Framework for cost-effective prediction of unregulated disinfection by-products in drinking water distribution using differential free chlorine
Unregulated disinfection by-products (U-DBPs) in drinking water are formed primarily by the reaction of disinfectants (e.g., chlorine) with natural organic matter (NOM). The formation can be dictated by basic water quality and operational parameters. The purpose of this study was to develop a cost-effective framework for predicting the U-DBP concentration in water distribution networks (WDNs) based on changes in free chlorine in the WDN with respect to that at the water treatment plant (WTP) outlet (called differential free chlorine, i.e., ΔCl2). Under the framework, mixed-effect models were developed to predict U-DBP families: haloacetonitriles (HANs) and haloketones (HKs) and individual compounds: chloropicrin (CPK), dichloroacetonitrile (DCAN), bromochloroacetonitrile (BCAN), and 1,1,1-trichloropropanone (TCP). Common water quality parameters at the WTP outlet were also included to increase the effectiveness of the models. The models developed were applied successfully to two applications: a single large-sized and combined medium-sized water systems with R2 values of the predictive models of up to 0.95 and a 95% confidence interval of the percent difference in the predictions of −27% to 32%. These applications show that ΔCl2 only or with basic water quality parameters at a WTP is an effective parameter for estimating the U-DBP concentration in WDNs. The proposed framework can be used by water utilities as a low-cost method to develop models for their water systems to identify potential locations to monitor U-DBPs. Moreover, the frameworks can be used by public health organizations to assess population exposure to U-DBPs in drinking water.