Index > 3 Characterisation of radioactively contaminated sites >

3.3.2 General site characterization survey design aspects

Contents
3.3.2.1 Introduction
3.3.2.2 Site classification by contamination potential.
3.3.2.3 Identification of survey units
3.3.2.4 Factors influencing the site characterization design quality
3.3.2.5 Survey design errors
3.3.2.6 Design considerations for sites with a relatively uniform distribution of contamination
3.3.2.7 Design considerations for small areas of elevated activity
3.3.2.8 Determining survey and investigation levels
3.3.2.9 Development of an integrated site characterization strategy
3.3.2.9.1 Land area surveys
3.3.2.9.2 Structure surveys
3.3.2.10 Other survey designs

3.3.2.1 Introduction

There is great variability in the details of characterization approaches appropriate to specific problems and sites. Differences among sites due to the heterogeneous character of the natural environment and to the nature and history of contamination are enough to require different approaches. However, the varieties of other important influences on the remediation problem definitely require that a characterization approach be designed to address all such issues. The guidance here focuses on the important elements that any strategy developed for characterization of a specific site should consider. It also addresses the value of flexibility and phasing of study components to allow revision of the strategy as new information becomes available. Characterization data are an important element in making effective remediation decisions. Clear specifications of the objectives and strategies for the characterization are important.
In the developed sampling and analysis plan, due consideration should be given to:

  • Spreading contamination. Characterization practices should be designed in such a way as not to contribute to the further spread of contamination at the site, or off-site. This is of particular concern when dealing with radioactive contamination. For example, contamination can be spread through uncontaminated aquifers as a result of poor drilling and well completion practices. Care should be exercised so that on-site workers do not inadvertently carry radioactive contamination off the site through inadequate decontamination processes.
  • Accessibility. During the planning process it is necessary to consider access logistics, including the ability to physically gain entry to the site, especially for any equipment that is brought in (e.g., drilling rigs, cone penetrometer trucks). It should also be considered whether there are any overhead or underground utilities which may impact the investigation. It may be necessary to limit access to a contaminated area to only specially trained site workers and to allow for a decontamination zone for equipment and personnel.
  • Jurisdictional concerns. Before initiating field work, it is essential to obtain any approvals necessary to access the area to be characterized. Authorization may be required from governmental or private parties. In addition, it may be necessary to obtain certain permits for digging, drilling, or installing any groundwater wells. A check list of requirements should be prepared to ensure preparedness.

Further, a field-based site characterization has to fulfill the following (suggested) decision sequence to determine the appropriate investigation strategy for a site:

  • Decision 1: Is there enough information to meet project objectives (for risk assessment, established data quality guideline level(s), options comparison, preferred option implementation or verification and validation)?
    If not, the objectives of the next phase of the investigation should be defined. If yes, the next phase of the characterization, remediation and restoration process should be defined.
  • Decision 2: For what should the (new) samples (soil, water, gas, etc.) to be collected be specified to obtain enough information to meet project objectives and what is the specification for the analyses?
    These samples have to provide additional information for one or more specific pathways of the conceptual model. Therefore specific requirements have to be set for the analysis results of those samples (e.g., sample size, accuracy, etc.) to obtain the correct and needed information.
  • Decision 3: What locations should be sampled, how many and with what frequency samples should be taken?
    Selection of the sample locations and the frequency of sampling have to fulfill the requirements to meet objectives, e.g., conceptual model.
  • Decision 4: From what depths should the samples (soil, water, gas, etc.) be collected, and what are the instrumental requirements?
    Selection is made at which depths (e.g., surface, 10 cm, etc.) the samples should be collected and the specific instrumental requirements defined to meet the required analysis results.
  • Decision 5: What form (non-intrusive and/or intrusive) investigation is necessary to obtain the specified samples to meet project objectives?
    Selection is made which non-intrusive and/or intrusive method(s) will be applied to collect the samples.
  • Decision 6: What techniques and (monitoring) installations/instrumentation should be employed to obtain the required analyze result from the samples (soil, water, gas, etc.)?
    Selection of the appropriate analyze technique (non-destructive and/or destructive) to obtain an analyzed result that fulfils the requirements.
  • Decision 7: What quality measures will be employed to ensure accurate data from the point of sample collection or monitoring to the laboratory and the data interpretation?
    What are quality control and quality assurance measures?

Each decision should be documented so that other stakeholders can understand why the design was selected. Examples of linkage between site investigation design aspects and conceptual models are presented in Table 3.2.

In the case that more than one initial site conceptual model for a site or part of a site has been developed, site characterization data should be obtained to test the various models and discriminate between them. Some of these models may be rejected because they are inconsistent with the new data, and uncertainty in the remaining model(s) will be reduced.

In many site characterizations, it is appropriate to phase the investigations. More detailed characterizations are deferred until the results of earlier phases of work have been evaluated. This approach ensures that the later investigations are focused on relevant areas with the appropriate degrees of accuracy and confidence employed.

Survey design to address potential contamination through identified pathway

Pathway identified in the conceptual model
* Air quality sampling.
* Surface sampling for radioactive and non-radioactive contaminants on an appropriate sampling pattern.
* For other contaminants, addressed by intrusive investigations on an appropriate sampling pattern.

Diffuse airborne contamination.
* Walk over radiation surveys.
* Soil vapor survey.
* Surface and shallow sampling adjacent to roads.

Spillage from vehicles during transport operations.
* Walk over radiation surveys.
* Soil vapor survey.
* Surface and shallow sampling adjacent to roads.
* Trial pits/boreholes located position of known buildings.

Disposals/spillages/losses associated with former buildings.
* A drain survey, including sampling of drain sediments.
* Trial pits/boreholes located along the line of the drain.

