VI
Performance Assessment

The statement of task charges the committee with evaluating: “… (2) any actions additional to those contained in current plans that the Department should consider to ensure that its plans to manage its radioactive waste streams will comply with the performance objectives of Part 61 of Title 10, Code of Federal Regulations; … (4) existing technology alternatives to the current management plan for the waste streams mentioned above and, for each such alternative, an assessment of the cost, consequences for worker safety, and long-term consequences for environmental and human health….” Because of the short time available for the completion of this report and the fact that some information from the Department of Energy (DOE) was not available, it was not possible to analyze cost, worker safety, or long-term human and environmental health consequences of alternatives to the current waste management plan. However, the committee did evaluate some of DOE’s performance assessments, which are meant to demonstrate compliance with the performance objectives and evaluate the long-term consequences of DOE’s plans.1

A performance assessment is a quantitative evaluation of the anticipated behavior of a disposal facility that projects the extent of contaminant migration from the facility and the potential impacts of releases on human health and the environment. Such a systematic examination of the engineered and natural environment can also give analysts a qualitative sense of the likelihood of different outcomes by allowing them to examine reasonable and bounding scenarios. DOE uses performance assessments to establish waste concentration limits, waste form requirements, and facility design requirements that are needed to protect long-term public health and safety and the environment (Mann et al., 2001). DOE also uses a facility’s performance assessment as its most prominent tool to demonstrate compliance with performance objectives, i.e., that dose limits2 will not be exceeded. The performance assessment that is used to seek approval for disposal plans may be somewhat different from other performance assessments for the same facility (see Sidebar VI-1), but performance assessment does not end with approval (or disapproval) of disposal plans.

Performance assessment and monitoring are connected before and after closure. Performance assessments use data from monitoring during operations and cleanup as input values for model parameters and to identify anomalies that may require revision of conceptual models. Monitoring programs use performance assessments to assist the selection and location of data acquisition systems. In considering this relationship, the committee benefited from the proceedings of a recent workshop on performance monitoring sponsored by DOE (DOE, 2005). Because there are ongoing requirements for performance assessment, the committee views a performance assessment as an evolutionary document that must be updated and changed as new information becomes available and as changes occur at the site (changes in both the physical situation and the site activities—e.g., operations, closure, post-closure monitoring and surveillance).

Performance assessments (see Sidebar VI-2) are based on conceptual and numerical models that simplify reality to predict contaminant releases, migration, and consequent exposures of people and biota under a set of assumptions and scenarios. As noted in an earlier National Research Council report, “[p]roperly done, risk assessment is a powerful tool for systematically organizing the information and understanding the behavior and impacts of radioactive waste at a particular location” (NRC, 2005b). The committee regards performance assessment as a valuable tool for examining the

1

The committee did not undertake a detailed review of every input parameter, feature of the model, or method used. The U.S. Nuclear Regulatory Commission is conducting just such a detailed review. The committee focused on the methods, key assumptions, and results.

2

DOE calls the numerical limits for inadvertent intruders “performance measures,” but for simplicity the committee refers to them as dose limits.



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 72
Tank Waste Retrieval, Processing, and On-Site Disposal at Three Department of Energy Sites: Final Report VI Performance Assessment The statement of task charges the committee with evaluating: “… (2) any actions additional to those contained in current plans that the Department should consider to ensure that its plans to manage its radioactive waste streams will comply with the performance objectives of Part 61 of Title 10, Code of Federal Regulations; … (4) existing technology alternatives to the current management plan for the waste streams mentioned above and, for each such alternative, an assessment of the cost, consequences for worker safety, and long-term consequences for environmental and human health….” Because of the short time available for the completion of this report and the fact that some information from the Department of Energy (DOE) was not available, it was not possible to analyze cost, worker safety, or long-term human and environmental health consequences of alternatives to the current waste management plan. However, the committee did evaluate some of DOE’s performance assessments, which are meant to demonstrate compliance with the performance objectives and evaluate the long-term consequences of DOE’s plans.1 A performance assessment is a quantitative evaluation of the anticipated behavior of a disposal facility that projects the extent of contaminant migration from the facility and the potential impacts of releases on human health and the environment. Such a systematic examination of the engineered and natural environment can also give analysts a qualitative sense of the likelihood of different outcomes by allowing them to examine reasonable and bounding scenarios. DOE uses performance assessments to establish waste concentration limits, waste form requirements, and facility design requirements that are needed to protect long-term public health and safety and the environment (Mann et al., 2001). DOE also uses a facility’s performance assessment as its most prominent tool to demonstrate compliance with performance objectives, i.e., that dose limits2 will not be exceeded. The performance assessment that is used to seek approval for disposal plans may be somewhat different from other performance assessments for the same facility (see Sidebar VI-1), but performance assessment does not end with approval (or disapproval) of disposal plans. Performance assessment and monitoring are connected before and after closure. Performance assessments use data from monitoring during operations and cleanup as input values for model parameters and to identify anomalies that may require revision of conceptual models. Monitoring programs use performance assessments to assist the selection and location of data acquisition systems. In considering this relationship, the committee benefited from the proceedings of a recent workshop on performance monitoring sponsored by DOE (DOE, 2005). Because there are ongoing requirements for performance assessment, the committee views a performance assessment as an evolutionary document that must be updated and changed as new information becomes available and as changes occur at the site (changes in both the physical situation and the site activities—e.g., operations, closure, post-closure monitoring and surveillance). Performance assessments (see Sidebar VI-2) are based on conceptual and numerical models that simplify reality to predict contaminant releases, migration, and consequent exposures of people and biota under a set of assumptions and scenarios. As noted in an earlier National Research Council report, “[p]roperly done, risk assessment is a powerful tool for systematically organizing the information and understanding the behavior and impacts of radioactive waste at a particular location” (NRC, 2005b). The committee regards performance assessment as a valuable tool for examining the 1 The committee did not undertake a detailed review of every input parameter, feature of the model, or method used. The U.S. Nuclear Regulatory Commission is conducting just such a detailed review. The committee focused on the methods, key assumptions, and results. 2 DOE calls the numerical limits for inadvertent intruders “performance measures,” but for simplicity the committee refers to them as dose limits.

