Return to the Main Document

EXECUTIVE SUMMARY

I. INTRODUCTION

In 1982, Congress gave the U.S. Department of Energy (DOE) the responsibility to site, construct, operate, and close a repository for the disposal of spent nuclear fuel and high-level radioactive waste. In 1987, Congress directed the DOE to investigate Yucca Mountain, Nevada, exclusively, to determine whether it would be a suitable site for development of a geologic repository. After studying Yucca Mountain for over 20 years, the Secretary of Energy is now considering whether to recommend the Yucca Mountain site to the President.

To support the Secretary of Energy's decision, the DOE, in accordance with the Nuclear Waste Policy Act of 1982, has developed a comprehensive set of technical documents that will serve as the basis of any recommendation. This Yucca Mountain Site Suitability Evaluation is one of those technical documents, and it specifically addresses the DOE site suitability guidelines set forth in 10 CFR Part 963.

The Yucca Mountain Preliminary Site Suitability Evaluation, issued in August 2001, provided the public with a preliminary evaluation of the suitability of the site for development of a repository. The evaluation was based on the DOE's proposed site suitability guidelines. This Yucca Mountain Site Suitability Evaluation provides an assessment of the site's performance against the DOE's final guidelines. The evaluation is based on information from the following reports:

These reports are supported by a comprehensive set of scientific and technical documents.

Structure of the Site Suitability Guidelines

The DOE's site suitability guidelines in 10 CFR Part 963 are divided into two sections corresponding to the repository's preclosure and postclosure periods. Within each of the preclosure and postclosure sections are three subsections: (1) the suitability determination, (2) the suitability evaluation method, and (3) the criteria to be used in the evaluation. The methods and criteria describe the scope of technical information that will support a suitability determination. The guidelines are consistent with licensing regulations issued by the U.S. Nuclear Regulatory Commission (NRC) (10 CFR Part 63) that implement the radiation protection standards established by the U.S. Environmental Protection Agency (EPA) (40 CFR Part 197).

The DOE preclosure guidelines provide a safety evaluation method that is consistent with a preclosure safety analysis that the NRC requires for licensing at 10 CFR 63.112. Like the NRC analysis process, the preclosure safety evaluation emphasizes performance requirements, analytical bases and technical justifications, and evaluations that assess the adequacy of the design and its intended safety functions.

The DOE postclosure guidelines provide a method for conducting a total system performance assessment (TSPA). This method involves DOE's obtaining field and laboratory data; accounting for uncertainties; considering alternative models; and assessing features, events, and processes (FEPs) that might affect the performance of the Yucca Mountain disposal system. Both the preclosure and postclosure guidelines reference the NRC radiation protection standards.

The Yucca Mountain Site—Geologic Setting and Underground Facility

The Yucca Mountain site is located on federal land on and adjacent to the Nevada Test Site in Nye County, Nevada, about 160 km (100 mi) northwest of Las Vegas (Figure 1). The mountain consists of a series of ridges extending 40 km (25 mi) from the southern rim of the Timber Mountain caldera in the north to the Amargosa Desert in the south. The water table is approximately 500 to 800 m (1,600 to 2,600 ft) below the surface of Yucca Mountain. The zone of soil or rock below the surface and above the water table is called the unsaturated zone. The geologic repository would be located in the unsaturated zone, about 200 to 500 m (660 to 1,600 ft) below the surface of the mountain and, on average, about 300 m (1,000 ft) above the water table.

The geologic, hydrologic, and geochemical system at Yucca Mountain is called the geologic setting. Some characteristics of the geologic setting include a deep water table, thick unsaturated zone, small flux of water flowing through the unsaturated zone, low annual rainfall, high runoff, and high rates of evaporation and plant transpiration. The geologic setting comprises volcanic rock, called tuff, that was deposited by a series of volcanic eruptions approximately 13 million years ago. The characteristics of the volcanic rock have been studied in underground excavations and boreholes and by geologic mapping on the surface of the Yucca Mountain site. Mapping and other studies show that no faults with significant displacement (i.e., movement of more than a few meters) occur within the area defined for waste emplacement. The location, relative age, and amount of movement on faults have been characterized as part of DOE's seismic hazard analysis.

The configuration and elevation of the underground facility depends upon the thickness of overlying rock and soil, the characteristics of the rock that would host the facility, the location of faults, and the depth to groundwater. The facility would be excavated in the Topopah Spring Tuff rock unit, deep enough underground to prevent erosion from exposing the waste and to discourage human intrusion. The host rock for the geologic repository would be stable enough to sustain excavated openings during repository operations.

Barriers and Processes Important to Isolating Waste

The Yucca Mountain disposal system would use a combination of natural and engineered barriers to prevent or substantially reduce releases from the waste to the accessible environment. The natural barriers would reduce the amount of water entering emplacement drifts (seepage flux) and reduce the transport of radionuclides through the geologic setting. The engineered barriers (the waste package, the drip shield, the waste form inside the waste package, and crushed rock and steel invert on the floor of the emplacement drifts) would complement the natural barriers and strengthen the ability of the Yucca Mountain disposal system to isolate waste. Figure 2 identifies the natural and engineered barriers and summarizes how each would limit or mitigate the transport of radionuclides to the accessible environment.

Physical processes pertinent to isolating waste include those that control the movement of water. These processes begin with precipitation falling as rain or snow on the surface with much of this water flowing off or evaporating and a small amount moving into the mountain. The water would move down through the unsaturated zone to the level of the geologic repository, where it could contact the drip shields and then the waste packages.