Leakage from drains.
* Walkover geophysics survey prior to intrusive sampling, in order to detect disturbed ground, buried objects and services.
* Soil vapor survey.
* Walk over radiation surveys.
* Intrusive investigations at positions identified by geophysical survey.

Burial of waste materials.
* Biota, e.g., meat, fish, diary products, vegetables, fruit, horticulture products, mushrooms, etc.

Food chain.

Table 3.2 Examples of linkages between site characterizations design aspects and conceptual model.

3.3.2.2 Site classification by contamination potential.

Classifying a site/survey unit is crucial to the survey design because this step determines the level of characterization/survey effort based on the potential for contamination. Sites are initially classified as impacted or non-impacted based on existing information and can be re-classified based on new information, e.g., preliminary investigation, historical site assessment.

Non-impacted areas have no reasonable potential for residual contamination and require no further evidence to demonstrate compliance with the release criterion. When planning the final status survey, impacted sites may be further divided into survey units. If a survey unit is classified incorrectly, the potential for making decision errors increases. For this reason, all impacted areas are initially assumed to be Class 1 (see Section 3.2 and Appendix D). Class 1 areas require the highest level of survey effort because they are known to have contaminant concentrations above the release criteria, or the contaminant concentrations are unknown.

Information indicating the potential or known contaminant concentration is less than the release criteria can be used to support re-classification of an area or survey unit as Class 2 or Class 3 (see Section 3.2).

There is a certain amount of information necessary to demonstrate compliance with the release criterion. The amount of this information that is available and the level of confidence in this information are reflected in the area classification. The initial assumption for affected areas is that none of the necessary information is available. This results in a default Class 1 classification. This corresponds with the statement of the null hypothesis that the survey unit is contaminated, and represents the most efficient case for the regulator. For this reason, the recommendations for a Class 1 final status survey represent the minimal amount of information necessary to demonstrate compliance.

Not all of the information available for an area will be collected for purposes of compliance demonstration. For example, data will be collected during characterization surveys to determine the extent, and not necessarily the amount, of contamination. This does not mean that the data do not meet the objectives of compliance demonstration, but may mean that appliance of statistical tests would be of little or no value because the data have not been collected using appropriate protocols or design. Rather than discard potentially valuable information, EURSSEM allows for a qualitative assessment of existing data (see Section 2.4, Historical site assessment).

Non-impacted areas represent areas where all of the information necessary to demonstrate compliance is available from existing sources. For these areas, no statistical tests are considered necessary. A classification as Class 2 or Class 3 indicates that some information on describing the potential for contamination is available for that survey unit. The data collection recommendations are modified to account for the information already available, and the statistical tests are performed on the data collected during the final status survey.
As previously stated, the conservative assumption that an area receives a classification of Class 1 is only applied to impacted sites. The historical site assessment (see Section 2.4) is used to provide an initial classification for the site of impacted or non-impacted based on existing data and professional judgment.

3.3.2.3 Identification of survey units

A survey unit is a physical area consisting of structures or land areas of specified size and shape for which a separate decision will be made as to whether or not that area exceeds the release criterion. This decision is made as a result of the final status survey. As a result, the survey unit is the primary entity for demonstrating compliance with the release criterion.

To facilitate survey design and ensure that the number of survey data points for a specific site is relatively uniformly distributed among areas of similar contamination potential, the site is divided into survey units that share a common history or other characteristics, or are naturally distinguishable from other portions of the site.

A site may be divided into survey units at any time before the final status survey. For example, a historical site assessment or scoping survey results may provide sufficient justification for partitioning the site into Class 1, 2, or 3 areas (see Figure 3.2 for an example). Note, however, that dividing the site into survey units is critical only for the final status survey – scoping, characterization, and remedial action support surveys may be performed without dividing the site into survey units.

A survey unit should, in principle, not include areas that have different classifications. The survey unit’s characteristics should also be generally consistent with exposure pathway modelling that is used to convert dose or risk into radionuclide concentrations. For indoor areas classified as Class 1, each room may be designated as a survey unit. Indoor areas may also be subdivided into several survey units of different classification, such as separating floors and lower walls from upper walls and ceilings (and other upper horizontal surfaces) or subdividing a large warehouse based on floor area.

Classification Suggested area
Land areas Structures

Class 1 Up to 2,000 m2 Up to 100 m2 floor area

Class 2 2,000 to 10,000 m2 100 to 1,000 m2

Class 3 No limit No limit

Table 3.3 Suggested areas for survey units

Survey units should be limited in size based on classification, exposure pathway modelling assumptions, and site-specific conditions. However, due to new instrumental developments for scanning surveys and insights the areas are increasing in practice. The suggested areas for survey units are indicated in Table 3.3.

Figure 3.2 Example showing how a site might be classified prior to clean-up based on preliminary investigations, historical site assessment and supplementary investigations
Figure 3.2 Example showing how a site might be classified prior to clean-up based on preliminary investigations, historical site assessment and supplementary investigations

The limitation on survey unit size for Class 1 and Class 2 areas ensures that each area is assigned an adequate number of data points. The rationale for selecting a larger survey unit area should be developed using the Data Quality Objective Process (Section 3.3) and fully documented. Because the number of data points (determined in Section 3.5) is independent of the survey unit size, disregarding locating small areas of elevated activity, the survey coverage in an area is determined by dividing the fixed (minimum) number of data points obtained from the statistical tests by the survey unit area. That is, if the statistical test estimates that 20 data points are necessary (minimum) to demonstrate compliance, then the survey coverage is determined by dividing 20 by the area over which the data points are distributed.

Special considerations may be necessary for survey units with structure surface areas less than 10 m2 or land areas less than 100 m2. In this case, the number of data points obtained from the statistical tests is unnecessarily large and not appropriate for smaller survey unit areas. Instead, some specified level of survey effort should be determined based on the DQO process and with the concurrence of the responsible regulatory agency. The data generated from these smaller survey units should be obtained based on judgment, rather than on systematic or random design, and compared individually to the DCGLs.