OCR for page 72
Tank Waste Retrieval, Processing, and On-Site Disposal at Three Department of Energy Sites: Final Report SIDEBAR VI-1 Conservative or Realistic Assessment Where this report mentions conservative or realistic assessments, the committee uses these terms in the context described below. A realistic performance assessment implies that the conceptual models on which the calculations are founded, and the values assigned to the model parameters, are structured to provide a “best estimate” of system performance, given the available data and current conceptual understanding. It is most appropriate to adopt this approach when using a performance assessment for planning or design approaches, such as improving facility design or establishing waste form requirements, or when confirming the performance of a disposal site. In demonstrating compliance in a regulatory context, it is often appropriate to adopt approaches that produce conservative estimates that overstate the magnitude of the impact. In this way, uncertainties in the description of certain physical or chemical processes, and in the estimates of model parameters, can be given less weight in evaluating the results of the performance assessment. SIDEBAR VI-2 The Performance Assessment Process During the last three decades, a generally agreed process has developed for carrying out the technical steps in a performance assessment (see Appendix I for a more thorough description of the performance assessment process), although different methods and models are used at each site and even for different facilities within a site. In many cases, the nontechnical steps, from identification of the consequences of concern to the approach used in characterizing risks, are just as important as the technical steps and indeed are inextricable parts of these steps (see, e.g., NRC, 1994, 2005b; PCCRARM 1997a, 1997b). The methodology for carrying out the nontechnical steps is not as standardized, and these factors are discussed in Chapter VIII. Even within the technical realm, there are debates about the best way to implement some parts of the technical methodology. However, the requisite components of the methodology are not really at issue. For example, some practitioners prefer to examine uncertainty and variability using Monte Carlo methods, which run computer models many times using probability- and frequency-based distributions of values for input parameters to yield a probability distribution of results. Other practitioners, including analysts that conducted all of DOE’s tank waste performance assessments, use deterministic methods and carry out sensitivity studies (with no embedded probability weighting) to illustrate the effects of uncertainty and variability. All, however, agree that uncertainty and variability must be accounted for and presented in an assessment. A previous National Research Council report, Risk and Decisions about Disposition of Transuranic and High-Level Radioactive Waste (NRC, 2005b), states that “The key feature of the risk analysis process described in this [report] is that the data, modeling, and any other calculations in estimating risk must be structured to inform a specific and well-defined decision,” and “analytical detail and complexity should be limited to the minimum necessary to distinguish the best option or options.” In this approach, complexity is added only as it provides needed greater fidelity to the behavior of the real system being modeled. waste form, the disposal facility, the disposal environment, and the likely future interactions among the waste, people, and surrounding environment. A good performance assessment can help identify which factors are most important in ensuring safe disposal of the waste, and even what actions may increase or reduce risk. The committee does not, however, believe that a performance assessment can accurately determine the concentrations and quantities of long-lived contaminants that will leach from a disposal facility far into the future. Accepting the absolute numerical results of a performance assessment entails accepting that the assumptions and scenarios adequately represent physical reality. Because the absolute numerical results, including the numerical predictions of doses from radionuclides, are approximations and subject to unquantifiable uncertainties, the committee views performance assessments as only a part of the demonstration of compliance. Performance assessment is a tool to support decision making; it is not a definitive statement of future conditions at a site. DOE’S PERFORMANCE ASSESSMENTS The committee has examined the performance assessments that were available before 2006 for tank heels (at all three sites) and the separated low-activity waste from the tanks (at the Savannah River Site and Hanford). DOE has

OCR for page 72
Tank Waste Retrieval, Processing, and On-Site Disposal at Three Department of Energy Sites: Final Report documented its code selection process (see, e.g., Mann et al., 1999), and DOE’s performance assessments use models that are recognized and accepted by the performance assessment community and specialists who model contaminant transport through different media. The committee has made no judgment whether DOE’s current tank closure plans now meet the performance objectives, which is a regulatory decision. Instead, the committee focused on technical aspects of several of DOE’s performance assessments. Although the committee did not have sufficient time or resources to confirm the results of the models used, it did examine the methods, key assumptions, and results to evaluate what confidence should be placed in the performance assessments. In preparing its interim report, which focused on the Savannah River Site, the committee identified several concerns about the modeling approach DOE was using to demonstrate compliance with the performance objectives in 10 CFR 61, some of which are discussed below. At each committee meeting at the three sites, the committee asked DOE to provide presentations making the case why its proposed approach for managing and disposing of tank wastes is safe and acceptable. Concurrent with the committee’s study, the U.S. Nuclear Regulatory Commission (USNRC) provided DOE with a long list of both general and detailed requests for additional information to support its review of DOE’s draft salt waste determination (DOE-SRS, 2005b) at the Savannah River Site. DOE and the USNRC also held supplemental discussions in public meetings to clarify technical points and reach a mutual understanding of their respective modeling philosophies for performance assessment. From these interactions, DOE clearly learned what information and what kinds of transparency independent analysts need for them to evaluate DOE’s proposed actions. The clarity, quality, and completeness of DOE’s analyses, presentation of the information, and reasoning have improved dramatically since the committee’s interim report was finalized. DOE has developed new documents called “performance objectives demonstration documents” (PODDs) that describe the objectives, conceptual models, assumptions, and reasoning that supports decisions far better than the committee had seen in prior DOE performance assessments. The PODDs also include sensitivity studies, which allow reviewers to evaluate the consequences of modeling assumptions and input parameter values. The PODDs address committee concerns, such as how the projected performance changes with grout durability, which is examined through a set of scenarios that constitute deterministic sensitivity studies. Some other concerns, such as questions about how long the pH and chemically reducing properties of grout need to persist in order to give acceptable performance, have not been addressed fully—the sensitivity studies provided in the September 30, 2005 Savannah River Site draft waste determination for closure of two tanks and its associated PODD (Buice et al., 2005; DOE-SRS, 2005a) are not entirely sufficient for this purpose. The committee’s concerns about the point of compliance (the location at which compliance with performance objectives must be demonstrated) that DOE has selected and DOE’s assumptions in intruder scenarios remain. An additional concern about the combined inventory estimated to remain in the tanks at Savannah River Site has arisen since the committee issued its interim report. These concerns are described in the discussion of the Savannah River Site results, later in this chapter. DOE’s plans to grout and close the tanks have raised a number of concerns among environmental groups (Makhijani et al., 1986; NRDC, 2003, 2004a, 2004b, 2005; Makhijani, 2004; Makhijani and Boyd, 2004; Perks, 2004; Smith, 2004). The main concern is that, given that the tank heels and grout do not mix, grout is not viewed as a form of waste immobilization, but rather as a layer on top of the tank heels. Even incorporating waste into grout will not prevent leaching of contaminants into groundwater. The only means to prevent leaching over the long term is to keep water out. Environmental public interest groups warned that the waste would eventually leach into the groundwater near the Columbia River in Washington, the Snake River Aquifer in Idaho, and the water table near the Savannah River (e.g., NRDC, 2003). Smith provides detailed concerns about the long-term performance of grout related to uncertainties in the evolution of leaching properties, degradation mechanisms, hydraulic properties, chemical properties, patterns of waste and grout distribution in the tanks, and temperature effects on grout and waste with time (Smith, 2004). Makhijani and others raise a number of further concerns related assumptions about long-term stewardship, long-term predictions of environmental changes, and human behavior at the Savannah River Site (Makhijani et al., 1986; Makhijani, 2004; Makhijani and Boyd, 2004). In its 2004 performance assessment for the Saltstone Vaults and in earlier assessments for closed tanks at the Savannah River Site (DOE-SRS 1997a,1997b), DOE assumed that the grout would maintain its physical integrity as a hydraulic barrier for 500 years. The committee was particularly concerned that: (a) there was scant scientific support for the 500-year assumption, and (b) DOE’s approach of treating modeling elements, such as all of the grout in a tank, as uniform or homogeneous ignores phenomena that are dominated by heterogeneities, such as fracture flow. In response to the USNRC’s request for additional information and supplemental discussions, DOE carried out sensitivity studies that included an “early failure” scenario in which the grout maintains a low hydraulic conductivity for only 100 years (with chemical reducing conditions remaining the same as in the original analysis). This scenario indicated little change in the quantities of radionuclides released (an 8 percent change in the projected dose) and, therefore, showed that assumptions about the physical integrity of the grout are not critical to the results of the performance assessment. In all subsequent calculations for tank closures, DOE did not

