The effect of integrative crop management on root growth and methane emission of paddy rice

Hao Zhang, Hailang Liu, Danping Hou, Yilei Zhou, Mengzhu Liu, Zhiqin Wang, Lijun Liu,Junfei Gu, Jianchang Yang*

Jiangsu Key Laboratory of Crop Genetics and Physiology, Jiangsu Key Laboratory of Crop Cultivation and Physiology, Jiangsu Co-Innovation Center for Modern Production Technology of Grain Crops,Yangzhou University,Yangzhou 225009, Jiangsu,China

Keywords:Rice Root traits Methane emission Integrative crop management

A B S T R A C T In previous studies, integrative crop management (ICM) improved shoot growth and grain yield of rice(Oryza sativa L.).However,little is known about the effect of ICM on root growth and methane (CH4) emission of paddy rice. In this study, two rice varieties, Wuyunjing 24 and Yongyou 2640, were grown. A field experiment was conducted with three crop management treatments including zero nitrogen fertilization (0N), local farmer practice(LFP),and ICM.Root morphophysiological traits and CH4 emission from the paddy field were investigated. ICM significantly increased mean grain yield by 29.9%, with the effect attributed mainly to an increase in mean total number of spikelets by 26.4% compared to LFP. ICM increased root and shoot biomass, root length, number of roots, root oxidation activity (ROA), root bleeding rate, and root total and active absorbing surface area by respectively 24.4%,25.7%,17.1%,9.3%,18.7%,29.5%,12.1%,and 24.7%.The concentrations of malic, succinic, and acetic acids in root exudates were respectively 5.8%, 6.0%, and 10.5%higher in ICM than in LFP. Compared to LFP, ICM significantly decreased the rate of CH4 emission during emission peak stages and reduced total CH4 emission by 17.1%. The root morphophysiological traits were positively and significantly correlated with grain yield,whereas root length, specific root length, ROA, and root total and active absorbing surface area were negatively and significantly correlated with total CH4 emission. These results suggest that ICM could achieve the dual goals of increasing grain yield and reducing the greenhouse gas effect by improving the root morphology and physiological traits of paddy rice.

1.Introduction

Rice(Oryza sativa L.) is a staple food for more than half of the world's population. To meet the ever-increasing demand for food due to population growth and improved living standards,world rice production needs to double by 2030 [1,2]. Considerable increases in rice yields have been attributed primarily to genetic improvement, innovations of crop management,and increases in resource input [3-7]. With the development of sustainable agriculture, Chinese agriculture will face great challenges to ensure high grain yield, high resource use efficiency, and low environmental costs [8-10]. To achieve high crop yield and high resource use efficiency,agronomists have developed many rice management techniques, including site-specific N management [11,12], optimized N application techniques [13-15], non-flooded mulching cultivation[16],alternate wetting and drying(AWD)irrigation[17,18],and precise and quantitative cultivation technique [19]. Although some of these optimizing techniques or integrative crop management practices have achieved the above purposes,the mechanism by which integrative crop management(ICM)increases yield and resource use efficiency is not fully understood, and little is known about the effect of ICM on root growth.

Roots are involved in acquisition of water and nutrients;synthesis of plant hormones, organic acids,and amino acids;and anchorage of plants, and their morphology is closely associated with the growth and development of aboveground organs and with grain yield and quality[20-23].Root biomass,structure in root tip cells, root activity and root-sourced hormones are considered to play a crucial role in root function, and much attention has been paid [24-29] to increasing grain yield by improving root distribution, structure and function in modern rice cultivars. However, little is known about the effect of ICM on root morphophysiological traits and their relationships with grain yield formation.

