Time-series analysis of atmospheric radiocesium at two SPM ... · Time-series analysis of atmospheric radiocesium at two SPM monitoring sites ... 2Department of Chemistry, ... 2011

103

Geochemical Journal, Vol. 52, pp. 103 to 121, 2018 doi:10.2343/geochemj.2.0520

*Corresponding author (e-mail: [email protected])

Copyright © 2018 by The Geochemical Society of Japan.

with 137Cs concentrations higher than 10 Bq m–3 weretransported to the Fukushima Prefecture or the Kantouarea including the Tokyo Metropolitan Area, during March12–23, 2011. Since then, the hourly atmosphericradionuclides in SPM collected on filter-tapes have beenmeasured at approximately 100 SPM monitoring stationsin the southern Tohoku area including the Fukushima Pre-fecture, and the Kantou area (Supplementary Fig. S1).This dataset has also been published online (Oura et al.,2015). These papers were evaluated as “Publications con-sidered to make a significant contribution to one or otheridentified research need” by the United Nations Scien-tific Committee on the Effects of Atomic Radiation(UNSCEAR, 2015, 2016). These data were used for thevalidation of the simulated results of two atmospheric

Time-series analysis of atmospheric radiocesium at two SPM monitoring sitesnear the Fukushima Daiichi Nuclear Power Plant just

after the Fukushima accident on March 11, 2011

HARUO TSURUTA,1* YASUJI OURA,2 MITSURU EBIHARA,2 YUICHI MORIGUCHI,3

TOSHIMASA OHARA4 and TERUYUKI NAKAJIMA5

1Remote Sensing Technology Center of Japan, 3-17-1 Toranomon, Minato-ku, Tokyo 105-0001, Japan2Department of Chemistry, Tokyo Metropolitan University, 1-1 Minami-Ohsawa, Hachioji, Tokyo 192-0397, Japan

3Graduate School of Engineering, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8656, Japan4Fukushima Project Office, National Institute for Environmental Studies, 16-2 Onogawa, Tsukuba 305-8506, Japan

5Earth Observation Research Center, Japan Aerospace Exploration Agency, 2-1-1 Sengen, Tsukuba, Ibaraki 305-8505, Japan

(Received September 6, 2017; Accepted January 25, 2018)

Using an hourly-resolution time series of the Fukushima radionuclides collected on used filter-tapes installed in sus-pended particulate matter (SPM) monitors, we measured the hourly radiocesium values at the SPM monitoring sites ofFutaba and Naraha located within 20 km of the Fukushima Daiichi Nuclear Power Plant (FD1NPP) during March 12–25,2011. The time-series of the 137Cs concentrations at the sites were analyzed and compared with radiation dose rates at themany monitoring posts/points of Fukushima Prefecture and the Tokyo Electric Power Company. At Futaba, nine plumes ofhigh 137Cs concentrations were found on March 12–13, 15–16, 18–20, and 24–25, 2011, when southeasterly winds pre-vailed. On March 12, the first peak of the 137Cs concentrations was detected at Futaba at 9:00 Japanese Standard Time(JST) due to the first release from reactor Unit 1 (U1) in the early morning. Furthermore, the highest 137Cs concentration,i.e., 13,600 Bq m–3 was observed at 15:00 JST after a vent operation at U1, just before the hydrogen explosion of U1 at15:36 JST. On the afternoon of March 15, plumes from the FD1NPP were observed at Futaba due to a constant southeast-erly wind and were then transported to downwind, resulting in the formation of a highly deposited zone of radionuclidesspanning more than 30 km from near the FD1NPP to the northwest. In contrast, seven plumes of high 137Cs concentrationswere found at Naraha on March 15–16, 18, 20–21, 2011, when northerly winds prevailed. On March 15, a plume causedby the first release from Unit 2 was observed at Naraha at 1:00 JST, and the highest concentration, i.e., 8,300 Bq m–3, wasobserved at 3:00 JST, and then were transported southward to the Kantou area. The activity ratios of 134Cs/137Cs in theplumes were divided into two groups. The plumes at Futaba on March 12–13, which had ratios of 0.92–0.94, are identifiedto be released from U1, compared to its ratio of 0.94, as derived from the inventory data. All other plumes with the ratiosof 1.02–1.04 at Futaba and Naraha during March 15–21 have not been determined to be released from U2 and/or Unit 3.

Keywords: hourly rediocesium concentrations, SPM monitoring sites, filter tape, activity ratio of 134Cs/137Cs, Fukushima Daiichi Nu-clear Power Plant

INTRODUCTION

The spatiotemporal distributions of atmospheric 137Csconcentrations were first retrieved by using filter-tapesinstalled in 40 suspended particulate matter (SPM) moni-tors in the air pollution monitoring network of easternJapan just after the Tokyo Electric Power Company(TEPCO) Fukushima Daiichi Nuclear Power Plant(FD1NPP) accident during March 12–23, 2011 (Tsurutaet al., 2014). Analysis of these data clarified that 9 plumes

104 H. Tsuruta et al.

transport models (Nakajima et al., 2016). The filter-tapesused at more than 600 SPM stations had been offered bylocal governments in eastern and central Japan throughthe Ministry of the Environment, Japan. In addition, wehave continued to measure the radionuclides of filter-tapesfrom the other SPM stations.

On the other hand, the data of atmospheric concentra-tions of radionuclides near the FD1NPP have been verylimited until now. The Fukushima Prefecture and theMinistry of Economy, Trade and Industry (METI) urgentlymeasured atmospheric radionuclides once a day duringMarch 12–13, 2011 with the sampling time of 10 minutesby dust samplers at several points located within 2–35km of the FD1NPP (METI, 2011). In addition, the Minis-try of Education, Sports, Culture, Science and Technol-ogy (MEXT) (MEXT, 2012) began the field measurementsof atmospheric radionuclides once or a few times per dayfrom March 18, 2011 with the sampling time of 10 or 20minutes by dust samplers at more than 10 points in theeastern Fukushima Prefecture. Furthermore, TEPCO

(2011, 2013) also started to measure atmosphericradionuclides by dust samplers once or two times per dayon March 19, 2011 at the site of the FD1NPP and theFukushima Daini Nuclear Power Plant (FD2NPP). Thesedata were valuable, but are too sparse spatiotemporallyto understand the detailed behavior of radionuclides nearthe FD1NPP just after their release into the atmosphere.

In the Kantou area located more than 100 km south ofthe FD1NPP, Nuclear Science Research Institute (NSRI),Japan Atomic Energy Agency (JAEA), located in the eastcoast of the Ibaraki Prefecture, started to measure atmos-pheric radionuclides from March 15, 2011 (Ohkura et al.,2012). The measurements by NSRI were made with a sam-pling time of 20 minutes until March 21, 2011, only whenthe radiation dose rates (RDs) simultaneously measuredby NaI (Tl) scintillation detectors increased. As a result,most of the plumes transported from the FD1NPP to theKantou area by northeasterly winds were first observedat this site. In addition, Tokyo Metropolitan IndustrialTechnology Research Institute (TIRI) started independ-

Fig. 1. Map of (a) the eastern part of Fukushima Prefecture and (b) SPM stations (�) in the air pollution monitoring network andmonitoring posts for environmental radioactivity (�) by Fukushima Prefecture in Hamadori.

Time-series of radiocesium at two SPM monitoring sites after Fukushima accident 105

ently to measure atmospheric radionuclides every houror two hours from March 12, 2011, at Fukasawa in Tokyo(Nagakawa et al., 2014). We have validated the accuracyof our measurements of radionuclides in SPM collectedon filter-tapes, comparing the 137Cs concentrations meas-ured at Kakinokizaka, an SPM monitoring site locatedwithin 2 km of the TIRI monitoring site, with those meas-ured at Fukasawa by TIRI, at the same hour in March 12–13, 2011 (Tsuruta et al., 2014; Oura et al., 2015).

