The high and low values of THMs in
the PP were reasonably predicted by the models. However, the
models showed relatively weak performance at few data
points. This may be explained by the fact that these data were
from the Danville (Q3) system, which has relatively high
concentrations of bromide ions. The presence of bromide ions
changes reaction patterns based on water pH, water temperature
and type and distribution of natural organic matter
(NOM) in water (Sohn et al., 2006; Hellur-Grossman et al.,
2001). For the THMs in the HWT, the models predicted most
of the peak values consistently, with the exception of the
highest value from the Q2 water system (THMs: 345 mg/L). The
observation of this single very high value may be associated
with an unusual temperature rise in the HWT, excessive
residual organics and free chlorine residuals in the HWT, as
well as a possible outlier effect resulting from instrument and
human errors during preservation and analysis.