Modeling Escherichia coli inactivation during solar disinfection: Effects of UV intensity, water temperature, and turbidity

The study aimed to develop a comprehensive regression model to estimate the inactivation rate constant of Escherichia coli during Solar Disinfection (SODIS) of drinking water. The model incorporates key parameters: UV intensity, water temperature, and turbidity, including their interactions and quadratic terms. The effects of expressing water temperature as maximum absolute temperature ($T_m$) and maximum temperature increase ($\Delta T_m$) on multicollinearity, significance, and model adequacy were also investigated. Experiments were conducted over 5 months to obtain the regression dataset, covering various combinations of these parameters. The results showed that the model using UV intensity and maximum temperature increase ($\Delta T_m$) as predictors achieved the highest performance (R-square: 0.93, predicted R-square: 0.91, RMSE: 0.261). Using $\Delta T_m$ improved predictive accuracy, reduced collinearity, and enhanced significance compared to $T_m$. Turbidity in the range of 1 – 30 NTU was significant in 40 % of models. Interactions were found between UV intensity and temperature, and temperature and turbidity, while no interaction was found between UV intensity and turbidity. The study highlights the importance of considering all possible regression models to avoid misleading interpretations of parameter significance. The developed model can estimate day-to-day fluctuations in SODIS efficiency, exposure period, and SODIS applicability in various regions, providing valuable insights for optimizing SODIS treatment strategies.