Surface Water Quality Assessment and Prediction Modelling of Kathajodi River

Abstract

Water quality index (WQI), a technique of rating water quality, is an effective tool to access quality and ensure sustainable safe use of water for drinking. The main objective of the present study is to access the surface water quality of Kathajodi river for knowing the suitability of drinking purpose by calculating the WQI. Samples were collected from selected locations during different seasons (winter, summer, rainy) over a period of 3 years (2011, 2012, 2013). Water quality assessment was carried out for the parameters like pH, total dissolved solids, total suspended solids, Alkalinity, Biological Oxygen Demand(BOD), Dissolved Oxygen(DO), Chloride, Nitrate, Alkalinity, Total Hardness, Calcium, Magnesium. The main objective is to develop a model to assess and predict the water quality changes of Kathajodi River Basin Odisha, India using neural networks and compared with the statistical methods. The result shows the proposed ANN prediction model has a great potential to simulate and predict the strongly correlated parameters like TSS (Total Suspended Solids), TDS(Total Dissolved Solids), Alkalinity, BOD(Biological Oxygen Demand)with Mean Square Error (MSE) : TSSMSE = 1.78 ; TDSMSE = 0; AlkalinityMSE = 3.77 and BODMSE = 8E-03.The Neural Network model has been compared with Linear Regression model to find out the best modelling approach for the study area. And it is concluded that the neural network model is superior to Linear Regression Model.