The Green Origio biological system, a two-stage, solar-powered constructed wetland mesocosm situated in Shanghai Jiao Tong University's plant park was used for nitrogen removal in the institution's landscape water. Whereas the first tank was anaerobic, the second tank containing fully grown Pontederia cordata was operated under aerobic conditions. The system was designed with an overflow by-pass pipe to prevent hydraulic overloading. Pump discharge values of 53, 74, 112, 120, 205 and 270 m3/day were recorded during the study. However, the corresponding flow rate values into the tanks were 16.25 m3/day, 29.47 m3/day, 36.79 m3/day, 50.57 m3/day, 71.38 m3/day and 80.52 m3/day, respectively. Experiments conducted between August and October, 2013 revealed a significant concentration decrease for total nitrogen (TN), ammonia (NH3-N), nitrate (NO3-N), nitrite (NO2-N) and total organic carbon (TOC). Whereas the Multi N/C 3100 Analyzer was used for simultaneous analysis of TN and TOC, the nitrogen fractions were analyzed using spectrophotometric methods. The TN, NO3-N, NH3-N, NO2-N and TOC removal rates under a flow rate of 16.25 m3/day were 55.65%, 75.19%, 74.03%, 95.43% and 29.87%, respectively. TN and NO2-N removal efficiencies of 4.92% and 83.64%, respectively, were observed under a flow rate of 80.52 m3/day. Concentration increase in the effluent caused by washout of accumulated concentrations was noted for NO3-N, NH3-N and TOC.

INTRODUCTION

Storm runoff from agricultural lands is one of the common sources of excessive nitrogen and phosphorus concentrations in the receiving waters. Consequently, oxygen depletion and eutrophication is observed in such water bodies. The recent utilization of constructed wetlands in the improvement of storm-water runoff quality is gaining popularity in both the developing and developed world (Kadlec & Knight 1996). A previous study proved that emergent and rooted floating-leaved macrophytes were capable of assimilating nutrients from the sediments. In addition, their ability to utilize sediment nutrients more efficiently compared to planktonic algae was observed (Wetzel 2001). Recent studies have revealed that total nitrogen (TN) removal from wastewater is not achieved in all biological removal systems. Rather, most processes convert nitrogen into its various forms (Reddy & Patrick 1984; Vymazal 2007). Nitrite (NO2-N), nitrate (NO3-N) and ammonia (NH3-N) are the most important inorganic nitrogen forms. It has also been noted that NH3-N and NO3-N are useful nitrogen forms responsible for nitrogen assimilation. However, the higher ability of NO3-N to be energetically reduced makes it a preferable source of nitrogen for assimilation (Kadlec & Knight 1996). Table 1 summarizes the different nitrogen transformation pathways in constructed wetlands. In a study involving five different macrophytes, whereas 4–11% of the nitrogen was removed by the planted wetland, 89–96% was due to denitrification (Lin et al. 2002). Adequate hydraulic retention time (HRT) and microbial residence time favors the denitrification process and plant uptake (assimilation) hence higher nitrogen removal (Akratos & Tsihrintzis 2007; Chung et al. 2008).

Table 1

Nitrogen transformation in constructed wetlands (Vymazal 2007)

Process Transformation Pathway 
Volatilization ammonia-N (aq) → ammonia-N (g) 
Ammonification organic-N → ammonia-N 
Nitrification ammonia-N → nitrite-N → nitrate-N 
Nitrate-ammonification nitrate-N → ammonia-N 
Denitrification nitrate-N → nitrite-N → gaseous N2,N2
Nitrogen Fixation gaseousN2 → ammonia-N (organic-N) 
Plant/microbial uptake (assimilation) ammonia-, nitrite-, nitrate-N → organic-N 
ANAMMOX (anaerobic ammonia oxidation) ammonia-N → gaseous N2 
Process Transformation Pathway 
Volatilization ammonia-N (aq) → ammonia-N (g) 
Ammonification organic-N → ammonia-N 
Nitrification ammonia-N → nitrite-N → nitrate-N 
Nitrate-ammonification nitrate-N → ammonia-N 
Denitrification nitrate-N → nitrite-N → gaseous N2,N2
Nitrogen Fixation gaseousN2 → ammonia-N (organic-N) 
Plant/microbial uptake (assimilation) ammonia-, nitrite-, nitrate-N → organic-N 
ANAMMOX (anaerobic ammonia oxidation) ammonia-N → gaseous N2 

