Jin Chen ,Xingheng Zhu ,Jing Xie ,Guoqing Deng ,Tinhu Tu ,Xinjio Gun ,Zhen Yng ,Shn Hung ,Xinmo Chen ,Cifei Qiu ,Yinfei Qin ,Cihong Sho ,Minggng Xu ,Chunrui Peng ,*

a Jiangxi Academy of Agricultural Sciences,Key Laboratory of Crop Ecophysiology and Farming System for the Middle and Lower Reaches of the Yangtze River,Ministry of Agriculture and Rural Affairs,National Engineering and Technology Research Center for Red Soil Improvement,Nanchang 330200,Jiangxi,China

b College of Life Science and Environmental Resources,Yichun University,Yichun 336000,Jiangxi,China

c Agricultural Technology Extension Center in Yifeng County of Jiangxi Province,Yichun 343100,Jiangxi,China

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

e Institute of Agricultural Resources and Regional Planning,Chinese Academy of Agricultural Sciences,National Engineering Laboratory for Improving Quality of Arable Land,Beijing 100081,China

ABSTRACT Rational nitrogen (N) application can greatly increase rice (Oryza sativa L.) yield.However,excessive N input can lead not only to low N use efficiency (NUE) but also to severe environmental pollution.Reducing N application rate with a higher planting density (RNHD) is recommended to maintain rice yield and improve NUE.The effects of RNHD on fertilizer N fate and rice root growth traits remain unclear.We accordingly conducted a two-year field experiment to investigate the influence of RNHD on rice yield,fertilizer 15N fate,and root growth in a double-rice cropping system in China.In comparison with the conventional practice of high N application with sparse planting,RNHD resulted in similar yield and biomass production as well as plant N uptake.RNHD increased agronomic NUEs by 23.3%–31.9%(P<0.05)and N recovery efficiency by 17.4%–24.1%(P<0.05).RNHD increased fertilizer 15N recovery rate by 14.5%–34.7%(P<0.05),but reduced 15N retention rate by 9.2%–12.0%(P<0.05).Although a reduced N rate led to significantly reduced root length,surface area,volume,and biomass,these root traits were significantly increased by higher planting density.RNHD did not affect these root morphological traits and reduced activities of nitrate reductase(NR)and glutamine synthetase(GS)only at tillering stage.Plant N uptake was significantly positively correlated with these root traits,but not correlated with NR and GS activities.Together,these findings show that reducing N application with dense planting can lead to high plant N uptake by maintaining rice root growth and thus increase NUE.

Keywords:Rice Planting density N recovery efficiency Root morphology South China

1.Introduction

Rice is the major staple food of more than half the global population.Global rice production must increase by 8–10 million Mg per year over the next decade to meet rising demand [1].Owing to the limitation in rice area expansion,the best way to further increase total rice production is to increase yield [2,3].Over the past three decades,an increase in nitrogen (N) fertilizer input is a key factor increasing rice yield in China[4].However,the current average rate of inorganic N application for rice production is 180 kg ha-1per cropping season in China,about 75% higher than the world average[5,6].Excessive N fertilizer input not only reduce rice yield,but results in lower N use efficiency(NUE)as well as severe environmental pollution [7].Elevated N deposition and soil N pool buildup from residual N fertilizer have also greatly increased the soil indigenous N supply in rice fields [8],likely further reducing NUE[9].It is desirable to establish sustainable agronomic practices for lower N application rates while maintaining adequate rice yield.

Both N application rate and planting density influence rice yield and NUE.To reduce the cost of manual transplanting and hybrid seeds,sparse planting is common in major Chinese rice cropping regions [10,11].In this practice,excessive N is often applied at basal and tillering stages to produce a sufficient panicle number for high yield [12].However,the combination of high N rate and sparse planting generally leads to low NUE[11].Reducing N application rates to minimize fertilizer N losses [12],may reduce rice yield under sparse planting [10].However,a moderate increase in planting density can increase plant N uptake as well as improve NUE [13,14].For this reason,in place of the conventional high N rate with sparse planting,a reduced N rate with higher planting density(RNHD)is now strongly recommended[15–18]for increasing NUE while maintaining yield.However,most the previous experiments [12–15] studied the NUE of RNHD based on differences in plant N uptake in fertilized and unfertilized plots.The quantitative effect of RNHD on the fate of fertilizer N (fertilizer N recovery efficiency) in rice–soil systems remains unclear.

