Full article: Air concentration estimation of radionuclides discharged from Fukushima Daiichi Nuclear Power Station using NaI(Tl) detector pulse height distribution measured in Ibaraki Prefecture

ABSTRACT
Formulae display:?Mathematical formulae have been encoded as MathML and are displayed in this HTML version using MathJax in order to improve their display. Uncheck the box to turn MathJax off. This feature requires Javascript. Click on a formula to zoom.

ABSTRACT

Air concentrations of ¹³³Xe, ¹³¹I, ¹³²I, ¹³³I, ¹³²Te, ¹³⁴Cs, ¹³⁶Cs, and ¹³⁷Cs in the early stage of the Fukushima Daiichi Nuclear Power Station (FNPS1) accident were estimated for six locations in Ibaraki Prefecture from pulse height distributions measured with NaI(Tl) scintillation detectors continuously operated as a part of monitoring station systems. Air concentration of ¹³³Xe in Ibaraki was estimated for the first time, which showed the maximum value of 150 kBq m⁻³ in the plume arrived at the southern part of Ibaraki around noon of 15 March 2011. The plume was found to consist mainly of noble gases without other nuclides at significant level and to correspond to the hydrogen explosion at Unit 3. Spatiotemporal distributions of ¹³¹I and other nuclides were discussed for the plumes detected during the periods of 15–16 and 20–21 March. Variations in radionuclide composition among the plumes and within each plume were also discussed.

KEYWORDS:

Fukushima Daiichi Nuclear Power Station accident NaI(Tl) detector pulse height distribution air concentration xenon-133 iodine-131 atmospheric transport

1. Introduction

Activity concentrations of radionuclides in air and on the ground surface are essential for dose evaluation in the case where a large amount of radioactive materials are released from a nuclear facility. However, it is extremely laborious and time-consuming to measure them as compared with gamma dose rate. Actually, in the early stage of the Fukushima Daiichi Nuclear Power Station (FNPS1) accident in March 2011, availability of activity concentration data was extremely limited. Especially, it was the case for air concentrations of short-lived nuclides such as ¹³¹I, which could not be retrieved later.

In the early stage of the FNPS1 accident, drastic increases of dose rates were observed in Ibaraki Prefecture. Figure 1 shows the location of Ibaraki Prefecture and the locations where the air concentration was evaluated. Tsuruta et al. [1], a few years after the accident, retrieved the air concentrations of ¹³⁴Cs and ¹³⁷Cs by analyzing filter-tapes from suspended particulate matter (SPM) monitoring stations in the eastern part of the main island of Japan. Although they revealed the spatiotemporal distribution of air concentrations in this area, concentrations of nuclides other than ¹³⁴Cs and ¹³⁷Cs were not obtained. Dust samplings were carried out in the early stage of FNPS1 accident by Ohkura et al. [2] at JAEA-1 denoted in Figure 1, and Furuta et al. [3] at JAEA-2. They obtained temporal change in air concentrations of many radionuclides including the short-lived ¹³¹I in Tokai-mura, Ibaraki Prefecture. Hirayama et al. [4] evaluated the temporal change of air concentration of ¹³¹I at JAEA-1 from pulse height distributions measured by NaI(Tl) scintillation detectors, but concentration of other nuclides was unavailable. Other dust sampling data are available in the Fukushima Prefecture, but the data are extremely limited to capture spatiotemporal evolution of plumes.

Figure 1. Locations of monitoring stations in the central part of Ibaraki Prefecture. The locations at which air concentrations have been evaluated by Tsuruta et al. (2014) are also shown by star marks.

Display full size

A huge amount of radioactive noble gases has been considered to be released from the FNPS1, and due to their gaseous nature, it would be highly probable that the mechanism and timing of their release were, to some extent, different from those of iodine and cesium. However, air concentrations of radioactive noble gases were not evaluated except for the data measured at Japan Chemical Analysis Center in Chiba and at the Comprehensive Nuclear Test Ban Treaty monitoring facility in Takasaki, which are distant from FNPS1 by more than 200 km. Therefore, the air concentrations of noble gases at nearer locations are essential to infer the source term of radioactive noble gases from FNPS1.

As shown in Figure 1, many monitoring stations (MSs) of environmental radiation were in operation in the central part of Ibaraki Prefecture at the time of the accident. At several MSs out of them, pulse height distributions measured with NaI(Tl) scintillation detectors were stored and available for later analyses. Since the main component that contributes to the pulse height distribution is the cloud shine radiation besides the ground shine radiation, pulse height distributions measured at MSs potentially have information on the air concentration of radionuclides. In our previous study, a method for evaluating the air and ground surface concentrations separately from pulse height distributions measured at MSs was developed [5], [6]. The estimated air concentrations of radioiodines and radiocesiums at MS in Tokai-mura showed good agreement with values measured near the MS [6].

The purpose of the present study is to estimate the air concentrations of key fission product (FP) nuclides released in the early stage of the FNPS1 accident from NaI(Tl) pulse height distribution measured at MSs in the central part of Ibaraki Prefecture by the method validated in our previous study. A focus was placed on the air concentrations of noble gases and iodine nuclides for the period of 15–16 and 20–21 March 2011, of which horizontal distributions have not been evaluated.

2. Methods

2.1. Estimation of radioactivity concentrations

Since the method for estimating radioactivity concentration used in this study has been described in our previous study [5], only a brief description is given here. The principle of estimation is to find the air and ground surface concentrations which reproduce a pulse height distribution that is closest to an observed one when they are used in a photon transport code. Count rates in energy bins, each of which is corresponding to a full energy peak by gamma radiation of a target nuclide as shown later, were calculated with a Monte-Carlo photon transport code EGS5 [7] for unit concentration of the nuclide. This procedure resulted in a matrix Γ_{i, j} standing for the count rate in the j-th bin contributed by unit concentration of the i-th nuclide, and the count rate A_j in the j-th bin can be expressed as (1) $A_{j} = \sum_{i} Γ_{i, j} C_{i},$ (1) where C_i is the activity concentration of the i-th nuclide either in the air or on the ground surface. The procedure of finding concentrations that best reproduce the measured pulse height distribution was carried out with a least-square method in which the square sum of relative difference in the count rate between observation A_obs and calculation A_cal of each energy range (bin) (2) $Δ = \sum_{j} {(\frac{A_{cal, j} - A_{obs, j}}{A_{obs, j}})}^{2},$ (2) was minimized.

