ww.sciencedirect.com

wat e r r e s e a r c h 6 0 ( 2 0 1 4 ) 1 5e2 7

Available online at w

ScienceDirect

journal homepage: www.elsevier .com/locate/watres

Distribution of dissolved and particulateradiocesium concentrations along rivers and therelations between radiocesium concentration anddeposition after the nuclear power plant accident inFukushima

Hideki Tsuji a, Tetsuo Yasutaka a,*, Yoshishige Kawabe a, Takeo Onishi b,Takeshi Komai c

aNational Institute of Advanced Industrial Science and Technology, 1-1-1 Higashi, Tsukuba, Ibaraki 305-8567, Japanb Faculty of Applied Biological Sciences, Gifu University, 1-1 Yanagito, Gifu 501-1193, JapancGraduate School of Environmental Studies, Tohoku University, 6-6-20 Aramaki-Aza-Aoba, Aoba-ku, Sendai,

Miyagi 980-8579, Japan

a r t i c l e i n f o

Article history:

Received 15 November 2013

Received in revised form

3 April 2014

Accepted 13 April 2014

Available online 24 April 2014

Keywords:

Dissolved radiocesium

Particulate radiocesium

Deposition

River water

Fukushima

Regression analysis

* Corresponding author. Tel.: þ81 29 849 154E-mail address: t.yasutaka@aist.go.jp (T.

http://dx.doi.org/10.1016/j.watres.2014.04.0240043-1354/ª 2014 Elsevier Ltd. All rights rese

a b s t r a c t

This study involved measurement of concentrations of dissolved and particulate radio-

cesium (134Cs and 137Cs) in river water, and determination of the quantitative relations

between the amount of deposited 137Cs and 137Cs concentrations in river waters after the

Fukushima Daiichi nuclear power plant accident. First, the current concentrations of dis-

solved and particulate 134Cs$137Cs were determined in a river watershed from 20 sampling

locations in four contaminated rivers (Abukuma, Kuchibuto, Shakado, and Ota).

Distribution characteristics of different 137Cs forms varied with rivers. Moreover, a

higher dissolved 137Cs concentration was observed at the sampling location where the137Cs deposition occurred much more heavily. In contrast, particulate 137Cs concentration

along the river was quite irregular, because fluctuations in suspended solids concentra-

tions occur easily from disturbance and heavy precipitation. A similar tendency with

dissolved 137Cs distribution was observed for the 137Cs concentration per unit weight of

suspended solids.

Regression analysis between deposited 137Cs and dissolved/particulate 137Cs concen-

trations was performed for the four rivers. The results showed a strong correlation be-

tween deposited 137Cs and dissolved 137Cs, and a relatively weak correlation between

deposited 137Cs and particulate 137Cs concentration for each river. However, if the partic-

ulate 137Cs concentration was converted to 137Cs concentration per unit weight of sus-

pended solid, the values showed a strong correlation with deposited 137Cs.

ª 2014 Elsevier Ltd. All rights reserved.

5; fax: þ81 29 861 8109.Yasutaka).

rved.

wat e r r e s e a r c h 6 0 ( 2 0 1 4 ) 1 5e2 716

1. Introduction

A significant amount of radiocesium (mainly 134Cs and 137Cs)

was released into the atmosphere and deposited on land from

the accident at the Fukushima Daiichi Nuclear Power Plant

(FDNPP). In the years following the accident, the radiocesium

stored mainly in forests and mountains gradually flowed into

mountain streams and rivers through rainfall events (Nagao

et al., 2013; Ueda et al., 2013). Since rivers can transfer

chemicals a great distance, continual monitoring of river

water is important for estimating downstream environmental

impacts.

Cesium concentrations in river waters can be divided into

two main forms: dissolved and particulate. Dissolved radio-

cesium exists as cesium ions and cesium ion hydrates, which

migrate relatively quickly with water flow and can contami-

nate plants via root uptake (Zhu and Smolders, 2000). In cur-

rent agricultural sites, some dissolved radiocesium can

migrate by irrigation into paddy fields or into water supplies

Fig. 1 e Sampling location and 137Cs deposition distribution aft

Kuchibuto River, Shakado River, and Ota River watersheds.

for hydroponics. In contrast, cesium ions from particulate

radiocesium are absorbed mainly onto soil particles (Evans

et al., 1983; Comans et al., 1991), which exist as suspended

solids (SS) in the water and may cause secondary contami-

nation due to sediment runoff (Walling and Bradley, 1990).

The difference in these environmental behaviors requires that

both forms of radiocesium must be monitored.

Several reports have described the distribution of dissolved

or particulate radioactive materials in environmental waters

after the Chernobyl accident using either field surveys

(Kryshev, 1995; Vakulovsky et al., 1994; Matsunaga et al., 1998)

or model simulations (Smith et al., 2004; McDougall et al.,

1991). These studies investigated actual conditions for the

temporal distribution of radiocesium, decay of radiocesium

with time, and behavior of soil containing radiocesium near

the Chernobyl Nuclear Power Plant. However, the climate and

topography around Fukushima are different from Chernobyl.

For example, annual precipitation in Fukushima is 3-fold

greater than that of Chernobyl, and the radiocesium was

er the Fukushima NPP accident in the Abukuma River,

Table 1 e Sampling locations and characteristics.

River Averagedflow rate

Samplingdate

Samplinglocation

Latitude Longitude Distance fromthe end of river

(km)

Watershedarea (km2)

Averaged 137Csdeposition(kBq/m2)

Sampledwater (L)

Abukuma

River

6.3 m3/s at B

121 m3/s at G

(2011 JaneDec)a

2013/1/28 A Kawatani 37.1478 140.1361 204.0 59.0 61.8 60

B Shirakawa 37.1167 140.2447 189.2 169.1 81.9 40

2013/1/27 C Sukagawa 37.2872 140.4133 151.4 884.4 60.4 40

D Akutsu 37.4080 140.4087 133.8 1830.3 80.1 40

2013/1/29 E Nihonmatsu 37.5940 140.4613 106.8 2391.1 91.3 40

F Fukushima-city 37.7588 140.4874 75.2 3245.2 115.7 40

G Yahata 37.8928 140.6436 51.8 4087.4 120.7 40

Kuchibuto

River

2.8 m3/s at the

vicinity of M (2011

JulyeAug)b

2013/1/27 H Nagayama 37.6058 140.6506 38.1 4.1 303.5 40

I Mizusakai 37.5847 140.7144 35.6 0.8 854.0 40

J Tazawa 37.5572 140.6531 25.2 55.0 341.5 40

K Doumeki 37.5306 140.6108 18.8 66.0 309.7 40

L Fukawada 37.5522 140.5825 14.0 82.6 278.2 60

M Nishitani 37.5810 140.5419 4.2 136.0 265.3 60

Shakado

River

7.9 m3/s at the

vicinity of P (2012

JaneDec)

2013/1/27 N Nishigou 37.2781 140.1667 24.6 13.5 69.3 60

O Shimomatsumoto 37.2525 140.2422 19.9 41.6 114.4 40

P Horisoko 37.2886 140.3547 3.2 284.9 128.3 40

Ota River 0.15 m3/s at the

vicinity of S (2013

JaneDec)

2013/5/7 Q 37.5969 140.9165 11.6 44.9 1714.4 20

R 37.5975 140.9249 10.5 47.0 1652.6 20

S 37.6046 140.9633 6.8 52.0 1514.4 20

T 37.6031 140.9904 3.7 55.6 1422.9 20

a Water Information System by the MLIT.b Japan Atomic Energy Agency (2013).

water

research

60

(2014)15e27

17

wat e r r e s e a r c h 6 0 ( 2 0 1 4 ) 1 5e2 718

deposited onto steeper mountains in Fukushima, which could

speed up radiocesium runoff.

