Artificial Neural Networks Modelling for Predicting Water Quality in the Surface Waters of Western Godavari Delta, India

Many human activities have been the main contributors to surface water contamination in recent years. However, the western Godavari delta region of Andhra Pradesh only seldom permits assessing the water quality's natural state due to the uneven distribution of industrial operations and agricultural farms. This study uses artificial neural networks (ANNs) to estimate surface waters’ water quality index (WQI) between 2014 and 2022. A prediction like this can reduce computing time, labour, and the chance of calculating errors. The ANN results demonstrate that convergent plots perform better when the coefficient of determination (R2) values is more significant. Electrical conductivity (EC) and total dissolved solids (TDS) are the most significant parameters in predicting WQI in the surface waters of the Godavari delta region, according to a sensitivity analysis used to demonstrate the significance of each parameter in the ANN's modelling process. The method described in this paper offers a practical and effective alternative to WQI assessment and prediction, particularly when compared to WQI calculation methods that require time-consuming calculations and multiple sub-index calculations for every single value, or range of standards, of the part water quality variables.