The performance of a hybrid phytosystem in landscape water purification and herbicides removal was investigated. The phytosystem operating in an arboretum is located in the Minhang Campus of Shanghai Jiao Tong University, China. The phytosystem is composed of two purification stages: sedimentation Stage 1 without external air supply; and Stage 2 with an external air supply. Stage 2 is also vegetated with three major kinds of plants, namely Pontederia cordata L., Typha latifolia L. and Cyperus alternifolius L. The system's hydraulic loading rate (HLR) was maintained at 1.632 m/day between December 2013 and November 2014. Sedimentation, filtration and adsorption by filter media, combined microbial processes in the rhizosphere (nitrification–denitrification) and plant uptake of the pollutants were all responsible for water purification in the phytosystem. The biological and physical parameters analyzed were total dissolved nitrogen (TDN), nitrate (NO3-N), nitrite (NO2-N), ammonia (NH3-N), total dissolved phosphorus (TDP), dissolved organic carbon (DOC), turbidity, chlorophyll-a and algal cells number. Highest removal efficiencies for TDN, TDP, turbidity, DOC, chlorophyll-a and algal cells were 56.9%, 73.3%, 92.4%, 29.9%, 94.3% and 91.0%, respectively. When the phytosystem was considered for herbicides removal, removal efficiencies of more than 25% were noted for all the herbicides.

INTRODUCTION

The use of constructed wetlands is one of the most cost-effective and energy saving technologies in wastewater treatment (Kadlec & Knight 1996; Lee et al. 2009). Such systems are composed of plants, microorganisms and filter media. Cyperus alternifolius L., Scirpus validus L, Typha latifolia L. and Pontederia cordata L. are some of the most widely used plants (Kadlec & Knight 1996). This is primarily due to their ability to adapt to water logged conditions, their high resistance to high pollutant loadings, large biomass and their deep roots that increase the oxygen release at the rhizosphere (Kadlec & Knight 1996). Denitrification within the rhizosphere and plant uptake contributes to about 60–70% and 20–30% nitrogen removal, respectively (Lin et al. 2002; Lee et al. 2009). Conversely, adsorption onto filter media is mainly responsible for phosphorus removal (Vymazal 2007; Jiang et al. 2014). Shale and biological ceramsite, which are common filter media, have high metal contents that render them appropriate for chemical adsorption, precipitation and biological adsorption (Jiang et al. 2014). Temperature, pollution concentrations, type of media and plant type affect the nitrogen and phosphorus removal in horizontal sub-surface flow constructed wetlands (Kuschk et al. 2003; Vymazal 2007).The sustainability of such systems in phosphorus removal may be limited as the root biomass fills the pore spaces of the filter media (Breen & Chick 1995).

Recent studies investigating plant uptake in nutrients removal have shown that plants contribute to water purification in constructed wetlands (Kadlec & Reddy 2001; Lin et al. 2002; Stottmeister et al. 2003). Nitrogen in the form of ammonia and nitrate is assimilated by plants to enhance plant growth (Kadlec & Reddy 2001; Vymazal 2007). Similarly, phytosystems have shown satisfactory degradation rates for commonly detected herbicides such as atrazine, metolachlor and simazine (Dietz & Schnoor 2001; George et al. 2003). Microbial degradation in the rhizosphere and uptake by the plant roots are the major herbicides removal processes (Dietz & Schnoor 2001; Stottmeister et al. 2003). Designing phytosystems that optimize the contribution of sedimentation, filtration and adsorption by filter media, plant uptake and microbial degradation may help control the increasing concentrations of these agricultural chemicals in water bodies.

In this study, the role of a hybrid phytosystem in landscape water purification was investigated. The objectives of the study were: (1) to examine the seasonal variation in the removal of physical and biological parameters and their corresponding removal mechanisms; and (2) to investigate the role of the phytosystem and individual plants in the removal of organic herbicides.

