1.3 Statistical analysis
The original data of SMBC and SMBN was ln-transformed to improve the normality of the dataset. Violin plot and column chart was used to report the mean and standard error graphically. One-way analysis of variance (ANOVA) followed by Duncan’s post hoc tests were used to detect differences in the amount of SMBC, SMBN, the ratio of SMBC to SMBN among the three drought classes and four ecosystem types (a level of p< 0.05 was accepted as significant). Scatter plots and linear fitting are used to report the relationship and trend between two parameters. Pearson correlation coefficient was used to verify the correlation between two factors (significant level: * means 0.05, ** means 0.01). Besides, Pearson correlation coefficient test, univariate analysis of variance and one-way analysis of variance test were used IBM SPSS Statistics 23 software. Origin 9.0 was used to generate graphs and Arc GIS 10.7 was used to generate the map of sample sites.