Search Results (819)

Recreational coastal waters are increasingly affected by urbanization, maritime activities, and tourism, creating a need for predictive tools that support proactive water quality management. This study proposes an explainable machine learning framework for forecasting near-future changes in the Tourism-Oriented Seawater Quality Index (SeaWQI-T) using high-frequency seawater monitoring data collected in the Gulf of Vlorë, Albania. A summer monitoring campaign (June–August 2025) produced 102,988 physicochemical observations from six monitoring stations using an unmanned surface vehicle equipped with a Horiba U53 multiparameter sonde. Following quality control and temporal aggregation, the data were used to formulate a multiclass forecasting problem (decrease, stable, or increase), and Logistic Regression, Random Forest, and Extreme Gradient Boosting (XGBoost) models were evaluated across multiple forecasting horizons. XGBoost achieved the best validation performance, while Random Forest demonstrated superior generalization on the independent test dataset and provided the most stable explainability results. SHapley Additive exPlanations (SHAP) identified SeaWQI-T dynamics, turbidity, dissolved oxygen, and oxidation–reduction potential as the most influential predictors. The proposed framework demonstrates that integrating explainable machine learning with autonomous high-frequency monitoring can provide accurate, interpretable forecasts to support intelligent coastal recreation management and sustainable tourism planning. Full article

Large yellow croaker (Larimichthys crocea) is an economically important marine fish, yet rapid and non-destructive discrimination between sea-caught and farmed collection-source groups remains challenging. This study developed a fish-level analytical workflow integrating YOLO11n-seg instance segmentation with visible–near-infrared hyperspectral imaging to extract relative reflectance ratios from six anatomical regions. The analytical cohort comprised 258 unique fish, including 171 sea-caught and 87 farmed individuals. Source categories were assigned according to the original capture or cage-culture collection channels. Fish identity was used as the grouping unit in five-fold internal cross-validation, and classification performance was evaluated based on out-of-fold fish-level predictions. Logistic-regression models achieved AUC values ranging from 0.985 to 1.000 across the six individual anatomical regions. The combined six-region model achieved an accuracy of 0.996 (95% CI, 0.988–1.000) and an AUC of 1.000 (95% CI, 1.000–1.000) within the present cohort. For anatomical-region segmentation, the 20-image validation set contained 140 annotated instances. Bounding-box precision, recall, mAP@0.5, and mAP@0.5–0.95 were 0.976, 0.983, 0.981, and 0.700, respectively, while the corresponding mask metrics were 0.969, 0.976, 0.972, and 0.669. These results indicate that region-specific hyperspectral information can support highly accurate internal discrimination between the two collection-source groups and provide an interpretable basis for characterizing source-associated spectral differences. However, the source labels were not independently verified, potentially influential covariates were incompletely recorded, and no independent external cohort was available. Therefore, the present findings should be interpreted as internally validated collection-source discrimination rather than verified provenance authentication or evidence of external generalizability. Full article

Reliable traffic flow forecasting is a core component of intelligent transportation systems; however, many current approaches are still unable to simultaneously model spatial interdependencies and long-term temporal correlations, particularly in cross-sea corridors that exhibit directional heterogeneity and pronounced temporal variability. This study aims to develop an accurate and stable traffic flow prediction framework for cross-sea corridors. To achieve this, an HGTransformer model was proposed that constructed a hypergraph from traffic nodes based on spatial proximity and correlated flow variations, and used hypergraph convolution to extract spatial node representations. These representations were then fed into a Transformer equipped with multi-head self-attention and positional encoding, enabling the model to capture global temporal dependencies in the evolution of traffic flow. Using hourly traffic flow data from the Huangmaohai cross-sea corridor, the model was tested on 1 to 4 h forecasting horizons and compared with long short-term memory (LSTM), multi-layer perceptron (MLP), random forest (RF), support vector regression (SVR), and Bayesian regression (BR) models. The proposed model achieved the best overall performance, with average mean absolute percentage error (MAPE), mean absolute error (MAE), weighted mean absolute percentage error (WMAPE), and root mean square error (RMSE) of 0.178, 13.375, 0.140, and 20.538, respectively. At the 1 h horizon, these values further decreased to 0.172, 12.678, 0.132, and 19.401, while preserving peak–valley structures more accurately under both short- and longer-horizon forecasting. The main contribution of this study lies in the systematic application and validation of the Huangmaohai Corridor dataset, including a reproducible hypergraph construction strategy tailored specifically for this particular infrastructure. Full article

