A bacteriological study of natural freshwater swimming areas in western Oregon.

Introduction

Historically, recreational swimming has been popular in nearly every country in the world. Most swimmers assume that the water is safe if it appears "clean." Bacteria, however, are not detectable to the naked eye, and they exist even in clear waters (1). As early as 1953, researchers were conducting epidemiological studies to assess the risk of illness associated with swimming in surface water; those studies were the first in a series of attempts to demonstrate that high levels of bacteria in natural waters can cause illness among swimmers (2). Recent reports confirm that, from 1991 to 1992, 1,825 people from 21 states became ill in outbreaks involving recreational waters (3). The majority of the outbreaks occurred in June or July, during the peak swimming season, and 45 percent of the cases were directly related to lake and river water exposure. The investigators also noted that "swimming and other recreational activities in which unintentional ingestion of water can occur are known to increase the risk of gastrointestinal illness, even in non-outbreak settings" (3). The disease-producing potential of freshwater recreational areas is currently being estimated according to certain bacteriological indicators associated with the presence of fecal contamination, such as fecal coliform, Escherichia coli (E. coli), and enterococcus (4,5). In most studies, the main factor for subsequent illness is the submersion of the swimmer's head or the total length of time spent underwater - in some cases, illness has been associated with as little as 10 minutes spent underwater (6).

One widely publicized incident occurred in the summer of 1991 at Blue Lake Park, near Portland, Oregon: 21 cases of E. coli O157:H7 and 38 cases of Shigella sonnei were traced to human fecal contamination of lake water (7). All cases occurred in children who had been swimming, and many of these children had reported swallowing lake water. At the time of the outbreak, however, the levels of indicator organisms did not consistently exceed U.S. Environmental Protection Agency (U.S. EPA) regulatory limits, which suggests that levels of indicator organisms are not always predictive of public health risk. As a result of that study, infants and toddlers who are not toilet-trained are restricted to a water spray play area that drains to city sewers and is separate from the lake. In addition, a pump has been installed to circulate the water in the lake (7).

Before 1998 - and at the time of the study reported in this article - the state of Oregon used a fecal-coliform standard to determine if water bodies met the criteria for the "beneficial use" of water-contact recreation. The limit set by the fecal-coliform standard was exceeded when the geometric mean of at least five data points was greater than 200 fecal coliforms per 100 milliliters (mL) of water (8). Recently, the state of Oregon replaced the fecal-coliform standard with an E. coli standard for the "beneficial use" of streams for water-contact recreation during the period of June 1 to September 31. The upper limit for the new standard is a geometric mean of 126 E. coli per 100 mL of water based on at least five samples collected over a 30-day period (9). The revised regulations also stipulate that in any single sample, E. coli concentrations shall not exceed 406 organisms per 100 mL (9). This standard is identical to federal water quality criteria (10,11). It is believed that at these levels, cases of gastroenteritis will not exceed eight per 1,000 freshwater swimmers (12).

With a few exceptions, such as Blue Lake Park, natural freshwater swimming areas in Oregon are not routinely tested or monitored, and there is only limited information from which to estimate the bacterial and chemical water quality of these areas. This study, therefore, sought to quantitatively measure bacterial and selected physical and chemical parameters at two popular natural swimming areas in Linn and Benton counties. The researchers hoped this information would be useful to state and local agencies that advise the public about possible health risks in natural swimming areas.

The following research questions directed the study:

1. What bacteria levels, represented by fecal-coliform and E. coli counts, are present in selected local natural swimming areas over the course of the summer?

2. Do bacterial indicators at these sites exceed Oregon state standards? If so, how often?

3. What factors might be contributing to the bacterial levels at these sites?

Areas of Study

Two local swimming areas in western Oregon were sampled throughout the summer of 1996. The sites were selected as the result of consultations with local citizens, university students, and public health professionals, who identified popular "swimming holes." The Avery Park swimming area is located in Corvallis, Oregon, on the Mary's River, which is a major tributary of the Willamette River. Avery Park is a public park maintained by the City of Corvallis Parks and Recreation Department, and the swimming area is close to a well-maintained group picnic area. The banks here are very high and are eroded from recent flooding. The river bottom is gravelly, consisting of everything from large rocks to pea gravel. Several feet from the beach, the fiver bottom drops off sharply Because of the sharp drop-off, most children swim near the shore. The banks have dense tree cover, and the water moves swiftly.

