Simulation of atmospheric radiocesium (¹³⁷Cs) from Fukushima nuclear accident using FLEXPART-WRF driven by ERA5 reanalysis data

Nuclear Science and Technology, Vol.10, No. 3 (2020), pp. 01-12

Simulation of atmospheric radiocesium (¹³⁷Cs) from Fukushima

nuclear accident using FLEXPART-WRF driven by ERA5

reanalysis data

Kieu Ngoc Dung^1,*, Nguyen Hao Quang², Hoang Huu Duc³, Nguyen Thị Hang¹,

Nguyen Thi Thoa³, Nguyen Quang Trung⁴

¹National Committee for Search and Rescue of Vietnam, Long Bien, Ha Noi, Vietnam

²Vietnam Atomic Energy Institute, 59 Ly Thuong Kiet, Ha Noi, Vietnam

³Military Environmental Chemistry Institute, An Khanh, Hoai Duc, Ha Noi, Vietnam

⁴Vietnam Institute of Meteorology, Hydrology and Climate change, Ha Noi, Vietnam

*Corresponding author, E-mail: kieudung.nbc@gmail.com

(Received 19 August 2020, accepted 18 November 2020)

Abstract: This study investigates short-range atmospheric transport of radiocesium (¹³⁷Cs) after

Fukushima nuclear accident using the Weather Research and Forecasting (WRF) model and the

Lagrangian particle dispersion FLEXPART-WRF model. The most up-to-date ERA5 reanalysis

dataset is used as initial and boundary condition for the WRF model for every hour. Four experiments

were carried out to examine the sensitivity of simulation results to micro-physics parameterizations in

the WRF model with two configured domains of 5 km and 1 km horizontal resolution. Compared with

observation at Futaba and Naraha station, all experiments reproduce reasonably the variation of ¹³⁷Cs

concentration from 11/03 to 26/03/2011. Statistical verification as shown in Taylor diagrams

highlights noticeable sensitivity of simulation results to different micro-physics choices. Three

configurations of the WRF model are also recommended for further study based on their better

performance among all.

Keywords: ¹³⁷Cs dispersion, Fukushima Daiichi nuclear power plant, FLEXPART–WRF model,

ERA5 reanalysis data, Futaba, Naraha.

The FDNPP consists of six units that

were strongly impacted by the earthquake and

tsunami, leading to a serious nuclear accident,

radioactive substances were released from the

plant area and released into soil, water and air

environments (Fig. 2). The most serious is that

radioactive materials are released into the air,

they will be spreaded under different weather

conditions and can be fell in continent and sea

areas that is very far from the accident site.

I. INTRODUCITON

The massive earthquake in Japan occurred

at 14:46 JST on 11/03/2011, with a magnitude of

9.0 [1] that caused heavy damage to infrastructure

along the east coast. It was followed by the

inundations of tsunami that caused power outages

and flooding in a large residential and industrial

area. This event had a major impact on five

nuclear power plants along Japan's northeast coast,

Higashidori, Onagawa, Fukushima Daiichi,

Fukushima Daini and Tokai Daini. Fig. 1 shows

the epicenter of the earthquake was far from

Fukushima Daiichi nuclear power plant (FDNPP)

180km in the northeast and Onagawa NPP 130km

in the east [2,3].

A number of computational models have

been used to study radioactive contamination

in the vicinity of the FDNPP from the

15/03/2011 [4], the regional simulation [5-8],

and global scale simulation [9-11]. The results

SIMULATION OF ATMOSPHERIC RADIOCESIUM (¹³⁷Cs) FROM FUKUSHIMA NUCLEAR ACCIDENT…

of these models are quite consistent in

reproducing high radioactive deposition at the

center of Fukushima prefecture. However,

these models have not yet accurately assessed

the radioactive matter deposition in the vicinity

of the factory, especially at radioactive

monitoring stations within a radius of 10 km.

