5.1 Inventory for Performance Calculations
DOE proposes to dispose of approximately 11,000 to 17,000 waste packages containing as much as 70,000 MTHM of spent nuclear fuel and high-level radioactive waste in a repository at Yucca Mountain. There are several different types of disposal containers for commercial spent nuclear fuel and different container designs for DOE spent nuclear fuel and high-level radioactive waste. The exact number of waste packages would depend on various options in the proposed design. This long-term consequence analysis identified the inventory by the source category of waste material to be disposed of (commercial spent nuclear fuel, DOE spent nuclear fuel, weapons-usable plutonium, and high-level radioactive waste). For purposes of modeling, the inventory for each of the categories was averaged into an appropriate number of packages, each with identical contents. The average of the modeled inventories resulted in a total of nearly 12,000 idealized waste packages (slightly higher than the actual number of waste packages that would be emplaced) in two basic types.Note that while the simulations were based on the nearly 12,000 packages, there would be no difference in the result if the smaller packages had been used (17,000 package case). This is because the use of smaller packages is merely a way of implementing the lower-temperature operating mode and would contain the same proposed inventory.
5.1.1 Inventory of Radioactive Materials
The number of radionuclides to be analyzed was determined by a screening analysis. The screening analysis identified those radionuclides that would collectively contribute at least 95 percent of the dose to a person living in the vicinity of Yucca Mountain. The list of radionuclides resulting from the screening process forms the basis for the analyses discussed in this chapter. Appendix I, Section I.3, contains more details of this screening analysis.
The total inventory was abstracted into two types of idealized waste packages: a codisposal package with high-level radioactive waste in a glass matrix and DOE spent nuclear fuel, and a commercial spent nuclear fuel package. Table 5-2 lists the abstracted inventory for the idealized waste packages. For analysis purposes, naval spent nuclear fuel is conservatively modeled as commercial spent nuclear fuel (DIRS 152059-BSC 2001, all; DIRS 153849-DOE 2001, Section 4.2.6.3.9, p. 4-257).[IDEALIZED WASTE PACKAGES]
5.1.2 Inventory of Chemically Toxic Material
A screening analysis reported in Appendix I (Section I.6.1) showed that the only chemical materials of concern for the 10,000-year analysis period were those released as the external wall of the waste package and the waste package support pallet materials corroded. The chemicals of concern would be chromium, nickel, molybdenum, and vanadium. The exposed surface areas that would corrode include Alloy-22 surfaces (drip shield rails, outer layer of waste packages, and portions of the emplacement pallets) and stainless steel 316NG surfaces (portions of the emplacement pallets).
The total quantities of materials would be 86,000,000 kilograms (190,000,000 pounds) of Alloy-22 (DIRS 150558-CRWMS M&O 2000, p. 6-6) containing 22.5 percent chromium, 14.5 percent molybdenum, 57.2 percent nickel, 0.35 percent vanadium (DIRS 104328-ASTM 1998, all) and 140,000,000 kilograms (310,000,000 pounds) of stainless steel, (DIRS 150558-CRWMS M&O 2000, p. 6-6) which is 17 percent chromium, 12 percent nickel and 2.5 percent molybdenum. A large percentage of the stainless steel would be inside the waste package (as an inner sleeve) and, therefore, much of this material would not be exposed until the Alloy-22 had corroded away.
5.2 System Overview
Thus, the access to, and flow of, contaminated water are the most important considerations in determining potential health hazards.
5.2.1 Components of the Natural System
The water table is the boundary between the unsaturated zone above and the saturated zone below. In the subsurface region above the water table, the rock contains water but the water does not fill all the open spaces in the rock. Because the open spaces are only partially filled, this region is called the unsaturated zone. Water in the unsaturated zone tends to move generally downward in response to capillary action and gravity. In contrast, water fills all the open spaces in the rock below the water table, so this region is called the saturated zone. Water in the saturated zone tends to flow laterally from higher to lower pressures. Both zones contain several different rock types, as shown in Figure 5-2. The layers of major rock types in the unsaturated zone at the Yucca Mountain site are the Tiva Canyon welded, Paintbrush nonwelded, Topopah Spring welded, Calico Hills nonwelded, and Crater Flat undifferentiated tuffs. Figure 5-2 shows two of the faults at the proposed site-the Ghost Dance fault that occurs within the repository block and the Solitario Canyon Fault that forms the western boundary of the repository block. Faults are slip zones where rock units have become displaced either vertically, laterally, or diagonally, resulting in the rock layers being discontinuous. These slip zones tend to form a thin plane in which there is more open space that acts as a channel for water. Some faults tend to fill with broken rock formed as they slip, so they take on a very different flow property from that of the surrounding rock. The proposed repository would be in the Topopah Spring welded tuff in the unsaturated zone, at least 200 meters (660 feet) below the surface and at least 160 meters (530 feet) above the water table (DIRS 154554-BSC 2001, pp. 28-29). [HYDROGEOLOGIC TERMS]
When it rains in the Yucca Mountain vicinity, most of the water runs off and a very limited amount infiltrates the rock on the surface of the mountain. Some of the water that remains on the surface or infiltrates the rock evaporates back into the atmosphere (directly or through plant uptake and evapotranspiration). The very small amount of water that infiltrates the rock and does not evaporate percolates down through the mountain to the saturated zone (DIRS 155950-BSC 2001, Section 3.3.2.1, p. 3-17). Water that flowed through the unsaturated zone into the proposed repository could dissolve some of the waste material, if there was a breach in the package containment, and could carry it through the groundwater system to the accessible environment, where exposure to humans could occur.
5.2.2 Components of the Waste Package and Drip Shield
5.2.3 Visualization of the Repository System for Analysis of Long-Term Performance
The first step in the visualization of the system is the development of a listing of all the possible features, events, and processes that could apply to the behavior of the system. An example of a feature is the existence of a fault, an example of an event is a seismic event (earthquake), and an example of a process is the gradual degradation of the waste package wall by general corrosion. The list is then screened using various types of analyses to determine what features, events, and processes should be included in the modeling. The chosen features, events, and processes are then assembled into scenarios, which are descriptions of how features, events, and processes link together to result in a certain outcome (see Appendix I, Section I.2.1, for further detail).
The elements of the TSPA model are organized into the following categories, which are generally related to various parts of the system:
The individual models associated with these elements are discussed in Appendix I, Sections I.2.2 through I.2.9 [I.2.2, I.2.3, I.2.4, I.2.5, I.2.6, I.2.7, I.2.8, I.2.9].
In addition, the following special scenarios are also discussed in Appendix I, Sections I.2.10 through I.2.13 [I.2.10, I.2.11, I.2.12, I.2.13]:
During the development of the TSPA model, DOE often had to make assumptions. The main reason for assumptions was to account for situations where there was limited data. With additional data, it may be possible to present a more "realistic" representation, usually as a statistical distribution. If data are limited, it is necessary to make assumptions and use associated conservative data values. The Nuclear Regulatory Commission and Environmental Protection Agency rulemaking processes acknowledged that uncertainty about physical processes acting over the large space and time scales of interest will remain, even after many years of site characterization. The long-term analysis does not seek an exact prediction but rather seeks to establish a representative projection. The list of assumptions is too large to include here. Table 5-3 is an index to a series of tables that describe in detail the assumptions in the model and associated key attributes. The detailed information is in the Total System Performance Assessment-Site Recommendation document (DIRS 153246-CRWMS M&O 2000, pp. F-2 to F-9).