Monteith Riverfront Park is located in Albany, Oregon, on the eastern bank of the Calapooia River, at its confluence with the Willamette River. Monteith Park is adjacent to the downtown area and is frequently used for summer concerts, weddings, picnicking, and swimming. It is maintained by the City of Albany Parks and Recreation Department. The samples were taken just upstream of the confluence of the two rivers. This location is ideal for young swimmers; across the swimming area, from beach to beach, the water is only 4 feet deep. The beach and river bottom are sandy, with large, easily visible rocks. The water is clear and swift. Both parks have outdoor toilet facilities, located about 200 yards from the swimming areas.

Methods

Sampling and Testing Procedures

The sites were first sampled in late June and early July 1996 to establish baseline bacterial counts and other baseline parameters before the start of the swimming season, which normally runs from early July through August. Samples were tested for fecal coliform and E. coli. The physical and chemical tests included air temperature, water temperature, conductivity, turbidity, pH, dissolved oxygen, nitrate-nitrogen, and ammonia-nitrogen. Daily air temperature and amounts of precipitation were obtained from a database maintained by a local weather-reporting station.

Sampling for the study began on July 6, 1996, and ended on August 31, 1996. Samples were collected twice a week, on Wednesday and either Saturday or Sunday, between noon and 3 p.m. The sites were always tested in the same order and at approximately the same time of day - during peak swimming hours.

Samples for physical and chemical tests were collected in clean glass bottles. The samples for bacterial analysis were collected in autoclaved, sterile glass bottles. Surface grab samples were taken from the top 12 inches of the swimming area, which is within the recommended sampling depth of approximately 3 feet and is also the area where a swimmer would be most likely to ingest water (13). The samples were then stored in a cooler with ice for delivery within six hours to the Environmental Health and Safety Laboratory at Oregon State University.

Water temperature, conductivity, and pH were all measured on site as soon as the sample was collected. Water temperature and pH were measured simultaneously with a Portable Hach One pH Meter. Conductivity was measured with the Hach Portable Conductivity/TDS Meter. Dissolved oxygen samples were drawn in clean, 60 mL bottles designed for the analysis of biochemical oxygen demand and were stabilized with two reagent powder pillows until the sample could be titrated with sodium thiosulfate in the laboratory. Dissolved oxygen, phosphorus, ammonia-nitrogen, and nitrate-nitrogen were analyzed with a Hach DR/2000 Portable Laboratory with spectrophotometer. Turbidity and apparent color were also analyzed with the Hach DR/2000 Portable Laboratory. Field measurements and procedures followed approved methods (14).

Duplicate bacterial samples were analyzed for fecal coliform and E. coli with the membrane filter technique, according to standardized methods (15). Data were reported as number of colonies per 100 mL sample. For plates showing more than 1,000 colonies, the colonies were recorded as too numerous to count (TNTC). Samples showing colonies TNTC are represented as samples with 1,000 colonies in Figure 1 and Figure 2.

Quality assurance and quality control procedures adhered to standardized methods (15). One random duplicate of each chemical or physical test was performed each sampling day, and every bacterial test was run in duplicate. In addition, the sites were sampled at approximately the same time each sampling day and in the same order. To ensure the precision of the technique, samples were split and run also at the Corvallis Treatment Plant Water Quality Laboratory, a state-certified laboratory, during the first two weeks of the sampling period. Results indicated that the values from the split samples correlated highly: r = .91 for E. coli colony counts, and r = .76 for fecal-coliform counts.

Treatment of Data

Bacteriological counts were averaged daily and transformed with the natural log. The natural log was used because the difference between the natural logarithms of two numbers approximately equals the difference between the nontransformed values, allowing direct proportional comparison between values in the transformed state (16). Transformation of the data was necessary because the values spanned several degrees of magnitude. Median values for bacterial counts per site also were calculated, and the data were transformed back for reporting.