Especially, the European Regional

Weather Forecast Center (ECMWF) provides

high resolution global forecasts with a

frequency of twice a day at 00 UTC and 12

UTC,

used

innovative

4D-Var

data

assimilation system with 91 different pressure

levels [16]. Recently, the ECMWF created a

new ERA5 reanalysis data with horizontal

resolution of 31 km and 137 different pressure

levels. In addition, the land surface and ocean

surface data are provided, including

Currently, the rapid development of

numerical dynamical weather model as well

as of particle dispersion models allows

high-resolution simulation of atmospheric

radionuclides [12]. An important factor for

these simulation is global meteorological

datasets driven the regional models.

precipitation, temperature at

atmospheric radiation [16].

2

m

and

In this study, the most up-to-date ERA5

reanalysis dataset will be used as initial and

boundary condition for a numerical weather

prediction model. Then, simulation output of

this weather model will force a particle

dispersion model, in order to simulate the

transport of radiocesium (¹³⁷Cs) after the

Fukushima nuclear accident. In addition,

different experiments are carried out to

evaluate the sensitivity of simulation results to

physics choices of the atmospheric model. The

sensitivity can be evaluated by discrepancies

among outputs of experiments, as analyzed in

the following sections.

Implementing

Multiscale

Global

model,

Environmental

the Canadian

Meteorological Center (CMC) provides

analysis data with horizontal resolution of

approximate 33 km and at 80 pressure levels

[13].

The

National

Center

for

Environmental Prediction (NCEP) runs the

global data assimilation system (GDAS)

four times per day (i.e. 00, 06, 12, and 18

UTC) to provide prediction issues [14]. The

UK Met Office gives a forecast data 6

hourly at horizontal resolution of 25km with

70 pressure levels [15].

(a)

(b)

Fig. 1. (a) The epicenter of the massive earthquake in Japan on 11/03/2011 [2] and (b) the satellite image of

the FDNPP site [3]

2

KIEU NGOC DUNG et al.

solver which implements flux-form equations

II. EXPERIMENT DESIGN

with variables that have conservation

properties and a terrain-following mass vertical

coordinate, is used in this study (Fig. 2). The

LPDM applied in this study is FLEXPART-

WRF model which is modified version of the

FLEXPART model to works with the WRF

model [12, 18]. The FLEXPART-WRF model

differs from its preceding versions in that it has

novel turbulence scheme for the convective

boundary layer [12].

In this study, the WRF model and a

Lagrangian particle dispersion model (LPDM)

are used to simulate short-range atmospheric

transport of radionuclides (i.e. ¹³⁷Cs). Figure 2

depicts the process to carry out experiments, in

form of a flow-chart. The WRF model is a

numerical dynamical atmospheric simulation

model, governed by compressible and non-

hydrostatic Euler equations [17]. The

Advanced Research WRF (ARW) dynamics

Fig. 2. Flow-chart of simulation processes in this study, implemented the WRF-ARW atmospheric model and

the FLEXPART-WRF dispersion model

(a)

(b)

Fig. 3. Simulation domains of the WRF model (a) with horizontal resolutions of 5 km (d01 – blue rectangle)

and 1 km (d02 – black rectangle) and (b) location of Futaba, Naraha station as well as Fukushima Daiichi

Nuclear Power Plant (FD1NPP)

3

SIMULATION OF ATMOSPHERIC RADIOCESIUM (¹³⁷Cs) FROM FUKUSHIMA NUCLEAR ACCIDENT…

The WRF model provides spatial and

temporal meteorological variables as forcing

for the FLEXPART-WRF to simulate the

spread of radionuclides. Two domains of the

WRF model are configured in this study with

horizontal resolutions of 05 km and 01 km

(Fig. 3.). Topography of the vicinity of the

FNPP is very complicated, with coastal lines

on the East and surrounded by mountains on

the West. The computational domain with a

high-resolution of 01 km inside the 100 km

vicinity of the plant (domain d02) and 05 km

for the outside of 100km region (domain d01)

is expected to reasonably capture the spead of

radiocesium plumes. The WRF model runs

with 51 vertical levels of the atmosphere and

04 soil layers of 10, 30, 60, 100 cm thick. The

ERA5 reanalysis data is used as initial and

boundary conditions for the WRF model with

hourly update timestep. Simulation time is

from 21:00 UTC on March 11, 2011 to 01:00

UTC on March 26, 2011.

to calculate the soil temperature and moisture

content in soil layers, taking into account snow

cover and freezing processes in the soil [19].

