Vertical and Lateral Transport of Particulate Radiocesium off Fukushima

Oct 13, 2014 - Observed 137Cs flux was highest in September 2011 (98 mBq m−2 day. −1. : decay-corrected to March 11, 2011), and decreased over tim...
0 downloads 0 Views 2MB Size
Download PDF
Article pubs.acs.org/est

Vertical and Lateral Transport of Particulate Radiocesium off Fukushima Shigeyoshi Otosaka,*,† Takahiro Nakanishi,‡ Takashi Suzuki,† Yuhi Satoh,† and Hisashi Narita§ †

Research Group for Environmental Science, Japan Atomic Energy Agency, 2-4 Shiraka-Shirane, Tokai-mura, Ibaraki 319-1195, Japan Fukushima Environmental Research Group, Japan Atomic Energy Agency, 1-29 Okitama-cho, Fukushima-shi, Fukushima 960-8034, Japan § School of Marine Science and Technology, Tokai University, Shimizu-ku, Shizuoka 424-8610, Japan ‡

S Supporting Information *

ABSTRACT: Transport processes of particulate radiocesium were investigated using a sediment trap deployed at about 100 km east of the Fukushima Daiichi Nuclear Power Plant. A sediment trap was installed at 873 m depth of the station (119 m above the bottom), and time-series sampling of sinking particles was carried out from August, 2011 to June, 2013. The accident-derived radiocesium was detected from sinking particles over two years after the accident. Observed 137Cs flux was highest in September 2011 (98 mBq m−2 day−1: decay-corrected to March 11, 2011), and decreased over time with seasonal fluctuations. Particulate fluxes of radiocesium were mainly affected by two principal processes. One was the rapid sinking of radiocesium-bound particles (moderate mode). This mode was dominant especially in the early postaccident stage, and was presumed to establish the distribution of radiocesium in the offshore seabed. Another mode was observed in winter, and secondary transport of particles attributed to turbulence near the seabed increased fluxes of particulate radiocesium (turbulence mode). Although the latter process would not drastically change the distribution of sedimentary radiocesium in the short term, attention should be paid as this key process redistributing the accident-derived radiocesium may cumulatively affect the long-term distribution.

■

INTRODUCTION The Great East Japan Earthquake, occurred on March 11, 2011, and induced the tsunami that triggered the accident at the Fukushima Daiichi Nuclear Power Plant (1FNPP). This accident released a large amount of radionuclides into the environment,1−3 and is still affecting the lives of people who are living around the plant. Among the accident-derived radionuclides, 137Cs (half-life: 30.1 years) is a representative to assess the effect of the accident on the environment. It is estimated that 11 PBq of 137Cs was discharged to the North Pacific by atmospheric deposition and direct discharge.2 The majority of the accident-derived radiocesium in the ocean has been advected to offshore regions, and concentrations of the radiocesium in coastal seawater have decreased by over 2 orders of magnitude.4−6 The accident-derived radiocesium has also been detected from seabed sediments. Although the abundance of the sedimentary radiocesium represents only 1−3% of the total discharge in the ocean and is gradually decreasing over time,7,8 it is estimated that the radiocesium will remain in the sediment over a time scale of decade along the Pacific coast of Japan.9,10 In fact, even three years after the accident, the accident-derived radiocesium is still detected from demersal fishes caught near Fukushima.11 In addition to the preferential accumulation of the radionuclides in the coastal region, accumulation of © 2014 American Chemical Society

radiocesium in the offshore regions (water depth >500 m) is also reported.8,12 However, the processes controlling the distribution of sedimentary radiocesium remain unknown and the temporal/spatial scale of the nearshore-offshore transport of particulate materials need clarification. Sinking particles are biological and geochemical debris which undergo various aggregation/disaggregation processes while sinking to the seabed. Sinking particles near the seabed, especially in the continental slope, contain a significant proportion of resuspended particles from the seabed.13,14 Since sinking particles transport contaminants in seawater, sediment trap experiments are helpful to investigate the migration of radionuclides in a postaccident stage of a nuclear accident. For example, after the Chernobyl accident in 1986, fluxes of accident-derived radionuclides were observed in the North Pacific,15 Black Sea,16,17 Mediterranean Sea,18 and North Sea.19 In recent research, radiocesium derived from the 1FNPP accident was detected from sinking particles collected in the North Pacific.20,21 From these studies, general characteristics of dynamics of particulate radiocesium can be summarized as Received: Revised: Accepted: Published: 12595

