Modeling studies on the atmospheric diffusion and deposition of the radiocesium associated with the Fukushima Dai-ichi Nuclear Power Plant accident is reviewed here, with a focus on a research collaboration between l’Institut de Radioprotection et de Sûreté Nucléaire (IRSN)­—the French institute in charge of evaluating the consequences of nuclear accidents and advising authorities in case of a crisis—and the Meteorological Research Institute (MRI) of the Japan Meteorological Agency—an operational weather forecasting center in Japan. While the modelers have come to know that wet deposition is one of the key processes, the size of its influence is unknown. They also know that the simulation results vary, but they do not know exactly why. Under the research collaboration, we aimed to understand the atmospheric processes, especially wet deposition, and to quantify the uncertainties of each component of our simulation using various numerical techniques, such as ensemble simulations, data assimilation, elemental process modeling, and inverse modeling. The outcomes of these collaborative research topics are presented in this paper. We also discuss the future directions of atmospheric modeling studies: data assimilation using the high temporal and spatial resolution surface concentration measurement data, and consideration of aerosol properties such as size and hygroscopicity into wet and dry deposition schemes.

Using an hourly-resolution time series of the Fukushima radionuclides collected on used filter-tapes installed in suspended particulate matter (SPM) monitors, we measured the hourly radiocesium values at the SPM monitoring sites of Futaba 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 the many monitoring posts/points of Fukushima Prefecture and the Tokyo Electric Power Company. At Futaba, nine plumes of high ¹³⁷Cs concentrations were found on March 12–13, 15–16, 18–20, and 24–25, 2011, when southeasterly winds prevailed. 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 at 15:36 JST. On the afternoon of March 15, plumes from the FD1NPP were observed at Futaba due to a constant southeasterly wind and were then transported to downwind, resulting in the formation of a highly deposited zone of radionuclides spanning more than 30 km from near the FD1NPP to the northwest. In contrast, seven plumes of high ¹³⁷Cs concentrations were found at Naraha on March 15–16, 18, 20–21, 2011, when northerly winds prevailed. On March 15, a plume caused by 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, was observed at 3:00 JST, and then were transported southward to the Kantou area. The activity ratios of ¹³⁴Cs/¹³⁷Cs in the plumes were divided into two groups. The plumes at Futaba on March 12–13, which had ratios of 0.92–0.94, are identified to be released from U1, compared to its ratio of 0.94, as derived from the inventory data. All other plumes with the ratios of 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.

Structures and distribution of constituent elements of non-spherical radioactive microparticles collected in the Fukushima Prefecture were investigated mainly using scanning transmission electron microscopy (STEM) with energy-dispersive X-ray spectroscopy of high detection efficiency. An angulated radioactive microparticle collected as an aerosol in 2015 had a chemical composition similar to that of the spherical microparticles collected in 2011 but with a different distribution of Sn and alkali ions, which may have been caused by partial dissolution of the microparticle in the field. Other non-spherical microparticles collected from the atmosphere in 2013 contained Al as a constituent of silicate glass, which was not detected in the spherical microparticles. Chromium-oxide glass containing Fe, Zn and Sn was also present, coexisting intricately with silicate glass in one of the microparticles. Finally, a microparticle collected in 2015 from plant tissue was substantially an aggregate of fine particulates composed of Cs-bearing silicate glass that did not contain either Fe or Zn, which were common in the silicate glass forming the other radioactive microparticles. These results suggest that the radioactive microparticles emitted from the Fukushima Daiichi nuclear power plant are of a considerable variety and that their structures likely change with time in the field.

Two types of radioactive particles were isolated from environmental samples collected at various distances from the Fukushima Dai-ichi Nuclear Power Station. “Type A” particles are 2–10 μm in diameter and display characteristic Cs X-ray emissions when analyzed using energy-dispersive X-ray spectrometry (EDS). “Type B” particles are considerably larger, up to 400 μm in diameter, with Cs concentrations too low to be detectable with EDS. These larger particles were isolated from the region north of the nuclear reactor site, which was contaminated on March 12, 2011. The specific activity of Type B particles is much lower than Type A, and the mean ¹³⁴Cs/¹³⁷Cs ratios are ~0.93 and 1.04, respectively. The Type B ratio indicates that power station Unit 1 is the source, implying that these larger radioactive particles were discharged on March 12. This study found that different type of radioactive particles were released not only on March 15 but also on March 12.