MATERIALS AND METHODS

Description of the system

The hybrid phytosystem (Figure 1) is located in the arboretum of Shanghai Jiao Tong University (31°5′ N, 121°25′ E) in Minhang District, Shanghai, China. During its installation phase in February 2011, Pontederia cordata L., Typha latifolia L. and Cyperus alternifolius L. were planted in the phytosystem's second tank. The phytosystem has since experienced natural plant growth cycles (rapid and dormant growth in spring and winter respectively). It comprises two watertight cylindrical tanks (each having a depth and diameter of 2.2 m and 1.9 m, respectively, 2.6 m apart, 0.2 m difference in tank levels). Each tank has an inner cylindrical tank (depth and diameter of 2.2 m and 0.6 m, respectively). A vertical wall connecting the outer and inner tanks prevents counter-clockwise flow. A clockwise vortex flow in the outer tanks is succeeded by upward flow in the inner tanks as water flows through 0.2 m-high, 50 mm-diameter gates. Stage 1 is represented by a closed sedimentation tank with no external aeration, while Stage 2 is vegetated (with a coverage area of 2.504 m2) and provided with external aeration. In order to provide clean energy to run the air pump (SECOH IP 45, 40 L/min, 11.8 kPa, 220–240 V, 50 Hz, and 30–40 W), a solar energy system composed of two solar panels (0.8 × 0.9 m, 100 Wp, 5.72 A, 17.5 V) and a solar-charge controller (12/24 V, 50 A) were installed in April 2013. The continuous aeration protects the roots against decay during winter. The phytosystem is also characterized by tubular blocks for plant support and deep root growth, the integration and optimization of microbial processes within the rhizosphere, and a filter media column (0.7 m deep) where fresh media can be introduced with minimal interference to the plants. At the start of autumn (Sept–Nov), the performance of the phytosystem was enhanced by introducing shale (35 kg, ø10–25 mm) and biological ceramsite (17 kg, ø3–5 mm) into the column. Furthermore, a submerged pump (UNILIFT KP 250, 50 Hz, 480 W, 2.2 A) was installed in the river to create a pressure head for the landscape water flowing into the phytosystem.
Figure 1

Inflow controlled by water valve. Manhole for sample collection in Stage 1. Tubular blocks that increase the surface area for sedimentation. 2.6 m-long polyvinyl chloride pipe (diameter of 50 mm). Solar energy system to run the air pump. Solar-powered air pump. Maximum aeration zone. Filter media (biological ceramsite and shale) column. Plants installed in tubular blocks. Outflow. Tubular blocks for deep root growth.

Figure 1

Inflow controlled by water valve. Manhole for sample collection in Stage 1. Tubular blocks that increase the surface area for sedimentation. 2.6 m-long polyvinyl chloride pipe (diameter of 50 mm). Solar energy system to run the air pump. Solar-powered air pump. Maximum aeration zone. Filter media (biological ceramsite and shale) column. Plants installed in tubular blocks. Outflow. Tubular blocks for deep root growth.

Water sampling

Water samples were collected from the river (inflow), Stage 1, Stage 2 and outflow sections (Figure 2). All samples were collected 5 cm below the water surface at all the sampling points throughout the study period. During preliminary tests to obtain the system's optimal hydraulic loading rate (HLR), 60 samples were collected for 60 days between April and June 2013. Each 20-day sampling period corresponded to a different flow rate. During the actual experiments in winter (Dec–Feb), spring (Mar–May), summer (Jun–Aug) and autumn (Sept–Nov), 120 samples (collected at 8:00 am and 5:00 pm for 60 days in each season) were used to obtain the mean seasonal concentrations. The corresponding mean seasonal temperatures were 6.5 °C, 16.5 °C, 27.5 °C and 22.5 °C, respectively.
Figure 2

The aerial view of the arboretum. Sampling point for the inflow. Submerged water pump pumps landscape water to the phytosystem through a sub-surface pipe. The bank of the river covered by grass, trees and litter. Landscape water enters the phytosystem. Sampling point for Stage 1. Sampling point for Stage 2. Sampling point for outflow. Solar-energy system.

Figure 2

The aerial view of the arboretum. Sampling point for the inflow. Submerged water pump pumps landscape water to the phytosystem through a sub-surface pipe. The bank of the river covered by grass, trees and litter. Landscape water enters the phytosystem. Sampling point for Stage 1. Sampling point for Stage 2. Sampling point for outflow. Solar-energy system.