Our dataset is unique in its time resolution and manymonitoring sites, because hourly atmospheric concentra-tions from March 12, 2011 have been mainly obtained inurban/industrial areas where many people have lived. Thepurpose of the SPM measurements of the air pollutionnetwork is to protect the population’s health in the pres-ence of air pollution. Among the datasets from the SPMmonitoring stations in the Fukushima Prefecture, hourly137Cs concentrations from March 12, 2011 are availableat 17 SPM monitoring sites in Nakadori, located in thecentral area and Hamadori in the east coast (Fig. 1a andFig. S1). The nearest SPM station in northern Hamadoriin a previous dataset (Oura et al., 2015) is located 25 kmnorth of the FD1NPP (Fig. 1a and Fig. S1). In contrast,the hourly 137Cs data in southern Hamadori were avail-able only from the afternoon of March 22, 2011 at 2 SPMsites (Fig. S1). Since no data are available from southernHamadori during March 12–22, 2011, the transport ofplumes to the south from the FD1NPP was only detectedat the SPM stations in central Kantou located more than170 km southwest of the FD1NPP (Fig. S1).

In contrast, the spatial distribution of radiocesium andradioiodine and/or radiation dose rates on the surface soiloutside the FD1NNP were measured in the field meas-urements by the MEXT (MEXT, 2011a). In addition, itwas measured by using airplanes (MEXT, 2011b; Torii etal., 2013) and an unmanned helicopter (Sanada and Torii,2015). These datasets have many data of radionuclides,however, are not available for the investigation of thehourly atmospheric radionuclides concentrations, andmight have missed if the polluted air masses passed with-out precipitation. Hence, the source term identificationof plumes still has a large uncertainty, although many stud-ies have been reported (Chino et al., 2011, 2016; Katataet al., 2015).

Recently, the filter-tapes at another two SPM moni-toring stations of Futaba and Naraha located within 20km from the FD1NPP (Fig. 1b) had been provided by theFukushima Prefecture, and then, the hourly atmosphericconcentrations of radiocesium at the two sites were meas-ured. A purpose of this paper is to reveal how theradionuclides near the FD1NPP were transported to thenorth/northwest and the south in Hamadori by showingthe time-series of radiocesium at the two sites. The otherpurpose of this paper is to make clear from which reactor

unit the radionuclides were released into the atmospherediscussing activity ratio of 134Cs/137Cs.

MATERIALS AND METHODS

Measurements of atmospheric radiocesium at two SPMsites near the FD1NPP

The two SPM monitoring stations of Futaba (37.44∞N,141.01∞E) and Naraha (37.27∞N, 141.00∞E) are locatedalong Hamadori, the east coast of the Fukushima Prefec-ture (Fig. 1b). The Futaba and Naraha SPM sites are lo-cated 3.2 km west-northwest and 17.5 km south-south-west of the center between reactor Unit 1 and Unit 2 ofthe FD1NPP, respectively. Before the measurement of theradionuclides in the SPM collected on these filter-tapes,we carefully checked whether these SPM monitors wereregularly operated even after the Great East Japan Earth-quake and the tsunami on March 11, 2011. Consideringthis from various points of view, we found that the SPMmonitors were operated normally, and SPM was properlycollected on the filter-tapes. Then, 134Cs and 137Cs in SPMwere measured during March 12–25, 2011 (March 12–23for Naraha) by a Ge semiconductor detector, as alreadydescribed by Tsuruta et al. (2014) and Oura et al. (2015).In a previous study, the hourly 134Cs and 137Cs concen-trations at the 99 sites have been shown in Appendix A(Journal of Nuclear and Radiochemical Sciences, 2015)followed by an article (Oura et al., 2015). Those at Futabaand Naraha are shown in Supplementary Tables S1 andS2 of this article. These concentrations are decay-cor-rected to the time when the collection of SPM samplesstopped. For example, the 137Cs concentration at 15:00Japanese Standard Time (JST; hereinafter JST is omit-ted) is derived from an averaged concentration of 137Csbetween 14:00–15:00. The detection limit of 134Cs and137Cs was approximately 0.1–0.4 Bq m–3 for the hourlymeasurements. The measurement errors due to gamma-ray counting statistics for each SMP sample were shownin Tables S1 and S2.

As previously reported by Tsuruta et al. (2014) andOura et al. (2015), part of the high SPM collected at asample spot might be attached on the backside of the newsample or the blank spot located one round after the pre-vious sample spot, which is defined as cross-contamina-tion. Analyzing all the data, the maximum cross-contami-nation error was 3–5% for the filter-tapes made of glassfiber (GF) (Oura et al., 2015). Although used filter-tapesat the two sites were made of GF, we carefully used auto-radiography imaging with an imaging plate to check forseriously affected data with cross-contamination, whichare treated as missing data and are represented with a “+”symbol in the Tables. Furthermore, data with “=” are notcertified values, but reference values.


Fig. 2. Time series of (a) the hourly 137Cs concentrations at Futaba and Haramachi, the radiation dose rates at Kamihatori andYamada, and the precipitation at Haramachi during March 12–25, 2011, (b) the wind direction from every ten minutes at theFD2NPP, and (c) the hourly 137Cs concentrations at Naraha, the radiation dose rates at the FD2NPP and Yamadaoka, and theprecipitation at Iwaki during March 12–23, 2011. p1-P11 mean the plume number as listed in Table 1. DL means the 137Csconcentration below detection limit.

Radiation dose rates, meteorological dataThe Environmental Radioactivity Monitoring Center

of the Fukushima Prefecture has a monitoring networkfor radioactivity measurements outside the FD1NPP and

FD2NPP (Fig. 1), and the radiation dose rates (RDs) wereroutinely measured at 24 monitoring posts (MPs)(Fukushima Prefecture, 2012). Although several MPsstopped just after March 11, 2011, the hourly RDs were


used for analysis at Kamihatori and Yamada, located atthe nearby Futaba SPM site, and at Shigeoka andYamadaoka, located at the nearby Naraha SPM site. Fur-thermore, RDs were also used at Namie, Ottozawa, Ohno,Mukaihata, Shimokoriyama, Matsudate, and Futatsunuma(Fukushima Prefecture, 2012) and at Minamisoma andIitatemura (Fukushima Prefecture, 2011, 2016). In themonitoring network, meteorological elements were alsomeasured at Koriyama and Shigeoka, and Ohno wheremeasurements stopped on March 15 and 16, 2011, respec-tively.

Other meteorological data were also used at the Iitate,Haramachi and Iwaki stations in the Automated Mete-orological Data Acquisition System (AMeDAS) networkof the Japan Meteorological Agency (JMA). In addition,a map of wind distribution at a height of 1000 hPa (e.g.,model’s lowermost layer winds) every three hours inMarch 2011 was used, which was calculated withmesoscale objective analysis by the JMA. Then, two-di-mensional wind vectors at 1000 hPa were superimposedon a map of atmospheric 137Cs concentrations at the SPMsites in the east coast of the Fukushima Prefecture.

TEPCO (2011) also measured RDs, and wind direc-

tion and speed at several monitoring points at a height ofaround 2 m above the ground by monitoring cars at thesite of the FD1NPP. The temporal monitoring points werethe main gate (MG), the west gate (WG), a point of thegymnasium (GM), the environmental management build-ing (EMB), a north point of the administration building(AB), and the monitoring posts of MP1–MP8 (Supple-mentary Fig. S2). On the site of the FD2NPP, located 10km south of the FD1NPP, TEPCO (2012) routinely meas-ured RDs at 6 monitoring posts (MP1–MP6) and the winddirection and speed at the top of the exhaust stack, 120 mabove ground level. The Fukushima Prefecture and theMETI (METI, 2011) measured atmospheric radionuclidesby dust samplers at several points located within 2–35km from the FD1NPP during March 12–13, 2011. Thesedata were also analyzed in this study.

RESULTS AND DISCUSSION

Plumes measured at the Futaba and Naraha SPM sitesduring March 12–25, 2011

Time-series of the hourly 137Cs concentrations atFutaba and Naraha, are shown in Fig. 2 for March 12–25

Table 1. Days and areas of the transport of plumes/polluted air masses (p1–P11)

No.: Number of major plumes transport from the FD1NPP just after the accident.Hamadori, Nakadori, and Kantou: The location of these areas is shown in Figs. 1b and 1c.�: Maximum 137Cs concentrations ≥ 100 Bq m–3.�: Maximum 137Cs concentrations < 100 Bq m–3.Æ: High 137Cs concentrations continued to the following morning.�: Precipitation was observed in the area.Ref. 1: Tsuruta et al. (2014).Ref. 2: Tsuruta et al. (2017).


and March 12–23, 2011, respectively. The time series ofthe wind direction measured every 10 minutes at theFD2NPP is also shown in Fig. 2b to clarify the prevailingwinds that transported plumes from the FD1NPP. Al-though the wind direction and wind speed were also meas-ured at a height of 2 m above the ground by a monitoring

car on the site of the FD1NPP just after the accident(TEPCO, 2011), these data are omitted in Fig. 2. The winddata at the FD1NPP, however, were used for detailedanalysis, because they presented more local informationthan those at the FD2NPP, especially during the nighttimewhen ambient air near the surface was stagnant.

