Improving modelling and prediction of the fate and removal of micropollutants during wastewater treatment

Treated municipal wastewater from Western Treatment Plant (WTP) is regarded as a sustainable and reliable source of recycled water. The plant, situated in Victoria, Australia, utilises activated sludge followed by sequential lagoon treatment. However, the presence of organic micropollutants (MPs) and other harmful compounds in the treated wastewater may lead to potential risks to public and environmental health. A quantitative risk assessment (QRA) program for the plant has the aim of identifying and evaluating the potential risks associated with MPs in the treated water. This study comprises part of this QRA program, and its primary aim was to develop and improve fugacity-based mass balance models for the prediction of the fate and removal of seven representative MPs (2,4-D, caffeine, carbamazepine, diuron, simazine, sulfamethoxazole and triclosan), in the combined activated sludge-lagoon system. To improve the current QRA fugacity model for the activated sludge process in the system, the biodegradation rate constants of MPs currently estimated from computer models were replaced with values determined experimentally in this study. Activated sludge and wastewater samples were collected during different seasons and the biodegradation constants obtained for MP spiked samples. The use of these data revealed the limited accuracy of the original CHEM-R WTP model. The original model predictions overestimated the biodegradation of carbamazepine and diuron, and significantly underestimated the biodegradation of simazine and triclosan, such that their biodegradation rate constants should be increased or reduced by 10 to 100 fold, respectively. The results also provided indications for further adjustment, such as incorporating the temperature dependence of the biodegradation of MPs in the model. In the development of a modified Quantitative Water Air Sediment Interaction (QWASI) model for the sequential lagoons, a lagoon was regarded as a four-layer system comprising the photolytic layer, the water column, the surface sediment layer and the buried sediment layer. The fate of the MPs was modelled in terms of (i) their sunlight-induced photodegradation in the photolytic layer, (ii) their transformation in the water column and the surface sediment layer, and (iii) their adsorption to the sediment in the surface sediment layer. The natural sunlight-induced photodegradation rate constants for the seven MPs in lagoon water were obtained bimonthly over a year to provide seasonal kinetic data to enable more representative model inputs. Biotransformation rate constants and sorption coefficients for the MPs were obtained using lagoon wastewater and sediment samples. Seasonal trends were observed for the photodegradation rates of all seven MPs in both pure water and the secondary effluent. In summer, the photodegradation rates obtained in the wastewater were approximately twice those in winter for caffeine and 2,4-D, and approximately 60% higher for diuron. Sunlight intensity was the primary influencing factor for triclosan and sulfamethoxazole, while for the other five compounds the wastewater composition also played a significant role. The depth at which these MPs occur in the wastewater lagoon also played an important role, the attenuation of sunlight with depth meaning that photolysis of these MPs is likely only within 10 cm of the water surface. Transformation rates of the MPs in the lagoon wastewater and sediment were determined using samples from the first pond downstream of the activated sludge process (Pond 5) in summer and autumn, when the greatest transformation of the MPs was expected. Much lower transformation rates than the biodegradation rates obtained during activated sludge treatment and photodegradation rates obtained in the photolysis study were obtained for all MPs in both lagoon wastewater (ranging from 0.001 to 0.005 h-1) and sediment samples (ranging from 0.003 to 0.006 h-1) for both sampling events. Overall, higher transformation rates were obtained in the sediment samples than in the lagoon wastewater, the highest was for caffeine. Organic carbon based sorption coefficients for the sediment samples collected from Pond 5 were determined for all MPs. Triclosan exhibited a significantly higher (by 100 fold) tendency to adsorb to the sediment than the other MPs. The sorption behaviour of the uncharged MPs, namely caffeine, carbamazepine and diuron, was dominated by the hydrophobic interactions between the compounds and the organic content in the sediments. Other mechanisms, such as electrochemical interactions, may be involved in the adsorption of the charged MPs. The higher tendency of triclosan to attach to the sediment could be attributed to a combined effect of the hydrophobic and electrostatic interactions. Using the experimentally obtained model inputs, a fugacity based QWASI model was initially developed for one of the six lagoons in the system and then the model was expanded for the sequential lagoon system. The single-lagoon model predicted that at least 65% of the concentration of the MPs remained in the outflow of the lagoon after treatment. The greatest removal was predicted for triclosan (35%) and the least for carbamazepine (5%). Caffeine, diuron and simazine were primarily removed via water transformation while the removal of sulfamethoxazole and triclosan was attributed mainly to their photodegradation. The multi-pond model predicted high removal of triclosan and 2,4-D (up to 90%) and these predictions were in good agreement with the site sampling data collected for another lagoon system which has the same design and influent source as the system under study. Uncertainty and sensitivity analysis (using @RISK) for the single-pond model revealed that the lagoon inflow rate was the most important input factor influencing several key model outputs for all seven MPs. The fate of the MPs in the lagoon system was also significantly influenced by the water transformation rates (except for carbamazepine) of these compounds in the lagoon and their sunlight-induced photodegradation rates (except for 2,4-D). Although the water transformation rates were much lower than the other degradation rates, due to the large volume of the lagoon wastewater residing in the water column, a greater overall impact of the water transformation rates of the MPs on the model output was observed. Another contribution of this study was investigation of the sunlight induced photolysis mechanisms of the selected MPs in the lagoon wastewater and the factors influencing their photolytic fate. Experiments indicated that triclosan and sulfamethoxazole primarily degraded via direct photolysis whereas the other five MPs degraded primarily via indirect photolysis. Hydroxyl radicals were shown to account for approximately 32-70% of the overall removal of these compounds. The presence of nitrate promoted the photochemical loss of all seven MPs. While humic acid enhanced the photolytic degradation of caffeine, sulfamethoxazole and diuron, while it hindered the photodegradation of the other four compounds by absorbing the available irradiation energy and/or reforming the parent compound. It was shown that there was only a small increase (up to 15%) in photodegradation of the compounds at 25 °C compared with that at 10 °C in the simulated system. Overall, the outcomes of this modelling study will contribute to the assessment of the treatability of the MPs in the target system at the WTP. These results can also be used as a part of the QRA program to facilitate the risk assessment of the treated wastewater