4.1.1.1 Total System Performance Assessment Methods and Objectives

TSPA is a systematic analysis that synthesizes information (data, analyses, and expert judgment) about the site and region with the design attributes of the engineered barriers of the repository system. As defined in 10 CFR 63.2 (66 FR 55732),

Performance assessment means an analysis that:

1. Identifies the features, events and processes (except human intrusion) and sequences of events and processes (except human intrusion) that might affect the Yucca Mountain disposal system and their probabilities of occurring during 10,000 years after disposal;

2. Examines the effects of those features, events, and processes and sequence of events and processes upon the performance of the Yucca Mountain disposal system; and

3. Estimates the dose incurred by the reasonably maximally exposed individual, including the associated uncertainties, as a result of releases caused by all significant features, events, and processes, and sequences of events and processes weighted by their probability of occurrence.

Features are the physical components of the total repository system, including both the natural system (e.g., the geologic setting) and the engineered system (e.g., the waste package). Processes typically act more or less continuously on the features; for example, moisture flow through the geologic materials and corrosion of the waste package. Events also act on the features but at discrete times. Examples include seismic and volcanic events.

The TSPA approach and models are designed to address the processes that could lead to release and migration of radionuclides, and the radiological consequences to potential human receptors. The approach is intended to provide a transparent analysis of the geologic repository in terms of the performance of the natural and engineered barriers over long periods of time.

40 CFR Part 197 provides that the DOE and NRC should determine compliance with the radiation protection standard of 40 CFR 197.20 based on the mean of the distribution of the highest doses resulting from the performance assessment. In the background information accompanying their final rule (66 FR 32074, p. 32125), the EPA noted that they believe that a thorough assessment of repository performance should examine the full range of reasonably foreseeable conditions and processes. However, they also stated that quantitative estimates of repository performance should not be dominated by unrealistic or extreme situations or assumptions. Therefore, the EPA believed the use of the mean was reasonable but still conservative. They further noted that the use of the mean was consistent with the literal mathematical interpretation of the term "reasonable expectation" and with the approach used to certify Waste Isolation Pilot Plant.

During their consideration of the appropriate performance measure, the EPA evaluated other possible measures, such as the median value of the distribution, or more extreme measures, such as the 95^th or 99^th percentile. Their analysis showed that the use of either the mean or the median was reasonably conservative because both are influenced by the high exposure estimates, without reflecting only the high dose results.

Although the EPA selected the mean for the compliance determination, both the EPA and NRC provide that the DOE consider the uncertainties inherent in performance assessment results. One way that the DOE's method addresses this concern is by presenting and analyzing the full range of doses resulting from the performance assessment. In addition, the DOE has performed numerous sensitivity and uncertainty analyses to characterize the properties and processes that are particularly important to dose calculations.

The TSPA described briefly in Section 4.4 and in more detail in Total System Performance Assessment for the Site Recommendation (CRWMS M&O 2000a), Volume 2 of FY01 Supplemental Science and Performance Analyses (BSC 2001b), and Total System Performance Assessment—Analyses for Disposal of Commercial and DOE Waste Inventories at Yucca Mountain—Input to Final Environmental Impact Statement and Site Suitability Evaluation (Williams 2001a) examines the performance of the potential repository for a broad range of potential subsurface and surface conditions (e.g., hydrologic, geologic, climatic, and biosphere) and evaluates potential radiation doses to future generations. The radiation protection standards that would apply to the postclosure performance of a Yucca Mountain repository are found in regulations promulgated by the EPA (40 CFR 197.20, 197.25, and 197.30) and the NRC (10 CFR 63.113(b), (c), and (d) [66 FR 55732]). Specific technical requirements for a comprehensive TSPA are prescribed in the NRC regulation, 10 CFR Part 63.

The TSPA method requires evaluation of postclosure performance of the Yucca Mountain disposal system where there is no human intrusion into the repository. The final regulations also require evaluation of the performance of the system where there is human intrusion, in accordance with NRC regulations. The TSPA evaluation method in both cases is the same, except that the TSPA method for human intrusion provides prescribed assumptions about the human intrusion scenario (10 CFR 63.322 [66 FR 55732]).

Human intrusion refers to inadvertent intrusion into the repository as a result of exploratory drilling for groundwater. Limited intrusion means a single borehole that penetrates the repository and the underlying groundwater aquifer.

To present the assessment results transparently, the first case (without human intrusion) is further subdivided into:

• A nominal scenario composed of the likely FEPs representing the most plausible evolution of the repository system, without the occurrence of unlikely disruptive FEPs

• Disruptive scenarios that include unlikely FEPs that could diminish the waste isolation capability of the repository system (e.g., igneous activity).

The result of a TSPA analysis is a distribution (range) of possible outcomes of future performance. Because of the probabilistic nature of the method, TSPA results can capture and display much of the uncertainty associated with complex models and unknown future conditions. For this reason, however, the results should be regarded as indicative of future performance, not as predictive in a precise way.

The TSPA methodology (i.e., approach and models) described in this report is the culmination of research and development conducted over more than a decade. Moreover, reviews of previous TSPAs by internal and external professional organizations have been invaluable in enhancing the rigor of the approach and guiding the improvements of the models. Most notably, important advances in the TSPA methodology have been made in response to review comments from:

• Nuclear Waste Technical Review Board (NWTRB 1998; NWTRB 1999a; NWTRB 2000)

• Total System Performance Assessment Peer Review Panel (Budnitz et al. 1999)

• NRC staff (Paperiello 1999)

• State of Nevada consultants and other independent expert review groups.

The TSPA methodology used for this report is very similar to that used in the compliance certification application for the Waste Isolation Pilot Plant (DOE 1996a, Section 6.1; Helton, Anderson et al. 1999), a bedded salt repository in southern New Mexico. The Waste Isolation Pilot Plant was certified by the EPA in 1998 and began receiving and disposing of transuranic nuclear waste in March 1999. The TSPA approach is also similar to approaches adopted by other countries currently conducting detailed siting studies for potential geologic repositories (NEA 1991; Thompson 1999). In addition, the computer techniques used in TSPAs to address uncertainties are rooted in the probabilistic risk assessment method applied in the safety assessments for commercial nuclear reactors (Rechard 1999).

4.1.1.2 Treatment of Uncertainty in the Performance Assessment

Inherent uncertainties will exist in any projections of the future performance of a deep geologic repository. These uncertainties must be addressed in a way that is both clear and understandable to ensure technical credibility and sound decision-making and must be reduced or eliminated if important. Most, but not all, of those uncertainties can be quantified and addressed in the TSPA; examples include:

• Potential changes in climate, seismicity, and other processes, such as coupled thermal-hydrologic-chemical processes, over the compliance period for geologic disposal (i.e., 10,000 years)

• Variability and lack of knowledge of the properties of geologic media over large spatial scales of the hydrogeologic setting (e.g., the flow path from the repository to a point of compliance)

• Incomplete knowledge about the long-term material behavior of engineered components (e.g., corrosion of metals over many thousands of years).

Both the EPA and the NRC have recognized that uncertainty about the future performance of the repository will remain even after site characterization is complete. In the licensing context, both EPA regulations at 40 CFR 197.14 and NRC regulations at 10 CFR 63.304 (66 FR 55732) have incorporated "reasonable expectation" as the standard for the NRC to determine whether the DOE complies with EPA and NRC regulations. Characteristics of reasonable expectation include that it:

• Requires less than absolute proof because absolute proof is impossible to attain for disposal due to the uncertainty of projecting long-term performance

• Accounts for the inherently greater uncertainties in making long-term projections of the performance of the Yucca Mountain disposal system

• Does not exclude important parameters from assessments and analyses simply because they are difficult to precisely quantify to a high degree of confidence

• Focuses performance assessments and analyses upon the full range of defensible and reasonable parameter distributions rather than only upon extreme physical situations and parameter values.

There are a number of ways to accommodate or address uncertainty in analyses of performance. The methods employed for the TSPA models include:

• Definition of parameter probability distributions

• Representation of spatial and temporal variability

• Definition of data bounds or conservative estimates

• Formulation and evaluation of alternative models permitted by current state of knowledge

• FEPs screening (screening out certain aspects from consideration because of low probability or consequence).

Whether uncertainties are incorporated quantitatively through probabilistic distributions, or by developing "conservative" or "bounding" estimates, it is important to define and document assumptions on how uncertainties are treated in performance analyses. Clear documentation of the rationale for the assumptions in the models will enable others to understand and evaluate their adequacy. For the total system model to be transparent and defensible, the selection of each probability distribution must also be defensible. In some instances, the insufficient information that exists on the subsystem model is so complex that a probability distribution cannot be defined defensibly. In these instances, conservative or bounding approaches are taken.

Individual sources of uncertainties may be either quantified or unquantified. Quantified uncertainty consists of those sources of uncertainties that can be and have been explicitly (i.e., mathematically) represented and evaluated through a probabilistic performance analysis. Unquantified uncertainties are those that are recognized but are not well suited for direct evaluation through a probabilistic analysis.