5.2.4 Uncertainty
DOE recognizes that uncertainties exist from the onset of an analysis; however, forecasts are valuable in the decisionmaking process because they provide insight based on the best information and scientific judgments available. The following section discusses uncertainties in the context of possible effects on the impact estimates reported in this chapter. The discussion is divided to address:
5.2.4.1 Uncertainty Associated with Societal Changes, Climate, and Other Long-Term Phenomena
Estimates of future climatic conditions are based on what is known about the past, with consideration given to climate impacts caused by human activities. Calcite in Devils Hole, a fissure in the ground approximately 40 kilometers (25 miles) southeast of Yucca Mountain, provides the best-dated record of climate changes over the past 500,000 years. The record shows continual variation, often with very rapid jumps, between cold glacial climates (for the Great Basin, these are called pluvial periods) and warm interglacial climates similar to the present. Fluctuations average 100,000 years in length (DIRS 153038-CRWMS M&O 2000, Section 6.4.1). The past climate cycles were idealized into a regular cycle of pulses, which were repeated throughout the period of the forecast. This method inherently assumes that the future will repeat the past. However, while current understanding of the causes of climate change allows some confidence in this approach, a considerable amount of conservatism was built into the models to account for possible climate uncertainties. For example, a large range of water fluxes was used to reflect the wide rainfall variations that could occur over thousands and hundreds of thousands of years (DIRS 155950-BSC 2001, Section 3.3.2.6). The analysis assumed that the current climate is the driest it will ever be at Yucca Mountain.
5.2.4.2 Uncertainty Associated with Currently Unavailable Data
As further discussed in Section 5.2.4.3, the use of parameter distributions and studies of alternative models can provide understanding of how the lack of data affects the range of the impact results. Furthermore, sensitivity studies (see Section 5.2.4.3.4) can also provide insight into the importance of particular parameters. The sensitivity studies sometimes identify data with a small contribution to the results, thereby mitigating concerns arising from their unavailability.
The fact that some data are currently unavailable does not necessarily preclude providing adequate assessment of long-term impacts. When the Draft EIS and the Supplement to the Draft EIS were prepared there was sufficient information to provide an adequate analysis of the long-term performance impacts.
However, additional data have been generated since the publication of those documents. These data have helped improve characterization of the range of impacts in this Final EIS over those reported in the Draft EIS. Some examples of the additional data and their uses are the following:
5.2.4.3 Uncertainty Associated with Models and Model Parameters
5.2.4.3.1 Variability Versus Uncertainty
Uncertainty and variability are sometimes related. The exact nature of the variability in a natural system cannot be known because all parts of the system cannot be observed. For example, DOE cannot dig up all the rock in Yucca Mountain and determine that the positioning of the rock layers is exactly as suggested by core sample data. Therefore, there is uncertainty about the properties of the rock at specific locations in the mountain because properties change with distance and it is not known how much they change at any given location. If the variability can be appropriately quantified or measured, a model usually can be developed to include this variability. If the variability cannot be physically quantified or estimated, it should be treated as uncertainty (lack of knowledge). However, the ability to model some types of spatial variability can be limited not only by lack of data but also by the capacity of a computer to complete calculations (for example, if one simulation took weeks or months to complete). In these instances, variability must be simplified in such a way as to be conservative (that is, the simulation would overestimate the impact).
Two basic tools were used in the analysis to deal with uncertainty and variability: alternative conceptual models and probability theory. Alternative conceptual models were used to handle uncertainty in the understanding of a key physical-chemical process controlling system behavior. Probability theory was used to understand the impacts of uncertainty in specific model parameters (that is, would results change if the parameter value was different). In particular, uncertain processes often required different conceptual models. For example, different conceptual models of how water in fractures communicates with water in the smaller pores or the matrix of the rock in the unsaturated zone lead to different flow and transport models. Sometimes conceptual models are not mutually exclusive (for example, both matrix and fracture flow might occur), and sometimes they do not exhaustively cover all possibilities (apparently matrix and fracture flow do cover all possibilities). These examples indicate that the use of alternative conceptual models, while often necessary to characterize some types of uncertainty, is not always as exact as desired.
A process of weighting alternative conceptual models (as described below) was used in the long-term consequence analysis to account for uncertainties in conceptual models. The Monte Carlo sampling technique was used for handling uncertainty in specific model parameters and for alternative conceptual models that were weighted beforehand with specific probabilities. The method involves random sampling of ranges of likely values, or distributions, for all uncertain input parameters. Distributions describe the probability of a particular value in the range. A common type of distribution is the familiar "bell-shaped" curve, also known as the normal distribution. Parameters in the consequence analysis are described by many different types of distributions appropriate for how the values and their probabilities are understood. Numerous realizations of the repository system behavior were calculated, each based on one set of samples of all the inputs. Each total system realization had an associated probability so that there is some perspective on the likelihood of that set of circumstances occurring. The Monte Carlo method yields a range for any chosen performance measure (for example, peak annual individual dose in a given period at a given location) along with a probability for each value in the range. In other words, it gives an estimate of repository performance and determines the possible errors based on the estimate. In this chapter, the impact estimates are expressed as the mean of all the realizations and the 95th-percentile value (that is, the value for which 95 percent of the results were smaller). [CALCULATING THE MEAN AND 95TH-PERCENTILE RESULTS]
5.2.4.3.2 Weighting of Alternative Conceptual Models
There were two possible approaches to incorporating discrete alternative models in the performance analysis: weighting all models into one comprehensive Monte Carlo simulation (lumping), or keeping the discrete models separate and performing multiple Monte Carlo simulations for each discrete model (splitting). The main results in Section 5.4 were developed using the splitting approach because they were based on a limited range of uncertainty. Based on expert judgment (and to some extent the finite time and resources that could be applied to the analysis effort), the analysis used a best estimate of the more likely ranges of model behavior and parameter ranges. Some alternative models were not included in the analysis, and some parameter ranges of the included models were narrowed. Because of this narrowed range of models and parameters, the results are conditional, meaning that they depend on certain models and parameters being held constant or having their variance restricted. One such condition is the specific design of the repository and the waste packages in the design evaluated in this EIS. Another important condition is that the cladding on the spent nuclear fuel can be depended on as a barrier. Other conditional results were used to characterize the effect of certain assumptions. For example, splitting was done to consider such events as human intrusion (Section 5.7.1), igneous activity (Section 5.7.2), and criticality (Section 5.8). The consequences of these types of events are not part of results given in Section 5.4; rather they are reported as added impacts with certain probabilities of occurrence.
5.2.4.3.3 Uncertainty and the Proposed Action
5.2.4.3.4 Uncertainty and Sensitivity
Many of the important uncertain parameters were examined in alternative models. The alternative models either expand the range of the parameters beyond the expected range of uncertainty or change the weighting of the parameter distribution. For example, this type of analysis was performed for alternative models of seepage (DIRS 101779-DOE 1998, Volume 3, pp. 5-1 to 5-9) and cladding degradation (DIRS 101779-DOE 1998, Volume 3, pp. 5-32 to 5-35). An example of alternative model studies for volcanic hazards is discussed in DIRS 155950-BSC (2001, Section 14.3.1, p. 14-6).
System performance could be sensitive to repository design options, but models and parameters for these various options do not have an assigned uncertainty. Therefore, although they can be important, they do not show up as key parameters based on an uncertainty analysis. The determination of the parameters or components that are most important depends on the particular performance measure being used. This point was demonstrated in the 1993 TSPA (DIRS 100111-CRWMS M&O 1994, all; DIRS 100191-Wilson et al. 1994, all) and the Total System Performance Assessment-1995 (DIRS 100198-CRWMS M&O 1995, all). For example, these two analyses showed that the important parameters would be different for 10,000-year peak doses than for 1-million-year peak doses.
