Estimation of Temporal and Spatial Variations of Groundwater Level by Combining Intelligent Models and Geostatistical Methods)Semnan Plain(

Due to the complexity of the nature of groundwater systems as well as the limitations of borehole drilling, temporal and spatial modeling of groundwater levels are not readily possible. Artificial intelligence methods such as RBF, ANN, SVR, ANFIS and ARIMA modeling and their combination with geostatistical methods have been used to find useful solutions for spatial prediction of groundwater level. The case study of this study is Semnan plain. The first step is the temporal modeling of groundwater level using different methods. The results showed that the ANFIS model provides more accurate prediction of the monthly groundwater level in the experimental stage than other methods (R2 = 0.994 and RMSE = 0.041). Next step, the ANFIS output data is used as input data for the geostatistical model and the linear kriging model is selected as the best model for spatial development of groundwater level. (R2 = 0.8889 and RMSE = 2.376). The results of this study showed that the combination of ANFIS model and linear kriging model is an appropriate method for temporal and spatial prediction of groundwater level.