The experimental site located in Shenyang Experimental Station of Ecology, Chinese Academy of Sciences (41°310′N, 123°220′E). This region has a continental monsoon climate with a mean annual temperature of 7.0–8.0°C, annual precipitation of 650–700 mm, and an annual non-frost period of 147–164 days. The soil at the study site is classified as an Alfisol (US Soil Taxonomy) with 11.28 g kg-1 organic C, 1.20 g kg-1 total N, 0.41 g kg-1 total P, pH (H2O) 6.7 at 0–15 cm depth.
Experimental design and sampling
The study was conducted on soybean plants grown in OTCs from June 16 to August 30 in 2017. The OTCs, which were established in 2008, had a diameter of 1.15 m and a height of 2.4 m with a 45° sloping frustum; the minimum distance between any two chambers was 4 m. Each chamber was made with an iron framework, clad with standard horticultural glass, with a plenum incorporated just below the mouth of the chamber (at 2.4 m from ground-level) to reduce the entrance of ambient air (Bao et al. 2014). From June to September, mean temperature of the day was about 25.6 ± 3.7 ℃ and mean relative air humidity in the OTCs throughout the day was 50.6 ± 19.9%. The mean value of O3 concentration was about 45 ± 5 ppb and the average value of AOT40 (the O3 concentration accumulated over a threshold O3 concentration of 40 ppb during daylight hours) was 3.5 ppm h-1 during clear sky conditions from June to September in the OTCs.
The soybean cultivation was fumigated in the OTCs by O3 for 2.5 months (from June 16 to August 30 in 2017). The experimental design was based on completely randomized plots including two O3 treatments and tree replicates per O3 treatment (overall 6 OTCs). Two O3 treatments were carried out: (1) non-filtered air treatment (control, hereinafter referred to as CK, O3 concentration 45 ± 10 ppb); (2) O3 stress treatment, non-filtered air with addition of O3 35 ppb (hereinafter referred to as O3, 80 ± 10 ppb). The O3 was produced from pure oxygen with an O3 generator (GP-5J Guolin Ltd., Qingdao, China) and then it was mixed with ambient air to achieve the target O3 concentration, and the mixture regulated by flow controllers in each OTC. The top of the OTCs is open. The O3 concentrations were continuously monitored by O3 analyzers (S-900 Aeroqual Ltd., Auckland, New Zealand) every day during the whole day from June 16 to August 30 in 2017, and controlled by computers using a professional program for O3 dispensing and monitoring (Bao et al. 2014).
Meanwhile, there were two straw treatments (tree replicates per straw treatment) for each O3 treatments. Two straw treatments were carried out: (1) no straw return (hereinafter referred to as S-); (2) The total amount of straw is returned to each pot (hereinafter referred to as S+). There were 18 pots per OTC, 3 collected periods (branching stage, flowering stage and podding stage) × 3 replications × 2 straw treatments. According to the average soybean yield and potted area, soybean straw (20 g), which was subjected to O3 fumigation in 2016, was crushed and applied in situ to 20cm depth per pot. Soil, used in each pot experiment, was collected from a cropland (at 0–15 cm layer) at the study site before crops were planted. The previous crop was soybean (Tiefeng 29) and the field did not receive any N fertilization because of the fallow management before the beginning of the pot. After sieving (<2 mm), the soil was immediately used to prepare the pot experiment. The potted soybean cultivar was Tiefeng 29, which was seeded in each pot (26 cm long ×36 cm wide × 45 cm deep) on May 09 in 2017. Before sowing, NH4H2PO4 at 300 kg ha-1 was applied to each plot. The plants were irrigated daily to avoid water stress and appropriate measures were taken to keep the plants free from any stresses of biotic, disease and grass. Five plants were planted in each pot.
Soil samples (containing rhizosphere and bulk soil) was collected at branching stage (July 10, 2017), flowering stage (August 03, 2017) and podding stage (August 30, 2017), respectively. Five soil samples from each pot were randomly collected at 0–10 cm depth by using soil-corer with an inner diameter of 4.5 cm and then they were pooled together to give one composite sample. The collected soil samples were immediately sieved (<2mm) to remove visible stones, roots and plant materials, and then divided into two sub-samples. One sub-sample was air-dried at 25°C for chemical analysis, one sub-sample was stored at 4°C for microbial biomass analysis. Furthermore, at harvest time, all plants were carefully removed from soil of each pot; plant sampling times were: branching stage (July 10, 2017), flowering stage (August 03, 2017) and podding stage (August 30, 2017); soil particles attached to roots were removed by water washing. After harvest, the root biomass was weighed after drying in the oven at 65 °C for 48 h and then were chemical analyzed. Meanwhile, the leaf samples and shoot samples were oven-dried at 65°C for chemical analysis.
