Hoyu Qin,Yuo Jin,Jin Chn,Shn Hun,Yunlon Liu,Jun Zhn,Aixin Dn,Jinwn Zou,Gnxin Pn, Ynn Din, Yu Jin,*, Ks Jn vn Gronin, Wijin Zhn,*

a Institute of Crop Sciences, Chinese Academy of Agricultural Sciences, Beijing 100081, China

b Jiangsu Key Laboratory of Low Carbon Agriculture and GHGs Mitigation, College of Resources and Environmental Sciences, Nanjing Agricultural University, Nanjing 210095,Jiangsu, China

c Soil and Fertilizer & Resources and Environmental Institute, Jiangxi Academy of Agricultural Science, Nanchang 330200, Jiangxi, China

d Jiangxi Key Laboratory of Crop Physiology, Ecology and Genetic Breeding, Jiangxi Agricultural University, Nanchang 330045, Jiangxi, China

e Center of Agriculture and Climate Change, Institute of Resource, Ecosystem and Environment of Agriculture, Nanjing Agricultural University, Nanjing 210095, Jiangsu, China

f Jiangsu Collaborative Innovation Center for Modern Crop Production/Key Laboratory of Crop Physiology and Ecology in Southern China, Nanjing Agricultural University,Nanjing 210095, Jiangsu, China

g Department of Geography, College of Life and Environmental Sciences, University of Exeter, Exeter EX4 4RJ, UK

Keywords:Elevated CO2 Methane emissions Rice Methanogens Methanotrophs

ABSTRACT Elevated levels of atmospheric CO2 (eCO2) promote rice growth and increase methane (CH4) emissions from rice paddies, because increased input of plant photosynthate to soil stimulates methanogenic archae.However, temporal trends in the effects of eCO2 on rice growth and CH4 emissions are still unclear.To investigate changes in the effects of eCO2 over time, we conducted a two-season pot experiment in a walk-in growth chamber.Positive effects of eCO2 on rice leaf photosynthetic rate,biomass,and grain yield were similar between growing seasons.However, the effects of eCO2 on CH4 emissions decreased over time.Elevated CO2 increased CH4 emissions by 48%-101% in the first growing season,but only by 28%-30%in the second growing season.We also identified the microbial process underlying the acclimation of CH4 emissions to atmospheric CO2 enrichment:eCO2 stimulated the abundance of methanotrophs more strongly in soils that had been previously exposed to eCO2 than in soils that had not been.These results emphasize the need for long-term eCO2 experiments for accurate predictions of terrestrial feedbacks.

1.Introduction

Rice is a staple food of half of the human race [1].Rice paddies are a major source of methane(CH4) emissions, contributing~11%of total anthropogenic CH4emissions [2].The global warming potential of greenhouse gas (GHG) emission from rice paddies is dominated by CH4, and is the highest of all cropping systems[3,4].Climate change strongly affects both rice yields and CH4emissions from rice paddies [5,6].

Fossil fuel consumption, deforestation, and other types of land use change have increased atmospheric CO2concentrations from~290 μmol mol-1before the Industrial Revolution to 415 μmol mol-1in 2020 [2,7].Elevated levels of atmospheric CO2(eCO2) generally increase rice plant growth and rice yield by increasing the rice leaf photosynthetic rate [8-10].In a global meta-analysis [5], eCO2increased rice yield by 24.6%.However, it is still unclear to what extent the effects of eCO2vary over time.In some studies [8,11-13], positive effects of eCO2on rice yield did not depend on growing season, but in other experiments[14,15], they varied among rice seasons.

Methane emissions are determined by the balance between CH4production and oxidation [16].Methane is produced mainly by methanogenic archaea under anaerobic conditions, whereas CH4oxidation in rice paddies is performed mostly by aerobic methanotrophic bacteria [16].Because rice root exudates and plant residues are the main substrates for methanogenesis [17,18], eCO2often increases CH4emissions from rice paddies [5,19].However,eCO2can also increase the growth of methanotrophs by stimulating root growth and O2transport into soils [20-22].Methanotrophs may respond to eCO2more slowly than methanogens, because CH4oxidation is limited by CH4concentration[23-25].The response of methanogens to substrate addition is also faster than that of methanotrophs [23,26].Finally,long-term eCO2can reduce soil N availability[27-29],which in turn may affect rice plant growth and soil microbes [30,31].Thus, eCO2affects rice growth and CH4emissions through various mechanisms that may operate on different time scales.