There are several techniques for analyzing uncertainties, including the use of qualitative scatter plots where the results (for example, annual individual dose) are plotted against the input parameters and visually inspected for trends. In addition, performance measures can be plotted against various subsystem outputs or surrogate performance measures (for example, waste package lifetime) to determine if that subsystem or performance surrogate would be important to performance. There are several formal mathematical techniques for analyzing the sets of realizations from a Monte Carlo analysis to extract information about the effects of parameters. Such an analysis determined the principal factors affecting the performance of the repository design.
5.2.4.3.5 Uncertainty Analysis for the TSPA-Site Recommendation
Despite the extensive scientific studies described in the Science and Engineering Report, DOE has always recognized that uncertainties will remain in any assessment of the performance of a repository over thousands of years, as discussed in that report (DIRS 153849-DOE 2001, Sections 1.5, 4.1, and 4.4). These uncertainties are attributable to a variety of causes, ranging from uncertainty regarding the fundamental processes that could affect radionuclide migration to uncertainty related to the design and operation of the repository. For this reason, one part of the DOE approach to dealing with uncertainty relies on multiple lines of evidence that can contribute to the understanding of the performance of the potential repository. Another part of the DOE approach is a commitment to continued testing, monitoring, and analysis beyond the possible recommendation of the site.
The TSPA-Site Recommendation model incorporated a number of uncertainties. These were uncertainties for which a realistic distribution of parameters is not identified, but rather a very conservative bounding value or bounding range was chosen. Additional studies have investigated effects of unquantified uncertainties and sensitivities in the TSPA model by better quantification of uncertainties and the affected processes. This research is documented in the Supplemental Science and Performance Analysis (DIRS 155950-BSC 2001, all). (See Appendix I, Section I.2 for more detailed discussion of the evolution of the TSPA model and application to this EIS.) A summary of areas in which the Supplemental Science and Performance Analysis model benefited from these additional uncertainty studies is provided below. The Supplemental Science and Performance Analysis (DIRS 155950-BSC 2001, all) contains full details of the studies.
Unquantified Uncertainty Analysis
5.2.4.3.6 Key Parameters and Uncertainty
Two types of analysis were used: stepwise linear rank regression and classification tree [in which parameters were classified in terms of the separation of outcomes into "high"-dose (top 10th-percentile) and "low"-dose (bottom 10th-percentile) categories] (DIRS 155934-Mishra 2001, all; DIRS 155936-Mon 2001, all).
Regression Analysis
Classification Tree Analysis
Results
A description of the important parameters identified in Table 5-5 follows:
The parameters in Table 5-5 that most affect the total uncertainty in the TSPA model are factors that would govern the degradation of the waste package for the first 250,000 years following repository closure. After 250,000 years, most waste packages would have failed and other factors become important. Even at 500,000 years after repository closure, waste package degradation is still important. At later times the important parameters would be related to factors that influenced the flow of water in the drifts, especially infiltration and episodic flow.
5.3 Locations for Impact Estimates
Figure 5-3 is a map with arrows showing the general direction of groundwater movement from Yucca Mountain. Shading indicates major areas of groundwater discharge through a combination of springs and evapotranspiration by plants. The general path of water that infiltrates through Yucca Mountain is south toward Amargosa Valley, into and through the area around Death Valley Junction in the lower Amargosa Desert. Natural discharge of groundwater from beneath Yucca Mountain probably occurs farther south at Franklin Lake Playa (DIRS 100376-Czarnecki 1990, pp. 1 to 12), and spring discharge in Death Valley is a possibility (DIRS 100131-D'Agnese et al. 1997, pp. 64 and 69).
Although groundwater from the Yucca Mountain vicinity flows under and to the west of Ash Meadows in the volcanic tuff or alluvial aquifers, the surface discharge areas at Ash Meadows and Devils Hole (see map in Figure 5-3 for locations) are fed from the carbonate aquifer. While these two aquifers are connected at some locations, the carbonate aquifer has a hydraulic head that is higher than that of the volcanic or alluvial aquifers. Because of this pressure difference, water from the volcanic aquifer does not flow into the carbonate aquifer; rather, the reverse occurs. Therefore, contamination from Yucca Mountain is not likely to mix with the carbonate waters and discharge to the surface at Ash Meadows or Devils Hole (DIRS 104983-CRWMS M&O 1999, all) under current conditions. This pressure difference could change under future climate conditions.
Because, under expected conditions, there would be no contamination of this discharge water, there would be no human health impacts. Furthermore, there would be no consequences to the endangered Ash Meadows Amargosa pupfish (Cyprinodon nevadensis mionectes) or Devils Hole pupfish (Cyprinodon diabolis) at those locations.
Figure 3-25 in Chapter 3 shows the projected population of 76,000 residents within 80 kilometers (50 miles) of Yucca Mountain in 2035. This map provides the information used to estimate population doses from radionuclides released to the atmosphere from the repository. The atmospheric analysis in Section 5.5 used the 80-kilometer (50-mile) population distribution described in Section 3.1.8.
In the Draft EIS, impacts were evaluated at 5-kilometer (3-mile), 20-kilometer (12-mile), and 30-kilometer (19-mile) distances from the repository as well as at the groundwater discharge point. The EPA regulation, 40 CFR 197.12, establishes a controlled area around the repository that must not extend farther south than 36 degrees, 40 minutes, 13.6661 north latitude, in the predominant direction of groundwater flow. For this EIS, DOE assumed the controlled area boundary to be the farthest point south. The predominant groundwater flow crosses this boundary approximately 18 kilometers (11 miles) from the repository. Therefore, the 5-kilometer (3-mile) distance would be inside the controlled area, would no longer be part of the accessible environment, and DOE did not evaluate impacts at this distance.
5.4 Waterborne Radiological Consequences
While the RMEI is a regulatory definition for a specific location, impacts to individuals at two additional locations were evaluated using the lifestyle characteristics of the RMEI.
The analysis converted the annual committed effective dose equivalent, referred to as the annual individual dose, to the probability of contracting a fatal cancer (referred to as a latent cancer fatality) due to exposure to radioactive materials in the water. In addition, the analysis calculated population doses in person-rem for two different periods: for the 70-year lifetime at the time of the peak dose during the first 10,000 years after repository closure, and integrated over the first 10,000 years after repository closure. The analysis also converted the population dose to the expected number of latent cancer fatalities in the population. DOE based the analysis on the radionuclide inventories discussed in Section 5.1. However, the analysis included the entire carbon-14 inventory of the commercial spent nuclear fuel as a solid in the groundwater release models. Actually, 2 percent of the carbon-14 exists as a gas in the fuel (see Section 5.5). Thus, the groundwater models slightly overestimate (by 2 percent) the potential impacts from carbon-14.
The analysis studied potential consequences to individuals at three impact locations arising from waste mobilization and waterborne transport. A set of 300 model simulations were run using the GoldSim model (DIRS 155182-BSC 2001, all) for the RMEI location [about 18 kilometers (11 miles) from Yucca Mountain]. Each simulation used separate sets of sampled uncertainty parameters and generated an annual individual-dose profile for the 1 million years following repository closure. This set of simulations for the RMEI location, and some additional groundwater simulations (DIRS 154659-BSC 2001, Enclosure 3) provided the basis for calculating doses at 30 kilometers (19 miles) from the repository and at the discharge location near Franklin Lake Playa.
