Effects of Plant Species on Methane and Nitrous Oxide Emissions from Constructed Wetlands Treating Municipal Wastewater

This study was conducted to quantify emissions of greenhouse gases (GHGs), methane (CH4) and Nitrous Oxide (N2O), from free water surface constructed wetlands used for domestic wastewater treatment. All constructed wetlands were monoculture and each plot was planted with Phragmites sp., Cyperus sp., or Canna sp. The average CH4 and N2O emissions were in the range of 5.9-11.2 and 0.9-1.8 g/m 2 /h, respectively. Seasonal fluctuations of CH4 and N2O emissions were observed. The highest fluxes of both GHGs occurred during hot rainy season (JulyOctober) followed by summer and the lowest found in cool season. The mean of CH4 and N2O emissions from different plants species were significantly different (p<0.05). Average CH4 emissions from constructed wetlands planted with Phragmites sp., Cyperus sp. and Canna sp. were 11.2, 6.0 and 5.9 mg/m 2 /h, respectively, while mean N2O emissions were 0.9, 1.0 and 1.8 mg/m 2 /h, respectively. Calculated of Global Warming Potential (GWP) found that GWP of CH4 and N2O flux from constructed wetlands planted with Cyperus sp., was the highest (669 mg CO2 equivalent/m 2 /h), followed by Phragmite sp., (524 mg CO2 equivalent/m 2 /h) and Canna sp., (434 mg CO2 equivalent/m 2 /h), respectively. These results suggested that municipal wastewater treatment by constructed wetlands planted with Canna sp. and Phragmite sp., had potential of lower GHGs emissions into the atmosphere and Phragmite sp., provided the highest removal rate of Biochemical Oxygen Demand (BOD) and Chemical Oxygen Demand (COD).


INTRODUCTION
Constructed Wetlands (CWs) systems are combinations of natural wetlands and conventional wastewater treatment processes and are constructed in order to reduce input of nutrients and organic pollutants to water bodies.Constructed wetland systems, a costeffective alternative, apply various technological designs, using natural wetland processes, associated with wetland hydrology, soils, microbes and plants (Kadlec and Wallace, 2009;Casselles-Osorio et al., 2011).Tropical countries, like Thailand, could have a great benefit of wastewater treatment by constructed wetlands.When constructed wetlands are used for purification of wastewater, the productivity of the ecosystem could increase as well as the production of greenhouse gases (GHGs), which are by-or endproducts of microbial decomposition processes (Inamori et al., 2007;Wang et al., 2008;Liu et al., 2009;Wu et al., 2009).Constructed wetlands, therefore, can be sources of important greenhouse gases.
In constructed wetland microcosm, soil-plant is a highly complex environmental system that acts as a reservoir for microorganisms with their activity varying over space and time.Plants release root exudation, which easily decomposable and preferentially used by microorganisms, increased carbon input into the system (Tanner, 2001).Microbial growth in wetlands soil has been believed to depend upon the plant species and substrate.Furthermore, many species of emergent plants CWs possess a convective flow mechanism when oxygen is transported to the roots and gaseous microbial by-products are emitted from plant roots to the atmosphere (Brix, 1989;Brix et al., 1996).The transport of gases by the convective mechanism is faster than diffusion through water.The presence of plants in constructed wetland system may increase gas emissions from the soil.Therefore, plant species could affect microbial ecology and gas emission from treatment processes.
Although constructed wetlands can be beneficial for wastewater treatment they may have an unfavorable environmental impact by increasing the fluxes of greenhouse gases to the atmosphere.Thus, objectives of this study were:

