Increased Soil Soluble Nitrogen Stocks and Decreased Nitrogen Leaching Loss in Rice Paddy Soil by Replacing Nitrogen Fertilizer with Chinese Milk Vetch

Abstract

Reducing soil nitrogen leaching losses and improving nitrogen-use efficiency with effective fertilization management strategies are extremely important for sustainable agricultural development. A 2-year field study was executed with the same nitrogen input in a subtropical rice production system in Southeast China, using chemical fertilizers as a control (CK), to study the influences of different application amounts of Chinese milk vetch (CMV), i.e., 15,000 kg hm⁻² (CL), 30,000 kg hm⁻² (CM), and 45,000 kg hm⁻² (CH), on soil soluble nitrogen stock and leaching risks in a clay paddy field. The results showed that the soil stocks of soluble inorganic nitrogen (SIN) and soluble organic nitrogen (SON) in a 0–60 cm soil profile under different application amounts of CMV significantly increased by 12.43–36.03% and 19.43–71.75% compared with CK, respectively, which was more favorable to soil SON accumulation. In the 2-year experiment, the total dissolved nitrogen leaching loss was 23.51–61.88 kg hm⁻² under different application rates of CMV, of which 50.08–62.69% was leached by dissolved inorganic nitrogen (DIN), and 37.31–49.92% was leached by dissolved organic nitrogen (DON). CMV application improved soil properties (pH, SOM, and urease/protease), increased SIN and SON stocks, and decreased surface water DIN and DON concentrations, thereby reducing DIN and DON leaching. The leachings of DIN and DON in different application rates of CMV were reduced by 11.37–66.23% and 13.39–52.07% compared with the CK treatment, respectively. Conclusively, nitrogen leaching loss in paddy fields was severe, and the DIN and DON leaching loss in CMV treatments were lower than those in the control under the same nitrogen input. Thus, replacing nitrogen fertilizer with CMV under the same nitrogen input could reduce the risk of nitrogen nonpoint pollution in clay paddy fields.

1. Introduction

Rice is one of the crucial food crops of the world, and its yield is heavily dependent on nitrogen (N) inputs [1,2]. The N fertilizer consumption by rice crops accounted for 21–25% of the world’s total N fertilizer application. However, the utilization efficiency of N fertilizer in flooded rice is very low, usually between 20–35% [3], mainly caused by ammonia volatilization, runoff, and N leaching. Rice plants with shallow roots limit their absorption of soil N, thus increasing the risk of N loss with gravity water in the paddy field. Ross et al. [4] showed that 5–50% of applied N is lost through leaching in N losses, which is an important way of N losses from paddy soil, resulting in lower N use efficiency and also harm to the environmental quality of the surrounding water bodies [5]. Therefore, it is important to reveal the dynamic changes of various N form concentrations and their leaching characteristics in paddy soil for controlling soil N leaching, improving N fertilizer utilization, and preventing water pollution.

N leaching is one of the main causes of water-quality deterioration [6], and the amount and form of N leached from the paddy soils were closely related to fertilization [7]. Generally, higher N fertilizers application would lead to higher N leaching [8]. For urea, as a conventional chemical N fertilizer, excessive application is an important factor leading to N leaching in paddy fields [9]. Compared with traditional N fertilizer, slow-control fertilizer and urea inhibitors can reduce N leaching by 35% and 30%, respectively [10]. Organic materials, i.e., straw [11], manure [12], sewage sludge compost [13], etc., could potentially be used as an alternative to N fertilizer and reduce nitrogen output through runoff and leaching. The effect of organic-materials application on soil N leaching was related to its carbon–nitrogen ratio (C/N). Organic-materials application with a low C/N ratio increased the risk of N leaching, while a high C/N ratio was the opposite [14,15]. Therefore, fertilization could significantly affect soil N leaching, but the composition and chemical properties of different N fertilizers were significantly different, and the effects on soil N leaching were also different. Chinese milk vetch (CMV), a leguminous green manure crop, has been widely cultivated in South China for centuries [16]. The N concentration of CMV at the full flowering stage was up to 28–37 g kg⁻¹, which could supply the soil with available N and reduce the input of chemical fertilizer. Economically and environmentally, CMV can be planted in paddy fields at a lower cost and produces less ammonia volatilization than traditional fertilization [17]. Currently, studies on the application of CMV mainly focus on the impact of CMV on crop yield, soil fertility, soil microorganisms, ammonia volatilization, etc. [18]. However, the role of CMV application on soil soluble N leaching and the environmental risks are not well known at the moment.