5.4.1 Extension of Groundwater Impacts to Other Distances
Appendix I, Section I.4.5, details the development of distance scale factors using a three-dimensional analytical advection and dispersion transport model. Scaling factors were developed based on two criteria: attenuation of the peak concentrations in the plume and general increase in the cross-sectional area of the plume (that is, reduction of the average plume concentration). Two sets of factors were developed based on a large source size (characteristic of the repository footprint) and a small source size [10 meters by 10 meters (33 feet by 33 feet)]. The scaling factors were used to estimate peak of the mean and peak of the 95th-percentile annual individual doses and the groundwater concentrations at the two additional distances reported in Sections 5.4.2.1 and 5.4.2.2.
For the 10,000-year period of the nominal scenario, the dose would be attributable to the failure of a few waste packages. In this case, scaling factors based on a small size source were used. For the 1-million-year period, the release would be attributable to general releases over the whole repository area, so large source size scale factors were used. The factors based on the cross-section of the plume were chosen for the estimates. This was appropriate because the effect of water usage by the communities would be to cause significant mixing, and the more characteristic parameter would be the plume average concentration. Appendix I, Section I.4.5, includes scale factors for both approaches for comparison.
5.4.2 Waterborne Radiological Results
5.4.2.1 Waterborne Radiological Results for the Higher-Temperature Repository Operating Mode
Table 5-7 lists the population consequences associated with the peak annual individual dose listed in Table 5-6. The population size was based on the projected population numbers for 2035 in Figure 3-25 in Chapter 3 of this EIS. For these calculations, the analysis assumed that no contaminated groundwater would reach populations in any regions to the north of Yucca Mountain. Therefore, populations in the sectors north of the due east and due west sectors in Figure 3-25 were not considered to be exposed.
Thus, approximately 74,000 people would be exposed to contaminated groundwater. This stylized population dose analysis assumed that people would continue to live in the locations being used at present. This assumption is consistent with the recommendation made by the National Academy of Sciences (DIRS 100018-National Research Council 1995, all) because it is impossible to make accurate predictions of lifestyles and residence locations far into the future.
The values in Table 5-7 include a scaling factor for water use. The performance assessment transport model calculated the annual individual dose assuming the radionuclides dissolved in water that flowed through the unsaturated zone of Yucca Mountain would mix in an average of 2.4 million cubic meters (1,940 acre feet) (DIRS 155950-BSC 2001, p. 13-42) per year in the saturated zone aquifer. This compares to an annual water use in the Amargosa Valley of about 17.1 million cubic meters (13,900 acre-feet)
(DIRS 155950-BSC 2001, p. 13-42). The analysis diluted the concentration of the nuclides in the 2.4 million cubic meters of water throughout the 17.1 million cubic meters of water before calculating the population dose.
The small consequences listed in Tables 5-6 and 5-7 would result from the durability of the waste packages; most of which would remain intact significantly longer than 10,000 years. The outer layer of the waste package would be subject to a very low average corrosion rate, but there is a high degree of uncertainty in the value of that average corrosion rate. Model simulations incorporated a small number of waste package failures within 10,000 years due to manufacturing defects; the dose results in Tables 5-6 and 5-7 during this period would result directly from these early failures.
The radionuclides that would contribute the most to individual dose in 10,000 years would be technetium-99, carbon-14 dissolved in groundwater, and iodine-129. For example, the mean consequence at 18 kilometers (11 miles) has technetium-99 contributing 77 percent of the total annual individual dose rate, carbon-14 contributing 16 percent, and iodine-129 contributing 7 percent. While the atmospheric analysis in this EIS assumed that 2 percent of the carbon-14 migrated as gas in the form of carbon dioxide (see Section 5.5 for more details), the groundwater modeling for this waterborne radiological consequences analysis conservatively assumed that all of the carbon-14 migrated in the groundwater.
Table 5-8 lists impacts for the post-10,000-year period. The table lists the mean and 95th-percentile peak annual individual dose and the times of the associated peaks at three locations. The mean and 95th-percentile annual individual doses during 1 million years following repository closure are shown in Figure 5-4. The multiple peaks occurring 200,000 years or more after repository closure are driven by transitions between climate states.
The simulations were ended after 1 million years largely because further radioactive decay would continue to decrease the annual individual dose even for very long-lived radionuclides. The peak annual individual dose usually coincided with the occurrence of a wetter climate period.
The radionuclides that would contribute the most to the peak annual individual dose in 1 million years would be neptunium-237 and plutonium-242. The mean peak annual individual dose at the RMEI location would have neptunium-237 contributing 61 percent of the total annual individual dose, plutonium-242 contributing 13 percent, actinium-227 contributing 5 percent, thorium-229 and uranium-234 each contributing 3 percent, and uranium-233, lead-210, and radium-226 each contributing 2 percent. The plutonium isotopes contributing to dose would be due to colloidal transport of plutonium, not transport of plutonium as a dissolved element in groundwater.
With respect to the groundwater protection standards in 40 CFR 197.30, both the mean and 95th-percentile estimated levels during the 10,000-year regulatory period would be hundreds of thousands of times less than the regulatory limits (see Table 5-9).
5.4.2.2 Waterborne Radiological Results for the Lower-Temperature Repository Operating Mode
Table 5-11 lists the population consequences associated with the peak annual individual dose listed in Table 5-10. The population size was based on the population numbers projected for the year 2035 in Figure 3-25 in Chapter 3 of this EIS. For these calculations, the analysis assumed that no contaminated groundwater would reach populations in any regions to the north of Yucca Mountain. Therefore, populations in the sectors north of the due east and due west sectors in Figure 3-25 were not considered to be exposed.
Thus, approximately 74,000 people would be exposed to contaminated groundwater. This stylized population dose analysis assumed that people would continue to live in the locations being used at present. This assumption is consistent with the recommendation made by the National Academy of Sciences (DIRS 100018-National Research Council 1995, all) because it is impossible to make accurate predictions of lifestyles and residence locations far into the future.
The values in Table 5-11 include a scaling factor for water use. The performance assessment transport model calculated the annual individual dose assuming the radionuclides dissolved in water that flowed through the unsaturated zone of Yucca Mountain would mix in an average of 2.4 million cubic meters (1,940 acre-feet) (DIRS 155950-BSC 2001, p. 13-42) per year in the saturated zone aquifer. This compares to an annual water use in the Amargosa Valley of about 17.1 million cubic meters (13,900 acre-feet) (DIRS 155950-BSC 2001, p. 13-42). The analysis diluted the concentration of the nuclides in the 2.4 million cubic meters of water throughout the 17.1 million cubic meters of water before calculating the population dose.
The small consequences listed in Tables 5-10 and 5-11 would result from the durability of the waste packages; most of which would remain intact significantly longer than 10,000 years. The outer layer of the waste package would be subject to a very low average corrosion rate, but there is a high degree of uncertainty in the value of that average corrosion rate. Model simulations incorporated a small number of waste package failures within 10,000 years due to manufacturing defects; the dose results in Table 5-10 and 5-11 during this period would result directly from these early failures.
The radionuclides that would contribute the most to individual dose in 10,000 years would be technetium-99, carbon-14 dissolved in groundwater, and iodine-129. For example, the mean consequence at 18 kilometers (11 miles) has technetium-99 contributing 63 percent of the total individual dose rate, carbon-14 contributing 25 percent, and iodine-129 contributing 10 percent. While the atmospheric analysis in this EIS assumed that 2 percent of the carbon-14 migrated as gas in the form of carbon dioxide (see Section 5.5 for more details), the groundwater modeling for this waterborne radiological consequences analysis conservatively assumed that all of the carbon-14 migrated in the groundwater.
Table 5-12 lists impacts for the post-10,000-year period as peak annual doses. The table lists the mean and 95th-percentile peak annual individual dose and the times of the associated peaks at three locations. The mean and 95th-percentile annual individual doses during 1 million years following repository closure are shown in Figure 5-5. The multiple peaks occurring 200,000 years or more after repository closure are driven by transitions between climate states.
