Cesium-134 and 137 activities in the central North Pacific Ocean after the Fukushima Dai-ichi Nuclear Power Plant accident

Surface seawater 134Cs and137Cs samples were collected in the central and western North Pacific Ocean during the 2 yr after the Fukushima Dai-ichi Nuclear Power Plant accident to monitor dispersion patterns of these radioisotopes towards the Hawaiian Islands. In the absence of other recent sources and due to its short half-life, only those parts of the Pacific Ocean would have detectable 134Cs values that were impacted by Fukushima releases. Between March and May 2011, 134Cs was not detected around the Hawaiian Islands and Guam. Here, most 137Cs activities (1.2–1.5 Bqm−3) were in the range of expected preexisting levels. Some samples north of the Hawaiian Islands (1.6– 1.8 Bqm−3) were elevated above the 23-month baseline established in surface seawater in Hawaii indicating that those might carry atmospheric fallout. The 23-month time-series analysis of surface seawater from Hawaii did not reveal any seasonal variability or trends, with an average activity of 1.46± 0.06 Bqm−3 (Station Aloha, 18 values). In contrast, samples collected between Japan and Hawaii contained 134Cs activities in the range of 1–4 Bqm −3, and137Cs levels were about 2–3 times above the preexisting activities. We found that the southern boundary of the Kuroshio and Kuroshio extension currents represented a boundary for radiation dispersion with higher activities detected within and north of the major currents. The radiation plume has not been detected over the past 2 yr at the main Hawaiian Islands due to the transport patterns across the Kuroshio and Kuroshio extension currents.


Introduction
The Tōhoku earthquake and subsequent tsunami on 11 March 2011 led to damages at the Fukushima Daiichi Nuclear Power Plant (F1-NPP) on the east coast of Japan.Significant amounts of radionuclides were released to the atmosphere and, by direct discharge or leakage, to the ocean.From these, 134 Cs (half-life 2.07 a) and 137 Cs (halflife 30.08 a) are important F1-NPP-derived radionuclides in the ocean because of their radioactive-conservative behavior and large discharged quantities.It is estimated that up to several tens of PBq (PBq = 10 15 Bq) of 134 Cs and 137 Cs entered the environment (e.g., IAEA, 2011;Bailly du Bois et al., 2012;Estournel et al., 2012;Stohl et al., 2012;Charette et al., 2013;Rypina et al., 2013).
Radioactive isotopes were released to the atmosphere from 12 March 2011 onward with a peak on 15 March (e.g., Stohl et al., 2012).Within days, accident-related radionuclides were registered by atmospheric monitoring stations across the Northern Hemisphere (e.g., Stohl et al., 2012).On the Pacific Islands (e.g., Hawaii, Guam, Commonwealth of the Northern Mariana Islands), an increase in gross beta activity in aerosols and trace amounts of radioactive iodine, cesium, and tellurium on air filters were identified between 19 and 23 March 2011 (EPA, 2011).
In mid-May 2011 activities in excess of 10 Bq m −3 of 134 Cs were found in the surface ocean in two isolated areas within the North Pacific (longitude 180 • W and 150 • W) indicating atmospheric fallout from the Fukushima accident (Aoyama et al., 2013a).Uniformly distributed 134 Cs within a 50-100 m thick mixed layer with activity of 10 Bq m −3 represents fallout of 500-1000 Bq m −2 .Although representative of only a very small area from the Pacific, it would be orders of magnitude higher than wet deposition of 0.47-180 Bq m −2 of 134 Cs measured on the North American continent (Wetherbee et al., 2012).Based on available data, these are the highest observed inventories and deposition rates in the far-field Pacific outside of the immediate vicinity of Fukushima.
Releases to the ocean peaked during the first two weeks of April 2011 and continued for months after the accident (Buesseler et al., 2011;Kanda, 2013).Numerous sampling efforts near and off the coast of Fukushima were conducted in order to quantify the released amounts and to study dispersion of radiation in the Pacific Ocean (e.g., Aoyama et al., 2013a, b;Buesseler et al., 2012;Honda et al., 2012;Kaeriyama et al., 2013).Several simulations of 137 Cs deposition and dispersion in the North Pacific were published (e.g., Behrens et al., 2012;Buesseler et al., 2012;Dietze and Kriest, 2012;Nakano and Povinec, 2012;Tsumune et al., 2012), but these were, among other parameters, limited by uncertainties of the source term of radioactivity releases.
In order to define the source term better and validate model predictions of dispersion and deposition patterns, direct observations need to be performed across the Pacific.Within this goal our efforts focused on the central Pacific Ocean around the Hawaiian Islands and between Hawaii and Japan.Our objectives were to detect any atmospheric fallout in the surface ocean by sampling immediately after the accident (March-May, 2011) and to monitor dispersion patterns of cesium towards and around Hawaii.

