Abstract: Disaster management and the health sector ought to be natural allies, but their different origins, culture, and priorities of the various agencies tasked with disaster response mean that communication and coordination between them is often lacking, leading to delayed, sub-standard, or inappropriate care for disaster victims. The potential of the new e-health technologies, such as the electronic health record, telehealth and mobile health, that are revolutionizing non-disaster healthcare, is also not being realised. These circumstances have led to an international project to develop a disaster e-health framework for the objectives of intelligent adaption to changing scenarios, presentation and management of information, and communication and collaboration. In this paper, we describe characteristics of disaster e-health support systems to achieve such set of objectives.

Abstract: Several studies on Twitter usage during disasters analyze tweets collected using ad-hoc keywords pre-defined by researchers. While recent efforts have worked to improve this methodology, open questions remain about which keywords can be used to uncover tweets contributing to situational awareness (SA) and the quality of tweets returned using different terms. Herein we consider a novel methodology for uncovering relevant keywords one can use to search for tweets containing situational awareness. We provide a description of the methodology and initial results which suggest our approach may lead to better keywords to use for keyword searching during disasters.

Abstract: This paper presents a domain-level user, task and context analysis for response technology, based on sociological studies of disaster and disaster response. The analysis examines three dimensions of disaster-scale, kind and anticipability-that have been linked to differences in response characteristics, including differences in individual and organizational responders and behaviors. It yields a number of implications for design, and reveals five domain-specific design requirements. It also offers systematic characterizations of users, tasks and contexts of response technology, that begin to structure the user interface design space, laying a foundation for a theory of design. This work provides a conceptual basis to help researchers and designers recognize and address possible limitations of design decisions, leading, ultimately, to more usable and effective response technologies.

Abstract: Information on large-scale disaster areas, like the location of affected civilians, is highly valuable for disaster relief efforts. This information can be collected by an Aerial Monitoring System, using UAVs to detect smart mobile devices carried by civilians. State-of-the-art systems typically rely on a purely passive detection approach. In this paper, we present a cooperative communication system between UAVs and ground-based devices to improve the detection performance of such an Aerial Monitoring System. We provide different approaches for the cooperative information collection and evaluate them in a simulated inner-city scenario. The results highlight the effectiveness of the cooperative system, being able to detect civilian devices in the disaster area faster and more comprehensively than a non-cooperative approach.

Abstract: Social media platforms play critical roles in information dissemination, communication and co-ordination during different phases of natural disasters as it is crucial to know the type of crisis information being disseminated and user concerns. Large-scale Twitter data from hurricanes Irma (Sept. 2017) and Michael (Oct. 2018) are used here to understand the topic dynamics over time by applying the Dynamic Topic Model, followed by a comparative analyses of the differences in such dynamics for these two hurricane scenarios. We performed a spatio-temporal analyses of user activities with reference to the hurricane center location and wind speed. The findings of spatio-temporal analyses show that differences in hurricane path and the affected regions influence user participation and social media activity. Besides, topic dynamics reveals that situational awareness, disruptions, relief action are among the patterns common for both hurricanes; unlike topics such as hurricane evacuation and political situation that are scenario dependent.

Abstract: Ethical, legal and social issues (ELSI) are widely recognised as important in IT innovation for crisis response and management. However, attention often struggles to get beyond theorising basic concepts, when the realities of how difficulties and opportunities manifest are complex and practical. Unless these realities are understood, solutions to ELSI will remain at the surface, missing opportunities to responsibly and creatively leverage the potential of IT in disaster response. This workshop brings together narratives of lived experiences of ethical, legal, and social issues encountered in the context of IT innovation in disaster response, and analyses of normative, policy and regulatory backgrounds. In this editorial, we motivate this turn to narrative, summarise the contributions that will be presented on the day, and set out some key questions.

Abstract: In China, storm surge disasters cause severe damages in coastal regions. One of the most important tasks is to predict affected regions and their relative damage levels to support decision-making. This study develops a two-stage retrieval model to search the most similar past disaster case to complete prediction. Based on spatial attributes of cases, the top-ranking past cases with a similar location to the target case are selected. Among these past cases, the most similar past case is selected by disaster attribute similarities. Three typical storm surge case studies have been used and implemented into this proposed model and the results show that all the most affected regions can be predicted. The proposed model simplifies the prediction process and updates results quickly. This study provides useful information for the government to make real-time response plans.

