- Email: [email protected]
Transport of radionuclides from the Mururoa and Fangataufa atolls through the marine environment
The Science of the Total Environment 237r238 Ž1999. 301]309
Transport of radionuclides from the Mururoa and Fangataufa atolls through the marine environment Ekkehard Mittelstaedt a,U , Iolanda Osvath b, Pavel P. Povinec b, Orihiko Togawab, E. Marian Scott c b
a Bundesamt fur ¨ Seeschiffahrt und Hydrographie, P.B. 301220, D-20305 Hamburg, Germany International Atomic Energy Agencyr Marine En¨ ironment Laboratory, 4 Quai Antoine 1er, B.P. 800, Monte Carlo 98012, Monaco c Department of Statistics, Uni¨ ersity of Glasgow, Glasgow G12 8QW, UK
Abstract A dispersion of radionuclides Ž 3 H, 90 Sr, 137Cs, 239 Pu. potentially released from the Mururoa and Fangataufa atolls through the South Pacific Ocean has been studied by means of computer models. The models used consisted of three differently structured compartmental models for the regional field and a hydrodynamic world ocean model for the far-field simulations. The outcome of regional modelling is predicted activity concentrations with time in different regions of French Polynesia Žover up to 10 000 years for plutonium.. The far-field model simulates large-scale dispersion in the South Pacific Ocean over periods of up to 50 years. The overall result suggests that there will not be radioactive contamination of any radiological interest at inhabited sites in French Polynesia or anywhere else in the ocean at present or in the future. Q 1999 Elsevier Science B.V. All rights reserved. Keywords: French Polynesia; Mururoa; Fangataufa; Marine modelling; Dispersion of radionuclides
1. Introduction In the framework of the IAEA study on ‘The Radiological Situation at the Atolls of Mururoa
U
Corresponding author. Tel.: q49-40-3190-3220; fax : q493190-5032. E-mail address: [email protected] ŽE. Mittelstaedt .
and Fangataufa’, the dispersion of radionuclides Ž 3 H, 90 Sr, 137 Cs, 239 Pu. potentially released from the atolls through the South Pacific Ocean has been studied by means of computer models ŽIAEA, 1998.. Mururoa and Fangataufa are two small neighbouring atolls at about 228S and 1398W in French Polynesia ŽFig. 1.. Both atolls rise just above the surface of the deep ocean. The question regarding ocean modelling in this project
0048-9697r99r$ - see front matter Q 1999 Elsevier Science B.V. All rights reserved. PII: S 0 0 4 8 - 9 6 9 7 Ž 9 9 . 0 0 1 4 4 - 8
302
E. Mittelstaedt et al. r The Science of the Total En¨ ironment 237 r 238 (1999) 301]309
Fig. 1. Location of French Polynesia, U MururoarFangataufa.
was: do the model results of radioactive concentrations in the ocean give rise to concern over the effects on human health and marine biota? Three classes of models were considered: 1. the near-field Žthe lagoons.; 2. the regional-fields Žbroadly French Polynesia.; and 3. the far-field Žthe South Pacific Ocean beyond the regional field.. The near-field models ŽIAEA, 1998; Tartinville, 1998. are hydrodynamic models describing the circulation and mixing in the lagoons of Mururoa and Fangataufa. The models were used to estimate the radioactive concentrations in lagoon water for given radionuclide releases and the flow rates of radionuclides from the lagoons into the surrounding ocean. Radionuclides in the lagoons arise from leaching of radioactivity deposited in lagoon sediments and potential migration from underground radioactive sources. The models
predict an average turnover time of lagoon water of 98 " 37 days for Mururoa and 33 " 12 days for Fangataufa. Another model has been developed for modelling the outflow of sediment from the Mururoa Lagoon suggesting that about 80 000 tons of sediment is removed per year during normal weather conditions and approximately 4 = 10 6 tons per storm ŽRajar and Zagar, 1999.. This paper concentrates on the description of results obtained by regional and far-field modelling.
