3
Estimating and Confirming The Source Term

THE SOURCE TERM IS THE amount of radionuclides released from a site to the environment over a specific period. The rate of release as a function of time should also be determined. Releases can be to the atmosphere, to surface waters, to groundwater, or to soil. In some cases, people might have been exposed directly.

A full description of the source term includes what was released and in what form and where and when the release occurred. These factors must be described with enough accuracy and detail both to satisfy the scientific requirements for the design and conduct of epidemiologic studies and to address the public's concern. Satisfying public concern can be difficult given the inherent uncertainty of retrospective dose reconstruction studies. In general, scientific requirements are most likely to be satisfied when it is possible to estimate the source term by several different methods. Usually, a scoping study should precede a full description of the source term; the objective of a scoping study is to provide a preliminary assessment of the source term and the magnitude of the estimated exposure to be used in further decision-making. Ideally, the results of a comprehensive source term analysis will contain all of the information needed for exposure or dose calculations, with a spatial and time resolution sufficient for the requirements of the epidemiologic study.

This chapter is concerned with general approaches to estimating releases rather than with the development of specific source terms. The following sections discuss locations and characteristics of releases; information needed for source term estimates and where that information can



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 16
3 Estimating and Confirming The Source Term THE SOURCE TERM IS THE amount of radionuclides released from a site to the environment over a specific period. The rate of release as a function of time should also be determined. Releases can be to the atmosphere, to surface waters, to groundwater, or to soil. In some cases, people might have been exposed directly. A full description of the source term includes what was released and in what form and where and when the release occurred. These factors must be described with enough accuracy and detail both to satisfy the scientific requirements for the design and conduct of epidemiologic studies and to address the public's concern. Satisfying public concern can be difficult given the inherent uncertainty of retrospective dose reconstruction studies. In general, scientific requirements are most likely to be satisfied when it is possible to estimate the source term by several different methods. Usually, a scoping study should precede a full description of the source term; the objective of a scoping study is to provide a preliminary assessment of the source term and the magnitude of the estimated exposure to be used in further decision-making. Ideally, the results of a comprehensive source term analysis will contain all of the information needed for exposure or dose calculations, with a spatial and time resolution sufficient for the requirements of the epidemiologic study. This chapter is concerned with general approaches to estimating releases rather than with the development of specific source terms. The following sections discuss locations and characteristics of releases; information needed for source term estimates and where that information can

OCR for page 16
be found; biases and uncertainties in historic measurements; and the uses of interpolation, extrapolation, and modeling for estimating source terms. APPROACH TO SOURCE TERM ANALYSIS There are, in principle, three ways to determine the source term in the case of specific facility—from engineering estimates of what the operation was expected to release, from historic reports of measured releases, and from reconstruction that uses independent measurements of environmental quantities related to the release. Not all of these ways will be possible in all cases, but to the extent they are, redundant analyses are desirable. A source term analysis consists of a mix of information about the physical, engineering, and chemical elements of a source. The information comes from models or from historic records provided by production data, from environmental monitoring of the amounts and forms of released materials and the environmental pathways into which the material was introduced (air, water, land), and from contemporary measurements made in the environment at the time of the dose reconstruction. The quality and completeness of the estimates of source terms, based on the physical and chemical characteristics of the source supplemented by historic, operational, and release data and those based on any available independent environmental measurements, vary with the facility or site being evaluated. In some instances it will be sufficient to bound the problem based on source estimates derived from engineering data and then to determine whether the public exposure, evaluated from such data, was great enough to warrant further epidemiologic investigation. If further investigation is justified, then, in general, it will be necessary to develop the source term more precisely. This will require use of all relevant data to identify what was actually released, and when source term contributions from different streams must be quantified, their statistical ranges must be determined for each period of interest. The accuracy of the estimate must be determined as realistically as possible. No purpose is achieved by exaggerating the accuracy of the estimate; public confidence is better served by being candid about uncertainty. It is also important to use source term data in concert with information about potential environmental pathways. For example, if stored solid materials have not leached to groundwater pathways, further characterization of release to groundwater is unnecessary because no off-site release has occurred. Such a process sets up a feedback mechanism for increasing or decreasing the scope of the source term analysis. Finally, the source term, which is quantified from a combination of historic records and engineering estimates, can be further confirmed or supplemented by estimates derived

