Single-Time Mechanical Deep Placement Fertilization Using Bulk Blending Fertilizer on Machine-Transplanted Rice: Balanced Yield, Nitrogen Utilization Efficiency, and Economic Benefits

Abstract

Despite growing interest in controlled-release N fertilizers (CRNFs) because of their potential for enhancing nitrogen use efficiency (NUE) and economic returns, their comprehensive impact on machine-transplanted rice remains to be understudied. To address this gap, here, we present a two-year field experiment that assessed the impact of CRNF using mechanical deep placement fertilization (DPF) on rice cultivation. The study involved three CRNF types (bulk blending fertilizer (BBF), polymer-coated urea (PCU), and sulfur-coated urea (SCU)) and two fertilization methods (DPF and broadcast application), with a high-yield split fertilization of urea as a control (CK). The results showed that DPF, especially with SCU, greatly enhanced soil NH₄⁺-N concentrations, NUE, rice yield, and economic benefits compared to broadcast application. BBF consistently exhibited superior NUE and notable economic benefits, regardless of the application method used. Conversely, single-time application of PCU was less favorable for rice growth. In conclusion, for optimal economic benefits and NUE, DPF combined with single-time application of SCU is recommended. However, if deep application is not feasible and only broadcasting is possible in rice cultivation, BBF emerges as the ideal choice for both high NUE and significant economic returns. This research offers insights for improved nitrogen management in machine-transplanted rice, effectively optimizing yield, NUE, and profitability.

1. Introduction

Growing populations and limited cultivated land resources require more nitrogen (N) fertilizer input, with an annual growth rate of ~2% to meet food demands [1]. As one of the staple food in the world, rice is planted in ~11% of the global arable land [2], and ~18% of the global rice harvest area is in China [3]. Rice production requires the highest N fertilizer rate among staple crop systems [4], while inefficient N management is often associated with low N use efficiency (NUE) and negative environmental impacts [5,6]. Thus, optimizing N fertilizer management is imperative to achieve higher economic benefits in rice production. Theoretically, balancing N release rate with the plant N requirements can effectively improve NUE [7], and split N fertilizer application with from three to four top-dressings during the rice growing season has been adopted as an agronomic practice to improve NUE and rice yields [8,9]. However, applying split applications of N fertilizer is often restricted by relevant expertise and high labor/time costs. Therefore, many efforts have been devoted to developing a simpler and more convenient N fertilization management practice for rice production.

Single-time application of controlled-release N fertilizer (CRNF) is considered to be a promising way to improve NUE and to reduce labor/time costs [10]. CRNF can gradually release N to meet plant growth requirements [11], and can increase rice yield and NUE by 5% and 20% [12]. Thus, CRNF is a better substitute for conventional split application of urea [13]. However, some previous studies have also reported negative effects of CRNF on rice yield and NUE, which may be due to an inappropriate N release pattern [14,15]. The CRNF type and the application method are often considered to be the main factors that control the effects on rice production under single-time application of CRNF. For CRNF types, the coating material decides the N release pattern, i.e., most of the N released from polymer-coated urea (PCU) is from the mid-tillering stage to the jointing stage, while sulfur-coated urea (SCU) quickly releases N after transplanting [16,17]. Because the N release patterns of these CRNF types cannot be totally synchronized to meet plant N requirements, fertilizer blending methods (the mixed use of CRNF and normal urea) have been developed [18]. Our previous study showed that BBF, which is a blend product with 20% N from a quick-acting N source and 80% N from two PCUs, significantly improved rice yield compared with split application of urea [19].

Mechanical deep placement fertilization (DPF) is an effective application method to minimize N losses and to improve rice production. DPF provides immediate plant nutrient uptake and prolongs N availability, since the N fertilizer is applied close to the plant roots. Recently, a meta-analysis showed that DPF significantly increased rice yield and NUE by 32% and 34% [20], respectively, but conventional N fertilizer may not be suitable for mechanical single-time deep application due to its light weight which limits uniform application and rapid N release [21]. CRNF with a balanced slow-release N pattern might be a promising alternative to conventional N fertilizer for DPF. DPF has often been applied combined with CRNF for higher NUE and rice yield [22,23], and DPF with CRNF have been shown to not reduce rice growth and NUE compared to split application [23]. With the development of agricultural machinery and new CRNF types (e.g., BBF), improving rice yield and farmer income by optimizing CRNF in combination with DPF management continues to be a major concern. However, it is still unclear which CRNF-DPF combination is better for rice production and economic benefits.

Machine-transplanted rice with single-time DPF is an innovative agricultural practice in which transplanting rice seedlings and applying CRNF close to rice roots can be done simultaneously [10]. Previous studies have reported that there are significant differences in rice growth among different CRNFs [15,24]. However, to the best of our knowledge, limited information is available about the effects of different CRNF types with DPF on rice production and economic benefits in machine-transplanted rice.