Experimental
Surface seawater samples (20 to 100 L) were collected on several scientific cruises and ships of opportunity between March 2011 and February 2013 (Fig. 1 All seawater samples were filtered using Micro-Wynd II cartridge (pore size 1 µm).The retention of 134 Cs and 137 Cs on these cartridges was negligible (< 0.1 %) due to high cesium solubility in seawater (Buesseler et al., 2012).Depending on the following chemical separation procedures, the samples were either acidified to pH 1 with nitric acid or were left untreated.Stable cesium tracer was added to all samples (0.04 mg L −1 ) for chemical recovery determination.
The 100 L samples were fed through 25 mL of ion exchanger at flow rates of 250-300 mL min −1 as described earlier (Kameník et al., 2013).Small-volume samples (16-25 L) were passed through 5 or 10 mL of ion exchanger at maximum flow rates of 30 and 100 mL min −1 , respectively.The ion exchanger was then dried and transferred to polyethylene containers for gamma counting.Gamma spectrometric measurements were performed using a coaxial HPGe detector at the University of Hawaii (relative efficiency 43 %, resolution 1.76 keV for 1.33 MeV gamma line of 60 Co).The gamma-ray spectra were evaluated using Hypermet-PC V5.01 software (Révay et al., 2001).Sample counting times were adjusted so the relative uncertainty of the area of the 662 keV gamma line of 137 Cs was typically 4-8 %.Several wet AMP-PAN samples were measured using a well-type HPGe detector (active crystal volume 182 mL) at the WHOI Radioanalytical Facility (Pike et al., 2013).The two detectors were crosscalibrated using NIST traceable radiocesium solutions.Minimum detectable activities (MDAs) were calculated using a formula by Currie (1968) for a limit of detection L D with a "well-known" blank (95 % confidence level), where the difference of the gross and net area of the peak was used as the background signal.The combined uncertainties (coverage factor k = 1) include uncertainties from counting statistics, detection efficiency, sample volume, recovery yield determination, and an additional uncertainty estimated at 3 % for small variability in ion exchanger volume.The relative combined uncertainty for 137 Cs activity was in the range of 5-10 % for surface seawater samples.For the deep seawater samples, lower activity but longer counting times resulted in 10-12 % relative uncertainty.All reported activities were decay-corrected to the date of the shutdown of the nuclear reactors at F1-NPP on 11 March 2011.
The chemical recovery of 134 Cs and 137 Cs was determined by comparing stable cesium concentrations in seawater aliquots taken before and after sample processing.The aliquots were 25-fold diluted by deionized water and analyzed on a high-performance double focusing magnetic sector field ICP-MS (Element 2, Thermo Finnigan).Cesium recovery was determined for each sample.The typical recovery was 85-92 % and 90-99 % for 100 L and 20 L volume samples, respectively.1 and 2).The contours of an average ocean surface velocity > 0.4 m s −1 were calculated by the SCUD model for June 2011 (red) and June 2012 (blue) to visualize the major Kuroshio and Kuroshio extension currents (Maximenko and Hafner, 2010).

