Investigating CH4 and N2O emissions from eco-engineering ...

Process Biochemistry 42 (2007) 363–373 www.elsevier.com/locate/procbio

Investigating CH4 and N2O emissions from eco-engineering wastewater treatment processes using constructed wetland microcosms R. Inamori b, P. Gui a,*, P. Dass a, M. Matsumura b, K.-Q. Xu a, T. Kondo c, Y. Ebie a, Y. Inamori a b

a National Institute for Environmental Studies, Tsukuba 305-0053, Japan Graduate School of Life and Environmental Sciences, Tsukuba University, Japan c Department of Chemical Engineering, Waseda University, Japan

Received 8 April 2006; received in revised form 7 September 2006; accepted 9 September 2006

Abstract Methane (CH4) and nitrous oxide (N2O) are important greenhouse gases, because of their contribution to the global greenhouse effect. The present study assessed emissions of N2O and CH4 from constructed wetland microcosms, planted with Phragmites australis and Zizania latifolia, when treating wastewater under different biological oxygen demand (BOD) concentration conditions. The removal rate was 95% for BOD and more than 80% for COD in all three pollutant concentrations, both plants’ removal rates of pollutants were at almost the same level, and both were found to resist BOD concentrations as high as 200 mg L1. When BOD concentrations fell below 200 mg L1, the soil plant units reached an average of 80–92% T-N and T-P removal rates; however, as the concentrations increased to 200 mg mg L1 or when during the initial phases of winter, the removal rates for T-N and T-P decreased to less than 70%. With NH3-N removal, the influences of BOD concentrations and air temperature were more obvious. When BOD concentrations increased to 100 mg L1 after October, an obvious decrease in NH3-N removal was detected; almost no nitrification occurred beginning in December at BOD concentrations of 200 mg mg L1. N2O and CH4 emissions showed obvious seasonal changes; higher emissions were observed with higher BOD concentrations, especially among Z. latifolia units. The enumeration of methane-oxidizing bacteria and methane-producing bacteria was also conducted to investigate their roles in impacting methane emissions and their relationships with plant species. The pollutant purification potentials of P. australis and Z. latifolia plant units during wastewater treatment of different pollutant concentrations occurred at almost the same levels. The nutrient outflow and methane flux were consistently higher with Z. latifolia units and higher concentrations of BOD. The more reductive status and higher biomass of methanogens may be the reason for the lower nitrification and higher CH4 emissions observed with Z. latifolia units and higher concentration systems. The Z. latifolia root system is shallow, and the activity of methanotrophs is primarily confined to the upper portion of the soil. However, the root system of P. australis is deeper and can oxidize methane to a greater depth. This latter structure is more favorable as it is better for reducing methane emissions from P. australis soil plant systems. # 2006 Elsevier Ltd. All rights reserved. Keywords: Eco-engineering; Constructed wetland microcosms; Phragmites australis; Zizania latifolia; CH4 and N2O emission; Methanotrophs; Methanogens; Fluorescent in situ hybridization

1. Introduction In 1957, Odum put forward the concept of ecological engineering [1]. In 1972, he introduced a self-organizing ecosystem of wetlands to receive wastewater, and the successful practice of arranging tertiary treatment of wetlands spread all across Florida and the rest of the World. It now is

* Corresponding author. Tel.: +81 29 850 2285; fax: +81 29 850 2285. E-mail address: [email protected] (P. Gui). 1359-5113/$ – see front matter # 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.procbio.2006.09.007

predicted that this system ultimately will replace traditional wastewater treatment technology [2,3]. Simulating a wetland ecosystem, soil plant systems can make use of the assimilation capacity of soil and aquatic plants to remove both pollutants and nutrients without additional energy demand. Being able to effectively remove or convert large quantities of pollutants from point sources (municipal and certain industrial effluents) and non-point sources (mine, agriculture and urban runoff), these systems are regarded to be a promising wastewater treatment technology for the future [4–9,39,42–44]. They have been used widely across the US and Europe, and are becoming