2. Approach The regional models are compartmental models using hydrodynamic model data as input for their simulations. Three different models were applied. The models differ from each other regarding their structure Žspatial resolution., model domain Žmodel extension., and input data. The reason for running all three models was to obtain
E. Mittelstaedt et al. r The Science of the Total En¨ ironment 237 r 238 (1999) 301]309
303
Fig. 2. Mean concentrations of 3 H in seawater in the upper layer Ž0]450 m. at various inhabited islands for a constant release of 6000 GBqrannum. over 30 years.
an indication of the uncertainty of the results by intercomparing the final simulated concentration values. The input flow data Žannual means. were taken from the large-scale Ocean General Circulation Model ŽOGCM. by Masumoto and Yamagata Ž1996. and from the World Ocean Model developed at the Max-Planck-Institut for Meteorology in Hamburg ŽSegschneider, 1996.. Regarding the regional models, it has been assumed that a radioactive release at Mururoa and Fangataufa enters the ocean through the surface layer of the lagoons, whose thicknesses differ for each model Žranging from surface to 450 m.. No attempt was made to differentiate between Mururoa and Fangataufa as separate sources, since on an oceanic scale, they are so close as to be indistinguishable. The model used for far-field simulations is a world ocean circulation model combined with a tracer dispersion model ŽSegschneider, 1996.. The model was originally developed to study global climate changes. It is a multi-layered prognostic OGCM and has a spatial resolution of 3.5= 3.58. The circulation model provides the input data for the Lagrangian Transport Model Žtracer dispersion model.. The horizontal and vertical spreading of the tracers considered, due to the largescale circulation and subgrid-scale mixing, is
Fig. 3. 3 H concentrations in seawater Žyearly means, depthaveraged over 0]450 m. at selected locations after an instantaneous release of 1 PBq of 3 H at Mururoa into the model surface layer: disruptive release scenario.
304
E. Mittelstaedt et al. r The Science of the Total En¨ ironment 237 r 238 (1999) 301]309
treated by means of Lagrangian particle tracking combined with a Monte-Carlo method. The model is forced by standard boundary conditions. Two scenarios were envisaged: 1. the total activity released enters the ocean through the lagoon; and 2. the total activity released enters the ocean sideways at subsurface depths through the karst layer Žthe karst represents a potential conducting medium for radionuclides coming from the volcanic formations down below.. Moreover, three different time scenarios of a release were considered for the regional and farfield simulations 1. a constant release of radionuclides; 2. an instantaneous release of a large amount of activity caused by a sudden landslide spilling
into the ocean Žthe disruptive release scenario.; and 3. a time-dependent release of slowly upwards migrating radionuclides from the deep underground where most of the tests had taken place. The transport through the rocky geosphere was modelled in the framework of the above-mentioned IAEA study ŽIAEA, 1998.. The results on the final releases of radionuclides to the ocean were then provided to the marine modellers in order to simulate their spreading in the ocean. Scavenging has been considered in the compartment modelling. But even for plutonium with relatively high K d Ža distribution coefficient for sediments. values, scavenging is fairly small because of the clear open ocean waters. The regional model outputs are presentations of radionuclide concentrations Ž 3 H, 90 Sr, 137 Cs, 239 Pu. in sea
Fig. 4. Predicted time-dependent release rates of 3 H, 90 Sr, 137 Cs, 239 Pu to the biosphere Žstart: 1980..
E. Mittelstaedt et al. r The Science of the Total En¨ ironment 237 r 238 (1999) 301]309
water at various inhabited sites mainly in French Polynesia and further abroad. The simulation time depends on the radionuclide considered and ex-
305
tends from 30 to 100 years for relatively short-lived nuclides and up to 10 000 years for plutonium. The far-field model results refer to simulations
Fig. 5. 239 Pu concentrations in seawater Žyearly means, averaged over 0]450-m depth. at selected locations: time-dependent release scenario.
E. Mittelstaedt et al. r The Science of the Total En¨ ironment 237 r 238 (1999) 301]309
306
at surface and at depth over time-scales over 50 years. The sites at which concentrations are presented are partly the same as those for the regional models. However, a few sites are at greater distances beyond French Polynesia. The more relevant outputs of the far-field model are largescale distribution patterns of radionuclide concentrations over maximum simulation periods of 50 years at the surface Ž0]50 m. and at depth Ž350]450 m..
