Replacing Nitrogen Fertilizers with Incorporation of Rice Straw and Chinese Milk Vetch Maintained Rice Productivity

Abstract

The cultivation of Chinese milk vetch (CMV) during the winter fallow season and the return of rice straw are important practices for increasing the soil fertility of paddy fields in southern China. In order to provide data-based evidence for the scientific strategy of nitrogen (N) fertilizer reduction through the incorporation of rice straw and CMV, a three-year field trial was conducted. The treatments included the three N application rates of 0%, 60%, and 100% of the local conventional rate (165 kg ha⁻¹), with the incorporation of CMV alone (MN0, MN60, and MN100) or with both CMV and rice straw (SMN60 and SMN100). The rice grain yield, N uptake, and dynamic changes in inorganic N in the soil and surface water were determined for the period from 2019 to 2021. The results show that both the rice grain yield and plant N uptake of the MN60 and SMN60 treatments were not significantly different from those of the treatment with only conventional N application (N100). Although the SMN100 treatment significantly increased the uptakes of N in the aboveground part in the tillering and shooting stages compared with SMN60, no significant differences were found between the grain yields in 2021. Meanwhile, the SMN60 treatment significantly increased the soil microbial biomass N and NH₄⁺-N contents during the maturity stage in 2020 and 2021, respectively, compared with MN60. Furthermore, the SMN100 treatment resulted in higher NO₃⁻-N concentrations in the surface water at days 3 and 6 after transplantation in 2020 than those under SMN60. In conclusion, the incorporation of CMV and rice straw with an application rate of 60% of conventional N fertilizer is an essential approach to reducing the risk of N loss while maintaining rice grain yields in the Jianghan Plain of China.

1. Introduction

Planting leguminous green manure in summer or winter in rotation with the main crop is a way of making full use of fallow fields. Also, it benefits the growth and development of leguminous green manure itself, as well as the subsequent main crop [1]. Previous studies have shown that leguminous green manure can increase N supply, improve N-use efficiency, and maintain the N balance in a crop rotation system through its biological N fixation, thus increasing the yield of the main crop and reducing the application rate of N chemical fertilizers [2,3,4]. As a high-quality source of organic fertilizer, Chinese milk vetch (CMV, Astragalus sinicus L.) can be planted to reduce soil erosion, optimize the soil’s physical structure, and increase the soil’s nutrient content to a certain extent [5,6].

When chemical fertilizers are applied to a field, the nutrients are rapidly released. After incorporation, the CMV decomposes at a fast rate during the early stage and at a slow rate during the later stage, so CMV and chemical fertilizers can be applied together to ensure both a high demand for nutrients in the early growth stage of a crop, and the supply of nutrients in the later stage [7]. Studies have shown that the biological N fixation by CMV ranged from 60 to 115 kg ha⁻¹ [8]. In southern China, the incorporation of CMV can replace 20–40% of N fertilizer without reducing the rice (Oryza sativa L.) yield [3].

It has been pointed out that returning straw to the soil may result in adverse effects [9,10]. When the C/N ratio of straw exceeds the demand of soil microorganisms, the addition of organic carbon will promote microbial reproduction, which causes N competition between microorganisms and crops [11,12,13]. Therefore, the incorporation of crop residues with an appropriate C/N ratio is essential for soil fertility and crop growth. Recently, the incorporation of green manure and rice straw has become a hot spot in research. The C/N ratio of rice straw (generally 50~70:1) is much higher than that of CMV (10~20:1); thus, the combination of them is helpful for the optimization of the C/N ratio [14]. The incorporation of CMV and rice straw promotes the activities of β-glucosidase and cellulose hydrolase, thus enhancing the decomposition of rice straw and nutrient release [15].

The N released through mineralization after the incorporation of CMV and rice straw into a field can replenish the soil’s N pool [16]. Combined with that, an appropriate amount of N fertilizer will meet the N demand of rice for the whole growth period. Previous studies have mostly focused on the nutrient release of CMV and rice straw when they are incorporated into a field alone. There is a lack of data on the N cycle during each growing stage of rice after incorporating both of them. The objective of the present study was to compare the N-supplying capacity of soils after the incorporation of CMV alone with that after the incorporation of CMV and rice straw at both 60% and 100% of the conventional N fertilizer application rate.

