Effects of eutrophication and Typha domingensis Pers on methanogenesis in tropical constructed wetland

Aim: Constructed wetlands (CWs) have shown to be an efficient way for removing nutrients from wastewaters. However, a negative effect is the emission of greenhouse gases (GHGs). Methane (CH4) is a GHG released from the organic matter degradation under anaerobic conditions by methanogenic bacteria. In CWs, the excessive nutrient inputs may stimulate both primary production and methanogenesis and then affect the overall cost-benefit balance of this system. Here, we assessed the role of cattail stands (Typha domingensis Pers) on CH4 concentrations in eutrophic and mesotrophic constructed wetlands. Methods: CWs were filled with water and sediment colonized by the aquatic macrophyte Typha domingensis Pers from two tropical freshwater lakes, one mesotrophic and another eutrophic. Results: Our results showed consistently higher CH4 concentrations in waters and sediment in the eutrophic CWs than in the mesotrophic CWs (two times), indicating a positive influence of eutrophication on methanogenesis. Conclusion: In conclusion, CH4 emissions from tropical CWs demand to be considered or even mitigated in any use of this system.


Introduction
The use of constructed wetlands (CWs) as an alternative to conventional wastewater treatment has increased in recent decades in order to reduce the excessive nutrient loading from human activities (Vymazal et al., 1998;Maltais-Landry et al., 2009).These systems can act as buffers between the main sources of pollution and natural watercourses by reducing the inputs of nitrogen (N) and phosphorous (P), nutrients that result in eutrophication of freshwater and saline systems (Schindler, 2006).However, CWs show high emissions of greenhouse gases, including methane (CH 4 ).This gas is produced during the last step of organic degradation under anaerobic conditions (Jones, 1991;Boone, 1991;Fonseca et al., 2004;Conrad et al., 2010).Methanogenesis is influenced by temperature (Kang and Freeman, 2002), redox potential, pH, organic matter contents ( Van-Hulzen et al., 1999;Bodegom and Stams, 1999), salinity, absence of oxygen (Keltjens and Vogels, 1993;Oren, 1994;Munson et al., 1997;Thauer, 1998) and other electrons acceptors (Lovley et al., 1982).
High emissions of CH 4 to the atmosphere have been observed especially from warm tropical inland waters (Bastviken et al., 2010).Tropical wetlands are broadly distributed in Brazil.These ecosystems are often degraded by the human activities.Nutrient-enriched inputs from sewage, agricultural fertilizers or residuals from the sugarcane processing, for biodiesel production, have been commonly related to eutrophication in inland waters and wetlands at low latitudes.Our hypothesis is that the high nutrient levels in eutrophic constructed wetlands with Typha domingensis Pers.favor the the methanogenesis.

2.1.Experimental design
We used twelve 40 L-CWs (length 41.0 cm, height 32.0 cm, depth 41.0 cm) with controlled temperature in a greenhouse.Details of the CWs design are illustrated in the Figure 1.The CWs were filled with water and sediment and replanted with Typha domingensis Pers.from two tropical lakes in Brazil (one eutrophic -Lake Vigário -and another mesotrophic -Lake Cima).Lake Vigário receives the inputs of sugarcane residuals and domestic waste, while Lake Cima only sugarcane residuals.Both lake are costal coastal, shallow (maximum depth <5.0 m) and relatively close to one another.Additionally, we determined the macrophyte growth measuring height, leaves, and a general characterization of sediment from all experimental CWs (granulometry -wet sieving and organic matter contents -dry weight) (Kimble et al., 2001).Data were not normally distributed even after transformations (Kolmogorov-Smirnov, p < 0.05).Non parametric tests were then used to compare differences (Kruskal-Wallis followed by the Dunn`s post test, significant p < 0.05 for both).