The N is dissolved in soil water, including dissolved inorganic nitrogen (DIN) and dissolved organic nitrogen (DON), which is collected in situ using devices such as a rhizon, lysimeter, or suction cup [19,20]. Numerous studies have shown that N leaching is the main pathway for N loss in agroecosystems [21,22]. However, most of these studies focused on the leaching of single N components from DIN or DON [23,24]. Few studies have addressed the speciation and composition of the leached N. By contrast, N in the soil that can be extracted by a water or salt solution is called soluble N, including soluble inorganic N (SIN, including NH₄⁺-N, NO₃⁻-N, and NO₂⁻-N) and soluble organic N (SON), which is one of the most active components of soil N pools [19]. Soluble N and dissolved N in the soil are assumed to be in equilibrium and influenced by adsorption–desorption [25]. Recent studies have suggested that soil adsorption of NO₃⁻ could decrease NO₃⁻ leaching [10,26,27]. However, it is not clear whether CMV application has the same effect on the accumulation of different N forms in the soil, and the leaching and influencing mechanisms of N accumulation on different N forms remain to be explored.

In this study, a field experiment with four different treatments (chemical fertilizer, replacing N fertilizer with a low amount of CMV, replacing N fertilizer with a medium amount of CMV, and replacing N fertilizer with a high amount of CMV) was conducted for two consecutive years under the same N, phosphorus, and potassium input base on the common application rates of CMV in the field in subtropical paddy soil in Southeast China. We hypothesized that (i) the application rate of CMV could have different effects on soil SIN and SON reserves; (ii) CMV application could reduce the leaching risks of N in paddy soil, but the effect on DON and DIN may be different, and (iii) CMV application could affect DIN and DON leaching in paddy soil by changing soil SIN and SON stocks. The results of this study would reveal insights for reducing N leaching and rational fertilization in subtropical paddy soils.

2. Materials and Methods

2.1. Experimental Site and Design

A rice field experiment was established in Baisha town, Fujian Province (119°04′12″ E, 26°13′33″ N), which belongs to a subtropical monsoon climate with an annual mean temperature of 19.5 °C, annual sunshine hours of 1812.5 h, a frost-free period of 311 days, and an annual mean precipitation of 1351 mm. The observation data of daily mean precipitation and temperature in the experimental area during the rice growing period in the research years are shown in Figure 1. The soil used in our study was gray–yellow paddy soil widely distributed in Fujian province. The characteristics of the 0–60 cm soils were as follows: texture, loamy clay; bulk density, 1.52 g cm⁻³; pH value, 5.93; organic matter, 12.04 g kg⁻¹; total nitrogen, 0.61 g kg⁻¹; available phosphorus, 5.38 mg kg⁻¹; and available potassium, 24.53 mg kg⁻¹.

In this experiment, based on the common CMV application rates in the field, four treatments with the same N input were set up in a random group arrangement: CK (chemical fertilizer), CL (replacing N fertilizer with 15,000 kg hm⁻² CMV), CM ((replacing N fertilizer with 30,000 kg hm⁻² CMV), and CH (replacing N fertilizer with 45,000 kg hm⁻² CMV). Three replicate experimental blocks (12 m²) were designed for each treatment (Figure S1). In this experiment, urea, superphosphate, and potassium chloride were applied in CK treatment, and the application rates were 481.67 kg hm⁻², 900 kg hm^−2, and 300 kg hm⁻², respectively. Among them, 100% superphosphate, 50% urea, and 50% potassium chloride were applied on the 9th day of CMV application (base fertilizer), and 50% urea and 50% potassium chloride were applied on the 25th day of CMV application (tillering fertilizer). The CMV variety tested was Minzi No. 7, and its carbon, nitrogen, phosphorus, and potassium contents were 436.63 g kg⁻¹, 30.94 g kg⁻¹, 5.91 g kg⁻¹, and 32.47 g kg⁻¹, respectively; acid-hydrolyzed amino acid, protein, cellulose, hemicellulose, and lignin contents were 82.35 mg kg⁻¹, 193.40 g kg⁻¹, 130.82 g kg⁻¹, 111.76 g kg⁻¹, and 40.81 g kg⁻¹ (dry-weight basis). Fresh CMV (90% water content) was harvested at the full flowering stage (mid-April) and immediately dispersed evenly in the corresponding plots according to the application rates of each treatment (0, 15,000, 30,000, and 45,000 kg hm⁻²) and turned them over into the topsoil. The insufficient part of nitrogen, phosphorus, and potassium in each CMV treatment was supplemented with chemical fertilizer based on the above experimental design. Irrigation was carried out before rice planting, and a 5 cm flooded layer was maintained.

