Predictability of swimming prohibitions by observational parameters: a proactive public health policy, Stamford, Connecticut, 1989-2004.

Introduction

The method used by the city of Stamford laboratory to determine what conditions warrant a preemptive closure may be applicable to any recreational water body. Knowing that bacterial elevations can be due to the flushing of urban areas through rivers, streams, and storm drains, as opposed to a point source (U.S. Environmental Protection Agency [U.S. EPA], 1995) helps direct resources. In 2000, U.S. EPA listed Stamford and two states as having established models for preemptive closure of bathing waters. The hope is that this will assist and prompt others to establish their own methods and parameters for recreational water closures (U.S. EPA, 1999).

Prior to 1990, beach closures were based on bacteria tests taken from designated points along the beach. If one or more samples from a beach were found to contain more than the recommended level of bacteria, the beach was closed and more samples were taken. Sampling continued until acceptable levels were obtained and then the beach was reopened. Samples were usually taken on Monday, so if test results were elevated, retesting could occur prior to the weekend. Samples were rarely collected if it was raining and rainfall amounts were generally not recorded. Before 1988, a few individual samples had been found with high bacteria counts but no cause was determined.

Several factors occurred to necessitate the previous procedures to be revised. First, certain areas apart from the bathing areas became "conditionally open" for recreational shellfish harvesting (State of Connecticut Marine Aquaculture Division, 1988). "Conditionally open" is a protocol based on rain monitoring. Second, elevated sample results in 1988 prompted a comprehensive sanitary survey that raised more questions than it answered. Finally, the primary indicator organism was changed to the enterococcus group (this was a 48-hour test in 1989), and during that indicator organism's evaluation in 1989 the summer season was extremely wet. Rainfall was collected and measured by a rain gauge operated at the Stamford Health Department and confirmed by the NOAA Station #5N-06-7970-3 at the Stamford Museum and Nature Center.

Methods

Standard methods for all sampling, transport, and holding methods were followed as referenced in the current edition of Standard Methods for the Examination of Water and Wastewater (Greenberg, Clesceri, & Eaton, 1992). All testing was performed within an hour of receipt in the lab. All bacteria testing used the membrane filtration technique using portions of a 100 mL sample. Typical aliquots would be 10 mL for total and fecal coliform testing and 20 mL for enterococcus testing. M-Endo broth was used for total coliform testing, m-FC broth with rosolic acid was used for fecal coliform testing, and mE agar was initially used for the enterococcus testing until the m-EI, U.S. EPA 1600 testing became available in 1998.

Until 1988, recreational waters at 11 points on four beaches were sampled weekly for approximately 15 weeks each summer.

Counts above 1,000 total coliform for 100 mL or 400 fecal coliform per 100 mL were considered elevated.

After a particularly rainy period at the end of July in 1988, several sets of samples failed, indicating a new problem. A broad survey was undertaken, with over 500 samples taken during the season to determine where the source of the problem might be. Samples were collected daily from the beaches, at marinas, points offshore from the beaches, known shoreline septic systems, storm drains, pump stations, and river outflows. Any unusual objects washing into the swimming area were noted, identified, and analyzed. Testing for fecal streptococcus was also added to total and fecal coliform testing.

Only enterococcus and fecal coliform were evaluated after 1989. An elevated value for enterococcus was greater than 61 colonies per 100 mL for a single sample and 33 for a geometric mean of five or more samples. Fecal coliform levels were used for continuity to past results. From 1989 to 1995, samples were collected when problems were thought to exist or under conditions that needed to be evaluated more extensively, which included high "moon" tides, storms, wind, and high temperatures. In 1995, additional samples were taken and evaluated after rain events less than one inch after a period of dry weather conditions.

The characteristics of each beach were evaluated in 1996 with special emphasis on flows from rivers and storm waters, tidal flows, and man-made and natural barriers.

In 1998, the bacterial analysis of the storm waters and rivers flowing near the beaches were checked. Also in 1998, the affects of discolored white Styrofoam-like objects washing ashore days after rain events greater than two inches were assessed.

In 2000, enterococcus limits were adjusted to 104 CFUs/100 mL for a single sample or a geometric mean limit at 35 to reflect the state of Connecticut and U.S. EPA recommendations (State of Connecticut Department of Health Services, 1989). These test results are evaluated yearly.

