Rotation with Soybean Improved Weed Control and Foxtail Millet Yield

Abstract

Foxtail millet is an important characteristic grain crop in northern China. However, weeds compete seriously with foxtail millet and have long been a biological factor that has plagued foxtail millet production. Rotation requires determining the species and sequence of crops, and reasonable rotation has many benefits for agriculture, including reducing the damage by weeds. In order to clarify the combination effects of foxtail millet–soybean rotation sequence and herbicide on weed control and crop yield, fixed-position experiments were designed in three growing seasons. Foxtail millet and soybean were planted following the sequence below in successive years (FFF, foxtail millet–wheat–foxtail millet–wheat–foxtail millet; SFF: soybean–wheat–foxtail millet–wheat–foxtail millet; SSF, soybean–wheat–soybean–wheat–foxtail millet; SSS, soybean–wheat–soybean–wheat–soybean), and weed density, biodiversity, weed seedbank, and crop yield were examined and analyzed. The results showed that the average weed density of SFF and SSF was reduced by 61.7% and 66.3% compared with FFF in the three years and by 16.5% and 26.6% compared with SSS, separately. Foxtail millet–soybean rotation (SFF and SSF) increased the species richness and the Margalef species richness index of the weed community and reduced the Simpson index compared with the continuous foxtail millet and the continuous soybean cropping (FFF and SSS). The weed seedbank of SFF and SSF was 45.7% and 55.8% smaller than that of FFF and increased by 92.7% and 56.7% compared with SSS, respectively. The weed density in the FFF treatment was significantly correlated with the 0–5 cm grass seedbank size, while there was no significant correlation in the other three treatments. Benefiting from the lower weed damage intensity, the yield of foxtail millet in SFF and SSF increased by 54.05% and 221.81% compared with FFF, respectively. The research results revealed that both SFF and SSF can effectively reduce the damage of weeds and help improve biodiversity. SSF has a higher weed control effect and higher foxtail millet yield than SFF. This study contributes to the understanding of crop–weed interactions in foxtail millet–soybean rotation and can be applied to areas with similar environments.

1. Introduction

Foxtail millet [Setaria italica (L.) Beauv] is an ancient crop originating from China with a planting history of about ca. 8700 cal yr BP [1]. It can resist environmental stress and is nutritious [2]. Foxtail millet is currently planted in many countries and regions, including China, India, Japan, Central Asia, the United States, and Africa, and is an important food and forage crop [3,4,5]. Weeds compete fiercely with foxtail millet for water, nutrients, and space, especially during the seedling stage of the crop and with weed relatives of Setaria Beauv., becoming the main biological factor that reduces foxtail millet yield and quality [6,7,8]. Giant foxtail (Setaria faberii Herrm.) was a close relative of foxtail millet by ISSR (inter-simple sequence repeats) data and dendrogram [9] and may have high competitive sensitivity with the crop because of their similar biological responses to environmental conditions. Soybean [Glycine max (L.) Merr.] is an important oil crop and source of high-quality vegetable protein for humans [10,11], which has important economic value.

Crop rotation is an indispensable element of sustainable agriculture. Rotating foxtail millet with other crops such as potatoes and corn can improve the ecological environment of foxtail millet fields, increase soil fertility, and increase soil saprophytes and pathogens compared to foxtail millet continuous cropping, thereby increasing foxtail millet yields [12,13]. The rotation of millet and soybean can result in more grain yield and larger millet plants, which is a commonly used rotation method [14]. Crop rotation is also an effective method for controlling harmful pests, including weeds [15]. There are also significant differences in weed community characteristics such as species richness, diversity index, and dominance index between different rotation patterns [16]. Therefore, crop rotation can be used as part of integrated weed management [17]. Gunsolus et al. [18] believe that non-chemical weed control technology cannot eliminate weeds, but it can limit the growth of weeds and thus reduce the harm of weeds. It is both a science and an art. Thus, integrated weed management strategies combining crop rotation and herbicide can control weeds more effectively [19].