Reliable modelling of monthly air temperature is relevant to station-scale climate assessment and the evaluation of meteorological data-driven models. However, station-scale monthly meteorological datasets often contain correlated and partially redundant predictors because thermal, moisture, precipitation, wind, and seasonal variables are jointly controlled by atmospheric and seasonal forcing. This study conducts an integrated comparative analysis of established regression and machine learning models for monthly mean air temperature modelling in Zonguldak, a humid coastal province in the Western Black Sea Region of Türkiye. Monthly meteorological observations from 2000 to 2022 were used to evaluate eight primary regression and machine-learning models: Partial Least Squares regression, Ridge, Lasso, ElasticNet, Support Vector Regression, Random Forest, Gradient Boosting, and Extreme Gradient Boosting. Ordinary Least Squares (OLS) and Huber regression were additionally included as reference models. The analysis retained the original meteorological predictors and jointly evaluated predictive accuracy, model stability, ablation sensitivity, and model-specific predictor relevance. Reduced-predictor and seasonality-only scenarios were examined to distinguish direct thermal reconstruction from broader climatological predictability. Model performance was assessed using repeated nested cross-validation, bootstrap summaries of performance variability, supplementary rolling-origin validation, and Wilcoxon signed-rank tests with Holm correction. Although the full-predictor models achieved high predictive accuracy, this performance largely reflected the direct thermal information contained in minimum and maximum air temperature. When these thermal predictors were excluded, MAE increased to approximately 1.13–1.22 °C and $R^2$ decreased to approximately 0.93–0.94. The seasonality-only scenario yielded MAE values of approximately 1.27–1.34 °C and R² values of approximately 0.92, indicating that the annual cycle accounted for a substantial proportion of monthly temperature predictability. The additional non-thermal meteorological predictors provided only limited improvement beyond the strong seasonal baseline. Overall, model performance depended on the predictor information available, and no single model family showed a consistent advantage across the evaluated scenarios. These findings highlight the importance of considering predictive accuracy together with model stability and predictor dependence in data-limited station-scale temperature modelling. Full article

Marine cold-air outbreaks (MCAOs) enhance air–sea exchange and are accompanied by substantial cloud-population reorganization over the North Atlantic. Previous satellite studies have documented MCAO-related cloud properties and International Satellite Cloud Climatology Project (ISCCP) weather-state distributions, but the continuous redistribution of the complete cloud population and its dependence on sampling scale remain less well quantified. We combine the merged H-series ISCCP (ISCCP-H) weather-state product with the ERA5 reanalysis over the North Atlantic sector of 35–50°N, 55–20°W during January–March and November–December of 2000–2009. Three complementary analytical strategies are used to examine scale-dependent cloud associations: intensity-binned analysis of positive-MCAO native ERA5 grid-point-days, domain-daily regressions, and ISCCP-H cell fixed-effect models. Across 1459 MCAO dates and 256,518 strict-ocean ISCCP-H cell-days, the combined frequency of shallow-cumulus-like and stratocumulus-like low-cloud weather states decreases systematically with MCAO intensity, with a domain-daily slope of −0.0562 K⁻¹. Compensating regional increases occur mainly in the midlatitude-storm and middle-to-high-cloud weather states, while the deep convective and anvil state shows no robust increase. After cell and calendar-month climatological differences are removed, the low-cloud reduction persists, together with positive associations for the midlatitude-storm and optically thick middle-top weather states. Cirrus is positively associated with MCAO intensity at the regional scale but negatively associated within fixed cells, consistent with geographical composition contributing to its regional response. MCAO intensity is therefore more consistently associated with total low-cloud weather-state occupancy than with the internal composition of the low-cloud subset. The closed weather-state framework provides an observational benchmark for satellite and climate model evaluation. Full article