An important consideration is that the above applied statistical method does not take into account the perception of different stakeholders. Therefore, it can be necessary to increase the number of data points to satisfy these stakeholders. It is advised to discuss and agree with the stakeholders first on the methodology to apply to calculate the number of data points and second on the calculation of this number.

The criteria used for designating areas as Class 1, 2, or 3 should be described in the final status survey plan. Compliance with the classification criteria should be demonstrated in the final status survey report. A thorough analysis of the historical site assessment (HAS) findings (Section 2.4) and the results of scoping and characterization surveys provide the basis for an area’s classification. As a survey progresses, re-evaluation of this classification may be necessary based on newly acquired survey data.

Example 3.1: Contamination identified in a Class 3 area

If contamination is identified in a Class 3 area, an investigation and re-evaluation of that area should be performed to determine if the Class 3 area classification is appropriate. Typically, the investigation will result in part or all of the area being reclassified as Class 1 or Class 2. If survey results identify residual contamination in a Class 2 area exceeding the DCGL or suggest that there may be a reasonable potential that contamination is present in excess of the DCGL, an investigation should be initiated to determine if all or part of the area should be reclassified to Class 1. More information on investigations and reclassifications is provided in Section 3.3.2.9.

3.3.2.4 Factors influencing the site characterization design quality

The quality of the site characterization arises primarily from:

  • The survey/investigation design errors.
  • Measurement errors.

3.3.2.5 Survey design errors

Survey design errors occur when the survey design is unable to capture the complete extent of variability that exists for the radionuclide distribution in a survey unit. Since it is impossible in every situation to measure the residual radioactivity at every point in space and time, the survey results will be incomplete to some degree. It is also impossible to know with complete certainty the residual radioactivity at locations that were not measured, so the incomplete survey results give rise to uncertainty. The greater the natural or inherent variation in residual radioactivity, the greater the uncertainty associated with a decision based on the survey results. The unanswered question is: “How well do the survey results represent the true level of residual radioactivity in the survey unit?”

Examples of possible areas of uncertainty are given in Table 3.4. Outstanding uncertainty should be recorded related to precision, bias, representativeness, completeness, comparability and sensitivity in order that the significance can be treated in the subsequent assessment of the data.

Site characterization activity

Examples of uncertainty Possible action to reduce uncertainty
Preliminary investigation; desk study Access or supply of historical information on site history limited by site owner/occupier, leading to failure to identify potential radioactive and chemical contaminants, jeopardishing health, safety and environmental management and scope of investigation (conceptual model uncertainty). Assume worst-case history, particularly for defense sites, and take client thorough an interactive process to try to establish all relevant sources of information.
Prepare contingency plans for health, safety and environmental management and site investigation procedures.

Inadequate information retained by client in plans and demolition records. Potential presence of in-situ buried structures (e.g., foundation, services) on the site (conceptual model uncertainty). Incorporate an exploratory investigation stage, using non-invasion geophysical surveying.
Limited intrusive investigations to prepare main investigation plans.

Poor conceptual model developed and/or lack of link with subsequent survey design. Results in poor quality investigation and poor quality health, safety and environmental management (conceptual model uncertainty).

Consult conceptual model checklist to ensure adequacy of model.
Review conceptual model and site investigation objectives at regular intervals throughout project.
Failure to set objectives, e.g., required risk target. Ensure that risk targets are set. Use conceptual model of site and required level of confidence in output to design an appropriate sampling strategy.

Failure to appreciate chemical and radioactive characteristics of waste that will be produced, possibly leading to production of waste (e.g., mixed radioactive and organics-contaminated waste) for which no regular disposal route exists.

Evaluate potential characteristics of waste, and ensure that disposal routes available.
Preliminary investigation; site reconnaissance. Failure to appreciate requirement of site operating procedures. Could limit technical scope of investigation (e.g., cannot investigate close to services) or could cause extensive delays to project schedule. Ensure that, during site visit, appropriate personnel are interviewed who can brief and supply contractors with necessary site operating instructions and documentation.

Site investigation:
- exploratory
- main
- supplementary

Uncertainty in conceptual model and therefore poor understanding of contamination occurrence. Use a phased investigation approach, real-time sampling to focus investigation, collect large number of samples, and or use Triad or Optimised Contaminated Land Investigation approaches.

Failure to locate services, both inside and outside site boundary. This could lead to damage to services, possibly resulting in injury/death to site personnel and/or disruption to site operations. Extensive delays and project schedule uncertainty. Ensure that excavation procedures on the client’s site are in accordance with site procedures and health, safety and environmental guidance. For off-site excavations, ensure that national utilities are contacted.

Inconsistent positioning information, leads to uncertainty in locations of contaminated ground, sampling points, services, etc. (data uncertainty). All investigations or surveys should be topographically surveyed to ordnance datum and national grid reference. The accuracy of the survey surveying method should be reported.

Poor quality management of investigation resulting in unreliable data (e.g., poor sampling and logging data). Further verification works may then be necessary to satisfy stakeholders (data uncertainty).

Ensure that all work is undertaken in accordance with quality management system.
Uncertainty in analytical data (data uncertainty). Check QA/QC procedures, analyse more samples, duplicate analyses, use different preparation methods, use different analytical methods with lower limits of detection, look for related contaminants.

Table 3.4 Examples of uncertainties arising during site investigation, and possible actions that can be taken to reduce uncertainties in the site charaterization and remediation [CIRIA-2009].

Measurement errors create uncertainty by masking the true level of residual radioactivity and may be classified as random or systematic errors. Random errors affect the precision of the measurement system, and show up as variations among repeated measurements. Systematic errors show up as measurements that are biased to give results that are consistently higher or lower than the true value.

