In a subsurface-flow constructed wetland the gravel/soil matrix is anoxic almost everywhere, but emergent macrophyte roots (e.g. Phragmites, Typha) transport O2 down through aerenchyma tissue and leak a fraction of it radially into the surrounding rhizosphere — radial oxygen loss (ROL). This creates thin, oxidised aerobic microzones only within a few centimetres of each root, while the bulk matrix stays anoxic.
Nitrification (aerobic, at the root surface):
NH4+ + 2 O2 -> NO3- + H2O + 2 H+
r_nit = k_nit [NH4+] · O2/(K_O2 + O2)
Denitrification (anoxic, needs a carbon donor):
5 CH2O + 4 NO3- -> 2 N2(g) + 4 HCO3- + CO2 + 3 H2O
r_denit = k_denit [NO3-] · C/(K_C + C) · (1 − O2 inhibition)
Net nitrogen removal needs a spatial cascade: an ammonium ion must first pass close enough to a root to be nitrified to nitrate in the aerobic microzone, then drift into the surrounding anoxic gravel where organic carbon lets denitrifiers reduce it all the way to N2 gas, which bubbles out to the atmosphere. Ammonium that never nears a root, or nitrate that never reaches carbon-rich anoxic gravel, simply leaves in the effluent unremoved.
- Radial oxygen loss — sets how far the aerobic microzone extends from each root; wider zones nitrify more NH4+ but leave less anoxic volume for denitrification.
- Organic carbon (BOD) — the electron donor denitrifiers need; too little and nitrate simply accumulates and exits as effluent NO3-.
- Hydraulic loading — faster flow means less residence time near roots and less nitrogen removed overall.
- Root density — more plant clusters mean more aerobic microsites distributed through the bed.
This coupled nitrification-denitrification pathway is the dominant nitrogen-removal mechanism documented in vegetated horizontal subsurface-flow constructed wetlands (Vymazal 2007; Brix 1997), and is why planted beds consistently outperform unplanted gravel filters for nitrogen.