Quantifying the Impact of Reduced Nitrogen Rates on Grain Yield and Nitrogen Use Efficiency in the Wheat and Rice Rotation System of the Yangtze River Region

Abstract

Nitrogen (N) plays a vital role in plant growth; however, the yield response to N fertilizer is regularly declining in the major cropping systems due to overuse and the upper limit of yields. Heavy losses of N are also documented due to its excessive use in the rice-wheat rotation system, resulting in low nitrogen-use efficiencies (NUE) and environmental problems. Therefore, a three-year field experiment was performed with different N managements to investigate the impact of reduced N input rates in this exhaustive cropping system. Reducing the N application rates did not affect the wheat and rice yields significantly, only during the second wheat season was the yield slightly reduced as compared to higher N input treatments. Decreasing the N input rates in the prior crop and present crop, and the interactions of both rice and wheat seasons (R × W) did not influence the yields of either crop. A reduction in N fertilizer had a considerable impact on dry matter production during the wheat seasons but demonstrated no effect during the rice season. The accumulation of N was significantly reduced during both crop seasons by lowering the rate of N application. In addition, the NUE indices were significantly influenced by N fertilizer application rates. In conclusion, N fertilizer input rates for both rice and wheat crops can be lowered as compared to present fertilization rates without any risk of yield decline. Lowering the N input increases the NUE and effectively reduces N losses, and soil N status can also be maintained.

1. Introduction

Rice and wheat are widely grown throughout the world and are a staple food for millions of people. Rice is a predominant source of calories for more than half of the world’s population [1]. China leads the world in producing both rice and wheat. The yields of rice and wheat in China during 2020 were approximately 213.61 m tones (30.341 m ha) and 134.25 m tones (23.38 m ha), respectively [2]. Crop rotation is widely practiced across the world to produce cereal grains. The rice-wheat rotation system covers 9.5 to 13.5 M ha of land and is practiced in sub-tropical to warm temperate climatic zones. It is the most important food production base in the Chinese agriculture system [3,4], and is commonly adopted by farmers around the Yangtze River Basin, where both crops have been grown for years in sequence [5].

Fertilizers are very important for cereal crop production, and among them, nitrogen (N) is an important fertilizer for uplifting the grain yields of cereal crops [6]. Therefore, high quantities of manufactured N are used throughout the world to boost the production of cereals such as rice and wheat crops [7]. China consumes about one third of the total N produced worldwide. In China, about 26.87 M tons of N was consumed during the year 2019, which accounts for about 25% of the world’s use [8]. The total application of N during rice-wheat rotation is 550 kg ha⁻¹ which may increase to 650 kg ha⁻¹ in some counties of the Taihu Lake region [9]. For the rice season only, the N application rate is increased to 360 kg ha⁻¹ in some regions of Jiangsu Province [10]. However, the yields do not increase in parallel with N consumption; from 1990 to 2008, N consumption increased by approximately 50%, while only a 10% increase in crop yields was observed [11]. The overuse of fertilizers, especially N, is very common, and the use of organic fertilizers has all but stopped except in open field vegetables [12,13]. The high nutrient input results in high N residues and a higher release of reactive N into the environment. Ju et al. [14] stated that N losses of even 329 kg ha⁻¹ were observed in a single double-cropped cycle under farmer’s fertilization practice in rice and wheat rotation, and this high application of N fertilizers has resulted in well-described deadly effects on the environment, such as eutrophication [15], acidification, and greenhouse gas emissions (GHGs) [16,17].

Nitrogen use efficiency (NUE) is an index generally utilized in evaluating fertilizer management. These excessive N inputs, besides harming the environment, also significantly decrease the NUE [18]. The NUE is further divided into various components, such as agronomic efficiency (AE_N), physiological efficiency (PE_N), partial factor productivity (PFP_N), and recovery efficiency (RE_N) [19,20]. The results of pooled research from the USA, Asia, and Europe have proposed the NUE catalogs under well-managed practices as follows: AE_N (25–30 kg kg⁻¹ N), PE_N (50–60 kg kg⁻¹ N), PFP_N (60–80 kg kg⁻¹ N), and RE_N (0.5–0.8 kg kg⁻¹ N) [21]. The agronomic efficiency reported in the study region is hardly 12 kg kg⁻¹ N, while in advanced countries it is double the value observed in China [22,23,24]. N fertilizer is usually carelessly used, and the N fertilizer’s recovery efficiency achieved after the cropping period was only 33% [25,26]. Nevertheless, various issues were recorded in the present rice-wheat rotation system, which include a loss of important nutrients and low nutrient utilization efficiency resulting from the extra input of nutrients and reduced soil nutrient supply ability [27,28,29].

The production of both rice and wheat is stable; however, the concerns associated with their production are the low NUE and environmental issues [30]. Excessive quantities and inappropriate methods of fertilizer input lead to a low NUE and higher losses through leaching, volatilization, runoff, and denitrification [31,32,33]. Thus, in the present scenario, it is critical to quantify the ongoing N fertilization rates and enhance N management strategies and policies in rice and wheat growing areas of China, as it is of domestic and international benefit [34]. According to a study performed by Zhang et al. [35], about 45% of rice fields in China receive an extra amount of N fertilizer.

Fertilizer management can help lessen the N wastage without reducing crop yield [36]. Through proper farmer training, a decrease of 15–20% in N fertilizer use was observed as compared to the control group in rice production [37,38]. According to the study, N application rates can be lowered by 30–60% in rice/wheat systems through knowledge-building optimal N management approaches in southern China [22]. As stated by Huang et al. [39], unnecessary N use in rice production can be minimized over shorter and longer periods without sacrificing yield through site-specific nutrient management practices. Ju [40] reported that unused N was somewhat consumed by the later crop, and 7.5% of the total applied N was recovered from residual N. Using suitable N fertilizer quantities and making complete use of the leftover N in the soil can help to maintain higher grain yields in various crops, while limiting the hunting of nutrients present in soil by crops and eradicating huge amounts of nutrient losses and remnants in soil. Dong et al. [41] stated that the concentrations of nutrients (organic C, K, P, and N) in soils are good indicators of soil health and productivity, as they can be directly taken up by plants and contribute impressively to soil fertility.

