Summary of temporal changes in air dose rates and radionuclide deposition densities in the 80 km zone over five years after the Fukushima Nuclear Power Plant accident

Journal of Environmental Radioactivity xxx (xxxx) xxxx

Contents lists available at ScienceDirect

Journal of Environmental Radioactivity journal homepage: www.elsevier.com/locate/jenvrad

Summary of temporal changes in air dose rates and radionuclide deposition densities in the 80 km zone over five years after the Fukushima Nuclear Power Plant accident Kimiaki Saitoa,∗, Satoshi Mikamib, Masaki Andohb, Norihiro Matsudac, Sakae Kinasec, Shuichi Tsudad, Tadayoshi Yoshidae, Tetsuro Satof,g, Akiyuki Sekic, Hideaki Yamamotoa, Yukihisa Sanadag, Haruko Wainwright-Murakamih, Hiroshi Takemiyac a

Japan Atomic Energy Agency, 178-4-4 Wakashiba, Kashiwa, Chiba, 227-0871, Japan Japan Atomic Energy Agency, 11601-13 Nishi-jusanbugyo, Hitachinaka-city, Ibaraki, 319-1206, Japan Japan Atomic Energy Agency, 2-4 Shirakata, Tokai-mura, Ibaraki, 319-1195, Japan d OECD Nuclear Energy Agency, 46, quai Alphonse Le Gallo, 92100, Boulogne-Billancourt, France e Japan Atomic Energy Agency, 4-33 Muramatsu, Tokai-mura, Ibaraki, 319-1194, Japan f Hitachi Solutions East Japan Ltd., 2-16-10 Honcho, Aoba-ku, Sendai, 980-0014, Japan g Japan Atomic Energy Agency, 45-169 Sukakeba, Kaihama, Haramachi-ku, Minamisoma, 975-0036, Japan h Lawrence Berkeley National Laboratory, 1 Cyclotron Road, MS 74R-316C, Berkeley, CA, 94720-8126, USA b c

ABSTRACT

We summarized temporal changes in air dose rates and radionuclide deposition densities over five years in the 80 km zone based on large-scale environmental monitoring data obtained continuously after the Fukushima Nuclear Power Plant (NPP) accident, including those already reported in the present and previous special issues. After the accident, multiple radionuclides deposited on the ground were detected over a wide area; radiocesium was found to be predominantly important from the viewpoint of long-term exposure. The relatively short physical half-life of 134Cs (2.06 y) has led to considerable reductions in air dose rates. The reduction in air dose rates owing to the radioactive decay of radiocesium was more than 60% over five years. Furthermore, the air dose rates in environments associated with human lives decreased at a considerably faster rate than expected for radioactive decay. The average air dose rate originating from the radiocesium deposited in the 80 km zone was lower than that predicted from radioactive decay by a factor of 2–3 at five years after the accident. Vertical penetration of radiocesium into the ground contributed greatly to the reduction in air dose rate because of an increase in the shielding of gamma rays; the estimated average reduction in air dose rate was approximately 25% with penetration compared to that without penetration. The average air dose rate measured in undisturbed fields in the 80 km zone was estimated to be reduced owing to decontamination by approximately 20% compared to that without decontamination. The average deposition density of radiocesium in undisturbed fields has decreased owing to radioactive decay, indicating that the migration of radiocesium in the horizontal direction has generally been slow. Nevertheless, in human living environments, horizontal radiocesium movement is considered to contribute significantly to the reduction in air dose rate. The contribution of horizontal radiocesium movement to the decrease in air dose rate was estimated to vary by up to 30% on average. Massive amounts of environmental data were used in extended analyses, such as the development of a predictive model or integrated air dose rate maps according to different measurement results, which facilitated clearer characterization of the contamination conditions. Ecological half-lives were evaluated in several studies by using a bi-exponential model. Short-term ecological half-lives were shorter than one year in most cases, while long-term ecological half-lives were different across the studies. Even though the general tendency of decrease in air dose rates and deposition densities in the 80 km zone were elucidated as summarized above, their trend was found to vary significantly according to location. Therefore, site-specific analysis is an important task in the future.

1. Introduction More than seven years have passed since the Fukushima Dai-ichi Nuclear Power Plant (NPP) accident. During the accident, many radionuclides were released into the atmosphere on a large scale and deposited on land. The concentrations of radionuclides with short halflives have decreased below detectable levels, while those of

∗

radionuclides with long half-lives have existed in the environment and subjected the public to constant exposure (IAEA, 2015). To clarify the impact of the accident, large-scale environmental monitoring activities have continued under several national and other projects. A so-called mapping project aimed at comprehensive elucidation of the contamination conditions was undertaken by Japan Atomic Energy Agency (JAEA) in collaboration with many other organizations (Saito and

Corresponding author. E-mail address: [email protected] (K. Saito).

https://doi.org/10.1016/j.jenvrad.2018.12.020 Received 15 November 2018; Received in revised form 13 December 2018; Accepted 16 December 2018 0265-931X/ © 2018 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/BY/4.0/).

Please cite this article as: Kimiaki Saito, et al., Journal of Environmental Radioactivity, https://doi.org/10.1016/j.jenvrad.2018.12.020

Journal of Environmental Radioactivity xxx (xxxx) xxxx

K. Saito, et al.

Onda, 2015). The results of past mapping projects have been already reported in previous special issues (Saito et al., 2015a, 2016a). Aerial monitoring over a wide area was initiated by JAEA along with the US Department of Energy, and it has been conducted repeatedly since, mostly by JAEA (Sanada et al., 2014, 2018). National and local governments have performed continuous environmental monitoring by using monitoring posts and other methods. Furthermore, several nonprofit organizations have performed independent environmental monitoring and accumulated an enormous amount of radiation data (e.g., Nursal et al., 2016). According to these environmental data, radiation environments in the affected areas have changed greatly compared to the initial conditions post the accident. At present, radiocesium is a dominant contributor to public exposure doses, even though multiple radionuclides were released and deposited on the ground (Saito et al., 2015b). The dose rates in air have decreased owing to radioactive decay of radiocesium and other factors. Radiocesium migration in natural environments depends on land use and vegetation, but generally, it has been found to be slow (Yoshimura et al., 2015). However, radiocesium movement in the environment associated with human activities is relatively fast (Yoshimura et al., 2017), and this has contributed to a reduction in air dose rates in living environments. Deposited radionuclides have gradually penetrated the ground (Matsuda et al., 2015), and this has significantly contributed to the decrease in air dose rates because of increasing shielding effects. The present study aims to summarize temporal changes in radiation conditions over five years in the areas affected by the Fukushima accident, mainly in the 80 km zone around Fukushima NPP, based on reliable and comprehensive environmental data. To this end, we selected data from the mapping projects and aerial monitoring, which were performed continuously over a wide region by using standardized methods. First, we review the types of environmental measurements performed in the two abovementioned projects. Then, the decreasing tendency of air dose rates under different conditions is discussed, along with the general features of radiocesium movement in different environments. Furthermore, we introduce a few applied studies to promote understanding of the contamination conditions.

density (Saito et al., 2015b). Soil was sampled to a depth of 5 cm by using specific tool and placed in a 100 cm3 plastic container after it was mixed well according to the soil sampling protocol (Onda et., 2015). In principle, five soil samples were collected from one location. The radioactivity concentrations of gamma-emitting radionuclides in the collected soil samples were determined with spectrometry by using a Ge detector. For plutonium and radiostrontium, alpha- and beta-ray measurements, respectively, were performed after chemical processing. About 100 soil samples were selected for this analysis in each campaign. All collected soil samples are stored safely in specified storage facilities with information such as collection location, date, radionuclide concentration, and analysis history, because these samples are important for obtaining novel information pertaining to the initial contamination conditions. Parts of the stored soil samples were analyzed later according to authorized demands. For example, 129I measurements were performed using many of the stored soil samples to improve the 131 I map in which the number of data points were limited because of its short half-life (Muramatsu et al., 2015). 2.1.1.2. In situ spectrometry using a portable Ge detector. Since the second campaign in the mapping project, in situ spectrometry by using a portable Ge detector has been employed to measure radionuclide deposition density (Mikami et al., 2015a; 2018a). This method can determine the average deposition density at a given location by detecting gamma rays originating from a wide area. In the second campaign, in situ spectrometry was performed at more than 1000 locations in eastern Japan. The measurements were then concentrated at approximately 380 locations in the 80 km zone. 2.1.2. Depth profile in ground Depth profiles in the ground were examined by collecting soil samples at different depths with a scraper plate, which peels off soil from the ground surface step by step (Matsduda et al., 2015). This method minimizes cross contamination among soil samples, thus facilitating the determination of accurate depth profiles. Soil samples were collected up to a depth of 10 cm, and the thickness of each sampling layer was varied from 0.5 cm to a few cm based on the distribution conditions. Radiocesium concentrations in the collected soil samples were measured using a Ge detector in the laboratory and expressed as a function of depth.

