Design and operation of drinking water treatment plants and associated distribution systems with long residence times are complicated by the formation of regulated disinfection byproducts (DBPs), comprised of total trihalomethanes (TTHM) and five haloacetic acids (HAA5). Treated water dissolved organic carbon (DOC) concentrations, the unit processes required to meet those DOC concentrations, and disinfection strategies (e.g., booster chlorination) are the primary design and operational considerations that can require extensive testing or modeling to determine. In this study, twelve different treated drinking waters were generated at the bench-scale using ferric chloride coagulation and granular activated carbon adsorption from four parent raw waters collected from the San Juan River representing spring runoff, monsoon, and low flow events. Treated drinking waters with DOC concentrations of 0.9, 1.4, and 1.9 mg/L were tested for regulated DBP formation under simulated distribution system (SDS) conditions over residence times as long as 56 days and compared to 7-day formation potential (FP) testing. SDS free chlorine concentrations were maintained between 0.2 and 1.0 mg/L as Cl2 through periodic booster chlorination. Maximum SDS TTHM and HAA5 concentrations were 0.208 and 0.074 mg/L, respectively, with formation consistently varying by approximately ±20% across the four different parent raw waters despite having been treated to the same three DOC concentrations. An average of four existing TTHM models consistently underpredicted TTHM formation by approximatively 20%. Long considered a conservative measure of DBP formation, FP testing also underpredicted SDS DBP formation at 56 days by approximately 40% on average. The DBP testing approach presented in this study allowed for the development of several significant linear relationships for predicting DBP concentrations based on treated water ultraviolet light absorbance at 254 nm, water temperature, and cumulative free chlorine demand.
Published by Elsevier Ltd.