Results
Table 2 indicates the variation of physicochemical parameters from Eutrophic CWs (Lake Vigário) planted and unplanted plots; and Table 3 indicates the variation of physico-chemical parameters from mesotrophic CWs (Lake Cima) planted and unplanted plots during experimental time.
pH, temperature, electrical conductivity, redox potential, alkalinity and dissolved oxygen concentration in waters were not significantly different comparing CWs within each eutrophic and mesotrophic lake (Kruskal-Wallis, p > 0.05).Granulometric fractions of the sediment and the being a receptor of wastewater.The lake is densely colonized by emergent and floating aquatic macrophytes, respectively Typha domingensis Pers and Eicchornia crassipes (Lannes, 2004), but the marginal area colonized by this vegetation has been substantially reduced by silting up from human activities.Lake Vigário shows pH around 7.5 and electrical conductivity around 0.7 mS.cm -1 (Lannes, 2004).On the other hand, Lake Cima has a lower human disturbance.This lake is located in the northern region of Rio de Janeiro state, in Campos dos Goytacazes (21° 45' 23" S, 41° 19' 49" W), showing freshwaters with a conductivity around 50 µS.cm - (Luz, 2003).
The experimental design was conducted using two hydraulic systems containing six 40 L-boxes of asbestos and one 200 L-box (general source of water).In each one of 40 L-boxes, a drip system on the surface to keep constant the water level with the continuous water inflow(water depth was maintained in 15 cm, accordingly to Figure 1).Three CWs filled with sediment and waters from each mesotrophic and eutrophic lake were planted with the aquatic macrophyte Typha domingensis Pers collected in the respective lakes (five young individuals), while other three were kept unplanted, in order to assess the effects of macrophytes on the methane concentrations in waters and sediments.. First measurements in our experimental CWs were conducted after 30 days for stabilization.
CH 4 concentrations in waters and the bottom sediment were analyzed 5 mL of surface waters which were collected in all CWs water surface using.Water samples (5 mL) were taken using plastic syringes with needle and stored in 10 mLglass vials closed with massive rubber stoppers.In addition, a small amount of NaCl was added in each vial to expurgate methane from waters to the headspace (Casper, 1992).These vials were kept under refrigeration before headspace analysis in a Gas Cromatographer (GC-14B Shimadzu), using a chromatographic packed column Porapak-Q at 60 °C; column injector at 120 °C, FID detector at 200 °C, and nitrogen was adopted as gas carrier.Then, 4 mL of NaOH 4% (v/v) was added using a syringe with needle to expurge CH 4 from the pore water to the headspace (Casper, 1992).After shaking, 0.2 mL of gas from headspace was also extracted using plastic syringes with needle, and then injected in the GC described above).Abiotic variables measured during the experiment in waters were temperature, electrical conductivity, pH, total alkalinity and dissolved oxygen using a calibrated  ranging from 50.2 to 79.6 mmol.L -1 and from 26.6 to 36.2 µmol.L -1 respectively (Figure 3).Planted and unplanted eutrophic CWs showed methane concentrations in the sediment significantly higher than those mesotrophic (Kruskal-Wallis test: p < 0.05 and Dunn´s test p < 0.05).
organic matter contents indicated that the silt-clay fraction was predominant for both lakes.Higher organic matter contents were observed in the eutrophic than in the mesotrophic lake sediments, respectively 14 and 8%.CH 4 concentrations in waters of the eutrophic CWs (Figure 2 and Table 4) ranged from 0.13 to 1.40 µmol.L -1 , showing higher values than in mesotrophic CWs (0.12-0.55 µmol.L -1 , Figure 2).Comparing planted and unplanted eutrophic CWs, CH 4 concentrations were higher planted than in unplanted plots (Kruskal-Wallis, p < 0.05 and Dunn, p < 0.05), but no significant differences were observed between mesotrophic CWs (Kruskal-Wallis, p < 0.05).Indeed, we did not find any significant relationship between presence of vegetation and methane concentrations in mesotrophic CWs, (linear regression, R 2 = 0.61; p < 0.05).Planted plots also showed higher methane concentrations in waters than unplanted plots, considering both eutrophic and mesotrophic CWs (Kruskal-Wallis, p < 0.05 and Dunn, p < 0.05).
Typha domingensis was not related to significant differences for CH 4 concentrations in eutrophic and mesotrophic CWs (Kruskal-Wallis, p < 0.05 and Dunn, p < 0.05).Higher methane concentrations in waters were found in eutrophic CWs, ranging from 63.4 to 69.7 µmol.L -1 , than in mesotrophic CWs (Kruskal-Wallis test: p < 0.05 and Dunn´s test p < 0.05), ranging from 29.3 to 34.1 µmol.L -1 .In the sediment, methane concentrations in the eutrophic CWs were higher than in mesotrophic ones, Table 4. Results obtained with Dunn´s test for evaluate the significant differences of treatments (with Typha domingensis Pers and without macrophyte) for Lake Vigário CWs (eutrophic) and Lake Cima CWs (mesotrophic).this study, macrophyte Typha domingensis Pers may be acting as methane source to the atmosphere, as is widely discussed in literature.This shows the importance of eutrophication and organic matter contents in eutrophic CWs, reducing dissolved oxygen and stimulating methanogenesis.The use of constructed wetlands is usually applied for wastewater treatment and nutrient removal, but the GHG's emissions to the atmosphere demand to be considered and subsequently mitigated.Our data confirms the contribution of constructed wetlands for GHG production; nonetheless further investigations mainly in tropical ecosystems are required.

