Turbidity as a surrogate for Cryptosporidium removal by filtration in drinking-water QMRA models

Abstract

A quantitative microbial risk assessment (QMRA) with Monte Carlo simulation was used to estimate risks of Cryptosporidium infection from drinking water. Four different approaches to assess Cryptosporidium oocyst removal by filtration were compared using three relationships between turbidity removal and oocyst removal proposed in the literature, and an empirical model based on oocyst log removal credits associated with filtered effluent turbidity. Using a large database of a full-scale water treatment plant in Brazil, two of the turbidity removal-based models returned 1–3 log unit removal of Cryptosporidium oocysts and median values of Cryptosporidium annual infection risk estimates of approximately 10−2pppy (per person per year). The other model based on turbidity removal and the empirical model resulted, respectively, in 2–3.5 and 0–4 log removal of Cryptosporidium oocysts and corresponding infection risk estimates around 10−2pppy, close therefore to the health-based targets of the WHO guidelines for drinking water. The results of this study indicate that the empirical model appears to be particularly suitable for QMRA studies and that similar models should be tested in other site-specific assessments of treatment performance.