Index > 2 Development of a strategy, implementation and execution program to remediate radioactively contaminated sites >

2.5.4 The risk assessment process: preliminary investigation

Contents
2.5.4.1 Establishing a preliminary view of the significance of the source term
2.5.4.2 System description
2.5.4.3 Selection of exposure scenarios
2.5.4.4 Developing/selecting and applying assessment models and data
2.5.4.5 Selecting input parameters
2.5.4.6 Selecting clean-up goals
2.5.4.7 Application of clean-up goals
2.5.4.8 Sensitivity analysis and risk management

A practical way forward in the assessment may be implemented based on a source, pathway, receptor approach [CIRIA-2005a]. This approach takes note that any harm arising from remaining contamination (the source), arises due to transfer (via various pathways) to those media including humans in which the harm may be expressed (receptors). It is consistent with the process of identification of hazards, the subsequent assessment of these hazards to estimate the risks and finally the evaluation of those risks. It also reflects a tiered approach to evaluation of the problem, so that the level of resources applied can be proportionate to the scale of the problem. If it reveals an unacceptable risk, the risk assessment process will feed into the options appraisal stage which then results in the implementation of a remediation strategy.

2.5.4.1 Establishing a preliminary view of the significance of the source term

This phase should begin with identification of the relevant contamination source term in terms of the activity levels of the main radionuclides, their physico-chemical forms, the size and activity of any particles present, and their spatial distribution over and under the land concerned, and an early view of the immediate near surface litho-stratigraphy of the site. This information will be available from site characterisation. The levels are then compared with those published for other purposes, in order to gain a preliminary view of the order of magnitude of potential committed effective doses to individual people. If doses seem likely to be of the order of microsievert then a simple assessment may be sufficient. If doses are of the order of hundreds of microsieverts or more, then much more detail will be needed.

It is also necessary in this phase to consider scientific uncertainties and stakeholder views. It would be unwise to conclude that levels of a particular radionuclide are of little significance, and to pay little attention to them in assessments, if recent evidence has called into question the scientific basis for the judgement of significance. Similarly, it is sensible to take into account stakeholder concerns about particular radionuclides, and particular physico-chemical forms of radionuclides, when judging source term significance and establishing an assessment methodology.

In addition, it should be noted that remediation work may result in a requirement to transport waste and to dispose of it elsewhere. This results in a need to consider the significance of a potentially wide range of issues marginal to the site being considered.

2.5.4.2 System description

In this phase, the features of the site and its environment should be defined. Relevant features should include soil type and land cover, surface and subsurface groundwater bodies, and the current land use. The amount of detail required will be influenced by the significance of the source term.

The system description needs to be understood sufficiently broadly to address all the environmental and human health risk endpoints of potential interest. Apart from radiation risks to exposed people, adequate emphasis should be given to protection of media.

2.5.4.3 Selection of exposure scenarios

In this phase scenarios should be developed, i.e. simple descriptions, for the evolution of the source term within the described system according to the assumed future land use associated with each option under evaluation. These descriptions should include:

  • Controls over land use;
  • Assumptions for land use;
  • Processes likely to results in migration and accumulation of radionuclides;
  • Processes likely to give rise to radiation exposure of people and non-human biota as a result of the presence, or migration and accumulation of radionuclides.

The concept of pollutant linkage from a source, via a pathway, to a receptor, applicable in the context of non radioactive contaminated land, has an equivalent approach in dealing with radioactively contaminated land. There can be a radiological impact from the contamination only if there is a source, pathway and receptor. The receptor will usually be a representative member of an exposed population receiving the highest dose, often termed the critical group. In reality pathways can be very complicated and may need to consider the impact of radioactive decay and in-growth of daughter radionuclides.

Modes of radiation exposure considered could include:

  • Ingestion of radioactively contaminated materials, including dust, aerosols, soil, foodstuffs and drinking water;
  • Inhalation of radioactively contaminated materials, including dust, aerosols and soil;
  • External irradiation from contaminated soils and other materials; and
  • Contact with contaminated materials.

As local people may be aware of local conditions, such possibilities should take into account local advice.

The key issue is to identify the more significant mechanisms by which people and other biota could come into contact with the more significant levels of radionuclides. Scenarios should include likely as well as unlikely events and processes.

