Nghiên cứu và ứng dụng chương trình FLEXPART trong đánh giá phát tán phóng xạ tầm xa tt tiếng anh

MINISTRY OF EDUCATION
AND
TRAINING

MINISTRY OF SCIENCE AND

TECHNOLOGY
VIETNAM ATOMIC ENERGY
INSTITUTE

PHAM KIM LONG

RESEARCH AND
APPLICATION OF
FLEXPART IN THE
LONG-RANGE
ATMOSPHERIC
DISPERSION OF
RADIONUCLIDES

Major: Nuclear and Atomic
Physics
Code: 9.44.01.06

SUMMARY OF


DOCTORAL
DISSERTATION OF
PHYSICS

Hanoi - 2019


The thesis has been completed at:
Nuclear Training Center, Vietnam Atomic Energy Institute.
140 Nguyen Tuan, Thanh Xuan, Hanoi.
Supervisors:
1. Prof. Dr. PHAM Duy Hien
Vietnam Atomic Energy Institute
2. Dr. NGUYEN Hao Quang
Vietnam Atomic Energy Institute
Referee 1:
Prof. Dr. LE Hong Khiem
Institute of Physics, Vietnam Academy of Science
and Technology
Referee 2:
Assoc. Prof. Dr. NGUYEN Tuan Khai
Vietnam Agency for Radiation and Nuclear Safety

The dissertation has been defended against the Institute-level
Doctoral Dissertation Defense of Nuclear Training Center,
Vietnam Atomic Energy Institute. At 14:00 on March 4,
2019

The thesis can be found at:
- National Library of Vietnam
- Library of Nuclear Training Center


OVERVIEW

1. The reason for choosing the thesis topic
Lessons learnt from the Chernobyl disaster in 1986 or the
Fukushima nuclear accident in 2011 show the importance of
environmental radioactivity monitoring, simulation, calculation and
evaluation of atmospheric dispersion of radionuclides from nuclear
power plants (NPP) in supporting the emergency preparedness in
response to nuclear reactor accidents. Among the NPPs near our
country, the Fangchenggang NPP is located near the border of our
country less than 50 km, about 250 km away from Hanoi capital. We
need to build an Environmental Radiation Warning and Monitoring
Network to continuously monitor artificial and natural radiation
levels, combined with simulation of atmospheric dispersion of
radionuclides in supporting the emergency preparedness.
Due to these urgent requirements, I chose the thesis topic
“Research and application of FLEXPART in the long-range
atmospheric dispersion of radionuclides” with my desire to
contribute a small part in the field of monitoring, warning and
responding to environmental radiation incidents. The aim of this
thesis is to provide a suitable model for evaluation of the long-range
atmospheric dispersion of radionuclides in supporting the emergency
preparedness.
2. Purpose of this study
To learn mathematical models and simulation programs
suitable for long-range dispersion of radioactivity.

1


To learn meteorological models that meet the requirements of
simulation programs and meteorological data analysis tools.

To validate of the atmospheric dispersion model through the
Fukushima nuclear power plant accident.
To apply of the model for assessing the impart of atmospheric
transport of radioactivity from China's nuclear power plants.
3. Research scope, research objective and research methods
The Lagrangian particle dispersion model FLEXPART.
FLEXPART meteorological input data: Climate Forecast
System version 2 (CFSv2), Global Forecast System (GFS). Panoply
software for meteorological data analysis.
High Performance Computing (PARAM-HUST) to run the
simulation using FLEXPART.
Statistical evaluation methods to verify compatibility between
simulation and observation.
Atmospheric transport of 131I and 137Cs from Fukushima
accident to tropical western Pacific (TWP) and Southeast Asia
(SEA).
Application of FLEXPART for assessing the impart of
atmospheric transport of radioactivity from Fangchenggang nuclear
power plant.
4. Significance of the Study
The thesis has studied the possibility of applying the dispersion
model used in FLEXPART to simulate the transport of radioactivity
into the atmosphere. Activity concentrations of radionuclides
measured at ten monitoring stations in TWP and SEA from

