Groundwater Level Fluctuations in Coastal Aquifer: Using Artificial Neural Networks to Predict the Impacts of Climatical CMIP6 Scenarios

摘要

Abstract Groundwater resources play a crucial role in supplying water for domestic, industrial, and agricultural use. In this study ACCESS-CM2, HadGEM3-GC31-LL, and NESM3 were selected for validation from Coupled Model Intercomparison Project Phase 6 (CMIP6). In the following, the feedforward neural network was employed to predict monthly groundwater level (GWL) based on the emission scenarios of the sixth IPCC report (SSP2-4.5 and SSp5-8.5) for the next two decades (2021–2040) in the Sari-Neka coastal aquifer near the Caspian Sea, Iran. In this regard, the monthly maximum and minimum temperature, precipitation, and water table of previous month from four piezometers from 2000 to 2019 were used as input variables to forecast GWL. The evaluation of the three GCM models demonstrated that the ACCESS-CM2 provided the best values of the R2 and RMSE with observation parameters. The results of r, R2, RMSE, and MAE were evaluated for the model and indicated good performance of the model. The results also illustrated that under such mentioned scenarios, the mean monthly temperature would rise approximately from 0.1–1.2?°C. In addition, the mean monthly precipitation is likely to witness changes from -10% to 78% in the next two decades. As a result, this seems to lead to improvement and recharge of groundwater level for the near future. The results can help managers and policymakers to identify adaptation strategies more precisely for basins with similar climates.