Although herbicide-resistant foxtail millet varieties are developing rapidly [20], non-herbicide-resistant conventional foxtail millet varieties cannot be replaced, and all foxtail millet varieties are still very sensitive to most herbicides [21]. Dong et al. [22] studied the application of the herbicide penoxsulam designed for rice fields on foxtail millet. The plant height of foxtail millet reduced 25.84% to 49.20%, physiological and biochemical indicators of foxtail millet were inhibited, and the risk of phytotoxicity was high. It is not recommended for field use. Li et al. studied the application potential of various herbicides in foxtail millet fields and found that most herbicides have serious phytotoxicity to foxtail millet [23,24]. In China, MCPA-isooctyl, 2,4-D butylate, bromoxynil octanoate, etc., can be used to control some broadleaf weeds in foxtail millet fields, but there is still an absence of safe and effective herbicides for grass weeds. Weed damage has become an important factor that seriously inhibits the development of the foxtail millet industry [24,25]. Therefore, crop rotation has become an important option for reducing the damage of gramineous weeds in common foxtail millet fields, especially for some difficult-to-control species, but there are few related studies. The objectives of this study are as follows: (1) to analyze and compare the effects of foxtail millet–soybean rotation combined with herbicides (hereinafter referred to as the “technology”) on weed density; (2) to clarify the effects of the technology on the biodiversity of weed communities; (3) to elucidate the effects of the technology on weed seedbanks and crop yields; and (4) to assess the application value of different rotation sequences. The above objectives can provide a basis for integrated weed management in foxtail millet fields.

2. Materials and Methods

2.1. Basic Information of the Experimental Site

The experimental site is located in Mazhuang Village, Shijiazhuang City, Hebei Province, China (114.787944° E, 37.929953° N; 56 m elevation). It is a warm temperate semi-humid continental monsoon climate, with an average annual temperature of 12.5 °C, annual maximum and minimum temperatures of 43.2 °C and −23.4 °C, and an average annual precipitation of 494 mm. The highest precipitation occurs in July and August, accounting for about 56.2% of the year. The annual sunshine hours are 2711.4 h, the sunshine rate is 61.2%, and the frost-free period is 190 days. The planting system in this area is a double-cropping system. Wheat is planted in early October every year. After the wheat is harvested the following year, corn, soybeans, foxtail millet, and other crops are planted in late June. After these crops are harvested in early October, wheat is planted again. This is typical climate type and planting method in the North China Plain.

The soil of the test site is loam, with a nitrogen content of 76.09 mg kg⁻¹, a phosphorus content of 22.09 mg kg⁻¹, a potassium content of 142.38 mg kg⁻¹, an organic matter content of 1.2556%, and a pH value of 8.4.

2.2. Experimental Design and Field Management

The foxtail millet variety planted in the experiment was Yugu 18, and the soybean variety was Jidou 12. Both of the two varieties are high-quality, high-yield, non-herbicide-resistant varieties and are widely planted in China. This experiment was designed in random blocks, with four planting patterns. From 2019 to 2021, foxtail millet–wheat–foxtail millet–wheat–foxtail millet (FFF), soybean–wheat–foxtail millet–wheat–foxtail millet (SFF), soybean–wheat–soybean–wheat–foxtail millet (SSF), and soybean–wheat–soybean–wheat–soybean (SSS) were planted in sequence where foxtail millet/soybean was planted from late June to October and wheat was planted from October to June each year. The position of each plot was fixed during the experiment, and each treatment was repeated four times. The plot area was 25 m² (5 m × 5 m). The row spacing of foxtail millet and soybean was 40 cm. The seedlings were thinned at the time of 4 leaves of foxtail millet and 1 compound leaf of soybean, and the plant spacing was 5.5 cm and 15 cm, respectively. There was a ridge with a width of 15 cm and a height of 8 cm at the junction of different treatments. A protection zone was set at both ends of the experiment and between the treatments. Weeds in the protection zone were removed by manual weeding.