2


Fukushima accident were used to validate the particle dispersion
model. Good agreement between the FLEXPART model and

observations yields confidence regarding its application to assess
radiation impacts and support emergency planning in response to a
possible future nuclear accident in the region.
5. The layout of the thesis
The thesis consists of 100 pages of content, 18 tables, 57
figures, 03 published works (02 articles and 01 national nuclear
conference), 79 references, 6 annexes, to be allocated as follows:
Overview: Introducing the reasons for choosing the thesis
topic, purpose, objective, scope and methods of research,
significance of the thesis (4 pages); Chapter 1: Overview of nuclear
power plants in East Asia, atmospheric dispersion model and
meteorological characteristics (30 pages); Chapter 2: Evaluation
methods

in

atmospheric

dispersion

of

radionuclides

using

FLEXPART (37 pages); Chapter 3: Results and Discussion (25
pages); Conclusion and recommendations for future research (3
pages); Finally, the list of publications related to the thesis,
references, and annexes.

3


CHAPTER 1. INTRODUCTION
Vietnam is located in Southeast Asia adjacent to East Asia,
There are currently 4 countries and territories with commercially
operating nuclear power plants, including: China, Taiwan, Japan and
South Korea (Fig. 1.1). A total of 48 commercially operating nuclear
power plants with 157 nuclear reactors [1], of which 116 units are in
operation, 20 units are under construction, 21 units have been
deactivated, and many units are in construction plans.

Fig. 1.1. Map of the nuclear power plants in East Asia
157 units of NPPs in East Asia connected the first time to the
grid by year as shown in Fig. 1.2. The technology and age of the
reactors are calculated at the end of 2018 in this area as shown in
Fig. 1.3 and 1.4 [1].

4


Fig. 1.2. Number of units connected to the grid by year in East Asia

Fig. 1.3. Nuclear power technology in East Asia

Fig. 1.4. The age of reactor types of operating NPPs in East Asia

5

Within a distance of 1000 km from our country border, 18 units
are operating and 4 units are under construction belong to China [1].
The units include generations of II, II +, III and III + reactors. In
particular, the Fangchenggang NPP is located in Quangxi, China.
With a distance less than 50 km from our country border, about
250 km away from Hanoi capital (Fig. 1.5). A total of six reactors
are planned to operate at Fangchenggang NPP (2 units are in
operation, 2 units are under construction, and 2 units are in the
construction plan) [2].

Fig. 1.5. Satellite image of Fangchenggang NPP
(Source: Google Earth, updated on May 10, 2016)
In fact in East Asia, the accident happened at Fukushima NPP
in March 2011 [4]. A huge amount of radioactive material was
released into the atmosphere and dispersed across the northern
hemisphere. Japan raised the disaster at Fukushima Daiichi to Level
7 on the INES scale [10].

6


As we all know when the nuclear accident occurs, an
inevitable byproduct of nuclear fission is the production of fission
products which are highly radioactive released into the environment,
especially the atmosphere. Radioactive isotopes are very useful in
environmental research to assess the extent of accidents affecting the
environment and people.
In such an urgent situation, we need to build an Environmental
Radiation Warning and Monitoring Network to continuously monitor

artificial and natural radiation levels, combined with simulation of
atmospheric dispersion of radionuclides in supporting the emergency

preparedness in response to nuclear reactor accidents.
We use the Lagrangian particle dispersion model FLEXPART
(Stohl et al., 1998, 2005) to simulate atmospheric transport of
radionuclides from nuclear power plants. FLEXPART can be used
for calculating the long-range and mesoscale dispersion of air
pollutants including dry deposition, wet deposition, radioactive
decay from a point source, line source, or area sources. FLEXPART
is a Lagrangian transport and dispersion model, compared with
Gaussian and Eulerian models, this is a model that is widely used in
simulation to evaluate the dispersion of pollutants into the
atmosphere. In addition, large-scale atmospheric circulation features
of atmospheric circulation are also being studied in order to better
explain the atmospheric dispersion of radioactivity.

7


CHAPTER 2. EVALUATION METHODS IN ATMOSPHERIC
DISPERSION OF RADIONUCLIDES USING FLEXPART
2.1. FLEXPART transport and dispersion model
This content delves into the theoretical basis and physical
methods used in FLEXPART to describe the process of radioactivity
released into the atmosphere, the process of removing such as wet
deposition, dry deposition or radioactive decay. Function of
calculating radioactive concentration.
2.2. Modelling the dispersion of radionuclides using FLEXPART
In this work, we use the Lagrangian particle dispersion model

FLEXPART (Stohl et al., 1998, 2005) to simulate the atmospheric
transport of radionuclides from NPPs. The block diagram is shown in
Fig. 2.1.