2. Measurement methods employed in large-scale environmental monitoring 2.1. Radionuclide ground deposition

2.2. Air dose rate

Two different quantities of radionuclide ground deposition were obtained in the mapping projects: 1) radionuclide deposition density in Bq/m2 and 2) depth profile of radiocesium, i.e., radiocesium concentration at different depths from the ground surface in Bq/kg.

In the mapping project, air dose rates were measured using four different methods targeting different situations: 1) measurement at a fixed location, 2) walk survey, 3) car-borne survey, and 4) unmanned helicopter survey. In another project, helicopter surveys covering wide areas were undertaken. These measurement methods are listed in Table 1, along with their qualitative scores about data accuracy, positional resolution, and area coverage. A larger score (on a scale of 1–5) indicates a higher rank with respect to each evaluated quality. In the present study, air dose rate refers to the ambient dose equivalent rate

2.1.1. Radionuclide deposition density 2.1.1.1. Soil sampling and analysis. In the first campaign of the mapping project, more than 11,000 soil samples were collected from approximately 2200 locations in the 100 km zone and the remainder of Fukushima Prefecture to determine the radionuclide deposition

Table 1 Evaluation of measurement methods employed in large-scale environmental monitoring after Fukushima accident. A higher number indicates a higher rank in the evaluation. Measuring method

Measurement at a fixed location Walk survey Car-borne survey Unmanned helicopter survey Helicopter survey

Evaluated score

Noteworthy

Data accuracy

Positional resolution

Coverage

5 4 3 2 1

5 4 3 2 1

1 2 4 4 5

2

Give reference value Related to living environment Enormous amount of data Supplment helicopter survey Cover whole area

Journal of Environmental Radioactivity xxx (xxxx) xxxx

K. Saito, et al.

Table 2 Periods in which large-scale environmental monitoring was conducted in national projects.

(μSv/h) at 1 m above the ground.

in a manner that helicopter surveys can.

2.2.1. Measurement at a fixed location Measurement at a fixed location was performed using a TCS-171 or TCS-172 survey meter (Mikami et al., 2015b, 2018a), which provides accurate air dose rates by using a NaI(Tl) detector and a spectrum-dose conversion function to compensate for the energy response (G(E) function) (Tsutsumi et al., 1991). For this measurement, undisturbed fields, where neither human disturbance nor high flooding were expected to change the conditions, were selected. The number of measurement locations was approximately 6500. A reference air dose rate for each region was obtained by performing this measurement using a conventional survey meter; nevertheless, the coverage area of this measurement is considerably smaller than that of other mobile survey methods.

2.2.4. Helicopter survey Helicopter surveys generally employed a large NaI(Tl) detector with high detection efficiency attached inside or under a helicopter (Sanada et al., 2014, 2018). The standard flight height was set to 300 m. The total counts measured in air were converted to the air dose rate at a height of 1 m above the ground by using a calibration factor determined in a calibration test performed in real fields. Continuously measured data were interpolated using an inverse distance weighted (IDW) method, and finally, the air dose rate distribution of 250 m grids was determined. In Monte Carlo simulations, assuming exponential distribution of radiocesium in the ground at a relaxation mass depth of 1 g/cm2, 90% of the air dose rates measured at a height of 300 m were attributed to a source within a 470 m radius, and the air dose rates attributable to a source within a 100 m radius accounted for only 10% (Malins et al., 2015). This suggests that it is difficult to connect the data measured at a height of 300 m directly to the data just below the flight position. However, an airborne survey can detect gamma rays over a rather wide area. Therefore, the measurement can cover extended regions within in a short time. Generally, the values in airborne survey data have been found to be systematically slightly higher than those measured on land in case of the Fukushima accident (Wainwright et al., 2017, 2018).

2.2.2. Walk survey Walk surveys were performed to obtain air dose rates in various human living environments (Andoh et al., 2018a). In these surveys, a person carrying a KURAMA-II system walked around a 1 × 1 km2 square area for as long a distance as possible, and about 600 areas were targeted in each campaign. The KURAMA-II mobile survey system was developed by Kyoto University. It is easy to operate and automatically transfers the data obtained through a cellular phone network (Tanigaki et al., 2015). The accuracy of the measured air dose rates was verified from different aspects (Tsuda et al., 2015). In the walk survey, continuously measured air dose rates were averaged over an area of 20 × 20 m2 square to reduce statistical fluctuations. Walk survey data are essential from the viewpoint of exposure doses to the public because they indicate radiation levels in various types of places inhabited by humans. However, the regional coverage of this method is small compared to that of car-borne and helicopter surveys.

2.2.5. Unmanned helicopter survey Unmanned helicopter surveys using a LaBr detector were performed within a 5 km radius of the Fukushima NPP site, where it is difficult to enter by land (Sanada and Torii, 2015). The flight height was set to 80–100 m, resulting in higher positional resolution than that of the manned helicopter surveys. However, unlike the manned helicopter surveys, the measurements cannot cover wide regions because of technical limits and flight restrictions. Unmanned helicopter surveys are regarded as a method to supplement manned helicopter surveys.

2.2.3. Car-borne survey Car-borne surveys in the mapping project employed more than 100 KURAMA-II systems, and the surveys were conducted in collaboration with many local municipalities (Andoh et al., 2015, 2018b). These collaborations enabled us to construct a detailed car-borne survey map in about one month. The car-borne surveys covered wide affected areas in eastern Japan. The total distance traveled in these surveys amounted to several tens of thousands of km in each campaign. Continuously measured air dose rates were averaged over an area of 100 × 100 m2 to reduce statistical fluctuations. The car-borne surveys collected an enormous amount of data on land, and these data have been used effectively in statistical analyses. For example, the predictive model for air dose rates in the 80 km zone was developed mainly based on analyses of car-borne survey data (Kinase et al., 2017). However, car-borne surveys are restricted to areas with roads and cannot cover entire targeted areas

2.3. Periods of large-scale environmental monitoring Periods in which large-scale environmental monitoring was performed are summarized in Table 2. As shown in this table, large-scale environmental monitoring activities have been performed repeatedly by using the unified standard methods, which enabled us to analyze the contamination conditions systematically. Walk surveys were started in 2013 according to the analyzed monitoring data obtained in 2011 and 2012. The air dose rates at fixed locations in undisturbed fields indicated significantly different trends from those above roads measured by car-borne surveys. We assumed that the air dose rates in undisturbed fields represent the upper bound of air dose rates in a certain area with a similar contamination level, while those above roads represent the 3

Journal of Environmental Radioactivity xxx (xxxx) xxxx

K. Saito, et al.

lower bound, and air dose rates in various human living environments measured by walk surveys lie between these values. Proving this assumption was one of the major objectives of the walk survey initiated in 2013. Unmanned helicopter surveys were launched at the end of 2012 because of a lack of air dose rate data within a 3 km radius of the Fukushima NPP given that helicopter flights are generally prohibited in this zone and the fact that it was difficult for measuring crews to enter this zone by land.

Prefecture, was contaminated to a relatively high level. The deposition density ratios of two different radionuclides indicated several specific features. For example, the ratios of 131I to 137Cs were specifically higher in the southern coastal area (Saito et al., 2015a). Since the released radionuclide ratios were different owing to burn-up conditions of the reactors and release pathways, deposition ratios between the two radionuclides can be used to identify the reactor and the plume that dominated the deposition. The 134Cs/137Cs ratio was successfully used to identify the reactor that mainly contributed to contamination in different areas over eastern Japan (Chino et al., 2016). In the first campaign, 131I was detected at approximately 400 locations of the more than 2200 locations at which soil was sampled. Considering the importance of 131I in terms of exposure doses to the public in the early phases after the accident, the original 131I map shown in Fig. 1(b) was improved based on 129I measurements performed using accelerator mass spectrometry (Muramatsu et al., 2015). For this purpose, parts of the soil samples collected and stored during the first campaign were used, as mentioned previously. Finally, 131I deposition densities were plotted for approximately 800 locations (NRA, 2014). The effective doses ascribable to the detected radionuclides for 50 y from June 2011 were approximated from the maximum deposition densities obtained using the dose conversion coefficients published by IAEA (2000). The effective doses due to radiocesium would be higher than those due to other radionuclides by more than two orders of magnitude (Saito et al., 2015a). Furthermore, more than 99% of the air dose rates were estimated to be attributable to radiocesium in June 2011. Thus, environmental monitoring in the mapping project has been focused mainly on radiocesium ever since.

