Assessment of gamma radiation from a limited area of forest floor using a cumulative personal dosimeter

Journal of Environmental Radioactivity 204 (2019) 95–103

Contents lists available at ScienceDirect

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

Assessment of gamma radiation from a limited area of forest floor using a cumulative personal dosimeter

T

Toshihiro Yoshiharaa,∗, Keisuke Kuritab, Hideyuki Matsumuraa, Vasyl Yoschenkoc, Naoki Kawachib, Shin-Nosuke Hashidaa, Alexei Konoplevc, Hirohisa Yoshidad a

Environmental Science Research Laboratory, Central Research Institute of Electric Power Industry (CRIEPI), 1646 Abiko, Abiko, Chiba, 270-1194, Japan Takasaki Advanced Radiation Research Institute, National Institutes for Quantum and Radiological Science and Technology (QST), 1233 Watanuki, Takasaki, Gunma, 370-1292, Japan c Institute of Environmental Radioactivity of Fukushima University (IER), 1 Kanayagawa, Fukushima, Fukushima, 960-1296, Japan d Urban Environmental Sciences, Tokyo Metropolitan University (TMU), 1-1 Minami-Osawa, Hachioji, Tokyo, 192-0397, Japan b

A R T I C LE I N FO

A B S T R A C T

Keywords: Forest floor Soil surface Ground Radiocesium Fukushima D-shuttle

To elucidate long term changes in gamma radiation from a limited region of interest of the forest floor, a simple monitoring procedure using a cumulative personal dosimeter (D-shuttle) was examined from 2016 to 2017. The test site was in a small forest in Abiko, Japan, where the initial radiocesium contamination from the Fukushima Dai-ichi Nuclear Power Plant was 60–100 kBq m−2. Three experimental plots basically containing a set of two 5 × 5 m2 observation areas were arranged at the site. The litterfall and decomposing organic layer of one area (D: decontaminated) were fully eliminated before the monitoring, whereas the other area (N: natural) was left unchanged. Five D-shuttle sets (i.e., D-shuttle, lead shield, and holder) per area were set up. One D-shuttle set could monitor the specific gamma radiation from radiocesium distributed within a limited area of ground (0.5 m radius of circle = ca. 0.8 m2 area of flat ground). The results indicated significant differences in the accumulated doses among each of the plots and areas, reflecting their soil radiocesium inventories. Interestingly, every index decreased with time, but the decreases were slower than the theoretical decay of radiocesium (134Cs and 137Cs). In addition, the accumulated dose decreased during heavy rainfall events. One possible explanation for these changes of the accumulated dose is a combination of meteorological and tree phenological phenomena, such as radiocesium from the forest canopy being newly added to the floor primarily by litterfall and soil moisture content disturbing radiation emitted from soils. This simple procedure enables long-term observation of gamma radiation from a limited area of forest floor non-invasively and semi-quantitatively.

1. Introduction The accident at the Fukushima Dai-ichi Nuclear Power Plant (FDNPP; 37.419°N, 141.022°E), following a huge earthquake on March 11, 2011, dispersed radionuclides across large areas of northern Japan. Since then, decontamination programs have been implemented hurriedly in both residential and agricultural areas. As a consequence of the integration of the decontamination efforts, physical decay, and meteorological/biological attenuation, the air dose in the contaminated areas has decreased drastically (MOE, 2012; Mikami et al., 2015). By April 2017, the gross area of evacuation zones had also been reduced to one-third (371 km2) of that in August 2013 (1149 km2; FRS, 2018). However, forest areas, except those within the vicinity of residential

areas (i.e., within 20 m), have been excluded from the decontamination plan (MOE, 2012), and this may be a major reason why returnees still feel considerable uneasiness about the fate of radionuclides in the natural environment. Possible discharge of radionuclides into drainage basins and/or drifts from decomposed particles of plant debris in flowing water are a large concern (Nagao et al., 2013; Forestry Agency, 2014; Sakuma et al., 2017). In addition, a phenomenon of woody plants recycling radiocesium in their biomass and/or accumulating it via roots from soil is not fully understood (Yoschenko et al., 2016; Imamura et al., 2017). Therefore, there is a need to discern small-scale topographical translocations/transitions of radio contaminants, mainly within soil particles, litterfall, and litterfall-derived organic matter. However, the majority of the current monitoring systems applicable to

Abbreviations:FDNPP, Fukushima Dai-ichi Nuclear Power Plant ∗ Corresponding author. E-mail address: [email protected] (T. Yoshihara). https://doi.org/10.1016/j.jenvrad.2019.03.023 Received 12 December 2018; Received in revised form 22 March 2019; Accepted 22 March 2019 0265-931X/ © 2019 Elsevier Ltd. All rights reserved.

Journal of Environmental Radioactivity 204 (2019) 95–103

T. Yoshihara, et al.

least 20 cm above the soil surface using a holder (Fig. 1(B)). The set of the D-shuttle, lead shield, and holder was designated as “the D-shuttle set” in this study. The lead shield (L130 × W95 × H45 mm3) included a cubic hole (L70 × W35 × D15 mm3) for mounting one D-shuttle (Fig. 1(D)). The lead shield had a thickness of at least 30 mm in any direction, meaning that approximately 98% of the direct gamma radiation from the covered sphere could be screened (Papagiannis et al., 2008). The holder was made of 1 mm thick stainless steel and shaped to a size (L130 × W97 × H41 mm3) suitable for the lead shield (Fig. 1(B)). The holder had four pillars (H265 mm × Φ8 mm) and a portion of each pillar was pegged into the soil to stabilize it (Fig. 1(C)). The D-shuttle was wrapped with two water-resistant plastic bags before (Fig. 1(D)) and after mounting the lead shield (Fig. 1(C)). All D-shuttles were calibrated before and after the monitoring to confirm the expected performance by the maker, Chiyoda Technol Corp.

