External Cesium-137 doses to humans from soil influenced by the Fukushima and Chernobyl nuclear power plants accidents: a comparative study.

External exposure to gamma-photon irradiation from soil contamination due to nuclear power plant (NPP) accidents has significant contribution to human radiation exposure in the proximity of the NPP. Detailed absorbed doses in human organs are rarely reported in the literature. We applied the Monte Carlo Neutron Particle (MCNP) transport code to calculate and compare the absorbed doses in different human organs. The absorbed doses by gamma-photon radiation were from cesium-137 (137Cs) in soil contaminated by the two major NPP accidents. More serious and wide-spread impacts of the Chernobyl NPP accident on soil contamination in Ukraine, Belarus, Russia and countries as far as Sweden and Greece were due to the inland location, radiative plume transport pathway and high 137Cs emission strength (9 times the Fukushima emission). Based on our MCNP calculations, the largest absorbed dose was found in skin. The maximum calculated external 137Cs annual effective dose received from the Chernobyl accident was 10 times higher relative to the Fukushima accident. Our calculated effective doses at various influenced areas were comparable to those available in the literature. The calculated annual effective doses at areas near the Fukushima and Chernobyl NPPs exceeded the ICRP recommendation of 1 mSv yr−1.

The external radiation doses in the human body irradiated by gamma photons from 137 Cs deposited in soil are of particular concern. Recent studies related to the FNPP accident for areas within the Fukushima prefecture, Japan have been undertaken 25,26 . Similar investigations at Bryansk, the most impacted areas by the CNPP accident in Russia, have been reported [27][28][29] . A study was carried out which focused on recent external doses in Minsk and Gomel, Belarus and Chernobyl, Ukraine in 2012 30 . It reported the estimated effective doses around CNPP to be still over the public dose limit of 1 mSv yr −1 recommended by the ICRP 31 26 years after the accident. The above studies were mainly undertaken at different local/regional-scale areas. A large-scale or country-scale comparative study on the soil distribution of 137 Cs and associated absorbed doses in the human organs between different impacted areas are rarely available in the literature, despite the fact that the determination of organ doses is important for cancer risk assessment 32,33 . The organ doses estimation has been conducted for populations near nuclear facilities 34 .
In the current study, we first summarized, compared and explained the literature values of the 137 Cs activity concentrations in soil measured in various areas influenced by the NPP fallouts. The external gamma-photon absorbed doses in different human organs due to derived 137 Cs concentrations in contaminated soil in these areas were then calculated by the MCNP code. The calculated human organ doses (in more than ten organs) around the NPPs were then compared, which is the novelty of our study. Finally, we evaluated our calculated annual effective doses with the reported values available in the literature for the FNPP and CNPP accidents. Given the available soil concentrations of 137 Cs, our study provided a method to calculate the annual effective doses at various areas where there were no estimates on effective doses available. The CNPP accident results were considered for the time immediately after the accident in order to have a meaningful comparison.

Results and discussion
Comparison of 137 Cs soil concentrations due to the NPP accidents. Table 1 summarizes the literature values of 137 Cs concentrations (kBq kg −1 ) in soil at various locations influenced by the FNPP and CNPP accidents. All the data reported here for CNPP accident are the values as in 1986. Not only the source-receptor distance, but also the meteorological setting for the radiative plume transport and soil property could affect the soil concentrations at a specific location. For the Fukushima event, the maximum soil concentration (66.7 kBq kg −1 ) was found in Minamisohma city 15 km from the NPP. Cities locating relatively far away from the NPP (>50 km) had smaller contamination of order of 10 kBq kg −1 or less. The data of 137 Cs soil concentrations influenced by the CNPP accident were difficult to obtain since the available data in the literature were reported in kBq m −2 . Extremely high soil concentration (>250 kBq kg −1 ) was found less than 30 km from the CNPP 14,18 . Within 100 km from the CNPP, the concentration was more than 20 kBq kg −1 but the contamination pattern was highly irregular and anisotropic. Total areas of 7200 and 116000 km 2 with maximum concentrations of 23 and 2.8 kBq kg −1 were located in Belarus, Russia and Ukraine 16 . For the far field, Sweden, Austria and Greece showed abnormally high soil contamination 14 , attributed to the long-range transport of radiative plume. It was noted that the 137 Cs deposition on Sweden was highly uneven 11 and reference therein, and therefore the concentrations at both lower and higher sides were presented in Table 1. The uneven deposition was a result of changing meteorological conditions during the dispersion of released materials 14 .
While some FNPP fallout occurred on the Japan landmass, the majority (80%) was on the northern Pacific Ocean along with the prevailing westerly 8,35,36 since the FNPP is located at the coastline. Therefore the influences of the radiative plumes to other countries such as those in Northern America and Europe were small 7-10 . On the contrary, the CNPP was located well within the European continent. The large initial release height (>1 km above ground) of the radiative plumes due to explosions at the CNPP and convective updraft of the plumes during long-range transport were the major reasons of the high 137 Cs deposition to areas in northern and southern European countries located >1000 km from the CNPP 1-3,37,38 . The large differences in soil contamination for the two accidents were also due to the large differences in atmospheric release of radionuclides 137 Cs, which were in the range of 74-98 PBq for the CNPP 39-41 and 12-17 PBq for the FNPP 7,42-44 .