Every year, when weeds are in the 4–6-leaf stage, weed control is carried out using the commonly used chemical weed control method for the crop fields. Soybean fields (at 2 compound leaves of soybean) are sprayed with 15% quizalofop-p-ethyl EC (Jilin Bada Pesticide Co., Ltd., Gongzhuling, China) 750 mL hm⁻² and 250 g L⁻¹ fomesafen (Tianjin Boke Technology Co., Ltd. Yum, Tianjin, China) AS 1200 mL hm⁻²; foxtail millet fields (at 5 leaves of foxtail millet) are sprayed with 56% MCPA-sodium WP (Anhui Yinong Chemical Co., Ltd., Liu’an, China) 750 g hm⁻² and 200 g/L fluroxypyr-meptyl EC (Jilin Jinqiu Pesticide Co., Ltd., Panshi, China) 750 mL hm⁻². The herbicides were applied using a hand-operated knapsack sprayer (207 kPa pressure) (Agrolex Pte Ltd., Gali Batu, Singapore) with a flat fan nozzle 8003EVS (TeeJet Technologies, Glendale Heights, IL, USA).

Wheat was planted as the previous crop in the experimental field. After the wheat was harvested every year, 3 tons hm⁻² of organic fertilizer (fully decomposed pig manure) were evenly applied in late June when the land was tilled twice by a rotary tiller. The tillage depth is 15 cm. Then, the land was leveled twice by a land leveler pulled by a tractor. The field management measures such as fertilization, watering, and prevention and control of pests and diseases were kept consistent.

2.3. Determination Indicators and Methods

2.3.1. Weed Density

Weed density was investigated annually before weed control of crops. Three sample sites were investigated on the diagonal of each plot, and the area of each sample site was 0.25 m² (0.5 m × 0.5 m), and the species and number of weeds were recorded.

2.3.2. Biodiversity and Similarity

The biodiversity indicators and similarity of the weed community were calculated based on the weed species and density from 2021 investigated in Section 2.3.1 [26,27].

Species richness is the number of species in each treatment. The Margalef species diversity index (D_Ma) was calculated by the following formula:

D_Ma = (S − 1)/ln (N)

(1)

where S is the number of weed species, and N is the sum of the densities of all weeds in the weed community.

The Shannon’s diversity index (H′) was calculated as follows:

$$H^{\prime }=-{\sum }_{i=1}^S{ P}_i ln{ P}_i$$

(2)

where S is the species richness, and P_i is the proportion of species i in the community.

Simpson index (D) was calculated by the following formula:

$$D={\sum }_{i=1}^S{ P}_i^2$$

(3)

Pielou evenness index (J) ranges from 0 to 1 and is used to explain the uniformity of species distribution in a community. A larger value indicates that the number of all species is closer or the uniformity is higher. Conversely, a smaller value indicates that a single or a few species dominate. Pielou evenness index was estimated as follows:

J = H′/ln (S)

(4)

Community similarity was measured by systematic cluster analysis following the method shown in Section 2.4 [28].

2.3.3. Weed Seedbank

After the harvest of crops in 2021, soil samples of 0–15 cm depth were collected 10 times for each treatment using a 5 cm diameter soil drill. The 0–5 cm, 5–10 cm, and 10–15 cm soil layers were mixed separately and spread into the culture box with a thickness of no more than 1.5 cm. The soil layer was sprayed with water every day to keep it moist and placed in a light culture room at 25 °C for germination (12 h of light/12 h of darkness). The types and numbers of weeds were recorded once a week and then removed until no new weeds germinated [29,30]. After the survey, the number of germinable weeds at three different depths was accumulated, and the weed seedbank (10,000 grains m⁻³) was calculated as follows:

Seedbank = Q/((10 × 3.14 × (0.05/2)² × 0.05/10,000))

(5)

where Q is the germinated number of weeds at this depth.

2.3.4. Crop Yield

On 12 October 2021, when the crops were matured, the whole plots were harvested and naturally dried, peeled, and weighed, and then the yield was converted according to the area.

