Premium
Guidelines for indicator bacteria in waters: uncertainties in applications
Author(s) -
ElShaarawi A. H.,
Marsalek J.
Publication year - 1999
Publication title -
environmetrics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.68
H-Index - 58
eISSN - 1099-095X
pISSN - 1180-4009
DOI - 10.1002/(sici)1099-095x(199907/08)10:4<521::aid-env372>3.0.co;2-3
Subject(s) - guideline , environmental science , statistics , limit (mathematics) , sample size determination , water quality , log normal distribution , recreation , quality (philosophy) , sample (material) , sampling (signal processing) , computer science , mathematics , ecology , biology , medicine , mathematical analysis , philosophy , chemistry , epistemology , pathology , chromatography , filter (signal processing) , computer vision
Microbiological water quality guidelines have been established in most countries to protect the users from the risk of waterborne diseases. A guideline requires that a minimum number of water samples be collected during a period of time and analysed for an indicator organism. One or more summary statistics are computed and compared to limits specified in the guidelines. When at least one of the limits is exceeded, then certain actions must be taken. Since uncertainty is involved in the implementation of each of these requirements, the use of the guideline is intended to control and not to eliminate the risk from the use of impaired waters. The sources of uncertainties are discussed using two Canadian guidelines for microbiological recreational water quality. However, the methods used are applicable in general to guidelines' specification and use. Criteria are proposed for the selection of an indicator organism and the specification of its limits. The probabilities that quality of recreational water meets the federal guideline are estimated by simulation from the lognormal distribution. These probabilities reveal that imposing a single limit on the maximum eliminates the need for also imposing a limit on the geometric mean. The distributions of the maximum in samples of various sizes are derived by simulation and used to estimate the number of samples needed to detect a predetermined increase in the mean bacterial level. These distributions can be also used to determine the increase in the mean that a sample of a specific size can detect. Finally, the Ontario provincial guideline is compared to its federal counterpart. When the number of samples analysed is less than 15, the Ontario guideline is more conservative, but the reverse is true for larger sample sizes. Copyright © 1999 John Wiley & Sons, Ltd.