2. Materials and Methods

2.1. Overview of the Field Trial

The present study was conducted at Taihu farm, Jianghan Plain, Hubei Province (N 30°22′1″, E 112°2′57″; altitude: 44.6 m), where the local climate was subtropical monsoon, with a monthly average temperature of 17.6 °C and an annual average precipitation of 1069 mm, during the 2019–2021 growing seasons (Figure 1). The paddy soil was developed from alluvial deposits with the following properties (0–20 cm): pH of 7.6, organic matter of 22.4 g kg⁻¹, total N of 2.0 g kg⁻¹, available phosphorus of 10.6 mg kg⁻¹, and available potassium of 156.0 mg kg⁻¹. The local traditional crop rotation pattern was mid-season rice, grown from May to September of every year, followed by a winter fallow.

The field trial was established in the fall of 2018. There were seven treatments, and these included two treatments with N fertilizer application rates of 0 (N0) and 165 kg ha⁻¹ (100% of local conventional rate, N100), without the addition of crop residues. With the other three treatments, only CMV was incorporated, and N fertilizer was applied at 0 (MN0), 99 (60% of conventional rate, MN60), and 165 kg ha⁻¹ (MN100). The remaining two treatments involved the incorporation of CMV and rice straw, with N fertilizer application rates of 99 (SMN60) and 165 kg ha⁻¹ (SMN100). The total N input of each treatment is shown in

Table 1. The nitrogen (N) input (kg ha⁻¹) of different treatments in 2020 and 2021.

Treatment	Urea	Chinese Milk Vetch (M)		Rice Straw (S)		Total
		2020	2021	2020	2021	2020	2021
N0
MN0		75.4	41.6			75.4	41.6
MN60	99.0	65.8	47.5			164.8	146.5
SMN60	99.0	62.3	47.4	35.2	43.4	196.5	189.9
N100	165.0					165.0	165.0
MN100	165.0	59.9	38.0			224.9	203.0
SMN100	165.0	57.6	62.2	40.0	50.0	262.6	277.2

The N input of Chinese milk vetch is biological N fixation with a N fixation rate of 66%, which was determined by the authors through the ¹⁵N Natural Abundance Method. The dry-weight biomass and N concentrations of Chinese milk vetch and rice straw are shown in Table S1.

. Phosphorus and potassium were applied at 60 kg ha⁻¹ of P₂O₅ and 90 kg ha⁻¹ of K₂O for all the treatments.

Treatments were arranged in a randomized complete block design with three replications and were applied to the same plot every year. Each plot was 4 m wide and 5 m long and was spaced with 30 cm-wide ridges. For the rice straw treatments, all the rice straws were incorporated into the soil after rice grain harvest; otherwise they were removed from the plots. The CMV seeds were sown in the first week of October at a rate of 30 kg ha⁻¹, with biomass incorporated on the spot during the full blooming stage (around 15 April in the following year). After CMV incorporation (with a depth of 10–15 cm), the field was subject to flooding for 4–5 weeks. The rice variety was “Fengliangyou No. 2” and 400 rice hills were transplanted for each plot. No fertilizers were added in the CMV growing season, with urea, calcium superphosphate, and potassium chloride applied in the rice season. On the day of rice transplantation, 70% of the total N fertilizers and all phosphorus and potassium fertilizers were spread on the surface of each plot. The remaining N fertilizers were applied at 13 and 70 days after transplantation (DAT). The pest control and irrigation management followed local practices.

2.2. Plant Sampling and Analysis

Before incorporation into soil, the aboveground fresh weight of CMV in the whole plot was measured; subsamples were randomly taken to measure the dry weight and N concentration. In the rice growing season, the aboveground biomasses of five hills per plot were collected in the tillering, shooting, grain filling, and maturity stages (i.e., approximately 30, 60, 90, and 120 DAT). In the latter two stages, grain and shoot were separated to measure dry weight individually. During the maturity stage, rice grains in the whole plot were harvested and the yield was determined after air-drying. In this study, rice yield data were from 2019 to 2021, and the rest of the data were from 2020 to 2021.