2.4. Quality assurance and control of obtained data The quality of the obtained data was assured in several ways. Fundamentally, measurement and sample collection were performed by following protocols determined beforehand. The measurement instruments were calibrated periodically by using corroborating sources traceable to National standards. Furthermore, in terms of the determination of radionuclide deposition density, inter-comparisons were performed to ensure that results from different Ge detector systems were within the bounds of acceptable uncertainty both for laboratory and in situ measurements (Mikami et al., 2018b). In the first campaign, environmental monitoring data that were not obtained by car-borne surveys were recorded on paper, resulting in rather high rates of miswriting and misreading. Since the second campaign, all data obtained in the fields were transferred to data collection servers through a cellular phone network soon after each measurement (Mikami et al., 2015b), and these data were screened using simple methods to detect outliers. For example, the data in question were compared to the data obtained at the same and surrounding locations during the previous campaign. The mapped data were carefully checked manually; if inappropriate data were found, measurements were reperformed or the data were removed. An ad-hoc committee formed to comment on adequate planning and implementation of the mapping projects contributed substantially toward ensuring the quality of the obtained results.

3.2. Total deposition of

137

Cs

The total deposition of 137Cs on land in Japan was estimated (Saito et al., 2016b) from the deposition densities obtained by in situ Ge spectrometry performed over wide regions in the second campaign (Mikami et al., 2015a), and these densities were integrated with helicopter survey data (Sanada et al., 2014). Fig. 2 shows the integrated deposition density map for 137Cs; the data from in situ Ge spectrometry were used for the areas in which the data existed, and helicopter survey data were used after they were adjusted to in situ data by using a correction factor for the areas in which in situ data did not exist. Total 137 Cs deposition in eastern Japan as of June 2011 was estimated to be about 2 PBq and that within the 80 km zone from the NPP about 1.6 PBq (Saito et al., 2016b). A recent study estimated the total release of 137 Cs into the atmosphere at 14.5 PBq (Katata et al., 2015); thus, 10%–20% of the 137Cs released was thought to be deposited on land in Japan. The proportions of 137Cs deposition on lands with different use types were estimated in the 80 km zone: 70% of 137Cs released into the atmosphere was estimated to be deposited in forested areas; 20% in farmlands, including paddy fields; and 5% in urban areas. These values

3. Features of deposited radionuclides 3.1. Radionuclides detected over wide areas The dominant radionuclides detected in the ground over wide areas after June 2011 when the mapping project started were 134Cs, 137Cs, 131 129m I, Te, 110mAg, 238Pu, 239+240Pu, 241Pu, 89Sr, and 90Sr. The periods in which these radionuclides were detected are listed in Table 3. Iodine-131, 129 mTe, and 89Sr with half-lives of 8.02 d, 33.6 d, and 50.5 d, respectively, were detected only in the first campaign in June 2011. Silver-110 m with a half-life of 250 d was detected over a wide area in the second campaign started in Dec 2011 as well. Deposition density maps were created for these detected radionuclides, and a few examples of such maps are shown in Fig. 1 (Saito et al., 2015b). The most significant deposition occurred in the region northwest of Fukushima NPP. Furthermore, the wide region stretching from the center of Fukushima Prefecture, through Tochigi Prefecture to Gunma

Table 3 Radionuclides detected over wide areas, and periods of detection and targeted areas. The bars in the table indicate that the radionuclide was not measured. Campaign No.

Period

Area

Detected radionuclide 134

137

131

129m

110m

89

○ ○ ○ ○ ○ ○ ○

○ ○ ○ ○ ○ ○ ○

○ X △ △ X X X

○ X X X X X X

○ ○ X X X X X

○ X – – – – –

Cs

1 2 3 4 5 6 7

2011/Jun - 2011/Nov 2011/Dec - 2012/Jun 2012/Jul - 2013/Mar 2013/Apr - 2014/Mar 2014/Apr - 2015/Mar 2015/Apr - 2016/Mar 2016/Apr - 2017/Mar

100 km zone, Fukushima pref. Eastern Japan 80 km zone 80 km zone 80 km zone Part of eastern Japan 80 km zone 80 km zone

○: detected, X: not detected, -: not measured, △: reconstructed according to

Cs

129

I

Te

I measurement

4

Ag

Sr

90

○ ○ – – – – –

Sr

238

○ ○ ○ ○ – – –

Pu

239 + 240

241

Natural

○ ○ ○ ○ – – –

– ○ ○ ○ – – –

– – – – ○ – ○

Pu

Pu

Journal of Environmental Radioactivity xxx (xxxx) xxxx

K. Saito, et al.

Fig. 1. Examples of radionuclide deposition density maps created in first campaign of Fukushima mapping project (Saito et al., 2015).

were almost proportional to the gross land areas for each use type. Radiocesium deposition in urban areas was relatively low. However, this is considered the dominant source of exposure to the public. In smaller regions, radiocesium deposition was not always proportional to the gross area.

the radiocesium deposited on artificial structures, especially roads, is generally run off quickly after deposition (Anderson et al., 2002). The same tendency was observed after the Fukushima accident as well (Yoshimura et al., 2017). Furthermore, relatively large radiocesium run-off was observed in paddy fields (Yoshimura et al., 2016). Thus, radiocesium migration depends greatly on the situation (Wakiyama et al., 2018). The overall migration rates of deposited radiocesium in the horizontal direction were generally considered small because the evaluated total radiocesium flux through rivers was small in comparison with the total deposition of radiocesium in the river catchment. Several studies have reported that the annual discharge rate was estimated to be of the order of 0.1% of the total deposited radiocesium in the catchment (Iwagami et al., 2017; Funaki et al., 2018).

3.3. Trend of radiocesium deposition density Fig. 3 shows the trend of the average deposition density of 137Cs and Cs in the 80 km zone obtained by combining the data from the first campaign (Saito et al., 2015b) and the data obtained since the second campaign (Mikami et al., 2015a, 201a). In this analysis, to observe the effect of general weathering, we excluded the locations at which decontamination was performed or the locations at which the ground conditions were significantly changed artificially or by other means. As shown in this figure, in undisturbed fields, the deposition density of radiocesium decreased almost in line with radioactive decay both for 137 Cs and 134Cs since the accident. This suggests that in undisturbed fields, radiocesium migration in the horizontal direction is slow. This finding is consistent with the result of a radiocesium migration study conducted in test plots (Yoshimura et al., 2015). Nevertheless, the deposition density decreased by a significantly greater margin upon the inclusion of decontaminated locations in the analysis (Mikami et al., 2018a). In forests, the radiocesium deposited on crowns migrated down to the ground, and at present, most of the deposited radiocesium exists near the ground surface (Kato et al., 2018a). However, radiocesium run-off from the forest was confirmed to be very small. On the contrary, it is known from studies performed after the Chernobyl accident that 134

3.4. Penetration of radiocesium into the ground Two types of typical depth profiles of radiocesium in the ground were observed in Fukushima: an exponential distribution, in which the radiocesium concentration decreases monotonously with depth, and a distribution with a concentration peak at a certain depth (peaked distribution). The latter type of profile was found to fit well to a hyperbolic secant function that asymptotically approaches an exponential function with increasing depth (Matsuda et al., 2015). The proportion of peaked distribution increased with time elapsed. These depth profiles were analyzed theoretically by applying an advection diffusion model considering three different radiocesium states (Kurikami et al., 2016). The model assumed that most radioceisum is strongly bound to fine soil particles, while a part of the bound 5

Journal of Environmental Radioactivity xxx (xxxx) xxxx

K. Saito, et al.

Fig. 2. Cesium-137 deposition density map of eastern Japan as of June 2011 created by integrating the data obtained by in situ spectrometry with a Ge detector (Mikami et al., 2015a) and by helicopter survey (Sanada et al., 2014).

