Values of δ13C were investigated of CH4 trapped in the soil pore water and floodwater of and emitted from a rice field under continuous flooding throughout the fallow and following rice seasons, and CH4 produced via different pathways and fraction of CH4 oxidized was calculated by using the isotopic data. Pore water CH4 was relatively 13C depleted, with δ13C values about −65‰ over the season except between July and August (around −55‰). Also, hydrogenotrophic methanogenesis was very important (around 50%) for most of the season, while acetoclastic methanogenesis dominated (about 70%) only between July and August. Floodwater CH4 was heavier in δ13C value (from −50‰ to −34‰) than pore water CH4 (from −68‰ to −54‰) over the season, demonstrating that it is highly influenced by methanotrophy. The δ13C value of emitted CH4was negatively correlated with flux in temporal variation (P <0.05), and it was more positive in the fallow season (between −56‰ and −44‰) than in the rice season (between −68‰ and −48‰). This indicates that plant-mediated CH4 transport is probably a more important pathway and causes less CH4 oxidation during the rice season than during the fallow season, which is further confirmed by the fraction of CH4 oxidized being generally greater in the fallow season (60%–90%) than in the rice season (10%–80%). These findings suggest a low contribution of acetoclastic methanogenesis and a high fraction of CH4 being oxidized in the field, especially in the fallow season.
 The increase in concentrations of greenhouse gases (CO2, CH4, and N2O) in the atmosphere is a main cause of global warming. CH4is the second largest radiative forcing of the long-lived greenhouse gas and 25 times greater than CO2 in global warming potential on a 100 year horizon [Intergovernmental Panel on Climate Change (IPCC), 2007]. Over the last 40 years of the twentieth century, atmospheric CH4 increased in concentration at an averaged rate of about 1% yr−1. However, the rate has been decreasing substantially in the last 8 years [Dlugokencky et al., 1998] and tended to be zero during 1999 to 2005 [Dlugokencky et al., 2003]. The cause of neither the rapid increase nor the recent slowing down is clearly understood, which is directly related to the large uncertainties in the magnitudes, temporal and spatial distributions of identified CH4 sources. As a consequence, considerable uncertainties remain in the global CH4 budget. Rice fields are an important anthropogenic source of atmospheric CH4, contributing about 31–112 Tg yr−1 or 5%–19% to the total global emission [IPCC, 2007]. More information is hence needed about the source of CH4 strength to obtain a more accurate estimation of the global emission from the rice fields.
 Carbon isotope study is useful for quantifying CH4 of different sources [Stevens and Engelkemeir, 1988; Miller, 2004]. Determining the carbon isotopic composition and fluxes of CH4 is helpful in measuring relative source strengths of the CH4 and providing constraints for the atmospheric CH4 budget [Miller, 2004]. Studies of carbon isotopic composition of CH4 from rice fields have been conducted, revealing that the values of δ13CH4 varied greatly (Table 1). There are also two reports available in China. Li et al.  reported that the δ13C values of emitted CH4 (just 6 gas samples) ranged from −65‰ to −54‰. Moreover, Bergamaschi  measured δ13C values of CH4 emitted directly from an intermittently irrigated rice field and CH4 in gas bubbles collected from stirred paddy soil of the field. Both CH4 were relatively depleted in 13C, with δ13C values ranging from −71‰ to −52‰ and −71‰ to −58‰, separately. No comparable study, however, has so far been done throughout the fallow and following rice seasons, related to the continuously flooded rice fields, a special kind of rice fields in China. Because of various reasons, such as poor drainage conditions caused by side leaching invasion in hilly and mountainous areas or very high groundwater table in plain, and poor irrigation conditions which could not ensure enough water for rice transplanting next year, about 12% of the rice fields in China is flooded all year-round [Cai, 1999; Cai et al., 2003]. Continuous flooding leads soils to intensive anaerobicity and high CH4 emission, and the contribution of such fields to the total CH4 emission is over 40% [Cai, 1999; Cai et al., 2003]. It is, therefore, essential to investigate rules of the CH4 emissions and corresponding 13C-isotopic signatures in particular.
