Scientific Analysis of Radiation Contamination at the Area around the
Fukushima-Daiichi Nuclear Power Station
Satoru TANAKA and Shinichiro KADO
School of Engineering, The University of Tokyo
The purpose of this article is to introduce some background knowledge to analyze the release of radioactive materials from the Fukushima Dai-Ichi (1F) nuclear power station (NPS) based on their inventory in the reactor core,
mechanisms of the release and the behavior of the released radionuclide from the scientific points of view. Basic
knowledge on contamination of the area and decontamination are also described.
1. Introduction
At the Fukushima Dai-Ichi (1F) nuclear power site, six nuclear power plants are located. Among them, units 1, 2, and 3
(1F1, 1F2 and 1F3 respectively, caused big accidents by earthquake and tsunami on Mar. 11, 2011. The NPSs - 1F1, - 1F2
and - 1F3 encountered the station blackout (SBO; loss of all A/C power including emergency diesel generator), back-up
battery depletion, and emergency cooling system failure, resulting in a core melt in the reactor pressure vessels (RPVs) at
1F1, 2 and 3, rupture of the suppression chamber (suspected) at 1F2, and hydrogen explosion in the reactor buildings 1F1
and 1F3. Image of these damages and the pathways of the radioactive materials are shown schematically in Fig. 1. These
events, together with the leakage of the primary containment vessels (PCVs), caused significant release of the radioactive
nuclides to the environment.
In this article we made trials to evaluate the release of radioactive materials from the NPS and their environmental behaviors based on the chemical and physical properties of the radioactive materials and based on the data analysis of the
radiation dose rate.*
Fig. 1
2. Methods of analysis
2.1 General ideas of the model
Figure 2 shows schematically the behavior of radioactive materials in the environment after its release from the reactor
facility. In order to evaluate the radiation dose to the people we must know inventory of radionuclides and chemical elements in fuel just before the accident, release from the fuel at the accident, existence states of radionuclides in RPV, PCV Fig.
and reactor building, release from the stack or reactor building, migration in the atmosphere, contamination of soil and
radiation dose from radionuclide in the soil and in the atmosphere.
There are basically two approaches to evaluate the amount of environmental release of the radionuclides. One is based
on the analysis of the physical and chemical conditions of the core fuel. In this approach, the fraction of the released
amount would be approximated. The other is based on the monitoring of the radiation dose increased due to the release
from the plant and deposition to the land.
2
2.2 Evaluation of the inventory
Amounts of radioactive nuclides, such as fission products (FPs), U, Pu and minor actinides (MAs) in the reactor fuel need
to be evaluated. Information about the chemical elements is also important in the stoichiometric estimation of the chemical
forms of released fission products. It can be calculated with the help of the ORIGEN code [1], which is based on the theory
of production and the following radioactive decay of the FPs and MAs.
2.3 Release from the fuel
A cause for release of radioactive materials at all reactors was that decay heat of fission products had not been eliminated
due to loss of the cooling function. Consequently the fuel rods were exposed to the steam and the fuel and cladding were
heated up, which resulted in generating hydrogen gas by chemical reaction between zirconium and the steam above about
900 ˚C. The reaction Zr+2H2O => ZrO2+2H2 produces hydrogen which caused subsequent hydrogen explosions. It also
produces heat because this reaction is exothermic. This heat accelerates the heating of the fuel combined with decay heat.
At high temperature uranium made eutectic compound with zirconium. Melting point of this eutectic is lower than ura- Fig.
nium oxide. Figure 3 shows high temperature phenomena of the fuel. Some radioactive materials in the fuel to be soluble
in UO2 were released following heating and melting of the fuel. A fraction of released radioactive materials from the heated
fuel depends on the vapor pressure (i.e., melting point) and diffusivity in the fuel. These behaviors are strongly temperature
dependent. A release rate constant k [min-1] as a function of the temperature T [K] is given by
*
Note that the data evaluated here have considerable ambiguities, so authors would like to suggest readers to take them as
examples for the study. Indeed, up to now, the given data from the government, TEPCO etc. has frequently updated.
1
3
k  k0 exp(Q / RT)
(1)
where Q is the activation energy [kcal/mol], and R = 0.001987 kcal/mol K the universal gas constant.
Although Q depends on the chemical species, ORNL and others proposed, in their CORSOR-O model [2], to use the
common Q of 55 kcal/mol for all species and the dependence on the species is represented by the empirically corrected k0.
For example, k0 =12,000 min-1 for Cs and Kr while 9,600 min-1(=0.8 x k0 (Cs)) for I and Te. The results are shown in Fig. 4.
Using CORSOR-O model, the fraction of inventory released from the fuel at time t is given by F 1exp(kt) . Taking
Cs as an example of calculation, the fractions of inventory released at 1,800 ˚C are: F = 90 % at t = 2 hours; and F = 100 %
at t = 4 hours.
Fig. 4
2.3 Severe accident progression analysis codes
In order to analyze or predict the severe accident progression, many computer codes have been developed. MAAP
(Modular Accident Analysis Program) was developed by U.S. industries while MELCOR has been developed by U.S.NRC
(United States Nuclear Regulatory Commission) [3].
These codes basically calculate the thermal response of the core, dealing with whole progression from the initiating event
to the radionuclide releases to the environment, which is called "source term". Therefore, initial inventory and the release
properties for each nuclide are required as input parameters. These values are usually calculated based, for example, on the
ORIGEN or CORSOR. The whole progression from the initial event includes damages in RPV and PCV and consequent
leakage of water and steam.
