A study of small public ground-water-supply wells that produce water from discontinuous sand and gravel aquifers was done from July 1999 through July 2001 in southeastern Michigan. Samples were collected to determine the occurrence of viral pathogens and microbiological indicators of fecal contamination (“indicators”), determine whether indicators are adequate predictors of the presence of enteric viruses, and determine the factors that affect the presence of enteric viruses. Small systems are those that serve less than 3,300 people. Samples were analyzed for specific enteric viruses by reverse transcriptase-polymerase chain reaction (RT-PCR), for culturable viruses by cell culture, and for the indicators total coliforms, Escherichia coli (E. coli), enterococci, and F-specific and somatic coliphage. Ancillary environmental and water-quality data were collected or compiled.

A total of 169 regular samples and 32 replicate pairs were collected from 38 wells. Replicate pairs were samples collected at the same well on the same date. One well was sampled 6 times, 30 wells were sampled five times, 6 wells were sampled twice, and 1 well was sampled once. By use of RT-PCR, enterovirus was found in four wells (10.5 percent) and hepatitis A virus (HAV) in five wells (13.2 percent). In two of these wells, investigators found both enterovirus and HAV, but on different sampling dates. Culturable viruses were found one time in two wells (5.9 percent), and neither of these wells was positive for viruses by use of RT-PCR on any sampling date. If results for all viruses are combined, 9 of the 38 small public-supply wells were positive for enteric viruses (23.7 percent) by either cell culture or RT-PCR.

One or more indicators were found in 18 of 38 wells. Total coliforms, E. coli, enterococci, and F-specific and somatic coliphage were found in 34.2, 10.5, 15.8, 5.9, and 5.9 percent, respectively, of the wells tested. In only 3 out of 18 wells were samples positive for an indicator on more than one date at the same well. The co-occurrence of enteric viruses and any indicator was 55.6 percent; five out of the nine virus-positive wells were also found to be positive for an indicator. Two wells with detections of viruses had a detection of total coliforms, one well had a detection of E. coli, one of enterococci, and one of F-specific coliphage. On a per sample basis, of 11 samples that were positive for enteric viruses, indicator bacteria co-occurred in only 2 samples, and coliphage were not present in any.

More virus-positive samples were found at sites served by septic systems than those served by sewerlines. Sampling condition (ground water or a mixture of tank and ground water), distance to septic system, type of and distance to nearest surface-water body, well characteristics, or land use were not related to the presence of viruses or indicators. Among continuous water-quality variables, statistically significant relations were found between total coliforms and dissolved organic carbon and between total coliforms and iron. There was a statistically significant relation between chloride concentrations >20 mg/L and detections of total coliforms. Presence of nitrate and nitrite was related to the presence of other indicators (E. coli, enterococci, and F-specific and somatic coliphage) or enteric viruses, but not to total coliforms. The data indicated that chloride-to-bromide (C1:Br) ratios may be useful as a screening tool for total coliforms and enteric viruses but not for E. coli, enterococci, and F-specific and somatic coliphage.

This study provides evidence for fecal contamination of ground water from small public-supply wells, at least on an intermittent basis. Collecting data on multiple lines of evidence would be needed to reliably predict the presence of enteric viruses and protect public health. Future data collection toward this end could include repeat sampling several times a year for different indicators, measuring dissolved-organic carbon, nitrate plus nitrite, and (or) chloride concentrations, or determining C1:Br ratios. The presence of a site served by a septic system is an indication that the well may be more vulnerable to contamination than a site served by a sewerline.