*3.1. Grain Yield and Its Components*

The grain yield and its components varied with different N fertilizer applications for both rice cultivars (Table 1). The highest grain yield for MAF treatment was found for the BF and N0 treatments in both years. Mean grain yields of both rice cultivars under MAF were 8.4 t ha<sup>−</sup>1, which was 52.7% higher than the BF treatment. Regarding yield components, deep placement produced the highest number of productive panicles ha−<sup>1</sup> and spikelet per panicle, which was 272.9 × 104 and 189.7, respectively. No significance was found between BF and N0 treatment in the 1000-grain-weight and grain filling. Significant differences were found in the number of productive panicles, spikelets per panicle, 1000-grain-weight, and grain yield between nitrogen treatments. Both rice cultivars differed significantly in 1000-grain-weight.


**Table 1.** Effects of mechanical deep placement of nitrogen fertilizer on average grain yield and its components for both rice cultivars in two-year (2018 and 2019).

MAF: mechanized deep placement of all fertilizer; BF: broadcasting fertilizer; N0: no fertilizer. Average values followed by different letter represent LSD significant differences at *p* < 0.05. \*\*: (*p* < 0.01); \*: (*p* < 0.05); ns: not significant variance.

#### *3.2. Nitrogen Use E*ffi*ciency*

The TNA, NGPE, NHI, NAE, NPFP, and NRE varied with different N fertilizer application in both rice cultivars (Table 2). In two-years, the N fertilizer application (MAF and BF) treatments significantly increased TNA relative to N0 treatment. Moreover, MAF treatment significantly increased TNA compared to BF treatment. The TNA of both rice cultivars for MAF was 173.6 kg ha<sup>−</sup>1, which was 27.7% higher than the BF treatment, respectively. NGPE was highest in MAF, followed by BF, and the lowest was the N0 treatment. MAF showed the maximum NRE and NAE, while the BF had the lowest value for NRE and NRE among all treatments, respectively. Furthermore, a significant difference was found between MAF and BF. The NPFP of both rice cultivars for MAF was 54.2 kg kg<sup>−</sup>1, which was

51.4 % higher than that in the BF treatment, respectively. Moreover, no remarkable difference among all treatment was found in NHI. There were notable differences in TNA, NGPE, NPFP, NAE, and NRE between nitrogen treatments. Moreover, the N × C (Nitrogen × Cultivar) factor interactions also had an obvious impact on TNA and NHI (Table 2).

**Table 2.** Effects of mechanical deep placement of nitrogen fertilizer on average nitrogen use efficiency for both rice cultivars in two-year (2018 and 2019).


Average values followed by different letter represent LSD significant differences at *p* < 0.05. \*\*: (*p* < 0.01); \*: (*p* < 0.05); ns: not significant variance. TNA: Total nitrogen accumulation; NAE: N agronomic efficiency, NRE: N recovery efficiency, NGPE: N grain production efficiency; NHI: N harvest index, NPFP: nitrogen partial factor productivity.

#### *3.3. Total Aboveground Biomass (TAB) and Leaf Area Index (LAI) at Di*ff*erent Growth Stage*

The N fertilizer application remarkably affected the LAI for both rice cultivars (Figure 2). For example, during the MT stage, the LAI for BF and MAF treatments were significantly higher than N0, but no significant difference was found between MAF and BF. At the PI and HS stages, MAF was significantly larger for LAI, especially when compared to N0 and BF. The result manifested that deep placement of the N application could modulate a sustainable longer growth period than surface broadcasting.

**Figure 2.** Effects of mechanical deep placement of nitrogen fertilizer on average for the total above-ground biomass and leaf area index for both rice cultivars (2018 and 2019). (**a**): total above-ground biomass, (**b)**: leaf area index. MT: Mid-tillering stage; PI: Panicle initiation stage; HS: Heading stage; MS: Maturity stage.

The N fertilizer application remarkably affected the TAB for both rice cultivars (Figure 2). For example, at the MT stage, the TAB for BF and MAF treatments were significantly higher than N0. There were no significant differences between MAF and BF, and the TAB for MAF was higher than BF. At the HS and MS stages, N fertilizer application (MAF and BF) treatments remarkably increased TBA, especially when compared to the N0 treatment. Moreover, there were remarkable differences between

MAF and BF treatments. In the whole growth period, a similar trend for TAB of both rice cultivars was observed.