such needs are based on measurements obtained using a Ge-semiconductor detector. Although the systems provide accurate data, they require physical samples for every measurement, which disturb the targeted trees and/or the soils. Furthermore, implementing such systems is labor intensive and expensive. Soon after the FDNPP accident in 2011, a personal electronic cumulative dosimeter (D-shuttle; Chiyoda Technol. Corp., Japan) was commercialized. The D-shuttle can measure γ-radiation with considerable accuracy at very fine resolution as an external exposure dose rate (> 0.1 μSv h−1) and store the data for more than 12 months (Čemusová et al., 2017; CTC, 2018). The data are recorded as at least hourly cumulative doses and can be read for discretionary durations. The Dshuttle has already been practically utilized to measure personal dosage in Fukushima and eleswhere (Adachi et al., 2016; Naito et al., 2015, 2016). Although the D-shuttle was developed to measure personal dosage, Murayama (2018) indicated the possibility of using it for measurement of environmental radiation instead of a passive dosimeter or monitoring post. The major question pertaining to the use of the Dshuttle for environmental monitoring is whether it can be applied to monitor the long-term transition of radionuclides in a specific area/ object in nature. In this regard, we have examined the detection of specific γ-radiation from living trees by a model experiment (Kurita et al., 2018) and a greenhouse experiment (Yoshihara et al., 2016a). Dshuttle measurements demonstrated specific changes in a limited observation area of catchment soil, such as a slight increase due to accumulation of contaminated sediments deposited during flooding and sharp reductions caused by slope erosion processes associated with heavy rainfall (Konoplev et al., 2018). Here, we present the results of a yearlong continuous measurement of γ-radiation using the D-shuttle in a small forest in Abiko, Japan, where radiocesium contamination from the FDNPP was apparent but comparatively lower than in areas closer to the plant. Our observation area was divided into two parts according to whether the litter layer was eliminated to determine whether the effect of decontamination and continuity would be discernible via this method. The radiocesium distribution in the soil, rainfall events, and tree phenology were considered together with the monitoring results to deduce the factors of changes in the accumulated dose.

2.3. Setup of observation areas and soil sampling From August 26, 2016 to August 9, 2017, three duplicates of monitoring plots containing a set of two 5 × 5 m2 observation areas were arranged at the test site (Figs. 1(E), 2 and 5). The litter layer and superficial decomposing organic layers of one area, designated as “D” (decontaminated), were fully eliminated before the monitoring, whereas the other area, designated as “N” (natural), was left in its natural condition. One area each in Plot 1 and 2 was allocated to the D and N areas, respectively; however, both areas in Plot 3 were left as N areas to determine the variation between neighboring areas without decontamination (Fig. 5). The upper and lower sides of these areas were partitioned with 50 cm high plywood boards (30 cm height above ground surface after setting up) to prevent soils and litterfalls from flowing in or out. Five D-shuttle sets were placed in two rows (i.e., two and three sets in the upper and lower rows, respectively) on the ground of each area (Figs. 2 and 5). However, even though five D-shuttle sets were placed in each area at the start of the monitoring, at most three of these (marked with crosses in Fig. 5) were not able to read data due to an accident (e.g., water-logging or tumbling) at the end of the monitoring and their data were omitted from the analyses. After the observations, soils up to 25 cm depth, litter layer, and superficial decomposing organic layers just beneath the D-shuttle sets were sampled using a soil core sampler (DIK-110C, Daiki Rika Kogyo Co., Ltd., Tokyo). Five replicates per area were obtained. Before the measurement of their radiocesium concentrations, the soils between 0 and 5 cm depth were separated into five parts (1 cm depth each), and those beneath 5 cm depth were separated into four parts (5 cm depth each). Soils and organic layers were screened with a 2 mm sieve, and size fractions of less than 2 mm were used as samples. Litter layers were pulverized with a mill. After drying for at least 3 d in an oven at 65 °C, every sample was weighed and measured for concentration of radiocesium.

2. Materials and methods 2.1. Test site The test site was in Abiko (35.87815° N, 140.02487° E, Laboratory of Environmental Science, CRIEPI, total area of 17.3 ha), approximately 200 km SSW of the FDNPP. Initial radionuclide fallout in Abiko was observed on March 16, 2011, in the form of dry deposition. However, the majority of the fallout was observed on March 21, 2011, occurring with rainfall. In total, 60–100 kBq m−2 of radiocesium (134Cs and 137 Cs) was recorded at Abiko following the latter deposition event (Morino et al., 2011; Terada et al., 2012; Doi et al., 2013). Additional details about the location are available in our previous reports (Yoshihara et al., 2013, 2014a, 2014b). The area has a moderate monsoon climate. The average of the 2013 and 2017 total annual precipitation and the average daily temperature were 1410.0 mm and 14.7 °C, respectively (JMA, 2018). The basal soil type was pale Ando soil overlain by a thin organic layer (NLA, 1990). The site was located on a slope (maximum gradient of 19.3–35°) in a small forest covered with densely planted Cryptomeria japonica (Plot 1) or a mixture of tree species (e.g., Pinus densiflora, C. japonica, Quercus myrsinifolia, and Eurya japonica) and understories (e.g., Fatsia Japonica, Sasa spp., Aucuba japonica, and Parthenocissus spp.) (Plots 2 and 3).

2.4. Radiocesium analysis The dried soil and litterfall samples were homogenized, weighed, and packed into screw-capped polystyrene containers (Φ56 mm × H68 mm, U-8 container, As-one Co. Ltd., Osaka, Japan) for radiocesium concentration measurements (MEXT, 1976). We made one (i.e., no replicate) 3600 s measurement per sample, unless data were beneath the detection limit calculated by Cooper's method (MEXT, 1992). For samples that were not detectable by the 3600 s measurement, an additional 30000 s measurement was conducted. A Ge-semiconductor detector (GEM20-70, ORTEC, Oak Ridge, TN, US) coupled with a multi-channel analyzer (MCA7600, SEIKO EG&G, Tokyo, Japan) was used to evaluate radiocesium (134Cs and 137Cs) concentrations in the samples. Before the measurements, the detector was calibrated using a serial standard volume source made of aluminum and mixed radionuclides (MX033U8PP, Japan Radioisotope Association, Tokyo,

2.2. Set of D-shuttle, lead shield, and holder The D-shuttle (Fig. 1(A)) was covered with a lead shield and kept at 96

Journal of Environmental Radioactivity 204 (2019) 95–103

T. Yoshihara, et al.

Fig. 1. Set-up of the monitoring plots. (A)D-shuttle, (B) Holder, A portion of the pillars of the holders was pegged into soil so that the height of the D-shuttle set was always 200 mm above the soil surface, (C) D-shuttle set (D-shuttle, lead shield, and holder), (D) Lead shield and D-shuttle wrapped in an inner water-resistant plastic bag, (E)View of the D-shuttle sets placed in plot 3 (red stars). Note that every numeric value in A−D is in millimeters (mm).

