Potted-Seedling Machine Transplantation Simultaneously Promotes Rice Yield, Grain Quality, and Lodging Resistance in China: A Meta-Analysis

Abstract

Potted-seedling machine transplantation (PSMT) is an innovative method of mechanical rice transplanting to improve seedling quality and reduce mechanical injury relative to blanket-seedling machine transplantation (BSMT). However, the responses of yield, grain quality, and risk of lodging in rice to PSMT have not yet been comprehensively defined. Here, we present a meta-analysis of 67 peer-reviewed studies with 382 field observations to investigate the impacts of PSMT on rice yield, grain quality, and lodging resistance in mainland China. The results indicated that compared to BSMT, PSMT increased grain yield, aboveground biomass, and nitrogen uptake by an average of 8.4%, 6.2%, and 7.2%, respectively. PSMT boosted grain yield with hybrid rice (+10.2%) more strongly than with inbred rice (+6.9%). PSMT improved the brown rice rate (+0.74%), milled rice rate (+1.1%), head rice rate (+2.3%), and gel consistency (+4.4%) while reducing the amylose content by 3.7% with no significant effects on the chalky grain rate, chalkiness, length/width ratio, or protein content. The increase in the milled rice rate under PSMT was greater with hybrid rice than with inbred rice. PSMT reduced the lodging index at the first (−5.1%), second (−9.4%), and third (−8.0%) internodes. In conclusion, PSMT is a promising practice for simultaneously improving rice yield, milling quality, cooking and eating quality, and lodging resistance in paddies. In addition, the grain yield and milling quality of hybrid rice under PSMT are higher than those of inbred rice.

1. Introduction

Rice (Oryza sativa L.) is one of the important staple food worldwide, and nearly 3.7 billion people consume rice daily [1]. Rice production must be improved by 28% to meet global food security by 2050 for the growing population [1]. China is the highest rice producer and consumer, contributing to 30% of global rice production and 28% of global rice consumption [2]. However, rice production has faced several critical constraints in China over the past two decades [3,4]. First, large agricultural input costs, rapid urbanization, and rural labor migration lead to less labor available for rice production, resulting in an adverse impact on the increase in grain yield [5]. Second, along with the improvement of people’s living standards, there is an asynchrony between high-quality rice consumption and production [6,7]. Third, extreme climate event-induced rice lodging instantly reduces grain yield and quality [8,9]. Therefore, synchronizing good yield, high quality, and lodging resistance in paddies is essential to address sustainable rice production in China [10,11].

Machine-transplanting technology plays a key role in achieving fully mechanized rice production in paddy soils [6,12]. As rural labor decreases, the method of rice transplanting has shifted progressively from traditional hand planting to machine transplanting [13]. A previous study reported that machine transplanting accounts for approximately 34% of China’s rice planting area [12]. Machine transplanting can greatly alleviate labor intensity and improve rice production efficiency and grain yield [14,15]. Currently, there are two methods of machine transplanting in rice production, blanket-seedling machine transplantation (BSMT) and potted-seedling machine transplantation (PSMT), in China (Figure S1) [6,16]. BSMT was introduced in Japan in the 1980s, and contributed up to 70% of the rice planting area in Jiangsu Province, China, in 2014 [17]. However, despite the standardization of seedling technology, BSMT with high sowing density and thin seedlings can lead to weak seedlings prior to transplanting, a high rate of seedling injury during transplanting, and a low rate of seedling growth after transplanting [6,12,14]. Developing an innovative method of mechanical rice transplanting to enhance seedling quality is critical for mitigating the negative effects of BSMT on rice production. As a developed machine-transplanting method in the 2010s, PSMT can promote seedling quality, reduce mechanical injury, and shorten the seedling establishment period with rapid early growth, thereby facilitating rice yield in paddies [15,16]. Although a large number of studies showed that PSMT increased rice yield, the magnitude of the improvement varied dramatically due to the difference in quantity and quality of seedlings [5,6,12]. The tillering capacity differs greatly between inbred rice and hybrid rice [8]. Generally, inbred rice with low tillering capacity promoted grain yield by increasing seeds quantity [18]. Hybrid rice with great tillering capacity and panicle size was adapted to the large planting density [12]. Thus, we hypothesized that the variation of genetic background could affect the response of rice yield to PSMT. However, the difference in response of rice types to PSMT was poorly investigated. Furthermore, PSMT directly regulates grain quality and rice lodging resistance [6,9]. For instance, PSMT reduced grain appearance quality in rice monoculture systems compared to BSMT, but improved it in rice-crayfish co-culture systems [19]. PSMT boosted gravity center height; however, it reduced mechanical injury and local necrosis in stems, which could avoid rice lodging risk [20]. Thus, it is necessary to conclusively understand the differences between BSMT and PSMT on rice yield, grain quality, and lodging resistance in paddies [21]. Meta-analysis is an effective statistical technique to assess the direction and magnitude of a treatment effect through aggregating independent studies [4,22]. Here, the goal of our study was to investigate how PSMT affected rice yield, grain quality, and lodging resistance relative to BSMT in paddies using meta-analytical techniques.

