Groundwater depth monitoring and short-term prediction: applied to El Hamma aquifer system, southeastern Tunisia

In arid and semi-arid areas, such as Tunisia, groundwater depth assessment and monitoring is an important task for groundwater resources sustainable management. The objective of this work is to assess and forecast groundwater depth (GWDs) at short-term, under different impact factors, using stochastic time series (STS) and artificial neural network (ANN). Each model was conceptualized and implemented. Groundwater depth, rainfall, pumping rate and temperature over the period 30years extending from 1986 to 2016 were used. Root mean squared error, bias, R-2, AIC, and BIC were used as sensitivity model performance criteria. Models were implemented, trained, and tested over an observation period and forecasting groundwater depth at short-term over a period 2017-2030. Results show that groundwater abstraction rates increase around 0.46Mm(3)/year, whereas a groundwater depth is in continuously drawdown with a rate of 0.3m/year which requires a monitoring network for sustainable management of water resources. The results are promising and indicate an outstanding suitability of stochastic time series and ANN models for groundwater depth assessment and predictions at short time. The predicted results at short-term will help to draw attention of local authority for clear groundwater sustainable management policy for studied area.