A Stochastic Modelling Technique for Groundwater Level Forecasting in an Arid Environment Using Time Series Methods

Abstract

In arid and semi-arid environments, groundwater plays a significant role in the ecosystem. In the last decades, groundwater levels have decreased due to the increasing demand for water, weak irrigation management and soil damage. For the effective management of groundwater, it is important to model and predict fluctuations in groundwater levels. In this study, several time series models were applied to predict groundwater level forecasting in Kashan plain, Isfahan province, Iran. At first, to reduce the calculation volume, the water table depths in 36 piezometric wells were clustered based on the Vard algorithm. Consequently, we categorized the 36 wells into five clusters. For each cluster, five time series models of autoregressive (AR), moving-average (MA), auto-regressive moving-average (ARMA), autoregressive integrated moving-average (ARIMA) and seasonal auto-regressive integrated moving-average (SARIMA) were applied. The results showed that the AR model with a two-times lag (AR(2)), shows the best forecasting of groundwater level for 60 months ahead of the five clusters, with a high accuracy of R2 (0.89, 0.89, 0.95, 0.95 and 0.75 in clusters 1 to 5, respectively). According to the results, the average groundwater level fluctuation in 2010 and 2016 was 74.58 and 80.71 m, respectively. With these conditions, the groundwater depletion rate would be 1.02 m per year in 2016. We combined several time series models for a better performance of prediction of groundwater level. We can conclude that combining time series models have an advantage in terms of groundwater level forecasting.