Sakaguchi et al. (2012) reported that dissolved radiocesium

concentrations in river water, 30e60 km from the FDNPP and

3e4 months after the accident were 0.05e0.52 Bq/L as 134Cs,

and 0.02e0.46 Bq/L as 137Cs. The Ministry of Education, Cul-

ture, Sports, Science and Technology (MEXT, 2011b) of Japan

reported that the average 134Cs and 137Cs concentrations of 51

rivers in Fukushima Prefecture one year after the FDNPP ac-

cident, were 0.54 Bq/L and 0.58 Bq/L, respectively. Ueda et al.

(2013) monitored dissolved and particulate radiocesium con-

centrations in two rivers in Fukushima Prefecture during a

precipitation event, and Yasutaka et al. (2013a) monitored

mountain stream and irrigation water in Fukushima Prefec-

ture. The results indicated that most of the measured radio-

cesium concentrations in river waters were several orders of

magnitude less than the standard for drinkable water (10 Bq/

L), as established by the World Health Organization (WHO).

However, monitoring low-level concentrations of radiocesium

in rivers is important for evaluating long-term environmental

behavior, transport into crops, and re-contamination of paddy

fields by irrigation. The spatial distribution of low levels of

radiocesium along a single river has not been thoroughly

examined.

In another survey of Fukushima rivers, total 137Cs con-

centration in water and of 137Cs deposited in the watershed

were examined using regression analysis (MEXT and Ministry

of Agriculture, Forestry and Fisheries (MAFF), 2012). In this

research, river water was sampled two times within 3e5

months after the FDNPP accident, and had R2 values of 0.62

and 0.68, indicating that concentrations in the river can be

predicted from the amount of 137Cs deposition and regression

analysis. However, the correlation between the amount of

radiocesium deposition and dissolved and particulate radio-

cesium concentrations has not been determined separately.

In the present study, dissolved and particulate 134Cs and137Cs concentrations in river water were measured separately

at 20 sampling locations on four rivers in Fukushima Prefec-

ture: Abukuma River, Kuchibuto River, Shakado River, and Ota

River. Characteristics of 137Cs concentrations along the river

courses also were determined. The relations between 137Cs

deposited by the FDNPP accident and observed dissolved/

particulate 137Cs concentrations in river waters were exam-

ined using regression analysis.

2. Materials and methods

2.1. Sampling locations and watershed characteristics

Fig. 1 shows the 20 sampling locations in Fukushima Prefec-

ture. River water was sampled at 7 locations (AeG) on the

Abukuma River (239 km long), 6 locations (HeM) on the

Kuchibuto River (35 km long), and 3 locations (NeP) on the

Shakado River (35 km long). The Kuchibuto River and the

Shakado River are two branches of the Abukuma River.

Additionally, river water was sampled at 4 locations on the

Ota River (23 km long), which flows throughMinamisoma City

to the Pacific Ocean (QeT). The Ota River partially originates in

Namie Town, where a large amount of radiocesium was

deposited and is currently designated a heavily contaminated

area.

Table 1 shows latitude, longitude, and distances from the

river mouths of the Abukuma and Ota Rivers, and the

confluence with the Abukuma River for the Kuchibuto and

Shakado Rivers to the sampling locations. The Abukuma River

flows southwest to northeast, the Shakado andOta Rivers flow

from west to east, and the Kuchibuto River from east to west.

The annual average flow rates were 6.3 m3/s (location B) to

121 m3/s (location G) for the Abukuma River in 2011, 7.9 m3/s

for the Shakado River (vicinity of location P) in 2012, and

0.15 m3/s for the Ota River (vicinity of location S) in 2013. For

the Kuchibuto River, 2.8 m3/s was observed on average at the

vicinity of location M during July 1eAugust 31.

The region shown in Fig. 1 presents thewatersheds, as well

as the sampling locations. The boundaries of each watershed

were determined from surface height data in the Global Dig-

ital Elevation Model GDEM, which has a spatial resolution of

30 m, and flow direction toward lower surfaces was deter-

mined for each grid. The area of each watershed is included in

Table 1. The watershed area for each sampling location was

50e4000 km2 for the Abukuma River, 10e300 km2 for the

Shakado River, 0.8e140 km2 for the Kuchibuto River, and

40e60 km2 for the Ota River.

2.2. The gamma spectroscopy methodology

134Cs and 137Cs concentrations were measured by a Ge semi-

conductor detector, SEG-EMS GEM 35-70 (detection efficiency:

22.7%; resolution: 1.76 keV Seiko EG&GCo., Ltd., Tokyo, Japan),

and a multichannel analyzer MCA 7600 (hereafter called MCA,

Seiko EG&G Co., Ltd.). MCA was set up to 4000 channels (each

channel has 0.5 keV) from about 0 to 2000 keV. Gammastudio

(Seiko EG&G Co., Ltd.) was used as analysis software. Effi-

ciency calibration was set by nine nuclide mixed activity

standard volume sources of a 2-L Marinelli container

(MX033MR), and nine nuclide mixed activity standard volume

sources of a U-8 container (MX033U8PP) (Japan Radioisotope

Association) during the installation of the semiconductor

detector.

The accuracy of the analysis equipment was calibrated by

background measurement for 24 h once a week, and the en-

ergy of the gamma raywas calibrated once aweek. The energy

peaks of the 60Co (1332.5 keV) and 137Cs (662 keV) were

adjusted to the prescribed channel (60Co:2665 ch, 137Cs:1324

ch) by nine nuclide mixed activity standard volume sources in

a 2-L Marinelli container. The semiconductor detector was

calibrated by standard volume sources once a week, and

decay compensation was applied. The minimum detection

count of each sample was set to threefold the ordinary

counting error, which is calculated from Cooper’s method

(Cooper, 1970), and the counts below the minimum detection

value were treated as “non-detected”.

2.3. Sampling and analysis of river water

River water was sampled during three periods. On 14 and 15

September 2012, river water was sampled at 7 locations on the

Abukuma River (AeG), and the concentrations were reported

(Yasutaka et al., 2012). A precipitation event was observed at

Table 2 e Air dose rate and water quality of sampled river water.