Fig. 3. Time series of (a) the hourly 137Cs concentrations at Futaba, and the radiation dose rates at Koriyama, Kamihatori,Yamada, and (b) the wind direction and speed at Koriyama during March 12–15, 2011. Horizontal bars (a, b, and c) show theperiods of sea breezes at Koriyama.


In this study, the definition of the number of plumesat Futaba and Naraha in Figs. 2a and 2c, respectively, isas follows. As we already found nine plumes of P1–P9(Tsuruta et al., 2014) and one plume of p2 (Tsuruta et al.,2017) from the spatiotemporal distributions of the hourly137Cs concentrations (Table 1), we have fixed these num-bers in this study. In the previous study, we found 5 plumes(P1, P3, P5, P6, and P8) transported to Nakadori and/ornorthern Hamadori. In Fig 2a, the hourly 137Cs concen-trations measured at the site of Haramachi in northernHamadori (Fig. 1a), located 25 km north of the FD1NPP,are also shown. For the two 137Cs peaks at Futaba meas-ured on the daytime of March 12, a few or several hoursbefore the plume of P1 measured at Haramachi on thenighttime of March 12 (Tsuruta et al., 2014), we namedthe new plumes of p1 and P1v, different from plume P1as described in detail later. Another new plume was meas-ured at Futaba on March 13, and was named P1¢ becauseit was found several hours after the plume of P1 atHaramachi in the previous study. In contrast, for the three137Cs peaks found at Futaba on March 15 and March 20,2011, (Fig. 2a), their general terms of plume are P3 andP5, respectively, because in a previous study (Tsuruta etal., 2014), the plumes or polluted air masses of P3 andP5 widely moved both spatially and temporally throughNakadori and Hamadori for one or two days.

On the other hand, no plumes transported to southernHamadori were listed in our previous studies, because nofilter-tapes at the SPM sites in southern Hamadori wereavailable for March 12–21, 2011 (Tsuruta et al., 2014;Oura et al., 2015). Another five plumes (p2, P2, P4, P7,

and P9) transported southwards, were found at the SPMsites in the Kantou area, 170 km south of the FD1NPP(Tsuruta et al., 2014, 2017). The atmospheric 137Cs con-centrations in southern Hamadori were first measured atNaraha in this study. In the same way as the case at Futaba,we named the same number of plumes for the 137Cs peaksat Naraha, as that already used in the Kantou area, com-paring the possible pathways one by one. Furthermore,for the 137Cs peaks measured at Naraha that were not trans-ported to the Kantou area, a new set of plumes was namedP4¢ and P9¢, which were measured just after the plume ofP4 and P9, respectively. We called the 137Cs peaks meas-ured at Naraha on March 18 and 20 that were not trans-ported to the Kantou area, P5¢ and P8¢, modifying thenumber of plumes (P5 and P8) that were measured atFutaba on the same day. Hence, one of the major findingsin this study is that 9 new plumes were identified, in ad-dition to the 10 plumes already found in the previous stud-ies: 5 (p1, P1v, P1¢, P10, P11) at Futaba, and 4 (P4¢, P5¢,P8¢, and P9¢) at Naraha. All the old and new plumes/pol-luted air masses are shown in Table 1.

A typical prevailing wind along Hamadori in the earlyspring season is from the west, and this wind pattern ap-peared on the afternoon of March 14, on March 17 andMarch 22 (Fig. 2b). Northerly winds also prevailed fromthe night of March 14 to the morning of March 15, on themorning of March 16 and March 21, and on March 23.As a result, the high 137Cs concentrations greater than 100Bq m–3 at Narana were observed on the early morningsof March 15, 16, and 21, corresponding to the plumes ofP2, P4 and P4¢, P9 and P9¢, respectively. In contrast, south-

Fig. 4. Time series of the hourly radiation dose rates at the monitoring points in the FD1NPP (TEPCO, 2011) and the precipita-tion at Ohno during March 12–16, 2011. Numbers mean the local hours (JST) at around the peak time.


erly or southeasterly winds prevailed most frequentlyduring this period in Hamadori, specifically during thedays of March 12, 13, 15, 18, 19, 20, 24, and 25, corre-sponding to the following plumes observed at Futaba: p1,P1v, P1¢, P3, P5, P6, P8, P10, and P11, respectively. Inthe following section, the transport of each plume isbriefly described in chronological order, regardless ofwhether a plume was transported to the north or the southfrom the FD1NPP.

[p1] at Futaba on the morning of March 12 The firstplume (p1) of the high 137Cs concentration at Futaba wasobserved to be 55.7 Bq m–3 at 9:00 on March 12, 2011.At Yamada, the RDs showed the first peak of 32.5 mSvh–1 at 10:00, one hour later than that at Futaba (Figs. 2aand 3a). At Koriyama, the RDs showed the first peak of2.9 mSv h–1 at 6:00 (A in Fig. 3a) under stagnant condi-tions, and the second peak of 7.8 mSv h–1 at 9:00, after asea breeze began to blow from the southeast (Fig. 3b). Atthe main gate (MG) and MP-8 in the FD1NPP, RDs be-gan to increase from 4:05 and showed a maximum of385.5 mSv h–1 at the MG at 10:30 (4 and 11 in Fig. 4).The atmospheric 137Cs concentrations at three emergentmonitoring points by METI within 2–10 km from theFD1NPP were in the range of 1.3–2.2 Bq m–3 (12–63 Bqm–3 for 131I) at around 9:00 and 13:30 (METI, 2011).Hence, these enhancements in the RDs or 137Cs outsidethe FD1NPP were caused by the earliest release ofradionuclides from the FD1NPP into the atmosphere bythe southeasterly wind. Furthermore, the RDs at Ottozawa,located south of the FD1NPP (Fig. 1b) also increased onthe morning of March 12 (Fig. 3a), because the wind di-rection was variable early in the morning at Koriyamaand the FD1NPP.

At the SPM site of Haramachi, located 25 km north ofthe FD1NPP, the first plume was already found duringthe nighttime on March 12, 2011 and named as P1 (Fig.2a) (Tsuruta et al., 2014). At Haramachi, the atmospheric137Cs concentrations were high with a maximum 137Csconcentration of 575 Bq m–3 from 19:00 on March 12 tothe following morning, after the hydrogen explosion ofUnit 1. Because the first plume at Futaba was observed at9:00 on March 12, 2011, and was caused by the directrelease from the FD1NPP, the number of this plume wasnamed as p1.

[P1v] at Futaba on the afternoon of March 12 A maxi-mum 137Cs concentration of 13,600 Bq m–3 was observedat Futaba at 15:00 on March 12 (Figs. 2a and 3a), justbefore the hydrogen explosion (H.E.) of Unit 1 at 15:36.It was the highest 137Cs concentration at Futaba duringMarch 12–25, 2011. A southeasterly wind continued toblow towards Koriyama during 8:00–18:00 on March 15(a in Fig. 3b). At the same time, the RDs at Kamihatori,

2.5 km northwest of Futaba (Fig. 1b), also showed a maxi-mum of 1.6 mSv h–1, indicating that the same plume (P1)was transported to Futaba and Kamihatori by the south-easterly wind. On the other hand, the RDs at Yamada, 2km southwest of Futaba, increased to only 6.7 mSv h–1 at14:00 (Figs. 2a and 3a). This result suggests that the west-ern edge of the plume was located near Yamada. Assum-ing that the 131I concentration is about ten times higherthan the 137Cs concentration in the plume (Katata et al.,2015), the sum of the radiocesium (134Cs+137Cs) and 131Iconcentrations is equivalent to approximately 17 mSv h–1

when using the conversion factors by the InternationalAtomic Energy Agency (2000). Hence, the radiation dosecorresponding to the three radionuclides at Futaba at 15:00is approximately 1% of the radiation dose at Kamihatori,supposing that the air masses at 15:00 at Futaba andKamihatori were the same ones. The major radionuclidein the air masses is estimated to be radioactive noble gases,such as radioactive xenon, as Hirayama et al. (2017) cal-culated by analyzing the data of the pulse height distri-bution from a NaI (T1) counter at Yamada on March 12,2011. It is reasonable that radioactive noble gases are themajor radionuclides in the plume released into the atmos-phere from a reactor unit during the first or early stagesjust after the accident.

