Indicator organisms are microorganisms whose presence in water indicates probable presence of pathogens (disease-causing organisms). Ideally, such microorganisms are nonpathogenic, occur consistently in pathogen-contaminated water, do not multiply in waters, are reliably detectable even at low concentrations, and are present in greater numbers than and have similar survival times to pathogens. The waterborne pathogens of interest during Phase II were bacterial, viral, and protozoal. Testing was limited to bacterial and viral indicators because, to date, no adequate indicator protozoa have been defined (Mara and Horan, 2003).

The indicators used during Phase II—fecal coliforms, Escherichia coli (E. coli), and coliphage—were chosen because of their efficacy at predicting pathogen presence and their widespread use in water-quality monitoring in the United States. Traditionally, the fecal coliform group has been used as an indicator of bacterial pathogen presence and general wastewater contamination (U.S. Environmental Protection Agency, 1976). In 1986, the USEPA updated its guidance to recommend that a particular fecal coliform member, E. coli, be monitored instead because of its stronger relation to the occurrence of swimming-associated gastrointestinal illness (Dufour, 1984). Coliphage are a class of viruses that infect E. coli. They have higher resistance to environmental stresses and disinfection than bacteria do and are therefore thought to mirror the survival rates of enteric viruses more closely than their bacterial counterparts (Mara and Horan, 2003).

The many detections in indicator data sets from the MMSD planning area generally allowed for the use of typical statistical descriptors in this report. Wherever possible, concentrations are given as medians. Where data were insufficient to do this and results were widely dispersed, results are compared in terms of ranges and maximum concentrations.

Fecal Coliforms

In 1976, the USEPA established a water-quality criterion stating that the acceptable limit for fecal coliform concentration in bathing waters was 200 col/100mL (U.S. Environmental Protection Agency, 1976). Historical data are available for fecal coliforms and were summarized in the Phase I report (Schneider and others, 2004).

Fecal coliform concentrations in stream samples ranged from less than 10 to 58,000 col/100 mL, with a median concentration of 350 col/100 mL. Median concentrations at the majority of stream sites (73 percent) were above the USEPA criterion of 200 col/100 mL. Median fecal coliform concentrations indicate a positive relation with increasing urban land use (fig. 27). The Honey Creek site had the highest median concentration (1,900 col/100 mL) (fig. 28), and the Milwaukee River at Mouth site had the lowest (55 col/100 mL). Median concentrations at remaining sites ranged from 130 to 695 col/100 mL.

Figure 27. Median concentrations of fecal coliform (log transformed) plotted against percent urban land use in site drainage basins for 15 stream sites in the Milwaukee Metropolitan Sewerage District planning area, Wis. Site abbreviations listed in table 1.

Figure 28. Distribution of fecal coliform concentrations, by site, in the Milwaukee Metropolitan Sewerage District planning area, Wis. Water-quality criterion line represents the U.S. Environmental Protection Agency acceptable limit for fecal coliform concentrations in recreational waters (200 col/100 mL) (U.S. Environmental Protection Agency, 1976). Site abbreviations listed in table 1.

Although higher maximum concentrations were found in stream samples collected during low-flow events (58,000 col/100 mL) than in those from high-flow events (6,600 col/100 mL), median concentrations indicated no consistent response in relation to flow (fig. 29A). No patterns were evident in relation to seasonality (fig. 29B); however, when flows were combined with seasonality, a consistent response was apparent (fig. 29C). Summer low-flow samples had the highest median concentration of fecal coliform, and summer high-flow events had the lowest; the opposite was true for winter samples.

Figure 29. Distributions of fecal coliform concentrations for stream and harbor sites, Milwaukee Metropolitan Sewerage District planning area, Wis. Stream sample distributions are grouped by flow (A), season (B), and flow and season combined (C). Harbor-sample distributions are grouped by season only (D) (no harbor samples were collected in winter). Water-quality criterion line represents the U.S. Environmental Protection Agency acceptable limit for fecal coliform concentrations in recreational waters (200 col/100 mL) (U.S. Environmental Protection Agency, 1976).