depth profiles. The deposited radiocesium has gradually penetrated the ground with elapsed time. The 90% depth, defined as the depth to which 90% of the deposited radiocesium exists, was calculated for all depth profiles observed in the mapping project. The average 90% depth has increased with elapsed time, but it was less than 5 cm in 2016 (NRA, 2017). However, in many individual cases, the 90% depth exceeded 5 cm, and a certain amount of radiocesium was found at the depth of 10 cm. Hereon, decontamination work must be performed considering these features of depth profiles. In the context of dose evaluation, the extent of radiocesium distribution in the ground should be expressed in terms of the relaxation mass depth β (g/cm2), which is a coefficient of an exponential function. The relaxation mass depth is an important parameter for connecting deposition density to the doses above ground. Dose conversion coefficients have been computed for exponential distribution as a function of the relaxation mass depth in terms of the kerma rate in air (ICRU, 1993), ambient dose equivalent rate (Saito et al., 2014), and effective dose equivalent rate (Saito et al., 2012; Satoh et al., 2016). For the exponential distributions, the relaxation mass depth can be determined directly by means of fitting depth profile to an exponential function. For the profiles approximated using a hyperbolic secant function, the relaxation mass can be determined by fitting to the function. However, the relaxation mass depth in a hyperbolic secant function cannot be used to evaluate the dose with the dose conversion

Fig. 3. Temporal changes in average deposition density of 134Cs and 137Cs in the 80 km zone evaluated from the data obtained in the first campaign of the mapping project (Saito et al., 2015) and those obtained since the second campaign (Mikami et. 2015a, 2018a). Deposition densities were normalized to those in June 2011.

radiocesium could dissolve reversibly and migrate. With this model, we could reconstruct both types of profiles by adjusting conditions. By contrast, the migration of radiocesium along the depth direction is considered a more complex phenomenon in real situations, including bioturbation (Bunzl, 2002) or movement of soil particles, and further investigation is necessary to thoroughly understand the associated 6

Journal of Environmental Radioactivity xxx (xxxx) xxxx

K. Saito, et al.

Fig. 4. Air dose rate maps of different occasions from 2011 to 2016 created by integration of the data obtained by fixed-location measurements in undisturbed fields (Mikami et al., 2015, 2018a) and those by helicopter surveys (Sanada et al., 2014, 2018).

(NRA, 2017). 4. Trend of air dose rate 4.1. Time-dependent air dose rate distribution Fig. 4 shows the air dose rate maps of the 80 km zone from 2011 to 2016. These maps were created by integrating the air dose rates at fixed locations (Mikami et al., 2015b, 2018a) with those obtained by helicopter surveys (Sanada et al., 2014, 2018). The integration was simply performed as follows. If the air dose rate data measured on the ground existed, they were accepted as the representative values of a 1 × 1 km2 area; if not, helicopter survey data were adjusted by multiplication with a correction factor and used instead. As can be seen in the figure, the high dose rate areas indicated by red, orange, and yellow have decreased to a great extent, and the low dose rate areas indicated by blue have become dominant. The obvious decrease in air dose rates can be ascribed mainly to the radioactive decay of radiocesium. Radiocesium movement in the vertical and horizontal directions accelerated this reduction in the dose rate. This will be discussed in the following section.

Fig. 5. Temporal changes in average air dose rates evaluated using the data from car-borne surveys (Andoh et al., 2015, 2018b), fixed-location measurements in undisturbed fields (Mikami et al., 2015b, 2018a), helicopter surveys (Sanada et al., 2015, 2018), and walk surveys (Andoh et al., 2018a). Air dose rates were normalized to those in June 2011.

coefficients computed for exponential distribution. Then, the effective relaxation mass depth βeff should be defined for peaked distributions (Matsuda et al., 2015). By using βeff, the three sorts of doses mentioned in the previous paragraph can be calculated with the dose conversion coefficients computed for exponential distributions. The relaxation mass depth or effective relaxation mass depth has been evaluated for all depth profiles observed in the mapping project, excluding decontaminated locations, and both kinds of relaxation mass depths were generically named “effective relaxation mass depth.” The average effective relaxation mass depth in June 2011 was estimated to be approximately 0.8 g/cm2 but it exceeded 3 g/cm2 in August 2016, and it was found to have increased linearly with elapsed time so far

4.2. Decreasing tendency of average air dose rate The decreasing tendency of the average air dose rate measured by different methods is shown in Fig. 5. The average air dose rates within the 80 km zone shown in this figure were calculated after the background air dose rate due to natural radionuclides was subtracted to observe the decreasing tendency of the air dose rates originating from the deposited radiocesium. The average air dose rate in each measurement was normalized to that in June 2011. Because an enormous amount of data covering the entire 80 km zone was analyzed, this dose rate is considered to reflect the general trend in the zone. A constant 7

Journal of Environmental Radioactivity xxx (xxxx) xxxx

K. Saito, et al.

background air dose rate of 0.05 μSv/h was assumed for the data obtained by fixed-location measurements and helicopter surveys because background data were not available for each location. On the contrary, location-dependent background levels were considered in car-borne surveys, in which the average air dose rates in each municipality were estimated by spectral analysis (Andoh et al., 2017). Differences between the methods for subtracting the background air dose rate did not significantly affect the results because the variation in the background air dose rate was small in comparison with those in the total air dose rate. Various method are available for evaluating the decreasing tendency of air dose rates. It is possible to calculate the temporal attenuation ratio of air dose rate at each measurement point and average them over all locations. Furthermore, the slope of the regression line in scattered plots that present a comparison of the air dose rates in two different measurement periods can be regarded as the attenuation rate. The attenuation tendency was not very sensitive to the evaluation method (Andoh et al., 2018b; Mikami et al., 2018a). In this study, we employed the change in average air dose rate to express the air dose rate attenuation because we considered it the direct indicator of the overall radiation level in the 80 km zone. In five years, the average air dose rate above roads owing to the deposited radiocesium decreased to about 12% of that in June 2011 (red line), that in undisturbed fields to about 18% (dark blue line), and that obtained by helicopter survey to about 22% (light blue line). These values were calculated based on the data obtained in the studies submitted to this special issue by Andoh et al. (2018b), Mikami et al. (2018a), and Sanada et al. (2018), respectively. Because the values are slightly different according to the evaluation method, they should be understood as generous indicators of the decrease in air dose rate after five years. We could not perform a similar analysis of the walk survey because this survey was started in 2013. Nevertheless, the air dose rates in various living environments measured by walk surveys decreased in a manner similar to the air dose rate obtained by car-borne surveys, and the average air dose rate was slightly greater than that obtained by carborne surveys (Andoh et al., 2018a). The air dose rate in Aug 2016 rate was estimated to be 37% of that in June 2011 simply owing to radioactive decay. Furthermore, the air dose rates in purely forested locations decreased in a similar manner owing to radioactive decay (Kato et al., 2018b). Therefore, the air dose rates in human living environments were concluded to have decreased much faster than radioactive decay, to 1/2 to 1/3 of the air dose rate expected from radioactive decay in average. The air dose rate curve obtained by helicopter surveys was between the radioactive decay curve and the curve obtained from fixed-location measurements. This is reasonable considering the fact that in the helicopter surveys, overall gamma rays emitted from various land-use areas, including large forested areas, were detected. Notably, the decreasing tendency shown in Fig. 5 varies greatly according to the conditions. The air dose rates recorded in car-borne surveys and fixed-location measurements demonstrated that the standard deviation of variation according to location was several tens of percentage. This variation is considered to include the variations caused by the fact that measurements were not carried out exactly at the same locations. Careful site-specific analysis will be necessary to clarify the contents of the variation.

Fig. 6. Temporal changes in air dose rates due to the decrease in radiocesium deposition density shown in Fig. 3 (Saito et al., 2015; Mikami et al., 2015a, 2018a) calculated using dose conversion coefficients under the assumption that the depth profile of radiocesium did not change with time. The dotted lines indicate theoretical values due to radioactive decay.