2.2. Soil and Soil–Water Sampling

A stainless-steel soil drill (5 cm diameter and 100 cm depth) was used to collect soil cores (0–60 cm) from each plot at 0 d (background soil) and 122 d (rice maturity stage) after CMV application using a multi-point sampling method. After pre-processing the collected soil samples (mixed, removed plant roots, animal and plant residues, stones, etc.), the background soil was used for the analysis of the related soil physical and chemical properties, and fresh soil samples at 122 d after CMV application were directly used to determine the soil chemical properties, SIN, and SON stocks.

Sand filter tubes (30 μm, 3.8 cm inner diameter, and 20 cm height) connected with covered stainless steel tubes were buried in each plot for in situ sampling of soil percolating water at 0–20 cm, 20–40 cm, and 40–60 cm before the application of CMV (Figure S2). Two sets of sand filter tubes were buried in opposite directions, and soil percolating water was mixed in each plot. Surface water (0 cm) and soil percolating water were sampled with an electric water loader at 5 d, 10 d, 10 d, 17 d, 24 d, 38 d, 59 d, 80 d, 101 d, and 122 d, respectively, according to the rice growth stage and the decomposition rate of CMV.

2.3. Soil and Soil Water Analysis

2.3.1. Analysis of Soil Basic Properties

The soil bulk density was measured by the ring-knife method [28]. Soil pH (soil–water ratio 1:2.5) was determined by a pH meter (PHS-3E, INESA, Shanghai, China). Soil organic matter (SOM) was measured by the potassium dichromate oxidation method described by Lu (2000) [29]. An elemental analyzer (LECO, TruMac, San Joseph, MI, USA) was used to analyze soil total N. Available phosphorus and available potassium were determined by the 0.5 M NaHCO₃ extraction colorimetric method [29] and the 1 M NH₄OAc extraction flame photometric method [29], respectively. Soil urease activities and protease activities were estimated by indophenol blue colorimetry and Folin colorimetry described by Guan (1986) [30].

2.3.2. Analysis of Soil SIN and SON

Soil SIN and SON concentrations were determined according to the method by Chen et al. [31] as follows: the soil and distilled water were mixed at the soil-water ratio of 1:5, heated in a 70 °C thermostat for 18 h, shaken for 5 min, and finally filtered through a 0.45 μm filter. A total organic carbon analyzer (TOC-L, Shimadzu, Japan) and a continuous flow analyzer (SAN++, Skalar, Breda, The Netherlands) were used to determine the TSN and SIN in the filtrate, respectively. The SON content in the filtrate was obtained as the difference between the TSN and SIN.

2.3.3. Analysis of DIN and DON in Soil Water

The concentrations of DIN and DON in soil water samples were directly determined by 0.45 μm microporous membrane filtration. The DIN and DON in soil water were determined by the same method as SIN and SON in the soil-sample extraction solution.

2.4. Statistical Analysis

All data statistical analyses were performed using Excel 2003. Correlation analyses were carried out by SPSS 19.0, and data plots were made with Origin 2022. The significance of differences among treatments was performed using analysis of variance (ANOVA). The driving factors and influencing paths of DIN and DON leaching were studied using a structural equation model (SEM) in AMOS 21.0.

The SIN or SON stocks of a profile were calculated using Equation (1) by Pan et al. [32]:

$$\left(S_{SIN} \mathrm{or} S_{SON}\right)=\left(C_{SIN} \mathrm{or} C_{SON}\right)\times B_s\times S_h÷10$$

(1)

where S_SIN or S_SON is the stocks of SIN or SON (kg hm⁻²), C_SIN or C_SON is the content of SIN or SON (mg kg⁻¹), Bs is the bulk density (g cm⁻³), and S_h is the soil thickness (cm).

It is generally considered that the amount of N leaching by soil water below the active root layer is called leaching loss. Because it is difficult for rice to absorb and utilize N below the 20 cm soil layer, the N concentration in the leachate of the 20–40 cm and 40–60 cm soil layers can be regarded as the amount of N leaching in paddy fields. The DIN or DON leaching loss was calculated using Equation (2) by Nie et al. [23]:

$$\begin{array}{l}\left({\mathrm{L}}_{\mathrm{DIN}} \mathrm{or} {\mathrm{L}}_{\mathrm{DON}}\right)={\displaystyle {\sum }_0^{122}({\mathrm{C}}_{\mathrm{DIN}} \mathrm{or} {\mathrm{C}}_{\mathrm{DON}})}\times {\mathrm{V}}_{\mathrm{water}}÷1000\\ ={\displaystyle {\sum }_0^{122}({\mathrm{C}}_{\mathrm{DIN}} \mathrm{or} {\mathrm{C}}_{\mathrm{DON}})}\times \frac{{\mathrm{B}}_{\mathrm{S}}\times \mathrm{S}\times {\mathrm{S}}_{\mathrm{h}}\times {\mathrm{M}}_{\mathrm{w}}}{{\mathrm{W}}_{\mathsf{\rho }}}÷10,000\end{array}$$

