Search Results (836)

To mitigate the frequent occurrence of rear-end collisions on urban expressways under rainy weather conditions, firstly, accident risk levels were classified using traffic conflict indicators. Secondly, three machine learning models were employed to predict the accident severity across different scenarios. Furthermore, key influencing factors of rear-end collisions were identified and analyzed based on SHAP values. Case studies were conducted by simulating vehicle trajectory data under light, moderate, and heavy rain scenarios, using an open urban expressway dataset and car-following parameters for rainy conditions. Next, the Modified Time-to-Collision (MTTC) metric was calculated. Risk thresholds for low-, medium-, and high-risk levels were established for each rainfall category using percentile-based cumulative distribution analysis. Finally, real-time risk prediction under the three rainfall scenarios was conducted using XGBoost, LightGBM, and Random Forest models. The model performances were evaluated in terms of accuracy, recall, precision, and AUC. Overall, the study finds that the LightGBM model achieves the highest predictive capability, with AUC values exceeding 0.78 under all weather conditions. Moreover, the study concludes that factors ranked by SHAP values reveal that the minimum distance has the greatest influence in light rain scenarios. As rainfall intensity increases, the influences of minimum headway time and average vehicle speed are found to grow, highlighting an interaction pattern characterized by “speed-distance-flow” coupling. Full article

Flood hazards in arid oil-producing regions result from both natural hydrological processes and terrain changes due to land subsidence. In the Yibal field in northern Oman, long-term hydrocarbon extraction has caused measurable ground deformation, altering surface gradients and drainage patterns. This study presents a deformation-adjusted flood hazard assessment by integrating a 2013 photogrammetric DEM with a 2023 subsidence-corrected DEM derived from multi-temporal PS-InSAR observations (RADARSAT-2 and TerraSAR-X). Key hydrological indicators—including slope, drainage networks, Height Above Nearest Drainage (HAND), floodplain depth, Curve Number, and extreme precipitation from the wettest monthly rainfall in a 10-year archive—were recalculated for both years. Flood hazard maps for 2013 and 2023 were generated using an AHP-based multi-criteria framework across five hydrologically motivated scenarios. Results indicate that while the total area of high- and very-high-hazard zones changed only slightly in most scenarios (within ±6%), these zones shifted into subsidence-affected depressions, reflecting deformation-driven redistribution of flood-prone areas. Low-hazard zones grew most significantly, especially in Scenarios S2–S4, with increases of 160–320% compared to 2013, while moderate-hazard areas showed smaller but consistent growth. Floodplain-dominated conditions (S5) produced the most pronounced nonlinear response, with a substantial increase in very low hazard and localized concentration of very high hazard in areas of deepest subsidence. Geomorphic analysis using the Geomorphic Flood Index (GFI) shows deepening of flow pathways and expansion of geomorphic depressions between 2013 and 2023, supporting the modeled redistribution of hazards. These findings demonstrate that even moderate subsidence can significantly alter hydrological susceptibility and underscore the importance of incorporating deformation-adjusted terrain modeling into flood hazard assessments in petroleum fields and other subsidence-prone areas. Full article

This study investigated the site-specific ionic composition and wet deposition loads of rainwater collected from eight actively cultivated agricultural regions across South Korea, with the aim of quantifying spatial and seasonal variability and interpreting how regional agricultural characteristics and surrounding site conditions influence major ion concentrations and deposition patterns. Rainfall samples were obtained using automated samplers and analyzed via high-performance ion chromatography for major cations (Na⁺, $\mathrm{N}{\mathrm{H}}_4^{+}$, K⁺, Ca²⁺, Mg²⁺) and anions (Cl⁻, $\mathrm{N}{\mathrm{O}}_3^{-}$, $\mathrm{S}{\mathrm{O}}_4^{2-}$, $\mathrm{N}{\mathrm{O}}_2^{-}$). The results revealed significant seasonal fluctuations in ion loads, with $\mathrm{N}{\mathrm{H}}_4^{+}$ (peak 1.13 kg/ha) and K⁺ (peak 0.25 kg/ha) reaching their highest levels during summer due to increased fertilizer use and crop activity. Conversely, Cl⁻ peaked in winter (2.11 kg/ha in December), particularly in coastal regions, likely influenced by de-icing salts and sea-salt aerosols. Correlation analysis showed a strong positive association among $\mathrm{N}{\mathrm{H}}_4^{+}$, $\mathrm{N}{\mathrm{O}}_3^{-}$, and $\mathrm{S}{\mathrm{O}}_4^{2-}$ (r = 0.89 and r = 0.84, respectively), indicating shared atmospheric transformation pathways from agricultural emissions. Ternary diagram analysis further revealed regional distinctions: coastal regions such as Gimhae and Muan exhibited Na⁺ and Cl⁻ dominance, while inland areas like Danyang and Hongcheon showed higher proportions of Ca²⁺ and Mg²⁺, reflecting differences in aerosol sources, land use, and local meteorological conditions. These findings underscore the complex interactions between agricultural practices, atmospheric processes, and local geography in shaping rainwater chemistry. The study provides quantitative baseline data for evaluating non-point source pollution and developing region-specific nutrient and soil management strategies in agricultural ecosystems. Full article