364

R. Inamori et al. / Process Biochemistry 42 (2007) 363–373

increasingly widespread in developing countries, like China and India. The sustainable operation of these systems depends upon high-rate conversion of organic and nitrogenous loads to gaseous end products. Among these gaseous products, N2O and CH4 are of special interest, because of their greenhouse effect and their high global-warming potential. As a report released by IPCC (intergovernmental panel on climate change) has mentioned, the greenhouse effects of CH4 and N2O actually are 20–30 and 200–300 times that of CO2, respectively [35]. Many studies have proven that soil plant systems with various kinds of aquatic plants, like Typha, Phragmites and water hyacinth, including rice, will produce methane [10–12,36] and nitrous oxide during denitrification [13,14,36]; such systems would include marshes, agro-ecosystems, and either natural or constructed wetlands. Consequently, the mechanisms for N2O and CH4 emission from plant soil systems and their control increasingly has been emphasized. Reviewing the internal structure of a soil plant system, it is not only the activity of the microorganisms that cleanses the wastewater, by removing nutrients and organic pollutants, it also is the hydrophytic vegetation inside that enhances nutrient uptake and the settling of solids [15,16]. At the same time, hydrophytic vegetation acts like a biological pump, converting sunlight into chemical energy and carrying oxygen from the leaves to the roots, thereby creating a small, oxygen-rich zone around the root, and promoting microbial growth [15] by providing a place for microbes like bacteria, fungi, algae, and protozoa to attach. Pollution-eating microbes colonize in the oxidized zone surrounding the root surface and convert pollutants into harmless gaseous CO2. However, in the anoxic zone further from the root, CH4 will be generated and N2O also will be emitted because of denitrification. It also has been reported that, at the same time, methane may be oxidized at the interface between the anoxic and oxic sites, where the concentration gradients of methane and oxygen overlap and oxygen diffusion through roots can reduce methane emission [10,11]. Such an interface can be located at the surface of flooded soil and in the rhizosphere of aquatic plants. The more that CH4-oxidizing bacteria inhabit the area, the less CH4 emission occurs. Although the existence of CH4-oxidizing bacteria in marshes has been believed to depend upon the kinds of plants there, few studies have compared the effect of different plant species on the efficiency of removing specific pollutants, or of CH4 or N2O emissions. Consequently, the primary aims of the present study are: (1) to compare the removal of pollutants and nutrients using two different types of soil plant systems with different BOD loads— considering the need for plants to be transplanted for future wetlands and ecology safety and the preference for plants in the local district, typical hydrophytes that inhabit the local Kasumigaura lakeside wetland area, Zizania latifolia and Phragmites australis, were selected as the plants of interest; (2) to compare CH4 and N2O emissions in different soil plant systems, and at different BOD loads; and (3) to enumerate methane-producing and oxidizing bacteria in different systems by fluorescent in situ hybridization (FISH), so as to clarify the

mechanisms by which plant type and BOD loads influence CH4 and N2O emissions. 2. Materials and methods 2.1. Design of experimental system The experiment was split into six independent units, each having a replicate sub-unit. Each cell was 88 cm in length and 56 cm in diameter. All cells received a 12 cm layer of gravel (1–2 cm in diameter), followed by a 6 cm layer of coarse sand (3–5 mm in diameter), and then a 43 cm layer of fine sand (1– 2 mm in diameter). The depth of the water above the sand was 6 cm, as shown in Fig. 1. Artificial wastewater composed of dextrin, bact peptone, yeast extract, meat extract, NaCl, MgSO47H2O, (NH2)2–CO, KH2PO4, and KCl, was used in this experiment. An influent system combined gravity flow and pump were used. High concentration wastewater with BOD of 1000, 2000, and 4000 mg L1 was made every 3 days and pumped to a small mixing column fixed near the surface of the wetland microcosm. At ten times the dilution, water was sent to the column by the flow of gravity; the volume was controlled by the water level. This mixing process took place once every 5 min, and at the end of the mixing process, the diluted water was discharged into the wetland microcosm in bulk. As a result, different BOD loads were created by introducing inlet wastewater with different BOD concentrations of 200, 100, and 50 mg L1 into two replicate sub-units of each plant. The inflow wastewater quality is identified in Table 1, which shows that, during the experiment period, the water quality of high concentration wastewater changed with the climate, temperature, and so on. Since the tendency for this change is the same regardless of the BOD concentration, the experiment cells experiencing the same change were comparable; this process also simulated the fluctuation of water quality for practical wastewater. Thirty centimeter sprouts of the aquatic plants Z. latifolia and P. australis, consisting of primarily the taproot, were transplanted from the nearby Kasumigaura (Fig. 2) lakeside into their respective cells (Fig. 3). Each cell had a hydraulic retention time (HRT) of 7 days.

2.2. Analytical items and methods Every 10 days, 1 L of influent and effluent experimental units were collected and immediatly measured for water quality. The parameters included total nitrogen (T-N), total phosphorus (T-P), ammonium nitrogen (NH3-N), chemical oxygen demand (COD), and biological oxygen demand (BOD). Among these, NH3-N, T-P, and T-N were analyzed using an auto-flow analyzer using the colorimetry TRAACS 2000. The others were analyzed according to standard method of Japan Sewage Works Association [38]. Gas sampling was done using a closed chamber. Gas samples were collected from the air outlet at the top of each chamber by means of a syringe, to determine the concentration of methane and nitrous oxide. The N2O concentration was analyzed by means of gas chromatography (Shimadzu Co., Japan) equipped with an electron capture detector (ECD) and a Poropak Q column, using 40 mL min1 argon-containing 5% methane as the carrier gas, and the

Fig. 1. Profile of constructed wetland microcosm unit.