Calculated equivalent dose in organs and effective dose in areas influenced by the NPP accidents.
The MCNP modeling results calculated for equivalent doses in various organs of the mathematical phantom are given in Table 2. Skin and bone surface, and thyroid received the largest and smallest equivalent doses, respectively. Figure 1 shows the calculated annual external effective doses due to the 137 Cs derived gamma photons irradiation calculated for various areas with contaminated soil influenced by the FNPP and CNPP accidents. The calculated annual effective doses for the CNPP accident were evaluated immediately after the accident as described above to have a meaningful comparison with the FNPP accident. The calculated maximum annual effective dose in areas affected by the CNPP accident was about 10 times higher compared with that in areas affected by the FNPP accident. Similarly, the near-field areas (in Ukraine, Belarus and Russia) received higher effective doses from the CNPP accident than those from the FNPP. Interestingly, the people habitat in Greece (~1400 km from the CNPP) received comparable effective dose as people in the Nishi-Shirakawa County, Japan located 85 km away from the FNPP. Even the lowest effective doses (0.01-0.02 mSv yr −1 ) calculated for Austria and Sweden for the current study, which were affected by the CNPP accident, were still several times higher than the background effective doses received in South Asian countries 45,46 .
For the accident in Japan, our result of calculated annual effective dose of 0.9 mSv yr −1 obtained for Fukushima City (Fig. 1) was well comparable to the external 137 Cs dose assessment for the same City (0.5 mSv yr −1 ) by Yoshida and Suzuki 25 . Taira et al. 26 reported the total external effective doses at the same location, which ranged from 2.2 to 7.6 mSv yr −1 after 3 month of the FNPP accident. Comparing with our dose calculations, higher dose levels were expected for the estimations by Taira et al. since they included the contribution from 134 Cs radionuclides, which was higher (for example, the dose contribution of 134 Cs was 3 times higher than 137 Cs in the early period after the accident 47 ). For the same reason, our calculated dose of 3.6 mSv yr −1 for the Minamisohma City was also comparable with that of 4.6 mSv yr −1 at a nearby village − Iitate Village 26 , to which 134 Cs and 137 Cs were the contributors. It is noted that the differences in dose levels could also be due to the adoption of different calculation methods. For instance, several empirical coefficients, such as occupancy-shielding factor, were adopted for the dose calculations for the work of Taira et al.
Regarding the accident in Ukraine, a recent study reported the external effective dose of 22 mSv yr −1 in soil samples taken in CNPP (Manany) 30 . The effective dose was calculated based on multiple radionuclides with predominated activity contribution from 137 Cs. Our 137 Cs dose calculation at the CNPP of 16 mSv yr −1 (Fig. 1) was thus reasonably compared with the reported value. There were large variations in the estimated external effective doses reported in the literature in western Bryansk, Russia -an area with the most significant soil contamination in Russia. For instance, Ramzaev et al. 28 reported effective doses from 0.6 to 1.9 mSv yr −1 but Thornberg et al. 27 reported values up to 2.8 mSv yr −1 . The large variations in the effective doses were reflected by the corresponding heterogeneity in soil activity concentrations from <45 kBq m −2 to >1806 kBq m −2 16 . Nevertheless, our calculated effective dose for Russia in areas with low soil concentration (0.19 mSv yr −1 in Fig. 1) was comparable to the minimum dose (0.6 mSv yr −1 ) evaluated in these studies. Our result calculated for Gomel, Belarus (1.3 mSv yr −1 , Fig. 1) was comparable to the estimation by Thornberg

