Search Results (857)

The increasingly frequent and severe natural disasters have posed significant challenges to the resilience of power systems worldwide, creating an urgent need to investigate the security issues associated with these extreme events and to develop effective risk mitigation strategies. Meanwhile, as one of the leading topics in current research, artificial intelligence (AI) has demonstrated outstanding performance across various domains, such as AI-driven smart grids and smart cities. In particular, its efficiency in processing big data and solving complex computational problems has made AI a powerful tool for supporting decision-making in complex scenarios. This article presents a focused overview of power system resilience against natural disasters, highlighting recent advancements in AI-based approaches aimed at enhancing system security and response capabilities. It begins by introducing various types of natural disasters and their corresponding impacts on power systems. Then, a systematic overview of AI applications in power systems under disaster scenarios is provided, with a classification based on the task categories, i.e., predictive, descriptive and prescriptive tasks. Following this, this article analyzes current research trends and finds a growing shift from knowledge-based models towards data-driven models. Furthermore, this paper discusses the major challenges in this research field, including data processing, data management, and data analytics; the challenges introduced by large language models in power systems; and the limitations related to AI model interpretability and generalization capability. Finally, this article outlines several potential future research directions. Full article

Urban underpasses are critical flood-prone hotspots during extreme rainfall, posing significant threats to urban resilience and infrastructure safety. However, a scale gap persists between catchment-scale hydrological models, which often oversimplify local geometry, and high-fidelity hydrodynamic models, which typically lack realistic boundary conditions. To bridge this gap, this study develops a multi-scale framework that integrates the Storm Water Management Model (SWMM) with 3D Computational Fluid Dynamics (CFD). The framework employs a unidirectional integration (one-way forcing), utilizing SWMM-simulated runoff hydrographs as dynamic inlet boundaries for a detailed CFD model of a 2.8 km underpass in Shanghai. Simulations across six design rainfall events (2- to 50-year return periods) revealed two distinct flooding mechanisms: a systemic response at the hydraulic low point, governed by cumulative inflow; and a localized response at entrance concavities, where water depth is rapidly capped by micro-topography. Informed by these mechanisms, an intensity-graded drainage strategy was developed. Simulation results show significant differences between different drainage strategies. Through this framework and optimized drainage system design, significant water accumulation within the underpass can be prevented, enhancing its flood resistance and reducing the severity of disasters. This integrated framework provides a robust tool for enhancing the flood resilience of urban underpasses and offers a basis for the design of proactive disaster mitigation systems. Full article

Effective disaster management is critical for safeguarding lives, infrastructure and economies in an era of escalating natural hazards like floods and landslides. Despite advanced early-warning systems and coordination frameworks, a persistent “last-mile” challenge undermines response effectiveness: transforming fragmented and unstructured multimodal data into timely and accountable field actions. This paper introduces ResQConnect, a human-centered, AI-powered multimodal multi-agent platform that bridges this gap by directly linking incident intake to coordinated disaster response operations in hazard-prone regions. ResQConnect integrates three key components. It uses an agentic Retrieval-Augmented Generation (RAG) workflow in which specialized language-model agents extract metadata, refine queries, check contextual adequacy and generate actionable task plans using a curated, hazard-specific knowledge base. The contribution lies in structuring the RAG for correctness, safety and procedural grounding in high-risk settings. The platform introduces an Adaptive Event-Triggered (AET) multi-commodity routing algorithm that decides when to re-optimize routes, balancing responsiveness, computational cost and route stability under dynamic disaster conditions. Finally, ResQConnect deploys a compressed, domain-specific language model on mobile devices to provide policy-aligned guidance when cloud connectivity is limited or unavailable. Across realistic flood and landslide scenarios, ResQConnect improved overall task-quality scores from 61.4 to 82.9 (+21.5 points) over a standard RAG baseline, reduced solver calls by up to 85% compared to continuous re-optimization while remaining within 7–12% of optimal response time, and delivered fully offline mobile guidance with sub-500 ms response latency and 54 tokens/s throughput on commodity smartphones. Overall, ResQConnect demonstrates a practical and resilient approach to AI-augmented disaster response. From a sustainability perspective, the proposed system contributes to Sustainable Development Goal (SDG) 11 by improving the speed and coordination of disaster response. It also supports SDG 13 by strengthening adaptation and readiness for climate-driven hazards. ResQConnect is validated using real-world flood and landslide disaster datasets, ensuring realistic incidents, constraints and operational conditions. Full article