3. Results 3.1. Regional modelling Fig. 2 shows simulated concentrations of 3 H in the ocean at various sites of French Polynesia due to a constant release of 6000 GBqrannum at Mururoa over 30 years. The figure suggests that once the release has started, a rapid increase in concentration takes place towards an equilibrium concentration within a few years. The highest concentration Ž0.15 Bqrm3 . emerges in this case at Tureia approximately 140 km to the north of Mururoa. Concentrations at the other sites are lower due to the increasing distance from the source. Around Tahiti, approximately 1200 km from Mururoa, the concentrations reach only 10% of those around Tureia. At no time during the simulations the concentrations reached the pre-
sent regional background level of 3 H in the surface layer Ž100]200 Bqrm3 .. The simulation curves for other radionuclides show similar shapes. The simulation of a landslide assumes an instantaneous release of radionuclides into the ocean. Fig. 3 illustrates the change with time for an annual depth-averaged concentration Žbetween 0 and 450 m. of 3 H up to 100 years at various selected sites after the sudden release of 1 PBq. The shapes of the curves are similar for all the radionuclides considered in this scenario. Differences are mainly due to differences in halflife. The highest concentration Žabove 1 Bqrm3 . occurs at Tureia. The only case indicating a distinct elevation of the concentration above the existing background level due to fallout is associated with a major release of plutonium into the ocean as consequence of the assumed landslide. Simulations of this case indicate that plutonium concentrations at Tureia would rise to 100 times the present background level but would fall again quickly within a few years. The time-dependent source terms are the sum of the underground releases, and contributions from sediment leaching. The releases via the lagoon are surface releases; releases to the ocean Ž‘direct’. are assumed to enter the ocean at a subsurface depth of 400 m. Fig. 4 shows that the total release rates of 3 H, 90 Sr and 137Cs will
Table 1 Maximum concentrations ŽBqrm3 . for the time-dependent source term at various sites Žregional models. Nuclide
Tureia
Tematangi
Hao
Tahiti
Tubuai
Rapa
Model 1 0]450 m
3
H Sr 137 Cs Pu
6 8 = 10y3 2 = 10y3 8 = 10y4
4 5 = 10y3 1 = 10y3 5 = 10y4
1 2 = 10y3 4 = 10y4 2 = 10y4
0.6 9 = 10y4 2 = 10y4 8 = 10y5
0.06 1 = 10y4 2 = 10y5 9 = 10y6
] ] ] ]
Model 2 0]450 m
3
H Sr 137 Cs Pu
2 3 = 10y3 5 = 10y4 2 = 10y4
4 = 10y1 6 = 10y4 1 = 10y4 5 = 10y5
2 = 10y1 2 = 10y4 5 = 10y5 2 = 10y5
1 = 10y1 2 = 10y4 4 = 10y5 2 = 10y5
1 = 10y1 2 = 10y4 3 = 10y5 1 = 10y5
4 = 10y2 7 = 10y5 1 = 10y5 4 = 10y6
Model 3 0]50 m
3
5 2 = 10y3 7 = 10y4
6 2 = 10y3 7 = 10y4
8 1 = 10y4 5 = 10y5
0.2 1 = 10y4 4 = 10y5
1 4 = 10y4 2 = 10y4
0.8 3 = 10y4 1 = 10y4
90
90
H Sr 137 Cs 90
E. Mittelstaedt et al. r The Science of the Total En¨ ironment 237 r 238 (1999) 301]309
decrease with time, with the dominant contribution from underground sources coming from releases directly into the ocean. In the case of plutonium, which will migrate extremely slowly from underground Žhigh K d ., the dominant source is lagoon sediment, and for a few tens of years, the release rates will decrease with a half-life of approximately 10 years. The contribution of 239 Pu Žhalf-life approximately 24 000 years. is dominant on timescales of the order of 10 000 years. Using the source terms of predicted time-dependent releases suggests that the model esti-
307
mates are very low and several orders of magnitude below present background levels. Simulations over 10 000 years of the local 239 Pu concentration variations exhibit two peaks Žaccording to the prescribed source term.. The first one occurs 5]8 years after the release of plutonium has started. The second peak is predicted to occur around AD 8000 ŽFig. 5.. In each case maximum concentrations stay below present-day regional background levels of plutonium Ž1]4 mBqrm3 .. Table 1 represents an overview of the various regional model results for a time-dependent
Fig. 6. Activity concentration 10 years after the time-dependent continuous release of plutonium has started into the surface layer Žabove. and at the depth of 400 m Žbelow.. Isolines: 1, 10, 100, 250, 500, 750, 1000, 2000 Ž10y4 mBqrm3 ..