Bacterial plate values that showed no apparent colonies were termed "nondetect" and were not necessarily zero values; usually these values occurred in highly diluted samples, and the bacteria may have been "missed" when water was drawn into the pipette. On the log mean scale, the graphical representation of these nondetect values is 1; this representation is designed to avoid giving the impression of missing zero-value data, which would not show on a log chart.

Results

A total of 19 samples were taken for each site over the course of the summer. Both sites might be classified as "healthy" water bodies; average dissolved oxygen, pH, conductivity, nitrate-nitrogen, phosphorus, and nitrogen-ammonia were within normal ranges and within the boundaries set by federal and state regulations (9,13,17). Baseline samples taken early in the summer were variable with respect to fecal coliform and E. coli but steady for all other tests. This variability, combined with the lack of adequate history about bacteria levels in these areas, made it difficult to determine what actual "normal" bacteria levels are at the sites. The mean and median instantaneous values were calculated, as were the log mean values for the entire sampling period, of June 28 to August 31. Monteith Park had higher mean and median instantaneous counts for both fecal coliform and E. coli; counts were roughly three times as high for fecal coliform and twice as high for E. coli. The median and mean values for both indicator organisms were comparable at the two sites.

Thirty-day log means were calculated to provide an evaluation of the bacterial quality over time; these calculations are used by both federal and state regulatory agencies in designating bacterial standards (9,10). Again, Monteith Park showed log mean values substantially higher than those found at Avery Park. During July, fecal-coliform log values for Monteith Park were more than three times as high as those for Avery Park; during August the fecal-coliform log values for Monteith Park were twice as high. Log values for E. coli were five times higher at Monteith Park during July, but only slightly higher during August.

Bacteria levels at the two sites also differed from sample to sample on any given day [ILLUSTRATION FOR FIGURE 1 OMITTED], [ILLUSTRATION FOR FIGURE 2 OMITTED]. During the course of the study, bacteria levels ranged from 0 to 1,000 colonies per 100 mL sample for E. coli, and from five to 1,000 colonies per 100 mL sample for fecal coliform. Tests at Monteith Park showed higher levels both of L coli and of fecal coliform, with nine-week means of 305 colonies per 100 mL for fecal coliform and 70 colonies per 100 mL for E. coli. When viewed over the entire sampling period, fecal-coliform and E. coli levels were not highly correlated at either site (r = .36 at Avery, and r = .03 at Monteith). From July 14 to August 17, however, fecal-coliform and E. coli levels were moderately to highly correlated at both parks (r = .74 at Avery, and r = .51 at Monteith).

The researchers also analyzed the data to determine if bacteria levels at the sites exceeded state and federal regulatory limits. Tests conducted on a daily basis, referred to as instantaneous counts, did find that existing state limits were exceeded at each site at least once during the summer, and, during several weeks, limits were exceeded regularly ([ILLUSTRATION FOR FIGURE 1 OMITTED] and [ILLUSTRATION FOR FIGURE 2 OMITTED]). Higher overall fecal-coliform counts were found at Monteith Park, with 79 percent of the colony counts at or exceeding the fecal-coliform state limits of 200 colonies per 100 mL sample. Lower fecal-coliform counts were documented at Avery Park, where values were at or exceeded the limit 21 percent of the time. Higher E. coli counts also were documented at Monteith Park, where values were at or exceeded the limit of 126 colonies per 100 mL sample 42 percent of the time. By contrast, E. coli counts at Avery Park were at or exceeded the limit in only 21 percent of the sampling period.

Computation of 30-day log mean values indicated that only Monteith Park exceeded the fecal-coliform limit of 200 colonies per 100 mL sample in July (293 colonies) and in August (314 colonies). The 30-day log mean E. coli limit of 126 colonies per 100 mL sample was not exceeded at either park in July or August.

For both sites, comparisons were made between fecal coliform and E. coli, fecal-coliform levels and regulatory limits, and E. coli levels and regulatory limits. The graphs showing these comparisons are presented by site ([ILLUSTRATION FOR FIGURE 1 OMITTED] and [ILLUSTRATION FOR FIGURE 2 OMITTED]). Both graphs are logarithmic so that the degrees of magnitude between counts can be clearly seen.