OCR for page 72
Tank Waste Retrieval, Processing, and On-Site Disposal at Three Department of Energy Sites: Final Report take credit for the concrete tank vault and carbon steel tank, and it is assumed that there are cracks in the basemat starting at 500 years after closure (Buice et al., 2005). (See Chapter V for a detailed discussion of tank grouting.) DOE made similar assumptions about the pH and chemically reducing capability of the grout in the tanks; i.e., that reducing conditions would be maintained for 10,000 years. DOE’s response to the USNRC’s request noted above was to carry out calculations in which concentric cracks form in the grout in the tank, thus creating preferred flow paths and locally diminishing the reducing conditions (Buice et al., 2005). DOE’s calculations indicate that most of the grout maintains its reducing capacity. However, the bulk of the grout is not particularly relevant, given that the waste does not substantially mix with the grout. Grout that comes in contact with water flowing through fractures may not maintain its reducing capacity indefinitely; hence leached radionuclides may not remain in a less mobile, chemically reduced valence state. Lukens and others (2005) have shown that the technetium species in grout are rapidly oxidized by oxygen if exposed to it, such as they could be along the cracks. According to DOE (Buice et al., 2005), Additional calculations were conducted to determine the rate at which the bottom seven inches (lowest modeling node) of reductant was consumed: with no cracks, 96% of the grout remained reduced after 10,000 years, whereas with three cracks, 83% of the grout remained reduced after 10,000 years. It is not clear whether such modeling can predict the grout’s reducing capabilities over thousands of years, or whether the impact of the quicker decline in grout reducing capability is important to the performance assessment results. While these assumptions seem reasonable when described by Savannah River Site scientists, there is no experience or other means to directly substantiate that such long-term assumptions are valid. In the construction industry the chemical reducing capabilities of a material are irrelevant, and “long term” means 75 to 100 years (with an active maintenance program). Long-term performance of tank fill material is site dependent due to soil conditions, climate, and near-field chemistry (i.e., interactions among the chemicals in the tank contaminants and effects of any residue from previous treatments). For example, because the sludge does not mix completely with the grout but remains interlayered with or at least partially encapsulated by it, the pH and reducing capability of water migrating through the grout are important factors in its ability to minimize the mobility of radionuclides and toxic heavy metals from the sludge into the groundwater. The first Savannah River Site tanks selected for closure do not contain some important complicating factors that affect tank cleaning, such as extensive vertical cooling coils (although they do contain zeolite). However, most of the tanks to be closed in the future contain cooling coils, which may become pathways for water infiltration to the residual waste if the grout inside or surrounding the coils shrinks significantly. Moreover, Tanks 9 through 12 are mostly submerged in the water table at all times. For these tanks, there is an additional possibility of water ingress from the sides or the bottoms.3 In these cases the radioactive waste residuals may not have the full protection of the layers of grout to reduce influx, and groundwater inflows may not be buffered by the overlying grout to the same degree (see Chapter V, Finding and Recommendation V-3). Performance Objectives and Exposure Scenarios As noted in Chapter I, Section 3116 of the 2005 NDAA requires removal of “highly radioactive radionuclides to the maximum extent practical” and compliance with the performance objectives in 10 CFR 61 for wastes disposed on-site. The performance objectives (and their supporting guidance documents) lay out a set of numerical dose limits for people who could be exposed (workers, members of the public, and inadvertent intruders); a requirement that releases to the environment and doses during operations be made as low as reasonably achievable (ALARA); and requirements concerning site stability (see Appendix C). Thus, Congress laid out an approach for waste management and disposal that requires minimization of the inventory of radioactive material that is to be disposed at the sites (within the limits of what is practical), meeting or surpassing a set of numerical dose criteria, and ensuring that contamination of the environment is as low as reasonably achievable. In addition to the assumed scenarios for release and transport of contaminants from the disposal location through the environment, DOE’s performance assessments rely on assumed exposure scenarios. For a member of the general population, the scenario typically involves exposures from drinking water, consuming fish and vegetables, bathing, and other activities that could bring people in contact with contaminants carried through groundwater. Although air and soil pathways must be considered, they typically are not as important for underground disposal facilities such as those considered here. The point of compliance is a critical element of a performance assessment used in support of a compliance decision. A point of compliance is a location some distance away from the disposal facility boundary. Inside the point of compliance is the facility operations area, which is subject to institutional controls for at least 100 years after closure. After site closure, anyone inside the boundary marked by these points of compliance (i.e., the controlled area) is considered an intruder. DOE must ensure that people in the controlled area (even after the period of institutional control) would not 3 DOE has not yet established closure plans for these tanks, so the committee has not commented on them in detail.