Root morphology and physiology strongly influence fertilizer N uptake and NUE in rice fields[19,20].A rice root system with larger biomass,greater length,and greater length density contributed to a higher NUE[21].The activities of nitrate reductase(NR)and glutamine synthetase(GS)are related to plant N uptake[22].N application rate may stimulate or weaken rice root growth,depending on the N rate [23].N addition also stimulates the activities of NR and GS[24,25].Increasing planting density added root length density [26] and root biomass per unit area [27],but reduced root NR activity[28].However,the effect of RNHD on rice root growth and the activities of NR and GS are still unknown.

Double-rice cropping is the dominant farming system in southern China,accounting for 33.3% of the total rice planting area in China in 2018[29].The double-rice cropping system typically consists of early rice (from March to July) and late rice (from June to October).Sparse planting with high N fertilizer(165–195 kg N ha-1in early rice and 195–225 kg N ha-1in late rice) is widely used by farmers[10,30,31].Most fertilizer N is applied at transplanting and tillering stages,resulting in low NUE [11].

We conducted a two-year field experiment to investigate grain yield,fertilizer N fate,and root growth traits under different N rates and planting densities in a double-rice cropping system.Our main objectives were (1) to investigate the effect of RNHD on fertilizer N fate and rice yield,and (2) to investigate the effect of RNHD on the morphological and physiological traits of rice roots.

2.Materials and methods

2.1.Experimental site

The field experiment was conducted from 2017 to 2018 in a farmer’s field in Yifeng county,Jiangxi province,China (28°39′N,114°80′E).The soil was loam and classified as Eutric Fluvisol.The initial soil properties in the 0–20 cm layer were as follows:pH 6.20,26.7 g kg-1soil organic matter,1.50 g kg-1total N,148.7 mg kg-1available N,32.4 mg kg-1Olsen P,and 135.6 mg kg-1available K.Daily mean temperature and precipitation are described in Fig.1.

2.2.Experimental design

The experiment was laid out in a split-plot design with three replications.The main plot was N rate with two treatments:standard N rate (SN) and reduced N rate (RN),and the subplot was planting density with three treatments:standard planting density(SD) and two higher planting densities (HD1 and HD2).The SNSD treatment was treated as the check(CK),determined from conventional cropping practices for high yield at the experimental site.The RNHD1 and RNHD2 treatments were considered as reducing N application rate with higher planting density.A zero-N treatment (N0SD) was included for NUE estimation.For the RN treatment,N rates at basal and tillering stages were reduced,whereas N rates at panicle initiation stage were kept the same as in the SN treatment.For HD1 and HD2,planting density was increased by reducing the distance between rows.N fertilizer was applied as urea (46% N),and N rate and planting density are described in Table 1.

Table 1 Details of the experimental design and treatments.

To investigate fertilizer-N fate,a fertilizer15N microplot was established in each main plot during the early and late seasons in 2017.Before transplanting,PVC frames were inserted 40 cm deep in the soil to prevent surface runoff and lateral diffusion.The length of each PVC frame was 50.1 cm in the early season and 60 cm in the late season.The width of each PVC frame was 60 cm for SD,50.1 cm for HD1,and 42.9 cm for HD2 in both early and late seasons.Six hills were planted in each microplot.The application rate and timing of15N-labeled urea(10%isotopic abundance;Shanghai Chem-Industry Institute,Shanghai,China) in15N microplots were the same as in the corresponding main plots.Other management practices in the15N microplot were consistent with those used in the main plots.