#### *3.4. Determination of Antioxidant Enzymatic Activity*

#### 3.4.1. POD Activity

The POD activity in the leaves at all critical growth stages including the MT, PI, and HS stages were shown in Figure 3. At the MT stage, the highest POD activities were observed in the MAF treatment. However, the POD activity of the BF treatment was higher than the N0 treatment while lower than the MAF treatment. At the PI stage, MAF treatment significantly improved POD activity, especially when compared to the BF treatment. However, the POD activity did not differ significantly between BF and N0 treatment. A significant difference was found among all treatments at the HS stage. In the whole growth period, the similar trend for POD activity of both rice cultivars was observed.

**Figure 3.** Effects of mechanical deep placement of nitrogen fertilizer on average for POD, CAT activity, and MDA content for both rice cultivars (2018 and 2019). (**a**): POD, (**b)**: CAT, (**c)**: MDA.

#### 3.4.2. CAT Activity

The CAT activity in the leaves at all critical growth stages including the MT, PI, and HS stages were shown in Figure 3. There were significant differences in CAT activity found between the MAF and BF treatments at the MT stage. However, the BF and N0 treatments did not differ significantly. A significant difference of CAT activity was found between MAF and N0 treatment at the PI stage. At the HS stage, the highest CAT activities were observed in the MAF treatment over two years. However, CAT activity was higher in the BF treatment than the N0 treatment, while remaining lower than MAF treatment. Furthermore, a significant difference was found among all treatments.

#### 3.4.3. MDA Content

The MDA content in the leaves at all critical growth stages including the MT, PI, and HS stages were shown in Figure 3. At the MT stage, the lowest MDA contents for both rice cultivars over two years were observed in the MAF treatment. However, the MDA content was higher in the BF treatment than the MAF treatment, while remaining lower than the N0 treatment. At the PI stage, compared to the N0 treatment, the BF and MAF treatments could significantly decrease MDA content. Marginal differences in MDA content was found between the MAF and BF treatments. Significant differences of MDA content were found among all treatments at the HS stage.

#### *3.5. Correlation Analysis*

The relationship among grain yield and its components TAB, TNA, LAI, and antioxidant enzyme activities for both rice cultivars over two years (Table 3). The variation of N treatment was focused on in the correlation analysis. Rice yield was significantly and positively correlated with the number of productive panicles ha<sup>−</sup>1, spikelet per panicle, TAB, LAI, and antioxidant enzyme activities, including POD and CAT activities. The spikelet per panicle also significantly correlated with LAI, TAB, and MDA content. Furthermore, TAB at the maturity stage and LAI at the HS stage significantly correlated with antioxidant enzyme activities and TNA.

**Table 3.** Relationship among grain yield, and its components, TAB, TNA, LAI, and antioxidant enzyme activities for both rice cultivars (2018 and 2019).


LAI: Leaf area index at HS stage; TAB: Total above-ground biomass at MS stage; POD: Peroxidase; CAT: Catalase, MDA: malonic dialdehyde at HS stage, TNA: Total nitrogen accumulation at MS stage. The same as below. \*\*: (*p* < 0.01); \*: (*p* < 0.05)

#### **4. Discussion**

#### *4.1. Grain Yield and Its Components*

Compared with broadcasting fertilizer treatment (BF), mechanized deep placement of all fertilizers at once (MAF) significantly enhanced the grain yields in mechanical pot-seedling transplanting (PST). The highest yields of MAF treatment for both rice cultivars were mainly due to the number of productive panicles ha−<sup>1</sup> and spikelet per panicle, which was in agreement with Bandaogo et al. [26]. Moreover, yield increase via a mechanized deep placement in PST was 52.7% for both rice cultivars. The increase was far larger than in previous studies on mechanized deep placement in mechanical carpet-seedling transplanting (6.3–11.6%) [15], suggesting that the mechanized deep placement method could be more effective under PST conditions than non-PST conditions. Moreover, Pan et al. [21] found that deep fertilization remarkably improved spikelet number per panicle in direct-seeded rice compared to manual surface broadcast fertilizers. Tracing it to the cause, the PST was used in our

experiment, which opened a fassula and then uniformly placed the N fertilizer at 10 cm depth. Finally, the fassula was covered by this applicator immediately. This method provided a continuous nitrogen supply for rice growth and the NH4 <sup>+</sup> absorbed by the root system remained in the soil for a longer period of time, thus promoting the growth of rice plants throughout the growth period and thereby increasing the nitrogen absorption and grain yield [27,28]. Deep placement fertilizer also maintained higher antioxidant enzyme activities at the heading stage, which was one of the reasons for the higher grain yield.

