Effects of mechanized deep placement of nitrogen fertilizer rate and type on rice yield and nitrogen use efficiency in Chuanxi Plain,China

ZHU Cong-hua,OUYANG Yu-yuan,DlAO You, ,YU Jun-qi,LUO Xi,ZHENG Jia-guo, ,Ll Xu-yi

1 Crop Research Institute,Sichuan Academy of Agricultural Sciences,Chengdu 610066,P.R.China

2 Sichuan Province Key Laboratory of Crop Ecophysiology and Cultivation,Chengdu 611130,P.R.China

Abstract This paper investigates the yield and nitrogen use efficiency (NUE) of machine-transplanted rice cultivated using mechanized deep placement of N fertilizer in the rice–wheat rotation region of Chuanxi Plain,China. It provides theoretical support for N-saving and improves quality and production efficiency of machine-transplanted rice. Using a single-factor complete randomized block design in field experiments in 2018 and 2019,seven N-fertilization treatments were applied,with the fertilizer being surface broadcast and/or mechanically placed beside the seedlings at (5.5±0.5) cm soil depth when transplanting.The treatments were:N0,no N fertilizer; U1,180 kg N ha–1 as urea,surface broadcast manually before transplanting;U2,108 kg N ha–1 as urea,surface broadcast manually before transplanting,and 72 kg N ha–1 as urea surface broadcast manually on the 10th d after transplanting,which is not only the local common fertilization method,but also the reference treatment; UD,180 kg N ha–1 as urea,mechanically deep-placed when transplanting; M1,81.6 kg N ha–1 as urea and 38.4 kg N ha–1 as controlled-release urea (CRU),mechanically deep-placed when transplanting; M2,102 kg N ha–1 as urea and 48 kg N ha–1 as CRU,mechanically deep-placed when transplanting; M3,122.4 kg N ha–1 as urea and 57.6 kg N ha–1 as CRU,mechanically deep-placed when transplanting. The effects of the N fertilizer treatments on rice yield and NUE were consistent in the 2 yr. With a N application rate of 180 kg ha–1,compared with U2,the N recovery efficiency (NRE),N agronomic use efficiency (NAE) and yield under the UD treatment were 20.6,3.5 and 1.1% higher in 2018,and 4.6,1.7 and 1.2% higher in 2019,respectively. Compared with urea alone (U1,U2 or UD),the NRE,NAE and yield achieved by M3 (combined application of urea and controlled-release urea) were higher by 9.2–73.3%,18.6–61.5% and 6.5–16.5%(2018),and 22.2–65.2%,25.6–75.0% and 5.9–13.9% (2019),respectively. Compared with M3,the lower-N treatments M1 and M2 significantly increased NRE by 4.0–7.8% in 2018 and 3.1–4.3% in 2019,respectively. Compared with urea surface application (U1 or U2),the yield under the M2 treatment was higher by 4.3–12.9% in 2018 and 3.6–10.1% in 2019,respectively. Compared with U2,the NRE and NAE under the M2 treatment was higher by 36.9 and 36.3% in 2018,and 33.2 and 37.4% in 2019,mainly because of higher N uptake. There was no significant difference in the concentration of nitrate in the top 0–20 cm soil under U1,U2 and M2 treatments during the full heading and maturity stages. During the full heading stage,U2 produced the highest concentration of nitrite in 0–20 cm and 20–40 cm soil among the N fertilizer treatments. In conclusion,mechanized deep placement of mixed urea and controlled-release urea (M2) at transplanting is a highly-efficient cultivation technology that enables increased yield of machine-transplanted rice and improved NUE,while reducing the amount of N-fertilization applied.

Keywords:rice,N-fertilization rate,controlled release urea,side deep fertilization,yield,nitrogen use efficiency

1.lntroduction

Nitrogen (N) is the main nutrient limiting the quality and yield of rice and nearly half of the world’s people live on rice as their basic food (Aslamet al.2015). The average of local common N application rate for single-crop rice in the rice–wheat rotation region of the Chuanxi Plain,China is 180 kg ha–1,which is 50% higher than the global average,but the N recovery efficiency (NRE) is only 30–35% (Moroet al.2015; Sunet al.2016; Jianget al.2017; Islamet al.2018).Excessive application of N fertilizer in pursuit of higher rice yield not only increases production costs,but also reduces rice quality and fertilizer use efficiency,as well as damaging rice field ecology (Juet al.2009). In the rice–wheat rotation region of the Chuanxi Plain,the rural labor force is tending to migrate to the city,increasingly causing seasonal shortages of age-appropriate labor for rice production,and thereby rapidly increasing labor costs,and 70–80% in total N as urea surface broadcast manually before transplanting,and 20–30% in total N as urea surface broadcast manually on the 10th d after transplanting widely used by famers to reduce production costs resulted in much lower N use efficiency(NUE) (Jianget al.2017). The development of N-saving and highly efficient planting technology,including mechanized operation,is urgent for large-scale rice production.

