Groundwater Levels Modeling and Forecasting for Ajabshir Plain by a Seasonal Autoregressive Integrated Moving Average (SARIMA)

Ajabshir plain is one of the most important agricultural areas in East Azarbaijan province, which groundwater resources were applied more than half the water requirement of Agricultural productions in this plain. Continuous utilization of groundwater caused a decline of 5.5 m from since 1984. Therefore, optimal and sustainable exploitation of groundwater resources in this plain is a management necessity. Consequently, modeling and prediction of the exploitation process could be accomplished by an appropriate technique. This research was conducted with the aim of analyzing the groundwater level variations in Ajabshir plain with time series statistical models due to the ability of time series techniques to model and predict the behavior of temporal variation in water engineering. Also, in this study, the groundwater level decline was modeled for 16 years with 16 models. A Seasonal Autoregressive Integrated Moving Average (SARIMA) was recognized as the most appropriate pattern. Modeling, testing and prediction model was as follows: about 50% of the data for modeling (from 1984 to 2006), 25% for the test (from 2006 to 2017) and 25% (from 2017 to 2028) was used for prediction. Results showed that the average volume of groundwater drop in the 11th year of the forecast (2028) is equal to 26.5×106 m3 under continued exploitation with the current conditions for agricultural production in the plain. Different saving scenarios, including 0, 10, 15, 20, 25 and 30 percent savings starting from 2018 were considered for optimal management of groundwater application. In the eleventh year, about ×106 m3 of groundwater will be saved with savings of only 10% from 2018.