Abstract: This paper summarises the contribution of EURATOM research to off-site emergency management in Europe over the past two decades. Effort initially focused on the development of methods and software that could be used to underpin the nature and extent of emergency management arrangements and policy. With time, and partially in response to accidents at TMI and Chernobyl, effort shifted to the development of a comprehensive decision support system that could find broad use in real time across Europe in order to better inform decisions on emergency management. The deployment of the developed system across Europe, largely so far at a pre-operational level, is described together with the opportunities this offers for more coherent response to any accident that may in future affect Europe and for better use of scarce resources, both human and otherwise. Indications are given of where further effort or initiatives should be directed with a view to ensuring that the major research achievements are fully and effectively exploited.

Abstract: Evacuation models can be used to determine evacuation capacity, by estimating the time required for evacuating populations to leave areas exposed to a hazard. Disaster management practices and evacuation modeling are generally carried out to prepare for ?worst-case? conditions. However, hazard severity is highly variable. Performing evacuation modeling for multiple hazard scenarios may provide flexibility and a comprehensive understanding of evacuation capacity. A case study was undertaken to analyze the merit of scenario-based evacuation modeling. Results demonstrate a difference in clearance time between maximum and historic tsunami scenario modeling. During a smaller-scale event, allowing the maximum scenario population to evacuate can add congestion and inhibit evacuation of at-risk populations. Managing evacuation can improve evacuation efficiency by preventing unneeded congestion. Results show that traditional worst-case-scenario modeling may lead to overestimation of time needed to evacuate. Planning under such a scenario may increase risk to smaller-scale hazards.

Abstract: International complex humanitarian emergencies present numerous challenges to aid organizations trying to manage data, information and knowledge about the situation or event. Humanitarian aid organizations should be able to identify what critical information they need, where to find it, what are the major gaps, and how best to share, present and disseminate this information. These challenges can be addressed through improved knowledge management. The faster and more efficiently humanitarian aid organizations are able to identify, collect, distill, analyze and manage the vast corpus of what they need to know, the more effectively they can plan for and respond to natural disasters and complex emergencies and the more lives are potentially saved.

Abstract: Social media plays a vital role as a communication source during large-scale disasters. The unstructured and informal nature of such short individual posts makes it difficult to extract useful information, often due to a lack of additional context. The potential of social media threads– sequences of posts– has not been explored as a source of adding context and more information to the initiating post. In this research, we explored Twitter threads as an information source and developed an information extraction model capable of extracting relevant information from threads posted during disasters. We used a crowdsourcing platform to determine whether a thread adds more information to the initial tweet and defined disaster-related information present in these threads into six themes– event reporting, location, time, intensity, casualty and damage reports, and help calls. For these themes, we created the respective thematic lexicons from WordNet. Moreover, we developed and compared four information extraction models trained on GloVe, word2vec, bag-of-words, and thematic bag-of-words to extract and summarize the most critical information from the threads. Our results reveal that 70 percent of all threads add information to the initiating post for various disaster-related themes. Furthermore, the thematic bag-of-words information extraction model outperforms the other algorithms and models for preserving the highest number of disaster-related themes.

Abstract: As of today, investments into early warning systems are, to a large extent, politically motivated and “disaster-driven”. This means that investments tend to increase significantly if a disaster strikes, but are often quickly reduced in the following disaster-free years. Such investment patterns make the continuous operation, maintenance and development of the early warning infrastructure a challenging task and may lead to sub-optimal investment decisions. The paper presented here proposes an economic assessment model for the tangible economic impact of early warning systems. The model places a focus on the false alert problematic and goes beyond previous approaches by incorporating some socio-cultural factors (qualitatively estimated as of now). By doing so, it supports policymakers (but also private investors) in their investment decisions related to early warning applications.

Abstract: The lack of availability of large and diverse labeled datasets is one of the most critical issues in the use of machine learning in disaster prevention. Natural disasters are rare occurrences, which makes it difficult to collect sufficient disaster data for training machine learning models. The imbalance between disaster and non-disaster data affects the performance of machine learning algorithms. This study proposes a generative adversarial network (GAN)- based data augmentation, which generates realistic synthesized disaster data to expand the disaster dataset. The effect of the proposed augmentation was validated in the roof damage rate classification task, which improved the recall score by 11.4% on average for classes with small raw data and a high ratio of conventional augmentations such as rotation of image, and the overall recall score improved by 3.9%.