Disinfection byproducts (DBPs) in municipal supply water are a concern because of their
possible risks to human health. Risk assessment studies often use DBP data in water
distribution systems (WDS). However, DBPs in tap water may be different because of
stagnation of the water in plumbing pipes (PP) and heating in hot water tanks (HWT). This
study investigated occurrences and developed predictive models for DBPs in the PP and the
HWT of six houses from three municipal water systems in Quebec (Canada) in a yearround
study. Trihalomethanes (THMs) in PP and HWT were observed to be 1.4e1.8 and 1.9
e2.7 times the THMs in the WDS, respectively. Haloacetic acid (HAAs) in PP and HWT were
observed to be variable (PP/WDS ¼ 0.23e2.24; HWT/WDS ¼ 0.53e2.61). Using DBPs occurrence
data from these systems, three types of linear models (main factors; main factors,
interactions and higher orders; logarithmic) and two types of nonlinear models (three
parameters Logistic and four parameters Weibull) were investigated to predict DBPs in the
PP and HWT. Significant factors affecting DBPs formation in the PP and HWT were identified
through numerical and graphical techniques. The R2 values of the models varied
between 0.77 and 0.96, indicating excellent predictive ability for THMs and HAAs in the PP
and the HWT. The models were found to be statistically significant. The models were
validated using additional data. These models can be used to predict DBPs increase from
WDS (water entry point of house) to the PP and HWT, and could thereby help gain a better
understanding of human exposure to DBPs and their associated risks.