Therefore, we conducted a two-year experiment to test the hypothesis that CRNF-DPF combination would balance the grain yield and NUE of rice and would increase the economic efficiency. The objectives of this study were (i) to quantify the effects of different CRNF-DPF combinations on the temporal and spatial distributions of soil mineral N, (ii) to evaluate the effects of different CRNF-DPF combinations on rice growth, and (iii) to explore the advantages of CRNF combined with mechanical single-time deep placement with respect to increasing yield and economic benefits.

2. Materials and Methods

2.1. Experimental Site

This 2-year field experiment was conducted at Danyang Experimental Station (32°00′ N, 119°32′ E), Jiangsu Province, China. At this site, rice (Oryza sativa L.) is grown in a rotation with winter wheat, and the cultivation system has been a rice–wheat rotation system for more than five years. This experimental area is characterized by a subtropical monsoon climate with a mean annual air temperature of 16.4 °C and a mean annual rainfall level of 882 mm. The meteorological data were obtained from a meteorological station (Watch Dog 2900ET, SPECTRUM, USA) positioned 100 m from the experimental station. The monthly average temperature and precipitation were recorded from rice transplantation to harvest over the 2015 and 2016 rice growing seasons (Figure 1). The average temperatures were 23.9 (17.9–27.4 °C) and 25.0 °C (18.4–29.2 °C) during the 2015 and 2016 rice growing seasons, respectively. The amounts of precipitation during the rice growing seasons were 433.8 and 1652.4 mm in 2015 and in 2016, respectively. The soil at the experimental site was classified as an Orthic Acrisol (FAO soil taxonomy). Before transplanting and fertilization, 0–20 cm of soil was taken for determination of soil basal fertility, following Bao’s method [25]. The soil basal fertility characteristics were as follows: a pH of 6.33 ± 0.04, an organic matter content of 19.56 ± 0.28 g kg⁻¹, a total N content of 1.25 g ± 0.02 kg⁻¹, a mineral N (NH₄⁺-N + NO₃⁻-N) content of 19.36 ± 0.31 mg kg⁻¹, an Olsen-P content of 27.92 ± 0.19 mg kg⁻¹, and an NH₄OAc-extractable K content of 168.00 ± 3.12 mg kg⁻¹.

2.2. Experimental Design

The experiment, in this study, was designed as a randomized complete block with three replications. Three types of CRNFs were used in this study: a polymer-coated urea (PCU, 4% (w/w) coating, 43% N), a sulfur-coated urea (SCU, 37% N), and a bulk blending fertilizer (BBF, 40% N). BBF is a product blend of 20% N from quick-acting N (ammonium phosphate) and 80% N from averaging two PCUs (urea with 2% and 4% (w/w) coating). The PCU, SCU, and BBF were provided by Hanfeng Slow Release Fertilizer Co., Ltd. (Jiangsu, China). Two fertilizer application methods were used for each CRNF: broadcast fertilization (BF) and mechanical deep placement fertilization (DPF). For BF, N fertilizer was broadcasted over the soil surface as a single basal application by hand, before rice machine transplanting. For DPF, mechanical-transplanted rice was synchronized with deep fertilization by machine, and the depth of fertilization was 5 cm [26]. High-yield split fertilization of urea (CK) was the control. For the CK treatment, urea was applied in 4 splits at the soil tillage (30%), tillering (20%), panicle initiation (30%), and heading (20%) stages. According to the recommended dose of N fertilizer by Hofmeier et al. [27], each treatment (except for N₀) was applied at a rate of 216 kg N ha⁻¹ in this study (~30% less than the conventional N dose of local farmers). P and K fertilizers were applied as basal dressings at rates of 108 kg P₂O₅ ha⁻¹ and 172 kg K₂O ha⁻¹, respectively. All the P and K fertilizers in all treatments were broadcast by hand as basal fertilizers, one day before transplanting. Ridges were built around each plot and covered with plastic film to form individual plots of 7.2 m × 20 m.

2.3. Crop and Water Management

For this study, Ningjing 7 (Oryza sativa L.), a high-yielding japonica rice bred by Nanjing Agricultural University, was selected. The seeds were soaked in water for 2 days; the volume ratio of the solution to seed was 3:1. After soaking, the seeds were kept in the dark at 30 °C for 1 day. Then, the germinated seeds (120 g) were sown in a seedling tray (57.5 × 27.5 cm) containing organic substrates, on 28 May 2015 and 31 May 2016. The nursery management adopted for seedling cultivation employed the application of micro-irrigation for the hard land seedling raising technique. Rice seedlings were transplanted using a rice blanket-seedling transplanter (PZ640, Iseki Agricultural Machinery Co., Ltd., Ehime-ken, Japan) on 20 June 2015 and 15 June 2016. During the rice growing season, the floodwater of all plots was maintained at a depth of 5 ± 2 cm, except for mid-season aeration. All other managements were consistent with local high-yield cultivation practices.