We found four radiocesium-bearing microparticles (CsMPs) with high cesium (Cs) radioactivity in suspended particles collected from the Kuchibuto River in Fukushima by filtering water during 2011–2016. The CsMPs were identified by autoradiography and subsequently were separated from other suspended particles by the “multiple wet separation method” using a NaI scintillation counter. The present four CsMPs contained 0.426–2.827 Bq of ¹³⁷Cs and had a chemical composition similar to that of particles released during the accident at the Fukushima Daiichi Nuclear Power Plant (FDNPP). The activity ratio of Cs (¹³⁴Cs/¹³⁷Cs) suggests that they originated from Unit 2 or 3 of the FDNPP. The ratio of the radioactivity of the separated CsMPs to the total radiocesium on the filters ranged from 0 to 46%. Moreover, we calculated the radioactivities of CsMPs with lower radioactivity (0.1 to 0.4 Bq ¹³⁷Cs) determined by autoradiography. When smaller CsMPs were included, the ratio of the radioactivity of the CsMPs to the total radioactivity of radiocesium on the filters ranged from 1.3 to 67%. It has been previously suggested that the solid-water distribution coefficient (K_d) of radiocesium in rivers is apparently increased due to the possible presence of CsMPs in the solid phase because the water solubility of radiocesium in CsMPs is small. However, this study reveals that higher K_d values in rivers in Fukushima compared with those in Chernobyl cannot be explained by the contribution of CsMPs alone. The temporal variation of the ratio of radioactivity of CsMPs to the total radioactivity of radiocesium in river water after the FDNPP accident is also discussed.

Rainwater has been sampled on a monthly basis at the University of Tsukuba, Japan, since 2009. The serious accident at the Fukushima Daiichi Nuclear Power Plant in March 2011 resulted in the atmospheric emission of large quantities of radionuclides. Monthly ¹²⁹I deposition was monitored for 19 months before and 41 months after the accident to estimate its influence at Tsukuba. Before the accident, ¹²⁹I concentration in rainwater was in the range (0.19–2.6) × 10⁸ atoms L^–1. There was a rapid increase after the accident, with the highest ¹²⁹I concentration recorded in March 2011 as (5.4 ± 0.8) × 10¹⁰ atoms L^–1. It took approximately 1 y for the ¹²⁹I concentration to return to pre-accident background levels, due to continuing resuspension and deposition. Measured decreases in ¹²⁹I concentration in rainfall showed good agreement with double-exponential modeling.

In this study, the potential natural sources of secondary radiocesium isotope (¹³⁴Cs and ¹³⁷Cs) emissions were investigated, with a focus on n-alkanes, a characteristic bioaerosol compound. Monitoring was performed to obtain a time series of aerosol, samples, from winter 2013 to summer 2014, and size-resolved aerosol samples in 2012 and 2014. Samples were collected from the area heavily contaminated after the Fukushima Daiichi Nuclear Power Plant accident in March 2011. A correlation analysis of radiocesium, n-alkanes, and black carbon concentrations was performed to identify the contributions of aerosols from biogenic and anthropogenic sources. Biogenic n-alkanes exhibited similar concentration ranges except for spring 2014. The continuous input of biogenic n-alkanes is characteristic of a sampling site surrounded by forest, where pollen dispersion increased the concentration of biogenic n-alkanes in spring 2014. On the other hand, anthropogenic n-alkane concentrations were significantly increased in spring and summer 2014 (>50 ng/m³), compared with those prior to winter 2014 (<20 ng/m³). This anthropogenic n-alkane increase represents the beginning of reconstruction near the area. The carbon preference index (CPI) clearly showed biogenic n-alkanes with coarse-sized particles (CPI > 3), and more anthropogenic n-alkanes were contained in fine particle aerosols. Our results showed that radiocesium and biogenic n-alkane concentrations in seasonal and size-resolved aerosol samples have a partially positive correlation, which supports the hypothesis that the secondary emissions of radionuclides occurred in the forested areas.