dose rate due to radioactive decay, as well as the contributions of 134Cs and 137Cs to the air dose rate as a function of time elapsed after the accident. The air dose rates calculated from the deposition density shown in Fig. 3 were plotted under the assumption that the relaxation mass depth of the deposited radiocesium did not change, and the dotted lines indicate the theoretical decrease tendency calculated from the physical half-lives. This indicates that the air dose rate in undisturbed fields would have decreased with the physical half-lives if penetration along the depth direction did not occur. The contribution of 134Cs to the air dose rates per unit deposition density is about 2.7 times higher than that of 137Cs, because 134Cs emits multiple gamma rays per disintegration (Saito et al., 2014). The average deposition ratio of 134Cs/137Cs in June 2011 was reported to be 0.91 (Saito et al., 2015b); at that time, about 71% of the air dose rate was attributed to 134Cs. In five years, 134Cs occupancy has decreased greatly due to its half-life, and the contribution of 134Cs to the air dose rate in August 2016 was estimated to be 32% of the total air dose rate. As a result, the total air dose decreased to less than 40% of its initial value after five years. The contribution of 134Cs to the total air dose rate decreased to be lower than that of 137Cs at approximately 1100 d after the accident. Furthermore, radiocesium movement decreased the air dose rates in two different ways. Radiocesium penetration into the ground enhanced the shielding effect of soil against gamma rays emitted in the ground, resulting in a reduction in the air dose rate above ground. This is the main reason why air dose rates in undisturbed fields decreased at a faster rate than that expected from radioactive decay, even when the horizontal radiocesium movement was small. This effect was reported in Ukrainian territories after the Chernobyl accident (Likhtarev et al., 2002) and in Russian territories (Golikov et al., 2002). In the studies cited in the preceding sentence, the attenuation of air dose rates was evaluated based on changes in the depth profile of raduiocesium with the passage of time by applying dose conversion coefficients. The same evaluation was performed after the Fukushima accident based on the depth profiles observed at about 80 locations within the 80 km zone around FDNPP. The evaluation is based on the presumption that 134Cs and 137Cs behave the same ways in the environment. Fig. 7 shows a comparison of the attenuation tendency of air dose rates evaluated after the Fukushima accident with those reported after the Chernobyl accident. The air dose rate reduction tendency due to ground penetration was similar between Ukraine and Russia. However, the pace of reduction has been slightly slower in the case of Fukushima. This is ascribed to the fact that the radiocesium in the soil in Fukushima tended to be adsorbed strongly by clay minerals with high probabilities. In Chernobyl, the situation was fundamentally the same: a large part of radiosecium was adsorbed by clay minerals. Nevertheless, the relatively high proportion of organic matter in the soil inhibited the adsorption of

4.3. Causes of decrease in air dose rate The air dose rate in the Fukushima area decreased because of three main reasons: 1) radioactive decay, 2) radiocesium movement, and 3) decontamination. The decrease in air dose rates owing to radiocesium movement can be classified into two categories: a) vertical penetration into the ground and b) migration along the horizontal direction. Radioactive decay of radiocesium, especially that of 134Cs with a half-life of 2.06 y, has contributed significantly to the decrease in air dose rates in the Fukushima area. Fig. 6 shows the decrease in the air 8

Journal of Environmental Radioactivity xxx (xxxx) xxxx

K. Saito, et al.

4.4.1. Land use One of the most important factors affecting the decrease in air dose rates in the 80 km zone was land use. The car-borne survey data were analyzed in connection with land use data in several studies, and it was found that the decrease in air dose rate in areas classified as forest was slow, that in urban and water areas was fast, and that in farmland was moderate (Andoh et al., 2015; Kinase et al., 2014). Differences in the trends of decrease in air dose rates owing to land use were observed in other analyses including helicopter survey data (Sanada et al., 2018; Wainright et al., 2018). The status of radiocesium binding and its related migration differ depending on land use and the environmental media involved, and these differences induce differences in the pace of dose rate reduction, as discussed before. In the mapping projects, car-borne survey data were averaged over an area of 100 × 100 m2 to reduce statistical fluctuation, and in many cases, multiple types of land use were prevalent in a 100 × 100 m2 area. For example, even if a 100 × 100 m2 area is classified into forest, in many cases, the area contains other environmental media, such as houses and roads. Therefore, the car-borne survey data do not provide representative air dose rates in pure environments, such as pure forest or pure farmland. Even so, the measured data apparently reflect the pace of dose rate reduction in the dominant surrounding environments.

Fig. 7. Evaluated attenuation of air dose rates due to increase in shielding effect with ground penetration of radiocesium for Fukushima and Chernobyl. The attenuation curves were normalized against those in June 2011. The curves for Ukraine and Russia were calculated according to the parameters published by Likhtarev et al. (2002) and Golikov et al. (2002), respectively.

radiocesium by clay minerals (Takahashi et al., 2017). Therefore, it is reasonable that cesium penetration and the associated reduction in dose rate have been slower at Fukushima. Radiocesium migration in the horizontal direction reduced the air dose rates due to decreasing deposition density per unit area. The red line and green areas are considered to lie below the dark blue line in Fig. 5 mainly because of the horizontal movement of radiocesium. We estimated that horizontal movement could reduce the air dose rate by a maximum of 30% in comparison with that without horizontal movement. As already discussed, radiocesium run-off depends strongly on land use, and the difference in air dose rate reduction due to land use could be explained qualitatively based on radiocesium movement. However, it is not easy to analyze quantitatively the effect of horizontal movement at concrete locations, and further investigation is desired. Decontamination could lead to a decrease in the air dose rate in the middle of a decontaminated field by a factor of 2–3 (MOE, 2018) compared to the air dose rates before decontamination. Furthermore, statistical analysis of fixed-location measurements at 6500 locations indicated that the average air dose rate in 2016 at all locations was about 20% lower than that at the non-decontaminated locations. This suggests that the decontamination work performed over wide areas decreased the average air dose rate at fixed locations by 20% or more, if we consider the fact that we could not identify all decontaminated locations in the statistical analysis. The effect of decontamination over wide areas, including those other than the fixed locations, should be investigated further. In summary, the most important cause of decrease in the air dose rate during the five years after the accident was radioactive decay, indicated by the black line in Fig. 5, which resulted in a reduction of more than 60%; the second cause was vertical penetration of radiocesium into the ground, which underlies the difference between the black and dark blue lines, which resulted in a reduction of approximately 25% on an average; the third was radiocesium movement in the horizontal direction, which caused the dose rate reduction below the dark blue line to the red line, resulting in reduction of up to 30%. Largescale decontamination activities certainly contributed to air dose rate reduction in human living environments, that is, depression of the blue and the red lines. This was observed clearly in the fixed-location measurements. It appears that decontamination reduced the average air dose rates in undisturbed fields by 20% or more. Further careful analyses of the effects of site-specific radiocesium movement and decontamination are needed.

4.4.2. Human activity Various human activities affect dose rate reduction. Car-borne survey data indicated that air dose rates outside of the evacuation zone where human activities were not restricted have decreased rapidly in comparison with those within the zone, especially in the early period after the accident (Andoh et al., 2018b). Even in the evacuation zone, the tendencies of decrease in air dose rates differ among the three zones based on their radiation levels and countermeasure policies: 1) difficultto-return zone,1 2) restricted residence area,2 3) zone in preparation for lifting of evacuation order.3 The first zone was affected by high levels of radioactivity, and large-scale decontamination activities have been started there only recently. In the second and third zones, evacuation orders have already been lifted in most areas after large-scale decontamination. In the first zone, the reduction in air dose rate was the slowest. In the second and third zones, the pace of decrease in air dose rates accelerated significantly when large-scale decontamination activities and accompanying human activities were started a few years after the accident (Andoh et al., 2018b). The ratios of the present air dose rate to that in June 2011 were almost the same in the second zone, third zone, and areas outside the evacuation zone. This indicated that human activities are an important factor for altering the reduction in air dose rate. Decontamination activities are a direct cause of dose rate reduction, as discussed in the previous section. Furthermore, various types of human activities are considered to accelerate the reduction in air dose rate. Cultivation of paddy fields and other farmland change vertical distributions of contaminants and generally reduce the air dose rate, in addition to stimulating radiocesium run-off. Car driving appears to be an important factor in accelerating both the removal of radiocesium from roads and other ground surfaces and discharge of the removed radiocesium. It was reported after the Chernobyl accident that the amount of traffic affected the rate of removal of radiocesium deposited on the streets (Anderson et al., 2002). A variety of human activities such as cutting grass, sweeping yards, gardening, and cleaning houses 1 Areas with a possibility that the annual integral radiation dose is over 50 mSv and may remain 20 mSv in five years. 2 Areas in which the estimated radiation dose for one year exceeds 20 mSv but is less than or equal to 50 mSv; entry is allowed for charitable purposes. 3 Areas in which the estimated radiation dose for one year is 20 mSv or less; businesses can resume operations, but it is prohibited to live in these areas or resume farming, forestry, etc.