River (samplingmonth)

Samplinglocation

Air doserate (mSv/h)a

Watertemperature

(�C)

pH Electricalconductivity(mS/cm)b

Concentration ofsuspended solid

(mg/L)

Abukuma River

(Sep. 2012)cA Kawatani 0.3 24.7 8.0 214 0.3

B Shirakawa 0.6 26.4 8.0 197 3.4

C Sukagawa 0.35 28.6 8.4 178 3.2

D Akutsu 1 27.3 8.1 221 6.1

E Nihonmatsu 1.8 28.7 7.9 239 6.1

F Fukushima-city 2 29.3 8.6 240 7.7

G Yahata 0.6 26.0 7.6 240 6.0

Abukuma River

(Jan. 2013)

A Kawatani 0.24 3.0 8.1 160 1.4

B Shirakawa 0.09 4.1 7.8 130 1.9

C Sukagawa 0.21 4.6 8.7 200 3.9

D Akutsu 0.24 2.5 8.0 230 4.7

E Nihonmatsu 0.32 3.1 7.9 250 3.3

F Fukushima-city 0.32 5.2 8.0 320 5.4

G Yahata 0.32 2.5 7.8 260 4.4

Kuchibuto River

(Jan. 2013)

H Nagayama 0.76 2.9 7.9 110 <0.5

I Mizusakai 2.30 2.2 8.2 60 <0.5

J Tazawa 0.31 2.3 7.9 110 0.8

K Doumeki 0.30 2.6 8.0 170 1.1

L Fukawada 0.40 2.6 8.1 180 0.7

M Nishitani 0.43 2.6 8.2 170 2.4

Shakado River

(Jan. 2013)

N Nishigou 0.40 2.7 7.9 110 1.3

O Shimomatsumoto 0.80 4.9 8.1 120 1.3

P Horisoko 0.24 4.7 8.7 180 6.0

Ota River

(May 2013)

Q 0.80 11.9 8.3 112 2.7

R 0.44 15.2 8.3 75 0.7

S 0.54 10.3 8.2 91 2.0

T 0.19 10.5 7.7 99 1.5

a Microsieverts per hour.b Microsiemens per centimeter.c Yasutaka et al. (2012).

wat e r r e s e a r c h 6 0 ( 2 0 1 4 ) 1 5e2 7 19

some sampling locations 2 days before the samples were

collected. On 27e29 January 2013, river water was sampled at

16 locations (AeP) on the Abukuma, Kuchibuto, and Shakado

Rivers, including the 7 sampling locations mentioned above.

Precipitation of less than 10 mm was observed on the day

before the samples were collected, but the water level on the

Abukuma River was almost unchanged (Ministry of Land,

Infrastructure and Transport, MLIT). However, snow covered

the ground near the river at the sampling time (e.g., snow

depth was 20 cm at location B and 23 cm around location F as

reported by the closest weather station of the Japan

Meteorological Agency). Thus, river water also contained

somemelted snow from the land surface around the sampling

locations. On 7 May 2013, river water was sampled from the

Ota River in Minamisoma City; the last precipitation event in

the watershed was 5 days before the sampling.

In the field, 20e60 L river water was collected by pumping

with a screw pump (Tempest DTW, Proactive Environmental

Products�, screen size was 2 mm), or by filling a bucket low-

ered from a rope from the center of a bridge. From the view of

appropriate sampling, at a large river (e.g. 100e1000 m3/s), a

depth- and width-integrated sample was preferred. In our

investigation, the flow rate of point G (the downstream of

Abukuma River) was over 100m3/s. It should be paid attention

that sample of point G was not well mixed sample of river

water.

In the laboratory, sampled water was filtered through a

0.45-mm membrane filter. The amount of suspended solids

(SS) was determined by weighing themembrane filters before

and after filtration. The filtrate was concentrated to 2 L by

evaporation. The membrane filter and concentrated water

were subjected to a germanium (Ge) semiconductor detector

to determine the particulate radiocesium concentration

(Cs_par) and dissolved radiocesium concentration (Cs_dis),

respectively. The membrane filter was analyzed in a U-8

container for 4000 s, while the concentrated water samples

were analyzed in 2-L Marinelli containers for 43,200 s to

determine the concentrations of 134Cs and 137Cs. 137Cs con-

centrations per unit weight of SS (137Cs_SS) [Bq/g] were

designated by dividing 137Cs_par [Bq/L] by SS concentration

[mg/L]. Water temperature and pH (D-51, Horiba Co., Ltd.) and

electric conductivity (ES-51, Horiba Co., Ltd.) also were

measured. The air dose rate in September 2012was estimated

through an airborne survey by the MEXT (2011c), and on

January 2013 and May 2013, the air dose rate was measured

using a surveymeter (TCS-172B, Hitachi Aloka Co., Ltd.)

located 1 m above the surface as an indicator of the ground

borne radiocesium contamination in our measurement lo-

cations. Actually in the survey by the MEXT (2011a), a relative

equation was reported between the total radiocesium con-

centration in soil particles and the air dose rate measured

above the ground surface.

Table 3 e Concentrations of 134Cs and 137Cs in dissolved and particulate form, ratios of 134Cs to 137Cs concentrations with counting error of the Ge semiconductor detector,and ratios of particulate 134Cs and 137Cs to total 134Cs and 137Cs concentrations in riverwater samples. An inequality sign indicates that the radiocesium concentrationwasbelow the detection limit of the Ge semiconductor detector.

River (sampling month) Sampling location Dissolved form Particulate form Proportion of particulate form134Cs_dis (Bq/L)

137Cs_dis (Bq/L)134Cs/137cs 134Cs_par (Bq/L)

137Cs_par (Bq/L)134Cs/137cs 134Cs 137Cs

aAbukuma River (Sep. 2012) A Kawatani <0.005 <0.003 e <0.002 <0.001 e e e

B Shirakawa 0.008 � 0.001 0.014 � 0.002 0.57 0.014 � 0.001 0.024 � 0.001 0.58 0.64 0.63

C Sukagawa 0.007 � 0.002 0.010 � 0.002 0.70 0.006 � 0.001 0.008 � 0.001 0.75 0.46 0.44

D Akutsu 0.013 � 0.002 0.027 � 0.003 0.48 0.031 � 0.001 0.052 � 0.001 0.60 0.70 0.66

E Nihonmatsu 0.011 � 0.002 0.025 � 0.002 0.44 0.053 � 0.001 0.086 � 0.001 0.62 0.83 0.77

F Fukushima-city 0.014 � 0.003 0.026 � 0.003 0.54 0.081 � 0.002 0.126 � 0.002 0.64 0.85 0.83

G Yahata 0.024 � 0.003 0.044 � 0.003 e 0.075 � 0.002 0.126 � 0.002 0.60 0.76 0.74

Abukuma River (Jan. 2013) A Kawatani <0.003 <0.003 e <0.005 <0.005 e e e

B Shirakawa <0.007 0.007 � 0.002 e <0.007 0.011 � 0.002 e e 0.61

C Sukagawa 0.008 � 0.001 0.014 � 0.002 0.57 0.031 � 0.003 0.056 � 0.004 0.55 0.79 0.80

D Akutsu 0.005 � 0.002 0.008 � 0.002 0.63 <0.007 0.017 � 0.003 e e 0.68

E Nihonmatsu 0.015 � 0.002 0.021 � 0.003 0.71 0.026 � 0.004 0.051 � 0.004 0.51 0.63 0.71

F Fukushima-city 0.015 � 0.002 0.020 � 0.003 0.75 0.031 � 0.005 0.061 � 0.006 0.51 0.67 0.75

G Yahata 0.013 � 0.001 0.023 � 0.002 0.57 0.020 � 0.003 0.043 � 0.004 0.47 0.61 0.65

Kuchibuto River (Jan. 2013) H Nagayama <0.005 0.008 � 0.002 e 0.002 � 0.0005 0.004 � 0.0005 0.50 e 0.33

I Mizusakai 0.018 � 0.002 0.027 � 0.002 0.67 0.014 � 0.001 0.027 � 0.001 0.52 0.43 0.51

