Discharges of fecal bacteria (fecal coliform bacteria and Escherichia coli ) to the middle main stem of the Cuyahoga River from storm water, combined sewers, and incompletely disinfected wastewater have resulted in frequent exceedances of bacteriological water-quality standards in a 23-mile reach of the river that flows through the Cuyahoga Valley National Recreation Area. Contamination of the middle main stem of the Cuyahoga River by bacteria of fecal origin and subsequent transport to downstream areas where water-contact recreation is an important use of the river are a concern because of the potential public-health risk from the presence of enteric pathogens.

Independent field investigations of bacterial decay, dilution, dispersion, transport, and sources, and bacterial contamination of streambed sediments, were completed in 1991-93 during periods of rainfall and runoff. The highest concentration of fecal coliform bacteria observed in the middle main stem during three transport studies exceeded the single-sample fecal coliform standard applicable to primary-contact recreation by a factor of approximately 1,300 and exceeded the Escherichia coli standard by a factor of approximately 8,000. The geometric-mean concentrations of fecal bacteria in the middle main stem were 6.7 to 12.3 times higher than geometric-mean concentrations in the monitored tributaries, and 1.8 to 7.0 times larger than the geometric-mean concentrations discharged from the Akron Water Pollution Control Station.

Decay rates of fecal bacteria measured in field studies in 1992 ranged from 0.0018 per hour to 0.0372 per hour for fecal coliform bacteria and from 0.0022 per hour to 0.0407 per hour for Escherichia coli. Most of the decay rates measured in June and August were significantly higher than decay rates measured in April and October. Results of field studies demonstrated that concentrations of fecal coliform bacteria were 1.2 to 58 times higher in streambed sediments than in the overlying water. Sediments are likely to be a relatively less important source of fecal bacteria during rainfall and runoff in the middle main stem relative to bacterial loading from point sources.

Numerical streamflow and transport simulation models were calibrated and verified with data collected during field studies. Of the constituents modeled, bacteria exhibited the poorest correspondence between observed and simulated values. The simulation results for a dye tracer indicated that the model reasonably reproduced the timing of dissolved constituents as well as dilution and dispersion effects. Calibrated and verified models for 1991 and 1992 data sets were used to simulate the improvements to bacteriological water quality that might result from reductions in concentrations of fecal bacteria discharged from two major sources.

The model simulation resulting in the greatest improvement in bacteriological water-quality was one in which concentrations of fecal coliform bacteria and Escherichia coli were reduced by 90 percent in the Cuyahoga River at the Old Portage gaging station, and to geometric-mean bathing-water standards in the effluent of the Akron Water Pollution Control Station (BWS/90 scenario). Compared to the results of the base-simulation, when the BWS/90 scenario was applied in the 1991 model simulation, Escherichia coli concentrations were reduced 98.5 percent at Botzum, 97.5 percent at Jaite, and 91.1 percent at Independence. For 1992 model simulations, similar percent reductions in the concentrations of Escherichia coli were predicted at the three stream sites when the same reductions were applied to sources. None of the model simulations resulted in attainment of bacteriological water-quality standards.

The potential benefits of source reductions to human health and recreational uses were estimated by comparing the number of illnesses per 1,000 people from concentrations of Escherichia coli associated with the BWS/90 simulation, with the base simulation, and with the geometric-mean standard for Escherichia coli. The predicted 22 to 26 illnesses per 1,000 people predicted by the E. coli concentrations resulting from BWS/90 simulation are 2.8 to 3.3 times higher than the 8 illnesses per 1,000 people associated with the geometric-mean primary-contact water-quality standard for Escherichia coli. Risks associated with the base simulation are 4.6 to 4.9 times higher than that associated with the geometric-mean primary- contact water-quality standard for Escherichia coli. The illness risks predicted from the BWS/90 scenario, although larger than acceptable, would nevertheless be an improvement over conditions that were encountered during field studies in 1991-93.