The inbred rice cultivars Ganxin 203 and Rongyou 225 were planted in the early and late seasons,respectively.Main plots were 30 m2and were separated by 30 cm-wide ridges covered with plastic film to prevent water leakage and nutrient losses.Before transplanting,90 and 105 kg P2O5ha-1were applied as calcium magnesium phosphate(14%P2O5)in early and late seasons,respectively.Applications of 180 and 210 kg K2O ha-1as potassium chloride (60% K2O) were split equally between basal and panicle initiation stages in early and late seasons,respectively.A 3–5 cm water depth was maintained after transplanting until the end of the tillering stage.The field was then drained for a week to reduce unproductive tillers.Thereafter,3–5 cm of flooding water was maintained until seven days before maturity.Weeds,diseases,and insects were intensively controlled to prevent yield losses.

2.3.Measurements

2.3.1.Aboveground plant sampling and N uptake

Plants were sampled at tillering,heading,and maturity stages in each growing season for the measurement of aboveground biomass and plant N uptake.At maturity,100 hills in each plot were harvested for determining grain yield.Grain moisture content was determined immediately after threshing (Riceter grain moisture meter,Kett Electric Laboratory,Tokyo,Japan).Grain yield was recorded at a standardized moisture content of 135 g H2O kg-1fresh weight.Six hills were randomly sampled to determine yield components including number of panicles per square meter,spikelets per panicle,filled-kernel rate and 1000-kernel weight.N content in leaves,stems,and panicles at maturity was determined by the micro-Kjeldahl method (FOSS Kjeltec 8400,FOSS Company,Copenhagen,Denmark) [32].N uptakes in grain and straw were calculated as the products of N content with grain and straw mass.

2.3.2.Root characteristic and enzyme activity

Root morphological characteristics and activities of NR and GS were recorded at tillering,heading,and maturity stages (or at filling stage for NR and GS activities) in 2017.Before transplanting,a nylon mesh bag (height,20 cm;mesh size,37 μm) was placed in each plot to facilitate root sampling.Roots were recovered from root bags and thoroughly washed with tap water.Root morphological characteristics (length,surface area,and volume) were determined using a scanner (Epson Expression 1680 Scanner,Seiko Epson Corp,Tokyo,Japan) and a WinRHIZO Root Analyzer System(Regent Instruments Inc,Quebec,Canada)[24].Root samples were oven-dried at 105 °C for 30 min and then at 70 °C to constant weight to determine root biomass.Fresh roots (0.5 g) were frozen immediately in liquid N and then extracted.The extracts were centrifuged at 12,000 r min-1for 20 min.The supernatants were then used to determine enzyme activity by enzyme-linked immunosorbent assays (ELISA).NR and GS activities in fresh roots were measured following the manufacturer’s instructions,using respectively the plant NR ELISA kit and the plant GS ELISA kit (Sinobestbio,Shanghai,China) [25].

Fig.1.Daily mean air temperature and precipitation near the field experimental site in 2017 and 2018.The climate data are from the local observational stations in the National Meteorological Networks of Central China Meteorological Agency.

2.3.3.The fate of fertilizer 15N

At maturity in 2017,all plants in the microplots were carefully collected to determine the fate of fertilizer15N.Plants were dried at 65 °C in a forced-air oven.Soil samples were collected from 0 to 40 cm depth in all microplots,and all soil samples were airdried,ground and passed through a 150-μm screen[32].N content in plant and soil samples was determined by the Kjeldahl method as described above.15N abundance in plant and soil samples was determined by a mass spectrometer (Delta Plus XP,Thermo Finnigan,Pittsburgh,PA,USA).

2.4.Calculations and statistical analysis

NUEs of agronomic N use efficiency (AEN) and apparent N recovery efficiency (REN) were calculated as follows [33]:

The fertilizer15N recovery efficiencies in rice plant and soil from the microplots were calculated using Eqs.(3)–(7) [34],where all15N atom % excesses (Nae) were corrected for background abundance (0.366%).

A two-way analysis of variance (ANOVA) was performed using SPSS 19.0 software (IBM SPSS Statistics,New York,USA,Version 19.0).ANOVA was used to evaluate the effects of N rate,planting density,and their interaction on grain yield,yield components,root characteristic,plant N uptake,fertilizer15N fate and NR and GS activities.Statistical significance was determined atP<0.05 andP<0.01.Pearson correlation coefficients were calculated using SPSS.