The reduction in N application in exhaustive rice and wheat cropping systems has not been widely investigated. Although a few studies have reported no effect on rice yield with a reduction in N fertilizer, the researchers did not consider the impact of reducing fertilizer in the current crop on the upcoming crop in a rigorous farming system (rice/wheat rotation). There may prevail a linkage in terms of fertilizer decline between the present crop and the succeeding crop. Therefore, we conducted this three-year field trial using reduced N application rates:

• to compare the implications of N reduction on yield, NUE, and soil fertility in rice and wheat crops;
• to evaluate whether the present crop affects the succeeding crop concerning the crop yield, NUE, and soil fertility;
• to ascertain the optimal N fertilizer rates for attaining high NUE and high grain yields in exhaustive wheat-rice production systems.

We hope that the findings of this study will provide a sound basis to lessen the use of N in an exhaustive rice/wheat rotation, and help to contribute towards the production of more grains with fewer environmental costs.

2. Materials and Methods

2.1. Site and Soil

The experiment was conducted from the wheat season in 2015 to the rice season in 2018 using randomized complete block design at experimental sites in Danyang city (34°36′ N, 119°10′ E) owned by Nanjing Agricultural University, P.R. China. Summer rice and winter rotation were practiced earlier in the field selected for the experiment. Data regarding the daily temperature extremes and average rainfall were collected using an automatic weather station (Watch Dog 2900Et, Spectrum, Aurora, IL, USA), and are given in Figure 1. The daily mean temperature was 17.15 °C and the average annual rainfall was 1269.5 mm, with almost two-thirds of the precipitation concentrated in the paddy season. The soil of the experimental location was categorized as yellow loam (FAO soil taxonomy, 1974), having key physico-chemical characteristics as presented in Table S1.

2.2. Experimental Design and Treatments

2.2.1. Wheat

The wheat cultivar Yangmai20# was used for this study and was sown in 2015–2016 on the 5 November, while during the next two years, seeds were planted in the second week of November. The wheat seeds were broadcasted at a sowing rate of 22.5 g m⁻² in each of the experimental plots. The treatments consist of three N levels: W1 (225 kg N ha⁻¹), W2 (180 kg N ha⁻¹), and W3 (135 kg N ha⁻¹). P (90 kg ha⁻¹) and K (90 kg ha⁻¹) were added at the time of sowing in each treatment, while 40% of the N fertilizer of each treatment was added during sowing and the remaining N fertilizer was split into different growth stages (

Table 1. Nitrogen application rates for the experimental treatments at various growth stages of wheat.

Treatments	Application Date for Wheat			Growth Stage	N Applied (kg ha−1)
	2015–2016	2016–2017	2017–2018		W1	W2	W3
1st Input	5 November	10 November	10 November	Basal	90	72	54
2nd Input	6 December	12 December	December	Tillering	45	36	27
3rd Input	9 March	15 March	15 March	Spikelet-initiation	45	36	27
4th Input	20 April	24 April	27 April	Booting	45	36	27
				Total N kg ha−1	225	180	135

W1, W2, and W3 represent the 1st, 2nd, and 3rd N application rate for the wheat season.

). The crop was harvested at maturity.

2.2.2. Rice

The rice variety Wuyunjing#24 was used as a test material during all three years. Seeds were sown in the last week of May and seedlings were shifted on the 20th June. The crop was harvested in October each year when it reached maturity. The plant spacing was 13 cm × 30 cm during all three years and three seedlings were planted per hill. The plot size was 18 m² (3 m × 6 m) and plots were parted by 0.5 m-wide lanes, embedded with a plastic film of 60 cm into the soil to form a fence. Three levels of N were used as treatments: R1 (300 kg N ha⁻¹), R2 (240 kg N ha⁻¹), and R3 (180 kg N ha⁻¹). Similar to the wheat crop, 40% of the N fertilizer was applied at the time of sowing (

Table 2. Nitrogen application rates for the experimental treatments at various growth stages of rice (kg ha⁻¹).

Treatments	Application Date for Rice			Growth Stage	N Applied (kg ha−1)
	2016	2017	2018		R1	R2	R3
1st Input	19 June	19 June	19 June	Basal	120	96	72
2nd Input	29 June	30 June	30 June	Tillering	60	48	36
3rd Input	8 August	10 August	9 August	Spikelet-promoting	60	48	36
4th Input	7 September	8 September	7 September	Spikelet-protecting	60	48	36
				Total N kg ha−1	300	240	180

R1, R2, and R3 represent the 1st, 2nd, and 3rd N application rate for the rice season.

) along with identical amounts of P and K fertilizers.

In addition, the treatments of both wheat and rice were pooled together to form a total of nine treatment: R1W1, R1W2, R1W3, R2W1, R2W2, R2W3, R3W1, R3W2, and R3W3. A treatment (CK) where no N fertilizer was used was also included. All the treatments were replicated thrice. Urea (N 60%) was used as a source of N fertilizer in the study and 60% of the N fertilizer was used basally as indicated in Table 1 and Table 2, while the remaining amount was top-dressed in splits. Dry-wet irrigation method (DW) was used during the rice season to attain more water-proficient irrigation. All the experimental plots were submerged alternatingly to attain a water layer of 40–60 mm, and not re-irrigated until the water level dropped to a certain depth beneath the topsoil; during the peak flowering period, a 50 mm water layer was maintained. Irrigation cycles were then continued after peak flowering until essential drainage was managed before crop harvesting [42]. The crop management practices were consistent with local farmers’ practices.