J Tazawa <0.004 0.006 � 0.001 e 0.008 � 0.003 0.016 � 0.003 0.50 e 0.73

K Doumeki <0.004 0.006 � 0.001 e 0.014 � 0.003 0.026 � 0.006 0.54 e 0.80

L Fukawada <0.003 0.004 � 0.001 e <0.007 0.011 � 0.002 e e 0.72

M Nishitani <0.004 <0.005 e 0.012 � 0.002 0.036 � 0.003 0.33 e e

Shakado River (Jan. 2013) N Nishigou <0.004 <0.004 e 0.003 � 0.0004 0.007 � 0.0005 0.43 e e

O Shimomatsumoto <0.004 0.005 � 0.001 e 0.004 � 0.0005 0.008 � 0.001 0.50 e 0.63

P Horisoko 0.007 � 0.002 <0.009 e 0.020 � 0.003 0.025 � 0.004 0.80 0.73 e

Ota River (May 2013) Q 0.051 � 0.004 0.101 � 0.006 0.50 0.207 � 0.009 0.444 � 0.013 0.47 0.80 0.82

R 0.062 � 0.005 0.163 � 0.006 0.38 0.053 � 0.005 0.080 � 0.006 0.66 0.46 0.33

S 0.058 � 0.003 0.112 � 0.004 0.52 0.102 � 0.006 0.210 � 0.009 0.49 0.64 0.65

T 0.095 � 0.005 0.198 � 0.007 0.48 0.044 � 0.004 0.086 � 0.006 0.51 0.31 0.30

a Yasutaka et al. (2012).

water

research

60

(2014)15e27

20

wat e r r e s e a r c h 6 0 ( 2 0 1 4 ) 1 5e2 7 21

2.4. Regression analysis of airborne 137Cs in watershedsand 137Cs concentrations in river water

The distribution of 137Cs deposition [Bq/m2] (Fig. 1) was ob-

tained from the 137Cs Concentration Distribution Map in Soil,

which was reported by the fourth airborne monitoring survey

for October 22eNovember 5, 2011 (MEXT, 2011c). The method

for preparing the distributions of deposited 137Cs concentra-

tions has been described previously (Yasutaka et al., 2013b).

Partial 1-km grid data for 137Cs deposition were added to the

data in the 137Cs Concentration Distribution Map in Soil

report, especially for location A, and the northwest area of

location F, where 137Cs depositionwas very low: only 6% of the

entire area.

The amount of 137Cs deposition (Fig. 1) was high in the

northeast area, near the FDNPP, and low in the northwest and

southeast areas, where values were in the range of

5.6 � 103e3.0 � 106 Bq/m2, which represents a 536-fold dif-

ference. The 137Cs deposition was greater in the downstream

portion of the Abukuma River than in the upstreamportion. In

contrast, the opposite was true in the Kuchibuto and Ota

Rivers.

Based on these data, the average 137Cs deposition in each

watershed through November 2011 was calculated by multi-

plying 137Cs deposition [Bq/m2] as measured by point data

with the regions [m2] for which they were representative (at a

spatial resolution of 100 m to 1 km), and the sum of these

values was divided by the watershed area [m2].

The relations between amount of 137Cs deposition and

measured dissolved and particulate 137Cs concentrations

(137Cs_dis and 137Cs_par) or the amount of 137Cs in SS per unit

weight (137Cs_SS) were examined by regression analysis.137Cs_dis of the Shakado River was eliminated from the anal-

ysis because of the limited sample numbers.

In this analysis, only 137Cs was used for widespread anal-

ysis because low concentration of 134Cs in water samples and

its relatively short half-life. Immediately after the FDNPP ac-

cident, mainly 134Cs and 137Cs were deposited on land near the

FDNPP. However, at the time of sampling (Sep. 2012eMay

2013), nearly half of the 134Cs had decayed due to its 2.1-year

half-life, and the concentrations of 134Cs in water samples

were lower than those of 137Cs (for some samples, 134Cs con-

centrations were below the detection limits). In addition,

almost all the 137Cs observed in the water samples in this

study were derived from the FDNPP accident, because 137Cs

concentrations in river water before the accident around the

sampling area (100 km from the FDNPP) were in the order of

magnitude of 10�4 Bq/L (Matsunaga et al., 1991), which is less

than one-tenth of the observed 137Cs concentrations in this

study.

3. Results

3.1. Air dose rate and water quality

Table 2 shows the air dose rate (inmicrosieverts per hour; mSv/

h) and water quality (pH, water temperature, electric con-

ductivity, SS concentrations) of river water samples. The air

dose rates were higher downstream of the Abukuma River

(locations EeG aswell as location I), where a high deposition of137Cs was observed, when compared to the upperstream re-

gions. In addition, the air dose rates at the sampling locations

of the Abukuma River decreased from September 2012 to

January 2013. This decrease could be due to not only nuclear

decay but also the shielding effect of the gamma ray by snow

cover, emitted from the radiocesium in the surface of river

bank soil.

The river temperature was 24e29 �C in September 2012,

2e5 �C in January 2013, and 10e15 �C in May 2013. The pH of

river water samples was 7.6e8.7. In 2013, electrical conduc-

tivity in the Abukuma River (160e320 mS/cm) was higher than

that in the Kuchibuto and Shakado Rivers (60e180 mS/cm). The

relatively high electrical conductivity of the Abukuma River

water indicates that the Abukuma River water contains some

drainage water from urban areas surrounding the river (see

Appendix Fig. A and Table A), because the water discharged

from residential areas is apt to have a larger amount of ions

than that discharged from farmland or forest area. The

amount of SS was low over the entire sampling period

(<0.5e7.7mg/L), but was lowest in the upstreamportion of the

Kuchibuto River and Shakado River. Themeasured range of SS

concentration (0.3e7.7 mg/L) suggests that the river on the

sampling day was relatively calm, compared to the SS con-

centration data by theMLIT (i.e., SS 1e22mg/L around location

G).

3.2. Dissolved and particulate radiocesiumconcentrations in river water

Table 3 shows dissolved and particulate 134Cs and 137Cs con-

centrations in water samples from the Abukuma, Kuchibuto,

Shakado, and Ota Rivers. The activity ratio of 134Cs to 137Cs

concentrations were 0.38e0.75 for the dissolved form, and

0.33e0.80 for the particulate form. The proportion of particu-

late 137Cs to the total 137Cs concentrations were 30e83%; in 21

out of 25 samples. The 137Cs_par concentrations were greater

than the 137Cs_dis concentrations (not including samples of

2012 A and 2013 A because particulate and dissolved 137Cs

were lower than the detection limit).

In the Abukuma River in September 2012, the 137Cs_disconcentrations were <0.005e0.044 Bq/L, and 137Cs_par con-

centrations were <0.005e0.126 Bq/L. In January 2013, the137Cs_dis concentrations decreased to <0.003e0.023 Bq/L and137Cs_par to <0.005e0.061 Bq/L. At location A, two samples had

both dissolved and particulate 134Cs and 137Cs concentrations

below the detection limits. In the Kuchibuto River, the 137Cs_disconcentrations were <0.005e0.027 Bq/L and particulate 137Cs

were 0.004e0.036 Bq/L, values slightly less than those from the

Abukuma River. In the Shakado River, 137Cs_dis concentrations

in samples from all locations were below the detection limit,

except for the sample from location O. The 137Cs_par concen-

trations were 0.007e0.025 Bq/L, which were less than those

from the Abukuma River. The 137Cs concentrations of river

water samples from the Ota River were 10-fold greater than

those from other rivers; 0.101e0.198 Bq/L in dissolved form

and 0.080e0.444 Bq/L in particulate form.