OCR for page 16
from environmental pollutant concentration data and dispersion calculations. Scoping studies are useful for determining the degree of effort required to conduct comprehensive source term studies, provided the goals of the dose reconstruction studies are clearly stated. A scoping study would use physical and engineering elements of the process or site and knowledge of operations over time to provide the initial input to a decision about whether to conduct an epidemiologic investigation. By its nature, a scoping study should be a conservative estimate that is used to preclude making an incorrect decision not to investigate the consequences of a particular release, either in terms of individual radiation exposures or as an epidemiologic study. Scoping studies also can provide guidance about which releases and pathways are of greatest importance in a more comprehensive analysis. One approach to the initial phase of a comprehensive source term study is to survey all available records and to extract and evaluate all relevant data. This is a time-consuming and labor-intensive effort, requiring the locating of records that often date back many decades and are in most cases poorly indexed or identified. It also requires trained and technically competent personnel to review many documents to identify and catalogue any significant information. Because of the sheer number of plant records for some sites, a searching strategy must be developed for the retrieval of all relevant data. The credibility of a comprehensive source term study depends on confirming that all pertinent documents have been seen and evaluated. Complete records are essential in identifying the source term. The major advantage of a comprehensive search is that it can eliminate bias in use of the data by allowing interpretation of all available data to determine the elements and conduct of the analysis of source term parameters, pathways, and persons exposed. A comprehensive search can be especially valuable when classified or proprietary information is involved. The major disadvantage of a comprehensive search is that it delays decisions on major elements of dose reconstruction and epidemiologic studies until after the data collection has been completed, and well after a significant expenditure of time and resources. Scoping studies conducted with care and public sensitivity could offset this potential disadvantage. Selective searches tend to be more efficient and less costly and for the most part provide the necessary data, because often much is known about the operations and history of the facility or site selected for study. Regardless of the method used to estimate the source term, it is essential that the process involve public dialogue, especially to address public reports on any events observed, remembered, or suspected. Each event

OCR for page 16
should be investigated and its contribution to the source term estimate included, either as a documented release or as a gap in the data set that is bracketed by other analyses. Proprietary or classified information must be reviewed. When exposure reconstruction is indicated for a site or facility, it is essential that the reported source term and the data upon which it is based be removed from the arena of proprietary information, including security classification, if the study is to satisfy public scrutiny and achieve acceptance. This means that decision-makers will need to strike a balance between potentially conflicting goals: protecting proprietary or security information and achieving public confidence in the analysis performed. In most cases it will be possible to make essential information available and to explain why the information is sufficient for the estimate. One effective method, when some material cannot be widely disseminated, is to have a team of people with the appropriate security clearance review the classified or proprietary information and report to the advisory or steering group whether that information is or is not essential to the source term estimate. It also is possible to agree on a cutoff date beyond which proprietary data would not be essential because of the availability of alternative confirming environmental or release measurements. DATA REQUIREMENTS FOR SOURCE TERM ANALYSIS For a study to reconstruct a source term, all aspects that affected the releases must be considered and no relevant data can be overlooked. Because of the sheer quantity of documents that exist for some sites, the search process, whether selective or comprehensive, must ensure that all relevant documents pertinent to the period of interest are inspected and all useful information is extracted. The process should, to the extent feasible, obtain information from at least two different sources, such as measurements and process inventory, to ensure independent confirmation of source term data. Independent confirmation of release data should be given a high priority. As far as possible, it is desirable to use original data, such as monitoring logs, rather than official summaries or reports. This might entail some judgment on the consistency or accuracy of records derived from different sources, but the reconstruction process itself should not attempt to edit the information unless there are obvious errors. The process of reconstructing data to fill in gaps should assess whether the contribution will be of epidemiologic significance. To assist in the modeling of the environmental transport of the released material, it is often important to include data about the physical and chemical characteristics of the released material. Unless the process