Table 1. Overview of Available Carbon Isotopic Composition of CH4 From Rice Fielda
Throughout the 2008 fallow season Throughout the 2009 rice season Throughout the 2008 fallow season Throughout the 2009 rice season Throughout the 2008 fallow season Throughout the 2009 rice season
−56 to −44 −68 to −48 −68 to −60 −68 to −54 −45 to −34 −50 to −40
 Measurement of carbon isotopic ratios is also useful for identifying processes of CH4 emission including production, oxidation and transport of CH4 from rice fields. It is principally possible to use values of 13C measured of CH4, CO2 and acetate to calculate relative contribution of each pathway to the total CH4 production and fraction of the CH4 oxidized [Tyler et al., 1997]. In rice paddies, CH4 is formed mainly by way of acetate fermentation and CO2 reduction [Takai, 1970]. However, the relative contribution of the two methanogenic pathways to total CH4 production can vary considerably from one environment to the other, such as different temperatures [Fey et al., 2004] or different habitats [Conrad, 1999; Conrad et al., 2002] and soils [Tyler et al., 1997; Yao and Conrad, 2000; Nakagawa et al., 2002], and is significantly affected by rice planting and rice cultivar [Tyler et al., 1997; Bilek et al., 1999]. Conversely, methanogenic pathways and effects of CH4 oxidation and transport are crucial for the extent of carbon isotope fractionation. Methanogenesis via CO2 reduction exhibits a much stronger fractionation factor than that via acetate fermentation [Whiticar, 1999], and methanotrophic bacteria consume 12CH4 faster than they do 13CH4, thus leaving the residual CH4 enriched in 13C [Whiticar, 1999; Venkiteswaran and Schiff, 2005]. Furthermore, as 12CH4 is more readily emitted than 13CH4, the remaining CH4 dissolved in the water becomes 13C enriched, as well as CH4 in the aerenchyma and the rhizosphere [Popp et al., 1999; Chanton, 2005].
 CH4 in soil pore water as a whole poses as a large CH4 reservoir in the paddy soil, contributing to CH4 emission from the rice fields [Alberto et al., 2000]. The production zone of CH4 in the bulk soil is considered to be the origin of CH4 in the paddy soil to some extent [Tyler et al., 1997; Bilek et al., 1999]. CH4 in floodwater, another kind of CH4entrapped in rice fields and transported through the oxidizing layer, namely, the soil-water interface, is supposed to be highly affected by methanotrophy [Tyler et al., 1997]. Measuring δ13C values of CH4 in various field compartments simultaneously should help us better understand the processes of CH4emissions from rice fields. In comparison of rice-growing season, on the other hand, it is expected to find large differences in pathways of CH4 production as well as CH4oxidation ratio in flooded rice fields all year-round, owing to no rice planting and relative low temperature in the fallow season.
 For these purposes, a field experiment was conducted from November 2008 to November 2009 in China to investigate carbon isotopic signatures of CH4 in soil pore water and floodwater of and emitted from a rice field under continuous flooding throughout the fallow and following rice seasons, and then to report its methanogenic pathway and fraction of the CH4 oxidized.
2. Methods and Measurements
2.1. Field Description and Experiment Design
 The field experiment was laid out with plots in triplicate, 3 × 4 m each, in a paddy field at Baitu Town, Jurong City, Jiangsu Province, China (31°58′N, 119°18′E). The soil of the rice field is classified as Typic Haplaquepts (USDA, 1975). Soil samples were collected from the top layer (0–0.15 m) for measurement of initial organic C, total N and isotopic signature of the soil organic carbon, which was found to be 0.96%, 0.10% and −27.4‰, respectively.
 After the rice in the plots was harvested on 2 November 2008, rice stubbles and weeds were all removed from the plots. Then the plots were kept flooded from 18 November 2008 to 15 October 2009 and drained prior to the next rice harvest on 16 October 2009. In 2009, all the plots were ploughed the way the local farmers do before rice transplanting. Rice seeds (Cultivar “Oryza sativa L. Huajing 3”) were sown in nursery beds on 25 May, seedlings transplanted into the plots on 26 June at their 3 to 4 leaf stage, and the crop harvested on 3 November. Urea was applied at a rate of 300 kg N ha−1, 50% as basal fertilizer on 26 June, 25% as tillering fertilizer on 17 July, and 25% as panicle fertilizer on 16 August. Both calcium superphosphate and potassium chloride were applied as basal fertilizer on 26 June at a rate of 450 kg ha−1 and 225 kg ha−1, respectively.