Leaf, shoot, root and soil analysis
Total N and P concentration of leaf, shoot, root and soil were determined by elemental analyzer (Vario MAX CNS, Elementar Analysensysteme GmbH, Hanau, Germany) (Zhao et al. 2017). Soil organic carbon (SOC) and total C concentration of leaf, shoot, root and soil were analyzed by the K2Cr2O7-H2SO4 calefaction and titration method (Nelson and Sommers 1982). The C/N and C/P ratios of leaf, shoot, root and soil were calculated from the values of SOC concentration, total C concentration, total N concentration and total P concentration.
Microbial biomass carbon (MBC) and Microbial biomass nitrogen (MBN) of soil cropped to soybean were analyzed by the chloroform fumigation-extraction method as described by Vance et al. (1987). Briefly, soil samples (25 g dry base) were fumigated with ethanol-free chloroform for 24 h at 25 °C. After removal of the chloroform, soluble C and N were extracted from fumigated and non-fumigated samples in 100 mL of 0.5M K2SO4 for 30 min on an orbital shaker. Total organic C in the filtered extract was determined by the K2Cr2O7-H2SO4 calefaction and titration method. Total organic N in the filtered extract was determined using the elemental analyzer (Vario MAX CNS, Elementar Analysensysteme GmbH, Hanau, Germany). We converted microbial C flush (the difference in extractable C between fumigated and non-fumigated samples) to MBC using a factor of 0.45 (Vance et al. 1987) and microbial N flush to MBN using a factor of 0.54 (Brookes et al. 1985). Microbial biomass phosphorus (MBP) of soil cropped to soybean was determined using a fumigation extraction method as described by Brookes et al. (1982). The pre-treatment was in accordance with MBC and MBN. Soluble P in fumigated and non-fumigated soil samples (5 g dry base) was extracted in 100 mL of 0.5M NaHCO3 (pH 8.5) for 30 min on an orbital shaker. We converted microbial P flush to MBP using a factor of 0.40 (Brookes et al. 1982).
The efficiency of conversion of nutrients taken up by the plant into crop biomass was calculated as follows (Tittonella et al. 2008):
Conversion efficiency of nutrient X = total aboveground biomass/total uptake of nutrient X,
where, the total aboveground biomass is the sum of the leaf biomass and shoot biomass at different stages, expressed on a dry weight basis. The conversion efficiencies for N and P have the units: g DM mg N−1, g DM mg P−1 taken up by the soybean per plant, respectively. The uptake of nutrients was calculated from measurements of N and P concentrations in leaf and shoot biomass.
The differences in C, N, P concentrations and stoichiometric ratios, root biomass, aboveground biomass and conversion efficiency of N and P between the two straw treatments and between the two O3 stress were evaluated by one-way analysis of variance (ANOVA) according to Tukey’s test (P < 0.05). Then, a multivariate analysis of variance (general linear model (GLM)) was used to evaluate the effects of the growth stage, O3 stress, straw return as well as their possible interactions on the C, N, P concentrations, stoichiometric ratios, root biomass, aboveground biomass and conversion efficiency of N and P (SPSS 16.0).
Path analyses were conducted to explore the direct and indirect influences of the O3 stress and straw return on soybean root growth. The values of C, N, P concentrations and stoichiometric ratios, soil microbial biomass and root biomass were log-transformed before the path analyses. The path analyses in straw return and no straw return were carried out separately. We started the path analyses procedure with the specification of a conceptual model of hypothetical relationships, based on a priori and theoretical knowledge (Carrillo et al. 2017; Wang et al. 2017b). Data were fitted to the models using the maximum likelihood estimation method (Boldea and Magnus 2009). The best-fitting model was selected by step-wise removal of non-significant paths (P＞0.05). The chi-squared tests (χ2; the model has a good fit when χ2 is low and the P-value＞0.05) were used to test the overall goodness of fit for the model (Grace and Keeley 2006). Standardized regression coefficients and significant level of each path were calculated. All the path analyses were conducted using Amos 22.0 (IBM, SPSS, New York, USA).