An update on EU doses3 November 2002
A new report commissioned by the EC examines the radiological impact of the European Union's nuclear facilities on the population. The data is compromised, however, by variations in data reporting.
In 1995, a report on "Radioactive effluents from nuclear power stations and nuclear fuel reprocessing plants in the European Community 1977-86" was published by the European Commission (EC). It included records of discharge data and an assessment of the resulting individual and collective dose.
Since then, a revised and updated methodology for assessing the radiological consequences of routine discharges has been published and implemented as a computer program called PC CREAM. This program includes revisions to dosimetry and calculates effective dose as defined by the International Commission for Radiation Protection in its 1990 Recommendations (known as ICRP 60).
In 1999, a revised radiological impact assessment of routine discharges from European Union (EU) nuclear sites between 1987 and 1996 was undertaken on behalf of the EC's Directorate General Environment. The project was co-ordinated by the UK's National Radiological Protection Board, which also carried out an assessment of doses arising from atmospheric discharges. The Nuclear Research and Consultancy group of KEMA in the Netherlands (NRG) took responsibility for assessing doses from aquatic discharges while Germany's GRS provided additional data.
The study used the discharge database Bilcom97.mdb as its data source and it included all nuclear power plants above 50MWe in capacity and reprocessing plants, including those Finland, Sweden and the German Democratic Republic, which joined the EU after the study began. Because of the large number of nuclear sites involved - the number of reactors was 118 in 1987 and 164 in 1996 - the assessment was carried out for the three specific years 1987, 1991 and 1996. For the sites of greatest radiological significance, namely the reprocessing plants at Cap de la Hague and Sellafield, doses were calculated for each year from 1987 to 1996.
METHODS AND DATA
The report examined the effective dose to individuals and the collective dose to the exposed population of the EC. The effective dose is the sum of the annual external effective dose and the committed effective dose to adults for intakes over one year. The dose to the most-exposed members of the population (the critical group) is modelled by calculating the annual dose received in the 50th year, assuming equilibrium conditions apply and that the dose is continuous and constant over the 50 years. The collective dose is calculated by summing the annual individual doses to 500 years.
The quantity of radionuclide discharged in liquid and airborne effluents from each nuclear site is available in the Bilcom97 database. The data is based on discharge reports from site operators and because methods of reporting vary there are inconsistencies in the data. In some cases, for example, aggregate discharges are recorded and assumptions had to be made on their radiological composition.
Airborne effluents from nuclear plants are generally H-3, C-14, noble gases (the largest component in terms of activity) and an aerosol component that includes a range of activation and fission products that can include actinides and gaseous halogens. Reprocessing plants have a wide spectrum of airborne discharges with Kr-85 and H-3 being highest in terms of activity.
In liquid effluents the most active component by one or two orders of magnitude is H-3. The GRS report identified the principal other components, which vary considerably from reactor to reactor. For reprocessing plants the liquid effluent is dominated by fission products such as Cs-137, Ru-106, Co-60 and Sr-90, and they contain significant quantities of actinides.
In addition to effluent data, PC CREAM requires information on the points of discharge, meteorological data, grids of population and agricultural consumption around each site and the habits of individuals.
To calculate the dose from liquid discharges the sites were split into "coastal" sites discharging directly into the marine environment and "inland" sites discharging into freshwater systems and being transported to the sea.
In the marine environment radionuclides are dispersed by currents and diffusion but can interact with suspended sediments. They can enter the aquatic food chain and contaminate foodstuffs consumed by man, or be dispersed to the terrestrial environment by seaspray. NRG used a compartmental marine model with generic and site-specific components to calculate the dispersion of radionuclides.
Inland sites were allocated to one of three modelled rivers - the Loire, Rhine and Rhone - to assess the dose. Ingestion doses can be important for river sites as the water is used for drinking and for irrigating crops. External dose from sediment is also considered.
The collective dose was calculated for the 377 million people in the EU area. It is not possible to calculate collective doses from discharges to rivers in PC CREAM, so collective doses from river water were considered in terms of exposures from the marine environment into which the river discharges.
There was a general reduction in discharges to river and sea in the period of the study, which means that in relative terms atmospheric discharges become more important towards the end of the period, rising from 48% to 88% of the whole.
Nuclear power stations accounted for 47% of the collective atmospheric dose in 1987 and 50% in 1996.
The estimated collective dose calculation had two components: data from the "first pass" of the radioactive plume and global summation of the dispersed radionuclides. Both components of the atmospheric dose have increased because of a change in reporting for UK plants, which required reporting of C-14 only after 1991.