The use of several emergent constructed wetland macrophytes in water pollution control for both research and industrial purposes is a common phenomenon. Typha latifolia, Scirpus validus, Cyperus alternifolius, Phragmites australis and Pontederia cordata represent a few plant species that are widely used in water pollution control (Vymazal 2011). Pontederia cordata (Pickerel weed) originated from North America. It is characterized by tall upright stems, bright green, spoon shaped lanceolate leaves and deep blue/purple flowers (Audlbach-Smith & deKozlowski 1996). The flowering process is usually from May to October. The optimal temperatures that support their growth range from 18 to 35 °C. At temperatures below 18 °C, slow growth rate is observed (Audlbach-Smith & deKozlowski 1996). Temperatures below 10 °C cause a halt to their growth. Several researchers have studied the role of the plant in water pollution control (Taylor et al. 2006; White et al. 2010). The plant was found to have a considerably high uptake of nitrates (Yan et al. 2009). Whereas 50–70% nitrate reduction was observed by planting Pontederia cordata in sub-surface constructed wetlands, 30–70% of total phosphorus removal was observed from planted storm water ponds (DeBust et al. 1995; Ou et al. 2006). Although proper choice of plants may achieve satisfactory nitrogen removal in CW, the CW configuration is also an important design consideration. Two-stage CW performed better than their single-stage counterparts (Langergraber et al. 2010). The poor performance in the latter is attributed to their inability to simultaneously provide both aerobic and anaerobic conditions (Vymazal 2007).

Based on the relatively higher water purification performance observed in two-stage constructed wetlands, a horizontal flow sub-surface constructed wetland mesocosm was considered for landscape water purification. The Green Origio biological system combines a set of physical, chemical and biological processes in the landscape purification. Its most important design features include solar-energy driven air pump, Pontederia cordata installed in substrate-free columns and a separate substrate column for enhanced adsorption process.

This study aimed to (a) evaluate the efficiency of a two-stage constructed wetland mesocosm in nitrogen and total organic carbon (TOC) removal from landscape water under various inflow rates; (b) evaluate the variation of dissolved oxygen (DO) concentration under varying flow conditions; (c) evaluate the relationship between varying inflow rates and the removal efficiencies for different parameters.

MATERIALS AND METHODS

Description of the Green Origio system

The Green Origio biological system is situated in Shanghai Jiao Tong University, (31°5′ N, 121°25′ E) in Minhang District, Shanghai, China. The two-stage solar-powered constructed wetland mesocosm combines a set of physical, chemical and biological processes, all which aim to make the system more effective and efficient. At the intake point (adjacent stationary stream) of the system is a submerged pump (GRUNDFOS UNILIFT KP 250, 50 Hz, 480 W, 2.2 A) that supplies the landscape water into the tanks. To prevent hydraulic overloading, an overflow pipe acts as a by-pass for excess flow. Water valves are used to control the water flow into the tanks. In addition, water flow meters are used to obtain the maximum flow discharge by the submerged pump and the flow rates through the overflow pipe. The difference in flow values represents the actual flow rates into the tanks. The design configuration of the system consists of two tanks that are connected in series as shown in Figure 1. The sub-surface tanks each having a depth and diameter of 2.2 m and 1.9 m, respectively, have varying DO concentrations. It is also important to note that each tank consists of both external and internal cylindrical tanks. Flow through the tanks initially assumes a vortex flow pattern through the external tank, followed by a vertical upward flow through the internal tank before the landscape water finally flows out of Tank 1 through the connecting pipe or out of Tank 2 through the outlet pipe. The first tank denoted as Tank 1 was operated under anaerobic conditions. On the other hand, a solar-driven air pump (SECOH IP 45, 40 l/min, 11.8 kPa, 220–240 V, 50 Hz, and 30–40 W) continuously supplied air to achieve aerobic degradation in Tank 2. In addition, the presence of fully grown self-growing Pontederia cordata (1.6–1.8 m) provided favorable conditions for nutrient assimilation. The computed plant area coverage was 2.504 m2. Compared to conventional constructed wetland systems where plants are supported by soil-filled columns, this mesocosm has substrate-free columns that allow unobstructed, deep lateral growth of the roots. However, a separate substrate column provided in the inner column of the tank enhances the landscape water purification by adsorption. Prior to the study, 35 kg and 17 kg of shale and biological ceramsite, respectively, were introduced into the substrate column. An outlet pipe into which the effluent flows out of the system is also provided. Figure 1 represents the various components of the system.