Agriculture accounts for 52% and 84% of global anthropogenic methane (CH4) and nitrous oxide (N2O) emissions.Agricultural soils may also act as sinks or sources of carbon dioxide(CO2),but the net flux is small.CH4exerts 25 times the greenhouse effect of CO2[30]. Because CH4is produced when organic materials decompose in oxygen-deprived conditions,long-term flooding of paddy fields is anaerobic environment[31]. The balance among the net exchanges of CO2, N2O, and CH4constitutes the net global warming potential (GWP). Net GWP per ton of crop yield is referred as greenhouse gas intensity(GHGI)[32].Agricultural greenhouse gas(GHG)fluxes are complex and heterogeneous, but many agricultural practices can potentially mitigate greenhouse gas emission[31].Future sustainable agriculture should employ cultivation management systems with low net GWP and GHGI at high crop productivity.However,the overall impacts of ICM on net GWP and GHGI, especially CH4emission, have not been assessed. Furthermore, most CH4is released by leaf blades,leaf sheaths, culm and roots in rice fields. The root system plays inhibiting and promoting roles in the production and emission of CH4in paddy fields [31], but the relationship between CH4emission and root morphological and physiological traits is not clear.

The objectives of the study were to investigate(1)whether ICM could improve root growth, and consequently, increase grain yield of rice,(2)whether ICM could reduce CH4emission from a paddy field, and (3) whether and how root morphophysiological traits are associated with grain yield and CH4emission. Root morphological and physiological traits were recorded. CH4emission was evaluated under three crop-management practices and the correlations of root traits with yield and total CH4emission were determined.This study would hope to provide an insight into the understanding the roles of root morphophysiological traits in yield formation and CH4emission.

2. Materials and methods

2.1. Plant materials and growing conditions

Field experiments were performed in 2015 and repeated in 2016 at an experimental farm in Yangzhou University, Jiangsu province,China(32°30′N(xiāo),119°25′E).The soil was a sandy loam(Typic Fluvaquent, Entisol, U.S. classification) with 1.30 g cm-3bulk density,0.192 g g-1soil moisture content at field capacity,23.2 g kg-1organic matter, 95.2 mg kg-1alkali-hydrolyzable N,22.5 mg kg-1available P, and 82.6 mg kg-1of available K.Weather records (average air temperature, sunshine hours,and precipitation during the rice growing period) over both years, obtained from a weather station located at the experimental site,are shown in Table 1.

A japonica rice variety Wuyunjing 24 (W24) and an indicajaponica hybrid rice variety Yongyou 2640 (Y2640) currently used in local production were grown in the field.In both years,seedlings were raised in a seedbed, sown on May 12 and transplanted on June 12. Heading occurred on August 23-25 for W24 and August 5-12 for Y2640. All plots were harvested on October 18-20.

2.2. Treatment

The treatments were laid out in a randomized complete block design with three replications in 5 m × 6 m plots. In all treatments, P (30 kg ha-1as single superphosphate) and K(40 kg ha-1as KCl) were applied as basal fertilizer. Fieldmanagement including weeds,insects,and diseases followed local high-yielding practice to avoid yield loss. Rice seedlings were transplanted at two seedlings per hill. The experiment received three crop management treatments: (i) zero N fertilization (0N), (ii) local farmer practice (LFP), and (iii)integrative crop management (ICM). The N rate, planting density, organic fertilizer application, and irrigation regime are described in Table 2.

Table 1 - Precipitation, sunshine hours, and mean temperature during the growing season of rice.

Table 2-Details of crop management treatments.

