Evaluation and Screening of Rapeseed Varieties (Brassica napus L.) Suitable for Mechanized Harvesting with High Yield and Quality

Abstract

Improving seed yield and quality and optimizing plant architecture to adapt to mechanized harvesting are essential strategies for rapeseed industry development in the Yangtze River basin. The present study selected 24 elite rapeseed varieties from the middle region of the Yangtze River basin as materials to investigate the growth period, plant architecture characteristics, lodging resistance, yield, and seed quality across 2 growing seasons. The results showed that plant biomass, silique number per plant, and seed yield showed a significant positive correlation with each other. A high plant growth rate was the prerequisite for early maturity varieties to achieve high yield. The path analysis illustrated that plant architecture can directly affect the seed yield (direct path efficiency = 0.17) or indirectly affect the yield through lodging (indirect path efficiency: −0.37 × 0.30 = −0.11). Therefore, modifying the plant architecture needs to balance the tradeoff between seed yield and lodging. The comprehensive performance of selected genotypes was evaluated by combining D-CRITIC (Distance-based inter-criteria correlation) and membership function methods. From the comprehensive performance across two cropping seasons, the varieties V24, V23, V22, V21, V12, V17, V19, and V20 had substantial potential for mechanized harvesting with high yield and good seed quality. These results provide a theoretical basis for farmers’ decisions and breeding of rapeseed suitable for mechanized harvesting in the Yangtze River basin.

1. Introduction

Rapeseed is a key component of China’s agricultural sector and plays a significant role in the country’s economy [1,2]. It is used in various products, including cooking oil, animal feed, and biodiesel [3]. The country’s growing population drives the demand for rapeseed in China with limited arable land. Rapeseed is grown primarily in the Yangtze River basin of China, typically planted in the fall and harvested in the spring, around May or June, depending on the specific region and cultivar characteristics [4]. Recently, elite genotypes with early maturity have been introduced to the Yangtze River basin to alleviate the temporal contradiction of the double-cropping system [5]. Meanwhile, the traditional planting technology, with high labor costs and low efficiency, restricts the development of China’s rapeseed industry. Mechanized harvesting is the key to improving the level of rapeseed production in China.

Due to the biological characteristics and planting agro-technology of rapeseed, the efficiency of mechanical harvesting is still less than 50% in recent years. Rapeseed has a racemose with indeterminate inflorescences developed from the shoot apical and axillary meristems [6]. The axillary meristem may remain dormant or develop into a lateral axis, depending on the equilibrium between endogenous growth processes and exogenous constraints exerted by the environment [7]. For rapeseed, the upper axillary meristems have a greater chance of developing into effective branches, related to the top flow of nutrients [8]. Longer branch lengths, larger branch angel, and lower branch heights easily cause winding and pulling between branches during machine harvesting, resulting in certain grain loss during mechanical threshing [9]. Plant height is a crucial agronomic trait of plant architecture that affects machine harvesting [10]. Dwarf or semi-dwarf plants have strong lodging resistance and benefit the machine harvesting by reducing the biomass of stalks in the threshing drum [11]. Rapeseed with long plant height is prone to root or stem lodging at the post-flowering stage, deteriorating the photosynthetic product accumulation and machine harvesting at maturity [12].

Selecting and breeding varieties with suitable architecture is an effective strategy to improve the mechanical harvesting ratio of rapeseed. The angle between the branch and stem reflects the compactness of the plant [13,14]. The rapeseed plants have been divided into compact, intermediate, and loose plant architectures, corresponding to the branch angles less than 30°, 30~40°, and more than 40°, respectively [15,16]. Generally, the varieties suitable for machine harvest are characterized by compact architecture with a short plant height, high branching position, and high harvest index. Due to substantial progress in genetic improvement and breeding, the advantageous characteristics that confer lodging resistance and high yield have been deployed in rapeseed varieties [17,18]. However, evaluating the comprehensive performance of rapeseed genotypes suitable for mechanical harvesting is still rare.

It is an arduous task to enlarge the area for the mechanized harvesting of rapeseed in China. On the one hand, we should strengthen the research and improvement of rapeseed harvesting machinery [19]. On the other hand, we should breed and select varieties suitable for mechanized harvesting with high yield [20]. In this scenario, the present study selected 24 elite rapeseed varieties from the middle region of the Yangtze River as materials to investigate the growth period, plant architecture characteristics, lodging resistance, yield, and seed quality. The objective of this study is to comprehensively evaluate the potential of varieties suitable for mechanized harvesting with high yield, thus providing theoretical support for high-yield and high-efficiency rapeseed cultivation in the Yangtze River basin.

