Notice: Wiley Online Library will be unavailable on Saturday 30th July 2016 from 08:00-11:00 BST / 03:00-06:00 EST / 15:00-18:00 SGT for essential maintenance. Apologies for the inconvenience.
 Methane (CH4) fluxes were measured in hypereutrophic Meiliang Bay of Taihu Lake with static chambers for 1 year. The results showed that the macrophyte-covered infralittoral zones were the “hotspots” of CH4 emission in water systems. There were large temporal variations for CH4 fluxes, ranging from −1.7 to 131 mg CH4 m−2 h−1, in the macrophyte-covered littoral zone. The highest CH4 emissions occurred during the period of the summer algal bloom. The amount of CH4 flux from June to September accounted for about 50–90% of the annual fluxes. CH4 fluxes from the bare infralittoral zone (−0.2∼4.2 mg CH4 m−2 h−1) were low and close to those in the pelagic zone. The difference in CH4 fluxes between macrophyte-covered and bare infralittoral zones indicated that vegetation in the inundated area played an essential role in CH4 production. In the infralittoral zone, the redox condition (DO, Eh), temperature, and primary production controlled CH4 fluxes; these variables explained 47% of flux variation, whereas such influences were not detected in the pelagic zone.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
 Methane, as an important radiative trace gas, accounts for about 20% of the greenhouse effect [Cicerone and Oremland, 1988; Wuebbles and Hayhoe, 2002]. Lakes and swamps contribute about 200 Tg y−1, which accounts for about 30% of the global methane flux to the atmosphere [Houghton et al., 2001]. However, documentation of emissions from lakes has been limited [Bastviken et al., 2004; Kankaala et al., 2004], and researches have mainly focused on boreal lakes [e.g., Huttunen et al., 2002a; Kankaala et al., 2004]. China has more than 2300 lakes greater than 1 km2 with a total area about 71,000 km2 and currently most of them are in a eutrophic state [Jin, 2001]. Researches on CH4 emissions from these lakes are very few. These eutrophic lakes with high sediment organic content may be potential important CH4 sources [e.g., Michmerhuizen et al., 1996; Casper et al., 2000]. Two recent studies showed that high CH4 emissions occurred in two eutrophic lakes in China [Duan et al., 2005; Xing et al., 2005]. Littoral zones are active biogeochemical areas of lake systems. In boreal lakes, high emissions of CH4 from littoral zones, exceeding those measured from boreal peatlands, have been reported [e.g., Nykänen et al., 1998; Juutinen et al., 2001, 2003]. There are no reports about CH4 emission from littoral zones in China. The neglect of eutrophic freshwater bodies, especially their littoral zones, raises serious uncertainties about estimates of the regional CH4 emissions. Field data on CH4 exchange are urgently needed to improve our understanding of regional CH4 budgets.
 In this research, we investigated CH4 fluxes from the pelagic and littoral zones of hypereutrophic Taihu Lake and also measured the major environmental variables in water and sediments, to examine their influences on CH4 flux.
2. Materials and Methods
2.1. Site Description
 Taihu Lake, a shallow hypereutrophic lake in a zone of northern subtropical monsoonal climate, covers a surface area of 2338 km2 with average water depth of 1.9 m. The mean annual air temperature is 14.9°–16.2°, and the growing season spans from March to November. Meiliang Bay (31°24′–31°32′N, 120°07′–120°14′E) is a semiclosed bay and is encompassed by hills to the north of Taihu Lake. In summer, heavy alga blooms break out from July to September, and numerous algae assemble along the littoral zone. Study sites were selected in the pelagic zone (site D) and the littoral zones (sections A and B) of Meiliang Bay. A series of field experiments were conducted within a perpendicular transect extending from the outer edge of the hillside to the exterior margin of floating-leaved plants. The transect was divided into three subzones. The terminology used follows that of Wetzel : the supralittoral zone lies entirely above the water level, the eulittoral zone encompasses the shoreline region affected by water level fluctuation, and the infralittoral zone is the macrophyte-covered inundated zone, generally with emerged, floating-leaved, and submerged vegetation. However, in Taihu Lake, only a few floating-leaved and submerged plants are alive in the infralittoral zone in spring, and they die when a severe Microcystis bloom occurs during summer. Therefore most of the infralittoral zones are bare or covered only by emerged plants. In our study area, section A is a macrophyte-covered littoral zone with three subzones and section B is bare (nonmacrophyte) with only an infralittoral zone.