2.2. Field Sampling and Measuring
 The static chamber method was used to sample CH4 flux. Plastic bases for the chambers (0.5 × 0.5 × 1 m each) were installed on 12 November 2008 and kept there until rice harvest in November 2009. Removable wooden boardwalks (2 m long) were set up on 18 November 2008 to avoid disturbance of the soil during sampling for measurement. To measure CH4 flux, four gas samples were collected using 18 mL vacuum vials from each chamber at 15 min intervals during the period between 08:00 A.M. and 12:00 noon on each sampling day. The collection was repeated once every 7–10 days in the fallow season and every 4–7 days during the rice season. CH4 flux was figured out based on the linear increase in gas concentration of the samples with each sampling and adjusted in light of the volume and the bottom area of the sampling chamber. Mean CH4flux is the average of the triplicates weighted by the interval between two adjoining measurements. Soil temperature at 0.1 m depth was measured on each sampling day with a hand-carried digital thermometer (Yokogawa Meter and Instruments Corporation, Japan).
 After flux of CH4became detectable, two gas samples were taken at 10–15 day intervals with special sampling bags (Aluminum foil compound membrane, 0.5 L, Delin gas packing Co., Ltd, Dalian, China) using a small battery-driven pump for carbon isotopic composition measurement. After the final drainage, sampling at 1–2 day intervals was conducted to measure CH4 flux and corresponding δ13C value. The first sample was taken directly after closure of the chambers, while the second was taken at the end of the 1–2 h closure period. Isotopic signature (S) of the emitted CH4 was calculated as follows:
where A and B stands for CH4 concentrations (μL L−1) in the samples at the beginning and at the end, respectively, of the sampling period and a and b for corresponding δ13CH4 values (‰) of the gas samples.
 Soil pore water samples were collected at 15–30 day intervals using a Rhizon soil moisture sampler (10 RHIZON SMSMOM, Eijkelkamp Agrisearch Equipment, Giesbeek, Netherlands), which consists of a 0.1 m long microporous polymer tube (2.5 mm OD × 1.5 mm ID) and a 1 m long PVC hose (2.7 mm OD × 1 mm ID). The microporous polymer tube was buried horizontally in the soil, 0.1 m in depth, in each plot on 18 November 2008 and then left there throughout the fallow and following rice seasons. Prior to sampling, about 5 mL soil solution was extracted using 18 mL vacuum vial to flush and purge the sampler. Then approximately 10 mL of soil solution was drawn into another vial for analysis. Simultaneously, 10 mL of floodwater was collected using a plastic syringe and transferred into a 18 mL vacuum vial. Finally, the pressure in all the vials was equilibrated by filling in pure N2 gas. After vigorous shaking by hand, the air in the headspace of the vials was directly analyzed for CH4on a GC-FID. Their correspondingδ13CH4 values were then determined using an isotope ratio mass spectrometer. CH4 concentration (CCH4) in pore water and floodwater was calculated using the following equation:
where m stands for mixing ratio of CH4 in the headspace of the vial (μL L−1), MV for gas volume of an ideal gas (24.78 L mol−1 at 25°C), GV for volume of the headspace of the vial (L), and GL for volume of the liquid in the vial (L).
2.3. Measurement of Gas Concentrations and Isotope Ratios
 Concentration of CH4was analyzed with a gas chromatograph (Shimadzu GC-12A, Kyoto, Japan) equipped with a flame ionization detector (FID) and a 2 m Poropak Q (80/100 mesh) column [Cai et al., 2000]. The carbon isotopic composition was determined using the continuous flow technique on a Finnigan MAT 253 isotope ratio mass spectrometer (Thermo Finnigan, Bremen, Germany), which has a fully automated interface for pre-GC concentration (PreCon) of trace gases [Cao et al., 2008]. CH4 in gas samples, preconcentrated on the PreCon by He (carrier gas, 99.999% in purity, at a rate of 20 mL min−1) and then converted to CO2 in a combustion furnace at 1000°C, was flowed into GC for separation. After CO2 was separated from other components in GC, it went into the mass spectrometer for analysis of isotopic signature. CO2 reference gas (99.999% in purity, with a δ13CPDB value of −23.73‰) was injected 3 times into the ionic source continuously, at 30 s intervals. The CO2 peak occurred at about 870 s and the ratio line was a positive peak. On the basis of the ratios of the CO2 peak of the working standard to the intensity of the three ionic flows, δ13C of CO2 derived from the CH4 in the samples was figured out relative to the international carbon isotope standard PDB (δ13CPDB‰). The compressed air with 2.02 μL L−1 CH4 was regarded as the sample from a stable source, and the standard deviations of δ13C was ± 0.196‰ (n = 9). Isotope ratios were expressed in the standard delta notation:
where Rsample and Rstandard denotes the ratio of 13C/12C in the sample and in standard methane, respectively.