2. Materials and Methods

2.1. Experimental Site and Materials

The experiment was carried out at the Agricultural Research Center of Huazhong Agricultural University, Wuhan, China, (114°22′ E, 30°29′ N) during the 2020–2021 and 2021–2022 growing seasons. The previous crop sown in the field was rice for both growing seasons. The soil samples at 0–30 cm depth were taken for soil chemical property analysis. The total nitrogen in the soil was 0.91 g kg⁻¹ and 1.29 g kg⁻¹ for the two growing seasons, respectively. The available nitrogen, phosphorus, and potassium were 24.67 mg kg⁻¹, 14.77 mg kg⁻¹, and 126.83 mg kg⁻¹, respectively, in the first year. The corresponding values were 29.68 mg kg⁻¹, 25.11 mg kg⁻¹, and 154.34 mg kg⁻¹, respectively, in the second year. Twenty-four elite genotypes (Brassica napus L.) were provided by six breeding institutions, representing a wide range of genetic backgrounds in the middle region of the Yangtze River basin (

Table 1. Information on 24 hybrid rapeseed varieties used in the study.

Variety Name	Variety Code	Breeding Institution	Variety Name	Variety Code	Breeding Institution
Shuangyou 195	V01	Henan-AAS	Xiangyouza 787	V13	HAU
Shuangyou 123	V02	Henan-AAS	Xiangyouza 512	V14	HAU
Shuangyou 1918	V03	Henan-AAS	Nongxiangyou 409	V15	HAU
Shuang you 1973	V04	Henan-AAS	Nongxiangyou 510	V16	HAU
Fengyou 737	V05	Hunan-AAS	Ganyouza 8	V17	JX-AAS
Fengyou 845	V06	Hunan-AAS	Ganyouza 9	V18	JX-AAS
Fengyou 306	V07	Hunan-AAS	Ganyouza 10	V19	JX-AAS
FY509	V08	Hunan-AAS	Ganfengyou 3	V20	JX-AAS
Huayouza 9	V09	HAZU	Zhongyouza 19	V21	OCR-CAAS
Huayouza 62	V10	HAZU	Zhongyouza 39	V22	OCR-CAAS
Huayouza 69	V11	HAZU	GDP-10	V23	OCR-CAAS
H20-09	V12	HAZU	Dadi 199	V24	OCR-CAAS

Note: Henan-AAS, Henan Academy of Agricultural Sciences; Hunan-AAS, Hunan Academy of Agricultural Sciences; HZAU, Huazhong Agricultural University; HAU, Hunan Agricultural University; JX-AAS, Jiangxi Academy of Agricultural Sciences; OCR-CAAS, Oil Crops Research Institute, Chinese Academy of Agricultural Sciences.

). FY509, H20-09, and GDP-10 are elite experimental lines, and the rest are local economic varieties, which farmers widely use.

2.2. Experimental Design

The experiment was performed in a randomized complete block design (RCBD) with three replicates. The plot area was 12.0 m² (2.0 m × 6.0 m), with a trench of 30 cm width. The sowing dates were 14 October 2020 and 9 October 2021. Plots were over-seeded in eight rows with 25 cm spacing. Seedlings in each plot were hand-thinned after emergence and determined to accommodate about 45 plants m⁻² at the three-leaf stage. For each plot, 600.00 kg ha⁻¹ compound fertilizer (N 15%, P₂O₅ 15%, K₂O 15%) and 11.25 kg ha⁻¹ boron fertilizer (Na₂B₈O₁₃·4H₂O) were applied as the base, and 112.50 kg ha⁻¹ urea (N 46%) was applied as topdressing at the overwintering stage. All the recommended agronomic techniques were practiced on plots at specific times throughout the two cropping seasons.

2.3. Measurement Index and Methods

2.3.1. Plant Biomass and Leaf Area Index

Five plants were sampled from each plot at the seedling, bolting, flowering, and podding stages, separately. The plants were oven-dried at 105 °C for 30 min to terminate the physiological process and re-oven dried at 75 °C for three days to obtain the dry biomass weight. The leaf area index (LAI) at these four stages was measured by the SunScan canopy analysis system (Delta-T Devices, Cambridge, UK) at 10:00–14:00 under clear and windless weather conditions, with five replicates in each plot.

2.3.2. Plant Lodging Investment

The degree of lodging measurement was carried out on ten randomly selected plants in each plot at maturity. Plant lodging was classified as root lodging (RLA) and stem lodging (SLA). The plant lodging angle was measured as the degree of the line connecting the highest point of the plant and the cotyledonary node diverging from the vertical direction. The root lodging was represented by the degree of the stem diverging from the vertical direction, measured by an inclinometer (IP65 Digital Inclinometer, Woosens Technology Co., Ltd., Shenzhen, China) at the stem, 20 cm above the cotyledonary node. The stem lodging, indicating the bending of the main stem, was calculated by the plant lodging degree, minus the root lodging degree [21,22].