4.4. Factors affecting decrease in air dose rate A few factors that affect the decrease in air dose rate were found by statistical analysis. These factors were considered to change radiocesium movement in the environment. 9

Journal of Environmental Radioactivity xxx (xxxx) xxxx

K. Saito, et al.

Table 4 Ecological half-lives of air dose rates obtained in different studies. A bi-exponential trend model consisting of fast-decay component and slow-decay component was assumed. The half-lives of the both components were not always determined. Reference

Basic mesurement data

Ecological half-life (y)

Method

Period

Area

Fast

Slow

Kinase (2017)

Car-borne

June 2011–January 2015

Andoh (2018a) Andoh (2018b) Sanada (2018) Wainwright (2018)

Car-borne Walk Helocopter Integrated

June 2011 - Novemer 2016 June 2013–December 2016 April 2011–October 2016 August 2012–October 2016

Inside the evacuation zone 80 km zone excluding the evacuation zone 80 km zone 80 km zone 80 km zone Inside the evacuation zone

1.03–1.10 0.46–0.64 0.44 (not observed) 0.61 (not observed)

92 (assumed) 92 (assumed) 6.7 4.1 57 4.6–5.6

are considered to accelerate radiocesium movement. Quantitative analysis of the effects of these human activities is an important challenge to be resolved in the future.

analyzed separately according to the conditions. The initial dose rate distribution in the targeted region was determined fundamentally by car-borne surveys, and in the areas for which car-borne survey data do not exist, helicopter survey data were used after simple correction. The advantage of this predictive model is that the parameters are based on a massive amount of monitoring data covering a wide area and several years. Therefore, it is anticipated that the predictions would be more realistic than those obtained in the past based on smaller amounts of environmental data. The predicted air dose rates were validated in different ways and were found to coincide with the observed data within expected uncertainties. According to a recent analysis (Kinase et al., 2017), the ecological half-life of the short-term component Tshort varies little with land use, while the fraction of the short-term component αshort varies considerably with land use. This indicates that the pace of air dose rate reduction in a 100 × 100 m2 area changes depending mostly on how wide an area with the short-term component is included in the said area. This is a natural deduction because areas with different land use types must be included in the 100 × 100 m2.

4.4.3. Season There is some evidence that a decrease in air dose rate due to weathering occurred at a specific period in a year (Mikami et al., 2018a; Yoshida et al., 2015). It is well known that snow cover significantly reduces air dose rates owing to increased shielding effects. Furthermore, an increase in the water content of soil due to rain or snow reduces air dose rates. However, in such cases, the reduction is temporary, and the air dose rates return to their original levels when the environmental conditions are restored. In measurements at fixed locations from spring to summer, some irreversible changes in air dose rates were observed, and these were considered non-temporary fluctuations in air dose rates. These changes in air dose rates were not related directly to the amount of precipitation, and the underlying cause was unclear. Radiocesium penetration may have occurred in this period, especially because of bioturbation or other reasons. 4.4.4. Indoor air dose rate Air dose rates inside houses in contaminated areas are generally lower than those outside houses, and the dose rate ratio between inside and outside is defined as the dose reduction factor. The dose reduction factor is important from the viewpoint of evaluation of dose exposure to inhabitants, since people spend a large proportion of time inside houses. A dose reduction factor of 0.4 has been widely used for dose evaluation in houses in Fukushima according to IAEA (1979). This value was validated in dwellings in Fukushima (Matsuda et al., 2017; Yoshida-Ouchi et al., 2014, 2018). The average dose reduction factor was found to be close to 0.4 if the contribution of natural radiation was excluded. However, the fluctuations due to conditions were so large that it was difficult to estimate indoor air dose rates accurately by using this factor. It is important to develop a method to estimate the indoor air dose rates more accurately.

5.2. Ecological half-life in complex environment Ecological half-lives of air dose rates have been evaluated by multiple researchers in ways similar to that in the prediction model described in the previous section. They employed a model with two exponential functions indicating a fast-decay component and a slowdecay component, and ecological half-lives, which are the dominant parameters of the exponential functions, were determined from an analysis of environmental monitoring data. In recent years, it has become possible to evaluate the ecological half-life of the long-term component Tlong because the current trend of air dose rates is considered to be subject to mainly the long-term component. Table 4 summarizes the ecological half-lives determined after the FDNPP accident in complex environments where diverse land use types exist together, that is, the actual environment. Kinase et al. found that the mean value of Tshort was 0.4–0.6 outside the evacuation zone and 1.0–1.1 inside the zone (Kinase et al., 2017) based on the car-borne survey data obtained until 2014. These values of Tshort seem to be consistent across studies. In any study in which environmental data were analyzed, including those corresponding to a few months after the accident, the presence of a fairly fast air dose decay component with a half-life of less than 1 y was confirmed. In the analysis of Kinase et al. (2014, 2017), the ecological half-life of the long-term component Tlong was fixed to 92 y based on a past study by Gale et al. (1964) since the duration covered by the car-borne survey data at the time of evaluation was not sufficiently long to determine the half-life of the long-term component. It was concluded by parametric research that the predicted air dose rates were not very sensitive to changes in the value of Tlong. By contrast, the results of most studies based on environmental data obtained in recent years suggest that Tlong may be shorter than that assumed by Kinase et al. (Andoh et al., 2018a, 2018b; Wainwright

5. Extended analysis of obtained environmental data 5.1. Prediction of air dose rate distribution An empirical model for predicting the trend of air dose rates was developed based on the statistical analysis of an enormous amount of car-borne survey data (Kinase et al., 2014, 2017). In the developed model, it was assumed that the average air dose rate in a 100 × 100 m2 area decreases in accordance with a formula containing two exponential functions that represent fast and slow ecological decay components other than radioactive decay of radiocesium due to the physical half-lives of 134Cs and 137Cs. The formula was fitted to timedependent air dose rates in a 100 × 100 m2 area. The fitting was performed for a large amount of air dose rate data, and the determined parameters, namely, ecological half-life of the short-term component Tshort and the fraction of the short-term component αshort, were 10

Journal of Environmental Radioactivity xxx (xxxx) xxxx

K. Saito, et al.

et al., 2018), that is, of the order of several years. This may demonstrate the possibility that air dose rates in complex environments tend to decrease according to a tri-exponential function rather than a bi-exponential function. In fact, the car-borne survey data, including the latest ones, indicate the tendency that the reduction in air dose rate is slowing down. The analysis of airborne data yielded a Tlong value of 57 y, which is closer to the 92 y assumed previously. Airborne monitoring measures the average air dose rate over a wide area, including forested areas, and it may be difficult to separate intermediate- and long-term components, even if they exist. The ecological half-lives need to be investigated further by using new monitoring data.

contributed to an air dose rate reduction of approximately 25% on average. Furthermore, radiocesium movement in the horizontal direction has accelerated the decrease by 30% at maximum. The effect of decontamination was clearly observed in the fixed-location measurements, which was estimated to be 20% or more. However, careful quantitative analyses of decontamination effects must be performed in the future. Enormous amounts of accumulated environmental data have been analyzed extensively. A predictive model of air dose rates in the Fukushima region was developed using an empirical formula, the parameters of which were determined by statistical analysis of enormous amounts of repeatedly obtained car-borne survey data. This statistical analysis made it possible to predict realistic air dose rates with appropriate confidence intervals. The analysis should be revised whenever new air dose rate sets are obtained to improve the accuracy of predictions. A data integration method was developed by utilizing Bayesian statistics to combine the air dose rate data obtained by helicopter, car-borne, and walk surveys with different specific features. The method facilitated the creation of air dose rate maps with high accuracy and resolution. The environmental data obtained in the mapping and helicopter survey projects have been released to the public as maps, numerical data, and graphs of analyzed data through the Internet. An overall picture of the temporal changes in contamination conditions was clarified by successive large-scale environmental monitoring and the extensive analyses. Even so, many challenges must be overcome to extract important information from the accumulated noble data, some of which were discussed in this paper.

5.3. Integration of air dose rate maps The integrated air dose rate maps created using helicopter survey data and fixed-location measurement data are shown in Fig. 4. Furthermore, a trial of the integration of air dose rate maps using a sophisticated method was launched in recent years (Wainwright et al., 2017a). The new trial employed multiscale Bayesian data integration and integrated monitoring data from helicopter, car-borne, and walk surveys. Walk survey data were presumed to express the air dose rates directly related to human inhabitation, and the integration aimed to construct the air dose rate map that would be obtained if walk surveys were performed over the entire targeted area. In the integration, the air dose rate distribution for which the contingent probability reached a maximum was determined with confidence intervals under the condition that certain aerial monitoring data, car-borne survey data, and walk survey data were obtained in the targeted area. The integration was first tested with the data obtained in Fukushima City (Wainwright et al., 2017a). Then, the target area was extended to the entire evacuation area (Wainwright et al., 2017b). In the paper published in this special issue (Wainwright et al., 2018), integration was performed for the data from 2014 to 2016, and it was demonstrated that time-dependent changes can be discussed precisely for the entire area. These types of integrated maps can offer better insights into the contamination conditions in the Fukushima region by assembling different pieces of information with different specific features.

Acknowledgment The data shown in this study were mostly obtained from the projects funded by the Nuclear Regulation Authority of Japan (NRA) and the Japanese Ministry of Education, Science, Sports and Technology (MEXT). We would like to express our sincere gratitude to Prof. T. Nakamura, Prof. N. Momoshima, and the other members of the ad-hoc committee of the Fukushima mapping projects. The noble environmental data discussed in the present study were obtained because of many people's efforts. We appreciate everyone who have helped and supported us, and thank them for contributing in various ways to conduct our research productively.