2.3. Sampling and Measurements

Twenty plants of rice or wheat were selected at three important growth phases of panicle initiation (PI), heading (HS), and maturity (MS) for observation of the tiller number in each plot. In order to reduce the boundary effect, three illustrative plants were tagged from every plot based on average tiller number and were selected from the third row. Plant samples selected from the field were parted into stems, sheaths, green leaf blades, and panicles/spikes. Samples were then desiccated in the oven(Labor-Mix, LP-321, Esztergom, Hungary) for half an hour at a temperature of 105 °C, and then dried at 80 °C for 72 h until the constant weight was achieved. After drying, samples were weighed, sieved, and utilized for N estimation. The collected samples were digested using H₂SO₄-H₂O₂, and the famous Kjeldahl method was used to calculate the N content at each growing stage [43]. At the maturity stage, samples were harvested from all plots to calculate the effective panicle number, spikelets per panicle (SN/P), filled grain percentage (FG%), and grain weight (1000 grains). FG% was calculated for the rice-growing season only, by dividing the number of filled grains by the total spikelet number. The soil samples were collected before the application of N fertilizer to determine the impact of the reduction in N rates. The collected samples were thoroughly mixed and sub-samples were extracted. Soil samples were collected using an auger (LX-1, Serve Real, Wuxi, China) with a 2 cm internal diameter at the key growth stages of PI, HS, and MS from the 0–20 cm soil layer during both the crop seasons to determine the soil’s available N contents. Fresh soil samples were analyzed for available N content in the form of ammonium (NH₄⁺) and nitrate (NO₃⁻) after extraction with 0.01 mol L⁻¹ calcium chloride using a continuous flow analyzer, or stored in a freezer at −20 °C until analysis by a similar method within 25 days.

2.4. Calculation of N Use Efficiencies

NUE is divided into the following parameters depending upon the objectives of the study: agronomic efficiency of N (AE_N), recovery efficiency of N (RE_N), partial factor productivity of N (PFP_N) and physiological efficiency of N (PE_N). These parameters were calculated using the formulas described in [5].

$${\mathrm{AE}}_{\mathrm{N}}={\mathrm{Y}}_{\mathrm{N}}-{\mathrm{Y}}_0/{\mathrm{F}}_{\mathrm{N}}\left(\mathrm{kg}{\mathrm{kg}}^{-1}\right)$$

(1)

$${\mathrm{RE}}_{\mathrm{N}}={\mathrm{U}}_{\mathrm{N}}-{\mathrm{U}}_0/{\mathrm{F}}_{\mathrm{N}}\times 100(\%)$$

(2)

$${\mathrm{PFP}}_{\mathrm{N}}={\mathrm{Y}}_{\mathrm{N}}/{\mathrm{F}}_{\mathrm{N}}\left(\mathrm{kg}{\mathrm{kg}}^{-1}\right)$$

(3)

$${\mathrm{PE}}_{\mathrm{N}}={\mathrm{Y}}_{\mathrm{N}}-{\mathrm{Y}}_0/{\mathrm{U}}_{\mathrm{N}}-{\mathrm{U}}_0\left(\mathrm{kg}{\mathrm{kg}}^{-1}\right)$$

(4)

where Y_N denotes the yield (kg ha⁻¹) achieved using a specific rate of N, Y₀ represents the yield (kg ha⁻¹) obtained from control treatment, U_N is the sum of N (kg ha⁻¹) taken up by the straw and grain, and U₀ represents the quantity of N absorbed in the treatment where no N was used. F_N typifies the quantity of N used (kg ha⁻¹).

2.5. Statistics

The data collected were analyzed by using the statistical software SPSS 21.0 (SPSS Inc. Chicago, IL, USA). A two-way analysis of variance (ANOVA) for the grain yields, N uptake, dry matter accumulation, NUE indices, and soil N contents from the three replicates was conducted at the harvest stage, considering these effects: rice N application rate, wheat N application rate, and rice N application rate × wheat N application rate. The differences between treatment means were compared for all the parameters using Duncan’s multiple range test at a p-value of <0.05. Sigma Plot (14.0) and R software were used to construct the figures.

3. Results

3.1. Impact of N Reduction on Grain Yield and Yield Related Attributes of Wheat and Rice

3.1.1. Wheat

The wheat grain yields and associated attributes from three consecutive seasons were analyzed and are presented in Figure 2. The spikes per m⁻² were significantly influenced by N rates during the first and third year of the experiment. On the other hand, the reduction in N rates did not significantly affected the SN/P and grain weight during the three years. The number of panicles per m⁻² remained higher in the 2016–2017 season under the same N rates, but the SN/P and grain weight were higher in the 2017–2018 wheat season. The grain yield ranged between 2.47–4.07 t ha⁻¹, 3.24–5.13 t ha⁻¹, and 5.87–6.73 t ha⁻¹ during the 2015–2016, 2016–2017, and 2017–2018 growing seasons, respectively. Interestingly, no obvious differences were noticed in the grain yields of wheat by lowering the N application rate to as low as 135 kg ha⁻¹ from the common farmer’s practice of 225 kg ha⁻¹ in any of the three years of the experiment, except for the treatments R2W3 and R3W3 where a decrease in the grain yield was noticed during the 2016–2017 season by reducing the N supply. Reducing the N fertilizer during previous rice seasons had no effect on wheat grain yield, while N reduction during the current wheat season significantly influenced the grain yield only in the 2016–2017 season. The interaction of R × W exhibited no significant differences for yield during all three wheat seasons.

3.1.2. Rice

The combined results depict that a reduction of 20–40% in the N application rate from the conventional farmers’ N fertilization level of 300 kg/ha caused no marked variations in rice grain yields among the treatments in all three crop seasons of the experiment (Figure 3), which indicates that decreasing the N fertilizer rate did not significantly reduce the rice grain yields. The grain yields during the rice seasons ranged between 12.49–13.94 t ha⁻¹ (2016), 9.74–11.08 t ha⁻¹ (2017), and 12.01–10.77 t ha⁻¹ (2018). The total panicle number decreased from 2016 to 2018, while the highest grain weight was observed during the third year of study. The FG% also reached a maximum during the 2018 rice season. Reducing the N fertilizer inputs in the previous wheat season and the present rice season had no significant effects on the yield or related attributes during the three consecutive years. The R × W interaction for the grain yield and yield traits between the different treatments was also non-significant, except for the SN/P in the 2018 rice season.