2.3.3. Plant Architecture, Seed Yield, and Quality Measurement

At maturity, five plants were sampled from each plot to evaluate the plant architecture traits, including plant height (PH), and the length and angle of branches at the lowest, middle, and utmost positions in the topological structure of the plant. The branch angle (BA) and branch length (BL) were the averaged values from these three different positions. The relative branch height (RBH) was calculated as the ratio of the lowest branch height to the plant height. The yield component traits, such as pod number per plant, seed yield per plant, and 1000-seed weight, were investigated. The siliques were divided into seed and pericarp, and the seed-to-pericarp weight ratio was calculated by dividing the seed yield by the pericarp weight per plant. The harvest index (HI) was calculated by dividing the plant seed yield by the up-root dry biomass. The plants in the central six rows of each plot were manually harvested and naturally dried, then threshed to get the seed yield per plot. The seed oil content (ratio of oil to seed weight) and oleic acid content (ratio of oleic acid to seed oil content) were analyzed using a Near Infrared Reflectance Spectroscopy System (NIRSystem 3750, Höganäs, Sweden).

2.3.4. Evaluation of the Comprehensive Performance of Variety Suitable for Machine Harvest with High Yield

The raw data of seed yield, growth period, lodging angle, plant height, relative branch height, branch length, branch angle, oil content, and oleic acid content of each tested rapeseed variety were used to evaluate the comprehensive performance. Firstly, these indices were mapped to 0–1 intervals to eliminate the impact of dimension by 0–1 normalization. For the positive indices (seed yield, relative branch height, oil content, and oleic acid content), the 0–1 normalized calculation formula is $U_{ik}=[Y_{ik}-\min (Y_k\left)\right]/\left[\max \right(Y_k)\hspace{0.17em}-\hspace{0.17em}\min (Y_k\left)\right]$; for the remaining negative indices, the corresponding formula is $U_{ik}=\left[\max \right(Y_k)\hspace{0.17em}-\hspace{0.17em}{Y}_{ik}]/\left[\max \right(Y_k)\hspace{0.17em}-\hspace{0.17em}\min (Y_k\left)\right]$, wherein i and k represent the i-th variety and k-th index, respectively. Secondly, the D-CRITIC (Distance-based inter-criteria correlation) method was used to calculate the objective weight $W_k$ of each index [23], which incorporated the distance correlation and standard deviation of each index to produce more valid criteria weights. The contained information of the k-th index ($I_k$) was calculated as follows: $I_k={\mathsf{\sigma }}_k{{\displaystyle \sum }}_{j=1}^n\left(1-p_{kj}\right)$, wherein ${\mathsf{\sigma }}_k$ is the standard deviation of each criterion, and $p_{kj}$ is the distance correlation coefficient between the k-th and j-th indices. Then, the objective weight was calculated as the formula: $W_k=I_k/{{\displaystyle \sum }}_{k=1}^n{I}_k$. In the final stage, the criteria weights and index values of each variety were aggregated into a comprehensive score by the membership function method, as follows: ${\mathrm{D}}_i={{\displaystyle \sum }}_{k=1}^n{U}_{ik}\ast W_k$, wherein ${\mathrm{D}}_i$ is the comprehensive score of i-th variety. Based on these comprehensive scores, the rapeseed varieties could be ranked from the most to the least preferred ones for decision making. The diagram graph for calculating the comprehensive score is provided in Figure 1.

2.4. Statistical Analysis

Analysis of variance (ANOVA) for growing seasons and genotypes was performed using the mixed linear model in SPSS version 20.0. Growing seasons and blocks were considered random effects, and genotypes were considered fixed effects. Significant differences in treatment means were compared by the LSD test at p < 0.05. Origin version 2023 was used to draw figures. The principle component analysis (PCA) and path analysis were conducted under the R environment by FactoMineR and lavaan packages, respectively [24]. We built a priori path analysis models based on current knowledge and hypotheses, representing cause-and-effect relationships between the variables. We expected that root lodging and shoot lodging both affect plant lodging conditions. Plant height, relative branch height, branch angle, and branch length have effects on plant architecture. Biomass and leaf area could indicate the plant growth rate.