6. Conclusions We summarized temporal changes in deposited radionuclides and air dose rates over five years after the Fukushima NPP accident based on large-scale environmental monitoring performed repeatedly by using reliable methods. Multiple radionuclides, including radioactive strontium and plutonium, were detected over a wide area in June 2011; however, radiocesium was found to be far more important than the other radionuclides from the viewpoint of long-term exposure. It was estimated that about 2 PBq of 137Cs was deposited on the ground in Japan. Radiocesium movement in the environment has been slow in general, and the average deposition density of both 134Cs and 137Cs in undisturbed fields has decreased almost in line with the pace of radioactive decay. Radiocesium has gradually penetrated the ground; nevertheless, in half of the cases, more than 90% of radiocesium existed within 5 cm from the ground surface in August 2016. The air dose rates in the 80 km zone have decreased noticeably, and the trends differ remarkably depending on the environment. The average air dose rate due to radiocesium in undisturbed fields after five years was approximately 18% of that in June 2011 and approximately 12% above roads, as measured by car-borne surveys. In forests, the air dose rates have decreased approximately in line with the pace of radioactive decay, that is, to 37% of the initial dose rate. Therefore, in the various environments inhabited by humans, the air dose rates decreased to 1/2–1/3 compared with those simply due to radioactive decay. The most important cause of air dose rate reduction so far is radioactive decay of radiocesium. In addition, an increase in gammaray shielding due to radiocesium penetration into the ground has

References Andersson, K.G., Roed, J., Fogh, C.L., 2002. Weathering of radiocaesium contamination on urban streets, walls and roofs. J. Environ. Radioact. 62, 49–60. Andoh, M., Nakahara, Y., Tsuda, S., Yoshida, T., Matsuda, N., Takahashi, F., Mikami, S., Kinouchi, N., Sato, T., Tanigaki, M., Takamiya, K., Sato, N., Okumura, R., Uchihori, Y., Saito, K., 2015. Measurement of air dose rates in wide area around the Fukushima Daiichi nuclear power plant through a series of car-borne surveys. J. Environ. Radioact. 139, 266–280. Andoh, M., Matsuda, N., Saito, K., 2017. Evaluation of ambient dose equivalent rates owing to natural radioactive nuclides in eastern Japan by car-borne surveys using KURAMA–II. Trans. Atom. Energy Soc. Jpn. 16, 63–80 ([in Japanese]). Andoh, M., Mikami, S., Tsuda, S., Yoshida, T., Matsuda, N., Saito, K., 2018a. Measurement of air dose rates by walk survey around the Fukushima Dai-ichi Nuclear Power Plant using KURAMA-II until 2016. J. For. Environ. Radioact. 190-191, 111–121. Andoh, M., Mikami, S., Tsuda, S., Yoshida, T., Matsuda, N., Saito, K., 2018b. Decreasing trend of ambient dose equivalent rates over a wide area in eastern Japan until 2016 evaluated by car-borne surveys using KURAMA Systems. J. Environ. Radioact. 192, 385–398. Bunzl, K., 2002. Transport of fallout radiocesium in the soil by bioturbation: a random walk model and application to a forest soil with a high abundance of earthworms. Sci. Total Environ. 293, 191–200. Chino, M., Terada, H., Naga, i H., Katata, G., Mikami, S., Torii, T., Saito, K., Nishizawa, Y., 2016. Utilization of 134Cs/137Cs in the environment to identify the reactor units that caused atmospheric releases during the Fukushima Daiichi accident. Sci. Rep. https://doi.org/10.1038/srep31376. 2016. Funaki, H., Yoshimura, K., Sakuma, K., Iri, S., Oda, Y., 2018. Evaluation of particulate 137 Cs discharge from a mountainous forested catchment using reservoir sediments and sinking particles. J. Environ. Radioact. 189, 48–56. Gale, H.L., Humphreys, D.L.O., Fisher, E.M.R., 1964. Weathering of caesium-137 in soil.

11

Journal of Environmental Radioactivity xxx (xxxx) xxxx

K. Saito, et al. Nature 4916, 257–261. Golikov, V.Yu, Balonov, M.I., Jacob, P., 2002. External exposure of the population living in areas of Russia contaminated due to the Chernobyl accident. Radiat. Environ. Biophys. 41, 185–193. IAEA (International Atomic Energy Agency), 1979. Planning for Off-site Response to Radiation Accidents in Nuclear Facilities. IAEA-TECDOC-225, Vienna, Austria. IAEA (International Atomic Energy Agency), 2000. Generic Procedures for Assessment and Response during a Radiological Emergency. IAEA-TECDOC-1162, Vienna, Austria. IAEA (International Atomic Energy Agency), 2015. The Fukushima Daiichi Accident. IAEA ISBN:978-92-0-107015-9. ICRU (International Commission on Radiation Units and Measurements), 1993. Gammaray Spectrometry in the Environment. ICRU Report 53, Bethesda, USA. . Iwagami, S., Onda, Y., Tsujimura, M., Abe, Y., 2017. Contribution of radioactive 137Cs discharge by suspended sediment, coarse organic matter, and dissolved fraction from a headwater catchment in Fukushima after the Fukushima Dai-ichi Nuclear Power Plant accident. J. Environ. Radioact. 166, 466–474. Katata, G., Chino, M., Kobayashi, T., Terada, H., Ota, M., Nagai, H., Kajino, M., Draxler, R., Hort, M.C., Malo, A., Torii, T., Sanada, Y., 2015. Detailed source term estimation of the atmospheric release for the Fukushima Daiichi Nuclear Power Station accident by coupling simulation of an atmospheric dispersion model with an improved deposition scheme and oceanic dispersion model. Atmos. Chem. Phys. 15, 1029–1070. Kato, H., et al., 2018a. Six-year monitoring study of radiocesium transfer in forest environments following the Fukushima nuclear power plant accident. J. Environ. Radioact (submitted for publication). Kato, H., et al., 2018b. Regional survey of the ambient dose rate in the forest environments of Fukushima Prefecture following the Fukushima reactor accident. J. Environ. Radioact (submitted for publication). Kinase, S., Sato, S., Takahashi, T., Sakamoto, R., Saito, K., 2014. Ecological Half-life of Radioactive Caesium within the 80 Km Radius of the Fukushima Daiichi Nuclear Power Plant. IRPA2014 Abstract Book 163-166. Kinase, S., Takahashi, T., Saito, K., 2017. Long-term predictions of ambient dose equivalent rates after the Fukushima Daiichi nuclear power plant accident. J. Nucl. Sci. Technol. 54, 1345–1354. Kurikami, H., Malins, A., Takeishi, M., Saito, K., Iijima, K., 2016. Coupling the advectiondispersion equation with fully kinetic reversible/irreversible sorption terms to model radiocesium soil profiles in Fukushima Prefecture. J. Environ. Radioact. 177, 1–12. Likhtarev, I.A., Kovgan, L.N., Jacob, P., 2002. Chernobyl accident: external exposure of Ukrainian population due to radioactive contamination of territory (retrospective and prospective dose estimations). Health Phys. 82, 290–303. Malins, A., Okumura, M., Machida, M., Takemiya, H., Saito, K., 2015. Field of view for environmental radioactivity. In: Proceedings of the 2015 International Symposium on Radiological Issues for Fukushima's Revitalized Future, Paruse Iizaka, Fykushima City, Japan, May 30-31. Matsuda, N., Mikami, S., Shimoura, S., Takahashi, J., Nakano, M., Shimada, K., Uno, K., Hagiwara, S., Saito, K., 2015. Depth profiles of radioactive cesium in soil using a scraper plate over a wide area surrounding the Fukushima Dai-ichi Nuclear Power Plant. J. Environ. Radioact. 139, 427–434. Matsuda, N., Mikami, S., Tetsuro, S., Saito, K., 2017. Measurement of air dos rates in and around houses in the Fukushima Prefecture in Japan after the Fukushima accident. J. Environ. Radioact. 166, 427–435. Mikami, S., Maeyama, T., Hoshide, Y., Sakamoto, R., Sato, S., Okuda, N., Demongeot, S., Gurriaran, R., Uwamino, Y., Kato, H., Fujiwara, M., Sato, T., Takemiya, H., Saito, K., 2015a. Spatial distributions of radionuclides deposited onto ground soil around the Fukushima Dai-ichi Nuclear Power Plant and their temporal change until December 2012. J. Environ. Radioact. 139, 320–343. Mikami, S., Maeyama, T., Hoshide, Y., Sakamoto, R., Sato, S., Okuda, N., Sato, T., Takemiya, H., Saito, K., 2015b. The air dose rate around the Fukushima Dai-ichi Nuclear Power Plant: its spatial characteristics and temporal changes until December 2012. J. Environ. Radioact. 139, 250–259 2015. Mikami, S., et al., 2018a. The deposition densities of radiocesium and the air dose rates in undisturbed fields around the Fukushima Dai-ichi Nuclear Power Plant; their temporal changes for five years after the accident. J. Environ. Radioact (submitted for publication). Mikami, S., et al., 2018b. Guidance for in situ spectrometry intercomparison based on the information obtained through five intercomparison during the Fukushima mapping project. J. Environ. Radioact (submitted for publication). Ministry of Environment (MOE), 2018. The transition of the spatial dose rate after decontamination model demonstration project (in Japanese). http://josen.env.go.jp/ area/pdf/transition_180323.pdf, Accessed date: 25 September 2018. Muramatsu, Y., Matsuzaki, H., Toyama, C., 2015. Analysis of 129I in the soils of Fukushima prefecture: preliminary reconstruction of 131I deposition related to the accident at Fukushima Daiichi nuclear plant (FDNPP). J. Environ. Radioact. 139, 344–350. NRA (Nuclear Regulation Authority), 2014. Report on radioactive substance distribution mapping project in FY2013 (in Japanese). http://radioactivity.nsr.go.jp/ja/contents/ 10000/9735/24/cover.pdf, Accessed date: 26 October 2017. NRA (Nuclear Regulation Authority), 2017. Report on radioactive substance distribution mapping project in FY2016 (in Japanese). http://radioactivity.nsr.go.jp/ja/contents/ 10000/9735/24/cover.pdf, Accessed date: 26 October 2017.

