AWWA WQTC64119

Enterococci and Escherichia coli (E. coli) analyses are recommended as a measure of recreational water quality. Epidemiological studies have led to the development of criteria that have been used to promulgate recreational water quality standards based on established relationships between health effects and water quality. Methods for enterococci (Methods 1106.1 and 1600) and E. coli (Methods 1103.1 and 1603) have been approved for ambient water quality monitoring. Bacterial analysis of non-potable waters (e.g., wastewater) for compliance monitoring requires that sample analysis begin within 6 hours of sample collection (40 CFR Part 136, Table II). The U.S. Environmental Protection Agency (USEPA) conducted a study to determine whether marine and freshwater samples analyzed for enterococci (Methods 1106.1 and 1600) and freshwater samples analyzed for E. coli (Methods 1103.1 and 1603) could be held for longer than 6 hours prior to sample analysis without a significant change in bacterial concentrations. Samples were analyzed at 0, 6, and 24 hours after sample collection or sample spiking. Time zero samples were analyzed within two hours of sample collection or sample spiking. Thirteen volunteer laboratories participated in the holding time study, with six to eight freshwater or marine matrices being evaluated using each of the methods noted above. Each laboratory evaluated "range-finding" fresh and/or marine water matrices to determine if sufficient ambient levels of enterococci and/or E. coli were present and whether sample spiking was required. If the target bacterial concentrations less than or equal to 150 CFU/100 mL, laboratories spiked samples. All results were natural log-transformed prior to performing any statistical analyses as a result of the skewed distribution of the results. All results also were stratified by method and matrix. For each method and matrix, a two-way Analysis of Variance (ANOVA) model was fit to assess the effect of holding time and laboratory on the log-transformed concentrations, and to assess whether there was a significant interaction between holding time and laboratory. For Methods 1600 and 1103.1 in freshwater, there was no significant interaction between holding time and laboratory, and, as a result, pairwise comparisons were performed in a single ANOVA model using data from all laboratories. For all other method/matrix combinations, there was a significant interaction between holding time and laboratory, and therefore an ANOVA model was fit for each laboratory separately. Within each ANOVA model, comparisons between each of the three holding times were performed using the Tukey-Kramer method for pairwise comparisons.