Neural Network Modelling of TDS Concentrations in Cauvery River Water, Tamilnadu, India

Monitoring and Modeling of river water quality is one of the key elements in the global environmental monitoring policy and management. Artificial Neural Networks (ANNs) have been used for modeling hydrological elements that are highly non-linear in both spatial and temporal scales. With the help of the available data on the physico-chemical characteristics, an attempt has been made to model the concentrations of Total Dissolved Solids (TDS) in the Cauvery River of approximately 40km stretch from Bhavani to Noyyal. The validation of the neural network model showed good agreement for predictions of the TDS concentrations between observed and predicted values. The coefficient of correlation during the validation process was found to be 0.951 and the mean squared error was 0.015. In conclusion, the work supports the concept that the neural network approach is a successful method of modelling such complex and nonlinear behaviour of TDS in the river with different environmental conditions.