In the 0N, no N was applied. Transplanting was at a hill spacing of 0.16 m × 0.25 m. Water management was continuous flooding (CF). Except for drainage at mid-season, the plots were continuously flooded to a 2-3 cm water depth until one week before harvest. In the LFP, application of fertilizer was mainly in line with local farmer practice. Nitrogen(300 kg ha-1as urea) was applied in fractions of 5:2:2:1 at pre-transplanting, mid-tillering, panicle initiation and spikelet differentiation stages,respectively.Irrigation management and plant density were the same as those in the 0N treatment.In the ICM, four techniques were adopted: increased plant density, optimized N management, alternate wetting and moderate soil drying (AWMD), and organic fertilizer application (Table 2). ICM increased planting density by 25%compared to LFP, with transplanting at a hill spacing of 0.128 m × 0.250 m. ICM reduced the N rate by 10% and increased the fraction of N applied during panicle differentiation and development stage. Nitrogen (270 kg ha-1as urea)was applied in fractions of 4:2:2:2 at the stages of pretransplanting, mid-tillering, panicle initiation, and spikelet differentiation, respectively. AWMD irrigation was applied from 10 days after transplanting until maturity. The plots were not irrigated until the soil water potential at 15-20 cm depth reached -15 kPa. At this threshold, the corresponding soil water content was 0.170 g g-1. After AWMD the soil was saturated and the soil water potential was 0 kPa. There were 18 and 16 irrigation applications under AWMD in 2015 and 2016. The threshold value (-15 kPa of soil water potential)chosen in the AWMD was based on our earlier work [33,34].The model, installation and use of tensiometer detecting soil water potential followed Zhang et al. [18]. The organic fertilizer used at pre-transplanting was rapeseed cake after fermentation, a byproduct of rapeseed after oil pressing. The contents of N, P, K, and organic matter were 5.00%, 1.05%,1.20%,and 81.60%,respectively.

2.3. Sampling and measurement

After the mean stem number in each plot was recorded,plants of 12 hills were sampled for measurements of shoot and root biomass,root length,number of roots,ROA,root total and active absorbing surface area,and root bleeding(the stem is cut off from the base of plant, droplets soon flow from the cut due to root pressure) at mid-tillering, panicle initiation,heading time,and maturity.

Root sampling followed Zhang et al. [18]. For each root sampling,a cube of soil(20 cm × 20 cm × 20 cm)around each individual hill was dug up using a sampling core.Such a cube contains approximately 95% of total root biomass [21,35]. A portion of the roots were used for measurements of root length and number, ROA, and root total and active absorbing surface area,and for root exudate collection and analysis.The rest were dried in an oven at 70 °C to constant weight and weighed.Aboveground plant samples were dried in an oven at 70 °C to constant weight and weighed. Root length and number were recorded following Zhang et al. [18]. The ROA was determined by measurement of oxidation of alphanaphthylamine following Ramasamy et al. [36]. Root total absorbing surface area and active absorbing surface area were determined by the methylthionine chloride dipping method[37].

Collection of root exudates followed Egle et al. [38] at the main growth and development stages. After collection, the root exudate solution was lyophilized and kept at-20 °C until analysis.The organic acids in root exudates were assayed on a high-performance liquid chromatograph (Waters 2695, Waters Corp., Milford, MA, USA) using a 2487 UV detector,reversed-phase Atlantis dC18analysis column of 3.9 mm ×150.0 mm, mobile phase A of 20 mmol L-1NaH2PO4(pH 2.7,adjusted with phosphoric acid),mobile phase B of acetonitrile,mobile phase C of extra-pure water, and mobile phase D of methanol (chromatographic grade). The flow velocity was 0.5 mL min-1, column temperature was 37 °C, UV detector wavelength was 248 nm, and sample volume was 10 μL. Five organic acid standards (malic, tartaric, succinic, citric, and acetic acids)were provided by Waters Corp.

Plants from five hills in each plot were cut 12 cm above soil level at 18:00, and a known weight of absorbent cotton was placed on the top of each decapitated stem and covered with a polyethylene sheet. At 6:00 on the following morning, the cotton was weighed and root bleeding rate was calculated according to the following formula:Gas samples for CH4was collected and assayed by the closed-chamber technique [39]. Gas samples were analyzed for concentrations of CH4with a 7890A gas chromatograph(Agilent, Santa Clara, CA, USA). The cumulative emissions of CH4during the rice growing season were calculated by integration. The CH4global warming potential (GWP) was calculated according to the following formula:

2.4. Final harvest

Number of panicles per square meter,percentage of filled kernels,and kernel weight were determined from 50 plants (excluding border ones)sampled randomly from each plot.Grain yield was determined from all plants in a 5 m2area(except border plants)in each plot and adjusted to a moisture content of 0.14 g H2O g-1fresh weight.The definition and calculation of percentage of filled kernels and number of spikelets per panicle followed Zhang et al.[18].Harvest Index is calculated by the grain yield divided by total dry matter produced above ground.

