Soil bacterial communities are pivotal in regulating terrestrial biogeochemical cycles and ecosystem functions. The increase in global nitrogen (N) deposition has impacted various aspects of terrestrial ecosystems, but we still have a rudimentary understanding of whether there is a threshold for N input level beyond which soil bacterial communities will experience critical transitions. Using high‐throughput sequencing of the 16S rRNA gene, we examined soil bacterial responses to a long‐term (13 yr), multi‐level, N addition experiment in a temperate steppe of northern China. We found that plant diversity decreased in a linear fashion with increasing N addition. However, bacterial diversity responded nonlinearly to N addition, such that it was unaffected by N input below 16 g N·m⁻²·yr⁻¹, but decreased substantially when N input exceeded 32 g N·m⁻²·yr⁻¹. A meta‐analysis across four N addition experiments in the same study region further confirmed this nonlinear response of bacterial diversity to N inputs. Substantial changes in soil bacterial community structure also occurred between N input levels of 16 to 32 g N·m⁻²·yr⁻¹. Further analysis revealed that the loss of soil bacterial diversity was primarily attributed to the reduction in soil pH, whereas changes in soil bacterial community were driven by the combination of increased N availability, reduced soil pH, and changes in plant community structure. In addition, we found that N addition shifted bacterial communities toward more putatively copiotrophic taxa. Overall, our study identified a threshold of N input level for bacterial diversity and community composition. The nonlinear response of bacterial diversity to N input observed in our study indicates that although bacterial communities are resistant to low levels of N input, further increase in N input could trigger a critical transition, shifting bacterial communities to a low‐diversity state.

Aim: As the two largest components of plant detritus input, leaf and root litter together determine ecosystem vegetation turnover and nutrient cycling rates. However, the similarities and differences between the factors controlling their decomposition remain unknown. We evaluated the relationship between leaf and fine root litter decomposition across biomes and analysed how litter traits, climate, soil conditions and decomposers shape their relationship.Location: Global.Time period: 1984–2020.Major taxa studied: Vascular plants.Methods: We collected 352 paired leaf and fine root decomposition rates (k values) and ancillary traits, climate, soil condition and decomposer abundance data from 88 sites spanning the major global biomes. Boosted regression trees (BRTs) were applied to partition the factors that control root and leaf decomposition rates.Results: Averaged across all biomes, leaf litter decomposes significantly faster (k_leaf = .72) than fine root litter (k_root = .42). The BRTs indicated that plant traits best explained the variance in both leaf and root litter decomposition. The key chemical traits of leaf litter and fine root litter, including C:N, [P], N:P, [lignin], [cellulose], [non‐structural carbohydrates] and [tannins], were positively correlated. Therefore, leaf and fine root k values were positively correlated within and across biomes, even after removing the influence of climate, soil conditions and decomposers. However, climate and decomposers had different impacts on leaf and fine root decomposition. Climate induced a greater impact on fine root litter decomposition, whereas decomposers had a greater influence on leaf litter decomposition.Main conclusions: Our finding indicates that plants evolve a coordinated nutrient supply‐and‐demand strategy. The plants with high nutrient demand produce labile leaf and fine root litter, which decomposes fast to meet their high nutrient requirements. However, leaf and fine root decomposition are also mediated by different combinations of trait, climate, soil condition and decomposer factors, which weakens the coordination between leaf and fine root decomposition.

We investigated soil bacterial and fungal communities, constructed co-occurrence networks, and estimated bacterial traits along a gradient of nitrogen (N) input. The results showed that soil bacterial co-occurrence networks complexity decreased with increasing N input. The ratio of negative to positive cohesion decreased with increasing N input, suggesting the declined competitive but strengthened cooperative interactions. However, soil fungal network complexity did not change under N enrichment. In addition, N input stimulated the copiotroph/oligotroph ratio, ribosomal RNA operon (rrn) copy number, and guanine-cytosine (GC) content of soil bacteria, shifting bacterial life history strategy toward copiotroph with increased r-/K-strategy ratio. Piecewise structural equation modeling results further revealed that the reduction in bacterial co-occurrence network complexity was directly regulated by the increased bacterial r-/K-strategy ratio, rather than reduced bacterial richness. Our study reveals the mechanisms through which microbial traits regulate interactions and shape co-occurrence networks under global changes.