Since the early 1990s, several studies have attempted to quantify the effects of eCO2on CH4emissions from rice paddies.But previous eCO2experiments often lasted for only one growing season [10,32,33].Several Free-air CO2Enrichment (FACE) and Opentop Chamber (OTC) experiments [15,34-36] provided multi-year data for CH4emissions, without explicitly considering temporal variation in the effect of eCO2.Although some of these experiments[15,34,36] reported an increase in eCO2effects over time, another[35] found that eCO2effects in the later seasons were similar to or lower than those in the earlier season.However, most previous multi-year experiments differed with respect to rice cultivar,straw or residue management, and fertilizer application rate among experimental seasons.Although all of these adjustments are common in practice,such differences have been shown[10,20,35,37]to affect eCO2effects on CH4emissions, possibly obscuring temporal trends.

Our previous FACE experiment with constant management practices [38] found that the effect of eCO2on seasonal CH4emissions decreased over time.Elevated CO2increased CH4emissions by 69.4%in year 1,but by only 44.0%in year 2 and by 25.6%in year 3(Fig.S1).However,eCO2effects on CH4emissions may depend on temperature [5,32], suggesting that year-to-year variation in climatic conditions may affect temporal trends.

To address these research gaps, we conducted a paddy microcosm experiment to investigate the effects of eCO2on rice growth and CH4emissions.To measure effects of eCO2on CH4emissions over time, we planted rice plants in soils that were previously exposed to eCO2and in soils that were not previously exposed.This experimental design was intended to assess effects of eCO2on CH4emissions over time under the same climatic conditions.We aimed to isolate the effect of experimental duration by keeping experimental and environmental factors constant between growing seasons.The objectives of our experiment were:1) to investigate effects of CO2on rice growth and CH4emission over time; and 2)to identify the mechanism underlying changes in the effects of eCO2over time.

2.Materials and methods

2.1.Experimental design

A two-season experiment to investigate the effects of eCO2on CH4emissions was conducted in two walk-in growth chambers(length, 4 m; width 2 m; height, 2.7 m) at the Jiangxi Academy of Agricultural Science (28.6°N, 115.9°E), Nanchang, China in 2019.CO2concentrations were set to 400 μmol mol-1in one chamber(ambient CO2,aCO2)and 600 μmol mol-1in the other(elevated CO2,eCO2).The soil used in the experiment was collected from the plow layer of a local rice paddy field and then was air dried,mixed,and sieved.Because CH4emissions from paddy soil are driven largely by labile soil C availability[16],statistical sensitivity to detect treatment effects on CH4emissions was increased by storing the soil outdoors for two years prior to the experiment.This pretreatment reduced soil labile C content, as it allowed most of the plant residues and other labile C in the soil to oxidize or mineralize[39].Soil properties were as follows:organic C 26.2 g kg-1,total N 2.8 g kg-1, total P 1.0 g kg-1, total K 14.4 g kg-1, available P 24.4 mg kg-1, and available K 99.4 mg kg-1.Because there are two major subspecies of rice cultivar(indicarice andjaponicarice),oneindica(Wuyou 308)and onejaponica(Ningjing 7)rice cultivars were used.

In the first season,plastic pots(diameter,22 cm;depth,20 cm)were filled with 4.5 kg of air-dried soil.For each treatment combination, 10 pots were used and two healthy rice seedlings were transplanted into each pot.On average,CO2concentrations during the daytime (12 h) were 410±40 μmol mol-1in the aCO2treatment and 615±18 μmol mol-1in the eCO2treatment;CO2concentrations during the nighttime were 501±42 μmol mol-1in the aCO2treatment and 657±22 μmol mol-1in the eCO2treatment.The Wuyou 308 and Ningjing 7 rice plants were harvested at 118 and 90 days after transplanting, respectively.After harvest, the soils of the first season were collected, combined within each CO2× cultivar combination, and air-dried.

At the start of the second season, plastic pots were filled with 4.5 kg of air-dried soil from the first season.For each cultivar,eight pots were planted in soils that had been exposed to aCO2in season 1.Four of these pots were placed in the aCO2growth chamber and the other four in the eCO2chamber.Similarly, four pots were planted in soils that had been exposed to eCO2in season 1 and placed in the eCO2growth chamber.Thus, three treatments per rice cultivar were applied:aCO2soils in the aCO2chamber (AA),aCO2soils in the eCO2chamber (AE), and eCO2soils in the eCO2chamber (EE).On average, daytime CO2concentrations were 425±39 μmol mol-1in the aCO2chamber and 618±20 μmol mol-1in the eCO2chamber,and nighttime CO2concentrations were 490±15 μmol mol-1in the aCO2chamber and 640±18 μmol mol-1in the eCO2chamber.The Wuyou 308 and Ningjing 7 rice plants were harvested at 100 and 90 days after transplanting, respectively.

