1

Electronic Supplementary Materials

Table S1. Results of split-plot ANOVA for the effects of grazing (the whole plot treatment) and removal (the subplot treatment) on the variables of temporal stability (ICV and TSI), diversity, and dominance hierarchies. Here, we also presented results of split-plot ANOVAfor the effects of grazing and removal on the relative abundance of B. gracilis and B. dactyloides, to assess the contribution of changes in the relative abundance of these species by removal treatment to the change in the relative abundance of dominant species.Significant P values are in bold.

Variable / Source / df / MS / F / P
Temporal stability
Inverse CV
Site / 5 / 1.100 / 1.417 / 0.265
Grazing / 1 / 0.001 / 0.002 / 0.964
Wholeplot error / 5 / 0.288
Removal / 1 / 10.271 / 47.886 / < 0.001
Grazing × Removal / 1 / 0.116 / 0.541 / 0.479
Subplot error / 10 / 0.215
TSI (temporal stability index)
Site / 5 / 0.831 / 2.102 / 0.112
Grazing / 1 / 0.015 / 0.128 / 0.735
Wholeplot error / 5 / 0.119
Removal / 1 / 4.604 / 24.500 / 0.001
Grazing × Removal / 1 / 0.016 / 0.087 / 0.774
Subplot error / 10 / 0.188
Diversity
Species richness
Site / 5 / 3.035 / 4.127 / 0.011
Grazing / 1 / 2.074 / 5.561 / 0.065
Wholeplot error / 5 / 0.373
Removal / 1 / 4.307 / 8.650 / 0.015
Grazing × Removal / 1 / 0.016 / 0.031 / 0.863
Subplot error / 10 / 0.498
Species diversity
Site / 5 / 0.016 / 8.109 / < 0.001
Grazing / 1 / 0.004 / 1.354 / 0.297
Wholeplot error / 5 / 0.003
Removal / 1 / 0.001 / 0.563 / 0.470
Grazing × Removal / 1 / 0.000 / 0.027 / 0.874
Subplot error / 10 / 0.002
Species evenness
Site / 5 / 0.010 / 3.596 / 0.020
Grazing / 1 / 0.000 / 0.090 / 0.776
Wholeplot error / 5 / 0.002
Removal / 1 / 0.005 / 1.684 / 0.224
Grazing × Removal / 1 / 0.001 / 0.308 / 0.591
Subplot error / 10 / 0.003
No. of rare species
Site / 5 / 1.557 / 3.117 / 0.034
Grazing / 1 / 2.241 / 16.098 / 0.010
Wholeplot error / 5 / 0.139
Removal / 1 / 2.817 / 8.847 / 0.014
Grazing × Removal / 1 / 0.051 / 0.162 / 0.696
Subplot error / 10 / 0.318
Functional diversity
Site / 5 / 0.137 / 6.367 / 0.001
Grazing / 1 / 0.032 / 3.347 / 0.127
Wholeplot error / 5 / 0.010
Removal / 1 / 0.000 / 0.010 / 0.921
Grazing × Removal / 1 / 0.028 / 1.021 / 0.336
Subplot error / 10 / 0.028
Dominance hierarchy
Relative abundance of dominant species
Site / 5 / 0.012 / 1.253 / 0.326
Grazing / 1 / 0.000 / 0.008 / 0.933
Wholeplot error / 5 / 0.002
Removal / 1 / 0.142 / 68.356 / < 0.001
Grazing × Removal / 1 / 0.008 / 3.863 / 0.078
Subplot error / 10 / 0.002
Relative abundance of subdominant species
Site / 5 / 0.023 / 2.859 / 0.045
Grazing / 1 / 0.004 / 1.290 / 0.308
Wholeplot error / 5 / 0.003
Removal / 1 / 0.066 / 11.804 / 0.006
Grazing × Removal / 1 / 0.001 / 0.179 / 0.681
Subplot error / 10 / 0.006
Relative abundance of rare species
Site / 5 / 0.010 / 2.386 / 0.079
Grazing / 1 / 0.005 / 1.126 / 0.337
Wholeplot error / 5 / 0.004
Removal / 1 / 0.011 / 3.756 / 0.081
Grazing × Removal / 1 / 0.013 / 4.429 / 0.062
Subplot error / 10 / 0.003
Relative abundance of Bouteloua gracilis
Site / 5 / 0.009 / 0.991 / 0.451
Grazing / 1 / 0.000 / 0.185 / 0.685
Wholeplot error / 5 / 0.001
Removal / 1 / 0.121 / 48.130 / < 0.001
Grazing × Removal / 1 / 0.004 / 1.657 / 0.227
Subplot error / 10 / 0.003
Relative abundance of Buchloë dactyloides
Site / 5 / 0.025 / 32.191 / < 0.001
Grazing / 1 / 0.000 / 0.620 / 0.467
Wholeplot error / 5 / 0.000
Removal / 1 / 0.000 / 0.800 / 0.392
Grazing × Removal / 1 / 0.001 / 0.565 / 0.470
Subplot error / 10 / 0.001

Fig. S1. Relationships between TSI (temporal stability index) and each variable of diversity (species richness, species diversity, species evenness, number of rare species, and functional diversity; a-e, respectively) and dominant hierarchy (relative abundance of dominant, subdominant and rare species; f-h, respectively) using the subset data from 2002 to 2006 (see Materials and methods). Solid lines indicate significant linear fits from single-variable mixed-effects models (P < 0.05). P values were obtained by using Markov chain Monte Carlo simulations. Higher TSI indicates greater stability.