Simulation of groundwater level through artificial intelligence system

Abstract

Individually applying intelligent calculating tools, such as artificial neural network and fuzzy logic techniques, to a variety of problems is confirmed to be efficient. Recently, a growing interest in a combination of these methods has resulted in the neuro-fuzzy calculating technique. The application of the artificial neural network (ANN) and the adaptive neuro-fuzzy inference system (ANFIS) to groundwater level simulation, over 7 years from 2007 to 2013, in the Langat Basin, Malaysia, is presented in this paper. Moreover, to the time series of groundwater levels, the time series of the five most effective parameters of groundwater level, that is rainfall, humidity, evaporation, minimum temperature and maximum temperature, were applied to obtain the best input parameters for the models. The performances of the different models were studied through evaluating the related values of the mean squared error and correlation coefficient to identify an optimal model that can simulate the decreasing trend of the groundwater level and provide passable simulation. In the model, excellent performance in different statistical indices was shown. Finally, a relatively good agreement between the calculated values and their corresponding measured values for the groundwater level were found. Evaluating the results of the various kinds of models, it has been shown that the obtained results of the ANFIS model are superior to those obtained from ANNs.