Yonsei University (YSU) planetary boundary

layer scheme [20], Dudhia short-wave

radiation scheme [21] and Rapid Radiative

Transfer Model (RRTM) long-wave radiation

scheme [22] are used in this research.

Radiation schemes are updated with time steps

of 5 minutes and 1 minute for 05 km and 01

km-resolution domains, respectively.

The sensitivity of the atmospheric

radionuclide simulation to different micro-

physics options will be investigated with

four experiments (Table I). Different

microphysical options will yield different

moisture variables, depending on the phase

transitions and interactions of water and ice

particles. The Kessler scheme is suitable for

a warm cloud consisting of water vapor,

water droplets and raindrops, and the other

processes including the generation, fall and

evaporation of raindrops [23]. The WSM 6-

class scheme includes snowfall and other

related processes and the phase transition of

ice [24]. This scheme is suitable for dealing

with grids that contain clouds while other

processes are similar to WSM 3-class

scheme [25]. Thompson’s scheme assumed

snow size distribution depends on both

temperature and ice water content and is

represented as a sum of exponential and

gamma distributions [26].

Physical processes are parameterized in

the WRF model include (1) micro-physics, (2)

cumulus parameterization, (3) planetary

boundary layer, (4) land surface model, and (5)

radiation

processes.

The

cumulus

parameterization is only valid for coarse grid

resolution (i.e. greater than 10 km) as

convective assumptions will be violated for

finer resolution. Therefore, in this study, the

convection parameterization schemes are not

activated. The Noah surface scheme was used

Table I. List of experiments in this study

Experiment

name

No.

Microphysics

Description

Reference

1

Exp 1

Kessler scheme

Warm rain (i.e. no ice or

idealized case)

Kessler (1969)

2

3

4

Exp 2

Exp 3

Exp 4

WRF Single-Moment (WSM)

3-class

Simple ice (3 arrays)

Hong et al., (2004)

WSM 6-class graupel scheme

Cloud scale, single moment (6

arrays, graupel)

Hong and Lim

(2006)

Thompson graupel scheme

Double moment (8-13 arrays)

Thompson et al.,

(2004)

4

KIEU NGOC DUNG et al.

In order to run a dispersion model, a

after the Fukushima accident are used [28]. In

addition to comparison on time-series plot and

map of concentrations, the Taylor diagram are

used to compare simulation results [29]. This

specific emission source is required to

indentify. The source term of the ¹³⁷Cs

radioactive nuclide released from the reactor

area by time and position is determined based

on the analysis report of Katata et al., (2015)

[27]. To evaluate simulation results, the time-

series analysis of atmospheric radiocesium at

two monitoring sites (i.e. Futuba and Naraha)

diagram provides

a

concise statistical

verification of how well simulation match

observation, in terms of Pearson’s correlation

coefficient, root-mean-square difference, and

the ratio of standard deviations [29].

10¹⁵

10¹⁴

10¹³

10¹²

10¹¹

KATATA

0

50

100 150 200 250 300 350 400 450 500

Time,h

Fig. 4. The source term of ¹³⁷Cs over the time after the accident,

retrieved from [27]

of wind barbs in Fig. 5a is thinned with factor

III. RESULTS AND DISCUSSIONS

A. Simulation of meteorological conditions

of 10. It means that the WRF model can

provide much more details of atmospheric

circulation over study area.

