Comparison of the UNSCEAR isodose maps for annual external exposure in Fukushima with those obtained based on airborne monitoring surveys

In 2016, in attachment 2 of UNSCEAR's Fukushima 2015 White Paper (hereafter referred to as Att-2) [1], as well as in an animation [2], UNSCEAR published maps showing isodose lines for an annual additional effective dose of 1 mSv from external exposure in years 1, 3, 5, 10, 30 and 50 years after the Fukushima Dai-ichi Nuclear Power Plant (FDNPP) accident. UNSCEAR estimated isodose lines based on radionuclide soil deposition data obtained from soil samples collected between 6 June 2011 and 8 July 2011 within 100 km from the FDNPP. The data were released by the Japanese Government in 2011 [3] and were reproduced by UNSCEAR with geospatial information on a 1 km grid [4]. Att-2 provides a table (table 2 of [1]) of the conversion coefficients from soil deposition density of radiocaesium to annual effective dose ab extra ($\mathrm{mSv}/{\mathrm{MBqm}}^{-2}$) for each year from years 1 to 10, then at every 10 years from years 10 to 50, and 100. Using the soil deposition data and the effective dose conversion coefficients, it is possible to estimate the annual additional effective dose from external exposure at each grid point around the FDNPP.

Meanwhile, Miyazaki and Hayano [5] analyzed data from a large-scale 'glass-badge' individual dose monitoring (with geographical information system (GIS) data of the residents' addresses) conducted by Date City, Fukushima Prefecture, together with the airborne monitoring data collected periodically by the Japanese Government, and found that the ratio c of ambient dose equivalent rate to the effective dose rate from external exposure is nearly constant between 5 and 51 months after the accident, at $c\sim 0.15$ [5]. Using this relationship, it becomes possible to draw annual 1 mSv isodose lines based on the airborne monitoring data.

In this paper, we compare the 1 mSv annual isodose lines predicted by UNSCEAR with those obtained based on the airborne monitoring for 3 and 5 years after the accident, and discuss the implications.

2.1. Reproduction of the UNSCEAR isodose lines

2.2. Isodose lines based on the airborne monitoring maps

The isodose lines based on the airborne monitoring maps were generated as follows.

For the annual additional effective dose 3 (5) years after the accident, we used the 8th (10th) airborne monitoring data for which the reference date of dose calculation is 19 November 2013 (2 November 2015), as indicated in figure 1. By multiplying the ambient dose equivalent rate ${\dot{H}}^{* }(10)$ ($\mu {\rm{Sv}}\,{{\rm{h}}}^{-1}$) at each airborne monitoring grid point by the factor $c=0.15$ [5], the median value of the external effective dose rate of the population living in the vicinity of the grid point was obtained.

Zoom In Zoom Out Reset image size

Since the ambient dose equivalent rate 1 m above the ground ${\dot{H}}^{* }(10)$ gradually decreased over time, we cannot simply use the dose rate on the reference date of airborne monitoring as a representative of the whole year. We therefore fitted, as shown in figure 1, the ratios of the ambient dose rate from the 5th through 11th airborne monitoring at each monitoring grid point to that from the 4th monitoring (AM5 through AM11 in figure 1), with a model function that considers physical and environmental attenuation,

$$\begin{eqnarray}&&f(t)=N\{{a}_{\mathrm{fast}}{2}^{-t/{T}_{\mathrm{fast}}}+(1-{a}_{\mathrm{fast}}){2}^{-t/{T}_{\mathrm{slow}}}\}\cdot \displaystyle \frac{(k\times {2}^{-t/{T}_{134}}+{2}^{-t/{T}_{137}})}{k+1}.\end{eqnarray} \tag{ 1 }$$

Here, ${T}_{134}=2.06\,{\rm{y}}$ and ${T}_{137}=30.17\,{\rm{y}}$ are, respectively, the physical half lives of ¹³⁴Cs and ¹³⁷Cs, ${T}_{\mathrm{fast}}(=0.43\pm 0.01\,{\rm{y}})$ and ${T}_{\mathrm{slow}}(\gt 400\,{\rm{y}})$ are the fast and slow half lives of environmental decay, and ${a}_{\mathrm{fast}}(=0.60\pm 0.01)$ is the fraction of the fast decaying component, and k = 2.95 is the ratio of air-kerma-rate constant of ¹³⁴Cs to ¹³⁷Cs [6]. In the fit, the pre-factor N was chosen to set the value of the model function to unity at the timing of the 4th monitoring (t = 0.65 y).

For year 3, the function f(t) (equation (1)) was integrated from $t=2\,{\rm{y}}$ to $t=3\,{\rm{y}}$ (shaded area A in figure 1) and the result was compared with the value at the reference date of the 8th airborne monitoring (red dot at y = 2.69). The correction factor for year 3 thus obtained was $F=1.05$. For year 5, the integration was done from $t=4\,{\rm{y}}$ to $t=5\,{\rm{y}}$ (shaded area B in figure 1), which yielded a factor $F=1.03$. This procedure is graphically presented in figure 1 by gray rectangles drawn on top of area A and on B. We used these factors to convert the ambient dose equivalent rate at each grid point ${\dot{H}}^{* }(10)$ (μSv h^–1) to the annual additional ambient dose equivalent ${H}^{* }(10)$ (mSv), as, ${H}^{* }(10)=F\times \tfrac{24\,\times \,365}{1000}\times {\dot{H}}^{* }(10)$. The annual additional effective dose from external exposure maps created using the factor c = 0.15 were interpolated to draw 1 mSv annual isodose lines.