3.2. Dry Matter Accumulation

The dry matter (DM) content accumulation at maturity in the wheat season ranged between 6.06–11.73 t ha⁻¹, 7.86–14.26 t ha⁻¹, and 7.68–10.88 t ha⁻¹ during 2015–2016, 2016–2017, and 2017–2018, respectively (

Table 3. Dry matter accumulation and N accumulation at maturity under different N rates in three wheat seasons (2015–2018).

Treatments	Dry Matter (t ha−1)			N Accumulation (kg ha−1)
	2015–2016	2016–2017	2017–2018	2015–2016	2016–2017	2017–2018
R1W1	11.62 ± 0.64	14.30 ± 0.78	9.93 ± 0.21	135.05 ± 7.92	138.28 ± 3.73	124.69 ± 0.98
R2W1	11.73 ± 0.85	12.36 ± 1.08	8.86 ± 0.35	140.65 ± 8.29	130.28 ± 7.52	103.71 ± 3.35
R3W1	9.94 ± 1.03	12.43 ± 2.06	9.61 ± 0.67	98.49 ± 9.66	119.76 ± 3.96	114.32 ± 7.54
R1W2	10.3 ± 0.46	11.45 ± 0.75	9.32 ± 0.69	119.65 ± 6.53	119.54 ± 5.52	110.88 ± 2.58
R2W2	10.16 ± 1.07	8.95 ± 1.09	8.56 ± 0.52	105.93 ± 2.97	79.35 ± 6.03	102.02 ± 2.42
R3W2	8.34 ± 1.10	9.75 ± 1.74	8.03 ± 0.20	94.49 ± 6.34	97.57 ± 2.98	95.18 ± 3.03
R1W3	8.14 ± 0.68	9.43 ± 0.93	8.5 ± 0.12	83.88 ± 5.15	98 ± 2.47	99.14 ± 5.52
R2W3	7.9 ± 0.58	7.86 ± 0.73	8.44 ± 0.40	87.28 ± 4.06	67.63 ± 4.59	93.77 ± 2.75
R3W3	6.06 ± 0.62	11.02 ± 1.25	9.7 ± 0.23	75.82 ± 3.07	95.03 ± 4.62	97.59 ± 2.09
R	*	ns	ns	***	***	***
W	***	**	ns	***	***	***
R × W	ns	ns	ns	ns	**	ns

R1 = 300 kg N ha⁻¹, R2 = 240 kg N ha⁻¹, R3 = 180 kg N ha⁻¹ represent the rice fertilizer rates and W1 = 225 kg N ha⁻¹, W2 = 180 kg N ha⁻¹, W3 = 135 kg N ha⁻¹ represent the wheat fertilizer rates. ±Standard error of mean (n = 3), * significant at p < 0.05; ** significant at p < 0.01; *** significant at p < 0.001; ns, means not significant.

). The treatment R2W1 demonstrated the maximum DM accumulation during the first study year, and RIWI during the second and third growing seasons of wheat. The N fertilizer reduction in wheat significantly influenced the DM content during the first two crop years. The R interactions significantly affected DM only during the 2015–2016 wheat season. The R × W interactions for all the treatments during the three consecutive wheat seasons showed no significant impact.

The DM content at rice crop harvest was accessed under different N application rates, and trends similar to yield were observed; no remarkable variations were noted between different N rates. The dry matter contents in three rice seasons from 2016–2018 ranged between 21.47–25.77 t ha⁻¹, 20.43–22.16 t ha⁻¹, and 19.9–24.18 t ha⁻¹, respectively (

Table 4. Dry matter accumulation and N accumulation at maturity under different N rates in three rice seasons (2016–2018).

Treatments	Dry Matter (t ha−1)			N Accumulation (kg ha−1)
	2016	2017	2018	2016	2017	2018
R1W1	25.77 ± 0.45	21.7 ± 0.74	21.87 ± 1.18	351.27 ± 9.3	260.92 ± 9.66	245.56 ± 7.75
R1W2	21.7 ± 1.22	20.83 ± 1.31	22.13 ± 2.53	308.24 ± 17.9	248.57 ± 17.61	267.96 ± 34.14
R1W3	23.47 ± 0.72	20.97 ± 0.76	19.9 ± 1.15	321.59 ± 9.64	247.83 ± 12.91	228.89 ± 11.17
R2W1	25.83 ± 1.92	22.03 ± 1.41	24.13 ± 2.82	286.99 ± 27.92	231.54 ± 19.18	258.4 ± 39.74
R2W2	23.07 ± 1.09	20.50 ± 1.22	20.63 ± 1.32	292.19 ± 6.26	234.22 ± 10.2	220.64 ± 17.87
R2W3	22.83 ± 1.38	22.16 ± 1.4	21.93 ± 3.4	278.21 ± 23.05	257.96 ± 22.61	235.37 ± 32.03
R3W1	21.47 ± 1.6	20.43 ± 1.14	20.87 ± 1.44	259.31 ± 19.57	215.72 ± 14.74	219.64 ± 18.04
R3W2	25.43 ± 1.13	21.30 ± 0.6	24.1 ± 1.51	275.66 ± 7.19	228.36 ± 3.17	235.12 ± 11.57
R3W3	23.39 ± 1.25	21.60 ± 0.86	24.2 ± 2.24	263.05 ± 9.98	243.32 ± 13.66	234.1 ± 24.79
R	ns	ns	ns	***	ns	ns
W	ns	ns	ns	ns	ns	ns
R × W	*	ns	ns	ns	ns	ns

R1 = 300 kg N ha⁻¹, R2 = 240 kg N ha⁻¹, R3 = 180 kg N ha⁻¹ represent the rice fertilizer rates and W1 = 225 kg N ha⁻¹, W2 = 180 kg N ha⁻¹, W3 = 135 kg N ha⁻¹ represent the wheat fertilizer rates. ±Standard error of mean (n = 3), * significant at p < 0.05; *** significant at p < 0.001; ns, means not significant.