The simulations were ended after 1 million years largely because further radioactive decay would continue to decrease annual individual dose even for very long-lived radionuclides. The peak annual individual dose usually coincided with the occurrence of a wetter climate period.
The radionuclides that would contribute the most to the peak annual individual dose in 1 million years would be neptunium-237 and plutonium-242. The mean peak dose at 18 kilometers (11 miles) would have neptunium-237 contributing 63 percent of the total individual dose rate, plutonium-242 contributing 12 percent, actinium-227 contributing 5 percent, thorium-229 and uranium-234 each contributing 3 percent, and uranium-233, lead-210, and radium-226 each contributing 2 percent. The plutonium isotopes contributing to dose would be due to colloidal transport of plutonium, not transport of plutonium as a dissolved element in groundwater.
With respect to the groundwater protection standards in 40 CFR 197.30, both the mean and 95th-percentile estimated levels during the 10,000-year regulatory period would be hundreds of thousands of times less than the regulatory limits (see Table 5-13).
5.4.2.3 Alternative Dosimetry Methods
5.5 Atmospheric Radiological Consequences
In addition, DOE considered the possible impacts from the release of radon from the repository. Radon is a decay product of uranium and would be generated for as long as any uranium remained in the repository. Based on gas flow studies, DOE believes that radon would decay before it reached the ground surface. Appendix I, Section I.7.3, contains a more detailed screening discussion.
5.5.1 Source Term
5.5.2 Atmospheric Consequences to the Local Population
5.5.3 Atmospheric Consequences to an Individual
5.6 Consequences from Chemically Toxic Materials
Appendix I, Section I.3 discusses the inventory of chemically toxic materials that would be emplaced in the repository under the Proposed Action by element. Based on this inventory, a screening analysis (described in Appendix I, Section I.6.1) identified which of the chemically toxic materials could pose a potential risk to human health. Chromium, molybdenum, nickel, and vanadium were identified as posing such a potential risk, and these elements were further evaluated in a bounding consequence analysis, as described in Appendix I, Section I.6.2. This analysis makes the conservative assumption that all chromium dissolves in hexavalent form.
It should also be noted that all of the chromium, molybdenum, nickel, and vanadium considered are elements contained in the metals used to package the waste and support the packages. None of the materials inside the waste packages were considered because, except for about three packages, all packages would last for more than 50,000 years.
Table 5-14 summarizes the results of the bounding analysis. In some cases a Maximum Contaminant Level or Maximum Contaminant Level Goal was available for comparison to the calculated concentration. In other cases, only an Oral Reference Dose was available. The Oral Reference Dose can be compared to the intake that would result for a 70-kilogram (154-pound) person drinking 2 liters (0.53 gallon) of water per day.
The bounding consequence analysis estimated that the maximum peak concentration of chromium in groundwater used at exposure locations would be 0.01 milligram per liter. There are two measures for comparing human health effects for chromium. When the Environmental Protection Agency established its Maximum Contaminant Level Goals, it considered safe levels of contaminants in drinking water and the ability to achieve these levels with the best available technology. The Maximum Contaminant Level Goal for chromium is 0.1 milligram per liter (40 CFR 141.51). The bounding concentration is well below the Maximum Contaminant Level Goal for chromium (about one-tenth of this limit). The other measure for comparison is the reference dose factor for chromium, which is an intake of 0.0004 milligram of chromium per kilogram of body mass per day (DIRS 148224-EPA 1999, all). The reference dose factor represents a level of intake that has no adverse effect on humans. It can be converted to a threshold concentration level for drinking water. The conversion yields essentially the same concentration for the reference dose factor as the Maximum Contaminant Level Goal. At present, the bounding estimate of groundwater concentration of hexavalent chromium cannot be expressed in terms of human health effects (for example, latent cancer fatalities). The carcinogenicity of hexavalent chromium by the oral route of exposure has not been determined because of a lack of sufficient epidemiological or toxicological data (DIRS 148224-EPA 1999, all; DIRS 101825-EPA 1998, p. 48).
The estimated bounding concentration of molybdenum in groundwater used at exposure locations would be 0.009 milligram per liter. There is no Maximum Contaminant Level Goal for molybdenum but intake can be compared to the Oral Reference Dose. The intake rate from drinking 2 liters (0.53 gallon) per day of contaminated water by a 70-kilogram (154-pound) person would be 0.0003 milligram per kilogram per day. This is well below the Oral Reference Dose of 0.005 milligram per kilogram per day (DIRS 148228-EPA 1999, all).
The estimated bounding concentration of nickel in groundwater used at exposure locations would be 0.04 milligram per kilogram. There is no Maximum Contaminant Level Goal available for nickel but intake can be compared against the Oral Reference Dose. The intake rate from drinking 2 liters (0.53 gallon) per day of contaminated water by a 70-kilogram (154-pound) person would be 0.001 milligram per kilogram per day. This is well below the Oral Reference Dose of 0.02 milligram per kilogram per day.
The estimated bounding concentration of vanadium in groundwater used at exposure locations would be 0.0002 milligram per liter. There is no Maximum Contaminant Level Goal available for vanadium, but intake can be compared to the Oral Reference Dose. The intake rate from drinking 2 liters (0.53 gallon) per day of contaminated water by a 70-kilogram (154-pound) person would be 0.000006 milligram per kilogram per day. This is well below the Oral Reference Dose of 0.007 milligram per kilogram per day.
Because the estimated bounding concentrations of chromium, molybdenum, nickel and vanadium in well water would be below the Maximum Contaminant Level Goal or yield intakes well below the Oral Reference Dose, there is no further need to refine the calculation to account for physical processes that would limit mobilization of those materials or delay and dilute them during transport in the geosphere.
5.7 Consequences from Disruptive Events
5.7.1 Human Intrusion Scenario
The human intrusion scenario analyzed in this EIS is consistent with the requirements of 40 CFR Part 197. The stylized human intrusion scenario is summarized as follows:
The human intrusion results were calculated probabilistically, analogous to the nominal scenario calculations for waterborne radioactive material releases. The calculations were carried out for the higher-temperature repository operating mode. For this stylized intrusion scenario, there would be no difference for the lower-temperature operating mode because exactly one waste package is intersected for both operating modes and its inventory is moved to the saturated zone where further transport does not depend on repository operating mode. Figure 5-6 shows the mean and 95th-percentile annual individual dose for 1 million years resulting from a human intrusion 30,000 years after repository closure for the set of 300 simulations. The values in Figure 5-6 represent the dose from a single waste package, and are not combined with releases for other waste packages that would fail due to other processes. The peak of the mean annual individual dose from human intrusion would be 0.002 millirem, occurring a short time after 100,000 years after repository closure. These results indicate that the repository would be sufficiently robust and resilient to limit releases caused by human intrusion to values well below the 15-millirem annual individual dose standard.
The analysis did not combine the results of the disruptive igneous event scenario with the results of the human intrusion scenario. However, combined results can be approximated by adding the results of the human intrusion analysis to that of the disruptive igneous event scenario, which would result in a total combined maximum dose. Based on the results presented in this section and Section 5.7.2, the highest mean annual individual dose that would result from an intrusion would be less than one-tenth of the radiological dose from a disruptive igneous event.
A sensitivity study where the human intrusion occurs at 100 years after repository closure has also been conducted (DIRS 157307-BSC 2001, Enclosure 1).