308
E. Mittelstaedt et al. r The Science of the Total En¨ ironment 237 r 238 (1999) 301]309
source term regarding maximum activity concentrations within the ocean model layer Žgiven in the first column of the table. at different sites. The intercomparison of the three regional models with respect to maximum concentrations vary largely within an order of magnitude. All values stay, however, distinctly below present day background levels in the ocean surface layer. 3.2. Far-field modelling With respect to the far-field modelling two
simulations are picked here referring to the spread of plutonium due to a time-dependent release. Fig. 6 suggests the simulated large-scale pattern of the plutonium concentration 10 years after the time-dependent release has started. The release takes place at the surface Ž25 m. and at subsurface depth Ž400 m., respectively. The subsurface release is assumed to arise from the highly permeable karstic channels below 200-m depth, which according to the model predictions of radionuclide transport through the solid geosphere ŽIAEA, 1998. are most likely the main outlet to
Fig. 7. Activity concentration 50 years after the time-dependent continuous release of plutonium has started into the surface layer Žabove. and at the depth of 400 m Žbelow.. Isolines: 1, 10, 100 and in the surface layer additionally 250, 500, 750, 1000 Ž10y4 mBqrm3 ..
E. Mittelstaedt et al. r The Science of the Total En¨ ironment 237 r 238 (1999) 301]309
the ocean for radioactivity coming from underground. Within the stratified thermocline layer between 100 and 500 m, subsurface flow and mixing are relatively weak. A radioactive injection released within this layer spreads and dilutes slower than the same release into the surface layer. After 10 years the contamination has moved with the general circulation in an easterly direction in the surface layer. In the subsurface layer it has approached the Australian continent. A striking feature of the model results is the concentration of the radionuclides around the region 1108W, 308S near the Easter Islands east of Mururoa in the surface layer. This is due to the kinematics Žconvergence region. and causes long-term trapping of potential contamination associated in this case with the absolute maximum of simulated concentrations Ž0.2 mBqrm3 .. This effect may be supported by the predicted plutonium source term Žtime function.,which reduces after the first decade ŽFig. 4.. Thus, maximum concentrations do not occur anymore around MururoarFangataufa but have moved to the area around the Easter Islands in the surface layer after 10 years of release. Fifty years after the beginning of the release contamination is widely dispersed in the ocean and still concentrates around the Easter Islands in the surface layer ŽFig. 7.. Also at 400-m depth the highest concentrations of the continuous time-dependent plutonium injection have propagated away from the source. The concentration pattern distinctly mirrors the various principal ocean circulation branches in this region Ži.e. subsurface westward current, East Australian current moving south, South Pacific currents moving east, Equatorial undercurrent and North Equa-
309
torial Counter Current moving east.. According to the simulations, maximum surface and subsurface concentrations Ž0.1 mBqrm3 . are strongly diluted and far below the present background levels. 4. Conclusion Model results on the dispersion of radionuclides in the South Pacific Ocean, which might potentially emanate from Mururoa and Fangataufa, suggest that regional and far-field radionuclide concentrations in ocean water attributable to this source will be distinctly below present background levels. Acknowledgements We thank the Max Planck Institute for Meteorology for making available the world ocean model and especially Dr. J. Segschneider, ŽECMWF. for running the large scale dispersion simulations by means of this model. References IAEA. The radiological situation at the atolls of Mururoa and Fangataufa. Vienna: IAEA, 1998. Masumoto Y, Yamagata T. Seasonal variations of the Indonesian throughflow in a general ocean circulation model. J Geophys Res 1996;101:12287]12293. Rajar R, Zagar D. Modelling the transport of sediments and plutonium from the Mururoa lagoon. Vienna: IAEA, 1999:in press. Segschneider J. Zur Ausbreitung geochemischer Spurenstoffe in der Tiefsee. Berichte aus dem ZMK. Reihe B 1996;24:95. Tartinville B. Modelisation tridimensionelle de la circulation dans le lagon de l’atoll de Mururoa, Polynesie francaise. Dissertation Universite Catholique de Louvain, 1998:184.
