Abstract
Introduction
On 22 September 1995, the General Conference of the IAEA called on all States concerned:
`to fulfil their responsibilities to ensure that sites where nuclear tests have been conducted are monitored scrupulously and to take appropriate steps to avoid adverse impacts on health, safety and the environment as a consequence of such nuclear testing'.
This led the Government of France to request the IAEA to undertake an independent assessment (`the Study') of the radiological situation at the atolls of Mururoa and Fangataufa in French Polynesia, where France had conducted a nuclear weapon testing programme between 1966 and 1996.
The IAEA established an International Advisory Committee (IAC), chaired by Dr E Gail de Planque of the United States, to oversee the Study, the aims of which were:
to assess the radiological situation at the two atolls and involved areas from the point of view radiological safety;
to ascertain whether there are any radiological hazards to people;
to make recommendations on the form, scale and duration of any monitoring, remedial action or other follow-up action that might be required.
The programme of work developed by the IAC was carried out by 2 Task Groups supported by 5 working groups. The French authorities made available much previously unpublished information that was invaluable to the Study and a major effort was expended in evaluating and, where practicable, validating these data. Task Group A, supported by 2 working groups concerned with evaluating the existing contamination in the surface terrestrial and aquatic environments of the atolls respectively, assessed the present and future radiation doses due to these radioactive materials. Working groups on the `Source Term', `Geosphere Radionuclide Transport' and `Marine Modelling' enabled Task Group B to estimate the current inventories, and release rates, of the residual radionuclides from the underground tests either to the lagoons or the open ocean, and their wider dispersion. These data allowed Task Group A to assess the longer-term doses that could arise from these materials. This meeting in Vienna was convened to present to the wider scientific community the results, conclusions and recommendations developed by the Study. There were some 120 attendees, inclusive of the Study participants who were giving the 26 invited papers summarising the Study.
Before the presentation of the Study results, two introductory papers provided a degree of scene setting. Barthoux (France) gave some details of the development of the testing programme from 1966-1996 (see below). Although it was not part of the Study to estimate the past human exposures in French Polynesia from the atmospheric component of the test programme, the Secretary of UNSCEAR (Bennett) did provide an assessment of their contribution (3.5 µSv a-1 in the peak years of 1969-1972, or about 15% of the total) to the dose rate in the area from global fallout. Localised fallout at Tureia Atoll, the Gambier Islands and Tahiti from specific tests has been estimated to have delivered maximum doses in the range 0.8-5.5 mSv in the first year of exposure. These values gave a useful context for the results of the Study, particularly as the entire radionuclide inventory of these tests was initially released into the accessible environment (i.e., the atmosphere).
The French programme at the Mururoa and Fangataufa atolls comprised a total of 193 nuclear tests and safety trials. There were 46 tests of weapons in the atmosphere - 4 at the surface on barges, 34 suspended from balloons, 3 dropped from aircraft and 5 safety trials (3 with no nuclear yield and 2 with an estimated fission yield of 0.001 kt). The four barge-mounted tests (3 at Mururoa and 1 at Fangataufa), in which the fire-ball intersected the surface, have been responsible for the greater part of the residual radioactivity in the accessible environment of the two atolls with the majority of the remainder arising from the safety trials. After reviewing the results of the environmental monitoring programme carried out over a number of years by the French scientists, the IAC concluded that the data appeared to be comprehensive and of high quality. In consequence, the terrestrial and aquatic environment working groups each devised sampling programmes that would, within the time and resource constraints on the Study, be the minimum necessary to validate the extensive French data.
The current environmental situation
The terrestrial sampling campaign, coordinated by the IAEA Seibersdorf Laboratory in Austria, collected 299 samples of aerosol, soil, coral sand, coral bedrock, and vegetation (including coconuts) and 106 in situ γ-spectra. The radionuclide analysis of the samples was carried out, under strict quality control procedures, by an international network of 11 laboratories that had been selected on the basis of good performance in an inter-comparison test (it should be noted that the French analytical laboratory at Montlhéry also performed very well in this, and previous, inter-comparison exercises). The radioanalytical results for the terrestrial samples showed general agreement (i.e., overlapping ranges of observations) with the more numerous data made available by the French authorities; discrepancies could reasonably be attributed to the spot nature of the samples taken for the Study, the known small-scale heterogeneity of the radionuclide distribution in the environment, and the biological variability in radionuclide uptake (Schönhofer, Austria).
