2. materials and methods

2.1. The study area and experimental design

The field experiment was conducted in the Yuyang District of Yulin City, Shaanxi Province, China (Fig. 1). The area belongs to semi-arid continental monsoon climate with an average elevation of 1000-1200 meters, annual average precipitation is 365.7 mm (Most occurs in July and August), annual average temperature of 10.7°C (monthly average temperature range is -13-29°C). The soil is a sandy loam with 66% silt and 32% clay, representing the local farmland soil. The landforms are mostly semi-fixed and fixed sand dunes, and the ecological environment is fragile due to the relatively barren soil and simple species structure.
Our study started in July 2019, according to field investigations and local government documents on alfalfa cultivation, 24 sites with similar geographical features and soil types, were selected as experimental plots, including seven age groups (1, 2, 3, 5, 7, 9 and 10 years) of the alfalfa with 3 sites per age group and 3 maize (Zea mays ) farmland (FL) as the control. Before the establishment of alfalfa, maize and potatoes (Solanum tuberosum ) were mainly grown in this area. Sites of different stand ages are more than 2 km apart, and 3 repeated sites of the same stand age are more than 1 km apart. Three plots (20×20 m) from each site were randomly selected for investigation and sampling, and the average measures of three replicated plots in the same site was used as the observation.

2.2. Plant and soil samplings

We selected five 1 m×1 m squares according to the five-point sampling method in each plot. All soil and plant samples were collected in squares, and represent each plot after fully mixed. All age groups were sampled at the early flowering stages of alfalfa in July and August, 2019. And 72 samples were collected (eight age group × three sites per age group × three replicate plots per site).
We used a 4 cm diameter soil drill to collect 0-10 cm soil in each square, and removed the plant roots and stones. Samples were divided into two parts after homogenization, a part of the soil sample was air-dried for the chemical property measurements and the other part was stored in the refrigerator (about 4°C). Then soil enzyme activity and microbial biomass were measured within 1 week. Aboveground and underground biomass of alfalfa has been investigated in each square. Plants were cut to ground-level, meanwhile, root system was sampled by a root drill (10 cm diameter) to take 0-100 cm depth of the soil profile. Then the samples were taken back to lab, rinse with water and oven-dried for 10 min at 105°C and then at 65°C for 72 hours, and weighed.

2.3. Measurements and calculations

Total C (TC) in the plant and soil was measured by the organic element analyzer (elementar vario MACRO cube, Elementar, Germany). Soil organic C (SOC) was measured using the potassium dichromate/sulfuric acid mixture titration method. Total N (TN) in the plant and soil was measured using the semimicro-Kjeldahl method with a Kjeldahl auto-analyzer (KjeltecTM 8400, FOSS, Denmark). Alkaline hydrolysis diffusion method was used to determine soil alkali nitrogen (AN). Total P (TP) in the plant and soil was measured by a spectrophotometer (UV-2450, Shimadzu, Japan). Soil available P (AP) was extracted in 0.5 mol·L−1NaHCO3 and the concentration was determined using the Olsen method. Soil available K (AK) was extracted in 1mol·L−1 CH3COONH4, and measured by Atomic Abserption Spectrometer (PinAAciie 900F, PE, US).
The C:N:P stoichiometric ratios of plant were calculated as TC vs. TN (C:N), TC vs. TP (C:P) and TN vs. TP (N:P). As for stoichiometric ratios in the soil, they were calculated as OC vs. TN (C:N), OC vs. TP (C:P) and TN vs. TP (N:P).
The determination method of soil urease was based on phenol-sodium hypochlorite sodium colorimetry, the 0.1 N KMnO4titration method was used to measure soil catalase activity and invertase was measured by 3,5-Dinitrosalicylic acid colorimetry (Xiao et al., 2018, Zhen et al., 2018). Soil microbial biomass C (MBC) was measured by Total Organic Carbon Analyzer (TOC-VCPH, Shimadzu, Japan) after chloroform fumigation and 0.5mol·L−1potassium-sulfate extracted. Soil microbial biomass N (MBN) was measured by auto-analyzer (Auto Analyzer 3-AA3, SEAL, US) after chloroform fumigation and 0.5mol·L−1 potassium-sulfate extracted.

2.4. Statistical analysis

Analysis of the differences in element concentrations and their stoichiometric ratios in all age groups were using ANOVAs and least significant difference (LSD) multiple comparison (P  < 0.05). The linear correlations of C, N, P concentrations and stoichiometric ratio were analyzed with the modely= ax + b . The correlation coefficient of aboveground and underground biomass with C, N, P content and stoichiometric ratio was calculated by Person correlation analysis.