The M_JMA 7.6 (Mw 7.5) Noto Peninsula Earthquake of 1 January 2024 in Japan triggered widespread slope failures across northern Noto region, but their spatial controls and susceptibility patterns remain poorly quantified. Most previous studies have focused mainly on fault rupture, ground deformation, and tsunami impacts, leaving a clear gap in machine learning based assessment of earthquake-induced slope failures. This study integrates 2323 mapped landslides with eleven conditioning factors to develop the first data-driven susceptibility framework for the 2024 event. Spatial analysis shows that 75% of the landslides are smaller than 3220 m² and nearly half occurred within about 23 km of the epicenter, reflecting concentrated ground shaking beyond the rupture zone. Terrain variables such as slope (mean 31.8°), southwest-facing aspects, and elevations of 100–300 m influenced the failure patterns, along with peak ground acceleration values of 0.8–1.1 g and proximity to roads and rivers. Six supervised machine learning models were trained, with Random Forest and Gradient Boosting achieving the highest accuracies (AUC = 0.95 and 0.94, respectively). Explainable AI using SHapley Additive exPlanations (SHAP) identified slope, epicentral distance, and peak ground acceleration as the dominant predictors. The resulting susceptibility maps align well with observed failures and provide an interpretable foundation for post-earthquake hazard assessment and regional risk reduction. Further work should integrate post-seismic rainfall, multi-temporal inventories, and InSAR deformation to support dynamic hazard assessment and improved early warning. Full article

Global warming is modifying precipitation patterns, and hence increasing the hazards of severe and extended rainstorms. Addressing the gap in integrating economic and environmental assessments into urban stormwater management—a key challenge in urban water resource analysis—this study utilizes the analytical hierarchy process (AHP) and SUSTAIN model to identify and evaluate low-impact development (LID) stormwater management strategies, assessing their impacts on runoff volume, peak flow reduction, chemical oxygen demand (COD), and suspended solids (SS) across four planning scenarios under five rainfall recurrence intervals, culminating in a cost–benefit analysis to ascertain the optimal scenario. The reduction rates for COD and SS varied from 41.85% to 87.11% across different scenarios, with Scenario Three (RM03) demonstrating the highest efficacy in pollutant management. (The four labels RM01–RM04 are used throughout the text to represent the four scenarios) Implementing the best plan may result in a reduction of yearly carbon emissions of 189.70 metric tons, with emissions from the operational load of the drainage network and COD pollution treatment potentially decreasing by 2.44% and 2.06%, respectively, indicating an overall annual reduction of 85.46%. This approach not only mitigates urban rainwater and flooding issues but also prevents resource wastage, optimizes resource utilization and benefits, offers a scientific foundation for urban construction and planning, and serves as a reference for sponge city development in other regions. Full article

The drainage of groundwater in mountainous tunnel projects always leads to substantial decline of the regional water table, which may induce numerous environmental issues, such as spring depletion, surface subsidence, vegetation degradation, and impacts on local water supplies, especially in the enclosed karst aquifers of anticlines in the area, such as the Jura mountain type. A systematic hydrological monitoring was conducted during the excavation of the Wufu Tunnel in Chongqing, China. The monitoring data includes discharge rate and water level collected from tunnels, boreholes, coal mines, springs, and ponds, respectively. Hydrological responses of karst aquifers and surface water bodies to tunnel drainage and precipitation were investigated by statistical analysis, Mann–Kendall test, heat map, and wavelet analysis. Results show that the enclosed karst water system has strong hydraulic connections and good water storage conditions. Tunnel drainage is the dominant factor causing dynamic changes at monitoring points, while the influence of rainfall is relatively limited. Borehole water levels and coal mine drainage have a close correlation with tunnel inflow, while springs are influenced by both rainfall and tunnel drainage. Few pond monitoring points are related to rainfall. Tunnel drainage has transformed the regional groundwater dynamic conditions, causing local groundwater flow direction reversal and reconstructing the groundwater recharge-flow-discharge pattern. Full article

