Characterizing Emergency Department Visits to Pediatric Hospitals After Local Disaster Declarations.

The impacts of climate-related disasters are shaped by the interaction between hazard intensity, exposure, and vulnerability. However, the influence of hazard intensity and within-country inequality on impact magnitudes remains poorly quantified. Here, we present a global multi-hazard study of over 7000 climate-related disasters reported by the Emergency Events Database from 1990 to 2020. Using subnational indicators, we show that human development drives major shifts in global exposure and impact patterns, with societal vulnerability outweighing hazard intensity in shaping impacts. Despite a declining share of global exposure over the past three decades, regions with low subnational Human Development Index scores experience disproportionately higher human losses across most disaster types. For instance, individuals in these regions face an 8.2-fold higher risk of fatality associated with storms (95% confidence interval: 2.16-23.06) compared to those in very high human development regions. Our findings also indicate that within-country inequality in human development exacerbates disaster risk in regions with low and medium levels of human development. These results underscore the critical role of human development in managing disaster risks and highlight the link between socioeconomic conditions and vulnerability to climate-related hazards.

The low level of students' understanding of disaster literacy and the lack of utilization of local history as a source of contextual learning are issues that require attention, especially in disaster-prone areas such as Palu City. This community service activity aims to strengthen students' disaster literacy through learning the local history of Palu City for 70 students of the History Education Study Program at Tadulako University. The implementation method uses a participatory approach through the stages of socialization, mentoring, discussion, practice of making educational media in the form of posters and infographics, and evaluation using questionnaires before and after the activity. Disaster literacy indicators measured include an understanding of disaster types and risks, mitigation and preparedness, the ability to understand disaster information, and the utilization of local history as a source of contextual learning. The results of the activity show an increase in disaster literacy, students' ability to understand disaster mitigation, and the utilization of the local history of Palu City as a contextual and relevant learning resource for the surrounding environment. This activity demonstrates that learning local history can be an effective tool in improving students' disaster literacy and fostering awareness and a culture of disaster preparedness within the university environment.

<sc></sc> This paper investigates the relationship between climate-related disasters and tax morale in 25 sub-Saharan African countries from 2011 to 2021. Using Afrobarometer survey data and disaster records from the International Disaster Database – Centre for Research on the Epidemiology of Disasters, we apply multilevel logistic regression and mediation analysis to assess direct and indirect relationships between five disaster types – droughts, extreme temperatures, floods, tropical storms, and wildfires – and tax morale. The results indicate heterogeneity across disaster types: floods are associated with higher tax morale, while droughts and extreme temperatures are associated with lower tax morale. The cumulative experience of multiple disasters is negatively associated with tax morale. Mediation analysis shows that disasters are associated with lower tax morale through increased economic inequality and eroded trust in public institutions, with these patterns more pronounced in rural areas and low-income countries. Further analysis shows that national disaster risk legislation in Benin, Kenya, and South Africa mitigates the negative relationship between disasters and tax morale. The findings highlight the importance of integrating climate resilience and equity considerations into tax policy, as well as strengthening public trust and addressing inequality, both necessary for sustaining tax compliance amid rising climate-related shocks.

Understanding disaster recovery is essential for effective risk management, yet recovery processes are often implicitly assumed to scale proportionally with disaster damage. This study examines disaster damage and recovery patterns in South Korea and Japan to assess whether recovery dynamics are linear, homogeneous, and comparable across contexts. Using official national statistics, we analyze annual disaster damage and recovery expenditure in South Korea from 2015 to 2024, disaggregated by hazard type and administrative region, and compare these patterns with disaster-related fatalities in Japan from 2015 to 2023. For South Korea, results show substantial interannual variability in recovery expenditure, while statistical analysis indicates a strong and approximately proportional relationship between damage and recovery at the aggregate level. Hazard-specific analysis reveals variation in recovery intensity across disaster types, while regional analysis highlights notable spatial disparities. Extreme disaster years exhibit deviations from proportional scaling, suggesting that large events can trigger additional recovery processes beyond direct asset restoration. The cross-national comparison illustrates that disaster response indicators show different temporal patterns between South Korea and Japan when aligned with national disaster management priorities. While recovery-to-damage ratios capture fiscal recovery intensity in South Korea, disaster-related fatalities reflect the human impacts emphasized in Japan. Comparisons using scaled indicators further show that temporal response patterns do not peak synchronously across the two countries, underscoring structural differences in disaster management systems. Overall, the findings suggest that disaster recovery is a context-dependent process shaped by hazard characteristics, spatial conditions, and institutional priorities, with variability emerging around an underlying proportional relationship. Meaningful international comparisons therefore require indicator choices sensitive to national frameworks rather than uniform metrics. This study provides empirical evidence supporting a more nuanced approach to analyzing and comparing disaster recovery across countries.