Study site:
The experimental scale of constructed wetlands was built outdoor at Suranaree University of Technology, Nakhon Ratchasima Province in northeastern Thailand (Fig. 1).The geo-location of the study area is 177879N and 1648339E.The region has three-season monsoonal climate, with a relatively cool dry season from November to late February, followed by a hot dry season from March to June and then a hot rainy season from July to October (TMD, 2012).
The experiment was employed on a regime of free water surface flow constructed wetland.Eight constructed wetlands with identical dimensions were built based primarily on criteria of Aspect Ratio (AR) or length to width of 4:1 to minimize short circuiting and force the flow to move closely to plug flow hydraulic regime (U.S. EPA, 2000).Each wetland had the dimension of 2.0×0.5×0.8 m (length×width×depth). Brick, cement and mortar were the materials used for the construction of the constructed wetlands.Permanent transparent roof made from clear plastic was also constructed to prevent rain getting into the experiment setup and allow direct sun light exposure.Three emergent plants were used.All constructed wetlands were monoculture with Phragmites sp., Cyperus sp. and Canna sp.Each had two replicate cells.
The compositions of synthetic domestic wastewater consisted of glucose, FeCl 3 , NaHCO 3 , KH 2 PO 4 , MgSO 4 •7H 2 O and urea (Sirianuntapiboon and Tondee, 2000), similar to the domestic wastewater from Thailand's Housing Estates.The concentration of each component is described below: Glucose 190 mg/L FeCl 3 0.31 mg/L NaHCO 3 6.7 mg/L KH 2 PO 4 6.0 mg/L MgSO 4 •7H 2 O 3.9 mg/L Urea 9.0 mg/L Synthetic domestic wastewater was fed daily into the constructed wetlands using gravimetric flow and the flow was controlled by needle valves.The characteristics of the prepared synthetic wastewater are shown in Table 1.Gas flux measurement and analysis: Gases emissions from the constructed wetlands were sampled monthly from June 2010 to May 2011.The measurements of GHGs fluxes were conducted between 09:00 and 15:00 periods.Gas emissions were measured using a static chamber method described by Hutchinson and Mosier (1981).The chambers consist of two parts.The upper part was constructed from 3.0 mm clear acrylic sheet and made gas-tight by heated glue doubling with silicon sealant.The chamber included two gas sampling points, a thermometer and a fan.Size of the acrylic chamber was 0.25×0.25×1.5 m (width×length×height).A small electronic fan was installed to ensure a thorough gas mixing inside the chamber during the measurement.The bottom part was made from aluminum rod.This frame was used as a base for the upper part.Four-side groove was made with 4.0 mm trench to accommodate the 3.0 mm acrylic chamber during the gas sampling.The aluminum frame was firmly inserted into the top soil overnight prior the measurement.During the measurements, the chamber was placed on top of the aluminum frame.In each constructed wetlands, three chambers were installed at the inlet, middle and outlet.Gas flux measurements were performed at these three sampling locations.The acrylic chambers were placed on the aluminum bases and gas measurements began at 0, 15, 30 and 45 min intervals.Soil temperature at the 5-cm depth was also monitored, whereas soil pH, soil ORP (at 5-cm depth) were measures by pH/ORP device.
Methane (CH 4 ) and nitrous oxide (N 2 O) samples were collected from the chambers and were analyzed later in laboratory with Gas Chromatography (GC) Fig. 2: Linear relationship of greenhouse gas flux response with time instrument.Gas samples were taken from each chamber with polypropylene syringes and transferred to evacuated glass vial.Air inside the glass vials were evacuated creating negative pressure prior to sampling.Vials were overfilled in order to minimize potential diffusion across the septa.All the samples were labeled and kept cool in ice-packed cooler immediately after the sampling.Then, the samples were transport to a freezer, <4°C, waiting for later GC analysis.Methane gas samples were analyzed in a laboratory using a gas chromatograph (Agilent GC 6890, USA) equipped with a flame ionization detector, a Poraplot Q capillary column (0.32 mm ID×10 m).Column, injector and detector temperatures were set at 40, 250 and 300°C, respectively, with split ratio 0.7:1 and split flow 15.0 mL/min.Nitrogen was use as carrier and flow rate of 20 mL/min was maintained during analysis.
Nitrous oxide concentrations were determined using a gas chromatograph (Agilent GC 6890, USA) equipped with a 0.53 mm ID×15 m HP-Plot Q column and a Micro Electron Capture Detector (µECD).The temperatures of µECD and the column were set at 300 and 180°C, respectively.Nitrogen was supplied as the carrier gas at a flow rate of 15 mL/min.Data were analyzed by Agilent Chemstation A0803 software (Agilent, USA).The concentrations of methane and nitrous oxide in gas samples were extrapolated by comparing chromatogram (peak area) of gas sample with standard gases (Air Liquid, Ltd. and Scott Specialty Gases, UK).
Gas emission rates were calculated based on the gas and time linear relationship (Fig. 2).Gas concentrations for each sample were plotted against sampling time.Gas emission rate (ppm/h) was converted to flux rates (mg/m 2 /h) and corrected for chamber volume and temperature (Healy et al., 1996).If the increase/decrease in the gas concentration was non-linear (r 2 <0.85), the measurement was rejected (Altor and Mitsch, 2006).Gas flux (mg/m 2 /h) was calculated by the following equation: where, E = Emission on the aerial basis (mg/m 2 /h) X = Gas concentration increase in chamber (ppm/h Data analysis: Statistical analysis was performed with SPSS ® and Microsoft Excel ® for Windows ® .All data entering statistical comparisons were tested for homogeneity of variance and normal distribution using Levene and Kolmogorov-smirnov Test.If assumptions were fulfilled, one-way ANOVA and LSD's post hoc test were carried out.Otherwise, non-parametric Chi-Square (χ 2 ) Test (Kruskal Wallis) was used instead, followed by Mann-Whitney as Post-hoc test.In all analysis, p<0.05 was considered as a significance level.