With coarse resolution of approximately

31 km, the ERA5 reanalysis data can not

reproduce meteorological variables over

complex terrain of Japanese region. The WRF

model can downscale dynamically to finner

mesh resolutions (i.e. 05 km and 01 km in this

study). Fig. 5 shows that the WRF model can

maintain well the spatial pattern of geo-

potential height over Japanese region from the

forcing ERA5 data. The high pressure system

located at the North of Japan as well as the

pressure gradient followed Northwest-

Southeast axis are reproduced well in the WRF

model (Fig. 5). It’s worthy to note that, for

facilitating “eyeball” verification, the number

Because precipitation is an important

factor for the wet deposition of radioactive

material in plumes, simulated precipitation

should be compared with observed data and

other previous published data. Fig. 6 presents

accumulated simulated precipitation from the

WRF model in Exp 1, from 09:00 to 15:00 on

March 15, 2011. Rainfall amount and intensity

in this case is highly similar to simulation

results from Katata et al., (2015) [27]. Heavy

rain, from 5 mm to 10 mm per 6 hours,

occurred over broad area around Fukushima

region. Due to the impacts of earthquake and

tsunami, almost all the meteorological

5

SIMULATION OF ATMOSPHERIC RADIOCESIUM (¹³⁷Cs) FROM FUKUSHIMA NUCLEAR ACCIDENT…

observation equipments were inoperable after

the nuclear accident. Therefore, it’s difficult to

obtain good quality meteorological observation

in this case. Large-scale meteorological

information during the occurrence of

radioactive material emissions into the air was

presented in the 2013 World Meteorological

Organization (WMO) report by previous

research groups such as Kinoshita et al., (2011)

[30], Stohl et al., (2012) [10] and Sugiyama et

al., (2012) [31]. In fact, rain occured over the

north area of Fukushima prefecture from 17:00

JST March 15 to 04:00 JST on the March 16

[30]. On the 20 to 22 March, sustainable low

pressure caused moderate rainfall in the

vicinity of Tokyo.

(a)

(b)

Fig. 5. Simulation of geo-potential height (shaded color) and wind field (barbs) on level of 850 mb, at 12h00

UTC 15/03/2011 from the WRF model in Exp 1 (a), in comparison with the ERA5 reanalysis data (b).

Notice: Factor of 10 is applied to thin the number of wind barbs in (a)

Fig. 6. Accumulated simulated precipitation from the WRF model in experiment Exp 1, from 09:00 to 15:00

on March 15, 2011

6

KIEU NGOC DUNG et al.

resolution. In this paper, high resolution of

B. Evaluation of atmospheric radiocesium

01 km is very suitable for considering the

geographical location of Futaba station, as

well as other neighboring stations (e.g.

Naraha station). Calculation results of the

concentration of atmospheric radiocesium

¹³⁷Cs for every hour at Futaba and Naraha

station are displayed in Fig. 7 and Fig. 8,

respectively. The observation data displayed

in these figures are retrieved from Tsuruta et

al., (2011) [28].

The Futaba station, an observatory

station in the town of Futaba, is very close to

the Fukushima NPP. The distance between

Futaba town and the plant is only about 3.2

km, where was severely affected by both

earthquakes, tsunamis and the effects of

radiation [28]. For the other researches of

global radioactive dispersions, the vicinity

areas of the plant are often not taken into

account, because of the limitation of the grid

Fig. 7. Hourly accumulated concentration of ¹³⁷Cs at Futaba station from observation (red dashed line) and

Exp 4 (blue solid line) with range of simulation results from all experiment (shaded light blue)

From Fig. 7 and Fig. 8, it can be seen

that simulation results have a good agreement

with the observed data, especially from 12 to

14/03/2011 at Futaba station and from 15 to

16/03/2011 at Naraha station. Peak values of

¹³⁷Cs concentration on 12 and 19/03/2011 at

Futaba station are reproduced well in all

experiments. Peak values on 15, 16 and

19/03/2011 at Naraha station are also captured

well by the FLEXPART-WRF model. The

range of simulated values from four

experiment (i.e. shaded light blue area in

Fig.7 and Fig. 8) can be recorgnized,

especially for concentrations of less than 10²

Bq.m^-3per hour. The uncertainty in simulation

or the sensitivity of calculation results to

different micro-physics option is more clear in

the case of Futaba station than in Naraha

station. This can be seen on simulated range

of 13-14/03/2011 and 19-21/03/2011 in Fig.

7. The Exp 4 was displayed due its better

performance, in comparison with others

experiments, which is confirmed by statistical

verification shown in Fig. 9.