Online visual object tracking is a critical component of remote sensing-based aerial vehicle physical tracking, enabling applications such as environmental monitoring, target surveillance, and disaster response. In real-world remote sensing scenarios, the inherent processing delay of tracking algorithms results in the tracker’s output lagging behind the actual state of the observed scene. This latency not only degrades the accuracy of visual tracking in dynamic remote sensing environments but also impairs the reliability of UAV physical tracking control systems. Although predictive trackers have shown promise in mitigating latency impacts by forecasting target future states, existing methods face two key challenges in remote sensing applications: weak correlation between trackers and predictors, where predictions rely solely on motion information without leveraging rich remote sensing visual features; and inadequate modeling of continuous historical memory from discrete remote sensing data, limiting adaptability to complex spatiotemporal changes. To address these issues, we propose TPC-Tracker, a Tracker-Predictor Correlation Framework tailored for latency compensation in remote sensing-based aerial tracking. A Visual Motion Decoder (VMD) is designed to fuse high-dimensional visual features from remote sensing imagery with motion information, strengthening the tracker-predictor connection. Additionally, the Visual Memory Module (VMM) and Motion Memory Module (M3) model discrete historical remote sensing data into continuous spatiotemporal memory, enhancing predictive robustness. Compared with state-of-the-art predictive trackers, TPC-Tracker reduces the Mean Squared Error (MSE) by up to 38.95% in remote sensing-oriented physical tracking simulations. Deployed on VTOL drones, it achieves stable tracking of remote sensing targets at 80 m altitude and 20 m/s speed. Extensive experiments on public UAV remote sensing datasets and real-world remote sensing tasks validate the framework’s superiority in handling latency-induced challenges in aerial remote sensing scenarios. Full article

Rapid and reliable building damage assessment (BDA) is crucial for post-disaster emergency response. However, existing methods face challenges such as complex background interference, the difficulty in jointly modeling local geometric details and global spatial dependencies, and adverse weather conditions. To address these issues, this paper proposes the Adaptive Difference State-Space Fusion Network (ADSFNet), capable of processing both optical and Synthetic Aperture Radar (SAR) data to alleviate weather-induced limitations. To achieve this, ADSFNet innovatively introduces the Adaptive Difference Attention Fusion (ADAF) module and the Hybrid Selective State-Space Convolution (HSSC) module. Specifically, ADAF integrates pre- and post-disaster features to guide the network to focus on building regions while suppressing background interference. Meanwhile, HSSC synergizes the local texture extraction of CNNs with the global modeling strength of Mamba, enabling the simultaneous capture of cross-building spatial relationships and fine-grained damage details. Experimental results on sub-meter high-resolution MultiModal (BRIGHT) and optical (xBD) datasets demonstrate that ADSFNet attains F1 scores of 71.36% and 73.98%, which are 1.29% and 0.6% higher than the state-of-the-art mainstream methods, respectively. Finally, we leverage the model outputs to construct a disaster-centric knowledge graph and integrate it with Large Language Models to develop an intelligent management system, providing a novel technical pathway for emergency decision-making. Full article

Mountain floods, one of the most common and destructive natural disasters worldwide, pose significant challenges to disaster prevention due to their sudden onset, high destructive power, and severe localized impacts. This study proposes an innovative flash flood early warning system based on a distributed hydrological model ensemble. The main objective is to improve the prediction and early warning accuracy of flash flood disasters by integrating multi-source data and regional modeling. The system simulates flood flow and risk levels under different rainfall scenarios to provide timely warnings in mountainous areas. A case study of a heavy rainfall event in Ma Jia Natural Village, Jiangxi Province was used to validate the system’s performance. Through regionalized parameter calibration within the ensemble, the system achieved Nash–Sutcliffe Efficiency (NSE) values exceeding 0.88, while the simulated peak discharges deviated from observed values by only 1.5%, 9.5%, and 4.8% under 3 h, 6 h, and 24 h rainfall scenarios, respectively, demonstrating the improved quantitative accuracy of flood prediction enabled by the ensemble-based framework. The system showed high consistency with observed data, accurately predicting flood responses at 3, 6, and 24 h time scales and providing reliable risk warnings. This approach not only enhances warning accuracy across multiple temporal scales but also supports risk-level early warnings at both river-section and village scales, offering significant practical value for the prevention of mountainous flood disasters. Full article