The N application rate and N application method are consistently considered as the key factors affecting rice yield and NUE. By optimizing N application in the forest agro-ecological zone of Ghana and in Northeast Spain(Moroet al.2015; Moreno-Garcíaet al.2017),20–30%reduction in N fertilizer input not only maintained rice yield,but also improved the NUE,which is beneficial for ecological and environmental protection (Xueet al.2014; Zhanget al.2017). Combining lower N application (by 20–30%)with controlled/slow-release urea maintains rice yield and improves the NRE while reducing the number of fertilizer application times (Huanget al.2017; Tianet al.2018a,b).Furthermore,urea deep placement is a proven technology that increases yield and NUE when compared with the conventional broadcast application of urea (Daset al.2015; Hasanet al.2016; Miahet al.2016). Mechanized transplanting with deep placement of fertilizer is an emerging and efficient planting technology. Specifically,the process is to apply the fertilizer into the soil at the side of the seedlings at the same time as mechanized transplanting,thereby achieving precision fertilization,reducing the labor intensity of fertilizer application,reducing N fertilizer loss,and improving the N supply capability of the soil,while improving planting efficiency and ultimately increasing rice yield and NUE (Keet al.2018; Zhuet al.2019).

Compared with conventional split applications comprising base fertilizer,tiller fertilizer and panicle fertilizer,applying urea as the N source with side deep placement of fertilizer reduced the total N applied by 10–15%; in combination with a single broadcast application of panicle fertilizer,this method increased the rice yield by 0.7–7.7% and the NUE by 18.2% (Zhaoet al.2017). Compared with conventional fertilization (150 kg N ha–1,with base fertilizer:tiller fertilizer=6:4),one-time side deep placement of N fertilizer(30% reduction in N application rate) treatment produced no significant reduction in yield,total N uptake and soil alkaline N concentration in maturity stage,and greatly increased the N agronomic use efficiency (NAE) and NRE for early and late rice (Zhonget al.2019). However,the N-saving potential of the combination of mechanized side deep placement fertilization with controlled-release N fertilizer in mechanized transplanting in the rice–wheat rotation region of the Chuanxi Plain has not been established. Therefore,in the current study,urea and/or controlled-release urea were applied as N sources in the rice season of the rice–wheat cropping region in the Chuanxi Plain. Using a single mechanical deep-placement of reduced-N fertilizer containing the two N sources combined,the effects on rice yield and N fertilizer use,as well as the potential for reducing N use,were investigated. This study provides guidelines for the promotion and application of simultaneous mechanized transplanting and deep-fertilization using N-saving and highly efficient technology.

2.Materials and methods

2.1.Experimental sites and weather conditions

The field experiments were conducted from April to October in 2018 and 2019,respectively,at the Experimental Research Farm of Crop Research Institute,Sichuan Academy of Agricultural Sciences,Xiaode Town,Mianzhu City,Sichuan Province,China (31.26°N,104.22°E). In general,this region located in the Chuanxi Plain has a monsoon and sub-tropical climate. The soil of the test field is sandy loam,which is easy to drain and irrigate; the previous crop was wheat. Prior to the trial in 2018,the properties of the 0–25 cm soil were as follows:pH 5.82,organic matter 20.63 g kg–1,available N 55.21 mg kg–1,available P 11.82 mg kg–1,available K 75.29 mg kg–1,total N 1.65 g kg–1,total P 0.93 g kg–1,and total K 17.99 g kg–1.

2.2.Experimental materials

The rice variety used was Huanghuazhan,a conventionalindicarice with an average growth period of 145.8 d. The N fertilizers used were urea (N content≥46.0%,Sichuan Meifeng Chemical Industry Co.,Ltd.,China) and controlledrelease urea (CRU) (N content≥41.6%,Jinzhengda Ecological Engineering Group Co.,Ltd.,average release period 80 d at 25°C).