Japan). The relative uncertainty of the measurements was approximately 4.8%. Both the 134Cs and 137Cs concentrations were corrected for physical decay to the end of the monitoring (August 9, 2017). 2.5. Test monitoring and simulation to estimate measuring range of the Dshuttle set During test monitoring, the D-shuttle set was placed on flat soil and encircled with lead bricks (L100 × D50 × H200 mm3) without gaps (Fig. 3). The radii of the circles were 10, 25, 50, 75, and 100 cm. The accumulated dose was measured for at least seven days for each radius to determine the variance. A measurement without the lead brick circle was conducted as a control. Finally, the measuring range of the Dshuttle set was estimated by comparing the values measured with and without the lead brick circle, as indicated by the RAD (Relative ratio of

Fig. 2. Schematic of the monitoring areas and setting of the D-shuttle set. The schematic shows the natural and decontaminated monitoring areas; the decontaminated area (D) and the natural area (N).

Fig. 3. Schematic of the test monitoring. A schematic of the measurement condition for the actual measurement is illustrated.

97

Journal of Environmental Radioactivity 204 (2019) 95–103

T. Yoshihara, et al.

accumulated dose measured with the lead brick circle against accumulated dose with out the lead brick circle). Monte Carlo simulations were performed using the Particle and Heavy Ion Transport code System (PHITS; Sato et al., 2018). First, a cylindrical soil mass (height of 100 cm; radii of 20, 50, 100, 150, and 200 cm; chemical formula of H (2.20)O (3.59)Al (0.32)Si (0.93)Fe (0.10); and density of 1.0 g cm−3) was placed in a virtual field for the simulations (ICRU, 1994). Thus, soil mass was increased in correspondence with the increase in radius of the lead brick walls as shown below. This was necessary to avoid an increase in statistical errors with the decreasing radius of the lead brick walls. The upper 5 cm depth of the virtual soil mass was assumed to be contaminated. It is noteworthy that two different soil water contents (19.8% and 50%) were supposed for this upper 5 cm depth soil mass as cases of ordinary and heavy rainfall conditions, respectively (Kondo, 1993), to estimate the shielding effect of rainfall water on soils. Five virtual walls, which differed in radius to imitate the annularly placed lead bricks, as shown in the actual measurement (height of 20 cm; thickness of 5 cm; radii of 10, 25, 50, 75, and 100 cm), were placed respectively on the surface center of the virtual cylinder. In addition, a thin silicon cuboid (W70 × D70 × H0.32 mm3) covered with a lead shield (W130 × D130 × H45 mm3) imitating the D-shuttle set was located at the center of the virtual soil mass and 20 cm above the surface. Then, 1.0212 × 1010 photons (662 keV) were assumed to be generated uniformly from the contaminated soil layer. From the energy characteristic curve of the D-Shuttle, the threshold energy was estimated to be 60 keV, where the relative sensitivity of the D-Shuttle was 0.5 (Murayama, 2018). Thus, the number of photon counts depositing 60 keV or more energy into the detector was simulated because the dosimeter integrated the number of detected photons with energies greater than 60 keV (CTC, 2018). A simulation of the photon detection was also conducted for the control.

Fig. 4. Estimations of the measuring range of the D-shuttle set. The results of the actual survey and the Monte Carlo simulation are shown as the relative ratios of the accumulated doses (RAD), which were mainly detected in the soils under D-shuttle and in the entire environment (i.e., the control values measured without the lead brick wall; see Fig. 3). The error bars of the actual survey and the simulation show the standard deviations of 72 h continuous measurement and the statistical errors of the simulation, respectively.

between the test measurements and the Monte Carlo simulation at larger radii was caused by differences between the preconditions of these two estimation method, such as the field assumed, depths of contaminated layer, and/or the sizes of detector and lead shields. A minor irregularity in the arrangement of the lead bricks and/or an imperceptible slope of the ground surface in the test measurements might have affected these results. Furthermore, it is probable that soil surface roughness and soil water content affected the test measurement results (Jacob and Paretzke, 1986), although the simulation did not contain these factors. These results mean that approximately 70% of the radiation detected by the D-shuttle set came from within the area of a circle with a radius of 50 cm (= ca. 0.8 m2 area of flat ground). In this regard, we previously demonstrated that D-shuttle monitoring with a collimator could record hourly dose rates inside the collimator, which correlated with the surface dose rates from spots of soil surface (with a diameter approximately equal to the installed height, i.e., 10–60 cm) directly below the collimation window (Konoplev et al., 2018). Thus, the present results verified the previous findings at more precise levels even though the collimation window was wider than that used previously.

2.6. Data treatment and statistical analysis The accumulated dose (originally counted as the 24 h cumulative dose) was basically shown as the 7 d moving average after correction with the temperature dependency. In brief, each index was multiplied by a factor of the temperature. The factor was 1.0 at 30 °C and the value was decreased with increase of temperature at 0.0025/°C (CTC, 2018; Kurita et al., 2018). The temperature and precipitation data for the calculation were obtained from a monitoring station located 7.8 km from the observation site in Abiko (JMA, 2018). For the statistical analyses, ANOVA, ANCOVA, Tukey's multiple comparison test, and Student's pairwise t-test were performed with the Excel 2010 (ver. 14.0; Microsoft Corp., WA, USA), the Statistics for Excel (ver. 1.12; Social Survey Research Information Co., Ltd., Tokyo, Japan), and/or the Kyplot (ver. 4.0; Kyens Lab Inc., Tokyo, Japan) software packages. Regression lines and correlation coefficients (Pearson's r) were calculated with the Origin software package (ver. 8.1J SR3; Origin Lab Co., MA, USA).