2.4 Atmospheric transport model
Behavior of the radioactive materials released from a NPS differs for their chemical properties and/or the weather conditions (e.g., wind direction, wind speed, rainfall, snowfall) during the accident and the geography around the plant. Noble
gases such as Kr or Xe are transported and dispersed by wind. If upward wind is predominant, the gases will be transported
to the stratosphere and delivered to the entire earth by the wind. Gases of volatile radioactive materials such as I2 are also
transported by the wind. CsI or Cs oxides can be transported by the wind if these nuclides float in the air as dust particles or
attach to the aerosols. This is called the "plume" as schematically shown in Fig. 2.
If rain or snow falls, some particles will fall out to the surface of the earth together with the raindrops (wash-out or
rain-out) and contaminate the land. Therefore, prediction of the transportation of the radionuclide plume is crucial from the
viewpoint of radiation protection.
Note, on the other hand, that relatively large/heavy particles such as fuel grains cannot transport far by the wind, so that
they tends to fall out by gravity near the NPS.
The time-integrated concentration of the released nuclides in the atmosphere, χ(x,y,z) [Bq/h], can be predicted based on
the Gaussian model developed by Pasquill's as :
(x, y, z) 

y2
z2
exp( 2 )exp( 2 )
2U y z
2 y
2 z
(2)
,
where Γ is the release rate at source [Bq/s], U the mean wind speed in the x direction [m/s] [4]. The diffusion parameters, σy
and σz, as a function of x and the air-stability (having 6 categories, A-E), represent the broadening in transverse and vertical
direction, respectively. They are larger than that deduced from molecular collisional diffusion, since the turbulent flows
enhance the net diffusion.
SPEEDI (System for Prediction of Environmental Emergency Dose Information) is the computer code in which the atmospheric transport of the specific particles can be predicted using meteorological conditions and topography based on the
basic diffusion equation above. SPEEDI predict the behavior of released radioactive materials for regional area using an
atmospheric dynamic equations model with meteorological data and topography and for local area using a mass-consistent
wind field model. Improved system (WSPEEDI) can predict in detail the process of the atmospheric dispersion of released
radioactive materials by turbulence and wet deposition by rainfall.
The behavior of radioactive materials released to the ocean is evaluated from transportation and dispersion along the
ocean current, dispersion by the tidal stream and wind, precipitation to the bottom of the sea and intake by fishes and their
migration. The compartment model is used for evaluation of the contamination in the ocean.
3. Occurrence of the accident and release, transport and washout of the radiation plume
Fig.5 shows a temporal evolution of radiation dose rate observed inside 1F site, nearby and distant cities together with the
wind conditions at 1F monitoring posts (MPs). The direction and the speed of the wind have been recorded in 16 point
compass, e.g. N, NS, N, S, NWN.... , at the same time with the radiation dose rate at the monitoring posts/car in 1F. In
order to compare the temporal evolution of the wind vector with other events, we represented the wind direction θ as the
sine and cosine components of the direction (orienting to the east be 0°, while orienting to the north be 90°). In the present
analysis, sin(θ) > 0 corresponds to the direction from the land to the ocean, while cos(θ) < 0 corresponds to the direction
to south (towards Tokyo).
2
Fig. 5
From the severe-accident analysis based on MAAP or MELCOR code, it is reported that the core damage incident for each
unit happened approximately at the period listed in Table 1.
In 1F site, the radiation dose began to increase at 5:10 on Mar. 12 (14.5 hours after scram), which coincides with the
incidents in 1F1 --- A. Vent and the following hydrogen explosion may be the cause of the increase in the radiation dose in
Minami-soma, 26 km north from 1F, on Mar. 12 --- B.
The core damage incident in 1F3 occurred on Mar.13. The venting of the PCV of 1F3 was operated several times to depressurize the RPV during Mar.13 and 14, and hydrogen explosion also occurred at 11:01 on Mar. 14 (68 hours after scram).
Fortunately, the wind directed to the East (sea direction) during the period --- C.
The incident at 1F2 caused the most serious release of the radioactive nuclides. The leakage of the PCV (suspected) caused
the release of the radioactive gasses around 21:30 on Mar. 14, which was several hours before the detection of the sound of
explosion or rupture at the suppression chamber (SC) of 1F2 (at 6:10) ---D. Note that at this moment, we have no enough
evidence to tell whether the event was the explosion or the rupture -- on Oct. 2, 2011, it has been reported that the accidents
investigation commission in Tokyo Electric Power COmpany (TEPCO) regarded from the signals recorded on a quake
mater that the hydrogen explosion might not occur.
The radioactive leakage from 1F2 could initiate the radiation plume toward the south direction, and the increase of the
radiation dose was observed as the plume propagated and passed through the locations at the speed of about 10 km/h-Fig.5, mark E. The radiation plume was observed even in Tokyo (SW 230 km south from 1F) or Sizuoka (SW 360 km). Fig.
6 shows the temporal evolution of the radiation dose observed at different locations after initiating the prominent radioactive release on Mar. 15.
[ North < 50 km from 1F ]
The plume on Mar. 15 was soon passed away and the radiation dose decreased rapidly in particular in far places. However,
plume initiated by the SC rupture propagated to the Northwest direction and causes a fallout/washout/rainout due to the
rainfalls and/or snowfalls. It contributed the significant increase in dose rate in these areas such as Iitate-mura (NW 40 km).
---- F
After Mar. 16, release of radioactive material was decreased but still continued. As a result, rainfalls over the wide area in
the South direction washed out the plume into the soils, leading to a significant increase in the radiation dose. This time the
decrease in the dose rate was dominated by the radiation decay of the radioactive nuclides.