OCR for page 16
or plant has changed over time, this information can be included generically in the source terms. Short periods of missing data can be bridged by interpolation between periods for which discharges are known. Longer periods might require extrapolation of release rates using changes in plant operations as a guide. For releases that were never measured or that were sampled only occasionally, it might be necessary to develop a model of the release that is related to the production process that caused it. The form of the model will likely depend on the data that are available to guide the calculations. Monte Carlo calculations can be used to derive best estimates, uncertainties, or probability distributions that characterize the particular source term. Points of releases are particularly important; these should be established from plans of the site and facilities, information on processing activities, process flow sheets, and facility drawings. On-site inspection and mapping of the premises and remaining facilities is an essential part of such determinations. Airborne releases of contaminants occur at a variety of locations, depending on the facility and the nature of its processes. Most routine releases of toxic chemicals or radionuclides have been discharged through chimneys or roof vents designed to exhaust process areas or equipment. The exhaust paths often used filtration systems to retain the pollutants of concern. However, other pathways, such as open doors, laboratory hoods, and air conditioners, could have contributed to emissions. Atmospheric releases require specification of the source term from the height, diameter, air flow rate, and temperature in the stack and the contaminant concentrations in the discharged air. Information about the size distribution and chemical form of discharged aerosols and particles is important. Chemical forms of discharged gases also are important. Releases at ground-level and in the open are characterized by release rates, by an effective height of the release point, and by information on chemical form and particle size, as appropriate. Uncontrolled processes, such as burning of contaminated materials, also lead to atmospheric contamination. Other examples of uncontrolled processes are evaporation of volatile solvents spilled outside or discharged into a surface liquid waste stream and resuspension of contaminated particles from waste disposal pits or dried-out areas surrounding disposal ponds. Liquid releases can occur from direct piping of waste products into a body of surface water or to a discharge point below the surface of a river, lake, or ocean. Disposal wells have been used to carry liquids into aquifers or to develop a water zone perched above an aquifer, and semisolids have been pumped into disposal ponds, lagoons and hydraulic fracture

OCR for page 16
zones from which migration has occurred. In the latter case, adsorption of pollutants might have taken place in the soil column above the aquifer. Unlined disposal ponds and settling basins that were used to reduce the quantity of pollutants that reached aquifers or downstream surface water bodies, might have produced contamination of underlying soil layers. Liquid effluent discharge rates and contaminant concentrations are two primary characteristics that should be documented. The fractions of the release that are dissolved and suspended are also important and should be identified if possible. Knowledge of the chemical forms of some pollutants can be critical to the exposure assessment. When liquid releases are initially directed to an on-site impoundment, assessing the off-site release requires documentation of the flow rate and concentrations leaving the impoundment. Similar considerations apply for evaluating concentrations of contaminants that enter an aquifer after they pass through a column of soil or rock above an aquifer. Solid material releases have included a variety of contaminated materials disposed of in pits or trenches. The method used to emplace the waste and the cover provided influence the amount of airborne contamination generated during or after disposal, potentially leading to soil contamination. Spills of process chemicals or radioactive materials also have produced soil contamination, part of which could remain to the date of the study and could be measured for confirmation of the release pathway. The physical and chemical forms of contaminants in solid wastes are important determinants of their environmental transport following emplacement. Some contaminants are incorporated in solids, in ion exchange resins, or in finely divided process waste materials. Knowledge of the distribution of the contaminants and the leachability of the waste form is highly desirable. Particle size is particularly important because of potential airborne transport. EPISODIC RELEASES It is important to distinguish episodic releases that justify special treatment in dose or exposure assessment, from the continuous releases at a facility. Episodic releases are those during which the release rate was at least 10 times greater than the average monthly or yearly release from the facility and that lasted fewer than 10 days. While use of average dispersion parameters is considered appropriate for high release rates of longer duration (more than 10 days), specific meteorologic or transport parameters for the time of the episodic events should be established and used if possible.

