The project MARIA (Methods for Assessing the Radiological Impact of Accidents) was initiated in 1983 as part of the CEC Radiation Protection Research Programme. The aim of the project was, firstly, to review and then to develop further methods and models for accident consequence assessment (ACA) within the EC. A European ACA code, COSYMA (COde SYstem from MAria), is now available as a result of the many investigations performed within the MARIA project.
There remain a number of areas within the field of ACA modelling where further development and refinements are required, in addition the COSYMA code and its databases also require maintenance and some further development. These are the aims of the current extension of the MARIA project under the 1990 to 1991 Radiation Protection Programme.
The work carried out under this topic is described in 4 sections:
the modelling of atmospheric dispersion and deposition;
deposition to skin;
the production of COSYMA (code system from MARIA (methods of assessing the radiological impact of accidents)).
The modelling of atmospheric dispersion and deposition:
The main objective of this work was to decide which sorts of atmospheric dispersion model are required in accident consequence assessments, and whether the models in general use at present are adequate. The work focussed on models for dispersion in conditions with low wind speeds, on different methods the spatial and temporal variation of rain, and the need to include trajectory models in accident consequence assessments (ACA).
The work done in MARIA-1 on meteorological sampling was carried out to determine an adequate sampling scheme for use with the National Radiological Protection Biraid's Computer program for accident consequence assessments (MARC) program. This used a Gaussian plume dispersion model with no allowance for changes of wind direction during plume travel. The work showed that the identification of rain sequences was important. This work has been continued, and a more refined system of identifying these sequences developed.
The importance of and modelling of deposition to skin:
Accident consequence programs consider early death from irradiation of several organs, including the skin and the bone marrow. The major contribution to skin doses comes from material deposited on the skin, while a major contributor to bone marrow doses is gamma exposure from material deposited on the ground. Recent changes to model parameter values (an increase in the shielding from deposited gamma dose assumed for people indoors, and a reduction of the dose to hill 50% (LD50) for death from skin exposure) have increased the relative importance of skin exposure compared to the exposure of other organs in term s of predicted numbers of early deaths. However there is very little information on the value to adopt for deposition velocity to skin.
This work showed that deposition to skin and the resulting skin doses are important, and should be considered in accident consequence assessments. However the current practice of calculating only the total deposition rather that the wet and dry components separately is adequate. The work also showed the need for research into radionuclide deposition onto skin and clothing.
The production of COSYMA:
One of the objectives of the MARIA programme is to develop a computer program system for assessing offsite consequences of accidental releases of radiactive material to atmosphere. The new program system, COSYMA, was developed jointly by NRPB and Kernforschungszentrum Karlsruhe (KFK). It is intended for use generally within the European Community (EC) by people who are not specialists in the area of accident consequence modelling but wish to use such a program package. The system is intended to provide a flexible package which will allow users to investigate special problems by appropriate choice of models, parameter values, and data sets. COSYMA incorporated ideas from the NRPB MARC programs, the KfK program UFOMOD, new developments, and contributions from other MARIA contractors.
External irradiation of people from radioactive material deposited onto the ground and surfaces following an accidental release to atmosphere is frequently an important exposure pathway. In the past, models for estimating the radiation doses from this pathway were based on rather simple assumptions because of the relatively limited understanding of the behaviour of radionuclides deposited on urban surfaces. However, in recent years a significant body of research work, much of it within the methods for assessing the radiological impact of accidents (MARIA) programme, has been carried out. This has enabled the National Radiological Protection Board (NRPB) to develop a more detailed and realistic model for calculating individual and population doses from gamma emitting material deposited in an urban environment. This model, given the acronym EXPURT (exposure from urban radionuclide transfer), simulates the movement of activity deposited on various surfaces in the urban environment. By taking into account the shielding properties of buildings and the habits of the population EXPURT then evaluates the external doses to people living in such urban environments as a function of time after deposition.
EXPURT is a linear compartmental model employing first order kinetics to calculate the transfer of activity from one urban surface to another. EXPURT has 17 compartments each representing various environmental pools in which activity may reside. 5 main surface types have been considered in the model: roofs, walls and indoor surface of buildings, impermeable ground surfaces (including roads, pavements, etc) and permeable surfaces (including grassland, parks, soil, gardens). The 3 compartments termed, 'ground water', 'loss' and 'sewers' have been introduced to hold activity lost from the soil column, from indoor surfaces of buildings and from the drains, respectively. This enables the potential for other exposure pathways to be identified and included at a later date if neces sary. An additional compartment labelled 'decontamination' has been included into which activity removed following decontamination of a surface may be transferred. This enables the amount of activity removed from the urban area by decontamination processes to be monitored and again allows for the possibility of studying other exposure pathways in the future, (eg doses to decontamination workers, or doses from the disposal of radioactive waste arising from decontamination).