2.4. Sampling and Analysis

2.4.1. N Release Patterns of CRNF and Soil NH₄⁺-N

The N release rates of the CRNFs in paddy soils were determined by using the burying method [28]. The CRNFs (10 ± 0.01 g) were placed in sealed polypropylene bags (12 cm × 8 cm) with 1.0 mm² mesh. For the BF treatments, the fertilizer bags were laid on the soil surface, whereas for the DPF treatments, they were buried 5 cm below the soil surface. All the fertilizer bags were placed on the same day when rice seedlings were transplanted. For each treatment, 4 bags were collected at each sampling time (10, 20, 30, 40, 55, 95, and 140 days after transplanting). The bags were rinsed with distilled water and were dried by using a maxi-dry instrument (Heto-Holten, Paris, France). Residual samples in bags were passed through a 0.25-mm sieve, and their N contents were determined by using the Kjeldahl method [29]. The amount of N loss from the fertilizer bag was assumed to be equivalent to the N release rate.

For the BF treatments, the soil samples (0–10 cm) were separated into two subsamples: rooted soil (at an approximate distance of 2 cm from the plant, B1) and bulk soil (in between the plant rows, B2) (Figure 2). For the DPF treatments, the soil samples were collected as three subsamples: rooted soil with fertilizer (at a distance of ~2 cm from the plants on the fertilizing side, S0), rooted soil without fertilizer (at a distance of ~2 cm from the plants on the non-fertilizing side, S1), and bulk soil without fertilizer (in between plant rows on the non-fertilizing side, S2) (Figure 2). Four soil cores (5 cm diameter) of each plot were taken randomly at 10- to 12-day intervals within 50 days after transplanting (DAT), and the last 2 soil samples were collected at 100 and 140 days after transplanting. After sampling, the fresh soil samples were immediately extracted with 2 M KCl solution. In the presence of sodium nitroferricyanide, NH₄⁺ reacts with ClO⁻ to form a blue compound, which can be detected in concentration by spectrophotometry at 660 nm. NO₃⁻ is reduced to NO₂⁻ by cadmium, NO₂⁻ and sulfanilamide can generate diazonium, and then couple with N-1-naphthylethylenediamine dihydrochloride to generate a red dye, which can be used to determine the concentration of the spectrophotometer at 540 nm. According to the two methods described above, soil NH₄⁺-N and NO₃⁻-N concentrations were measured using an auto-analyzer (Model AA3, Bran-Luebbe, Hamburg, Germany). Exceptionally, NO₃⁻-N concentrations were undetectable, which might be due to the long-term flooding [30,31]. Thus, we only analyzed the NH₄⁺-N dynamics in this study.

2.4.2. Plant Sampling and Analysis

The plant samples were collected at the tillering, panicle initiation, heading, and maturity stages. Five representative hills were sampled per plot based on the mean tiller number. The plants were divided into two parts (stems and leaves) before the heading stage and into three parts (stems, leaves, and panicles) after the heading stage.

Once the rice had matured, the plant yield components (panicles per m², spikelets per panicle, filled grain rate, and filled grain weight) were determined from an area of 1.0 m², randomly sampled from each plot. The measurement of rice yield followed the method by Li et al. [32].

To obtain intact rice root systems, three rice plants were dug up along with the representative rhizosphere soil block; the size of the soil blocks was 60 cm long × 30 cm wide × 15 cm deep. The side strip fertilization treatment soil blocks were divided into two blocks according to the fertilization side and the non-fertilization side, and then marked. The blocks were placed in a 40-mesh nylon mesh bag and rinsed under floating water, during which the left and right adjacent root systems were gradually separated. The intact roots were further washed, scanned (Epson Expression 1680 Scanner, Seiko Epson Corp., Tokyo, Japan), and root morphology was quantified using a WinRHIZO Root Analyzer System (Regent Instruments Inc., Québec, QC, Canada). After the measurement, the roots were dried at 80 °C to a constant weight, and the root dry matter was weighed.

All the plant samples were oven-dried at 105 °C for 30 min, followed by drying at 80 °C to a constant weight. Dried samples were weighed, ground, and sifted through a 2-mm sieve. Tissue N contents (%) were measured following the Kjeldahl method, described by Bremner et al. [29].

2.4.3. NUE Calculation

The N use efficiency was defined as apparent N recovery (NUE, %), and was calculated as follows [33]:

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

(1)

where U_N and U₀ are the total N uptake (kg N ha⁻¹) at maturity with and without N fertilizer input, respectively, and F_N is the dose of N fertilizer input (kg N ha⁻¹).

