Plutonium isotopes present in today’s oceans originate from various sources1,2,3. The major portion can be attributed to nuclear weapons testing in the atmosphere between 1952 and 1980. The most abundant plutonium isotopes in the marine environment are 239Pu (T1/2 = 24,110 yr) and 240Pu (T1/2 = 6,561 yr). The short-lived 241Pu (T1/2 = 14 yr) has in large part decayed, and its progeny 241Am (T1/2 = 433 yr) is now another prominent isotope associated with nuclear fallout.

In the 1950’s a series of nuclear tests was performed by the United States in the Marshall Islands, namely at the Bikini and Enewetak Atolls, an area known as the Pacific Proving Grounds (PPG). These tests released substantial quantities of radioactive material into the environment. Between 1946–1958 the total estimated fission yield from the PPG tests was 57.7 Mt, which corresponds to over 30% of the estimated 189 Mt total global fission yield from all atmospheric tests4. The distinct 240Pu/239Pu atom ratio provides a fingerprint of PPG-derived material5, and ratios in the sediments of the Northwest Pacific indicate a notable contribution from the PPG6,7. To this day significant amounts of plutonium continue to be released from the PPG into the surrounding waters. Ocean water from the test area is transported by the North Equatorial Current (NEC) towards the Philippines and further north with the Kuroshio Current. Indeed, coral records from the Pacific island Guam, ~2,000 km west of PPG and along the NEC pathway, show spikes in the 14C that can be linked to individual PPG tests8. It has been known for several years that plutonium from the PPG can be found in the Northwest Pacific marginal seas, particularly in the East China Sea9,10,11 and along the Japanese west coast12. Detectable PPG signals are also reported in sediments from the Sulu and South China Seas13,14.

Part of NEC water also heads south with the Mindanao Current and further on via the Indonesian Throughflow (ITF, Fig. 1). Coral-based 14C records from the Makassar Strait area suggest that a large proportion of Pacific-derived radiouclides is being transported southwards15. Surface water samples from the Sulu Sea and the Makassar Strait also showed a PPG plutonium contribution16. The significance of any PPG derived plutonium in ITF sediments is not presently known, however high 241Am concentrations were recently reported in ITF sediments from the Lombok Basin off Sumba Island, Indonesia17. Gamma spectrometric analysis demonstrated the sediment core contained a high inventory of 241Am. The aims of this study, performed on a sediment core taken at the same Lombok Basin station17, were to

  1. 1

    determine, if high 241Am concentrations at the Lombok Basin site are correlated with similarly high concentrations of plutonium isotopes, providing definitive evidence of transport of the transuranic radioisotopes through the ITF,

  2. 2

    ascertain the fraction of PPG derived plutonium based on known end-members from the PPG and global weapons fallout, and

  3. 3

    evaluate the depositional histories of the anthropogenic isotopes.

Figure 1: The location of the study site GeoB 10065, marked by a star, within the Indonesian Throughflow (ITF).
figure 1

Major ocean currents relevant for the Pacific - Indian Ocean water exchange are displayed schematically with arrows: North Equatorial Current (NEC) bifurcating into Kuroshio Current (KC) and Mindanao Current (MC), North Equatorial Countercurrent (NECC), New Guinea Coastal Current (NGCC). The U.S. nuclear test site Pacific Proving Grounds (PPG) is marked by a circle. Other locations mentioned in the text are marked by numbers: 1 East China Sea, 2 Sulu Sea, 3 South China Sea, 4 Island of Guam and 5 Makassar Strait. This map was created by using Ocean Data View software39.

The available 50 cm long deep sea sediment multicorer (MUC) core, GeoB 10065-9 MUC-B, covering the Anthropocene period with high resolution, was recognized as a potentially suitable material for achieving these aims. Notably, the sampling site is located in a region affected by seasonal coastal upwelling during the southeast summer monsoon. Core sediments consist of olive to dark olive foraminiferral mud, comprising mainly of a lithogenic portion (73%)18. The surface sediment consist largely of clay minerals19 and CaCO3 (16%)18. Total organic carbon (Corg) and nitrogen (N) are high at this site (2.3% and 0.31%, respectively) owing to the upwelling-related high primary productivity. C/N ratios of 7.5 and δ13C of −20.3 point to a predominantly marine origin of deposited organic matter18.


