Is there a nitrogen fertilizer threshold emitting less N2O with the prerequisite of high wheat production?

Excessive use of synthetic nitrogen (N) fertilizer and lower nitrogen use efficiency (NUE) are threatening the wheat production in the middle and lower reaches of Yangtze River. Excess input of N fertilizers also results in severe environmental pollution, climate change and biodiversity loss. However, the study on reasonable nitrogen application and NUE improvement with the prerequisite of stable and high yield remains unexplored. In our study, the four different levels of nitrogen were applied to find out the nitrogen threshold which could be both friendly to environment and promise the stable and high yield. The experiment was carried out in Yangzhou University (Yangzhou, China). The wheat cultivar Yangmai 23 was selected as the research material. The four nitrogen levels were as follows: 0, 189, 229.5, and 270 kg ha-1. The results showed that the grain yield under the application of 229.5 kg ha-1 N was as high as that under 270 kg ha-1 N level, with the observation of 20.3% increase in agronomic efficiency. The N2O emission of 229.5 kg ha-1 N application was as low as that of 189 kg ha-1 N, but the grain yield and agronomic efficiency were significantly higher (11.9%) under 229.5 kg ha-1 treatment than the lower one. Taken together, this indicated the nitrogen level at 229.5 kg ha-1 could be identified as the fertilizer threshold, which will be beneficial for the future fieldwork.


Introduction
Wheat is a dominant crop used for human food and livestock feed in temperate countries [1]. China is the largest wheat producer in the world, with an annual sowing area of approximately 23.4 million ha and production of 105 million tons [2]. Due to the ubiquitous utilization of synthetic nitrogen (N) fertilizer, it is easier for China to feed 22% of the world's population using only 9% of the available arable land [3]. The high input of synthetic N fertilizer has contributed to a substantial increase in wheat production in China [4]. Consequently, China has become the largest consumer of N fertilizer in the world, and over 30% of the world's total consumption is used by China [5]. Based

Materials and methods
A field experiment was conducted at the experimental station of Yangzhou University, China (32.39˚N, 119.42˚E). The site is located in the middle and lower reaches of the Yangtze River, which has a subtropical monsoon climate. The soil was a sandy-loam, and the soil properties (0-20 cm soil layer) were characterized using the methods previously described by Lu [29]. Before land preparation, composite soil samples (0-20 cm depth) were collected and analyzed using the methods described by Lu [29]. The soil contained 1.7% organic matter, 0.7 g kg -1 total N, 75.2 mg kg -1 available N, 54.8 mg kg -1 available P, and 181.2 mg kg -1 available K in the 2013/2014 growing season; additionally, the soil contained 1.3% organic matter, 0.6 g kg -1 total N, 67.2 mg kg -1 available N, 45.5 mg kg -1 available P, and 99.3 mg kg -1 available K in the 2014/ 2015 growing season. The main meteorological data from two wheat growing seasons were measured and are summarized in Table 1. The stages of wheat growing were classified and referred to previous study as follow [30]: sowing, over-wintering, jointing, stem elongation, booting, and maturing stages, which corresponded to 0, 41, 115, 129, 156 and 211 d after sowing, respectively, during the 2013/2014 growing season; and 0, 43, 120, 136, 162 and 213 d after sowing, respectively, during the 2014/2015 growing season.

Experimental design and field management
The field trial used a randomized complete block design with three replicates per treatment.

Sampling and data collection
2.2.1 Grain yield and N content. During the entire experiment, the dates of the key growth stages of crops were recorded. Wheat plants were harvested from 1m 2 subplots to determine the number of effective spikes. The grain numbers per spike were counted from 50 selected spikes. Three samples were weighed to obtain the mean thousand-grain weight for each plot. All harvested samples were threshed, and the grain yield was standardized at 13% moisture content. The concentrations of N in grain and straw were determined by micro-Kjeldahl [31], followed by digestion in a H 2 SO 4 -H 2 O 2 solution. The yield response was calculated as follows [32]: where Y N is the grain yield (kg ha -1 ) at a certain level of applied N fertilizer, and Y 0 is the grain yield (kg ha -1 ) without N application.