Figure 1

(a). (1) Inlet (influent from the adjacent stream flows into Tank 1) (2) The water sampling point for Tank 1 (3) Vortex flow in the external tank of Tank 1 (4) Pipe connecting Tank 1 and Tank 2 (5) Solar-Energy System (6) Air Pump (7) Air nozzle (8) Biological ceramsite and shale column (9) Fully grown Pontederia cordata (10) System's Outlet (11) Substrate-free columns for unobstructed, lateral root growth. (b) From top-left to bottom (clockwise): system's inlet (water sampling point for the influent), Tank 1, System's outlet (water sampling point for the effluent), Tank 2.

Figure 1

(a). (1) Inlet (influent from the adjacent stream flows into Tank 1) (2) The water sampling point for Tank 1 (3) Vortex flow in the external tank of Tank 1 (4) Pipe connecting Tank 1 and Tank 2 (5) Solar-Energy System (6) Air Pump (7) Air nozzle (8) Biological ceramsite and shale column (9) Fully grown Pontederia cordata (10) System's Outlet (11) Substrate-free columns for unobstructed, lateral root growth. (b) From top-left to bottom (clockwise): system's inlet (water sampling point for the influent), Tank 1, System's outlet (water sampling point for the effluent), Tank 2.

Collection of water samples

Water samples from the landscape water stream (inlet), Tank 1, Tank 2 and outlet sections of the system were collected and analyzed from August 2013 to October 2013. Grab water samples were collected twice (8 a.m. and 5 p.m.) daily during the experimental period and average daily values recorded. Samples from the stream (inlet) representing the system's influent were collected at the suction point of the submerged pump (about 20 cm below the water surface). On the other hand, samples collected from Tank 1 represented landscape water that had experienced sedimentation after a given HRT. The samples were collected from the internal tank of Tank 1. Samples collected from the internal tank of Tank 2 represented landscape water that had experienced complete purification by the Pontederia cordata but incomplete purification by the substrate media. Samples collected from the outlet pipe represented the system's effluent. Temperature variations observed during each sampling month were recorded. The mean air and water temperatures during the study period are summarized in Figure 2.

Figure 2

Mean air and water temperatures (error bars correspond to the 30-day standard deviation).

Figure 2

Mean air and water temperatures (error bars correspond to the 30-day standard deviation).

Determination of HRT

The HRT, also referred to as the detention time refers to the period of time spent by water in a tank (Davis & Masten 2004). It is inversely proportional to the incoming flow rate for any given system. A high inflow rate is matched by a low HRT and vice versa. In this study, the following mathematical expression was used to determine the HRT of the system. The actual flow into the system (Qin) was obtained by obtaining the difference between the regulated pump discharge values (Qp) and the flow rate through the overflow pipe. 
formula
1
where V1 = volume of tank 1, V2 = Volume of tank 2, = flow rate into the tanks.

Analyses of water samples

Table 2 summarizes the various analyses methods used to monitor the parameters' concentrations during the study. All laboratory experiments including the preparation of standard solutions and concentration monitoring of the parameters were carried out under room temperature. In addition, all standard solutions were prepared using ultra-pure water (18.2 MΩ/cm). Pre-filtered water samples (0.45 μm) were used for the analysis of TN, NO3-N, NO2-N, NH3-N and TOC concentrations. The pH values of the analyzed water samples ranged between 7.3 and 7.5 throughout the study period.