The plume with the maximum 137Cs concentration atFutaba at 15:00 was caused by a vent operation of U1before the hydrogen explosion, as described later. As aresult, a composition ratio in radionuclides released intothe atmosphere from the FD1NPP after the vent opera-tion is considered to be significantly different from thatafter the hydrogen explosion. For these reasons, thenumber of this plume was named as P1v.

[P1¢] at Futaba on March 13 A plume of P1¢ with a 137Csconcentration of 511 Bq m–3 was detected at Futaba at9:00 on March 13 (Figs. 2a and 3a), and the RDs at 10:00at Yamada (Fig. 2a) and at the MP4 of the FD1NPP (9 inFig. 4) also recorded a peak of 25.7 and 122.3 mSv h–1,respectively. As a sea breeze from the southeast was ob-served at Koriyama from 8:30 to 9:30 (b in Fig. 3b), theplume was transported from the FD1NPP to Futaba bythe sea breeze. In addition, at Ottozawa and Mukaihatalocated south and southwest of the FD1NPP, respectively,the peaks of the RDs were found at approximately 8:30and 9:30 (data were not shown), when the wind directionwas variable at the FD1NPP. This plume is a newly fea-ture and named as P1¢ in this study, because it was notfound on the morning of March 13 at Haramachi locatednorth of Futaba in the previous study (Tsuruta et al., 2014)(Fig. 2a). The 137Cs concentration at Futaba also showedthe second maximum of 50.2 Bq m–3 at 18:00 (B in Fig.3a). At Koriyama, a sea breeze blew again between 14:00–17:30 when the RDs also increased to 33 and 28 mSv h–1


Fig. 5. Time series of (a) the hourly 137Cs concentrations at Naraha and the hourly radiation dose rates at Shigeoka, (b) the winddirection and wind speed every 20 seconds at Shigeoka, during March 14–15, 2011. (c) The atmospheric 137Cs concentrations(colored dot) at the SPM sites including Haramachi, Futaba, and Naraha, in Hamadori and wind vectors (black arrow) at 1000hPa at 0:00, 3:00, 6:00, and 9:00 (JST), March 15, 2011.


at 17:00 and 18:00 (c in Fig. 3b), and the RDs at the MGand MP4 of the FD1NPP also increased in the afternoon(15 in Fig. 4). These data suggest that the air masses witha high 137Cs concentration at 18:00 at Futaba (B in Fig.3a) were also transported from the FD1NPP.

[P2] at Naraha from the night of March 14 to the morn-ing of March 15 A westerly wind shifted clockwise and anortherly wind began to blow at the FD1NPP, FD2NPPand Shigeoka at approximately 22:00 on March 14 (Figs.2b and 5b). The 137Cs concentrations at Naraha started to

increase from 23:00 on March 14, and reached maximumsof 2,780 and 8,300 Bq m–3 at 1:00 and 3:00 on March 15,respectively (Figs. 2c and 5a). The highest 137Cs concen-tration was comparable to that at Futaba on March 12,and at the same time, the RD peaks of 119 and 146 mSvh–1 were observed at Shigeoka and Yamadaoka, respec-tively (Figs. 2c and 5a). Assuming that the 131I concen-tration was about ten times higher than the 137Cs concen-tration in the plume (Katata et al., 2015), the sum ofradiocesium (134Cs+137Cs) and 131I concentrations isequivalent to approximately 11 mSv h–1. Hence, the ra-

Fig. 6. Time series of (a) the hourly 137Cs concentrations at Futaba, the radiation dose rates at Yamada, Ohno, Iitatemura, andthe MG and WG of the FD1NPP, and the precipitation at Ohno, (b) the wind direction and wind speed at Ohno, during March 14–16, 2011.


diation dose corresponding to the three radionuclides atNaraha at 3:00 is approximately 7% of the RD atYamadaoka, supposing that the air masses arrived at 3:00at Naraha and Yamadaoka were the same ones. This plumewas transported southward to the Kantou area as the plumeof P2 by a northeasterly wind through the east coast ofthe Ibaraki Prefecture (Tsuruta et al., 2014; Nakajima etal., 2016). Another 137Cs maximum of 278 Bq m–3 wasobserved at Naraha at 7:00 on March 15, but it was nottransported to the Kantou area but to southern Nakadoriby a northeasterly or easterly wind (Fig. 5c).

In the Kantou area, the first plume of p2 with a maxi-mum 137Cs concentration of 14 Bq m–3 was observed onthe early morning of March 12, 2011 by a northeasterlywind, a few hours before the arrival of plume P2 (Tsurutaet al., 2017). Part of the air masses with a 137Cs concen-tration of 20 Bq m–3 at Naraha at 23:00 (p2 in Fig. 5a)would be transported in the Kanto area as the plume ofp2, and further analysis is needed.

[P3] at Futaba on the afternoon of March 15 After thewind direction shifted clockwise to a southeasterly windat approximately 12:00 at the MG of the FD1NPP andOhno (Fig. 6b), high 137Cs concentrations of 607, 1,600,and 1,560 Bq m–3 were observed at Futaba at 14:00, 17:00,and 21:00, respectively (Fig. 2a and Fig. 6a), while theRDs at Ohno and the MG of the FD1NPP did not increaseduring the afternoon (Figs. 4 and 6a). In contrast, atYamada, the high RDs were observed at 13:00 and 19:00,

Fig. 7. Time series of the hourly 137Cs concentrations at Futaba, and the radiation dose rates at Minamisoma, Namie, and theFD1NPP (MG, WG, AB) during March 17–25, 2011.

when the 137Cs concentrations were low at Futaba (Fig.6a). These data indicate that the air masses with high 137Csconcentrations were constantly transported from theFD1NPP to the northwest during the afternoon. Further-more, the RDs at Iitatemura, located 39 km northwest ofthe FD1NPP (Fig. 1a), rapidly increased from 15:00 onMarch 15 to reach a maximum of 44.7 mSv h–1 at 18:20and then gradually decreased (Fig. 6a). At the AMeDASIitate station, an easterly or southeasterly wind prevailedduring 9:00–20:00 and precipitation started at 16:30 onMarch 15 (the data were not shown). In addition, the 137Csconcentrations at the Fukushima city, located approxi-mately 63 km northwest of the FD1NPP (Fig. 1a and Fig.S1b), started to increase from 18:00 (Tsuruta et al., 2014).These data indicate that the plumes observed at Futabawere transported downwind by the southeasterly wind,and part of radionuclides in the plumes was deposited onthe surface on the afternoon of March 15. Hence, themechanism of how the highly deposited zone ofradionuclides was produced for a 30 km span from nearthe FD1NPP to the northwest, is partly confirmed by sur-face-soil measurements (MEXT, 2011a) and airbornemeasurements (MEXT, 2011b). Part of the high 137Csconcentrations at Futaba at 21:00, however, might not betransported to the northwest, because the wind directionat Ohno shifted clockwise to the southwest at 19:00 andto the northwest at 22:00 (Fig. 6b). A sudden increase inRDs was observed at Yamada, Ohno, and the MG of theFD1NPP from approximately 23:00 (C in Fig. 6a). Moreo-


ver, a high 137Cs concentration of 116 Bq m–3 was ob-served at Futaba at 0:00 on March 16. Meanwhile, pre-cipitation started at 22:30 at the MG of the FD1NPP andat Ohno, resulting in a large amount of radionuclides be-ing deposited around the FD1NPP, as estimated by Chinoet al. (2016). These data can contribute to clarifying howthe deposition area located west of the FD1NPP was pro-

duced (Nishizawa et al., 2016; Chino et al., 2016), andfurther analysis will be done in future. As the plume ofP3 was widely observed in Nakadori and northernHamadori from the morning to the night on March 15(Tsuruta et al., 2014), the three plumes at Futaba describedabove were also included to plume P3.

Fig. 8. Time series of (a) the hourly 137Cs concentrations at Futaba and the radiation dose rates at Yamada, (b) the winddirection and speed at the AB or the WG (2:00–11:30 on March 19) of the FD1NPP during March 19–20, 2011.