The increasing severity of extreme weather events necessitates rapid methods for post-disaster damage assessment. Current remote sensing datasets often lack the spatial resolution required for a detailed evaluation of critical waterfront infrastructure, which is vulnerable during hurricanes. To address this limitation, we introduce the Flood and Waterfront Infrastructure Segmentation Dataset (FWISD), a new dataset constructed from high-resolution unmanned aerial vehicle imagery captured after a major hurricane, comprising 3750 annotated 1024 × 1024 pixel image patches. The dataset provides semantic labels for 11 classes, specifically designed to distinguish between intact and damaged structures. We conducted comprehensive experiments to evaluate the performance of both convolution and Transformer-based models. Our results indicate that hybrid models integrating Transformer encoders with convolutional decoders achieve a superior balance of contextual understanding and spatial precision. Regression analysis indicates that the distance to water has the maximum influence on the detection success rate, while comparative experiments emphasize the unique complexity of waterfront infrastructure compared to homogenous datasets. In summary, FWISD provides a valuable resource for developing and evaluating advanced models, establishing a foundation for automated systems that can improve the timeliness and precision of post-disaster response. Full article

To align with disaster monitoring and sustainable risk assessment, the low-carbon transition of fisheries necessitates comprehensive carbon emission management throughout the supply chain. As China advances supply-side structural reform, transitioning from traditional to low-carbon fisheries is vital for the green development of the industry and its associated sectors. This study employs input–output models and LMDI decomposition to examine the trends and drivers of embodied carbon emissions within China’s fishery production system from 2010 to 2019. By constructing a cross-sectoral full-emission accounting system, the research calculates total direct and indirect emissions, exploring how accounting scopes influence regional responsibility and reduction strategies. Empirical results indicate that while China’s aquatic trade and processing have steadily developed, the sector remains dominated by low-value-added primary products. This structure highlights vast potential for deep processing development amidst shifting global dietary habits. Factor decomposition reveals that economic and technological development are the primary drivers of carbon emissions. Notably, technological progress within fisheries emerges as the most significant factor, playing a pivotal role in both driving and potentially mitigating emissions. Consequently, to effectively lower carbon intensity, the study concludes that restructuring the fishery industry is crucial. Promoting low-carbon development and enhancing the R&D of green technologies are essential strategies to navigate the dual challenges of industrial upgrading and environmental protection. Full article

Accurate short-lead river discharge forecasting is critical for effective flood risk mitigation in small mountainous basins, where rapid hydrological responses pose significant challenges. In this study, we focus on the Fuhu Stream in Emeishan City, China, and utilize high-resolution 5-min time series of upstream precipitation, stage, and discharge to predict downstream flow. We benchmark three data-driven models—SARIMAX, XGBoost, and LSTM—using a dataset spanning from 7 June 2024 to 25 October 2024. The data were split chronologically, with observations from October 2024 reserved exclusively for testing to ensure rigorous out-of-sample evaluation. Lagged correlation analysis was employed to estimate travel times from upstream to the basin outlet and to inform the selection of time-lagged input features for model training. Results during the test period demonstrate that the LSTM model significantly outperformed both XGBoost and SARIMAX across all regression metrics: it achieved the highest coefficient of determination (R² = 0.994) and the lowest prediction errors (RMSE = 0.016, MAE = 0.011). XGBoost exhibited moderate performance, while SARIMAX showed a tendency toward mean reversion and failed to capture low-flow variability. Accuracy evaluation reveals that LSTM accurately reproduced both baseflow conditions and multiple flood peaks, whereas XGBoost and SARIMAX failed. These results highlight the advantage of sequence-learning architectures in modeling nonlinear hydrological propagation and memory effects in short-term discharge dynamics. Feature importance analysis indicates that the LSTM model was highly effective for real-time forecasting and that the WSQ/LY sites served as critical monitoring nodes for achieving reliable predictions. This research contributes to the operationalization of early warning systems and provides support for decision-making regarding downstream flood disaster prevention. Full article

