Bacterial Pollution of Waters in Pristine and Agricultural Lands

Concentrations of thermotolerant coliform bacteria, presumptive E. coll, and presumptive fecal streptococci were determined from ditches, brooks, and natural ponds in six agricultural areas and 22 uninhabited pristine areas in southern Finland in the summer of 1987. For comparison, the same fecal indicators were enumerated from the effluents of three wastewater t eatment plants. The objective was to compare the importance of these waters as sources of fecal indicators in receiving waters. The numbers of bacteria in waters in agricultural areas often exceeded the limit of acceptable swimming water (1000 bacteria per 100 mL), especially during wet periods, which shows that diffuse loading can be a significant source of fecal pollution. Fecal indicators were detected in about half of the samples of pristine areas, sometimes in concentrations exceeding the limit of good swimming water (100 bacteria per 100 mL). This contamination was probably caused by wild animals, especially by elk (Alces alces) and deer (Odocoileus virginlanus) living in the areas. The concentrations of bacteria were higher in running waters than in ponds. The reliabilities of routine enumeration methods for the bacteria were evaluated by carrying out confirmation tests for isolated strains. Thermotolerant coliforms were reliable indicators in waters contaminated by diffuse loading. The reliability of enumeration of fecal streptococci in these waters should be studied further. B LOADS discharged to receiving waters from wastewater treatment plants appear to be decreasing in countries where purification processes have improved. Smith et al. (1987) studied water quality trends in the rivers of the USA in 1974-1981 and found that the concentrations of fecal indicator bacteria had decreased. Poikolainen (1988) observed similar decrease of these bacteria in Finnish inland waters when analyzing 16456 fecal coliform and 96 348 fecal streptococci measurements carried out from 1962 to 1984 from water samples of the national monitoring network. The decreases in these two examples are due to improved waste treatment. Because wastewater treatment plants discharge less bacteria to waters than earlier, the role of diffuse bacterial loading has simultaneously increased. For example, Smith et al. (1987) found that the occasionally observed increases in fecal coliform counts were positively associated with the density of cattle population and with feedlot activity in the drainage basins. Bacterial pollution in runoff from agricultural lands has been reviewed, e.g., by Khaleel et al. (1980), Crane et al. (1983), and Baxter-Potter and Gilliand (1988). Gary et al. (1983) found that cattle (Bos sp.) grazing in pastures bisected by a small perennial stream had only minor effects on the water quality during a study of 2 yr. However, the concentrations of indicator bacteria in the stream increased when the number of grazing cattle was high. When the cattle were removed the bacterial concentrations decreased to the levels obNational Board of Waters and the Environment, Water and Environment Research Institute, P.O. Box 250, SF 00101 Helsinki, Finland. Received 27 Apr. 1990. *Corresponding author. Published in J. Environ. Qual. 20:620-627 (1991). served in adjacent ungrazed pasture. Jawson et al. (1982) monitored indicator bacteria for 3 yr in grazed and nongrazed watersheds and found that the concentrations of total coliforms and fecal streptococci in runoff from the two watersheds did not differ significantly but fecal coliforms were greater from the grazed watershed. Walter and Bottman (1967) found that the concentrations of coliforms and enterococci were affected by wildlife. Coliforms have been enumerated from the waters of pristine areas affected only by wild animal populations. For example, Bohn and Buckhouse (1985) discussed the occurrence of coliforms and the problems of using them as indicators in wildland streams. Hazen (1988) and Fujioka et al. (1988) demonstrated the currence of high numbers of fecal coliforms and E. coil in pristine tropical environments. Niemelii and Niemi (1989) studied the species distribution and temperature relations of coliform populations from uninhabited watershed areas. They found that Serratia fonticola and Hafnia alvei were common among the species of total coliform bacteria isolated in pristine watersheds. Total coliform analysis therefore had no indication value in these waters. E. coli was the only coliform isolated from pristine areas that could grow at 44.5 °C, demonstrating the value of elevated temperature as a selective factor in monitoring fecal pO llution of surface waters by coliform analysis. Great temporal variation is observed in the concentrations of bacteria discharged from land to receiving waters. These variations are caused by characteristics of the individual drainage basins and by hydrological phenomena (Kunkle, 1970; Faust, 1976; Patni et al., 1985; Niemi and Niemi, 1988). The maximum bacterial concentrations that occur in water only for short periods of time may therefore remain undetected. Occasionally the concentrations of indicator bacteria in runoff from agricultural lands exceed the recommended water quality standards of point sources (Doran and Linn, 1979). This has led to discussions of whether the water quality standards developed for waters receiving point source pollution are applicable to waters under the influence of diffuse population (Doran and Linn, 1979; Bohn and Buckhouse, 1985; Milligan, 1987). The objective of this work was to compare the concentrations of thermotolerant coliform bacteria, presumptive E. coli, and presumptive fecal streptococci in receiving waters of pristine and agricultural areas in southern Finland. For further comparison, the same bacteria were enumerated from treated wastewater. The reliabilities of routine enumeration methods for these bacteria were evaluated by carrying out additional tests for isolated strains. MATERIALS AND METHODS