OCR for page 16
SOURCES OF INFORMATION As previously stated, three independent data sets can be used to determine the source term: historic records, engineering estimates, and environmental monitoring data. Use of historic records involves collecting information on releases from stacks and other discharge structures or effluent pathways. This information can be partially confirmed or supplemented by engineering estimates, which are derived primarily from process information and environmental control technology performance. The resulting source term, based on a combination of historic records and engineering estimates, can be further confirmed or supplemented by estimates derived from environmental monitoring data. In some cases, the source term can be reconstructed by use of environmental pollutant concentration data and dispersion calculations. An important first step in the source term estimation process is to locate measurements and estimates made by plant operators during the period of interest. After 1960, such estimates were frequently documented in environmental monitoring reports. More detailed information is available for years after the passage of the National Environmental Policy Act of 1969. Routine reports of release estimates are sometimes available in the files of a facility's health physics or industrial hygiene department. However, it is recommended that the formal reports of effluent releases be validated by examination of the original, handwritten logbooks, files, or analytic data sheets that contain sample mass, activity, or concentration measurements. These original data also can indicate, directly or indirectly, the uncertainty associated with the measurements. Daily waste-processing logs often contain detailed information about transfers of liquid effluents. Unplanned releases may be documented in several ways. The least formal source is found in routine production reports or in periodic reports of the cognizant safety organization. Events involving larger releases are usually discussed in some detail in incident reports or accident investigation reports. In recent years, data bases of unusual occurrences have been established and updated routinely. Very large releases and their causes are almost certain to have been evaluated by an investigation team and formally documented. Interviews with current and former employees who witnessed or investigated such events can provide useful additional information. These persons often have information not found in formal reports. They also can offer information about routine plant operations and about incidents of lesser consequence that might have occurred more frequently. For some chemical releases, it is possible to estimate the source term by using logbooks of material usage or transfer and information about the

OCR for page 16
physical and chemical characteristics of the process and effluent treatment systems. For long-lived radionuclides or chemicals, source terms can be ''back calculated" using transport models from measurements of soil or sediment deposition, if such data are available. The magnitudes of episodic releases can be estimated from measurements of the contaminant in air, water, or biota during or shortly after the event. The success of these approaches depends on the quality of the data and the transport model used in the calculation and on the mobility and persistence either of the contaminant of interest or of a surrogate tracer in the environment. Some of the information obtained in source term studies is useful for model development, especially data found in production files, specific process data, purchasing records, materials accountability reports, and logbooks of material use or transfer. In some cases, detailed information on reactor operating cycles and fuel cooling times will be needed to quantify estimates of short-lived nuclides. Other models will not require such detailed information. In many cases, operational data will no longer be available and the modeling effort will reflect the limitations imposed by the data base. Research and development reports that describe prototype facilities and their operation can provide useful information for understanding releases from the process equipment. These reports typically describe initial production runs and the problems that were encountered. Measurements of release fractions during pilot plant or early start-up operations are particularly useful, but records of release fractions of particles and volatile radionuclides measured in later years can be helpful as well. Reports that describe effluent treatment systems or changes in them are especially valuable, as are operational measurements. BIAS AND UNCERTAINTY IN RELEASE ESTIMATES Release estimates inevitably are subject to uncertainty and can be biased because of unrepresentative effluent sampling or because of the occurrence of accidents in which material was released under less-than-ideal conditions for measurement. For example, a collected sample might not have been representative if the effluent discharge stream or the contents of the holding tank being sampled were not well mixed. For airborne effluents, another source of possible bias is unequal rates of sample withdrawal. Biases can be positive or negative, and they can lead to over-or underestimation of the releases. Corrections for such biases can be inadequate if information about the relevant parameters is limited. Lack of monitoring data is another source of uncertainty in release estimates. In some cases, releases were not measured, and sometimes the