2.3.Experimental design and treatments

Using a single-factor randomized-block design,seven N-fertilization treatments were applied:N0,no N fertilizer;U1,180 kg N ha–1as urea,surface broadcast manually before transplanting; U2,108 kg N ha–1as urea surface broadcast manually before transplanting,and 72 kg N ha–1as urea,surface broadcast manually on the 10th d after transplanting,which is not only the local common fertilization method,but also the reference treatment; UD,180 kg N ha–1as urea,mechanically deep-placed when transplanting (for details,see next paragraph); M1,81.6 kg N ha–1as urea and 38.4 kg N ha–1as CRU,mechanically deep-placed when transplanting; M2,102 kg N ha–1as urea and 48 kg N ha–1as CRU,mechanically deep-placed when transplanting; M3,122.4 kg N ha–1as urea and 57.6 kg N ha–1as CRU,mechanically deep-placed when transplanting.In treatments M1,M2 and M3,the two types of fertilizers were mixed then used as base fertilizer that was applied during transplantingviamechanized deep-placement.Each treatment was repeated three times,the area of each block was 60 m2,and ridges with height of 30 cm and width of 30 cm were built among the blocks. Each block was covered with plastic film,and drained and irrigated independently. Each treatment received 93.6 kg P2O5ha–1as superphosphate and 93.6 kg K2O ha–1as potassium chloride. Both phosphate and potassium fertilizers were used as base fertilizers and applied before transplanting.

On March 25,2018 and March 29,2019,seeds were sown in seedling substrate (in 58-cm long and 28-cm wide seedling trays) with (110±5) g seeds per tray. After emergence,the seedlings were irrigated and cultivated as required. On May 1,2018 and May 9,2019,seedlings were transplanted using a Kubota NSPU-68CMDF rice transplanter. At the same time as transplanting,ditch fertilization was conducted using a robot arm at(5.0±0.5) cm on one side of the seedlings. The ditch depth was (5.5±0.5) cm and the soil was backfilled to cover the fertilizer. The row spacing and plant spacing were 30 and 16 cm,respectively,with 3–4 seedlings being planted in each hole. All the blocks were subjected to frequent shallow irrigation in the early growth stage. When the tiller number in the field reached 80% of the expected panicle number,the field was drained and allowed to dry for 7 d. Afterwards,alternate irrigation and drying was applied to the field week by week,until irrigation was terminated 7 d before harvest.The control of diseases,insects and weeds was the same as used in ordinary field cultivation in the region.

2.4.Sampling and analyses

Plant N content,N uptake and NUEDuring panicle initiation,full heading and maturity stages,three bundles of representative plants were selected,each with the average number of tillers in the respective block. The leaves,stems (including leaf sheaths) and panicles (at full heading and maturity stages) were subjected to high-temperature desiccation at 105°C for 30 min and then dried at 80°C to constant weight. The dry weights were measured.

The dried samples of panicles,leaves and stems with leaf sheath were milled and sieved through an 80-mesh sieve. After digestion with concentrated H2SO4+H2O2at 420°C for 2 h,the N concentrations were measured using a fully automatic Kjeldahl azotometer (KjeltecTM8400,FOSS,Denmark).

N uptake was calculated using the formula TDM×NC,where TDM represents the total dry matter of panicles,leaves and stems with leaf sheaths,and NC represents the N concentration in panicles,leaves and stems with leaf sheaths.

Aspects of NUE,namely N dry matter production efficiency (NDMPE),N grain production efficiency (NGPE),N recovery efficiency (NRE),and N agronomic use efficiency(NAE) were calculated using the following formulas:

NDMPE=TBm/TNup

NGPE=GY/Nup

NRE=(Nup–N0up)/FN

NAE=(GY–GY0)/FN

whereTBm,the total dry matter accumulation at maturity stage;TNup,total N accumulation at maturity stage; GY,grain yields in N-fertilized plots;NupandN0up,total N uptake above-ground in N-fertilized plots and N0 plots at maturity stage,respectively; GY0,grain yields in N0 plots; FN,the total N application rate in N-fertilized plots.