Age model and sedimentation rates

The depth profile of the naturally occurring 210Pb activity concentrations (Fig. 2a) reflects a very high sedimentation rate at the study site. The sediment core chronology based on the excess 210Pb, developed within this study, is presented in Fig. 2b. The first appearance of 241Am and 239Pu, expected around 1952, appears at a depth of ~27.5 cm, in an excellent agreement with the 210Pb age model. The average sedimentation rates vary between 0.61 ± 0.02 cm · yr−1 in the top layers and 0.27 ± 0.06 cm · yr−1 towards the bottom of the core, which covers approximately the last 160 years. These values correspond to mass accumulation rates of 0.14 ± 0.01 and 0.09 ± 0.02 g · cm−2 · yr−1. These high sedimentation rates are similar to those determined at this site over the last ~6,000 years from the parallel gravity core GeoB 10065–7 which, based on the radiocarbon record of mixed layer and thermocline dwelling planktonic foraminifera, gave sedimentation rates between 0.11 ± 0.05 and 0.35 ± 0.05 cm · yr−117.

Figure 2: Chronology.
figure 2

(a) Depth profile of excess 210Pb activity in dry mass measured by gamma spectrometry in the GeoB 10065–9 MUC-B core17 plotted at mass depth scale for compaction correction. Horizontal error bars reflect combined uncertainties (1 σ), including counting statistics and detector calibration uncertainty. Vertical error bars represent the thickness of the depth increment. An exponential fit, depicted as a dashed line, was used for estimation of the residual inventory below 50 cm depth, required for the application of the Constant Rate of Supply (CRS) age model. (b) Chronology of the MUC-B core based on the CRS model. Horizontal error bars represent model age uncertainties (1 σ). The onset of 241Am and 239Pu (1952) is indicated by arrows. (c) Sedimentation and accumulation rates derived from the chronology.

Anthropogenic isotopes

Gamma spectra from the uppermost samples of the core show high concentrations of 241Am. The detection of this radionuclide by gamma spectrometry is difficult, owing to the rather low gamma-ray energy (59.5 keV) and moderate emission probability, but for these samples evidence for 241Am was clearly above the decision threshold of 1–5 Bq · kg−1 for the experimental setup used. The threshold depends mainly on the sample mass, measurement time and the activities of other gamma emitters in the sample and detector system. Comparable activities with lower experimental uncertainty were confirmed within this study by means of accelerator mass spectrometry (AMS) and are plotted in Fig. 3a. Clear evidence for 137Cs was not present in the individual gamma spectra (<1.5 Bq · kg−1), but had been quantified in the parallel core MUC-A by summing up gamma spectra of 5 successive 1 cm depth intervals (Fig. 3e)17.

Figure 3: Depth profiles of anthropogenic radionuclide activity concentrations.
figure 3

(a) 241Am measured by gamma spectrometry17 (reference date: October 2010; for samples in which no Am was detected, decision thresholds (DT) are plotted instead) and by AMS (reference date: July 2015). (b) 239Pu measured by AMS. (c) 241Am/239Pu activity ratios (calculated from the AMS dataset). (d) 240Pu/239Pu atom ratios. (e) 137Cs measured by gamma spectrometry in the parallel core MUC-A17 decay corrected to the sampling date (September 2005). Vertical error bars represent the thickness of the depth increment. Horizontal error bars represent combined counting and calibration uncertainties (1 σ) for gamma spectrometry data and combined statistical and systematic errors (1 σ) for AMS data. The right axis shows ages derived from the CRS age model.

Plutonium-239 activity concentrations determined in sediment layers by AMS are plotted in Fig. 3b. The isotope was detected in the top 30 cm of the core, with high activity concentrations, comparable to those from deep ocean sediments between the major test sites of the PPG6. Below this depth traces of Pu were found, these however are almost two orders of magnitude lower than activities above 30 cm and are probably the result of natural replacement (e.g., due to biological or physical mixing) or reflect displacement of material during sampling. These values were therefore not considered for further interpretation. A broad maximum in the 239Pu concentration at approximately 10–20 cm depth can be observed. The activity is slightly lower in the top 10 cm.

