Mechanized Hybrid Rice Seed Production: Planting Density, the Flight Height of an Unmanned Aerial Vehicle, Fertilizer Application, and the Row-Ratio of Parents

: The mechanized seed production of hybrid rice ( Oryza sativa L.) represents signiﬁcant progress in modern agriculture. However, the technologies and the crop management strategies in mechanized hybrid rice seed production are still immature. The present study was conducted with three ﬁeld experiments to explore the effects of different planting densities, the ﬂight height of an agricultural unmanned aerial vehicle (AUAV) for assisting the pollination, fertilization techniques, and the row-ratio of the restore line and the sterile line on seed yield in hybrid rice seed production. In experiment 1, three planting densities “DS1: 250000 hills per ha; DS2: 285000 hills per ha; DS3: 3,33,000 hills per ha” and three ﬂight heights of an AUAV (FH1: above ground 2 m; FH2: above ground 3 m; FH3: above ground 4 m) were adopted. The results showed that DS1 and DS2 treatments produced a higher seed yield than the DS3 treatment, and the seed yield in the FH2 treatment was signiﬁcantly higher than FH1 and FH3. In experiment 2, two fertilizer application methods, mechanized deep placement of fertilizer and traditional manual broadcasted fertilizer, were adopted. The results showed that deep placement of fertilizer signiﬁcantly increased seed yield by increasing the effective panicle number and the grain number per panicle. In experiment 3, three row-ratios of sterile line and restore line at 8:1 (R1), 9:1 (R2), 10:1(R3) were adopted, and the highest seed yield was recorded in the R1 treatment.


Introduction
Rice (Oryza sativa L.) is an important cereal crop, and it constitutes on average more than half of the energy intake of the human population in China, similar to many other countries in the world [1].As before, 60% of the energy intake in the diet is on average from rice, and 40% of the total national grain production was also from rice [2].Therefore, improving rice productivity is important to feed the increasing population of China and the rest of the world.In recent years, the planting area of hybrid rice has been increasing in China [3].Xie et al. [4] indicated that a big accomplishment in modern agriculture was the breeding and the large-scale adoption of hybrid rice, which contributed significantly to the global food supply.In hybrid rice production, seed production is a critical part, but it is very complicated, and it still relies on manual labor rather than mechanization or automation methods [4].
In the last decade, the rural population and the available labor for rice production declined because of the development of urbanization and industrialization in China.Seed production with high manual labor is no longer suitable for farmers due to severe labor scarcity and low efficiency [5,6].Therefore, mechanization is necessary for rice production to decrease production costs and to promote efficiency.
Currently, there are several steps during hybrid rice seed production available for mechanization, such as transplanting, tillage, fertilization, pollination, and harvest.The research by Li et al. [7] indicated that mechanical transplanting was more efficient and less labor intensive compared to traditional manual transplanting in rice production.A previous study showed that mechanized deep placement of fertilizer significantly increased the grain yield of rice and improved both the gross return and the beneficial cost ratio [7].Du et al. [8] demonstrated that deep placement of fertilizer by machine could compensate for the yield loss of rice in zero tillage.In recent years, the rapid development of agricultural unmanned aerial vehicles (AUAV) has made it possible to realize the mechanization of supplementary pollination in hybrid rice seed production [9].However, the system of mechanized hybrid rice seed production is far from perfect because the suitable amendments of many details and parts, including planting density, the ratio of the sterile line and the restorer line, and the flight height of the UAV during pollination are still unknown.
Therefore, the present study was conducted with three field experiments in Guangdong Province (a major rice production province in South China) to explore the effects of different planting densities, different ratios of sterile lines and restorer lines, and different flight heights of an AUAV for assisting pollination on seed yield in hybrid rice seed production.