(2)

where L_DIN or L_DON is the leaching losses of DIN or DON (kg hm⁻²), C_DIN and C_DON are the content of DIN and DON in the soil water (mg L⁻¹), V_water is the volume of the soil water (m³), Bs is the soil bulk density (g cm⁻³), S is the area, S_h is the thickness of the soil layer (cm), M_w is soil water content, and W_ρ is soil water density (g cm⁻³).

3. Results

3.1. Soil Properties in Loamy Clay Paddy Soil under Different Treatments

CMV substitution for partial N fertilizer significantly increased soil pH, SOM contents, urease activity, and protease activity at the maturity stage compared to CK treatment (Figure 2). In the two experimental years, the average pH values of the soil in the CL, CM, and CH treatments were 3.53%, 4.45%, and 6.26% higher than that in the CK treatment, respectively (p < 0.05); the average SOM contents of soil in the CL, CM, and CH treatments were 5.40%, 11.51% and 15.65% higher than that in CK treatment, respectively (p < 0.05); the average urease activities of soil in the CL, CM, and CH treatments were 15.76%, 37.04%, and 33.52% higher than that in the CK treatment, respectively (p < 0.05); and the average protease activities of soil in the CL, CM, and CH treatments were 22.89%, 42.55%, and 50.03% higher than that in the CK treatment, respectively (p < 0.05). Compared with the first year, the urease activity of different treatments in the second year significantly increased by 21.56–36.03%, and protease significantly increased by 35.70% and 18.86% only in CK and CL treatments. The soil pH value and SOM contents were not significantly increased between the two experimental years.

3.2. SIN and SON Stocks in Loamy Clay Paddy Soil under Different Treatments

In the 2-year experiment, SIN accounted for 38.36–45.94% of TSN, and SON accounted for 54.06–61.64% of TSN (Figure 3). Compared with the first year, SIN stocks in CM treatment in the second year were significantly increased by 18.50%, and SON stocks in the CK, CL, and CM treatments were increased by 12.96%, 29.34%, and 31.47%, respectively. In the first year, the stocks of SIN and SON under different fertilization treatments showed the tendency of CH > CM ≈ CL > CK. The SIN stocks in the CL, CM, and CH treatments were significantly increased by 15.34%, 21.46%, and 36.03% compared with the CK treatment, while the SON stocks were significantly increased by 19.43%, 34.17%, and 71.75% (p < 0.05), respectively. In the second year, the SIN and SON stocks under different fertilization treatments showed the tendency of CH ≈ CM > CL > CK. Compared with CK, the SIN stocks in the CL, CM, and CH treatments were significantly increased by 12.43%, 40.30%, and 27.65%, while the SON stocks were significantly increased by 36.75%, 56.17%, and 58.32%, respectively. CMV substitution increased the proportion of SON in the soil soluble N pool, indicating that CMV was more conducive to the improvement of SON content.

