Elevated atmospheric CO2 reduces CH4 and N2O emissions under two contrasting rice cultivars from a subtropical paddy field in China

Haiyang YUGuangbin ZHANGJing MATianyu WANGKaifu SONGQiong HUANGChunwu ZHUQian JIANGJianguo ZHU and Hua XU

1State KeyLaboratoryof Soil and Sustainable Agriculture,Institute of Soil Science,Chinese Academyof Sciences,Nanjing 210008(China)2Universityof Chinese Academyof Sciences,Beijing 100049(China)

ABSTRACT Elevated CO2 (eCO2)and rice cultivars can strongly alter CH4 and N2O emissions from paddy fields.However,detailed information on how their interaction affects greenhouse gas fluxes in the field is still lacking. In this study, we investigated CH4 and N2O emissions and rice growth under two contrasting rice cultivars(the strongly and weakly responsive cultivars)in response to eCO2,200 μmol mol-1 higher than the ambient CO2 (aCO2),in Chinese subtropical rice systems relying on a multi-year in-situ free-air CO2 enrichment platform from 2016to 2018.The results showed that compared to aCO2,eCO2 increased rice yield by 7%–31%,while it decreased seasonal cumulative CH4 and N2O emissions by 11%–59%and 33%–70%,respectively,regardless of rice cultivar.The decrease in CH4 emissions under eCO2 was possibly ascribed to the lower CH4 production potential(MPP)and the higher CH4 oxidation potential(MOP)correlated with the higher soil redox potential(Eh)and O2 concentration([O2])in the surface soil.The mitigating effect of eCO2 on N2O emissions was likely associated with the reduction of soil soluble N content.The strongly responsive cultivars had lower CH4 and N2O emissions than the weakly responsive cultivars,and the main reason might be that the former induced higher soil Eh and[O2]in the surface soil and had larger plant biomass and greater N uptake.The findings indicated that breeding strongly responsive cultivars with the potential for greater rice production and lower greenhouse gas emissions is an effective agricultural practice to ensure food security and environmental sustainability under future climate change scenarios.

KeyWords: climate change,free-air CO2 enrichment,greenhouse gas emission,methane oxidation potential,methane production potential,soil oxygen,soil redox potential

INTRODUCTION

Rice fileds produce staple food for more than 50% of the global population,and the rice demand is projected to continually increase by approximately 30%by 2050 due to population growth and economic development(Alexandratos and Bruinsma,2012).However,rice fileds are also important sources of CH4and N2O emissions, accounting for 11%and 10%of anthropogenic emissions,respectively(Saunoiset al.,2020;Wanget al.,2020).As the critical greenhouse gas(GHG),the atmospheric CO2concentration has reached 410 μmol mol-1in 2020,which exceeds the pre-industrial levels by about 168%(http://www.esrl.noaa.gov/gmd/ccgg/trends/). This elevated CO2(eCO2) concentration could stimulate grain yield (Allenet al., 2020) and alter CH4and N2O emissions (Zhenget al., 2006; Wang Cet al.,2018a, b; Yaoet al., 2021). For example, a 13%increase in rice production (Ainsworth, 2008; Allenet al., 2020)was followed by 34%and 10%increases in CH4and N2O emissions, respectively, under eCO2(Liuet al., 2018).In other words,rice cultivation practices under eCO2are expected to improve rice yields,accompanied by an increase in CH4and N2O emissions.

Breeding high-yielding rice cultivars is a key measure to meet the increasing global food demand (Jianget al.,2017).Some rice cultivars have been found to enhance grain yield in response to eCO2, and breeding such ‘positively responsive’rice cultivars has received considerable attention recently as a strategy to increase rice production in response to climate change(Huet al.,2020).For instance,the cultivars weakly responsive to eCO2showed a yield growth rate of 10%–15%(Kimet al.,2003;Yanget al.,2006,2007),while the strongly responsive cultivars significantly increased rice grain yield by more than 30%(Liuet al.,2008;Yanget al.,2009). Accordingly, the strongly responsive cultivars will be probably preferred in the future to meet the rising rice consumption due to the growing population.However,the breeding strategies, as well as eCO2, may influence CH4and N2O emissions from fields(Louet al.,2008;Wang Bet al.,2018).Previous studies have mainly focused on the response of CH4and N2O emissions to eCO2from rice fields of weakly responsive cultivars(Inubushiet al.,2003;Zhenget al.,2006;Xieet al.,2012).Whether the effects of eCO2on CH4and N2O emissions with the strongly responsive cultivar and the underlying mechanisms differed from those of the weakly responsive cultivars has not yet been documented.