Figure 2 shows examples of pulse height distribution measured at Tokai-mura, Ibaraki during passage of an FNPS1 plume, in which full-energy peaks of anthropogenic radionuclides are clearly seen. In this study, ¹³³Xe, ¹³¹I, ¹³²I, ¹³³I, ¹³²Te, ¹³⁴Cs, ¹³⁶Cs, and ¹³⁷Cs were selected as the target nuclides, since they have large potential contributions to the air dose rate, and most of them were actually detected in Ibaraki prefecture during and after the FNPS1 accident. The energy ranges used to evaluate the count rates A_obs and A_cal in the analysis are listed in Table 1.

Figure 2. Pulse height distributions measured at a monitoring station in Tokai-mura, Ibaraki Prefecture, and target nuclides that contribute to full energy peaks.

Display full size

Table 1. Energy ranges used in the analysis of this study. Those written in bold letters are the main nuclides to be estimated from count rate of corresponding energy range.

CSV Display Table

2.2. Calculation of pulse height distribution

The calculations of pulse height distribution were carried out with the EGS5 to evaluate Γ_{i, j}. Calculation domain was a 1-km cube of air layer and a 2-m deep soil layer with the same horizontal extent as the air layer. The MS of interest was located at the center of the base of air cube. To efficiently calculate the photon transport, the technique developed by Namito et al. [8] was applied in this study, in which the geometries of the detector and the source were exchanged. In our previous study [5], only a NaI crystal was considered in the calculation, and this simplification might cause errors in calculated pulse height distributions and hence in estimated concentrations, especially for ¹³³Xe which emits only a low energy gamma ray. In the present study, an aluminum detector cover with 9-mm thickness and a concrete-made monitoring station building with a dimension of 4 m × 3 m × 2.6 m were considered in addition to the 2''ϕ × 2'' cylindrical NaI crystal at the 3.5-m height above the ground. The density of air was set to be 1.2 × 10⁻³ g cm⁻³, and the composition and density of soil were set to be SiO₂ and 1.6 g cm⁻³. Radionuclides were assumed to exist uniformly in the air, on the ground surface and on the roof of building. An exception was that the radionuclides deposited on the soil surface were assumed to exist in the soil layer following an exponential profile, (3) $C = C_{0} exp (- \frac{d}{β}),$ (3) where d is the mass depth in g cm⁻², and β is the relaxation mass per unit area (g cm⁻²). The values of β were determined for every plume and every MS so that the calculated count rate in the 120–180 keV range best agreed with the observed one. The result showed the values in the range from 0 to 3 g cm⁻².

2.3. Separation of airborne and deposited radionuclides

To separately estimate concentrations of airborne radionuclides and those of radionuclides deposited on the ground surface, the method described in our previous study [5] was used. This method relies on the fact that the count rate in a low energy range differs depending on whether nuclides are in the air or on the ground surface. If a radionuclide exists in the air, the ratio of the count rate in a low energy range contributed by scattered gamma rays to that of its full energy peaks becomes larger compared with the case where the radionuclide is deposited on the ground surface, because the paths of gamma rays are longer in the air-borne case causing larger number of photons scattered by air. It is assumed that the air concentration C_air is in proportion to the ground surface concentration C_sfc except for ¹³³Xe, (4) $C_{air} = F \times C_{sfc},$ (4) and that the value of F is common to all nuclides. The latter assumption may no longer hold for the second and later plumes arrived, since the radionuclide composition may change plume by plume. This shortcoming caused by the assumption was overcome by subtracting the pulse height distribution measured just before the arrival of a plume of interest from the objective pulse height distribution during the passage of the plume. With this procedure, C_sfc in Equation (4) can be regarded as the increment in the surface concentration caused only by the plume of interest. Following our previous study [5], the energy range of 120–180 keV was used to evaluate the count rate of the scattered component.

2.4. Lower limit of air concentration estimation

Since our method of separating the cloud shine component from the ground shine component relies on their different contributions to the count rate in the 120–180 keV range, the separation is possible if the contribution from the air-borne nuclide of interest to this range significantly exceeds the uncertainty in the count rate in this range contributed by nuclides both in the air and on the ground surface. Therefore, the criterion for the significant separation can be expressed as (5) $A_{air} > 3 σ_{sfc} + 3 σ_{air},$ (5) where, A_air is the count rate in the 120–180 keV range from the objective nuclide, and (6) $σ_{sfc} = \sqrt{A_{sfc}}, σ_{air} = \sqrt{A_{air}},$ (6)

The ground shine count rate A_sfc accounted for contributions by all the nuclides on the ground surface. The surface concentrations was assumed to be 100 kBq m⁻² for ¹³¹I, 10 kBq m⁻² for ¹³²I, and ¹³²Te, and 1 kBq m⁻² for the other nuclides. These values were typical for the surface concentrations in the central part of Ibaraki during the period from 15 to 20 March [5,6]. It was also assumed for simplicity that only a single nuclide existed in the air. The minimum concentrations that meet the criterion expressed by Equation (5) are listed in Table 2. Although it is obvious that coexistence of air-borne nuclides other than the objective one raises the significance level of separation, values in this table can be regarded as reference values for data screening.

Table 2. Minimum concentrations C_air that suffice Equation (5).

CSV Display Table

2.5. Selection of MS

Figure 1 shows the location of MSs that we focused on. They were located in the central part of Ibaraki prefecture, about 100 km to the south from FNPS1. The density of the monitoring network was fairly high and there were some 40 stations in total.