OCR for page 72
Tank Waste Retrieval, Processing, and On-Site Disposal at Three Department of Energy Sites: Final Report receive environmental exposures that exceed the limits in DOE and USNRC guidance concerning inadvertent intruders. Beyond the point of compliance, DOE must ensure that members of the public would not receive environmental exposures that exceed the dose limits in DOE Order 435.1 and 10 CFR 61. The point of compliance most commonly is set 100 m (109 yards) from the disposal facility boundary based on DOE Order 435.1 (DOE, 2001a) and USNRC guidance (USNRC, 1997b). Other locations based on plans for future land use and physical and institutional controls that may be in place are allowed under both commercial low-level waste disposal regulations and DOE Order 435.1 to prevent decision makers from missing important risks.4 In all cases, institutional controls are assumed to prevent inadvertent intrusion for 100 years after closure of the disposal facility. In principle, the time of compliance, too, can restrict consideration of risks in a way that neglects potentially important aspects of the disposal decisions, but the committee found no cases where DOE’s analyses cut off important results because of the time of compliance. The time of compliance required for low-level waste facilities under DOE Order 435.1 is 1,000 years, but many of DOE’s analyses extend up to 10,000 years to account for dose peaks that arrive after the mandated time of compliance. Various intruder scenarios are considered in the DOE assessments: a construction scenario, a recreation scenario, and an agriculture scenario. In some cases, it is assumed that an intruder drills directly through a waste vault, tank, or transfer line. Some cases assume that drinking water and irrigation water are drawn from an aquifer that underlies the disposal facility. Finally, for low-level waste disposal facilities, DOE Order 435.1 states (DOE, 2001a): [t]he performance assessment and composite analysis shall be maintained to evaluate changes that could affect the performance, design, and operating bases for the facility … maintenance shall include the conduct of research, field studies, and monitoring needed to address uncertainties or gaps in existing data…. Additional iterations of the performance assessment and composite analysis shall be conducted as necessary during the post-closure period. The manual explicitly notes that review and revision are required when changes “alter the conclusions or the conceptual model(s) of the existing performance assessment or composite analysis.” It also requires an annual “determination of the continued adequacy of the performance assessment and composite analysis [based on] the results of data This is perfectly in accord with the committee’s view of a performance assessment as a living document. It is not clear in practice, however, that DOE’s performance assessments and composite analyses are treated as living documents after disposal authorization is received. For example, the Savannah River Site’s composite analysis for the E-Area Vaults Saltstone Disposal Facilities (WSRC, 1997) was last revised in 1999 (Cook et al., 1999) and does not reflect DOE’s current understanding of inventories and expected performance of disposal facilities at the site. This means that the composite analysis has not been up to date as DOE, the State of South Carolina, and the USNRC have evaluated DOE’s plans for disposal of tank waste on the site. DOE has in place a program for maintaining the adequacy of the performance assessments and composite analysis (WSRC, 2000). DOE informed the committee that the composite analysis is scheduled for revision in the near future. The need for current assessments of all the contaminants and facilities that contribute to risk at the site is discussed in the context of decision making in Chapter VIII. The Savannah River Site Performance Assessment Results DOE assumes that the Savannah River Site will remain under federal government ownership within its current boundaries in perpetuity. (See executive summary of DOE-SRS, 2005f for assumptions about future land use; see Appendix J of this report for a map of the site and the General Separations Area.) DOE further assumes that the site will continue to remain zoned for industrial, industrial support, and general uses. DOE intends to restrict the area around the facilities where tank waste will remain on-site from residential use for 10,000 years. In its performance assessment, DOE assumes that active institutional controls to prevent inadvertent human intrusion are effective for 100 years. In its performance assessments, DOE has considered the possibility that after 100 years, an inadvertent intruder could construct a residence near one of the disposal facilities (DOE-SRS, 2005b; Ross, 2005). DOE uses different points of compliance for the different tank waste disposal facilities. Table VI-1 summarizes the points of compliance that DOE uses for members of the public and inadvertent intruders at each of the facilities considered at the Savannah River Site. For the closed tanks, the general population is defined as people residing outside the General Separations Area (GSA), which includes both the F and H Tank Farms, and the Z Area Saltstone Disposal Facility, where tank wastes will remain, in addition to the Defense Waste Processing Facility, the Old Burial Ground, the E Area low-level waste disposal facilities, and the canyons themselves (see Appendix J). DOE states in its performance assessment demonstration document for the tank closures, “[a] key assumption to the modeling analysis is 4 “The point of compliance shall correspond to the point of highest projected dose or concentration beyond a 100 meter buffer zone surrounding the disposed waste. A larger or smaller buffer zone may be used if adequate justification is provided” (DOE, 2001a). collection and analysis from research, field studies, and monitoring.”