Most studies indicate that eCO2stimulates CH4emissions from rice fileds due to enhanced plant growth,more release of root exudates, and increased abundance of methanogens for CH4production(Inubushiet al.,2003;Bhattacharyyaet al.,2013; Wang Cet al.,2018a; Qianet al.,2020).However,eCO2could also lead to the attenuation of CH4production due to more O2being delivered to the rice rhizosphere,which is caused by the increased underground biomass and porosity(Schropeet al.,1999).The quantity and quality of root exudates and soil porosity are different from rice cultivars,which play an important role in the production,oxidation,and transportation of CH4from rice fields(Lin and You, 1989; Lin, 1993). As such, eCO2and rice cultivars are likely to interact in determining CH4emissions from rice fileds.Moreover,the strongly responsive cultivars might deliver more O2to the rice rhizosphere because of their higher biomass and porosity.Thus,we hypothesized that,compared with the weakly responsive cultivars,CH4emissions under the strongly responsive cultivars would be decreased by eCO2because of the higher O2concentration([O2])in the surface soil.

The effects of eCO2on N2O emissions are inconsistent.For instance,N2O emissions have been reported to increase(Bhattacharyyaet al., 2013; Pereiraet al., 2013; Wang Bet al.,2018;Wang Cet al.,2018b),decrease(Sunet al.,2018;Yaoet al., 2021), or remain unchanged (Xuet al., 2002;Chenget al.,2006)in response to eCO2.It should be noted that previous studies on the response of N2O emissions to eCO2have primarily focused on weakly responsive cultivars(Xuet al.,2002;Chenget al.,2006;Bhattacharyyaet al.,2013;Pereiraet al.,2013;Sunet al.,2018;Wang Bet al.,2018;Wang Cet al.,2018b),whereas studies on the strongly responsive cultivars are lacking. Under eCO2conditions,strongly responsive cultivars have higher nitrogen(N)uptake capacity,net photosynthetic assimilation,and larger plant biomass than the weakly responsive cultivars (Zhuet al.,2014). Thus, we hypothesized that eCO2might decrease much more N2O emissions under strongly responsive cultivars than under weakly responsive cultivars due to increased N uptake by plant roots.

Therefore,the goal of this study was to investigate the effects of eCO2on CH4and N2O emissions from rice fields under strongly responsive cultivars based on a free-air CO2enrichment(FACE)platform initiated in 2004.We expected that breeding strongly responsive cultivars would meet rice yield demands and weaken the contributions of CH4and N2O emissions from paddy fields to global climate change.

MATERIALS AND METHODS

Site description

A field experiment was conducted from 2016to 2018 at the FACE system located in Xiaoji Town(119°42′0′′E,32°35′34′′N),Yangzhou City,Jiangsu Province,in a typical Chinese rice-growing region with a typical northern subtropical monsoon climate(Zhuet al.,2014).The FACE platform,which was created in 2004,consists of three identical 14-m diameter octagonal rings receiving eCO2that is 200 μmol mol-1above the ambient CO2(aCO2)concentration.The target eCO2was achieved by injecting pure CO2into the FACE rings during daylight hours.More detailed descriptions of the design and operation of the FACE facility can be found in the following studies:Xieet al.(2012),Huet al.(2020),and Liet al.(2020).

Rice cultivation

Four strongly responsive cultivars selected in this experiment were Yangdao6(Y6,anindicarice cultivar planted in 2016and 2017),YIIyou900(Y900,a hybridindicarice cultivar planted in 2017),LongIIyou1988(L1988,a hybridindicarice cultivar planted in 2018),and Yongyou1540(Y1540,a hybridjaponicarice cultivar planted in 2018).Three weakly responsive cultivars were selected:Wuyungeng23(W23,ajaponicarice cultivar planted in 2016and 2017),Wuyungeng27(W27,ajaponicarice cultivar planted in 2017 and 2018),and Huaidao5(H5,ajaponicarice cultivar planted in 2018)(Table SI,see Supplementary Material for Table SI).All management practices,including water and fertilization regimes,in the experimental plots were consistent with those in the local area.

