Bacterial source tracking in the karstic Duck River Watershed of Middle Tennessee: A comparison of methods

James Jay Farmer, Tennessee State University


To remediate bacterial contaminated waters the source of the contamination must be known. Tracking environmental bacteria to the host source origin is often referred to as bacterial source tracking. In this study Escherichia coli were collected from cow, deer, human, septic systems, sewage influent, and waterfowl and fingerprinted using the molecular method of pulsed-field gel electrophoresis and the biochemical methods of antibiotic resistance patterns and carbon utilization profiles. Discriminant analysis and cross-validation studies of fingerprints from cow, septic systems, and sewage influent indicate that combining variables from all three methods gives the best results and that carbon utilization profiles are more accurate than antibiotic resistance patterns or pulsed-field gel electrophoresis. The Biolog® identification system was used to develop the carbon utilization profile and calculate a similarity to a database of known organisms. The four different parameters utilized by the Biolog® system for recording the reactions were compared, and it was concluded that optical density and colorimetric methods are superior to the presence or absence of a reaction method. This study concluded that for some sources use of E. coli with similarities between 0.3 and 0.5 results in a greater accuracy in predicting sources than sources with similarities greater than 0.5. Of the 96 carbon sources, 47 show a reaction in E. coli in at least 20% of the reactions. When these carbon sources were selected a priori to discriminant analysis the ability to accurately predict sources was decreased. Environmental E. coli collected from the Duck River were classified against a database containing carbon utilization profile fingerprints of E. coli from cow, human, deer, and waterfowl host sources. A flow duration curve was constructed from 70 years of discharge data. When the fingerprints of all bacteria collected from site 1 were compared to the fingerprints of all the bacteria from site 2, the percentage attributed to each source were similar. When the samples were compared on the basis of flow conditions differences were noticed between the sites. This type of analysis emphasizes the importance of incorporating hydrology in bacteria source tracking studies.

Subject Area

Microbiology|Environmental science

Recommended Citation

James Jay Farmer, "Bacterial source tracking in the karstic Duck River Watershed of Middle Tennessee: A comparison of methods" (2006). ETD Collection for Tennessee State University. Paper AAI3203848.