Nitrogen utilization parameters.
The calculations for the nitrogen utilization parameters were as follows [33,34]: where U N is the total N uptake (kg ha -1 ) in the shoot, U 0 is the total N uptake measured without N application, and F N is the rate of applied N fertilizer (kg ha -1 ). AE N ðagronomic efficiency of applied NÞ ðkg kg À 1 Þ ¼ ðY N À Y 0 Þ=F N : PFP N ðpartial factor productivity of applied NÞ ðkg kg À 1 Þ¼ Y N =F N : PE N ðphysiological efficiency of applied NÞ ðkg kg À 1 Þ ¼ ðY N À Y 0 Þ=ðU N À U 0 Þ:

CO 2 , CH 4 and N 2 O fluxes.
After sowing, the dark static chamber/GC method was used to detect the CO 2 , CH 4 and N 2 O fluxes between 9:00 am and 11:00 am every 7 days from November 9 to May 30 during the 2014-2015 season. At the same time, the soil temperature and soil moisture content were also measured (Fig 1 and S1 Table). The chamber covered a field area of 0.25 m 2 and was placed on a fixed PVC frame located on each plot. The chamber was wrapped with a layer of sponge and aluminum foil to minimize the air temperature changes inside the chamber during the sampling period. The chamber was 0.5 or 1.1 m high and was adapted based on crop growth and plant height. Each sampling was subdivided five times in 10-min intervals. A fan was used to mix the gases in the chamber, which were then drawn off using a 20-ml gas-sampling syringe. The concentrations of CO 2 , CH 4 , and N 2 O were simultaneously detected using a gas chromatograph (Agilent 7890A, Shanghai, China) in the laboratory.
The increase in the GHG concentration in the static chamber was calculated by linear regression. Fluxes were calculated based on the following formula [35].
Here, dc/dt is acquired from the linear regression equation. The value m is the molecular weight of trace gas, P indicates the atmospheric pressure (P = 1.013×10 5 Pa), R is the gas constant (R = 8.314 J/mol/K), and T is the air temperature in the chamber. V, H, and A are the volume, height, and area of the static chamber, respectively. Sample sets were rejected unless linear regression yielded an r 2 value greater than 0.90. The seasonal CH 4 , N 2 O and CO 2 emissions were sequentially linearly determined based on the emissions between every two adjacent intervals in the measurements. The air temperature inside the chamber was monitored during gas collection, and it was calibrated for the flux calculation.
The emission factor (EF-N 2 O) refers to the percentage of N that is released in the form of N 2 O to the applied N nutrients.
where E N and E 0 are the cumulative NO 2 -N emissions (kg N ha -1 ) from the fertilized and unfertilized plots, respectively, and NAR represents the N application rate (kg N ha -1 )

GWP and GHGI values.
The global warming potential (GWP) of a greenhouse gas depends on its life time. Considering a time horizon of 100 years, the N 2 O and CH 4 warming potentials are estimated to be 298 and 25 times higher than the CO 2 warming potential, respectively [36]. The net global warming potential (net GWP) excluded CO 2 [36].
GHGI is related to grain yield, as described in Mosier et al. and Shang et al. [37,38].

Statistical analysis
Data were subjected to statistical analysis (ANOVA) using the IBM SPSS 21.0 statistical package (SPSS, 2012). Emissions of CO 2 , CH 4 , and N 2 O followed a logarithmic distribution, and log transformations of these emissions were used for statistical analysis. Significant differences among means were determined by Duncan's multiple range tests at P 0.05. Pearson's bivariate correlation analysis was used to evaluate the relationships between GHG emission and both yield and nitrogen utilization parameters.

Grain yield and protein content
As seen from Table 2, in two crop years, the grain yields increased significantly due to the application of more nitrogen fertilizer; however, yields reached a plateau at 229.5 kg N ha -1 , after which the wheat yield was hardly affected by the NAR. Compared to the HN plot, the grain yield in the MN plot was almost the same (2014/2015 growing season) or even higher (2013/2014 growing season); however, the NAR could be efficiently reduced by 15%. In contrast, the wheat yields in the two growing seasons significantly decreased by 13.5% and 13.0%, respectively, when the NARs were reduced by 30% in the LN treatment, which negatively affected wheat production. Similarly, yield responses were almost the same between the MN and HN plots, and both were significantly higher than that of the LN plot. The responses of grains per spike to the NARs were positive; however, there were no significant differences in the number of effective spikes among the different nitrogen application treatments. The improvements in grain yield were mainly due to the interaction of grains per spike and thousand-grain weight. The protein content increased as more N was applied in the 2014/2015 growing season, but the protein content was even higher in the MN plot than in the HN plot during the 2013/2014 growing season.

