Drought forecasting is an essential component of efficient water resource management that helps water planners mitigate the severe consequences of water shortages. This is especially important in hyper-arid climates, where drought consequences are more drastic due to the limited water resources and harsh environments. This paper presents a data-driven approach based on an artificial neural network algorithm for predicting droughts. Initially, the observed drought events in the State of Kuwait were tested for autocorrelation using the correlogram test. Due to the cyclic nature of the observed drought time series, nonlinear autoregressive neural networks (NARs) were used to predict the occurrence of drought events using the Levenberg–Marquardt algorithm to train the NAR models. This approach was tested for the forecasting of 12- and 24-month droughts using the recently developed precipitation index (PI). Four statistical measures were used to assess the model’s performance during training and validation. The performance metrics indicated that the drought prediction was reliable, with Nash–Sutcliffe values of 0.761–0.878 during the validation period. Additionally, the computed R2 values for model forecasts ranged between 0.784–0.883, which indicated the quality of the model predictions. These findings contribute to the development of more efficient drought forecasting tools for use by water managers in hyper-arid regions.

The formation of subsurface freshwater lenses on top of brackish groundwater is a fascinating hydrologic phenomenon that creates groundwater supplies of great potential value in arid regions. Information on the recharge quantity and mechanism of these lenses is both scarce and uncertain. This study examines the formation and