3.3. DIN and DON Concentrations in Soil Water at Different Depths

3.3.1. DIN Concentration in Soil Water

The temporal variation of the DIN concentration in soil water at different depths showed similar trends under different treatments (Figure 4), with the concentration increasing from background to peak at 17 d after CMV application, rapidly decreasing to 24 d, increasing gradually to a second peak at 38 d, and then dropping to close to zero. In general, the DIN concentrations in soil water decreased with depth, especially at 17 d. In the 2-year experiment, the DIN mean concentration in the surface water at 17 d was significantly increased by 0.80–7.14 mg L⁻¹, 1.09–8.76 mg L⁻¹, and 2.06–10.36 mg L⁻¹ compared with the 0–20 cm, 20–40 cm, and 40–60 cm layers, but there was no significant difference in the DON concentration among the 0–20 cm, 20–40 cm, and 40–60 cm percolation water. CMV substitution for partial N fertilizer decreased the DIN concentration in the soil water, and the DIN mean concentrations in the CL, CM, and CH treatments were significantly decreased by 28.75%, 54.38%, and 71.52% in surface water, decreased by 19.25%, 38.29%, and 52.41% in 0–20 cm, decreased by 11.27%, 15.22%, and 25.92% in 20–40 cm, and decreased by 7.92%, 10.10%, and 18.92% in 40–60 cm during the rice growth period in the first year, respectively. The DIN mean concentrations in the CL, CM, and CH treatments were significantly decreased by 23.89%, 43.67%, and 53.23% in surface water, decreased by 16.22%, 39.44%, and 53.13% in 0–20 cm, decreased by 27.08%, 55.92%, and 60.57% in 20–40 cm, decreased by 24.13%, 51.65%, and 72.61% in 40–60 cm during the rice growth period in the second year, respectively. Overall, compared with the first experiment year, the DIN mean concentrations at different layers under CK and CL treatments in the second year were significantly increased by 26.80% and 22.16%, while the DIN mean concentrations under CM and CH treatments had no significant difference between the two years.

3.3.2. DON Concentration in Soil Water

The temporal variation of DON concentration varied with different depths (Figure 5). The DON concentration in the surface water under different fertilization treatments increased rapidly to the peak at 10 d after CMV application and then rapidly decreased and gradually approached zero after 59 d. The DON concentration in 0–20 cm percolation water reached the peak at 17 d and 38 d after CMV application, while the DON concentration in 20–40 cm and 40–60 cm percolation water increased to the peak at 24 d after CMV application. In the 2-year experiment, the DON mean concentration during the rice growth period in different percolation layers was shown as surface water >0–20 cm > 20–40 cm ≈ 40–60, and that in surface water was significantly increased by 0.48–15.58 mg L⁻¹, 1.36–17.25 mg L⁻¹, and 1.60–17.50 mg L⁻¹ compared with those in 0–20 cm, 20–40 cm, and 40–60 cm percolation water under different fertilization treatments, but there was no significant difference between the DON concentrations in 20–40 cm and 40–60 cm percolation water. The CMV substitution for partial N fertilizer decreased the DON concentration in the soil water, and the DON mean concentrations in the CL, CM, and CH treatments were significantly decreased by 34.69%, 64.81%, and 88.89% in the surface water; decreased by 26.91%, 34.29%, and 38.31% in 0–20 cm; decreased by 8.61%, 20.33% and 19.52% in 20–40 cm; and decreased by 10.23%, 17.01%, and 27.48% in 40–60 cm during the rice growth period in the first year, respectively. The DON mean concentrations in the CL, CM, and CH treatments were significantly decreased by 15.12%, 29.43%, and 75.15% in the surface water; decreased by 11.01%, 35.00%, and 50.90% in 0–20 cm; decreased by 31.69%, 47.35%, and 59.64% in 20–40 cm; and decreased by 23.55%, 33.99%, and 48.87% in 40–60 cm during the rice growth period in the second year, respectively. Overall, compared with the first experiment year, the DON mean concentration at different layers of CK treatments in the second year was significantly reduced by 28.42%, while the DON mean concentrations of CL, CM, and CH treatments had no significant differences between the two years.

3.4. Potential Leaching Loss of DIN and DON in Loamy Clay Paddy Soil under Different Treatments

3.4.1. Leaching Rates of DIN and DON in Loamy Clay Paddy Soil

The dynamics of DIN and DON leaching loss in paddy soil were fitted by unary linear, polynomial, and logistic models, respectively (Table S2). The results showed that the R² fitted by the logistic model was closest to one, which indicates that the logistic model had the best fitting effect on the leaching dynamics of DIN and DON in the tested soil. Logistic model fitting results showed that the application of CMV significantly reduced the leaching rate of DIN and DON (Figure 6). Compared with CK treatment, the maximum leaching rate of DIN was significantly reduced by 15.85–55.45%, and the maximum rate of DON leaching was significantly reduced by 16.41–65.75% in the experimental years. The maximum rates of DIN leaching in the second year under CK, CL, CM, and CH treatments significantly increased by 32.37%, 37.15%, 91.75%, and 111.20% compared with those in the first year, and the maximum rates of DON leaching significantly increased by 34.72%, 102.31%, 127.86%, and 142.25% compared with those in the first year. The maximum leaching rates of DIN and DON also showed significant differences, and the maximum leaching rates of DIN were significantly higher than those of SON by 91.57–182.57%. In the 2-year experiment, the time corresponding to the maximum leaching rate of DIN and DON under different treatments ranged from 17 d to 29.25 d, but there were no obvious patterns among different treatments.