Transport of radionuclides from the Mururoa and Fangataufa atolls through the marine environment Ekkehard Mittelstaedt a,U , Iolanda Osvath b, Pavel P. Povinec b, Orihiko Togawab, E. Marian Scott c b
a Bundesamt fur ¨ Seeschiffahrt und Hydrographie, P.B. 301220, D-20305 Hamburg, Germany International Atomic Energy Agencyr Marine En¨ ironment Laboratory, 4 Quai Antoine 1er, B.P. 800, Monte Carlo 98012, Monaco c Department of Statistics, Uni¨ ersity of Glasgow, Glasgow G12 8QW, UK
Abstract A dispersion of radionuclides Ž 3 H, 90 Sr, 137Cs, 239 Pu. potentially released from the Mururoa and Fangataufa atolls through the South Pacific Ocean has been studied by means of computer models. The models used consisted of three differently structured compartmental models for the regional field and a hydrodynamic world ocean model for the far-field simulations. The outcome of regional modelling is predicted activity concentrations with time in different regions of French Polynesia Žover up to 10 000 years for plutonium.. The far-field model simulates large-scale dispersion in the South Pacific Ocean over periods of up to 50 years. The overall result suggests that there will not be radioactive contamination of any radiological interest at inhabited sites in French Polynesia or anywhere else in the ocean at present or in the future. Q 1999 Elsevier Science B.V. All rights reserved. Keywords: French Polynesia; Mururoa; Fangataufa; Marine modelling; Dispersion of radionuclides
1. Introduction In the framework of the IAEA study on ‘The Radiological Situation at the Atolls of Mururoa
U
Corresponding author. Tel.: q49-40-3190-3220; fax : q493190-5032. E-mail address: [email protected] ŽE. Mittelstaedt .
and Fangataufa’, the dispersion of radionuclides Ž 3 H, 90 Sr, 137 Cs, 239 Pu. potentially released from the atolls through the South Pacific Ocean has been studied by means of computer models ŽIAEA, 1998.. Mururoa and Fangataufa are two small neighbouring atolls at about 228S and 1398W in French Polynesia ŽFig. 1.. Both atolls rise just above the surface of the deep ocean. The question regarding ocean modelling in this project
0048-9697r99r$ - see front matter Q 1999 Elsevier Science B.V. All rights reserved. PII: S 0 0 4 8 - 9 6 9 7 Ž 9 9 . 0 0 1 4 4 - 8
302
E. Mittelstaedt et al. r The Science of the Total En¨ ironment 237 r 238 (1999) 301]309
Fig. 1. Location of French Polynesia, U MururoarFangataufa.
was: do the model results of radioactive concentrations in the ocean give rise to concern over the effects on human health and marine biota? Three classes of models were considered: 1. the near-field Žthe lagoons.; 2. the regional-fields Žbroadly French Polynesia.; and 3. the far-field Žthe South Pacific Ocean beyond the regional field.. The near-field models ŽIAEA, 1998; Tartinville, 1998. are hydrodynamic models describing the circulation and mixing in the lagoons of Mururoa and Fangataufa. The models were used to estimate the radioactive concentrations in lagoon water for given radionuclide releases and the flow rates of radionuclides from the lagoons into the surrounding ocean. Radionuclides in the lagoons arise from leaching of radioactivity deposited in lagoon sediments and potential migration from underground radioactive sources. The models
predict an average turnover time of lagoon water of 98 " 37 days for Mururoa and 33 " 12 days for Fangataufa. Another model has been developed for modelling the outflow of sediment from the Mururoa Lagoon suggesting that about 80 000 tons of sediment is removed per year during normal weather conditions and approximately 4 = 10 6 tons per storm ŽRajar and Zagar, 1999.. This paper concentrates on the description of results obtained by regional and far-field modelling.