All plant samples were oven-dried at 105 °C for 30 min and then at 65 °C to a constant weight. After drying, they were ground to pass a 0.84 mm sieve and digested by H₂SO₄-H₂O₂. The N concentration was measured using the semi-micro Kjeldahl method [17]. Plant N uptake was calculated by multiplying N concentration by dry weight.

2.3. Soil Sampling and Analysis

Soil samples were taken one day before rice transplantation (18 May 2020 and 20 May 2021) and in the rice tillering, shooting, and maturity stages. Five subsamples were collected at a depth of 0 to 20 cm from each plot. The soil samples were air-dried, ground, and passed through a 2 mm sieve for the measurement of microbial biomass N (MBN), ammonium-N (NH₄⁺-N), and nitrate-N (NO₃⁻-N) contents. The MBN content was measured by the chloroform fumigation–UV spectrophotometer method [18]. Soil NH₄⁺-N and NO₃⁻-N were extracted by 1 mol L⁻¹ KCl and determined using an AA3 Continuous Flow Injection Analyzer (SEAL Analytical, Norderstedt, Germany).

2.4. Water Sampling and Analysis

Field surface water samples were taken by syringe before fertilizer application, and at 3, 6, 13, 16, and 28 DAT in 2020 and 2, 8, 13, 15, and 29 DAT in 2021. Five subsamples were randomly collected in each plot and mixed as one sample. These samples were sent to the laboratory in ice boxes, filtered through a 0.45 μm membrane, and stored in a 4 °C refrigerator before analysis. The NH₄⁺-N and NO₃⁻-N concentrations in water samples were determined by an AA3 Continuous Flow Injection Analyzer.

2.5. Statistical Analysis

The data from each year were analyzed separately because of the different weather conditions. One-way analysis of variance was performed through SAS (Version 9.2). Duncan’s Multiple Range Test was used to compare the significance of differences at the p < 0.05 level.

3. Results

3.1. Rice Grain Yield

The rice grain yields of all the treatments in 2021 were lower than those in 2019 and 2020, which might be due to the higher rainfall accumulation in the grain filling stage of 2021, but a similar trend was shown among the treatments over the three growing seasons (Figure 2). The yield of the MN0 treatment was significantly higher than that of N0 over three years, but significantly lower than that of MN60 in 2019 and 2020. No significant differences were found among the MN60, SMN60, and N100 treatments, indicating that CMV alone or in combination with rice straw incorporated into the field could replace 40% of conventional N fertilizer without reducing rice yield. There were no significant differences in rice yields between the MN60 and MN100 treatments in three years. Additionally, no significant differences were found between the SMN60 and SMN100 treatments except in 2020, indicating that increasing the N application rate from 60% to 100% of conventional rate could not improve rice yield any more when CMV was incorporated with or without rice straw.

3.2. Nitrogen Uptake

The MN60 and SMN60 treatments did not significantly decrease the N concentrations in rice plant compared with the N100 treatment (except rice straw in the grain filling stage in 2020) (

Table 2. Plant nitrogen (N) concentrations (%) during different rice growing stages in response to the incorporation of Chinese milk vetch (M) and rice straw (S) at different N application rates.

Year	Treatments	Tillering Stage	Shooting Stage	Grain Filling Stage		Maturity Stage
		Aboveground	Aboveground	Panicle	Straw	Grain	Straw
2020	N0	2.28 d	1.50 c	0.90 b	0.68 d	0.98 b	0.40 d
	MN0	2.53 cd	1.72 bc	0.94 b	0.72 d	1.05 b	0.48 cd
	MN60	3.21 b	1.94 ab	1.09 ab	1.00 c	1.20 a	0.59 bc
	SMN60	2.98 b	1.67 bc	0.99 ab	0.99 c	1.17 a	0.62 bc
	N100	2.63 c	1.95 ab	1.18 a	1.20 b	1.17 a	0.67 ab
	MN100	2.98 b	2.13 a	1.18 a	1.55 a	1.22 a	0.77 a
	SMN100	3.52 a	2.05 a	1.19 a	1.39 a	1.26 a	0.64 ab
2021	N0	2.80 b	1.31 d	0.97 c	0.46 e	0.81 c	0.57 b
	MN0	2.90 b	1.43 cd	1.04 c	0.61 de	1.01 b	0.61 b
	MN60	3.30 ab	1.85 abc	1.04 c	0.91 bcd	1.11 ab	0.62 b
	SMN60	3.04 ab	1.65 bcd	1.13 bc	0.95 bc	1.11 ab	0.68 b
	N100	3.25 ab	1.68 bcd	1.08 c	0.83 cd	1.09 ab	0.69 b
	MN100	3.56 a	2.15 a	1.45 a	1.48 a	1.25 a	1.04 a
	SMN100	3.51 a	1.95 ab	1.35 ab	1.22 ab	1.20 ab	1.07 a