2.4. Data Processing

The field experiments were conducted in a completely randomized design with four replications. Microsoft Excel 2016 was used to calculate data and plot the charts. Variance analysis, systematic cluster analysis, and correlation analysis were performed by IBM SPSS Statistics (Version 26). The variance analysis method was a one-way randomized block, and intra-group and inter-group comparisons were performed separately. The post hoc comparison method was Duncan’s multiple comparisons. Systematic cluster analysis was performed based on the linkage between groups with squared Euclidean distance. The correlation analysis method was Pearson’s bivariate correlation analysis (significance was considered at the p < 0.05 level). The influence regulation and correlations of rotation sequence on weed density and seedbanks were determined by the analysis.

3. Results

3.1. Effect of Crop Rotation on Weed Density

The weed survey in 2019 showed that the total weed density and density of each weed species (grass weeds include Setaria faberii Herrm., Digitaria sanguinalis (Linn.) Scop., Eleusine indica Linn. Gaertn.; broadleaf weeds include Acalypha australis Linn., Portulaca oleracea Linn., Amaranthus retroflexus Linn.) at the beginning of the experiment were similar for each treatment, except for S. faberii (Figure 1 and Figure 2). In 2020, the total density of weeds in the foxtail millet continuous treatment (FFF) increased significantly compared with the previous year, while SSF decreased significantly, and SFF and SSS did not decrease significantly. In 2021, the total weed density of FFF decreased significantly compared with 2020, but it was still significantly higher than that in 2019 and higher than that of other treatments. The weed density of SFF in 2021 decreased significantly compared with 2020 and was lower than FFF and SSS. The total weed density of SSF and SSS did not change significantly from 2020, but SSS still decreased significantly compared with 2019.

Crop rotation treatment has a significant control effect on grass weeds (S. faberii and E. indica) compared with foxtail millet continuous cropping (FFF), but the changing density patterns of Digitaria sanguinalis (Linn.) Scop., Eleusine indica Linn. Gaertn., and S. faberii under different treatments are different. S. faberii is the dominant population in the weed community. The density changes significantly between different years during foxtail millet continuous cropping, showing a wave shape of first increasing and then decreasing. The density in 2021 is still significantly higher than that in 2019. The density of S. faberii in other treatments has shown a downward trend year by year, and the density of S. faberii in 2021 is significantly lower than that in 2019 (Figure 2C). The density of D. sanguinalis and E. indica in each treatment showed a decreasing trend year by year, and the density in 2021 was significantly lower than that in 2019 (Figure 2A,B). The reason is the dominant species S. faberii in FFF inhibits other species. This also leads to a decline in its species richness.

There was no significant fluctuation in the density of broadleaf weeds between different treatments and different years, and there was no significant difference in the control effect of crop rotation compared with foxtail millet continuous cropping (FFF) on broadleaf weeds (Figure 2D).

3.2. Effect of Crop Rotation on Species Diversity in Weed Communities

After continuous cropping of foxtail millet (FFF) and soybean (SSS), the number of weed species in the field decreased, with four species and three species, respectively, and the Margalef species diversity index was lower, at 0.51 and 0.42, respectively. The number of weed species in the rotation treatment (SFF and SSF) was more, with five and six species, respectively, and the Margalef species diversity index was also higher than that in the continuous cropping treatment, at 0.86 and 1.08, respectively (

Table 1. Effect of foxtail millet–soybean rotation on weedy community biodiversity.

Treatment	Species Richness (S)	Margalef Species Diversity Index (DMa)	Shannon’s Diversity Index (H′)	Simpson Index (D)	Pielou Evenness Index (J)
FFF	4	0.51	0.2250	0.9107	0.1623
SFF	5	0.86	0.3710	0.8478	0.2305
SSF	6	1.08	0.6071	0.6666	0.3388
SSS	3	0.42	0.4957	0.6955	0.4512

Note: The data are from 2021. Weed species in FFF were E. indica, S. faberii, P. oleracea, and A. retroflexus; weed species in SFF were E. indica, S. faberii, P. oleracea, A. retroflexus, and A. australis; weed species in SSF were D. sanguinalis, E. indica, S. faberii, P. oleracea, A. retroflexus, and A. australis; weed species in SSF were E. indica, S. faberii, and A. retroflexus.