OCR for page 72
Tank Waste Retrieval, Processing, and On-Site Disposal at Three Department of Energy Sites: Final Report TABLE VI-1 Points of Compliance for Tank Wastes Planned for On-site Disposal at the Savannah River Site   F Tank Farm H Tank Farm Saltstone Vaults Members of the public Resident who constructs and lives in a dwelling at the point of compliance “Just downstream” and on the opposite bank from a seepline where groundwater outcrops to a stream. The seepline is located 1.8 km downgradient of the tank farm. Exposure is via surface water.a “Just downstream” and on the opposite bank of a seepline where groundwater outcrops to a stream. The seepline is located 1.2 km downgradient of the tank farm. Exposure is via surface water. a 100 m from the boundaries of the facility. Note that only doses arising from air and groundwater pathways have currently been considered at the point of compliance. Inadvertent intruders Resident who constructs and lives in a dwelling and drills a well inside facility boundary For soils,b inside the facility boundary. For groundwater, assumes a well drilled inside the facility boundary through a transfer line into the third aquifer below the surface. For soils,b inside the facility boundary. For groundwater, assumes a well drilled inside the facility boundary through a transfer line into the third aquifer below the surface. For soils, at the facility boundary. For groundwater, assumes a well drilled into the top aquifer (water table) at the facility boundary. a Exposure pathways for member of public: (1) incidental ingestion of soil from shoreline deposits (during recreational activities), (2) direct radiation from seepline, (3) air inhalation (I-129 volatilization) at seepline, (4) dermal contact with Four Mile Branch, (5) drinking water from Four Mile Branch, (6) ingestion of fish from Four Mile Branch, (7) direct radiation from Four Mile Branch, (8) ingestion of milk from cows fed vegetation grown on soil irrigated with Four Mile Branch water, (9) ingestion of meat from cows fed vegetation grown on soil irrigated with Four Mile Branch water, and (10) ingestion of produce irrigated with Four Mile Branch water. Only 5, 6, and 10 are significant contributors to the peak dose. b Contaminated soils considered in the intruder scenario. that no unrestricted use of the land or groundwater for the GSA will be permitted as presented in the SRS End State Vision [6] and Savannah River Site (SRS) Long Range Comprehensive Plan….” (Buice et al., 2005). The committee notes that the GSA is a rather large area (approximately 20 km2 [about 8 square miles], judging from maps) and the points of compliance for the tank farms are rather distant from the boundaries of the facilities (1.8 km [1.1 mile]). In contrast, the point of compliance for the Saltstone Disposal Facility is set 100 m from the disposal facility boundary. Concentrations of contaminants in the aquifers at the Savannah River Site (and therefore doses from the consumption of groundwater) decrease with distance from a leaking facility because of mixing, dilution, sorption, and decay of contaminants as they migrate away from the facility. As a result, the estimated radiation dose received by a member of the public at a point of compliance that is far from the disposal facility will be lower, and thus more likely to meet performance objectives. Other factors being equal, a larger area with contaminated groundwater has a greater likelihood than a smaller area does of an inadvertent intruder drilling a well and drawing contaminated water after active institutional controls cease to be enforced. The selection of the point of compliance has both policy and technical dimensions. Good policy, however, requires that there be justification, including at least technical coherence, for the selection. DOE justifies its point of compliance for the tanks by turning to the land-use plans for the site, which envision government control and only industrial uses of the General Separations Area, in perpetuity. The South Carolina Department of Health and Environmental Control approved these points of compliance in closing Tanks 17 and 20, but recently raised questions about whether they are acceptable for future tank closures (SCDHEC, 2005a). Nonetheless, the points selected at the Savannah River Site tank farms afford the lower, intruder level of protection (i.e., higher dose limit) over a large area (approximately 20 square km2), and even the intruder dose might be much higher than expected if some assumptions in the intruder scenarios, which are discussed below, do not prove to be correct (see Chapter VIII). The agricultural intruder scenario for the Tanks 18 and 19 performance assessment assumes that an inadvertent intruder would draw household water from the Congaree aquifer, the deepest of three aquifers below the tank farms where estimated potential contaminant levels are lowest, rather than from the upper (water-table) or Barnwell-McBean aquifers where estimated potential contaminant levels are higher. (See Figure J-3 in Appendix J for a diagram showing these hydrologic units.) DOE supports this assumption by reasoning that an intruder would not choose to draw well water from the water-table or the Barnwell-McBean aquifers, which have a yield of 11 to 19 liters (3 to 5 gallons) per minute.5 Instead, an intruder would know that a higher-yield aquifer lies below the Barnwell-McBean aquifer and would drill into that. DOE also assumes that the intruder constructs the household water well in a manner ensuring that cross-contamination of the aquifers does not occur during either 5 Any well in the United States that yields less than six gallons per minute is generally considered a low-yield well. In performance assessments at other sites, DOE has considered exposure scenarios based on its interpretation of land uses by local Native American peoples. Alternative uses that could entail, for example, use of lower yield wells were not considered at the Savannah River Site.

OCR for page 72
Tank Waste Retrieval, Processing, and On-Site Disposal at Three Department of Energy Sites: Final Report TABLE VI-2 Dose Predictions at Different Locations from the Savannah River Site F Tank Farm   Intruder   General Population Source Drinking Water Dosea from Well in Barnwell-McBean Aquifer 100 m from Tank Farm Boundary Drinking Water Dosea from Well at Barnwell-McBean Aquifer Seepline All Pathways Whole-Body Doseb Tank 18       Technetium-99 2.1 mrem/yr at 525 yr 0.14 mrem/yr at 665 yr   Neptunium-237 23 mrem/yr at 4935 yr 1.3 mrem/yr at 6755 yr 0.04 mrem/yr at 6405 yr Tank 19       Technetium-99 5.8 mrem/yr at 595 yr 0.29 mrem/yr at 735 yr 0.009 mrem/yr at 735 yr Neptunium-237 3.0 mrem/yr at 5355 yr 0.13 mrem/yr at 7245 yr   F Tank Farm       Technetium-99 57 mrem/yr at 525 yr 2.5 mrem/yr at 595 yr 0.07 mrem/yr Neptunium-237 30 mrem/yr at 4795 yr 0.13 mrem/yr at 7245 yr   Iodine-129 3.2 mrem/yr at 385 yr 0.26 mrem/yr at 455 yr 0.01 mrem/yr NOTE: 1 mrem = 0.01 mSv. a 50-year committed effective dose equivalent. b DOE describes this dose as the 50-year committed dose equivalent to the whole body. SOURCE: Buice et al., 2005. well drilling or well use. Thus, the intruder’s actions do not result in migration of contamination from the upper aquifer to the lower ones and do not increase the radiation dose received by the intruder as a result of aquifer cross-contamination. While one might hope that an inadvertent resident would both construct his or her well properly and draw from the more productive, cleaner aquifer, it seems only prudent to examine what the consequences would be if these assumptions were incorrect. The contaminant concentrations that DOE predicts in the water-table aquifer are higher than those in the Congaree aquifer by factors ranging from about 60 to about 250, while those in the intermediate Barnwell-McBean aquifer are higher than those in the Congaree aquifer by factors ranging from about 250 to about 1,000. The more conservative approach of considering exposures from the more contaminated aquifers is more consistent with standard practice. Table VI-2 presents projected doses for people drawing drinking water from the Barnwell-McBean aquifer at two locations within the boundary defined by the chosen point of compliance, alongside the results for a member of the public residing across the Four Mile Branch. This scenario also assumes that drinking water is the only source of radiation exposure for the intruder. Although other scenarios take into account the possibility that radionuclides could be concentrated in animals and plants consumed by the intruder, this one does not (see Findings and Recommendation VI-2). All of the doses are within the limits for intruders described in the guidance for implementing the performance objectives, although the higher concentrations of neptunium-237 and technetium-99 in the Barnwell-McBean Aquifer at 100 meters bear watching to ensure that they do not exceed the performance objectives. The committee was surprised to note that a large fraction (perhaps more than 75 percent) of the estimated neptunium-237 drinking water dose from a well in the Barnwell-McBean Aquifer 100 meters from the tank farm boundary is attributed to neptunium-237 migration from just one tank: Tank 18. The PODD for Tanks 18 and 19 (Buice et al., 2005) presents the estimated residual radionuclide inventory in tanks that have already been cleaned out (Tanks 17-20) and the projected residual inventory for F Tank Farm tanks that have not been cleaned out (Tanks 1-8, 25-28, 33-34, and 44-47). These inventories were used to calculate the concentrations and doses from the whole F Tank Farm, as presented in Table VI-2. The estimated remaining quantity of neptunium-237 in Tank 18 is 0.118 Ci (4.37 x 109 Bq) and the combined total estimate of neptunium-237 in all of the other tanks in the F Tank Farm is 0.0535 Ci (1.98 x 109 Bq) or approximately 45 percent of the inventory in Tank 18. DOE explains the imbalance in the neptunium-237 inventory based on three factors (the following numbered points are quoted from Ross, 2006a): The F-Area laboratory was used to analyze H Canyon Np-237 processing samples. The waste from the F-Area Laboratory was discarded to Tank 18 (and carried over to Tank 19 through evaporator processing) and, therefore, these tanks would have higher than average concentrations of Np-237.6 6 DOE notes that the addition of the laboratory waste was not tracked in the waste characterization system (WCS) tables (DOE-SRS, 1999).