RESULTS
Constructed wetlands performance in wastewater treatment: Average removal rates of BOD, COD, NH 3 -N and TP were in the ranges of 57-70%, 51-67%, 25-43% and 39-45%, respectively (Table 2).Similar removal rate of BOD was reported by Coleman et al. (2001) that used Typha latifolia in constructed wetlands.Higher removal rate of COD was found in constructed wetlands planted with Eriochloa aristata and Eleocharis mutate in Columbia reported by Caselles-Osorio et al. (2010) at about 75%.In this present study, the removal efficiencies varied in constructed wetland units with different plants species.High removal rate of BOD occurred in the units planted with Phragmite sp., while low removal rate occurred in Canna sp.The highest COD removal rate found in the unit planted with Phragmite sp., while removal rate was lowest in Cyperus sp.NH 3 -N was removed in the unit planted with Canna sp., more effectively than Phragmite sp.Total phosphorus removal rate was high in the unit planted with Canna sp., while the rate was lower in the units planted with Cyperus sp.Statistical analyses of removal efficiency by one-way ANOVA followed by LSD's post hoc test indicated that average removal rates of BOD, COD and NH 3 were significantly different for all plant species (Table 3).

Seasonal variation of environmental factors in constructed wetlands:
Monthly average of air temperature reached maximum in September (36.7°C) while the lowest mean was found in March (24.7°C).It should be noted that lower air temperature in March was unusual in the area of northeastern Thailand but it occurred in March 2011 during the experiment.Soil temperatures at all constructed wetlands were measured at 5 cm depth.The graphic shown in Fig. 3 illustrates that soil temperatures were in the range of 20-29°C, which was in optimum range for plants and microorganisms to grow (Iasur-Kruh et al., 2010) and narrowly changed during three seasons.Soil temperature peaked in December with the average of 29°C and reached its minimum value in January (19.9°C) before it rise again in summer season (March-June).
Average soil pH during the experiments ranged between 6.7 and 8.0, in different constructed wetlands.However, soil pH remained in the optimum range for methanogenic bacteria (Iasur-Kruh et al., 2010).There were differences in soil Oxidation-Reduction Potential (ORP) across the seasons.The lowest soil ORP occurred in January with the average of -258±5 mV whereas the highest soil ORP occurred in May with the average of -127±6 mV.