Americium-241 shows a qualitatively similar depth profile (Fig. 3a). The Spearman’s rank-order correlation was used to determine the relationship between 241Am (AMS derived dataset) and 239Pu activities and yielded a statistically significant (rs(18) = 0.707, p = 0.001) positive correlation.

The 240Pu/239Pu atom ratio (Fig. 3d) varies between 0.20 at the top to 0.36 at 30 cm depth, and a single measurements of the 241Pu/239Pu ratio at 18–20 cm depth yields an atom ratio of (1.72 ± 0.20) · 10−3 (July 2015, equivalent to (2.77 ± 0.32) · 10−3 at September 2005). 236U content in all investigated sediment samples was below our detection limit.


The calculated inventories (Eq. 1) of the particle reactive radionuclides studied are much higher than would be expected from atmospheric fallout deposition alone (Table 1). For example, the excess 210Pb inventory is 61,300 ± 1,700 Bq · m−2, which leads to a mean annual flux (Eq. 2) of 1,914 ± 54 Bq · m−2 · yr−1. This value is 20 times higher than the flux observed for Darwin, Australia20. The other radionuclide inventories (239Pu, 240Pu and 241Am) are also high. This strongly suggests that sediment focusing is occurring at the study site, which enables study of deposition conditions with high time resolution. This finding is in agreement with the high accumulation rates observed at the study site, which are exceptional for deep sea sediments.

Table 1 Radionuclide inventories measured at the study site.

In contrast, the observed inventory for 137Cs of 84 ± 15 Bq · m−2 (September 2005) is lower than what would be expected to be deposited at this site from global fallout. A reference value to compare the 137Cs inventory to is about 250 Bq · m−2 as a mean value for 0–10°S latitude band, decay corrected to the year 20054. A possible explanation for this observed mismatch in inventories is the different biochemical behaviour of caesium, which is rather soluble in the marine environment. Plutonium and americium, being particle reactive, are more rapidly scavenged in areas with sufficient particle availability11.

The 240Pu/239Pu atom ratio can provide information about the origin of plutonium in the environment. Early atmospheric testing in the years 1952–1954 was dominated by high 240Pu/239Pu ratio U.S. tests at the Marshall Islands. In October 1952 the first thermonuclear test, Ivy Mike, was conducted at Enewetak Atoll. Airborne and condensed debris from the test yielded a 240Pu/239Pu atom ratio of 0.363 ± 0.00421. A recent high resolution study of plutonium accumulation history in a coral from Enewetak Atoll lagoon provides refined information about Pu ratios in the local PPG fallout22. This uncovered a wide variation in 240Pu/239Pu atom ratios with time; during the period 1952–1955 the values decreased from 0.424 ± 0.026 to 0.234 ± 0.016. The value related to the largest test, Ivy Mike, was found to be very close to that previously reported: 0.379 ± 0.03822. Furthermore, the largest U.S. thermonuclear test at Bikini Atoll, Castle Bravo in March 1954, heavily contaminated the atoll with high (>0.3) 240Pu/239Pu atom ratio fallout5,23.

In the time period between 1955 and 1958 the U.S. tests continued to produce high 240Pu/239Pu ratio material24, however a number of smaller tests yielded local fallout with much lower ratios22. For instance, tests at Enewetak Atoll (e.g., Redwing Apache in 1956 or Hardtack I Oak in 1958) were characterized by 240Pu/239Pu atom ratios of ~0.17 and ~0.08, respectively22. A substantial contribution to the total explosive yield in this period also came from tests conducted by the U.S.S.R. and the U.K., and overall there was a general decrease in the isotopic ratio to ~0.22, recorded in Arctic and Antarctic glaciers25.

Atmospheric testing was briefly interrupted worldwide during the three year moratorium (1958–1961). During the post-moratorium atmospheric testing period in 1961–1963 the total yield peaked25 with a series of very large tests by the U.S.S.R. at Novaya Zemlya, and the isotopic ratio decreased further. The Partial Nuclear Test Ban Treaty came into effect in October 1963 leading to a subsequent significant decrease in atmospheric fallout. While French and Chinese atmospheric tests continued until the 1980’s, these were of much lower yield than the earlier tests of the U.S.A. and U.S.S.R. Although the atmospheric radionuclide concentrations decreased, radionuclides in the oceans remain to great part in the water column.