365

Table 1 Quality of inflow wastewater used soil plant systems Items

T (8C)

BOD5 (mg L1)

T-N (mg L1)

NH3-N (mg L1)

T-P (mg L1)

Influent BOD50 Influent BOD100 Influent BOD200

5–28.5 5–28.5 5–28.5

18–82 (average 38) 30–118 (average 60) 90–257 (average 163)

0.8–17.3 (average 9.8) 8.2–35.2 (average 18.4) 13.7–53.8 (average 36.7)



2.3. Enumeration of methanogens and methanotrophs

Fig. 2. Location of Kasumigaura Lake in Japan. temperature of the detector and oven set at 340 and 80 8C, respectively. CH4 was analyzed by gas chromatography (Shimadzu Co., Japan) with a flame ionic detector and a Poropak Q column, using nitrogen as the carrier gas, and the temperature of the oven and injector port set at 60 and 80 8C, respectively. A modified portable handheld ORP (oxidation reduction potential) meter with separate comparing electrode and platinum electrode made by DKK-TOA Co. Ltd. of Japan was used for ORP measurement. The modified platinum electrodes were inserted into the soil beforehand; once the comparing electrode was inserted below the soil’s surface, ORP values could be read from the portable ORP meter.

In water environments, which are polluted by wastewater, organic matter accumulates in the sediment. Methane is produced during mineralization of organic matter under strictly anaerobic conditions. About 70% of the methane is formed from acetate and the remaining 30% from carbon dioxide and hydrogen [17]. The complete conversion of organic matter into methane in an anaerobic system requires at least three functionally distinct trophic groups of bacteria, whose coordinated activities are required for overall function [18]. These bacteria are hydrolytic-fermentative, syntrophic-acetogenic and methanogenic bacteria. For the final step in methane generation, it is important to study the relationships between methane emission from the sediment and the population dynamics of the methanogens in the sediment. Various techniques have been used to study the ecology of methanogens and methanotrophs in environmental samples, including cell culture, autofluorescence, electron and confocal scanning laser microscopy, and molecular biological techniques. Fluorescence in situ hybridization (FISH) has been used to study the distribution of methanogens and methanotrophs in wetland land units; it was used in this study as well. Table 2 lists the oligonuleotide probes used in this study. Soil material was fixed with 4% paraformaldehyde in PBS. It was left for 16 h at 4 8C and was then exposed to 50% ethanol in PBS and stored at 20 8C until used. The material was spotted on gelatine-coated slides (0.1% gelatine, 0.01% KCr(SO4)2) in 3 ml aliquots per well and dried at room temperature. After dehydration with 50, 80, and 99.8% ethanol for 3 min each, the hybridization was carried out at 46 8C for 2 h in the hybridization buffer (0.9 M NaCl, 20 mM Tris–HCl, 10 mM EDTA, 0.01% SDS) in the presence of x% formamide

Fig. 3. Constructed wetland microcosm systems for treatment of domestic wastewater.

366


Table 2 Fluorescently labeled oligonucleotide probes used in this study Probe name

Target group probe

Sequence (50 –30 )

Reference

MG1200 MB1174 MS1414 MX825 MG-64 MA621

Methanomicrobiales Methanobacteriaceae Methanosarcinaceae Methanosaetaceae Type I methanotrophs Type II methanotrophs

CGGATAATTCGGGGCATGCTG TACCGTCGTCCACTCCTTCCTC CTCACCCATACCTCACTCGGG TCGCACCGTGGCCGACACCTAGC CCGAAGGCCTRTTACCGTTC TCAAAGGCAGTTCCGAGGTT

[19] [19] [19] [19] [20] [20]

for hybridization, determined by the probe. After hybridization, the slides were immersed in a washing buffer for 15 min at 48 8C (20 mM Tris–HCl, 10 mM EDTA, 0.01% SDS, and y mM NaCl, depending on the formamide concentration during hybridization); they were subsequently rinsed with distilled water and air-dried. The slides were mounted with Fluoro Guard solution and examined with a Leica TCS 4D microscope. Microbial cells were counted at 100 times magnification. Twenty fields, randomly selected, covering an area of 0.01 mm2 were examined from a sample distributed over circular area of 19.6 mm2 each; this sample was then calculated to the amount of soil. Probe MG1200 can detect Methanomicrobiales; probe MB1174 can detect the Methanobacteriaceae of Methanomicrobiales; and probe MS1414 and MX825 can detect the Methanosarcinaceae and Methanosaetaceae of Methanosarcinales. These probes thereby comprise order, family, and genus-specific probes for phylogenetic groups of methanogens. The total numbers of methanogens in the experimental units was calculated by summing up the enumeration results for these four probes. The sequence of the probes and the hybridization conditions were in accordance with those recommended by Raskin et al. [19]. On the other hand, methanotrophs utilize methane as their sole carbon and energy source. They are divided into two distinct physiological groups. Type I methanotrophs assimilate formaldehyde produced from the oxidation of methane, using the ribulose monophosphate pathway; they contain predominantly 16-carbon fatty acids; and they possess bundles of intracytoplasmic membranes. Type II methanotrophs utilize the serine pathway for formaldehyde assimilation; they contain predominantly 18-carbon fatty acids; and they have intracytoplasmic membranes arranged around the periphery of the cell. For whole cell hybridization, probes MG-64 and MA-621 were used for types I and II methanotrophs, respectively. The sequence of the probes and the hybridization conditions were in accordance with those recommended by Bourne et al. [20].