3.4.2. Leaching Loss of DIN and DON in Loamy Clay Paddy Soil

In the 2-year experiment, DIN accounted for 50.08–62.69% of TDN, and DON accounted for 37.31–49.92% of TDN (Figure 7). Compared with the first year, TDN leaching losses in CK, CL, and CM treatments in the second year were significantly increased by 76.80%, 56.49%, and 27.67%, respectively. DON leaching losses in the CK, CL, CM, and CH treatments in the second year were significantly increased by 37.71%, 69.72%,76.43%, and 88.70%, respectively, while DIN was significantly increased by 69.29% and 44.20% only in CK and CL. In the first year, compared with CK, the leaching losses of TDN in the CL, CM, and CH treatments significantly decreased by 12.15%, 22.03%, and 32.82%; the leaching losses of DIN significantly decreased by 11.37%, 20.50% and 31.86%; and DON leaching losses in the CM and CH treatments were significantly reduced by 24.91% and 34.33%, respectively. But, there was no significant difference in DON leaching loss between the CK and CL treatments. In the second year, the leaching losses of TDN in the CL, CM, and CH treatments significantly decreased by 22.24%, 43.82%, and 60.39%; the leaching losses of DIN significantly decreased by 24.51%, 51.80%, and 66.23%; and DON leachings were significantly decreased by 19.01%, 32.46%, and 52.07%, respectively. CMV substitution for partial N fertilizer could effectively reduce the dissolved N leaching in paddy soil, and the effect on DIN leaching was more significant.

3.5. Influencing Factors of DIN and DON Leaching in Loamy Clay Paddy Soil under Different Treatments

The effect of CMV application on N leaching is consistent with our hypothesis of causal relationships, as showed by the SEM properties (χ²/df < 3, RMSEA < 0.05, NFI > 0.9, CFI > 0.9, Figure 8) and was able to explain 83% of SIN stocks and 90% of DIN leaching (Figure 8a). CMV application showed a direct positive effect on SIN accumulation, with a path coefficient of 0.56, whereas there was no significant direct effect on DIN leaching. On the other hand, CMV application can indirectly influence SIN stocks and DIN leaching by affecting soil properties (pH, SOM, and urease) and DIN concentration in surface water, with path coefficients of 0.25 and −0.25, respectively. Moreover, CMV application can indirectly influence DIN leaching by affecting SIN stocks, with a path coefficient of −0.50. The SEM analysis showed that the effects of CMV application explained 78% of SON stocks and 89% of DON leaching (Figure 8b). The application of CMV had a direct positive effect on SON accumulation (path coefficient = 0.42), but no significant direct effect on DON leaching. On the other hand, CMV application can indirectly influence SON stocks and DON leaching by affecting soil properties (pH, SOM, and protease) and DON concentration in surface water, with path coefficients of 0.29 and −0.48, respectively. Moreover, CMV application can indirectly influence DON leaching by affecting SON stocks, with a path coefficient of −0.27.

4. Discussion

4.1. Effects of CMV Application on SIN and SON Stocks in Loamy Clay Paddy Soil

Soluble N is one of the most active components of soil N pools in paddy fields. Previous studies have shown that CMV application was beneficial to soil N accumulation [33,34]. Similarly, the results of this study also showed that CMV substitution for partial N fertilizer significantly increased SON and SIN stocks in paddy soil. This is probably due to the fact that CMV application could bring a large number of nitrogenous compounds to the soil, which significantly increases the soil SIN and SON stocks. The acid-hydrolyzed amino acid and protein contents of the tested CMV were 82.35 mg kg⁻¹ and 193.4 g kg⁻¹, respectively. These N-containing substances can directly increase the concentration of SIN and SON in the soil. The application of CMV can increase the soil organic matter content (Figure 2), improve soil adsorption of SIN and SON [35], and, thus, indirectly increase SIN and SON stocks. On the other hand, the SEM results in our study showed that CMV application could indirectly affect SIN and SON stocks through urease and protease activities (Figure 8). Tang et al. [36] reported that the application of green manure could not only release enzymes into the soil but also provide nutrients for soil microorganisms, thus significantly improving soil-enzyme activities. Similarly, in our study, soil urease and protease activity increased significantly after CMV application (Figure 2), which promoted macromolecular organic N decomposition to produce SIN and SON [37], thus increasing soil SIN and SON stocks.