Ammonium,nitrate and nitrite concentrations in surface waterOn the 20th d after transplanting,100-mL surface water was collected at 20 locations in each block using a 100-mL syringe through the S-type sampling method,and mixed to make one composite surface water sample.The water samples were brought back to the laboratory and centrifuged (5 000 r min–1) to remove solid impurities.The remaining solutions were refrigerated for future use.The concentrations of nitrate,nitrite and ammonium were determined by salicylic acid nitrification,sulfanilamide colorimetry and indophenol-blue colorimetry,respectively.

Ammonium,nitrate and nitrite concentrations in soilDuring the full heading and maturity stages,soil samples at 0–20,20–40 and 40–60 cm depths were collected from each block with an earth drill in accordance with the five-point sampling method,and thoroughly mixed. Some sub-samples were air-dried and sieved through a 0.149-mm sieve then the ammonium content was determined. Other sub-samples were stored refrigerated until determination of the nitrate content.Further sub-samples were stored at −20°C until determination of nitrite. The concentrations of nitrate,nitrite and ammonium were determined by salicylic acid nitrification,sulfanilamide colorimetry and indophenol-blue colorimetry,respectively.

Yield and its componentsOn the day before harvest,the effective panicles of 30 rice bundles in each block were sampled for yield. Six bundles of representative plants were selected,each with the average panicle number in the respective block,to measure yield indicators,namely seed setting rate,grain number per panicle and thousandgrain weight. A further 15 m2of rice was threshed manually and then dried. Afterwards,it was adjusted to a standard moisture content of 13.5% for yield calculation. Each treatment was sampled in triplicate.

2.5.Data analysis and figure production

The datasets in 2018 and 2019 were analyzed using SAS®9.1. Pairwise comparisons of means among treatments were conducted by the least significant difference (LSD)tests at the 0.05 level of probability. The figures were produced by OriginPro 2017.

3.Results

3.1.Grain yield and yield components

In 2018 and 2019,the yield of the M3 treatment was significantly higher than those of U1,U2,UD,and M1(Table 1). Compared with U1,the yields of U2,UD,M2,and M3 treatments were significantly higher by 8.2,9.4,12.9,and 16.5% in 2018; and 6.3,7.6,10.1,and 13.9% in 2019,respectively. Compared with U2,the yields of UD,M2 and M3 treatments were higher by 1.1,4.3 and 7.6% in 2018; and 1.2,3.6 and 7.1% in 2019,respectively. While there was no significance among the yields of U2,UD and M2,the yield of M3 was significantly higher than that of M1 by 13.8% in 2018 and 12.5% in 2019,respectively. The effective panicle numbers of M3 were significantly higher than those of U1 and U2. Compared with U2,the effective panicle numbers of UD and M3 treatments were significantly higher by 3.6 and 8.3% in 2018; and 4.3 and 4.4% in 2019,respectively. The grain number per panicle,seed setting rate and 1 000-grain weight of U2,UD,M2,and M3 did not differ significantly.

3.2.N uptake and distribution

The method and rate of N application affected N uptakeof rice (Fig.1). During the panicle initiation stage,among all N application methods,M3 induced the highest total N uptake,followed by UD,with the lowest observed for N0.There was no significant difference in the total N uptake among U2,UD and M2. At the full heading stage,the highest total N uptake was again observed for M3,followed by UD,and the lowest for N0. Compared with U2,UD,M2 and M3 treatments produced much higher N uptake in 2018,and M1,M2 and M3 treatments produced much higher N uptake in 2019. At the maturity stage,M3 still had the highest total N uptake among all treatments,followed by UD and M2,with N0 the lowest. Compared with U2,UD,M2 and M3 treatments produced much higher N uptake in 2018,and M3 treatment produced much higher N uptake in 2019. At maturity,the N distribution ratio in the panicles ranged from 64.0 to 71.6%,with the highest and lowest values in the N0 and M3 treatments,respectively.

Table 1 Effects of nitrogen (N) fertilizer treatments on rice yield and its components

3.3.NUE

There were significant differences in the effects of N treatments on NUE of rice (Table 2). There was no significant difference in the N dry matter production efficiency and N grain production efficiency among U2,UD and M2.Compared with manual broadcast fertilizer application(treatments U1 and U2),mechanized deep-placement fertilizer application significantly improved the NRE and NAE of machine-transplanted rice. Compared with U2,the treatments of UD,M1,M2,and M3 increased NRE by 20.6,42.0,36.9,and 31.7%,and enhanced NAE by 3.5,26.9,36.3,and 22.8%,respectively,in 2018. Compared with U2,the treatments of UD,M1,M2,and M3 increased NRE by 4.6,31.8,33.2,and 27.8%,and enhanced NAE by 1.7,15.7,37.4,and 27.8%,respectively,in 2019.

