Return to the Main Document

(Section 4, continued)

4.3.2 Scenarios Considered in Total System Performance Assessment

The potential repository's capability to isolate radioactive waste is evaluated by modeling scenarios of the evolution of the geologic system and the occurrence of unlikely adverse conditions. Using the systematic procedure described in Section 4.3.1, earth scientists and engineers have developed scenarios of future system evolution and unlikely adverse conditions. These scenarios, which were developed by combining FEPs relevant to the site, are grouped into two basic categories: a nominal scenario and disruptive scenarios.

The nominal scenario includes the most likely FEPs expected to occur in the future (e.g., climate change, repository heating). The disruptive scenario category includes adverse conditions that are extremely unlikely (e.g., volcanism) but that could, if they were to happen, significantly reduce the capability of the repository to isolate waste. In addition to analyses of scenarios in these two categories, the TSPA analyzes a separate scenario for human intrusion, which assumes a drill hole penetrating the repository during a hypothetical groundwater exploration operation.

The following are examples of specific disruptive conditions that were considered in the TSPA evaluation:

Igneous activity (or, alternatively, volcanism) was the principal disruptive scenario retained after the comprehensive screening/selection process described in Section 4.3.1. Potential seismic effects on the underground facilities and waste packages were screened out (i.e., excluded from consideration) because the waste packages would not be adversely damaged by design basis rockfalls or vibratory ground motion (CRWMS M&O 2000ex, Section 6.2). However, because vibratory ground motion from seismic events might damage the commercial spent nuclear fuel cladding, the effect of this potential damage to cladding by a discrete seismic event is included in the TSPA model in the nominal scenario. The frequency assumed for this event is 1.1 math symbol, multiply 10-6 per year. For this analysis, when the vibratory ground motion event occurs, all commercial spent nuclear fuel cladding in the repository is assumed to fail by perforation, and further cladding degradation is calculated according to the cladding degradation. A drilling scenario of inadvertent human intrusion is considered in a separate TSPA calculation.

The following sections provide background information relevant to the disruptive scenarios that have been considered in the TSPA.

4.3.2.1 Volcanic/Igneous Activity

For more than two decades, scientists have performed extensive volcanism studies at Yucca Mountain. Their studies identified the location, age, volume, geochemistry, and geologic setting of past volcanic activity in the area. The results from these studies are described in Yucca Mountain Site Description (CRWMS M&O 2000b, Section 12.2). Although scientists cannot predict future volcanic activity with total certainty, the resulting data provide a comprehensive basis for estimating the probability of future volcanic activity and for determining the effects on people and the environment if volcanic activity were to disrupt the potential repository (CRWMS M&O 2000f, Section 3.1).

4.3.2.1.1 Past Volcanic Activity in the Yucca Mountain Region

Volcanoes have played an important role in the development of the Yucca Mountain region. From 15 million to about 7.5 million years ago, a series of large, silicic, explosive volcanic eruptions occurred in the region. These eruptions produced dense clouds of incandescent volcanic glass (silica), ash, and rock fragments, which melted or compressed together to create layers of rock called tuff. Some of the explosive volcanoes in the Yucca Mountain region formed large calderas as much as 20 km (12 mi) in diameter. Calderas are circular depressions that form when large volumes of magma erupt rapidly, causing the volcano's surface to collapse.

The large-volume, silicic type of volcanic activity no longer occurs in the Yucca Mountain region and has not occurred for more than 7.5 million years. The layers of ash-fall and ash-flow tuff formed by these eruptions have since been disrupted by faulting and erosion. The subsurface bedrock units of Yucca Mountain are made up of ash-fall and ash-flow tuffs that were deposited approximately 13 million years ago, during the episode of silicic volcanism.

Basaltic volcanism began during the latter part of the caldera-forming phase, as rates of extension of the earth's crust waned; small-volume basaltic volcanism continued in the Quaternary Period (the past approximately 2 million years). Collectively, the calderas and basaltic eruptions in the Yucca Mountain region are called the southwestern Nevada volcanic field (Sawyer et al. 1994). Approximately 99.9 percent of the volume of the southwestern Nevada volcanic field erupted between 15 and 7.5 million years ago. The last 0.1 percent of eruptive volume of the volcanic field, consisting entirely of basalt, erupted since 7.5 million years ago. Considered in terms of total eruption volume, frequency of eruptions, and duration of volcanism, basaltic volcanic activity in the Yucca Mountain region defines one of the least active basaltic volcanic fields in the western United States (CRWMS M&O 2000ez, Section 6.2).

In the Yucca Mountain region, there are more than 30 basaltic volcanoes that were formed between 9 million and 80,000 years ago. These volcanoes can be separated into two distinct periods of volcanism that are separated both temporally and spatially. The older basaltic volcanoes formed between about 9 and 7.3 million years ago (during the Miocene epoch). The younger (post-Miocene) basaltic volcanoes erupted between approximately 5 million and 80,000 years ago. As shown in Figure 4-158, the general location of the younger, post-Miocene volcanoes shifted substantially to the southwest (CRWMS M&O 2000ez, Section 6.2).

The post-Miocene volcanoes were formed during at least six different episodes that occurred within 50 km (31 mi) of the proposed Yucca Mountain repository (Figure 4-158). Three of these episodes produced six cinder cones that are in or near the Crater Flat basin, within 20 km (12 mi) of Yucca Mountain (Figure 4-159). The latest volcanic episode, about 80,000 years ago, created the Lathrop Wells Cone, about 18 km (11 mi) south of the potential repository site.

Basaltic volcanoes form from magma that has a low silica and water content. In the southern Great Basin, basaltic volcanoes generally do not erupt as violently as magmas with higher silica and water content. Basaltic volcanoes in the Yucca Mountain region typically form cinder cones associated with small-volume lava flows. However, if the water content is high enough or the magma encounters groundwater during its ascent, explosive phases can also occur, resulting in eruption of an ash plume and deposition of an ash blanket. The youngest basaltic volcanoes in the Yucca Mountain region contain deposits that record moderately violent eruptions that may have produced ash plumes 5 to 10 km (3 to 6 mi) high, along with evidence for less violent and even nonexplosive eruptions.

Most observed basaltic eruptions begin as fissure eruptions, discharging magma where a dike (a vertical, tabular sheet of intrusive magma) intersects the earth's surface. They rapidly become focused at one or more vents into semicircular conduit eruptions. Volcanoes in the Yucca Mountain region are each fed by one main dike, along which a central cone and other vents may form, although subsidiary dikes are also present (CRWMS M&O 2000fa, Section 6.1). Typical basaltic dikes in the Yucca Mountain region are approximately 1.5 m (5 ft) wide (CRWMS M&O 2000fa, Section 6.1) and about 1,000 to 5,000 m (3,000 to 16,000 ft) long (CRWMS M&O 2000ez, Figure 13).

4.3.2.1.2 Probabilistic Volcanic Hazard Analysis

To assess the probability of volcanic activity disrupting a repository, the DOE has performed numerous analyses and conducted extensive volcanic hazard assessments. A panel of ten experts representing a wide range of expertise in the fields of physical volcanology, volcanic hazards, geophysics, and geochemistry were assembled to evaluate the volcanic hazard. The scientists reviewed extensive information presented by representatives of the DOE, U.S. Geological Survey (USGS), State of Nevada, NRC, and others regarding the timing and location of possible future volcanic activity near Yucca Mountain. This study included a careful evaluation of the uncertainties in all the analyses. The expert panel devoted considerable effort to evaluating the existing data, testing alternative models and hypotheses, and ultimately, incorporating a wide variety of alternative models and parameters in their evaluations. Their evaluations (elicitations) were then combined to produce an integrated assessment of the volcanic hazard that reflects a range of alternative interpretations.

To estimate the probability of future volcanic activity at Yucca Mountain, the panel of experts used sophisticated modeling techniques documented in Probabilistic Volcanic Hazard Analysis for Yucca Mountain, Nevada (CRWMS M&O 1996b). As defined in Probabilistic Volcanic Hazard Analysis for Yucca Mountain, Nevada, a volcanic event is the formation of a volcano (with one or more vents) from the ascent of basaltic magma through the crust as a dike or system of dikes (CRWMS M&O 1996b, Appendix E). For the hazard analysis, the panel of experts considered a volcanic event as a point in space representing a volcano and an associated dike having length, azimuth, and location relative to the point event (CRWMS M&O 2000ez, Figures 10 and 12). This hazard analysis evaluated the annual probability of a future basaltic dike intersecting the subsurface area of the potential repository, based on considerations of the locations and recurrence rates of past volcanic activity in the region.

The hazard analysis models are based on data from Yucca Mountain volcanic studies, along with other data and observations from analogue studies of both modern and ancient volcanic eruptions. From these studies, scientists infer how and where magmas form and the processes that control the timing and location of magma ascent through the earth's crust. The panel of experts agreed that future volcanism is more likely to occur within or near existing clusters of geologically recent volcanism than elsewhere in the Yucca Mountain region. While the experts considered the entire 15-million-year history of volcanism in the Yucca Mountain region, they assigned the highest weights for assessing the volcanic hazard in the probabilistic analysis to the past 5 million years. They also emphasized the Crater Flat basin because of the frequency of past volcanic activity there and its proximity to the potential repository (Figure 4-159). One difficulty, however, in evaluating past basaltic volcanic activity in the area is that evidence of basaltic volcanoes, particularly those that are older than 2 million years, may have been eroded or buried by younger sediments. The panel of experts recognized the possibility that there may be additional undetected basaltic volcanoes in the Yucca Mountain region and factored this into their uncertainty estimates for the number of volcanic events that have occurred.

The results of the probabilistic volcanic hazard analysis (CRWMS M&O 1996b) form the basis for probabilistic volcanic risk estimates that account for the repository layout described in this report and the probability of eruption through the repository, conditional on dike intrusion within the repository footprint. These latter results are included in Total System Performance Assessment for the Site Recommendation (CRWMS M&O 2000a).

The results of the hazard analysis estimate that 1.6 math symbol, multiply 10-8 igneous events per year could be expected to disrupt the potential repository (CRWMS M&O 2000ez, Section 6.5.3). This translates to approximately one chance in 6,250 of an igneous event disrupting the repository during the first 10,000 years after repository closure. This is the probability of a future basaltic dike intersecting the subsurface area of the potential repository (intrusive scenario). If a dike does intersect the repository, analysts estimate about a 77 percent chance that a volcano would form at the surface with magma flowing through the repository (eruptive scenario) (CRWMS M&O 2000ez, Table 13a). This translates to approximately 1 chance in 7,700 of a volcano forming above the repository during the first 10,000 years. The annualized probabilities for both disruptive events are just slightly greater than the probability cutoff of 10-8/yr; therefore, both intrusive and eruptive igneous scenarios have been included in the TSPA. Earlier versions of the TSPA analyses used lower estimates of the likelihood that magma contacting the repository would erupt at the surface (a probability of 0.36). Nevertheless, both intrusive and eruptive scenarios were also included in those analyses.