). The R × W interactions produced a significant difference only during the 2016 rice season, while the N reduction during the current crop (R) and earlier crop (W) remained non-significant throughout the experiment (Table 3).

3.3. Nitrogen Accumulation

The results from three years of experimentation illustrate that, along with the decline in the N fertilizer application rates, the N accumulation also decreases. In the wheat crop, the reduction in N application significantly reduced the quantity of N accumulation during the whole experiment. During all three years, the highest amount of N accumulated in the treatments where N application was higher, and the lowest was detected in treatments where the least amount of N was used. In comparison with R3W3, the N accumulation in treatment R1W1 enhanced from 27% to 80% in three wheat-growing seasons (Table 3). The reduction in the N rate during the current wheat season (W) and previous rice season (R) significantly reduced the N accumulation during all three wheat seasons. However, R × W interactions were non-significant during the 2015–2016 and 2017–2018 wheat season (Table 3).

A significant decrease in N accumulation was observed in 2016 with reduced N application, while in the 2017 and 2018 rice seasons the differences were not statistically different from each other in various treatments (Table 4). In the 2016, 2017, and 2018 rice seasons, increases of 33%, 6.99%, and 4.70%, respectively, were observed in N accumulation in treatment R1W1 as compared with R3W3. Reducing the N application rate in the current rice crop (R) only produced significant effects in the 2016 growing season, while during 2017 and 2018, no evident differences were observed. However, the N decline in the wheat seasons (W) and the R × W interactions remained non-significant during all three rice seasons (Table 4).

3.4. Nitrogen Use Efficiency

The NUE indicators also demonstrate trends similar to other parameters, which show a negative correlation with N fertilizer rates. The various components of NUE are explained further below.

3.4.1. Nitrogen Agronomic Efficiency (AE_N)

The AE_N during the wheat seasons remained between 10.88–19.56 kg kg⁻¹, 16.50–24.67 kg kg⁻¹, and 22.26–36.38 kg kg⁻¹ during the years 2015–2016, 2016–2017, and 2017–2018, respectively (

Table 5. Indices of NUE under different N input rates during wheat seasons in the three-year experiment (2016–2018).

Treatments	2015–2016				2016–2017				2017–2018
	AEN (kg kg−1)	PFPN (kg kg−1)	PEN (kg kg−1)	REN (%)	AEN (kg kg−1)	PFPN (kg kg−1)	PEN (kg kg−1)	REN (%)	AEN (kg kg−1)	PFPN (kg kg−1)	PEN (kg kg−1)	REN (%)
R1W1	13.30 ± 0.19	18.13 ± 0.14	27.29 ± 1.12	58.46 ± 2.01	16.64 ± 1.51	21.07 ± 1.91	33.62 ± 3.13	49.99 ± 0.66	22.69 ± 2.48	29.14 ± 2.39	52.18 ± 5.47	44.49 ± 3.41
R2W1	13.19 ± 3.40	17.69 ± 3.08	25.59 ± 5.33	51.19 ± 3.22	19.19 ± 1.46	23.62 ± 1.32	42.11 ± 10.65	49.40 ± 7.47	23.56 ± 1.51	30.01 ± 1.31	74.16 ± 3.11	32.05 ± 3.42
R3W1	12.16 ± 3.34	16.61 ± 3.38	36.92 ± 6.12	32.35 ± 1.78	18.44 ± 2.51	22.88 ± 2.37	45.72 ± 3.48	41.77 ± 7.91	22.26 ± 0.88	28.71 ± 0.99	62.03 ± 6.21	34.92 ± 4.82
R1W2	15.59 ± 1.69	21.40 ± 1.85	31.23 ± 5.22	52.07 ± 3.43	19.04 ± 1.92	24.61 ± 2.30	40.62 ± 5.31	48.96 ± 9.24	29.37 ± 1.52	37.42 ± 1.25	66.65 ± 2.61	44.05 ± 1.31
R2W2	15.65 ± 3.72	21.48 ± 3.65	35.47 ± 8.42	44.4. ± 2.94	20.86 ± 1.03	26.40 ± 1.43	71.56 ± 7.94	29.75 ± 2.03	27.79 ± 1.84	35.79 ± 1.72	78.67 ± 3.47	38.58 ± 8.98
R3W2	14.25 ± 0.60	19.86 ± 0.18	38.04 ± 4.19	38.15 ± 3.17	20.94 ± 3.54	26.48 ± 1.64	52.50 ± 3.57	39.88 ± 3.59	25.12 ± 1.15	33.17 ± 1.35	78.47 ± 9.9	35.15 ± 7.69
R1W3	19.56 ± 2.84	26.91 ± 1.05	44.62 ± 4.15	43.09 ± 2.24	24.67 ± 2.51	32.06 ± 3.08	46.03 ± 2.49	53.49 ± 3.75	35.42 ± 0.91	46.16 ± 1.34	73.06 ± 8.57	50.04 ± 6.90
R2W3	17.54 ± 1.50	24.96 ± 1.92	38.77 ± 2.19	45.14 ± 2.44	16.5 ± 4.26	23.89 ± 4.26	52.25 ± 12.7	30.99 ± 1.73	36.38 ± 1.59	47.13 ± 1.16	78.97 ± 0.91	46.06 ± 1.83
R3W3	10.88 ± 1.63	18.27 ± 2.13	29.79 ± 4.61	37.22 ± 3.12	20.23 ± 4.31	27.61 ± 4.77	37.50 ± 7.51	56.71 ± 3.92	32.77 ± 0.94	43.51 ± 0.55	68.16 ± 5.72	48.89 ± 4.89
R	ns	ns	ns	***	ns	ns	ns	*	ns	ns	*	ns
W	ns	*	ns	ns	ns	ns	ns	ns	***	***	*	*
R × W	ns	ns	ns	ns	ns	ns	ns	ns	ns	ns	ns	ns

