Wastewater treatment using a constructed wetland and recirculating gravel filter system

The use of constructed wetlands (CW) and recirculating gravel filters (RGF) as a wastewater treatment system was demonstrated at a municipal treatment plant. The demonstration was a proof-of-concept test for the use of RGFs to treat CW effluent for ammonia. Inadequate ammonia conversion by nitrification is characteristic of most CWs in operation. Additionally, total nitrogen removal was planned by discharging the nitrified RGF effluent at the head of the CW which was typically anoxic. Three existing CWs (two free water surface and one subsurface flow), each with an area of 1.46 ha, were used to treat effluent from a 6.47 ha stabilization pond serving Benton, Kentucky, a small community with a population of 4,700. Recirculating gravel filters were designed having an area of 0.2 ha, a filter bed depth of 0.61 m, and pea gravel filter media with a nominal size of 7 mm. Based on design system flow of 3,785 m³/d and an RGF recirculation ratio, r, of 3, a recirculation flow, Q_r of 11,355 m³/d was planned with a resulting hydraulic loading rate (HLR) of 5.66 m³/m²-d. The HLR was doubled during testing to 11.3 m³/m²-d to allow alternating dosing of filter halves. At the test HLR and a specific filter bed ammonia nitrogen loading of 74.4 g/m³-d (74.4 kg/1000m³-d), the filter mass removal efficiency was ≈ 80 % with an influent concentration of about 5 mg/l and an effluent concentration of about 1 mg/l. For a full range of test conditions, the relation of mass loading to mass removal was well represented (r² = 0.92) by a linear regression. The system performance as measured from the stabilization pond effluent to the system discharge was also well represented by a linear model of mass loading versus mass removal for both TKN and BOD₅ (r² = 0.96 and 0.89 respectively). Also, TKN mass removal efficiency increased with increasing recirculation ratio from about 65 % at r = 2 to 90 % at r = 6. With r < 6, the removal efficiency declined. The TKN was primarily organic as algae in the stabilization pond effluent. Ammonification of this algae was assumed to occur in the wetland based on past performance published on this wetland where a net production of ammonia rather than reduction was frequently measured. The dominant flow entering the CW was the aerobic RGF effluent flow draining into the CWs. Apparently sufficient oxygen was supplied for nitrification of wetland ammonia by this mechanism. The nitrate concentrations in the discharge were typically below detection limits, indicating that the nitrified RGF effluent as well as any nitrified wetland products experienced denitrification and, therefore, total nitrogen removal. As the recircualtion ratio increased, internal or "per pass" efficiency of the SF CW/RGF subsystem was significantly reduced for both BOD₅ and TSS removal. This effect was at least partially because of greatly increased surface flow across the gravel bed that allowed bypassing of treatment area. At the same time TKN mass removal increased and mass removal efficiency was unaffected. Recirculating gravel filters were found to be an effective complement to constructed wetlands for removal of both ammonia nitrogen and total nitrogen. The RGF area was equal to only 5 % of the wetland area. This relatively small area and the simple design and materials of an RGF argues for their cost-effectiveness as compared to options that require significant increases in wetland area.