2.5. Statistical analysis

Analyses of variance were performed using the SAS/STAT statistical analysis package (version 9.2, SAS Institute; Cary,NC, USA). Data from each sampling date were analyzed separately, and the resulting means were tested by least significant difference at P ＜0.05(LSD0.05).

3.Results

3.1. Shoot and root biomasses and root:Shoot ratio

Shoot and root biomasses were significantly greater in ICM than in LFP at the main growth stages across the two years(Fig.1-A-H).The mean root biomasses of the two rice varieties in ICM were 33.5%,8.9%,9.8%,and 9.3%greater than in LFP at mid-tillering, panicle initiation, heading time and maturity,respectively (Fig. 1-A-D). Shoot biomasses in ICM were respectively 41.9%, 18.9%, 20.5%, and 21.9% greater than those in LFP at the four main growth stages (Fig. 1-E-H). At each growth stage, the shoot and root biomasses of Y2640 under the three treatments were higher or significantly higher than that of W24 in both years(Fig.1-A-H).

The root:shoot ratio decreased gradually with the advance of growth stage (Fig. 1-I-L). Except at the mid-tillering stage,when the root:shoot ratio was significantly lower in both LFP and ICM than in the 0N treatment, the differences in root:shoot ratio among the three treatments were not significant(Fig. 1-I-L), suggesting that both shoot and root biomasses were simultaneously increased in ICM relative to LFP.

3.2. Root length and number

The changes in root length and number are shown in Fig. 2.Root length and number were greater or significantly greater in ICM than in LFP throughout the growth season in both years (Fig. 2-A-H). The maximum root length and number occurred at panicle initiation and heading time, respectively.The maximum root length of the two rice varieties in ICM,on average,was 160.6 m hill-1in 2015 and 162.5 m hill-1in 2016,and was respectively 16.2% and 17.7% more than that in LFP(Fig. 2-A-D). For maximum number of roots, both rice varieties, on average, recorded 10.1 × 103m-2in both years in ICM,7.6% more than in LFP(Fig.2-E-H).

3.3. ROA and root bleeding rate

As with root length, maximum ROA occurred at panicle initiation, and was significantly greater in ICM than in LFP at the main growth stages in both years(Fig.3-A-D).The changes in root bleeding rate were similar to those of the number of roots.The maximum root bleeding rate occurred at the heading time and the rate was significantly greater in ICM than in LFP at the main growth stages in both years (Fig. 3-E-H). At each growth stage,the ROA and root bleeding rate of Y2640 under all crop management treatments were higher or significantly higher than those of W24 in both years(Fig.3-A-H).

3.4. Root total and active absorbing surface area

The maximum values of root total and active absorbing surface area appeared simultaneously at heading time. Except for the mid-tillering stage,at which there was no significant difference in root total absorbing surface area between LFP and ICM, the two parameters were significantly greater in ICM than in LFP throughout the growth season(Fig.4-A-H).The mean root total absorbing surface areas of the two rice varieties were 14.5,11.8,and 11.3%greater in ICM than that in LFP at panicle initiation,heading time and maturity, respectively (Fig. 4-A-D). For root active absorbing surface area, a more significant increase was observed in ICM than in LFP. It was 24.3%, 27.2%, 23.8%, and 23.3%in ICM more than in LFP at four main growth stages(Fig.4-E-H). The root total absorbing surface area and active absorbing surface area of Y2640 at each growth stage were greater than those of W24 in both years(Fig.4-A-H).