3. Results

3.1. Climate Data

The trail field is located in the region of a subtropical monsoon climate, with significant non-periodic changes in temperature and seasonal changes in precipitation. The monthly average temperature, maximum temperature, and minimum temperature showed a similar trend between the growing season of 2020–2021 and 2021–2022 (Figure 2). The lowest monthly average temperature occurs in January (6.0 °C) and February (5.7 °C) for 2020–2021 and 2021–2022, respectively. The lowest monthly rainfall occurred in December, with 14.3 mm and 4.9 mm for 2020–2021 and 2021–2022, respectively. In comparison, the highest monthly rainfall occurred in May (220.4 mm) and March (212.2 mm) for 2020–2021 and 2021–2022, respectively. The total rainfall in the two growing seasons (October to May of the next cropping year) was 582.2 mm and 539.7 mm, respectively, and most of the precipitation came from Mar to May. The total solar radiation in the growing season of 2021–2022 was higher than that in 2020–2021. The lowest solar radiation in the two years occurred in December (168.6 MJ/m²) and January (159.7 MJ/m²), respectively. The highest solar radiation occurred in May, with 419.3 MJ/m² and 501.7 MJ/m² for the growing season 2020–2021 and 2021–2022, respectively.

3.2. Growth Period

The total growth period of the two growing seasons for each variety is shown in Figure 3. The whole growth period among the varieties ranged from 196 to 204 days and from 199 to 210 days for 2020–2021 and 2021–2022, respectively. The accumulated temperature during the whole growth period ranged from 2208~2391 °Cd and from 2284~2517 °Cd for the two growing seasons, respectively. The solar radiation during the whole growth period ranged from 1587~1735 MJ/m² and from 1869~2071 MJ/m², respectively. Considering the two growing seasons, the varieties V24, V04, V22, V20, V23, and V17 showed a shorter growth period, while the varieties V13, V03, V16, V01, V10, and V02 showed a longer growth period (Figure 3).

3.3. Yield Component and Seed Quality

The yield and yield components varied among years and genotypes. The ANOVA analysis indicated that all the yield-related traits showed significant differences among the genotypes, while only the siliques per plant and seed quality traits differed significantly between the two growing seasons (

Table 2. Analyses of variance for the seed yield, yield components, and seed quality.