As described in Section 4.4.1.2, many uncertainties have been quantified and incorporated directly into the TSPA models. The analyses done to address quantified uncertainties in the TSPA-SR model include a variety of sensitivity studies that address how total system performance might be affected if individual or groups of processes or parameters differed significantly from the way they were represented in the TSPA-SR model. These are described further in Section 4.4.5 and in Total System Performance Assessment for the Site Recommendation (CRWMS M&O 2000a). Additional uncertainties have been identified, analyzed, and quantified in FY01 Supplemental Science and Performance Analyses (BSC 2001a; BSC 2001b). These analyses are described, as appropriate, throughout Section 4.

As noted by the National Research Council (1990, p. 13) and others, there are residual uncertainties with deep geologic disposal that cannot easily be quantified and incorporated into performance analyses. Nevertheless, their potential impact must be, to the extent practicable, addressed and, if important, mitigated to provide confidence in postclosure performance. Examples of residual uncertainties associated with geologic disposal that are difficult to quantify include:

• The potential for currently unknown processes to affect performance

• The possibility that incompletely characterized processes have been incorporated in the TSPA in a manner that results in the underestimation of radionuclide releases; examples of incompletely characterized processes include thermal, chemical, hydrologic, and mechanical processes that are coupled in complex ways that cannot be completely tested at the scale of a repository, as well as processes that are difficult to observe and test, such as colloidal transport of radionuclides

• Uncertainty associated with the projections of engineered barrier performance over 10,000 years based on data from short-term (e.g., several years) laboratory testing

• Uncertainty associated with the large spatial scale of the three-dimensional groundwater flow system, which makes it difficult to characterize flow paths and processes.

The DOE has made a substantial effort to identify, characterize, and mitigate the potential impacts of residual uncertainties that could significantly affect long-term performance. Where practicable, additional tests have been conducted to collect information that would provide insight to analysts. For example, the DOE, along with Nye County and the National Park Service, is continuing research on the regional groundwater system. The program includes cooperative and joint analysis of soil, rock, and water samples provided by Nye County, which is drilling additional boreholes to characterize the saturated zone as part of its review of the project.

To address uncertainties associated with coupled thermal-hydrologic-geochemical processes, the DOE has performed numerous tests that were not envisioned when site characterization began. Where additional testing was not feasible (e.g., it is not possible to run tests over the same time period as the repository must perform) or of limited benefit (e.g., no amount of excavation or drilling could completely characterize the natural system), modelers used conservative assumptions to "bound" their analyses of uncertain processes. To do this, they have incorporated assumptions in their models that represent the observed range of properties and processes and that also include parameter values or processes likely to result in calculated dose assessments that are greater than actual dose. The DOE has also used empirical observations and qualitative lines of evidence from natural analogues to address uncertainties (see Section 4.1.2).

In some cases, more than one conceptual model may be consistent with available data and observations; if so, the analysis of uncertainty includes the identification of the basis for model selection. In the absence of definitive data sets or compelling technical arguments for any specific conceptual model, analyses or sensitivity studies may be performed to test whether the selection of a given model is likely to substantially affect analyses of system performance. When model uncertainty is unavoidable, analysts may develop simplified performance models for use in TSPA that reflect the range of outcomes predicted by more detailed and specific conceptual models. The selection of models, model parameters, and scenarios of potential future behavior are described in Total System Performance Assessment for the Site Recommendation (CRWMS M&O 2000a) and its supporting documents, as well as in FY01 Supplemental Science and Performance Analyses (BSC 2001a; BSC 2001b) and Total System Performance Assessment—Analyses for Disposal of Commercial and DOE Waste Inventories at Yucca Mountain—Input to Final Environmental Impact Statement and Site Suitability Evaluation (Williams 2001a).

In addition, consistent with the postclosure safety case described in Section 4.1.3, the DOE has implemented a safety margin/defense-in-depth approach to the design of repository facilities. This approach mitigates some of the residual uncertainties by providing additional confidence in the performance of the total system. The iterative approach to characterization and testing, design, and performance analysis remains a fundamental part of the DOE approach in evaluating whether Yucca Mountain is a safe site to host a repository.

One disadvantage of an approach that combines both conservative and representative parameter distributions is that it is difficult to assess the extent to which the total system results are conservative or realistic. More realistic representations may be drawn from literature data, analogue systems or processes, and the technical judgment of the broader scientific and engineering community. In addition, the DOE initiated several activities to improve the treatment of uncertainty in current models. These are described in FY01 Supplemental Science and Performance Analyses (BSC 2001a; BSC 2001b). These activities:

• Identified and evaluated the degree of conservatism introduced by current approaches, quantified key unquantified uncertainties, and characterized the effect of explicitly quantifying ranges of possible parameters

• Developed alternative representations of key unquantified uncertainties and evaluated the impact on TSPA results

• Considered the uncertainties associated with alternative models into which assumptions or conservatisms were introduced when they were abstracted for use in the TSPA, in addition to assumptions and conservatisms in the process model

• Developed more realistic representations of models and parameters, based on all available information and the scientific judgment of individuals who are experts in the appropriate disciplines

• Developed, for each model in the TSPA, alternate lines of evidence, such as additional analogues and defense in depth, that are less dependent on the TSPA computational tool than those given in Repository Safety Strategy: Plan to Prepare the Safety Case to Support Yucca Mountain Site Recommendation and Licensing Considerations (CRWMS M&O 2001a).

This approach provides recommendations for consistent treatment of uncertainties, which will lead to a more rigorous treatment of the different types of uncertainty. It will also lead to a more transparent description of uncertainties in the process models and in the overall system models.

4.1.1.3 Explicit Consideration of Disruptive Processes and Events that Could Affect Repository Performance

Most of this report, as well as the Total System Performance Assessment for the Site Recommendation (CRWMS M&O 2000a), the FY01 Supplemental Science and Performance Analyses (BSC 2001a; BSC 2001b), the Total System Performance Assessment—Analyses for Disposal of Commercial and DOE Waste Inventories at Yucca Mountain—Input to Final Environmental Impact Statement and Site Suitability Evaluation (Williams 2001a), and associated documentation, focuses on the processes that are expected to operate in and near the repository over time (e.g., water flow, the effects of heat on conditions near waste packages, and the slow degradation of the engineered barrier system). Because of the time frames involved, however, it is also important for analysts to understand the safety consequences of unexpected events or behavior. Consequently, the performance analysis also describes explicitly how disruptive events (i.e., possible but unlikely events that could negatively affect performance) and alternative models of processes could affect the performance of the total system. This thorough consideration of what could go wrong, how wrong the models could be, and what the effect of inaccuracy in the models would be is a key element of the postclosure safety case.

A comprehensive set of potentially disruptive events, ranging from meteor or comet impacts to unexpected flooding of the repository, has been identified and evaluated. Similarly, a wide variety of potentially harmful processes, such as unanticipated failure of the waste packages and damage to repository systems by seismic activity, have been identified and evaluated. Section 4.3 describes the identification and screening of these potentially adverse processes and events. Depending on the results of this analysis, the events and processes have been treated in one of three ways:

1. Events or processes with very low probabilities (e.g., less than 1 chance in 10,000 of occurring in 10,000 years, such as meteor impacts) or very small consequences (e.g., potential radionuclide releases much lower than regulatory standards) have been screened out and not analyzed further.

2. Events or processes that have probabilities greater than 1 chance in 10,000 of occurring in 10,000 years (e.g., intrusive igneous activity), have been included in the performance assessment probabilistically—that is, the consequence of each event is weighted by its probability of occurrence.

3. Events or processes expected to occur during the period of regulatory compliance, such as climate change or ground shaking associated with earthquakes, are included directly in models of the performance of the repository.

The potential consequences of each event and process considered in the performance assessment models are analyzed. These analyses are summarized in Section 4.2, which considers alternative models of subsystem processes that could affect performance, and Section 4.3, which considers specific events or processes that are not part of nominal site behavior. Section 4.4.5 also describes a number of sensitivity studies that enable analysts to assess the importance of alternative models or unexpected events on total system performance.

Many different events and processes have been considered to assess whether they might have the potential to disrupt repository performance. Examples of potential events considered include:

• The potential for future volcanism near the repository

• The potential for future seismic activity near the repository

• The potential for future human intrusion into the repository

• The potential for climate change in the region of the repository

• The potential for criticality to occur during the postclosure period.

Climate change and cladding damage caused by earthquake ground motion is directly incorporated into the nominal case performance analyses (CRWMS M&O 2000c) because climate change and seismic activity are expected to occur during the period of regulatory compliance.

Many other events and processes have been considered but screened out because their impacts are so inconsequential that inclusion in the TSPA-SR is not warranted. Examples include the potential for future rises in the water table that could inundate the repository and the potential for nuclear criticality in spent nuclear fuel after repository closure. These examples are discussed in Section 4.3.

It is not possible to foresee every event or process that could affect the potential repository. However, by recognizing and explicitly analyzing all identified events and processes that could affect repository performance, the DOE has provided a sound basis for evaluating the performance of the repository system.