To avoid artefacts arising from differences in microclimate,pots were moved between chambers every week in both seasons and the CO2concentrations were switched between the growth chambers at the same time.The mean day/night air temperature was set to 30/26 °C and the relative air humidity at 75%.These temperatures were roughly similar to local mean day/night temperatures of 32/26 °C during the rice growing season at our experimental site; the day air temperature was set slightly lower to avoid the risk of heat stress.Each chamber was lit by 8 high-pressure metal halide lamps and 8 high-pressure sodium lamps,providing an illumination intensity of about 3×104lx during the daytime.Nitrogen fertilizer as urea, phosphorus fertilizer, and potassium fertilizer were applied as basal dressing at the rates of 108 kg N ha-1,120 kg P2O5ha-1and 80 kg K2O ha-1,respectively.Side-dressed N fertilizer as urea was added at 36 kg N ha-1each at the tillering and panicle initiation stages.A 3-4 cm water layer was maintained on the soil surface during the rice growth period.

2.2.Sampling and measurement

The photosynthetic rate of the flag leaf was measured with a LI-6400 Portable Photosynthesis System(LI-COR Bio-Science,Lincoln,NE, USA) at heading stage.After harvest, rice plants were ovendried to constant weight and root and aboveground biomass were measured.Effective panicle number, filled-kernel number, and total kernel number were recorded.Grain yield and 1000-kernel weight were expressed on the basis of 15% moisture content.

The static closed chamber technique [40] was used to measure CH4fluxes from 1 week after rice transplanting to harvest at 7-10-day intervals.Prior to each measurement, the pot was transferred to a large container containing 2-3 cm of water, and an openbottom PVC chamber (diameter, 30 cm; height, 80 or 130 cm)was placed on top of each pot.On each sampling day,four gas samples were collected for each pot at 5-min intervals between 9:00 and 11:00 AM.A gas chromatograph (GC-2010 PLUS, SHIMADZU,Kyoto,Japan)was used to measure CH4concentrations.CH4fluxes were calculated as follows:

where ΔC/ΔTis the rate of change of CH4concentration(mg L-1h-1)in the chamber, calculated by linear regression.Vis the volume of the chamber (L) andAis surface area of the pot (m2).For the CH4flux calculation, only measurements for whichR2> 0.90 were accepted.Fewer than 5% of measurements failed to meet this threshold.

Soil samples were collected at harvest and at 45 days after rice transplanting, when CH4emissions were significantly different among treatments in both seasons.For each treatment, four pots were used to collect soil samples.In each pot,a 2-cm diameter soil corer was used to take three soil cores to 15 cm depth.Soil samples were mixed by hand and then passed through a 2-mm sieve to remove all visible roots and stones.Soil dissolved organic carbon(DOC) concentrations, soil extractable nitrogen (-N and-N) concentrations, and the abundance of methanogens and methanotrophs were measured.A total organic carbon analyzer(multi N/C UV,Analytik Jena AG,Jena,Germany)was used to measure soil DOC concentrations, and a flow autoanalyzer (Auto Analyzer 3, BRAN LUEBBE, Norderstedt, Germany) was used to measure soilandconcentrations.Soil DNA was extracted with a Power Soil DNA Isolation Kit (MoBio, USA).ThemcrAandpmoAgenes were then quantified using the primer pairs mcrAf/mcrAr and A189f/A682r, respectively [41,42].The number of copies ofmcrAgenes represents the abundance of methanogens,and those ofpmoAthat of methanotrophs.Quantitative real-time PCR was performed with CFX96 Touch Real-Time PCR Detection System (Bio-Rad, Foster City, CA, USA).

2.3.Statistical analysis

Means were compared by independent sample t-test in the first-season experiment and by one-way ANOVA in the secondseason experiment.Multiple comparisons were performed using the least significant difference (LSD) test.All analyses were performed with the statistical package SPSS 22.0.Differences between treatments were considered significant atP <0.05.