5.7.2 Igneous Activity Scenario
5.7.2.1 Volcanic Eruption Events
Model development included the selection of conservative assumptions about the event, selection of input parameter distributions characterizing important physical properties of the system, and use of a computational model to calculate entrainment of waste in the erupting ash. Each intrusive event (a swarm of one or more dikes) was assumed to generate one or more volcanoes somewhere along its length, but eruptions would not need to occur within the repository footprint. Approximately 77 percent of intrusive events that intersected the repository would be associated with one or more surface eruptions within the repository footprint. The number of eruptive conduits (volcanoes) is independent of the number of dikes in a swarm. Characteristics of the eruption such as eruptive power, style (violent versus normal), velocity, duration, column height, and total volume of erupted material, are included in the analysis.
5.7.2.2 Groundwater Transport of Radionuclides Following Igneous Intrusion Event
The analysis assumed that waste packages close to the point of intrusion would be so damaged that they would provide no further protection for the waste. Actual conditions would be uncertain, and damage probably would range from moderate to extensive. Nominal models for radionuclide mobilization and transport were used even though conditions would change in the drift following intrusion. All waste in the most severely damaged packages would be immediately available for transport in the unsaturated zone, depending on solubility limits and the availability of water, which was determined using the seepage model for nominal performance. The thermal, chemical, and mechanical effects of the intrusion on the drift environment were neglected. No credit was taken for water diversion by the remnants of the drip shield or waste package, and cladding was assumed to be fully degraded. Actual thermal, chemical, hydrological, and mechanical conditions in the drift following igneous intrusion are unknown, although conservatively assuming that the engineered barriers would have completely failed is sufficient to compensate for the uncertainty associated with conditions in the drift.
5.7.2.3 Results for Igneous Activity Scenario
The average doses from igneous activity calculated in this manner incorporate uncertainties regarding the time at which the igneous event could occur, and account for the reality that, as time passed, the likelihood would increase that igneous disruptions could have already occurred. For example, a person living downwind from Yucca Mountain 10,000 years after repository closure would have a mean probability of 1.6
Figure 5-7 shows the mean probability-weighted dose histories representing possible doses to an individual for the higher-temperature repository operating mode. The figure also shows the nominal scenario for comparison. The igneous activity scenario is only simulated to 100,000 years because the nominal scenario impacts dominate after that time. These summary curves are based on 5,000 individual dose histories calculated using different sets of uncertain input parameters in the model. For approximately the first 20,000 years, the dose history is a smooth curve that is dominated by the effects of volcanic eruption. The probability-weighted mean annual individual dose during this period would reach a peak of approximately 0.1 millirem about 300 years after repository closure, and then decline because of radioactive decay of the relatively shorter-lived radionuclides that contributed to doses from the ash fall exposure pathway. The major contributors to the eruptive dose would be americium-241, plutonium-238, plutonium-239, and plutonium-240. Strontium-90 would be a significant contributor at extremely early times, but would drop off rapidly because of radioactive decay (half-life of 29.1 years). Inhalation of resuspended particles in the ash layer would be the primary exposure pathway during this period, and the smooth decline of the mean dose curve from approximately 300 to 2,000 years would result from decay of americium-241 (half-life of 432 years). From approximately 20,000 years after closure, the mean igneous dose would be dominated by groundwater releases from packages damaged by igneous intrusions that did not erupt to the surface. The irregular shape of the curve from this point forward is in part a result of the groundwater transport processes, and in part reflects the occurrence of intrusive events at random times, rather than the prescribed intervals used for extrusive simulations. The intrusive event could occur at any time, and the first appearance of groundwater doses in the mean curve at approximately 20,000 years reflects retardation during transport, rather than the absence of intrusions at earlier times. Results for the lower-temperature operating mode would be essentially identical to those for the higher-temperature mode because the probability of an igneous intrusion interacting with waste packages is reduced for the wider waste package spacing. However, the overall probability of an igneous intrusion intersecting the potential repository would increase because of a larger repository emplacement area.
The dose history for the igneous activity scenario in Figure 5-7 is presented as a probability-weighted annual dose resulting from events occurring at uncertain times throughout the period of simulation. This approach to calculating and displaying the probability-weighted annual doses is consistent with the approach specified by 40 CFR Part 197 and is required for determination of the overall expected annual dose. However, displays of the probability-weighted annual dose do not allow direct interpretation of the conditional annual dose, which is the annual dose an individual would receive if a volcanic event occurred at a specified time. For conditional analyses, the probability of the event is set equal to one, and the time of the event is specified. Conditional results do not provide a meaningful estimate of the overall risk associated with igneous activity at Yucca Mountain, but they provide insights into the magnitude of possible consequences for specific sets of assumptions. A sensitivity calculation was performed to provide results for this conditional case (DIRS 154659-BSC 2001, pp. 3-47 to 3-48). Conditional mean annual dose histories were calculated for eruptive events at 100, 500, 1,000, and 5,000 years. The conditional mean dose in the first year after an eruptive event at 100 years after repository closure is approximately 13 rem. The conditional dose in the first year after an eruption decreases to approximately one-half this level for an eruption 500 years after closure, and is approximately 10 percent of this value for an eruption 5,000 years after closure. This calculation was made with a previous TSPA model (DIRS 153246-CRWMS M&O 2000, all) that has some differences from the model used elsewhere in this EIS for long-term performance (DIRS 157307-BSC 2001, Enclosure 1). The differences that affect the analysis described above are that dose factors were revised to conform to 40 CFR Part 197 and the distance analyzed is 20 kilometers rather than 18 kilometers from the repository. These changes would be expected to increase the dose values at 100 years and 500 years by a factor of between 2 and 3. The results at the later times would increase by about 20 percent.
5.8 Nuclear Criticality
5.8.1 Probability of Internal Criticality for Commercial Spent Nuclear Fuel
5.8.2 Probability of Internal Criticality for Codisposed DOE Spent Nuclear Fuel and High-level Radioactive Waste
5.8.3 Probability of Criticality for the Immobilized Plutonium Waste Form
5.8.4 Probability of External Criticality
5.8.5 Effect of a Steady-state Criticality on Radionuclide Inventory
5.8.6 Transient Criticality Consequences
5.8.7 Autocatalytic Criticality
5.8.8 Disruptive Natural Events Influencing Criticality
The potential adverse criticality considerations of igneous intrusion into the repository include: (1) the possibility of immediate waste package breach, (2) the separation of a significant fraction of the fissile material from the neutron absorber by magma transport, and (3) the accumulation of a critical mass of fissile material from, or within, the transporting magma. The potential for criticality following igneous intrusion has been evaluated for commercial spent nuclear fuel under extremely conservative assumptions, and no sufficiently probable mechanism for accumulating a critical mass has been identified.
5.9 Consequences to Biological Resources and Soils
After closure, heat from the radioactive decay of the waste could cause temperatures in the rock near the waste packages to rise above the boiling point of water at this altitude [96°C (205°F)] (DIRS 101779-DOE 1998, Volume 3, p. 3-36). The period the subsurface temperature could remain above the boiling point would vary from a few hundred years to a few thousand years, depending on the operating mode. Conduction and the flow of heated air and water through the rock (advection) would carry the heat from the waste packages through the rock to the surface and to the aquifer.
Although the atmosphere would remove excess heat when it reached the ground surface, the temperature of near-surface soils probably would increase slightly. Predicted increases in surface soil temperatures range from approximately 10°C (18°F) at the bedrock-soil interface (DIRS 100627-Bodvarsson and Bandurraga 1996, p. 510) to 6°C (10.8°F) for dry soil at a depth of 2 meters (6.6 feet) (Table 5-15). To address soil heterogeneity (differences in depth and water content), a recent study (DIRS 103618-CRWMS M&O 1999, all) modeled soil temperature increases at various depths under wet (saturated) and dry (no water at all) soil conditions for the high thermal load. They predicted that temperatures of near-surface soils would be unlikely to rise more than a few degrees (Table 5-14) but would increase with depth from the surface. Surface soil temperatures would start to increase approximately 200 years after repository closure and would peak more than 1,000 years after repository closure. Later, the temperature would gradually decline and would approximate prerepository conditions after 10,000 years (DIRS 103618-CRWMS M&O 1999, Figure 30 and p. 41).