Variety Number	Seed Yield (kg/ha)		Siliques per Plant		Seeds per Silique		1000-Seed Weight (g)		Harvest Index		Seed-to-Pericarp Ratio		Oil Content (%)		Oleic Acid (%)
	Y1	Y2	Y1	Y2	Y1	Y2	Y1	Y2	Y1	Y2	Y1	Y2	Y1	Y2	Y1	Y2
V01	2628hi	2446o	113jk	115j	20.95cdefgh	18.12hi	3.48efg	3.27fgh	0.32ab	0.27ef	1.18ab	0.91defg	47.25g	47.04hi	68.08fghi	69.09d
V02	2554i	2574mn	106k	105l	22.54abcd	18.23hi	3.49efg	3.50bcdef	0.32abc	0.28def	0.90cdef	0.92defg	50.43cde	47.41gh	68.09fghi	70.15cd
V03	2864def	2814ij	131fghi	118hij	21.20cdefgh	21.5bcdef	3.40fgh	3.29fgh	0.30abcde	0.31abcd	1.05abcde	0.99bcdef	49.52def	48.23gh	69.19efgh	69.22d
V04	2820def	2608mn	122hij	117ij	17.78i	17.44hi	4.80a	4.20a	0.27ef	0.30abcde	0.82efg	1.05abcde	48.70f	50.84de	67.8fghij	71.81bcd
V05	2854def	2745jk	154bc	140d	18.70ghi	17.02i	3.41fgh	3.43cdef	0.30abcde	0.28def	0.90cdef	0.90defg	45.11j	47.07hi	67.9fghij	69.76d
V06	3127a	3283a	156bc	160b	21.86cdef	19.48efgh	3.39fgh	3.31efg	0.32abcd	0.33abc	1.17ab	1.11abcd	47.24g	48.31gh	66.79hijk	69.26d
V07	2937bcd	3209b	164b	153c	21.34cdefgh	21.83bcde	3.16hij	3.06gh	0.32ab	0.31abcde	1.02abcde	1.05abcde	47.06gh	48.72fg	71.44cde	73.75ab
V08	3047ab	3061cde	144cdef	142d	21.62cdefg	22.67abcd	3.56efg	3.34defg	0.31abcd	0.33abc	1.03abcde	1.13abc	45.70ij	47.81gh	67.88fghij	71.13bcd
V09	3106a	3023.e	178a	169a	25.16a	20.72defg	2.92j	3.07gh	0.32abc	0.33abc	1.12abcd	1.08abcd	46.67ghi	47.61gh	68.64fgh	69.76d
V10	2774efg	2907fg	146cde	130f	20.97cdefgh	20.89cdefg	3.14hij	3.65bc	0.33a	0.29cde	1.15abc	0.87efgh	45.27j	46.96hi	65.6jk	65.67e
V11	2859def	2791ij	143cdef	124g	21.83cdef	21.94bcde	3.09ij	3.26fgh	0.31abcd	0.28def	1.06abcde	0.83fgh	44.84j	46.07i	65.36k	63.41e
V12	3133a	3111cd	145cdef	130f	21.42cdefgh	23.21abc	3.67ef	3.60bcd	0.28cde	0.33ab	0.95bcdef	1.05abcde	52.27a	54.27a	68.33fghi	74.25ab
V13	2742fgh	2583mn	133efgh	115jk	24.78ab	21.63bcde	2.93j	3.02h	0.28def	0.28def	0.97bcdef	0.86efgh	49.04f	50.81de	65.95ijk	69.19d
V14	2876def	2847ghi	135efgh	112k	21.59cdefg	22.55abcd	3.53efg	3.53bcdef	0.29bcde	0.27ef	0.87def	0.83fgh	50.93bc	51.2cd	68.13fghi	69.99cd
V15	2581i	2691kl	112jk	101m	22.67abcd	21.88bcde	3.49efg	3.41cdef	0.25f	0.25f	0.74fg	0.73fg	47.44g	50.21de	67.93fghij	70.20cd
V16	2427j	2644lm	152bc	118hi	11.54j	18.44ghi	3.48efg	3.42cdef	0.21g	0.28def	0.64g	0.68g	50.68bcd	52.39bc	76.63a	75.85a
V17	2948bcd	2934f	124ghij	121h	23.85abc	23.62ab	3.61ef	3.59bcde	0.32abc	0.33a	1.10abcd	1.14ab	47.29g	50.42de	68.70fgh	68.63d
V18	2813def	2761jk	129ghi	113k	20.81defgh	21.65bcde	3.49efg	3.5bcdef	0.28cde	0.29def	0.90cdef	0.94cdef	45.84hij	49.58ef	67.03ghijk	69.79d
V19	2910cde	2893fgh	124ghij	107l	22.87abcd	24.99a	3.73de	3.62bcd	0.31abcd	0.31abcd	1.04abcde	0.99bcdef	48.82f	50.34de	71.97bc	71.85bcd
V20	2943bcd	2917fg	124ghij	117hij	23.19abcd	23.37ab	3.92cd	3.75b	0.33a	0.31abcd	1.26a	1.04abcde	49.16ef	50.64de	69.76cdef	73.25abc
V21	2663ghi	2549n	119ijk	93i	18.48hi	23.04abcd	3.64ef	3.59bcde	0.29bcde	0.29bcde	0.94bcdef	0.93cdef	50.36cde	51.4bcd	68.41fghi	70.37cd
V22	2813def	2884fgh	129ghi	121h	22.21bcde	22.49bcd	3.30ghi	3.28fgh	0.31abcd	0.30abcde	1.04abcde	0.94bcdef	51.06abc	52.57b	71.52cd	73.95ab
V23	3130a	3136bc	149cd	133ef	18.89fghi	19.16fghi	4.11bc	4.3a	0.32ab	0.33ab	1.07abcde	1.24a	50.43cde	50.92de	69.38defg	69.95cd
V24	3022abc	3039de	137defj	135e	19.26efghi	19.08fghi	4.21b	4.25a	0.3abcde	0.33abc	1.00bcde	1.05abcde	51.86ab	52.4bc	73.99b	74.37ab
Significant analyses
Y	0.22 NS		94.35 **		0.01 NS		1.07 NS		0.12 NS		0.41 NS		109.62 **		46.06 **
V	60.88 **		35.36 **		12.05 **		25.47 **		7.71 **		3.15 **		50.60 **		17.83 **
Y × V	5.47 **		2.46 **		4.09 **		1.80 *		3.04 **		1.58 NS		5.60 **		2.07 **

Note: Means followed by a different letter indicate significant differences at p < 0.05. “^NS” indicates non-significant; “*” and “**” indicate significant difference at p < 0.05 and p < 0.01, respectively. Y1 and Y2 indicate the growing seasons of 2020–2021 and 2021–2022, respectively. Y: Year; V: Variety; Y × V: interaction between year and variety.