AE_N, PFP_N, PE_N, and RE_N represent the agronomy efficiency, the physiological efficiency of applied N, the partial factor productivity from applied N, the recovery efficiency of applied N, and the harvest index of applied N, respectively. R1 = 300 kg N ha⁻¹, R2 = 240 kg N ha⁻¹, R3 = 180 kg N ha⁻¹ represent the rice fertilizer rates and W1 = 225 kg N ha⁻¹, W2 = 180 kg N ha⁻¹, W3 = 135 kg N ha⁻¹ represent the wheat fertilizer rates. ±Standard error of mean (n = 3), * significant at p < 0.05; *** significant at p < 0.001. ns, means not significant.

) The R and R × W interactions remained non-significant during all the wheat seasons (W). In the rice crop, the lowest N application rate, R3, showed the highest AEN in all three growing seasons, ranging between 15.16–26.34 kg kg⁻¹, 7.33–15.93 kg kg⁻¹, and 9.8–20.5 kg kg⁻¹ in years 2016, 2017, and 2018, respectively (

Table 6. Indices of NUE under different N input rates during rice seasons in the three-year experiment (2016–2018).

Treatments	2016				2017				2018
	AEN (kg kg−1)	PFPN (kg kg−1)	PEN (kg kg−1)	REN (%)	AEN (kg kg−1)	PFPN (kg kg−1)	PEN (kg kg−1)	REN (%)	AEN (kg kg−1)	PFPN (kg kg−1)	PEN (kg kg−1)	REN (%)
R1W1	15.16 ± 1.52	43.85 ± 2.20	21.20 ± 2.34	71.72 ± 1.35	7.82 ± 1.61	34.36 ± 1.73	19.47 ± 2.98	40.04 ± 4.97	9.87 ± 0.88	36.9 ± 1.40	24.26 ± 1.13	40.60 ± 2.36
R1W2	17.80 ± 2.15	46.48 ± 3.93	31.03 ± 0.7	57.37 ± 7.26	7.33 ± 1.34	33.92 ± 0.79	25.76 ± 11.1	35.92 ± 7.64	11.40 ± 1.51	38.40 ± 2.0	32.17 ± 13.4	48.06 ± 15.5
R1W3	17.63 ± 0.88	46.32 ± 1.55	28.88 ± 2.66	61.82 ± 4.72	8.29 ± 1.49	34.87 ± 0.19	24.81 ± 6.30	35.67 ± 5.66	11.96 ± 1.5	38.97 ± 2.08	33.91 ± 2.34	35.04 ± 1.94
R2W1	19.65 ± 5.14	55.51 ± 8.07	31.13 ± 8.56	62.86 ± 9.67	8.74 ± 1.14	41.97 ± 2.06	24.53 ± 5.76	37.80 ± 5.76	16.27 ± 2	50.03 ± 3.18	30.55 ± 5.29	56.18 ± 7.73
R2W2	19.35 ± 3.04	55.21 ± 5.95	29.57 ± 3.76	65.03 ± 2.38	8.41 ± 1.83	41.65 ± 2.13	21.60 ± 4.47	38.92 ± 2.22	11.57 ± 2.32	45.33 ± 0.62	32.04 ± 6.86	40.36 ± 13.02
R2W3	18.33 ± 1.01	54.20 ± 3.88	33.41 ± 6.5	59.20 ± 11.67	7.34 ± 0.42	40.57 ± 2.06	16.77 ± 3.52	48.81 ± 9.46	12.83 ± 3.68	46.60 ± 1.94	30.92 ± 9.58	46.50 ± 17.53
R3W1	21.59 ± 2.72	69.40 ± 4.49	33.13 ± 6.47	68.44 ± 9.42	11.08 ± 1.36	55.39 ± 1.44	29.16 ± 8.22	41.62 ± 6.45	17.97 ± 3.4	62.99 ± 6.9	33.70 ± 6.69	53.81 ± 2.39
R3W2	26.34 ± 3.57	75.48 ± 8.69	35.78 ± 6.79	77.52 ± 2.40	15.93 ± 1.92	61.58 ± 0.25	36.15 ± 6.75	48.64 ± 2.71	14.80 ± 4.38	59.83 ± 1.43	23.38 ± 5.28	61.87 ± 13.26
R3W3	26 ± 2.91	73.82 ± 6.97	38.26 ± 6.82	70.51 ± 8.78	13.52 ± 1.03	57.84 ± 3.50	24.65 ± 3.61	56.95 ± 8.86	20.50 ± 3.82	65.50 ± 2.26	34.62 ± 12.2	74.62 ± 17.02
R	*	***	ns	ns	***	***	ns	ns	*	***	ns	ns
W	ns	ns	ns	ns	ns	ns	ns	ns	ns	ns	ns	ns
R × W	ns	ns	ns	ns	ns	ns	ns	ns	ns	ns	ns	ns

AE_N, PFP_N, PE_N, and RE_N represent the agronomy efficiency, the physiological efficiency of applied N, the partial factor productivity from applied N, the recovery efficiency of applied N, and the harvest index of applied N, respectively. R1 = 300 kg N ha⁻¹, R2 = 240 kg N ha⁻¹, R3 = 180 kg N ha⁻¹ represent the rice fertilizer rates and W1 = 225 kg N ha⁻¹, W2 = 180 kg N ha⁻¹, W3 = 135 kg N ha⁻¹ represent the wheat fertilizer rates. ±Standard error of mean (n = 3), * significant at p < 0.05; *** significant at p < 0.001; ns, means not significant.