Data in the table are the means of three replications and different lowercase letters in each column indicate significant difference at the p < 0.05 level.

). The N concentrations in the rice straw under the MN100 and SMN100 treatments were significantly higher than those under N100 in the grain filling stage of both growing seasons and the maturity stage in 2021. When CMV was incorporated alone, there were no significant differences in the N concentrations of rice plant between the MN60 and MN100 treatments in rice tillering and shooting stages. However, the N concentration in rice straw under the MN100 treatment was significantly higher than that under MN60 in the grain filling and maturity stages. When CMV and rice straw were incorporated, the N concentrations in both the whole aboveground part in the tillering and shooting stages and the rice straw in the grain filling stage under the SMN100 treatment were significantly higher than those under SMN60 in 2020.

The variation in N uptake by rice was similar to that in N concentration, with the lowest N uptake seen in the N0 treatment in both years (

Table 3. Plant nitrogen (N) uptake (kg ha⁻¹) during different rice growing stages in response to the incorporation of Chinese milk vetch (M) and rice straw (S) at different N application rates.

Year	Treatments	Tillering Stage	Shooting Stage	Grain Filling Stage		Maturity Stage
		Aboveground	Aboveground	Panicle	Straw	Grain	Straw
2020	N0	24.5 ± 1.5 e	43.6 ± 2.7 c	29.4 ± 2.9 a	21.2 ± 2.0 c	66.5 ± 5.0 d	21.9 ± 3.2 d
	MN0	37.0 ± 0.8 d	70.7 ± 8.4 b	33.6 ± 1.9 a	24.7 ± 2.1 c	78.3 ± 3.0 c	31.0 ± 4.2 c
	MN60	45.3 ± 1.3 c	78.7 ± 13.1 ab	48.0 ± 6.9 a	45.4 ± 6.0 b	99.0 ± 3.4 b	40.8 ± 0.7 b
	SMN60	50.1 ± 4.5 c	68.2 ± 6.6 b	41.7 ± 7.3 a	43.3 ± 3.7 b	94.3 ± 0.3 b	43.4 ± 2.1 b
	N100	47.2 ± 1.7 c	91.2 ± 2.6 ab	44.4 ± 3.0 a	55.6 ± 3.2 b	94.6 ± 3.1 b	47.0 ± 5.8 b
	MN100	62.5 ± 3.0 b	80.9 ± 11.9 ab	43.6 ± 7.1 a	80.2 ± 3.1 a	104.4 ± 8.1 ab	59.1 ± 0.5 a
	SMN100	70.3 ± 3.7 a	100.8 ± 8.3 a	50.0 ± 2.8 a	71.2 ± 7.3 a	112.3 ± 2.3 a	50.0 ± 2.6 b
2021	N0	27.3 ± 5.5 d	53.2 ± 3.0 d	51.9 ± 1.9 a	24.2 ± 2.3 c	42.2 ± 3.3 d	26.2 ± 1.9 c
	MN0	39.3 ± 5.2 cd	77.0 ± 8.3 cd	65.2 ± 5.7 a	39.7 ± 3.5 bc	66.7 ± 5.0 c	33.0 ± 0.6 bc
	MN60	50.8 ± 3.7 bc	120.5 ± 21.1 ab	59.4 ± 6.5 a	64.3 ± 9.9 b	80.0 ± 1.4 b	42.7 ± 1.4 bc
	SMN60	57.0 ± 3.6 b	105.1 ± 3.9 bc	71.6 ± 7.9 a	75.8 ± 10.2 ab	78.1 ± 4.4 bc	47.7 ± 5.2 b
	N100	58.2 ± 1.4 b	102.3 ± 24.0 bc	71.0 ± 11.5 a	61.1 ± 0.1 bc	74.3 ± 3.1 bc	45.0 ± 4.3 b
	MN100	53.2 ± 6.4 b	123.5 ± 18.7 ab	61.3 ± 4.9 a	103.9 ± 21.6 a	82.4 ± 4.2 ab	64.0 ± 5.5 a
	SMN100	80.4 ± 9.7 a	154.4 ± 9.9 a	73.4 ± 8.4 a	105.4 ± 11.7 a	92.2 ± 3.8 a	72.5 ± 10.0 a