Compared with BF treatment, MAF significantly enhanced the leaf area index (LAI) and total above-ground biomass (TAB) at the PI and HS stages in PST. The main reason for this was that MAF treatment promoted rice growth in the middle and late stages by reducing nutrition loss. Broadcasting fertilizers did not meet this demand and led to insufficient nutrient supply. Moreover, LAI and TAB increased by mechanized deep placement in PST, reaching 58.5% and 40.2% for both rice cultivars, respectively. The increase was larger than in previous studies on mechanized deep placement in mechanical carpet-seedling transplanting (36.1–38.9% and 8.7–10.6%) [15,29]. In addition, a larger TAB was beneficial to the transportation of dry matter to the panicle, leading to more spikelet per panicle and 1000-grain-weight of rice, so as to improve the yield of rice. Our results showed that TAB at the MS stage remarkably correlated with LAI at the HS stage, and both of them were positively related to grain yield. The result indicated that MAF had some superiority in larger LAI and TAB, which thus resulted in a higher grain yield.

#### *4.2. Nitrogen Use E*ffi*ciency and Antioxidant Enzyme Activities*

We discovered that MAF led to a substantial increase in total N accumulation (TNA), N recovery efficiency (NRE), and agronomic efficiency (NAE) compared to BF treatment. Moreover, NAE and NRE increase by mechanized deep placement in PST was 104.3% and 123.7% for both rice cultivars, respectively. The increase was far larger than in previous studies on mechanized deep placement in mechanical carpet-seedling transplanting (17.9–43.1% and 19.6–37.4%), suggesting that the mechanized deep placement method could be more effective to improve NRE and NAE under PST conditions than non-PST conditions [29,30]. Some researchers have showed that deep fertilization could improve NUE by reducing nitrogen loss and prolonging the duration of fertilizer, which was compared with surface broadcasting [31,32]. Previous results also found that the deep N fertilization could reduce urease activity by increasing NH4 <sup>+</sup> concentration in soil depth, thereby reducing NH4 <sup>+</sup> concentration in flood water [33,34]. The reason was that the deep nitrogen fertilizer in the anaerobic soil layer caused the NH4 <sup>+</sup> to move slowly from the depth to soil surface, thereby the NH4 <sup>+</sup> content in the flood was low. Moreover, the decrease of N concentration in the flood reduced the loss of N through runoff, ammonia volatilization, and denitrification, thus improving the NUE. Deep placement of fertilizers was a concentrated application of fertilizer near the roots of rice, which was conducive to the absorption of roots and improved the NUE [18,35]. We also found that TNA content was significantly related to grain yield and antioxidant enzyme activities in a positive correlation, because the higher antioxidant enzyme activity at the full heading stage was decisive to the transfer of nutrients to grains, and increased the accumulation of nitrogen in grains, thus increasing the NAE and NRE of rice plants.

Our results showed that MAF notably increased antioxidant enzyme activities including POD and CAT and reduced the MDA content of both rice cultivars. Some reports have showed that the appropriate application of nitrogen fertilizer could maintain high antioxidant enzyme activities at the MS stage, which would be beneficial when delaying the senescence of functional leaves of rice [36,37]. The reason was that the deep application of fertilizer could fulfill nutrient requirements of rice growth and development in time, providing sufficient energy for the antioxidant enzyme activities of the plant, enhancing the activity of the plant leaf protection enzyme system, accelerating the scavenging of free radicals, and reducing membrane lipid peroxidation and the MDA content in the rice plant [38]. Moreover, the rice plant had a stronger root system and longer green leaf duration under deep N fertilization [4]. We also found a significant positive correlation between POD and CAT activities

and grain yield. N-fertilized plants need such effective antioxidant machinery to cope with excessive reactive oxygen species production. In this way, it can delay the senescence of leaves in the late growth stage [23,38]. Therefore, maintaining high POD and CAT activities at the HS stage was conducive to a higher rice yield.