3.2. Variation in accumulated doses at various monitoring extents The variations in the accumulated doses were significant at every level of monitoring extent: plot, area, and row (p ≤ 0.005; data not shown), although the actual differences in mean doses were only approximately 1–1.5 μSv. The variations among positions in each plot were also significant (p ≤ 0.001; data not shown). In particular, the doses in Plot 1 were obviously higher than those of the other two plots (Fig. 5). Such apparent variations can mainly be attributed to differences in the specific radiocesium distribution in the soils where the Dshuttles were set up. In fact, the significant differences in radiocesium concentration between Plot 1 and the other plots were confirmed by the soil sampling after the monitoring (p ≤ 0.05; Figs. 6 and 7). The average radiocesium concentrations in the upper 5 cm of the soils of Plots 2 and 3 were 36 and 45% (Fig. 6) and the respective inventories were 73 and 55% (Fig. 7) of those of Plot 1. The observed differences between the plots were based mainly on the initial interception of the fallout (Kinnersley et al., 1997; Pröhl, 2009) and/or tree phenology (Nord and Lynch, 2009), both of which are governed mainly by

3. Results and Discussion 3.1. Estimation of D-shuttle set measuring range The measuring range of the D-shuttle set was determined by actual measurements (test measurements) and the Monte Carlo simulation (Fig. 3). The RAD of the test measurement increased significantly with increasing encircled area at radii of smaller than 50 cm (Fig. 4). The RAD reached its maximum of approximately 0.72 at a radius of 75 cm, and the differences among values were not significant at radii of larger than 50 cm. On the other hand, the RAD of the simulation increased significantly to a radius of 100 cm (approximately 0.94), although the values at radii of smaller than 50 cm were not significantly different from those in the case of the test measurement. The discrepancy of RAD 98

Journal of Environmental Radioactivity 204 (2019) 95–103

T. Yoshihara, et al.

Fig. 5. Variances of the accumulated doses during the measurement. Variances of the actual accumulated dose (24 h cumulative dose, μSv) in each plot/area during the measurement (from August 26, 2016, to August 9, 2017) are illustrated using box charts (maximum, 99%, 75%, average, median, 25%,1%, minimum). The numbers in the figure (e.g., 38, 56, …) identify the individual D-shuttle sets. The relative locations are shown in the schematic figure at the right side. Even though five D-shuttle sets were placed in each area at the start of the monitoring, at most three of these (marked with crosses) were not able to read the data because of an accident (e.g., water-logged or tumbled) during the measurement.

tendency is usually considered to be primarily due to the physical decay of radiocesium, and physical, chemical, and biological processes, such as erosion, sedimentation, wash-off, sorption, ion exchange, reduction/ oxidation, and methylation/volatilization, accelerate the natural attenuation of radionuclides (IAEA, 2006). Natural attenuation of accident-derived radiocesium that is faster than expected from the physical decay rate has often been observed in the environment affected by the FDNPP accident (Andoh et al., 2015; Mikami et al., 2015). The cause of faster attenuation has been linked to land use, such that the trend was more pronounced for built-up areas than for land used as forest areas, or radiocesium mobility in the depth direction (Malins et al., 2016). On the other hand, Konoplev et al. (2018) observed the expected attenuation of radiocesium depending on the physical decay rate in a floodplain in Fukushima. However, some invalid attenuations were measured during specific events where the top soil layer with high radionuclide contamination was eroded and/or buried beneath fresh, cleaner sediments from bank erosion and sediment movement. The present monitoring data showed a slower attenuation than the hypothetical attenuation, calculated exclusively for the radioactive

vegetation. The major tree species in the plot and/or the planting density could affect the interception status (i.e., dense planting of C. Japonica in Plot 1). Typically, the total annual translocation from forest canopy to floor in cedar forest is higher than that in deciduous or mixed forest (Endo et al., 2015). Some differences may be attributable to physical/meteorological events, depending primarily on the inclination angle of the slope and rainfall intensity (i.e., the gently sloping surface in Plot 1) (Mabit et al., 2008; Sakuma et al., 2017). In this regard, Koarashi et al. (2014) demonstrated a clear topographic heterogeneity (i.e., higher accumulation at the bottom of a slope than at the top or in the middle) in Fukushima-accident derived 137Cs on the forest floor. These authors supposed the involvement of biological processes in the accumulation after physical/meteorological events.

3.3. Overall tendency of accumulated doses during observation period The accumulated doses decreased over time irrespective of plot, area, row, or position. Representative data for the D and N areas (i.e., average of four each in plot 1) are shown in Fig. 8A. This decreasing

Fig. 6. Radiocesium concentration in soils beneath the D-shuttle sets. Soils beneath the D-shuttle sets were sampled at the end of the monitoring and the depth distributions of radiocesium is indicated (134Cs + 137Cs, Bq g-DW−1). From the left to right in the figure, data for the average of all plots/areas (All), P1-D (plot1-D), P1-N (plot1-N), P2-D (plot 2-D), P2-N (plot 2-N), and P3-N (plot 3-N), are indicated, respectively. Ol, litter layer; Of, superficial decomposing organic layer. 99

Journal of Environmental Radioactivity 204 (2019) 95–103

T. Yoshihara, et al.

Fig. 7. Depth distributions in soils beneath the D-shuttle sets. Depth distributions of inventory (134Cs+137Cs, kBq m−2) in the soils beneath the D-shuttle sets are indicated. See Fig. 6 for details.