Results

Prior to 1988 (i.e., 1967 to 1986), an average of 14 samples out of 170 samples per season were elevated, but in 1988, more than 20 elevated samples occurred in a single two-week period. An additional 400 samples were taken and 48 of those showed high bacterial levels. Results clearly indicated that a simple point source of the elevated bacteria counts did not exist.

In 1989, evaluating enterococcus bacteria as the primary indicator organism showed a correlation between elevated enterococcus levels and rainfall. This correlation was not the same with total and fecal coliforms. Using enterococcus values with multiple regression and Chi-squared analysis determined that tide, wind, and water temperature were nonsignificant (p > .05). The data was stratified by rain (less than one inch and one inch or more), bacterial level (less than 60 colonies/ 100 mL and greater than or equal to 61 colonies/100 mL), and days after the rain event (less than two and greater than or equal to two), which revealed a significant relationship (p < .01). A higher level of enterococcus bacteria in the bathing water occurred with more rain (Glassman & O'Brien, 1990).

[FIGURE 1 OMITTED]

The evaluations taken from 1988 to 1995 showed significant rainfall events (defined as events greater than one inch or more) were the primary condition that elevated the geometric mean above 33/100 mL for enterococcus bacteria (Glassman & O'Brien, 1990) (Table 1).

Elevated temperature and high tides show an elevated geometric mean until rainfall is factored out, then they show geometric means below 33/100 mL.

In 1993 and 1995, the total amount of rain during the beach season was very light and rain greater than one inch did not occur in the months of June, July, or August. Bacterial counts were greatly influenced by storm water runoff with a rain of 0.7", and some areas were affected with as little as a half inch of rain. The effect caused by less rain was not as widespread and predictable as for that observed after a rainfall greater than an inch, but it can prove to be significant in times of decreased rain during the summer months (Table 2).

In 1996, the physical characteristic of each beach was examined and cumulated data was reviewed (Figure 1).

Cove Island beach is bordered on the east by the outflow of a sizable salt marsh that is fed by a small river draining the eastern portion of Stamford (light industrial, urban, and residential) and the western portion of a strictly residential town.

Horseshoe beach is located in a small cove and is sheltered from direct runoff but has the outlet of a small marina to the west.

Cummings beach has an outlet from a man-made drainage basin and a marina to the west separated by a rock jetty.

West beach is located to the west of the same drainage basin and marina as Cummings beach but has no barriers to the flow of the outlet.

The east and midsample points at Cummings beach and the Horseshoe beach at Cove Island did not show the same increase in bacteria after significant rain.

Of all the beaches, West beach showed the greatest influence after significant rainfall with the Cove beach area and one sampling point from Cummings showing a smaller but still significant increase in bacteria after rains.

The discolored white Styrofoam-like objects washing ashore days after rain events greater than two inches were analyzed several times. In 1989, total and fecal coliform bacteria were found in quantities from 1.2 to 5 million colonies per gram. Analysis in other years were consistent with bacteria counts greater than one million CFUs per gram for both fecal coliform and enterococcus bacteria.

The primary chemical constituent is a moderately high hydrocarbon not of petroleum origin determined by high pressure liquid chromatograph and referenced to oleic acid by infrared spectrometry (see photo on page 20).

Discussion

In Stamford, storm water is not combined with the sanitary sewers and therefore is not treated, but does not discharge directly to the recreational swimming areas (Bell & Romick, 2003). Storm drains were considered a point source in the early eighties because large populations of raccoons were living within the system. Raccoon habitation in the storm drains was well documented during "Operation Shellfish," a program that checked all homes and businesses in the area for improper sanitary sewer connections. Raccoon habitation was also considered a primary cause of the high bacterial levels in Stamford's storm drains. This changed in 1991 when a rabies epidemic came to the northeast United States and the raccoon population was reduced by up to 95% (State of Connecticut Department of Health Services, 1991).

[ILLUSTRATION OMITTED]

Bacterial levels continued to be elevated after rainfall of one inch or greater even after the collapse of the raccoon population, showing again that the source was human, not animal.

Normal tidal fluctuations do not appear to affect the water quality at the beaches. Unusually high tides, particularly the perigee tide occurring in summer, might have an influence by washing out storm drains in the same way as rain. The geometric mean of samples taken after the perigee tide was not above 35.