OCR for page 16
records of measurements have been lost or destroyed. Data might be missing for only a brief period because of a sampler failure or there can be substantial uncertainty in the release estimates that must therefore be developed without effluent-monitoring data. Incomplete information about processes and about the historic details of facility operations also contributes to uncertainty. Even when the samples extracted were representative of the effluent streams, there are other sources of bias and uncertainty. Part of the sample might have been lost in the sampling line, leading to an underestimate of the release. Incomplete collection of the sample by the filter also would cause an underestimate. Uncertainties also arise when estimating correction factors for line loss and collection efficiency. Uncertainty also can come from the measurement of the quantity of material collected in the sample and of the volume of the sample. In many cases, only the analytic measurement uncertainty was recorded and included in reports of releases. This sometimes changed over time because of improvements in measurement equipment and in the development of new analytic techniques that led to more isotope-specific analyses. GAPS IN RELEASE DATA For some periods of time or for some types of operations, production, release, or emission data might not be available. This could result from loss of documents, destruction of documents considered obsolete, or from security concerns. Gaps in information should be filled, as far as possible, by extrapolation of release data obtained for comparable operating periods or by reconstruction from other data such as exposure records, environmental monitoring results, or waste shipment records. In some cases steady-state conditions of operations and releases will be sufficiently well established to permit interpolation or extrapolation to undocumented periods. Whenever this is done, the methodology used should be fully described and justified. Fundamental physical principles can be used when the facility's processes are well understood. If a reactor was the original source of the release, fuel burn-up calculations can be made to estimate core inventories, and releases can be estimated once the cooling time is determined and the separation process is described. This methodology was used successfully for reconstructing 131I releases from the Hanford, Washington, reactor (Heeb and Morgan 1991). For simpler situations, release fractions can be estimated from records that detail the throughput of the material, its physical and chemical characteristics, and the facility's effluent treatment systems. Research and development reports can provide

OCR for page 16
facility descriptions, information about effluent control technology, and plant operation data. SUMMARY AND RECOMMENDATIONS The source-term estimate reported in a dose reconstruction study should provide information about the quantities of materials released in a form that is suitable for the environmental pathway models used. For this reason there should be a connection between the source term analysis and the development of the pathway models so that the arrangement and presentation of the various source streams will be appropriate. Because the data analysis might lead to a decision not to pursue further study of some minor releases, it could be important to report their magnitude and when they occurred. The same could apply to releases that mostly affected the plant site itself. In a detailed study, completeness of data and comprehensive estimates of reported releases should be the goal of the reported source term; this is ensured by using all relevant data, by investigating all reported events and data gaps, and by determining each release value by as many independent means as possible. The following six recommendations are made by this committee to ensure that the source term is as complete as necessary for the goals of a study. The source term often provides the foundation of a dose reconstruction study. Scoping studies should be the primary approach to initiating a source term evaluation. These should be followed, if appropriate, by more comprehensive studies. The scoping study of the source term should seek to generate the data needed to identify the environmental pathways of potential importance and to permit estimation of the concentrations of radionuclides to which the public might have been exposed. To ensure maximum confidence in the source term analysis, proprietary or classified information should be made available for the analysis, or a mechanism should be developed to determine whether such data are essential to the accuracy and consistency of the analysis. The source term should be derived chiefly from available original data in as many different ways as practical. The source term should be confirmed, wherever possible, by comparison with independent environmental monitoring data from another source of information. To ensure completeness and accuracy of the estimated source term, all relevant data should be evaluated. Any gaps in the data should be analyzed carefully for their significance and filled by reconstruction from existing data if appropriate. Episodic events should be documented as separate releases for

OCR for page 16
specific consideration in environmental transport and dose calculations. An event is considered episodic if it lasted for less than 10 days and if the release rate was at least 10 times the average monthly or annual rate. The release quantities provided for use in a comprehensive study of the source term should be complete, unbiased estimates of all amounts and forms of relevant materials released to the environment.