Laboratory testing protocol to identify critical factors in bacterial compliance monitoring

This research focused on providing guidelines for water utilities on the collection and handling of routine bacteriological samples and in developing scientifically-based approaches in selecting the most representative sampling locations. A laboratory-scale pilot distribution system was designed comprising two parallel loops, one using unlined cast-iron pipe and one using PVC pipe. Each loop contained six sampling ports, including (1) a distribution main dead end faucet, (2) one long (5.5 m; 18 feet) and (3) one short (0.3 m; 1 foot) household copper service line with threaded hose-bibb taps, (4) one hose-bibb with welded faucet, (5) one dedicated sampling port (modeled after a manufacturer's specifications) and (6) one laboratory-style (PVC) stop-cock sampling port. Residual chlorine concentrations were maintained at 0, 0.5, 1.5 and 2.0 mg/L stages during the course of the experiment. Bacterial samples were collected from the different sampling ports and assayed by membrane filtration and/or spread plate. Nutrient and R2A agars were used for heterotrophic plate counts (HPC), m-Endo agar for total coliform (TC) counts and Chromocult agar for injured bacterial analyses. Several methods of sample collection were tested using various combinations of flushing and tap disinfection, including "first flush" (no flushing, without tap disinfection), flushing only, tap disinfection only (using alcohol or hypochlorite solution) and flushing coupled with tap disinfection. The results indicated that the bacterial counts in samples drawn from dead ends were not significantly different from counts in samples from the other sample port configurations. First flush samples consistently produced the highest bacterial count results. Bacterial counts in samples from the long household copper service line were typically three orders of magnitude higher than in samples from the other sample ports. Thus, there is evidence that long copper household service connections may be unsuitable sample tap configurations for collecting samples intended to represent microbial quality in the distribution system.