A major Study effort was expended on monitoring the surface of the small islet (motu) of Colette in the northern area of the Mururoa atoll that had been one of the sites for the atmospheric safety trials (the two adjacent motus, Ariel and Vesta were also used). The area had been contaminated with plutonium from the nuclear cores destroyed in the conventional explosions. After each safety trial, the surface of the motu was monitored, the greater part of the contaminated debris collected for disposal (in deep boreholes), and the contaminated area covered with asphalt. In the period up to 1987, the French authorities had carried out several clean-up operations to reduce the level of plutonium contamination on the Colette motu with the final decontamination criterion being 1 MBqm-2 averaged over a 20m × 20m area. Such a criterion is operationally defined by the instrumentation available but could also mean that there were a few hot particles in an area of otherwise relatively low contamination. The Study results for limited areas on the Colette motu, obtained with different instruments, indicated that the current level of plutonium contamination is about 3 times higher than the average reported by the French, and that hot spots were very unevenly distributed and contributed about 5% of the total contamination. During the survey, two active particles were identified in situ and collected; subsequent laboratory analysis showed that these had 239Pu activities in excess of 0.6 MBq. A further 18 particles with 239Pu activities less than 5000 Bq were later identified and isolated from coral sand and rock samples in the laboratory (Danesi, IAEA).
The aquatic environment programme, coordinated by the IAEA Marine Environment Laboratory in Monaco, also entailed the collection of over 300 samples (lagoon and ocean water, sediment pore water, sediment, corals and biota). In situ gamma spectrometric surveys of the seabed were used to optimise the sediment sampling in the vicinity of the sites of the 4 barge tests where it was known from the French data that the highest levels of contamination were likely to be found. To reduce the weight for transport, the individual seawater samples (13000 litres in total) were processed on site, as appropriate, to concentrate the radionuclides. The final analyses were carried out by an international group of 9 accredited laboratories in addition to the IAEA Monaco Laboratory. As was the case for the terrestrial environment, the values for the radionuclide concentrations in the various marine samples were found to be in generally good agreement with the more numerous French data. The complete data set was used to provide recommended, site specific Concentration Factor (CF) and Distribution Coefficient (kd) values for use in the dose assessment.
As expected, the sediments adjacent to the Colette motu were found to contain high concentrations of plutonium, up to 1 MBqkg-1 dry weight, from the safety trials. A study using track etch detectors indicated that the plutonium could be present in the form of small particles at densities up to 104 kg-1, with activities in the range 10-2000 Bq and sizes less than about 0.5 mm (Woodhead, UK).
The data obtained by the Study team were used to estimate the inventories of certain radionuclides in the lagoon waters (a current source for the wider ocean) and the sediments (a continuing future source). The average tritium concentration in the lagoon waters was found to be 2-3 times higher than that in the open ocean. Together with the finding of even higher concentrations in sediment pore waters and the known exchange of water between the lagoons and the ocean, this was interpreted as evidence of the release of tritium from an underground source. The complete data set, i.e., from the French monitoring programme and the Study, indicates that the 3H release may now have stabilised or be declining. It was concluded that the data for 90Sr indicated a balance between remobilisation from the lagoon sediments and exchange with the open ocean. For 137Cs and 239Pu, the available information is clear in showing a continuing decline in the average concentrations in the lagoon waters and, therefore, of release to the ocean. The aggregated source terms from the two atolls to the open ocean are estimated to be: 3H, 5.6 TBqa-1; 90Sr, 27 GBqa-1; 137CS, 11 GBqa-1; and 239+240Pu, 11GBqa-1 (Povinec, IAEA, Monaco).
