Search Results (788)

Climate-driven extreme rainfall is increasing landslide risk, whereas conventional regional rainfall thresholds inadequately represent spatial variations in slope preconditioning across complex mountainous terrain. This study develops a dynamic early-warning framework that couples slope unit landslide susceptibility with rainfall triggering for Jinyang County, Southwest China. The county was divided into 7249 slope units, and eight conditioning factors were used to train support vector machine (SVM) and extreme gradient boosting (XGBoost) models with Bayesian hyperparameter optimization. Bayesian-optimized XGBoost (BO-XGBoost) achieved the best discrimination, with an area under the curve (AUC) of 0.893. For temporal triggering, 113 historical rainfall-landslide events were analyzed in logarithmic intensity–duration (I–D) space. A power-law fit described the central trend, and the fifth percentile of residuals defined a 5% non-exceedance threshold exceeded by approximately 95% of triggering events. Residual-based thresholds were used to establish four rainfall-trigger levels. A gated risk matrix coupled these levels with susceptibility classes. Validation using 162 online monitoring sites during the 21 August 2023 extreme rainfall event yielded 85.2% accuracy, 83.6% precision, 81.2% recall, 88.2% specificity, and an F1 score of 82.4%, demonstrating practical potential for spatially refined regional landslide early warning. Full article

Climate change and rapid urbanisation are intensifying the urban heat island effect, increasing thermal stress, degrading air quality, and leading to rising energy demand. Monitoring neighbourhood-scale heat requires Land Surface Temperature (LST) observations at fine spatial and temporal resolutions, yet satellite thermal products are limited by revisit frequency, acquisition time, and cloud cover. This study developed a novel approach integrating satellite-derived land cover characteristics with continuous contact-based temperature measurements from low-cost LoRaWAN sensors and geostatistical modelling to generate hourly LST maps at 10 m resolution. The technique provides communities with simpler, affordable methods for measuring heat islands and supporting mitigation strategies. Observations from 52 locations across Curtin University’s Bentley campus in Perth, Western Australia, were combined with land cover indices. Empirical Bayesian Kriging captured spatial and temporal urban heat patterns with a root mean square error of approximately 3 °C, representing a bias of near 1 °C. Predictions were consistent with Landsat-derived LST, revealing persistent heat retention over asphalt and cooler conditions associated with vegetation. Integrating satellite-derived predictors with ground measurements provides continuous fine-scale information to identify local heat hotspots and inform targeted mitigation. Unlike satellite data, these low-cost ground measurements could be collected with the help of urban practitioners, developers, and academic institutions. Full article

Persistent late-time variation can remain in physics-informed neural network (PINN) solutions after the governing transient has effectively decayed, while conventional error norms do not reveal whether this variation has a systematic temporal–frequency structure. This study develops an offline temporal–spectral diagnostic and postprocessing workflow for a one-dimensional advection–diffusion benchmark. A high-accuracy analytical reference and three fixed-resolution finite-difference baselines are used to assess a PINN whose architecture is selected by a fully supervised neural architecture search and whose parameters are trained with progressive temporal windowing. Candidate late-time intervals are selected without using the reference solution by applying the Bayesian information criterion (BIC) to a breakpoint model for the inter-reconstruction sensitivity; the selected field is subsequently reconstructed by retaining a prescribed fraction of its temporal spectral energy and is evaluated independently through reference-error and physics-consistency measures. For $[t_{\mathrm{cut}},t_{\max }]=[1.8,5]$, the zero-frequency component contains $0.9999996$ of the raw-field energy, so the $q=0.95$ reconstruction retains only the temporal mean. This projection reduces the final-time spatial error norm from $2.70\times {10}^{-3}$ to $1.06\times {10}^{-3}$, a factor of approximately $2.5$, while changing the discrete governing-equation residual by less than $0.3\%$ over the filtered window. Mean-removed tests for $q=0.90,0.95,0.99$ show that the discarded fluctuation is dominated by low-frequency approximation error rather than high-frequency noise. The result supports the proposed selection–validation workflow for this controlled benchmark but does not establish a universally transferable filter. Full article

