Nonlinear response of soil microbial network complexity to long-term nitrogen addition in a semiarid grassland: Implications for soil carbon processes

Increased atmospheric nitrogen (N) deposition alters the structure and function of soil microbial communities in terrestrial ecosystems, consequently exerting a profound influence on ecosystem processes. However, the effects of N deposition on soil microbial network complexity and its regulation of soil carbon (C) processes in semiarid grassland ecosystems are poorly understood. In this study, based on a 13-year multilevel field N addition experiment in a semiarid grassland on the Loess Plateau, together with metagenomic sequencing and co-occurrence network analysis methods, we observed that the complexity of microbial co-occurrence network, characterized by the number of nodes and edges and the average path length, increased first and then decreased in a nonlinear response to N addition, with thresholds between 4.60 g N m⁻² yr⁻¹ and 9.20 g N m⁻² yr⁻¹ in both the topsoil and subsoil. Meanwhile, soil microbial network complexity was significantly positively correlated with plant root traits (e.g., root biomass), soil microbial properties (e.g., fungal community composition and bacterial Shannon diversity and community composition) and most physicochemical properties (e.g., soil water content, NH₄⁺-N, and Fe_p). Structural equation model analysis (SEM) revealed that the major determinants of the soil microbial network complexity shifted from soil physicochemical properties to bacterial community composition along the N addition gradient. Further analysis revealed that N-induced alterations in microbial network complexity could modulate soil organic C (SOC) formation, preservation, and decomposition by affecting the functional potential of microbial communities. For instance, the microbial network complexity, abundance of functional genes involved in starch and hemicellulose degradation, and microbial C use efficiency decreased significantly under high levels of N addition. These results provide empirical evidence for the close linkages between soil microbial network complexity and soil C processes and highlight the need to disentangle the mechanisms underlying the nonlinear response of soil microbial interactions to atmospheric N deposition to improve soil C projections.