2. Approach The regional models are compartmental models using hydrodynamic model data as input for their simulations. Three different models were applied. The models differ from each other regarding their structure Žspatial resolution., model domain Žmodel extension., and input data. The reason for running all three models was to obtain
E. Mittelstaedt et al. r The Science of the Total En¨ ironment 237 r 238 (1999) 301]309
303
Fig. 2. Mean concentrations of 3 H in seawater in the upper layer Ž0]450 m. at various inhabited islands for a constant release of 6000 GBqrannum. over 30 years.
an indication of the uncertainty of the results by intercomparing the final simulated concentration values. The input flow data Žannual means. were taken from the large-scale Ocean General Circulation Model ŽOGCM. by Masumoto and Yamagata Ž1996. and from the World Ocean Model developed at the Max-Planck-Institut for Meteorology in Hamburg ŽSegschneider, 1996.. Regarding the regional models, it has been assumed that a radioactive release at Mururoa and Fangataufa enters the ocean through the surface layer of the lagoons, whose thicknesses differ for each model Žranging from surface to 450 m.. No attempt was made to differentiate between Mururoa and Fangataufa as separate sources, since on an oceanic scale, they are so close as to be indistinguishable. The model used for far-field simulations is a world ocean circulation model combined with a tracer dispersion model ŽSegschneider, 1996.. The model was originally developed to study global climate changes. It is a multi-layered prognostic OGCM and has a spatial resolution of 3.5= 3.58. The circulation model provides the input data for the Lagrangian Transport Model Žtracer dispersion model.. The horizontal and vertical spreading of the tracers considered, due to the largescale circulation and subgrid-scale mixing, is
Fig. 3. 3 H concentrations in seawater Žyearly means, depthaveraged over 0]450 m. at selected locations after an instantaneous release of 1 PBq of 3 H at Mururoa into the model surface layer: disruptive release scenario.
304
E. Mittelstaedt et al. r The Science of the Total En¨ ironment 237 r 238 (1999) 301]309
treated by means of Lagrangian particle tracking combined with a Monte-Carlo method. The model is forced by standard boundary conditions. Two scenarios were envisaged: 1. the total activity released enters the ocean through the lagoon; and 2. the total activity released enters the ocean sideways at subsurface depths through the karst layer Žthe karst represents a potential conducting medium for radionuclides coming from the volcanic formations down below.. Moreover, three different time scenarios of a release were considered for the regional and farfield simulations 1. a constant release of radionuclides; 2. an instantaneous release of a large amount of activity caused by a sudden landslide spilling
into the ocean Žthe disruptive release scenario.; and 3. a time-dependent release of slowly upwards migrating radionuclides from the deep underground where most of the tests had taken place. The transport through the rocky geosphere was modelled in the framework of the above-mentioned IAEA study ŽIAEA, 1998.. The results on the final releases of radionuclides to the ocean were then provided to the marine modellers in order to simulate their spreading in the ocean. Scavenging has been considered in the compartment modelling. But even for plutonium with relatively high K d Ža distribution coefficient for sediments. values, scavenging is fairly small because of the clear open ocean waters. The regional model outputs are presentations of radionuclide concentrations Ž 3 H, 90 Sr, 137 Cs, 239 Pu. in sea
Fig. 4. Predicted time-dependent release rates of 3 H, 90 Sr, 137 Cs, 239 Pu to the biosphere Žstart: 1980..
E. Mittelstaedt et al. r The Science of the Total En¨ ironment 237 r 238 (1999) 301]309
water at various inhabited sites mainly in French Polynesia and further abroad. The simulation time depends on the radionuclide considered and ex-
305
tends from 30 to 100 years for relatively short-lived nuclides and up to 10 000 years for plutonium. The far-field model results refer to simulations
Fig. 5. 239 Pu concentrations in seawater Žyearly means, averaged over 0]450-m depth. at selected locations: time-dependent release scenario.
E. Mittelstaedt et al. r The Science of the Total En¨ ironment 237 r 238 (1999) 301]309
306
at surface and at depth over time-scales over 50 years. The sites at which concentrations are presented are partly the same as those for the regional models. However, a few sites are at greater distances beyond French Polynesia. The more relevant outputs of the far-field model are largescale distribution patterns of radionuclide concentrations over maximum simulation periods of 50 years at the surface Ž0]50 m. and at depth Ž350]450 m..