Selection of MSs was made to ensure quality of the concentration values estimated in this study. According to visits to the MSs of interest, some of them were found to be closely surrounded by trees or even overhung by branches of trees. These MSs were excluded from the analysis as described below since the nuclides deposited on the trees give rise to unwanted and difficult-to-eliminate but substantial contribution to pulse height distributions.

Figure 3 shows examples of the temporal changes in the air-absorbed dose rate measured at the MSs during multiple passages of the FNPS1 plumes on 15 March 2011. The smooth decreasing trend in the dose rate starting just after the passage of major plumes during the period denoted by τ corresponds to the decrease in the ground shine component due to decay of deposited nuclides. The relative differences in the ground shine dose rate among the MSs are much larger than those in the cloud shine dose rates shown as the peaks. Ratio of the increment in the ground shine dose rate caused by the plumes during the time period τ to the time-integrated dose rate for the period was calculated for each MS; (7) $R = \frac{D_{d} - D_{0}}{\int_{τ} (D - D_{0}) d t},$ (7) where D₀ and D_d are the ground shine dose rates before and after a given plume passage. Among the MSs located in the northern part of the MS network where three plume passed during the time period τ, MSs with R ⩾ 0.016 h⁻¹ were found by the visits to the MSs to be significantly affected by surrounding trees and were excluded from the analysis. As for the MSs in the southern part, where only one plume passed during the time period τ, ones with R ⩾ 0.026 h⁻¹ were excluded from the analysis for the same reason. Also excluded were the MSs with significant undulations of the ground surface, e.g. MSs located in the vicinity of complex slopes, for which geometrical representation of terrain in the photon transport simulations were difficult. As a result of these procedures, six MSs were selected for further analyses (Figure 1). The objective time period of concentration estimation was 15—31 March 2011. Data interval of pulse height distribution analyzed was 10 minutes during plume passage, and 1 h for the rest. The shorter time interval of 10 minutes was adopted to resolve the rapid change in the air concentrations during plume passage as shown by the change in the dose rate (Figure 3).

Figure 3. Temporal changes in air absorbed dose rate at six MSs in Tokai-mura (left panel, corresponding to the circle A in Figure 1) and three MSs in Hokota-shi (right panel, the circle B).

Display full size

3. Results and discussion

3.1. Air concentrations

The estimated temporal changes in air concentrations of ¹³³Xe, ¹³¹I, and ¹³²Te during the periods from 15 to 17 and 20 to 22 March 2011 are shown in Figures 4–7. There were eight plumes for which air concentrations were estimated. Although there were small increases in air dose rate after 22 March, air concentrations could not be estimated because of interference from nuclides on the ground surface with higher concentrations than in the earlier periods.

Figure 4. Temporal change in the estimated air concentration in 15 March 2011. Measured values at the point JAEA-1 are also plotted for ¹³¹I and ¹³²Te.

Display full size

Air concentration directly measured by Ohkura et al. [2], with a filter sampling method plus gamma spectrometry at the point denoted as JAEA-1 in Figure 1, close to Muramatsu MS, is also plotted in Figures 4–7 by dots. The estimated ¹³¹I concentration is in good agreement with the measured values, although it was slightly underestimated. The estimated concentration of ¹³²Te is in quite good agreement with the measured values without underestimation. The estimated concentrations of ¹³²I and ¹³³I, which are not depicted in Figures 4–7, are also in good agreement with the measured values. Air concentrations of ¹³⁴Cs and ¹³⁷Cs could not be estimated reasonably for Plumes 1–6 because of the co-existence of ¹³¹I and ¹³²I [6]. They are only available after 20 March 2011. Estimated values of ¹³⁴Cs and ¹³⁷Cs after 20 March showed good agreement with measured values, but slightly underestimated within a factor of 2. Estimated values of ¹³⁶Cs, which was also only available after 20 March because the full-energy peak of ¹³⁶Cs was not clearly seen before 20 March, showed good agreement with measured values.

The credibility of ¹³³Xe air concentration in Figure 4–7 is discussed in this paragraph. According to the comparison of measured and calculated pulse height distribution during the plume passage, the calculated count rate of full energy peak of ¹³³Xe showed good agreement with measured one (Figure 8). The pulse height distribution was calculated with sets of estimated ground surface and air concentrations of the target nuclides. In addition, as mentioned in later discussion in Section 3.6.2, the ratio of estimated air concentrations of xenon isotopes including ¹³³Xe showed good agreement with activity ratio calculated from inventory of FNPS1 Unit 3 [9]. Consequently, it can be concluded that the estimated air concentrations of ¹³³Xe in Figure 4–7 are reasonable. However, it must be pointed out that air concentration of ¹³³Xe might be somewhat overestimated by the present method because of relatively poor reproducibility of pulse height distribution in the 70–90 keV range in the absence of plume. Figure 9 shows the comparison between measured and calculated pulse height distributions in the absence of plume (18 March 0:00). The calculation was carried out from an estimated set of the ground surface concentrations of the target nuclides. The calculated count rate in the 70–90 keV energy range is smaller than that measured for Muramatsu (Figure 9(a)). This was the case for the other MSs except for Ishikawa (Figure 9(b)). In the estimation of ¹³³Xe concentration, this discrepancy in the count rate is falsely accounted for as a contribution from ¹³³Xe, resulting in overestimation in its concentration. Although the reason for the slightly poor reproducibility for the 70–90 keV range has not clearly been identified so far, insufficiency in the source geometry representation in the EGS5 calculation is probably one of the reasons. The maximum values of overestimation of the ¹³³Xe air concentration are listed in Table 3. The values in Table 3 were calculated from the differences in the 70–90 keV count rate between calculation and measurement after each plume passage. Overestimation in Plumes 1–3 could not reasonably be evaluated due to contribution from the cloud shine, but their maximum values are considered to be smaller than that of Plume 6 because the ground surface concentration was lower on 15 March than on 16 March.