A quality control (QC) program can both lower the chances of making an incorrect decision and help the data user understand the level of uncertainty that surrounds the decision. Quality control data are collected and analyzed during implementation of the site characterisation to provide an estimate of the uncertainty associated with the survey results. Quality control measurements (scans, direct measurements, and samples, etc.) are technical activities performed to measure the attributes and performance of the survey. During any survey, a certain number of measurements should be taken for quality control purposes.

3.3.2.6 Design considerations for sites with a relatively uniform distribution of contamination

The survey design for areas with relatively uniform distributions of contamination is primarily controlled by classification and the requirements of the statistical test. The guidance and recommendations provided for Class 1 survey units are designed to minimize the decision error. Guidance and recommendations for Class 2 or Class 3 surveys may be appropriate based on the existing information and the level of confidence associated with this information.

The first consideration is the identification of survey units. The identification of survey units may be accomplished early (e.g., scoping) or late (e.g., final status) in the survey process, but must be accomplished prior to performing a final status survey. Early identification of survey units can help in planning and performing surveys throughout the Radiation Site Survey Investigation Process. Late identification of survey units can prevent misconceptions and problems associated with reclassification of areas based on results of subsequent surveys. The area of an individual survey unit is determined based on the area classification and modelling assumptions used to develop the release criteria or derived concentration guideline level (DCGLW). Identification of survey units is discussed below.

Another consideration is the estimated number of measurements to demonstrate compliance using the statistical tests. Section 3.5 describes the calculations used to estimate the number of measurements. These calculations use information that is usually available from planning or from preliminary surveys (i.e., scoping, characterization, remedial action support).

The information needed to perform these calculations is:

  1. Acceptable values for the probabilities of making Type I (α) or Type II (β) decision errors,
  2. The estimates of the measurement variability in the survey unit (σs) and the reference area (σr) if necessary, and
  3. The shift (Δ).

EURSSEM recommends that site-specific values be determined for each of these parameters. To assist the user in selecting site-specific values for decision error rates and Δ, EURSSEM recommends that an initial value be selected and adjusted to develop a survey design that is appropriate for a specific site. An arbitrary initial value of one half the DCGLW is selected for the lower bound of the gray region. This value is adjusted to provide a relative shift (Δ/σ) value between one and three as described in Section 3.5. For decision error rates, a value that minimizes the risk of making a decision error is recommended for the initial calculations. The number of measurements can be recalculated using different decision error rates until an optimum survey design is obtained. A prospective power curve (see Appendix B, Section B.2 and Appendix E, Sections E1.3 and E2.4) that considers the effects of these parameters can be very helpful in designing a survey and considering alternative values for these parameters, and is highly recommended.

To ensure that the desired power is achieved with the statistical test and to account for uncertainties in the estimated values of the measurement variability’s, EURSSEM recommends that the estimated number of measurements calculated using the formulas in Section 3.5 be increased by 20%. Insufficient numbers of measurements may result in failure to achieve the DQO for power and result in increased Type II decision errors, where survey units below the release criterion fail to demonstrate compliance.

Once survey units are identified and the number of measurements is determined, measurement locations should be selected. The statistical tests assume that the measurements are taken from random locations within the survey unit. A random survey design is used for Class 3 survey units, and a random starting point for the systematic grid is used for Class 2 and Class 1 survey units.

3.3.2.7 Design considerations for small areas of elevated activity

Scanning surveys are typically used to identify small areas of elevated activity. The size of the area of elevated activity that the survey is designed to detect affects the release criteria DCGLEMC, which in turn determines the ability of a scanning technique to detect these areas. Larger areas have a lower DCGLEMC and are more difficult to detect than smaller areas.

The percentage of the survey unit to be covered by scans is also an important consideration. 100% coverage means that the entire surface area of the survey unit has been covered by the field of view of the scanning instrument. 100% scanning coverage provides a high level of confidence that all areas of elevated activity have been identified. If the available information concerning the survey unit provides information demonstrating that areas of elevated activity may not be present, the survey unit may be classified as Class 2 or Class 3. Because there is already some level of confidence that areas of elevated activity are not present, 100% coverage may not be necessary to demonstrate compliance. The scanning survey coverage may be adjusted based on the level of confidence supplied by the existing data. If there is evidence providing a high level of confidence, that areas of elevated activity are not present, 10% scanning coverage may meet the objectives of the survey. If the existing information provides a lower level of confidence, the scanning coverage may be adjusted between 10 and 100% based on the level of confidence and the objectives of the survey. A general recommendation is: always try to minimize the decision error. In general, scanning the entire survey unit is less expensive than finding areas of elevated activity later in the survey process. Finding such areas will lead to performing additional surveys due to survey unit misclassification.

Another consideration for scanning surveys is the selection of scanning locations. This is not an issue when 100% of the survey unit is scanned. Whenever less than 100% of the survey unit is scanned, a decision must be made on what areas are scanned. The general recommendation is that, when large amounts of the survey unit are scanned (e.g., > 50%), the scans should be systematically performed along transects of the survey unit. When smaller amounts of the survey unit are scanned, selecting areas based on professional judgment may be more appropriate and efficient for locating areas of elevated activity (e.g., drains, ducts, piping, ditches). A combination of 100% scanning in portions of the survey unit selected based on professional judgment and less coverage (e.g., 20-50%) for all remaining areas may result in an efficient scanning survey design for some survey units.

3.3.2.8 Determining survey and investigation levels

An important aspect of a (e.g., final status) survey is the design and implementation of investigation levels. Investigation levels are radionuclide-specific levels of radioactivity used to indicate when additional investigations may be necessary. Investigation levels also serve as a quality control check to determine when a measurement process begins to get out of control. For example, a measurement that exceeds the investigation level may indicate that the survey unit has been improperly classified – definitions of applied area classes are given in Section 3.2 – or it may indicate a failing instrument.