The seeds were grown under the aCO2conditions until they were transplanted at the three-leaf stage. According to the recommended density of rice cultivation,they were manually transplanted to the aCO2and eCO2rings at a density of two seedlings of the weakly responsive cultivars and one seedling of the strongly responsive cultivars per hill in late June in each rice season.The spacing of the hills was 16.7 cm× 25 cm (equivalent to 24 hills m-2). Nitrogen was supplied in the form of urea at 22.5 g m-2. In each season,N was applied as basal fertilizer one day before rice transplanting(40%of the total),as a top dressing at the early tillering stage(30%of the total),and at the panicle initiation stage(30%of the total).Phosphorus(9 g P2O5m-2)and potassium(9 g K2O m-2)were applied as basal fertilizers.

Biomass(aboveground and underground biomass)and grain yield were measured at crop harvest.To ensure representative sampling,16hills were randomly selected from each subplot.Thereafter,6hills were counted for effective tiller numbers.For each sampling,the biomass samples were oven-dried at 80°C for 72 h to a constant weight,and grain yields were determined by subtracting a moisture content of 0.14 g H2O g-1fresh weight(Liuet al.,2008).

Gas and soil sampling and measurements

The CH4and N2O fluxes were measured using a static chamber-gas chromatography(GC)method(Caiet al.,2009).The chambers were cuboid,35 cm in length and width,60 or 120 cm in height(according to plant height),and with a plastic base(35 cm×35 cm)installed before the initiation of the experiment.Gas was sampled at 4–5 d intervals during drainage and re-flooding, and at 7–10 d intervals for the rest of the sampling period. Four gas samples from each chamber were injected into the prepared 21-mL vacuum vials at an interval of 12 min between 8:30 and 11:30 a.m.on every sampling day.The samples were then taken to the laboratory to measure the concentrations of CH4and N2O using GC(Agilent 7890B,Agilent Technologies,Santa Clara,USA).The fluxes of CH4and N2O were calculated from the change in gas concentrations in the enclosed chamber over time. Seasonal cumulative CH4and N2O emissions were calculated directly from the measured fluxes.

When gas fluxes were monitored, soil redox potential(Eh)of the surface soils was simultaneously measured using a portable oxidation-reduction potential meter(Hirose Rika Co.Ltd.,Tokyo,Japan).Soil samples(0–15 cm)were collected during the main rice growth stage in each plot to analyze soil mineral N(NH+4-N and NO-3-N),dissolved organic C(DOC),CH4production potential(MPP),and CH4oxidation potential(MOP).

Soil NH+4-N and NO-3-N were extracted with 2 mol L-1KCl solution(soil:solution=1:5,weight:weight)by shaking for 60 min at 250 r min-1and were quantified colorimetrically using a continuous flow autoanalyzer(San++System,Skalar Analytical BV,Breda,the Netherlands).The DOC was extracted with 0.5 mol L-1K2SO4extracts(soil:solution=1:4,weight:weight)by shaking for 30 min at 250 r min-1,centrifuged for 15 min at 4 000 r min-1,and then filtered through a 0.45-μm polyethersulfone membrane filter.The extracts were determined using a TOC analyzer(Vario TOC Cube,Elementar,Hanau,Germany).

The MPP of fresh soil samples was determined anaerobically and calculated using the linear regression of CH4increasing with the incubation time. Approximately 25 g(dry weight) of soil was quickly transferred into 150-mL flasks. Sterile water flushed by N2was added into each flask to prepare the soil slurry with a soil/water ratio of 1:1 (weight:weight). All flasks were sealed with rubber plugs with a silicone septum,which allowed the sampling of headspace gas. The flasks used for MPP were flushed with N2six times to purge the remaining CH4and O2.They were then incubated at 25°C for 50 h in dark.Gas samples were collected twice with a pressure lock syringe at 1 and 50 h after the flasks were heavily shaken by hand,and then analyzed for CH4.

The MOP was determined aerobically using the same devices as described above but with headspace air in the flasks.About 1 mL of pure CH4was injected into each flask to obtain a high CH4concentration inside(approximately 10 000 μL L-1). The flasks were then incubated in dark at 25°C and shaken at 120 r min-1. The CH4depletion was measured by sampling the headspace gas in the flask after vigorous shaking for subsequent GC-flame ionization detection analysis. The first sample was collected 30 min after pure CH4was injected and homogeneously distributed inside the flask. Samples were then taken at 2 h intervals during the first 8 h of the experiment.The flasks were left overnight and sampled in the next day at 2 h intervals.CH4oxidation was calculated using linear regression of CH4depletion with incubation time.