In any remediation project, selecting appropriate current and future land use scenarios is a critical step in calculating clean-up levels [ITRC-2002]. Scenarios are descriptions of various lifestyles and activity patterns that approximate an individual’s exposure to contaminants in environmental media. Conceptual site models display the exposure pathways inherent in a scenario and are useful tools to convey which pathways are reasonable and complete, i.e., capable of transferring harmful effects from radionuclides in surface soil to exposed individuals. By developing conceptual site models, it is possible to estimate representative modes of exposure for target populations, allowing those exposures to be quantified.

Depending on the regulatory framework, a reasonable maximum exposure of the average member of the critical group should be defined based on current land use as a starting point for establishing exposure scenarios. Alternative future land uses may be considered if they seem possible or likely based on available information and professional judgment. It should not be necessary to assume catastrophic events, but rather reasonable land uses and human activities and that the current physical characteristics (i.e., important surface features, soils, geology, hydrogeology, meteorology, and ecology) will exist at the site for the next 1,000 years.

Generally, clean-up based on a residential scenario (suburban resident, rural resident, resident farmer, or rancher) will allow unrestricted use of a site. Choosing a less conservative scenario may invoke institutional controls and inherent long-term stewardship issues. The considerable difference in half-lives among various radionuclides is an important consideration in deciding whether long-term controls are feasible and therefore may affect exposure scenario selection.

2.5.4.4 Developing/selecting and applying assessment models and data

Mathematical models are used to approximate human and ecological exposure at a site [ITRC-2002]. The basic equations used to assess health effects due to radiological exposure are relatively straightforward and can be computed with a hand calculator or a spreadsheet. These equations generally sum the exposure from the ingestion, inhalation, and external irradiation pathways, each of which has an intake or source term, an exposure period, and either a dose conversion factor or a cancer slope factor. Modifying factors can be added, which adjust exposure periods and account for fate and transport of radionuclides in the environment. These factors may add considerably to the number of interacting terms and therefore to the complexity of the calculations.

The models are normally developed in stages, including a conceptual description, a mathematical representation of that description and the selection of data for the mathematical models 17. In general, new models will not be required; rather, based on the output of previous phases and the choice of endpoints, models can be chosen from the literature. Furthermore, many models can be implemented on spreadsheets and do not require sophisticated techniques or software.

The assessment process typically involves some iteration. For example, suitable data may not be available for the initial choice of model, or some variant exposure pathway which is locally relevant may have been identified, and so a variation in the model may be appropriate. Any such developments should be transparently documented and justified. Preliminary results may be used to identify the more significant impacts and hence guide assessment iterations.

In the case of more significant contamination it may be appropriate to apply more sophisticated models, e.g. for the long-term migration of contamination through the ground. Several multimedia/multiple-pathway computer models have been developed to handle these more complex calculations [ITRC-2002]:

  • RESRAD family of codes (DOE-Argonne National Laboratory);
  • MEPAS (Multimedia Environmental Pollutant Assessment Systems)/-GENII/FRAMES/SUM3 set of codes (Pacific Northwest Laboratories);
  • MMSOILS (EPA);
  • DandD (NRC);
  • Presto-EPA-CPG (EPA);
  • PATHRAE-EPA (EPA).

Computer codes can be evaluated or compared through processes known as “benchmarking,” “verification,” and “validation.” Benchmarking compares the results from several different computer codes using the same set of problems. Verification is the procedure that tests for internal mathematical consistency and accuracy. Validation is the process that tests a mathematical model against actual field measurements.

Several criteria can be considered during the computer code selection process:

  • Does the code incorporate key processes from the conceptual site model?
  • Does the code satisfy study objectives?
  • Has the code been verified using published analytical equations in scientific and technical journals?
  • Has the code been validated against known site conditions?
  • Does the code have the capability of inputting probabilistic analyses?
  • Is the code well documented?
  • Is the model available in the public domain?