The waste packages, fabricated from a corrosion-resistant nickel-based alloy (Alloy 22) are designed to prevent water from contacting the waste form for thousands of years. Eventually, however, water could corrode and breach the waste packages, slowly dissolve the waste form, and transport a small fraction of the radionuclides that are mobile (soluble in water or able to form colloids) out of the waste packages and downward through the unsaturated zone. The most prevalent waste form, commercial spent nuclear fuel, provides two additional barriers that would delay the release of radionuclides: the Zircaloy cladding and the matrix of the ceramic pellets that contain the radioactive material. After radionuclides are eventually released from the waste form, water would transport them out of the package and through invert materials below the package. While the drip shield is intact, or in instances of limited water seepage into the drifts, the release of radionuclides through the invert would be limited to diffusive processes that would further delay radionuclide migration. Finally, after passing though the engineered barriers, water would transport any mobile radionuclides down through about 300 m (1,000 ft) of unsaturated rock to the water table and then laterally to the accessible environment.

II. Preclosure Suitability Evaluation

In accordance with 10 CFR 963.12, the DOE has completed a preclosure suitability evaluation of Yucca Mountain's preclosure performance using the preclosure safety evaluation method described in 10 CFR 963.13. In the first step of the evaluation, a preliminary description of the site and the repository facilities was used to identify and categorize event sequences that could occur, even if the probability of such occurrence is low.

In the next step, design and operational concepts were used to assess how well the repository's surface and underground facilities would perform during Category 1 and Category 2 event sequences (as discussed in the next two paragraphs). The facilities' structures, systems, and components were evaluated using the criteria described in 10 CFR 963.14. These criteria include the ability to (1) contain radioactive material and limit releases, (2) implement control and emergency systems, (3) maintain the system and its components so they can perform their safety functions, and (4) maintain the option to retrieve wastes during the preclosure period. Part of the safety evaluation is a quality assurance classification analysis that ensures that quality controls are implemented in accordance with each repository feature's importance to safety (i.e., the system's impact on preventing or mitigating a radiation exposure to a worker or a member of the public).

The surface and subsurface areas where radioactive waste would be handled (i.e., the geologic repository operations area [10 CFR 963.2]) would be designed to withstand event sequences that could occur during the preclosure (operational) period. Waste emplacement is anticipated to last for approximately 24 to 50 years. The NRC's current licensing regulations require that the waste be able to be retrieved from the repository starting at any time up to 50 years after waste emplacement begins, unless a different time period is established by the NRC. Using the guideline's preclosure safety evaluation method, the DOE identified five external event sequences, such as floods and earthquakes, and 23 internal event sequences, such as accidentally dropping a fuel assembly, that could occur during the preclosure period.

Using the method set out in 10 CFR 963.13, potential event sequences were categorized according to their probability of occurrence during the preclosure period. Assuming a higher-temperature operating mode, Category 1 event sequences have an annual frequency that is greater than or equal to one chance in 100 per year, and Category 2 event sequences have an annual frequency that is less than one chance in 100 per year but greater than one chance in a million per year. Category 1 and Category 2 event sequence frequencies for a lower-temperature operating mode with a longer preclosure period were also considered. Both frequencies are based on an assumed 100-year preclosure period. Event sequences that have less than one chance in 10,000 of occurring during the preclosure period do not need further analysis.

Table 1 compares the results of the preclosure safety evaluation to the NRC's preclosure standards at 10 CFR 63.204 and objectives at 10 CFR 63.111(a) and (b) as referenced in DOE's suitability guidelines as the applicable radiation protection standards for the preclosure period. The information in the Preliminary Preclosure Safety Assessment for Monitored Geologic Repository Site Recommendation and the Yucca Mountain Science and Engineering Report and the results of the preclosure suitability evaluations described in this report show that doses are likely to be below the applicable radiation protection standards for the preclosure period.

III. Postclosure Suitability Evaluation

For over 20 years, the DOE has studied the processes important to postclosure total system performance. The data collected during this period have been used to develop conceptual and numerical models of the hydrologic, geochemical, thermal, and mechanical processes (as shown in Figure 3) that are linked together into a TSPA model. The TSPA model is then used to determine how the Yucca Mountain disposal system is projected to perform over the regulatory compliance period of 10,000 years, while taking into account uncertainties in the supporting process models and data. For this purpose, the guidelines set out analytical methods (10 CFR 963.16) and criteria (10 CFR 963.17) designed to allow a comprehensive evaluation of the postclosure performance of the Yucca Mountain disposal system. Figure 4 illustrates the sequence of abstractions that support a TSPA model.

The DOE has completed an evaluation of Yucca Mountain's postclosure performance in accordance with 10 CFR Part 963. The process used is a systematic analysis that identifies the FEPs that might affect the long-term performance of the Yucca Mountain disposal system. FEPs include features observed today, events such as those that could suddenly and randomly occur, and processes that gradually and continuously evolve. By simulating interactions between the FEPs and the radioactive waste, the TSPA model projects the amount of the waste that the disposal system would release over 10,000 years. The releases and release limits are expressed as a radiation dose to an individual and as a concentration of radionuclides in groundwater.

The NRC licensing regulations adopt three EPA standards derived from EPA rules for releases from a Yucca Mountain disposal system. Individuals are protected by two standards: one applies to releases where the system is undisturbed by human intrusion, and the other applies to releases that might be caused by inadvertent human intrusion. The third standard protects groundwater.

The suitability evaluations described in this report evaluate whether the site is likely to meet the applicable radiation protection standards for individual and groundwater protection. An inadvertent human intrusion was considered but is not projected to occur during the regulatory compliance period of 10,000 years, so dose limits for the human intrusion scenario do not apply in the postclosure suitability evaluation. The results of the human intrusion, however, will be included in the final environmental impact statement as required by the NRC human intrusion standard (10 CFR 63.321 and 10 CFR 63.322).