). There was a significant interaction between growing season and genotypes for the yield-related traits, except for the seed-to-pericarp ratio. Averaging across the two growing seasons, V06 has the highest seed yield of 3205 kg/hm², which could mainly be attributed to a larger silique number per plant and high harvest index; V01 and V16 obtained the lowest seed yield of about 2530 kg/hm². There was a 1.5-fold difference in the 1000-seed weight among the genotypes, and the highest (4.498 g) and lowest (2.975 g) were obtained by V04 and V13, respectively. The harvest index showed a 1.2-fold difference among the genotypes, and V17 obtained the highest value (0.33), and the lower value (0.28) was obtained by V13 and V14. The seed-to-pericarp ratio ranged from 0.73 to 1.16 among the genotypes, and V23 and V15, respectively, obtained the highest and lowest values. The oil and oleic acid contents ranged from 45.5% to 53.3% and from 67.6% to 76.2%, respectively. V12 had the highest oil content, and V16 had the highest oleic acid content (Supplementary Figure S1). The correlation analysis showed that the seed yield showed a significant positive correlation with the siliques per plant, plant dry biomass, and harvest index across the two years (Figure 4). Meanwhile, the harvest index showed a significant positive correlation with the seed-to-pericarp ratio.

3.4. Plant Architecture

The principal component analysis was conducted for the plant architecture traits of plant height, branch angle, branch length, and the relative branch height, and the result is shown in Figure 5. PC1 and PC2 could explain about 50% and 30%, respectively, of the total variance for plant architecture traits across the two growing seasons. As the cosine value of the angle between the trait vectors represents the correlation between the two variables, the branch length had a higher negative correlation with relative branch height. The genotypes showed a very similar distribution on the PCA-Biplot for the growing season of 2020–2021 (Figure 5A) and 2021–2022 (Figure 5B). V19 had a large relative branch height and branch angle, as its projection values on these traits were high. V02, V04, V12, and V16 had large plant heights across the two growing seasons. The varieties V05 and V07 had large branch length.

3.5. Lodging Resistance

Figure 6 shows the varieties’ plant lodging, root lodging, and stem lodging angles in the growing seasons of 2020–2021 and 2021–2022. The plant lodging angle range ranged from 8.1~70.9° and 8.7~69.6° for the two growing seasons, respectively. The root lodging range ranged from 3.7~40.6° and 4.6~23.7° for the two growing seasons, respectively. The stem lodging ranged from 3.4~32.8° and 4.0~37.9° for the two growing seasons, respectively. Considering the sorting trend across the two years, the varieties with smaller plant lodging angles were V01, V03, V12, V21, V22, V23, and V24, while the varieties V13, V14, and V15 were not resistant to lodging.

3.6. Path Analysis for the Seed Yield, Plant Architecture, Lodging, and Growth Rate

The path analysis showed the direct contribution of plant architecture, lodging, and growth rate to seed yield (Figure 7). The plant biomass and leaf area index at different developmental stages were used to reveal the plant growth rate. The result showed that the plant biomass at the podding stage (Bio_P) and the leaf area index at the bolting stage (LAI_B) and flowering stage (LAI_F) had a larger contribution to the plant growth rate. The plant growth rate had the largest positive contribution to seed yield with a direct path efficiency of 0.83. Lodging showed a negative contribution to seed yield with a direct path efficiency of −0.37. The plant height and branch length had positive contributions to architecture, while the relative branch height and angles had negative contributions to architecture. The plant architecture had a direct positive contribution to seed yield (direct path efficiency = 0.17). It also had an indirect negative contribution to seed yield by influencing the lodging (indirect path efficiency: −0.37 × 0.30 = −0.11).

3.7. A Comprehensive Evaluation for Mechanized Harvesting and High Yield

The objective weights of seed yield, growth period, lodging angle, plant architecture, and seed quality were obtained by the D-CRITIC calculation method, and the weights for these traits are listed in

Table 3. Weight of each indicator based on the D-CRITIC method for the growing seasons of 2020–2021 and 2021–2022.

Index	2020–2021			2021–2022
	Standard Deviation	Information Content	Objective Weight	Standard Deviation	Information Content	Objective Weight
Seeds yield	0.271	2.601	18.54%	0.267	2.518	20.34%
Growth period	0.273	2.708	19.45%	0.317	2.280	21.86%
Lodging angle	0.318	2.647	22.09%	0.282	2.250	19.20%
Plant architecture	0.282	2.692	19.96%	0.254	2.523	19.39%
Seed quality	0.274	2.773	19.96%	0.242	2.623	19.21%

. The lodging angle had a higher objective weight in 2020–2021, while the growth period had a higher objective weight in 2021–2022; this could be attributed to the large standard deviation of the lodging angle in 2020–2021 and the growth period in 2021–2022. The comprehensive scores of each variety in 2020–2021 and 2021–2022 ranged from 0.24 to 0.93 and 0.26 to 0.86, respectively (

Table 4. Comprehensive scores for each indicator of the 24 tested varieties from the growing seasons of 2020–2021 and 2021–2022.