On Mar. 18 and 19, wind directed to north direction and several peaks of the dose rate were observed in Minami-soma(N 30
km). However, presumably because there was no rainfall, these plumes did not deposit onto the ground. ---G
On Mar. 21, although rain felled in Fukushima, plume did not deposit to Minami-soma, because wind directed to South.
---- H
It suggested that the ground contamination occurred by both the plume and rainfalls.
[South 50 ~ 100 km from 1F]
Ibaraki prefecture, located South of Fukushima, underwent considerable degree of washout/rainout on 16 and 20, that can
be found from the increase of the baseline of the radiation dose, having the decay timescale of 131I, 8.02d. --- I
Just after the delivery of the plume, the decay of the short-lifetime radioactive nucleus was also observed, such as 135I
(6.7h), or 132I in radiative equilibrium with 132Te (78h) -- J
[South > 100 km from 1F]
In Tokyo, the rain on Mar. 21 washed out the plume and increase in the radiation dose, which led to a kind of small panic
when 131I was detected from the source of tap water. ---- K
In Shizuoka, at 360 km from 1F, one can see from the time difference of the rain and the increase in the dose rate, plume
arrived during the rainy weather. ---L
It suggests that the plume stay no longer than a few days.
This speculation agrees with the observation of radioactive material level of fallout in Tokyo per day [5]. Usually the
fallout last around 3-4 days in March and April.
Namely, for the purpose of self-protection from the radioactive exposure, it is preferable to watch the dose rate near one's
location and the rain for a few days after passing through the plume. Especially, the cesium deposition onto soils is caused
by the rain, while removal of the deposit cesium is difficult. Therefore, the covering of the playground with plastic sheet
before rain might have been effective as an emergency protection of the ground contamination -- even a mattress/blanket is
better than nothing. Some prefecture office and the nuclear power plant provide the real-time dose rate. It might be preferable if one could watch these data together with the rain and wind speed in the weather forecast. At the same time, it
might be required that not only the government but also the scientists provide appropriate information about how to interpret the monitored data.
4. Evaluations
4.1 Approach based on the radionuclide release analysis
Behavior of released radioactive nuclides is complicated because it is closely related to the evolution of the accident.
However, we roughly evaluated it assuming that a certain proportion of the inventory in the fuel existing one day after the
scram was released.
3
Fig. 6
Figure 7 is an illustrative image of behavior of radioactive materials in the reactor and release to the environment. Environment to the release is basically composed of two steps: release from fuel (A) and release to the environment after
release from the fuel (B). The latter release mechanism is complicated because detailed information of reactor damage
and RI behavior in the damaged reactor is not simple. Radionuclides exist in various chemical forms in RPV, PCV and
reactor building such as gas, dissolved in water, aerosol, solid particle. Entrainment of the species to the gas phase is also
important in the dynamic or boiling state.
The electric output powers, the number of fuel assemblies and the average burn-up at the scram of Unit 1(1F1), 2(1F2) and
3(1F3) of the 1F-NPS are (460 MWe, 400, 26 GWd/t), (784 MWe, 548, 23 GWd/t) and (784 MWe, 548, 22 GWd/t), respectively. Using these data, amounts of radionuclides and chemical elements for FP and MA at one day after the scram
were calculated using ORIGEN code. Figures 8 and 9 show the inventories of radionuclide and chemical elements in 1F1
at one day after the scram, respectively. Inventories in 1F2 and 1F3 at one day after the scram are about 1.5 times of 1F1.
The following nuclides were found to be important from the produced amount and half-life: 239Np (2.4d), 133Xe (5.25d),
140
La (40.3h), 141Ce (32.5d), 131I (8.04d), 137Cs (30.17y), 134Cs (2.06y) 89Sr (50.0d), 90Sr (29.1y) 132Te (78.2h), 129mTe
(33.6d), 238Pu (87.7y), 239Pu (24,000y), 241Pu (14.4y), 140Ba (12.7d), 95Zr (64d), 91Y (58.5d), 127Sb (3.9d), 99Mo (66.0h),
3
H (12.3y) and 85Kr (10.7y). Chemical inventory shown in Fig. 9 gives us important information. For example the inventory of Cs is about ten times larger than that of I.
The chemical state can typically be categorized into noble gases (Kr, Xe), volatile materials (I, Cs, Te, H) and low volatile
materials (Sr, Y, Pu). The degree of volatilization is a key to understand the release at the accident.
The chemical forms and their location of radioactive materials released from the fuel depend on their chemical properties.
Noble gases such as Kr and Xe exist in the gas phase and are released to the atmosphere by venting operation. Iodine was
released as CsI and dissolved in the water. However, some exist in the gas phase as I attached to aerosol, I2 and organic
iodine. Cs takes the chemical forms of CsOH and oxide as well as CsI in the gas phases or the water. Te exists as oxide in
the gas phase or is dissolved in the water. Sr is dissolved in the water as cation or exists as oxide in the gas phase or the
water. Therefore aerosols in the gas phase might carry these kinds of species.
In order to evaluate radionuclide release from fuel, we assumed the two cases of the temperature and the duration time
as (i) 2,800˚C and one hour, and (ii) 2,000 ˚C and four hours. We used source terms based on the inventory at one hour
after scram for 1F1, 1F2 and 1F3 for simplicity. The fraction of inventory released from the fuel is calculated by the release rate constants in as shown for these two cases in Table. 2., As can be seen from the comparison, all noble gasses and
volatile materials are released for both cases. However, the difference in the fraction is remarkable in Ba, Sr, La and Pu.