3.Results

3.1.Plant traits

In both seasons, eCO2significantly increased the flag-leaf photosynthetic rate, root biomass, aboveground biomass, yield, and effective panicle number (Tables 1 and 2).In the first season,eCO2significantly increased these parameters by respectively 40%, 83%, 43%, 26% and 45% for Wuyou 308 and by 29%, 34%,59%, 17% and 29% for Ningjing 7.In the second season, the effects of eCO2on the parameters were similar between soils with and without an eCO2history.For the Wuyou 308 cultivar, eCO2increased the parameters by respectively 29%, 31%, 22%, 26% and 31% from soils without an eCO2history, and by 30%, 24%, 14%,17% and 38% from soils with an eCO2history.Similarly, for the Ningjing 7 cultivar, eCO2increased the parameters by 31%, 24%,53%, 29% and 21% from soils without an eCO2history, and by 36%, 16%, 66%, 20% and 14% from soils with an eCO2history.

In neither season did eCO2affect the seed setting rate of Wuyou 308, but it increased the seed setting rate of Ningjing 7 by 20% in the first season.eCO2increased the 1000-kernel weight of both cultivars in the first but not the second season.

3.2.CH4 emissions

Elevated CO2significantly increased CH4emissions in both growing seasons (Figs.1,2;Table 3).In the first season,CH4emission peaked at 30 days after transplanting for both rice cultivars,and Wuyou 308 showed a second peak at 70 days after transplanting (Fig.1a and b).eCO2stimulated mean CH4emissions by 101%for Wuyou 308 and by 48% for Ningjing 7 (Fig.2a).In the second season,there were two CH4emission peaks at about 30 and 80 days after transplanting(Fig.1c and d).eCO2increased mean and cumulative CH4emissions of both cultivars less strongly in soils that had been previously exposed to eCO2(Fig.2b;Table 3).For Wuyou 308,eCO2stimulated cumulative CH4emissions by 96%from soils without an eCO2history, but by only 28% from soils with an eCO2history.Similarly, for Ningjing 7, eCO2increased cumulative CH4emissions by 67% from soils without an eCO2history but by only 26% from soils with an eCO2history (Table 3).These results strongly suggested that the effects of eCO2on CH4emissions decrease over time.

Table 1 Photosynthetic rate,aboveground biomass,root biomass,grain yield,effective panicle number,seed setting rate,and 1000-kernel weight as affected by elevated CO2 in the firstseason experiment.

Table 2 Photosynthetic rate,aboveground biomass,root biomass,grain yield,effective panicle number,seed setting rate,and 1000-kernel weight as affected by elevated CO2 and previous exposure to eCO2 in the second-season experiment.

Table 3 Cumulative CH4 emission, soil -N, and soil -N as affected by elevated CO2 and previous exposure to eCO2 in the second-season experiment.

Table 3 Cumulative CH4 emission, soil -N, and soil -N as affected by elevated CO2 and previous exposure to eCO2 in the second-season experiment.

Mean±standard error (n = 4).AA, ambient CO2 soils in ambient-CO2 chamber; AE, ambient CO2 soils in elevated-CO2 chamber; EE, elevated CO2 soils in elevated-CO2 chamber.Different letters represent significant difference within CO2 treatments (P < 0.05).

3.3.Soil , , and DOC concentrations

Elevated CO2reduced the soilandconcentrations for both cultivars at the tillering and mature stages.However, eCO2reduced soilconcentrations more strongly in soils that had been previously exposed to eCO2in Ningjing 7.eCO2reduced soilconcentrations for both Wuyou 308 and Ningjing 7 by 10%in soils without an eCO2history (AE) and by 9% and 19% in soils with an eCO2history (EE) respectively at the tillering stage.Similarly, eCO2reduced soilconcentrations of Wuyou 308 and Ningjing by respectively 12% and 13% in AE and by 12% and 20%in EE at mature stage.eCO2reduced soilconcentrations more strongly in soils with an eCO2history in both stages.For Wuyou 308, eCO2reduced soilconcentrations by only respectively 5% and 9% from soils without an eCO2history but by 21% and 23% from soils with an eCO2history at the tillering and mature stages.Similar results were found for the Ningjing 7 cultivar.Thus,soil,concentrations were lower at the mature stage,showing a decreasing effect of eCO2on soilover time.