Determination of optimal hydraulic loading rate

Water control valves and water flow meters were used to regulate and monitor the flow rates. The daily mean flow rates were determined by obtaining the difference between the preceding and successive flow values. The HLR and the corresponding removal rates were estimated by Equations (1) and (2). Table 1 summarizes the flow rates, surface areas and the HLR. 
formula
1
 
formula
2
where QIN is the system's flow rate, and A is the surface area of each tank.
Table 1

Flow rates and HLR conditions

QIN (m3/d) ATANK 1 (m2ATANK 2 (m2HLR (m/d) 
9.6 17.647 17.647 1.088 
14.4 17.647 17.647 1.632 
24.0 17.647 17.647 2.720 
QIN (m3/d) ATANK 1 (m2ATANK 2 (m2HLR (m/d) 
9.6 17.647 17.647 1.088 
14.4 17.647 17.647 1.632 
24.0 17.647 17.647 2.720 

Analyses for biological and physical parameters

The Multi N/C 3100 Analyzer (analytikjena, Germany) was used to analyze both total dissolved nitrogen (TDN) and dissolved organic carbon (DOC) while PHYTO-PAM plankton analyzer (Water-PAM, Walz, Germany) was used for chlorophyll-a analysis. Whereas turbidity was monitored using a portable 2100Q HACH turbidity meter, the algal cells were counted by a HACH Hydrolab Multi-Parameter Probe. The ascorbic-acid method, phenol disulphonic acid spectrophotometry, N-(1-naphthyl)-ethylenediamine and the salicylic acid spectrophotometry methods were used for the analysis of total dissolved phosphorus (TDP), NO3-N, NO2-N and NH3-N respectively (APHA 1992; Ministry of Environmental Protection PRC 1993, 2010; China Environmental Protection Industrial Standards 2007). Pre-filtered water samples (0.45 μm) were used for the analysis of TDP, TDN, NO3-N, NO2-N, NH3-N and DOC. TDP was determined after 30 min of autoclave-mediated digestion (120 °C, 100 kPa, with K2S2O8 and H2SO4) (Jiang et al. 2014). Ultrapure water (18.2 MΩ/cm) used to make standard solutions was produced by a water filtration system (Hitachi, Japan). Equation (3) was used to estimate the parameters' removal efficiencies. 
formula
3

Analyses for organic herbicides

Investigations were conducted to determine the role of the phytosystem in the removal of herbicides. Laboratory investigations were also carried out to determine the role of selected plants in the removal of atrazine under controlled conditions. Herbicides were extracted from water samples by solid phase extraction (SPE) process and concentration of herbicides determined by liquid chromotography mass spectrometry (LC-MS).

C18-cartridges were first pre-conditioned by passing 5 mL of water and 5 mL of methanol. Pre-filtered samples (0.45 μm) were then passed through the cartridges at a loading rate of 5 mL/min (Rodriguez-Mozaz et al. 2004). The SPE kit was connected to a vacuum pump. Elution was conducted by passing 8 mL of methanol in two phases (4 mL in each phase after 5 minutes interval). The eluate was then concentrated by nitrogen gas evaporation and reduced to a volume of 200 μL. Sodium sulphate was then used to remove any water in the eluate.

The LC analysis was conducted by 1,260 Infinity (Agilent, USA) where a ZORBAX Eclipse plus C18, Analytical 4.6 × 250 mm. 5-micron (Agilent, USA) column was installed. Whereas the flow rate was 1 mL/min, the inject volume was 20 μL. The column temperature was maintained at 30 °C. The MS analysis was conducted by the 6120 Quadrupole LC/MS (Agilent, USA). The analysis conditions included: positive polarity mode, fragmentor voltage of 70 eV, gas temperature of 350 °C, drying gas rate of 12 L/min and capillary voltage of 3,000 V (Rodriguez-Mozaz et al. 2004). Table 2 summarizes the correlation coefficients and the recovery ratios of the herbicides analyzed.