Fig. 2aee shows the 137Cs_dis and 137Cs_par in river water

and the air dose rate with distance from the river mouths or

wat e r r e s e a r c h 6 0 ( 2 0 1 4 ) 1 5e2 722

the confluence with the Abukuma River. For the Abukuma

River in 2012, both the 137Cs_dis and 137Cs_par concentrations

were greater in the downstream area. A similar pattern was

observed in January 2013, but the trend was weaker. In 2012,

200 150 100 50 0

Abukuma River (Sep. 2012)

Kuchibuto River (Jan. 2013)

a b

c d

Con

cent

ratio

nsof

137 C

s (B

q/L

)C

once

ntra

tions

of

137 C

s (B

q/L

)

Air

dos

e ra

te (µ

Sv/

h)

Con

cent

ratio

ns o

f 13

7 Cs

(Bq/

L)

e

Distance from the confluence with the Abukuma River (km)

Ota River (May 2013)

A B C D E F G

upstream downstreamDistance from mouth (km)

3210

I J K L M

upstream downstream u

Q R S T

1

0.5

0

upstream downstreamDistance from the mouth (km)

0.2

0.1

0

0.2

0.1

0

Air

dos

e ra

te (µ

Sv/

h)

3210

Air

dos

e ra

te (µ

Sv/

h)

3210

50 40 30 20 10 0

15 10 5 0

Fig. 2 e Concentrations of dissolved and particulate 137Cs (137Cs

Abukuma River in September 2012, (b) Abukuma River in Janua

River in January 2013, and (e) Ota River in May 2013. The horizo

(mouth of the Abukuma and Ota Rivers, or the confluence with

Error bars indicate the counting error of the Ge semiconductor d

the detection limit of the detector were not included (Table 3 sho

H was eliminated because this point is located on a different ch

the air dose rate maximum occurred near location C or D;

however, the 137Cs pattern did not correspond to that of the air

dose rate. In the Kuchibuto River, 137Cs_dis and air dose rate

were particularly high at location I (0.027 Bq/L and 2.3 mSv/h),

Abukuma River (Jan. 2013)

Shakado River (Jan. 2013)

Con

cent

ratio

ns o

f 13

7 Cs

(Bq/

L)

Con

cent

ratio

ns o

f 13

7 Cs

(Bq/

L)

(Particulate 137Cs)(Dissolved 137Cs)

Air dose rate

137Cs dis

137Cs par

upstream downstreamDistance from mouth (km)

N O P

Air

dos

e ra

te (µ

Sv/

h)

Distance from the confluence with the Abukuma River (km)

pstream downstream

A B C D E F G

Air

dos

e ra

te (µ

Sv/

h)

32100.2

0.1

0

32100.2

0.1

0

200 150 100 50 0

30 20 10 0

_dis and 137Cs_par) in water, and the air dose rate for (a)

ry 2013, (c) Kuchibuto River in January 2013, (d) Shakado

ntal axis indicates the distance from the end of each river

the Abukuma River of the Kuchibuto and Shakado Rivers).

etector. Data points with radiocesium concentrations below

ws detection limits). In Fig. 2c, the concentration at location

annel from that of location I (see Fig. 1).

wat e r r e s e a r c h 6 0 ( 2 0 1 4 ) 1 5e2 7 23

but these values decreased by over a factor of 4 downstream.

In the Shakado River, the 137Cs formwasmainly as particulate

and increased downstream. In the Ota River, 137Cs_dis con-

centrations fluctuated along the river channel.

3.3. 137Cs concentration per unit weight of suspendedsolid (137Cs_SS)

Fig. 3aed shows the 137Cs_SS in river waters at sampling loca-

tions with distance from the rivermouths or the confluence of

the Abukuma River. For the Abukuma River, 137Cs_SS concen-

trations were 7e21 Bq/g in September 2012 and 5e16 Bq/g in

January 2013. At locations C and E, amounts of 137Cs_SSincreased, but the average value for all of the locations on the

Abukuma River decreased. The 137Cs_SS were 15e54 Bq/g for

the Kuchibuto River, 4.3e6.5 Bq/g for the Shakado River, and

59e160 Bq/g for the Ota River. These results indicate that137Cs_SS valueswerehighwhere 137Csdepositionwasalsohigh.

3.4. Relations between 137Cs deposition and measured137Cs concentrations in river water

Fig. 4 shows the relations between 137Cs deposition in each

watershed and measured 137Cs concentrations. Points for137Cs or SS amounts below the detection limits were elimi-

nated from the plots. Regression lines were applied and slopes

and correlation coefficients (R2) were calculated (Table 4).

Sep. 2012

A B C D E F G

137 C

s_SS

(Bq/

g)

30

20

10

0

upstream downstreamDistance from mouth (km)

200 150 100 50 0

137 C

s_(B

q/g)

up

Abukuma Rivera

N O P

30 20 10 0

137 C

s_SS

(Bq/

g)

30

20

10

0

upstream downstream

137

uDistance from the confluence with the Abukuma River (km)

Shakado River (Jan. 2013)c

Jan. 2013

Fig. 3 e Concentrations of 137Cs in suspended solids (SS) (137Cs_Shakado River, and (d) Ota River. The horizontal axis indicates t

the minimum value was obtained by dividing 137Cs_par by the u

For the Abukuma River, the relations between average137Cs deposition and: 137Cs_dis (Fig. 4a), 137Cs_par (Fig. 4b), or137Cs_SS (Fig. 4c) in river water were positively correlated, with

R2 values of 0.76, 0.92, and 0.93, respectively. However, in

January 2013, the R2 values decreased to 0.49, 0.07, and <0.01,

respectively. The slope of the regression line of 137Cs_dis in

January 2013 was lower than that in September 2012. For the

Kuchibuto River, a positive correlation was observed between137Cs deposition and 137Cs_dis, but the

137Cs_par did not appear

to correlate with 137Cs deposition (R2 ¼ 0.07); however, a

regression line of R2 ¼ 0.47 was obtained for 137Cs_SS. For the

Ota River, a negative correlation was observed between137Cs_dis in river waters and 137Cs deposition. In contrast, a

positive correlation was observed between particulate 137Cs_SSand 137Cs deposition, with an R2 value of 0.91.

4. Discussion

4.1. Spatial distribution of 137Cs concentration in riverwater

Regarding the spatial distribution of 137Cs concentration

along the rivers, relatively high 137Cs_dis were observed at

locations in watersheds that experienced relatively high 137Cs

deposition after the FDNPP accident. High 137Cs_dis were

SS

30

20

10

050 40 30 20 10 0

(> 54 Bq/g)

(> 16 Bq/g)

stream downstream

Kuchibuto River (Jan. 2013)bI J K L M

Cs_

SS(B

q/g)

200

150

100

50

015 10 5 0

pstream downstreamDistance from mouth (km)

Distance from the confluence with the Abukuma River (km)

Q R S T

d Ota River (May 2013)

SS)[Bq/g] from (a) Abukuma River, (b) Kuchibuto River, (c)

he distance from the end of each river. At locations I and L,

pper limit of SS concentration (0.5 mg/L).

