Bayesian modeling to assess populated areas impacted by radiation from Fukushima
Abstract
Citizen-led movements producing spatio-temporal big data are increasingly important sources of information about populations that are impacted by natural disasters. Citizen science can be used to fill gaps in disaster monitoring data, in addition to inferring human exposure and vulnerability to extreme environmental impacts. As a response to the 2011 release of radiation from Fukushima, Japan, the Safecast project began collecting open radiation data which grew to be a global dataset of over 70 million measurements to date. This dataset is spatially distributed primarily where humans are located and demonstrates abnormal patterns of population movements as a result of the disaster. Previous work has demonstrated that Safecast is highly correlated in comparison to government radiation observations. However, there is still a scientific need to understand the geostatistical variability of Safecast data and to assess how reliable the data are over space and time. The Bayesian hierarchical approach can be used to model the spatial distribution of datasets and flexibly integrate new flows of data without losing previous information. This enables an understanding of uncertainty in the spatio-temporal data to inform decision makers on areas of high levels of radiation where populations are located. Citizen science data can be scientifically evaluated and used as a critical source of information about populations that are impacted by a disaster.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2017
- Bibcode:
- 2017AGUFMIN53A0080H
- Keywords:
-
- 1640 Remote sensing;
- GLOBAL CHANGE;
- 1910 Data assimilation;
- integration and fusion;
- INFORMATICS;
- 4323 Human impact;
- NATURAL HAZARDS;
- 4329 Sustainable development;
- NATURAL HAZARDS