3.5. Concentrations of organic acids in root exudates

tneserper MCI dna,PFL,N0.6102 dna 5102 ni stnemtaert tnemeganam porc eerht rednu ecir fo)L-I(oitar toohs:toor dna,)H-E(thgiew yrd toohs,)D-A(thgiew yrd tooR-1.giFdna,emit gnidaeh,noitaitini elcinap,gnirellit-dim tneserper AM dna,TH,IP ,TM .ylevitcepser,tnemeganam porc evitargetni dna,ecitcarp remraf lacol ,noitazilitref N orez emas eht nihtiw 50.0 = ta ecnacifingis lacitsitats etacidni srab evoba srettel tnereffiD.naem eht fo rorre dradnats ± tneserper srab lacitreV.ylevitcepser,ytirutamP .50.0 = ta tnacifingis ton snaem SN dna,egats tnemerusaemP

Amounts of five organic acids in root exudates are presented in Table 3. The concentrations of organic acids at panicle initiation were higher than those at heading time. At panicle initiation,ICM significantly increased concentrations of malic,succinic, and acetic acids in comparison with LFP. The difference in concentrations of tartaric and citric acids among the three treatments was not significant. The organic acid concentrations of Y2640 were higher than those of W24 in both years and at the same measurement stage(Table 3).

Fig.2-Root length(A-D)and root number(E-H)of rice under three crop management treatments in 2015 and 2016.0N,LFP,and ICM represent zero N fertilization,local farmer's practice,and integrative crop management,respectively.MT,PI,HT,and MA represent mid-tillering,panicle initiation,heading time and maturity,respectively.Vertical bars represent ± standard error of the mean.Different letters above bars indicate statistical significance at P = 0.05 within the same measurement stage,and NS means not significant at P = 0.05.

Fig.3-Root oxidation activity(A-D)and root bleeding rate(E-H)of rice under three crop management treatments in 2015 and 2016.0N,LFP,and ICM represent zero N fertilization,local farmer practice,and integrative crop management,respectively.MT,PI,HT and MA represent mid-tillering, panicle initiation,heading time,and maturity,respectively.Vertical bars represent ±standard error of the mean.Different letters above bars indicate statistical significance at P = 0.05 within the same measurement stage,and NS means not significant at P = 0.05.

Fig.4- Root total absorbing surface area (A-D)and active absorbing surface area (E-H)of rice under three crop management treatments in 2015 and 2016.0N,LFP, and ICM represent zero N fertilization,local farmer practice,and integrative crop management,respectively.MT, PI,HT,and MA represent mid-tillering,panicle initiation,heading time, and maturity,respectively.Vertical bars represent ± standard error of the mean.Different letters above bars indicate statistical significance at P = 0.05 within the same measurement stage,and NS means not significant at P = 0.05.

Table 3-Organic acid concentrations (μmol g-1 DW) in root exudates of rice under three crop management treatments in 2015 and 2016.

3.6. CH4 emission

CH4emission exhibited a similar trend throughout the rice growth season for all the treatments (Fig. 5-A-D). CH4emission increased at rice transplanting, peaked at 30 days after transplanting, and gradually decreased thereafter. Another small peak was observed at 63 days after transplanting(Fig. 5-A-D). Compared to 0N, both LFP and ICM significantly increased CH4emission peak values at two emission peaks.However, ICM significantly decreased CH4emission peak values compared to LFP. CH4emission peak values of Y2640 were lower than those of W24 in both years(Fig.5-A-D).

3.7. Grain yield and harvest index

Table 4 shows the grain yield and its components and harvest indexes of the two rice varieties. ICM significantly enhanced yield by 27.6% for W24 and 32.3% for Y2640 compared to LFP, on average. The high grain yield in ICM was attributed mainly to a larger sink size (total number of spikelets) as a result of a larger panicle. Increases in filled kernels and kernel weight also contributed to higher grain yield. The grain yield of Y2640 was higher than that of W24 in both years (Table 4). ICM also significantly increased the harvest indexes of the two rice varieties compared to LFP(Table 4).

3.8. CH4 global warming potential (GWP)

Compared to LFP and on average,ICM significantly decreased the total amount of CH4emission or equivalent CO2by 16.2%for W24 and 18.1%for Y2640.ICM also significantly decreased GHGI compared to LFP(Table 5).

