A mathematical model for removal of human pathogenic viruses and bacteria by slow sand filtration under variable operational conditions

Slow sand filtration (SSF) in drinking water production removes pathogenic microorganisms, but detection limits and variable operational conditions complicate assessment of removal efficiency. Therefore, a model was developed to predict removal of human pathogenic viruses and bacteria as a function of the operational conditions. Pilot plant experiments were conducted, in which bacteriophage MS2 and Escherichia coli WR1 were seeded as model microorganisms for pathogenic viruses and bacteria onto the filters under various temperatures, flow rates, grain sizes and ages of the Schmutzdecke. Removal of MS2 was 0.082–3.3 log10 and that of E. coli WR1 0.94–4.5 log10 by attachment to the sand grains and additionally by processes in the Schmutzdecke. The contribution of the Schmutzdecke to the removal of MS2 and E. coli WR1 increased with its ageing, with sticking efficiency and temperature, decreased with grain size, and was modelled as a logistic growth function with scale factor f0 and rate coefficient f1. Sticking efficiencies were found to be microorganism and filter specific, but the values of f0 and f1 were independent of microorganism and filter. Cross-validation showed that the model can be used to predict log removal of MS2 and ECWR1 within ±0.6 log. Within the range of operational conditions, the model shows that removal of microorganisms is most sensitive to changes in temperature and age of the Schmutzdecke. ► A model was developed to predict microorganism removal by slow sand filtration. ► Temperature, Schmutzdecke age, grain size and filtration rate are incorporated. ► The model can be used as part of quantitative microbial risk assessment.