 Nine sampling sites in section A (Figure 1) and three in section B were established along the gradients from open water to shore, marked as A1–A9 and B1–B3, respectively. The interval distances of sites A1–A9 were 30, 20, 10, 10, 10, 30, 30, and 30 m, respectively. In section A, A1–A4 belong to the infralittoral zone with 30–40 cm deep soft sediment, A5–A7 belongs to the eulittoral zone, and A8–A9 are in the supralittoral zone. Few submerged or floating-leaved vegetation survived in spring from A1 to A3, and A3 was at the outer fringe of the reed belt. A3–A7 is characterized by Phragmites australis with an above-ground dry biomass density of 3–4 kg m−2. The dominant species in the supralittoral zone was Miscanthus saccharifloous. In section B, all sites with an interval of 25 m were within the infralittoral zone, where the bottom was eroded by waves, and only gravel existed here. Site D is located in the pelagic zone of Meiliang Bay and is more than 500 m far away from the shore.
2.2. CH4 Fluxes
 CH4 fluxes were measured by the closed-chamber technique approximately once or twice a month from August 2003 to August 2004. Duplicate chambers were used at sites of sections A and B and triplicate chambers at Site D. The stainless steel chambers consisted of two parts: pedestal and upper chamber. The pedestal is 25 cm high with an internal diameter of 40 cm. The lower rim was sharpened so that it could be driven into the soil, and the upper rim had a 2 cm by 2 cm gutter around the outside that could be filled with water to make an airtight seal with the upper chambers. When water depth exceeds 25 cm, a height-adjustable buoy with an internal diameter of 40.5 cm keeps the pedestal floating on the water surface. The upper chamber (h = 45 cm, I.D. = 40 cm) is equipped with two battery-driven (12 V) brushless fans for air circulation and one temperature probe. In this study the chamber was not as high as the reeds (3–4 m). If the reeds were folded into many parts, the gas in the hollow tissue (aerenchyma) is in direct contact with air; then the increased concentration of CH4 that we measured in the chamber may not reflect the true value of CH4 fluxes. Therefore the chambers were installed between the reeds, and our measurement did not include the plant-mediated CH4 flux.
 Sampling bridges (50 cm wide, 100 m long) were installed from the upper limit of high water level to the open water, to avoid disturbing the soil/sediment. During the gas sampling period, the upper chamber was inserted into the water-filled gutter of the pedestal. Four gas samples were taken at 5 min intervals from the headspace of the chambers through polytetrafluoroethylene tubes into the 500 mL gas sampling bags (multilayer polymer with aluminum foil) by an oil-free air pump (12V, 1.8 L min−1, KNF, Germany). Pressure changes during gas sampling were eliminated by using a long capillary to allow air to enter the chamber as a sample was withdrawn [Regina et al., 1999].
 CH4 concentrations were determined within two days after sampling, by a gas chromatograph (Agilent 4890D) equipped with a flame ionization detector. The CH4 flux was calculated from the linear increase of CH4 concentrations in the headspace of chambers, and the annual CH4 emission from every subzone in the lake was calculated based on monthly emissions. The accuracy of the analyses was maintained by calibrating gas chromatographs against a standard gas mixture after every eight samples, which kept the coefficient of variation of replicated concentration determinations below 1%. Measurements with an r2 < 0.9 in a linear regression of concentration change over time were considered erroneous and excluded.
