This study was conducted to (1) determine appropriate conditions for nitrogen removal by a combined
constructed wetland (CW) system and (2) evaluate the effects of hydraulic loading rate (HLR), nitrogen loading
rate (NLR), effluent recirculation, plant uptake, and seasonal change on the removal efficiency of nitrogen. A
three-stage experimental CW system consisting of a vertical flow-gravel filtration (v-GF) bed without plants, a
horizontal-subsurface flow (h-SSF) bed planted with Iris australis, and a vertical-subsurface flow (v-SSF) bed
planted with Phragmites australis in series were fed with primary-treated domestic wastewaters. The beds
with plants produced effluents of better quality than the v-GF bed without plants. It was observed that the average
removal efficiencies increased as the HLR (from 100 to 30 L/m2 d) decreased and the recycle ratios (to 100%)
increased. The average removal efficiencies in the summer period were found higher of approximately 6% for
the v-GF and h-SSF beds and 11% for v-SSF bed. Nitrogen removal by plant uptake was very low and only
3% (Iris) to 7%(Phragmites) of the N-input compared to other mechanisms such as nitrification and denitrification.
The volume-base first-order nitrification and denitrification removal constants (1/d) were 0.388 and 0.785
for the h-SSF bed, and 0.412 and 0.293 for the v-SSF bed, respectively.