Fig. 9. The atmospheric 137Cs concentrations (colored dot) at the SPM sites in Hamadori and wind vectors (black arrow) at 1000hPa at 0:00, 3:00, 6:00, and 9:00 (JST), March 21, 2011.

concentrations of 287–440 Bq m–3 were observed atHaramachi during 19:00–21:00 (Tsuruta et al., 2014). Onthe other hand, at Naraha, a plume with a 137Cs concen-tration of 40 Bq m–3 for 137Cs (P5¢ in Fig. 2c) was ob-served at 15:00 when an easterly or southeasterly windblew at the FD2NPP. After the plume was initially trans-ported off the coast from the FD1NPP in the morning bya northwesterly wind, and the plume was likely to returnto the east coast to leach Naraha in the afternoon (Sup-plementary Fig. S4). Because the arrival time of the plumeat Naraha was almost equal to that of plume P5 at Futaba,the number of the plume was named as P5¢

[P6] at Futaba on the morning of March 19 On March19, high 137Cs concentrations were observed during 9:00–13:00 at Futaba with the maximum concentration of 3,080Bq m–3 occurring at 10:00 (P6 in Fig. 2a and Figs. 7 and8). At Yamada, the RDs increased to 315 mSv h–1 at 10:00,compared to approximately 280 mSv h–1 before or afterthe peak (Fig. 8a). At the WG of the FD1NPP, an easterlyor southeasterly wind blew during 8:20–10:30 (Fig. 8b),and the RDs increased by 540 mSv h–1 at 8:10 (Fig. 7).These increases also suggest the transport of radionuclidesfrom the FD1NPP to Futaba in the morning. At Haramachi,a maximum 137Cs concentration of 192 Bq m–3 was ob-served at 12:00 (Fig. 2a), and this plume was alreadynamed P6 (Tsuruta et al., 2014). Accordingly, theseplumes measured at Futaba and Haramachi were consid-ered to be the same ones.

[P8] at Futaba and Naraha on March 20 At Futaba onMarch 20, three plumes were observed at 7:00, 10:00,and 14:00, with the maximums of 156, 382, and 5,680 Bq

[P4 and P4¢] at Naraha on the morning of March 16 AtNaraha, a plume with the 137Cs concentration of 219 Bqm–3 was observed at 2:00 on March 16, when a northerlywind prevailed (Figs. 2b and 2c). This plume was trans-ported southward as plume P4 to the east coast of theKantou area (Tsuruta et al., 2014; Nakajima et al., 2016).A precipitation measured at Iwaki early in the morning(Fig. 2c) suggests that some amounts of radionuclideswere deposited around the southern part of Hamadori.Several hours later, the wind direction shifted to the north-east from the north, and the plume of P4¢, with a maxi-mum 137Cs concentration of 988 Bq m–3, also reachedNaraha at 10:00, due to its direct transport from theFD1NPP or via a return path from the east coast after theplume from the FD1NPP initially went off the coast (Sup-plementary Fig. S3). The plume of P4¢ was not transportedto the Kantou area where a strong northwesterly windprevailed.

[P5 and P5¢] at Futaba and Naraha on the afternoon ofMarch 18 On March 17, no plumes nor high RDs wereobserved at any monitoring points/MPs of the FD1NPPand Fukushima Prefecture, because a strong westerly windprevailed all day (Fig. 2b). A plume with maximum 137Csconcentrations of 3,760–3,920 Bq m–3 was measured atFutaba for the three hours from 14:00 to 16:00 on March18 following the plume of P5 in the evening at Haramachi(Figs. 2a and 7). This three-hours continuation of high137Cs concentration at Futaba was caused by a southeast-erly wind constantly prevailing from 12:20 to 16:20 atthe MG of the FD1NPP, and during 11:30–16:20 at theFD2NPP (Figs. 2b and 7). Then, the wind direction shiftedto the south along Hamadori, such that the maximum 137Cs


m–3, respectively (Figs. 7 and 8a). During 6:20–7:20 and9:00–10:20 an easterly or southeasterly wind blew for ashort period at the northern point of the AB of theFD1NPP, and the RDs at Yamada increased (Fig. 8a), in-dicating that the radionuclides were transported to Futaba

from the FD1NPP. Then, the wind direction shifted clock-wise, and a southeasterly wind blew from 12:20 to 14:30at the FD1NPP (Fig. 8b), resulting in the transport of airmasses with high 137Cs concentrations to Futaba at 14:00.The RDs at the AB of the FD1NPP increased by approxi-

Fig. 10. Time series of the atmospheric 137Cs concentrations (red solid lines) and the ratios of 134Cs/137Cs (�) at (a) Futaba and(b) Naraha, during March 12–25, 2011. Bars mean the measurement errors due to gamma-ray counting statistics for each SPMsamples. The numerical values of 0.94, 1.05, and 1.08 are the inventory values of the ratios of 134Cs/137Cs at Unit 1, Unit 3, andUnit 2, respectively (Nishihara et al., 2012).


mately 800 mSv h–1 during 14:00–16:00 (Fig. 7), whichmight be a direct effect of the plume from the FD1NPP.Three hours later, the plume of P8 with the maximum137Cs of 200 Bq m–3 was observed at 17:00–19:00 atHaramachi (Tsuruta et al., 2014). Hence, the third plumemeasured at Futaba at 14:00 was considered to be trans-ported to Haramachi as plume P8 (Fig. 2a). At Namie andMinamisoma located north of Futaba in northernHamadori (Fig. 1a), the RDs also increased a few hourslater than the peak time at Futaba (Fig. 7). In contrast, apume with a 137Cs concentration of 272 Bq m–3 was ob-served at 7:00 at Naraha (P8¢ in Fig. 2c). Furthermore,the RDs at Yamadaoka and Matsudate also increased byapproximately 2 and 10 mSv h–1 at 7:00 and 8:00, respec-tively. At the FD2NPP, a northeasterly or north-northeast-erly wind blew during 5:50–7:40 when the RDs also in-creased by approximately 10 mSv h–1. Hence, the plumeof P8¢ at Naraha was directly transported from theFD1NPP by a north-northeasterly wind (Fig. 2b). Afterthat, the plume was transported to Nakadori by an east-erly wind, but not to the Kantou area.

[P9 and P9¢] at Futaba and Naraha on the morning ofMarch 21 A northerly or north-northwesterly wind startedto blow approximately at 21:00 on March 20 at theFD2NPP, and the 137Cs concentration at Naraha startedto increase from 22:00, showing a maximum of 100 Bqm–3 at 1:00 on March 21 (P9 in Fig. 2c), although no in-crease in the RDs was found at the monitoring points ofthe FD1NPP (Fig. 7) or at the MPs of FD2NPP. As theRDs at Yamadaoka and Futatsunuma a little bit increased

to 3.1 and 4.0 mSv h–1 at 1:00 on March 21, compared to2.8 and 3.7 mSv h–1 at 23:00 on March 20, respectively(Fig. 2c), the center of this plume would pass far the eastof these MPs. In addition, the plume was transported asP9 to the Kantou area on the morning of March 21 by anortheasterly wind (Fig. 9) (Tsuruta et al., 2014; Nakajimaet al., 2016). Another plume with the 137Cs concentrationof 106 mSv h–1 was also observed at 9:00 at Naraha, andthe RDs at Yamadaoka and Futatsunuma also increasedto 12 mSv h–1. This plume named as P9¢, however, wasnot transported to the Kantou area, because the wind di-rection shifted clockwise to the northeast in southernHamadori (Fig. 9). As shown in Fig. 2c, the increasingrate of RDs at Yamadaoka between 1:00 (P9) and 9:00(P9¢) was approximately 4 times, although the 137Cs con-centrations at Naraha at 1:00 and 9:00 on March 21 werealmost unchanged. A possible reason why the rates of in-crease were so different between Naraha and Yamadaokais that the plume (P9¢) had high concentrations ofradionuclides other than radiocesium, which were notincluded in the plume of P9. Another possibility is thatthe increase at 9:00 might be caused by ground shinethrough precipitation, but this possibility does not seemto be reasonable, because the RDs after the peak atYamadaoka rapidly decreased to the levels (3.6 and 3.4mSv h–1 at 10:00 and 11:00) a little bit higher than thosebefore precipitation.

