Guidelines for Canadian Recreational Water Quality

6. Microbiological Sampling and Analysis

6.1 Sampling

In recreational water quality investigations, the purpose of sampling is to obtain aliquots that are as representative as possible with respect to the microbiological properties of the area. Sampling should be conducted during the bathing season, but it is most appropriate when recreational waters are suspected as a source of waterborne disease. Regular sampling may not be necessary at all recreational water use areas. Historical data, combined with an annual environmental health assessment, may indicate that only occasional sampling is necessary. If a deterioration of water quality has occurred, then routine monitoring of the area should be carried out. Such an approach will allow health officials to concentrate their resources on beaches of questionable quality. By taking the following factors into consideration, an effective sampling program can be designed to optimize the estimation of fecal indicator bacteria in recreational waters.

6.1.1 Sampling locations

Most bodies of water used for recreational purposes frequently lack homogeneity with respect to their microbiological properties, thus making multi-point sampling necessary. The sites should be selected on the basis of information gathered during the environmental health assessment. Ideally, the sites chosen should be representative of the water quality throughout the whole bather exposure area. The selection of sites should pay particular attention to site-specific conditions that may influence the levels and distribution of indicator organisms and pathogens.

The sampling sites should include points of greatest bather activity as well as peripheral points subject to external fecal pollution. Natural or artificial streams discharging storm water and sewage can give certain sections of a body of water very different microbiological qualities than the body at large. The degree of heterogeneity can also be affected by rainfall, wind velocity and direction, and tides. In larger bodies, the contribution of local events is somewhat diminished by the large volumes of water involved.

The importance of sampling location has been considered in a number of studies. Brenniman et al. (1981), in a study of water sampling design at two Lake Erie beaches, observed that levels of indicator bacteria varied significantly with the time and day of collection, but not within the various sampling sites in the bathing areas. The authors concluded that at beaches where the dispersion of any fecal input is incomplete, sampling at various locations as well as during periods of maximum bather load should be conducted.

The collection of subsurface samples at wading depth should be considered where the water has been stirred up either by bather activity or by the person collecting the sample (Warrington 1989). Two recent epidemiological studies have discussed the importance of sampling in shallow water frequented by young children (Fattal et al. 1986; Seyfried 1987).

6.1.2 Frequency of sampling

A water sample provides a quantitative estimate of the bacteria present at a particular site and instant. As the total number of samples increases, the more representative the data will be of the overall water quality.

Samples should be collected at random intervals and at times of greatest user activity (i.e., mid-afternoon on weekends or holidays, as recommended by Sherry 1986), as well as at times when maximum fecal contamination can be expected (e.g., periods of storm water runoff and high onshore winds that upwell bottom sediments). If concentrations of fecal indicator bacteria fluctuate cyclically (e.g., effluent discharges at regular intervals or tidal variations), as shown by Churchland and Kan (1982), then samples should be collected during all phases of the cycle in addition to periods of high bather density.

The minimum recommended sampling frequency for routine investigations is five samples in not more than 30 days from each sampling location. At beaches with higher bather densities or at those known to have poor water quality, or in cases of suspected waterborne diseases associated with bathing, then the sampling frequency should be increased. The number of samples will be determined from the identification of sampling locations described above. Occasional sampling should be adequate in areas that historically have had acceptable water quality. However, if information gathered prior to the bathing season indicates that the water quality may have deteriorated, then routine monitoring should be initiated.

When analyses indicate that a single sample contains more than 4000 Escherichia coli or fecal coliforms/L or more than 700 enterococci/L, resampling of the area is required. The number of samples collected and their location should be sufficient to indicate the possible sources of contamination.

6.1.3 Sampling procedures for water

Samples for microbiological examination should be collected in sterile, 200- to 500-mL "environmental sensitive" containers. When sampling is done by hand, the bottle should be held near the base with one hand, the cap removed, and the bottle mouth plunged downward into the water. The bottle is tilted slightly upward to displace the air, then pushed forward against the current away from the hand and the boat or sampling platform (if used) to avoid contamination. The sampling depth should be 15 to 30 cm below the surface in both deep and shallow waters. When collecting is done with a sampling pole, the bottle should be fit into the holder in the recommended manner, the cover removed, and the sample collected, upstream away from the collector, by simulating the scooping motion of the hand-collected sample.

With either method, a small amount of sample should be poured out, leaving an airspace to allow for proper mixing prior to analysis. The cap should be replaced and the bottle labelled and stored in an ice chest. The samples must be collected and processed individually. Composite samples are not acceptable.