The fraction of PPG derived plutonium in the studied core was calculated from the PPG and global fallout end-members (Eqs 4 and 5). In previous studies, values of 0.36 or 0.30 were used as a PPG end-member for distinguishing between different plutonium sources6,10,12,13,26. In the light of the latest results22, we consider the value 0.36 as one representing well the earliest testing period at the PPG. For the younger sediment core section, fractions of PPG plutonium calculated using this ratio might be underestimated. For more recent PPG derived plutonium, we have used an alternative end-member value of 0.27 ± 0.03, to better reflect the material being continuously released from the PPG sediments. This ratio represents an average value from a coral-based record of plutonium deposition at the Guam Island between years 1981 and 199927. The PPG fraction quantified using this ratio might well represent the upper estimate. A 240Pu/239Pu atom ratio of 0.176 ± 0.008, calculated from 15 soil samples collected between 30°N and 30°S28, is considered characteristic of integrated global fallout and this value was used as the global fallout end-member in the mixing model.

The 240Pu/239Pu atom ratios (Fig. 3d) and calculated fractions (Supplementary Table S2) show, that virtually all of plutonium in the deepest sediment slices of our core can be assigned to be of PPG origin. In Fig. 4 combined 239+240Pu yearly depositions over the studied time interval together with PPG fraction estimates are displayed. The first arrival of PPG radionuclide enriched waters could be expected after 1952: a high resolution 14C coral record from the Makassar Strait15 shows elevated 14C in 1955 following the Castle Bravo test. This is in a good agreement with the observed appearance of Pu and Am in our sediment core within the period of the first large tests; the bottom and top of the first Pu-rich layer in the sediment core are dated at 1938 ± 6 and 1955 ± 6 AD, respectively.

Figure 4: The time-line of combined 239Pu and 240Pu fluxes to the Lombok Basin sediments.
figure 4

The temporal axis was derived from the 210Pb age model. The minimum fraction of PPG derived plutonium for each analysed sediment core interval is distinguished by the dark orange colour (derived from Model 1), while the dotted area shows the limit of PPG derived fraction (calculated using Model 2). The remaining fraction in green represents the global fallout plutonium. The error bars (1 σ) consist of combined Pu activity and accumulation rate uncertainties.

The 240Pu/239Pu atom ratio and fractions derived from PPG versus integrated global fallout remain at elevated, rather constant levels during 1960’s. The plutonium isotope ratio is well above that recognised value for global fallout and strongly suggests that the material is originating almost exclusively from the PPG. Above 18 cm depth the 240Pu/239Pu atom ratio decreases significantly from ~0.34 to ~0.24, presumably largely due to the influence of the post-moratorium tests at Novaya Zemlya leading to global fallout with a lower ratio.

In the upper part of the core covering the 1970’s to the 2000’s, approximately 30–60% can be attributed to integrated global fallout. The remaining portion is likely transported from the PPG as a result of remobilization of plutonium deposited from the close-in fallout in the Marshall Islands area. The plutonium flux to the sediment remains high (2.5–4.5 Bq · m−2yr−1) over the entire 1950’s–2000’s period. The slow decrease in the transuranic isotope deposition rates is especially remarkable, because it contrasts the much more rapid decrease observed in high resolution coral archives at multiple Northwest Pacific sites22,27,29. The biogeochemical reasons for this mismatch deserve further investigation.

Activity ratios of 241Am/239Pu increase in the upper part of the core, from around 2.5 at 30 cm depth to close to 4 towards the surface (Fig. 3c). This increase is attributed to the steady increase in the proportion of Am to Pu, as a consequence of radioactive decay and the relatively short 14.35 year half-life of 241Pu, that occurs in the PPG sediments and also during transport to the study site, combined with different behaviour of the two elements during re-mobilisation (from the sediment back into the water column at the PPG), transport and subsequent scavenging processes prior to deposition in the Lombok Basin.