3.4.Ammonium,nitrate and nitrite concentrations in surface water

Fig.1 Effects of nitrogen (N) fertilizer treatments on N uptake and distribution of rice. N0,no N fertilizer; U1,180 kg N ha–1 as urea,surface broadcast manually before transplanting; U2,108 kg N ha–1 as urea surface broadcast manually before transplanting,and 72 kg N ha–1 as urea surface broadcast manually on the 10th d after transplanting,which is not only the local common fertilization method,but also the reference treatment; UD,180 kg N ha–1 as urea,mechanically deep-placed when transplanting; M1,81.6 kg N ha–1 as urea and 38.4 kg N ha–1 as controlled-release urea (CRU),mechanically deep-placed when transplanting; M2,102 kg N ha–1 as urea and 48 kg N ha–1 as CRU,mechanically deep-placed when transplanting; M3,122.4 kg N ha–1 as urea and 57.6 kg N ha–1 as CRU,mechanically deep-placed when transplanting. Columns marked with same lowercase letters were not significantly different among all treatments in 2018 and 2019 (P>0.05,LSD multiple test). Error bars indicate standard deviation (n=3).

Table 2 Effects of nitrogen (N) fertilizer treatments on N use efficiency of rice

Fig.2 Effects of nitrogen (N) fertilizer treatments on ammonium,nitrite and nitrate concentrations in surface water 20 d after transplanting. N0,no N fertilizer; U1,180 kg N ha–1 as urea,surface broadcast manually before transplanting; U2,108 kg N ha–1 as urea surface broadcast manually before transplanting,and 72 kg N ha–1 as urea surface broadcast manually on the 10th d after transplanting,which is not only the local common fertilization method,but also the reference treatment; UD,180 kg N ha–1 as urea,mechanically deep-placed when transplanting; M1,81.6 kg N ha–1 as urea and 38.4 kg N ha–1 as controlled-release urea (CRU),mechanically deep-placed when transplanting; M2,102 kg N ha–1 as urea and 48 kg N ha–1 as CRU,mechanically deep-placed when transplanting; M3,122.4 kg N ha–1 as urea and 57.6 kg N ha–1 as CRU,mechanically deep-placed when transplanting. Columns marked with same lowercase letters were not significantly different among all treatments in 2018 and 2019 (P>0.05,LSD multiple test). Error bars indicate standard deviation (n=3).

The method and rate of N application significantly affected the N content in the surface water of the field 20 d after transplantation (Fig.2). The concentrations of ammonium,nitrite and nitrate in the surface water after N application were significantly higher than those in N0 treatment. Among the urea treatments,the concentrations of ammonium,nitrite and nitrate in the surface water after U1 treatment were significantly higher than those after U2 or UD treatments.Compared with U2,M1,M2 and M3 treatments significantly reduced ammonium and nitrite concentrations in the surface water,and significantly increased nitrate concentration in the surface water. Among the controlled-release urea treatments,M3 resulted in significantly higher concentration of ammonium in the surface water than M1 or M2. The ammonium and nitrite concentrations in the surface water after M1,M2 and M3 treatments were significantly lower than those after UD treatment,whereas M2 and M3 treatments led to significantly higher nitrate concentration in the surface water than did UD. Therefore,it can be deduced that the deep placement of mixed urea and controlled-release urea restricts the volatilization of NH3from surface water.

3.5.Ammonium,nitrate and nitrite concentrations in soil

Fig.3 Effects of nitrogen (N) fertilizer treatments on ammonium,nitrate and nitrite concentration in different soil layers at full heading and maturity stages. N0,no N fertilizer; U1,180 kg N ha–1 as urea,surface broadcast manually before transplanting; U2,108 kg N ha–1 as urea surface broadcast manually before transplanting,and 72 kg N ha–1 as urea surface broadcast manually on the 10th d after transplanting,which is not only the local common fertilization method,but also the reference treatment; UD,180 kg N ha–1 as urea,mechanically deep-placed when transplanting; M1,81.6 kg N ha–1 as urea and 38.4 kg N ha–1 as controlled-release urea(CRU),mechanically deep-placed when transplanting; M2,102 kg N ha–1 as urea and 48 kg N ha–1 as CRU,mechanically deepplaced when transplanting; M3,122.4 kg N ha–1 as urea and 57.6 kg N ha–1 as CRU,mechanically deep-placed when transplanting.Within the same soil layer,columns marked with same lowercase letters were not significantly different among all treatments in 2018 and 2019 (P>0.05,LSD multiple test). Error bars indicate standard deviation (n=3).

