|Home | About | Journals | Submit | Contact Us | Français|
Although bubble ebullition through water in rice paddy fields dominates direct methane (CH4) emissions from paddy soil to the atmosphere in tropical regions, the temporal changes and regulating factors of this ebullition are poorly understood. Bubbles in a submerged paddy soil also contain high concentrations of carbon dioxide (CO2), implying that CO2 ebullition may occur in addition to CH4 ebullition. We investigated the dynamics of CH4 and CO2 ebullition in tropical rice paddy fields using an automated closed chamber installed between rice plants. Abrupt increases in CH4 concentrations occurred by bubble ebullition. The CO2 concentration in the chamber air suddenly increased at the same time, which indicated that CO2 ebullition was also occurring. The CH4 and CO2 emissions by bubble ebullition were correlated with falling atmospheric pressure and increasing soil surface temperature. The relative contribution of CH4 and CO2 ebullitions to the daily total emissions was 95–97% and 13–35%, respectively.
Understanding the dynamics of methane (CH4) and carbon dioxide (CO2) fluxes in rice paddy fields is crucial for improving the accuracy of estimating CH4 and CO2 emissions from global rice paddy fields. In particular, flooded rice paddies are considered to be a major source of anthropogenic CH4. Methane emissions from rice paddies in tropical Asian countries account for 90% of global annual CH4 emissions from rice paddies [1, 2].
Methane produced in an anaerobic-flooded paddy soil is mainly transported to the atmosphere through the aerenchyma of rice plants [3–5]. Such emissions through the aerenchyma are estimated to account for 48–85% of net CH4 emissions throughout the rice-cropping season . In addition, CO2 exchange in paddy fields mainly results from photosynthesis and respiration of rice plants, as well as soil microbial respiration.
Also, some of the CH4 and CO2 produced in rice field soil is directly emitted to the atmosphere through paddy water. In one study, when rice straw was applied to a paddy field, CH4 emissions via bubble ebullition from the soil accounted for 35–62% of total CH4 emissions . However, research on the direct CH4 and CO2 exchanges between paddy soil and the atmosphere, via paddy water, is limited and so further studies are required on these emissions, as has been noted by other researchers [7, 8].
Methane in paddy soil is transported to the atmosphere through paddy water by two pathways: (1) diffusion between soil and atmosphere and (2) bubble ebullition . Methane emission by bubble ebullition is considered to be greater than that by diffusion from paddy water . The bubbles usually contain a high concentration of CH4 ranging between 1 and 82% (v/v) [10, 11] and comprise most of the total CH4 pool in flooded paddy soil . Bubble production and ebullition are enhanced by applied organic materials during the initial plant growth period [6, 13, 14] and by organic substances originating from rice roots during later growth stages [6, 12, 14]. Although the variation of CH4 bubble ebullition during the cultivation period has been studied previously, the factors controlling the diurnal changes in CH4 ebullition remain unclear .
Methane ebullition from submerged peatlands, which are similar to flooded paddy soil in that they contain many bubbles, is controlled by atmospheric pressure, soil temperature, and water table level [16–19]. Falling atmospheric pressure has been shown to be the most important contributor to CH4 bubbling in peatlands [18, 19]. A study in rice paddy fields in Thailand also suggested that CH4 ebullitions occurred when atmospheric pressure dropped, but further research is needed to clarify this .
In contrast, CO2 exchange through paddy water is the result of photosynthesis of aquatic plants and respiration of both the plants and the soil microorganisms . Emission due to soil respiration is suppressed by paddy water during flood irrigation [21, 22], but the CO2 concentration in soil bubbles is between 2.2 and 13.0% (v/v) [11, 23], which suggests that bubble ebullition will release both CH4 and CO2 from paddy soil into the atmosphere.
Therefore, in this paper, we examined the dynamics of both CH4 and CO2 ebullition in tropical rice paddy fields in Thailand using an automatically closing chamber method.
Gas field measurements were conducted on September 20th and 21st, 2014, in a rice field of Kasetsart University, Kamphaeng Saen campus (14°00′33′′N, 99°59′03′′E) located in Nakhon Pathom Province, Thailand. The soil had a clay texture (65.7% clay, 23.30% silt, and 11.0% sand) with a dry bulk density of 1.69 gm−3. The soil was sampled on September 17 and had a pH of 6.0 (1:1 for soil:water), 4.32% organic matter, 1.81% total carbon, and 1.85% total nitrogen. Seedlings of the rice variety “Homcholasit” were transplanted on June 30 at 18 × 30cm spacing with 4-5 seedlings per hill, after the soil had been plowed on June 17 and 26 when weeds and rice plants that had grown during the fallow period were plowed into the soil. The rice plants headed on September 22 and were harvested on October 28. The paddy field was continuously flooded from June 17 until harvest, with flooding water depth maintained at 2–20cm. During the gas measurement period, the water depth slowly decreased from 5.5 to 4cm because there was no precipitation or irrigation.