2.4.4. Economic Benefit Analysis

Gross and net incomes were calculated as follows [18]:

Gross income (USD ha⁻¹) = Grain yield (kg ha⁻¹) × Price of grain (USD kg⁻¹)

(2)

Net income = Gross income − N fertilizer cost − Cost of N application − Other costs

(3)

The local market prices (2015) were as follows: seed 380.3 USD t⁻¹, urea 264.5 USD t⁻¹, SCU 367.4 USD t⁻¹, PCU 436.9 USD t⁻¹, BBF 393.1 USD t⁻¹, labor costs of broadcasting urea (four times) 176.4 USD ha⁻¹, labor costs of broadcasting CRNF (one time) 133.8 USD t⁻¹, and costs of mechanical deep placement 33.3 USD ha⁻¹. Other costs included P and K fertilizer, irrigation, pesticide and insecticide application, seeding, mechanical tillage, and harvesting (1672.8 USD ha⁻¹).

2.5. Statistical Analysis

All data were analyzed using the statistical program SPSS 18.0 (SPSS Statistics, SPSS Inc., Chicago, IL, USA). Datasets of each year were examined separately by two-way analysis of variance. The effects of year, CRNF type (SCU vs. PCU vs. BBF), fertilization application method (BF vs. DPF), and their interactions on dry matter accumulation, N uptake, NUE, rice yield, and yield components, as well as economic benefits were estimated by three-way ANOVA. The means of treatments were compared based on the least significant difference (LSD0.05) test.

3. Results

3.1. Cumulative N Release Curve of CRNF

The three CRNFs showed significantly different cumulative N release curves, while the application method did not affect the N release patterns of each CRNF (Figure 3). The cumulative N release exceeded 70% for SCU, 29% for PCU, and 38% for BBF, during the rice tillering stage (0–30 DAT). The cumulative N release was only about 3% for SCU, 30% for PCU, and 12% for BBF, from rice tillering to the panicle initiation stage (30–55 DAT). Finally, 8%, 17%, and 17% of cumulative N were released from SCU, PCU, and BBF from the panicle initiation stage to the maturity stage (55–140 DAT), respectively. Overall, PCU showed the slowest N release among the three CRNFs, and BBF had a medium N release pattern.

3.2. Soil NH₄⁺-N Concentrations

The soil NH₄⁺-N contents among different treatments displayed unique dynamics with respect to their corresponding N release patterns; DPF significantly increased the soil NH₄⁺-N content (Figure 4 and Figure 5). For the BF treatments, the NH₄⁺-N concentration in rooted soil (B1) was consistently higher than in bulk soil (B2) for all types of fertilizers during 15–80 DAT. CK showed three NH₄⁺-N peaks at the initial, mid-tillering, and panicle initiation stages, which might have been due to the urea addition. The mean NH₄⁺-N concentration during 0–9 DAT for BBF-BF (17.2 mg kg⁻¹ in 2015 and 33.8 mg kg⁻¹ in 2016) was less than those of SCU-BF (27.3 mg kg⁻¹ in 2015 and 53.7 mg kg⁻¹ in 2016). However, the opposite trend was identified during 20–60 DAT, i.e., the mean NH₄⁺-N concentrations for BBF-BF (44.8 mg kg⁻¹ in 2015 and 47.2 mg kg⁻¹ in 2016) were higher than those of SCU-BF (29.4 mg kg⁻¹ in 2015 and 38.2 mg kg⁻¹ in 2016). PCU-BF showed low NH₄⁺-N concentrations (14.5 mg kg⁻¹ in 2015 and 25.5 mg kg⁻¹ in 2016) during 0–15 DAT, and high concentrations (28.6 mg kg⁻¹ in 2015 and 38.1 mg kg⁻¹ in 2016) during 15–55 DAT, and then decreased quickly during 55–95 DAT.

For CRNF with DPF treatments, the NH₄⁺-N concentrations of rooted soil with fertilizer (S0) were greater than those of rooted soil without fertilizer (S1) and bulk soil without fertilizer (S2) (Figure 5). Notably, the NH₄⁺-N patterns of S0 in DPF were like those of B2 in BF for each CRNF. The mean NH₄⁺-N concentrations of S0 in SCU, PCU, and BBF were 47.8–53.4 mg kg⁻¹, 43.0–45.2 mg kg⁻¹, and 43.4–45.4 mg kg⁻¹ from transplanting to the heading stage, which were significantly higher than those of B2 soil under BF by 73.2–221.7%, 102.7–246.8%, and 97.4–156.8%, respectively.

