Data collection
Data were collected from peer-reviewed journal articles. These articles were searched by ISI Web of Science, ScienceDirect, Springer, Wiley, and Google Scholar. The search terms were “(nitrogen addition OR nitrogen deposition OR nitrogen fertilization) AND (microbial biomass OR fungal biomass OR bacterial biomass).” If these papers were selected for further analysis, they must meet the following criteria: (1) Only original research papers were included. (2) Experiments were conducted in terrestrial ecosystems. Laboratory incubation and agro-system experiment were excluded, as it is hard to estimate the background N deposition level and ecological factors. (3) Data only related to control and N addition treatment were extracted. Data with treatments of warming, CO2, water or phosphorous addition plus N addition were excluded. (4) Means, sample sizes, and standard deviations or standard errors of soil microbial characteristics of both control and N fertilized treatments were extracted. (5) For N types, NH4Cl, (NH4)2SO4 and (NH4)3PO4 were classified as NH4+-N, while NaNO3, KNO3 and Ca(NO3)2 were classified as NO3--N. N fertilization rate was measured as N per unit area per year (kg N ha-1yr-1).
Soil microbial characteristics, such as total microbial biomass, fungal biomass, bacterial biomass, actinomycete biomass, saprophytic fungal biomass, arbuscular mycorrhizal (AM) fungal biomass, gram-positive (G+) bacterial biomass, gram-negative (G) bacterial biomass, as well as fungi to bacteria (F/B) ratio and G+ bacteria to Gbacteria (G+/G) ratio, were extracted. Data of field experiment such as N application rate (kg N ha-1 yr-1) and duration (yr) were also extracted. Information of sample site, such as ecosystem types, background N deposition level (kg N ha-1yr-1), mean annual temperature (MAT, °C) and mean annual precipitation (MAP, mm yr−1) were separately recorded as well. If the data of background N deposition of sample sites were missing, we used the “Nitrogen deposition onto the United States and Western Europe” dataset to estimate the background level (Hollandet al. 2005) and extracted the information fromhttp://nadp.slh.wisc.edu/. If the sample site was located in China, we extracted the information from “National Science & Technology Infrastructure” fromhttp://www.cnern.org.cn. If the MAT and MAP were not reported in the papers, we extracted the information from the database athttp://www.worldclim.org/ by latitude and longitude. Finally, our database included 1585 paired observations (Supplementary Online Material 1 ) from 178 articles (Supplementary Online Material 2 ) were classified into six ecotypes: desert, forest (including broadleaved and coniferous forest), grassland, tundra, shrub, and wetland around the world (Supplementary Online sMaterial 3 ).