Additional information
Supplementary Information accompanies this paper at https://doi.org/…
Fig. S1 Correlation between NP-induced changes in soil total N content and NP-induced changes in plant N content. ΔPTN equals plant total nitrogen under concurrent N+P treatment minus plant total nitrogen under N only condition. Additions; ΔSTN equals soil total nitrogen under concurrent N+P treatment minus soil total nitrogen under N only condition. The log-transformation was used to producing variables following normal distribution.
Fig. S2 Bivariate correlations between variables used in structure equation models (observational dataset)
Fig. S3 Structural equation models of meteorology, plant, edaphic and microbial biomass as predictors of CH4 uptake for (a) control (n=96, χ2=2.041, DF=2, CFI=1, NFI=0.99), (b) N addition only (n=96, χ2=0.667, DF=2, CFI=1, NFI=0.997), (c) P addition only (n=96, χ2=1.946, DF=2, CFI=1, NFI=0.991) and (d) concurrent N+P additions (n=96, χ2=2.079, DF=2, CFI=1, NFI=0.99) (black lines represent positive paths; red lines represent negative paths; solid lines indicate significant impacts [p<0.05]; dotted lines indicate non-significant impacts [p>0.05]).
Fig. S4. Spatial distribution of the data points for the whole dataset used for meta-analysis; 382 data points distributed across the global grasslands, covering major grassland types, spanning major climate zones
Fig. S5. The correlation matrix of the key variables in the dataset used for meta-analysis (noted most variables are not in normal distribution, indicating the needs of log-transformation in further analysis and model development)
Fig. S6. Correlation between NH4+ and total inorganic N (NH4++NO3-).
Fig. S7. Comparison of modeled CH4 with the compiled CH4 uptake.
Fig. S8. Frequency distribution of the best estimates of CH4 oxidation across global grasslands (CK: pre-industrial N and P; N: the model simulation with pre-industrial P and contemporary N inputs; P: N: the model simulation with contemporary P and pre-industrial N inputs; NP: model simulations with contemporary N and P depositions).
Table S1 Correlation coefficients for the regression between independent variables and CH4 uptake rates based on the comprehensive data (R2 = 0.44; P < 0.001); it is used as the empirical model for global estimation of N and P impacts on CH4 uptake in grasslands
Competing interests: The authors declare no competing interests.