Monitoring and modeling of trihalomethanes (THMs) for a water treatment plant in Istanbul

Because of increasing concern for both microbial control and disinfection by-products (DBPs) formation, water utilities are strictly examining and optimizing disinfection practices. In this study, modeling of trihalomethanes (THMs) formation at processed water of the Kagithane water treatment plant in Istanbul City was conducted. Data for THMs and other water quality and operational parameters were generated through a 12-month sampling program between January and December 2003. A multiple linear regression model was developed to predict THMs concentrations in processed water. Routinely measured parameters including total organic carbon (TOC), pH, temperature, and chlorine dose were used to develop the model for the prediction of THMs. Both pH (r = 0.963) and temperature (r = 0.921) were found to be the parameters of the highest statistical significance as predictors for THMs occurrence. The regression analysis resulted in a model that is directly applicable to the chlorination of raw waters. This indicated that the linear models developed could be used to estimate THMs concentration for different water quality and treatment processes with different operational conditions.

THMs are formed during the chlorination of waters containing humic substances. Changes in temperature, pH, TOC concentration, and applied chlorine dose directly influence THM concentra- ions in processed water. This article discusses the formulation of a model for predicting THM levels in processed water of a water treatment plant in Istanbul subjected to chlorination. With the use of multiple linear regression techniques, it is possible to develop models for simulating and predicting THMs during water treatment. This research resulted in a model that is directly applicable to the chlorination of raw waters. The model described above can be used to estimate THM concentration for different water qualities and operational conditions in processed water where chlorination is the only treatment in the process. Based on the validation results obtained, it can be stated that the model is the most applicable to chlorinated waters. This indicates that the linear model developed could be used to estimate THM concentration for different water quality and operational conditions where chlorination is the only treatment in the process. Due to the fact that very little data about THMs are currently available for the distribution system in Istanbul, the generation of data using a modeling approach is very useful. However, with given complexity of water quality and evolution in distribution systems, modeling results must be interpreted with caution.