Development of a Method to Determine Sources of Fecal Pollution in Water

Crop and Soil Environmental News, September 1998

Charles Hagedorn
    Professor, Waste Management and Water Quality
R. B. Reneau, Jr.,
    Professor and Biotechnology Specialist

Introduction
Many surface and ground waters in Virginia and the U.S. are contaminated by fecal pollution. This contamination results in increased health risks to persons exposed to the water, degradation of recreational and drinking water quality, and nutrient loss from the watershed to the Chesapeake Bay. Unfortunately, the source(s) of this pollution often cannot be readily determined. Knowledge of the type of pollution source could aid in the restoration of the water quality, reduce the amounts of nutrients leaving the watershed, and reduce the danger of infectious disease from exposure to these waters. The following quote was taken from the Virginia 305b water quality report in the National Watershed Database (maintained by EPA and located at http://www.epa.gov/surf/surf_search.html), "Virginia: Fecal coliform bacteria are the most widespread problem in rivers and streams. Agriculture and pasture land contribute much of the fecal coliform bacteria in Virginia's waters." Also, according to a Virginia Department of Environmental Quality report, in 1998 there were 1,165 stream miles in Virginia that were "impaired" because they exceeded bacterial standards.

Rationale for Using the Fecal Streptococci
While most emphasis on bacterial indicators of water quality has been placed on fecal coliforms, there are several good reasons to consider the fecal streptococci as well. Using natural antibiotic resistance patterns as identifying markers has been tried, with little success, to determine sources of fecal coliforms. To date, however, these patterns do appear to have potential with the fecal streptococci. There are standards for recreational water quality with the enterococci (33 cells per 100 ml for fresh water), so these organisms are still useful as indicators. The fecal streptococci tend to persist longer in the environment than fecal coliforms and this may limit their usefullness as indicators of recent fecal pollution. However, for determining the sources of fecal pollution, an indicator with a longer survival time can be an advantage. Lastly, there are some potential sources of fecal contamination (e.g. composted animal and poultry litters, and advanced-treatment Class B biosolids) where it is difficult to isolate fecal coliforms while there is no difficulty in isolating fecal streptococci. Fecal coliforms would not be suitable for sourcing contamination from these types of materials.

Previous Research on Source Differentiation
Numerous attempts have been made to develop methods to determine sources of fecal pollution and to date most have not proven useful. These include the ratio of fecal coliforms to fecal streptococci, source-specific bacteriophages, differences in the species composition of fecal streptococci between various types of animals, and patterns of antibiotic resistance in fecal coliforms. Dr. G. Simmons (at Virginia Tech) has successfully used fatty acid profiles and DNA genetic fingerprinting in Escherichia coli to resolve non-point fecal coliform sources in tidal inlets in the Chesapeake Bay. The possibility that antibiotic resistance patterns could serve as markers to determine sources with the fecal streptococci was first investigated by Hagedorn 20 years ago. Dr. B. Wiggins (at JMU) has successfully used antibiotic resistance in the fecal streptococci to differentiate between human and animal sources from a small watershed in Shenandoah County, Va. The recent development of DNA fingerprinting to identify individual strains and substrains of bacteria is an indication that this technology may be suitable in source differentiation of the fecal streptococci as well.

Methodology
Six types of known sources have been collected to build the Va Tech isolate database: beef cattle, dairy cattle, deer, chicken, Canada goose, and human (influent of the local municipal sewage treatment plant and septic tank effluent). For each animal source, fresh manure was the source material while the human sources were obtained from the local wastewater treatment facility and an onsite waste treatment system from an individual home. After collection, all samples were placed on ice and processed within 12 hours. The fecal streptococci were isolated and identified based on procedures in "Standard Methods for the Examination of Water and Wastewater." Details of these procedures can be seen at: http://www.bsi.vt.edu/facultyfiles/biol_4684/water.html. For source differentiation, each isolate was tested for antibiotic resistance patterns, salt tolerance, starch hydrolysis, and growth at different temperatures (20 tests per isolate).

Two types of statistical analysis were used to separate the isolates by source. The first is Cluster Analysis (CA). Clustering is a technique of grouping variables (isolates) together that share similar values. A database is built where all known isolates are grouped into a cluster by source. Unknowns are then placed in the most likely cluster by the statistical program. The second is Discriminate Analysis (DA). The DA procedure compares each set of isolates from an unknown source against the database of known sources and then classifies each isolate into one of the possible sources. The rate of correct classification for each set of unknowns is then determined by averaging the percentages of correctly classified isolates from known sources.

Database Development
Approximately 6,600 fecal streptococcus isolates from samples of cattle (beef and dairy), human, Canada goose, deer, and chicken wastes were collected around the Blacksburg, Va., area and characterized (Table 1). It has taken the past two years to build the isolate database and to test many different antibiotics and other biochemical tests to determine which provided the best source separations. Additional isolates are being added to the database weekly. Cluster Analysis (CA) was able to classify all the isolates in Table 1 into six separate groups, and successfully differentiate between human and animal sources of fecal pollution. Images of the results from CA will soon be available on the Va Tech website listed above (see Methodology).

Table 1. Isolate database showing the numbers of fecal streptococci from known sources.

Source	Number of Isolates
Human		1,579
Canada Goose		1,284
Beef Cow		1,398
Deer		1,245
Dairy Cow		528
Chicken		624
Total		6,658

Database Validation
The first validation test of the database involved collecting isolates from known sources at other locations in Virginia and using Discriminate Analysis (DA) to determine the level of correct classification (Table 2). The separations were significant on the basis of Duncan's multiple range test (0.05), and the lowest correct classification rate was 84% for deer and the highest was 94% for human. While the number of isolates from other locations tested to date is small, the results are encouraging as the overall correct classification rate was 91%.

Table 2. Discriminate analysis results for classification of known sources from other locations.

Source	Number of     Isolates			Number Correctly Identified			Correct Classification Rate (%)
Human		384			361			94
Canada Goose		95			86			90
Beef Cow		126			108			86
Deer		114			96			84
Total		719			660			91

Conclusions
The Virginia Tech database did not need to be adjusted (changing the antibiotic selection or concentrations), indicating its suitability for correctly sourcing fecal streptococci isolates from other locations. The initial validation test was successful, based on an overall correct classification rate of 91%. The statistical analysis methods provided reasonable answers that were in agreement on both database development and the initial validation test for fecal source identification. Next month's article will describe the first large-scale validation test that includes greater numbers of isolates and both knowns and unknowns from a watershed in Clarke County, Va.

Visit Virginia Cooperative Extension