Analysis of new data collected in the Yucca Mountain region since the completion of the hazard assessment demonstrates the thoroughness of the results. Post-elicitation studies by the NRC (Connor et al. 1997) provided evidence to support the possibility of a greater volume for one of the volcanic centers in the Crater Flat basin and an additional volcanic center in the Amargosa Valley. Sensitivity studies showed that these new data did not significantly affect the results of DOE's hazard assessment (Brocoum 1997). The DOE will continue to monitor data and will incorporate significant new information into future technical and licensing documents.

4.3.2.1.3 Consequence Analyses for Volcanic/Igneous Disruptive Events

The DOE evaluated the possible consequences of volcanic activity disrupting the potential repository and estimated the risk to people and the environment that would result from such a disruption. Models of the consequences of an igneous intrusion into, or a volcanic eruption through, a repository at Yucca Mountain require specific information about the nature of the igneous event and the response of the repository to intrusion. This analysis considered the full range of basaltic eruptive processes that have occurred in the Yucca Mountain area during the Quaternary Period, from explosive to nonviolent basaltic eruptions. Information used in the TSPA-SR model to characterize these intrusive and eruptive processes comes from three sources: (1) examination of the geologic record of past intrusive and eruptive events in the Yucca Mountain region, (2) observations of eruptive processes during analogous modern volcanic events elsewhere in the world, and (3) consideration of the range of physical processes that might occur during the interaction between the repository and an igneous dike. The first two sources of information are described in the analysis model reports Characterize Framework for Igneous Activity at Yucca Mountain, Nevada (CRWMS M&O 2000ez) and Characterize Eruptive Processes at Yucca Mountain, Nevada (CRWMS M&O 2000fa). The analysis model report Dike Propagation Near Drifts (CRWMS M&O 2000fb) and the calculation Waste Package Behavior in Magma (CRWMS M&O 1999p) provide additional information regarding dike and repository interactions.

Figure 4-160 is a schematic illustration of the hypothetical igneous activity modeled in the analysis. Two possible scenarios for igneous disruption are included in the consequence analysis: a model for volcanic eruptions that intersect drifts and bring waste to the surface; and a model for igneous intrusions that damage waste packages and expose radionuclides to groundwater transport processes. These models are described in detail in Igneous Consequence Modeling for the TSPA-SR (CRWMS M&O 2000fc) and represented schematically in Figure 4-161.

The scenario for a volcanic eruption assumes that magma erupts through a section of the repository, forming a volcano at the surface (Figure 4-161a). This scenario assumes that an igneous dike rises through the earth's crust and intersects one or more drifts in the repository. An eruptive conduit (the vertical cylindrical passageway through which magma and pyroclasts move upward) forms somewhere along the dike as it nears the land surface, feeding a volcano at the surface. Waste packages in the path of the conduit are destroyed, and the waste is available to be entrained in the eruption. Volcanic ash is contaminated, erupted, and then transported by wind. Ash settles out of the plume as it is transported downwind, resulting in an ash layer on the land surface. The receptor receives a radiation dose from various pathways associated with the contaminated ash layer.

The scenario begins with an eruptive event, which is characterized in the TSPA by both its probability (CRWMS M&O 2000ez, Section 6.5.3) and its physical properties, such as energy and volume of the eruption, composition of the magma, and properties of the pyroclastic ash. Interactions of the eruption with the repository are described in terms of the damage to the engineered barrier system and the waste package. Characteristics of the waste form in the eruptive environment are described in terms of waste particle size. Atmospheric transport of waste in the volcanic ash plume begins with entrainment of waste particles in the pyroclastic eruption and is affected by wind speed and direction. Biosphere dose conversion factors are developed specifically for exposure pathways relevant to atmospheric deposition of contaminated ash, rather than the groundwater pathways considered for nominal performance. As a final step, the volcanic eruption biosphere dose conversion factors are used to determine radiation doses resulting from exposure to contaminated volcanic ash in the accessible environment, approximately 18 km (11 mi) from the repository.

The other scenario in the analysis models the possible effects of a basaltic dike that intersects a section of the repository and partially or completely engulfs waste packages in magma (Figure 4-161b). This may or may not be accompanied by an eruption from the surface of the mountain. Radionuclide releases from waste packages damaged by the intrusion are then available for transport in groundwater. The rate of transport depends on the solubility limits of the waste and the availability of water. The movements of radionuclides released by this type of an event are modeled directly in the TSPA using existing flow-and-transport models.

This scenario begins with an intrusive event (dike intersection), which is characterized in the TSPA by its probability and physical properties. Although the intrusion damages waste packages and other components of the engineered barrier system, it does not significantly alter the long-term flow of water through the mountain. Possible effects of a dike intrusion on the mountain hydrology in the unsaturated and saturated zones are discussed in Features, Events, and Processes: Disruptive Events (CRWMS M&O 2000ex, Section 6.2.8). Based on natural analogue sites, there is no indication of extensive hydrothermal circulation and alteration related to magmatic intrusion. In particular, natural analogue studies at sites on the Nevada Test Site show that alteration is limited to less than 10 m (30 ft) into the country rock adjacent to a dike intrusion (CRWMS M&O 2000ex, Section 6.2.9). Therefore, the disruptive event scenario uses nominal models to describe groundwater flow and radionuclide transport through the mountain.

For calculation of the annual dose resulting from radionuclides that are transported in groundwater following a disruptive igneous event, conditions in the biosphere at the location of the receptor are assumed to be the same as the conditions for nominal performance. Biosphere dose conversion factors for this pathway are therefore the same as those used in the nominal scenario. Because the total expected annual dose in the TSPA models is the sum of probability-weighted doses from both the nominal and disruptive scenarios, any additional increment of dose due to nominal processes that might occur following an igneous disruption is appropriately included in the overall analysis.

The TSPA-SR conceptual models take a conservative approach to modeling uncertainty in several respects. For the purposes of the analysis, the contents of all packages that are fully or partially damaged by an eruption (i.e., that lie in part or entirely within the circumference of the conduit, as described in Section 4.4.1) are assumed to be fully available for entrainment in the eruption. Likewise, it is conservatively assumed that any waste package that is partially or completely intersected by an intrusive dike is fully destroyed. The model does not take credit for the possibility of the magma encapsulating the waste and waste package, which could slow or even prevent water from reaching the waste. In the eruptive scenario, the entire volume of erupted material is conservatively assumed to have been involved in a violent phase of eruption. Observations of both modern and past analogue volcanoes indicate that the violent phases account for only a portion of the total eruption. This assumption overestimates, perhaps significantly, the amount of energy in the eruption and, therefore, the amount of ash transported away from the site. The TSPA evaluates the risk to people and the environment from both disruptive igneous event scenarios. These include (1) doses from the direct release of contaminated ash from an explosive volcanic eruption and (2) doses resulting from the release of radionuclides into the groundwater from waste packages damaged by magma intrusion. The results of these analyses are discussed in Section 4.4.3. Indirect effects that result from volcanic activity outside the repository (e.g., changes to the hydrologic and mineralogical properties of the rock or alteration of water flow and transport) have such low consequences that they are not evaluated further (CRWMS M&O 2000ex).

4.3.2.2 Seismic Activity

The geologic setting of a region influences potential earthquake effects by controlling:

During site characterization, the DOE performed extensive studies at Yucca Mountain to estimate the potential sizes and frequencies of future earthquakes and to determine the level of ground motion and fault displacement that might affect potential repository facilities, both on the surface and underground. The DOE has used the results from these studies to design repository facilities that will withstand future earthquakes and to assess the long-term performance of the total repository system.

Scientists cannot predict the exact location and timing of future earthquakes. However, through intensive study of an area's surface and underground geology, with particular attention to past faulting, scientists can estimate the frequency and size of future earthquakes, the potential intensity of ground movement, and the possible effects from earthquakes on the area's geologic features and man-made structures.

The Yucca Mountain region is one of the most intensively studied areas in the world. Since 1978, scientists have collected and analyzed data on the geologic features of the surface and underground environments. These features provide information on the area's past seismic activity, which is important in estimating the characteristics of future earthquakes. Using these studies, scientists rate the Yucca Mountain region as having low to moderate seismicity (Figure 4-162) (CRWMS M&O 2000fd, Section 6.3.1).

Most of the movement in the earth's crust that formed the mountains and valleys of the southern Nevada landscape ended about 10 million years ago, but slow movement has continued into the present. Some of this movement occurs along faults, which are zones of weakness in the earth's crust. Future movement on some faults will produce future earthquakes. Therefore, intensive study has focused on the faults in the Yucca Mountain region to determine which faults are likely to move in the future and to estimate the potential magnitude and frequency of those earthquakes.

Scientists have used the data from site characterization studies to assess the seismic hazard at the site by defining the potential for ground shaking and fault rupture related to earthquakes. The type of analysis used is called a probabilistic seismic hazard analysis. This type of analysis provides estimates on where and how often earthquakes might occur and on how much fault displacement and ground motion could result. To assess the seismic hazard for Yucca Mountain, scientists factored in different interpretations to account for uncertainties in evaluation of the data for each geologic feature.

Recent earthquakes near Yucca Mountain have been consistent with the seismic hazard analysis for the Yucca Mountain site. For example, in 1999 a magnitude 4.7 Mw (moment magnitude) earthquake occurred at Frenchman Flat about 45 km (28 mi) east of Yucca Mountain (for an earthquake of this size, moment magnitude is approximately equal to magnitude on the Richter scale). This was a moderate-magnitude event in a zone that scientists had identified as a seismic source. This earthquake did not exceed the ground motion estimates from the seismic hazard analysis. (See Sections 6.3 and 6.5 in Characterize Framework for Seismicity and Structural Deformation at Yucca Mountain, Nevada [CRWMS M&O 2000fd] for a discussion of seismic source characterization.) The results from the seismic hazard analysis are a sound basis for designing repository facilities to withstand potential earthquakes and for assessing future repository performance.

Engineers use proven techniques to design structures to withstand the potential earthquakes within that area. The repository surface facilities, where waste would be received, prepared for emplacement, then moved into the repository, would be subject to stronger earthquake ground shaking than subsurface facilities. (See Section 4.3.2.2.1 for more on the effects of seismic activity underground.) Extensive experience in designing and operating critical facilities, such as nuclear power plants, will be relied on in the design of the structures used during repository operations, so they will perform their safety functions during and after an earthquake. After all the waste is emplaced underground, the risks related to seismic activity would decrease. However, a strong earthquake could cause rock falls in the emplacement drifts or alter the pathways followed by groundwater. Scientists have studied these potential consequences and concluded that it is extremely unlikely that the consequences would be significant to repository performance. (See Section 4.3.2.2.3 for a discussion of fault displacement effects.)

4.3.2.2.1 Effects of Seismic Activity Underground

Underground openings are less likely to sustain damage from earthquakes than structures on the surface. This is because ground shaking is stronger on the surface than underground. When an earthquake occurs deep underground, locations at the surface receive ground motion energy simultaneously from the upward-traveling waves and from waves reflected back at the earth's surface. In addition, subsurface rock generally is stronger than weathered near-surface rock, which results in a smaller ground motion at depth.