As noted earlier, the barriers and the processes that isolate waste control the movement of water through the unsaturated zone to the underground facility and from there to the accessible environment. Water moving into the underground facility would be affected by the physical and chemical processes associated with the decay heat produced by radioactive decay and could interact with waste packages and waste forms. These processes could lead to the mobilization of radionuclides. Eventually, water containing radionuclides could move out of the underground facility and downward through the unsaturated zone. Subsequently, it could move into the saturated zone, where it could be transported to the accessible environment where the reasonably maximally exposed individual resides.

The terms "accessible environment" and "reasonably maximally exposed individual" are regulatory terms. According to DOE's site suitability guidelines, which adopt the definition of a reasonably maximally exposed individual set out in NRC regulations, all postclosure releases are evaluated within the accessible environment at a point specified at 10 CFR 63.302. This point, which is where the reasonably maximally exposed individual is assumed to live, lies approximately 18 km (11 mi) from within the repository footprint. The individual is assumed to withdraw and use water from a well located above the highest concentration of radionuclides in the plume of contamination. These and other characteristics of the reasonably maximally exposed individual are specified by the NRC at 10 CFR 63.312, which is adopted by reference in DOE's site suitability guidelines at 10 CFR 963.2.

The results presented in the postclosure suitability evaluation are based on the TSPA completed in accordance with DOE's site suitability guidelines, which cross-reference the radiation protection standards set out in the final NRC rules (which adopt EPA's final radiation protection standards). The TSPA comprises models and sensitivity analyses that are described in the following reports:

  1. Total System Performance Assessment for the Site Recommendation describes the TSPA-SR model that evaluates the possible performance of the repository for a 10,000-year compliance period and through the period of geologic stability. The TSPA-SR model evaluation is based on a higher-temperature operating mode. The model evaluates performance using provisions contained in proposed EPA and NRC regulations.

  2. FY01 Supplemental Science and Performance Analyses describes a supplemental TSPA model that quantifies conservatisms and uncertainties in the TSPA-SR model. It also evaluates new and updated information that improved DOE's scientific understanding of the potential effects of a range of thermal operating modes. The supplemental TSPA model evaluates performance using provisions contained in proposed EPA and NRC regulations. However, key aspects of the final EPA standard were examined as were possible impacts on assessments of system performance due to differences between the EPA's proposed and final standards. In particular, Appendix A of the analyses evaluates when the waste package would degrade sufficiently that a human intrusion could occur without recognition by the drillers as a result of exploratory drilling for groundwater, in a manner consistent with the provisions set out in the final EPA (40 CFR 197.25) and NRC (10 CFR 63.321 and 10 CFR 63.322) regulations.

  3. Total System Performance Assessment—Analysis for Disposal of Commercial and DOE Waste Inventories at Yucca Mountain—Input to Final Environmental Impact Statement and Site Suitability Evaluation—referred to as the "TSPA Report for Final Environmental Impact Statement and Suitability Evaluation"—describes the revised supplemental TSPA model. Model results evaluate performance using provisions contained in the final EPA standards (40 CFR Part 197). The report includes results of performance assessments for the 10,000-year regulatory compliance period, as well as after 10,000 years through the period of geologic stability. The analyses also consider the potential effects of a range of thermal operating modes.

  4. Total System Performance Assessment Sensitivity Analyses for Final Nuclear Regulatory Commission Regulations—referred to as the "TSPA Sensitivity Analyses for NRC Regulations"—provides additional sensitivity analyses conducted following the promulgation of the final NRC regulations. Specifically, the report includes an evaluation of the effect of the final NRC provision that specifies a water demand of 3,000 acre-ft/yr be used for the dose calculation to the reasonably maximally exposed individual for evaluation against the individual protection standard (10 CFR 63.312(c)). TSPA model calculations used an average water demand of approximately 2,000 acre-ft/yr, consistent with proposed NRC regulations.

Nominal Scenario—The nominal scenario analyzed by the TSPA-SR model and both the supplemental and revised supplemental TSPA models incorporates the important effects of processes illustrated conceptually in Figure 3, including climate change to a wetter environment, seismic activity, and repository heating. To respond to changes in the regulations, the nominal scenario analyses have been updated and are discussed in the TSPA Report for Final Environmental Impact Statement and Suitability Evaluation and in the TSPA Sensitivity Analyses for NRC Regulations.

Disruptive Events—The DOE site suitability guidelines provide criteria for the disruptive events that could affect ongoing processes or directly interrupt and permanently alter the repository system. In contrast to the expected processes considered in the nominal scenario, disruptive events are unexpected events that could directly release radionuclides to the surface or accelerate releases caused by ongoing processes. Three criteria, 10 CFR 963.17(b)(1) through (3), identify three potentially disruptive events for consideration:

  1. Volcanism
  2. Seismic events
  3. Nuclear criticality.

The evaluation of volcanism in the disruptive scenario considers both igneous intrusion and volcanic eruption. In a scenario initiated by an igneous intrusion, molten rock would contact one or more waste packages and release radionuclides into the groundwater. The estimated mean probability of such an intrusive scenario is 1.6 chances in 10,000 over 10,000 years. In a scenario initiated by a volcanic eruption, molten rock would contact one or more waste packages and release radionuclides into the atmosphere. The estimated mean probability of an eruptive scenario is 1.2 chances in 10,000 over 10,000 years. Although the probabilities of these scenarios are very low, they are not below the DOE site suitability threshold for very unlikely events (i.e., one chance in 10,000 over 10,000 years) (10 CFR 963.16(b)(4)). The TSPA, therefore, evaluated releases caused by both igneous intrusion and volcanic eruption in the disruptive scenario.