Discussion
Our results consistently indicate a positive influence from eutrophication and the presence of aquatic macrophyte Typha domingensis Pers on methanogenesis.Additionally, high alkalinity of eutrophic CWs is likely caused by the sewage inputs into this lake, supporting anaerobic processes of organic matter cycling (Brix et al., 1996;Singh, 2001;Chang and Yang, 2003;Wang and Han, 2005).
For unplanted eutrophic CWs, CH 4 concentrations in waters and sediments increased over the experiment, confirming lower CH 4 production related to absence of aquatic macrophyte (Chan et al., 2002).In contrast, higher CH 4 concentrations found in eutrophic CWs were likely caused by the sewage inputs into Lake Vigário, increasing the availability of organic substrates to methanogenesis (Neue et al., 1995, Singh et al., 2000;Liikanen and Martikainen, 2003).
Lake Vigário is a freshwater lake, nonetheless in this lake is important detach the role of sulphatereduction competing with methanogenesis by organic substrates in Lake Vigário (Lannes, 2004); high methane concentrations in waters and sediments of CWs also suggested potentially elevated rates of methanogenesis in this eutrophic lake.This confirms that besides sulphate-reduction, the methanogenesis might be also an important organic degradation processes under highly eutrophic conditions that are typically related to substantial sewage inputs.
Planted plots also showed higher CH 4 concentrations in waters than those unplanted, confirming the role of the aquatic vegetation to increase methanogenesis like in artificial rice fields (Schütz et al., 1989) or natural tropical lakes (Bastviken et al., 2010).The CH 4 concentrations observed in experimental CWs are within the range for other tropical lakes in Brazil (Table 5), with the exception was a brackish tropical coastal lagoons (e.g.Lagoon Cabiúnas), which showed lower CH 4 concentrations probably by the higher salinity limiting CH 4 production (Fonseca et al., 2004).
In conclusion, we clearly support our hypothesis as the eutrophic conditions led to elevate CH 4 in eutrophic CWs.Our data suggested that sewage discharges enriched in organic substrates and nutrients besides other subsequent hydrochemical conditions, such as redox potential and pH, could positively affect CH 4 production in aquatic environments.Although the measures of methane flux through macrophytes have not been made in

Figure 2 .
Figure2.CH 4 concentrations in waters of eutrophic (Lake Vigário) and mesotrophic (Lake Cima) CW`s with and without macrophyte.a) Lake Vigário CWs with macrophyte Typha domingensis Pers; b) Lake Vigário CWs without macrophyte; c) Lake Cima CWs with Typha domingensis Pers; and d) Lake Cima unplanted plots.Distinct letters (a,b) show significant differences among treatments, for each lake separately (Kruskal-wallis test (95% s level of significance) and Dunn´s test (95% of level of significance).

Table 2 .
Variation of physico-chemical parameters from Eutrophic CWs (Lake Vigário) planted and unplanted plots during experimental time.

Table 3 .
Variation of physico-chemical parameters from mesotrophic CWs (Lake Cima) planted and unplanted plots.

Table 5 .
Methane concentration in waters for others aquatic ecosystems.