July 31, 2014 October 4, 2014 October 13, 2014 October 13, 2014 dx.doi.org/10.1021/es503736d | Environ. Sci. Technol. 2014, 48, 12595−12602

Environmental Science & Technology

Article

more than 10 cm in length) were subsampled by cutting into 1 cm thickness sections on board and transferred to laboratories on land for further processing. Radiochemical/Chemical Analysis. After recovery of the sediment trap and removal of “swimmers”, samples were collected on a preweighed 0.6 μm pore size membrane filter (Millipore Co.). Samples were rinsed with deionized water (>18.2 MΩ) during the filtration procedures, and were dried under a vacuum at 60 °C. After weight measurement of the dried samples for calculation of the total mass flux, sinking particles were crushed and sealed in a plastic container. Sediment samples were dried at 80 °C, pulverized, sealed in a plastic container, and stored until measurement. For sediment samples, water content and dry bulk density were measured with a given volume of plastic tube. Specific gamma-rays of 134Cs (604 and 795 keV) and 137Cs (661 keV) were measured using a coaxial Ge detector (ORTEC GEM20P4, 1.7 keV/1.33 MeV of resolution and 29−31% of relative efficiencies). Specific gamma-rays of 210Pb (46.5 keV) and 214Pb (352 keV) were measured using a low-energy photon detector (ORTEC LOAx-51370/20P, 0.625 keV/122 keV of resolution), and activities of the excess-210Pb (210Pbex) were calculated by subtracting 214Pb activities from the 210Pb activities on the assumption that the supported 210Pb from 226 Ra is equal to 214Pb. Detailed procedures for calibration and correction of summing effects are described in Otosaka and Kato.7 Under our analytical conditions (136 800−468 000 s counting), the lowest amount of 134Cs and 137Cs that could be determined in a sediment trap sample was ∼33 mBq and ∼27 mBq, corresponding to ∼34 Bq kg−1 and ∼28 Bq kg−1, respectively. Concentrations of the radionuclides reported in the following sections are represented as Bq kg−1 by dry weight. Activities of radiocesium were decay-corrected to March 11, 2011, and activities of 210Pbex were decay-corrected to the date of sampling. An aliquot (∼60 mg) of powdered sample was decomposed with a mixed acid solution in a Teflon-sealed vessel,22 and concentrations of aluminum (Al) and calcium (Ca) were measured by ICP atomic emission spectrometry. Concentrations of lithogenic matter and biogenic carbonate were calculated with eqs 1 and 2 assuming that the Al and Ca concentrations of the upper crust of the Japanese arc are 7.8% and 2.7%, respectively.23

follows, (1) radiocesium is removed from the surface water at rates of 0.2−1% year−1,16,19,20 (2) particulate radiocesium sinks rapidly during the early post accidental stage,15−21 and (3) secondary (lateral) transport of radiocesium-bound lithogenic particles occurs.17 These findings were obtained from stations located more than 500 km away from the accident sites, but little is known about dynamics of the radionuclides at closer region to the contaminant source. In this paper, accumulation and redistribution processes of particulate radiocesium near Fukushima were investigated by a sediment trap experiment carried out at about 100 km east of the 1FNPP over two years after the 1FNPP accident.

■

EXPERIMENTAL SECTION Field Sampling. A mooring array with a sediment trap was deployed at station FS1 (37° 20.1′ N 142° 10.0′ E, bottom depth 992 m) (Figure 1). A sediment trap (Nichiyu Giken

Lithogenic matter(%) = [Al]/7.8 × 100 Figure 1. Location of the sediment trap mooring in this study (FS1: triangle). Circle, square and rhombus symbols indicate stations of sediment coring by Black and Buesseler,8 Otosaka and Kato7 and this study, respectively. Black and white symbols indicate locations in the northern and southern area of the mooring station, respectively.