Fig.1.CH4 emissions from Wuyou 308(a,c)and Ningjing 7(b,d)as affected by elevated CO2 in the first-season experiment(a,b)and by elevated CO2 and previous exposure to eCO2 in the second-season experiment(c,d).aCO2,ambient CO2;eCO2,elevated CO2;AA,ambient CO2 soils in ambient-CO2 chamber;AE,ambient CO2 soils in elevated-CO2 chamber; EE, elevated CO2 soils in elevated-CO2 chamber.Error bars indicate standard errors (n = 4).

Fig.2.Mean CH4 emissions from two rice cultivars as affected by elevated CO2 in the first-season experiment(a)and by elevated CO2 and previous exposure to eCO2 in the second-season experiment(b).aCO2,ambient CO2;eCO2,elevated CO2;AA,ambient CO2 soils in ambient-CO2 chamber;AE,ambient CO2 soils in elevated-CO2 chamber;EE,elevated CO2 soils in elevated-CO2 chamber.Different letters represent significant difference between treatments (P < 0.05).Error bars indicate standard errors (n = 4).

Elevated CO2increased soil DOC in both growing seasons(Fig.3).In the first season, eCO2increased soil DOC by 13% for Wuyou 308 and by 17% for Ningjing 7 (Fig.3a).In the second season,eCO2increased soil DOC more strongly in soils that were previously exposed to eCO2for both cultivars.For Wuyou 308, eCO2increased soil DOC by only 18%from soils without an eCO2historybut by 38% from soils with an eCO2history.Similarly, for Ningjing 7, eCO2increased soil DOC by only 9% from soils without an eCO2history, but 17% from soils with an eCO2history.

Fig.3.Dissolved organic carbon (DOC) from two rice cultivars as affected by elevated CO2 in the first-season experiment (a) and by elevated CO2 and previous exposure to eCO2 in the second-season experiment (b).aCO2, ambient CO2; eCO2, elevated CO2; AA, ambient CO2 soils in ambient-CO2 chamber; AE, ambient CO2 soils in elevated-CO2 chamber;EE,elevated CO2 soils in elevated-CO2 chamber.Different letters represent significant difference between treatments(P<0.05).Error bars indicate standard errors(n = 4).

3.4.Methanogens and methanotrophs

Elevated CO2increased the abundance of methanogens and methanotrophs in soils both with and without a history of eCO2(Fig.4).The effect of eCO2on the abundance of methanogens was similar between soils with and without an eCO2history(Fig.4a).However, eCO2increased methanotrophs more in soils with a history of eCO2than in soils without such a history(Fig.4b).For Wuyou 308, eCO2increased methanotrophic abundance by 67%in soils without an eCO2history,but by 159%in soils with an eCO2history.For Ningjing 7, eCO2increased methanotrophic abundance by 10% in soils without an eCO2history, but by 53% in soils with an eCO2history.

4.Discussion

In our study, eCO2increased rice biomass and grain yield by increasing leaf photosynthetic rate, in agreement with previous studies [5,8].The positive effects of eCO2on rice grain yield were similar between soils with and without an eCO2history,indicating that the positive effects of eCO2on rice yield are not affected by experimental duration, corroborating numerous field studies[8,11-13].However,Yang et al.[14]and Wang et al.[15]reported that positive effects of eCO2on rice yield decreased over time, a finding that may be explained by large variation in plant density[14] and meteorological parameters [15] among rice seasons.

Fig.4.The abundance of methanogens (a) and methanotrophs (b) as affected by eCO2 and previous exposure to eCO2.AA, ambient CO2 soils in ambient-CO2 chamber; AE,ambient CO2 soils in elevated-CO2 chamber;EE,elevated CO2 soils in elevated-CO2 chamber.Different letters represent significant difference between treatments(P<0.05).Error bars indicate standard errors (n = 4).

Elevated CO2increased CH4emissions from both rice cultivars,corroborating field and laboratory studies [5,19,20].Production of CH4is limited primarily by substrate availability [16].eCO2increased root biomass and soil DOC in our experiment and previous studies [5,22,43,44].Thus, our findings strongly suggest that eCO2increased CH4emissions by increasing substrate availability for CH4production by stimulating plant productivity.

However, the effect of eCO2on CH4emissions decreased with experimental duration.Our results suggest that this decrease was caused by the difference in the temporal response of methanogens and methanotrophs to eCO2.Methanogens tend to respond quickly to increases in root exudation.For instance,Aulakh et al.[26]found that soil CH4production rates peaked at 2-4 h after root-exudate addition.This rapid response of methanogens was confirmed by our experiment, as eCO2increased methanogens to the same extent in soils with or without a history of eCO2.