Table 2

Herbicides analyses parameters

Herbicide Recovery (%) Correlation (R2
Atrazine 65.19 0.992 
Simazine 100.02 0.999 
Prometryn 84.59 0.999 
Metolachlor 69.23 0.998 
Herbicide Recovery (%) Correlation (R2
Atrazine 65.19 0.992 
Simazine 100.02 0.999 
Prometryn 84.59 0.999 
Metolachlor 69.23 0.998 

RESULTS AND DISCUSSION

Optimal system's HLR conditions

Maximum TDP and NH3-N removal rates were noted under an optimal HLR of 1.632 m/day (Figure 3). Under high HLR conditions, a short hydraulic retention time (HRT) was realized (Jiang et al. 2014). In our study, although highest HRT was noted at a HLR of 1.088 m/day, relatively lower removal rates were observed. The reduced phosphorus removal may be attributed to physical/chemical clogging or loss of binding capacity by the filter media (Liira et al. 2009; Vohla et al. 2011). The long HRT also favored NH3-N release by the sediments into the overlying water (Kadlec & Knight 1996). Increase of HRT was found to increase the TDP and NH3-N removal in previous studies (Huang et al. 2000; Jiang et al. 2014). It can be noted that in the HLR range of 1.088–1.632 m/day, the TDP and NH3-N removal rates increased as the HRT decreased (Figure 3). However, when the HLR exceeded 1.632 m/day, the removal rates began to decline. This may be a result of wash out of microorganisms by the flowing water and inadequate contact time for TDP and NH3-N removal (Huang et al. 2000; Jiang et al. 2014).
Figure 3

The effect of varying HLR on TDP and NH3-N removal rates.

Figure 3

The effect of varying HLR on TDP and NH3-N removal rates.

Seasonal concentrations and removal

Nitrogen

TDN is the combination of all the inorganic fractions (NO3-N, NO2-N and NH3-N) and organic nitrogen (Lee et al. 2009). Highest TDN and NO3-N inflow concentrations (3.09 mg/L and 2.4 mg/L, respectively) were observed during winter (Dec–Feb) period (Figures 4(a) and (b)). The corresponding concentrations finally decreased to 1.45 mg/L and 1.02 mg/L, respectively in autumn (Sept–Nov). This was probably due to the relatively high temperatures that enhanced nitrogen utilization by the algal matter. Low temperatures observed during the winter (Dec–Feb) period increased the dissolved oxygen (DO) concentrations (Wetzel 2001) leading to increase in NO3-N concentration (Vymazal 2007). However, highest NH3-N concentrations (0.34 mg/L) noted during this period was attributed to increased ammonification, mineralization of ammonia from decay products or release from sediments (Kadlec & Knight 1996). Whereas highest inflow NO2-N concentrations (0.11 mg/L) observed during autumn (Sept–Nov) were probably due to increasing NH3-N concentrations (Figures 4(c) and (d)), the least NO2-N and NH3-N concentrations (0.03 mg/L and 0.17 mg/L, respectively) observed during summer (Jun–Aug) were as a result of increased microbial activities.
Figure 4

Mean seasonal concentrations and removal rates for: (a) total dissolved nitrogen; (b) nitrate; (c) nitrite; and (d) ammonia. The error bars correspond to 3-month standard deviation.

Figure 4

Mean seasonal concentrations and removal rates for: (a) total dissolved nitrogen; (b) nitrate; (c) nitrite; and (d) ammonia. The error bars correspond to 3-month standard deviation.

Highest TDN and NH3-N removal efficiencies (56.9% and 78.8% respectively) were observed during summer (Jun–Aug). Highest NO3-N and NO2-N removal efficiencies (73.1% and 96.2%, respectively) were observed during autumn and spring, respectively. The high removal efficiencies were as a result of high temperatures (16.5–27.5 °C) that enhanced the denitrification process and the plant uptake capacities (Kadlec & Reddy 2001; Kuschk et al. 2003). Low temperatures during winter (Dec–Aug) led to low TDN and NO2-N removal efficiencies (6.46% and 49.6%, respectively). However, low NO3-N removal efficiencies observed during spring may be attributed to excessive nitrification (Lee et al. 2009). When the combined seasonal NO3-N, NO2-N, NH3-N removal efficiencies were considered, outflow TDN concentrations (<1.0 mg/L) were observed during autumn (Sept–Nov) and summer (Jun–Aug). Based on these results, the outflow conformed to Grade III standards (An & Wang 2014) as shown in Appendix 1 (available in the online version of this paper).