3.3. Rice Tillering Dynamics

Both the fertilizer application method and the fertilizer type significantly affected the rice tillering dynamics (Figure 6). The number of tillers peaked at ~40 DAT among all treatments, and then significantly decreased in the second year. DPF increased the number of tillers for all types of CRNF in both years, and BBF had the greatest increase compared with CK. Notably, PCU showed fewer tillers during the early tillering stage in 2016.

3.4. Rice Plant Dry Matter Accumulation

Both the fertilizer application method and the fertilizer type significantly affected rice dry matter accumulation, and application × type interaction was also significant for plant dry matter accumulation at the tillering, heading, and maturity stages (

Table 1. Effect of different fertilizer treatments on rice dry matter accumulations in 2015 and in 2016.

Year	Fertilizer Application Method	Fertilizer Type	Dry Matter Accumulations (t ha−1)
			Tillering	Panicle Initiation	Heading	Maturity
2015		CK	0.70 ± 0.02	4.19 ± 0.22	10.73 ± 0.13	20.10 ± 0.11
	Broadcast	SCU	0.63 ± 0.01	3.68 ± 0.04	11.01 ± 0.08	19.06 ± 0.40
		PCU	0.48 ± 0.02	4.56 ± 0.08	11.19 ± 0.05	19.82 ± 0.48
		BBF	0.58 ± 0.01	5.01 ± 0.07	11.67 ± 0.33	22.39 ± 0.53
	Deep placement	SCU	0.86 ± 0.03	4.49 ± 0.31	11.97 ± 0.08	23.21 ± 0.41
		PCU	0.75 ± 0.02	5.96 ± 0.22	12.40 ± 0.08	18.93 ± 0.57
		BBF	0.70 ± 0.01	5.90 ± 0.14	11.45 ± 0.04	22.39 ± 0.30
		LSD0.05	0.07	0.69	0.57	1.49
2016		CK	0.93 ± 0.04	5.35 ± 0.04	12.60 ± 0.34	18.56 ± 0.33
	Broadcast	SCU	0.84 ± 0.01	3.91 ± 0.03	11.21 ± 0.11	17.84 ± 0.08
		PCU	0.43 ± 0.02	4.17 ± 0.05	10.17 ± 0.06	16.88 ± 0.13
		BBF	0.70 ± 0.01	5.33 ± 0.01	12.50 ± 0.23	18.93 ± 0.17
	Deep placement	SCU	1.10 ± 0.04	5.98 ± 0.12	12.89 ± 0.48	20.71 ± 0.23
		PCU	0.66 ± 0.04	5.08 ± 0.18	11.88 ±	17.18 ± 0.41
		BBF	0.80 ± 0.01	6.74 ± 0.02	12.60 ±	19.29 ± 0.04
		LSD0.05	0.12	0.32	1.40	0.93
p-value	Year (Y)		**	ns	ns	**
Fertilizer placement (P)			**	**	**	**
Fertilizer type (T)			**	**	*	**
	Y × P		ns	ns	ns	ns
	Y × T		**	**	**	ns
	P × T		**	ns	**	**
	Y × P × T		ns	*	ns	ns

Notes: CK, high-yield split fertilization of urea; PCU, a polymer-coated urea (4% (w/w) coating, 43% N); SCU, a sulfur-coated urea (37% N); BBF, a bulk blending fertilizer (40% N); ns, not significant. (p > 0.05), * significant at p ≤ 0.05, and ** significant at p ≤ 0.01.

). Plant dry matter accumulations of SCU and PCU were lower than those of CK in BF, while they were higher in DPF. BBF had greater dry matter accumulation than CK under the two fertilizer application methods. Plant dry matter accumulation of SCU-DPF was the highest level among all treatments at maturity, followed by BBF-DPF without significance.

3.5. Rice Yield and NUE

Both the fertilizer application method and the fertilizer type significantly affected rice yield; NUE and application × type interaction was also significant (Figure 7). DPF significantly improved the rice yield and NUE of CRFN compared with those of BF. The rice yield and NUE of SCU were lower than those of all treatments in BF, while they were higher in DPF. However, the fertilizer application methods did not affect the rice yield of PCU; single-time application of PCU under both fertilizer application method treatments had significantly lower yields than CK. DPF significantly increased the NUE of SCU by 66.1–100.8% and of PCU by 24.4–26.3%. However, no significant differences in NUE were observed in BBF between the two fertilizer application method treatments. Moreover, SCU-DPF was similar to BBF-DPF in NUE, and both were greater than CK. Notably, single-time application of BBF resulted in higher rice yield and NUE than those of the CK under both fertilizer application method treatments.

3.6. Rice Yield Components

Fertilizer significantly affected rice yield components, and different CRNFs also affected rice panicles and total spikelets (

Table 2. Effect of different fertilizer treatments on rice yield components in 2015 and in 2016.