Methods
Calculation of the conversion coefficients. The conversion coefficients for different human organs irradiated by gamma photons from 137 Cs in soil were calculated by the MCNP5/X version 2.6.0 code 48 . Detailed descriptions of the coefficients for different organs are available elsewhere (e.g. Krstić and Nikezić 49 ). Briefly, an Oak Ridge National Laboratory (ORNL) mathematical phantom described in ICRU Report 48 50 and Eckerman et al. 51 was adopted for the absorbed dose calculations 52 . Similar methods have been used for dose calculations in many studies [53][54][55][56] . The phantom consists of elliptical cylinders, truncated circular cones, circular cylinders, half ellipsoids, etc. to represent the trunk, arms, legs, feet, neck, head and female breasts. These "organs" were described by mathematical equations and inequalities 51 , which were then programmed in the input files for the MCNP code. Totally 66 (68) cells and 180 (188) surfaces were used for a male (female) phantom (Fig. 2). Transport of gamma photons with energy 662 keV from soil to target organs was simulated by the MCNP code. The soil source (pure SiO 2 ) was assumed to be cylindrical shape with a radius of 3 m. Photons emitted farther than 3 m from the phantom have small probability to hit the target and were neglected here. The 137 Cs vertical migration is known to be very slow and most of the accident-derived 137 Cs was found within the top 15 cm of the soil 16 . The profiles reported in IAEA 16 in Ukraine and Taira et al. 26 in Iitate village, Japan were adopted respectively  (Table 1), 24 (Table S1) Minamisohma city 15 66.7 23 (Table 1), 24 (Table S1) J Village 20 9.4 20 ( Table 1) Motomiya city 50 8.4 23 (Table 1), 24 (Table S1) Fukushima city 60 15.8 21 (Fig. 4), 22 (Fig. 1), 23 (Table 1), 24 (Table S1) Koriyama city 60 11.4 22 (Fig. 1), 23 (Table 2), 24 (Table S1) Nishi-Shirakawa county 85 1.6 22 (Fig. 1), 24 (Table S1) CNPP accident Chistogalovka, Ukraine 3 87.1 **** 17 (Table 1 Table 1. Summary of soil 137 Cs concentrations contaminated by the FNPP and CNPP accidents. *For the FNPP accident, arithmetic mean was reported when more than one reference reported the soil concentration for the same location. Concentrations at top soil (0-10 cm) were presented. ** A typical soil density of 1300 kg m −3 (21) and top 5 cm sampled surface soil 11,17,30 were assumed. *** A typical soil density of 1300 kg m −3 (21) was assumed. **** Value corrected to 1 May 1986. ***** See text.
(Supplementary Materials S2) as the 137 Cs distribution in soil cylinder impacted by the CNPP and FNPP accidents in the current study despite the information of site-specific 137 Cs vertical profile in soil is not always available. The 137 Cs concentration for the top 5 cm soil was about twice as that for the next 5 cm soil for both profiles. The soil profile adopted here is consistent with those reported elsewhere 57,58 . The cylinder representing the source was split into smaller cylinders 2 cm in height. Homogenous distribution of the 137 Cs radionuclides in smaller cylinders were assumed and uniform sampling of initial points in these small cylinders was applied. Totally, 10 8 simulations were run for each source to ensure small calculation uncertainties (relative error is less than 10%). The MCNP energy deposition tally F6 was used for dose estimation. Conversion of units was applied to the MCNP results to obtain the conversion coefficient (in fGy per Bq s kg −1 ) for all major organs as a function of the source depth. The conversion coefficient is the absorbed dose rate in target organ per unit activity concentration in soil. The coefficient as a function of soil depth, which has been previously presented 49 , is given in Supplementary Materials S3.

Calculation of the effective dose due to soil contamination from FNPP and CNPP accidents.
The equivalent dose (H t ) in a tissue or organ (t) was calculated through the sum: where D t,r is the dose absorbed in that organ from 137 Cs in soil for 1 year (in Gy y −1 ) from the radiation of type r; the radiation weighting factor is w r = 1 for photons of all energies according to ICRP Publication 60 31 . The soil concentrations of 137 Cs were obtained from Table 1 below. Contributions to the equivalent dose in organs from a soil layer were calculated by multiplying the above conversion coefficients with the number of disintegrations occurring during one year. The contributions from different layers were then summed up to obtain the equivalent dose in that organ. The effective dose (E) for an individual standing on the ground was calculated according to ICRP Publication 116 59 : where H t,male and H t,female are equivalent doses in male and female phantoms, respectively. The values of w t are given in ICRP Publication 103 60 . An average outdoor occupancy factor of 0.3 61,62 was included to the annual effective dose calculation. The factor used is comparable to the UNSCEAR's value of 0.2, which is suggested to have differences around the world 63 . The differences are due to that, for instance, human is considered likely to spend more (less) time indoors for industrialized (agricultural) countries in temperate (warm) climates. More details of the occupancy factor could be referred to Hinrichsen et al. 64 It is noted that the external dose is likely to be changed when the radionuclides migrate deeper into the soil 64 .

Concluding remarks
Our study summarized the literature values of 137 Cs soil concentrations influenced by the CNPP and FNPP and rationalized their spatial distributions. The soil distributions were affected by the location of NPPs (either inland or coastal), initial release heights and emission strengths of the contaminants, as well as the atmospheric transport pathways. We then used these values and the conversion coefficients determined by the MCNP code to calculate the organs doses (µSv yr −1 ) due to 137 Cs in soil in various areas contaminated by the NPP accidents, which are important for cancer risk assessment but were seldom reported. Finally, we discussed the similarity and differences of our calculated effective doses with values reported in the literature. Given the available soil concentrations of 137 Cs, the annual effective doses at various locations could be calculated, where estimations of the effective doses at these locations were not available.