The results of this study indicated that CMV substitution had different effects on SON and SIN stocks, which was more conducive to the accumulation of SON in soil. As legume green fertilizer, CMV increased the number and activity of soil microorganisms after being applied to soil, and benefited the synthesis of organic N through assimilation, thus increasing soil SON stocks [38]. On the other hand, studies have shown that rice, as an NH₄⁺ loving crop, when abundant inorganic N exists, will reduce SON uptake, resulting in SON accumulation [39]. Furthermore, the adsorption of N by the soil also results in the different accumulations of N in the soil. Gu et al. [40] showed that soil had a strong adsorption capacity for larger molecular weight components in the N pool. The large molecular weight organic N released after CMV application was preferentially absorbed by the loamy clay paddy soil, thus increasing SON stocks.

4.2. Effects of CMV Application on DIN and DON Concentrations in Soil Water

DIN and DON concentrations in soil water declined with an increased soil depth under different fertilization treatments (Figure 4 and Figure 5). Similar findings were reported by Nie et al. [23], who found that the concentration of DON in the upper layer was significantly higher than that in the deeper layer after long-term fertilizer application. Several mechanisms could explain the decrease in N leaching with increasing depth, including N application, root exudates, and N storage via abiotic/biotic uptake. The fertilizer and crop residues applied in successive years are mainly accumulated in the surface soil, and the enzyme substrate is sufficient to provide abundant energy for surface-soil microorganisms, promote the decomposition of organic matter, and improve the concentration of DIN and DON in surface water [37]. Moreover, CMV application is beneficial for rice growth and promotes root secretion production, which is an important source of DIN and DON [25]. Another possible reason may be the absorption capacity of soil for N, which affects the mobility of N in the soil [41]. The soil tested in this study was loamy clay with rich surface charges, showing strong absorption of DIN and DON, leading to a relatively weak penetration of DIN and DON. Furthermore, the plow pan of paddy soil has an intercept effect on soil water penetration [23], so only part of the DIN and DON leached to the bottom.

Dissolved N has strong mobility and is easily leached down vertically with soil water [42]. The results showed that the peak time of DIN concentration at different depths was consistent (Figure 4), while DON concentration peaked at different depths with a delay (Figure 5), indicating different migration rates of DIN and DON in paddy soil. This may be due to the different adsorptions of different N forms in soil. DON usually consists of a series of compounds ranging from high molecular weight nitrogenous compounds (tannins, humus, polyphenols, etc.) to low molecular weight nitrogenous compounds (amides, amino acids, amino sugars, etc.) [43], whose average molecular weight is greater than DIN (NH₄⁺-N and NO₃⁻-N, NO₂⁻-N). The soil had a stronger adsorption capacity for the larger molecular weight components [40]. While the small molecular weight and hydrophilic compounds are not easy to be absorbed by the soil, desorption and replacement adsorption by large molecular weight and hydrophobic compounds tend to occur after adsorption [44,45].

In this study, CMV substitution for partial N fertilizer significantly reduced the concentrations of DIN and DON in percolation water. The urea applied in the CK treatment was low molecular weight DON [46], which may be leached with complex and gradual processes, including adsorption, desorption, and exchange. When the soil-adsorption capacity is saturated, the urea in the overlying water will infiltrate directly into the lower soil under the influence of gravity, increasing the N concentration in percolation water [47]. Moreover, due to the rapid rate of hydrolysis, a large amount of DIN was produced by urea after application, increasing the source of the concentrations of N in percolation water [48]. However, compared with urea, most of the N in CMV exists in the organic form, and the slow release of N is beneficial to be absorbed by the rice, thus preventing N infiltration into the lower soil. The N released by CMV was mostly hydrophobic macromolecular organic N, which was easily retained by the soil to reduce the N concentration in the leachate [49].

4.3. Effects of CMV Application on DIN and DON Leaching in Loamy Clay Paddy Soil

The wetland environment of the paddy field plays an important role in N leaching from the farmland ecosystem [50]. The application of chemical N fertilizer to tilled soils would result in N leaching in paddy soil [47]. CMV application can significantly reduce DIN and DON leaching in paddy soil under equal N conditions. Two mechanisms could explain the reductions of DIN and DON leaching by the CMV treatment. First, N in the surface water is the direct source of leaching loss, and its content can directly reflect the risk of N loss. It has been found that the combined application of organic fertilizers could increase soil organic matter content and fix fertilizer N into the soil, thereby reducing the N concentration in surface water [51,52]. Similarly, in our study, the application of CMV has been demonstrated to lower the concentrations of DIN and DON in surface water by improving soil properties, including pH, soil organic matter (SOM), and urease/protease activities, thereby reducing the leaching of DIN and DON (the structural equation model path coefficients were −0.13 and −0.24, respectively, Figure 8). Secondly, CMV application could increase soil organic matter and N cycling-related enzyme activities (Figure 2), thus promoting soil N adsorption and soluble N production, which was conducive to N accumulation [53]. CMV application can, either directly or by changing the soil properties, increase the storage of soil SIN and SON in our study, thus inhibiting the leaching of DIN and DON (the structural equation model path coefficients were −0.50 and −0.27, respectively, Figure 8). Similar results were also reported by Wang et al. [54] who found that organic fertilization significantly reduced the leaching of soil mineral N by enhancing soil-adsorption capacity. The study by Dong et al. [27] also showed that NO₃⁻ adsorption could decrease the NO₃⁻ leaching at the subsoil in subtropical red-soil regions. Therefore, under the condition of equal N, CMV application could reduce N leaching, and the reduction effect of high dosage was more obvious.