OCR for page 72
Tank Waste Retrieval, Processing, and On-Site Disposal at Three Department of Energy Sites: Final Report TABLE VI-3 Assumed Volume of Residual Waste Remaining in Closed HLW Tanks at the Savannah River Site.a Tank Number Area Tank Type Residual Material Volume (gallons) Tank Number Area Tank Type Residual Material Volume (gallons) 1 F I 100 27 F III 1,000 2 F I 100 28 F III 1,000 3 F I 100 29 H III 100 4 F I 100 30 H III 100 5 F I 100 31 H III 100 6 F I 100 32 H III 100 7 F I 100 33 F III 100 8 F I 100 34 F III 100 9 H I 100 35 H III 100 10 H I 100 36 H III 100 11 H I 100 37 H III 100 12 H I 100 38 H III 100 13 H II 100 39 H III 100 14 H II 100 40 H III 100 15 H II 100 41 H III 100 16 H II 100 42 H III 100 17b F IV 2,200 43 H III 100 18 F IV 1,000 44 F III 1,000 19 F IV 1,000 45 F III 1,000 20b   FIV 1,000 46 F III 1,000 21 H IV 100 47 F III 1,000 22 H IV 100 48 H III 100 23 H IV 1,000 49 H III 100 24 H IV 100 50 H III 1,000 25 F III 1,000 51 H III 100 26 F III 1,000         a These volumes are an assumption for modeling purposes only and do not represent a commitment or goal for waste removal. b Tank has been closed. SOURCE: DOE-SRS, 2002. The mass of material remaining in Tank 18 is 15,335 kg (16,357 kg using the upper 95% confidence density). For the purposes of estimating the future F-Tank Farm residual material radionuclide inventories, it was assumed that future waste removal activities would be successful in reducing residual material inventories to the [environmental impact statement (EIS)] predicted final inventories of 88.4 kg for high heat tanks and 884 kg for low heat tanks. As future tanks are closed, the mass of the residual materials will be adjusted as necessary for those tanks and the predicted radionuclide inventories at closure will be replaced with inventories based on analysis of residual materials. Several conservative assumptions were made that also contributed to the elevated Np-237 inventory in Tank 18 (and Tank 19) … : (a) The upper 95% confidence sample results were used instead of the average. This increased the Np-237 inventory in Tank 18 by almost 16%. (b) The upper 95% confidence sample density was used instead of the average density. This increased the Np-237 inventory by another 7%. (c) The residual liquid Np-237 inventory was calculated based on a less-than-detection sample result for Np-237. (d) An additional 0.033 Ci [1.2×109 Bq] of Np-237 was calculated to be in tank wall corrosion products. This increased the Np-237 inventory by another 40%. The committee examined some of the details of point two above, the prediction of final tank inventories quoted from the environmental impact statement. Table VI-3 shows the data from the Savannah River Site High Level Waste Tank Closure Environmental Impact Statement (DOE-SRS, 2002, Table C.3.1-2 in Appendix C). The committee observes that these predictions are characterized in the tank closure environmental impact statement as assumptions, rather than commitments or even goals. Further, the assumed volumes of residual material in Tanks 17 and 20 (2,200 gallons [8.3 m3] for Tank 17 and 1,000 gallons [3.8 m3] for Tank 20), which are the only tanks that had been closed at the time the tank closure environmental impact statement was issued (2002), represent the upper bound of the assumed residuals among all of the tanks. Indeed, DOE assumed that Tanks 18 and 19 would each have 1,000 gallons

