Experimental site and design
This long-term experiment was conducted in a field orchard located at the Experimental Station of Northwest A&F University (109°56ʹE, 35°21ʹN, altitude 850 m) in Baishui County, Shaanxi Province, China. This site was a typical Loess Plateau area, with a hot and moist summer, but the winter and early spring are always cold. The average annual precipitation is 577.8 mm and 60% falls in the middle of the summer and early autumn (July, August, and September). The soil of the apple orchard was silt loam (with 8% clay, 67% silt, and 25% sand, USDA textural classification system).
The underplanting experiments designed with the following treatments: 4-yr white clover (4-yr WC), 8-yr white clover (8-yr WC), 12-yr white clover (12-yr WC), with clover intercropping between rows of apple trees, and clear tillage used as the control (CT). Each treatment was established in three plots with two rows and covering an area of approximately 200 m2. The intercropping width was 2 m, and the edge was 1 m away from the row of fruit trees. The Fuji apple tree stock (Malus pumila Mil)) was grafted onto M26 root stock in 1998, 2006, and 2012, respectively, and the orchard management regime involved clear tillage. White clover (Trifolium repens) was planted in 2006, 2011, and 2015, sown in between two tree rows at 0.75g m− 2 after deep tillage. Field management for the white clover treatments involved aboveground mowing in early July, August, and September, and the residue was retained on the soil surface as mulch. The surface of the CT treatment (up to 15–20 cm soil depth) was weeded using a rotavator to keep the soil clear of vegetation. Each treatment received an application of chemical fertilizer at a rate of 3660 kg per hectare each year, containing N 18%, P2O5 10%, and K2O 18%. All other apple agricultural management followed traditional farming methods.
Soil sampling and processing
All soil samples were collected in mid-April 2018 using a five-point sampling method, that is, five soil cores were randomly taken from each plot, and split into four depths (0–10 cm, 10–20 cm, 20–40 cm and 40–60 cm). Soils from the same depth were mixed in situ for each plot as one composite sample and each plot served as a replicate. In total, 48 (4 treatments ×3 plots ×4 depths of soil) samples were collected. One subsample was immediately transported to the laboratory and then stored at -20 ⁰C for analysis of microbial biomass carbon and nitrogen. The other subsample was air dried, then passed through 0.1mm and 0.25mm sieves prior to the analysis soil basic physical and chemical soil properties.
Soil physicochemical property analyses
The soil total nitrogen (TN) was determined by acid digestion according to the Kjeldahl method (Bremner and Mulvaney. 1982). Soil organic carbon (SOC) was measured using the K2CrO7-H2SO4 oxidation method (Walkley and Black. 1934). The Persulfate Digestion Method was used to determine total dissolved nitrogen. Nitrate (NO3−N) and ammonium (NH4+-N) nitrogen were determined using a continuous flow analytical system (AA3, SEAL Company, Germany) (Song et al. 2015). Dissolved organic nitrogen (DON) was calculated as: DON = total dissolved nitrogen (TDN) − NO3−N−NH4+-N (Franklin et al., 2019). Microbial biomass carbon (MBC) or nitrogen (MBN) were determined by the chloroform fumigation extraction method (Vance et al. 1987; Wu et al. 1990). The values for the basic soil physicochemical properties in this study are presented in the supplementary materials (Table S1).
Soil dissolved organic matter extraction and analysis
Air-dried soil samples that had been passed through a 0.15 mm sieve were used to determine dissolved organic matter by water–soil oscillation. Briefly, the soil samples were extracted with deionized water at solid/liquid at ratios of 1:10 (w/v) in a shaker at 60°C and 300 rpm for 30 min. The extraction solutions were centrifuged at 8000 rpm for 6min, and filtered using a 0.45µm acetate fibre membrane. Finally, the filtrate was transferred and stored at -20°C until analysis. The filtered solution that had passed through the 0.45µm acetate fibre membrane was analysed for DOC content. DOM is usually quantified as DOC because carbon represents about 67% of the elemental composition of organic matter (Li et al. 2019). DOC was measured using a total organic carbon analyser (Shimadzu, TOC-L, Japan).
UV–Vis and fluorescence spectral analysis
UV–Vis absorbance at a scanning wavelength of 250–400 nm was measured in a 10-mm quartz cuvette using a UV spectrophotometer (Shimadzu, UV-1780, Japan). UV-Vis spectral parameters included absorbance at 254 nm (SUVA254) and slope ratio (SR). SUVA254 was calculated as the UV absorbance at 254 nm per 1 mg C− 1 m− 1; a higher SUVA254 value indicates that the DOM is more aromatic and humic (Xie et al. 2019). SUVA254 > 4 represents hydrophilic components with high aromaticity and high molecular weight, but SUVA254 < 3 represents hydrophobic components (Scott and Rothstein. 2014). The slope ratio (SR) is the slope ratio of the total absorbance values at 275–295 nm relative to the total absorbance values at 350–400 nm. SR increased with decreasing DOM molecular weight (Li et al. 2019).
Fluorescence EEMs were analysed using a fluorescence spectrometer (RF-6000; Shimadzu). Prior to testing, DOC concentrations were diluted to < 8 mg L− 1 to minimize the influence of any internal filtering. The EEMs were created using the following parameters: emission wavelength was set between 200 and 500 nm with 5-nm increments, and the excitation wavelength was between 250 and 550 nm with 2-nm increments, with a scan speed of 2000 nm min− 1. Absorbance measurements were used to correct the internal filtering effects of all EEMs, with most of the effects associated with Raman scatter eliminated by subtracting the Milli-Q blank. Humification (HIX) and fluorescence (FI) optical indicies were determined on the basis of EEM data for separately describing the DOM humification degree and source (Li et al. 2019; Zhang et al. 2019). FI ≤ 1.2 represents the plant residue source, FI ≥ 1.8 represents the microorganism source, and 1.2 < FI ≤ 1.8, mixed resources (Fellman et al. 2010). Humification degree is positively related to HIX value.
Data processing and statistical analysis.
Fluorescence EEM spectra of the filtered soil samples (48 samples) were used in parallel factor (PARAFAC) analysis. PARAFAC models were created in MATLAB R2015a (MathWorks, Natick, MA, USA) using the DOM Flour toolbox (Stedmon and Bro. 2008). The characteristic variables of the fluorescence components of DOM can be derived by this type of processing. The fluorescence spectrum was divided into independent components based on core consistency diagnostics and a split-half validation. Fluorescence intensities of all components were represented by the maximum fluorescence intensity (Fmax). Fluorescence EEM spectra of the DOM five-component model are shown in Fig. 2.
The averages and standard deviations were calculated using Excel 2013 (Microsoft Office 2013, Microsoft, USA). Differences in DOC and DON concentrations, UV-Vis indices and optical indices were tested by one-way ANOVA and Tukey’s test (p < 0.05) (version 23; SPSS Inc), comparing different soil depths and intercropping treatments, after determining that the data were generally normally distributed and exhibited variance homogeneity. The figures were drawn using Origin 9.0. Redundancy analysis (RDA) was conducted on data representing DOM characteristics, DOC content, UV-Vis indices, components, optical indices and physicochemical properties from all soil layers using Canoco5.0. Linear regression was used to examine the relationship between DOC and MBC or SOC using Origin 9.0. Pearson’s correlation analyses were carried out to examine the relationship between DOM characteristics and physicochemical properties before and after eight years of white clover growth.