A Multiscale Bayesian Data Integration
Approach for Mapping Radionuclide
Contamination
Haruko Murakami Wainwright, Masahiko Okumura
and Kimiaki Saito
Abstract This chapter presents a multiscale data fusion method to estimate the
spatial distribution of radiation dose rates at regional scale around the Fukushima
Daiichi nuclear power plant. We integrate various types of radiation measurements,
such as ground-based hand-held monitors, car-borne surveys, and airborne surveys,
all of which have different resolutions, spatial coverage, and accuracy. This method
is based on geostatistics to represent spatial heterogeneous structures, and also on
Bayesian hierarchical models to integrate multiscale, multitype datasets in a consistent manner. Although this approach is primarily data-driven, it has great flexibility, enabling it to include mechanistic models for representing radiation transport
or other complex processes and correlations. As a first demonstration, we show a
simple case study in which we integrate two datasets over Fukushima City, Japan:
(1) coarse-resolution airborne survey data covering the entire city and
(2) high-resolution ground-based car-borne data along major roads. Results show
that the method can successfully integrate two datasets in a consistent manner and
generate an integrated map of air dose rates over the domain in high resolution.
A further advantage of this method is that it can quantify estimation errors and
estimate confidence intervals, which are necessary for modeling and for robust
policy planning. In addition, evaluating correlations among different datasets provides us with various insights into the characteristics of each dataset, as well as
radionuclide transport and distribution. The resulted maps have started being used
by local governments to plan the residents’ return, and they are expected to be used
for additional policy decisions in the future such as decontamination planning.
Keywords Radiation dose rate mapping Á Bayesian hierarchical models Á
Geostatistics
H.M. Wainwright (&) Á K. Saito
Lawrence Berkeley National Laboratory, 1 Cyclotron Road, Berkeley, CA 94720, USA
e-mail: hmwainwright@lbl.gov
M. Okumura
Japan Atomic Energy Agency, 2-2-2 Uchisawai-Cho, Chiyoda, Tokyo 100-0011, Japan
© The Author(s) 2017
J. Ahn et al. (eds.), Resilience: A New Paradigm of Nuclear Safety,
DOI 10.1007/978-3-319-58768-4_5
57
Approach for Mapping Radionuclide
Contamination
Haruko Murakami Wainwright, Masahiko Okumura
and Kimiaki Saito
Abstract This chapter presents a multiscale data fusion method to estimate the
spatial distribution of radiation dose rates at regional scale around the Fukushima
Daiichi nuclear power plant. We integrate various types of radiation measurements,
such as ground-based hand-held monitors, car-borne surveys, and airborne surveys,
all of which have different resolutions, spatial coverage, and accuracy. This method
is based on geostatistics to represent spatial heterogeneous structures, and also on
Bayesian hierarchical models to integrate multiscale, multitype datasets in a consistent manner. Although this approach is primarily data-driven, it has great flexibility, enabling it to include mechanistic models for representing radiation transport
or other complex processes and correlations. As a first demonstration, we show a
simple case study in which we integrate two datasets over Fukushima City, Japan:
(1) coarse-resolution airborne survey data covering the entire city and
(2) high-resolution ground-based car-borne data along major roads. Results show
that the method can successfully integrate two datasets in a consistent manner and
generate an integrated map of air dose rates over the domain in high resolution.
A further advantage of this method is that it can quantify estimation errors and
estimate confidence intervals, which are necessary for modeling and for robust
policy planning. In addition, evaluating correlations among different datasets provides us with various insights into the characteristics of each dataset, as well as
radionuclide transport and distribution. The resulted maps have started being used
by local governments to plan the residents’ return, and they are expected to be used
for additional policy decisions in the future such as decontamination planning.
Keywords Radiation dose rate mapping Á Bayesian hierarchical models Á
Geostatistics
H.M. Wainwright (&) Á K. Saito
Lawrence Berkeley National Laboratory, 1 Cyclotron Road, Berkeley, CA 94720, USA
e-mail: hmwainwright@lbl.gov
M. Okumura
Japan Atomic Energy Agency, 2-2-2 Uchisawai-Cho, Chiyoda, Tokyo 100-0011, Japan
© The Author(s) 2017
J. Ahn et al. (eds.), Resilience: A New Paradigm of Nuclear Safety,
DOI 10.1007/978-3-319-58768-4_5
57