Saleh H. Alhathloul

High-turbidity exceedance events are important indicators of abnormal coastal water quality conditions, but their relatively low frequency compared with normal observations makes reliable classification challenging. This study developed a scenario-based machine learning framework to classify high-turbidity exceedance using water quality and meteorological predictors from an automated seawater monitoring station along the Arabian Gulf. After quality control screening and temporal matching with NASA POWER meteorological variables, turbidity was converted into a binary target using a 2 NTU threshold, resulting in 4408 low-turbidity observations (85.3%) and 758 high-turbidity observations (14.7%). Four predictor scenarios were evaluated: meteorological-only selected variables, combined selected variables, water-quality-only all variables, and combined all variables. Random Forest, XGBoost, a Support Vector Machine, and K-Nearest Neighbors were tested under three class-imbalance treatments: undersampling, oversampling, and SMOTE. The meteorological-only scenario showed moderate classification ability, with ROC-AUC values of 0.81–0.86, but relatively low precision, indicating limited control of false alarms. Adding water quality predictors substantially improved classification performance, with the combined selected-variable scenario achieving ROC-AUC values of 0.96–0.98. Under oversampling, Random Forest achieved an accuracy = 0.95, precision = 0.83, specificity = 0.97, weighted F1-score = 0.95, and ROC-AUC = 0.98. The water-quality-only scenario also performed strongly, with accuracy values of 0.90–0.95 and ROC-AUC values of 0.95–0.96, confirming that direct water quality measurements carried most of the predictive signal. The combined all-variable scenario produced similarly high performance, with ROC-AUC values of 0.96–0.98, while variable selection reduced redundancy without loss of predictive skill. Computational analysis showed that the top-ranked XGBoost model under the combined all-variable oversampling scenario required only 0.08 s for training and 0.003 ms per sample for prediction. Overall, the results demonstrate that integrating selected water quality and meteorological predictors provides an accurate, efficient, and operationally practical framework for high-turbidity early-warning classification.

Continuous monitoring of seawater alkalinity is essential for maintaining chemical stability in coastal environments and supporting efficient operation of desalination and water-treatment systems; however, direct alkalinity measurements are often limited in temporal resolution. This study develops and evaluates a machine learning framework for estimating seawater alkalinity using quality-controlled daily and sub-daily water-quality observations collected from a coastal monitoring station along the Arabian Gulf coast of eastern Saudi Arabia during 2017–2023. Five machine learning models, Random Forest (RF), Gradient Boosting (GB), Extreme Gradient Boosting (XGB), Support Vector Regression (SVR), and K-Nearest Neighbors (KNN), are assessed under two configurations: a baseline setup relying on the original predictor variables and an enhanced setup incorporating wavelet-decomposed features to represent multiscale temporal variability. Model performance is evaluated using five-fold cross-validation and quantified using R², root mean square error (RMSE), and mean absolute error (MAE). Under the baseline configuration, ensemble-based models outperform single-estimator and distance-based approaches, with RF achieving the best performance (R² = 0.77, RMSE = 2.57 ppm, MAE = 1.71 ppm). The incorporation of wavelet-based feature enrichment leads to consistent performance improvements across all models, reflected by higher R² values and reduced RMSE and MAE. The wavelet-enhanced RF model exhibits the strongest overall performance, attaining a mean R² of approximately 0.91 together with an RMSE of about 1.6 ppm and an MAE of around 1.0 ppm, while also showing reduced variability across cross-validation folds. The XGB model shows notable improvement with wavelet enrichment, whereas SVR and KNN benefit mainly through moderate error reduction. Overall, the findings show that wavelet-based feature enrichment improves the accuracy and stability of ML models for seawater alkalinity estimation, with RF providing the most reliable performance for coastal monitoring applications.