When large-scale disasters exhaust civilian federal agencies' assets, the U.S. Federal Emergency Management Agency (FEMA) can issue mission assignments to the Department of War (DoW) to leverage its unique, self-sustaining logistical and operational capabilities. This study characterized DoW contributions to U.S. disaster response by analyzing 4,065 FEMA mission assignments issued between 2012 and 2024. Mission assignments were analyzed by geographic location, disaster type, event timing, personnel types deployed, and associated Emergency Support Function. The DoW supported 37% of all original FEMA mission assignments during the study period. Activity peaked during the 2017 hurricane season and the COVID-19 pandemic. Tropical cyclones and hurricanes in low-lying coastal states and island territories, and biological incidents in densely populated areas, were the primary drivers for DoW assistance. The DoW primarily supported Emergency Support Functions relating to Public Works and Engineering, Logistics, and Information and Planning activities. As an essential federal partner in managing increasingly complex, large-scale disasters, a data-informed picture of the DoW's contributions is vital for strategically optimizing operational planning and resource deployments to protect the health and well-being of the American public.

People have been increasingly using social media to post messages during a natural disaster, and describe the locations of victims, damages, difficult situations, and relief resources. Many of these location descriptions are in the forms of detailed and multi-entity descriptions, such as door number addresses, road intersections, and highway exits. Currently, there is limited availability of datasets that contain these detailed location descriptions labeled in disaster-related messages. A lack of these datasets hinders the understanding of how people describe locations during disasters and the automatic extraction of these location descriptions. This paper fills this gap by providing a dataset that covers ten disasters in the United States and in five disaster types: hurricanes, floods, wildfires, tornados, and winter storms. The messages containing location descriptions are collected from the social media platform Twitter/X, and we describe the collection, labeling, and validation of this dataset. This dataset can be used for studying the ways people describe locations under disaster contexts and for training AI models to extract these important locations.

Extreme weather events are increasing in frequency and severity, disrupting health care services across the cancer care continuum. Oncology professionals and patients are directly affected by these events, yet preparedness and resilience planning specific to cancer care remain limited. This commentary synthesizes current knowledge on the impacts of extreme weather events on oncology care and draws on lessons from prior disasters to highlight opportunities for action. Three illustrative case studies regarding Hurricane Katrina, California wildfires, and Texas Winter Storm Uri highlight commonalities between different types of disasters and geographic locations, including patterns of disruption related to infrastructure damage, power outages, workforce strain, and challenges to continuity of care. Existing evidence-informed strategies to enhance resilience across clinical, institutional, and policy domains include integrating climate resilience into clinical training and practice, adapting infrastructure and operations, reviewing local climate vulnerabilities, and strengthening partnerships with community organizations. As extreme weather events increasingly threaten cancer care delivery and patient outcomes, proactive, coordinated efforts by oncology professionals and health systems are essential to maintain high-quality care in a changing environment.

Automated classification of crisis-related social media posts is widely used to support humanitarian response; however, models trained on historical disasters often degrade when applied to new events due to cross-disaster domain shift. In emerging crises, labeled data is scarce while large volumes of unlabeled content accumulate rapidly, making effective domain adaptation critical for reliable deployment. In this work, we investigate semi-supervised domain adaptation for cross-disaster tweet classification under temporally realistic transfer settings, where each target event occurs strictly later than its source event. We evaluate adaptation performance across multiple humanitarian disasters under low-data regimes (5–50 labeled examples per class), distinguishing between within-disaster and cross-disaster transfer. We compare fully supervised fine-tuning, self-training, and unsupervised domain adaptation (UDA) against a structured co-training framework that leverages dual-view source–target supervision and cross-view pseudo-label exchange. We further study a family of controlled design variations that modify individual components—such as pseudo-label selection and mixup regularization—to analyze their impact on cross-event generalization and calibration. Results show these co-training variants consistently outperform UDA alone in low-resource settings, while different pseudo-label utilization strategies exhibit distinct trade-offs across disaster types and label budgets. By providing a temporally grounded benchmark and a structured analysis of adaptation mechanisms, this work contributes empirical guidance for designing more robust cross-disaster classification systems for crisis informatics.