Variety Number	2020–2021						2021–2022						Averaged Comprehensive Score
	Seed Yield	Calendar Days	Lodging Angle	Plant Architecture	Seed Quality	Comprehensive Score	Seed Yield	Calendar Days	Lodging Angle	Plant Architecture	Seed Quality	Comprehensive Score
V01	0.28	0.38	0.93	0.82	0.33	0.56	0.00	0.09	0.86	0.77	0.29	0.39	0.48
V02	0.18	0.38	0.53	0.01	0.61	0.35	0.15	0.36	0.41	0.00	0.36	0.26	0.30
V03	0.62	0.13	0.98	0.64	0.57	0.60	0.44	0.09	0.94	0.58	0.38	0.48	0.54
V04	0.56	1.00	0.14	0.05	0.44	0.43	0.19	1.00	0.45	0.15	0.66	0.50	0.46
V05	0.60	0.63	0.22	0.14	0.13	0.34	0.36	0.45	0.47	0.16	0.32	0.35	0.35
V06	0.99	0.63	0.23	0.19	0.27	0.45	1.00	0.36	0.52	0.27	0.39	0.51	0.48
V07	0.72	0.63	0.34	0.16	0.46	0.45	0.91	0.45	0.53	0.26	0.59	0.55	0.50
V08	0.88	0.63	0.42	0.64	0.18	0.54	0.73	0.36	0.63	0.51	0.42	0.53	0.54
V09	0.96	0.63	0.64	0.15	0.30	0.53	0.69	0.45	0.85	0.30	0.36	0.53	0.53
V10	0.49	0.38	0.72	0.19	0.05	0.37	0.55	0.27	0.85	0.26	0.15	0.41	0.39
V11	0.61	0.38	0.77	0.54	0.00	0.46	0.41	0.27	0.74	0.38	0.00	0.36	0.41
V12	1.00	0.38	0.82	0.19	0.78	0.63	0.79	0.45	0.91	0.18	1.00	0.66	0.65
V13	0.45	0.00	0.21	0.48	0.39	0.30	0.16	0.00	0.23	0.44	0.56	0.27	0.29
V14	0.64	0.63	0.05	0.67	0.65	0.51	0.48	0.09	0.15	0.51	0.62	0.36	0.44
V15	0.22	0.38	0.00	0.31	0.34	0.24	0.29	0.09	0.00	0.41	0.56	0.27	0.25
V16	0.00	0.13	0.28	0.00	1.00	0.29	0.24	0.09	0.31	0.03	0.93	0.31	0.30
V17	0.74	0.88	0.76	0.45	0.36	0.64	0.58	0.82	0.78	0.29	0.51	0.60	0.62
V18	0.55	0.63	0.32	0.36	0.16	0.40	0.38	0.36	0.34	0.25	0.50	0.36	0.38
V19	0.69	0.88	0.59	0.42	0.63	0.64	0.53	0.64	0.63	0.57	0.63	0.60	0.62
V20	0.73	0.88	0.48	0.17	0.57	0.56	0.56	1.00	0.63	0.22	0.71	0.63	0.60
V21	0.33	0.63	0.93	0.90	0.61	0.69	0.12	0.73	0.90	0.92	0.65	0.66	0.67
V22	0.55	0.88	1.00	0.38	0.81	0.73	0.52	0.73	1.00	0.34	0.87	0.69	0.71
V23	1.00	0.63	0.88	0.52	0.66	0.74	0.82	1.00	0.91	0.59	0.60	0.79	0.76
V24	0.84	1.00	0.84	1.00	0.99	0.93	0.71	0.82	0.90	1.00	0.87	0.86	0.90

). Considering the average comprehensive score ranking across the two years, the varieties with high yield, good quality, and suitable for mechanized harvesting were V24, V23, V22, V21, V12, V17, V19, and V20.

4. Discussion

Screening the varieties with high yield, high quality, and suitable for mechanized harvesting is an important avenue to improving the production efficiency and planting benefit of rapeseed in the Yangtze River basin of China [25]. The key finding from this study revealed the existence of substantial genetic variation and genotype × environment interactions of rapeseed in the growth period, yield components, seed quality, and suitability for mechanized harvesting and the linkage between these traits. The high genotype × environment interaction has been observed in rapeseed productivity in several cropping systems around the world [26,27,28]. Incorporating all the advantageous traits in an ideotype is potentially hazardous. Thus, a comprehensive evaluation of the elite varieties could be a valuable resource for farmers and breeders.