The concentration of DIN and DON in soil water was decreased in CMV treatments, but the reduction effect of DIN leaching was better. This may be due to higher consumption and immobilization of inorganic N in the soil after CMV application. Microbes are an important factor in the decomposition and mineralization of green manure [38], and they need to assimilate a large amount of inorganic N to meet their N demand during green-manure decomposition, thereby reducing the leaching of DIN. In addition, CMV application increased soil microbial activity and biomass, and promoted the immobilization of DIN by microorganisms [55], thereby significantly reducing DIN leaching. Similarly, Zhang et al. [10] also showed that inorganic N in the soil is immobilized in microbial cells and stable organic N, reducing the risk of potential loss of DIN. Therefore, NH₄⁺ immobilization was accelerated with increasing rates of green-manure incorporation, and nitrification is expected to be constrained due to NH₄⁺ assimilation and immobilization [56], thus reducing DIN leaching more effectively. Nitrate, which carries a negative charge like soil particles, is not easily adsorbed by soil and is considered to be the major form of DIN leaching [57].

4.4. Potential Risk of Nitrogen Leaching in Loamy Clay Paddy Soil

The rate variation of the N loss caused by leaching could provide a better comprehension of the N leaching risk [58]. Therefore, DIN and DON leaching loss rates during the rice growth period were investigated in this study. The logistic equations best described the DIN and DON leaching and predicted that the DIN and DON leaching occurred mainly in the early stage of rice growth (Figure 6). The maximum leaching rate occurred within 30 days after CMV application, during which all CMV and 50% urea were applied as basal fertilizer. Similar findings were also reported by Zhao et al. [59], who found that N leaching tended to occur in the early stages of rice growth. Therefore, attention should be paid to field water management in the early stage of rice growth.

The overuse of N fertilizer can reduce N efficiency and increase N loss from soils to groundwater [60], leading to severe N pollution and posing a direct risk to human health [61]. In the two experimental years, the total N loads in leaching under different fertilization treatments were 23.51–61.88 kg hm⁻², indicating that 8.59–22.61% of the N fertilizer applied was leached, which was close to 9.0–22.0% of the total applied as estimated by Amin et al. [62]. This was also far higher than that reported by Kopácek et al. [63], who found that 13.4 kg hm⁻² yr⁻¹ was lost by leaching, accounting for 5.7% of the N applied. The possible reason for these differences may be attributed to the large amount of N applied in this study and the lack of drainage during the rice growth period, resulting in a large amount of N leaching. Within the experimental year of this study, when more than 20% of urea application was replaced by CMV, N leaching in paddy fields was significantly reduced compared with CK treatment. High-dose CMV substitution for partial N fertilizer in paddy soils could save up to 37.37 kg N hm⁻². The optimization of the chemical N fertilization application rate could greatly reduce N leaching [47]. Therefore, in actual production, replacing N fertilizer with 45,000 kg hm⁻² of green manure and fractional fertilization can improve nitrogen-use efficiency and reduce nutrient loss and environmental risk generated by the farmland system.

5. Conclusions

The leaching rate of DIN was higher than that of DON in loamy clay paddy soil. Replacing N fertilizer with CMV could decrease N leaching loss by improving soil properties to increase soil SON and SIN stocks and reduce surface water DIN and DON concentrations under the same nitrogen, phosphorus, and potassium input in the 2-year experiment. Replacing N fertilizer with different amounts of CMV in loamy clay paddy soil could save up to 37.37 kg N hm⁻², and the effect was more obvious with a higher CMV application rate. Under the premise of the application rates used in this study, the CMV application rate of 45,000 kg hm⁻² in loamy clay paddy soil is more friendly to soil and environmental sustainability.