Site description
The field experiment was initiated in 2013 at the Yongchun County, Fujian Province, China (25°12′37″ N, 118°10′24″ E), using the two rotations of vegetables and rice (Fig. 1). The site is in the north of the Tropic of Cancer, with a typical subtropical marine monsoon climate, sufficient sunshine, and average annual solar radiation 462.26 kJ/cm2. The climate is mild and humid, with average annual temperature 16–21 °C and average annual rainfall about 1400 mm. Agricultural production allows for the cultivation of three crops annually. The soil of the test field was lateritic red soil.
Location of the field experiment site.
Experiment design
The experiment was conducted over 9 years from 2013 to 2021. Soil samples were collected before the experiment began to determine the main physical and chemical properties of the soil in the test plot, which were: organic matter content 19.96 g/kg, total nitrogen 2.25 g/kg, total phosphorus 1.31 g/kg, total potassium 27.86 g/kg, alkaline hydrolyzable nitrogen 107.73 mg/kg, available phosphorus 60.35 mg/kg, available potassium 116 mg/kg and soil pH 5.54. The test site was a rectangular field, 26 m long and 9 m wide, divided into 15 test blocks, each 5 m long and 2.8 m wide. Cement ridges were used to separate the test blocks, and irrigation drainage ditches were set outside the blocks. A protective isolation strip 1 m wide was formed around the test site. The experiment included two crop rotations: (I) rotation P–B–O: P, kidney bean (Phaseolus vulgaris L.), B, mustard (Brassica juncea L.), O, rice (Oryza sativa L.); and II) rotation P–B–V: P, kidney bean (P. vulgaris L.), B, mustard (B. juncea L.), V, cowpea (Vigna unguiculata L.). Four fertilizer treatments were selected: (1) recommended fertilization (RF) used with rotation P–B–O; (2) recommended fertilization (RF) used with P–B–V; (3) conventional fertilization (CF) used with P–B–O; (4) conventional fertilization (CF) used with P–B–V. A randomized complete block experimental design with three replications was used in the field study. The fertilization amounts used for treatments RF and CF are shown in Table 1. Under the CF, the amount of fertilizer applied to crops in each season is determined according to the years of fertilization habits of local farmers. The fertilization amount of crops in each season under the RF was calculated according to the measured basic soil fertility combined with the fertilization model of previous studies. The fertilization amount of crops in each season under the CF in this study is obtained by investigating the local farmers. The data on the fertilization amount of crops in each season under the RF is cited from the research report of Zhang et al.23. Urea (N 46%) was the nitrogen fertilizer, calcium superphosphate (P2O5 12%) was the phosphorus fertilizer, and potassium chloride (K2O 60%) was the potassium fertilizer. All phosphorus fertilizer applied to crops in each season was used as base fertilizer, and nitrogen and potassium fertilizer were applied separately as base fertilizer (40% of the total fertilization) and topdressing (60% of the total fertilization). The topdressing method was that nitrogen and potassium fertilizer for kidney bean and cowpea were applied twice, 30% of the fertilization amount each time; nitrogen and potassium fertilizer for mustard was applied three times, 20% of the fertilization amount each time; nitrogen fertilizer for rice was applied at two different growing stages, 50% of the fertilization amount at the tillering stage and 10% of the fertilization amount at the panicle stage; potassium fertilizer was applied once, using 60% of the fertilization amount. The first crop, kidney bean, was sown in early September and harvested in November. The second crop mustard, was sown in early December and harvested in February of the following year. The third crop, rice or cowpea, was sown in early April and harvested in July.
Data analysis and methods
Yield stability analysis was conducted for the 9 years period using three different approaches. First, the coefficient of variation (CV) was calculated to give a measure of the temporal variability of yield for each treatment:
$$CV=frac{upsigma }{Y}*100 {%}$$
(1)
where σ is the standard deviation of average crop yield in each year, and Y is the average crop yield in each year. A low value of CV indicates little variation, which implies that interannual difference in crop yield in the experimental plot is small and the yield is relatively stable over the years of the experimental period.
A second yield stability indicator is the sustainable yield index (SYI), which is calculated by Singh et al.25:
$$SYI=frac{mathrm{Y}-upsigma }{{Y}_{max}}$$
(2)
where Y is the average annual crop yield, σ is the standard deviation of the average annual crop yield, and YMax is the maximum annual crop yield. A high value of SYI indicates a greater capacity of the soil to sustain a particular crop yield over time.
The third stability measure is Wricke’s ecovalence index (Wi2), which was calculated individually for each crop management system by Wricke26:
$${Wi}^{2}={sum }_{j=1}^{q}({x}_{ij}-{{m}_{i}-{m}_{j}+m)}^{2}$$
(3)
where xij is the yield for treatment i in year j, mi is the yield for treatment i across all years, mj is the yield for year j across all treatments, and m is the average yield for all treatments across all years. When Wi2 is close to 0, the yield for treatment i is very stable.
Analysis of crop yield trends
A simple linear regression analysis of grain yield (slopes and P values) over the years was performed to identify the yield trend (Choudhary et al.27):
$$Y=a+bt$$
(4)
where Y is the crop yield (t/ha), a is a constant, t is the time in years, and b is the slope, or magnitude of the yield trend (annual rate of change in yield).
Analysis of variance (ANOVA) was performed using MATLAB R2019b in order to compare crop yields in the long term experiment. Yield stability and univariate linear regression equations were created and statistically analyzed using the software toolbox. The coefficients of variation for yields, yield sustainability indexes, and graphs presented in this paper were calculated and drawn using MATLAB; differences were considered to be significant when P < 0.05.
Interaction effects on crops yield
The long term crop yield data from this study provide a unique opportunity to assess yield and yield stability. Crops yield was influenced by the interaction of rotation tillage and fertilizer application, and environmental conditions (interannual change). To assess the complex interaction effect of factors on crop yield, analysis of variance (ANOVA) was performed for interaction effects that included interannual change × crop rotation, interannual variation × fertilization, crop rotation crop × fertilization, and interannual change × crop rotation × fertilization.
Ethics approval
The experimental research and field studies on plants, including the collection of plant material, complied with relevant institutional, national, and international guidelines and legislation. The appropriate permissions and/or licenses for collection of plant or seed specimens were obtained for the study.
Source: Ecology - nature.com
