Properties and ecological assembly mechanisms of microbial communities across salinity levels in coastal saline-alkaline soils

Microorganisms play key roles in nutrient cycling and soil remediation in saline-alkaline ecosystems. However, how their assembly rules and interaction networks change during natural succession remains unclear. We used a space-for-time substitution approach to investigate soil microbial communities along a salinity chronosequence in the Yellow River Delta, covering three successional stages: high salinity (HS), medium salinity (MS), and low salinity (LS). We performed 16S rRNA and ITS amplicon sequencing on 24 composite soil samples and analyzed community composition, functional traits, assembly processes, and biotic association networks. The results showed that, environmental filtering—especially by soil organic carbon (SOC) and electrical conductivity (EC)—was the primary driver of community structure. Functionally, low-salinity soils were enriched in nitrogen-cycling genera (e.g., Nitrospira) and arbuscular mycorrhizal fungi, whereas high-salinity soils favored saprotrophs and sulfur-metabolizing taxa. Importantly, null-model analysis revealed a fundamental shift in assembly mechanisms. In harsh high-salinity habitats, stochastic processes dominated community assembly (normalized stochasticity ratio, NST > 0.5), and bacterial communities formed highly positive, cooperative networks. By contrast, as salinity decreased, deterministic processes prevailed (NST < 0.5). Meanwhile, negative edges increased significantly (P < 0.05), suggesting that relaxed environmental stress intensified interspecific competition and exclusion. Cross-domain network analysis further indicated that under high-salinity conditions, bacteria and fungi adopted a more tightly coordinated survival strategy. These findings suggest that restoration of saline-alkaline land should not only reduce salinity, but also steer the balance between deterministic selection and stochastic colonization to build stable, multifunctional microbial networks. This provides a theoretical basis for the targeted development of salt-tolerant microbial inoculants and for optimizing vegetation restoration schemes.