decay of 134Cs (T1/2 = 2.06 years) and 137Cs (T1/2 = 30.17 years) (Fig. 8A). The slopes of both regression lines calculated after the logarithmic conversion were significantly smaller than those of the theoretical regression lines (p ≤ 0.001), and the differences between the accumulated doses and the hypothetical attenuation increased over time. Notably, the equation of Yoschenko et al. (2016) was used, which neglected the pre-accident radiation background at the sites, assuming

that the observed dose rate could be exclusively attributed to 134Cs and Cs, assumuing that the initial isotope ratio of 134Cs/137Cs in the fallout immediately after the accident was 1 (Hirose, 2012), and including a ratio of 134Cs and 137Cs gamma kerma (kinetic energy released unit mass) of 2.687 (Gusev and Belyaev, 1991). The possible primary cause of the “slower than hypothetical attenuation” is radiocesium from the forest canopy newly added to the floor by throughfall, 137

Fig. 8. Comparison of accumulated doses between the decontaminated area (D) and the natural area (N) in plot 1. (A) The average of four accumulated doses (seven days moving average after correction with the temperature dependency, μSv) each in the decontaminated area (D; 38, 56, 95, and 98) and the natural area (N; 63, 71, 82, and 86) in plot 1 is indicated with each hypothetical reduction (see text). (B) The differences in the average values in Fig. 8A for the N areas from those for the D areas are indicated separately for the first duration between August 26, 2016, and April 10, 2017 (narrow line), and the second duration between April 11, 2017, and August 9, 2017 (bold line), respectively. Narrow and bold dotted lines indicate the regression line for the first (y = −2.3E-4x −570, r = −0.28, p = 1.7E-5) and the second duration (y = 1.1E-3x −2790, r = 0.60, p = 2.1E-13), respectively. 100

Journal of Environmental Radioactivity 204 (2019) 95–103

T. Yoshihara, et al.

deposited from the air with rain and accumulating on the soil surface are less subject to water content, and the dose rate increases with rainfall. In contrast, artificial radionuclides (radiocesium in our case), which mainly penetrate the first 0–5 cm of soils, were influenced by the soil water content after the nuclear accident, and the dose rate decreased with rainfall. These results could be quantitfied and clarified if correlated with changes in soil moisture content.

stemflow, and litterfall after the start of the monitoring. The contributions to total annual translocation of throughfall and litterfall are always higher than those of stemflow (Loffredo et al., 2014; Teramage et al., 2014; Endo et al., 2015). In addition, movement and accumulation/sedimentation of contaminated soil and litterfall to the range of the D-shuttle sets could also have contributed to the slower attenuation. The deviations between the actual monitoring and hypothetical reductions were approximately 6.0% in both the D and N areas at the end of the monitoring (Fig. 8A). In other words, our results demonstrated that secondary contamination of radiocesium by translocation still continued and changed the radiocesium distribution in the forest five to six years after the accident and that the amount reached approximately 6.0% of the total radioactivity at the soil surface per year. In previous observations of the FDNPP accident, the leachable 137Cs loss from the canopy by throughfall and stemflow was as high as 72% of the total plume deposition within three years from the accident, but the absolute amount and occupancy in the total leachate decreased over time (Teramage et al., 2014; Loffredo et al., 2014, 2015). Therefore, it is expected that the contribution of throughfall and stemflow during translocation would not be much in the present observation compared to that of litterfall. Other than the overall tendency during the observation period, the accumulated doses showed a remarkable decline corresponding to heavy rainfall irrespective of decontamination (Fig. 9A−C). This may mean that water has a shielding effect to radioactivity in soils. For example, the present simulation results indicated that increased water content in the first 5 cm of the contaminated surface soil, from 19.8 to 50%, can reduce radiocesium gamma radiation by as much as 10% (data not shown). On the other hand, the ambient gamma dose rate usually increases with rainfall (Fukuda, 1982). Ground deposition of atmospheric 222Rn decay products is related to this phenomenon (Fukuda, 1982; Takeyasu et al., 2010). The possible contradiction in our observation can be explained by a difference in the position of radionuclides (in/on the soils). Ordinarily, natural radionuclides

3.4. Effect of decontamination Basically, the accumulated doses in the D area were lower than those in the N area during the entire duration of observation (p ≤ 0.001; Fig. 8A). This means that the effect of decontamination by the removal of the litter layer and superficial decomposing organic layers was noticeable throughout the observation period, as shown in other cases of decontamination in forests (Ayabe et al., 2017). However, the differences between the D and N areas were not stable and changed after mid-April 2016 (Fig. 8B). Specifically, the effect of decontamination (i.e., the ratio of the difference in the accumulated doses between the D and N areas to that in the N area) was approximately 1.2% at the beginning, reached a maximum of approximately 3.3% in mid-April, and decreased to approximately 1.3% at the end of the observation period. It is possible that capturing of newly added radiocesium was easier in N areas than in D areas, and this might caused the increased decontamination effect until mid-April. However, this trend was converted after mid-April. Newly added radiocesium could be accumulated similarly in both the N and D areas and this might lead the decrease in the decontamination effect after mid-April. This conversion of the trend after mid-April could have been induced by the falling of large branches and/or by something effective in damming up newly added radiocesium attached to litterfall and soils. Observations supporting such a hypothesis have been reported. For example, in a cedar plantation from October 2012 to September 2013, although the major part of the litterfall occurred from November to December, litterfall also Fig. 9. Alignment of averages of the D and N areas in plot 1 with daily cumulative precipitation. (A) Averages of four accumulated doses (seven days moving average after correction with the temperature dependency) each in the decontaminated area (D; 38, 56, 95, and 98) in plot 1 and (B) the natural area (N; 63, 71, 82, and 86) in plot 1 are indicated. (C) The seven days moving average of daily cumulative precipitation is shown.