In recent years, virtual geographic environments have played a crucial role in enhancing public risk perception through disaster simulations. Previous studies, such as Zhu et al.1, have proposed knowledge-driven visualization frameworks, offering valuable insights into public risk perception. Building on this foundation, this paper further focuses on the automated construction of knowledge organization and cross-platform adaptive presentation to better meet the public’s needs for understanding and participation in settings without professional guidance. This framework begins with a thorough analysis of the public’s specific needs for disaster visualization, using large language model (LLM) to extract triples and construct a detailed knowledge graph containing disaster geographic information. Under this knowledge framework, we realized precise modeling of virtual scenes and context-adaptive representation based on the theory of virtual geographic environments (VGEs). Then, we designed a cross-platform data organization and dynamic scheduling algorithm to enable content presentation across diverse devices. Finally, we conducted three groups of experiments using typical disaster cases, with user cognition assessed via eye-tracking. The experiment results indicate our method supports adaptive, smooth visualization on multiple platforms effectively. Compared to traditional approaches, it significantly improves debris flow disaster information dissemination efficiency by integrating scene context, user interest, demand response and spatial intelligence, offering advantages in standardized modeling, personalization and adaptive optimization, thereby enhancing public debris flow disaster information perception.

Despite the increasing frequency and severity of natural disasters, their macroeconomic impact remains debated in the economic literature. Therefore, this study re-examines the macroeconomic costs of natural disasters by assessing their impact on both overall and sector-specific growth across four major disaster types in the short- to medium-term. Drawing on panel data of Lower Middle-Income Countries (LMCs) for the period 1980–2024, this study employed a twofold empirical strategy: the system GMM to estimate the short-term growth effects while controlling for endogeneity, and an event study approach to capture the adjustment process of economic growth following large natural disasters. The findings reveal that natural disasters, regardless of type, tend to exert short-term adverse effects on economic activity, particularly in agricultural sector. However, considerable heterogeneity exists in the medium-term recovery patterns. Due to limited resilience capacity, one-fourth of these countries remain below pre-disaster growth levels, even five years after severe events.

On the possible meteorological factors contributing to lightning-ignited wildfires in West Sichuan: A case study of MuLi wildfire in March 2019

A Technique for Constructing a Typical Disaster Knowledge Graph Based on Spatiotemporal Features

Natural disasters and human migration in the United States: Insights from automated machine learning and explainable AI.

This study investigates the claim that women are disproportionately more likely to die in disasters by reviewing existing data sources and compiling new datasets on sex-differentiated disaster fatalities in the twenty-first century. The analysis is structured by disaster type, covering geophysical, meteorological, climatological, hydrological, and biological hazards, as well as broader national-level patterns based on global databases. It examines high-impact events across these disaster types and validates sex-disaggregated fatality patterns by integrating and assessing multiple data sources and demographic proxies. The findings do not support the widely cited claim of consistently higher female mortality. Instead, sex-disaggregated data remain very limited, and the evidence is largely inconclusive, except for biological disasters where male fatalities are consistently higher. Rather than assuming disproportionate effects in advance, sex-specific patterns should be assessed empirically. The study recommends mandatory, systematic reporting of gender-disaggregated fatalities and greater attention to differences in gender-based vulnerabilities across disaster types and contexts.

This study investigates the integration of ecological literacy into Sundanese language learning through a game-based medium, ecological literacy quartet cards, designed for elementary school students. The research responds to the need for contextual, interdisciplinary education that fosters environmental awareness while strengthening language skills, in line with the Kurikulum Merdeka and Sustainable Development Goals (SDGs). The cards embed nine ecological domains (endangered animals, energy sources, types of waste, pollution, green actions, climate change, ecosystems, marine animals, and natural disasters) combined with Sundanese language literacy tasks such as sentence construction, vocabulary enrichment, and role-play. Employing a descriptive quantitative approach, the study involved 32 fifth-grade students in Lebak district, with ecological literacy measured through a 20-item multiple-choice test after the learning intervention. Results showed high achievement across most domains, with mean scores of 90.00 (SD = 6.48), and strongest performance in “Endangered Animals” and “Marine Animals,” indicating the effectiveness of visual narratives and cultural relevance in enhancing retention. Lower performance in “Types of Waste” and “Green Actions” highlights the need for more application-oriented activities to translate knowledge into behavior. The findings demonstrate that integrating ecological and language literacy through simple, low-cost, and culturally embedded game-based media can promote cognitive and affective learning outcomes. This model offers potential scalability for other regional languages and ecological themes, contributing to language preservation and environmental education

