Reflectingthe desire for equitable control of disinfection byproducts (DBPs) at the tap, the Stage 2 DBP rule brings a new focus todifferentiating among distribution system (DS) monitoring locations with respect to DBPoccurrence levels. However, studies showing that DBP concentrations can vary significantlywithin a 24-hour period at both point of entry (POE) and DS sites raises important questionsrelated to sampling design and data interpretation. These findings suggest that observeddifferences in DBP concentrations among DS locations could derive from the timing of samplecollection rather than from location attributes. The research presented here involved monitoring short-termfluctuations in DBP concentrations at the POE of one Philadelphia Water Department(PWD) treatment plant, and examined corresponding water quality and operational data in detailto explain observed variations. This work was part of a broader study designed to determine theprevalence and magnitude of diurnal variations in finished water DBPs from differentgeographical regions in the U.S., to identify their causes, to evaluate their significance withrespect to monitoring programs, and to provide a foundation for further study of temporalvariation in DBPs at DS sites. The project involved 2 sampling rounds to capture seasonal waterquality variations. Each round included Trihalomethane (THM) and Haloacetic Acid (HAA) sample collection at the POE and atselected locations through the treatment train every three hours, continuously over a five-dayperiod. Temporal variability patterns observed for THMs differed markedly from those forHAAs. During both sampling events, THM4 concentrations transitioned dramatically betweenlow and high values in a seemingly random and binary pattern. These changes did notcorrespond to any known shifts in source or treated water quality or in applied chemical doses. Incontrast, HAA6 levels exhibited only gradual changes in concentration over both samplingperiods and these changes corresponded with observed source water quality trends. Subsequentinvestigation suggested post-filter pH adjustment practices as the cause of the THM fluctuations.Follow-up controlled laboratory experiments confirmed this hypothesis. Operational changesnow being implemented as a consequence of the study findings are expected to result insubstantially lower and more consistent average finished water THM4 concentrations. This high-resolutionexamination of DBP and treatment data at PWD yielded critical insights aboutoperational impacts on DBP formation that were not previously apparent. Includes tables, figures.