Potential impacts of seismic events were excluded from the disruptive scenario because of low consequences and low probability of occurrence—the waste packages would not be damaged by rockfalls or vibratory ground motion. However, because vibratory ground motion may damage the cladding that encases spent nuclear fuel and make it more susceptible to other types of degradation, the TSPA-SR model and both supplemental TSPA models include seismically induced failure of cladding in the nominal scenario.

The evaluation of nuclear criticality considers whether the geometric configuration of spent nuclear fuel could produce a self-sustaining chain reaction that releases large amounts of radiation and energy. Nuclear criticality was not included in the TSPA because it has less than one chance in 10,000 of occurring in 10,000 years and because the consequences would not significantly change the projected releases (10 CFR 63.342). The low probability and consequences are primarily attributed to the characteristics of the geologic setting. The absence of water reduces the likelihood that nuclear fission could occur within the waste packages and produce the neutrons that would initiate a chain reaction. After the packages release the fissile isotopes, such as plutonium-239, into the unsaturated zone, the absence of organic material that could attract and accumulate the isotopes reduces the likelihood that the isotopes could reach a critical configuration.

Uncertainty—There are recognized limitations to TSPA models and their ability to forecast the future behavior of the Yucca Mountain disposal system. One of the most important uncertainties in the TSPA analyses is in projecting the long-term performance of natural and engineered barriers using data derived from short-term, multiyear tests. Also important are the inherent uncertainties in forecasting changes in climate and in other processes over the 10,000-year postclosure period. Because of the long time frames over which the disposal system must perform, the natural variability in features of a geologic repository, and limitations on the amount of data that can be collected, uncertainties will always remain.

The number and detail of process models developed for the Yucca Mountain site and the complex coupling among those models make direct incorporation of all model uncertainties difficult. Model uncertainties arise from several sources, including (1) parameter values, (2) conceptual model representations, and (3) the mathematical models used to implement the conceptual models. In many cases the value of model input parameters is uncertain. Parameter value uncertainty may result from imperfect knowledge, limited data, and treatment of variability as uncertainty. Parameter value uncertainty can be addressed by developing a probability distribution that captures the full range of potential values or by using a single value that is conservative. Conceptual model uncertainty arises from incomplete understanding or characterization of FEPs that will affect a potential repository. There may be several equally plausible ways to conceptualize a specific process being modeled; that is, multiple alternative conceptual models may be considered to explain the current data equally well. Mathematical model uncertainty may be introduced by simplifications and approximations that generally are introduced to represent conceptual models to make the problem tractable and to implement the model in a computer program.

Conceptual model uncertainties can also arise from an incomplete knowledge of complex physical processes. Although current conceptual models are consistent with existing knowledge, there have been questions raised by some in the technical community regarding how some processes are conceptualized in process models feeding the performance assessment. Examples of such uncertainties include detailed coupling of thermal-hydrologic processes and stability over very long periods of the passive oxide films that will develop on the Alloy 22 waste package surface. Portions of the current and planned testing and analysis are aimed at gaining a better understanding of such issues such that their representation in conceptual models is facilitated. In turn, these models can be included in the performance assessment, as appropriate. Remaining uncertainties of this type have been identified and provide a basis for part of DOE's ongoing and future testing, analysis, and modeling.

These issues have been addressed in a number of ways through an uncertainty strategy that is focused on quantifying uncertainties in a defensible manner. In some cases, conservative assumptions have been made or parameter values have been bounded, creating what has been referred to as an "unquantified uncertainty." If the entire probability distribution for a parameter value is used, this is referred to as a "quantified uncertainty" because the Monte Carlo analysis used by TSPA will sample the full range of potential parameter values. If a single conservative value is used to represent a parameter value, this introduces an uncertainty of unknown magnitude into the Monte Carlo simulations, although the TSPA results are likewise believed to be conservative. More than one conceptual model may be consistent with available data, and in the absence of definitive data sets or compelling technical arguments, the most conservative model is normally used. This is another case where an uncertainty of unknown magnitude (i.e., an unquantified uncertainty) is introduced into the TSPA analyses. Efforts to address uncertainties have included enhancement of engineered barriers to provide additional defense in depth and the investigation of natural analogues to provide additional information to support the model representations used. These activities are discussed in FY01 Supplemental Science and Performance Analyses, which also explores the significance of these unquantified uncertainties.

The end result is that analyses of total system performance have used a mix of probabilistic representations, single-value conservative estimates, and conservative assumptions and models. This approach has been used in other projects, but it complicates the determination of quantification of the degree of conservatism associated with the projected margin relative to regulatory standards. In addition, the mixing of varying degrees of conservatism in models and parameter representations could be viewed as complicating the analysis. On the other hand, it increases the confidence that the results of the analysis conservatively bound any future performance.

