From February 2004 through September 2005, USGS personnel, in cooperation with MMSD, were involved in an extensive data-collection effort as part of Phase II of the MMSD Corridor Study. The purpose of this data-collection effort was to address spatial, temporal, and technological gaps identified in the Phase I data analysis. This effort included the addition of harbor sites in Phase II, resulting in data collection at 15 stream and 6 harbor sites within the MMSD planning area. Data gathering included the collection of stream-discharge data for selected streams, the sampling and analysis for over 220 water-quality properties and constituents, and collection of additional suspended sediment, total phosphorus, and chloride data for computation of loads, yields, and volumetrically weighted concentrations at 4 sites. Bioassessment during autumn 2004 included gathering biologic-community data and stream-habitat data at wadeable Phase II stream sites.

Summary of Relations between Biological Metrics, Site Characteristics, and Water-Quality Data

Data for selected biological metrics were used to divide the 14 wadeable stream sites into four groups to investigate relations between bioassessment data and site characteristic and water-quality data. Aggregate bioassessment rankings of Phase II data were used to divide the sites into four quartiles reflecting relative water quality. Quartile 1 contained sites where bioassessment data indicated the least-degraded water quality among those sampled, and quartile 4 contained sites that indicated the most-degraded water quality. Quartiles contained the following stream sites:

• Quartile 1: Milwaukee River near Cedarburg, Milwaukee River at Milwaukee, Jewel Creek, and Menomonee River at Menomonee Falls
• Quartile 2: Willow Creek, Root River near Franklin, and Root River at Grange Avenue
• Quartile 3: Menomonee River at Wauwatosa, Oak Creek, and Little Menomonee River
• Quartile 4: Honey Creek, Underwood Creek, Lincoln Creek, and Kinnickinnic River

Site characteristics (in this case, drainage area and land use) and selected water- and sediment-quality constituent results were summarized based on the four bioassessment quartiles to determine if there were relations with the aggregate bioassessment rankings. In general, sites having the largest drainage basins with lowest proportion of urban land use were in quartile 1, and the smallest drainage basins with highest proportion of urban land use were in quartile 4. Major ions, indicator organisms, and wastewater compounds generally had the lowest overall results in quartile 1, and highest overall results in quartile 4. Results for other constituent types (nutrients, mercury, pathogenic organisms, and bed sediment) indicated mixed results, with results for some constituents decreasing from quartile 1 to quartile 4. Since neither biological metrics nor water-quality data provided a complete description of stream quality, the Phase II assessment of streams was strengthened by the use of both approaches to establish a holistic baseline assessment of stream quality.

Findings for Harbor Sites and the Milwaukee River at Mouth at Milwaukee Site

Harbor results from Phase II data collection indicated that most of the water-quality constituents highlighted in this report were also detected in the harbor. In the inner harbor, approximately half of the constituents highlighted had similar results when compared to results from stream sites; remaining constituents had results that were generally lower than those from stream sites. In the outer harbor, most constituents highlighted had lower results when compared to results from stream sites and inner-harbor sites.

Phase II findings indicated that 13 of the 15 wastewater compound (WWC) classes were present in the Milwaukee Harbor. In the inner harbor, results for seven WWC classes indicated DFs similar to those in stream samples: detergent metabolites, fire retardants, flavors and fragrances, herbicides, human drugs (nonprescription), insecticides, and miscellaneous. The antimicrobial disinfectant and antioxidant classes were not detected in the inner harbor. In the outer harbor, results for two WWC classes indicated DFs similar to those in stream samples: herbicides and miscellaneous. Eight classes were not detected in the outer harbor: antimicrobial disinfectants, antioxidants, detergent metabolites, dyes and pigments, fuel, plasticizers, solvents, and sterols.

The Milwaukee River at Mouth site integrates flow from multiple streams and is affected by seiches from Lake Michigan. Median results for certain water-quality constituents at this site were similar to those for inner-harbor samples, and median results for other constituents were more like those for stream samples. The median chloride concentration at Milwaukee River at Mouth was less than half than that observed at most stream sites, and was similar to median concentrations observed at harbor sites. Median concentrations for nutrients were similar to those for stream sites. The Milwaukee River at Mouth site had the lowest median concentrations of all stream sites for dissolved mercury and dissolved and particulate methylmercury (similar to harbor samples), but median concentrations of particulate mercury were similar to samples from the other stream sites. Median concentrations of indicator organisms (fecal coliform, E. coli, and coliphage) were lower than those observed at stream sites, but were higher than those observed in harbor samples. DFs of coliphage groups II and III indicated that, of the coliphage detected, the proportion of probable human sources was more similar to stream sites than harbor sites. With the exception of E. coli O157:H7, pathogen DFs were similar to those observed in stream samples. DFs of WWCs were similar to those observed at stream sites. Concentrations of trace elements and PCBs in bed sediment at Milwaukee at Mouth were among the highest observed at Phase II stream sites.

return to top

Phase I–Phase II Comparisons

General comparisons between Phase I and Phase II median concentrations were made where data were available for both phases and were sufficient to discern an overall pattern. Phase I median concentrations were calculated from available data; depending on the constituent, the amount of data ranged from a small dataset with only a few results to a large dataset with numerous results. Comparisons could not be made for a number of Phase II constituents (for example, the WWCs), as technology to analyze for many of these compounds has only recently been developed. For many of the constituents for which comparable data were available for Phase I and Phase II, notable differences were evident (appendix 5). All comparisons between Phase I and Phase II were for manually collected samples; automatically collected samples used for Phase II loads determinations had no Phase I counterparts available for comparison.