The[O2]in floodwater and surface soil was measuredinsituusing a Unisense field microprofiling system(Unisense,Aarhus,Denmark)on July 18,July 28,and August 10 in 2018.The O2microelectrode is a miniaturized Clark-type sensor connected to a field microsensor multimeter(Li and Wang,2013;Huanget al.,2020).Then,the O2microelectrode was placed in the sense holder and inserted into the soil near the rice plant approximately 20 mm from the field motor.The stepwise sequence was set up in advance using a field microsensor multimeter.In our sequence,the path size was set to 2 mm,and the end depth was 20 mm below the soilwater interface.The periods for“wait before measure”and“measure”were both set to 3 s.

Statistical analysis

All statistical analyses were performed using SPSS software version 25.0 (SPSS Inc., Chicago, USA). Statistical significance was determined at 0.05 probability level,and the differences between means under eCO2and aCO2of the same rice cultivar were examined with Tukey’s honestly significant difference test.A two-way analysis of variance(CO2and rice cultivar)was used to identify factors associated with CH4and N2O emissions.

RESULTS

Rice biomass and tiller number

Relative to aCO2,eCO2increased the grain yield of the strongly responsive cultivars by 19%–31%(P ＜0.01)and that of the weakly responsive cultivars by 7%–14%(P ＞0.05)(Fig.1a,b).The eCO2increased aboveground and underground biomass of the strongly responsive cultivars by 20%and 18%,respectively,and that of the weakly responsive cultivars by 14%and 16%,respectively(Fig.1c,f).On average,grain yield,aboveground biomass,and underground biomass of the strongly responsive cultivars under eCO2were 15%(P ＜0.05),6%(P ＞0.05),and 2%(P ＞0.05),respectively, higher than those of the weakly responsive cultivars.The eCO2increased the tiller number(P ＞0.05),irrespective of the rice cultivar(Figs.1g,h,and S1,see the Supplementary Material for Fig.S1).

Fig.1 Effect of elevated CO2 (eCO2)on grain yield,aboveground biomass,underground biomass,and effective tiller number of strongly(a,c,e,and g)and weakly(b,d,f,and h)responsive rice cultivars planted in 2016–2018.Strongly responsive cultivars selected were Yangdao6(Y6),YIIyou900(Y900),LongIIyou1988(L1988),and Yongyou1540(Y1540),and weakly responsive cultivars were Wuyungeng23(W23),Wuyungeng27(W27),and Huaidao5(H5).Vertical bars indicate standard deviations of the means(n=3).Significant differences between eCO2 and ambient CO2 (aCO2)levels were determined by Tukey’s honestly significant difference test.The asterisks*and**indicate significant differences at P ＜0.05 and P ＜0.01,respectively.

CH4 and N2O emissions

Compared with the weakly responsive cultivars, CH4fluxes under the strongly responsive cultivars showed similar temporal patterns but varying amplitudes (Fig. 2). They increased rapidly from the time of rice transplanting until the end of July,when it reached the maximum,and sharply declined during the mid-season drainage stage.Thereafter,CH4fluxes reached the second peak during the intermittent irrigation stage and then declined to zero during the harvest stage.For all the strongly responsive cultivars,the highest CH4fluxes under eCO2conditions(17.9 to 24.0 mg CH4m-2h-1) were generally lower than those under aCO2conditions(20.5 to 31.0 mg CH4m-2h-1).

Fig.2 CH4 fluxes from fields of strongly(a,c,and e)and weakly(b,d,and f)responsive rice cultivars under ambient CO2 (aCO2)and elevated CO2(eCO2)conditions in 2016(a and b),2017(c and d),and 2018(e and f).Strongly responsive cultivars selected were Yangdao6(Y6),YIIyou900(Y900),LongIIyou1988(L1988),and Yongyou1540(Y1540),and weakly responsive cultivars were Wuyungeng23(W23),Wuyungeng27(W27),and Huaidao5(H5).Vertical bars indicate standard deviations of the means(n=3).