Charybdis (Gonioneptunus) bimaculata (Miers, 1886), C. (Charybdis) japonica (A. Milne-Edwards, 1861), and C. (Charybdis) miles (De Haan, 1835) are dominant target species harvested by small-scale artisanal fisheries in the coastal waters of China. Characterizing their spatial distribution and interannual population variability is critical for sustainable crustacean stock conservation. Previous investigations of these three Charybdis congeners have been limited to short-duration, single-season, or geographically restricted descriptive surveys, with no systematic cross-species comparison of seasonal habitat use. This has created major knowledge gaps regarding their long-term spatiotemporal dynamics and ecological niche partitioning. We conducted standardized bottom-trawl surveys aboard research vessels across the southern Yellow Sea and East China Sea from 2018 to 2021 to quantify numerical abundance and biomass of the three congeneric crabs. Two species distribution modeling frameworks—Generalized Additive Models (GAMs) and Boosted Regression Trees (BRTs)—were implemented to generate fine-resolution seasonal habitat suitability maps for each species. We identified minimum thermal thresholds for species presence: bottom water temperature > 8 °C for C. bimaculata and >10 °C for C. miles. Core aggregation zones of C. bimaculata and C. japonica occurred in waters adjacent to the Yangtze River Estuary (32–34° N), whereas C. miles predominated in offshore habitats of the southern East China Sea (27–31° N). In this study, C. bimaculata favors shallow seas with warm bottom water and euryhaline conditions, C. japonica prefers low-salinity coastal waters with cold bottom water, and C. miles thrives in deep, high-salinity offshore waters with eurythermal tolerance. The findings of this study can contribute to the sustainable fisheries management and conservation of these species. Full article

Accurate Estimated Time of Arrival (ETA) forecasting is essential for improving operational planning and decision-making in modern ports. While machine learning has significantly enhanced ETA prediction using Automatic Identification System data, the impact of data preprocessing on forecasting performance remains underexplored. This study investigates how AIS data preprocessing and machine learning model selection jointly affect ETA forecasting for short-sea shipping, using the Port of Sines as an empirical case study. A reproducible forecasting workflow was developed, integrating dataset construction, voyage selection, feature engineering, predictive modelling and performance evaluation. Three supervised machine learning algorithms, K-Nearest Neighbors, Random Forest Regression, and Multilayer Perceptron, were trained and compared under identical experimental conditions. The results show that forecasting performance depends not only on model selection but also on the quality of the modelling dataset. In particular, Random Forest Regression achieved the strongest and most consistent performance. It was found to be invariant to feature scaling, whereas scaling had a small negative effect on K-Nearest Neighbors and increased training variance for the Multilayer Perceptron. Although all three models achieved accurate ETA predictions, they exhibited different strengths regarding predictive performance, computational efficiency, and operational applicability. The proposed workflow contributes to the development of transparent and reproducible ETA forecasting methodologies and provides practical guidance for implementing AIS-based decision-support systems in short-sea port operations. Full article

The concentrations of DIN and DIP are important indicators in an offshore ecosystem; although they do not have optical activity, their concentrations are affected by optically active substances, such as sediment, chlorophyll, and dissolved organic matter, an association that is especially close in coastal waters. This study aimed to identify this relationship to provide a theoretical basis for using remote sensing to monitor DIN/DIP concentrations. This study first used correlation analysis to analyze the relationship between water quality indicators and the field-measured spectrum in the Sheyang estuary. The results show a strong positive correlation between the DIN and DIP concentrations and spectrum in near-infrared range, similar to that between the suspended sediment concentrations and spectrum; this indicates a close relationship between DIN/DIP concentrations and sediment concentration in this sea area. Traditional regression models for DIN and DIP concentrations were constructed using the sensitive bank factors of a Hyperion image. By comparing the physical meaning of the factors and the precision and stability of the models, the quadratic model established by the ratio factor of 45th and 10th bands was selected as the DIN concentration inversion model, the quadratic model established by the ratio factor of the 45th and 9th bands was selected as the DIP inversion model, and the inversion results of the image conformed to the actual distribution pattern of DIN and DIP concentrations. In order to fully utilize the spectral information of the Hyperion data, the model coupled using partial least squares (PLS) and support vector machine (SVM) was used to construct regression models of DIN and DIP concentrations. By comparing the standardized coefficients of PLS regression, the 8~16th bands and 37~57th bands of the Hyperion image were selected; all these bands were extracted as two orthogonal components to construct the SVM regression model. Finally, the parameter combinations of radial basis model with C = 10, γ = 0.05, and ε = 0.1 and C = 1, γ = 0.1, and ε = 0.001 were determined as the inversion models for DIN and DIP concentrations, respectively. The prediction accuracy of the models was significantly improved compared to the traditional regression models, and the inversion results were superior to those of the traditional regression models, demonstrating the potential of this algorithm in hyperspectral image modeling. Full article