where δ13CH4(original) stands for carbon isotopic composition of primarily produced CH4, δ13CH4(oxidized) for carbon isotopic composition of residual CH4 after methanotrophy, which is calculated in terms of a semiempirical equation [Tyler et al., 1997]:
and αoxfor isotope fractionation factor, which is consumed by methane-oxidizing bacteria. Generally,αox is 1.013–1.049 [Zhang et al., 2009].
2.5. Statistical Analysis
 Statistical analysis was conducted with the aid of Software SPSS 15.0 for Windows (SPSS Inc., Chicago), and Standard deviation of means was calculated with software Microsoft Excel 2003 for Windows. Sample sets were rejected unless they yielded a linear regression value of r2 ≥ 0.90.
3.1. CH4 Concentration and Corresponding δ13CH4 and δ13CO2 in Pore Water
 CH4 concentration in pore water increased obviously in the fallow season and peaked on 5 June (Figure 1a). After rice transplanting, it descended markedly but soon turned upward to a second peak on 12 October. It ranged from 6 to 398 μM L−1 in the fallow season and from 21 to 336 μM L−1 during the following rice season. Pore water CH4 was relatively constant in δ13C value, around −65‰ in the early fallow season but increased to −60‰ in the late period (Figure 1b). After rice transplanting, the value increased markedly again to about −55‰ in the beginning and nearly unchanged between July and August. Nevertheless, it declined rapidly to −68‰ at the end. In general, pore water CH4 was relatively 13C depleted, with δ13C values about −65‰ for the most of the season. CO2 in pore water exhibited a rather heterogeneous temporal variation pattern in δ13C value (Figure 1c). In the fallow season, it gradually became 13C depleted and its δ13C value decreased from −15‰ to −21‰. However, in the rice season, the value increased again to −16‰.
3.2. CH4 Concentration and Corresponding δ13CH4 in Floodwater
 As expected, CH4 concentration in floodwater was very low (<7 μM L−1) over the season and was significantly lower than that in pore water (Figure 1a). CH4 in floodwater was also measured for δ13C value (Figure 1b). After a short 13C depletion in the fallow season, floodwater CH4 first became heavier, with δ13C values from −45‰ to −34‰. Subsequently, the value decreased to −45‰ at the end. After rice transplanting, it again increased during the season from −50‰ to −40‰. Obviously, floodwater CH4 was more 13C enriched in contrast to pore water CH4 during the whole observational period (Figure 1b).
3.3. CH4 Emission and Corresponding δ13CH4
 Pronounced temporal variation of CH4 flux was observed (Figure 2). In the fallow season, a large CH4 flux was observed just 20 days before rice transplanting. Afterward, it increased quickly at the beginning of the rice season and varied significantly, peaking (23.1–29.8 mg CH4 m−2 h−1) twice in July and August, separately. Final drainage before rice harvest led to a brief increase in flux, which was followed by a rapid decrease, probably because of the fact that the CH4 stored in the soil got released after disappearance of the floodwater layer (Figure 1a). The total CH4 emission in the fallow season was 3.3 g CH4 m−2, accounting for 8.9% of the annual total emission. Significant changes in soil temperature were also recorded (Figure 2), with a mean value of 13.2°C in the fallow season and 24.4°C during the rice season. The flux tallied well with soil temperature in temporal variation, and statistical analysis also showed a significant positive correlation between the two over the season (r = 0.635 and 0.732, respectively, P <0.01).