7

SIMULATION OF ATMOSPHERIC RADIOCESIUM (¹³⁷Cs) FROM FUKUSHIMA NUCLEAR ACCIDENT…

Fig. 8. Hourly accumulated concentration of ¹³⁷Cs at Naraha station from observation (red dashed line) and

Exp 4 (blue solid line) with range of simulation results from all experiment (shaded light blue)

(a)

(b)

Fig. 9. Taylor diagram compares simulation results from 04 experiments using Pearson correlation

coefficient and Normalized Standard Deviation for (a) Futaba and (b) Naraha station. Observation value is

depicted by black star

Fig. 9 demonstrates high sensitivity of

simulation results to different micro-physics

options of the WRF model. Scatter of

experiment’s points on the Taylor diagram

highlights the significant variations of not

only correlation coefficients (CC) but also

standard deviations (σ) of simulated

atmospheric radiocesium retrieved from

four experiments. For example, at Futaba

station, simulation result from experiment

Exp 1 has CC value of 0.28 and normalized

σ of 0.48. While respective verification

metrics for Exp 4 are 0.77 and 0.36 which

means better capture of hourly observed

8

KIEU NGOC DUNG et al.

release of ¹³⁷Cs air concentration. At Naraha

maps explain the peak of concentration

shown in Fig. 7 and Fig. 8. At level of 100

m, atmospheric radionuclide propagated to

the North on 12/03/2011 which plumes

station, the higher CC values can be seen, in

comparison with simulation results at

station Futaba (i.e. 0.92 for Exp 4 or 0.89

for Exp 1). Based on this Taylor diagram,

spreaded

widely

to

Southwest

on

the experiment Exp

3

show worse

15/03/2011. Smaller plumes in both area

and intensity blowed along coastal line to

the South are simulated on 19/03/2011.

These results show a similarity to the results

of Tsuyoshi et al., (2015) [1] in which

different horizontal grid resolutions are used

to calculate radioactivity concentration on

15/03/2011.

simulation results than Exp 1, Exp 2 and

Exp 4. Therefore, the configuration of Exp

1, Exp 2 or Exp 4 can be recommended for

further study in the future.

From Fig. 10, dispersion plume of

137Cs concentration at 100 m can be seen

for three different days. These distrubition

(a)

(b)

(c)

Fig. 10. Local-scale spatial distributions of accumulated concentrations of 137Cs at 100 meter from Exp 4

retrieved (a) from 00 UTC 12 to 00 UTC 13/03/2011, (b) from 00 UTC 15 to 00 UTC 16/03/2011 and (c)

from 00 UTC 19 to 00 UTC 20/03/2011. Unit: Bq.m-3

configured with two domains of 05 km and 01

km. Both meteorological conditions and

dispersion of atmospheric radiocesium (¹³⁷Cs)

are evaluated. In comparison with observation

at Futaba and Naraha station, all experiments

captured reasonably the variation of ¹³⁷Cs

concentration from 11/03 to 26/03/2011.

Analysis on Taylor diagram confirm the

noticeable sensitivity of simulation results to

four selected micro-physics parameterizations.

The configurations of Exp 1, Exp 2 and Exp 4

IV. CONCLUSIONS

This study investigates short-range

atmospheric transport of radionuclides after

Fukushima nuclear accident using a numerical

weather model and a Lagrangian particle

dispersion model. Four different experiments

were carried out using the FLEXPART-WRF

model coupled with the WRF model. The

ERA5 reanalysis data is used as initial and

boundary conditions for the WRF model with

hourly update time step. The WRF model is

9

SIMULATION OF ATMOSPHERIC RADIOCESIUM (¹³⁷Cs) FROM FUKUSHIMA NUCLEAR ACCIDENT…

[5]. Morino et al., “Atmospheric behavior,

deposition, and budget of radioactive materials

from the Fukushima Daiichi nuclear power

plant in March 2011”, Geophysical Research

Letters, VOL. 38, L00G11, 2011.

are recommended for further study due to their

better performance among all.

