Context-dependent links between microbial co-occurrence topology and nitrogen mineralization in two adjacent subalpine Picea asperata stands
摘要
Nitrogen (N) mineralization underpins plant-available N and forest productivity. In subalpine forests, root-associated fungal indicators and microbial association patterns may co-vary with belowground N dynamics, yet how legacy and localized disturbance contexts shape these relationships remains unclear.
MethodsWe compared two adjacent Picea asperata stands in western Sichuan, China, representing a natural forest–plantation legacy contrast, and applied a localized root-trenching treatment to reduce root-derived inputs within each stand. We integrated phospholipid fatty acids (PLFA) profiling, shotgun metagenomics, extracellular enzyme assays, litter decomposition, and net N mineralization measurement.
ResultsRelative to the natural forest, the plantation showed lower denitrification and assimilatory nitrate transport gene abundances, lower net N mineralization, and higher C- and N-acquiring hydrolytic enzyme activities. Trenching responses were forest-type- specific: in the natural forest, trenching was associated with increased path length and lower nitrification; in the plantation, it co-varied with higher values of several co-occurrence topology metrics and elevated oxidative enzyme activities. However, most measured microbial functional groups and most N-metabolism-related gene categories showed limited trenching-related differences. Nonetheless, structural equation modeling described context-dependent statistical associations: higher N-acetyl-β-D-glucosaminidase activity was associated with lower net N mineralization in the forest-type contrast model, whereas lower co-occurrence network modularity was associated with lower net N mineralization in the trenching-based model.
ConclusionsBoth the legacy and local disturbance context was associated with different microbial indicators and co-occurrence patterns linked to selected N-related responses. These stand-specific findings reflect context-dependent statistical associations and require broader multi-site replication before generalization.