Using enterococcus bacteria only, it was clear that rainfall was the primary influence. The use of total coliform and fecal coliform bacteria did not make the contribution of storm water apparent because some coliform organisms are saprophytes and can bloom in higher water temperatures without being linked to a pollution source (Greenberg, Clesceri, & Eaton, 1992).

Throughout this study, any rain of one inch or greater would cause a number of samples to be elevated, but when conditions had been dry for a period of time, less rain was needed to affect water quality. Figure 2 represents a map of Stamford beaches and Table 3 illustrates the effect of a 0.7-inch rainfall in samples taken several hours after the rain ended. The results of tests taken two days after the initial profile showed no significant bacteria counts at any area. With this issue identified, dry weather conditions were defined as less than two inches of rain in 30 days and less than one inch of rain in 10 days; a half inch of rain is the amount needed to trigger a preemptive closure.

[FIGURE 2 OMITTED]

Sewage grease was an issue in 1988 when it was observed washing into bathing areas, and was initially believed to be the main cause of beach closures during that summer.

More information concerning floatables (Greenberg, Clesceri, & Eaton, 1992) was obtained in 1998, when several of them were observed before they washed onto a beach. Lifeguards reported the appearance of reddish water several days after a problem occurred at a neighboring sewage treatment plant due to almost four inches of rain over a two-day period. The town where the problem originated closed their beaches for several weeks. Beaches from a neighboring town located between where the problem originated and Stamford had one beach closed for one week with enterococcus colonies per 100 mL averaging 151 and fecal coliform colonies per 100 mL averaging 868. Stamford's beaches were analyzed and found to contain a significant number of rotifers and jellyfish but a low level of enterococcus and fecal coliform bacteria. Identification and quantification of these rotifers and comparing this to the saprobic index indicated the water to be moderately to heavily polluted at the time (Foissner, 1992; Sladecek, 1983) (Table 4). When the sewage grease did wash ashore, the enterococcus levels became elevated until the floatables were no longer present.

If floatables are found in open water, no significant elevation occurs in bacterial counts of the water where they are found. This unexplained characteristic caused some confusion in identifying the floatables as being a component of sewage, but as they washed up on shore and were subject to wave action they broke apart, releasing bacterial contamination.

Lifeguards and marine police were trained to recognize sewage grease and other suspicious events in the waters and to report them to the health department. Beach closures can occur when these items are observed and remain in effect until they stop washing ashore. Usually the marine police observe a slick approaching and are able to give one days' notice before these floatables wash ashore. Lifeguards also report and send samples of any unusual substance found in the water to the health department.

The presence of sewage grease was found to be significant only at the beach where it was found and not at neighboring beaches. The geometric mean of samples taken from beaches where the sewage grease was present was greater than 35/100 mL. The appearance of floatables usually occurred between one and 10 days after a sewage discharge from treatment plants with combined storm water and sewer connections. Some of these discharges were reported and some were assumed, since the plants' capacities were exceeded by rains greater than two inches (Massachusetts Water Resources Authority, 1996; U.S. EPA, 1997). The floatable components of these discharges are "grease balls" and a thin, highly visible film or slick that is wind driven.

Perhaps the most difficult issue in using predictive closures is dealing with the public. Their perception is that the water is getting "dirtier" when bathing areas are closed, but in reality it means that the issues and problems of water quality are becoming more apparent. If the water were rarely tested, swimming would almost never be prohibited. The media can be extremely helpful by assisting in explaining the situation ("The Good and the Bad," 1998; Waldman, 1999).

Conclusion

A policy was established to prohibit bathing at beaches using immediately observable and measurable parameters. Monitoring the beach areas continues on a weekly basis, with special emphasis on sampling immediately after a rain, at low tide or during any other adverse condition that might affect the water quality.

In summary, runoff after a significant rain event affects the quality of water at the recreational beaches in Stamford. What constitutes significant rain may vary according to the drainage and topography of each beach and according to the rainfall pattern for the season. If more sources of drainage around a beach or the occurrence of drought or near-drought conditions occurs, then less rain is needed to influence water quality. Under conditions of normal or above normal rainfall, one inch of rain is needed to elevate bacterial levels in bathing areas with a single source of storm water influx. Under dry weather conditions, a half inch of rain is all that is necessary to affect the bacterial level in the recreational water area. If multiple sources of storm runoff exist, a mere quarter inch of rain can have a significant impact on water quality (State of Connecticut Department of Environmental Protection, 1998).