2.3. Environmental Variables
 Environmental data were collected in order to probe the relationship between CH4 fluxes and environmental variables and to produce a regression model for CH4 flux. In the inundated area, on each gas sampling date, triplicate water and sediment samples were collected at every gas sampling site, placed in pre-acid-washed polyethylene bottles or bags, transported to the laboratory in an icebox, and stored in the dark below 4°. Water samples were taken using a 3 m long sampling tube to provide a vertical sample from the whole water column. In the eulittoral and supralittoral zones, we measured soil variables twice in both the growing and nongrowing seasons. Dissolved oxygen (DO), pH, Eh were measured in situ by oxygen meter (YSI Model 57), pH, and Eh meter (HI 8424 Microcomputer HANNA), respectively. Nutrients, including ammonium nitrogen (NH4), nitrate nitrogen (NO3) and nitrite nitrogen (NO2), and chlorophyll-a (chl-a) were determined using APHA standard methods (4500-NH3 F. Phenate Method; 4500-NO2− B. Colorimetric Method; 4500-NO3− E. Cadmium Reduction Method; 10020 H. Chlorophyll, respectively) [American Public Health Association, 1998]. Total nitrogen (TN) and total phosphorus (TP) were determined simultaneously by peroxodisulfate oxidation [Ebina et al., 1983] of the original water samples.
 About 300–400 g wet weight surface sediment (3–5 cm) was collected using an Ekman-bottom sampler (Hydro-Bios, Germany). In section B, sediment samples were not collected every time. Ten to 15 g of fresh soil were dried at 65° for 48 hours to determine gravimetric water content and about 200 g of fresh soil were air-dried in the laboratory for subsequent analyses. The dried sediments were sieved (≤150 μm) to measure soil organic matter (SOM) and total soil nitrogen (TSN) contents. SOM was obtained from the loss on ignition (LOI %, 500°, 2 h) [Zhu and Carreiro, 2004]. TSN was determined by the semimicrodistillation method using a catalyst of K2SO4-CuSO4-Se [Lao, 1996].
2.4. Estimates of Methane Concentrations in Surface Water
 During our sampling processes, bubbles arising from the bottom were not found and almost no stepwise increase of CH4 concentration was observed in the chambers. These phenomena suggested that the measured CH4 fluxes mainly came from diffusion. However, ebullition has been reported for eutrophic lakes [Kankaala et al., 2004, 2005]. Therefore assuming diffusive flux only, the CH4 concentrations in the surface water were estimated using results of several empirical studies. If the estimated concentrations appear unrealistic, based on given literature data on the range of surface CH4 concentrations previously observed, ebullitions are assumed to occur. The estimate procedure is outlined by the following five steps.
 1. Schmidt numbers (Sc) for CH4 are calculated by the surface water temperature (t) in °C and can be described as the equation derived by Wannikhof  for CH4 in freshwater:
 2. Wind speeds (Uz, m s−1) measured at z = 2 m above water surface were converted to wind speeds at 10 m (U10) using the following equation [Amorocho and Devries, 1980; equation (21)]:
where C10 equals the surface FLdrag coefficient for wind at 10 m (1.3 × 10−3) [Stauffer, 1980], K is the von Karman constant (0.41), and z = height of wind speed measurement in m above water surface.
 3. K600 (the piston velocity of CO2 at 20°C) was calculated using wind speed at 10 m above water surface (U10) [Cole and Caraco, 1998], then piston velocities for CH4 at each water temperature (KCH4,t) were obtained using the following two formulae:
where Sc600 equals the Schmidt number for CO2 at 20°C, n = 1/3 when U10 < 2 m s−1, n = 1/2 when U10 > 2 m s−1.
 4. The Ostwald solubility coefficient (α) for CH4 in freshwater was calculated using Bunsen solubility coefficients (β) at each surface water temperature (T in degrees Kelvin) [Wiesenburg and Guinasso, 1979].
where R is the gas constant.
 5. The concentration of dissolved CH4 in the surface water (Cw) was estimated using the expression [Wannikhof, 1992]:
where F is the measured CH4 flux and Ca is the measured concentration of CH4 in air at the water surface.
