Soil carbon and inundation regimes jointly regulate denitrification potential in restored riverine floodplains

Denitrification plays a key role in nitrogen removal in restored floodplains, yet the relative importance of inundation regimes, hydrogeomorphic setting, and soil properties remains uncertain. In this study, we evaluated the environmental controls on denitrification potential across four restored sites using a combination of machine learning, structural equation modeling (SEM), and Copula-based analysis. Among the tested models, Random Forest provided the best predictive performance and explained 71% of the variance. Variable importance and partial dependence results identified total carbon (TC), soil moisture content (MC), and bulk density (BD) as the most influential predictors of denitrification potential. Specifically, a threshold in TC between 3% and 5% was associated with a sharp increase in denitrification potential. SEM results indicated that hydrological variables primarily influence denitrification indirectly through their effects on soil moisture and carbon accumulation, accounting for up to 44.4% of the variance at the site level. Copula analysis further quantified the joint probability of elevated denitrification enzyme activity (DEA) under combinations of favorable hydrologic conditions and sufficient carbon availability. Based on these patterns, the sites were grouped into three functional regimes: hydrology-carbon coupled regulation, carbon-limited regulation, and substrate-driven regulation with hydrological enhancement. Together, these results indicate that carbon availability establishes the biogeochemical potential for denitrification, whereas hydrological and physical conditions regulate its expression, with the relative importance of these controls varying among sites. The findings highlight the need for site-specific restoration strategies, including rebuilding soil carbon in degraded agricultural areas and maintaining appropriate inundation regimes while preserving favorable soil conditions in carbon-rich systems to enhance nitrogen removal.