Establishing a predictive model for beach closures is resource intensive and requires many samples taken under a variety of different conditions. Unusual occurrences of elevated bacterial levels should not be discounted until all possible explanations have been exhausted. Once a predictive model is established, then health of bathers is promoted by prohibiting swimming and potentially preventing disease when studies have shown the risk is highest. This also gives those who go to the recreational bathing areas maximum opportunity to enjoy these areas, since they are not needlessly closed while awaiting a bacterial retest.

Acknowledgements: Iliana Ayala, M.S., Research Assistant for Save the Sound, Inc., provided information and direct assistance on rotifer identification. Thomas J. Murray, Assistant Director of Waste Water Treatment for Westchester County, New York, from 1967 to 1988, provided insight into the problems treatment plants have handling large rain events. Connecticut Agricultural Station provided HPLC analysis on floatable material submitted July, 1994.

Corresponding Author: Joseph E. Kuntz, Lab Technician, City of Stamford, Health Department Laboratory, 888 Washington Blvd., Stamford, CT 06901. E-mail: JKuntz@ ci.stamford.ct.us.

REFERENCES

Bell, D., & Romick, L. (2003). 2000 & 2003 Triennial Evaluation. Hartford, CT: State of Connecticut Department of Agriculture, Bureau of Aquaculture, and Laboratory.

Foissner, W. (1992). Evaluating water quality using protozoa and saprobity indexes. Protocols in Protozoology, B(11), 1?20.

The good and bad about beach closing. (1998, July 20). The Stamford Advocate, p A14.

Glassman, M.J., & O'Brien, G. (1990). Rainfall as a predictor of unacceptable bathing water bacteria level. Paper presented at the Stamford Health Department.

Greenberg, A.E., Clesceri, L.S., & Eaton, A.D. (Eds.) (1992). Standard methods for the examination of water and wastewater (18th ed.). Washington, DC: American Public Health Association.

Massachusetts Water Resources Authority. (1996). Combined sewer overflows. Boston, MA: Author.

Sladecek, V. (1983). Rotifers as indicators of water quality. Hydrobiologia, 100, 169-201.

State of Connecticut Department of Environmental Protection (1998). Connecticut water bodies not meeting water quality standards, state of Connecticut department of environmental protection report to the United States Environmental Protection Agency [section 303(d) list]. Hartford, CT: Author.

State of Connecticut Department of Health Services (1989). Guidelines for monitoring bathing waters and closure protocol. Hartford, CT: Author.

State of Connecticut Department of Health Services. (1991). Connecticut epidemiologist, epidemiology section, 11(3). Hartford, CT: Author.

State of Connecticut Marine Aquaculture Division. (1988). Memorandum of understanding between the state of Connecticut Marine Aquaculture Division and the city of Stamford. Hartford, CT: Author. U.S. Environmental Protection Agency Office of Enforcement and Compliance. (1995). Pro-Act fact sheet on storm water pollution prevention, report TI#6727. Washington, DC: Author.

U.S. Environmental Protection Agency Long Island Sound Office, Long Island Sound Study. (1997). Floatable debris. Stamford, CT: Author.

U.S. Environmental Protection Agency Beach Watch Program. (1999). Review of potential modeling tools and approaches to support the beach program. Washington, DC: Author.

Waldman, L. (1999, June 9). Testing the sound's water, Stamford health department monitors shoreline for bacteria. The Stamford Advocate, p. A3.

Joseph Kuntz, R.S.

Robert Murray, M.S.

TABLE 1
Evaluation Using the Geometric Mean under Various Conditions

Condition          Bacteria               No.        Geometric
Tested                                  Samples        Mean

All samples
                Total coliform            291          27.5
                Fecal coliform            328          11.9
                Enterococcus              569           6.5
Rain < 1.00"
                Total coliform             71           186
                Fecal coliform             71            50
                Enterococcus               71            43
Low tide
                Total coliform             51          13.8
                Fecal coliform             58           4.8
                Enterococcus               58           3.3
Flood tide
                Total coliform            127          26.8
                Fecal coliform            146            11
                Enterococcus              146           6.1
High tide
                Total coliform             84          47.9
                Fecal coliform             88          24.1
                Enterococcus               88          13.7
Ebb tide
                Total coliform             28          19.3
                Fecal coliform             35          11.7
                Enterococcus               35           5.1
Water temp. < 51[degrees]F
                Total coliform             11           3.7
                Fecal coliform             11           1.9
                Enterococcus               14           2.0
Water temp. 51[degrees]F-61[degrees]F
                Total coliform             84          39.4
                Fecal coliform             84          11.8
                Enterococcus               84           7.6
Water temp. > 61[degrees]F
                Total coliform            186          31.1
                Fecal coliform            220          14.3
                Enterococcus              377           6.2