In contrast, eCO2increased the growth of methanotrophs more strongly in soils with a history of eCO2, suggesting that the response of methanotrophs to eCO2increases over time.eCO2stimulated root growth of both cultivars, which in turn can increase root O2release and stimulate CH4oxidation [20,21].However,CH4oxidation rates in rice paddies depend on the availability of CH4, suggesting that the response of methanotrophs to eCO2trails the response of methanogens.Indeed, ‘‘high-affinity” CH4oxidation is often induced at high (10,000 μmol mol-1) CH4concentrations [23].Without straw incorporation, soil CH4concentrations are typically low(<5000 μmol mol-1)during most of the rice growing season [10,20,39,45].Thus, the effect of short-term eCO2on CH4oxidation may be limited.

Longer-term eCO2reduced soilavailability more strongly than short-term eCO2in Ningjing 7.The reduction of ammonium availability could be associated with plant uptake or ammonium fixation [46].This response can increase CH4oxidation rates,because methane monooxygenase (the enzyme responsible for CH4consumption) preferentially binds and reacts withover CH4[47,48].eCO2also reduced soilconcentrations.Becauseis an effective electron acceptor for anaerobic oxidation of CH4, reductions inaddition can reduce CH4oxidation rates[49].However,soilconcentrations in rice paddies are typically low, owing to anaerobic conditions and the predominant use of ammonium nitrogen fertilizer in rice production.Thus, the effect of reducedconcentrations on CH4oxidation rates in rice paddies is likely limited.In summary,eCO2stimulated the growth and activity of methanotrophs more strongly in the long than in the short term, resulting in lower CH4emissions over time.

Because effects of eCO2on CH4emissions depend on management practices [20], it is not yet clear to what extent our results can be extrapolated to the global rice paddy area.However, our findings also corroborate an experiment[37]in a rice-wheat cropping system.In this experiment, the effect of eCO2decreased over two consecutive years in paddy fields with straw incorporation,whereas straw was removed in our present and previous studies[38].Our own experiment yielded similar results forindicaandjaponicacultivars.These results suggest that the acclimation of CH4emissions to eCO2occurs across a wide range of management practices.

Our conclusions come with one important caveat:the exact mechanism underlying the delayed response of methanotrophs to eCO2is unknown.Knowing this mechanism is important,because it will determine to what extent CH4emissions acclimate to eCO2in real-world systems.Wide-scale reductions in CH4emissions due to eCO2-induced reductions in soil N availability seem unlikely [50], as rice paddies are typically managed to maintain high soil N availability.To address this research gap, we suggest that future eCO2experiments study temporal trends in CH4emissions under a range of fertilizer practices.

5.Conclusions

The positive effects of eCO2on rice leaf photosynthetic rate,biomass, and grain yield were similar between growing seasons.The effect size of eCO2on CH4emissions from rice paddies appeared to decrease over time.eCO2increased the growth of methanotrophs more strongly in soils with an eCO2history than in soils without an eCO2history.These findings emphasize that accurate estimates of CH4emissions from rice agriculture in a high-CO2world require long-term experiments.

CRediT authorship contribution statement

Haoyu Qian:Data curation, Writing - original draft, Writing -review & editing.Yaguo Jin:Data curation.Jin Chen:Data curation.Shan Huang:Data curation.Yunlong Liu:Data curation.Jun Zhang:Data curation.Aixing Deng:Data curation.Jianwen:Writing - review & editing.Genxing Pan:Writing - review &editing.Yanfeng Ding:Writing - review & editing.Yu Jiang:Conceptualization, Funding acquisition, Writing - original draft,Writing - review & editing.Kees Janvan Groenigen:Conceptualization,Writing-review&editing.Weijian Zhang:Conceptualization, Funding acquisition, Supervision, Writing - review & editing.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgments

This work was supported by the National Key Research and Development Program of China (2017YFD0300104,2016YFD0300903, 2015BAC02B02), the National Natural Science Foundation of China (32022061), the Special Fund for Agroscientific Research in the Public Interest (201503118,201503122), the Agricultural Science and Technology Innovation Program of CAAS (Y2016PT12, Y2016XT01), the Modern Agricultural Development of Jiangsu Province (2019-SJ-039-07), and the GEF Project of Climate Smart Staple Crop Production in China(P144531).

Appendix A.Supplementary data

Supplementary data for this article can be found online at https://doi.org/10.1016/j.cj.2021.03.017.