The chemical properties such as vapor pressure of each element must be taken into account on the release rate from RPV
and PCV as shown in Fig. 7. Some appropriate assumptions must be taken for estimate of released amount by the accident
except noble gases (fully released). Released radioactive materials exist in the water or the steam of RVP and PCV due to
their chemical properties and damages of RPV and PCV. Particles in the fuel generated by rapid cooling after melting
might be dispersed in the water. Some parts of the radionuclide in the gas phase are released in the accident. Some part
of the species in the liquid phase is considered to be transported to the gas phase by the entrainment. Release fraction to the
environment from the reactor is very complicated because it reflects various events causing RI release such as seat leakage from the flanges or bulbs (including SRV release to a leaking vessel), venting, hydrogen explosion, damage in the
suppression chamber. Therefore we tentatively assumed the release fraction ((B) in Fig. 7) from the RPV+PCV+Reactor
Building to atmosphere considering vapor pressure of the elements: Xe 100%; Kr 100%; Cs 1 %; I 1%; Te 0.1%; Sr 10-4;
Ba 10-4; Zr 10-4; Np 10-4; Pu 10-5; 3H 25 % etc. In this evaluation we also assumed these releases occurred at one day after
the scram by the earthquake. Although these assumptions are different from the actual accident scheme, our estimation
can give us fundamental understanding of RI release,
Amounts released to the environment can be calculated by "inventory" time "release fraction from the fuel ((A) in Fig.
7)" times "release fraction from the reactor ((B) in Fig. 7)". The calculation results of amounts released are: 3H 9.4 x 1014
Bq; 85Kr 7.6 x 1016 Bq; 89Sr 3.9 x 1014 Bq; 90Sr 3.5 x 1013 Bq; 129mTe 2.9 x 1014 Bq; 131I 6.0 x 1016 Bq ; 133Xe 1.3 x 1019 Bq;
137
Cs 7.6 x 1015 Bq; 134Cs 7.4 x 1015 Bq; 249Np 8.8 x 1013 Bq; 241Pu 1.4 x 1010 Bq; 241Am 8.9 x 107 Bq for fuel damage at
2800 ºC for one hour. The 131I equivalent amount of radionuclide is evaluated as 4.9 x 1017 Bq using the conversion factor
by INES manual.
NISA (Nuclear and Industrial Safety Agency) reported it as 7.7 x 1017 Bq calculated by MELCOR code [6].
As to the release points of radioactive materials, these are released from the venting stack in not severely damaged accident. However, in 1F accident, these were also released from the disrupted points of PCV, duct pipes and reactor building.
Moreover contaminated water was released into the sea through the tunnel of 1F2 from a crack in a concrete pit.
4.2 Approach based on the radiation monitor
4.2.1 Result of the standard method based on the SPEEDI simulation
NSC (Nuclear Safety Commission of Japan) reported the source term of 131I and 137Cs released between Mar. 12 and Apr.
5 based on the atmospheric dispersion simulation such as SPEEDI or WSPEEDI. The simulation result for the unit release
rate (1 Bq h-1) was compared with that obtained with the dust-sampler to normalize the absolute value. They obtained: 150
PBq for 131I, and 12 PBq for 137Cs (1 PBq = 1015 Bq ). In order to obtain the radiological equivalence to 131I release, the
4
Fig. 7
Fig. 8
Fig. 9
value for 137Cs was multiplied by 40, yielding the total 131I equivalent release of 630 PBq [6]. ---- These data was made
minor correction to 570 PBq (131I: 130 PBq and 137Cs: 11 PBq) on Aug. 22.
However, the results of the SPEEDI calculation have only disclosed one on Mar. 23, another one on Apr. 11, then finally
the government admit that more than 5000 evaluation results had exist from the beginning of the accident and had not be
opened being afraid of the panic.
Because these evaluations rely on the simulation code and detailed weather data, that is inaccessible by the public, we
proposed a simple but straightforward estimation from the radiation dose data, or radiation map, available at many locations.
4.2.2 Alternative method based on the ground shine
The total release of the radioactive material, integral of the source term with respect to the period of release, can be
roughly evaluated from the ground contamination after the plume passed through and being fallout/rainout/washout, based
on the following equations.
tobs tcom / j
 1
Dj (tobs )   Aj (tcom )  
CF gnd, j
(3)
 2
D̂ j (t obs )  D j (t obs ) /
D (t
j
obs
)
Fig. 10
(4)
j
 1
Aj (ts )  D̂j (tobs ) 
 2
(ts tobs )/ j
1
1
CFgnd,
j  SF
(5)
Land  Ocean
(6)
Land
2
where D and A represent the dose rate [Sv/h] and the radio activity of the surface area [Bq/m ], respectively.
CFgnd is the conversion factor from the ground contamination to the dose rate at 1m high above the ground,
[(Sv/h)/(Bq/m2)] shown in Fig. 10, while SF is the shielding factor depending on the ground condition or location or
buildings. We examined that SF=0.7 is a plausible value to be applied in the present situation(see Appendix B). τ is the half
life of the radioactivity. tcom, tobs and ts are the time when the species ratio is determined, when the dose rate was measured,
and when the radioactive species are released, respectively. Note that the subscript j is the label of the species and 131I (τ =
8.02 d), 134Cs (τ = 752.4 d) and 137Cs (τ = 11019.3 d) in the present case.
The ratio of the radioactivity A' can be determined at the specific time when all species of interest can be commonly
determined from the measurement such as dust sampling or the soil analysis or from the simulations.