Providing effective post-disaster housing remains a globally complex challenge shaped by interrelated constraints, including environmental sustainability, socio-cultural compatibility, logistical capacity, and economic feasibility. Contemporary responses therefore require housing solutions that extend beyond rapid deployment to incorporate flexibility, adaptability, and long-term spatial transformation. In this context, this study advances a design-oriented, computational framework that positions parametric design at the core of post-disaster housing production within the broader digital transformation of the construction sector. The research proposes an adaptive parametric–modular housing system in which standardized architectural units are governed by a rule-based aggregation logic capable of generating context-responsive spatial configurations across multiple scales and typologies. The methodology integrates a qualitative synthesis of global post-disaster housing literature with a quantitative computational workflow developed in Grasshopper for Rhinoceros 3D (version 8). Algorithmic scripting defines a standardized spatial grid and parametrically regulates key building components structural systems, façade assemblies, and site-specific environmental parameters, enabling real-time configuration, customization, and optimization of housing units in response to diverse user needs and varying climatic, social, and economic conditions while maintaining constructability. The applicability of the framework is examined through a case study of the Düzce Permanent Housing context, where limitations of existing post-disaster stock, such as spatial rigidity, restricted growth capacity, and fragmented public-space integration, are contrasted with alternative settlement scenarios generated by the proposed system. The findings demonstrate that the framework supports multi-scalar and multi-typological reconstruction, extending beyond individual dwellings to include public, service, and open-space components. Overall, the study contributes a transferable computational methodology that integrates modular standardization with configurational diversity and user-driven adaptability, offering a sustainable pathway for transforming temporary post-disaster shelters into permanent, resilient, and socially integrated community assets. Full article

Effective disaster management depends on rapidly understanding earthquake damage, yet traditional methods struggle to operate at scale and rely on expert inspections that become difficult when access is limited or time is critical. Satellite-based damage detection also faces limitations, particularly under adverse weather conditions and delays associated with satellite overpass schedules. This study introduces a machine learning-based approach to assess post-earthquake building damage using real observations collected after the event. The aim is to develop fast and reliable estimation techniques that can be deployed immediately after the mainshock by integrating structural, seismic, and geographic data. Three machine learning models—Random Forest, Histogram Gradient Boosting, and Bagging Classifier—are evaluated across both reinforced concrete and masonry buildings and across multiple spatial levels, including building, district, and city scales. Damage is categorized using practical three-class (traffic light) and detailed four-class systems. The models generally perform better in simpler classifications, with the Bagging Classifier offering the most consistent results across different scales. Although detecting severely damaged buildings remains challenging in some cases, the three-class system proves especially effective for supporting rapid decision-making during emergency response. Overall, this study demonstrates how machine learning can provide faster, scalable, and practical earthquake damage assessments that benefit emergency teams and urban planners. Full article

The unique natural environment and climate of tropical island regions present significant challenges to construction. Under these variable natural conditions and complex construction processes, identifying and analyzing potential risks that could lead to vulnerabilities in construction safety systems and clarifying their transmission pathways remains a pressing issue. To fill this research gap, a GV-IB model for vulnerability analysis of construction safety systems in tropical island building projects (CSSTIBPs) was established. This model constructs a vulnerability analysis index system for tropical island construction safety systems based on the Grey Relational Analysis (GRA) and Vulnerability Scoping Diagram (VSD), considering exposure, sensitivity, and adaptability. By combining the artificial fish swarm algorithm with the K2 algorithm and the EM algorithm, an Improved Bayesian Network (IBN) is constructed to analyze and infer the influencing factors and disaster chains of vulnerability in tropical island construction safety systems. The IBN can effectively overcome the dependence on node order and data gaps in traditional Bayesian Network construction methods. The effectiveness of the model is verified by analyzing Hainan Island, China. The research results show that (a) The IBN stability verification showed an Area Under ROC Curve (AUC) of 0.783 > 0.7, indicating high effectiveness in identifying vulnerability factors. (b) Within the vulnerability measurement nodes of the CSSTIBPs, the influence on the system decreases in the following order is exposure (0.41), sensitivity (0.31), and adaptability (0.03). (c) Emergency response time, safety training, hazard identification time, accident response time, and duration of severe weather are key factors affecting the vulnerability of CSSTIBPs. Full article

Against the backdrop of intensifying global climate change and expanding human encroachment into mountainous regions, landslides have increased markedly in both frequency and destructiveness, emerging as a key risk to socio-ecological security and development in mountain areas. Rigorous assessment and forward-looking prediction of landslide disaster vulnerability (LDV) are essential for targeted disaster risk reduction and regional sustainability. However, existing studies largely center on landslide susceptibility or risk, often overlooking the dynamic evolution of adaptive capacity within affected systems and its nonlinear responses across temporal and spatial scales, thereby obscuring the complex mechanisms underpinning LDV. To address this gap, we examine Yunnan Province, a landslide-prone region of China where intensified extreme rainfall and the expansion of human activities in recent years have exacerbated landslide risk. Drawing on the vulnerability scoping diagram (VSD), we construct an exposure–sensitivity–adaptive capacity assessment framework to characterize the spatiotemporal distribution of LDV during 2000–2020. We further develop a multi-model, multi-scale integrated prediction framework, benchmarking the predictive performance of four machine learning algorithms—backpropagation neural network (BPNN), support vector machine (SVM), random forest (RF), and XGBoost—across sample sizes ranging from 2500 to 360,000 to identify the optimal model–scale combination. From 2000 to 2020, LDV in Yunnan declined overall, exhibiting a spatial pattern of “higher in the northwest and lower in the southeast.” High-LDV areas decreased markedly, and sustained enhancement of adaptive capacity was the primary driver of the decline. At approximately the 90,000-cell grid scale, XGBoost performed best, robustly reproducing the observed spatiotemporal evolution and projecting continued declines in LDV during 2030–2050, albeit with decelerating improvement; low-LDV zones show phased fluctuations of “expansion followed by contraction”, whereas high-LDV zones continue to contract northwestward. The proposed multi-model, multi-scale fusion framework enhances the accuracy and robustness of LDV prediction, provides a scientific basis for precise disaster risk reduction strategies and resource optimization in Yunnan, and offers a quantitative reference for resilience building and policy design in analogous regions worldwide. Full article