The maximum change in temperature would occur directly above the repository, affecting approximately 5 square kilometers (1,250 acres) under the higher-temperature operating mode. The effects of repository heat on the surface soil temperatures would gradually decline with distance from the repository (DIRS 103618-CRWMS M&O 1999, p. 43). Although not modeled, the increase in surface soil temperature would be lower under the lower-temperature operating mode, and the area that could be affected would be larger [as much as 6.2 square kilometers (1,550 acres) above the repository for the lower-temperature operating mode].
There is considerable uncertainty in the estimates of soil temperature increases due to uncertainties in the thermal properties of the soil at Yucca Mountain, particularly thermal conductivity (the amount of heat that can be conducted through a unit of soil per unit time) (DIRS 103618-CRWMS M&O 1999, p. 50). The predicted temperature increase for dry soil provides a conservative estimate of the temperature increase that could occur because even partially saturated soil has a much greater thermal conductivity than dry soil. Soil moisture content recorded at a depth of 15 centimeters (6 inches) was as low as 3 percent on some study sites during some months, but the soil was never completely dry (DIRS 105031-CRWMS M&O 1999, p. 14).
A depth of 1 meter (3.3 feet) is within the root zone for many desert shrubs. A temperature increase of 3°C (5.4°F) could affect root growth and other soil parameters such as the growth of microbes or nutrient availability. Studies at Yucca Mountain (DIRS 105031-CRWMS M&O 1999, pp. 11 to 46) show that due to natural variations some plant species experienced a spatial range in soil temperatures of 4°C (7.2°F) at a depth of 0.45 meter (18 inches), which is comparable to the 0.5-meter (20-inch) depth used by DIRS 103618-CRWMS M&O (1999, pp. 37-41). Impacts to biological resources probably would consist of an increase of heat-tolerant species over the repository and a decrease of less tolerant species. In general, areas affected by repository heating could experience a loss of shrub species and an increase in annual species. A gradual (over 1,000 years) temperature increase of the magnitude predicted (DIRS 103618-CRWMS M&O 1999, all) probably would have less effect on the plant community than a more rapid change.
The predicted increase in temperature would extend as far as 500 meters (1,600 feet) beyond the edge of the repository, with the greatest increase in temperature occurring in soils directly above the repository. A shift in the plant species composition, if any, would be limited to the area within 500 meters of the repository footprint [that is, as much as 8 square kilometers (2,000 acres)].
A shift in the plant community probably would lead to localized changes in the animal community that depends on it for food and shelter. Specific plant and animal species and community changes cannot be predicted with certainty because changes in climate or seasonal episodic events (droughts, high rainfall) can substantially change species responses to single factors. However, the variation in surface soil temperatures at Yucca Mountain that are caused by elevation, slope, aspect, and other natural attributes suggest that soil temperature increases of the magnitude predicted (DIRS 103618-CRWMS M&O 1999, pp. 44 to 48) are probably within the adaptive range of some plant species now at Yucca Mountain (DIRS 105031-CWRMS M&O 1999, pp. 11 to 46).
Some reptiles, including the desert tortoise, exhibit temperature-dependent sex determination (DIRS 103463-Spotila et al. 1994, all). Nest temperatures have a direct effect on sex determination, with low temperatures resulting in predominately male hatchlings and high temperatures resulting in predominately females. Although existing experimental data do not adequately represent the large fluctuations in nest temperatures in natural settings, an increase in soil temperature due to repository operations could influence the sex ratio and other aspects of the life history of the desert tortoise population residing over the repository footprint. However, depth to the top eggs of 23 nests at Yucca Mountain during 1994 averaged 11 centimeters (4.3 inches). Predicted temperature increases of clutches at that depth based on modeling results (DIRS 103618-CRWMS M&O 1999, pp. 37 to 42) would be less than 0.5°C (0.9°F). Given the ranges of critical temperatures reported by DIRS 103463-Spotila et al. (1994, all), an increase of this magnitude would be unlikely to cause adverse effects.
Changes in plant nutrient uptake, growth, and species composition, as a result of increases in soil temperature over long periods of time, could influence vegetation community dynamics and possibly alter desert tortoise habitat structure in areas immediately above the repository. However, little is known about the effects that minor alterations in habitat would have on desert tortoise population dynamics.
As discussed in Sections 5.4 and 5.6, in the distant future water at certain discharge points would be likely to carry concentrations of radionuclides and chemically toxic substances. DOE did not quantify impacts to biological resources from irrigation water extracted at the RMEI location, from irrigation water extracted at 30 kilometers (19 miles) downgradient from the potential repository, or for the evaporation of water at Franklin Lake Playa (where there is no surface water at present). The estimated doses to humans exposed to this water would be very small. Expected dose rates to plants and animals would be much less than 100 millirad per day. The International Atomic Energy Agency concluded that chronic dose rates less than 100 millirad per day are unlikely to cause measurable detrimental effects in populations of the more radiosensitive species in terrestrial ecosystems (DIRS 103277-IAEA 1992, p. 53).
The desert tortoise is the only threatened or endangered species in the analyzed repository land withdrawal area (DIRS 104593-CRWMS M&O 1999, p. 3-14). Desert tortoises are rare or absent on or around playas (DIRS 101914-Rautenstrauch and O'Farrell 1998, pp. 407 to 411; DIRS 103160-Bury and Germano 1994, pp. 64 and 65); therefore, DOE anticipates no impacts to this species from contaminated water resources at Franklin Lake Playa in the future.
Impacts to surface soils would be possible. Changes in the plant community as a result of the presence of the repository could lead to an increase in the amount of rainfall runoff and, therefore, an increase in the erosion of surface soils, thereby increasing the sediment load in ephemeral surface water in the immediate Yucca Mountain vicinity.
5.10 Summary
Tables 5-6 and 5-10 list individual doses from groundwater releases of radionuclides during 10,000 years after repository closure. The mean annual individual doses at the RMEI location are summarized in Table 5-16. The mean annual individual doses in Table 5-16 are much less than the limit of 15 millirem in 40 CFR Part 197.
Tables 5-7 and 5-11 list estimated lifetime and 10,000-year integrated radiation dose impacts for members of the affected population from the groundwater release pathway during the first 10,000 years after repository closure. Table 5-17 summarizes the health effects for the affected population of 74,000 persons based on a 10,000-year integrated basis.
The average mortality rate for cancer deaths per 100,000 persons in Nevada is 202 (DIRS 153066-Murphy 2000, p. 83). Using the Nevada cancer death rate, about 154 cancer fatalities would normally occur each year in the population affected by groundwater potentially contaminated by a repository at Yucca Mountain (74,000 persons). All of the values in Table 5-17 are much smaller than 1, meaning that it is most likely than no person would die due to groundwater contamination by radiological material in the 10,000-year period after repository closure. This comparison clearly indicates that human health impacts associated with effects on groundwater from the Proposed Action would be very small for the affected population. Using the Nevada cancer death rate, about 140 cancer fatalities would normally occur each year in the population within an 80-kilometer radius of Yucca Mountain (assuming a population of about 76,000 persons). All of the values in Table 5-17 are much smaller than 1.0, meaning that it is most likely that no person would die due to groundwater contamination by radiological material in the 10,000-year period after repository closure. This comparison clearly indicates that human health impacts associated with the Proposed Action would be very small for the population in general.