The contribution of Stage 1 and Stage 2 in the removal of nitrogen fractions was investigated. In Stage 1, the relatively high temperatures during summer (Jun–Aug) and autumn (Sept–Nov) enhanced the removal of NO3-N, NO2-N through denitrification. Although NO3-N concentration increase was observed during spring (Mar–May), TDN removal observed in Stage 1 was probably due to ammonification of organic nitrogen followed by denitrification (Vymazal 2007; Lee et al. 2009). When Stage 2 was considered, plant uptake, denitrification, ammonia volatilization and ammonia adsorption onto filter media contributed to TDN removal (Vymazal 2007). The relatively high temperatures during spring, summer and autumn favored plant growth (Figure 7(a)) as well as enhanced the microbial activities (Kuschk et al. 2003). The plants absorbed NO3-N and NH3-N and converted them to organic compounds (Vymazal 2007). Elevation of temperature from 10–25 °C increased the nitrogen use efficiency of the Typha species from 5% to 38% (Kadlec & Reddy 2001).

Dissolved organic carbon

DOC inflow concentration remained low during the December–May period (8.2–8.5 mg/L) but increased during the June–August period (11.0 mg/L) as shown in Figure 5(a). Increased temperature and sunlight intensity enhanced algal photosynthesis leading to an increased release of organic compounds (Jing et al. 2001). Increase in temperatures was also responsible for the highest DOC removal efficiency (29.9%) observed during summer. The high DOC removal during this period was attributed to the enhanced denitrification process responsible for the increased TDN removal during the same period (Figure 4(a)).
Figure 5

Mean seasonal concentrations and removal efficiencies for: (a) dissolved organic carbon; (b) total dissolved phosphorus; and (c) turbidity.

Figure 5

Mean seasonal concentrations and removal efficiencies for: (a) dissolved organic carbon; (b) total dissolved phosphorus; and (c) turbidity.

DOC removal in Stage 1 was mainly through the biodegradation of organic compounds, providing energy in the nitrification and denitrification processes (Lee et al. 2009). The highest DOC removal was observed during spring (Mar–May). Although the denitrification efficiency was low as is shown by Figure 4(b), the conversion of DOC to CO2 may have occurred to support the nitrification process (Lee et al. 2009). In Stage 2, higher DO (Figure 7(b)) that led to increased oxidation of organic compounds into CO2, degradation of organic substances by the microorganisms in the rhizosphere, plant uptake of organic substances all contributed to DOC removal (Lin et al. 2002; Lee et al. 2009).

Total dissolved phosphorus

Highest TDP inflow concentrations (0.59 mg/L) observed during winter (Dec–Feb) may be attributed to the low TDP uptake by the algae due to inadequate sunlight intensity (Jing et al. 2001). Adequate sunlight intensity and temperatures during the summer (Jun–Aug) contributed to the lowest concentration (0.29 mg/L) as phosphorus was consumed by the algae (Figure 5(b)). Although high temperatures were observed during summer (Jun–Aug), the lowest TDP removal efficiencies (17.1%) were observed. This may be attributed to the loss of binding capacity by the filter media (Liira et al. 2009) after a continuous TDP loading. When fresh media was introduced during autumn (Sept–Nov), highest TDP removal efficiencies (73.3%) were noted. When the inflow (0.3–0.59 mg/L) and outflow (0.08–0.3 mg/L) TDP concentrations were compared, the phytosystem improved the landscape water quality from Grade V to Grade III (Appendix 1).