3.Results

3.1.Aboveground biomass and grain yield

In the two years,RN significantly decreased aboveground biomass,but HD increased aboveground biomass in both early and late seasons (Fig.2,P<0.05).Compared to the CK treatment,RNHD1 did not affect the aboveground biomass of early or late seasons.RNHD2 increased aboveground biomass by 7.4%(P<0.05)in the early season of 2018,but did not affect aboveground biomass in other seasons.RN did not affect rice grain yield in either season(Fig.2).HD increased grain yield by 4.7%(P<0.05)in 2017,but did not affect rice yield in 2018.RNHD1 and RNHD2 did not affect rice yield in either year.

During both early and late seasons,panicle number was decreased with reduced N rate,but increased with planting density(Table 2,P<0.05).Compared to the CK treatment,RNHD1 and RNHD2 had no effect on panicle number in either season.RN significantly reduced spikelets per panicle in the late season,but did not affect this trait in the early season (Table 2).HD significantly reduced spikelets per panicle in both seasons.Neither RN nor HD influenced filled-kernel rate or 1000-kernel weight(Table 2).Compared to the CK treatment,RNHD1 and RNHD2 did not affect spikelets per panicle,filled-kernel rate,or 1000-kernel weight in either season.

3.2.Plant N uptake and NUE

Plant N uptake was significantly reduced by RN,but significantly elevated by HD in both early and late seasons (Fig.3,P<0.05).Compared to the CK treatment,RNHD1 and RNHD2 had no influence on plant N uptake in either season.RN significantly increased AEN,but HD did not affect AEN in either season(Fig.3).Both RN and HD significantly increased REN in both seasons (Fig.3).Compared to the CK treatment,RNHD1 increased AEN by 29.2% (P<0.05) and REN by 18.4% (P<0.05),and RNHD2 increased AEN by 27.6% (P<0.05) and REN by 23.1% (P<0.05).

3.3.The fate of fertilizer 15N

During both early and late seasons in the microplot experiment,RN significantly reduced and HD significantly increased15N uptake from fertilizer (Table 3,P<0.05).Compared to the CK treatment,RNHD1 and RNHD2 did not affect15N uptake from fertilizer in the early season,but significantly reduced it the late season.During both early and late seasons,N uptake from soil was unaffected by N rate,but was significantly increased by HD (Table 3).Compared to the CK treatment,RNHD1 and RNHD2 did not affect N uptake from soil in either season.

During both early and late seasons,both RN and HD significantly increased15N recovery rate.Compared to the CK treatment,RNHD1 and RNHD2 increased15N recovery by 16.4%(P<0.05)and 25.6% (P<0.05),respectively.RN significantly reduced15N retention rate (Table 3),whereas HD did not affect15N retention in either season.RNHD1 and RNHD2 reduced15N retention by 9.5%(P<0.05) and 10.9% (P<0.05),respectively.Neither RN nor HD affected15N loss rate(Table 3).RNHD1 did not affect15N loss rate.RNHD2 reduced15N loss by 19.6%(P<0.05)in the early season but did not affect15N loss in the late season.

Table 2 Yield components under different N application rates and planting densities.

3.4.Root morphology and enzyme activity

Fig.2.Aboveground biomass and grain yield under different N application rates and planting densities.Error bars represent SEM.Different letters indicate statistical significance among treatments in the same growing season at P <0.05.* and ** represent statistical significance at P <0.05 and P <0.01,respectively.NS,not significant.There was no significant N × D interaction effect on aboveground biomass and grain yield in either early or late rice seasons.N,nitrogen;D,density;SD,standard planting density;HD1 and HD2,two higher planting densities;SN,standard N rate;RN,reduced N rate.