Preexisting 137 Cs levels in the surface North Pacific
Global and local fallout from atmospheric nuclear weapons tests conducted in the 20th century was the major source of 137 Cs in the North Pacific Ocean (Aarkrog, 2003).In 2000, measured and estimated surface seawater 137 Cs activities across the North Pacific were in the range of 1.7-2.8Bq m −3 (Hirose and Aoyama, 2003a).To the best of our knowledge, no observations have been made for the Hawaiian Islands region since the 1980s.The estimates for 2011 were made based on measurements from neighboring areas (Povinec et al., 2005) and an effective half-life of 11-24 yr.This effective half-life integrates cesium removal from the surface ocean by radioactive decay and lateral and vertical removal processes (e.g., Hirose and Aoyama, 2003b;Povinec et al., 2005).Based on these removal rates, the estimated 137 Cs activity range in 2011 was 0.9-2.4Bq m −3 for the entire North Pacific.No preexisting 134 Cs should be present in the surface ocean due to its short half-life and the absence of recent sources.

Footprint of the F1-NPP-derived atmospheric fallout near the Hawaiian Islands and Guam
Fukushima-derived radionuclides in the atmosphere were detected around the Hawaiian Islands and Guam beginning 19 March 2011 (EPA, 2011).Surface seawater samples collected between March and May 2011 (G1-G6, M1-M6, Sta-tion Aloha, Honolulu; Table 1) at these locations (Fig. 1) were expected to carry only atmospheric fallout signature because direct discharges to the sea could not have yet reached these areas.In all our March-May 2011 samples, 134 Cs was below MDA (Table 1).Depending on sample volumes and counting times, the MDAs for both 134 Cs and 137 Cs were in the range of 0.1-0.8Bq m −3 .The lower value represented detection limits for the 100 L samples.At Station Aloha, the first sampling was on 13 April 2011, about three weeks after the F1-NPP-derived radionuclides were detected on air filters in Hawaii.Assuming uniform radionuclide distribution within the mixed layer, the product of the MDA and mixed layer depth gives a detection limit of surface deposition (Bq m −2 ).The mixed layer at Station Aloha was 50 m deep (HOT, 2012).Because 134 Cs was not detected and MDA was 0.2 Bq m −3 , the atmospheric fallout at Station Aloha had to be below 10 Bq m −2 at the time of sampling.This however does not rule out possible higher deposition rates earlier that were diluted by mixing by the time sample collection took place.
Cesium-137 activities in the March-May 2011 samples were in the range of 1.2-1.8Bq m −3 (Table 1 and Fig. 2).Higher values (1.6-1.8Bq m −3 ) were exclusively observed north of the main Hawaiian Islands (M4-M6). 137Cs activity in sample M6 (latitude 32 • N) from 19 April 2011 was 1.8±0.1 Bq m −3 , which was about 20 % higher than the baseline determined at Station Aloha (1.46 ± 0.06 Bq m −3 ) at latitude 22.75 • N. Should this difference be due to F1-NPP atmospheric fallout rather than variability in preexisting 137  activities, the sample would also contain about 0.3 Bq m −3 of 134 Cs based on the reported activities ratio of 1 in the F1-NPP source (e.g., Buesseler et al., 2012).Such a small level could however not be confirmed because the MDA for 134 Cs was two times higher than the expected activity.Nevertheless, activity of 137 Cs at stations M4-M6 is higher than at Station Aloha and potentially includes up to 0.3 Bq m −3 of F1-NPP-derived 137 Cs fallout.These sites overlapped with locations sampled by Aoyama et al. (2013a), who did not detect 134 Cs in the same part of the Pacific on 9-10 April 2011 and found 137 Cs activities of 1.4-1.9Bq m −3 .In their study detectable 134 Cs activities were found in the range of 1-14 Bq m −3 only north of latitude 40 • N between longitudes 170 • W and 150 • W in April and May 2011 (Aoyama et al., 2013a).Our results confirm that the southernmost extent of detectable increase in 137 Cs due to atmospheric fallout in the central north Pacific was at latitude 32 • N. Cesium-137 activities in surface seawater samples from Guam (G1-G7) were up to 15 % lower (1.2-1.4Bq m −3 ) in comparison to Station Aloha in the central North Pacific.Although a slight decreasing trend over 6 weeks between March and May 2011 was observed, this was probably a statistical Fig. 2. 137 Cs activities in surface seawater around the Hawaiian Islands (M1-M6), Guam (G1-G7), and time series for Station Aloha and Honolulu.The solid line is an average for Station Aloha, and dashed lines are ±2 standard deviations of the average.variability due to measurement uncertainties (Fig. 2).Additional sampling in the area 16 months later (G7) re-confirmed lower activities than at Station Aloha.It is interesting to note that, in contrast to our results, preexisting 137 Cs activities estimated using effective half-lives (Povinec et al., 2005) are same for Guam and the main Hawaiian Islands in 2011.