3.9. Correlations of roots traits with yield and total CH4 emission

Root morphological and physiological traits were positively and highly significantly correlated with grain yield,whereas root length, specific root length (root length/root dry weight), ROA, and root total and active absorbing surface area were negatively and significantly or highly significantly correlated with total CH4emission (Fig. 6, Table S1). Concentrations of malic and acetic acids at panicle initiation and concentrations of succinic and acetic acids at heading time were positively and significantly correlated with total CH4emission (Fig. 6, Table S1).

4. Discussion

It is a great challenge to achieve the goals of high yield and quality, high N and water use efficiency, and environmental friendliness, and high grain yield usually needs high N input[12,16-18]. Can high yield and N and water use efficiency be achieved simultaneously? This has been a research question in China and abroad [8,12,15,17,18]. Our earlier work showed that integrative and optimized crop management could achieve the dual goal of increasing grain yield and N and water use efficiency [18]. In the present study, ICM significantly increased grain yield for two rice varieties in two years compared to LFP (Table 4). The results indicate that adoption of an ICM in rice could not only reduce N application but also increase grain yield.

How could an ICM increase crop yield and reduce N and water input? There are some reports showing that ICM improves the agronomic and physiological traits of rice shoots[12,15,17,18]. However, little is known about the effect of an ICM on root morphological and physiological traits. Our results showed that an ICM could improve root morphology(increases in root biomass,root length,and root number)and physiology(increases in ROA,root bleeding rate,and root total and active absorbing surface area)compared to LFP(Figs.1-4).Many investigations have been made of the relationship between rice root traits with yield formation [21,24,40]. Root biomass and ROA are considered the two most important traits in root morphology and physiology. Root biomass is closely related to root absorbing ability and aboveground biomass, and a high ROA is necessary to maintain root and shoot growth and nutrition uptake[24,34,36,41].Root bleeding rate is closely related to active water absorption of the root system, and reflects physiological root activity. A higher bleeding rate is associated with higher amounts of N and dry matter, leading to an increase in spikelet number [42]. Our finding that root morphological and physiological traits were positively and highly significantly correlated with grain yield(Fig.6,Table S1)suggests that improved root morphology and physiology contribute to increases in rice grain yield under ICM.

The differences in root:shoot ratio among all the treatments were not significant after mid-tillering stage (Fig. 1). It is known that the root and shoot are interdependent: active shoots supply sufficient carbohydrates to roots to develop and maintain root functions, while active roots supply nutrients,water, and phytohormones to shoots to increase shoot biomass productivity [34,40,43]. We conclude that increases in both root and shoot growth result in higher grain yield under ICM.

Organic acids, as important components of root exudates,are involved in many rhizosphere processes, including mineral weathering, nutrient uptake and detoxification [29].It is believed [29] that the composition and concentration of organic acids in root exudates change with environment.Chang et al.[44]reported that organic acids exuded from roots were also related to grain quality: malic and succinic acids exuded from roots were significantly correlated with breakdown values and significantly and negatively correlated with setback values in starch profile. However, little is known about the effect of crop management on the components and concentrations of organic acids from root exudates. In the present study,ICM significantly increased the concentrations of malic,succinic,and acetic acids compared to LFP(Table 3).The difference in concentrations of tartaric and citric acids among the three crop management treatments was not significant (Table 3). These results suggest that organic acid secretion can be changed by cultivation management. However, the mechanism by which organic acids regulate the rhizosphere environment and plant growth needs further study.

Table 4-Grain yield, yield components, and harvest index of rice under three crop-management treatments in 2015 and 2016.