ACKNOWLEDGEMENT

This article is supported by the project

titled “Study the effects of the floating nuclear

power plant on the sea and the nuclear power

plants on Hainan Island on Vietnam’s marine

environment by modelling and developing

response plans for a nuclear accident occurred

over the sea”, project code: KC.AT, in the

framework of technical research program,

nuclear safety to ensure combat readiness for

the Army in the 2016-2020 period. This article

is also supported by the project titled “Study

and calculating the spread of atmospheric

radionuclides and modernization of stations

for radioactive analysis” (PACT-1).

[6]. Yasunari et al., “Cesium-137 deposition and

contamination of Japanese soils due to the

Fukushima

nuclear

accident”,

PNAS

December 6, 108 (49) 19530-19534, 2011.

[7]. Katata et al., “Atmospheric discharge and

dispersion of radionuclides during the Fukushima

Dai-ichi Nuclear Power Plant accident. Part I:

Source term estimation and local-scale

atmospheric dispersion in early phase of the

accident”,

Journal

of

Environmental

Radioactivity 109, 103-113, 2012.

[8]. Le Petit et al., “Analysis of radionuclide

releases from the Fukushima Dai-ichi nuclear

power plant accident part I”, Pure Appl.

Geophys. , 2012.

REFERENCES

[9]. Takemura et al., “A Numerical Simulation of

Global Transport of Atmospheric Particles

Emitted from the Fukushima Daiichi Nuclear

Power Plant” , SOLA, Vol. 7, 101−104,

doi:10.2151/sola.2011-026, 2011.

[1]. Tsuyoshi T. Sekiyama, Masaru Kunii, Mizuo

Kajino, and Toshiki Shimbori, “Horizontal

Resolution Dependence of Atmospheric

Simulations of the Fukushima Nuclear Accident

Using 15-km, 3-km, and 500-m Grid Models”,

Journal of the Meteorological Society of Japan,

Vol. 93, No. 1, pp. 49−64, 2015.

[10].Stohl et al., 2012, “Xenon-133 and caesium-

137 releases into the atmosphere from the

Fukushima Dai-ichi nuclear power plant:

determination of the source term, atmospheric

dispersion, and deposition” , Atmos. Chem.

Phys., 12, 2313–2343, 2012.

[2]. Japan Meteorological Agency, “Information on

the 2011 off the Pacific Coast of Tohoku

Earthquake”, 2015.

[11].Christoudias and Lelieveld, “Modelling the

global atmospheric transport and deposition of

radionuclides from the Fukushima Dai-ichi

nuclear accident”, Atmos. Chem. Phys., 13,

1425–1438, 2013.

[3]. IAEA Library Cataloguing in Publication

Data, “The Fukushima Daiichi accident,

Technical Volume 1: Description and Context

of the accident”, IAEA, 2015.

[4]. Chino et al., , “Preliminary Estimation of

Release Amounts of ¹³¹I and ¹³⁷Cs

Accidentally Discharged from the Fukushima

Daiichi Nuclear Power Plant into the

Atmosphere”, Journal of Nuclear Science and

Technology, 48:7, 1129-1134, 2011.

[12].Brioude, Jerome, Delia Arnold, Andreas Stohl,

Massimo Cassiani, Don Morton, P. Seibert, W.

Angevine et al., “The Lagrangian particle

dispersion model FLEXPART-WRF version

3.1.”, Geoscientific Model Development, 6.6:

1889-1904, 2013.

10

KIEU NGOC DUNG et al.

[13].Charron et al., “The stratospheric extension of

experiment using a mesoscale two-dimensional

model", Journal of the atmospheric sciences

46.20: 3077-3107, 1989.

the Canadian global deterministic medium-

range weather forecasting system and its

impact on tropospheric forecasts” , Mon. Wea.

Rev., 140 (2012), pp. 1924-1944, 2012.

[22].Mlawer, Eli J., et al., "Radiative transfer for

inhomogeneous atmospheres: RRTM,

a

validated correlated-k model for the longwave

(Paper 97JD00237)", Journal of Geophysical

Researche-All Series-102: 16-663, 1997.

[14].Kanamitsu

et

al.,

“Recent

changes

implemented into the global forecast system at

NMC” , Wea. Forecasting, 6 (1991), pp. 425-

435, 1991.