Data in the table are mean ± standard error and different lowercase letters in each column indicate significant difference at the p < 0.05 level.

). Plant N uptakes among the MN60, SMN60, and N100 treatments did not differ significantly during all stages, indicating that N fertilizer reduction by 40% with crop residues incorporation did not significantly affect N accumulation in the rice plant. In the maturity stage, N uptakes in rice grain and straw under MN100 and SMN100 increased by 10.4–24.1% and 6.4–61.1%, respectively, compared with those under N100, indicating that CMV incorporation alone or in combination with rice straw at the conventional N application rate could improve N accumulation. Nonetheless, the absorbed extra N tended to accumulate in rice straw, especially when the grain yield was low (in 2021). There were no significant differences in the grain N uptakes between the MN60 and MN100 treatments in the grain filling and maturity stages. However, the MN100 treatment resulted in a significantly higher N uptake in rice straw compared with MN60, showing that the higher N application rate only promoted N accumulation in rice straw when CMV was incorporated alone. The SMN100 treatment significantly increased N uptake in the aboveground part during the tillering and shooting stages and in the rice grain in the maturity stage compared with SMN60, indicating that more N fertilizers promoted N uptake in the whole rice plant, including grain, when CMV and rice straw were incorporated together.

3.3. Soil Nitrogen Content

In 2020, there were no significant differences in soil MBN contents among all the treatments before the maturity stage (

Table 4. Soil microbial biomass nitrogen content (mg kg⁻¹) during different stages in response to the incorporation of Chinese milk vetch (M) and rice straw (S) at different N application rates.

Year	Treatment	Before Transplantation	Tillering Stage	Shooting Stage	Maturity Stage
2020	N0	19.4 ± 4.1 a	19.3 ± 0.8 a	20.7 ± 4.2 a	12.7 ± 0.1 d
	MN0	18.5 ± 4.5 a	39.9 ± 10.1 a	26.3 ± 1.9 a	20.3 ± 0.7 c
	MN60	33.8 ± 4.5 a	40.9 ± 7.6 a	40.8 ± 2.7 a	27.9 ± 4.4 b
	SMN60	35.3 ± 4.7 a	26.8 ± 3.2 a	43.5 ± 7.8 a	34.6 ± 1.1 a
	N100	22.8 ± 4.1 a	20.9 ± 0.5 a	28.0 ± 1.1 a	21.6 ± 2.4 bc
	MN100	34.8 ± 5.1 a	32.8 ± 2.3 a	33.1 ± 8.0 a	26.5 ± 1.8 bc
	SMN100	28.3 ± 3.9 a	32.4 ± 2.3 a	33.0 ± 4.7 a	20.6 ± 1.3 c
2021	N0	44.7 ± 0.8 a	36.9 ± 11.4 c	41.7 ± 5.2 b	14.0 ± 3.1 c
	MN0	47.4 ± 4.4 a	65.6 ± 11.7 ab	61.2 ± 4.5 a	24.9 ± 5.1 bc
	MN60	40.0 ± 1.7 a	44.6 ± 3.2 bc	35.3 ± 0.7 b	32.2 ± 6.2 ab
	SMN60	46.5 ± 4.9 a	34.8 ± 3.2 c	52.9 ± 7.0 ab	35.3 ± 1.2 ab
	N100	40.0 ± 5.1 a	48.9 ± 3.7 abc	61.0 ± 2.0 a	32.0 ± 7.5 ab
	MN100	45.7 ± 7.6 a	61.5 ± 3.5 ab	65.0 ± 1.9 a	46.5 ± 3.8 a
	SMN100	53.5 ± 10.9 a	68.7 ± 4.3 a	68.8 ± 11.6 a	30.2 ± 1.1 b

Data in the table are mean ± standard error and different lowercase letters in each column indicate significant difference at the p < 0.05 level.

