Premium
Statistical limitations in the usefulness of total coliform data
Author(s) -
Rosen Jeffrey S.,
Sobrinho José A.H.,
Lechevallier Mark
Publication year - 2009
Publication title -
journal ‐ american water works association
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.466
H-Index - 74
eISSN - 1551-8833
pISSN - 0003-150X
DOI - 10.1002/j.1551-8833.2009.tb09859.x
Subject(s) - pooling , statistics , sample (material) , sampling (signal processing) , environmental science , mathematics , chemistry , computer science , filter (signal processing) , chromatography , artificial intelligence , computer vision
Most water systems— rates of coliform‐positive samples are well below the 5% trigger for the monthly maximum contaminant level (MCL) set by the Total Coliform Rule (TCR). This study reviewed results of 540,357 samples collected from 15 systems and represented more than 12 years of sampling. Of these samples, 3,912 (0.72%) tested positive for total coliform and 79 (0.01%) tested positive for Escherichia coli. No correlations were observed between samples positive for total coliform and any available environmental data, including chlorine residuals. Small systems (serving 501‐3,300 people) and very small systems (serving 500 or fewer people) had a significantly higher percentage of positive samples, but values in general remained far below the monthly MCL. No clear patterns in the number of violations by system type were discerned. According to a power analysis of these data, under current sampling strategies, few systems will be able to detect the effects of any mitigation following TCR violations because sample sizes are too small. Water systems taking fewer than 250 samples per month cannot detect changes smaller than 2.5%. For most water providers, detecting the effects of mitigation will require pooling multiple months of data before and after a corrective action is taken.