When an investigation level is exceeded, the first step is to confirm that the initial measurement/sample actually exceeds the particular investigation level. This may involve taking further measurements to determine that the area and level of the elevated residual radioactivity are such that the resulting dose or risk meets the release criterion1. Depending on the results of the investigation actions, the survey unit may require reclassification, remediation, and/or resurvey. Table 3.5 illustrates an example of how investigation levels can be developed.

Survey unit classification Flag direct measurement or sample result when: Flag scanning measurement result when:

Class 1 > DCGLEMC or
> DCGLW and
> a statistical parameter-based value

> DCGLEMC
Class 2 > DCGLW > DCGLW or MDC

Class 3 > fraction of DCGLW > DCGLW or MDC

Table 3.5 Example of final status survey investigation levels

When determining an investigation level using a statistical-based parameter (e.g., standard deviation) one should consider:

  • Data quality objectives for this survey objectives;
  • Underlying radionuclide distributions and an understanding of corresponding types (e.g., normal, log normal, non-parametric);
  • Statistical descriptors (e.g., standard deviation, mean, median), population stratifications (i.e., are there sub-groups present?);
  • Other prior survey and historical information. For example, a level might be arbitrarily established at the mean + 3 standard deviation of the survey unit, assuming a normal distribution. A higher value might be used if locating discrete sources of higher activity was a primary survey objective.

By the time the final status survey is conducted, survey units should be defined. Estimates of the mean, variance, and standard deviation of the radionuclide activity levels within the survey units should also be available.

For a Class 1 survey unit, measurements above the DCGLW are not necessarily unexpected. However, a measurement above the DCGLW at one of the discrete measurement locations might be considered unusual if it were much higher than all of the other discrete measurements performed during a survey. Thus, any discrete measurement that is both above the DCGLW and above the statistical-based parameter for the measurements should be investigated further. Any measurement, either at a discrete location or from a scan that is above the DCGLEMC should be flagged for further investigation.

For Class 2 or Class 3 areas, neither measurements above the DCGLW nor areas of elevated activity are expected. Any measurement at a discrete location exceeding the DCGLW in these areas should be flagged for further investigation. Because the survey design for Class 2 and Class 3 survey units is not driven by the elevated measurement criterion (EMC), the scanning minimum detectable concentration (MDC) might exceed the DCGLW. In this case, any indication of residual radioactivity during the scan would warrant further investigation.

The basis for using the DCGLEMC rather than the more conservative criteria for Class 2 and Class 3 areas should be justified in survey planning documents. For example, where there is high uncertainty in the reported scanning MDC, a more conservative criterion would be warranted.
Similarly, data quality assessment (DQA) for scanning may warrant a more conservative flag, as would greater uncertainty from historical site assessment or other surveys on the size of potential areas of elevated activity. In some cases, it may even be necessary to agree in advance with the regulatory agency responsible for the site on which site-specific investigation will be used if other than those presented in Table 3.5.

For a Class 3 area, there is a low expectation for residual radioactivity. It may be prudent to investigate any measurement exceeding even a fraction of the DCGLW. The level selected in these situations depends on the site, the radio-nuclides of concern, and the measurement and scanning methods selected. This level should be set using the DQO Process during the survey design phase of the Data Life Cycle. In some cases, the user may also wish to follow this procedure for Class 2 and even Class 1 survey units.

3.3.2.9 Development of an integrated site characterization strategy

The final step in survey design is to integrate the different selected survey techniques (see Section 3.6 and 3.7) with the number of measurements and measurement spacing (see Section 3.5). This integration along with the guidance provided in other portions of this manual produce an overall strategy for performing the survey. Table 3.6 provides a summary of the recommended survey coverage for structures and land areas. This survey coverage for different areas is the subject of this section.

Area classification Structures Land areas
Surface scans Surface activity measurements

Surface scans Soil samples
Class 1 100% Number of data points from statistical tests (Sections 3.5.1); additional measurements may be necessary for small areas of elevated activity (Sections 3.5.1). 100% Number of data points from statistical tests (Sections 3.5.1); additional measurements may be necessary for small areas of elevated activity (Sections 3.5.1).

Class 2 10 to 100%
(10 to 50% for upper walls and ceilings) Systematic and judgmental

Number of data points from statistical tests (Sections 3.5.1); 10 to 100% Systematic and judgmental Number of data points from statistical tests (Sections 3.5.1).
Class 3 Judgmental Number of data points from statistical tests (Sections 3.5.1);

Judgmental Number of data points from statistical tests (Sections 3.5.1).

Table 3.6 Recommended survey coverage for structures and land areas

To account for assumptions used to develop the DCGLW 2 and the realistic possibility of small areas of elevated activity, an integrated survey design should be developed to include all of the design considerations. An integrated survey design combines a scanning survey for areas of elevated activity with random measurements for relatively uniform distributions of contamination. Table 3.7 presents the recommended conditions for demonstrating compliance for a final status survey based on classification.

Random measurement patterns are used for Class 3 survey units to ensure that the measurements are independent and meet the requirements of the statistical tests. Systematic grids are used for Class 2 survey units because there is an increased probability of small areas of elevated activity. The use of a systematic grid allows the decision maker to draw conclusions about the size of any potential areas of elevated activity based on the area between measurement locations, while the random starting point of the grid provides an unbiased method for determining measurement locations for the statistical tests.

Survey unit classification Statistical test Elevated measurement comparison Sampling and/or direct measurements

Scanning
Impacted Class 1 Yes Yes Systematic 100% Coverage

Class 2 Yes Yes Systematic 10-100% Systematic

Class 3 Yes Yes Random Judgemental

Non-impacted No No No None

Table 3.7 Recommended conditions for demonstrating compliance based on survey unit classification for a final survey

Class 1 survey units have the highest potential for small areas of elevated activity, so the areas between measurement locations are adjusted to ensure that these areas can be identified by the scanning survey if the area of elevated activity is not detected by the direct measurements or samples.

