significantly higher MBC than beds planted with P. australis (P < 0.05). This might be caused by the fast growth rate of H. littoralis and greater amount of litter created by C. indica compared to, for example, P. australis.

Microbial community structure

Seventeen different bands were obtained from the DGGE gel and sequenced successfully (Figure 2(a)). Six bands appeared in all samples; three bands only appeared in extracts from the surface substrate layer and two bands only appeared in substrate from the C. flabelliformis wetlands. Table 2 shows that the SDI of the microbial community structure varied between the different layers of the substrate as well as the seasons. Because samples from July and December were combined, we only had a total n of 2 to reflect the variation in SDI with substrate depth. Similarly, samples from the two substrate depths were combined to reflect the variation with season. The substrate in the beds with C. flabelliformis showed the highest diversity, while the substrate in the beds with P. australis exhibited the lowest. The diversity in July was generally higher than in December, and the diversity was significantly higher in the surface layer than in the deeper layer for all the plant species (P < 0.05).
Table 2

SDI of the microbial community structure (mean ± SD, n = 2)

    C. flabelliformis H. littoralis C. indica P. australis 
Depth 0–5 cm 3.3 ± 0.1 3.0 ± 0.1 3.1 ± 0.1 2.9 ± 0.1 
15–20 cm 2.8 ± 0.2 2.8 ± 0.1 2.6 ± 0.3 2.7 ± 0.1 
Season Summer (July) 3.1 ± 0.2 3.0 ± 0.2 3.0 ± 0.2 2.8 ± 0.1 
Winter (December) 3.0 ± 0.4 2.8 ± 0.2 2.7 ± 0.5 2.8 ± 0.2 
    C. flabelliformis H. littoralis C. indica P. australis 
Depth 0–5 cm 3.3 ± 0.1 3.0 ± 0.1 3.1 ± 0.1 2.9 ± 0.1 
15–20 cm 2.8 ± 0.2 2.8 ± 0.1 2.6 ± 0.3 2.7 ± 0.1 
Season Summer (July) 3.1 ± 0.2 3.0 ± 0.2 3.0 ± 0.2 2.8 ± 0.1 
Winter (December) 3.0 ± 0.4 2.8 ± 0.2 2.7 ± 0.5 2.8 ± 0.2 
Figure 2

Schematic picture of detected bands in the DGGE gel (a) and cluster chart of bacterial community structure (b) in different wetlands samples (S: the surface 0–5 cm depth layer; B: the deeper 15–20 cm depth layer; Jul: July; Dec: December; Pa: P. australis; Ci: C. indica; Hl: H. littoralis; Cf: C. flabelliformis).

Figure 2

Schematic picture of detected bands in the DGGE gel (a) and cluster chart of bacterial community structure (b) in different wetlands samples (S: the surface 0–5 cm depth layer; B: the deeper 15–20 cm depth layer; Jul: July; Dec: December; Pa: P. australis; Ci: C. indica; Hl: H. littoralis; Cf: C. flabelliformis).

The DGGE atlas cluster analysis showed that all the samples clustered into two groups (Figure 2(b)): cluster I included samples from the 15–20 cm depth layer in December and cluster II included the remaining samples. In cluster II, all the December samples were in the same subgroup, and the July samples clustered in the other subgroup, except the 0–5 cm depth sample from the C. flabelliformis wetlands.

The soil microbial community structure can be influenced by environmental factors and plant species (Calheiros et al. 2009). In our experiment, the cluster analysis showed that all the samples were divided into two groups, indicating that the bacterial community structure had a strong seasonal influence. Moreover, the SDI in the C. flabelliformis wetlands was the highest, implying a more complex microbial community in these beds.

The higher SDI values obtained from DGGE profiles appeared to be in July and at the 0–5 cm depth layer. Calheiros et al. (2009) used SDI to evaluate the diversity of bacterial communities from two-stage CWs planted with Typha latifolia and P. australis treating tannery wastewater, and noted that the SDI in all CWs was in general above 1.0. Yin et al. (2009) found SDI values above 2.0 in horizontal subsurface flow CWs treating water from a polluted landscape lake, which is similar to our results. The difference might be related to the wastewater source, as Calheiros et al. (2009) worked with tannery industrial wastewater and showed that there is higher concentration of COD in comparison with the polluted lake water studied by Yin et al. (2009), which may inhibit the growth of bacteria and result in a lower SDI.