Cumulative CH4emissions under the strongly and weakly responsive cultivars varied from 115 to 197 kg CH4ha-1and from 93 to 312 kg CH4ha-1,respectively,under eCO2conditions and from 224 to 280 kg CH4ha-1and from 144 to 351 kg CH4ha-1,respectively,under aCO2conditions(Table I).On average,in the 3-year experiment,CH4emissions decreased more under eCO2from rice fields of the strongly responsive cultivars(-39%)than under eCO2from rice fields of the weakly responsive cultivars(-30%)(Table I).Cumulative CH4emissions decreased correlatively with increasing biomass and tiller numbers,induced by eCO2and affected by rice cultivar(Table SII,see Supplementary Material for Table SII).

Very low N2O fluxes were measured throughout most of the rice-growing season,except for a flux peak at the midseason drainage stage(Fig.3).Compared with the weakly responsive cultivars,the highest N2O fluxes under the strongly responsive cultivars were observed under aCO2conditions,ranging from 314 to 597 μg N2O-N m-2h-1.The fluxes of N2O for the strongly and weakly responsive cultivars under eCO2conditions were lower than those under aCO2conditions(Fig.3).The cumulative N2O emissions under eCO2conditions ranged from 0.25 to 0.41 kg N2O-N ha-1for the strongly responsive cultivars,and from 0.25 to 0.70 kg N2O-N ha-1for the weakly responsive cultivars. These values were lower than those observed for the strongly(0.63–0.92 kg N2O-N ha-1) and weakly (0.42–1.05 kg N2O-N ha-1)responsive cultivars under aCO2conditions(Table I).On average,eCO2significantly decreased cumulative N2O emissions for the strongly and weakly responsive cultivars by 60%and 43%,respectively(Table I).The cumulative N2O emissions were significantly correlated with rice biomass and tiller numbers,which were affected by CO2concentration and rice cultivar(Table SII).

Soil parameters

The seasonal mean surface soil Eh under eCO2for the strongly(-114 to-84 mV)and weakly(-125 to-84 mV)responsive cultivars was higher than that under aCO2for the strongly(-150 to-90 mV)and weakly(-164 to-93 mV)responsive cultivars(Figs.4a,b and S2,see Supplementary Material for Fig. S2). The soil Eh on average increased by 7%–9%for the strongly responsive cultivars compared with that for the weakly responsive cultivars. Generally,regardless of eCO2or aCO2,CH4fluxes for the strongly and weakly responsive cultivars were negatively associated with the dynamics of soil Eh(Table SII).

The eCO2increased soil DOC of the strongly and weakly responsive cultivars by-2% to 14% and by 1% to 32%,respectively(Fig.4c,d).There was no correlation between cumulative CH4emissions and the average concentration of DOC, but a significant negative correlation was observed between N2O emissions and DOC across all rice cultivars during the rice-growing season from 2017 to 2018(Table SII).

TABLE I Accumulative CH4 and N2O emissions from fields of strongly and weakly responsive rice cultivars affected by elevated CO2 (eCO2)relative to ambient CO2(aCO2)in 2016–2018

Fig.3 N2O fluxes from fields of strongly(a and c)and weakly(b and d)responsive rice cultivars under ambient CO2 (aCO2)and elevated CO2 (eCO2)conditions in 2017(a and b)and 2018(c and d).Strongly responsive cultivars selected were Yangdao6(Y6),YIIyou900(Y900),LongIIyou1988(L1988),and Yongyou1540(Y1540),and weakly responsive cultivars were Wuyungeng23(W23),Wuyungeng27(W27),and Huaidao5(H5).Vertical bars indicate standard deviations of the means(n=3).

Compared to aCO2,eCO2reduced the average content of soil NH+4-N of the strongly and weakly responsive cultivars by 1%to 44%and by 23%to 33%,respectively(Fig.4e,f).Generally,the content of NO-3-N was one order of magnitude lower than that of NH+4-N, and there was a significant difference in NO-3-N content between eCO2and aCO2for the strongly responsive cultivars L1988 and Y1540 in 2018(Fig.4 g,h).The N2O emissions were positively correlated(P ＜0.01)with NH+4-N content(Table SII).

MPP and MOP

Fig. 4 Effect of elevated CO2 (eCO2) on mean surface soil redox potential (Eh), dissolved organic carbon (DOC), NH+4 -N, and NO-3 -N in fields of strongly(a,c,e,and g)and weakly(b,d,f,and h)responsive rice cultivars.Strongly responsive cultivars selected were Yangdao6(Y6),YIIyou900(Y900),LongIIyou1988(L1988),and Yongyou1540(Y1540),and weakly responsive cultivars were Wuyungeng23(W23),Wuyungeng27(W27),and Huaidao5(H5).Vertical bars indicate standard deviations of the means(n=3).Significant differences between eCO2 and ambient CO2 (aCO2)levels were determined by Tukey’s honestly significant difference test.The asterisks*and**indicate significant differences at P ＜0.05 and P ＜0.01,respectively.

