A pilot study on municipal wastewater treatment using a constructed wetland in Uganda

The potential of using constructed v wetlands as a cheaper and yet effective alternative method for treating domestic wastewater in tropical environments was investigated in this study from May 1996 - April 1999. The major aim was to determine their technical viability with respect to treatment performance under different operating conditions and the economic competitiveness of the technology in Uganda and within the region. A pilot constructed wetland design, based on horizontal flow criteria and receiving pre-settled sewage from the Jinja Kirinya Sewage anaerobic lagoons was used in the study. The wetland had a total surface area of 32 m ² wich was sub-divided into eight individual units. Four of the units were planted with Cyperus papyrus floating without a substratum base and two with Phragmites mauritianus anchored on a substratum base. Two units were used as controls, one of them had a substratum base of similar volume as the planted ones. The wetlands were operated over three consecutive phases. In the initial phase, all planted units remained intact but in the 2 ^nd phase, plant biomass was removed from a quarter of the area of the two Cyperus papyrus and one Phragmites mauritianus wetland units. In the last phase, two wetland units of each plant type and one control (open) were joined in series; two other papyrus wetland units with smaller areas of alternating planted and unplanted sections were joined in series. The hydraulic loading rates applied to the different units ranged between 1.3 - 12 cm day ^-1 over the whole experimental period.

The shoot density and size of open areas controlled the physical-chemical conditions in the wetland units. When the shoot density was greater than 50 shoots/m ² , low oxygen concentrations < 2 mg/l, pH values of 7 - 7.5 and water temperatures of 22 ± 0.5 °C prevailed. In the open ponds with relatively larger open areas, peak oxygen concentration of 27 mg/l, pH of 10.4 and water temperature 29 °C were obtained. In the reduced open surface areas applied in the 2 ^nd and 3 ^rd phases, peak oxygen concentration of 12 mg/l, pH values of 7.5-8 and temperature of 23 °C were registered. These conditions influenced the extent of reduction of the pollutants achieved in each case. Removal efficiencies in both C. papyrus and P. mauritianus wetland units' of just over 70% of the input settled COD (maximum 350 kg ha ^-1 ) and settled BOD (maximum 100 kg ha ^-1 ) were obtained when the shoot cover was intense. It increased to over 80% in the third phase. TSS reduction above 80% of the input (maximum 250 kg ha ^-1 ) was obtained in all the vegetated wetland units and in all the phases. Similar findings were observed in the household wetland. A significant faecal coliform removal of 4 log units ivas obtained in the control ponds as compared to 3 log units derived trom the vegetated wetland units or 1.13 log units from the household wetland. The faecal coliform reduction in the planted units was correlated with TSS and particulate organic matter removal. The lethal impact of direct sunlight and its secondary effects such as high pH were considered more significant in ponds with open areas. All the results show that Ciperus papyrus planted units performed better than Phragmites mauritianus units. The low residual background BOD radiation of 12 mg/l in papyrus units as compared to 17 mg/l in Phragmites units further confirmed the observed trends. In addition, the rate of BOD decay in papyrus units was higher as indicated by the areal first order rate constant (0.084 m d ^-1 ) than in Phragmites (0.039 m d ^-1 ). Throughout, the effluent concentrations of these parameters from the vegetated wetland units were consistently below the Uganda wastewater discharge standards.

Nutrient (N & P) removal from the wastewater via plant uptake showed extreme variability at different growth phases; uptake was correlated with the biomass yields exhibited in the different phases. Uptake rates of 7.1 kg N ha ^-1 day ^-1 and 0.24 kg P ha ^-1 day ^-1 in papyrus and 10.4 kg N ha ^-1 day ^-1 and 0.26 kg P ha ^-1 day ^-1 in P. mauritianus were derived in the exponential growth. Mass balance considerations over exponential phase showed plant uptake contribution of 15% N and 10 % P of the total input to papyrus wetland and 58% N and 37% P of the total input to P. mauritianus units. This contribution declined to less than 4% for both plant types at the slower growth phase. It is concluded that nitrogen and phosphorus removal via plant uptake is only significant at the exponential growth phase and more so in P. mauritianus which had nearly 90% of its total biomass as above ground. Nitrogen removal by other routes was influenced by the environmental conditions. Low oxygen concentrations which, were prevalent when the shoot density was high, minimised ammonium removal via the sequential nitrification-denitrification pathway. The pH (<10) and temperature (max 29 °C) characterising the open units favoured ammonium loss by volatilisation of ammonia gas. In the papyrus wetland series with small alternating vegetated and non-vegetated zones applied in the third phase, up to 90% of total ammonium N input (26 kg N ha ^-1 ) was removed, 77% of this loss was attributed to nitrification - denitrification process. Effluent with ammonium concentration below the Uganda regulatory discharge limit of 10 mg/l was only obtained in the last phase. The wetland serial configuration applied in this phase is concluded to be the most suitable for nitrogen removal from wastewater in the tropical environments.

The economic viability of using constructed wetlands in Uganda was deduced from the total annua' costs of the wetland and the waste stabilisation ponds designed for a population equivalent of 4000. The total annual cost for waste stabilisation ponds was 21 % more than that of the constructed wetlands (US $ 11,400). The recurrent cost or both systems were similar but nearly eight times lower than that estimated for conventional treatment systems.

Based on the overall results of the treatment performance and economic costs, it is concluded that application of constructed wetlands in Uganda and in the sub-region can be considered both technically and economically viable option for municipal wastewater treatment.