Multivariate Analysis of Disinfection By-products Formation in Drinking Water Distribution Systems

Drinking water distribution systems (DWDSs) have historically been looked upon like a black box, in which various chemical and microbial processes affect disinfectant residuals stability. This results in disinfectant residuals decay and unintended disinfection by-product (DBP) formations, which are associated with increased adverse health outcomes. To improve drinking water quality, in-depth understandings of disinfectant stability are required. A systematic literature review was undertaken to investigate disinfectant residuals stability affected by various influencing parameters. Nine important water quality and operational parameters were identified, including temperature, pH, water age, piping material, corrosion products, hydraulic conditions, microbial activity, and the type and dose of disinfectant residuals. A key conclusion from this review was that understanding and predicting DBP formation under a variety of conditions in DWDSs remain challenging due to complications of parameter co-dependence and feedback amplification of some key parameters. In order to address these complications, a multivariate analysis approach was adapted to disinfectant residual stability in chloraminated DWDSs. To support lab-scale studies, analytical methods were developed and optimised using gas chromatography tandem mass spectrometry to quantitatively determine a wide range of DBPs. Bayesian network models were assessed as an approach for predicting DBPs formation using water quality parameters routinely collected from 15 chloraminated DWDSs in Australia. These networks were applied to successfully determine the interrelationships among influencing factors and predicted DBP concentrations via backcasting and conditional scenario studies. However, as some water quality data (e.g. bromide concentration) were lacking from full-scale DWDSs, lab-scale experiments were needed to further investigate relationships between other water quality parameters and DBP formation. Lab-scale batch experiments were conducted using the Doehlert experimental design to evaluate the co-effects of important influencing factors on disinfectant stability in chloraminated drinking water. The optimal conditions for maintaining disinfectant stability were evaluated from the results. Furthermore, re-chloramination was found to be effective for reducing one class of chloramination DBP (N-nitrosamine) formation in a lab-scale DWDS under various conditions. Findings drawn from this research will provide guidance for water utilities to design suitable disinfection strategies for improved management of disinfectant decay and DBP formation in chloraminated DWDSs