The CH4 and CO2 fluxes were measured using the automatic closed chamber method. A customized-bottomless polycarbonate chamber (50 × 20cm at the base and 40cm height, Green Blue Corp., Tokyo, Japan) was placed between the rows of rice plants on August 8; the base part was inserted 4.5cm deep into the paddy soil (Figure 1). The lid of the chamber was automatically closed for 10min every 1h by a pneumatic piston, with the lid kept open for the rest of the time. A small electric fan was installed on the upper sidewall inside the chamber and was kept running throughout the experiment to uniformly mix the air within the chamber. The chamber headspace air was circulated at 500mLmin−1 (using a diaphragm pump; TD-4X2N, Brailsford Co., Rye, NY, USA) between the chamber and a 250mL buffer tank placed in a shed located approximately 4m away from the chamber to minimize the high frequency noise. A loop line was installed between the buffer tank and a wavelength-scanned cavity ring-down spectroscopy CH4/CO2 analyzer (G2201-i, Picarro Inc., Santa Clara, CA, USA). Air in the buffer tank was withdrawn to the analyzer at a flow rate of ~25mLmin−1 using another diaphragm pump (UN84.4 ANDC-B, KNF Neuberger Inc., NJ, USA) and then returned to the loop line. Concentrations of CH4 and CO2 were analyzed at approximately 3.6s intervals by the gas analyzer. The sampled air was dried before entering the gas analyzer using a reflux method with a membrane dryer (SWG-A01-06, Asahi Glass Engineering Co., Chiba, Japan) so that the water vapor concentration in the air was kept <0.1%. Based on the internal volumes of the buffer tank and connecting tube and the flow rate the air inside the chamber was calculated to first reach the gas analyzer 2min after closing the chamber lid. The measurements of CH4 and CO2 concentrations in the chamber air stopped when the chamber lid opened meaning that a measurement cycle of gas flux measurements lasted 8min every hour.
Temporal changes in CH4 concentration in the chamber during a measurement cycle were categorized into either a sudden increase (Figures 2(a), 2(c), and 2(e)) or a slow-constant increase (Figures 2(c) and 2(e)). Emission by bubble ebullition events was defined as a sudden increase in concentration (ΔC/Δt) of ≥0.29ppmmin−1, whereas emission by diffusion was defined as a slow-constant increase (ΔC/Δt) of <0.29ppmmin−1.
Changes in CO2 concentration in the chamber showed either an episodic increase accompanied by CH4 ebullition events (Figures 2(b) and 2(f)), a steady increase (Figure 2(d)), or a decrease by plant uptake (Figure 2(f)). CO2 emission by bubble ebullition was defined as episodic CO2 concentration increases accompanied by CH4 ebullition, whereas emission by diffusion was defined as a constant CO2 increase. The CO2 uptake by photosynthetically active aquatic plants was defined by a decrease in CO2 concentration (Figure 2(f)) observed during the daytime on both days.
Since CH4 and CO2 concentrations in the chamber often changed episodically with time due to bubble ebullition events (Figures 2(a), 2(b), 2(c), 2(e), and 2(f)), CH4 and CO2 fluxes were calculated for each single flux event and then summed proportionately for the time of each event to give a total flux for each 8min measurement period. The start of each flux event was determined as the intersection between tangent lines at the inflection point of the time series of CH4 or CO2 concentrations (Figure 2). The end of each event was the time just before the start of the next flux event or the end of the 8min measurement period (Figure 2). The gas flux F (mgm−2h−1) was calculated using temporal changes in gas concentrations as 
where V is the headspace volume within the chamber (m3), A is the water-surface area covered by the chamber (m2), t is elapsed time (h), and C(t) is temporal changes in gas concentration (mgm−3) expressed as
where C max is the maximum gas concentration (mgm−3), C 0 is the initial gas concentration (mgm−3), and k is a rate constant. The values of C max, C 0, and k were iteratively obtained using the data of observed gas concentration versus time. Substituting (2) at t = 0 into (1) means that the gas flux F (mgm−2h−1) can be calculated as 
Atmospheric pressure and air temperature were measured with a barometer (MPXAZ6115A and MPXV7007DP, Freescale Inc., TX, USA) and a thermometer (HMP45A, Vaisala Inc., Helsinki, Finland), respectively. Water depth in the rice field was measured with a water level sensor (eTape Continuous Fluid Level Sensor, Milone Technologies Inc., NJ, USA). Soil surface temperature was measured with a type T thermocouple.
Bubbles in soil were collected directly with a syringe by disturbing the topsoil at 3 p.m. local time on September 20. The CH4 and CO2 concentrations in the bubbles were measured using the CH4/CO2 gas analyzer after the sampled air was diluted 101 times with high-purity nitrogen gas.
