Research papersConjunction of radial basis function interpolator and artificial intelligence models for time-space modeling of contaminant transport in porous media

•Hybrid black box-physical-based model developed for GFCT.•ANN and ANFIS were used for temporal modeling of GL and CC.•MQ-RBF meshless method was employed for spatial computation of GL and CC.•De-noised data enhanced the performance of the hybrid AI-meshless model.•The efficiency of ANFIS-meshless model was more than ANN-meshless model up to 13%.

As an innovation, both black box and physical-based models were incorporated into simulating groundwater flow and contaminant transport. Time series of groundwater level (GL) and chloride concentration (CC) observed at different piezometers of study plain were firstly de-noised by the wavelet-based de-noising approach. The effect of de-noised data on the performance of artificial neural network (ANN) and adaptive neuro-fuzzy inference system (ANFIS) was evaluated. Wavelet transform coherence was employed for spatial clustering of piezometers. Then for each cluster, ANN and ANFIS models were trained to predict GL and CC values. Finally, considering the predicted water heads of piezometers as interior conditions, the radial basis function as a meshless method which solves partial differential equations of GFCT, was used to estimate GL and CC values at any point within the plain where there is not any piezometer. Results indicated that efficiency of ANFIS based spatiotemporal model was more than ANN based model up to 13%.