Geological hazards pose persistent challenges to sustainable land use planning, infrastructure safety, and community resilience in alpine valley regions. Taking Tongren City on the northeastern margin of the Qinghai–Tibet Plateau as the study area, this study assessed composite susceptibility to landslides, rockfalls, and debris flows using 18,136 slope units, 189 historical hazard points, and 92 independent verification points, with the three hazard types combined into a single positive class for regional screening. Twelve conditioning factors were retained after collinearity testing, and negative-sample buffer distances of 200, 500, 1000, and 1500 m were compared before applying Bayesian-optimized random forest (Bo-RF), Bayesian-optimized CatBoost (Bo-CatBoost), and TabPFN models. Model performance was evaluated through repeated buffered spatial block cross-validation, probability calibration, spatial error analysis, independent verification, and SHAP interpretation. The 1000 m buffer produced the best negative-sample performance, with accuracy of 0.905 and an ROC-AUC of 0.972. Under spatial validation, the three models showed comparable discrimination, with ROC-AUC values of 0.9152–0.9161 and accuracy values of 0.8509–0.8549; Bo-CatBoost performed better in probability calibration, whereas TabPFN exhibited weaker spatial clustering of residuals. The TabPFN very-high-susceptibility zone covered 17.03% of the study area and contained 84.66% of historical hazard points, while the corresponding Bo-CatBoost zone captured 72.83% of independent verification points. SHAP analysis identified the mean annual rainfall, distance to water systems, distance to roads, NDVI, and lithology as the principal factors shaping the composite susceptibility pattern. Overall, the proposed framework provides spatial decision support for sustainable land use planning, resilient infrastructure management, targeted hazard investigation, and the efficient allocation of disaster prevention resources in alpine valley regions. Full article

Vegetation–hydrothermal trend patterns in semiarid mountains are inherently multidimensional across altitudinal gradients, requiring probabilistic frameworks to characterize overlapping vegetation–hydrothermal trend patterns. We develop a Gaussian mixture model (GMM)-based framework to classify vegetation trend regimes from multidimensional trend variables and to characterize their climate–regime associations across elevation gradients in the semiarid Al Baha region, Saudi Arabia. Landsat time-series data (2014–2025) were used to derive trends in the Normalized Difference Vegetation Index (NDVI), normalized difference water index (NDWI), and land surface temperature (LST) using the Mann–Kendall test. These standardized trend variables were integrated within a probabilistic GMM framework, with candidate models evaluated using Bayesian information criterion (BIC), Integrated Completed Likelihood (ICL), classification entropy, resampling-based partition stability, and held-out predictive performance. Although BIC favored the eight-component VVV model, the six-component solution was retained as a balanced intermediate-complexity representation based on partition reproducibility, near-optimal predictive performance, classification ambiguity, and parsimony. The results reveal a pronounced elevation-dependent reorganization of vegetation–hydrothermal trends, characterized by High Relative LST Trend dominance at low elevations (~44%), Mixed Trend State at mid elevations (~40.8%), and Low Relative NDWI Trend at higher elevations (>48%), with High Relative LST Trend becoming negligible at the highest altitudes. This spatial organization showed a moderate association with elevation (Cramér’s V = 0.299; χ² = 74,265.54, p < 0.001), with statistical significance interpreted cautiously given the large, spatially autocorrelated sample. Climate–regime analysis further reveals that precipitation variability exhibits the strongest association at mid elevations, whereas temperature shows stronger associations at both low- and high-elevation extremes. Cross-sensor comparison with VIIRS showed substantial classification consistency after class-label alignment (overall agreement ≈ 0.76; κ ≈ 0.64). These findings demonstrate systematic variation in vegetation–hydrothermal trend regimes and their climate associations along elevation gradients. The proposed GMM-based framework provides a scalable and uncertainty-aware approach for assessing climate–vegetation associations in semiarid mountain systems and other data-scarce dryland environments. Full article