We adopted the Becquerel ratio [131I]:[134Cs]:[137Cs] = 1:1:1 on tcom= Apr. 10 from the air sampling data at the Comprehensive Nuclear-Test-Ban Treaty(CTBT) National Operation System of Japan in Takasaki Gunma prefecture [7].
In this case, taking tobs = Apr. 5 for instance, the normalized dose ratio can be calculated to be 0.21: 0.57: 0.22 from the
Eq(1) followed by the normalization using Eq (2). This ratio agrees with the species-sensitive dose monitoring in
Ichihara-city, Chiba prefecture reported by Japan Chemical Analysis Center [8].
About Eq (3), we corrected data from the dose rate the places of which were categorized into the following several groups.
Si  Aj (ts )
(i) Inside 20 km no-go zone:
For inside the 20 km no-go zone, TEPCO monitored the dose rate in Mar.30-Apr. 2 and Apr.18-19, the result of which are
listed with the distance from 1F [9]. The data is shown in Fig. 11 as a function of the distance from the 1F site. Although the
scattering of the data was significant having directional dependence, the general trend exhibited decaying property.
Therefore, we performed an exponential decay fitting to determine the rough integral of the dose rate in this area based on
the following equation,


r
Dj (0) 4 r exp( )dr  4 Dj (0)L2
L
(7)
0
where, L is the characteristic length of the spatial decay while Dj(0) is the dose rate extrapolated to the origin, namely 1F
plant.
(ii) East part of Fukushima prefecture outside the 20 km no-go zone, and
(iii) West part of Fukushima prefecture:
The authority of Fukushima prefecture conducted a gamma dose rate monitoring at more than 1,600 schools all over
Fukushima except inside the 20 km circle from 1F in Apr. 4-7, 2011 [10].
In this period, the contribution of 131I to the dose rate was still obvious, so can be used to evaluate the radioactive release.
5
Fig. 11
A dose rate map of these data is shown in Fig. 12 together with the 1-dimentioal distribution projected along the latitude
direction.
It may safely be said that the Fukushima prefecture outside the 20 km no-go zone was divided into two parts, East and
West areas of which are approximately equal.
We distributed the 1370 points to (ii) and 267 points to (iii) depending on the location. As a result, one data point can be
regarded as being representing around 4.57 and 28.6 km2, respectively (area of the half disk 20 km in radius was removed
from the half area of the prefecture in (ii)).
(iv) Ibaraki prefecture:
Ibaraki prefecture, located South of Fukushima prefecture exhibits relatively higher dose rate among the adjacent prefecture. We assumed the excess dose rate of the Mito-city on Apr. 10, 0.17 µSv/h as being representing the averaged values
of Ibaragi prefecture.
(v) Tochigi, Chiba, Gunma prefectures and others:
For the minor correction of the evaluation the area of these three prefectures were regarded as representing the ground
contamination in the distant places. The excess dose rate was assumed to be 0.07 µSv/h on Apr. 10.
The largest ambiguity in evaluating the released radioactive materials is in the fraction of the ground deposition against the
total release. Atmospheric simulations also have considerable amount of error depending on the modeling.
We had first approximated the fraction was about half. Actually, ref.[11] reported from the atmospheric dispersion model
GEARN in WSPEEDI-II that these fractions for 131I and 137Cs are 0.44 and 0.46, respectively - i.e. can be suspected that the
fraction for 134Cs is also 0.46. However, ref [12], on the other hand, implies, from a three‐ dimensional chemical transport
model, Models‐ 3 Community Multi-scale Air Quality (CMAQ), that the fractions for 131I, 134Cs and 137Cs deposit on land
were 0.13, 0.22 and 0.22, respectively. We have updated the evaluation using these value range. The results were shown in
the 3rd raw in Table 3. Smaller values are the result adopting ref [12] as the fraction for the land deposition, while larger
values are those adopting ref. [11].
From these results, 131I equivalent released radioactive nucleus was evaluated to be 336-780 PBq.
The databases necessary to evaluate the ground contamination can be compiled in Appendix B. Note that the ambiguity in
137
Cs has a large effect on the 131I equivalent value, since the multiplying factor is as large as 40.
In this rough evaluation, different from NSC, we did not care about the daily change of the release rate and based on the
assumption that almost all released species had been fall-out/rain-out/wash-out and the fraction of those on the land has
been contributing to the dose rate. Ambiguity in the fraction that was deposited on land also has a direct effect on the
evaluation of the source terms. One needs to remind that it is also true for the other method based on the atmospheric
transport simulation.
4.2.3 Crosscheck of the evaluation:
The result in the previous subsection was compared to the cesium radiation map of June 14 over Fukushima and adjacent
prefectures and was presented by MEXT on Sep.30. 1732 points are located in Fukushima Prefecture[13,14].
By correcting the data to that on Mar. 15, using Eq.(1), total ground contamination scaled on Mar. 15 of Fukushima prefecture of 134Cs and 137Cs are 2.5 and 2.6 PBq, yielding source term of 32 and 470 PBq, respectively, which agree well
with the sum of our evaluation (i), (ii) and (iii).
Fall-out of 131I, 134Cs and 137Cs has been monitored at specific cities in every prefectures in Japan except Fukushima and
Miyagi[15] .We summed up the monthly fall-out in Bq/m2 in each prefecture multiplied by the area of each prefecture in
m2., yielding 0.4 PBq for 137Cs. Value roughly agree with our evaluation for the some of (iv) and (v).
Moreover, we would like to emphasize that these different methods lead to consistent results with each other for source
terms, release rates, dust sampling, fall-outs, and dose rate by ground shine with SF=0.7 (Appendix B).