Field Experiment 1
This experiment was conducted in Xuwen (20 • 33 N, 110 • 17 E, altitude 10 m), Guangdong Province, China, and Zengcheng (23 • 13 N, 113 • 81 E, altitude 11 m), Guangdong Province, China, during March and July 2019.Both sites enjoy a subtropical monsoon climate, and they have been planting rice for many years.The experimental soil at Xuwen was sandy loam, consisting of 12.46 g kg −1 organic matter, 1.04 g kg −1 total nitrogen, 0.68 g kg −1 total phosphorus, 10.32 mg kg −1 total potassium, and the pH was 6.68.The experimental soil at Zengcheng was sandy loam consisting of 15.86 g kg −1 organic matter, 1.33 g kg −1 total nitrogen, 0.93 g kg −1 total phosphorus, 11.67 mg kg −1 total potassium, and the pH was 6.56.The sterile line "Shengyou A" and the restore line "Jinhuangzhan", provided by the Guangdong Agricultural Technology Extension Center, were used as plant materials in the experiment.The rice seeds were sown in polyvinyl chloride (PVC) trays for nursery after the germination, and then 15-day-old seedlings were transplanted into a paddy field at three planting hill densities: (DS1) 16 × 25 cm, about 250,000 hills ha −1 ; (DS2) 14 × 25 cm, about 285,000 hills ha −1 ; and (DS3) 16 × 25 cm, about 333,000 hills ha −1 .When the rice reached the flowering stage, the AUAV was used to assist pollination by manufacturing wind to better spread the pollen.The operation of the AUAV was carried out at 11:00-12:00 a.m. and the flight path of the AUAV is shown in Figure 1.The weather during pollination was sunny.It is common sense that rice blossoms between 9:00 a.m. and 2:00 p.m.The best pollination time was 11:00 a.m.-12:00 a.m.The AUAV was an intelligent, oil-powered, single rotor, agricultural UAV (3WQF120-12, 2.13 mL × 0.70 mB × 0.67 mH, 30 kg, main rotor diameter 2.41 m, the fully autonomous flight control system can fly at a fixed height and speed, pushrod at a fixed distance, and start with one click), produced by QUANFENG AVIATION Co., Ltd., Anyang City, Henan Province, China.Three flight heights of the AUAV were set as: above-ground 2 m (FH1), above-ground 3 m (FH2), and above-ground 4 m (FH3).The flight heights were measured between ground and the rotor of the AUAV.The treatments were arranged in a randomized complete block design (RCBD) in triplicate in each site with a gross plot size of 120 m 2 .At the maturity stage, the rice grains were harvested from the three-unit area of 10 m 2 in each plot, threshed, and sun-dried to determine the seed yield.Meanwhile, 28 presentative hills were collected to determine the effective panicle number, grain number per panicle, seed-setting rate, and 1000-grain weight.The seed-setting rate was the ratio of the full-filling seed number per panicle to the total seed number per panicle.
(RCBD) in triplicate in each site with a gross plot size of 120 m 2 .At the maturity stage, the rice grains were harvested from the three-unit area of 10 m 2 in each plot, threshed, and sun-dried to determine the seed yield.Meanwhile, 28 presentative hills were collected to determine the effective panicle number, grain number per panicle, seed-setting rate, and 1000-grain weight.The seed-setting rate was the ratio of the full-filling seed number per panicle to the total seed number per panicle.

Field Experiment 2
This experiment was conducted in Xuwen (20°33′ N, 110°17′ E, altitude 10 m), Guangdong Province, China, and Zengcheng (23°13′ N, 113°81′ E, altitude 11 m), Guangdong Province, China, during March and July 2019.Both sites enjoy a subtropical monsoon climate, and they have been planting rice for many years.The experimental soil at Xuwen was sandy loam, consisting of 12.79 g kg −1 organic matter, 1.36 g kg −1 total nitrogen, 0.85 g kg-1 total phosphorus, 11.49 mg kg −1 total potassium, and the pH was 6.60.The experimental soil at Zengcheng was sandy loam consisting of 16.03 g kg −1 organic matter, 1.41 g kg −1 total nitrogen, 1.16 g kg −1 total phosphorus, 12.69 mg kg −1 total potassium, and the pH was 6.55.The sterile line "Shengyou A" and restore line "Jinhuangzhan," provided by the Guangdong Agricultural Technology Extension Center, were used as plant materials in the experiment.The commercial compound fertilizer (YARA, N:P2O5:K2O = 15%:15%:15%) was applied at the same amount of 600 kg ha −1 according to the tradition of local farmers.Two application methods, manual surface broadcast three days before transplanting (CK, also the traditional fertilization method of local farmers) and 10 cm depth mechanized placement by machine during the transplanting (DF) were used in the present study.A mechanical hill transplanted rice machine (Kubota, SPV-8C25) synchronous with deep placement of a fertilizer device (Kubota, 2FH-2) were produced by Kubotian agricultural machinery (Suzhou) Co., Ltd., China.In the DF treatment, the fertilizer was deep placed in rows.At the maturity stage, the rice grains were harvested from the three-unit area of 10 m 2 in each plot, threshed, and sun-dried to determine the seed yield.Meanwhile, 28 presentative hills were collected to determine the effective panicle number, grain number per panicle, seed-setting rate, and 1000-grain weight.The seed-setting rate was the ratio of the full-filling seed number per panicle to the total seed number per panicle.The treatments were arranged in a randomized complete block design (RCBD) in triplicate in each site with a gross plot size of 132 m 2 .