[P10] at Futaba on the morning of March 24 The 137Csconcentration at Futaba increased to 116 Bq m–3 at 11:00,when a southeasterly wind blew at the FD1NPP, althoughthe RDs at Yamada and the MG of the FD1NPP did notincrease on the morning of March 24 (Figs. 2 and 7). Thisplume is newly found as the plume of 10 in this study. AtNamie and Minamisoma (Fig. 1a), the RDs increased atapproximately this time (Fig. 7), indicating the transportof plume P10 to northern Hamadori.

[P11] at Futaba on the afternoon of March 25 On March25, the 137Cs concentration also increased at Futaba be-tween 13:00–17:00 with a maximum of 280 Bq m–3 at14:00 (Figs. 2 and 7) when a southeasterly wind contin-ued to blow at the FD1NPP and FD2NPP. This plume isalso newly found as the plume of 11 in this study. At theMG of the FD1NPP, the RDs increased to 540 mSv h–1 at10:30 (Fig. 7), and which might affect the high 137Cs con-centrations at Futaba.

Activity ratio of 134Cs/137Cs134Cs/137Cs at Futaba and Naraha The atmospheric con-centration ratios of 134Cs/137Cs (R(Cs)) in the plumes,which were decay-corrected to the accident date, werecalculated at Futaba and Naraha, and their time seriesvariations (Fig. 10) and the daily means (Table 2) were

Table 2. Daily mean ratios of 134Cs/137Cs at Futaba andNaraha during March 12–25, 2011. The ratio was only usedfor the data with the 137Cs concentrations higher than 30Bq m–3.

N: Number of data for the ratio per day.S.D.: standard deviation.


analyzed because R(Cs) is an indicator of the reactor unitfrom which the radionuclides were released into the at-mosphere. The ratios measured from the filter-tapes wereonly used for the data with 137Cs concentrations higherthan 30 Bq m–3 because R(Cs) became more diverse whenthe 137Cs concentrations were less than 30 Bq m–3 (Sup-plementary Fig. S5). At Futaba and Natraha, the R(Cs)swere roughly divided into two groups; A and B. For the Agroup corresponding to the plumes of p1, P1v, and P1¢ atFutaba during March 12–13, 2011, the R(Cs)s were inthe range of 0.92–0.93, very close to 0.94, which is theR(Cs) of Unit 1 based on the core inventory data esti-mated by Nishihara et al. (2012). In particular, the R(Cs)sin plume P1v released from Unit 1 were 0.92 and 0.93, at15:00 and 16:00 on March 12, respectively.

In contrast, the R(Cs)s for the B group at Futaba andNaraha during March 15–21, 2011, were in the range of1.01–1.05, which is significantly higher than the R(Cs)range for group A. At this stage, we have not been able todetermine the possible source units for the plumes ingroup B, as discussed below, whereas Nishihara et al.(2012) estimated that the R(Cs)s are 1.08 and 1.05, forUnit 2 and Unit 3, respectively. The R(Cs)s in plume P2at Naraha on the morning of March 15, 2011 were in therange of 1.01–1.04, closer to 1.05 (Unit 3) than 1.08 (Unit2), although the source of the FD1NPP on the morning ofMarch 15 was estimated to be Unit 2 (TEPCO, 2013;Katata et al., 2015). Furthermore, the R(Cs)s in the plumeof P4 and P4¢ measured at Naraha on the morning ofMarch 16 were in the range of 0.98–1.05, with an aver-age of 1.02, and were also close to 1.05 (Unit 2), althoughthe source term was estimated to be Unit 2 and/or Unit 3on the morning of March 16 (TEPCO, 2013; Katata etal., 2015). At Futaba, a daily average R(Cs) of 1.03–1.05was found for four plumes (P3, P5, P6, and P8) duringMarch 15–20, 2011, and was also closer to 1.05 (Unit 3)than 1.08 (Unit 2). Because there are no large differencesin the R(Cs) values among these 7 plumes during March15–20, 2011, we could not identify whether theradionuclides were released from Unit 2 or Unit 3. It isnoted that the average R(Cs) of the first two datasets fromplume P9¢ at Naraha on March 21 and of plume 11 atFutaba on March 25 was equal to 1.07, which is closer tothe R(Cs) of Unit 2 based on the inventory data. It is alsonoted that the average R(Cs) in plume P10 at Futaba onMarch 24 was 0.96, which is closer to the R(Cs) of Unit1 in the inventory data. Nevertheless, these data are thefirst direct attempts to show that R(Cs) is determined byatmospheric aerosols collected within a few hours aftertheir release into the atmosphere from the FD1NPP. Ourmeasurement of the radionuclides was made for only ap-proximately 1 hour via Ge semiconductor detectors, andthe accuracy of the R(Cs) values, especially for theradiocesium concentrations less than 100 Bq m–3, might

not be enough to discuss the small difference in R(Cs).134Cs/137Cs in other studies All the R(Cs)s previouslyreported were analyzed by using samples deposited onthe earth’s surface (Komori et al., 2013; Chino et al.,2016; Satou et al., 2016). Komori et al. (2013) measuredthe radionuclides of the soil and vegetation samples col-lected in southern Tohoku (Fukushima, Miyagi, and IwatePrefectures) and the Kantou area. The R(Cs) values ofthose samples were divided to three groups, i.e., 0.91–0.98 in Oshika peninsula, located in the eastern Miyagiprefecture; 0.98–1.01 near the FD1NPP; and 1.00–1.04among the Kantou area, the Iwate Prefecture, and north-western Fukushima. This work concluded that the sourceunit of the first group was Unit 1, because the RDs showedhigh values at the MPs of the Onagawa Nuclear PowerStation near the sampling sites on the early morning ofMarch 13 (Tohoku Electric Power Company, 2013) dueto the transport of radioactive materials released from theFD1NPP. They also concluded that the possible sourceunit of the third group was estimated to be Units 2 and/or3, and that the R(Cs) of the second group was a result ofthe mixture of all three Units. These results are consist-ent with our results.

Chino et al. (2016) analyzed R(Cs) using all the dataof in situ ground surface measurements from gamma-rayspectrometers in the southern Tohoku and the Kantou ar-eas (Mikami et al., 2015). They found that the R(Cs)sfrom the frequency distribution had the two peaks of 1.09and 1.04, which are close to the inventory data of Unit 2and Unit 3, respectively. They also found another groupof R(Cs) that had a peak value of 0.95, corresponding tothe inventory data of Unit 1. Nishizawa et al. (2016)measured the detailed distribution of the radiocesiumdeposition within a 3-km area of the FD1NPP using anautonomous unmanned helicopter monitoring system andfound four plume traces by analyzing the distribution mapof high air dose rates. In addition, they estimated that thenorth plume traces with R(Cs) of 0.95 and the other threeplume traces with the R(Cs) values of 1.05–1.06 werederived from Unit 1 and from Unit 2 or Unit 3, respec-tively. The results in these two studies indicate that theR(Cs) of Unit 1 was close to that of the data from ourstudy and from Komori et al. (2013), and to the inven-tory data. However, the R(Cs) values of Unit 2 and/orUnit 3 differed from the data in our study and from thatof Komori et al. (2013). At this stage, we have not clari-fied the cause of the discrepancies in the R(Cs) valuesfor Unit 2 and Unit 3 among these four studies.

The source term of the plumes at Futaba on March 12–13As previously described, the plumes measured at Futabaon March 12–13 were identified as released from Unit 1.TEPCO (2015a) estimated that the earlier leak ofradionuclides into the atmosphere started from Unit 1 at


3:10 on March 12 because the pressure in the drywell(DW) of the reactor containment vessel (RCV) in Unit 1remained high from 2:30 to 14:30–14:50 when a vent op-eration was finally made. As a result, the first release intothe atmosphere from the reactor building of Unit 1 wasobserved at the MG and MP-8 of the FD1NPP after 4:05(4 in Fig. 4). In addition, one of the largest release ratesof radionuclides (Katata et al., 2015) was caused by avent operation at Unit 1 after 14:00 (TEPCO, 2015a). Fur-thermore, this result was confirmed by a photograph takenfrom an airplane at 14:46 on March 12 (Tokyo Shinbun,2016), which clearly showed that a plume of white smokeflowed horizontally just after the release from the exhauststack of Unit 1. The RDs at the FD1NPP did not increaseat the MG and MP8 on the afternoon until 19:00 (Fig. 4)possibly because plume P1v did not pass over these points.However, the RDs at the MP4 recorded a high value of162 mSv h–1 at 15:13 when the RDs measurements startedand before the hydrogen explosion (16 in Fig. 4), com-pared to less than 10 mSv h–1 at the MG and the MP8.This high value was possibly affected by plume P1v be-cause the MP4 point was located downwind from Unit 1due to the southeasterly wind (Fig. S2). It is noted thatthe surface wind direction at Koriyama was from thesoutheast, with a wind speed of 3 m s–1 (Fig. 3b). AtKoriyama, the RDs showed a maximum of 1,500 mSvh–1 at 15:43 (Fig. 3), suggesting the direct effect of theplume by the vent operation or the hydrogen explosion inUnit 1 at 15:36 on Mach 12.