While models are extensively used in risk assessment, the selection and interpretation of results need close examination. Relying excessively on models in the context of waste disposal and site contamination issues should be considered with care, taking into account that:

  • Existing major differences between models may be due to differing objectives – where the capabilities of the models overlap, such differences may be due to the formulation of transport components.
  • Spreadsheets (or pen-and-pencil calculations) are much more flexible than computer models. The effect of using a computer programme rather than a spreadsheet to implement the risk assessment may be that the assumptions that most need review are hidden where they are not accessible.
  • Deterministic models are unable to account for uncertainties in input data and therefore yield outputs (such as contaminant concentrations, exposure doses and risks) of unknown reliability.
  • The principle of parsimony should be used to differentiate between alternative operational models. This principle states that among all operational models that can be used to explain a given set of experimental data, this model should be selected that is conceptually least complex and involves the smallest number of unknown (fitting) parameters.
  • Models are appropriate, often essential, tools for risk assessment and decision-making concerning clean-up and management of contaminated or potentially contaminated sites. However, it is inappropriate to use models as “black boxes” without tailoring them to site conditions and basing them firmly on-site data. Neither disregard of models nor overreliance on them is desirable.
  • The environment constitutes a complex system that can be described neither with perfect accuracy nor with complete certainty. It is imperative that uncertainties in system conceptualisation and model parameters and inputs be properly assessed and translated into corresponding uncertainties in risk and decisions concerning risk management. The quantification of uncertainties requires a statistically meaningful amount of quality site data. Where sufficient site data are not obtainable, uncertainty must be assessed through a rigorous critical review and sensitivity analyses.
  • Models and their applications must be transparent to avoid hidden assumptions. Model results must not be accepted at face value, because hidden assumptions are easily manipulated to achieve desired outcomes.
  • Decisions concerning site disposition and risk management should account explicitly and realistically for lack of information and uncertainty.
  • The monitoring of site conditions and contamination is an imperfect art. It is important that uncertainty associated with monitoring results be assessed a priori and factored explicitly into site remedial design and post-closure management.

2.5.4.5 Selecting input parameters

Many of the key parameters used in calculating clean-up levels are bounded within certain ranges once an exposure scenario is established. For example, typical exposure periods and breathing and ingestion rates for various scenarios have been determined for use in risk or dose calculations. In some cases, especially for sensitive parameters, distributions may be available and used in place of discrete values. Using distributions enables the entire range of possible values to be considered for a parameter and helps to account for the uncertainty and variability inherent in parameter selection. Relatively few input parameters used in computer codes or risk equations have significant influence on the resultant clean-up level. These include inhalation rate, dose conversion factors, soil ingestion rate, mass loading for inhalation, and others.

When assessing human exposure, input parameters should be selected so that the combination of all intake variables results in an estimate of the “reasonable maximum exposure” expected to occur at a site for a given scenario. Exposure is mainly addressed in terms of the “average member of the critical group,” which means “the group of individuals reasonably expected to receive the greatest exposure to residual radioactivity for any applicable set of circumstances.”

Behavioural parameters are generally determined, or at least bounded, by the selected exposure scenario. Physical parameters are determined by measurements at or near a particular site, if available. Site-specific values should always be used whenever possible. Differences in physical settings from site to site, or between site-specific and default values, account for some of the variations in calculated risk levels.

2.5.4.6 Selecting clean-up goals

In a risk assessment process, dose-based and/or risk-based values are calculated. In a subsequent risk management process, clean-up goals are established using calculated soil concentrations as a basis.

Various terms are used, sometimes interchangeably, to describe numbers that guide remedial actions at radioactively contaminated sites, such as “action levels”, “ALARA goal levels”, “allowable residual soil concentrations”, “clean-up levels”, “clean-up standards”, “derived concentration guideline levels”, “guideline concentrations”, “remedial goal options”, “remedial goals”, “remediation levels”, “risk-based concentrations”, “soil clean-up concentrations”, and “soil clean-up criteria”. Clean-up levels from site to site, or even at a single site, cannot be compared without knowing their purpose, how they were derived, and how they will be applied.

An “action level” may refer to the existence of a contaminant concentration in the environment high enough to warrant action or trigger a response such as removal, treatment, containment, stabilization, or institutionally controlling exposure.

“Derived concentration guideline levels” may be examples of specific investigation levels derived by converting dose or risk from a release criterion into concentration or activity levels that are directly measurable.