The data quality objectives of the scanning surveys are different. Scanning is used to identify locations within the survey unit that exceed the investigation level. These locations are marked and receive additional investigations to determine the concentration, area, and extent of the contamination.

For Class 1 areas, scanning surveys are designed to detect small areas of elevated activity that are not detected by the measurements using the systematic grids. For this reason, the measurement locations and the number of measurements may need to be adjusted based on the sensitivity of the scanning technique (see Section 3.5). This is also the reason for recommending 100% coverage for the scanning survey. 100% coverage means that the entire surface area of the survey unit is covered by the field of view of the scanning instrument. If the field of view is two meters wide, the survey instrument can be moved along parallel paths/spacing of two meters apart to provide 100% coverage. If the field of view of the detector is 5 cm, the parallel paths/spacing should be 5 cm apart.

Scanning surveys in Class 2 areas are also performed primarily to find areas of elevated activity not detected by the measurements using the systematic pattern. However, the measurement locations are not adjusted based on sensitivity of the scanning technique, and scanning is only performed in portions of the survey unit. The level of scanning effort should be proportional to the potential for finding areas of elevated activity based on the conceptual model. In Class 2 survey units that have residual radioactivity close to the release criterion a larger portion of the survey unit would be scanned, but for survey units that are closer to background scanning a smaller portion of the survey unit may be appropriate. Class 2 survey units have a lower probability for areas of elevated activity than Class 1 survey units, but some portions of the survey unit may have a higher potential than others. Judgmental scanning surveys would focus on the portions of the survey unit with the highest probability for areas of elevated activity. If the entire survey unit has an equal probability for areas of elevated activity, or the judgmental scans don’t cover at least 10% of the area, systematic scans along transects of the survey unit or scanning surveys of randomly selected grid blocks are performed.

Class 3 areas have the lowest potential for areas of elevated activity. For this reason, EURSSEM recommends that scanning surveys be performed in areas of highest potential (e.g., corners, ditches, drains) based on professional judgment. Such recommendations may be typically provided by a health physics professional with radiation survey experience. This provides a qualitative level of confidence that no areas of elevated activity were missed by the random measurements or that there were no errors made in the classification of the area.

The sensitivity for scanning techniques used in Class 2 and Class 3 areas is not tied to the area between measurement locations, as they are in a Class 1 area (see Section 3.5). The scanning techniques selected should represent the best reasonable effort based on the survey data quality objectives. Structure surfaces are generally scanned for alpha, beta, and gamma emitting radio-nuclides. Scanning for alpha emitters or low-energy (< 100 keV) beta emitters for land area survey units is generally not considered effective because of problems with attenuation and media interferences. If one can reasonably expect to find any residual radioactivity, it is prudent to perform a judgmental scanning survey.

If the equipment and methodology used for scanning is capable of providing data of the same quality as direct measurements (e.g., detection limit, location of measurements, ability to record and document results), then scanning may be used in place of direct measurements. Results should be documented for at least the number of locations estimated for the statistical tests. The same logic can be applied for using direct measurements instead of sampling. In addition, some direct measurement systems may be able to provide scanning data.

As previously discussed, investigation levels are determined and used to indicate when additional investigations may be necessary or when a measurement process begins to get out of control. The results of all investigations should be documented in the final status survey report, including the results of scan surveys that may have potentially identified areas of elevated direct radiation.

3.3.2.9.1 Land area surveys

Class 1 areas
100% scanning coverage of Class 1 land areas is recommended. Locations of scanning survey results above the investigation level are identified and evaluated. Results of initial and follow-up direct measurements and sampling at these locations are recorded. Soil sampling is performed at locations identified by scans and at previously determined locations (Section 3.5.1). Where gamma emitting radio-nuclides are contaminants, in situ gamma spectroscopy may be used to confirm the absence of specific radio-nuclides or to demonstrate compliance.

Direct measurement or sample investigation levels for Class 1 areas should establish a course of action for individual measurements that approach or exceed the DCGLW. Because measurements above the DCGLW are not necessarily unexpected in a Class 1 survey unit, additional investigation levels may be established to identify discrete measurements that are much higher than the other measurements. Any discrete measurement that is both above the DCGLW and exceeds three standard deviations above the mean should be investigated further (see Table 3.5). Any measurement (direct measurement, sample, or scan) that exceeds the DCGLEMC should be flagged for further investigation. The results of the investigation and any additional remediation that was performed should be included in the final status survey report. Data are reviewed as described in Section 3.10.8.4, additional data are collected as necessary, and the final complete data set evaluated as described in Section 3.10.3 or Section 3.10.4.

Class 2 areas
Surface scans are performed over 10 to 100% of open land surfaces. Locations of direct radiation above the scanning survey investigation level are identified and evaluated. If small areas of elevated activity are identified, the survey unit should be reclassified as “Class 1” and the survey strategy for that survey unit redesigned accordingly.

If small areas of elevated activity above DCGL values are not identified, direct measurement or soil sampling is performed at previously determined locations (Section 3.5.1). Where gamma emitting radio-nuclides are contaminants, in situ gamma spectroscopy may be used to confirm the absence of specific radio-nuclides or to demonstrate compliance. Data are reviewed as described in Section 3.10.8.4, additional data are collected as necessary, and the final complete data set evaluated as described in Section 3.10.3 or Section 3.10.4.

Investigation levels for Class 2 areas should establish levels for investigation of individual measurements close to but below the DCGLW. The results of the investigation of the positive measurements and basis for reclassifying all or part of the survey unit as Class 1 should be included in the final status survey report.

