Phosphate and ammonium removal from waste water, using constructed wetland systems
Phosphorus and nitrogen in waste water from sewerage systems contribute to excessive nutrient enrichment of surface waters, presenting a threat to nature conservation, domestic and industrial water supplies, and recreation. The general objective of this research was to investigate phosphate and ammonium removal from waste water by constructed wetland systems (CWS), which are increasingly being used for low-cost water treatment. Phosphate (P) adsorption capacity and other properties of potential CWS substrate materials (bauxite, shale, burnt oil shale, limestone, zeolite, light expanded clay aggregates (LECA) and fly ash) were investigated. Fly ash and shale had the highest P adsorption values, which were significantly correlated with porosity and hydraulic conductivity. Longer - term experiments with shale and bauxite gave maximum P uptake values of 730 and 355 mg P kg- I, respectively. Phragmites australis (common reed) seedlings grew satisfactorily in shale, bauxite, LECA and fly ash. Shale was selected as the most suitable substrate, and used in a pilot-scale CWS in plastic tanks in a greenhouse, with and without P. australis, at two input nutrient concentrations and a loading rate of 0.02 m3 m-2 d-1. Both planted and unplanted systems removed 98 - I 00% of P from a synthetic sewage over ll months. Removal of ammonium N was also complete in the planted tanks, but only 40 - 75% was removed in the unplanted ones. Corresponding nitrate N removal was 85 - 95% and 45 - 75%. The systems performed as well at high as at low concentration for both phosphate and ammonium. The variations in P and N removal could not be attributed to differences in pH, Eh and temperature, which did not differ significantly between planted and unplanted tanks. During the experiment, P and N concentrations were determined at 3 depths and 4 positions along the length of the tanks. H2P04- - P and NH/ - N concentrations were low ( < I. 0 g m-3) at all locations in the planted systems, whereas the P concentrations were sometimes twice as high in the unplanted ones. NH4 + -N in the unplanted systems was relatively high (l 0 - 30 g m"3) throughout the experiment. N03--N concentrations were very low by comparison. P. precipitation on shale and P. australis root and rhizome surfaces was examined by X-ray fluorescence analysis, and by chemical extraction with ammonium acetate, 0.1 M HCI and 2%> citric acid. This showed that P, Fe and Al had precipitated on all these surfaces. However. it was not possible to quantify the surface deposits, and further research is necessary. The hydraulic residence time. flow characteristics and permeability of the shale was investigated by a bromide tracer. The tracer breakthrough curves showed a similar pattern in all tanks, with ca 66% of the flow occurring through the bottom zone. However, the actual hydraulic residence time (6 days) was slightly higher than the theoretical one Although there was a significant difference (p < 0.02) between the distribution of flow in planted and unplanted tanks. there was no reduction in the reactive pore volume observed in any of the tanks. This confirmed that shale has good permeability properties. Monitoring of the full-scale systems was carried out during June - September 1995. Although P removal in a planted bed was between 50-75%, the overall performance of the full-scale systems was disappointing, especially for ammonium removal. This was attributed to high loading rates, visibly non-uniform flow and clogged gabions. A bromide tracer study carried out on these systems confirmed the short hydraulic retention times and heterogeneous flow mechanisms in both the unplanted and planted systems. Results obtained from the pilot scale study do not necessarily provide a quantitative prediction of the performance of larger-scale systems. However, the potential value of a shale-based system has been demonstrated, and this opens a new direction in the design of CWS; most systems built to date in the UK use gravel as a substrate. Shale has proved to have superior properties for P removal and is cheap and readily available in Scotland. Its application as a substrate in a full-scale system remains a subject for further investigation.