Field Experiment 2
This experiment was conducted in Xuwen (20 • 33 N, 110 • 17 E, altitude 10 m), Guangdong Province, China, and Zengcheng (23 • 13 N, 113 • 81 E, altitude 11 m), Guangdong Province, China, during March and July 2019.Both sites enjoy a subtropical monsoon climate, and they have been planting rice for many years.The experimental soil at Xuwen was sandy loam, consisting of 12.79 g kg −1 organic matter, 1.36 g kg −1 total nitrogen, 0.85 g kg −1 total phosphorus, 11.49 mg kg −1 total potassium, and the pH was 6.60.The experimental soil at Zengcheng was sandy loam consisting of 16.03 g kg −1 organic matter, 1.41 g kg −1 total nitrogen, 1.16 g kg −1 total phosphorus, 12.69 mg kg −1 total potassium, and the pH was 6.55.The sterile line "Shengyou A" and restore line "Jinhuangzhan," provided by the Guangdong Agricultural Technology Extension Center, were used as plant materials in the experiment.The commercial compound fertilizer (YARA, N:P 2 O 5 :K 2 O = 15%:15%:15%) was applied at the same amount of 600 kg ha −1 according to the tradition of local farmers.Two application methods, manual surface broadcast three days before transplanting (CK, also the traditional fertilization method of local farmers) and 10 cm depth mechanized placement by machine during the transplanting (DF) were used in the present study.A mechanical hill transplanted rice machine (Kubota, SPV-8C25) synchronous with deep placement of a fertilizer device (Kubota, 2FH-2) were produced by Kubotian agricultural machinery (Suzhou) Co., Ltd., China.In the DF treatment, the fertilizer was deep placed in rows.At the maturity stage, the rice grains were harvested from the three-unit area of 10 m 2 in each plot, threshed, and sun-dried to determine the seed yield.Meanwhile, 28 presentative hills were collected to determine the effective panicle number, grain number per panicle, seed-setting rate, and 1000-grain weight.The seed-setting rate was the ratio of the full-filling seed number per panicle to the total seed number per panicle.The treatments were arranged in a randomized complete block design (RCBD) in triplicate in each site with a gross plot size of 132 m 2 .

Field Experiment 3
This experiment was conducted in Lianjiang (21 • 61 N, 110 • 28 E, altitude 10 m), Guangdong Province, China, and Zengcheng (23 • 13 N, 113 • 81 E, altitude 11 m), Guangdong Province, China during March and July 2019.Both sites enjoy a subtropical monsoon climate, and they have been planting rice for many years.The experimental soil at Xuwen was sandy loam, consisting of 11.86 g kg −1 organic matter, 1.64 g kg −1 total nitrogen, 0.57 g kg −1 total phosphorus, 10.98 mg kg −1 total potassium, and the pH was 6.63.The experimental soil at Zengcheng was sandy loam consisting of 15.49 g kg −1 organic matter, 1.66 g kg −1 total nitrogen, 1.07 g kg −1 total phosphorus, 11.86 mg kg −1 total potassium, and the pH was 6.58.The sterile line "Shengyou A" and the restore line "Jinhuangzhan," which were provided by the Guangdong Agricultural Technology Extension Center, Guangzhou, China.were grown at both sites under three row-ratios of sterile line and restore line at 8:1 (R1), 9:1 (R2), 10:1 (R3).At the maturity stage, the rice grains were harvested from the three-unit area of 10 m 2 in each plot, threshed, and sun-dried to determine the seed yield.The treatments were arranged in a randomized complete block design (RCBD) in triplicate in each site with a gross plot size of 150 m 2 .

Statistical Analysis
The experimental data was tested by employing one-way ANOVA using statistical package Statistix 8.1 (Analytical software, Tallahassee, Florida, FL, USA), according to the methods of Zhou et al. [1] and Pan et al. [5].Mean variances were separated by an LSD-test (p < 0.05).Sigma Plot 14.0 (Systat Software Inc., San Jose, CA, USA) was used for graphical representations.