Intensity–Duration–Frequency (IDF) curves are essential tools in the design of stormwater management systems and are often used over long periods without frequent updates. However, the continuous collection of rainfall data and the expansion of monitoring networks call for regular revisions of these curves. In Romania, current engineering and hydrological practices still rely on regionalized IDF graphs developed in 1973. Given the ongoing effects of climate change—particularly the increased frequency and, more significantly, intensity of extreme rainfall events—updating these curves has become critical. Incorporating recent observations is essential not only for methodological accuracy but also to support climate-resilient infrastructure design. This study employs updated IDF curves provided by the National Administration of Meteorology, based on 30 years of precipitation records from 68 meteorological stations across Romania. The main objective is to evaluate alternative regionalization approaches—including clustering methods, geographic proximity analysis, and hourly precipitation isolines for a 1:10 Annual Exceedance Frequency—to develop a new regionalization model and the corresponding nationwide IDF relationships. A comparative analysis using raster-based regional rainfall datasets from both the 1973 and 2025 regionalizations revealed significant changes in precipitation patterns. Short-duration rainfall events (5, 10, and 30 min) have increased in intensity across most regions, while long-duration events (3, 6, 12, and 24 h) have generally decreased in magnitude in several areas. These findings highlight a growing trend toward more intense short-term convective storms, underlining the urgent need for improved flash flood prevention and urban stormwater management strategies. Full article

Pakistan, one of the world’s most water-stressed countries, is extremely vulnerable to climate change. This study analyzes the prospective effects of climate change on the rainfall in the Hub River Watershed (HRW), Sulaiman-Kirthar mountainous range of Pakistan. The projections of five global climate models (GCMs), from the Coupled Model Intercomparison Project phase 6 (CMIP6), were used. Analysis of future changes in rainfall patterns was performed under two shared socioeconomic pathways (SSPs). Results showed that the historical annual average rainfall was increasing in the HRW. The annual average rainfall is expected to decrease under SSP2–4.5 in HRW. However, under SSP5-8.5, an increasing trend over the next three decades is expected, particularly over the southern part of the HRB. The findings should further our knowledge of how climate change affects the Hub River Basin and motivate stakeholders and planners to develop the best mitigation plans. Full article

Organic micropollutants in agricultural soils pose significant ecological and health risks. This study conducted the first large-scale, integrated non-targeted screening and targeted analysis across China’s major food-producing regions. Using high-resolution mass spectrometry, 498 micropollutants were identified, including pesticides, industrial chemicals, pharmaceuticals, personal care products, food additives, natural products, and emerging contaminants. Spatial analysis revealed strong correlations in pesticide detections between Henan and Hebei, as well as between Hebei and Shandong, indicating pronounced regional similarities in pesticide occurrence patterns. Concentrations of 50 quantified micropollutants showed clear spatial variability, which was associated with precipitation, water use, and agricultural output, reflecting climate–agriculture–socioeconomic synergies. Greenhouse soils accumulated higher micropollutant levels than open fields, driven by intensive agrochemical inputs, plastic-film confinement, and reduced phototransformation. Co-occurrence patterns indicated similar pathways for personal care products, industrial chemicals, and pesticides, whereas natural products and pharmaceuticals showed lower levels of co-occurrence due to crop-specific exudates, fertilization, and rainfall-driven leaching. Among cropping systems, orchard soils had the highest micropollutant accumulation, followed by paddy and vegetable soils, consistent with frequent pesticide use and minimal tillage. Risk quotients indicated moderate-to-high ecological risks at over half of the sites. These results reveal complex soil pollution patterns and highlight the need for dynamic inventories and spatially differentiated, crop- and system-specific mitigation strategies. Full article