The mean MPP for the strongly responsive cultivars decreased from 0.15–0.39 μg CH4g-1d-1under aCO2to 0.06–0.28 μg CH4g-1d-1under eCO2,and for the weakly responsive cultivars,it decreased from 0.23–0.63 μg CH4g-1d-1under aCO2to 0.06–0.25 μg CH4g-1d-1under eCO2in 2016–2018(Table II).Compared to aCO2,eCO2decreased the MPP for the strongly and weakly responsive cultivars by 17%–81%and 54%–86%,respectively.The MPP was significantly positively correlated with CH4emissions(Table SII).Contrary to changes in MPP,eCO2tended to increase the mean MOP in rice paddies of strongly and weakly responsive cultivars by 37%and 22%,respectively(Table II).Although MOP was not significantly affected by eCO2or cultivar,it was significantly negatively correlated with CH4emissions(Table SII).

Concentration of O2 in floodwater and surface soil

The eCO2significantly increased(P ＜0.01)the mean[O2]in 0–20 mm soil of the strongly and weakly responsive cultivars by 24%–37%and 22%–29%,respectively(Fig.5).Compared with the weakly responsive cultivars, the [O2]under the strongly responsive cultivars increased(P ＜0.01)by 19% under aCO2and 23% under eCO2. In the floodwater of the rice fields,the[O2]was relatively stable(Fig.S3, see Supplementary Material for Fig. S3). However, it rapidly decreased with soil depth,reaching near 0 μmol L-1at approximately 10 mm below the soil-water interface(Fig.S3).

TABLE II Mean CH4 production potential(MPP)and CH4 oxidation potential(MOP)in fields of strongly and weakly responsive rice cultivars affected by elevated CO2 (eCO2)relative to ambient CO2 (aCO2)in 2016–2018

Fig.5 Effect of elevated CO2 (eCO2)on mean surface soil O2 concentration([O2])in fields of strongly(LongIIyou1988,L1988;Yongyou1540,Y1540)and weakly(Wuyungeng27,W27;Huaidao5,H5)responsive rice cultivars.Vertical bars indicate standard deviations of the means(n=3).Significant differences between eCO2 and ambient CO2(aCO2)levels were determined by Tukey’s honestly significant difference test.The asterisk**indicates significant difference at P ＜0.01.

DISCUSSION

Effects of stronglyand weaklyresponsive rice cultivars on CH4 emissions under eCO2

In this first-ever study on the effects of eCO2on GHG emissions from rice fields with strongly responsive cultivars using FACE technology,eCO2reduced more CH4emissions from rice fields of the strongly responsive cultivars(16%–59%)than of weakly responsive cultivars(11%–54%)(Fig.2,Table I).However, regardless of rice cultivar,this decline in CH4emissions under eCO2conditions was in contrast with previous studies(Inubushiet al.,2003; Zhenget al.,2006;Bhattacharyyaet al.,2013;Pereiraet al.,2013;Wang Bet al.,2018;Wang Cet al.,2018a),and even differed from a non-significant increase(15%)of the seasonal total CH4emissions from the same experimental fields reported by Xieet al.(2012).The reduction in CH4emissions induced by eCO2from rice field soil might be attributed to higher surface soil Eh and[O2]due to eCO2-induced production of more tillers and rice biomass than those under aCO2(Figs.4,5,and S1–S3).These responses were significantly correlated with the lower MPP and higher MOP under eCO2relative to those under aCO2(Table SII).Although there was only an upward trend for MOP in all treatments under eCO2,the significant positive relationship between the lower MPP and reduced CH4emissions further suggests that decreased CH4production is the key factor for decreased CH4emissions.Meanwhile,CH4emissions were negatively correlated with MOP, although there was only an upward trend for MOP in all treatments under eCO2(Table II). This significant correlation of CH4emission with MPP and MOP might support the hypothesis that eCO2reduces CH4emission from paddy fields.In addition,the effect of eCO2on CH4emissions may be altered by different soil types.The present study was conducted in sandy loam paddies,which might result in higher hydraulic conductivity and soil Eh (Xieet al.,2012).These related soil properties can inhibit MPP and stimulate MOP.It has been reported that MPP can be accurately calculated from soil Eh development(Mitraet al.,2002),whereas MOP is influenced by soil moisture content,the plant-mediated diffusion rate of CH4,and O2availability(Zhanget al.,2012).Therefore,the eCO2-induced decrease in CH4emissions is likely caused by the changes in MPP and MOP and by many other factors due to the direct and indirect effects of eCO2(Malyanet al.,2016).