2. Materials and Methods

2.1. Data Selection

In the present study, a literature survey of the impacts of PSMT on rice yield, grain quality, and lodging resistance was carried out by collecting peer-reviewed publications through the China National Knowledge Infrastructure (http://www.cnki.net/, accessed on 1 August 2022) and Web of Science (http://apps.webofknowledge.com/, accessed on 1 August 2022). We used the phrases and keywords “rice” and “pot seedling” OR “pot-hole seedling” OR “pothole seedling” OR “potted-seedling” OR “bowl seedling” for the article topic. The following three criteria were included in this meta-analysis: (1) studies must include a BSMT method as a control and a PSMT method treatment, which were arranged in a side-by-side paired block; (2) paddy field experiments were included, while pot experiments were excluded; and (3) the sample sizes in the experimental plots were clearly described. We used a normal hill spacing of 30.0 cm × 13.3 cm as the control when the BSMT method reported multiple hill spacing in an independent study [15]. Raw data were extracted directly from texts and tables or digitized from graphical representations according to GetData software (version 2.24, http://getdata-graph-digitizer.com, Moscow, Russia, accessed on 1 August 2022). A total of 382 field observations from 67 peer-reviewed studies were included based on the above criteria (Appendix data). The Appendix data have been submitted to Mendeley Data with DOI: 10.17632/vnww2gyzkg.1. The experimental plots were conducted in Jiangsu Province (52 sites), Sichuan Province (8 sites), Anhui Province (6 sites), Jiangxi Province (2 sites), Liaoning Province (1 site), Shandong Province (1 site), Henan Province (1 site), Zhejiang Province (1 site), Fujian Province (1 site), and Shanghai City (1 site) in the meta-analysis (Figure 1).

In each study, we collected rice yield, yield components (panicle number per m², spikelets per panicle, total spikelets, filled grain percentage, and grain weight), aboveground biomass, harvest index, nitrogen (N) uptake, grain quality (brown rice rate, milled rice rate, head rice rate, chalky grain rate, chalkiness, length/width ratio, amylose content, gel consistency, and protein content), or lodging resistance (lodging index, bending moment, and breaking strength at the first, second, and third internodes). We employed the lodging index to assess rice lodging risk because only two studies reported the rate of actual lodging [9,18]. Previous studies reported that rice types can affect the responses of grain yield, grain quality, and lodging resistance to PSMT treatment [12,14,15]. Thus, studies in this category were divided into two subgroups according to rice types (i.e., hybrid rice or inbred rice). The rice types were confirmed from the China Rice Data Center (https://www.ricedata.cn/, accessed on 10 September 2022) database. Each category should include more than 10 sample sizes to boost statistical power [23,24]. The total spikelets, harvest index, and lodging index can be calculated based on the following formulas if unavailable [6,9]:

Total spikelets (10⁴ ha⁻¹) = panicle number (10⁴ ha⁻¹) × spikelets per panicle

(1)

Harvest index = grain yield (kg ha⁻¹)/aboveground biomass (kg ha⁻¹)

(2)

Lodging index = bending moment (g cm)/breaking strength (g cm)

(3)

2.2. Meta-Analysis

We employed the natural logarithm of the response ratio (ln RR) to quantify the effects of PSMT on rice yield, yield components, aboveground biomass, harvest index, N uptake, grain quality, and lodging resistance [25,26,27,28]:

$$\ln RR=\ln \left(\frac{X_p}{X_b}\right)=\ln \left(X_p\right)-\ln \left(X_b\right)$$