 Similar to the flux of CH4, δ13C value of emitted CH4 also varied apparently as is shown in Figure 1b. Since little flux was detected until April 2009, measurement of δ13CH4 began in the late fallow season. Before rice transplanting, it increased from −51‰ to −44‰ at the beginning and then dropped rapidly at the end. After rice transplanting, the emitted CH4 was 13C depleted with δ13C values being around −65‰ in most of the rice season. Late in the rice season, however, the emitted CH4 became heavier. The impact of the final drainage on δ13C value of emitted CH4 was remarkable (Figure 1b). Emitted CH4 was significantly enriched in 13C 1 day after the final drainage, with δ13C value increased markedly to −51‰. The value declined rapidly to −62‰ just 2 days later but ascended again to −48‰ in the 2 days that followed. Taken as a whole, the δ13C value of emitted CH4 was more positive in the fallow season (from −56‰ to −44‰) than in the following rice season (from −68‰ to −48‰). Notably, temporal variation of CH4 flux was negatively related to δ13C value of emitted CH4 (3Figure 3, r = −0.875, P <0.01).
3.4. Contribution of Acetate to Total CH4 Production (fac) and Fraction of CH4 Oxidized (fox)
 For comparison with what was obtained by Tyler et al.  and Bilek et al. , δ13CH4 (acetate) was set to be −40‰ and αCO2/CH4 to be 1.079–1.082 [Fey et al., 2004] in equations (5) and (7); furthermore, δ13C values of pore water CH4 (Figure 1b) were used as proxy for total produced CH4. It was found that hydrogenotrophic methanogenesis was an important methanogenic pathway over the entire fallow season (Figure 4a), with fac value being <0.5 in the early and about 0.6 in the late. After rice transplanting, acetoclastic methanogenesis took the place as the dominant pathway just between July and August, and its fac value was close to 0.7 but then decreased rapidly toward the end of the rice season. Corresponding to the δ13C value of pore water CH4, hydrogenotrophic methanogenesis was very important (around 50%) for the most of the season (Figure 4a).
 The δ13C value of floodwater CH4 (Figure 1b) in the fallow season was considered to be δ13CH4(oxidized). During the rice season, however, equation (9) and εtransport = −12.2‰ [Tyler et al., 1997] were used for estimating δ13CH4(oxidized) (why different δ13C values were used for δ13CH4(oxidized), please see section 4 below). Moreover, a value of 1.025–1.038 was used for αox [Coleman et al., 1981; Chanton and Liptay, 2000]. It was calculated in equation (8) that, in terms of percentage, approximately 60%–90% in the fallow season and 10%–80% of CH4 during the rice season was oxidized before it entered the atmosphere (Figure 4b). Obviously, fox value was high over the entire fallow season and at the end of rice season but relatively low between July and August. In general, it was higher in the fallow season than in the following rice season (Table 2).
Table 2. Contribution of Acetate to Total CH4 Production (fac) and Fraction of CH4 Oxidized (fox) in Comparison With Previous Measurements
 Numerous measurements show that the carbon isotopic composition of CH4 from rice fields varies greatly with the methods and types of sampling (Table 1). In principle, CH4 is more 13C depleted in pore water or gas bubbles from stirred bulk soil than in the aerenchyma of rice plants or in naturally released gas bubbles, because the former is less influenced by CH4 oxidation [Chanton et al., 1997; Krüger et al., 2002]. Furthermore, aerenchymatic CH4 is heavier than emitted CH4, mainly because of the fact that rice plants preferentially transport 12CH4, making the residual CH413C enriched in the aerenchyma [Chanton, 2005]. However, these results were mainly about the intermittently flooded rice fields or the rice-growing season (Table 1), and little about rice fields flooded for a whole year. Here we systematically investigated the carbon isotopic composition of CH4 from a field continuously flooded throughout the fallow and following rice seasons and quantified the pathways of CH4 production and the fraction of CH4 oxidized.
 In this study, although the δ13C value of pore water CH4 varied temporally (Figure 1b), it was negative and relatively lighter than most of the earlier measurements (Table 1), which might be attributed to changes of methanogenic pathways. The δ13C value of pore water CH4 varied slightly around −65‰ in the early fallow season, which suggests that acetoclastic methanogenesis during this period was relatively stable, and the fac value being constantly lower than 0.5 confirmed this. Subsequently, pore water CH4 was 13C enriched promptly in the late fallow season, indicating an obvious increase in the contribution of acetoclastic methanogenesis (Figure 4a). During the rice season it increased significantly and then remained at a high level between July and August, which is probably ascribed to an increase in CH4 production via acetate fermentation with the increase in root exudates. The fac value was found to be close to 0.7 as well (Figure 4a). Another potential reason would be the rice plants preferential transport 12CH4 of soil pore water into the atmosphere [Popp et al., 1999], because the emitted CH4 was more 13C depleted than the pore water CH4 during this period (Figure 1b). Late in the rice season, pore water CH4 became lighter gradually, indicating that though CH4 concentration in pore water increased during this period (Figure 1a) probably as a result of plant debris supplying abundant substrates for methanogenesis, it still depended mainly on hydrogenotrophic methanogenesis (Figure 4a).
