Runoff is a Key Constraint Toward Water Table Fluctuation Using Neural Networks: A Case Study

Monitoring water table fluctuations is essential, and it is even more important to predict the groundwater level to plan for the future needs. Though it is challenging task to model groundwater fluctuations due to its nonlinear nature, Artificial Neural Network (ANN) is the most robust tool to monitor and forecast groundwater level, where other empirical model fails. In this study, a casecade forward neural network and co-adaptive neuro-fuzzy inference system architecture has been designed and trained to learn the past water table fluctuations, to predict the future groundwater level. The model efficiency is analyzed and validated through the field observed value. After proper validation, the model is applied to predict the future water levels in the wells.