A seismometer (an instrument that records earthquake waves) at the ground level and another in the Exploratory Studies Facility recently measured seismic waves from a 3.5 Mw earthquake at Frenchman Flat (this event was an aftershock of the 4.7 Mw earthquake described previously). Recordings from this earthquake, as shown in Figure 4-163, clearly indicate the decrease in amplitude of ground motion deep within the mountain (Savino et al. 1999). Scientists have observed this effect around the world for underground structures. For example, tunnels near the 1995 Kobe, Japan earthquake (6.9 Mw) experienced no major damage despite high surface ground motion and extensive damage to surface facilities (Savino et al. 1999).

4.3.2.2.2 Probabilistic Seismic Hazard Analysis

The probabilistic seismic hazard analysis identifies the potential earthquake ground motion and fault displacement that could occur at the Yucca Mountain site. The results from this analysis were incorporated into estimates of seismic consequences and used in designing the repository's facilities. The following paragraphs describe the DOE's methods for performing the seismic hazard analysis (CRWMS M&O 2000fd).

The probabilistic seismic hazard analysis involved a multistep process. First, scientists identified the location of potential earthquake sources and defined their characteristics. Next, they estimated how frequently earthquakes of various magnitudes might occur at each source location. In the third step, experts calculated potential earthquake effects at Yucca Mountain, including the level of vibratory ground motion and amount of fault displacement, given earthquakes of a particular magnitude. Finally, they combined all of the information from the previous steps into "hazard curves" that show the probability of exceeding different levels of ground motion or fault displacement at a particular location during a specific period in time.

The method for performing the seismic hazard analysis of Yucca Mountain is state-of-the-practice and consistent with recent guidance developed by government agencies (e.g., the U.S. Army Corps of Engineers, the USGS). The NRC uses this method to evaluate the safety of existing nuclear power plants, and it would be the basis for licensing new plants (CRWMS M&O 2000fd, Section 6.1.4; CRWMS M&O 2000f). Specific information provided by the probabilistic seismic hazard analysis includes:

Many scientists and engineers contributed to these activities. These individuals were associated with universities (including the University of Nevada, Reno and the University of Utah), government agencies (including the DOE, USGS, and U.S. Bureau of Reclamation), and private companies.

Seismic Source Location and Geometry—A seismic source is a region of the earth's crust that has certain characteristics indicating that earthquakes could originate there. In the probabilistic seismic hazard analysis, seismic sources can be of two basic types: fault sources and areal sources.

Fault sources are known fractures in the earth's crust containing characteristics that indicate past movement along the fault's surfaces. A fault is considered to have the potential to move in the future if it has moved in the past 2 million years (the Quaternary Period). Evidence from historically active faults around the world indicates that, in general, faults that show characteristics of having moved in the recent past, or "active faults," are those most likely to sustain future movement. Studies of faults in the Yucca Mountain region have focused on active and potentially active faults.

To represent the full range of possible fault rupture patterns and interactions near Yucca Mountain, experts used various combinations of possible fault orientations and behaviors in their analyses of each potentially active fault. They also used alternate lengths for each fault to compensate for uncertainties in geologic mapping and different rupture patterns. Figure 4-164 shows the faults that were considered in the seismic analysis. Table 4-32 shows selected fault parameters for potentially significant faults within 10 km (6.2 mi.) of the repository that potentially could rupture independently (that is, rupture of a single fault and not simultaneous rupture of multiple faults). The slip rates are relatively low in comparison with rates for significant regional faults (such as the Furnace Creek and Death Valley faults, which have slip rates of 2.5 to 8 mm/yr [0.1 to 0.3 in./yr]).

Areal sources represent areas where experts believe that potential earthquakes could occur on faults not singularly recognized as significant earthquake sources, such as buried faults. These areas are treated as having relatively uniform earthquake frequency and maximum earthquake magnitudes.

Maximum Earthquake Magnitudes—Experts used two basic approaches to estimate the maximum magnitudes of potential earthquakes in the Yucca Mountain region. The primary approach was based on estimates of the maximum dimensions of fault rupture. Multiple sources of uncertainties were considered in estimating physical dimensions of maximum rupture on faults (e.g., uncertainties in rupture length, rupture area, and displacement per event). Experts then compared these estimated dimensions with the estimated dimensions of ruptured faults and known magnitudes of earthquakes observed throughout the world. The second approach considered historical data on the seismicity of the region and the potential size of unrecognized faults in that region.

Earthquake Recurrence—Earthquake recurrence relationships express the rate or annual frequency of earthquakes occurring on a seismic source. Seismic sources generate a range of earthquake magnitudes, up to some maximum magnitude. A magnitude-distribution model defines the relative number of earthquakes having particular magnitudes. Methods for developing these relationships are usually different for fault sources than for areal sources. Experts estimated recurrence rates for fault sources from geologic data and rates for areal sources from historical seismicity data.

Vibratory Ground Motion Hazard—The level, or amplitude, of ground shaking is influenced by three main elements. The first element is how the size and nature of an earthquake controls the generation of earthquake waves. The second element is the travel path of seismic waves from the source of the earthquake to a particular site. The length of this path is important because the amplitude of ground motion will decrease (attenuate) with distance. The third element is the local site condition, or the effect of the uppermost several hundred feet of rock and soil, and the surface topography. Experts explicitly addressed all three of these elements in the Yucca Mountain seismic hazard analysis. Combining the ground motion analysis with seismic source characteristics produces a probabilistic representation of vibratory ground motion hazard that gives annual exceedance probabilities associated with different levels of earthquake ground motion.

Fault Displacement Hazard—Fault displacement hazard is the hazard related to rupture along a fault triggered by an earthquake. Based on the careful study of fault ruptures associated with earthquakes around the world, the potential for fault displacement is categorized as either principal or distributed faulting. Principal faulting occurs along the main plane of a fault. Distributed faulting is rupture that occurs on other faults in the vicinity of an earthquake in response to the principal displacement.

The DOE's method for assessing probabilistic fault displacement hazard is very similar to that for vibratory ground motion hazard and relies heavily on the detailed geologic studies of individual faults in the Yucca Mountain vicinity. Experts evaluated the fault displacement hazard at nine locations within the Yucca Mountain site. These locations span the range of known faulting conditions in the area, which include recognized faults to small fractures and unfaulted rock. For the period of repository operations (i.e., before closure of the repository surface facilities), results of the hazard assessment show that, in the areas where waste would be emplaced, fault displacements of 0.1 cm (0.04 in.) have less than one chance in 100,000 of being exceeded each year (CRWMS M&O 2000fd, Section 6.6.3). DOE studies have also shown that the consequences of fault displacement in the repository after the repository is closed will not significantly affect performance (see Section 4.3.2.2.3 for a discussion of fault displacement effects).

4.3.2.2.3 Application of Seismic Hazard Analyses

Based on the results of the probabilistic seismic hazard analyses, a team of scientists computed the vibratory ground motion inputs to be used for preclosure design analyses. These inputs were developed for the following areas:

Analysts selected these three areas, shown in Figure 4-165, because they represent the range of locations and conditions where the potential repository's facilities would be located. The ground motion for the three areas was derived from the ground motion developed for a "reference rock outcrop" (Point A) during the probabilistic seismic hazard analysis. The ground motion at each of the three areas is different because of local site conditions. The estimated ground motion at these areas will be used in designing repository facilities and in performance assessments of the potential repository during the postclosure period. Figure 4-166 shows calculated preliminary design ground motion for the Point B and C areas only. These are defined for a probability of 1 chance in 10,000 of being exceeded each year. The design ground motion is represented by response spectra (curves) that indicate the strength of ground shaking at different frequencies of motion (frequency, in Hertz [Hz], is the number of cycles per second of a seismic wave or the vibration of a structure). The shape and stiffness of engineered structures determine which frequencies of seismic waves are important. In addition, the frequencies of strong seismic waves at a site are determined by the size of the earthquake, its distance from the site, and the physical properties of the earth between the site and the earthquake. Two design earthquake spectra are shown on each panel of Figure 4-166 because two kinds of earthquakes control the ground shaking hazard at Yucca Mountain (CRWMS M&O 1998l). The shaking hazard at moderate frequencies (5 to 10 Hz) is due predominately to earthquakes of magnitude Mw 5.5 to 6.5 within 15 km (9 mi) of the site. The shaking hazard at low frequencies (1 to 2 Hz) is due predominately to earthquakes of approximately magnitude Mw 7 and larger at a distance of about 50 km (31 mi). The sources for these larger, more distant earthquakes are the Death Valley and Furnace Creek faults. The DOE has collected additional geotechnical data to refine analyses of ground motion at the Point B, C, and D sites.

The DOE has also studied the potential effects and consequences of fault displacements associated with both known and unknown faults. Experts conducted analyses to evaluate potential effects of fault displacement on emplacement drifts and the engineered barriers in the drifts (i.e., drip shields and waste packages) (CRWMS M&O 2000fe). They also examined how fault displacement could affect water movement within Yucca Mountain, which could affect the movement of radionuclides into the groundwater (CRWMS M&O 2000ff). These detailed studies show that hydrologic changes resulting from fault displacement are highly unlikely, and that faulting events will not affect the safety of the potential repository.

4.3.2.3 Human Intrusion Scenario

The DOE conducted a separate TSPA (see Section 4.4.4) to evaluate how well the repository system would limit radiological exposures under the hypothetical condition of a human intrusion. This scenario assumes a drill penetrating the repository during a possible exploration for groundwater resources. To determine the impacts from this type of disruptive event, the scenario involves site-specific data that represent the repository's natural and engineered barriers. Specifications for this scenario also include such information as the type of drilling, the size of the borehole, and the location of the borehole. By mathematically modeling this scenario, experts evaluated the repository's performance under this disrupted condition and projected the effects. This section describes the basis and assumptions for the TSPA evaluation of the human intrusion scenario.

4.3.2.3.1 Scenario Basis and Assumptions

Human intrusion analyses were performed in Total System Performance Assessment for the Site Recommendation (CRWMS M&O 2000a), 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, Section 5.2.7), and Total System Performance Assessment Sensitivity Analyses for Final Nuclear Regulatory Commission Regulations (Williams 2001b).

The purpose of the TSPA for human intrusion scenario is to provide a basis for judging the resilience of the geologic repository to inadvertent human intrusion (National Research Council 1995, Chapter 4). Unlike analyses of other potential disruptive scenarios, the human intrusion scenario does not evaluate the probability of such an event occurring but instead assumes the event probability of 1. The TSPA analysis uses a "stylized" human intrusion scenario, which refers to specific characteristics that define the scenario.

The human intrusion scenario assumes:

The DOE has estimated that this event could happen no sooner than about 30,000 years after closure without recognition by the driller that he had penetrated a waste package. Therefore, human intrusion is assumed to occur at 30,000 years in 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, Section 5.2.7). For conservatism and consistent with NRC proposed regulations, the DOE also analyzed a human intrusion scenario that assumed (TSPA-SR) that the intrusion occurs at 100 years.