Rice cultivar breeding is a win-win strategy,as it simultaneously increases grain yield and decreases CH4emissions(Jiaet al.,2006).It is well recognized that high-yielding rice cultivars strongly decrease CH4emissions from paddy soils with high organic C content(Jianget al.,2017).In this study,eCO2increased aboveground and underground biomass and tiller number of the strongly and weakly responsive cultivars(Fig.1),which is consistent with previous studies(Inubushiet al.,2003;Wang Bet al.,2018).Meanwhile,the negative linear relationship between seasonal cumulative CH4emissions and aboveground and underground biomass (Table SII)likely indicates that CH4emissions decreased with an increase in rice biomass.Indeed,regardless of rice cultivar,higher biomass under eCO2(Fig.1)was accompanied with higher[O2]in the surface soil(Fig.5),which favored CH4oxidation(Maet al.,2010;Jianget al.,2017,2019).CH4emissions for the strongly responsive cultivar were much lower than those for the weakly responsive cultivar,regardless of the CO2level(Table I).This likely depended on the more rice biomass, higher soil Eh, and more [O2]in the surface soil(Figs.4,5,and S3,Table SII),suggesting that CH4production is lower and CH4oxidation is higher for the strongly responsive cultivar.These results were demonstrated by the decreased MPP and increased MOP for the strongly responsive cultivars relative to weakly responsive cultivars under eCO2conditions(Table II).

Compared to aCO2,eCO2did not significantly increase soil DOC content of the strongly and weakly responsive cultivars,while decreased CH4emissions(Fig.4,Table I).This result indicates that the difference in CH4emissions may not depend on soil DOC content.Jianget al.(2017)reported that high-yielding rice cultivars facilitated CH4oxidation by increasing O2transport and promoting methanotrophic organisms when DOC was high.Therefore,compared with other factors,soil DOC content may not be the main factor affecting CH4emissions. In our FACE system, previous studies have shown that soil organic C(SOC)sequestration increased with high N input through a dissolved organic matter (DOM)-microbial pathway in 2015 under 11-year eCO2conditions(Huet al.,2020).The eCO2induced higher microbial C use efficiency and lower effectiveness of the plant to prime DOM,thus increasing SOC sequestration and reducing soil C loss (Huet al., 2020). In addition, these processes probably led to more methanogenic substrates but less CH4production and emissions under eCO2conditions(Tables I and II,Fig.2).Moreover,the cropping system in our FACE was changed from rice-wheat to rice-fellow in 2010,which may increase SOC sequestration and thereby mitigate the effects of high-yielding rice cultivars on CH4emissions(Jianget al.,2017).

Consequently,consistent with our hypothesis,we propose a new inference that,regardless of rice cultivar,eCO2may reduce CH4emissions from rice fields in the context of continuous climate change and global warming.First,soil conditions and microbial communities may change under eCO2.In this case,the increase in rice biomass under eCO2promotes root aerenchyma tissue for transporting O2from air to paddy soil through the increased number of tillers(Inubushiet al., 2003; Kimet al., 2003). This enhances the transmission capacity of O2, resulting in higher soil Eh,which inhibits MPP and stimulates MOP,and thereby decreasing CH4emissions(Table SII,Figs.4a, b, and 5).Second,eCO2inhibits the abundance of methanogens due to the utilization balance of CH4substances for rice plant growth and soil microbial activity,although it increases CH4substrate availability with the contribution of root exudates by photosynthesis (Tokidaet al., 2011). Additionally, the stimulation of rice growth requires more irrigated water,which changes soil conditions and indirectly affects CH4emission from paddy fields. This change in hydrological conditions may be exacerbated by CO2.

