Autoregressive integrated moving average (Arima) model for simulation of groundwater level at Devasuguru Nala watershed, Raichur district

Abstract

Groundwater is a significant source of water in India, with approximately 65-70% of irrigation and 85-90% of the rural domestic water supply dependent on groundwater. India is one of the world's largest groundwater users, with an estimated annual groundwater extraction of over 230 cubic kilometers (km³). This high rate of extraction raises concerns about over-exploitation in many regions. For the effective management of groundwater, it is important to model and predict fluctuations in groundwater levels. The most widely used technique for time series analysis is, the Box Jenkins’ Autoregressive Integrated Moving Average (ARIMA) model is adopted for the study. Results showed that the groundwater levels had significantly declined from January 1999 to March 2017. The results indicated that seasonal decline in groundwater level for the observation well was 0.034 m/year and average annual decline was 0.7424 m/yr. The ARIMA candidate model [3, 0, 2] was identified as the best fit model for groundwater level time series modelling and forecasting in Devasuguru nala watershed region.