For the purposes of analysis, it was necessary for the DOE to provide supplemental information and data to completely specify the human intrusion scenario. This information and data includes:

Therefore, all of the DOE's human intrusion scenarios assume that someone unknowingly drills for groundwater at Yucca Mountain and the drill penetrates a waste package containing commercial spent nuclear fuel or DOE high-level radioactive waste glass. The drill continues down approximately several hundred meters to the water table. The exposed radioactive waste is mobilized by degradation of the waste form and mixing with the infiltrating water, which then travels down the uncased drill hole and into the water table. After entering the water table, the radionuclides, in dissolved and colloidal forms, are transported by the groundwater to the receptor location. This scenario is illustrated in Figure 4-167.

For an intrusion at 100 years after repository closure, a waste package containing commercial spent nuclear fuel would still be thermally hot, so that any water entering the waste package would convert to vapor (CRWMS M&O 2000fg, Table 10). Thus, the absence of liquid water would prevent waste form degradation and transport of radionuclides down the drill hole. At that time, any water entering the package would convert to vapor, and the absence of liquid water would prevent the transport of radionuclides down the drill hole. However, to add a conservative bias to the analysis, the scenario does not reflect this fact and instead assumes liquid water flows continuously through the penetrated waste package and mobilizes radionuclides. A lower-temperature repository might not be hot enough to vaporize liquid water which then would flow continuously through the penetrated waste package. In either case, the TSPA for a 100-year intrusion would be similar. Additional specifications and assumptions for this human intrusion scenario are presented in Section 4.4.4.1.

4.3.2.3.2 Total System Performance Assessment Evaluation and Sensitivity Analyses

The TSPA-SR and supplemental TSPA evaluations and sensitivity analyses of the human intrusion scenario focused on such aspects as:

To compensate for uncertainty, the TSPA-SR and supplemental TSPA evaluation of the human intrusion scenario does not account for factors within the unsaturated zone (the rock between the repository and the water table) that would limit the transport of some radionuclides. The TSPA-SR evaluated the consequences of the scenario (with human intrusion at 100 years) during the remaining part of the 10,000-year compliance period following the hypothetical intrusion. The analysis assumed the same climatic and hydrogeologic conditions as the nominal scenario (i.e., the representation of the most plausible evolution of the repository system).

The TSPA-SR analysis (intrusion at 100 years after closure) focused on:

The DOE performed a sensitivity analysis (see Section 4.4.5) of this evaluation to identify the factors that have the greatest influence on the release, transport, and radiation exposure to the hypothetical receptor group.

Section 4.4.4 summarizes the results of the TSPA-SR model evaluation of the 100-year human intrusion scenario. In addition, Section 4.4 of Total System Performance Assessment for the Site Recommendation (CRWMS M&O 2000a) documents the results in detail. 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, Section 5.2.7) summarizes the results of the human intrusion scenario occurring at 30,000 years.

4.3.3 Scenarios Addressed and Screened Out of Total System Performance Assessment

Over the past decade, scientists have hypothesized and considered many possible scenarios of conditions that could occur in the distant future and have the potential to affect the repository's capability to isolate radioactive waste. They selected the most probable of these conditions and developed scenarios for further analysis in the TSPA. Many conditions and scenarios were screened out according to probability or consequence. The technical basis for excluding these scenarios from consideration in analyses of the performance of the design described in this report is presented in the various process model and analysis model reports and in the DOE FEPs database (Freeze et al. 2001).

Two particular scenarios that the DOE considered but ultimately excluded from further TSPA analysis have been the focus of past scientific interest. They are:

  1. Rise of the water table—a hypothesis that speculates that tectonic and hydrothermal events could cause the water table to rise to the repository level

  2. Nuclear criticality—the possibility of the spent nuclear fuel in the repository causing a self-sustaining chain reaction (as a result of chemical or physical processes affecting the waste) either in or after release from breached waste packages.

As explained in the following sections, the DOE and others have determined that the water table rise scenario is not scientifically credible and that the nuclear criticality scenario is highly improbable. A model detailed analysis has shown that the probability of a nuclear criticality event falls below the screening threshold of less than 1 chance in 10,000 in the first 10,000 years following emplacement (see Section 4.3.3.2.3). This is primarily because there are few failures of waste packages, which are designed to prevent breach before 10,000 years. Therefore, it was screened out. However, because these scenarios have been the subject of much discussion, Sections 4.3.3.1 and 4.3.3.2 present the scientific investigations that led to the DOE's conclusions.

4.3.3.1 Long-Term Stability of the Water Table

The current elevation of the water table beneath Yucca Mountain is about 730 m (2,400 ft) above sea level, and the potential repository horizon would be an average of about 300 m (1,000 ft) above this elevation. The minimum height of an emplacement drift above the water table is about 210 m (690 ft) in the northernmost section of the repository. Geologic, geochemical, and hydrologic investigations of the site indicate that the elevation of the water table has fluctuated in the past and that the maximum elevation in the Yucca Mountain region was about 120 m (390 ft) above the current level, primarily as a result of variation in climate. Paces and Whelan (2001) have recently re-estimated the maximum climate-caused water table rise during the Pleistocene, including glacial climates, to be 17 to 30 m (56 to 98 ft). Based on this evidence and on numerical modeling, and to conservatively account for uncertainties related to future climate states, the DOE's analyses of repository performance include future variations in water table elevation of up to 120 m (390 ft). In 1987, a DOE scientist hypothesized that much larger variation in the water table could occur as a result of tectonic processes (Szymanski 1989). Szymanski (1989) postulated that tectonic stress could cause fracture apertures to be opened wider than they would be under normal stress and that during an earthquake the extensional stress could be released, causing the fracture apertures to decrease. He further postulated that, in a process called "seismic pumping," the water in the fractures would be compressed, forcing the water upward to the level proposed for storage of nuclear waste (Szymanski 1987, 1989). A group of 23 project scientists (DOE 1989) reviewed this hypothesis and concluded it was not supported by available information.

Since that initial internal review, Szymanski and colleagues have written several additional papers in support of the hypothesis that the repository could be inundated by a rising water table (Davies and Archambeau 1997a, 1997b; Dublyansky and Reutsky 1995; Hill et al. 1995). This new information has been systematically evaluated by the DOE and in a series of external peer reviews, which are discussed in this section. These external peer reviews included:

Except for the minority report of Archambeau and Price (1991), these reviews all concluded that the available evidence does not support the proposed upwelling groundwater hypothesis.

This section will discuss:

4.3.3.1.1 Evidence for the Origin of Surficial Deposits of Calcite and Opaline Silica

A trench, designated Trench 14, was excavated to a depth of 1.8 m (6 ft) in 1982 to study the Bow Ridge fault along the west side of Exile Hill. The trench exposed a vein-like deposit of calcium carbonate and subordinate opaline silica, as well as a breccia deposit (angular fragments of older rock cemented within a fine-grained matrix) in the bedrock on the footwall. The origin of these deposits was the subject of considerable debate, so in 1984 the trench was deepened to 3.6 m (12 ft) to help elucidate the origin of the deposits. Taylor and Huckins (1986) proposed that the deposits were formed by evaporation of precipitation, leaving behind dissolved minerals (a pedogenic, or soil forming, origin). After examination of deposits in Trench 14, Szymanski (1987) hypothesized that the potential repository was at risk because of upwelling water. Varying hypotheses on the origin of the deposits and peer reviews of the hypotheses are discussed in this section.

Field Observations—Stuckless, Peterman et al. (1992) examined several lines of evidence and concluded that a pedogenic origin provided the best explanation for the deposits observed at Trench 14. This evidence includes field relationships, such as the slope-parallel orientation and great lateral extent of the calcic or caliche deposits below the surface of soils in the desert Southwest. Additionally, some carbonate layers continue upslope beyond their supposed feeder-vein source. A second deepening of Trench 14 in 1989 showed (Taylor and Huckins 1995) that the veins pinch out with depth (Figure 4-168). In addition, the veins are not visible in the Exploratory Studies Facility where the tunnel intersects the Bow Ridge fault, although minor calcite may occur on the footwall. These observations contrast sharply with the morphology observed at Travertine Point (Figure 4-169), a site of known groundwater discharge along Furnace Creek on the east side of Death Valley.

Textural and Mineralogical Evidence—Vaniman, Bish et al. (1988) have shown that the deposit at Trench 14 is similar mineralogically and texturally to soil deposits and distinct from spring deposits. For example, the vein material in Trench 14 is poorly indurated (soft and not well cemented together), porous, and fine grained. In contrast, feeder veins, like those at Travertine Point, are coarse grained and well indurated (having a solid, hard structure). The deposit at Trench 14 exhibits intimate intergrowths of calcite and opal, as well as abundant ooids (egg-shaped particles) and pellets. Both features are atypical of any type of spring deposits, but they are common in soils. Springs and seeps have abundant biogenic evidence of their past higher water content, including fossils of aquatic animals, such as ostracodes and mollusks, and algal or diatomaceous deposits of opal-A. These biogenic features are not found at Trench 14. Biogenic evidence found at springs and seeps also includes poorly preserved root casts of phreatophytic plants (i.e., plants that obtain moisture from below the water table). In contrast, Trench 14 has well-preserved, very delicate filaments that are common in pedogenic calcretes and are associated with roots of xerophytic plants (i.e., plants adapted to low moisture conditions, typically with deep roots).

Monger and Adams (1996) examined the microscopic structure of the deposits at Trench 14 and elsewhere and compared the textures with known pedogenic and phreatic (from groundwater) deposits. They used a combination of thin sections, scanning electron microscopy, x-ray diffraction, and cathodoluminescence, and concluded that the vein deposits appear to be of pedogenic origin.

Geochemical Evidence—Vaniman, Chipera et al. (1994) note that the major- and trace-element chemistry of the deposit at Trench 14 is distinct from the chemistry of spring deposits and similar to the chemistry of soils. The iron and scandium abundance data are most distinctive. The weathering process that forms soils elevates the concentrations of these elements. The elevated iron and scandium content in the veins has the same ratio as the local soils. Lanthanum/ytterbium ratio values in the Trench 14 deposit also match those of the soils. This correspondence can be accounted for only if the material in the veins is derived from soil.