Episodic and rapid increases in CH4 concentration were identified in 21 out of the 46 measurements (Figures 2(a), 2(c), and 2(e)). These sudden increases in CH4 concentration are likely to be from bubbles released from the soil to the atmosphere because the CH4 concentration in topsoil bubbles was as high as 63.73%v/v. In the other 25 measurements, the CH4 concentration in the chamber air increased gradually with time (ΔCH4/Δt < 0.29ppmmin−1) during the closure period, indicating that CH4 was released from the water surface to the atmosphere by molecular diffusion. The CH4 fluxes at the water surface fluctuated between 0.7 and 218.7mgm−2h−1 on the observation days, which are similar to previously reported values of −0.6–192.0mgm−2h−1 .
The large CH4 emissions via bubble ebullition mainly occurred between 10:00 a.m. and 5:00 p.m. local time (Figure 3(a)). During this period, atmospheric pressure markedly decreased and reached a minimum value (Figure 3(b)). A night-time CH4 ebullition event also occurred at 2:50 a.m. local time on September 21 (Figures 2(c), 3(a), and 3(b)), once again when air pressure decreased. There was a significant negative linear correlation between atmospheric pressure and log10-CH4 emission by bubble ebullition (Figure 4; r = −0.77, p < 0.001). These results suggesting that decreases in atmospheric pressure triggered the CH4 ebullitions in the tropical rice paddy field are supported by the findings of Tokida et al. [18, 19] who reported that decreases in atmospheric pressure triggered CH4 ebullitions in peatlands.
In peatlands, air pressure reduction expands bubble volume and thereby enhances bubble buoyancy which causes the bubbles to rise to the water surface . Reduced air pressure also increases the CH4 concentration of gas bubbles by degassing dissolved CH4 in soil solution [16, 26, 27]. These factors probably caused the higher CH4 emissions via ebullition that were found in the current study. Moreover, the higher CH4 ebullition emissions in the daytime, compared with nighttime, are probably due to larger decreases in daytime atmospheric pressure which would increase the volume of the bubbles and the CH4 concentration.
Rising soil temperature also increases the buoyancy and CH4 concentration of bubbles as barometric pressure decreases [17, 27]. In the current study, soil surface temperature increased from around 6:30 a.m. and reached a maximum value at 3:00–3:30 p.m. on each day (Figure 3(b)). This period approximately corresponded to that when CH4 ebullition events frequently occurred. The positive and significant correlation between soil surface temperature and log10-CH4 emission via bubble ebullition (r = 0.66; p < 0.005; Figure 4(b)) indicates that the increase in soil surface temperature contributed to CH4 ebullitions in the daytime. Ebullition events occurred at 8:50 a.m. on both days and at 9:50 a.m. on September 21, even though atmospheric pressure did not fall between 6:30 a.m. and 10:00 a.m. on either day. These ebullitions indicate that the rising soil temperature principally triggered the release of bubbles at those times. Rising soil temperature also has a role in enhancing methanogenic activities, leading to increases in CH4 production in soil . Therefore such increased biological activities might have also increased the CH4 concentration in the bubbles.
CH4 emission via bubble ebullition (546–617mgm−2d−1) contributed 95-96% of total daily CH4 emission (567–647mgm−2d−1) through flooded water (Table 1). These CH4 ebullitions mainly occurred in the daytime and were associated with falling atmospheric pressure and increasing soil temperature, as discussed above (Figures 3(a) and 3(b)). In contrast, CH4 emission by diffusion (21–30mgm−2d−1) accounted for only 3.7–4.7% of total daily CH4 emission from flooded water (Table 1). The CH4 emissions by diffusion were mostly observed at nighttime when soil temperature decreased (Figures 3(a) and 3(b)). Therefore, these results clearly show that CH4 emission in rice paddy fields is predominant by daytime ebullition from flooded water with much lower CH4 emissions at nighttime by diffusion.
Episodic increases in CO2 concentration were found in 14 of the 21 measurements when CH4 ebullition events occurred. During these 14 chamber closure periods, the CO2 concentration in the chamber air increased abruptly (Figures 2(b) and 2(f)) at about the same time as CH4 concentration increased (Figures 2(a) and 2(e)). These similar patterns indicate that CO2 was released to the atmosphere in the bubbles along with the CH4. In the other 7 measurements, there was a steady increase in CO2 concentration but no episodic increase, as shown in Figure 2(d), while CH4 concentration abruptly increased (Figure 2(c)). This suggests these bubbles did not contain much CO2.