101

Journal of Environmental Radioactivity 204 (2019) 95–103

T. Yoshihara, et al.

showed a peak between April and May with a significant flux of radiocesium (Endo et al., 2015). In addition, the litterfall during November and December consisted mainly of leaves but that during April to May was mainly non-leaves (e.g., branches). Furthermore, the radiocesium concentrations in old branches and dead parts, particularly in those that had emerged before 2010 and received direct radioesium deposition, were critically higher than those in younger parts, including newly developed leaves (Yoshihara et al., 2016b). These results suggest that different types of litterfall (i.e., leaves and non-leaves) could possibly cause differences in the accumulation status of radiocesium and might finally lead to a disappearance of the decontamination effect. On the other hand, Ramzaev et al. (2006) demonstrated the longterm stability of the decontamination effect after the Chernobyl accident in Russia. It is noteworthy that their decontamination was conducted more than 10 years after the accident (i.e., 1995–1997), when the migration of radionuclides from the canopy to the ground was mostly over, and was mainly targeted at residential areas. In addition, in the forest zone, the removal of newly formed litterfall was repeated once each year. In other words, the stability of the decontamination effect would depend on site specificity, the timing of decontamination, and the additional effort to maintain the effect. This emphasizes the importance of monitoring/assesssing the decontamination effect in the Fukushima restoration zone, particularly in forest areas in the vicinity of residential areas.

doi.org/10.1016/j.jenvrad.2019.03.023. References 220 authorsAdachi, N., Adamovitch, V., et al., 2016. Measurement and comparison of individual external doses of high-school students living in Japan, France, Poland and Belarus—the ‘D-shuttle’project—. J. Radiol. Prot. 36, 49–66. https://doi.org/10. 1088/0952-4746/36/1/49. 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, T., Saito, K., 2015. Measurement of air dose rates over a wide area around the Fukushima Dai-ichi Nuclear Power Plant through a series of car-borne surveys. J. Environ. Radioact. 139, 266–280. https://doi.org/10.1016/j.jenvrad.2014.05.014. Ayabe, Y., Hijii, N., Takenaka, C., 2017. Effects of local-scale decontamination in a secondary forest contaminated after the Fukushima nuclear power plant accident. Environ. Pollut. 228, 344–353. https://doi.org/10.1016/j.envpol.2017.05.041. Čemusová, Z., Ekendahl, D., Judas, L., 2017. Testing of the D-Shuttle personal dosemeter. Radiat. Meas. 106, 214–217. https://doi.org/10.1016/j.radmeas.2017.03.011. Chiyoda Technol Corp. (CTC), 2018. Technical Specification for D-Shuttle. http://www. c-technol.co.jp/eng/e-dshuttle, Accessed date: 19 October 2018. Doi, T., Masumoto, K., Toyoda, A., Tanaka, A., Shibata, Y., Hirose, K., 2013. Anthropogenic radionuclides in the atmosphere observed at Tsukuba: characteristics of the radionuclides derived from Fukushima. J. Environ. Radioact. 122, 55–62. https://doi.org/10.1016/j.jenvrad.2013.02.001. Endo, I., Ohte, N., Iseda, K., Tanoi, K., Hirose, A., Kobayashi, N.I., Murakami, M., Tokuchi, N., Ohashi, M., 2015. Estimation of radioactive 137-cesium transportation by litterfall, stemflow and throughfall in the forests of Fukushima. J. Environ. Radioact. 149, 176–185. https://doi.org/10.1016/j.jenvrad.2015.07.027. Forestry Agency, 2014. Forests/timbersand Radionuclides – Revival of Forests and Forestry in Fukushima. In Japanese. https://www.pref.fukushima.lg.jp/uploaded/ attachment/100065.pdf, Accessed date: 17 March 2016. Fukuda, K., 1982. On the factors affecting variation of gamma-ray exposure rates observed with the monitoring post. Tenki, Meteorol. Soc. Japan 29, 781–785 (In Japanese). Fukushima Revitalization Station (FRS) Fukushima Prefectural Government, Japan. 2018. Transition of Evacuation Designated Zones. http://www.pref.fukushima.lg.jp/site/ portal-english/en03-08.html, Accessed date: 23 October 2018. Gusev, N.G., Belyaev, V.A., 1991. Radioactive Releases into the Biosphere. Reference book. second ed. Energoatomizdat, Moscow, pp. 256 (in Russian). Hirose, K., 2012. 2011 Fukushima Dai-ichi nuclear power plant accident: summary of regional radioactive deposition monitoring results. J. Environ. Radioact. 111, 13–17. https://doi.org/10.1016/j.jenvrad.2011.09.003. Imamura, N., Komatsu, M., Ohashi, S., Hashimoto, S., Kajimoto, T., Kaneko, S., Takano, T., 2017. Temporal changes in the radiocesium distribution in forests over the five years after the Fukushima Daiichi Nuclear Power Plant accident. Sci. Rep. 7, 8179. https://doi.org/10.1038/s41598-017-08261-x. International Atomic Energy Agency (IAEA), 2006. Applicability of Monitored Natural Attenuation at Radioactively Contaminated Sites. Technical Reports Series No. 445. IAEA, Vienna, pp. 105. International Commission on Radiation Units (ICRU), 1994. International Commission on Radiation Units and Measurements: "Gamma-Ray Spectrometry in the Environment". ICRU Report 53. . Jacob, P., Paretzke, H.G., 1986. Gamma-Ray exposure from contaminated soil. Nucl. Sci. Eng. 93, 248–261. Japan Meteorological Agency (JMA), 2018. The latest 30 y observation in Tsushima and Iitate observation sites, http://www.jma.go.jp/jma/menu/report.html (accessed 16 Oct 2018). Kinnersley, R.P., Goddard, A.J.H., Minski, M.J., Shaw, G., 1997. Interception of caesiumcontaminated rain by vegetation. Atmos. Environ. 31, 1137–1145. https://doi.org/ 10.1016/S1352-2310(96)00312-3. Koarashi, J., Atarashi-Andoh, M., Takeuchi, E., Nishimura, S., 2014. Topographic heterogeneity effect on the accumulation of Fukushima-derived radiocesium on forest floor driven by biologically mediated processes. Sci. Rep. 4, 6853. https://doi.org/10. 1038/srep06853. Kondo, J., 1993. Time variation in soil water content of the surface soil layer after a rainfall. J. Japan Soc. Hydrol. Water Resour. 6, 336–343 (In Japanese). Konoplev, A., Golosov, V., Wakiyama, Y., Takase, T., Yoschenko, V., Yoshihara, T., Parenyuk, O., Cresswell, A., Ivanov, M., Carradine, M., Nanba, K., Onda, Y., 2018. Natural attenuation of Fukushima-derived radiocesium in soils due toits vertical and lateral migration. J. Environ. Radioact. 186, 23–33. https://doi.org/10.1016/j. jenvrad.2017.06.019. Kurita, K., Yamaguchi, M., Nagao, Y., Suzui, N., Yin, Y.-G., Yoshihara, T., Kawachi, N., 2018. Development of an easy and simple method to measure the environmental radioactivity in trees with efficient personal dosimeters. Radioisotopes 67, 427–434. https://doi.org/10.3769/radioisotopes.67.427. Loffredo, N., Onda, Y., Kawamori, A., Kato, H., 2014. Modeling of leachable 137Cs in throughfall and stemflow for Japanese forest canopies after Fukushima Daiichi Nuclear Power Plant accident. Sci. Total Environ. 493, 701–707. https://doi.org/10. 1016/j.scitotenv.2014.06.059. Loffredo, N., Onda, Y., Hurtevent, P., Coppin, F., 2015. Equation to predict the 137Cs leaching dynamic from evergreen canopies after a radio-cesium deposit. J. Environ. Radioact. 147, 100–107. https://doi.org/10.1016/j.jenvrad.2015.05.018. Mabit, L., Benmansour, M., Walling, D.E., 2008. Comparative advantages and limitations of the fallout radionuclides 137Cs, 210Pbex and 7Be for assessing soil erosion and sedimentation. J. Environ. Radioact. 99, 1799–1807. https://doi.org/10.1016/j.