Figure 5. Same as Figure 4 except for the period in 16 March 2011.

Display full size

Figure 6. Same as Figure 4 except for the period in 20 March 2011.

Display full size

Figure 7. Same as Figure 4 except for the period in 21 March 2011.

Display full size

Figure 8. Comparison of the measured and calculated pulse height distribution at Muramatsu during the plume passage.

Display full size

Figure 9. Comparison of the measured and calculated pulse height distributions at Muramatsu and Ishikawa in the absence of plume.

Display full size

Table 3. Maximum values of overestimation in the ¹³³Xe air concentration (Bq m⁻³).

CSV Display Table

3.2. Characteristics of plumes

According to the results in Figure 4, the estimated air concentrations of Plumes 2 and 3 were substantially higher than those of Plume 1 and the highest among the plumes in March for ¹³¹I, ¹³²I, ¹³³I, and ¹³²Te. Only one exception is the ¹³³Xe concentration of Plume 4 at Ishikawa and Araji, which was as high as about 150 kBq m⁻³ around noon of March 15 at Araji. Plume 2 was detected only at the northern MSs. Since the arrival of this plume was the earliest and its coverage was short at Muramatsu as compared with the inland MSs (Sugaya and Ishikawa), it can be inferred that the plume travelled southward along the coast line but its axis was located inland and that the concentrations at the plume axis were around or more than the estimated maximum concentrations of 81 kBq m⁻³ for ¹³³Xe, 1.3 kBq m⁻³ for ¹³¹I and 1.7 kBq m⁻³ for ¹³²Te.

As for Plume 3, the northern MSs (Muramatsu, Sugaya, Ishikawa) were affected for longer duration than the southern MSs (Ajigaura, Ohnuki, Araji). Air concentrations were more than several-fold higher at the MSs along the coastline (Muramatsu, Ajigaura, Ohnuki, Araji) than at the inland MSs, indicating passage of the main part of this plume along or off-shore the coastline. It should also be noted that the concentrations of ¹³¹I, ¹³²I, ¹³³I and ¹³²Te in this plume were highest among the plumes except for the inland MSs.

Plumes 4 and 5 are characteristic in the sense that only ¹³³Xe was estimated. This implies these plumes were dominated by noble gases, mostly by ¹³³Xe, and concentrations of nuclides other than noble gases were below detectable levels. The air concentration of ¹³³Xe was higher at the southern MSs in Plume 4 (Figure 10), and the concentration was estimated only at the inland MSs in Plume 5.

Figure 10. Maximum air concentrations in kBq m⁻³ of ¹³³Xe in plume No.4.

Display full size

According to the results in Figure 5, concentrations of Plume 6 were slightly lower than that of Plume 2 or 3 for each nuclide. The air concentration of ¹³³Xe reached 51-70 kBq m⁻³ at the early stage of plume passage and decreased more gradually than did the other nuclides. The duration of plume coverage was as long as six hours or more for all MSs. It is interesting to notice that the fastest decrease in concentration was found for ¹³²Te (and its progeny ¹³²I) and following by other radionuclides and the slowest for ¹³³Xe, implying the slower decrease for more gaseous nuclides. It is difficult to tell whether the release mechanisms or the deposition properties are responsible for this result.

According to the results in Figure 6, the concentrations of all nuclides in Plume 7 were estimated to be highest at Ohnuki around 11:00, which were followed by smaller peaks at other locations a few tens of minutes later. From these features and the relatively short time period of this plume, it can be inferred that this radioactive cloud have had a puff-shaped distribution rather than a plume-shaped structure and that the center of this puff crossed the coastline around the MS at Ohnuki or just south of it. It should be pointed out that, like in Plume 6, the concentration decreased most slowly for ¹³³Xe and most rapidly for ¹³²Te with intermediate rate of decrease in the concentration of ¹³¹I.

According to the results in Figure 7, air concentration of ¹³³Xe in Plume 8 were estimated to be about 22 kBq m⁻³ at the maximum, which is much lower than that of Plumes 1–6. Air concentration of ¹³¹I is comparable with that of the plumes in 15 to 17 March, about 1.9 kBq m⁻³ at highest. At Sugaya, air concentrations of all nuclides in Plume 8 are clearly lower than those at MSs near the coastline and have three peaks. At another inland MS, Ishikawa, the concentrations of all nuclides were not estimated because the pulse height distribution data were unavailable during this time period. This feature of multiple peaks was also found at northern MSs. On the other hand, at southern MSs, there is only a single large peak at 5:10 at Ohnuki (5:20 at Araji) in the concentrations of all nuclides, whose level is comparable to those at the northern MSs. These features in the temporal change could have been caused by a plume stretching in the north–south direction along or just off-shore of the coastline, which moved westward in the early morning crossing the coastline and sweeping the MSs. The multiple peaks may correspond to swing motions of the tail of the plume which obliquely crossed the coastline in the northeast–southwest direction at the northern part of the MS network. It is noted again that the feature of slower decrease in the ¹³³Xe concentration holds for this plume.

3.3. Radionuclide composition

Table 4 shows the ¹³³Xe/¹³¹I air concentration ratio of each plume. Values for Plumes 4 and 5 are not shown since they did not have detectable concentrations of ¹³¹I. The ratios were calculated from the time-integrated air concentration of each plume. The ratios of Plumes 2 and 3 are about 50–60 and substantially larger than that of Plume 1 at 14–25. The ratios of Plume 6 are 100–130, which are larger than that of the plumes in the previous day. These differences may indicate possibilities that deposition of ¹³¹I to the ground surface had been enhanced by the rain observed in 16 March, and/or that the plumes in 15 and 16 March originated from different units in FNPS1. The ratios of Plume 7 and 8 are significantly smaller than those of plumes in 15 and 16 March. This can be attributed partially to the shorter physical half-life of ¹³³Xe than that of ¹³¹I and partially to the easily-escaping nature of the noble gases. It is, however, interesting to notice that ¹³³Xe still had substantial or at least noticeable contribution to the activity composition several days after the start of release. This may be due to the production of ¹³³Xe in the reactor from the decay of ¹³³I.

