Operational Robustness of Drinking Water Treatment Plants with Respect to Turbidity Representing Normal, Severe and Unprecedented Weather Events

Abstract

Drinking water treatment plants (DWTPs) are required to supply safe drinking water continuously to the consumers to protect public health and sanitation. The adverse effects of climate change can influence raw water quality, which is likely to worsen in the future as predicted by numerous climate models. The intensity, frequency and duration of precipitation events have been observed to be changed throughout the world as a consequence of natural and anthropogenic climate change. Severe and untimely precipitation events have the potential to deteriorate the water quality in surface water bodies directly and have been associated with water-borne diseases. Many DWTPs in Canada use surface water as their raw water source. Heavy precipitation events can lead to a significant increase in suspended and dissolved particles in surface water bodies by fluvial erosion and transportation of particles, which can result in raw water with elevated turbidity at the intake. Most of the DWTPs are designed based on historical data including past weather events. However, with the rapid change in precipitation patterns leading to very high turbidity levels in raw water more frequently, it can be quite challenging for the DWTPs to maintain regulated water quality during these heavy storm events.

To control such turbidity spikes in raw water, a DWTP should be robust. Robustness of DWTPs is defined as the ability to provide excellent performance under normal conditions and deviate minimally during periods of upsets and challenges, maintaining a set finished water quality. The robustness of the affected treatment steps needs to be quantified to evaluate the robustness of the DWTPs under normal and historical weather scenarios and be improved for future weather scenarios that may occur due to climate change. A robustness framework was applied to two full-scale DWTPs (Plant A and Plant B) from Southern Ontario to assess their robustness with respect to turbidity for three raw water scenarios: (a) baseline turbidity representing normal weather, (b) elevated turbidity representing historical precipitation events, and (c) extremely high turbidity representing future precipitation events that is beyond general experience. For evaluating scenarios (a) and (b), on-line turbidity data for the calendar years 2019 and 2020 were provided by the two plants which have different raw water sources and treatment methods. To quantify the robustness of the affected treatment steps for turbidity removal, the turbidity robustness index (TRI) was used. A lower value of TRI is desired as it implies that the treatment step was robust for the given period. Using the TRI method has the advantage to quantify the robustness of treatment units with one index and one classification system irrespective of the different geographic locations, raw water sources, treatment techniques, and intensity and duration of precipitation events experienced in the two DWTPs. The weekly TRIs were calculated for each unit of the selected treatment steps during normal weather conditions using the on-line data. A method was developed to distinguish the elevated turbidity events representing heavy precipitation from background turbidity data and the TRIs corresponding to these periods were separated. However, no correlation was observed between higher TRIs and weather events characterized by elevated raw water turbidity, which is an indication of robustness with respect to raw water turbidity. The overall robustness of the two plants was assessed during the study period. Plant A was found to be more robust than Plant B in general. The higher TRIs observed in both plants can be a good tool to evaluate their operational regime retroactively and improve the robustness of the treatment steps.

To assess scenario (c), the full-scale coagulation and sand ballasted clarification (SBC) process of Plant A was simulated using modified bench-scale jar tests where spiked water samples with very high turbidity were assessed in addition to controls at normal turbidity. A factorial design experiment was conducted to determine the significant factors for turbidity removal and optimize the process. The outcome of these experiments suggested that the polymer dosage used in the plant is optimum for extremely high turbidities, but the coagulant and microsand dosages can be increased for better removal. The outcome of the bench-scale simulation can aid in potential pilot- or full-scale studies.

This study focuses on elevated raw water turbidity caused by heavy precipitation events. It is recommended to explore the effects of other climatic events on various raw water quality parameters to evaluate and improve the robustness of DWTPs.