The underground sources of radionuclides
Of the total of 147 underground tests, 137 were conducted at depths between 500 and 1100 m beneath the atoll rims or the lagoons (127 at Mururoa and 10 at Fangataufa) and 10 safety tests were conducted at depths greater than 280 m beneath the rim of Mururoa. For the 140 tests with a nuclear yield, the Study used the available seismic records to estimate the actual yield (in practice, 11 of the tests, including the 3 safety tests with some fission, did not produce a sufficiently strong seismic signal to be detected and were arbitrarily assigned a nominal yield of 1 kt). The estimated aggregate yield produced by the Study (3.19 Mt) was in very good agreement with the French information derived from radiochemical analyses of the explosion products. From available information on generic nuclear weapons and the data on residual inventories of specific radionuclides provided by the French authorities, estimates were made of the proportions of the total nuclear yields due to fission of 239Pu, 235U and 235U, and fusion. This information provided the basis for estimates of the total fission product yields and, together with assumptions concerning the materials in the nuclear devices, the composition of the surrounding rocks and neutron activation cross-sections, was also used to estimate the total inventories of the activation products. Applying lower and upper limits of 1 and 1010 years on the half-lives of radionuclides that could be reasonably expected to leave the cavities and reach the biosphere to deliver significant doses, the underground inventories of 36 radionuclides were estimated and found, in general, to be in good agreement with French values. On the basis of the radionuclide production mechanisms and their physico-chemical properties, the inventories were partitioned between the lava produced by the explosions, the rubble in the collapse chimney and the seawater that saturates the basalt and quickly seeps back into the cavity. This provided the Study source term for the modelling of radionuclide transport in the geosphere (De Geer, Sweden).
For the purposes of the Study, the underground tests were separated, on the basis of information provided by the French, into 6 categories depending on their perceived potential for releasing radionuclides to the ocean:
121 `regular' tests for which the overlying basalt was considered to have sufficient integrity to give good confinement of the radionuclides, i.e., the fracturing above the explosion cavity and the collapse chimney did not penetrate to the carbonate capping;
4 tests in which the overlying basalt had been assessed as of low integrity;
12 tests in which the post-explosion collapse chimney penetrated to the carbonate capping;
3 safety trials at least 280 m deep in the carbonate rock in which there was a very small fission yield;
4 safety trials at least 280 m deep in the carbonate rock in which there was no fission yield; and
3 safety trials in the volcanic rock in which there was no fission yield.
Radionuclide transport in the geosphere
Fairhurst (USA) provided a lucid account of the geological origin and evolution of the atolls and their present stability. The volcanic production of the basalts under the sea generates intense horizontal and vertical fracturing from thermal shock and it is, thus, close to isotropic in respect of transport pathways. The aerial volcanics have been subject to weathering and the erosion products can then be covered by additional layers of lava as can coral reefs on the coastal flanks. After the end of active volcanism, erosion, lava collapse and the sinking of the mass of the volcano under its own weight on the oceanic crust give rise to conditions in which coral growth can produce increasing thicknesses of consolidated carbonate sediments that cap the volcano to depths up to 450 m. The carbonate sediments have been subject to significant chemical modification - dolomitisation (Ca-Mg exchange), dissolution by seawater at depth to produce a porous chalky limestone and karstification through freshwater action in periods of low sea-level during glaciations. The result is that the carbonate caps of the atolls show a diversity of discrete stratigraphic sequences. These factors have implications for the migration of the radionuclides through the geosphere.
The major force generating water movement in the atoll basement is the residual heat from the volcano. This drives a thermal syphon that draws seawater in through the flanks of the atoll and forces it upwards to exit through the lagoon. A generic model has been developed with water flows adjusted to give a match with the temperature distributions observed in the numerous drill holes. The estimated water velocities are highest in the karsts, lower in the overlying carbonates and lowest in the basalts where the flow is governed by the presence of small fractures. In the vicinity of the test cavities/chimneys, the residual heat anomaly from the explosion distorts the flow field with a local thermal syphon cell, but the enhanced transport of radionuclides is limited to some extent by the necessity to draw water from undamaged rock (de Marsily, France).
Models were developed to determine the releases of radionuclides from the lava and chimney rubble into the seawater in the cavities, and thence into the surrounding basalt and the overlying carbonates (Hadermann, Germany and Levins, Australia). The model output for the concentrations of 3H, 90Sr and 137Cs in the chimney water were in reasonable agreement with available measurements both from French studies and those made specifically for the Study (Warnecke, IAEA). The model for the transport of the radionuclides through the basalt into the carbonates and thence into the oceans took account of the fractured and porous nature of the rock, i.e., the processes of dispersion in the fractures and diffusion in the rock matrix, sorption and radionuclide decay chains. It was concluded that the 137Cs from the 121 category 1 tests would not be released into the carbonate, but that releases would be likely to occur for the 16 category 2 and 3 tests. The uncertainty in the predictions was mainly due to the lack of site-specific values for the model parameters and, hence, the necessity to make conservative assumptions. Overall, it was concluded that radionuclides from the 121 category 1 tests would be effectively contained with the greater part of the actual releases coming from the remaining 26 tests; sorbing radionuclides with T1/2<50 years would decay before release; longer lived radionuclides would survive transport; and plutonium would be released at low levels that would peak at 4-5000 years and extend out to 105 years.