In Total System Performance Assessment for the Site Recommendation, the critical assumptions, data limitations, and residual uncertainties in the TSPA-SR model were identified, and the significance of uncertainties included in the TSPA-SR model were evaluated. A number of activities were undertaken to accomplish this evaluation:

The TSPA-SR model included a mixture of conservative and realistic inputs. Using this approach, model results for the nominal scenario, for example, showed significant uncertainty, approximately two orders of magnitude at the time of peak dose (see Figure 5a). Although the DOE considered this a defensible approach, some reviewers of the work thought the approach masked information and understanding and called for a reevaluation of the TSPA-SR model. In response to these comments and observations, the model was revised and documented in FY01 Supplemental Science and Performance Analyses. The revised model used in the analyses, referred to as the supplemental TSPA model, included quantification of some important uncertainties that were previously unquantified for nominal scenario performance. The supplemental TSPA model showed wider ranges of doses at earlier times (prior to 20,000 years as compared with the TSPA-SR model results for that same time period), but this added indication of uncertainty was the result of incorporating new information into the waste package modeling and not of simply replacing single-point conservative values with distributions. For the nominal scenario, the net change in terms of dose during the period of 10,000 years after disposal was small, moving from a mean annual dose of zero (in the TSPA-SR model) to a peak mean annual dose of less than 2.0 x 10-4 mrem/yr in the supplemental TSPA model (see Figure 5b). This dose was a result of a few early waste package failures, which resulted from changes made to the model taking into consideration the uncertainty associated with possible improper heat treatment of the waste package lid welds. In terms of mean annual dose beyond 10,000 years, however, additional uncertainty was introduced by changes made to parameter distributions and models that support the supplemental TSPA analyses. The net effect beyond 10,000 years was a lower, more defensible mean dose projection. Because the supplemental TSPA model incorporated additional quantified uncertainties, the projection was also associated with a wider uncertainty band.

The supplemental TSPA model also evaluated the effects on performance of operating the repository at lower temperatures. From the standpoint of the resulting calculated uncertainty distribution in the supplemental TSPA model results, both higher- and lower-temperature operating modes showed similar performance and range of uncertainty for the nominal scenario. This conclusion acknowledges that the existing conceptual models supporting the TSPA may or may not have the capacity to discern effects from the temperature or moisture differences between the operating modes. While acknowledging that the current TSPA models may not have the capability to discern all temperature effects, FY01 Supplemental Science and Performance Analyses shows that many process models do have the capability to discern differences between the operating modes at the subsystem level. Incorporation of additional temperature-dependent processes not currently captured in the TSPA might allow for distinguishable differences at the system level between the operating modes. The supplemental TSPA model was subsequently modified upon issuance of the final EPA rule at 40 CFR Part 197. The modified model, known as the revised supplemental TSPA model, differs slightly from the supplemental TSPA model and is described in detail in the TSPA Report for Final Environmental Impact Statement and Suitability Evaluation. The report includes comparative diagrams showing the results of the TSPA-SR model, the supplemental TSPA model, and the revised supplemental TSPA model.

To capture the full detail of the uncertainty and variability in the behavior of the Yucca Mountain disposal system, the reports described above display graphical representations of the results of the TSPA analyses that highlight the mean, the median, and the 5th and 95th percentiles of the distribution of results. The results explicitly incorporate uncertainty by calculating estimates of statistical measures of the output as a means to evaluate performance, thus capturing the probabilistic treatment of uncertainty in the total system performance. In the same manner described in Total System Performance Assessment for the Site Recommendation, these statistical measures are calculated at each time step of the dose histories and include data from all realizations of the probabilistic simulations. These statistical measures—the mean, median, and 5th and 95th percentiles—are superimposed on the diagrams showing all realizations; each realization has some definite possibility of representing the actual radiological exposure from the disposal system. Because of their appearance, such diagrams showing all realizations (usually hundreds of separate realizations) are called "horsetail diagrams." The uncertainty in projecting the exposure is represented by the range of outcomes, expressed by the range of realizations produced by the TSPA model (for example, the spread between the 5th and 95th percentiles). By calculating a distribution of exposures, the models reflect the range of parameter values and models that could be appropriate, knowing that the actual exposure cannot be reasonably predicted except in probabilistic terms. Such projections of possible outcomes, given quantified uncertainties in the inputs, is a common method of displaying expected (mean) risk and associated uncertainties in probabilistic risk analyses.

As an example, Figure 5 shows two horsetail diagrams for the nominal scenario generated by the TSPA-SR model and the supplemental TSPA model. Both diagrams represent 300 separate realizations of performance and reveal important information about forecasted performance of the modeled system. The uncertainty band in the range of realizations shown in Figure 5b is a function of the quantified uncertainty in the inputs to the system representation. For example, in the supplemental TSPA model, uncertainty has been quantified regarding the possibility of improper heat treatment of the waste package lid welds, which resulted in the possibility of waste package failures prior to 10,000 years. The additional quantification of uncertainty represented in Figure 5b results in realizations of dose histories that are different from those resulting from the conservative assumptions made in the TSPA-SR model used to generate Figure 5a. The wider range of quantified uncertainty in the supplemental TSPA model, in this case, leads to broader uncertainty in the performance assessment results beyond 100,000 years, expressed by the range of realizations shown in Figure 5b. However, this added uncertainty resulted in negligible change during the period of 10,000 years after disposal. The diagrams also show that the additional quantified uncertainties and updated models in the supplemental TSPA model lead not only to a reduction in the mean peak dose beyond 10,000 years but to a broader uncertainty band in the range of the peak annual doses. An alternative way to express this result is that the conservative models of the TSPA-SR lead to a higher peak dose with a narrower range of annual doses.

Comparison of dose histories over a million years using the TSPA-SR model and the supplemental TSPA model shows the following two characteristics. First, the supplemental TSPA model shows significantly wider ranges of doses at a given time and of times to reach given doses. Second, except at early times, the magnitude of the mean annual dose is less for the supplemental TSPA model, and it occurs later in time.