2.5. Statistical Analyses
 The normal distribution of the fluxes and environmental variables was tested using the Kolmogorov-Smirnov test, and most variables were right-skewed. The data were normalized by natural logarithm transformation. Nonparametrical Mann-Whitney U and Kruskal–Wallis Post Hoc tests was used to compare the fluxes between different sites. After normalization, the CH4 fluxes were related to environmental variables by Pearson correlation analysis. Pearson correlations were also used to analyze autocorrelations between environmental variables. Mutual relationships among variables were then investigated by principal component analysis. Stepwise multiple linear regression analysis was used to correlate CH4 fluxes with the principal components [Camdevýren et al., 2005]. The alpha value was set to 0.05. All residuals were investigated for independence, constancy, and normality. The above analyses were performed using the SPSS statistical package (SPSS Inc., release 11.0.0).
3.1. CH4 Fluxes
 Spatial variation of CH4 emissions mainly occurred in section A. From open water to lakeshore, CH4 emissions increased gradually, with peak values, at A3 (mean ± SD 33.5 ± 48.2 mg CH4 m−2 h−1) where floating algae easily assembled (chl-a 200,000–300,000 μg L−1) during the algae-bloom period, and then decreased in the eulittoral and supralittoral zones (Figure 2). Influx occurred and accounted for about 54% of the total measurements in the supralittoral zone. The infralittoral zone in section A was the key area of CH4 emissions (10.0 ± 23.0 mg CH4 m−2 h−1), with the eulittoral (2.6 ± 5.5 mg CH4 m−2 h−1), pelagic (0.5 ± 1.9 mg CH4 m−2 h−1), and infralittoral zones of section B (0.4 ± 0.9 mg CH4 m−2 h−1), and the supralittoral zone (0.1 ± 0.5 mg CH4 m−2 h−1) following in order of decreasing CH4 emissions during our investigation period. The emissions from the bare infralittoral littoral zone (section B) were significantly lower than those from the macrophyte-covered infralittoral zone (p < 0.001) and were approximately equal to the fluxes from the pelagic zone (p = 0.919).
 CH4 fluxes in the littoral zone displayed a strong temporal variation (Figure 3) and followed a seasonal trend. The highest fluxes occurred in summer except for the supralittoral zone when algal bloom broke out. In wintertime, the lowest air and sediment temperatures were about 1° and 5°, respectively. CH4 fluxes in winter were also fairly low, and ranged from −0.09 to 0.33 mg CH4 m−2 h−1 in the littoral zones and from −0.03 to 0.25 mg CH4 m−2 h−1 in the pelagic zones. There was no statistical difference for CH4 fluxes between littoral and pelagic zones during winter.
3.2. Relationships Between CH4 Fluxes and Environmental Variables
 The ln(CH4 flux) values from two infralittoral zones (sections A and B) were plotted against environmental variables (Figure 4). Significant relationships exist between CH4 fluxes and DO, Eh, TP, chl-a, air and sediment temperature. Among these environmental variables, chl-a had a significantly positive relationship with TP (r = 0.820, p < 0.001). SOM and TSN in the sediment had little effect on CH4 fluxes. In the pelagic zone, CH4 effluxes were not related to measured environmental variables. Mutual correlations were common among environmental variables in the water and sediment (data not shown). Therefore multiple linear regression analysis of CH4 could not be performed directly, and principal components (PCs) were derived to analyze the relationships among these variables in the infralittoral and pelagic zones. The small quantity of data limited such analysis in the other two zones. The principal component analysis (PCA) on environmental (water and sediment) variables for the infralittoral zones resulted in four components with eigenvalues larger than 1, which explained 88% of total variance (Table 1). The first component explained about 43% of the observed variance. Several variables were significantly correlated with this first component, namely DO, Eh, chl-a, TP, air and sediment temperature. The first component could be seen as influenced by the redox condition, primary production, and seasonal variation, in which primary production had a close relationship with seasonal variation. The second component explained about 22% of observed variance. The two variables correlated strongly with this component were TN in water and SOM in sediment. The third component only had a significant correlation with pH and no variables strongly related to the fourth component.