 The process of CH4 production exhibits seasonal variation in the pathways of methanogenesis, namely, hydrogenotrophic methanogenesis was a very important pathway of CH4 production for the most of the season, and acetoclastic methanogenesis dominated only between July and August (Figure 4a). It might be related to the differences in soil temperature and rice cultivation. Since the field was unplanted and continuously flooded over the entire fallow season, CH4 production mainly occurred in the process of decomposition and degradation of endogenous soil organic matters, and a relatively low soil temperature (Figure 2) would be a limited factor for formation of acetate and hence CH4 production [Cai et al., 2009]. Conrad et al.  speculated that the degradation of organic matter which was incompletely mineralized would cause dominance of hydrogenotrophic methanogenesis. When the temperature rose high during the following rice season, especially between July and August (Figure 2), rice planting should become a more important determinant for pathway of methanogensis [Cai et al., 2009]. On the other hand, investigation of this study found something different from the previous measurements, that is, the contribution of acetoclastic methanogenesis was relatively lower in the rice field that was flooded year-round than in the fields flooded only during the rice-growing season (Table 2). It may be ascribed to the dominance of different methanogenic pathways in regions different in climates and environments, farming and cropping conditions, agricultural managements, rice varieties and soil properties, etc. [Cai et al., 2009].
 In the rice fields, CH4oxidation occurs mainly in the rhziosphere and at the soil-water interface [Conrad and Rothfuss, 1991; Denier van der Gon and Neue, 1996]. Consequently, the CH4 dissolved in soil pore water will be partially oxidized when it diffuses into the floodwater. Both Tyler et al.  and Bilek et al.  found dissolved CH4 increased in 13C enrichment with the decrease in soil depth. This suggests that for CH4produced in the zone, the closer to the soil surface, the more likely to be consumed by methane-oxidizing bacteria. Similarly, the effect of transport on isotopic fractionation would be the other probable reason [Popp et al., 1999]. In our measurements, floodwater CH4 was more 13C enriched than pore water CH4 over the season (Figure 1b), which was in agreement with the measurement of a greenhouse experiment conducted by Cheng et al.  during the rice season in Japan.
 The CH4 oxidation is expected to be a key natural process mitigating CH4 emission from rice fields, and earlier studies through analysis of stable carbon isotopic signatures have shown that 3%–71% of produced CH4 was oxidized (Table 2). As it is highly influenced by CH4oxidation at the soil-water interface, theδ13C value of CH4 in floodwater in the fallow season is regarded as δ13CH4(oxidized). During the rice season, however, rice plants become the leading conduit for CH4 emission, and CH4 oxidation occurs mainly in the rhizosphere [Denier van der Gon and Neue, 1996; Jia et al., 2001, 2002]. Equation (9) was hence used for estimating δ13CH4 (oxidized) [Tyler et al., 1997]. Moreover, αox is significantly affected by temperature and is 1.030–1.033 and 1.034–1.038 at 24 and 12°C, respectively [Chanton and Liptay, 2000; Chanton et al., 2008]. Coleman et al.  also found αox = 1.025 at 26°C, which was later on applied to field data [Tyler et al., 1997; Bilek et al., 1999]. Therefore, it was calculated using αox = 1.038 in the fallow season and 1.025 during the following rice season that around 60%–90% of CH4 was oxidized in the fallow season, which might be an important reason for the very low CH4 flux (Figure 2). Similarly, a relatively lower fraction of CH4 being oxidized, especially between July and August, was found to be the key factor for the considerable flux during the rice season (Figure 2). Additionally, compared to what was in the former reports, a generally higher fraction of CH4 oxidized was found (Table 2), which might be interpreted as more CH4 being emitted by rice plants than by bubble ebullition and molecular diffusion, thus causing less CH4 oxidation during the rice season than during the fallow season. A previous study showed that CH4emission flux was positively related with the efficiency of rice plant-mediated CH4 transport, but negatively with rhizospheric CH4 oxidation [Jia et al., 2002].