The large-scale integration of renewable energy has compelled coal-fired power plants to operate under deep peak-shaving conditions, significantly increasing the failure risk of Circulating Fluidized Bed (CFB) boiler heating surfaces due to severe thermal and pressure fluctuations. To address the limitations of traditional static risk evaluations, this study proposes a novel dynamic risk assessment framework integrating a Spatial–Temporal Graph Convolutional Network (STGCN) and a Dynamic Bayesian Network (DBN). The STGCN, enhanced with an operation-adaptive dynamic cross-attention delay module, predicts spatiotemporal temperature and pressure variations across the high-temperature heating surfaces. The predicted variables are incorporated into the DBN as dynamic evidence, which utilizes Noisy-OR logic and an embedded Weibull physical degradation model to continuously quantify cumulative failure probabilities. Case study results demonstrate that the STGCN outperforms traditional LSTM and RNN baselines in prediction accuracy. Furthermore, the DBN effectively maps the distinct degradation characteristics of individual boiler components, accurately identifying the water wall and superheater as having the highest failure risks and the largest fluctuations in marginal failure probability during rapid load cycling. This integrated data-driven approach provides highly accurate, real-time risk predictions, offering essential decision-making support for the predictive maintenance and safe flexible operation of CFB boilers. Full article

Data-driven site characterization in geotechnical engineering increasingly relies on high-dimensional waveform data and computationally intensive inverse modeling. Full waveform inversion and finite element model updating typically rely on gradient-based or Bayesian optimization, requiring many serial forward simulations, making large-scale applications computationally expensive. In this pilot study, we propose an alternative methodology based on physics-informed neural surrogate models—FE–PINN (continuous time) and FE–NODE (discrete time)—that embed fully assembled spatial finite-element mass–damping–stiffness matrices directly into physics loss residuals. This framework infers subsurface material parameters from sparse geodata without requiring repeated forward solver execution during iteration. Using a two-dimensional synthetic site characterization example, we demonstrate that the proposed surrogates can recover soil parameters with good accuracy, exhibit robustness to measurement noise, and reveal sensitivity to parameter initialization. Across a 240-case sensitivity sweep varying data sparsity, noise (0–50%), parameter initialization (±30%), and material heterogeneity, the proposed surrogates successfully recover soil parameters from as few as 2 measured surface degrees of freedom. Specifically, FE–PINN demonstrates strong noise tolerance, maintaining relative parameter errors below 5% for up to 25% measurement noise and 30% initialization error. In contrast, FE–NODE achieves fast convergence in clean, homogeneous cases but fails to reliably converge in heterogeneous media, where parameter errors reach 90–113% due to discrete finite-difference error propagation. The results highlight the potential and operational limits of finite-element-based neural surrogates as an auto-differentiable framework for geotechnical inverse analysis. Full article

Accurate monitoring of surface water quality remains challenging due to pronounced spatial heterogeneity and limited ground observations. Satellite remote sensing offers scalable solutions, yet the extent to which environmental drivers enhance predictive performance, particularly in complex river–lake systems, remains insufficiently understood. This study develops a parallel comparative river–lake modeling framework to estimate permanganate index (CODmn), total phosphorus (TP), and total nitrogen (TN) by integrating satellite spectral data with climatic and land-use variables. Four model configurations were evaluated: spectral predictors alone (M₁), spectral predictors combined with climate variables (M₂), spectral predictors combined with land-use information (M₃), and full integration of all predictors (M₄). LightGBM models were optimized using Bayesian hyperparameter tuning (Optuna) and trained over rivers (January 2021–July 2025) and lakes (January 2021–December 2024) datasets in the Dongliao Basin, China. Spectral predictors alone (M₁) provided robust performance for CODmn ($R^2$ = 0.78), with marginal improvement when land-use variables were included (M₂) in rivers ($R^2$ = 0.80). In contrast, nutrient predictions showed stronger dependence on environmental covariates. TN predictions improved substantially with land-use inputs ($M3$) (R² = 0.75 in rivers and 0.63 in lakes with M₂), with further gains with full integration ($M4$) in lakes (${\mathrm{R}}^2$ = 0.66). TP predictions exhibited marked improvements with land-use variables in rivers (R² = 0.76) and with full integration in lakes ($R^2$ = 0.72). Model interpretability analysis using SHAP revealed that spectral features dominate CODmn estimation, while climatic and watershed characteristics exert greater influence on TN variability. Seasonal analysis indicated that hydrological drivers dominate during wet seasons, while land-use effects and internal biogeochemical processes become more important in dry seasons. The proposed framework advances predictive accuracy and process understanding, supporting more effective monitoring and management of water quality in complex river–lake systems. Full article