6.1.4 Sampling procedures for sediments

When evidence indicates that bathing beaches could be the source of waterborne diseases among bathers, sediment sampling and analysis for suspected pathogens are also recommended. Many investigations have demonstrated that pollution indicator bacteria and pathogenic bacteria survive for extended periods in sediments (e.g., Burton et al. 1987).

Sediment samples may be collected using sterile, 250- to 500-mL wide-mouth jars, observing the same precautions used with water sampling to ensure aseptic collection. In shallow waters, the jars are pushed along the bottom, collecting the material at the sediment-water interface, until half full. The excess water is poured off, and the sample is stored as described above. In deeper waters, sediment samplers used for collecting benthic invertebrates, such as the Ponar or Ekman grabs, can also be used (American Public Health Association 1989). When sediments are brought to the surface, a subsample is aseptically transferred from the centre of the material to the sterile jar.

6.1.5 Sample preservation and storage

Water and sediment samples should be maintained at 1 to 5°C and processed within 30 hours after collection. For transport to the laboratory, the sample bottles should be placed in an insulated ice chest containing melting ice or freezer packs; to prevent the possibility of contamination, total immersion of bottles in the water should be avoided. The samples should never be frozen. If freezer packs are used, the samples should be protected from direct contact to avoid freezing. Storage in the dark under these conditions (or at 4 to 5°C in a refrigerator) minimizes die-off and multiplication for at least 30 hours after collection. A sample preservation experiment conducted by Dutka and El-Shaarawi (1980) indicated that when water samples were stored at 1.5°C, the concentrations of indicator organisms were stable for at least 24 hours. Neither the original water temperature nor bacteria and nutrient load appeared to affect preservation.

If the microbiological counts are to be used in legal action, the samples must be delivered to the laboratory within six hours of collection and processed within two hours of receipt with proof of continuity of possession (American Public Health Association 1989).

6.2 Methods for Microbiological Analysis

6.2.1 Escherichia coli and fecal coliforms

The 16th edition of Standard Methods for the Examination of Water and Wastewater (American Public Health Association 1989) contains two official methods for the determination of fecal coliforms: the multiple tube fermentation or most probable number (MPN) procedure, and the membrane filtration (MF) technique.

Multiple Tube Fermentation or Most Probable Number (MPN) Method

This procedure is not an actual enumeration of bacteria but is a statistically derived index that gives the probable number of bacteria present in a sample. A series of replicate tubes of lauryl tryptose broth (usually five) is inoculated with decimal quantities of the sample and examined for growth or gas production after a 48-hour incubation at 35°C. Positive tubes are transferred to EC broth and incubated at 44.5°C for 24 hours. The presence of gas within 24 hours or less is considered a positive reaction and indicates coliforms of fecal origin. The number of positive tubes per dilution is referred to a standard MPN (Most Probable Number) table, which provides an estimate of the fecal coliforms per 100 mL of sample. If an estimate of E. coli is desired, positive EC tubes are streaked on EMB agar and incubated at 35°C for 24 hours, and isolated colonies are subcultured and identified by routine IMViC procedures.

The method is time-consuming to perform, requires large amounts of media, glassware, and incubator space, and requires 72 hours for a confirmed test. Except for turbid waters or water suspected of containing stressed organisms, the method has been replaced by the membrane filtration technique. However, the addition of 4-methylumbelliferone glucuronide (MUG) to coliform and fecal coliform broths for the direct detection of E. coli, described by Feng and Hartman (1982), should make it easier and quicker to enumerate E. coli in turbid samples.

Membrane Filtration (MF) Technique

In this procedure, fecal coliforms in water are measured directly. The sample (usually 100 mL) is passed through a filter that retains the bacteria; the filter is placed on the surface of an appropriate medium (mFC broth for fecal coliforms) and incubated. After 24 hours' incubation in a water bath at 44.5°C, blue colonies typical of fecal coliforms are counted and recorded as the number of fecal coliforms per 100 mL. In recreational water contaminated with sewage, dilution of the sample to avoid confluent growth may be necessary. The anaerobic incubation of membranes to suppress the growth of non-coliform bacteria has also been described (Doyle et al. 1984). One disadvantage with the mFC broth is its inability to distinguish between E. coli and other thermotolerant, lactose-fermenting species. To overcome this problem, an MF method for enumerating E. coli in water has been developed (Dufour et al. 1981). The procedure uses a medium (mTEC) for lactose-fermenting, Gram-negative bacteria, a resuscitation step for stressed organisms, and an in situ urease test to differentiate E. coli (urease-negative) from other thermotolerant fecal coliforms (mostly urease- positive). This method should be useful for counting E. coli in most surface waters. However, because some Klebsiella pneumoniae subspecies are also urease-negative, the method may not be useful in waters receiving industrial effluents known to contain high levels of Klebsiella pneumoniae. In this case, the mTEC medium incorporating indoxyl ß-D-glucoside might be more appropriate (Shaw and Cabelli 1980). Most recently, the incorporation of 4-methylumbelliferone glucuronide (MUG) into the various fecal coliform media to increase their specificity for E. coli has also been examined (Freier and Hartman 1987; Brodsky 1989; Young 1989). Various laboratories in Canada may wish to evaluate the applicability of these methods to recreational waters in their regions.

