Contaminant removal in a constructed wetland for industrial wastewater treatment.

MAINE, M. A.; HADAD, H. R.; SÁNCHEZ, G.; CAFFARATTI, S.; PEDRO, M. C.; MUFARREGE, M.; DI LUCA, G.; BONETTO, C.

Barcelona

Congreso; 3rd Wetland Pollutant Dynamics and Control; 2009

WETPOL

INTRODUCTION Constructed wetlands (CW) are often used for nutrient and organic matter retention in domestic and municipal sewage, storm water and agricultural runoff (Kadlec et al., 2000). The application of CW for industrial wastewater treatment is a promising alternative in Argentina, because the central and northern areas of the country have mild winters, allowing extended growing periods for plants, and the low population density determines the availability of marginal land around cities. A free water surface wetland was constructed at Bahco Argentina metallurgic plant, in Santo Tomé, Argentina. It has been in operation since 2003. The wastewater receives a primary treatment. However, it contains Cr, Ni and Zn and shows high pH and conductivity. An assemblage of locally common macrophytes was transplanted. Vegetation was initially dominated by Eichhornia crassipes (water hyacinth), but later replaced by Typha domingensis (cattail), since the end of the first year of operation. Efficient metal and nutrient retention throughout the first stages of vegetation development has already been reported (Maine et al., 2009). Sediment and biomass metal concentrations at the inlet increased significantly, giving raise to the question whether removal could be sustained on a long term basis. Standing dry shoots are harvested and superficial bottom sediments are pumped out in the inlet area by the end of each winter. The present contribution reports metal and nutrient retention in the last years of operation, when the vegetation was stabilized, and compares it with the retention at the initial vegetation development. METHODS A free water surface wetland 50 m long, 40 m wide and 0.5-0.8 m deep with a central baffle was constructed. Water samples were taken monthly at the inlet and outlet. BOD, COD, DO , total phosphorus (TP), soluble reactive phosphorous (SRP), NO3-, NO2-, NH4+, Na+, K+, Ca2+, Mg2+, Cl-,SO42-, alkalinity and Cr, Ni, Zn and Fe total concentrations were determined in water following APHA (1998). Sediment TP, Ni, Cr and Zn concentrations were determined in the inlet and outlet areas of the wetland. Sediment samples were collected in triplicate using a 4-cm diameter PVC corer, dried and analyzed following APHA (1998). Macrophytes were sampled monthly with a 0.50 x 0.50m square sampler. Four replicates were taken randomly at the inlet and outlet at each monthly sampling. Macrophytes were harvested, washed, separated between aerial parts and roots, dried and weighed. TP, Total Kjedahl Nitrogen (TKN), Cr, Ni and Zn in roots and leaves were determined in the same way as in the sediment samples. RESULTS AND DISCUSSION DO concentration was low at the inlet and lower at the outlet, close to exhaustion in roughly half of the samplings. Lower COD and BOD at the outlet suggest that mineralization of incoming organic mater caused the observed oxygen depletion and pH decrease. Ca2+ and alkalinity decrease at the outlet suggests calcium carbonate precipitation. NO3- and NO2- removal is likely to be caused by diffusion from the water column towards the anoxic sediments. Denitrification seems to cause the main nitrate loss. SRP and ammonia removal were low due to low OD. SO42- and Fe reductions at the outlet suggest pyrite formation in the bottom sediment. Cr, Ni and Zn were efficiently retained within the wetland. Bottom sediment concentrations were higher at the inlet than at the outlet and showed large variation ranges. Metal uptake by vegetation was also observed. Concentrations were higher in roots and rhizomes than in aerial parts, suggesting scarce translocation. Large differences were observed along the studied period without any discernible pattern. Maximum Cr and Ni root concentrations were attained simultaneously in July 2008 (1.9 and 1.2 mg/g dw, respectively), decreased later and remained low since then (0.1-1 mg/g Cr and 0.1-0.5 mg/g Ni). Selective allocation to senescent roots and shoots is hypothesized. T. domingensis showed a luxuriant growth attaining a maximum biomass (8 kg dw/m2) higher than in natural environments nearby (Maine et al., 2006). T. domingensis cover in the last two years ranged from 55% in late winter and early spring to 95% in summer. Present results are consistent with previous studies showing that complete vegetation development requires 3-5 years, while the CW performance improves with wetland maturity (Kadlec et al., 2000). CONCLUSIONS The CW efficiently retained metals and nitrate, the main N source. Retention efficiency was generally higher or similar than that reported for the previous stages of vegetation development (Maine et al., 2009). Plant biomass and cover attained stability showing only seasonal variations. Metal concentration in the biomass did not show an increasing trend over the last years of operation, while sediment concentrations at the outlet remained lower than at the inlet. Present results suggest that contaminant retention efficiency could be maintained for long under the present managerial practice.