Assessing long-term hydroclimatic variability in central Mexico is essential to understand regional water availability and groundwater recharge for urban centers such as Querétaro. This study developed a multi-century winter–spring precipitation reconstruction for the Sierra Gorda Biosphere Reserve (SGBR) using ring width chronologies of Douglas-fir, Pseudotsuga menziesii (Mirb.) Franco. Standard dendrochronological techniques were applied to develop a 284-year master chronology (1731–2015). Following the accepted Subsample Signal Strength criterion (SSS ≥ 0.85) for chronology reliability, the reconstruction was restricted to the 1744–2015 period, yielding a statistically robust 271-year December–April precipitation record. A bootstrapped ordinary least-squares regression model relating tree-ring indices to instrumental December–April precipitation was calibrated and validated using split-sample cross-validation, explaining 46% of the instrumental precipitation variance (R² = 0.46) and yielding positive verification statistics (RE = 0.38–0.58; CE = 0.37–0.57). Spatial field correlations against gridded climate data (CRU TS4.08) confirmed a broad regional hydroclimatic signal centered over the Sierra Madre Oriental. Continuous wavelet transform (CWT), spectral analysis, superposed epoch analysis (SEA), and wavelet coherence (WTC) revealed significant interannual (2–8 years) and decadal (10–20 years) variability associated with large-scale ocean–atmosphere climate modes, including the El Niño–Southern Oscillation (ENSO), North Atlantic Oscillation (NAO), Atlantic Multidecadal Oscillation (AMO), and Tropical North Atlantic (TNA) index. The pronounced sensitivity of these conifer forests to pre-monsoonal moisture deficits highlights their vulnerability to projected warming and increasing spring evapotranspiration stress. Although the reconstruction is limited to pre-monsoonal (December–April) precipitation, it provides a robust centuries-long baseline for contextualizing regional hydroclimatic variability and supports water-resource management, groundwater conservation, and climate-adaptation strategies in central Mexico. Full article

Medicanes are mesoscale cyclones that develop over the Mediterranean Sea and display tropical-like cyclone characteristics, including a warm core, spiral cloud organization, and deep convection over warm sea surfaces. Since their structure and position can change rapidly on short lead times before coastal impact, robust near-real-time tracking algorithms are essential for timely warning and operational decision support. To advance this research direction, this work introduces the Deep Learning Medicane Tracking (DeMeTrA) Algorithm, an end-to-end deep learning framework for medicane detection and rotation-center localization from SEVIRI Rapid Scan Airmass RGB imagery. The proposed methodology consists of a three-stage VideoMAE v2 architecture encompassing the following: (i) self-supervised domain specialization on unlabeled satellite image sequences, (ii) supervised binary classification of cyclone versus non-cyclone events, and (iii) supervised coordinate regression for rotation-center tracking. The training corpus spans several time windows of Meteosat Second-Generation observations across the Mediterranean basin, with ground-truth annotations derived from a consensus cyclone-track reference. On event-based splits, cyclone detection reaches 91% balanced accuracy on a balanced validation set and 89% on an unbalanced test set representative of operational conditions. The tracking results show generally low localization errors (mostly below 20 km), with limited outliers in the most complex cases. These findings support the use of Transformer-based video models for operational medicane monitoring and establish a baseline for future developments. Full article