Following development of the EXPURT model a preliminary sensitivity analysis was performed with the following objectives:
to identify those parameters in the model whose uncertainty contributes most to the uncertainty in the model predictions and which would therefore warrant more detailed study;
to identify those parameters whose uncertainty does not greatly affect the model predictions, for which more detailed study is not required;
to compare the NRPB's existing model for calculating external gamma doses from deposited activity with the range of values that might reasonably be predicted using the EXPURT model.
EXPURT has also been used to assess the effect of decontamination in reducing external doses to populations living in urban environments. Decontamination of the various surfaces modelled is simulated by the transfer of activity from one compartment to another, or by the removal of activity from a particular surface.
At the National Radiological Protection Board (NRPB), work centred around the continuing development of an improved model for calculating the economic consequences of accidents, and plans for the incorporation of this model into COSYMA (code system from MARIA (methods of assessing the radiological impact of accidents)). Work on the latter project was undertaken at both NRPB and Kernforschungszentrum Karlsruhe (KfK). The work described in this progress report is limited to the parts of the study that have been undertaken at NRPB.
Several models for predicting the economic impact of accidents have been developed, both in Europe and in the United States. These models have, however, either been incomplete in the economic aspects considered, or have been inappropriate for application in the United Kingdom because of national economic differences. Under this contract, a new model, COCO-1 (cost of consequences offsite) has been developed, for estimating the impact of an accident in monetary terms. COCO-1 is intended for application in studies of the offsite impact of accidents at nuclear installations. It excludes consideration of the cost of onsite consequences of accidents, which are beyond the scope of the applications planned for the model.
Most of the consequences of an accident can, at least theoretically, be associated with an economic cost. These may be broadly summarised as resulting from:
the application of countermeasures to reduce doses;
radiation induced health effects in the exposed population;
impact on the activity with which the installation is associated, for example the power programme;
long term social and political impact;
Ideally, a model for estimating the economic impact of an accident should consider the cost of all benefit foregone as a result of the accident, and should therefore include consideration of all of the above categories. However, because of the difficulties in calculating the economic impact of intangi ble effects, the approach generally adopted in the COCO-1 model is to calculate the tangible costs that have a direct and measurable effect on the economy. In 1 or 2 areas, the calculation is broadened to include less tangible costs, to represent lost benefits. The COCO-1 approach is thought to offer a broadly applicable and robust technique for estimating the economic impact of most accidents.
The principal application of the COCO-1 model is likely to be in the general areas of emergency planning, where information on the costs and effectiveness of remedial actions will provide input into decisions on the type and extent of countermeasures to be imposed after an accident, and into studies of siting policy and plant safety design features. To enable thee model to be used for such applications, it has been constructed in a form appropriate for use in standard accident consequence codes.
In addition to the development of a detailed economic consequence model, work was also carried out under this project to improve the countermeasures models in the NRPB's accident consequence code, MARC. The models for evacuation, sheltering and relocation were made more flexible and general and the code was modified so that the effect of introducing food restrictions at 1 dose criterion and withdrawing them at another could be assessed. An illustrative analysis was undertaken to demonstrate the use of the method to determine the optimum dose criterion for withdrawing food restrictions.
A version of MARC was developed in which the criteria for banning food is if the contamination exceeds a given level, rather than being related to the annual dose. The effect of introducing restrictions using the 2 methods show that, for the criteria considered, restrictions based on a contamination level are more extensive than those based on an annual individual dose of 5 mSv used in earlier studies.
All the work on uncertainty analysis at National Radiological Protection (NRPB) has been carried out using Monte-Carlo methods, in which values are assigned to each of the parameters considered to be uncertain and results of the accident consequence assessment (ACA) programs for those values of the input variables obtained. There are 2 ways of sampling values for the input variables from the distributions specified for each, and the relative merits of the 2 systems have been investigated. The variables can be selected at random when particular quantitative statements are to be made about the probability of the output quantity exceeding particular values (the statistical tolerance limit approach). This method has the disadvantage that random samples obtained do not necessarily cover the full range of the input distributions unless a large number of input variables are considered. The alternative method is to use the Latin hypercube sampling method, in which the values selected for the input variables must cover the whole of the parameter value distribution. This method, therefore, has the advantage of covering the whole of the parameter space but the disadvantage that quantitative statements about the probability of the output quantity exceeding particular values cannot be made. However the identification of important variables is easier with the Latin hypercube sampling method than with the statistical tolerance limit approach. Both methods have been used in studies of the uncertainty in predictions of modules of the NRPB's computer program for ACA's, MARC.