). The Shannon index, Simpson index, and Pielou evenness index were not obvious. The Shannon index, the Simpson index, and the Pielou evenness index have no obvious regularity.

Foxtail millet–soybean rotation affected the structure of the weed community after three years; thus, the similarity was changed too. The weed communities were similar in the SFF, SSF, and SSS treatments. Their weed communities were different from FFF treatment (Figure 3).

3.3. Effect of Crop Rotation on Weed Seedbank

Crop rotation had a significant effect on the size of the weed seedbank. The weed seedbank at a soil depth of 0 to 15 cm in the FFF treatment was the largest, at 176,700 seeds m⁻³; with the increase in the number of soybean plantings, the seedbank decreased significantly, with the seedbank of the SFF, SSF, and SSS treatments being 95,200 seeds m⁻³, 77,400 seeds m⁻³, and 49,400 seeds m⁻³, respectively (

Table 2. Effect of rotation sequence on the seedbanks of different weed species (10,000 seeds m⁻³).

Treatment	E. indica	S. faberii	A. retroflexus	P. oleracea	A. australis
FFF	0.61 ± 0.82 Abc	16.15 ± 1.02 Aa	0.89 ± 0.36 Ab	0.00 ± 0.00 Ac	0.03 ± 0.05 Ac
SFF	0.18 ± 0.23 Ab	8.82 ± 0.92 Ba	0.48 ± 0.23 Ab	0.00 ± 0.00 Ab	0.05 ± 0.06 Ab
SSF	0.13 ± 0.15 Ab	7.00 ± 0.66 Ca	0.54 ± 0.31 Ab	0.03 ± 0.05 Ab	0.05 ± 0.06 Ab
SSS	0.03 ± 0.05 Ac	3.87 ± 0.36 Da	0.89 ± 0.65 Ab	0.10 ± 0.15 Ac	0.05 ± 0.06 Ac

Note: The data in the table are the mean values and standard errors of four replications. Different uppercase letters in the table indicate column differences, and different lowercase letters indicate row differences at p < 0.05.

). The seedbank of S. faberii was the largest in the field, and its size decreased significantly with the increase in the number of rotations with soybeans, indicating that the technology can significantly reduce the accumulation of S. faberii in foxtail millet fields. The second largest seedbank was A. retroflexus, accounting for 4.94–7.02% of the total size.

The weed seedbank size in all treatments was decreased with the increase in soil depth. The weed seedbank was mainly distributed in the 0–5 cm soil layer, with a seedbank size of 31,300 seeds m⁻³–116,600 seeds m⁻³, accounting for 61.60–76.65% of the total. In the 5–10 cm soil layer, the seedbank size was 11,500 seeds m⁻³–47,100 seeds m⁻³, accounting for 20.89–33.391% of the total. The weed seedbank in the 10–15 cm soil layer was the smallest, with a seedbank size of 2300 seeds m⁻³–5100 seeds m⁻³, accounting for 1.53–10.61% of the total (Figure 4).

The effects of crop rotation on seedbanks at different depths are significantly different (Figure 4). Among them, crop rotation had the greatest effect on the seedbank of total weed and grass weed at the soil depth of 0–5 cm, which decreased significantly with the increase in the frequency of soybean plantings (Figure 4A,B). At the soil depth of 5–10 cm, the seedbank of the foxtail millet continuous cropping treatment (FFF) was the largest, and the seedbank of the soybean continuous cropping treatment (SSS) was the smallest. There was no significant effect on the seedbank size between the crop rotation treatments (SFF and SSF, Figure 4A,B). At the soil depth of 10–15 cm, the seedbank of the foxtail millet continuous cropping treatment (FFF) was significantly larger than that of the other three treatments, while there was no significant difference between the crop rotation treatments (SFF and SSF) and the soybean continuous cropping treatment (SSS) (Figure 4A). The distribution pattern of the seedbank of grass weed in the soil is similar to the total seedbank of weeds, but there is no significant difference in the seedbank of grass weeds at a depth of 10–15 cm among the treatments (Figure 4A,B). And there is no significant difference in the seedbank of broadleaf weeds at different soil depths among the treatments (Figure 4C).