Historical earthquake records are inherently incomplete: many sites that were likely damaged were never documented, leaving spatial gaps in the macroseismic record. This study evaluates whether intensity values at unreported sites can be estimated from the spatial relationships of surrounding reports, framing the task as a spatial data imputation problem. Three spatial imputation methods, Linear regression, K-Nearest Neighbors (KNN), and Kriging, were applied to eight macroseismic datasets, comprising two historical Dead Sea Transform earthquakes (1927 Dead Sea, 1837 South Lebanon) and six instrumental events from major strike-slip fault systems. Model performance was assessed with 5-fold cross-validation under random and spatial-block designs, using Mean Squared Error (MSE) and success rate, defined as the percentage of predictions falling within ±0.5 and ±1.0 intensity units of observed values. Under random cross-validation, simple and locally focused models performed on par with the complex geostatistical approaches. For the geographically concentrated historical data, success rates reached up to 90% within ±1.0 intensity units. San Andreas events yielded the strongest results among instrumental datasets, while Caribbean events showed the weakest performance due to spatial reporting biases. Under spatial-block cross-validation, performance declined across all models, with linear regression and Universal Kriging proving most robust to spatial extrapolation. These findings provide a methodological basis for estimating intensity at undocumented sites. While continuous intensity mapping from sparse data remains inadvisable, point-based imputation offers a practical tool for enriching historical earthquake records, with direct implications for seismic research along poorly documented fault systems. Full article

The embedded finance paradigm is fundamentally restructuring digital economies by seamlessly integrating financial services into non-financial digital infrastructures. This study dynamically evaluates the productivity frontiers of Sea Limited’s embedded fintech operations (SeaMoney) across six core geographic markets (Indonesia, Thailand, Vietnam, the Philippines, Malaysia, and Brazil) over the 2023–2026 temporal horizon. Employing a rigorous Panel Data Envelopment Analysis (DEA) Malmquist Productivity Index framework, the research measures systemic performance by analyzing Sales & Marketing (S & M) Expenses and the undesirable Non-Performing Loan (NPL) rate as inputs, against Gross Loan Outstanding as the primary credit output. Before model execution, robust isotonicity was empirically validated using Pearson correlation matrices. The empirical findings reveal profoundly robust systemic performance across the global ecosystem, driven primarily by overarching algorithmic innovations captured by the Technical Change (TC) index. However, this technological boundary exhibits a stabilizing deceleration over time, indicative of a maturing ecosystem transitioning from explosive, frontier-shifting innovation to optimized refinement. Furthermore, localized managerial optimization, measured by the Efficiency Change (EC) index, displays significant regional heterogeneity. While markets like Brazil demonstrated aggressive late-stage efficiency spikes, and core Southeast Asian markets (such as Vietnam and the Philippines) maintained highly stable, competitive trajectories, other regions, such as Thailand, experienced notable managerial regression. This regression signals severe internal frictions in optimizing local marketing budgets against rising credit defaults. Managerial Implications: These findings provide critical strategic insights for orchestrators of the multinational e-commerce ecosystem. The empirical evidence suggests that relying exclusively on centralized technological scaling, such as unified platform infrastructure and global AI architectures, is insufficient for sustained operational growth. To maintain a competitive advantage, operations executives must deploy hyper-localized resource-allocation and customer-acquisition frameworks tailored to specific regional market dynamics and consumer behavior. Sustainable scaling in cross-border digital commerce requires a precise dynamic equilibrium: leveraging robust global technological infrastructure while executing highly adaptive, market-specific operational and marketing optimizations to maximize customer lifetime value (CLV), eliminate customer acquisition waste, and streamline localized transaction and engagement cycles. Full article