Data were available for Phase I and Phase II fecal coliform concentration comparisons at ten sites (appendix 5). At most sites (7 of 10), median fecal coliform concentrations decreased from Phase I to Phase II, where percent differences ranged from -97 percent at Underwood Creek to -29 percent at Menomonee River at Menomonee Falls. However, there was one site with a notable increase: Jewel Creek, which had no concentrations above the reporting level (10 col/100 mL) in the two samples summarized for Phase I and a median concentration of 165 col/100 mL in samples collected for Phase II. The remaining two sites, Root River near Franklin and Root River at Grange Avenue, indicated no notable differences.

The overall median fecal coliform concentration at harbor sites (less than 10 col/100 mL) was lower than that for stream sites (350 col/100 mL), and may be attributed to organism die-off due to environmental stress or dilution when mixing with water from Lake Michigan. Median concentrations in samples from the inner-harbor sites ranged from less than 10 to 20 col/100 mL (fig 28). No fecal coliform were detected in samples collected from any of the three sites in the outer harbor. Harbor samples indicated no seasonal response (fig. 29D).

return to top

Escherichia coli

In its 1986 guidance on E. coli as an indicator of contamination in freshwater systems, USEPA established a criterion for designated beach areas wherein the single sample maximum density allowed is 235 colonies of E. coli per 100 mL (U.S. Environmental Protection Agency, 1986). Historical data are available for E. coli and were summarized in the Phase I report (Schneider and others, 2004).

E. coli concentrations in stream samples ranged from less than 1 to 34,000 MPN/100 mL, with a median of 420 MPN/100 mL. Median concentrations at the majority of stream sites (67 percent) were above the USEPA criterion of 235 col/100 mL⁴. Median E. coli concentrations indicated a positive relation with increasing urban land use (fig. 30). The Honey Creek site had the highest median (2,000 MPN/100 mL) (fig. 31), and Milwaukee River near Cedarburg and Milwaukee River at Mouth had the lowest (44 and 57 MPN/100 mL, respectively). Median concentrations at remaining sites ranged from 180 to 990 MPN/100 mL.

⁴ In this report units of col/100 mL were considered equivalent to units of MPN/100 mL (Wisconsin Department of Natural Resources, [n.d.]).

Figure 30. Median concentrations of E. coli (log transformed) plotted against percent urban land use in site drainage basins for 15 stream sites in the Milwaukee Metropolitan Sewerage District planning area, Wis. Site abbreviations listed in table 1.

Figure 31. Distribution of E. coli concentrations, by site, in the Milwaukee Metropolitan Sewerage District planning area, Wis. Water-quality criterion line represents the U.S. Environmental Protection Agency single sample maximum density of E. coli allowed in designated beach areas (235 col/100 mL) (U.S. Environmental Protection Agency, 1986). Units of col/100 mL were considered equivalent to units of MPN/100 mL (Wisconsin Department of Natural Resources, 2006a). Site abbreviations listed in table 1.

Median concentrations indicated no consistent response in relation to flow (fig. 32A), but a consistent response was evident in relation to seasonality (fig. 32B). Median concentrations were generally higher in summer and winter than during other times of the year. A consistent response was also observed when flows were combined with seasonality (fig. 32C). Summer low-flow samples had the highest median concentration of E. coli, and summer high-flow samples had the lowest; the opposite was true for winter samples.

Figure 32. Distributions of E. coli concentrations for stream and harbor sites, Milwaukee Metropolitan Sewerage District planning area, Wis. Stream sample distributions are grouped by flow (A), season (B), and flow and season combined (C). Harbor-sample distributions are grouped by season only (D) (no harbor samples were collected in winter).Water-quality criterion line represents the U.S. Environmental Protection Agency single sample maximum density of E. coli allowed in designated beach areas (235 col/100 mL) (U.S. Environmental Protection Agency, 1986). Units of col/100 mL were considered equivalent to units of MPN/100 mL (Wisconsin Department of Natural Resources, 2006a).