3. Results 3.1. Performance of pollutant removal During the operation of soil plant units, the degree of removal of BOD, COD, T-N, T-P and NH3-N was investigated, from July 2001 to December in 2002. Water samples were taken

every 10 days and Fig. 4 summarizes the average removal rate of the various pollutants. Since the removal rate was 95% for BOD and more than 80% for COD in all three pollutant concentrations and both plant units most of the time, this shows that the removal rate of pollutants by both plants were at almost the same level and both were found to resist BOD concentrations as high as 200 mg L1. Soil plant system studies generally have reported significant reductions in BOD due to the additional detention provided by further storage, and to the presence of plants assisting with sedimentation and filtration, thereby accounting for the high organic pollutant removal performance even when influent BOD reaches 200 mg L1. However, the removal efficiencies of the different soil plant units for nitrogen and phosphorus were quite variable. At lower concentrations of 50 and 100 mg L1, soil plant units reached an average of 80–92% T-N and T-P removal rates; but, as the concentration increased to 200 mg L1, the removal rates for TN and T-P decreased markedly, to lower than 70%. We also observed temporal changes in T-N and T-P removal (Fig. 5), in that the percentage of T-N and T-P removal, with both the Z. latifolia and P. australis systems, began to decrease from December through to the next April. This especially was noted for Z. latifolia and when BOD concentration was greater than 200 mg L1, at which times T-N and T-P removal decreased to nearly 10%. With NH3-N removal, the influences of BOD concentrations and air temperature were more obvious. At a BOD concentration of 50 mg L1, NH3-N removal exhibited small incremental increases as the plants grew. But when the BOD concentration increased to 100 mg L1, an obvious decrease in NH3-N removal was detected after October. The latitude of change was greater in the second year than in the first year. The effluent of

Fig. 4. Pollutant removal efficiencies of constructed wetland microcosm units at different BOD concentration conditions.


367

Fig. 5. Nutrients removal efficiencies of constructed wetland microcosm units at different BOD concentration conditions.

both plant systems was discharged without any nitrification at the beginning of December at BOD concentrations of 200 mg L1; this phenomenon was noted in P. australis systems even earlier. After the beginning of December, the average ambient temperature per 24 h decreased to lower than 10 8C. Many reports have discussed the activity of ammonia oxidizing bacteria [40,41], which are responsible for the key process of nitrification; these bacteria are also supposed to be the reason for temperatures that result in the sharp decline of nitrification in units with different BOD concentrations. For BOD of 200 mg L1 units, the degradation of BOD needs to consume more oxygen; however, the plant’s ability to transport oxygen declined because the plant withered. As a result, the soil became anaerobic, dropping nitrification to nearly zero after December. This behavior indicates that, during winter, nitrification is the controlling process for nitrogen removal. Moreover, the increase in BOD concentration and correspondingly higher BOD removal demand increase the amount of oxygen available for organic breakdown and decreases the

amount of oxygen available for nitrification. Furthermore, because the growth of for the P. australis root system is much slower than that of Z. latifolia in the first year, the roots tend to be vulnerable. When BOD concentrations increased to 200 mg L1, the growth of the P. australis root system was further hindered because of the insufficient growth in the first year. Finally, the average NH3-N removal of P. australis was only 21%, while that of Z. latifolia was nearly 78%. It was not until the next May that nutrient removal efficiency began to increase again, though it had quickly normalized by June. By the time the temperature started to increase in April, nutrient removal was at its lowest. This indicates that nutrients accumulated over the whole winter inside the constructed wetland, and that this persisted until full microbial growth resumed in May and June. Monthly data reveal that the rate of phosphorus removal decreased over time and, on occasion, the wetland released phosphorus. While phosphorus may be immobilized in soil plant systems by plant uptake, absorption, precipitation and

368


Fig. 6. Seasonal changes in nitrous oxide emission from different constructed wetland microcosm units at different BOD concentration conditions.

incorporation into biological films, P removal is highly sensitive to loading rate and is considered to be finitely subject to the ability of the substrate to adsorb phosphorus. 3.2. Temporal changes in CH4 and N2O emissions CH4 and N2O emissions from constructed wetland microcosm units were measured monthly from July 2001 to December 2002. For N2O emission, four peaks were observed during the course of observation (Fig. 6). In the first year, one peak was observed in August and another in October. Thereafter, a sharp decline began, a decline that continued until the temperature started to rise and microbial activity resumed the next April. In April, then, a sharp increase led to the third peak. The fourth peak was observed during September of the second year. Higher N2O emission was observed with higher BOD concentration units, especially among Z. latifolia units. Measurement of CH4 flux also showed considerable variation with sampling date (Fig. 7) and four peaks also were observed. The first peak was observed during July in the first summer. The second peak was observed in October when the

plant reached maturity and most of the oxygen in the rhizosphere was consumed via root respiration. In winter, when the plant withered, methane emission was very low. During the second March, when the plants started to grow again, their biomass increased and emission also increased, generating a third peak. The fourth peak was observed when the plants again matured in September. 3.3. Distribution of methanotrophs and methanogens In soil plant units, aquatic plants develop an air ventilation system to survive in the anaerobic environment [21]. Methane may be oxidized at the interface between anoxic and oxic sites, where the concentration gradients of methane and oxygen overlap, such as at the surface of flooded soil or in the rhizosphere of aquatic plants [22]. Methane emission from soil plant units is the net effect of methane production by methanogens and oxidation by methanotrophs to CO2. Because the biological oxidation of methane often is associated with sites of methanogenesis, an understanding of the relationship between these two activities is essential, if their contribution to global methane emission is to be

Fig. 7. Seasonal changes in methane emission from different constructed wetland microcosm units at different BOD concentration conditions.