 The δ13C value of emitted CH4 was more positive in the fallow season than most of the available observations conducted during the rice season (Table 1). One possible explanation for the measurements in this study is that a fraction of CH4 is oxidized. Emitted CH4 was relatively enriched in 13C, probably because of a high fox value in the fallow season, whereas a low fox value in the most part might be a major cause for 13C depletion during the rice season (Figure 1b). With an increase in CH4 oxidation toward the end of the rice season, δ13C values of emitted CH4 increased again. Krüger et al.  found that a seasonal decrease in fox value resulted in a seasonal decrease in 13CH4. Another probability might lie in the influence of CH4 transport. CH4 transport is mainly accomplished through bubble ebullition and molecular diffusion in the fallow season [Tyler et al., 1997], and CH4 crossing the water–air interface results in low (2–3‰) carbon isotope fractionation [Happell et al., 1994, 1995; Chanton, 2005]. The difference between floodwater and emitted CH4 in δ13C value on 27 May and 5 June was calculated to be 3‰ and 7‰ (Figure 1b), respectively, indicating an analogous fractionation of 3–7‰ in CH4 transport. On the contrary, CH4 transport by plants brings about significant carbon isotope fractionation [Chanton, 2005], which would make emitted CH4 more 13C depleted. Previous studies have showed that CH4 flux through rice plants could lead to a fractionation of 11–18‰ by measuring δ13C values of emitted and aerenchymatic CH4 [Chanton et al., 1997; Tyler et al., 1997; Bilek et al., 1999; Krüger et al., 2002; Krüger and Frenzel, 2003]. Another experiment by the authors found a similar result (G. Zhang et al., unpublished data). Furthermore, flux was negatively related to δ13C value of emitted CH4 in temporal variation (Figure 3), suggesting that plant transport is probably a more important pathway and less CH4 oxidation occurs during the rice season than in the fallow season.
 After the final drainage, a sharp fluctuation in δ13C value of emitted CH4 should be paid considerable attention to (Figure 1b). It drastically increased 1 day after the drainage, which is mainly ascribed to the influence of CH4 oxidation. It is expected that draining of the rice field results in 13C enrichment [Tyler et al., 1994; Bergamaschi, 1997; Marik et al., 2002; Han et al., 2005; Rao et al., 2008] because of increased activity of methanotrophs both in and around the rhizosphere. Two days later, a significant decrease in δ13C value of emitted CH4was observed. A possible explanation for this might be that a substantial amount of soil-entrapped CH4 was released instantaneously at that time (Figure 2) and too quickly to be influenced by oxidation. The emitted CH4 was 13C enriched again as the days after drainage went on, which is most likely attributable to CH4 oxidation as well.
 By measuring δ13C values of different CH4 pools in a continuously flooded rice field throughout the fallow and following rice seasons, it was found that pore water CH4 was more 13C depleted over the season except between July and August; The δ13C value of CH4 was more positive in floodwater (from −50‰ to −34‰) than in pore water (from −68‰ to −54‰) over the season, which indicates that floodwater CH4 is highly oxidized; δ13C value of emitted CH4 was negatively related to flux in temporal variation. In addition, it was heavier in the fallow season (between −56‰ and −44‰) than in the rice season (between −68‰ and −48‰). This demonstrates that CH4 transport through rice plants is more important than through molecular diffusion, which brings about less methanotrophy during the rice season than during the fallow season.
 Further calculation of methanogenic pathway and fraction of CH4 oxidized by using isotopic data shows that hydrogenotrophic methanogenesis was very important for the most of the season, and acetoclastic methanogenesis became dominant only between July and August during the period of vigorous growth of rice plants; Approximately 60%–90% of produced CH4 was oxidized in the fallow season, while only 10%–80% during the following rice season. In short, a low contribution of acetoclastic methanogenesis and a high fraction of CH4 oxidized in the field were found, especially in the fallow season.
 The authors are grateful to Yong Han and Yacheng Cao, etc., at the Institute of Soil Science, Chinese Academy of Sciences, Nanjing, China, for their help in isotopic data analysis. The authors are also grateful to our local collaborators for field assistance and to our anonymous reviewers for their helpful comments on the article. Funding for this research was provided by the National Natural Sciences Foundation of China (grants 40921061 and 40971154) and the Ministry of Science and Technology of China (S2010GR0080).