Diffuse nitrate (NO₃⁻) contamination is a critical environmental concern threatening the quality of coastal groundwater resources, particularly in regions undergoing agricultural intensification and rapid land use changes. This study presents an explainable deep learning framework for predicting nitrate concentrations and identifying areas at risk of elevated contamination. The framework integrates key hydrochemical parameters electrical conductivity (EC), chloride (Cl⁻), organic matter (OM), and fecal coliforms (FC) with remote-sensing derived indicators, including the Normalized Difference Vegetation Index (NDVI) and land use/land cover (LU/LC). Two deep learning models were evaluated in this study: a Multilayer Perceptron (MLP) and TabNet, a novel attention-based architecture for interpretable tabular data. TabNet outperformed MLP, achieving an overall accuracy of 81.60% and a Macro-averaged recall of 84.13%, while providing transparent feature attribution. LASSO regression identified FC (0.52) and EC (0.48) as dominant predictors, highlighting the combined influence of domestic wastewater and agricultural runoff on nitrate contamination. The output risk maps revealed spatially heterogeneous contamination patterns, with hotspots concentrated in agricultural and peri-urban areas. This research highlights the importance of integrating explainable AI with geospatial analysis to guide targeted groundwater monitoring and management strategies. This approach is transferable to other vulnerable coastal aquifers, supporting sustainable groundwater governance under diffuse pollution conditions.

Urban Flooding is one of the most prevalent natural hazards worldwide, leading to substantial human and economic losses. Therefore, the assessment and mapping of flood hazard levels are essential for reducing the impact of future flood disasters. This study develops and integrates a methodology to evaluate urban flood susceptibility in Nangarhar Province, Afghanistan, a semi-arid region with limited prior research. Landsat imagery from 2004 to 2024 was used to analyze land use land cover change (LULCC), indicating that built-up areas increased from 124 to 180 km² in 2004 to 2024, respectively, while agricultural land decreased from 1978 km² to 1883 km² during the same period. Climate data exhibit increases in temperatures and intensifying rainfall, exacerbating flood hazards. Geospatial analysis of elevation, slope, drainage density, and proximity to water bodies highlights the high susceptibility of low-lying areas. The Analytical Hierarchy Process (AHP) was employed to integrate diverse flood risk factors and produce accurate flood hazard maps. The findings show that very-high flood susceptibility zones expanded from 1537 to 1699 km² in 2004 to 2024, whereas low-susceptibility zones declined from 131 km² to 110 km². Socioeconomic indicators such as population density, built-up density, and education accessibility were also incorporated into the assessment. This study underscores the need for adaptive land use planning, resilient drainage systems, and community-based flood risk reduction strategies. The findings provide actionable insights for sustainable flood management and demonstrate the value of combining GIS, remote sensing, and multi-criteria analysis in data-scarce, conflict-affected regions.

Morocco’s Témara Plain relies heavily on its aquifer system as a critical resource for drinking water, irrigation, and industrial activities. However, this essential groundwater reserve is increasingly threatened by over-extraction, seawater intrusion, and complex hydrogeochemical processes driven by the region’s geological characteristics and anthropogenic pressures. This study aims to assess groundwater quality and its vulnerability to pollution risks and map the spatial distribution of key hydrochemical processes through an integrated approach combining Geographic Information System (GIS) techniques and multivariate statistical analysis, as well as applying the DRASTIC model to evaluate water vulnerability. A total of fifty-eight groundwater samples were collected across the plain and analyzed for major ions to identify dominant hydrochemical facies. Spatial interpolation using Inverse Distance Weighting (IDW) within GIS revealed distinct patterns of sodium chloride (Na-Cl) facies near the coastal areas with chloride concentrations exceeding the World Health Organization (WHO) drinking water guideline of 250 mg/L—indicative of seawater intrusion. In addition to marine intrusion, agricultural pollution constitutes a major diffuse pressure across the aquifer. Shallow groundwater zones in agricultural areas show heightened vulnerability to salinization and nitrate contamination, with nitrate concentrations reaching up to 152.3 mg/L, far surpassing the WHO limit of 45 mg/L. Furthermore, other anthropogenic pollution sources—such as wastewater discharges from septic tanks in peri-urban zones lacking proper sanitation infrastructure and potential leachate infiltration from informal waste disposal sites—intensify stress on the aquifer. Principal Component Analysis (PCA) identified three key factors influencing groundwater quality: natural mineralization due to carbonate rock dissolution, agricultural inputs, and salinization driven by seawater intrusion. Additionally, The DRASTIC model was used within the GIS environment to create a vulnerability map based on seven key parameters. The map revealed that low-lying coastal areas are most vulnerable to contamination.