4. Conclusion The results of this study demonstrated the ability of the D-shuttle set to monitor changes in specific radioactivity from the forest floor over the long term. The actual monitoring and the Monte Carlo simulation demonstrated that one D-shuttle set enables the monitoring of gamma radiation from radiocesium distributed over a limited area (0.5 m radius ca. 0.8 m2 area of flat ground). The yearlong monitoring using the D-shuttle sets distinguished every level of monitoring (i.e., plot, area, row, and position), reflecting the differences in radiocesium distribution on the forest floor. The results suggested that long-term monitoring using the D-shuttle set provides novel findings that could not be detected by existing method. For example, the accumulated doses decreased over time, but the attenuation was slower than that expected by the physical decay of radiocesium. This suggests that radiocesium was newly added to the forest floor from the canopy by litterfall, and the amount was approximately 6% of the total radioactivity at the soil surface per year. In addition, the difference in the radioactivity between the D and N areas was obvious. However, these differences did not transition stably, and there was an apparent change in the trend after mid-April 2016. This may have been caused by quantitative/qualitative changes in the litterfall at the time. The accumulated doses showed a remarkable decrease in accordance with a heavy rainfall event. This may have been attributable to a shielding effect of water to radioactivity in soils. Application of this monitoring method to a wider range of forests and/or decontaminated areas with some environmental indices, such as soil moisture content, would yield quantitative and meaningful results and elucidate additional implications. Acknowledgements We thank Ms. Mari Sato (Civil Engineering Research and Environmental Studies, Co. Ltd.) and Mr. Keita Yamaguchi (Electric Power Engineering Systems, Co. Ltd.) for their skillful assistance in analyses of radionuclides. This research was supported in part by a Grant-in-Aid for Scientific Research from the Japan Society for the Promotion of Science (JSPS, 15H04621). Appendix A. Supplementary data Supplementary data to this article can be found online at https:// 102

Journal of Environmental Radioactivity 204 (2019) 95–103

T. Yoshihara, et al.

Ramzaev, V., Andersson, K.G., Barkovsky, A., Fogh, C.L., Mishine, A., Roed, J., 2006. Long-term stability of decontamination effect in recreational areas near the town Novozybkov, Bryansk Region, Russia. J. Environ. Radioact. 85, 280–298. https://doi. org/10.1016/j.jenvrad.2004.05.023. Sakuma, K., Kitamura, A., Malins, A., Kurikami, H., Machida, M., Mori, K., Tada, K., Kobayashi, T., Tawara, Y., Tosaka, H., 2017. Characteristics of radio-cesium transport and discharge between different basins near to the Fukushima Dai-ichi Nuclear Power Plant after heavy rainfall events. J. Environ. Radioact. 169, 137–150. https://doi. org/10.1016/j.jenvrad.2016.12.006. Sato, T., Iwamoto, Y., Hashimoto, S., Ogawa, T., et al., 2018. Features of particle and heavy ion transport code system (PHITS) version 3.02. J. Nucl. Sci. Technol. 55, 684–690. https://doi.org/10.1080/00223131.2017.1419890. Takeyasu, M., Onuma, T., Takeishi, M., 2010. Determination of the variation of environmental radiation due to the ground deposition of atmospheric 222Rn decay products during rainfall: utilization of the correlation between the variations observed at neighboring monitoring posts. J. Radioanal. Nucl. Chem. 284, 635–639. https://doi.org/10.1007/s10967-010-0532-6. Terada, H., Katata, G., Chino, M., Haruyasu, N., 2012. Atmospheric discharge and dispersion of radionuclides during the Fukushima Dai-ichi Nuclear Power Plant accident. Part II: verification of the source term and analysis of regional−scale atmospheric dispersion. J. Environ. Radioact. 112, 141–154. https://doi.org/10.1016/j. jenvrad.2012.05.023. Teramage, M.T., Onda, Y., Kato, H., Gomi, T., 2014. The role of litterfall in transferring Fukushima-derived radiocesium to a coniferous forest floor. Sci. Total Environ. 490, 435–439. https://doi.org/10.1016/j.scitotenv.2014.05.034. Yoschenko, V., Nanba, K., Yoshida, S., Watanabe, Y., Takase, T., Sato, N., Keitoku, K., 2016. Morphological abnormalities in Japanese red pine (Pinus densiflora) at the territories contaminated as a result of the accident at Fukushima Dai-Ichi Nuclear Power Plant. J. Environ. Radioact. 165, 60–67. https://doi.org/10.1016/j.jenvrad. 2016.02.017. Yoshihara, T., Matsumura, H., Hashida, S.-N., Nagaoka, T., 2013. Radiocesium contaminations of 20 wood species and the corresponding gamma−ray dose rates around the canopies at 5 months after the Fukushima nuclear power plant accident. J. Environ. Radioact. 115, 60–68. https://doi.org/10.1016/j.jenvrad.2012.07.002. Yoshihara, T., Hashida, S.-N., Abe, K., Ajito, H., 2014a. A time dependent behavior of radiocesium from the Fukushima fallout in Litterfall of Japanese flowering cherry tree. J. Environ. Radioact. 127, 34–39. https://doi.org/10.1016/j.jenvrad.2013.09. 007. Yoshihara, T., Matsumura, H., Kobayashi, T., Hashida, S.-N., Nagaoka, T., Tsuzaki, M., Wakamatsu, T., Goto, F., 2014b. Changes in radiocesium contamination from Fukushima in foliar parts of 10 common tree species in Japan between 2011 and 2013. J. Environ. Radioact. 138, 220–226. https://doi.org/10.1016/j.jenvrad.2014. 09.002. Yoshihara, T., Kominato, A., Nagao, Y., Kawachi, N., Hashida, S.-N., Suzui, N., Yin, Y.-G., Fujimaki, S., 2016a. Measurement of environmental radioactivity using a cumulative gamma radiation dosimeter: fertilization-induced changes in young fruit trees. In: Ioannidou, A., Povinec, P.P. (Eds.), Proc. ENVIRA2015, New Challenges with New Analytical Technologies, pp. 134–137 Sep. 21−25, 2015, Thessaloniki, Greece. Yoshihara, T., Matsumura, H., Hashida, S.-N., Nakaya, K., 2016b. Radiocesium contamination in living and dead foliar parts of Japanese cedar during 2011-2015. J. Environ. Radioact. 164, 291–299. https://doi.org/10.1016/j.jenvrad.2016.08.010.