Fecal-coliform and E. coli counts were highly correlated (r = .74) at Avery Park between July 14 and August 17, 1996. Although the levels at this site did not often exceed regulatory limits, fecal-coliform counts were very close to those limits through most of the summer, and E. coli counts were close to the limits throughout August. Higher fecal-coliform and E. coli levels were found at Monteith Park. Figure 2 indicates that at Monteith Park, fecal-coliform and E. coli levels were moderately correlated (r = .51) from July 14 to August 17 but were not correlated for the remainder of the sampling period.

Other parameters that might have been contributing to elevated bacteria levels were investigated. Correlation coefficients were obtained for the following analyses: air temperature and water temperature, daily precipitation and fecal coliform, daily precipitation and E. coli, number of swimmers and fecal coliform, number of swimmers and E. coli, turbidity and fecal coliform, and turbidity and E. coli. Of these pairs, only air and water temperature correlated strongly at each site (r = .67). Because swimmers would be more likely to use air temperature as a "barometer" for going swimming, correlation coefficients were obtained for air temperature and E. coli, as well as for air temperature and fecal coliform. The results, however, showed that air temperature appeared not to be correlated with E. coli (r = -.05) and to be only moderately correlated with fecal coliform (r = .42).

Discussion

Higher mean and median levels for both fecal coliform and E. coil were found at Monteith Park. This result may be due to a variety of factors, including higher levels of turbidity and a larger number of younger swimmers. The sandy beaches and river bottom at Monteith Park may be one factor in the greater turbidity; Avery Park has a gravel beach and river bottom. Higher turbidity levels in Monteith Park also may be attributed to the number of swimmers, which was as high as 33 per sampling day. Swimmers tend to stir up the water, potentially releasing bacteria that have settled into the sediment on the river bottom (18). In addition, turbid waters tend to harbor bacteria because the bacteria cling to suspended particles (13). These bacteria may be of a natural origin, since the upstream area is rich in relatively undisturbed woodlands and fields. A number of warm-blooded animals, including beaver, deer, nutria, and other small mammals, inhabit this region, as do numerous waterfowl. Higher levels of fecal coliform and E. coli also may be due to the larger population of younger, diapered swimmers, who have been noted to have contributed to previous outbreaks (7).

Although studies have found temperature to be a factor in E. coli growth in natural waters, temperature did not appear to be a factor in this study, which found no correlation between water or air temperature and E. coli levels (19). In the course of the study, however, a ruptured City of Albany sanitary sewer line was found approximately one-third of a mile upstream from the park on the south bank of the Calapooia River. A substantial quantity of sewage was observed around the damaged site, and a depressed swale led directly to the river channel approximately 15 feet away. During wet weather flooding, this location would have been underwater, and the sewage would have mixed freely with the current. During the summer, however, the sewer line was not underwater and would not be expected to affect the swimming area. The damage was promptly reported to and repaired by the Albany Public Works Department, which forwarded a copy of the report to the Oregon Department of Environmental Quality.

In this study, the 30-day log means for bacteria exceeded the limit set by state standards for fecal coliform only at Monteith Park. Nevertheless, the

possibility of a public health risk for swimmers at both parks cannot be dismissed. As demonstrated by the 1992 E. coli and Shigella sonnei outbreak that occurred in Portland, Oregon, swimmers can contract illness even when bacterial tests generally meet U.S. EPA regulatory standards for fecal coliform. The incident raises the question of whether instantaneous bacteria counts instead of 30-day log mean values should be used as bacteria standards. This study found that, although the log mean values never exceeded regulatory limits at Avery Park and exceeded only fecal-coliform limits at Monteith Park, the daily bacterial levels often exceeded the daily maximum limit, particularly at Monteith Park. For example, the instantaneous fecal-coliform value at Monteith Park on July 14 was 600 colonies per 100 mL (which exceeded the 200 colonies per 100 mL regulatory limit), but the 30-day log mean value for the month of July was 92 colonies, which was far below the limit. Swimmers who ingested water on July 14 might have been exposed to water with high levels of fecal bacteria, although the log mean value was acceptable.

The revised state bacteria standards specify that no single sample shall have more than 406 E. coli organisms per 100 mL (9). The data from this study suggest that this standard would have been exceeded only once during the summer, at Monteith Park on August 28, when the E. coli count was 500 colonies per 100 mL.