3. Results 3.1. Regional modelling Fig. 2 shows simulated concentrations of 3 H in the ocean at various sites of French Polynesia due to a constant release of 6000 GBqrannum at Mururoa over 30 years. The figure suggests that once the release has started, a rapid increase in concentration takes place towards an equilibrium concentration within a few years. The highest concentration Ž0.15 Bqrm3 . emerges in this case at Tureia approximately 140 km to the north of Mururoa. Concentrations at the other sites are lower due to the increasing distance from the source. Around Tahiti, approximately 1200 km from Mururoa, the concentrations reach only 10% of those around Tureia. At no time during the simulations the concentrations reached the pre-
sent regional background level of 3 H in the surface layer Ž100]200 Bqrm3 .. The simulation curves for other radionuclides show similar shapes. The simulation of a landslide assumes an instantaneous release of radionuclides into the ocean. Fig. 3 illustrates the change with time for an annual depth-averaged concentration Žbetween 0 and 450 m. of 3 H up to 100 years at various selected sites after the sudden release of 1 PBq. The shapes of the curves are similar for all the radionuclides considered in this scenario. Differences are mainly due to differences in halflife. The highest concentration Žabove 1 Bqrm3 . occurs at Tureia. The only case indicating a distinct elevation of the concentration above the existing background level due to fallout is associated with a major release of plutonium into the ocean as consequence of the assumed landslide. Simulations of this case indicate that plutonium concentrations at Tureia would rise to 100 times the present background level but would fall again quickly within a few years. The time-dependent source terms are the sum of the underground releases, and contributions from sediment leaching. The releases via the lagoon are surface releases; releases to the ocean Ž‘direct’. are assumed to enter the ocean at a subsurface depth of 400 m. Fig. 4 shows that the total release rates of 3 H, 90 Sr and 137Cs will
Table 1 Maximum concentrations ŽBqrm3 . for the time-dependent source term at various sites Žregional models. Nuclide
Tureia
Tematangi
Hao
Tahiti
Tubuai
Rapa
Model 1 0]450 m
3
H Sr 137 Cs Pu
6 8 = 10y3 2 = 10y3 8 = 10y4
4 5 = 10y3 1 = 10y3 5 = 10y4
1 2 = 10y3 4 = 10y4 2 = 10y4
0.6 9 = 10y4 2 = 10y4 8 = 10y5
0.06 1 = 10y4 2 = 10y5 9 = 10y6
] ] ] ]
Model 2 0]450 m
3
H Sr 137 Cs Pu
2 3 = 10y3 5 = 10y4 2 = 10y4
4 = 10y1 6 = 10y4 1 = 10y4 5 = 10y5
2 = 10y1 2 = 10y4 5 = 10y5 2 = 10y5
1 = 10y1 2 = 10y4 4 = 10y5 2 = 10y5
1 = 10y1 2 = 10y4 3 = 10y5 1 = 10y5
4 = 10y2 7 = 10y5 1 = 10y5 4 = 10y6
Model 3 0]50 m
3
5 2 = 10y3 7 = 10y4
6 2 = 10y3 7 = 10y4
8 1 = 10y4 5 = 10y5
0.2 1 = 10y4 4 = 10y5
1 4 = 10y4 2 = 10y4
0.8 3 = 10y4 1 = 10y4
90
90
H Sr 137 Cs 90
E. Mittelstaedt et al. r The Science of the Total En¨ ironment 237 r 238 (1999) 301]309
decrease with time, with the dominant contribution from underground sources coming from releases directly into the ocean. In the case of plutonium, which will migrate extremely slowly from underground Žhigh K d ., the dominant source is lagoon sediment, and for a few tens of years, the release rates will decrease with a half-life of approximately 10 years. The contribution of 239 Pu Žhalf-life approximately 24 000 years. is dominant on timescales of the order of 10 000 years. Using the source terms of predicted time-dependent releases suggests that the model esti-
307
mates are very low and several orders of magnitude below present background levels. Simulations over 10 000 years of the local 239 Pu concentration variations exhibit two peaks Žaccording to the prescribed source term.. The first one occurs 5]8 years after the release of plutonium has started. The second peak is predicted to occur around AD 8000 ŽFig. 5.. In each case maximum concentrations stay below present-day regional background levels of plutonium Ž1]4 mBqrm3 .. Table 1 represents an overview of the various regional model results for a time-dependent
Fig. 6. Activity concentration 10 years after the time-dependent continuous release of plutonium has started into the surface layer Žabove. and at the depth of 400 m Žbelow.. Isolines: 1, 10, 100, 250, 500, 750, 1000, 2000 Ž10y4 mBqrm3 ..