A comparison of Figures 5a and 5b shows that the supplemental TSPA model produces a broader range of annual doses or times to specific annual dose values than does the TSPA-SR model. This is represented quantitatively by the distribution of realizations at particular dose rates and at particular times. The broader range is a result of the additional uncertainties and updated models that have been incorporated into the supplemental TSPA model. In many cases, simplified or bounding models have been replaced with more physically representative models that include quantified uncertainties in their parameters. For example, a bounding solubility model for neptunium in the TSPA-SR model has been replaced with a more complex model that accounts for the solubility of secondary phases that control the solubility. The updated solubility model is believed to be more realistic, but the uncertainties in the model lead to a broader range of neptunium concentrations than the previous model. Propagation of these uncertainties, as well as those of all of the other updated process models, results in the broad ranges that are seen in results of the supplemental TSPA model.

A second observation is based on a comparison of the estimates of mean performance (dose rate and time to dose) using the TSPA-SR model and the supplemental TSPA model. This comparison shows that after approximately 10,000 years, the mean annual dose using the supplemental TSPA model is always less than the mean using the TSPA-SR model. The difference between the mean estimates is one measure of the magnitude of the conservatism in the TSPA-SR model. For example, at 30,000 years, the difference between the mean estimates of annual dose is about three orders of magnitude, and at the time of peak mean dose, the difference is about one order of magnitude.

During the period prior to 10,000 years, the small annual doses (less than about 2 x 10-4 mrem/yr) indicated by the supplemental TSPA nominal model exceed the zero annual dose calculated by the TSPA-SR model, and the TSPA-SR model could be interpreted as being nonconservative with respect to the supplemental TSPA model during this time. These small doses, resulting from the revised treatment of uncertainty regarding the potential for improper heat treatment of lid welds on waste packages, are more than a factor of 1,000 smaller than the doses forecasted for the disruptive scenario (0.1 mrem/yr). As discussed in the following section on postclosure results, the combined nominal and disruptive scenario dose is approximated as the probability-weighted mean annual dose from the disruptive scenario. Differences between the supplemental TSPA model and the TSPA-SR model for the first 10,000 years after disposal would have essentially no impact on this mean dose.

At 10 CFR 63.303, the NRC states that in the case of the specific numerical requirements for individual, human intrusion, and groundwater protection, compliance with the NRC's numerical standards for licensing is to be based upon the mean of the distribution of projected doses of DOE's performance assessments for 10,000 years after disposal. The mean is conservative because it is sensitive to the number of realizations having zero and nonzero annual doses. Because the mean is an average of all realizations at any given point in time, if any realizations have a nonzero dose, the mean will likewise be a nonzero number. Consequently, the mean will trend towards the higher percentiles and will stay above the median (see Figure 5b); performing greater numbers of calculations does not materially affect the location of the mean. Because the quantitative estimates of repository performance should not be dominated by unrealistic or extreme situations or assumptions, the mean value is an appropriate measure of performance for comparison with regulatory standards because it conservatively represents hundreds of realizations that collectively capture parameter and model uncertainty. For the evaluation of site suitability, the mean values of the probabilistic simulations in the TSPA Report for Final Environmental Impact Statement and Suitability Evaluation have been used.

Despite significant efforts to reduce and quantify uncertainties in the TSPA inputs, unquantified uncertainties remain and, in most cases, are addressed through the use of conservative assumptions and bounded parameter values. Their potential implications to performance and risk have been examined, and strategies for managing them have been developed. This information and a discussion of key remaining uncertainties are presented in Uncertainty Analyses and Strategy. In some cases, conceptual model process uncertainties have been identified for which gaining a better understanding is appropriate (e.g., long-term stability of passive films on the waste package surface), and testing, analysis, and modeling activities are continuing to expand and refine these models.

The DOE recognizes that the results of TSPA modeling do not constitute absolute certainty with respect to future outcomes because absolute certainty is unobtainable, as both EPA and NRC explicitly acknowledge in their regulations. Even in the presence of the remaining uncertainties, however, the DOE is satisfied with the level of treatment and understanding of these uncertainties in the current TSPA analyses supporting the site recommendation decision process and has confidence in the overall safety of the repository.

Additional confidence is gained through the analysis and investigations of anthropogenic (man-made) and natural analogues to the processes and materials—natural and engineered—used in the repository. For example, geothermal reservoirs at Yellowstone, Wyoming, and at Broadlands, New Zealand, have provided some insight into the thermal-hydrologic and chemical processes that would occur in the repository environment. Analogue systems, which have occurred over time periods of decades to millennia and over spatial scales of up to tens of kilometers, are beyond the scale of laboratory experiments such as metal corrosion tests. Therefore, careful observation and analysis of these analogue systems provide increased confidence in the performance of the Yucca Mountain disposal system.

Further, in addition to the inclusion of multiple natural and engineered barriers in the disposal system and the consideration of analogues in the analyses, additional provisions are being implemented to increase confidence that the postclosure performance objectives will be met. These provisions include model validation, implementation of a rigorous quality assurance program, and continuation of a performance confirmation program. However, because uncertainty is integrated into the assessment of total system performance, the DOE does not expect that additional information will significantly change the conclusions reached in this site suitability evaluation.

Postclosure Results

Table 2 compares the limits of the applicable radiation protection standards to the results of the postclosure suitability evaluation. The results using the revised supplemental TSPA model and sensitivity analyses support the conclusion reached in the Yucca Mountain Preliminary Site Suitability Evaluation that releases for the 10,000-year regulatory compliance period would likely fall below the applicable radiation protection standards. After the compliance period, the supplemental TSPA model and revised supplemental TSPA model results show lower doses than the TSPA-SR model and a lower peak dose at or before 1 million years. These differences are primarily the result of more realistic treatments of waste package degradation and the solubility characteristics of neptunium, plutonium, and thorium.