Particle size and mineralogy are important factors controlling the distribution of radiocesium in soils and sediments contaminated by nuclear weapon testing and nuclear power plant accidents. However, it is often difficult to distinguish the influence of particle size and mineralogical composition on the size distribution of radiocesium because they are closely related. The objective of this study was to elucidate the influence of mineralogical composition on the distribution of stable Cs and radiocesium in river sediments. We analyzed size-fractioned samples of riverbed sediments collected at two sites, a pasture and Kuroiwa in the Abukuma River system in Fukushima after the Fukushima Daiichi Nuclear Power Plant (FDNPP) accident. The size distributions of K, Rb, and ¹³³Cs reflected the mineralogy of sediments, where primary host minerals for these alkali elements would be biotite, K-feldspar, and clay minerals. Silt-size fractions contained high ¹³³Cs and ¹³⁷Cs concentrations possibly due to adsorption on clay minerals. Their concentrations decreased with increasing particle size at the pasture site. In contrast, coarse and very coarse sand fractions from the Kuroiwa site showed higher ¹³³Cs and ¹³⁷Cs concentrations in comparison to fine-medium sand fractions. The coarse sand fractions contained many weathered biotite grains, whereas hornblende was the major constituent in the fine-medium sand fractions. Overall, the size distributions of ¹³³Cs and ¹³⁷Cs were similar in the sediments, suggesting that the FDNPP-derived radiocesium was distributed into each particle size fraction in response to the distribution of the stable Cs that was controlled by mineralogical composition. Leaching experiments using 1 M CH₃COONH₄ indicated that ¹³⁷Cs was fixed more strongly in silt-size fractions than in sand-size fractions. However, it should be noted that sand-size fractions retained more than 90% of the ¹³⁷Cs in the leaching experiments. Illite and weathered biotite are likely the hosts of ¹³⁷Cs in silt- and sand-size fractions, respectively. In this study, we demonstrated the strong retention of the FDNPP-derived radiocesium in wide range of particle size fractions of sediments.

To understand the dissolutive behaviour of radio Cs discharged to the ocean environment as a result of the Fukushima Dai-ichi Nuclear Power Plant accident, an aerosol sample collected on the 15th of March 2011 at Kawasaki City (Kanagawa) was sequentially leached with seawater for 30 days. In addition, a surface soil sample collected from Kawamata Town (Fukushima) two months after the accident, was leached for three days with natural seawater, diluted seawaters and ultrapure water to observe the effect of the ionic strength of the waters on the respective leaching ratios and apparent distribution coefficient (K_d) values. Furthermore, the soil sample was subjected to a 223-day continuous sequential leaching with a natural seawater and with a 1:1 mixture of ultrapure water and seawater. When leaching the aerosol sample in seawater, about 40% of the total ¹³⁷Cs was extracted in the first three days, and a further 20% of the total ¹³⁷Cs was extracted within 30 days. Lower K_d values for ¹³⁷Cs between the soil and leachates were obtained with seawater and diluted seawater compared to ultrapure water. For the long-term experiment (223 days) using the three leaching solutions, approximately 0.1–2% of the total ¹³⁷Cs was leached in the first three days. Eventually, more than 15% of total ¹³⁷Cs in the surface soil sample was efficiently desorbed by seawater leaching. In comparison, about 9% of the total ¹³⁷Cs was leached with 1:1 diluted seawater and less than 1% of the total ¹³⁷Cs was leached with ultrapure water over the 223 days. In general, there were some similarities between the leaching behaviour for natural ¹³³Cs and radio Cs. In the surface soil, radio Cs species was eventually incorporated into the clays after undergoing solubilisation as fallout aerosols in natural waters. Thereafter, the insoluble or less soluble forms of radio Cs in the soil would be partially extracted by seawater after the transport of contaminated surface soils to the ocean via rivers.

To elucidate the activity concentration and behavior of radiocaesium, we observed the spatial and vertical distributions of radiocaesium in sediment collected at 12 monitoring stations off the coast of Fukushima Prefecture in 2013–2015. In bulk surface-layer sediment (core depth: 0–3 cm), high ¹³⁴Cs activity concentrations were observed at stations around the water depth of 100 m, where the sediment was rich in silt to clay particles and organic matter. The activity concentrations generally decreased with increasing core depth and the extent of the decrease was different between the stations. The difference trend for temporal change of ¹³⁴Cs activity concentrations between the two zones of off Onahama and within 30 km of the FDNPS was partly attributed to the mobility of sediment particles reflecting water content, porosity and permeability. At some stations, shaped peaks for activity concentrations were temporarily observed in upper-layer sediment (core depth: 0–1 cm) or sediment below that. The ¹³⁴Cs activity concentrations in the surface-layer sediment were not always accompanied by an increase in the contribution from fine (i.e., silt to clay) particle-bound ¹³⁴Cs in the sediment. In October 2014, sediment collected at a station about 1.5 km from FDNPS was found to have broad ¹³⁴Cs peaks in the middle-layer sediment (core depth: 5–16 cm). In this middle-layer sediment, both sand and silt to clay fractions contributed to the increased ¹³⁴Cs activity concentrations. On the other hand, such broad peaks were not found in October 2015. These results suggest that the activity concentrations in sediment had changed significantly by a complicated process of sediment mixing, erosion or re-sedimentation that affected the broad peak for the activity concentration in the middle-layer sediment.