4.3 Comparison between approaches
Evaluation results based on these approaches are compared in Table 4. All of them exceeded the criteria of INES accident
level 7 (>1016 Bq) . For the comparison the results for the Chernobyl accident is also listed. One may understand that the
rough evaluations, nearly hand calculations, could obtain approximated release amount of radioactive materials.
It is worthwhile to mention that the total amounts of radionuclide of Chernobyl accident was 1800 PBq for 131I, and 85
PBq for 137Cs, yielding the radiological equivalence to 131I of 5200 PBq. It is considerably larger than that of Fukushima-Daiichi accident. Reason for this fact if not only because the PCV covers the RPV in the present case, but also because
in the Chernobyl case, the massive amount of Cs, I, Sr and Pu was released to the environment by steam explosion of the
melted fuel.
5. Contamination and clean up the environment
The air dose, surface dose and radioactivity in soil have been measured around the site after the accident. The detailed
distribution maps of radiation dose are made with smaller mesh than as before. Fig. 13 presents that massive amounts of
radiation fell to the surface of the ground by snowfall after the plume, which contained radioactive materials released by
6
Fig. 14
Fig. 12
Fig. 13
rupture at 1F2, had been transported towards north-east by the wind of south-west. Although 131I with a half-life of 8 days
was predominant in an early stage, 134Cs (2.06y), 137Cs (30y) and 129mTe (33.6d) are main now. 137Cs with a half-life of 30
years will have been mainly a target of cleanup in the future.
The radioactive materials absorbed to the particles in the air are detected by the dust sampling in some points in and out
the site. 89Sr (50.5d), 90Sr (29y), 234U and 235U are detected in soil within the range of 20km from the site. The value of
these elements have the range between 1/10 – 1/10,000 of that of Cs. These elements may be an evidence of the fact that
these were released in this accident because the half-life of 89Sr is as short as 50.5 days. Moreover 140La, 95Nb and 110mAg
are detected slightly in soil towards north-west with a distance of 30km from the site. A small amount of Pu is detected
within the site and it is likely to be released in this accident. MEXT says that the difference of behavior of these elements
might make the wide range of detected value and more detailed investigation is needed.
The amount of released Sr and Pu is estimated to be extremely smaller than that of the Chernobyl accident in which
contamination by Sr and Pu was severe problem. Specifically, the highest value of 90Sr/137Cs was 8.2 % at Soma-city but
in average, it is about 0.37 % regardless of the location [16], shown in Fig. 14. Simple scaling of the radiological equivalence of averaged 90Sr release yields about 0.7 PBq from both analysis in §4.1 and §4.2.2 (=10 PBq x 0.37 % x 20), that
is much smaller than that in Chernobyl accident of 160 PBq ( 131I-eq.), where 90Sr/137Cs was 1/10.
Monitoring of radioactivity under the sea has been executed. Although concentration of radioactivity at a sampling point
within the harbor of the plant is high because highly concentrated radioactive water was released from the concrete near the
sluice gate of 1F2 (i.e., 131I 2.8 x 1015 Bq, 134Cs 9.4 x 1014 Bq, 137Cs 9.4 x1014 Bq), that of the outside plant, especially in
the area with a distance of more than 30km from the plant, is low[6].
Mechanism of soil contamination by Cs depends on the fraction of it absorbed on the outer surface of minerals or on the
layer structure of clay. While an effective method for desorption of Cs from clay has not been found yet and it will be
expected in the future, various ways for cleanup of paddy soils should be taken as for the level of contamination which are
stripping surface soil, elimination of clay particles by plow and vegetation of some plants. Effective decontamination way
for soil, which contains the way for disposal of second-waste, is indispensible for evacuated residents to return to their
home if evacuation zone will be released. For secondary waste from decontamination, temporary keeping, interim storage
and final disposal are required depending on radiation level. For determining sited for these, communication with stakeholders, people living there, local government and central Government, are very important. Academic societies should
play important roles by supplying scientific information on RI behavior and safety evaluation for storage and disposal.
6. Summary and Conclusion
We have seen that the release of the radionuclides is subject to the physical and chemical properties and composition of
the fuel core that is highly dependent on its temperature. By assuming the fraction of the release to the environment,
source term can be evaluated. The integrated source term can also be evaluated alternatively based on the radiation monitor by assuming the fraction of the land deposition, or by making use of the atmospheric simulation. Although the exact
value of the radioactive release has considerable ambiguity, the amount of the release is roughly consistent with each
other, and is considerably smaller than that in Chernobyl accident.
Atmospheric diffusion/transport mechanist of each nuclide has not fully understood yet. However, in the present situation, Cs is the most serious radionuclides to be considered and the other nuclides may have minor effect on the environment. The environmental behavior of each species are still needed to be investigated from both scientific and political
points of the view to find a better roadmap of the decontamination procedures.
Acknowledgement:
The authors wish to acknowledge Dr. Takuji Oda for providing the data based on the ORIGEN code, and Ms. Ayumi Ito
for the support of compiling data.
References:
[1] A.G. Groff, Nuclear Technology, 62 (1983) 335-352.
[2] R.A. Lorenz, M.F. Osborne, A Summary of ORNL Fission Product Release Test with Recommended Release Rates
and Diffusion Coefficients, Oak Ridge National Laboratory, May 1995 (NUREG/CR-6261).
[3] K. Vierow et al., Nuclear Engineering and Design, 234 (2004) 129-145.
[4] Frank Pasquill, "Atmospheric Diffusion", Ellis Horwood Ltd , 2nd Rev. (1974).