Individual Protection—For the evaluation against the NRC's individual protection standard (10 CFR 63.311, referenced in DOE's site suitability guidelines at 10 CFR 963.2), the results of the nominal scenario are combined with the results of the disruptive scenario (which included both igneous intrusion and volcanic eruption). Because the calculated peak mean annual doses over the 10,000-year regulatory compliance period for the nominal scenario are thousands of times smaller than the probability-weighted mean annual dose for the disruptive scenario, the combined mean annual dose for comparison with individual protection standards is approximated as the probability-weighted mean annual dose from the disruptive scenario.

The horsetail diagram in Figure 6 shows the results of the revised supplemental TSPA model analyses for the disruptive scenario (for the higher-temperature operating mode) as described in the TSPA Report for Final Environmental Impact Statement and Suitability Evaluation. The diagram shows the results for 500 of 5,000 realizations calculated by the revised supplemental TSPA model; the 500 realizations include the realization with the highest probability-weighted dose within the regulatory period. In addition, statistical measures (i.e., the mean, median, and 5th and 95th percentile dose curves), based on all 5,000 realizations, are also shown. The mean annual probability of an igneous event disrupting a repository at Yucca Mountain is 1.6 x 10-8, or approximately 1 in 60 million per year. The calculated results show a probability-weighted peak mean annual dose of 0.1 mrem/yr over the 10,000-year regulatory compliance period. This dose is probability-weighted in accordance with the TSPA definition in DOE's site suitability guidelines at 10 CFR 963.2 to quantify the risk associated with igneous activity. The calculated mean dose falls below the applicable radiation protection standard for individual protection (15 mrem/yr), as do all the calculated values of probability-weighted annual dose (all 5,000 realizations within the regulatory time period).

Another way to display this information is to modify the horsetail diagram with the individual realizations removed and the statistical measures displayed separately (Figure 7). This representation emphasizes the general features of the plot without all of the details. Examining the annual dose histories from the standpoint of the probability distribution of realizations can also provide insights into the aggregate or system-level significance of uncertainties of the inputs. For instance, the horsetail diagrams can be sliced vertically, at particular times, or horizontally, at particular dose rates, to reveal details of the distributions. A variant of this approach is to produce a slice at the time of the peak of the mean of the realizations. Using this approach, a cumulative distribution function can be created by summing the number of realizations at particular probability-weighted annual doses, or a histogram created by summing the number of realizations within various dose-rate increments. These types of plots provide additional insights into the details of the TSPA results. The cumulative distribution functions and histograms are most useful for comparisons between different TSPA models. For instance, the slope of the cumulative distribution function is a function of the amount of uncertainty captured in the TSPA calculation. Comparison of cumulative distribution functions constructed for different TSPA models, such as the TSPA-SR model and the supplemental TSPA model, will immediately reveal differences in the degree of uncertainty quantification because of differences in the slopes of the cumulative distribution functions. Similarly, differences between histograms can reveal differences in the statistical nature of the results from different models.

For the assessment of individual protection, the TSPA-SR model and both supplemental and revised supplemental TSPA models calculated radionuclide doses assuming an average annual water demand of approximately 2,000 acre-ft, consistent with the NRC's proposed rule. In its final rule, the NRC specifies that an annual water demand of 3,000 acre-ft be used in the assessment of individual protection. In projecting the dose, the TSPA analyses conservatively assumed that groundwater volume for use and consumption contains 100 percent of the annual radionuclide releases reaching the accessible environment. The mean doses calculated for the evaluation of individual protection, therefore, represent conservative (higher) estimates. Calculated values of dose for groundwater-pathway-dominated scenarios (e.g., igneous intrusion scenario and nominal scenario) would be reduced by approximately one-third of that shown in Table 2 for individual protection using the increased NRC-specified annual water demand of 3,000 acre-ft since the water volume remains the same.

Human Intrusion—In Total System Performance Assessment for the Site Recommendation, Yucca Mountain Preliminary Site Suitability Evaluation and the TSPA Report for Final Environmental Impact Statement and Suitability Evaluation, the DOE evaluated a stylized human intrusion scenario based on the NRC's proposed regulations. Those proposed regulations specified that the prescribed human intrusion scenario be assumed to occur at 100 years after repository closure (proposed 10 CFR 63.113(d)).

Following an intrusion at 100 years, the revised supplemental TSPA model calculated a mean annual dose of 4.8 x 10-3 mrem/yr over the 10,000-year regulatory compliance period. The final NRC regulation (10 CFR 63.321) adopted the EPA human intrusion standard and prescribed a set of assumptions for the human intrusion scenario. The final NRC rule requires the DOE to determine the earliest time that the waste package would degrade sufficiently that a human intrusion could occur without recognition by the drillers (as a result of exploratory drilling for groundwater). In accordance with the final NRC rule, if exposure from a human intrusion is projected to occur more than 10,000 years after disposal, then the results of the analysis of the human intrusion scenario would be presented in the Yucca Mountain environmental impact statement, and the dose limits for the human intrusion standard would not apply in this suitability evaluation.

