Biogeochemical and genomic drivers of groundwater DNRA: predictability of ammonium accumulation risk

Widespread ammonium (NH₄⁺-N) contamination in groundwater, with local concentrations exceeding 20 mg/L, challenges traditional nitrogen removal, which is dominated by denitrification. Dissimilatory nitrate reduction to ammonium (DNRA), a nitrogen-retaining pathway, competes with denitrification. However, the environmental conditions that promote DNRA in aquifers and its quantitative contribution to NH₄⁺-N loading remain critical knowledge gaps in groundwater nitrogen cycling research. We investigated how C/N ratios, Fe²⁺ concentrations, initial NO₃^--N, hydrochemical types, total dissolved solids (TDS), and pH influenced DNRA efficiency and its competition with denitrification. We found that DNRA efficiency peaked in organic-rich aquifers (15.70-26.69%) and was minimum in industrially and agriculturally contaminated groundwater (0.71%). High Fe²⁺, high initial NO₃^--N, and HCO₃^--type environments markedly promoted DNRA, whereas Cl^--type water inhibited it. Competition analysis revealed that elevated C/N ratios, Fe²⁺, and TDS increased the relative contribution of DNRA to nitrate reduction. Metagenomic analysis further demonstrated that high C/N ratios and Fe²⁺-rich conditions promoted DNRA dominance via selective enrichment of Enterobacteriaceae carrying the abundant nrfA gene. Conversely, high mineralization and Na-Cl water types drastically reduced DNRA efficiency by suppressing the expression of key functional genes. Furthermore, under intermittent nitrate input, NH₄⁺-N accumulated even with low DNRA efficiency, posing long-term water quality risks. We developed a DNRA efficiency prediction model using the Extreme Gradient Boosting algorithm (R² = 0.92), thereby enabling accurate assessment across diverse groundwater conditions. This work advances mechanistic understanding and provides an innovative predictive methodology for assessing DNRA-driven NH₄⁺-N enrichment risks in groundwater.