4.1.2 Observations from Natural and Man-Made Analogues

An alternative means of analyzing the reliability of repository performance models is by comparing them with natural or anthropogenic (man-made) analogues. As defined in this report, analogues are systems in which processes similar to those that could occur in a nuclear waste repository have occurred over long time periods (decades to millennia) and large spatial scales (up to tens of kilometers) not suited to laboratory or field experiments. The concept of geologic disposal is based on an analogue observation that certain geologic environments have intrinsic properties that contribute to the isolation of waste and will continue to do so for a long time. Analogue studies can also help scientists understand how specific repository systems might behave by allowing them to compare possible future behavior with the known past behavior of an analogue system. A variety of specific analogue sites are described throughout Section 4.2, including several in settings that provide insight into the flow of water in the unsaturated zone, as described in Section 4.2.1.2.13.

For example, observations of the hydrologic behavior of ancient man-made tunnels and natural caves can provide relevant information about water seepage into mined openings in an unsaturated zone over thousands of years. Similarly, observations of the past migration of radioactive contaminants in groundwater in similar environments can provide insight into the possible future transport of radionuclides away from a repository. Such information can be obtained from both natural analogues (e.g., natural deposits of uranium and other minerals) or from anthropogenic analogues (e.g., the Nevada Test Site, Hanford, or the Idaho National Engineering and Environmental Laboratory), where the movement of radionuclides in groundwater caused by past releases is currently being monitored. The archaeological and historical record can also provide qualitative information on the degradation of materials that may be relevant to the performance of the repository (e.g., the preservation of materials in Egyptian pyramids and tombs more than 5,000 years old).

The value of natural analogues is not restricted to locations that may mimic aspects of repository behavior. The study of natural analogues is an intrinsic part of scientific studies, particularly in the earth sciences. For example, in order to undertake studies of basaltic volcanism in the vicinity of the Yucca Mountain site, an investigator must be versed in basaltic volcanism and especially in basaltic volcanism in the southern Great Basin. All other known occurrences of basaltic volcanism become, to some degree, natural analogues for volcanism near Yucca Mountain. Each occurrence can tell the investigator something about the mechanisms and controls on volcanic episodes. The overall understanding of volcanic processes gained from regional studies provides a basis for analyzing trends and patterns in the Yucca Mountain area that are essential to evaluating the possibility that future volcanic events could occur and affect a repository. In fact, it would be impossible to understand basaltic volcanism as a site characterization issue without recourse to natural analogues. One of the fundamental tasks for the investigator is to recognize and appreciate the value of the information offered by this kind of analogue study.

The scientific community has endorsed the use of analogues as a means of assessing the potential future performance of systems, components, and processes related to nuclear waste disposal (Chapman and Smellie 1986). The National Academy of Sciences/National Research Council (National Research Council 1990) and the NRC (10 CFR 63.101(a)(2) [66 FR 55732]) have also encouraged analogue studies.

There are no close analogues to a total repository system at Yucca Mountain. Nevertheless, studies of a variety of analogues can and have been used to assess how well repository models represent processes known to be important to performance, as well as the magnitude and duration of the phenomena. Analogue information has also been used (1) to evaluate the validity of extrapolating from short-term field-scale experiments to longer time scales in which field-scale experiments are impractical and (2) to add confidence when spatially extrapolating processes studied at laboratory and intermediate-scale experiments to tests at larger spatial scales. Knowledge gained from natural analogues has helped refine performance assessment model assumptions and parameter ranges and has improved the robustness and consistency of process models.

Given the imprecise nature of the information gained from investigations of similar, but not identical, processes and sites, analogue studies alone cannot prove that process or total system performance models are valid in a strict sense. However, natural analogue observations can confirm that a model takes into account the relevant processes in appropriate ways. In this way, the analogues can build confidence in models of future behavior. This is consistent with the expectations of NRC regulations in 10 CFR 63.101(a)(2) (66 FR 55732), which state: "Demonstrating compliance will involve the use of complex predictive models that are supported by limited data from field and laboratory tests, site-specific monitoring, and natural analog studies that may be supplemented by prevalent expert judgment."

Throughout this report and its supporting documents, numerous analogues are analyzed to provide information on processes that may affect both engineered and natural system features of a geologic repository at Yucca Mountain. Specific examples of relevant analogues are presented in Table 4-1. Additional discussion of analogues is provided as appropriate throughout Sections 4.2 and 4.3, which discuss in greater detail the understanding of the Yucca Mountain site. Although the direct applicability of the analogues for each process model varies, the analogue observations generally suggest that the conceptual and numerical models that form the basis for analyses of repository performance are reasonable to conservative. For many of the analogues, a large body of literature exists.

4.1.3 Use of Defense in Depth and Safety Margin to Increase Confidence in System Performance

The extensive testing program at Yucca Mountain and the thorough assessments of the future performance of the potential repository do not "prove," in the usual sense of the word, that the potential repository will be safe. To provide additional assurance of long-term safety, the third major element of the postclosure safety case relies on a complementary, but less analytical, approach that is based on engineering principles that have a proven track record for safety. These principles are known as "safety margin" (or design margin) and "defense in depth."

Safety margin refers to the standard engineering practice of including safety factors on the performance of engineered components to account for uncertainty and variability in material, fabrication, and use. These safety factors are typically expressed as a ratio of the calculated level of performance to an allowable or laboratory measured level of performance. They are developed to ensure that the component or system has ample reserve performance capability. A simple example of a safety factor would be the limiting of the stress in a component to a fraction of what would cause failure. The safety margin is then the reserve strength over and above what would actually be applied. A Yucca Mountain-specific example of safety margin is the design and selection of the waste package material. The assumed corrosion resistance of Alloy 22, used in the outer shell of the waste packages, was decreased (or the corrosion rate was increased) over the laboratory values to account for potential environmental conditions. The corrosion rate of Alloy 22 was increased by 2.5 times to address potential heat-accelerated corrosion and an additional 2.0 times to address potential microbial corrosion.

The defense-in-depth approach (or reliance on multiple system attributes) complements design margin in that it provides a method of ensuring overall performance if one or more components of the repository system fail to perform as expected. Defense in depth is provided by having safety components that do not share common failure modes. In other words, the processes or conditions (such as the geochemical environment) that might cause a degradation of performance of one component of the design will not similarly affect other components. In a repository, defense in depth is provided by the attributes of both the natural barriers and the engineered barrier system.

The safety margin/defense-in-depth approach is not specifically required by the regulations for a repository. However, the NRC regulation (10 CFR Part 63 [66 FR 55732]) does contain statements that are based on a similar philosophy, adapted to the long-term requirements for postclosure safety. 10 CFR 63.113(a) provides that a repository include multiple barriers, consisting of both natural barriers and an engineered barrier system.

At Yucca Mountain, the potential geologic repository system would contain several different barriers. As defined in 10 CFR 63.2 (66 FR 55732), a barrier is any material, structure, or feature that prevents or substantially reduces the rate of movement of water or radioactive material from the Yucca Mountain repository to the accessible environment. Table 4-2 presents a summary of the natural and engineered barriers present at Yucca Mountain, along with a brief description of their intrinsic and intended functions for the design analyzed in this section. An analysis of each barrier's contribution to performance is presented in Section 4.5.

Implementation of the safety margin/defense-in-depth approach has resulted in several improvements in the repository design since the design described in 1998 in the Viability Assessment (VA) (DOE 1998). Examples include:

• Modification of the waste package design to place the corrosion-resistant Alloy 22 layer on the outside. Having the corrosion-resistant layer outside provides sufficient margin that performance assessments indicate complete containment (in the absence of unlikely disruptive processes) of nuclear wastes throughout the period of regulatory compliance.

• Evaluating thermal loading strategies that may reduce the complexity of coupled processes in the near-field zone and simultaneously maximizes the redirection of water away from the repository drifts.

• Adding drip shields to the engineered barrier design to minimize the potential for dripping water to contact the waste packages (and waste form).

More detailed descriptions of the barriers and their performance functions are summarized in Section 4.5. Additional information is available in Volume 2 of Repository Safety Strategy: Plan to Prepare the Safety Case to Support Yucca Mountain Site Recommendation and Licensing Considerations (CRWMS M&O 2001a) and Total System Performance Assessment for the Site Recommendation (CRWMS M&O 2000a) as well as FY01 Supplemental Science and Performance Analyses (BSC 2001a; BSC 2001b) and Total System Performance Assessment—Analyses for Disposal of Commercial and DOE Waste Inventories at Yucca Mountain—Input to Final Environmental Impact Statement and Site Suitability Evaluation (Williams 2001a).

4.1.4 Mitigation of Uncertainties by Selection of a Thermal Operating Mode

One way of mitigating the uncertainties in modeling long-term repository performance is to operate the repository so the temperature of the host rock stays below the boiling point of water. Uncertainties in thermally driven processes are of special interest because of their complexity and because the current modeling approach may mask the importance of thermal effects on performance. Two key uncertainties about thermal effects on potential repository performance are (1) the way coupled processes in the mountain will respond to the heat generated by emplaced waste and (2) the long-term performance of waste package materials in the potential repository environment.