Table 4. Air concentration ratio of ¹³³Xe to ¹³¹I calculated from time integrated air concentration of each plume.

CSV Display Table

The ratio of air concentration of each nuclide with respect to that of ¹³¹I is listed in Table 5. The values in Table 5 are the geometric means of the values at the six MSs. The activity concentration ratios of each nuclide to ¹³¹I are generally consistent with the measured values in Ohkura's report [2], but they are slightly larger than measured values by Ohkura in the Plumes 1, 2, 3, and 6, due to the slight underestimation of ¹³¹I concentration in this study. In contrast to the fairly constant value of ¹³³I/¹³¹I for Plumes 1–3, the ratios of ¹³²I/¹³¹I and ¹³²Te/¹³¹I of Plumes 2 and 3 seem to differ from that of Plume 1.

Table 5. Air concentration ratio of each nuclide to ¹³¹I. Values in parentheses are the values measured by Ohkura et al. [2].

CSV Display Table

3.4. Concentration of ¹³¹I

Horizontal distributions of hourly averaged air concentration of ¹³¹I estimated for Plumes 1–3 in this study are shown in Figure 11 together with air concentration of ¹³¹I converted from the ¹³⁷Cs concentration data analyzed from SPM monitoring filter tapes by Tsuruta et al. [1]. The conversion was done by multiplying ¹³¹I/¹³⁷Cs ratios of 12.9 for Plume 1 and 7.8 for Plumes 2 and 3 measured by Ohkura et al. [2]. Although the figures are given hourly, the following discussion refers to 10 minutes concentration values estimated in this study.

Figure 11. Horizontal distribution of ¹³¹I air concentrations in Bq m⁻³ estimated in this study (diamonds). Concentrations with star marks are values converted from the ¹³⁷Cs concentrations from filter tape analysis [1]. Values in parentheses are the minimum and maximum of the 10 minute average air concentrations during the corresponding time period.

Display full size

Plume 1 arrived first at Muramatsu and Ajigaura at 1:00 of 15 March, then at the other four MSs 10 to 20 minutes later. The concentration was higher at the MSs near the coast line with the maximum of 380 Bq m⁻³ at the southmost MS Araji at 1:50. This plume left the six MSs around 3:30. The southwest tip of the plume reached the SPM monitoring network at 2:00–3:00 and caused the first maximum of about 100 Bq m⁻³ in the next hour at the SPM monitoring stations just the south of Kasumiga-ura lake.

Plume 2 reached the northern part of the MS network around 4:00. It is obvious from the horizontal distributions at 4:00–5:00 and the next hour that Plume 2 was horizontally separated from Plume 1 and moved mainly over the inland MSs and caused the second maximum of 1310 Bq m⁻³ at Sugaya (4:50) and 1150 Bq m⁻³ at Ishikawa (5:00). The same figures also indicate the steep horizontal gradient of concentration at the southeast edge of the plume, which separated Plume 2 from Plume 1. It is also worthwhile noting that this plume did not significantly affected the SPM network except for the two westmost stations showing concentrations a little more than 100 Bq m⁻³ at 7:00–8:00. These 10-fold lower concentrations than those at Sugaya and Ishikawa indicate that the plume shifted more westward than did Plume 1.

Plume 3 was separated from Plume 2 by the period of low concentration well below 100 Bq m⁻³ at Muramatsu, Sugaya, and Ishikawa. The concentration exceeded 100 Bq m⁻³ again during the period of 6:10–7:10. The maximum concentration at each MS was estimated to be clearly higher at the coastal MSs, e.g. 2300 Bq m⁻³ at Araji (7:40). Like Plume 2, the east edge of Plume 3 seems very sharp in the horizontal distribution at 8:00–9:00. It would be reasonable to consider that this horizontal structure was caused by the onshore wind of fresh maritime air without radionuclides, causing a sea-breeze front across which the concentration gradient was extremely steep.

3.5. Concentration of ¹³⁷Cs

Concentrations of ¹³⁴Cs and ¹³⁷Cs in Plume 7 and 8 were estimated for the five MSs although they could not be estimated for the other earlier plumes due to the interference from co-existing nuclides such as ¹³¹I and ¹³²I at high concentration. To depict the horizontal evolution of plumes in terms of the ¹³⁷Cs air concentration, horizontal distributions of ¹³⁷Cs concentration estimated for Plume 8 in this study are plotted in Figure 12 with values determined by analyzing air filter tapes [1]. The concentration at Ishikawa was not available due to lack of pulse height distribution data. It was pointed out by our previous study that the ¹³⁷Cs concentration estimated in this study had tendency of underestimation by a factor of about 2 [6].

Figure 12. Horizontal distribution of ¹³⁷Cs air concentrations in Bq m⁻³ estimated in this study (diamonds) and determined by filter tape analysis (stars) [1]. Values in parentheses are the minimum and maximum of the 10 minute average air concentrations during the corresponding time period.

Display full size

It can be inferred from the air concentration in the central part of Ibaraki that Plume 8 approaching from the north had an axis which crossed the coast line onshore around Ajigaura at 4:00–5:00 (maximum of 390 Bq m⁻³ at 4:40) and Ohnuki next hour (270 Bq m⁻³ at 5:10). This axis reached the array of the SPM monitoring stations in the southern part of Ibaraki, first at Tsuchiura and Narita with concentrations more than 100 Bq m⁻³ at 6:00–7:00, then causing the maxima around 300 Bq m⁻³ at 7:00–8:00.