Potential evapotranspiration (PET) and water balance (WB) are key indicators of hydroclimatic conditions and water availability, particularly in arid and semi-arid regions. This study investigates the interannual variability, long-term trends, and abrupt regime shifts in annual PET and WB across Saudi Arabia using multi-station observational data spanning 1985–2022. PET was estimated using a temperature-based approach suitable for data-scarce arid environments, and WB was calculated as the difference between precipitation and PET. Non-parametric statistical methods were applied to assess trend magnitude and significance, while Pettitt’s change-point test was used to identify abrupt shifts at both regional and station scales. The main findings show a widespread and spatially coherent increase in atmospheric evaporative demand, with predominantly positive PET trends at both regional and station scales, accompanied by persistently negative and increasingly declining WB values, indicating a long-term intensification of water deficit across much of the country. Spatial patterns of PET and WB closely follow gradients in energy availability and temperature, confirming the dominant influence of warming-driven processes on hydroclimatic conditions in this arid environment. Change-point analysis identifies a statistically significant regional hydroclimatic regime shift during the late 1990s, characterized by an abrupt increase in PET and a concurrent deterioration of WB, marking the onset of a more water-limited climatic regime. At the station scale, the timing and significance of detected change points display pronounced spatial heterogeneity, reflecting the modulation of regional climatic forcing by local climatic and geographic factors. Overall, the results demonstrate that increasing evaporative demand, rather than precipitation variability alone, has become a primary control on water availability across Saudi Arabia, highlighting the importance of explicitly accounting for hydroclimatic non-stationarity in water resource assessment and long-term planning under continued warming conditions.

Understanding the historical patterns of drought changes is important to effectively manage and mitigate drought. This paper aims to provide a quantitative assessment of the spatiotemporal drought patterns in Saudi Arabia from 1985 to 2022. The study used the Standardized Precipitation Index (SPI) to examine drought patterns on both monthly and yearly timescales. The findings indicate a significant trend of increasing drought conditions in certain regions of the Kingdom from 1985 to 2022. The average rates of change for SPI-03, SPI-06, and SPI-12 were found to be −0.003 yr⁻¹, −0.0034 yr⁻¹, and −0.0099 yr⁻¹, respectively. Droughts were more frequent and persistent in the northern regions of the country, while the western region experienced severe and intense droughts. There were fewer drought occurrences before 2000, but droughts became more frequent after 2000, with large-scale impacts occurring during 2007–2008 and 2013–2014. These findings have important implications for water management strategies and can help mitigate the effects of drought, as they identify hotspot regions across Saudi Arabia at different timescales. Overall, it is important to implement province-specific efforts to reduce environmental vulnerabilities to droughts.

Many dammed reservoirs in dry climate conditions witness high sediment inflow rates due to higher soil erodibility, yet there are limited actual sediment influx measurements. Therefore, this study first applies the Soil and Water Assessment Tool (SWAT) hydrologic model to simulate reservoir sedimentation inflow to the Haditha Reservoir. Next, utilizing sediment inflows estimated by the SWAT model, the Trap Efficiency Function (TEF) is employed to estimate its remaining storage capacity and its useful life at multiple reservoir water levels. Calibration (1986–1997) and validation (1998–2007) of the SWAT model were conducted at three streamflow gaging stations and one sediment station located upstream of the reservoir. Results show that the SWAT model performed better during calibration than during the validation period for all streamflow and sediment gaging stations. In addition, modeled streamflow and sediment predictions were relatively more accurate on a monthly scale than on a daily scale. Simulated daily sediment inflow to the reservoir demonstrates slightly lower accuracy than daily streamflow, where the Coefficient of Determination (R²) and Nash-Sutcliffe Efficiency values are 0.34 and 0.32 in the case of sediment load, compared to 0.39 and 0.33 for streamflow, respectively. Reservoir storage capacity for the period (1986–2005) shows a continuous decrease with time at all reservoir water levels, which indicates an increase in sediment accumulation. According to measurements taken between 1986 and 2005, sediment accumulation has reduced the reservoir’s capacity by approximately 15% at a water level of 112 m (the lowest water level in the reservoir). During the same period, the storage capacity loss at 147 m (the design working water level in the reservoir) was calculated to be 35%. Over 19 years of operation (1986–2005) at the 147-m water level, the total sediment buildup in the reservoir is estimated at 3.2 million tons. Notably, about one-third of this sediment was deposited in the five-year span from 2000 to 2005.