In figure 2 top (bottom), isodose lines for an annual additional effective dose of 1 mSv for 3 (5) years after the accident are overlaid on the 8th (10th) airborne monitoring map. The 1 mSv isodose lines shown in blue are the recalculated UNSCEAR estimate, and the contours in red are the 1 mSv isodose line estimated from the airborne monitoring results, converted to annual effective dose rates with the factor c = 0.15. Also superimposed are (hatched regions) the isodose bands between c = 0.10 (25th percentile) and c = 0.22 (75th percentile), reflecting the distribution of the factor c as presented in figure 5 of [5]. Presumably, uncertainties may be also inherent in the UNSCEAR's estimate (blue line), but they are not explicitly discussed in [1].

Zoom In Zoom Out Reset image size

In both years, the isodose lines independently calculated with different methods are in fairly close agreement, but on closer examination, the UNSCEAR predicted isodose line (blue) encloses a smaller area than the line based on actual measurements derived in this study (red), the difference being slightly greater for the 5th year.

After the FDNPP accident, an airborne monitoring method has been established and carried out regularly [7], and the ambient dose equivalent rates have been released as maps and numerical data [8]. Using the data of the large-scale individual dose monitoring conducted by Date City, Fukushima Prefecture, Miyazaki et al [5] found that the personal dose equivalent (≈the effective dose) monitored by glass-badges, and the ambient dose equivalent of the residential area from airborne monitoring are closely correlated.

Naito et al [9] used the airborne monitoring database and individual dosemeters (D-Shuttle) along with a global positioning system, possessed by approximately 100 voluntary participants, and obtained a conversion factor of $c\sim 0.2$. In a more recent study by Naito et al [10], targeting 38 voluntary residents of Iitate village, Fukushima Prefecture, the coefficient was $c\sim 0.13$ for time spent at home and $c\sim 0.18$ for time spent outdoors.

The results of these well-controlled studies targeting a small number of participants are thus consistent with the large-scale 'glass-badge' result of [5]. These works have established that the results of the airborne monitoring can be used to assess the median effective doses of the groups residing in areas defined by the airborne-survey grid points. In the present study, we used c = 0.15, the coefficient obtained from the Date City glass-badge survey data.

UNSCEAR, in Att-2 of their 2015 White Paper [1], predicted the long-term cumulative effective dose ab extra after the Fukushima accident. In Att-2, estimates of cumulative effective dose up to 100 years after the accident are shown in terms of ¹³⁷Cs deposition density as of 14 June 2011. They also published a GIF animation their website at the same time; isodose lines for an annual additional effective dose ab extra in the 1st, 3rd, 5th, 10th, 30th and 50th years after the accident were presented. Comparison of the isodose line of annual additional effective dose of 1 mSv estimated by UNSCEAR and the isodose line estimated from the data obtained by airborne monitoring with the factor c = 0.15 [5] shows relatively good agreement, despite the widely differing methods employed.

However, on a closer examination, a trend in which UNSCEAR's isodose line encloses a smaller area than that generated in the present work is evident. The reason for this small but noticeable difference can be understood by comparing the attenuation of the ambient dose rate deduced from the airborne monitoring, with that assumed in the UNSCEAR's prediction. The attenuation curve obtained by analyzing the airborne monitoring data (in red), already shown in figure 1, is reproduced in figure 3. In comparison, a curve (in blue) for the attenuation of effective dose rates of representative population group calculated from the levels of deposition of radionuclides on the soil is superimposed using the parameters assumed in UNSCEAR's prediction, as described in detail in [11]. In UNSCEAR's model, the attenuation g(t) of the effective dose rate is factorized into three parts,

$$\begin{eqnarray}&&g(t)=\mathrm{physical}(t)\times \mathrm{environment}(t)\times \mathrm{location}(t).\end{eqnarray} \tag{ 2 }$$

Here, $\mathrm{physical}(t)$ is the physical decay curve of ¹³⁴Cs and ¹³⁷Cs, $\mathrm{environment}(t)$ is the environmental attenuation with a 50% fast component half life of 1.5 y and a 50% slow component half life of 50 y, and $\mathrm{location}(t)$ is the location factor for typical adults (estimated to spend 0.6 of their time in wooden one-to-two-storey houses and 0.3 of their time at work in concrete multi-storey buildings), as described in [11].

Zoom In Zoom Out Reset image size

The reason for the difference between the two attenuation curves can be understood by showing environment(t) and local(t) of each curve individually (figure 4). Equation (1) uses a smaller ${T}_{\mathrm{fast}}$ value and does not have $\mathrm{local}(t)$, which makes the curve a relatively rapid asymptote to the physical decay curve. The UNSCEAR's g(t) has two, almost similar magnitude of relatively slow attenuation functions, which together make a greater reduction. The authors deem that both deposition density-to-dose conversion coefficient and $\mathrm{local}(t)$ based on the study of the Chernobyl accident do not fit the Fukushima case. Although UNSCEAR's prediction is useful for policy-makers, it is necessary to update it continuously based on the latest diachronic monitoring data.

Zoom In Zoom Out Reset image size

In this paper, we compared two different methods to estimate the individual external doses of the public residing in the Fukushima Prefecture caused by radioactive fallout following the FDNPP accident in 2011. One is the method adopted by UNSCEAR, which makes use of the official soil deposition map. The other makes use of airborne monitoring maps, together with an empirical factor c = 0.15 obtained by comparing the large-scale 'glass-badge' dosemeter measurements and airborne surveys. Although the two are independent and are based on different data and methodologies, the resultant isodose lines for an annual additional dose of 1 mSv 3 and 5 years after the FDNPP accident show relatively good agreement. However, to improve the accuracy of long-term annual effective isodose lines, feedback from continuous measurements such as airborne monitoring is important.

The authors are grateful to Dr J Tada of Radiation Safety Forum for valuable discussions. This work was partially supported by donations by many individuals to RH through The University of Tokyo Foundation.