The N in the 0–60 cm depth of soil was mainly in the forms of ammonium and nitrate (Fig.3). The ammonium,nitrate and nitrite concentrations in different soil layers were significantly affected by the method and rate of N fertilization. In a given soil layer,the ammonium and nitrate concentrations were higher at the full heading stage than at the maturity stage,whereas the nitrite content was lower than that at the maturity stage. At the full heading and maturity stages,for the same N treatment,the ammonium content decreased gradually at greater soil depths. The ammonium concentrations in the 0–20,20–40 and 40–60 cm layers of U1-,U2-,UD-,M1-,and M2-treated soil did not vary much.However,the ammonium concentrations in the 0–20 and 20–40 cm M3-treated soil layers were significantly higher than those in U1-or U2-treated soil. The nitrate content increased gradually with the depth of the soil layer at the full heading and maturity stages. No significant difference was observed for the nitrate concentration in the 0–20 and 20–40 cm soil layers among U1,U2,UD,and M1 treatments.However,the M3 treatment resulted in higher nitrate content in the 0–20,20–40 and 40–60 cm soil layers compared to U1,U2,UD,and M1. At the full heading and maturity stages,the nitrite concentration in the 0–20 and 20–40 cm M2-treated soil layers were lower than those in UD-treated soil. The nitrite concentration was higher in the 20–40 and 40–60 cm soil layers in M1,M2 and M3 treatments compared to U1-or U2-treated soil at the full heading stage,while the opposite trend was observed at the maturity stage.

4.Discussion

Fertilizer placement methods affected grain yields and NUE differently with rate and type of N fertilizer. Compared with U1,UD increased yield by 7.6–9.4% (Table 1),and increased NRE and NAE by 35.2–58.7% and 36.2–39.3%(Table 2),respectively. Compared with U2 (the local common fertilization method),UD increased yield by 1.1–1.2%,and increased NRE and NAE by 4.6–20.6% and 1.7–3.5%,respectively. In 180 kg N ha–1as urea treatment,the trend of grain yield,NRE and NAE was UD>U2>U1. It is likely that broadcast urea increased N loss,causing a limited supply of N to the plants,which resulted in lower yields. Generally,particularly with 180 kg N ha–1as urea,mechanized deep placement of urea increased grain yields,NRE and NAE consistently compared to urea,surface broadcast manually,which are consistent with previous studies conducted in Bangladesh (Mamunet al.2017; Reaet al.2019). Compared with UD,M3 increased yield by 5.9–6.5%,and increased NRE and NAE by 9.2–22.2% and 18.6–25.6%,respectively. Increased grain yields with M3 is mainly associated with an increased effective panicle. It is likely that controlled-release urea reduced N loss,causing a continuous supply of N to the plants,that resulted in higher yields. This is consistent with previous study conducted in Jiangsu Province (Keet al.2018). Compared with U2,M3 significantly increased yield by 7.1–7.6%,and increased NRE and NAE by 27.8–31.7% and 22.8–27.8%,respectively.This implied that mechanized deep placement of urea and controlled-release urea produced much higher yiled by significantly reducing N losses compared with urea,surface broadcast manually. Furthermore,the magnitude of grain yield,NRE and NAE increase varied with N fertilizer rate and type. M2 with 16.7% less N input than UD or U2 increased grain yields by 2.4–4.3%,NRE by 13.6–36.9% and NAE by 31.6–37.4%,respectively,which is basically consistent with previous studies conducted in Heilongjiang Province(Zhaoet al.2017).