TDP removal in Stage 1 was by adsorption onto the settling sediments (Liira et al. 2009) while that in Stage 2 was by adsorption onto filter media and plant uptake (Vymazal 2007). Based on Figure 5(b), despite the low temperatures, highest TDP removal in Stage 1 was observed during winter (Dec–Feb). Low temperatures had little negative effect on phosphorus removal as observed previously (Kadlec & Reddy 2001). Increased temperatures during summer (Jun–Aug) and autumn (Sept–Nov) converted Stage 1 to a phosphorus source rather than a sink (Jin et al. 2013) hence the increase in TDP concentration. In Stage 2, highest removal efficiencies observed during autumn (Sept–Nov) are attributed to the adequate concentrations of calcium (Ca), iron (Fe) or aluminum (Al) in the freshly introduced shale and biological ceramsite (Jiang et al. 2014). The role of Stage 2 as a phosphorus source was however observed during winter (Dec–Nov) and spring (Mar–May) as is shown in Figure 5(b). Physical/chemical clogging or loss of binding capacity by the filter media may have occurred (Liira et al. 2009; Vohla et al. 2011).

Turbidity

Turbidity was caused by suspended matter of clay, silt or organic matter (Jin et al. 2013). Based on Figure 5(c), whereas the highest turbidity concentration (44.3 NTU) was observed in autumn (Sept–Nov), the lowest (18.7 NTU) was observed during spring (Mar–May). The variations may have been due to changes in the silt content or decayed matter in the river. Highest and least turbidity removal efficiencies (92.4% and 83.3%, respectively) were observed during autumn and spring, respectively.

Turbidity removal was primarily by sedimentation (Jin et al. 2013) in Stage 1 (Figure 5(c)) with the highest removal efficiency being observed during summer (Jun–Aug). The contribution of Stage 2 in turbidity removal was lower compared to that of Stage 1. This may be attributed to the decreasing background concentration in Stage 1 compared to that of the inflow. The relatively high turbidity removal observed in Stage 2 during autumn (Sept–Nov) was attributed to the optimum plant growth (Figure 7(a)) and fresh filter media that provided adequate conditions for filtration by the stems and roots (Vymazal 2007).

Chlorophyll-a and algal cells number

Whereas chlorophyll was used as an indicator for algal growth, the algal cells number represented the population of the algal cells. Based on Figure 6(a), highest chlorophyll-a concentration (40.9 μg/L) was observed in spring (Mar–May) while the lowest (20.9 μg/L) was observed during autumn (Sept–Oct). High inflow chlorophyll-a concentration was also observed between April and June in a previous study (Jing et al. 2001). The high number of algal cells noted in summer (Jun–Aug) reveals that temperature influenced algal growth due to increased algal photosynthesis (Figure 6(b)). The highest chlorophyll-a removal efficiency (94.3%) was observed during spring (Mar–May) while the least (69.6%) was during winter (Dec–Nov). Increased temperatures led to increased competition for nutrients between the plants and algae cells hence high chlorophyll-a removal during spring (Jing et al. 2001; Wetzel 2001). Comparison analysis of the inflow and outflow chlorophyll-a concentrations (Figure 6(a)) revealed significant landscape water quality improvement (<10 μg/L) based on the eutrophication standards (Appendix 2, available in the online version of this paper).
Figure 6

Mean seasonal concentrations and removal efficiencies for: (a) chlorophyll-a; and (b) algal cells.

Figure 6

Mean seasonal concentrations and removal efficiencies for: (a) chlorophyll-a; and (b) algal cells.

Light attenuation contributed to chlorophyll-a and algal cells removal in Stage 1 (Jing et al. 2001). Conversely, competition for nutrients between the plants and the algae enhanced algal cells removal in Stage 2. Based on Figure 7(a), an exponential growth in the height of the plants is observed implying that as more nutrients were absorbed by plants, a condition of limited nutrients was created hence the reduced algal growth (Wetzel 2001). Similarly, relatively higher temperatures during the warm periods (Mar–Nov) led to increased removal of chlorophyll and algal cells as both the microbial activity rates and oxygen diffusion rates increased favoring the nutrients depletion in both stages (Lee et al. 2009).
Figure 7

(a) The monthly growth characteristics of the plants (dormant growth during the January–February period, rapid growth during the March–August period, optimum growth during the September–December period). (b) Mean seasonal dissolved oxygen in Stage 2.

Figure 7

(a) The monthly growth characteristics of the plants (dormant growth during the January–February period, rapid growth during the March–August period, optimum growth during the September–December period). (b) Mean seasonal dissolved oxygen in Stage 2.

