Sampling and experiment
The soils were sampled from Gongga Snow Mountain, on the eastern edge of the Tibet plateau, Sichuan province, China. We selected five different ecosystems, bare soil (29°34’37”N, 101°59’25”E, 2961m), grass (29°34’48”N, 101°59’33”E, 2955m), shrub (29°34’60”N, 101°59’42”E, 2934m), deciduous forest (29°34’16”N, 101°59’55”E, 2924m), and coniferous forest (29°34’23”N, 102°00’06”E, 2893m) (Li & Xiong 1995; Jiang et al. 2018). Three samples were taken in each type ecosystem (total of 15). All samples were passed through 2mm sieves with large plant litters and small animal bodies (like insects and earthworms) picked out by hand. Samples were transported on ice to Beijing as soon as possible. The main vegetation species and types of the sampling sites were described in detail in previous studies (Li & Xiong 1995; Jiang et al. 2018).
Fifteen incubations were set up for each ecosystem (5 per soil sample). For each incubation, 5g (dry weight) of soil were incubated in 50ml centrifuge tubes under 60% water holding capacity and at 20°C. Every 12h the lids of the tubes were opened to allow complete exchange of the air and to adjust water content. Incubations were allowed to acclimate for 3 days before beginning the experiment. At the beginning of the incubation, we sacrificed 3 original samples (O) for each ecosystem, the other incubations were divided into high temperature stress treatment and control groups (6 incubation per group for each ecosystem). The high temperature treatment groups were transferred to 35°C for 3 days while the control groups were incubating under the original temperature condition. We then sampled 3 incubations from each ecosystem for treatment (T1) and control (U1) groups. The remaining temperature stress treatment incubations were returned to the original temperature condition (20°C) and the incubations for both treatment and control groups were incubated an additional 9 days. Then we sampled both treatment (T2) and control (U2) groups at the end 12thday.