Sep. 2012

ML

KJ

0

5

10

15

20

25

0 200 400

Kuchibuto River Ota River

137 C

s _di

s (B

q/L

)13

7 Cs _

par (B

q/L

)13

7 Cs _

SS

(Bq/

g)

Abukuma River

Q

R

S

T

0

0.1

0.2

0.3

0 1,000 2,000

Q

R

S

T0

0.2

0.4

0.6

0.8

0 1,000 2,000

Q

RS

T

0

50

100

150

200

0 1,000 2,000

LK J

I

H

0

0.01

0.02

0.03

0.04

0.05

0 500 1,000

M

LK

JI

H0

0.05

0.1

0.15

0 500 1,000

Deposited 137Cs (kBq/m2)

FED

CB

GF

E

D

C

B0

0.01

0.02

0.03

0.04

0.05

0 50 100 150

G

GF

E

D

C BG

FE

D

C

B0

0.05

0.1

0.15

0 50 100 150

G

FE

BC

D

GF

E

D

C

B0

10

20

30

0 50 100 150

Jan. 2013 Jan. 2013 May 2013a1 a2 a3

b1 b2 b3

c1 c2 c3

Fig. 4 e Scatter plots showing deposited 137Cs and (a) 137Cs_dis (dissolved form), (b) 137Cs_par (particulate form), and (c) 137Cs_SS(in suspended solids per unit weight) in water from the Abukuma River (subscript 1), Kuchibuto River (subscript 2), and Ota

River (subscript 3). Solid line, dash line, and dot-dash line indicate the regression lines of sample water from Sept. 2012, Jan.

2013, and May 2013, respectively. Results from the Shakado River are not displayed because of inadequate sample numbers.

wat e r r e s e a r c h 6 0 ( 2 0 1 4 ) 1 5e2 724

observed in the downstream areas for the Abukuma River

(Fig. 1, DeF), midstream area for the Shakado River (O), and

upstream areas of the Kuchibuto River (I). In particular for the

Kuchibuto River, more 137Cs concentrated water at location I

was diluted by relatively clean river water discharged from an

area that received less 137Cs deposition; therefore, 137Cs_disdecreased suddenly downstream of the dilution point (from I

to J).

For the Ota River, higher 137Cs_dis were monitored

comparing to the other three rivers (Table 3 and Fig. 2),

Table 4 e Slopes and coefficients of determination for regressioin each watershed). R2 value represents the determination coe

Samplingmonth

River

Dissolved 137Csconcentration

Slopes (m�1) R2

Sep. 2012 Abukuma 4.8 � 10�4 0.76

Jan. 2013 Abukuma 2.1 � 10�4 0.49

Kuchibuto 3.7 � 10�5 0.98

May 2013 Ota �2.1 � 10�4 0.36

nevertheless the air dose rate and concentration of deposited137Cs in our sampling points were almost the same level as the

Abukuma River (Figs. 1 and 2). Thismight be because themain

source of dissolved 137Cswas upstreamof the Ota River, where

higher amounts of 137Cs deposited (2.9 � 105e3.0 � 106 Bq/m2,

Fig. 1) compared with the Abukuma River (5.6 � 103e1.4 �106 Bq/m2, Fig. 1). Accordingly, the trend of 137Cs_dis along the

river was not observed because of little the difference of each

watershed area (QeT) compared to the area of main 137Cs

source in upstream.

nmodels (linear expression to the average 137Cs depositionfficient.

Dependent variable

Particulate 137Csconcentration

137Cs concentrationper unit weight of SS

Slopes (m�1) R2 Slopes (m2/kg) R2

2.1 � 10�3 0.92 2.9 � 10�4 0.93

2.4 � 10�4 0.07 �4.8 � 10�7 <0.01

1.3 � 10�5 0.07 7.8 � 10�5 0.47

7.9 � 10�4 0.38 3.2 � 10�4 0.91

wat e r r e s e a r c h 6 0 ( 2 0 1 4 ) 1 5e2 7 25

In contrast to 137Cs_dis, the137Cs_par pattern did not show a

strong trend along the stream lengths, especially for the

Kuchibuto and Ota Rivers in 2013 (Fig. 2), because 137Cs_pardepended directly on the SS concentration, which fluctuated

due to disturbance and heavy precipitation. Therefore, 137Cs

concentrations per unit weight of SS were determined to

eliminate the effect of fluctuating SS concentrations. This

revealed that 137Cs_SS was greater in the upstream areas of the

Kuchibuto and Ota Rivers, where the 137Cs deposition was

high (Fig. 3), which is a trend similar to that found for 137Cs_dis.

The 137Cs_dis and137Cs_par decreased from Sep. 2012 to Jan.

2013 in the Abukuma River, and this change was the similar to

the results obtained by Spezzano et al. (1994). However, the

long-term trend 137Cs_dis and 137Cs_par must be clarified by

continuous monitoring.

4.2. Quantitative relations between 137Cs deposition and137Cs concentration in watersheds

In the regression analysis for the Abukuma River in September

2012 (Fig. 4), both 137Cs_dis and137Cs_par showed strong linear

relations (R2 value ¼ 0.92 and 0.76, respectively) with 137Cs

deposition, and the correlation coefficient increased when137Cs_par was converted to 137Cs_SS (R2 value ¼ 0.93). In

contrast, a smaller correlation value (137Cs_dis: R2 value ¼ 0.49;

137Cs_par: R2 value ¼ 0.07) was found for the samples collected

in January 2013. However, elimination of the data from

137 C

s _di

s or

137 C

s _pa

r(B

q/L

)

1

1

137 C

s _S

S ( B

q/g)

GF

BD

G

F

BD

GF

B

D

0

4

8

12

0

0.05

0.1

0 50 100 150Deposited 137Csp (kBq/m2)

137Cs_par = 9.7 10-4 137Csp – 0.064, R² = 0.82

137Cs_SS = 0.15 137Csp – 7.53, R² = 0.88137Cs_dis = 3.8 10-4 137Csp – 0.023, R² > 0.99

G

F

DB

M

LKJ

I

H

P

ON0

0.05

0.1

0 200 400 600 800 1,000

137Cs_par = 1.3 10-6 137Csp + 0.022, R² < 0.01

Deposited 137Csp (kBq/m2)

137 C

s _pa

r (B

q/L

)

(Particulate)(Dissolved)(Unit weight of SS)

137Cs_dis

137Cs_par

137Cs_SS

a

c d

.

.

.

.

Fig. 5 e (a) Relationships between deposited 137Cs (137Csp) and1

137Cs_SS (in SS per unit weight) in the Abukuma River alone, an137Cs_par, and (d) 137Cs_SS in the Abukuma, Kuchibuto, and Shak

locations C and E.

locations C and E resulted in good correlation between

deposited 137Cs_dis and137Cs_par from the other four locations

(Fig. 5a, 137Cs_dis: R2 value ¼ 0.99; 137Cs_par: R

2 value ¼ 0.82).

From the given regression line, the slope values (137Cs_dis:

3.8 � 10�4 m2/kL; 137Cs_par: 9.7 � 10�4 m2/kL) were slightly less

than those of regression lines created from 2012 data (137Cs_dis:

4.8 � 10�4 m2/kL; 137Cs_par: 2.1 � 10�3 m2/kL). At locations C

and E, the ratios of 137Cs_dis and 137Cs_SS to the amount of

deposited 137Cs were particularly higher than the other four

locations. These results suggest that river water at locations C

and E may have unique runoff characteristics, and there is

possibility thatmunicipal effluent from the neighboring urban

areas of Nihonmatsu (locations C) and Sukagawa (location E)

might significantly influence the river quality. In fact, the flow

on the sampling day in January was the lowest level over the

past year (monitored by MLIT). Further research should be

undertaken to evaluate the effect ofmunicipal effluent to river

water.