Year	Fertilizer Application Method	Fertilizer Type	Panicles m−2	Spikelets Panicle−1	Total Spikelets	Filled Grain Rate	Filled Grain Weight
					(×103 m−2)	(%)	(mg)
2015		CK	263 ± 3.76	207 ± 0.85	54.6 ± 0.65	75.9 ± 0.54	28.2 ± 0.01
	Broadcast	SCU	273 ± 5.53	188 ± 1.24	51.5 ± 1.22	73.1 ± 0.83	27.5 ± 0.18
		PCU	287 ± 8.18	193 ± 2.13	55.4 ± 0.99	71.7 ± 0.70	26.8 ± 0.12
		BBF	313 ± 5.08	197 ± 0.24	61.9 ± 1.08	72.7 ± 0.63	27.1 ± 0.14
	Deep placement	SCU	308 ± 2.36	207 ± 2.83	63.9 ± 0.39	72.3 ± 0.54	27.7 ± 0.05
		PCU	332 ± 9.24	185 ± 1.67	61.5 ± 1.24	64.4 ± 2.38	27.0 ± 0.04
		BBF	325 ± 0.11	197 ± 1.66	64.1 ± 0.53	72.1 ± 0.95	27.4 ± 0.1
		LSD0.05	20.65	7.47	2.34	4.86	0.45
2016		CK	279 ± 5.22	162 ± 0.90	45.4 ± 0.72	91.5 ± 0.06	27.9 ± 0.01
	Broadcast	SCU	280 ± 0.67	153 ± 0.54	43.0 ± 0.25	87.0 ± 1.05	28.1 ± 0.01
		PCU	316 ± 2.34	142 ± 1.48	45.0 ± 0.78	84.9 ± 1.04	27.4 ± 0.12
		BBF	320 ± 2.99	157 ± 3.30	50.2 ± 1.31	86.8 ± 0.96	27.6 ± 0.20
	Deep placement	SCU	302 ± 2.47	161 ± 3.75	48.6 ± 0.96	92.2 ± 0.38	28.0 ± 0.09
		PCU	348 ± 1.90	142 ± 4.69	49.6 ± 1.39	79.1 ± 0.28	26.5 ± 0.17
		BBF	333 ± 1.91	156 ± 0.82	52.0 ± 0.28	86.0 ± 3.03	27.0 ± 0.35
		LSD0.05	10.96	11.35	3.71	2.14	0.67
p-value	Year (Y)		*	**	**	**	ns
	Fertilizer placement (P)		**	ns	**	ns	ns
	Fertilizer type (T)		**	**	**	**	*
	Y × P		ns	ns	ns	ns	**
	Y × T		*	ns	ns	ns	**
	P × T		ns	*	ns	**	ns
	Y × P × T		ns	ns	ns	ns	ns

Notes. CK, high-yield split fertilization of urea; PCU, a polymer-coated urea (4% (w/w) coating, 43% N); SCU, a sulfur-coated urea (37% N); BBF, a bulk blending fertilizer (40% N). ns: not significant (p > 0.05), *: significant at p ≤ 0.05, **: significant at p ≤ 0.01.

). BBF had both higher panicles and total spikelets than CK with no significant differences under the two fertilizer application method treatments. DPF significantly increased the rice yield components of SCU and PCU. Compared to BF, DPF significantly increased the number of panicles and total spikelets by 3.8–15.7% and 3.6–24.1%, respectively, while filled grain rate decreased by 0.8–11.3%.

3.7. Economic Benefits

A significant interaction effect of the application method and fertilizer type were found with respect to net incomes (

Table 3. Economic benefits under different fertilizer treatments in 2015 and in 2016.

Year	Fertilizer Application Method	Fertilizer Type	Gross Income	N Fertilizer Cost	Cost of N Application	Other Costs	Net Income
			(USD ha−1)
2015		CK	4462.2 ± 68.9	124.2	176.4	1672.8	2488.8 ± 68.9
	Broadcast	SCU	3942.4 ± 39.2	214.5	78.1	1672.8	1977.0 ± 39.2
		PCU	4045.5 ± 15.8	224.7	68.8	1672.8	2079.2 ± 15.8
		BBF	4652.3 ± 87.3	212.3	72.3	1672.8	2694.9 ± 87.3
	Deep placement	SCU	4867.8 ± 15.9	214.5	33.3	1672.8	2947.2 ± 15.9
		PCU	4056.5 ± 79.7	224.7	33.3	1672.8	2125.7 ± 79.7
		BBF	4842.5 ± 63.5	212.3	33.3	1672.8	2924.1 ± 63.5
		LSD0.05	965.7	-	-	-	373.9
2016		CK	4424.2 ± 56.6	124.2	176.4	1672.8	2450.8 ± 56.6
	Broadcast	SCU	4005.8 ± 20.5	214.5	78.1	1672.8	2040.4 ± 20.5
		PCU	3994.8 ± 50.7	224.7	68.8	1672.8	2028.5 ± 50.7
		BBF	4576.3 ± 32.5	212.3	72.3	1672.8	2618.9 ± 32.5
	Deep placement	SCU	4766.4 ± 59.7	214.5	33.3	1672.8	2845.8 ± 59.7
		PCU	3955.1 ± 116.5	224.7	33.3	1672.8	2024.3 ± 116.5
		BBF	4601.6 ± 73.8	212.3	33.3	1672.8	2683.2 ± 73.8
		LSD0.05	830.9	-	-	-	401.6
p-value	Year (Y)		**				ns
Fertilizer placement (P)			**				**
Fertilizer type (T)			**				**
	Y × P		**				ns
	Y × T		**				ns
	P × T		**				**
	Y × P × T		**				ns