In the Western Anti-Atlas, the stratiform copper mineralisations of the Tabia Member are distinctly hosted within sedimentary rocks, forming deposits comparable to the sediment-hosted stratabound copper (SSC) type. In Tizert, Ouarmdaz, and Talat n’Ouamane, the sandstones and clays host primary sulphides as disseminations and copper carbonates occurring as small continuous or lenticular beds, small geodes, and coatings on clay laminations. In Tiferki, the conglomerates are more mineralised; sulphides are represented by chalcopyrite and bornite, locally altered into chalcocite. Copper carbonates appear as malachite and azurite coating pebbles and deeply impregnate the granular matrix. From a sedimentological and sequence-stratigraphic perspective, the conglomeratic level at Tiferki represents a low-sea-level prism (LST) formed during a stage of maximum platform exposure allowing erosion products to accumulate at the base of the slope. By contrast, the silty–sandy complex known as the “Talat n’ Ouamane level” corresponds to a prograding sedimentary sequence (Highstand Systems Tract, HST) that developed after a relative regression where the platform was only partially exposed. In both cases, relative sea-level fall promoted the emergence of the hinterland and active erosion, supplying the depositional environment with coarse detrital during the maximum drops in sea level and sandy-to-micro-conglomeratic deposits during less pronounced regressions. This detrital supply probably brought copper in the form of grains and especially dissolved copper, which is largely deposited at the bottom of slopes and in marine environments where microbial communities contribute to the precipitation of copper. The selective distribution of sulphides within the Tabia Member shows that only the low-sea-level and high-sea-level prisms are enriched in stratiform copper, while the transgressive and maximum flooding systems tracts lack significant mineralisation. This suggests that sea-level variations played a controlling role in copper deposition. The regressive trend of the high sea-level suite continued until emergence, reflected by the invasion of the environment by red or ochre siltstones from the alluvial plain. The emergence at the transition from sandstones to dolomites, and the “Red Beds”-type mineralisation embedded in these subaerial facies, support the syngenetic origin of the sulphides in the Tabia Member. After this main emplacement of stratiform copper, the sulphides would have undergone remobilisation and alteration during a significant episode of vertical water escape that deformed the host facies and enhanced mineral concentration within permeable sandstones. Even the clays were delaminated and coated with a thin films of copper carbonates on their surfaces. In the Tamjout dolomites, carbonates are percolated by acidic solutions, contributing to the brecciation and silicification of the stromatolitic layers. Copper mineralisation accompanies this silicification and fills the karstification pockets. Full article

In this study, an electrochemical immunosensor based on polyaniline/nano-Fe₃O₄ (PANI/Nano-Fe₃O₄) nanocomposite (PANI/Nano-Fe₃O₄/Anti-FLA/BSA/GCE) was developed for the highly sensitive and selective detection of trace levels of fluoranthene (FLA) in marine environments. Fluoranthene antibodies (Anti-FLA) were covalently immobilized on a glassy carbon electrode (GCE) modified with PANI/Nano-Fe₃O₄ via an EDC/NHS activation strategy, enabling specific recognition of FLA based on the antigen–antibody binding mechanism. The performance of the sensor was systematically optimized using cyclic voltammetry (CV), electrochemical impedance spectroscopy (EIS), linear sweep voltammetry (LSV), and differential pulse voltammetry (DPV). The results demonstrated a linear inverse relationship between peak current (I_p) and FLA concentration in the range of 0.5~80 ng/mL, with a regression equation of I = −1.55C + 174.602 (R² = 0.996). The limit of detection (LOD) was as low as 0.354 ng/mL (S/N = 3). In real seawater sample analysis, spiked recovery tests at three representative sites in the Maowei Sea, Guangxi, yielded recoveries of 95.44%~97.51%, with RSDs below 3%, confirming the sensor’s resistance to matrix interference. The synergistic effect of the porous conductive network of PANI and the high specific surface area of Nano-Fe₃O₄ significantly amplified the electrochemical signal, while the molecular specificity of the antibody ensured targeted recognition. This sensor provides a novel and effective approach for the on-site rapid detection of polycyclic aromatic hydrocarbon (PAH) pollutants in complex marine environments, offering both high sensitivity and selectivity. Full article

Background: Extreme drought and flood events continuously threaten the stability of forest and crop production. Long-term hydrological records derived from tree-ring anatomical proxies provide critical evidence for revealing historical drought hazard differentiation. Methods: Cell wall thickness chronologies of Betula platyphylla (northern forest) and Picea koraiensis (southern agro-pastoral zone) were developed to reconstruct nearly 300-year annual runoff sequences. Pearson correlation, quadratic regression, wavelet transform and superposed epoch analysis (SEA) were applied to quantify hydrological evolution, periodic signals, large-scale climate forcing and statistical coupling between dry/wet extremes and historical yield reduction records. Results: The northern watershed showed stronger interannual runoff oscillation. Both regions entered persistent low-flow phases post-1950. Pacific Decadal Oscillation (PDO) acted as the dominant driver, while solar radiation exerted weak secondary regulation. Severe drought events corresponded closely to historical forest and grain yield losses, with far higher agricultural vulnerability in the southern agro-pastoral ecotone. Conclusions: This study reconstructed the long-term historical runoff of the Greater Khingan Range from the thickness of the cell wall, analyzed the different impacts of PDO on it, and clarified the differentiated effects of drought and flood on agricultural and forestry production losses and the interrelated impact of land use on hydrology and the value of agricultural output. Full article