Many authors agree that monitoring of recreational waters needs to be increased, particularly to correlate bacteria levels with the actual risk of illness to swimmers (20,21). Some researchers suggest that, "primary prevention requires the elimination of fecal contamination at swimming areas - an effort that may depend more on public cooperation than on governmental regulation" (7).

Conclusions and Recommendations

Bacteria levels fluctuated during the nine-week sampling period from July 6 to August 31, 1997. Counts ranged from zero to 1,000 colonies per 100 mL sample for E. coli and from five to 1,000 colonies per 100 mL sample for fecal coliform. Tests at Monteith Park showed higher levels both of E. coli and of fecal coliform, with nine-week means of 305 colonies per 100 mL for fecal coliform and 70 colonies per 100 mL for E. coli; these levels were more than double those found at Avery Park.

E. coli levels were much more variable on a day-to-day basis than were fecal-coliform levels. The fecal-coliform and E. coli levels were closely correlated over short periods of time, but they did not correlate when viewed over the entire summer. To better assess risk to swimmers, additional research might focus on evaluating the percentage of E. coli O157:H7 in fecal coliform relative to the percentage of nonpathogenic organisms.

At Avery Park, the instantaneous fecal-coliform values exceeded the limits set by existing state standards during 21 percent of the sampling period. At Monteith Park, these values exceeded the limits during 79 percent of the same period. Similarly, instantaneous E. coli counts taken at Avery Park exceeded the regulatory limit during 11 percent of the sampling period, and E. coli counts at Monteith Park were at or exceeded the limit during 42 percent of the period. Although instantaneous values exceeded the daily maximum limits for fecal coliform and E. coli at both parks, the 30-day log means show that only Monteith Park exceeded the 30-day state fecal-coliform limits in place at the time of the study. These somewhat conflicting data raise the question of whether standards based on 30-day log means are sufficiently protective of swimmers who may be swimming on days when the bacterial colony counts exceed the daily maximum limits by several orders of magnitude. On the basis of these data, the authors recommend better surveillance of freshwater recreational areas, including follow-up studies with swimmers to determine if certain recreational waters are associated with increased morbidity.

Factors that may have contributed to the sporadically high bacteria levels have not been identified. The high fecal-coliform and E. coli levels may be due to a combination of factors that are beyond the scope of this study. Septic systems along the waterways, high turbidity, water temperature, and stream flow all may have influenced the overall bacteria levels. Although the study locations were not randomly selected and inferences cannot necessarily be drawn for other bodies of water in the state, the results are still valuable; they have provided specific information to local and state public health officials about bacteria levels at two popular and previously unsampled swimming areas.

The authors suggest that additional research be conducted in areas where bacteria levels are high, such as at Monteith Park - and, periodically, at Avery Park. Contributing influences, such as tributaries or failing on-site sewage systems might be investigated as potential contributors to fecal-bacteria loading. Recommendations from staff at the City of Corvallis Water Treatment Plant include sampling upstream of the swimming area and measuring stream flow through a swimming area to see if the bacteria are being carried past the area quickly or slowly, thereby decreasing or increasing potential swimmer exposure.

Because the two highest swimmer use areas in this study were also parks, it might be beneficial for the public to be informed by local public health officials of the bacteria levels present in these areas and of illnesses that might develop from ingesting water that contains high levels of fecal bacteria. It is advised that signs be posted to inform parents that diapers soiled with fecal material may contaminate swimming areas. Diapered children should be changed regularly, and diapers should be disposed of properly - not left on the beach. In addition, parents should be alert for any connections between gastrointestinal illness and recent swimming by their children. The information could be effectively presented to the public in a flyer that lists the areas with high bacteria counts and that mentions the prevention methods parents can use. These methods include the need to be aware of potential E. coli infection if symptoms of gastrointestinal illness arise. Finally, it might be necessary for public health authorities to provide public notice of periods when they expect bacteria levels to exceed limits set by state standards and to recommend that children under four years of age not be allowed in swimming areas while these conditions prevail.