The age of the Pu source term was calculated using the measured activities of 241Pu and 241Am for the 18–20 cm depth slice. The 210Pb dates of the bottom and top of the sediment core interval are 1967 ± 5 and 1970 ± 4, respectively, and the 240Pu/239Pu atom ratio indicates a minimum of 90% of PPG derived plutonium. Using Eq. 7 and the 241Pu – 241Am data from this slice, the plutonium age6,30 was determined as 1945 ± 5. This apparent age is older than 1952 Ivy Mike test, possibly a result of preferential transport and/or scavenging of americium over plutonium, as indicated by the observed increase in the 241Am/239Pu ratio noted above, and as documented for instance in the Atlantic Ocean31 and the Mediterranean Sea32.

Uranium ions have shown to be strongly conserved in seawater33 and anthropogenic 236U has recently seen interest as a conservative oceanic tracer. Currently there are few studies confirming negligible scavenging of anthropogenic 236U to the seabed34. The absence of 236U in our sediment with high concentrations of other anthropogenic radionuclides supports the suggestion of this isotope being a suitable conservative oceanic tracer.

In summary, our study has shown that the previously observed high 241Am concentrations at the Lombok Basin site are correlated with similarly high concentrations of plutonium isotopes. Our results demonstrate that the transport of the transuranic radioisotopes through the ITF continues to this day, and that over the period since the cessation of atmospheric nuclear weapons testing, the rate of their deposition to the sediments has decreased more slowly than would be expected from coral records. Based on the end-member analysis of plutonium atom ratios we infer that the portion of PPG derived plutonium was high until 1970’s and remains significant.


Sediment coring

Sediment core GeoB10065–9 was taken with a multicorer (MUC) on 5. September 2005 during the R/V SONNE SO-184 cruise in the Lombok Basin (9°13.41′ 118°53,55′, 1,284 m, length 50 cm). The material used for the analysis within this study is identical to the sediment core described as MUC-B previously17. The inner diameter of the transparent sampling tube was 10 cm. The core was recovered with a visibly undisturbed sediment - water interface.

Gamma spectrometry

The sediment samples were measured freeze dried. Material was taken in 1 cm slices between 0–10 cm, 2 cm slices between 10–20 cm and 5 cm slices between 20–50 cm core depth; a total number of 21 samples was measured. Individual sample masses varied between 1–3 g. Transparent polystyrene cylindrical dishes with an inner diameter of 34.7 mm and height of 12.6 mm were filled with the samples to heights of 2–5 mm. All samples in their containers were sealed by welding using a gas tight foil and the measurement was started more than 3 weeks after sealing, in order to ensure, that the radioactive equilibrium between 226Ra, 222Rn and its progeny had been established.

A coaxial HPGe detector (Canberra Industries, 50% rel. efficiency) housed in 10 cm Pb shielding with Cu, Cd and plastic linings operated under Genie 2000 software was used for gamma spectrometry. The detection efficiencies were calculated using LabSOCS software (Canberra) and verified using measurements of certified materials in similar geometries. The overall uncertainties combine errors arising from counting statistics, efficiency calibration, sample mass determination and decay data. More details on analytical uncertainties are available in the Supplementary material.

Radiochemical extraction

The method is based on Srncik et al.35 with minor modifications which are described briefly. The samples (1–3 g) were ashed at 450 °C and then spiked with ~1010 atoms each of 242Pu (SRM 4334H, NIST), 233U (CRM 111-A, NBL) and 243Am (SRM 4332C, NIST) in solutions derived from the given standards. After leaching in 8 M HNO3 the undissolved residues were filtered and about 0.2 g NaNO2 was added to adjust the oxidation state of Pu to Pu(IV).

Am separation from Pu and U

A column filled with 0.5 g UTEVA© (100–150 μm) was mounted underneath a column filled with 2 g of BioRad AG© 1 × 8 (100–200 mesh, Cl form). After conditioning with 8 M HNO3 the samples were transferred onto the resins followed by rinses with 8 M HNO3 and 4 M HNO3. Then the columns were separated and all transmitted solutions were retained as these contain the Am. 37% HCl was added to the BioRad AG© 1 × 8 resins to remove Th, and Pu was then eluted with 0.1 M NH4I-9M HCl. 6 M HCl was added to the UTEVA© resins prior the U elution with 1 M HCl. Both the Pu and U fractions were evaporated and 1.5 mg Fe was added as FeCl3.

