Modeling water quality in rivers: A case study of Beylerderesi river in Turkey

Abstract

River pollution is a major environmental problem that has negative consequences for humans and wildlife alike. To prevent its consequences, the sources and severity of pollution must be determined by monitoring water quality in river basins, followed by the measures necessary to control the contamination. Models and computer simulation of water quality are important tools for predicting adverse effects of pollution along a stream, and they can help guide practical investments in stream health. In water quality models, parameters that are determined through optimization rather than through trial and error are required to ensure the reliability of the model. In this study, a continuous stirred tank reactor (CSTR) approach was used to model Beylerderesi stream as a dynamic model, and the kinetic parameters were determined through optimization. For the optimization step, the Sequential Quadratic Programming method was used. The model predictions indicated good agreement with experimental data. The Mean Absolute Percentage error values for dissolved oxygen and biochemical oxygen demand were calculated as 0.95% and 1.39%, respectively. Statistical analysis showed differences between river and effluent samples for all parameters measured.