Based on analyses presented in FY01 Supplemental Science and Performance Analyses, the DOE has determined that the earliest time after disposal that the waste packages would degrade sufficiently that a human intrusion could occur without recognition by the drillers is 30,000 years. Before that time, the drillers would recognize that they had drilled into a waste package because the compressive strength and ductility of the metals from which the waste packages and drip shields are fabricated differ significantly from the rock that would surround them. Drillers would notice these differences. For example, the drilling assembly would buckle and bend when the bit attempts to penetrate the titanium drip shield and waste package (drill bits that are designed for rock do not easily penetrate metal, particularly titanium). The drillers should, therefore, recognize that they have attempted to drill into some material other than rock for at least as long as the drip shields or waste packages remain intact, which is approximately 30,000 years. This is true even if manufacturing defects cause a few waste packages to fail prematurely, as the failures would not weaken the overall structural integrity of the waste packages and their resistance to drilling. Because the human intrusion is not expected to occur without recognition by the drillers within the 10,000-year compliance period, the suitability evaluation does not consider releases from a human intrusion. Admittedly, there is some uncertainty about the timing of when penetrating a waste package would not be detected. However, the results above, based upon an assumption of penetration at 100 years, show that the resulting dose is orders of magnitude less than the applicable radiation protection standard. Therefore, even if penetration was much earlier than expected, the resulting doses are of little consequence for the suitability evaluation.

Groundwater Protection—The TSPA models evaluated the releases of radionuclides from waste in the Yucca Mountain disposal system into groundwater. In accordance with 10 CFR 963.2, which references 10 CFR 63.331, concentration limits for combined radium-226 and radium-228 and gross alpha activity—expressed as picocuries per liter (pCi/L)—include natural background, but the limit for combined beta- and photon-emitting radionuclides—expressed as millirem per year (mrem/yr)—does not.

For the evaluation of groundwater protection, the TSPA models evaluated the level of radioactivity that the disposal system would release into a representative volume of 3,000 acre-ft of groundwater that would be withdrawn and used annually at a location within the accessible environment (10 CFR 63.332). This is the groundwater volume that would be withdrawn annually from a well located above the highest concentration of radionuclides in the plume of contamination at the same location as the reasonably maximally exposed individual, which is approximately 18 km (11 mi) from within the repository footprint. The TSPA conservatively assumed that the groundwater contains 100 percent of the released radionuclides reaching that point.

In the revised supplemental TSPA model analyses described in the TSPA Report for Final Environmental Impact Statement and Suitability Evaluation, the releases during the 10,000-year regulatory compliance period come from early waste package failures as a result of assumed improper heat treatment of the waste package lid welds. The calculated peak mean concentrations during the 10,000-year regulatory compliance period for the gross alpha activity (which do not include background radiation) are 1.8 x 10-8 pCi/L of gross alpha-emitting radionuclides for the higher-temperature operating mode and 3.3 x 10-8 pCi/L of gross alpha-emitting radionuclides for the lower-temperature operating mode. The calculated peak mean concentration for total radium (combined radium-226 and radium-228) is less than 10-10 pCi/L for both the higher-temperature and lower-temperature operating modes.

Available data indicate that the gross alpha background concentration is no greater than 1.1 pCi/L. Total radium background concentration is no greater than 1.04 pCi/L. Because calculated concentrations for gross alpha and total radium are orders of magnitude lower than their natural background concentration, the combined totals (background and calculated concentrations), when rounded, are the same as natural background. Therefore, the natural background radiation concentrations for gross alpha-emitting radionuclides and total radium are used for comparison to NRC radiation protection standards as shown in Table 2.

For a full understanding of the results of the TSPA calculations, it is useful to look at the results for the entire 1-million-year period used in the calculations. Figure 8 shows the horsetail diagram for gross alpha activity concentrations for the regulatory period and for the time period out to 1 million years for the lower-temperature operating mode. Figure 9 shows the same diagram for total radium activity concentrations. Both figures show the mean, median, 5th and 95th percentiles of the distribution of results. It is apparent from these figures that the magnitude of the calculated gross alpha activity concentration, total radium activity concentration, and the uncertainty in the calculated values for both increase significantly after approximately 70,000 years. A cumulative distribution function and histogram of gross alpha and total radium activity concentrations at the time of peak mean dose during the regulatory period showed that over 75 percent of the realizations have a zero activity concentration value. This is because no waste packages have failed in those realizations during the regulatory period.

The calculated maximum mean annual dose to any critical organ from combined beta- and photon-emitting radionuclides over the period of regulatory compliance is 2.3 x 10-5 mrem/yr for the higher-temperature operating mode and 1.3 x 10-5 mrem/yr for the lower-temperature operating mode. These calculated doses, as shown in Table 2, are below the applicable radiation protection standards. Iodine-129 is one of the predominant contributors to organ dose. Figure 10 shows a horsetail diagram with the mean, median, 5th and 95th percentiles for the annual dose to a critical organ from iodine-129; for iodine-129, the critical organ is the thyroid. This figure includes calculated values for realizations out to 1 million years and shows a significant increase in the values of annual dose from iodine-129 after approximately 70,000 years and an increase in the magnitude of the uncertainty in those values. A cumulative distribution function and histogram of this information at the time of peak mean dose during the regulatory period showed that over 75 percent of the realizations have a zero critical organ dose value from iodine-129. This is because no waste packages have failed in those realizations during the regulatory period. The maximum critical organ dose value from iodine-129 during the regulatory period is 1.4 x 10-3 mrem/yr, which is orders of magnitude less than the regulatory standard. The results for technetium-99 are similar to iodine-129.

Summary of Results

DOE has applied the preclosure suitability evaluation method and criteria set forth in 10 CFR 963.13 and 10 CFR 963.14 to evaluate the suitability of the site for the preclosure period. The results of the preclosure safety evaluation show that the Yucca Mountain site is likely to meet applicable radiation protection standards for the preclosure period. DOE has also applied the postclosure evaluation method and criteria set forth in 10 CFR 963.16 and 10 CFR 963.17 to evaluate the suitability of the site for the postclosure period. The results of the TSPA show that the Yucca Mountain site is likely to meet applicable radiation protection standards for the postclosure period, as well.

Previous Section | Next Section