(1)

Biogenic carbonate(%) = ([Ca] − 2.7/7.8 × [Al]) × 100/40

(2)

where, [Al] and [Ca] is concentration of aluminum (%) and calcium (%) in sinking particles, respectively. The loss of ignition method was used to determine the organic matter content. In this method, samples were heated in a muffle furnace at 500 °C for 24 h. When an adequate amount of the sample was obtained, measurements were taken in duplicate, and the precision was within 5% for all elements. As reference materials, JG-1a (granodiorite) and JLk-1 (lake sediment) from National Institute of Advanced Industrial Science and Technology, Japan were analyzed. The concentrations obtained using this method coincided with certified values within 6% for all elements. Concentrations of biogenic silicate (opal) were analyzed using a modified alkaline extraction method.22 The opal

Kogyo, SMD6000-13W) was installed at 873 m depth (119 m above the bottom). Time-series sampling was conducted from August 5, 2011 to June 23, 2013. Samples were collected for 26 periods with cup-opening periods of 26 days (SI Table S1). Due to the replacement of the mooring array, the sampling was interrupted from July 8 to 20, 2012. Sampling cups of the sediment trap were filled with 38‰ salinity-controlled, 5% formalin neutral buffer solution to prevent biological degradation. Seabed sediment was collected using a multiple corer at Sta. FS1 and seven surrounding stations (open and closed rhombuses in Figure 1). Sediment coring at Sta. FS1 was conducted three times. Core samples (80 mm in diameter; 12596

dx.doi.org/10.1021/es503736d | Environ. Sci. Technol. 2014, 48, 12595−12602

Environmental Science & Technology

Article

Figure 2. Temporal changes in (a) total mass flux and sea surface temperature, (b) component ratio, (c) 137Cs flux and concentration in sinking particles, and (d) 134Cs/137Cs activity ratio. Vertical broken lines indicate turn of the year. Activities are decay-corrected to March 11, 2011. For SST data, NOAA local area coverage data (MCSST) was obtained from the Agriculture, Forestry and Fisheries Research Information Center, Tsukuba, Japan.

sediment trap experiments in the western North Pacific,25,26 and can be regarded as a general trend in this region. In the area dominated by the Oyashio current, representative cold current in the western North Pacific, it is known that salinity stratification takes place from March to April which enables blooms of diatoms in the surface water.27 As shown in Figure 2(a) and (b), concentrations of biogenic opal in sinking particles increased in this season followed by peaks of total mass flux. This result indicates that the production of diatoms and subsequent sinking of biogenic opal is a potential cause to increase total mass flux from April to June. Radiocesium in Sinking Particles. Even just before the 1FNPP accident, significant amount of 137Cs mainly originating from global fallout occurred in seawater (∼2 mBq L−1)28 and seabed sediments (∼3 Bq kg−1)8 in the North Pacific. On the other hand, 134Cs (half-life: 2.06 years) was not detected because there was no recent deposition of 134Cs to the environment. Therefore, it might be appropriate to use 134Cs as an indicator of the accident-derived radiocesium in the environment. Nevertheless, concentrations of 134Cs in sinking particles are quite low or below the detection limit (SI Table

content was calculated from silicon concentration, extracted with 0.1 M sodium carbonate at 80 °C for 8 h, assuming that the relative water content in the biogenic opal was 10% (SiO2· 0.4H2O).24

■

RESULTS Mass Flux and Characteristics of Sinking Particles. Total mass fluxes observed with sediment trap ranged between 67 and 1791 mg m−2 day−1 (SI Table S1, Figure 2a) and showed three peaks a year. From August to September, a flux maximum, corresponding to the SST maximum, was observed. Sinking particles in this season had a high content of organic matter (Figure 2b), and production and rapid sinking of biogenic particles is presumed to increase mass fluxes in this season. Other flux peaks, observed from January to March and April to June in each year, did not correspond with the seasonal variation of SST. Remarkably high mass fluxes, exceeding 500 mg m−2 day−1, were observed from January to March. During this season, lithogenic matter was the dominant component of sinking particles. Such an increase of mass flux due to the input of lithogenic matter has also been observed by previous 12597