369

Fig. 8. Amount of methanogens and methanotrophs along the depth of different vegetation’s rhizosphere at different BOD concentration conditions.

understood and controlled. Therefore, the distribution of methane-oxidizing bacteria and methane-producing bacteria in the rhizosphere was enumerated using fluorescence in situ hybridization (FISH). Fig. 8 presents the cell numbers of methanogens, type I methanotrophs and type II methanotrophs at different depths of rhizosphere in the experimental systems. The biomass of methanogens in soil plant units inhabiting different depths of P. australis and Z. latifolia are shown in

Fig. 8(a). In all three BOD-concentration soil plant units, the number of methanogens was higher in the Z. latifolia units than in the P. australis units. Moreover, as the BOD load increased, the number of methanogens also increased. Conversely, as shown in Fig. 8(b) and (c), except for type I methanotrophs in the upper 10 cm of the rhizosphere, the number of methanotrophs was higher in P. australis units. Also, as BOD concentration increased, the number of methanotrophs decreased.

370


4. Discussion 4.1. Seasonal changes in pollutant removal and gas emissions During water purification, oxygen concentration determines organic degradation, the oxidation of ammonia, and the subsequent process of denitrification. As a comprehensive system consisting of plant, soil and water, re-aeration of oxygen into a system primarily transpires via the water surface or through the extensive oxygen-transport system of aerenchyma tissue that exist in roots, stems and leaves [34]. This system allows a plant to transport needed oxygen to the rhizosphere and offer the oxygen for microbial growth there. Decided by plant growth status, re-areation through arrenchyma also changes along with the season, and the corresponding redox also changes and influences pollutant removal and CH4 and N2O generation. Fig. 9 presents the ORP values at different depths of the plant rhizosphere, in experimental units in August (Fig. 9a) and October (Fig. 9b) 2001. It is obvious that ORPs in August were in the aerobic range, but most of the ORPs in October were in the anoxic and anaerobic range. Reed growth is active in August, but plants start to wither starting in September; and this is the main reason for ORP change. As a result of ORP change, although air temperature did not decrease much in October, nonetheless pollutant removal still began to decline.

Different from NH3-N and T-N removal, which are concerned mainly with the microbial process, T-P removal primarily is achieved through sand adsorption and plant uptake. But both the T-N and T-P removal processes experienced a sharp decline after February and reached their trough in April. Their removal efficiency increased again from May and reached a peak in August. The biomass of both P. australis and Z. latifolia peaked in August and September. The peaks for T-N and T-P removal in August suggest that plant uptake is the key process in T-N and T-P removal. Starting in November, both plants begin to lose their leaves, and the leaves begin to accumulate within the wetland. From the middle of February, air temperature begins to increase, and the accumulated nitrogen and phosphorus start to be released from the fallen leaves; but the plant growth is not sufficient to adsorb the nutrients and these results in a sharp increase in nitrogen and phosphorus concentrations in the effluent, and a resultant sharp decrease in nitrogen and phosphorus removal. Both the aerobic process of nitrification and the anaerobic process of denitrification contribute to the generation of N2O [23–26]. The aerobic condition with soil redox values less than 400 mV and reductive conditions between 100 and 200 mV may result in high N2O emission. The former process was thought to be an aerobic process and the latter was thought to be an anaerobic process. As a result, when N2O emission showed the same tendency as CH4 emission, as in October 2001 and September 2002, an anaerobic process was felt to be

Fig. 9. Redox potential along the depth of different vegetation’s rhizosphere at different BOD concentration conditions.