Earthquakes remain among the most destructive natural hazards, necessitating reliable monitoring and early warning systems for effective risk mitigation. Recent advances in machine learning (ML) and deep learning (DL) have significantly improved seismic signal analysis, enabling more accurate event detection, phase picking, classification, and magnitude estimation. This study presents a systematic review of ML- and DL-based approaches for earthquake monitoring, with particular emphasis on Distributed Acoustic Sensing (DAS) as an emerging technology for high-resolution, real-time seismic observation. Following the PRISMA 2020 guidelines, a systematic literature search was conducted across Scopus, Web of Science, IEEE Xplore, and Google Scholar, yielding 252,223 initial records. After applying the predefined publication period, removing duplicate records, conducting relevance screening, and performing eligibility assessment, 138 peer-reviewed studies published between 2021 and 2025 were retained for detailed analysis and synthesis. The review reveals a significant transition from conventional signal-processing techniques to advanced artificial intelligence-based approaches, including Convolutional Neural Networks (CNNs), Long Short-Term Memory (LSTM) networks, Bidirectional Long Short-Term Memory (BiLSTM) networks, Transformer-based architectures, hybrid models, and Bayesian learning methods for uncertainty quantification. The findings further demonstrate that Distributed Acoustic Sensing (DAS) has emerged as a transformative sensing technology because of its dense spatial coverage, high spatial resolution, and continuous monitoring capability. However, several challenges remain, including the lack of standardized datasets, limited model generalization across diverse geological settings, insufficient model interpretability, high computational complexity, and the limited integration of uncertainty-aware approaches for real-time seismic monitoring. This review identifies these critical research gaps and highlights promising future research directions, including multimodal data fusion, interpretable artificial intelligence, physics-informed learning, self-supervised learning, and robust uncertainty quantification for next-generation intelligent seismic monitoring systems. Unlike previous review studies that primarily focus on individual machine learning techniques or conventional seismic monitoring, this review provides a comprehensive and systematic synthesis of recent advances in machine learning, deep learning, and Distributed Acoustic Sensing (DAS), identifies current research gaps, and offers practical recommendations to guide future research on intelligent earthquake monitoring systems. Full article

Bounded observations occur in a variety of applied scenarios such as reliability probabilities, degradation measures, rates, and other functions defined on the unit interval for which flexible distributional and hazard-rate behavior are desired. This paper presents the Unit Harris Extended Exponential Distribution (UHEED), a three-parameter unit distribution derived from the Harris Extended Exponential Distribution to offer more flexibility in these types of data, including the ability to model bathtub-shaped hazard rates. Several mathematical properties are developed, such as parameter identifiability, quantile elasticity, moments, order statistics, and Shannon entropy. Maximum likelihood estimation is used as classical inference; the squared error and LINEX loss functions are used to develop the Bayesian estimation with a Metropolis–Hastings within Gibbs algorithm. A representative-point approximation is also proposed for evaluating important distributional quantities, including moments and reliability measures. The Monte Carlo results indicate that the more samples, the more accurate the results of the estimation. A useful example of the UHEED is presented with naturally bounded bramble cane spatial-coordinate data, and goodness-of-fit comparisons show that it performs competitively relative to the competing unit distributions used in the analysis. Full article