). The AE_N increased in 2017 as compared to the previous year’s rice season, but in 2018, it again showed a decreasing trend relative to the earlier two seasons. Reducing the N input significantly improved the AE_N for both crops, as higher N rates possessed lower values for AE_N than reduced N application rates in both crops. The W and R × W interactions were non-significant for the rice crop during the entire experiment, while the reduction during the current rice season (R) significantly affected AE_N in all years.

3.4.2. Nitrogen Partial Factor Productivity (PFP_N)

In the present experiment, decreasing the N application rates increased the PFP_N gradually in the wheat crop. The PFP_N during the wheat seasons 2015–2016, 2016–2017, and 2017–2018 increased by 18.47–48.4%, 12.10–35.73%, and 24.69–57.04%, respectively, in N treatments W2 and W3 as compared to W1 (Table 5). Lowering the N in the rice season (R) and R × W interactions did not significantly influence the PFP_N in the wheat crop. However, the decline in N use during the wheat seasons (W) had a significant effect during the years 2015–2016 and 2017–2018.

The peak value for PFP_N in rice was observed for the minimum N application rate R3. In the rice seasons, the PFP_N of R2 and R3 increased by 19.42–62.36% (2016), 20.36–76.58% (2017), and 28.38–68.01% (2018) as compared to R1 (Table 6). The N reduction during the wheat (W) season and R × W interactions for all three years remained non-significant, while N reduction during the current seasons (R) significantly affected the PFP_N during all the three years. The PFP_N remained lower in the wheat-growing seasons as compared to rice, which suggested that rice had notably higher grain yields in our whole experiment as compared to wheat.

3.4.3. Nitrogen Physiological Efficiency (PE_N)

In the three wheat seasons, the PE_N ranged between 25.59–44.62 kg kg⁻¹, 33.75–71.56 kg kg⁻¹, and 52.18–78.97 kg kg⁻¹, respectively (Table 5). The PE_N was not significantly affected during wheat seasons 2015–2016 and 2016–2017 by the N application rates. However, during the 2017–2018 growing season, a significant difference was observed among different N input treatments, where the uppermost PE_N was obtained in the R2W3 treatment and the lowest was observed for R1W1. The R × W interactions remained non-significant during all three years of study (Table 5).

The PE_N during the three rice seasons 2016, 2017, and 2018 ranged between 21.20–38.26 kg kg⁻¹, 16.77–36.15 kg kg⁻¹, and 23.38–34.62 kg kg⁻¹, respectively (Table 6). The R3 treatment had the maximum PE_N value during all three rice growing seasons. The reduction in the N application rate during previous wheat (W) seasons, current rice (R) seasons, and R × W interactions remained non-significant. The effects of N application rates on PE_N were consistent, as lower N rates exhibited higher values for PE_N for both crops during the whole experiment.

3.4.4. Nitrogen Recovery Efficiency (RE_N)

During the first wheat season, the maximum RE_N was observed in the RIW1 treatment, while during the next two years, R3W3 and R1W3 exhibited the highest values of RE_N, respectively (Table 5). The reduction in N in the current wheat (W) season had a significant impact only during 2017–2018, while the R × W interactions remained non-significant throughout. In the rice season, the treatment R3 (where the least amount of N fertilizer was applied) had the highest RE_N values during all three years. Increases of 7.81%, 29.76%, and 35%% were observed in treatment R3 as compared to treatment R1 from the 2016–2018 rice seasons, respectively (Table 6). Lowering the N fertilizer rate during the present crops and R × W interactions had no dominant effect on the RE_N of either the rice or wheat crops.

3.5. Soil N Status

Ammonium (NH₄⁺) and nitrate (NO₃⁻) are the chief N forms present in the soil, which represent the soil capacity to supply N to plants. The NH₄⁺ and NO₃⁻ contents present in the soil at various growth stages (PI, HS, and MS) of wheat and rice crops during the three-year trial are presented in Tables S2–S5. It is evident from the data that higher NH₄⁺-N and NO₃⁻-N contents were present in the soil during the early growth stages of both the crops, and rapid reduction was noted in the soil N contents from PI to MS. Similarly, more NH₄⁺-N and NO₃⁻-N contents were present during the wheat crop season as compared to rice. No significant differences were noted among various treatments regarding the soils’ available N contents. Lower soil N contents were observed during rice seasons as compared to wheat, this is attributable to flooding conditions and the anaerobic soil environment. The reduction in the N rate during the current season and previous season mostly remained non-significant, R × W interactions for NH₄⁺ were significant only at PI in the first year of the trial, while for NO₃⁻ contents, it was significant at HS in both the second and third year of the experiment. For the rice season, the interaction of R × W produced a significant difference for both NH₄⁺ and NO₃⁻ only at MS in the first year of the experiment.

4. Discussion

Nitrogen plays a vital role in the growth and development of plants. In our study, the three consecutive rice-wheat rotations depicted no substantial decline in the grain yield of either wheat or rice with a reduction in N fertilizer application rates of 20–40%. The average annual yield across three years of rice-wheat rotation in our experiment was 17.45 t ha⁻¹, which was higher than the average annual yield (13.17 t ha⁻¹) documented in this region [44]. We used the controlled soil drying irrigation method (DW) during the rice crop and the improvement in the paddy yield may attributed to this [45]. The boost in crop yields during the last year could be due to greater sunshine hours in the rice growing season and a prolonged dry season in winter as these conditions favor crop development and yield formation [46]. The increase in the N utilization ratio by reducing the N input at the early growth stage could be another reason behind higher crop production [18].