OCR for page 72
Tank Waste Retrieval, Processing, and On-Site Disposal at Three Department of Energy Sites: Final Report residual. The current estimates in the performance objective demonstration document (Buice et al., 2005) are 4,300 gallons (16.3 m3) in Tank 18 and 15,100 gallons (57.2 m3) in Tank 19. There is zeolite in Tanks 18 and 19, which proved difficult to remove, so DOE proposes to leave it in place, but the performance objective demonstration document reports that Tanks 7, 25, and 27, all of which are in the F Tank Farm, also contain zeolite. Of the tanks that have not been cleaned out, some of them have large quantities of sludge; several of them do not. Access within Tanks 18 and 19 should be better than others because neither of the tanks has cooling coils in contrast to all of the tanks in the F Tank Farm yet to be cleaned, which do. For these reasons, Tanks 18 and 19 should be easier to clean than some of the other future tanks. The committee views the assumptions about residual inventories as both optimistic and unsupported. It would be more prudent for DOE to base its assumptions on what is known about the actual tanks and their contents, and on what is known about the effectiveness of the tank waste removal technologies that DOE plans to employ in those tanks (see Chapter III). With this more informed estimate, DOE and others could then examine sensitivities to see what estimated consequences arise if assumptions in DOE’s scenarios and models turn out to be inaccurate. Without such an approach, the committee does not have confidence in the results of the performance assessment for the tank farm (see Finding and Recommendation 3). The Idaho National Laboratory Performance Assessment Results In 2003, the Idaho National Laboratory released a performance assessment for the INTEC tank farm facility (TFF; DOE-ID, 2003b). The performance assessment is a comprehensive document that contains essentially all of the elements that the committee has identified as forming the blueprint for an acceptable performance assessment (see the flowchart in Appendix I). Widely accepted simulation codes were used to predict the release of radionuclides from the vaults and tanks, and to model flow and transport both in the vadose zone above the water table and in the saturated zone. Infiltration through the vadose zone was modeled using equivalent porous medium concepts, with estimates of hydraulic conductivity of the basalt layers linked to geologic descriptions of the individual flows. The hydrogeology model embedded within the performance assessment was calibrated in an attempt to match site conditions as they were best understood, data uncertainties were discussed, and a wide-ranging sensitivity study was presented. This performance assessment led the Idaho National Laboratory to conclude that closure plans for the tank farm could proceed safely and would be protective of the environment. The sensitivity study indicated that the potential for the doses to exceed the performance objectives was low and that a combination of worst-case assumptions would have to be realized to result in doses that exceeded the performance objectives for members of the public. The performance assessment was reviewed by the USNRC, acting in an advisory role (USNRC, 2003), unlike its current role under Section 3116 of the 2005 NDAA. In addition, the USNRC developed its own performance assessment models for use in this review. The USNRC came to the conclusion that the Idaho National Laboratory had developed a reasonable source term estimate and had adequately modeled engineering system degradation, release, hydrology, and transport. Several recommendations were provided to increase confidence in the model predictions, such as including an effort to better estimate the radionuclide inventory in the sand pads beneath several tanks and to evaluate the sensitivity of the model results to the possibility of oxidizing conditions in the grout that would affect solid-solution distribution coefficients (Kds). The 2003 performance assessment used a conservative estimate of the radionuclide inventory in the tank heels, because the Idaho National Laboratory had yet to develop experience with residual waste removal. Tank cleaning that was carried out subsequent to the development of the performance assessment has shown that the radioactive inventories in the tank heels could be reduced to a value significantly less than a “best-case” condition that was modeled in the performance assessment as part of its sensitivity study. Updates of projected doses calculated on the basis of current estimates of the waste residuals that were developed from new operational experience are given in the Draft Section 3116 Determination Idaho Nuclear Technology and Engineering Center Tank Farm Facility (DOE-ID, 2005a). These calculations yield lower doses than the conservative estimates provided in the 2003 performance assessment (see Table VI-4). Concerning the point of compliance at the Idaho tank farm facility for a member of the public, DOE states (DOE-ID, 2005a), The groundwater model analysis shows that the contamination plume center (where the highest concentrations enter the regional aquifer) would be 600 m (1,969 ft) southward in the downgradient direction from the center of the southernmost TFF tank. The contamination plume center is taken as the source of drinking water after the institutional control period of 100 years has expired. Using this location appears to be a conservative approach. A key element of the performance assessment calculations involves modeling of infiltration and radionuclide transport through the deep vadose zone beneath the Idaho National Laboratory. It is on the basis of this modeling that the maximum impacts on groundwater from radionuclide release are predicted to occur at a point on the water table some 600 meters to the south of the tank farm facility. The hydrostratigraphy beneath the tank farm facility is complex, as are the inferred flow patterns with multiple, discontinuous,

OCR for page 72
Tank Waste Retrieval, Processing, and On-Site Disposal at Three Department of Energy Sites: Final Report TABLE VI-4 The Idaho National Laboratory Tank Farm Closure Performance Assessment Resultsa Performance Objectives (dose limit) Performance Assessment Results (early estimate of inventory at closure) Performance Assessment Results (current estimate of inventory at closure)b,c Doses under worst-case Kds, elevated infiltration, and conservative inventory All-pathways dose to the public (not exceeding 25 mrem/yr) 1.86 mrem/yrd 0.46 mrem/yr 15.0 mrem/yr (at 342 years due to strontium-90) Acute drilling scenario (less than 500 mrem) 232 mrem 152 mrem   Acute construction scenario (less than 500 mrem) 0.80 mrem 0.23 mrem   Chronic post-drilling scenario (less than 100 mrem/yr) 91.1 mrem/yr 25 mrem/yr   Chronic post-construction scenario (less than 100 mrem/yr) 26.1 mrem/yr 3.15 mrem/yr   NOTES: 1 mrem = 0.01 mSv. a All doses are described by DOE as 50-year committed organ dose equivalents (ht,50). b The peak annual dose to the thyroid is approximately 6 mrem/yr compared to the 10 CFR 61.41 limit of 75 mrem/yr. c The peak annual dose to any other organ is approximately 0.15 mrem/yr compared to the 10 CFR 61.41 limit of 25 mrem/yr. d The groundwater pathway contributed 1.35 mrem/yr. perched zones forming on lower-permeability units in the subsurface. The flow system may be undergoing transient readjustments to long-term changes in the infiltration regime at the facility. Numerical simulations of flow and transport can be exceedingly challenging when conducted over scales of hundreds of meters, particularly when conducted over scales of hundreds and thousands of years, especially within the vadose zone. Further complexity is introduced in the vadose zone by the presence of open fractures in the basalt flows. Predictions are likely to be non-unique (i.e., multiple solutions satisfy the boundary conditions and other requirements) and subject to considerable uncertainty. Given the core role of this pathway element in the performance assessment, and the inherent complexity of the vadose zone, the committee sees merit in a detailed, independent evaluation of flow and radionuclide transport to the water table at the tank farm facility. The intent would not be simply to reproduce the computations in the performance assessment with a different software package, but to construct an independently derived conceptual model, numerical simulation, calibration, and prediction. This modeling task may benefit from the recent data analysis and insight gained in the model calculations associated with the ongoing soil remediation program at Idaho Nuclear Technology and Engineering Center and work at Box Canyon (see, e.g., Unger et al., 2004). This latter program has developed a groundwater flow model that extends from the ground surface to the Snake River Plain aquifer and a detailed geochemical model of the near-surface zone. A strategy of independent confirmation could provide considerable support to the performance assessment calculations that have been developed to date or raise important questions about the results. Hanford Performance Assessment Results DOE issued the Preliminary Performance Assessment for Waste Management Area C at the Hanford Site, Washington in 2003 (Mann and Connelly, 2003). DOE, the Washington Department of Ecology, and the contractor that generated the performance assessment, CH2M-Hill Hanford Group, Inc., convened a panel to review the performance assessment (Kosson et al., 2004). The panel criticized several aspects of the report, which reflected the preliminary nature of the analysis. Hanford Site personnel said that the review panel provided valuable feedback that will improve the quality of the revised performance assessment significantly. DOE is working toward a major update of the performance assessment based on the review, but the report had not been issued by the end of 2005. The committee concluded it was not worthwhile to review the preliminary tank farm performance assessment in depth because the work would be outdated immediately upon issuance of DOE’s revised performance assessment, which the committee was told is imminent. The committee commends DOE and others involved for seeking peer review. Just as the Savannah River Site has the Saltstone Disposal Facility, Hanford has a facility for disposal of immobi-