The advantages of the MF technique include reduction of space, labour, and equipment necessary, ability to test large volumes of water, rapidity and ease of testing, and a high degree of reproducibility. With the use of portable equipment, it can also be employed directly in the field.

Widespread use of the MF technique has revealed some unforeseen problems. Many investigators have shown that there are significant differences between various membranes in their ability to recover fecal coliforms from natural waters. For example, Tobin and Dutka (1977) concluded that membrane filters were not equivalent in their ability to recover bacteria from water samples and pointed out the great need for standardization. Apart from brand differences in filters, many studies have also shown that the MPN method often yielded better recovery of fecal coliforms than the MF technique. It is believed that the broths used in the MPN test provide a more favourable environment than the selective medium and membrane structure of the MF test for the recovery and growth of stressed fecal coliforms. However, despite this problem, the MF technique, particularly when used with resuscitation techniques, is probably sufficiently precise to detect small differences in the pollution index of a given area when sampled regularly.

6.2.2 Enterococci

Enterococci in marine and freshwater recreational areas are usually enumerated by the MF technique described by the U.S. Environmental Protection Agency (1985). After filtering a portion of the sample, the membrane is placed on mE agar and incubated at 41°C for 48 hours. The membranes are then transferred to EIA plates and re-incubated for an additional 20 minutes. All pink to red colonies with black or reddish-brown precipitates are considered enterococci. A 1-step modification of this MF procedure is also being investigated (Dufour 1989). Highly turbid waters and those directly influenced by chlorinated sewage should be examined by an MPN method, using azide dextrose broth followed by confirmation with Pfizer selective enterococcus agar (American Public Health Association 1989).

6.2.3 Pseudomonas aeruginosa

A variety of enumeration procedures for Pseudomonas aeruginosa in natural waters is available. Levin and Cabelli (1972) described an MF technique and medium (mPA) that were more efficient and accurate than the MPN methods in use. Dutka and Kwan (1977) confirmed their findings and, by slightly modifying the mPA medium and using a longer incubation period, increased the sensitivity of the test. Brodsky and Ciebin (1978) further modified the medium and reported recoveries of P. aeruginosa comparable with those of Dutka and Kwan after only 24 hours' incubation.

However, if turbid waters or sediments are examined, the MPN method must be employed (Environment Canada 1978; American Public Health Association 1989). This procedure requires extended incubation periods and confirmation of presumptive positive tubes. Using an MPN procedure, Seyfried et al. (1985b) reported higher recoveries of P. aeruginosa from sediments than from ambient waters.

6.2.4 Staphylococcus aureus

The American Public Health Association (1989) lists a tentative MPN method for enumeration of Staphylococcus aureus from waters. An MF technique designed to count S. aureus in swimming pools (Alico and Dragonjac 1978) was also found useful for recreational waters (Seyfried 1980). Recently, a new MF medium for enumerating total staphylococci as well as S aureus has been formulated (Borrego et al. 1987a).

6.2.5 Salmonella and Shigella

Many methods are available for the isolation of Salmonella and Shigella from water and sediments using concentration and enrichment techniques followed by identification procedures or detection by fluorescent antibody techniques (Environment Canada 1978; American Public Health Association 1989). MPN and MF techniques for the quantitative determination of Salmonella have also been described (American Public Health Association 1989).

6.2.6 Aeromonas

There are a few methods available for the enumeration of Aeromonas in fresh and marine waters. MPN methods were used to count A. hydrophila in an estuary (Kaper et al. 1981). MF techniques for both fresh and marine waters have also been described (Rippey and Cabelli 1979; Havelaar et al. 1987).

6.2.7 Campylobacter jejuni

At present, there are no standard methods for the enumeration or detection of Campylobacter jejuni in water (American Public Health Association 1989). However, the enumeration of Campylobacter spp. using MPN methods followed by steps to identify C. jejuni has been reported (Bolton et al. 1987; Carter et al. 1987).