Median chloride concentrations generally increased between Phase I and Phase II, likely because winter sampling and event sampling were included as part of Phase II. This finding indicated that winter contributions of chloride to streams may be substantial. Increased chloride concentrations typically occur in winter when road salts were used as deicers on smaller snowfall events (typically 1 to 2 inches) and road salts are used to clear roads and parking lots. Meltwater from deicing activities of this kind produce small volumes of runoff with high concentrations of chloride. This type of runoff has the most effect on urban streams where the proportion of runoff is high due to the concentration of impervious surfaces, and where roads are prevalent and frequently deiced. Phase II stream sites with median chloride concentrations above the USEPA national chronic freshwater-quality criterion all had drainage basins with over 80 percent urban land use.

Nutrients, chlorophyll a, and suspended sediment concentrations indicated varying results from Phase I to Phase II. Total nitrogen and suspended sediment concentrations decreased or indicated no notable difference. Nitrate, total phosphorus, and chlorophyll a concentrations increased or indicated no notable difference from Phase I to Phase II.

Indicator organisms, fecal coliform, and E. coli comparisons also indicated varying results from Phase I to Phase II. Fecal coliform concentrations were generally decreasing. E. coli concentrations indicated no notable difference, but this may be due to the limited data set available for Phase I.

Although bed-sediment data for Phase I was limited, changes were still observed. Median concentrations of most trace elements (cadmium, chromium, copper, lead, mercury, nickel, and zinc) in bed sediment decreased from Phase I to Phase II; however, arsenic concentrations increased. PCB concentrations in bed sediment indicated no notable difference once the increased analytical sensitivity of Phase II analyses had been considered.

PCB data in sediment and fish tissue were available for comparison at three sites between Phase I and Phase II, because Phase II fish tissue samples were only collected at Milwaukee River at Milwaukee, Menomonee River at Wauwatosa, and Root River near Franklin. Although direct comparisons between the Lower Milwaukee River subwatershed and the three Milwaukee River Phase II sites were generally avoided, contamination of this subwatershed with synthetic organic contaminants (especially PCBs) is well documented and was discussed in general terms using data from the Milwaukee River at Milwaukee site. The Milwaukee River continued to have PCB detections in fish tissue, which could be due to contamination from the upstream Cedar Creek Superfund alternative site. At the Menomonee River at Wauwatosa site, PCBs were detected in bed sediment in both Phase I and Phase II; PCBs were not detected in fish tissue in Phase I but were detected in Phase II, which could imply that PCBs persisting in bed sediments are now moving up into biological systems. At the Root River near Franklin site, no PCBs were detected in bed sediment during Phase II. No data were available for Phase I. There were also no PCB detections in fish tissue in either Phase I or Phase II at the Root River near Franklin site. Upstream, at the Root River at Grange Avenue site, PCBs were detected in bed sediment during Phase II.

Fish-tissue data also were analyzed for historically used pesticides. Pesticide detections in fish tissue were observed at the Milwaukee River at Milwaukee site during Phase I; however concentrations observed during Phase II were below the Phase I reporting level. Although not detected in fish-tissue samples at Menomonee River at Wauwatosa during Phase I, historically used pesticides were observed in the samples collected during Phase II at concentrations above the Phase I reporting level. Root River near Franklin had no pesticide detections in fish tissue in Phase I or Phase II.

Bioassessments of streams indicated varying results when Phase I and Phase II data were compared. EPT taxa representation at most sites remained constant or improved from Phase I to Phase II; declines in EPT taxa representation were observed at Oak Creek and Root River near Franklin. In addition, the majority of Phase II HBI-10 scores remained constant or improved when compared to Phase I water-quality ratings; increases in HBI-10 ratings (from fairly poor to fair) were observed at Oak Creek and Root River at Grange Avenue. Fish IBI data had fewer sites available for comparison; however, where available, Phase II biotic integrity ratings at compared sites either remained constant or worsened when compared with Phase I. Although results for macroinvertebrate assessments were generally inconsistent with the fish IBI data, changes in water quality are generally reflected at the lower trophic levels first; therefore, this inconsistency may be an indication of improving water quality in these urban streams.

return to top
continue to References