Modeling temporal of groundwater level using basic techniques of time series analysis (Case Study: Ardabil Plain)

In most areas, groundwater resources play a major role in water supply of the needs in parts of agriculture, drinking and industry. The scope of studied area of Ardebil is 4804.7 square kilometers and the extent of Ardebil plain in this region is 820 square kilometers. This aquifer feeds through the direct infiltration of surface downfalls, returning water from agriculture, drinking and industry expenditures and also the underground drains. In this study, the basic methods of the analysis of time series includes models of autoregressive (AR), moving average (MA) and the combination of autoregressive and moving average (ARMA) were implemented on the data and the best model was chosen by using the test of Akaike coefficient data autocorrelation functions. Then, it was observed the results of model AR (2) are better than other models according to these tests and coefficients. Finally, a prediction was accomplished for a period of thirty years by using the same model. The results of the predicted values by time series shows a reduction of approximately 11 meters into the current situation of water table of aquifer in the case of constant consumption patterns and also without change in the procedure of the feeding in the aquifer during the next thirty years. The decision to manage the groundwater in this area is required, due to the restriction of the resources and the drawdown of water table and the particular sensitivity of this region about drinking water supply in the coming years.