[5] http://monitoring.tokyo-eiken.go.jp/monitoring/f-past_data.html
[6] Nuclear Emergency Response Headquarters Government of Japan, "Report of the Japanese Government to the IAEA
Ministerial Conference on Nuclear Safety - The Accident at TEPCO’s Fukushima Nuclear Power Stations - " June
2011. (http://www.kantei.go.jp/foreign/kan/topics/201106/iaea_houkokusho_e.html)
[7] http://www.cpdnp.jp/eng/English/
[8] Japan Chemical Analysis Center, http://www.jcac.or.jp/senryoritu_kekka.html [in Japanese]
[9] http://radioactivity.mext.go.jp/ja/1100/2011/04/1305284_0421.pdf
[10] http://www.pref.fukushima.jp/j/schoolmonitamatome.pdf
[11] Hideyuki KAWAMURA, Takuya KOBAYASHI, Akiko FURUNO et al., "Preliminary Numerical Experiments on
7
Fig. 15
Oceanic Dispersion of 131I and 137Cs Discharged into the Ocean because of the Fukushima Daiichi Nuclear Power Plant
Disaster", J. Nuc. Sci and Tech, 48, 1349–1356 (2011).
[12] Yu Morino, Toshimasa Ohara, and Masato Nishizawa, " 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).
[13] http://www.mext.go.jp/b_menu/shingi/chousa/gijyutu/017/shiryo/__icsFiles/afieldfile/2011/09/02/1310688_1.pdf
[14] http://www.mext.go.jp/b_menu/shingi/chousa/gijyutu/017/shiryo/__icsFiles/afieldfile/2011/09/02/1310688_2.pdf
[15] Tokyo Metropolitan Institute of Public Health, http://monitoring.tokyo-eiken.go.jp/monitoring/f-past_data.html
[16] http://www.mext.go.jp/b_menu/shingi/chousa/gijyutu/017/shiryo/1311753.htm [in Japanese]
8
Fig. 1 Schematic drawing of the reactor damage and behavior of radioactive materials
Fig. 2
Radioactive materials in the environment
9
Fig. 3 High temperature phenomena in the core
Fig. 4 Temperature dependence of release rate constants from UO2 fuel
10
Fig. 5 Temporal evolution of the radiation dose rate at (a) 1F monitoring post, (c) nearby place and (d) far place. (b) wind
direction is indicated by the sine and cosine components. [Data other than in the Univ. Tokyo were mainly from the public release by TEPCO and MEXT].
11
Fig. 6
Temporal evolution of the radiation dose rate of far places. Right axis corresponds to the rain. Note that the data
for office 0.5 km far from 1F was devided by 10 to scale for its temporal behavior to be consistent with that of main gate
(GateM) and the west gate (GateW).
12
Fig.7
Fig. 8
Release fraction of leakage from reactor
The inventories of radionuclide in 1F1 at one day after the scram
Fig. 9 The inventories of chemical elements in 1F1 at one day after the scram
13
Fig. 10 Schematic drawing of evaluation of dose rate based on the ground shine.
14
Fig. 11 Dose rates at the location inside 20 km no-go zone, scaled to that on Mar. 15 when dominant radioactive release
occured.
Fig. 12 Dose rate mapping of schools in Fukushima performed in April. The figure above is the projection along the
latitude. Fukushima map from "National Land numerical information (Administrative Divisions, 2011),
Ministry of Land, Infrastructure, Transport and Tourism, file: japan_ver71, processed by esri Japan"
15
Fig. 13 Map of deposition of radioactive cesium (sum of Cs-134 and Cs-137) for the land area within 80 km of
the Fukushima Daiichi plant, reported by MEXT.
Fig. 14 Becquerel Ratio of Sr90/Cs137 [%] as a function of the distance from 1F. Color of dots corresponds to the latitude, but no clear tendency has been observed.
16
Simulation Analysis
MAAP Code (TEPCO)
1F1
Unit
3h
Core Exposure
4h
Core Damage
RPV melt-through 15 h
MELCOR Code (NISA)
1F1
Unit
2h
Core Exposure
3h
Core Damage
RPV melt-through 5 h
1F2
75 h
77 h
109 h
1F3
40 h
42 h
66 h
1F2
75 h
77 h
80 h
1F3
41 h
44 h
79 h
70h39
?
77 h
35h56
42h30
68h12
Actual Events
IC/ RCIC stopped
Vent (AO bulb)
Explosion/Rupture
2h50
23h44
24h50
Table 1 Hours from scram (2011-Mar-11 14:46 JST)
2800˚C 1hour
100%
Xe, Kr
100%
H
100%
Cs
100%
I
100%
Te
83%
Ba
58%
Sr
1.7%
Zr
1%
Np
100%
Mo
1.7%
La
0.17%
Pu
1%
Am
2000˚C 4hour
100%
Xe, Kr
100%
H
100%
Cs
100%
I
100%
Te
29%
Ba
16%
Sr
0.34%
Zr
0.5%
Np
99%
Mo
0.34%
La
0.034%
Pu
0.5%
Am
Table 2 The fraction of inventory released from the fuel
(i)
(ii)
(iii)
(iv)
(v)
Total
land
fraction
Source
I-eq.