Background/Objectives: Immunization coverage in the Democratic Republic of the Congo (DRC) consistently falls short of global standards, placing the country among those with the highest prevalence of unvaccinated or “zero-dose” children. This study aimed to generate high-resolution (1 km × 1 km) geospatial estimates of coverage for the first to third doses of diphtheria–tetanus–pertussis vaccine (DTP1–3) and the first dose of measles-containing vaccine (MCV1), along with estimates of zero-dose and under-vaccinated children to support vaccination programming in DRC. Methods: We used Bayesian geostatistical modelling to integrate data from the 2023 Enquête de Couverture Vaccinale (ECV) survey with geospatial covariates and harmonized population datasets. The model generated vaccine coverage and dropout rates at 1 × 1 km resolution while accounting for spatial dependence and prediction uncertainty. Grid-level estimates were aggregated to health areas, health zones, and provinces using population-weighted averages, and combined with under-one population estimates to quantify zero-dose and DTP-under-vaccinated children. Results: The grid-level unconstrained coverage estimates (estimation over all land grid squares) showed substantial variability across the country (DTP1: 13.3 to 99.7%; DTP2: 12.3 to 96.8%; DTP3: 3.1 to 93.6%; and MCV1: 7.5 to 93.6%). In total, 56.4% of mapped health areas achieved DTP1 coverage above 80%, but this declined to 32% for DTP2, 15% for DTP3, and 8.3% for MCV1. Around 60.5% of the health zones achieved DTP1 coverage ≥ 80%, while only 2.7% fell below 40%, indicating that substantial disparities persist across these zones. Nationally, based on constrained coverage estimates (only areas with identified built settlements) applied to an estimated 4.2 million children under 1 year in 2024, 18.4% had not received DTP1, 20.7% were DTP-under-vaccinated (received DTP1 but not DTP3), and 46.8% remained unvaccinated for MCV1. Conclusions: The prevalence of large numbers of zero-dose children, together with increased dropout rates, underscores systemic issues in access and follow-up initiatives. These findings are critical for microplanning and targeted outreach to achieve equitable immunization coverage in line with the Immunization Agenda 2030’s goal of leaving no child behind. Full article

The widespread use of location-based services (LBSs) has led to the continuous collection of user location and trajectory data, increasing the risk of privacy leakage and creating a persistent tradeoff between privacy protection and data utility. To address this problem in discrete location query scenarios, this paper proposes a single-point location privacy protection method based on Q-R tree retrieval and differential privacy, termed QRDPP. QRDPP combines the adaptive spatial partitioning capability of a Q-tree with the minimum bounding rectangle (MBR)-based indexing capability of an R-tree. It applies an improved geometric privacy budget allocation strategy to leaf nodes and an arithmetic allocation strategy to non-leaf nodes, followed by Laplace perturbation of the corresponding location data and node information. For continuous trajectory query scenarios, this paper proposes a spatiotemporal generalization and differential privacy method, termed STG-DPTP, to address inadequate temporal protection, inappropriate generalization, and trajectory distortion. STG-DPTP performs hierarchical spatiotemporal clustering, separately models temporal and spatial distributions using Gaussian kernel density estimation, dynamically optimizes bandwidth parameters through Bayesian optimization, selects representative candidate subsets using the exponential mechanism, and generates protected trajectories through constrained sampling. Experiments on the GeoLife dataset evaluate the proposed methods in terms of query accuracy, computational efficiency, spatial trajectory similarity, reconstruction error, adversarial uncertainty, and temporal preservation. The results show that QRDPP improves the utility and efficiency of privacy-preserving spatial queries, while STG-DPTP better preserves the spatial distribution, trajectory structure, and temporal characteristics of the original data under the adopted differential privacy framework. Full article