Notes. CK, high-yield split fertilization of urea; PCU, a polymer-coated urea (4% (w/w) coating, 43% N); SCU, a sulfur-coated urea (37% N); BBF, a bulk blending fertilizer (40% N). ns: not significant (p > 0.05), ** significant at p ≤ 0.01.

). SCU-DPF significantly increased net incomes by 39.5–49.1% compared with SCU-BF. No significant differences in net incomes were observed for PCU and BBF, between the two application method treatments. Furthermore, the net incomes of BBF under the two application method treatments increased compared with those of CK, but SCU had the highest net incomes among all treatments under DPF.

4. Discussion

4.1. Different Fertilizer Types and Application Methods Affected the Temporal and Spatial Distributions of Soil NH₄⁺-N

Our study showed that the NH₄⁺-N concentration in B2 soil was significantly higher than that in B1 soil for each CRNF, during 15–80 DAT under BF (Figure 4), which was similar to the results of Xu et al. [34]. This may be explained by the high levels of plant N uptake from the rhizosphere during the rice growth period [35]. The high NH₄⁺-N concentration in B2 soil could not be readily utilized by the root system; however, there was an increased risk of N loss through NH₃ volatilization, N₂O emissions, and runoff [12,20]. Thus, conventional broadcasting of N fertilizer, often, could not ensure a balance between N release and its uptake by crops at the right time and place in order to maintain a high NUE [36]. Actually, the patterns of NH₄⁺-N concentrations in B2 soil correlate well with the cumulative N release curve of each CRNF (Figure 3 and Figure 4). Therefore, single-time broadcast application of CRNF still needs to be improved to achieve high NUE.

CRNF gradually releases N to meet plant growth requirements [11], but some previous studies have also reported negative effects of CRNF on rice growth, which might be due to inappropriate N release patterns [14,15,19]. Similar to these studies, we found that the mean NH₄⁺-N concentration during 0–9 DAT for BBF-BF was less than that for SCU-BF, while the opposite trend was identified during 20–60 DAT. PCU-BF showed low NH₄⁺-N concentrations during 0–15 DAT and high concentrations during 15–55 DAT. N release that is too quick (SCU) or slow (PCU) is not conducive to rice growth. Li et al. [19] reported that SCU had higher NH₄⁺ concentrations for quick N release and the N release of PUC was slower than other CRNFs, but these CRNFs both decreased rice yield and NUE compared to split application of urea.

Many previous studies have found that N deep placement was an efficient alternative practice for broadcasting [13,20,23]. Broadcasting N fertilizer often leads to high NH₄⁺-N in surface water and non-rooted soil, with low NUE [37,38,39]. However, DPF can directly provide N to emerging rice roots [40] and prolong N availability [41]. Although the application method did not affect the N release patterns of the CRNFs in our field experiment, CRNF-DPF increased NH₄⁺-N concentration and prolonged N availability on rooted soil compared to CRNF-BF (Figure 3 and Figure 5). In addition, the NH₄⁺-N concentration of rooted soil with fertilizer (S0) was significantly higher than rooted soil without fertilizer (S1) and bulk soil without fertilizer (S2) under CRNF-DPF treatments, mainly due to the localized placement of CRNF (Figure 5). Thus, DPF prolonged the timing of fertilizer contact with soil, which improved NH₄⁺-N fixed on the surface of clay minerals and reduced NH₄⁺-N in the surface water [42,43,44].

4.2. Single-Time Broadcast Application of CRNF Cannot Be a Substitute for Split Application of Urea

Urea is the main N fertilizer in agriculture [45], but the rapid N release of urea leads to 20–70% of the N not being used by plants [46], resulting in yields inconsistent with expectations and low NUE. Split application of urea is considered to be an effective practice to achieve high yield and NUE. In general, this practice involves applying urea two or more times (in splits), according to the crop N requirements, to maintain synchronization between N demand and supply [13]. Split application of urea has certainly improved crop yield and NUE, but it cannot be ignored that the additional operations will cost more labor input due to the lack of topdressing machinery [47]. In theory, single-time application of CRNF can replace split application of urea because of slow N release, which can reduce labor [10].

