2.2 Experimental design
Our field study was conducted in May 2019 in fenced grassland and had a complete randomised block design. This field experiment was conducted in five 2 m × 23 m blocks. We created 40 plots, each measuring 2.2 m × 2.2 m, in each of these blocks, and then allocated one of eight different treatment regimens to each of the plots. The buffer zones around the plots and blocks measured 1 m. The following treatments were combined in a cross factorial experiment: application of 0 or 10 g N m−2·year−1(NH4NO3), application of 0 or 10 g P m−2·year−1[Ca(H2PO4)2] and application of fungicide to suppress mycorrhizal fungi [12 g benomyl (50% active substance) in 10 L of deionized water per m2] or plain water. The levels of N and P added exceeded the natural deposition rates of these nutrients in East Asia (Cui et al., 2020b; H. Yang et al., 2012), but conformed to predictions of N and P deposition from land use change and atmosphere (Cui et al., 2020b). In the early summer months, NH4NO3 and Ca(H2PO4)2 were inputted by manual work and distributed after sunset (when soil moisture levels were higher). The mycorrhizal suppression plots were treated with fungicide twice per month in growing season, and the mycorrhizal plots were inputed through the same amount of plain water at the same time points. Benomyl was used as a fungicide because it can effectively reduce AM fungal colonisation and extra-radical biomass, but it has slightest side effects on host plants and other soil microorganisms in native grasslands (Jia, Walder, Wagg, & Feng, 2021; Qiao et al., 2019; Xue Yang, Mariotte, Guo, Hautier, & Zhang, 2021).