At most growth stages in early and late season,root length,surface area,volume,and biomass were significantly reduced by RN,but significantly increased by HD(Table 4,Fig.4,P<0.05).The difference of root morphological indexes were not significant between RNHD1 and CK.RNHD2 increased root length,surface area,volume and biomass by 8.6%,5.8%,10.8% and 9.0%,respectively.RN significantly decreased NR activity at tillering and filling stages in the late season(Fig.5).HD did not significantly affect NR activity in most growing stages of two seasons.Compared to the CK treatment,RNHD2 significantly decreased NR activity at tillering stages of two seasons and at filling stage of late season,but there was no significant difference in NR activity between CK and RNHD1 at any stage.RN significantly decreased GS activity at tillering and heading stages of two seasons (Fig.5).HD did not significantly affect GS activity at any stage of two seasons.Compared to the CK,RNHD1 and RNHD2 significantly decreased GS activity at tillering and heading stages,but did not affect GS activities at filling stage of two seasons.

Fig.3.Plant N uptake,AEN,and REN under different N application rates and planting densities.Error bars represent SEM.Different letters indicate statistical significance among treatments in the same growing season at P<0.05.*and**represent statistical significance at P<0.05 and P<0.01,respectively.NS,not significant.AEN,agronomic N use efficiency;REN,N recovery efficiency.These data were averaged across years,as there was no significant N×D×year interaction.There was no significant N×D interaction effect on plant N uptake,AEN and REN in both early and late rice seasons.N,nitrogen;D,density;SD,standard planting density;HD1 and HD2,two higher planting densities;SN,standard N rate;RN,reduced N rate.

3.5.Correlation of root traits with plant N uptake

In both early and late seasons,root length,surface area,volume,and biomass at tillering,heading,and maturity stages were significantly correlated with plant N uptake (Table 5).In contrast,therewas no significant correlation among NR and GS activity and plant N uptake (Table 5).

Table 3 The fate of fertilizer 15N under different N application rates and planting densities in a microplot experiment.

Table 4 Root morphological traits under different N application rates and planting densities.

4.Discussion

Fig.4.Root biomass under different N application rates and planting densities.Error bars represent SEM.Different letters indicate statistical significance among treatments in the same growing stage at P <0.05.* and ** represent statistical significance at P <0.05 and P <0.01,respectively.NS,not significant.There was no significant N × D interaction at any growth stage in either early or late rice seasons.N,nitrogen;D,density;SD,standard planting density;HD1 and HD2,two higher planting densities;SN,standard N rate;RN,reduced N rate.

Reduced N application at basal and tillering stages combined with higher planting density (RNHD) maintained high rice yield(Fig.2).It is believed that fertilizer N application at basal and tillering stages is important for tiller production and panicle number.In the present study,the 20% reduction in N rate at basal and tillering stages reduced panicle number in both early and late seasons (Table 2),leading to declines in aboveground biomass and grain yield (Fig.2).Moderately dense planting can increase rice panicle number and biomass production [13,15].The reduction in panicle number and aboveground biomass was compensated by increased planting density (Table 2;Fig.2).Sufficient fertilizer N at panicle initiation is an effective practice for increasing spikelet number and biomass production to ensure high yield[35,36].However,in some previous studies,RNHD treatment reduced N rate in all N fertilization periods,and thus caused a reduction in rice yield [21,37,38].For instance,Zhou et al.[38]reported that RNHD reduced grain yield,owing mainly to a decline in the number of spikelets per unit area.However,we increased planting density with reduced N rate only at early growth stages and not at panicle initiation stage,a practice that maintains rice productivity.