In the models and design described in this report, uncertainties related to the higher-temperature operating mode have been recognized and addressed. Since the VA, the design has evolved to include a thermal management strategy that limits the region of rock with temperatures above the boiling point of water, along with other features (such as drip shields) that mitigates uncertainty. Current models attempt to capture the remaining uncertainties well enough to understand their impacts; however, the DOE has considered additional options for mitigation. In particular, the performance characteristics of lower-temperature operating mode concepts such as those described in Section 2.1 have been investigated in FY01 Supplemental Science and Performance Analyses (BSC 2001a; BSC 2001b) and Total System Performance Assessment—Analyses for Disposal of Commercial and DOE Waste Inventories at Yucca Mountain—Input to Final Environmental Impact Statement and Site Suitability Evaluation (Williams 2001a).

Because uncertainties due to thermally induced coupled processes cannot be eliminated through additional testing, the DOE's approach is to consider options for mitigating thermal uncertainties by lowering temperatures in the emplacement drifts and on the waste package surface. Keeping the host rock temperature below boiling may reduce uncertainties associated with coupled processes (Anderson et al. 1998, p. 18; Cohon 1999). In addition, uncertainties in localized corrosion rates may be mitigated by avoiding the conservatively defined window of increased susceptibility by keeping the temperature of the waste package below 85°C (185°F) or the relative humidity in the emplacement drifts below 50 percent. Section 2.1 provides additional discussions of lower-temperature operating modes.

4.1.5 Performance Confirmation, Postclosure Monitoring, and Site Stewardship

The EPA, NRC, and DOE have recognized that some uncertainty about repository performance cannot be eliminated. Furthermore, the DOE understands that ensuring public safety requires continued site stewardship, including a program for evaluating new information discovered during the construction and operation phases. Therefore, the final component of the postclosure safety case is a program of performance confirmation, monitoring, and site stewardship that accomplishes several goals related to the DOE's obligation to protect public health and safety and the environment. The program addresses 10 CFR Part 63 (66 FR 55732) Subpart F provisions for performance confirmation to ensure consistency with license specifications after waste emplacement and before permanent closure. The program also includes activities necessary for the DOE to provide postclosure oversight, as specified by Section 801(c) of the Energy Policy Act of 1992 (Public Law No. 102-486), and post-permanent closure monitoring, consistent with 10 CFR 63.51(a)(2). Specifically, the DOE will continue to observe and test the performance of the repository during and after waste emplacement and will maintain the integrity and security of the repository through a variety of institutional controls. The DOE will also continue to participate in research on geologic disposal, in cooperation with other international programs. These activities will ensure that any new information discovered at Yucca Mountain (or elsewhere) that is relevant to future repository decisions is considered appropriately.

The performance confirmation program is the most important monitoring activity. NRC regulations provide for performance confirmation to continue for at least 50 years after the initiation of waste emplacement. The DOE will continue its performance confirmation program until the repository is permanently closed. The amount of information collected during this period may be more relevant for long-term analyses of the repository than any experiment or test that could be conducted now or in the near future. Any decision to close the repository would be based on the increased understanding and confidence derived from decades of testing and observation.

The performance confirmation program will provide data on the actual performance of the key natural and engineered systems and components of the repository as conditions evolve. The program will also provide data to confirm (after repository construction and operation) that subsurface conditions encountered, and any changes in those conditions during repository construction and waste emplacement, are consistent with the expected performance of the repository. A primary goal of the program will be to confirm, through observation, monitoring, and analysis, that the repository is performing in a manner that will contain and isolate waste.

As described in Section 4.6, the performance confirmation program will monitor the processes important to future waste isolation in the repository. Examples include the flow of water past and near the repository, the geochemical environment in and near emplaced waste, coupled thermal-hydrologic-chemical-mechanical processes, and the performance characteristics of engineered materials in the repository environment (e.g., drip shield and waste package degradation). In addition to technical monitoring of the performance of the site, the DOE will maintain the security and integrity of the site throughout the performance confirmation period and beyond as required by the Energy Policy Act of 1992, Section 801(c) (Public Law No. 102-486). A program will be developed to prevent any human activity, including deliberate or inadvertent human intrusion, that could affect engineered or geologic barriers.

Section 122 of the NWPA (42 U.S.C. 10142) requires the DOE to maintain the ability to retrieve any and all emplaced wastes "for any reason pertaining to the public health and safety, or the environment, or for the purpose of permitting the recovery of the economically valuable contents of such spent fuel" prior to closure. For example, if it were learned from the monitoring program that an engineered barrier had been damaged, the waste packages could be removed and repairs could be made, as necessary.

NRC regulations (10 CFR Part 63 [66 FR 55732]) anticipate that the repository could be closed as early as 50 years after initial waste receipt. Closing the repository would involve the sealing of shafts, ramps, exploratory boreholes, and other underground openings. These actions would discourage any human intrusion into the repository and prevent water from entering through these openings. If a decision to close the repository were made, the DOE would still be responsible for a program of site stewardship under the Energy Policy Act of 1992 (Public Law No. 102-486).

At the surface, all radiological areas would be decontaminated, all structures removed, and all wastes and debris disposed at approved sites. All disturbed areas would be restored as close as practicable to their preconstruction condition.

NRC regulations (10 CFR Part 63 [66 FR 55732]) require the DOE to submit a plan for postclosure monitoring with any application to close the repository. The DOE has also committed to maintain security and continue monitoring at the Nevada Test Site for the foreseeable future. A network of permanent monuments and markers would be erected around the site to warn future generations of the presence and nature of the buried waste. Detailed records of the repository would be placed in the archives and land records of local, state, and federal government agencies and archives elsewhere in the world that future generations would be likely to consult. These records would identify the location and layout of the repository and the nature and hazard of the waste it contains.

4.2 DESCRIPTION OF SITE CHARACTERIZATION DATA AND ANALYSES RELATED TO POSTCLOSURE SAFETY

This section discusses the data obtained during site characterization activities, as well as analyses of the safety of a potential Yucca Mountain repository. The DOE planned and conducted its site characterization program to collect data about the site and about those physical and chemical processes that would affect the ability of the repository system to isolate waste.

Section 1.3 presented a brief summary of the geology of Yucca Mountain based on the results of site investigations. It provided a framework for the descriptions of the repository and waste package designs contained in Sections 2 and 3. In this section, the results of studies focused on the characteristics and potential future behavior of the repository system are presented in additional detail. The discussion is organized to provide a description of the major processes that control the waste isolation capability of the potential repository. As shown schematically in Figure 4-2, Section 4.2 describes in sequence the data and analyses relevant to the processes that affect the movement of water through Yucca Mountain and relevant to the potential for that water to contact and mobilize radionuclides. Disruptive events could potentially affect these processes and, therefore, also need to be considered. The data and analyses related to potential disruptive events are presented in Section 4.3, and the combined analysis of the potential performance of the repository is summarized in Section 4.4.

The processes pertinent to performance include those physical processes that control the movement of water, beginning with precipitation as rain and snow at the surface, followed by infiltration into the mountain, flow through the unsaturated zone to the potential repository level, flow from the repository level to the saturated zone, and from there to the accessible environment. At the repository level, water moving past the engineered barriers would be affected by the physical and chemical processes associated with the decay heat and could interact with waste packages and waste forms. These processes could lead to the mobilization of radionuclides. Eventually, the water could move out of the repository horizon and further downward through the unsaturated zone. Subsequently, it could move into the saturated zone, where it could be transported to the accessible environment where humans could be exposed.

The data collected during site characterization have been used to develop conceptual and numerical models of the hydrologic, geochemical, thermal, and mechanical processes that will determine how a repository at Yucca Mountain may behave over the next 10,000 years. These process models have, in turn, been used to develop a TSPA model that has been used to assess quantitatively the potential for radionuclide releases to the public and, consequently, the safety of the Yucca Mountain site.

Attributes Important to Long-Term Performance—The potential repository system can be described in terms of five key attributes that would be important to long-term performance: (1) limited water entering waste emplacement drifts; (2) long-lived waste package and drip shield; (3) limited release of radionuclides from the engineered barriers; (4) delay and dilution of radionuclide concentrations by the natural barriers; and (5) low peak mean annual dose considering potentially disruptive events. These attributes are summarized below. The first four reflect the interactions of natural barriers and the engineered barriers in prolonging the containment of radionuclides within the repository and limiting their release. The fifth attribute reflects the likelihood that disruptive events would not affect repository performance over 10,000 years.

