Simulation of groundwater level fluctuations in response to main climate parameters using a wavelet-ANN hybrid technique for the Shabestar Plain, Iran

In recent decades, increasing global water demand, coupled with the effects of climate change, has led to increased variation in groundwater level depletion. In this work, the effect of climate parameters is investigated with respect to groundwater levels in the Shabestar Plain, Iran. In the first step, the best models for the study region were selected from the general circulation models provided under the Fifth Assessment Report of the United Nations Intergovernmental Panel on Climate Change. To increase the spatial resolution of the precipitation data, downscaling of the models was performed using the Long Ashton Research Station weather generator for three representative concentration pathway (RCP) scenarios (RCP2.6, RCP4.5, RCP8.5) for the future period 2020-2049. The results of these models illustrated an increase in temperature and a decrease in precipitation for the study region. In the next step, an artificial neural network (ANN) technique for studying aquifer behavior was used. To increase the efficiency of the model, spatial and temporal preprocessing of data was performed using k-means clustering and wavelet transform de-noising, respectively. A fuzzy inference system was also used as a tool for estimating groundwater extraction and reducing uncertainty of illegal extraction. The results of ANN for five selected observation wells showed correlation coefficients of 0.92, 0.86, 0.76, 0.57 and 0.94 for the simulation. The model simulation under the three above-mentioned scenarios and the trend in groundwater decline in the Shabestar Plain for the base and future periods illustrated that the groundwater level dynamics were not related solely to climate parameters and that the impact of anthropogenic factors would be high.