The Ulva prolifera green tide in the South Yellow Sea has erupted annually for many years, posing significant threats to coastal ecology, the economy, and society. While environmental factors are widely acknowledged as prerequisites for these outbreaks, the asynchrony and complex coupling between their variations and disaster events have challenged traditional studies that rely on instantaneous correlations to uncover the underlying dynamic mechanisms. This study focuses on the Ulva prolifera disaster in the South Yellow Sea, systematically analyzing its spatiotemporal distribution patterns, the temporal accumulation and lag effects of environmental factors, and the coupled driving mechanisms using the Floating Algae Index (FAI). The results indicate that: (1) The disaster shows significant interannual variability, with 2019 experiencing the most severe outbreak. Monthly, the disaster begins offshore of Jiangsu in May, moves northward and peaks in June, expands northward with reduced scale in July, and largely dissipates in August. Years with large-scale outbreaks exhibit higher distribution frequency and broader spatial extent. (2) Environmental factors demonstrate significant accumulation and lag effects on Ulva prolifera disasters, with a mixed temporal mode of both accumulation and lag effects being dominant. Temporal parameters vary across different factors—nutrients generally have longer lag times, while light and temperature factors show longer accumulation times. These parameters change dynamically across disaster stages and display a clear inshore–offshore gradient, with shorter effects in coastal areas and longer durations in offshore waters, revealing significant spatiotemporal heterogeneity in temporal response patterns. (3) The driving mechanism of Ulva prolifera disasters follows a “nutrient-dominated, temporally relayed” pattern. Nutrient accumulation (PO₄, NO₃, SI) from the previous autumn and winter serves as the decisive factor, explaining 86.8% of interannual variation in disaster scale and 56.1% of the variation in first outbreak timing. Light and heat conditions play a secondary modulating role. A clear temporal relay occurs through three distinct stages: the initial outbreak triggered by nutrients, the peak outbreak governed by light–temperature–nutrient synergy, and the system decline characterized by the dissipation of all driving forces. These findings provide a mechanistic basis for developing predictive models and targeted control strategies. Full article

Unmanned helicopters play a critical role in various fields including defense, disaster response, and infrastructure inspection. Military platforms such as the MQ-8C Fire Scout represent high-value assets exceeding $40 million per unit including development costs, particularly when compared to expendable multicopter drones costing approximately $500–2000 per unit. Unexpected failures of these high-value assets can lead to substantial economic losses and mission failures, making the implementation of Health and Usage Monitoring Systems (HUMS) essential. However, the scarcity of failure data in unmanned helicopters presents significant challenges for HUMS development, while the economic feasibility of investing resources comparable to manned helicopter programs remains questionable. This study presents a novel cross-scale transfer learning framework for vibration-based anomaly detection in unmanned helicopters. The framework successfully transfers knowledge from a source domain (Airbus large manned helicopter) using publicly available data to a target domain (Stanford small RC helicopter), achieving excellent anomaly detection performance without labeled target domain data. The approach consists of three key processes. First, we developed a multi-task learning transformer model achieving an F-β score of 0.963 (β = 0.3) using only Airbus vibration data. Second, we applied CORAL (Correlation Alignment) domain adaptation techniques to reduce the distribution discrepancy between source and target domains by 79.7%. Third, we developed a Control Effort Score (CES) based on control input data as a proxy labeling metric for 20 flight maneuvers in the target domain, achieving a Spearman correlation coefficient ρ of 0.903 between the CES and the Anomaly Index measured by the transfer-learned model. This represents a 95.5% improvement compared to the non-transfer learning baseline of 0.462. Full article