The response of grain yield towards N in an extensive rice-wheat rotation system is decreasing because of the overuse of nitrogenous fertilizers in China [47]. The plant demand is below the level of fertilizer application in a growing region, which is why no notable differences were observed among the treatments in our experiment. The split input of N fertilizer played a vital part in increasing the grain yield in this study and various studies have reported an increase in grain yields by postponing N fertilizer use to different crop stages [48]. The fertilizer input increases grain yields but, after a certain limit, it fails to respond and the same was observed in our study. Delaying N application increases the grain sink size and total dry matter which helps indirectly in improving crop yields [6]. The outcomes from our trial are also near the recommended levels, as no noteworthy differences were seen between the highest N treatments and others in either crop [5,14]. Over-fertilization also causes lodging and delayed maturity in treatments with a higher input of N, which limits plants’ ability to produce. The splitting of N fertilizer helped in maintaining the yields of rice and wheat as this approach prolonged the availability of N to plants for longer periods. Dry matter (DM) accumulation is one of the main growth factors of wheat and rice. The reason for the lack of difference between reduced and high fertilizer rates might be that under reduced N rates, proper growth of crops is maintained by enhancing the DM of the leaves, stem, and ears at different growth stages of rice and wheat, thus promoting total dry matter accumulation [49]. In our study, a reduction in fertilizer application at later growth stages of wheat and rice had no notable consequences on DM accumulation. The DW method showed a positive impact on the DM accumulation of rice by enhancing the uptake of nutrients to the stem, leaves, and ears, improving its dry weight and in turn helping in dry matter accumulation which was in line with Kundu et al. [50].

The way the plant responds to applied fertilizer is crucial to determine the proper levels of fertilizer for crops. The quantity of N uptake varies with N application rates as expected [51]. Along with higher N application rates, a delay in N input may also be the reason for increased N accumulation in our study [6]. In early growth stages, plant roots are less developed and unable to uptake more N, which results in higher N losses at early stages [52], thus the reduction in N application at early plant stages is feasible. The N accumulation was greater at later stages of the crops. The summer rice crop exhibited more N accumulation than the winter wheat in our study, despite abundant fertilizer supplies, which might be due to the poor N assimilation efficiency of the wheat variety selected for this study. NH₄⁺ and NO₃⁻ are the inorganic forms in which N is mostly absorbed from the soil. Lower N contents during the rice season as compared to wheat can be attributed to more rainfall which can result in persistent leaching, volatilization, and denitrification [53]. No significant differences in N contents were found in this study by the reduction of N fertilizer among treatments, but great variation was found among the N contents of soil at various growth periods (Tables S2–S5). Higher soil available N during the wheat season is credited to the least availability of surface water and less precipitation. During the growth of both crops, the early growth stages exhibit higher soil N contents as compared to maturity, which is compatible with plant growth. N reduction having no remarkable impact on the soil available N indicates that soil fertility is not affected by reducing N fertilizer, rather it minimizes the N loss [5]. The C:N ratio analysis of the experimental soil was not performed, which is a limitation of our study, as the C:N ratio (if determined) gives a clear picture of soil N relations.

The NUE is the index used to describe the rise in grain yield per unit of N available. The NUE indices largely depend upon the quantity of N used, N accumulation, and grain yield. Better synchrony among crop N requirements and accessible N resources is a key factor to increase NUE. AE_N is the term used to describe a surge in grain yield due to fertilizer input. The findings of this study disclosed the positive impact of N reduction on AE_N, as higher figures were obtained with the reduced N rates for both the crops (Table 5 and Table 6. Dobermann [21] recommended an AE_N value of 25–30 kg kg⁻¹ under good management practices, while in our study, even with a 20–40% reduction in N fertilizer, the AE_N is below the proposed value which predicts that higher N fertilizer levels are in practice in the growing region. Decent values of AE_N can be attained by lowering the input rate of N in both wheat and rice crops. The constant reduction in N fertilizer for longer periods at the cost of yield decrease is not feasible. The major reasons for lower NUE values were the high grain yields from control treatments and higher N fertilizer inputs.

PE_N is another key index that describes the effectiveness with which the applied N fertilizer is converted to economical yield. Compared to the 50–60 kg kg⁻¹ recommended PE_N, the values in our experiment during wheat seasons were optimum but remained lower for rice because of higher yields from the plots with zero N fertilizer input during the rice season [5]. The higher values of AE_N and PE_N in reduced N application treatments were likely artifacts of lower N accumulation and somewhat lower yields as compared to the highest N application rate. PFP_N is vital for agronomists as it provides a relation between N input from all available sources and economic gains. The N input gradually raises the yield and the capability of consuming the native N supplies from the soil-flood water system. In the present trial, the highest PFP_N values were obtained from the lowest N input rates during the three-year experiment. This suggests that reducing the N input and delaying its application will help to improve the NUE [48,54]. The higher PFP_N values observed in our experiment might be due to the higher grain yields observed in our study. The RE_N during the wheat seasons in our experiment remained lower, which might be due to the higher N accumulation reported in previous experiments at similar N input rates, and the fact that yield in the control treatments remained very low. The lower grain yields in control plots in this trial were ambiguous. A greater uptake of N in rice at the MS resulted in a higher RE_N during all rice seasons compared to that reported by others. The lower values of RE_N in wheat indicate a poor N uptake from applied N. Thus, the outcomes of this research suggest that high N input rates should not be encouraged in a rigorous rice-wheat rotation system.

5. Conclusions

It can be concluded that the application of 20–40% less N as compared to the recommended rate can significantly increase the N use efficiency without affecting the yield in both rice and wheat crops. Our study revealed that the application of 180 kg N per hectare resulted in a good yield and NUE in the rice crop, whereas, in the wheat season, a reduction in the grain yield was observed during the second crop season. The residual N in the soil aided in sustaining the yield for up to one year in wheat; thus, it would be a good strategy to reduce the N application rate and subsequently apply the recommended N rate in the next season to achieve sustainable rice and wheat production and reduce environmental pollution by increasing N use efficiency. Regarding the soil fertility, a reduced N application rate had no effect. Overall, it is suggested that N fertilization should be managed in a sustainable way to reduce the environmental impact and to increase the crop production without influencing the soil health and biodiversity. Our study and other related studies represent a good source of information for farmers and the institutes responsible for the management of chemical N fertilizers. Future studies should also include observations of the C:N status and soil temperature, as these are also potential soil effectors.