An analysis of the uncertainty in the food chain modules of MARC-1 using the statistical tolerance limit approach was carried out as part of MARIA-1 for a very large hypothetical accident (that designated UK1) at a pressurised water reactor (PWR) in the United Kingdom. A similar analysis was carried out as part of MARIA-2 for a source term equal to 1% of that for UK1. The same random sample for the input varia bles was used in the 2 sets of analyses. The analysis showed that the range of the output values for the small accident was slightly larger than that of the large accident, both in terms of the range at the higher percentiles of the distribution and in terms of the probability of exceeding relatively low values of the consequences.
An analysis was carried out as part of the MARIA-1 program into the uncertainty in air concentration and deposition of selected nuclides at specific points. This work was extended in the MARIA-2 program by continuing the analysis to obtain the uncertainty on the consequences of a large hypothetical release (that designated UK1). 59 runs of MARC-1 were carried out with input parameter values selected at random from their distribution. The range of uncertainty for the numbers of health effects and countermeasures was typically about a factor of 8. The uncertainty on the amounts of food banned resulting from the uncertainty on the atmospheric dispersion model parameters is comparable to that found in the earlier analyses of the uncertainty in the food chain parameter values. The parameters contributing most to the uncertainty were identified by considering partial rank correlation coefficients (PRCC) between the input and output values. Unfortunately the results were not clear, with some parameters having a very high PRCC for 1 end point but a very low value for other end points which would be expected to have similar results (eg number of early deaths and number of people evacuated).
Some of the parameters which are considered to be uncertain are read into MARC from data libraries. There are 2 ways in which the uncertainty on these parameters can be included, and both methods are used in this analysis. A data library of dose per unit intake was created for each MARC run from their standard value and an uncertain multiplying factor. However the food chain data library was calculated for each run of the MARC program from values for each of the basic transfer coefficients describing transfer of material along food chains. This was necessary to describe adequately the correlations between concentrations in the same food at different times or in different foods.
The National Radiological Protection Board's (NRPB) set of foodchain models, now called FARMLAND, has been further developed and various model validation studies carried out. In addition, work has been carried out to look at the patterns of production and consumption of terrestrial foods in the European community and the distribution of foods between different regions.
An improved root crop model has been developed as part of FARMLAND. The processes modelled include the interception and retention of material deposited on the plant's surface and its translocation from the plant's surface to the edible roots. Also modelled is the uptake of radionuclides from the soil to the edible parts of the plant. The model can predict the concentration of activity in the root crop as a function of time following deposition from atmosphere at any time of the year. Account is taken of the times of planting and harvesting of the crop and the subsequent pattern of consumption. The model is based on transfers and agricultural practices applicable to potatoes, as this is the main root crop consumed in the United Kingdom. It could easily be adapted to be more applicable for other types of root vegetables. For potatoes, a distinction is made between early and late varieties, where the early varieties are consumed in a short period after harvesting while a proportion of late varieties are stored and consumed over a longer period. The revised root crop model predicts higher concentrations of most radionuclides in root crops than the previous model used at NRPB. This is particularly the case for radioisotopes of elements such as caesium, which are readily translocated within plants. For radionuclides such as those of strontium or the actinides, where translocation is less important, the differences between the previous and revised root crop model are smaller. However, using the results of the revised root crop model does not significantly affect the predicted agricultural consequences of accidental releases. This is due to the importance of the transfer of radionuclides to other foods, in particular milk and green vegetables, together with grain for releases at certain times of the year.
Following the Chernobyl nuclear reactor accident in April 1986, a variety of environmental data became available. Some of these data were suitable for validating parts of the NRPB's model for the transfer of radionuclides through terrestrial foodchains. Initially, the FARMLAND predictions were compared with measurement data averaged over large areas of the United Kingdom and for 2 specific farms. The aim was to test the ability of FARMLAND to represent the general conditions for which it was developed, and also to test its ability to simulate site specific conditions. The comparisons were carried out for iodine-131, caesium-134 and caesium-137. For the general, large area case, the foods considered were milk, green vegetables and lamb. The 2 specific farms were dairy farms in Cumbria and Berkshire and detailed measurement data were available for the deposition to ground and activity concentrations in animal fodder and milk.
Comparisons of FARMLAND results with measurement data applicable to large regions of the United Kingdom have shown that the models predict the time dependence of radionuclide concentrations well. In using the post-Chernobyl monitoring data there are considerable uncertainties in the compatibility of the deposition measurements with those in food. It is therefore difficult to draw any real conclusions from a comparison with these data on the ability of the model to predict the scale of transfer to a particular food. From this point of view many of the environmental monitoring information collected after Chernobyl were disappointing for detailed model validation.