Isotopic data—Isotopes are atoms of a chemical element with the same atomic number (i.e., the same number of protons) but a different atomic mass (containing a different number of neutrons). Several types of isotopic data can be used to demonstrate that local groundwater was not involved in the formation of the Trench 14 veins. The isotopic evidence and references to isotopic work are summarized by Stuckless, Marshall et al. (1998). Oxygen isotopes in the calcite found in veins at Trench 14 indicate that their origin would have been at unreasonably low temperatures if they had been precipitated from either of the aquifers beneath Yucca Mountain. Carbon isotopes preclude involvement of the deepest groundwater aquifer at Yucca Mountain. Strontium in the vein deposit has a greater proportion of the isotope strontium-87 than strontium found in either aquifer beneath Yucca Mountain, which precludes the involvement of either aquifer in the formation of the veins. In contrast, the isotopic composition of strontium in the Trench 14 deposits is within the range measured for soil deposits, which is permissive evidence for formation of the deposits by soil-forming processes. The uranium-234/uranium-238 activity ratio in groundwater beneath Yucca Mountain is anomalously large relative to most groundwaters (greater than 5 in the volcanic aquifer). In comparison, the uranium-234/uranium-238 activity ratio values for both soils and the calcite from Trench 14 are less than 1.5. This isotopic system again shows that the local groundwater could not have formed the deposits, but that they may be closely related to soil-forming processes. Finally, lead isotope data also support a pedogenic origin for the calcite and identify a detrital origin as the probable source of the lead.

Radioactive elements decay at known rates, so the relative abundance of radioactive elements and their decay products can sometimes be used to determine the age of a sample. Alternatively, if one knows the amount of a radioactive element initially present, one can measure the amount currently remaining to calculate an age. Carbon-14 dating is an example of the latter. Samples of modern groundwater from volcanic tuffs sampled within 1 km (0.6 mi) of the potential repository block have apparent carbon-14 ages of 12,000 to 18,500 years (Benson and McKinley 1985). In contrast, apparent thorium-230/uranium-234 ratio ages of calcite found at Trench 14 range from about 80,000 to more than 400,000 years. Data from a drill hole into calcite at Devils Hole (a fault-controlled cavern in limestone in Ash Meadows) show little variation in the isotopic composition of strontium, uranium, oxygen, and carbon for the last 500,000 years (Stuckless, Peterman et al. 1992; Ludwig, Simmons et al. 1992; Winograd et al. 1992). Because the isotopic composition of groundwater beneath Yucca Mountain has not likely varied significantly during the last 500,000 years, the comparison of modern groundwater with that of the deposits provides a valid means of assessing the origin of the deposits.

4.3.3.1.2 Peer Reviews and Evaluations of New Evidence

In 1990, the DOE requested that the National Academy of Sciences/National Research Council consider whether the water table had risen in the geologically recent past to the level of the potential repository and whether a water table rise was likely to occur over the life of the repository in the manner proposed by Szymanski (1989). The National Research Council established a panel that reviewed the pertinent literature and data available up to 1992. This panel consulted with scientists involved in related field and laboratory studies. The panel's conclusion (National Research Council 1992, p. 3) was that none of the evidence offered as proof of groundwater upwelling in and around Yucca Mountain could reasonably be attributed to that process. Furthermore, the panel stated (p. 130): "The preponderance of features [ascribed to ascending water] (1) were clearly related to the much older (13-10 Ma) volcanic eruptive processes that produced the tuffs in which the features appear, (2) contained contradictions or inconsistencies that made an upwelling groundwater origin geologically impossible or unreasonable, or (3) were classic pedogenic features recognized worldwide." The panel concluded that the physical and textural evidence from the veins in the trenches indicated a sedimentary, low-temperature origin from descending meteoric water (infiltration) rather than an origin involving upwelling of thermal water from deep within the crust. The panel also concluded (p. 56): "...to date the preponderance of evidence supports the view that the calcretes and other secondary carbonates in veins of the area formed from meteoric water and surface processes."

Other independent reviewers reached similar conclusions, which were considered by the National Research Council. In the majority report of the independent peer review panel convened in 1991, Powers et al. (1991, p. s-2) concluded: "Surficial deposits cited by Szymanski as evidence of upwelling fluids are consistent in isotopic and physical character with surficial, pedogenic processes; these deposits are not consistent isotopically with known groundwater in the area of Yucca Mountain." Although Archambeau and Price (1991), in their minority report, supported Szymanski's hypothesis, a colleague they selected to review their report concluded the model was "a set of unsupported and unsupportable assertions" (Evernden 1992, p. 65).

After the National Academy of Sciences/National Research Council reached its conclusions, Szymanski and others, as documented by Hill et al. (1995), asserted in 1995 that much of the data available for the calcite/opal deposits could be explained by the upwelling of warm, carbon dioxide-rich water along faults. Hill et al. postulate that once this upwelling water reaches the surface, it flows downhill back into the ground. Therefore, they argued, the resulting deposits are slope-parallel and acquire a pedogenic character. In a critique of this hypothesis, Stuckless, Marshall et al. (1998, p. 70) concluded the hypothesis was based on misstatements, omissions of available information, and misleading generalizations that lead to an erroneous conclusion. In addition, Vaniman, Carey et al. (1999) have examined the chemistry of zeolitic strata beneath the potential repository horizon and concluded that flow in the unsaturated zone has been downwards and predominately gravity-driven for the last 12 million years.

The NWTRB (Cohon 1998) reviewed a group of reports submitted by the Nevada Nuclear Waste Project Office. One of the reports cited evidence from fluid inclusion studies that the authors asserted were an indication of a purported high-temperature origin for secondary calcite collected in the Exploratory Studies Facility (Dublyansky and Reutsky 1995). However, the age of the secondary calcite was unknown, and there was a well-documented thermal period that affected the volcanic rock for a long time after its formation. Because of this and other evidence from the Exploratory Studies Facility excavation, discussed below, the NWTRB concluded that the fluid inclusion data does not significantly affect the conclusions of the National Academy of Sciences/National Research Council report (National Research Council 1992). The NWTRB did, however, recommend further studies of this type in conjunction with age determinations.

In response to this recommendation, the DOE sponsored research at the University of Nevada, Las Vegas and the USGS on age and thermal history indicated by fluid inclusions. Representatives of the state of Nevada participated in the sampling program and were given splits of all materials collected. In addition, the representatives participated in biannual meetings to review and interpret the data. The general conclusion reached by USGS and University of Nevada, Las Vegas researchers was that the fluid inclusions formed in the vadose environment by downward percolating meteoric water and that the environment was warmer than ambient until at least 5 million years ago; however, there is no evidence for above-ambient temperature precipitation during the last 1.9 million years (Paces, Whelan et al. 2000; Paces, Neymark, Persing et al. 2000; Whelan, Roedder, Paces, Neymark et al. 2000; Wilson, N.S.F., Cline, Rotert, and Amelin 2000; and Wilson, N.S.F., Cline, and Amelin 2001).

Evidence for precipitation in a vadose environment includes (1) only 1 to 40 percent of the cavities are mineralized in a given area, whereas precipitation in a saturated environment would predict that most, if not all sites would be mineralized (Marshall, Neymark et al. 2000); (2) mineralization is restricted to floors of cavities and footwalls of fractures (Marshall and Whelan 2000); and (3) the fluid inclusion assemblage of all liquid, all vapor, liquid and vapor in variable proportions is most consistent with a vadose environment (Whelan, Roedder, Paces, Neymark et al. 2000).

Fluid inclusion temperatures for all but the earliest calcites range from 35° to 75°C (95° to 167°F) (Whelan, Roedder, Paces, Neymark et al. 2000; Wilson, N.S.F., Cline, Rotert, and Amelin 2000). Most of these temperatures were determined for calcite that is clearly older than 4 to 5.3 million years (Wilson, N.S.F., Cline, and Amelin 2001). The chemical composition of calcite changed between 2.8 and 1.9 million years ago to include a few percent magnesium, and this calcite lacks two-phase inclusions, thereby indicating precipitation at ambient temperatures (Wilson, N.S.F., Cline, and Amelin 2001).

The isotopic composition of strontium in secondary calcite collected in the Exploratory Studies Facility is more radiogenic (containing more of the isotope strontium-87) in successively younger layers, which is consistent with an origin of meteoric water reacting with rocks that are getting older and accumulating radiogenic strontium (Marshall and Whelan 2000). Carbon isotopes also show an evolution in time that can be related to a change in plant community at the earth's surface, which in turn reflects changing climate (Whelan and Moscati 1998). In contrast to evidence for downward percolating water, groundwater beneath Yucca Mountain is undersaturated with respect to calcite and could not precipitate that mineral without significant degassing or evaporation.

Considerable geochronology has been done on the secondary minerals. Uranium-lead dating shows a long-term slow average growth rate, which is consistent both for an entire secondary mineral coating or for parts of a coating where layers of opal are separated by calcite (Neymark et al. 1998). Uranium series dating yields similar growth rates for the outer layers of secondary minerals (Paces, Whelan et al. 2000; Paces, Neymark, Persing et al. 2000). Disagreement in apparent ages between techniques with different half-lives and the correlation between initial uranium-234/uranium-238 and apparent age is best explained by a continuous growth model and slow growth rates (Neymark et al. 1998; Neymark and Paces 2000).

4.3.3.1.3 Evidence Related to Past Water Table Elevation

Field Observation—Although the calcite and opaline silica deposits provide no direct information on past water table elevations, estimates can be made from other evidence for former high levels of the water table in the region surrounding Yucca Mountain during the past 2 million years. The evidence includes:

Most of these features are attributable to the carbonate aquifer, which is the deep and probably confined aquifer beneath Yucca Mountain. Nonetheless, this aquifer and the one hosted by the volcanic rocks and valley fill likely have responded similarly to past conditions. Available evidence indicates (1) an apparent lowering of the water table of 50 to 70 m (160 to 230 ft) for the carbonate aquifer in the east-central Amargosa Desert since the middle Pleistocene and (2) an apparent lowering of perhaps as much as 130 m (430 ft) since the end of the Pliocene. This estimated general lowering does not preclude small water table fluctuations in response to climate changes superimposed on the general trend.

There are also several discharge deposits from the volcanic and valley-fill aquifer, including three small deposits near the south end of Crater Flat. Root casts from these deposits have yielded uranium-series ages of approximately 10,000 to 20,000 years (Paces, Mahan et al. 1995). Four carbon-14 ages for root casts range from approximately 6,000 to 16,000 years and confirm the young (and possibly Holocene) age for groundwater discharge in Crater Flat. The initial uranium-234/uranium-238 ratio values and the initial strontium-87/strontium-86 values (Marshall, Stuckless et al. 1993) agree with the values obtained for samples from the volcanic and valley-fill aquifer in Crater Flat (Ludwig, Peterman et al. 1993; Peterman and Stuckless 1993). This correspondence confirms the origin of the Crater Flat deposits as groundwater discharge from a regional aquifer.

The largest and most extensive past discharge site for the volcanic and valley-fill aquifer is near Stateline, Nevada. This site is particularly significant in that it occurs where the regional groundwater flow model predicts that—under conditions of nearly a 15-fold increase in recharge and a water table rise of about 100 m (330 ft) beneath Yucca Mountain—discharge would occur (Czarnecki 1984); note that the amount of water table rise predicted to initiate spring discharge is less than the conservative, maximum water table rise calculated by Czarnecki and discussed in Section 4.3.3.1.4. Eight uranium-series ages on five different samples from this site range from approximately 11,000 to 110,000 years, but the ages cluster around the times of the last two pluvial periods (times with higher rainfall) (Paces, Neymark, Marshall et al. 1996). This suggests that maximum water table rises correspond to the wettest past climates. Recent work by Paces and Whelan (2001) on spring discharge deposits from Nye County Early Warning Drilling Program wells is interpreted to indicate that the maximum Pleistocene groundwater table rise was only 17 to 30 m (56 to 98 ft).