1. Limited Water Entering Emplacement Drifts—The climate at Yucca Mountain is dry and arid, with precipitation averaging about 190 mm (7.5 in.) per year. Little of this precipitation percolates into the mountain; nearly all of it (above 95 percent) either runs off or is lost to evaporation, limiting the amount of water available to seep into emplacement drifts. For the higher-temperature operating mode described in this report, a thermal management strategy was developed that would take advantage of the heat of the emplaced wastes to drive water away from the emplacement drifts. The heat generated by the waste would dry out the rock surrounding the drift and decrease the amount of water available to contact the waste packages until the wastes have cooled substantially. Drainage of water in the rock pillars between drifts would be encouraged by keeping much of the pillar rock between the drifts below the boiling temperature of water. As long as the drift walls remained at temperatures above the boiling point of water, there would no liquid water in the emplacement drifts and very little in the nearby rock. Even after the drift walls cooled below the boiling point of water, the residual heat would increase evaporation in and near emplacement drifts, thereby continuing to limit the amount of water present in the rocks adjacent to waste packages. In lower-temperature operating modes, the waste packages would be exposed to water earlier. Because the rock would eventually cool in any operating mode, there does not appear to be a significant difference in the amount of water to which the waste packages would eventually be exposed.
   Over the range of operating modes, the design also takes advantage of the mechanical and hydrologic processes that divert water around emplacement drift openings in the unsaturated zone. Because of capillary forces, water flowing in narrow fractures tends to remain in the fractures rather than flow into large openings, such as drifts. If any water reaches an emplacement drift, it could flow down the drift wall to the floor and drain without contacting the drip shield or the waste packages (however, in taking a more conservative approach in modeling, this potential process is not included in the models). Thus, the natural and engineered features of a repository in the unsaturated zone will combine to limit the potential for water to enter the emplacement drifts.

2. Long-Lived Waste Package and Drip Shield—To further reduce the possibility of water contacting waste, the DOE has designed a robust, dual-wall waste package with an outer cylinder of corrosion-resistant material, Alloy 22. Alloy 22 was selected because it will remain stable in the geochemical environment that would be expected in a repository at Yucca Mountain. In the operating mode described in this report, the repository environment would be warm, with temperatures at the waste package surface initially rising above the boiling point of water. Waste package surface temperatures are expected to gradually decrease to below boiling after a period of hundreds to thousands of years, depending on the waste package's location within the repository layout, spatial variation in the infiltration of water at the ground surface, and variability in the heat output of individual waste packages.
   Chemically, the environment is expected to be near-neutral pH (mildly acidic to mildly alkaline) and mildly oxidizing. Because most corrosion would occur only in the presence of water and because highly corrosive chemical conditions are not expected, both the titanium drip shield and the Alloy 22 outer barrier of the waste package are expected to have long lifetimes in an unsaturated zone repository. In lower-temperature operating modes, the waste packages would be exposed to water earlier.

3. Limited Release of Radionuclides from the Engineered Barriers—Because of the characteristics of the natural system, the drip shields, and the waste packages, scientists do not expect water to come into contact with the waste forms for over 10,000 years. Even if water were to penetrate a breached waste package before 10,000 years, several characteristics of the waste form and the repository would limit radionuclide releases. First, because of the warm temperatures, much of the water that penetrates the waste package will evaporate before it can dissolve or transport radionuclides. Neither spent nuclear fuel nor glass waste forms will dissolve rapidly in the expected repository environment. Further, a large majority of the radionuclide inventory is insoluble in the geochemical environment expected within the repository. Although the performance of the cladding (metallic outer sheath of a fuel element) as a barrier may vary because of possible degradation, it is expected to limit contact between water and waste. The component of the engineered barrier system below the waste package, called the invert, contains crushed tuff that would also limit the transport of radionuclides into the host rock.

4. Delay and Dilution of Radionuclide Concentrations by the Natural Barriers—Eventually, the engineered barrier systems in the repository are expected to degrade, and small amounts of water may contact waste. Even then, features of the geologic environment and the repository system will help decrease radionuclide migration to the accessible environment and slow it by hundreds to thousands of years. Processes that could be important to the movement of radionuclides include sorption, matrix diffusion, dispersion, and dilution. Rock units in both the unsaturated zone and the saturated zone at Yucca Mountain contain minerals that can adsorb many types of radionuclides (i.e., radionuclides would attach to and collect on the mineral surfaces). As water flows through fractures, dissolved radionuclides can diffuse into and out of the pores of the rock matrix, increasing both the time it takes for radionuclides to move from the repository and the likelihood that radionuclides will be exposed to sorbing minerals. Dispersion and dilution will occur naturally as potentially contaminated groundwater flows and mixes with other groundwater and lowers the concentration of contaminants.

5. Low Mean Annual Dose Considering Potentially Disruptive Events—Yucca Mountain provides an environment in which hydrogeologic conditions important to waste isolation (e.g., a thick unsaturated zone with low rates of water movement) have not changed very much for at least hundreds of thousands of years. Analysts have identified and evaluated a wide variety of potentially disruptive processes and events that could affect the performance of the design and operating mode described in this report. These range from extremely unlikely events, such as meteor impacts, to events that are likely to occur, such as regional climate change. Although the probability of volcanic activity in or near the potential repository is low, volcanic activity was a consideration in TSPA in the disruptive scenario case. Performance assessment results to date show that potentially disruptive events are not likely to compromise the performance of the repository, and the probability-weighted mean dose for an igneous disruption is low.

Table 4-3 shows the relationship between the key attributes of the site and the physical barriers that comprise the repository system, the processes important to waste isolation, and the descriptions presented in Sections 4.2.1 through 4.2.10 [4.2.1, 4.2.2, 4.2.3, 4.2.4, 4.2.5, 4.2.6, 4.2.7, 4.2.8, 4.2.9, 4.2.10]and in Section 4.3. The table is based on similar information developed and presented in Repository Safety Strategy: Plan to Prepare the Safety Case to Support Yucca Mountain Site Recommendation and Licensing Considerations (CRWMS M&O 2001a) and the Total System Performance Assessment for the Site Recommendation (CRWMS M&O 2000a). The table incorporates the results of lower-temperature operating mode evaluations (BSC 2001b). In addition to the information presented in these sources, Table 4-3 relates the key attributes and processes to the natural and engineered barriers that would contribute to waste isolation at Yucca Mountain. The last column of the table also lists process model reports that contain more detailed descriptions of the processes summarized here and included in the TSPA. The column showing the processes important to repository performance is not intended to be comprehensive: it is presented at a summary level. More detailed definitions of the proposed repository system and relevant specific processes are included in the individual sections that follow.

Organization of Section 4.2—As organized below, and shown in Figure 4-2, the total system consists of numerous interdependent subsystems that are described in ten subsections, as follows:

• Section 4.2.1 describes the data and analyses relevant to the movement of water in the unsaturated zone and describes the development of process models related to the potential movement of water into the repository.

• Section 4.2.2 describes the data and analyses related to the effects of decay heat on the movement of water through the unsaturated zone.

• Section 4.2.3 describes information that provides the basis for the DOE's understanding of the physical and chemical environment within the repository drifts, which influences the expected lifetimes of the drip shields and waste packages.

• Section 4.2.4 summarizes the data and evaluations that support models of the degradation of drip shields and waste packages, which will affect the lifetimes of the engineered barrier system.

• Section 4.2.5 describes the information that supports the analysis of the performance of engineered barriers in diverting water away from the waste package. After the eventual degradation of the waste packages, and in the presence of liquid water, it would be possible for chemical reactions to occur to mobilize radionuclides within the waste packages.

• Section 4.2.6 describes the experimental data and analyses of the chemical and hydrologic processes within the breached waste packages that would result in the release of radionuclides from the waste form.

• Section 4.2.7 describes information related to the potential transport of mobilized radionuclides through the engineered barriers.

• Section 4.2.8 describes the data and analyses related to the potential flow and transport of radionuclides through the unsaturated zone.

• Section 4.2.9 describes potential flow and transport in the saturated zone.

• Section 4.2.10 describes the biosphere model used to model the uptake and biological consequences of released radionuclides that eventually reach the human environment.

Within each of these sections, a similar internal organization is used to present the DOE's understanding of the key processes and to present the data and analyses that provide the basis for that understanding. Each discussion includes:

1. A summary description of how each process would operate in a potential repository

2. A description of the investigations, tests, and other data (including analogue information) that provide the technical basis for the DOE's understanding

3. A description of the conceptual and numerical models that have been developed to allow the DOE to assess potential future performance, including specific information about the sources and treatment of uncertainties, alternative models, and model calibration and validation

4. A description of how the conceptual and numerical models have been abstracted (or represented) in the TSPA-SR (CRWMS M&O 2000a).

4.2.1 Unsaturated Zone Flow

This section summarizes the current understanding of water movement (i.e., percolation flux) through the unsaturated zone and into a repository (i.e., seepage into drifts) at Yucca Mountain. Fluid flow through the unsaturated zone at Yucca Mountain is described at length in Unsaturated Zone Flow and Transport Model Process Model Report (CRWMS M&O 2000c, Sections 3.2 to 3.4 and 3.6 to 3.9), which is supported by 24 detailed analysis model reports. Yucca Mountain Site Description (CRWMS M&O 2000b, Sections 8.2 to 8.10) also provides a comprehensive summary of investigations performed to characterize flow and seepage in the unsaturated zone. Figure 4-3 shows the relationships between the main unsaturated zone processes, with those relevant to unsaturated zone flow highlighted.