In the present study, the results from the sequencing bands showed that the bacterial species of the wetlands belonged to the families of Cyanobacteria, Proteobacteria, Chloroflexi and Acidobacteria. Plant root is an important factor for forming the community structure of the bacteria (Faulwetter et al. 2013). More specifically, Zhang et al. (2011) demonstrated that aqueous extracts of roots of the macrophyte Thalia dealbata inhibited the growth of the Cyanobacteria. In our study, bands 3 and 9 from the Cyanobacteria were found in all the CWs, indicating that the four plant species studied here did not release root exudates capable of inhibiting the growth of the Cyanobacteria. The Cyanobacteria are a Gram-negative group of bacteria, which occur ubiquitously in freshwaters, seawaters and soils, and even in extreme environments. The Cyanobacteria play an important role in the nitrogen fixation in ecosystems (Asplund & Wardle 2012). The Proteobacteria are also a group of Gram-negative bacteria, which include a high number of nitrogen fixing species. In the present study, the band 4 from the Proteobacteria only occurred in the 0–5 cm depth samples of the C. flabelliformis wetlands. The band 2, however, occurred in all wetlands, indicating that different Proteobacteria have different environmental requirements. Bacteria of the Chloroflexi phylum are widely spread in the environment, including hot springs and aerobic and anaerobic sludge. Bjornsson et al. (2002) have shown that the Chloroflexi also occurred in wastewater treatment processes. In the present study, band 1 from the Chloroflexi was present in all wetland samples. The Acidobacteria are a physiologically diverse group of bacteria that are consistently detected in many different habitats around the globe. However, the physiological information about the Acidobacteria is scarce. Shange et al. (2013) found that the Acidobacteria positively responded to the lack of carbon in the soil environment. In the present study, band 16, from the Acidobacteria, was present in all C. flabelliformis wetlands.

Correlation analysis

Tables 3 and 4 show the correlations between MBC, SDI, and pollutant removal efficiencies. Biunique data from each plant species obtained in July and December for these microbial parameters and pollutant removal efficiencies were analyzed. The MBC in both substrate depth layers was positively correlated with the removal efficiencies of BOD5, NH4-N and NO3-N (Table 3). The SDI at the 15–20 cm substrate depth layer positively correlated with the same pollutants, but with higher correlation coefficients. Negative correlations between SP and MBC in both layers, and SDI at the 15–20 cm depth layer were also observed. There was a positive correlation between MBC and SDI at the 15–20 cm depth layer and in July (Table 4).

Table 3

Correlations between pollutant removal efficiencies and the MBC, and the SDI in July and December in CWs planted with four different wetland species

    COD BOD5 TP SP TN NH4-N NO3-N 
MBC 0–5 cm 0.429 0.778* −0.167 −0.714* 0.587 0.714* 0.833* 
15–20 cm 0.405 0.743* −0.048 −0.833* 0.455 0.714* 0.714* 
SDI 0–5 cm 0.455 0.373 −0.036 −0.216 0.139 0.108 0.635 
15–20 cm 0.647 0.916** −0.204 −0.790* 0.584 0.826* 0.802* 
    COD BOD5 TP SP TN NH4-N NO3-N 
MBC 0–5 cm 0.429 0.778* −0.167 −0.714* 0.587 0.714* 0.833* 
15–20 cm 0.405 0.743* −0.048 −0.833* 0.455 0.714* 0.714* 
SDI 0–5 cm 0.455 0.373 −0.036 −0.216 0.139 0.108 0.635 
15–20 cm 0.647 0.916** −0.204 −0.790* 0.584 0.826* 0.802* 

*Significant correlation at the 95% probability level.

**Significant correlation at the 99% probability level.

Table 4

Correlations between the SDI of the microbes and the MBC in different substrate depths, and in July and December in CWs planted with four different wetland species

  SDI
 
 0–5 cm 15–20 cm July December 
MBC 0.359 0.743* 0.868** −0.359 
  SDI
 
 0–5 cm 15–20 cm July December 
MBC 0.359 0.743* 0.868** −0.359 

*Significant correlation at the 95% probability level.

**Significant correlation at the 99% probability level.

MBC has been suggested as a suitable biogeochemical indicator of pollutant removal efficiency, and a positive relation between MBC and the removal of carbon and nitrogen in CWs has been reported (Reddy & D'Angelo 1997). With the development of molecular techniques, more sensitive indicators, such as quantitative PCR, have been applied as a bioindicator for the evaluation of pollution assessment in water systems from the Pearl River Delta to the South China Sea (Cao et al. 2012). However, for CWs, this type of technique has been rarely studied. In the present study, the MBC at both substrate depths and the SDI at the deep 15–20 cm substrate depth were positively correlated with the removal of BOD5 NH4-N and NO3-N, which are mainly removed through microbial pathways in CWs. On the other hand, MBC and SDI were negatively correlated with the removal of phosphorus, which is known to be more recalcitrant to treatment in CWs. Hence, our results suggest that the SDI and MBC of the bed substrate might work as bioindicators of the pollutant removal ability in CWs. However, the MBC only correlated significantly with the SDI at the 15–20 cm depth layer and in July. Therefore, the relation between the microbial community, as characterized by parameters like MBC and SDI, and the treatment performance of CWs is complex. A more comprehensive understanding of the microbial density, activity and diversity is needed for further understanding of the relationship between the microbial community and the treatment performance of CWs.