Effects of Planting Densities and Flight Heights on Hybrid Rice Seed Production
There were significant differences in both seed yield and yield-related traits among different planting density and flight height treatments in both sites (Tables 1 and 2).The highest seed yields and seed-setting rates were recorded in DS1FH2 and DS2FH2 treatments in both sites.The lowest or equally lowest seed yields and seed-setting rates were recorded in DS3FH1 and DS3FH3 treatments.The highest effective panicle number was recorded in DS3FH1, DS3FH2, and DS3FH3 treatments.The lowest effective panicle number was recorded in DS1FH1, DS1FH2, and DS1FH3 treatments.The highest grain number per panicle was recorded in DS1FH1, DS1FH2, and DS1FH3 treatments.The lowest grain number per panicle was recorded in DS3FH1, DS3FH2, and DS3FH3 treatments.There was no remarkable difference in 1000-grain weight among all treatments.

Effects of Different Fertilizer Application Methods on Hybrid Rice Seed Production
As shown in Figure 2, mechanized deep placement of fertilizer significantly increased the seed yield of hybrid rice compared with conventional manual broadcasting of fertilizer; 33.74% and 38.58% higher seed yields were recorded in the DF treatment than in CK in Xuwen and Zengcheng, respectively.Compared with CK, DF treatment significantly increased the effective panicle number and the seed-setting rate by 11.0-15.1% and 13.3-18.3%,respectively.There was no significant difference between CK and DF treatment in grain number per panicle, and 1000-grain weight in both sites.

Effects of Different Parents' Ratios on Hybrid Rice Seed Production
As shown in Figure 3, different parents' ratios remarkably affected the seed yield of hybrid rice seed production.The highest seed yield of hybrid rice was recorded in R1 treatment at both sites.In Lianjiang, the seed yield of the R3 treatment was lower than the R2 treatment, but the difference didn't reach a significant level.In Zengcheng, the seed yield under the R3 treatment was significantly lower than under the R2 treatment.

Effects of Different Parents' Ratios on Hybrid Rice Seed Production
As shown in Figure 3, different parents' ratios remarkably affected the seed yield of hybrid rice seed production.The highest seed yield of hybrid rice was recorded in R1 treatment at both sites.In Lianjiang, the seed yield of the R3 treatment was lower than the R2 treatment, but the difference didn't reach a significant level.In Zengcheng, the seed yield under the R3 was significantly lower than under the R2 treatment.

Discussion
The production of hybrid rice seeds is one of the key steps in increasing rice yield and satisfying the increasing demand for food [10,11].In the 1970s, hybrid rice seed production technology was first applied in China.For the supplementary pollination mainly ropes, bamboo sticks, or wooden clubs to shake the male parents manually and thus to support pollination [12,13].In recent years, agricultural machines, such as an AUAV and transplanters, have developed rapidly, which makes it possible to achieve the goal of high yield and low labor cost in rice production [5].The present study explored the effects of different crop management strategies, including planting density, the flight height of an AUAV, fertilization mode, and the ratio of the parent on the seed yield of hybrid rice.
In Experiment 1, the highest effective panicle number was observed in the DS3 treatment followed by the DS2 treatment and the lowest was observed in the DS1 treatment.The trend of grain number per panicle in different planting densities was the opposite of the effective panicle number.Our results agreed with the study by Huang et al. [14], which demonstrated that increasing hill density substantially increased the panicle number per area of rice.The study of He et al. [15] revealed that increasing planting density was able to remarkably increase the effective panicle number per area of rice.Our results also agreed with the study by Duan et al. [16], which showed that high planting density would