I-131
8.03
12.5
2.33
1.35
1.63
25.8
0.13/0.44
Cs-134
0.95
1.47
0.27
0.16
0.19
3.05
0.22/0.46
Cs-137
0.93
1.44
0.27
0.16
0.19
2.98
0.22/0.46
199/59
199/59
13.8/6.6
38.8/18
13.6/6.5
542/258
Table 3 Evaluated ground contamination and source terms in PBq
17
Method
Radionuclide release
analysis
§ 4.1
Radiation
monitor
§ 4.2.2
Chernobyl
NISA
NSC
ORIGEN
CORSOR-O
Chemical analysis
MELCOR
Radiation map
(Ground shine)
SPEEDI
Dust Sampling
Core inventory
analysis code
I-131
equivalent
I-131
x1
Cs-134
x 2.8
Cs-137
x 40
490
60
7.4
7.6
370
(770)
130
-----
6.1
336-780
59-199
6.6-13.8
6.5-13.6
630(570)
150(130)
-----
12(11)
5200
1800
-----
85
Table 4 Comparison between different approaches to evaluate the total release in PBq. Some data have been updated
from the initial publication, indicated in the round brackets.
18
Appendix. A Vapor pressure of water
Water tends to condense more at higher pressures. The saturated vapor Psat [atm] above 1 atm can be empirically approximated as a function of the temperature T [ºC] by the Antoine equation proposed in 1888,
log Psat =c0 -c1/(c2+T)
(A1)
where {c0, c1, c2} are {8.07131, 1730.63, 233.426} for 1 < T < 100, while {8.14019, 1810.94, 244.485} for 99 < T <
374.
The result of the T as a function of Psat is shown in Fig. A. For example, sudden decrease of the pressure, say from 70 to
1 atm, by valving on the SRV, leads to the sudden drop of the vaporization temperature from 290 to 100 ºC. As a result
quite rapid evaporation of the water in RPV might happen. Therefore, water injection needs to be started at the same time
as, or at least as soon as the SRV opens.
Fig. A Saturated vapor pressure.
Appendix B
Ground shine
Half-value layer
External exposure from the contaminated ground is called "ground shine". Required database is compiled in ref. [B1].
Due to the absorption/scattering by the air, the radiation was attenuated. The characteristic lengths at which the radiation
become half, half-value layer (HVL), of lead, iron, aluminum, water, air and concrete are compiled in Table E2 [B1].
Roughly speaking, the dose rate by the ground shine reflects the area inside the circle of several times the air HVL in
radius. Each point of Figs. 10 and 14 represents at least 4.57 and 4 km2, respectively. Therefore, the influence of the area
for the adjacent points in the MEXT monitoring that could cause a double-counting of the dose rate are eliminated.
Dose factor
19
Conversion factors from [Bq/m2] to [Sv/h] listed in Table E3 [B1] include effective dose rate for external dose and
committed inhalation due to resuspension resulting from remaining on contaminated ground. However, since the inhalation dose for the present situation is considerably small, this database can also be used for the pure external dose.
The shielding factor (SF) needs to be considered, since the evaluation above is the ideal case where the ground is
smooth spread over the infinite disk. In the ordinary ground ref[B1] proposed to use SF of 0.47-0.85(representative of
0.7).
In addition, the dose factors for the plume submersion, that is crucial in the early stage of accidents, is calculated in the
Table III.1 of ref.[B2].
Radiological equivalence
Radiological equivalence is the ratio of the activity released of a specified radionuclide to the case for 131I. This value is
used to classify the scale of the accident as describe in ref [B3]. It consider the above mentioned ground contamination
and the plume submersion. The database for the total effect on the public are given in Table 15 in Appendix I of ref.[B3].
These are listed in Table B1 for species of interest. Although 134Cs is the present leading radiator, 137Cs is the most crucial nucleus in the evaluation of the rating of the accident. On the other hand, 90Sr is less effective due to both the low
concentration (approximately below 1 % of 137Cs) and the low radiological equivalence (half of 137Cs).
Experimental determination of the Shielding Factor
By comparing the ground contamination with the dose rate for each point in the radiation MAP, shielding factor SF can
also be evaluated, as shown in Fig. B.
In the fitting procedure, we fixed the baseline corresponding to the background natural radiation dose.
One can see that the SF=0.7 is a plausible value in the general discussion, but the scattering of the data is considerably
large. Therefore, it may mislead if one calculate the external exposure from the local ground contamination. On the other
hand, dose rate was the averaged, namely effective, value around the measurement point.
[B1] IAEA, "Generic procedures for assessment and response during a radiological emergency" [IAEA-TECDOC-1162]
(2000). http://www-pub.iaea.org/mtcd/publications/pdf/te_1162_prn.pdf
[B2] Keith F. Eckerman and Jeffrey C. Ryman "External exposure to radionuclides in air, water, and soil" FEDERAL
GUIDANCE REPORT NO. 12, EPA-402-R-93-081, OAK RIDGE NATIONAL LABORATORY (1993).
http://nnsa.energy.gov/sites/default/files/seis/fgr12.pdf
[B3] INES, The International Nuclear and Radiological Event Scale "the international nuclear and radiological event
scale user’s manual 2008 edition" (international atomic energy agency vienna, 2009)"
http://www-pub.iaea.org/MTCD/publications/PDF/INES-2009_web.pdf
Table B1. Dataset to access radiological equivalence to I-131
nucleus
τ
I-131
Cs-134
Cs-137
Sr-89
Sr-90
8.0d
2.06y
50y
50.5d
29.1y
Air
HVL[m]
55.9
71.9
69.2
80.5
-
[Sv/h]/[Bq/m2]
I-131 eq.
1.30E-12
5.40E-12
2.10E-12
8.0E-15
1.0E-15
1
2.8
40
0.5
20
20
Fig. B Measured dose rates vs. those evaluated from the ground shine conversion with the ideal case, SF =1. Background dose rate of 0.06 µSv/h was assumed.
21