Role of phytosystem in herbicides removal

The efficiency of the phytosystem in the removal atrazine, simazine, prometryn and metolachlor under low temperatures was investigated during winter (Dec–Feb). Based on Figure 8(a), overall removal efficiencies of 29.5%, 32.4%, 33.1%, 26.9% were noted for atrazine, simazine, metolachlor and prometryn, respectively. Further batch experiments were conducted under controlled laboratory conditions to investigate the performance of Typha latifolia L., Scirpus validus L. and Cyperus alternifolius L. in the removal of atrazine under hydroponic conditions (Figure 8(b)). The solution (initial concentration of 10 μg/L) was introduced in 1 L-plant columns (triplicate sets for each plant type). The water samples were extracted and analyzed after a HRT of 3 days. The columns were coated with a thin sheet of black polythene to minimize the effect of photodegradation. Continuous air supply was maintained to obtain a DO concentration of 7.5–8.5 mg/L. Micro-climatic conditions included light intensity of 6,000 lux, air temperatures of 20–25 °C, humidity of 55–65% and 12 h light/dark conditions. The results revealed atrazine removal efficiencies of 85.1%, 84.2% and 83.8% for Cyperus alternifolius L., Typha latifolia L. and Scirpus validus L. respectively (Figure 8(b)). The higher herbicide removal efficiencies under controlled laboratory conditions were due to relatively warmer temperatures and higher HRT providing adequate time for plant uptake. This proved that plant uptake contributed to herbicide removal.
Figure 8

(a) Mean concentration and removal rates for atrazine, simazine, metolachlor and prometryn by the hybrid phytosystem. (b) The mean removal efficiencies for atrazine by Cyperus alternifolius L., Typha latifolia L., and Scirpus validus L., under controlled laboratory conditions.

Figure 8

(a) Mean concentration and removal rates for atrazine, simazine, metolachlor and prometryn by the hybrid phytosystem. (b) The mean removal efficiencies for atrazine by Cyperus alternifolius L., Typha latifolia L., and Scirpus validus L., under controlled laboratory conditions.

Mechanisms for herbicide removal

Phytodegradation, biodegradation, adsorption onto sediments, volatilization, hydrolysis and photodegradation are dependent on the physicochemical properties (Table 3) of the herbicides (Navarro et al. 2004; Rice et al. 2004). The major herbicide removal mechanisms in Stage 1 were biodegradation and adsorption onto sediments. In a previous study, atrazine and metolachlor were detected in the sediments. The sediments significantly reduced the concentrations of these herbicides in the surface water by enhancing the bioavailability of the pesticides for microbial degradation (Rice et al. 2004). However, in Stage 2, plant uptake and adsorption on the filter media played a great role in the removal of these herbicides. Based on the physicochemical properties shown in Table 3, all the herbicides analyzed were of moderate hydrophobicity (log Kow = 0.5–3).

Table 3

The physicochemical properties of common herbicides

Herbicide Solubility in water (mg/L) Log Kow Molecular mass (g) 
Atrazine 33 2.50 215.69 
Simazine 6.2 2.18 201.66 
Prometryn 33 3.10 241.37 
Metolachlor 530 3.13 283.46 
Herbicide Solubility in water (mg/L) Log Kow Molecular mass (g) 
Atrazine 33 2.50 215.69 
Simazine 6.2 2.18 201.66 
Prometryn 33 3.10 241.37 
Metolachlor 530 3.13 283.46 

Log Kow-Octanol water partition.

Direct plant uptake played a significant role in the degradation of the moderately hydrophobic herbicides. Sufficient sorption onto the roots followed by translocation within the plants occurred (Dietz & Schnoor 2001). The large and dense root systems with high levels of degrading enzymes enhanced important physiological and biochemical process within the rhizosphere for the degradation of the herbicides (Stottmeister et al. 2003). Adsorption onto the filter media may also have occurred in Stage 2 based on studies that revealed the contribution of adsorption in herbicides removal (Navarro et al. 2004; Rice et al. 2004).