Table 1. Results of Principal Component Analysis (PCA) for the Water and Sediment Variables in the Infralittoral Zone
 A stepwise multiple linear regression analysis between PCA components and CH4 fluxes in the infralittoral zones showed that the first component had significant influence (r2 = 0.470, p < 0.05) and could be depicted as ln(CH4,flux) = 0.42 + 1.56 × component 1 + ɛ, where “ɛ” was the residual of model. We also performed PCA and stepwise multiple linear regression analyses for the pelagic zone (data not shown). CH4 flux, however, was not strongly related to measured environmental variables which explained less than 5% of the variability of CH4 fluxes.
3.3. Estimation of CH4 Concentration in the Surface Water and Annual Emission
 Considering only diffusive flux, the estimated CH4 concentrations in the surface water (5% trimmed mean 4.3 μM) was supersaturated, ranging from 0.4 to 141 μM. The highest concentration was 99.5 ± 36.5 μM in the macrophyte-covered infralittoral zone in summer. The annual CH4 emissions from different subzones are shown in Table 2. The contribution of the infralittoral zone was up to 104 g m−2, about 23 times higher than that in the pelagic zone.
Table 2. Mean and Total CH4 Fluxes From the Littoral and Pelagic Zones During Ice-Free Period in Comparison With References Reported
CH4 Flux, mg m−2 h−1
Total Flux, g m−2
SOM LOI, %
Recalculate the mean CH4 flux according to the reference.
 In the macrophyte-covered littoral zone, CH4 fluxes differed considerably within short (10–30 m) distances. The great spatial variation suggests that fine-scale investigation should be done. In this study the summer CH4 fluxes in the macrophyte-covered infralittoral zone ranged from 9.3 to 131 mg CH4 m−2 h−1 and annually ranged from −1.7 to 131 mg CH4 m−2 h−1. The underestimated fluxes (not including plant-mediated flux) from the macrophyte-covered infralittoral zone were still much higher than those from other parts of Meiliang Bay and other boreal lakes (Table 2) and even higher than emissions from paddy fields (1.5–14.4 mg CH4 m−2 h−1) in the same region [Wang, 2001], and the fluxes from pelagic zones were in a similar range (Table 2). The same phenomenon was also found in an outer reed area [Kankaala et al., 2004], which approximately was similar to the macrophyte-covered infralittoral zone in our study, in a flooded transgression shore (Table 2). These results suggest that macrophyte-covered infralittoral zones are the “hotspots” of CH4 fluxes in water systems.
 The difference in CH4 fluxes between macrophyte-covered and bare infralittoral zones indicates that vegetation in the inundated area may have played an essential role in CH4 production. Vegetation along the littoral zone decreased the energy of wave and current, causing most particles from open water and land to be intercepted and combined with accumulated reed culms of the previous growing seasons. The SOM in the littoral zone was higher than that in the pelagic zone, and the soft sediment (about 30–40 cm) in the littoral zone was also thicker. These physical aspects may have caused the macrophyte-covered infralittoral zone to produce larger amounts of CH4 than the bare infralittoral zone. In addition, the amounts of accumulated algae in the macrophyte-covered littoral zone were unusually large (chl-a 200,000–300,000 μg L−1). Decomposing littoral algae can maintain significantly higher CH4 production than detritus from P. australis [Kankaala et al., 2003b].
 Large amounts of organic matter accumulated in the macrophyte-covered littoral zone in this study, suggesting that ebullition might be important emission mechanism. However, only about 6% of the data were excluded for r2 < 0.9 in a linear regression of CH4 concentrations change over time, and the increase in concentration was seldom stepwise in the 6% data. We conclude that large ebullition seldom took place during our sampling period. Thus most of the measured emissions may have been due to diffusion. However, sites only 10 m apart occasionally had widely different emissions, which is surprising since the diffusive flux is positively related to the methane concentration in the water. Assuming diffusive flux only, the estimated concentration in the inundated area was supersaturated, ranging from 0.4 to 141 μM. The 5% trimmed mean was 4.3 μM, which suggested that most lower emission rates were mainly due to diffusive flux. However, the highest rate in the macrophyte-covered infralittoral zone in summer appeared unrealistic given literature data on the range of surface methane concentration (0.1–2.3 μM) previously observed in more than 100 lakes [Kling et al., 1992; Engle and Melack, 2000; Huttunen et al., 2002b; Bastviken et al., 2004; Huttunen et al., 2004]. We conclude that the high emission by diffusion alone was unlikely and that the emission mechanism must have included continuous ebullition, which in turn indicates that large amounts of CH4 were produced. An explanation for continuous ebullition was that wave interactions with the reed belt created turbulence making small ebullition more frequent.