Controlled-release N fertilizer enables progressive release and reduces loss of fertilizer N,and ensures that the fertilizer N released matches the nutrient absorption of plants to the maximum extent,promoting effective N absorption by the root system and improving NUE. Therefore,controlledrelease N fertilizer is suitable for a single application as base fertilizer. A15N isotope study found that the NRE was 39%higher with a single application of 90 kg N ha–1controlledrelease N fertilizer than when urea was applied twice (Jiet al.2011). Application of controlled-release N fertilizer on river sand and mud reduced N use by 30% and enhanced NUE by 39% compared with urea application (Luet al.2016).With deep placement of controlled-release N fertilizer,rice yield and NUE were further increased (Keet al.2018; Zhuet al.2019). In this study,with a N application rate of 180 kg ha–1,the deep placement of mixed urea and controlledrelease urea improved the NRE by 9.2–73.3% compared with application of urea alone,which is consistent with the results of the aforementioned studies.With increased N application rate,the NRE of rice displayed a downward trend while the N harvest index increased,which is also in accordance with previous studies (Zhu and Chen 2002; Luet al.2008). In addition,the NRE under the M2 treatment(deep application of a reduced amount of mixed urea and controlled-release urea with N application rate 150 kg ha–1)reached a maximum of 51.9%,which was 37.0 and 80.0%higher,respectively,than the treatments with one or two applications of urea (U1 and U2). The NAE under the M2 treatment was as high as 15.8–23.3 kg kg–1,which was 36.3–88.1% higher than that with application of urea alone. Two possible reasons for this might be as follows:First,with the mechanized deep application of mixed controlled-release N,fertilizer N is delivered directly near the root zone of the seedlings. Urea is rapidly converted to ammonium under the action of microbial urease; the ammonium is partially adsorbed and fixed by the soil. The rest gradually diffuses to the root system and is absorbed and utilized,reducing loss N by volatilization and runoff in the surface water and promoting rapid and efficient N uptake by seedlings (Liet al.2013; Houet al.2019). In this way,the greening process is shortened and the requirements of N nutrition in the early and middle stages of rice growth can be fully supplied (Maet al.2017); second,the slow release of N from controlledrelease fertilizer ensures a steady N supply in the late stage of rice growth and enhances N transfer and reuse in stems and leaves (Fanet al.2007). This in turn not only avoids the tendency for the plants to remain green and shows a growth spurt,but also prevents premature senescence of the leaves. In summary,by reducing the amount of N fertilizer by 30 kg ha–1and applying the fertilizer by mechanized deep placement,N uptake by rice can be enhanced,leading to higher nutrient efficiency and stable or even increased yield.

N loss from fertilizers applied to paddy fields is mainly through volatilization,runoff and leakage. N volatilization and runoff loss in paddy fields are significantly positively related to the concentrations of various forms of N in the surface water after transplanting (Yuet al.2013; Tianet al.2018a,b). In this study,the concentrations of ammonium and nitrite in the surface water after broadcast application of urea (U1 and U2) increased rapidly after 20 d from transplantation,and were significantly higher than those of the mechanized deep placement treatments (UD,M1,M2,and M3). However,the opposite was observed for the nitrate concentration. The concentrations of ammonium,nitrite and nitrate in the surface water after broadcast application of urea were higher than those after the deep placement treatments,which means that mechanized deep placement can reduce N loss from paddy fields through runoff and volatilization. The N concentration in the leakage water increases rapidly within 1–2 d after fertilizer application (Sun and Liu 2007),and the key period to regulate N leaching is within 7 d after treatment (Jiet al.2007a,b). Climatic conditions,soil characteristics,crop types,farming systems,irrigation methods,and fertilizer management are major factors that affect N leaching (Huet al.2011; Yinet al.2013). The concentrations of ammonium and nitrate in the root-zone soil are closely related to N uptake by rice plants(Liet al.2008). In this study,at deeper soil layers (within the range 0–60 cm),the ammonium content gradually decreased; in contrast,the nitrate content showed a slight increasing trend,while the nitrite content was broadly similar at 20–40 and 40–60 cm depths. These trends indicate that nitrate was the main form of soil leakage loss,which is basically consistent with previous studies (Tianet al.2018b,c). The nutrient release rate from controlled-release fertilizer is slow in the early growth stage and relatively high in the middle and late stages,which is beneficial in enabling rice plants to absorb more N during the vigorous vegetative growth and reproductive growth stages,thus reducing the N concentration in the soil leakage water (Duet al.2007) and the extent of N loss from the fertilizer (Zhenget al.2004). In addition,the application of mixed urea and controlled-release urea resulted in higher ammonium and nitrate concentrations in 0–20 and 20–40 cm soil layers during the heading and maturity stages,compared with urea alone. Moreover,the M3 treatment led to the highest soil ammonium and nitrate concentrations among the treatments,coupled with lower nitrite content in the 0–20 cm soil layer; and the concentrations of nitrite in all soil layers during the heading stage under the UD treatment were significantly higher than those under the other treatments,which was different from previous studies (Zhanget al.2017;Keet al.2018; Zhonget al.2019). The reason might be that,in this study,urea and controlled-release urea were mixed and applied using the mechanized deep placement method (at depth (5.5±0.5) cm). Urea is converted into ammonium by soil microbial urease and enters the soil.While most of the ammonium is fixed by the soil,it is partially absorbed by the plant roots,and partially enters the soil through nitrification and denitrification. However,nitrate and nitrite cannot be easily adsorbed and fixed by the soil,and will move downward with the leakage water.The hypoxic conditions in soil deeper than 40 cm strengthen the denitrification pathway,resulting in the stable or even slightly increased content of nitrite in the soil deeper than 40 cm. Therefore,the deep placement of urea might increase the risk of leakage loss while reducing the runoff and volatilization loss of N. In contrast,the mechanized deep placement of mixed urea and controlled-release urea cannot only fulfill the N requirement of rice,but also greatly reduce the risk of fertilizer N runoff,volatilization and leakage loss. Compared with a single application of urea to the surface,the ammonium and nitrate concentrations in the 0–20 and 20–40 cm soil layers at the maturity stage did not vary significantly in the M2 treatment. This further confirms that the mechanized deep placement of mixed urea and controlled-release urea is an efficient technology that reduces loss of N and increases NUE and yield.