West Nile virus (WNV) transmission has intensified and expanded in Italy, but quantitative evidence on how local climatic and environmental conditions influence human risk in southern regions remains limited. This study examined the association between meteorological, environmental, and host-related factors and West Nile virus (WNV) cases at the municipal level in Apulia in 2023. Human WNV cases were georeferenced at the municipal level and linked to monthly indicators (minimum and maximum temperature, precipitation, surface water extent, green area coverage, land-use change, avian occurrence). A Bayesian spatio-temporal Poisson model with a conditional autoregressive structure was fitted to monthly counts of human WNV cases, with equine WNV cases, climatic, environmental and avian indicators included as covariates. Eight human WNV cases were reported between August and October and four equine WNV cases between September and November, with partial spatial and temporal overlap. In univariable analyses, the strongest associations were observed for meteorological variables, particularly temperature. In the final multivariable model, higher maximum temperature at a two-month lag was associated with increased WNV risk (RR = 1.53; 95% CrI: 1.18–2.23), while minimum temperature was excluded due to collinearity with maximum temperature. Green area coverage and water body extent showed uncertain effects. Model-based maps indicated that elevated fitted risk was concentrated in a narrow temporal window between August and October, expanding sharply across the region in September before receding, rather than describing a stable, spatially fixed hotspot. This exploratory analysis suggests that reported human WNV infections in Apulia in 2023 were temporally concentrated during the late summer/early autumn period and that maximum temperature at a two-month lag was positively associated with the outcome in the selected model. The findings are hypothesis-generating and should be interpreted with caution given the very small number of events, but they illustrate the feasibility of integrating multisource epidemiological, climatic, environmental, and veterinary data to support locally tailored early-warning efforts in southern Italy. Full article

Generative data augmentation has been widely explored in Wi-Fi fingerprint-based indoor localization to reduce the cost of dense radiomap construction, with many studies reporting substantial localization improvements. However, these comparisons typically rely on default or weakly optimized baseline regressors, making it difficult to determine whether reported gains reflect genuine synthesis quality or merely compensate for suboptimal baselines. In this paper, we systematically investigate under which conditions generative augmentation is actually justified. We fine-tune four widely used localization regressors—kNN, SVR, XGBoost, and DNN—using Bayesian hyperparameter optimization and establish strong non-augmented baselines across radiomaps with controlled levels of spatial sparsity, constructed via farthest-point sampling. We then train five representative generative models—VAE, GAN, DDPM, DiT, and TDPM—within a unified augmentation pipeline that includes quality filtering and pseudo-labeling, and benchmark them against these baselines. Using two publicly available datasets, we show that none of the generative models consistently outperforms a non-augmented, fine-tuned baseline regressor such as XGBoost or kNN, across a wide range of sparsity levels. We further show that these conclusions are robust to three potential confounders: various proportions of synthetic data, the choice of localization regressor (ruling out circularity with the pseudo-labeling model), and the dataset itself, since the findings on the first dataset replicate on a second, structurally different building. These findings suggest that reported augmentation benefits in prior work may partly reflect under-optimized baselines rather than genuine synthesis quality, and that generative augmentation should be treated as a conditional last resort rather than a universal improvement strategy. Full article

Electromobility is central to urban decarbonisation, but its charging infrastructure must be sized under substantial uncertainty about user behaviour that varies across stations, time of day and user type. This study couples a hierarchical Bayesian model with an agent-based, discrete-event simulation of a charging network. It is fitted by Markov chain Monte Carlo to the public ACN-Data dataset (13,694 sessions across 52 stations; 16,468 user requests), with partial pooling across stations. Posterior parameters drive a 24 h simulation of 500 vehicles across nine configurations and 30 to 180 slots. Service success rises from 19% to 81% and mean waiting falls from 110 to 62 min; long workplace dwell times limit turnover, so capacity rather than energy binds. TOPSIS with a paired bootstrap selects 160 slots under balanced weighting, but that optimum holds for only a tenth of the weight simplex, and the recommendation spans 140–180 slots. Spreading the arrival peak at fixed hardware raises service from 67% to 83%, matching a 29% expansion. Spatially explicit assignment costs three percentage points when stations are evenly sited, and six when clustered. The framework makes the cost of over-provisioning explicit and preference-conditional rather than naming a single optimum, giving a reproducible basis for sustainable capacity planning. Full article