Response of the Shabestar Plain aquifer to climate-change scenarios through statistical and hybrid soft computing techniques

Abstract

Groundwater-level prediction has a significant impact on sustainable water-resources management. In this research, the impact of climate change on groundwater level on Shabestar Plain, Iran, was investigated. First, through General Circulation Models (GCM), temperature and precipitation values were projected under future scenarios of Representative Concentration Pathway (i.e., RCP2.6, RCP4.5, RCP8.5). In the next phase, groundwater-level variations in the Shabestar Plain through the Autoregressive Integrated Moving Average (ARIMA) statistical model and Artificial Neural Network (ANN) and Nonlinear Autoregressive Neural Network with Exogenous Inputs (NARX) soft-computing techniques were evaluated. The climate projections showed that the temperature would increase while precipitation would decrease in the future period (2020–2050). Comparing performance criteria among statistical and soft-computing models in simulation groundwater levels demonstrated that the Wavelet-NARX has superior performance compared to the others. Prediction of future groundwater levels showed that the average groundwater levels would decline in a future period (2020–2025) with 1.2 m, 2.2 m, and 3.0 m, under RCP2.6, RCP4.5, and RCP8.5 scenarios, respectively. These results emphasize an urgent need for dynamic management for the conservation of water resources in the study area.