Devitrification and Other Mineralogical Evidence—The maximum water table rise suggested by past discharge deposits south of Yucca Mountain is supported by data from drill holes at Yucca Mountain. Bish and Vaniman (1985) noted that the volcanic tuffs were devitrified (i.e., glass was altered to minerals) below the current water table and for as much as 80 to 100 m (260 to 330 ft) above that surface. Devitrification is attributed to prolonged contact with water; therefore, the top of the devitrified zone marks the high stand of the water table. Dibble and Tiller (1981) have concluded that the time required to form zeolites from glass is on the order of 10,000 years, so larger short-term fluctuations are not precluded by the elevation of devitrified tuff. This suggests that the water table must have remained about 100 m (330 ft) higher than at present over the thousands of years required for zeolitization of the glassy tuffs (NRC 1999b, pp. 27 and 28).

Levy (1991) reevaluated the base of the vitric tuff as an indicator of the hydrologic past, considering structures and uplift occurring after devitrification. She noted that the uppermost boundary of the devitrified tuff, which occurred farthest above the current water table, was not parallel to the water table, suggesting that it formed before tilting of the tuffs. Therefore, at least some of the devitrification must have occurred more than 12 million years ago. The most recent water level fluctuations would postdate major tectonic events, but would be difficult to date. The base of the vitric tuff shown on Bish and Vaniman's (1985) cross sections is 10 m (33 ft) to more than 100 m (330 ft) below the base of the potential repository horizon. Thus, it appears likely that the maximum water table elevation is well below the potential repository. If the structural interpretations proposed by Levy (1991) are taken into account, the maximum past rise in the water table, for enough time to cause devitrification, was about 60 m (200 ft).

Because the shallow groundwater beneath Yucca Mountain is undersaturated with calcite, calcite can dissolve in the groundwater. Consequently, secondary calcite is rare below the current water table, occurring only where it is armored, such as in a tightly sealed fracture. Secondary calcite is also rare in a zone as much as 85 m (280 ft) above the water table (Marshall, Stuckless et al. 1993), suggesting that the water table could have risen as much as 85 m (280 ft) in the past.

Carbon isotopes in unsaturated zone calcite are consistent with the carbon isotopic signature acquired in the soil zone by downward percolating water (Marshall, Whelan et al. 1992). This consistency is also corroborated by strontium isotopic compositions for calcite and pore water that are closely similar to values in the soil zone at shallow depth and become more similar to values for the volcanic rock as water reacts with the rock during downward percolation (Marshall, Futa et al. 1998).

Evidence from the Exploratory Studies Facility Excavation—Studies of secondary minerals from the Exploratory Studies Facility show that these minerals formed in the unsaturated zone (Paces, Neymark, Marshall et al. 1996). For example, mineralization in the Exploratory Studies Facility is found only on the floors of cavities and footwalls of open fractures, which contrasts with the locus of mineralization for geodes or veins that formed in the saturated zone. Most cavities are devoid of the mineralization that would be expected in a saturated environment. Uranium-lead and uranium-series dating have shown that the environment of deposition for the secondary minerals has stayed the same for millions of years (Neymark et al. 1998; Neymark and Paces 2000).

4.3.3.1.4 Future Water Table Elevations

Although nothing in the geologic record suggests that the water table might rise sufficiently to flood the potential repository horizon, there are two general mechanisms proposed as capable of doing so in the future: climate change and tectonic activity. A careful examination of both of these mechanisms shows that neither is likely to be able to cause flooding of the repository level.

The long-term climate history has been examined and correlated with variations in the Earth's orbit over the past 400,000 years (USGS 2000a). This climate study concluded that most of the next 10,000 years will be wetter, cooler, and more like the glacial periods in the past and that the climate will be much like the current climate in northern Nevada. Czarnecki (1984) modeled the regional groundwater flow system to examine water table rises due to increased precipitation (i.e., pluvial periods). He calculated a maximum increase in water table altitude of about 130 m (430 ft) beneath the emplacement horizon. This analysis is considered conservative because a 100-percent increase in precipitation during pluvial periods was assumed. This in turn results in a 15-fold increase in recharge from the assumed modern rate of 0.5 mm (0.02 in.) per year to about 8 mm (0.3 in.) per year at Yucca Mountain. Half of the calculated recharge flux in the model was applied directly east of the potential repository site, along a segment of Fortymile Wash, which is consistent with the proposal of Claassen (1985) that Fortymile Wash is an area of recharge under wetter climatic conditions. The flux at Fortymile Wash causes about three-quarters of the computed water table rise of 130 m (430 ft). Czarnecki (1984) notes, however, that under a 100-percent increase in precipitation, large quantities of runoff might flow away from the area down Fortymile Wash and other drainage ways. This would have the effect of decreasing the effective groundwater recharge to less than the calculated values.

Based on past discharge deposits, Quade et al. (1995) calculated a maximum past water table elevation. They concluded that the maximum increase in water table elevation under Yucca Mountain in response to climate change was less than or equal to 115 m (380 ft) during the last full glacial period. They estimated the paleo-water table rise at the Lathrop Wells diatomite spring deposits based on water table information from Winograd and Thordarson (1975). The rise of 115 m (380 ft) was needed to establish spring flow in the past. Nye County drilling at these spring deposit sites established that the water table is currently about 16 to 30 m (52 to 98 ft) below the land surface (NRC 1999b, pp. 25, 26, and 31). This 115 m (380 ft) value is now recognized as a significant overestimate. Recent work by Paces and Whelan (2001) on results from the most recent Nye County drilling indicates that the maximum water table rise during the Pleistocene was between 17 to 30 m (56 to 98 ft). Thus, multiple lines of evidence suggest that a 120-m (390-ft) rise in water table elevation is a reliable estimate of the potential maximum increase due to climate change, based on a projected wetter future climate (CRWMS M&O 2000c, Section 3.7.5.2). Future climate changes are not expected to cause the emplacement drifts to flood because they would be located at a distance ranging between approximately 210 m (690 ft) and 390 m (1,300 ft) above the water table. The average distance of the repository above the water table would be about 300 m (1,000 ft).

Possible tectonic-induced changes in water table elevation include volcanic and seismic effects. If basaltic intrusion formed barriers downstream from Yucca Mountain, those barriers could raise the elevation of the water table, much like a dam impounds water. Similarly, a dike that is formed upgradient from the repository could lower the water table at Yucca Mountain. The heat associated with a volcanic intrusion could also cause the water table to rise. However, the only potential volcanism near Yucca Mountain would be basaltic dikes that tend to form along faults in extensional tectonic areas and are typically less than two, but potentially up to four, meters in width (Carrigan, King, and Barr 1990). Thus, future intrusion would be expected to be small and to trend in a northerly direction, which coincides with the principal strike of faulting. Movement of groundwater at Yucca Mountain is parallel to the direction of possible dike formation, so an intrusive dike would be unlikely to form a barrier to flow. As described previously, Szymanski (1989) and Archambeau and Price (1991) have postulated seismic activity as a driving mechanism to raise the water table to the level of the potential repository. Carrigan, King, Barr, and Bixler (1991) and the National Research Council (1992) examined this mechanism and concluded that it would not raise the water table more than a few meters. In fact, the opposite effect is more common. Increased stream flow and decreased wellheads have been observed after moderate earthquakes (Rojstaczer and Wolf 1992; National Research Council 1992).

Davies and Archambeau (1997a, 1997b) developed a variation of the seismic pumping model. Their model posits that variations in the state of stress cause changes in water table elevation and that these changes may be as great as several hundred meters. The NWTRB asked an independent expert to review this hypothesis and other material. The reviewer concluded, "The interpretations depend significantly upon theoretical models that have never been tested or previously used and run counter to observations in nature and in the laboratory" (Rojstaczer 1998; see also Rojstaczer 1999). Based on this review, the Board concluded that the material reviewed: "...does not make a credible case for the assertion that there has been ongoing, intermittent hydrothermal activity at Yucca Mountain or that large earthquake-induced changes in the water table are likely at Yucca Mountain. This material does not significantly affect the conclusions of the 1992 NAS report" (NWTRB 1999a, p. 20).

4.3.3.1.5 Conclusion

Despite the suggestion that the water table could rise to the level of the potential repository, the preponderance of evidence shows that the water table has not risen to this level in the past and is not likely to do so in the future. Earthquakes are known to have an impact on the water table, but water table excursions are likely to be small and short-lived. For these reasons, tectonically induced fluctuations of the water table have not been included in the TSPA-SR. Future, wetter climates have the potential to cause the water table to rise to levels that were experienced during past, wetter climate periods. Multiple lines of evidence suggest that 120 m (390 ft) is a reliable estimate of the potential increase in water table elevation for a future, wetter climate (CRWMS M&O 2000c, Section 3.7.5.2). Paces and Whelan (2001) estimate that the maximum water table rise during the Pleistocene was only 17 to 30 m (56 to 98 ft). Future climate changes are not likely to result in flooding of the waste emplacement drifts due to a water table rise because the emplacement drifts would be located a distance above the water table ranging between approximately 210 m (690 ft) and 390 m (1,300 ft) (see Figure 1-13 in Section 1). At the northernmost end of the repository block, the design described in this report has portions of the layout outside the emplacement area that would be less than 120 m (390 ft) above the water table. The repository layout shown in Figure 1-13 includes a ramp at the northernmost areas that serves as an access to an observation drift located below the emplacement area (see Figure 2-38 in Section 2.3.1 for labeled repository components). Sections of that ramp and the east repository main north extension that it connects to are located such that portions of the excavations could be flooded if the water table rose as much as 120 m (390 ft). Water would not be expected to flow to the emplacement drifts. Future assessment of performance and design work will consider this situation, together with more specific details of the water table and its potential rise, and ensure that the design is developed to take advantage of opportunities to contribute positively to performance. Variation in the water table caused by climate is included within the nominal dose for the TSPA-SR.

4.3.3.2 Nuclear Criticality

This section describes the FEPs that could lead to a nuclear criticality event, together with the possible consequences. It begins with a discussion of the physics of nuclear criticality, followed by a brief summary of the methods used to evaluate the postclosure criticality potential of the waste forms that may be emplaced in the potential repository. These methods ensure that all the critical configurations that could result from the degradation of the waste forms are identified and the results quantified. Criticality control is an important objective of waste package design because a criticality event can result in an increased radionuclide inventory, potentially increasing the radiation dose to the receptor. For the 10,000-year period after disposal, nuclear criticality is screened out from the TSPA nominal case analysis based on its low probability. Nevertheless, because of significant interest in this topic, the DOE investigated the potential consequences of a criticality event. The results of the investigation indicate that a nuclear criticality would not have a significant impact on repository performance.