dx.doi.org/10.1021/es503736d | Environ. Sci. Technol. 2014, 48, 12595−12602

Environmental Science & Technology

Article

considered as key parameters to understand the transport mechanism of particulate radiocesium. The relationship between 137Cs concentration in sinking particles and total mass flux is shown in SI Figure S4. From the relationship, observed data can be classified into three groups. The first data group (group I), consists of data obtained in periods 1−6 and 15, is characterized by the high 137Cs concentration (>125 Bq kg−1, n = 7, p-value of Student’s t-test = 0.001). The second group (group II), consists of data obtained in periods 7−9, 11, 12, 21−24, is characterized by the high total mass flux (TMF > 450 mg m−2 day−1, n = 9, p = 0.001). The third group (group III) consists of data that are not classified into any group (characterized by low total mass flux and low 137 Cs concentration). Averaged total mass flux, 210Pbex flux and concentration of major components with the three data for each group are listed in SI Table S5. Sinking particles of the group “I” were characterized with high organic and low lithogenic contents, as well as high 137Cs concentrations. Sinking particles of the group “III” showed similar characteristics with those of the group “I” (SI Table S5). It can be considered that mass fluxes of the group “III” are regulated by a similar mechanism as the group “I”. The averaged 210Pbex fluxes for the group “I” and “III” were 0.56 ± 0.24 and 0.46 ± 0.38 Bq m−2 day−1, respectively. The deposition rate of 210Pb onto the surface water near the station is reported to be 0.54−0.55 Bq m−2 day−1.30,31 Applying distributions of 226Ra and 210Pb in the deep water32 into a onedimensional scavenging model, the approximate removal rate of 210 Pbex at the depth of sediment trap (873 m) is estimated to be 0.1 Bq m−2 day−1. Accordingly, assuming a steady-state condition, expected sinking flux of 210Pbex (0.64−0.65 Bq m−2 day−1) approximately agrees with 210Pbex fluxes of the groups “I” and “III”. From these results, groups “I” and “III” are hardly affected by lateral import of particles from neighboring continental slope or seabed areas. We therefore define these groups as the “moderate mode”. In contrast, the data group “II” was characterized by the high total mass flux attributed to high content of lithogenic matter. The averaged 210Pbex flux of this group was more than three times higher than those of the other groups. The high 210Pbex flux and high lithogenic content indicate that the lithogenic particles were imported over the trap layer by lateral transport or resuspension of sediment. It is reported that inventories of 210 Pbex in seabed sediment near the mooring site (36−38°N, 500−1200 m depth) is more than 10 kBq m−2.8 Assuming that residence time of 210Pb in sediment is 32 years (1/λ, λ = decay constant), the sedimentation rate of 210Pbex in this region is estimated to be more than 0.8 Bq m−2 day−1. This rate exceeds the expected 210Pbex flux from water column scavenging and atmospheric sources (0.64−0.65 Bq m−2 day−1), and shows evidence of enhanced boundary scavenging.33 In general, the dynamics of particles near the seabed is driven by wind, storm, and tidal currents.34 Several sediment trap experiments in the western North Pacific have reported that the effect of resuspension near the seabed is propagated to the offshore.26,35,36 The high 137Cs fluxes from January to March in each year (Figure 2(c)), correspond to the group “II”, would be affected by efficient lateral transport and the resuspension of 137 Cs-bound sediment in this season. We therefore define the group “II” as the “turbulence mode”. After the Chernobyl accident in 1986, the accident-derived radiocesium was transported to the deep layers associate with