For the Kuchibuto River in January 2013, both 137Cs_dis and137Cs_SS correlated with 137Cs deposition (R2 ¼ 0.98 and 0.47,

respectively). By contrast, data from the Abukuma River in

January 2013, indicated no locations showed distinct 137Cs_disor 137Cs_SS levels. This is probably because forest and culti-

vated land dominate nearly the entire area in the Kuchibuto

River watershed. Therefore the runoff ratio of 137Cs is nearly

the same as the amount of 137Cs deposition. A comparison of

slopes of the regression lines in January 2013 shows that the

137 C

s _di

s (B

q/L

)

GF

B

D

ML

K

J

NOP

0

5

0

5

20

25

0 100 200 300 400

Deposited 137Csp (kBq/m2)

Deposited 137Csp (kBq/m2)

137 C

s _SS

(Bq/

g) GF

D

BL

JK

I

H

O0

0.01

0.02

0.03

0 500 1,000

137Cs_dis = 1.5 10-5 137Csp + 0.008, R² = 0.09

137Cs_SS = 0.060 137Csp + 0.54, R² = 0.86

b

.

.

37Cs_dis (dissolved form), 137Cs_par (particulate form), and

d the relation between deposited 137Cs and (b) 137Cs_dis, (c)

ado Rivers combined for January 2013, excluding data from

wat e r r e s e a r c h 6 0 ( 2 0 1 4 ) 1 5e2 726

slope of the 137Cs_dis for the Kuchibuto River (3.7 � 10�5 m2/kL,

Table 4) was one-tenth of that for the Abukuma River except

for locations C and E (3.8 � 10�4 m2/kL, Fig. 5a). These results

indicates that the runoff ratio of 137Cs_dis was larger in the

Abukuma River than in the Kuchibuto River. Additionally, the

runoff from urban areas, especially those located along the

Abukuma River, was greater than that from forest and culti-

vated land. The area ratio of buildings lands along the Abu-

kuma River basin (Table A, point G) was greater than that of

Kuchibuto River basin (Table A, point M). There is a possibility

that the runoff ratio of 137Cs_dis from urban areas might be

greater than that from forest and paddy field because of its

high adsorption characteristics of 137Cs_dis.

For the Ota River, a negative correlation was found be-

tween 137Cs deposition and 137Cs_dis (Fig. 4a3). Asmentioned in

Section 4.1, this might be because the high 137Cs deposited

area upstream of sampling locations, which occupies almost

half of the watershed, was the dominant source of dissolved137Cs, and the distance from the highly contaminated area to

each sampling pointwas short. However, the 137Cs_SS was very

strongly correlated with 137Cs deposition.

Next, data for all of the river water sampled from January

2013, except for samples from locations C and E (Fig. 5bed),

was analyzed together. Regression analysis between 137Cs

deposition and river concentrations showed that the R2

value was 0.09 for 137Cs_dis, <0.01 for 137Cs_par, and 0.86 for137Cs_SS. The lower correlation value of 137Cs_dis shown in

Fig. 5b than that in Fig. 4a1 and 4a2 indicates that the

mechanism for the runoff of dissolved 137Cs for the Abu-

kuma River was different than that for the Kuchibuto and

Shakado Rivers. The correlation of 137Cs_par shown in Fig. 5c

was weak, because SS concentrations from samples

collected at the sampling locations were more than 12-fold

difference (Table 2).

The correlation coefficient value for 137Cs_SS shown in

Fig. 5d was high (R2 ¼ 0.86). Additionally, the slope of the line

of 137Cs_SS plotted against deposited 137Cs decreased from

summer to winter (2.9 � 10�4e1.5 � 10�4 m2/kg; Table 4 and

Fig. 5a). These results mentioned above indicates that 137Cs

deposition and 137Cs_SS in river water might have high corre-

lation regardless of the river characteristics for different pe-

riods, but slope values could vary at different periods due to

nuclear decay, discharge or runoff coefficient. The MEXT and

MAFF (2012) measured the distribution of 137Cs in the

riverbed sediment 30e50 km north of the FDNPP (including

the whole watershed of the Ota River, and part of the Abu-

kuma and Kuchibuto Rivers), and found a relatively high R2

value (0.83) to the total 137Cs concentrations in water.

Data show that the spatial and temporal variation in SS

concentrations were consistent with the relatively weak re-

lations of 137Cs_par to the amount of deposited 137Cs. Consid-

ering that the dominant form of 137Cs in rivers was the

particulate form (Table 3), predicting total 137Cs concentration

in river water directly from the amount of deposited 137Cs is

not recommended. In contrast, the 137Cs_dis and 137Cs_SSshowed a stronger correlation with 137Cs deposition at each

sampling point. Therefore, only 137Cs_dis and 137Cs_SS can be

predicted from data on deposited 137Cs. The SS concentrations

would need to be obtained separately to predict total 137Cs

concentration in river water.

5. Conclusions

In this study, current levels of dissolved and particulate 137Cs

in river waters after the FDNPP accident were measured, and

the relations between 137Cs concentration and deposition

were determined by regression analysis.

1) In 21 of 25 water samples, 134Cs and 137Cs concentration in

the particulate formwere greater than that in the dissolved

form.

2) Relatively high dissolved 137Cs concentrationwas observed

where greater 137Cs deposition occurred. The trend in dis-

solved 137Cs concentration along the rivers followed this

relation; decreased downstream for the Abukuma River,

and the reverse was observed for the Kuchibuto River. In

contrast, particulate 137Cs concentrations along the stream

varied. Similar trends with dissolved 137Cs were observed

between downstream distance along the rivers and the137Cs concentration per unit weight of SS.

3) Results of regression analysis for each river showed a

strong correlation between 137Cs deposition and dissolved137Cs concentration, and also for the 137Cs concentration

per unit weight of SS.

4) The dissolved 137Cs concentrations were regressed against137Cs deposition for the Kuchibuto River, thereby the slope

was one-tenth of the slope for the Abukuma River. This

result suggests that the drainage from urban areas (rather

than forest and cultivated fields) was a dominant source of

the dissolved 137Cs contamination.

5) 137Cs concentrations per unit weight of SS plotted against137Cs deposition for the Abukuma and Kuchibuto River

had high R2 value, indicating that 137Cs deposition

and 137Cs concentrations per unit weight of SS in river

water have high correlation regardless of the river

characteristics.

6) Because particulate 137Cs was dominant in river water and

particulate 137Cs concentration was not highly correlated

with the amount of 137Cs deposition, it is difficult to predict

total 137Cs concentration directly from the amount of

deposited 137Cs. The particulate 137Cs concentration must

be predicted by estimating SS concentrations.

Acknowledgments

We would like to thank the Fukushima Agricultural Technol-

ogy Centre and EAC corporation for their significant support of

water sampling. Analysis of water samples was conducted by

EAC corporation, Japan Environment Science Co., Ltd., and the

Department of Chemical Biology and Applied Chemistry of

Nihon University.

Appendix A. Supplementary data

Supplementary data related to this article can be found at

http://dx.doi.org/10.1016/j.watres.2014.04.024.

wat e r r e s e a r c h 6 0 ( 2 0 1 4 ) 1 5e2 7 27

r e f e r e n c e s

Comans, R.N.J., Haller, M., Depreter, P., 1991. Sorption of cesiumon illite: nonequilibrium behavior and reversibility. Geochim.Cosmochim. Acta 55 (2), 433e440.

