Nitrogen removal and microbial community responses in anammox-enhanced biofilters across diverse urban waters

Biofilters offer decentralized nitrogen treatment but often show variable performance across influent types. This study evaluated anammox-enhanced biofilters (A-BF, inoculated with enriched anammox biomass) and control biofilters (C-BF) treating wastewater (WW), stormwater (SW), and greywater (GW) under synthetic and real conditions for 54 weeks. Under synthetic conditions, A-BF outperformed C-BF, with the greatest improvement in ammonium-rich WW, where TN removal reached 65.7±5.1% compared with 17.2±2.4% in C-BF. This was supported by enrichment of Candidatus Brocadia (1.19%) and high anammox 16S rRNA gene abundance (4.9×10⁷ copies/g soil). In low-nitrogen SW, A-BF achieved 52.9±4.9% TN removal versus 34.5±3.8% in C-BF, despite limited anammox bacterial abundance (0.34%) and lower anammox 16S rRNA gene copies (4.5×10⁵ copies/g soil). In carbon-rich GW, A-BF reached 54.6±14.4% TN removal compared with 43.1±11.4% in C-BF, but lower anammox bacterial abundance and gene copies suggested that anammox-associated potential may have been constrained by heterotrophic competition. Under real influents, water-type-specific responses emerged. Under real WW conditions, TN removal in A-BF decreased to 47.4±4.2%, coinciding with lower NO₂^--N availability, higher Nitrospira abundance, and lower anammox bacterial abundance and gene copies, suggesting more complete nitrification and stronger competition for NO₂^--N. In contrast, real SW and GW maintained A-BF TN removal at 57.8±5.5% and 55.8±2.7%, respectively, likely through different compensatory mechanisms: substrate limitation was partly relieved in SW, whereas the lower COD in real GW likely reduced heterotrophic competition and slightly favoured anammox-associated bacteria. Overall, anammox-enhanced biofilters improved nitrogen removal, but their effectiveness was regulated by influent composition and microbial competition.