Nitrogen utilization parameters
As shown in Table 3, in the two crop years, the AE N significantly increased by 20.3% and 16.2%, respectively, in the MN plot compared to the HN plot. Additionally, the AE N of the MN plot was higher than that in the LN plot in both years. The PFP N decreased significantly due to the increasing NAR, which reflected the law of diminishing returns. Thus, the 15% reduction in the NAR was an effective measure that improved the NUE without reducing the grain yield. AE N can be further decomposed into the AR N and PE N of applied N. The AR N improved as the NAR increased, while the PE N was negatively affected by the increase in the NAR. In the 2013/2014 growing season, the AR N in the MN plot only had a slight reduction of 1.7%, which was not significant relative to the HN plot; however, the PE N in the MN plot was significantly higher than that in the HN plot. Table 4, the CO 2 released from soil and plants was the largest source of greenhouse gas emission in all treatments. It was observed that the cumulative CO 2 emissions significantly increased with increases in the NAR during the 2014/2015 wheat growing season. There was no obvious relationship between the CH 4 emissions and NAR, and the lowest cumulative emissions were measured in the MN plot. The cumulative N 2 O emissions gradually increased with increases in the NAR, and the values varied from 0.621 to 1.32 kg N ha -1 , which were equivalent to 0.41%-0.48% of the N fertilizer that was applied. Relative to the HN plot, the seasonal N 2 O emissions significantly decreased by 19.1% in the MN plot. The MN practices emitted 16.6% more N 2 O because they received additional N through the higher application relative to the LN treatment; however, this difference was not significant. The emission factor (EF-N 2 O) relative to the applied N was measured to range from 0.144 to 0.258% in all nitrogen treatments. Compared to the HN plot, the EF-N 2 O decreased by 39.9% and 24.8% in the LN and MN plots, respectively.  Table. Higher net GWP fluxes occurred in the early of growing season, and the highest peak of net GWP fluxes were recorded at the 28 days after sowing. Peak net GWP flux increased with NAR, with rates ranging from 24.56 to 71.47 mg N m -2 h -1 . The net GWP flux under all treatments were concentrated on the sowing-before wintering   There were several small emission peaks in the next days. After stem elongation, the period of rapid wheat growth, fast uptake and utilization of soil N occurred, resulting in the slight peak of net GWP flux in the week after top-dressing under all N application conditions. At later stages, large amounts of soil N were absorbed and utilized for wheat growth along with rising temperature, leading to minor changes in net GWP under all N conditions. Although CH 4 emissions were not obviously affected by N fertilization, N 2 O emissions significantly increased as N increased. Accordingly, the net GWP significantly increased with the increase in the NAR. Relative to the net GWP (631.8 kg CO 2 eq ha −1 yr −1 ) from the HN plot, the net GWP was reduced by 61.1% and 62.1%, respectively, in the LN and MN plots. Compared to LN plot, the net GWP value was even lower than that in the MN plot.

2 Net
The lowest GHGI was observed in the MN plot, which was 12.8% lower than the value in the 0N plot and 9.7% lower than the value in the HN plot. Hence, the MN practices provided ecological solutions for wheat production and, therefore, deserve considerable attention.

Correlation analysis between GHG emission and nitrogen utilization parameters
The results of correlation analysis between the nitrogen utilization parameters and the GHG emission index are shown in Table 5 and S3 Table. The N 2 O emission was negatively correlated with PFP N (r = -0.999 Ã , p < 0.05) and PE N (r = -0.999 Ã , p < 0.05), indicating that a reduction in N 2 O emissions improved the PFP N and PE N to some extent. Similarly, both CO 2 and CH 4 emissions were negatively correlated with the PFP N and PE N , and the trends were similar to those of the N 2 O emission. All these findings resulted in a negative correlation between the net GWP and the PFP N (r = -0.940, p > 0.05) and PE N (r = -0.904, p > 0.05). Furthermore, a negative correlation was also observed between the GHGI and the AE N (r = -0.865, p > 0.05), PFP N (r = -0.814, p > 0.05) and PE N (r = -0.756, p > 0.05).