4.2. Temporal Variation of CH4 Fluxes
 A pronounced seasonal variation of methane emissions, i.e., maximum in July and August and low but clearly detectable in winter, typically observed in boreal peatlands and vegetated littoral areas [e.g., Nykänen et al., 1998; Kankaala et al., 2004], was also found in the littoral zone of Meiliang Bay. During the period from June to September, the amount of CH4 fluxes accounted for about 50–90% of the annual fluxes in our study. Numerous studies have shown that plant biomass or primary production is a master variable in controlling CH4 in water systems [e.g., Whiting and Chanton, 1993; Huttunen et al., 2002a; Kankaala et al., 2003a]. In our study area, reed was the dominant species in the littoral zone and algae were the main primary producer in the whole water system. Previous studies have shown that CH4 emissions were correlated positively to chl-a in Lake Donghu (r2 = 0.12, p = 0.04) in China [Xing et al., 2005] and in Lake Mekrijärvi (r2 = 0.48, p < 0.05) in Finland [Juutinen et al., 2003]. In Meiliang Bay, a similar relationship between chl-a and CH4 fluxes was also found (r2 = 0.34, p < 0.0001). The algal assemblage could play an essential role in CH4 fluxes. The CH4 content in bubbles extracted from the sediment of shallow, highly productive reservoirs and lakes were of modern origin [Huttunen et al. 2002a]. A 14C pulse-labeling experiment under field conditions indicated that photosynthate carbon can be converted to CH4 quickly [King and Reeburgh, 2002]. These studies in eutrophic lakes indicate that fresh organic C derived from primary production rather than old peat deposits are very important for CH4 production. Therefore organic matter production by algae, the main primary producers, may control the seasonal pattern of CH4 production and fluxes in eutrophic lakes. Temperature was another important factor affecting seasonal variation of CH4 production and fluxes because a significant correlation between CH4 efflux and temperature (air and sediment) was found in our study.
 Stepwise multiple linear regression against PCs describes 47% of the CH4 flux variation in the infralittoral zone. Only one component, which contained temperature, DO, Eh, chl-a, and TP, affected the regression outcome. DO and Eh represented redox conditions. A significant correlation between TP and CH4 fluxes has been found in many freshwater studies [e.g., Huttunen et al., 2002a; Bastviken et al., 2004]. Generally, phosphorus is the limiting factor of primary production in freshwater systems. In this research, TP significantly related to chl-a (r2 = 0.67, p < 0.001), and chl-a coupled with TP was depicted as a factor representing primary production. We conclude that CH4 fluxes in the infralittoral zone were controlled by temperature, reducing conditions, and primary production, with temperature and primary production being the main factors affecting the seasonal patterns of CH4 fluxes.
4.3. Importance of Littoral CH4 Emission in Eutrophic Lakes
 Drift algae accumulate naturally near the downwind shore in eutrophic water systems [Verhagen, 1994] and have become a common phenomenon along the world's shorelines [Valiela et al., 1997]. Fresh organic C derived from algae would be enriched nearshore, especially in the macrophyte-covered area. The lack of available data precludes an accurate estimation of total CH4 emission from littoral regions of Chinese lakes. However, our data hint that littoral regions and lake-associated wetlands may be important to the total atmospheric CH4 load. We conclude that the littoral CH4 emission could be important in China because more than 2300 Chinese lakes (most eutrophic) cover about 71,000 km−2.
 Yujing Mu, Yuesi Wang, and Wen Zhang are appreciated for their technical assistance. We also thank Alice S. Honig in Syracuse University and Wayne S. Gardner in University of Texas for language edit. This work was funded by the Key Project of Knowledge Innovation Program of CAS (KZCX1-SW-12) and the National High Technology Research and Development Program of China (2002AA601011-05).