operational. If a contrary tendency is detected – for example, when CH4 emission is low but a CH4 peak takes place, as in August 2001 and April 2002 – then an aerobic process should be considered. Fig. 9(a) depicts measurements performed in August 2001, when peak emission of N2O and low emission of CH4 were observed. Readings all were found to be in the range of 50–200 mV, which is favorable enough for N2O generation, but not reductive enough to generate methane. This means that an aerobic process is contributing to N2O emission here, as we presumed. Fig. 9(b) depicts measurements performed in October 2001. All readings were found to be lower than 100 mV, and most of them were lower than 100 mV. This is the range of ORP that favors both N2O and CH4 generation. Air temperature is another parameter that affects net CH4 flux, by influencing the CH4-oxidizing and CH4-producing microbial community and its level of activity [25]. High temperatures result in high consumption of oxygen via organic removal and greater reductive redox status. The summer peak of CH4 in the first year was correlated with higher temperatures that season. The spring peak was the result of organic matter accumulating over the winter months, when low temperatures reduced microbial activity. As the temperature increased, the level of microbial activity also increased, which increased decomposition and the resultant supply of substrate for methanogens. 4.2. Influence of BOD concentration and plant species on pollutant removal and methane emission The higher the BOD concentration, the more substrate decomposes and the more reductive ORP will result. Being very sensitive to redox potential, ammonia oxidation was shown to be influenced by BOD concentration the most. This explains the deterioration in ammonia nitrogen removal when BOD concentration reaches 200 mg L1. Since organic pollutants can be degraded in both aerobic and anaerobic conditions, BOD removal did not exhibit obvious differences at different degrees of BOD concentration. For T-N and T-P removal, at lower BOD concentrations of 50 mg L1, the T-N influent also was lowest at around 10 mg L1; sand adsorption was sufficient and no obvious fluctuation was identified across the different seasons. When BOD concentration is increased to 100 mg L1 or higher, plant uptake also plays an important role in T-N and T-P removal, and then the removal exhibits obvious fluctuation with the seasons. The higher the concentration, the more the latitude of fluctuation is. For gas emissions, methane flux consistently was higher with Z. latifolia units than P. australis units. Methane flux also was higher at higher degrees of BOD concentration. Shown in Fig. 9, ORP values of Z. latifolia units all were lower than those of P. australis units at the same BOD concentration. Moreover, ORP values decreased with increases in BOD concentration. This explains the higher methane emissions observed with higher BOD concentration, and in Z. latifolia units. These results suggest that P. australis, with its complex root zone, and lower BOD conentration both tend to weaken the activity of methane-producing microorganisms or

371

strengthen the decomposition of methane in favor of methane oxidization. Considering the effect of macrophytes on CH4 emission, when macrophytes develop an aerenchymatous structure to avoid deficiency in their roots, the aerenchyme cells then form an important route for the transport of CH4 from the anaerobic layer into the atmosphere [27–29]. On the other hand, this also will favor the transport of O2 released from such roots into the rhizosphere, and increase methane oxidation in the anaerobic layer [10,11,21–23,37]. This O2 rapidly may be consumed by the respiration of plant roots or aerobic microbes. Furthermore, macrophytes enhance methane emission by providing easily degradable substrate for anaerobic decomposition, e.g. root litter and exudates. These differences in the root and stem architecture, aerenchymous tissue, and oxygen availability for rhizospheric bacteria of macrophytes result in ultimate differences in the biomass of methanogens and methanotrophs. The higher number of methanogens in the Z. latifolia units and the higher number of methanotrophs in the Phragmite units reflect the promoting effect of certain vascular plants on methane emission and methane oxidation. In this study, clearly Z. latifolia was favorable for the growth of methanogens and P. australis for the growth of methanotrophs. When all the experiments were completed, all the experimental units were split in the middle to expose their rhizosphere. These differences in the root and stem architecture, aerenchymous tissue, and oxygen availability for rhizospheric bacteria of macrophytes result in ultimate differences in the biomass of methanogens and methanotrophs. The higher number of methanogens in the Z. latifolia units and the higher number of methanotrophs in the Phragmite units reflect the promoting effect of certain vascular plants on methane emission and methane oxidation. In this study, clearly Z. latifolia was favorable for the growth of methanogens and P. australis for the growth of methanotrophs. When all the experiments were completed, all the experimental units were split in the middle to expose their rhizospheres. A very different rhizosphere structure was observed for the Z. latifolia versus P. australis systems. The root of Z. latifolia is shallow, and 90% of the root biomass is concentrated in the upper 10 cm of the experimental unit. Conversely, the root of P. australis is deeper and the root biomass more evenly distributed from near the soil surface to the bottom of the rhizosphere. With less area for oxygen diffusion from the root, more methanogen cells were observed in Z. latifolia units. The distribution of methanogens along the depth of the rhizosphere in P. australis units is even, whereas, it increases along the depth of the rhizosphere in Z. latifolia units. On the other hand, in Z. latifolia units, the number of methanotrophs was greatest at a depth of 10 cm and then decreased along the depth of the rhizosphere. Contrary to this, in P. australis units, a smaller number of methanotrophs was detected at the 10 cm soil depth, but a larger number was observed at depths of 20 and 30 cm. The shallow root of Z. latifolia confines oxygen’s availability and the activity of methanotrophs in the upper portion of the soil, while the root of P. australis is deeper and can oxidize methane to a greater depth. This latter favorable structure is good for reducing

372


Table 3 Ratio of types I and II methanotrophs in different plant units at different BOD concentration conditions Type I methanotrophs

BOD 50 (%) BOD100 (%) BOD200 (%)

Type II methanotrophs

Zizania latifolia

Phragmites australis

Zizania latifolia

Phragmites australis

42.02 46.06 44.62

50.19 54.26 47.36

57.98 53.94 55.38

49.81 45.74 52.64

Fig. 10. Amount of methanogens and methanotrophs in constructed wetland microcosm units at different BOD concentration conditions.