Table 2

Analyses Methods for different parameters

Parameter Analysis Method/Equipment Reference 
DO HACH Portable LDO meter  
pH HACH Portable pH meter  
TN/TOC Multi N/C 3100 analyzer (analytikjena, Germany)  
Nitrate-nitrogen (NO3-N) Ultraviolet Spectrophotometry (China Environmental Protection Industrial Standards 2007
Nitrite-nitrogen (NO2-N) N-(1- naphthyl) – ethylenediamine method (Ministry of Environmental Protection PRC 1993
Ammonia Nitrogen (NH3-N) Salicylic acid spectrophotometry (Ministry of Environmental Protection PRC 2009
Parameter Analysis Method/Equipment Reference 
DO HACH Portable LDO meter  
pH HACH Portable pH meter  
TN/TOC Multi N/C 3100 analyzer (analytikjena, Germany)  
Nitrate-nitrogen (NO3-N) Ultraviolet Spectrophotometry (China Environmental Protection Industrial Standards 2007
Nitrite-nitrogen (NO2-N) N-(1- naphthyl) – ethylenediamine method (Ministry of Environmental Protection PRC 1993
Ammonia Nitrogen (NH3-N) Salicylic acid spectrophotometry (Ministry of Environmental Protection PRC 2009

Statistical analysis

Linear regression analysis to evaluate the relationship between the removal efficiencies for different parameters and the flow rate conditions was carried out using Microsoft Excel, Analysis ToolPak-VBA. Daily mean and standard deviation concentration values were plotted using SIGMAPLOT 10.0.

RESULTS AND DISCUSSION

The system's average flow conditions

The various inflow rates were obtained directly from the water flow meters. The volume of the tanks and the inflow rates were used for the estimation of the system's HRT as represented by Equation (1). A summary of the flow duration, acclimation duration, pump discharge values (Qp), system's total volume, inflow rates and HRT is represented in Table 3. Whereas the flow duration in Table 2 represents the period of time corresponding to each flow rate condition, the acclimation duration represents the system's adjustment time between the preceding and the successive flow rate condition.

Table 3

The system's average flow conditions

  Flow Acclimation Qp Average flow conditions
 
Flow Duration (Days) Duration (Days) (m3/d) V1(m3V2(m3Qin(m3/d) HRT(Hrs) 
10 53 6.238 5.671 16.25 17.6 
10 74 6.238 5.671 16.25 9.69 
10 112 6.238 5.671 16.25 7.77 
10 120 6.238 5.671 16.25 5.65 
10 205 6.238 5.671 16.25 4.01 
10 270 6.238 5.671 16.25 3.55 
  Flow Acclimation Qp Average flow conditions
 
Flow Duration (Days) Duration (Days) (m3/d) V1(m3V2(m3Qin(m3/d) HRT(Hrs) 
10 53 6.238 5.671 16.25 17.6 
10 74 6.238 5.671 16.25 9.69 
10 112 6.238 5.671 16.25 7.77 
10 120 6.238 5.671 16.25 5.65 
10 205 6.238 5.671 16.25 4.01 
10 270 6.238 5.671 16.25 3.55 

Comparison of the removal efficiencies under varying flow rate conditions

Nitrogen removal

Figures 3456 summarize the removal rates for TN, nitrate (NO3-N), nitrite (NO2-N) and ammonia (NH3-N). The removal efficiencies represent the percentage concentration decrease between the influent and the effluent. Based on the results, highest removal rates were observed under least flow conditions (16.25 m3/day). The removal efficiencies for TN, NO3-N, NO2-N and NH3-N were 55.65%, 75.19%, 95.43% and 74.03%, respectively. When the system was subjected to this flow rate, the longest HRT (17.6 hours) was realized (Table 3). Consequently, there was adequate time for microbial degradation, adsorption onto substrate and plant assimilation. In addition, the temperatures were favorable for microbial degradation and plant uptake. Results obtained from previous studies showed higher nitrogen removal efficiencies under long HRT (Akratos & Tsihrintzis 2007; Chung et al. 2008). In addition, temperatures above 15 °C were favorable for optimal nitrogen removal efficiencies as was observed in another study (Kadlec & Reddy 2001). In our study, the progressive increase in flow rate contributed to the wash out of nitrifiers and denitrifiers hence the reduced removal efficiencies. In addition, the already accumulated nitrogen in the system was washed out as the flow rate was progressively increased. Although an acclimation period of 3 days was provided for each flow rate condition, the system's removal efficiency decreased as increasing deviation from the optimal flow conditions occurred. When maximum flow rate conditions were attained in October, least HRT conditions were noted as indicated in Table 3. TN and NO2-N removal efficiencies of 4.92% and 83.64%, respectively, were noted under this flow condition. On the other hand, concentration increase of 17.86% and 24.11% for NO3-N and NH3-N, respectively, was noted in the effluent under maximum flow conditions. This implies that wash out of accumulated NO3-N and NH3-N fractions occurred within the system.