OCR for page 72
Tank Waste Retrieval, Processing, and On-Site Disposal at Three Department of Energy Sites: Final Report TABLE VI-5 Hanford Low-Activity Waste Disposal Facility (IDF) Performance Assessment Resultsa Performance Objectives (dose limit) Performance Assessment Results All-pathways dose to the public (not exceeding 25 mrem/yr) 1.8 mrem/yrb Acute drilling scenario (less than 500 mrem) 1.06 mrem Chronic post-drilling scenario (less than 100 mrem/yr) 26.8 mrem/yr Groundwater protection (less than 4 mrem from beta and gamma emitters) 0.7 mrem/yr NOTES: 1 mrem = 0.01 mSv. a All doses are 50-year committed effective dose equivalents. b Peak occurs 2,400 years after closure. lized low-activity waste from its waste tanks. DOE issued a performance assessment for the waste in 2001 (Mann et al., 2001). A revision or update was made in 2003 to account for: (1) a DOE decision not to separate technetium-99 from the low-activity waste streams, and (2) the likelihood that the waste would be disposed in an integrated disposal facility,7 and (3) the different possible waste forms for the supplemental low-activity waste treatment.8 The results of the update are presented in Table VI-5. The doses to members of the public are controlled by only three radionuclides: iodine-129, technetium-99, and neptunium-237. For the first 5,000 years after facility closure, iodine-129 contributes approximately 90 percent of the dose in a scenario in which the member of the public is considered to be a farmer (the most restrictive of three scenarios examined for members of the public). Technetium-99 contributes about 10 percent of the dose in that same time frame. Further out in time, neptunium-237 increasingly contributes to the dose (44 percent at 10,000 years after closure). Important to the plume arrival times is the set of models and assumptions used to assess release from different waste forms. Low-activity waste from the Hanford tanks is considered either Category 1 waste (low-concentrations of radionuclides but the waste is not stabilized) or Category 3 waste (higher concentrations and stabilized or encased with grout). The “early” dose peak (at 2,400 years) results mostly from transport of iodine-129 and technetium-99 from, “only a relatively few Category 1 packages … (i.e., those packages with high technetium/iodine content)” (Mann, 2003). DOE notes that those packages could be disposed of as Category 3 packages, if necessary. Peer Review of Performance Assessments As noted above, DOE’s assessments have improved as a result of the reviews carried out under the 2005 NDAA. The committee commends DOE for the improved quality of its presentation of the waste determinations and the supporting documentation (including performance assessments) and reasoning. The dramatic improvement reinforces the committee’s judgment concerning the value of independent peer review. The National Research Council has issued several reports on peer review (see, e.g., NRC, 1995, 2000b, 2002a), including reports specifically advising the DOE’s Office of Environmental Management (NRC, 1996a, 1997b, 1998). Two critical elements that are worth noting in this context are independence and the level of effort. DOE and its contractors described having used only internal reviews for many or most of their performance assessments until recently. One contractor told the committee that it was an adjustment to write for outsiders rather than for themselves. That adjustment is not just about how information is presented, but also about supporting assumptions and reasoning for those who may not already think in the same way as the authors. The USNRC’s reviews in its legislated role under Section 3116 of the 2005 NDAA provide such independence. To fulfill its potential, a peer review must be carried out at an appropriate level of effort. The scale of information (and the sheer number of documents) to be reviewed for the draft waste determinations requires a much greater effort than was provided in, for example, previous USNRC consultations. The level of effort must be matched to the task (see Finding and Recommendation VI-1). FINDINGS AND RECOMMENDATIONS Finding VI-1: Independent peer review, which is a cornerstone of good scientific and engineering practice, has helped DOE to improve the clarity, quality, and completeness of its waste determinations and performance assessments. 7 The integrated disposal facility is a single disposal facility for immobilized low-activity waste from tank cleanup and other waste from the site. 8 The different waste forms correspond to the different supplemental treatment options: glass from bulk vitrification, a granular solid from steam reforming, and grout from cast stone.

OCR for page 72
Tank Waste Retrieval, Processing, and On-Site Disposal at Three Department of Energy Sites: Final Report Recommendation VI-1: DOE should: (1) seek independent peer review of the data collection and analysis relevant to risk done to support the waste determinations and performance assessments before submitting draft waste determinations under Section 3116 of the 2005 NDAA; (2) arrange for independent reviews similar to those under Section 3116 for any waste determinations made under DOE Order 435.1; and (3) publish more of its data and analyses in peer-reviewed literature and accessible reports so that they can be reviewed by the technical community. Finding VI-2a: Performance assessments are complex structures of models and assumptions. The objective of the assessment is to provide a conservative or realistic projection of contaminant release and movement and of human exposure over time to determine compliance with performance objectives. The performance assessment provides a reasonable method for predicting the ability of a site to meet the performance objectives of 10 CFR 61, provided that: (1) the assessment includes deterministic or probabilistic modeling coupled with uncertainty analysis; (2) the site characterization data enable analysts to model likely pathways for movement of potential contaminants from the site; (3) parameter values based on data and assumptions (e.g., those concerning the longevity of institutional controls) are conservative or realistic; and (4) the assessment is updated as site conditions change and as DOE’s knowledge of site conditions and other factors improves. Finding VI-2b: Considering the range of assumptions, the limitations of predictions regarding engineered systems with a finite design life in dynamic environments, the challenges in modeling contaminant transport through the environment (especially in the vadose zone), and unpredictable future human behavior, the performance assessment’s numerical output is insufficient in itself to determine compliance with the performance objectives. Finding VI-2c: Accepting the performance assessment results entails accepting the scenarios and assumptions that underlie those results. Recommendation VI-2: DOE should continue to make clearer the basis for decision making and for assumptions in its performance assessments; it should ensure that the features of the models used are explicit and transparent, that the models are verifiable, and that assessments present an evaluation of uncertainty. One important example is that when DOE uses a nonstandard point of compliance, it should state clearly the potential exposures closer to the disposal facility in case assumptions about human behavior and institutions do not turn out to be true. Finding VI-3: The committee views the assumptions about future residual inventories in the F Tank Farm as both optimistic and unsupported. Given these assumptions, the committee does not have confidence in the results of the performance assessment for the tank farm. Recommendation VI-3: DOE should base its projections of future residual inventories on what is known about the actual tanks and their contents, and on what is known about the effectiveness of the tank waste removal technologies that DOE plans to employ in those tanks.