The tail of the plume, defined as the time when the concentration decreased below 10 Bq m⁻³, passed Araji at 6:40, Muramatsu at 7:00, and then Ajigaura and Ohnuki at 7:20, supporting the inferred location of plume axis around Ajigaura and Ohnuki. The concentration at Sugaya remained higher than 10 Bq m⁻³ until around 8:00, indicating westward shift of the plume at the central part of Ibaraki as discussed in Section 3.2. This shift was not seen at the southern part of Ibaraki.

3.6. Noble gas plume

3.6.1. Release and transport

As mentioned in Section 3.2, Plume 4 was found to be consisting mainly of noble gases (hereafter referred to as noble gas plume) without other nuclides at significant levels. The atmospheric transport process of the noble gas plume was analyzed to find time of release from FNPS1 by using atmospheric dispersion simulations with the model system developed by Hirao et al. [10]. The model system consisted of the weather research and forecasting model WRF [11] for three-dimensional meteorological simulation and a Lagrangian dispersion model. Model configuration and calculation conditions were identical with those of Hirao et al. except for the wider calculation area over the Pacific Ocean in the present study.

According to the results of simulation with an assumption of continuous release, it was found that only the effluents released around noon of 14 March had possibility to reach the southern part of Ibaraki around noon of the next day. It was found from the simulations, each of which assumed one-hour release with the release rate of 1 TBq h⁻¹ at different time around noon of 14 March, that the effluents during 10:00–11:00 and 11:00–12:00 were first transport northeastward during the afternoon, and then returned southwestward during the following half day, resulting in coverage of the southern part of Ibaraki by the plume as shown in Figure 13. The feature shown in Figure 10 that the higher concentrations have been estimated for southern MSs is consistent with the location of the plume calculated by the simulation. It is worthwhile noting that the hydrogen explosion of Unit 3 was on 14 March 11:01. Furthermore, according to our calculation, the deposition of radionuclides to the ocean during the transport was insignificant due to the absence of precipitation, and did not substantially change the nuclide composition of plume during the 24 hour travel. Therefore, it is highly likely that the hydrogen explosion at Unit 3 has not released radionuclides other than noble gases in large amount. The source term estimations [12,13] which assumed the very high release rates of ¹³¹I and ¹³⁷Cs due to the hydrogen explosion may have to be revised in this regards. It should also be pointed out that, according to the atmospheric dispersion simulation results, the maximum concentration of 150 kBq m⁻³ estimated in this study does not correspond to and is smaller than the maximum concentration of the noble gas plume.

Figure 13. Atmospheric dispersion simulation results of surface concentration for an atmospheric discharge from FNPS1 during 11:00–12:00 of 14 March (JST).

Display full size

3.6.2. Composition of noble gases

Figure 14 shows the pulse height distributions measured at Araji in 15 March 10:40 and 11:40. The count rate in full energy peak of ¹³³Xe at about 80 keV shows drastic increase during the last one hour. In addition, there was a peak at 220–240 keV. It is obviously unlikely that the peak was caused by isotopes of radioiodine or radiocesium because the peaks of their main members such as ¹³¹I and ¹³⁴Cs did not increase during the passage of the noble gas plume. The likeliest nuclide causing this peak is ¹³²Te since it has a full energy peak at 228 keV. However, this candidate should be turned down since there is no sign of increase in the count rate corresponding to its progeny ¹³²I. According to the inventory estimation for the reactors [9], noble gas nuclides other than xenon are also unlikely since their abundance in the reactor or emission probabilities of gamma ray in this energy range is low. Therefore, it can be concluded that nuclides that can contribute to this peak are ^131mXe, ^133mXe, and ¹³⁵Xe.

Figure 14. Pulse height distributions measured at Araji.

Display full size

Concentrations of four xenon nuclides including ¹³³Xe were estimated with the procedure described in Section 2 from the pulse height distribution measured at Araji on 15 March 11:40, which corresponds to the highest concentration of ¹³³Xe. The result is shown in Table 6, which also shows the activity ratios of nuclides on 15 March 12:00 calculated from the inventory in Unit 3 on 14 March 12:00 [9] by assuming only physical decay. Since ^133mXe and ¹³⁵Xe could not be separated by the energy resolution of the NaI(Tl) scintillation detector, their concentrations were estimated as a sum by assuming their activity ratio of inventory on 15 March 12:00 in Table 6. The energy ranges used in the analysis are 70–90 keV for ¹³³Xe, 155–175 keV for ^131mXe, and 220–240 keV for ^133mXe+¹³⁵Xe.

Table 6. Estimated concentrations of xenon isotopes (15 Mar. 11:40) and comparison with activity ratios R of inventory in Unit 3 [9] with respect to ¹³³Xe. The inventory was decay corrected to 12:00 Mar. 15.

CSV Display Table

As shown in Table 6, the activity ratios of estimated air concentrations of ^133mXe and ¹³⁵Xe are in reasonable agreement with those calculated from the inventory. On the other hand, the estimated activity ratio of ^131mXe is much larger than that calculated from the inventory. It was pointed out by comparing the reproduced pulse height distribution in the energy range from 100 to 180 keV for ^131mXe with the measured one, this erroneous estimation of concentration for ^131mXe was caused by the rather poor reproducibility of pulse height distribution in this energy range.

4. Conclusions

Air concentrations of key radionuclides in the plumes released from FNPS1 were estimated from the pulse height distribution data measured with NaI(Tl) scintillation detectors at MSs in the central part of Ibaraki prefecture for the period of 15–16 and 20–21 March, 2011. Characteristics of eight plumes that covered the objective area were discussed in terms of nuclide composition of the plumes and horizontal distribution of the air concentration. The activity ratios estimated in this study were found to be consistent with those evaluated from the direct measurements. It was pointed out that temporal changes in the air concentrations were different among nuclides even within a plume, with slower decrease for more gaseous nuclides.