Data were available for Phase I and Phase II E. coli concentration comparisons at five sites (appendix 5). The Kinnickinnic River was the only site with an increase from Phase I to Phase II (28 percent difference). Two sites indicated a notable decrease: Lincoln Creek (-35 percent difference) and Menomonee River at Wauwatosa (-21 percent difference). The two remaining sites, Honey Creek and Menomonee River at Menomonee Falls, indicated no notable differences.

The overall median E. coli concentration at harbor sites (2 MPN/100 mL) was lower than that for stream sites (420 MPN/100 mL) and may be attributed to organism die-off or dilution when mixing with water from Lake Michigan. Median concentrations in samples from the inner-harbor sites ranged from 3 to 19 MPN/100 mL (fig. 31), whereas medians in samples from the outer-harbor sites were lower than those from the inner harbor, ranging from less than 1 to 1 MPN/100 mL. Harbor samples did not indicate a consistent seasonal response (fig. 32D).

return to top

Coliphage

The group of coliphages tested for during Phase II of the MMSD Corridor Study were male-specific and therefore characterized by their infection of E. coli bacteria containing F-pili. Despite their high correlation with enteric virus concentrations in contaminated rivers and lakes, and their comparable resistance to environmental factors and disinfecting agents, coliphage fall short of being an ideal indicator of enteric virus presence, and a need for additional field studies of ecology and survival time still exists (Mara and Horan, 2003). Coliphage data are not present in the Phase I database and were not analyzed in conjunction with the Phase I report.

Coliphage concentrations in stream samples ranged from less than 1 to 4,400 plaques/100 mL, with an overall median of 4 plaques/100 mL. Median coliphage concentrations indicate a positive relation with increasing urban land use (fig. 33). The Menomonee River at Wauwatosa site had the highest median concentration (31 plaques/ 100 mL) (fig. 34), and the Willow Creek, Milwaukee River at Mouth, and Jewel Creek sites had the lowest (less than 1 plaque/100 mL). Median concentrations at remaining sites ranged from 1 to 22 plaques/100 mL.

Figure 33. Median concentrations of coliphage (log transformed) plotted against percent urban land use in site drainage basins for 15 stream sites in the Milwaukee Metropolitan Sewerage District planning area, Wis. Site abbreviations listed in table 1.

Figure 34. Distribution of coliphage concentrations, by site, in the Milwaukee Metropolitan Sewerage District planning area, Wis. Site abbreviations listed in table 1.

Median concentrations for stream samples indicated no consistent response to flow (fig. 35A), however the maximum concentration during low-flow events was much higher (4,400 plaques/100 mL) than that for high-flow events (990 plaques/100 mL). Median concentrations in stream samples indicated a weak seasonal response (fig. 35B); the highest median was for summer samples (10 plaques/100 mL), whereas medians for the rest of the year ranged from 3 to 4 plaques/100 mL. A consistent response became more apparent when flows were combined with seasonality (fig. 35C). The summer low-flow events had the highest median concentration (12 plaques/100 mL) of any season, and summer high-flow events had the lowest (1 plaque/100 mL). This pattern of higher median concentrations in the low-flow-event samples (6 plaques/100 mL) and lower median concentrations in the high-flow-event samples (2 plaques/100 mL) held for samples collected in the spring, though the difference between medians was less dramatic. Winter concentrations indicated the opposite pattern: the lowest median concentrations occurred during low-flow events (2 plaques/100 mL) and the highest median concentrations occurred during high-flow events (6 plaques/100 mL).

Figure 35. Distributions of coliphage concentrations for stream and harbor sites, Milwaukee Metropolitan Sewerage District planning area, Wis. Stream sample distributions are grouped by flow (A), season (B), and flow and season combined (C). Harbor-sample distributions are grouped by season only (D) (no harbor samples were collected in winter).

