Testing and validation of a multiple nonlinear regression model for predicting trihalomethane formation potential

Abstract

This paper summarizes a testing and validation analysis of a previously developed model for predicting trihalomethane formation potential (THMFP) in chlorinated waters containing THM precursors. The original model, in the form of a nonlinear multiple regression equation, tended to overpredict THM formation potential at lower chlorine concentrations and underpredict at higher chlorine concentrations. The model proved to be more accurate in simulating THM formation potential in waters with moderate levels of organic carbon/THM precursors. While the model did not provide a high degree of predictive accuracy, the general format of the nonlinear model represents a rational framework for developing source specific models applicable to a given water source.