RNHD has been recommended[35,36]for increasing NUE in rice field.In the present study,RNHD increased NUE for AEN and REN(Fig.3).Moderately dense planting increases N uptake [39,40].Similarly,we found that HD increased N accumulation,offsetting the reduction in N uptake resulting from RN,so that RNHD achieved relatively high N uptake and markedly improved REN compared with CK.15N tracing showed that HD increased N absorption from both fertilizer and soil (Table 3),thus leading to a higher N accumulation.At present,the N pool is increased by elevated N deposition and residual N fertilizer,which increase the soil indigenous N supply in rice fields [8].For this reason,moderately dense planting is of benefit to meet the N nutrient demand for high rice yield when fertilizer N application is reduced.Biomass and tissue N concentration are main factors determining plant N accumulation [10].In the present study,the high N absorption amount observed for RNHD produced high biomass without leading to a sharp reduction in plant N concentration.Toriyama et al.[41]suggested that dense planting with high N rate reduced dry matter accumulation and N uptake owing to high-density stresses.However,in our study,20% or 40% increases in planting density increased aboveground biomass regardless of N rate,indicating that the N rate and the increase of planting density might not cause high-density stresses.RNHD also reduced15N retention and15N loss (Table 3).RNHD thus offers the benefit of reducing environmental pollution by N fertilizer.However,given that HD reduced fertilizer N residue in soil while increasing the absorption of N from soil,the long-term effect of RNHD on soil N pools remains an open question.

Root morphology influences N uptake of the rice plant [20].A better-developed rice root system effectively increased N absorption by flooded rice[42].Our results showed that RN reduced root length,surface area,volume,and biomass(Table 4;Fig.4).But the negative effects of RN on these root traits could be compensated by increasing planting density.Accordingly,these root morphological attributes of RNHD2 were generally higher than those of the CK.Root length,surface area,volume and biomass at all three growth stages were strongly positively correlated with plant N accumulation in most growth stages (Table 5).In addition to root morphology,the activities of NR and GS are closely associated with plant N uptake [43,44].N application at panicle initiation is necessary to maintain NR and GS activities in the reproductive stage [36].In our study,constant N rate at panicle initiation avoided the clear reduction in the activities of NR and GS at heading and grain filling stages (Fig.5).The activities of NR and GS showed no significant correlation with plant N uptake.Overall,HD produced a larger root system that increased plant N uptake under a reduced N rate.Thus,RNHD maintained root growth and plant N uptake as well as increasing NUE.

Table 5 Correlations between plant N uptake and root traits.

Fig.5.Activities of nitrate reductase and glutamine synthetase under different N application rates and planting densities.Error bars represent SEM.Different letters indicate statistical significance among treatments in the same growing stage at P <0.05.* and ** represent statistical significance at P <0.05 and P <0.01,respectively.NS,not significant.There was no significant N×D interaction at any growth stage in either early or late rice seasons.N,nitrogen;D,density;SD,standard planting density;HD1 and HD2,two higher planting densities;SN,standard N rate;RN,reduced N rate.

5.Conclusions

In summary,our results supported that RNHD increased agronomic NUEs and N recovery efficiency of rice,but its effects on the rice yield,biomass production and plant N uptake were not obvious.Moreover,RNHD significantly improved fertilizer15N recovery,while it reduced15N retention.However,the reduced N rate resulted the decline of root length,surface area,volume,and biomass,but higher planting density improved these root characteristics.Taken together,RNHD presented the weak influence on the morphological characteristics of roots.Furthermore,RNHD reduced the activities of NR and GS only at tillering stage.The correlation analysis indicated that plant N uptake was positively correlated with root length,surface area,volume,and biomass.Ourresearch results support that RNHD may be an environmentfriendly cultivation pattern with high yield and high NUE in a double-rice cropping system.

CRediT authorship contribution statement

Jin Chen,Minggang Xu,andChunrui Pengdesigned the research.Jiang Xie,Guoqiang Deng,Tianhua Tu,Xianjiao Guan,Zhen Yang,Xianmao Chen,Caifei Qiu,Yinfei Qian,andCaihong Shaoconducted the experiments and data collection.Jin Chen,Xiangcheng Zhu,Minggang Xu,Chunrui Peng,andShan Huangcontributed to data analysis and interpretation and wrote the manuscript.All authors commented on and approved the final manuscript.

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 Natural Science Foundation of China(31601263),the National Key Research and Development Program of China (2018YFD0301103),the China Postdoctoral Science Foundation(2017M622100),the Jiangxi Province Postdoctoral Science Foundation(2017KY11),and the Open Foundation of Guangxi Key Laboratory of Rice Genetics and Breeding(160-380-16-2).