The Innovative Combination of Time Series Analysis Methods for the Forecasting of Groundwater Fluctuations

Groundwater fluctuations forecasting is a valuable tool for the intelligent management of groundwater resources that prevents additional costs to the system. In this research, time series analysis methods, including Auto-Regressive Integrated Moving Average (ARIMA) and Holt-Winters Exponential Smoothing (HWES), were used for developing a short-term model and the forecasting of the groundwater fluctuations in six piezometer wells around the Gohar Zamin Iron Ore Mine. To take advantage of all the features of the robustly developed ARIMA and HWES models, the innovative combination of these methods with specific weights was used to predict groundwater levels. The results of Diebold and Mariano’s test were applied for the innovative combination of the developed models in six piezometers, and well no. 2 with a p-value higher than 5% was selected due to stable environmental conditions and was analyzed and interpreted. For this purpose, 250 non-seasonal data of daily groundwater level were used, of which 200 data are for modeling, and 50 data are for water level forecasting. The research results show that among all the forecasting, the least-squares method in the innovation of combining models has the highest accuracy for forecasting the groundwater fluctuations in the short term. All the forecasts in this period time show a decrease in the groundwater level, indicating the effects of pumping wells in the Gohar Zamin Iron Ore Mine area.