However, some recent studies have reported that single-time application of CRNF could limit rice growth due to inappropriate release of N [19]. Similar to these studies, SCU-BF and PCU-BF significantly reduced rice yield by 9.5–11.6% and 9.4–9.7%, respectively, compared with CK in our 2-year field experiments. On the one hand, SCU released N quickly, which led to a short supply of nitrogen during late growth stages. Limited N supply inhibited rice root growth (Table S1) and decreased rice plant N content (Table S2), which might be detrimental to competing grain filling [19]. On the other hand, PCU with slow N release did not meet rice N uptake demands due to the N deficit during the early tillering stage and N excess during the ineffective tillering stage [14].

Indeed, sufficient N supply at critical growth stages is the key to high yield [48,49]. Thus, SCU-BF and PCU-BF significantly reduced photosynthetic rate and dry matter accumulation and both CRNFs reduced spikelets per panicle and grain filling due to the limited supply of N under BF treatments (Table 2). Notably, BBF partly compensated for the deficiencies of soil NH₄⁺-N contents in SCU and PCU at either the later or the early stage, and BBF resulted in a yield comparable to CK in both years.

4.3. Optimization N Management Synchronized N Supply and Requirements and Raised Net Income

In our study, DPF actually significantly improved root growth, and SCU-DPF had the greatest root system among all treatments. More roots often led to stronger N uptake. Plant dry matter accumulation of SCU and PCU was lower than that of CK under BF, while it was higher under DPF in our experiment, similar to a previous study [23].

DPF did not affect N release patterns of the CRNFs (Figure 3), but increased the levels of localized NH₄⁺-N in soil during the rice growth stage, especially at 0–60 DAT (Figure 5), which might help to induce productive tillers [50]. Our results showed that DPF increased rice tiller numbers and dry matter accumulation at the tillering and panicle initiation stages (Figure 6 and Table 1), consistent with many previous studies [23,51,52,53]. Compared to SCU-BF, SCU-DPF increased rice yields by 19.0–24.3% and NUE levels by 24.4–26.3%. Above all, the yield and NUE of SCU-DPF were greater than those of CK by 7.7–9.1% and 6.0–24.4%, respectively. The increase in tiller number with DPF as compared to BF is consistent with many previous studies [24,32,33,34]. Subsurface or deep placement can supply more N to the root rhizosphere and preserve fertilizer N availability, especially during the early stage, which can help to induce productive tillers and N uptake [50]. Therefore, for SCU-DPF, its large amount of N release in the early stage increased N leaching loss, but still increased the content of soil NH₄⁺-N in the root zone, so overall, it promoted growth of the rice population in the early stage and increased the amount of N accumulated in the early stage, which in turn promoted the yield and NUE. DPF increased the panicle number, while significantly decreasing the filled grain of PCU, and single-time application of PCU decreased rice yield as compared with CK (Table 1). Wang et al. [54] reported that rice leaf N content was closely related to the number of filled grains and ideal leaf N content of japonica rice at the heading stage was 2.7–2.8%. But the N contents of leaves in PCU-DPF (3.25% in 2015 and 3.39% in 2016) were far higher than the normal content and other treatments (Table S1), which might be against grain filling.

In general, adjustment of fertilizer type and application method to synchronize N supply and demand can promote root growth and N uptake, thus improving NUE and rice yield. Actually, the fertilization cost of CRNF (33.3~78.1 USD ha⁻¹) was lower than those of urea split application (176.4 USD ha⁻¹), and applying DPF could further reduce fertilization cost due to the use of machinery (Table 3). Moreover, DPF was also beneficial to inhibit weeds which might reduce the production cost [55]. In the present study, SCU was suitable to combine with deep placement application, which could significantly increase rice yields and NUE and had higher economic advantages than urea split application. Above all, BBF was a good choice to substitute for split application of urea irrespective of the fertilizer application method used.

5. Conclusions

While the types of CRNF significantly influenced the nitrogen release patterns, the method of application further dictated the effectiveness of these fertilizers. Specifically, DPF was found to notably enhance NUE, especially when combined with SCU, leading to augmented rice yields and economic returns. Conversely, PCU did not yield favorable outcomes. BBF showcased versatility, yielding both high NUE and economic benefits irrespective of the application method. Importantly, the application of SCU with DPF not only optimized nitrogen management but also maximized economic benefits, suggesting a potential paradigm shift in N fertilizer management for machine-transplanted rice. This research underscores the necessity of a dual-focused approach, emphasizing both the selection of appropriate CRNFs and their optimized application methods to ensure enhanced crop yields, increased farmer profits, and sustainable agriculture practices.