Landslide disasters are complex spatiotemporal phenomena. Existing deep learning (DL) models for remote sensing (RS) image analysis primarily exploit shallow visual features, inadequately incorporating critical geological, geographical, and environmental knowledge. This limitation impairs detection accuracy and generalization, especially in complex terrains and diverse vegetation conditions. We propose Knowledge Graph Embedding in Swin Feature Pyramid Networks (KGE–SwinFpn), a novel RS landslide segmentation framework that integrates explicit domain knowledge with deep features. First, a comprehensive landslide knowledge graph is constructed, organizing multi-source factors (e.g., lithology, topography, hydrology, rainfall, land cover, etc.) into entities and relations that characterize controlling, inducing, and indicative patterns. A dedicated KGE Block learns embeddings for these entities and discretized factor levels from the landslide knowledge graph, enabling their fusion with multi-scale RS features in SwinFpn. This approach preserves the efficiency of automatic feature learning while embedding prior knowledge guidance, enhancing data–knowledge–model coupling. Experiments demonstrate significant outperformance over classic segmentation networks: on the Yuan-yang dataset, KGE–SwinFpn achieved 96.85% pixel accuracy (PA), 88.46% mean pixel accuracy (MPA), and 82.01% mean intersection over union (MIoU); on the Bijie dataset, it attained 96.28% PA, 90.72% MPA, and 84.47% MIoU. Ablation studies confirm the complementary roles of different knowledge features and the KGE Block’s contribution to robustness in complex terrains. Notably, the KGE Block is architecture-agnostic, suggesting broad applicability for knowledge-guided RS landslide detection and promising enhanced technical support for disaster monitoring and risk assessment. Full article

We integrated retrieval-augmented generation (RAG) with a multimodal large language model (LLM) to cluster large-scale circulation patterns associated with extreme rainfall events (>80 mm day⁻¹) in Taiwan. Employing an event-perspective approach on 62 years (1960–2022) of ERA5 reanalysis data, we identified discrete rainfall events by season. The LLM-based classification effectively captured intra-seasonal variability and mixed-regime events, outperforming conventional seasonal categorizations. The results of this analysis indicate that 56.7% of extreme rainfall events occurred between July and October. During this period, the LLM–RAG framework performed event-based clustering and identified four representative synoptic patterns: typhoon systems, subtropical high perturbations, southwesterly monsoonal flow, and southeasterly flow regimes. Full article

Landscapes in semiarid regions are highly sensitive to climate change and anthropogenic activities, and their evolution directly influences ecosystem services and regional ecological security. Although previous research has examined land use changes, systematic quantitative analyses of long-term evolutionary trends and driving mechanisms, particularly the comprehensive relationships between key hydrological elements and landscape pattern evolution in water-scarce, semiarid watersheds, remain limited. To address the research gap in long-term, multifactor, and hydro–landscape integrated analysis, China’s Tuwei River watershed was selected as the study area in this study, and methods such as landscape pattern indices and gray relational analysis were employed to quantitatively reveal the spatiotemporal evolution of watershed landscape fragmentation from 1980 to 2020 and identify its dominant driving forces. The results revealed that (1) over the 40-year period, the land use structure of the watershed underwent significant restructuring, with developed land expanding by 1282%, cropland and bare land areas decreasing by 14.2% and 32.01%, respectively, and grassland and forestland areas increasing by 24.5% and 14.9%, respectively; (2) land-scape fragmentation continued to intensify, with the landscape fragmentation composite index (FCI) increasing by 37.6%, patch density (PD) continuously increasing, edge density (ED) and landscape shape index (LSI) increasing significantly, and landscape connectivity weakening; (3) natural and socioeconomic factors jointly drove landscape evolution, with temperature and mean annual flow contributing the most among natural factors and the urbanization rate and secondary industry output value serving as the core drivers among socioeconomic factors; and (4) the trend of landscape fragmentation was synchronized with changes in annual rainfall and runoff and exhibited a significant negative correlation with the groundwater level. In summary, through long-term, multifactor comprehensive analysis, the evolution characteristics and driving mechanisms of landscape patterns in the Tuwei River watershed were systematically revealed in this study. These findings not only deepen the understanding of landscape fragmentation processes under the dual pressures of climate change and anthropogenic activities but also provide scientific evidence for the sustainable management of landscapes and associated ecosystems in semiarid watersheds. Full article