CONCLUSIONS

The substrate in CWs planted with H. littoralis and C. indica had higher MBC than substrates in beds planted with P. australis, while CWs planted with C. flabelliformis generally had a higher diversity of the microbial community. The MBC in both the surface and deep substrate layers and the SDI of the deep substrate layer correlated with the removal of BOD5, NH4-N and NO3-N. Our results indicate that SDI and MBC can both be regarded as bioindicators of the pollutant removal ability in CWs.

REFERENCES

REFERENCES
Calheiros
C. S. C.
Duque
A. F.
Moura
A.
Henriques
I. S.
Correia
A.
Rangel
A. O. S. S.
Castro
P. M. L.
2009
Substrate effect on bacterial communities from constructed wetlands planted with Typha latifolia treating industrial wastewater
.
Ecol. Eng.
35
(
5
),
744
753
.
Evans
F. F.
Rosado
A. S.
Sebastian
G. V.
Casella
R.
Machado
P. L. O. A.
Holmstrom
C.
Kjelleberg
S.
van Elsas
J. D.
Seldin
L.
2004
Impact of oil contamination and biostimulation on the diversity of indigenous bacterial communities in soil microcosms
.
FEMS Microbiol. Ecol.
49
(
2
),
295
305
.
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
),
987
1004
.
Gagnon
V.
Chazarenc
F.
Comeau
Y.
Brisson
J.
2007
Influence of macrophyte species on microbial density and activity in constructed wetlands
.
Water Sci. Technol.
56
(
3
),
249
254
.
Kong
L.
Wang
Y. B.
Zhao
L. N.
Chen
Z. H.
2009
Enzyme and root activities in surface-flow constructed wetlands
.
Chemosphere
76
(
5
),
601
608
.
Muyzer
G.
De Waal
E. C.
Uitterlinden
A. G.
1993
Profiling of complex microbial populations by denaturing gradient gel electrophoresis analysis of polymerase chain reaction-amplified genes coding for 16S rRNA
.
Appl. Environ. Microbiol.
59
(
3
),
695
700
.
Stottmeister
U.
Wießner
A.
Kuschk
P.
Kappelmeyer
U.
Kästner
M.
Bederski
O.
Müller
R. A.
Moormann
H.
2003
Effects of plants and microorganisms in constructed wetlands for wastewater treatment
.
Biotechnol. Adv.
22
(
1–2
),
93
117
.
Tanner
C. C.
2001
Plants as ecosystem engineers in subsurface-flow treatment wetlands
.
Water Sci. Technol.
44
(
11–12
),
11
12
.
Taylor
C. R.
Hook
P. B.
Stein
O. R.
Zabinski
C. A.
2011
Seasonal effects of 19 plant species on COD removal in subsurface treatment wetland microcosms
.
Ecol. Eng.
37
(
5
),
703
710
.
Thomas
G.
Clay
D.
2000
Bio-Dap, ecological diversity and its measurement
. .
Truu
M.
Juhanson
J.
Truu
J.
2009
Microbial biomass, activity and community composition in constructed wetlands
.
Sci. Total Environ.
407
(
13
),
3958
3971
.
Wei
F. S.
Qi
W. Q.
Sun
Z. G.
Huang
Y. R.
Shen
Y. W.
2002
Water and Wastewater Monitoring and Analysis Methods
.
China Environmental Science Press
,
Beijing
.
Witt
C.
Gaunt
J. L.
Galicia
C. C.
Ottow
J. C. G.
Neue
H. U.
2000
A rapid chloroform-fumigation extraction method for measuring soil microbial biomass carbon and nitrogen in flooded rice soils
.
Biol. Fert. Soils
30
(
5–6
),
510
519
.
Yin
J.
Jiang
L. Y.
Wen
Y.
Yao
Z. L.
Zhou
Q.
2009
Treatment of polluted landscape lake water and community analysis of ammonia-oxidizing bacteria in constructed wetland
.
J. Environ. Sci. Health Part A
44
(
7
),
722
731
.