The analysis indicates (as listed in Table 5-17 and the peak dose values) that there is no significant difference in impacts due to the operating mode, even though the impacts for the higher-temperature mode appear to be slightly larger than those impacts for the lower-temperature mode. One reason for the similarity in annual individual dose between the operating modes is that most waste packages would still be intact beyond the time at which the repository temperature would be elevated much above ambient rock temperatures (DIRS 155950-BSC 2001, p. 7-85). Thus, most radionuclides would not be released until long after the thermal effects had subsided and, therefore, the operating modes would not have a large effect on the peak doses.
The EPA has set annual dose limits of 15 millirem to an individual for human intrusion and igneous disruption events (40 CFR Part 197). As shown in Figure 5-7, the peak of the mean annual dose rate from a human intrusion 30,000 years after repository closure would be 0.002 millirem. The probability weighted mean annual dose to an individual for the igneous intrusion scenario would have a peak of 0.1 millirem. Both of these results are well below the regulatory limits.
The peak mean annual individual doses at the RMEI location in the first 1 million years after repository closure would be 150 millirem for the higher-temperature operating mode and 120 millirem for the lower-temperature operating mode. These doses do not specifically include the effects of disruptive events. The effects of disruptive events would be very small compared to the 1-million-year peak annual dose. These effects are evaluated separately and reported in Section 5.7.
As listed in Table 5-14, human impacts from chemically toxic materials would be unlikely because water concentrations would be below Maximum Contaminant Level Goals (40 CFR 191.51) or Oral Reference Doses (chromium, DIRS 148224-EPA 1999, all; molybdenum, DIRS 148228-EPA 1999, all; nickel, DIRS 148229-EPA 1999, all; and vanadium, DIRS 103705-EPA 1997, all). Estimated concentrations of radionuclides in groundwater (see Table 5-9) would be hundreds of thousands of times less than regulatory limits (40 CFR 197.30). Atmospheric release of carbon-14 would yield an estimated 80-kilometer (50-mile) population impact of 5.3
Part of the work described in the Supplemental Science and Performance Analysis (DIRS 155950-BSC 2001, all) included analysis of unquantified uncertainties. Table 5-4 summarizes the elements of the model that DOE studied and indicates whether or not revised model elements were included in the Supplemental Science and Performance Analysis model. The Supplemental Science and Performance Analysis model, with additional modifications, was used for the long-term performance analysis for this Final EIS (see Appendix I, Sections I.2 and I.4). The first column of Table 5-4 lists the major process models and a reference to the appropriate section in the Science and Engineering Report (DIRS 153849-DOE 2001, all). The second column lists the individual model elements analyzed in the unquantified uncertainties report. The third column lists sections of Volume 1 of the Supplemental Science and Performance Analysis report (DIRS 155950-BSC 2001, all) that contain additional details on the analysis. The analyses included sensitivity studies or other analysis methods to determine how significant the uncertainty might be. If warranted and possible, changes were made to the Supplemental Science and Performance Analysis model to better characterize the uncertainties; this is noted in the fourth column.
Regression analysis is a tool for quantifying the strength of input-output relationships in the TSPA model. To this end, a stepwise linear rank regression model is fitted between individual dose at a given time (or some other performance measure) and all randomly sampled input variables. Parameters are ranked on the basis of how much their exclusion would degrade the explanatory power of the regression model. The importance ranking measure used for this purpose is the uncertainty importance factor, which is defined as the loss in explanatory power divided by the coefficient of determination of the regression model. The uncertainty importance factor quantifies the proportion of the total spread (variance) in total dose explained by the regression model that can be attributed to the variable of interest.
Classification tree analysis, a subset of classification and regression tree analysis, is a method for determining variables or interactions of variables that drive output into particular categories. Classification and regression tree analyses can be used to generate decision rules that determine whether a particular realization would produce "high" or "low" dose depending on the values of the most important variables. Unlike regression analysis, which is based on the total range of model outcomes, classification tree analysis focuses on extreme values of model results and tries to relate them to specific ranges of values for the important variables.
For different time frames in the analysis, different parameters emerge as important to the overall variability of the results (DIRS 155934-Mishra 2001, all and DIRS 155936-Mon 2001, all). Table 5-5 lists the results of the analysis.
10-9 person-rem per microcurie per year of release. For a 3.3-microcurie-per-year release, this corresponds to a maximum 80-kilometer annual population dose of 1.5 10-8 person-rem. This dose corresponds to 7.5 10-12 latent cancer fatality in the regional population of 76,000 persons during each year at the maximum carbon-14 release rate. This annual population radiological dose corresponds to a 70-year lifetime radiological population dose of 1.1 10-6 person-rem, which corresponds to 5.3 10-10 latent cancer fatality during the 70-year period of the maximum release. 10-14 rem per microcurie per year of release was calculated using the GENII code for this location. For a 3.3-microcurie-per-year maximum release rate, the individual maximum radiological dose rate would be 1.8 10-13 rem per year, corresponding to a 9.2 10-17 probability of a latent cancer fatality. The 70-year lifetime dose would be 1.3 10-11 rem, representing a 6.4 10-15 probability of a latent cancer fatality.
10-8), and the probability of more than one igneous disruption occurring during the next 100,000 years is far below the level of concern. Therefore, the analysis considered only a single igneous eruption within the repository during the next 100,000 years, occurring with a mean 100,000-year probability of 1.6 10-3. The year in which that eruption could occur is uncertain; therefore, the igneous eruption scenario was evaluated as if it were many different eruptive scenarios, each occurring in a different 25-year time interval, and each occurring with a probability 25 times that of the annual probability. The average dose resulting from igneous disruption was determined by calculating doses resulting from igneous events in each 25-year period, multiplying by the probability (mean 25-year probability of 4.0 10-7), and adding the doses from each disruptive event. For computational efficiency, igneous intrusions that would not result in a surface eruption were simulated using a simpler approach in which the time of intrusion was sampled randomly from the 100,000-year period, and the probability associated with each simulation is the full 100,000-year probability of 1.6 10-3. Probability-weighted doses from both eruptive and intrusive events were added together to give the total dose from igneous disruption. 10-4 of receiving a radiation dose from soil contaminated by an igneous event sometime in the past. The probability-weighted average dose emphasizes the overall risk to a person living downwind from Yucca Mountain, in terms of both the likelihood and consequences of the igneous activity scenario. 10-7 in 10,000 years (after closure of the repository). The 21-PWR Absorber Plate waste package was chosen for criticality calculations because it is the design for fuel with the highest reactivity and thus would be expected to have the highest probability of criticality. 10-7 for commercial spent nuclear fuel. The primary reasons are the lower fissile loading per waste package and the greater flexibility to install neutron absorber due to smaller fuel mass per waste package. 10-12 in 10,000 years following repository closure. This low probability is primarily a result of the following Yucca Mountain characteristics: (1) limited dripping water to transport enough fissile material out of the waste package and into a geometry favorable for criticality; (2) a limited number of regions in the rock below the drifts to allow for fissile material accumulation in a geometry favorable for criticality; (3) a low concentration of fissile material in the water exiting out of a breached waste package due to low waste form solubility; and (4) lack of a chemical means to accumulate fissile materials and lack of a reducing environment to encourage precipitation. 10-10 latent cancer fatality during the 70-year period of maximum release, much lower than the groundwater-borne population impacts. Finally, as discussed in Section 5.9, there are no anticipated adverse impacts to biological resources from either repository heating effects or the migration of radioactive materials.