Radionuclide transport in the sea
These release rates provided the source terms for modelling the dispersion of the radionuclides in the ocean on scales extending from the local (the lagoons), to regional (French Polynesia) and to global (the remainder of the South Pacific Ocean). The lagoon model, constrained by the available information on the variation of the 3H inventory in the lagoons, provided the basis for estimating the time trends in radionuclide concentrations in the lagoon waters from both the underground sources and due to leaching from the present sediment inventory. The current concentrations of 3H, 90Sr, 137Cs and 239Pu in the lagoons are above the corresponding concentrations arising in the open ocean from global fallout. It is predicted that 137Cs and 239Pu concentrations are unlikely to exceed present levels at any time in the future. 90Sr concentrations could rise marginally but only during the next few decades. The tritium concentrations are likely to remain fairly constant for a few decades before declining slowly. These concentration data were one of the inputs for the dose assessment. A second model was used to estimate the export of contaminated sediment from the lagoons under both normal and storm conditions. Four compartmental models, employing a variety of available information to develop the required basic water circulation patterns, were used to estimate the radionuclide concentrations at a selection of inhabited Islands in French Polynesia. For the estimated time-dependent releases of radionuclides, the predicted concentrations at the Islands from the four models were in reasonable agreement (within a factor of about 10), and below the current ambient concentrations from global fallout. In only one case considered, that of a hypothetical disruptive event (see below) leading to a major, instantaneous release of plutonium, was the concentration at the nearest inhabited island of Tureia predicted to increase (by a factor of 100) above the local ambient level, and that only for a few years. As might be expected from these results, the predicted radionuclide concentrations in the wider south Pacific Ocean are extremely low (Osvath, IAEA, Monaco and Mittelstaedt, Germany).
Levins (Australia) described a number of disruptive events that could potentially lead to substantially different patterns of radionuclide release from the atolls. It was clear from the information presented by Fairhurst (see above) that future volcanic activity at the atolls was extremely unlikely. Global warming would lead to the flooding of the atolls, but would result in little change in the pattern of radionuclide release to the ocean from the underground cavities. Past history indicates that there is likely to be a period of glaciation in the future and the atolls would then become islands. This would increase the number of potential pathways for human exposure to plutonium, e.g., inhalation of resuspended, dried lagoon sediment, consumption of crops grown on the dried lagoon sediments, and the drinking of contaminated water from freshwater lenses under the atolls. An earthquake, a severe storm or a tsunami could trigger a slide of carbonate rock along fractures generated by the testing programme. As a worst case, it was assumed that the slip surface would intersect both a safety trial site and one of the 12 category 3 test sites for which the chimney reached the carbonate layer and that the corresponding plutonium inventory would be instantaneously released to the ocean (this latter assumption was acknowledged to be very pessimistic, but was used because it would be difficult to justify any particular fractional release). Only the last two scenarios were considered further from the viewpoint of the potential radiological consequences.
Dose assessments
As there is no firm historical evidence that either Mururoa or Fangataufa has been permanently inhabited, a hypothetical population was assumed to have established a village at the widest and most protected area of Mururoa (at the eastern end of the atoll). In addition, this population was assumed to have adopted a diet identical to that determined for the present population of Tureia, the nearest permanently inhabited atoll, i.e., mainly subsistence but supplemented with imported bread, rice and noodles in place of the traditional staples. The radiological assessment considered a total of 6 potential pathways of exposure: external from contaminant radionuclides in the soil; inhalation of radionuclides in resuspended soil or sediments; ingestion of radionuclides taken up by locally grown foods or in foods harvested from the local marine environment; incorporation of a plutonium particle in a wound; ingestion of soil (pica) by children; and external irradiation from beaches and fishing gear.
The most important source of potential exposure to the hypothetical population from the currently existing contamination is the radionuclide content of the locally produced foods: this results in an estimated dose rate of less than 6 µSv a-1. External exposure (including visits amounting to 4 h in every 20 d to the relatively higher dose rate area on the Colette motu) might amount to 1 µSv a-1, and pica could result in an exposure of perhaps 2 µSv a-1 to children for short periods. The visits to the Colette motu entail the risk of plutonium particle incorporation into a wound caused by the coral rocks. It was decided that this was a possibility that was more appropriately assessed in terms of risk rather than dose. The assessed risk of fatal cancer induction was <10-6 at the 99% confidence level, a level of risk that is commonly regarded as trivial. The estimated dose rates from inhalation and exposure during fishing were very much smaller and could be neglected. As noted above, the future releases of radionuclides from the underground sources are not predicted to increase the concentrations in the local environment above those that currently exist. It follows, therefore, that future dose rates to the hypothetical population on Mururoa also will not show any increase.