methane emissions from wetlands. Upon entering October, the plants began to wither; oxygen transportation through plants obviously declined. The aerobic area was concentrated mainly near the hair roots. The lower the BOD concentration, the more hair root. The hair root grows the most near 10 cm in the rhizosphere at BODs of 50 and 100 mg L1; it grows the most near 20 cm at BOD 200 mg L1. The abnormally high ORP at 5 cm in the rhizosphere results from the oxygen transfer from surface water. Algae began to grow after the middle of September, increasing more in BOD 100 mg L1 units due to the higher concentration of nitrogen and phosphorus than BOD 50 mg L1. But when the BOD is higher than 200 mg L1, the degradation of organic material consumed most of the oxygen, making it difficult for the algae to grow. The algae released oxygen, resulting in an ORP value of BOD 100 mg L1—the highest level. This determined the ORP distribution as well as the methanotrophs distribution. A supply of both methane and oxygen is essential for methanotroph populations. Type I methanotrophs out-competed type II methanotrophs at low methane concentrations, whereas, the growth of type II methanotrophs was favored under low-oxygen, high-methane conditions [30,31]. Recently, it has been shown that type I methanotrophs are able to proliferate rapidly and dominate CH4 oxidation in aerated rice field soil, when CH4 is supplied at a low (1000 ppmv) mixing ratio; whereas, type II methanotrophs only proliferate at a high (10,000 ppmv) CH4 mixing ratio [32]. Table 3 displays the ratio of types I and II methanotrophs to total methanotrophs within both plant units at different depths into the rhizosphere under different BOD concentration conditions. In this study, the ratio of type II to total methanotrophs was higher in Z. latifolia units. This indicates that, in the

rhizosphere of aquatic plants that are favorable for methane emission, type II methanotrophs dominate. Other investigators have come to the same conclusion [33]. On the other hand, the ratio of type I to total methanotrophs is higher in P. australis units, which have a favorable rhizosphere structure to reduce methane emission. Assessing both methane and oxygen together, the ratios of types I and II did not demonstrate a clear relationship with depth into the rhizosphere or BOD concentration, but we still found that the ratio of type I methanotrophs tended to decrease with BOD concentration. Analyzing the relationship between the cell number of methanogens and methanotrophs inside whole plant units under different BOD concentration conditions (Fig. 10), a linear relationship between BOD load and the amount of methanogens and methanotrophs indicates that BOD loading is the key factor controlling methane, via bacterial growth and methane emission. Acknowledgements This work was supported by the Foundation for Earth Environment Promotion from the Association of International Research Initiatives for Environmental Studies. References [1] Odum HT. Environment, power and society. New York: Wiley; 1971. p. 331. [2] Odum HT, Ewel KC, Mitsch WJ, Ordway JW. Recycling treated sewage through cypress wetlands in Florida. In: D’Itri F, editor. Wastewater renovation and reuse. New York: Dekker (Marcel); 1977. [3] Oduma HT, Odumb B. Concepts and methods of ecological engineering. Ecol Eng 2003;20:339–61.

R. Inamori et al. / Process Biochemistry 42 (2007) 363–373 [4] Dunbabin JS, Pokorny J, Bowmer KH. Rhizosphere oxygenation by Typha Domingensis per. In miniature artificial wetland filters used for metal removal from wastewater. Aquat Bot 1988;29:303–17. [5] Hamiton H, Nix PG, Sobolewski A. An overview of constructed wetlands as alternative to conventional waste treatment systems. Water Pollut Res J Can 1993;28:529–48. [6] Mitsch WJ, Gosselink JG. Wetlands, 2nd ed., New York: Van NostrandReinhold; 1993. [7] Kadlec R, Knight RL. Treatment wetlands. Boca Raton, FL: Lewis Publishers; 1996. [8] Vrhovsek D, Kukanja V, Bulc T. Constructed wetland for wastewater treatment. Water Res 1996;30:2287–92. [9] Haberl R, Perfler R, Laber J, Cooper P. Wetland systems for water pollution control. Wat Sci Tech 1997;35(5):347. [10] Conrad R. Control of methane production in terrestrial ecosystem. In: Andrea MO, Schimel DS, editors. Exchange of trace gases between terrestrial ecosystem and the atmosphere. Chichester: Wiley; 1989. p. 39–58. [11] Chanton JP, Dacey JWH. Effects of vegetation on methane flux, reservoir and carbon isotopic composition. In: Sharkey TD, Holland EA, Mooney HA, editors. Trace gas emissions by plants. San Diego: Academic Press; 1991. p. 65–92. [12] Guo LP, Lin E. Carbon sink in cropland soils and the emission of greenhouse gases from paddy soils: a review of work in China. Chemos Global Change Sci 2001;3:413–8. [13] Aulakh MS, Rennie DA, Paul EA. Acetylene and N-serve effects upon N2O emissions from NH4+ and NO3 treated soils under aerobic and anaerobic conditions. Soil Biol Biochem 1984;16:351–6. [14] Russow R, Sich I, Neue HU. The formation of the trace gases NO and N2O in soils by the coupled processes of nitrification and denitrification: results of kinetic 15N tracer investigation. Chemos Global Change Sci 2000;2:359–66. [15] Hammer DA. Constructed wetlands for wastewater treatment—municipal, industrial, and agricultural. Chelsea, MI: Lewis Publisher; 1989. [16] Brix H. Do macrophytes play a role in constructed treatment wetlands? Wat Sci Tech 1997;35(5):11–7. [17] Gaudy AF, Gaudy ET. Microbiology for environmental scientist and engineers. New York: McGraw-Hill; 1980. [18] Chartrain M, Zeikus JG. Microbial ecophysiology of whey biomethanation: intermediate metabolism of lactose degradationin continuous culture. Appl Environ Microbiol 1986;51:180–7. [19] Raskin L, Stromley JM, Rittmann BE, Stahl DA. Group-specific 16S rRNA hybridization probes to describe natural communites of methanogens. Appl Environ Microbiol 1994;60:1232–40. [20] Bourne DG, Holmes AN, Iversen N, Murell JC. Fluorescent oligonucleotide rDNA pobes for specific detection of methane oxidizing bacteria. J FEMS Microbiol Ecol 2000;31:29–38. [21] Allen LH. Mechanics and rate of O2 transfer to and through submerged rhizome and roots via aerenchyma. Soil Crop Sci Soc Florida Proc 1996;56:74–81. [22] Jespersen DN, Sorell BK, Brix H. Growth and root oxygen release by Typha latifolia and its effects on sediment methanogenesis. Aquat Bot 1998;61:165–80. [23] Annstrong J, Armstrong W. Pathways and mechanisms of oxygen transport in Phragmites australis. In: Cooper PF, Findlater BC, editors. Constructed wetlands in water pollution control. Oxford: Pergamon Press; 1990. p. 529–34.