The upper Xin’an River basin serves as a critical ecological barrier and water-conservation area for the Yangtze River Delta. However, with rapid economic development, nitrogen pollution in the surface waters of this region has become increasingly pronounced. This study analyzed river water samples collected on four occasions from the upper Xin’an River basin for ammonium (NH₄⁺–N), nitrate-nitrogen (NO₃⁻–N), total nitrogen (TN), and nitrate isotopic (δ¹⁵N–NO₃⁻ and δ¹⁸O–NO₃⁻). The sources of nitrate (NO₃⁻) were apportioned using the MixSIAR stable-isotope mixing model, and the spatial distribution of these sources across the basin was characterized. Across the four sampling rounds, the mean TN concentration exceeded 1.3 mg/L, with NO₃⁻–N accounting for over 45% of TN, indicating that nitrate was the dominant inorganic nitrogen species. The δ¹⁵N–NO₃⁻ values ranged from 2.17‰ to 13.0‰, with mean values following the order summer > winter > autumn > spring. The δ¹⁸O–NO₃⁻ values varied from −5.20‰ to −3.48‰, and the average value showed a completely opposite seasonal variation pattern to that of δ¹⁵N–NO₃⁻. Process-based analysis of nitrogen transformations revealed that nitrification predominates during nitrate transport and transformation, whereas denitrification is comparatively weak. MixSIAR-based estimates indicate marked seasonal differences in the source composition of nitrate pollution in the upper Xin’an River basin; NO₃⁻ derives primarily from soil nitrogen (SN) and livestock/sewage manure nitrogen (LSN). LSN was the dominant contributor in spring and summer (49.2% and 59.9%, respectively). SN dominated in autumn (49.2%) and winter (54.1%). Fertilizer nitrogen (FN) contributed more during summer and autumn, when fertilization is concentrated and rainfall is higher. Atmospheric deposition (AN) contributed approximately 1% across all seasons and was thus considered negligible. These findings provide a scientific basis for source control of nitrogen pollution and water-quality management in the upper Xin’an River. Full article

This study presents an integrated stochastic framework for assessing and forecasting drought dynamics in the western Bhagirathi–Hooghly River Basin, encompassing the districts of Bankura, Birbhum, Burdwan, Medinipur, and Purulia. Employing multiple probabilistic and statistical techniques, including the gamma-based standardized precipitation index (SPI), effective drought index (EDI), rainfall anomaly index (RAI), and the auto-regressive integrated moving average (ARIMA) model, the research quantifies spatio-temporal variability and projects drought risk under non-stationary climatic conditions. The analysis of century-long rainfall records (1905–2023), coupled with LANDSAT-derived vegetation and moisture indices, reveals escalating drought frequency and severity, particularly in Purulia, where recurrent droughts occur at roughly four-year intervals. Stochastic evaluation of rainfall anomalies and SPI distributions indicates significant inter-annual variability and complex temporal dependencies across all districts. ARIMA-based forecasts (2025–2045) suggest persistent negative SPI trends, with Bankura and Purulia exhibiting heightened drought probability and reduced predictability at longer timescales. The integration of remote sensing and time-series modelling enhances the robustness of drought prediction by combining climatic stochasticity with land-surface responses. The findings demonstrate that a hybrid stochastic modelling approach effectively captures uncertainty in drought evolution and supports climate-resilient water resource management. This research contributes a novel, region-specific stochastic framework that advances risk-based drought assessment, aligning with the broader goal of developing adaptive and probabilistic environmental management strategies under changing climatic regimes. Full article

Citrus orchards are especially vulnerable owing to low inter-row vegetation cover, and frequent tillage. Here, we combine controlled field experiments with proximal remote sensing–derived geomorphometric variables and machine learning (ML) to identify key factors of erosion in a Mediterranean climate citrus plantation located close to Seville and the National Park of Doñana (Southern Spain) on Gleyic Regosols (clayic, arenic). We conducted rainfall simulations with 30 s sampling, measured infiltration (mini-disc infiltrometer), saturated hydraulic conductivity (Kfs; Guelph permeameter), compaction (penetrologger), and soil respiration (gas analyzer) at multiple points, and derived high resolution morphometric indices from proximal sensing (UAV-LiDAR). Linear models and Random Forests were trained to explain three responses: soil loss, sediment concentration (SC), and runoff. Results show that soil loss is most strongly associated with maximum compaction and Kfs (multiple regression: R² = 0.68; adjusted R² = 0.52; p = 0.063), while SC increases with surface compaction and exhibits weak relationships with topographic metrics. Runoff decreases with average infiltration, which is related to compaction (β = −4.83 ± 2.38; R² = 0.34; p = 0.077). Diagnostic checks indicate centered residuals with mild heteroscedasticity and a few high leverage observations. Random Forests captured part of the variance for soil loss (≈29%) but performed poorly for runoff, consistent with limited sample size and modest nonlinear signal. Morphometric analysis revealed gentle relief but pronounced convergent–divergent patterns that modulate hydrological connectivity. There were strong differences in the experiments conducted close to the trees and in the tractor trails. We conclude that compaction and near surface hydraulic properties are the most influential and measurable controls of erosion at plot scale and the UAV-LiDAR could not give us extra-insights. We highlight that integrating standardized field protocols with proximal morphometrics and ML can be the best method to prioritize a small set of explanatory variables, helping to reduce experimental effort while maintaining explanatory power. Full article