Full Scale Data Analysis for Treatment System Validation and Treatment Modeling

Assessment of treatment efficacy is a critical part of quantitative microbiological risk assessment (QMRA) of drinking water. To develop a protocol for treatment assessment, historical data on treatment efficacy of 12 water treatment plants was collected in the MicroRisk project. By analyzing the data typical treatment efficacy was determined for several treatment processes and organisms including variation of the efficacy. The treatment efficacy for pathogen, indicator organism and surrogate removal was compared. The ratio between E. coli and Campylobacter reduction by filtration or ozonation varies around 1:1, therefore E. coli was used as an indicator of treatment efficacy for Campylobacter. Turbidity can only be applied as a surrogate for organism removal by physical processes when their relation is verified by full-scale data. At several treatment sites no relation between turbidity and organism removal was apparent. Models are used to estimate or predict treatment efficacy from process conditions. Commonly used models strongly overestimate ozonation efficacy for E. coli inactivation at full scale. This shows that these models need to be verified at full scale before applying them to calculate pathogen inactivation. Pathogen, indicator, surrogate and model data that best describes the efficacy of a treatment step was combined in a treatment assessment. Treatment variation was included by using all the available data in a Monte Carlo analysis. The case study showed the added value of using surrogate data to calculate the variation of pathogen concentration in treated water.