4.3.3.2.1 Physics of Nuclear Criticality

Nuclear fission is the splitting of an atomic nucleus by an impacting neutron. Only certain isotopes of very heavy elements can be split; the most commonly known of these are the uranium-235 and plutonium-239 isotopes. The nucleus of each element has a constant number of protons but may have a varying number of neutrons; these variations identify isotopes of an element.

The key factors associated with the potential fissioning of a nucleus are the energy or speed of the impacting neutron and the characteristics of the nucleus, which affect the manner in which it reacts to the neutron. The nuclear reaction most important for criticality occurs when a nucleus absorbs a free neutron, creating an unstable compound nucleus. The compound nucleus splits into fission fragments (lighter elements), releases energy in the form of gamma rays, and emits two or three additional neutrons in the process. The released neutrons are free to collide with other nuclei, which may absorb one of them and fission as well (Duderstadt and Hamilton 1976, Chapter 2).

There are other possible outcomes from a neutron interacting with a nucleus besides fissioning. The neutron may simply collide and be scattered, transferring kinetic energy to the nucleus in the process, or it may be absorbed by the nucleus without fissioning. The likelihood of any of these interactions depends heavily on the energy of the neutron. Neutrons emitted from fission have high energy. However, the likelihood of an interaction resulting in fission increases as the energy of the neutron decreases. As a result, fission is more likely if the energy or speed of a neutron is reduced.

Each time a neutron collides with a nucleus but is not absorbed, it loses energy, its speed decreases, and its potential for fissioning a nucleus increases. This slowing down of energetic neutrons is referred to as moderation; the material that provides the nuclei that cause this slowing down is called a "moderator." In general, the lighter the element, the better it works as a moderator because it scatters neutrons much better than it absorbs them. Hydrogen, the lightest element, is a very efficient moderator. Since water contains hydrogen nuclei, it also is a very effective moderator. The silicon in sand and rocks, such as those at the potential repository site, can also serve as a moderator, but silicon is a much less efficient moderator than water.

A nuclear chain reaction can be initiated once a neutron interacts with a fissile isotope, inducing it to fission. This releases additional neutrons that are, in turn, available to cause more fissions. Of the neutrons resulting from each fission, some will be absorbed in a nucleus without a resulting fission, some will escape from the system before colliding with a potentially absorbing nucleus, and some will be absorbed and cause fission. When each fission releases just enough neutrons to cause one additional fission, the effective neutron multiplication factor (keff) equals one, the chain reaction becomes self-sustaining, and the system is considered critical.

For a nuclear criticality to occur, there must be the proper combination of material and geometric configuration, known as the critical mass and critical geometry. Fissile material is essential for criticality. The only naturally occurring fissile isotope is uranium-235. However, it is not the only isotope that can support a critical reaction. Uranium-233 and plutonium-239 can also support criticality; these isotopes are produced in a reactor by fast neutrons that are absorbed by thorium-232 and uranium-238, respectively. In general, the greater the amount of fissile material, the more likely it is to have a criticality. The presence of neutron absorbers—materials that absorb neutrons without causing fission—is also important. The greater the amount of neutron absorber material, the less likely it is to have a criticality. Isotopes of boron and gadolinium are good neutron absorbers.

The geometrical arrangements of the fissile material and any neutron-absorbing material are of equal importance in determining whether an accumulation of a fissile material can achieve criticality. Because neutrons can leak out of a system without causing fission, the most efficient geometry for the occurrence of criticality is a sphere, which has the least surface area per unit volume of any shape.

With the presence of moderating materials like water, the likelihood of fission can be greatly increased, and a reduced amount of fissile material is necessary to form a critical mass. However, criticality is possible without the presence of a moderator. These situations are known as fast criticalities because they can occur in the presence of predominantly high-energy neutrons.

In addition to the release of energy and additional neutrons, the fission process effectively removes fissile nuclei from the system by splitting them into other, lighter elements. This is referred to as fuel depletion, or burnup. Also, many of the fission products—the lighter elements resulting from the fragmenting of nuclei—are strong neutron absorbers. Their existence in the area where the critical reaction is ongoing dampens the reaction. Thus, as commercial spent nuclear fuel is depleted, its ability to support a criticality is reduced from the combination of the removal of fissile material and the creation of materials that absorb neutrons through the fission process. This is the primary reason fuel is discharged from operating reactors after several years: it can no longer effectively support a critical reaction. Incorporating fuel depletion in the criticality evaluation of systems is called "burnup credit."

A repository would contain fissile material, mostly from commercial spent nuclear fuel in the form of a reduced amount of uranium-235 (compared to what was initially loaded into the reactor) and plutonium-239. Moderating materials, primarily consisting of silicon dioxide, would be present; some water may be present as well. As a result, a criticality event is possible: (1) inside one or more degraded waste packages; (2) in the rock surrounding the emplacement drifts, as a result of water transporting fissile materials from degraded waste packages; or (3) in the rock at some distance from the emplacement drifts as a result of an accumulation of fissile materials under favorable conditions in the nearby rock. An example of the last situation would be a deposit of organic material, which could accumulate dissolved fissile material from a flow of water that had originated in or previously passed through the repository area. The probability of a criticality depends on the probability of occurrence of three conditions: (1) a waste package breach, (2) the separation of the neutron absorber from the fissile material, and (3) the presence of sufficient amounts of a moderator. The spacing between fuel rods will affect the probability of criticality for certain waste forms.

If a critical event were to occur in a repository, the characteristics and consequences of the event may be considered as falling into two categories: steady-state and transient.

In a steady-state criticality, the reaction remains constant, with the keff equal to one. The magnitude of this type of criticality is limited by the fact that the energy released as a result of the criticality serves, in part, to evaporate surrounding water. Since the water is necessary for moderation of the neutrons, removing the water through boiling restrains the reaction and would, over time, make the system no longer critical.

One of the principal consequences of a steady-state criticality would be an increase in radionuclide inventory, particularly in fission products and actinides (i.e., elements heavier than uranium, formed primarily from the initial absorption of neutrons by uranium-238 for those waste forms which contain a significant fraction of this isotope). Another result would be the production of heat from the fissioning of nuclei, which would incrementally increase the heat generated from the ongoing decay of fission products already present in the waste. The additional heat could create localized hot spots in the repository, potentially affecting nearby geochemical and other conditions. These changes could marginally affect the amount and rate of release of radioactive materials from the repository. However, evaluations have shown that neither the inventory increase nor the thermal effects are significant enough to adversely impact repository performance (CRWMS M&O 1998m, Sections 9 and 10).

A natural analogue for steady-state criticality is the family of natural reactors that occurred in a sedimentary formation approximately 2 billion years ago at Oklo, a site in what is now Gabon, Africa. These natural reactors occurred in several closely linked, very rich uranium ore bodies that were saturated with water, which served as a moderator. Criticality at Oklo continued periodically for approximately one million years, pausing when the heat from the reaction evaporated the water and resuming when the site reflooded. Such a natural reactor was possible because, two billion years ago, the natural enrichment of uranium was greater than 3 percent (by weight) (enrichment is the percentage of uranium-235 by weight contained in the total uranium content). The radioactive decay process of uranium-235 has since dropped the worldwide natural uranium enrichment from greater than 3 percent at that time to its current value of 0.72 percent. The natural reactors at Oklo at that time period are unique; no other place in the world has been found where natural reactors existed (Smellie 1995). A reactor like the one that occurred at Oklo is not probable in a repository at Yucca Mountain because (1) the prevailing geochemical conditions for the Oklo deposit allow a higher concentration of uranium than any known deposit in the world, while the geochemistry and lithology characteristics of the Yucca Mountain site are not conducive to such concentrated accumulations of fissile materials, and (2) the effective enrichment (uranium-235 plus plutonium-239) of most commercial spent nuclear fuel that would be emplaced in the potential repository is less than the enrichment of the Oklo reactors during the time they were critical (Smellie 1995).

Transient criticalities could occur if there were a sudden change in the geometry or material composition of a disposed waste. A transient criticality would increase the rate of fission rapidly so that a very large burst of neutrons would be produced for a brief period; the accompanying increased rate of energy release could lead to a large increase in pressure and temperature. A transient criticality ends when the neutron multiplication factor decreases below unity, usually because of the loss of the moderating effects of water, which boils off during heightened temperatures. If some water were to return (recharge) after the criticality ended, it could increase neutron moderation, and the critical reaction could start again. However, the recharge of water would require a relatively long time compared with the duration of the criticality, so the total radionuclide increment from a transient criticality would be much smaller than the total for a steady-state criticality enduring over the same total elapsed period. The potential for damage to waste packages in the repository because of the peak temperature pulse from a transient criticality has been investigated, and evaluations (as presented in Section 4.3.3.2.3) indicate that such effects are not sufficient to significantly alter repository performance (CRWMS M&O 1999q).

A special subset of transient criticality is autocatalytic criticality, which is a transient criticality where the usual mechanisms that tend to shut down a criticality, such as the loss of moderator or loss of confinement, are delayed until a high fission rate is achieved. Although the probability of occurrence of an autocatalytic criticality at the potential repository is so low as to be considered not credible (Paperiello 1995), it is addressed for completeness and for evaluation of any hypothetical consequences. Scientists have pointed out that an autocatalytic criticality could occur when silica is the primary moderator (Bowman and Venneri 1996), but this would require extraordinarily unusual conditions that cannot occur in the waste package or at all outside the waste package unless (1) the entire fissile content of the waste package is spread uniformly in a nearly spherical shape or (2) fissile material from multiple waste packages accumulates in a small region of rock. Either configuration is essentially impossible to achieve. If such an event were to occur, it could create a much higher peak temperature and pressure than would result from transient criticality. If such an event were to result in a large and rapid release of kinetic energy, it could be considered an explosion. The possibility of an explosion as a result of an autocatalytic criticality in the potential repository has been examined by the scientific community and found not to be credible (Canavan et al. 1995; Kastenberg et al. 1996).

4.3.3.2.2 Methodology for Criticality Evaluation

The complete evaluation process for postclosure criticality has been documented in Disposal Criticality Analysis Methodology Topical Report (YMP 1998), which has been reviewed by the NRC. Some aspects of the methodology have already been accepted by the NRC (Reamer 2000), which is providing additional information and refinement to allow for the acceptance of remaining aspects of the methodology (YMP 2000c).

The postclosure criticality methodology is based primarily on a risk-informed approach. The objective of the risk-informed methodology is to provide assurance that all known waste forms and associated degradation products that can result in configurations that can support criticality are considered and evaluated based on probability of occurrence and potential consequences. The bases of this risk-informed methodology are the estimated probabilities of the potential configurations that could lead to a criticality and the evaluation of the consequences of those criticalities. This methodology was designed to, and can, encompass time frames much greater than 10,000 years. This methodology requires a redesign of any proposed design concept that may result in a critical event that has a frequency of occurrence of greater than 1 in 10,000 per year for the entire repository in any of the first 10,000 years. The methodology also identifies potentially critical configurations that have such a low probability of occurrence that they are deemed not credible.