In this work we report tritium concentrations determined in seawater samples collected offshore of Fukushima in the northwestern Pacific Ocean, immediately after the Fukushima Dai-ichi Nuclear Power Plant (F1-NPP) accident. We found surface seawater to have high concentrations of tritium (³H) and cesium-137 (¹³⁷Cs) with respect to the concentrations expected for the investigated region. Tritium concentrations were up to six times the pre-accident level. However, these concentrations were found to be relatively low compared to those of ¹³⁷Cs and ¹²⁹I. This is most likely because of the very low production ratio of tritium to ¹³⁷Cs in the F1-NPP, and the inherently high background concentration of tritium in these environmental waters (mainly ascribed to global fallout from nuclear weapons tests conducted during the 1960s). The tritium distribution in surface seawater showed higher concentrations close to the shore and lower concentrations offshore. Higher tritium concentration areas had spread both northward and southward from the F1-NPP along the coast, indicating that large amounts of tritium were carried by coastal currents. A positive correlation was found to exist between ³H and ¹³⁷Cs concentrations in the seawater. The calculated ³H/¹³⁷Cs ratio was similar to the production ratio of these isotopes reported for the broken reactors in the F1-NPP, which indicates that both radionuclides might have originated from the F1-NPP. Direct emission of tritium to the ocean from the F1-NPP was estimated to be approximately 0.05 PBq immediately after the accident, which is much smaller than the total inventory in the environment.

The ¹³⁵Cs/¹³⁷Cs atom ratio has been proved to be a reliable tracer for radiocesium source identification in the studies on the Fukushima Daiichi Nuclear Power Plant (FDNPP) accident. However, due to the technical challenge to measure ¹³⁵Cs, no ¹³⁵Cs data are available for Japanese river or ocean sediments. In the present study, the vertical distributions of ¹³⁴Cs and ¹³⁷Cs activities and their ratios in two marine sediment cores, collected offshore from the FDNPP site immediately after the accident, were measured by γ spectrometry. A conventional introduction system was replaced by an APEX-Q system, by which it was possible to get 6.5 times higher Cs intensities and then ¹³⁵Cs was analyzed by triple-quadrupole inductively coupled plasma-mass spectrometry. From the vertical distributions, it was seen that the deposition of ¹³⁴Cs and ¹³⁷Cs presented an increasing trend, which indicated continuous radiocesium input into the ocean and deposition from sea water onto the sea floor up to the collection date. Therefore, the ¹³⁴Cs/¹³⁷Cs activity ratios (0.866–1.16) and ¹³⁵Cs/¹³⁷Cs atom ratios (0.249–0.343) (all decay-corrected to March 11, 2011) in marine sediment core samples showed the fingerprints of radiocesium in the oceanic environment were mainly derived from the FDNPP.

We report new results for Cu²⁺ adsorption on δ-MnO₂ and the resulting changes in Cu isotopic composition due to isotopic fractionation. The adsorption experiments were designed as a pH series (3.22–6.94) and an adsorption-desorption series. ⁶³Cu adsorbs preferentially on δ-MnO₂, and the degree of isotopic fractionation is 0.45 ± 0.18‰ (2σ, n = 12). Preferential desorption of ⁶⁵Cu from δ-MnO₂ was also observed. The isotopic data are consistent with a closed isotopic equilibrium model, and the degree of fractionation indicates no systematic variation with pH or reaction time. Previously reported extended X-ray absorption fine structure study predicted the preferential adsorption of ⁶⁵Cu on δ-MnO₂, because of the formation of Cu²⁺ surface complexes (with 3–4 ligands) from dissolved Cu²⁺ (with 5 ligands); however, the opposite was observed in this study. The equilibrium in isotopic fractionation is affected by the bonding environment; hence, we suggest that dissolved Cu²⁺ species have a stronger bond than adsorbed ones.

To evaluate post-depositional loss from the snow surface and subsequent redistribution of nitrate, we determined the spatial variation of nitrate concentrations, as well as δ¹⁵N and Δ¹⁷O values of nitrate in surface snow, sampled along a latitudinal transect in eastern Dronning Maud Land, East Antarctica. The NO₃^– concentrations of surface snow ranged from 40.0 to 130.8 μg L^–1, showing no obvious trends with latitude, while the δ¹⁵N(NO₃^–) in surface snow increased from coastal sites to inland sites, ranging from –19.4 to 165.5‰. The relationship between the isotopic values (δ¹⁵N and Δ¹⁷O) of nitrate and snow accumulation rate are consistent with other traverses studied in East Antarctica (e.g., from Dumont d’Urville station to Vostok, and from East Antarctic coast to Dome Argus), implying that post-depositional loss and redistribution occur similarly throughout East Antarctica.