373

[24] Knowels R. Denitrification. Microbiol Rev 1992;46(1):43–70. [25] Moore TR, Dalva. The influence of temperature and water table position on CO2 and CH4 emission from laboratory columns of peatland soils. J Soil Sci 1993;44:651–64. [26] Devi I, Delaune RD. Light hydrocarbon production in freshwater marsh soil as influenced by soil redox conditions. Wat Air Soil Pollut 1996;88:39–46. [27] Schutz H, Holzapfel-Pschorn A, Rennenberg H, Seiler W. Methane emission from Italian irrigated rice fields. Int Rice Res Newslett 1990;14:35–42. [28] Torn MS, Chapin III FS. Environmental and biotic controls over methane flux from Arctic tundra. Chemosphere 1993;26:357–68. [29] Shannon RD, White JR, Lawson JE, Gilmour BS. Methane efflux from emergent vegetation in peat lands. J Ecol 1996;84:239–46. [30] Amaral JA, Archambault C, Richard SR, Knowles R. Denitrification associated with groups I and II methanotrophs in a gradient enrichment system. FEMS Micrbiol Ecol 1995;18:289–98. [31] Amaral JA, Knowles R. Growth of methanotrophs in methane and oxygen counter gradients. FEMS Microbiol Lett 1995;126:215–20. [32] Henckel T, Friedrich M, Conrad R. Molecular analyses of the methaneoxidizing microbial community in rice field soil by targeting the genes of the 16S rRNA, particulate methane monooxygenase, and methanol dehydrogenase. Appl Environ Microbiol 1999;65:1980–90. [33] King GM. Associations of methanotrophs with the roots and rhizomes of aquatic vegetation. Appl Environ Microbiol 1994;60:3220–7. [34] Brix H. Macrophyte medicated oxygen transfer in wetlands; transport mechanism and rates. In: Moshiri GA, editor. Constructed wetlands for water quality improvement. Boca Raton, FL: Lewis Publishers; 1993. p. 391–8. [35] IPCC. Climate change. Cambridge: Cambridge University Press; 1994. p. 337. [36] Patrick HM, Ronald DD, William HP. Methane and nitrous oxide emissions from laboratory measurement of rice soil suspension; effect of soil oxidation–reduction status. Chemosphere 1993;26(1–4):251–60. [37] Steinberg SL, Coonrod HS. Oxidation of the root zone by aquatic plants growing in gravel-nutrient solution. J Environ Qual 1994;23:907–13. [38] Japan Sewage Works Association. Testing methods for sewage works. Tokyo: Japan Sewage Works Association; 1997. [39] Sun G, Gray KR, Biddlestone AJ, Allen SJ, Cooper DJ. Effect of effluent recirculation on the performance of a reed bed system treating agricultural wastewater. Proc Biochem 2003;39(3):351–7. [40] George A, Kowalchuk L, John RS. Ammonia-oxidizing bacteria: a model for molecular microbial ecology. Annu Rev Microbiol 2001;55: 485–529. [41] Lutz B, Ralf K, Klaus BB. Temperature and moisture effects on nitrification rates in tropical rain-forest soils. Soil Sci Soc Am J 2001;66:834–44. [42] Yang B, Lan CY, Yang CS, Liao WB, Chang H, Shu WS. Long-term efficiency and stability of wetlands for treating wastewater of a lead/zinc mine and the concurrent ecosystem development. Environ Pollut 2006;143(3):499–512. [43] Greenway M. The role of constructed wetlands in secondary effluent treatment and water reuse in subtropical and arid Australia. Ecol Eng 2005;25(5):501–9. [44] Poach ME, Hunt PG, Reddy GB, Stone KC, Johnson MH, Grubbs A. Swine wastewater treatment by marsh–pond–marsh constructed wetlands under varying nitrogen loads. Ecol Eng 2004;23(3):165–75.