(4)

where X_p indicates the arithmetic mean values of dependent variables under the PSMT method, and X_b indicates the arithmetic mean values of dependent variables under the BSMT method. The inverse of the variance was employed to weight the ln RR, and the mean coefficient of variation was used to calculate the unavailable variance for a specific dataset [4]. The variance of ln RR was estimated by the following formula:

$$\mathrm{variance}=\frac{S{D}_p^2}{N_p{X}_p^2}+\frac{S{D}_b^2}{N_b{X}_b^2}$$

(5)

where SD_p and N_p indicate the standard deviations and sample sizes under the PSMT method, and SD_b and N_b indicate the standard deviations and sample sizes under the BSMT method, respectively.

The overall effect and 95% confidence interval (CI) values were estimated using a mixed-effects model based on the function rma.mv in the R package “metafor” [29]. We employed a Wald-type test to statistically assess the responses of dependent variables (e.g., grain yield) to machine-transplanting methods. In addition, we included the variable “experimental site” as a random effect if the experimental site could obtain multiple observations. The impacts of PSMT on dependent variables were determined to be significant when the 95% CI of effect sizes did not overlap “0” lines. For instance, the grain yield was significantly reduced (<0) or increased (>0) under PSMT relative to BSMT. To ease interpretation, the results were presented as percentage changes (100 × (RR − 1)) according to methods of machine transplanting.

2.3. Analysis of Publication Bias

We employed Rosenberg’s fail-safe number (Nfs) method to test whether there was publication bias [30]. Our results observed that the 5n + 10 (n represents the number of observations) of all dependent variables were far lower than Nfs values (Table S1). In addition, we used the funnel plot to assess potential publication bias (Figures S2–S4), Egger’s regression test was used to further test for asymmetry in the funnel plot and performed a statistical test (Table S1) [31]. Overall, the results of the present meta-analysis were robust [32].

3. Results

3.1. Grain Yield, Yield Components, Aboveground Biomass, Harvest Index, and N Uptake

Averaged among our dataset, PSMT significantly increased grain yield, aboveground biomass, harvest index, and N uptake by 8.4%, 6.2%, 2.5%, and 7.2% relative to BSMT, respectively (Figure 2). PSMT boosted the number of spikelets per panicle (+10.2%), the total number of spikelets (+6.9%), filled grain percentage (+1.5%), and grain weight (+0.66%), while reducing panicle number (−3.0%).

The grain yield, total spikelets, filled grain percentage, and harvest index of hybrid rice (+10.2%, +7.9%, +1.9%, and +4.0%, respectively) were greater than those of inbred rice (+6.9%, +6.0%, +1.2%, and +1.2%, respectively) (Figure 3), whereas no significant differences were observed in panicle number, spikelets per panicle, grain weight, aboveground biomass, or N uptake (Figure S5).

3.2. Grain Quality

On average, compared to BSMT, PSMT increased the brown rice rate (+0.74%), milled rice rate (+1.1%), head rice rate (+2.3%), and gel consistency (+4.4%) while reducing the amylose content (−3.7%), with no significant effects on the chalky grain rate, chalkiness, length/width ratio, or protein content (Figure 4).

The magnitude of the improvement in the milled rice rate of hybrid rice (+1.4%) was higher than that of inbred rice (+1.1%) (Figure 5). However, rice types did not affect the responses of brown rice rate, head rice rate, chalky grain rate, chalkiness, length/width ratio, amylose content, and gel consistency to PSMT treatment (Figure S6a–g). Compared to BSMT, neither hybrid rice nor inbred rice altered the response of protein content to PSMT (Figure S6h).

3.3. Lodging Resistance

Overall, compared to BSMT, PSMT reduced the lodging index at the first (−5.1%), second (−9.4%), and third (−8.0%) internodes while enhancing the bending moment and breaking strength at the first (+5.8% and +10.9%), second (+9.5% and +20.0%), and third (+12.1% and +21.8%) internodes, respectively (Figure 6). In addition, PSMT significantly increased plant height by 2.7% (Figure S7).

Rice types did not affect the responses of plant height (Figure S7) and lodging index, bending moment, or breaking strength at any of the three internodes (Figure S8) to PSMT treatment.