Fig. 2.1. The block diagram of FLEXPART

8


2.3. High performance computing
In this study, the simulations were designed to run on the
PARAM-HUST supercomputer at the Centre for High-Performance
Computing, Hanoi University of Science and Technology.
2.4. Source term estimation of the atmospheric release
There are two main contents in this study:
1) Validation of the atmospheric dispersion model through the
Fukushima nuclear power plant accident. The source term has been
published by the many studies, such as Chino (2011), Stohl (2012), Terada
(2012), and Katata (2015) [5, 34, 35, 37]. In this study, the source term for
131

I and 137Cs were used by referring to the study of

Katata et al. (2015) [27] and WMO report (2013) [33].
2) Application of FLEXPART for assessing the impart of
atmospheric transport of radioactivity from Fangchenggang nuclear
power plant with regional characteristics of meteorological. The
source term is assumed to correspond to level 7 of the INES scale.
2.5. Meteorological input data
2.5.1. Meteorological input data for FLEXPART

In this study, CFSv2 model of NCEP [55] will be used
primarily as input meteorological data for FLEXPART in the
simulation of radionuclides released from Fukushima.
2.5.2. Collecting meteorological data
CFSv2 meteorological data is collected at the address:
" />
9


2.5.3. Meteorological data analysis
NASA's Panoply was used to analyze meteorological data such
as pressure, temperature, wind speed and wind direction. Thereby
statistical analysis of meteorological data to see the main effects of
atmospheric circulation affecting the dispersion of radionuclides in
the atmosphere. In addition, other meteorological data analysis
software is used, such as NCL, WGRIB.
2.6. FLEXPART postprocessing tools
The output format of FLEXPART is binary files, in this study
using Quicklook developed by Radek Hofman to create regional and
global concentrations maps. In addition, the output data processing
software for FLEXPART such as Quickdose, Pflexible, and
Reflexible was developed by FLEXPART‘s community.
2.7. Statistical evaluation methods
Procedures

for

evaluating

atmospheric

transport

and

disperision model calculations have a long history [42], [67]–[71]. In
the following evaluation, a new ranking method (Draxler, 2006) [72]
was defined by giving equal weight to the normalized (0 to 1) sum of
the correlation coefficient (R), the fractional bias (FB), the figure-ofmerit in space (FMS), and the Kolmogorov-Smirnov parameter
(KSP).

10


CHAPTER 3. RESULTS AND DISCUSSION
3.1. Atmospheric transport of radionuclides from Fukushima to
Southeast Asia
3.1.2. Design in simulation of radionuclides dispersed from
Fukushima to TWP and SEA
In this work, we use the Lagrangian particle dispersion model
FLEXPART (Stohl et al., 1998, 2005) combined with the
meteorological data provided by NCEP (with a spatial resolution of
0.5°) (Saha et al., 2014), to simulate atmospheric transport of

131

I và

137

Cs from FNPP to TWP and SEA. The results are evaluated with
observational data from the region. We assess the contribution of the
northeast monsoon and its associated meteorological conditions to
regional transport of Fukushima-derived radionuclides because this
could inform the potential impacts of radiological emissions from
other NPPs in northeast Asia (NEA).
For the emission scenarios of 131I and 137Cs, source terms from
Katata et al. (2015) were adapted for this simulation (Fig. 3.2). There
were two sets of particle median diameter (d p) and particle density
(ρp), namely, dp = 0.6 μm, ρp = 2500 kg/m3 (Arnold et al., 2015;
Geng et al., 2017) in the simulation I, and d p = 0.4 μm, ρp = 2500
kg/m3 in the simulation II. The simulations were designed to run on
the PARAM-HUST supercomputer.

11


Fig. 3.1. The locations of radiation monitoring stations operating
during the FNPP accident in TWP and SEA (red points). The FNPP
is marked with a star in yellow.

Fig. 3.2. The emission rates of 131I and 137Cs during the FNPP
accident according to Katata et al. (2015).

12


Table 3.1. FNPP-derived 131I observed at monitoring stations in TWP
and SEA.

