| There are no indicators of human enteric virus presence in surface waters and, at present, the water quality industry relies upon reducing fecal load which identifies pathogenic risk from both animal and human sources. Two potentially non-biological indicators of viable enteric virus are fecal sterols, coprostanol and epicoprostanol. This study investigates fecal sterols in conjunction with various other parameters to generate an enteric virus presence predictive model using multivariate logistic regression. A location along the Kentucky River at an inlet to a potable water supply was monitored once a week for two and a half years using a total of 17 chemical, microbial, and viral parameters. Applying a new paradigm combining fecal load, source, and age indicators resulted in an enteric virus predictive model using fecal coliforms, epicoprostanol, and the atypical/total coliform (AC/TC) ratio that resulted in 84.5% and 78.6% correct positive and negative predictions, respectively. The predictive ability of enteric virus presence was not impacted when coprostanol replaced fecal coliforms as the load parameter in the model. In both cases, epicoprostanol presence was an important input parameter but the AC/TC ratio dominated the predictive power of the model.; The behavior of the AC/TC ratio in septage changed with a shift in microbial population over time, resulting in an environment where atypicals (AC) dominated confirming the use of the ratio as a fecal age indicator. A fecal sterol extraction method was developed using Soxhlet techniques that showed improved recovery from raw waters. Results indicate that future studies should utilize the fecal load, source, and age paradigm including fecal sterols, fecal coliforms, and the AC/TC ratio when evaluating waters for elevated pathogenic risk.; Keywords. Coprostanol, Epicoprostanol, AC/TC Ratio, Human Enteric Virus, Kentucky River. |