Calculation of the distribution of doses arising from the contamination of foodstuffs following an accidental release requires information on the distribution of food from the places of production to those of consumption. There has been a review of available data on food distribution in the United Kingdom, for the food categories of milk, milk products, beef, sheep meat, offals, grain, green vegetables and root crops. The information obtained from this review, together with data on the distribution of foodstuff production and population in the United Kingdom, has been used to estimate distribution patterns for these food categories, between 9 regions in the United Kingdom. The data are appropriate for use in accident consequence assessment (ACA) programs. As part of the Commission of European communities (CEC) post-Chernobyl research programme on underlying data for Derived Emergency Reference Levels NRPB and the Gesellschaft fuer Strahlen and Umweltforschung (GSF) have carried out a programme of work on terrestrial foodchain models. One of the aims of this work was to make recommendations on a general model suitable for use in the European community (EC). As a first step to recommending a model, predictions of the GSF foodchain model ECOSYS and FARMLAND were compared with sets of environmental measurements. Secondly, the predictions of the 2 models were compared for a range of situations. Based on this work NRPB and GSF were able to recommend a general model for use in the EC in the absence of site specific information. This general model will be used to generate data sets for foodchain concentration for a number of radionuclides for use with the COSYMA (code system from MARIA (methods of assessing the radiological impact of accidents)) package.
A computer system has been developed for calculating the consequences of hypothetical nuclear accident, as part of the MARIA project. A number of studies have been carried out to determne the most appropriate models to use in such a system.
An aim of these studies was to identify atmospheric dispersion models that could be used in place of the straight line Gaussian models normally used. Comparisons have been undertaken between a Musemet,a Gaussian model, RIMPUFF, a puff trajectory model and TRANSLOC, a three-dimensional Eularian grid model. The study concluded that Gaussian trajectory models can be used in place of the straight line models. The study also showed that the system should be divided into different programs for considering consequences ner to and far from the site, using different dispersion models.
A model for the dispersion of tritium, including its deposition and subsequent release to the atmosphere has been developed and included in the system.
Different ways of modelling the economic consequences of accidents have been considered, and comparisons made between the NRPB model COCO-1 and the MECA model developed by the Universidad Polttechnic of Madrid.
These studies led to accident consequence code system COSYMA, for use on a mainframe computer. Investigations have been carried out to determine how thw system can be simplified for use on a PC. A PC version of the system has subsequently been developed.
Topics to be studied under the 1990 to 1991 MARIA project
This summary describes the work being undertaken as a part of the MARIA project by the Nuclear Research Centre Karlsruhe (KfK), FRG, and the National Radiological Protection Board (NRPB), UK.
COSYMA development and maintenance
The COSYMA system is an important tool for research and application, and as such requires continuous updating, testing and improvement. The COSYMA code will be maintained by KfK and NRPB throughout the period covered by the contract, and there will be further development of its databases by both organizations.
The increasing availability of small computers, and the need by inexperienced ACA code users of the COSYMA code as a tool, necessitate the development of a reduced and simpler version of the code. This version will be more limited in its flexibility, but will be consistent in its basic assumptions and data with the full COSYMA code, and will produce compatible results. This work will be undertaken largely at NRPB, in cooperation with KfK.
Under the MARIA contract, NRPB and KfK will participate jointly, with the COSYMA code, in the international ACA code intercomparison exercise being organized by the CEC and NEA. It is anticipated that the participation of the code in this exercise will indicate any weak aspects in coding and in data which will be improved in future releases of COSYMA.
Uncertainty analysis techniques
Continuing research will be undertaken into the application of available uncertainty analysis techniques to ACA code systems. An important task is research into techniques for the reliable and defensible quantification of the variation in model parameters, including the use of expert judgement. Both NRPB and KfK will be involved in these studies, together with other CEC contractors.
Model and code development
A number of areas have been identified where further modelling development is needed; these will be considered during the period of this contract. These include:
at NRPB there will be investigations with COSYMA to examine the effects of using models of different complexity in the area of atmospheric dispersion;
at KfK, wet deposition processes will be included in the foodchain module of COSYMA;
at NRPB, there will be further development of the urban contamination model EXPURT; onsideration will be given to incorporating EXPURT results in a COSYMA database;
at KfK, the economics module will be extended to incorporate specific data on the economic productivity of the region near to the release point;
at KfK, there will be further development of the health effects models used in COSYMA; in particular the method of calculating loss of life expectancy will be improved, and the dependence on dose and dose rate of both the dose/risk relationships for nonstochastic health effects and of the cancer risk factors will be included.