The primary purpose of Unsaturated Zone Flow and Transport Model Process Model Report (CRWMS M&O 2000c) is to develop models for the TSPA that evaluate the postclosure performance of the unsaturated zone. Unsaturated Zone Flow and Transport Model Process Model Report supplies to the TSPA (1) ambient and predicted (i.e., future) three-dimensional flow fields based on different climate states and infiltration scenarios and (2) seepage rates into potential waste emplacement drifts. The flow fields are used directly in TSPA calculations of transport, while the seepage rates are used to calculate distributions of the fraction of waste packages in contact with seepage water and the volumetric flow rate to a waste package segment, taking into account possible flow focusing effects from site scale to drift scale (see Section 4.2.1.4).

As noted in Section 4.1.4, the DOE is evaluating the possibility of mitigating uncertainties in modeling long-term repository performance by operating the design described in this report at lower temperatures. Some analyses described in this section have been updated and expanded to capture the features and processes relevant to operating the design at lower temperatures. The updated analyses are described in Volume 2, Section 4 of FY01 Supplemental Science and Performance Analyses (BSC 2001b).

4.2.1.1 Conceptual Basis

On the most fundamental level, the important factors affecting unsaturated groundwater flow at Yucca Mountain are climate and rock hydrologic properties. Derived from these two basic components are estimates of percolation flux and seepage into potential waste emplacement drifts, both of which are key unsaturated zone processes.

Located in southern Nevada, one of the most arid regions of the United States, Yucca Mountain is underlain by a thick unsaturated zone (CRWMS M&O 2000c, Section 3.2.1). A dry climate and a deep water table are considered favorable characteristics for waste isolation. In a desert environment, the total amount of available water is small. The potential repository would be designed to complement the hydrologic environment by diverting the small flow of water that does occur away from the waste packages. Multiple natural and engineered barriers are expected to limit contact between water and waste forms, and retard radionuclide migration. The climate data and unsaturated zone characteristics are discussed in Sections 4.2.1.2.1 and 4.2.1.2.2, respectively.

The major components of the unsaturated zone and the emplacement drifts that affect water movement are illustrated in Figure 4-4. Water movement starts with rainfall in the arid environment, which is subject to runoff, evaporation, and plant uptake, such that much of the rainfall never reaches the potential repository host rock units. The infiltration of water that penetrates into the rock units of the unsaturated zone is redistributed by flow processes in the fractured and faulted, welded and nonwelded tuff layers. When percolating water encounters an underground drift, much of it will be diverted by capillary barrier mechanisms around the opening and never contact the engineered barriers within.

Major issues related to unsaturated zone flow processes include:

• Climate and infiltration

• Fracture flow versus matrix flow within major hydrogeologic units
  • Flow above the potential repository

  • Flow below the potential repository, including the formation and hydrologic significance of perched water

• Fracture–matrix interaction

• Effects of major faults

• Seepage into drifts.

Using these major issues as subheadings, a summary of the current conceptual understanding of unsaturated zone flow beneath Yucca Mountain is first presented. Following this summary are subsections that describe the data supporting the conceptual understanding of flow in the unsaturated zone (Section 4.2.1.2), the development and integration of numerical process models based on the conceptual interpretation of the available data (Section 4.2.1.3), and how the results of unsaturated zone numerical modeling studies have been abstracted for the TSPA (Section 4.2.1.4).

4.2.1.1.1 Climate and Infiltration

Climate is defined by the variation of meteorological conditions (such as temperature, pressure, humidity, precipitation rate and prevailing winds) over time. At Yucca Mountain, climate is important because it provides the boundary conditions for the hydrologic system—specifically, the amount of water available at the surface. Estimates of the precipitation rate and temperature taken from climate models have been used as information to determine the net infiltration of water into Yucca Mountain and the percolation flux at the repository horizon. Percolation flux within the unsaturated zone, governed by climate and rock hydrogeologic properties, is a key process affecting seepage in waste emplacement drifts and transport of radioactive particles below the repository. Three representative climates are forecast to occur within the next 10,000 years (i.e., the period of regulatory compliance): the modern (present-day) climate, a monsoon climate, and a glacial-transition climate (USGS 2000a, Section 6.6). Beyond 10,000 years, the TSPA-SR extends the glacial-transition climate for base-case simulations and includes a revised long-term climate model in a sensitivity study, as discussed in Section 4.4.2.4 (CRWMS M&O 2000a, Section 3.2.1.2) and Volume 1, Section 3.3.1 of FY01 Supplemental Science and Performance Analyses (BSC 2001a).

Net infiltration refers to the penetration of liquid water through the ground surface and to a depth where it can no longer be removed by evaporation or transpiration by plants. Net infiltration is the source of groundwater recharge and water percolation at the potential repository horizon; it provides the water for flow and transport mechanisms that could move radionuclides from the potential repository to the water table. The overall framework of the conceptual model for net infiltration is based on the hydrologic cycle. Important processes that affect net infiltration include precipitation (rain and snow), runoff and run-on (flow of surface water off one place and onto another), evaporation, transpiration, and moisture redistribution by flow in the shallow subsurface (USGS 2000b, Section 6.1.3). Infiltration is temporally and spatially variable because of the nature of the storm events that supply precipitation and because of variation in soil cover and topography (CRWMS M&O 2000bq, Section 6.1.1). Surficial soils and topography are considered part of the natural barrier system because they reduce the amount of water entering the unsaturated zone by surficial processes (e.g., precipitation lost to runoff, evaporation, and plant transpiration) (Table 4-2). Net infiltration rates are believed to be high on sideslopes and ridge tops where bedrock is exposed and fracture flow in the bedrock is able to move liquid water away from zones of active evaporation (Flint, L.E. and Flint 1995, p. 15).

Within the limits of extant monitoring data, significant net infiltration occurs only every few years under the present climate (CRWMS M&O 2000bq, Section 6.1.1). In these years, the amount of net infiltration still varies greatly, depending on storm amplitude, duration, and frequency. In very wet years, net infiltration pulses into Yucca Mountain may occur over a period of a few hours to a few days. A detailed discussion of net infiltration processes can be found in Simulation of Net Infiltration for Modern and Potential Future Climates (USGS 2000b) and in Volume 1, Section 3.3.1 of FY01 Supplemental Science and Performance Analyses (BSC 2001a).

4.2.1.1.2 Fracture Flow and Matrix Flow within Major Hydrogeologic Units

An early conceptual hydrologic model of unsaturated flow at Yucca Mountain, developed by Montazer and Wilson (1984), identified five major hydrogeologic units within the unsaturated zone. From the land surface to the water table, these units are the Tiva Canyon welded (TCw), the Paintbrush nonwelded (PTn), the Topopah Spring welded (TSw), the Calico Hills nonwelded (CHn), and the Crater Flat undifferentiated (CFu) units. Table 4-4 correlates major hydrogeologic, lithostratigraphic, and detailed hydrogeologic units with the layering scheme used in unsaturated zone modeling activities. These units are described in greater detail in the Development of Numerical Grids for UZ Flow and Transport Modeling (CRWMS M&O 2000br, Section 6.4.1), in the Geologic Framework Model (GFM3.1) (CRWMS M&O 2000bs, Section 6.4.1), and by Flint, L.E. (1998).

The texture of Yucca Mountain tuffs ranges from nonwelded to densely welded (CRWMS M&O 2000c, Section 3.2.1). Typically, the porosity and permeability of the rock mass are inversely proportional to the degree of welding, and the degree of fracturing is directly proportional to the degree of welding. The degree of welding observed in the individual tuff units is primarily controlled by their cooling history. Generally speaking, the slower a rock cools, the more densely welded the material becomes. This densely welded material (matrix) is usually quite brittle in nature and develops well-connected fracture networks. These extensive, well-connected fracture networks, in turn, provide numerous pathways for the flow of liquids and gases. Conversely, the nonwelded rocks, which experienced rapid heat dissipation, display high matrix porosity and possess few fractures. Flow through these rocks is dominated by matrix flow processes (CRWMS M&O 2000c, Section 3.3.3).

The partitioning of total flux between the fracture flow component and the matrix flow component is one of the most important processes to determine in the unsaturated zone. Percolation distribution determines the amount of water that could potentially contact the waste packages and other components of the engineered barrier system. Determination of the flow components is also important for chemical transport processes. Water flow in fractures is typically much faster than flow in the matrix, leading to much faster movements for radionuclides and other chemicals in fractures compared to the matrix (CRWMS M&O 2000c, Section 3.3.3). The characteristic flow behavior in each of the major hydrogeologic units is described in the following paragraphs.