Human life, infrastructures, and environment can all suffer greatly from disasters, whether man-made or natural. The growing prevalence of Big Data Analytics (BDA) techniques and Internet of Things (IoT) technologies as well as Artificial Intelligence (AI) methods presents significant opportunities and solutions for Emergency Management (EM) authorities. These technologies provide cutting-edge supports and foundations that enable a better understanding of events. In this paper, we review relevant existing research works dealing with EM and we propose an innovative two-level classification: the first level refers to the type of disasters and the second one represents the specific emergency situations for each level-one type. This classification serves as a systematic framework to organize and structure the current state-of-the-art in this field, in particular, it summarizes the main challenges encountered and addressed in the literature. Furthermore, the discussion and the comparison of the main literature works according to this proposed classification allowed us to easier identify key and relevant trends, issues and gaps of the existing contributions. Thus, we identified and highlighted various open issues and proposed future research directions that would be very interesting to explore and investigate in order to improve different aspects related to this important and highly concerning area.

Central America and Nicaragua are frequently affected by disasters due to their physical and environmental characteristics, exposure, and vulnerability. This study begins by offering a physical-geographic characterization of Nicaragua, followed by a national-scale analysis using the EM-DAT database to examine the most intense disasters from 1930 to 2023. We then focus on a municipal-level analysis using the available DesInventar database between 1992 and 2013 to identify municipalities with the highest disaster incidence. We filled the last decade (2014–2023) with available information extracted from secondary sources such as large events reports and scientific papers. Key findings highlight the most impactful years, months, and types of disasters, such as floods, tropical cyclones, droughts, wildfires, and epidemics. Subsequently, statistical analyses are conducted, including a correlation matrix, multiple linear regression, Poisson regression, and random forest modeling. These methods help determine which socioeconomic variables best explain the occurrence of disasters by municipality throughout history. Our results are crucial for understanding future disaster risk, explaining the spatiotemporal patterns, and identifying disaster hot spots. This study provides valuable insights into disaster risk management and serves as a methodological example for countries with similar conditions—tropical regions and developing nations that often lack comprehensive baseline data.

Natural hazards such as earthquakes, floods, and tropical cyclones pose significant threats to the operation of critical infrastructure systems (CISs) in urban environments. Rapid recovery of post-disaster CISs is essential not only for mitigating immediate socio-economic impacts but also for strengthening urban resilience against future shocks. A key challenge in this recovery process is the efficient scheduling of resources to repair damaged infrastructure, a task complicated by the dynamic and uncertain post-disaster environment, the interdependencies within infrastructure networks, and the diverse priorities and demands of various stakeholders. Given the multifaceted nature of these challenges, numerous repair resource scheduling models have been developed, each incorporating distinct algorithmic strategies tailored to different disaster types and infrastructure systems. Despite a growing body of literature on optimization problems in disaster recovery, a comprehensive understanding of the variations in these models and methods remains lacking. This review aims to systematically explore and synthesize the landscape of repair resource scheduling models, highlighting model variants and their solution algorithms. In particular, it addresses the emerging challenges in post-disaster recovery, exacerbated by the coupled effects of climate change and rapid urbanization. By categorizing the variants and extensions of existing models, this study seeks to refine current frameworks and inspire the development of more comprehensive models, ultimately contributing to more informed restoration decisions and enhanced resilience of urban infrastructure systems.

Abstract This study investigates the impact of natural disasters on socioeconomic indicators across multiple countries from 2018 to 2022. The present study collected a comprehensive dataset encompassing disaster types, total deaths, affected populations, and key development indicators such as GDP, healthcare spending, and access to clean water and electricity. Through the application of Artificial Intelligence techniques combined with Python programming, the analysis identified drought as the most impactful disaster type. Furthermore, the research concurrently specified the top five countries most affected by drought: Zimbabwe, Spain, China, India, and Germany, and conducted a statistical analysis of their conditions. This study utilized Machine learning, and a time-series clustering technique called Dynamic Time Warping (DTW) to group these countries based on trends in six key indicators. The results revealed two distinct clusters, reflecting variations in development patterns and disaster resilience regardless of the countries’ income level. Despite the challenges, this research highlights the importance of time-aware data analysis in understanding disaster vulnerability and guiding targeted policy interventions. Additionally, the findings provide data-driven structure for disasters comparison and response capabilities.