Groundwater Modelling. A Comparison Between Multiple Regression and Artificial Neural Network Approaches

      Аннотация: Groundwater is being exploited indiscriminately to meet our ever increasing demand of water in different parts of the world. In India, the Gangetic plane is amongst the most fertile land of the country. Over the past couple of decades, intensive growth in agriculture, industries, and human population have increased the water demand substantially. As a result severe problems of groundwater table declination have taken place causing threat to future availability of water. Keeping these in view, a study was undertaken in the Ramganga-Bahgul interbasin of Uttar Pradesh, India, to investigate the groundwater behaviour, analyse the causes behind water table declination, and estimate the stages of groundwater development. The collected field data were used to develop groundwater models using multiple regression and artificial neural network (ANN) approaches for the prediction of seasonal water table depths below ground level in the study area. The performance of both multiple regression and ANN models were compared and evaluated.

Ключевые слова: Modelling, multiple regression, Artificial Neural Network, Groundwater hydrology