Seasonal variation of CH 4 gases emissions from constructed wetlands:
Methane emissions from all constructed wetlands observed during August and September were noticeable higher than those in other months (Fig. 3).The highest CH 4 emissions from all constructed wetlands with different plants occurred in September with the average of 58.3, 44.9 and 21.6 mg/m 2 /h for Phragmite sp., Canna sp. and Cyperus sp., respectively.The lowest CH 4 emissions occurred during December and March.CH 4 emissions from constructed wetlands planted with Phragmite sp. and Canna sp., were lowest in March with the average of 2.4 and 1.1 mg/m 2 /h, respectively, whereas low CH 4 emissions from Cyperus sp., occurred in December with the average of 0.9 mg/m 2 /h.When seasonal variation was taken into account, CH 4 emissions from all constructed wetlands were highest in hot rainy season (July-October) with the average of 10.3-19.9mg/m 2 /h and lower in summer season (March-June) with the average of 2.1-6.3 mg/m 2 /h and cool season (November-February) with the average of 2.4-3.9 mg/m 2 /h.The CH 4 emission was lower in summer season that may because the lower air temperature in March was unusually low indicating that local climate fluctuation was influence CH 4 emissions from constructed wetlands in the area.Seasonal variation of CH 4 emissions is summarized in Table 4.
Seasonal variations of N 2 O gases emissions from constructed wetlands: High Nitrous oxide emissions from all constructed wetlands were found during April and October (Fig. 4).The highest N 2 O emissions from constructed wetlands planted with Phragmite sp., were highest in June with the average of 3.6 mg/m 2 /h while high emissions from Canna sp. and Cyperus sp., occurred in April with the average of 2.9 and 3.7 mg/m 2 /h, respectively.The lowest N 2 O emission from constructed wetlands planted with Phragmite sp. and Canna sp., occurred in November with the average of 0.3 and 0.2 mg/m 2 /h, respectively.Low N 2 O emissions from Cyperus sp., occurred in February with the average of 0.3 mg/m 2 /h.When seasonal variation was taken into account, N 2 O emission from all constructed wetlands were highest in hot rainy season (July-October) with the average of 1.3-2.6 mg/m 2 /h, followed by summer season (March-June) with the average of 0.9-2.5 mg/m 2 /h and lowest in cool season (November-February) with the average of 0.3 mg/m 2 /h.

Influence of plant species on CH 4 and N 2 O emissions:
Data on CH 4 and N 2 O emissions from seasonal study were used to evaluate the influence of plant species within constructed wetlands on greenhouse gas emissions.Descriptive statistics for CH 4 and N 2 O emissions are shown in Table 5.Data screening showed that gas emissions were neither a normal distribution nor homogeneity of variance.Thus, chi-square was suitable for differentiating greenhouse gas fluxes among different plant species within constructed wetlands.The means of CH 4 and N 2 O emissions from different plants species were compared by chi-square (χ 2 ) test (Kruskal Wallis).The results showed that the mean of CH 4 and N 2 O emissions from different plants species were significantly different (p<0.05).Constructed wetlands planted with Phragmite sp., emitted the highest CH 4 with the average of 11.2 mg/m 2 /h but emitted the lowest N 2 O with the average of 0.9 mg/m 2 /h.Constructed wetlands planted with Cannas sp., emitted CH 4 and N 2 O with the average of 6.0 and 1.0 mg/m 2 /h.Constructed wetlands planted with Cyperus sp., emitted CH 4 with the average of 5.9 and highest N 2 O with the average of 1.8 mg/m 2 /h.(IPCC, 2001).Our estimate showed that the GWP of constructed wetlands planted with Phragmite sp., had the average CH 4 and N 2 O emissions about 11.2 and 0.9 mg/m 2 /h, respectively.These numbers were about 258 and 266 mg CO 2 equivalent/m 2 /h (or 524 mg CO 2 equivalent/m 2 /h in total).Constructed wetland planted with Canna sp., had the average CH 4 and N 2 O emissions of 6.0 and 1.0 mg/m 2 /h, respectively, corresponded to the average GWP of 138 and 296 mg CO 2 equivalent/m 2 /h, for CH 4 and N 2 O emissions, respectively (or 434 mg CO 2 equivalent/m 2 /h in total).