Station

Fukuoka

Okinawa

Nankang

Guam

Hong
Kong
Manila

Hanoi

Dalat

HCMC

KL


3.1.2. Atmospheric transport of radioactive material from
Fukushima to TWP and SEA
The meteorological conditions during the FNPP accident (e.g.,
mean sea level pressure, wind speed and wind direction) in East Asia
are shown in Fig. 3.4. The modeled FNPP radioactive plumes are
shown in Fig. 3.6a–d and 3.10a-d in terms of mean activity

13


concentrations in the atmospheric columns from 0 to 2 km and 2–10
km. These represent the planetary boundary and troposphere layers,
respectively.

Fig. 3.4. The meteorological conditions (mean sea level pressure
(Pa), wind speed (m/s) and wind direction) driven the first regional
plumes on March 18 and the second regional plume on April 4
3.1.2.1. Hermispherical transport
After the first release of radioactivity at FNPP on March 12,
the radioactive cloud was transported towards the Pacific Ocean,
where it was captured by the extra-tropical low-pressure system
located over the Bering Sea (the Aleutian Low). The plume was
lifted up to the troposphere and then rapidly transported
northeastward via the jet stream (Fig. 3.9 and 3.10). Cyclonic and
frontal systems following the jet stream led to vertical mixing
resulting in surface detection of radionuclides across the northern
hemisphere (Mészáros et al., 2016). The hemispherical plume
progressively arrived from North America (16–18/3) to the north
Atlantic (19/3), Scandinavia (22/3), western Russia (23/3), NEA and
then the western Pacific (30–31/3) (Fig. 3.9).

14


Fig. 3.9. Mean airborne concentration (in Bq/m3) of 131I in the 0–20
km layer.
3.1.2.2. Regional transport

The first FNPP regional plume was transported towards TWP
and SEA on March 18th when the south-eastward moving Siberian
anticyclone appeared over southern Japan (Fig. 3.9a), forcing the
radioactive

cloud

in

a

clockwise

direction,

then

traveled

southwestward by the northeast monsoon (Fig. 3.4a). The modeled
plume arrived in Guam, Philippines, Okinawa, Taiwan, and northern
Vietnam on March 20, 22, 24, 25, and 27, respectively (Fig. 3.9b).
The strong release of radioactivity on March 23 (Fig. 3.9) and the
anticyclone coming back over southern Japan on March 26 (Fig. 3.4)
transported this plume further into the equatorial western Pacific
(Fig. 3.9c), landing on southern Vietnam and Malaysia from March
28–29.

15

The second regional plume departed from Japan on
approximately April 4, when two eastward moving anticyclones
occurred over the western Pacific with one approaching Japan and
the other further east (Fig. 3.4b). The FNPP radioactive plume was
blocked from tropospheric transport for several days by the eastern
anticyclone, while the anticyclone approaching Japan forced the
plume to move in a southwest direction and subsequently traveled
under the influence of the northeast monsoon. The radioactive air
mass traveled almost entirely within the marine boundary layer (Fig.
3.9d) and was not observed at higher altitudes (Fig. 3.10d).

Fig. 3.11. 131I and 137Cs peak activity concentrations as a function of
distance from Fukushima (km) during the second regional plume
The peak activity concentration associated with the arrival of
the second regional plume decreased exponentially with distance
from FNPP and had decreased by half after a distance of 577 km for
131

I and 433 km for

of

137

137

Cs, as shown in Fig. 3.11. The faster decrease

Cs peak activity concentrations with distance indicates greater

deposition loss of 137Cs relative to
due to its greater particle size.

131

16

I during atmospheric transport


3.1.3. Comparison between modeled and observed surface activity
concentrations
131I

Figs. 3.12 and 3.13 compare the measured and modeled
and 137Cs activity concentrations, respectively, at nine monitoring
stations located in TWP and SEA. The Hong Kong station is
removed due to lack of 137Cs measurement data (Lee et al., 2012).
The arrival times of the plumes and the dates of peak activity
concentrations were predicted rather well, within an error of ± 2
days. The correlation coefficients were statistically significant for
most monitoring stations (Table 3.1). The results show good
agreement between the observed andmodeled concentrations in the
simulation I for

137

Cs (dp = 0.6 μm) and in the simulation II for

131

(dp = 0.4 μm).

