Many constructed wetlands adequately treat BOD5, TSS, and bacteria. Ho
wever, a review of nitrogen (N) data from 52 constructed and natural w
etlands in the North American data base confirmed that N removal was v
ariable. Nitrification and denitrification require aerobic and anaerob
ic conditions. This paper presents case histories of systems that use
alternating shallow and deep water zones to create both environments.
Regression analysis of N removal and N loadings in 18 shallow-deep wat
er systems suggested that NH4+ loading (kg N/ha/day) could be used to
predict effluent NH4+ values. Combinations of shallow water-emergent v
egetation and deep water-submergent vegetation with low NH4+ (and TKN)
loading rates can produce very low levels of discharged NH4+.