High seed yield has been a consequence of large plant biomass accumulation, coupled with a high harvest index [29]. Increasing the accumulation of dry matter can lead to an increase in the number of siliques in the plant population, thus improving the yield [30]. The correlation analysis for the two cropping seasons showed that plant biomass, silique number per plant, and seed yield showed a significant positive correlation with each other. The seeds per silique and 1000-seed weight showed no significant correlation with seed yield. Meanwhile, the correlation between the number of seeds per pod and the harvest index was significant in the first year, but not significant in the second year, indicating that environmental differences tend to lead to an unstable relationship between these two traits. Siliques act as the primary photosynthetic organs to support the development of the pericarp and seed at the post-flowering stage [31]. The pericarp plays dual roles in the source–sink relationship during seed development, and the translocation of dry matter from the pericarp to the seed largely determines the seed yield [32]. This conformed with our result that the seed-to-pericarp weight ratio showed a significant positive correlation with the harvest index.

Introducing early-maturing rapeseed genotypes to the Yangtze River basin could alleviate the temporal conflict between double-season rice and rapeseed [33,34]. Previous studies had identified a series of candidate genetic variations to control the earlier flowering, which could facilitate breeding for early maturity genotypes in rapeseed [35,36]. In general conditions, high yield and early maturity were negatively correlated. A transcriptional regulator, OsDREB1C, had been identified in rice to boost grain yield and shorten the growth duration [37], implying that the prerequisite for early maturity varieties to achieve high yield is to improve carbon fixation and nitrogen assimilation. In our experiment, the varieties V23 and V24 were both early maturing and high yielding, which was attributed to a larger amount of biomass accumulation in a shorter growth cycle. The order of the growth period of the varieties was not completely consistent between the two years. As the minimum temperature in Dec and Jan in 2021–2022 was higher than that in 2020–2021, the varieties with rigid vernalization requirements needed more days to initiate reproductive growth. The path analysis also showed that the plant growth rate had the largest positive contribution to seed yield. It agreed with a previous study that early-maturing genotypes of rapeseed could have a better tradeoff between early maturation and high seed yield by increasing the light energy and temperature production efficiencies [38].

Plant architecture can directly affect the seed yield or indirectly affect the yield through lodging. The path analysis illustrated that the plant height and branch length had positive contributions to architecture, while the relative branch height and branch angles had negative contributions to architecture. It implied that the positive effect of plant architecture on seed yield derived from the high plant height, long branch length, low relative branch height, and small branch angle. However, this type of plant architecture was prone to lodging. A significant positive correlation existed between the plant height and seed yield per plant of the winter rapeseed [39]. Height reduction was enhanced by applying plant growth regulators, and a significant reduction in lodging accompanied this change [40]. Lodging not only inhibited the yield formation of rapeseed, but also increased the loss rate of machine harvesting [41]. Increasing stem diameter and vascular bundle thickness can enhance the stem’s physical strength to improve rapeseed lodging resistance. Thus, these traits should be considered in further research [42,43].

Taken together, the breeding and selecting of rapeseed varieties should emphasize the agronomic functional characteristics of early maturity, suitable plant architecture, high lodging resistance, high seed yield, and good seed quality. Integrated assessment of varieties’ performance using these multiple traits is a crucial strategy for rapeseed mechanized harvesting development in the Yangtze River basin. The membership function method has been widely used to evaluate the comprehensive performance ability of crops [44,45]. The present study used the D-CRITIC method to calculate the objective weights of target traits to obtain the comprehensive score of each variety. The lodging angle and growth period had a higher objective weight in 2020–2021 and 2021–2022, respectively, suggesting that the objective weight was determined by the genotype and environment interaction. From the comprehensive performance of the two cropping seasons, the varieties V24, V23, V22, V21, V12, V17, V19, and V20 had substantial potential for mechanized harvesting with high yield and good seed quality.

5. Conclusions

Mechanized harvesting of rapeseed requires the systematic combination of varieties, machinery, and cultivation technology to improve planting efficiency. It is urgent to breed and select varieties with early maturation, compact plant architecture, short plant height, and high lodging resistance and high seed yield and seed quality. This experiment revealed the phenotypic diversity of different genotypes regarding the growth period, plant architecture characteristics, lodging resistance, yield, and seed quality. High yield and early maturity are not mutually exclusive, and early maturity varieties with high growth rates are crucial to obtain high seed yields. The comprehensive score of the varieties evaluated in the present research provides candidate criteria for selecting genotypes suitable for mechanized harvesting. Further study should be conducted to investigate the comprehensive scores of rapeseed germplasm resources in multiple year × environment trials to determine outstanding genotypes with broad adaption and high stability suitable for mechanized harvesting.