Corresponding Author: Anna K. Harding, Ph.D., R.S., Chair and Associate Professor, Dept. of Public Health, Oregon State Univ., Corvallis, OR 97331-6406.

REFERENCES

1. Brock, T.D., and M.T. Madigan (1991), Biology of Microorganisms, 6th ed., Englewood Cliffs, N.J.: Prentice Hall.

2. Stevenson, A.H. (1953), "Studies of Bathing Water Quality and Health, American Journal of Public Health, 43:529-538.

3. Moore, A.C., R.L. Calderon, G.F. Craun, B.L. Herwaldt, A.K. Highsmith, and D.D. Juranek (1993), "Surveillance for Waterborne Disease Outbreaks - United States, 19911992," Morbidity and Mortality Weekly Report, 42(SS-5): 1-22.

4. Nuzzi, R., and R. Burhans (1997), "The Use of Enterococcus and Coliform in Characterizing Bathing-Beach 'Waters," Journal of Environmental Health, 60(1):16-22.

5. Dufour, A.P (1984), "Health Effects Criteria for Fresh Recreational Waters," EPA-600/1-84-004, Cincinnati. Ohio: U.S. Environmental Protection Agency.

6. Fleisher, J.M., A.F. Godfree, D. Kay, R.L. Salmon, and M.D. Wyer (1996), "Marine Waters Contaminated with Domestic Sewage: Nonenteric Illnesses Associated with Bather Exposure in the United Kingdom," American Journal of Public Health, 86(9):1228-1234.

7. Keene, W.E., T.J. Barrett, D.W. Fleming, K. Hedberg, F.C. Hoesly, J.M. McAnulty, F.L. Oxman, M.A. Pfaller, and L.P. Williams (1994), "Swimming-Associated Outbreak of Hemorrhagic Colitis Caused by Escherichia coli O157:H7 and Shigella sonnei," New England Journal of Medicine, 331(9):579-584.

8. "Oregon's Water Quality Status Assessment Report, 305(b)" (1994), Portland, Oreg.: Oregon Department of Environmental Quality.

9. "Statewide Water Quality Management Plan: Beneficial Uses, Policies, Standards, and Treatment Criteria for Oregon" (1998), Oregon Administrative Rules, Chapter 340 (OAR 340-41).

10. Office of Water Regulation and Standards (1986), "Ambient Water Quality Criteria for Bacteria - 1986," EPA-440/5-84-002, Washington, D.C.: U.S. Environmental Protection Agency.

11. Oregon Association of Clean Water Agencies (1996), "EQC Adopts Revisions to Water Quality Standards," Oregon Association of Clean Water Agencies Newsletter, Winter: 1,4.

12. Pike, E.B. (1993), "Recreational Use of Coastal Waters: Development of Health-Related Standards," Journal of the Institute of Water and Environmental Management, 7(2):162-169.

13. Csuros, M. (1994), Environmental Sampling and Analysis for Technicians, Boca Raton, Fla.: CRC Press, Inc.

14. Water Analysis Handbook (1992), 2nd ed., Loveland, Colo.: Hach Company.

15. Standard Methods for the Examination of Water and Wastewater (1995), 19th ed., Washington, D.C.: American Public Health Association.

16. Darlington, R.B. (1990), Regression and Linear Models, New York, N.Y.: McGraw-Hill Publishing Co.

17. Salvato, J.A. (1992), Environmental Engineering and Sanitation, 4th ed., New York, N.Y.: John Wiley & Sons.

18. Seyfreid, P.L., N.E. Brown, P.F. Ness, and R.S. Tobin (1985). "Microbiological Study of Swimming-Related Illness: Morbidity and the Microbiological Quality of Water," American Journal of Public Health, 75(9): 1071-1075.

19. Neimi, R.M., and J.S. Neimi (1991), "Bacterial Pollution of Waters in Pristine and Agricultural Lands," Morbidity and Mortality Weekly Report, 42(SS-5):1-22.

20. Kebabjian, R. (1994), "Monitoring the Effects of Urban Runoff on Recreational Waters," Journal of Environmental Health, 56(9):15-18.

21. Blostein, J. (1991), "Shigellosis from Swimming in a Park Pond in Michigan," Public Health Reports, 106(3):317-321. in Pristine