Cooper, J.A., 1970. Factors determining the ultimate detectionsensitivity of Ge(Li) gamma-ray spectrometers. Nucl. Instrum.Methods 82, 273e277.

Evans, D.W., Alberts, J.J., Clark, R.A., 1983. Reversible ion-exchange fixation of cesium-137 leading to mobilization fromreservoir sediments. Geochim. Cosmochim. Acta 47 (6),1041e1049.

Kryshev, I.I., 1995. Radioactive contamination of aquaticecosystems following the Chernobyl accident. J. Environ.Radioact. 27 (3), 207e219.

Matsunaga, T., Amano, H., Yanase, N., 1991. Discharge ofdissolved and particulate 137Cs in the Kuji River, Japan. Appl.Geochem. 6, 159e167.

Matsunaga, T., Ueno, T., Amano, H., Tkatchenko, Y., Kovalyov, A.,Watanabe, M., Onuma, Y., 1998. Characteristics of Chernobyl-derived radionuclides in particulate form in surface waters inthe exclusion zone around the Chernobyl Nuclear Power Plant.J. Contam. Hydrol. 35 (1e3), 101e113.

McDougall, S., Hilton, J., Jenkins, A., 1991. A dynamic model ofcesium transport in lakes and their catchments. Water Res. 25(4), 437e445.

Nagao, S., Kanamori, M., Ochiai, S., Tomihara, S., Fukushi, K.,Yamamoto, M., 2013. Export of 134Cs and 137Cs in theFukushima river systems at heavy rains by Typhoon Roke inSeptember 2011. Biogeosci. Discuss. 10, 2767e2790.

Sakaguchi, A., Kadokura, A., Steier, P., Tanaka, K., Takahashi, Y.,Chiga, H., Matsushima, A., Nakashima, S., Onda, Y., 2012.Isotopic determination of U, Pu and Cs in environmentalwaters following the Fukushima Daiichi Nuclear Power Plantaccident. Geochem. J. 46 (4), 355e360.

Smith, J.T., Wright, S.M., Cross, M.A., Monte, L., Kudelsky, A.V.,Saxen, R., Vakulovsky, S.M., Timms, D.N., 2004. Globalanalysis of the riverine transport of Sr-90 and Cs-137. Environ.Sci. Technol. 38 (3), 850e857.

Spezzano, P., Bortoluzzi, S., Giacomelli, R., Massironi, L., 1994.Seasonal variations of Cs-137 activities in the Dora BalteaRiver (northwest Italy) after the Chernobyl accident. J. Environ.Radioact. 22 (1), 77e88.

Ueda, S., Hasegawa, H., Kakiuchi, H., Akata, N., Ohtsuka, Y.,Hisamatsu, S., 2013. Fluvial discharges of radiocaesium fromwatersheds contaminated by the Fukushima Dai-ichiNuclear Power Plant accident, Japan. J. Environ. Radioact.118, 96e104.

Vakulovsky, S.M., Nikitin, A.I., Chumichev, V.B., Katrich, I.Y.,Voitsekhovich, O.A., Medinets, V.I., Pisarev, V.V.,Bovkum, L.A., Khersonsky, E.S., 1994. Cesium-137 and Sr-90contamination of water bodies in the areas affected byreleases from the Chernobyl-nuclear-power-plant accident:an overview. J. Environ. Radioact. 23 (2), 103e122.

Walling, D.E., Bradley, S.B., 1990. Some applications of caesium-137 measurements in the study of erosion, transport anddeposition. Eros. Transp. Deposit. Process. 189, 179e203.

Yasutaka, T., Kawamoto, T., Kawabe, Y., Sato, T., Sato, M.,Suzuki, Y., Nakamura, K., Komai, T., 2013a. Rapidmeasurement of radiocesium in water using a Prussian blue

impregnated nonwoven fabric: Fukushima NPP accidentrelated. J. Nucl. Sci. Technol. 50 (7), 674e681.

Yasutaka, T., Iwasaki, Y., Hashimoto, S., Naito, W., Kishimoto, A.,Yoshida, K., Murakami, M., Kawaguchi, I., Oka, T.,Nakanishi, J., 2013b. A GIS-based evaluation of the effect ofdecontamination on effective doses due to long-term externalexposures in Fukushima. Chemosphere 93 (6), 1222e1229.

Zhu, Y.G., Smolders, E., 2000. Plant uptake of radiocaesium: areview of mechanisms, regulation and application. J. Exp. Bot.51 (351), 1635e1645.

r e f e r r e d w e b s i t e s

Emergency Operation Center, Ministry of Education, Culture,Sports, Science and Technology (MEXT), Agriculture, Forestryand Fisheries Research Council, Ministry of Agriculture,Forestry and Fisheries (MAFF) in Japan, 2012. SummarizedVersion of the “Results of the Research on Distribution ofRadioactive Substances Discharged by the Accident atTEPCO’s Fukushima Dai-ichi NPP”. http://radioactivity.nsr.go.jp/en/contents/1000/294/24/PressR04%200802s.pdf (accessed28.01.14.).

Japan Atomic Energy Agency, 2013. Comprehensive migrationsurvey of radioactive materials (in Japanese). http://fukushima.jaea.go.jp/initiatives/cat01/pdf05/02-05.pdf(accessed 04.02.14.).

Japan Meteorological Agency. Information of weather statistics(in Japanese). http://www.data.jma.go.jp/obd/stats/etrn/index.php (accessed 28.01.14.).

Ministry of Education, Culture, Sports, Science and Technology(MEXT) in Japan, 2011a. Preparation of Distribution Map ofRadiation Doses, Etc. (Map of Radioactive CesiumConcentration in Soil) by MEXT. http://radioactivity.nsr.go.jp/en/contents/5000/4165/24/1750_083014.pdf (accessed28.01.14.).

Ministry of Education, Culture, Sports, Science and Technology(MEXT) in Japan, 2011b. Results of the Survey on theDistribution of Radioactive Substances by MEXT (Survey onthe Movement of Radioactive Substances in River Water andWell Water). http://radioactivity.nsr.go.jp/en/contents/5000/4173/24/1750_102018.pdf (accessed 28.01.14.).

Ministry of Education, Culture, Sports, Science and Technology(MEXT) in Japan, 2011c. Results of the Fourth AirborneMonitoring Survey by MEXT. http://radioactivity.nsr.go.jp/en/contents/4000/3179/24/1270_1216.pdf (accessed 28.01.14.).

Ministry of Land, Infrastructure, Transport and Tourism (MLIT) inJapan. Water information system (in Japanese). http://www1.river.go.jp (accessed 31.10.13.).

World Health Organization, 2011. Guidelines for Drinking-WaterQuality, fourth ed. http://whqlibdoc.who.int/publications/2011/9789241548151_eng.pdf (accessed 26.01.14.).

Yasutaka, T., Kawabe, Y., Kusorawa, A., Komai, T., 2012.Monitoring dissolved radioactive cesium in Abukuma River inFukushima Prefecture. In: Proceedings of InternationalSymposium on Environmental Monitoring and DoseEstimation of Residents after Accident of TEPCO’s FukushimaDaiichi Nuclear Power Station. In: http://www.rri.kyoto-u.ac.jp/anzen_kiban/outcome/Proceedings_for_Web/Topics_2-17.pdf (accessed 31.10.13.).