For plume P1¢ at 9:00 and a maximum concentrationof 50 Bq m–3 at 18:00 at Futaba on March 13, a possiblesource might be Unit 3 because TEPCO (2015b) reportedthat water vapor was released from the exhaust stack ofUnit 3 at 10:00 and 13:00 on March 13 by the first andsecond vent operations of Unit 3, respectively. As previ-ously described, the average R(Cs) of 0.92 on March 13leads to the conclusion that the source unit of the plumeson March 13 was Unit 1, as was the case on March 12.This suggests that an amount of radionuclides releasedfrom the two vent operations of Unit 3 would not be largecompared to that from Unit 1.

CONCLUSION

Time-series analysis of the hourly 137Cs concentrationsat the Futaba and Naraha SMP stations near the FD1NPPrevealed that many plumes were found during March 12–25, 2011, just after the Fukushima accident. Nine plumeswere observed at Futaba located 3.2 km northwest of theFD1NPP during March 12–13, 15–16, 18–20, and 24–25,2011, i.e., every time a southeasterly/southerly wind pre-vailed over Hamadori, Fukushima Prefecture, indicatingthe transport of radionuclides released from the FD1NPP.The first plume at Futaba was observed at 9:00 on the

morning of March 12, due to the direct release from theFD1NPP in the early morning. The highest 137Cs concen-tration was measured at 15:00 on March 12 after a ventoperation just before the hydrogen explosion of Unit 1 at15:36. This was confirmed by a photograph taken froman airplane, showing a white plume streaming downwindfrom the exhaust stack of Unit 1. The second highest 137Csconcentration, approximately 43% of the highest concen-tration, was measured on the afternoon of March 20, andwas still high even 8 days after the first release by theaccident. In contrast, at Naraha located 17.5 km south ofthe FD1NPP, seven plumes were observed on March 15–16, 20–21, 2011, i.e., every time when a northerly windprevailed, and at least half of these plumes were trans-ported southward to the Kantou area including the TokyoMetropolitan area by a northeasterly wind. The highest137Cs concentration of 8,300 Bq m–3 at Naraha was meas-ured at 3:00 on March 15, and this plume was possiblycaused by the first release from Unit 2. According to theratios of 134Cs/137Cs in the atmospheric concentrations,the plumes measured at Futaba on March 12–13 wereidentified from Unit 1 because the ratios were in the rangeof 0.92–0.94, close to the inventory data. In contrast, theratios at Futaba and Naraha during March 15–21, 2011,were in the range of 1.02–1.04, but the source unit ofthese plumes has not been determined to be Unit 2 and/orUnit 3 at this stage. These analyses strongly suggest thatthe hourly radiocesium data measured at Futaba andNaraha will be a useful tool to speculate what occurredin the reactor units of the FD1NPP and to reevaluate therelease rates from the FD1NPP, in addition to validatingand improving the simulation results by atmospherictransport and deposition models.

Acknowledgments—We thank to the staff of the FukushimaPrefecture for providing the used filter-tapes at two SPM moni-toring stations and for technical advice how to handle the data-base measured at monitoring posts. We are also thankful to S.Mizokami in TEPCO Fukushima Daiichi DecommissioningPromotion Company for productive comments. We acknowl-edge N. Matsumoto and H. Yamada in Earth Observation Re-search Center, Japan Aerospace Exploration Agency, and Y.Yoshida and T. Arai in Atmosphere and Ocean Research Insti-tute, the University of Tokyo, for mapping and handling a largeamount of data. This work was partly supported by the “Inter-disciplinary Study on Environmental Transfer of Radionuclidesfrom the Fukushima Daiichi NPP Accident” project by JSPSKAKENHI (grant numbers 24110002, 24110008, and24110009). It was also partly supported by the “Interdiscipli-nary study for exposure assessment and risk assessment of air-borne radionuclides released by the nuclear power plant acci-dent” project (5-1501), and by the “Evaluation of SLCP Envi-ronmental Impacts and Promotion of Climate Change Counter-measures through Seeking the Optimal Pathway” project (S-12), in the Environmental Research and Technology Develop-ment Fund by the Ministry of the Environment, Japan.


REFERENCES

Chino, M., Nakayama, H., Nagai. H., Terada, H., Katata, G.and Yamazawa, H. (2011) Preliminary estimation of releaseamounts of 131I and 137Cs accidentally discharged from theFukushima Daiichi nuclear power plant into the atmosphere.J. Nucl. Sci. Technol. 48, 1129–1134.

Chino, M., Terada, H., Nagai, H., Katata, G., Mikami, S., Torii,T., Saito, K. and Nishizawa, Y. (2016) Utilization of 134Cs/137Cs in the environment to identify the reactor units thatcaused atmospheric releases during the Fukushima Daiichiaccident. Sci. Rep. 6, 31376; doi:10.1038/srep31376.

Fukushima Prefecture (2011) Results of environmental radio-activity measurements in seven districts of Fukushima Pre-fecture during March 11–31, 2011. http:/ /www.pref.fukushima.lg.jp/sec/16025d/h22-24-7houbu.html(last access: May 2016).

Fukushima Prefecture (2012) Results of radiation dose rates inMarch 2011. Data collected from monitoring posts inFukushima Prefecture by the Environmental RadioactivityMonitoring Center of Fukushima, Sep. 21, 2012. http://www.pref.fukushima.lg.jp/sec_file/monitoring/etc/post3-20120921.pdf (original dataset is available: http://www.atom-moc.pref.fukushima.jp/old/monitoring/monitor-ing201103/201103_mpdata.html) (last access: July 2017).

Fukushima Prefecture (2016) Results of environmental radio-activity measurements within 20–50 km from the FukushimaDaiichi Nuclear power plant during March 12–31, 2011.https:/ /www.pref.fukushima.lg.jp/sec/16025d/kako-monitoring.html (last access: May 2016).

Hirayama, H., Matsumura, H., Namito, Y. and Sanami T. (2017)Estimation of Xe-135, I-131, I-132, I-133, Te-132 concen-trations in plumes at monitoring posts in Fukushima Pre-fecture using pulse height distribution obtained fromNaI(T1) Detector. Trans. At. Energy. Soc. Jpn. 16, 1–14 (inJapanese).

International Atomic Energy Agency (2000) Generic proceduresfor assessment and response during a radiological emer-gency. IAEA-TECDOC-1162.

Journal of Nuclear and Radiochemical Sciences (2015) Appen-dix A, 15 http:/ /www.radiochem.org/paper/JN152/jn15201_Appendix_A_rev.pdf

Katata, G., Chino, M., Kobayashi, T., Terada, H., Ota, M., Nagai,H., Kajino, M., Draxler, R., Hort, M. C., Malo, A., Torii, T.and Sanada, Y. (2015) Detailed source term estimation ofthe atmospheric release for the Fukushima Daiichi NuclearPower Station accident by coupling simulations of an at-mospheric dispersion model with an improved depositionscheme and oceanic dispersion model. Atmos. Chem. Phys.15, 1029–1070.

Komori, M., Shozugawa, K., Nogawa, N. and Matsuo, M.(2013) Evaluation of radioactive contamination caused byeach plant of Fukushima Daiichi Nuclear Power Stationusing 134Cs/137Cs activity ratio as an index. Bunseki Kagaku62, 475–483 (in Japanese).