Figure 3

TN Concentration and removal efficiencies under various inflow rates (error bars correspond to the 10-day standard deviation).

Figure 3

TN Concentration and removal efficiencies under various inflow rates (error bars correspond to the 10-day standard deviation).

Figure 4

Nitrate concentration and removal efficiencies under various inflow rates (error bars correspond to the 10-day standard deviation) (negative removal efficiencies correspond to concentration increase in the effluent).

Figure 4

Nitrate concentration and removal efficiencies under various inflow rates (error bars correspond to the 10-day standard deviation) (negative removal efficiencies correspond to concentration increase in the effluent).

Figure 5

Nitrite Concentration and removal efficiencies under various inflow rates (error bars correspond to the 10-day standard deviation).

Figure 5

Nitrite Concentration and removal efficiencies under various inflow rates (error bars correspond to the 10-day standard deviation).

Figure 6

Ammonia Concentration and removal efficiencies under various inflow rates (error bars correspond to the 10-day standard deviation) (negative removal efficiencies correspond to concentration increase in the effluent).

Figure 6

Ammonia Concentration and removal efficiencies under various inflow rates (error bars correspond to the 10-day standard deviation) (negative removal efficiencies correspond to concentration increase in the effluent).

Several processes were responsible for the removal efficiencies observed. In Tank 1, denitrification was the main biological process. Previous studies have shown that denitrification is the main nitrogen removal in constructed wetlands with warm temperatures leading to high nitrogen removal (Kadlec & Reddy 2001; Lin et al. 2002). Ammonia adsorption onto settling sediments may also have contributed to nitrogen removal in Tank 1 (Kadlec & Knight 1996). Increase in the flow conditions negatively affected these processes as the micro-organisms responsible for denitrification were washed out. In addition, there was inadequate time for ammonia adsorption in Tank 1. The relative increase in nitrogen concentration under high flow conditions in Tank 1 caused by inadequate HRT for denitrification process lowered the overall nitrogen removal efficiency of the system. The contribution of Tank 2 in nitrogen removal included plant assimilation, denitrification and ammonia adsorption onto the biological ceramsite and shale. Plant assimilation of NO3-N and NH3-N fractions by Pontederia cordata played an important role in the generation of new plant tissues (Vymazal 2007). In addition, the roots provided adequate attachment sites for microbial communities responsible for denitrification process. The nitrification and denitrification processes are summarized in Equations (2)‒(4). 
formula
2
 
formula
3
 
formula
4

TOC removal

The effect of increasing flow rates on the TOC removal efficiency was also investigated. The percentage removal efficiencies were calculated by comparing the TOC concentrations between the influent and effluent. TOC removal efficiencies and respective concentrations are summarized in Figure 7. A maximum removal efficiency of 29.87% was observed under a flow rate of 16.25 m3/day. Although, variation in TOC concentration was observed during the study period, decrease in the TOC removal efficiency was observed as the flow rate was increased progressively. The role of TOC as an energy source for nitrification and denitrification has been studied previously (Lin et al. 2002). Increase in flow rate was responsible for the wash out of the accumulated TOC. Consequently, when maximum flow conditions were maintained, concentration increase of 10.67% in the effluent was noted. TOC consumption in the denitrification process was the main process in Tank 1 while plant assimilation and microbial uptake of the organic compounds in Tank 2 contributed to reduced TOC removal in the effluent.