Condition                                                No.
Tested             Bacteria             Arithmatic     Elevated
                                           Mean         Samples
All samples
                Total coliform              583           24
                Fecal coliform              160           21
                Enterococcus               96.9           48
Rain < 1.00"
                Total coliform            1,134           11
                Fecal coliform              242           10
                Enterococcus                146           32
Low tide
                Total coliform              373            2
                Fecal coliform              112            2
                Enterococcus                 30            5
Flood tide
                Total coliform              480           14
                Fecal coliform              147           11
                Enterococcus                123           15
High tide
                Total coliform              778            6
                Fecal coliform              143            5
                Enterococcus                 97           20
Ebb tide
                Total coliform              543            2
                Fecal coliform              334            3
                Enterococcus                 63            7
Water temp. < 51[degrees]F
                Total coliform              101            0
                Fecal coliform               75            0
                Enterococcus                 51            1
Water temp. 51[degrees]F-61[degrees]F
                Total coliform              815           13
                Fecal coliform              296           10
                Enterococcus                212           13
Water temp. > 61[degrees]F
                Total coliform              512           11
                Fecal coliform              124           11
                Enterococcus                 62           34

TABLE 2
Evaluation of Drought Conditions and the Amount of Rain Causing
Closure Conditions

                                        Enterococcus/
                Summer      Samples        100 mL
Beach            Rain     after 0.5"      Geometric
               Amounts     to 0.99"         Mean

Cove Island    Normal         21              8.4
               Drought         8             29.1
Cummings       Normal         14             11.5
               Drought         6             54.3
West           Normal          8             26.6
               Drought         4            399.8

Note. The effects of abnormal events have been removed from this
table.

TABLE 3
Single Example of Significant Rain in Dry Weather Conditions, 0.73" of
Rain

Samples                      Cove Island          Horseshoe

                        E       M         W

Low tide samples,      110     90        <5        <5
  5 hours after
  rain ended
Low tide samples,      <5      <5     No sample    <5
  2 days later

Samples                    Cummings Beach           West Beach

                        E       M         W         N       S

Low tide samples,      15      <5        605       410     300
  5 hours after
  rain ended
Low tide samples,      <5      <5        <5        <5      <5
  2 days later

Note. All values are reported as enterococcus CFUs/100 mL.

TABLE 4
Rotifer Determinations Before an Incident of Sewage Grease Washing
onto the Shore

Date    Event(s)

6/17    A two-day rain event of 3.940" occurs and neighboring sewage
        treatment plants with combined storm and sewer systems are
        overwhelmed.

6/24    1. Routine sampling occurs in morning.

        2. Lifeguards call to report "red water" and samples are
        taken in the afternoon.

        3. Microscopic examination of the water showed large
        quantities of rotifers and a person familiar with this
        examination was contacted for assistance.

        4. Water determinations before the red color was reported:
        Enterococcus, 2; Fecal coliforms, 10.

        5. Water determinations after red color:

        Enterococcus      Fecal
                        coliforms    E. coli

             35            180         170
        Saprobic Index (Si)

                        Brachionus   Encentrum   Asplancha
                          Rubens      Lupus     Brightwelli

                            85         170          204
        Saprobic
        Index (Si)         3.2         2.4          2.3

6/25    Bacteria determinations: Water

                          Fecal
        Enterococcus    coliforms    E. coli

            77.5           335         270

        Bacteria determination: Sewage grease results in CFUs/g

        Enterococcus

        1,000,000

6/29    Bacteria determinations: Water

                          Fecal
        Enterococcus    coliforms

              5            <10

Note. Rotifer determined microscopically by Save the Sound's Research
Assistant Iliana Ayala and is reported as number of organisms per 10
fields. Bacterial testing on water was performed by Stamford's Health
Department Laboratory and is reported in CFUs/100 mL. Bacterial
testing on the solid material was performed by Stamford's Health
Department Laboratory and is reported in CFUs/g.