https://doi.org/10.5061/dryad.brv15dvm1

Description of the data and file structure

We have submitted our raw data (dataset.zip). The original data corresponding to each figure (including the supplementary figures) are reflected in the table below.

• n/a: Not applicable;
• null: Cannot be calculated (divisor = 0);
• RII: Relative interaction intensity of shrub;
• G: Grazed conditions;
• UG: Ungrazed conditions;
• S: Shrub;
• I: Interspace;
• SE: Standard error;
• pH: Soil pH;

Figure 1 data

NMDS1 and NMDS2 data were obtained by analyzing bacterial and fungal communities using non-metric multidimensional scaling (NMDS).

Figure 2 data

Number of functions above the threshold represents the multithreshold multifunctional index under different thresholds.

Figures 3 and 5 data

Standardized direct and indirect effects generated by structural equation modeling.

Figure 4 data

The Relative Interaction Intensity (RII) of shrubs on soil microbial diversity, environmental factors, average multifunctionality, and the functional indices related to plant productivity, soil carbon (C), nitrogen (N), and phosphorus (P) cycling, as well as their mean ± 95% Confidence Interval (CI) under grazed and ungrazed conditions.

Table S1

Background characteristics of shrubs and interspaces under different grazing conditions at three sites (Mean ± SE).

Table S2

Effects of shrubs and grazing on the relative abundance of the top 10 soil bacterial taxonomic groups. Bold text indicates significant results (P<0.05).

• a denotes that the relative abundance of this taxonomic group significantly increased in shrub patches or under ungrazed conditions.
• b signifies that the relative abundance of this taxonomic group significantly increased in interspaces or under grazed conditions.

Table S3

Effects of shrubs and grazing on the relative abundance of the top 10 soil fungal taxonomic groups. Bold text indicates significant results (P<0.05).

• a signifies that the relative abundance of this taxonomic group significantly increased in shrub patches or under ungrazed conditions.
• b denotes that the relative abundance of this taxonomic group significantly increased in interspaces or under grazed conditions.
• The F-value refers to the ratio of the treatment effect to the error term in a linear mixed-effects model.
• The P-value represents the statistical significance of this effect in a linear mixed-effects model.

Table S4

Significance tests of shrubs and grazing effects on bacterial and fungal community compositions using three non-parametric multivariate analyses. The interaction of shrubs and grazing on bacterial and fungal community compositions was evaluated with the Adonis method.

• Adonis: Permutational multivariate analysis of variance.
• ANOSIM:  Analysis of Similarities.
• MRPP:  Multiple Response Permutation Procedure.
• R2: the proportion of total variation explained by the effect term.
• R: standardized measure of differences between and within groups, ranging from -1 to 1.
• δ： the actual observed average intra-group distance within the entire dataset.
• P : the significance of the difference.

Figure S2 data

The raw data of dim1 of the multidimensional multifunctionality, average multifunctionality, and weighted multifunctionality.

• G1S1 represents data from shrub patches under grazed conditions in site 1;
• G1I1 represents data from interspaces under grazed conditions in site 1;
• U1S1 represents data from shrub patches under ungrazed conditions in site 1;
• U1I1 represents data from interspaces under ungrazed conditions in site 1.

Figure S3 data

Bacterial and fungal β diversity based on their NMDS1.

Figure S4 data

Effect of shrubs on the microbial Shannon index, soil bulk density, and soil pH under grazed and ungrazed conditions. SE: standard error.

Figure S5 data

Effect of shrubs on the observed variables of plant productivity and soil carbon, nitrogen, and phosphorus cycling under grazed and ungrazed conditions. SE: standard error.

Figure S6 data

Bacterial and fungal β diversity based on their NMDS1. Number of functions above the threshold represents the multithreshold multifunctional index under different thresholds.

Figure S7 data

The raw data of average multifunctionality, soil bulk density, and soil pH.

Figure S8 data

The raw data of average multifunctionality and 24 bacterial classes.

Figure S9 data

Years represent grazing exclusion years.

Figure S10 data

The Relative Interaction Intensity (RII) of shrubs on significant bacterial predictors that are remarkably related to multifunctionality under grazed and ungrazed conditions (mean ± 95% CI).

Figure S11 data

• G: under grazed conditions;
• UG: under ungrazed conditions.