6.2.8 Legionella

Media and methods for the isolation and enumeration of Legionella in water have been described (Calderson and Dufour 1984; Hsu et al. 1984; Voss et al. 1984). The American Public Health Association (1989) has also summarized available information on specimen collection, identification, and isolation of Legionella species.

6.2.9 Protozoa

The technique for recovery of protozoa from waters is complex and involves two steps: the concentration of large volumes of water, and a microscopic identification using ordinary, phase-contrast, or fluorescent microscopy.

The American Public Health Association (1989) details a sampling device used for the detection of Giardia lamblia. Quantification of Giardia cysts by membrane filtration has been suggested by Spaulding et al. (1983). Jakubowski and Ericksen (1979) reviewed the methods for the detection of Giardia cysts in water, and Sauch (1985) detailed their microscopic identification.

Cryptosporidium spp. can be concentrated from waters using polypropylene cartridge filters, as described by Musial et al. (1987). The identification of oocysts in river water has been reported by Ongerth and Stibbs (1987) and by Gallaher et al. (1989).

6.2.10 Viruses and coliphages

The development of methods for the concentration of viruses from large volumes of water (Wallis et al. 1972; Payment et al. 1976; Sobsey et al. 1980; Gerba and Goyal 1982; Block and Schwartzbrod 1982; Gerba 1983; Payment and Trudel 1988) and their detection by highly sensitive methods (Payment and Trudel 1985; Margolin et al. 1986) now allow the virological analysis of surface waters. The methodology for concentration and isolation of viruses in large volumes of water has been standardized to some extent (American Public Health Association 1989), so that monitoring of recreational waters is possible if epidemiological data indicate a need. The use of positively charged micro-porous filters has been recommended by Sobsey and Jones (1979), although wound fibreglass depth filters were found to be less expensive by Payment and Trudel (1988).

There are now available some rather simple, quick, and inexpensive procedures for the monitoring of coliphages and bacteriophages in water. One of the most sensitive procedures for enumerating coliphages from water or effluents is described in Standard Methods for the Examination of Water and Wastewater (American Public Health Association 1989) using E. coli (ATCC 13706) as host.

6.2.11 Toxic phytoplankton

The presence of potentially toxic blue-green species can be determined microscopically, but this technique cannot distinguish toxic from non-toxic strains because the strains look alike.

Rapid chemical analyses using reversed-phase, high performance liquid chromatography (Harada et al. 1988), HPLC and internal surface reversed-phase columns (Meriluoto and Eriksson 1988) and high performance, thin-layer chromatography (Jamel Al-Layl et al. 1988) have been proposed, for toxins that affect the liver from Microcystis aeruginosa and Anabaena flos-aquae to replace the previous, more time- consuming methodology using gel filtration (Krishnamurthy et al. 1986).

The standard mouse bioassay (Bishop et al. 1959; Elleman et al. 1978) provides a rapid general assessment of the presence and toxicity of hepatotoxins. The survival time is a measure of toxicity. Falconer et al. (1981) and Siegelman et al. (1984) also provide guides to interpreting the results. Codd et al. (1989) describe in vitro cytotoxicity tests, immunoassays and other procedures now emerging to supplement the mouse bioassay.

Definite chemical analysis for the alkaloid neurotoxin anatoxin-a from Anabaena flos-aquae can be performed within a matter of hours (Smith and Lewis 1987). Ikawa et al. (1982) and Sasner et al. (1984) describe chemical analyses for aphantoxins from Aphanizomenon flos-aquae.

A 100-mL glass jar with a snap cap lid should be used to collect water samples for microscopic identification and enumeration of algal species. When the sample is taken, it should be preserved by the addition of Lugol's solution from an eye dropper until the sample is the colour of tea. The sample should be kept refrigerated.

For the toxicity assay and the extraction and identification of toxins, samples should be collected in two 1-L Nalgene containers. Enough algal mass should be collected to fill each container about three-quarters full. The samples should be frozen immediately after collection and kept frozen to preserve the toxins from decomposition.

Animals suspected of dying from blue-green algae poisoning should be autopsied by the local veterinary surgeon. It may also be appropriate during any investigation to sample and test the water for priority pollutants and pesticides, bacteriological quality, routine chemical constituents, and nutrients.

7. Posting of Recreational Waters

When the appropriate authority has determined that a beach or body of water is not suitable for recreational use, the public should be notified. Normally this involves placing one or more signs in conspicuous places along the beach or shoreline. These signs should be clear and concise as to the health risk and recommended course of action. They should be written in simple understandable text and symbols. The authority making the determination should be clearly indicated on the signs. The signs should be left in place only as long as necessary and promptly removed when the health hazard no longer exists.