For Sellafield there was little change in the estimate of collective dose. For Cap de la Hague there was a consistent increase in collective dose, primarily due to an increase in reported discharges of C-14, I-129 and Kr-85. The increase may be an artefact of the data, however, as the Bilcom97 database does not include C-14 and I-129 before 1992, so the dose may have been at post-1992 levels before 1992.
For other fuel cycle facilities, collective doses from Dounreay decreased from 5.8x10-2manSv in 1987 to 6.8x10-3manSv in 1996. Those at WAK, which is being decommissioned, dropped from 1.9x10-1manSv to 1.3x10-2man Sv by 1992, while for Marcoule doses have had to be estimated at 4.9manSv.
Individual doses are generally higher for UK sites, which discharge C-14, S-35 and Ar-41. The dose at these plants is also underestimated before 1992 because C-14 discharges were not reported until after 1991.
At Sellafield and Cap de la Hague the individual doses only amount to a few tens of microSieverts throughout the period. At Sellafield the most important nuclide is I-129 in milk, milk products and fruit. Ar-41 in the plume also makes a significant contribution, particularly at 0.5km distance. I-129 is also important at Cap La Hague but it is not recorded in the database prior to 1992. Before that time Kr-85 dominates the exposures. At Marcoule the halogen release is recorded: it is assumed that this is mostly made up of I-129 and again this dominates the exposure.
Total collective doses from Sellafield remained fairly steady until 1994 when it more than doubled to 10manSv. This was due to an increase in the reported discharges of C-14. A reduction in discharges of Ru-106 helped Cap de la Hague to a steady reduction in collective dose over the period.
For nuclear power plant sites the estimated collective dose decreased from 0.74manSv in 1987 to 0.13manSv in 1996. There are significant factors in the data in 1987 and 1991 from the Trawsfynydd plant in Wales, now shut down, which is unique in being sited on a lake. This cannot be modelled in PC CREAM. There are also significant exposures at Paluel and elsewhere from Ag-110, arising from the ingestion of seafood.
Individual doses at coastal sites generally are of the order of a few microsieverts, except for the following: Bradwell, where Cs-137 in fish contributed to a dose of 10µSv in 1996; Heysham 1, when Co-60 dominated a 10µSv release in 1991; and Paluel, where 70µSv was recorded in 187 due to Ag-110 in seafood.
The more significant exposures were due to reprocessing plants. At Cap de la Hague individual doses dropped steadily from 107µSv in 1987 to 19µSv in 1996. They are dominated by consumption of molluscs contaminated by Ru-106 and Pu-241, along with external exposure to gamma rays from Co-60 in sediments. C-14 is not recorded until late in the period but is expected to be around 10µSv in the period 1994-1996.
Around Sellafield individual doses have decreased from 187µSv in 1987 to 114µSv in 1996. The sources vary considerably over time but important sources include Tc-99 in crustaceans, C-14 and Cs-137 in fish, Ru-106 and Pu-241 in molluscs and external exposure to Co-60, Zn-65, Zr-95, Nb-95 and Eu-152.
Exposures at Dounreay are estimated at 15µSv dropping to 5µSv by 1996 and are dominated by Cs-137 in fish.
In calculating individual doses for inland sites three groups were considered: those living by the river; consumers of seafood; and individuals living around the river estuary.
The first group generally had doses of a few µSv. The second group had its highest doses due to power plants at Bugey in 1987 (20µSv), Chinon in 1987 (12µSv), Dampierre in 1987 (39µSv) and Le Blayais in 1987 (16µSv). Marcoule provided the highest dose at 300µSv in 1996.
The third group were in some cases identical to the second. In cases where the estuary was in a different country to the source they had doses between five and ten times lower.
SEEKING BETTER DATA
Overall the study found that the estimated collective dose of the EC population had increased by about 76% between 1987 and 1996, with the most important sources being the reprocessing plants at Cap de la Hague and Sellafield. The increase would be likely to appear less significant if C-14 discharges had been included in earlier assessments.
The authors of the report noted that while the Bilcom97 database was a "valuable tool" its weakness was that there was inconsistency in data reporting across sites. There were many cases in the study where assumptions had to be made regarding the radionuclide composition of aggregated discharges because breakdowns were missing or incomplete. To overcome this problem, they advised that a consistent method of reporting discharges needs to be adopted by operators.
The PC CREAM program was found to be an extremely useful tool for radiological impact assessments. But it had limitations, for example it cannot measure atmospheric dispersion from stacks higher than 100m. Pathways like the ingestion of terrestrial foods are not covered. A PC CREAM user group is discussing these and other issues.