Correlation analysis showed that the grass weed seedbank at a soil depth of 0–5 cm in the FFF treatment was significantly correlated with the grass weed seedbank and weed density. The grass weed seedbank at a soil depth of 0–5 cm in the SFF, SSF, and SSS treatments was significantly correlated only with the grass weed seedbank size, the broadleaf weed seedbank in the SSS treatment was significantly correlated with its broadleaf weed seedbank, and there was no significant correlation between the weed seedbank size and weed density in the three treatments.

3.4. Effect of Crop Rotation on Crop Yield

The foxtail millet yield of the SSF treatment was significantly higher than that of other treatments, followed by SFF (Figure 5). The yield of the foxtail millet continuous cropping treatment (FFF) was the lowest. This indicated that the competition intensity between weed and foxtail millet was reduced, and the stress of weeds on foxtail millet yield was weakened because the rotation decreased the weed density and seedbank size significantly.

4. Discussion

Different planting, tillage, and weed control methods have complex effects on weed communities [31]. In general, continuous cropping will lead to an increase in certain types of weeds, changes in the structure of weed communities, and continuous cropping problems. Crop rotation has been proven to have many benefits for agriculture, including improving the ecological environment of farmland, reducing the impact of continuous cropping problems, reducing fertilizer and pesticide inputs, inhibiting weed growth and damage, etc., which is beneficial to crop growth and increasing yields [32]. Therefore, selecting an appropriate rotation model based on local production practices is a basic element of sustainable agriculture.

Taking advantage of the more intense competition between grassland and weeds can achieve better weed control effects than crop rotation [33]. Foxtail millet and soybeans are both crops that are subject to intense competition from weeds [34]. However, under the current situation of increasingly tense grain production, the adoption of comprehensive weed control measures such as crop rotation and corresponding chemical control can achieve higher grain yields and better economic benefits, thereby ensuring food security and social stability. Weed density is generally positively related to the size of a weed seedbank, and a larger germinable weed seedbank will predict more weed density [35], and which is in line with the results of this study.

The use of a variety of weed management measures, including crop rotation, can more effectively manage malignant weeds in the field [36]. This study showed that, combined with conventional chemical weed control technology, the weed density in foxtail millet and soybean rotation fields experiences a significant downward trend, and the weed seedbank was significantly lower than that in foxtail millet continuous cropping, especially the weeds of the genus Setaria, which are seriously harmful to foxtail millet [25]. However, the weed density in foxtail millet continuous cropping fields will not increase significantly all the time, nor will it decrease significantly after crop rotation. Instead, it will fluctuate under the combined effects of intraspecific competition, interspecific competition, and other factors [37]. In other words, the weeds cannot be completely eliminated or grow indefinitely. Therefore, after planting foxtail millet one–two times, it should be rotated with soybean one–two times, which can effectively reduce the weed density, decrease the weed seedbank size, and increase the species richness in the agro-ecosystem. This study conducted a three-year crop rotation experiment and obtained many results that are beneficial to agriculture. Therefore, research on the effects of more crop rotation times and different sites on crops, weeds, and farmland ecological environment can be carried out.

5. Conclusions

A three-year fixed-plot experiment for foxtail millet–soybean rotation was conducted in China. The aim of this study was to determine the optimal sequence of foxtail millet–soybean rotation based on weed control effects by field weed density, weed community biodiversity, weed seedbank, and crop yield after combining rotation with local chemical weeding.

After continuous cropping of foxtail millet, the density of grass weeds increased significantly, especially S. faberii, which led to an increase in density and seedbank of weed, resulting in a serious reduction in yield loss and a decrease in biodiversity. Foxtail millet–soybean rotation decreased the weed density and seedbank size in a foxtail millet field significantly, including S. faberii, which increased the yield of foxtail millet and the field’s biodiversity. In summary, the best rotation pattern sequence was soybean–soybean–foxtail millet, which had the highest weed control effect and foxtail millet yield.