4. Discussion

4.1. Effects of PSMT on Grain Yield

Our meta-analysis showed that PSMT increased rice yield by 8.4% in paddies, consistent with a previous study (6.0~12.6%) [33]. First, although PSMT reduced the panicle number, PSMT with high-quality seedlings significantly increased the number of spikelets per panicle and the total number of spikelets (Figure 2). The higher number of spikelets contributed to a larger sink size in rice [34,35,36]. In terms of panicle structure in rice, PSMT also increased panicle length, grain weight per panicle, grain density, and the number of primary and secondary branches [37]. Second, PSMT reduced mechanical damage to the root and exhibited great root morpho-physiology (e.g., root length, root volume, and root activity) after the heading, resulting in an increase in N uptake (Figure 2) [12]. Raised rice N uptake can facilitate panicle initiation, thereby boosting spikelet number in paddies [38]. Third, by improving seedling quality and reducing mechanical damage to the lower or upper parts of individual leaves, PSMT can boost solar radiation and thermal use efficiency, leading to the facilitation of carbohydrate accumulation and accelerating its translocation from vegetative tissues to grain during the grain-filling periods [15]. Indeed, we found that PSMT increased aboveground biomass, filled grain percentage, grain weight, and harvest index. Thus, PSMT improved rice yield mainly by balancing the relationship between source and sink [6,16]. Last, as discussed below, PSMT enhanced rice lodging resistance (Figure 6), thereby avoiding lodging-induced grain yield loss [9].

In confirmation of our hypothesis, our results indicated that compared to inbred rice, hybrid rice had a greater increase in grain yield under PSMT. The main reason was that hybrid rice presented greater responses of total spikelets and filled grain percentage to PSMT (Figure 3b,c). Compared to inbred rice, hybrid rice can boost the leaf photosynthetic rate and leaf area index after the heading, which benefits improve photosynthetic assimilation production [12,39]. Indeed, we found that hybrid rice tended to increase aboveground biomass (p = 0.08; Figure S5d). In addition, we found that the increase in the harvest index under PSMT was higher with hybrid rice than with inbred rice (Figure 3d), which was closely associated with high photosynthetic assimilates transplanting into panicles and ultimately raised grain yield [6,15].

4.2. Effects of PSMT on Grain Quality

The present results indicated that compared to BSMT, PSMT significantly increased the brown rice rate, milled rice rate, and head rice rate in paddies. With low planting density and mechanical injury, PSMT improved ventilation and light transmission and mitigated leaf aging, thereby enhancing the leaf photosynthetic rate and root morpho-physiological traits after the heading [16,33]. Thus, PSMT had strong capacities for nutrient uptake and photosynthetic assimilate production that led to an improvement in grain filling and then raised rice milling quality [6,40]. Indeed, both the filled grain percentage and grain weight were significantly increased by the PSMT treatment (Figure 2). In addition, more than 86% of studies in the present dataset reported that rice was harvested at the beginning of November or the end of October (Appendix data with DOI: 10.17632/vnww2gyzkg.1). At this time, the rice canopy with low temperatures could limit the grain filling rate [7,41]. However, PSMT improved seedling quality and shortened rice growth period before heading [6], while facilitating effective accumulative temperature at the grain filling phases, improving rice milling quality [15]. Our finding supports that the increase in milled rice rate under PSMT was stronger with hybrid rice than with inbred rice. As described above, hybrid rice has great nutrient uptake and radiation use efficiency characteristics that benefits grain filling, thereby increasing the milled rice rate [12,39].

The ripening temperature is the most important environmental factor affecting rice appearance quality [7,42,43]. The method of PSMT-induced high effective accumulative temperature at the ripening stage could raise the grain filling rate and shorten the grain filling duration, thereby loosening the packing of amyloplasts [44]. In addition, high ripening temperature could accelerate rice respiration rate and benefit α-amylase degrading starch in the rice endosperm, which weakened the availability of assimilates to the developing endosperm [7,45]. Thus, PSMT deteriorated grain appearance quality in the rice monoculture system [6]. In contrast, a previous study indicated that PSMT reduced the chalky grain rate and chalkiness in a rice-crayfish co-culture system [18]. Continuous flooding management in the rice-crayfish co-culture system reduced rice canopy temperature, resulting in high activities of starch synthase, starch branching enzymes, and sucrose synthase, leading to boost grain appearance quality [40]. Thus, across our dataset, the great variation (i.e., high 95% CI) induced by different rice cultivation regimes may reduce the overall effects of PSMT on the chalky grain rate and chalkiness (Figure 4).