Flow Above the Potential Repository—The TCw is the most prevalent hydrogeologic unit exposed at the land surface (CRWMS M&O 2000c, Section 3.2.2.1). The unit is of variable thickness because of erosion and is composed of moderately to densely welded, highly fractured pyroclastic flow deposits of the Tiva Canyon Tuff. The high density of interconnected fractures and low matrix permeability of the unit (CRWMS M&O 2000bt, Sections 6.1 and 6.2) are considered to give rise to significant water flow in fractures and limited matrix imbibition (water flow from fractures to the matrix). Therefore, episodic infiltration pulses are expected to move rapidly through the fracture system into the underlying PTn unit with little attenuation by the matrix.

At the interface between hydrogeologic units TCw and PTn, tuffs grade downward over a few tens of centimeters from densely welded to nonwelded, accompanied by an increase in matrix porosity and a decrease in fracture frequency (Figure 4-5a) (CRWMS M&O 2000c, Section 3.2.2.1). The relatively high matrix permeabilities and porosities, and low fracture densities, of the PTn (CRWMS M&O 2000bt, Sections 6.1 and 6.2) should convert the predominant fracture flow in the TCw to dominant matrix flow within the PTn unit (CRWMS M&O 2000bq, Section 6.1). This, along with the relatively large storage capacity of the matrix resulting from its high porosity and low saturation, is expected to give the PTn significant capability to attenuate infiltration pulses and smooth areal differences in infiltration from the overlying welded unit and result in approximately steady-state water flow below the PTn. Through-going fracture networks within the PTn unit are rare and typically associated with faults (Rousseau, Kwicklis et al. 1999, pp. 53 to 54), so only a small amount of water is expected to pass through the PTn by way of fast flow paths. Recent analyses indicates that some lateral diversion on the PTn is probable (BSC 2001a, Section 3.3.3).

Conceptualization of the character of flow at the interface between the PTn and the overlying TCw for the TSPA-SR model was based on findings from estimates of flux rates in the PTn from geochemistry (Yang and Peterman 1999) and on hydrogeological properties described by Flint, L.E. (1998). The implication of that characterization was that little or no lateral flow diversion of downward percolating unsaturated zone water could be expected as a result of a contrast of hydrologic properties across this surface and its gentle dip to the east. If diversion of downward percolating water to permeable fault zones outside the potential repository footprint at that elevation occurs, it implies that water might be diverted around a repository, thus benefiting repository performance. Earlier work by Montazer and Wilson (1984) supported the conceptual model of lateral diversion on the PTn. The TSPA-SR model did not include lateral diversion on the PTn on the assumption that if any such diversion exists, it is small and that to leave it out of the model is conservative.

Geochemical evidence collected since the conceptualization of the original TSPA-SR modeling appears to support the existence of lateral diversion in the PTn based on chloride abundance (BSC 2001a, Section 3.2.3). Recent modeling approaches that combine pneumatic data from above, below, and within the PTn, saturation and water potential data, and geochemical data were used to calibrate unsaturated characteristic parameters and to differentiate alternative conceptual models (BSC 2001a, Section 3.2.3). Using a fine grid spacing, calibrated to match the chloride distributions and the estimated percolation flux data in the unsaturated zone below the PTn, supports the lateral diversion of water around a potential repository and the relatively uniform distribution of percolation flux in the deep unsaturated zone (BSC 2001a, Section 3.3.3). The PTn exhibits inhomogeneous lithologic character and distribution and some evidence, in the form of inferred occurrences of bomb-pulse chlorine-36 at depth, that fast flow paths for relatively small volumes of water may exist along faults and perhaps in zones of fracture flow focusing. In most of the region of the potential repository footprint, the PTn is expected to damp flow surges in the percolation flux rate and smooths areal differences in flow, which originate in the temporal and spatial patterns of infiltration at the surface (CRWMS M&O 2000c, Section 3.3.3.2). Smoothing of flow is supported by evenly distributed chloride mass balance data (BSC 2001a, Section 3.3.3).

Although the PTn is predominantly nonwelded, rock-hydrologic properties are highly heterogeneous because of differing depositional environments, lateral variations in welding, and the variable distribution of mineralogically altered (e.g., smectitic and zeolitic) intervals within the individual PTn subunits (CRWMS M&O 2000c, Section 3.2.2.2).

The transition from the lower PTn into the upper TSw is marked by a decrease in matrix porosity and an increase in fracture frequency (Figure 4-5b) (CRWMS M&O 2000c, Section 3.2.2.2) as the tuffs grade sharply downward from nonwelded to densely welded. These changes in porosity and fracture characteristics may create saturated conditions above this contact that could initiate fracture flow into the TSw.

Lithostratigraphic units within the TSw (including the middle nonlithophysal, lower lithophysal, and lower nonlithophysal potential repository host units) are moderately to densely welded and are primarily distinguished by the relative abundance of lithophysae (cavities formed by bubbles of volcanic gases trapped in the tuff matrix during cooling), crystal content, mineral composition, pumice and rock fragment abundance, and fracture characteristics (CRWMS M&O 2000c, Section 3.2.2.3). Differences in lithophysal abundance and fracture characteristics are shown in Figure 4-6.

Unsaturated flow in the TSw is primarily through the fractures because of the magnitude of matrix hydraulic conductivity of the TSw relative to the estimated average infiltration rate. If the hydraulic gradient is assumed to be one (i.e., flow is vertically downward and gravity driven), the maximum matrix percolation rate is the same as the matrix hydraulic conductivity. Because the estimated matrix hydraulic conductivity of some TSw subunits is much lower than the average estimated infiltration rate (CRWMS M&O 2000bq, Section 6.1.2), the remainder of the flow must be distributed in the fracture network.

Flow Below the Potential Repository—Flow behavior below the TSw is important for modeling radionuclide transport from the repository horizon to the water table because transport paths follow the water flow pattern. The main hydrogeologic units below the TSw are the CHn and CFu (CRWMS M&O 2000c, Sections 3.2.2.4 and 3.2.2.5). The CHn contains primarily nonwelded layers whose initial vitric composition has been variably transformed by high and low temperature alteration to clays and zeolites. A portion of the lower half of the CHn (corresponding to the interior of the Prow Pass Tuff) is characterized by moderately welded to densely welded layers that have undergone devitrification (high-temperature conversion of glass to crystalline material). Devitrified, welded rocks show greater fracture intensity than the nonwelded layers and typically do not contain alteration minerals (Flint, L.E. 1998, p. 9). In the southern half of the potential repository footprint and to the south, a portion of the upper CHn (corresponding principally to the Calico Hills Formation, Tac) is largely unaltered (i.e., vitric). This volume of vitric material is believed to represent the part of the CHn that remained above past elevated saturated zone water levels (CRWMS M&O 2000c, Section 3.2.4).

The CFu unit (consisting of portions of the Bullfrog and Tram tuffs that occur above the water table) is a subset of the Crater Flat Group, which contains the Prow Pass, Bullfrog, and Tram tuffs (CRWMS M&O 2000b, Section 4.5.4.5). Lithostratigraphic units within the CFu are nonwelded to densely welded, with the nonwelded tuffs being pervasively altered to zeolites. The Prow Pass, Bullfrog, and Tram tuffs are all similar in that they each contain devitrified, densely welded interiors that grade above and below into nonwelded, zeolitically altered tuffs.

The nonwelded vitric, nonwelded zeolitic, and welded devitrified tuffs have significantly different properties and flow characteristics. The zeolitic rocks have very low matrix permeability and slightly greater fracture permeability; therefore, a relatively small amount of water may flow through the zeolitic units (primarily through fractures), while most of the water is diverted around these low-permeability bodies (Figure 4-7b) (CRWMS M&O 2000c, Section 3.3.3.4). Conversely, vitric portions of the CHn have relatively high matrix porosity and permeability and are characterized by low fracture frequencies (similar to layers within the PTn); therefore, matrix flow dominates, and fracture flow is believed to be limited in the vitric units. Devitrified tuffs have slightly lower matrix porosities than the nonwelded tuffs and increased fracturing (because of increased welding), yet little or no alteration, giving them relatively higher permeabilities than the zeolitic tuffs (Flint, L.E. 1998, pp. 29 and 32).

The high storage capacity of the vitric (unaltered) CHn matrix will attenuate the rate of water movement through the unsaturated zone (Figure 4-7a). Where extensive mineralogic alteration has occurred—for example, at the TSw–CHn contact—the downward flux of water may exceed the rock's transmissive capacity, leading to ponding above the flow barrier and the formation of perched water (Figure 4-7c). The presence of a low-permeability barrier to vertical flow can lead to lateral flow diversion, especially if the flow barrier is dipping and saturated moisture conditions (i.e., perched water bodies) exist above the barrier. Therefore, not all flow paths below the potential repository horizon are expected to be vertical. Lateral diversion of water at perching horizons may lead to flow focusing if the vertical flow barrier is intersected by a high-permeability feature, such as a fault, that could channel flow to the water table (CRWMS M&O 2000c, Sections 3.3.3.4 and 3.3.5).

4.2.1.1.3 Effects of Major Faults

Different kinds of faults with varying amounts of displacement exist at Yucca Mountain (CRWMS M&O 2000c, Section 3.3.5). Fault hydrologic properties are variable and generally controlled by rock type and stratigraphic displacement. Because major faults have the potential to significantly affect the flow processes at Yucca Mountain, they are important features of the unsaturated zone.