Nursal, N.I., Okuda, T., Yamada, T., 2016. Potential of Crowdsourcing approach on monitoring radioactivity in Fukushima prefecture. Global Environ. Res. 20, 111–117. Onda, Y., Kato, H., Hoshi, M., Takahashi, K., Nguyen, M.L., 2015. Soil sampling and analytical strategies for mapping fallout in nuclear emergencies based on the Fukushima Dai-ichi Nuclear Power Plant accident. J. Environ. Radioact. 139, 300–307. Saito, K., Onda, Y., 2015. Outline of the national mapping projects implemented after the Fukushima accident. J. Environ. Radioact. 139, 240–249. Saito, K., Petoussi-Henss, N., 2014. Ambient dose equivalent conversion coefficients for radionuclides distributed in the ground. J. Nucl. Sci. Technol. 51, 1274–1287. Saito, K., Ishigure, N., Petoussi-Henss, N., Schlattl, H., 2012. Effective dose conversion coefficients for radionuclides exponentially distributed in the ground. Radiat. Environ. Biophys. 51, 411–423. Saito, K., Onda, Y., Hisamatsu, S., 2015a. SPECAL ISSUE: Japanese national projects on large-scale environmental monitoring and mapping in Fukushima Volume 1 (Edt.). J. Environ. Radioact. 139, 240–434. Saito, K., Tanihata, I., Fujiwara, M., Saito, T., Shimoura, S., Otsuka, T., Onda, Y., Hoshi, M., Ikeuchi, Y., Takahashi, F., Kinouchi, N., Saegusa, J., Seki, A., Takemiya, H., Shibata, T., 2015b. Detailed deposition density maps constructed by large-scale soil sampling for gamma-ray emitting radioactive nuclides from the Fukushima Daiichi Nuclear Power Plant accident. J. Environ. Radioact. 139, 308–319. Saito, K., Onda, Y., Hisamatsu, S., 2016a. SPECAL ISSUE: Japanese national projects on large-scale environmental monitoring and mapping in Fukushima Volume 2 (Edt.). J. Environ. Radioact. 166, 417–474. Saito, K., Yamamoto, H., Mikami, S., Andoh, M., Matsuda, N., Kinase, S., Tsuda, S., Yoshida, T., Matsumoto, S., Sato, T., Seki, A., Takemiya, H., 2016b. Radiological conditions in the environment around the Fukushima Daiichi nuclear power plant site. Global Environ. Res. 20, 15–22. Sanada, Y., Torii, T., 2015. Aerial monitoring around the Fukushima Dai-ichi nuclear power plant using an unmanned helicopter. J. Environ. Radioact. 139, 294–299. Sanada, Y., Sugira, T., Nishizawa, Y., Kondo, A., Torii, T., 2014. The aerial radiation monitoring in Japan after the Fukushima Daiichi nuclear power plant accident. J. Prog. In Nucl. Sci. Technol. 4, 76–80. Sanada, Y., Urabe, Y., Sasaki, M., Ochi, K., Torii, T., 2018. Evaluation of ecological halflife of dose rate based on airborne radiation monitoring following the Fukushima Daiichi nuclear plant accident. J. Environ. Radioact. 192, 417–425. Satoh, D., Furuya, T., Takahashi, F., Endo, A., Lee, C., Bolch, W.E., 2016. Age-dependent dose conversion coefficients for external exposure to radioactive cesium in soil. J. Nucl. Sci. Technol. 53, 69–81. Takahashi, Y., Fan, Q., Suga, H., Tanaka, K., Sakaguchi, A., Takeichi, Y., Ono, K., Mase, K., Kato, K., Kanivets, V.V., 2017. Comparison of soil-water partitions of radiocesium in river waters in Fukushima and Chernobyl areas. Sci. Rep. 7, 21407. Tanigaki, M., Okumura, R., Takamiya, K., et al., 2015. Development of KURAMA-II and its operation in Fukushima. Nucl. Instrum. Methods Phys. Res. 781, 57–64. Tsuda, S., Yushida, T., Tsutsumi, M., Saito, K., 2015. Characteristics and verification of a car-borne survey system for dose rates in air: KURAMA-II. J. Environ. Radioact. 139, 260–265. Tsutsumi, M., Saito, K., Moriuchi, S., 1991. Spectrum-dose Conversion Operators, G(E) Functions of NaI(Tl) Scintillators Adopted for Effective Dose Equivalent Quantities. Japan Atomic Energy Research Institute JAERI-M 91-204. Wainwright, H.M., Seki, A., Chen, J., Saito, K., 2017a. A multiscale Bayesian data integration approach for mapping air dose rates around the Fukushima Daiichi Nuclear Power Plant. J. Environ. Radioact. 167, 62–69. Wainwright, H.M., Seki, A., Chen, J., Saito, K., 2017b. A multiscale Bayesian data integration approach for mapping air dose rates around the Fukushima Daiichi Nuclear Power Plant. In: Proceeding of WM2017 Conference, March 5-9, 2017, Poenix, Arisona, USA. Wainwright, H.M., Seki, A., Mikami, S., Saito, K., 2018. Characterizing regional-scale temporal evolution of air dose rates after the Fukushima Daiichi Nuclear Power Plant accident. J. Environ. Radioact. 189, 213–220. Wakiyama, Y., et al., 2018. Land use type controls the wash-off rate, entrainment coefficient, and time dependence of Fukushima-derived cesium-137. J. Environ. Radioact (submitted for publication). Yoshida, H., et al., 2015. Seasonal change in the reduction of the air dose rate observed by daily measurements. In: 2015 Fall Meeting. Atomic Energy Society of Japan (in Japanese). Yoshida-Ohuchi, H., Hosoda, M., Kanagami, T., Uegaki, M., Tashima, H., 2014. Reduction factor for wooden houses due to external g-radiation based on in situ measurements after the Fukushima accident. Sci. Rep. 4, 7541. Yoshida-Ohuchi, H., Matsuda, N., Saito, K., 2018. Review of reduction factors by buildings for gamma radiation from radiocesium deposited on the ground due to fallout. J. Environ. Radioact. 187, 32–39. Yoshimura, K., Onda, Y., Kato, H., 2015. Evaluation of radiocesium wash-off by soil erosion from various land use using USLE plots. J. Environ. Radioact. 139, 362–369. Yoshimura, K., Onda, Y., Wakahara, T., 2016. Time dependence of the 137Cs concentration in particles discharged from rice paddies to freshwater bodies after the Fukushima Daiichi NPP accident. Environ. Sci. Technol. 50, 4186–4193. Yoshimura, K., Saito, K., Fujiwara, 2017. Distribution of 137Cs on components in urban area four years after the Fukushima Dai-ichi Nuclear Powere Plant accident. J. Environ. Radioact. 178–179, 48–54.

12