Estimated
In the case of constructed wetland planted with Cyperus sp., the average CH 4 and N 2 O emission were 5.9 and 1.8 mg/m 2 /h, respectively.Estimated mean GWP of CH 4 and N 2 O emissions were 61 and 325 mg CO 2 equivalent/m/h, respectively, or approximately 669 mg CO 2 equivalent/m 2 /h.Therefore, the results indicated that estimated GWP of CH 4 and N 2 O emissions from constructed wetland planted with Cyperus sp., was the highest, followed by Phragmite sp. and Canna sp., constructed wetlands, respectively.

DISCUSSION
Average CH 4 and N 2 O fluxes from wastewater treatment constructed wetlands, planted with three emergent plants, ranged between 5.9-11.2and 0.9-1.8mg/m 2 /h, respectively.The average fluxes were comparable to those reported by Inamori et al. (2007) which used experimental scale free water surface flow constructed wetland treating non-point sewage at Tsukuba, Japan.However, average N 2 O emission was lower than the studies of 2.39 and 3.4 mg/m 2 /h reported by Wu et al. (2009) and Inamori et al. (2007).Seasonal fluctuations of CH 4 and N 2 O emissions were observed in this present study.The average CH 4 and N 2 O emissions were highest in hot rainy season, followed by summer season and lowest in cool season.However, the pattern of seasonal CH 4 and N 2 O emission variations and their relations to environmental factors were not obvious.These may because important environmental factors such as soil temperature, soil pH and soil ORP were in the optimum range for gas production and changed in a narrow range during the experiment.Additionally, CH 4 and N 2 O emissions were not the result of one-factor action but of the interaction of more biotic and abiotic factors.
For pollutant removal, it was found that constructed wetland planted with different plants had significantly different in efficiency of BOD, COD and NH 3 -N removal although TP removal rates indicated no significantly different.The constructed wetland planted with Phragmite sp., had the highest removal rate of BOD and COD but had lowest efficiency on NH 3 -N removal.Cyperus sp., or Canna sp., had the higher removal efficiency on NH 3 -N.Despite the benefit of constructed wetlands on wastewater treatment, it is important to considered by-product of constructed wetlands in low carbon era.The use of Phragmite sp., should be of benefit in reducing GWP from these greenhouse gases when it is used in constructed wetlands for wastewater treatment.

Fig. 1 :
Fig. 1: Map of study area in northeastern Thailand ) h = Height of chamber (m) M = Molecular weight R = Gas constant = 0.0821 (atm/K/mol) T = Absolute temperature (K) Air pressure = 1 atm Wastewater analysis: Daily wastewater samples were collected from the influent and effluent points and analyzed for Biochemical Oxygen Demand (BOD) until the steady-state conditions were reached.After the steady-state conditions have achieved, monthly analyses were performed on influent and effluent samples for BOD, Chemical Oxygen Demand (COD), ammonia nitrogen (NH 3 -N) and Total Phosphorus (TP).All samples were analyzed according to the standard methods (APHA-AWWA-WEF, 2005).

Table 1 :
Characteristics of synthetic wastewater

Table 2 :
Pollutants removal rates in constructed wetlands with different plant species

Table 3 :
Comparison of pollutants removal rates among constructed wetlands with different plant species by one-way ANOVA

Global Warming Potential (GWP) of CH 4 and N 2 O emissions from constructed wetland planted with different plants:
Comparison of CH 4 and N 2 O characteristics in terms of Global Warming Potential (GWP), were corresponded to the average GWP of CH 4 and N 2 O at 23 and 296 times of CO 2 , respectively