A fault can act as a fast flow conduit for liquid water (Figure 4-7d). In this case, transient water flow may occur within a fault as a result of temporally variable infiltration. Major faults cut through the PTn unit, possibly reducing the attenuating effect of the PTn on transient water flow. However, fast flow along major faults is expected to carry only a small amount of water and may not contribute significantly to the flow of water above the potential repository horizon in the unsaturated zone (see Section 4.2.1.3.1.1). Faults intercepting the perched water bodies, however, can correspond to significant vertical water flow if fault permeability is relatively high because of the locally saturated conditions existing in the surrounding rock (see Figure 4-7c) (CRWMS M&O 2000c, Sections 3.3.5 and 3.7.3.2).

If faults within the CHn are relatively permeable features, they may provide a direct flow pathway to the water table. This is particularly significant because radionuclides released from the potential repository could bypass zeolitic or vitric layers within the CHn unit, where they could be retarded by sorption. Conversely, faults might be considered a positive feature of the site if they divert water around waste emplacement drifts or prevent laterally flowing water from focusing at the area of waste emplacement.

Alternatively, a fault can act as a barrier for water flow (CRWMS M&O 2000c, Section 3.3.5). Where a fault zone is highly fractured, the corresponding coarse openings will create a capillary barrier for lateral flow. On the other hand, a fault can displace the surrounding geologic units such that a unit with low permeability abuts one with relatively high permeability within the fault zone. In this case, the fault will act as a permeability barrier to lateral flow within the units with relatively high permeability. Montazer and Wilson (1984, p. 20) hypothesized that permeability would vary along faults, with higher permeability in the brittle, welded units and lower permeability in the nonwelded units where gouge or sealing material may be produced. While a fault sealed with gouge or other fine-grained material has much higher capillary suction (i.e., driving imbibition), it also has low permeability, retarding the movement of water.

Large lateral flow to the faults and/or focusing of infiltration near the fault zone on the ground surface are required to generate significant water flow in faults. Below the repository, low-permeability (zeolitic) layers in the CHn may channel some flow to faults that act as pathways to the water table. However, it is also possible that alteration of faulted rocks in the CHn and CFu causes the faults to be of low permeability (Figure 4-7e), slowing water movement from the TSw to the water table.

4.2.1.1.4 Fracture–Matrix Interaction

Fracture–matrix interaction refers to flow and transport between fractures and the rock matrix (CRWMS M&O 2000c, Section 3.3.4). Owing to their different hydrologic properties, distinct flow and transport behavior occurs in each hydrogeologic unit. The extent of fracture–matrix interaction is therefore a key factor in assessing flow and transport processes in the unsaturated zone (BSC 2001a, Section 4.2.1).

4.2.1.1.5 Seepage into Drifts

Potential seepage of water into waste emplacement drifts is important to the overall performance of the potential repository system. The corrosion of drip shields and waste packages, the mobilization of radioactive contaminants from breached waste packages, and the migration of radionuclides to a receptor location all depend on the distribution of water seepage into the emplacement drifts.

Seepage is defined as flow of liquid water into an underground opening, such as a waste emplacement drift or exploratory tunnel (CRWMS M&O 2000c, Section 3.9.1). Seepage does not include water vapor movement into openings or condensation of water vapor within openings. Seepage flux is the rate of seepage flow per unit area. The seepage percentage is the ratio of seepage flux to percolation flux in the surrounding host rock unit. Seepage threshold is defined as a critical percolation flux below which seepage into the opening is unlikely to occur. The seepage fraction is the proportion of waste packages that are located where drift seepage occurs. The drift shadow zone is a zone of reduced water saturation beneath the emplacement drift as a result of diversion of seepage around the drift opening by capillary forces.

Estimating seepage into underground openings excavated from an unsaturated fractured formation requires an understanding of processes on a wide range of scales (CRWMS M&O 2000c, Section 3.9.1). These scales range from the mountain-scale distribution of percolation flux, to the intermediate-scale channeling or dispersion of flow in an unsaturated fracture network, to the small-scale capillary-barrier effect, to the microscale phenomena within fractures, and specifically at the drift wall. Moreover, the thermodynamic environment in the drift (temperature, relative humidity, ventilation regime, etc.) must be considered. Figure 4-8 illustrates and summarizes seepage-related processes. The factors affecting drift seepage highlighted in Figure 4-8 are outlined below.

Capillary-Barrier Effect, Flow Diversion, and Seepage Threshold—For unsaturated conditions, the seepage flux is expected to be less than the percolation flux because the drift opening acts as a capillary barrier (Philip et al. 1989). When percolating water encounters the opening, capillary forces retain the water in the rock, preventing it from seeping into the drift. Water accumulates in the rock around the opening, and if the rock permeability is sufficient, the water flows around the drift opening. If the incident percolation flux is very high or the rock permeability is insufficient, complete water saturation occurs in the rock above the opening, and seepage occurs. The effectiveness of the capillary barrier to seepage is determined by the capillarity of fractures surrounding the drift and by the permeability and connectivity of the fracture network in the horizontal direction (BSC 2001a, Section 4.2.2). Note that even if seepage occurs, the seepage flux is generally less than the percolation flux unless flow is focused by fractures or other features. The seepage threshold indicates whether or not water seeps into the opening for a given average percolation flux in the surrounding rock. Seepage threshold behavior is controlled by drift geometry, fracture geometry, capillary properties of fractures, and the hydrologic properties of the fracture network (BSC 2001a, Section 4.2.2).

Distribution of Percolation Flux, Flow Channeling, and Episodic Flow—The magnitude and spatial distribution of percolating water in the potential repository host rock is the most important factor affecting seepage. The distribution of flow channels in the fracture network and the hydrologic properties of individual flowing fractures determine how seepage occurs (CRWMS M&O 2000c, Section 3.9.1). Depending on the flow of water within an individual channel, the seepage threshold may or may not be exceeded locally. This is important because seepage is sensitive to the magnitude of the percolation flux, which is moderated by flow processes between the ground surface and the potential repository horizon. For repository thermal loading conditions, the percolation flux will include downward flow of condensed water, in addition to water that infiltrates at the ground surface.

Hierarchal Fracture Network—The characteristics of the fracture network affect seepage because they determine the spatial distribution of percolation flux and the effectiveness of the capillary barrier (CRWMS M&O 2000c, Section 3.9.1). Intermediate-scale characteristics (between mountain-scale and drift-scale) of the fracture network control the potential focusing of flux in the unsaturated zone. Heterogeneity of the fracture network affects local percolation flux and, therefore, seepage. The capillary-barrier effect depends on the connectivity of the fracture network near drift openings and the capillary properties of individual fractures. Small fractures and microfractures, if interconnected, can decrease seepage because they have sufficient capillarity to hold water, but (unlike the rock matrix) sufficient permeability for flow diversion around the openings.

Drift Opening Geometry and Rock Surface Characteristics—The shape and size of underground openings also affects the likelihood of seepage. Partial collapse of the opening because of rockfall can affect seepage. Analytical solutions demonstrating the impact of drift geometry on seepage were developed by Philip et al. (1989). In addition, the geometry of the rock roof in drift openings and the characteristics of the rock surface control processes that could lead to dripping of seepage onto the engineered barriers below.

Ventilation, Evaporation, and Condensation—Until permanent closure of the potential repository, the emplacement drifts will be ventilated. The resulting temperature and humidity conditions in the drift will determine evaporation and condensation effects. Evaporation at the drift wall will generally decrease droplet formation and dripping (Ho 1997a) and create a dryout zone around the drift. When relative humidity in the drift is kept well below 100 percent by ventilation, seepage of liquid water will decrease, while water vapor movement into the drift will increase. Seepage flux and the moisture from increased vapor influx will be effectively removed by ventilation. After the thermal period, the relative humidity in emplacement drifts may be high enough to support condensation within the drift and engineered barrier system whose thermal properties are such that their temperature may be below the dew point. This would mostly occur after the diminution of all or most of the waste heat and the cessation of drift ventilation.

Excavation-Disturbed Zone and Dryout Zone Effects—The capillary-barrier effect that produces seepage diversion around openings occurs within a limited region around the opening (CRWMS M&O 2000c, Section 3.9.1). The extent of this zone is approximately given by the height to which water rises on account of capillarity. It is probably smaller than the zone affected by the mechanical effects of excavation; therefore, these effects may modify seepage behavior. Thermally induced stress changes may also cause changes in fracture permeability. These stress changes are currently under investigation in field-scale thermal testing at Yucca Mountain (BSC 2001a, Section 4.3.1.5). In addition, drift ventilation and heating will produce a dryout zone with associated dissolution and precipitation of minerals (CRWMS M&O 2000c, Section 3.9.1). The consequent alteration of hydrologic properties and its extent are also under investigation (BSC 2001a, Section 4.3.5).

Design—The layout and design of the potential repository and the engineered barrier system affect the probability of seepage water contacting waste packages. Orientation of the emplacement drifts with