AWWA WQTC58810

This work utilized Information Collection Rule (ICR) data to compare the percent halogen as bromine observed in four classes of disinfection byproducts (DBPs) for which a full spectra of brominated and chlorinated species were available: dihaloacetic acids (DHAAs), trihaloacetic acids (THAAs), trihalomethanes (THMs), and dihaloacetonitriles (DHANs). As expected, the extent of bromine substitution within each of the four DBP compound classes ranged widely for the data set overall. However, a strong interdependence between the extents of bromine substitution in the four DBP classes was demonstrated such that, for a given water sample, the value in any one of the four classes could be well predicted by that for any of the other classes. The extent of bromine substitution in DHAAs closely mirrored that in THMs on a sample-by-sample basis. In contrast, the extent of bromine substitution in THAAs was approximately 20% lower than that in THMs and the extent of bromine substitution in DHANs was approximately 30% higher than that in THMs. Four bromine incorporation measures corresponding to the four DBP classes were treated as a multivariate response on individual water samples and observations with anomalous interclass bromine substitution patterns were identified using a multivariate outlier detection method. This method was successful at capturing samples that appeared discordant with respect to bromine incorporation upon graphic inspection. Only a small fraction (less than 1%) of the ICR samples were identified as outliers in this manner. In about half of these cases, the reported species concentration data were likely a result of data entry error. These results suggest a simple approach for augmenting routine DBP data quality assurance protocols. Includes 12 references, tables, figures.