Figure 7

TOC Concentration and removal efficiencies under various inflow rates (error bars correspond to the 10-day standard deviation) (negative removal efficiencies correspond to concentration increase in the effluent).

Figure 7

TOC Concentration and removal efficiencies under various inflow rates (error bars correspond to the 10-day standard deviation) (negative removal efficiencies correspond to concentration increase in the effluent).

Effect of varying flow rates on DO

Figure 8 summarizes the DO concentrations under varying flows. Oxygen depletion led to relatively lower DO values in Tank 1. Increase in flow led to shorter HRT hence lower oxygen depletion. Whereas the highest DO depletion was noted under a flow of 16.25 m3/day, the least depletion rate was noted under a flow rate of 80.52 m3/day. Based on the influent concentrations, oxygen depletion rates of 49.72%, 42.71%, 44.3%, 28.21%, 35.02%, and 2.46% were noted under flow conditions of 16.25 m3/day, 29.47 m3/day, 36.79 m3/day, 50.57 m3/day, 71.38 m3/day and 80.52 m3/day, respectively. Increase in DO levels was observed in both Tank 2 and in the effluent. Based on Tank 1 concentrations, DO concentration increase rates of 45.36%, 46.31%, 48.46%, 29.94%, 32.75% and 6.52% were noted under flow rates of 16.25 m3/day, 29.47 m3/day, 36.79 m3/day, 50.57 m3/day, 71.38 m3/day and 80.52 m3/day, respectively. Relatively higher DO values were noted in the effluent. Root oxygen loss (ROL) has been cited as an important process that aids in nitrogen removal from constructed wetlands (Armstrong et al. 2000). It has particularly been shown that ROL has a significant impact on the redox potential, which plays an important role in determining the nitrogen fate (Bialowiec et al. 2012) and enhancing the microbial activity (Ueckert et al. 1990). In this study, the ROL was estimated by the equation below. 
formula
5
Figure 8

DO concentration and removal efficiencies under various inflow rates (error bars correspond to the 10-day standard deviation).

Figure 8

DO concentration and removal efficiencies under various inflow rates (error bars correspond to the 10-day standard deviation).

The calculated ROL values were 3.89%, 2.94%, 3.59%, 2.79% and 0.64% under flow conditions of 16.25 m3/day, 29.47 m3/day, 36.79 m3/day, 50.57 m3/day and 71.38 m3/day, respectively. A negative ROL value of −2.12% (decrease in concentration when Tank 2 and outlet DO concentrations were compared) was observed under a flow of 80.52 m3/day. Regression analysis revealed a strong correlation between ROL and flow rate conditions (R2=0.823: P < 0.01), with an increase in flow causing a decrease in ROL.

Relationship between the flow rate and parameters removal efficiencies

A linear regression analysis was done to ascertain the correlation between the flow rate conditions and the removal efficiencies of both nitrogen fractions and TOC. The regression coefficients (R2) are summarized in Table 4.

Table 4

Relationship between flow rate and parameters' removal efficiencies

Parameter R2 
TN 0.972 
Nitrate 0.959 
Nitrite 0.965 
Ammonia 0.942 
TOC 0.862 
Parameter R2 
TN 0.972 
Nitrate 0.959 
Nitrite 0.965 
Ammonia 0.942 
TOC 0.862 

Increase in flow had a negative effect on nitrogen and TOC removal. At high flows, the HRT is short as shown in Table 3. Higher HRT favored higher removal of all the parameters studied. This observation is consistent with previous studies that have concluded that increase in HRT is responsible for higher nitrogen removal (Akratos & Tsihrintzis 2007; Chung et al. 2008). Figure 9 summarizes the nitrogen and TOC removal efficiencies under varying flow rates.

Figure 9

Effect of varying flow rates on parameter removal efficiencies.

Figure 9

Effect of varying flow rates on parameter removal efficiencies.

