1. Radiocesium (134Cs and 137Cs) derived from the damaged FNPP1 caused radioactive contamination of the islands of Japan and the North Pacific Ocean2. Most of the Fukushima-derived radiocesium deposited on land has remained in soils. Within about 100 km of the FNPP1, where contamination was serious, the radiocesium in soils has been measured intensively3. The decay-corrected ratio of 134Cs/137Cs in soils has been calculated to be 1.0, which suggests that the total amounts of 134Cs and 137Cs released from FNPP1 were equivalent. The relationship between the radiocesium activity in the soil and the air dose rate derived from airborne monitoring has provided a map of the density of radiocesium deposition throughout the islands of Japan4. The sum of the deposition, the total inventory of 137Cs (or 134Cs) on the islands of Japan, has been estimated to be 2.4 PBq5. However, the total amount of Fukushima-derived radiocesium in the North Pacific remains uncertain, because it has been difficult to obtain sufficient samples of water, especially from subsurface and deep waters, in the vast North Pacific Ocean, except from the coastal area near the FNPP16, 7, 8.
Radiocesium isotopes were released into the North Pacific through two major pathways, direct discharges of radioactive water and atmospheric deposition. About ten days after the earthquake, Tokyo Electric Power Company and the Ministry of Education, Culture, Sports, Science and Technology of Japan (MEXT) began marine monitoring in the coastal area within about 50 km from the FNPP16, 7, 8. These high-frequency measurements have facilitated an evaluation of the total amount of radiocesium derived from the directly discharged radioactive water. The values estimated in several studies were in the range 4–6 PBq1, 9, 10, 11, 12, 13, although one study calculated the value to be 27 PBq (12–41 PBq)14. The total direct release of 27 PBq was somewhat of an overestimate11, 15 and resulted in activities in a model ocean that were unrealistically high compared to activities measured in the real ocean16. However, radiocesium activities measured during a cruise in June 2011, mainly in the open ocean17, indicated that the total activity of 137Cs (or 134Cs) directly discharged to the ocean equaled 11–16 PBq18, 19.
A large portion of the radiocesium released to the atmosphere from the FNPP1 was deposited onto the North Pacific Ocean, because the winds over Japan usually blow from the west in the spring20. However, the small number of observational data in the open ocean cannot estimate the total oceanic deposition directly. Alternatively, that could be calculated indirectly from the total amount of radiocesium released to the atmosphere, which was derived primarily from measurements on land. Estimations of the total amount released to the atmosphere range widely, from 8.8 to 37 PBq1, 5, 9, 11, 14, 21, 22, 23, 24, 25. The 2.4 PBq deposited onto the islands of Japan suggests that most of the remaining radiocesium, 6.4–35 PBq, found its way into the North Pacific through atmospheric deposition. Atmospheric models have estimated independently the total oceanic deposition to be 5.8–30 PBq5, 9, 11, 12, 23, 25, similar to the range of 6.4–35 PBq. However, the deposition on land has been overestimated in many of the models.
Efforts to obtain observational data from the open ocean have continued. The marine monitoring from March 2011 by MEXT or the Nuclear Regulation Authority was extended eastward to the 144°E meridian in August 20117. Radiocesium measurements in the area further east have been reported in several publications8, 17, 26, 27, 28, 29, 30, 31. Seawater sampling from April 2011 during commercial ship cruises has produced a valuable dataset across the North Pacific28, although as in many other previous studies, most of the samples were collected only at the surface. In June 2011 vertical profiles of the Fukushima-derived radiocesium were measured at stations along 147°E between 34.5°N and 38°N, and it was found that the radiocesium had penetrated to a depth of about 200 m roughly two months after the disaster17. Although these observational data are still insufficient for direct estimation of the total amount of radiocesium in the whole North Pacific, these data can be used to validate ocean model simulations that have predicted vertical and horizontal spreading of the radiocesium in the ocean13, 15, 16, 25, 32, 33.
Here we report the vertical distributions of the Fukushima-derived radiocesium at stations along 149°E between 10°N and 42°N in the winter of 2012, about ten months after the accident. Our preliminary reports, which have already been published31, 34, revealed that (1) the Fukushima-derived radiocesium activity was highest in the transition area between the subarctic and subtropical regions and (2) the radiocesium was transported southward across the Kuroshio Extension (KE) through subsurface layers. In this study, we discuss the causes of the southward spreading of the radiocesium based on temporal changes in the activity of surface waters. Secondly, we have estimated the vertical water-column inventory of radiocesium. These results will contribute to determination of the total inventory of radiocesium and will facilitate prediction of the spreading of the Fukushima-derived radiocesium in the North Pacific Ocean in the future. We measured both 134Cs and 137Cs activities (Methods). The ratio of decay-corrected 134Cs/137Cs in samples in which the 137Cs activity was higher than 20 Bq m−3 was about 0.95. The small excess of 137Cs was derived from another source of 137Cs, global fallout due to the nuclear bomb testing in the 1950s and 1960s35. The excess 137Cs in surface waters (about 1.5 Bq m−3) in the winter of 2012 corresponds to bomb-produced 137Cs activities (about 1.9 Bq m−3) in surface water of the North Pacific before the accident (about 2.4 Bq m−3 in 2000)36. Therefore, only results for 134Cs, which is a unique tracer of the FNPP1 accident, are presented in later sections.