Class 3 areas
Class 3 areas may be uniformly scanned for radiations from the radio-nuclides of interest, or the scanning may be performed in areas with the greatest potential for residual contamination based on professional judgment and the objectives of the survey. In some cases a combination of these approaches may be the most appropriate. Locations exceeding the scanning survey investigation level are evaluated, and, if the presence of contamination not occurring in background is identified, re-evaluation of the classification of contamination potential should be performed.

Investigation levels for Class 3 areas should be established to identify areas of elevated activity that may indicate the presence of residual radioactivity. Scanning survey locations that exceed the investigation level should be flagged for further investigation. The results of the investigation and basis for reclassifying all or part of the survey unit as Class 1 or Class 2 should be included in the final status survey report. The data are tested relative to the pre-established criteria. If additional data are needed, they should be collected and evaluated as part of the entire data set. Soil sampling is performed at randomly selected locations (Section 3.5.1); if the contaminant can be measured at DCGL levels by in-situ techniques, this method may be used to replace or supplement the sampling and laboratory analysis approach. For gamma emitting radio-nuclides, the above data should be supplemented by several exposure rate and/or in-situ gamma spectrometry measurements. Survey results are tested for compliance with DCGLs and additional data are collected and tested, as necessary.

3.3.2.9.2 Structure Surveys

Class 1 areas
Surface scans are performed over 100% of structure surfaces for radiations which might be emitted from the potential radionuclide contaminants. Locations of direct radiation, distinguishable above background radiation, are identified and evaluated [USNRC-2006]. Results of initial and follow-up direct measurements and sampling at these locations are recorded and documented in the final status survey report. Measurements of total and removable contamination are performed at locations identified by scans and at previously determined locations (Section 3.5.1). Where gamma emitting radio-nuclides are present, in-situ gamma spectroscopy may be used to identify the presence of specific radio-nuclides or to demonstrate compliance with the release criterion.

Direct measurement or sample investigation levels for Class 1 areas should establish a course of action for individual measurements that approach or exceed the DCGLW. Because measurements above the DCGLW are not necessarily unexpected in a Class 1 survey unit, additional investigation levels may be established to identify discrete measurements that are much higher than the other measurements. Any discrete measurement that is both above the DCGLW and exceeds three times the standard deviation (s) of the mean should be investigated further (Section 3.3.2.7). Any measurement (direct measurement, sample, or scan) that exceeds the DCGLEMC should be flagged for further investigation. The results of the investigation and any additional remediation that was performed should be included in the final status survey report. Data are reviewed as described in Section 3.10.8.4, additional data are collected as necessary, and the final complete data set evaluated as described in Section 3.10.3 or Section 3.10.4.

Class 2 areas
Surface scans are performed over 10 to 100% of structure surfaces. Generally, upper wall surfaces and ceilings should receive surface scans over 10 to 50% of these areas. Locations of scanning survey results above the investigation level are identified and investigated. If small areas of elevated activity are confirmed by this investigation, all or part of the survey unit should be reclassified as Class 1 and the survey strategy for that survey unit redesigned accordingly.

Investigation levels for Class 2 areas should establish a course of action for individual measurements that exceed or approach the DCGL~W~. The results of the investigation of the positive measurements and basis for reclassifying all or part of the survey unit as Class 1 should be included in the final status survey report. Where gamma emitting radio-nuclides are contaminants, in-situ gamma spectroscopy may be used to identify the presence of specific radio-nuclides or to demonstrate compliance with the release criterion. Data are reviewed as described in Section 3.10.8.4, additional data are collected as necessary, and the final complete data set evaluated as described in Section 3.10.3 or Section 3.10.4.

Class 3 areas
Scans of Class 3 area surfaces should be performed for all radiations which might be emitted from the potential radionuclide contaminants. EURSSEM recommends that the surface area be scanned. Locations of scanning survey results above the investigation level are identified and evaluated. Measurements of total and removable contamination are performed at the locations identified by the scans and at the randomly selected locations that are chosen in accordance with Section 3.5.1. Identification of contamination suggests that the area may be incorrectly classified. If so, a re-evaluation of the Class 3 area classification should be performed and, if appropriate, all or part of the survey unit should be resurveyed as a Class 1 or Class 2 area. In some cases the investigation may include measurements by in-situ gamma spectroscopy at a few locations in each structure in a Class 3 area. A gamma spectroscopy system might even be an appropriate substitution for surface scans.

Because there is a low expectation for residual radioactivity in a Class 3 area, it may be prudent to investigate any measurement exceeding even a fraction of the DCGLW. The investigation level selected will depend on the site, the radio-nuclides of concern, and the measurement and scanning methods chosen. This level should be determined using the DQO Process during survey planning. In some cases, the user may wish to follow this procedure for Class 2 survey units.

The results of the investigation of the measurements that exceed the investigation level and the basis for reclassifying all or part of the survey unit as Class 1 or Class 2 should be included in the final status survey report. The data are tested relative to the pre-established criteria. If additional data are needed, they should be collected and evaluated as part of the entire data set.

3.3.2.10 Other survey designs

The survey design in EURSSEM is based on six principal steps. Although the process is described sequential, EURSSEM is not intended to be a serial process that would slow site clean-ups. Rather, EURSSEM supports existing programs and encourages approaches to expedite site clean-ups. Part of the significant emphasis on planning in EURSSEM is meant to promote saving time and resources.

There are a number of approaches designed to expedite site clean-ups and characterizations. These approaches/methodologies can save time and resources by reducing sampling, preventing duplication of effort, and reducing inactive time periods between steps in a clean-up process. While differing in details, these methodologies have some common features:

  • Decision making processes that affect sampling are determined before going to the field, but actual sampling decisions of “where” and “how many” are made in the field in “real time” by experts on the basis of evolving sampling results (such sampling and analysis plans are known as “dynamic” or “flexible” sampling plans).
  • Regulator approval for the “science-based” approach over the traditional step-by-step approach in which regulators approve each phase of sampling before it is undertaken.
  • Use of a suite of non-invasive and minimally-invasive technologies and field screening supported, when possible, by high quality on-site sample analysis with smaller amounts of verification sample analysis in off-site laboratories.
  • Technology for efficient management, visualization, and interpretation of data to facilitate on-site, “real time” decision making.

