Forecasting of groundwater level fluctuations using a hybrid of multi-discrete wavelet transforms with artificial intelligence models

Abstract Groundwater is often one of the significant natural sources freshwater supply, especially in arid and semi-arid regions, paramount importance. This study provides a new high accurate technique for forecasting groundwater level (GWL). The artificial intelligence (AI) models include neural network (ANN) multi-layer perceptron (MLP) radial basis function (RBF), adaptive neural-fuzzy inference system (ANFIS) models. Input data to monthly average GWL 17 piezometers. In this study, preprocessing including discrete wavelet transform (DWT) multi-discrete (M-DWT) simultaneously was utilized. results showed that hybrid M-DWT-ANN, M-DWT-RBF, M-DWT-ANFIS compared DWT-ANN, DWT-RBF, DWT-ANFIS as well than regular ANN, RBF, ANFIS models, had highest accuracy 1-, 2-, 3-, 6-months ahead. Also, M-DWT-ANN model best performance. Overall, illustrated using M-DWT method input can be valuable tool increase predictive model's efficiency. indicate potential M-DWT-AI improve forecasting.