Fig. 3.12. Time series of observed (blue column) and simulated 131I
surface activity concentration in the simulation I (red line) and
simulation II (green line)

17

I


3.1.5. New findings of this study
Our

research

was

published

in

December

2018

as

“Atmospheric transport of 131I and 137Cs from Fukushima by the East
Asian northeast monsoon” [79]. We found that in addition to the
hemispherical plume, there were also 2 regional plumes. The
regional transport was more important in contributing Fukushimaderived radioactivity to the regions. The radioactivity of the
hemispherical plume became depleted after a 20-day journey
circumnavigating the northern hemisphere before reaching TWP. The
FNPP radioactive level recorded in TWP and SEA would have been
much greater if the accident occurred at the same time as those
meteorological conditions that generate regional southwestward
plumes (i.e. on March 18 or April 4 instead of March 12).
3.2. Application of FLEXPART to assess the imparts of
atmospheric

transport

of

radioactivity

from

China‘s

Fangchenggang nulear power plant to Vietnam
The northeast monsoon (NEAM) dominates the weather and
climate of Vietnam in winter, it also brings to our country air
pollutants, including radioactive plumes on the windy road when a
nuclear power accident occurs. There are 3-5 monsoon winds due to
disputes and conflicts between continental cold air (Siberian high
pressure) and Pacific hot air (low pressure). Therefore, the southeast

monsoon winds blows not only in the northeast direction but also in
the southeast at the end of the phase and intermediate directions. This
makes it difficult to guess the direction of the radiation plumes. The
thesis only initially mentioned some typical scenarios in January

18


2018 when the Siberian was most active high pressure is the highest
level pressure.
The atmospheric dispersion of radionuclides released from
China’s Fangchenggang NPP to Vietnam was simulated by using the
Lagrangian particle dispersion model FLEXPART with the
meteorological data CFSv2 [5]. The input parameters for particle
size, dry deposition, wet deposition through the Fukushima study
(Long et al., 2019) were used for this simulation with level 7 of the
INES scale. Atmospheric dispersion of radionuclides was simulated
from January 01 to 31, 2018. At this time, there are two main
northeast monsoon winds coming to Vietnam. The cold front began
to appear above Fangchenggang NPP on January 8 and 28 and then
moved through it. The simulation results of the concentration of
in the atmosphere are shown in Fig. 3.18. The concentration of

131

131

I

I at

3 monitoring stations in Hai Phong, Vinh and Da Nang is shown in
Figures 3.21 and 3.22.

Fig. 3.15. The first cold front appears in January 2018 (a) and moves
to Fangchenggang NPP (b)

19


Fig 3.17. Concentration of 131I (µBq/m3) in the boundary layer
when the first cold front moved through Fangchenggang
The simulation results show the impact assessment of
atmospheric transport of radioactivity from Fangchenggang NPP via
the East Asian NE monsoon entering our country in winter. Before
cold front moved to Fangchenggang, the plume were mostly travel
towards the mainland of China due to low pressure in this area.

20


CONCLUSION
In this work, we used the particle dispersion model
FLEXPART to simulate long-range transport of

131

I and

137

Cs

derived from the Fukushima Daiichi nuclear power plant (FNPP)
accident to the tropical western Pacific (TWP) and southeast Asia
(SEA). Measurement data at ten monitoring stations located in this
region were used for validation of the model results. There was good
agreement between the model and observations, and the following
conclusion can be drawn.
The airborne radioactivity observed in this region came from
both the hemispherical transport following the jet stream and the
regional transport in the boundary layer by the East Asian northeast
monsoon. Due to the late arrivals of both hemispherical and regional
plumes, the TWP and SEA region had recorded much lower airborne
radioactivity than other regions in the northern hemisphere, which
were affected earlier by the FNPP radioactivity. The regional
transport, however, was more important in contributing Fukushimaderived radioactivity to the regions. The

131

I and

137

Cs activity

concentrations in the regional plume decreased exponentially with
distance from Fukushima and had decreased by half after a distance
of 577 km and 433 km, respectively. The faster decrease of

137

Cs

peak activity concentrations with distance indicates greater
deposition loss of 137Cs relative to 131I during atmospheric transport
that may be due to its greater particle size. Results of analysis of the
four statistical indicators also shows better agreement between

21