Am purification

The Am fractions were evaporated, dissolved in 2 M HNO3 and few grains of NH4SCN were added to indicate the presence of iron which was then reduced with ascorbic acid to Fe(II). After conditioning 0.74 g TRU© (100–150 μm) resin per sample with 2 M HNO3 the samples were loaded onto the columns, followed by washes with 2 M HNO3 and 9 M HCl before Am was eluted with 4 M HCl. The Am fractions were evaporated and 1.5 mg Fe was added.

Final preparation for AMS measurement

The Pu, U and Am fractions were dried for at least 48 h at 90 °C and then combusted at 800 °C for 8 h. The Pu and Am fractions were mixed with Ag powder (4x the sample weight), while the U samples were mixed with 1 mg of Al powder. Each mixture was then pressed into individual aluminium sample holders for the AMS measurement. The Ag and Al powders are added to ensure electrical and thermal conductivity.

Accelerator mass spectrometry

Isotope concentrations were determined with AMS using the 14UD pelletron accelerator at the ANU based on the methods described in Fifield36. The method provided above effectively separates americium, plutonium and uranium, however the 236U content in these sediment samples was below our detection limit, and results for only americium and plutonium isotopes are reported.

Negatively-charged monoxide ions from the purified Am (Pu) samples were extracted using a Cs sputter ion source. The ions were accelerated to 105 keV, mass-selected with a high resolution (M/ΔM ≥ 300) magnet, focussed and injected into the 14UD accelerator, which was operated at ~4 MV. A low-pressure gas stripper in the high-voltage terminal at the centre of the accelerator dissociates the 4 MeV molecular ions and removes outer electrons from the atom species. The positively charged atoms were then again accelerated and Am5+ (Pu5+) ions of energy ~24 MeV selected by an analysing magnet.

The isotopes were measured using a slow cycling sequence to determine the relative count rates for each isotope. A propane-filled ionisation chamber was used to identify and count individual atoms. Measurement times for americium isotopes were typically 1 min for 243Am and 2 min for 241Am, and for the plutonium isotopes 1 min for 242Pu, 2 min for 239Pu and 3 min for 240Pu. At least two cycles of the sequence, and concluding with a 243Am (242Pu) measurement, made for each sample. 241Am (239,240Pu) concentrations were deduced from the known mass of the 243Am (242Pu) spike added to the sample.

Analytical errors are quantified in the Supplementary material.

Age model

The chronology of the sediment core was constructed using the constant rate of supply (CRS) model37, assuming constant flux of 210Pb to the sediment. The chronology data are provided in Supplementary Table S1.

Inventories and fluxes

The radionuclide inventories (Bq · m−2) were calculated as a sum of measured values of activity concentration Ci in the sediment core as:

where zi is the thickness and ρi the dry bulk density of the ith core slice. If a constant flux of the radionuclide is presumed, the average yearly flux (Bq · m−2 · yr−1) can be calculated as:

where λ (yr−1) is the radionuclide’s decay constant.

Dates from the CRS age model allow for the determination of a mass accumulation rate, ri (g · m2yr−1), for each core slice i. We have used these to determine the associated plutonium fluxes, Fi, viz:

Source attribution

Assuming that the plutonium activity in the core is a mixture of Global fallout and PPG,

the contributions of global and PPG derived plutonium was resolved from the measured 240Pu/239Pu atom ratios RM using a two end-member mixing model38:

where 3.674 is a factor for conversion between atom and activity ratios of 240Pu/239Pu. Two scenarios were considered to estimate both the initial and ongoing PPG fractions:

  • Model 1: RGlo = 0.18 ± 0.01 for integrated global fallout28, and RPPG = 0.36 ± 0.02 for PPG21,

  • Model 2: RGlo = 0.18 ± 0.01 for global fallout end-member28, and RPPG = 0.27 ± 0.03 for PPG27.

The portions of activity due to PPG from total Pu activities are then calculated by solving the system of Eqs 4 and 5:

Am-Pu age

The age of the Pu source term was calculated using the measured activities of 241Pu and 241Am,

assuming that initial 241Am = 0. λPu and λAm are decay constants of 241Pu and 241Am, respectively.

Additional Information

How to cite this article: Pittauer, D. et al. Continuous transport of Pacific-derived anthropogenic radionuclides towards the Indian Ocean. Sci. Rep. 7, 44679; doi: 10.1038/srep44679 (2017).

Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.