The optimum N application rate for rice is closely related to factors such as variety characteristics (Sunet al.2016),paddy soil fertility (Zhanget al.2011),ecological conditions(Jianget al.2017),planting methods (Wanget al.2015),and N fertilizer type (Zhuet al.2019). Numerous studies have shown that in areas with high N application rates,such as Jiangsu Province,the rice yield does not decrease significantly if N fertilizer use during the rice growing season is reduced by up to 30% (Qiaoet al.2012; Guoet al.2019).The study has indicated that,through the combination of mechanized deep placement of N fertilizer and a single surface application of panicle fertilizer,the yield can be maintained despite a 15% reduction in N applied,and the NUE can be significantly improved (Liu and Fan 2019). In the current study,compared with the U1 and U2 treatments(N application rate 180 kg ha–1),the yield was enhanced by 10.1–12.9% and 3.6–4.3%,respectively,in the M2 treatment(single deep application of mixed urea and controlledrelease urea with N application rate 150 kg ha–1). With the M1 treatment (single deep application of mixed urea and controlled-release urea with N application rate 120 kg ha–1),the yield was 1.3–2.4% higher than that in U1 treatment and 5–5.7% lower than that in U2 treatment. Therefore,it is suggested that,in the rice–wheat rotation region of the Chuanxi Plain,in order to ensure a high yield of rice and the efficient use of N fertilizer,the N application rate should be kept above 150 kg ha–1with a single application of fertilizer by deep placement of mixed urea and controlled-release urea.

5.Conclusion

A single application of mixed urea and controlled-release urea fertilizer by mechanized deep placement avoids one top dressing operation and reduces the risk of N fertilizer loss through runoff,volatilization and leakage. At the same time,it maintains a high N-supply capacity of the soil,promotes efficient N uptake by rice,enhances N reuse in stems and leaves,and significantly increases rice yield and NUE. In the rice–wheat rotation region of the Chuanxi Plain,compared with the local common fertilization method (U2,108 kg N ha–1as urea surface broadcast manually before transplanting and 72 kg N ha–1as urea surface broadcast manually on the 10th d after transplanting),a single mechanized deep application of mixed urea and controlled-release urea with the amount of N applied reduced by up to 30 kg ha–1increased the yield by 3.6–4.3%,the NRE by 33.2–36.9%and the NAE by 36.3–37.4%.

Acknowledgements

This study was supported by the National Key Research and Development Program of China (2016YFD0300108),the Application and Basic Research Project of Sichuan Province,China (2018JY0630) and the Financial Innovation Capacity Improvement of Sichuan Province,China (2017QNJJ-031).We thank Drs.Jennifer Smith and Huw Tyson,from Liwen Bianji,Edanz Group China (www.liwenbianji.cn/ac),for editing the English text of a draft of this manuscript.