Long-term observation of 137 Cs in seawater around Hawaii
Seawater sampling was initiated at two locations in Hawaii soon after the F1-NPP accident.Surface coastal seawater (20 L) was collected in Honolulu on the south shore of Oahu (Table 1) approximately every month starting 27 March 2011.Larger open-ocean samples (100 L) were collected at Station Aloha by the monthly HOT cruises (HOT, 2012). 134Cs was not detected in any of these samples collected between March 2011 and February 2013.The MDAs of 134 Cs and 137 Cs were about 0.2 and 0.6 Bqm −3 for 100 and 20 L samples, respectively.The average and standard deviation of 137 Cs activities at Station Aloha (18 observations) was 1.46 ± 0.06 Bq m −3 .In coastal seawater (11 observations) the average and standard deviation of 137 Cs activities was 1.49 ± 0.07 Bq m −3 .The two averages agree within uncertainties and represent a baseline of preexisting 137 Cs in the central North Pacific.These values are also well within the expected range estimated using cesium levels from pre-2000 values and effective half-lives described earlier in Sect.3.1.
No seasonal fluctuation of 137 Cs activities was observed at the two sampling locations (Fig. 2).These data also confirm that the radioactive contamination released at F1-NPP directly to the ocean did not reach the main Hawaiian Islands by February 2013.Additional sampling (Ne1-Ne3) was performed from seawater intakes at the Natural Energy Laboratory of Hawaii Authority (NELHA) facility on the Kona coast of Hawaii Island.The surface water (depth 24 m) 137 Cs activity (1.53 ± 0.08 Bq m −3 ) was in excellent agreement with established baseline for coastal Oahu and Station Aloha indicating a uniform 137 Cs distribution around the Hawaiian Islands.Cesium-137 activities were 0.19 ± 0.02 Bq m −3 and below MDA (< 0.1 Bq m −3 ) for seawater collected at 674 and 915 m depths, respectively.These values are in agreement with depth profiles reported for the central North Pacific (Duran et al., 2004).Cesium-134 was not detected in any of these samples.