). Compared with the N0 treatment, MN0 significantly increased the MBN by 59.8% in the maturity stage. No significant differences were found in the MBN for the MN60 and N100 treatments, but the SMN60 treatment significantly increased the MBN by 60.2% and 24.0% compared with N100 and MN60. There were no significant differences in MBN for the three treatments with 100% of the conventional N application rate. In 2021, the MBN values under MN0 were significantly higher than those under N0 in both tillering and shooting stages. Throughout the whole growing season, the MBN values in soil with SMN60 or SMN100 were not significantly different from those with N100. These results indicate that the incorporation of CMV and rice straw maintained a high soil N supply capacity by improving MBN content even when using 60% of the conventional N application rate, but when the conventional rate was applied, the extra N fertilizers were not reserved in the soil in the form of MBN.

There were no significant differences in soil NH₄⁺-N contents among all the treatments before rice transplantation and in the tillering stages of the two growing seasons (

Table 5. Soil ammonium–nitrogen content (mg kg⁻¹) in different stages in response to the incorporation of Chinese milk vetch (M) and rice straw (S) at different N application rates.

Year	Treatment	Before Transplantation	Tillering Stage	Shooting Stage	Maturity Stage
2020	N0	4.4 ± 0.9 a	5.1 ± 0.8 a	4.4 ± 0.9 c	3.8 ± 0.6 d
	MN0	5.1 ± 0.6 a	6.7 ± 1.1 a	7.7 ± 0.5 bc	6.2 ± 0.9 cd
	MN60	3.1 ± 0.5 a	4.7 ± 0.7 a	12.0 ± 1.1 a	10.2 ± 0.6 ab
	SMN60	3.7 ± 0.2 a	5.8 ± 0.3 a	10.3 ± 1.6 ab	8.0 ± 1.1 bc
	N100	3.3 ± 0.1 a	4.6 ± 0.8 a	8.2 ± 1.4 b	6.6 ± 0.9 cd
	MN100	4.3 ± 0.7 a	5.1 ± 0.4 a	10.9 ± 1.5 ab	8.5 ± 1.8 bc
	SMN100	3.5 ± 0.1 a	5.5 ± 0.1 a	10.5 ± 0.4 ab	12.8 ± 0.4 a
2021	N0	3.8 ± 0.6 a	9.2 ± 2.5 a	9.0 ± 0.3 a	4.4 ± 0.7 bc
	MN0	2.8 ± 0.4 a	11.4 ± 1.2 a	10.6 ± 0.4 a	4.9 ± 0.5 abc
	MN60	3.2 ± 0.9 a	8.9 ± 0.7 a	12.9 ± 0.4 a	4.1 ± 0.2 c
	SMN60	2.8 ± 0.5 a	9.2 ± 1.8 a	11.6 ± 1.9 a	6.1 ± 0.4 a
	N100	2.9 ± 0.4 a	9.7 ± 1.8 a	6.9 ± 2.1 a	5.7 ± 0.5 ab
	MN100	2.4 ± 0.9 a	10.8 ± 1.5 a	9.1 ± 1.4 a	5.6 ± 0.0 abc
	SMN100	2.4 ± 0.2 a	13.2 ± 1.1 a	10.3 ± 1.2 a	6.3 ± 0.8 a

Data in the table are mean ± standard error and different lowercase letters in each column indicate significant difference at the p < 0.05 level.