308
E. Mittelstaedt et al. r The Science of the Total En¨ ironment 237 r 238 (1999) 301]309
source term regarding maximum activity concentrations within the ocean model layer Žgiven in the first column of the table. at different sites. The intercomparison of the three regional models with respect to maximum concentrations vary largely within an order of magnitude. All values stay, however, distinctly below present day background levels in the ocean surface layer. 3.2. Far-field modelling With respect to the far-field modelling two
simulations are picked here referring to the spread of plutonium due to a time-dependent release. Fig. 6 suggests the simulated large-scale pattern of the plutonium concentration 10 years after the time-dependent release has started. The release takes place at the surface Ž25 m. and at subsurface depth Ž400 m., respectively. The subsurface release is assumed to arise from the highly permeable karstic channels below 200-m depth, which according to the model predictions of radionuclide transport through the solid geosphere ŽIAEA, 1998. are most likely the main outlet to
Fig. 7. Activity concentration 50 years after the time-dependent continuous release of plutonium has started into the surface layer Žabove. and at the depth of 400 m Žbelow.. Isolines: 1, 10, 100 and in the surface layer additionally 250, 500, 750, 1000 Ž10y4 mBqrm3 ..
E. Mittelstaedt et al. r The Science of the Total En¨ ironment 237 r 238 (1999) 301]309
the ocean for radioactivity coming from underground. Within the stratified thermocline layer between 100 and 500 m, subsurface flow and mixing are relatively weak. A radioactive injection released within this layer spreads and dilutes slower than the same release into the surface layer. After 10 years the contamination has moved with the general circulation in an easterly direction in the surface layer. In the subsurface layer it has approached the Australian continent. A striking feature of the model results is the concentration of the radionuclides around the region 1108W, 308S near the Easter Islands east of Mururoa in the surface layer. This is due to the kinematics Žconvergence region. and causes long-term trapping of potential contamination associated in this case with the absolute maximum of simulated concentrations Ž0.2 mBqrm3 .. This effect may be supported by the predicted plutonium source term Žtime function.,which reduces after the first decade ŽFig. 4.. Thus, maximum concentrations do not occur anymore around MururoarFangataufa but have moved to the area around the Easter Islands in the surface layer after 10 years of release. Fifty years after the beginning of the release contamination is widely dispersed in the ocean and still concentrates around the Easter Islands in the surface layer ŽFig. 7.. Also at 400-m depth the highest concentrations of the continuous time-dependent plutonium injection have propagated away from the source. The concentration pattern distinctly mirrors the various principal ocean circulation branches in this region Ži.e. subsurface westward current, East Australian current moving south, South Pacific currents moving east, Equatorial undercurrent and North Equa-
309
torial Counter Current moving east.. According to the simulations, maximum surface and subsurface concentrations Ž0.1 mBqrm3 . are strongly diluted and far below the present background levels. 4. Conclusion Model results on the dispersion of radionuclides in the South Pacific Ocean, which might potentially emanate from Mururoa and Fangataufa, suggest that regional and far-field radionuclide concentrations in ocean water attributable to this source will be distinctly below present background levels. Acknowledgements We thank the Max Planck Institute for Meteorology for making available the world ocean model and especially Dr. J. Segschneider, ŽECMWF. for running the large scale dispersion simulations by means of this model. References IAEA. The radiological situation at the atolls of Mururoa and Fangataufa. Vienna: IAEA, 1998. Masumoto Y, Yamagata T. Seasonal variations of the Indonesian throughflow in a general ocean circulation model. J Geophys Res 1996;101:12287]12293. Rajar R, Zagar D. Modelling the transport of sediments and plutonium from the Mururoa lagoon. Vienna: IAEA, 1999:in press. Segschneider J. Zur Ausbreitung geochemischer Spurenstoffe in der Tiefsee. Berichte aus dem ZMK. Reihe B 1996;24:95. Tartinville B. Modelisation tridimensionelle de la circulation dans le lagon de l’atoll de Mururoa, Polynesie francaise. Dissertation Universite Catholique de Louvain, 1998:184.