134 Cs and 137 Cs levels between Japan and the Hawaiian Islands
Dispersion of radionuclides and the potential for the radioactive plume to reach the main Hawaiian Islands was evaluated based on surface ocean samples collected on two cruises be-  Rypina et al., 2013).The contaminated water from F1-NPP was discharged north of the Kuroshio current, which acted as a southern boundary for transport of this oceanic source (Buesseler et al., 2012;Rypina et al., 2013).The atmospheric transport models (e.g., Stohl et al., 2012) predicted that some radionuclides from the atmospheric source were spread south of the Kuroshio current.This was later confirmed by seawater (Aoyama et al., 2013a;Honda et al., 2012) and sinking particle analysis (Honda et al., 2013).
The samples T1-01 to T1-08 from 2011 were on the southern side of the Kuroshio and Kuroshio extension currents with seawater salinity in the range of 34.0-34.5.The position on the southern boundary of the main currents was indicated also by contours of higher surface ocean velocity from June 2011 (Fig. 1, red contours).Because of the overlap between oceanic releases and atmospheric fallout in this region, the source of F1-NPP-accident-derived 134 Cs on the transect might be both direct oceanic discharge and fallout from atmospheric releases.
In June 2011, 134 Cs was not detected in the two westernmost samples of the transect (T1-01, T1-02).The oceanic releases could not reach these areas due to Kuroshio transport barrier, and the footprint of the atmospheric fallout detected in the area in April 2011 (Aoyama et al., 2013a;Honda et al., 2012) was rapidly advected from this area by the time of our sampling, as suggested by model simulations (Rypina et al., 2013).In samples between longitudes 155 • E and 180 • E (T1-03 to T1-08), activities of 134 Cs were 1-10 Bq m −3 (Fig. 3a).The highest value was observed in the sample with the lowest salinity (T1-06) indicating substantial influence by oceanic releases distributed mainly in the mixed waters north of the Kuroshio extension.The three sites clos- est to Hawaii (T1-09 to T1-11) had no measurable 134 Cs activities.Generally, activities in transect samples were lower than activities observed closer to Japan and north of the major currents (several hundreds and thousands of Bq m −3 measured in June 2011; Buesseler et al., 2012) indicating that we did not sample the core but rather the southeastern leading edge of the radiation plume.
Most of the 2012 transect samples were collected several degrees south of the 2011 transect.The higher salinities (34.6-35.1)indicate that those samples were collected south of the major currents (Fig. 1, Table 2).Since these currents are deflected north of Hawaii, this sampling pattern would reveal the potential for dispersion of radionuclides south of the currents towards the main islands.
In June 2012 (T2-01 to T2-10), 134 Cs was detected over a wider geographic area than in 2011 (between longitudes 150 • E and 175 • W), but the activities in surface seawater were lower, 1-4 Bq m −3 (Fig. 3b).The highest 134 Cs activities were associated with lower salinities (34.5-34.7),but measurable levels were observed in samples with salinities up to 35.1.This might indicate that during about 15 months after the accident the distribution of 134 Cs became more uniform in this area.Again the two sites closest to Hawaii (T2-09, T2-10) had no detectable 134   happened to have activities at the same level as the southeastern leading edge identified in 2011); and (2) radioactive releases from F1-NPP into the ocean continued over a longer period of time creating an extensive plume of radiation that continued moving through the study site for over 15 months.We analyzed additional samples collected in 2012.Samples N6 and N7 originating from the location of the 2011 transect had 134 Cs activities of 2.4 and 1.1 Bq m −3 , respectively.These results support our conclusion that the radiation became more uniformly distributed between the 2011 and 2012 samplings.Similar activities were found in samples N1 and N2 north of the main Hawaiian Islands (longitude 160 • W, latitude 40 • N), and from all our samples these were the easternmost that had a Fukushima signature, perhaps representing the eastern boundary of the radiation plume in June 2012. 134Cs was not detected at any station around longitude 150 • W in August 2012 (N8, N11, N12).
Samples N4 and N5 from the southwest part of the North Pacific (latitude 24 and 27 • N) from 2012 did not contain detectable amounts of 134 Cs, and 137 Cs activities were in the range of expected preexisting levels.This is in agreement with previous findings that the dispersion of oceanic releases in the south direction is limited by the Kuroshio current.
Generally, the samples between Japan and Hawaii contained 137 Cs activities lower than 6 Bq m −3 (with the exception for T1-06).When cesium-134 activity was higher than MDA, it was usually 1-2 Bq m −3 lower than 137 Cs activity (decay-corrected to 11 March 2011).The activity ratio in these samples holds true for higher activities but breaks down for samples that have 137 Cs levels comparable to preexisting activities.Figure 4 shows the 134 Cs and 137  plotted along a straight line.Linear regression of the data (95 % confidence level) gives a slope of 1.05±0.07as the activity ratio of the Fukushima-derived isotopes and estimates 1.6 ± 0.3 Bq m −3 as the preexisting 137 Cs activity.These values are in good agreement with the published ratio close to one (Buesseler et al., 2012) and the expected preexisting 137 Cs activity in the North Pacific of 0.9-2.4Bq m −3 (effective half-life, Povinec et al., 2005).

Cs activities
The results also confirm previous findings that the southern boundary of the Kuroshio and Kuroshio extension currents represents a boundary for radiation dispersion of direct oceanic discharges, and higher activities can be found within and north of the major currents.The Kuroshio extension and North Pacific current were deflected north of the main Hawaiian Islands throughout March 2011 to October 2012 (N.Maximenko, personal communication, 2012) leaving no F1-NPP isotopic signature around the islands.