Mikami, S., Maeyama, T., Hoshide, Y., Sakamoto, R., Sato, S.,Okuda, N., Demongeot, S., Gurriaran, R., Uwamino, Y.,Kato, H., Fujiwara, F., Sato, T., Takemiya, H. and Saito, K.(2015) Spatial distributions of radionuclides deposited onto

ground soil around Fukushima Dai-ichi Nuclear Power Plantand their temporal change until December 2012. J. Environ.Radioact. 139, 320–343.

Ministry of Economy, Trade and Industry (2011) Results ofurgent monitoring of radioactive materials aroundFukushima Daiich and Daini Nuclear Power Stations (inJapanese). http://warp.da.ndl.go.jp/info:ndljp/pid/8422823/w w w. m e t i . g o . j p / p r e s s / 2 0 1 1 / 0 6 / 2 0 1 1 0 6 0 3 0 1 9 /20110603019.html (last access: July 2017).

Ministry of Education, Culture, Sports, Science and Technol-ogy (2011a) Preparation of distribution map of radiationdoses, etc. (Map of radioactive cesium concentration in soil)by MEXT. http://radioactivity.nsr.go.jp/en/contents/5000/4165/24/1750_083014.pdf (last access: December 2014).

Ministry of Education, Culture, Sports, Science and Technol-ogy (2011b) Results of the fourth airborne monitoring sur-vey by MEXT (2011). http://radioactivity.nsr.go.jp/en/con-tents/4000/3179/24/1270_1216.pdf (last access: March2014).

Ministry of Education, Culture, Sports, Science and Technol-ogy (2012) Readings of dust sampling by MEXT, as of Sep.12, 2012 (in Japanese). http://radioactivity.nsr.go.jp/ja/con-tents/8000/7572/24/dust%20sampling_All%20Results%20for%20May%202011.pdf (last access: May 2015).

Nagakawa, Y., Sotodate, T., Kinjo, Y. and Suzuki, T. (2014)One-year time variations of anthropogenic radionuclides inaerosols in Tokyo after the Fukushima Dai-ichi NuclearPower Plant reactor failures. J. Nucl. Sci. Technol.,doi:10.1080/00223131.2014.985279.

Nakajima, T., Misawa, S., Morino, Y., Tsuruta, H., Goto, D.,Uchida, J., Takemura, T., Ohara, T., Oura, Y., Ebihara, M.and Satoh, M. (2016) Model depiction of the atmosphericflows of radioactive cesium emitted from the FukushimaDaiichi Nuclear Power Station accident. Prog. Earth Planet.Sci. 4, 1–18, doi:10.1186/s40645-017-0117-x.

Nishihara, K., Iwamoto, H. and Suyama, K. (2012) Estimationof fuel compositions in Fukushima-Daiichi nuclear powerplant. JAEA-Data/Code 2012-018, Japan Atomic EnergyAgency.

Nishizawa, Y., Yoshida, M., Sanada, Y. and Torii, T. (2016)Distribution of the 134Cs/137Cs ratio around the FukushimaDaiichi nuclear power plant using an unmanned helicopterradiation monitoring system. J. Nucl. Sci. Technol. 53, 468–474.

Ohkura, T., Oishi, T., Taki, M., Shibanuma, Y., Kikuchi, M.,Akino, H., Kikuta, Y., Kawasaki, M., Saegusa, J., Tsutsumi,M., Ogose, H., Tamura, S. and Sawahata, T. (2012) Emer-gency monitoring of environmental radiation and atmos-pheric radionuclides at Nuclear Science Research Institute,JAEA following the accident of Fukushima Daiichi NuclearPower Plant, JAEA-Data/Code 2012-010.

Oura, Y., Ebihara, M., Tsuruta, H., Nakajima, T., Ohara, T.,Ishimoto, M., Sawahata, H., Katsumura, Y. and Nitta, W.(2015) A Database of hourly atmospheric concentrations ofradiocesium (134Cs and 137Cs) in suspended particulate mat-ter collected in March 2011 at 99 air pollution monitoringstations in eastern Japan. J. Nuclear Radiochem. Sci. 15,1–12.


Sanada, N. and Torii, T. (2015) Aerial radiation monitoringaround the Fukushima Dai-ichi Nuclear Power Plant usingan unmanned helicopter. J. Environ. Radioact. 139, 294–299.

Satou, Y., Sueki, K., Sasa, K., Adachi, K. and Igarashi, Y. (2016)First successful isolation of radioactive particles from soilnear the Fukushima Daiichi Nuclear Power Plant.Anthropocene 14, 71–76.

TEPCO (2011) Press release (May 28, 2011): Additional moni-toring data at Fukushima Daiichi Nuclear Power station.ht tp: / /www.tepco.co. jp/en/press/corp-com/release/11052811-e.html (last access: July 2017).

TEPCO (2012) Radiation dose measured in the Fukushima DainiNuclear Power Station 2011 | Archives. http:/ /www.tepco.co.jp/en/nu/fukushima-np/f2/data/2011/index-e.html (last access: July 2017).

TEPCO (2013) Influence to surrounding environment | Ar-chives. Air, 2011, March. The results of nuclide analyses ofradioactive materials in the air at the site of FukushimaDaiichi Nuclear Power Station (from the first to the tenthreleases). http://www.tepco.co.jp/en/nu/fukushima-np/f1/index2-e.html (last access: July 2017).

TEPCO (2015a) Press release (Dec. 17, 2015): The 4th progressreport on the investigation and examination of unconfirmedand unresolved issues on the development mechanism ofthe Fukushima Daiichi Nuclear Accident [Main body (UP-DATE)] https://www4.tepco.co.jp/en/press/corp-com/re-lease/betu15_e/images/151217e0102.pdf (last access: July2017).

TEPCO (2015b) Press release (Dec. 17, 2015): The 4th progressreport on the investigation and examination of unconfirmedand unresolved issues on the development mechanism ofthe Fukushima Daiichi Nuclear Accident [Attachment 3-8].Leaks from Unit-3 PCV and stream release in a largeamount. http://www.tepco.co.jp/en/press/corp-com/release/betu15_e/images/151217e0137.pdf (last access: July 2017).

Tohoku Electric Power Company (2013) Emergency informa-tion (March 13, 2011). http://www.tohoku-epco.co.jp/emer-gency/8/1182599_1800.html (last access: March 2013).

Tokyo Shinbun (2016) White smoke by a vent operation beforethe hydrogen explosion at Unit 1 on March 12, 2011. Anewspaper on February 19, 2016 (in Japanese).

Torii, T., Sugita, T., Okada, C., Reed, M. and Blumenthal, D.(2013) Enhanced analysis methods to derive the spatial dis-tribution of 131I deposition on the ground by airborne sur-veys at an early stage after the Fukushima Daiichi NuclearPower Plant accident. Health Phys. 105, 192–200.

Tsuruta, H., Oura, Y., Ebihara, M., Ohara, T. and Nakajima, T.(2014) First retrieval of hourly atmospheric radionuclidesjust after the Fukushima accident by analyzing filter-tapesof operational air pollution monitoring stations. Sci. Rep.4, 6717, doi:10.1038/srep06717.

Tsuruta, H., Oura, Y., Ebihara, M., Moriguchi, Y., Ohara, T.and Nakajima, T. (2017) Spatio-temporal distribution ofatmospheric radiocesium in eastern Japan just after theTEPCO Fukushima Daiichi Nuclear Power Plant accident—Analysis of used filter-tapes of SPM monitors in air qualitymonitoring stations—. Earozoru Kenkyu 32, 244–255 (inJapanese).

UNSCEAR (2015) Developments since the 2013 UNSCEARrepot on the levels and effects of radiation exposure to thenuclear accident following the great east-Japan earthquakeand tsunami. A 2015 white paper to guide the ScientificCommittee’s future programme of work (last access: De-cember 2016).

UNSCEAR (2016) Developments since the 2013 UNSCEARreport on the levels and effects of radiation exposure to thenuclear accident following the great east-Japan earthquakeand tsunami. A 2016 white paper to guide the ScientificCommittee’s future programme of work (last access: Au-gust 2017).

SUPPLEMENTARY MATERIALS

URL (http://www.terrapub.co.jp/journals/GJ/archives/data/52/MS520.pdf)

Figures S1 to S5Tables S1 and S2

Documents

Time-series analysis of atmospheric radiocesium at two SPM ... · Time-series analysis of atmospheric radiocesium at two SPM monitoring sites ... 2Department of Chemistry, ... 2011