Our sampling stations were located in the western North Pacific from cold subarctic to warm tropical regions, although information on sea surface temperatures estimated by satellite sensors was patchy in the northern area due to cloudy conditions during the sampling cruise (Figure 1a). The image of sea surface height (SSH) implied that our observational line along 149°E crossed eastward-flowing currents around 35°N and 40°N where SSH gradient was relatively steep (Figure 1b). The northern and southern currents correspond to the subarctic and KE fronts, respectively. Here we define areas north of the subarctic and south of the KE fronts as the subarctic and subtropical regions, respectively. In addition, we designate the area between the two fronts as the transition area, in which the FNPP1 is situated (Figure 1). Although a boundary between the subtropical and tropical regions is not clear in Figure 1, we provisionally regarded the area south of 20°N as the tropical region because of the subtropical front around 20°N37. The distribution of SSH also suggests that the observational line crossed a southward meander of the KE front around 148°E (A in Figure 1b).
To the south of the KE, the surface activity was less than a few Bq m−3 in the winter of 2012 (Figure 2). Figure 3a indicates that the 134Cs activity was also low (but significantly above the detection limit) in the surface mixed layer from the surface to a depth of 150–200 m between approximately 25°N and 35°N. In contrast, to the south of 20°N the activity was not detected in the surface mixed layer to a depth of 100–150 m, except in surface waters collected with a bucket. Below the surface mixed layer, we found a conspicuous subsurface maximum centered at a depth of about 300 m throughout the subtropical region between 20°N and 35°N. This subsurface tongue-shaped maximum appeared in a pycnostad between potential density anomalies of approximately 25.0 and 25.6 σθ (Figure 3d), which corresponds to water temperatures of 15–18 °C (Figure 3b) and salinities of approximately 34.60–34.75 (Figure 3c). The pycnostad resulted in a subsurface minimum of potential vorticity (Figure 3e). Higher activities in the subsurface maximum were observed at 32°N and 34°N (10–20 Bq m−3), and the activity decreased at lower latitudes. We also note that the 134Cs had penetrated into deeper layers, to depths of at least 600 m, between 32°N and 35°N.
We calculated vertically integrated (i.e., areal) 134Cs inventories from the surface to a depth of 800 m in the winter of 2012 (Figure 4). The areal inventories were corrected for radioactive decay to the date of the earthquake, 11 March 2011. High areal inventories were observed in the transition area, where surface activities were also high. Although the surface activities were low in the subtropical region between 30°N and 35°N, the areal inventories were comparable to those in the transition area because of the subsurface activity maximum. The areal inventories of 134Cs activity in the subarctic region (40°N–42°N), transition area (35°N–40°N), and subtropical region (20°N–35°N) were calculated to be 0.8 ± 0.1, 4.6 ± 0.3, and 1.6 ± 0.1 kBq m−2, respectively, where the error bounds indicate standard deviations. We compared the areal inventories in the winter of 2012 with those calculated about 8 month earlier, in June 201117 (Figure 4). The areal inventory in the transition area (36°N–38°N) in June 2011, 7.9 ± 0.3 kBq m−2, implies about a 40% decrease in the areal inventory between June 2011 and the winter of 2012, although the spatial variation in June 2011 was larger than in the winter of 2012. The mean of the decay-corrected radioactivity in the surface water also decreased by about 70%, from 73 to 21 Bq m−3, in the transition area during the same period. The higher rate of decline in the surface 134Cs radioactivity was caused by its deeper penetration during the winter of 2012 (to a depth of about 300 m) than in June 2011 (to a depth of about 200 m). A relatively large areal inventory at the southernmost station (36°N) to the south of the KE in June 2011 was caused by a subsurface 134Cs maximum at depths of 150–450 m.
Figure 2). In April 2011, 134Cs activity was also observed at stations in the subarctic and subtropical regions, more than 1000 km distant from the plant26, 28. The wide dispersal of Fukushima-derived 134Cs in the western North Pacific within about two months of the accident is consistent with patterns of atmospheric deposition of 134Cs simulated by atmospheric models13, 25, 38. A low-pressure system traveling across Japan from 14–15 March 2011 was found to be effective in lifting particles containing 134Cs from the surface layer to the altitude of the westerly jet stream, which carried the particles across the North Pacific within 3–4 days39.
In the transition area between 35°N and 40°N, the 134Cs activities in surface waters during June–August 2011 were significantly higher than in April–May 2011 (Figure 2), which implied that contaminated waters discharged from the FNPP1 had been transported by the eastward-flowing North Pacific Current (Figure 5). The radiocesium activities in surface seawater collected by commercial cruise ships revealed an eastward propagation of the main plume of the directly discharged 134Cs. The zonal speed of the plume was estimated to be about 200 km month−1, a speed that was consistent with trajectories of Argo floats launched near the FNPP128. Therefore, arrival of the directly discharged 134Cs water in June–August 2011 was delayed by about two months relative to the atmospheric deposition in April–May 2011. The activity decrease in September–December 2011 indicated that the main body of the plume had passed to the east between April–May and September–December 2011. The radiocesium, however, also had spread vertically and penetrated deeper in the winter of 2012 (a depth of about 300 m) compared to June 2011 (a depth of about 200 m).