Compared to BSMT, PSMT reduced amylose content, but increased gel consistency with no significant effects on protein content. Several reasons can be given to explain the present results. First, the method of PSMT improved the effective accumulative temperature during grain filling phases and thus reduced the effective time on accumulating amylose [6,15]. Second, with high canopy temperature after heading in rice, PSMT could reduce the activity of granule-bound starch synthase and consequently mitigate amylose synthesis, while increasing gel consistency [6,7]. Rice flour with low amylose content and high gel consistency can give sticky and soft cooked grains excellent palatability [46]. Third, the RVA parameters are the main predictor of the texture of cooked rice [47]. A previous study pointed out that compared to BSMT, PSMT increased the rice flour’s final viscosity, peak viscosity, trough viscosity, and breakdown while reducing pasting temperature and setback, resulting in a high stickiness and low hardness of cooked rice [44]. In addition, a low pasting temperature can obviously reduce the cooking time and temperature [7]. Therefore, according to the changes in RVA parameters, amylose content, and gel consistency, our results demonstrated that PSMT can increase rice cooking and eating quality.

4.3. Effects of PSMT on Lodging Resistance

Compared to BSMT, PSMT significantly increased rice yield and plant height, thereby increasing gravity center height and stem bending moment [19]. PSMT reduced mechanical injury to the stems, but enhanced the thickness of the culm wall and stem diameter and cross-section modulus, resulting in a great stem breaking strength [9,19]. In addition, PSMT promoted content of lignin and cellulose in cell walls and the number of vascular bundles in culm, and thus enhanced stem mechanical strength, which presented a high shock-absorbing capacity under extreme environments [9,48]. Our results also indicated that both the bending moment and breaking strength were clearly increased by PSMT (Figure 6). However, the magnitude of the improvement in the bending moment was less than that in the breaking strength at all the three internodes, thereby leading to a lower lodging index for PSMT than for BSMT.

4.4. Limitations and Implications

Three limitations should be noted in the present meta-analysis. First, over 92% of the experimental sites were geographically clustered in the rice–wheat rotation system, while experimental sites from the single- and double-cropped rice systems were rarely included in China (Figure 1). The geographical bias may reduce the evaluation regarding the impacts of PSMT on rice yield, grain quality, and lodging resistance in paddies [49,50]. Thus, comparative studies between BSMT and PSMT on rice production should be further explored in single- and double-cropped rice systems. Second, we did not investigate the interaction between the method of machine-transplanting and other agronomic practices on rice production due to the small number of observations [51]. For instance, the combination of PSMT and slow-release N fertilizers can strongly boost grain yield and N uptake in rice paddies [16]. Third, small sample sizes likely lead to biases as resampling from the same small set of values [4,23]. Therefore, the subgroups of length/width ratio as affected by rice types in our meta-analysis may reduce the statistical power due to only a few observations (Figure S6e).

As an innovative rice cultivation technique, the method of PSMT can be also resistant to abiotic stress (e.g., saline sodic stress, flooding stress, and chilling stress), while the price of PSMT machines is more expensive than BSMT [52]; moreover, PSMT has a complex process during seedling cultivation relative to BSMT [19]. Therefore, we recommend that the government to raise financial subsidies for the purchase of PSMT machines and provide technical guidance to farmers. In addition, with the development of seed germination technology in darkness and micro-sprinkler irrigation system, rice planting area of PSMT increased rapidly in the recent years in China. Despite these shortcomings, our results showed that PSMT can simultaneously increase rice yield, grain quality, and lodging resistance in paddies.

5. Conclusions

The present meta-analysis pointed out that compared with BSMT, PSMT increased grain yield, aboveground biomass, and N uptake. PSMT boosted the brown rice rate, milled rice rate, head rice rate, and gel consistency and reduced amylose content, but no significant effects on the chalky grain rate, chalkiness, length/width ratio, or protein content were observed. PSMT reduced the lodging index at the first, second, and third internodes. PSMT increased yield and milled rice rate with hybrid rice greater than with inbred rice. Thus, developing the PSMT method can simultaneously promote rice yield, milling quality, cooking and eating quality, and lodging resistance, thereby ensuring food security in China. In addition, we suggest that breeding high-quality seedlings and low mechanical damage during transplanting are the urgent priority for mechanized rice production.