S1). As described later, most of radiocesium detected from sinking particles in this study originated from the 1FNPP accident, and we used 137Cs as a representative radiocesium in the following discussion because it has lower detection limits and uncertainties. The 137Cs concentrations in sinking particles generally decreased with time with three temporary peaks in December 2011, August and November 2012 (Figure 2(c)). The highest concentration (412 ± 18 Bq kg−1) was observed in December 2011. 137Cs flux was highest in September 2011 (98 ± 3 mBq m−2 day−1), and generally decreased to 0.9 ± 0.2 mBq m−2 day−1 in June 2013 with fluctuations (Figure 2(c)). During the sampling period, the 137Cs flux showed seven peaks in AugustSeptember 2011, December 2011, February 2012, April−June 2012, August−September 2012, January−February 2013, and March−April 2013. The cumulative 137Cs flux throughout the sampling period was 22 Bq m−2 (SI Table S3). The 134Cs/137Cs activity ratio in sinking particles collected by this study (decay-corrected to March 11, 2011) was 1.01 ± 0.06 (Figure 2(d)), and agreed well with the signature of 1FNPPderived radiocesium (1.0).29 This indicates that most of radiocesium in sinking particles of this study originated from 1FNPP. Seabed Sediments. Vertical changes in 134Cs concentration and 134Cs/137Cs ratio are shown in SI Figure S2. 134Cs penetrated to the 1−2 cm sedimentary layer, and was not detected below the 3 cm layer. Higher concentrations were observed in November 2011. The 134Cs/137Cs ratio in sediment was 0.92 ± 0.12 at the surface (0−1 cm layer) and 0.72 ± 0.16 in the 1−2 cm layer. This vertical change in 134Cs/137Cs ratio did not vary with time significantly. The lower 134Cs/137Cs ratio in the subsurface sediment than the 1FNPP signature (1.0) indicates that part of 137Cs in this layer was derived from global fallout. Accordingly, in contrast to radiocesium in sinking particles, it would be appropriate to apply 134Cs data in assessing the inventory and distribution of the accident-derived radiocesium in the sediment. We therefore define the concentration of the accident-derived 137Cs (137Csacc) in sediment by eq 3, and ensure the comparability with 137Cs data of sinking particles. Equation 3 is also applied for 137Cs in seawater. 137

Csacc =

134

Cs × 1.0

(3)

The inventories of 137Csacc (Bq m−2) in seabed of the mooring station in August 2011, October 2011, and July 2012 were 47 ± 5, 81 ± 10 and 46 ± 6 Bq m−2, respectively (SI Table S3). The 137Csacc inventory seemed to change with time, but the variability (±34%) is larger than that of 210Pbex (±10%). The variation of 137Csacc in sediment would indicate a lack of uniformity of accident-derived radionuclides in sediment rather than a seasonal variation. Even the uniformity is considered, the inventory of 137Csacc is much larger than the 137Cs flux cumulative from August 2011 to June 2013 (22 Bq m−2). This result indicates that the 137Cs flux after August 2011 hardly changed the 137Csacc inventory in the sediment, and the distribution of radiocesium to the sediment established before the starting date of the sediment trap experiment.

■

DISCUSSION Two Modes Controlling Fluxes of Particulate Radiocesium. Since the vertical flux of radiocesium is a product of total mass flux and radiocesium concentration, these two can be 12598

dx.doi.org/10.1021/es503736d | Environ. Sci. Technol. 2014, 48, 12595−12602

Environmental Science & Technology

Article

rapid sinking of particles from the surface, followed by a lateral transport from continental shelves with a time lag of several months.17 Such dynamics of particulate radiocesium is consistent with those observed near Fukushima. Moderate Mode: Removal of the Accident-Derived Radiocesium from Surface Water. Concentration of radiocesium in sinking particles varies with adsorption/ desorption through the seawater−particle interface or incorporation into biogenic particles. Since these processes mainly take place in the surface waters, concentration of radiocesium in the surface seawater is regarded as a key factor controlling radiocesium concentration in sinking particles. In addition, in the moderate mode, the removal rate of radiocesium-bound particles from the surface would be another key factor regulating fluxes of particulate radiocesium to the deep water. In order to assess the particulate flux of radiocesium removed from the surface, the following eq 4 is defined. FCs(mBq m−2day −1) = Cw × K * × Fv

(4)

where, Cw, K*, and Fv are concentration of radiocesium in the surface seawater (Bq L−1), a factor converting concentration of radiocesium in seawater to that in sinking particles (L g−1), and mass flux of particles directly transported from the surface layer (mg m−2 day−1), respectively. Although the K* value depends on kinetics of radiocesium uptake/release and particle concentration, it would be regarded as a constant because the variation is smaller than those of Cw and Fv. Figure 3(a) shows the temporal variation of 137 Csacc concentration in the surface (