Improving Data on the Built Infrastructure for Disaster Management: Progress and Opportunities
Robert CHEN (USA)
Sustainable Data Infrastructures For Integrated Disaster Science Research
Sara GRAVES (USA)
Big Data Potentials for Early Warning and Crises Management
Milan KONEČNÝ (Czech Republic)
The Data Application of SJ-9 Satellite in Disaster Monitoring
Yiwei LIU (China)
A Gap Analysis of Using Social Science Data in Natural Disaster Research (Representing CODATA LODGD Task Group)
Xinlu XIE (China)
Robert S. CHEN
Former Secretary General of CODATA; Director and Senior Research Scientist, Center for International Earth Science Information Network (CIESIN), The Earth Institute, Columbia University, USA
The location, design, and condition of the human built infrastructure — including buildings, dams, reservoirs, storm drains, roads, railroads, airports, hospitals, schools, power grids, industrial facilities, and other structures — are key factors in assessing the risks of both natural and technological hazards. Unfortunately, data on infrastructure are often incomplete, inconsistent, out of date, poorly georeferenced, and/or inaccessible to those involved in disaster risk assessment or management. Progress has been made recently in a few areas, e.g., in improving global-scale data on building fragility in relationship to earthquake hazards (through the Global Earthquake Model initiative), in assembling open access data on intercity road networks (through the CODATA Global Roads Data Development Task Group and OpenStreetMap), and in making data on reservoirs, dams, and nuclear power plants more accessible (through the NASA Socioeconomic Data and Applications Center). Remote sensing data from medium-resolution instruments also have the potential to greatly improve assessment of key building characteristics as well as patterns of human settlement and infrastructure. Efforts are needed to improve the interoperability and accessibility of existing infrastructure databases, to explore new methods such as crowd sourcing for gathering and updating infrastructure data, and to better integrate infrastructure data and information with hazard, exposure, and vulnerability data in risk models and disaster management tools.
Secretary General, CODATA; Director, Information Technology and Systems Center, University of Alabama in Huntsville, USA
Many diverse types of data necessary for research in various aspects of disaster science create challenges for the international data science community. The collection, organisation, stewardship and dissemination of invaluable and unique for disaster science research require sustainable data infrastructures across the globe. International data science groups, such as ICSU’s CODATA, are integral in leading discussions and activities to address the challenges by considering the frontiers of data science and emerging data issues and trends. This presentation will describe some sustainable data and software infrastructure activities that address the demanding needs of integrated disaster science research.
Laboratory on Geoinformatics and Cartography, Department of Geography, Masaryk University, Kotlářská 2, 611 37 BRNO, Czech Republic
This talk describes some approaches and visions of Big Data (BD) and discusses certain potential new strategies in the field of early warning (EW) and crises management (CM). It is recognised that, in many respects, BD is still an open question. Nevertheless, the concept of BD is now reaching many areas of science. ”There is no rigorous definition of big data. Initially the idea was that the volume of information had grown so large that the quantity being examined no longer fit into the memory that computers use for processing, so engineers needed to revamp the tools they used for analyzing it all” (Mayer-Schönberger and Cukier, 2013).
The talk will consider and describe the main generic BD characteristics: volume, velocity, variety, variability and complexity. In addition to these various disciplines add other specifications, e.g. in cartographical research, the following characteristics are significant: accuracy, dimensionality, quality and interactivity, map scale, map dimensionality, map generalisation (Bandrova, Konecny, Yotova, 2014).
The group of specialists who accept BD as a genuinely new concept is growing and the necessity of coming to terms with BD becomes apparent when specialists from different communities work together on interdisciplinary projects. As so often, the human aspects of research collaboration become very important. The talk will also consider orientations towards new kinds of statistical methods.
There are several challenges coming with BD, including discussion about the role of spatial data infrastructures (SDIs) in new conditions, development of the geoinfostrategies in individual states or continents and the formulation of new research topics relating to BD.
The development of the BD concept also creates a new qualitative and quantitative basis for better approaches in the field of EW and CM. One of the challenges should be improvement of the approaches covering the ‘Disaster Management Cycle’ with the support of BD approaches. The talk will also briefly compare approaches to EW and CM in Europe and China and try to formulate some stimulating thoughts about the role of geoinformatics and cartography in EW and CM in the future BD environment.
Vice Director of SJ-9A Satellite and Senior Researcher, DFH Satellite Co., Ltd., China
The SJ-9 mission consists of the SJ-9A small-satellite and the SJ-9B micro-satellite and was launched in a one rocket successfully on October 14, 2012. The SJ-9A satellite adopts a panchromatic and four multispectral integration design. The good images can be simultaneously acquired in PA and MS mode at the same point, enabling fast-response for disaster monitoring. The resolution is superior to 2.5m (PA)/10m (MS) and the swath is 30km for nadir orbit height 645km. The dynamic MTF in orbit is better than 0.12. The SJ-9A satellite can work in maneuver motive imaging mode to meet the needs for disaster monitoring. The image fusion algorithm is optimised to get better image quality from the ground application system. The result shows that image is more accurate and clear than before.
China provided images in response to the floods in Oman in December, 2013, and three times in response to the floods in Brazil in April, 2014. The SJ-9A images use in urban disaster evaluations such as earthquake monitoring in Gansu province.
This paper will make a proposal for a virtual networking formation flying plan of satellite Earth observation for disaster monitoring in fast response action.
Xinlu XIE (China)
Representing the CODATA LODGD Task Group; Institute for Urban and Environment Studies, Chinese Academy of Social Science (CASS), China
Natural disasters have both natural and social attributes, so natural disaster research needs the cooperation of natural and social scientists. As to data, both social and natural scientists feel that there are gaps in the use of social science data in natural disaster research. Why are there gaps? What are the gaps? Based on a stakeholder workshop with experts and scholars from disaster management, local emergency sector, insurance industry, hydrological sector, meteorological sector, seismic sector, and social science, our analysis found gaps in the use of social science data for natural disaster research, which relate to various issues: the research paradigm, the data collection pattern, the focus of research as well as institutional and technological barriers. Many social scientists use data from the statistical yearbooks and reports, which are in large scale and may not match their precise research purposes. Context analysis is of great importance in disaster research because local knowledge and cultures influence people’s behaviours and perception of disaster risks. Some social scientists may do fieldwork, questionnaires and interviews with stakeholders in their research. Data quality is influenced by many factors, such as whether the questionnaire is well designed, whether the interviewers are trained, whether the interviewees are willing to tell the truth and so on. The costs of data collection are high. Social science data are scattered in reports, newspapers, websites and other sources, which are not collected and organised in a systematic way for data mining. Statistical data, such as population, infrastructure, GDP, houses, are collected in large administrative scales and hard to distribute in small spatial scales due to the imbalanced development of communities. Some data are not well recorded, are not digitised or are inaccessible for various reasons. There are also lack of standards of data and laws to protect intellectual property or guarantee sharing.