[23].Kessler, Edwin. “On the distribution and

continuity of water substance in atmospheric

[15].Davies et al., “A new dynamical core for the

Met Office's global and regional modelling of

the atmosphere” , Q. J. R. Meteorol. Soc., 131

(2005), pp. 1759-1782, 2005.

circulations.”

pp.

1-84.

American

Meteorological Society, Boston, MA, 1969.

[24].Hong, S.-Y., and J.-O. J. Lim. “The WRF

Single-Moment 6-Class Microphysics Scheme

(WSM6)”, J. Korean Meteor. Soc., 42, 129–

151, 2006.

[16].Hersbach, Hans, Bill Bell, Paul Berrisford,

Shoji Hirahara, András Horányi, Joaquín

Muñoz‐Sabater, Julien Nicolas et al. “The

ERA5 global reanalysis.” Quarterly Journal of

the Royal Meteorological Society 146, no. 730,

1999-2049, 2020.

[25].Hong, S.-Y., J. Dudhia, and S.-H. Chen. “A

Revised Approach to Ice Microphysical

Processes for the Bulk Parameterization of

Clouds and Precipitation”, Mon. Wea. Rev.,

132, 103–120, 2004.

[17].Skamarock, William C., Joseph B. Klemp, Jimy

Dudhia, David O. Gill, Dale M. Barker, Michael

G. Duda, Xiang-Yu Huang, Wei Wang, and

Jordan G. Powers., “A description of the

Advanced Research WRF version 3.” In NCAR

Tech. Note NCAR/TN-475+ STR. 2008.

[26].Thompson, G., R. M. Rasmussen, and K.

Manning. “Explicit forecasts of winter

precipitation using an improved bulk

microphysics scheme. Part I: Description and

sensitivity analysis”. Mon. Wea. Rev., 132,

519–542, 2004.

[18].Stohl, Andreas, C. Forster, A. Frank, P.

Seibert, and G. Wotawa. “Technical note: The

Lagrangian

particle

dispersion

model

[27].Katata, G., et al., “Detailed source term

estimation of the atmospheric release for the

Fukushima Daiichi Nuclear Power Station

accident by coupling simulations of an

atmospheric dispersion model with an

improved deposition scheme and oceanic

dispersion model”, Atmospheric Chemistry &

Physics 15.2, 2015.

FLEXPART version 6.2.”, Atmos. Chem.

Phys. Discuss., 62 pages, 2005.

[19].Ek, M. B., et al., "Implementation of Noah land

surface model advances in the National Centers

for Environmental Prediction operational

mesoscale Eta model", Journal of Geophysical

Research: Atmospheres 108.D22, 2003.

[28].Tsuruta, Haruo, Yasuji Oura, Mitsuru Ebihara,

Yuichi Moriguchi, Toshimasa Ohara, and

Teruyuki Nakajima. “Time-series analysis of

atmospheric radiocesium at two SPM

monitoring sites near the Fukushima Daiichi

Nuclear Power Plant just after the Fukushima

accident on March 11, 2011.” Geochemical

Journal 52, no. 2, 103-121, 2018.

[20].Hong, Song-You, Yign Noh, and Jimy Dudhia.

“A new vertical diffusion package with an

explicit treatment of entrainment processes.”

Monthly weather review 134, no. 9, 2318-

2341, 2006.

[21].Dudhia, Jimy, "Numerical study of convection

observed during the winter monsoon

11

SIMULATION OF ATMOSPHERIC RADIOCESIUM (¹³⁷Cs) FROM FUKUSHIMA NUCLEAR ACCIDENT…

[29].Taylor, Karl E. "Summarizing multiple aspects

of model performance in a single diagram."

Journal of Geophysical Research: Atmospheres

106, no. D7, 7183-7192, 2001.

nuclear accident covering central-east Japan”,

Proc. Natl. Acad. Sci. U. S. A., 108, pp.

19526-19529, 2011.

[31]. Sugiyama et al., “Atmospheric dispersion

modeling: challenges of the Fukushima Daiichi

response”, Health Phys., 102, pp. 493-508, 2012.

[30].Kinoshita et al., “Assessment of individual

radionuclide distributions from the Fukushima

12