jenvrad.2008.08.009. Malins, A., Kurikami, H., Nakama, S., Saito, T., Okumura, M., Machida, M., Kitamura, A., 2016. Evaluation of ambient dose equivalent rates influenced by vertical and horizontal distribution of radioactive cesium in soil in Fukushima Prefecture. J. Environ. Radioact. 151, 38–49. https://doi.org/10.1016/j.jenvrad.2015.09.014. Ministry of Education, Culture, Sport, Science, and technology (MEXT), 1976. Radioactivity Measurement Series No. 3 (Revision 1), Radiocesium Analytical Method. Japan chemical analysis center, Chiba, Japan, pp. 21–27. Ministry of Education, Culture, Sport, Science, and technology (MEXT), 1992. Radioactivity Measurement Series No. 7 (Revision 3), Gamma Ray Spectrometry with a Germanium Semiconductor Detector. Japan chemical analysis center, Chiba, Japan, pp. 179–182. Ministry of the environment (MOE), 2012. Decontamination Plans for Forests. In Japanese. http://josen.env.go.jp/about/efforts/forest.html, Accessed date: 30 October 2018. Mikami, S., Maeyama, T., Hoshide, Y., Sakamoto, R., Sato, S., Okuda, N., Sato, T., Takemiya, H., Saito, K., 2015. 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. https://doi.org/10.1016/j.jenvrad.2014. 08.020. Morino, Y., Ohara, T., Nishizawa, M., 2011. Atmospheric behavior, deposition, and budget of radioactive materials from the Fukushima Daiichi nuclear power plant in March 2011. Geophys. Res. Lett. 38 L00G11. https://doi.org/10.1029/ 2011GL048689. Murayama, K., 2018. D-shuttle electric dosimeter and its applications. Radioisotopes 67, 453–460. In Japanese except abstract. https://doi.org/10.3769/radioisotopes.67. 453. Nagao, S., Kanamori, M., Ochiai, S., Tomihara, S., Fukushi, K., Yamamoto, M., 2013. Export of 134Cs and 137Cs in the Fukushima river systems at heavy rains by typhoon roke in september 2011. Biogeosciences 10, 6215–6223. https://doi.org/10.5194/bg10-6215-2013. Naito, W., Uesaka, M., Yamada, C., Ishii, H., 2015. Evaluation of dose from external irradiation for individuals living in areas affected by the Fukushima daiichi nuclear plant accident. Radiat. Protect. Dosim. 163, 353–361. https://doi.org/10.1093/rpd/ ncu201. Naito, W., Uesaka, M., Yamada, C., Kurosawa, T., Yasutaka, T., Ishii, H., 2016. Relationship between individual external doses, ambient dose rates and individuals' activity-patterns in affected areas in Fukushima following the Fukushima Daiichi nuclear power plant accident. PLoS One 11, e0158879. https://doi.org/10.1371/ journal.pone.0158879. National Land Agency, Japan (NLA), 1990. Soil Map, Namie-Iwaki-Tomioka. In Japanese. http://nrb-www.mlit.go.jp/kokjo/tochimizu/F3/ZOOMA/0720/index. html, Accessed date: 31 July 2018. Nord, E.A., Lynch, J.P., 2009. Plant phenology: a critical controller of soil resource acquisition. J. Exp. Bot. 60, 1927–1937. https://doi.org/10.1093/jxb/erp018. Papagiannis, P., Baltas, D., Granero, d., Pérez-Calatayud, J., Gimeno, J., Ballester, F., Venselaar, J.L.M., 2008. Radiation transmission data for radionuclides and materials relevantto brachytherapy facility shielding. Med. Phys. 35, 4898–4906. https://doi. org/10.1118/1.2986153. Pröhl, G., 2009. Interception of dry and wet deposited radionuclides by vegetation. J. Environ. Radioact. 100, 675–682. https://doi.org/10.1016/j.jenvrad.2008.10.006.

103