Summaries of alternate clean-up approaches/methodologies are given below:

  • Observational approach. The observational approach draws on tenets of geotechnical engineering in which it is accepted that the subsurface environment can never be reasonably sampled enough to create a conceptual model that contains no uncertainty. Geotechnical engineering deals with this uncertainty by designing subsurface building structures based on the “nominal” conditions and preparing contingency plans to handle the uncertainties should they be encountered in construction. This approach uses an iterative process of sample collection and real-time data evaluation to characterize a site. This process allows early field results to guide later data collection in the field. Data collection is limited to only that required for selecting a unique remedy for a site 3.The application of this approach to remediation of contaminated site stresses accelerating characterization to determine only the nominal conditions needed for design of a specific remediation system and providing remedial contingency designs to be employed should nominal conditions not pertain. Applications of the observational approach have been made to both radiological and non-radiological contamination problems [IAEA-1998a], [Wallace], [Smyth], [Belcher].
  • The Superfund Accelerated Clean-up Model (SACM), which includes a module called integrated site assessment, has as its objectives increased efficiency and shorter response times [EPA-1992], [EPA-1993], [EPA-1997].
  • Tri-Parti Agreement Negotiation Approach. At DOE’s Hanford Site, the parties to the Tri-Party Agreement negotiated a method to implement the CERCLA process in order:
    • To accelerate the assessment phase;
    • To co-ordinate RCRA and CERCLA requirements whenever possible, thereby resulting in cost savings. The Hanford Past Practice Strategy (HPPS) was developed in 1991 to accelerate decision making and initiation of remediation through activities that include maximizing the use of existing data consistent with data quality objectives4, 5.
  • Streamlined approach for environmental restoration (SAFER). The US Department of Energy created the so-called SAFER approach (streamlined approach for environmental restoration) which combined the bias for implementing remediation with accelerated characterization that relies heavily on the data quality objective (DQO) approach. The results of applications have been faster and less costly characterization (and potentially smaller total remediation costs) [Gianti], [Bottrel].
  • Expedited site characterization (ESC) stresses taking a multi-disciplinary team of technical experts to the field to minimize the number of phases of characterization. The team members are very well versed in the site history, have an initial conceptual model of the site environment, are equipped with a suite of non-invasive and invasive technologies, and are prepared to carry out a dynamic sampling effort that may be adjusted daily as sampling results become available. ESC has been particularly effective in accelerating and improving the characterization of the subsurface environment in cases of groundwater contamination. An appropriate combination of geological, geophysical, hydrogeological, and geochemical investigations is bought to bear concurrently as the study identifies and focuses on critical parameters [Aggerwal], [Burton-1994, 1994a, 1994b, 1995].
  • Adaptive sampling and analysis (or, as sometimes referred to as ASAP, for “adaptive sampling and analysis programme developed at the Environmental Assessment Division (EAD) of Argonne National Laboratory5”.) fuse soft data (for example, historical records, aerial photos, non-intrusive geophysical data) with hard sampling results to estimate contaminant extent, measure the uncertainty associated with these estimates, determine the benefits from collecting additional samples, and assist in siting new sample locations to maximize the information gained. ASAP exploits the opportunity for in-the-field decision making when field analytical and screening instrumentation can provide rapid results regarding contamination levels. The decision-making regarding sample location and number is facilitated by a decision support system that uses the results of radiological or chemical analyses and other site information to estimate the extent of contamination [Johnson-1993]. It also calculates the level of uncertainty associated with the estimate of extent. The system provides visualization of the data, contamination extent, and uncertainty. Just as important, it indicates where the next sampling should occur to have the greatest impact on reducing the uncertainty in the estimate of the extent of contamination. The system successively updates the prediction of new sampling locations after each set of new data is gathered and the estimate of contamination is refined. In several cases of soil contamination, rapid rounds of iterative sampling guided by the adaptive sampling and analysis system have resulted in delineation of contamination with costs as low as 25-40% of the originally predicted sampling and analysis costs for a traditional uniform-grid sampling program [Johnson-1993a-b, 1994, 1996], [Robbat].

1 Rather than, or in addition to, taking further measurements the investigation may involve assessing the adequacy of the exposure pathway model used to obtain the DCGL’s and area factors, and the consistency of the results obtained with the Historical Site Assessment and the scoping, characterization and remedial action support surveys.

2 Note that the DCGL itself is not free of error. The assumptions made in any model used to develop DCGLs for a site should be examined carefully. The results of this examination should determine if the use of site-specific parameters results in large changes in the DCGLs, or whether a site-specific model should be developed to obtain DCGLs more relevant to the exposure conditions at the site. Appendix A provides additional information about the uncertainty associated with the DCGL and other considerations for developing an integrated survey design using the DQO Process.

3 Information on the Observational Approach recommended by Sandia National Laboratories is available on the internet here.

4 Information on the Hanford Past Practice Strategy is available on the internet here.

5 Information on the Argonne National Laboratory adaptive sampling programs can be obtained on the internet here.

When the null hypothesis is "the survey unit is contaminated2, acceptable values for the probabilities of making Type I (α) decision errors are 0.05 or less ". However, acceptable values for the probabilities of making Type II (β) decision errors are not fixed. Usual beta values are from 0.5 to 0.05. according the formulae used to determine the sample size (n) if beta is a high value , the sample size will be lower making more difficult the satisfactory (clearance) result of the test.
– by Rafael Garcia-Bermejo Fernandez about 6 years ago