Figure 4-170 provides a visual representation of the postclosure criticality methodology. The starting point is the establishment of the range of:

Some of these parameters can be specified deterministically, but some are most readily characterized probabilistically. Based on the established parameters, scenarios are developed encompassing how the emplaced material may degrade, which results in various degraded configurations. These configurations are grouped into classes reflecting a range of related parameters. Detailed degradation analyses are then performed on these configuration classes to identify the parameters that affect criticality potential: when the calculated keff of the configuration exceeds an established analytical value called the critical limit, the configuration is considered potentially critical (Section 3.5.2). These parameters include the respective quantities of fissile material, neutron absorber material, corrosion products, and moderator present in the configuration. Evaluations are then performed to determine the potential for criticality in the various proposed configuration classes (Section 3.5.2). Initially, the worst combination of parameters that define the range is used to determine whether that configuration class is potentially critical. If no combination of parameters can be constructed such that probability of exceeding the critical limit is below the screening value, then that configuration class is considered acceptable for inclusion in the repository. Configuration classes that show a potential for criticality are evaluated further to more precisely define which parametric combinations result in a keff that exceeds the critical limit criterion.

The results of these keff calculations are tabulated to cover the range of parameter values. These tables are then used to determine whether a specific combination of parameters within a configuration class can be considered potentially critical.

The probability of exceeding the critical limit is estimated for each configuration class by comparing the characteristics of the waste stream to the established parameter ranges. A probability criterion is established to identify when the probability of criticality is considered so high that a redesign of the waste package or engineered barrier system is needed to reduce the probability of criticality. The probability criterion requires that the overall frequency of criticality be less than 1 in 10,000 per year for the entire repository in any of the first 10,000 years. The criticality limit and probability criterion form design criteria for limiting the potential for criticality in the repository during postclosure (CRWMS M&O 2000au; CRWMS M&O 2000bg).

If any configurations are determined to be capable of supporting criticality events and have an estimated probability of occurrence below the probability criterion, but contribute to a total probability of criticality for the entire repository inventory above the 10 CFR 63.114(d) (66 FR 55732) screening probability threshold of 10-4 in 10,000 years, consequence analyses are performed. The probability tested against this screening threshold is the sum of the probabilities for all the scenarios that can lead to criticality for all waste forms in the repository.

The consequence analyses are primarily concerned with the increase in the radionuclide inventory resulting from a criticality, but they include other potential consequences, such as increased waste package degradation due to the increased heat generated from the criticality.

In addition to the establishment of probability criteria, the overall risk of potential critical events is determined. The probabilities and associated consequences of all configurations that would contribute to the probability of exceeding the screening probability threshold (1 chance in 10,000 in 10,000 years) are collected into an input set to perform an additional TSPA. This TSPA, which includes the total criticality risk, is performed to determine if the health and safety of the public would be maintained within the proposed regulatory limit. If necessary, additional design features for reducing the potential level of criticality would be implemented.

As Figure 4-170 shows and the preceding paragraphs describe, the postclosure disposal analysis criticality methodology includes both deterministic and probabilistic aspects. The evaluation of the various long-term processes, the combination of events, any potential criticality, and the consequences resulting from a potential criticality are all deterministic analyses. However, probabilistic analyses are performed to establish the likelihood of a criticality occurrence.

The methodology as described in this section pertains to commercial and DOE spent nuclear fuel, DOE high-level radioactive waste, and immobilized plutonium. Criticality analyses for naval spent nuclear fuel employ a separate methodology that is primarily deterministic (YMP 1998, Section 1.2) and show that intact naval spent nuclear fuel cannot go critical for any credible configuration within the repository.

4.3.3.2.3 Criticality Results: Probability and Consequences

The methodology described in the previous section has been applied to postulated configurations that could result from the degradation of commercial spent nuclear fuel waste packages and their contents (CRWMS M&O 2000fh), as well as from several of the possible DOE spent nuclear fuel types (CRWMS M&O 1999j, 2000bh, 2000bi). The results, which demonstrate a very low probability of criticality and minimal consequences, are summarized below. These results apply to commercial spent nuclear fuel, except as otherwise noted. The 21-PWR waste package was chosen for these calculations because it is the design for fuel with the highest reactivity (CRWMS M&O 2000bg). Other waste forms with lower reactivity planned for repository disposal are expected to have a lower probability of criticality.

Probability of Criticality within 10,000 Years Following Repository Closure—The postulated earliest time to breaches in the waste package barrier that could allow criticality exceeds 10,000 years because of the limited water available in the natural system to contact the waste packages, the use of an extremely corrosion-resistant material for the waste package outer barrier, and a titanium drip shield that covers the tops of the waste packages and is designed to divert water away from the waste packages. Even with postulated early breaches, the initial conditions required for criticality would have a very low probability. The probability of a critical event, internal or external to the waste package, is expected to be less than 1 chance in 10,000 within 10,000 years (CRWMS M&O 2000fh, Section 6). The main possible exception to this pre-10,000-year conclusion is in the event of igneous intrusion. In that case, the combination of low probability for breach by igneous intrusion and the other conditions required for criticality (filling with water moderator and separating fissile material from neutron absorber) have been shown, for commercial spent nuclear fuel, to be well below the screening probability threshold (CRWMS M&O 2000fh, Section 6). Because of the combination of low probability for breach by igneous intrusion and other conditions required for criticality, other waste forms are expected to show similar results.

Probability of Internal Criticality—As mentioned previously, the probability of a waste package breach and subsequent loss of neutron absorber will increase with time after 10,000 years. Figure 4-171 depicts the different stages of internal degradation for a typical 21-PWR Absorber Plate waste package (YMP 1998, Appendix C). The final stages of degradation, shown in this figure, illustrate the collapse of the assemblies, which reduces the probability of criticality because of the reduced volume between fuel rods available for the moderator to fill. Another factor tending to reduce the probability of criticality with time is the eventual breach of the bottom of the waste package, which can drain most of the moderating water. Evaluations have shown that the potential for criticality of commercial spent nuclear fuel is maximized when the internal basket is fully degraded, but the assemblies remain intact (CRWMS M&O 1997d) and there is no breach of the bottom of the waste package. Using the TSPA-SR waste package degradation models, the probability of a critical event within the total inventory of the 21-PWR Absorber Plate waste packages is calculated to be 2 math symbol, multiply 10-7 in 10,000 years (CRWMS M&O 2000fh, Section 6).

Probability of Internal Criticality for Codisposed U.S. Department of Energy Spent Nuclear Fuel and High-Level Radioactive Waste—Evaluations have been performed of the criticality potential of waste packages that would contain high-level radioactive waste glass and certain types of codisposed DOE spent nuclear fuel (CRWMS M&O 1996c, 1999j, 2000bh, 2000bi). These evaluations have generally shown the probability of criticality to be less than that for commercial spent nuclear fuel, which is very small to begin with. The primary reasons are the lower fissile loading per waste package for many of the DOE fuel types and the greater flexibility to install insoluble neutron absorber. This flexibility arises because of the smaller fuel mass per waste package.

Probability of Criticality for the Immobilized Plutonium Waste Form—The range of possible degradation scenarios for this waste form has been examined to identify potentially critical configurations. It was found that the design of the immobilized plutonium waste form itself provides several layers of defense in depth, so that criticality is virtually impossible (CRWMS M&O 2000ba). The degradation rate of the ceramic waste form is so slow that in the unlikely event that the waste package is breached and filled by a continuous dripping of water, it will be nearly 50,000 years after emplacement before enough of this waste form has degraded to permit any significant separation of the uranium and plutonium from the gadolinium and hafnium neutron absorbers. Even after degradation of the waste form, the gadolinium and hafnium are generally less soluble than the fissile material, so they cannot be transported out of the waste package while the fissile material remains. Even with the additional conservative assumption of extremely unlikely chemistry conditions that would make the gadolinium sufficiently soluble to be removed before the fissile material, enough of the completely insoluble hafnium would remain to prevent criticality (CRWMS M&O 2000ba).

Probability of External Criticality—The probability for an external criticality to occur before 10,000 years is based on the igneous intrusion scenario. The probability for igneous intrusion is 1.6 math symbol, multiply 10-4   in 10,000 years, which is just above the criticality-screening threshold of 1 math symbol, multiply 10-4 in 10,000 years. However, subsequent processes to transport and accumulate the fissile material have a very small probability, so the combined probability of criticality before 10,000 years from igneous intrusion is less than the screening threshold (CRWMS M&O 2000fh, Section 6.2). Evaluations have shown that the probability, not considering igneous intrusion, of an external criticality even in the rock beneath the waste packages containing immobilized plutonium waste forms is less than 4 math symbol, multiply 10-4 (McClure and Alsaed 2001, Section 8.4) and occurs beyond 10,000 years. This is primarily a result of (1) limited dripping water available to transport enough fissile mass out of the waste package and into a geometry favorable for criticality; (2) limited fracture intensity and lithophysae frequency to allow for fissile material accumulation in a geometry favorable for criticality; (3) low concentration of fissile material in the water trickling out of a breached waste package due to low waste form solubility; and (4) limited fresh water needed to dilute the effluent water and neutralize and precipitate the fissile minerals (McClure and Alsaed 2001, Section 8.4).

Steady-State Criticality Consequences—If a steady-state criticality were to occur, it would be very unlikely to have a power level greater than 5 kW because the power is limited by the evaporation of the water moderator, which increases with power level or temperature. The duration is likely to be limited to 10,000 years, which is the average period of a climate cycle that might have a high enough rainfall or drip rate to sustain the required level of water moderation against evaporation (YMP 1998, Appendix C, Section 5.1). For a typical commercial spent nuclear fuel waste package, a continuous steady-state criticality would result in an increase of the radionuclide inventory in that waste package by less than 30 percent at approximately 10,000 years (YMP 1998, Appendix C, Section 5.1). The return of a moist climate cycle, upwards of 10,000 years after the shutdown, would be very unlikely because continued degradation of the waste package would have removed some conditions necessary for criticality (e.g., intact waste package bottom that supports water ponding, or optimum spacing between fuel rods). When this moderate increase is spread over all the waste packages likely to be degraded and not critical, the increment in radionuclide inventory is likely to be much less than the 10 percent stated in the criticality consequence criterion in Section 4.3.3.2.2. The temperature increase during steady-state criticality will be too small to measurably enhance waste form or waste package degradation. Steady-state criticalities would not operate above boiling temperatures because such temperatures would boil away the water moderator and shut down the criticality (YMP 1998, Appendix C, Section 5.1), and no increased degradation rates have been observed in iron-based materials at 150°C (302°F) (CRWMS M&O 1996d, Volume III, p. 8-4). As a result, the incremental impact of steady-state criticality events on the total inventory for the repository is expected to be insignificant.

Transient Criticality Consequences—In the unlikely event that a transient criticality were to occur, a worst-case rapid initiating event would provide a uniform ramp of positive reactivity insertion lasting 90 seconds and producing a peak power level of 10 MW for less than 60 seconds (CRWMS M&O 1999q