). In 2020, the NH₄⁺-N content under the MN60 treatment was significantly higher than that under MN0 and N100 in the shooting and maturity stages, but not significantly different from that under SMN60 and MN100. In the maturity stage, the SMN100 treatment significantly increased soil NH₄⁺-N content by 93.9% and 50.6% compared with the N100 and MN100 treatments, respectively. No significant differences were found in the NH₄⁺-N contents between the two treatments with the incorporation of CMV and rice straw in the shooting stage. However, in the maturity stage, the NH₄⁺-N content under SMN100 was significantly higher than that under SMN60. In the maturity stage in 2021, the NH₄⁺-N content in soil with SMN60 was not significantly different from that with N100 and SMN100, but was significantly higher than that with MN60 and N0. These results indicate that the SMN60 and SMN100 treatments could maintain higher soil NH₄⁺-N contents, especially in the season with a higher rainfall accumulation (in 2020, Figure 1). However, the soil NH₄⁺-N content under the SMN100 treatment tended to be surplus.

There were no significant differences in soil NO₃⁻-N contents among all the treatments throughout the rice growing season in 2020 (

Table 6. Soil nitrate–nitrogen content (mg kg⁻¹) during different stages in response to the incorporation of Chinese milk vetch (M) and rice straw (S) at different N application rates.

Year	Treatment	Before Transplantation	Tillering Stage	Shooting Stage	Maturity Stage
2020	N0	4.5 ± 0.4 a	5.7 ± 0.9 a	9.0 ± 2.3 a	2.4 ± 0.2 a
	MN0	4.2 ± 0.7 a	6.6 ± 0.5 a	7.5 ± 1.1 a	2.5 ± 0.1 a
	MN60	4.4 ± 0.6 a	6.5 ± 0.9 a	6.4 ± 1.2 a	2.7 ± 0.1 a
	SMN60	5.0 ± 0.4 a	5.1 ± 0.4 a	11.8 ± 0.3 a	3.1 ± 0.1 a
	N100	4.1 ± 0.7 a	4.3 ± 1.0 a	6.0 ± 0.3 a	3.1 ± 0.1 a
	MN100	4.5 ± 0.5 a	5.6 ± 0.8 a	6.5 ± 1.7 a	3.0 ± 0.5 a
	SMN100	4.0 ± 0.1 a	5.5 ± 1.1 a	6.4 ± 1.4 a	3.0 ± 0.2 a
2021	N0	2.3 ± 0.0 b	3.9 ± 0.2 a	5.9 ± 1.3 c	4.8 ± 1.9 a
	MN0	7.3 ± 0.5 a	3.9 ± 1.8 a	5.9 ± 1.9 c	4.5 ± 0.6 a
	MN60	3.3 ± 0.5 b	5.6 ± 1.6 a	8.1 ± 0.9 bc	6.6 ± 1.8 a
	SMN60	3.8 ± 0.4 b	4.7 ± 0.2 a	11.9 ± 1.6 ab	5.8 ± 1.0 a
	N100	3.5 ± 0.7 b	4.0 ± 0.5 a	8.8 ± 2.4 bc	4.5 ± 0.3 a
	MN100	3.7 ± 0.2 b	5.0 ± 0.7 a	4.0 ± 1.2 c	7.2 ± 1.5 a
	SMN100	7.6 ± 0.5 a	3.7 ± 1.0 a	14.9 ± 1.0 a	5.6 ± 1.2 a

Data in the table are mean ± standard error and different lowercase letters in each column indicate significant difference at the p < 0.05 level.

). In the shooting stage in 2021, the NO₃⁻-N contents under MN60, SMN60, and N100 treatments were not significantly different. Compared with the N100 treatment, SMN100 significantly increased the NO₃⁻-N content, but not compared to the MN100 treatment. This indicates that the conventional N application rate with the incorporation of CMV and rice straw increased the risk of NO₃⁻-N loss.

3.4. NH₄⁺-N and NO₃⁻-N in Surface Water

The changes in NH₄⁺-N concentrations in surface water were consistent over the two seasons (Figure 3). Compared with N100, the treatments with 60% of conventional N application rate significantly reduced the NH₄⁺-N concentrations by 24.4–47.3% and 66.3–85.1% at 2–3 and 15–16 DAT, respectively. The MN100 and SMN100 treatments significantly increased the NH₄⁺-N concentrations by 26.3% and 72.4%, respectively, compared with the N100 treatment at 2 DAT in 2021.