CONCLUSION

Based on the seasonal performance of a hybrid phytosystem in landscape water purification, high removal efficiencies for total dissolved nitrogen, nitrate, nitrite, ammonia, total dissolved phosphorus, dissolved organic carbon, turbidity, chlorophyll-a and algal cells were observed. In addition, the removal of atrazine, simazine, prometryn and metolachlor was also noted. Relatively high system performance was noted during summer and autumn as microbial activities and plant uptake were temperature dependent. The small scale and large scale of similar phytosystems may be considered a sustainable technology in the ecological restoration of landscape water. This may help to reduce the level of eutrophication and herbicides contamination in such waters.

ACKNOWLEDGEMENTS

The study was funded 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 the authors are also appreciated for making this study a success.

REFERENCES

REFERENCES
An
S.
2014
Wetland Restoration: Shanghai Dalian Lake Project
.
Springer
,
Berlin
,
162
164
.
APHA
1992
Standard Methods for the Examination of Water and Wastewater
,
18th edn
.
American Public Health Association, American Water Works Association, Water Environment Federation
,
Washington DC, USA
.
China Environmental Protection Industrial Standards
2007
Water quality—Determination of nitrate nitrogen—Ultraviolet spectrophotometry method (HJ/T 346-2007)
.
Dietz
A. C.
Schnoor
J. L.
2001
Advances in phytoremediation
.
Environmental Health Perspectives
109
(
1
),
163
.
George
D.
Stearman
G. K.
Carlson
K.
Lansford
S.
2003
Simazine and metolachlor removal by subsurface flow constructed wetlands
.
Water Environment Research
7
(
2
),
101
112
.
Jing
S. R.
Lin
Y. F.
Lee
D. Y.
Wang
T. W.
2001
Using constructed wetland systems to remove solids from highly polluted river water
.
Water Science and Technology: Water Supply
1
(
1
),
89
95
.
Kadlec
R. H.
Knight
R. L.
1996
Treatment Wetlands
.
CRC Press
,
Boca Raton
,
893
.
Kadlec
R. H.
Reddy
K. R.
2001
Temperature effects in treatment wetlands
.
Water Environmental Research
73
(
5
),
543
557
.
Kuschk
P.
Wiessner
A.
Kappelmeyer
U.
Weissbrodt
E.
Kastner
M.
Stottmeister
U.
2003
Annual cycle of nitrogen removal by a pilot-scale subsurface horizontal flow in a constructed wetland under moderate climate
.
Water Research
37
(
17
),
4236
4242
.
Lee
C. G.
Fletcher
T. D.
Sun
G.
2009
Nitrogen removal in constructed wetland systems
.
Engineering in Life Sciences
9
(
1
),
11
22
.
Liira
M.
Kõiv
M.
Mander
Ü.
Mõtlep
R.
Vohla
C.
Kirsimäe
K.
2009
Active filtration of phosphorus on Ca-rich hydrated oil shale ash: Does longer retention time improve the process?
.
Environmental Science & Technology
43
(
10
),
3809
3814
.
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)
.
Rice
P. J.
Anderson
T. A.
Coats
J. R.
2004
Effect of sediment on the fate of metolachlor and atrazine in surface water
.
Environmental Toxicology and Chemistry
3
(
5
),
1145
1155
.
Stottmeister
U.
Wießner
A.
Kuschk
P.
Kappelmeyer
U.
Kästner
M.
Bederski
O.
Muller
R.
Moormann
H.
2003
Effects of plants and microorganisms in constructed wetlands for wastewater treatment
.
Biotechnology Advances
22
(
1
),
93
117
.
Vohla
C.
Kõiv
M.
Bavor
H. J.
Chazarenc
F.
Mander
U.
2011
Filter materials for phosphorus removal from wastewater in treatment wetlands – a review
.
Ecological. Engineering
37
(
1
),
70
89
.
Vymazal
J.
2007
Removal of nutrients in various types of constructed wetlands
.
Science of the Total Environment
380
(
1–3
),
48
65
.
Wetzel
R. G.
2001
Limnology: lake and river ecosystems
.
Gulf Professional Publishing
,
Academic Press, San Diego, CA, USA
,
159
167
,
1006
.

Supplementary data