The overall median coliphage concentration at harbor sites (less than 1 plaque/100 mL) was slightly lower than that for stream sites (4 plaques/100 mL), and may be attributed to dilution from mixing with water from Lake Michigan. Median concentrations in samples from the inner-harbor sites ranged from less than 1 to 2 plaques/100 mL (fig. 34). Median concentrations in the outer-harbor samples were less than the inner-harbor samples, at less than 1 plaque/100 mL at each of the three sites. Harbor samples collected in spring had the highest median concentration (2 plaques/100 mL, compared to less than 1 plaque/100 mL summer and autumn) and the highest detection frequency (DF) (67 percent) among all other seasons sampled (average DF of 27 percent) (fig. 35D).

In addition to its use as an indicator of enteric virus levels, coliphage can be used for tracking sources of fecal contamination. For Phase II of the MMSD Corridor Study, small subsets of detected coliphage were subsequently serotyped in an effort to categorize potential contamination sources. More specifically, five plaques isolated from each coliphage sample were classified as either deoxyribonucleic acid (DNA) or ribonucleic acid (RNA) phages, depending on the chemical composition of their genetic material. Any RNA coliphages detected were then serotyped into one of four general serogroups: group I, group II, group III, or group IV.

Individual serogroups originate from a limited number of sources: group I is found in human and animal feces, groups II and III are found predominantly in human feces, and group IV is found predominantly in nonhuman animal feces (Simpson and others, 2002). Therefore, the serotyping of any RNA coliphages present allowed for a broad source-tracking signal. For the purposes of this report, groups I and IV were not considered for analysis, and groups II and III were considered together as a probable indicator of human fecal contamination.

Because the method was to select only five of the coliphage plaques for further analysis, and only a fraction of these were serotyped into RNA coliform serogroups, data on various serogroup presences in samples were scarce by design. In fact, only 49 percent of all the samples collected contained a coliphage classified into any of the four serogroups. Therefore, in contrast to the general indicator data discussed previously in the report, serogroup detections were much less frequent. Whereas general indicator data contained ample detections for the description of constituent levels by standard statistical methods (for example, medians), detections of various serogroups are too scarce to be analyzed in this manner. These data are reported instead as the frequency with which detections of the combined groups were found in samples.

return to top

Groups II and III

When groups II and III were combined, the overall DF at stream sites was 14 percent. Detections for coliphage in these groups were found at 11 of the 15 sites (fig. 36). At four sites, DF was greater than or equal to 25 percent: Honey Creek (33 percent), Menomonee River at Wauwatosa (33 percent), Lincoln Creek (25 percent), and Underwood Creek (25 percent). Sites with no detections were the Milwaukee River at Milwaukee, Little Menomonee River, Root River at Grange Avenue, and Jewel Creek. DFs at remaining sites ranged from 8 to 17 percent. DFs indicated no relation to land use.

Figure 36. Detection frequency of coliphage groups II and (or) III in samples, by site, in the Milwaukee Metropolitan Sewerage District planning area, Wis. Site abbreviations listed in table 1.

Consistent responses in DFs were observed in relation to flow and season. Flow-related DFs for stream samples were higher during low-flow events (16 percent) than during high-flow events (12 percent) (fig. 37A). For stream samples, the highest seasonal DF was in the autumn (23 percent), and the lowest was in the spring (8 percent) (fig. 37B). DFs for the rest of the year ranged from 13 to 17 percent. When flows were combined with seasonality, DFs were higher during low-flow events than during high-flow events for corresponding seasons, with the exception of the autumn high-flow-event category (DF of 67 percent) (fig. 37C).

Figure 37. Detection frequencies of coliphage groups II and (or) III detections for stream and harbor sites, Milwaukee Metropolitan Sewerage District planning area, Wis. Stream samples are grouped by flow (A), season (B), and flow and season combined (C). Harbor samples are grouped by season only (D) (no harbor samples were collected in winter).

The overall DF at harbor sites (16 percent) was similar to that of stream sites (14 percent). All harbor sites had detections in at least one sample (fig. 36). DFs at all inner-harbor sites were 20 percent. DFs at all outer-harbor sites were 11 percent. The highest seasonal DF at harbor sites was in the autumn (33 percent), and the lowest was in the summer (4 percent) (fig. 37D). The DF in spring was 19 percent. No harbor samples were collected in winter.

return to top
continue to Pathogenic Organisms