Effects of stronglyand weaklyresponsive rice cultivars on N2O emissions as affected byeCO2

The increased N2O emissions from rice soil as affected by eCO2are primarily regulated by plant-mediated N and C availability induced by eCO2(Carnolet al.,2002;Caiet al.,2009; Bhattacharyyaet al., 2013; Wang Cet al., 2018b).Contrary to those previous studies, our multi-year FACE results clearly showed that N2O emissions from rice paddies of strongly and weakly responsive cultivars were lower at eCO2conditions than at aCO2conditions(Table I).These results suggest that,regardless of rice cultivar,N2O fluxes mainly occurred at the midseason drainage stage and the decrease in total N2O emissions induced by eCO2was due to a sharp reduction of N2O emissions at this stage(Fig.3),which is consistent with the results of previous studies(Wang Cet al.,2018b;Yaoet al.,2021).In the present study,the reduction of N2O emissions under eCO2was significantly correlated with a reduction in soil NH+4-N content and an increase in rice biomass (Table SII). Thus, the inhibitory effect of eCO2on N2O emissions from rice systems was due to the lower soil N availability for plant uptake (Sunet al.,2018;Yaoet al.,2021).Meanwhile,the eCO2caused substantial losses of NH+4-N(Fig.3),which resulted from the anaerobic oxidation of ammonium coupled to the reduction of iron(Xuet al.,2020).Thus,the high demand for available N for rice plant growth and the eCO2-induced losses of NH+4-N would inhibit N2O production and emissions from rice soil under eCO2conditions.

In this study,the lower N2O emissions for the strongly responsive cultivars compared to those for the weakly responsive cultivars under eCO2conditions(Table I)were mostly attributed to more NH+4-N content for plant growth and reduced N availability for N2O production for the strongly responsive cultivars (Jianget al., 2020). In addition, the enhanced soil C sequestration under strongly responsive cultivars suggested that this type of cultivar may provide little C source for denitrification, thus curbing N2O emissions(Yaoet al.,2021).Meanwhile,the significant decrease in the ratio of NO-2-reducing bacteria(nirS/nirK-type denitrifiers)abundance to N2O-reducing bacteria(nosZ-type denitrifiers)abundance under eCO2conditions suggested that eCO2reduced N2O emission by facilitating the consumption of N2O and favoring N2production through denitrification (Sun

et al.,2018).

Therefore, consistent with our hypothesis, the results of this study verified that eCO2significantly reduced N2O emissions from the rice fields,and the strongly responsive cultivars induced lower N2O emissions than the weakly responsive cultivars.This finding might help to better understand the different negative effects of eCO2and rice cultivars on N2O emissions from paddy fields(Yaoet al.,2021).In addition to eCO2, the reduction of N2O emissions is also closely related to rice cultivar and the interaction between rice cultivar and eCO2.

CONCLUSIONS

The CH4and N2O emissions varied among the strongly and weakly responsive rice cultivars.They decreased significantly and were positively correlated with the stimulated rice biomass under eCO2conditions compared to aCO2conditions.The decrease in CH4emissions under eCO2conditions was strongly linked to the decreased MPP and the increased MOP associated with higher surface soil Eh and[O2].The reduction in N2O emissions under eCO2conditions was mainly related to the decrease in soil available N and the increase in N uptake caused by rice biomass production.The greater reduction of CH4and N2O emissions for the strongly responsive cultivars compared with the weakly responsive cultivars was due to the higher soil Eh and[O2]in the surface soil,lower MPP and higher MOP,and greater N uptake for the former.However,considering the limitations of microbial testing of C and N cycles, the limitations of irrigation water variation,and the limited synergy of elevated temperature with eCO2on the emissions of CH4and N2O,further research is needed to better understand the long-term mechanisms involved.

ACKNOWLEDGEMENTS

This work was financially supported by the National Key Research and Development Program of China (No.2017YFD0300105),the National Natural Science Foundation of China (No. 41877325), and the Youth Innovation Promotion Association of Chinese Academy of Sciences(No.2018349).We thank Prof.Lianxin YANG from Yangzhou University, China for providing the rice seeds of LongIIyou1988,Yongyou1540,and Wuyungeng27 cultivars.We would like to express our appreciation to the editor and anonymous reviewers for their useful comments and insightful suggestions on the manuscript.We are also grateful to Ms. Qing LÜ, Mr. Rong ZHU, and Mr. Guoxin ZHU for their help in sampling and measurements.

SUPPLEMENTARY MATERIAL

Supplementary material for this article can be found in the online version.