Conclusions
The data presented in this study provide some constraints on the southeastern extent of radionuclides released from F1-NPP, and can be used for verification of 137 Cs dispersion models in the North Pacific Ocean.A visual comparison with published model simulation results indicates that our easternmost detection of 134 Cs from June 2012 (latitude 162 • W) is slightly north (Behrens et al., 2012) and east (Nakano and Povinec, 2012) of the predicted areas.The time-series measurements at Station Aloha and in Honolulu confirm model predictions by Behrens et al. (2012) that the radiation plume would not reach the main Hawaiian Islands at least until two years after their release from F1-NPP.
The major conclusions that we can draw from this study are as follows: (1) atmospheric fallout did not leave a significant radiocesium footprint in the surface ocean at the investigated regions of Hawaii and Guam.(2) The easternmost extent of the radiation plume between Japan and Hawaii was at 180 • E and 174 • W in June 2011 and 2012, respectively.The Kuroshio and Kuroshio extension currents were effective boundaries against the southward spreading of radiation, so the plume has not been detected over the past 2 yr at the main Hawaiian Islands.
). Coastal sampling was performed periodically in Honolulu on the south shore of Oahu starting 27 March 2011.At Station Aloha, about 150 km north of Oahu, offshore samples have been collected monthly beginning 13 April 2011.Samples were also gathered from around the Hawaiian Islands between 29 March and 19 April 2011.Several expeditions covered the area between Japan and Hawaii in June 2011 and in June-September 2012.Near-shore samples were collected in Guam starting 26 March 2011 until mid-May 2011; additional sampling was performed in September 2012.

Fig. 1 .
Fig. 1.Map of the western North Pacific between Japan and Hawaii illustrating sampling locations.The red and blue filled symbols represent sampling in 2011 and 2012, respectively; empty symbols are used for stations where sampling was performed in both years (for details see Tables1 and 2).The contours of an average ocean surface velocity > 0.4 m s −1 were calculated by the SCUD model for June 2011 (red) and June 2012 (blue) to visualize the major Kuroshio and Kuroshio extension currents(Maximenko and Hafner, 2010).

Fig. 3 .
Fig. 3. Activities of 134 Cs (red pattern) and 137 Cs (blue solid) in transect surface seawater samples between Japan and Hawaii from June 2011 (a) and June 2012 (b).The estimated preexisting 137 Cs activities are drawn by the empty boxes.Activities are decaycorrected to 11 March 2011.Salinity of the samples is shown by triangles.
Cs activities.The persistence of similar 134 Cs activities in the same area detected 12 months apart may be an indication of two things: (1) in 2012 we detected the trailing edge of the radiation plume (which www

Fig. 4 .
Fig. 4. The activities of 134 Cs and 137 Cs in analyzed surface seawater from the North Pacific Ocean.The linear fit of the data is represented by straight line and 95 % confidence limits (dashed lines).
(3) The easternmost detection of 134 Cs within the North Pacific in June 2012 was north of the Hawaiian Islands at longitude 162 • W and latitude 40 • N.

Table 1 .
Activities of134Cs and137Cs in seawater around Hawaiian Islands (M1-M6, Ne1-Ne3, Station Aloha, Honolulu) and Guam (G1-G7) after the Fukushima Dai-ichi Nuclear Power Plant accident.Average and standard deviation is presented for samples from Station Aloha and Honolulu.All samples were collected from the surface except Ne1, Ne2, and Ne3.Activities are decay-corrected to 11 March 2011.
tween Japan and Hawaii in June 2011 and June 2012.The area of interest is dominated by east-flowing warmer waters of Kuroshio and Kuroshio extension currents at a latitude of about 35 • N (for general circulation patterns, see, for example, Qiu, 2001; for dominant current directions in June 2011, see

Table 2 .
Activities of134Cs and137Cs in seawater between Japan and Hawaii after the Fukushima Dai-ichi Nuclear Power Plant accident.All samples were collected from the surface except N3.Activities are decay-corrected to 11 March 2011.
a Positive value for • E, negative value for • W. b Sampling depth 639 m.