Forecasting Groundwater Production and Rain Amounts Using ARIMA-Hybrid ARIMA: Case Study of Deir El-Balah City in GAZA

Abstract

Forecasting as a data mining technique offers the opportunity to leverage the numerous sources of time-series data, to support business decision-makers for effective planning and derive value from historical data. The lack and fluctuating rainfall and the Israeli occupation caused a crisis in Palestine and in Gaza in particular. Water resources are decreasing by the increase of population as the passage of time. Groundwater is the only water resource used in Gaza. Moreover, the increase in the demand for groundwater and the decrease in rainfall water, which is the main source of groundwater, will lead to depletion of groundwater wells, thus increase the salinity rate. In this paper, we conducted the forecasting technique on both the groundwater and the rain amounts in Dear El-Balah city of Gaza using the following forecasting algorithms: Auto-Regressive Integrated Moving Average (ARIMA), Hybrid ARIMA:(ARIMA+ NN), (ARIMA+ ets), (ARIMA+ tbats). The best performance of applied algorithms on rainfall data according to Mean Absolute Percentage Error (MAPE) measure was ARIMA combined with NN which gave the MAPE = 21%. On the other hand, ARIMA combined with tbats was the best algorithm applied on wells production data which achieved MAPE= 8.9%. The results showed that after 5 years the amounts of rainfall and groundwater production in comparison with the period from (2013 to 2017) will decrease by 8.4%, 0.03%, respectively.