CONCLUSION

The study conducted from August to October 2013 to investigate the efficiency of Pontederia cordata in nitrogen and TOC removal from landscape under varying flow rates revealed that an increase in flow had a negative effect on the system's removal efficiency due to wash out of the accumulated nitrogen and TOC concentrations. In addition, a decrease in ROL was observed with increasing flow rate. The two-stage constructed wetland mesocosm recorded highest removal efficiencies of 55.65%, 75.19%, 74.03%, 95.43% and 29.87% for TN, NH3-N, NO3-N, NO2-N and TOC, respectively, under a flow of 16.25 m3/day. Whereas removal efficiencies of 4.92% and 83.64% were observed for TN and NO2-N, respectively, under a flow of 80.52 m3/day, effluent concentrations increase of 21.44%, 17.86% and 10.67% were noted for NH3-N, NO3-N and TOC, respectively. Progressive increase in the system's flow rates was detrimental in realizing optimal removal efficiency. Such variations in flow progressively decreased the microbial residence time, the time for adequate plant assimilation and the adsorption time by the substrate media. Oxygen depletion and ROL were also affected by the changes in flow. The study confirms that increased flow rate or short HRT negatively affects the efficiency of Pontederia cordata in the removal of nitrogen from landscape water. Although high flow rates were associated with increased effluent volume, optimal flow rate conditions should be maintained in the system to realize optimal nitrogen and TOC removal efficiencies.

ACKNOWLEDGEMENTS

The publication is supported by Campus for Research Excellence and Technological Enterprise (CREATE) and the National Science and Technology Major Projects of Water Pollution Control and Management of China (2014ZX07206001). All contributions made in making the study successful are highly appreciated.

REFERENCES

REFERENCES
Audlbach-Smith
C. A.
deKozlowski
S. J.
1996
Aquatic and Wetland Plants of South Carolina
.
South Carolina Department of Natural Resources
. .
China Environmental Protection Industrial Standards
.
2007
Water quality- Determination of nitrate nitrogen-Ultraviolet spectrophotometry method (HJ /T 346-2007)
.
Davis
M.
Masten
S.
2004
Principles of Environmental Engineering and Science
.
McGraw Hill
,
New York
.
DeBust
T. A.
Peterson
J. E.
Jensen
K. R.
1995
Phosphorus removal from Agricultural run-off. An assessment of macrophyte periphyten-based treatment species
,
619–627
.
Kadlec
R. H.
Knight
R. L.
1996
Treatment Wetlands
.
CRC Press
,
Boca Raton
.
Kadlec
R. H.
Reddy
K. R.
2001
Temperature effects in treatment wetlands
.
Water Environ. Res.
73
(
5
),
543
557
.
Langergraber
G.
Pressl
A.
Leroch
K.
Rohrhofer
R.
Haberl
R.
2010
Comparison of single-stage and a two-stage vertical flow constructed wetland systems for different load scenarios
.
Water Sci. Technol.
64
(
5
),
1122
1129
.
Ministry of Environmental Protection, PRC
1993
Determination of nitrite in the wet precipitation-N-(1-naphthyl)-1, 2-diaminoethane dihydrochloride spectrophotometry. (GB/T13580.7-1992)
.
Ministry of Environmental Protection, PRC
2010
Water Quality—Determination of ammonia nitrogen—Salicylic acid spectrophotometry (HJ 536-2009)
.
Reddy
K. R.
Patrick
W. H.
1984
Nitrogen transformations and loss in flooded soils and sediments
.
CRC Crit. Rev. Environ. Control
13
,
273
309
.
Taylor
M. D.
White
S. A.
Chandler
S. L.
Whitwel
T.
Klaine
S. J.
2006
Nutrient management of nursery runoff water using constructed wetland systems
.
HortTechnology
16
,
548
549
.
Wetzel
R. G.
2001
Fundamental processes within natural and constructed wetland ecosystems: short-term versus long-term objectives
.
Water Sci Technol
44
(
11–12
),
1
8
.
White
S. A.
Taylor
M. D.
Chandler
S. L.
Whitwel
T.
Klaine
S. J.
2010
Remediation of nitrogen and phosphorus from nursery runoff during the spring via free water surface constructed wetlands
.
J. Environ. Hort.
28
(
4
),
209
217
.
Yan
C.
Bracy
R. P.
Owings
A. D.
2009
Nitrogen and phosphorous removal by ornamental and wetland plants in a greenhouse recirculation research system
.
Hortscience
44
(
6
),
1704
1711
.