Comparison of grain yield, AE N and N 2 O emission in wheat grown under 229.5 kg N ha -1 and 270 kg N ha -1
N application cannot promise a substantial increase in crop productivity due to the principle of diminishing returns [39]. Our grain yield under the application of 229.5 kg N ha -1 was as high as that under 270 kg ha -1 N level, which was accordant with previous studies [40,41]. N application amount to winter wheat in Wuxi County has now been reduced by 20-45 kg N ha -1 or 10-20% without concomitant yield decreases, from formerly around 230 kg N ha -1 [42,43]. Additionally, synthesized from an economic and ecological point of view, 150-225 kg N ha -1 is recommended in southern China [44]. Furthermore, significant increase (20.3%) of AE N was reported under 229.5 kg N ha -1 condition relative to 270 kg N ha -1 in our study,  [45,46]. Thus, a quantum leap in the AE N is possible by simply reducing the N rate to 229.5 kg ha −1 in winter wheat; this reduction is primarily possible because the high N inputs of 270 kg N ha −1 in this region are excessive, and not all nitrogen is absorbed. The cumulative N 2 O emission gradually increased with N increasing in this paper, which was consistent with previous studies [47,48]. Relative to 270 kg N ha −1 , the seasonal N 2 O emissions was significantly decreased by 19.1% under 229.5 kg N ha -1 condition. Previously, a 10-30% reduction in N fertilizer would decrease N 2 O emissions by 11-22% in wheat [49], which were comparable with our findings. Meanwhile, N fertilizer reduction can lead to GHG emission reductions according to Kahrl's estimation [50]. Controlling the overall N rates to meet the needs of crop growth can help minimize N losses through N 2 O emissions. Hence, it is suggested that N 2 O emissions from wheat production can be reduced under the condition of 229.5 kg N ha -1 without yield reduction while AE N improved significantly. These findings demonstrated that the N loss could be greatly reduced due to the increased crop uptake.

Comparison of N 2 O emission, grain yield and AE N in wheat grown under 229.5 kg N ha -1 and 189 kg N ha -1
N 2 O is produced naturally in the soil through nitrification and denitrification and depends on soil mineral N contents [51]. The input of N fertilizers into agricultural systems is considered to be the dominant source of N 2 O emissions from agricultural soils [52,53]. Since there exits significant correlations between N 2 O emissions and the amount of N applied [47,48], N 2 O emissions gradually increased as more N input. However, in this study, no significant difference was observed on N 2 O emissions between 229.5 kg N ha -1 and 189 kg N ha -1 . Hence, the practices of 229.5 kg N ha -1 provided ecological solutions for wheat production, therefore it deserved considerable attention. Grain yield was negatively affected with significant decrease under 189 kg N ha -1 condition compared to 229.5 kg N ha -1 , which was consistent with previous studies [40,41]. Based on the results of other studies, the AE N declines when rates exceed 150 kg N ha -1 [54]. However, the highest values were achieved when 229.5 kg N ha -1 was applied. Although our highest AE N (15.4 kg kg -1 ) was greater than the mean AE N for winter wheat in China (9.4 kg kg -1 ), as reported by Chuan et al. [32], it was still lower than the world average AE N for cereal crop production, i.e., 18 kg kg -1 , as calculated by Ladha et al. [55]. It indicated that there still exists great potential for grain yield under the condition of 229.5 kg N ha -1 . Generally, significant improvement in wheat production and AE N were achieved with fewer potential threats to environment and ecology under 229.5 kg N ha -1 condition.

Comparison of AR N and PE N in wheat grown under three nitrogen Levels
Reported by previous studies, AR N and PE N were negatively affected as increasing N input [11,12], which reflected the principle of diminishing returns. It is well known that the PE N is very important to the AE N because improvements in the PE N directly result in greater plant biomass or grain yields. Normally, the stimulation effects of N fertilizer on PE N dramatically reduced due to excessive nitrogen fertilizer. An interesting finding showed that under 229.5 kg N ha -1 condition, the AR N was as high as it under 270 kg N ha -1 , but the PE N was significantly higher, indicating that the N absorbed by the plant was utilized more efficiently under 229.5 kg N ha -1 condition. Though a slight reduction in PE N was found under 229.5 kg N ha -1 condition relative to 189 kg N ha -1 condition, the AR N was significantly increased, which explained that more N could be absorbed from soil under 229.5 kg N ha -1 . All these findings illustrated that wheat plant could uptake and utilize more efficiently under 229.5 kg N ha -1 , which was the reason why the highest AE N was achieved under 229.5 kg N ha -1 .

Conclusions
Grain yields increased significantly due to higher nitrogen fertilizer input, but reach a plateau at 229.5 kg N ha -1 , in which wheat yield is hardly affected by higher nitrogen input. Meanwhile, under 229.5 kg N ha -1 condition, the highest AE N value were achieved, and N 2 O emission was as low as that of 189 kg ha -1 N. These findings demonstrated the practice of 229.5 kg N ha -1 could be identified as a fertilizer threshold which was conductive to enhancing the sustainability of crop production in our research region. Whether it is also beneficial for other region or other cultivar is still remains unexplored.
Supporting information S1