Discussion
The production of hybrid rice seeds is one of the key steps in increasing rice yield and satisfying the increasing demand for food [10,11].In the 1970s, hybrid rice seed production technology was first applied in China.For the supplementary pollination mainly ropes, bamboo sticks, or wooden clubs to shake the male parents manually and thus to support pollination [12,13].In recent years, agricultural machines, such as an AUAV and transplanters, have developed rapidly, which makes it possible to achieve the goal of high yield and low labor cost in rice production [5].The present study explored the effects of different crop management strategies, including planting density, the flight height of an AUAV, fertilization mode, and the ratio of the parent on the seed yield of hybrid rice.
In Experiment 1, the highest effective panicle number was observed in the DS3 treatment followed by the DS2 treatment and the lowest was observed in the DS1 treatment.The trend of grain number per panicle in different planting densities was the opposite of the effective panicle number.Our results agreed with the study by Huang et al. [14], which demonstrated that increasing hill density substantially increased the panicle number per area of rice.The study of He et al. [15] revealed that increasing planting density was able to remarkably increase the effective panicle number per area of rice.Our results also agreed with the study by Duan et al. [16], which showed that high planting density would significantly reduce the grain number per panicle of rice.On the other hand, our results showed that the flight height of an AUAV had no significant influence on yield-related traits except seed-setting rate, while the highest seed-setting rate was recorded in the FH2 treatment in both sites.Overall, the highest seed yields were recorded in DS1FH2 and DS2FH2 treatments in both sites.Thus, we suggested that the planting densities of 250,000 hills per ha and 285,000 hills per ha and the flight height of an AUAV at 3 m could be adopted in hybrid rice seed production.
Many studies have revealed the benefits of mechanized deep placement of fertilizer in rice production.For example, the study by Pan et al. [5] revealed that deep placement of nitrogen fertilizer in direct-seeded rice substantially enhanced the absorption and the utilization of nitrogen in rice plants and hence improved grain yield.Liu et al. [17] demonstrated the deep placement of urea remarkably reduced the ammonia volatilization and significantly increased the nitrogen utilization efficiency of rice in a zero-tillage rice system.In our study (experiment 2), deep placement of fertilizer treatment significantly increased the seed yield of hybrid rice, which was attributed to the increment in the effective panicle number and the grain number per panicle.Our results agreed with the study by Guo et al. [18], which showed that deep placement of fertilizer remarkably enhanced the yield formation of maize.A previous study also revealed that the combination of a rice transplanted machine and a harrow and/or a furrower opener simultaneously helped transplanting and fertilizer application operations and led to a well-aerated micro-environment, favorable soil conditions, crop establishment, and even decreased pest and disease incidence [19].The results of the present study showed that mechanized deep placement of fertilizer also had benefits in hybrid rice seed production.
In experiment 3, the different ratios of parents significantly affected the seed yield of hybrid rice by affecting the seed-setting rate.We observed that the seed-setting rate and the seed yield decreased with the increase of row ration between the restore and the sterile lines.The highest seed-setting rate and the seed yield were recorded in the R1 treatment at both sites.The lower seed setting rates in R2 and R3 treatments might be attributed to the pollen from the restore line being unable to reach the long distant rice plants of the sterile line, even with the help of an UAVA, and it thus led to the decline of the average pollination rate, finally causing the decrement of the average seed-setting rate and seed yield.Thus, the best ratio of parents in the present study was considered to be the R1 treatment.

Conclusions
The present study estimated the effects of multiple cultivation approaches on the production of hybrid rice seeds.According to our results, the recommended planting densities were DS1 and DS2, which were 250,000 hills per ha and 285,000 hills per ha, respectively, compared with DS3, which was 333,000 hills per ha.The recommended flight height of an AUAV was 3 m compared with 2 m and 4 m.The recommended row ratio of the sterile line and the restore line was 8:1 in comparison with 9:1 and 10:1.Using these recommendations, yields in hybrid rice seed production can be improved.

Figure 1 .
Figure 1.Mechanized hybrid rice seed production with unmanned aerial vehicles at pollination.(a): The flight path of the AUAV for assisting pollination; (b): the photograph of the AUAV during assisting pollination.

Figure 1 .
Figure 1.Mechanized hybrid rice seed production with unmanned aerial vehicles at pollination.(a): The flight path of the AUAV for assisting pollination; (b): the photograph of the AUAV during assisting pollination.

Agronomy 2022 , 9 Figure 2 .
Figure 2. Effects of different fertilizer application methods on hybrid rice seed production.CK: conventional manual broadcasting of fertilizer; DF: deep placement of fertilizer.Seed yield was at 14% moisture.The different letters above the column indicate a difference at p < 0.05 by LSD tests.Error bars represent standard error.

Figure 2 .
Figure 2. Effects of different fertilizer application methods on hybrid rice seed production.CK: conventional manual broadcasting of fertilizer; DF: deep placement of fertilizer.Seed yield was at 14% moisture.The different letters above the column indicate a difference at p < 0.05 by LSD tests.Error bars represent standard error.

Agronomy 2022 , 9 Figure 3 .
Figure 3. Effects of different parents' ratios on seed yield of hybrid rice seed production.Seed yield was at 14% moisture.The different letters above the column indicate a difference at p < 0.05 by LSD tests.Error bars represent standard error.

Figure 3 .
Figure 3. Effects of different parents' ratios on seed yield of hybrid rice seed production.Seed yield was at 14% moisture.The different letters above the column indicate a difference at p < 0.05 by LSD tests.Error bars represent standard error.

Table 1 .
Effects of different planting densities and flight heights on hybrid rice seed production in Xuwen.

Table 2 .
Effects of different planting densities and flight heights on hybrid rice seed production in Zengcheng.
Same letters after the values indicate not significant differences between treatments (p ≤ 0.05) according to the least significant difference (LSD) test.