In the wider environment of French Polynesia, the highest dose rates from the radionuclides being dispersed from Mururoa and Fangataufa are predicted to be experienced by the residents of Tureia who have a high intake of seafoods. At present, the dose rate is estimated to be 0.03 µSv a-1; this will decline over the next few thousand years until the peak release of plutonium from the underground sources produces an increase to 0.008 µSv a-1 at about 5000 years into the future.
For the postulated release of plutonium following a carbonate rock slide, Tureia would again be the site of the highest estimated radiation exposures - <7 µSv a-1 in the first year and declining rapidly thereafter. At more distant sites, the processes of transport and dispersion would delay and reduce the peak exposure. At Tahiti, for example, the peak exposure has been estimated to be 0.41 µSv a-1 in the third year after the rock slide. During a period of glaciation, assumed to be 50000 years into the future, the dose rates from plutonium via the inhalation and drinking water routes have been estimated to total about 1 µSv a-1; if potable water were to be extracted from the immediate vicinity of a safety trial cavity, much higher dose rates would be possible, but this was considered to be a very improbable scenario (McEwan, New Zealand).
Impacts on marine biota
The current levels of radionuclides in the various compartments of the marine environment were used to estimate the radiation exposures of a range of marine organisms. The highest dose rates to lagoon organisms in general were estimated to be about 0.2 µGy h-1 for zooplankton and benthic crustaceans from internal sources of α-radiation; the exposure of most other organisms was at least a factor of 10 lower. Overall, it appeared that the incremental radiation exposures of the majority of the marine organisms were less than, or of the same order as, that from the natural background. In the immediate vicinity of the sites of the barge tests, the γ-ray exposure of benthic organisms from the contaminated sediment is estimated to be up to µGy h-1, i.e., about 5 times higher than the previous highest estimate of the natural background dose rate from sediments. These dose rates are, however, lower than those that could cause any effects at the population level.
In the specific case of the particulate plutonium contamination of the sediments adjacent to the Colette motu, it is estimated that high doses of α-radiation could be delivered to very small volumes of either the surface tissues of benthic in-fauna or to the gut lining in the case of particle ingestion. The total doses to the small tissue volumes could be several gray (or even more, depending on the movements of the animal), and such as to cause cell killing. The consequences of this for the individual organisms are unknown, but it is unlikely that there would be any impact at the population level (Woodhead, UK).
Conclusions
Overall, the assessments have shown that the exposures from the residual radioactivity at the atolls from the nuclear testing programme, both now and in the future, are small fractions of that from the local natural background. This assessment was taken as the basis of the final conclusion:
`that there will be no radiation health effects which could be medically diagnosed in an individual or epidemiologically discerned in a group of people and which would be attributable to the estimated radiation doses that are now being received or that would be received in the future by people as a result of the residual radioactive material at Mururoa and Fangataufa Atolls.'
It was also concluded that the magnitude of the predicted dose rates was such that no remedial action is needed on radiological protection grounds, either now or in the future. It was continually emphasised that these conclusions concerning the predicted minimal radiological consequences of the French underground nuclear testing programme at Mururoa and Fangataufa were not to be considered in any way a mitigating argument in respect of any possible future programme of nuclear testing.
Notwithstanding the conclusions of the Study, the French authorities are proposing to maintain a limited environmental programme of radiological monitoring to acquire data to validate the predictive assessment. In addition, the geomechanical stability of the NE sector of Mururoa and the flanks of both atolls will be subject to continuous remote monitoring with daily data being transmitted to base in France (Sornein, France). This programme, together with the Study Reports, may go some way to reassure the population of French Polynesia who, if the perceptive questions of the representative of the South Pacific Forum (Shorten) provide any indication, remain concerned about the realities of the situation.
This meeting report can do no more than give the briefest glimpse of the work that was carried out in the assessment. In addition to the proceedings of the meeting, the resulting documentation extends to an Executive Summary, a Summary Report, the Main Study Report and six volumes of supporting technical material - in all, a total of some 1350 pages - that will be published by the IAEA in the near future.
Dennis Woodhead