“Preliminary remediation goals” may be the initial remedial guidelines usually developed early in the remediation phase to provide risk-reduction targets. Numerical “preliminary remediation goals” for radionuclides are typically based on the upper-bound carcinogenic risk of one in a million (10-6). Until the final remedy is selected and documented, “preliminary remediation goals” constitute initial guidelines, not final cleanup goals.

“Remediation goals” may be media-specific clean-up goals for a selected remedial action. Numerical “remediation goals”, which are part of the remedial action objectives, can be based on existing standards or on risk calculations. These two criteria are the “threshold criteria” for evaluating both remedial alternatives and remedial action objectives.

As risk-based “preliminary remediation goals” do not necessarily represent realistic exposure and risk, those numbers may not be appropriate clean-up levels. “Preliminary remediation goals” can be proportionally adjusted upward to become “remediation goals” using a level higher in the acceptable carcinogenic risk range to account for the conservatism inherent in the “preliminary remediation goals”. Other factors related to technical limitations (e.g., detection or quantification limits) can also be applied. In addition, the “balancing criteria” and the “modifying criteria” for analysing remedial alternatives, such as cost and state and community acceptance, should also be considered. In some cases, “remediation goals” may be adjusted downward to account for multiple radionuclides or co-occurring non-radionuclide chemicals. Final “remediation goals” should be documented as radionuclide-specific “remediation levels” or as qualitative definition of the risk-reduction clean-up objective to be achieved for the non-numerical “remediation goals”.

A specific approach for the implementation of remediation criteria has been discussed in Section 2.2.2.3, Definition of a remediation process, initial decision making, based on the form of the reference levels indicated in Table 2.1.

2.5.4.7 Application of clean-up goals

Once a clean-up level has been established, differences may still remain in how the value is applied. The application of a clean-up level, whether risk- or dose-based, should be tied in some way to characterisation data points. The location and density of these data points may be determined by a variety of characterisation sampling schemes:

  • Biased sampling – locations where process knowledge, limited analytical data, visible staining, topography, vegetation, etc. suggest the possibility of contamination.
  • Standard statistical sampling – a regular, systematic plot of locations on sites of little or no data; triangular grids and protocols for determining appropriate grid spacing may have to be recommended.
  • Geostatistical sampling – an iterative process based on the remediation of a contaminated site to a required clean-up level at a specified level of confidence; sampling results are used to determine the optimal number and locations of samples to be collected in the next iteration, if necessary.

If multiple radionuclides are present in the environment, the sum-of-ratios (or sum-of-fractions) method should be used to account for the contribution of each single isotope towards the dose- or risk-based limit. Measured values of all radionuclides present should be compared to clean-up levels by dividing the measured value of each radionuclide by its respective clean-up level, then adding the ratios. If the sum of the individual ratios is greater than 1, then the limit is considered to be exceeded:

Formula 2.1
Formula 2.1

where:
Cj = soil concentration of radionuclide j,
CGj = clean-up goal for radionuclide j.

Exceedances of clean-up levels may be determined by comparing those levels to aggregations of sampling data over specified areas of concern or exposure units. Clean-up criteria at most sites may also include hot-spot methodologies, which will require evaluation of small areas of elevated sample results within larger areas, which have been determined to require no further remedial action. These hot spot methodologies usually incorporate an area-weighted factor, which – when applied to clean-up levels – provides an upper limit on the amount of activity that can be left in these small isolated spots.
Setting more restrictive clean-up levels will necessarily lead to more clean-up at a higher cost, but for specific projects at some sites, those increased costs may be incrementally small or may reduce long-term stewardship costs.

2.5.4.8 Sensitivity analysis and risk management

In most cases a sensitivity analysis should be carried out to address variations in assumptions and parameter values, and perhaps models [CIRIA-2005a]. The analysis could be quantitative or semi-quantitative, and need not involve complex calculations. The aim should be to produce a range of results so that it can be seen whether the comparison of options has a different outcome if very different assumptions and parameter values are used in estimating risks.

The apliation of cleaunp levels is confused . it is necessary to establish if the cleanup target is on single values (as DCGL emc are) or on averages (mean or median) (as the DCGL´s are). Other issue is how the uncertainties are taking in account when Sum-of-ratios (or Residual Activity Index) isfaormual is used. Plese clarify
– by Rafael Garcia-Bermejo Fernandez about 6 years ago