The application of a nonlinear autoregressive modeling approach with exogenous input (NARX) neural networks for modeling groundwater level fluctuation has been examined by several researchers. However, the suitability of NARX in modeling groundwater level dynamics in urbanized and arid aquifer systems has not been comprehensively investigated. In this study, a NARX-based modeling approach is presented to establish a robust water management tool to aid urban water managers in controlling the development of shallow water tables induced by artificial recharge activity. Temperature data series are used as exogenous inputs for the NARX network, as they better reflect the intensity of artificial recharge activities, such as excessive lawns irrigation. Input delays and feedback delays for the NARX networks are determined based on the autocorrelation and cross-correlation analyses of detrended groundwater levels and monthly temperature averages. The validation of the proposed approach is assessed through a rolling validation procedure. Four observation wells in Kuwait City are selected to test the applicability of the proposed approach. The results showed the superiority of the NARX-based approach in modeling groundwater levels in such an urbanized and arid aquifer system, with coefficient of determination (R²) values ranging between 0.762 and 0.994 in the validation period. Comparison with other statistical models applied to the same study area shows that NARX models presented here reduced the mean absolute error (MAE) of groundwater levels forecasts by 50%. The findings of this paper are promising and provide a valuable tool for the urban city planner to assist in controlling the problem of shallow water tables for similar climatic and aquifer systems.

Evaporation is the major water-loss component of the hydrologic cycle and thus requires efficient management. This study aims to model daily pan evaporation rates in hyper-arid climates using artificial neural networks (ANNs). Hyper-arid climates are characterized by harsh environmental conditions where annual precipitation rates do not exceed 3% of annual evaporation rates. For the first time, ANNs were applied to model such climatic conditions in the State of Kuwait. Pan evaporation data from 1993–2015 were normalized to a 0–1 range to boost ANN performance and the ANN structure was optimized by testing various meteorological input combinations. Levenberg–Marquardt algorithms were used to train the ANN models. The proposed ANN was satisfactorily efficient in modeling pan evaporation in these hyper-arid climatic conditions. The Nash–Sutcliffe coefficients ranged from 0.405 to 0.755 over the validation period. Mean air temperatures and average wind speeds were identified as meteorological variables that most influenced the ANN performance. A sensitivity analysis showed that the number of hidden layers did not significantly impact the ANN performance. The ANN models demonstrated considerable bias in predicting high pan evaporation rates (>25 mm/day). The proposed modeling method may assist water managers in Kuwait and other hyper-arid regions in establishing resilient water-management plans.

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.

Efficient water management plans should rely on quantitative metrics for assessing water resource shortage scenarios. This study develops a simplified precipitation index (PI) requiring precipitation data only in order to assess hydrometeorological droughts affecting various hydrological systems. The PI index is inspired by the famous Standardized Precipitation Index (SPI), and it aims to provide the same indication for drought severity and duration while overcoming the disadvantage of needing hydrological data normalization. Avoiding hydrological data normalization overcomes the non-satisfactory results of this procedure that were reported in previous studies. Analysis of groundwater drought drivers in the arid region of Kuwait is presented to test the index applicability at timescales 12 and 24 months using available historical precipitation data from 1958 to 2017. A bivariate joint probability analysis was conducted by Clayton copula to assess the occurrence of certain drought severities and durations. The results showed that PI is comparable to the original SPI and provides drought severity linearly propagating with respect to time. This index constitutes a simple means to help water managers assess and describe the impact of droughts in precipitation-controlled systems and establish appropriate water management plans.