CO2 uptake via the photosynthetic activities of the aquatic plants was also observed in these measurements. In the other 25 measurements, there was a transfer of CO2 from flooded water to the atmosphere by diffusion, likely due to the gradient in CO2 concentration at the interface between the flooded water and the atmosphere and also due to respiration of the aquatic plants . The values of CO2 fluxes ranged between −120.4 and 196.2mgm−2h−1 which are within the previously reported range of −285.1 to 459.4mgm−2h−1 .
On September 20, most of the CO2 fluxes were outgoing emissions due to bubble ebullitions. The highest CO2 emission (196.2mgm−2h−1) occurred at 2:50 p.m. (Figures 2(d) and 3(a)), coinciding with a high CO2 concentration in the bubbles of up to 11.74% (v/v). However, at 1:50 p.m., there was a negative (incoming) CO2 flux, even though there was a CO2 ebullition event. This overall negative flux must have been due to the fact that CO2 uptake by photosynthesis of the aquatic plants exceeded emissions by bubble ebullition, as shown in Figure 2(f) for CO2 transfer.
During the daytime on September 21, the CO2 fluxes mainly showed negative values even though CO2 ebullition events were observed. Therefore, this indicates that CO2 assimilation by the aquatic plants dominated CO2 fluxes on that day.
The log10-CO2 emissions by bubble ebullitions, omitting measurements with evidence of absorption by plant photosynthesis, were significantly correlated to changes in atmospheric pressure (r = −0.72; p < 0.05; Figure 4(c)) and soil surface temperature (r = 0.72; p < 0.05; Figure 4(d)). This indicates that these two environmental factors control CO2 ebullition in addition to CH4 ebullition. As previously discussed, these two triggered expanding bubble volume and degassing of gas dissolved in soil solution [16, 17]. In addition, the soil surface temperature was between 27 and 40°C during the measuring period which was optimal for respiratory soil microbes in the submerged paddy soil . Therefore, all these factors probably enhanced CO2 bubble ebullitions.
CO2 emission by bubble ebullition, accounted for only 13–35% of total CO2 emissions, compared with 65–87% from CO2 diffusion (Table 2), indicating that CO2 ebullition did not dominate CO2 emissions from flooded water unlike CH4 ebullition. This is probably due to the fact that CO2 uptake by aquatic plants would have exceeded CO2 emission by bubble ebullition. Moreover, the very low concentration of CO2 in the bubbles also contributed to lower CO2 emission by bubble ebullition. The CO2 emissions by diffusion mostly occurred at nighttime just like for CH4. The nighttime CO2 emissions by diffusion were mostly attributed to the gradient in CO2 concentrations between the atmosphere and the flooded water and also to CO2 respiration by small aquatic plants .
Our study found that daytime CH4 ebullition events in tropical rice paddy fields occurred due to falling atmospheric pressure and increasing soil surface temperature. At nighttime, the drop in atmospheric pressure predominately triggered the CH4 ebullition because soil temperature was low compared with that in the daytime. The fact that CH4 and CO2 concentrations in the chamber air increased abruptly when bubbles were released suggests that bubble ebullition events caused not only CH4 emission but also CO2 emission. The CO2 ebullition events were also controlled by decreases in air pressure and increases in soil temperature. Therefore, diurnal changes in atmospheric pressure and soil temperature play major roles in regulating CH4 and CO2 ebullitions in tropical rice paddy fields.
We also found that CH4 emission was predominant due to daytime ebullition, whereas only a small proportion of CO2 emissions was due to daytime ebullition. The low CO2 ebullition throughout the day was due to CO2 photosynthesis and respiration by aquatic plants, meaning that CO2 emission was mainly by diffusion between flooded water and the atmosphere.
This research was partly supported by Grant-in-Aid for Scientific Research (A) (25252044, PI: K. Noborio), a JSPS Fellowship (DC1, 12J10924, for S. Komiya) by the Japan Society for the Promotion of Science, and a Program for Establishing Strategic Research Foundations in Private Universities (S0901028, PI: K. Noborio) by MEXT of Japan. The authors are grateful to Dr. Jonaliza Siangliw (BIOTEC) and Ms. Rungthip Kohkhoo for their support at our experimental field, to Dr. Fumiyoshi Kondo (National Institute for Environmental Studies), Dr. Takeshi Tokida, and Dr. Seiichiro Yonemura (National Institute for Agro-Environmental Sciences) for their valuable comments, to Dr. Iain McTaggart (Meiji University) for reviewing a draft, to Dr. Masaru Mizoguchi (University of Tokyo) for making the water level sensor, to Mr. Ryoji Taniyama (Takumi Technical Laboratory Inc., Japan) for assisting with data analysis, to Mr. Ryo Higuchi for making the sensors, and to Mr. Shinsuke Aoki, Mr. Naoto Sato, and Mr. Ryuta Honda for analyzing soil samples. Experimental information and data are available on request to K. Noborio.
The authors declare that there is no conflict of interests regarding the publication of this paper.