CH4emission from irrigated rice fields is the result of complex interactions between rice plants and soil microbes.The flooding of rice fields initiates a series of events involving anaerobic degradation of organic substrates under oxygendeprived conditions, leading to the production of fermentation products that finally drive CH4formation [31]. Up to 90%of the CH4emission from rice fields is plant-mediated through well-developed intracellular air spaces in leaf blades, leaf sheaths,culms,and roots[45].To mitigate the production and emission of CH4in rice fields, many technologies and practices have been advocated, including the choice of low-CH4emission rice cultivars, balanced fertilization, adjusting the timing of organic residue additions,and draining wetland rice once or several times during the growing season [31]. In the present study, ICM significantly reduced CH4emission compared to LFP. One of the major reasons was that AWMD irrigation avoided placing the rice under oxygen-deprived conditions, so that the activity of methanogenic bacteria was severely inhibited and the production of CH4reduced. Also,root activity was enhanced significantly. The higher capacity of oxidation and oxygen secretion of the root system and the higher rhizospheric redox state repressed the activity of methanogenic bacteria,whereas the activity of CH4-oxidizing bacteria was enhanced, promoting oxidation of CH4and leading to the reduction of CH4emission [46]. Root length,specific root length, ROA, and root total and active absorbing surface area were negatively and significantly or very significantly correlated with total CH4emission(Fig.6,Table S1).In another study, components and concentrations of organic acids in root exudates exerted different effects on CH4production and emission [47]. Concentrations of malic and acetic acids at panicle initiation and concentrations of succinic and acetic acids at heading time were positively and significantly correlated with total CH4emission (Fig. 6, Table S1).Because CH4emissions are determined mainly by the net balance between the activities of methanogens and methanotrophs, any factors influencing these microbes could influence CH4metabolism [48,49]. The relationship between organic acids and CH4emission is complicated and needs further study.

Table 5-Amount of CH4 emission and GHGI of rice under three crop management treatments in 2015 and 2016.

Usually, applying organic fertilizers increases CH4emission from paddy fields [31,45]. However, no increase in CH4emission was observed when oilseed cake fertilizer was applied in ICM (Fig. 5). A possible explanation is that the organic fertilizer had been fully decomposed before application and CH4had already been emitted, so that advanced fermentation and decomposition avoided an increase of CH4emissions in rice fields under ICM.

Fig.6-Correlation heat map of root traits with yield and total CH4 emission(TME)of rice.?Specific root length = root length/root dry weight.

Generally, a flooded paddy emits more CH4than a nonflooded paddy on a sunny day. However, a similar pattern of CH4emission was observed in the two study years, although July rainfall was quite different between 2015 and 2016.We do not know the reason. CH4emission was lower on a rainy day than on a sunny day under the same field conditions(data not shown), owing probably to less opening of leaf stomata on a rainy day.More rainfall in July of 2016 may have reduced CH4emission from the paddy field, possibly contributing to the similar pattern of CH4emission between the two study years despite the longer drying time in the field in 2015 than in 2016.The question of whether and how a rainy day could result in lower CH4emission than a sunny day awaits further study.

5. Conclusions

An ICM featuring increase in plant density, decrease in N application, application of a higher proportion of N at later growth stages,and use of AWMD and rapeseed cake fertilizer significantly increased grain yield and reduced CH4emission from the paddy field in comparison with LFP. The increase and the reduction were attributed mainly to improvement in root morphophysiological traits, including increases in root and shoot biomass, root length, root number, ROA, root bleeding rate, root total and active absorbing surface area,and five organic acid concentrations in root exudates. The mechanism underlying root-shoot and root-soil interactions for high grain yield and low environment costs,and the role of the root-supplied hormones in regulating yield formation and greenhouse gas emission, await further study.

Supplementary data for this article can be found online at https://doi.org/10.1016/j.cj.2018.12.011.

Acknowledgments

This work was supported by the National Basic Research Program of China (2015CB150404), the National Key Research and Development Program of China (2016YFD0300206-4,2018YFD0300801), the National Natural Science Foundation of China (31871559, 31671614), Young Elite Scientists Sponsorship Program by CAST(2016QNRC001),the Natural Science Foundation of the Jiangsu Higher Education Institutions(15KJA210005), the Project Funded by the Priority Academic Program Development of Jiangsu Higher Education Institutions, and the Top Talent Supporting Program of Yangzhou University (2015-01).