Introduction: This study investigates the relationship between land use/land cover (LULC) changes and land surface temperature (LST) dynamics in Al-Khobar, Saudi Arabia, a rapidly urbanizing desert city, from 1999 to 2024. The research addresses the critical need to understand urban thermal environment changes driven by anthropogenic modification in arid regions. Methods: The analysis utilized Landsat satellite imagery. LULC changes were quantified using post-classification comparison and spectral indices, including the Normalized Difference Vegetation Index (NDVI), Normalized Difference Built-up Index (NDBI), Normalized Difference Water Index (NDWI), and Barren Soil Index (BSI). LST was retrieved using a radiative transfer equation. Statistical correlation analyses were performed to examine the relationships between these spectral indices and LST for the study years. Results: Significant LULC transformations were observed: urban area increased by 71.5%, while barren land decreased by 50.6%. Vegetation cover grew by 12.4%, and water bodies shrank by 20.1%. Concurrently, the average LST rose from 40.2°C to 42.5°C. Correlation analyses revealed: 1) a strong positive correlation between NDBI and LST, with NDBI increasing from 0.395 to 0.635; 2) a consistent positive correlation between NDVI and LST, challenging the typical cooling paradigm; 3) a strong negative correlation between NDWI and LST, with NDWI declining from 0.518 to 0.294; and 4) a negative correlation between BSI and LST, with BSI decreasing from 0.374 to 0.267. Discussion: The results identify urban expansion (indicated by NDBI) as the primary driver of surface heating. The positive NDVI-LST correlation suggests that the cooling efficacy of irrigated urban greenery in arid settings can be compromised by the surrounding heat-retaining built-up matrix. The strong negative NDWI-LST correlation underscores the disproportionate cooling importance of water bodies, whose decline contributed to warming. The findings provide an evidence-based refinement of LULC-LST dynamics for arid environments, concluding that sustainable urban planning in desert cities must prioritize water body conservation, mandate high-albedo materials, and re-evaluate greening strategies based on proven cooling efficacy rather than mere area coverage.

Coastal zones in arid regions are particularly vulnerable to climate change because of their limited sediment supply and high sensitivity to marine and aeolian forces. This study provides probabilistic projections of coastal evolution for a 130 km segment of the Duba shoreline, Saudi Arabia, a rapidly developing region that includes the NEOM mega-project. An integrated modeling framework was developed by combining a four-decade (1985–2024) diachronic analysis of shoreline evolution from Landsat imagery with a cascade of numerical models. Specifically, climate projections from CMIP6 (under SSP1-2.6, SSP2-4.5, and SSP5-8.5 scenarios) were dynamically downscaled using the regional climate models COSMO-CLM and RegCM, which provided boundary conditions for the SWAN hydrodynamic model to simulate the wave dynamics. The SWAN outputs were then used to force the Delft3D morphodynamic model to project future shoreline evolution. A Bayesian framework was applied to systematically quantify and integrate the uncertainties across all modeling steps, enabling robust probabilistic forecasts. Results indicate an accelerated trend of shoreline retreat, with mean Net Shoreline Movement (NSM) by 2100 ranging from −8.1 m under the low-emission SSP1-2.6 scenario to a critical −25.6 m under the high-emission SSP5-8.5 scenario, with 95% confidence intervals reaching −47.9 m. This erosion is mainly driven by a projected relative sea-level rise of up to 48.3 cm (±15.8 cm) and an increase in significant wave height of up to 40% (mean of 1.95 m). By delivering probabilistic rather than deterministic results, this study provides a solid scientific basis to guide sustainable coastal management, inform the design of risk-sensitive infrastructure, and support the development of climate-resilient adaptation strategies in one of the world’s most rapidly transforming coastal regions.

Rapid urbanization and climate change pose significant challenges to air quality in arid metropolitan areas, with critical implications for public health and sustainable development. This study projects the evolution of air pollution in Riyadh, Saudi Arabia, through 2070 using an integrated modeling approach that combines CMIP6 climate projections with localized air quality data. We analyzed daily concentrations of major pollutants (SO₂, NO₂) across 15 strategically selected monitoring stations representing diverse urban environments, including traffic corridors, residential areas, healthcare facilities, and semi-natural zones. Climate data from two Earth System Models (CNRM-ESM2-1 and MPI-ESM1.2) were bias-corrected and integrated with historical pollution measurements (2000–2015) using hierarchical Bayesian statistical modeling under SSP2-4.5 and SSP5-8.5 emission scenarios. Our results revealed substantial deterioration in air quality, with projected increases of 80–130% for SO₂ and 45–55% for NO₂ concentrations by 2070 under high-emission scenarios. Spatial analysis demonstrated pronounced pollution gradients, with traffic corridors (Eastern Ring Road, Northern Ring Road, Southern Ring Road) and densely urbanized areas (King Fahad Road, Makkah Road) experiencing the most severe increases, exceeding WHO guidelines by factors of 2–3. Even semi-natural areas showed significant increases in pollution due to regional transport effects. The hierarchical Bayesian framework effectively quantified uncertainties while revealing consistent degradation trends across both climate models, with the MPI-ESM1.2 model showing a greater sensitivity to anthropogenic forcing. Future concentrations are projected to reach up to 70 μg m⁻³ for SO₂ and exceed 100 μg m⁻³ for NO₂ in heavily trafficked areas by 2070, representing 2–3 times the Traffic corridors showed concentration increases of 21–24% compared to historical baselines, with some stations (R5, R13, and R14) recording projected levels above 4.0 ppb for SO₂ under the SSP5-8.5 scenario. These findings highlight the urgent need for comprehensive emission reduction strategies, accelerated renewable energy transition, and reformed urban planning approaches in rapidly developing arid cities.