Horizontal distribution maps of air concentrations of ¹³¹I and ¹³⁷Cs in Ibaraki were depicted, in which movements of plumes were clearly shown with concentration values. The highest air concentrations of ¹³¹I, as of 10 minute average values, were estimated to be around 2 kBq m⁻³ for the MSs in the coastal area in the morning of 15 March. The air concentrations of ¹³⁷Cs in the central part of Ibaraki estimated in this study are consistent with those from the filter tape analysis in the southern part of Ibaraki [1], although the values in this study are considered to be slightly underestimated. Very steep horizontal gradients of air concentration were found at the east to southeast edges of plumes, which were considered to be caused by onshore advection of fresh maritime air without significant contamination.

The air concentration of ¹³³Xe in Ibaraki was estimated for the first time, which showed the maximum air concentration of about 150 kBq m⁻³ around noon of 15 March. The air concentration ratio of ¹³³Xe to ¹³¹I during 15 March was found to be lower than that during the next day, indicating changes in processes and route of release and/or change in units from which the nuclides were released. It was also found that the plume around noon of 15 March was consisted mainly of noble gases, and the air concentration in the southern part of Ibaraki was substantially higher than that in the central part. According to the atmospheric dispersion simulation, it is highly likely that the plume corresponded to the hydrogen explosion at Unit 3. This result suggests that the explosion did not release radioiodine and radiocesium in large amount. It was also pointed out that the estimated air concentrations of xenon isotopes had activity ratios reasonably corresponding to the inventory of the reactors.

Table 1. Energy ranges used in the analysis of this study. Those written in bold letters are the main nuclides to be estimated from count rate of corresponding energy range.

Energy range (keV)	Radionuclides
70–90	¹³³Xe
210–250	¹³²Te
340–390	¹³¹I, ¹³⁶Cs
500–560	¹³³I, ¹³²I
630–700	¹³⁷Cs, ¹³¹I, ¹³²I
760–840	¹³⁴Cs, ¹³²I, ¹³⁶Cs
900–1000	¹³²I
1020–1080	¹³⁶Cs

Nuclide	C_air (Bq m⁻³)
¹³¹I	12.2
¹³²I	4.2
¹³³I	11.7
¹³²Te	15.2
¹³⁴Cs	5.2
¹³⁶Cs	4.3
¹³⁷Cs	13.5

	Plume
Muramatsu	1370	800	2220
Ajigaura	630	510	1320
Ohnuki	1040	730	2980
Araji	–	–	1600
Sugaya	300	240	450
Ishikawa	0	0	0

Table 4. Air concentration ratio of ¹³³Xe to ¹³¹I calculated from time integrated air concentration of each plume.

	Plume
MS	1	2	3	6	7	8
Muramatsu	25	60	53	102	5.5	13
Ajigaura	21	–	51	99	2.3	8.0
Ohnuki	19	–	61	130	6.1	19
Araji	17	–	49	–	3.6	24
Sugaya	15	62	58	130	2.4	14
Ishikawa	14	50	51	130	–	–

Table 5. Air concentration ratio of each nuclide to ¹³¹I. Values in parentheses are the values measured by Ohkura et al. [2].

	Plume
	1	2	3	6	7	8
¹³²I/¹³¹I	0.59 (0.52)	0.96 (0.71)	1.2 (0.93)	0.66 (0.30)	0.37 (0.26)	0.36 (0.16)
¹³³I/¹³¹I	0.22 (0.17)	0.22 (0.15)	0.23 (0.14)	0.11 (0.073)	–	–
¹³²Te/¹³¹I	0.61 (0.47)	1.1 (0.58)	0.95 (0.72)	0.37 (0.094)	0.19 (0.20)	0.27 (0.11)
¹³⁴Cs/¹³¹I	–	–	–	–	0.11 (0.18)	0.13 (0.14)
¹³⁶Cs/¹³¹I	–	–	–	–	0.029 (0.037)	0.039 (0.020)
¹³⁷Cs/¹³¹I	–	–	–	–	0.073 (0.19)	0.11 (0.14)

Table 6. Estimated concentrations of xenon isotopes (15 Mar. 11:40) and comparison with activity ratios R of inventory in Unit 3 [9] with respect to ¹³³Xe. The inventory was decay corrected to 12:00 Mar. 15.

			Estimated in this study		Inventory
Nuclides	Half life	Energy in keV (emission probability)	Concentration in kBq m⁻³	R	R
^131mXe	11.84 d	163.9 (1.9%)	26.3	0.18	0.0073
^133mXe	2.19 d	233.2 (10%)	3.19	0.022	0.018
¹³³Xe	5.24 d	80.9 (38%)	150	1	1
¹³⁵Xe	9.14 h	249.7 (90%)	0.444	0.0031	0.0025

Acknowledgments

This work was supported by the JSPS KAKENHI [grant number 24110002].

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

Japan Society for the Promotion of Science [24110002].

References

Tsuruta H, Oura Y, Ebihara M, et al. First retrieval of hourly atmospheric radionuclides just after the Fukushima accident by analyzing filter-tapes of operational air pollution monitoring stations. Sci. Rep. 2014;4:6717; doi:10.1038/srep06717. [Crossref], [Google Scholar]

Okura T, Oishi T, Taki M, et al. Emergency monitoring of environmental radiation and atmospheric radionuclides at Nuclear Science Research Institute, JAEA following the accident of Fukushima Daiichi Nuclear Power Plant. Tokai (Japan): Japan Atomic Energy Agency (Japan); 2012 May. 37p. Report No.: JAEA Data/Code 2012 010. [Google Scholar]

Furuta S, Sumiya S, Watanabe H, et al. Results of the environmental radiation monitoring following the accident at the Fukushima Daiichi Nuclear power plant –Interim Report (Ambient Radiation Dose Rate, Radioactivity Concentration in the Air and Radioactivity Concentration in the Fallout)-, JAEA-Review 2011-035. [Google Scholar]

Hirayama H, Kawasaki M, Matsumura H, et al. Estimation of I-131 concentration using time history of pulse height distribution at monitoring post and detector response for radionuclide in Plume. J Nucl Sci Technol. 2014;13:119–126. [Google Scholar]