In 2020, the NO₃⁻-N concentrations in surface water were significantly higher under the SMN100 treatment than those under N100 and SMN60 at 3 and 6 DAT. At 16 DAT, the SMN60 treatment significantly decreased the NO₃⁻-N concentration by 59.2% compared with N100. In 2021, the SMN100 treatment also resulted in a significantly higher NO₃⁻-N concentration compared with N100 and SMN60 at 8 DAT. However, at 2 and 15 DAT, there were no significant differences in the NO₃⁻-N concentrations among the treatments with 0% and 100% of the conventional N application rate.

4. Discussion

The present study found that the treatments incorporating CMV alone or combined with rice straw and with 60% of N fertilizer did not reduce rice grain yield compared with N100. This is consistent with the previous results, which showed that planting CMV could replace 20–40% of N fertilizer in the paddy fields in southern China [3,19]. The CMV, as a type of leguminous green manure, released a large amount of N after incorporation, which met the N demand of rice in the early growth stage and thus replaced the partial N fertilizer [20,21]. In addition, the C sources provided by the fresh organic residues increased the biomass and activity of soil microorganisms, leading to a positive priming effect and organic N mineralization [22]. When organic materials were incorporated, the treatments with the conventional N application rate did not significantly increase rice grain yield compared with the N-reduced treatments, but significantly increased the N uptake in the rice plant.

The results show that the plant N uptake increased with increasing N application rate. However, the N uptakes under the treatments with a 40% reduction in N fertilizer (MN60 and SMN60) were not significantly different from that under N100. This confirms that the incorporation of CMV alone or in combination with rice straw could replace 40% of N fertilizers. Comparing with CMV alone, the incorporation of both CMV and rice straw with chemical fertilizers could harmonize the supply and demand of N nutrient, promoting rice growth [20,23]. The incorporation of CMV and rice straw at the conventional N application rate promoted the N uptake in rice straw and grain, but did not significantly improve rice grain yield. Thus, such a high amount of N accumulation in rice plants could be considered as surplus uptake. Abe et al. [24] found that excessive N fertilizer application even reduced rice grain yield. Therefore, a proper application rate of N fertilizer is crucial for improving rice grain yield and minimizing the waste of resources and environmental risk.

In the tillering and shooting stages, the incorporation of CMV and rice straw with a 40% reduction in N fertilizer did not significantly reduce the soil inorganic N content compared with the N100 treatment. The input of nutrients from CMV and rice straw increased the soil organic N pool and promoted the mineralization of organic matter, which increased N release and thus the inorganic N concentration [25]. In addition, the incorporation of CMV and rice straw optimized the C/N ratio of the inputs [26], increasing the activity of soil microorganisms and the MBN content [27].

Inorganic N in the surface water can be directly absorbed by rice and represent the primary source of N loss from paddy field [28]. When N was applied at 60% of the conventional rate, the NH₄⁺-N concentrations in the surface water with CMV alone and in combination with rice straw were lower than that under N100. This indicates that the 40% N fertilizer replacement reduced the risk of N loss, especially at 2–3 days after N fertilizer application. The changes in NO₃⁻-N concentration in the surface water generally lagged behind changes in NH₄⁺-N [29]. The first peak of NO₃⁻-N concentration occurred at 2 and 8 DAT in 2020 and 2021, respectively, which might be related to the sampling time. The first sampling date after initial fertilizer application in 2021 was one day earlier than that in 2020, probably resulting in the missed peak of NO₃⁻-N concentration. Chu et al. [30] and Ai et al. [31] showed that an increased N application rate could enhance nitrification and soil NO₃⁻-N content. Therefore, the high N input is one of the key factors causing N loss from the paddy field. In the present study, especially in 2020, a 40% N fertilizer replacement reduced the NO₃⁻-N concentrations in the surface water, thus reducing the potential for N loss through leaching and runoff.

5. Conclusions

Based on the rice grain yield and N uptake in rice plant, both the incorporation of CMV alone and its combination with rice straw could be used to replace 40% of conventional N fertilizer, but the latter maintained a higher level of soil N supply capacity. Moreover, the conventional N application rate with the incorporation of crop residues could result in a luxury N uptake and a high risk of N loss. Consequently, the incorporation of both CMV and rice straw with a N application rate at 60% of the conventional rate would be sufficient for use in the mid-season rice–green manure rotation system.