Hirouchi J. Development of method for nuclide identification and radioactivity concentration with observation at monitoring station [Doctoral dissertation]. Nagoya (Japan): Nagoya University; 2015. 126p. Japanese. [Google Scholar]

Terasaka Y. Estimation of air and surface radioactivity concentrations from pulse height distribution measured by NaI(Tl) scintillation detector. Paper presented at: International Conference of Environmental Radioactivity; 2015 Sept 21–25; Thessaloniki, Greece. (To be published). [Google Scholar]

Hirayama H, Namito Y, Bielajew AF, et al. The EGS5 code system. Tsukuba (Japan): High Energy Accelerator Research Organization (KEK) (Japan); 2005. Report No.: KEK Report 2005 8, SLAC Report SLAC-R-730. [Crossref], [Google Scholar]

Namito Y, Nakamura H, Toyoda A, et al. Transformation of a system consisting of plane isotropic source and unit sphere detector into a system consisting of point isotropic source and plane detector in Monte Carlo calculation. J Nucl Sci Technol. 2012;49:167–172. [Taylor & Francis Online], [Web of Science ®], [Google Scholar]

Nishihara K, Iwamoto H, Suyama K. Estimation of fuel compositions in Fukushima-Daiichi nuclear power plant. Tokai (Japan): Japan Atomic Energy Agency (Japan); 2012. Report No.: JAEA Data/Code 2012 018. [Google Scholar]

Hirao S, Yamazawa H, Nagae T. Estimation of release rate of iodine-131 and cesium-137 from the Fukushima Daiichi nuclear power plant. J Nucl Sci Technol. 2013;50:139–147. [Taylor & Francis Online], [Web of Science ®], [Google Scholar]

Michalakes J, Dudhia J, Gill D, et al. The weather research and forecast model: software architecture and performance. Paper presented at: 11th ECMWF Workshop on the Use of High Performance Computing in Meteorology; 2004 Oct 25–29; Reading (UK). [Google Scholar]

Katata G, Chino M, Kobayashi T, et al. Detailed source term estimation of the atmospheric release for the Fukushima Daiichi Nuclear Power Station accident by coupling simulations of an atmospheric dispersion model with an improved deposition scheme and oceanic dispersion model. Atmos Chem Phys. 2015;15:1029–1070. [Crossref], [Web of Science ®], [Google Scholar]

Terada H, Katata G, Chino M, et al. 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. 2012;112:141–154. [Crossref], [PubMed], [Web of Science ®], [Google Scholar]

Article: Fukushima NPP Accident Related

Air concentration estimation of radionuclides discharged from Fukushima Daiichi Nuclear Power Station using NaI(Tl) detector pulse height distribution measured in Ibaraki Prefecture

Article: Fukushima NPP Accident Related

Air concentration estimation of radionuclides discharged from Fukushima Daiichi Nuclear Power Station using NaI(Tl) detector pulse height distribution measured in Ibaraki Prefecture

ABSTRACTFormulae display:?Mathematical formulae have been encoded as MathML and are displayed in this HTML version using MathJax in order to improve their display. Uncheck the box to turn MathJax off. This feature requires Javascript. Click on a formula to zoom.

1. Introduction

Published online:

2. Methods

2.1. Estimation of radioactivity concentrations

Published online:

Published online:

Table 1. Energy ranges used in the analysis of this study. Those written in bold letters are the main nuclides to be estimated from count rate of corresponding energy range.

2.2. Calculation of pulse height distribution

2.3. Separation of airborne and deposited radionuclides

2.4. Lower limit of air concentration estimation

Published online:

Table 2. Minimum concentrations Cair that suffice Equation (5).

2.5. Selection of MS

Published online:

3. Results and discussion

3.1. Air concentrations

Published online:

Published online:

Published online:

Published online:

Published online:

Published online:

Published online:

Table 3. Maximum values of overestimation in the 133Xe air concentration (Bq m−3).

3.2. Characteristics of plumes

Published online:

3.3. Radionuclide composition

Published online:

Table 4. Air concentration ratio of 133Xe to 131I calculated from time integrated air concentration of each plume.

Published online:

Table 5. Air concentration ratio of each nuclide to 131I. Values in parentheses are the values measured by Ohkura et al. [2].

3.4. Concentration of 131I

Published online:

3.5. Concentration of 137Cs

Published online:

3.6. Noble gas plume

3.6.1. Release and transport

Published online:

3.6.2. Composition of noble gases

Published online:

Published online:

Table 6. Estimated concentrations of xenon isotopes (15 Mar. 11:40) and comparison with activity ratios R of inventory in Unit 3 [9] with respect to 133Xe. The inventory was decay corrected to 12:00 Mar. 15.

4. Conclusions

Acknowledgments

Disclosure statement

Additional information

Funding

References

Related articles

Information for

Open access

Help and info

Keep up to date

ABSTRACT
Formulae display:?Mathematical formulae have been encoded as MathML and are displayed in this HTML version using MathJax in order to improve their display. Uncheck the box to turn MathJax off. This feature requires Javascript. Click on a formula to zoom.

Table 2. Minimum concentrations C_air that suffice Equation (5).

Table 3. Maximum values of overestimation in the ¹³³Xe air concentration (Bq m⁻³).

Table 4. Air concentration ratio of ¹³³Xe to ¹³¹I calculated from time integrated air concentration of each plume.

Table 5. Air concentration ratio of each nuclide to ¹³¹I. Values in parentheses are the values measured by Ohkura et al. [2].

3.4. Concentration of ¹³¹I

3.5. Concentration of ¹³⁷Cs

Table 6. Estimated concentrations of xenon isotopes (15 Mar. 11:40) and comparison with activity ratios R of inventory in Unit 3 [9] with respect to ¹³³Xe. The inventory was decay corrected to 12:00 Mar. 15.